Merge tag 'soc' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc

Pull arm soc-specific updates from Arnd Bergmann:
 "This is stuff that does not fit well into another category and in
  particular is not related to a particular board.  The largest part in
  here is extending the am33xx support in the omap platform."

Fix up trivial conflicts in arch/arm/mach-{imx/mach-mx35_3ds.c, tegra/Makefile}

* tag 'soc' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (74 commits)
  ARM: LPC32xx: Add PWM support
  ARM: LPC32xx: Add PWM clock
  ARM: LPC32xx: Set system serial based on cpu unique id
  ARM: vexpress: Config option for early printk console
  ARM: vexpress: Add Device Tree for V2P-CA15_CA7 core tile
  ARM: vexpress: Convert V2P-CA15 Device Tree to 64 bit addresses
  ARM: vexpress: Add fixed regulator for SMSC
  ARM: vexpress: Add missing SP804 interrupt in motherboard's DTS files
  ARM: vexpress: Initial common clock support
  ARM: SAMSUNG: Introduce Kconfig variable for Samsung custom clk API
  ARM: EXYNOS: Add missing static storage class specifier in pmu.c file
  ARM: EXYNOS: Make combiner_init function static
  ARM: EXYNOS: Update HSOTG PHY clock setting for EXYNOS4X12
  ARM: versatile: Make plat-versatile clock optional
  ARM: vexpress: Check master site in daughterboard's sysctl operations
  ARM: vexpress: remove automatic errata workaround selection
  ARM: LPC32xx: Adjust to pl08x DMA interface changes
  ARM: EXYNOS: Clear SYS_WDTRESET bit to use watchdog reset
  ARM: imx: fix mx51 ehci setup errors
  ARM: imx: make ehci power/oc polarities configurable
  ...

4 files changed, 193 insertions, 26 deletions

diff --git a/arch/arm/mach-tegra/Makefile b/arch/arm/mach-tegra/Makefile
index 35253fd..90aae34 100644
--- a/arch/arm/mach-tegra/Makefile
+++ b/arch/arm/mach-tegra/Makefile
@@ -9,6 +9,7 @@ obj-y					+= fuse.o
 obj-y					+= pmc.o
 obj-y					+= flowctrl.o
 obj-y					+= powergate.o
+obj-y					+= apbio.o
 obj-$(CONFIG_CPU_IDLE)			+= cpuidle.o
 obj-$(CONFIG_CPU_IDLE)			+= sleep.o
 obj-$(CONFIG_ARCH_TEGRA_2x_SOC)         += tegra2_clocks.o
@@ -18,7 +19,7 @@ obj-$(CONFIG_ARCH_TEGRA_3x_SOC)		+= tegra30_clocks.o
 obj-$(CONFIG_SMP)			+= platsmp.o headsmp.o
 obj-$(CONFIG_SMP)                       += reset.o
 obj-$(CONFIG_HOTPLUG_CPU)               += hotplug.o
-obj-$(CONFIG_TEGRA_SYSTEM_DMA)		+= dma.o apbio.o
+obj-$(CONFIG_TEGRA_SYSTEM_DMA)		+= dma.o
 obj-$(CONFIG_CPU_FREQ)                  += cpu-tegra.o
 obj-$(CONFIG_TEGRA_PCI)			+= pcie.o
 obj-$(CONFIG_USB_SUPPORT)		+= usb_phy.o
diff --git a/arch/arm/mach-tegra/apbio.c b/arch/arm/mach-tegra/apbio.c
index e75451e..dc0fe38 100644
--- a/arch/arm/mach-tegra/apbio.c
+++ b/arch/arm/mach-tegra/apbio.c
@@ -15,6 +15,9 @@
 
 #include <linux/kernel.h>
 #include <linux/io.h>
+#include <mach/iomap.h>
+#include <linux/of.h>
+#include <linux/dmaengine.h>
 #include <linux/dma-mapping.h>
 #include <linux/spinlock.h>
 #include <linux/completion.h>
@@ -22,17 +25,21 @@
 #include <linux/mutex.h>
 
 #include <mach/dma.h>
-#include <mach/iomap.h>
 
 #include "apbio.h"
 
+#if defined(CONFIG_TEGRA_SYSTEM_DMA) || defined(CONFIG_TEGRA20_APB_DMA)
 static DEFINE_MUTEX(tegra_apb_dma_lock);
-
-static struct tegra_dma_channel *tegra_apb_dma;
 static u32 *tegra_apb_bb;
 static dma_addr_t tegra_apb_bb_phys;
 static DECLARE_COMPLETION(tegra_apb_wait);
 
+static u32 tegra_apb_readl_direct(unsigned long offset);
+static void tegra_apb_writel_direct(u32 value, unsigned long offset);
+
+#if defined(CONFIG_TEGRA_SYSTEM_DMA)
+static struct tegra_dma_channel *tegra_apb_dma;
+
 bool tegra_apb_init(void)
 {
 	struct tegra_dma_channel *ch;
@@ -72,13 +79,13 @@ static void apb_dma_complete(struct tegra_dma_req *req)
 	complete(&tegra_apb_wait);
 }
 
-u32 tegra_apb_readl(unsigned long offset)
+static u32 tegra_apb_readl_using_dma(unsigned long offset)
 {
 	struct tegra_dma_req req;
 	int ret;
 
 	if (!tegra_apb_dma && !tegra_apb_init())
-		return readl(IO_TO_VIRT(offset));
+		return tegra_apb_readl_direct(offset);
 
 	mutex_lock(&tegra_apb_dma_lock);
 	req.complete = apb_dma_complete;
@@ -108,13 +115,13 @@ u32 tegra_apb_readl(unsigned long offset)
 	return *((u32 *)tegra_apb_bb);
 }
 
-void tegra_apb_writel(u32 value, unsigned long offset)
+static void tegra_apb_writel_using_dma(u32 value, unsigned long offset)
 {
 	struct tegra_dma_req req;
 	int ret;
 
 	if (!tegra_apb_dma && !tegra_apb_init()) {
-		writel(value, IO_TO_VIRT(offset));
+		tegra_apb_writel_direct(value, offset);
 		return;
 	}
 
@@ -143,3 +150,176 @@ void tegra_apb_writel(u32 value, unsigned long offset)
 
 	mutex_unlock(&tegra_apb_dma_lock);
 }
+
+#else
+static struct dma_chan *tegra_apb_dma_chan;
+static struct dma_slave_config dma_sconfig;
+
+bool tegra_apb_dma_init(void)
+{
+	dma_cap_mask_t mask;
+
+	mutex_lock(&tegra_apb_dma_lock);
+
+	/* Check to see if we raced to setup */
+	if (tegra_apb_dma_chan)
+		goto skip_init;
+
+	dma_cap_zero(mask);
+	dma_cap_set(DMA_SLAVE, mask);
+	tegra_apb_dma_chan = dma_request_channel(mask, NULL, NULL);
+	if (!tegra_apb_dma_chan) {
+		/*
+		 * This is common until the device is probed, so don't
+		 * shout about it.
+		 */
+		pr_debug("%s: can not allocate dma channel\n", __func__);
+		goto err_dma_alloc;
+	}
+
+	tegra_apb_bb = dma_alloc_coherent(NULL, sizeof(u32),
+		&tegra_apb_bb_phys, GFP_KERNEL);
+	if (!tegra_apb_bb) {
+		pr_err("%s: can not allocate bounce buffer\n", __func__);
+		goto err_buff_alloc;
+	}
+
+	dma_sconfig.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+	dma_sconfig.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+	dma_sconfig.slave_id = TEGRA_DMA_REQ_SEL_CNTR;
+	dma_sconfig.src_maxburst = 1;
+	dma_sconfig.dst_maxburst = 1;
+
+skip_init:
+	mutex_unlock(&tegra_apb_dma_lock);
+	return true;
+
+err_buff_alloc:
+	dma_release_channel(tegra_apb_dma_chan);
+	tegra_apb_dma_chan = NULL;
+
+err_dma_alloc:
+	mutex_unlock(&tegra_apb_dma_lock);
+	return false;
+}
+
+static void apb_dma_complete(void *args)
+{
+	complete(&tegra_apb_wait);
+}
+
+static int do_dma_transfer(unsigned long apb_add,
+		enum dma_transfer_direction dir)
+{
+	struct dma_async_tx_descriptor *dma_desc;
+	int ret;
+
+	if (dir == DMA_DEV_TO_MEM)
+		dma_sconfig.src_addr = apb_add;
+	else
+		dma_sconfig.dst_addr = apb_add;
+
+	ret = dmaengine_slave_config(tegra_apb_dma_chan, &dma_sconfig);
+	if (ret)
+		return ret;
+
+	dma_desc = dmaengine_prep_slave_single(tegra_apb_dma_chan,
+			tegra_apb_bb_phys, sizeof(u32), dir,
+			DMA_PREP_INTERRUPT |  DMA_CTRL_ACK);
+	if (!dma_desc)
+		return -EINVAL;
+
+	dma_desc->callback = apb_dma_complete;
+	dma_desc->callback_param = NULL;
+
+	INIT_COMPLETION(tegra_apb_wait);
+
+	dmaengine_submit(dma_desc);
+	dma_async_issue_pending(tegra_apb_dma_chan);
+	ret = wait_for_completion_timeout(&tegra_apb_wait,
+		msecs_to_jiffies(50));
+
+	if (WARN(ret == 0, "apb read dma timed out")) {
+		dmaengine_terminate_all(tegra_apb_dma_chan);
+		return -EFAULT;
+	}
+	return 0;
+}
+
+static u32 tegra_apb_readl_using_dma(unsigned long offset)
+{
+	int ret;
+
+	if (!tegra_apb_dma_chan && !tegra_apb_dma_init())
+		return tegra_apb_readl_direct(offset);
+
+	mutex_lock(&tegra_apb_dma_lock);
+	ret = do_dma_transfer(offset, DMA_DEV_TO_MEM);
+	if (ret < 0) {
+		pr_err("error in reading offset 0x%08lx using dma\n", offset);
+		*(u32 *)tegra_apb_bb = 0;
+	}
+	mutex_unlock(&tegra_apb_dma_lock);
+	return *((u32 *)tegra_apb_bb);
+}
+
+static void tegra_apb_writel_using_dma(u32 value, unsigned long offset)
+{
+	int ret;
+
+	if (!tegra_apb_dma_chan && !tegra_apb_dma_init()) {
+		tegra_apb_writel_direct(value, offset);
+		return;
+	}
+
+	mutex_lock(&tegra_apb_dma_lock);
+	*((u32 *)tegra_apb_bb) = value;
+	ret = do_dma_transfer(offset, DMA_MEM_TO_DEV);
+	if (ret < 0)
+		pr_err("error in writing offset 0x%08lx using dma\n", offset);
+	mutex_unlock(&tegra_apb_dma_lock);
+}
+#endif
+#else
+#define tegra_apb_readl_using_dma tegra_apb_readl_direct
+#define tegra_apb_writel_using_dma tegra_apb_writel_direct
+#endif
+
+typedef u32 (*apbio_read_fptr)(unsigned long offset);
+typedef void (*apbio_write_fptr)(u32 value, unsigned long offset);
+
+static apbio_read_fptr apbio_read;
+static apbio_write_fptr apbio_write;
+
+static u32 tegra_apb_readl_direct(unsigned long offset)
+{
+	return readl(IO_TO_VIRT(offset));
+}
+
+static void tegra_apb_writel_direct(u32 value, unsigned long offset)
+{
+	writel(value, IO_TO_VIRT(offset));
+}
+
+void tegra_apb_io_init(void)
+{
+	/* Need to use dma only when it is Tegra20 based platform */
+	if (of_machine_is_compatible("nvidia,tegra20") ||
+			!of_have_populated_dt()) {
+		apbio_read = tegra_apb_readl_using_dma;
+		apbio_write = tegra_apb_writel_using_dma;
+	} else {
+		apbio_read = tegra_apb_readl_direct;
+		apbio_write = tegra_apb_writel_direct;
+	}
+}
+
+u32 tegra_apb_readl(unsigned long offset)
+{
+	return apbio_read(offset);
+}
+
+void tegra_apb_writel(u32 value, unsigned long offset)
+{
+	apbio_write(value, offset);
+}
diff --git a/arch/arm/mach-tegra/apbio.h b/arch/arm/mach-tegra/apbio.h
index 8b49e8c..f05d71c 100644
--- a/arch/arm/mach-tegra/apbio.h
+++ b/arch/arm/mach-tegra/apbio.h
@@ -16,24 +16,7 @@
 #ifndef __MACH_TEGRA_APBIO_H
 #define __MACH_TEGRA_APBIO_H
 
-#ifdef CONFIG_TEGRA_SYSTEM_DMA
-
+void tegra_apb_io_init(void);
 u32 tegra_apb_readl(unsigned long offset);
 void tegra_apb_writel(u32 value, unsigned long offset);
-
-#else
-#include <asm/io.h>
-#include <mach/io.h>
-
-static inline u32 tegra_apb_readl(unsigned long offset)
-{
-        return readl(IO_TO_VIRT(offset));
-}
-
-static inline void tegra_apb_writel(u32 value, unsigned long offset)
-{
-        writel(value, IO_TO_VIRT(offset));
-}
-#endif
-
 #endif
diff --git a/arch/arm/mach-tegra/common.c b/arch/arm/mach-tegra/common.c
index 204a5c8..96fef6b 100644
--- a/arch/arm/mach-tegra/common.c
+++ b/arch/arm/mach-tegra/common.c
@@ -33,6 +33,7 @@
 #include "clock.h"
 #include "fuse.h"
 #include "pmc.h"
+#include "apbio.h"
 
 /*
  * Storage for debug-macro.S's state.
@@ -127,6 +128,7 @@ static void __init tegra_init_cache(u32 tag_latency, u32 data_latency)
 #ifdef CONFIG_ARCH_TEGRA_2x_SOC
 void __init tegra20_init_early(void)
 {
+	tegra_apb_io_init();
 	tegra_init_fuse();
 	tegra2_init_clocks();
 	tegra_clk_init_from_table(tegra20_clk_init_table);
@@ -138,6 +140,7 @@ void __init tegra20_init_early(void)
 #ifdef CONFIG_ARCH_TEGRA_3x_SOC
 void __init tegra30_init_early(void)
 {
+	tegra_apb_io_init();
 	tegra_init_fuse();
 	tegra30_init_clocks();
 	tegra_clk_init_from_table(tegra30_clk_init_table);