From f930ddd0583c1a9e68d80a27d4e5077e795007b1 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Wed, 21 Nov 2012 15:55:21 +0100
Subject: drm/i915: remove duplicate register #defines

Somehow a chunk of unused register defines ended up in the middle of
the PLL defines. They go back to the original kms merging.

The only used #define is SR01, move it to the register name together
with the other legacy vga stuff.

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 97fbd9d..462b50a 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -142,6 +142,7 @@
 #define   VGA_MSR_CGA_MODE (1<<0)
 
 #define VGA_SR_INDEX 0x3c4
+#define SR01			1
 #define VGA_SR_DATA 0x3c5
 
 #define VGA_AR_INDEX 0x3c0
@@ -938,23 +939,6 @@
 #define   DPLL_LOCK_VLV			(1<<15)
 #define   DPLL_INTEGRATED_CLOCK_VLV	(1<<13)
 
-#define SRX_INDEX		0x3c4
-#define SRX_DATA		0x3c5
-#define SR01			1
-#define SR01_SCREEN_OFF		(1<<5)
-
-#define PPCR			0x61204
-#define PPCR_ON			(1<<0)
-
-#define DVOB			0x61140
-#define DVOB_ON			(1<<31)
-#define DVOC			0x61160
-#define DVOC_ON			(1<<31)
-#define LVDS			0x61180
-#define LVDS_ON			(1<<31)
-
-/* Scratch pad debug 0 reg:
- */
 #define   DPLL_FPA01_P1_POST_DIV_MASK_I830	0x001f0000
 /*
  * The i830 generation, in LVDS mode, defines P1 as the bit number set within
-- 
cgit v0.10.2


From ca9c46c5c77987acf1bf7137bf85e9221bc459ba Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Tue, 27 Nov 2012 20:34:58 +0200
Subject: drm/i915: Kill i915_gem_execbuffer_wait_for_flips()
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

As per Chris Wilson's suggestion make
i915_gem_execbuffer_wait_for_flips() go away.

This was used to stall the GPU ring while there are pending
page flips involving the relevant BO. Ie. while the BO is still
being scanned out by the display controller.

The recommended alternative is to use the page flip events to
wait for the page flips to fully complete before reusing the BO
of the old front buffer. Or use more buffers.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Kristian Høgsberg <krh@bitplanet.net>
Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
[danvet: don't remove obj->pending_flips, still required due to
reorder patches.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index ee8f97f..802d925 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -602,44 +602,11 @@ err:
 }
 
 static int
-i915_gem_execbuffer_wait_for_flips(struct intel_ring_buffer *ring, u32 flips)
-{
-	u32 plane, flip_mask;
-	int ret;
-
-	/* Check for any pending flips. As we only maintain a flip queue depth
-	 * of 1, we can simply insert a WAIT for the next display flip prior
-	 * to executing the batch and avoid stalling the CPU.
-	 */
-
-	for (plane = 0; flips >> plane; plane++) {
-		if (((flips >> plane) & 1) == 0)
-			continue;
-
-		if (plane)
-			flip_mask = MI_WAIT_FOR_PLANE_B_FLIP;
-		else
-			flip_mask = MI_WAIT_FOR_PLANE_A_FLIP;
-
-		ret = intel_ring_begin(ring, 2);
-		if (ret)
-			return ret;
-
-		intel_ring_emit(ring, MI_WAIT_FOR_EVENT | flip_mask);
-		intel_ring_emit(ring, MI_NOOP);
-		intel_ring_advance(ring);
-	}
-
-	return 0;
-}
-
-static int
 i915_gem_execbuffer_move_to_gpu(struct intel_ring_buffer *ring,
 				struct list_head *objects)
 {
 	struct drm_i915_gem_object *obj;
 	uint32_t flush_domains = 0;
-	uint32_t flips = 0;
 	int ret;
 
 	list_for_each_entry(obj, objects, exec_list) {
@@ -650,18 +617,9 @@ i915_gem_execbuffer_move_to_gpu(struct intel_ring_buffer *ring,
 		if (obj->base.write_domain & I915_GEM_DOMAIN_CPU)
 			i915_gem_clflush_object(obj);
 
-		if (obj->base.pending_write_domain)
-			flips |= atomic_read(&obj->pending_flip);
-
 		flush_domains |= obj->base.write_domain;
 	}
 
-	if (flips) {
-		ret = i915_gem_execbuffer_wait_for_flips(ring, flips);
-		if (ret)
-			return ret;
-	}
-
 	if (flush_domains & I915_GEM_DOMAIN_CPU)
 		i915_gem_chipset_flush(ring->dev);
 
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 78d12c4..8cdf3a1 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -6945,8 +6945,6 @@ static void do_intel_finish_page_flip(struct drm_device *dev,
 
 	obj = work->old_fb_obj;
 
-	atomic_clear_mask(1 << intel_crtc->plane,
-			  &obj->pending_flip.counter);
 	wake_up(&dev_priv->pending_flip_queue);
 
 	queue_work(dev_priv->wq, &work->work);
@@ -7292,10 +7290,6 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc,
 
 	work->enable_stall_check = true;
 
-	/* Block clients from rendering to the new back buffer until
-	 * the flip occurs and the object is no longer visible.
-	 */
-	atomic_add(1 << intel_crtc->plane, &work->old_fb_obj->pending_flip);
 	atomic_inc(&intel_crtc->unpin_work_count);
 
 	ret = dev_priv->display.queue_flip(dev, crtc, fb, obj);
@@ -7312,7 +7306,6 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc,
 
 cleanup_pending:
 	atomic_dec(&intel_crtc->unpin_work_count);
-	atomic_sub(1 << intel_crtc->plane, &work->old_fb_obj->pending_flip);
 	drm_gem_object_unreference(&work->old_fb_obj->base);
 	drm_gem_object_unreference(&obj->base);
 	mutex_unlock(&dev->struct_mutex);
-- 
cgit v0.10.2


From 04b97b3422c3403696f3bcb52140b7ebeaee9d3c Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Tue, 27 Nov 2012 17:06:53 +0000
Subject: drm/i915/debugfs: Prune a couple of superfluous leading zeros from bo
 domains

As we do not have any domains occupying the high bits, there is no point
in always printing the leading 00.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index 8afc0dd..62619e3 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -102,7 +102,7 @@ static const char *cache_level_str(int type)
 static void
 describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj)
 {
-	seq_printf(m, "%p: %s%s %8zdKiB %04x %04x %d %d %d%s%s%s",
+	seq_printf(m, "%p: %s%s %8zdKiB %02x %02x %d %d %d%s%s%s",
 		   &obj->base,
 		   get_pin_flag(obj),
 		   get_tiling_flag(obj),
@@ -608,7 +608,7 @@ static void print_error_buffers(struct seq_file *m,
 	seq_printf(m, "%s [%d]:\n", name, count);
 
 	while (count--) {
-		seq_printf(m, "  %08x %8u %04x %04x %x %x%s%s%s%s%s%s%s",
+		seq_printf(m, "  %08x %8u %02x %02x %x %x%s%s%s%s%s%s%s",
 			   err->gtt_offset,
 			   err->size,
 			   err->read_domains,
-- 
cgit v0.10.2


From dafd226c4f54eded10ba43c37789a6aa20b59c32 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Mon, 26 Nov 2012 17:22:07 +0100
Subject: drm/i915: add encoder->pre_pll_enable callback

Currently we have two encoder specific bits in the common mode_set
functions:
- lvds pin pair enabling
- dp m/n setting and computation

Now the lvds stuff needs to happen before the pll is enabled. Since
that is done in the crtc_mode_set functions, we need to add a new
callback to be able to move them to the encoder code (where they
belong). The dp m/n stuff is a giant mess anyway (since it also
confuses itself with the fdi link m/n handling), so that needs to be
handled separately.

I think that we can move the pll enabling down quite a bit, which
might allow us to eventually merge encoder->pre_enable with this new
pre_pll_enable callback. But for now this will allow us to clean
things up a bit.

Note that vlv doesn't support lvds, hence we don't need to change
anything in there.

v2: Fixup commit message, both suggested from Paulo Zanoni.
- dp m/n doesn't need to happen before pll enabling
- lvds doesn't exist on vlv, hence no changes required in the vlv pll
  function.

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 8cdf3a1..55d3e3d 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -4464,6 +4464,7 @@ static void i9xx_update_pll(struct drm_crtc *crtc,
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct intel_encoder *encoder;
 	int pipe = intel_crtc->pipe;
 	u32 dpll;
 	bool is_sdvo;
@@ -4532,6 +4533,10 @@ static void i9xx_update_pll(struct drm_crtc *crtc,
 	POSTING_READ(DPLL(pipe));
 	udelay(150);
 
+	for_each_encoder_on_crtc(dev, crtc, encoder)
+		if (encoder->pre_pll_enable)
+			encoder->pre_pll_enable(encoder);
+
 	/* The LVDS pin pair needs to be on before the DPLLs are enabled.
 	 * This is an exception to the general rule that mode_set doesn't turn
 	 * things on.
@@ -4576,6 +4581,7 @@ static void i8xx_update_pll(struct drm_crtc *crtc,
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct intel_encoder *encoder;
 	int pipe = intel_crtc->pipe;
 	u32 dpll;
 
@@ -4609,6 +4615,10 @@ static void i8xx_update_pll(struct drm_crtc *crtc,
 	POSTING_READ(DPLL(pipe));
 	udelay(150);
 
+	for_each_encoder_on_crtc(dev, crtc, encoder)
+		if (encoder->pre_pll_enable)
+			encoder->pre_pll_enable(encoder);
+
 	/* The LVDS pin pair needs to be on before the DPLLs are enabled.
 	 * This is an exception to the general rule that mode_set doesn't turn
 	 * things on.
@@ -5537,6 +5547,10 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 		I915_WRITE(TRANSDPLINK_N1(pipe), 0);
 	}
 
+	for_each_encoder_on_crtc(dev, crtc, encoder)
+		if (encoder->pre_pll_enable)
+			encoder->pre_pll_enable(encoder);
+
 	if (intel_crtc->pch_pll) {
 		I915_WRITE(intel_crtc->pch_pll->pll_reg, dpll);
 
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 522061c..a74c572 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -153,6 +153,7 @@ struct intel_encoder {
 	bool cloneable;
 	bool connectors_active;
 	void (*hot_plug)(struct intel_encoder *);
+	void (*pre_pll_enable)(struct intel_encoder *);
 	void (*pre_enable)(struct intel_encoder *);
 	void (*enable)(struct intel_encoder *);
 	void (*disable)(struct intel_encoder *);
-- 
cgit v0.10.2


From a210b028f07690c127733addbbe137e8f4cad30c Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Mon, 26 Nov 2012 17:22:08 +0100
Subject: drm/i915: replace ad-hoc dual-link lvds checks

... with is_dual_link_lvds introduced in

commit b03543857fd75876b96e10d4320b775e95041bb7
Author: Takashi Iwai <tiwai@suse.de>
Date:   Tue Mar 20 13:07:05 2012 +0100

    drm/i915: Check VBIOS value for determining LVDS dual channel mode, too

All these checks predate this commit and have simply been overlooked.
Since we don't support switching between single-link and dual-link
modes anyway, this different checks could at best only get in the way
of refactorings, and in the worst case cause inconsistencies.

v2: Update the comment, we now have a solid way to figure out whether
we need dual-link lvds or not (falling back to vbt values as a last
resort). We still don't know how to switch between dual-link and
single link so leave that part intact. I'm not sure though whether
switching between these two modes makes any sense - we always drive
the panel at its fixed mode (with a fixed bpc) anyway ...

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 55d3e3d..be19b6d 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -690,13 +690,11 @@ intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 	intel_clock_t clock;
 	int err = target;
 
-	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS) &&
-	    (I915_READ(LVDS)) != 0) {
+	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS)) {
 		/*
-		 * For LVDS, if the panel is on, just rely on its current
-		 * settings for dual-channel.  We haven't figured out how to
-		 * reliably set up different single/dual channel state, if we
-		 * even can.
+		 * For LVDS just rely on its current settings for dual-channel.
+		 * We haven't figured out how to reliably set up different
+		 * single/dual channel state, if we even can.
 		 */
 		if (is_dual_link_lvds(dev_priv, LVDS))
 			clock.p2 = limit->p2.p2_fast;
@@ -766,8 +764,7 @@ intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 			lvds_reg = PCH_LVDS;
 		else
 			lvds_reg = LVDS;
-		if ((I915_READ(lvds_reg) & LVDS_CLKB_POWER_MASK) ==
-		    LVDS_CLKB_POWER_UP)
+		if (is_dual_link_lvds(dev_priv, lvds_reg))
 			clock.p2 = limit->p2.p2_fast;
 		else
 			clock.p2 = limit->p2.p2_slow;
@@ -5359,7 +5356,7 @@ static uint32_t ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 	if (is_lvds) {
 		if ((intel_panel_use_ssc(dev_priv) &&
 		     dev_priv->lvds_ssc_freq == 100) ||
-		    (I915_READ(PCH_LVDS) & LVDS_CLKB_POWER_MASK) == LVDS_CLKB_POWER_UP)
+		    is_dual_link_lvds(dev_priv, PCH_LVDS))
 			factor = 25;
 	} else if (is_sdvo && is_tv)
 		factor = 20;
-- 
cgit v0.10.2


From 1974cad0ee4ce84e5cb792e49c4f0d9421e0312c Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Mon, 26 Nov 2012 17:22:09 +0100
Subject: drm/i915: move is_dual_link_lvds to intel_lvds.c

Just a prep patch to make this a property of intel_lvds. Makes more
sense, removes clutter from intel_display.c and eventually I want to
move all the encoder special cases wrt clock handling to encoders
anyway.

v2: Add an intel_ prefixe to is_dual_link_lvds since it's non-static
now.

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index be19b6d..2705da3 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -472,61 +472,14 @@ static void vlv_init_dpio(struct drm_device *dev)
 	POSTING_READ(DPIO_CTL);
 }
 
-static int intel_dual_link_lvds_callback(const struct dmi_system_id *id)
-{
-	DRM_INFO("Forcing lvds to dual link mode on %s\n", id->ident);
-	return 1;
-}
-
-static const struct dmi_system_id intel_dual_link_lvds[] = {
-	{
-		.callback = intel_dual_link_lvds_callback,
-		.ident = "Apple MacBook Pro (Core i5/i7 Series)",
-		.matches = {
-			DMI_MATCH(DMI_SYS_VENDOR, "Apple Inc."),
-			DMI_MATCH(DMI_PRODUCT_NAME, "MacBookPro8,2"),
-		},
-	},
-	{ }	/* terminating entry */
-};
-
-static bool is_dual_link_lvds(struct drm_i915_private *dev_priv,
-			      unsigned int reg)
-{
-	unsigned int val;
-
-	/* use the module option value if specified */
-	if (i915_lvds_channel_mode > 0)
-		return i915_lvds_channel_mode == 2;
-
-	if (dmi_check_system(intel_dual_link_lvds))
-		return true;
-
-	if (dev_priv->lvds_val)
-		val = dev_priv->lvds_val;
-	else {
-		/* BIOS should set the proper LVDS register value at boot, but
-		 * in reality, it doesn't set the value when the lid is closed;
-		 * we need to check "the value to be set" in VBT when LVDS
-		 * register is uninitialized.
-		 */
-		val = I915_READ(reg);
-		if (!(val & ~(LVDS_PIPE_MASK | LVDS_DETECTED)))
-			val = dev_priv->bios_lvds_val;
-		dev_priv->lvds_val = val;
-	}
-	return (val & LVDS_CLKB_POWER_MASK) == LVDS_CLKB_POWER_UP;
-}
-
 static const intel_limit_t *intel_ironlake_limit(struct drm_crtc *crtc,
 						int refclk)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	const intel_limit_t *limit;
 
 	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS)) {
-		if (is_dual_link_lvds(dev_priv, PCH_LVDS)) {
+		if (intel_is_dual_link_lvds(dev)) {
 			/* LVDS dual channel */
 			if (refclk == 100000)
 				limit = &intel_limits_ironlake_dual_lvds_100m;
@@ -550,11 +503,10 @@ static const intel_limit_t *intel_ironlake_limit(struct drm_crtc *crtc,
 static const intel_limit_t *intel_g4x_limit(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	const intel_limit_t *limit;
 
 	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS)) {
-		if (is_dual_link_lvds(dev_priv, LVDS))
+		if (intel_is_dual_link_lvds(dev))
 			/* LVDS with dual channel */
 			limit = &intel_limits_g4x_dual_channel_lvds;
 		else
@@ -686,7 +638,6 @@ intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	intel_clock_t clock;
 	int err = target;
 
@@ -696,7 +647,7 @@ intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 		 * We haven't figured out how to reliably set up different
 		 * single/dual channel state, if we even can.
 		 */
-		if (is_dual_link_lvds(dev_priv, LVDS))
+		if (intel_is_dual_link_lvds(dev))
 			clock.p2 = limit->p2.p2_fast;
 		else
 			clock.p2 = limit->p2.p2_slow;
@@ -749,7 +700,6 @@ intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 			intel_clock_t *best_clock)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	intel_clock_t clock;
 	int max_n;
 	bool found;
@@ -764,7 +714,7 @@ intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 			lvds_reg = PCH_LVDS;
 		else
 			lvds_reg = LVDS;
-		if (is_dual_link_lvds(dev_priv, lvds_reg))
+		if (intel_is_dual_link_lvds(dev))
 			clock.p2 = limit->p2.p2_fast;
 		else
 			clock.p2 = limit->p2.p2_slow;
@@ -5356,7 +5306,7 @@ static uint32_t ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 	if (is_lvds) {
 		if ((intel_panel_use_ssc(dev_priv) &&
 		     dev_priv->lvds_ssc_freq == 100) ||
-		    is_dual_link_lvds(dev_priv, PCH_LVDS))
+		    intel_is_dual_link_lvds(dev))
 			factor = 25;
 	} else if (is_sdvo && is_tv)
 		factor = 20;
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index a74c572..7ca7772 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -441,6 +441,7 @@ extern void intel_mark_idle(struct drm_device *dev);
 extern void intel_mark_fb_busy(struct drm_i915_gem_object *obj);
 extern void intel_mark_fb_idle(struct drm_i915_gem_object *obj);
 extern bool intel_lvds_init(struct drm_device *dev);
+extern bool intel_is_dual_link_lvds(struct drm_device *dev);
 extern void intel_dp_init(struct drm_device *dev, int output_reg,
 			  enum port port);
 extern void intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c
index b9a660a..4158a88 100644
--- a/drivers/gpu/drm/i915/intel_lvds.c
+++ b/drivers/gpu/drm/i915/intel_lvds.c
@@ -903,6 +903,59 @@ static bool lvds_is_present_in_vbt(struct drm_device *dev,
 	return false;
 }
 
+static int intel_dual_link_lvds_callback(const struct dmi_system_id *id)
+{
+	DRM_INFO("Forcing lvds to dual link mode on %s\n", id->ident);
+	return 1;
+}
+
+static const struct dmi_system_id intel_dual_link_lvds[] = {
+	{
+		.callback = intel_dual_link_lvds_callback,
+		.ident = "Apple MacBook Pro (Core i5/i7 Series)",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Apple Inc."),
+			DMI_MATCH(DMI_PRODUCT_NAME, "MacBookPro8,2"),
+		},
+	},
+	{ }	/* terminating entry */
+};
+
+bool intel_is_dual_link_lvds(struct drm_device *dev)
+{
+	unsigned int val;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 lvds_reg;
+
+	if (HAS_PCH_SPLIT(dev)) {
+		lvds_reg = PCH_LVDS;
+	} else {
+		lvds_reg = LVDS;
+	}
+
+	/* use the module option value if specified */
+	if (i915_lvds_channel_mode > 0)
+		return i915_lvds_channel_mode == 2;
+
+	if (dmi_check_system(intel_dual_link_lvds))
+		return true;
+
+	if (dev_priv->lvds_val)
+		val = dev_priv->lvds_val;
+	else {
+		/* BIOS should set the proper LVDS register value at boot, but
+		 * in reality, it doesn't set the value when the lid is closed;
+		 * we need to check "the value to be set" in VBT when LVDS
+		 * register is uninitialized.
+		 */
+		val = I915_READ(lvds_reg);
+		if (!(val & ~(LVDS_PIPE_MASK | LVDS_DETECTED)))
+			val = dev_priv->bios_lvds_val;
+		dev_priv->lvds_val = val;
+	}
+	return (val & LVDS_CLKB_POWER_MASK) == LVDS_CLKB_POWER_UP;
+}
+
 static bool intel_lvds_supported(struct drm_device *dev)
 {
 	/* With the introduction of the PCH we gained a dedicated
-- 
cgit v0.10.2


From 13c7d8703127334890c894ddab13b3a92a26580a Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Mon, 26 Nov 2012 17:22:10 +0100
Subject: drm/i915: track is_dual_link in intel_lvds

Yeah, all users (both the clock selection special cases and the lvds
pin pair stuff) are still in common code, but this will change.

v2: Rebase on top of Jani Nikula's panel rework.

v3: Incorporate review from Paulo Zanoni:
- s/__is_dual_link_lvds/compute_is_dual_link_lvds
- kill dev_priv->lvds_val
- drop spurious whitespace change

v4: Add a debug printk to display the dual-link status, as suggested
by Paulo Zanoni in review.

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com> (v3)
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 79589bb..9be7efd 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -705,7 +705,6 @@ typedef struct drm_i915_private {
 	unsigned int display_clock_mode:1;
 	int lvds_ssc_freq;
 	unsigned int bios_lvds_val; /* initial [PCH_]LVDS reg val in VBIOS */
-	unsigned int lvds_val; /* used for checking LVDS channel mode */
 	struct {
 		int rate;
 		int lanes;
diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c
index 4158a88..31fdbfd 100644
--- a/drivers/gpu/drm/i915/intel_lvds.c
+++ b/drivers/gpu/drm/i915/intel_lvds.c
@@ -52,6 +52,7 @@ struct intel_lvds_encoder {
 	u32 pfit_control;
 	u32 pfit_pgm_ratios;
 	bool pfit_dirty;
+	bool is_dual_link;
 
 	struct intel_lvds_connector *attached_connector;
 };
@@ -923,6 +924,23 @@ static const struct dmi_system_id intel_dual_link_lvds[] = {
 
 bool intel_is_dual_link_lvds(struct drm_device *dev)
 {
+	struct intel_encoder *encoder;
+	struct intel_lvds_encoder *lvds_encoder;
+
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
+			    base.head) {
+		if (encoder->type == INTEL_OUTPUT_LVDS) {
+			lvds_encoder = to_lvds_encoder(&encoder->base);
+
+			return lvds_encoder->is_dual_link;
+		}
+	}
+
+	return false;
+}
+
+static bool compute_is_dual_link_lvds(struct drm_device *dev)
+{
 	unsigned int val;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 lvds_reg;
@@ -940,19 +958,15 @@ bool intel_is_dual_link_lvds(struct drm_device *dev)
 	if (dmi_check_system(intel_dual_link_lvds))
 		return true;
 
-	if (dev_priv->lvds_val)
-		val = dev_priv->lvds_val;
-	else {
-		/* BIOS should set the proper LVDS register value at boot, but
-		 * in reality, it doesn't set the value when the lid is closed;
-		 * we need to check "the value to be set" in VBT when LVDS
-		 * register is uninitialized.
-		 */
-		val = I915_READ(lvds_reg);
-		if (!(val & ~(LVDS_PIPE_MASK | LVDS_DETECTED)))
-			val = dev_priv->bios_lvds_val;
-		dev_priv->lvds_val = val;
-	}
+	/* BIOS should set the proper LVDS register value at boot, but
+	 * in reality, it doesn't set the value when the lid is closed;
+	 * we need to check "the value to be set" in VBT when LVDS
+	 * register is uninitialized.
+	 */
+	val = I915_READ(lvds_reg);
+	if (!(val & ~(LVDS_PIPE_MASK | LVDS_DETECTED)))
+		val = dev_priv->bios_lvds_val;
+
 	return (val & LVDS_CLKB_POWER_MASK) == LVDS_CLKB_POWER_UP;
 }
 
@@ -1162,6 +1176,10 @@ bool intel_lvds_init(struct drm_device *dev)
 		goto failed;
 
 out:
+	lvds_encoder->is_dual_link = compute_is_dual_link_lvds(dev);
+	DRM_DEBUG_KMS("detected %s-link lvds configuration\n",
+		      lvds_encoder->is_dual_link ? "dual" : "single");
+
 	/*
 	 * Unlock registers and just
 	 * leave them unlocked
-- 
cgit v0.10.2


From 7dec060675125f17a547950d715018a63131a437 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Tue, 11 Sep 2012 14:12:25 +0200
Subject: drm/i915: add intel_lvds->reg

To ditch at least some of the PCH_SPLIT ? PCH_LVDS : LVDS code ...

v2: Rebase on top of Jani Nikula's panel rework.

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c
index 31fdbfd..6b6ed64 100644
--- a/drivers/gpu/drm/i915/intel_lvds.c
+++ b/drivers/gpu/drm/i915/intel_lvds.c
@@ -53,6 +53,7 @@ struct intel_lvds_encoder {
 	u32 pfit_pgm_ratios;
 	bool pfit_dirty;
 	bool is_dual_link;
+	u32 reg;
 
 	struct intel_lvds_connector *attached_connector;
 };
@@ -72,15 +73,10 @@ static bool intel_lvds_get_hw_state(struct intel_encoder *encoder,
 {
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 lvds_reg, tmp;
-
-	if (HAS_PCH_SPLIT(dev)) {
-		lvds_reg = PCH_LVDS;
-	} else {
-		lvds_reg = LVDS;
-	}
+	struct intel_lvds_encoder *lvds_encoder = to_lvds_encoder(&encoder->base);
+	u32 tmp;
 
-	tmp = I915_READ(lvds_reg);
+	tmp = I915_READ(lvds_encoder->reg);
 
 	if (!(tmp & LVDS_PORT_EN))
 		return false;
@@ -102,19 +98,17 @@ static void intel_enable_lvds(struct intel_encoder *encoder)
 	struct intel_lvds_encoder *lvds_encoder = to_lvds_encoder(&encoder->base);
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->base.crtc);
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 ctl_reg, lvds_reg, stat_reg;
+	u32 ctl_reg, stat_reg;
 
 	if (HAS_PCH_SPLIT(dev)) {
 		ctl_reg = PCH_PP_CONTROL;
-		lvds_reg = PCH_LVDS;
 		stat_reg = PCH_PP_STATUS;
 	} else {
 		ctl_reg = PP_CONTROL;
-		lvds_reg = LVDS;
 		stat_reg = PP_STATUS;
 	}
 
-	I915_WRITE(lvds_reg, I915_READ(lvds_reg) | LVDS_PORT_EN);
+	I915_WRITE(lvds_encoder->reg, I915_READ(lvds_encoder->reg) | LVDS_PORT_EN);
 
 	if (lvds_encoder->pfit_dirty) {
 		/*
@@ -133,7 +127,7 @@ static void intel_enable_lvds(struct intel_encoder *encoder)
 	}
 
 	I915_WRITE(ctl_reg, I915_READ(ctl_reg) | POWER_TARGET_ON);
-	POSTING_READ(lvds_reg);
+	POSTING_READ(lvds_encoder->reg);
 	if (wait_for((I915_READ(stat_reg) & PP_ON) != 0, 1000))
 		DRM_ERROR("timed out waiting for panel to power on\n");
 
@@ -145,15 +139,13 @@ static void intel_disable_lvds(struct intel_encoder *encoder)
 	struct drm_device *dev = encoder->base.dev;
 	struct intel_lvds_encoder *lvds_encoder = to_lvds_encoder(&encoder->base);
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 ctl_reg, lvds_reg, stat_reg;
+	u32 ctl_reg, stat_reg;
 
 	if (HAS_PCH_SPLIT(dev)) {
 		ctl_reg = PCH_PP_CONTROL;
-		lvds_reg = PCH_LVDS;
 		stat_reg = PCH_PP_STATUS;
 	} else {
 		ctl_reg = PP_CONTROL;
-		lvds_reg = LVDS;
 		stat_reg = PP_STATUS;
 	}
 
@@ -168,8 +160,8 @@ static void intel_disable_lvds(struct intel_encoder *encoder)
 		lvds_encoder->pfit_dirty = true;
 	}
 
-	I915_WRITE(lvds_reg, I915_READ(lvds_reg) & ~LVDS_PORT_EN);
-	POSTING_READ(lvds_reg);
+	I915_WRITE(lvds_encoder->reg, I915_READ(lvds_encoder->reg) & ~LVDS_PORT_EN);
+	POSTING_READ(lvds_encoder->reg);
 }
 
 static int intel_lvds_mode_valid(struct drm_connector *connector,
@@ -939,17 +931,11 @@ bool intel_is_dual_link_lvds(struct drm_device *dev)
 	return false;
 }
 
-static bool compute_is_dual_link_lvds(struct drm_device *dev)
+static bool compute_is_dual_link_lvds(struct intel_lvds_encoder *lvds_encoder)
 {
+	struct drm_device *dev = lvds_encoder->base.base.dev;
 	unsigned int val;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 lvds_reg;
-
-	if (HAS_PCH_SPLIT(dev)) {
-		lvds_reg = PCH_LVDS;
-	} else {
-		lvds_reg = LVDS;
-	}
 
 	/* use the module option value if specified */
 	if (i915_lvds_channel_mode > 0)
@@ -963,7 +949,7 @@ static bool compute_is_dual_link_lvds(struct drm_device *dev)
 	 * we need to check "the value to be set" in VBT when LVDS
 	 * register is uninitialized.
 	 */
-	val = I915_READ(lvds_reg);
+	val = I915_READ(lvds_encoder->reg);
 	if (!(val & ~(LVDS_PIPE_MASK | LVDS_DETECTED)))
 		val = dev_priv->bios_lvds_val;
 
@@ -1076,6 +1062,12 @@ bool intel_lvds_init(struct drm_device *dev)
 	connector->interlace_allowed = false;
 	connector->doublescan_allowed = false;
 
+	if (HAS_PCH_SPLIT(dev)) {
+		lvds_encoder->reg = PCH_LVDS;
+	} else {
+		lvds_encoder->reg = LVDS;
+	}
+
 	/* create the scaling mode property */
 	drm_mode_create_scaling_mode_property(dev);
 	drm_object_attach_property(&connector->base,
@@ -1176,7 +1168,7 @@ bool intel_lvds_init(struct drm_device *dev)
 		goto failed;
 
 out:
-	lvds_encoder->is_dual_link = compute_is_dual_link_lvds(dev);
+	lvds_encoder->is_dual_link = compute_is_dual_link_lvds(lvds_encoder);
 	DRM_DEBUG_KMS("detected %s-link lvds configuration\n",
 		      lvds_encoder->is_dual_link ? "dual" : "single");
 
-- 
cgit v0.10.2


From fc683091eb57692679f629784f42dce453877430 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Mon, 26 Nov 2012 17:22:12 +0100
Subject: drm/i915: move intel_update_lvds to intel_lvds->pre_pll_enable

A few things needed to change:
- HAS_PCH_SPLIT since ilk+ is not yet converted to this.
- s/LVDS/intel_lvds->reg/ to prep for ilk conversion
- replace the clock.p2 == 7 check with a is_dual_link check
- s/adjusted_mode/intel_lvds->fixed_mode

v2: Rebase on top of Jani Nikula's panel rework. I'm wondering whether
we shouldn't add an attached_panel pointer to intel_encoder, to
replace the encoder private ->attached_connector pointers, since
that's essentially what we need.

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 2705da3..bb080fb 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -4257,51 +4257,6 @@ static void i9xx_update_pll_dividers(struct drm_crtc *crtc,
 	}
 }
 
-static void intel_update_lvds(struct drm_crtc *crtc, intel_clock_t *clock,
-			      struct drm_display_mode *adjusted_mode)
-{
-	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int pipe = intel_crtc->pipe;
-	u32 temp;
-
-	temp = I915_READ(LVDS);
-	temp |= LVDS_PORT_EN | LVDS_A0A2_CLKA_POWER_UP;
-	if (pipe == 1) {
-		temp |= LVDS_PIPEB_SELECT;
-	} else {
-		temp &= ~LVDS_PIPEB_SELECT;
-	}
-	/* set the corresponsding LVDS_BORDER bit */
-	temp |= dev_priv->lvds_border_bits;
-	/* Set the B0-B3 data pairs corresponding to whether we're going to
-	 * set the DPLLs for dual-channel mode or not.
-	 */
-	if (clock->p2 == 7)
-		temp |= LVDS_B0B3_POWER_UP | LVDS_CLKB_POWER_UP;
-	else
-		temp &= ~(LVDS_B0B3_POWER_UP | LVDS_CLKB_POWER_UP);
-
-	/* It would be nice to set 24 vs 18-bit mode (LVDS_A3_POWER_UP)
-	 * appropriately here, but we need to look more thoroughly into how
-	 * panels behave in the two modes.
-	 */
-	/* set the dithering flag on LVDS as needed */
-	if (INTEL_INFO(dev)->gen >= 4) {
-		if (dev_priv->lvds_dither)
-			temp |= LVDS_ENABLE_DITHER;
-		else
-			temp &= ~LVDS_ENABLE_DITHER;
-	}
-	temp &= ~(LVDS_HSYNC_POLARITY | LVDS_VSYNC_POLARITY);
-	if (adjusted_mode->flags & DRM_MODE_FLAG_NHSYNC)
-		temp |= LVDS_HSYNC_POLARITY;
-	if (adjusted_mode->flags & DRM_MODE_FLAG_NVSYNC)
-		temp |= LVDS_VSYNC_POLARITY;
-	I915_WRITE(LVDS, temp);
-}
-
 static void vlv_update_pll(struct drm_crtc *crtc,
 			   struct drm_display_mode *mode,
 			   struct drm_display_mode *adjusted_mode,
@@ -4484,13 +4439,6 @@ static void i9xx_update_pll(struct drm_crtc *crtc,
 		if (encoder->pre_pll_enable)
 			encoder->pre_pll_enable(encoder);
 
-	/* The LVDS pin pair needs to be on before the DPLLs are enabled.
-	 * This is an exception to the general rule that mode_set doesn't turn
-	 * things on.
-	 */
-	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS))
-		intel_update_lvds(crtc, clock, adjusted_mode);
-
 	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT))
 		intel_dp_set_m_n(crtc, mode, adjusted_mode);
 
@@ -4566,13 +4514,6 @@ static void i8xx_update_pll(struct drm_crtc *crtc,
 		if (encoder->pre_pll_enable)
 			encoder->pre_pll_enable(encoder);
 
-	/* The LVDS pin pair needs to be on before the DPLLs are enabled.
-	 * This is an exception to the general rule that mode_set doesn't turn
-	 * things on.
-	 */
-	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS))
-		intel_update_lvds(crtc, clock, adjusted_mode);
-
 	I915_WRITE(DPLL(pipe), dpll);
 
 	/* Wait for the clocks to stabilize. */
diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c
index 6b6ed64..c93ec03 100644
--- a/drivers/gpu/drm/i915/intel_lvds.c
+++ b/drivers/gpu/drm/i915/intel_lvds.c
@@ -89,6 +89,62 @@ static bool intel_lvds_get_hw_state(struct intel_encoder *encoder,
 	return true;
 }
 
+/* The LVDS pin pair needs to be on before the DPLLs are enabled.
+ * This is an exception to the general rule that mode_set doesn't turn
+ * things on.
+ */
+static void intel_pre_pll_enable_lvds(struct intel_encoder *encoder)
+{
+	struct intel_lvds_encoder *lvds_encoder = to_lvds_encoder(&encoder->base);
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->base.crtc);
+	struct drm_display_mode *fixed_mode =
+		lvds_encoder->attached_connector->base.panel.fixed_mode;
+	int pipe = intel_crtc->pipe;
+	u32 temp;
+
+	/* pch split platforms are not yet converted. */
+	if (HAS_PCH_SPLIT(dev))
+		return;
+
+	temp = I915_READ(lvds_encoder->reg);
+	temp |= LVDS_PORT_EN | LVDS_A0A2_CLKA_POWER_UP;
+	if (pipe == 1) {
+		temp |= LVDS_PIPEB_SELECT;
+	} else {
+		temp &= ~LVDS_PIPEB_SELECT;
+	}
+	/* set the corresponsding LVDS_BORDER bit */
+	temp |= dev_priv->lvds_border_bits;
+	/* Set the B0-B3 data pairs corresponding to whether we're going to
+	 * set the DPLLs for dual-channel mode or not.
+	 */
+	if (lvds_encoder->is_dual_link)
+		temp |= LVDS_B0B3_POWER_UP | LVDS_CLKB_POWER_UP;
+	else
+		temp &= ~(LVDS_B0B3_POWER_UP | LVDS_CLKB_POWER_UP);
+
+	/* It would be nice to set 24 vs 18-bit mode (LVDS_A3_POWER_UP)
+	 * appropriately here, but we need to look more thoroughly into how
+	 * panels behave in the two modes.
+	 */
+	/* set the dithering flag on LVDS as needed */
+	if (INTEL_INFO(dev)->gen >= 4) {
+		if (dev_priv->lvds_dither)
+			temp |= LVDS_ENABLE_DITHER;
+		else
+			temp &= ~LVDS_ENABLE_DITHER;
+	}
+	temp &= ~(LVDS_HSYNC_POLARITY | LVDS_VSYNC_POLARITY);
+	if (fixed_mode->flags & DRM_MODE_FLAG_NHSYNC)
+		temp |= LVDS_HSYNC_POLARITY;
+	if (fixed_mode->flags & DRM_MODE_FLAG_NVSYNC)
+		temp |= LVDS_VSYNC_POLARITY;
+
+	I915_WRITE(lvds_encoder->reg, temp);
+}
+
 /**
  * Sets the power state for the panel.
  */
@@ -1041,6 +1097,7 @@ bool intel_lvds_init(struct drm_device *dev)
 			 DRM_MODE_ENCODER_LVDS);
 
 	intel_encoder->enable = intel_enable_lvds;
+	intel_encoder->pre_pll_enable = intel_pre_pll_enable_lvds;
 	intel_encoder->disable = intel_disable_lvds;
 	intel_encoder->get_hw_state = intel_lvds_get_hw_state;
 	intel_connector->get_hw_state = intel_connector_get_hw_state;
-- 
cgit v0.10.2


From 62810e5a9df3f47b7261e5d78fc1c33e550f2171 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Mon, 26 Nov 2012 17:22:13 +0100
Subject: drm/i915: enable intel_lvds->pre_pll_enable for ilk+, too

Only two things needed adjustment:
- pipe select for PCH_CPT
- There's no dithering bit on ilk+ in the lvds ctl reg

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index bb080fb..5958b3b 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -5322,7 +5322,6 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 	bool ok, has_reduced_clock = false;
 	bool is_lvds = false, is_dp = false, is_cpu_edp = false;
 	struct intel_encoder *encoder;
-	u32 temp;
 	int ret;
 	bool dither, fdi_config_ok;
 
@@ -5386,45 +5385,6 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 	} else
 		intel_put_pch_pll(intel_crtc);
 
-	/* The LVDS pin pair needs to be on before the DPLLs are enabled.
-	 * This is an exception to the general rule that mode_set doesn't turn
-	 * things on.
-	 */
-	if (is_lvds) {
-		temp = I915_READ(PCH_LVDS);
-		temp |= LVDS_PORT_EN | LVDS_A0A2_CLKA_POWER_UP;
-		if (HAS_PCH_CPT(dev)) {
-			temp &= ~PORT_TRANS_SEL_MASK;
-			temp |= PORT_TRANS_SEL_CPT(pipe);
-		} else {
-			if (pipe == 1)
-				temp |= LVDS_PIPEB_SELECT;
-			else
-				temp &= ~LVDS_PIPEB_SELECT;
-		}
-
-		/* set the corresponsding LVDS_BORDER bit */
-		temp |= dev_priv->lvds_border_bits;
-		/* Set the B0-B3 data pairs corresponding to whether we're going to
-		 * set the DPLLs for dual-channel mode or not.
-		 */
-		if (clock.p2 == 7)
-			temp |= LVDS_B0B3_POWER_UP | LVDS_CLKB_POWER_UP;
-		else
-			temp &= ~(LVDS_B0B3_POWER_UP | LVDS_CLKB_POWER_UP);
-
-		/* It would be nice to set 24 vs 18-bit mode (LVDS_A3_POWER_UP)
-		 * appropriately here, but we need to look more thoroughly into how
-		 * panels behave in the two modes.
-		 */
-		temp &= ~(LVDS_HSYNC_POLARITY | LVDS_VSYNC_POLARITY);
-		if (adjusted_mode->flags & DRM_MODE_FLAG_NHSYNC)
-			temp |= LVDS_HSYNC_POLARITY;
-		if (adjusted_mode->flags & DRM_MODE_FLAG_NVSYNC)
-			temp |= LVDS_VSYNC_POLARITY;
-		I915_WRITE(PCH_LVDS, temp);
-	}
-
 	if (is_dp && !is_cpu_edp) {
 		intel_dp_set_m_n(crtc, mode, adjusted_mode);
 	} else {
diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c
index c93ec03..7781069 100644
--- a/drivers/gpu/drm/i915/intel_lvds.c
+++ b/drivers/gpu/drm/i915/intel_lvds.c
@@ -104,17 +104,20 @@ static void intel_pre_pll_enable_lvds(struct intel_encoder *encoder)
 	int pipe = intel_crtc->pipe;
 	u32 temp;
 
-	/* pch split platforms are not yet converted. */
-	if (HAS_PCH_SPLIT(dev))
-		return;
-
 	temp = I915_READ(lvds_encoder->reg);
 	temp |= LVDS_PORT_EN | LVDS_A0A2_CLKA_POWER_UP;
-	if (pipe == 1) {
-		temp |= LVDS_PIPEB_SELECT;
+
+	if (HAS_PCH_CPT(dev)) {
+		temp &= ~PORT_TRANS_SEL_MASK;
+		temp |= PORT_TRANS_SEL_CPT(pipe);
 	} else {
-		temp &= ~LVDS_PIPEB_SELECT;
+		if (pipe == 1) {
+			temp |= LVDS_PIPEB_SELECT;
+		} else {
+			temp &= ~LVDS_PIPEB_SELECT;
+		}
 	}
+
 	/* set the corresponsding LVDS_BORDER bit */
 	temp |= dev_priv->lvds_border_bits;
 	/* Set the B0-B3 data pairs corresponding to whether we're going to
@@ -129,8 +132,11 @@ static void intel_pre_pll_enable_lvds(struct intel_encoder *encoder)
 	 * appropriately here, but we need to look more thoroughly into how
 	 * panels behave in the two modes.
 	 */
-	/* set the dithering flag on LVDS as needed */
-	if (INTEL_INFO(dev)->gen >= 4) {
+
+	/* Set the dithering flag on LVDS as needed, note that there is no
+	 * special lvds dither control bit on pch-split platforms, dithering is
+	 * only controlled through the PIPECONF reg. */
+	if (INTEL_INFO(dev)->gen == 4) {
 		if (dev_priv->lvds_dither)
 			temp |= LVDS_ENABLE_DITHER;
 		else
-- 
cgit v0.10.2


From a39a68054f63da0ea3b4806e1bfad79670a93d9f Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 15 Nov 2012 15:40:05 +0100
Subject: drm/i915: simplify shmem pwrite/pread slowpath handling

The shmem paths for pwrite/pread used a clever trick to hold onto a
single page when dropping the big dev->struct_mutex for the slowpath.
But this ran the risk of reinstating (or not completely purging) the
backing storage when dropping purgeable objects.

Hence the code needed to keep track of whether it ever dropped the
lock, and if it did, manually check whether it needs to re-purge the
backing storage. But thanks to the pages pin count introduced in

commit a5570178c059cec59e9835be20bc8546377fa7b5
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Tue Sep 4 21:02:54 2012 +0100

    drm/i915: Pin backing pages whilst exporting through a dmabuf vmap

which allowed us to pin the backing storage and remove that page
reference trick from shmem_pwrite/read in

commit f60d7f0c1d55a935475ab394955cafddefaa6533
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Tue Sep 4 21:02:56 2012 +0100

    drm/i915: Pin backing pages for pread

and

commit 755d22184f1e5015b040acee794542d9cf8a16c5
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Tue Sep 4 21:02:55 2012 +0100

    drm/i915: Pin backing pages for pwrite

we can now abolish this check. The slowpath cleanup completely
disappears from pread, and for pwrite we're only left with the domain
fixup in case someone moved the object out of the cpu domain from
under us. A follow-on patch will optimize that a notch more.

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 4ef8b57..36f629a 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -407,7 +407,6 @@ i915_gem_shmem_pread(struct drm_device *dev,
 	loff_t offset;
 	int shmem_page_offset, page_length, ret = 0;
 	int obj_do_bit17_swizzling, page_do_bit17_swizzling;
-	int hit_slowpath = 0;
 	int prefaulted = 0;
 	int needs_clflush = 0;
 	struct scatterlist *sg;
@@ -469,7 +468,6 @@ i915_gem_shmem_pread(struct drm_device *dev,
 		if (ret == 0)
 			goto next_page;
 
-		hit_slowpath = 1;
 		mutex_unlock(&dev->struct_mutex);
 
 		if (!prefaulted) {
@@ -502,12 +500,6 @@ next_page:
 out:
 	i915_gem_object_unpin_pages(obj);
 
-	if (hit_slowpath) {
-		/* Fixup: Kill any reinstated backing storage pages */
-		if (obj->madv == __I915_MADV_PURGED)
-			i915_gem_object_truncate(obj);
-	}
-
 	return ret;
 }
 
@@ -838,11 +830,8 @@ out:
 	i915_gem_object_unpin_pages(obj);
 
 	if (hit_slowpath) {
-		/* Fixup: Kill any reinstated backing storage pages */
-		if (obj->madv == __I915_MADV_PURGED)
-			i915_gem_object_truncate(obj);
-		/* and flush dirty cachelines in case the object isn't in the cpu write
-		 * domain anymore. */
+		/* Fixup: Flush dirty cachelines in case the object isn't in the
+		 * cpu write domain anymore. */
 		if (obj->base.write_domain != I915_GEM_DOMAIN_CPU) {
 			i915_gem_clflush_object(obj);
 			i915_gem_chipset_flush(dev);
-- 
cgit v0.10.2


From 8dcf015eb967c718962c0690330d9a94d56f2c5d Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 15 Nov 2012 16:53:58 +0100
Subject: drm/i915: optimize the shmem_pwrite slowpath handling

Since we drop dev->struct_mutex when going through the slowpath, the
object might have been moved out of the cpu domain. Hence we need to
clflush the entire object to ensure that after the ioctl returns,
everything is coherent again (interwoven writes are ill-defined
anyway).

But we only need to do this if we start in the cpu domain and the
object requires flushing for coherency. So don't do the flushing if
the object is coherent anyway or if we've done in-line clfushing
already.

v2: i915_gem_clflush_object already checks whether the object is
coherent and if so, drops the flushing. Hence we don't need to check
that ourselves, simplifying the condition.

v3: Reorder the checks for better clarity (and adjust the comment
accordingly), suggested by Chris Wilson.

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 36f629a..5c8df57 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -830,9 +830,13 @@ out:
 	i915_gem_object_unpin_pages(obj);
 
 	if (hit_slowpath) {
-		/* Fixup: Flush dirty cachelines in case the object isn't in the
-		 * cpu write domain anymore. */
-		if (obj->base.write_domain != I915_GEM_DOMAIN_CPU) {
+		/*
+		 * Fixup: Flush cpu caches in case we didn't flush the dirty
+		 * cachelines in-line while writing and the object moved
+		 * out of the cpu write domain while we've dropped the lock.
+		 */
+		if (!needs_clflush_after &&
+		    obj->base.write_domain != I915_GEM_DOMAIN_CPU) {
 			i915_gem_clflush_object(obj);
 			i915_gem_chipset_flush(dev);
 		}
-- 
cgit v0.10.2


From c1f63f9d688e68088ef743a48a3aa8e0dc01a160 Mon Sep 17 00:00:00 2001
From: Paulo Zanoni <paulo.r.zanoni@intel.com>
Date: Fri, 23 Nov 2012 15:30:37 -0200
Subject: drm/i915: intel_prepare_ddi_buffers should be static

It's not even declared on header files.

Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c
index 852012b..16d3e78 100644
--- a/drivers/gpu/drm/i915/intel_ddi.c
+++ b/drivers/gpu/drm/i915/intel_ddi.c
@@ -84,7 +84,8 @@ static enum port intel_ddi_get_encoder_port(struct intel_encoder *intel_encoder)
  * in either FDI or DP modes only, as HDMI connections will work with both
  * of those
  */
-void intel_prepare_ddi_buffers(struct drm_device *dev, enum port port, bool use_fdi_mode)
+static void intel_prepare_ddi_buffers(struct drm_device *dev, enum port port,
+				      bool use_fdi_mode)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 reg;
-- 
cgit v0.10.2


From 20749730e39bba1c6100bec0e0d1a45c99db559e Mon Sep 17 00:00:00 2001
From: Paulo Zanoni <paulo.r.zanoni@intel.com>
Date: Fri, 23 Nov 2012 15:30:38 -0200
Subject: drm/i915: remove Haswell code from ironlake_fdi_pll_enable

This function is not called on Haswell anymore.

Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 5958b3b..511b397 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -2787,18 +2787,14 @@ static void ironlake_fdi_pll_enable(struct intel_crtc *intel_crtc)
 	POSTING_READ(reg);
 	udelay(200);
 
-	/* On Haswell, the PLL configuration for ports and pipes is handled
-	 * separately, as part of DDI setup */
-	if (!IS_HASWELL(dev)) {
-		/* Enable CPU FDI TX PLL, always on for Ironlake */
-		reg = FDI_TX_CTL(pipe);
-		temp = I915_READ(reg);
-		if ((temp & FDI_TX_PLL_ENABLE) == 0) {
-			I915_WRITE(reg, temp | FDI_TX_PLL_ENABLE);
+	/* Enable CPU FDI TX PLL, always on for Ironlake */
+	reg = FDI_TX_CTL(pipe);
+	temp = I915_READ(reg);
+	if ((temp & FDI_TX_PLL_ENABLE) == 0) {
+		I915_WRITE(reg, temp | FDI_TX_PLL_ENABLE);
 
-			POSTING_READ(reg);
-			udelay(100);
-		}
+		POSTING_READ(reg);
+		udelay(100);
 	}
 }
 
-- 
cgit v0.10.2


From affa935440733a79c5a9eb0e5357e2564ca4b355 Mon Sep 17 00:00:00 2001
From: Paulo Zanoni <paulo.r.zanoni@intel.com>
Date: Fri, 23 Nov 2012 15:30:39 -0200
Subject: drm/i915: add HAS_DDI check

And use it whenever we call code that uses the DDIs. We already have
intel_ddi.c and prefix every function with intel_ddi_something instead of
haswell_something, so I think replacing the checks with HAS_DDI makes more
sense. Just a cosmetical change, yes I know, but I have this OCD...

Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 9be7efd..e908c49 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -1208,6 +1208,8 @@ struct drm_i915_file_private {
 
 #define HAS_PIPE_CONTROL(dev) (INTEL_INFO(dev)->gen >= 5)
 
+#define HAS_DDI(dev)		(IS_HASWELL(dev))
+
 #define INTEL_PCH_DEVICE_ID_MASK		0xff00
 #define INTEL_PCH_IBX_DEVICE_ID_TYPE		0x3b00
 #define INTEL_PCH_CPT_DEVICE_ID_TYPE		0x1c00
diff --git a/drivers/gpu/drm/i915/intel_crt.c b/drivers/gpu/drm/i915/intel_crt.c
index 5c77743..bc07b3f 100644
--- a/drivers/gpu/drm/i915/intel_crt.c
+++ b/drivers/gpu/drm/i915/intel_crt.c
@@ -771,7 +771,7 @@ void intel_crt_init(struct drm_device *dev)
 
 	crt->base.disable = intel_disable_crt;
 	crt->base.enable = intel_enable_crt;
-	if (IS_HASWELL(dev))
+	if (HAS_DDI(dev))
 		crt->base.get_hw_state = intel_ddi_get_hw_state;
 	else
 		crt->base.get_hw_state = intel_crt_get_hw_state;
diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c
index 16d3e78..197dede 100644
--- a/drivers/gpu/drm/i915/intel_ddi.c
+++ b/drivers/gpu/drm/i915/intel_ddi.c
@@ -115,7 +115,7 @@ void intel_prepare_ddi(struct drm_device *dev)
 {
 	int port;
 
-	if (IS_HASWELL(dev)) {
+	if (HAS_DDI(dev)) {
 		for (port = PORT_A; port < PORT_E; port++)
 			intel_prepare_ddi_buffers(dev, port, false);
 
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 511b397..52b6b0e 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -1072,8 +1072,8 @@ static void assert_fdi_tx(struct drm_i915_private *dev_priv,
 	enum transcoder cpu_transcoder = intel_pipe_to_cpu_transcoder(dev_priv,
 								      pipe);
 
-	if (IS_HASWELL(dev_priv->dev)) {
-		/* On Haswell, DDI is used instead of FDI_TX_CTL */
+	if (HAS_DDI(dev_priv->dev)) {
+		/* DDI does not have a specific FDI_TX register */
 		reg = TRANS_DDI_FUNC_CTL(cpu_transcoder);
 		val = I915_READ(reg);
 		cur_state = !!(val & TRANS_DDI_FUNC_ENABLE);
@@ -1117,7 +1117,7 @@ static void assert_fdi_tx_pll_enabled(struct drm_i915_private *dev_priv,
 		return;
 
 	/* On Haswell, DDI ports are responsible for the FDI PLL setup */
-	if (IS_HASWELL(dev_priv->dev))
+	if (HAS_DDI(dev_priv->dev))
 		return;
 
 	reg = FDI_TX_CTL(pipe);
@@ -7976,7 +7976,7 @@ static const struct drm_crtc_funcs intel_crtc_funcs = {
 
 static void intel_cpu_pll_init(struct drm_device *dev)
 {
-	if (IS_HASWELL(dev))
+	if (HAS_DDI(dev))
 		intel_ddi_pll_init(dev);
 }
 
@@ -8112,11 +8112,10 @@ static void intel_setup_outputs(struct drm_device *dev)
 		I915_WRITE(PFIT_CONTROL, 0);
 	}
 
-	if (!(IS_HASWELL(dev) &&
-	      (I915_READ(DDI_BUF_CTL(PORT_A)) & DDI_A_4_LANES)))
+	if (!(HAS_DDI(dev) && (I915_READ(DDI_BUF_CTL(PORT_A)) & DDI_A_4_LANES)))
 		intel_crt_init(dev);
 
-	if (IS_HASWELL(dev)) {
+	if (HAS_DDI(dev)) {
 		int found;
 
 		/* Haswell uses DDI functions to detect digital outputs */
@@ -8360,7 +8359,7 @@ static void intel_init_display(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
 	/* We always want a DPMS function */
-	if (IS_HASWELL(dev)) {
+	if (HAS_DDI(dev)) {
 		dev_priv->display.crtc_mode_set = haswell_crtc_mode_set;
 		dev_priv->display.crtc_enable = haswell_crtc_enable;
 		dev_priv->display.crtc_disable = haswell_crtc_disable;
@@ -8849,7 +8848,7 @@ void intel_modeset_setup_hw_state(struct drm_device *dev,
 	struct intel_encoder *encoder;
 	struct intel_connector *connector;
 
-	if (IS_HASWELL(dev)) {
+	if (HAS_DDI(dev)) {
 		tmp = I915_READ(TRANS_DDI_FUNC_CTL(TRANSCODER_EDP));
 
 		if (tmp & TRANS_DDI_FUNC_ENABLE) {
@@ -8890,7 +8889,7 @@ void intel_modeset_setup_hw_state(struct drm_device *dev,
 			      crtc->active ? "enabled" : "disabled");
 	}
 
-	if (IS_HASWELL(dev))
+	if (HAS_DDI(dev))
 		intel_ddi_setup_hw_pll_state(dev);
 
 	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index d76258d..f290438 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -379,7 +379,7 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 	 * clock divider.
 	 */
 	if (is_cpu_edp(intel_dp)) {
-		if (IS_HASWELL(dev))
+		if (HAS_DDI(dev))
 			aux_clock_divider = intel_ddi_get_cdclk_freq(dev_priv) >> 1;
 		else if (IS_VALLEYVIEW(dev))
 			aux_clock_divider = 100;
@@ -1791,7 +1791,7 @@ intel_dp_start_link_train(struct intel_dp *intel_dp)
 	int voltage_tries, loop_tries;
 	uint32_t DP = intel_dp->DP;
 
-	if (IS_HASWELL(dev))
+	if (HAS_DDI(dev))
 		intel_ddi_prepare_link_retrain(encoder);
 
 	/* Write the link configuration data */
@@ -1981,7 +1981,7 @@ intel_dp_link_down(struct intel_dp *intel_dp)
 	 *   intel_ddi_prepare_link_retrain will take care of redoing the link
 	 *   train.
 	 */
-	if (IS_HASWELL(dev))
+	if (HAS_DDI(dev))
 		return;
 
 	if (WARN_ON((I915_READ(intel_dp->output_reg) & DP_PORT_EN) == 0))
@@ -2742,7 +2742,7 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 	intel_connector_attach_encoder(intel_connector, intel_encoder);
 	drm_sysfs_connector_add(connector);
 
-	if (IS_HASWELL(dev))
+	if (HAS_DDI(dev))
 		intel_connector->get_hw_state = intel_ddi_connector_get_hw_state;
 	else
 		intel_connector->get_hw_state = intel_connector_get_hw_state;
diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c
index 2ee9821..53df0a8 100644
--- a/drivers/gpu/drm/i915/intel_hdmi.c
+++ b/drivers/gpu/drm/i915/intel_hdmi.c
@@ -48,7 +48,7 @@ assert_hdmi_port_disabled(struct intel_hdmi *intel_hdmi)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	uint32_t enabled_bits;
 
-	enabled_bits = IS_HASWELL(dev) ? DDI_BUF_CTL_ENABLE : SDVO_ENABLE;
+	enabled_bits = HAS_DDI(dev) ? DDI_BUF_CTL_ENABLE : SDVO_ENABLE;
 
 	WARN(I915_READ(intel_hdmi->sdvox_reg) & enabled_bits,
 	     "HDMI port enabled, expecting disabled\n");
@@ -1013,7 +1013,7 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
 		intel_hdmi->set_infoframes = cpt_set_infoframes;
 	}
 
-	if (IS_HASWELL(dev))
+	if (HAS_DDI(dev))
 		intel_connector->get_hw_state = intel_ddi_connector_get_hw_state;
 	else
 		intel_connector->get_hw_state = intel_connector_get_hw_state;
-- 
cgit v0.10.2


From 0d536cb4e911c2496073f268e9730a9badaafa55 Mon Sep 17 00:00:00 2001
From: Paulo Zanoni <paulo.r.zanoni@intel.com>
Date: Fri, 23 Nov 2012 16:46:41 -0200
Subject: drm/i915: invert the log inside intel_prepare_ddi

Do an early return in case we don't have DDI instead of having the
whole function inside an "if" statement.

Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c
index 197dede..ad936c6 100644
--- a/drivers/gpu/drm/i915/intel_ddi.c
+++ b/drivers/gpu/drm/i915/intel_ddi.c
@@ -115,16 +115,17 @@ void intel_prepare_ddi(struct drm_device *dev)
 {
 	int port;
 
-	if (HAS_DDI(dev)) {
-		for (port = PORT_A; port < PORT_E; port++)
-			intel_prepare_ddi_buffers(dev, port, false);
+	if (!HAS_DDI(dev))
+		return;
 
-		/* DDI E is the suggested one to work in FDI mode, so program is as such by
-		 * default. It will have to be re-programmed in case a digital DP output
-		 * will be detected on it
-		 */
-		intel_prepare_ddi_buffers(dev, PORT_E, true);
-	}
+	for (port = PORT_A; port < PORT_E; port++)
+		intel_prepare_ddi_buffers(dev, port, false);
+
+	/* DDI E is the suggested one to work in FDI mode, so program is as such
+	 * by default. It will have to be re-programmed in case a digital DP
+	 * output will be detected on it
+	 */
+	intel_prepare_ddi_buffers(dev, PORT_E, true);
 }
 
 static const long hsw_ddi_buf_ctl_values[] = {
-- 
cgit v0.10.2


From 9fa5f6522e6eecb5ab20192a264a29ba4f2f4e85 Mon Sep 17 00:00:00 2001
From: Paulo Zanoni <paulo.r.zanoni@intel.com>
Date: Thu, 29 Nov 2012 11:31:29 -0200
Subject: drm/i915: kill intel_dp_link_clock()

Use drm_dp_bw_code_to_link_rate insead. It's the same thing, but
supports DP_LINK_BW_5_4 and is also used by the other drivers.

Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index f290438..b51043e 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -148,15 +148,6 @@ intel_dp_max_link_bw(struct intel_dp *intel_dp)
 	return max_link_bw;
 }
 
-static int
-intel_dp_link_clock(uint8_t link_bw)
-{
-	if (link_bw == DP_LINK_BW_2_7)
-		return 270000;
-	else
-		return 162000;
-}
-
 /*
  * The units on the numbers in the next two are... bizarre.  Examples will
  * make it clearer; this one parallels an example in the eDP spec.
@@ -191,7 +182,8 @@ intel_dp_adjust_dithering(struct intel_dp *intel_dp,
 			  struct drm_display_mode *mode,
 			  bool adjust_mode)
 {
-	int max_link_clock = intel_dp_link_clock(intel_dp_max_link_bw(intel_dp));
+	int max_link_clock =
+		drm_dp_bw_code_to_link_rate(intel_dp_max_link_bw(intel_dp));
 	int max_lanes = drm_dp_max_lane_count(intel_dp->dpcd);
 	int max_rate, mode_rate;
 
@@ -722,12 +714,15 @@ intel_dp_mode_fixup(struct drm_encoder *encoder,
 
 	for (clock = 0; clock <= max_clock; clock++) {
 		for (lane_count = 1; lane_count <= max_lane_count; lane_count <<= 1) {
-			int link_avail = intel_dp_max_data_rate(intel_dp_link_clock(bws[clock]), lane_count);
+			int link_bw_clock =
+				drm_dp_bw_code_to_link_rate(bws[clock]);
+			int link_avail = intel_dp_max_data_rate(link_bw_clock,
+								lane_count);
 
 			if (mode_rate <= link_avail) {
 				intel_dp->link_bw = bws[clock];
 				intel_dp->lane_count = lane_count;
-				adjusted_mode->clock = intel_dp_link_clock(intel_dp->link_bw);
+				adjusted_mode->clock = link_bw_clock;
 				DRM_DEBUG_KMS("DP link bw %02x lane "
 						"count %d clock %d bpp %d\n",
 				       intel_dp->link_bw, intel_dp->lane_count,
-- 
cgit v0.10.2


From 99057c81037ee45e94e5c7e6b403b73caccbaf9e Mon Sep 17 00:00:00 2001
From: "Luis R. Rodriguez" <mcgrof@do-not-panic.com>
Date: Thu, 29 Nov 2012 12:45:06 -0800
Subject: i915: convert struct spinlock to spinlock_t
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

spinlock_t should always be used.

  LD      drivers/gpu/drm/i915/built-in.o
  CHECK   drivers/gpu/drm/i915/i915_drv.c
  CC [M]  drivers/gpu/drm/i915/i915_drv.o
  CHECK   drivers/gpu/drm/i915/i915_dma.c
  CC [M]  drivers/gpu/drm/i915/i915_dma.o
  CHECK   drivers/gpu/drm/i915/i915_irq.c
  CC [M]  drivers/gpu/drm/i915/i915_irq.o
  CHECK   drivers/gpu/drm/i915/i915_debugfs.c
drivers/gpu/drm/i915/i915_debugfs.c:558:31: warning: dereference of noderef expression
drivers/gpu/drm/i915/i915_debugfs.c:558:39: warning: dereference of noderef expression
drivers/gpu/drm/i915/i915_debugfs.c:558:51: warning: dereference of noderef expression
drivers/gpu/drm/i915/i915_debugfs.c:558:63: warning: dereference of noderef expression
  CC [M]  drivers/gpu/drm/i915/i915_debugfs.o
  CHECK   drivers/gpu/drm/i915/i915_suspend.c
  CC [M]  drivers/gpu/drm/i915/i915_suspend.o
  CHECK   drivers/gpu/drm/i915/i915_gem.c
drivers/gpu/drm/i915/i915_gem.c:3703:14: warning: incorrect type in assignment (different base types)
drivers/gpu/drm/i915/i915_gem.c:3703:14:    expected unsigned int [unsigned] [usertype] mask
drivers/gpu/drm/i915/i915_gem.c:3703:14:    got restricted gfp_t
drivers/gpu/drm/i915/i915_gem.c:3706:22: warning: invalid assignment: &=
drivers/gpu/drm/i915/i915_gem.c:3706:22:    left side has type unsigned int
drivers/gpu/drm/i915/i915_gem.c:3706:22:    right side has type restricted gfp_t
drivers/gpu/drm/i915/i915_gem.c:3707:22: warning: invalid assignment: |=
drivers/gpu/drm/i915/i915_gem.c:3707:22:    left side has type unsigned int
drivers/gpu/drm/i915/i915_gem.c:3707:22:    right side has type restricted gfp_t
drivers/gpu/drm/i915/i915_gem.c:3711:39: warning: incorrect type in argument 2 (different base types)
drivers/gpu/drm/i915/i915_gem.c:3711:39:    expected restricted gfp_t [usertype] mask
drivers/gpu/drm/i915/i915_gem.c:3711:39:    got unsigned int [unsigned] [usertype] mask
  CC [M]  drivers/gpu/drm/i915/i915_gem.o
  CHECK   drivers/gpu/drm/i915/i915_gem_context.c
  CC [M]  drivers/gpu/drm/i915/i915_gem_context.o
  CHECK   drivers/gpu/drm/i915/i915_gem_debug.c
  CC [M]  drivers/gpu/drm/i915/i915_gem_debug.o
  CHECK   drivers/gpu/drm/i915/i915_gem_evict.c
  CC [M]  drivers/gpu/drm/i915/i915_gem_evict.o
  CHECK   drivers/gpu/drm/i915/i915_gem_execbuffer.c
  CC [M]  drivers/gpu/drm/i915/i915_gem_execbuffer.o
  CHECK   drivers/gpu/drm/i915/i915_gem_gtt.c
  CC [M]  drivers/gpu/drm/i915/i915_gem_gtt.o
  CHECK   drivers/gpu/drm/i915/i915_gem_stolen.c
  CC [M]  drivers/gpu/drm/i915/i915_gem_stolen.o
  CHECK   drivers/gpu/drm/i915/i915_gem_tiling.c
  CC [M]  drivers/gpu/drm/i915/i915_gem_tiling.o
  CHECK   drivers/gpu/drm/i915/i915_sysfs.c
  CC [M]  drivers/gpu/drm/i915/i915_sysfs.o
  CHECK   drivers/gpu/drm/i915/i915_trace_points.c
  CC [M]  drivers/gpu/drm/i915/i915_trace_points.o
  CHECK   drivers/gpu/drm/i915/intel_display.c
drivers/gpu/drm/i915/intel_display.c:1736:9: warning: mixing different enum types
drivers/gpu/drm/i915/intel_display.c:1736:9:     int enum transcoder  versus
drivers/gpu/drm/i915/intel_display.c:1736:9:     int enum pipe
drivers/gpu/drm/i915/intel_display.c:3659:48: warning: mixing different enum types
drivers/gpu/drm/i915/intel_display.c:3659:48:     int enum pipe  versus
drivers/gpu/drm/i915/intel_display.c:3659:48:     int enum transcoder
  CC [M]  drivers/gpu/drm/i915/intel_display.o
  CHECK   drivers/gpu/drm/i915/intel_crt.c
  CC [M]  drivers/gpu/drm/i915/intel_crt.o
  CHECK   drivers/gpu/drm/i915/intel_lvds.c
  CC [M]  drivers/gpu/drm/i915/intel_lvds.o
  CHECK   drivers/gpu/drm/i915/intel_bios.c
drivers/gpu/drm/i915/intel_bios.c:706:60: warning: incorrect type in initializer (different address spaces)
drivers/gpu/drm/i915/intel_bios.c:706:60:    expected struct vbt_header *vbt
drivers/gpu/drm/i915/intel_bios.c:706:60:    got void [noderef] <asn:2>*vbt
drivers/gpu/drm/i915/intel_bios.c:726:42: warning: incorrect type in argument 1 (different address spaces)
drivers/gpu/drm/i915/intel_bios.c:726:42:    expected void const *<noident>
drivers/gpu/drm/i915/intel_bios.c:726:42:    got unsigned char [noderef] [usertype] <asn:2>*
drivers/gpu/drm/i915/intel_bios.c:727:40: warning: cast removes address space of expression
drivers/gpu/drm/i915/intel_bios.c:738:24: warning: cast removes address space of expression
  CC [M]  drivers/gpu/drm/i915/intel_bios.o
  CHECK   drivers/gpu/drm/i915/intel_ddi.c
drivers/gpu/drm/i915/intel_ddi.c:87:6: warning: symbol 'intel_prepare_ddi_buffers' was not declared. Should it be static?
drivers/gpu/drm/i915/intel_ddi.c:1036:34: warning: mixing different enum types
drivers/gpu/drm/i915/intel_ddi.c:1036:34:     int enum pipe  versus
drivers/gpu/drm/i915/intel_ddi.c:1036:34:     int enum transcoder
  CC [M]  drivers/gpu/drm/i915/intel_ddi.o
drivers/gpu/drm/i915/intel_ddi.c: In function ‘intel_ddi_setup_hw_pll_state’:
drivers/gpu/drm/i915/intel_ddi.c:1129:2: warning: ‘port’ may be used uninitialized in this function [-Wmaybe-uninitialized]
drivers/gpu/drm/i915/intel_ddi.c:1111:12: note: ‘port’ was declared here
  CHECK   drivers/gpu/drm/i915/intel_dp.c
  CC [M]  drivers/gpu/drm/i915/intel_dp.o
  CHECK   drivers/gpu/drm/i915/intel_hdmi.c
  CC [M]  drivers/gpu/drm/i915/intel_hdmi.o
  CHECK   drivers/gpu/drm/i915/intel_sdvo.c
  CC [M]  drivers/gpu/drm/i915/intel_sdvo.o
  CHECK   drivers/gpu/drm/i915/intel_modes.c
  CC [M]  drivers/gpu/drm/i915/intel_modes.o
  CHECK   drivers/gpu/drm/i915/intel_panel.c
  CC [M]  drivers/gpu/drm/i915/intel_panel.o
  CHECK   drivers/gpu/drm/i915/intel_pm.c
drivers/gpu/drm/i915/intel_pm.c:2173:1: warning: symbol 'mchdev_lock' was not declared. Should it be static?
  CC [M]  drivers/gpu/drm/i915/intel_pm.o
  CHECK   drivers/gpu/drm/i915/intel_i2c.c
  CC [M]  drivers/gpu/drm/i915/intel_i2c.o
  CHECK   drivers/gpu/drm/i915/intel_fb.c
  CC [M]  drivers/gpu/drm/i915/intel_fb.o
  CHECK   drivers/gpu/drm/i915/intel_tv.c
  CC [M]  drivers/gpu/drm/i915/intel_tv.o
  CHECK   drivers/gpu/drm/i915/intel_dvo.c
  CC [M]  drivers/gpu/drm/i915/intel_dvo.o
  CHECK   drivers/gpu/drm/i915/intel_ringbuffer.c
  CC [M]  drivers/gpu/drm/i915/intel_ringbuffer.o
  CHECK   drivers/gpu/drm/i915/intel_overlay.c
  CC [M]  drivers/gpu/drm/i915/intel_overlay.o
  CHECK   drivers/gpu/drm/i915/intel_sprite.c
  CC [M]  drivers/gpu/drm/i915/intel_sprite.o
  CHECK   drivers/gpu/drm/i915/intel_opregion.c
  CC [M]  drivers/gpu/drm/i915/intel_opregion.o
  CHECK   drivers/gpu/drm/i915/dvo_ch7xxx.c
  CC [M]  drivers/gpu/drm/i915/dvo_ch7xxx.o
  CHECK   drivers/gpu/drm/i915/dvo_ch7017.c
  CC [M]  drivers/gpu/drm/i915/dvo_ch7017.o
  CHECK   drivers/gpu/drm/i915/dvo_ivch.c
  CC [M]  drivers/gpu/drm/i915/dvo_ivch.o
  CHECK   drivers/gpu/drm/i915/dvo_tfp410.c
  CC [M]  drivers/gpu/drm/i915/dvo_tfp410.o
  CHECK   drivers/gpu/drm/i915/dvo_sil164.c
  CC [M]  drivers/gpu/drm/i915/dvo_sil164.o
  CHECK   drivers/gpu/drm/i915/dvo_ns2501.c
  CC [M]  drivers/gpu/drm/i915/dvo_ns2501.o
  CHECK   drivers/gpu/drm/i915/i915_gem_dmabuf.c
  CC [M]  drivers/gpu/drm/i915/i915_gem_dmabuf.o
  CHECK   drivers/gpu/drm/i915/i915_ioc32.c
  CC [M]  drivers/gpu/drm/i915/i915_ioc32.o
  CHECK   drivers/gpu/drm/i915/intel_acpi.c
  CC [M]  drivers/gpu/drm/i915/intel_acpi.o
  LD [M]  drivers/gpu/drm/i915/i915.o
  Building modules, stage 2.
  MODPOST 1 modules
  CC      drivers/gpu/drm/i915/i915.mod.o
  LD [M]  drivers/gpu/drm/i915/i915.ko

Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: intel-gfx@lists.freedesktop.org
Cc: dri-devel@lists.freedesktop.org
Reported-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index e908c49..31ab43b 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -628,7 +628,7 @@ typedef struct drm_i915_private {
 	/** forcewake_count is protected by gt_lock */
 	unsigned forcewake_count;
 	/** gt_lock is also taken in irq contexts. */
-	struct spinlock gt_lock;
+	spinlock_t gt_lock;
 
 	struct intel_gmbus gmbus[GMBUS_NUM_PORTS];
 
@@ -1128,7 +1128,7 @@ struct drm_i915_gem_request {
 
 struct drm_i915_file_private {
 	struct {
-		struct spinlock lock;
+		spinlock_t lock;
 		struct list_head request_list;
 	} mm;
 	struct idr context_idr;
-- 
cgit v0.10.2


From acd15b6cc20f85bcef9e08b6ed4f142c34791c32 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Fri, 30 Nov 2012 11:24:50 +0100
Subject: drm/i915: optimize ilk/snb irq handler

We only need to read/write the south interrupt register if the
corresponding bit is set in the north master interrupt register.
Noticed while reading our interrupt handling code.

Same optimization has already been applied on ivb in

commit 0e43406bcc1868a316eea6012a0a09d992c53521
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Wed May 9 21:45:44 2012 +0100

    drm/i915: Simplify interrupt processing for IvyBridge

    We can take advantage that the PCH_IIR is a subordinate register to
    reduce one of the required IIR reads, and that we only need to clear
    interrupts handled to reduce the writes. And by simply tidying the code
    we can reduce the line count and hopefully make it more readable.

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 6cd3dc9..26753ee 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -758,7 +758,7 @@ static irqreturn_t ironlake_irq_handler(int irq, void *arg)
 	struct drm_device *dev = (struct drm_device *) arg;
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
 	int ret = IRQ_NONE;
-	u32 de_iir, gt_iir, de_ier, pch_iir, pm_iir;
+	u32 de_iir, gt_iir, de_ier, pm_iir;
 
 	atomic_inc(&dev_priv->irq_received);
 
@@ -769,11 +769,9 @@ static irqreturn_t ironlake_irq_handler(int irq, void *arg)
 
 	de_iir = I915_READ(DEIIR);
 	gt_iir = I915_READ(GTIIR);
-	pch_iir = I915_READ(SDEIIR);
 	pm_iir = I915_READ(GEN6_PMIIR);
 
-	if (de_iir == 0 && gt_iir == 0 && pch_iir == 0 &&
-	    (!IS_GEN6(dev) || pm_iir == 0))
+	if (de_iir == 0 && gt_iir == 0 && (!IS_GEN6(dev) || pm_iir == 0))
 		goto done;
 
 	ret = IRQ_HANDLED;
@@ -804,10 +802,15 @@ static irqreturn_t ironlake_irq_handler(int irq, void *arg)
 
 	/* check event from PCH */
 	if (de_iir & DE_PCH_EVENT) {
+		u32 pch_iir = I915_READ(SDEIIR);
+
 		if (HAS_PCH_CPT(dev))
 			cpt_irq_handler(dev, pch_iir);
 		else
 			ibx_irq_handler(dev, pch_iir);
+
+		/* should clear PCH hotplug event before clear CPU irq */
+		I915_WRITE(SDEIIR, pch_iir);
 	}
 
 	if (IS_GEN5(dev) &&  de_iir & DE_PCU_EVENT)
@@ -816,8 +819,6 @@ static irqreturn_t ironlake_irq_handler(int irq, void *arg)
 	if (IS_GEN6(dev) && pm_iir & GEN6_PM_DEFERRED_EVENTS)
 		gen6_queue_rps_work(dev_priv, pm_iir);
 
-	/* should clear PCH hotplug event before clear CPU irq */
-	I915_WRITE(SDEIIR, pch_iir);
 	I915_WRITE(GTIIR, gt_iir);
 	I915_WRITE(DEIIR, de_iir);
 	I915_WRITE(GEN6_PMIIR, pm_iir);
-- 
cgit v0.10.2


From 5973c7ee519e2a240c68b290a1836bdb25ed3701 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Thu, 15 Nov 2012 11:32:16 +0000
Subject: drm: Introduce drm_mm_create_block()

To be used later by i915 to preallocate exact blocks of space from the
range manager.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Dave Airlie <airlied@redhat.com>
Acked-by: Dave Airlie <airlied@redhat.com>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_mm.c b/drivers/gpu/drm/drm_mm.c
index 0761a03..bd203b6 100644
--- a/drivers/gpu/drm/drm_mm.c
+++ b/drivers/gpu/drm/drm_mm.c
@@ -161,6 +161,56 @@ static void drm_mm_insert_helper(struct drm_mm_node *hole_node,
 	}
 }
 
+struct drm_mm_node *drm_mm_create_block(struct drm_mm *mm,
+					unsigned long start,
+					unsigned long size,
+					bool atomic)
+{
+	struct drm_mm_node *hole, *node;
+	unsigned long end = start + size;
+
+	list_for_each_entry(hole, &mm->hole_stack, hole_stack) {
+		unsigned long hole_start;
+		unsigned long hole_end;
+
+		BUG_ON(!hole->hole_follows);
+		hole_start = drm_mm_hole_node_start(hole);
+		hole_end = drm_mm_hole_node_end(hole);
+
+		if (hole_start > start || hole_end < end)
+			continue;
+
+		node = drm_mm_kmalloc(mm, atomic);
+		if (unlikely(node == NULL))
+			return NULL;
+
+		node->start = start;
+		node->size = size;
+		node->mm = mm;
+		node->allocated = 1;
+
+		INIT_LIST_HEAD(&node->hole_stack);
+		list_add(&node->node_list, &hole->node_list);
+
+		if (start == hole_start) {
+			hole->hole_follows = 0;
+			list_del_init(&hole->hole_stack);
+		}
+
+		node->hole_follows = 0;
+		if (end != hole_end) {
+			list_add(&node->hole_stack, &mm->hole_stack);
+			node->hole_follows = 1;
+		}
+
+		return node;
+	}
+
+	WARN(1, "no hole found for block 0x%lx + 0x%lx\n", start, size);
+	return NULL;
+}
+EXPORT_SYMBOL(drm_mm_create_block);
+
 struct drm_mm_node *drm_mm_get_block_generic(struct drm_mm_node *hole_node,
 					     unsigned long size,
 					     unsigned alignment,
diff --git a/include/drm/drm_mm.h b/include/drm/drm_mm.h
index 06d7f79..4020f96 100644
--- a/include/drm/drm_mm.h
+++ b/include/drm/drm_mm.h
@@ -102,6 +102,10 @@ static inline bool drm_mm_initialized(struct drm_mm *mm)
 /*
  * Basic range manager support (drm_mm.c)
  */
+extern struct drm_mm_node *drm_mm_create_block(struct drm_mm *mm,
+					       unsigned long start,
+					       unsigned long size,
+					       bool atomic);
 extern struct drm_mm_node *drm_mm_get_block_generic(struct drm_mm_node *node,
 						    unsigned long size,
 						    unsigned alignment,
-- 
cgit v0.10.2


From 9e8944ab564f2e3dde90a518cd32048c58918608 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Thu, 15 Nov 2012 11:32:17 +0000
Subject: drm: Introduce an iterator over holes in the drm_mm range manager

This will be used i915 in forthcoming patches in order to measure the
largest contiguous chunk of memory available for enabling chipset
features.

v2: Try to make the macro marginally safer and more readable by not
depending upon the drm_mm_hole_node_end() being non-zero. Note that we
need to open code list_for_each() in order to update the hole_start,
hole_end variable on each iteration and keep the macro sane.

v3: Tidy up few BUG_ONs that fell foul of adding additional tests to
drm_mm_hole_node_start().

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Dave Airlie <airlied@redhat.com>
Acked-by: Dave Airlie <airlied@redhat.com>
Cc: dri-devel@lists.freedesktop.org
Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_mm.c b/drivers/gpu/drm/drm_mm.c
index bd203b6..b751b8e 100644
--- a/drivers/gpu/drm/drm_mm.c
+++ b/drivers/gpu/drm/drm_mm.c
@@ -102,20 +102,6 @@ int drm_mm_pre_get(struct drm_mm *mm)
 }
 EXPORT_SYMBOL(drm_mm_pre_get);
 
-static inline unsigned long drm_mm_hole_node_start(struct drm_mm_node *hole_node)
-{
-	return hole_node->start + hole_node->size;
-}
-
-static inline unsigned long drm_mm_hole_node_end(struct drm_mm_node *hole_node)
-{
-	struct drm_mm_node *next_node =
-		list_entry(hole_node->node_list.next, struct drm_mm_node,
-			   node_list);
-
-	return next_node->start;
-}
-
 static void drm_mm_insert_helper(struct drm_mm_node *hole_node,
 				 struct drm_mm_node *node,
 				 unsigned long size, unsigned alignment,
@@ -127,7 +113,7 @@ static void drm_mm_insert_helper(struct drm_mm_node *hole_node,
 	unsigned long adj_start = hole_start;
 	unsigned long adj_end = hole_end;
 
-	BUG_ON(!hole_node->hole_follows || node->allocated);
+	BUG_ON(node->allocated);
 
 	if (mm->color_adjust)
 		mm->color_adjust(hole_node, color, &adj_start, &adj_end);
@@ -155,7 +141,7 @@ static void drm_mm_insert_helper(struct drm_mm_node *hole_node,
 	BUG_ON(node->start + node->size > adj_end);
 
 	node->hole_follows = 0;
-	if (node->start + node->size < hole_end) {
+	if (__drm_mm_hole_node_start(node) < hole_end) {
 		list_add(&node->hole_stack, &mm->hole_stack);
 		node->hole_follows = 1;
 	}
@@ -168,15 +154,10 @@ struct drm_mm_node *drm_mm_create_block(struct drm_mm *mm,
 {
 	struct drm_mm_node *hole, *node;
 	unsigned long end = start + size;
+	unsigned long hole_start;
+	unsigned long hole_end;
 
-	list_for_each_entry(hole, &mm->hole_stack, hole_stack) {
-		unsigned long hole_start;
-		unsigned long hole_end;
-
-		BUG_ON(!hole->hole_follows);
-		hole_start = drm_mm_hole_node_start(hole);
-		hole_end = drm_mm_hole_node_end(hole);
-
+	drm_mm_for_each_hole(hole, mm, hole_start, hole_end) {
 		if (hole_start > start || hole_end < end)
 			continue;
 
@@ -293,7 +274,7 @@ static void drm_mm_insert_helper_range(struct drm_mm_node *hole_node,
 	BUG_ON(node->start + node->size > end);
 
 	node->hole_follows = 0;
-	if (node->start + node->size < hole_end) {
+	if (__drm_mm_hole_node_start(node) < hole_end) {
 		list_add(&node->hole_stack, &mm->hole_stack);
 		node->hole_follows = 1;
 	}
@@ -358,12 +339,13 @@ void drm_mm_remove_node(struct drm_mm_node *node)
 	    list_entry(node->node_list.prev, struct drm_mm_node, node_list);
 
 	if (node->hole_follows) {
-		BUG_ON(drm_mm_hole_node_start(node)
-				== drm_mm_hole_node_end(node));
+		BUG_ON(__drm_mm_hole_node_start(node) ==
+		       __drm_mm_hole_node_end(node));
 		list_del(&node->hole_stack);
 	} else
-		BUG_ON(drm_mm_hole_node_start(node)
-				!= drm_mm_hole_node_end(node));
+		BUG_ON(__drm_mm_hole_node_start(node) !=
+		       __drm_mm_hole_node_end(node));
+
 
 	if (!prev_node->hole_follows) {
 		prev_node->hole_follows = 1;
@@ -421,6 +403,8 @@ struct drm_mm_node *drm_mm_search_free_generic(const struct drm_mm *mm,
 {
 	struct drm_mm_node *entry;
 	struct drm_mm_node *best;
+	unsigned long adj_start;
+	unsigned long adj_end;
 	unsigned long best_size;
 
 	BUG_ON(mm->scanned_blocks);
@@ -428,17 +412,13 @@ struct drm_mm_node *drm_mm_search_free_generic(const struct drm_mm *mm,
 	best = NULL;
 	best_size = ~0UL;
 
-	list_for_each_entry(entry, &mm->hole_stack, hole_stack) {
-		unsigned long adj_start = drm_mm_hole_node_start(entry);
-		unsigned long adj_end = drm_mm_hole_node_end(entry);
-
+	drm_mm_for_each_hole(entry, mm, adj_start, adj_end) {
 		if (mm->color_adjust) {
 			mm->color_adjust(entry, color, &adj_start, &adj_end);
 			if (adj_end <= adj_start)
 				continue;
 		}
 
-		BUG_ON(!entry->hole_follows);
 		if (!check_free_hole(adj_start, adj_end, size, alignment))
 			continue;
 
@@ -465,6 +445,8 @@ struct drm_mm_node *drm_mm_search_free_in_range_generic(const struct drm_mm *mm,
 {
 	struct drm_mm_node *entry;
 	struct drm_mm_node *best;
+	unsigned long adj_start;
+	unsigned long adj_end;
 	unsigned long best_size;
 
 	BUG_ON(mm->scanned_blocks);
@@ -472,13 +454,11 @@ struct drm_mm_node *drm_mm_search_free_in_range_generic(const struct drm_mm *mm,
 	best = NULL;
 	best_size = ~0UL;
 
-	list_for_each_entry(entry, &mm->hole_stack, hole_stack) {
-		unsigned long adj_start = drm_mm_hole_node_start(entry) < start ?
-			start : drm_mm_hole_node_start(entry);
-		unsigned long adj_end = drm_mm_hole_node_end(entry) > end ?
-			end : drm_mm_hole_node_end(entry);
-
-		BUG_ON(!entry->hole_follows);
+	drm_mm_for_each_hole(entry, mm, adj_start, adj_end) {
+		if (adj_start < start)
+			adj_start = start;
+		if (adj_end > end)
+			adj_end = end;
 
 		if (mm->color_adjust) {
 			mm->color_adjust(entry, color, &adj_start, &adj_end);
diff --git a/include/drm/drm_mm.h b/include/drm/drm_mm.h
index 4020f96..cd45365 100644
--- a/include/drm/drm_mm.h
+++ b/include/drm/drm_mm.h
@@ -89,6 +89,29 @@ static inline bool drm_mm_initialized(struct drm_mm *mm)
 {
 	return mm->hole_stack.next;
 }
+
+static inline unsigned long __drm_mm_hole_node_start(struct drm_mm_node *hole_node)
+{
+	return hole_node->start + hole_node->size;
+}
+
+static inline unsigned long drm_mm_hole_node_start(struct drm_mm_node *hole_node)
+{
+	BUG_ON(!hole_node->hole_follows);
+	return __drm_mm_hole_node_start(hole_node);
+}
+
+static inline unsigned long __drm_mm_hole_node_end(struct drm_mm_node *hole_node)
+{
+	return list_entry(hole_node->node_list.next,
+			  struct drm_mm_node, node_list)->start;
+}
+
+static inline unsigned long drm_mm_hole_node_end(struct drm_mm_node *hole_node)
+{
+	return __drm_mm_hole_node_end(hole_node);
+}
+
 #define drm_mm_for_each_node(entry, mm) list_for_each_entry(entry, \
 						&(mm)->head_node.node_list, \
 						node_list)
@@ -99,6 +122,19 @@ static inline bool drm_mm_initialized(struct drm_mm *mm)
 	     entry != NULL; entry = next, \
 		next = entry ? list_entry(entry->node_list.next, \
 			struct drm_mm_node, node_list) : NULL) \
+
+/* Note that we need to unroll list_for_each_entry in order to inline
+ * setting hole_start and hole_end on each iteration and keep the
+ * macro sane.
+ */
+#define drm_mm_for_each_hole(entry, mm, hole_start, hole_end) \
+	for (entry = list_entry((mm)->hole_stack.next, struct drm_mm_node, hole_stack); \
+	     &entry->hole_stack != &(mm)->hole_stack ? \
+	     hole_start = drm_mm_hole_node_start(entry), \
+	     hole_end = drm_mm_hole_node_end(entry), \
+	     1 : 0; \
+	     entry = list_entry(entry->hole_stack.next, struct drm_mm_node, hole_stack))
+
 /*
  * Basic range manager support (drm_mm.c)
  */
-- 
cgit v0.10.2


From e12a2d53ae45a69aea499b64f75e7222cca0f12f Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Thu, 15 Nov 2012 11:32:18 +0000
Subject: drm/i915: Fix detection of base of stolen memory

The routine to query the base of stolen memory was using the wrong
registers and the wrong encodings on virtually every platform.

It was not until the G33 refresh, that a PCI config register was
introduced that explicitly said where the stolen memory was. Prior to
865G there was not even a register that said where the end of usable
low memory was and where the stolen memory began (or ended depending
upon chipset). Before then, one has to look at the BIOS memory maps to
find the Top of Memory. Alas that is not exported by arch/x86 and so we
have to resort to disabling stolen memory on gen2 for the time being.

Then SandyBridge enlarged the PCI register to a full 32-bits and change
the encoding of the address, so even though we happened to be querying
the right register, we read the wrong bits and ended up using address 0
for our stolen data, i.e. notably FBC.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 31ab43b..21360a7 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -765,6 +765,7 @@ typedef struct drm_i915_private {
 		unsigned long gtt_start;
 		unsigned long gtt_mappable_end;
 		unsigned long gtt_end;
+		unsigned long stolen_base; /* limited to low memory (32-bit) */
 
 		struct io_mapping *gtt_mapping;
 		phys_addr_t gtt_base_addr;
diff --git a/drivers/gpu/drm/i915/i915_gem_stolen.c b/drivers/gpu/drm/i915/i915_gem_stolen.c
index 8e91083..be24312 100644
--- a/drivers/gpu/drm/i915/i915_gem_stolen.c
+++ b/drivers/gpu/drm/i915/i915_gem_stolen.c
@@ -42,56 +42,50 @@
  * for is a boon.
  */
 
-#define PTE_ADDRESS_MASK		0xfffff000
-#define PTE_ADDRESS_MASK_HIGH		0x000000f0 /* i915+ */
-#define PTE_MAPPING_TYPE_UNCACHED	(0 << 1)
-#define PTE_MAPPING_TYPE_DCACHE		(1 << 1) /* i830 only */
-#define PTE_MAPPING_TYPE_CACHED		(3 << 1)
-#define PTE_MAPPING_TYPE_MASK		(3 << 1)
-#define PTE_VALID			(1 << 0)
-
-/**
- * i915_stolen_to_phys - take an offset into stolen memory and turn it into
- *                       a physical one
- * @dev: drm device
- * @offset: address to translate
- *
- * Some chip functions require allocations from stolen space and need the
- * physical address of the memory in question.
- */
-static unsigned long i915_stolen_to_phys(struct drm_device *dev, u32 offset)
+static unsigned long i915_stolen_to_physical(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct pci_dev *pdev = dev_priv->bridge_dev;
 	u32 base;
 
-#if 0
 	/* On the machines I have tested the Graphics Base of Stolen Memory
-	 * is unreliable, so compute the base by subtracting the stolen memory
-	 * from the Top of Low Usable DRAM which is where the BIOS places
-	 * the graphics stolen memory.
+	 * is unreliable, so on those compute the base by subtracting the
+	 * stolen memory from the Top of Low Usable DRAM which is where the
+	 * BIOS places the graphics stolen memory.
+	 *
+	 * On gen2, the layout is slightly different with the Graphics Segment
+	 * immediately following Top of Memory (or Top of Usable DRAM). Note
+	 * it appears that TOUD is only reported by 865g, so we just use the
+	 * top of memory as determined by the e820 probe.
+	 *
+	 * XXX gen2 requires an unavailable symbol and 945gm fails with
+	 * its value of TOLUD.
 	 */
-	if (INTEL_INFO(dev)->gen > 3 || IS_G33(dev)) {
-		/* top 32bits are reserved = 0 */
+	base = 0;
+	if (INTEL_INFO(dev)->gen >= 6) {
+		/* Read Base Data of Stolen Memory Register (BDSM) directly.
+		 * Note that there is also a MCHBAR miror at 0x1080c0 or
+		 * we could use device 2:0x5c instead.
+		*/
+		pci_read_config_dword(pdev, 0xB0, &base);
+		base &= ~4095; /* lower bits used for locking register */
+	} else if (INTEL_INFO(dev)->gen > 3 || IS_G33(dev)) {
+		/* Read Graphics Base of Stolen Memory directly */
 		pci_read_config_dword(pdev, 0xA4, &base);
-	} else {
-		/* XXX presume 8xx is the same as i915 */
-		pci_bus_read_config_dword(pdev->bus, 2, 0x5C, &base);
-	}
-#else
-	if (INTEL_INFO(dev)->gen > 3 || IS_G33(dev)) {
-		u16 val;
-		pci_read_config_word(pdev, 0xb0, &val);
-		base = val >> 4 << 20;
-	} else {
+#if 0
+	} else if (IS_GEN3(dev)) {
 		u8 val;
+		/* Stolen is immediately below Top of Low Usable DRAM */
 		pci_read_config_byte(pdev, 0x9c, &val);
 		base = val >> 3 << 27;
-	}
-	base -= dev_priv->mm.gtt->stolen_size;
+		base -= dev_priv->mm.gtt->stolen_size;
+	} else {
+		/* Stolen is immediately above Top of Memory */
+		base = max_low_pfn_mapped << PAGE_SHIFT;
 #endif
+	}
 
-	return base + offset;
+	return base;
 }
 
 static void i915_warn_stolen(struct drm_device *dev)
@@ -116,7 +110,7 @@ static void i915_setup_compression(struct drm_device *dev, int size)
 	if (!compressed_fb)
 		goto err;
 
-	cfb_base = i915_stolen_to_phys(dev, compressed_fb->start);
+	cfb_base = dev_priv->mm.stolen_base + compressed_fb->start;
 	if (!cfb_base)
 		goto err_fb;
 
@@ -129,7 +123,7 @@ static void i915_setup_compression(struct drm_device *dev, int size)
 		if (!compressed_llb)
 			goto err_fb;
 
-		ll_base = i915_stolen_to_phys(dev, compressed_llb->start);
+		ll_base = dev_priv->mm.stolen_base + compressed_llb->start;
 		if (!ll_base)
 			goto err_llb;
 	}
@@ -148,7 +142,7 @@ static void i915_setup_compression(struct drm_device *dev, int size)
 	}
 
 	DRM_DEBUG_KMS("FBC base 0x%08lx, ll base 0x%08lx, size %dM\n",
-		      cfb_base, ll_base, size >> 20);
+		      (long)cfb_base, (long)ll_base, size >> 20);
 	return;
 
 err_llb:
@@ -180,6 +174,13 @@ int i915_gem_init_stolen(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	unsigned long prealloc_size = dev_priv->mm.gtt->stolen_size;
 
+	dev_priv->mm.stolen_base = i915_stolen_to_physical(dev);
+	if (dev_priv->mm.stolen_base == 0)
+		return 0;
+
+	DRM_DEBUG_KMS("found %d bytes of stolen memory at %08lx\n",
+		      dev_priv->mm.gtt->stolen_size, dev_priv->mm.stolen_base);
+
 	/* Basic memrange allocator for stolen space */
 	drm_mm_init(&dev_priv->mm.stolen, 0, prealloc_size);
 
-- 
cgit v0.10.2


From ed2f3452677e46a4270c25c1b7fa3e060fdd501e Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Thu, 15 Nov 2012 11:32:19 +0000
Subject: drm/i915: Avoid clearing preallocated regions from the GTT

As yet we do not do any preallocation (chicken-and-egg problem), but we
may like to preserve anything already allocated by the BIOS or grub and
reuse for own purposes after initialising the driver.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 21360a7..f16101f 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -935,6 +935,8 @@ enum i915_cache_level {
 	I915_CACHE_LLC_MLC, /* gen6+, in docs at least! */
 };
 
+#define I915_GTT_RESERVED ((struct drm_mm_node *)0x1)
+
 struct drm_i915_gem_object_ops {
 	/* Interface between the GEM object and its backing storage.
 	 * get_pages() is called once prior to the use of the associated set
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index f7ac61e..cc1be53 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -531,20 +531,46 @@ void i915_gem_init_global_gtt(struct drm_device *dev,
 			      unsigned long end)
 {
 	drm_i915_private_t *dev_priv = dev->dev_private;
+	struct drm_mm_node *entry;
+	struct drm_i915_gem_object *obj;
+	unsigned long hole_start, hole_end;
 
-	/* Substract the guard page ... */
+	/* Subtract the guard page ... */
 	drm_mm_init(&dev_priv->mm.gtt_space, start, end - start - PAGE_SIZE);
 	if (!HAS_LLC(dev))
 		dev_priv->mm.gtt_space.color_adjust = i915_gtt_color_adjust;
 
+	/* Mark any preallocated objects as occupied */
+	list_for_each_entry(obj, &dev_priv->mm.bound_list, gtt_list) {
+		DRM_DEBUG_KMS("reserving preallocated space: %x + %zx\n",
+			      obj->gtt_offset, obj->base.size);
+
+		BUG_ON(obj->gtt_space != I915_GTT_RESERVED);
+		obj->gtt_space = drm_mm_create_block(&dev_priv->mm.gtt_space,
+						     obj->gtt_offset,
+						     obj->base.size,
+						     false);
+		obj->has_global_gtt_mapping = 1;
+	}
+
 	dev_priv->mm.gtt_start = start;
 	dev_priv->mm.gtt_mappable_end = mappable_end;
 	dev_priv->mm.gtt_end = end;
 	dev_priv->mm.gtt_total = end - start;
 	dev_priv->mm.mappable_gtt_total = min(end, mappable_end) - start;
 
-	/* ... but ensure that we clear the entire range. */
-	i915_ggtt_clear_range(dev, start / PAGE_SIZE, (end-start) / PAGE_SIZE);
+	/* Clear any non-preallocated blocks */
+	drm_mm_for_each_hole(entry, &dev_priv->mm.gtt_space,
+			     hole_start, hole_end) {
+		DRM_DEBUG_KMS("clearing unused GTT space: [%lx, %lx]\n",
+			      hole_start, hole_end);
+		i915_ggtt_clear_range(dev,
+				      hole_start / PAGE_SIZE,
+				      (hole_end-hole_start) / PAGE_SIZE);
+	}
+
+	/* And finally clear the reserved guard page */
+	i915_ggtt_clear_range(dev, end / PAGE_SIZE - 1, 1);
 }
 
 static int setup_scratch_page(struct drm_device *dev)
-- 
cgit v0.10.2


From 11be49eb4db24c5e971edef160afb87788cc270c Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Thu, 15 Nov 2012 11:32:20 +0000
Subject: drm/i915: Delay allocation of stolen space for FBC

As FBC is commonly disabled due to limitations of the chipset upon
output configurations, on many systems FBC is never enabled. For those
systems, it is advantageous to make use of the stolen memory for other
objects and so we defer allocation of the FBC chunk until we actually
require it. This increases the likelihood of that allocation failing,
but that in turns means that we are already taking advantage of the
stolen memory!

As well as delaying the allocation from driver initialisation until the
first use of FBC, we also return the stolen block after we finish using
it - allowing greater flexibility in our usage of stolen space. A side
effect of this is that we can then attempt to allocate only the required
amount of space (with a little slack to reduce reallocation rate and
avoid fragmentation).

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index f16101f..6969c72 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -1571,6 +1571,8 @@ int i915_gem_evict_everything(struct drm_device *dev);
 
 /* i915_gem_stolen.c */
 int i915_gem_init_stolen(struct drm_device *dev);
+int i915_gem_stolen_setup_compression(struct drm_device *dev, int size);
+void i915_gem_stolen_cleanup_compression(struct drm_device *dev);
 void i915_gem_cleanup_stolen(struct drm_device *dev);
 
 /* i915_gem_tiling.c */
diff --git a/drivers/gpu/drm/i915/i915_gem_stolen.c b/drivers/gpu/drm/i915/i915_gem_stolen.c
index be24312..10ca473 100644
--- a/drivers/gpu/drm/i915/i915_gem_stolen.c
+++ b/drivers/gpu/drm/i915/i915_gem_stolen.c
@@ -88,33 +88,27 @@ static unsigned long i915_stolen_to_physical(struct drm_device *dev)
 	return base;
 }
 
-static void i915_warn_stolen(struct drm_device *dev)
-{
-	DRM_INFO("not enough stolen space for compressed buffer, disabling\n");
-	DRM_INFO("hint: you may be able to increase stolen memory size in the BIOS to avoid this\n");
-}
-
-static void i915_setup_compression(struct drm_device *dev, int size)
+static int i915_setup_compression(struct drm_device *dev, int size)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_mm_node *compressed_fb, *uninitialized_var(compressed_llb);
-	unsigned long cfb_base;
-	unsigned long ll_base = 0;
 
-	/* Just in case the BIOS is doing something questionable. */
-	intel_disable_fbc(dev);
-
-	compressed_fb = drm_mm_search_free(&dev_priv->mm.stolen, size, 4096, 0);
+	/* Try to over-allocate to reduce reallocations and fragmentation */
+	compressed_fb = drm_mm_search_free(&dev_priv->mm.stolen,
+					   size <<= 1, 4096, 0);
+	if (!compressed_fb)
+		compressed_fb = drm_mm_search_free(&dev_priv->mm.stolen,
+						   size >>= 1, 4096, 0);
 	if (compressed_fb)
 		compressed_fb = drm_mm_get_block(compressed_fb, size, 4096);
 	if (!compressed_fb)
 		goto err;
 
-	cfb_base = dev_priv->mm.stolen_base + compressed_fb->start;
-	if (!cfb_base)
-		goto err_fb;
-
-	if (!(IS_GM45(dev) || HAS_PCH_SPLIT(dev))) {
+	if (HAS_PCH_SPLIT(dev))
+		I915_WRITE(ILK_DPFC_CB_BASE, compressed_fb->start);
+	else if (IS_GM45(dev)) {
+		I915_WRITE(DPFC_CB_BASE, compressed_fb->start);
+	} else {
 		compressed_llb = drm_mm_search_free(&dev_priv->mm.stolen,
 						    4096, 4096, 0);
 		if (compressed_llb)
@@ -123,56 +117,68 @@ static void i915_setup_compression(struct drm_device *dev, int size)
 		if (!compressed_llb)
 			goto err_fb;
 
-		ll_base = dev_priv->mm.stolen_base + compressed_llb->start;
-		if (!ll_base)
-			goto err_llb;
+		dev_priv->compressed_llb = compressed_llb;
+
+		I915_WRITE(FBC_CFB_BASE,
+			   dev_priv->mm.stolen_base + compressed_fb->start);
+		I915_WRITE(FBC_LL_BASE,
+			   dev_priv->mm.stolen_base + compressed_llb->start);
 	}
 
+	dev_priv->compressed_fb = compressed_fb;
 	dev_priv->cfb_size = size;
 
-	dev_priv->compressed_fb = compressed_fb;
-	if (HAS_PCH_SPLIT(dev))
-		I915_WRITE(ILK_DPFC_CB_BASE, compressed_fb->start);
-	else if (IS_GM45(dev)) {
-		I915_WRITE(DPFC_CB_BASE, compressed_fb->start);
-	} else {
-		I915_WRITE(FBC_CFB_BASE, cfb_base);
-		I915_WRITE(FBC_LL_BASE, ll_base);
-		dev_priv->compressed_llb = compressed_llb;
-	}
+	DRM_DEBUG_KMS("reserved %d bytes of contiguous stolen space for FBC\n",
+		      size);
 
-	DRM_DEBUG_KMS("FBC base 0x%08lx, ll base 0x%08lx, size %dM\n",
-		      (long)cfb_base, (long)ll_base, size >> 20);
-	return;
+	return 0;
 
-err_llb:
-	drm_mm_put_block(compressed_llb);
 err_fb:
 	drm_mm_put_block(compressed_fb);
 err:
-	dev_priv->no_fbc_reason = FBC_STOLEN_TOO_SMALL;
-	i915_warn_stolen(dev);
+	return -ENOSPC;
+}
+
+int i915_gem_stolen_setup_compression(struct drm_device *dev, int size)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (dev_priv->mm.stolen_base == 0)
+		return -ENODEV;
+
+	if (size < dev_priv->cfb_size)
+		return 0;
+
+	/* Release any current block */
+	i915_gem_stolen_cleanup_compression(dev);
+
+	return i915_setup_compression(dev, size);
 }
 
-static void i915_cleanup_compression(struct drm_device *dev)
+void i915_gem_stolen_cleanup_compression(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	drm_mm_put_block(dev_priv->compressed_fb);
+	if (dev_priv->cfb_size == 0)
+		return;
+
+	if (dev_priv->compressed_fb)
+		drm_mm_put_block(dev_priv->compressed_fb);
+
 	if (dev_priv->compressed_llb)
 		drm_mm_put_block(dev_priv->compressed_llb);
+
+	dev_priv->cfb_size = 0;
 }
 
 void i915_gem_cleanup_stolen(struct drm_device *dev)
 {
-	if (I915_HAS_FBC(dev) && i915_powersave)
-		i915_cleanup_compression(dev);
+	i915_gem_stolen_cleanup_compression(dev);
 }
 
 int i915_gem_init_stolen(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	unsigned long prealloc_size = dev_priv->mm.gtt->stolen_size;
 
 	dev_priv->mm.stolen_base = i915_stolen_to_physical(dev);
 	if (dev_priv->mm.stolen_base == 0)
@@ -182,21 +188,7 @@ int i915_gem_init_stolen(struct drm_device *dev)
 		      dev_priv->mm.gtt->stolen_size, dev_priv->mm.stolen_base);
 
 	/* Basic memrange allocator for stolen space */
-	drm_mm_init(&dev_priv->mm.stolen, 0, prealloc_size);
-
-	/* Try to set up FBC with a reasonable compressed buffer size */
-	if (I915_HAS_FBC(dev) && i915_powersave) {
-		int cfb_size;
-
-		/* Leave 1M for line length buffer & misc. */
-
-		/* Try to get a 32M buffer... */
-		if (prealloc_size > (36*1024*1024))
-			cfb_size = 32*1024*1024;
-		else /* fall back to 7/8 of the stolen space */
-			cfb_size = prealloc_size * 7 / 8;
-		i915_setup_compression(dev, cfb_size);
-	}
+	drm_mm_init(&dev_priv->mm.stolen, 0, dev_priv->mm.gtt->stolen_size);
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 52b6b0e..d303f2a 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -8655,6 +8655,9 @@ void intel_modeset_init(struct drm_device *dev)
 	/* Just disable it once at startup */
 	i915_disable_vga(dev);
 	intel_setup_outputs(dev);
+
+	/* Just in case the BIOS is doing something questionable. */
+	intel_disable_fbc(dev);
 }
 
 static void
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index f595b8d..abfff29 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -440,12 +440,6 @@ void intel_update_fbc(struct drm_device *dev)
 		dev_priv->no_fbc_reason = FBC_MODULE_PARAM;
 		goto out_disable;
 	}
-	if (intel_fb->obj->base.size > dev_priv->cfb_size) {
-		DRM_DEBUG_KMS("framebuffer too large, disabling "
-			      "compression\n");
-		dev_priv->no_fbc_reason = FBC_STOLEN_TOO_SMALL;
-		goto out_disable;
-	}
 	if ((crtc->mode.flags & DRM_MODE_FLAG_INTERLACE) ||
 	    (crtc->mode.flags & DRM_MODE_FLAG_DBLSCAN)) {
 		DRM_DEBUG_KMS("mode incompatible with compression, "
@@ -479,6 +473,14 @@ void intel_update_fbc(struct drm_device *dev)
 	if (in_dbg_master())
 		goto out_disable;
 
+	if (i915_gem_stolen_setup_compression(dev, intel_fb->obj->base.size)) {
+		DRM_INFO("not enough stolen space for compressed buffer (need %zd bytes), disabling\n", intel_fb->obj->base.size);
+		DRM_INFO("hint: you may be able to increase stolen memory size in the BIOS to avoid this\n");
+		DRM_DEBUG_KMS("framebuffer too large, disabling compression\n");
+		dev_priv->no_fbc_reason = FBC_STOLEN_TOO_SMALL;
+		goto out_disable;
+	}
+
 	/* If the scanout has not changed, don't modify the FBC settings.
 	 * Note that we make the fundamental assumption that the fb->obj
 	 * cannot be unpinned (and have its GTT offset and fence revoked)
@@ -526,6 +528,7 @@ out_disable:
 		DRM_DEBUG_KMS("unsupported config, disabling FBC\n");
 		intel_disable_fbc(dev);
 	}
+	i915_gem_stolen_cleanup_compression(dev);
 }
 
 static void i915_pineview_get_mem_freq(struct drm_device *dev)
-- 
cgit v0.10.2


From c1ad11fce86405c955873974f58ab305d894be78 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Thu, 15 Nov 2012 11:32:21 +0000
Subject: drm/i915: Allow objects to be created with no backing pages, but
 stolen space

In order to accommodate objects that are not backed by struct pages, but
instead point into a contiguous region of stolen space, we need to make
various changes to avoid dereferencing obj->pages or obj->base.filp.

First introduce a marker for the stolen object, that specifies its
offset into the stolen region and implies that it has no backing pages.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index 62619e3..58e6676 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -124,6 +124,8 @@ describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj)
 	if (obj->gtt_space != NULL)
 		seq_printf(m, " (gtt offset: %08x, size: %08x)",
 			   obj->gtt_offset, (unsigned int)obj->gtt_space->size);
+	if (obj->stolen)
+		seq_printf(m, " (stolen: %08lx)", obj->stolen->start);
 	if (obj->pin_mappable || obj->fault_mappable) {
 		char s[3], *t = s;
 		if (obj->pin_mappable)
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 6969c72..1646d03 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -962,6 +962,8 @@ struct drm_i915_gem_object {
 
 	/** Current space allocated to this object in the GTT, if any. */
 	struct drm_mm_node *gtt_space;
+	/** Stolen memory for this object, instead of being backed by shmem. */
+	struct drm_mm_node *stolen;
 	struct list_head gtt_list;
 
 	/** This object's place on the active/inactive lists */
-- 
cgit v0.10.2


From 960e3564bfa464f92549383d41659f2aaeee1420 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Thu, 15 Nov 2012 11:32:23 +0000
Subject: drm/i915: Support readback of stolen objects upon error

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 26753ee..7a0ddee 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -929,6 +929,14 @@ i915_error_object_create(struct drm_i915_private *dev_priv,
 						     reloc_offset);
 			memcpy_fromio(d, s, PAGE_SIZE);
 			io_mapping_unmap_atomic(s);
+		} else if (src->stolen) {
+			unsigned long offset;
+
+			offset = dev_priv->mm.stolen_base;
+			offset += src->stolen->start;
+			offset += i << PAGE_SHIFT;
+
+			memcpy_fromio(d, (void *)offset, PAGE_SIZE);
 		} else {
 			struct page *page;
 			void *s;
-- 
cgit v0.10.2


From 0104fdbb84d7adb0e377ed05bf75eba97b007544 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Thu, 15 Nov 2012 11:32:26 +0000
Subject: drm/i915: Introduce i915_gem_object_create_stolen()

Allow for the creation of GEM objects backed by stolen memory. As these
are not backed by ordinary pages, we create a fake dma mapping and store
the address in the scatterlist rather than obj->pages.

v2: Mark _i915_gem_object_create_stolen() as static, as noticed by Jesse
Barnes.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 1646d03..d1d68f0 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -1576,6 +1576,9 @@ int i915_gem_init_stolen(struct drm_device *dev);
 int i915_gem_stolen_setup_compression(struct drm_device *dev, int size);
 void i915_gem_stolen_cleanup_compression(struct drm_device *dev);
 void i915_gem_cleanup_stolen(struct drm_device *dev);
+struct drm_i915_gem_object *
+i915_gem_object_create_stolen(struct drm_device *dev, u32 size);
+void i915_gem_object_release_stolen(struct drm_i915_gem_object *obj);
 
 /* i915_gem_tiling.c */
 void i915_gem_detect_bit_6_swizzle(struct drm_device *dev);
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 5c8df57..3de62b0 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -3772,6 +3772,7 @@ void i915_gem_free_object(struct drm_gem_object *gem_obj)
 	obj->pages_pin_count = 0;
 	i915_gem_object_put_pages(obj);
 	i915_gem_object_free_mmap_offset(obj);
+	i915_gem_object_release_stolen(obj);
 
 	BUG_ON(obj->pages);
 
diff --git a/drivers/gpu/drm/i915/i915_gem_stolen.c b/drivers/gpu/drm/i915/i915_gem_stolen.c
index 10ca473..7299d63 100644
--- a/drivers/gpu/drm/i915/i915_gem_stolen.c
+++ b/drivers/gpu/drm/i915/i915_gem_stolen.c
@@ -192,3 +192,128 @@ int i915_gem_init_stolen(struct drm_device *dev)
 
 	return 0;
 }
+
+static struct sg_table *
+i915_pages_create_for_stolen(struct drm_device *dev,
+			     u32 offset, u32 size)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct sg_table *st;
+	struct scatterlist *sg;
+
+	DRM_DEBUG_DRIVER("offset=0x%x, size=%d\n", offset, size);
+	BUG_ON(offset > dev_priv->mm.gtt->stolen_size - size);
+
+	/* We hide that we have no struct page backing our stolen object
+	 * by wrapping the contiguous physical allocation with a fake
+	 * dma mapping in a single scatterlist.
+	 */
+
+	st = kmalloc(sizeof(*st), GFP_KERNEL);
+	if (st == NULL)
+		return NULL;
+
+	if (sg_alloc_table(st, 1, GFP_KERNEL)) {
+		kfree(st);
+		return NULL;
+	}
+
+	sg = st->sgl;
+	sg->offset = offset;
+	sg->length = size;
+
+	sg_dma_address(sg) = (dma_addr_t)dev_priv->mm.stolen_base + offset;
+	sg_dma_len(sg) = size;
+
+	return st;
+}
+
+static int i915_gem_object_get_pages_stolen(struct drm_i915_gem_object *obj)
+{
+	BUG();
+	return -EINVAL;
+}
+
+static void i915_gem_object_put_pages_stolen(struct drm_i915_gem_object *obj)
+{
+	/* Should only be called during free */
+	sg_free_table(obj->pages);
+	kfree(obj->pages);
+}
+
+static const struct drm_i915_gem_object_ops i915_gem_object_stolen_ops = {
+	.get_pages = i915_gem_object_get_pages_stolen,
+	.put_pages = i915_gem_object_put_pages_stolen,
+};
+
+static struct drm_i915_gem_object *
+_i915_gem_object_create_stolen(struct drm_device *dev,
+			       struct drm_mm_node *stolen)
+{
+	struct drm_i915_gem_object *obj;
+
+	obj = kzalloc(sizeof(*obj), GFP_KERNEL);
+	if (obj == NULL)
+		return NULL;
+
+	if (drm_gem_private_object_init(dev, &obj->base, stolen->size))
+		goto cleanup;
+
+	i915_gem_object_init(obj, &i915_gem_object_stolen_ops);
+
+	obj->pages = i915_pages_create_for_stolen(dev,
+						  stolen->start, stolen->size);
+	if (obj->pages == NULL)
+		goto cleanup;
+
+	obj->has_dma_mapping = true;
+	obj->pages_pin_count = 1;
+	obj->stolen = stolen;
+
+	obj->base.write_domain = I915_GEM_DOMAIN_GTT;
+	obj->base.read_domains = I915_GEM_DOMAIN_GTT;
+	obj->cache_level = I915_CACHE_NONE;
+
+	return obj;
+
+cleanup:
+	kfree(obj);
+	return NULL;
+}
+
+struct drm_i915_gem_object *
+i915_gem_object_create_stolen(struct drm_device *dev, u32 size)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_gem_object *obj;
+	struct drm_mm_node *stolen;
+
+	if (dev_priv->mm.stolen_base == 0)
+		return NULL;
+
+	DRM_DEBUG_KMS("creating stolen object: size=%x\n", size);
+	if (size == 0)
+		return NULL;
+
+	stolen = drm_mm_search_free(&dev_priv->mm.stolen, size, 4096, 0);
+	if (stolen)
+		stolen = drm_mm_get_block(stolen, size, 4096);
+	if (stolen == NULL)
+		return NULL;
+
+	obj = _i915_gem_object_create_stolen(dev, stolen);
+	if (obj)
+		return obj;
+
+	drm_mm_put_block(stolen);
+	return NULL;
+}
+
+void
+i915_gem_object_release_stolen(struct drm_i915_gem_object *obj)
+{
+	if (obj->stolen) {
+		drm_mm_put_block(obj->stolen);
+		obj->stolen = NULL;
+	}
+}
-- 
cgit v0.10.2


From 0ffb0ff283cca16f72caf29c44496d83b0c291fb Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Thu, 15 Nov 2012 11:32:27 +0000
Subject: drm/i915: Allocate fbcon from stolen memory

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Acked-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_fb.c b/drivers/gpu/drm/i915/intel_fb.c
index 7b30b5c..b7773e5 100644
--- a/drivers/gpu/drm/i915/intel_fb.c
+++ b/drivers/gpu/drm/i915/intel_fb.c
@@ -83,7 +83,9 @@ static int intelfb_create(struct intel_fbdev *ifbdev,
 
 	size = mode_cmd.pitches[0] * mode_cmd.height;
 	size = ALIGN(size, PAGE_SIZE);
-	obj = i915_gem_alloc_object(dev, size);
+	obj = i915_gem_object_create_stolen(dev, size);
+	if (obj == NULL)
+		obj = i915_gem_alloc_object(dev, size);
 	if (!obj) {
 		DRM_ERROR("failed to allocate framebuffer\n");
 		ret = -ENOMEM;
-- 
cgit v0.10.2


From ebc052e0c65f84f68626e388ffa1704b3a190ef7 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Thu, 15 Nov 2012 11:32:28 +0000
Subject: drm/i915: Allocate ringbuffers from stolen memory

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Acked-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index bc7cf7c..36e1e13a 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -1113,7 +1113,11 @@ static int intel_init_ring_buffer(struct drm_device *dev,
 			return ret;
 	}
 
-	obj = i915_gem_alloc_object(dev, ring->size);
+	obj = NULL;
+	if (!HAS_LLC(dev))
+		obj = i915_gem_object_create_stolen(dev, ring->size);
+	if (obj == NULL)
+		obj = i915_gem_alloc_object(dev, ring->size);
 	if (obj == NULL) {
 		DRM_ERROR("Failed to allocate ringbuffer\n");
 		ret = -ENOMEM;
-- 
cgit v0.10.2


From 8040513870399f1cb032cb8bc805df5042fedcdf Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Thu, 15 Nov 2012 11:32:29 +0000
Subject: drm/i915: Allocate overlay registers from stolen memory

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Acked-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_overlay.c b/drivers/gpu/drm/i915/intel_overlay.c
index d7bc817..fabe0ac 100644
--- a/drivers/gpu/drm/i915/intel_overlay.c
+++ b/drivers/gpu/drm/i915/intel_overlay.c
@@ -1333,8 +1333,10 @@ void intel_setup_overlay(struct drm_device *dev)
 
 	overlay->dev = dev;
 
-	reg_bo = i915_gem_alloc_object(dev, PAGE_SIZE);
-	if (!reg_bo)
+	reg_bo = i915_gem_object_create_stolen(dev, PAGE_SIZE);
+	if (reg_bo == NULL)
+		reg_bo = i915_gem_alloc_object(dev, PAGE_SIZE);
+	if (reg_bo == NULL)
 		goto out_free;
 	overlay->reg_bo = reg_bo;
 
-- 
cgit v0.10.2


From 42dcedd4f2e715dc0313e359c8288e6397843fff Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Thu, 15 Nov 2012 11:32:30 +0000
Subject: drm/i915: Use a slab for object allocation

The primary purpose of this was to debug some use-after-free memory
corruption that was causing an OOPS inside drm/i915. As it turned out
the corruption was being caused elsewhere and i915.ko as a major user of
many objects was being hit hardest.

Indeed as we do frequent the generic kmalloc caches, dedicating one to
ourselves (or at least naming one for us depending upon the core) aids
debugging our own slab usage.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index 80ed751..2635ee6 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -1736,6 +1736,9 @@ int i915_driver_unload(struct drm_device *dev)
 
 	destroy_workqueue(dev_priv->wq);
 
+	if (dev_priv->slab)
+		kmem_cache_destroy(dev_priv->slab);
+
 	pci_dev_put(dev_priv->bridge_dev);
 	kfree(dev->dev_private);
 
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index d1d68f0..e2944e9 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -614,6 +614,7 @@ struct intel_l3_parity {
 
 typedef struct drm_i915_private {
 	struct drm_device *dev;
+	struct kmem_cache *slab;
 
 	const struct intel_device_info *info;
 
@@ -1379,12 +1380,15 @@ int i915_gem_get_aperture_ioctl(struct drm_device *dev, void *data,
 int i915_gem_wait_ioctl(struct drm_device *dev, void *data,
 			struct drm_file *file_priv);
 void i915_gem_load(struct drm_device *dev);
+void *i915_gem_object_alloc(struct drm_device *dev);
+void i915_gem_object_free(struct drm_i915_gem_object *obj);
 int i915_gem_init_object(struct drm_gem_object *obj);
 void i915_gem_object_init(struct drm_i915_gem_object *obj,
 			 const struct drm_i915_gem_object_ops *ops);
 struct drm_i915_gem_object *i915_gem_alloc_object(struct drm_device *dev,
 						  size_t size);
 void i915_gem_free_object(struct drm_gem_object *obj);
+
 int __must_check i915_gem_object_pin(struct drm_i915_gem_object *obj,
 				     uint32_t alignment,
 				     bool map_and_fenceable,
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 3de62b0..dfe7174 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -192,6 +192,18 @@ i915_gem_get_aperture_ioctl(struct drm_device *dev, void *data,
 	return 0;
 }
 
+void *i915_gem_object_alloc(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	return kmem_cache_alloc(dev_priv->slab, GFP_KERNEL | __GFP_ZERO);
+}
+
+void i915_gem_object_free(struct drm_i915_gem_object *obj)
+{
+	struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
+	kmem_cache_free(dev_priv->slab, obj);
+}
+
 static int
 i915_gem_create(struct drm_file *file,
 		struct drm_device *dev,
@@ -215,7 +227,7 @@ i915_gem_create(struct drm_file *file,
 	if (ret) {
 		drm_gem_object_release(&obj->base);
 		i915_gem_info_remove_obj(dev->dev_private, obj->base.size);
-		kfree(obj);
+		i915_gem_object_free(obj);
 		return ret;
 	}
 
@@ -3695,12 +3707,12 @@ struct drm_i915_gem_object *i915_gem_alloc_object(struct drm_device *dev,
 	struct address_space *mapping;
 	u32 mask;
 
-	obj = kzalloc(sizeof(*obj), GFP_KERNEL);
+	obj = i915_gem_object_alloc(dev);
 	if (obj == NULL)
 		return NULL;
 
 	if (drm_gem_object_init(dev, &obj->base, size) != 0) {
-		kfree(obj);
+		i915_gem_object_free(obj);
 		return NULL;
 	}
 
@@ -3783,7 +3795,7 @@ void i915_gem_free_object(struct drm_gem_object *gem_obj)
 	i915_gem_info_remove_obj(dev_priv, obj->base.size);
 
 	kfree(obj->bit_17);
-	kfree(obj);
+	i915_gem_object_free(obj);
 }
 
 int
@@ -4101,8 +4113,14 @@ init_ring_lists(struct intel_ring_buffer *ring)
 void
 i915_gem_load(struct drm_device *dev)
 {
-	int i;
 	drm_i915_private_t *dev_priv = dev->dev_private;
+	int i;
+
+	dev_priv->slab =
+		kmem_cache_create("i915_gem_object",
+				  sizeof(struct drm_i915_gem_object), 0,
+				  SLAB_HWCACHE_ALIGN,
+				  NULL);
 
 	INIT_LIST_HEAD(&dev_priv->mm.active_list);
 	INIT_LIST_HEAD(&dev_priv->mm.inactive_list);
diff --git a/drivers/gpu/drm/i915/i915_gem_dmabuf.c b/drivers/gpu/drm/i915/i915_gem_dmabuf.c
index 773ef77..defb888 100644
--- a/drivers/gpu/drm/i915/i915_gem_dmabuf.c
+++ b/drivers/gpu/drm/i915/i915_gem_dmabuf.c
@@ -276,8 +276,7 @@ struct drm_gem_object *i915_gem_prime_import(struct drm_device *dev,
 	if (IS_ERR(attach))
 		return ERR_CAST(attach);
 
-
-	obj = kzalloc(sizeof(*obj), GFP_KERNEL);
+	obj = i915_gem_object_alloc(dev);
 	if (obj == NULL) {
 		ret = -ENOMEM;
 		goto fail_detach;
@@ -285,7 +284,7 @@ struct drm_gem_object *i915_gem_prime_import(struct drm_device *dev,
 
 	ret = drm_gem_private_object_init(dev, &obj->base, dma_buf->size);
 	if (ret) {
-		kfree(obj);
+		i915_gem_object_free(obj);
 		goto fail_detach;
 	}
 
diff --git a/drivers/gpu/drm/i915/i915_gem_stolen.c b/drivers/gpu/drm/i915/i915_gem_stolen.c
index 7299d63..f817b0c 100644
--- a/drivers/gpu/drm/i915/i915_gem_stolen.c
+++ b/drivers/gpu/drm/i915/i915_gem_stolen.c
@@ -252,7 +252,7 @@ _i915_gem_object_create_stolen(struct drm_device *dev,
 {
 	struct drm_i915_gem_object *obj;
 
-	obj = kzalloc(sizeof(*obj), GFP_KERNEL);
+	obj = i915_gem_object_alloc(dev);
 	if (obj == NULL)
 		return NULL;
 
@@ -277,7 +277,7 @@ _i915_gem_object_create_stolen(struct drm_device *dev,
 	return obj;
 
 cleanup:
-	kfree(obj);
+	i915_gem_object_free(obj);
 	return NULL;
 }
 
-- 
cgit v0.10.2


From 4239ca779dbe47a310a3106d9f4cd5458014bdb6 Mon Sep 17 00:00:00 2001
From: Damien Lespiau <damien.lespiau@intel.com>
Date: Mon, 3 Dec 2012 16:26:16 +0000
Subject: drm/i915: Fix dieing -> dying typo

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index dfe7174..b68bc02 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -2818,7 +2818,7 @@ static bool i915_gem_valid_gtt_space(struct drm_device *dev,
 
 	/* On non-LLC machines we have to be careful when putting differing
 	 * types of snoopable memory together to avoid the prefetcher
-	 * crossing memory domains and dieing.
+	 * crossing memory domains and dying.
 	 */
 	if (HAS_LLC(dev))
 		return true;
-- 
cgit v0.10.2


From 2a2d548240ff5f5bd826640bdf8baa016ab6cec4 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Mon, 3 Dec 2012 11:49:06 +0000
Subject: drm/i915: Tighten the checks for invalid relocation domains

Be specific for the GPU domains so that we can detect if userspace ever
passed in an invalid combination, as well as accurately reflect the
known GPU domains when printing state.

Fixes i-g-t/gem_exec_bad_domains

References: https://bugs.freedesktop.org/show_bug.cgi?id=57826
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index e2944e9..3473298 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -83,7 +83,12 @@ enum port {
 };
 #define port_name(p) ((p) + 'A')
 
-#define I915_GEM_GPU_DOMAINS	(~(I915_GEM_DOMAIN_CPU | I915_GEM_DOMAIN_GTT))
+#define I915_GEM_GPU_DOMAINS \
+	(I915_GEM_DOMAIN_RENDER | \
+	 I915_GEM_DOMAIN_SAMPLER | \
+	 I915_GEM_DOMAIN_COMMAND | \
+	 I915_GEM_DOMAIN_INSTRUCTION | \
+	 I915_GEM_DOMAIN_VERTEX)
 
 #define for_each_pipe(p) for ((p) = 0; (p) < dev_priv->num_pipe; (p)++)
 
-- 
cgit v0.10.2


From c1f093e09c4ceb583b04d11e767bb3201812e4d2 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Mon, 3 Dec 2012 11:49:07 +0000
Subject: drm/i915: Remove check for conflicting relocation write-domains

Simply use the last write-domain set for the object in the batch,
trusting userspace to have correctly flushed the caches between usage as
a write target. This check dates back from the golden age of having only
a single operation per batch with the kernel repeating it for each
cliprect, and conflicts both with userspace trying to efficiently batch
multiple operations and with reducing the kernel overhead of relocation
processing.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 802d925..6cd3e1c 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -150,17 +150,6 @@ i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj,
 			  reloc->write_domain);
 		return ret;
 	}
-	if (unlikely(reloc->write_domain && target_obj->pending_write_domain &&
-		     reloc->write_domain != target_obj->pending_write_domain)) {
-		DRM_DEBUG("Write domain conflict: "
-			  "obj %p target %d offset %d "
-			  "new %08x old %08x\n",
-			  obj, reloc->target_handle,
-			  (int) reloc->offset,
-			  reloc->write_domain,
-			  target_obj->pending_write_domain);
-		return ret;
-	}
 
 	target_obj->pending_read_domains |= reloc->read_domains;
 	target_obj->pending_write_domain |= reloc->write_domain;
-- 
cgit v0.10.2


From 1a240d4de2ccf40de5796a4d1dbb3a0236051fc9 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 29 Nov 2012 22:18:51 +0100
Subject: drm/i915: fixup sparse warnings

- __iomem where there is none (I love how we mix these things up).
- Use gfp_t instead of an other plain type.
- Unconfuse one place about enum pipe vs enum transcoder - for the pch
  transcoder we actually use the pipe enum. Fixup the other cases
  where we assign the pipe to the cpu transcoder with explicit casts.
- Declare the mch_lock properly in a header.

There is still a decent mess in intel_bios.c about __iomem, but heck,
this is x86 and we're allowed to do that.

Makes-sparse-happy: Chris Wilson <chris@chris-wilson.co.uk>
[danvet: Use a space after the cast consistently and fix up the
newly-added cast in i915_irq.c to properly use __iomem.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index 58e6676..ce60506 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -546,11 +546,11 @@ static int i915_hws_info(struct seq_file *m, void *data)
 	struct drm_device *dev = node->minor->dev;
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	struct intel_ring_buffer *ring;
-	const volatile u32 __iomem *hws;
+	const u32 *hws;
 	int i;
 
 	ring = &dev_priv->ring[(uintptr_t)node->info_ent->data];
-	hws = (volatile u32 __iomem *)ring->status_page.page_addr;
+	hws = ring->status_page.page_addr;
 	if (hws == NULL)
 		return 0;
 
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 3473298..bb0eb0f 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -577,6 +577,9 @@ struct intel_gen6_power_mgmt {
 	struct mutex hw_lock;
 };
 
+/* defined intel_pm.c */
+extern spinlock_t mchdev_lock;
+
 struct intel_ilk_power_mgmt {
 	u8 cur_delay;
 	u8 min_delay;
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index b68bc02..cb94144 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -3705,7 +3705,7 @@ struct drm_i915_gem_object *i915_gem_alloc_object(struct drm_device *dev,
 {
 	struct drm_i915_gem_object *obj;
 	struct address_space *mapping;
-	u32 mask;
+	gfp_t mask;
 
 	obj = i915_gem_object_alloc(dev);
 	if (obj == NULL)
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 7a0ddee..68e7caf 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -300,9 +300,6 @@ static void i915_hotplug_work_func(struct work_struct *work)
 	drm_helper_hpd_irq_event(dev);
 }
 
-/* defined intel_pm.c */
-extern spinlock_t mchdev_lock;
-
 static void ironlake_handle_rps_change(struct drm_device *dev)
 {
 	drm_i915_private_t *dev_priv = dev->dev_private;
@@ -936,7 +933,7 @@ i915_error_object_create(struct drm_i915_private *dev_priv,
 			offset += src->stolen->start;
 			offset += i << PAGE_SHIFT;
 
-			memcpy_fromio(d, (void *)offset, PAGE_SIZE);
+			memcpy_fromio(d, (void __iomem *) offset, PAGE_SIZE);
 		} else {
 			struct page *page;
 			void *s;
diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c
index ad936c6..f02b3fe 100644
--- a/drivers/gpu/drm/i915/intel_ddi.c
+++ b/drivers/gpu/drm/i915/intel_ddi.c
@@ -1044,7 +1044,7 @@ bool intel_ddi_connector_get_hw_state(struct intel_connector *intel_connector)
 	if (port == PORT_A)
 		cpu_transcoder = TRANSCODER_EDP;
 	else
-		cpu_transcoder = pipe;
+		cpu_transcoder = (enum transcoder) pipe;
 
 	tmp = I915_READ(TRANS_DDI_FUNC_CTL(cpu_transcoder));
 
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index d303f2a..093a163 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -1671,7 +1671,7 @@ static void lpt_enable_pch_transcoder(struct drm_i915_private *dev_priv,
 	BUG_ON(dev_priv->info->gen < 5);
 
 	/* FDI must be feeding us bits for PCH ports */
-	assert_fdi_tx_enabled(dev_priv, cpu_transcoder);
+	assert_fdi_tx_enabled(dev_priv, (enum pipe) cpu_transcoder);
 	assert_fdi_rx_enabled(dev_priv, TRANSCODER_A);
 
 	/* Workaround: set timing override bit. */
@@ -1759,7 +1759,7 @@ static void intel_enable_pipe(struct drm_i915_private *dev_priv, enum pipe pipe,
 {
 	enum transcoder cpu_transcoder = intel_pipe_to_cpu_transcoder(dev_priv,
 								      pipe);
-	enum transcoder pch_transcoder;
+	enum pipe pch_transcoder;
 	int reg;
 	u32 val;
 
@@ -1779,7 +1779,8 @@ static void intel_enable_pipe(struct drm_i915_private *dev_priv, enum pipe pipe,
 		if (pch_port) {
 			/* if driving the PCH, we need FDI enabled */
 			assert_fdi_rx_pll_enabled(dev_priv, pch_transcoder);
-			assert_fdi_tx_pll_enabled(dev_priv, cpu_transcoder);
+			assert_fdi_tx_pll_enabled(dev_priv,
+						  (enum pipe) cpu_transcoder);
 		}
 		/* FIXME: assert CPU port conditions for SNB+ */
 	}
@@ -3598,7 +3599,7 @@ static void haswell_crtc_off(struct drm_crtc *crtc)
 
 	/* Stop saying we're using TRANSCODER_EDP because some other CRTC might
 	 * start using it. */
-	intel_crtc->cpu_transcoder = intel_crtc->pipe;
+	intel_crtc->cpu_transcoder = (enum transcoder) intel_crtc->pipe;
 
 	intel_ddi_put_crtc_pll(crtc);
 }
-- 
cgit v0.10.2


From 40633219a0ff4953601c7b2326f8faabc0171fd9 Mon Sep 17 00:00:00 2001
From: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Date: Tue, 4 Dec 2012 15:12:00 +0200
Subject: drm/i915: Add debugfs entry to read/write next_seqno

This is needed for testing seqno wrapping. Be careful to not bump next_seqno
more than 0x7FFFFFFF at a time (between some handled requests) as
i915_seqno_passed() can't handle bigger difference in between.

v2: Address review comments from Chris Wilson.

Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
[danvet: Squash in fixup to properly remove the debugfs file on driver
unload again.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index ce60506..602058f 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -843,6 +843,86 @@ static const struct file_operations i915_error_state_fops = {
 	.release = i915_error_state_release,
 };
 
+static ssize_t
+i915_next_seqno_read(struct file *filp,
+		 char __user *ubuf,
+		 size_t max,
+		 loff_t *ppos)
+{
+	struct drm_device *dev = filp->private_data;
+	drm_i915_private_t *dev_priv = dev->dev_private;
+	char buf[80];
+	int len;
+	int ret;
+
+	ret = mutex_lock_interruptible(&dev->struct_mutex);
+	if (ret)
+		return ret;
+
+	len = snprintf(buf, sizeof(buf),
+		       "next_seqno :  0x%x\n",
+		       dev_priv->next_seqno);
+
+	mutex_unlock(&dev->struct_mutex);
+
+	if (len > sizeof(buf))
+		len = sizeof(buf);
+
+	return simple_read_from_buffer(ubuf, max, ppos, buf, len);
+}
+
+static ssize_t
+i915_next_seqno_write(struct file *filp,
+		      const char __user *ubuf,
+		      size_t cnt,
+		      loff_t *ppos)
+{
+	struct drm_device *dev = filp->private_data;
+	drm_i915_private_t *dev_priv = dev->dev_private;
+	char buf[20];
+	u32 val = 1;
+	int ret;
+
+	if (cnt > 0) {
+		if (cnt > sizeof(buf) - 1)
+			return -EINVAL;
+
+		if (copy_from_user(buf, ubuf, cnt))
+			return -EFAULT;
+		buf[cnt] = 0;
+
+		ret = kstrtouint(buf, 0, &val);
+		if (ret < 0)
+			return ret;
+	}
+
+	if (val == 0)
+		return -EINVAL;
+
+	ret = mutex_lock_interruptible(&dev->struct_mutex);
+	if (ret)
+		return ret;
+
+	if (i915_seqno_passed(val, dev_priv->next_seqno)) {
+		dev_priv->next_seqno = val;
+		DRM_DEBUG_DRIVER("Advancing seqno to %u\n", val);
+	} else {
+		ret = -EINVAL;
+	}
+
+	mutex_unlock(&dev->struct_mutex);
+
+	return ret ?: cnt;
+}
+
+static const struct file_operations i915_next_seqno_fops = {
+	.owner = THIS_MODULE,
+	.open = simple_open,
+	.read = i915_next_seqno_read,
+	.write = i915_next_seqno_write,
+	.llseek = default_llseek,
+};
+
 static int i915_rstdby_delays(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
@@ -2107,6 +2187,12 @@ int i915_debugfs_init(struct drm_minor *minor)
 	if (ret)
 		return ret;
 
+	ret = i915_debugfs_create(minor->debugfs_root, minor,
+				 "i915_next_seqno",
+				 &i915_next_seqno_fops);
+	if (ret)
+		return ret;
+
 	return drm_debugfs_create_files(i915_debugfs_list,
 					I915_DEBUGFS_ENTRIES,
 					minor->debugfs_root, minor);
@@ -2130,6 +2216,8 @@ void i915_debugfs_cleanup(struct drm_minor *minor)
 				 1, minor);
 	drm_debugfs_remove_files((struct drm_info_list *) &i915_error_state_fops,
 				 1, minor);
+	drm_debugfs_remove_files((struct drm_info_list *) &i915_next_seqno_fops,
+				 1, minor);
 }
 
 #endif /* CONFIG_DEBUG_FS */
-- 
cgit v0.10.2


From 43a7b924a920729ff775e6a847f671d0ea456801 Mon Sep 17 00:00:00 2001
From: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Date: Tue, 4 Dec 2012 15:12:01 +0200
Subject: drm/i915: Fix debugfs seqno info print to use uint

seqno's are u32 so print accordingly

Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index 602058f..7e516ee 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -389,7 +389,7 @@ static void i915_ring_seqno_info(struct seq_file *m,
 				 struct intel_ring_buffer *ring)
 {
 	if (ring->get_seqno) {
-		seq_printf(m, "Current sequence (%s): %d\n",
+		seq_printf(m, "Current sequence (%s): %u\n",
 			   ring->name, ring->get_seqno(ring, false));
 	}
 }
-- 
cgit v0.10.2


From cbcc80dff3896015385c67d6be0beb3399999e5c Mon Sep 17 00:00:00 2001
From: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Date: Tue, 4 Dec 2012 15:12:03 +0200
Subject: drm/i915: Split intel_ring_begin

In preparation for handling ring seqno wrapping, split
intel_ring_begin into helper part which doesn't allocate
seqno.

Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index 36e1e13a..01a660a 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -1363,11 +1363,31 @@ intel_ring_alloc_seqno(struct intel_ring_buffer *ring)
 	return i915_gem_get_seqno(ring->dev, &ring->outstanding_lazy_request);
 }
 
+static int __intel_ring_begin(struct intel_ring_buffer *ring,
+			      int bytes)
+{
+	int ret;
+
+	if (unlikely(ring->tail + bytes > ring->effective_size)) {
+		ret = intel_wrap_ring_buffer(ring);
+		if (unlikely(ret))
+			return ret;
+	}
+
+	if (unlikely(ring->space < bytes)) {
+		ret = ring_wait_for_space(ring, bytes);
+		if (unlikely(ret))
+			return ret;
+	}
+
+	ring->space -= bytes;
+	return 0;
+}
+
 int intel_ring_begin(struct intel_ring_buffer *ring,
 		     int num_dwords)
 {
 	drm_i915_private_t *dev_priv = ring->dev->dev_private;
-	int n = 4*num_dwords;
 	int ret;
 
 	ret = i915_gem_check_wedge(dev_priv, dev_priv->mm.interruptible);
@@ -1379,20 +1399,7 @@ int intel_ring_begin(struct intel_ring_buffer *ring,
 	if (ret)
 		return ret;
 
-	if (unlikely(ring->tail + n > ring->effective_size)) {
-		ret = intel_wrap_ring_buffer(ring);
-		if (unlikely(ret))
-			return ret;
-	}
-
-	if (unlikely(ring->space < n)) {
-		ret = ring_wait_for_space(ring, n);
-		if (unlikely(ret))
-			return ret;
-	}
-
-	ring->space -= n;
-	return 0;
+	return __intel_ring_begin(ring, num_dwords * sizeof(uint32_t));
 }
 
 void intel_ring_advance(struct intel_ring_buffer *ring)
-- 
cgit v0.10.2


From 498d2ac15ce0fc08edb005a7faf9ed6b5aa028d8 Mon Sep 17 00:00:00 2001
From: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Date: Tue, 4 Dec 2012 15:12:04 +0200
Subject: drm/i915: Add intel_ring_handle_seqno wrap

If there are pre-wrap values in semaphore-mbox registers after wrap,
syncing against some after-wrap request will complete immediately.
Fix this by emitting ring commands to set mbox registers to zero
when the wrap happens.

v2: Use __intel_ring_begin to emit ring commands, from
Chris Wilson.

Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
[danvet: Add a small comment to handle_seqno_wrap.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index cb94144..02d3151 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1952,6 +1952,10 @@ i915_gem_handle_seqno_wrap(struct drm_device *dev)
 
 	i915_gem_retire_requests(dev);
 	for_each_ring(ring, dev_priv, i) {
+		ret = intel_ring_handle_seqno_wrap(ring);
+		if (ret)
+			return ret;
+
 		for (j = 0; j < ARRAY_SIZE(ring->sync_seqno); j++)
 			ring->sync_seqno[j] = 0;
 	}
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index 01a660a..0d03dc6 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -1402,6 +1402,28 @@ int intel_ring_begin(struct intel_ring_buffer *ring,
 	return __intel_ring_begin(ring, num_dwords * sizeof(uint32_t));
 }
 
+int intel_ring_handle_seqno_wrap(struct intel_ring_buffer *ring)
+{
+	int ret;
+
+	BUG_ON(ring->outstanding_lazy_request);
+
+	if (INTEL_INFO(ring->dev)->gen < 6)
+		return 0;
+
+	ret = __intel_ring_begin(ring, 6 * sizeof(uint32_t));
+	if (ret)
+		return ret;
+
+	/* Leaving a stale, pre-wrap seqno behind in the mboxes will result in
+	 * post-wrap semaphore waits completing immediately. Clear them. */
+	update_mboxes(ring, ring->signal_mbox[0]);
+	update_mboxes(ring, ring->signal_mbox[1]);
+	intel_ring_advance(ring);
+
+	return 0;
+}
+
 void intel_ring_advance(struct intel_ring_buffer *ring)
 {
 	struct drm_i915_private *dev_priv = ring->dev->dev_private;
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h
index d4b7416..b4a533e 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.h
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.h
@@ -196,7 +196,7 @@ static inline void intel_ring_emit(struct intel_ring_buffer *ring,
 }
 void intel_ring_advance(struct intel_ring_buffer *ring);
 int __must_check intel_ring_idle(struct intel_ring_buffer *ring);
-
+int __must_check intel_ring_handle_seqno_wrap(struct intel_ring_buffer *ring);
 int intel_ring_flush_all_caches(struct intel_ring_buffer *ring);
 int intel_ring_invalidate_all_caches(struct intel_ring_buffer *ring);
 
-- 
cgit v0.10.2


From 4a06e201dae98ed893bae677b7b05bff29fcbab0 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sat, 1 Dec 2012 13:53:40 +0100
Subject: drm/i915: haswell has the same irq handlers as ivb

No need to have the exaxt same code twice.

Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 68e7caf..2aa23e7 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -2711,7 +2711,7 @@ void intel_irq_init(struct drm_device *dev)
 		dev->driver->irq_uninstall = valleyview_irq_uninstall;
 		dev->driver->enable_vblank = valleyview_enable_vblank;
 		dev->driver->disable_vblank = valleyview_disable_vblank;
-	} else if (IS_IVYBRIDGE(dev)) {
+	} else if (IS_IVYBRIDGE(dev) || IS_HASWELL(dev)) {
 		/* Share pre & uninstall handlers with ILK/SNB */
 		dev->driver->irq_handler = ivybridge_irq_handler;
 		dev->driver->irq_preinstall = ironlake_irq_preinstall;
@@ -2719,14 +2719,6 @@ void intel_irq_init(struct drm_device *dev)
 		dev->driver->irq_uninstall = ironlake_irq_uninstall;
 		dev->driver->enable_vblank = ivybridge_enable_vblank;
 		dev->driver->disable_vblank = ivybridge_disable_vblank;
-	} else if (IS_HASWELL(dev)) {
-		/* Share interrupts handling with IVB */
-		dev->driver->irq_handler = ivybridge_irq_handler;
-		dev->driver->irq_preinstall = ironlake_irq_preinstall;
-		dev->driver->irq_postinstall = ivybridge_irq_postinstall;
-		dev->driver->irq_uninstall = ironlake_irq_uninstall;
-		dev->driver->enable_vblank = ivybridge_enable_vblank;
-		dev->driver->disable_vblank = ivybridge_disable_vblank;
 	} else if (HAS_PCH_SPLIT(dev)) {
 		dev->driver->irq_handler = ironlake_irq_handler;
 		dev->driver->irq_preinstall = ironlake_irq_preinstall;
-- 
cgit v0.10.2


From d83779a9cb9374977c1c05364b4d7dfe9ec68ef3 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sat, 1 Dec 2012 13:53:41 +0100
Subject: drm/i915: don't handle PIPE_LEGACY_BLC_EVENT_STATUS on vlv

This is for legacy legacy stuff, and checking with the leftover
pipe from the previous loop is propably not what we want. Since
pipe == 2 after the loop ... Then we only assing a variable and do
nothing with it.

Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 2aa23e7..935b6b7 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -530,7 +530,6 @@ static irqreturn_t valleyview_irq_handler(int irq, void *arg)
 	unsigned long irqflags;
 	int pipe;
 	u32 pipe_stats[I915_MAX_PIPES];
-	bool blc_event;
 
 	atomic_inc(&dev_priv->irq_received);
 
@@ -587,9 +586,6 @@ static irqreturn_t valleyview_irq_handler(int irq, void *arg)
 			I915_READ(PORT_HOTPLUG_STAT);
 		}
 
-		if (pipe_stats[pipe] & PIPE_LEGACY_BLC_EVENT_STATUS)
-			blc_event = true;
-
 		if (pm_iir & GEN6_PM_DEFERRED_EVENTS)
 			gen6_queue_rps_work(dev_priv, pm_iir);
 
-- 
cgit v0.10.2


From 61bac78e03d4385e225cb2837e33974feda489c2 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sat, 1 Dec 2012 21:03:21 +0100
Subject: drm/i915: setup the hangcheck timer early

... together with all the other irq related resources in
intel_irq_init. I've managed to oops in the notify_ring function on my
ilk, presumably because of the powerctx setup call to i915_gpu_idle.

Note that this is only a problem with the reorder irq setup sequence
for irq-driver gmbus/dp aux.

Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index 2635ee6..901f8fc 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -1611,9 +1611,6 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 	intel_opregion_init(dev);
 	acpi_video_register();
 
-	setup_timer(&dev_priv->hangcheck_timer, i915_hangcheck_elapsed,
-		    (unsigned long) dev);
-
 	if (IS_GEN5(dev))
 		intel_gpu_ips_init(dev_priv);
 
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 935b6b7..568820b 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -2687,6 +2687,9 @@ void intel_irq_init(struct drm_device *dev)
 	INIT_WORK(&dev_priv->rps.work, gen6_pm_rps_work);
 	INIT_WORK(&dev_priv->l3_parity.error_work, ivybridge_parity_work);
 
+	setup_timer(&dev_priv->hangcheck_timer, i915_hangcheck_elapsed,
+		    (unsigned long) dev);
+
 	dev->driver->get_vblank_counter = i915_get_vblank_counter;
 	dev->max_vblank_count = 0xffffff; /* only 24 bits of frame count */
 	if (IS_G4X(dev) || INTEL_INFO(dev)->gen >= 5) {
-- 
cgit v0.10.2


From 52d7ecedac3f96fb562cb482c139015372728638 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sat, 1 Dec 2012 21:03:22 +0100
Subject: drm/i915: reorder setup sequence to have irqs for output setup

Otherwise the new&shiny irq-driven gmbus and dp aux code won't work that
well. Noticed since the dp aux code doesn't have an automatic fallback
with a timeout (since the hw provides for that already).

v2: Simple move drm_irq_install before intel_modeset_gem_init, as
suggested by Ben Widawsky.

v3: Now that interrupts are enabled before all connectors are fully
set up, we might fall over serving a HPD interrupt while things are
still being set up. Instead of jumping through massive hoops and
complicating the code with a separate hpd irq enable step, simply
block out the hotplug work item from doing anything until things are
in place.

v4: Actually, we can enable hotplug processing only after the fbdev is
fully set up, since we call down into the fbdev from the hotplug work
functions. So stick the hpd enabling right next to the poll helper
initialization.

v5: We need to enable irqs before intel_modeset_init, since that
function sets up the outputs.

v6: Fixup cleanup sequence, too.

Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index 901f8fc..14d679a 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -1294,19 +1294,21 @@ static int i915_load_modeset_init(struct drm_device *dev)
 	if (ret)
 		goto cleanup_vga_switcheroo;
 
+	ret = drm_irq_install(dev);
+	if (ret)
+		goto cleanup_gem_stolen;
+
+	/* Important: The output setup functions called by modeset_init need
+	 * working irqs for e.g. gmbus and dp aux transfers. */
 	intel_modeset_init(dev);
 
 	ret = i915_gem_init(dev);
 	if (ret)
-		goto cleanup_gem_stolen;
-
-	intel_modeset_gem_init(dev);
+		goto cleanup_irq;
 
 	INIT_WORK(&dev_priv->console_resume_work, intel_console_resume);
 
-	ret = drm_irq_install(dev);
-	if (ret)
-		goto cleanup_gem;
+	intel_modeset_gem_init(dev);
 
 	/* Always safe in the mode setting case. */
 	/* FIXME: do pre/post-mode set stuff in core KMS code */
@@ -1314,7 +1316,10 @@ static int i915_load_modeset_init(struct drm_device *dev)
 
 	ret = intel_fbdev_init(dev);
 	if (ret)
-		goto cleanup_irq;
+		goto cleanup_gem;
+
+	/* Only enable hotplug handling once the fbdev is fully set up. */
+	dev_priv->enable_hotplug_processing = true;
 
 	drm_kms_helper_poll_init(dev);
 
@@ -1323,13 +1328,13 @@ static int i915_load_modeset_init(struct drm_device *dev)
 
 	return 0;
 
-cleanup_irq:
-	drm_irq_uninstall(dev);
 cleanup_gem:
 	mutex_lock(&dev->struct_mutex);
 	i915_gem_cleanup_ringbuffer(dev);
 	mutex_unlock(&dev->struct_mutex);
 	i915_gem_cleanup_aliasing_ppgtt(dev);
+cleanup_irq:
+	drm_irq_uninstall(dev);
 cleanup_gem_stolen:
 	i915_gem_cleanup_stolen(dev);
 cleanup_vga_switcheroo:
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index bb0eb0f..e7411f3 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -673,6 +673,7 @@ typedef struct drm_i915_private {
 
 	u32 hotplug_supported_mask;
 	struct work_struct hotplug_work;
+	bool enable_hotplug_processing;
 
 	int num_pipe;
 	int num_pch_pll;
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 568820b..02a13ab 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -287,6 +287,10 @@ static void i915_hotplug_work_func(struct work_struct *work)
 	struct drm_mode_config *mode_config = &dev->mode_config;
 	struct intel_encoder *encoder;
 
+	/* HPD irq before everything is fully set up. */
+	if (!dev_priv->enable_hotplug_processing)
+		return;
+
 	mutex_lock(&mode_config->mutex);
 	DRM_DEBUG_KMS("running encoder hotplug functions\n");
 
-- 
cgit v0.10.2


From 61168c53f5be309332006dbd967aaac501cde8b2 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sat, 1 Dec 2012 13:53:43 +0100
Subject: drm/i915: extract gmbus_wait_hw_status

The gmbus interrupt generation is rather fiddly: We can only ever
enable one interrupt source (but we always want to check for NAK
in addition to the real bit). And the bits in the gmbus status
register don't map at all to the bis in the irq register.

To prepare for this mess, start by extracting the hw status wait
loop into it's own function, consolidate the NAK error handling a
bit. To keep things flexible, pass in the status bit we care about
(in addition to any NAK signalling).

v2: I've failed to notice that the sense of GMBUS_ACTIVE is inverted,
Chris Wilson gladly pointed that out for me. To keep things simple,
ignore that case for  now (we only need to idle the gmbus controller
at the end of an entire i2c transaction, not after every message).

Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_i2c.c b/drivers/gpu/drm/i915/intel_i2c.c
index 3ef5af1..a16eecd 100644
--- a/drivers/gpu/drm/i915/intel_i2c.c
+++ b/drivers/gpu/drm/i915/intel_i2c.c
@@ -203,6 +203,24 @@ intel_gpio_setup(struct intel_gmbus *bus, u32 pin)
 }
 
 static int
+gmbus_wait_hw_status(struct drm_i915_private *dev_priv,
+		     u32 gmbus2_status)
+{
+	int ret;
+	int reg_offset = dev_priv->gpio_mmio_base;
+	u32 gmbus2;
+
+	ret = wait_for((gmbus2 = I915_READ(GMBUS2 + reg_offset)) &
+		       (GMBUS_SATOER | gmbus2_status),
+		       50);
+
+	if (gmbus2 & GMBUS_SATOER)
+		return -ENXIO;
+
+	return ret;
+}
+
+static int
 gmbus_xfer_read(struct drm_i915_private *dev_priv, struct i2c_msg *msg,
 		u32 gmbus1_index)
 {
@@ -219,15 +237,10 @@ gmbus_xfer_read(struct drm_i915_private *dev_priv, struct i2c_msg *msg,
 	while (len) {
 		int ret;
 		u32 val, loop = 0;
-		u32 gmbus2;
 
-		ret = wait_for((gmbus2 = I915_READ(GMBUS2 + reg_offset)) &
-			       (GMBUS_SATOER | GMBUS_HW_RDY),
-			       50);
+		ret = gmbus_wait_hw_status(dev_priv, GMBUS_HW_RDY);
 		if (ret)
-			return -ETIMEDOUT;
-		if (gmbus2 & GMBUS_SATOER)
-			return -ENXIO;
+			return ret;
 
 		val = I915_READ(GMBUS3 + reg_offset);
 		do {
@@ -261,7 +274,6 @@ gmbus_xfer_write(struct drm_i915_private *dev_priv, struct i2c_msg *msg)
 		   GMBUS_SLAVE_WRITE | GMBUS_SW_RDY);
 	while (len) {
 		int ret;
-		u32 gmbus2;
 
 		val = loop = 0;
 		do {
@@ -270,13 +282,9 @@ gmbus_xfer_write(struct drm_i915_private *dev_priv, struct i2c_msg *msg)
 
 		I915_WRITE(GMBUS3 + reg_offset, val);
 
-		ret = wait_for((gmbus2 = I915_READ(GMBUS2 + reg_offset)) &
-			       (GMBUS_SATOER | GMBUS_HW_RDY),
-			       50);
+		ret = gmbus_wait_hw_status(dev_priv, GMBUS_HW_RDY);
 		if (ret)
-			return -ETIMEDOUT;
-		if (gmbus2 & GMBUS_SATOER)
-			return -ENXIO;
+			return ret;
 	}
 	return 0;
 }
@@ -345,8 +353,6 @@ gmbus_xfer(struct i2c_adapter *adapter,
 	I915_WRITE(GMBUS0 + reg_offset, bus->reg0);
 
 	for (i = 0; i < num; i++) {
-		u32 gmbus2;
-
 		if (gmbus_is_index_read(msgs, i, num)) {
 			ret = gmbus_xfer_index_read(dev_priv, &msgs[i]);
 			i += 1;  /* set i to the index of the read xfer */
@@ -361,13 +367,11 @@ gmbus_xfer(struct i2c_adapter *adapter,
 		if (ret == -ENXIO)
 			goto clear_err;
 
-		ret = wait_for((gmbus2 = I915_READ(GMBUS2 + reg_offset)) &
-			       (GMBUS_SATOER | GMBUS_HW_WAIT_PHASE),
-			       50);
+		ret = gmbus_wait_hw_status(dev_priv, GMBUS_HW_WAIT_PHASE);
+		if (ret == -ENXIO)
+			goto clear_err;
 		if (ret)
 			goto timeout;
-		if (gmbus2 & GMBUS_SATOER)
-			goto clear_err;
 	}
 
 	/* Generate a STOP condition on the bus. Note that gmbus can't generata
-- 
cgit v0.10.2


From 515ac2bb95f609bc4a0d2ad5f7011b3264b2bb21 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sat, 1 Dec 2012 13:53:44 +0100
Subject: drm/i915: wire up gmbus irq handler

Only enables the interrupt and puts a irq handler into place, doesn't
do anything yet.

Unfortunately there's no gmbus interrupt support for gen2/3 (safe for
pnv, but there the irq is marked as "Test mode").

v2: Wire up the irq handler for vlv and gen4 properly.

v3: i915_enable_pipestat expects the mask bit, not the status bits ... and
for added hilarity those are rather inconsistently named.

Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 02a13ab..956b95f 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -525,6 +525,11 @@ static void gen6_queue_rps_work(struct drm_i915_private *dev_priv,
 	queue_work(dev_priv->wq, &dev_priv->rps.work);
 }
 
+static void gmbus_irq_handler(struct drm_device *dev)
+{
+	DRM_DEBUG_DRIVER("GMBUS interrupt\n");
+}
+
 static irqreturn_t valleyview_irq_handler(int irq, void *arg)
 {
 	struct drm_device *dev = (struct drm_device *) arg;
@@ -590,6 +595,9 @@ static irqreturn_t valleyview_irq_handler(int irq, void *arg)
 			I915_READ(PORT_HOTPLUG_STAT);
 		}
 
+		if (pipe_stats[0] & PIPE_GMBUS_INTERRUPT_STATUS)
+			gmbus_irq_handler(dev);
+
 		if (pm_iir & GEN6_PM_DEFERRED_EVENTS)
 			gen6_queue_rps_work(dev_priv, pm_iir);
 
@@ -616,7 +624,7 @@ static void ibx_irq_handler(struct drm_device *dev, u32 pch_iir)
 				 SDE_AUDIO_POWER_SHIFT);
 
 	if (pch_iir & SDE_GMBUS)
-		DRM_DEBUG_DRIVER("PCH GMBUS interrupt\n");
+		gmbus_irq_handler(dev);
 
 	if (pch_iir & SDE_AUDIO_HDCP_MASK)
 		DRM_DEBUG_DRIVER("PCH HDCP audio interrupt\n");
@@ -662,7 +670,7 @@ static void cpt_irq_handler(struct drm_device *dev, u32 pch_iir)
 		DRM_DEBUG_DRIVER("AUX channel interrupt\n");
 
 	if (pch_iir & SDE_GMBUS_CPT)
-		DRM_DEBUG_DRIVER("PCH GMBUS interrupt\n");
+		gmbus_irq_handler(dev);
 
 	if (pch_iir & SDE_AUDIO_CP_REQ_CPT)
 		DRM_DEBUG_DRIVER("Audio CP request interrupt\n");
@@ -1880,12 +1888,14 @@ static int ironlake_irq_postinstall(struct drm_device *dev)
 		hotplug_mask = (SDE_CRT_HOTPLUG_CPT |
 				SDE_PORTB_HOTPLUG_CPT |
 				SDE_PORTC_HOTPLUG_CPT |
-				SDE_PORTD_HOTPLUG_CPT);
+				SDE_PORTD_HOTPLUG_CPT |
+				SDE_GMBUS_CPT);
 	} else {
 		hotplug_mask = (SDE_CRT_HOTPLUG |
 				SDE_PORTB_HOTPLUG |
 				SDE_PORTC_HOTPLUG |
 				SDE_PORTD_HOTPLUG |
+				SDE_GMBUS |
 				SDE_AUX_MASK);
 	}
 
@@ -1945,7 +1955,8 @@ static int ivybridge_irq_postinstall(struct drm_device *dev)
 	hotplug_mask = (SDE_CRT_HOTPLUG_CPT |
 			SDE_PORTB_HOTPLUG_CPT |
 			SDE_PORTC_HOTPLUG_CPT |
-			SDE_PORTD_HOTPLUG_CPT);
+			SDE_PORTD_HOTPLUG_CPT |
+			SDE_GMBUS_CPT);
 	dev_priv->pch_irq_mask = ~hotplug_mask;
 
 	I915_WRITE(SDEIIR, I915_READ(SDEIIR));
@@ -1999,6 +2010,7 @@ static int valleyview_irq_postinstall(struct drm_device *dev)
 	POSTING_READ(VLV_IER);
 
 	i915_enable_pipestat(dev_priv, 0, pipestat_enable);
+	i915_enable_pipestat(dev_priv, 0, PIPE_GMBUS_EVENT_ENABLE);
 	i915_enable_pipestat(dev_priv, 1, pipestat_enable);
 
 	I915_WRITE(VLV_IIR, 0xffffffff);
@@ -2483,6 +2495,7 @@ static int i965_irq_postinstall(struct drm_device *dev)
 
 	dev_priv->pipestat[0] = 0;
 	dev_priv->pipestat[1] = 0;
+	i915_enable_pipestat(dev_priv, 0, PIPE_GMBUS_EVENT_ENABLE);
 
 	/*
 	 * Enable some error detection, note the instruction error mask
@@ -2636,6 +2649,9 @@ static irqreturn_t i965_irq_handler(int irq, void *arg)
 		if (blc_event || (iir & I915_ASLE_INTERRUPT))
 			intel_opregion_asle_intr(dev);
 
+		if (pipe_stats[0] & PIPE_GMBUS_INTERRUPT_STATUS)
+			gmbus_irq_handler(dev);
+
 		/* With MSI, interrupts are only generated when iir
 		 * transitions from zero to nonzero.  If another bit got
 		 * set while we were handling the existing iir bits, then
-- 
cgit v0.10.2


From 28c70f162a315bdcfbe0bf940a740ef8bfb918d6 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sat, 1 Dec 2012 13:53:45 +0100
Subject: drm/i915: use the gmbus irq for waits

We need two special things to properly wire this up:
- Add another argument to gmbus_wait_hw_status to pass in the
  correct interrupt bit in gmbus4.
- Since we can only get an irq for one of the two events we want,
  hand-roll the wait_event_timeout code so that we wake up every
  jiffie and can check for NAKs. This way we also subsume gmbus
  support for platforms without interrupts (or where those are not
  yet enabled).

The important bit really is to only enable one gmbus interrupt source
at the same time - with that piece of lore figured out, this seems to
work flawlessly.

Ben Widawsky rightfully complained the lack of measurements for the
claimed benefits (especially since the first version was actually
broken and fell back to bit-banging). Previously reading the 256 byte
hdmi EDID takes about 72 ms here. With this patch it's down to 33 ms.
Given that transfering the 256 bytes over i2c at wire speed takes
20.5ms alone, the reduction in additional overhead is rather nice.

v2: Chris Wilson wondered whether GMBUS4 might contain some set bits
when booting up an hence result in some spurious interrupts. Since we
clear GMBUS4 after every wait and we do gmbus transfer really early in
the setup sequence to detect displays the window is small, but still
be paranoid and clear it properly.

v3: Clarify the comment that gmbus irq generation can only support one
kind of event, why it bothers us and how we work around that limit.

Cc: Daniel Kurtz <djkurtz@chromium.org>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index e7411f3..3843383 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -641,6 +641,7 @@ typedef struct drm_i915_private {
 
 	struct intel_gmbus gmbus[GMBUS_NUM_PORTS];
 
+
 	/** gmbus_mutex protects against concurrent usage of the single hw gmbus
 	 * controller on different i2c buses. */
 	struct mutex gmbus_mutex;
@@ -650,6 +651,8 @@ typedef struct drm_i915_private {
 	 */
 	uint32_t gpio_mmio_base;
 
+	wait_queue_head_t gmbus_wait_queue;
+
 	struct pci_dev *bridge_dev;
 	struct intel_ring_buffer ring[I915_NUM_RINGS];
 	uint32_t next_seqno;
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 956b95f..917a48d 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -527,7 +527,11 @@ static void gen6_queue_rps_work(struct drm_i915_private *dev_priv,
 
 static void gmbus_irq_handler(struct drm_device *dev)
 {
+	struct drm_i915_private *dev_priv = (drm_i915_private_t *) dev->dev_private;
+
 	DRM_DEBUG_DRIVER("GMBUS interrupt\n");
+
+	wake_up_all(&dev_priv->gmbus_wait_queue);
 }
 
 static irqreturn_t valleyview_irq_handler(int irq, void *arg)
diff --git a/drivers/gpu/drm/i915/intel_i2c.c b/drivers/gpu/drm/i915/intel_i2c.c
index a16eecd..8b71892 100644
--- a/drivers/gpu/drm/i915/intel_i2c.c
+++ b/drivers/gpu/drm/i915/intel_i2c.c
@@ -63,6 +63,7 @@ intel_i2c_reset(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	I915_WRITE(dev_priv->gpio_mmio_base + GMBUS0, 0);
+	I915_WRITE(dev_priv->gpio_mmio_base + GMBUS4, 0);
 }
 
 static void intel_i2c_quirk_set(struct drm_i915_private *dev_priv, bool enable)
@@ -204,20 +205,38 @@ intel_gpio_setup(struct intel_gmbus *bus, u32 pin)
 
 static int
 gmbus_wait_hw_status(struct drm_i915_private *dev_priv,
-		     u32 gmbus2_status)
+		     u32 gmbus2_status,
+		     u32 gmbus4_irq_en)
 {
-	int ret;
+	int i;
 	int reg_offset = dev_priv->gpio_mmio_base;
-	u32 gmbus2;
+	u32 gmbus2 = 0;
+	DEFINE_WAIT(wait);
+
+	/* Important: The hw handles only the first bit, so set only one! Since
+	 * we also need to check for NAKs besides the hw ready/idle signal, we
+	 * need to wake up periodically and check that ourselves. */
+	I915_WRITE(GMBUS4 + reg_offset, gmbus4_irq_en);
 
-	ret = wait_for((gmbus2 = I915_READ(GMBUS2 + reg_offset)) &
-		       (GMBUS_SATOER | gmbus2_status),
-		       50);
+	for (i = 0; i < msecs_to_jiffies(50) + 1; i++) {
+		prepare_to_wait(&dev_priv->gmbus_wait_queue, &wait,
+				TASK_UNINTERRUPTIBLE);
+
+		gmbus2 = I915_READ(GMBUS2 + reg_offset);
+		if (gmbus2 & (GMBUS_SATOER | gmbus2_status))
+			break;
+
+		schedule_timeout(1);
+	}
+	finish_wait(&dev_priv->gmbus_wait_queue, &wait);
+
+	I915_WRITE(GMBUS4 + reg_offset, 0);
 
 	if (gmbus2 & GMBUS_SATOER)
 		return -ENXIO;
-
-	return ret;
+	if (gmbus2 & gmbus2_status)
+		return 0;
+	return -ETIMEDOUT;
 }
 
 static int
@@ -238,7 +257,8 @@ gmbus_xfer_read(struct drm_i915_private *dev_priv, struct i2c_msg *msg,
 		int ret;
 		u32 val, loop = 0;
 
-		ret = gmbus_wait_hw_status(dev_priv, GMBUS_HW_RDY);
+		ret = gmbus_wait_hw_status(dev_priv, GMBUS_HW_RDY,
+					   GMBUS_HW_RDY_EN);
 		if (ret)
 			return ret;
 
@@ -282,7 +302,8 @@ gmbus_xfer_write(struct drm_i915_private *dev_priv, struct i2c_msg *msg)
 
 		I915_WRITE(GMBUS3 + reg_offset, val);
 
-		ret = gmbus_wait_hw_status(dev_priv, GMBUS_HW_RDY);
+		ret = gmbus_wait_hw_status(dev_priv, GMBUS_HW_RDY,
+					   GMBUS_HW_RDY_EN);
 		if (ret)
 			return ret;
 	}
@@ -367,7 +388,8 @@ gmbus_xfer(struct i2c_adapter *adapter,
 		if (ret == -ENXIO)
 			goto clear_err;
 
-		ret = gmbus_wait_hw_status(dev_priv, GMBUS_HW_WAIT_PHASE);
+		ret = gmbus_wait_hw_status(dev_priv, GMBUS_HW_WAIT_PHASE,
+					   GMBUS_HW_WAIT_EN);
 		if (ret == -ENXIO)
 			goto clear_err;
 		if (ret)
@@ -473,6 +495,7 @@ int intel_setup_gmbus(struct drm_device *dev)
 		dev_priv->gpio_mmio_base = 0;
 
 	mutex_init(&dev_priv->gmbus_mutex);
+	init_waitqueue_head(&dev_priv->gmbus_wait_queue);
 
 	for (i = 0; i < GMBUS_NUM_PORTS; i++) {
 		struct intel_gmbus *bus = &dev_priv->gmbus[i];
-- 
cgit v0.10.2


From 2c438c0273b76d6cb158f8bdd0aa3ebf66e48a28 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sat, 1 Dec 2012 13:53:46 +0100
Subject: drm/i915: use gmbus irq to wait for gmbus idle

GMBUS_ACTIVE has inverted sense and so doesn't fit into the
wait_hw_status helper, hence create a new gmbus_wait_idle functions.
Also, we only care about the idle irq event and nothing else, which
allows us to use the wait_event_timeout helper directly without
jumping through hoops to catch NAKs.

Since gen2/3 don't have gmbus interrupts, handle them separately with
the old wait_for macro.

This shaves another few ms off reading EDID from a hdmi screen on my
testbox here. EDID reading with interrupt driven gmbus is now as fast
as with busy-looping gmbus at 28 ms here (with negligible cpu
overhead).

Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_i2c.c b/drivers/gpu/drm/i915/intel_i2c.c
index 8b71892..1fc3119 100644
--- a/drivers/gpu/drm/i915/intel_i2c.c
+++ b/drivers/gpu/drm/i915/intel_i2c.c
@@ -203,6 +203,7 @@ intel_gpio_setup(struct intel_gmbus *bus, u32 pin)
 	algo->data = bus;
 }
 
+#define HAS_GMBUS_IRQ(dev) (INTEL_INFO(dev)->gen >= 4)
 static int
 gmbus_wait_hw_status(struct drm_i915_private *dev_priv,
 		     u32 gmbus2_status,
@@ -240,6 +241,31 @@ gmbus_wait_hw_status(struct drm_i915_private *dev_priv,
 }
 
 static int
+gmbus_wait_idle(struct drm_i915_private *dev_priv)
+{
+	int ret;
+	int reg_offset = dev_priv->gpio_mmio_base;
+
+#define C ((I915_READ(GMBUS2 + reg_offset) & GMBUS_ACTIVE) == 0)
+
+	if (!HAS_GMBUS_IRQ(dev_priv->dev))
+		return wait_for(C, 10);
+
+	/* Important: The hw handles only the first bit, so set only one! */
+	I915_WRITE(GMBUS4 + reg_offset, GMBUS_IDLE_EN);
+
+	ret = wait_event_timeout(dev_priv->gmbus_wait_queue, C, 10);
+
+	I915_WRITE(GMBUS4 + reg_offset, 0);
+
+	if (ret)
+		return 0;
+	else
+		return -ETIMEDOUT;
+#undef C
+}
+
+static int
 gmbus_xfer_read(struct drm_i915_private *dev_priv, struct i2c_msg *msg,
 		u32 gmbus1_index)
 {
@@ -406,8 +432,7 @@ gmbus_xfer(struct i2c_adapter *adapter,
 	 * We will re-enable it at the start of the next xfer,
 	 * till then let it sleep.
 	 */
-	if (wait_for((I915_READ(GMBUS2 + reg_offset) & GMBUS_ACTIVE) == 0,
-		     10)) {
+	if (gmbus_wait_idle(dev_priv)) {
 		DRM_DEBUG_KMS("GMBUS [%s] timed out waiting for idle\n",
 			 adapter->name);
 		ret = -ETIMEDOUT;
@@ -431,8 +456,7 @@ clear_err:
 	 * it's slow responding and only answers on the 2nd retry.
 	 */
 	ret = -ENXIO;
-	if (wait_for((I915_READ(GMBUS2 + reg_offset) & GMBUS_ACTIVE) == 0,
-		     10)) {
+	if (gmbus_wait_idle(dev_priv)) {
 		DRM_DEBUG_KMS("GMBUS [%s] timed out after NAK\n",
 			      adapter->name);
 		ret = -ETIMEDOUT;
-- 
cgit v0.10.2


From ce99c2569dfcec9662993961db7f433e160c50f3 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sat, 1 Dec 2012 13:53:47 +0100
Subject: drm/i915: wire up do aux channel done interrupt

Doesn't do anything yet than call dp_aux_irq_handler.

Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 917a48d..1e245e5 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -534,6 +534,11 @@ static void gmbus_irq_handler(struct drm_device *dev)
 	wake_up_all(&dev_priv->gmbus_wait_queue);
 }
 
+static void dp_aux_irq_handler(struct drm_device *dev)
+{
+	DRM_DEBUG_DRIVER("AUX channel interrupt\n");
+}
+
 static irqreturn_t valleyview_irq_handler(int irq, void *arg)
 {
 	struct drm_device *dev = (struct drm_device *) arg;
@@ -627,6 +632,9 @@ static void ibx_irq_handler(struct drm_device *dev, u32 pch_iir)
 				 (pch_iir & SDE_AUDIO_POWER_MASK) >>
 				 SDE_AUDIO_POWER_SHIFT);
 
+	if (pch_iir & SDE_AUX_MASK)
+		dp_aux_irq_handler(dev);
+
 	if (pch_iir & SDE_GMBUS)
 		gmbus_irq_handler(dev);
 
@@ -671,7 +679,7 @@ static void cpt_irq_handler(struct drm_device *dev, u32 pch_iir)
 				 SDE_AUDIO_POWER_SHIFT_CPT);
 
 	if (pch_iir & SDE_AUX_MASK_CPT)
-		DRM_DEBUG_DRIVER("AUX channel interrupt\n");
+		dp_aux_irq_handler(dev);
 
 	if (pch_iir & SDE_GMBUS_CPT)
 		gmbus_irq_handler(dev);
@@ -712,6 +720,9 @@ static irqreturn_t ivybridge_irq_handler(int irq, void *arg)
 
 	de_iir = I915_READ(DEIIR);
 	if (de_iir) {
+		if (de_iir & DE_AUX_CHANNEL_A_IVB)
+			dp_aux_irq_handler(dev);
+
 		if (de_iir & DE_GSE_IVB)
 			intel_opregion_gse_intr(dev);
 
@@ -790,6 +801,9 @@ static irqreturn_t ironlake_irq_handler(int irq, void *arg)
 	else
 		snb_gt_irq_handler(dev, dev_priv, gt_iir);
 
+	if (de_iir & DE_AUX_CHANNEL_A)
+		dp_aux_irq_handler(dev);
+
 	if (de_iir & DE_GSE)
 		intel_opregion_gse_intr(dev);
 
@@ -1858,7 +1872,8 @@ static int ironlake_irq_postinstall(struct drm_device *dev)
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
 	/* enable kind of interrupts always enabled */
 	u32 display_mask = DE_MASTER_IRQ_CONTROL | DE_GSE | DE_PCH_EVENT |
-			   DE_PLANEA_FLIP_DONE | DE_PLANEB_FLIP_DONE;
+			   DE_PLANEA_FLIP_DONE | DE_PLANEB_FLIP_DONE |
+			   DE_AUX_CHANNEL_A;
 	u32 render_irqs;
 	u32 hotplug_mask;
 
@@ -1893,7 +1908,8 @@ static int ironlake_irq_postinstall(struct drm_device *dev)
 				SDE_PORTB_HOTPLUG_CPT |
 				SDE_PORTC_HOTPLUG_CPT |
 				SDE_PORTD_HOTPLUG_CPT |
-				SDE_GMBUS_CPT);
+				SDE_GMBUS_CPT |
+				SDE_AUX_MASK_CPT);
 	} else {
 		hotplug_mask = (SDE_CRT_HOTPLUG |
 				SDE_PORTB_HOTPLUG |
@@ -1930,7 +1946,8 @@ static int ivybridge_irq_postinstall(struct drm_device *dev)
 		DE_MASTER_IRQ_CONTROL | DE_GSE_IVB | DE_PCH_EVENT_IVB |
 		DE_PLANEC_FLIP_DONE_IVB |
 		DE_PLANEB_FLIP_DONE_IVB |
-		DE_PLANEA_FLIP_DONE_IVB;
+		DE_PLANEA_FLIP_DONE_IVB |
+		DE_AUX_CHANNEL_A_IVB;
 	u32 render_irqs;
 	u32 hotplug_mask;
 
@@ -1960,7 +1977,8 @@ static int ivybridge_irq_postinstall(struct drm_device *dev)
 			SDE_PORTB_HOTPLUG_CPT |
 			SDE_PORTC_HOTPLUG_CPT |
 			SDE_PORTD_HOTPLUG_CPT |
-			SDE_GMBUS_CPT);
+			SDE_GMBUS_CPT |
+			SDE_AUX_MASK_CPT);
 	dev_priv->pch_irq_mask = ~hotplug_mask;
 
 	I915_WRITE(SDEIIR, I915_READ(SDEIIR));
-- 
cgit v0.10.2


From 9ee32fea5fe810ec06af3a15e4c65478de56d4f5 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sat, 1 Dec 2012 13:53:48 +0100
Subject: drm/i915: irq-drive the dp aux communication

At least on the platforms that have a dp aux irq and also have it
enabled - vlvhsw should have one, too. But I don't have a machine to
test this on. Judging from docs there's no dp aux interrupt for gm45.

Also, I only have an ivb cpu edp machine, so the dp aux A code for
snb/ilk is untested.

For dpcd probing when nothing is connected it slashes about 5ms of cpu
time (cpu time is now negligible), which agrees with 3 * 5 400 usec
timeouts.

A previous version of this patch increases the time required to go
through the dp_detect cycle (which includes reading the edid) from
around 33 ms to around 40 ms. Experiments indicated that this is
purely due to the irq latency - the hw doesn't allow us to queue up
dp aux transactions and hence irq latency directly affects throughput.
gmbus is much better, there we have a 8 byte buffer, and we get the
irq once another 4 bytes can be queued up.

But by using the pm_qos interface to request the lowest possible cpu
wake-up latency this slowdown completely disappeared.

Since all our output detection logic is single-threaded with the
mode_config mutex right now anyway, I've decide not ot play fancy and
to just reuse the gmbus wait queue. But this would definitely prep the
way to run dp detection on different ports in parallel

v2: Add a timeout for dp aux transfers when using interrupts - the hw
_does_  prevent this with the hw-based 400 usec timeout, but if the
irq somehow doesn't arrive we're screwed. Lesson learned while
developing this ;-)

v3: While at it also convert the busy-loop to wait_for_atomic, so that
we don't run the risk of an infinite loop any more.

v4: Ensure we have the smallest possible irq latency by using the
pm_qos interface.

v5: Add a comment to the code to explain why we frob pm_qos. Suggested
by Chris Wilson.

v6: Disable dp irq for vlv, that's easier than trying to get at docs
and hw.

v7: Squash in a fix for Haswell that Paulo Zanoni tracked down - the
dp aux registers aren't at a fixed offset any more, but can be on the
PCH while the DP port is on the cpu die.

Reviewed-by: Imre Deak <imre.deak@intel.com> (v6)
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index 14d679a..9363066 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -1737,6 +1737,7 @@ int i915_driver_unload(struct drm_device *dev)
 	intel_teardown_mchbar(dev);
 
 	destroy_workqueue(dev_priv->wq);
+	pm_qos_remove_request(&dev_priv->pm_qos);
 
 	if (dev_priv->slab)
 		kmem_cache_destroy(dev_priv->slab);
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 3843383..a00ee3d 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -40,6 +40,7 @@
 #include <linux/backlight.h>
 #include <linux/intel-iommu.h>
 #include <linux/kref.h>
+#include <linux/pm_qos.h>
 
 /* General customization:
  */
@@ -665,6 +666,9 @@ typedef struct drm_i915_private {
 	/* protects the irq masks */
 	spinlock_t irq_lock;
 
+	/* To control wakeup latency, e.g. for irq-driven dp aux transfers. */
+	struct pm_qos_request pm_qos;
+
 	/* DPIO indirect register protection */
 	spinlock_t dpio_lock;
 
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 1e245e5..58bb11b 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -536,7 +536,11 @@ static void gmbus_irq_handler(struct drm_device *dev)
 
 static void dp_aux_irq_handler(struct drm_device *dev)
 {
+	struct drm_i915_private *dev_priv = (drm_i915_private_t *) dev->dev_private;
+
 	DRM_DEBUG_DRIVER("AUX channel interrupt\n");
+
+	wake_up_all(&dev_priv->gmbus_wait_queue);
 }
 
 static irqreturn_t valleyview_irq_handler(int irq, void *arg)
@@ -2732,6 +2736,8 @@ void intel_irq_init(struct drm_device *dev)
 	setup_timer(&dev_priv->hangcheck_timer, i915_hangcheck_elapsed,
 		    (unsigned long) dev);
 
+	pm_qos_add_request(&dev_priv->pm_qos, PM_QOS_CPU_DMA_LATENCY, 0);
+
 	dev->driver->get_vblank_counter = i915_get_vblank_counter;
 	dev->max_vblank_count = 0xffffff; /* only 24 bits of frame count */
 	if (IS_G4X(dev) || INTEL_INFO(dev)->gen >= 5) {
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index b51043e..ada1b31 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -322,6 +322,48 @@ intel_dp_check_edp(struct intel_dp *intel_dp)
 	}
 }
 
+static uint32_t
+intel_dp_aux_wait_done(struct intel_dp *intel_dp, bool has_aux_irq)
+{
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct drm_device *dev = intel_dig_port->base.base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t ch_ctl = intel_dp->output_reg + 0x10;
+	uint32_t status;
+	bool done;
+
+	if (IS_HASWELL(dev)) {
+		switch (intel_dig_port->port) {
+		case PORT_A:
+			ch_ctl = DPA_AUX_CH_CTL;
+			break;
+		case PORT_B:
+			ch_ctl = PCH_DPB_AUX_CH_CTL;
+			break;
+		case PORT_C:
+			ch_ctl = PCH_DPC_AUX_CH_CTL;
+			break;
+		case PORT_D:
+			ch_ctl = PCH_DPD_AUX_CH_CTL;
+			break;
+		default:
+			BUG();
+		}
+	}
+
+#define C (((status = I915_READ(ch_ctl)) & DP_AUX_CH_CTL_SEND_BUSY) == 0)
+	if (has_aux_irq)
+		done = wait_event_timeout(dev_priv->gmbus_wait_queue, C, 10);
+	else
+		done = wait_for_atomic(C, 10) == 0;
+	if (!done)
+		DRM_ERROR("dp aux hw did not signal timeout (has irq: %i)!\n",
+			  has_aux_irq);
+#undef C
+
+	return status;
+}
+
 static int
 intel_dp_aux_ch(struct intel_dp *intel_dp,
 		uint8_t *send, int send_bytes,
@@ -333,11 +375,17 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	uint32_t ch_ctl = output_reg + 0x10;
 	uint32_t ch_data = ch_ctl + 4;
-	int i;
-	int recv_bytes;
+	int i, ret, recv_bytes;
 	uint32_t status;
 	uint32_t aux_clock_divider;
 	int try, precharge;
+	bool has_aux_irq = INTEL_INFO(dev)->gen >= 5 && !IS_VALLEYVIEW(dev);
+
+	/* dp aux is extremely sensitive to irq latency, hence request the
+	 * lowest possible wakeup latency and so prevent the cpu from going into
+	 * deep sleep states.
+	 */
+	pm_qos_update_request(&dev_priv->pm_qos, 0);
 
 	if (IS_HASWELL(dev)) {
 		switch (intel_dig_port->port) {
@@ -400,7 +448,8 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 	if (try == 3) {
 		WARN(1, "dp_aux_ch not started status 0x%08x\n",
 		     I915_READ(ch_ctl));
-		return -EBUSY;
+		ret = -EBUSY;
+		goto out;
 	}
 
 	/* Must try at least 3 times according to DP spec */
@@ -413,6 +462,7 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 		/* Send the command and wait for it to complete */
 		I915_WRITE(ch_ctl,
 			   DP_AUX_CH_CTL_SEND_BUSY |
+			   (has_aux_irq ? DP_AUX_CH_CTL_INTERRUPT : 0) |
 			   DP_AUX_CH_CTL_TIME_OUT_400us |
 			   (send_bytes << DP_AUX_CH_CTL_MESSAGE_SIZE_SHIFT) |
 			   (precharge << DP_AUX_CH_CTL_PRECHARGE_2US_SHIFT) |
@@ -420,12 +470,8 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 			   DP_AUX_CH_CTL_DONE |
 			   DP_AUX_CH_CTL_TIME_OUT_ERROR |
 			   DP_AUX_CH_CTL_RECEIVE_ERROR);
-		for (;;) {
-			status = I915_READ(ch_ctl);
-			if ((status & DP_AUX_CH_CTL_SEND_BUSY) == 0)
-				break;
-			udelay(100);
-		}
+
+		status = intel_dp_aux_wait_done(intel_dp, has_aux_irq);
 
 		/* Clear done status and any errors */
 		I915_WRITE(ch_ctl,
@@ -443,7 +489,8 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 
 	if ((status & DP_AUX_CH_CTL_DONE) == 0) {
 		DRM_ERROR("dp_aux_ch not done status 0x%08x\n", status);
-		return -EBUSY;
+		ret = -EBUSY;
+		goto out;
 	}
 
 	/* Check for timeout or receive error.
@@ -451,14 +498,16 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 	 */
 	if (status & DP_AUX_CH_CTL_RECEIVE_ERROR) {
 		DRM_ERROR("dp_aux_ch receive error status 0x%08x\n", status);
-		return -EIO;
+		ret = -EIO;
+		goto out;
 	}
 
 	/* Timeouts occur when the device isn't connected, so they're
 	 * "normal" -- don't fill the kernel log with these */
 	if (status & DP_AUX_CH_CTL_TIME_OUT_ERROR) {
 		DRM_DEBUG_KMS("dp_aux_ch timeout status 0x%08x\n", status);
-		return -ETIMEDOUT;
+		ret = -ETIMEDOUT;
+		goto out;
 	}
 
 	/* Unload any bytes sent back from the other side */
@@ -471,7 +520,11 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 		unpack_aux(I915_READ(ch_data + i),
 			   recv + i, recv_bytes - i);
 
-	return recv_bytes;
+	ret = recv_bytes;
+out:
+	pm_qos_update_request(&dev_priv->pm_qos, PM_QOS_DEFAULT_VALUE);
+
+	return ret;
 }
 
 /* Write data to the aux channel in native mode */
-- 
cgit v0.10.2


From ef04f00d12312e43b13ba8f79435763e25aaff5c Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sat, 1 Dec 2012 21:03:59 +0100
Subject: drm/i915: use _NOTRACE for gmbus/dp aux wait loops

Less clutter in the traces. And in both cases we yell rather loud
into the logs if we time out. Patch suggested by Chris Wilson.

v2: Annotate another I915_READ in dp_aux to be consistent - we filter
out all register io in wait_for and similar loops. Chris also
suggested to mark all dp_aux register access as _NOTRACE, but I think
we should keep all functionally relevant access around, and filter
unneeded bits in userspace after the trace is captured.

Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index ada1b31..3f633ca 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -351,7 +351,7 @@ intel_dp_aux_wait_done(struct intel_dp *intel_dp, bool has_aux_irq)
 		}
 	}
 
-#define C (((status = I915_READ(ch_ctl)) & DP_AUX_CH_CTL_SEND_BUSY) == 0)
+#define C (((status = I915_READ_NOTRACE(ch_ctl)) & DP_AUX_CH_CTL_SEND_BUSY) == 0)
 	if (has_aux_irq)
 		done = wait_event_timeout(dev_priv->gmbus_wait_queue, C, 10);
 	else
@@ -439,7 +439,7 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 
 	/* Try to wait for any previous AUX channel activity */
 	for (try = 0; try < 3; try++) {
-		status = I915_READ(ch_ctl);
+		status = I915_READ_NOTRACE(ch_ctl);
 		if ((status & DP_AUX_CH_CTL_SEND_BUSY) == 0)
 			break;
 		msleep(1);
diff --git a/drivers/gpu/drm/i915/intel_i2c.c b/drivers/gpu/drm/i915/intel_i2c.c
index 1fc3119..7f09041 100644
--- a/drivers/gpu/drm/i915/intel_i2c.c
+++ b/drivers/gpu/drm/i915/intel_i2c.c
@@ -223,7 +223,7 @@ gmbus_wait_hw_status(struct drm_i915_private *dev_priv,
 		prepare_to_wait(&dev_priv->gmbus_wait_queue, &wait,
 				TASK_UNINTERRUPTIBLE);
 
-		gmbus2 = I915_READ(GMBUS2 + reg_offset);
+		gmbus2 = I915_READ_NOTRACE(GMBUS2 + reg_offset);
 		if (gmbus2 & (GMBUS_SATOER | gmbus2_status))
 			break;
 
@@ -246,7 +246,7 @@ gmbus_wait_idle(struct drm_i915_private *dev_priv)
 	int ret;
 	int reg_offset = dev_priv->gpio_mmio_base;
 
-#define C ((I915_READ(GMBUS2 + reg_offset) & GMBUS_ACTIVE) == 0)
+#define C ((I915_READ_NOTRACE(GMBUS2 + reg_offset) & GMBUS_ACTIVE) == 0)
 
 	if (!HAS_GMBUS_IRQ(dev_priv->dev))
 		return wait_for(C, 10);
-- 
cgit v0.10.2


From 6ef6a450b9a4c8851d63e58aa62ee6a34c01c156 Mon Sep 17 00:00:00 2001
From: Dexuan Cui <dexuan.cui@intel.com>
Date: Wed, 5 Dec 2012 22:37:17 +0800
Subject: drm/i915: Remove duplicate and unused register #defines in i915_reg.h

TRANS_DP_VIDEO_AUDIO is not used at all.
The other 3 has duplicated #defines.

Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 462b50a..f2a5ea6 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -1875,8 +1875,6 @@
 #define   PFIT_SCALING_PILLAR	(2 << 26)
 #define   PFIT_SCALING_LETTER	(3 << 26)
 #define PFIT_PGM_RATIOS	0x61234
-#define   PFIT_VERT_SCALE_MASK			0xfff00000
-#define   PFIT_HORIZ_SCALE_MASK			0x0000fff0
 /* Pre-965 */
 #define		PFIT_VERT_SCALE_SHIFT		20
 #define		PFIT_VERT_SCALE_MASK		0xfff00000
@@ -3799,8 +3797,6 @@
 #define  TRANS_FSYNC_DELAY_HB2  (1<<27)
 #define  TRANS_FSYNC_DELAY_HB3  (2<<27)
 #define  TRANS_FSYNC_DELAY_HB4  (3<<27)
-#define  TRANS_DP_AUDIO_ONLY    (1<<26)
-#define  TRANS_DP_VIDEO_AUDIO   (0<<26)
 #define  TRANS_INTERLACE_MASK   (7<<21)
 #define  TRANS_PROGRESSIVE      (0<<21)
 #define  TRANS_INTERLACED       (3<<21)
-- 
cgit v0.10.2


From 36dacf5b8b3c0bab7010943c34d2dbcc09a0d6f3 Mon Sep 17 00:00:00 2001
From: Paulo Zanoni <paulo.r.zanoni@intel.com>
Date: Thu, 6 Dec 2012 09:44:52 -0200
Subject: drm/i915: be less verbose when handling gmbus/aux irqs

Having 9500 lines repeated on dmesg does not help me at all.

Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 58bb11b..2e1d80d 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -529,8 +529,6 @@ static void gmbus_irq_handler(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = (drm_i915_private_t *) dev->dev_private;
 
-	DRM_DEBUG_DRIVER("GMBUS interrupt\n");
-
 	wake_up_all(&dev_priv->gmbus_wait_queue);
 }
 
@@ -538,8 +536,6 @@ static void dp_aux_irq_handler(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = (drm_i915_private_t *) dev->dev_private;
 
-	DRM_DEBUG_DRIVER("AUX channel interrupt\n");
-
 	wake_up_all(&dev_priv->gmbus_wait_queue);
 }
 
-- 
cgit v0.10.2


From ed7ef43989b3a8235b1cbf1e2b025aa4af3368a6 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 6 Dec 2012 14:24:21 +0100
Subject: drm/i915: rip out pre-DDI stuff from haswell_crtc_mode_set

Especially getting rid of all things lvds is ... great!

v2: Drop the two additional pre-hsw hunks noticed by Paulo Zanoni.

v3:
- handle DP ports correctly (spoted by Paulo)
- don't leave {} behind for a single-line block (again spotted by
  Paulo)
- kill another if (IBX || CPT) block

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 093a163..5865cd9 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -5461,20 +5461,13 @@ static int haswell_crtc_mode_set(struct drm_crtc *crtc,
 	int pipe = intel_crtc->pipe;
 	int plane = intel_crtc->plane;
 	int num_connectors = 0;
-	intel_clock_t clock, reduced_clock;
-	u32 dpll = 0, fp = 0, fp2 = 0;
-	bool ok, has_reduced_clock = false;
-	bool is_lvds = false, is_dp = false, is_cpu_edp = false;
+	bool is_dp = false, is_cpu_edp = false;
 	struct intel_encoder *encoder;
-	u32 temp;
 	int ret;
 	bool dither;
 
 	for_each_encoder_on_crtc(dev, crtc, encoder) {
 		switch (encoder->type) {
-		case INTEL_OUTPUT_LVDS:
-			is_lvds = true;
-			break;
 		case INTEL_OUTPUT_DISPLAYPORT:
 			is_dp = true;
 			break;
@@ -5508,143 +5501,26 @@ static int haswell_crtc_mode_set(struct drm_crtc *crtc,
 	if (!intel_ddi_pll_mode_set(crtc, adjusted_mode->clock))
 		return -EINVAL;
 
-	if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev)) {
-		ok = ironlake_compute_clocks(crtc, adjusted_mode, &clock,
-					     &has_reduced_clock,
-					     &reduced_clock);
-		if (!ok) {
-			DRM_ERROR("Couldn't find PLL settings for mode!\n");
-			return -EINVAL;
-		}
-	}
-
 	/* Ensure that the cursor is valid for the new mode before changing... */
 	intel_crtc_update_cursor(crtc, true);
 
 	/* determine panel color depth */
 	dither = intel_choose_pipe_bpp_dither(crtc, fb, &intel_crtc->bpp,
 					      adjusted_mode);
-	if (is_lvds && dev_priv->lvds_dither)
-		dither = true;
 
 	DRM_DEBUG_KMS("Mode for pipe %d:\n", pipe);
 	drm_mode_debug_printmodeline(mode);
 
-	if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev)) {
-		fp = clock.n << 16 | clock.m1 << 8 | clock.m2;
-		if (has_reduced_clock)
-			fp2 = reduced_clock.n << 16 | reduced_clock.m1 << 8 |
-			      reduced_clock.m2;
-
-		dpll = ironlake_compute_dpll(intel_crtc, adjusted_mode, &clock,
-					     fp);
-
-		/* CPU eDP is the only output that doesn't need a PCH PLL of its
-		 * own on pre-Haswell/LPT generation */
-		if (!is_cpu_edp) {
-			struct intel_pch_pll *pll;
-
-			pll = intel_get_pch_pll(intel_crtc, dpll, fp);
-			if (pll == NULL) {
-				DRM_DEBUG_DRIVER("failed to find PLL for pipe %d\n",
-						 pipe);
-				return -EINVAL;
-			}
-		} else
-			intel_put_pch_pll(intel_crtc);
-
-		/* The LVDS pin pair needs to be on before the DPLLs are
-		 * enabled.  This is an exception to the general rule that
-		 * mode_set doesn't turn things on.
-		 */
-		if (is_lvds) {
-			temp = I915_READ(PCH_LVDS);
-			temp |= LVDS_PORT_EN | LVDS_A0A2_CLKA_POWER_UP;
-			if (HAS_PCH_CPT(dev)) {
-				temp &= ~PORT_TRANS_SEL_MASK;
-				temp |= PORT_TRANS_SEL_CPT(pipe);
-			} else {
-				if (pipe == 1)
-					temp |= LVDS_PIPEB_SELECT;
-				else
-					temp &= ~LVDS_PIPEB_SELECT;
-			}
-
-			/* set the corresponsding LVDS_BORDER bit */
-			temp |= dev_priv->lvds_border_bits;
-			/* Set the B0-B3 data pairs corresponding to whether
-			 * we're going to set the DPLLs for dual-channel mode or
-			 * not.
-			 */
-			if (clock.p2 == 7)
-				temp |= LVDS_B0B3_POWER_UP | LVDS_CLKB_POWER_UP;
-			else
-				temp &= ~(LVDS_B0B3_POWER_UP |
-					  LVDS_CLKB_POWER_UP);
-
-			/* It would be nice to set 24 vs 18-bit mode
-			 * (LVDS_A3_POWER_UP) appropriately here, but we need to
-			 * look more thoroughly into how panels behave in the
-			 * two modes.
-			 */
-			temp &= ~(LVDS_HSYNC_POLARITY | LVDS_VSYNC_POLARITY);
-			if (adjusted_mode->flags & DRM_MODE_FLAG_NHSYNC)
-				temp |= LVDS_HSYNC_POLARITY;
-			if (adjusted_mode->flags & DRM_MODE_FLAG_NVSYNC)
-				temp |= LVDS_VSYNC_POLARITY;
-			I915_WRITE(PCH_LVDS, temp);
-		}
-	}
-
-	if (is_dp && !is_cpu_edp) {
+	if (is_dp && !is_cpu_edp)
 		intel_dp_set_m_n(crtc, mode, adjusted_mode);
-	} else {
-		if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev)) {
-			/* For non-DP output, clear any trans DP clock recovery
-			 * setting.*/
-			I915_WRITE(TRANSDATA_M1(pipe), 0);
-			I915_WRITE(TRANSDATA_N1(pipe), 0);
-			I915_WRITE(TRANSDPLINK_M1(pipe), 0);
-			I915_WRITE(TRANSDPLINK_N1(pipe), 0);
-		}
-	}
 
 	intel_crtc->lowfreq_avail = false;
-	if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev)) {
-		if (intel_crtc->pch_pll) {
-			I915_WRITE(intel_crtc->pch_pll->pll_reg, dpll);
-
-			/* Wait for the clocks to stabilize. */
-			POSTING_READ(intel_crtc->pch_pll->pll_reg);
-			udelay(150);
-
-			/* The pixel multiplier can only be updated once the
-			 * DPLL is enabled and the clocks are stable.
-			 *
-			 * So write it again.
-			 */
-			I915_WRITE(intel_crtc->pch_pll->pll_reg, dpll);
-		}
-
-		if (intel_crtc->pch_pll) {
-			if (is_lvds && has_reduced_clock && i915_powersave) {
-				I915_WRITE(intel_crtc->pch_pll->fp1_reg, fp2);
-				intel_crtc->lowfreq_avail = true;
-			} else {
-				I915_WRITE(intel_crtc->pch_pll->fp1_reg, fp);
-			}
-		}
-	}
 
 	intel_set_pipe_timings(intel_crtc, mode, adjusted_mode);
 
 	if (!is_dp || is_cpu_edp)
 		ironlake_set_m_n(crtc, mode, adjusted_mode);
 
-	if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev))
-		if (is_cpu_edp)
-			ironlake_set_pll_edp(crtc, adjusted_mode->clock);
-
 	haswell_set_pipeconf(crtc, adjusted_mode, dither);
 
 	/* Set up the display plane register */
-- 
cgit v0.10.2


From ea9b6006b51b79cfbb87c1ca81923761b7799c0f Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 29 Nov 2012 15:59:31 +0100
Subject: drm/i915: move set_pll_edp to intel_dp.c

Now that we enable the cpu edp pll in intel_dp->pre_enable and no
longer in crtc_mode_set, we can also move the modeset part to the
intel_dp->mode_set callback. Previously this was not possible because
the encoder ->mode_set callbacks are called after the crtc mode set
callback.

v2: Rebase on top of copy&pasted hsw crtc_mode_set.

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 5865cd9..f3e58a4 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -2294,43 +2294,6 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
 	return 0;
 }
 
-static void ironlake_set_pll_edp(struct drm_crtc *crtc, int clock)
-{
-	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 dpa_ctl;
-
-	DRM_DEBUG_KMS("eDP PLL enable for clock %d\n", clock);
-	dpa_ctl = I915_READ(DP_A);
-	dpa_ctl &= ~DP_PLL_FREQ_MASK;
-
-	if (clock < 200000) {
-		u32 temp;
-		dpa_ctl |= DP_PLL_FREQ_160MHZ;
-		/* workaround for 160Mhz:
-		   1) program 0x4600c bits 15:0 = 0x8124
-		   2) program 0x46010 bit 0 = 1
-		   3) program 0x46034 bit 24 = 1
-		   4) program 0x64000 bit 14 = 1
-		   */
-		temp = I915_READ(0x4600c);
-		temp &= 0xffff0000;
-		I915_WRITE(0x4600c, temp | 0x8124);
-
-		temp = I915_READ(0x46010);
-		I915_WRITE(0x46010, temp | 1);
-
-		temp = I915_READ(0x46034);
-		I915_WRITE(0x46034, temp | (1 << 24));
-	} else {
-		dpa_ctl |= DP_PLL_FREQ_270MHZ;
-	}
-	I915_WRITE(DP_A, dpa_ctl);
-
-	POSTING_READ(DP_A);
-	udelay(500);
-}
-
 static void intel_fdi_normal_train(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
@@ -5429,9 +5392,6 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 
 	fdi_config_ok = ironlake_check_fdi_lanes(intel_crtc);
 
-	if (is_cpu_edp)
-		ironlake_set_pll_edp(crtc, adjusted_mode->clock);
-
 	ironlake_set_pipeconf(crtc, adjusted_mode, dither);
 
 	intel_wait_for_vblank(dev, pipe);
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 3f633ca..e525f03 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -899,6 +899,43 @@ void intel_dp_init_link_config(struct intel_dp *intel_dp)
 	}
 }
 
+static void ironlake_set_pll_edp(struct drm_crtc *crtc, int clock)
+{
+	struct drm_device *dev = crtc->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 dpa_ctl;
+
+	DRM_DEBUG_KMS("eDP PLL enable for clock %d\n", clock);
+	dpa_ctl = I915_READ(DP_A);
+	dpa_ctl &= ~DP_PLL_FREQ_MASK;
+
+	if (clock < 200000) {
+		u32 temp;
+		dpa_ctl |= DP_PLL_FREQ_160MHZ;
+		/* workaround for 160Mhz:
+		   1) program 0x4600c bits 15:0 = 0x8124
+		   2) program 0x46010 bit 0 = 1
+		   3) program 0x46034 bit 24 = 1
+		   4) program 0x64000 bit 14 = 1
+		   */
+		temp = I915_READ(0x4600c);
+		temp &= 0xffff0000;
+		I915_WRITE(0x4600c, temp | 0x8124);
+
+		temp = I915_READ(0x46010);
+		I915_WRITE(0x46010, temp | 1);
+
+		temp = I915_READ(0x46034);
+		I915_WRITE(0x46034, temp | (1 << 24));
+	} else {
+		dpa_ctl |= DP_PLL_FREQ_270MHZ;
+	}
+	I915_WRITE(DP_A, dpa_ctl);
+
+	POSTING_READ(DP_A);
+	udelay(500);
+}
+
 static void
 intel_dp_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
 		  struct drm_display_mode *adjusted_mode)
@@ -998,6 +1035,9 @@ intel_dp_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
 	} else {
 		intel_dp->DP |= DP_LINK_TRAIN_OFF_CPT;
 	}
+
+	if (is_cpu_edp(intel_dp))
+		ironlake_set_pll_edp(crtc, adjusted_mode->clock);
 }
 
 #define IDLE_ON_MASK		(PP_ON | 0 	  | PP_SEQUENCE_MASK | 0                     | PP_SEQUENCE_STATE_MASK)
-- 
cgit v0.10.2


From 1ce17038093f02e708e5816f645bbec63aff8bd2 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 29 Nov 2012 15:59:32 +0100
Subject: drm/i915: rip out pre-production ilk cpu edp w/a
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

While reading docs I've noticed that this special workaround to select
the 1.6 GHz DP clock only applies to pre-production ilk machines.
Since the registers we're touching here are rather undocumented and
might be harmful on later chips, rip it out.

For the Bspec reference of this w/a look in "vol4g CPU Display
Registers [DevILK]", Section 4.1.7.1 "DP_A—DisplayPort A
Control Register", "DP_PLL_Frequency_Select".

v2: Keep a debug message as a hint in case something regresses.
Requested by Chris Wilson.

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index e525f03..c01163d 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -910,26 +910,15 @@ static void ironlake_set_pll_edp(struct drm_crtc *crtc, int clock)
 	dpa_ctl &= ~DP_PLL_FREQ_MASK;
 
 	if (clock < 200000) {
-		u32 temp;
+		/* For a long time we've carried around a ILK-DevA w/a for the
+		 * 160MHz clock. If we're really unlucky, it's still required.
+		 */
+		DRM_DEBUG_KMS("160MHz cpu eDP clock, might need ilk devA w/a\n");
 		dpa_ctl |= DP_PLL_FREQ_160MHZ;
-		/* workaround for 160Mhz:
-		   1) program 0x4600c bits 15:0 = 0x8124
-		   2) program 0x46010 bit 0 = 1
-		   3) program 0x46034 bit 24 = 1
-		   4) program 0x64000 bit 14 = 1
-		   */
-		temp = I915_READ(0x4600c);
-		temp &= 0xffff0000;
-		I915_WRITE(0x4600c, temp | 0x8124);
-
-		temp = I915_READ(0x46010);
-		I915_WRITE(0x46010, temp | 1);
-
-		temp = I915_READ(0x46034);
-		I915_WRITE(0x46034, temp | (1 << 24));
 	} else {
 		dpa_ctl |= DP_PLL_FREQ_270MHZ;
 	}
+
 	I915_WRITE(DP_A, dpa_ctl);
 
 	POSTING_READ(DP_A);
-- 
cgit v0.10.2


From ab527efc2feadcab1cad0740be307cb4153b493f Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 29 Nov 2012 15:59:33 +0100
Subject: drm/i915: use wait_for_vblank instead of msleep(17)

17 ms is eerily close to 60 Hz ^-1

Unfortunately this goes back to the original DP enabling for ilk, and
unfortunately does not come with a reason for it's existance attached.

Some closer inspection of the code and DP specs shows that we set the
idle link pattern before we disable the port. And it seems like that
the DP spec (or at least our hw) only switch to the idle pattern on
the next vblank. Hence a vblank wait at this spot makes _much_ more
sense than a really long wait.

v2: Rebase fixup.

v3: Add comment requested by Paulo Zanoni saying that we don't really
know what this wait is for.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index c01163d..89164c5 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -2041,6 +2041,8 @@ intel_dp_link_down(struct intel_dp *intel_dp)
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = intel_dig_port->base.base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc =
+		to_intel_crtc(intel_dig_port->base.base.crtc);
 	uint32_t DP = intel_dp->DP;
 
 	/*
@@ -2075,7 +2077,8 @@ intel_dp_link_down(struct intel_dp *intel_dp)
 	}
 	POSTING_READ(intel_dp->output_reg);
 
-	msleep(17);
+	/* We don't really know why we're doing this */
+	intel_wait_for_vblank(dev, intel_crtc->pipe);
 
 	if (HAS_PCH_IBX(dev) &&
 	    I915_READ(intel_dp->output_reg) & DP_PIPEB_SELECT) {
@@ -2107,7 +2110,7 @@ intel_dp_link_down(struct intel_dp *intel_dp)
 			POSTING_READ(intel_dp->output_reg);
 			msleep(50);
 		} else
-			intel_wait_for_vblank(dev, to_intel_crtc(crtc)->pipe);
+			intel_wait_for_vblank(dev, intel_crtc->pipe);
 	}
 
 	DP &= ~DP_AUDIO_OUTPUT_ENABLE;
-- 
cgit v0.10.2


From ff50afe9aceb6264a4fbe40459da75170fb9a2a2 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 29 Nov 2012 15:59:34 +0100
Subject: drm/i915: WARN on !crtc in intel_dp_link_down

This could have happened with the old crtc helper based modeset code,
but can't happen any longer with the new code.

Hence put in a WARN and adjust the comment. If no one hits this, we
can eventually remove it (like a few other such cases across our
code).

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 89164c5..84652ca 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -2098,15 +2098,10 @@ intel_dp_link_down(struct intel_dp *intel_dp)
 		/* Changes to enable or select take place the vblank
 		 * after being written.
 		 */
-		if (crtc == NULL) {
-			/* We can arrive here never having been attached
-			 * to a CRTC, for instance, due to inheriting
-			 * random state from the BIOS.
-			 *
-			 * If the pipe is not running, play safe and
-			 * wait for the clocks to stabilise before
-			 * continuing.
-			 */
+		if (WARN_ON(crtc == NULL)) {
+			/* We should never try to disable a port without a crtc
+			 * attached. For paranoia keep the code around for a
+			 * bit. */
 			POSTING_READ(intel_dp->output_reg);
 			msleep(50);
 		} else
-- 
cgit v0.10.2


From 2f0c2ad18b88691496e23d1ddbc2d0af8f6df5fa Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 29 Nov 2012 15:59:35 +0100
Subject: drm/i915: drop unnecessary clearing of pch dp transcoder timings

This has originally been added in

commit 8db9d77b1b14fd730561f64beea8c00e4478d7c5
Author: Zhenyu Wang <zhenyuw@linux.intel.com>
Date:   Wed Apr 7 16:15:54 2010 +0800

    drm/i915: Support for Cougarpoint PCH display pipeline

probably to combat issues with hw state left behind by the BIOS. And
indeed, I've checked out that specific revision, and there is no DP
support yet. So the pch dp transcoder won't be correctly disabled, and
that's important since it requires a rether special disable dance:
Just writing 0 to TRANS_DP_CTL won't cut it, since we need to select
the NONE port when disabling, too.

And indeed, things seem to still work, so let's just remove this.

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index f3e58a4..953796c 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -5345,15 +5345,8 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 	} else
 		intel_put_pch_pll(intel_crtc);
 
-	if (is_dp && !is_cpu_edp) {
+	if (is_dp && !is_cpu_edp)
 		intel_dp_set_m_n(crtc, mode, adjusted_mode);
-	} else {
-		/* For non-DP output, clear any trans DP clock recovery setting.*/
-		I915_WRITE(TRANSDATA_M1(pipe), 0);
-		I915_WRITE(TRANSDATA_N1(pipe), 0);
-		I915_WRITE(TRANSDPLINK_M1(pipe), 0);
-		I915_WRITE(TRANSDPLINK_N1(pipe), 0);
-	}
 
 	for_each_encoder_on_crtc(dev, crtc, encoder)
 		if (encoder->pre_pll_enable)
-- 
cgit v0.10.2


From e69d0bc1c67520c302e070ac078975ea9c786de8 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 29 Nov 2012 15:59:36 +0100
Subject: drm/i915: extract common link_m_n helpers

Both the dp and fdi code use the exact same computations (ignore minor
differences in conversion between bits and bytes).

This makes it even more apparent that we have a _massive_ mess between
cpu transcoder/fdi link/pch transcoder and pch link settings. And also
that we have hilarious amounts of confusion between edp and dp
(despite that they're identical at a link level).

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index a00ee3d..e9ac360 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -107,6 +107,19 @@ struct intel_pch_pll {
 };
 #define I915_NUM_PLLS 2
 
+/* Used by dp and fdi links */
+struct intel_link_m_n {
+	uint32_t	tu;
+	uint32_t	gmch_m;
+	uint32_t	gmch_n;
+	uint32_t	link_m;
+	uint32_t	link_n;
+};
+
+void intel_link_compute_m_n(int bpp, int nlanes,
+			    int pixel_clock, int link_clock,
+			    struct intel_link_m_n *m_n);
+
 struct intel_ddi_plls {
 	int spll_refcount;
 	int wrpll1_refcount;
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 953796c..34832bc0 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -3951,16 +3951,8 @@ static int i830_get_display_clock_speed(struct drm_device *dev)
 	return 133000;
 }
 
-struct fdi_m_n {
-	u32        tu;
-	u32        gmch_m;
-	u32        gmch_n;
-	u32        link_m;
-	u32        link_n;
-};
-
 static void
-fdi_reduce_ratio(u32 *num, u32 *den)
+intel_reduce_ratio(uint32_t *num, uint32_t *den)
 {
 	while (*num > 0xffffff || *den > 0xffffff) {
 		*num >>= 1;
@@ -3968,20 +3960,18 @@ fdi_reduce_ratio(u32 *num, u32 *den)
 	}
 }
 
-static void
-ironlake_compute_m_n(int bits_per_pixel, int nlanes, int pixel_clock,
-		     int link_clock, struct fdi_m_n *m_n)
+void
+intel_link_compute_m_n(int bits_per_pixel, int nlanes,
+		       int pixel_clock, int link_clock,
+		       struct intel_link_m_n *m_n)
 {
-	m_n->tu = 64; /* default size */
-
-	/* BUG_ON(pixel_clock > INT_MAX / 36); */
+	m_n->tu = 64;
 	m_n->gmch_m = bits_per_pixel * pixel_clock;
 	m_n->gmch_n = link_clock * nlanes * 8;
-	fdi_reduce_ratio(&m_n->gmch_m, &m_n->gmch_n);
-
+	intel_reduce_ratio(&m_n->gmch_m, &m_n->gmch_n);
 	m_n->link_m = pixel_clock;
 	m_n->link_n = link_clock;
-	fdi_reduce_ratio(&m_n->link_m, &m_n->link_n);
+	intel_reduce_ratio(&m_n->link_m, &m_n->link_n);
 }
 
 static inline bool intel_panel_use_ssc(struct drm_i915_private *dev_priv)
@@ -5095,7 +5085,7 @@ static void ironlake_set_m_n(struct drm_crtc *crtc,
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	enum transcoder cpu_transcoder = intel_crtc->cpu_transcoder;
 	struct intel_encoder *intel_encoder, *edp_encoder = NULL;
-	struct fdi_m_n m_n = {0};
+	struct intel_link_m_n m_n = {0};
 	int target_clock, pixel_multiplier, lane, link_bw;
 	bool is_dp = false, is_cpu_edp = false;
 
@@ -5153,8 +5143,7 @@ static void ironlake_set_m_n(struct drm_crtc *crtc,
 
 	if (pixel_multiplier > 1)
 		link_bw *= pixel_multiplier;
-	ironlake_compute_m_n(intel_crtc->bpp, lane, target_clock, link_bw,
-			     &m_n);
+	intel_link_compute_m_n(intel_crtc->bpp, lane, target_clock, link_bw, &m_n);
 
 	I915_WRITE(PIPE_DATA_M1(cpu_transcoder), TU_SIZE(m_n.tu) | m_n.gmch_m);
 	I915_WRITE(PIPE_DATA_N1(cpu_transcoder), m_n.gmch_n);
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 84652ca..b2130bc 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -790,39 +790,6 @@ intel_dp_mode_fixup(struct drm_encoder *encoder,
 	return false;
 }
 
-struct intel_dp_m_n {
-	uint32_t	tu;
-	uint32_t	gmch_m;
-	uint32_t	gmch_n;
-	uint32_t	link_m;
-	uint32_t	link_n;
-};
-
-static void
-intel_reduce_ratio(uint32_t *num, uint32_t *den)
-{
-	while (*num > 0xffffff || *den > 0xffffff) {
-		*num >>= 1;
-		*den >>= 1;
-	}
-}
-
-static void
-intel_dp_compute_m_n(int bpp,
-		     int nlanes,
-		     int pixel_clock,
-		     int link_clock,
-		     struct intel_dp_m_n *m_n)
-{
-	m_n->tu = 64;
-	m_n->gmch_m = (pixel_clock * bpp) >> 3;
-	m_n->gmch_n = link_clock * nlanes;
-	intel_reduce_ratio(&m_n->gmch_m, &m_n->gmch_n);
-	m_n->link_m = pixel_clock;
-	m_n->link_n = link_clock;
-	intel_reduce_ratio(&m_n->link_m, &m_n->link_n);
-}
-
 void
 intel_dp_set_m_n(struct drm_crtc *crtc, struct drm_display_mode *mode,
 		 struct drm_display_mode *adjusted_mode)
@@ -833,7 +800,7 @@ intel_dp_set_m_n(struct drm_crtc *crtc, struct drm_display_mode *mode,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int lane_count = 4;
-	struct intel_dp_m_n m_n;
+	struct intel_link_m_n m_n;
 	int pipe = intel_crtc->pipe;
 	enum transcoder cpu_transcoder = intel_crtc->cpu_transcoder;
 
@@ -856,8 +823,8 @@ intel_dp_set_m_n(struct drm_crtc *crtc, struct drm_display_mode *mode,
 	 * the number of bytes_per_pixel post-LUT, which we always
 	 * set up for 8-bits of R/G/B, or 3 bytes total.
 	 */
-	intel_dp_compute_m_n(intel_crtc->bpp, lane_count,
-			     mode->clock, adjusted_mode->clock, &m_n);
+	intel_link_compute_m_n(intel_crtc->bpp, lane_count,
+			       mode->clock, adjusted_mode->clock, &m_n);
 
 	if (IS_HASWELL(dev)) {
 		I915_WRITE(PIPE_DATA_M1(cpu_transcoder),
-- 
cgit v0.10.2


From e9b73c67390a5d4faec1d22cbdf24cd6fcca53f6 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Mon, 3 Dec 2012 21:03:14 +0000
Subject: drm/i915: Reduce memory pressure during shrinker by preallocating
 swizzle pages

On a machine with bit17 swizzling, we need to store the bit17 of the
physical page address in put-pages. This requires a memory allocation,
on average less than a page, which may be difficult to satisfy is the
request to put-pages is on behalf of the shrinker. We could allow that
allocation to pull from the reserved memory pools, but it seems much
safer to preallocate the array for tiled objects on affected machines.

v2: Export i915_gem_object_needs_bit17_swizzle() for reuse.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index e9ac360..2ab476d 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -30,6 +30,8 @@
 #ifndef _I915_DRV_H_
 #define _I915_DRV_H_
 
+#include <uapi/drm/i915_drm.h>
+
 #include "i915_reg.h"
 #include "intel_bios.h"
 #include "intel_ringbuffer.h"
@@ -1614,6 +1616,14 @@ i915_gem_object_create_stolen(struct drm_device *dev, u32 size);
 void i915_gem_object_release_stolen(struct drm_i915_gem_object *obj);
 
 /* i915_gem_tiling.c */
+inline static bool i915_gem_object_needs_bit17_swizzle(struct drm_i915_gem_object *obj)
+{
+	drm_i915_private_t *dev_priv = obj->base.dev->dev_private;
+
+	return dev_priv->mm.bit_6_swizzle_x == I915_BIT_6_SWIZZLE_9_10_17 &&
+		obj->tiling_mode != I915_TILING_NONE;
+}
+
 void i915_gem_detect_bit_6_swizzle(struct drm_device *dev);
 void i915_gem_object_do_bit_17_swizzle(struct drm_i915_gem_object *obj);
 void i915_gem_object_save_bit_17_swizzle(struct drm_i915_gem_object *obj);
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 02d3151..e4b233d 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -271,14 +271,6 @@ i915_gem_create_ioctl(struct drm_device *dev, void *data,
 			       args->size, &args->handle);
 }
 
-static int i915_gem_object_needs_bit17_swizzle(struct drm_i915_gem_object *obj)
-{
-	drm_i915_private_t *dev_priv = obj->base.dev->dev_private;
-
-	return dev_priv->mm.bit_6_swizzle_x == I915_BIT_6_SWIZZLE_9_10_17 &&
-		obj->tiling_mode != I915_TILING_NONE;
-}
-
 static inline int
 __copy_to_user_swizzled(char __user *cpu_vaddr,
 			const char *gpu_vaddr, int gpu_offset,
diff --git a/drivers/gpu/drm/i915/i915_gem_tiling.c b/drivers/gpu/drm/i915/i915_gem_tiling.c
index cedbfd7..65f1d4f 100644
--- a/drivers/gpu/drm/i915/i915_gem_tiling.c
+++ b/drivers/gpu/drm/i915/i915_gem_tiling.c
@@ -396,6 +396,18 @@ i915_gem_set_tiling(struct drm_device *dev, void *data,
 	/* we have to maintain this existing ABI... */
 	args->stride = obj->stride;
 	args->tiling_mode = obj->tiling_mode;
+
+	/* Try to preallocate memory required to save swizzling on put-pages */
+	if (i915_gem_object_needs_bit17_swizzle(obj)) {
+		if (obj->bit_17 == NULL) {
+			obj->bit_17 = kmalloc(BITS_TO_LONGS(obj->base.size >> PAGE_SHIFT) *
+					      sizeof(long), GFP_KERNEL);
+		}
+	} else {
+		kfree(obj->bit_17);
+		obj->bit_17 = NULL;
+	}
+
 	drm_gem_object_unreference(&obj->base);
 	mutex_unlock(&dev->struct_mutex);
 
-- 
cgit v0.10.2


From 97a19a247c23e286814a5ac7ec0825d0ff82a16c Mon Sep 17 00:00:00 2001
From: Tomas Janousek <tomi@nomi.cz>
Date: Sat, 8 Dec 2012 13:48:13 +0100
Subject: drm/i915: don't prevent CPU idle states

Commit 9ee32fea5f unconditionally prevents the CPU from entering idle states
until intel_dp_aux_ch completes for the first time, which never happens on my
DisplayPort-less intel gfx, causing the CPU to get rather hot.

Signed-off-by: Tomas Janousek <tomi@nomi.cz>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 2e1d80d..914ecf4 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -2732,7 +2732,7 @@ void intel_irq_init(struct drm_device *dev)
 	setup_timer(&dev_priv->hangcheck_timer, i915_hangcheck_elapsed,
 		    (unsigned long) dev);
 
-	pm_qos_add_request(&dev_priv->pm_qos, PM_QOS_CPU_DMA_LATENCY, 0);
+	pm_qos_add_request(&dev_priv->pm_qos, PM_QOS_CPU_DMA_LATENCY, PM_QOS_DEFAULT_VALUE);
 
 	dev->driver->get_vblank_counter = i915_get_vblank_counter;
 	dev->max_vblank_count = 0xffffff; /* only 24 bits of frame count */
-- 
cgit v0.10.2


From 378a6a77ae59d3312627b996ded94e23166d9e63 Mon Sep 17 00:00:00 2001
From: Sachin Kamat <sachin.kamat@linaro.org>
Date: Fri, 7 Dec 2012 14:18:29 +0530
Subject: drm/i915: Remove duplicate inclusion of drm/drm_edid.h

drm/drm_edid.h was included twice.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_modes.c b/drivers/gpu/drm/i915/intel_modes.c
index b00f1c8..49249bb 100644
--- a/drivers/gpu/drm/i915/intel_modes.c
+++ b/drivers/gpu/drm/i915/intel_modes.c
@@ -28,7 +28,6 @@
 #include <linux/fb.h>
 #include <drm/drm_edid.h>
 #include <drm/drmP.h>
-#include <drm/drm_edid.h>
 #include "intel_drv.h"
 #include "i915_drv.h"
 
-- 
cgit v0.10.2


From 3ac18232946aacf0a9807c3143f8449ed4aa68f4 Mon Sep 17 00:00:00 2001
From: Tim Gardner <tim.gardner@canonical.com>
Date: Fri, 7 Dec 2012 07:54:26 -0700
Subject: i915: intel_set_mode: Reduce stack allocation from 500 bytes to 2
 pointers

smatch warning:

drivers/gpu/drm/i915/intel_display.c:7019 intel_set_mode() warn: function puts
500 bytes on stack

Refactor so that saved_mode and saved_hwmode are dynamically allocated as opposed
to being automatic variables. 500 bytes seems like it could run the potential for blowing
the kernel stack.

Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: David Airlie <airlied@linux.ie>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 34832bc0..115bf62 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -7398,11 +7398,18 @@ bool intel_set_mode(struct drm_crtc *crtc,
 {
 	struct drm_device *dev = crtc->dev;
 	drm_i915_private_t *dev_priv = dev->dev_private;
-	struct drm_display_mode *adjusted_mode, saved_mode, saved_hwmode;
+	struct drm_display_mode *adjusted_mode, *saved_mode, *saved_hwmode;
 	struct intel_crtc *intel_crtc;
 	unsigned disable_pipes, prepare_pipes, modeset_pipes;
 	bool ret = true;
 
+	saved_mode = kmalloc(2 * sizeof(*saved_mode), GFP_KERNEL);
+	if (!saved_mode) {
+		DRM_ERROR("i915: Could not allocate saved display mode.\n");
+		return false;
+	}
+	saved_hwmode = saved_mode + 1;
+
 	intel_modeset_affected_pipes(crtc, &modeset_pipes,
 				     &prepare_pipes, &disable_pipes);
 
@@ -7412,8 +7419,8 @@ bool intel_set_mode(struct drm_crtc *crtc,
 	for_each_intel_crtc_masked(dev, disable_pipes, intel_crtc)
 		intel_crtc_disable(&intel_crtc->base);
 
-	saved_hwmode = crtc->hwmode;
-	saved_mode = crtc->mode;
+	*saved_hwmode = crtc->hwmode;
+	*saved_mode = crtc->mode;
 
 	/* Hack: Because we don't (yet) support global modeset on multiple
 	 * crtcs, we don't keep track of the new mode for more than one crtc.
@@ -7424,7 +7431,8 @@ bool intel_set_mode(struct drm_crtc *crtc,
 	if (modeset_pipes) {
 		adjusted_mode = intel_modeset_adjusted_mode(crtc, mode);
 		if (IS_ERR(adjusted_mode)) {
-			return false;
+			ret = false;
+			goto out;
 		}
 	}
 
@@ -7476,12 +7484,14 @@ bool intel_set_mode(struct drm_crtc *crtc,
 done:
 	drm_mode_destroy(dev, adjusted_mode);
 	if (!ret && crtc->enabled) {
-		crtc->hwmode = saved_hwmode;
-		crtc->mode = saved_mode;
+		crtc->hwmode = *saved_hwmode;
+		crtc->mode = *saved_mode;
 	} else {
 		intel_modeset_check_state(dev);
 	}
 
+out:
+	kfree(saved_mode);
 	return ret;
 }
 
-- 
cgit v0.10.2


From f72b3435c1a75406d82d6e252bb78f009efd4bd9 Mon Sep 17 00:00:00 2001
From: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Date: Mon, 10 Dec 2012 15:41:48 +0200
Subject: drm/i915: Don't emit semaphore wait if wrap happened

If wrap just happened we need to prevent emitting waits for
pre wrap values. Detect this and emit no-ops instead.

v2: Use olr > seqno to detect wrap instead of *seqno == 0
as suggested by Chris Wilson.

v3: Use last used seqno to detect the wraparound. From
Chris Wilson

v4: Fixed unnecessary last_seqno assigment

References: https://bugs.freedesktop.org/show_bug.cgi?id=57967
Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 2ab476d..9da6782 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -671,7 +671,7 @@ typedef struct drm_i915_private {
 
 	struct pci_dev *bridge_dev;
 	struct intel_ring_buffer ring[I915_NUM_RINGS];
-	uint32_t next_seqno;
+	uint32_t last_seqno, next_seqno;
 
 	drm_dma_handle_t *status_page_dmah;
 	struct resource mch_res;
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index e4b233d..a81b78a 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1969,7 +1969,7 @@ i915_gem_get_seqno(struct drm_device *dev, u32 *seqno)
 		dev_priv->next_seqno = 1;
 	}
 
-	*seqno = dev_priv->next_seqno++;
+	*seqno = dev_priv->last_seqno = dev_priv->next_seqno++;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index 0d03dc6..69bbe7b 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -596,6 +596,13 @@ gen6_add_request(struct intel_ring_buffer *ring)
 	return 0;
 }
 
+static inline bool i915_gem_has_seqno_wrapped(struct drm_device *dev,
+					      u32 seqno)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	return dev_priv->last_seqno < seqno;
+}
+
 /**
  * intel_ring_sync - sync the waiter to the signaller on seqno
  *
@@ -626,11 +633,20 @@ gen6_ring_sync(struct intel_ring_buffer *waiter,
 	if (ret)
 		return ret;
 
-	intel_ring_emit(waiter,
-			dw1 | signaller->semaphore_register[waiter->id]);
-	intel_ring_emit(waiter, seqno);
-	intel_ring_emit(waiter, 0);
-	intel_ring_emit(waiter, MI_NOOP);
+	/* If seqno wrap happened, omit the wait with no-ops */
+	if (likely(!i915_gem_has_seqno_wrapped(waiter->dev, seqno))) {
+		intel_ring_emit(waiter,
+				dw1 |
+				signaller->semaphore_register[waiter->id]);
+		intel_ring_emit(waiter, seqno);
+		intel_ring_emit(waiter, 0);
+		intel_ring_emit(waiter, MI_NOOP);
+	} else {
+		intel_ring_emit(waiter, MI_NOOP);
+		intel_ring_emit(waiter, MI_NOOP);
+		intel_ring_emit(waiter, MI_NOOP);
+		intel_ring_emit(waiter, MI_NOOP);
+	}
 	intel_ring_advance(waiter);
 
 	return 0;
-- 
cgit v0.10.2


From 107f27a5df8ad33a351d6e82fe95ff92b428f72e Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Mon, 10 Dec 2012 13:56:17 +0200
Subject: drm/i915: Open-code i915_gpu_idle() for handling seqno wrapping

The complication is that during seqno wrapping we must be extremely
careful not to write to any ring as that will require a new seqno, and
so would recurse back into the seqno wrap handler. So we cannot call
i915_gpu_idle() as that does additional work beyond simply retiring the
current set of requests, and instead must do the minimal work ourselves
during seqno wrapping.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index a81b78a..bc73f98 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1938,11 +1938,15 @@ i915_gem_handle_seqno_wrap(struct drm_device *dev)
 	if (ret == 0)
 		return ret;
 
-	ret = i915_gpu_idle(dev);
-	if (ret)
-		return ret;
-
+	/* Carefully retire all requests without writing to the rings */
+	for_each_ring(ring, dev_priv, i) {
+		ret = intel_ring_idle(ring);
+		if (ret)
+			return ret;
+	}
 	i915_gem_retire_requests(dev);
+
+	/* Finally reset hw state */
 	for_each_ring(ring, dev_priv, i) {
 		ret = intel_ring_handle_seqno_wrap(ring);
 		if (ret)
-- 
cgit v0.10.2


From 9e8e36879f268f1652e11b1c8560bbb67cf4f08e Mon Sep 17 00:00:00 2001
From: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Date: Tue, 4 Dec 2012 15:12:05 +0200
Subject: drm/i915: Set initial seqno value close to wrap boundary

To gain confidence in the wrap handling, make it happen quite
soon after the boot.

Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index bc73f98..6380c60 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -3943,7 +3943,7 @@ i915_gem_init_hw(struct drm_device *dev)
 			goto cleanup_bsd_ring;
 	}
 
-	dev_priv->next_seqno = 1;
+	dev_priv->next_seqno = (u32)-1 - 0x1000;
 
 	/*
 	 * XXX: There was some w/a described somewhere suggesting loading
-- 
cgit v0.10.2


From 20afbda209d708be66944907966486d0c1331cb8 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Tue, 11 Dec 2012 14:05:07 +0100
Subject: drm/i915: Fixup hpd irq register setup ordering

For GMCH platforms we set up the hpd irq registers in the irq
postinstall hook. But since we only enable the irq sources we actually
need in PORT_HOTPLUG_EN/STATUS, taking dev_priv->hotplug_supported_mask
into account, no hpd interrupt sources is enabled since

commit 52d7ecedac3f96fb562cb482c139015372728638
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Sat Dec 1 21:03:22 2012 +0100

    drm/i915: reorder setup sequence to have irqs for output setup

Wrongly set-up interrupts also lead to broken hw-based load-detection
on at least GM45, resulting in ghost VGA/TV-out outputs.

To fix this, delay the hotplug register setup until after all outputs
are set up, by moving it into a new dev_priv->display.hpd_irq_callback.
We might also move the PCH_SPLIT platforms to such a setup eventually.

Another funny part is that we need to delay the fbdev initial config
probing until after the hpd regs are setup, for otherwise it'll detect
ghost outputs. But we can only enable the hpd interrupt handling
itself (and the output polling) _after_ that initial scan, due to
massive locking brain-damage in the fbdev setup code. Add a big
comment to explain this cute little dragon lair.

v2: Encapsulate all the fbdev handling by wrapping the move call into
intel_fbdev_initial_config in intel_fb.c. Requested by Chris Wilson.

v3: Applied bikeshed from Jesse Barnes.

v4: Imre Deak noticed that we also need to call intel_hpd_init after
the drm_irqinstall calls in the gpu reset and resume paths - otherwise
hotplug will be broken. Also improve the comment a bit about why
hpd_init needs to be called before we set up the initial fbdev config.

Bugzilla: Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54943
Reported-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> (v3)
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index 9363066..0c2ab40 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -1319,6 +1319,21 @@ static int i915_load_modeset_init(struct drm_device *dev)
 		goto cleanup_gem;
 
 	/* Only enable hotplug handling once the fbdev is fully set up. */
+	intel_hpd_init(dev);
+
+	/*
+	 * Some ports require correctly set-up hpd registers for detection to
+	 * work properly (leading to ghost connected connector status), e.g. VGA
+	 * on gm45.  Hence we can only set up the initial fbdev config after hpd
+	 * irqs are fully enabled. Now we should scan for the initial config
+	 * only once hotplug handling is enabled, but due to screwed-up locking
+	 * around kms/fbdev init we can't protect the fdbev initial config
+	 * scanning against hotplug events. Hence do this first and ignore the
+	 * tiny window where we will loose hotplug notifactions.
+	 */
+	intel_fbdev_initial_config(dev);
+
+	/* Only enable hotplug handling once the fbdev is fully set up. */
 	dev_priv->enable_hotplug_processing = true;
 
 	drm_kms_helper_poll_init(dev);
diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index a129218..fbd0b28 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -566,6 +566,7 @@ static int __i915_drm_thaw(struct drm_device *dev)
 		intel_modeset_init_hw(dev);
 		intel_modeset_setup_hw_state(dev, false);
 		drm_irq_install(dev);
+		intel_hpd_init(dev);
 	}
 
 	intel_opregion_init(dev);
@@ -871,6 +872,7 @@ int i915_reset(struct drm_device *dev)
 
 		drm_irq_uninstall(dev);
 		drm_irq_install(dev);
+		intel_hpd_init(dev);
 	} else {
 		mutex_unlock(&dev->struct_mutex);
 	}
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 9da6782..e55303e 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -297,6 +297,7 @@ struct drm_i915_display_funcs {
 			  struct drm_i915_gem_object *obj);
 	int (*update_plane)(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 			    int x, int y);
+	void (*hpd_irq_setup)(struct drm_device *dev);
 	/* clock updates for mode set */
 	/* cursor updates */
 	/* render clock increase/decrease */
@@ -1343,6 +1344,7 @@ void i915_hangcheck_elapsed(unsigned long data);
 void i915_handle_error(struct drm_device *dev, bool wedged);
 
 extern void intel_irq_init(struct drm_device *dev);
+extern void intel_hpd_init(struct drm_device *dev);
 extern void intel_gt_init(struct drm_device *dev);
 extern void intel_gt_reset(struct drm_device *dev);
 
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 914ecf4..551f370 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -1995,7 +1995,6 @@ static int valleyview_irq_postinstall(struct drm_device *dev)
 {
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
 	u32 enable_mask;
-	u32 hotplug_en = I915_READ(PORT_HOTPLUG_EN);
 	u32 pipestat_enable = PLANE_FLIP_DONE_INT_EN_VLV;
 	u32 render_irqs;
 	u16 msid;
@@ -2024,6 +2023,9 @@ static int valleyview_irq_postinstall(struct drm_device *dev)
 	msid |= (1<<14);
 	pci_write_config_word(dev_priv->dev->pdev, 0x98, msid);
 
+	I915_WRITE(PORT_HOTPLUG_EN, 0);
+	POSTING_READ(PORT_HOTPLUG_EN);
+
 	I915_WRITE(VLV_IMR, dev_priv->irq_mask);
 	I915_WRITE(VLV_IER, enable_mask);
 	I915_WRITE(VLV_IIR, 0xffffffff);
@@ -2053,6 +2055,15 @@ static int valleyview_irq_postinstall(struct drm_device *dev)
 #endif
 
 	I915_WRITE(VLV_MASTER_IER, MASTER_INTERRUPT_ENABLE);
+
+	return 0;
+}
+
+static void valleyview_hpd_irq_setup(struct drm_device *dev)
+{
+	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
+	u32 hotplug_en = I915_READ(PORT_HOTPLUG_EN);
+
 	/* Note HDMI and DP share bits */
 	if (dev_priv->hotplug_supported_mask & HDMIB_HOTPLUG_INT_STATUS)
 		hotplug_en |= HDMIB_HOTPLUG_INT_EN;
@@ -2070,8 +2081,6 @@ static int valleyview_irq_postinstall(struct drm_device *dev)
 	}
 
 	I915_WRITE(PORT_HOTPLUG_EN, hotplug_en);
-
-	return 0;
 }
 
 static void valleyview_irq_uninstall(struct drm_device *dev)
@@ -2301,6 +2310,9 @@ static int i915_irq_postinstall(struct drm_device *dev)
 		I915_USER_INTERRUPT;
 
 	if (I915_HAS_HOTPLUG(dev)) {
+		I915_WRITE(PORT_HOTPLUG_EN, 0);
+		POSTING_READ(PORT_HOTPLUG_EN);
+
 		/* Enable in IER... */
 		enable_mask |= I915_DISPLAY_PORT_INTERRUPT;
 		/* and unmask in IMR */
@@ -2311,8 +2323,18 @@ static int i915_irq_postinstall(struct drm_device *dev)
 	I915_WRITE(IER, enable_mask);
 	POSTING_READ(IER);
 
+	intel_opregion_enable_asle(dev);
+
+	return 0;
+}
+
+static void i915_hpd_irq_setup(struct drm_device *dev)
+{
+	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
+	u32 hotplug_en;
+
 	if (I915_HAS_HOTPLUG(dev)) {
-		u32 hotplug_en = I915_READ(PORT_HOTPLUG_EN);
+		hotplug_en = I915_READ(PORT_HOTPLUG_EN);
 
 		if (dev_priv->hotplug_supported_mask & HDMIB_HOTPLUG_INT_STATUS)
 			hotplug_en |= HDMIB_HOTPLUG_INT_EN;
@@ -2333,10 +2355,6 @@ static int i915_irq_postinstall(struct drm_device *dev)
 
 		I915_WRITE(PORT_HOTPLUG_EN, hotplug_en);
 	}
-
-	intel_opregion_enable_asle(dev);
-
-	return 0;
 }
 
 static irqreturn_t i915_irq_handler(int irq, void *arg)
@@ -2496,7 +2514,6 @@ static void i965_irq_preinstall(struct drm_device * dev)
 static int i965_irq_postinstall(struct drm_device *dev)
 {
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
-	u32 hotplug_en;
 	u32 enable_mask;
 	u32 error_mask;
 
@@ -2538,6 +2555,19 @@ static int i965_irq_postinstall(struct drm_device *dev)
 	I915_WRITE(IER, enable_mask);
 	POSTING_READ(IER);
 
+	I915_WRITE(PORT_HOTPLUG_EN, 0);
+	POSTING_READ(PORT_HOTPLUG_EN);
+
+	intel_opregion_enable_asle(dev);
+
+	return 0;
+}
+
+static void i965_hpd_irq_setup(struct drm_device *dev)
+{
+	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
+	u32 hotplug_en;
+
 	/* Note HDMI and DP share hotplug bits */
 	hotplug_en = 0;
 	if (dev_priv->hotplug_supported_mask & HDMIB_HOTPLUG_INT_STATUS)
@@ -2572,10 +2602,6 @@ static int i965_irq_postinstall(struct drm_device *dev)
 	/* Ignore TV since it's buggy */
 
 	I915_WRITE(PORT_HOTPLUG_EN, hotplug_en);
-
-	intel_opregion_enable_asle(dev);
-
-	return 0;
 }
 
 static irqreturn_t i965_irq_handler(int irq, void *arg)
@@ -2754,6 +2780,7 @@ void intel_irq_init(struct drm_device *dev)
 		dev->driver->irq_uninstall = valleyview_irq_uninstall;
 		dev->driver->enable_vblank = valleyview_enable_vblank;
 		dev->driver->disable_vblank = valleyview_disable_vblank;
+		dev_priv->display.hpd_irq_setup = valleyview_hpd_irq_setup;
 	} else if (IS_IVYBRIDGE(dev) || IS_HASWELL(dev)) {
 		/* Share pre & uninstall handlers with ILK/SNB */
 		dev->driver->irq_handler = ivybridge_irq_handler;
@@ -2780,13 +2807,23 @@ void intel_irq_init(struct drm_device *dev)
 			dev->driver->irq_postinstall = i915_irq_postinstall;
 			dev->driver->irq_uninstall = i915_irq_uninstall;
 			dev->driver->irq_handler = i915_irq_handler;
+			dev_priv->display.hpd_irq_setup = i915_hpd_irq_setup;
 		} else {
 			dev->driver->irq_preinstall = i965_irq_preinstall;
 			dev->driver->irq_postinstall = i965_irq_postinstall;
 			dev->driver->irq_uninstall = i965_irq_uninstall;
 			dev->driver->irq_handler = i965_irq_handler;
+			dev_priv->display.hpd_irq_setup = i965_hpd_irq_setup;
 		}
 		dev->driver->enable_vblank = i915_enable_vblank;
 		dev->driver->disable_vblank = i915_disable_vblank;
 	}
 }
+
+void intel_hpd_init(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (dev_priv->display.hpd_irq_setup)
+		dev_priv->display.hpd_irq_setup(dev);
+}
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 7ca7772..22728f2 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -587,6 +587,7 @@ extern int intel_framebuffer_init(struct drm_device *dev,
 				  struct drm_mode_fb_cmd2 *mode_cmd,
 				  struct drm_i915_gem_object *obj);
 extern int intel_fbdev_init(struct drm_device *dev);
+extern void intel_fbdev_initial_config(struct drm_device *dev);
 extern void intel_fbdev_fini(struct drm_device *dev);
 extern void intel_fbdev_set_suspend(struct drm_device *dev, int state);
 extern void intel_prepare_page_flip(struct drm_device *dev, int plane);
diff --git a/drivers/gpu/drm/i915/intel_fb.c b/drivers/gpu/drm/i915/intel_fb.c
index b7773e5..82289678 100644
--- a/drivers/gpu/drm/i915/intel_fb.c
+++ b/drivers/gpu/drm/i915/intel_fb.c
@@ -243,10 +243,18 @@ int intel_fbdev_init(struct drm_device *dev)
 	}
 
 	drm_fb_helper_single_add_all_connectors(&ifbdev->helper);
-	drm_fb_helper_initial_config(&ifbdev->helper, 32);
+
 	return 0;
 }
 
+void intel_fbdev_initial_config(struct drm_device *dev)
+{
+	drm_i915_private_t *dev_priv = dev->dev_private;
+
+	/* Due to peculiar init order wrt to hpd handling this is separate. */
+	drm_fb_helper_initial_config(&dev_priv->fbdev->helper, 32);
+}
+
 void intel_fbdev_fini(struct drm_device *dev)
 {
 	drm_i915_private_t *dev_priv = dev->dev_private;
-- 
cgit v0.10.2


From 09153000b8ca32a539a1207edebabd0d40b6c61b Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Wed, 12 Dec 2012 14:06:44 +0100
Subject: drm/i915: rework locking for intel_dpio|sbi_read|write

Spinning for up to 200 us with interrupts locked out is not good. So
let's just spin (and even that seems to be excessive).

And we don't call these functions from interrupt context, so this is
not required. Besides that doing anything in interrupt contexts which
might take a few hundred us is a no-go. So just convert the entire
thing to a mutex. Also move the mutex-grabbing out of the read/write
functions (add a WARN_ON(!is_locked)) instead) since all callers are
nicely grouped together.

Finally the real motivation for this change: Dont grab the modeset
mutex in the dpio debugfs file, we don't need that consistency. And
correctness of the dpio interface is ensured with the dpio_lock.

Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index 7e516ee..7047c4a9f 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -1633,7 +1633,7 @@ static int i915_dpio_info(struct seq_file *m, void *data)
 		return 0;
 	}
 
-	ret = mutex_lock_interruptible(&dev->mode_config.mutex);
+	ret = mutex_lock_interruptible(&dev_priv->dpio_lock);
 	if (ret)
 		return ret;
 
@@ -1662,7 +1662,7 @@ static int i915_dpio_info(struct seq_file *m, void *data)
 	seq_printf(m, "DPIO_FASTCLK_DISABLE: 0x%08x\n",
 		   intel_dpio_read(dev_priv, DPIO_FASTCLK_DISABLE));
 
-	mutex_unlock(&dev->mode_config.mutex);
+	mutex_unlock(&dev_priv->dpio_lock);
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index 0c2ab40..a99c46d 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -1599,7 +1599,7 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 	spin_lock_init(&dev_priv->irq_lock);
 	spin_lock_init(&dev_priv->error_lock);
 	spin_lock_init(&dev_priv->rps.lock);
-	spin_lock_init(&dev_priv->dpio_lock);
+	mutex_init(&dev_priv->dpio_lock);
 
 	mutex_init(&dev_priv->rps.hw_lock);
 
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index e55303e..514aee8 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -686,7 +686,7 @@ typedef struct drm_i915_private {
 	struct pm_qos_request pm_qos;
 
 	/* DPIO indirect register protection */
-	spinlock_t dpio_lock;
+	struct mutex dpio_lock;
 
 	/** Cached value of IMR to avoid reads in updating the bitfield */
 	u32 pipestat[2];
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 115bf62..2a01b09 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -416,13 +416,11 @@ static const intel_limit_t intel_limits_vlv_dp = {
 
 u32 intel_dpio_read(struct drm_i915_private *dev_priv, int reg)
 {
-	unsigned long flags;
-	u32 val = 0;
+	WARN_ON(!mutex_is_locked(&dev_priv->dpio_lock));
 
-	spin_lock_irqsave(&dev_priv->dpio_lock, flags);
 	if (wait_for_atomic_us((I915_READ(DPIO_PKT) & DPIO_BUSY) == 0, 100)) {
 		DRM_ERROR("DPIO idle wait timed out\n");
-		goto out_unlock;
+		return 0;
 	}
 
 	I915_WRITE(DPIO_REG, reg);
@@ -430,24 +428,20 @@ u32 intel_dpio_read(struct drm_i915_private *dev_priv, int reg)
 		   DPIO_BYTE);
 	if (wait_for_atomic_us((I915_READ(DPIO_PKT) & DPIO_BUSY) == 0, 100)) {
 		DRM_ERROR("DPIO read wait timed out\n");
-		goto out_unlock;
+		return 0;
 	}
-	val = I915_READ(DPIO_DATA);
 
-out_unlock:
-	spin_unlock_irqrestore(&dev_priv->dpio_lock, flags);
-	return val;
+	return I915_READ(DPIO_DATA);
 }
 
 static void intel_dpio_write(struct drm_i915_private *dev_priv, int reg,
 			     u32 val)
 {
-	unsigned long flags;
+	WARN_ON(!mutex_is_locked(&dev_priv->dpio_lock));
 
-	spin_lock_irqsave(&dev_priv->dpio_lock, flags);
 	if (wait_for_atomic_us((I915_READ(DPIO_PKT) & DPIO_BUSY) == 0, 100)) {
 		DRM_ERROR("DPIO idle wait timed out\n");
-		goto out_unlock;
+		return;
 	}
 
 	I915_WRITE(DPIO_DATA, val);
@@ -456,9 +450,6 @@ static void intel_dpio_write(struct drm_i915_private *dev_priv, int reg,
 		   DPIO_BYTE);
 	if (wait_for_atomic_us((I915_READ(DPIO_PKT) & DPIO_BUSY) == 0, 100))
 		DRM_ERROR("DPIO write wait timed out\n");
-
-out_unlock:
-       spin_unlock_irqrestore(&dev_priv->dpio_lock, flags);
 }
 
 static void vlv_init_dpio(struct drm_device *dev)
@@ -1455,13 +1446,12 @@ static void intel_disable_pll(struct drm_i915_private *dev_priv, enum pipe pipe)
 static void
 intel_sbi_write(struct drm_i915_private *dev_priv, u16 reg, u32 value)
 {
-	unsigned long flags;
+	WARN_ON(!mutex_is_locked(&dev_priv->dpio_lock));
 
-	spin_lock_irqsave(&dev_priv->dpio_lock, flags);
 	if (wait_for((I915_READ(SBI_CTL_STAT) & SBI_BUSY) == 0,
 				100)) {
 		DRM_ERROR("timeout waiting for SBI to become ready\n");
-		goto out_unlock;
+		return;
 	}
 
 	I915_WRITE(SBI_ADDR,
@@ -1475,24 +1465,19 @@ intel_sbi_write(struct drm_i915_private *dev_priv, u16 reg, u32 value)
 	if (wait_for((I915_READ(SBI_CTL_STAT) & (SBI_BUSY | SBI_RESPONSE_FAIL)) == 0,
 				100)) {
 		DRM_ERROR("timeout waiting for SBI to complete write transaction\n");
-		goto out_unlock;
+		return;
 	}
-
-out_unlock:
-	spin_unlock_irqrestore(&dev_priv->dpio_lock, flags);
 }
 
 static u32
 intel_sbi_read(struct drm_i915_private *dev_priv, u16 reg)
 {
-	unsigned long flags;
-	u32 value = 0;
+	WARN_ON(!mutex_is_locked(&dev_priv->dpio_lock));
 
-	spin_lock_irqsave(&dev_priv->dpio_lock, flags);
 	if (wait_for((I915_READ(SBI_CTL_STAT) & SBI_BUSY) == 0,
 				100)) {
 		DRM_ERROR("timeout waiting for SBI to become ready\n");
-		goto out_unlock;
+		return 0;
 	}
 
 	I915_WRITE(SBI_ADDR,
@@ -1504,14 +1489,10 @@ intel_sbi_read(struct drm_i915_private *dev_priv, u16 reg)
 	if (wait_for((I915_READ(SBI_CTL_STAT) & (SBI_BUSY | SBI_RESPONSE_FAIL)) == 0,
 				100)) {
 		DRM_ERROR("timeout waiting for SBI to complete read transaction\n");
-		goto out_unlock;
+		return 0;
 	}
 
-	value = I915_READ(SBI_DATA);
-
-out_unlock:
-	spin_unlock_irqrestore(&dev_priv->dpio_lock, flags);
-	return value;
+	return I915_READ(SBI_DATA);
 }
 
 /**
@@ -2924,6 +2905,8 @@ static void lpt_program_iclkip(struct drm_crtc *crtc)
 	u32 divsel, phaseinc, auxdiv, phasedir = 0;
 	u32 temp;
 
+	mutex_lock(&dev_priv->dpio_lock);
+
 	/* It is necessary to ungate the pixclk gate prior to programming
 	 * the divisors, and gate it back when it is done.
 	 */
@@ -3005,6 +2988,8 @@ static void lpt_program_iclkip(struct drm_crtc *crtc)
 	udelay(24);
 
 	I915_WRITE(PIXCLK_GATE, PIXCLK_GATE_UNGATE);
+
+	mutex_unlock(&dev_priv->dpio_lock);
 }
 
 /*
@@ -4222,6 +4207,8 @@ static void vlv_update_pll(struct drm_crtc *crtc,
 	bool is_sdvo;
 	u32 temp;
 
+	mutex_lock(&dev_priv->dpio_lock);
+
 	is_sdvo = intel_pipe_has_type(crtc, INTEL_OUTPUT_SDVO) ||
 		intel_pipe_has_type(crtc, INTEL_OUTPUT_HDMI);
 
@@ -4305,6 +4292,8 @@ static void vlv_update_pll(struct drm_crtc *crtc,
 			temp |= (1 << 21);
 		intel_dpio_write(dev_priv, DPIO_DATA_CHANNEL2, temp);
 	}
+
+	mutex_unlock(&dev_priv->dpio_lock);
 }
 
 static void i9xx_update_pll(struct drm_crtc *crtc,
-- 
cgit v0.10.2


From b696519e51910943f084ea0fc8bdceceae8383ef Mon Sep 17 00:00:00 2001
From: Damien Lespiau <damien.lespiau@intel.com>
Date: Thu, 13 Dec 2012 16:08:59 +0000
Subject: drm/i915: Cleanup SHOTPLUG_CTL status bits definitions

Those status bits don't follow the usual pattern: _MASK (those bits are
write 1 to clear, useful to select the value we want to read) and the
values shifted by the same amount.

Cleaned that that up when poking at the register for testing purposes,
might as well upstream that cleanup.

Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index f2a5ea6..f834804 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -3558,27 +3558,30 @@
 #define PORTD_PULSE_DURATION_6ms        (2 << 18)
 #define PORTD_PULSE_DURATION_100ms      (3 << 18)
 #define PORTD_PULSE_DURATION_MASK	(3 << 18)
-#define PORTD_HOTPLUG_NO_DETECT         (0)
-#define PORTD_HOTPLUG_SHORT_DETECT      (1 << 16)
-#define PORTD_HOTPLUG_LONG_DETECT       (1 << 17)
+#define PORTD_HOTPLUG_STATUS_MASK	(0x3 << 16)
+#define  PORTD_HOTPLUG_NO_DETECT	(0 << 16)
+#define  PORTD_HOTPLUG_SHORT_DETECT	(1 << 16)
+#define  PORTD_HOTPLUG_LONG_DETECT	(2 << 16)
 #define PORTC_HOTPLUG_ENABLE            (1 << 12)
 #define PORTC_PULSE_DURATION_2ms        (0)
 #define PORTC_PULSE_DURATION_4_5ms      (1 << 10)
 #define PORTC_PULSE_DURATION_6ms        (2 << 10)
 #define PORTC_PULSE_DURATION_100ms      (3 << 10)
 #define PORTC_PULSE_DURATION_MASK	(3 << 10)
-#define PORTC_HOTPLUG_NO_DETECT         (0)
-#define PORTC_HOTPLUG_SHORT_DETECT      (1 << 8)
-#define PORTC_HOTPLUG_LONG_DETECT       (1 << 9)
+#define PORTC_HOTPLUG_STATUS_MASK	(0x3 << 8)
+#define  PORTC_HOTPLUG_NO_DETECT	(0 << 8)
+#define  PORTC_HOTPLUG_SHORT_DETECT	(1 << 8)
+#define  PORTC_HOTPLUG_LONG_DETECT	(2 << 8)
 #define PORTB_HOTPLUG_ENABLE            (1 << 4)
 #define PORTB_PULSE_DURATION_2ms        (0)
 #define PORTB_PULSE_DURATION_4_5ms      (1 << 2)
 #define PORTB_PULSE_DURATION_6ms        (2 << 2)
 #define PORTB_PULSE_DURATION_100ms      (3 << 2)
 #define PORTB_PULSE_DURATION_MASK	(3 << 2)
-#define PORTB_HOTPLUG_NO_DETECT         (0)
-#define PORTB_HOTPLUG_SHORT_DETECT      (1 << 0)
-#define PORTB_HOTPLUG_LONG_DETECT       (1 << 1)
+#define PORTB_HOTPLUG_STATUS_MASK	(0x3 << 0)
+#define  PORTB_HOTPLUG_NO_DETECT	(0 << 0)
+#define  PORTB_HOTPLUG_SHORT_DETECT	(1 << 0)
+#define  PORTB_HOTPLUG_LONG_DETECT	(2 << 0)
 
 #define PCH_GPIOA               0xc5010
 #define PCH_GPIOB               0xc5014
-- 
cgit v0.10.2


From b0ea7d37a8f63eeec5ae80b4a6403cfba01da02f Mon Sep 17 00:00:00 2001
From: Damien Lespiau <damien.lespiau@intel.com>
Date: Thu, 13 Dec 2012 16:09:00 +0000
Subject: drm/i915/hdmi: Read the HPD status before trying to read the EDID

If you unplug the hdmi connector slowly enough, the hotplug interrupt
fires but then the kernel code tries to read the EDID and succeeds
(because the connector is still half connected, the HPD pin is shorter
than the others, and DDC works). Since EDID succeeds it thinks the
monitor is still connected.

To prevent that, read the live HPD status in the hotplug handler before
trying to read the EDID.

v2: Rename the function to ibx_ (Chris Wilson)

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=55372
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 2a01b09..6adeabd 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -985,6 +985,39 @@ void intel_wait_for_pipe_off(struct drm_device *dev, int pipe)
 	}
 }
 
+/*
+ * ibx_digital_port_connected - is the specified port connected?
+ * @dev_priv: i915 private structure
+ * @port: the port to test
+ *
+ * Returns true if @port is connected, false otherwise.
+ */
+bool ibx_digital_port_connected(struct drm_i915_private *dev_priv,
+				struct intel_digital_port *port)
+{
+	u32 bit;
+
+	/* XXX: IBX has different SDEISR bits */
+	if (HAS_PCH_IBX(dev_priv->dev))
+		return true;
+
+	switch(port->port) {
+	case PORT_B:
+		bit = SDE_PORTB_HOTPLUG_CPT;
+		break;
+	case PORT_C:
+		bit = SDE_PORTC_HOTPLUG_CPT;
+		break;
+	case PORT_D:
+		bit = SDE_PORTD_HOTPLUG_CPT;
+		break;
+	default:
+		return true;
+	}
+
+	return I915_READ(SDEISR) & bit;
+}
+
 static const char *state_string(bool enabled)
 {
 	return enabled ? "on" : "off";
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 22728f2..53d4c8f 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -545,6 +545,9 @@ hdmi_to_dig_port(struct intel_hdmi *intel_hdmi)
 	return container_of(intel_hdmi, struct intel_digital_port, hdmi);
 }
 
+bool ibx_digital_port_connected(struct drm_i915_private *dev_priv,
+				struct intel_digital_port *port);
+
 extern void intel_connector_attach_encoder(struct intel_connector *connector,
 					   struct intel_encoder *encoder);
 extern struct drm_encoder *intel_best_encoder(struct drm_connector *connector);
diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c
index 53df0a8..9f834d3 100644
--- a/drivers/gpu/drm/i915/intel_hdmi.c
+++ b/drivers/gpu/drm/i915/intel_hdmi.c
@@ -793,16 +793,21 @@ static bool g4x_hdmi_connected(struct intel_hdmi *intel_hdmi)
 static enum drm_connector_status
 intel_hdmi_detect(struct drm_connector *connector, bool force)
 {
+	struct drm_device *dev = connector->dev;
 	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 	struct intel_digital_port *intel_dig_port =
 		hdmi_to_dig_port(intel_hdmi);
 	struct intel_encoder *intel_encoder = &intel_dig_port->base;
-	struct drm_i915_private *dev_priv = connector->dev->dev_private;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct edid *edid;
 	enum drm_connector_status status = connector_status_disconnected;
 
-	if (IS_G4X(connector->dev) && !g4x_hdmi_connected(intel_hdmi))
+
+	if (IS_G4X(dev) && !g4x_hdmi_connected(intel_hdmi))
 		return status;
+	else if (HAS_PCH_SPLIT(dev) &&
+		 !ibx_digital_port_connected(dev_priv, intel_dig_port))
+		 return status;
 
 	intel_hdmi->has_hdmi_sink = false;
 	intel_hdmi->has_audio = false;
-- 
cgit v0.10.2


From 1b4696394aeb8a550f8537e92ae6cc65f444dca0 Mon Sep 17 00:00:00 2001
From: Damien Lespiau <damien.lespiau@intel.com>
Date: Thu, 13 Dec 2012 16:09:01 +0000
Subject: drm/i915/dp: Read the HPD status before trying to read the DPCD

Just like:

  Author: Damien Lespiau <damien.lespiau@intel.com>
  Date:   Wed Dec 12 19:37:22 2012 +0000

      drm/i915/hdmi: Read the HPD status before trying to read the EDID

But this time for DiplayPort.

v2: Adapt to the ibx_ name change and don't add commit hash (Chris
Wilson, Jani Nikula)

Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index b2130bc..19a0d89 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -2248,6 +2248,8 @@ static enum drm_connector_status
 ironlake_dp_detect(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	enum drm_connector_status status;
 
 	/* Can't disconnect eDP, but you can close the lid... */
@@ -2258,6 +2260,9 @@ ironlake_dp_detect(struct intel_dp *intel_dp)
 		return status;
 	}
 
+	if (!ibx_digital_port_connected(dev_priv, intel_dig_port))
+		return connector_status_disconnected;
+
 	return intel_dp_detect_dpcd(intel_dp);
 }
 
-- 
cgit v0.10.2


From 577c7a505b5601a9a441039dd37543775f9af8f5 Mon Sep 17 00:00:00 2001
From: Damien Lespiau <damien.lespiau@intel.com>
Date: Thu, 13 Dec 2012 16:09:02 +0000
Subject: drm/i915/dp: Log the DPCD only if we have successfully retrieved one

Moving the DPCD just after a successful read will allow to:
  - log all DPCD reads (eDP ones, changes signalled by HPD IRQ)
  - don't log it if we haven't been able to read it

v2: Be sure to log the DPCD when a downstream port does not have HPD
    support and the branch device asserts HPD (Jani Nikula)

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 19a0d89..5ed8fb3 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -2084,10 +2084,16 @@ intel_dp_link_down(struct intel_dp *intel_dp)
 static bool
 intel_dp_get_dpcd(struct intel_dp *intel_dp)
 {
+	char dpcd_hex_dump[sizeof(intel_dp->dpcd) * 3];
+
 	if (intel_dp_aux_native_read_retry(intel_dp, 0x000, intel_dp->dpcd,
 					   sizeof(intel_dp->dpcd)) == 0)
 		return false; /* aux transfer failed */
 
+	hex_dump_to_buffer(intel_dp->dpcd, sizeof(intel_dp->dpcd),
+			   32, 1, dpcd_hex_dump, sizeof(dpcd_hex_dump), false);
+	DRM_DEBUG_KMS("DPCD: %s\n", dpcd_hex_dump);
+
 	if (intel_dp->dpcd[DP_DPCD_REV] == 0)
 		return false; /* DPCD not present */
 
@@ -2353,7 +2359,6 @@ intel_dp_detect(struct drm_connector *connector, bool force)
 	struct drm_device *dev = connector->dev;
 	enum drm_connector_status status;
 	struct edid *edid = NULL;
-	char dpcd_hex_dump[sizeof(intel_dp->dpcd) * 3];
 
 	intel_dp->has_audio = false;
 
@@ -2362,10 +2367,6 @@ intel_dp_detect(struct drm_connector *connector, bool force)
 	else
 		status = g4x_dp_detect(intel_dp);
 
-	hex_dump_to_buffer(intel_dp->dpcd, sizeof(intel_dp->dpcd),
-			   32, 1, dpcd_hex_dump, sizeof(dpcd_hex_dump), false);
-	DRM_DEBUG_KMS("DPCD: %s\n", dpcd_hex_dump);
-
 	if (status != connector_status_connected)
 		return status;
 
-- 
cgit v0.10.2


From c36346e302b3a36bfb10b58167b69ab7ac95f10a Mon Sep 17 00:00:00 2001
From: Damien Lespiau <damien.lespiau@intel.com>
Date: Thu, 13 Dec 2012 16:09:03 +0000
Subject: drm/i915: Implement ibx_digital_port_connected() for IBX

CPT+ PCHs have different bit definition to read the HPD live status. I
don't have an ILK with digital ports handy, which is why this patch is
separate from the CPT+ implementation. If the docs don't lie, it should
all be fine though.

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 6adeabd..7f1b07c 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -997,22 +997,34 @@ bool ibx_digital_port_connected(struct drm_i915_private *dev_priv,
 {
 	u32 bit;
 
-	/* XXX: IBX has different SDEISR bits */
-	if (HAS_PCH_IBX(dev_priv->dev))
-		return true;
-
-	switch(port->port) {
-	case PORT_B:
-		bit = SDE_PORTB_HOTPLUG_CPT;
-		break;
-	case PORT_C:
-		bit = SDE_PORTC_HOTPLUG_CPT;
-		break;
-	case PORT_D:
-		bit = SDE_PORTD_HOTPLUG_CPT;
-		break;
-	default:
-		return true;
+	if (HAS_PCH_IBX(dev_priv->dev)) {
+		switch(port->port) {
+		case PORT_B:
+			bit = SDE_PORTB_HOTPLUG;
+			break;
+		case PORT_C:
+			bit = SDE_PORTC_HOTPLUG;
+			break;
+		case PORT_D:
+			bit = SDE_PORTD_HOTPLUG;
+			break;
+		default:
+			return true;
+		}
+	} else {
+		switch(port->port) {
+		case PORT_B:
+			bit = SDE_PORTB_HOTPLUG_CPT;
+			break;
+		case PORT_C:
+			bit = SDE_PORTC_HOTPLUG_CPT;
+			break;
+		case PORT_D:
+			bit = SDE_PORTD_HOTPLUG_CPT;
+			break;
+		default:
+			return true;
+		}
 	}
 
 	return I915_READ(SDEISR) & bit;
-- 
cgit v0.10.2


From 3f8c65d60467c6ccf4d690f5266567d6c423ae7d Mon Sep 17 00:00:00 2001
From: Damien Lespiau <damien.lespiau@intel.com>
Date: Thu, 13 Dec 2012 16:09:04 +0000
Subject: drm/i915: Remove stale comment about intel_dp_detect()

The function doesn't use any of the registers mentioned, nor does it
return true or false. Hard to do worse. Remove it, the function is
absolutely descriptive enough to not need any comment.

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 5ed8fb3..185bf4e 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -2343,13 +2343,6 @@ intel_dp_get_edid_modes(struct drm_connector *connector, struct i2c_adapter *ada
 	return intel_ddc_get_modes(connector, adapter);
 }
 
-
-/**
- * Uses CRT_HOTPLUG_EN and CRT_HOTPLUG_STAT to detect DP connection.
- *
- * \return true if DP port is connected.
- * \return false if DP port is disconnected.
- */
 static enum drm_connector_status
 intel_dp_detect(struct drm_connector *connector, bool force)
 {
-- 
cgit v0.10.2


From eb119bd612906519cacef2536a9a524c2da5f7fb Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Sun, 16 Dec 2012 12:43:36 +0000
Subject: drm/i915: Access to snooped system memory through the GTT is
 incoherent

We ignore all the user requests to handle flushing to the GTT domain if
the user requests such on a snoopable bo, and as such access through the
GTT to such pages remains incoherent. The specs even warn that such
behaviour is undefined - a strong reason never to do so.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 6380c60..d15c862 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1341,6 +1341,12 @@ int i915_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 
 	trace_i915_gem_object_fault(obj, page_offset, true, write);
 
+	/* Access to snoopable pages through the GTT is incoherent. */
+	if (obj->cache_level != I915_CACHE_NONE && !HAS_LLC(dev)) {
+		ret = -EINVAL;
+		goto unlock;
+	}
+
 	/* Now bind it into the GTT if needed */
 	ret = i915_gem_object_pin(obj, 0, true, false);
 	if (ret)
-- 
cgit v0.10.2


From 88afe715dd5469bc24ca7a19ac62dd3c241cab48 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Sun, 16 Dec 2012 12:15:41 +0000
Subject: drm/i915: Clear the stolen fb before enabling

As the stolen memory region will contain the contents of whatever was
last there, it invariably contains garbage. To be consistent with the
shmemfs backed fb and the expectations of the fb layer, we need to clear
the fb prior to installing it as an fbcon.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=58111
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
[danvet: Fixup sparse __iomem confusion reported by Wu Fengguang.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_fb.c b/drivers/gpu/drm/i915/intel_fb.c
index 82289678..71d5580 100644
--- a/drivers/gpu/drm/i915/intel_fb.c
+++ b/drivers/gpu/drm/i915/intel_fb.c
@@ -155,6 +155,13 @@ static int intelfb_create(struct intel_fbdev *ifbdev,
 	drm_fb_helper_fill_fix(info, fb->pitches[0], fb->depth);
 	drm_fb_helper_fill_var(info, &ifbdev->helper, sizes->fb_width, sizes->fb_height);
 
+	/* If the object is shmemfs backed, it will have given us zeroed pages.
+	 * If the object is stolen however, it will be full of whatever
+	 * garbage was left in there.
+	 */
+	if (ifbdev->ifb.obj->stolen)
+		memset_io(info->screen_base, 0, info->screen_size);
+
 	/* Use default scratch pixmap (info->pixmap.flags = FB_PIXMAP_SYSTEM) */
 
 	DRM_DEBUG_KMS("allocated %dx%d fb: 0x%08x, bo %p\n",
-- 
cgit v0.10.2


From 681e5811f88b7ecca77734421adafb62f19563be Mon Sep 17 00:00:00 2001
From: Paulo Zanoni <paulo.r.zanoni@intel.com>
Date: Thu, 6 Dec 2012 11:12:38 -0200
Subject: drm/i915: check for the PCH when setting pch_transcoder

Don't check the CPU, it doesn't have any PCH transcoder.

Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 7f1b07c..fa73b4a 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -1789,7 +1789,7 @@ static void intel_enable_pipe(struct drm_i915_private *dev_priv, enum pipe pipe,
 	int reg;
 	u32 val;
 
-	if (IS_HASWELL(dev_priv->dev))
+	if (HAS_PCH_LPT(dev_priv->dev))
 		pch_transcoder = TRANSCODER_A;
 	else
 		pch_transcoder = pipe;
-- 
cgit v0.10.2


From a0e63c22ee6daae8fda98e1c6eaff5f5f1e2ba31 Mon Sep 17 00:00:00 2001
From: Paulo Zanoni <paulo.r.zanoni@intel.com>
Date: Thu, 6 Dec 2012 11:12:39 -0200
Subject: drm/i915: remove leftover display.update_wm assignment

This was moved to intel_init_pm.

Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index fa73b4a..f02ec7b 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -8284,8 +8284,7 @@ static void intel_init_display(struct drm_device *dev)
 		} else if (IS_HASWELL(dev)) {
 			dev_priv->display.fdi_link_train = hsw_fdi_link_train;
 			dev_priv->display.write_eld = haswell_write_eld;
-		} else
-			dev_priv->display.update_wm = NULL;
+		}
 	} else if (IS_G4X(dev)) {
 		dev_priv->display.write_eld = g4x_write_eld;
 	}
-- 
cgit v0.10.2


From f0a3424e964ec8e934497a0724cd1d58690fd9ab Mon Sep 17 00:00:00 2001
From: Paulo Zanoni <paulo.r.zanoni@intel.com>
Date: Thu, 6 Dec 2012 16:51:50 -0200
Subject: drm/i915: add intel_dp_set_signal_levels

So we can de-duplicate code that's inside intel_dp_start_link_train
and intel_dp_complete_link_train.

V2: Rebase since patch 3/5 was discarded.

Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 185bf4e..2d3b268 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -1587,7 +1587,7 @@ intel_get_adjust_train(struct intel_dp *intel_dp, uint8_t link_status[DP_LINK_ST
 }
 
 static uint32_t
-intel_dp_signal_levels(uint8_t train_set)
+intel_gen4_signal_levels(uint8_t train_set)
 {
 	uint32_t	signal_levels = 0;
 
@@ -1685,7 +1685,7 @@ intel_gen7_edp_signal_levels(uint8_t train_set)
 
 /* Gen7.5's (HSW) DP voltage swing and pre-emphasis control */
 static uint32_t
-intel_dp_signal_levels_hsw(uint8_t train_set)
+intel_hsw_signal_levels(uint8_t train_set)
 {
 	int signal_levels = train_set & (DP_TRAIN_VOLTAGE_SWING_MASK |
 					 DP_TRAIN_PRE_EMPHASIS_MASK);
@@ -1717,6 +1717,34 @@ intel_dp_signal_levels_hsw(uint8_t train_set)
 	}
 }
 
+/* Properly updates "DP" with the correct signal levels. */
+static void
+intel_dp_set_signal_levels(struct intel_dp *intel_dp, uint32_t *DP)
+{
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct drm_device *dev = intel_dig_port->base.base.dev;
+	uint32_t signal_levels, mask;
+	uint8_t train_set = intel_dp->train_set[0];
+
+	if (IS_HASWELL(dev)) {
+		signal_levels = intel_hsw_signal_levels(train_set);
+		mask = DDI_BUF_EMP_MASK;
+	} else if (IS_GEN7(dev) && is_cpu_edp(intel_dp) && !IS_VALLEYVIEW(dev)) {
+		signal_levels = intel_gen7_edp_signal_levels(train_set);
+		mask = EDP_LINK_TRAIN_VOL_EMP_MASK_IVB;
+	} else if (IS_GEN6(dev) && is_cpu_edp(intel_dp)) {
+		signal_levels = intel_gen6_edp_signal_levels(train_set);
+		mask = EDP_LINK_TRAIN_VOL_EMP_MASK_SNB;
+	} else {
+		signal_levels = intel_gen4_signal_levels(train_set);
+		mask = DP_VOLTAGE_MASK | DP_PRE_EMPHASIS_MASK;
+	}
+
+	DRM_DEBUG_KMS("Using signal levels %08x\n", signal_levels);
+
+	*DP = (*DP & ~mask) | signal_levels;
+}
+
 static bool
 intel_dp_set_link_train(struct intel_dp *intel_dp,
 			uint32_t dp_reg_value,
@@ -1853,24 +1881,8 @@ intel_dp_start_link_train(struct intel_dp *intel_dp)
 	for (;;) {
 		/* Use intel_dp->train_set[0] to set the voltage and pre emphasis values */
 		uint8_t	    link_status[DP_LINK_STATUS_SIZE];
-		uint32_t    signal_levels;
-
-		if (IS_HASWELL(dev)) {
-			signal_levels = intel_dp_signal_levels_hsw(
-							intel_dp->train_set[0]);
-			DP = (DP & ~DDI_BUF_EMP_MASK) | signal_levels;
-		} else if (IS_GEN7(dev) && is_cpu_edp(intel_dp) && !IS_VALLEYVIEW(dev)) {
-			signal_levels = intel_gen7_edp_signal_levels(intel_dp->train_set[0]);
-			DP = (DP & ~EDP_LINK_TRAIN_VOL_EMP_MASK_IVB) | signal_levels;
-		} else if (IS_GEN6(dev) && is_cpu_edp(intel_dp)) {
-			signal_levels = intel_gen6_edp_signal_levels(intel_dp->train_set[0]);
-			DP = (DP & ~EDP_LINK_TRAIN_VOL_EMP_MASK_SNB) | signal_levels;
-		} else {
-			signal_levels = intel_dp_signal_levels(intel_dp->train_set[0]);
-			DP = (DP & ~(DP_VOLTAGE_MASK|DP_PRE_EMPHASIS_MASK)) | signal_levels;
-		}
-		DRM_DEBUG_KMS("training pattern 1 signal levels %08x\n",
-			      signal_levels);
+
+		intel_dp_set_signal_levels(intel_dp, &DP);
 
 		/* Set training pattern 1 */
 		if (!intel_dp_set_link_train(intel_dp, DP,
@@ -1926,7 +1938,6 @@ intel_dp_start_link_train(struct intel_dp *intel_dp)
 void
 intel_dp_complete_link_train(struct intel_dp *intel_dp)
 {
-	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	bool channel_eq = false;
 	int tries, cr_tries;
 	uint32_t DP = intel_dp->DP;
@@ -1936,8 +1947,6 @@ intel_dp_complete_link_train(struct intel_dp *intel_dp)
 	cr_tries = 0;
 	channel_eq = false;
 	for (;;) {
-		/* Use intel_dp->train_set[0] to set the voltage and pre emphasis values */
-		uint32_t    signal_levels;
 		uint8_t	    link_status[DP_LINK_STATUS_SIZE];
 
 		if (cr_tries > 5) {
@@ -1946,19 +1955,7 @@ intel_dp_complete_link_train(struct intel_dp *intel_dp)
 			break;
 		}
 
-		if (IS_HASWELL(dev)) {
-			signal_levels = intel_dp_signal_levels_hsw(intel_dp->train_set[0]);
-			DP = (DP & ~DDI_BUF_EMP_MASK) | signal_levels;
-		} else if (IS_GEN7(dev) && is_cpu_edp(intel_dp) && !IS_VALLEYVIEW(dev)) {
-			signal_levels = intel_gen7_edp_signal_levels(intel_dp->train_set[0]);
-			DP = (DP & ~EDP_LINK_TRAIN_VOL_EMP_MASK_IVB) | signal_levels;
-		} else if (IS_GEN6(dev) && is_cpu_edp(intel_dp)) {
-			signal_levels = intel_gen6_edp_signal_levels(intel_dp->train_set[0]);
-			DP = (DP & ~EDP_LINK_TRAIN_VOL_EMP_MASK_SNB) | signal_levels;
-		} else {
-			signal_levels = intel_dp_signal_levels(intel_dp->train_set[0]);
-			DP = (DP & ~(DP_VOLTAGE_MASK|DP_PRE_EMPHASIS_MASK)) | signal_levels;
-		}
+		intel_dp_set_signal_levels(intel_dp, &DP);
 
 		/* channel eq pattern */
 		if (!intel_dp_set_link_train(intel_dp, DP,
-- 
cgit v0.10.2


From dfd07d72cf70cc8845006b698b55369f5d9c2733 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Mon, 17 Dec 2012 11:21:38 +0100
Subject: drm/i915: clean up PIPECONF bpc #defines

Ilk+ somehow used #defines in near the PIPESTAT definitions, which
decently confused me. Earlier platforms called it BPP instead of
BPC. Clean this all up.

Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index f834804..5b3020f 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -2648,11 +2648,11 @@
 #define   PIPECONF_INTERLACED_DBL_ILK		(4 << 21) /* ilk/snb only */
 #define   PIPECONF_PFIT_PF_INTERLACED_DBL_ILK	(5 << 21) /* ilk/snb only */
 #define   PIPECONF_CXSR_DOWNCLOCK	(1<<16)
-#define   PIPECONF_BPP_MASK	(0x000000e0)
-#define   PIPECONF_BPP_8	(0<<5)
-#define   PIPECONF_BPP_10	(1<<5)
-#define   PIPECONF_BPP_6	(2<<5)
-#define   PIPECONF_BPP_12	(3<<5)
+#define   PIPECONF_BPC_MASK	(0x7 << 5)
+#define   PIPECONF_8BPC		(0<<5)
+#define   PIPECONF_10BPC	(1<<5)
+#define   PIPECONF_6BPC		(2<<5)
+#define   PIPECONF_12BPC	(3<<5)
 #define   PIPECONF_DITHER_EN	(1<<4)
 #define   PIPECONF_DITHER_TYPE_MASK (0x0000000c)
 #define   PIPECONF_DITHER_TYPE_SP (0<<2)
@@ -2696,11 +2696,6 @@
 #define   PIPE_START_VBLANK_INTERRUPT_STATUS	(1UL<<2) /* 965 or later */
 #define   PIPE_VBLANK_INTERRUPT_STATUS		(1UL<<1)
 #define   PIPE_OVERLAY_UPDATED_STATUS		(1UL<<0)
-#define   PIPE_BPC_MASK				(7 << 5) /* Ironlake */
-#define   PIPE_8BPC				(0 << 5)
-#define   PIPE_10BPC				(1 << 5)
-#define   PIPE_6BPC				(2 << 5)
-#define   PIPE_12BPC				(3 << 5)
 
 #define PIPESRC(pipe) _PIPE(pipe, _PIPEASRC, _PIPEBSRC)
 #define PIPECONF(tran) _TRANSCODER(tran, _PIPEACONF, _PIPEBCONF)
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index f02ec7b..1bb6816 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -1669,8 +1669,8 @@ static void ironlake_enable_pch_transcoder(struct drm_i915_private *dev_priv,
 		 * make the BPC in transcoder be consistent with
 		 * that in pipeconf reg.
 		 */
-		val &= ~PIPE_BPC_MASK;
-		val |= pipeconf_val & PIPE_BPC_MASK;
+		val &= ~PIPECONF_BPC_MASK;
+		val |= pipeconf_val & PIPECONF_BPC_MASK;
 	}
 
 	val &= ~TRANS_INTERLACE_MASK;
@@ -2764,7 +2764,7 @@ static void ironlake_fdi_pll_enable(struct intel_crtc *intel_crtc)
 	temp = I915_READ(reg);
 	temp &= ~((0x7 << 19) | (0x7 << 16));
 	temp |= (intel_crtc->fdi_lanes - 1) << 19;
-	temp |= (I915_READ(PIPECONF(pipe)) & PIPE_BPC_MASK) << 11;
+	temp |= (I915_READ(PIPECONF(pipe)) & PIPECONF_BPC_MASK) << 11;
 	I915_WRITE(reg, temp | FDI_RX_PLL_ENABLE);
 
 	POSTING_READ(reg);
@@ -2845,7 +2845,7 @@ static void ironlake_fdi_disable(struct drm_crtc *crtc)
 	reg = FDI_RX_CTL(pipe);
 	temp = I915_READ(reg);
 	temp &= ~(0x7 << 16);
-	temp |= (I915_READ(PIPECONF(pipe)) & PIPE_BPC_MASK) << 11;
+	temp |= (I915_READ(PIPECONF(pipe)) & PIPECONF_BPC_MASK) << 11;
 	I915_WRITE(reg, temp & ~FDI_RX_ENABLE);
 
 	POSTING_READ(reg);
@@ -2876,7 +2876,7 @@ static void ironlake_fdi_disable(struct drm_crtc *crtc)
 	}
 	/* BPC in FDI rx is consistent with that in PIPECONF */
 	temp &= ~(0x07 << 16);
-	temp |= (I915_READ(PIPECONF(pipe)) & PIPE_BPC_MASK) << 11;
+	temp |= (I915_READ(PIPECONF(pipe)) & PIPECONF_BPC_MASK) << 11;
 	I915_WRITE(reg, temp);
 
 	POSTING_READ(reg);
@@ -3115,7 +3115,7 @@ static void ironlake_pch_enable(struct drm_crtc *crtc)
 	if (HAS_PCH_CPT(dev) &&
 	    (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT) ||
 	     intel_pipe_has_type(crtc, INTEL_OUTPUT_EDP))) {
-		u32 bpc = (I915_READ(PIPECONF(pipe)) & PIPE_BPC_MASK) >> 5;
+		u32 bpc = (I915_READ(PIPECONF(pipe)) & PIPECONF_BPC_MASK) >> 5;
 		reg = TRANS_DP_CTL(pipe);
 		temp = I915_READ(reg);
 		temp &= ~(TRANS_DP_PORT_SEL_MASK |
@@ -4686,10 +4686,10 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 	}
 
 	/* default to 8bpc */
-	pipeconf &= ~(PIPECONF_BPP_MASK | PIPECONF_DITHER_EN);
+	pipeconf &= ~(PIPECONF_BPC_MASK | PIPECONF_DITHER_EN);
 	if (is_dp) {
 		if (adjusted_mode->private_flags & INTEL_MODE_DP_FORCE_6BPC) {
-			pipeconf |= PIPECONF_BPP_6 |
+			pipeconf |= PIPECONF_6BPC |
 				    PIPECONF_DITHER_EN |
 				    PIPECONF_DITHER_TYPE_SP;
 		}
@@ -4697,7 +4697,7 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 
 	if (IS_VALLEYVIEW(dev) && intel_pipe_has_type(crtc, INTEL_OUTPUT_EDP)) {
 		if (adjusted_mode->private_flags & INTEL_MODE_DP_FORCE_6BPC) {
-			pipeconf |= PIPECONF_BPP_6 |
+			pipeconf |= PIPECONF_6BPC |
 					PIPECONF_ENABLE |
 					I965_PIPECONF_ACTIVE;
 		}
@@ -4907,19 +4907,19 @@ static void ironlake_set_pipeconf(struct drm_crtc *crtc,
 
 	val = I915_READ(PIPECONF(pipe));
 
-	val &= ~PIPE_BPC_MASK;
+	val &= ~PIPECONF_BPC_MASK;
 	switch (intel_crtc->bpp) {
 	case 18:
-		val |= PIPE_6BPC;
+		val |= PIPECONF_6BPC;
 		break;
 	case 24:
-		val |= PIPE_8BPC;
+		val |= PIPECONF_8BPC;
 		break;
 	case 30:
-		val |= PIPE_10BPC;
+		val |= PIPECONF_10BPC;
 		break;
 	case 36:
-		val |= PIPE_12BPC;
+		val |= PIPECONF_12BPC;
 		break;
 	default:
 		/* Case prevented by intel_choose_pipe_bpp_dither. */
-- 
cgit v0.10.2


From 4d7bb01162b768ca0f7e50aa3f31790d8d08402c Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Tue, 18 Dec 2012 15:24:37 +0100
Subject: drm/i915: fixup overlay stolen memory leak

We need to clean up the overlay first, before taking down the
stolen memory allocator.

This regression has been introducec in

commit 8040513870399f1cb032cb8bc805df5042fedcdf
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Thu Nov 15 11:32:29 2012 +0000

    drm/i915: Allocate overlay registers from stolen memory

v2: Rework the patch a bit as suggested by Chris Wilson:
- move the overlay teardown up, into the modeset cleanup
- move the stolen mm takedown into i915_gem_cleanup_stolen

Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index a99c46d..272e500 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -1737,9 +1737,6 @@ int i915_driver_unload(struct drm_device *dev)
 		mutex_unlock(&dev->struct_mutex);
 		i915_gem_cleanup_aliasing_ppgtt(dev);
 		i915_gem_cleanup_stolen(dev);
-		drm_mm_takedown(&dev_priv->mm.stolen);
-
-		intel_cleanup_overlay(dev);
 
 		if (!I915_NEED_GFX_HWS(dev))
 			i915_free_hws(dev);
diff --git a/drivers/gpu/drm/i915/i915_gem_stolen.c b/drivers/gpu/drm/i915/i915_gem_stolen.c
index f817b0c..f21ae17 100644
--- a/drivers/gpu/drm/i915/i915_gem_stolen.c
+++ b/drivers/gpu/drm/i915/i915_gem_stolen.c
@@ -173,7 +173,10 @@ void i915_gem_stolen_cleanup_compression(struct drm_device *dev)
 
 void i915_gem_cleanup_stolen(struct drm_device *dev)
 {
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
 	i915_gem_stolen_cleanup_compression(dev);
+	drm_mm_takedown(&dev_priv->mm.stolen);
 }
 
 int i915_gem_init_stolen(struct drm_device *dev)
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 1bb6816..f7f4ef1 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -8869,6 +8869,8 @@ void intel_modeset_cleanup(struct drm_device *dev)
 	flush_scheduled_work();
 
 	drm_mode_config_cleanup(dev);
+
+	intel_cleanup_overlay(dev);
 }
 
 /*
-- 
cgit v0.10.2


From 7dbf9d6e0fd8d42398955de119ccba6c35813c88 Mon Sep 17 00:00:00 2001
From: Ben Widawsky <benjamin.widawsky@intel.com>
Date: Tue, 18 Dec 2012 10:31:22 -0800
Subject: drm/i915: BUG() if fences are used on unsupported platform

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index d15c862..ad03dea 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -2649,7 +2649,7 @@ static void i915_gem_write_fence(struct drm_device *dev, int reg,
 	case 4: i965_write_fence_reg(dev, reg, obj); break;
 	case 3: i915_write_fence_reg(dev, reg, obj); break;
 	case 2: i830_write_fence_reg(dev, reg, obj); break;
-	default: break;
+	default: BUG();
 	}
 }
 
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 551f370..6ba0573 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -1106,6 +1106,8 @@ static void i915_gem_record_fences(struct drm_device *dev,
 			error->fence[i] = I915_READ(FENCE_REG_830_0 + (i * 4));
 		break;
 
+	default:
+		BUG();
 	}
 }
 
-- 
cgit v0.10.2


From 8782e26c0cf3444a0d11400277bb44de710a0142 Mon Sep 17 00:00:00 2001
From: Ben Widawsky <benjamin.widawsky@intel.com>
Date: Tue, 18 Dec 2012 10:31:23 -0800
Subject: drm/i915: Bug on unsupported swizzled platforms

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index ad03dea..6685861 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -3897,8 +3897,10 @@ void i915_gem_init_swizzling(struct drm_device *dev)
 	I915_WRITE(TILECTL, I915_READ(TILECTL) | TILECTL_SWZCTL);
 	if (IS_GEN6(dev))
 		I915_WRITE(ARB_MODE, _MASKED_BIT_ENABLE(ARB_MODE_SWIZZLE_SNB));
-	else
+	else if (IS_GEN7(dev))
 		I915_WRITE(ARB_MODE, _MASKED_BIT_ENABLE(ARB_MODE_SWIZZLE_IVB));
+	else
+		BUG();
 }
 
 static bool
-- 
cgit v0.10.2


From 079a43f67f41274b137d8be1fb6093c8f6b6a762 Mon Sep 17 00:00:00 2001
From: Ben Widawsky <benjamin.widawsky@intel.com>
Date: Tue, 18 Dec 2012 10:31:24 -0800
Subject: drm/i915: Missed conversion to gtt_pte_t

commit f61c0609073133375ace61f74b0e4e7cf631406b
Author: Ben Widawsky <ben@bwidawsk.net>
Date:   Mon Oct 22 11:44:43 2012 -0700

    drm/i915: introduce gtt_pte_t

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index cc1be53..ea646b4 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -282,7 +282,7 @@ void i915_gem_init_ppgtt(struct drm_device *dev)
 	uint32_t pd_offset;
 	struct intel_ring_buffer *ring;
 	struct i915_hw_ppgtt *ppgtt = dev_priv->mm.aliasing_ppgtt;
-	uint32_t __iomem *pd_addr;
+	gtt_pte_t __iomem *pd_addr;
 	uint32_t pd_entry;
 	int i;
 
@@ -290,7 +290,7 @@ void i915_gem_init_ppgtt(struct drm_device *dev)
 		return;
 
 
-	pd_addr = dev_priv->mm.gtt->gtt + ppgtt->pd_offset/sizeof(uint32_t);
+	pd_addr = dev_priv->mm.gtt->gtt + ppgtt->pd_offset/sizeof(gtt_pte_t);
 	for (i = 0; i < ppgtt->num_pd_entries; i++) {
 		dma_addr_t pt_addr;
 
-- 
cgit v0.10.2


From b70ec5bf439b35a18e702f88078d393261c3e3f2 Mon Sep 17 00:00:00 2001
From: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Date: Wed, 19 Dec 2012 11:13:05 +0200
Subject: drm/i915: Introduce ring set_seqno

In preparation for setting per ring initial seqno values
add ring::set_seqno().

Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index 69bbe7b..f536a99 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -727,6 +727,12 @@ ring_get_seqno(struct intel_ring_buffer *ring, bool lazy_coherency)
 	return intel_read_status_page(ring, I915_GEM_HWS_INDEX);
 }
 
+static void
+ring_set_seqno(struct intel_ring_buffer *ring, u32 seqno)
+{
+	intel_write_status_page(ring, I915_GEM_HWS_INDEX, seqno);
+}
+
 static u32
 pc_render_get_seqno(struct intel_ring_buffer *ring, bool lazy_coherency)
 {
@@ -734,6 +740,13 @@ pc_render_get_seqno(struct intel_ring_buffer *ring, bool lazy_coherency)
 	return pc->cpu_page[0];
 }
 
+static void
+pc_render_set_seqno(struct intel_ring_buffer *ring, u32 seqno)
+{
+	struct pipe_control *pc = ring->private;
+	pc->cpu_page[0] = seqno;
+}
+
 static bool
 gen5_ring_get_irq(struct intel_ring_buffer *ring)
 {
@@ -1602,6 +1615,7 @@ int intel_init_render_ring_buffer(struct drm_device *dev)
 		ring->irq_put = gen6_ring_put_irq;
 		ring->irq_enable_mask = GT_USER_INTERRUPT;
 		ring->get_seqno = gen6_ring_get_seqno;
+		ring->set_seqno = ring_set_seqno;
 		ring->sync_to = gen6_ring_sync;
 		ring->semaphore_register[0] = MI_SEMAPHORE_SYNC_INVALID;
 		ring->semaphore_register[1] = MI_SEMAPHORE_SYNC_RV;
@@ -1612,6 +1626,7 @@ int intel_init_render_ring_buffer(struct drm_device *dev)
 		ring->add_request = pc_render_add_request;
 		ring->flush = gen4_render_ring_flush;
 		ring->get_seqno = pc_render_get_seqno;
+		ring->set_seqno = pc_render_set_seqno;
 		ring->irq_get = gen5_ring_get_irq;
 		ring->irq_put = gen5_ring_put_irq;
 		ring->irq_enable_mask = GT_USER_INTERRUPT | GT_PIPE_NOTIFY;
@@ -1622,6 +1637,7 @@ int intel_init_render_ring_buffer(struct drm_device *dev)
 		else
 			ring->flush = gen4_render_ring_flush;
 		ring->get_seqno = ring_get_seqno;
+		ring->set_seqno = ring_set_seqno;
 		if (IS_GEN2(dev)) {
 			ring->irq_get = i8xx_ring_get_irq;
 			ring->irq_put = i8xx_ring_put_irq;
@@ -1672,6 +1688,7 @@ int intel_render_ring_init_dri(struct drm_device *dev, u64 start, u32 size)
 	else
 		ring->flush = gen4_render_ring_flush;
 	ring->get_seqno = ring_get_seqno;
+	ring->set_seqno = ring_set_seqno;
 	if (IS_GEN2(dev)) {
 		ring->irq_get = i8xx_ring_get_irq;
 		ring->irq_put = i8xx_ring_put_irq;
@@ -1732,6 +1749,7 @@ int intel_init_bsd_ring_buffer(struct drm_device *dev)
 		ring->flush = gen6_ring_flush;
 		ring->add_request = gen6_add_request;
 		ring->get_seqno = gen6_ring_get_seqno;
+		ring->set_seqno = ring_set_seqno;
 		ring->irq_enable_mask = GEN6_BSD_USER_INTERRUPT;
 		ring->irq_get = gen6_ring_get_irq;
 		ring->irq_put = gen6_ring_put_irq;
@@ -1747,6 +1765,7 @@ int intel_init_bsd_ring_buffer(struct drm_device *dev)
 		ring->flush = bsd_ring_flush;
 		ring->add_request = i9xx_add_request;
 		ring->get_seqno = ring_get_seqno;
+		ring->set_seqno = ring_set_seqno;
 		if (IS_GEN5(dev)) {
 			ring->irq_enable_mask = GT_BSD_USER_INTERRUPT;
 			ring->irq_get = gen5_ring_get_irq;
@@ -1776,6 +1795,7 @@ int intel_init_blt_ring_buffer(struct drm_device *dev)
 	ring->flush = blt_ring_flush;
 	ring->add_request = gen6_add_request;
 	ring->get_seqno = gen6_ring_get_seqno;
+	ring->set_seqno = ring_set_seqno;
 	ring->irq_enable_mask = GEN6_BLITTER_USER_INTERRUPT;
 	ring->irq_get = gen6_ring_get_irq;
 	ring->irq_put = gen6_ring_put_irq;
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h
index b4a533e..4a7cd67 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.h
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.h
@@ -79,6 +79,8 @@ struct  intel_ring_buffer {
 	 */
 	u32		(*get_seqno)(struct intel_ring_buffer *ring,
 				     bool lazy_coherency);
+	void		(*set_seqno)(struct intel_ring_buffer *ring,
+				     u32 seqno);
 	int		(*dispatch_execbuffer)(struct intel_ring_buffer *ring,
 					       u32 offset, u32 length,
 					       unsigned flags);
@@ -166,6 +168,13 @@ intel_read_status_page(struct intel_ring_buffer *ring,
 	return ring->status_page.page_addr[reg];
 }
 
+static inline void
+intel_write_status_page(struct intel_ring_buffer *ring,
+			int reg, u32 value)
+{
+	ring->status_page.page_addr[reg] = value;
+}
+
 /**
  * Reads a dword out of the status page, which is written to from the command
  * queue by automatic updates, MI_REPORT_HEAD, MI_STORE_DATA_INDEX, or
-- 
cgit v0.10.2


From f7e98ad4d4a8afa043126a6f24d0a154a684e081 Mon Sep 17 00:00:00 2001
From: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Date: Wed, 19 Dec 2012 11:13:06 +0200
Subject: drm/i915: Initialize hardware semaphore state on ring init

Hardware status page needs to have proper seqno set
as our initial seqno can be arbitrary. If initial seqno is close
to wrap boundary on init and i915_seqno_passed() (31bit space)
refers to hw status page which contains zero, errorneous result
will be returned.

v2: clear mboxes and set hws page directly instead of going
through rings. Suggested by Chris Wilson.

v3: hws needs to be updated for all gens. Noticed by Chris
Wilson.

References: https://bugs.freedesktop.org/show_bug.cgi?id=58230
Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 6685861..51282b2 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1954,9 +1954,7 @@ i915_gem_handle_seqno_wrap(struct drm_device *dev)
 
 	/* Finally reset hw state */
 	for_each_ring(ring, dev_priv, i) {
-		ret = intel_ring_handle_seqno_wrap(ring);
-		if (ret)
-			return ret;
+		intel_ring_init_seqno(ring, 0);
 
 		for (j = 0; j < ARRAY_SIZE(ring->sync_seqno); j++)
 			ring->sync_seqno[j] = 0;
@@ -3935,6 +3933,8 @@ i915_gem_init_hw(struct drm_device *dev)
 
 	i915_gem_init_swizzling(dev);
 
+	dev_priv->next_seqno = dev_priv->last_seqno = (u32)~0 - 0x1000;
+
 	ret = intel_init_render_ring_buffer(dev);
 	if (ret)
 		return ret;
@@ -3951,8 +3951,6 @@ i915_gem_init_hw(struct drm_device *dev)
 			goto cleanup_bsd_ring;
 	}
 
-	dev_priv->next_seqno = (u32)-1 - 0x1000;
-
 	/*
 	 * XXX: There was some w/a described somewhere suggesting loading
 	 * contexts before PPGTT.
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index f536a99..2bd074a 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -1184,6 +1184,8 @@ static int intel_init_ring_buffer(struct drm_device *dev,
 	if (IS_I830(ring->dev) || IS_845G(ring->dev))
 		ring->effective_size -= 128;
 
+	intel_ring_init_seqno(ring, dev_priv->last_seqno);
+
 	return 0;
 
 err_unmap:
@@ -1431,26 +1433,18 @@ int intel_ring_begin(struct intel_ring_buffer *ring,
 	return __intel_ring_begin(ring, num_dwords * sizeof(uint32_t));
 }
 
-int intel_ring_handle_seqno_wrap(struct intel_ring_buffer *ring)
+void intel_ring_init_seqno(struct intel_ring_buffer *ring, u32 seqno)
 {
-	int ret;
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
 
 	BUG_ON(ring->outstanding_lazy_request);
 
-	if (INTEL_INFO(ring->dev)->gen < 6)
-		return 0;
-
-	ret = __intel_ring_begin(ring, 6 * sizeof(uint32_t));
-	if (ret)
-		return ret;
-
-	/* Leaving a stale, pre-wrap seqno behind in the mboxes will result in
-	 * post-wrap semaphore waits completing immediately. Clear them. */
-	update_mboxes(ring, ring->signal_mbox[0]);
-	update_mboxes(ring, ring->signal_mbox[1]);
-	intel_ring_advance(ring);
+	if (INTEL_INFO(ring->dev)->gen >= 6) {
+		I915_WRITE(RING_SYNC_0(ring->mmio_base), 0);
+		I915_WRITE(RING_SYNC_1(ring->mmio_base), 0);
+	}
 
-	return 0;
+	ring->set_seqno(ring, seqno);
 }
 
 void intel_ring_advance(struct intel_ring_buffer *ring)
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h
index 4a7cd67..e7b9a6a 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.h
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.h
@@ -205,7 +205,7 @@ static inline void intel_ring_emit(struct intel_ring_buffer *ring,
 }
 void intel_ring_advance(struct intel_ring_buffer *ring);
 int __must_check intel_ring_idle(struct intel_ring_buffer *ring);
-int __must_check intel_ring_handle_seqno_wrap(struct intel_ring_buffer *ring);
+void intel_ring_init_seqno(struct intel_ring_buffer *ring, u32 seqno);
 int intel_ring_flush_all_caches(struct intel_ring_buffer *ring);
 int intel_ring_invalidate_all_caches(struct intel_ring_buffer *ring);
 
-- 
cgit v0.10.2


From ba1a7067c00f4cbdd99b00a570ff072639ae59f2 Mon Sep 17 00:00:00 2001
From: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Date: Wed, 19 Dec 2012 11:13:07 +0200
Subject: drm/i915: Always clear semaphore mboxes on seqno wrap

In preparation for setting the seqno to arbitrary value on init or
through debugfs. We need to always clear the semaphores and set the
hws page seqno index by calling intel_ring_init_seqno().

v2: rewrote the commit message as suggested by Chris Wilson.

Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 51282b2..bb38bca 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1932,18 +1932,6 @@ i915_gem_handle_seqno_wrap(struct drm_device *dev)
 	struct intel_ring_buffer *ring;
 	int ret, i, j;
 
-	/* The hardware uses various monotonic 32-bit counters, if we
-	 * detect that they will wraparound we need to idle the GPU
-	 * and reset those counters.
-	 */
-	ret = 0;
-	for_each_ring(ring, dev_priv, i) {
-		for (j = 0; j < ARRAY_SIZE(ring->sync_seqno); j++)
-			ret |= ring->sync_seqno[j] != 0;
-	}
-	if (ret == 0)
-		return ret;
-
 	/* Carefully retire all requests without writing to the rings */
 	for_each_ring(ring, dev_priv, i) {
 		ret = intel_ring_idle(ring);
-- 
cgit v0.10.2


From fca26bb45375266d9d9b8b4b57fee905ac38fe3c Mon Sep 17 00:00:00 2001
From: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Date: Wed, 19 Dec 2012 11:13:08 +0200
Subject: drm/i915: Introduce i915_gem_set_seqno()

This function can be used to set the driver's next_seqno
to arbitrary value.

i915_gem_set_seqno() will idle the gpu, retire outstanding
requests, clear the semaphore mailboxes and set the hardware
status page's seqno index.

Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 514aee8..b0faa91 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -1479,8 +1479,8 @@ i915_seqno_passed(uint32_t seq1, uint32_t seq2)
 	return (int32_t)(seq1 - seq2) >= 0;
 }
 
-extern int i915_gem_get_seqno(struct drm_device *dev, u32 *seqno);
-
+int __must_check i915_gem_get_seqno(struct drm_device *dev, u32 *seqno);
+int __must_check i915_gem_set_seqno(struct drm_device *dev, u32 seqno);
 int __must_check i915_gem_object_get_fence(struct drm_i915_gem_object *obj);
 int __must_check i915_gem_object_put_fence(struct drm_i915_gem_object *obj);
 
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index bb38bca..23b883a 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1926,7 +1926,7 @@ i915_gem_object_move_to_inactive(struct drm_i915_gem_object *obj)
 }
 
 static int
-i915_gem_handle_seqno_wrap(struct drm_device *dev)
+i915_gem_init_seqno(struct drm_device *dev, u32 seqno)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_ring_buffer *ring;
@@ -1942,7 +1942,7 @@ i915_gem_handle_seqno_wrap(struct drm_device *dev)
 
 	/* Finally reset hw state */
 	for_each_ring(ring, dev_priv, i) {
-		intel_ring_init_seqno(ring, 0);
+		intel_ring_init_seqno(ring, seqno);
 
 		for (j = 0; j < ARRAY_SIZE(ring->sync_seqno); j++)
 			ring->sync_seqno[j] = 0;
@@ -1951,6 +1951,32 @@ i915_gem_handle_seqno_wrap(struct drm_device *dev)
 	return 0;
 }
 
+int i915_gem_set_seqno(struct drm_device *dev, u32 seqno)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int ret;
+
+	if (seqno == 0)
+		return -EINVAL;
+
+	/* HWS page needs to be set less than what we
+	 * will inject to ring
+	 */
+	ret = i915_gem_init_seqno(dev, seqno - 1);
+	if (ret)
+		return ret;
+
+	/* Carefully set the last_seqno value so that wrap
+	 * detection still works
+	 */
+	dev_priv->next_seqno = seqno;
+	dev_priv->last_seqno = seqno - 1;
+	if (dev_priv->last_seqno == 0)
+		dev_priv->last_seqno--;
+
+	return 0;
+}
+
 int
 i915_gem_get_seqno(struct drm_device *dev, u32 *seqno)
 {
@@ -1958,7 +1984,7 @@ i915_gem_get_seqno(struct drm_device *dev, u32 *seqno)
 
 	/* reserve 0 for non-seqno */
 	if (dev_priv->next_seqno == 0) {
-		int ret = i915_gem_handle_seqno_wrap(dev);
+		int ret = i915_gem_init_seqno(dev, 0);
 		if (ret)
 			return ret;
 
-- 
cgit v0.10.2


From e94fbaa8750a8f20c14718633764fba2e6755825 Mon Sep 17 00:00:00 2001
From: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Date: Wed, 19 Dec 2012 11:13:09 +0200
Subject: drm/i915: Make next_seqno debugs entry to use i915_gem_set_seqno

This debugs entry can be used to set arbitrary value to next_seqno.
Use i915_gem_set_seqno instead of poking next_seqno.

v2: nasty details of next_seqno and last_seqno handling
moved inside i915_gem_set_seqno as suggested by Chris Wilson.

Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index 7047c4a9f..882a735 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -878,7 +878,6 @@ i915_next_seqno_write(struct file *filp,
 		      loff_t *ppos)
 {
 	struct drm_device *dev = filp->private_data;
-	drm_i915_private_t *dev_priv = dev->dev_private;
 	char buf[20];
 	u32 val = 1;
 	int ret;
@@ -896,19 +895,11 @@ i915_next_seqno_write(struct file *filp,
 			return ret;
 	}
 
-	if (val == 0)
-		return -EINVAL;
-
 	ret = mutex_lock_interruptible(&dev->struct_mutex);
 	if (ret)
 		return ret;
 
-	if (i915_seqno_passed(val, dev_priv->next_seqno)) {
-		dev_priv->next_seqno = val;
-		DRM_DEBUG_DRIVER("Advancing seqno to %u\n", val);
-	} else {
-		ret = -EINVAL;
-	}
+	ret = i915_gem_set_seqno(dev, val);
 
 	mutex_unlock(&dev->struct_mutex);
 
-- 
cgit v0.10.2


From d7e5008f7c2077d856e40a3af746f1a47028b5f2 Mon Sep 17 00:00:00 2001
From: Ben Widawsky <benjamin.widawsky@intel.com>
Date: Tue, 18 Dec 2012 10:31:25 -0800
Subject: drm/i915: Move even more gtt code to i915_gem_gtt

This really should have been part of the kill agp series.

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index b0faa91..15799e7 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -1587,10 +1587,9 @@ void i915_gem_gtt_bind_object(struct drm_i915_gem_object *obj,
 				enum i915_cache_level cache_level);
 void i915_gem_gtt_unbind_object(struct drm_i915_gem_object *obj);
 void i915_gem_gtt_finish_object(struct drm_i915_gem_object *obj);
-void i915_gem_init_global_gtt(struct drm_device *dev,
-			      unsigned long start,
-			      unsigned long mappable_end,
-			      unsigned long end);
+void i915_gem_init_global_gtt(struct drm_device *dev);
+void i915_gem_setup_global_gtt(struct drm_device *dev, unsigned long start,
+			       unsigned long mappable_end, unsigned long end);
 int i915_gem_gtt_init(struct drm_device *dev);
 void i915_gem_gtt_fini(struct drm_device *dev);
 static inline void i915_gem_chipset_flush(struct drm_device *dev)
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 23b883a..ad98db5 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -163,8 +163,8 @@ i915_gem_init_ioctl(struct drm_device *dev, void *data,
 		return -ENODEV;
 
 	mutex_lock(&dev->struct_mutex);
-	i915_gem_init_global_gtt(dev, args->gtt_start,
-				 args->gtt_end, args->gtt_end);
+	i915_gem_setup_global_gtt(dev, args->gtt_start, args->gtt_end,
+				  args->gtt_end);
 	mutex_unlock(&dev->struct_mutex);
 
 	return 0;
@@ -3981,58 +3981,13 @@ cleanup_render_ring:
 	return ret;
 }
 
-static bool
-intel_enable_ppgtt(struct drm_device *dev)
-{
-	if (i915_enable_ppgtt >= 0)
-		return i915_enable_ppgtt;
-
-#ifdef CONFIG_INTEL_IOMMU
-	/* Disable ppgtt on SNB if VT-d is on. */
-	if (INTEL_INFO(dev)->gen == 6 && intel_iommu_gfx_mapped)
-		return false;
-#endif
-
-	return true;
-}
-
 int i915_gem_init(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	unsigned long gtt_size, mappable_size;
 	int ret;
 
-	gtt_size = dev_priv->mm.gtt->gtt_total_entries << PAGE_SHIFT;
-	mappable_size = dev_priv->mm.gtt->gtt_mappable_entries << PAGE_SHIFT;
-
 	mutex_lock(&dev->struct_mutex);
-	if (intel_enable_ppgtt(dev) && HAS_ALIASING_PPGTT(dev)) {
-		/* PPGTT pdes are stolen from global gtt ptes, so shrink the
-		 * aperture accordingly when using aliasing ppgtt. */
-		gtt_size -= I915_PPGTT_PD_ENTRIES*PAGE_SIZE;
-
-		i915_gem_init_global_gtt(dev, 0, mappable_size, gtt_size);
-
-		ret = i915_gem_init_aliasing_ppgtt(dev);
-		if (ret) {
-			mutex_unlock(&dev->struct_mutex);
-			return ret;
-		}
-	} else {
-		/* Let GEM Manage all of the aperture.
-		 *
-		 * However, leave one page at the end still bound to the scratch
-		 * page.  There are a number of places where the hardware
-		 * apparently prefetches past the end of the object, and we've
-		 * seen multiple hangs with the GPU head pointer stuck in a
-		 * batchbuffer bound at the last page of the aperture.  One page
-		 * should be enough to keep any prefetching inside of the
-		 * aperture.
-		 */
-		i915_gem_init_global_gtt(dev, 0, mappable_size,
-					 gtt_size);
-	}
-
+	i915_gem_init_global_gtt(dev);
 	ret = i915_gem_init_hw(dev);
 	mutex_unlock(&dev->struct_mutex);
 	if (ret) {
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index ea646b4..61b3e72 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -525,10 +525,10 @@ static void i915_gtt_color_adjust(struct drm_mm_node *node,
 	}
 }
 
-void i915_gem_init_global_gtt(struct drm_device *dev,
-			      unsigned long start,
-			      unsigned long mappable_end,
-			      unsigned long end)
+void i915_gem_setup_global_gtt(struct drm_device *dev,
+			       unsigned long start,
+			       unsigned long mappable_end,
+			       unsigned long end)
 {
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	struct drm_mm_node *entry;
@@ -573,6 +573,57 @@ void i915_gem_init_global_gtt(struct drm_device *dev,
 	i915_ggtt_clear_range(dev, end / PAGE_SIZE - 1, 1);
 }
 
+static bool
+intel_enable_ppgtt(struct drm_device *dev)
+{
+	if (i915_enable_ppgtt >= 0)
+		return i915_enable_ppgtt;
+
+#ifdef CONFIG_INTEL_IOMMU
+	/* Disable ppgtt on SNB if VT-d is on. */
+	if (INTEL_INFO(dev)->gen == 6 && intel_iommu_gfx_mapped)
+		return false;
+#endif
+
+	return true;
+}
+
+void i915_gem_init_global_gtt(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	unsigned long gtt_size, mappable_size;
+	int ret;
+
+	gtt_size = dev_priv->mm.gtt->gtt_total_entries << PAGE_SHIFT;
+	mappable_size = dev_priv->mm.gtt->gtt_mappable_entries << PAGE_SHIFT;
+
+	if (intel_enable_ppgtt(dev) && HAS_ALIASING_PPGTT(dev)) {
+		/* PPGTT pdes are stolen from global gtt ptes, so shrink the
+		 * aperture accordingly when using aliasing ppgtt. */
+		gtt_size -= I915_PPGTT_PD_ENTRIES*PAGE_SIZE;
+
+		i915_gem_setup_global_gtt(dev, 0, mappable_size, gtt_size);
+
+		ret = i915_gem_init_aliasing_ppgtt(dev);
+		if (ret) {
+			mutex_unlock(&dev->struct_mutex);
+			return;
+		}
+	} else {
+		/* Let GEM Manage all of the aperture.
+		 *
+		 * However, leave one page at the end still bound to the scratch
+		 * page.  There are a number of places where the hardware
+		 * apparently prefetches past the end of the object, and we've
+		 * seen multiple hangs with the GPU head pointer stuck in a
+		 * batchbuffer bound at the last page of the aperture.  One page
+		 * should be enough to keep any prefetching inside of the
+		 * aperture.
+		 */
+		i915_gem_setup_global_gtt(dev, 0, mappable_size, gtt_size);
+	}
+}
+
 static int setup_scratch_page(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-- 
cgit v0.10.2


From 06e5598fce5ce89fe8bf081398296e5b08d993dd Mon Sep 17 00:00:00 2001
From: Ben Widawsky <benjamin.widawsky@intel.com>
Date: Tue, 18 Dec 2012 10:31:26 -0800
Subject: drm/i915: Move GSM mapping into dev_priv

This removes an unused field from the AGP structure and moves it into
the dev_priv structure (with a slightly better name). This builds upon
the kill-agp series already merged.

GSM is a well defined term in the bspec:
GSM: Graphics Stolen Memory

GTT stolen space is defined for storage of the GFX GTT entries in
physical memory. IA can not access GSM directly , it can only access via
GTTMMADR. GT can access GSM directly or through GTTMMADR.

This is not the entire stolen space.

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/char/agp/intel-gtt.c b/drivers/char/agp/intel-gtt.c
index dbd901e..c8d9dcb 100644
--- a/drivers/char/agp/intel-gtt.c
+++ b/drivers/char/agp/intel-gtt.c
@@ -602,7 +602,6 @@ static int intel_gtt_init(void)
 		iounmap(intel_private.registers);
 		return -ENOMEM;
 	}
-	intel_private.base.gtt = intel_private.gtt;
 
 	global_cache_flush();   /* FIXME: ? */
 
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 15799e7..ae88d87 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -800,6 +800,9 @@ typedef struct drm_i915_private {
 		unsigned long gtt_end;
 		unsigned long stolen_base; /* limited to low memory (32-bit) */
 
+		/** "Graphics Stolen Memory" holds the global PTEs */
+		uint32_t __iomem *gsm;
+
 		struct io_mapping *gtt_mapping;
 		phys_addr_t gtt_base_addr;
 		int gtt_mtrr;
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 61b3e72..912389e 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -290,7 +290,7 @@ void i915_gem_init_ppgtt(struct drm_device *dev)
 		return;
 
 
-	pd_addr = dev_priv->mm.gtt->gtt + ppgtt->pd_offset/sizeof(gtt_pte_t);
+	pd_addr = dev_priv->mm.gsm + ppgtt->pd_offset/sizeof(gtt_pte_t);
 	for (i = 0; i < ppgtt->num_pd_entries; i++) {
 		dma_addr_t pt_addr;
 
@@ -367,7 +367,7 @@ static void i915_ggtt_clear_range(struct drm_device *dev,
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	gtt_pte_t scratch_pte;
-	gtt_pte_t __iomem *gtt_base = dev_priv->mm.gtt->gtt + first_entry;
+	gtt_pte_t __iomem *gtt_base = (gtt_pte_t __iomem *) dev_priv->mm.gsm + first_entry;
 	const int max_entries = dev_priv->mm.gtt->gtt_total_entries - first_entry;
 	int i;
 
@@ -432,7 +432,7 @@ static void gen6_ggtt_bind_object(struct drm_i915_gem_object *obj,
 	struct scatterlist *sg = st->sgl;
 	const int first_entry = obj->gtt_space->start >> PAGE_SHIFT;
 	const int max_entries = dev_priv->mm.gtt->gtt_total_entries - first_entry;
-	gtt_pte_t __iomem *gtt_entries = dev_priv->mm.gtt->gtt + first_entry;
+	gtt_pte_t __iomem *gtt_entries = dev_priv->mm.gsm + first_entry;
 	int unused, i = 0;
 	unsigned int len, m = 0;
 	dma_addr_t addr;
@@ -747,9 +747,9 @@ int i915_gem_gtt_init(struct drm_device *dev)
 		goto err_out;
 	}
 
-	dev_priv->mm.gtt->gtt = ioremap_wc(gtt_bus_addr,
-					   dev_priv->mm.gtt->gtt_total_entries * sizeof(gtt_pte_t));
-	if (!dev_priv->mm.gtt->gtt) {
+	dev_priv->mm.gsm = ioremap_wc(gtt_bus_addr,
+				      dev_priv->mm.gtt->gtt_total_entries * sizeof(gtt_pte_t));
+	if (!dev_priv->mm.gsm) {
 		DRM_ERROR("Failed to map the gtt page table\n");
 		teardown_scratch_page(dev);
 		ret = -ENOMEM;
@@ -773,7 +773,7 @@ err_out:
 void i915_gem_gtt_fini(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	iounmap(dev_priv->mm.gtt->gtt);
+	iounmap(dev_priv->mm.gsm);
 	teardown_scratch_page(dev);
 	if (INTEL_INFO(dev)->gen < 6)
 		intel_gmch_remove();
diff --git a/include/drm/intel-gtt.h b/include/drm/intel-gtt.h
index 6eb76a1..3e3a166 100644
--- a/include/drm/intel-gtt.h
+++ b/include/drm/intel-gtt.h
@@ -18,8 +18,6 @@ struct intel_gtt {
 	/* Share the scratch page dma with ppgtts. */
 	dma_addr_t scratch_page_dma;
 	struct page *scratch_page;
-	/* for ppgtt PDE access */
-	u32 __iomem *gtt;
 	/* needed for ioremap in drm/i915 */
 	phys_addr_t gma_bus_addr;
 } *intel_gtt_get(void);
-- 
cgit v0.10.2


From 1c45140d3da0c98fa7eda2cd10f9805232a5e9ac Mon Sep 17 00:00:00 2001
From: Ben Widawsky <benjamin.widawsky@intel.com>
Date: Tue, 18 Dec 2012 10:31:27 -0800
Subject: drm/i915: Make GSM void

The iomapping of the register region has historically been a uint32_t
for the obvious reason that our PTE size was always 4b. In the future
however, we cannot make this assumption.

By making the type void, it makes the upcoming pointer math we will do
much easier, and hopefully gives the compiler opportunities to warn us
when we do stupid things.

v2: Cast to __iomem, caught by Ville

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
[danvet: Fixup __iomem issue for real.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index ae88d87..d2b93a4 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -801,7 +801,7 @@ typedef struct drm_i915_private {
 		unsigned long stolen_base; /* limited to low memory (32-bit) */
 
 		/** "Graphics Stolen Memory" holds the global PTEs */
-		uint32_t __iomem *gsm;
+		void __iomem *gsm;
 
 		struct io_mapping *gtt_mapping;
 		phys_addr_t gtt_base_addr;
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 912389e..eac2cec 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -290,7 +290,7 @@ void i915_gem_init_ppgtt(struct drm_device *dev)
 		return;
 
 
-	pd_addr = dev_priv->mm.gsm + ppgtt->pd_offset/sizeof(gtt_pte_t);
+	pd_addr = (gtt_pte_t __iomem*)dev_priv->mm.gsm + ppgtt->pd_offset/sizeof(gtt_pte_t);
 	for (i = 0; i < ppgtt->num_pd_entries; i++) {
 		dma_addr_t pt_addr;
 
@@ -432,7 +432,8 @@ static void gen6_ggtt_bind_object(struct drm_i915_gem_object *obj,
 	struct scatterlist *sg = st->sgl;
 	const int first_entry = obj->gtt_space->start >> PAGE_SHIFT;
 	const int max_entries = dev_priv->mm.gtt->gtt_total_entries - first_entry;
-	gtt_pte_t __iomem *gtt_entries = dev_priv->mm.gsm + first_entry;
+	gtt_pte_t __iomem *gtt_entries =
+		(gtt_pte_t __iomem *)dev_priv->mm.gsm + first_entry;
 	int unused, i = 0;
 	unsigned int len, m = 0;
 	dma_addr_t addr;
-- 
cgit v0.10.2


From c0c36b941b6f0be6ac74f340040cbb29d6a0b06c Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Wed, 19 Dec 2012 16:08:43 +0000
Subject: drm/i915: Return the real error code from intel_set_mode()

Note: This patch also adds a little helper intel_crtc_restore_mode for
the common case where we do a full modeset but with the same
parameters, e.g. to undo bios damage or update a property.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@gmail.com>
[danvet: Added note.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index f7f4ef1..1464e47 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -6313,7 +6313,7 @@ bool intel_get_load_detect_pipe(struct drm_connector *connector,
 		return false;
 	}
 
-	if (!intel_set_mode(crtc, mode, 0, 0, fb)) {
+	if (intel_set_mode(crtc, mode, 0, 0, fb)) {
 		DRM_DEBUG_KMS("failed to set mode on load-detect pipe\n");
 		if (old->release_fb)
 			old->release_fb->funcs->destroy(old->release_fb);
@@ -7426,22 +7426,20 @@ intel_modeset_check_state(struct drm_device *dev)
 	}
 }
 
-bool intel_set_mode(struct drm_crtc *crtc,
-		    struct drm_display_mode *mode,
-		    int x, int y, struct drm_framebuffer *fb)
+int intel_set_mode(struct drm_crtc *crtc,
+		   struct drm_display_mode *mode,
+		   int x, int y, struct drm_framebuffer *fb)
 {
 	struct drm_device *dev = crtc->dev;
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	struct drm_display_mode *adjusted_mode, *saved_mode, *saved_hwmode;
 	struct intel_crtc *intel_crtc;
 	unsigned disable_pipes, prepare_pipes, modeset_pipes;
-	bool ret = true;
+	int ret = 0;
 
 	saved_mode = kmalloc(2 * sizeof(*saved_mode), GFP_KERNEL);
-	if (!saved_mode) {
-		DRM_ERROR("i915: Could not allocate saved display mode.\n");
-		return false;
-	}
+	if (!saved_mode)
+		return -ENOMEM;
 	saved_hwmode = saved_mode + 1;
 
 	intel_modeset_affected_pipes(crtc, &modeset_pipes,
@@ -7465,7 +7463,7 @@ bool intel_set_mode(struct drm_crtc *crtc,
 	if (modeset_pipes) {
 		adjusted_mode = intel_modeset_adjusted_mode(crtc, mode);
 		if (IS_ERR(adjusted_mode)) {
-			ret = false;
+			ret = PTR_ERR(adjusted_mode);
 			goto out;
 		}
 	}
@@ -7492,11 +7490,11 @@ bool intel_set_mode(struct drm_crtc *crtc,
 	 * on the DPLL.
 	 */
 	for_each_intel_crtc_masked(dev, modeset_pipes, intel_crtc) {
-		ret = !intel_crtc_mode_set(&intel_crtc->base,
-					   mode, adjusted_mode,
-					   x, y, fb);
-		if (!ret)
-		    goto done;
+		ret = intel_crtc_mode_set(&intel_crtc->base,
+					  mode, adjusted_mode,
+					  x, y, fb);
+		if (ret)
+			goto done;
 	}
 
 	/* Now enable the clocks, plane, pipe, and connectors that we set up. */
@@ -7517,7 +7515,7 @@ bool intel_set_mode(struct drm_crtc *crtc,
 	/* FIXME: add subpixel order */
 done:
 	drm_mode_destroy(dev, adjusted_mode);
-	if (!ret && crtc->enabled) {
+	if (ret && crtc->enabled) {
 		crtc->hwmode = *saved_hwmode;
 		crtc->mode = *saved_mode;
 	} else {
@@ -7529,6 +7527,11 @@ out:
 	return ret;
 }
 
+void intel_crtc_restore_mode(struct drm_crtc *crtc)
+{
+	intel_set_mode(crtc, &crtc->mode, crtc->x, crtc->y, crtc->fb);
+}
+
 #undef for_each_intel_crtc_masked
 
 static void intel_set_config_free(struct intel_set_config *config)
@@ -7798,11 +7801,11 @@ static int intel_crtc_set_config(struct drm_mode_set *set)
 			drm_mode_debug_printmodeline(set->mode);
 		}
 
-		if (!intel_set_mode(set->crtc, set->mode,
-				    set->x, set->y, set->fb)) {
-			DRM_ERROR("failed to set mode on [CRTC:%d]\n",
-				  set->crtc->base.id);
-			ret = -EINVAL;
+		ret = intel_set_mode(set->crtc, set->mode,
+				     set->x, set->y, set->fb);
+		if (ret) {
+			DRM_ERROR("failed to set mode on [CRTC:%d], err = %d\n",
+				  set->crtc->base.id, ret);
 			goto fail;
 		}
 	} else if (config->fb_changed) {
@@ -7819,8 +7822,8 @@ fail:
 
 	/* Try to restore the config */
 	if (config->mode_changed &&
-	    !intel_set_mode(save_set.crtc, save_set.mode,
-			    save_set.x, save_set.y, save_set.fb))
+	    intel_set_mode(save_set.crtc, save_set.mode,
+			   save_set.x, save_set.y, save_set.fb))
 		DRM_ERROR("failed to restore config after modeset failure\n");
 
 out_config:
@@ -8804,11 +8807,8 @@ void intel_modeset_setup_hw_state(struct drm_device *dev,
 	}
 
 	if (force_restore) {
-		for_each_pipe(pipe) {
-			crtc = to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
-			intel_set_mode(&crtc->base, &crtc->base.mode,
-				       crtc->base.x, crtc->base.y, crtc->base.fb);
-		}
+		for_each_pipe(pipe)
+			intel_crtc_restore_mode(dev_priv->pipe_to_crtc_mapping[pipe]);
 	} else {
 		intel_modeset_update_staged_output_state(dev);
 	}
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 2d3b268..1dd89d5 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -2483,11 +2483,8 @@ intel_dp_set_property(struct drm_connector *connector,
 	return -EINVAL;
 
 done:
-	if (intel_encoder->base.crtc) {
-		struct drm_crtc *crtc = intel_encoder->base.crtc;
-		intel_set_mode(crtc, &crtc->mode,
-			       crtc->x, crtc->y, crtc->fb);
-	}
+	if (intel_encoder->base.crtc)
+		intel_crtc_restore_mode(intel_encoder->base.crtc);
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 53d4c8f..116580b 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -501,9 +501,10 @@ struct intel_set_config {
 	bool mode_changed;
 };
 
-extern bool intel_set_mode(struct drm_crtc *crtc, struct drm_display_mode *mode,
-			   int x, int y, struct drm_framebuffer *old_fb);
+extern int intel_set_mode(struct drm_crtc *crtc, struct drm_display_mode *mode,
+			  int x, int y, struct drm_framebuffer *old_fb);
 extern void intel_modeset_disable(struct drm_device *dev);
+extern void intel_crtc_restore_mode(struct drm_crtc *crtc);
 extern void intel_crtc_load_lut(struct drm_crtc *crtc);
 extern void intel_crtc_update_dpms(struct drm_crtc *crtc);
 extern void intel_encoder_noop(struct drm_encoder *encoder);
diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c
index 9f834d3..6387f9b 100644
--- a/drivers/gpu/drm/i915/intel_hdmi.c
+++ b/drivers/gpu/drm/i915/intel_hdmi.c
@@ -917,11 +917,8 @@ intel_hdmi_set_property(struct drm_connector *connector,
 	return -EINVAL;
 
 done:
-	if (intel_dig_port->base.base.crtc) {
-		struct drm_crtc *crtc = intel_dig_port->base.base.crtc;
-		intel_set_mode(crtc, &crtc->mode,
-			       crtc->x, crtc->y, crtc->fb);
-	}
+	if (intel_dig_port->base.base.crtc)
+		intel_crtc_restore_mode(intel_dig_port->base.base.crtc);
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c
index 7781069..8c61876 100644
--- a/drivers/gpu/drm/i915/intel_lvds.c
+++ b/drivers/gpu/drm/i915/intel_lvds.c
@@ -646,8 +646,7 @@ static int intel_lvds_set_property(struct drm_connector *connector,
 			 * If the CRTC is enabled, the display will be changed
 			 * according to the new panel fitting mode.
 			 */
-			intel_set_mode(crtc, &crtc->mode,
-				       crtc->x, crtc->y, crtc->fb);
+			intel_crtc_restore_mode(crtc);
 		}
 	}
 
diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c
index 0e03985..ea2e79f6 100644
--- a/drivers/gpu/drm/i915/intel_sdvo.c
+++ b/drivers/gpu/drm/i915/intel_sdvo.c
@@ -1997,11 +1997,8 @@ set_value:
 
 
 done:
-	if (intel_sdvo->base.base.crtc) {
-		struct drm_crtc *crtc = intel_sdvo->base.base.crtc;
-		intel_set_mode(crtc, &crtc->mode,
-			       crtc->x, crtc->y, crtc->fb);
-	}
+	if (intel_sdvo->base.base.crtc)
+		intel_crtc_restore_mode(intel_sdvo->base.base.crtc);
 
 	return 0;
 #undef CHECK_PROPERTY
diff --git a/drivers/gpu/drm/i915/intel_tv.c b/drivers/gpu/drm/i915/intel_tv.c
index ea93520..984a113 100644
--- a/drivers/gpu/drm/i915/intel_tv.c
+++ b/drivers/gpu/drm/i915/intel_tv.c
@@ -1479,8 +1479,7 @@ intel_tv_set_property(struct drm_connector *connector, struct drm_property *prop
 	}
 
 	if (changed && crtc)
-		intel_set_mode(crtc, &crtc->mode,
-			       crtc->x, crtc->y, crtc->fb);
+		intel_crtc_restore_mode(crtc);
 out:
 	return ret;
 }
-- 
cgit v0.10.2


From 42f6e3da974dc8ad81775110c8d06835acdf375e Mon Sep 17 00:00:00 2001
From: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Date: Tue, 15 Jan 2013 14:53:18 +0100
Subject: drm/vmwgfx: always use ttm_bo_is_reserved

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>

diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
index e01a17b..1655617 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
@@ -959,13 +959,13 @@ void vmw_resource_unreserve(struct vmw_resource *res,
 	if (new_backup && new_backup != res->backup) {
 
 		if (res->backup) {
-			BUG_ON(atomic_read(&res->backup->base.reserved) == 0);
+			BUG_ON(!ttm_bo_is_reserved(&res->backup->base));
 			list_del_init(&res->mob_head);
 			vmw_dmabuf_unreference(&res->backup);
 		}
 
 		res->backup = vmw_dmabuf_reference(new_backup);
-		BUG_ON(atomic_read(&new_backup->base.reserved) == 0);
+		BUG_ON(!ttm_bo_is_reserved(&new_backup->base));
 		list_add_tail(&res->mob_head, &new_backup->res_list);
 	}
 	if (new_backup)
-- 
cgit v0.10.2


From 979ee290ff0a543352243145dc3654af5a856ab8 Mon Sep 17 00:00:00 2001
From: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Date: Tue, 15 Jan 2013 14:54:22 +0100
Subject: drm/nouveau: increase reservation sequence every retry

This is temporary until the fence framework can be used. With the
lru/reservation atomicity removal it is possible to see your old
sequence number and the buffer being reserved, leading to erroneously
reporting -EDEADLK.

Workaround it by bumping the sequence number every retry.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>

diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c b/drivers/gpu/drm/nouveau/nouveau_gem.c
index 8bf695c..9fcfcb2 100644
--- a/drivers/gpu/drm/nouveau/nouveau_gem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_gem.c
@@ -321,8 +321,8 @@ validate_init(struct nouveau_channel *chan, struct drm_file *file_priv,
 	int trycnt = 0;
 	int ret, i;
 
-	sequence = atomic_add_return(1, &drm->ttm.validate_sequence);
 retry:
+	sequence = atomic_add_return(1, &drm->ttm.validate_sequence);
 	if (++trycnt > 100000) {
 		NV_ERROR(drm, "%s failed and gave up.\n", __func__);
 		return -EINVAL;
-- 
cgit v0.10.2


From 63d0a4195560362e2e00a3ad38fc331d34e1da9b Mon Sep 17 00:00:00 2001
From: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Date: Tue, 15 Jan 2013 14:56:37 +0100
Subject: drm/ttm: remove lru_lock around ttm_bo_reserve

There should no longer be assumptions that reserve will always succeed
with the lru lock held, so we can safely break the whole atomic
reserve/lru thing. As a bonus this fixes most lockdep annotations for
reservations.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>

diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index 33d20be..e8e4814 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -213,14 +213,13 @@ int ttm_bo_del_from_lru(struct ttm_buffer_object *bo)
 	return put_count;
 }
 
-int ttm_bo_reserve_locked(struct ttm_buffer_object *bo,
+int ttm_bo_reserve_nolru(struct ttm_buffer_object *bo,
 			  bool interruptible,
 			  bool no_wait, bool use_sequence, uint32_t sequence)
 {
-	struct ttm_bo_global *glob = bo->glob;
 	int ret;
 
-	while (unlikely(atomic_read(&bo->reserved) != 0)) {
+	while (unlikely(atomic_xchg(&bo->reserved, 1) != 0)) {
 		/**
 		 * Deadlock avoidance for multi-bo reserving.
 		 */
@@ -241,26 +240,36 @@ int ttm_bo_reserve_locked(struct ttm_buffer_object *bo,
 		if (no_wait)
 			return -EBUSY;
 
-		spin_unlock(&glob->lru_lock);
 		ret = ttm_bo_wait_unreserved(bo, interruptible);
-		spin_lock(&glob->lru_lock);
 
 		if (unlikely(ret))
 			return ret;
 	}
 
-	atomic_set(&bo->reserved, 1);
 	if (use_sequence) {
+		bool wake_up = false;
 		/**
 		 * Wake up waiters that may need to recheck for deadlock,
 		 * if we decreased the sequence number.
 		 */
 		if (unlikely((bo->val_seq - sequence < (1 << 31))
 			     || !bo->seq_valid))
-			wake_up_all(&bo->event_queue);
+			wake_up = true;
 
+		/*
+		 * In the worst case with memory ordering these values can be
+		 * seen in the wrong order. However since we call wake_up_all
+		 * in that case, this will hopefully not pose a problem,
+		 * and the worst case would only cause someone to accidentally
+		 * hit -EAGAIN in ttm_bo_reserve when they see old value of
+		 * val_seq. However this would only happen if seq_valid was
+		 * written before val_seq was, and just means some slightly
+		 * increased cpu usage
+		 */
 		bo->val_seq = sequence;
 		bo->seq_valid = true;
+		if (wake_up)
+			wake_up_all(&bo->event_queue);
 	} else {
 		bo->seq_valid = false;
 	}
@@ -289,14 +298,14 @@ int ttm_bo_reserve(struct ttm_buffer_object *bo,
 	int put_count = 0;
 	int ret;
 
-	spin_lock(&glob->lru_lock);
-	ret = ttm_bo_reserve_locked(bo, interruptible, no_wait, use_sequence,
-				    sequence);
-	if (likely(ret == 0))
+	ret = ttm_bo_reserve_nolru(bo, interruptible, no_wait, use_sequence,
+				   sequence);
+	if (likely(ret == 0)) {
+		spin_lock(&glob->lru_lock);
 		put_count = ttm_bo_del_from_lru(bo);
-	spin_unlock(&glob->lru_lock);
-
-	ttm_bo_list_ref_sub(bo, put_count, true);
+		spin_unlock(&glob->lru_lock);
+		ttm_bo_list_ref_sub(bo, put_count, true);
+	}
 
 	return ret;
 }
@@ -510,7 +519,7 @@ static void ttm_bo_cleanup_refs_or_queue(struct ttm_buffer_object *bo)
 	int ret;
 
 	spin_lock(&glob->lru_lock);
-	ret = ttm_bo_reserve_locked(bo, false, true, false, 0);
+	ret = ttm_bo_reserve_nolru(bo, false, true, false, 0);
 
 	spin_lock(&bdev->fence_lock);
 	(void) ttm_bo_wait(bo, false, false, true);
@@ -603,7 +612,7 @@ static int ttm_bo_cleanup_refs_and_unlock(struct ttm_buffer_object *bo,
 			return ret;
 
 		spin_lock(&glob->lru_lock);
-		ret = ttm_bo_reserve_locked(bo, false, true, false, 0);
+		ret = ttm_bo_reserve_nolru(bo, false, true, false, 0);
 
 		/*
 		 * We raced, and lost, someone else holds the reservation now,
@@ -667,7 +676,14 @@ static int ttm_bo_delayed_delete(struct ttm_bo_device *bdev, bool remove_all)
 			kref_get(&nentry->list_kref);
 		}
 
-		ret = ttm_bo_reserve_locked(entry, false, !remove_all, false, 0);
+		ret = ttm_bo_reserve_nolru(entry, false, true, false, 0);
+		if (remove_all && ret) {
+			spin_unlock(&glob->lru_lock);
+			ret = ttm_bo_reserve_nolru(entry, false, false,
+						   false, 0);
+			spin_lock(&glob->lru_lock);
+		}
+
 		if (!ret)
 			ret = ttm_bo_cleanup_refs_and_unlock(entry, false,
 							     !remove_all);
@@ -815,7 +831,7 @@ static int ttm_mem_evict_first(struct ttm_bo_device *bdev,
 
 	spin_lock(&glob->lru_lock);
 	list_for_each_entry(bo, &man->lru, lru) {
-		ret = ttm_bo_reserve_locked(bo, false, true, false, 0);
+		ret = ttm_bo_reserve_nolru(bo, false, true, false, 0);
 		if (!ret)
 			break;
 	}
@@ -1796,7 +1812,7 @@ static int ttm_bo_swapout(struct ttm_mem_shrink *shrink)
 
 	spin_lock(&glob->lru_lock);
 	list_for_each_entry(bo, &glob->swap_lru, swap) {
-		ret = ttm_bo_reserve_locked(bo, false, true, false, 0);
+		ret = ttm_bo_reserve_nolru(bo, false, true, false, 0);
 		if (!ret)
 			break;
 	}
diff --git a/drivers/gpu/drm/ttm/ttm_execbuf_util.c b/drivers/gpu/drm/ttm/ttm_execbuf_util.c
index cd9e452..bd37b5c 100644
--- a/drivers/gpu/drm/ttm/ttm_execbuf_util.c
+++ b/drivers/gpu/drm/ttm/ttm_execbuf_util.c
@@ -153,7 +153,7 @@ retry:
 		struct ttm_buffer_object *bo = entry->bo;
 
 retry_this_bo:
-		ret = ttm_bo_reserve_locked(bo, true, true, true, val_seq);
+		ret = ttm_bo_reserve_nolru(bo, true, true, true, val_seq);
 		switch (ret) {
 		case 0:
 			break;
diff --git a/include/drm/ttm/ttm_bo_driver.h b/include/drm/ttm/ttm_bo_driver.h
index e3a43a4..6fff432 100644
--- a/include/drm/ttm/ttm_bo_driver.h
+++ b/include/drm/ttm/ttm_bo_driver.h
@@ -790,16 +790,7 @@ extern void ttm_mem_io_unlock(struct ttm_mem_type_manager *man);
  * to make room for a buffer already reserved. (Buffers are reserved before
  * they are evicted). The following algorithm prevents such deadlocks from
  * occurring:
- * 1) Buffers are reserved with the lru spinlock held. Upon successful
- * reservation they are removed from the lru list. This stops a reserved buffer
- * from being evicted. However the lru spinlock is released between the time
- * a buffer is selected for eviction and the time it is reserved.
- * Therefore a check is made when a buffer is reserved for eviction, that it
- * is still the first buffer in the lru list, before it is removed from the
- * list. @check_lru == 1 forces this check. If it fails, the function returns
- * -EINVAL, and the caller should then choose a new buffer to evict and repeat
- * the procedure.
- * 2) Processes attempting to reserve multiple buffers other than for eviction,
+ * Processes attempting to reserve multiple buffers other than for eviction,
  * (typically execbuf), should first obtain a unique 32-bit
  * validation sequence number,
  * and call this function with @use_sequence == 1 and @sequence == the unique
@@ -832,7 +823,7 @@ extern int ttm_bo_reserve(struct ttm_buffer_object *bo,
 
 
 /**
- * ttm_bo_reserve_locked:
+ * ttm_bo_reserve_nolru:
  *
  * @bo: A pointer to a struct ttm_buffer_object.
  * @interruptible: Sleep interruptible if waiting.
@@ -840,9 +831,7 @@ extern int ttm_bo_reserve(struct ttm_buffer_object *bo,
  * @use_sequence: If @bo is already reserved, Only sleep waiting for
  * it to become unreserved if @sequence < (@bo)->sequence.
  *
- * Must be called with struct ttm_bo_global::lru_lock held,
- * and will not remove reserved buffers from the lru lists.
- * The function may release the LRU spinlock if it needs to sleep.
+ * Will not remove reserved buffers from the lru lists.
  * Otherwise identical to ttm_bo_reserve.
  *
  * Returns:
@@ -855,7 +844,7 @@ extern int ttm_bo_reserve(struct ttm_buffer_object *bo,
  * -EDEADLK: Bo already reserved using @sequence. This error code will only
  * be returned if @use_sequence is set to true.
  */
-extern int ttm_bo_reserve_locked(struct ttm_buffer_object *bo,
+extern int ttm_bo_reserve_nolru(struct ttm_buffer_object *bo,
 				 bool interruptible,
 				 bool no_wait, bool use_sequence,
 				 uint32_t sequence);
-- 
cgit v0.10.2


From 7a1863084c9d90ce4b67d645bf9b0f1612e68f62 Mon Sep 17 00:00:00 2001
From: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Date: Tue, 15 Jan 2013 14:56:48 +0100
Subject: drm/ttm: cleanup ttm_eu_reserve_buffers handling

With the lru lock no longer required for protecting reservations we
can just do a ttm_bo_reserve_nolru on -EBUSY, and handle all errors
in a single path.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>

diff --git a/drivers/gpu/drm/ttm/ttm_execbuf_util.c b/drivers/gpu/drm/ttm/ttm_execbuf_util.c
index bd37b5c..c7d3236 100644
--- a/drivers/gpu/drm/ttm/ttm_execbuf_util.c
+++ b/drivers/gpu/drm/ttm/ttm_execbuf_util.c
@@ -82,22 +82,6 @@ static void ttm_eu_list_ref_sub(struct list_head *list)
 	}
 }
 
-static int ttm_eu_wait_unreserved_locked(struct list_head *list,
-					 struct ttm_buffer_object *bo)
-{
-	struct ttm_bo_global *glob = bo->glob;
-	int ret;
-
-	ttm_eu_del_from_lru_locked(list);
-	spin_unlock(&glob->lru_lock);
-	ret = ttm_bo_wait_unreserved(bo, true);
-	spin_lock(&glob->lru_lock);
-	if (unlikely(ret != 0))
-		ttm_eu_backoff_reservation_locked(list);
-	return ret;
-}
-
-
 void ttm_eu_backoff_reservation(struct list_head *list)
 {
 	struct ttm_validate_buffer *entry;
@@ -152,19 +136,23 @@ retry:
 	list_for_each_entry(entry, list, head) {
 		struct ttm_buffer_object *bo = entry->bo;
 
-retry_this_bo:
 		ret = ttm_bo_reserve_nolru(bo, true, true, true, val_seq);
 		switch (ret) {
 		case 0:
 			break;
 		case -EBUSY:
-			ret = ttm_eu_wait_unreserved_locked(list, bo);
-			if (unlikely(ret != 0)) {
-				spin_unlock(&glob->lru_lock);
-				ttm_eu_list_ref_sub(list);
-				return ret;
-			}
-			goto retry_this_bo;
+			ttm_eu_del_from_lru_locked(list);
+			spin_unlock(&glob->lru_lock);
+			ret = ttm_bo_reserve_nolru(bo, true, false,
+						   true, val_seq);
+			spin_lock(&glob->lru_lock);
+			if (!ret)
+				break;
+
+			if (unlikely(ret != -EAGAIN))
+				goto err;
+
+			/* fallthrough */
 		case -EAGAIN:
 			ttm_eu_backoff_reservation_locked(list);
 			spin_unlock(&glob->lru_lock);
@@ -174,18 +162,13 @@ retry_this_bo:
 				return ret;
 			goto retry;
 		default:
-			ttm_eu_backoff_reservation_locked(list);
-			spin_unlock(&glob->lru_lock);
-			ttm_eu_list_ref_sub(list);
-			return ret;
+			goto err;
 		}
 
 		entry->reserved = true;
 		if (unlikely(atomic_read(&bo->cpu_writers) > 0)) {
-			ttm_eu_backoff_reservation_locked(list);
-			spin_unlock(&glob->lru_lock);
-			ttm_eu_list_ref_sub(list);
-			return -EBUSY;
+			ret = -EBUSY;
+			goto err;
 		}
 	}
 
@@ -194,6 +177,12 @@ retry_this_bo:
 	ttm_eu_list_ref_sub(list);
 
 	return 0;
+
+err:
+	ttm_eu_backoff_reservation_locked(list);
+	spin_unlock(&glob->lru_lock);
+	ttm_eu_list_ref_sub(list);
+	return ret;
 }
 EXPORT_SYMBOL(ttm_eu_reserve_buffers);
 
-- 
cgit v0.10.2


From 5e45d7dfd74100d622f9cdc70bfd1f9fae1671de Mon Sep 17 00:00:00 2001
From: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Date: Tue, 15 Jan 2013 14:57:05 +0100
Subject: drm/ttm: add ttm_bo_reserve_slowpath

Instead of dropping everything, waiting for the bo to be unreserved
and trying over, a better strategy would be to do a blocking wait.

This can be mapped a lot better to a mutex_lock-like call.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>

diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index e8e4814..4dd6f9e 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -310,6 +310,53 @@ int ttm_bo_reserve(struct ttm_buffer_object *bo,
 	return ret;
 }
 
+int ttm_bo_reserve_slowpath_nolru(struct ttm_buffer_object *bo,
+				  bool interruptible, uint32_t sequence)
+{
+	bool wake_up = false;
+	int ret;
+
+	while (unlikely(atomic_xchg(&bo->reserved, 1) != 0)) {
+		WARN_ON(bo->seq_valid && sequence == bo->val_seq);
+
+		ret = ttm_bo_wait_unreserved(bo, interruptible);
+
+		if (unlikely(ret))
+			return ret;
+	}
+
+	if ((bo->val_seq - sequence < (1 << 31)) || !bo->seq_valid)
+		wake_up = true;
+
+	/**
+	 * Wake up waiters that may need to recheck for deadlock,
+	 * if we decreased the sequence number.
+	 */
+	bo->val_seq = sequence;
+	bo->seq_valid = true;
+	if (wake_up)
+		wake_up_all(&bo->event_queue);
+
+	return 0;
+}
+
+int ttm_bo_reserve_slowpath(struct ttm_buffer_object *bo,
+			    bool interruptible, uint32_t sequence)
+{
+	struct ttm_bo_global *glob = bo->glob;
+	int put_count, ret;
+
+	ret = ttm_bo_reserve_slowpath_nolru(bo, interruptible, sequence);
+	if (likely(!ret)) {
+		spin_lock(&glob->lru_lock);
+		put_count = ttm_bo_del_from_lru(bo);
+		spin_unlock(&glob->lru_lock);
+		ttm_bo_list_ref_sub(bo, put_count, true);
+	}
+	return ret;
+}
+EXPORT_SYMBOL(ttm_bo_reserve_slowpath);
+
 void ttm_bo_unreserve_locked(struct ttm_buffer_object *bo)
 {
 	ttm_bo_add_to_lru(bo);
diff --git a/include/drm/ttm/ttm_bo_driver.h b/include/drm/ttm/ttm_bo_driver.h
index 6fff432..5af71af 100644
--- a/include/drm/ttm/ttm_bo_driver.h
+++ b/include/drm/ttm/ttm_bo_driver.h
@@ -821,6 +821,36 @@ extern int ttm_bo_reserve(struct ttm_buffer_object *bo,
 			  bool interruptible,
 			  bool no_wait, bool use_sequence, uint32_t sequence);
 
+/**
+ * ttm_bo_reserve_slowpath_nolru:
+ * @bo: A pointer to a struct ttm_buffer_object.
+ * @interruptible: Sleep interruptible if waiting.
+ * @sequence: Set (@bo)->sequence to this value after lock
+ *
+ * This is called after ttm_bo_reserve returns -EAGAIN and we backed off
+ * from all our other reservations. Because there are no other reservations
+ * held by us, this function cannot deadlock any more.
+ *
+ * Will not remove reserved buffers from the lru lists.
+ * Otherwise identical to ttm_bo_reserve_slowpath.
+ */
+extern int ttm_bo_reserve_slowpath_nolru(struct ttm_buffer_object *bo,
+					 bool interruptible,
+					 uint32_t sequence);
+
+
+/**
+ * ttm_bo_reserve_slowpath:
+ * @bo: A pointer to a struct ttm_buffer_object.
+ * @interruptible: Sleep interruptible if waiting.
+ * @sequence: Set (@bo)->sequence to this value after lock
+ *
+ * This is called after ttm_bo_reserve returns -EAGAIN and we backed off
+ * from all our other reservations. Because there are no other reservations
+ * held by us, this function cannot deadlock any more.
+ */
+extern int ttm_bo_reserve_slowpath(struct ttm_buffer_object *bo,
+				   bool interruptible, uint32_t sequence);
 
 /**
  * ttm_bo_reserve_nolru:
-- 
cgit v0.10.2


From f2d476a110bc24fde008698ae9018c99e803e25c Mon Sep 17 00:00:00 2001
From: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Date: Tue, 15 Jan 2013 14:57:10 +0100
Subject: drm/ttm: use ttm_bo_reserve_slowpath_nolru in ttm_eu_reserve_buffers,
 v2

This requires re-use of the seqno, which increases fairness slightly.
Instead of spinning with a new seqno every time we keep the current one,
but still drop all other reservations we hold. Only when we succeed,
we try to get back our other reservations again.

This should increase fairness slightly as well.

Changes since v1:
 - Increase val_seq before calling ttm_bo_reserve_slowpath_nolru and
   retrying to take all entries to prevent a race.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>

diff --git a/drivers/gpu/drm/ttm/ttm_execbuf_util.c b/drivers/gpu/drm/ttm/ttm_execbuf_util.c
index c7d3236..7b90def 100644
--- a/drivers/gpu/drm/ttm/ttm_execbuf_util.c
+++ b/drivers/gpu/drm/ttm/ttm_execbuf_util.c
@@ -129,13 +129,17 @@ int ttm_eu_reserve_buffers(struct list_head *list)
 	entry = list_first_entry(list, struct ttm_validate_buffer, head);
 	glob = entry->bo->glob;
 
-retry:
 	spin_lock(&glob->lru_lock);
 	val_seq = entry->bo->bdev->val_seq++;
 
+retry:
 	list_for_each_entry(entry, list, head) {
 		struct ttm_buffer_object *bo = entry->bo;
 
+		/* already slowpath reserved? */
+		if (entry->reserved)
+			continue;
+
 		ret = ttm_bo_reserve_nolru(bo, true, true, true, val_seq);
 		switch (ret) {
 		case 0:
@@ -155,11 +159,26 @@ retry:
 			/* fallthrough */
 		case -EAGAIN:
 			ttm_eu_backoff_reservation_locked(list);
+
+			/*
+			 * temporarily increase sequence number every retry,
+			 * to prevent us from seeing our old reservation
+			 * sequence when someone else reserved the buffer,
+			 * but hasn't updated the seq_valid/seqno members yet.
+			 */
+			val_seq = entry->bo->bdev->val_seq++;
+
 			spin_unlock(&glob->lru_lock);
 			ttm_eu_list_ref_sub(list);
-			ret = ttm_bo_wait_unreserved(bo, true);
+			ret = ttm_bo_reserve_slowpath_nolru(bo, true, val_seq);
 			if (unlikely(ret != 0))
 				return ret;
+			spin_lock(&glob->lru_lock);
+			entry->reserved = true;
+			if (unlikely(atomic_read(&bo->cpu_writers) > 0)) {
+				ret = -EBUSY;
+				goto err;
+			}
 			goto retry;
 		default:
 			goto err;
-- 
cgit v0.10.2


From c354c893dd57aac11f5d96ada7c47a20fe090a6e Mon Sep 17 00:00:00 2001
From: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Date: Tue, 15 Jan 2013 14:57:20 +0100
Subject: drm/nouveau: use ttm_bo_reserve_slowpath in validate_init, v2

Similar rationale to the identical commit in drm/ttm.
Instead of only waiting for unreservation, we make sure we actually
own the reservation, then retry to get the rest.

Changes since v1:
 - Increase the seqno before calling ttm_bo_reserve_slowpath

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c b/drivers/gpu/drm/nouveau/nouveau_gem.c
index 9fcfcb2..7fa6882 100644
--- a/drivers/gpu/drm/nouveau/nouveau_gem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_gem.c
@@ -320,9 +320,10 @@ validate_init(struct nouveau_channel *chan, struct drm_file *file_priv,
 	uint32_t sequence;
 	int trycnt = 0;
 	int ret, i;
+	struct nouveau_bo *res_bo = NULL;
 
-retry:
 	sequence = atomic_add_return(1, &drm->ttm.validate_sequence);
+retry:
 	if (++trycnt > 100000) {
 		NV_ERROR(drm, "%s failed and gave up.\n", __func__);
 		return -EINVAL;
@@ -340,6 +341,11 @@ retry:
 			return -ENOENT;
 		}
 		nvbo = gem->driver_private;
+		if (nvbo == res_bo) {
+			res_bo = NULL;
+			drm_gem_object_unreference_unlocked(gem);
+			continue;
+		}
 
 		if (nvbo->reserved_by && nvbo->reserved_by == file_priv) {
 			NV_ERROR(drm, "multiple instances of buffer %d on "
@@ -352,15 +358,19 @@ retry:
 		ret = ttm_bo_reserve(&nvbo->bo, true, false, true, sequence);
 		if (ret) {
 			validate_fini(op, NULL);
-			if (unlikely(ret == -EAGAIN))
-				ret = ttm_bo_wait_unreserved(&nvbo->bo, true);
-			drm_gem_object_unreference_unlocked(gem);
+			if (unlikely(ret == -EAGAIN)) {
+				sequence = atomic_add_return(1, &drm->ttm.validate_sequence);
+				ret = ttm_bo_reserve_slowpath(&nvbo->bo, true,
+							      sequence);
+				if (!ret)
+					res_bo = nvbo;
+			}
 			if (unlikely(ret)) {
+				drm_gem_object_unreference_unlocked(gem);
 				if (ret != -ERESTARTSYS)
 					NV_ERROR(drm, "fail reserve\n");
 				return ret;
 			}
-			goto retry;
 		}
 
 		b->user_priv = (uint64_t)(unsigned long)nvbo;
@@ -382,6 +392,8 @@ retry:
 			validate_fini(op, NULL);
 			return -EINVAL;
 		}
+		if (nvbo == res_bo)
+			goto retry;
 	}
 
 	return 0;
-- 
cgit v0.10.2


From cc4c0c4de3c775be22072ec3251f2e581b63d9a0 Mon Sep 17 00:00:00 2001
From: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Date: Tue, 15 Jan 2013 14:57:28 +0100
Subject: drm/ttm: unexport ttm_bo_wait_unreserved

All legitimate users of this function outside ttm_bo.c are gone, now
it's only an implementation detail.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>

diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index 4dd6f9e..4df47f7 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -158,7 +158,8 @@ static void ttm_bo_release_list(struct kref *list_kref)
 	ttm_mem_global_free(bdev->glob->mem_glob, acc_size);
 }
 
-int ttm_bo_wait_unreserved(struct ttm_buffer_object *bo, bool interruptible)
+static int ttm_bo_wait_unreserved(struct ttm_buffer_object *bo,
+				  bool interruptible)
 {
 	if (interruptible) {
 		return wait_event_interruptible(bo->event_queue,
@@ -168,7 +169,6 @@ int ttm_bo_wait_unreserved(struct ttm_buffer_object *bo, bool interruptible)
 		return 0;
 	}
 }
-EXPORT_SYMBOL(ttm_bo_wait_unreserved);
 
 void ttm_bo_add_to_lru(struct ttm_buffer_object *bo)
 {
diff --git a/include/drm/ttm/ttm_bo_driver.h b/include/drm/ttm/ttm_bo_driver.h
index 5af71af..0fbd046 100644
--- a/include/drm/ttm/ttm_bo_driver.h
+++ b/include/drm/ttm/ttm_bo_driver.h
@@ -898,18 +898,6 @@ extern void ttm_bo_unreserve(struct ttm_buffer_object *bo);
  */
 extern void ttm_bo_unreserve_locked(struct ttm_buffer_object *bo);
 
-/**
- * ttm_bo_wait_unreserved
- *
- * @bo: A pointer to a struct ttm_buffer_object.
- *
- * Wait for a struct ttm_buffer_object to become unreserved.
- * This is typically used in the execbuf code to relax cpu-usage when
- * a potential deadlock condition backoff.
- */
-extern int ttm_bo_wait_unreserved(struct ttm_buffer_object *bo,
-				  bool interruptible);
-
 /*
  * ttm_bo_util.c
  */
-- 
cgit v0.10.2


From 2c10d571169abab9181e0c91b259e0c4e392bc5b Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 20 Dec 2012 21:24:07 +0100
Subject: drm/i915: wake up all pageflip waiters

Otherwise it seems like we can get stuck with concurrent waiters.
Right now this /shouldn't/ be a problem, since all pending pageflip
waiters are serialized by the one mode_config.mutex, so there's at
most on waiter. But better paranoid than sorry, since this is tricky
code.

v2: WARN_ON(waitqueue_active) before waiting, as suggested by Chris
Wilson.

Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 8c36a11..40b6b5e9 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -2220,6 +2220,8 @@ intel_finish_fb(struct drm_framebuffer *old_fb)
 	bool was_interruptible = dev_priv->mm.interruptible;
 	int ret;
 
+	WARN_ON(waitqueue_active(&dev_priv->pending_flip_queue));
+
 	wait_event(dev_priv->pending_flip_queue,
 		   atomic_read(&dev_priv->mm.wedged) ||
 		   atomic_read(&obj->pending_flip) == 0);
@@ -2887,6 +2889,8 @@ static void intel_crtc_wait_for_pending_flips(struct drm_crtc *crtc)
 	if (crtc->fb == NULL)
 		return;
 
+	WARN_ON(waitqueue_active(&dev_priv->pending_flip_queue));
+
 	wait_event(dev_priv->pending_flip_queue,
 		   !intel_crtc_has_pending_flip(crtc));
 
@@ -6824,7 +6828,7 @@ static void do_intel_finish_page_flip(struct drm_device *dev,
 
 	obj = work->old_fb_obj;
 
-	wake_up(&dev_priv->pending_flip_queue);
+	wake_up_all(&dev_priv->pending_flip_queue);
 
 	queue_work(dev_priv->wq, &work->work);
 
-- 
cgit v0.10.2


From af5163acd8319dbc901bb59a8d503d8bb774d88b Mon Sep 17 00:00:00 2001
From: Egbert Eich <eich@suse.de>
Date: Thu, 10 Jan 2013 10:02:39 -0500
Subject: drm/i915: Remove pch_rq_mask from struct drm_i915_private.

This variable is only used locally in the irq postinstall
functions for ivybridge and ironlake.

Signed-off-by: Egbert Eich <eich@suse.de>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index b1b1b73..910cc0b 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -697,7 +697,6 @@ typedef struct drm_i915_private {
 	u32 pipestat[2];
 	u32 irq_mask;
 	u32 gt_irq_mask;
-	u32 pch_irq_mask;
 
 	u32 hotplug_supported_mask;
 	struct work_struct hotplug_work;
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 6689a61..2028131 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -1892,6 +1892,7 @@ static int ironlake_irq_postinstall(struct drm_device *dev)
 			   DE_AUX_CHANNEL_A;
 	u32 render_irqs;
 	u32 hotplug_mask;
+	u32 pch_irq_mask;
 
 	dev_priv->irq_mask = ~display_mask;
 
@@ -1935,10 +1936,10 @@ static int ironlake_irq_postinstall(struct drm_device *dev)
 				SDE_AUX_MASK);
 	}
 
-	dev_priv->pch_irq_mask = ~hotplug_mask;
+	pch_irq_mask = ~hotplug_mask;
 
 	I915_WRITE(SDEIIR, I915_READ(SDEIIR));
-	I915_WRITE(SDEIMR, dev_priv->pch_irq_mask);
+	I915_WRITE(SDEIMR, pch_irq_mask);
 	I915_WRITE(SDEIER, hotplug_mask);
 	POSTING_READ(SDEIER);
 
@@ -1966,6 +1967,7 @@ static int ivybridge_irq_postinstall(struct drm_device *dev)
 		DE_AUX_CHANNEL_A_IVB;
 	u32 render_irqs;
 	u32 hotplug_mask;
+	u32 pch_irq_mask;
 
 	dev_priv->irq_mask = ~display_mask;
 
@@ -1995,10 +1997,10 @@ static int ivybridge_irq_postinstall(struct drm_device *dev)
 			SDE_PORTD_HOTPLUG_CPT |
 			SDE_GMBUS_CPT |
 			SDE_AUX_MASK_CPT);
-	dev_priv->pch_irq_mask = ~hotplug_mask;
+	pch_irq_mask = ~hotplug_mask;
 
 	I915_WRITE(SDEIIR, I915_READ(SDEIIR));
-	I915_WRITE(SDEIMR, dev_priv->pch_irq_mask);
+	I915_WRITE(SDEIMR, pch_irq_mask);
 	I915_WRITE(SDEIER, hotplug_mask);
 	POSTING_READ(SDEIER);
 
-- 
cgit v0.10.2


From d865110cc2345d67752cd7e1350b391c34feb2aa Mon Sep 17 00:00:00 2001
From: Imre Deak <imre.deak@intel.com>
Date: Mon, 7 Jan 2013 21:47:33 +0200
Subject: drm/i915: merge get_gtt_alignment/get_unfenced_gtt_alignment()

The two functions are rather similar, so merge them.

Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 910cc0b..0b09361 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -1565,9 +1565,8 @@ void i915_gem_free_all_phys_object(struct drm_device *dev);
 void i915_gem_release(struct drm_device *dev, struct drm_file *file);
 
 uint32_t
-i915_gem_get_unfenced_gtt_alignment(struct drm_device *dev,
-				    uint32_t size,
-				    int tiling_mode);
+i915_gem_get_gtt_alignment(struct drm_device *dev, uint32_t size,
+			    int tiling_mode, bool fenced);
 
 int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj,
 				    enum i915_cache_level cache_level);
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index e6cc020..2166b61 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1463,16 +1463,15 @@ i915_gem_get_gtt_size(struct drm_device *dev, uint32_t size, int tiling_mode)
  * Return the required GTT alignment for an object, taking into account
  * potential fence register mapping.
  */
-static uint32_t
-i915_gem_get_gtt_alignment(struct drm_device *dev,
-			   uint32_t size,
-			   int tiling_mode)
+uint32_t
+i915_gem_get_gtt_alignment(struct drm_device *dev, uint32_t size,
+			   int tiling_mode, bool fenced)
 {
 	/*
 	 * Minimum alignment is 4k (GTT page size), but might be greater
 	 * if a fence register is needed for the object.
 	 */
-	if (INTEL_INFO(dev)->gen >= 4 ||
+	if (INTEL_INFO(dev)->gen >= 4 || (!fenced && IS_G33(dev)) ||
 	    tiling_mode == I915_TILING_NONE)
 		return 4096;
 
@@ -1483,35 +1482,6 @@ i915_gem_get_gtt_alignment(struct drm_device *dev,
 	return i915_gem_get_gtt_size(dev, size, tiling_mode);
 }
 
-/**
- * i915_gem_get_unfenced_gtt_alignment - return required GTT alignment for an
- *					 unfenced object
- * @dev: the device
- * @size: size of the object
- * @tiling_mode: tiling mode of the object
- *
- * Return the required GTT alignment for an object, only taking into account
- * unfenced tiled surface requirements.
- */
-uint32_t
-i915_gem_get_unfenced_gtt_alignment(struct drm_device *dev,
-				    uint32_t size,
-				    int tiling_mode)
-{
-	/*
-	 * Minimum alignment is 4k (GTT page size) for sane hw.
-	 */
-	if (INTEL_INFO(dev)->gen >= 4 || IS_G33(dev) ||
-	    tiling_mode == I915_TILING_NONE)
-		return 4096;
-
-	/* Previous hardware however needs to be aligned to a power-of-two
-	 * tile height. The simplest method for determining this is to reuse
-	 * the power-of-tile object size.
-	 */
-	return i915_gem_get_gtt_size(dev, size, tiling_mode);
-}
-
 static int i915_gem_object_create_mmap_offset(struct drm_i915_gem_object *obj)
 {
 	struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
@@ -2934,11 +2904,11 @@ i915_gem_object_bind_to_gtt(struct drm_i915_gem_object *obj,
 					   obj->tiling_mode);
 	fence_alignment = i915_gem_get_gtt_alignment(dev,
 						     obj->base.size,
-						     obj->tiling_mode);
+						     obj->tiling_mode, true);
 	unfenced_alignment =
-		i915_gem_get_unfenced_gtt_alignment(dev,
+		i915_gem_get_gtt_alignment(dev,
 						    obj->base.size,
-						    obj->tiling_mode);
+						    obj->tiling_mode, false);
 
 	if (alignment == 0)
 		alignment = map_and_fenceable ? fence_alignment :
diff --git a/drivers/gpu/drm/i915/i915_gem_tiling.c b/drivers/gpu/drm/i915/i915_gem_tiling.c
index 65f1d4f..cb71ded 100644
--- a/drivers/gpu/drm/i915/i915_gem_tiling.c
+++ b/drivers/gpu/drm/i915/i915_gem_tiling.c
@@ -374,9 +374,9 @@ i915_gem_set_tiling(struct drm_device *dev, void *data,
 		/* Rebind if we need a change of alignment */
 		if (!obj->map_and_fenceable) {
 			u32 unfenced_alignment =
-				i915_gem_get_unfenced_gtt_alignment(dev,
-								    obj->base.size,
-								    args->tiling_mode);
+				i915_gem_get_gtt_alignment(dev, obj->base.size,
+							    args->tiling_mode,
+							    false);
 			if (obj->gtt_offset & (unfenced_alignment - 1))
 				ret = i915_gem_object_unbind(obj);
 		}
-- 
cgit v0.10.2


From 56c844e539f1f6f5768c5f73f119e6f4aed9d320 Mon Sep 17 00:00:00 2001
From: Imre Deak <imre.deak@intel.com>
Date: Mon, 7 Jan 2013 21:47:34 +0200
Subject: drm/i915: merge {i965, sandybridge}_write_fence_reg()

The two functions are rather similar, so merge them.

Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 2166b61..718574e 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -2507,52 +2507,38 @@ int i915_gpu_idle(struct drm_device *dev)
 	return 0;
 }
 
-static void sandybridge_write_fence_reg(struct drm_device *dev, int reg,
-					struct drm_i915_gem_object *obj)
-{
-	drm_i915_private_t *dev_priv = dev->dev_private;
-	uint64_t val;
-
-	if (obj) {
-		u32 size = obj->gtt_space->size;
-
-		val = (uint64_t)((obj->gtt_offset + size - 4096) &
-				 0xfffff000) << 32;
-		val |= obj->gtt_offset & 0xfffff000;
-		val |= (uint64_t)((obj->stride / 128) - 1) <<
-			SANDYBRIDGE_FENCE_PITCH_SHIFT;
-
-		if (obj->tiling_mode == I915_TILING_Y)
-			val |= 1 << I965_FENCE_TILING_Y_SHIFT;
-		val |= I965_FENCE_REG_VALID;
-	} else
-		val = 0;
-
-	I915_WRITE64(FENCE_REG_SANDYBRIDGE_0 + reg * 8, val);
-	POSTING_READ(FENCE_REG_SANDYBRIDGE_0 + reg * 8);
-}
-
 static void i965_write_fence_reg(struct drm_device *dev, int reg,
 				 struct drm_i915_gem_object *obj)
 {
 	drm_i915_private_t *dev_priv = dev->dev_private;
+	int fence_reg;
+	int fence_pitch_shift;
 	uint64_t val;
 
+	if (INTEL_INFO(dev)->gen >= 6) {
+		fence_reg = FENCE_REG_SANDYBRIDGE_0;
+		fence_pitch_shift = SANDYBRIDGE_FENCE_PITCH_SHIFT;
+	} else {
+		fence_reg = FENCE_REG_965_0;
+		fence_pitch_shift = I965_FENCE_PITCH_SHIFT;
+	}
+
 	if (obj) {
 		u32 size = obj->gtt_space->size;
 
 		val = (uint64_t)((obj->gtt_offset + size - 4096) &
 				 0xfffff000) << 32;
 		val |= obj->gtt_offset & 0xfffff000;
-		val |= ((obj->stride / 128) - 1) << I965_FENCE_PITCH_SHIFT;
+		val |= (uint64_t)((obj->stride / 128) - 1) << fence_pitch_shift;
 		if (obj->tiling_mode == I915_TILING_Y)
 			val |= 1 << I965_FENCE_TILING_Y_SHIFT;
 		val |= I965_FENCE_REG_VALID;
 	} else
 		val = 0;
 
-	I915_WRITE64(FENCE_REG_965_0 + reg * 8, val);
-	POSTING_READ(FENCE_REG_965_0 + reg * 8);
+	fence_reg += reg * 8;
+	I915_WRITE64(fence_reg, val);
+	POSTING_READ(fence_reg);
 }
 
 static void i915_write_fence_reg(struct drm_device *dev, int reg,
@@ -2636,7 +2622,7 @@ static void i915_gem_write_fence(struct drm_device *dev, int reg,
 {
 	switch (INTEL_INFO(dev)->gen) {
 	case 7:
-	case 6: sandybridge_write_fence_reg(dev, reg, obj); break;
+	case 6:
 	case 5:
 	case 4: i965_write_fence_reg(dev, reg, obj); break;
 	case 3: i915_write_fence_reg(dev, reg, obj); break;
-- 
cgit v0.10.2


From 0fa877965194fa79fe87944844185d90cfc35352 Mon Sep 17 00:00:00 2001
From: Imre Deak <imre.deak@intel.com>
Date: Mon, 7 Jan 2013 21:47:35 +0200
Subject: drm/i915: use gtt_get_size() instead of open coding it

Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 0b09361..35ecabc 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -1565,6 +1565,8 @@ void i915_gem_free_all_phys_object(struct drm_device *dev);
 void i915_gem_release(struct drm_device *dev, struct drm_file *file);
 
 uint32_t
+i915_gem_get_gtt_size(struct drm_device *dev, uint32_t size, int tiling_mode);
+uint32_t
 i915_gem_get_gtt_alignment(struct drm_device *dev, uint32_t size,
 			    int tiling_mode, bool fenced);
 
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 718574e..313bdbab 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1435,7 +1435,7 @@ i915_gem_release_mmap(struct drm_i915_gem_object *obj)
 	obj->fault_mappable = false;
 }
 
-static uint32_t
+uint32_t
 i915_gem_get_gtt_size(struct drm_device *dev, uint32_t size, int tiling_mode)
 {
 	uint32_t gtt_size;
diff --git a/drivers/gpu/drm/i915/i915_gem_tiling.c b/drivers/gpu/drm/i915/i915_gem_tiling.c
index cb71ded..e76f0d8 100644
--- a/drivers/gpu/drm/i915/i915_gem_tiling.c
+++ b/drivers/gpu/drm/i915/i915_gem_tiling.c
@@ -272,18 +272,7 @@ i915_gem_object_fence_ok(struct drm_i915_gem_object *obj, int tiling_mode)
 			return false;
 	}
 
-	/*
-	 * Previous chips need to be aligned to the size of the smallest
-	 * fence register that can contain the object.
-	 */
-	if (INTEL_INFO(obj->base.dev)->gen == 3)
-		size = 1024*1024;
-	else
-		size = 512*1024;
-
-	while (size < obj->base.size)
-		size <<= 1;
-
+	size = i915_gem_get_gtt_size(obj->base.dev, obj->base.size, tiling_mode);
 	if (obj->gtt_space->size != size)
 		return false;
 
-- 
cgit v0.10.2


From dd624afd533bdc87b8c10835515a0c8b2b9868b1 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Tue, 15 Jan 2013 12:39:35 +0000
Subject: drm/i915: Add a debug interface to forcibly evict and shrink our
 object caches

As a means to investigate some bad system behaviour related to the
purging of the active, inactive and unbound lists, it is useful to be
able to manually control when those lists should be cleared.

v2: use _safe list iterators as we kick objects from the list as we
walk.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
[danvet: Add a small comment explaining why we don't need to check and
wait for gpu resets, acked by Chris on irc.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index f7d88e9..f94418b 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -1776,6 +1776,102 @@ static const struct file_operations i915_ring_stop_fops = {
 	.llseek = default_llseek,
 };
 
+#define DROP_UNBOUND 0x1
+#define DROP_BOUND 0x2
+#define DROP_RETIRE 0x4
+#define DROP_ACTIVE 0x8
+#define DROP_ALL (DROP_UNBOUND | \
+		  DROP_BOUND | \
+		  DROP_RETIRE | \
+		  DROP_ACTIVE)
+static ssize_t
+i915_drop_caches_read(struct file *filp,
+		      char __user *ubuf,
+		      size_t max,
+		      loff_t *ppos)
+{
+	char buf[20];
+	int len;
+
+	len = snprintf(buf, sizeof(buf), "0x%08x\n", DROP_ALL);
+	if (len > sizeof(buf))
+		len = sizeof(buf);
+
+	return simple_read_from_buffer(ubuf, max, ppos, buf, len);
+}
+
+static ssize_t
+i915_drop_caches_write(struct file *filp,
+		       const char __user *ubuf,
+		       size_t cnt,
+		       loff_t *ppos)
+{
+	struct drm_device *dev = filp->private_data;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_gem_object *obj, *next;
+	char buf[20];
+	int val = 0, ret;
+
+	if (cnt > 0) {
+		if (cnt > sizeof(buf) - 1)
+			return -EINVAL;
+
+		if (copy_from_user(buf, ubuf, cnt))
+			return -EFAULT;
+		buf[cnt] = 0;
+
+		val = simple_strtoul(buf, NULL, 0);
+	}
+
+	DRM_DEBUG_DRIVER("Dropping caches: 0x%08x\n", val);
+
+	/* No need to check and wait for gpu resets, only libdrm auto-restarts
+	 * on ioctls on -EAGAIN. */
+	ret = mutex_lock_interruptible(&dev->struct_mutex);
+	if (ret)
+		return ret;
+
+	if (val & DROP_ACTIVE) {
+		ret = i915_gpu_idle(dev);
+		if (ret)
+			goto unlock;
+	}
+
+	if (val & (DROP_RETIRE | DROP_ACTIVE))
+		i915_gem_retire_requests(dev);
+
+	if (val & DROP_BOUND) {
+		list_for_each_entry_safe(obj, next, &dev_priv->mm.inactive_list, mm_list)
+			if (obj->pin_count == 0) {
+				ret = i915_gem_object_unbind(obj);
+				if (ret)
+					goto unlock;
+			}
+	}
+
+	if (val & DROP_UNBOUND) {
+		list_for_each_entry_safe(obj, next, &dev_priv->mm.unbound_list, gtt_list)
+			if (obj->pages_pin_count == 0) {
+				ret = i915_gem_object_put_pages(obj);
+				if (ret)
+					goto unlock;
+			}
+	}
+
+unlock:
+	mutex_unlock(&dev->struct_mutex);
+
+	return ret ?: cnt;
+}
+
+static const struct file_operations i915_drop_caches_fops = {
+	.owner = THIS_MODULE,
+	.open = simple_open,
+	.read = i915_drop_caches_read,
+	.write = i915_drop_caches_write,
+	.llseek = default_llseek,
+};
+
 static ssize_t
 i915_max_freq_read(struct file *filp,
 		   char __user *ubuf,
@@ -2173,6 +2269,12 @@ int i915_debugfs_init(struct drm_minor *minor)
 		return ret;
 
 	ret = i915_debugfs_create(minor->debugfs_root, minor,
+				  "i915_gem_drop_caches",
+				  &i915_drop_caches_fops);
+	if (ret)
+		return ret;
+
+	ret = i915_debugfs_create(minor->debugfs_root, minor,
 				  "i915_error_state",
 				  &i915_error_state_fops);
 	if (ret)
@@ -2203,6 +2305,8 @@ void i915_debugfs_cleanup(struct drm_minor *minor)
 				 1, minor);
 	drm_debugfs_remove_files((struct drm_info_list *) &i915_cache_sharing_fops,
 				 1, minor);
+	drm_debugfs_remove_files((struct drm_info_list *) &i915_drop_caches_fops,
+				 1, minor);
 	drm_debugfs_remove_files((struct drm_info_list *) &i915_ring_stop_fops,
 				 1, minor);
 	drm_debugfs_remove_files((struct drm_info_list *) &i915_error_state_fops,
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 35ecabc..8f5b281 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -1445,6 +1445,7 @@ int __must_check i915_gem_object_pin(struct drm_i915_gem_object *obj,
 				     bool nonblocking);
 void i915_gem_object_unpin(struct drm_i915_gem_object *obj);
 int __must_check i915_gem_object_unbind(struct drm_i915_gem_object *obj);
+int i915_gem_object_put_pages(struct drm_i915_gem_object *obj);
 void i915_gem_release_mmap(struct drm_i915_gem_object *obj);
 void i915_gem_lastclose(struct drm_device *dev);
 
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 313bdbab..9d33fb3 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1662,7 +1662,7 @@ i915_gem_object_put_pages_gtt(struct drm_i915_gem_object *obj)
 	kfree(obj->pages);
 }
 
-static int
+int
 i915_gem_object_put_pages(struct drm_i915_gem_object *obj)
 {
 	const struct drm_i915_gem_object_ops *ops = obj->ops;
-- 
cgit v0.10.2


From 43e28f092b2fa4ebc46bdc210134a80610815785 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Tue, 8 Jan 2013 10:53:09 +0000
Subject: drm/i915: Bail if we attempt to allocate pages for a purged object

Move the existing checking inside bind_to_gtt() to the more appropriate
layer in order to prevent recreation of the pages after they have been
explicitly truncated.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 9d33fb3..a2bb189 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1828,6 +1828,11 @@ i915_gem_object_get_pages(struct drm_i915_gem_object *obj)
 	if (obj->pages)
 		return 0;
 
+	if (obj->madv != I915_MADV_WILLNEED) {
+		DRM_ERROR("Attempting to obtain a purgeable object\n");
+		return -EINVAL;
+	}
+
 	BUG_ON(obj->pages_pin_count);
 
 	ret = ops->get_pages(obj);
@@ -2440,7 +2445,7 @@ int
 i915_gem_object_unbind(struct drm_i915_gem_object *obj)
 {
 	drm_i915_private_t *dev_priv = obj->base.dev->dev_private;
-	int ret = 0;
+	int ret;
 
 	if (obj->gtt_space == NULL)
 		return 0;
@@ -2880,11 +2885,6 @@ i915_gem_object_bind_to_gtt(struct drm_i915_gem_object *obj,
 	bool mappable, fenceable;
 	int ret;
 
-	if (obj->madv != I915_MADV_WILLNEED) {
-		DRM_ERROR("Attempting to bind a purgeable object\n");
-		return -EINVAL;
-	}
-
 	fence_size = i915_gem_get_gtt_size(dev,
 					   obj->base.size,
 					   obj->tiling_mode);
-- 
cgit v0.10.2


From 419fa72a1973c9ba2fd2c2505dc889f54b857459 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Tue, 8 Jan 2013 10:53:13 +0000
Subject: drm/i915: Mark a temporary allocation for copy-from-user as such

The difference is that the kernel will then know that this memory will
be reclaimable in the near future.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 163bb52b..986eb98 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -1113,7 +1113,7 @@ i915_gem_execbuffer2(struct drm_device *dev, void *data,
 	}
 
 	exec2_list = kmalloc(sizeof(*exec2_list)*args->buffer_count,
-			     GFP_KERNEL | __GFP_NOWARN | __GFP_NORETRY);
+			     GFP_TEMPORARY | __GFP_NOWARN | __GFP_NORETRY);
 	if (exec2_list == NULL)
 		exec2_list = drm_malloc_ab(sizeof(*exec2_list),
 					   args->buffer_count);
-- 
cgit v0.10.2


From 3b96eff447b4ca34ca8ccd42e2651be2955f34b4 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Tue, 8 Jan 2013 10:53:14 +0000
Subject: drm/i915: Take the handle idr spinlock once for looking up the exec
 objects

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 986eb98..da103c1 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -69,6 +69,46 @@ eb_add_object(struct eb_objects *eb, struct drm_i915_gem_object *obj)
 		       &eb->buckets[obj->exec_handle & eb->and]);
 }
 
+static int
+eb_lookup_objects(struct eb_objects *eb,
+		  struct drm_i915_gem_exec_object2 *exec,
+		  int count,
+		  struct drm_file *file,
+		  struct list_head *objects)
+{
+	int i;
+
+	spin_lock(&file->table_lock);
+	for (i = 0; i < count; i++) {
+		struct drm_i915_gem_object *obj;
+
+		obj = to_intel_bo(idr_find(&file->object_idr, exec[i].handle));
+		if (obj == NULL) {
+			spin_unlock(&file->table_lock);
+			DRM_DEBUG("Invalid object handle %d at index %d\n",
+				   exec[i].handle, i);
+			return -ENOENT;
+		}
+
+		if (!list_empty(&obj->exec_list)) {
+			spin_unlock(&file->table_lock);
+			DRM_DEBUG("Object %p [handle %d, index %d] appears more than once in object list\n",
+				   obj, exec[i].handle, i);
+			return -EINVAL;
+		}
+
+		drm_gem_object_reference(&obj->base);
+		list_add_tail(&obj->exec_list, objects);
+
+		obj->exec_handle = exec[i].handle;
+		obj->exec_entry = &exec[i];
+		eb_add_object(eb, obj);
+	}
+	spin_unlock(&file->table_lock);
+
+	return 0;
+}
+
 static struct drm_i915_gem_object *
 eb_get_object(struct eb_objects *eb, unsigned long handle)
 {
@@ -550,21 +590,9 @@ i915_gem_execbuffer_relocate_slow(struct drm_device *dev,
 
 	/* reacquire the objects */
 	eb_reset(eb);
-	for (i = 0; i < count; i++) {
-		obj = to_intel_bo(drm_gem_object_lookup(dev, file,
-							exec[i].handle));
-		if (&obj->base == NULL) {
-			DRM_DEBUG("Invalid object handle %d at index %d\n",
-				   exec[i].handle, i);
-			ret = -ENOENT;
-			goto err;
-		}
-
-		list_add_tail(&obj->exec_list, objects);
-		obj->exec_handle = exec[i].handle;
-		obj->exec_entry = &exec[i];
-		eb_add_object(eb, obj);
-	}
+	ret = eb_lookup_objects(eb, exec, count, file, objects);
+	if (ret)
+		goto err;
 
 	ret = i915_gem_execbuffer_reserve(ring, file, objects);
 	if (ret)
@@ -872,31 +900,9 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 
 	/* Look up object handles */
 	INIT_LIST_HEAD(&objects);
-	for (i = 0; i < args->buffer_count; i++) {
-		struct drm_i915_gem_object *obj;
-
-		obj = to_intel_bo(drm_gem_object_lookup(dev, file,
-							exec[i].handle));
-		if (&obj->base == NULL) {
-			DRM_DEBUG("Invalid object handle %d at index %d\n",
-				   exec[i].handle, i);
-			/* prevent error path from reading uninitialized data */
-			ret = -ENOENT;
-			goto err;
-		}
-
-		if (!list_empty(&obj->exec_list)) {
-			DRM_DEBUG("Object %p [handle %d, index %d] appears more than once in object list\n",
-				   obj, exec[i].handle, i);
-			ret = -EINVAL;
-			goto err;
-		}
-
-		list_add_tail(&obj->exec_list, &objects);
-		obj->exec_handle = exec[i].handle;
-		obj->exec_entry = &exec[i];
-		eb_add_object(eb, obj);
-	}
+	ret = eb_lookup_objects(eb, exec, args->buffer_count, file, &objects);
+	if (ret)
+		goto err;
 
 	/* take note of the batch buffer before we might reorder the lists */
 	batch_obj = list_entry(objects.prev,
-- 
cgit v0.10.2


From bcffc3faa692d6b2ef734e4f0c8f097175284db6 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Tue, 8 Jan 2013 10:53:15 +0000
Subject: drm/i915: Move the execbuffer objects list from the stack into the
 tracker

Instead of passing around the eb-objects hashtable and a separate object
list, we can include the object list into the eb-objects structure for
convenience.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index da103c1..386677f 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -34,6 +34,7 @@
 #include <linux/dma_remapping.h>
 
 struct eb_objects {
+	struct list_head objects;
 	int and;
 	struct hlist_head buckets[0];
 };
@@ -53,6 +54,7 @@ eb_create(int size)
 		return eb;
 
 	eb->and = count - 1;
+	INIT_LIST_HEAD(&eb->objects);
 	return eb;
 }
 
@@ -73,8 +75,7 @@ static int
 eb_lookup_objects(struct eb_objects *eb,
 		  struct drm_i915_gem_exec_object2 *exec,
 		  int count,
-		  struct drm_file *file,
-		  struct list_head *objects)
+		  struct drm_file *file)
 {
 	int i;
 
@@ -98,7 +99,7 @@ eb_lookup_objects(struct eb_objects *eb,
 		}
 
 		drm_gem_object_reference(&obj->base);
-		list_add_tail(&obj->exec_list, objects);
+		list_add_tail(&obj->exec_list, &eb->objects);
 
 		obj->exec_handle = exec[i].handle;
 		obj->exec_entry = &exec[i];
@@ -129,6 +130,15 @@ eb_get_object(struct eb_objects *eb, unsigned long handle)
 static void
 eb_destroy(struct eb_objects *eb)
 {
+	while (!list_empty(&eb->objects)) {
+		struct drm_i915_gem_object *obj;
+
+		obj = list_first_entry(&eb->objects,
+				       struct drm_i915_gem_object,
+				       exec_list);
+		list_del_init(&obj->exec_list);
+		drm_gem_object_unreference(&obj->base);
+	}
 	kfree(eb);
 }
 
@@ -328,8 +338,7 @@ i915_gem_execbuffer_relocate_object_slow(struct drm_i915_gem_object *obj,
 
 static int
 i915_gem_execbuffer_relocate(struct drm_device *dev,
-			     struct eb_objects *eb,
-			     struct list_head *objects)
+			     struct eb_objects *eb)
 {
 	struct drm_i915_gem_object *obj;
 	int ret = 0;
@@ -342,7 +351,7 @@ i915_gem_execbuffer_relocate(struct drm_device *dev,
 	 * lockdep complains vehemently.
 	 */
 	pagefault_disable();
-	list_for_each_entry(obj, objects, exec_list) {
+	list_for_each_entry(obj, &eb->objects, exec_list) {
 		ret = i915_gem_execbuffer_relocate_object(obj, eb);
 		if (ret)
 			break;
@@ -531,7 +540,6 @@ static int
 i915_gem_execbuffer_relocate_slow(struct drm_device *dev,
 				  struct drm_file *file,
 				  struct intel_ring_buffer *ring,
-				  struct list_head *objects,
 				  struct eb_objects *eb,
 				  struct drm_i915_gem_exec_object2 *exec,
 				  int count)
@@ -542,8 +550,8 @@ i915_gem_execbuffer_relocate_slow(struct drm_device *dev,
 	int i, total, ret;
 
 	/* We may process another execbuffer during the unlock... */
-	while (!list_empty(objects)) {
-		obj = list_first_entry(objects,
+	while (!list_empty(&eb->objects)) {
+		obj = list_first_entry(&eb->objects,
 				       struct drm_i915_gem_object,
 				       exec_list);
 		list_del_init(&obj->exec_list);
@@ -590,15 +598,15 @@ i915_gem_execbuffer_relocate_slow(struct drm_device *dev,
 
 	/* reacquire the objects */
 	eb_reset(eb);
-	ret = eb_lookup_objects(eb, exec, count, file, objects);
+	ret = eb_lookup_objects(eb, exec, count, file);
 	if (ret)
 		goto err;
 
-	ret = i915_gem_execbuffer_reserve(ring, file, objects);
+	ret = i915_gem_execbuffer_reserve(ring, file, &eb->objects);
 	if (ret)
 		goto err;
 
-	list_for_each_entry(obj, objects, exec_list) {
+	list_for_each_entry(obj, &eb->objects, exec_list) {
 		int offset = obj->exec_entry - exec;
 		ret = i915_gem_execbuffer_relocate_object_slow(obj, eb,
 							       reloc + reloc_offset[offset]);
@@ -756,7 +764,6 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 		       struct drm_i915_gem_exec_object2 *exec)
 {
 	drm_i915_private_t *dev_priv = dev->dev_private;
-	struct list_head objects;
 	struct eb_objects *eb;
 	struct drm_i915_gem_object *batch_obj;
 	struct drm_clip_rect *cliprects = NULL;
@@ -899,28 +906,26 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 	}
 
 	/* Look up object handles */
-	INIT_LIST_HEAD(&objects);
-	ret = eb_lookup_objects(eb, exec, args->buffer_count, file, &objects);
+	ret = eb_lookup_objects(eb, exec, args->buffer_count, file);
 	if (ret)
 		goto err;
 
 	/* take note of the batch buffer before we might reorder the lists */
-	batch_obj = list_entry(objects.prev,
+	batch_obj = list_entry(eb->objects.prev,
 			       struct drm_i915_gem_object,
 			       exec_list);
 
 	/* Move the objects en-masse into the GTT, evicting if necessary. */
-	ret = i915_gem_execbuffer_reserve(ring, file, &objects);
+	ret = i915_gem_execbuffer_reserve(ring, file, &eb->objects);
 	if (ret)
 		goto err;
 
 	/* The objects are in their final locations, apply the relocations. */
-	ret = i915_gem_execbuffer_relocate(dev, eb, &objects);
+	ret = i915_gem_execbuffer_relocate(dev, eb);
 	if (ret) {
 		if (ret == -EFAULT) {
 			ret = i915_gem_execbuffer_relocate_slow(dev, file, ring,
-								&objects, eb,
-								exec,
+								eb, exec,
 								args->buffer_count);
 			BUG_ON(!mutex_is_locked(&dev->struct_mutex));
 		}
@@ -943,7 +948,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 	if (flags & I915_DISPATCH_SECURE && !batch_obj->has_global_gtt_mapping)
 		i915_gem_gtt_bind_object(batch_obj, batch_obj->cache_level);
 
-	ret = i915_gem_execbuffer_move_to_gpu(ring, &objects);
+	ret = i915_gem_execbuffer_move_to_gpu(ring, &eb->objects);
 	if (ret)
 		goto err;
 
@@ -997,20 +1002,11 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 
 	trace_i915_gem_ring_dispatch(ring, intel_ring_get_seqno(ring), flags);
 
-	i915_gem_execbuffer_move_to_active(&objects, ring);
+	i915_gem_execbuffer_move_to_active(&eb->objects, ring);
 	i915_gem_execbuffer_retire_commands(dev, file, ring);
 
 err:
 	eb_destroy(eb);
-	while (!list_empty(&objects)) {
-		struct drm_i915_gem_object *obj;
-
-		obj = list_first_entry(&objects,
-				       struct drm_i915_gem_object,
-				       exec_list);
-		list_del_init(&obj->exec_list);
-		drm_gem_object_unreference(&obj->base);
-	}
 
 	mutex_unlock(&dev->struct_mutex);
 
-- 
cgit v0.10.2


From ed5982e6ce5f106abcbf071f80730db344a6da42 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 17 Jan 2013 22:23:36 +0100
Subject: drm/i915: Allow userspace to hint that the relocations were known

Userspace is able to hint to the kernel that its command stream and
auxiliary state buffers already hold the correct presumed addresses and
so the relocation process may be skipped if the kernel does not need to
move any buffers in preparation for the execbuffer. Thus for the common
case where the allotment of buffers is static between batches, we can
avoid the overhead of individually checking the relocation entries.

Note that this requires userspace to supply the domain tracking and
requests for workarounds itself that would otherwise be computed based
upon the relocation entries.

Using copywinwin10 as an example that is dependent upon emitting a lot
of relocations (2 per operation), we see improvements of:

c2d/gm45: 618000.0/sec to 632000.0/sec.
i3-330m: 748000.0/sec to 830000.0/sec.

(measured relative to a baseline with neither optimisations applied).

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Imre Deak <imre.deak@intel.com>
[danvet: Fixup merge conflict in userspace header due to different
baseline trees.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index 6d8a1dc..a6e047d 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -992,6 +992,9 @@ static int i915_getparam(struct drm_device *dev, void *data,
 	case I915_PARAM_HAS_PINNED_BATCHES:
 		value = 1;
 		break;
+	case I915_PARAM_HAS_EXEC_NO_RELOC:
+		value = 1;
+		break;
 	default:
 		DRM_DEBUG_DRIVER("Unknown parameter %d\n",
 				 param->param);
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 386677f..34f6cdf 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -373,7 +373,8 @@ need_reloc_mappable(struct drm_i915_gem_object *obj)
 
 static int
 i915_gem_execbuffer_reserve_object(struct drm_i915_gem_object *obj,
-				   struct intel_ring_buffer *ring)
+				   struct intel_ring_buffer *ring,
+				   bool *need_reloc)
 {
 	struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
 	struct drm_i915_gem_exec_object2 *entry = obj->exec_entry;
@@ -414,7 +415,20 @@ i915_gem_execbuffer_reserve_object(struct drm_i915_gem_object *obj,
 		obj->has_aliasing_ppgtt_mapping = 1;
 	}
 
-	entry->offset = obj->gtt_offset;
+	if (entry->offset != obj->gtt_offset) {
+		entry->offset = obj->gtt_offset;
+		*need_reloc = true;
+	}
+
+	if (entry->flags & EXEC_OBJECT_WRITE) {
+		obj->base.pending_read_domains = I915_GEM_DOMAIN_RENDER;
+		obj->base.pending_write_domain = I915_GEM_DOMAIN_RENDER;
+	}
+
+	if (entry->flags & EXEC_OBJECT_NEEDS_GTT &&
+	    !obj->has_global_gtt_mapping)
+		i915_gem_gtt_bind_object(obj, obj->cache_level);
+
 	return 0;
 }
 
@@ -440,7 +454,8 @@ i915_gem_execbuffer_unreserve_object(struct drm_i915_gem_object *obj)
 static int
 i915_gem_execbuffer_reserve(struct intel_ring_buffer *ring,
 			    struct drm_file *file,
-			    struct list_head *objects)
+			    struct list_head *objects,
+			    bool *need_relocs)
 {
 	struct drm_i915_gem_object *obj;
 	struct list_head ordered_objects;
@@ -468,7 +483,7 @@ i915_gem_execbuffer_reserve(struct intel_ring_buffer *ring,
 		else
 			list_move_tail(&obj->exec_list, &ordered_objects);
 
-		obj->base.pending_read_domains = 0;
+		obj->base.pending_read_domains = I915_GEM_GPU_DOMAINS & ~I915_GEM_DOMAIN_COMMAND;
 		obj->base.pending_write_domain = 0;
 		obj->pending_fenced_gpu_access = false;
 	}
@@ -508,7 +523,7 @@ i915_gem_execbuffer_reserve(struct intel_ring_buffer *ring,
 			    (need_mappable && !obj->map_and_fenceable))
 				ret = i915_gem_object_unbind(obj);
 			else
-				ret = i915_gem_execbuffer_reserve_object(obj, ring);
+				ret = i915_gem_execbuffer_reserve_object(obj, ring, need_relocs);
 			if (ret)
 				goto err;
 		}
@@ -518,7 +533,7 @@ i915_gem_execbuffer_reserve(struct intel_ring_buffer *ring,
 			if (obj->gtt_space)
 				continue;
 
-			ret = i915_gem_execbuffer_reserve_object(obj, ring);
+			ret = i915_gem_execbuffer_reserve_object(obj, ring, need_relocs);
 			if (ret)
 				goto err;
 		}
@@ -538,16 +553,18 @@ err:		/* Decrement pin count for bound objects */
 
 static int
 i915_gem_execbuffer_relocate_slow(struct drm_device *dev,
+				  struct drm_i915_gem_execbuffer2 *args,
 				  struct drm_file *file,
 				  struct intel_ring_buffer *ring,
 				  struct eb_objects *eb,
-				  struct drm_i915_gem_exec_object2 *exec,
-				  int count)
+				  struct drm_i915_gem_exec_object2 *exec)
 {
 	struct drm_i915_gem_relocation_entry *reloc;
 	struct drm_i915_gem_object *obj;
+	bool need_relocs;
 	int *reloc_offset;
 	int i, total, ret;
+	int count = args->buffer_count;
 
 	/* We may process another execbuffer during the unlock... */
 	while (!list_empty(&eb->objects)) {
@@ -602,7 +619,8 @@ i915_gem_execbuffer_relocate_slow(struct drm_device *dev,
 	if (ret)
 		goto err;
 
-	ret = i915_gem_execbuffer_reserve(ring, file, &eb->objects);
+	need_relocs = (args->flags & I915_EXEC_NO_RELOC) == 0;
+	ret = i915_gem_execbuffer_reserve(ring, file, &eb->objects, &need_relocs);
 	if (ret)
 		goto err;
 
@@ -660,6 +678,9 @@ i915_gem_execbuffer_move_to_gpu(struct intel_ring_buffer *ring,
 static bool
 i915_gem_check_execbuffer(struct drm_i915_gem_execbuffer2 *exec)
 {
+	if (exec->flags & __I915_EXEC_UNKNOWN_FLAGS)
+		return false;
+
 	return ((exec->batch_start_offset | exec->batch_len) & 0x7) == 0;
 }
 
@@ -673,6 +694,9 @@ validate_exec_list(struct drm_i915_gem_exec_object2 *exec,
 		char __user *ptr = (char __user *)(uintptr_t)exec[i].relocs_ptr;
 		int length; /* limited by fault_in_pages_readable() */
 
+		if (exec[i].flags & __EXEC_OBJECT_UNKNOWN_FLAGS)
+			return -EINVAL;
+
 		/* First check for malicious input causing overflow */
 		if (exec[i].relocation_count >
 		    INT_MAX / sizeof(struct drm_i915_gem_relocation_entry))
@@ -680,9 +704,6 @@ validate_exec_list(struct drm_i915_gem_exec_object2 *exec,
 
 		length = exec[i].relocation_count *
 			sizeof(struct drm_i915_gem_relocation_entry);
-		if (!access_ok(VERIFY_READ, ptr, length))
-			return -EFAULT;
-
 		/* we may also need to update the presumed offsets */
 		if (!access_ok(VERIFY_WRITE, ptr, length))
 			return -EFAULT;
@@ -704,8 +725,10 @@ i915_gem_execbuffer_move_to_active(struct list_head *objects,
 		u32 old_read = obj->base.read_domains;
 		u32 old_write = obj->base.write_domain;
 
-		obj->base.read_domains = obj->base.pending_read_domains;
 		obj->base.write_domain = obj->base.pending_write_domain;
+		if (obj->base.write_domain == 0)
+			obj->base.pending_read_domains |= obj->base.read_domains;
+		obj->base.read_domains = obj->base.pending_read_domains;
 		obj->fenced_gpu_access = obj->pending_fenced_gpu_access;
 
 		i915_gem_object_move_to_active(obj, ring);
@@ -770,14 +793,12 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 	struct intel_ring_buffer *ring;
 	u32 ctx_id = i915_execbuffer2_get_context_id(*args);
 	u32 exec_start, exec_len;
-	u32 mask;
-	u32 flags;
+	u32 mask, flags;
 	int ret, mode, i;
+	bool need_relocs;
 
-	if (!i915_gem_check_execbuffer(args)) {
-		DRM_DEBUG("execbuf with invalid offset/length\n");
+	if (!i915_gem_check_execbuffer(args))
 		return -EINVAL;
-	}
 
 	ret = validate_exec_list(exec, args->buffer_count);
 	if (ret)
@@ -916,17 +937,18 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 			       exec_list);
 
 	/* Move the objects en-masse into the GTT, evicting if necessary. */
-	ret = i915_gem_execbuffer_reserve(ring, file, &eb->objects);
+	need_relocs = (args->flags & I915_EXEC_NO_RELOC) == 0;
+	ret = i915_gem_execbuffer_reserve(ring, file, &eb->objects, &need_relocs);
 	if (ret)
 		goto err;
 
 	/* The objects are in their final locations, apply the relocations. */
-	ret = i915_gem_execbuffer_relocate(dev, eb);
+	if (need_relocs)
+		ret = i915_gem_execbuffer_relocate(dev, eb);
 	if (ret) {
 		if (ret == -EFAULT) {
-			ret = i915_gem_execbuffer_relocate_slow(dev, file, ring,
-								eb, exec,
-								args->buffer_count);
+			ret = i915_gem_execbuffer_relocate_slow(dev, args, file, ring,
+								eb, exec);
 			BUG_ON(!mutex_is_locked(&dev->struct_mutex));
 		}
 		if (ret)
diff --git a/include/uapi/drm/i915_drm.h b/include/uapi/drm/i915_drm.h
index c4d2e9c..2430b6a 100644
--- a/include/uapi/drm/i915_drm.h
+++ b/include/uapi/drm/i915_drm.h
@@ -308,6 +308,7 @@ typedef struct drm_i915_irq_wait {
 #define I915_PARAM_RSVD_FOR_FUTURE_USE	 22
 #define I915_PARAM_HAS_SECURE_BATCHES	 23
 #define I915_PARAM_HAS_PINNED_BATCHES	 24
+#define I915_PARAM_HAS_EXEC_NO_RELOC	 25
 
 typedef struct drm_i915_getparam {
 	int param;
@@ -628,7 +629,11 @@ struct drm_i915_gem_exec_object2 {
 	__u64 offset;
 
 #define EXEC_OBJECT_NEEDS_FENCE (1<<0)
+#define EXEC_OBJECT_NEEDS_GTT	(1<<1)
+#define EXEC_OBJECT_WRITE	(1<<2)
+#define __EXEC_OBJECT_UNKNOWN_FLAGS -(EXEC_OBJECT_WRITE<<1)
 	__u64 flags;
+
 	__u64 rsvd1;
 	__u64 rsvd2;
 };
@@ -687,6 +692,15 @@ struct drm_i915_gem_execbuffer2 {
  */
 #define I915_EXEC_IS_PINNED		(1<<10)
 
+/** Provide a hint to the kernel that the command stream and auxilliary
+ * state buffers already holds the correct presumed addresses and so the
+ * relocation process may be skipped if no buffers need to be moved in
+ * preparation for the execbuffer.
+ */
+#define I915_EXEC_NO_RELOC		(1<<11)
+
+#define __I915_EXEC_UNKNOWN_FLAGS -(I915_EXEC_NO_RELOC<<1)
+
 #define I915_EXEC_CONTEXT_ID_MASK	(0xffffffff)
 #define i915_execbuffer2_set_context_id(eb2, context) \
 	(eb2).rsvd1 = context & I915_EXEC_CONTEXT_ID_MASK
-- 
cgit v0.10.2


From eef90ccb8a4d50b219a95cc53878ebb007315b32 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Tue, 8 Jan 2013 10:53:17 +0000
Subject: drm/i915: Use the reloc.handle as an index into the execbuffer array

Using copywinwin10 as an example that is dependent upon emitting a lot
of relocations (2 per operation), we see improvements of:

c2d/gm45: 618000.0/sec to 623000.0/sec.
i3-330m: 748000.0/sec to 789000.0/sec.

(measured relative to a baseline with neither optimisations applied).

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index a6e047d..4421182 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -995,6 +995,9 @@ static int i915_getparam(struct drm_device *dev, void *data,
 	case I915_PARAM_HAS_EXEC_NO_RELOC:
 		value = 1;
 		break;
+	case I915_PARAM_HAS_EXEC_HANDLE_LUT:
+		value = 1;
+		break;
 	default:
 		DRM_DEBUG_DRIVER("Unknown parameter %d\n",
 				 param->param);
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 34f6cdf..f5a11ec 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -36,24 +36,40 @@
 struct eb_objects {
 	struct list_head objects;
 	int and;
-	struct hlist_head buckets[0];
+	union {
+		struct drm_i915_gem_object *lut[0];
+		struct hlist_head buckets[0];
+	};
 };
 
 static struct eb_objects *
-eb_create(int size)
+eb_create(struct drm_i915_gem_execbuffer2 *args)
 {
-	struct eb_objects *eb;
-	int count = PAGE_SIZE / sizeof(struct hlist_head) / 2;
-	BUILD_BUG_ON(!is_power_of_2(PAGE_SIZE / sizeof(struct hlist_head)));
-	while (count > size)
-		count >>= 1;
-	eb = kzalloc(count*sizeof(struct hlist_head) +
-		     sizeof(struct eb_objects),
-		     GFP_KERNEL);
-	if (eb == NULL)
-		return eb;
-
-	eb->and = count - 1;
+	struct eb_objects *eb = NULL;
+
+	if (args->flags & I915_EXEC_HANDLE_LUT) {
+		int size = args->buffer_count;
+		size *= sizeof(struct drm_i915_gem_object *);
+		size += sizeof(struct eb_objects);
+		eb = kmalloc(size, GFP_TEMPORARY | __GFP_NOWARN | __GFP_NORETRY);
+	}
+
+	if (eb == NULL) {
+		int size = args->buffer_count;
+		int count = PAGE_SIZE / sizeof(struct hlist_head) / 2;
+		BUILD_BUG_ON(!is_power_of_2(PAGE_SIZE / sizeof(struct hlist_head)));
+		while (count > 2*size)
+			count >>= 1;
+		eb = kzalloc(count*sizeof(struct hlist_head) +
+			     sizeof(struct eb_objects),
+			     GFP_TEMPORARY);
+		if (eb == NULL)
+			return eb;
+
+		eb->and = count - 1;
+	} else
+		eb->and = -args->buffer_count;
+
 	INIT_LIST_HEAD(&eb->objects);
 	return eb;
 }
@@ -61,26 +77,20 @@ eb_create(int size)
 static void
 eb_reset(struct eb_objects *eb)
 {
-	memset(eb->buckets, 0, (eb->and+1)*sizeof(struct hlist_head));
-}
-
-static void
-eb_add_object(struct eb_objects *eb, struct drm_i915_gem_object *obj)
-{
-	hlist_add_head(&obj->exec_node,
-		       &eb->buckets[obj->exec_handle & eb->and]);
+	if (eb->and >= 0)
+		memset(eb->buckets, 0, (eb->and+1)*sizeof(struct hlist_head));
 }
 
 static int
 eb_lookup_objects(struct eb_objects *eb,
 		  struct drm_i915_gem_exec_object2 *exec,
-		  int count,
+		  const struct drm_i915_gem_execbuffer2 *args,
 		  struct drm_file *file)
 {
 	int i;
 
 	spin_lock(&file->table_lock);
-	for (i = 0; i < count; i++) {
+	for (i = 0; i < args->buffer_count; i++) {
 		struct drm_i915_gem_object *obj;
 
 		obj = to_intel_bo(idr_find(&file->object_idr, exec[i].handle));
@@ -101,9 +111,15 @@ eb_lookup_objects(struct eb_objects *eb,
 		drm_gem_object_reference(&obj->base);
 		list_add_tail(&obj->exec_list, &eb->objects);
 
-		obj->exec_handle = exec[i].handle;
 		obj->exec_entry = &exec[i];
-		eb_add_object(eb, obj);
+		if (eb->and < 0) {
+			eb->lut[i] = obj;
+		} else {
+			uint32_t handle = args->flags & I915_EXEC_HANDLE_LUT ? i : exec[i].handle;
+			obj->exec_handle = handle;
+			hlist_add_head(&obj->exec_node,
+				       &eb->buckets[handle & eb->and]);
+		}
 	}
 	spin_unlock(&file->table_lock);
 
@@ -113,18 +129,24 @@ eb_lookup_objects(struct eb_objects *eb,
 static struct drm_i915_gem_object *
 eb_get_object(struct eb_objects *eb, unsigned long handle)
 {
-	struct hlist_head *head;
-	struct hlist_node *node;
-	struct drm_i915_gem_object *obj;
+	if (eb->and < 0) {
+		if (handle >= -eb->and)
+			return NULL;
+		return eb->lut[handle];
+	} else {
+		struct hlist_head *head;
+		struct hlist_node *node;
 
-	head = &eb->buckets[handle & eb->and];
-	hlist_for_each(node, head) {
-		obj = hlist_entry(node, struct drm_i915_gem_object, exec_node);
-		if (obj->exec_handle == handle)
-			return obj;
-	}
+		head = &eb->buckets[handle & eb->and];
+		hlist_for_each(node, head) {
+			struct drm_i915_gem_object *obj;
 
-	return NULL;
+			obj = hlist_entry(node, struct drm_i915_gem_object, exec_node);
+			if (obj->exec_handle == handle)
+				return obj;
+		}
+		return NULL;
+	}
 }
 
 static void
@@ -615,7 +637,7 @@ i915_gem_execbuffer_relocate_slow(struct drm_device *dev,
 
 	/* reacquire the objects */
 	eb_reset(eb);
-	ret = eb_lookup_objects(eb, exec, count, file);
+	ret = eb_lookup_objects(eb, exec, args, file);
 	if (ret)
 		goto err;
 
@@ -919,7 +941,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 		goto pre_mutex_err;
 	}
 
-	eb = eb_create(args->buffer_count);
+	eb = eb_create(args);
 	if (eb == NULL) {
 		mutex_unlock(&dev->struct_mutex);
 		ret = -ENOMEM;
@@ -927,7 +949,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 	}
 
 	/* Look up object handles */
-	ret = eb_lookup_objects(eb, exec, args->buffer_count, file);
+	ret = eb_lookup_objects(eb, exec, args, file);
 	if (ret)
 		goto err;
 
diff --git a/include/uapi/drm/i915_drm.h b/include/uapi/drm/i915_drm.h
index 2430b6a..07d5941 100644
--- a/include/uapi/drm/i915_drm.h
+++ b/include/uapi/drm/i915_drm.h
@@ -309,6 +309,7 @@ typedef struct drm_i915_irq_wait {
 #define I915_PARAM_HAS_SECURE_BATCHES	 23
 #define I915_PARAM_HAS_PINNED_BATCHES	 24
 #define I915_PARAM_HAS_EXEC_NO_RELOC	 25
+#define I915_PARAM_HAS_EXEC_HANDLE_LUT   26
 
 typedef struct drm_i915_getparam {
 	int param;
@@ -699,7 +700,12 @@ struct drm_i915_gem_execbuffer2 {
  */
 #define I915_EXEC_NO_RELOC		(1<<11)
 
-#define __I915_EXEC_UNKNOWN_FLAGS -(I915_EXEC_NO_RELOC<<1)
+/** Use the reloc.handle as an index into the exec object array rather
+ * than as the per-file handle.
+ */
+#define I915_EXEC_HANDLE_LUT		(1<<12)
+
+#define __I915_EXEC_UNKNOWN_FLAGS -(I915_EXEC_HANDLE_LUT<<1)
 
 #define I915_EXEC_CONTEXT_ID_MASK	(0xffffffff)
 #define i915_execbuffer2_set_context_id(eb2, context) \
-- 
cgit v0.10.2


From c70af1e4b60a1853feeaae2bf6e2aa98de68acd9 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Wed, 16 Jan 2013 19:59:03 +0200
Subject: drm/i915: Fix SPRITE0_FLIP_DONE_INT_EN_VLV and
 SPRITE0_FLIPDONE_INT_STATUS_VLV
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Fix up some copypaste errors in the PIPESTAT register for VLV.

SPRITE0_FLIP_DONE_INT_EN_VLV is bit 22, not bit 26.

SPRITE0_FLIPDONE_INT_STATUS_VLV is bit 14, not bit 15.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 3b039f4..4c33bd2 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -2672,7 +2672,7 @@
 #define   PIPE_VSYNC_INTERRUPT_ENABLE		(1UL<<25)
 #define   PIPE_DISPLAY_LINE_COMPARE_ENABLE	(1UL<<24)
 #define   PIPE_DPST_EVENT_ENABLE		(1UL<<23)
-#define   SPRITE0_FLIP_DONE_INT_EN_VLV		(1UL<<26)
+#define   SPRITE0_FLIP_DONE_INT_EN_VLV		(1UL<<22)
 #define   PIPE_LEGACY_BLC_EVENT_ENABLE		(1UL<<22)
 #define   PIPE_ODD_FIELD_INTERRUPT_ENABLE	(1UL<<21)
 #define   PIPE_EVEN_FIELD_INTERRUPT_ENABLE	(1UL<<20)
@@ -2682,7 +2682,7 @@
 #define   PIPEA_HBLANK_INT_EN_VLV		(1UL<<16)
 #define   PIPE_OVERLAY_UPDATED_ENABLE		(1UL<<16)
 #define   SPRITE1_FLIPDONE_INT_STATUS_VLV	(1UL<<15)
-#define   SPRITE0_FLIPDONE_INT_STATUS_VLV	(1UL<<15)
+#define   SPRITE0_FLIPDONE_INT_STATUS_VLV	(1UL<<14)
 #define   PIPE_CRC_ERROR_INTERRUPT_STATUS	(1UL<<13)
 #define   PIPE_CRC_DONE_INTERRUPT_STATUS	(1UL<<12)
 #define   PIPE_GMBUS_INTERRUPT_STATUS		(1UL<<11)
-- 
cgit v0.10.2


From c1fc6521efbed7d3aca4e37ac61b23300f161cd1 Mon Sep 17 00:00:00 2001
From: Ben Widawsky <ben@bwidawsk.net>
Date: Thu, 17 Jan 2013 12:45:12 -0800
Subject: drm/i915: Kill gtt_end

It's duplicated in the more useful gtt_total.

Reviewed-by: Rodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 8f5b281..d58189e 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -801,7 +801,6 @@ typedef struct drm_i915_private {
 		/** Usable portion of the GTT for GEM */
 		unsigned long gtt_start;
 		unsigned long gtt_mappable_end;
-		unsigned long gtt_end;
 		unsigned long stolen_base; /* limited to low memory (32-bit) */
 
 		/** "Graphics Stolen Memory" holds the global PTEs */
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index a4af0f7..3ffcf52 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -394,7 +394,7 @@ void i915_gem_restore_gtt_mappings(struct drm_device *dev)
 
 	/* First fill our portion of the GTT with scratch pages */
 	i915_ggtt_clear_range(dev, dev_priv->mm.gtt_start / PAGE_SIZE,
-			      (dev_priv->mm.gtt_end - dev_priv->mm.gtt_start) / PAGE_SIZE);
+			      dev_priv->mm.gtt_total / PAGE_SIZE);
 
 	list_for_each_entry(obj, &dev_priv->mm.bound_list, gtt_list) {
 		i915_gem_clflush_object(obj);
@@ -556,7 +556,6 @@ void i915_gem_setup_global_gtt(struct drm_device *dev,
 
 	dev_priv->mm.gtt_start = start;
 	dev_priv->mm.gtt_mappable_end = mappable_end;
-	dev_priv->mm.gtt_end = end;
 	dev_priv->mm.gtt_total = end - start;
 	dev_priv->mm.mappable_gtt_total = min(end, mappable_end) - start;
 
-- 
cgit v0.10.2


From 35451cb6fbd3655ede4694b11761a587d400d647 Mon Sep 17 00:00:00 2001
From: Ben Widawsky <ben@bwidawsk.net>
Date: Thu, 17 Jan 2013 12:45:13 -0800
Subject: drm/i915: Mappable_end can't ever be > end

Both DRI1 and DRI2 can never specify a mappable size which goes past the
GTT size.  Don't pretend otherwise.

Reviewed-by: Rodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 3ffcf52..e66c700 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -536,6 +536,8 @@ void i915_gem_setup_global_gtt(struct drm_device *dev,
 	struct drm_i915_gem_object *obj;
 	unsigned long hole_start, hole_end;
 
+	BUG_ON(mappable_end > end);
+
 	/* Subtract the guard page ... */
 	drm_mm_init(&dev_priv->mm.gtt_space, start, end - start - PAGE_SIZE);
 	if (!HAS_LLC(dev))
@@ -557,7 +559,7 @@ void i915_gem_setup_global_gtt(struct drm_device *dev,
 	dev_priv->mm.gtt_start = start;
 	dev_priv->mm.gtt_mappable_end = mappable_end;
 	dev_priv->mm.gtt_total = end - start;
-	dev_priv->mm.mappable_gtt_total = min(end, mappable_end) - start;
+	dev_priv->mm.mappable_gtt_total = mappable_end - start;
 
 	/* Clear any non-preallocated blocks */
 	drm_mm_for_each_hole(entry, &dev_priv->mm.gtt_space,
-- 
cgit v0.10.2


From 00fc2c3c53d7bfc9a29e5f4bdf2677f0c399f3bc Mon Sep 17 00:00:00 2001
From: Ben Widawsky <ben@bwidawsk.net>
Date: Thu, 17 Jan 2013 12:45:14 -0800
Subject: drm/i915: Remove gtt_mappable_total

With the assertion from the previous patch in place, it should be safe
to get rid gtt_mappable_total. Keeps things saner to not have to track
the same info in two places.

In order to keep the diff as simple as possible and keep with the
existing gtt_setup semantics we opt to keep gtt_mappable_end. It's not
as consistent with the 'total' used in the previous patch, but that can
be fixed later.

Reviewed-by: Rodrigo Vivi <rodrigo.vivi@gmail.com>
[Ben modified commit message]
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index f94418b..35d326d 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -259,7 +259,8 @@ static int i915_gem_object_info(struct seq_file *m, void* data)
 		   count, size);
 
 	seq_printf(m, "%zu [%zu] gtt total\n",
-		   dev_priv->mm.gtt_total, dev_priv->mm.mappable_gtt_total);
+		   dev_priv->mm.gtt_total,
+		   dev_priv->mm.gtt_mappable_end - dev_priv->mm.gtt_start);
 
 	mutex_unlock(&dev->struct_mutex);
 
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index d58189e..6e2c10b 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -886,7 +886,6 @@ typedef struct drm_i915_private {
 
 		/* accounting, useful for userland debugging */
 		size_t gtt_total;
-		size_t mappable_gtt_total;
 		size_t object_memory;
 		u32 object_count;
 	} mm;
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index e66c700..8c42ddd 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -559,7 +559,6 @@ void i915_gem_setup_global_gtt(struct drm_device *dev,
 	dev_priv->mm.gtt_start = start;
 	dev_priv->mm.gtt_mappable_end = mappable_end;
 	dev_priv->mm.gtt_total = end - start;
-	dev_priv->mm.mappable_gtt_total = mappable_end - start;
 
 	/* Clear any non-preallocated blocks */
 	drm_mm_for_each_hole(entry, &dev_priv->mm.gtt_space,
-- 
cgit v0.10.2


From 5d4545aef561ad47f91bcf75814af20c104b5a9e Mon Sep 17 00:00:00 2001
From: Ben Widawsky <ben@bwidawsk.net>
Date: Thu, 17 Jan 2013 12:45:15 -0800
Subject: drm/i915: Create a gtt structure

The purpose of the gtt structure is to help isolate our gtt specific
properties from the rest of the code (in doing so it help us finish the
isolation from the AGP connection).

The following members are pulled out (and renamed):
gtt_start
gtt_total
gtt_mappable_end
gtt_mappable
gtt_base_addr
gsm

The gtt structure will serve as a nice place to put gen specific gtt
routines in upcoming patches. As far as what else I feel belongs in this
structure: it is meant to encapsulate the GTT's physical properties.
This is why I've not added fields which track various drm_mm properties,
or things like gtt_mtrr (which is itself a pretty transient field).

Reviewed-by: Rodrigo Vivi <rodrigo.vivi@gmail.com>
[Ben modified commit messages]
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index 35d326d..773b23e 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -259,8 +259,8 @@ static int i915_gem_object_info(struct seq_file *m, void* data)
 		   count, size);
 
 	seq_printf(m, "%zu [%zu] gtt total\n",
-		   dev_priv->mm.gtt_total,
-		   dev_priv->mm.gtt_mappable_end - dev_priv->mm.gtt_start);
+		   dev_priv->gtt.total,
+		   dev_priv->gtt.mappable_end - dev_priv->gtt.start);
 
 	mutex_unlock(&dev->struct_mutex);
 
diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index 4421182..bb62213 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -1076,7 +1076,7 @@ static int i915_set_status_page(struct drm_device *dev, void *data,
 	ring->status_page.gfx_addr = hws->addr & (0x1ffff<<12);
 
 	dev_priv->dri1.gfx_hws_cpu_addr =
-		ioremap_wc(dev_priv->mm.gtt_base_addr + hws->addr, 4096);
+		ioremap_wc(dev_priv->gtt.mappable_base + hws->addr, 4096);
 	if (dev_priv->dri1.gfx_hws_cpu_addr == NULL) {
 		i915_dma_cleanup(dev);
 		ring->status_page.gfx_addr = 0;
@@ -1543,17 +1543,17 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 	}
 
 	aperture_size = dev_priv->mm.gtt->gtt_mappable_entries << PAGE_SHIFT;
-	dev_priv->mm.gtt_base_addr = dev_priv->mm.gtt->gma_bus_addr;
+	dev_priv->gtt.mappable_base = dev_priv->mm.gtt->gma_bus_addr;
 
-	dev_priv->mm.gtt_mapping =
-		io_mapping_create_wc(dev_priv->mm.gtt_base_addr,
+	dev_priv->gtt.mappable =
+		io_mapping_create_wc(dev_priv->gtt.mappable_base,
 				     aperture_size);
-	if (dev_priv->mm.gtt_mapping == NULL) {
+	if (dev_priv->gtt.mappable == NULL) {
 		ret = -EIO;
 		goto out_rmmap;
 	}
 
-	i915_mtrr_setup(dev_priv, dev_priv->mm.gtt_base_addr,
+	i915_mtrr_setup(dev_priv, dev_priv->gtt.mappable_base,
 			aperture_size);
 
 	/* The i915 workqueue is primarily used for batched retirement of
@@ -1658,11 +1658,11 @@ out_gem_unload:
 out_mtrrfree:
 	if (dev_priv->mm.gtt_mtrr >= 0) {
 		mtrr_del(dev_priv->mm.gtt_mtrr,
-			 dev_priv->mm.gtt_base_addr,
+			 dev_priv->gtt.mappable_base,
 			 aperture_size);
 		dev_priv->mm.gtt_mtrr = -1;
 	}
-	io_mapping_free(dev_priv->mm.gtt_mapping);
+	io_mapping_free(dev_priv->gtt.mappable);
 out_rmmap:
 	pci_iounmap(dev->pdev, dev_priv->regs);
 put_gmch:
@@ -1696,10 +1696,10 @@ int i915_driver_unload(struct drm_device *dev)
 	/* Cancel the retire work handler, which should be idle now. */
 	cancel_delayed_work_sync(&dev_priv->mm.retire_work);
 
-	io_mapping_free(dev_priv->mm.gtt_mapping);
+	io_mapping_free(dev_priv->gtt.mappable);
 	if (dev_priv->mm.gtt_mtrr >= 0) {
 		mtrr_del(dev_priv->mm.gtt_mtrr,
-			 dev_priv->mm.gtt_base_addr,
+			 dev_priv->gtt.mappable_base,
 			 dev_priv->mm.gtt->gtt_mappable_entries * PAGE_SIZE);
 		dev_priv->mm.gtt_mtrr = -1;
 	}
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 6e2c10b..f3f2e5e 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -364,6 +364,25 @@ struct intel_device_info {
 	u8 has_llc:1;
 };
 
+/* The Graphics Translation Table is the way in which GEN hardware translates a
+ * Graphics Virtual Address into a Physical Address. In addition to the normal
+ * collateral associated with any va->pa translations GEN hardware also has a
+ * portion of the GTT which can be mapped by the CPU and remain both coherent
+ * and correct (in cases like swizzling). That region is referred to as GMADR in
+ * the spec.
+ */
+struct i915_gtt {
+	unsigned long start;		/* Start offset of used GTT */
+	size_t total;			/* Total size GTT can map */
+
+	unsigned long mappable_end;	/* End offset that we can CPU map */
+	struct io_mapping *mappable;	/* Mapping to our CPU mappable region */
+	phys_addr_t mappable_base;	/* PA of our GMADR */
+
+	/** "Graphics Stolen Memory" holds the global PTEs */
+	void __iomem *gsm;
+};
+
 #define I915_PPGTT_PD_ENTRIES 512
 #define I915_PPGTT_PT_ENTRIES 1024
 struct i915_hw_ppgtt {
@@ -781,6 +800,8 @@ typedef struct drm_i915_private {
 	/* Register state */
 	bool modeset_on_lid;
 
+	struct i915_gtt gtt;
+
 	struct {
 		/** Bridge to intel-gtt-ko */
 		struct intel_gtt *gtt;
@@ -799,15 +820,8 @@ typedef struct drm_i915_private {
 		struct list_head unbound_list;
 
 		/** Usable portion of the GTT for GEM */
-		unsigned long gtt_start;
-		unsigned long gtt_mappable_end;
 		unsigned long stolen_base; /* limited to low memory (32-bit) */
 
-		/** "Graphics Stolen Memory" holds the global PTEs */
-		void __iomem *gsm;
-
-		struct io_mapping *gtt_mapping;
-		phys_addr_t gtt_base_addr;
 		int gtt_mtrr;
 
 		/** PPGTT used for aliasing the PPGTT with the GTT */
@@ -885,7 +899,6 @@ typedef struct drm_i915_private {
 		struct drm_i915_gem_phys_object *phys_objs[I915_MAX_PHYS_OBJECT];
 
 		/* accounting, useful for userland debugging */
-		size_t gtt_total;
 		size_t object_memory;
 		u32 object_count;
 	} mm;
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index a2bb189..51fdf16 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -186,7 +186,7 @@ i915_gem_get_aperture_ioctl(struct drm_device *dev, void *data,
 			pinned += obj->gtt_space->size;
 	mutex_unlock(&dev->struct_mutex);
 
-	args->aper_size = dev_priv->mm.gtt_total;
+	args->aper_size = dev_priv->gtt.total;
 	args->aper_available_size = args->aper_size - pinned;
 
 	return 0;
@@ -637,7 +637,7 @@ i915_gem_gtt_pwrite_fast(struct drm_device *dev,
 		 * source page isn't available.  Return the error and we'll
 		 * retry in the slow path.
 		 */
-		if (fast_user_write(dev_priv->mm.gtt_mapping, page_base,
+		if (fast_user_write(dev_priv->gtt.mappable, page_base,
 				    page_offset, user_data, page_length)) {
 			ret = -EFAULT;
 			goto out_unpin;
@@ -1362,7 +1362,7 @@ int i915_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 
 	obj->fault_mappable = true;
 
-	pfn = ((dev_priv->mm.gtt_base_addr + obj->gtt_offset) >> PAGE_SHIFT) +
+	pfn = ((dev_priv->gtt.mappable_base + obj->gtt_offset) >> PAGE_SHIFT) +
 		page_offset;
 
 	/* Finally, remap it using the new GTT offset */
@@ -1544,7 +1544,7 @@ i915_gem_mmap_gtt(struct drm_file *file,
 		goto unlock;
 	}
 
-	if (obj->base.size > dev_priv->mm.gtt_mappable_end) {
+	if (obj->base.size > dev_priv->gtt.mappable_end) {
 		ret = -E2BIG;
 		goto out;
 	}
@@ -2910,7 +2910,7 @@ i915_gem_object_bind_to_gtt(struct drm_i915_gem_object *obj,
 	 * before evicting everything in a vain attempt to find space.
 	 */
 	if (obj->base.size >
-	    (map_and_fenceable ? dev_priv->mm.gtt_mappable_end : dev_priv->mm.gtt_total)) {
+	    (map_and_fenceable ? dev_priv->gtt.mappable_end : dev_priv->gtt.total)) {
 		DRM_ERROR("Attempting to bind an object larger than the aperture\n");
 		return -E2BIG;
 	}
@@ -2931,7 +2931,7 @@ i915_gem_object_bind_to_gtt(struct drm_i915_gem_object *obj,
 	if (map_and_fenceable)
 		ret = drm_mm_insert_node_in_range_generic(&dev_priv->mm.gtt_space, node,
 							  size, alignment, obj->cache_level,
-							  0, dev_priv->mm.gtt_mappable_end);
+							  0, dev_priv->gtt.mappable_end);
 	else
 		ret = drm_mm_insert_node_generic(&dev_priv->mm.gtt_space, node,
 						 size, alignment, obj->cache_level);
@@ -2971,7 +2971,7 @@ i915_gem_object_bind_to_gtt(struct drm_i915_gem_object *obj,
 		(node->start & (fence_alignment - 1)) == 0;
 
 	mappable =
-		obj->gtt_offset + obj->base.size <= dev_priv->mm.gtt_mappable_end;
+		obj->gtt_offset + obj->base.size <= dev_priv->gtt.mappable_end;
 
 	obj->map_and_fenceable = mappable && fenceable;
 
diff --git a/drivers/gpu/drm/i915/i915_gem_evict.c b/drivers/gpu/drm/i915/i915_gem_evict.c
index 776a322..c86d5d9 100644
--- a/drivers/gpu/drm/i915/i915_gem_evict.c
+++ b/drivers/gpu/drm/i915/i915_gem_evict.c
@@ -80,7 +80,7 @@ i915_gem_evict_something(struct drm_device *dev, int min_size,
 	if (mappable)
 		drm_mm_init_scan_with_range(&dev_priv->mm.gtt_space,
 					    min_size, alignment, cache_level,
-					    0, dev_priv->mm.gtt_mappable_end);
+					    0, dev_priv->gtt.mappable_end);
 	else
 		drm_mm_init_scan(&dev_priv->mm.gtt_space,
 				 min_size, alignment, cache_level);
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index f5a11ec..2726910 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -281,7 +281,7 @@ i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj,
 
 		/* Map the page containing the relocation we're going to perform.  */
 		reloc->offset += obj->gtt_offset;
-		reloc_page = io_mapping_map_atomic_wc(dev_priv->mm.gtt_mapping,
+		reloc_page = io_mapping_map_atomic_wc(dev_priv->gtt.mappable,
 						      reloc->offset & PAGE_MASK);
 		reloc_entry = (uint32_t __iomem *)
 			(reloc_page + (reloc->offset & ~PAGE_MASK));
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 8c42ddd..61bfb12 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -290,7 +290,7 @@ void i915_gem_init_ppgtt(struct drm_device *dev)
 		return;
 
 
-	pd_addr = (gtt_pte_t __iomem*)dev_priv->mm.gsm + ppgtt->pd_offset/sizeof(gtt_pte_t);
+	pd_addr = (gtt_pte_t __iomem*)dev_priv->gtt.gsm + ppgtt->pd_offset/sizeof(gtt_pte_t);
 	for (i = 0; i < ppgtt->num_pd_entries; i++) {
 		dma_addr_t pt_addr;
 
@@ -367,7 +367,7 @@ static void i915_ggtt_clear_range(struct drm_device *dev,
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	gtt_pte_t scratch_pte;
-	gtt_pte_t __iomem *gtt_base = (gtt_pte_t __iomem *) dev_priv->mm.gsm + first_entry;
+	gtt_pte_t __iomem *gtt_base = (gtt_pte_t __iomem *) dev_priv->gtt.gsm + first_entry;
 	const int max_entries = dev_priv->mm.gtt->gtt_total_entries - first_entry;
 	int i;
 
@@ -393,8 +393,8 @@ void i915_gem_restore_gtt_mappings(struct drm_device *dev)
 	struct drm_i915_gem_object *obj;
 
 	/* First fill our portion of the GTT with scratch pages */
-	i915_ggtt_clear_range(dev, dev_priv->mm.gtt_start / PAGE_SIZE,
-			      dev_priv->mm.gtt_total / PAGE_SIZE);
+	i915_ggtt_clear_range(dev, dev_priv->gtt.start / PAGE_SIZE,
+			      dev_priv->gtt.total / PAGE_SIZE);
 
 	list_for_each_entry(obj, &dev_priv->mm.bound_list, gtt_list) {
 		i915_gem_clflush_object(obj);
@@ -433,7 +433,7 @@ static void gen6_ggtt_bind_object(struct drm_i915_gem_object *obj,
 	const int first_entry = obj->gtt_space->start >> PAGE_SHIFT;
 	const int max_entries = dev_priv->mm.gtt->gtt_total_entries - first_entry;
 	gtt_pte_t __iomem *gtt_entries =
-		(gtt_pte_t __iomem *)dev_priv->mm.gsm + first_entry;
+		(gtt_pte_t __iomem *)dev_priv->gtt.gsm + first_entry;
 	int unused, i = 0;
 	unsigned int len, m = 0;
 	dma_addr_t addr;
@@ -556,9 +556,9 @@ void i915_gem_setup_global_gtt(struct drm_device *dev,
 		obj->has_global_gtt_mapping = 1;
 	}
 
-	dev_priv->mm.gtt_start = start;
-	dev_priv->mm.gtt_mappable_end = mappable_end;
-	dev_priv->mm.gtt_total = end - start;
+	dev_priv->gtt.start = start;
+	dev_priv->gtt.mappable_end = mappable_end;
+	dev_priv->gtt.total = end - start;
 
 	/* Clear any non-preallocated blocks */
 	drm_mm_for_each_hole(entry, &dev_priv->mm.gtt_space,
@@ -752,9 +752,9 @@ int i915_gem_gtt_init(struct drm_device *dev)
 		goto err_out;
 	}
 
-	dev_priv->mm.gsm = ioremap_wc(gtt_bus_addr,
-				      dev_priv->mm.gtt->gtt_total_entries * sizeof(gtt_pte_t));
-	if (!dev_priv->mm.gsm) {
+	dev_priv->gtt.gsm = ioremap_wc(gtt_bus_addr,
+				       dev_priv->mm.gtt->gtt_total_entries * sizeof(gtt_pte_t));
+	if (!dev_priv->gtt.gsm) {
 		DRM_ERROR("Failed to map the gtt page table\n");
 		teardown_scratch_page(dev);
 		ret = -ENOMEM;
@@ -778,7 +778,7 @@ err_out:
 void i915_gem_gtt_fini(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	iounmap(dev_priv->mm.gsm);
+	iounmap(dev_priv->gtt.gsm);
 	teardown_scratch_page(dev);
 	if (INTEL_INFO(dev)->gen < 6)
 		intel_gmch_remove();
diff --git a/drivers/gpu/drm/i915/i915_gem_tiling.c b/drivers/gpu/drm/i915/i915_gem_tiling.c
index e76f0d8..abcba2f 100644
--- a/drivers/gpu/drm/i915/i915_gem_tiling.c
+++ b/drivers/gpu/drm/i915/i915_gem_tiling.c
@@ -357,7 +357,7 @@ i915_gem_set_tiling(struct drm_device *dev, void *data,
 
 		obj->map_and_fenceable =
 			obj->gtt_space == NULL ||
-			(obj->gtt_offset + obj->base.size <= dev_priv->mm.gtt_mappable_end &&
+			(obj->gtt_offset + obj->base.size <= dev_priv->gtt.mappable_end &&
 			 i915_gem_object_fence_ok(obj, args->tiling_mode));
 
 		/* Rebind if we need a change of alignment */
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 2028131..cc49a6d 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -939,7 +939,7 @@ i915_error_object_create(struct drm_i915_private *dev_priv,
 			goto unwind;
 
 		local_irq_save(flags);
-		if (reloc_offset < dev_priv->mm.gtt_mappable_end &&
+		if (reloc_offset < dev_priv->gtt.mappable_end &&
 		    src->has_global_gtt_mapping) {
 			void __iomem *s;
 
@@ -948,7 +948,7 @@ i915_error_object_create(struct drm_i915_private *dev_priv,
 			 * captures what the GPU read.
 			 */
 
-			s = io_mapping_map_atomic_wc(dev_priv->mm.gtt_mapping,
+			s = io_mapping_map_atomic_wc(dev_priv->gtt.mappable,
 						     reloc_offset);
 			memcpy_fromio(d, s, PAGE_SIZE);
 			io_mapping_unmap_atomic(s);
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 40b6b5e9..e4c5067 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -8687,7 +8687,7 @@ void intel_modeset_init(struct drm_device *dev)
 		dev->mode_config.max_width = 8192;
 		dev->mode_config.max_height = 8192;
 	}
-	dev->mode_config.fb_base = dev_priv->mm.gtt_base_addr;
+	dev->mode_config.fb_base = dev_priv->gtt.mappable_base;
 
 	DRM_DEBUG_KMS("%d display pipe%s available.\n",
 		      dev_priv->num_pipe, dev_priv->num_pipe > 1 ? "s" : "");
diff --git a/drivers/gpu/drm/i915/intel_fb.c b/drivers/gpu/drm/i915/intel_fb.c
index 71d5580..ce02af8 100644
--- a/drivers/gpu/drm/i915/intel_fb.c
+++ b/drivers/gpu/drm/i915/intel_fb.c
@@ -142,7 +142,7 @@ static int intelfb_create(struct intel_fbdev *ifbdev,
 	info->fix.smem_len = size;
 
 	info->screen_base =
-		ioremap_wc(dev_priv->mm.gtt_base_addr + obj->gtt_offset,
+		ioremap_wc(dev_priv->gtt.mappable_base + obj->gtt_offset,
 			   size);
 	if (!info->screen_base) {
 		ret = -ENOSPC;
diff --git a/drivers/gpu/drm/i915/intel_overlay.c b/drivers/gpu/drm/i915/intel_overlay.c
index fabe0ac..ba978bf 100644
--- a/drivers/gpu/drm/i915/intel_overlay.c
+++ b/drivers/gpu/drm/i915/intel_overlay.c
@@ -195,7 +195,7 @@ intel_overlay_map_regs(struct intel_overlay *overlay)
 	if (OVERLAY_NEEDS_PHYSICAL(overlay->dev))
 		regs = (struct overlay_registers __iomem *)overlay->reg_bo->phys_obj->handle->vaddr;
 	else
-		regs = io_mapping_map_wc(dev_priv->mm.gtt_mapping,
+		regs = io_mapping_map_wc(dev_priv->gtt.mappable,
 					 overlay->reg_bo->gtt_offset);
 
 	return regs;
@@ -1434,7 +1434,7 @@ intel_overlay_map_regs_atomic(struct intel_overlay *overlay)
 		regs = (struct overlay_registers __iomem *)
 			overlay->reg_bo->phys_obj->handle->vaddr;
 	else
-		regs = io_mapping_map_atomic_wc(dev_priv->mm.gtt_mapping,
+		regs = io_mapping_map_atomic_wc(dev_priv->gtt.mappable,
 						overlay->reg_bo->gtt_offset);
 
 	return regs;
-- 
cgit v0.10.2


From dabb7a91ae6442d642a52e0de22d781a8ad2af12 Mon Sep 17 00:00:00 2001
From: Ben Widawsky <ben@bwidawsk.net>
Date: Thu, 17 Jan 2013 12:45:16 -0800
Subject: drm/i915: Remove use on gma_bus_addr on gen6+

We have enough info to not use the intel_gtt bridge stuff.

v2: Move setup of mappable_base above the legacy init stuff because we
still need that on older platforms. (Daniel)

v3: Remove the dev_priv hunk which was rebased in by accident

Reviewed-by: Rodrigo Vivi <rodrigo.vivi@gmail.com> (v2)
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index bb62213..468d2a0 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -1426,7 +1426,7 @@ static void i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
 	if (!ap)
 		return;
 
-	ap->ranges[0].base = dev_priv->mm.gtt->gma_bus_addr;
+	ap->ranges[0].base = dev_priv->gtt.mappable_base;
 	ap->ranges[0].size =
 		dev_priv->mm.gtt->gtt_mappable_entries << PAGE_SHIFT;
 	primary =
@@ -1543,7 +1543,6 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 	}
 
 	aperture_size = dev_priv->mm.gtt->gtt_mappable_entries << PAGE_SHIFT;
-	dev_priv->gtt.mappable_base = dev_priv->mm.gtt->gma_bus_addr;
 
 	dev_priv->gtt.mappable =
 		io_mapping_create_wc(dev_priv->gtt.mappable_base,
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 61bfb12..0b89305 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -691,6 +691,8 @@ int i915_gem_gtt_init(struct drm_device *dev)
 	u16 snb_gmch_ctl;
 	int ret;
 
+	dev_priv->gtt.mappable_base = pci_resource_start(dev->pdev, 2);
+
 	/* On modern platforms we need not worry ourself with the legacy
 	 * hostbridge query stuff. Skip it entirely
 	 */
@@ -723,7 +725,6 @@ int i915_gem_gtt_init(struct drm_device *dev)
 
 	/* For GEN6+ the PTEs for the ggtt live at 2MB + BAR0 */
 	gtt_bus_addr = pci_resource_start(dev->pdev, 0) + (2<<20);
-	dev_priv->mm.gtt->gma_bus_addr = pci_resource_start(dev->pdev, 2);
 
 	/* i9xx_setup */
 	pci_read_config_word(dev->pdev, SNB_GMCH_CTRL, &snb_gmch_ctl);
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index 59e0269..ce1d074 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -1203,7 +1203,7 @@ static int intel_init_ring_buffer(struct drm_device *dev,
 		goto err_unpin;
 
 	ring->virtual_start =
-		ioremap_wc(dev_priv->mm.gtt->gma_bus_addr + obj->gtt_offset,
+		ioremap_wc(dev_priv->gtt.mappable_base + obj->gtt_offset,
 			   ring->size);
 	if (ring->virtual_start == NULL) {
 		DRM_ERROR("Failed to map ringbuffer.\n");
-- 
cgit v0.10.2


From bac4b7c3b5c0660c08dc4949fe40e08e20364ee3 Mon Sep 17 00:00:00 2001
From: Carsten Emde <C.Emde@osadl.org>
Date: Thu, 19 Jul 2012 15:54:25 +0000
Subject: drm: Load EDID: Explain better how to write your own EDID firmware

A description was lacking how to write an EDID firmware file that
corresponds to a given X11 setting.

Signed-off-by: Carsten Emde <C.Emde@osadl.org>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/Documentation/EDID/HOWTO.txt b/Documentation/EDID/HOWTO.txt
index 75a9f2a..2d0a8f0 100644
--- a/Documentation/EDID/HOWTO.txt
+++ b/Documentation/EDID/HOWTO.txt
@@ -28,11 +28,30 @@ Makefile environment are given here.
 To create binary EDID and C source code files from the existing data
 material, simply type "make".
 
-If you want to create your own EDID file, copy the file 1024x768.S and
-replace the settings with your own data. The CRC value in the last line
+If you want to create your own EDID file, copy the file 1024x768.S,
+replace the settings with your own data and add a new target to the
+Makefile. Please note that the EDID data structure expects the timing
+values in a different way as compared to the standard X11 format.
+
+X11:
+HTimings:  hdisp hsyncstart hsyncend htotal
+VTimings:  vdisp vsyncstart vsyncend vtotal
+
+EDID:
+#define XPIX hdisp
+#define XBLANK htotal-hdisp
+#define XOFFSET hsyncstart-hdisp
+#define XPULSE hsyncend-hsyncstart
+
+#define YPIX vdisp
+#define YBLANK vtotal-vdisp
+#define YOFFSET (63+(vsyncstart-vdisp))
+#define YPULSE (63+(vsyncend-vsyncstart))
+
+The CRC value in the last line
   #define CRC 0x55
-is a bit tricky. After a first version of the binary data set is
-created, it must be be checked with the "edid-decode" utility which will
+also is a bit tricky. After a first version of the binary data set is
+created, it must be checked with the "edid-decode" utility which will
 most probably complain about a wrong CRC. Fortunately, the utility also
 displays the correct CRC which must then be inserted into the source
 file. After the make procedure is repeated, the EDID data set is ready
-- 
cgit v0.10.2


From 93d187993b783c68383a884091a600d9ad499ea6 Mon Sep 17 00:00:00 2001
From: Ben Widawsky <ben@bwidawsk.net>
Date: Thu, 17 Jan 2013 12:45:17 -0800
Subject: drm/i915: Remove use of gtt_mappable_entries

Mappable_end, ie. size is almost always what you want as opposed to the
number of entries. Since we already have that information, we can scrap
the number of entries and only calculate it when needed.

If gtt_start is !0, this will have slightly different behavior. This
difference can only occur in DRI1, and exists when we try to kick out
the firmware fb. The new code seems like a bugfix to me.

The other case where we've changed the behavior is during init we check
the mappable region against our current known upper and lower limits
(64MB, and 512MB). This now matches the comment, and makes things more
convenient after removing gtt_mappable_entries.

Also worth noting is the setting of mappable_end is taken out of setup
because we do it earlier now in the DRI2 case and therefore need to add
that tiny hunk to support the DRI1 IOCTL.

v2: Move up mappable end to before legacy AGP init

v3: Add the dev_priv inclusion here from previous rebase error in patch
5

Reviewed-by: Rodrigo Vivi <rodrigo.vivi@gmail.com> (v2)
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
[danvet: squash in fix for a printk format flag mismatch warning.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index 773b23e..90a6fc5 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -258,7 +258,7 @@ static int i915_gem_object_info(struct seq_file *m, void* data)
 	seq_printf(m, "%u fault mappable objects, %zu bytes\n",
 		   count, size);
 
-	seq_printf(m, "%zu [%zu] gtt total\n",
+	seq_printf(m, "%zu [%lu] gtt total\n",
 		   dev_priv->gtt.total,
 		   dev_priv->gtt.mappable_end - dev_priv->gtt.start);
 
diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index 468d2a0..3f70178 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -1427,8 +1427,8 @@ static void i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
 		return;
 
 	ap->ranges[0].base = dev_priv->gtt.mappable_base;
-	ap->ranges[0].size =
-		dev_priv->mm.gtt->gtt_mappable_entries << PAGE_SHIFT;
+	ap->ranges[0].size = dev_priv->gtt.mappable_end - dev_priv->gtt.start;
+
 	primary =
 		pdev->resource[PCI_ROM_RESOURCE].flags & IORESOURCE_ROM_SHADOW;
 
@@ -1542,7 +1542,7 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 		goto put_gmch;
 	}
 
-	aperture_size = dev_priv->mm.gtt->gtt_mappable_entries << PAGE_SHIFT;
+	aperture_size = dev_priv->gtt.mappable_end;
 
 	dev_priv->gtt.mappable =
 		io_mapping_create_wc(dev_priv->gtt.mappable_base,
@@ -1699,7 +1699,7 @@ int i915_driver_unload(struct drm_device *dev)
 	if (dev_priv->mm.gtt_mtrr >= 0) {
 		mtrr_del(dev_priv->mm.gtt_mtrr,
 			 dev_priv->gtt.mappable_base,
-			 dev_priv->mm.gtt->gtt_mappable_entries * PAGE_SIZE);
+			 dev_priv->gtt.mappable_end);
 		dev_priv->mm.gtt_mtrr = -1;
 	}
 
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 51fdf16..e4132ff 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -149,6 +149,7 @@ int
 i915_gem_init_ioctl(struct drm_device *dev, void *data,
 		    struct drm_file *file)
 {
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_i915_gem_init *args = data;
 
 	if (drm_core_check_feature(dev, DRIVER_MODESET))
@@ -165,6 +166,7 @@ i915_gem_init_ioctl(struct drm_device *dev, void *data,
 	mutex_lock(&dev->struct_mutex);
 	i915_gem_setup_global_gtt(dev, args->gtt_start, args->gtt_end,
 				  args->gtt_end);
+	dev_priv->gtt.mappable_end = args->gtt_end;
 	mutex_unlock(&dev->struct_mutex);
 
 	return 0;
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 0b89305..0f0db02 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -557,7 +557,6 @@ void i915_gem_setup_global_gtt(struct drm_device *dev,
 	}
 
 	dev_priv->gtt.start = start;
-	dev_priv->gtt.mappable_end = mappable_end;
 	dev_priv->gtt.total = end - start;
 
 	/* Clear any non-preallocated blocks */
@@ -596,7 +595,7 @@ void i915_gem_init_global_gtt(struct drm_device *dev)
 	int ret;
 
 	gtt_size = dev_priv->mm.gtt->gtt_total_entries << PAGE_SHIFT;
-	mappable_size = dev_priv->mm.gtt->gtt_mappable_entries << PAGE_SHIFT;
+	mappable_size = dev_priv->gtt.mappable_end;
 
 	if (intel_enable_ppgtt(dev) && HAS_ALIASING_PPGTT(dev)) {
 		/* PPGTT pdes are stolen from global gtt ptes, so shrink the
@@ -692,6 +691,7 @@ int i915_gem_gtt_init(struct drm_device *dev)
 	int ret;
 
 	dev_priv->gtt.mappable_base = pci_resource_start(dev->pdev, 2);
+	dev_priv->gtt.mappable_end = pci_resource_len(dev->pdev, 2);
 
 	/* On modern platforms we need not worry ourself with the legacy
 	 * hostbridge query stuff. Skip it entirely
@@ -735,14 +735,13 @@ int i915_gem_gtt_init(struct drm_device *dev)
 	else
 		dev_priv->mm.gtt->stolen_size = gen7_get_stolen_size(snb_gmch_ctl);
 
-	dev_priv->mm.gtt->gtt_mappable_entries = pci_resource_len(dev->pdev, 2) >> PAGE_SHIFT;
 	/* 64/512MB is the current min/max we actually know of, but this is just a
 	 * coarse sanity check.
 	 */
-	if ((dev_priv->mm.gtt->gtt_mappable_entries >> 8) < 64 ||
-	    dev_priv->mm.gtt->gtt_mappable_entries > dev_priv->mm.gtt->gtt_total_entries) {
-		DRM_ERROR("Unknown GMADR entries (%d)\n",
-			  dev_priv->mm.gtt->gtt_mappable_entries);
+	if ((dev_priv->gtt.mappable_end < (64<<20) ||
+	    (dev_priv->gtt.mappable_end > (512<<20)))) {
+		DRM_ERROR("Unknown GMADR size (%lx)\n",
+			  dev_priv->gtt.mappable_end);
 		ret = -ENXIO;
 		goto err_out;
 	}
@@ -764,7 +763,7 @@ int i915_gem_gtt_init(struct drm_device *dev)
 
 	/* GMADR is the PCI aperture used by SW to access tiled GFX surfaces in a linear fashion. */
 	DRM_INFO("Memory usable by graphics device = %dM\n", dev_priv->mm.gtt->gtt_total_entries >> 8);
-	DRM_DEBUG_DRIVER("GMADR size = %dM\n", dev_priv->mm.gtt->gtt_mappable_entries >> 8);
+	DRM_DEBUG_DRIVER("GMADR size = %ldM\n", dev_priv->gtt.mappable_end >> 20);
 	DRM_DEBUG_DRIVER("GTT stolen size = %dM\n", dev_priv->mm.gtt->stolen_size >> 20);
 
 	return 0;
diff --git a/drivers/gpu/drm/i915/intel_fb.c b/drivers/gpu/drm/i915/intel_fb.c
index ce02af8..ce5f544 100644
--- a/drivers/gpu/drm/i915/intel_fb.c
+++ b/drivers/gpu/drm/i915/intel_fb.c
@@ -135,8 +135,7 @@ static int intelfb_create(struct intel_fbdev *ifbdev,
 		goto out_unpin;
 	}
 	info->apertures->ranges[0].base = dev->mode_config.fb_base;
-	info->apertures->ranges[0].size =
-		dev_priv->mm.gtt->gtt_mappable_entries << PAGE_SHIFT;
+	info->apertures->ranges[0].size = dev_priv->gtt.mappable_end;
 
 	info->fix.smem_start = dev->mode_config.fb_base + obj->gtt_offset;
 	info->fix.smem_len = size;
-- 
cgit v0.10.2


From 3685a8f38f2c54bb6058c0e66ae0562f8ab84e66 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 17 Jan 2013 16:31:28 +0200
Subject: drm/i915: Fix RGB color range property for PCH platforms
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The RGB color range select bit on the DP/SDVO/HDMI registers
disappeared when PCH was introduced, and instead a new PIPECONF bit
was added that performs the same function.

Add a new INTEL_MODE_LIMITED_COLOR_RANGE private mode flag, and set
it in the encoder mode_fixup if limited color range is requested.
Set the the PIPECONF bit 13 based on the flag.

Experimentation showed that simply toggling the bit while the pipe is
active doesn't work. We need to restart the pipe, which luckily already
happens.

The DP/SDVO/HDMI bit 8 is marked MBZ in the docs, so avoid setting it,
although it doesn't seem to do any harm in practice.

TODO:
- the PIPECONF bit too seems to have disappeared from HSW. Need a
  volunteer to test if it's just a documentation issue or if it's really
  gone. If the bit is gone and no easy replacement is found, then I suppose
  we may need to use the pipe CSC unit to perform the range compression.

v2: Use mode private_flags instead of intel_encoder virtual functions
v3: Moved the intel_dp color_range handling after bpc check to help
    later patches

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=46800
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 4c33bd2..2521617 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -2650,6 +2650,7 @@
 #define   PIPECONF_INTERLACED_DBL_ILK		(4 << 21) /* ilk/snb only */
 #define   PIPECONF_PFIT_PF_INTERLACED_DBL_ILK	(5 << 21) /* ilk/snb only */
 #define   PIPECONF_CXSR_DOWNCLOCK	(1<<16)
+#define   PIPECONF_COLOR_RANGE_SELECT	(1 << 13)
 #define   PIPECONF_BPC_MASK	(0x7 << 5)
 #define   PIPECONF_8BPC		(0<<5)
 #define   PIPECONF_10BPC	(1<<5)
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index e4c5067..b35902e 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -5096,6 +5096,11 @@ static void ironlake_set_pipeconf(struct drm_crtc *crtc,
 	else
 		val |= PIPECONF_PROGRESSIVE;
 
+	if (adjusted_mode->private_flags & INTEL_MODE_LIMITED_COLOR_RANGE)
+		val |= PIPECONF_COLOR_RANGE_SELECT;
+	else
+		val &= ~PIPECONF_COLOR_RANGE_SELECT;
+
 	I915_WRITE(PIPECONF(pipe), val);
 	POSTING_READ(PIPECONF(pipe));
 }
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 5f12eb2..d995627 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -763,6 +763,10 @@ intel_dp_mode_fixup(struct drm_encoder *encoder,
 		return false;
 
 	bpp = adjusted_mode->private_flags & INTEL_MODE_DP_FORCE_6BPC ? 18 : 24;
+
+	if (intel_dp->color_range)
+		adjusted_mode->private_flags |= INTEL_MODE_LIMITED_COLOR_RANGE;
+
 	mode_rate = intel_dp_link_required(adjusted_mode->clock, bpp);
 
 	for (clock = 0; clock <= max_clock; clock++) {
@@ -967,7 +971,8 @@ intel_dp_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
 		else
 			intel_dp->DP |= DP_PLL_FREQ_270MHZ;
 	} else if (!HAS_PCH_CPT(dev) || is_cpu_edp(intel_dp)) {
-		intel_dp->DP |= intel_dp->color_range;
+		if (!HAS_PCH_SPLIT(dev))
+			intel_dp->DP |= intel_dp->color_range;
 
 		if (adjusted_mode->flags & DRM_MODE_FLAG_PHSYNC)
 			intel_dp->DP |= DP_SYNC_HS_HIGH;
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 54a034c..4df47be 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -109,6 +109,11 @@
  * timings in the mode to prevent the crtc fixup from overwriting them.
  * Currently only lvds needs that. */
 #define INTEL_MODE_CRTC_TIMINGS_SET (0x20)
+/*
+ * Set when limited 16-235 (as opposed to full 0-255) RGB color range is
+ * to be used.
+ */
+#define INTEL_MODE_LIMITED_COLOR_RANGE (0x40)
 
 static inline void
 intel_mode_set_pixel_multiplier(struct drm_display_mode *mode,
diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c
index 6387f9b..f194d75 100644
--- a/drivers/gpu/drm/i915/intel_hdmi.c
+++ b/drivers/gpu/drm/i915/intel_hdmi.c
@@ -766,6 +766,11 @@ bool intel_hdmi_mode_fixup(struct drm_encoder *encoder,
 			   const struct drm_display_mode *mode,
 			   struct drm_display_mode *adjusted_mode)
 {
+	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
+
+	if (intel_hdmi->color_range)
+		adjusted_mode->private_flags |= INTEL_MODE_LIMITED_COLOR_RANGE;
+
 	return true;
 }
 
diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c
index 153377b..3b8491a 100644
--- a/drivers/gpu/drm/i915/intel_sdvo.c
+++ b/drivers/gpu/drm/i915/intel_sdvo.c
@@ -1064,6 +1064,9 @@ static bool intel_sdvo_mode_fixup(struct drm_encoder *encoder,
 	multiplier = intel_sdvo_get_pixel_multiplier(adjusted_mode);
 	intel_mode_set_pixel_multiplier(adjusted_mode, multiplier);
 
+	if (intel_sdvo->color_range)
+		adjusted_mode->private_flags |= INTEL_MODE_LIMITED_COLOR_RANGE;
+
 	return true;
 }
 
@@ -1153,7 +1156,7 @@ static void intel_sdvo_mode_set(struct drm_encoder *encoder,
 		/* The real mode polarity is set by the SDVO commands, using
 		 * struct intel_sdvo_dtd. */
 		sdvox = SDVO_VSYNC_ACTIVE_HIGH | SDVO_HSYNC_ACTIVE_HIGH;
-		if (intel_sdvo->is_hdmi)
+		if (!HAS_PCH_SPLIT(dev) && intel_sdvo->is_hdmi)
 			sdvox |= intel_sdvo->color_range;
 		if (INTEL_INFO(dev)->gen < 5)
 			sdvox |= SDVO_BORDER_ENABLE;
-- 
cgit v0.10.2


From 55bc60db5988c8366751d3d04dd690698a53412c Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 17 Jan 2013 16:31:29 +0200
Subject: drm/i915: Add "Automatic" mode for the "Broadcast RGB" property
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Add a new "Automatic" mode to the "Broadcast RGB" range property.
When selected the driver automagically selects between full range and
limited range output.

Based on CEA-861 [1] guidelines, limited range output is selected if the
mode is a CEA mode, except 640x480. Otherwise full range output is used.
Additionally DVI monitors should most likely default to full range
always.

As per DP1.2a [2] DisplayPort should always use full range for 18bpp, and
otherwise will follow CEA-861 rules.

NOTE: The default value for the property will now be "Automatic"
so some people may be affected in case they're relying on the
current full range default.

[1] CEA-861-E - 5.1 Default Encoding Parameters
[2] VESA DisplayPort Ver.1.2a - 5.1.1.1 Video Colorimetry

v2: Use has_hdmi_sink to check if a HDMI monitor is present
v3: Add information about relevant spec chapters

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index f3f2e5e..0a2a18b8 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -1811,5 +1811,9 @@ __i915_write(64, q)
 #define POSTING_READ(reg)	(void)I915_READ_NOTRACE(reg)
 #define POSTING_READ16(reg)	(void)I915_READ16_NOTRACE(reg)
 
+/* "Broadcast RGB" property */
+#define INTEL_BROADCAST_RGB_AUTO 0
+#define INTEL_BROADCAST_RGB_FULL 1
+#define INTEL_BROADCAST_RGB_LIMITED 2
 
 #endif
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index d995627..1492706 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -764,6 +764,18 @@ intel_dp_mode_fixup(struct drm_encoder *encoder,
 
 	bpp = adjusted_mode->private_flags & INTEL_MODE_DP_FORCE_6BPC ? 18 : 24;
 
+	if (intel_dp->color_range_auto) {
+		/*
+		 * See:
+		 * CEA-861-E - 5.1 Default Encoding Parameters
+		 * VESA DisplayPort Ver.1.2a - 5.1.1.1 Video Colorimetry
+		 */
+		if (bpp != 18 && drm_mode_cea_vic(adjusted_mode) > 1)
+			intel_dp->color_range = DP_COLOR_RANGE_16_235;
+		else
+			intel_dp->color_range = 0;
+	}
+
 	if (intel_dp->color_range)
 		adjusted_mode->private_flags |= INTEL_MODE_LIMITED_COLOR_RANGE;
 
@@ -2462,10 +2474,21 @@ intel_dp_set_property(struct drm_connector *connector,
 	}
 
 	if (property == dev_priv->broadcast_rgb_property) {
-		if (val == !!intel_dp->color_range)
-			return 0;
-
-		intel_dp->color_range = val ? DP_COLOR_RANGE_16_235 : 0;
+		switch (val) {
+		case INTEL_BROADCAST_RGB_AUTO:
+			intel_dp->color_range_auto = true;
+			break;
+		case INTEL_BROADCAST_RGB_FULL:
+			intel_dp->color_range_auto = false;
+			intel_dp->color_range = 0;
+			break;
+		case INTEL_BROADCAST_RGB_LIMITED:
+			intel_dp->color_range_auto = false;
+			intel_dp->color_range = DP_COLOR_RANGE_16_235;
+			break;
+		default:
+			return -EINVAL;
+		}
 		goto done;
 	}
 
@@ -2606,6 +2629,7 @@ intel_dp_add_properties(struct intel_dp *intel_dp, struct drm_connector *connect
 
 	intel_attach_force_audio_property(connector);
 	intel_attach_broadcast_rgb_property(connector);
+	intel_dp->color_range_auto = true;
 
 	if (is_edp(intel_dp)) {
 		drm_mode_create_scaling_mode_property(connector->dev);
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 4df47be..1a698c6 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -343,6 +343,7 @@ struct intel_hdmi {
 	u32 sdvox_reg;
 	int ddc_bus;
 	uint32_t color_range;
+	bool color_range_auto;
 	bool has_hdmi_sink;
 	bool has_audio;
 	enum hdmi_force_audio force_audio;
@@ -362,6 +363,7 @@ struct intel_dp {
 	bool has_audio;
 	enum hdmi_force_audio force_audio;
 	uint32_t color_range;
+	bool color_range_auto;
 	uint8_t link_bw;
 	uint8_t lane_count;
 	uint8_t dpcd[DP_RECEIVER_CAP_SIZE];
diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c
index f194d75..db67be6 100644
--- a/drivers/gpu/drm/i915/intel_hdmi.c
+++ b/drivers/gpu/drm/i915/intel_hdmi.c
@@ -768,6 +768,15 @@ bool intel_hdmi_mode_fixup(struct drm_encoder *encoder,
 {
 	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
 
+	if (intel_hdmi->color_range_auto) {
+		/* See CEA-861-E - 5.1 Default Encoding Parameters */
+		if (intel_hdmi->has_hdmi_sink &&
+		    drm_mode_cea_vic(adjusted_mode) > 1)
+			intel_hdmi->color_range = SDVO_COLOR_RANGE_16_235;
+		else
+			intel_hdmi->color_range = 0;
+	}
+
 	if (intel_hdmi->color_range)
 		adjusted_mode->private_flags |= INTEL_MODE_LIMITED_COLOR_RANGE;
 
@@ -912,10 +921,21 @@ intel_hdmi_set_property(struct drm_connector *connector,
 	}
 
 	if (property == dev_priv->broadcast_rgb_property) {
-		if (val == !!intel_hdmi->color_range)
-			return 0;
-
-		intel_hdmi->color_range = val ? SDVO_COLOR_RANGE_16_235 : 0;
+		switch (val) {
+		case INTEL_BROADCAST_RGB_AUTO:
+			intel_hdmi->color_range_auto = true;
+			break;
+		case INTEL_BROADCAST_RGB_FULL:
+			intel_hdmi->color_range_auto = false;
+			intel_hdmi->color_range = 0;
+			break;
+		case INTEL_BROADCAST_RGB_LIMITED:
+			intel_hdmi->color_range_auto = false;
+			intel_hdmi->color_range = SDVO_COLOR_RANGE_16_235;
+			break;
+		default:
+			return -EINVAL;
+		}
 		goto done;
 	}
 
@@ -964,6 +984,7 @@ intel_hdmi_add_properties(struct intel_hdmi *intel_hdmi, struct drm_connector *c
 {
 	intel_attach_force_audio_property(connector);
 	intel_attach_broadcast_rgb_property(connector);
+	intel_hdmi->color_range_auto = true;
 }
 
 void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
diff --git a/drivers/gpu/drm/i915/intel_modes.c b/drivers/gpu/drm/i915/intel_modes.c
index 49249bb..0e860f3 100644
--- a/drivers/gpu/drm/i915/intel_modes.c
+++ b/drivers/gpu/drm/i915/intel_modes.c
@@ -100,8 +100,9 @@ intel_attach_force_audio_property(struct drm_connector *connector)
 }
 
 static const struct drm_prop_enum_list broadcast_rgb_names[] = {
-	{ 0, "Full" },
-	{ 1, "Limited 16:235" },
+	{ INTEL_BROADCAST_RGB_AUTO, "Automatic" },
+	{ INTEL_BROADCAST_RGB_FULL, "Full" },
+	{ INTEL_BROADCAST_RGB_LIMITED, "Limited 16:235" },
 };
 
 void
diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c
index 3b8491a..3e34a35 100644
--- a/drivers/gpu/drm/i915/intel_sdvo.c
+++ b/drivers/gpu/drm/i915/intel_sdvo.c
@@ -103,6 +103,7 @@ struct intel_sdvo {
 	 * It is only valid when using TMDS encoding and 8 bit per color mode.
 	 */
 	uint32_t color_range;
+	bool color_range_auto;
 
 	/**
 	 * This is set if we're going to treat the device as TV-out.
@@ -1064,6 +1065,15 @@ static bool intel_sdvo_mode_fixup(struct drm_encoder *encoder,
 	multiplier = intel_sdvo_get_pixel_multiplier(adjusted_mode);
 	intel_mode_set_pixel_multiplier(adjusted_mode, multiplier);
 
+	if (intel_sdvo->color_range_auto) {
+		/* See CEA-861-E - 5.1 Default Encoding Parameters */
+		if (intel_sdvo->has_hdmi_monitor &&
+		    drm_mode_cea_vic(adjusted_mode) > 1)
+			intel_sdvo->color_range = SDVO_COLOR_RANGE_16_235;
+		else
+			intel_sdvo->color_range = 0;
+	}
+
 	if (intel_sdvo->color_range)
 		adjusted_mode->private_flags |= INTEL_MODE_LIMITED_COLOR_RANGE;
 
@@ -1900,10 +1910,21 @@ intel_sdvo_set_property(struct drm_connector *connector,
 	}
 
 	if (property == dev_priv->broadcast_rgb_property) {
-		if (val == !!intel_sdvo->color_range)
-			return 0;
-
-		intel_sdvo->color_range = val ? SDVO_COLOR_RANGE_16_235 : 0;
+		switch (val) {
+		case INTEL_BROADCAST_RGB_AUTO:
+			intel_sdvo->color_range_auto = true;
+			break;
+		case INTEL_BROADCAST_RGB_FULL:
+			intel_sdvo->color_range_auto = false;
+			intel_sdvo->color_range = 0;
+			break;
+		case INTEL_BROADCAST_RGB_LIMITED:
+			intel_sdvo->color_range_auto = false;
+			intel_sdvo->color_range = SDVO_COLOR_RANGE_16_235;
+			break;
+		default:
+			return -EINVAL;
+		}
 		goto done;
 	}
 
@@ -2200,13 +2221,16 @@ intel_sdvo_connector_init(struct intel_sdvo_connector *connector,
 }
 
 static void
-intel_sdvo_add_hdmi_properties(struct intel_sdvo_connector *connector)
+intel_sdvo_add_hdmi_properties(struct intel_sdvo *intel_sdvo,
+			       struct intel_sdvo_connector *connector)
 {
 	struct drm_device *dev = connector->base.base.dev;
 
 	intel_attach_force_audio_property(&connector->base.base);
-	if (INTEL_INFO(dev)->gen >= 4 && IS_MOBILE(dev))
+	if (INTEL_INFO(dev)->gen >= 4 && IS_MOBILE(dev)) {
 		intel_attach_broadcast_rgb_property(&connector->base.base);
+		intel_sdvo->color_range_auto = true;
+	}
 }
 
 static bool
@@ -2254,7 +2278,7 @@ intel_sdvo_dvi_init(struct intel_sdvo *intel_sdvo, int device)
 
 	intel_sdvo_connector_init(intel_sdvo_connector, intel_sdvo);
 	if (intel_sdvo->is_hdmi)
-		intel_sdvo_add_hdmi_properties(intel_sdvo_connector);
+		intel_sdvo_add_hdmi_properties(intel_sdvo, intel_sdvo_connector);
 
 	return true;
 }
-- 
cgit v0.10.2


From b1edd6a6ecd436af33f210a5c75e0249466fd200 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 17 Jan 2013 16:31:30 +0200
Subject: drm/edid: Add drm_rgb_quant_range_selectable()
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

drm_rgb_quant_range_selectable() will report whether the monitor
claims to support for RGB quantization range selection.

The information can be found in the CEA Video capability block.

v2: s/quantzation/quantization/ in the comment

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Acked-by: David Airlie <airlied@linux.ie>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
index 5a3770f..a3a3b61 100644
--- a/drivers/gpu/drm/drm_edid.c
+++ b/drivers/gpu/drm/drm_edid.c
@@ -1483,9 +1483,11 @@ add_detailed_modes(struct drm_connector *connector, struct edid *edid,
 #define VIDEO_BLOCK     0x02
 #define VENDOR_BLOCK    0x03
 #define SPEAKER_BLOCK	0x04
+#define VIDEO_CAPABILITY_BLOCK	0x07
 #define EDID_BASIC_AUDIO	(1 << 6)
 #define EDID_CEA_YCRCB444	(1 << 5)
 #define EDID_CEA_YCRCB422	(1 << 4)
+#define EDID_CEA_VCDB_QS	(1 << 6)
 
 /**
  * Search EDID for CEA extension block.
@@ -1902,6 +1904,37 @@ end:
 EXPORT_SYMBOL(drm_detect_monitor_audio);
 
 /**
+ * drm_rgb_quant_range_selectable - is RGB quantization range selectable?
+ *
+ * Check whether the monitor reports the RGB quantization range selection
+ * as supported. The AVI infoframe can then be used to inform the monitor
+ * which quantization range (full or limited) is used.
+ */
+bool drm_rgb_quant_range_selectable(struct edid *edid)
+{
+	u8 *edid_ext;
+	int i, start, end;
+
+	edid_ext = drm_find_cea_extension(edid);
+	if (!edid_ext)
+		return false;
+
+	if (cea_db_offsets(edid_ext, &start, &end))
+		return false;
+
+	for_each_cea_db(edid_ext, i, start, end) {
+		if (cea_db_tag(&edid_ext[i]) == VIDEO_CAPABILITY_BLOCK &&
+		    cea_db_payload_len(&edid_ext[i]) == 2) {
+			DRM_DEBUG_KMS("CEA VCDB 0x%02x\n", edid_ext[i + 2]);
+			return edid_ext[i + 2] & EDID_CEA_VCDB_QS;
+		}
+	}
+
+	return false;
+}
+EXPORT_SYMBOL(drm_rgb_quant_range_selectable);
+
+/**
  * drm_add_display_info - pull display info out if present
  * @edid: EDID data
  * @info: display info (attached to connector)
diff --git a/include/drm/drm_crtc.h b/include/drm/drm_crtc.h
index 00d78b5..30892dc 100644
--- a/include/drm/drm_crtc.h
+++ b/include/drm/drm_crtc.h
@@ -1033,6 +1033,7 @@ extern u8 *drm_find_cea_extension(struct edid *edid);
 extern u8 drm_match_cea_mode(struct drm_display_mode *to_match);
 extern bool drm_detect_hdmi_monitor(struct edid *edid);
 extern bool drm_detect_monitor_audio(struct edid *edid);
+extern bool drm_rgb_quant_range_selectable(struct edid *edid);
 extern int drm_mode_page_flip_ioctl(struct drm_device *dev,
 				    void *data, struct drm_file *file_priv);
 extern struct drm_display_mode *drm_cvt_mode(struct drm_device *dev,
-- 
cgit v0.10.2


From abedc077b45eff0b5a8630af8431ad5d59213582 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 17 Jan 2013 16:31:31 +0200
Subject: drm/i915: Provide the quantization range in the AVI infoframe
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The AVI infoframe is able to inform the display whether the source is
sending full or limited range RGB data.

As per CEA-861 [1] we must first check whether the display reports the
quantization range as selectable, and if so we can set the approriate
bits in the AVI inforframe.

[1] CEA-861-E - 6.4 Format of Version 2 AVI InfoFrame

v2: Give the Q bits better names, add spec chapter information

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 1a698c6..aeff0d1 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -289,6 +289,9 @@ struct cxsr_latency {
 #define DIP_LEN_AVI     13
 #define DIP_AVI_PR_1    0
 #define DIP_AVI_PR_2    1
+#define DIP_AVI_RGB_QUANT_RANGE_DEFAULT	(0 << 2)
+#define DIP_AVI_RGB_QUANT_RANGE_LIMITED	(1 << 2)
+#define DIP_AVI_RGB_QUANT_RANGE_FULL	(2 << 2)
 
 #define DIP_TYPE_SPD	0x83
 #define DIP_VERSION_SPD	0x1
@@ -347,6 +350,7 @@ struct intel_hdmi {
 	bool has_hdmi_sink;
 	bool has_audio;
 	enum hdmi_force_audio force_audio;
+	bool rgb_quant_range_selectable;
 	void (*write_infoframe)(struct drm_encoder *encoder,
 				struct dip_infoframe *frame);
 	void (*set_infoframes)(struct drm_encoder *encoder,
diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c
index db67be6..d53b731 100644
--- a/drivers/gpu/drm/i915/intel_hdmi.c
+++ b/drivers/gpu/drm/i915/intel_hdmi.c
@@ -331,6 +331,7 @@ static void intel_set_infoframe(struct drm_encoder *encoder,
 static void intel_hdmi_set_avi_infoframe(struct drm_encoder *encoder,
 					 struct drm_display_mode *adjusted_mode)
 {
+	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
 	struct dip_infoframe avi_if = {
 		.type = DIP_TYPE_AVI,
 		.ver = DIP_VERSION_AVI,
@@ -340,6 +341,13 @@ static void intel_hdmi_set_avi_infoframe(struct drm_encoder *encoder,
 	if (adjusted_mode->flags & DRM_MODE_FLAG_DBLCLK)
 		avi_if.body.avi.YQ_CN_PR |= DIP_AVI_PR_2;
 
+	if (intel_hdmi->rgb_quant_range_selectable) {
+		if (adjusted_mode->private_flags & INTEL_MODE_LIMITED_COLOR_RANGE)
+			avi_if.body.avi.ITC_EC_Q_SC |= DIP_AVI_RGB_QUANT_RANGE_LIMITED;
+		else
+			avi_if.body.avi.ITC_EC_Q_SC |= DIP_AVI_RGB_QUANT_RANGE_FULL;
+	}
+
 	avi_if.body.avi.VIC = drm_mode_cea_vic(adjusted_mode);
 
 	intel_set_infoframe(encoder, &avi_if);
@@ -825,6 +833,7 @@ intel_hdmi_detect(struct drm_connector *connector, bool force)
 
 	intel_hdmi->has_hdmi_sink = false;
 	intel_hdmi->has_audio = false;
+	intel_hdmi->rgb_quant_range_selectable = false;
 	edid = drm_get_edid(connector,
 			    intel_gmbus_get_adapter(dev_priv,
 						    intel_hdmi->ddc_bus));
@@ -836,6 +845,8 @@ intel_hdmi_detect(struct drm_connector *connector, bool force)
 				intel_hdmi->has_hdmi_sink =
 						drm_detect_hdmi_monitor(edid);
 			intel_hdmi->has_audio = drm_detect_monitor_audio(edid);
+			intel_hdmi->rgb_quant_range_selectable =
+				drm_rgb_quant_range_selectable(edid);
 		}
 		kfree(edid);
 	}
diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c
index 3e34a35..f01063a 100644
--- a/drivers/gpu/drm/i915/intel_sdvo.c
+++ b/drivers/gpu/drm/i915/intel_sdvo.c
@@ -126,6 +126,7 @@ struct intel_sdvo {
 	bool is_hdmi;
 	bool has_hdmi_monitor;
 	bool has_hdmi_audio;
+	bool rgb_quant_range_selectable;
 
 	/**
 	 * This is set if we detect output of sdvo device as LVDS and
@@ -947,7 +948,8 @@ static bool intel_sdvo_write_infoframe(struct intel_sdvo *intel_sdvo,
 				    &tx_rate, 1);
 }
 
-static bool intel_sdvo_set_avi_infoframe(struct intel_sdvo *intel_sdvo)
+static bool intel_sdvo_set_avi_infoframe(struct intel_sdvo *intel_sdvo,
+					 const struct drm_display_mode *adjusted_mode)
 {
 	struct dip_infoframe avi_if = {
 		.type = DIP_TYPE_AVI,
@@ -956,6 +958,13 @@ static bool intel_sdvo_set_avi_infoframe(struct intel_sdvo *intel_sdvo)
 	};
 	uint8_t sdvo_data[4 + sizeof(avi_if.body.avi)];
 
+	if (intel_sdvo->rgb_quant_range_selectable) {
+		if (adjusted_mode->private_flags & INTEL_MODE_LIMITED_COLOR_RANGE)
+			avi_if.body.avi.ITC_EC_Q_SC |= DIP_AVI_RGB_QUANT_RANGE_LIMITED;
+		else
+			avi_if.body.avi.ITC_EC_Q_SC |= DIP_AVI_RGB_QUANT_RANGE_FULL;
+	}
+
 	intel_dip_infoframe_csum(&avi_if);
 
 	/* sdvo spec says that the ecc is handled by the hw, and it looks like
@@ -1134,7 +1143,7 @@ static void intel_sdvo_mode_set(struct drm_encoder *encoder,
 		intel_sdvo_set_encode(intel_sdvo, SDVO_ENCODE_HDMI);
 		intel_sdvo_set_colorimetry(intel_sdvo,
 					   SDVO_COLORIMETRY_RGB256);
-		intel_sdvo_set_avi_infoframe(intel_sdvo);
+		intel_sdvo_set_avi_infoframe(intel_sdvo, adjusted_mode);
 	} else
 		intel_sdvo_set_encode(intel_sdvo, SDVO_ENCODE_DVI);
 
@@ -1526,6 +1535,8 @@ intel_sdvo_tmds_sink_detect(struct drm_connector *connector)
 			if (intel_sdvo->is_hdmi) {
 				intel_sdvo->has_hdmi_monitor = drm_detect_hdmi_monitor(edid);
 				intel_sdvo->has_hdmi_audio = drm_detect_monitor_audio(edid);
+				intel_sdvo->rgb_quant_range_selectable =
+					drm_rgb_quant_range_selectable(edid);
 			}
 		} else
 			status = connector_status_disconnected;
@@ -1577,6 +1588,7 @@ intel_sdvo_detect(struct drm_connector *connector, bool force)
 
 	intel_sdvo->has_hdmi_monitor = false;
 	intel_sdvo->has_hdmi_audio = false;
+	intel_sdvo->rgb_quant_range_selectable = false;
 
 	if ((intel_sdvo_connector->output_flag & response) == 0)
 		ret = connector_status_disconnected;
-- 
cgit v0.10.2


From a81cc00c11ab6816fbcb7dd99a60b50e71765d25 Mon Sep 17 00:00:00 2001
From: Ben Widawsky <ben@bwidawsk.net>
Date: Fri, 18 Jan 2013 12:30:31 -0800
Subject: drm/i915: Cut out the infamous ILK w/a from AGP layer

And, move it to where the rest of the logic is.

There is some slight functionality changes. There was extra paranoid
checks in AGP code making sure we never do idle maps on gen2 parts. That
was not duplicated as the simple PCI id check should do the right thing.

v2: use IS_GEN5 && IS_MOBILE check instead. For now, this is the same as
IS_IRONLAKE_M but is more future proof. The workaround docs hint that
more than one platform may be effected, but we've never seen such a
platform in the wild. (Rodrigo, Daniel)

Reviewed-by: Rodrigo Vivi <rodrigo.vivi@gmail.com> (v1)
Cc: Dave Airlie <airlied@redhat.com>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/char/agp/intel-gtt.c b/drivers/char/agp/intel-gtt.c
index c8d9dcb..12c3102 100644
--- a/drivers/char/agp/intel-gtt.c
+++ b/drivers/char/agp/intel-gtt.c
@@ -840,9 +840,6 @@ static int intel_fake_agp_insert_entries(struct agp_memory *mem,
 {
 	int ret = -EINVAL;
 
-	if (intel_private.base.do_idle_maps)
-		return -ENODEV;
-
 	if (intel_private.clear_fake_agp) {
 		int start = intel_private.base.stolen_size / PAGE_SIZE;
 		int end = intel_private.base.gtt_mappable_entries;
@@ -907,9 +904,6 @@ static int intel_fake_agp_remove_entries(struct agp_memory *mem,
 	if (mem->page_count == 0)
 		return 0;
 
-	if (intel_private.base.do_idle_maps)
-		return -ENODEV;
-
 	intel_gtt_clear_range(pg_start, mem->page_count);
 
 	if (intel_private.base.needs_dmar) {
@@ -1069,24 +1063,6 @@ static void i965_write_entry(dma_addr_t addr,
 	writel(addr | pte_flags, intel_private.gtt + entry);
 }
 
-/* Certain Gen5 chipsets require require idling the GPU before
- * unmapping anything from the GTT when VT-d is enabled.
- */
-static inline int needs_idle_maps(void)
-{
-#ifdef CONFIG_INTEL_IOMMU
-	const unsigned short gpu_devid = intel_private.pcidev->device;
-
-	/* Query intel_iommu to see if we need the workaround. Presumably that
-	 * was loaded first.
-	 */
-	if ((gpu_devid == PCI_DEVICE_ID_INTEL_IRONLAKE_M_HB ||
-	     gpu_devid == PCI_DEVICE_ID_INTEL_IRONLAKE_M_IG) &&
-	     intel_iommu_gfx_mapped)
-		return 1;
-#endif
-	return 0;
-}
 
 static int i9xx_setup(void)
 {
@@ -1115,9 +1091,6 @@ static int i9xx_setup(void)
 		break;
 	}
 
-	if (needs_idle_maps())
-		intel_private.base.do_idle_maps = 1;
-
 	intel_i9xx_setup_flush();
 
 	return 0;
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 0a2a18b8..1fe802f 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -381,6 +381,8 @@ struct i915_gtt {
 
 	/** "Graphics Stolen Memory" holds the global PTEs */
 	void __iomem *gsm;
+
+	bool do_idle_maps;
 };
 
 #define I915_PPGTT_PD_ENTRIES 512
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 0f0db02..84c521f 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -338,11 +338,27 @@ void i915_gem_init_ppgtt(struct drm_device *dev)
 	}
 }
 
+extern int intel_iommu_gfx_mapped;
+/* Certain Gen5 chipsets require require idling the GPU before
+ * unmapping anything from the GTT when VT-d is enabled.
+ */
+static inline bool needs_idle_maps(struct drm_device *dev)
+{
+#ifdef CONFIG_INTEL_IOMMU
+	/* Query intel_iommu to see if we need the workaround. Presumably that
+	 * was loaded first.
+	 */
+	if (IS_GEN5(dev) && IS_MOBILE(dev) && intel_iommu_gfx_mapped)
+		return true;
+#endif
+	return false;
+}
+
 static bool do_idling(struct drm_i915_private *dev_priv)
 {
 	bool ret = dev_priv->mm.interruptible;
 
-	if (unlikely(dev_priv->mm.gtt->do_idle_maps)) {
+	if (unlikely(dev_priv->gtt.do_idle_maps)) {
 		dev_priv->mm.interruptible = false;
 		if (i915_gpu_idle(dev_priv->dev)) {
 			DRM_ERROR("Couldn't idle GPU\n");
@@ -356,11 +372,10 @@ static bool do_idling(struct drm_i915_private *dev_priv)
 
 static void undo_idling(struct drm_i915_private *dev_priv, bool interruptible)
 {
-	if (unlikely(dev_priv->mm.gtt->do_idle_maps))
+	if (unlikely(dev_priv->gtt.do_idle_maps))
 		dev_priv->mm.interruptible = interruptible;
 }
 
-
 static void i915_ggtt_clear_range(struct drm_device *dev,
 				 unsigned first_entry,
 				 unsigned num_entries)
@@ -709,6 +724,9 @@ int i915_gem_gtt_init(struct drm_device *dev)
 			intel_gmch_remove();
 			return -ENODEV;
 		}
+
+		dev_priv->gtt.do_idle_maps = needs_idle_maps(dev);
+
 		return 0;
 	}
 
diff --git a/include/drm/intel-gtt.h b/include/drm/intel-gtt.h
index 3e3a166..1747aa09 100644
--- a/include/drm/intel-gtt.h
+++ b/include/drm/intel-gtt.h
@@ -13,8 +13,6 @@ struct intel_gtt {
 	unsigned int gtt_mappable_entries;
 	/* Whether i915 needs to use the dmar apis or not. */
 	unsigned int needs_dmar : 1;
-	/* Whether we idle the gpu before mapping/unmapping */
-	unsigned int do_idle_maps : 1;
 	/* Share the scratch page dma with ppgtts. */
 	dma_addr_t scratch_page_dma;
 	struct page *scratch_page;
-- 
cgit v0.10.2


From 9c61a32d31a55c8c6e590d83ae5645e14fde09f2 Mon Sep 17 00:00:00 2001
From: Ben Widawsky <ben@bwidawsk.net>
Date: Fri, 18 Jan 2013 12:30:32 -0800
Subject: drm/i915: Remove scratch page from shared

We already had a mapping in both (minus the phys_addr in AGP).

Reviewed-by: Rodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/char/agp/intel-gtt.c b/drivers/char/agp/intel-gtt.c
index 12c3102..7fcee5c 100644
--- a/drivers/char/agp/intel-gtt.c
+++ b/drivers/char/agp/intel-gtt.c
@@ -75,6 +75,7 @@ static struct _intel_private {
 	struct resource ifp_resource;
 	int resource_valid;
 	struct page *scratch_page;
+	phys_addr_t scratch_page_dma;
 	int refcount;
 } intel_private;
 
@@ -297,9 +298,9 @@ static int intel_gtt_setup_scratch_page(void)
 		if (pci_dma_mapping_error(intel_private.pcidev, dma_addr))
 			return -EINVAL;
 
-		intel_private.base.scratch_page_dma = dma_addr;
+		intel_private.scratch_page_dma = dma_addr;
 	} else
-		intel_private.base.scratch_page_dma = page_to_phys(page);
+		intel_private.scratch_page_dma = page_to_phys(page);
 
 	intel_private.scratch_page = page;
 
@@ -546,7 +547,7 @@ static unsigned int intel_gtt_mappable_entries(void)
 static void intel_gtt_teardown_scratch_page(void)
 {
 	set_pages_wb(intel_private.scratch_page, 1);
-	pci_unmap_page(intel_private.pcidev, intel_private.base.scratch_page_dma,
+	pci_unmap_page(intel_private.pcidev, intel_private.scratch_page_dma,
 		       PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
 	put_page(intel_private.scratch_page);
 	__free_page(intel_private.scratch_page);
@@ -891,7 +892,7 @@ void intel_gtt_clear_range(unsigned int first_entry, unsigned int num_entries)
 	unsigned int i;
 
 	for (i = first_entry; i < (first_entry + num_entries); i++) {
-		intel_private.driver->write_entry(intel_private.base.scratch_page_dma,
+		intel_private.driver->write_entry(intel_private.scratch_page_dma,
 						  i, 0);
 	}
 	readl(intel_private.gtt+i-1);
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 1fe802f..3189034 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -383,6 +383,8 @@ struct i915_gtt {
 	void __iomem *gsm;
 
 	bool do_idle_maps;
+	dma_addr_t scratch_page_dma;
+	struct page *scratch_page;
 };
 
 #define I915_PPGTT_PD_ENTRIES 512
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 84c521f..3c1f66e 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -162,7 +162,7 @@ int i915_gem_init_aliasing_ppgtt(struct drm_device *dev)
 		}
 	}
 
-	ppgtt->scratch_page_dma_addr = dev_priv->mm.gtt->scratch_page_dma;
+	ppgtt->scratch_page_dma_addr = dev_priv->gtt.scratch_page_dma;
 
 	i915_ppgtt_clear_range(ppgtt, 0,
 			       ppgtt->num_pd_entries*I915_PPGTT_PT_ENTRIES);
@@ -396,7 +396,7 @@ static void i915_ggtt_clear_range(struct drm_device *dev,
 		 first_entry, num_entries, max_entries))
 		num_entries = max_entries;
 
-	scratch_pte = pte_encode(dev, dev_priv->mm.gtt->scratch_page_dma, I915_CACHE_LLC);
+	scratch_pte = pte_encode(dev, dev_priv->gtt.scratch_page_dma, I915_CACHE_LLC);
 	for (i = 0; i < num_entries; i++)
 		iowrite32(scratch_pte, &gtt_base[i]);
 	readl(gtt_base);
@@ -659,8 +659,8 @@ static int setup_scratch_page(struct drm_device *dev)
 #else
 	dma_addr = page_to_phys(page);
 #endif
-	dev_priv->mm.gtt->scratch_page = page;
-	dev_priv->mm.gtt->scratch_page_dma = dma_addr;
+	dev_priv->gtt.scratch_page = page;
+	dev_priv->gtt.scratch_page_dma = dma_addr;
 
 	return 0;
 }
@@ -668,11 +668,11 @@ static int setup_scratch_page(struct drm_device *dev)
 static void teardown_scratch_page(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	set_pages_wb(dev_priv->mm.gtt->scratch_page, 1);
-	pci_unmap_page(dev->pdev, dev_priv->mm.gtt->scratch_page_dma,
+	set_pages_wb(dev_priv->gtt.scratch_page, 1);
+	pci_unmap_page(dev->pdev, dev_priv->gtt.scratch_page_dma,
 		       PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
-	put_page(dev_priv->mm.gtt->scratch_page);
-	__free_page(dev_priv->mm.gtt->scratch_page);
+	put_page(dev_priv->gtt.scratch_page);
+	__free_page(dev_priv->gtt.scratch_page);
 }
 
 static inline unsigned int gen6_get_total_gtt_size(u16 snb_gmch_ctl)
diff --git a/include/drm/intel-gtt.h b/include/drm/intel-gtt.h
index 1747aa09..c787ee4 100644
--- a/include/drm/intel-gtt.h
+++ b/include/drm/intel-gtt.h
@@ -13,9 +13,6 @@ struct intel_gtt {
 	unsigned int gtt_mappable_entries;
 	/* Whether i915 needs to use the dmar apis or not. */
 	unsigned int needs_dmar : 1;
-	/* Share the scratch page dma with ppgtts. */
-	dma_addr_t scratch_page_dma;
-	struct page *scratch_page;
 	/* needed for ioremap in drm/i915 */
 	phys_addr_t gma_bus_addr;
 } *intel_gtt_get(void);
-- 
cgit v0.10.2


From 8d2e630899165d413ae8a2adc36846ac0b71bada Mon Sep 17 00:00:00 2001
From: Ben Widawsky <ben@bwidawsk.net>
Date: Fri, 18 Jan 2013 12:30:33 -0800
Subject: drm/i915: Needs_dmar, not

The reasoning behind our code taking two paths depending upon whether or
not we may have been configured for IOMMU isn't clear to me. It should
always be safe to use the pci mapping functions as they are designed to
abstract the decision we were handling in i915.

Aside from simpler code, removing another member for the intel_gtt
struct is a nice motivation.

I ran this by Chris, and he wasn't concerned about the extra kzalloc,
and memory references vs. page_to_phys calculation in the case without
IOMMU.

v2: Update commit message

v3: Remove needs_dmar addition from Zhenyu upstream

This reverts (and then other stuff)
commit 20652097dadd9a7fb4d652f25466299974bc78f9
Author: Zhenyu Wang <zhenyuw@linux.intel.com>
Date:   Thu Dec 13 23:47:47 2012 +0800

    drm/i915: Fix missed needs_dmar setting

Reviewed-by: Rodrigo Vivi <rodrigo.vivi@gmail.com> (v2)
Cc: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
[danvet: Squash in follow-up fix to remove the bogus hunk which
deleted the dma_mask configuration for gen6+.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/char/agp/intel-gtt.c b/drivers/char/agp/intel-gtt.c
index 7fcee5c..b452414 100644
--- a/drivers/char/agp/intel-gtt.c
+++ b/drivers/char/agp/intel-gtt.c
@@ -77,6 +77,8 @@ static struct _intel_private {
 	struct page *scratch_page;
 	phys_addr_t scratch_page_dma;
 	int refcount;
+	/* Whether i915 needs to use the dmar apis or not. */
+	unsigned int needs_dmar : 1;
 } intel_private;
 
 #define INTEL_GTT_GEN	intel_private.driver->gen
@@ -292,7 +294,7 @@ static int intel_gtt_setup_scratch_page(void)
 	get_page(page);
 	set_pages_uc(page, 1);
 
-	if (intel_private.base.needs_dmar) {
+	if (intel_private.needs_dmar) {
 		dma_addr = pci_map_page(intel_private.pcidev, page, 0,
 				    PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
 		if (pci_dma_mapping_error(intel_private.pcidev, dma_addr))
@@ -608,7 +610,7 @@ static int intel_gtt_init(void)
 
 	intel_private.base.stolen_size = intel_gtt_stolen_size();
 
-	intel_private.base.needs_dmar = USE_PCI_DMA_API && INTEL_GTT_GEN > 2;
+	intel_private.needs_dmar = USE_PCI_DMA_API && INTEL_GTT_GEN > 2;
 
 	ret = intel_gtt_setup_scratch_page();
 	if (ret != 0) {
@@ -866,7 +868,7 @@ static int intel_fake_agp_insert_entries(struct agp_memory *mem,
 	if (!mem->is_flushed)
 		global_cache_flush();
 
-	if (intel_private.base.needs_dmar) {
+	if (intel_private.needs_dmar) {
 		struct sg_table st;
 
 		ret = intel_gtt_map_memory(mem->pages, mem->page_count, &st);
@@ -907,7 +909,7 @@ static int intel_fake_agp_remove_entries(struct agp_memory *mem,
 
 	intel_gtt_clear_range(pg_start, mem->page_count);
 
-	if (intel_private.base.needs_dmar) {
+	if (intel_private.needs_dmar) {
 		intel_gtt_unmap_memory(mem->sg_list, mem->num_sg);
 		mem->sg_list = NULL;
 		mem->num_sg = 0;
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 3c1f66e..a0ba4a9 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -138,28 +138,23 @@ int i915_gem_init_aliasing_ppgtt(struct drm_device *dev)
 			goto err_pt_alloc;
 	}
 
-	if (dev_priv->mm.gtt->needs_dmar) {
-		ppgtt->pt_dma_addr = kzalloc(sizeof(dma_addr_t)
-						*ppgtt->num_pd_entries,
-					     GFP_KERNEL);
-		if (!ppgtt->pt_dma_addr)
-			goto err_pt_alloc;
+	ppgtt->pt_dma_addr = kzalloc(sizeof(dma_addr_t) *ppgtt->num_pd_entries,
+				     GFP_KERNEL);
+	if (!ppgtt->pt_dma_addr)
+		goto err_pt_alloc;
 
-		for (i = 0; i < ppgtt->num_pd_entries; i++) {
-			dma_addr_t pt_addr;
+	for (i = 0; i < ppgtt->num_pd_entries; i++) {
+		dma_addr_t pt_addr;
 
-			pt_addr = pci_map_page(dev->pdev, ppgtt->pt_pages[i],
-					       0, 4096,
-					       PCI_DMA_BIDIRECTIONAL);
+		pt_addr = pci_map_page(dev->pdev, ppgtt->pt_pages[i], 0, 4096,
+				       PCI_DMA_BIDIRECTIONAL);
 
-			if (pci_dma_mapping_error(dev->pdev,
-						  pt_addr)) {
-				ret = -EIO;
-				goto err_pd_pin;
+		if (pci_dma_mapping_error(dev->pdev, pt_addr)) {
+			ret = -EIO;
+			goto err_pd_pin;
 
-			}
-			ppgtt->pt_dma_addr[i] = pt_addr;
 		}
+		ppgtt->pt_dma_addr[i] = pt_addr;
 	}
 
 	ppgtt->scratch_page_dma_addr = dev_priv->gtt.scratch_page_dma;
@@ -294,11 +289,7 @@ void i915_gem_init_ppgtt(struct drm_device *dev)
 	for (i = 0; i < ppgtt->num_pd_entries; i++) {
 		dma_addr_t pt_addr;
 
-		if (dev_priv->mm.gtt->needs_dmar)
-			pt_addr = ppgtt->pt_dma_addr[i];
-		else
-			pt_addr = page_to_phys(ppgtt->pt_pages[i]);
-
+		pt_addr = ppgtt->pt_dma_addr[i];
 		pd_entry = GEN6_PDE_ADDR_ENCODE(pt_addr);
 		pd_entry |= GEN6_PDE_VALID;
 
@@ -730,16 +721,12 @@ int i915_gem_gtt_init(struct drm_device *dev)
 		return 0;
 	}
 
-	dev_priv->mm.gtt = kzalloc(sizeof(*dev_priv->mm.gtt), GFP_KERNEL);
-	if (!dev_priv->mm.gtt)
-		return -ENOMEM;
-
 	if (!pci_set_dma_mask(dev->pdev, DMA_BIT_MASK(40)))
 		pci_set_consistent_dma_mask(dev->pdev, DMA_BIT_MASK(40));
 
-#ifdef CONFIG_INTEL_IOMMU
-	dev_priv->mm.gtt->needs_dmar = 1;
-#endif
+	dev_priv->mm.gtt = kzalloc(sizeof(*dev_priv->mm.gtt), GFP_KERNEL);
+	if (!dev_priv->mm.gtt)
+		return -ENOMEM;
 
 	/* For GEN6+ the PTEs for the ggtt live at 2MB + BAR0 */
 	gtt_bus_addr = pci_resource_start(dev->pdev, 0) + (2<<20);
diff --git a/include/drm/intel-gtt.h b/include/drm/intel-gtt.h
index c787ee4..984105c 100644
--- a/include/drm/intel-gtt.h
+++ b/include/drm/intel-gtt.h
@@ -11,8 +11,6 @@ struct intel_gtt {
 	/* Part of the gtt that is mappable by the cpu, for those chips where
 	 * this is not the full gtt. */
 	unsigned int gtt_mappable_entries;
-	/* Whether i915 needs to use the dmar apis or not. */
-	unsigned int needs_dmar : 1;
 	/* needed for ioremap in drm/i915 */
 	phys_addr_t gma_bus_addr;
 } *intel_gtt_get(void);
-- 
cgit v0.10.2


From e5c653777986b40e2986d2c918847fddbcba3a34 Mon Sep 17 00:00:00 2001
From: Ben Widawsky <ben@bwidawsk.net>
Date: Fri, 18 Jan 2013 12:30:34 -0800
Subject: agp/intel: Add gma_bus_addr

It is no longer used in the i915 code, so isolate it from the shared
struct.

This was originally part of:
commit 0e275518f325418d559c05327775bff894b237f7
Author: Ben Widawsky <ben@bwidawsk.net>
Date:   Mon Jan 14 13:35:33 2013 -0800

    agp/intel: decouple more of the agp-i915 sharing

    Reviewed-by: Rodrigo Vivi <rodrigo.vivi@gmail.com>
    Signed-off-by: Ben Widawsky <ben@bwidawsk.net>

That commit had some other hunks which can't be used due to issues
Daniel found in previous commits.

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
[danvet: drop squash notice from the commit since it's imo ok to keep
this one separate.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/char/agp/intel-gtt.c b/drivers/char/agp/intel-gtt.c
index b452414..ff5f348 100644
--- a/drivers/char/agp/intel-gtt.c
+++ b/drivers/char/agp/intel-gtt.c
@@ -79,6 +79,7 @@ static struct _intel_private {
 	int refcount;
 	/* Whether i915 needs to use the dmar apis or not. */
 	unsigned int needs_dmar : 1;
+	phys_addr_t gma_bus_addr;
 } intel_private;
 
 #define INTEL_GTT_GEN	intel_private.driver->gen
@@ -625,7 +626,7 @@ static int intel_gtt_init(void)
 		pci_read_config_dword(intel_private.pcidev, I915_GMADDR,
 				      &gma_addr);
 
-	intel_private.base.gma_bus_addr = (gma_addr & PCI_BASE_ADDRESS_MEM_MASK);
+	intel_private.gma_bus_addr = (gma_addr & PCI_BASE_ADDRESS_MEM_MASK);
 
 	return 0;
 }
@@ -781,7 +782,7 @@ static int intel_fake_agp_configure(void)
 	    return -EIO;
 
 	intel_private.clear_fake_agp = true;
-	agp_bridge->gart_bus_addr = intel_private.base.gma_bus_addr;
+	agp_bridge->gart_bus_addr = intel_private.gma_bus_addr;
 
 	return 0;
 }
diff --git a/include/drm/intel-gtt.h b/include/drm/intel-gtt.h
index 984105c..769b6c7 100644
--- a/include/drm/intel-gtt.h
+++ b/include/drm/intel-gtt.h
@@ -11,8 +11,6 @@ struct intel_gtt {
 	/* Part of the gtt that is mappable by the cpu, for those chips where
 	 * this is not the full gtt. */
 	unsigned int gtt_mappable_entries;
-	/* needed for ioremap in drm/i915 */
-	phys_addr_t gma_bus_addr;
 } *intel_gtt_get(void);
 
 int intel_gmch_probe(struct pci_dev *bridge_pdev, struct pci_dev *gpu_pdev,
-- 
cgit v0.10.2


From 4b5aed62121eddfc47fd8f2739ca6b802b97390e Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Wed, 14 Nov 2012 17:14:03 +0100
Subject: drm/i915: move dev_priv->mm out of line

Tha one is really big, since it contains tons of comments explaining
how things work. Which is nice ;-)

Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 3189034..ea32268 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -665,6 +665,107 @@ struct intel_l3_parity {
 	struct work_struct error_work;
 };
 
+struct i915_gem_mm {
+	/** Bridge to intel-gtt-ko */
+	struct intel_gtt *gtt;
+	/** Memory allocator for GTT stolen memory */
+	struct drm_mm stolen;
+	/** Memory allocator for GTT */
+	struct drm_mm gtt_space;
+	/** List of all objects in gtt_space. Used to restore gtt
+	 * mappings on resume */
+	struct list_head bound_list;
+	/**
+	 * List of objects which are not bound to the GTT (thus
+	 * are idle and not used by the GPU) but still have
+	 * (presumably uncached) pages still attached.
+	 */
+	struct list_head unbound_list;
+
+	/** Usable portion of the GTT for GEM */
+	unsigned long stolen_base; /* limited to low memory (32-bit) */
+
+	int gtt_mtrr;
+
+	/** PPGTT used for aliasing the PPGTT with the GTT */
+	struct i915_hw_ppgtt *aliasing_ppgtt;
+
+	struct shrinker inactive_shrinker;
+	bool shrinker_no_lock_stealing;
+
+	/**
+	 * List of objects currently involved in rendering.
+	 *
+	 * Includes buffers having the contents of their GPU caches
+	 * flushed, not necessarily primitives.  last_rendering_seqno
+	 * represents when the rendering involved will be completed.
+	 *
+	 * A reference is held on the buffer while on this list.
+	 */
+	struct list_head active_list;
+
+	/**
+	 * LRU list of objects which are not in the ringbuffer and
+	 * are ready to unbind, but are still in the GTT.
+	 *
+	 * last_rendering_seqno is 0 while an object is in this list.
+	 *
+	 * A reference is not held on the buffer while on this list,
+	 * as merely being GTT-bound shouldn't prevent its being
+	 * freed, and we'll pull it off the list in the free path.
+	 */
+	struct list_head inactive_list;
+
+	/** LRU list of objects with fence regs on them. */
+	struct list_head fence_list;
+
+	/**
+	 * We leave the user IRQ off as much as possible,
+	 * but this means that requests will finish and never
+	 * be retired once the system goes idle. Set a timer to
+	 * fire periodically while the ring is running. When it
+	 * fires, go retire requests.
+	 */
+	struct delayed_work retire_work;
+
+	/**
+	 * Are we in a non-interruptible section of code like
+	 * modesetting?
+	 */
+	bool interruptible;
+
+	/**
+	 * Flag if the X Server, and thus DRM, is not currently in
+	 * control of the device.
+	 *
+	 * This is set between LeaveVT and EnterVT.  It needs to be
+	 * replaced with a semaphore.  It also needs to be
+	 * transitioned away from for kernel modesetting.
+	 */
+	int suspended;
+
+	/**
+	 * Flag if the hardware appears to be wedged.
+	 *
+	 * This is set when attempts to idle the device timeout.
+	 * It prevents command submission from occurring and makes
+	 * every pending request fail
+	 */
+	atomic_t wedged;
+
+	/** Bit 6 swizzling required for X tiling */
+	uint32_t bit_6_swizzle_x;
+	/** Bit 6 swizzling required for Y tiling */
+	uint32_t bit_6_swizzle_y;
+
+	/* storage for physical objects */
+	struct drm_i915_gem_phys_object *phys_objs[I915_MAX_PHYS_OBJECT];
+
+	/* accounting, useful for userland debugging */
+	size_t object_memory;
+	u32 object_count;
+};
+
 typedef struct drm_i915_private {
 	struct drm_device *dev;
 	struct kmem_cache *slab;
@@ -806,106 +907,7 @@ typedef struct drm_i915_private {
 
 	struct i915_gtt gtt;
 
-	struct {
-		/** Bridge to intel-gtt-ko */
-		struct intel_gtt *gtt;
-		/** Memory allocator for GTT stolen memory */
-		struct drm_mm stolen;
-		/** Memory allocator for GTT */
-		struct drm_mm gtt_space;
-		/** List of all objects in gtt_space. Used to restore gtt
-		 * mappings on resume */
-		struct list_head bound_list;
-		/**
-		 * List of objects which are not bound to the GTT (thus
-		 * are idle and not used by the GPU) but still have
-		 * (presumably uncached) pages still attached.
-		 */
-		struct list_head unbound_list;
-
-		/** Usable portion of the GTT for GEM */
-		unsigned long stolen_base; /* limited to low memory (32-bit) */
-
-		int gtt_mtrr;
-
-		/** PPGTT used for aliasing the PPGTT with the GTT */
-		struct i915_hw_ppgtt *aliasing_ppgtt;
-
-		struct shrinker inactive_shrinker;
-		bool shrinker_no_lock_stealing;
-
-		/**
-		 * List of objects currently involved in rendering.
-		 *
-		 * Includes buffers having the contents of their GPU caches
-		 * flushed, not necessarily primitives.  last_rendering_seqno
-		 * represents when the rendering involved will be completed.
-		 *
-		 * A reference is held on the buffer while on this list.
-		 */
-		struct list_head active_list;
-
-		/**
-		 * LRU list of objects which are not in the ringbuffer and
-		 * are ready to unbind, but are still in the GTT.
-		 *
-		 * last_rendering_seqno is 0 while an object is in this list.
-		 *
-		 * A reference is not held on the buffer while on this list,
-		 * as merely being GTT-bound shouldn't prevent its being
-		 * freed, and we'll pull it off the list in the free path.
-		 */
-		struct list_head inactive_list;
-
-		/** LRU list of objects with fence regs on them. */
-		struct list_head fence_list;
-
-		/**
-		 * We leave the user IRQ off as much as possible,
-		 * but this means that requests will finish and never
-		 * be retired once the system goes idle. Set a timer to
-		 * fire periodically while the ring is running. When it
-		 * fires, go retire requests.
-		 */
-		struct delayed_work retire_work;
-
-		/**
-		 * Are we in a non-interruptible section of code like
-		 * modesetting?
-		 */
-		bool interruptible;
-
-		/**
-		 * Flag if the X Server, and thus DRM, is not currently in
-		 * control of the device.
-		 *
-		 * This is set between LeaveVT and EnterVT.  It needs to be
-		 * replaced with a semaphore.  It also needs to be
-		 * transitioned away from for kernel modesetting.
-		 */
-		int suspended;
-
-		/**
-		 * Flag if the hardware appears to be wedged.
-		 *
-		 * This is set when attempts to idle the device timeout.
-		 * It prevents command submission from occurring and makes
-		 * every pending request fail
-		 */
-		atomic_t wedged;
-
-		/** Bit 6 swizzling required for X tiling */
-		uint32_t bit_6_swizzle_x;
-		/** Bit 6 swizzling required for Y tiling */
-		uint32_t bit_6_swizzle_y;
-
-		/* storage for physical objects */
-		struct drm_i915_gem_phys_object *phys_objs[I915_MAX_PHYS_OBJECT];
-
-		/* accounting, useful for userland debugging */
-		size_t object_memory;
-		u32 object_count;
-	} mm;
+	struct i915_gem_mm mm;
 
 	/* Kernel Modesetting */
 
-- 
cgit v0.10.2


From 99584db33ba4f864777e2cfef5329ed1bf13f714 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Wed, 14 Nov 2012 17:14:04 +0100
Subject: drm/i915: extract hangcheck/reset/error_state state into substruct

This has been sprinkled all over the place in dev_priv. I think
it'd be good to also move all the code into a separate file like
i915_gem_error.c, but that's for another patch.

Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index 90a6fc5..3b1bf4e 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -814,11 +814,11 @@ static int i915_error_state_open(struct inode *inode, struct file *file)
 
 	error_priv->dev = dev;
 
-	spin_lock_irqsave(&dev_priv->error_lock, flags);
-	error_priv->error = dev_priv->first_error;
+	spin_lock_irqsave(&dev_priv->gpu_error.lock, flags);
+	error_priv->error = dev_priv->gpu_error.first_error;
 	if (error_priv->error)
 		kref_get(&error_priv->error->ref);
-	spin_unlock_irqrestore(&dev_priv->error_lock, flags);
+	spin_unlock_irqrestore(&dev_priv->gpu_error.lock, flags);
 
 	return single_open(file, i915_error_state, error_priv);
 }
@@ -1727,7 +1727,7 @@ i915_ring_stop_read(struct file *filp,
 	int len;
 
 	len = snprintf(buf, sizeof(buf),
-		       "0x%08x\n", dev_priv->stop_rings);
+		       "0x%08x\n", dev_priv->gpu_error.stop_rings);
 
 	if (len > sizeof(buf))
 		len = sizeof(buf);
@@ -1763,7 +1763,7 @@ i915_ring_stop_write(struct file *filp,
 	if (ret)
 		return ret;
 
-	dev_priv->stop_rings = val;
+	dev_priv->gpu_error.stop_rings = val;
 	mutex_unlock(&dev->struct_mutex);
 
 	return cnt;
diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index 3f70178..11c7aa8 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -1605,7 +1605,7 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 		pci_enable_msi(dev->pdev);
 
 	spin_lock_init(&dev_priv->irq_lock);
-	spin_lock_init(&dev_priv->error_lock);
+	spin_lock_init(&dev_priv->gpu_error.lock);
 	spin_lock_init(&dev_priv->rps.lock);
 	mutex_init(&dev_priv->dpio_lock);
 
@@ -1725,8 +1725,8 @@ int i915_driver_unload(struct drm_device *dev)
 	}
 
 	/* Free error state after interrupts are fully disabled. */
-	del_timer_sync(&dev_priv->hangcheck_timer);
-	cancel_work_sync(&dev_priv->error_work);
+	del_timer_sync(&dev_priv->gpu_error.hangcheck_timer);
+	cancel_work_sync(&dev_priv->gpu_error.work);
 	i915_destroy_error_state(dev);
 
 	if (dev->pdev->msi_enabled)
diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index c8cbc32..3ff8e73 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -779,9 +779,9 @@ int intel_gpu_reset(struct drm_device *dev)
 	}
 
 	/* Also reset the gpu hangman. */
-	if (dev_priv->stop_rings) {
+	if (dev_priv->gpu_error.stop_rings) {
 		DRM_DEBUG("Simulated gpu hang, resetting stop_rings\n");
-		dev_priv->stop_rings = 0;
+		dev_priv->gpu_error.stop_rings = 0;
 		if (ret == -ENODEV) {
 			DRM_ERROR("Reset not implemented, but ignoring "
 				  "error for simulated gpu hangs\n");
@@ -820,12 +820,12 @@ int i915_reset(struct drm_device *dev)
 	i915_gem_reset(dev);
 
 	ret = -ENODEV;
-	if (get_seconds() - dev_priv->last_gpu_reset < 5)
+	if (get_seconds() - dev_priv->gpu_error.last_reset < 5)
 		DRM_ERROR("GPU hanging too fast, declaring wedged!\n");
 	else
 		ret = intel_gpu_reset(dev);
 
-	dev_priv->last_gpu_reset = get_seconds();
+	dev_priv->gpu_error.last_reset = get_seconds();
 	if (ret) {
 		DRM_ERROR("Failed to reset chip.\n");
 		mutex_unlock(&dev->struct_mutex);
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index ea32268..dfe0e74 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -766,6 +766,28 @@ struct i915_gem_mm {
 	u32 object_count;
 };
 
+struct i915_gpu_error {
+	/* For hangcheck timer */
+#define DRM_I915_HANGCHECK_PERIOD 1500 /* in ms */
+#define DRM_I915_HANGCHECK_JIFFIES msecs_to_jiffies(DRM_I915_HANGCHECK_PERIOD)
+	struct timer_list hangcheck_timer;
+	int hangcheck_count;
+	uint32_t last_acthd[I915_NUM_RINGS];
+	uint32_t prev_instdone[I915_NUM_INSTDONE_REG];
+
+	/* For reset and error_state handling. */
+	spinlock_t lock;
+	/* Protected by the above dev->gpu_error.lock. */
+	struct drm_i915_error_state *first_error;
+	struct work_struct work;
+	struct completion completion;
+
+	unsigned long last_reset;
+
+	/* For gpu hang simulation. */
+	unsigned int stop_rings;
+};
+
 typedef struct drm_i915_private {
 	struct drm_device *dev;
 	struct kmem_cache *slab;
@@ -829,16 +851,6 @@ typedef struct drm_i915_private {
 	int num_pipe;
 	int num_pch_pll;
 
-	/* For hangcheck timer */
-#define DRM_I915_HANGCHECK_PERIOD 1500 /* in ms */
-#define DRM_I915_HANGCHECK_JIFFIES msecs_to_jiffies(DRM_I915_HANGCHECK_PERIOD)
-	struct timer_list hangcheck_timer;
-	int hangcheck_count;
-	uint32_t last_acthd[I915_NUM_RINGS];
-	uint32_t prev_instdone[I915_NUM_INSTDONE_REG];
-
-	unsigned int stop_rings;
-
 	unsigned long cfb_size;
 	unsigned int cfb_fb;
 	enum plane cfb_plane;
@@ -886,11 +898,6 @@ typedef struct drm_i915_private {
 
 	unsigned int fsb_freq, mem_freq, is_ddr3;
 
-	spinlock_t error_lock;
-	/* Protected by dev->error_lock. */
-	struct drm_i915_error_state *first_error;
-	struct work_struct error_work;
-	struct completion error_completion;
 	struct workqueue_struct *wq;
 
 	/* Display functions */
@@ -949,7 +956,7 @@ typedef struct drm_i915_private {
 	struct drm_mm_node *compressed_fb;
 	struct drm_mm_node *compressed_llb;
 
-	unsigned long last_gpu_reset;
+	struct i915_gpu_error gpu_error;
 
 	/* list of fbdev register on this device */
 	struct intel_fbdev *fbdev;
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index e4132ff..95e022e 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -90,7 +90,7 @@ static int
 i915_gem_wait_for_error(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct completion *x = &dev_priv->error_completion;
+	struct completion *x = &dev_priv->gpu_error.completion;
 	unsigned long flags;
 	int ret;
 
@@ -943,7 +943,7 @@ i915_gem_check_wedge(struct drm_i915_private *dev_priv,
 		     bool interruptible)
 {
 	if (atomic_read(&dev_priv->mm.wedged)) {
-		struct completion *x = &dev_priv->error_completion;
+		struct completion *x = &dev_priv->gpu_error.completion;
 		bool recovery_complete;
 		unsigned long flags;
 
@@ -2045,7 +2045,7 @@ i915_add_request(struct intel_ring_buffer *ring,
 
 	if (!dev_priv->mm.suspended) {
 		if (i915_enable_hangcheck) {
-			mod_timer(&dev_priv->hangcheck_timer,
+			mod_timer(&dev_priv->gpu_error.hangcheck_timer,
 				  round_jiffies_up(jiffies + DRM_I915_HANGCHECK_JIFFIES));
 		}
 		if (was_empty) {
@@ -3803,7 +3803,7 @@ i915_gem_idle(struct drm_device *dev)
 	 * And not confound mm.suspended!
 	 */
 	dev_priv->mm.suspended = 1;
-	del_timer_sync(&dev_priv->hangcheck_timer);
+	del_timer_sync(&dev_priv->gpu_error.hangcheck_timer);
 
 	i915_kernel_lost_context(dev);
 	i915_gem_cleanup_ringbuffer(dev);
@@ -4064,7 +4064,7 @@ i915_gem_load(struct drm_device *dev)
 		INIT_LIST_HEAD(&dev_priv->fence_regs[i].lru_list);
 	INIT_DELAYED_WORK(&dev_priv->mm.retire_work,
 			  i915_gem_retire_work_handler);
-	init_completion(&dev_priv->error_completion);
+	init_completion(&dev_priv->gpu_error.completion);
 
 	/* On GEN3 we really need to make sure the ARB C3 LP bit is set */
 	if (IS_GEN3(dev)) {
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index cc49a6d..c768ebd 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -356,8 +356,8 @@ static void notify_ring(struct drm_device *dev,
 
 	wake_up_all(&ring->irq_queue);
 	if (i915_enable_hangcheck) {
-		dev_priv->hangcheck_count = 0;
-		mod_timer(&dev_priv->hangcheck_timer,
+		dev_priv->gpu_error.hangcheck_count = 0;
+		mod_timer(&dev_priv->gpu_error.hangcheck_timer,
 			  round_jiffies_up(jiffies + DRM_I915_HANGCHECK_JIFFIES));
 	}
 }
@@ -863,7 +863,7 @@ done:
 static void i915_error_work_func(struct work_struct *work)
 {
 	drm_i915_private_t *dev_priv = container_of(work, drm_i915_private_t,
-						    error_work);
+						    gpu_error.work);
 	struct drm_device *dev = dev_priv->dev;
 	char *error_event[] = { "ERROR=1", NULL };
 	char *reset_event[] = { "RESET=1", NULL };
@@ -878,7 +878,7 @@ static void i915_error_work_func(struct work_struct *work)
 			atomic_set(&dev_priv->mm.wedged, 0);
 			kobject_uevent_env(&dev->primary->kdev.kobj, KOBJ_CHANGE, reset_done_event);
 		}
-		complete_all(&dev_priv->error_completion);
+		complete_all(&dev_priv->gpu_error.completion);
 	}
 }
 
@@ -1255,9 +1255,9 @@ static void i915_capture_error_state(struct drm_device *dev)
 	unsigned long flags;
 	int i, pipe;
 
-	spin_lock_irqsave(&dev_priv->error_lock, flags);
-	error = dev_priv->first_error;
-	spin_unlock_irqrestore(&dev_priv->error_lock, flags);
+	spin_lock_irqsave(&dev_priv->gpu_error.lock, flags);
+	error = dev_priv->gpu_error.first_error;
+	spin_unlock_irqrestore(&dev_priv->gpu_error.lock, flags);
 	if (error)
 		return;
 
@@ -1341,12 +1341,12 @@ static void i915_capture_error_state(struct drm_device *dev)
 	error->overlay = intel_overlay_capture_error_state(dev);
 	error->display = intel_display_capture_error_state(dev);
 
-	spin_lock_irqsave(&dev_priv->error_lock, flags);
-	if (dev_priv->first_error == NULL) {
-		dev_priv->first_error = error;
+	spin_lock_irqsave(&dev_priv->gpu_error.lock, flags);
+	if (dev_priv->gpu_error.first_error == NULL) {
+		dev_priv->gpu_error.first_error = error;
 		error = NULL;
 	}
-	spin_unlock_irqrestore(&dev_priv->error_lock, flags);
+	spin_unlock_irqrestore(&dev_priv->gpu_error.lock, flags);
 
 	if (error)
 		i915_error_state_free(&error->ref);
@@ -1358,10 +1358,10 @@ void i915_destroy_error_state(struct drm_device *dev)
 	struct drm_i915_error_state *error;
 	unsigned long flags;
 
-	spin_lock_irqsave(&dev_priv->error_lock, flags);
-	error = dev_priv->first_error;
-	dev_priv->first_error = NULL;
-	spin_unlock_irqrestore(&dev_priv->error_lock, flags);
+	spin_lock_irqsave(&dev_priv->gpu_error.lock, flags);
+	error = dev_priv->gpu_error.first_error;
+	dev_priv->gpu_error.first_error = NULL;
+	spin_unlock_irqrestore(&dev_priv->gpu_error.lock, flags);
 
 	if (error)
 		kref_put(&error->ref, i915_error_state_free);
@@ -1482,7 +1482,7 @@ void i915_handle_error(struct drm_device *dev, bool wedged)
 	i915_report_and_clear_eir(dev);
 
 	if (wedged) {
-		INIT_COMPLETION(dev_priv->error_completion);
+		INIT_COMPLETION(dev_priv->gpu_error.completion);
 		atomic_set(&dev_priv->mm.wedged, 1);
 
 		/*
@@ -1492,7 +1492,7 @@ void i915_handle_error(struct drm_device *dev, bool wedged)
 			wake_up_all(&ring->irq_queue);
 	}
 
-	queue_work(dev_priv->wq, &dev_priv->error_work);
+	queue_work(dev_priv->wq, &dev_priv->gpu_error.work);
 }
 
 static void i915_pageflip_stall_check(struct drm_device *dev, int pipe)
@@ -1723,7 +1723,7 @@ static bool i915_hangcheck_hung(struct drm_device *dev)
 {
 	drm_i915_private_t *dev_priv = dev->dev_private;
 
-	if (dev_priv->hangcheck_count++ > 1) {
+	if (dev_priv->gpu_error.hangcheck_count++ > 1) {
 		bool hung = true;
 
 		DRM_ERROR("Hangcheck timer elapsed... GPU hung\n");
@@ -1782,25 +1782,29 @@ void i915_hangcheck_elapsed(unsigned long data)
 			goto repeat;
 		}
 
-		dev_priv->hangcheck_count = 0;
+		dev_priv->gpu_error.hangcheck_count = 0;
 		return;
 	}
 
 	i915_get_extra_instdone(dev, instdone);
-	if (memcmp(dev_priv->last_acthd, acthd, sizeof(acthd)) == 0 &&
-	    memcmp(dev_priv->prev_instdone, instdone, sizeof(instdone)) == 0) {
+	if (memcmp(dev_priv->gpu_error.last_acthd, acthd,
+		   sizeof(acthd)) == 0 &&
+	    memcmp(dev_priv->gpu_error.prev_instdone, instdone,
+		   sizeof(instdone)) == 0) {
 		if (i915_hangcheck_hung(dev))
 			return;
 	} else {
-		dev_priv->hangcheck_count = 0;
+		dev_priv->gpu_error.hangcheck_count = 0;
 
-		memcpy(dev_priv->last_acthd, acthd, sizeof(acthd));
-		memcpy(dev_priv->prev_instdone, instdone, sizeof(instdone));
+		memcpy(dev_priv->gpu_error.last_acthd, acthd,
+		       sizeof(acthd));
+		memcpy(dev_priv->gpu_error.prev_instdone, instdone,
+		       sizeof(instdone));
 	}
 
 repeat:
 	/* Reset timer case chip hangs without another request being added */
-	mod_timer(&dev_priv->hangcheck_timer,
+	mod_timer(&dev_priv->gpu_error.hangcheck_timer,
 		  round_jiffies_up(jiffies + DRM_I915_HANGCHECK_JIFFIES));
 }
 
@@ -2769,11 +2773,12 @@ void intel_irq_init(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
 	INIT_WORK(&dev_priv->hotplug_work, i915_hotplug_work_func);
-	INIT_WORK(&dev_priv->error_work, i915_error_work_func);
+	INIT_WORK(&dev_priv->gpu_error.work, i915_error_work_func);
 	INIT_WORK(&dev_priv->rps.work, gen6_pm_rps_work);
 	INIT_WORK(&dev_priv->l3_parity.error_work, ivybridge_parity_work);
 
-	setup_timer(&dev_priv->hangcheck_timer, i915_hangcheck_elapsed,
+	setup_timer(&dev_priv->gpu_error.hangcheck_timer,
+		    i915_hangcheck_elapsed,
 		    (unsigned long) dev);
 
 	pm_qos_add_request(&dev_priv->pm_qos, PM_QOS_CPU_DMA_LATENCY, PM_QOS_DEFAULT_VALUE);
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index ce1d074..d6b06aa 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -1491,7 +1491,7 @@ void intel_ring_advance(struct intel_ring_buffer *ring)
 	struct drm_i915_private *dev_priv = ring->dev->dev_private;
 
 	ring->tail &= ring->size - 1;
-	if (dev_priv->stop_rings & intel_ring_flag(ring))
+	if (dev_priv->gpu_error.stop_rings & intel_ring_flag(ring))
 		return;
 	ring->write_tail(ring, ring->tail);
 }
-- 
cgit v0.10.2


From 33196deddacc7790defb9a7e84659e0362d4da7a Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Wed, 14 Nov 2012 17:14:05 +0100
Subject: drm/i915: move wedged to the other gpu error handling stuff

And to make Ben Widawsky happier, use the gpu_error instead of
the entire device as the argument in some functions.

Drop the outdated comment on ->wedged for now, a follow-up patch will
change the semantics and add a proper comment again.

Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index 3b1bf4e..e1b7eaf 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -1672,7 +1672,7 @@ i915_wedged_read(struct file *filp,
 
 	len = snprintf(buf, sizeof(buf),
 		       "wedged :  %d\n",
-		       atomic_read(&dev_priv->mm.wedged));
+		       atomic_read(&dev_priv->gpu_error.wedged));
 
 	if (len > sizeof(buf))
 		len = sizeof(buf);
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index dfe0e74..62da6c7 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -744,15 +744,6 @@ struct i915_gem_mm {
 	 */
 	int suspended;
 
-	/**
-	 * Flag if the hardware appears to be wedged.
-	 *
-	 * This is set when attempts to idle the device timeout.
-	 * It prevents command submission from occurring and makes
-	 * every pending request fail
-	 */
-	atomic_t wedged;
-
 	/** Bit 6 swizzling required for X tiling */
 	uint32_t bit_6_swizzle_x;
 	/** Bit 6 swizzling required for Y tiling */
@@ -784,6 +775,8 @@ struct i915_gpu_error {
 
 	unsigned long last_reset;
 
+	atomic_t wedged;
+
 	/* For gpu hang simulation. */
 	unsigned int stop_rings;
 };
@@ -1548,7 +1541,7 @@ i915_gem_object_unpin_fence(struct drm_i915_gem_object *obj)
 
 void i915_gem_retire_requests(struct drm_device *dev);
 void i915_gem_retire_requests_ring(struct intel_ring_buffer *ring);
-int __must_check i915_gem_check_wedge(struct drm_i915_private *dev_priv,
+int __must_check i915_gem_check_wedge(struct i915_gpu_error *error,
 				      bool interruptible);
 
 void i915_gem_reset(struct drm_device *dev);
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 95e022e..04b2f92 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -87,14 +87,13 @@ static void i915_gem_info_remove_obj(struct drm_i915_private *dev_priv,
 }
 
 static int
-i915_gem_wait_for_error(struct drm_device *dev)
+i915_gem_wait_for_error(struct i915_gpu_error *error)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct completion *x = &dev_priv->gpu_error.completion;
+	struct completion *x = &error->completion;
 	unsigned long flags;
 	int ret;
 
-	if (!atomic_read(&dev_priv->mm.wedged))
+	if (!atomic_read(&error->wedged))
 		return 0;
 
 	/*
@@ -110,7 +109,7 @@ i915_gem_wait_for_error(struct drm_device *dev)
 		return ret;
 	}
 
-	if (atomic_read(&dev_priv->mm.wedged)) {
+	if (atomic_read(&error->wedged)) {
 		/* GPU is hung, bump the completion count to account for
 		 * the token we just consumed so that we never hit zero and
 		 * end up waiting upon a subsequent completion event that
@@ -125,9 +124,10 @@ i915_gem_wait_for_error(struct drm_device *dev)
 
 int i915_mutex_lock_interruptible(struct drm_device *dev)
 {
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	int ret;
 
-	ret = i915_gem_wait_for_error(dev);
+	ret = i915_gem_wait_for_error(&dev_priv->gpu_error);
 	if (ret)
 		return ret;
 
@@ -939,11 +939,11 @@ unlock:
 }
 
 int
-i915_gem_check_wedge(struct drm_i915_private *dev_priv,
+i915_gem_check_wedge(struct i915_gpu_error *error,
 		     bool interruptible)
 {
-	if (atomic_read(&dev_priv->mm.wedged)) {
-		struct completion *x = &dev_priv->gpu_error.completion;
+	if (atomic_read(&error->wedged)) {
+		struct completion *x = &error->completion;
 		bool recovery_complete;
 		unsigned long flags;
 
@@ -1025,7 +1025,7 @@ static int __wait_seqno(struct intel_ring_buffer *ring, u32 seqno,
 
 #define EXIT_COND \
 	(i915_seqno_passed(ring->get_seqno(ring, false), seqno) || \
-	atomic_read(&dev_priv->mm.wedged))
+	atomic_read(&dev_priv->gpu_error.wedged))
 	do {
 		if (interruptible)
 			end = wait_event_interruptible_timeout(ring->irq_queue,
@@ -1035,7 +1035,7 @@ static int __wait_seqno(struct intel_ring_buffer *ring, u32 seqno,
 			end = wait_event_timeout(ring->irq_queue, EXIT_COND,
 						 timeout_jiffies);
 
-		ret = i915_gem_check_wedge(dev_priv, interruptible);
+		ret = i915_gem_check_wedge(&dev_priv->gpu_error, interruptible);
 		if (ret)
 			end = ret;
 	} while (end == 0 && wait_forever);
@@ -1081,7 +1081,7 @@ i915_wait_seqno(struct intel_ring_buffer *ring, uint32_t seqno)
 	BUG_ON(!mutex_is_locked(&dev->struct_mutex));
 	BUG_ON(seqno == 0);
 
-	ret = i915_gem_check_wedge(dev_priv, interruptible);
+	ret = i915_gem_check_wedge(&dev_priv->gpu_error, interruptible);
 	if (ret)
 		return ret;
 
@@ -1146,7 +1146,7 @@ i915_gem_object_wait_rendering__nonblocking(struct drm_i915_gem_object *obj,
 	if (seqno == 0)
 		return 0;
 
-	ret = i915_gem_check_wedge(dev_priv, true);
+	ret = i915_gem_check_wedge(&dev_priv->gpu_error, true);
 	if (ret)
 		return ret;
 
@@ -1379,7 +1379,7 @@ out:
 		/* If this -EIO is due to a gpu hang, give the reset code a
 		 * chance to clean up the mess. Otherwise return the proper
 		 * SIGBUS. */
-		if (!atomic_read(&dev_priv->mm.wedged))
+		if (!atomic_read(&dev_priv->gpu_error.wedged))
 			return VM_FAULT_SIGBUS;
 	case -EAGAIN:
 		/* Give the error handler a chance to run and move the
@@ -3390,7 +3390,7 @@ i915_gem_ring_throttle(struct drm_device *dev, struct drm_file *file)
 	u32 seqno = 0;
 	int ret;
 
-	if (atomic_read(&dev_priv->mm.wedged))
+	if (atomic_read(&dev_priv->gpu_error.wedged))
 		return -EIO;
 
 	spin_lock(&file_priv->mm.lock);
@@ -3978,9 +3978,9 @@ i915_gem_entervt_ioctl(struct drm_device *dev, void *data,
 	if (drm_core_check_feature(dev, DRIVER_MODESET))
 		return 0;
 
-	if (atomic_read(&dev_priv->mm.wedged)) {
+	if (atomic_read(&dev_priv->gpu_error.wedged)) {
 		DRM_ERROR("Reenabling wedged hardware, good luck\n");
-		atomic_set(&dev_priv->mm.wedged, 0);
+		atomic_set(&dev_priv->gpu_error.wedged, 0);
 	}
 
 	mutex_lock(&dev->struct_mutex);
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index c768ebd..f2c0016 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -871,11 +871,11 @@ static void i915_error_work_func(struct work_struct *work)
 
 	kobject_uevent_env(&dev->primary->kdev.kobj, KOBJ_CHANGE, error_event);
 
-	if (atomic_read(&dev_priv->mm.wedged)) {
+	if (atomic_read(&dev_priv->gpu_error.wedged)) {
 		DRM_DEBUG_DRIVER("resetting chip\n");
 		kobject_uevent_env(&dev->primary->kdev.kobj, KOBJ_CHANGE, reset_event);
 		if (!i915_reset(dev)) {
-			atomic_set(&dev_priv->mm.wedged, 0);
+			atomic_set(&dev_priv->gpu_error.wedged, 0);
 			kobject_uevent_env(&dev->primary->kdev.kobj, KOBJ_CHANGE, reset_done_event);
 		}
 		complete_all(&dev_priv->gpu_error.completion);
@@ -1483,7 +1483,7 @@ void i915_handle_error(struct drm_device *dev, bool wedged)
 
 	if (wedged) {
 		INIT_COMPLETION(dev_priv->gpu_error.completion);
-		atomic_set(&dev_priv->mm.wedged, 1);
+		atomic_set(&dev_priv->gpu_error.wedged, 1);
 
 		/*
 		 * Wakeup waiting processes so they don't hang
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index b35902e..160aa5f 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -2223,7 +2223,7 @@ intel_finish_fb(struct drm_framebuffer *old_fb)
 	WARN_ON(waitqueue_active(&dev_priv->pending_flip_queue));
 
 	wait_event(dev_priv->pending_flip_queue,
-		   atomic_read(&dev_priv->mm.wedged) ||
+		   atomic_read(&dev_priv->gpu_error.wedged) ||
 		   atomic_read(&obj->pending_flip) == 0);
 
 	/* Big Hammer, we also need to ensure that any pending
@@ -2871,7 +2871,7 @@ static bool intel_crtc_has_pending_flip(struct drm_crtc *crtc)
 	unsigned long flags;
 	bool pending;
 
-	if (atomic_read(&dev_priv->mm.wedged))
+	if (atomic_read(&dev_priv->gpu_error.wedged))
 		return false;
 
 	spin_lock_irqsave(&dev->event_lock, flags);
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index d6b06aa..9438bcd 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -1371,7 +1371,8 @@ static int ring_wait_for_space(struct intel_ring_buffer *ring, int n)
 
 		msleep(1);
 
-		ret = i915_gem_check_wedge(dev_priv, dev_priv->mm.interruptible);
+		ret = i915_gem_check_wedge(&dev_priv->gpu_error,
+					   dev_priv->mm.interruptible);
 		if (ret)
 			return ret;
 	} while (!time_after(jiffies, end));
@@ -1460,7 +1461,8 @@ int intel_ring_begin(struct intel_ring_buffer *ring,
 	drm_i915_private_t *dev_priv = ring->dev->dev_private;
 	int ret;
 
-	ret = i915_gem_check_wedge(dev_priv, dev_priv->mm.interruptible);
+	ret = i915_gem_check_wedge(&dev_priv->gpu_error,
+				   dev_priv->mm.interruptible);
 	if (ret)
 		return ret;
 
-- 
cgit v0.10.2


From 308887aad14c4ecc3fc10a3c58ec42641c5e4423 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Wed, 14 Nov 2012 17:14:06 +0100
Subject: drm/i915: fix reset handling in the throttle ioctl

While auditing the code I've noticed one place (the throttle ioctl)
which does not yet wait for the reset handler to complete and doesn't
properly decode the wedge state into -EAGAIN/-EIO. Fix this up by
calling the right helpers. This might explain the oddball "my
compositor just died in a successfull gpu reset" reports. Or maybe not, since
current mesa doesn't use this ioctl to throttle command submission.

The throttle ioctl doesn't take the struct_mutex, so to avoid busy-looping
with -EAGAIN while a reset is in process, check for errors first and wait
for the handler to complete if a reset is pending by calling
i915_gem_wait_for_error.

Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 04b2f92..c96a501 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -3390,8 +3390,13 @@ i915_gem_ring_throttle(struct drm_device *dev, struct drm_file *file)
 	u32 seqno = 0;
 	int ret;
 
-	if (atomic_read(&dev_priv->gpu_error.wedged))
-		return -EIO;
+	ret = i915_gem_wait_for_error(&dev_priv->gpu_error);
+	if (ret)
+		return ret;
+
+	ret = i915_gem_check_wedge(&dev_priv->gpu_error, false);
+	if (ret)
+		return ret;
 
 	spin_lock(&file_priv->mm.lock);
 	list_for_each_entry(request, &file_priv->mm.request_list, client_list) {
-- 
cgit v0.10.2


From 1f83fee08d625f8d0130f9fe5ef7b17c2e022f3c Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 15 Nov 2012 17:17:22 +0100
Subject: drm/i915: clear up wedged transitions

We have two important transitions of the wedged state in the current
code:

- 0 -> 1: This means a hang has been detected, and signals to everyone
  that they please get of any locks, so that the reset work item can
  do its job.

- 1 -> 0: The reset handler has completed.

Now the last transition mixes up two states: "Reset completed and
successful" and "Reset failed". To distinguish these two we do some
tricks with the reset completion, but I simply could not convince
myself that this doesn't race under odd circumstances.

Hence split this up, and add a new terminal state indicating that the
hw is gone for good.

Also add explicit #defines for both states, update comments.

v2: Split out the reset handling bugfix for the throttle ioctl.

v3: s/tmp/wedged/ sugested by Chris Wilson. Also fixup up a rebase
error which prevented this patch from actually compiling.

v4: To unify the wedged state with the reset counter, keep the
reset-in-progress state just as a flag. The terminally-wedged state is
now denoted with a big number.

v5: Add a comment to the reset_counter special values explaining that
WEDGED & RESET_IN_PROGRESS needs to be true for the code to be
correct.

v6: Fixup logic errors introduced with the wedged+reset_counter
unification. Since WEDGED implies reset-in-progress (in a way we're
terminally stuck in the dead-but-reset-not-completed state), we need
ensure that we check for this everywhere. The specific bug was in
wait_for_error, which would simply have timed out.

v7: Extract an inline i915_reset_in_progress helper to make the code
more readable. Also annote the reset-in-progress case with an
unlikely, to help the compiler optimize the fastpath. Do the same for
the terminally wedged case with i915_terminally_wedged.

Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index e1b7eaf..384f193 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -1672,7 +1672,7 @@ i915_wedged_read(struct file *filp,
 
 	len = snprintf(buf, sizeof(buf),
 		       "wedged :  %d\n",
-		       atomic_read(&dev_priv->gpu_error.wedged));
+		       atomic_read(&dev_priv->gpu_error.reset_counter));
 
 	if (len > sizeof(buf))
 		len = sizeof(buf);
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 62da6c7..c84743b 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -771,11 +771,37 @@ struct i915_gpu_error {
 	/* Protected by the above dev->gpu_error.lock. */
 	struct drm_i915_error_state *first_error;
 	struct work_struct work;
-	struct completion completion;
 
 	unsigned long last_reset;
 
-	atomic_t wedged;
+	/**
+	 * State variable controlling the reset flow
+	 *
+	 * Upper bits are for the reset counter.
+	 *
+	 * Lowest bit controls the reset state machine: Set means a reset is in
+	 * progress. This state will (presuming we don't have any bugs) decay
+	 * into either unset (successful reset) or the special WEDGED value (hw
+	 * terminally sour). All waiters on the reset_queue will be woken when
+	 * that happens.
+	 */
+	atomic_t reset_counter;
+
+	/**
+	 * Special values/flags for reset_counter
+	 *
+	 * Note that the code relies on
+	 * 	I915_WEDGED & I915_RESET_IN_PROGRESS_FLAG
+	 * being true.
+	 */
+#define I915_RESET_IN_PROGRESS_FLAG	1
+#define I915_WEDGED			0xffffffff
+
+	/**
+	 * Waitqueue to signal when the reset has completed. Used by clients
+	 * that wait for dev_priv->mm.wedged to settle.
+	 */
+	wait_queue_head_t reset_queue;
 
 	/* For gpu hang simulation. */
 	unsigned int stop_rings;
@@ -1543,6 +1569,16 @@ void i915_gem_retire_requests(struct drm_device *dev);
 void i915_gem_retire_requests_ring(struct intel_ring_buffer *ring);
 int __must_check i915_gem_check_wedge(struct i915_gpu_error *error,
 				      bool interruptible);
+static inline bool i915_reset_in_progress(struct i915_gpu_error *error)
+{
+	return unlikely(atomic_read(&error->reset_counter)
+			& I915_RESET_IN_PROGRESS_FLAG);
+}
+
+static inline bool i915_terminally_wedged(struct i915_gpu_error *error)
+{
+	return atomic_read(&error->reset_counter) == I915_WEDGED;
+}
 
 void i915_gem_reset(struct drm_device *dev);
 void i915_gem_clflush_object(struct drm_i915_gem_object *obj);
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index c96a501..2ca9011 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -89,36 +89,32 @@ static void i915_gem_info_remove_obj(struct drm_i915_private *dev_priv,
 static int
 i915_gem_wait_for_error(struct i915_gpu_error *error)
 {
-	struct completion *x = &error->completion;
-	unsigned long flags;
 	int ret;
 
-	if (!atomic_read(&error->wedged))
+#define EXIT_COND (!i915_reset_in_progress(error))
+	if (EXIT_COND)
 		return 0;
 
+	/* GPU is already declared terminally dead, give up. */
+	if (i915_terminally_wedged(error))
+		return -EIO;
+
 	/*
 	 * Only wait 10 seconds for the gpu reset to complete to avoid hanging
 	 * userspace. If it takes that long something really bad is going on and
 	 * we should simply try to bail out and fail as gracefully as possible.
 	 */
-	ret = wait_for_completion_interruptible_timeout(x, 10*HZ);
+	ret = wait_event_interruptible_timeout(error->reset_queue,
+					       EXIT_COND,
+					       10*HZ);
 	if (ret == 0) {
 		DRM_ERROR("Timed out waiting for the gpu reset to complete\n");
 		return -EIO;
 	} else if (ret < 0) {
 		return ret;
 	}
+#undef EXIT_COND
 
-	if (atomic_read(&error->wedged)) {
-		/* GPU is hung, bump the completion count to account for
-		 * the token we just consumed so that we never hit zero and
-		 * end up waiting upon a subsequent completion event that
-		 * will never happen.
-		 */
-		spin_lock_irqsave(&x->wait.lock, flags);
-		x->done++;
-		spin_unlock_irqrestore(&x->wait.lock, flags);
-	}
 	return 0;
 }
 
@@ -942,23 +938,14 @@ int
 i915_gem_check_wedge(struct i915_gpu_error *error,
 		     bool interruptible)
 {
-	if (atomic_read(&error->wedged)) {
-		struct completion *x = &error->completion;
-		bool recovery_complete;
-		unsigned long flags;
-
-		/* Give the error handler a chance to run. */
-		spin_lock_irqsave(&x->wait.lock, flags);
-		recovery_complete = x->done > 0;
-		spin_unlock_irqrestore(&x->wait.lock, flags);
-
+	if (i915_reset_in_progress(error)) {
 		/* Non-interruptible callers can't handle -EAGAIN, hence return
 		 * -EIO unconditionally for these. */
 		if (!interruptible)
 			return -EIO;
 
-		/* Recovery complete, but still wedged means reset failure. */
-		if (recovery_complete)
+		/* Recovery complete, but the reset failed ... */
+		if (i915_terminally_wedged(error))
 			return -EIO;
 
 		return -EAGAIN;
@@ -1025,7 +1012,7 @@ static int __wait_seqno(struct intel_ring_buffer *ring, u32 seqno,
 
 #define EXIT_COND \
 	(i915_seqno_passed(ring->get_seqno(ring, false), seqno) || \
-	atomic_read(&dev_priv->gpu_error.wedged))
+	i915_reset_in_progress(&dev_priv->gpu_error))
 	do {
 		if (interruptible)
 			end = wait_event_interruptible_timeout(ring->irq_queue,
@@ -1379,7 +1366,7 @@ out:
 		/* If this -EIO is due to a gpu hang, give the reset code a
 		 * chance to clean up the mess. Otherwise return the proper
 		 * SIGBUS. */
-		if (!atomic_read(&dev_priv->gpu_error.wedged))
+		if (i915_terminally_wedged(&dev_priv->gpu_error))
 			return VM_FAULT_SIGBUS;
 	case -EAGAIN:
 		/* Give the error handler a chance to run and move the
@@ -3983,9 +3970,9 @@ i915_gem_entervt_ioctl(struct drm_device *dev, void *data,
 	if (drm_core_check_feature(dev, DRIVER_MODESET))
 		return 0;
 
-	if (atomic_read(&dev_priv->gpu_error.wedged)) {
+	if (i915_reset_in_progress(&dev_priv->gpu_error)) {
 		DRM_ERROR("Reenabling wedged hardware, good luck\n");
-		atomic_set(&dev_priv->gpu_error.wedged, 0);
+		atomic_set(&dev_priv->gpu_error.reset_counter, 0);
 	}
 
 	mutex_lock(&dev->struct_mutex);
@@ -4069,7 +4056,7 @@ i915_gem_load(struct drm_device *dev)
 		INIT_LIST_HEAD(&dev_priv->fence_regs[i].lru_list);
 	INIT_DELAYED_WORK(&dev_priv->mm.retire_work,
 			  i915_gem_retire_work_handler);
-	init_completion(&dev_priv->gpu_error.completion);
+	init_waitqueue_head(&dev_priv->gpu_error.reset_queue);
 
 	/* On GEN3 we really need to make sure the ARB C3 LP bit is set */
 	if (IS_GEN3(dev)) {
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index f2c0016..4562c54 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -862,8 +862,10 @@ done:
  */
 static void i915_error_work_func(struct work_struct *work)
 {
-	drm_i915_private_t *dev_priv = container_of(work, drm_i915_private_t,
-						    gpu_error.work);
+	struct i915_gpu_error *error = container_of(work, struct i915_gpu_error,
+						    work);
+	drm_i915_private_t *dev_priv = container_of(error, drm_i915_private_t,
+						    gpu_error);
 	struct drm_device *dev = dev_priv->dev;
 	char *error_event[] = { "ERROR=1", NULL };
 	char *reset_event[] = { "RESET=1", NULL };
@@ -871,14 +873,18 @@ static void i915_error_work_func(struct work_struct *work)
 
 	kobject_uevent_env(&dev->primary->kdev.kobj, KOBJ_CHANGE, error_event);
 
-	if (atomic_read(&dev_priv->gpu_error.wedged)) {
+	if (i915_reset_in_progress(error)) {
 		DRM_DEBUG_DRIVER("resetting chip\n");
 		kobject_uevent_env(&dev->primary->kdev.kobj, KOBJ_CHANGE, reset_event);
+
 		if (!i915_reset(dev)) {
-			atomic_set(&dev_priv->gpu_error.wedged, 0);
+			atomic_set(&error->reset_counter, 0);
 			kobject_uevent_env(&dev->primary->kdev.kobj, KOBJ_CHANGE, reset_done_event);
+		} else {
+			atomic_set(&error->reset_counter, I915_WEDGED);
 		}
-		complete_all(&dev_priv->gpu_error.completion);
+
+		wake_up_all(&dev_priv->gpu_error.reset_queue);
 	}
 }
 
@@ -1482,11 +1488,12 @@ void i915_handle_error(struct drm_device *dev, bool wedged)
 	i915_report_and_clear_eir(dev);
 
 	if (wedged) {
-		INIT_COMPLETION(dev_priv->gpu_error.completion);
-		atomic_set(&dev_priv->gpu_error.wedged, 1);
+		atomic_set(&dev_priv->gpu_error.reset_counter,
+			   I915_RESET_IN_PROGRESS_FLAG);
 
 		/*
-		 * Wakeup waiting processes so they don't hang
+		 * Wakeup waiting processes so that the reset work item
+		 * doesn't deadlock trying to grab various locks.
 		 */
 		for_each_ring(ring, dev_priv, i)
 			wake_up_all(&ring->irq_queue);
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 160aa5f..7725446 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -2223,7 +2223,7 @@ intel_finish_fb(struct drm_framebuffer *old_fb)
 	WARN_ON(waitqueue_active(&dev_priv->pending_flip_queue));
 
 	wait_event(dev_priv->pending_flip_queue,
-		   atomic_read(&dev_priv->gpu_error.wedged) ||
+		   i915_reset_in_progress(&dev_priv->gpu_error) ||
 		   atomic_read(&obj->pending_flip) == 0);
 
 	/* Big Hammer, we also need to ensure that any pending
@@ -2871,7 +2871,7 @@ static bool intel_crtc_has_pending_flip(struct drm_crtc *crtc)
 	unsigned long flags;
 	bool pending;
 
-	if (atomic_read(&dev_priv->gpu_error.wedged))
+	if (i915_reset_in_progress(&dev_priv->gpu_error))
 		return false;
 
 	spin_lock_irqsave(&dev->event_lock, flags);
-- 
cgit v0.10.2


From d0a57789d5ec807fc218151b2fb2de4da30fbef5 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Tue, 9 Oct 2012 19:24:37 +0100
Subject: drm/i915: Only insert the mb() before updating the fence parameter

With a fence, we only need to insert a memory barrier around the actual
fence alteration for CPU accesses through the GTT. Performing the
barrier in flush-fence was inserting unnecessary and expensive barriers
for never fenced objects.

Note removing the barriers from flush-fence, which was effectively a
barrier before every direct access through the GTT, revealed that we
where missing a barrier before the first access through the GTT. Lack of
that barrier was sufficient to cause GPU hangs.

v2: Add a couple more comments to explain the new barriers

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 2ca9011..ce70655 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -2611,9 +2611,22 @@ static void i830_write_fence_reg(struct drm_device *dev, int reg,
 	POSTING_READ(FENCE_REG_830_0 + reg * 4);
 }
 
+inline static bool i915_gem_object_needs_mb(struct drm_i915_gem_object *obj)
+{
+	return obj && obj->base.read_domains & I915_GEM_DOMAIN_GTT;
+}
+
 static void i915_gem_write_fence(struct drm_device *dev, int reg,
 				 struct drm_i915_gem_object *obj)
 {
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	/* Ensure that all CPU reads are completed before installing a fence
+	 * and all writes before removing the fence.
+	 */
+	if (i915_gem_object_needs_mb(dev_priv->fence_regs[reg].obj))
+		mb();
+
 	switch (INTEL_INFO(dev)->gen) {
 	case 7:
 	case 6:
@@ -2623,6 +2636,12 @@ static void i915_gem_write_fence(struct drm_device *dev, int reg,
 	case 2: i830_write_fence_reg(dev, reg, obj); break;
 	default: BUG();
 	}
+
+	/* And similarly be paranoid that no direct access to this region
+	 * is reordered to before the fence is installed.
+	 */
+	if (i915_gem_object_needs_mb(obj))
+		mb();
 }
 
 static inline int fence_number(struct drm_i915_private *dev_priv,
@@ -2652,7 +2671,7 @@ static void i915_gem_object_update_fence(struct drm_i915_gem_object *obj,
 }
 
 static int
-i915_gem_object_flush_fence(struct drm_i915_gem_object *obj)
+i915_gem_object_wait_fence(struct drm_i915_gem_object *obj)
 {
 	if (obj->last_fenced_seqno) {
 		int ret = i915_wait_seqno(obj->ring, obj->last_fenced_seqno);
@@ -2662,12 +2681,6 @@ i915_gem_object_flush_fence(struct drm_i915_gem_object *obj)
 		obj->last_fenced_seqno = 0;
 	}
 
-	/* Ensure that all CPU reads are completed before installing a fence
-	 * and all writes before removing the fence.
-	 */
-	if (obj->base.read_domains & I915_GEM_DOMAIN_GTT)
-		mb();
-
 	obj->fenced_gpu_access = false;
 	return 0;
 }
@@ -2678,7 +2691,7 @@ i915_gem_object_put_fence(struct drm_i915_gem_object *obj)
 	struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
 	int ret;
 
-	ret = i915_gem_object_flush_fence(obj);
+	ret = i915_gem_object_wait_fence(obj);
 	if (ret)
 		return ret;
 
@@ -2752,7 +2765,7 @@ i915_gem_object_get_fence(struct drm_i915_gem_object *obj)
 	 * will need to serialise the write to the associated fence register?
 	 */
 	if (obj->fence_dirty) {
-		ret = i915_gem_object_flush_fence(obj);
+		ret = i915_gem_object_wait_fence(obj);
 		if (ret)
 			return ret;
 	}
@@ -2773,7 +2786,7 @@ i915_gem_object_get_fence(struct drm_i915_gem_object *obj)
 		if (reg->obj) {
 			struct drm_i915_gem_object *old = reg->obj;
 
-			ret = i915_gem_object_flush_fence(old);
+			ret = i915_gem_object_wait_fence(old);
 			if (ret)
 				return ret;
 
@@ -3068,6 +3081,13 @@ i915_gem_object_set_to_gtt_domain(struct drm_i915_gem_object *obj, bool write)
 
 	i915_gem_object_flush_cpu_write_domain(obj);
 
+	/* Serialise direct access to this object with the barriers for
+	 * coherent writes from the GPU, by effectively invalidating the
+	 * GTT domain upon first access.
+	 */
+	if ((obj->base.read_domains & I915_GEM_DOMAIN_GTT) == 0)
+		mb();
+
 	old_write_domain = obj->base.write_domain;
 	old_read_domains = obj->base.read_domains;
 
-- 
cgit v0.10.2


From 97c809fd9cf5e914322b53773ad0d67efe503fde Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Tue, 9 Oct 2012 19:24:38 +0100
Subject: drm/i915: Only apply the mb() when flushing the GTT domain during a
 finish

Now that we seem to have brought order to the GTT barriers, the last one
to review is the terminal barrier before we unbind the buffer from the
GTT. This needs to only be performed if the buffer still resides in the
GTT domain, and so we can skip some needless barriers otherwise.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index ce70655..5bb370f 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -2407,15 +2407,15 @@ static void i915_gem_object_finish_gtt(struct drm_i915_gem_object *obj)
 {
 	u32 old_write_domain, old_read_domains;
 
-	/* Act a barrier for all accesses through the GTT */
-	mb();
-
 	/* Force a pagefault for domain tracking on next user access */
 	i915_gem_release_mmap(obj);
 
 	if ((obj->base.read_domains & I915_GEM_DOMAIN_GTT) == 0)
 		return;
 
+	/* Wait for any direct GTT access to complete */
+	mb();
+
 	old_read_domains = obj->base.read_domains;
 	old_write_domain = obj->base.write_domain;
 
-- 
cgit v0.10.2


From 8faf6b18a2a6bece008de1e6bb80f0c608e58483 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sat, 1 Dec 2012 23:43:11 +0100
Subject: drm: review locking rules in drm_crtc.c

- config_cleanup was confused: It claimed that callers need to hold
  the modeset lock, but the connector|encoder_cleanup helpers grabbed
  that themselves (note that crtc_cleanup did _not_ grab the modeset
  lock). Which resulted in all drivers _not_ hodling the lock. Since
  this is for single-threaded cleanup code, drop the requirement from
  docs and also drop the lock_grabbing from all _cleanup functions.

- Kill the LOCKING section in the doctype, since clearly we're not
  good enough to keep them up-to-date. And misleading locking
  documentation is worse than useless (see e.g. the comment in the
  vmgfx driver about the cleanup mess). And since for most functions
  the very first line either grabs the lock or has a WARN_ON(!locked)
  the documentation doesn't really add anything.

- Instead put in some effort into explaining the only two special
  cases a bit better: config_init and config_cleanup are both called
  from single-threaded setup/teardown code, so don't do any locking.
  It's the driver's job though to enforce this.

- Where lacking, add a WARN_ON(!is_locked). Not many places though,
  since locking around fbdev setup/teardown is through-roughly screwed
  up, and so will break almost every single WARN annotation I've tried
  to add.

- Add a drm_modeset_is_locked helper - the Grate Modset Locking Rework
  will use the compiler to assist in the big reorg by renaming the
  mode lock, so start encapsulating things. Unfortunately this ended
  up in the "wrong" header file since it needs the definition of
  struct drm_device.

v2: Drop most WARNS again - we hit them all over the place, mostly in
the setup and teardown sequences. And trying to fix it up leads to
nice deadlocks, since the locking in the setup code is really
inconsistent.

Reviewed-by: Rob Clark <rob@ti.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index f2d667b..b970e41 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -208,8 +208,6 @@ char *drm_get_connector_status_name(enum drm_connector_status status)
  * @ptr: object pointer, used to generate unique ID
  * @type: object type
  *
- * LOCKING:
- *
  * Create a unique identifier based on @ptr in @dev's identifier space.  Used
  * for tracking modes, CRTCs and connectors.
  *
@@ -247,9 +245,6 @@ again:
  * @dev: DRM device
  * @id: ID to free
  *
- * LOCKING:
- * Caller must hold DRM mode_config lock.
- *
  * Free @id from @dev's unique identifier pool.
  */
 static void drm_mode_object_put(struct drm_device *dev,
@@ -279,9 +274,6 @@ EXPORT_SYMBOL(drm_mode_object_find);
  * drm_framebuffer_init - initialize a framebuffer
  * @dev: DRM device
  *
- * LOCKING:
- * Caller must hold mode config lock.
- *
  * Allocates an ID for the framebuffer's parent mode object, sets its mode
  * functions & device file and adds it to the master fd list.
  *
@@ -317,15 +309,12 @@ static void drm_framebuffer_free(struct kref *kref)
 
 /**
  * drm_framebuffer_unreference - unref a framebuffer
- *
- * LOCKING:
- * Caller must hold mode config lock.
  */
 void drm_framebuffer_unreference(struct drm_framebuffer *fb)
 {
 	struct drm_device *dev = fb->dev;
 	DRM_DEBUG("FB ID: %d\n", fb->base.id);
-	WARN_ON(!mutex_is_locked(&dev->mode_config.mutex));
+	WARN_ON(!drm_modeset_is_locked(dev));
 	kref_put(&fb->refcount, drm_framebuffer_free);
 }
 EXPORT_SYMBOL(drm_framebuffer_unreference);
@@ -344,15 +333,13 @@ EXPORT_SYMBOL(drm_framebuffer_reference);
  * drm_framebuffer_cleanup - remove a framebuffer object
  * @fb: framebuffer to remove
  *
- * LOCKING:
- * Caller must hold mode config lock.
- *
  * Scans all the CRTCs in @dev's mode_config.  If they're using @fb, removes
  * it, setting it to NULL.
  */
 void drm_framebuffer_cleanup(struct drm_framebuffer *fb)
 {
 	struct drm_device *dev = fb->dev;
+
 	/*
 	 * This could be moved to drm_framebuffer_remove(), but for
 	 * debugging is nice to keep around the list of fb's that are
@@ -370,9 +357,6 @@ EXPORT_SYMBOL(drm_framebuffer_cleanup);
  * drm_framebuffer_remove - remove and unreference a framebuffer object
  * @fb: framebuffer to remove
  *
- * LOCKING:
- * Caller must hold mode config lock.
- *
  * Scans all the CRTCs and planes in @dev's mode_config.  If they're
  * using @fb, removes it, setting it to NULL.
  */
@@ -384,6 +368,8 @@ void drm_framebuffer_remove(struct drm_framebuffer *fb)
 	struct drm_mode_set set;
 	int ret;
 
+	WARN_ON(!drm_modeset_is_locked(dev));
+
 	/* remove from any CRTC */
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
 		if (crtc->fb == fb) {
@@ -421,9 +407,6 @@ EXPORT_SYMBOL(drm_framebuffer_remove);
  * @crtc: CRTC object to init
  * @funcs: callbacks for the new CRTC
  *
- * LOCKING:
- * Takes mode_config lock.
- *
  * Inits a new object created as base part of an driver crtc object.
  *
  * RETURNS:
@@ -460,9 +443,6 @@ EXPORT_SYMBOL(drm_crtc_init);
  * drm_crtc_cleanup - Cleans up the core crtc usage.
  * @crtc: CRTC to cleanup
  *
- * LOCKING:
- * Caller must hold mode config lock.
- *
  * Cleanup @crtc. Removes from drm modesetting space
  * does NOT free object, caller does that.
  */
@@ -484,9 +464,6 @@ EXPORT_SYMBOL(drm_crtc_cleanup);
  * @connector: connector the new mode
  * @mode: mode data
  *
- * LOCKING:
- * Caller must hold mode config lock.
- *
  * Add @mode to @connector's mode list for later use.
  */
 void drm_mode_probed_add(struct drm_connector *connector,
@@ -501,9 +478,6 @@ EXPORT_SYMBOL(drm_mode_probed_add);
  * @connector: connector list to modify
  * @mode: mode to remove
  *
- * LOCKING:
- * Caller must hold mode config lock.
- *
  * Remove @mode from @connector's mode list, then free it.
  */
 void drm_mode_remove(struct drm_connector *connector,
@@ -521,9 +495,6 @@ EXPORT_SYMBOL(drm_mode_remove);
  * @funcs: callbacks for this connector
  * @name: user visible name of the connector
  *
- * LOCKING:
- * Takes mode config lock.
- *
  * Initialises a preallocated connector. Connectors should be
  * subclassed as part of driver connector objects.
  *
@@ -577,9 +548,6 @@ EXPORT_SYMBOL(drm_connector_init);
  * drm_connector_cleanup - cleans up an initialised connector
  * @connector: connector to cleanup
  *
- * LOCKING:
- * Takes mode config lock.
- *
  * Cleans up the connector but doesn't free the object.
  */
 void drm_connector_cleanup(struct drm_connector *connector)
@@ -596,11 +564,9 @@ void drm_connector_cleanup(struct drm_connector *connector)
 	list_for_each_entry_safe(mode, t, &connector->user_modes, head)
 		drm_mode_remove(connector, mode);
 
-	mutex_lock(&dev->mode_config.mutex);
 	drm_mode_object_put(dev, &connector->base);
 	list_del(&connector->head);
 	dev->mode_config.num_connector--;
-	mutex_unlock(&dev->mode_config.mutex);
 }
 EXPORT_SYMBOL(drm_connector_cleanup);
 
@@ -721,9 +687,6 @@ EXPORT_SYMBOL(drm_plane_cleanup);
  * drm_mode_create - create a new display mode
  * @dev: DRM device
  *
- * LOCKING:
- * Caller must hold DRM mode_config lock.
- *
  * Create a new drm_display_mode, give it an ID, and return it.
  *
  * RETURNS:
@@ -751,9 +714,6 @@ EXPORT_SYMBOL(drm_mode_create);
  * @dev: DRM device
  * @mode: mode to remove
  *
- * LOCKING:
- * Caller must hold mode config lock.
- *
  * Free @mode's unique identifier, then free it.
  */
 void drm_mode_destroy(struct drm_device *dev, struct drm_display_mode *mode)
@@ -978,11 +938,13 @@ EXPORT_SYMBOL(drm_mode_create_dirty_info_property);
  * drm_mode_config_init - initialize DRM mode_configuration structure
  * @dev: DRM device
  *
- * LOCKING:
- * None, should happen single threaded at init time.
- *
  * Initialize @dev's mode_config structure, used for tracking the graphics
  * configuration of @dev.
+ *
+ * Since this initializes the modeset locks, no locking is possible. Which is no
+ * problem, since this should happen single threaded at init time. It is the
+ * driver's problem to ensure this guarantee.
+ *
  */
 void drm_mode_config_init(struct drm_device *dev)
 {
@@ -1057,12 +1019,13 @@ EXPORT_SYMBOL(drm_mode_group_init_legacy_group);
  * drm_mode_config_cleanup - free up DRM mode_config info
  * @dev: DRM device
  *
- * LOCKING:
- * Caller must hold mode config lock.
- *
  * Free up all the connectors and CRTCs associated with this DRM device, then
  * free up the framebuffers and associated buffer objects.
  *
+ * Note that since this /should/ happen single-threaded at driver/device
+ * teardown time, no locking is required. It's the driver's job to ensure that
+ * this guarantee actually holds true.
+ *
  * FIXME: cleanup any dangling user buffer objects too
  */
 void drm_mode_config_cleanup(struct drm_device *dev)
@@ -1112,9 +1075,6 @@ EXPORT_SYMBOL(drm_mode_config_cleanup);
  * @out: drm_mode_modeinfo struct to return to the user
  * @in: drm_display_mode to use
  *
- * LOCKING:
- * None.
- *
  * Convert a drm_display_mode into a drm_mode_modeinfo structure to return to
  * the user.
  */
@@ -1151,9 +1111,6 @@ static void drm_crtc_convert_to_umode(struct drm_mode_modeinfo *out,
  * @out: drm_display_mode to return to the user
  * @in: drm_mode_modeinfo to use
  *
- * LOCKING:
- * None.
- *
  * Convert a drm_mode_modeinfo into a drm_display_mode structure to return to
  * the caller.
  *
@@ -1193,9 +1150,6 @@ static int drm_crtc_convert_umode(struct drm_display_mode *out,
  * @cmd: cmd from ioctl
  * @arg: arg from ioctl
  *
- * LOCKING:
- * Takes mode config lock.
- *
  * Construct a set of configuration description structures and return
  * them to the user, including CRTC, connector and framebuffer configuration.
  *
@@ -1381,9 +1335,6 @@ out:
  * @cmd: cmd from ioctl
  * @arg: arg from ioctl
  *
- * LOCKING:
- * Takes mode config lock.
- *
  * Construct a CRTC configuration structure to return to the user.
  *
  * Called by the user via ioctl.
@@ -1441,9 +1392,6 @@ out:
  * @cmd: cmd from ioctl
  * @arg: arg from ioctl
  *
- * LOCKING:
- * Takes mode config lock.
- *
  * Construct a connector configuration structure to return to the user.
  *
  * Called by the user via ioctl.
@@ -1618,9 +1566,6 @@ out:
  * @data: ioctl data
  * @file_priv: DRM file info
  *
- * LOCKING:
- * Takes mode config lock.
- *
  * Return an plane count and set of IDs.
  */
 int drm_mode_getplane_res(struct drm_device *dev, void *data,
@@ -1667,9 +1612,6 @@ out:
  * @data: ioctl data
  * @file_priv: DRM file info
  *
- * LOCKING:
- * Takes mode config lock.
- *
  * Return plane info, including formats supported, gamma size, any
  * current fb, etc.
  */
@@ -1735,9 +1677,6 @@ out:
  * @data: ioctl data*
  * @file_prive: DRM file info
  *
- * LOCKING:
- * Takes mode config lock.
- *
  * Set plane info, including placement, fb, scaling, and other factors.
  * Or pass a NULL fb to disable.
  */
@@ -1867,9 +1806,6 @@ out:
  * @cmd: cmd from ioctl
  * @arg: arg from ioctl
  *
- * LOCKING:
- * Takes mode config lock.
- *
  * Build a new CRTC configuration based on user request.
  *
  * Called by the user via ioctl.
@@ -2125,9 +2061,6 @@ EXPORT_SYMBOL(drm_mode_legacy_fb_format);
  * @cmd: cmd from ioctl
  * @arg: arg from ioctl
  *
- * LOCKING:
- * Takes mode config lock.
- *
  * Add a new FB to the specified CRTC, given a user request.
  *
  * Called by the user via ioctl.
@@ -2309,9 +2242,6 @@ static int framebuffer_check(const struct drm_mode_fb_cmd2 *r)
  * @cmd: cmd from ioctl
  * @arg: arg from ioctl
  *
- * LOCKING:
- * Takes mode config lock.
- *
  * Add a new FB to the specified CRTC, given a user request with format.
  *
  * Called by the user via ioctl.
@@ -2375,9 +2305,6 @@ out:
  * @cmd: cmd from ioctl
  * @arg: arg from ioctl
  *
- * LOCKING:
- * Takes mode config lock.
- *
  * Remove the FB specified by the user.
  *
  * Called by the user via ioctl.
@@ -2430,9 +2357,6 @@ out:
  * @cmd: cmd from ioctl
  * @arg: arg from ioctl
  *
- * LOCKING:
- * Takes mode config lock.
- *
  * Lookup the FB given its ID and return info about it.
  *
  * Called by the user via ioctl.
@@ -2549,9 +2473,6 @@ out_err1:
  * drm_fb_release - remove and free the FBs on this file
  * @filp: file * from the ioctl
  *
- * LOCKING:
- * Takes mode config lock.
- *
  * Destroy all the FBs associated with @filp.
  *
  * Called by the user via ioctl.
diff --git a/include/drm/drmP.h b/include/drm/drmP.h
index fad21c9..3c609ab 100644
--- a/include/drm/drmP.h
+++ b/include/drm/drmP.h
@@ -1276,6 +1276,11 @@ static inline int drm_device_is_unplugged(struct drm_device *dev)
 	return ret;
 }
 
+static inline bool drm_modeset_is_locked(struct drm_device *dev)
+{
+	return mutex_is_locked(&dev->mode_config.mutex);
+}
+
 /******************************************************************/
 /** \name Internal function definitions */
 /*@{*/
-- 
cgit v0.10.2


From 065a50ed3ef75cb265e12e3e1b615db0835150bc Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sun, 2 Dec 2012 00:09:18 +0100
Subject: drm/doc: integrate drm_crtc.c kerneldoc

And do a quick pass to adjust them to the last few (years?) of changes
...

This time actually compile-tested ;-)

Reviewed-by: Rob Clark <rob@ti.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl
index 4ee2304..caab791 100644
--- a/Documentation/DocBook/drm.tmpl
+++ b/Documentation/DocBook/drm.tmpl
@@ -1609,6 +1609,10 @@ void intel_crt_init(struct drm_device *dev)
         make its properties available to applications.
       </para>
     </sect2>
+    <sect2>
+      <title>KMS API Functions</title>
+!Edrivers/gpu/drm/drm_crtc.c
+    </sect2>
   </sect1>
 
   <!-- Internals: kms helper functions -->
diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index b970e41..8154554 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -203,10 +203,10 @@ char *drm_get_connector_status_name(enum drm_connector_status status)
 }
 
 /**
- * drm_mode_object_get - allocate a new identifier
+ * drm_mode_object_get - allocate a new modeset identifier
  * @dev: DRM device
- * @ptr: object pointer, used to generate unique ID
- * @type: object type
+ * @obj: object pointer, used to generate unique ID
+ * @obj_type: object type
  *
  * Create a unique identifier based on @ptr in @dev's identifier space.  Used
  * for tracking modes, CRTCs and connectors.
@@ -241,9 +241,9 @@ again:
 }
 
 /**
- * drm_mode_object_put - free an identifer
+ * drm_mode_object_put - free a modeset identifer
  * @dev: DRM device
- * @id: ID to free
+ * @object: object to free
  *
  * Free @id from @dev's unique identifier pool.
  */
@@ -273,6 +273,8 @@ EXPORT_SYMBOL(drm_mode_object_find);
 /**
  * drm_framebuffer_init - initialize a framebuffer
  * @dev: DRM device
+ * @fb: framebuffer to be initialized
+ * @funcs: ... with these functions
  *
  * Allocates an ID for the framebuffer's parent mode object, sets its mode
  * functions & device file and adds it to the master fd list.
@@ -309,6 +311,9 @@ static void drm_framebuffer_free(struct kref *kref)
 
 /**
  * drm_framebuffer_unreference - unref a framebuffer
+ * @fb: framebuffer to unref
+ *
+ * This functions decrements the fb's refcount and frees it if it drops to zero.
  */
 void drm_framebuffer_unreference(struct drm_framebuffer *fb)
 {
@@ -321,6 +326,7 @@ EXPORT_SYMBOL(drm_framebuffer_unreference);
 
 /**
  * drm_framebuffer_reference - incr the fb refcnt
+ * @fb: framebuffer
  */
 void drm_framebuffer_reference(struct drm_framebuffer *fb)
 {
@@ -493,7 +499,7 @@ EXPORT_SYMBOL(drm_mode_remove);
  * @dev: DRM device
  * @connector: the connector to init
  * @funcs: callbacks for this connector
- * @name: user visible name of the connector
+ * @connector_type: user visible type of the connector
  *
  * Initialises a preallocated connector. Connectors should be
  * subclassed as part of driver connector objects.
@@ -1145,10 +1151,9 @@ static int drm_crtc_convert_umode(struct drm_display_mode *out,
 
 /**
  * drm_mode_getresources - get graphics configuration
- * @inode: inode from the ioctl
- * @filp: file * from the ioctl
- * @cmd: cmd from ioctl
- * @arg: arg from ioctl
+ * @dev: drm device for the ioctl
+ * @data: data pointer for the ioctl
+ * @file_priv: drm file for the ioctl call
  *
  * Construct a set of configuration description structures and return
  * them to the user, including CRTC, connector and framebuffer configuration.
@@ -1330,10 +1335,9 @@ out:
 
 /**
  * drm_mode_getcrtc - get CRTC configuration
- * @inode: inode from the ioctl
- * @filp: file * from the ioctl
- * @cmd: cmd from ioctl
- * @arg: arg from ioctl
+ * @dev: drm device for the ioctl
+ * @data: data pointer for the ioctl
+ * @file_priv: drm file for the ioctl call
  *
  * Construct a CRTC configuration structure to return to the user.
  *
@@ -1387,10 +1391,9 @@ out:
 
 /**
  * drm_mode_getconnector - get connector configuration
- * @inode: inode from the ioctl
- * @filp: file * from the ioctl
- * @cmd: cmd from ioctl
- * @arg: arg from ioctl
+ * @dev: drm device for the ioctl
+ * @data: data pointer for the ioctl
+ * @file_priv: drm file for the ioctl call
  *
  * Construct a connector configuration structure to return to the user.
  *
@@ -1675,7 +1678,7 @@ out:
  * drm_mode_setplane - set up or tear down an plane
  * @dev: DRM device
  * @data: ioctl data*
- * @file_prive: DRM file info
+ * @file_priv: DRM file info
  *
  * Set plane info, including placement, fb, scaling, and other factors.
  * Or pass a NULL fb to disable.
@@ -1801,10 +1804,9 @@ out:
 
 /**
  * drm_mode_setcrtc - set CRTC configuration
- * @inode: inode from the ioctl
- * @filp: file * from the ioctl
- * @cmd: cmd from ioctl
- * @arg: arg from ioctl
+ * @dev: drm device for the ioctl
+ * @data: data pointer for the ioctl
+ * @file_priv: drm file for the ioctl call
  *
  * Build a new CRTC configuration based on user request.
  *
@@ -2056,10 +2058,9 @@ EXPORT_SYMBOL(drm_mode_legacy_fb_format);
 
 /**
  * drm_mode_addfb - add an FB to the graphics configuration
- * @inode: inode from the ioctl
- * @filp: file * from the ioctl
- * @cmd: cmd from ioctl
- * @arg: arg from ioctl
+ * @dev: drm device for the ioctl
+ * @data: data pointer for the ioctl
+ * @file_priv: drm file for the ioctl call
  *
  * Add a new FB to the specified CRTC, given a user request.
  *
@@ -2237,10 +2238,9 @@ static int framebuffer_check(const struct drm_mode_fb_cmd2 *r)
 
 /**
  * drm_mode_addfb2 - add an FB to the graphics configuration
- * @inode: inode from the ioctl
- * @filp: file * from the ioctl
- * @cmd: cmd from ioctl
- * @arg: arg from ioctl
+ * @dev: drm device for the ioctl
+ * @data: data pointer for the ioctl
+ * @file_priv: drm file for the ioctl call
  *
  * Add a new FB to the specified CRTC, given a user request with format.
  *
@@ -2300,10 +2300,9 @@ out:
 
 /**
  * drm_mode_rmfb - remove an FB from the configuration
- * @inode: inode from the ioctl
- * @filp: file * from the ioctl
- * @cmd: cmd from ioctl
- * @arg: arg from ioctl
+ * @dev: drm device for the ioctl
+ * @data: data pointer for the ioctl
+ * @file_priv: drm file for the ioctl call
  *
  * Remove the FB specified by the user.
  *
@@ -2352,10 +2351,9 @@ out:
 
 /**
  * drm_mode_getfb - get FB info
- * @inode: inode from the ioctl
- * @filp: file * from the ioctl
- * @cmd: cmd from ioctl
- * @arg: arg from ioctl
+ * @dev: drm device for the ioctl
+ * @data: data pointer for the ioctl
+ * @file_priv: drm file for the ioctl call
  *
  * Lookup the FB given its ID and return info about it.
  *
@@ -2471,7 +2469,7 @@ out_err1:
 
 /**
  * drm_fb_release - remove and free the FBs on this file
- * @filp: file * from the ioctl
+ * @priv: drm file for the ioctl
  *
  * Destroy all the FBs associated with @filp.
  *
@@ -2581,10 +2579,9 @@ EXPORT_SYMBOL(drm_mode_detachmode_crtc);
 
 /**
  * drm_fb_attachmode - Attach a user mode to an connector
- * @inode: inode from the ioctl
- * @filp: file * from the ioctl
- * @cmd: cmd from ioctl
- * @arg: arg from ioctl
+ * @dev: drm device for the ioctl
+ * @data: data pointer for the ioctl
+ * @file_priv: drm file for the ioctl call
  *
  * This attaches a user specified mode to an connector.
  * Called by the user via ioctl.
@@ -2636,10 +2633,9 @@ out:
 
 /**
  * drm_fb_detachmode - Detach a user specified mode from an connector
- * @inode: inode from the ioctl
- * @filp: file * from the ioctl
- * @cmd: cmd from ioctl
- * @arg: arg from ioctl
+ * @dev: drm device for the ioctl
+ * @data: data pointer for the ioctl
+ * @file_priv: drm file for the ioctl call
  *
  * Called by the user via ioctl.
  *
-- 
cgit v0.10.2


From c7d73f6a8ad71f9d9f58c86981322c6e48093a4f Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 13 Dec 2012 23:38:38 +0100
Subject: drm/<drivers>: reorder framebuffer init sequence

With more fine-grained locking we can no longer rely on the big
mode_config lock to prevent concurrent access to mode resources
like framebuffers. Instead a framebuffer becomes accessible to
other threads as soon as it is added to the relevant lookup
structures. Hence it needs to be fully set up by the time drivers
call drm_framebuffer_init.

This patch here is the drivers part of that reorg. Nothing really fancy
going on safe for three special cases.

- exynos needs to be careful to properly unref all handles.
- nouveau gets a resource leak fixed for free: one of the error
  cases didn't cleanup the framebuffer, which is now moot since
  the framebuffer is only registered once it is fully set up.
- vmwgfx requires a slight reordering of operations, I'm hoping I didn't
  break anything (but it's refcount management only, so should be safe).

v2: Split out exynos, since it's a bit more hairy than expected.

v3: Drop bogus cirrus hunk noticed by Richard Wilbur.

v4: Split out vmwgfx since there's a small change in return values.

Reviewed-by: Rob Clark <rob@ti.com> (core + omapdrm)
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/ast/ast_main.c b/drivers/gpu/drm/ast/ast_main.c
index f668e6c..d5ba709 100644
--- a/drivers/gpu/drm/ast/ast_main.c
+++ b/drivers/gpu/drm/ast/ast_main.c
@@ -266,13 +266,13 @@ int ast_framebuffer_init(struct drm_device *dev,
 {
 	int ret;
 
+	drm_helper_mode_fill_fb_struct(&ast_fb->base, mode_cmd);
+	ast_fb->obj = obj;
 	ret = drm_framebuffer_init(dev, &ast_fb->base, &ast_fb_funcs);
 	if (ret) {
 		DRM_ERROR("framebuffer init failed %d\n", ret);
 		return ret;
 	}
-	drm_helper_mode_fill_fb_struct(&ast_fb->base, mode_cmd);
-	ast_fb->obj = obj;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/cirrus/cirrus_main.c b/drivers/gpu/drm/cirrus/cirrus_main.c
index 6a9b12e..364474c 100644
--- a/drivers/gpu/drm/cirrus/cirrus_main.c
+++ b/drivers/gpu/drm/cirrus/cirrus_main.c
@@ -42,13 +42,13 @@ int cirrus_framebuffer_init(struct drm_device *dev,
 {
 	int ret;
 
+	drm_helper_mode_fill_fb_struct(&gfb->base, mode_cmd);
+	gfb->obj = obj;
 	ret = drm_framebuffer_init(dev, &gfb->base, &cirrus_fb_funcs);
 	if (ret) {
 		DRM_ERROR("drm_framebuffer_init failed: %d\n", ret);
 		return ret;
 	}
-	drm_helper_mode_fill_fb_struct(&gfb->base, mode_cmd);
-	gfb->obj = obj;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/drm_fb_cma_helper.c b/drivers/gpu/drm/drm_fb_cma_helper.c
index fd9d0af..e1e0cb0 100644
--- a/drivers/gpu/drm/drm_fb_cma_helper.c
+++ b/drivers/gpu/drm/drm_fb_cma_helper.c
@@ -85,6 +85,11 @@ static struct drm_fb_cma *drm_fb_cma_alloc(struct drm_device *dev,
 	if (!fb_cma)
 		return ERR_PTR(-ENOMEM);
 
+	drm_helper_mode_fill_fb_struct(&fb_cma->fb, mode_cmd);
+
+	for (i = 0; i < num_planes; i++)
+		fb_cma->obj[i] = obj[i];
+
 	ret = drm_framebuffer_init(dev, &fb_cma->fb, &drm_fb_cma_funcs);
 	if (ret) {
 		dev_err(dev->dev, "Failed to initalize framebuffer: %d\n", ret);
@@ -92,11 +97,6 @@ static struct drm_fb_cma *drm_fb_cma_alloc(struct drm_device *dev,
 		return ERR_PTR(ret);
 	}
 
-	drm_helper_mode_fill_fb_struct(&fb_cma->fb, mode_cmd);
-
-	for (i = 0; i < num_planes; i++)
-		fb_cma->obj[i] = obj[i];
-
 	return fb_cma;
 }
 
diff --git a/drivers/gpu/drm/gma500/framebuffer.c b/drivers/gpu/drm/gma500/framebuffer.c
index afded54..38e7e75 100644
--- a/drivers/gpu/drm/gma500/framebuffer.c
+++ b/drivers/gpu/drm/gma500/framebuffer.c
@@ -260,13 +260,13 @@ static int psb_framebuffer_init(struct drm_device *dev,
 	default:
 		return -EINVAL;
 	}
+	drm_helper_mode_fill_fb_struct(&fb->base, mode_cmd);
+	fb->gtt = gt;
 	ret = drm_framebuffer_init(dev, &fb->base, &psb_fb_funcs);
 	if (ret) {
 		dev_err(dev->dev, "framebuffer init failed: %d\n", ret);
 		return ret;
 	}
-	drm_helper_mode_fill_fb_struct(&fb->base, mode_cmd);
-	fb->gtt = gt;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index da1ad9c..26fa6a7 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -8666,14 +8666,15 @@ int intel_framebuffer_init(struct drm_device *dev,
 	if (mode_cmd->offsets[0] != 0)
 		return -EINVAL;
 
+	drm_helper_mode_fill_fb_struct(&intel_fb->base, mode_cmd);
+	intel_fb->obj = obj;
+
 	ret = drm_framebuffer_init(dev, &intel_fb->base, &intel_fb_funcs);
 	if (ret) {
 		DRM_ERROR("framebuffer init failed %d\n", ret);
 		return ret;
 	}
 
-	drm_helper_mode_fill_fb_struct(&intel_fb->base, mode_cmd);
-	intel_fb->obj = obj;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/mgag200/mgag200_main.c b/drivers/gpu/drm/mgag200/mgag200_main.c
index 70dd3c5..266438a 100644
--- a/drivers/gpu/drm/mgag200/mgag200_main.c
+++ b/drivers/gpu/drm/mgag200/mgag200_main.c
@@ -40,13 +40,15 @@ int mgag200_framebuffer_init(struct drm_device *dev,
 			     struct drm_mode_fb_cmd2 *mode_cmd,
 			     struct drm_gem_object *obj)
 {
-	int ret = drm_framebuffer_init(dev, &gfb->base, &mga_fb_funcs);
+	int ret;
+	
+	drm_helper_mode_fill_fb_struct(&gfb->base, mode_cmd);
+	gfb->obj = obj;
+	ret = drm_framebuffer_init(dev, &gfb->base, &mga_fb_funcs);
 	if (ret) {
 		DRM_ERROR("drm_framebuffer_init failed: %d\n", ret);
 		return ret;
 	}
-	drm_helper_mode_fill_fb_struct(&gfb->base, mode_cmd);
-	gfb->obj = obj;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_display.c b/drivers/gpu/drm/nouveau/nouveau_display.c
index 508b00a..d42c9e8 100644
--- a/drivers/gpu/drm/nouveau/nouveau_display.c
+++ b/drivers/gpu/drm/nouveau/nouveau_display.c
@@ -78,11 +78,6 @@ nouveau_framebuffer_init(struct drm_device *dev,
 	struct drm_framebuffer *fb = &nv_fb->base;
 	int ret;
 
-	ret = drm_framebuffer_init(dev, fb, &nouveau_framebuffer_funcs);
-	if (ret) {
-		return ret;
-	}
-
 	drm_helper_mode_fill_fb_struct(fb, mode_cmd);
 	nv_fb->nvbo = nvbo;
 
@@ -125,6 +120,11 @@ nouveau_framebuffer_init(struct drm_device *dev,
 		}
 	}
 
+	ret = drm_framebuffer_init(dev, fb, &nouveau_framebuffer_funcs);
+	if (ret) {
+		return ret;
+	}
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/radeon/radeon_display.c b/drivers/gpu/drm/radeon/radeon_display.c
index 1da2386..12ec3ef 100644
--- a/drivers/gpu/drm/radeon/radeon_display.c
+++ b/drivers/gpu/drm/radeon/radeon_display.c
@@ -1089,12 +1089,12 @@ radeon_framebuffer_init(struct drm_device *dev,
 {
 	int ret;
 	rfb->obj = obj;
+	drm_helper_mode_fill_fb_struct(&rfb->base, mode_cmd);
 	ret = drm_framebuffer_init(dev, &rfb->base, &radeon_fb_funcs);
 	if (ret) {
 		rfb->obj = NULL;
 		return ret;
 	}
-	drm_helper_mode_fill_fb_struct(&rfb->base, mode_cmd);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/udl/udl_fb.c b/drivers/gpu/drm/udl/udl_fb.c
index d4ab3be..829c4a7 100644
--- a/drivers/gpu/drm/udl/udl_fb.c
+++ b/drivers/gpu/drm/udl/udl_fb.c
@@ -435,8 +435,8 @@ udl_framebuffer_init(struct drm_device *dev,
 	int ret;
 
 	ufb->obj = obj;
-	ret = drm_framebuffer_init(dev, &ufb->base, &udlfb_funcs);
 	drm_helper_mode_fill_fb_struct(&ufb->base, mode_cmd);
+	ret = drm_framebuffer_init(dev, &ufb->base, &udlfb_funcs);
 	return ret;
 }
 
diff --git a/drivers/staging/omapdrm/omap_fb.c b/drivers/staging/omapdrm/omap_fb.c
index 09028e9..bf6421f 100644
--- a/drivers/staging/omapdrm/omap_fb.c
+++ b/drivers/staging/omapdrm/omap_fb.c
@@ -424,14 +424,6 @@ struct drm_framebuffer *omap_framebuffer_init(struct drm_device *dev,
 	}
 
 	fb = &omap_fb->base;
-	ret = drm_framebuffer_init(dev, fb, &omap_framebuffer_funcs);
-	if (ret) {
-		dev_err(dev->dev, "framebuffer init failed: %d\n", ret);
-		goto fail;
-	}
-
-	DBG("create: FB ID: %d (%p)", fb->base.id, fb);
-
 	omap_fb->format = format;
 
 	for (i = 0; i < n; i++) {
@@ -462,6 +454,14 @@ struct drm_framebuffer *omap_framebuffer_init(struct drm_device *dev,
 
 	drm_helper_mode_fill_fb_struct(fb, mode_cmd);
 
+	ret = drm_framebuffer_init(dev, fb, &omap_framebuffer_funcs);
+	if (ret) {
+		dev_err(dev->dev, "framebuffer init failed: %d\n", ret);
+		goto fail;
+	}
+
+	DBG("create: FB ID: %d (%p)", fb->base.id, fb);
+
 	return fb;
 
 fail:
-- 
cgit v0.10.2


From 80f0b5aff8f49f63eaf08bdae243de3a8ea3f77e Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 13 Dec 2012 23:39:01 +0100
Subject: drm/vmwgfx: reorder framebuffer init sequence

vmwgfx has an oddity, when failing to reference the surface it'll
return 0, since that's what the successfull drm_framebuffer_init will
leave behind in ret. Fix this up by returning -EINVAL.

Split out from all the other driver updates due to the above tiny
semantic change. Shouldn't matter though since the reference grabbing
seemingly can't fail.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
index 5474394..edc9792 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
@@ -681,14 +681,10 @@ static int vmw_kms_new_framebuffer_surface(struct vmw_private *dev_priv,
 		goto out_err1;
 	}
 
-	ret = drm_framebuffer_init(dev, &vfbs->base.base,
-				   &vmw_framebuffer_surface_funcs);
-	if (ret)
-		goto out_err2;
-
 	if (!vmw_surface_reference(surface)) {
 		DRM_ERROR("failed to reference surface %p\n", surface);
-		goto out_err3;
+		ret = -EINVAL;
+		goto out_err2;
 	}
 
 	/* XXX get the first 3 from the surface info */
@@ -707,10 +703,15 @@ static int vmw_kms_new_framebuffer_surface(struct vmw_private *dev_priv,
 
 	*out = &vfbs->base;
 
+	ret = drm_framebuffer_init(dev, &vfbs->base.base,
+				   &vmw_framebuffer_surface_funcs);
+	if (ret)
+		goto out_err3;
+
 	return 0;
 
 out_err3:
-	drm_framebuffer_cleanup(&vfbs->base.base);
+	vmw_surface_unreference(&surface);
 out_err2:
 	kfree(vfbs);
 out_err1:
@@ -1053,14 +1054,10 @@ static int vmw_kms_new_framebuffer_dmabuf(struct vmw_private *dev_priv,
 		goto out_err1;
 	}
 
-	ret = drm_framebuffer_init(dev, &vfbd->base.base,
-				   &vmw_framebuffer_dmabuf_funcs);
-	if (ret)
-		goto out_err2;
-
 	if (!vmw_dmabuf_reference(dmabuf)) {
 		DRM_ERROR("failed to reference dmabuf %p\n", dmabuf);
-		goto out_err3;
+		ret = -EINVAL;
+		goto out_err2;
 	}
 
 	vfbd->base.base.bits_per_pixel = mode_cmd->bpp;
@@ -1077,10 +1074,15 @@ static int vmw_kms_new_framebuffer_dmabuf(struct vmw_private *dev_priv,
 	vfbd->base.user_handle = mode_cmd->handle;
 	*out = &vfbd->base;
 
+	ret = drm_framebuffer_init(dev, &vfbd->base.base,
+				   &vmw_framebuffer_dmabuf_funcs);
+	if (ret)
+		goto out_err3;
+
 	return 0;
 
 out_err3:
-	drm_framebuffer_cleanup(&vfbd->base.base);
+	vmw_dmabuf_unreference(&dmabuf);
 out_err2:
 	kfree(vfbd);
 out_err1:
-- 
cgit v0.10.2


From 7147573a5ce499dec3979e6b524691d47e1288d5 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sun, 2 Dec 2012 21:55:41 +0100
Subject: drm/gma500: move fbcon restore to lastclose

Doing this within the fb->destroy callback leads to a locking
nightmare. And all other drm drivers that restore the fbcon do
it in lastclose, too.

With this adjustments all fb->destroy callbacks optionally drop
references to any gem objects used as backing storage, call
drm_framebuffer_cleanup and then kfree the struct. Which nicely
simplifies the locking for framebuffer unreferencing and freeing,
since this doesn't require that we hold the mode_config lock. A
slight exception is the vmwgfx surface backed framebuffer, it also
calls drm_master_put and removes the object from a device-private
framebuffer list. Both seem to have solid locking in place already.

Conclusion is that now it is no longer required to hold the
mode_config lock while freeing a framebuffer.

v2: Drop the corresponding mutex_lock WARN check from
drm_framebuffer_unreference.

v3: Use just the mode_config lock not modeset_lock_all, due to patch
reordering.

Acked-by: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index 8154554..8d665fa 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -317,9 +317,7 @@ static void drm_framebuffer_free(struct kref *kref)
  */
 void drm_framebuffer_unreference(struct drm_framebuffer *fb)
 {
-	struct drm_device *dev = fb->dev;
 	DRM_DEBUG("FB ID: %d\n", fb->base.id);
-	WARN_ON(!drm_modeset_is_locked(dev));
 	kref_put(&fb->refcount, drm_framebuffer_free);
 }
 EXPORT_SYMBOL(drm_framebuffer_unreference);
diff --git a/drivers/gpu/drm/gma500/framebuffer.c b/drivers/gpu/drm/gma500/framebuffer.c
index 38e7e75..49800d2 100644
--- a/drivers/gpu/drm/gma500/framebuffer.c
+++ b/drivers/gpu/drm/gma500/framebuffer.c
@@ -668,30 +668,6 @@ static void psb_user_framebuffer_destroy(struct drm_framebuffer *fb)
 {
 	struct psb_framebuffer *psbfb = to_psb_fb(fb);
 	struct gtt_range *r = psbfb->gtt;
-	struct drm_device *dev = fb->dev;
-	struct drm_psb_private *dev_priv = dev->dev_private;
-	struct psb_fbdev *fbdev = dev_priv->fbdev;
-	struct drm_crtc *crtc;
-	int reset = 0;
-
-	/* Should never get stolen memory for a user fb */
-	WARN_ON(r->stolen);
-
-	/* Check if we are erroneously live */
-	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head)
-		if (crtc->fb == fb)
-			reset = 1;
-
-	if (reset)
-		/*
-		 * Now force a sane response before we permit the DRM CRTC
-		 * layer to do stupid things like blank the display. Instead
-		 * we reset this framebuffer as if the user had forced a reset.
-		 * We must do this before the cleanup so that the DRM layer
-		 * doesn't get a chance to stick its oar in where it isn't
-		 * wanted.
-		 */
-		drm_fb_helper_restore_fbdev_mode(&fbdev->psb_fb_helper);
 
 	/* Let DRM do its clean up */
 	drm_framebuffer_cleanup(fb);
diff --git a/drivers/gpu/drm/gma500/psb_drv.c b/drivers/gpu/drm/gma500/psb_drv.c
index dd1fbfa..dbcefe9 100644
--- a/drivers/gpu/drm/gma500/psb_drv.c
+++ b/drivers/gpu/drm/gma500/psb_drv.c
@@ -149,6 +149,16 @@ static struct drm_ioctl_desc psb_ioctls[] = {
 
 static void psb_lastclose(struct drm_device *dev)
 {
+	int ret;
+	struct drm_psb_private *dev_priv = dev->dev_private;
+	struct psb_fbdev *fbdev = dev_priv->fbdev;
+
+	mutex_lock(&dev->mode_config.mutex);
+	ret = drm_fb_helper_restore_fbdev_mode(&fbdev->psb_fb_helper);
+	if (ret)
+		DRM_DEBUG("failed to restore crtc mode\n");
+	mutex_unlock(&dev->mode_config.mutex);
+
 	return;
 }
 
-- 
cgit v0.10.2


From 59ad1465423d968f06f243bc52cc3b1a320a27cf Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sun, 2 Dec 2012 14:49:44 +0100
Subject: drm/nouveau: protect evo_wait/evo_kick sections with a channel mutex

With per-crtc locks modeset operations can run in parallel, and the
cursor code uses the device-global evo master channel for hw frobbing.
But the pageflip code can also sync with the master under some
circumstances. Hence just wrap things up in a mutex to ensure that
pushbuf access doesn't intermingle.

The approach here is a bit overkill since the per-crtc channels used
to schedule the pageflips could probably be used without this pushbuf
locking, but I'm not familiar enough with the nouveau codebase to be
sure of that.

v2: Add missing mutex_init to avoid angering lockdep.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/nouveau/nv50_display.c b/drivers/gpu/drm/nouveau/nv50_display.c
index 3587408..d4cbea1 100644
--- a/drivers/gpu/drm/nouveau/nv50_display.c
+++ b/drivers/gpu/drm/nouveau/nv50_display.c
@@ -128,6 +128,11 @@ struct nv50_dmac {
 	struct nv50_chan base;
 	dma_addr_t handle;
 	u32 *ptr;
+
+	/* Protects against concurrent pushbuf access to this channel, lock is
+	 * grabbed by evo_wait (if the pushbuf reservation is successful) and
+	 * dropped again by evo_kick. */
+	struct mutex lock;
 };
 
 static void
@@ -271,6 +276,8 @@ nv50_dmac_create(struct nouveau_object *core, u32 bclass, u8 head,
 	u32 pushbuf = *(u32 *)data;
 	int ret;
 
+	mutex_init(&dmac->lock);
+
 	dmac->ptr = pci_alloc_consistent(nv_device(core)->pdev, PAGE_SIZE,
 					&dmac->handle);
 	if (!dmac->ptr)
@@ -395,11 +402,13 @@ evo_wait(void *evoc, int nr)
 	struct nv50_dmac *dmac = evoc;
 	u32 put = nv_ro32(dmac->base.user, 0x0000) / 4;
 
+	mutex_lock(&dmac->lock);
 	if (put + nr >= (PAGE_SIZE / 4) - 8) {
 		dmac->ptr[put] = 0x20000000;
 
 		nv_wo32(dmac->base.user, 0x0000, 0x00000000);
 		if (!nv_wait(dmac->base.user, 0x0004, ~0, 0x00000000)) {
+			mutex_unlock(&dmac->lock);
 			NV_ERROR(dmac->base.user, "channel stalled\n");
 			return NULL;
 		}
@@ -415,6 +424,7 @@ evo_kick(u32 *push, void *evoc)
 {
 	struct nv50_dmac *dmac = evoc;
 	nv_wo32(dmac->base.user, 0x0000, (push - dmac->ptr) << 2);
+	mutex_unlock(&dmac->lock);
 }
 
 #define evo_mthd(p,m,s) *((p)++) = (((s) << 18) | (m))
-- 
cgit v0.10.2


From 0ae6d7bc0e70dafc1739d50b2b8d9d7c61968395 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Tue, 11 Dec 2012 21:52:30 +0100
Subject: drm/nouveau: try to protect nbo->pin_refcount

... by moving the bo_pin/bo_unpin manipulation of the pin_refcount
under the protection of the ttm reservation lock. pin/unpin seems
to get called from all over the place, so atm this is completely racy.

After this patch there are only a few places in cleanup functions
left which access ->pin_refcount without locking. But I'm hoping that
those are safe and some other code invariant guarantees that this
won't blow up.

In any case, I only need to fix up pin/unpin to make ->pageflip work
safely, so let's keep it at that.

Add a comment to the header to explain the new locking rule.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 69d7b1d..64d6e304 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -300,17 +300,18 @@ nouveau_bo_pin(struct nouveau_bo *nvbo, uint32_t memtype)
 	struct ttm_buffer_object *bo = &nvbo->bo;
 	int ret;
 
+	ret = ttm_bo_reserve(bo, false, false, false, 0);
+	if (ret)
+		goto out;
+
 	if (nvbo->pin_refcnt && !(memtype & (1 << bo->mem.mem_type))) {
 		NV_ERROR(drm, "bo %p pinned elsewhere: 0x%08x vs 0x%08x\n", bo,
 			 1 << bo->mem.mem_type, memtype);
-		return -EINVAL;
+		ret = -EINVAL;
+		goto out;
 	}
 
 	if (nvbo->pin_refcnt++)
-		return 0;
-
-	ret = ttm_bo_reserve(bo, false, false, false, 0);
-	if (ret)
 		goto out;
 
 	nouveau_bo_placement_set(nvbo, memtype, 0);
@@ -328,10 +329,8 @@ nouveau_bo_pin(struct nouveau_bo *nvbo, uint32_t memtype)
 			break;
 		}
 	}
-	ttm_bo_unreserve(bo);
 out:
-	if (unlikely(ret))
-		nvbo->pin_refcnt--;
+	ttm_bo_unreserve(bo);
 	return ret;
 }
 
@@ -342,13 +341,13 @@ nouveau_bo_unpin(struct nouveau_bo *nvbo)
 	struct ttm_buffer_object *bo = &nvbo->bo;
 	int ret;
 
-	if (--nvbo->pin_refcnt)
-		return 0;
-
 	ret = ttm_bo_reserve(bo, false, false, false, 0);
 	if (ret)
 		return ret;
 
+	if (--nvbo->pin_refcnt)
+		goto out;
+
 	nouveau_bo_placement_set(nvbo, bo->mem.placement, 0);
 
 	ret = nouveau_bo_validate(nvbo, false, false);
@@ -365,6 +364,7 @@ nouveau_bo_unpin(struct nouveau_bo *nvbo)
 		}
 	}
 
+out:
 	ttm_bo_unreserve(bo);
 	return ret;
 }
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.h b/drivers/gpu/drm/nouveau/nouveau_bo.h
index 25ca379..81d00fe 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.h
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.h
@@ -28,6 +28,8 @@ struct nouveau_bo {
 	struct nouveau_drm_tile *tile;
 
 	struct drm_gem_object *gem;
+
+	/* protect by the ttm reservation lock */
 	int pin_refcnt;
 
 	struct ttm_bo_kmap_obj dma_buf_vmap;
-- 
cgit v0.10.2


From af26ef3b3978349cfbd864163a6ebb4906b733b5 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 13 Dec 2012 23:07:50 +0100
Subject: drm/<drivers>: Unified handling of unimplemented fb->create_handle

Some drivers don't have real ->create_handle callbacks.

- cirrus/ast/mga200: Returns either 0 or -EINVAL.

- udl: Didn't even bother with a callback, leading to a nice
  userspace-triggerable OOPS.

- vmwgfx: This driver bothered with an implementation to return 0 as
  the handle (which is the canonical no-obj gem handle).

All have in common that ->create_handle doesn't really make too much
sense for them - that ioctl is used only for seamless fb takeover in
the radeon/nouveau/i915 ddx drivers. So allow drivers to not implement
this and return a consistent -ENODEV.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/ast/ast_main.c b/drivers/gpu/drm/ast/ast_main.c
index d5ba709..f60fd7b 100644
--- a/drivers/gpu/drm/ast/ast_main.c
+++ b/drivers/gpu/drm/ast/ast_main.c
@@ -246,16 +246,8 @@ static void ast_user_framebuffer_destroy(struct drm_framebuffer *fb)
 	kfree(fb);
 }
 
-static int ast_user_framebuffer_create_handle(struct drm_framebuffer *fb,
-					      struct drm_file *file,
-					      unsigned int *handle)
-{
-	return -EINVAL;
-}
-
 static const struct drm_framebuffer_funcs ast_fb_funcs = {
 	.destroy = ast_user_framebuffer_destroy,
-	.create_handle = ast_user_framebuffer_create_handle,
 };
 
 
diff --git a/drivers/gpu/drm/cirrus/cirrus_main.c b/drivers/gpu/drm/cirrus/cirrus_main.c
index 364474c..35cbae8 100644
--- a/drivers/gpu/drm/cirrus/cirrus_main.c
+++ b/drivers/gpu/drm/cirrus/cirrus_main.c
@@ -23,16 +23,8 @@ static void cirrus_user_framebuffer_destroy(struct drm_framebuffer *fb)
 	kfree(fb);
 }
 
-static int cirrus_user_framebuffer_create_handle(struct drm_framebuffer *fb,
-						 struct drm_file *file_priv,
-						 unsigned int *handle)
-{
-	return 0;
-}
-
 static const struct drm_framebuffer_funcs cirrus_fb_funcs = {
 	.destroy = cirrus_user_framebuffer_destroy,
-	.create_handle = cirrus_user_framebuffer_create_handle,
 };
 
 int cirrus_framebuffer_init(struct drm_device *dev,
diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index 8d665fa..a9abf49 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -2384,7 +2384,10 @@ int drm_mode_getfb(struct drm_device *dev,
 	r->depth = fb->depth;
 	r->bpp = fb->bits_per_pixel;
 	r->pitch = fb->pitches[0];
-	fb->funcs->create_handle(fb, file_priv, &r->handle);
+	if (fb->funcs->create_handle)
+		ret = fb->funcs->create_handle(fb, file_priv, &r->handle);
+	else
+		ret = -ENODEV;
 
 out:
 	mutex_unlock(&dev->mode_config.mutex);
diff --git a/drivers/gpu/drm/mgag200/mgag200_main.c b/drivers/gpu/drm/mgag200/mgag200_main.c
index 266438a..64297c7 100644
--- a/drivers/gpu/drm/mgag200/mgag200_main.c
+++ b/drivers/gpu/drm/mgag200/mgag200_main.c
@@ -23,16 +23,8 @@ static void mga_user_framebuffer_destroy(struct drm_framebuffer *fb)
 	kfree(fb);
 }
 
-static int mga_user_framebuffer_create_handle(struct drm_framebuffer *fb,
-						 struct drm_file *file_priv,
-						 unsigned int *handle)
-{
-	return 0;
-}
-
 static const struct drm_framebuffer_funcs mga_fb_funcs = {
 	.destroy = mga_user_framebuffer_destroy,
-	.create_handle = mga_user_framebuffer_create_handle,
 };
 
 int mgag200_framebuffer_init(struct drm_device *dev,
diff --git a/drivers/gpu/drm/udl/udl_fb.c b/drivers/gpu/drm/udl/udl_fb.c
index 829c4a7..f8904b4 100644
--- a/drivers/gpu/drm/udl/udl_fb.c
+++ b/drivers/gpu/drm/udl/udl_fb.c
@@ -422,7 +422,6 @@ static void udl_user_framebuffer_destroy(struct drm_framebuffer *fb)
 static const struct drm_framebuffer_funcs udlfb_funcs = {
 	.destroy = udl_user_framebuffer_destroy,
 	.dirty = udl_user_framebuffer_dirty,
-	.create_handle = NULL,
 };
 
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
index edc9792..9c0876b 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
@@ -373,16 +373,6 @@ void vmw_kms_cursor_post_execbuf(struct vmw_private *dev_priv)
  * Generic framebuffer code
  */
 
-int vmw_framebuffer_create_handle(struct drm_framebuffer *fb,
-				  struct drm_file *file_priv,
-				  unsigned int *handle)
-{
-	if (handle)
-		*handle = 0;
-
-	return 0;
-}
-
 /*
  * Surface framebuffer code
  */
@@ -610,7 +600,6 @@ int vmw_framebuffer_surface_dirty(struct drm_framebuffer *framebuffer,
 static struct drm_framebuffer_funcs vmw_framebuffer_surface_funcs = {
 	.destroy = vmw_framebuffer_surface_destroy,
 	.dirty = vmw_framebuffer_surface_dirty,
-	.create_handle = vmw_framebuffer_create_handle,
 };
 
 static int vmw_kms_new_framebuffer_surface(struct vmw_private *dev_priv,
@@ -961,7 +950,6 @@ int vmw_framebuffer_dmabuf_dirty(struct drm_framebuffer *framebuffer,
 static struct drm_framebuffer_funcs vmw_framebuffer_dmabuf_funcs = {
 	.destroy = vmw_framebuffer_dmabuf_destroy,
 	.dirty = vmw_framebuffer_dmabuf_dirty,
-	.create_handle = vmw_framebuffer_create_handle,
 };
 
 /**
-- 
cgit v0.10.2


From 2d13b6796e420ed00389b7399a5d5ac7b1fed436 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Tue, 11 Dec 2012 13:47:23 +0100
Subject: drm: encapsulate crtc->set_config calls

With refcounting we need to adjust framebuffer refcounts at each
callsite - much easier to do if they all call the same little helper
function.

Reviewed-by: Rob Clark <rob@ti.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index a9abf49..7ca2f28 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -381,7 +381,7 @@ void drm_framebuffer_remove(struct drm_framebuffer *fb)
 			memset(&set, 0, sizeof(struct drm_mode_set));
 			set.crtc = crtc;
 			set.fb = NULL;
-			ret = crtc->funcs->set_config(&set);
+			ret = drm_mode_set_config_internal(&set);
 			if (ret)
 				DRM_ERROR("failed to reset crtc %p when fb was deleted\n", crtc);
 		}
@@ -1801,6 +1801,21 @@ out:
 }
 
 /**
+ * drm_mode_set_config_internal - helper to call ->set_config
+ * @set: modeset config to set
+ *
+ * This is a little helper to wrap internal calls to the ->set_config driver
+ * interface. The only thing it adds is correct refcounting dance.
+ */
+int drm_mode_set_config_internal(struct drm_mode_set *set)
+{
+	struct drm_crtc *crtc = set->crtc;
+
+	return crtc->funcs->set_config(set);
+}
+EXPORT_SYMBOL(drm_mode_set_config_internal);
+
+/**
  * drm_mode_setcrtc - set CRTC configuration
  * @dev: drm device for the ioctl
  * @data: data pointer for the ioctl
@@ -1963,7 +1978,7 @@ int drm_mode_setcrtc(struct drm_device *dev, void *data,
 	set.connectors = connector_set;
 	set.num_connectors = crtc_req->count_connectors;
 	set.fb = fb;
-	ret = crtc->funcs->set_config(&set);
+	ret = drm_mode_set_config_internal(&set);
 
 out:
 	kfree(connector_set);
diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index 954d175..82c3a9f 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -245,7 +245,7 @@ bool drm_fb_helper_restore_fbdev_mode(struct drm_fb_helper *fb_helper)
 	int i, ret;
 	for (i = 0; i < fb_helper->crtc_count; i++) {
 		struct drm_mode_set *mode_set = &fb_helper->crtc_info[i].mode_set;
-		ret = mode_set->crtc->funcs->set_config(mode_set);
+		ret = drm_mode_set_config_internal(mode_set);
 		if (ret)
 			error = true;
 	}
@@ -675,7 +675,7 @@ int drm_fb_helper_set_par(struct fb_info *info)
 	mutex_lock(&dev->mode_config.mutex);
 	for (i = 0; i < fb_helper->crtc_count; i++) {
 		crtc = fb_helper->crtc_info[i].mode_set.crtc;
-		ret = crtc->funcs->set_config(&fb_helper->crtc_info[i].mode_set);
+		ret = drm_mode_set_config_internal(&fb_helper->crtc_info[i].mode_set);
 		if (ret) {
 			mutex_unlock(&dev->mode_config.mutex);
 			return ret;
@@ -711,7 +711,7 @@ int drm_fb_helper_pan_display(struct fb_var_screeninfo *var,
 		modeset->y = var->yoffset;
 
 		if (modeset->num_connectors) {
-			ret = crtc->funcs->set_config(modeset);
+			ret = drm_mode_set_config_internal(modeset);
 			if (!ret) {
 				info->var.xoffset = var->xoffset;
 				info->var.yoffset = var->yoffset;
diff --git a/drivers/gpu/drm/i2c/ch7006_drv.c b/drivers/gpu/drm/i2c/ch7006_drv.c
index b865d07..51fa323 100644
--- a/drivers/gpu/drm/i2c/ch7006_drv.c
+++ b/drivers/gpu/drm/i2c/ch7006_drv.c
@@ -364,7 +364,7 @@ static int ch7006_encoder_set_property(struct drm_encoder *encoder,
 				.crtc = crtc,
 			};
 
-			crtc->funcs->set_config(&modeset);
+			drm_mode_set_config_internal(&modeset);
 		}
 	}
 
diff --git a/drivers/gpu/drm/nouveau/nv04_display.c b/drivers/gpu/drm/nouveau/nv04_display.c
index 2cd6fb8..4c6e9f8 100644
--- a/drivers/gpu/drm/nouveau/nv04_display.c
+++ b/drivers/gpu/drm/nouveau/nv04_display.c
@@ -140,7 +140,7 @@ nv04_display_destroy(struct drm_device *dev)
 			.crtc = crtc,
 		};
 
-		crtc->funcs->set_config(&modeset);
+		drm_mode_set_config_internal(&modeset);
 	}
 
 	/* Restore state */
diff --git a/drivers/gpu/drm/nouveau/nv17_tv.c b/drivers/gpu/drm/nouveau/nv17_tv.c
index 2ca276a..977e42b 100644
--- a/drivers/gpu/drm/nouveau/nv17_tv.c
+++ b/drivers/gpu/drm/nouveau/nv17_tv.c
@@ -768,7 +768,7 @@ static int nv17_tv_set_property(struct drm_encoder *encoder,
 				.crtc = crtc,
 			};
 
-			crtc->funcs->set_config(&modeset);
+			drm_mode_set_config_internal(&modeset);
 		}
 	}
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
index 161f8b2..07dfd82 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
@@ -829,7 +829,7 @@ static void vmw_lastclose(struct drm_device *dev)
 
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
 		set.crtc = crtc;
-		ret = crtc->funcs->set_config(&set);
+		ret = drm_mode_set_config_internal(&set);
 		WARN_ON(ret != 0);
 	}
 
diff --git a/include/drm/drm_crtc.h b/include/drm/drm_crtc.h
index 00d78b5..665289a6 100644
--- a/include/drm/drm_crtc.h
+++ b/include/drm/drm_crtc.h
@@ -985,6 +985,7 @@ extern int drm_mode_getcrtc(struct drm_device *dev,
 			    void *data, struct drm_file *file_priv);
 extern int drm_mode_getconnector(struct drm_device *dev,
 			      void *data, struct drm_file *file_priv);
+extern int drm_mode_set_config_internal(struct drm_mode_set *set);
 extern int drm_mode_setcrtc(struct drm_device *dev,
 			    void *data, struct drm_file *file_priv);
 extern int drm_mode_getplane(struct drm_device *dev,
-- 
cgit v0.10.2


From 848499032504b1defdebdcf930aa70bd01a45ac1 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sun, 2 Dec 2012 00:28:11 +0100
Subject: drm: add drm_modeset_lock|unlock_all

This is the first step towards introducing the new modeset locking
scheme. The plan is to put helper functions into place at all the
right places step-by-step, so that the final patch to switch on the
new locking scheme doesn't need to touch every single driver.

This helper here will serve as the shotgun solutions for all places
where a more fine-grained locking isn't (yet) implemented.

v2: Fixup kerneldoc for unlock_all.

Reviewed-by: Rob Clark <rob@ti.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index 7ca2f28..a82ec05 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -37,6 +37,29 @@
 #include <drm/drm_edid.h>
 #include <drm/drm_fourcc.h>
 
+/**
+ * drm_modeset_lock_all - take all modeset locks
+ * @dev: drm device
+ *
+ * This function takes all modeset locks, suitable where a more fine-grained
+ * scheme isn't (yet) implemented.
+ */
+void drm_modeset_lock_all(struct drm_device *dev)
+{
+	mutex_lock(&dev->mode_config.mutex);
+}
+EXPORT_SYMBOL(drm_modeset_lock_all);
+
+/**
+ * drm_modeset_unlock_all - drop all modeset locks
+ * @dev: device
+ */
+void drm_modeset_unlock_all(struct drm_device *dev)
+{
+	mutex_unlock(&dev->mode_config.mutex);
+}
+EXPORT_SYMBOL(drm_modeset_unlock_all);
+
 /* Avoid boilerplate.  I'm tired of typing. */
 #define DRM_ENUM_NAME_FN(fnname, list)				\
 	char *fnname(int val)					\
@@ -425,7 +448,7 @@ int drm_crtc_init(struct drm_device *dev, struct drm_crtc *crtc,
 	crtc->funcs = funcs;
 	crtc->invert_dimensions = false;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 
 	ret = drm_mode_object_get(dev, &crtc->base, DRM_MODE_OBJECT_CRTC);
 	if (ret)
@@ -437,7 +460,7 @@ int drm_crtc_init(struct drm_device *dev, struct drm_crtc *crtc,
 	dev->mode_config.num_crtc++;
 
  out:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 
 	return ret;
 }
@@ -512,7 +535,7 @@ int drm_connector_init(struct drm_device *dev,
 {
 	int ret;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 
 	ret = drm_mode_object_get(dev, &connector->base, DRM_MODE_OBJECT_CONNECTOR);
 	if (ret)
@@ -542,7 +565,7 @@ int drm_connector_init(struct drm_device *dev,
 				      dev->mode_config.dpms_property, 0);
 
  out:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 
 	return ret;
 }
@@ -592,7 +615,7 @@ int drm_encoder_init(struct drm_device *dev,
 {
 	int ret;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 
 	ret = drm_mode_object_get(dev, &encoder->base, DRM_MODE_OBJECT_ENCODER);
 	if (ret)
@@ -606,7 +629,7 @@ int drm_encoder_init(struct drm_device *dev,
 	dev->mode_config.num_encoder++;
 
  out:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 
 	return ret;
 }
@@ -615,11 +638,11 @@ EXPORT_SYMBOL(drm_encoder_init);
 void drm_encoder_cleanup(struct drm_encoder *encoder)
 {
 	struct drm_device *dev = encoder->dev;
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	drm_mode_object_put(dev, &encoder->base);
 	list_del(&encoder->head);
 	dev->mode_config.num_encoder--;
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 }
 EXPORT_SYMBOL(drm_encoder_cleanup);
 
@@ -631,7 +654,7 @@ int drm_plane_init(struct drm_device *dev, struct drm_plane *plane,
 {
 	int ret;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 
 	ret = drm_mode_object_get(dev, &plane->base, DRM_MODE_OBJECT_PLANE);
 	if (ret)
@@ -665,7 +688,7 @@ int drm_plane_init(struct drm_device *dev, struct drm_plane *plane,
 	}
 
  out:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 
 	return ret;
 }
@@ -675,7 +698,7 @@ void drm_plane_cleanup(struct drm_plane *plane)
 {
 	struct drm_device *dev = plane->dev;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	kfree(plane->format_types);
 	drm_mode_object_put(dev, &plane->base);
 	/* if not added to a list, it must be a private plane */
@@ -683,7 +706,7 @@ void drm_plane_cleanup(struct drm_plane *plane)
 		list_del(&plane->head);
 		dev->mode_config.num_plane--;
 	}
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 }
 EXPORT_SYMBOL(drm_plane_cleanup);
 
@@ -963,9 +986,9 @@ void drm_mode_config_init(struct drm_device *dev)
 	INIT_LIST_HEAD(&dev->mode_config.plane_list);
 	idr_init(&dev->mode_config.crtc_idr);
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	drm_mode_create_standard_connector_properties(dev);
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 
 	/* Just to be sure */
 	dev->mode_config.num_fb = 0;
@@ -1185,7 +1208,7 @@ int drm_mode_getresources(struct drm_device *dev, void *data,
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 
 	/*
 	 * For the non-control nodes we need to limit the list of resources
@@ -1327,7 +1350,7 @@ int drm_mode_getresources(struct drm_device *dev, void *data,
 		  card_res->count_connectors, card_res->count_encoders);
 
 out:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
@@ -1355,7 +1378,7 @@ int drm_mode_getcrtc(struct drm_device *dev,
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 
 	obj = drm_mode_object_find(dev, crtc_resp->crtc_id,
 				   DRM_MODE_OBJECT_CRTC);
@@ -1383,7 +1406,7 @@ int drm_mode_getcrtc(struct drm_device *dev,
 	}
 
 out:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
@@ -1426,7 +1449,7 @@ int drm_mode_getconnector(struct drm_device *dev, void *data,
 
 	DRM_DEBUG_KMS("[CONNECTOR:%d:?]\n", out_resp->connector_id);
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 
 	obj = drm_mode_object_find(dev, out_resp->connector_id,
 				   DRM_MODE_OBJECT_CONNECTOR);
@@ -1523,7 +1546,7 @@ int drm_mode_getconnector(struct drm_device *dev, void *data,
 	out_resp->count_encoders = encoders_count;
 
 out:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
@@ -1538,7 +1561,7 @@ int drm_mode_getencoder(struct drm_device *dev, void *data,
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	obj = drm_mode_object_find(dev, enc_resp->encoder_id,
 				   DRM_MODE_OBJECT_ENCODER);
 	if (!obj) {
@@ -1557,7 +1580,7 @@ int drm_mode_getencoder(struct drm_device *dev, void *data,
 	enc_resp->possible_clones = encoder->possible_clones;
 
 out:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
@@ -1581,7 +1604,7 @@ int drm_mode_getplane_res(struct drm_device *dev, void *data,
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	config = &dev->mode_config;
 
 	/*
@@ -1603,7 +1626,7 @@ int drm_mode_getplane_res(struct drm_device *dev, void *data,
 	plane_resp->count_planes = config->num_plane;
 
 out:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
@@ -1628,7 +1651,7 @@ int drm_mode_getplane(struct drm_device *dev, void *data,
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	obj = drm_mode_object_find(dev, plane_resp->plane_id,
 				   DRM_MODE_OBJECT_PLANE);
 	if (!obj) {
@@ -1668,7 +1691,7 @@ int drm_mode_getplane(struct drm_device *dev, void *data,
 	plane_resp->count_format_types = plane->format_count;
 
 out:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
@@ -1696,7 +1719,7 @@ int drm_mode_setplane(struct drm_device *dev, void *data,
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 
 	/*
 	 * First, find the plane, crtc, and fb objects.  If not available,
@@ -1795,7 +1818,7 @@ int drm_mode_setplane(struct drm_device *dev, void *data,
 	}
 
 out:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 
 	return ret;
 }
@@ -1850,7 +1873,7 @@ int drm_mode_setcrtc(struct drm_device *dev, void *data,
 	if (crtc_req->x > INT_MAX || crtc_req->y > INT_MAX)
 		return -ERANGE;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	obj = drm_mode_object_find(dev, crtc_req->crtc_id,
 				   DRM_MODE_OBJECT_CRTC);
 	if (!obj) {
@@ -1983,7 +2006,7 @@ int drm_mode_setcrtc(struct drm_device *dev, void *data,
 out:
 	kfree(connector_set);
 	drm_mode_destroy(dev, mode);
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
@@ -2001,7 +2024,7 @@ int drm_mode_cursor_ioctl(struct drm_device *dev,
 	if (!req->flags || (~DRM_MODE_CURSOR_FLAGS & req->flags))
 		return -EINVAL;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	obj = drm_mode_object_find(dev, req->crtc_id, DRM_MODE_OBJECT_CRTC);
 	if (!obj) {
 		DRM_DEBUG_KMS("Unknown CRTC ID %d\n", req->crtc_id);
@@ -2029,7 +2052,7 @@ int drm_mode_cursor_ioctl(struct drm_device *dev,
 		}
 	}
 out:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
@@ -2108,7 +2131,7 @@ int drm_mode_addfb(struct drm_device *dev,
 	if ((config->min_height > r.height) || (r.height > config->max_height))
 		return -EINVAL;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 
 	/* TODO check buffer is sufficiently large */
 	/* TODO setup destructor callback */
@@ -2125,7 +2148,7 @@ int drm_mode_addfb(struct drm_device *dev,
 	DRM_DEBUG_KMS("[FB:%d]\n", fb->base.id);
 
 out:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
@@ -2293,7 +2316,7 @@ int drm_mode_addfb2(struct drm_device *dev,
 	if (ret)
 		return ret;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 
 	fb = dev->mode_config.funcs->fb_create(dev, file_priv, r);
 	if (IS_ERR(fb)) {
@@ -2307,7 +2330,7 @@ int drm_mode_addfb2(struct drm_device *dev,
 	DRM_DEBUG_KMS("[FB:%d]\n", fb->base.id);
 
 out:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
@@ -2337,7 +2360,7 @@ int drm_mode_rmfb(struct drm_device *dev,
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	obj = drm_mode_object_find(dev, *id, DRM_MODE_OBJECT_FB);
 	/* TODO check that we really get a framebuffer back. */
 	if (!obj) {
@@ -2358,7 +2381,7 @@ int drm_mode_rmfb(struct drm_device *dev,
 	drm_framebuffer_remove(fb);
 
 out:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
@@ -2386,7 +2409,7 @@ int drm_mode_getfb(struct drm_device *dev,
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	obj = drm_mode_object_find(dev, r->fb_id, DRM_MODE_OBJECT_FB);
 	if (!obj) {
 		ret = -EINVAL;
@@ -2405,7 +2428,7 @@ int drm_mode_getfb(struct drm_device *dev,
 		ret = -ENODEV;
 
 out:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
@@ -2424,7 +2447,7 @@ int drm_mode_dirtyfb_ioctl(struct drm_device *dev,
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	obj = drm_mode_object_find(dev, r->fb_id, DRM_MODE_OBJECT_FB);
 	if (!obj) {
 		ret = -EINVAL;
@@ -2478,7 +2501,7 @@ int drm_mode_dirtyfb_ioctl(struct drm_device *dev,
 out_err2:
 	kfree(clips);
 out_err1:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
@@ -2499,11 +2522,11 @@ void drm_fb_release(struct drm_file *priv)
 	struct drm_device *dev = priv->minor->dev;
 	struct drm_framebuffer *fb, *tfb;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	list_for_each_entry_safe(fb, tfb, &priv->fbs, filp_head) {
 		drm_framebuffer_remove(fb);
 	}
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 }
 
 /**
@@ -2618,7 +2641,7 @@ int drm_mode_attachmode_ioctl(struct drm_device *dev,
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 
 	obj = drm_mode_object_find(dev, mode_cmd->connector_id, DRM_MODE_OBJECT_CONNECTOR);
 	if (!obj) {
@@ -2642,7 +2665,7 @@ int drm_mode_attachmode_ioctl(struct drm_device *dev,
 
 	drm_mode_attachmode(dev, connector, mode);
 out:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
@@ -2671,7 +2694,7 @@ int drm_mode_detachmode_ioctl(struct drm_device *dev,
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 
 	obj = drm_mode_object_find(dev, mode_cmd->connector_id, DRM_MODE_OBJECT_CONNECTOR);
 	if (!obj) {
@@ -2688,7 +2711,7 @@ int drm_mode_detachmode_ioctl(struct drm_device *dev,
 
 	ret = drm_mode_detachmode(dev, connector, &mode);
 out:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
@@ -2934,7 +2957,7 @@ int drm_mode_getproperty_ioctl(struct drm_device *dev,
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	obj = drm_mode_object_find(dev, out_resp->prop_id, DRM_MODE_OBJECT_PROPERTY);
 	if (!obj) {
 		ret = -EINVAL;
@@ -3012,7 +3035,7 @@ int drm_mode_getproperty_ioctl(struct drm_device *dev,
 		out_resp->count_enum_blobs = blob_count;
 	}
 done:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
@@ -3063,7 +3086,7 @@ int drm_mode_getblob_ioctl(struct drm_device *dev,
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	obj = drm_mode_object_find(dev, out_resp->blob_id, DRM_MODE_OBJECT_BLOB);
 	if (!obj) {
 		ret = -EINVAL;
@@ -3081,7 +3104,7 @@ int drm_mode_getblob_ioctl(struct drm_device *dev,
 	out_resp->length = blob->length;
 
 done:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
@@ -3223,7 +3246,7 @@ int drm_mode_obj_get_properties_ioctl(struct drm_device *dev, void *data,
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 
 	obj = drm_mode_object_find(dev, arg->obj_id, arg->obj_type);
 	if (!obj) {
@@ -3260,7 +3283,7 @@ int drm_mode_obj_get_properties_ioctl(struct drm_device *dev, void *data,
 	}
 	arg->count_props = props_count;
 out:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
@@ -3277,7 +3300,7 @@ int drm_mode_obj_set_property_ioctl(struct drm_device *dev, void *data,
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 
 	arg_obj = drm_mode_object_find(dev, arg->obj_id, arg->obj_type);
 	if (!arg_obj)
@@ -3315,7 +3338,7 @@ int drm_mode_obj_set_property_ioctl(struct drm_device *dev, void *data,
 	}
 
 out:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
@@ -3377,7 +3400,7 @@ int drm_mode_gamma_set_ioctl(struct drm_device *dev,
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	obj = drm_mode_object_find(dev, crtc_lut->crtc_id, DRM_MODE_OBJECT_CRTC);
 	if (!obj) {
 		ret = -EINVAL;
@@ -3418,7 +3441,7 @@ int drm_mode_gamma_set_ioctl(struct drm_device *dev,
 	crtc->funcs->gamma_set(crtc, r_base, g_base, b_base, 0, crtc->gamma_size);
 
 out:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 
 }
@@ -3436,7 +3459,7 @@ int drm_mode_gamma_get_ioctl(struct drm_device *dev,
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	obj = drm_mode_object_find(dev, crtc_lut->crtc_id, DRM_MODE_OBJECT_CRTC);
 	if (!obj) {
 		ret = -EINVAL;
@@ -3469,7 +3492,7 @@ int drm_mode_gamma_get_ioctl(struct drm_device *dev,
 		goto out;
 	}
 out:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
@@ -3489,7 +3512,7 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,
 	    page_flip->reserved != 0)
 		return -EINVAL;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	obj = drm_mode_object_find(dev, page_flip->crtc_id, DRM_MODE_OBJECT_CRTC);
 	if (!obj)
 		goto out;
@@ -3567,7 +3590,7 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,
 	}
 
 out:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
diff --git a/drivers/gpu/drm/drm_crtc_helper.c b/drivers/gpu/drm/drm_crtc_helper.c
index 7b2d378..400ef86 100644
--- a/drivers/gpu/drm/drm_crtc_helper.c
+++ b/drivers/gpu/drm/drm_crtc_helper.c
@@ -980,7 +980,7 @@ static void output_poll_execute(struct work_struct *work)
 	if (!drm_kms_helper_poll)
 		return;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
 
 		/* Ignore forced connectors. */
@@ -1010,7 +1010,7 @@ static void output_poll_execute(struct work_struct *work)
 			changed = true;
 	}
 
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 
 	if (changed)
 		drm_kms_helper_hotplug_event(dev);
@@ -1070,7 +1070,7 @@ void drm_helper_hpd_irq_event(struct drm_device *dev)
 	if (!dev->mode_config.poll_enabled)
 		return;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
 
 		/* Only handle HPD capable connectors. */
@@ -1088,7 +1088,7 @@ void drm_helper_hpd_irq_event(struct drm_device *dev)
 			changed = true;
 	}
 
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 
 	if (changed)
 		drm_kms_helper_hotplug_event(dev);
diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index 82c3a9f..be0f2d6 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -337,7 +337,7 @@ static void drm_fb_helper_dpms(struct fb_info *info, int dpms_mode)
 	/*
 	 * For each CRTC in this fb, turn the connectors on/off.
 	 */
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	for (i = 0; i < fb_helper->crtc_count; i++) {
 		crtc = fb_helper->crtc_info[i].mode_set.crtc;
 
@@ -352,7 +352,7 @@ static void drm_fb_helper_dpms(struct fb_info *info, int dpms_mode)
 				dev->mode_config.dpms_property, dpms_mode);
 		}
 	}
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 }
 
 int drm_fb_helper_blank(int blank, struct fb_info *info)
@@ -672,16 +672,16 @@ int drm_fb_helper_set_par(struct fb_info *info)
 		return -EINVAL;
 	}
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	for (i = 0; i < fb_helper->crtc_count; i++) {
 		crtc = fb_helper->crtc_info[i].mode_set.crtc;
 		ret = drm_mode_set_config_internal(&fb_helper->crtc_info[i].mode_set);
 		if (ret) {
-			mutex_unlock(&dev->mode_config.mutex);
+			drm_modeset_unlock_all(dev);
 			return ret;
 		}
 	}
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 
 	if (fb_helper->delayed_hotplug) {
 		fb_helper->delayed_hotplug = false;
@@ -701,7 +701,7 @@ int drm_fb_helper_pan_display(struct fb_var_screeninfo *var,
 	int ret = 0;
 	int i;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	for (i = 0; i < fb_helper->crtc_count; i++) {
 		crtc = fb_helper->crtc_info[i].mode_set.crtc;
 
@@ -718,7 +718,7 @@ int drm_fb_helper_pan_display(struct fb_var_screeninfo *var,
 			}
 		}
 	}
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 }
 EXPORT_SYMBOL(drm_fb_helper_pan_display);
@@ -1375,7 +1375,7 @@ int drm_fb_helper_hotplug_event(struct drm_fb_helper *fb_helper)
 	if (!fb_helper->fb)
 		return 0;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
 		if (crtc->fb)
 			crtcs_bound++;
@@ -1385,7 +1385,7 @@ int drm_fb_helper_hotplug_event(struct drm_fb_helper *fb_helper)
 
 	if (bound < crtcs_bound) {
 		fb_helper->delayed_hotplug = true;
-		mutex_unlock(&dev->mode_config.mutex);
+		drm_modeset_unlock_all(dev);
 		return 0;
 	}
 	DRM_DEBUG_KMS("\n");
@@ -1397,7 +1397,7 @@ int drm_fb_helper_hotplug_event(struct drm_fb_helper *fb_helper)
 	count = drm_fb_helper_probe_connector_modes(fb_helper, max_width,
 						    max_height);
 	drm_setup_crtcs(fb_helper);
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 
 	return drm_fb_helper_single_fb_probe(fb_helper, bpp_sel);
 }
diff --git a/include/drm/drm_crtc.h b/include/drm/drm_crtc.h
index 665289a6..9f0524d 100644
--- a/include/drm/drm_crtc.h
+++ b/include/drm/drm_crtc.h
@@ -842,6 +842,9 @@ struct drm_prop_enum_list {
 	char *name;
 };
 
+extern void drm_modeset_lock_all(struct drm_device *dev);
+extern void drm_modeset_unlock_all(struct drm_device *dev);
+
 extern int drm_crtc_init(struct drm_device *dev,
 			 struct drm_crtc *crtc,
 			 const struct drm_crtc_funcs *funcs);
-- 
cgit v0.10.2


From a0e99e68c12ac6dc5d6b1da7942b5e05d5f848af Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sun, 2 Dec 2012 01:05:46 +0100
Subject: drm/i915: use drm_modeset_lock_all

Two exceptions:
- debugfs files only read information which is not related to crtc, so
  can stay on the modeset_config lock.
- Same holds for the edp vdd work in intel_dp.c. Add a corresponding
  WARN_ON and a comment next to the intel_dp struct fields for
  documentation.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 1b63d55..6912d29 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -1057,6 +1057,8 @@ static void ironlake_panel_vdd_off_sync(struct intel_dp *intel_dp)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 pp;
 
+	WARN_ON(!mutex_is_locked(&dev->mode_config.mutex));
+
 	if (!intel_dp->want_panel_vdd && ironlake_edp_have_panel_vdd(intel_dp)) {
 		pp = ironlake_get_pp_control(dev_priv);
 		pp &= ~EDP_FORCE_VDD;
diff --git a/drivers/gpu/drm/i915/intel_fb.c b/drivers/gpu/drm/i915/intel_fb.c
index 7b30b5c..a9f7280 100644
--- a/drivers/gpu/drm/i915/intel_fb.c
+++ b/drivers/gpu/drm/i915/intel_fb.c
@@ -280,7 +280,7 @@ void intel_fb_restore_mode(struct drm_device *dev)
 	struct drm_mode_config *config = &dev->mode_config;
 	struct drm_plane *plane;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 
 	ret = drm_fb_helper_restore_fbdev_mode(&dev_priv->fbdev->helper);
 	if (ret)
@@ -290,5 +290,5 @@ void intel_fb_restore_mode(struct drm_device *dev)
 	list_for_each_entry(plane, &config->plane_list, head)
 		plane->funcs->disable_plane(plane);
 
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 }
diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c
index 17aee74..279e9d7 100644
--- a/drivers/gpu/drm/i915/intel_lvds.c
+++ b/drivers/gpu/drm/i915/intel_lvds.c
@@ -531,9 +531,9 @@ static int intel_lid_notify(struct notifier_block *nb, unsigned long val,
 
 	dev_priv->modeset_on_lid = 0;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	intel_modeset_setup_hw_state(dev, true);
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 
 	return NOTIFY_OK;
 }
diff --git a/drivers/gpu/drm/i915/intel_overlay.c b/drivers/gpu/drm/i915/intel_overlay.c
index d7bc817..5be09d1 100644
--- a/drivers/gpu/drm/i915/intel_overlay.c
+++ b/drivers/gpu/drm/i915/intel_overlay.c
@@ -1045,13 +1045,13 @@ int intel_overlay_put_image(struct drm_device *dev, void *data,
 	}
 
 	if (!(put_image_rec->flags & I915_OVERLAY_ENABLE)) {
-		mutex_lock(&dev->mode_config.mutex);
+		drm_modeset_lock_all(dev);
 		mutex_lock(&dev->struct_mutex);
 
 		ret = intel_overlay_switch_off(overlay);
 
 		mutex_unlock(&dev->struct_mutex);
-		mutex_unlock(&dev->mode_config.mutex);
+		drm_modeset_unlock_all(dev);
 
 		return ret;
 	}
@@ -1075,7 +1075,7 @@ int intel_overlay_put_image(struct drm_device *dev, void *data,
 		goto out_free;
 	}
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	mutex_lock(&dev->struct_mutex);
 
 	if (new_bo->tiling_mode) {
@@ -1157,7 +1157,7 @@ int intel_overlay_put_image(struct drm_device *dev, void *data,
 		goto out_unlock;
 
 	mutex_unlock(&dev->struct_mutex);
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 
 	kfree(params);
 
@@ -1165,7 +1165,7 @@ int intel_overlay_put_image(struct drm_device *dev, void *data,
 
 out_unlock:
 	mutex_unlock(&dev->struct_mutex);
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	drm_gem_object_unreference_unlocked(&new_bo->base);
 out_free:
 	kfree(params);
@@ -1241,7 +1241,7 @@ int intel_overlay_attrs(struct drm_device *dev, void *data,
 		return -ENODEV;
 	}
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	mutex_lock(&dev->struct_mutex);
 
 	ret = -EINVAL;
@@ -1307,7 +1307,7 @@ int intel_overlay_attrs(struct drm_device *dev, void *data,
 	ret = 0;
 out_unlock:
 	mutex_unlock(&dev->struct_mutex);
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 
 	return ret;
 }
diff --git a/drivers/gpu/drm/i915/intel_sprite.c b/drivers/gpu/drm/i915/intel_sprite.c
index d7b060e..f829306 100644
--- a/drivers/gpu/drm/i915/intel_sprite.c
+++ b/drivers/gpu/drm/i915/intel_sprite.c
@@ -593,7 +593,7 @@ int intel_sprite_set_colorkey(struct drm_device *dev, void *data,
 	if ((set->flags & (I915_SET_COLORKEY_DESTINATION | I915_SET_COLORKEY_SOURCE)) == (I915_SET_COLORKEY_DESTINATION | I915_SET_COLORKEY_SOURCE))
 		return -EINVAL;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 
 	obj = drm_mode_object_find(dev, set->plane_id, DRM_MODE_OBJECT_PLANE);
 	if (!obj) {
@@ -606,7 +606,7 @@ int intel_sprite_set_colorkey(struct drm_device *dev, void *data,
 	ret = intel_plane->update_colorkey(plane, set);
 
 out_unlock:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
@@ -622,7 +622,7 @@ int intel_sprite_get_colorkey(struct drm_device *dev, void *data,
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -ENODEV;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 
 	obj = drm_mode_object_find(dev, get->plane_id, DRM_MODE_OBJECT_PLANE);
 	if (!obj) {
@@ -635,7 +635,7 @@ int intel_sprite_get_colorkey(struct drm_device *dev, void *data,
 	intel_plane->get_colorkey(plane, get);
 
 out_unlock:
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
-- 
cgit v0.10.2


From 0a819515fc346b4e79d6e3fc01d837a660452c74 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sun, 2 Dec 2012 01:33:17 +0100
Subject: drm/gma500: use drm_modeset_lock_all

Only two places:
- suspend/resume
- Some really strange mode validation tool with too much funny-lucking
  hand-rolled conversion code.
- The recently-added lastclose fbdev restore code.

Better safe than sorry, so convert both places to keep the locking
semantics as much as possible.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/gma500/psb_device.c b/drivers/gpu/drm/gma500/psb_device.c
index b58c470..f6f534b 100644
--- a/drivers/gpu/drm/gma500/psb_device.c
+++ b/drivers/gpu/drm/gma500/psb_device.c
@@ -194,7 +194,7 @@ static int psb_save_display_registers(struct drm_device *dev)
 	regs->saveCHICKENBIT = PSB_RVDC32(DSPCHICKENBIT);
 
 	/* Save crtc and output state */
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
 		if (drm_helper_crtc_in_use(crtc))
 			crtc->funcs->save(crtc);
@@ -204,7 +204,7 @@ static int psb_save_display_registers(struct drm_device *dev)
 		if (connector->funcs->save)
 			connector->funcs->save(connector);
 
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return 0;
 }
 
@@ -234,7 +234,7 @@ static int psb_restore_display_registers(struct drm_device *dev)
 	/*make sure VGA plane is off. it initializes to on after reset!*/
 	PSB_WVDC32(0x80000000, VGACNTRL);
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head)
 		if (drm_helper_crtc_in_use(crtc))
 			crtc->funcs->restore(crtc);
@@ -243,7 +243,7 @@ static int psb_restore_display_registers(struct drm_device *dev)
 		if (connector->funcs->restore)
 			connector->funcs->restore(connector);
 
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/gma500/psb_drv.c b/drivers/gpu/drm/gma500/psb_drv.c
index dbcefe9..111e3df 100644
--- a/drivers/gpu/drm/gma500/psb_drv.c
+++ b/drivers/gpu/drm/gma500/psb_drv.c
@@ -153,11 +153,11 @@ static void psb_lastclose(struct drm_device *dev)
 	struct drm_psb_private *dev_priv = dev->dev_private;
 	struct psb_fbdev *fbdev = dev_priv->fbdev;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	ret = drm_fb_helper_restore_fbdev_mode(&fbdev->psb_fb_helper);
 	if (ret)
 		DRM_DEBUG("failed to restore crtc mode\n");
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 
 	return;
 }
@@ -486,7 +486,7 @@ static int psb_mode_operation_ioctl(struct drm_device *dev, void *data,
 	case PSB_MODE_OPERATION_MODE_VALID:
 		umode = &arg->mode;
 
-		mutex_lock(&dev->mode_config.mutex);
+		drm_modeset_lock_all(dev);
 
 		obj = drm_mode_object_find(dev, obj_id,
 					DRM_MODE_OBJECT_CONNECTOR);
@@ -535,7 +535,7 @@ static int psb_mode_operation_ioctl(struct drm_device *dev, void *data,
 		if (mode)
 			drm_mode_destroy(dev, mode);
 mode_op_out:
-		mutex_unlock(&dev->mode_config.mutex);
+		drm_modeset_unlock_all(dev);
 		return ret;
 
 	default:
-- 
cgit v0.10.2


From 795f1426b484ddfc136e723a26bfd6a46c8ab085 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sun, 2 Dec 2012 01:34:59 +0100
Subject: drm/ast: use drm_modeset_lock_all

Just a call to drm_helper_resume_force_mode, obviously wants full
locking for that.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/ast/ast_drv.c b/drivers/gpu/drm/ast/ast_drv.c
index 2d2c2f8..df0d0a0 100644
--- a/drivers/gpu/drm/ast/ast_drv.c
+++ b/drivers/gpu/drm/ast/ast_drv.c
@@ -94,9 +94,9 @@ static int ast_drm_thaw(struct drm_device *dev)
 	ast_post_gpu(dev);
 
 	drm_mode_config_reset(dev);
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	drm_helper_resume_force_mode(dev);
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 
 	console_lock();
 	ast_fbdev_set_suspend(dev, 0);
-- 
cgit v0.10.2


From b13f59804918ad584a97a639064a24d14e2fd740 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sun, 2 Dec 2012 01:36:59 +0100
Subject: drm/shmobile: use drm_modeset_lock_all

Only a resume method to account for.

v2: Fixup compilation.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/shmobile/shmob_drm_drv.c b/drivers/gpu/drm/shmobile/shmob_drm_drv.c
index d1d5306..f6e0b53 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_drv.c
+++ b/drivers/gpu/drm/shmobile/shmob_drm_drv.c
@@ -313,9 +313,9 @@ static int shmob_drm_pm_resume(struct device *dev)
 {
 	struct shmob_drm_device *sdev = dev_get_drvdata(dev);
 
-	mutex_lock(&sdev->ddev->mode_config.mutex);
+	drm_modeset_lock_all(sdev->ddev);
 	shmob_drm_crtc_resume(&sdev->crtc);
-	mutex_unlock(&sdev->ddev->mode_config.mutex);
+	drm_modeset_unlock_all(sdev->ddev);
 
 	drm_kms_helper_poll_enable(sdev->ddev);
 	return 0;
-- 
cgit v0.10.2


From bbe4b99ff2443e305598768ae8eac6bc3516b7c9 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sun, 2 Dec 2012 01:48:38 +0100
Subject: drm/vmwgfx: use drm_modeset_lock_all

Ok, this one here is a bit more complicated, and I can't really claim
to fully understand the locking and lifetime rules of the vmwgfx
driver. So just convert ever mutex_lock call, including the
interruptible one. Since other places (e.g. in the execbuf ioctl) take
the mode_config.mutex without bothering with interruptible handling,
I've figured I should be able to get away with this in a few more
places ...

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c
index d9fbbe1..a135498 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c
@@ -163,11 +163,7 @@ int vmw_present_ioctl(struct drm_device *dev, void *data,
 		goto out_no_copy;
 	}
 
-	ret = mutex_lock_interruptible(&dev->mode_config.mutex);
-	if (unlikely(ret != 0)) {
-		ret = -ERESTARTSYS;
-		goto out_no_mode_mutex;
-	}
+	drm_modeset_lock_all(dev);
 
 	obj = drm_mode_object_find(dev, arg->fb_id, DRM_MODE_OBJECT_FB);
 	if (!obj) {
@@ -200,8 +196,7 @@ out_no_surface:
 	ttm_read_unlock(&vmaster->lock);
 out_no_ttm_lock:
 out_no_fb:
-	mutex_unlock(&dev->mode_config.mutex);
-out_no_mode_mutex:
+	drm_modeset_unlock_all(dev);
 out_no_copy:
 	kfree(clips);
 out_clips:
@@ -251,11 +246,7 @@ int vmw_present_readback_ioctl(struct drm_device *dev, void *data,
 		goto out_no_copy;
 	}
 
-	ret = mutex_lock_interruptible(&dev->mode_config.mutex);
-	if (unlikely(ret != 0)) {
-		ret = -ERESTARTSYS;
-		goto out_no_mode_mutex;
-	}
+	drm_modeset_lock_all(dev);
 
 	obj = drm_mode_object_find(dev, arg->fb_id, DRM_MODE_OBJECT_FB);
 	if (!obj) {
@@ -282,8 +273,7 @@ int vmw_present_readback_ioctl(struct drm_device *dev, void *data,
 	ttm_read_unlock(&vmaster->lock);
 out_no_ttm_lock:
 out_no_fb:
-	mutex_unlock(&dev->mode_config.mutex);
-out_no_mode_mutex:
+	drm_modeset_unlock_all(dev);
 out_no_copy:
 	kfree(clips);
 out_clips:
-- 
cgit v0.10.2


From d5d2636ed7990b93c7216f6a4d323f6b0eee08af Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sun, 20 Jan 2013 15:50:41 +0100
Subject: omapdrm: use modeset_lock_all

I've left the locking in the debugfs code as-is, it's essentially just
used to keep the framebuffer object alive (which won't be necessary
any more later on). We don't need fb refcounting either, since the new
mode_config.fb_lock ensures that the framebuffers can't disappear
(once mode_config.mutex doesn't guarantee this any more later on in
the series).

The fbcon restore needs all modeset locks. The crtc callbacks seem to
only need the crtc locks, but I've quickly discussed things with Rob
Clark and he's fine with just using modeset_lock_all for those, too.
He'll look into converting things over later.

Reviewed-by: Rob Clark <rob@ti.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/staging/omapdrm/omap_crtc.c b/drivers/staging/omapdrm/omap_crtc.c
index 5c6ed60..510942e 100644
--- a/drivers/staging/omapdrm/omap_crtc.c
+++ b/drivers/staging/omapdrm/omap_crtc.c
@@ -278,13 +278,13 @@ static void page_flip_worker(struct work_struct *work)
 	struct drm_display_mode *mode = &crtc->mode;
 	struct drm_gem_object *bo;
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	omap_plane_mode_set(omap_crtc->plane, crtc, crtc->fb,
 			0, 0, mode->hdisplay, mode->vdisplay,
 			crtc->x << 16, crtc->y << 16,
 			mode->hdisplay << 16, mode->vdisplay << 16,
 			vblank_cb, crtc);
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 
 	bo = omap_framebuffer_bo(crtc->fb, 0);
 	drm_gem_object_unreference_unlocked(bo);
@@ -417,7 +417,7 @@ static void apply_worker(struct work_struct *work)
 	 * the callbacks and list modification all serialized
 	 * with respect to modesetting ioctls from userspace.
 	 */
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	dispc_runtime_get();
 
 	/*
@@ -462,7 +462,7 @@ static void apply_worker(struct work_struct *work)
 
 out:
 	dispc_runtime_put();
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 }
 
 int omap_crtc_apply(struct drm_crtc *crtc,
diff --git a/drivers/staging/omapdrm/omap_drv.c b/drivers/staging/omapdrm/omap_drv.c
index ae5ecc2..dfdb4ba 100644
--- a/drivers/staging/omapdrm/omap_drv.c
+++ b/drivers/staging/omapdrm/omap_drv.c
@@ -449,9 +449,9 @@ static void dev_lastclose(struct drm_device *dev)
 		}
 	}
 
-	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock_all(dev);
 	ret = drm_fb_helper_restore_fbdev_mode(priv->fbdev);
-	mutex_unlock(&dev->mode_config.mutex);
+	drm_modeset_unlock_all(dev);
 	if (ret)
 		DBG("failed to restore crtc mode");
 }
-- 
cgit v0.10.2


From 29494c174dc4793ebd236aa522a2a1ed73b7180e Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sun, 2 Dec 2012 02:18:25 +0100
Subject: drm: add per-crtc locks

*drumroll*

The basic idea is to protect per-crtc state which can change without
touching the output configuration with separate mutexes, i.e.  all the
input side state to a crtc like framebuffers, cursor settings or plane
configuration. Holding such a crtc lock gives a read-lock on all the
other crtc state which can be changed by e.g. a modeset.

All non-crtc state is still protected by the mode_config mutex.
Callers that need to change modeset state of a crtc (e.g. dpms or
set_mode) need to grab both the mode_config lock and nested within any
crtc locks.

Note that since there can only ever be one holder of the mode_config
lock we can grab the subordinate crtc locks in any order (if we need
to grab more than one of them). Lockdep can handle such nesting with
the mutex_lock_nest_lock call correctly.

With this functions that only touch connectors/encoders but not crtcs
only need to take the mode_config lock. The biggest such case is the
output probing, which means that we can now pageflip and move cursors
while the output probe code is reading an edid.

Most cases neatly fall into the three buckets:
- Only touches connectors and similar output state and so only needs
  the mode_config lock.
- Touches the global configuration and so needs all locks.
- Only touches the crtc input side and so only needs the crtc lock.

But a few cases that need special consideration:

- Load detection which requires a crtc. The mode_config lock already
  prevents a modeset change, so we can use any unused crtc as we like
  to do load detection. The only thing to consider is that such
  temporary state changes don't leak out to userspace through ioctls
  that only take the crtc look (like a pageflip). Hence the load
  detect code needs to grab the crtc of any output pipes it touches
  (but only if it touches state used by the pageflip or cursor
  ioctls).

- Atomic pageflip when moving planes. The first case is sane hw, where
  planes have a fixed association with crtcs - nothing needs to be
  done there. More insane^Wflexible hw needs to have plane->crtc
  mapping which is separately protect with a lock that nests within
  the crtc lock. If the plane is unused we can just assign it to the
  current crtc and continue. But if a plane is already in use by
  another crtc we can't just reassign it.

  Two solution present themselves: Either go back to a slow-path which
  takes all modeset locks, potentially incure quite a hefty delay. Or
  simply disallowing such changes in one atomic pageflip - in general
  the vblanks of two crtcs are not synced, so there's no sane way to
  atomically flip such plane changes accross more than one crtc. I'd
  heavily favour the later approach, going as far as mandating it as
  part of the ABI of such a new a nuclear pageflip.

  And if we _really_ want such semantics, we can always get them by
  introducing another pageflip mutex between the mode_config.mutex and
  the individual crtc locks. Pageflips crossing more than one crtc
  would then need to take that lock first, to lock out concurrent
  multi-crtc pageflips.

- Optimized global modeset operations: We could just take the
  mode_config lock and then lazily lock all crtc which are affected by
  a modeset operation. This has the advantage that pageflip could
  continue unhampered on unaffected crtc. But if e.g. global resources
  like plls need to be reassigned and so affect unrelated crtcs we can
  still do that - nested locking works in any order.

This patch just adds the locks and takes them in drm_modeset_lock_all,
no real locking changes yet.

v2: Need to initialize the new lock in crtc_init and lock it righ
away, for otherwise the modeset_unlock_all below will try to unlock a
not-locked mutex.

Reviewed-by: Rob Clark <rob@ti.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index a82ec05..b7c6168 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -46,7 +46,12 @@
  */
 void drm_modeset_lock_all(struct drm_device *dev)
 {
+	struct drm_crtc *crtc;
+
 	mutex_lock(&dev->mode_config.mutex);
+
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head)
+		mutex_lock_nest_lock(&crtc->mutex, &dev->mode_config.mutex);
 }
 EXPORT_SYMBOL(drm_modeset_lock_all);
 
@@ -56,6 +61,11 @@ EXPORT_SYMBOL(drm_modeset_lock_all);
  */
 void drm_modeset_unlock_all(struct drm_device *dev)
 {
+	struct drm_crtc *crtc;
+
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head)
+		mutex_unlock(&crtc->mutex);
+
 	mutex_unlock(&dev->mode_config.mutex);
 }
 EXPORT_SYMBOL(drm_modeset_unlock_all);
@@ -449,6 +459,8 @@ int drm_crtc_init(struct drm_device *dev, struct drm_crtc *crtc,
 	crtc->invert_dimensions = false;
 
 	drm_modeset_lock_all(dev);
+	mutex_init(&crtc->mutex);
+	mutex_lock_nest_lock(&crtc->mutex, &dev->mode_config.mutex);
 
 	ret = drm_mode_object_get(dev, &crtc->base, DRM_MODE_OBJECT_CRTC);
 	if (ret)
diff --git a/include/drm/drm_crtc.h b/include/drm/drm_crtc.h
index 9f0524d..c89b116 100644
--- a/include/drm/drm_crtc.h
+++ b/include/drm/drm_crtc.h
@@ -390,6 +390,15 @@ struct drm_crtc {
 	struct drm_device *dev;
 	struct list_head head;
 
+	/**
+	 * crtc mutex
+	 *
+	 * This provides a read lock for the overall crtc state (mode, dpms
+	 * state, ...) and a write lock for everything which can be update
+	 * without a full modeset (fb, cursor data, ...)
+	 */
+	struct mutex mutex;
+
 	struct drm_mode_object base;
 
 	/* framebuffer the connector is currently bound to */
-- 
cgit v0.10.2


From bfb899282f500eeb9dff2600729904aad0fd39e7 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sun, 2 Dec 2012 13:48:21 +0100
Subject: drm: only take the crtc lock for ->cursor_set

First convert ->cursor_set to only take the crtc lock, since that
seems to be the function with the least amount of state - the core
ioctl function doesn't check anything which can change at runtime, so
we don't have any object lifetime issues to contend.

The only thing which is important is that the driver's implementation
doesn't touch any state outside of that single crtc which is not yet
properly protected by other locking:

- ast: access the global ast->cache_kmap. Luckily we only have on crtc
  on this driver, so this is fine. Add a comment.

- gma500: calls gma_power_begin|and and psb_gtt_pin|unpin, both which
  have their own locking to protect their state. Everything else is
  crtc-local.

- i915: touches a bit of global gem state, all protected by the One
  Lock to Rule Them All (dev->struct_mutex).

- nouveau: Pre-nv50 is all nice, nv50+ uses the evo channels to queue
  up all display changes. And some of these channels are device
  global. But this is fine now since the previous patch introduced an
  evo channel mutex.

- radeon: Uses some indirect register access for cursor updates, but
  with the previous patches to protect these indirect 2-register
  access patterns with a spinlock, this should be fine now, too.

- vmwgfx: I have no idea how that works - update_cursor_position
  doesn't take any per-crtc argument and I haven't figured out any
  other place where this could be set in some form of a side-channel.
  But vmwgfx definitely has more than one crtc (or at least can
  register more than one), so I have no idea how this is supposed to
  not fail with the current code already. Hence take the easy way out
  and simply acquire all locks (which requires dropping the crtc lock
  the core acquired for us). That way it's not worse off for
  consistency than the old code.

Reviewed-by: Rob Clark <rob@ti.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/ast/ast_drv.h b/drivers/gpu/drm/ast/ast_drv.h
index 5ccf984..5284292 100644
--- a/drivers/gpu/drm/ast/ast_drv.h
+++ b/drivers/gpu/drm/ast/ast_drv.h
@@ -98,6 +98,8 @@ struct ast_private {
 
 	struct drm_gem_object *cursor_cache;
 	uint64_t cursor_cache_gpu_addr;
+	/* Acces to this cache is protected by the crtc->mutex of the only crtc
+	 * we have. */
 	struct ttm_bo_kmap_obj cache_kmap;
 	int next_cursor;
 };
diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index b7c6168..58fa69e 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -2036,7 +2036,6 @@ int drm_mode_cursor_ioctl(struct drm_device *dev,
 	if (!req->flags || (~DRM_MODE_CURSOR_FLAGS & req->flags))
 		return -EINVAL;
 
-	drm_modeset_lock_all(dev);
 	obj = drm_mode_object_find(dev, req->crtc_id, DRM_MODE_OBJECT_CRTC);
 	if (!obj) {
 		DRM_DEBUG_KMS("Unknown CRTC ID %d\n", req->crtc_id);
@@ -2051,20 +2050,23 @@ int drm_mode_cursor_ioctl(struct drm_device *dev,
 			goto out;
 		}
 		/* Turns off the cursor if handle is 0 */
+		mutex_lock(&crtc->mutex);
 		ret = crtc->funcs->cursor_set(crtc, file_priv, req->handle,
 					      req->width, req->height);
+		mutex_unlock(&crtc->mutex);
 	}
 
 	if (req->flags & DRM_MODE_CURSOR_MOVE) {
 		if (crtc->funcs->cursor_move) {
+			drm_modeset_lock_all(dev);
 			ret = crtc->funcs->cursor_move(crtc, req->x, req->y);
+			drm_modeset_unlock_all(dev);
 		} else {
 			ret = -EFAULT;
 			goto out;
 		}
 	}
 out:
-	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
index 9c0876b..8d82e63 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
@@ -180,16 +180,29 @@ int vmw_du_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv,
 	struct vmw_dma_buffer *dmabuf = NULL;
 	int ret;
 
+	/*
+	 * FIXME: Unclear whether there's any global state touched by the
+	 * cursor_set function, especially vmw_cursor_update_position looks
+	 * suspicious. For now take the easy route and reacquire all locks. We
+	 * can do this since the caller in the drm core doesn't check anything
+	 * which is protected by any looks.
+	 */
+	mutex_unlock(&crtc->mutex);
+	drm_modeset_lock_all(dev_priv->dev);
+
 	/* A lot of the code assumes this */
-	if (handle && (width != 64 || height != 64))
-		return -EINVAL;
+	if (handle && (width != 64 || height != 64)) {
+		ret = -EINVAL;
+		goto out;
+	}
 
 	if (handle) {
 		ret = vmw_user_lookup_handle(dev_priv, tfile,
 					     handle, &surface, &dmabuf);
 		if (ret) {
 			DRM_ERROR("failed to find surface or dmabuf: %i\n", ret);
-			return -EINVAL;
+			ret = -EINVAL;
+			goto out;
 		}
 	}
 
@@ -197,7 +210,8 @@ int vmw_du_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv,
 	if (surface && !surface->snooper.image) {
 		DRM_ERROR("surface not suitable for cursor\n");
 		vmw_surface_unreference(&surface);
-		return -EINVAL;
+		ret = -EINVAL;
+		goto out;
 	}
 
 	/* takedown old cursor */
@@ -225,14 +239,20 @@ int vmw_du_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv,
 					       du->hotspot_x, du->hotspot_y);
 	} else {
 		vmw_cursor_update_position(dev_priv, false, 0, 0);
-		return 0;
+		ret = 0;
+		goto out;
 	}
 
 	vmw_cursor_update_position(dev_priv, true,
 				   du->cursor_x + du->hotspot_x,
 				   du->cursor_y + du->hotspot_y);
 
-	return 0;
+	ret = 0;
+out:
+	drm_modeset_unlock_all(dev_priv->dev);
+	mutex_lock(&crtc->mutex);
+
+	return ret;
 }
 
 int vmw_du_crtc_cursor_move(struct drm_crtc *crtc, int x, int y)
-- 
cgit v0.10.2


From dac35663cef4ca7f572d430bb54b14be8f03cb10 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sun, 2 Dec 2012 15:24:10 +0100
Subject: drm: only take the crtc lock for ->cursor_move

->cursor_move uses mostly the same facilities in drivers as
->cursor_set, so pretty much nothing to fix up:

- ast/gma500/i915: They all use per-crtc registers to update the
  cursor position. ast again touches the global cursor cache, but
  that's ok since there's only one crtc.

- nouveau: nv50+ is again special, updates happen through the per-crtc
  channel (without pushbufs), so it's not protected by the new evo
  lock introduced earlier. But since this channel is per-crtc, we
  should be fine anyway.

- radeon: A bit a mess: avivo asics need a workaround when both output
  pipes are enabled, which means it'll access the crtc list. Just
  reading that flag is ok though as long as radeon _always_ grabs all
  locks when changing the crtc configuration. Which means with the
  current scheme it cannot do an optimized modeset which only locks
  the relevant crtcs. This can be fixed though by introducing a bit of
  global state with separate locks and ensure in the modeset code that
  the cursor will be updated appropriately when enabling the 2nd pipe
  (on affected asics).

- vmwgfx: I still don't understand what it's doing exactly, so apply
  the same trick for now.

v2: Fixup unlocking for the error cases, spotted by Richard Wilbur.

v3: Another error-case fixup.

Reviewed-by: Rob Clark <rob@ti.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index 58fa69e..4af6a3d 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -2039,34 +2039,32 @@ int drm_mode_cursor_ioctl(struct drm_device *dev,
 	obj = drm_mode_object_find(dev, req->crtc_id, DRM_MODE_OBJECT_CRTC);
 	if (!obj) {
 		DRM_DEBUG_KMS("Unknown CRTC ID %d\n", req->crtc_id);
-		ret = -EINVAL;
-		goto out;
+		return -EINVAL;
 	}
 	crtc = obj_to_crtc(obj);
 
+	mutex_lock(&crtc->mutex);
 	if (req->flags & DRM_MODE_CURSOR_BO) {
 		if (!crtc->funcs->cursor_set) {
 			ret = -ENXIO;
 			goto out;
 		}
 		/* Turns off the cursor if handle is 0 */
-		mutex_lock(&crtc->mutex);
 		ret = crtc->funcs->cursor_set(crtc, file_priv, req->handle,
 					      req->width, req->height);
-		mutex_unlock(&crtc->mutex);
 	}
 
 	if (req->flags & DRM_MODE_CURSOR_MOVE) {
 		if (crtc->funcs->cursor_move) {
-			drm_modeset_lock_all(dev);
 			ret = crtc->funcs->cursor_move(crtc, req->x, req->y);
-			drm_modeset_unlock_all(dev);
 		} else {
 			ret = -EFAULT;
 			goto out;
 		}
 	}
 out:
+	mutex_unlock(&crtc->mutex);
+
 	return ret;
 }
 
diff --git a/drivers/gpu/drm/radeon/radeon_cursor.c b/drivers/gpu/drm/radeon/radeon_cursor.c
index ad6df62..c1680e6 100644
--- a/drivers/gpu/drm/radeon/radeon_cursor.c
+++ b/drivers/gpu/drm/radeon/radeon_cursor.c
@@ -245,8 +245,14 @@ int radeon_crtc_cursor_move(struct drm_crtc *crtc,
 		int i = 0;
 		struct drm_crtc *crtc_p;
 
-		/* avivo cursor image can't end on 128 pixel boundary or
+		/*
+		 * avivo cursor image can't end on 128 pixel boundary or
 		 * go past the end of the frame if both crtcs are enabled
+		 *
+		 * NOTE: It is safe to access crtc->enabled of other crtcs
+		 * without holding either the mode_config lock or the other
+		 * crtc's lock as long as write access to this flag _always_
+		 * grabs all locks.
 		 */
 		list_for_each_entry(crtc_p, &crtc->dev->mode_config.crtc_list, head) {
 			if (crtc_p->enabled)
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
index 8d82e63..3e3c7ab 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
@@ -264,10 +264,23 @@ int vmw_du_crtc_cursor_move(struct drm_crtc *crtc, int x, int y)
 	du->cursor_x = x + crtc->x;
 	du->cursor_y = y + crtc->y;
 
+	/*
+	 * FIXME: Unclear whether there's any global state touched by the
+	 * cursor_set function, especially vmw_cursor_update_position looks
+	 * suspicious. For now take the easy route and reacquire all locks. We
+	 * can do this since the caller in the drm core doesn't check anything
+	 * which is protected by any looks.
+	 */
+	mutex_unlock(&crtc->mutex);
+	drm_modeset_lock_all(dev_priv->dev);
+
 	vmw_cursor_update_position(dev_priv, shown,
 				   du->cursor_x + du->hotspot_x,
 				   du->cursor_y + du->hotspot_y);
 
+	drm_modeset_unlock_all(dev_priv->dev);
+	mutex_lock(&crtc->mutex);
+
 	return 0;
 }
 
-- 
cgit v0.10.2


From 4b096ac10da0b63f09bd123b86fed8deb80646ce Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Mon, 10 Dec 2012 21:19:18 +0100
Subject: drm: revamp locking around fb creation/destruction

Well, at least step 1. The goal here is that framebuffer objects can
survive outside of the mode_config lock, with just a reference held
as protection. The first step to get there is to introduce a special
fb_lock which protects fb lookup, creation and destruction, to make
them appear atomic.

This new fb_lock can nest within the mode_config lock. But the idea is
(once the reference counting part is completed) that we only quickly
take that fb_lock to lookup a framebuffer and grab a reference,
without any other locks involved.

vmwgfx is the only driver which does framebuffer lookups itself, also
wrap those calls to drm_mode_object_find with the new lock.

Also protect the fb_list walking in i915 and omapdrm with the new lock.

As a slight complication there's also the list of user-created fbs
attached to the file private. The problem now is that at fclose() time
we need to walk that list, eventually do a modeset call to remove the
fb from active usage (and are required to be able to take the
mode_config lock), but in the end we need to grab the new fb_lock to
remove the fb from the list. The easiest solution is to add another
mutex to protect this per-file list.

Currently that new fbs_lock nests within the modeset locks and so
appears redudant. But later patches will switch around this sequence
so that taking the modeset locks in the fb destruction path is
optional in the fastpath. Ultimately the goal is that addfb and rmfb
do not require the mode_config lock, since otherwise they have the
potential to introduce stalls in the pageflip sequence of a compositor
(if the compositor e.g. switches to a fullscreen client or if it
enables a plane). But that requires a few more steps and hoops to jump
through.

Note that framebuffer creation/destruction is now double-protected -
once by the fb_lock and in parts by the idr_lock. The later would be
unnecessariy if framebuffers would have their own idr allocator. But
that's material for another patch (series).

v2: Properly initialize the fb->filp_head list in _init, otherwise the
newly added WARN to check whether the fb isn't on a fpriv list any
more will fail for driver-private objects.

v3: Fixup two error-case unlock bugs spotted by Richard Wilbur.

Reviewed-by: Rob Clark <rob@ti.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index 4af6a3d..13a3d34 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -262,15 +262,21 @@ again:
 
 	mutex_lock(&dev->mode_config.idr_mutex);
 	ret = idr_get_new_above(&dev->mode_config.crtc_idr, obj, 1, &new_id);
+
+	if (!ret) {
+		/*
+		 * Set up the object linking under the protection of the idr
+		 * lock so that other users can't see inconsistent state.
+		 */
+		obj->id = new_id;
+		obj->type = obj_type;
+	}
 	mutex_unlock(&dev->mode_config.idr_mutex);
+
 	if (ret == -EAGAIN)
 		goto again;
-	else if (ret)
-		return ret;
 
-	obj->id = new_id;
-	obj->type = obj_type;
-	return 0;
+	return ret;
 }
 
 /**
@@ -312,6 +318,12 @@ EXPORT_SYMBOL(drm_mode_object_find);
  * Allocates an ID for the framebuffer's parent mode object, sets its mode
  * functions & device file and adds it to the master fd list.
  *
+ * IMPORTANT:
+ * This functions publishes the fb and makes it available for concurrent access
+ * by other users. Which means by this point the fb _must_ be fully set up -
+ * since all the fb attributes are invariant over its lifetime, no further
+ * locking but only correct reference counting is required.
+ *
  * RETURNS:
  * Zero on success, error code on failure.
  */
@@ -320,16 +332,20 @@ int drm_framebuffer_init(struct drm_device *dev, struct drm_framebuffer *fb,
 {
 	int ret;
 
+	mutex_lock(&dev->mode_config.fb_lock);
 	kref_init(&fb->refcount);
+	INIT_LIST_HEAD(&fb->filp_head);
+	fb->dev = dev;
+	fb->funcs = funcs;
 
 	ret = drm_mode_object_get(dev, &fb->base, DRM_MODE_OBJECT_FB);
 	if (ret)
-		return ret;
+		goto out;
 
-	fb->dev = dev;
-	fb->funcs = funcs;
 	dev->mode_config.num_fb++;
 	list_add(&fb->head, &dev->mode_config.fb_list);
+out:
+	mutex_unlock(&dev->mode_config.fb_lock);
 
 	return 0;
 }
@@ -385,8 +401,10 @@ void drm_framebuffer_cleanup(struct drm_framebuffer *fb)
 	 * this.)
 	 */
 	drm_mode_object_put(dev, &fb->base);
+	mutex_lock(&dev->mode_config.fb_lock);
 	list_del(&fb->head);
 	dev->mode_config.num_fb--;
+	mutex_unlock(&dev->mode_config.fb_lock);
 }
 EXPORT_SYMBOL(drm_framebuffer_cleanup);
 
@@ -406,6 +424,7 @@ void drm_framebuffer_remove(struct drm_framebuffer *fb)
 	int ret;
 
 	WARN_ON(!drm_modeset_is_locked(dev));
+	WARN_ON(!list_empty(&fb->filp_head));
 
 	/* remove from any CRTC */
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
@@ -432,8 +451,6 @@ void drm_framebuffer_remove(struct drm_framebuffer *fb)
 		}
 	}
 
-	list_del(&fb->filp_head);
-
 	drm_framebuffer_unreference(fb);
 }
 EXPORT_SYMBOL(drm_framebuffer_remove);
@@ -989,6 +1006,7 @@ void drm_mode_config_init(struct drm_device *dev)
 {
 	mutex_init(&dev->mode_config.mutex);
 	mutex_init(&dev->mode_config.idr_mutex);
+	mutex_init(&dev->mode_config.fb_lock);
 	INIT_LIST_HEAD(&dev->mode_config.fb_list);
 	INIT_LIST_HEAD(&dev->mode_config.crtc_list);
 	INIT_LIST_HEAD(&dev->mode_config.connector_list);
@@ -1091,6 +1109,9 @@ void drm_mode_config_cleanup(struct drm_device *dev)
 		drm_property_destroy(dev, property);
 	}
 
+	/* Single-threaded teardown context, so it's not requied to grab the
+	 * fb_lock to protect against concurrent fb_list access. Contrary, it
+	 * would actually deadlock with the drm_framebuffer_cleanup function. */
 	list_for_each_entry_safe(fb, fbt, &dev->mode_config.fb_list, head) {
 		drm_framebuffer_remove(fb);
 	}
@@ -1220,8 +1241,8 @@ int drm_mode_getresources(struct drm_device *dev, void *data,
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	drm_modeset_lock_all(dev);
 
+	mutex_lock(&file_priv->fbs_lock);
 	/*
 	 * For the non-control nodes we need to limit the list of resources
 	 * by IDs in the group list for this node
@@ -1229,6 +1250,23 @@ int drm_mode_getresources(struct drm_device *dev, void *data,
 	list_for_each(lh, &file_priv->fbs)
 		fb_count++;
 
+	/* handle this in 4 parts */
+	/* FBs */
+	if (card_res->count_fbs >= fb_count) {
+		copied = 0;
+		fb_id = (uint32_t __user *)(unsigned long)card_res->fb_id_ptr;
+		list_for_each_entry(fb, &file_priv->fbs, filp_head) {
+			if (put_user(fb->base.id, fb_id + copied)) {
+				mutex_unlock(&file_priv->fbs_lock);
+				return -EFAULT;
+			}
+			copied++;
+		}
+	}
+	card_res->count_fbs = fb_count;
+	mutex_unlock(&file_priv->fbs_lock);
+
+	drm_modeset_lock_all(dev);
 	mode_group = &file_priv->master->minor->mode_group;
 	if (file_priv->master->minor->type == DRM_MINOR_CONTROL) {
 
@@ -1252,21 +1290,6 @@ int drm_mode_getresources(struct drm_device *dev, void *data,
 	card_res->max_width = dev->mode_config.max_width;
 	card_res->min_width = dev->mode_config.min_width;
 
-	/* handle this in 4 parts */
-	/* FBs */
-	if (card_res->count_fbs >= fb_count) {
-		copied = 0;
-		fb_id = (uint32_t __user *)(unsigned long)card_res->fb_id_ptr;
-		list_for_each_entry(fb, &file_priv->fbs, filp_head) {
-			if (put_user(fb->base.id, fb_id + copied)) {
-				ret = -EFAULT;
-				goto out;
-			}
-			copied++;
-		}
-	}
-	card_res->count_fbs = fb_count;
-
 	/* CRTCs */
 	if (card_res->count_crtcs >= crtc_count) {
 		copied = 0;
@@ -1765,8 +1788,10 @@ int drm_mode_setplane(struct drm_device *dev, void *data,
 	}
 	crtc = obj_to_crtc(obj);
 
+	mutex_lock(&dev->mode_config.fb_lock);
 	obj = drm_mode_object_find(dev, plane_req->fb_id,
 				   DRM_MODE_OBJECT_FB);
+	mutex_unlock(&dev->mode_config.fb_lock);
 	if (!obj) {
 		DRM_DEBUG_KMS("Unknown framebuffer ID %d\n",
 			      plane_req->fb_id);
@@ -1908,8 +1933,10 @@ int drm_mode_setcrtc(struct drm_device *dev, void *data,
 			}
 			fb = crtc->fb;
 		} else {
+			mutex_lock(&dev->mode_config.fb_lock);
 			obj = drm_mode_object_find(dev, crtc_req->fb_id,
 						   DRM_MODE_OBJECT_FB);
+			mutex_unlock(&dev->mode_config.fb_lock);
 			if (!obj) {
 				DRM_DEBUG_KMS("Unknown FB ID%d\n",
 						crtc_req->fb_id);
@@ -2151,16 +2178,17 @@ int drm_mode_addfb(struct drm_device *dev,
 	fb = dev->mode_config.funcs->fb_create(dev, file_priv, &r);
 	if (IS_ERR(fb)) {
 		DRM_DEBUG_KMS("could not create framebuffer\n");
-		ret = PTR_ERR(fb);
-		goto out;
+		drm_modeset_unlock_all(dev);
+		return PTR_ERR(fb);
 	}
 
+	mutex_lock(&file_priv->fbs_lock);
 	or->fb_id = fb->base.id;
 	list_add(&fb->filp_head, &file_priv->fbs);
 	DRM_DEBUG_KMS("[FB:%d]\n", fb->base.id);
-
-out:
+	mutex_unlock(&file_priv->fbs_lock);
 	drm_modeset_unlock_all(dev);
+
 	return ret;
 }
 
@@ -2333,16 +2361,18 @@ int drm_mode_addfb2(struct drm_device *dev,
 	fb = dev->mode_config.funcs->fb_create(dev, file_priv, r);
 	if (IS_ERR(fb)) {
 		DRM_DEBUG_KMS("could not create framebuffer\n");
-		ret = PTR_ERR(fb);
-		goto out;
+		drm_modeset_unlock_all(dev);
+		return PTR_ERR(fb);
 	}
 
+	mutex_lock(&file_priv->fbs_lock);
 	r->fb_id = fb->base.id;
 	list_add(&fb->filp_head, &file_priv->fbs);
 	DRM_DEBUG_KMS("[FB:%d]\n", fb->base.id);
+	mutex_unlock(&file_priv->fbs_lock);
 
-out:
 	drm_modeset_unlock_all(dev);
+
 	return ret;
 }
 
@@ -2373,27 +2403,34 @@ int drm_mode_rmfb(struct drm_device *dev,
 		return -EINVAL;
 
 	drm_modeset_lock_all(dev);
+	mutex_lock(&dev->mode_config.fb_lock);
 	obj = drm_mode_object_find(dev, *id, DRM_MODE_OBJECT_FB);
 	/* TODO check that we really get a framebuffer back. */
 	if (!obj) {
+		mutex_unlock(&dev->mode_config.fb_lock);
 		ret = -EINVAL;
 		goto out;
 	}
 	fb = obj_to_fb(obj);
+	mutex_unlock(&dev->mode_config.fb_lock);
 
+	mutex_lock(&file_priv->fbs_lock);
 	list_for_each_entry(fbl, &file_priv->fbs, filp_head)
 		if (fb == fbl)
 			found = 1;
-
 	if (!found) {
 		ret = -EINVAL;
+		mutex_unlock(&file_priv->fbs_lock);
 		goto out;
 	}
 
-	drm_framebuffer_remove(fb);
+	list_del_init(&fb->filp_head);
+	mutex_unlock(&file_priv->fbs_lock);
 
+	drm_framebuffer_remove(fb);
 out:
 	drm_modeset_unlock_all(dev);
+
 	return ret;
 }
 
@@ -2422,7 +2459,9 @@ int drm_mode_getfb(struct drm_device *dev,
 		return -EINVAL;
 
 	drm_modeset_lock_all(dev);
+	mutex_lock(&dev->mode_config.fb_lock);
 	obj = drm_mode_object_find(dev, r->fb_id, DRM_MODE_OBJECT_FB);
+	mutex_unlock(&dev->mode_config.fb_lock);
 	if (!obj) {
 		ret = -EINVAL;
 		goto out;
@@ -2460,7 +2499,9 @@ int drm_mode_dirtyfb_ioctl(struct drm_device *dev,
 		return -EINVAL;
 
 	drm_modeset_lock_all(dev);
+	mutex_lock(&dev->mode_config.fb_lock);
 	obj = drm_mode_object_find(dev, r->fb_id, DRM_MODE_OBJECT_FB);
+	mutex_unlock(&dev->mode_config.fb_lock);
 	if (!obj) {
 		ret = -EINVAL;
 		goto out_err1;
@@ -2535,9 +2576,12 @@ void drm_fb_release(struct drm_file *priv)
 	struct drm_framebuffer *fb, *tfb;
 
 	drm_modeset_lock_all(dev);
+	mutex_lock(&priv->fbs_lock);
 	list_for_each_entry_safe(fb, tfb, &priv->fbs, filp_head) {
+		list_del_init(&fb->filp_head);
 		drm_framebuffer_remove(fb);
 	}
+	mutex_unlock(&priv->fbs_lock);
 	drm_modeset_unlock_all(dev);
 }
 
@@ -3542,7 +3586,9 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,
 	if (crtc->funcs->page_flip == NULL)
 		goto out;
 
+	mutex_lock(&dev->mode_config.fb_lock);
 	obj = drm_mode_object_find(dev, page_flip->fb_id, DRM_MODE_OBJECT_FB);
+	mutex_unlock(&dev->mode_config.fb_lock);
 	if (!obj)
 		goto out;
 	fb = obj_to_fb(obj);
diff --git a/drivers/gpu/drm/drm_fops.c b/drivers/gpu/drm/drm_fops.c
index 133b413..13fdcd1 100644
--- a/drivers/gpu/drm/drm_fops.c
+++ b/drivers/gpu/drm/drm_fops.c
@@ -276,6 +276,7 @@ static int drm_open_helper(struct inode *inode, struct file *filp,
 
 	INIT_LIST_HEAD(&priv->lhead);
 	INIT_LIST_HEAD(&priv->fbs);
+	mutex_init(&priv->fbs_lock);
 	INIT_LIST_HEAD(&priv->event_list);
 	init_waitqueue_head(&priv->event_wait);
 	priv->event_space = 4096; /* set aside 4k for event buffer */
diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index e6a11ca8..a40c674 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -1374,7 +1374,9 @@ static int i915_gem_framebuffer_info(struct seq_file *m, void *data)
 		   fb->base.bits_per_pixel);
 	describe_obj(m, fb->obj);
 	seq_printf(m, "\n");
+	mutex_unlock(&dev->mode_config.mutex);
 
+	mutex_lock(&dev->mode_config.fb_lock);
 	list_for_each_entry(fb, &dev->mode_config.fb_list, base.head) {
 		if (&fb->base == ifbdev->helper.fb)
 			continue;
@@ -1387,8 +1389,7 @@ static int i915_gem_framebuffer_info(struct seq_file *m, void *data)
 		describe_obj(m, fb->obj);
 		seq_printf(m, "\n");
 	}
-
-	mutex_unlock(&dev->mode_config.mutex);
+	mutex_unlock(&dev->mode_config.fb_lock);
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c
index a135498..0d6a161 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c
@@ -165,7 +165,9 @@ int vmw_present_ioctl(struct drm_device *dev, void *data,
 
 	drm_modeset_lock_all(dev);
 
+	mutex_lock(&dev->mode_config.fb_lock);
 	obj = drm_mode_object_find(dev, arg->fb_id, DRM_MODE_OBJECT_FB);
+	mutex_unlock(&dev->mode_config.fb_lock);
 	if (!obj) {
 		DRM_ERROR("Invalid framebuffer id.\n");
 		ret = -EINVAL;
@@ -248,7 +250,9 @@ int vmw_present_readback_ioctl(struct drm_device *dev, void *data,
 
 	drm_modeset_lock_all(dev);
 
+	mutex_lock(&dev->mode_config.fb_lock);
 	obj = drm_mode_object_find(dev, arg->fb_id, DRM_MODE_OBJECT_FB);
+	mutex_unlock(&dev->mode_config.fb_lock);
 	if (!obj) {
 		DRM_ERROR("Invalid framebuffer id.\n");
 		ret = -EINVAL;
diff --git a/drivers/staging/omapdrm/omap_debugfs.c b/drivers/staging/omapdrm/omap_debugfs.c
index 2f122e0..e95540b 100644
--- a/drivers/staging/omapdrm/omap_debugfs.c
+++ b/drivers/staging/omapdrm/omap_debugfs.c
@@ -72,6 +72,7 @@ static int fb_show(struct seq_file *m, void *arg)
 	seq_printf(m, "fbcon ");
 	omap_framebuffer_describe(priv->fbdev->fb, m);
 
+	mutex_lock(&dev->mode_config.fb_lock);
 	list_for_each_entry(fb, &dev->mode_config.fb_list, head) {
 		if (fb == priv->fbdev->fb)
 			continue;
@@ -79,6 +80,7 @@ static int fb_show(struct seq_file *m, void *arg)
 		seq_printf(m, "user ");
 		omap_framebuffer_describe(fb, m);
 	}
+	mutex_unlock(&dev->mode_config.fb_lock);
 
 	mutex_unlock(&dev->struct_mutex);
 	mutex_unlock(&dev->mode_config.mutex);
diff --git a/include/drm/drmP.h b/include/drm/drmP.h
index 3c609ab..e74731c 100644
--- a/include/drm/drmP.h
+++ b/include/drm/drmP.h
@@ -446,7 +446,15 @@ struct drm_file {
 	int is_master; /* this file private is a master for a minor */
 	struct drm_master *master; /* master this node is currently associated with
 				      N.B. not always minor->master */
+
+	/**
+	 * fbs - List of framebuffers associated with this file.
+	 *
+	 * Protected by fbs_lock. Note that the fbs list holds a reference on
+	 * the fb object to prevent it from untimely disappearing.
+	 */
 	struct list_head fbs;
+	struct mutex fbs_lock;
 
 	wait_queue_head_t event_wait;
 	struct list_head event_list;
diff --git a/include/drm/drm_crtc.h b/include/drm/drm_crtc.h
index c89b116..c35a807 100644
--- a/include/drm/drm_crtc.h
+++ b/include/drm/drm_crtc.h
@@ -254,6 +254,10 @@ struct drm_framebuffer {
 	 * userspace perspective.
 	 */
 	struct kref refcount;
+	/*
+	 * Place on the dev->mode_config.fb_list, access protected by
+	 * dev->mode_config.fb_lock.
+	 */
 	struct list_head head;
 	struct drm_mode_object base;
 	const struct drm_framebuffer_funcs *funcs;
@@ -780,8 +784,18 @@ struct drm_mode_config {
 	struct mutex idr_mutex; /* for IDR management */
 	struct idr crtc_idr; /* use this idr for all IDs, fb, crtc, connector, modes - just makes life easier */
 	/* this is limited to one for now */
+
+
+	/**
+	 * fb_lock - mutex to protect fb state
+	 *
+	 * Besides the global fb list his also protects the fbs list in the
+	 * file_priv
+	 */
+	struct mutex fb_lock;
 	int num_fb;
 	struct list_head fb_list;
+
 	int num_connector;
 	struct list_head connector_list;
 	int num_encoder;
-- 
cgit v0.10.2


From 786b99ed13223d8ac58a937dd348aead45eb8191 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sun, 2 Dec 2012 21:53:40 +0100
Subject: drm: create drm_framebuffer_lookup

And replace all fb lookups with it. Also add a WARN to
drm_mode_object_find since that is now no longer the blessed interface
to look up an fb. And add kerneldoc to both functions.

This only updates all callsites, but immediately drops the acquired
refence again. Hence all callers still rely on the fact that a mode fb
can't disappear while they're holding the struct mutex. Subsequent
patches will instate proper use of refcounts, and then rework the rmfb
and unref code to no longer serialize fb destruction with the
mode_config lock. We don't want that since otherwise a compositor
might end up stalling for a few frames in rmfb.

v2: Don't use kref_get_unless_zero - Greg KH doesn't like that kind of
interface.

Reviewed-by: Rob Clark <rob@ti.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index 13a3d34..f2ccda8 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -294,11 +294,24 @@ static void drm_mode_object_put(struct drm_device *dev,
 	mutex_unlock(&dev->mode_config.idr_mutex);
 }
 
+/**
+ * drm_mode_object_find - look up a drm object with static lifetime
+ * @dev: drm device
+ * @id: id of the mode object
+ * @type: type of the mode object
+ *
+ * Note that framebuffers cannot be looked up with this functions - since those
+ * are reference counted, they need special treatment.
+ */
 struct drm_mode_object *drm_mode_object_find(struct drm_device *dev,
 		uint32_t id, uint32_t type)
 {
 	struct drm_mode_object *obj = NULL;
 
+	/* Framebuffers are reference counted and need their own lookup
+	 * function.*/
+	WARN_ON(type == DRM_MODE_OBJECT_FB);
+
 	mutex_lock(&dev->mode_config.idr_mutex);
 	obj = idr_find(&dev->mode_config.crtc_idr, id);
 	if (!obj || (obj->type != type) || (obj->id != id))
@@ -359,6 +372,40 @@ static void drm_framebuffer_free(struct kref *kref)
 }
 
 /**
+ * drm_framebuffer_lookup - look up a drm framebuffer and grab a reference
+ * @dev: drm device
+ * @id: id of the fb object
+ *
+ * If successful, this grabs an additional reference to the framebuffer -
+ * callers need to make sure to eventually unreference the returned framebuffer
+ * again.
+ */
+struct drm_framebuffer *drm_framebuffer_lookup(struct drm_device *dev,
+					       uint32_t id)
+{
+	struct drm_mode_object *obj = NULL;
+	struct drm_framebuffer *fb;
+
+	mutex_lock(&dev->mode_config.fb_lock);
+
+	mutex_lock(&dev->mode_config.idr_mutex);
+	obj = idr_find(&dev->mode_config.crtc_idr, id);
+	if (!obj || (obj->type != DRM_MODE_OBJECT_FB) || (obj->id != id))
+		fb = NULL;
+	else
+		fb = obj_to_fb(obj);
+	mutex_unlock(&dev->mode_config.idr_mutex);
+
+	if (fb)
+		kref_get(&fb->refcount);
+
+	mutex_unlock(&dev->mode_config.fb_lock);
+
+	return fb;
+}
+EXPORT_SYMBOL(drm_framebuffer_lookup);
+
+/**
  * drm_framebuffer_unreference - unref a framebuffer
  * @fb: framebuffer to unref
  *
@@ -1788,17 +1835,15 @@ int drm_mode_setplane(struct drm_device *dev, void *data,
 	}
 	crtc = obj_to_crtc(obj);
 
-	mutex_lock(&dev->mode_config.fb_lock);
-	obj = drm_mode_object_find(dev, plane_req->fb_id,
-				   DRM_MODE_OBJECT_FB);
-	mutex_unlock(&dev->mode_config.fb_lock);
-	if (!obj) {
+	fb = drm_framebuffer_lookup(dev, plane_req->fb_id);
+	if (!fb) {
 		DRM_DEBUG_KMS("Unknown framebuffer ID %d\n",
 			      plane_req->fb_id);
 		ret = -ENOENT;
 		goto out;
 	}
-	fb = obj_to_fb(obj);
+	/* fb is protect by the mode_config lock, so drop the ref immediately */
+	drm_framebuffer_unreference(fb);
 
 	/* Check whether this plane supports the fb pixel format. */
 	for (i = 0; i < plane->format_count; i++)
@@ -1933,17 +1978,16 @@ int drm_mode_setcrtc(struct drm_device *dev, void *data,
 			}
 			fb = crtc->fb;
 		} else {
-			mutex_lock(&dev->mode_config.fb_lock);
-			obj = drm_mode_object_find(dev, crtc_req->fb_id,
-						   DRM_MODE_OBJECT_FB);
-			mutex_unlock(&dev->mode_config.fb_lock);
-			if (!obj) {
+			fb = drm_framebuffer_lookup(dev, crtc_req->fb_id);
+			if (!fb) {
 				DRM_DEBUG_KMS("Unknown FB ID%d\n",
 						crtc_req->fb_id);
 				ret = -EINVAL;
 				goto out;
 			}
-			fb = obj_to_fb(obj);
+			/* fb is protect by the mode_config lock, so drop the
+			 * ref immediately */
+			drm_framebuffer_unreference(fb);
 		}
 
 		mode = drm_mode_create(dev);
@@ -2392,7 +2436,6 @@ int drm_mode_addfb2(struct drm_device *dev,
 int drm_mode_rmfb(struct drm_device *dev,
 		   void *data, struct drm_file *file_priv)
 {
-	struct drm_mode_object *obj;
 	struct drm_framebuffer *fb = NULL;
 	struct drm_framebuffer *fbl = NULL;
 	uint32_t *id = data;
@@ -2403,16 +2446,13 @@ int drm_mode_rmfb(struct drm_device *dev,
 		return -EINVAL;
 
 	drm_modeset_lock_all(dev);
-	mutex_lock(&dev->mode_config.fb_lock);
-	obj = drm_mode_object_find(dev, *id, DRM_MODE_OBJECT_FB);
-	/* TODO check that we really get a framebuffer back. */
-	if (!obj) {
-		mutex_unlock(&dev->mode_config.fb_lock);
+	fb = drm_framebuffer_lookup(dev, *id);
+	if (!fb) {
 		ret = -EINVAL;
 		goto out;
 	}
-	fb = obj_to_fb(obj);
-	mutex_unlock(&dev->mode_config.fb_lock);
+	/* fb is protect by the mode_config lock, so drop the ref immediately */
+	drm_framebuffer_unreference(fb);
 
 	mutex_lock(&file_priv->fbs_lock);
 	list_for_each_entry(fbl, &file_priv->fbs, filp_head)
@@ -2451,7 +2491,6 @@ int drm_mode_getfb(struct drm_device *dev,
 		   void *data, struct drm_file *file_priv)
 {
 	struct drm_mode_fb_cmd *r = data;
-	struct drm_mode_object *obj;
 	struct drm_framebuffer *fb;
 	int ret = 0;
 
@@ -2459,14 +2498,13 @@ int drm_mode_getfb(struct drm_device *dev,
 		return -EINVAL;
 
 	drm_modeset_lock_all(dev);
-	mutex_lock(&dev->mode_config.fb_lock);
-	obj = drm_mode_object_find(dev, r->fb_id, DRM_MODE_OBJECT_FB);
-	mutex_unlock(&dev->mode_config.fb_lock);
-	if (!obj) {
+	fb = drm_framebuffer_lookup(dev, r->fb_id);
+	if (!fb) {
 		ret = -EINVAL;
 		goto out;
 	}
-	fb = obj_to_fb(obj);
+	/* fb is protect by the mode_config lock, so drop the ref immediately */
+	drm_framebuffer_unreference(fb);
 
 	r->height = fb->height;
 	r->width = fb->width;
@@ -2489,7 +2527,6 @@ int drm_mode_dirtyfb_ioctl(struct drm_device *dev,
 	struct drm_clip_rect __user *clips_ptr;
 	struct drm_clip_rect *clips = NULL;
 	struct drm_mode_fb_dirty_cmd *r = data;
-	struct drm_mode_object *obj;
 	struct drm_framebuffer *fb;
 	unsigned flags;
 	int num_clips;
@@ -2499,14 +2536,13 @@ int drm_mode_dirtyfb_ioctl(struct drm_device *dev,
 		return -EINVAL;
 
 	drm_modeset_lock_all(dev);
-	mutex_lock(&dev->mode_config.fb_lock);
-	obj = drm_mode_object_find(dev, r->fb_id, DRM_MODE_OBJECT_FB);
-	mutex_unlock(&dev->mode_config.fb_lock);
-	if (!obj) {
+	fb = drm_framebuffer_lookup(dev, r->fb_id);
+	if (!fb) {
 		ret = -EINVAL;
 		goto out_err1;
 	}
-	fb = obj_to_fb(obj);
+	/* fb is protect by the mode_config lock, so drop the ref immediately */
+	drm_framebuffer_unreference(fb);
 
 	num_clips = r->num_clips;
 	clips_ptr = (struct drm_clip_rect __user *)(unsigned long)r->clips_ptr;
@@ -3586,12 +3622,11 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,
 	if (crtc->funcs->page_flip == NULL)
 		goto out;
 
-	mutex_lock(&dev->mode_config.fb_lock);
-	obj = drm_mode_object_find(dev, page_flip->fb_id, DRM_MODE_OBJECT_FB);
-	mutex_unlock(&dev->mode_config.fb_lock);
-	if (!obj)
+	fb = drm_framebuffer_lookup(dev, page_flip->fb_id);
+	if (!fb)
 		goto out;
-	fb = obj_to_fb(obj);
+	/* fb is protect by the mode_config lock, so drop the ref immediately */
+	drm_framebuffer_unreference(fb);
 
 	hdisplay = crtc->mode.hdisplay;
 	vdisplay = crtc->mode.vdisplay;
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c
index 0d6a161..1b8f428 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c
@@ -131,7 +131,7 @@ int vmw_present_ioctl(struct drm_device *dev, void *data,
 	struct vmw_master *vmaster = vmw_master(file_priv->master);
 	struct drm_vmw_rect __user *clips_ptr;
 	struct drm_vmw_rect *clips = NULL;
-	struct drm_mode_object *obj;
+	struct drm_framebuffer *fb;
 	struct vmw_framebuffer *vfb;
 	struct vmw_resource *res;
 	uint32_t num_clips;
@@ -165,15 +165,15 @@ int vmw_present_ioctl(struct drm_device *dev, void *data,
 
 	drm_modeset_lock_all(dev);
 
-	mutex_lock(&dev->mode_config.fb_lock);
-	obj = drm_mode_object_find(dev, arg->fb_id, DRM_MODE_OBJECT_FB);
-	mutex_unlock(&dev->mode_config.fb_lock);
-	if (!obj) {
+	fb = drm_framebuffer_lookup(dev, arg->fb_id);
+	if (!fb) {
 		DRM_ERROR("Invalid framebuffer id.\n");
 		ret = -EINVAL;
 		goto out_no_fb;
 	}
-	vfb = vmw_framebuffer_to_vfb(obj_to_fb(obj));
+	/* fb is protect by the mode_config lock, so drop the ref immediately */
+	drm_framebuffer_unreference(fb);
+	vfb = vmw_framebuffer_to_vfb(fb);
 
 	ret = ttm_read_lock(&vmaster->lock, true);
 	if (unlikely(ret != 0))
@@ -217,7 +217,7 @@ int vmw_present_readback_ioctl(struct drm_device *dev, void *data,
 	struct vmw_master *vmaster = vmw_master(file_priv->master);
 	struct drm_vmw_rect __user *clips_ptr;
 	struct drm_vmw_rect *clips = NULL;
-	struct drm_mode_object *obj;
+	struct drm_framebuffer *fb;
 	struct vmw_framebuffer *vfb;
 	uint32_t num_clips;
 	int ret;
@@ -250,16 +250,16 @@ int vmw_present_readback_ioctl(struct drm_device *dev, void *data,
 
 	drm_modeset_lock_all(dev);
 
-	mutex_lock(&dev->mode_config.fb_lock);
-	obj = drm_mode_object_find(dev, arg->fb_id, DRM_MODE_OBJECT_FB);
-	mutex_unlock(&dev->mode_config.fb_lock);
-	if (!obj) {
+	fb = drm_framebuffer_lookup(dev, arg->fb_id);
+	if (!fb) {
 		DRM_ERROR("Invalid framebuffer id.\n");
 		ret = -EINVAL;
 		goto out_no_fb;
 	}
+	/* fb is protect by the mode_config lock, so drop the ref immediately */
+	drm_framebuffer_unreference(fb);
 
-	vfb = vmw_framebuffer_to_vfb(obj_to_fb(obj));
+	vfb = vmw_framebuffer_to_vfb(fb);
 	if (!vfb->dmabuf) {
 		DRM_ERROR("Framebuffer not dmabuf backed.\n");
 		ret = -EINVAL;
diff --git a/include/drm/drm_crtc.h b/include/drm/drm_crtc.h
index c35a807..7dc1b31 100644
--- a/include/drm/drm_crtc.h
+++ b/include/drm/drm_crtc.h
@@ -958,6 +958,8 @@ extern void drm_framebuffer_set_object(struct drm_device *dev,
 extern int drm_framebuffer_init(struct drm_device *dev,
 				struct drm_framebuffer *fb,
 				const struct drm_framebuffer_funcs *funcs);
+extern struct drm_framebuffer *drm_framebuffer_lookup(struct drm_device *dev,
+						      uint32_t id);
 extern void drm_framebuffer_unreference(struct drm_framebuffer *fb);
 extern void drm_framebuffer_reference(struct drm_framebuffer *fb);
 extern void drm_framebuffer_remove(struct drm_framebuffer *fb);
-- 
cgit v0.10.2


From 362063619cf67c2c2fc2eb90951b2623cbb69a7c Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Mon, 10 Dec 2012 20:42:17 +0100
Subject: drm: revamp framebuffer cleanup interfaces

We have two classes of framebuffer
- Created by the driver (atm only for fbdev), and the driver holds
  onto the last reference count until destruction.
- Created by userspace and associated with a given fd. These
  framebuffers will be reaped when their assoiciated fb is closed.

Now these two cases are set up differently, the framebuffers are on
different lists and hence destruction needs to clean up different
things. Also, for userspace framebuffers we remove them from any
current usage, whereas for internal framebuffers it is assumed that
the driver has done this already.

Long story short, we need two different ways to cleanup such drivers.
Three functions are involved in total:
- drm_framebuffer_remove: Convenience function which removes the fb
  from all active usage and then drops the passed-in reference.
- drm_framebuffer_unregister_private: Will remove driver-private
  framebuffers from relevant lists and drop the corresponding
  references. Should be called for driver-private framebuffers before
  dropping the last reference (or like for a lot of the drivers where
  the fbdev is embedded someplace else, before doing the cleanup
  manually).
- drm_framebuffer_cleanup: Final cleanup for both classes of fbs,
  should be called by the driver's ->destroy callback once the last
  reference is gone.

This patch just rolls out the new interfaces and updates all drivers
(by adding calls to drm_framebuffer_unregister_private at all the
right places)- no functional changes yet. Follow-on patches will move
drm core code around and update the lifetime management for
framebuffers, so that we are no longer required to keep framebuffers
alive by locking mode_config.mutex.

I've also updated the kerneldoc already.

vmwgfx seems to again be a bit special, at least I haven't figured out
how the fbdev support in that driver works. It smells like it's
external though.

v2: The i915 driver creates another private framebuffer in the
load-detect code. Adjust its cleanup code, too.

Reviewed-by: Rob Clark <rob@ti.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/ast/ast_fb.c b/drivers/gpu/drm/ast/ast_fb.c
index d9ec779..3e6584b 100644
--- a/drivers/gpu/drm/ast/ast_fb.c
+++ b/drivers/gpu/drm/ast/ast_fb.c
@@ -290,6 +290,7 @@ static void ast_fbdev_destroy(struct drm_device *dev,
 	drm_fb_helper_fini(&afbdev->helper);
 
 	vfree(afbdev->sysram);
+	drm_framebuffer_unregister_private(&afb->base);
 	drm_framebuffer_cleanup(&afb->base);
 }
 
diff --git a/drivers/gpu/drm/cirrus/cirrus_fbdev.c b/drivers/gpu/drm/cirrus/cirrus_fbdev.c
index 6c6b4c8..3daea0f 100644
--- a/drivers/gpu/drm/cirrus/cirrus_fbdev.c
+++ b/drivers/gpu/drm/cirrus/cirrus_fbdev.c
@@ -258,6 +258,7 @@ static int cirrus_fbdev_destroy(struct drm_device *dev,
 
 	vfree(gfbdev->sysram);
 	drm_fb_helper_fini(&gfbdev->helper);
+	drm_framebuffer_unregister_private(&gfb->base);
 	drm_framebuffer_cleanup(&gfb->base);
 
 	return 0;
diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index f2ccda8..3eddfab 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -68,6 +68,7 @@ void drm_modeset_unlock_all(struct drm_device *dev)
 
 	mutex_unlock(&dev->mode_config.mutex);
 }
+
 EXPORT_SYMBOL(drm_modeset_unlock_all);
 
 /* Avoid boilerplate.  I'm tired of typing. */
@@ -430,11 +431,34 @@ void drm_framebuffer_reference(struct drm_framebuffer *fb)
 EXPORT_SYMBOL(drm_framebuffer_reference);
 
 /**
+ * drm_framebuffer_unregister_private - unregister a private fb from the lookup idr
+ * @fb: fb to unregister
+ *
+ * Drivers need to call this when cleaning up driver-private framebuffers, e.g.
+ * those used for fbdev. Note that the caller must hold a reference of it's own,
+ * i.e. the object may not be destroyed through this call (since it'll lead to a
+ * locking inversion).
+ */
+void drm_framebuffer_unregister_private(struct drm_framebuffer *fb)
+{
+}
+EXPORT_SYMBOL(drm_framebuffer_unregister_private);
+
+/**
  * drm_framebuffer_cleanup - remove a framebuffer object
  * @fb: framebuffer to remove
  *
- * Scans all the CRTCs in @dev's mode_config.  If they're using @fb, removes
- * it, setting it to NULL.
+ * Cleanup references to a user-created framebuffer. This function is intended
+ * to be used from the drivers ->destroy callback.
+ *
+ * Note that this function does not remove the fb from active usuage - if it is
+ * still used anywhere, hilarity can ensue since userspace could call getfb on
+ * the id and get back -EINVAL. Obviously no concern at driver unload time.
+ *
+ * Also, the framebuffer will not be removed from the lookup idr - for
+ * user-created framebuffers this will happen in in the rmfb ioctl. For
+ * driver-private objects (e.g. for fbdev) drivers need to explicitly call
+ * drm_framebuffer_unregister_private.
  */
 void drm_framebuffer_cleanup(struct drm_framebuffer *fb)
 {
@@ -460,7 +484,8 @@ EXPORT_SYMBOL(drm_framebuffer_cleanup);
  * @fb: framebuffer to remove
  *
  * Scans all the CRTCs and planes in @dev's mode_config.  If they're
- * using @fb, removes it, setting it to NULL.
+ * using @fb, removes it, setting it to NULL. Then drops the reference to the
+ * passed-in framebuffer.
  */
 void drm_framebuffer_remove(struct drm_framebuffer *fb)
 {
diff --git a/drivers/gpu/drm/drm_fb_cma_helper.c b/drivers/gpu/drm/drm_fb_cma_helper.c
index e1e0cb0..3742bc9 100644
--- a/drivers/gpu/drm/drm_fb_cma_helper.c
+++ b/drivers/gpu/drm/drm_fb_cma_helper.c
@@ -266,6 +266,7 @@ static int drm_fbdev_cma_create(struct drm_fb_helper *helper,
 	return 0;
 
 err_drm_fb_cma_destroy:
+	drm_framebuffer_unregister_private(fb);
 	drm_fb_cma_destroy(fb);
 err_framebuffer_release:
 	framebuffer_release(fbi);
@@ -370,8 +371,10 @@ void drm_fbdev_cma_fini(struct drm_fbdev_cma *fbdev_cma)
 		framebuffer_release(info);
 	}
 
-	if (fbdev_cma->fb)
+	if (fbdev_cma->fb) {
+		drm_framebuffer_unregister_private(&fbdev_cma->fb->fb);
 		drm_fb_cma_destroy(&fbdev_cma->fb->fb);
+	}
 
 	drm_fb_helper_fini(&fbdev_cma->fb_helper);
 	kfree(fbdev_cma);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
index 71f8673..90d335c 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
@@ -326,8 +326,10 @@ static void exynos_drm_fbdev_destroy(struct drm_device *dev,
 	/* release drm framebuffer and real buffer */
 	if (fb_helper->fb && fb_helper->fb->funcs) {
 		fb = fb_helper->fb;
-		if (fb)
+		if (fb) {
+			drm_framebuffer_unregister_private(fb);
 			drm_framebuffer_remove(fb);
+		}
 	}
 
 	/* release linux framebuffer */
diff --git a/drivers/gpu/drm/gma500/framebuffer.c b/drivers/gpu/drm/gma500/framebuffer.c
index 49800d2..c1ef37e 100644
--- a/drivers/gpu/drm/gma500/framebuffer.c
+++ b/drivers/gpu/drm/gma500/framebuffer.c
@@ -590,6 +590,7 @@ static int psb_fbdev_destroy(struct drm_device *dev, struct psb_fbdev *fbdev)
 		framebuffer_release(info);
 	}
 	drm_fb_helper_fini(&fbdev->psb_fb_helper);
+	drm_framebuffer_unregister_private(&psbfb->base);
 	drm_framebuffer_cleanup(&psbfb->base);
 
 	if (psbfb->gtt)
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 26fa6a7..df51203 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -6789,8 +6789,10 @@ void intel_release_load_detect_pipe(struct drm_connector *connector,
 		intel_encoder->new_crtc = NULL;
 		intel_set_mode(crtc, NULL, 0, 0, NULL);
 
-		if (old->release_fb)
-			old->release_fb->funcs->destroy(old->release_fb);
+		if (old->release_fb) {
+			drm_framebuffer_unregister_private(old->release_fb);
+			drm_framebuffer_unreference(old->release_fb);
+		}
 
 		return;
 	}
diff --git a/drivers/gpu/drm/i915/intel_fb.c b/drivers/gpu/drm/i915/intel_fb.c
index a9f7280..c7b8316 100644
--- a/drivers/gpu/drm/i915/intel_fb.c
+++ b/drivers/gpu/drm/i915/intel_fb.c
@@ -212,6 +212,7 @@ static void intel_fbdev_destroy(struct drm_device *dev,
 
 	drm_fb_helper_fini(&ifbdev->helper);
 
+	drm_framebuffer_unregister_private(&ifb->base);
 	drm_framebuffer_cleanup(&ifb->base);
 	if (ifb->obj) {
 		drm_gem_object_unreference_unlocked(&ifb->obj->base);
diff --git a/drivers/gpu/drm/mgag200/mgag200_fb.c b/drivers/gpu/drm/mgag200/mgag200_fb.c
index 2f48648..5c69b43 100644
--- a/drivers/gpu/drm/mgag200/mgag200_fb.c
+++ b/drivers/gpu/drm/mgag200/mgag200_fb.c
@@ -247,6 +247,7 @@ static int mga_fbdev_destroy(struct drm_device *dev,
 	}
 	drm_fb_helper_fini(&mfbdev->helper);
 	vfree(mfbdev->sysram);
+	drm_framebuffer_unregister_private(&mfb->base);
 	drm_framebuffer_cleanup(&mfb->base);
 
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/nouveau_fbcon.c b/drivers/gpu/drm/nouveau/nouveau_fbcon.c
index 67a1a06..d4ecb4d 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fbcon.c
+++ b/drivers/gpu/drm/nouveau/nouveau_fbcon.c
@@ -433,6 +433,7 @@ nouveau_fbcon_destroy(struct drm_device *dev, struct nouveau_fbdev *fbcon)
 		nouveau_fb->nvbo = NULL;
 	}
 	drm_fb_helper_fini(&fbcon->helper);
+	drm_framebuffer_unregister_private(&nouveau_fb->base);
 	drm_framebuffer_cleanup(&nouveau_fb->base);
 	return 0;
 }
diff --git a/drivers/gpu/drm/radeon/radeon_fb.c b/drivers/gpu/drm/radeon/radeon_fb.c
index cc8489d..515e5ee 100644
--- a/drivers/gpu/drm/radeon/radeon_fb.c
+++ b/drivers/gpu/drm/radeon/radeon_fb.c
@@ -293,6 +293,7 @@ out_unref:
 	}
 	if (fb && ret) {
 		drm_gem_object_unreference(gobj);
+		drm_framebuffer_unregister_private(fb);
 		drm_framebuffer_cleanup(fb);
 		kfree(fb);
 	}
@@ -339,6 +340,7 @@ static int radeon_fbdev_destroy(struct drm_device *dev, struct radeon_fbdev *rfb
 		rfb->obj = NULL;
 	}
 	drm_fb_helper_fini(&rfbdev->helper);
+	drm_framebuffer_unregister_private(&rfb->base);
 	drm_framebuffer_cleanup(&rfb->base);
 
 	return 0;
diff --git a/drivers/gpu/drm/udl/udl_fb.c b/drivers/gpu/drm/udl/udl_fb.c
index f8904b4..caa84f1 100644
--- a/drivers/gpu/drm/udl/udl_fb.c
+++ b/drivers/gpu/drm/udl/udl_fb.c
@@ -555,6 +555,7 @@ static void udl_fbdev_destroy(struct drm_device *dev,
 		framebuffer_release(info);
 	}
 	drm_fb_helper_fini(&ufbdev->helper);
+	drm_framebuffer_unregister_private(&ufbdev->ufb.base);
 	drm_framebuffer_cleanup(&ufbdev->ufb.base);
 	drm_gem_object_unreference_unlocked(&ufbdev->ufb.obj->base);
 }
diff --git a/drivers/staging/omapdrm/omap_fbdev.c b/drivers/staging/omapdrm/omap_fbdev.c
index 8a027bb..2728e37 100644
--- a/drivers/staging/omapdrm/omap_fbdev.c
+++ b/drivers/staging/omapdrm/omap_fbdev.c
@@ -275,8 +275,10 @@ fail:
 	if (ret) {
 		if (fbi)
 			framebuffer_release(fbi);
-		if (fb)
+		if (fb) {
+			drm_framebuffer_unregister_private(fb);
 			drm_framebuffer_remove(fb);
+		}
 	}
 
 	return ret;
@@ -400,8 +402,10 @@ void omap_fbdev_free(struct drm_device *dev)
 	fbdev = to_omap_fbdev(priv->fbdev);
 
 	/* this will free the backing object */
-	if (fbdev->fb)
+	if (fbdev->fb) {
+		drm_framebuffer_unregister_private(fbdev->fb);
 		drm_framebuffer_remove(fbdev->fb);
+	}
 
 	kfree(fbdev);
 
diff --git a/include/drm/drm_crtc.h b/include/drm/drm_crtc.h
index 7dc1b31..66b2732 100644
--- a/include/drm/drm_crtc.h
+++ b/include/drm/drm_crtc.h
@@ -964,6 +964,7 @@ extern void drm_framebuffer_unreference(struct drm_framebuffer *fb);
 extern void drm_framebuffer_reference(struct drm_framebuffer *fb);
 extern void drm_framebuffer_remove(struct drm_framebuffer *fb);
 extern void drm_framebuffer_cleanup(struct drm_framebuffer *fb);
+extern void drm_framebuffer_unregister_private(struct drm_framebuffer *fb);
 extern int drmfb_probe(struct drm_device *dev, struct drm_crtc *crtc);
 extern int drmfb_remove(struct drm_device *dev, struct drm_framebuffer *fb);
 extern void drm_crtc_probe_connector_modes(struct drm_device *dev, int maxX, int maxY);
-- 
cgit v0.10.2


From 2b677e8c08eed11e4ebe66a7c334f03e389a19a3 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Mon, 10 Dec 2012 21:16:05 +0100
Subject: drm: reference framebuffers which are on the idr

Since otherwise looking and reference-counting around
drm_framebuffer_lookup will be an unmanageable mess. With this change,
an object can either be found in the idr and will stay around once we
incremented the reference counter. Or it will be gone for good and
can't be looked up using its id any more.

Atomicity is guaranteed by the dev->mode_config.fb_lock. The
newly-introduce fpriv->fbs_lock looks a bit redundant, but the next
patch will shuffle the locking order between these two locks and all
the modeset locks taken in modeset_lock_all, so we'll need it.

Also, since userspace could do really funky stuff and race e.g. a
getresources with an rmfb, we need to make sure that the kernel
doesn't fall over trying to look-up an inexistent fb, or causing
confusion by having two fbs around with the same id. Simply reset the
framebuffer id to 0, which marks it as reaped. Any lookups of that id
will fail, so the object is really gone for good from userspace's pov.

Note that we still need to protect the "remove framebuffer from all
use-cases" and the final unreference with the modeset-lock, since most
framebuffer use-sites don't implement proper reference counting yet.
We can only lift this once _all_ users are converted.

With this change, two references are held on alife, but unused
framebuffers:
- The reference for the idr lookup, created in this patch.
- For user-created framebuffers the fpriv->fbs reference, for
  driver-private fbs the driver is supposed to hold it's own last
  reference.

Note that the dev->mode_config.fb_list itself does _not_ hold a
reference onto the framebuffers (this list is essentially only used
for debugfs files). Hence if there's anything left there when the
driver has cleaned up all it's modeset resources, this is a ref-leak.
WARN about it.

Now we only need to fix up all other places to properly reference
count framebuffers.

v2: Fix spelling fail in a comment spotted by Rob Clark.

Reviewed-by: Rob Clark <rob@ti.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index 3eddfab..a50f755 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -356,6 +356,9 @@ int drm_framebuffer_init(struct drm_device *dev, struct drm_framebuffer *fb,
 	if (ret)
 		goto out;
 
+	/* Grab the idr reference. */
+	drm_framebuffer_reference(fb);
+
 	dev->mode_config.num_fb++;
 	list_add(&fb->head, &dev->mode_config.fb_list);
 out:
@@ -372,6 +375,23 @@ static void drm_framebuffer_free(struct kref *kref)
 	fb->funcs->destroy(fb);
 }
 
+static struct drm_framebuffer *__drm_framebuffer_lookup(struct drm_device *dev,
+							uint32_t id)
+{
+	struct drm_mode_object *obj = NULL;
+	struct drm_framebuffer *fb;
+
+	mutex_lock(&dev->mode_config.idr_mutex);
+	obj = idr_find(&dev->mode_config.crtc_idr, id);
+	if (!obj || (obj->type != DRM_MODE_OBJECT_FB) || (obj->id != id))
+		fb = NULL;
+	else
+		fb = obj_to_fb(obj);
+	mutex_unlock(&dev->mode_config.idr_mutex);
+
+	return fb;
+}
+
 /**
  * drm_framebuffer_lookup - look up a drm framebuffer and grab a reference
  * @dev: drm device
@@ -384,22 +404,12 @@ static void drm_framebuffer_free(struct kref *kref)
 struct drm_framebuffer *drm_framebuffer_lookup(struct drm_device *dev,
 					       uint32_t id)
 {
-	struct drm_mode_object *obj = NULL;
 	struct drm_framebuffer *fb;
 
 	mutex_lock(&dev->mode_config.fb_lock);
-
-	mutex_lock(&dev->mode_config.idr_mutex);
-	obj = idr_find(&dev->mode_config.crtc_idr, id);
-	if (!obj || (obj->type != DRM_MODE_OBJECT_FB) || (obj->id != id))
-		fb = NULL;
-	else
-		fb = obj_to_fb(obj);
-	mutex_unlock(&dev->mode_config.idr_mutex);
-
+	fb = __drm_framebuffer_lookup(dev, id);
 	if (fb)
 		kref_get(&fb->refcount);
-
 	mutex_unlock(&dev->mode_config.fb_lock);
 
 	return fb;
@@ -430,6 +440,24 @@ void drm_framebuffer_reference(struct drm_framebuffer *fb)
 }
 EXPORT_SYMBOL(drm_framebuffer_reference);
 
+static void drm_framebuffer_free_bug(struct kref *kref)
+{
+	BUG();
+}
+
+/* dev->mode_config.fb_lock must be held! */
+static void __drm_framebuffer_unregister(struct drm_device *dev,
+					 struct drm_framebuffer *fb)
+{
+	mutex_lock(&dev->mode_config.idr_mutex);
+	idr_remove(&dev->mode_config.crtc_idr, fb->base.id);
+	mutex_unlock(&dev->mode_config.idr_mutex);
+
+	fb->base.id = 0;
+
+	kref_put(&fb->refcount, drm_framebuffer_free_bug);
+}
+
 /**
  * drm_framebuffer_unregister_private - unregister a private fb from the lookup idr
  * @fb: fb to unregister
@@ -441,6 +469,12 @@ EXPORT_SYMBOL(drm_framebuffer_reference);
  */
 void drm_framebuffer_unregister_private(struct drm_framebuffer *fb)
 {
+	struct drm_device *dev = fb->dev;
+
+	mutex_lock(&dev->mode_config.fb_lock);
+	/* Mark fb as reaped and drop idr ref. */
+	__drm_framebuffer_unregister(dev, fb);
+	mutex_unlock(&dev->mode_config.fb_lock);
 }
 EXPORT_SYMBOL(drm_framebuffer_unregister_private);
 
@@ -464,14 +498,6 @@ void drm_framebuffer_cleanup(struct drm_framebuffer *fb)
 {
 	struct drm_device *dev = fb->dev;
 
-	/*
-	 * This could be moved to drm_framebuffer_remove(), but for
-	 * debugging is nice to keep around the list of fb's that are
-	 * no longer associated w/ a drm_file but are not unreferenced
-	 * yet.  (i915 and omapdrm have debugfs files which will show
-	 * this.)
-	 */
-	drm_mode_object_put(dev, &fb->base);
 	mutex_lock(&dev->mode_config.fb_lock);
 	list_del(&fb->head);
 	dev->mode_config.num_fb--;
@@ -1181,9 +1207,15 @@ void drm_mode_config_cleanup(struct drm_device *dev)
 		drm_property_destroy(dev, property);
 	}
 
-	/* Single-threaded teardown context, so it's not requied to grab the
+	/*
+	 * Single-threaded teardown context, so it's not required to grab the
 	 * fb_lock to protect against concurrent fb_list access. Contrary, it
-	 * would actually deadlock with the drm_framebuffer_cleanup function. */
+	 * would actually deadlock with the drm_framebuffer_cleanup function.
+	 *
+	 * Also, if there are any framebuffers left, that's a driver leak now,
+	 * so politely WARN about this.
+	 */
+	WARN_ON(!list_empty(&dev->mode_config.fb_list));
 	list_for_each_entry_safe(fb, fbt, &dev->mode_config.fb_list, head) {
 		drm_framebuffer_remove(fb);
 	}
@@ -2464,39 +2496,41 @@ int drm_mode_rmfb(struct drm_device *dev,
 	struct drm_framebuffer *fb = NULL;
 	struct drm_framebuffer *fbl = NULL;
 	uint32_t *id = data;
-	int ret = 0;
 	int found = 0;
 
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	drm_modeset_lock_all(dev);
-	fb = drm_framebuffer_lookup(dev, *id);
-	if (!fb) {
-		ret = -EINVAL;
-		goto out;
-	}
-	/* fb is protect by the mode_config lock, so drop the ref immediately */
-	drm_framebuffer_unreference(fb);
-
 	mutex_lock(&file_priv->fbs_lock);
+	mutex_lock(&dev->mode_config.fb_lock);
+	fb = __drm_framebuffer_lookup(dev, *id);
+	if (!fb)
+		goto fail_lookup;
+
 	list_for_each_entry(fbl, &file_priv->fbs, filp_head)
 		if (fb == fbl)
 			found = 1;
-	if (!found) {
-		ret = -EINVAL;
-		mutex_unlock(&file_priv->fbs_lock);
-		goto out;
-	}
+	if (!found)
+		goto fail_lookup;
+
+	/* Mark fb as reaped, we still have a ref from fpriv->fbs. */
+	__drm_framebuffer_unregister(dev, fb);
 
 	list_del_init(&fb->filp_head);
+	mutex_unlock(&dev->mode_config.fb_lock);
 	mutex_unlock(&file_priv->fbs_lock);
 
+	drm_modeset_lock_all(dev);
 	drm_framebuffer_remove(fb);
-out:
 	drm_modeset_unlock_all(dev);
 
-	return ret;
+	return 0;
+
+fail_lookup:
+	mutex_unlock(&dev->mode_config.fb_lock);
+	mutex_unlock(&file_priv->fbs_lock);
+
+	return -EINVAL;
 }
 
 /**
@@ -2639,7 +2673,15 @@ void drm_fb_release(struct drm_file *priv)
 	drm_modeset_lock_all(dev);
 	mutex_lock(&priv->fbs_lock);
 	list_for_each_entry_safe(fb, tfb, &priv->fbs, filp_head) {
+
+		mutex_lock(&dev->mode_config.fb_lock);
+		/* Mark fb as reaped, we still have a ref from fpriv->fbs. */
+		__drm_framebuffer_unregister(dev, fb);
+		mutex_unlock(&dev->mode_config.fb_lock);
+
 		list_del_init(&fb->filp_head);
+
+		/* This will also drop the fpriv->fbs reference. */
 		drm_framebuffer_remove(fb);
 	}
 	mutex_unlock(&priv->fbs_lock);
-- 
cgit v0.10.2


From 7d331595b05d6f3c38567ad7031b75075557ce2a Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Mon, 10 Dec 2012 23:44:22 +0100
Subject: drm: nest modeset locks within fpriv->fbs_lock

Atm we still need to unconditionally take the modeset locks in the
rmfb paths. But eventually we only want to take them if there are
other users around as a slow-path. This way sane userspace avoids
blocking on edid reads and other stuff in rmfb if it ensures that the
fb isn't used anywhere by a crtc/plane.

We can do a quick check for such other users once framebuffers are
properly refcounting by locking at the refcount - if it's more than 1,
there are other users left. Again, rmfb racing against other ioctls
isn't a real problem, userspace is allowed to shoot its foot.

This patch just prepares this by moving the modeset locks to nest
within fpriv->fbs_lock. Now the distinction between the fbs_lock and
the device-global fb_lock is clear, since we need to hold the fbs_lock
outside of any modeset_locks in fb_release.

Reviewed-by: Rob Clark <rob@ti.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index a50f755..09e02a7 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -2282,13 +2282,13 @@ int drm_mode_addfb(struct drm_device *dev,
 		drm_modeset_unlock_all(dev);
 		return PTR_ERR(fb);
 	}
+	drm_modeset_unlock_all(dev);
 
 	mutex_lock(&file_priv->fbs_lock);
 	or->fb_id = fb->base.id;
 	list_add(&fb->filp_head, &file_priv->fbs);
 	DRM_DEBUG_KMS("[FB:%d]\n", fb->base.id);
 	mutex_unlock(&file_priv->fbs_lock);
-	drm_modeset_unlock_all(dev);
 
 	return ret;
 }
@@ -2465,6 +2465,7 @@ int drm_mode_addfb2(struct drm_device *dev,
 		drm_modeset_unlock_all(dev);
 		return PTR_ERR(fb);
 	}
+	drm_modeset_unlock_all(dev);
 
 	mutex_lock(&file_priv->fbs_lock);
 	r->fb_id = fb->base.id;
@@ -2472,7 +2473,6 @@ int drm_mode_addfb2(struct drm_device *dev,
 	DRM_DEBUG_KMS("[FB:%d]\n", fb->base.id);
 	mutex_unlock(&file_priv->fbs_lock);
 
-	drm_modeset_unlock_all(dev);
 
 	return ret;
 }
@@ -2670,7 +2670,6 @@ void drm_fb_release(struct drm_file *priv)
 	struct drm_device *dev = priv->minor->dev;
 	struct drm_framebuffer *fb, *tfb;
 
-	drm_modeset_lock_all(dev);
 	mutex_lock(&priv->fbs_lock);
 	list_for_each_entry_safe(fb, tfb, &priv->fbs, filp_head) {
 
@@ -2682,10 +2681,11 @@ void drm_fb_release(struct drm_file *priv)
 		list_del_init(&fb->filp_head);
 
 		/* This will also drop the fpriv->fbs reference. */
+		drm_modeset_lock_all(dev);
 		drm_framebuffer_remove(fb);
+		drm_modeset_unlock_all(dev);
 	}
 	mutex_unlock(&priv->fbs_lock);
-	drm_modeset_unlock_all(dev);
 }
 
 /**
-- 
cgit v0.10.2


From 468174f748603497e73dba9b5c6d1d9f71121486 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Tue, 11 Dec 2012 00:09:12 +0100
Subject: drm: push modeset_lock_all into ->fb_create driver callbacks

And drop it where it's not needed. Most driver just lookup the gem
object, allocate an fb struct, fill in all the useful fields and then
register it with drm_framebuffer_init.

All of these operations are already separately locked, and since we
only put the fb into the fpriv->fbs list _after_ having called
->fb_create, we can't also race with rmfb. We can otoh race with other
ioctls that put the framebuffer to use, but all drivers have been
reorganized already to call drm_framebuffer_init last in the fb
creation sequence.

So essentially, we can completely remove any modeset locks from the
addfb ioctl paths. Yeah!

Also, reference-counting is solid - we get a reference from fb_create
which we transfer to the fpriv->fbs list. And after unlocking the
fpriv->fbs_lock we don't touch the framebuffer any longer. Furthermore
drm_framebuffer_init has added a 2nd reference for the idr lookup, and
any access through that table will do it's own refcounting.

Reviewed-by: Rob Clark <rob@ti.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index 09e02a7..2e6103c 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -2271,18 +2271,12 @@ int drm_mode_addfb(struct drm_device *dev,
 	if ((config->min_height > r.height) || (r.height > config->max_height))
 		return -EINVAL;
 
-	drm_modeset_lock_all(dev);
-
-	/* TODO check buffer is sufficiently large */
-	/* TODO setup destructor callback */
-
 	fb = dev->mode_config.funcs->fb_create(dev, file_priv, &r);
 	if (IS_ERR(fb)) {
 		DRM_DEBUG_KMS("could not create framebuffer\n");
 		drm_modeset_unlock_all(dev);
 		return PTR_ERR(fb);
 	}
-	drm_modeset_unlock_all(dev);
 
 	mutex_lock(&file_priv->fbs_lock);
 	or->fb_id = fb->base.id;
@@ -2457,15 +2451,12 @@ int drm_mode_addfb2(struct drm_device *dev,
 	if (ret)
 		return ret;
 
-	drm_modeset_lock_all(dev);
-
 	fb = dev->mode_config.funcs->fb_create(dev, file_priv, r);
 	if (IS_ERR(fb)) {
 		DRM_DEBUG_KMS("could not create framebuffer\n");
 		drm_modeset_unlock_all(dev);
 		return PTR_ERR(fb);
 	}
-	drm_modeset_unlock_all(dev);
 
 	mutex_lock(&file_priv->fbs_lock);
 	r->fb_id = fb->base.id;
-- 
cgit v0.10.2


From 58c0dca10614117cf4b385e3314e79e3b37fa66b Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 13 Dec 2012 23:06:08 +0100
Subject: drm: don't take modeset locks in getfb ioctl

We only need to push the fb unreference a bit down. While at it,
properly pass the return value from ->create_handle back to userspace.

Most drivers either return -ENODEV if they don't have a concept of
buffer objects (ast, cirrus, ...) or just install a handle for the
underlying gem object (which is ok since we hold a reference on that
through the framebuffer).

v2: Split out the ->create_handle rework in the individual drivers.

Reviewed-by: Rob Clark <rob@ti.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index 2e6103c..cba8c8b 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -2542,19 +2542,14 @@ int drm_mode_getfb(struct drm_device *dev,
 {
 	struct drm_mode_fb_cmd *r = data;
 	struct drm_framebuffer *fb;
-	int ret = 0;
+	int ret;
 
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	drm_modeset_lock_all(dev);
 	fb = drm_framebuffer_lookup(dev, r->fb_id);
-	if (!fb) {
-		ret = -EINVAL;
-		goto out;
-	}
-	/* fb is protect by the mode_config lock, so drop the ref immediately */
-	drm_framebuffer_unreference(fb);
+	if (!fb)
+		return -EINVAL;
 
 	r->height = fb->height;
 	r->width = fb->width;
@@ -2566,8 +2561,8 @@ int drm_mode_getfb(struct drm_device *dev,
 	else
 		ret = -ENODEV;
 
-out:
-	drm_modeset_unlock_all(dev);
+	drm_framebuffer_unreference(fb);
+
 	return ret;
 }
 
-- 
cgit v0.10.2


From 4ccf097f1935321f03ad36218588a9e446006b6a Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Tue, 11 Dec 2012 00:38:18 +0100
Subject: drm: fb refcounting for dirtyfb_ioctl

We only need to ensure that the fb stays around for long enough. While
at it, only grab the modeset locks when we need them (since most
drivers don't implement the dirty callback, this should help jitter
and stalls when using the generic modeset driver).

Reviewed-by: Rob Clark <rob@ti.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index cba8c8b..d4f8fa5 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -2580,14 +2580,9 @@ int drm_mode_dirtyfb_ioctl(struct drm_device *dev,
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	drm_modeset_lock_all(dev);
 	fb = drm_framebuffer_lookup(dev, r->fb_id);
-	if (!fb) {
-		ret = -EINVAL;
-		goto out_err1;
-	}
-	/* fb is protect by the mode_config lock, so drop the ref immediately */
-	drm_framebuffer_unreference(fb);
+	if (!fb)
+		return -EINVAL;
 
 	num_clips = r->num_clips;
 	clips_ptr = (struct drm_clip_rect __user *)(unsigned long)r->clips_ptr;
@@ -2625,17 +2620,19 @@ int drm_mode_dirtyfb_ioctl(struct drm_device *dev,
 	}
 
 	if (fb->funcs->dirty) {
+		drm_modeset_lock_all(dev);
 		ret = fb->funcs->dirty(fb, file_priv, flags, r->color,
 				       clips, num_clips);
+		drm_modeset_unlock_all(dev);
 	} else {
 		ret = -ENOSYS;
-		goto out_err2;
 	}
 
 out_err2:
 	kfree(clips);
 out_err1:
-	drm_modeset_unlock_all(dev);
+	drm_framebuffer_unreference(fb);
+
 	return ret;
 }
 
-- 
cgit v0.10.2


From 6c2a75325c800de286166c693e0cd33c3a1c5ec8 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Tue, 11 Dec 2012 00:59:24 +0100
Subject: drm: refcounting for sprite framebuffers

Now plane->fb holds a reference onto it's framebuffer. Nothing too
fancy going on here:
- Extract __drm_framebuffer_unreference to be called when we know
  we're not dropping the last reference, e.g. useful in the fb cleanup
  code.
- Reduce the locked sections in the set_plane ioctl to only protect
  plane->fb/plane->crtc and the driver callback (i.e. hw state).
  Everything either doesn't disappear (crtc, plane) or is refcounted
  (fb), and all the data we check is invariant over the respective
  object's lifetimes.

Reviewed-by: Rob Clark <rob@ti.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index d4f8fa5..64ef120 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -445,6 +445,12 @@ static void drm_framebuffer_free_bug(struct kref *kref)
 	BUG();
 }
 
+static void __drm_framebuffer_unreference(struct drm_framebuffer *fb)
+{
+	DRM_DEBUG("FB ID: %d\n", fb->base.id);
+	kref_put(&fb->refcount, drm_framebuffer_free_bug);
+}
+
 /* dev->mode_config.fb_lock must be held! */
 static void __drm_framebuffer_unregister(struct drm_device *dev,
 					 struct drm_framebuffer *fb)
@@ -455,7 +461,7 @@ static void __drm_framebuffer_unregister(struct drm_device *dev,
 
 	fb->base.id = 0;
 
-	kref_put(&fb->refcount, drm_framebuffer_free_bug);
+	__drm_framebuffer_unreference(fb);
 }
 
 /**
@@ -544,6 +550,7 @@ void drm_framebuffer_remove(struct drm_framebuffer *fb)
 			if (ret)
 				DRM_ERROR("failed to disable plane with busy fb\n");
 			/* disconnect the plane from the fb and crtc: */
+			__drm_framebuffer_unreference(plane->fb);
 			plane->fb = NULL;
 			plane->crtc = NULL;
 		}
@@ -1850,7 +1857,7 @@ int drm_mode_setplane(struct drm_device *dev, void *data,
 	struct drm_mode_object *obj;
 	struct drm_plane *plane;
 	struct drm_crtc *crtc;
-	struct drm_framebuffer *fb;
+	struct drm_framebuffer *fb = NULL, *old_fb = NULL;
 	int ret = 0;
 	unsigned int fb_width, fb_height;
 	int i;
@@ -1858,8 +1865,6 @@ int drm_mode_setplane(struct drm_device *dev, void *data,
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	drm_modeset_lock_all(dev);
-
 	/*
 	 * First, find the plane, crtc, and fb objects.  If not available,
 	 * we don't bother to call the driver.
@@ -1869,16 +1874,18 @@ int drm_mode_setplane(struct drm_device *dev, void *data,
 	if (!obj) {
 		DRM_DEBUG_KMS("Unknown plane ID %d\n",
 			      plane_req->plane_id);
-		ret = -ENOENT;
-		goto out;
+		return -ENOENT;
 	}
 	plane = obj_to_plane(obj);
 
 	/* No fb means shut it down */
 	if (!plane_req->fb_id) {
+		drm_modeset_lock_all(dev);
+		old_fb = plane->fb;
 		plane->funcs->disable_plane(plane);
 		plane->crtc = NULL;
 		plane->fb = NULL;
+		drm_modeset_unlock_all(dev);
 		goto out;
 	}
 
@@ -1899,8 +1906,6 @@ int drm_mode_setplane(struct drm_device *dev, void *data,
 		ret = -ENOENT;
 		goto out;
 	}
-	/* fb is protect by the mode_config lock, so drop the ref immediately */
-	drm_framebuffer_unreference(fb);
 
 	/* Check whether this plane supports the fb pixel format. */
 	for (i = 0; i < plane->format_count; i++)
@@ -1946,18 +1951,25 @@ int drm_mode_setplane(struct drm_device *dev, void *data,
 		goto out;
 	}
 
+	drm_modeset_lock_all(dev);
 	ret = plane->funcs->update_plane(plane, crtc, fb,
 					 plane_req->crtc_x, plane_req->crtc_y,
 					 plane_req->crtc_w, plane_req->crtc_h,
 					 plane_req->src_x, plane_req->src_y,
 					 plane_req->src_w, plane_req->src_h);
 	if (!ret) {
+		old_fb = plane->fb;
+		fb = NULL;
 		plane->crtc = crtc;
 		plane->fb = fb;
 	}
+	drm_modeset_unlock_all(dev);
 
 out:
-	drm_modeset_unlock_all(dev);
+	if (fb)
+		drm_framebuffer_unreference(fb);
+	if (old_fb)
+		drm_framebuffer_unreference(old_fb);
 
 	return ret;
 }
-- 
cgit v0.10.2


From b0d1232589df5575c5971224ac4cb30e7e525884 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Tue, 11 Dec 2012 01:07:12 +0100
Subject: drm: refcounting for crtc framebuffers

With the prep patch to encapsulate ->set_crtc calls, this is now
rather easy. Hooray for inconsistent semantics between ->set_crtc and
->page_flip, where the driver callback is supposed to update the fb
pointer, and ->update_plane, where the drm core does the same.

Also, since the drm core functions check crtc->fb before calling into
driver callbacks, we can't really reduce the critical sections
protected by the mode_config locks.

Reviewed-by: Rob Clark <rob@ti.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index 64ef120..6dc75ee 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -1984,8 +1984,21 @@ out:
 int drm_mode_set_config_internal(struct drm_mode_set *set)
 {
 	struct drm_crtc *crtc = set->crtc;
+	struct drm_framebuffer *fb, *old_fb;
+	int ret;
+
+	old_fb = crtc->fb;
+	fb = set->fb;
 
-	return crtc->funcs->set_config(set);
+	ret = crtc->funcs->set_config(set);
+	if (ret == 0) {
+		if (old_fb)
+			drm_framebuffer_unreference(old_fb);
+		if (fb)
+			drm_framebuffer_reference(fb);
+	}
+
+	return ret;
 }
 EXPORT_SYMBOL(drm_mode_set_config_internal);
 
@@ -2046,6 +2059,8 @@ int drm_mode_setcrtc(struct drm_device *dev, void *data,
 				goto out;
 			}
 			fb = crtc->fb;
+			/* Make refcounting symmetric with the lookup path. */
+			drm_framebuffer_reference(fb);
 		} else {
 			fb = drm_framebuffer_lookup(dev, crtc_req->fb_id);
 			if (!fb) {
@@ -2054,9 +2069,6 @@ int drm_mode_setcrtc(struct drm_device *dev, void *data,
 				ret = -EINVAL;
 				goto out;
 			}
-			/* fb is protect by the mode_config lock, so drop the
-			 * ref immediately */
-			drm_framebuffer_unreference(fb);
 		}
 
 		mode = drm_mode_create(dev);
@@ -2156,6 +2168,9 @@ int drm_mode_setcrtc(struct drm_device *dev, void *data,
 	ret = drm_mode_set_config_internal(&set);
 
 out:
+	if (fb)
+		drm_framebuffer_unreference(fb);
+
 	kfree(connector_set);
 	drm_mode_destroy(dev, mode);
 	drm_modeset_unlock_all(dev);
@@ -3656,7 +3671,7 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,
 	struct drm_mode_crtc_page_flip *page_flip = data;
 	struct drm_mode_object *obj;
 	struct drm_crtc *crtc;
-	struct drm_framebuffer *fb;
+	struct drm_framebuffer *fb = NULL, *old_fb = NULL;
 	struct drm_pending_vblank_event *e = NULL;
 	unsigned long flags;
 	int hdisplay, vdisplay;
@@ -3687,8 +3702,6 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,
 	fb = drm_framebuffer_lookup(dev, page_flip->fb_id);
 	if (!fb)
 		goto out;
-	/* fb is protect by the mode_config lock, so drop the ref immediately */
-	drm_framebuffer_unreference(fb);
 
 	hdisplay = crtc->mode.hdisplay;
 	vdisplay = crtc->mode.vdisplay;
@@ -3734,6 +3747,7 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,
 			(void (*) (struct drm_pending_event *)) kfree;
 	}
 
+	old_fb = crtc->fb;
 	ret = crtc->funcs->page_flip(crtc, fb, e);
 	if (ret) {
 		if (page_flip->flags & DRM_MODE_PAGE_FLIP_EVENT) {
@@ -3742,9 +3756,18 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,
 			spin_unlock_irqrestore(&dev->event_lock, flags);
 			kfree(e);
 		}
+		/* Keep the old fb, don't unref it. */
+		old_fb = NULL;
+	} else {
+		/* Unref only the old framebuffer. */
+		fb = NULL;
 	}
 
 out:
+	if (fb)
+		drm_framebuffer_unreference(fb);
+	if (old_fb)
+		drm_framebuffer_unreference(old_fb);
 	drm_modeset_unlock_all(dev);
 	return ret;
 }
-- 
cgit v0.10.2


From 623f9783027ef0a948205f17792c9e1fcedb61c6 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Tue, 11 Dec 2012 16:21:38 +0100
Subject: drm/i915: dump refcount into framebuffer debugfs file

Useful for checking whether the new refcounting works as advertised.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index a40c674..8a11085 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -1367,11 +1367,12 @@ static int i915_gem_framebuffer_info(struct seq_file *m, void *data)
 	ifbdev = dev_priv->fbdev;
 	fb = to_intel_framebuffer(ifbdev->helper.fb);
 
-	seq_printf(m, "fbcon size: %d x %d, depth %d, %d bpp, obj ",
+	seq_printf(m, "fbcon size: %d x %d, depth %d, %d bpp, refcount %d, obj ",
 		   fb->base.width,
 		   fb->base.height,
 		   fb->base.depth,
-		   fb->base.bits_per_pixel);
+		   fb->base.bits_per_pixel,
+		   atomic_read(&fb->base.refcount.refcount));
 	describe_obj(m, fb->obj);
 	seq_printf(m, "\n");
 	mutex_unlock(&dev->mode_config.mutex);
@@ -1381,11 +1382,12 @@ static int i915_gem_framebuffer_info(struct seq_file *m, void *data)
 		if (&fb->base == ifbdev->helper.fb)
 			continue;
 
-		seq_printf(m, "user size: %d x %d, depth %d, %d bpp, obj ",
+		seq_printf(m, "user size: %d x %d, depth %d, %d bpp, refcount %d, obj ",
 			   fb->base.width,
 			   fb->base.height,
 			   fb->base.depth,
-			   fb->base.bits_per_pixel);
+			   fb->base.bits_per_pixel,
+			   atomic_read(&fb->base.refcount.refcount));
 		describe_obj(m, fb->obj);
 		seq_printf(m, "\n");
 	}
-- 
cgit v0.10.2


From 2fd5eabab02d9cdade04397eae0bfd49f452cdba Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Tue, 11 Dec 2012 16:28:34 +0100
Subject: drm/vmwgfx: add proper framebuffer refcounting

Afact vmwgfx already has all the right refcounting implemented on the
backing storage, and we only need to ensure that the drm fb doesn't
disappear untimely. So holding onto the fb reference from _lookup
until vmw_kms_present has completed should be enough.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c
index 1b8f428..c509d40 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c
@@ -171,8 +171,6 @@ int vmw_present_ioctl(struct drm_device *dev, void *data,
 		ret = -EINVAL;
 		goto out_no_fb;
 	}
-	/* fb is protect by the mode_config lock, so drop the ref immediately */
-	drm_framebuffer_unreference(fb);
 	vfb = vmw_framebuffer_to_vfb(fb);
 
 	ret = ttm_read_lock(&vmaster->lock, true);
@@ -197,6 +195,7 @@ int vmw_present_ioctl(struct drm_device *dev, void *data,
 out_no_surface:
 	ttm_read_unlock(&vmaster->lock);
 out_no_ttm_lock:
+	drm_framebuffer_unreference(fb);
 out_no_fb:
 	drm_modeset_unlock_all(dev);
 out_no_copy:
@@ -256,14 +255,12 @@ int vmw_present_readback_ioctl(struct drm_device *dev, void *data,
 		ret = -EINVAL;
 		goto out_no_fb;
 	}
-	/* fb is protect by the mode_config lock, so drop the ref immediately */
-	drm_framebuffer_unreference(fb);
 
 	vfb = vmw_framebuffer_to_vfb(fb);
 	if (!vfb->dmabuf) {
 		DRM_ERROR("Framebuffer not dmabuf backed.\n");
 		ret = -EINVAL;
-		goto out_no_fb;
+		goto out_no_ttm_lock;
 	}
 
 	ret = ttm_read_lock(&vmaster->lock, true);
@@ -276,6 +273,7 @@ int vmw_present_readback_ioctl(struct drm_device *dev, void *data,
 
 	ttm_read_unlock(&vmaster->lock);
 out_no_ttm_lock:
+	drm_framebuffer_unreference(fb);
 out_no_fb:
 	drm_modeset_unlock_all(dev);
 out_no_copy:
-- 
cgit v0.10.2


From b62584e366ebcb3adffefad373a5abc4c4b677ca Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Tue, 11 Dec 2012 16:51:35 +0100
Subject: drm: optimize drm_framebuffer_remove

Now that all framebuffer usage is properly refcounted, we are no
longer required to hold the modeset locks while dropping the last
reference. Hence implemented a fastpath which avoids the potential
stalls associated with grabbing mode_config.lock for the case where
there's no other reference around.

Explain in a big comment why it is safe. Also update kerneldocs with
the new locking rules around drm_framebuffer_remove.

Reviewed-by: Rob Clark <rob@ti.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index 6dc75ee..9492e7e 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -517,7 +517,11 @@ EXPORT_SYMBOL(drm_framebuffer_cleanup);
  *
  * Scans all the CRTCs and planes in @dev's mode_config.  If they're
  * using @fb, removes it, setting it to NULL. Then drops the reference to the
- * passed-in framebuffer.
+ * passed-in framebuffer. Might take the modeset locks.
+ *
+ * Note that this function optimizes the cleanup away if the caller holds the
+ * last reference to the framebuffer. It is also guaranteed to not take the
+ * modeset locks in this case.
  */
 void drm_framebuffer_remove(struct drm_framebuffer *fb)
 {
@@ -527,33 +531,51 @@ void drm_framebuffer_remove(struct drm_framebuffer *fb)
 	struct drm_mode_set set;
 	int ret;
 
-	WARN_ON(!drm_modeset_is_locked(dev));
 	WARN_ON(!list_empty(&fb->filp_head));
 
-	/* remove from any CRTC */
-	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
-		if (crtc->fb == fb) {
-			/* should turn off the crtc */
-			memset(&set, 0, sizeof(struct drm_mode_set));
-			set.crtc = crtc;
-			set.fb = NULL;
-			ret = drm_mode_set_config_internal(&set);
-			if (ret)
-				DRM_ERROR("failed to reset crtc %p when fb was deleted\n", crtc);
+	/*
+	 * drm ABI mandates that we remove any deleted framebuffers from active
+	 * useage. But since most sane clients only remove framebuffers they no
+	 * longer need, try to optimize this away.
+	 *
+	 * Since we're holding a reference ourselves, observing a refcount of 1
+	 * means that we're the last holder and can skip it. Also, the refcount
+	 * can never increase from 1 again, so we don't need any barriers or
+	 * locks.
+	 *
+	 * Note that userspace could try to race with use and instate a new
+	 * usage _after_ we've cleared all current ones. End result will be an
+	 * in-use fb with fb-id == 0. Userspace is allowed to shoot its own foot
+	 * in this manner.
+	 */
+	if (atomic_read(&fb->refcount.refcount) > 1) {
+		drm_modeset_lock_all(dev);
+		/* remove from any CRTC */
+		list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
+			if (crtc->fb == fb) {
+				/* should turn off the crtc */
+				memset(&set, 0, sizeof(struct drm_mode_set));
+				set.crtc = crtc;
+				set.fb = NULL;
+				ret = drm_mode_set_config_internal(&set);
+				if (ret)
+					DRM_ERROR("failed to reset crtc %p when fb was deleted\n", crtc);
+			}
 		}
-	}
 
-	list_for_each_entry(plane, &dev->mode_config.plane_list, head) {
-		if (plane->fb == fb) {
-			/* should turn off the crtc */
-			ret = plane->funcs->disable_plane(plane);
-			if (ret)
-				DRM_ERROR("failed to disable plane with busy fb\n");
-			/* disconnect the plane from the fb and crtc: */
-			__drm_framebuffer_unreference(plane->fb);
-			plane->fb = NULL;
-			plane->crtc = NULL;
+		list_for_each_entry(plane, &dev->mode_config.plane_list, head) {
+			if (plane->fb == fb) {
+				/* should turn off the crtc */
+				ret = plane->funcs->disable_plane(plane);
+				if (ret)
+					DRM_ERROR("failed to disable plane with busy fb\n");
+				/* disconnect the plane from the fb and crtc: */
+				__drm_framebuffer_unreference(plane->fb);
+				plane->fb = NULL;
+				plane->crtc = NULL;
+			}
 		}
+		drm_modeset_unlock_all(dev);
 	}
 
 	drm_framebuffer_unreference(fb);
@@ -2538,9 +2560,7 @@ int drm_mode_rmfb(struct drm_device *dev,
 	mutex_unlock(&dev->mode_config.fb_lock);
 	mutex_unlock(&file_priv->fbs_lock);
 
-	drm_modeset_lock_all(dev);
 	drm_framebuffer_remove(fb);
-	drm_modeset_unlock_all(dev);
 
 	return 0;
 
@@ -2691,9 +2711,7 @@ void drm_fb_release(struct drm_file *priv)
 		list_del_init(&fb->filp_head);
 
 		/* This will also drop the fpriv->fbs reference. */
-		drm_modeset_lock_all(dev);
 		drm_framebuffer_remove(fb);
-		drm_modeset_unlock_all(dev);
 	}
 	mutex_unlock(&priv->fbs_lock);
 }
-- 
cgit v0.10.2


From b4d5e7d1dbdd6a758a4f7717beef7bd6b007bd66 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Tue, 11 Dec 2012 16:59:31 +0100
Subject: drm: only grab the crtc lock for pageflips

The pagelip ioctl itself is rather simply, so the hard work for this
patch is auditing all the drivers:

- exynos: Pageflip is protect with dev->struct_mutex and ...
  synchronous. But nothing fancy going on, besides a check whether the
  crtc is enabled, which should probably be somewhere in the drm core
  so that we have unified behaviour across all drivers.

- i915: hw-state is protected with dev->struct_mutex, the delayed
  unpin work together with the other stuff the pageflip complete irq
  handler needs is protected by the event_lock spinlock.

- nouveau: With the pin/unpin functions fixed, everything looks safe:
  A bit of ttm wrestling and refcounting, and a few channel accesses.
  The later are either already proteced sufficiently, or are now safe
  with the channel locking introduced to make cursor updates safe.

- radeon: The irq_get/put functions look a bit race, since the
  atomic_inc/dec isn't protect with locks. Otoh they're all per-crtc,
  so we should be safe with per-crtc locking from the drm core. Then
  there's tons of per-crtc register access, which could potentially go
  through the indirect reg acces. But that's fixed to make cursor
  updates concurrent. Bookeeping for the drm even is also protected
  with the even_lock, which also protects against the pageflip irq
  handler since radeon hw seems to have no way to queue these up
  asynchronously. Otherwise just a bit of ttm-based buffer handling
  and fencing, which is now safe with the previous patch to hold
  bdev->fence_lock while grabbing the ttm fence.

- shmob: Only one crtc. That's an easy one ...

- vmwgfx: As usual a bit special with tons different things:
  - Flippable check using is_implicit and num_implicit. Changes to
    those seem to be nicely covered with the global modeset lock, so
    we should be fine.
  - Some dirty cliprect handling stuff, or at least that is my guess.
    Looks like it's fine since either it's per-crtc, invariant or
    (like the execbuf stuff launched) protected otherwise.
  - Adding the actual flip to the fence_event list. On a quick look
    this seems to have solid locking in place, too.
  ... but generally this is all way over my head.

- imx: Impressive display of races between the page_flip
  implementation and the irq handler. Also, ipu_drm_set_base which
  gets eventually called from the irq handler to update the display
  base isn't really protected against concurrent set_config calls from
  process context.  In any case, going for per-crtc locking won't make
  this worse, so nothing to do.

- omap: The new async callback code merged into 3.8 seems to have
  solid locking in place, and there doesn't seem to be any shared
  state at risk. Especially since the callbacks still use
  modeset_lock_all and are so not converted.

v2: Update omapdrm analysis to 3.8 code per the discussion with Rob
Clark.

Reviewed-by: Rob Clark <rob@ti.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index 9492e7e..fd3e9a1 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -3699,12 +3699,12 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,
 	    page_flip->reserved != 0)
 		return -EINVAL;
 
-	drm_modeset_lock_all(dev);
 	obj = drm_mode_object_find(dev, page_flip->crtc_id, DRM_MODE_OBJECT_CRTC);
 	if (!obj)
-		goto out;
+		return -EINVAL;
 	crtc = obj_to_crtc(obj);
 
+	mutex_lock(&crtc->mutex);
 	if (crtc->fb == NULL) {
 		/* The framebuffer is currently unbound, presumably
 		 * due to a hotplug event, that userspace has not
@@ -3786,7 +3786,8 @@ out:
 		drm_framebuffer_unreference(fb);
 	if (old_fb)
 		drm_framebuffer_unreference(old_fb);
-	drm_modeset_unlock_all(dev);
+	mutex_unlock(&crtc->mutex);
+
 	return ret;
 }
 
-- 
cgit v0.10.2


From 7b24056be6db7ce907baffdd4cf142ab774ea60c Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Wed, 12 Dec 2012 00:35:33 +0100
Subject: drm: don't hold crtc mutexes for connector ->detect callbacks

The coup de grace of the entire journey. No more dropped frames every
10s on my testbox!

I've tried to audit all ->detect and ->get_modes callbacks, but things
became a bit fuzzy after trying to piece together the umpteenth
implemenation. Afaict most drivers just have bog-standard output
register frobbing with a notch of i2c edid reading, nothing which
could potentially race with the newly concurrent pageflip/set_cursor
code. The big exception is load-detection code which requires a
running pipe, but radeon/nouveau seem to to this without touching any
state which can be observed from page_flip (e.g. disabled crtcs
temporarily getting enabled and so a pageflip succeeding).

The only special case I could find is the i915 load detect code. That
uses the normal modeset interface to enable the load-detect crtc, and
so userspace could try to squeeze in a pageflip on the load-detect
pipe. So we need to grab the relevant crtc mutex in there, to avoid
the temporary crtc enabling to sneak out and be visible to userspace.

Note that the sysfs files already stopped grabbing the per-crtc locks,
since I didn't want to bother with doing a interruptible
modeset_lock_all. But since there's very little in-between breakage
(essentially just the ability for userspace to pageflip on load-detect
crtcs when it shouldn't on the i915 driver) I figured I don't need to
bother.

Reviewed-by: Rob Clark <rob@ti.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index fd3e9a1..9c797f6 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -1617,7 +1617,7 @@ int drm_mode_getconnector(struct drm_device *dev, void *data,
 
 	DRM_DEBUG_KMS("[CONNECTOR:%d:?]\n", out_resp->connector_id);
 
-	drm_modeset_lock_all(dev);
+	mutex_lock(&dev->mode_config.mutex);
 
 	obj = drm_mode_object_find(dev, out_resp->connector_id,
 				   DRM_MODE_OBJECT_CONNECTOR);
@@ -1714,7 +1714,8 @@ int drm_mode_getconnector(struct drm_device *dev, void *data,
 	out_resp->count_encoders = encoders_count;
 
 out:
-	drm_modeset_unlock_all(dev);
+	mutex_unlock(&dev->mode_config.mutex);
+
 	return ret;
 }
 
diff --git a/drivers/gpu/drm/drm_crtc_helper.c b/drivers/gpu/drm/drm_crtc_helper.c
index 400ef86..7b2d378 100644
--- a/drivers/gpu/drm/drm_crtc_helper.c
+++ b/drivers/gpu/drm/drm_crtc_helper.c
@@ -980,7 +980,7 @@ static void output_poll_execute(struct work_struct *work)
 	if (!drm_kms_helper_poll)
 		return;
 
-	drm_modeset_lock_all(dev);
+	mutex_lock(&dev->mode_config.mutex);
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
 
 		/* Ignore forced connectors. */
@@ -1010,7 +1010,7 @@ static void output_poll_execute(struct work_struct *work)
 			changed = true;
 	}
 
-	drm_modeset_unlock_all(dev);
+	mutex_unlock(&dev->mode_config.mutex);
 
 	if (changed)
 		drm_kms_helper_hotplug_event(dev);
@@ -1070,7 +1070,7 @@ void drm_helper_hpd_irq_event(struct drm_device *dev)
 	if (!dev->mode_config.poll_enabled)
 		return;
 
-	drm_modeset_lock_all(dev);
+	mutex_lock(&dev->mode_config.mutex);
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
 
 		/* Only handle HPD capable connectors. */
@@ -1088,7 +1088,7 @@ void drm_helper_hpd_irq_event(struct drm_device *dev)
 			changed = true;
 	}
 
-	drm_modeset_unlock_all(dev);
+	mutex_unlock(&dev->mode_config.mutex);
 
 	if (changed)
 		drm_kms_helper_hotplug_event(dev);
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index df51203..c8f417b 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -6700,6 +6700,8 @@ bool intel_get_load_detect_pipe(struct drm_connector *connector,
 	if (encoder->crtc) {
 		crtc = encoder->crtc;
 
+		mutex_lock(&crtc->mutex);
+
 		old->dpms_mode = connector->dpms;
 		old->load_detect_temp = false;
 
@@ -6729,6 +6731,7 @@ bool intel_get_load_detect_pipe(struct drm_connector *connector,
 		return false;
 	}
 
+	mutex_lock(&crtc->mutex);
 	intel_encoder->new_crtc = to_intel_crtc(crtc);
 	to_intel_connector(connector)->new_encoder = intel_encoder;
 
@@ -6756,6 +6759,7 @@ bool intel_get_load_detect_pipe(struct drm_connector *connector,
 		DRM_DEBUG_KMS("reusing fbdev for load-detection framebuffer\n");
 	if (IS_ERR(fb)) {
 		DRM_DEBUG_KMS("failed to allocate framebuffer for load-detection\n");
+		mutex_unlock(&crtc->mutex);
 		return false;
 	}
 
@@ -6763,6 +6767,7 @@ bool intel_get_load_detect_pipe(struct drm_connector *connector,
 		DRM_DEBUG_KMS("failed to set mode on load-detect pipe\n");
 		if (old->release_fb)
 			old->release_fb->funcs->destroy(old->release_fb);
+		mutex_unlock(&crtc->mutex);
 		return false;
 	}
 
@@ -6777,14 +6782,13 @@ void intel_release_load_detect_pipe(struct drm_connector *connector,
 	struct intel_encoder *intel_encoder =
 		intel_attached_encoder(connector);
 	struct drm_encoder *encoder = &intel_encoder->base;
+	struct drm_crtc *crtc = encoder->crtc;
 
 	DRM_DEBUG_KMS("[CONNECTOR:%d:%s], [ENCODER:%d:%s]\n",
 		      connector->base.id, drm_get_connector_name(connector),
 		      encoder->base.id, drm_get_encoder_name(encoder));
 
 	if (old->load_detect_temp) {
-		struct drm_crtc *crtc = encoder->crtc;
-
 		to_intel_connector(connector)->new_encoder = NULL;
 		intel_encoder->new_crtc = NULL;
 		intel_set_mode(crtc, NULL, 0, 0, NULL);
@@ -6800,6 +6804,8 @@ void intel_release_load_detect_pipe(struct drm_connector *connector,
 	/* Switch crtc and encoder back off if necessary */
 	if (old->dpms_mode != DRM_MODE_DPMS_ON)
 		connector->funcs->dpms(connector, old->dpms_mode);
+
+	mutex_unlock(&crtc->mutex);
 }
 
 /* Returns the clock of the currently programmed mode of the given pipe. */
-- 
cgit v0.10.2


From 5d7a951537927555fa1286a338e1b91c3b8b7445 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Fri, 4 Jan 2013 22:31:20 +0100
Subject: drm/doc: updates for new framebuffer lifetime rules

Now that framebuffer are reference-counted for all use-sites, update
the documentation accordingly to stress the new rules for
initialization and teardown.

Also add a short paragraph about the implications for drivers of the
new locking rules.

Reviewed-by: Rob Clark <rob@ti.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl
index caab791..6c11d77 100644
--- a/Documentation/DocBook/drm.tmpl
+++ b/Documentation/DocBook/drm.tmpl
@@ -978,10 +978,25 @@ int max_width, max_height;</synopsis>
         If the parameters are deemed valid, drivers then create, initialize and
         return an instance of struct <structname>drm_framebuffer</structname>.
         If desired the instance can be embedded in a larger driver-specific
-        structure. The new instance is initialized with a call to
-        <function>drm_framebuffer_init</function> which takes a pointer to DRM
-        frame buffer operations (struct
-        <structname>drm_framebuffer_funcs</structname>). Frame buffer operations are
+	structure. Drivers must fill its <structfield>width</structfield>,
+	<structfield>height</structfield>, <structfield>pitches</structfield>,
+        <structfield>offsets</structfield>, <structfield>depth</structfield>,
+        <structfield>bits_per_pixel</structfield> and
+        <structfield>pixel_format</structfield> fields from the values passed
+        through the <parameter>drm_mode_fb_cmd2</parameter> argument. They
+        should call the <function>drm_helper_mode_fill_fb_struct</function>
+        helper function to do so.
+      </para>
+
+      <para>
+	The initailization of the new framebuffer instance is finalized with a
+	call to <function>drm_framebuffer_init</function> which takes a pointer
+	to DRM frame buffer operations (struct
+	<structname>drm_framebuffer_funcs</structname>). Note that this function
+	publishes the framebuffer and so from this point on it can be accessed
+	concurrently from other threads. Hence it must be the last step in the
+	driver's framebuffer initialization sequence. Frame buffer operations
+	are
         <itemizedlist>
           <listitem>
             <synopsis>int (*create_handle)(struct drm_framebuffer *fb,
@@ -1022,16 +1037,16 @@ int max_width, max_height;</synopsis>
         </itemizedlist>
       </para>
       <para>
-        After initializing the <structname>drm_framebuffer</structname>
-        instance drivers must fill its <structfield>width</structfield>,
-        <structfield>height</structfield>, <structfield>pitches</structfield>,
-        <structfield>offsets</structfield>, <structfield>depth</structfield>,
-        <structfield>bits_per_pixel</structfield> and
-        <structfield>pixel_format</structfield> fields from the values passed
-        through the <parameter>drm_mode_fb_cmd2</parameter> argument. They
-        should call the <function>drm_helper_mode_fill_fb_struct</function>
-        helper function to do so.
-      </para>
+	The lifetime of a drm framebuffer is controlled with a reference count,
+	drivers can grab additional references with
+	<function>drm_framebuffer_reference</function> </para> and drop them
+	again with <function>drm_framebuffer_unreference</function>. For
+	driver-private framebuffers for which the last reference is never
+	dropped (e.g. for the fbdev framebuffer when the struct
+	<structname>drm_framebuffer</structname> is embedded into the fbdev
+	helper struct) drivers can manually clean up a framebuffer at module
+	unload time with
+	<function>drm_framebuffer_unregister_private</function>.
     </sect2>
     <sect2>
       <title>Output Polling</title>
@@ -1043,6 +1058,22 @@ int max_width, max_height;</synopsis>
         operation.
       </para>
     </sect2>
+    <sect2>
+      <title>Locking</title>
+      <para>
+        Beside some lookup structures with their own locking (which is hidden
+	behind the interface functions) most of the modeset state is protected
+	by the <code>dev-&lt;mode_config.lock</code> mutex and additionally
+	per-crtc locks to allow cursor updates, pageflips and similar operations
+	to occur concurrently with background tasks like output detection.
+	Operations which cross domains like a full modeset always grab all
+	locks. Drivers there need to protect resources shared between crtcs with
+	additional locking. They also need to be careful to always grab the
+	relevant crtc locks if a modset functions touches crtc state, e.g. for
+	load detection (which does only grab the <code>mode_config.lock</code>
+	to allow concurrent screen updates on live crtcs).
+      </para>
+    </sect2>
   </sect1>
 
   <!-- Internals: kms initialization and cleanup -->
-- 
cgit v0.10.2


From 20c60c35de3285222b3476c3445c66bedf0c449c Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Mon, 17 Dec 2012 12:13:23 +0100
Subject: drm/fb_helper: check whether fbcon is bound

We need to make sure that the fbcon is still bound when touching the
hw, since otherwise we might corrupt the modeset state of kms clients.
X mostly works around that with VT switching and setting the VT into
raw mode, which disables most fbcon events.

Raw kms test programs though don't do that dance, and in the future
we might want to aim to abolish CONFIG_VT anyway. So improve preventive
measures a bit. To do so, extract the existing logic for handling hotplug
events (which X can't block with the current set of tricks) and reuse
it for the fbdev blanking helper.

Long-term we really need to either scrap this all and only have a OOPS
console, or come up with a saner model for device ownership sharing
between fbdev/fbcon and kms userspace.

Reviewed-by: Rob Clark <rob@ti.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index be0f2d6..0c6e25e 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -305,6 +305,24 @@ void drm_fb_helper_restore(void)
 }
 EXPORT_SYMBOL(drm_fb_helper_restore);
 
+static bool drm_fb_helper_is_bound(struct drm_fb_helper *fb_helper)
+{
+	struct drm_device *dev = fb_helper->dev;
+	struct drm_crtc *crtc;
+	int bound = 0, crtcs_bound = 0;
+
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
+		if (crtc->fb)
+			crtcs_bound++;
+		if (crtc->fb == fb_helper->fb)
+			bound++;
+	}
+
+	if (bound < crtcs_bound)
+		return false;
+	return true;
+}
+
 #ifdef CONFIG_MAGIC_SYSRQ
 static void drm_fb_helper_restore_work_fn(struct work_struct *ignored)
 {
@@ -338,6 +356,11 @@ static void drm_fb_helper_dpms(struct fb_info *info, int dpms_mode)
 	 * For each CRTC in this fb, turn the connectors on/off.
 	 */
 	drm_modeset_lock_all(dev);
+	if (!drm_fb_helper_is_bound(fb_helper)) {
+		drm_modeset_unlock_all(dev);
+		return;
+	}
+
 	for (i = 0; i < fb_helper->crtc_count; i++) {
 		crtc = fb_helper->crtc_info[i].mode_set.crtc;
 
@@ -702,6 +725,11 @@ int drm_fb_helper_pan_display(struct fb_var_screeninfo *var,
 	int i;
 
 	drm_modeset_lock_all(dev);
+	if (!drm_fb_helper_is_bound(fb_helper)) {
+		drm_modeset_unlock_all(dev);
+		return -EBUSY;
+	}
+
 	for (i = 0; i < fb_helper->crtc_count; i++) {
 		crtc = fb_helper->crtc_info[i].mode_set.crtc;
 
@@ -1369,21 +1397,12 @@ int drm_fb_helper_hotplug_event(struct drm_fb_helper *fb_helper)
 	struct drm_device *dev = fb_helper->dev;
 	int count = 0;
 	u32 max_width, max_height, bpp_sel;
-	int bound = 0, crtcs_bound = 0;
-	struct drm_crtc *crtc;
 
 	if (!fb_helper->fb)
 		return 0;
 
 	drm_modeset_lock_all(dev);
-	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
-		if (crtc->fb)
-			crtcs_bound++;
-		if (crtc->fb == fb_helper->fb)
-			bound++;
-	}
-
-	if (bound < crtcs_bound) {
+	if (!drm_fb_helper_is_bound(fb_helper)) {
 		fb_helper->delayed_hotplug = true;
 		drm_modeset_unlock_all(dev);
 		return 0;
-- 
cgit v0.10.2


From f69061bedd6ea63f271fe97914364def2f33fc6b Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 6 Dec 2012 09:01:42 +0100
Subject: drm/i915: create a race-free reset detection

With the previous patch the state transition handling of the reset
code itself is now (hopefully) race free and solid. But that still
leaves out everyone else - with the various lock-free wait paths
we have there's the possibility that the reset happens between the
point where we read the seqno we should wait on and the actual wait.

And if __wait_seqno then never sees the RESET_IN_PROGRESS state, we'll
happily wait for a seqno which will in all likelyhood never signal.

In practice this is not a big problem since the X server gets
constantly interrupted, and can then submit more work (hopefully) to
unblock everyone else: As soon as a new seqno write lands, all waiters
will unblock. But running the i-g-t reset testcase ZZ_hangman can
expose this race, especially on slower hw with fewer cpu cores.

Now looking forward to ARB_robustness and friends that's not the best
possible behaviour, hence this patch adds a reset_counter to be able
to detect any reset, even if a given thread never observed the
in-progress state.

The important part is to correctly order things:
- The write side needs to increment the counter after any seqno gets
  reset.  Hence we need to do that at the end of the reset work, and
  again wake everyone up. We also need to place a barrier in between
  any possible seqno changes and the counter increment, since any
  unlock operations only guarantee that nothing leaks out, but not
  that at later load operation gets moved ahead.
- On the read side we need to ensure that no reset can sneak in and
  invalidate the seqno. In all cases we can use the one-sided barrier
  that unlock operations guarantee (of the lock protecting the
  respective seqno/ring pair) to ensure correct ordering. Hence it is
  sufficient to place the atomic read before the mutex/spin_unlock and
  no additional barriers are required.

The end-result of all this is that we need to wake up everyone twice
in a reset operation:
- First, before the reset starts, to get any lockholders of the locks,
  so that the reset can proceed.
- Second, after the reset is completed, to allow waiters to properly
  and reliably detect the reset condition and bail out.

I admit that this entire reset_counter thing smells a bit like
overkill, but I think it's justified since it makes it really explicit
what the bail-out condition is. And we need a reset counter anyway to
implement ARB_robustness, and imo with finer-grained locking on the
horizont this is the most resilient scheme I could think of.

v2: Drop spurious change in the wait_for_error EXIT_COND - we only
need to wait until we leave the reset-in-progress wedged state.

v3: Don't play tricks with barriers in the throttle ioctl, the
spin_unlock is barrier enough.

I've also considered using a little helper to grab the current
reset_counter, but then decided that hiding the atomic_read isn't a
great idea, since having it explicitly show up in the code is a nice
remainder to reviews to check the memory barriers.

v4: Add a comment to explain why we need to fall through in
__wait_seqno in the end variable assignments.

v5: Review from Damien:
- s/smb/smp/ in a comment
- don't increment the reset counter after we've set it to WEDGED. Now
  we (again) properly wedge the gpu when the reset fails.

Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index c84743b..56ece50 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -775,9 +775,16 @@ struct i915_gpu_error {
 	unsigned long last_reset;
 
 	/**
-	 * State variable controlling the reset flow
+	 * State variable and reset counter controlling the reset flow
 	 *
-	 * Upper bits are for the reset counter.
+	 * Upper bits are for the reset counter.  This counter is used by the
+	 * wait_seqno code to race-free noticed that a reset event happened and
+	 * that it needs to restart the entire ioctl (since most likely the
+	 * seqno it waited for won't ever signal anytime soon).
+	 *
+	 * This is important for lock-free wait paths, where no contended lock
+	 * naturally enforces the correct ordering between the bail-out of the
+	 * waiter and the gpu reset work code.
 	 *
 	 * Lowest bit controls the reset state machine: Set means a reset is in
 	 * progress. This state will (presuming we don't have any bugs) decay
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 5bb370f..5226ebc 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -976,13 +976,22 @@ i915_gem_check_olr(struct intel_ring_buffer *ring, u32 seqno)
  * __wait_seqno - wait until execution of seqno has finished
  * @ring: the ring expected to report seqno
  * @seqno: duh!
+ * @reset_counter: reset sequence associated with the given seqno
  * @interruptible: do an interruptible wait (normally yes)
  * @timeout: in - how long to wait (NULL forever); out - how much time remaining
  *
+ * Note: It is of utmost importance that the passed in seqno and reset_counter
+ * values have been read by the caller in an smp safe manner. Where read-side
+ * locks are involved, it is sufficient to read the reset_counter before
+ * unlocking the lock that protects the seqno. For lockless tricks, the
+ * reset_counter _must_ be read before, and an appropriate smp_rmb must be
+ * inserted.
+ *
  * Returns 0 if the seqno was found within the alloted time. Else returns the
  * errno with remaining time filled in timeout argument.
  */
 static int __wait_seqno(struct intel_ring_buffer *ring, u32 seqno,
+			unsigned reset_counter,
 			bool interruptible, struct timespec *timeout)
 {
 	drm_i915_private_t *dev_priv = ring->dev->dev_private;
@@ -1012,7 +1021,8 @@ static int __wait_seqno(struct intel_ring_buffer *ring, u32 seqno,
 
 #define EXIT_COND \
 	(i915_seqno_passed(ring->get_seqno(ring, false), seqno) || \
-	i915_reset_in_progress(&dev_priv->gpu_error))
+	 i915_reset_in_progress(&dev_priv->gpu_error) || \
+	 reset_counter != atomic_read(&dev_priv->gpu_error.reset_counter))
 	do {
 		if (interruptible)
 			end = wait_event_interruptible_timeout(ring->irq_queue,
@@ -1022,6 +1032,13 @@ static int __wait_seqno(struct intel_ring_buffer *ring, u32 seqno,
 			end = wait_event_timeout(ring->irq_queue, EXIT_COND,
 						 timeout_jiffies);
 
+		/* We need to check whether any gpu reset happened in between
+		 * the caller grabbing the seqno and now ... */
+		if (reset_counter != atomic_read(&dev_priv->gpu_error.reset_counter))
+			end = -EAGAIN;
+
+		/* ... but upgrade the -EGAIN to an -EIO if the gpu is truely
+		 * gone. */
 		ret = i915_gem_check_wedge(&dev_priv->gpu_error, interruptible);
 		if (ret)
 			end = ret;
@@ -1076,7 +1093,9 @@ i915_wait_seqno(struct intel_ring_buffer *ring, uint32_t seqno)
 	if (ret)
 		return ret;
 
-	return __wait_seqno(ring, seqno, interruptible, NULL);
+	return __wait_seqno(ring, seqno,
+			    atomic_read(&dev_priv->gpu_error.reset_counter),
+			    interruptible, NULL);
 }
 
 /**
@@ -1123,6 +1142,7 @@ i915_gem_object_wait_rendering__nonblocking(struct drm_i915_gem_object *obj,
 	struct drm_device *dev = obj->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_ring_buffer *ring = obj->ring;
+	unsigned reset_counter;
 	u32 seqno;
 	int ret;
 
@@ -1141,8 +1161,9 @@ i915_gem_object_wait_rendering__nonblocking(struct drm_i915_gem_object *obj,
 	if (ret)
 		return ret;
 
+	reset_counter = atomic_read(&dev_priv->gpu_error.reset_counter);
 	mutex_unlock(&dev->struct_mutex);
-	ret = __wait_seqno(ring, seqno, true, NULL);
+	ret = __wait_seqno(ring, seqno, reset_counter, true, NULL);
 	mutex_lock(&dev->struct_mutex);
 
 	i915_gem_retire_requests_ring(ring);
@@ -2297,10 +2318,12 @@ i915_gem_object_flush_active(struct drm_i915_gem_object *obj)
 int
 i915_gem_wait_ioctl(struct drm_device *dev, void *data, struct drm_file *file)
 {
+	drm_i915_private_t *dev_priv = dev->dev_private;
 	struct drm_i915_gem_wait *args = data;
 	struct drm_i915_gem_object *obj;
 	struct intel_ring_buffer *ring = NULL;
 	struct timespec timeout_stack, *timeout = NULL;
+	unsigned reset_counter;
 	u32 seqno = 0;
 	int ret = 0;
 
@@ -2341,9 +2364,10 @@ i915_gem_wait_ioctl(struct drm_device *dev, void *data, struct drm_file *file)
 	}
 
 	drm_gem_object_unreference(&obj->base);
+	reset_counter = atomic_read(&dev_priv->gpu_error.reset_counter);
 	mutex_unlock(&dev->struct_mutex);
 
-	ret = __wait_seqno(ring, seqno, true, timeout);
+	ret = __wait_seqno(ring, seqno, reset_counter, true, timeout);
 	if (timeout) {
 		WARN_ON(!timespec_valid(timeout));
 		args->timeout_ns = timespec_to_ns(timeout);
@@ -3394,6 +3418,7 @@ i915_gem_ring_throttle(struct drm_device *dev, struct drm_file *file)
 	unsigned long recent_enough = jiffies - msecs_to_jiffies(20);
 	struct drm_i915_gem_request *request;
 	struct intel_ring_buffer *ring = NULL;
+	unsigned reset_counter;
 	u32 seqno = 0;
 	int ret;
 
@@ -3413,12 +3438,13 @@ i915_gem_ring_throttle(struct drm_device *dev, struct drm_file *file)
 		ring = request->ring;
 		seqno = request->seqno;
 	}
+	reset_counter = atomic_read(&dev_priv->gpu_error.reset_counter);
 	spin_unlock(&file_priv->mm.lock);
 
 	if (seqno == 0)
 		return 0;
 
-	ret = __wait_seqno(ring, seqno, true, NULL);
+	ret = __wait_seqno(ring, seqno, reset_counter, true, NULL);
 	if (ret == 0)
 		queue_delayed_work(dev_priv->wq, &dev_priv->mm.retire_work, 0);
 
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 4562c54..f833f2c 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -867,9 +867,11 @@ static void i915_error_work_func(struct work_struct *work)
 	drm_i915_private_t *dev_priv = container_of(error, drm_i915_private_t,
 						    gpu_error);
 	struct drm_device *dev = dev_priv->dev;
+	struct intel_ring_buffer *ring;
 	char *error_event[] = { "ERROR=1", NULL };
 	char *reset_event[] = { "RESET=1", NULL };
 	char *reset_done_event[] = { "ERROR=0", NULL };
+	int i, ret;
 
 	kobject_uevent_env(&dev->primary->kdev.kobj, KOBJ_CHANGE, error_event);
 
@@ -877,13 +879,31 @@ static void i915_error_work_func(struct work_struct *work)
 		DRM_DEBUG_DRIVER("resetting chip\n");
 		kobject_uevent_env(&dev->primary->kdev.kobj, KOBJ_CHANGE, reset_event);
 
-		if (!i915_reset(dev)) {
-			atomic_set(&error->reset_counter, 0);
-			kobject_uevent_env(&dev->primary->kdev.kobj, KOBJ_CHANGE, reset_done_event);
+		ret = i915_reset(dev);
+
+		if (ret == 0) {
+			/*
+			 * After all the gem state is reset, increment the reset
+			 * counter and wake up everyone waiting for the reset to
+			 * complete.
+			 *
+			 * Since unlock operations are a one-sided barrier only,
+			 * we need to insert a barrier here to order any seqno
+			 * updates before
+			 * the counter increment.
+			 */
+			smp_mb__before_atomic_inc();
+			atomic_inc(&dev_priv->gpu_error.reset_counter);
+
+			kobject_uevent_env(&dev->primary->kdev.kobj,
+					   KOBJ_CHANGE, reset_done_event);
 		} else {
 			atomic_set(&error->reset_counter, I915_WEDGED);
 		}
 
+		for_each_ring(ring, dev_priv, i)
+			wake_up_all(&ring->irq_queue);
+
 		wake_up_all(&dev_priv->gpu_error.reset_queue);
 	}
 }
@@ -1488,8 +1508,8 @@ void i915_handle_error(struct drm_device *dev, bool wedged)
 	i915_report_and_clear_eir(dev);
 
 	if (wedged) {
-		atomic_set(&dev_priv->gpu_error.reset_counter,
-			   I915_RESET_IN_PROGRESS_FLAG);
+		atomic_set_mask(I915_RESET_IN_PROGRESS_FLAG,
+				&dev_priv->gpu_error.reset_counter);
 
 		/*
 		 * Wakeup waiting processes so that the reset work item
-- 
cgit v0.10.2


From 7db0ba242b3a7ddcfc1450bc50eb3102c68f0244 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 6 Dec 2012 16:23:37 +0100
Subject: drm/i915: clarify concurrent hang detect/gpu reset consistency

Damien Lespiau wondered how race the gpu reset/hang detection code is
against concurrent gpu resets/hang detections or combinations thereof.
Luckily the single work item is guranteed to never run concurrently,
so reset handling is already single-threaded.

Hence we only have to worry about concurrent hang detections, or a
hang detection firing off while we're still processing an older gpu
reset request. Due to the new mechanism of setting the reset in
progress flag and the ordering guaranteed by the schedule_work
function there's nothing to do but add a comment explaining why we're
safe.

The only thing I've noticed is that we still try to reset the gpu now,
even when it is declared terminally wedged. Add a check for that to
avoid continous warnings about failed resets, in case the hangcheck
timer ever gets stuck.

Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index f833f2c..943db10 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -875,9 +875,20 @@ static void i915_error_work_func(struct work_struct *work)
 
 	kobject_uevent_env(&dev->primary->kdev.kobj, KOBJ_CHANGE, error_event);
 
-	if (i915_reset_in_progress(error)) {
+	/*
+	 * Note that there's only one work item which does gpu resets, so we
+	 * need not worry about concurrent gpu resets potentially incrementing
+	 * error->reset_counter twice. We only need to take care of another
+	 * racing irq/hangcheck declaring the gpu dead for a second time. A
+	 * quick check for that is good enough: schedule_work ensures the
+	 * correct ordering between hang detection and this work item, and since
+	 * the reset in-progress bit is only ever set by code outside of this
+	 * work we don't need to worry about any other races.
+	 */
+	if (i915_reset_in_progress(error) && !i915_terminally_wedged(error)) {
 		DRM_DEBUG_DRIVER("resetting chip\n");
-		kobject_uevent_env(&dev->primary->kdev.kobj, KOBJ_CHANGE, reset_event);
+		kobject_uevent_env(&dev->primary->kdev.kobj, KOBJ_CHANGE,
+				   reset_event);
 
 		ret = i915_reset(dev);
 
-- 
cgit v0.10.2


From 1ffff60320879830e469e26062c18f75236822ba Mon Sep 17 00:00:00 2001
From: Jani Nikula <jani.nikula@intel.com>
Date: Tue, 22 Jan 2013 12:50:34 +0200
Subject: drm/i915: add quirk to invert brightness on eMachines G725

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=59628
Reported-by: Roland Gruber <post@rolandgruber.de>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 7725446..44f9d8f 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -8599,6 +8599,9 @@ static struct intel_quirk intel_quirks[] = {
 
 	/* Acer Aspire 5734Z must invert backlight brightness */
 	{ 0x2a42, 0x1025, 0x0459, quirk_invert_brightness },
+
+	/* Acer/eMachines G725 */
+	{ 0x2a42, 0x1025, 0x0210, quirk_invert_brightness },
 };
 
 static void intel_init_quirks(struct drm_device *dev)
-- 
cgit v0.10.2


From 01e3a8feb40e54b962a20fa7eb595c5efef5e109 Mon Sep 17 00:00:00 2001
From: Jani Nikula <jani.nikula@intel.com>
Date: Tue, 22 Jan 2013 12:50:35 +0200
Subject: drm/i915: add quirk to invert brightness on eMachines e725

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=31522#c35
[Note: There are more than one broken setups in the bug. This fixes one.]
Reported-by: Martins <andrissr@inbox.lv>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 44f9d8f..8575a62 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -8602,6 +8602,9 @@ static struct intel_quirk intel_quirks[] = {
 
 	/* Acer/eMachines G725 */
 	{ 0x2a42, 0x1025, 0x0210, quirk_invert_brightness },
+
+	/* Acer/eMachines e725 */
+	{ 0x2a42, 0x1025, 0x0212, quirk_invert_brightness },
 };
 
 static void intel_init_quirks(struct drm_device *dev)
-- 
cgit v0.10.2


From 5559ecadad5a73b27f863e92f4b4f369501dce6f Mon Sep 17 00:00:00 2001
From: Jani Nikula <jani.nikula@intel.com>
Date: Tue, 22 Jan 2013 12:50:36 +0200
Subject: drm/i915: add quirk to invert brightness on Packard Bell NCL20

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=44156
Reported-by: Alan Zimmerman <alan.zimm@gmail.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 8575a62..7262786 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -8605,6 +8605,9 @@ static struct intel_quirk intel_quirks[] = {
 
 	/* Acer/eMachines e725 */
 	{ 0x2a42, 0x1025, 0x0212, quirk_invert_brightness },
+
+	/* Acer/Packard Bell NCL20 */
+	{ 0x2a42, 0x1025, 0x034b, quirk_invert_brightness },
 };
 
 static void intel_init_quirks(struct drm_device *dev)
-- 
cgit v0.10.2


From 99433931950f33039d9e1a52b4ed9af3f1b58e84 Mon Sep 17 00:00:00 2001
From: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Date: Tue, 22 Jan 2013 14:12:17 +0200
Subject: drm/i915: use gem_set_seqno() on hardware init

When machine was rebooted or module was reloaded,
gem_hw_init() set last_seqno to be identical to next_seqno.
This lead to situation that waits for first ever request
always passed immediately regardless if it was actually
executed.

Use gem_set_seqno() to be consistent how hw is
initialized on init, wrap and on resume.

Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 5226ebc..801f77e 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -3941,8 +3941,6 @@ i915_gem_init_hw(struct drm_device *dev)
 
 	i915_gem_init_swizzling(dev);
 
-	dev_priv->next_seqno = dev_priv->last_seqno = (u32)~0 - 0x1000;
-
 	ret = intel_init_render_ring_buffer(dev);
 	if (ret)
 		return ret;
@@ -3959,6 +3957,10 @@ i915_gem_init_hw(struct drm_device *dev)
 			goto cleanup_bsd_ring;
 	}
 
+	ret = i915_gem_set_seqno(dev, ((u32)~0 - 0x1000));
+	if (ret)
+		return ret;
+
 	/*
 	 * XXX: There was some w/a described somewhere suggesting loading
 	 * contexts before PPGTT.
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index 9438bcd..dc6ae2f 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -1223,8 +1223,6 @@ static int intel_init_ring_buffer(struct drm_device *dev,
 	if (IS_I830(ring->dev) || IS_845G(ring->dev))
 		ring->effective_size -= 128;
 
-	intel_ring_init_seqno(ring, dev_priv->last_seqno);
-
 	return 0;
 
 err_unmap:
-- 
cgit v0.10.2


From 7b9f35a6dd72f89452c58bbdbaf063027bf857ec Mon Sep 17 00:00:00 2001
From: Wang Xingchao <xingchao.wang@intel.com>
Date: Tue, 22 Jan 2013 23:25:25 +0800
Subject: drm/i915: HDMI/DP - ELD info refresh support for Haswell
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

ELD info should be updated dynamically according to hot plug event.
For haswell chip, clear/set the eld valid bit and output enable bit
from callback intel_disable/eanble_ddi().

Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: Wang Xingchao <xingchao.wang@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c
index 2e904a5..33b9112 100644
--- a/drivers/gpu/drm/i915/intel_ddi.c
+++ b/drivers/gpu/drm/i915/intel_ddi.c
@@ -677,6 +677,7 @@ static void intel_ddi_mode_set(struct drm_encoder *encoder,
 	DRM_DEBUG_KMS("Preparing DDI mode for Haswell on port %c, pipe %c\n",
 		      port_name(port), pipe_name(pipe));
 
+	intel_crtc->eld_vld = false;
 	if (type == INTEL_OUTPUT_DISPLAYPORT || type == INTEL_OUTPUT_EDP) {
 		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
 
@@ -1287,10 +1288,14 @@ static void intel_ddi_post_disable(struct intel_encoder *intel_encoder)
 static void intel_enable_ddi(struct intel_encoder *intel_encoder)
 {
 	struct drm_encoder *encoder = &intel_encoder->base;
+	struct drm_crtc *crtc = encoder->crtc;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	int pipe = intel_crtc->pipe;
 	struct drm_device *dev = encoder->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	enum port port = intel_ddi_get_encoder_port(intel_encoder);
 	int type = intel_encoder->type;
+	uint32_t tmp;
 
 	if (type == INTEL_OUTPUT_HDMI) {
 		/* In HDMI/DVI mode, the port width, and swing/emphasis values
@@ -1303,18 +1308,34 @@ static void intel_enable_ddi(struct intel_encoder *intel_encoder)
 
 		ironlake_edp_backlight_on(intel_dp);
 	}
+
+	if (intel_crtc->eld_vld) {
+		tmp = I915_READ(HSW_AUD_PIN_ELD_CP_VLD);
+		tmp |= ((AUDIO_OUTPUT_ENABLE_A | AUDIO_ELD_VALID_A) << (pipe * 4));
+		I915_WRITE(HSW_AUD_PIN_ELD_CP_VLD, tmp);
+	}
 }
 
 static void intel_disable_ddi(struct intel_encoder *intel_encoder)
 {
 	struct drm_encoder *encoder = &intel_encoder->base;
+	struct drm_crtc *crtc = encoder->crtc;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	int pipe = intel_crtc->pipe;
 	int type = intel_encoder->type;
+	struct drm_device *dev = encoder->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t tmp;
 
 	if (type == INTEL_OUTPUT_EDP) {
 		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
 
 		ironlake_edp_backlight_off(intel_dp);
 	}
+
+	tmp = I915_READ(HSW_AUD_PIN_ELD_CP_VLD);
+	tmp &= ~((AUDIO_OUTPUT_ENABLE_A | AUDIO_ELD_VALID_A) << (pipe * 4));
+	I915_WRITE(HSW_AUD_PIN_ELD_CP_VLD, tmp);
 }
 
 int intel_ddi_get_cdclk_freq(struct drm_i915_private *dev_priv)
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 7262786..305be11 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -3721,10 +3721,12 @@ static void intel_crtc_disable(struct drm_crtc *crtc)
 	struct drm_device *dev = crtc->dev;
 	struct drm_connector *connector;
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 
 	/* crtc should still be enabled when we disable it. */
 	WARN_ON(!crtc->enabled);
 
+	intel_crtc->eld_vld = false;
 	dev_priv->display.crtc_disable(crtc);
 	intel_crtc_update_sarea(crtc, false);
 	dev_priv->display.off(crtc);
@@ -5792,6 +5794,7 @@ static void haswell_write_eld(struct drm_connector *connector,
 	struct drm_i915_private *dev_priv = connector->dev->dev_private;
 	uint8_t *eld = connector->eld;
 	struct drm_device *dev = crtc->dev;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	uint32_t eldv;
 	uint32_t i;
 	int len;
@@ -5833,6 +5836,7 @@ static void haswell_write_eld(struct drm_connector *connector,
 	DRM_DEBUG_DRIVER("ELD on pipe %c\n", pipe_name(pipe));
 
 	eldv = AUDIO_ELD_VALID_A << (pipe * 4);
+	intel_crtc->eld_vld = true;
 
 	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT)) {
 		DRM_DEBUG_DRIVER("ELD: DisplayPort detected\n");
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index aeff0d1..66619d8 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -211,6 +211,7 @@ struct intel_crtc {
 	 * some outputs connected to this crtc.
 	 */
 	bool active;
+	bool eld_vld;
 	bool primary_disabled; /* is the crtc obscured by a plane? */
 	bool lowfreq_avail;
 	struct intel_overlay *overlay;
-- 
cgit v0.10.2


From c00db2463978ffab59d731773aae1a4f4e11d78c Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Tue, 22 Jan 2013 15:33:27 +0100
Subject: drm/i915: fixup sbi_read/write locking

commit 09153000b8ca32a539a1207edebabd0d40b6c61b
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Wed Dec 12 14:06:44 2012 +0100

    drm/i915: rework locking for intel_dpio|sbi_read|write

reworked the locking around sbi_read/write functions for 3.8-fixes.
But

commit dde86e2db54545ef981b64805097a7b4c3156d6e
Author: Paulo Zanoni <paulo.r.zanoni@intel.com>
Date:   Sat Dec 1 12:04:25 2012 -0200

    drm/i915: add lpt_init_pch_refcl

Added new use-cases in the -next tree which has not been updated in
the merge. Fix it up.

Reported-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Tested-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 305be11..886124a 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -4873,6 +4873,8 @@ static void lpt_init_pch_refclk(struct drm_device *dev)
 	if (!has_vga)
 		return;
 
+	mutex_lock(&dev_priv->dpio_lock);
+
 	/* XXX: Rip out SDV support once Haswell ships for real. */
 	if (IS_HASWELL(dev) && (dev->pci_device & 0xFF00) == 0x0C00)
 		is_sdv = true;
@@ -5015,6 +5017,8 @@ static void lpt_init_pch_refclk(struct drm_device *dev)
 	tmp = intel_sbi_read(dev_priv, SBI_DBUFF0, SBI_ICLK);
 	tmp |= SBI_DBUFF0_ENABLE;
 	intel_sbi_write(dev_priv, SBI_DBUFF0, tmp, SBI_ICLK);
+
+	mutex_unlock(&dev_priv->dpio_lock);
 }
 
 /*
-- 
cgit v0.10.2


From ea4f3111ef0daffa1d11fded6f375227febca458 Mon Sep 17 00:00:00 2001
From: Steffen Trumtrar <s.trumtrar@pengutronix.de>
Date: Thu, 22 Nov 2012 13:11:32 +0100
Subject: viafb: rename display_timing to via_display_timing

The struct display_timing is specific to the via subsystem. The naming leads to
collisions with the new struct display_timing, which is supposed to be a shared
struct between different subsystems.
To clean this up, prepend the existing struct with the subsystem it is specific
to.

Signed-off-by: Steffen Trumtrar <s.trumtrar@pengutronix.de>

diff --git a/drivers/video/via/hw.c b/drivers/video/via/hw.c
index 80233da..22450908 100644
--- a/drivers/video/via/hw.c
+++ b/drivers/video/via/hw.c
@@ -1467,10 +1467,10 @@ void viafb_set_vclock(u32 clk, int set_iga)
 	via_write_misc_reg_mask(0x0C, 0x0C); /* select external clock */
 }
 
-struct display_timing var_to_timing(const struct fb_var_screeninfo *var,
+struct via_display_timing var_to_timing(const struct fb_var_screeninfo *var,
 	u16 cxres, u16 cyres)
 {
-	struct display_timing timing;
+	struct via_display_timing timing;
 	u16 dx = (var->xres - cxres) / 2, dy = (var->yres - cyres) / 2;
 
 	timing.hor_addr = cxres;
@@ -1491,7 +1491,7 @@ struct display_timing var_to_timing(const struct fb_var_screeninfo *var,
 void viafb_fill_crtc_timing(const struct fb_var_screeninfo *var,
 	u16 cxres, u16 cyres, int iga)
 {
-	struct display_timing crt_reg = var_to_timing(var,
+	struct via_display_timing crt_reg = var_to_timing(var,
 		cxres ? cxres : var->xres, cyres ? cyres : var->yres);
 
 	if (iga == IGA1)
diff --git a/drivers/video/via/hw.h b/drivers/video/via/hw.h
index a820575..3be073c 100644
--- a/drivers/video/via/hw.h
+++ b/drivers/video/via/hw.h
@@ -637,7 +637,7 @@ extern int viafb_LCD_ON;
 extern int viafb_DVI_ON;
 extern int viafb_hotplug;
 
-struct display_timing var_to_timing(const struct fb_var_screeninfo *var,
+struct via_display_timing var_to_timing(const struct fb_var_screeninfo *var,
 	u16 cxres, u16 cyres);
 void viafb_fill_crtc_timing(const struct fb_var_screeninfo *var,
 	u16 cxres, u16 cyres, int iga);
diff --git a/drivers/video/via/lcd.c b/drivers/video/via/lcd.c
index 980ee1b..5d21ff4 100644
--- a/drivers/video/via/lcd.c
+++ b/drivers/video/via/lcd.c
@@ -549,7 +549,7 @@ void viafb_lcd_set_mode(const struct fb_var_screeninfo *var, u16 cxres,
 	int panel_hres = plvds_setting_info->lcd_panel_hres;
 	int panel_vres = plvds_setting_info->lcd_panel_vres;
 	u32 clock;
-	struct display_timing timing;
+	struct via_display_timing timing;
 	struct fb_var_screeninfo panel_var;
 	const struct fb_videomode *mode_crt_table, *panel_crt_table;
 
diff --git a/drivers/video/via/share.h b/drivers/video/via/share.h
index 3158dfc..65c65c6 100644
--- a/drivers/video/via/share.h
+++ b/drivers/video/via/share.h
@@ -319,7 +319,7 @@ struct crt_mode_table {
 	int refresh_rate;
 	int h_sync_polarity;
 	int v_sync_polarity;
-	struct display_timing crtc;
+	struct via_display_timing crtc;
 };
 
 struct io_reg {
diff --git a/drivers/video/via/via_modesetting.c b/drivers/video/via/via_modesetting.c
index 0e431ae..0b414b0 100644
--- a/drivers/video/via/via_modesetting.c
+++ b/drivers/video/via/via_modesetting.c
@@ -30,9 +30,9 @@
 #include "debug.h"
 
 
-void via_set_primary_timing(const struct display_timing *timing)
+void via_set_primary_timing(const struct via_display_timing *timing)
 {
-	struct display_timing raw;
+	struct via_display_timing raw;
 
 	raw.hor_total = timing->hor_total / 8 - 5;
 	raw.hor_addr = timing->hor_addr / 8 - 1;
@@ -88,9 +88,9 @@ void via_set_primary_timing(const struct display_timing *timing)
 	via_write_reg_mask(VIACR, 0x17, 0x80, 0x80);
 }
 
-void via_set_secondary_timing(const struct display_timing *timing)
+void via_set_secondary_timing(const struct via_display_timing *timing)
 {
-	struct display_timing raw;
+	struct via_display_timing raw;
 
 	raw.hor_total = timing->hor_total - 1;
 	raw.hor_addr = timing->hor_addr - 1;
diff --git a/drivers/video/via/via_modesetting.h b/drivers/video/via/via_modesetting.h
index 06e09fe..f6a6503 100644
--- a/drivers/video/via/via_modesetting.h
+++ b/drivers/video/via/via_modesetting.h
@@ -33,7 +33,7 @@
 #define VIA_PITCH_MAX	0x3FF8
 
 
-struct display_timing {
+struct via_display_timing {
 	u16 hor_total;
 	u16 hor_addr;
 	u16 hor_blank_start;
@@ -49,8 +49,8 @@ struct display_timing {
 };
 
 
-void via_set_primary_timing(const struct display_timing *timing);
-void via_set_secondary_timing(const struct display_timing *timing);
+void via_set_primary_timing(const struct via_display_timing *timing);
+void via_set_secondary_timing(const struct via_display_timing *timing);
 void via_set_primary_address(u32 addr);
 void via_set_secondary_address(u32 addr);
 void via_set_primary_pitch(u32 pitch);
-- 
cgit v0.10.2


From 8714c0cecfc28f7ce73a520be4831f09743c4fd7 Mon Sep 17 00:00:00 2001
From: Steffen Trumtrar <s.trumtrar@pengutronix.de>
Date: Mon, 17 Dec 2012 14:20:17 +0100
Subject: video: add display_timing and videomode

Add display_timing structure and the according helper functions. This allows
the description of a display via its supported timing parameters.

Also, add helper functions to convert from display timings to a generic videomode
structure.

The struct display_timing specifies all needed parameters to describe the signal
properties of a display in one mode. This includes
    - ranges for signals that may have min-, max- and typical values
    - single integers for signals that can be on, off or are ignored
    - booleans for signals that are either on or off

As a display may support multiple modes like this, a struct display_timings is
added, that holds all given struct display_timing pointers and declares the
native mode of the display.

Although a display may state that a signal can be in a range, it is driven with
fixed values that indicate a videomode. Therefore graphic drivers don't need all
the information of struct display_timing, but would generate a videomode from
the given set of supported signal timings and work with that.

The video subsystems all define their own structs that describe a mode and work
with that (e.g. fb_videomode or drm_display_mode). To slowly replace all those
various structures and allow code reuse across those subsystems, add struct
videomode as a generic description.

This patch only includes the most basic fields in struct videomode. All missing
fields that are needed to have a really generic video mode description can be
added at a later stage.

Signed-off-by: Steffen Trumtrar <s.trumtrar@pengutronix.de>
Reviewed-by: Thierry Reding <thierry.reding@avionic-design.de>
Acked-by: Thierry Reding <thierry.reding@avionic-design.de>
Tested-by: Thierry Reding <thierry.reding@avionic-design.de>
Tested-by: Philipp Zabel <p.zabel@pengutronix.de>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-by: Afzal Mohammed <Afzal@ti.com>
Tested-by: Rob Clark <robclark@gmail.com>
Tested-by: Leela Krishna Amudala <leelakrishna.a@gmail.com>

diff --git a/drivers/video/Kconfig b/drivers/video/Kconfig
index e7068c5..09a8f0d 100644
--- a/drivers/video/Kconfig
+++ b/drivers/video/Kconfig
@@ -33,6 +33,12 @@ config VIDEO_OUTPUT_CONTROL
 	  This framework adds support for low-level control of the video 
 	  output switch.
 
+config DISPLAY_TIMING
+       bool
+
+config VIDEOMODE
+       bool
+
 menuconfig FB
 	tristate "Support for frame buffer devices"
 	---help---
diff --git a/drivers/video/Makefile b/drivers/video/Makefile
index 768a137..e0dd820 100644
--- a/drivers/video/Makefile
+++ b/drivers/video/Makefile
@@ -168,3 +168,5 @@ obj-$(CONFIG_FB_VIRTUAL)          += vfb.o
 
 #video output switch sysfs driver
 obj-$(CONFIG_VIDEO_OUTPUT_CONTROL) += output.o
+obj-$(CONFIG_DISPLAY_TIMING) += display_timing.o
+obj-$(CONFIG_VIDEOMODE) += videomode.o
diff --git a/drivers/video/display_timing.c b/drivers/video/display_timing.c
new file mode 100644
index 0000000..5e1822c
--- /dev/null
+++ b/drivers/video/display_timing.c
@@ -0,0 +1,24 @@
+/*
+ * generic display timing functions
+ *
+ * Copyright (c) 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>, Pengutronix
+ *
+ * This file is released under the GPLv2
+ */
+
+#include <linux/export.h>
+#include <linux/slab.h>
+#include <video/display_timing.h>
+
+void display_timings_release(struct display_timings *disp)
+{
+	if (disp->timings) {
+		unsigned int i;
+
+		for (i = 0; i < disp->num_timings; i++)
+			kfree(disp->timings[i]);
+		kfree(disp->timings);
+	}
+	kfree(disp);
+}
+EXPORT_SYMBOL_GPL(display_timings_release);
diff --git a/drivers/video/videomode.c b/drivers/video/videomode.c
new file mode 100644
index 0000000..21c47a2
--- /dev/null
+++ b/drivers/video/videomode.c
@@ -0,0 +1,39 @@
+/*
+ * generic display timing functions
+ *
+ * Copyright (c) 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>, Pengutronix
+ *
+ * This file is released under the GPLv2
+ */
+
+#include <linux/errno.h>
+#include <linux/export.h>
+#include <video/display_timing.h>
+#include <video/videomode.h>
+
+int videomode_from_timing(const struct display_timings *disp,
+			  struct videomode *vm, unsigned int index)
+{
+	struct display_timing *dt;
+
+	dt = display_timings_get(disp, index);
+	if (!dt)
+		return -EINVAL;
+
+	vm->pixelclock = display_timing_get_value(&dt->pixelclock, TE_TYP);
+	vm->hactive = display_timing_get_value(&dt->hactive, TE_TYP);
+	vm->hfront_porch = display_timing_get_value(&dt->hfront_porch, TE_TYP);
+	vm->hback_porch = display_timing_get_value(&dt->hback_porch, TE_TYP);
+	vm->hsync_len = display_timing_get_value(&dt->hsync_len, TE_TYP);
+
+	vm->vactive = display_timing_get_value(&dt->vactive, TE_TYP);
+	vm->vfront_porch = display_timing_get_value(&dt->vfront_porch, TE_TYP);
+	vm->vback_porch = display_timing_get_value(&dt->vback_porch, TE_TYP);
+	vm->vsync_len = display_timing_get_value(&dt->vsync_len, TE_TYP);
+
+	vm->dmt_flags = dt->dmt_flags;
+	vm->data_flags = dt->data_flags;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(videomode_from_timing);
diff --git a/include/video/display_timing.h b/include/video/display_timing.h
new file mode 100644
index 0000000..71e9a38
--- /dev/null
+++ b/include/video/display_timing.h
@@ -0,0 +1,124 @@
+/*
+ * Copyright 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>
+ *
+ * description of display timings
+ *
+ * This file is released under the GPLv2
+ */
+
+#ifndef __LINUX_DISPLAY_TIMING_H
+#define __LINUX_DISPLAY_TIMING_H
+
+#include <linux/bitops.h>
+#include <linux/types.h>
+
+/* VESA display monitor timing parameters */
+#define VESA_DMT_HSYNC_LOW		BIT(0)
+#define VESA_DMT_HSYNC_HIGH		BIT(1)
+#define VESA_DMT_VSYNC_LOW		BIT(2)
+#define VESA_DMT_VSYNC_HIGH		BIT(3)
+
+/* display specific flags */
+#define DISPLAY_FLAGS_DE_LOW		BIT(0)	/* data enable flag */
+#define DISPLAY_FLAGS_DE_HIGH		BIT(1)
+#define DISPLAY_FLAGS_PIXDATA_POSEDGE	BIT(2)	/* drive data on pos. edge */
+#define DISPLAY_FLAGS_PIXDATA_NEGEDGE	BIT(3)	/* drive data on neg. edge */
+#define DISPLAY_FLAGS_INTERLACED	BIT(4)
+#define DISPLAY_FLAGS_DOUBLESCAN	BIT(5)
+
+/*
+ * A single signal can be specified via a range of minimal and maximal values
+ * with a typical value, that lies somewhere inbetween.
+ */
+struct timing_entry {
+	u32 min;
+	u32 typ;
+	u32 max;
+};
+
+enum timing_entry_index {
+	TE_MIN = 0,
+	TE_TYP = 1,
+	TE_MAX = 2,
+};
+
+/*
+ * Single "mode" entry. This describes one set of signal timings a display can
+ * have in one setting. This struct can later be converted to struct videomode
+ * (see include/video/videomode.h). As each timing_entry can be defined as a
+ * range, one struct display_timing may become multiple struct videomodes.
+ *
+ * Example: hsync active high, vsync active low
+ *
+ *				    Active Video
+ * Video  ______________________XXXXXXXXXXXXXXXXXXXXXX_____________________
+ *	  |<- sync ->|<- back ->|<----- active ----->|<- front ->|<- sync..
+ *	  |	     |	 porch  |		     |	 porch	 |
+ *
+ * HSync _|¯¯¯¯¯¯¯¯¯¯|___________________________________________|¯¯¯¯¯¯¯¯¯
+ *
+ * VSync ¯|__________|¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯|_________
+ */
+struct display_timing {
+	struct timing_entry pixelclock;
+
+	struct timing_entry hactive;		/* hor. active video */
+	struct timing_entry hfront_porch;	/* hor. front porch */
+	struct timing_entry hback_porch;	/* hor. back porch */
+	struct timing_entry hsync_len;		/* hor. sync len */
+
+	struct timing_entry vactive;		/* ver. active video */
+	struct timing_entry vfront_porch;	/* ver. front porch */
+	struct timing_entry vback_porch;	/* ver. back porch */
+	struct timing_entry vsync_len;		/* ver. sync len */
+
+	unsigned int dmt_flags;			/* VESA DMT flags */
+	unsigned int data_flags;		/* video data flags */
+};
+
+/*
+ * This describes all timing settings a display provides.
+ * The native_mode is the default setting for this display.
+ * Drivers that can handle multiple videomodes should work with this struct and
+ * convert each entry to the desired end result.
+ */
+struct display_timings {
+	unsigned int num_timings;
+	unsigned int native_mode;
+
+	struct display_timing **timings;
+};
+
+/* get value specified by index from struct timing_entry */
+static inline u32 display_timing_get_value(const struct timing_entry *te,
+					   enum timing_entry_index index)
+{
+	switch (index) {
+	case TE_MIN:
+		return te->min;
+		break;
+	case TE_TYP:
+		return te->typ;
+		break;
+	case TE_MAX:
+		return te->max;
+		break;
+	default:
+		return te->typ;
+	}
+}
+
+/* get one entry from struct display_timings */
+static inline struct display_timing *display_timings_get(const struct
+							 display_timings *disp,
+							 unsigned int index)
+{
+	if (disp->num_timings > index)
+		return disp->timings[index];
+	else
+		return NULL;
+}
+
+void display_timings_release(struct display_timings *disp);
+
+#endif
diff --git a/include/video/videomode.h b/include/video/videomode.h
new file mode 100644
index 0000000..a421562
--- /dev/null
+++ b/include/video/videomode.h
@@ -0,0 +1,48 @@
+/*
+ * Copyright 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>
+ *
+ * generic videomode description
+ *
+ * This file is released under the GPLv2
+ */
+
+#ifndef __LINUX_VIDEOMODE_H
+#define __LINUX_VIDEOMODE_H
+
+#include <linux/types.h>
+#include <video/display_timing.h>
+
+/*
+ * Subsystem independent description of a videomode.
+ * Can be generated from struct display_timing.
+ */
+struct videomode {
+	unsigned long pixelclock;	/* pixelclock in Hz */
+
+	u32 hactive;
+	u32 hfront_porch;
+	u32 hback_porch;
+	u32 hsync_len;
+
+	u32 vactive;
+	u32 vfront_porch;
+	u32 vback_porch;
+	u32 vsync_len;
+
+	unsigned int dmt_flags;	/* VESA DMT flags */
+	unsigned int data_flags; /* video data flags */
+};
+
+/**
+ * videomode_from_timing - convert display timing to videomode
+ * @disp: structure with all possible timing entries
+ * @vm: return value
+ * @index: index into the list of display timings in devicetree
+ *
+ * DESCRIPTION:
+ * This function converts a struct display_timing to a struct videomode.
+ */
+int videomode_from_timing(const struct display_timings *disp,
+			  struct videomode *vm, unsigned int index);
+
+#endif
-- 
cgit v0.10.2


From cc3f414cf2e404130584b63d373161ba6fd24bc2 Mon Sep 17 00:00:00 2001
From: Steffen Trumtrar <s.trumtrar@pengutronix.de>
Date: Thu, 4 Oct 2012 15:32:52 +0200
Subject: video: add of helper for display timings/videomode

This adds support for reading display timings from DT into a struct
display_timings. The of_display_timing implementation supports multiple
subnodes. All children are read into an array, that can be queried.

If no native mode is specified, the first subnode will be used.

For cases where the graphics driver knows there can be only one
mode description or where the driver only supports one mode, a helper
function of_get_videomode is added, that gets a struct videomode from DT.

Signed-off-by: Steffen Trumtrar <s.trumtrar@pengutronix.de>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Stephen Warren <swarren@nvidia.com>
Reviewed-by: Thierry Reding <thierry.reding@avionic-design.de>
Acked-by: Thierry Reding <thierry.reding@avionic-design.de>
Tested-by: Thierry Reding <thierry.reding@avionic-design.de>
Tested-by: Philipp Zabel <p.zabel@pengutronix.de>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-by: Afzal Mohammed <Afzal@ti.com>
Tested-by: Rob Clark <robclark@gmail.com>
Tested-by: Leela Krishna Amudala <leelakrishna.a@gmail.com>

diff --git a/Documentation/devicetree/bindings/video/display-timing.txt b/Documentation/devicetree/bindings/video/display-timing.txt
new file mode 100644
index 0000000..1500385
--- /dev/null
+++ b/Documentation/devicetree/bindings/video/display-timing.txt
@@ -0,0 +1,109 @@
+display-timing bindings
+=======================
+
+display-timings node
+--------------------
+
+required properties:
+ - none
+
+optional properties:
+ - native-mode: The native mode for the display, in case multiple modes are
+		provided. When omitted, assume the first node is the native.
+
+timing subnode
+--------------
+
+required properties:
+ - hactive, vactive: display resolution
+ - hfront-porch, hback-porch, hsync-len: horizontal display timing parameters
+   in pixels
+   vfront-porch, vback-porch, vsync-len: vertical display timing parameters in
+   lines
+ - clock-frequency: display clock in Hz
+
+optional properties:
+ - hsync-active: hsync pulse is active low/high/ignored
+ - vsync-active: vsync pulse is active low/high/ignored
+ - de-active: data-enable pulse is active low/high/ignored
+ - pixelclk-active: with
+			- active high = drive pixel data on rising edge/
+					sample data on falling edge
+			- active low  = drive pixel data on falling edge/
+					sample data on rising edge
+			- ignored     = ignored
+ - interlaced (bool): boolean to enable interlaced mode
+ - doublescan (bool): boolean to enable doublescan mode
+
+All the optional properties that are not bool follow the following logic:
+    <1>: high active
+    <0>: low active
+    omitted: not used on hardware
+
+There are different ways of describing the capabilities of a display. The
+devicetree representation corresponds to the one commonly found in datasheets
+for displays. If a display supports multiple signal timings, the native-mode
+can be specified.
+
+The parameters are defined as:
+
+  +----------+-------------------------------------+----------+-------+
+  |          |        ↑                            |          |       |
+  |          |        |vback_porch                 |          |       |
+  |          |        ↓                            |          |       |
+  +----------#######################################----------+-------+
+  |          #        ↑                            #          |       |
+  |          #        |                            #          |       |
+  |  hback   #        |                            #  hfront  | hsync |
+  |   porch  #        |       hactive              #  porch   |  len  |
+  |<-------->#<-------+--------------------------->#<-------->|<----->|
+  |          #        |                            #          |       |
+  |          #        |vactive                     #          |       |
+  |          #        |                            #          |       |
+  |          #        ↓                            #          |       |
+  +----------#######################################----------+-------+
+  |          |        ↑                            |          |       |
+  |          |        |vfront_porch                |          |       |
+  |          |        ↓                            |          |       |
+  +----------+-------------------------------------+----------+-------+
+  |          |        ↑                            |          |       |
+  |          |        |vsync_len                   |          |       |
+  |          |        ↓                            |          |       |
+  +----------+-------------------------------------+----------+-------+
+
+Example:
+
+	display-timings {
+		native-mode = <&timing0>;
+		timing0: 1080p24 {
+			/* 1920x1080p24 */
+			clock-frequency = <52000000>;
+			hactive = <1920>;
+			vactive = <1080>;
+			hfront-porch = <25>;
+			hback-porch = <25>;
+			hsync-len = <25>;
+			vback-porch = <2>;
+			vfront-porch = <2>;
+			vsync-len = <2>;
+			hsync-active = <1>;
+		};
+	};
+
+Every required property also supports the use of ranges, so the commonly used
+datasheet description with minimum, typical and maximum values can be used.
+
+Example:
+
+	timing1: timing {
+		/* 1920x1080p24 */
+		clock-frequency = <148500000>;
+		hactive = <1920>;
+		vactive = <1080>;
+		hsync-len = <0 44 60>;
+		hfront-porch = <80 88 95>;
+		hback-porch = <100 148 160>;
+		vfront-porch = <0 4 6>;
+		vback-porch = <0 36 50>;
+		vsync-len = <0 5 6>;
+	};
diff --git a/drivers/video/Kconfig b/drivers/video/Kconfig
index 09a8f0d..807c7fa 100644
--- a/drivers/video/Kconfig
+++ b/drivers/video/Kconfig
@@ -39,6 +39,21 @@ config DISPLAY_TIMING
 config VIDEOMODE
        bool
 
+config OF_DISPLAY_TIMING
+	bool "Enable device tree display timing support"
+	depends on OF
+	select DISPLAY_TIMING
+	help
+	  helper to parse display timings from the devicetree
+
+config OF_VIDEOMODE
+	bool "Enable device tree videomode support"
+	depends on OF
+	select VIDEOMODE
+	select OF_DISPLAY_TIMING
+	help
+	  helper to get videomodes from the devicetree
+
 menuconfig FB
 	tristate "Support for frame buffer devices"
 	---help---
diff --git a/drivers/video/Makefile b/drivers/video/Makefile
index e0dd820..f592f3b 100644
--- a/drivers/video/Makefile
+++ b/drivers/video/Makefile
@@ -169,4 +169,6 @@ obj-$(CONFIG_FB_VIRTUAL)          += vfb.o
 #video output switch sysfs driver
 obj-$(CONFIG_VIDEO_OUTPUT_CONTROL) += output.o
 obj-$(CONFIG_DISPLAY_TIMING) += display_timing.o
+obj-$(CONFIG_OF_DISPLAY_TIMING) += of_display_timing.o
 obj-$(CONFIG_VIDEOMODE) += videomode.o
+obj-$(CONFIG_OF_VIDEOMODE) += of_videomode.o
diff --git a/drivers/video/of_display_timing.c b/drivers/video/of_display_timing.c
new file mode 100644
index 0000000..13ecd98
--- /dev/null
+++ b/drivers/video/of_display_timing.c
@@ -0,0 +1,239 @@
+/*
+ * OF helpers for parsing display timings
+ *
+ * Copyright (c) 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>, Pengutronix
+ *
+ * based on of_videomode.c by Sascha Hauer <s.hauer@pengutronix.de>
+ *
+ * This file is released under the GPLv2
+ */
+#include <linux/export.h>
+#include <linux/of.h>
+#include <linux/slab.h>
+#include <video/display_timing.h>
+#include <video/of_display_timing.h>
+
+/**
+ * parse_timing_property - parse timing_entry from device_node
+ * @np: device_node with the property
+ * @name: name of the property
+ * @result: will be set to the return value
+ *
+ * DESCRIPTION:
+ * Every display_timing can be specified with either just the typical value or
+ * a range consisting of min/typ/max. This function helps handling this
+ **/
+static int parse_timing_property(struct device_node *np, const char *name,
+			  struct timing_entry *result)
+{
+	struct property *prop;
+	int length, cells, ret;
+
+	prop = of_find_property(np, name, &length);
+	if (!prop) {
+		pr_err("%s: could not find property %s\n",
+			of_node_full_name(np), name);
+		return -EINVAL;
+	}
+
+	cells = length / sizeof(u32);
+	if (cells == 1) {
+		ret = of_property_read_u32(np, name, &result->typ);
+		result->min = result->typ;
+		result->max = result->typ;
+	} else if (cells == 3) {
+		ret = of_property_read_u32_array(np, name, &result->min, cells);
+	} else {
+		pr_err("%s: illegal timing specification in %s\n",
+			of_node_full_name(np), name);
+		return -EINVAL;
+	}
+
+	return ret;
+}
+
+/**
+ * of_get_display_timing - parse display_timing entry from device_node
+ * @np: device_node with the properties
+ **/
+static struct display_timing *of_get_display_timing(struct device_node *np)
+{
+	struct display_timing *dt;
+	u32 val = 0;
+	int ret = 0;
+
+	dt = kzalloc(sizeof(*dt), GFP_KERNEL);
+	if (!dt) {
+		pr_err("%s: could not allocate display_timing struct\n",
+			of_node_full_name(np));
+		return NULL;
+	}
+
+	ret |= parse_timing_property(np, "hback-porch", &dt->hback_porch);
+	ret |= parse_timing_property(np, "hfront-porch", &dt->hfront_porch);
+	ret |= parse_timing_property(np, "hactive", &dt->hactive);
+	ret |= parse_timing_property(np, "hsync-len", &dt->hsync_len);
+	ret |= parse_timing_property(np, "vback-porch", &dt->vback_porch);
+	ret |= parse_timing_property(np, "vfront-porch", &dt->vfront_porch);
+	ret |= parse_timing_property(np, "vactive", &dt->vactive);
+	ret |= parse_timing_property(np, "vsync-len", &dt->vsync_len);
+	ret |= parse_timing_property(np, "clock-frequency", &dt->pixelclock);
+
+	dt->dmt_flags = 0;
+	dt->data_flags = 0;
+	if (!of_property_read_u32(np, "vsync-active", &val))
+		dt->dmt_flags |= val ? VESA_DMT_VSYNC_HIGH :
+				VESA_DMT_VSYNC_LOW;
+	if (!of_property_read_u32(np, "hsync-active", &val))
+		dt->dmt_flags |= val ? VESA_DMT_HSYNC_HIGH :
+				VESA_DMT_HSYNC_LOW;
+	if (!of_property_read_u32(np, "de-active", &val))
+		dt->data_flags |= val ? DISPLAY_FLAGS_DE_HIGH :
+				DISPLAY_FLAGS_DE_LOW;
+	if (!of_property_read_u32(np, "pixelclk-active", &val))
+		dt->data_flags |= val ? DISPLAY_FLAGS_PIXDATA_POSEDGE :
+				DISPLAY_FLAGS_PIXDATA_NEGEDGE;
+
+	if (of_property_read_bool(np, "interlaced"))
+		dt->data_flags |= DISPLAY_FLAGS_INTERLACED;
+	if (of_property_read_bool(np, "doublescan"))
+		dt->data_flags |= DISPLAY_FLAGS_DOUBLESCAN;
+
+	if (ret) {
+		pr_err("%s: error reading timing properties\n",
+			of_node_full_name(np));
+		kfree(dt);
+		return NULL;
+	}
+
+	return dt;
+}
+
+/**
+ * of_get_display_timings - parse all display_timing entries from a device_node
+ * @np: device_node with the subnodes
+ **/
+struct display_timings *of_get_display_timings(struct device_node *np)
+{
+	struct device_node *timings_np;
+	struct device_node *entry;
+	struct device_node *native_mode;
+	struct display_timings *disp;
+
+	if (!np) {
+		pr_err("%s: no devicenode given\n", of_node_full_name(np));
+		return NULL;
+	}
+
+	timings_np = of_find_node_by_name(np, "display-timings");
+	if (!timings_np) {
+		pr_err("%s: could not find display-timings node\n",
+			of_node_full_name(np));
+		return NULL;
+	}
+
+	disp = kzalloc(sizeof(*disp), GFP_KERNEL);
+	if (!disp) {
+		pr_err("%s: could not allocate struct disp'\n",
+			of_node_full_name(np));
+		goto dispfail;
+	}
+
+	entry = of_parse_phandle(timings_np, "native-mode", 0);
+	/* assume first child as native mode if none provided */
+	if (!entry)
+		entry = of_get_next_child(np, NULL);
+	/* if there is no child, it is useless to go on */
+	if (!entry) {
+		pr_err("%s: no timing specifications given\n",
+			of_node_full_name(np));
+		goto entryfail;
+	}
+
+	pr_debug("%s: using %s as default timing\n",
+		of_node_full_name(np), entry->name);
+
+	native_mode = entry;
+
+	disp->num_timings = of_get_child_count(timings_np);
+	if (disp->num_timings == 0) {
+		/* should never happen, as entry was already found above */
+		pr_err("%s: no timings specified\n", of_node_full_name(np));
+		goto entryfail;
+	}
+
+	disp->timings = kzalloc(sizeof(struct display_timing *) *
+				disp->num_timings, GFP_KERNEL);
+	if (!disp->timings) {
+		pr_err("%s: could not allocate timings array\n",
+			of_node_full_name(np));
+		goto entryfail;
+	}
+
+	disp->num_timings = 0;
+	disp->native_mode = 0;
+
+	for_each_child_of_node(timings_np, entry) {
+		struct display_timing *dt;
+
+		dt = of_get_display_timing(entry);
+		if (!dt) {
+			/*
+			 * to not encourage wrong devicetrees, fail in case of
+			 * an error
+			 */
+			pr_err("%s: error in timing %d\n",
+				of_node_full_name(np), disp->num_timings + 1);
+			goto timingfail;
+		}
+
+		if (native_mode == entry)
+			disp->native_mode = disp->num_timings;
+
+		disp->timings[disp->num_timings] = dt;
+		disp->num_timings++;
+	}
+	of_node_put(timings_np);
+	/*
+	 * native_mode points to the device_node returned by of_parse_phandle
+	 * therefore call of_node_put on it
+	 */
+	of_node_put(native_mode);
+
+	pr_debug("%s: got %d timings. Using timing #%d as default\n",
+		of_node_full_name(np), disp->num_timings,
+		disp->native_mode + 1);
+
+	return disp;
+
+timingfail:
+	if (native_mode)
+		of_node_put(native_mode);
+	display_timings_release(disp);
+entryfail:
+	kfree(disp);
+dispfail:
+	of_node_put(timings_np);
+	return NULL;
+}
+EXPORT_SYMBOL_GPL(of_get_display_timings);
+
+/**
+ * of_display_timings_exist - check if a display-timings node is provided
+ * @np: device_node with the timing
+ **/
+int of_display_timings_exist(struct device_node *np)
+{
+	struct device_node *timings_np;
+
+	if (!np)
+		return -EINVAL;
+
+	timings_np = of_parse_phandle(np, "display-timings", 0);
+	if (!timings_np)
+		return -EINVAL;
+
+	of_node_put(timings_np);
+	return 1;
+}
+EXPORT_SYMBOL_GPL(of_display_timings_exist);
diff --git a/drivers/video/of_videomode.c b/drivers/video/of_videomode.c
new file mode 100644
index 0000000..5b8066c
--- /dev/null
+++ b/drivers/video/of_videomode.c
@@ -0,0 +1,54 @@
+/*
+ * generic videomode helper
+ *
+ * Copyright (c) 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>, Pengutronix
+ *
+ * This file is released under the GPLv2
+ */
+#include <linux/errno.h>
+#include <linux/export.h>
+#include <linux/of.h>
+#include <video/display_timing.h>
+#include <video/of_display_timing.h>
+#include <video/of_videomode.h>
+#include <video/videomode.h>
+
+/**
+ * of_get_videomode - get the videomode #<index> from devicetree
+ * @np - devicenode with the display_timings
+ * @vm - set to return value
+ * @index - index into list of display_timings
+ *	    (Set this to OF_USE_NATIVE_MODE to use whatever mode is
+ *	     specified as native mode in the DT.)
+ *
+ * DESCRIPTION:
+ * Get a list of all display timings and put the one
+ * specified by index into *vm. This function should only be used, if
+ * only one videomode is to be retrieved. A driver that needs to work
+ * with multiple/all videomodes should work with
+ * of_get_display_timings instead.
+ **/
+int of_get_videomode(struct device_node *np, struct videomode *vm,
+		     int index)
+{
+	struct display_timings *disp;
+	int ret;
+
+	disp = of_get_display_timings(np);
+	if (!disp) {
+		pr_err("%s: no timings specified\n", of_node_full_name(np));
+		return -EINVAL;
+	}
+
+	if (index == OF_USE_NATIVE_MODE)
+		index = disp->native_mode;
+
+	ret = videomode_from_timing(disp, vm, index);
+	if (ret)
+		return ret;
+
+	display_timings_release(disp);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(of_get_videomode);
diff --git a/include/video/of_display_timing.h b/include/video/of_display_timing.h
new file mode 100644
index 0000000..8016eb7
--- /dev/null
+++ b/include/video/of_display_timing.h
@@ -0,0 +1,20 @@
+/*
+ * Copyright 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>
+ *
+ * display timings of helpers
+ *
+ * This file is released under the GPLv2
+ */
+
+#ifndef __LINUX_OF_DISPLAY_TIMING_H
+#define __LINUX_OF_DISPLAY_TIMING_H
+
+struct device_node;
+struct display_timings;
+
+#define OF_USE_NATIVE_MODE -1
+
+struct display_timings *of_get_display_timings(struct device_node *np);
+int of_display_timings_exist(struct device_node *np);
+
+#endif
diff --git a/include/video/of_videomode.h b/include/video/of_videomode.h
new file mode 100644
index 0000000..a07efcc
--- /dev/null
+++ b/include/video/of_videomode.h
@@ -0,0 +1,18 @@
+/*
+ * Copyright 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>
+ *
+ * videomode of-helpers
+ *
+ * This file is released under the GPLv2
+ */
+
+#ifndef __LINUX_OF_VIDEOMODE_H
+#define __LINUX_OF_VIDEOMODE_H
+
+struct device_node;
+struct videomode;
+
+int of_get_videomode(struct device_node *np, struct videomode *vm,
+		     int index);
+
+#endif /* __LINUX_OF_VIDEOMODE_H */
-- 
cgit v0.10.2


From 2db54c72395298a58f29c75ae880be9e478fdbbd Mon Sep 17 00:00:00 2001
From: Steffen Trumtrar <s.trumtrar@pengutronix.de>
Date: Wed, 14 Nov 2012 10:55:04 +0100
Subject: fbmon: add videomode helpers

Add a function to convert from the generic videomode to a fb_videomode.

Signed-off-by: Steffen Trumtrar <s.trumtrar@pengutronix.de>
Reviewed-by: Thierry Reding <thierry.reding@avionic-design.de>
Acked-by: Thierry Reding <thierry.reding@avionic-design.de>
Tested-by: Thierry Reding <thierry.reding@avionic-design.de>
Tested-by: Philipp Zabel <p.zabel@pengutronix.de>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-by: Afzal Mohammed <Afzal@ti.com>
Tested-by: Rob Clark <robclark@gmail.com>
Tested-by: Leela Krishna Amudala <leelakrishna.a@gmail.com>

diff --git a/drivers/video/fbmon.c b/drivers/video/fbmon.c
index cef6557..17ce135 100644
--- a/drivers/video/fbmon.c
+++ b/drivers/video/fbmon.c
@@ -31,6 +31,7 @@
 #include <linux/pci.h>
 #include <linux/slab.h>
 #include <video/edid.h>
+#include <video/videomode.h>
 #ifdef CONFIG_PPC_OF
 #include <asm/prom.h>
 #include <asm/pci-bridge.h>
@@ -1373,6 +1374,57 @@ int fb_get_mode(int flags, u32 val, struct fb_var_screeninfo *var, struct fb_inf
 	kfree(timings);
 	return err;
 }
+
+#if IS_ENABLED(CONFIG_VIDEOMODE)
+int fb_videomode_from_videomode(const struct videomode *vm,
+				struct fb_videomode *fbmode)
+{
+	unsigned int htotal, vtotal;
+
+	fbmode->xres = vm->hactive;
+	fbmode->left_margin = vm->hback_porch;
+	fbmode->right_margin = vm->hfront_porch;
+	fbmode->hsync_len = vm->hsync_len;
+
+	fbmode->yres = vm->vactive;
+	fbmode->upper_margin = vm->vback_porch;
+	fbmode->lower_margin = vm->vfront_porch;
+	fbmode->vsync_len = vm->vsync_len;
+
+	/* prevent division by zero in KHZ2PICOS macro */
+	fbmode->pixclock = vm->pixelclock ?
+			KHZ2PICOS(vm->pixelclock / 1000) : 0;
+
+	fbmode->sync = 0;
+	fbmode->vmode = 0;
+	if (vm->dmt_flags & VESA_DMT_HSYNC_HIGH)
+		fbmode->sync |= FB_SYNC_HOR_HIGH_ACT;
+	if (vm->dmt_flags & VESA_DMT_HSYNC_HIGH)
+		fbmode->sync |= FB_SYNC_VERT_HIGH_ACT;
+	if (vm->data_flags & DISPLAY_FLAGS_INTERLACED)
+		fbmode->vmode |= FB_VMODE_INTERLACED;
+	if (vm->data_flags & DISPLAY_FLAGS_DOUBLESCAN)
+		fbmode->vmode |= FB_VMODE_DOUBLE;
+	fbmode->flag = 0;
+
+	htotal = vm->hactive + vm->hfront_porch + vm->hback_porch +
+		 vm->hsync_len;
+	vtotal = vm->vactive + vm->vfront_porch + vm->vback_porch +
+		 vm->vsync_len;
+	/* prevent division by zero */
+	if (htotal && vtotal) {
+		fbmode->refresh = vm->pixelclock / (htotal * vtotal);
+	/* a mode must have htotal and vtotal != 0 or it is invalid */
+	} else {
+		fbmode->refresh = 0;
+		return -EINVAL;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(fb_videomode_from_videomode);
+#endif
+
 #else
 int fb_parse_edid(unsigned char *edid, struct fb_var_screeninfo *var)
 {
diff --git a/include/linux/fb.h b/include/linux/fb.h
index c7a9571..100a176 100644
--- a/include/linux/fb.h
+++ b/include/linux/fb.h
@@ -19,6 +19,7 @@ struct vm_area_struct;
 struct fb_info;
 struct device;
 struct file;
+struct videomode;
 
 /* Definitions below are used in the parsed monitor specs */
 #define FB_DPMS_ACTIVE_OFF	1
@@ -714,6 +715,9 @@ extern void fb_destroy_modedb(struct fb_videomode *modedb);
 extern int fb_find_mode_cvt(struct fb_videomode *mode, int margins, int rb);
 extern unsigned char *fb_ddc_read(struct i2c_adapter *adapter);
 
+extern int fb_videomode_from_videomode(const struct videomode *vm,
+				       struct fb_videomode *fbmode);
+
 /* drivers/video/modedb.c */
 #define VESA_MODEDB_SIZE 34
 extern void fb_var_to_videomode(struct fb_videomode *mode,
-- 
cgit v0.10.2


From 790890444f591145d3a00623af461c9006e17d51 Mon Sep 17 00:00:00 2001
From: Steffen Trumtrar <s.trumtrar@pengutronix.de>
Date: Wed, 14 Nov 2012 11:12:38 +0100
Subject: fbmon: add of_videomode helpers

Add helper to get fb_videomode from devicetree.

Signed-off-by: Steffen Trumtrar <s.trumtrar@pengutronix.de>
Reviewed-by: Thierry Reding <thierry.reding@avionic-design.de>
Acked-by: Thierry Reding <thierry.reding@avionic-design.de>
Tested-by: Thierry Reding <thierry.reding@avionic-design.de>
Tested-by: Philipp Zabel <p.zabel@pengutronix.de>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-by: Afzal Mohammed <Afzal@ti.com>
Tested-by: Rob Clark <robclark@gmail.com>
Tested-by: Leela Krishna Amudala <leelakrishna.a@gmail.com>

diff --git a/drivers/video/fbmon.c b/drivers/video/fbmon.c
index 17ce135..94ad0f7 100644
--- a/drivers/video/fbmon.c
+++ b/drivers/video/fbmon.c
@@ -31,6 +31,7 @@
 #include <linux/pci.h>
 #include <linux/slab.h>
 #include <video/edid.h>
+#include <video/of_videomode.h>
 #include <video/videomode.h>
 #ifdef CONFIG_PPC_OF
 #include <asm/prom.h>
@@ -1425,6 +1426,47 @@ int fb_videomode_from_videomode(const struct videomode *vm,
 EXPORT_SYMBOL_GPL(fb_videomode_from_videomode);
 #endif
 
+#if IS_ENABLED(CONFIG_OF_VIDEOMODE)
+static inline void dump_fb_videomode(const struct fb_videomode *m)
+{
+	pr_debug("fb_videomode = %ux%u@%uHz (%ukHz) %u %u %u %u %u %u %u %u %u\n",
+		 m->xres, m->yres, m->refresh, m->pixclock, m->left_margin,
+		 m->right_margin, m->upper_margin, m->lower_margin,
+		 m->hsync_len, m->vsync_len, m->sync, m->vmode, m->flag);
+}
+
+/**
+ * of_get_fb_videomode - get a fb_videomode from devicetree
+ * @np: device_node with the timing specification
+ * @fb: will be set to the return value
+ * @index: index into the list of display timings in devicetree
+ *
+ * DESCRIPTION:
+ * This function is expensive and should only be used, if only one mode is to be
+ * read from DT. To get multiple modes start with of_get_display_timings ond
+ * work with that instead.
+ */
+int of_get_fb_videomode(struct device_node *np, struct fb_videomode *fb,
+			int index)
+{
+	struct videomode vm;
+	int ret;
+
+	ret = of_get_videomode(np, &vm, index);
+	if (ret)
+		return ret;
+
+	fb_videomode_from_videomode(&vm, fb);
+
+	pr_debug("%s: got %dx%d display mode from %s\n",
+		of_node_full_name(np), vm.hactive, vm.vactive, np->name);
+	dump_fb_videomode(fb);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(of_get_fb_videomode);
+#endif
+
 #else
 int fb_parse_edid(unsigned char *edid, struct fb_var_screeninfo *var)
 {
diff --git a/include/linux/fb.h b/include/linux/fb.h
index 100a176..58b9860 100644
--- a/include/linux/fb.h
+++ b/include/linux/fb.h
@@ -20,6 +20,7 @@ struct fb_info;
 struct device;
 struct file;
 struct videomode;
+struct device_node;
 
 /* Definitions below are used in the parsed monitor specs */
 #define FB_DPMS_ACTIVE_OFF	1
@@ -715,6 +716,9 @@ extern void fb_destroy_modedb(struct fb_videomode *modedb);
 extern int fb_find_mode_cvt(struct fb_videomode *mode, int margins, int rb);
 extern unsigned char *fb_ddc_read(struct i2c_adapter *adapter);
 
+extern int of_get_fb_videomode(struct device_node *np,
+			       struct fb_videomode *fb,
+			       int index);
 extern int fb_videomode_from_videomode(const struct videomode *vm,
 				       struct fb_videomode *fbmode);
 
-- 
cgit v0.10.2


From ebc64e453857e93b230881f48126257be9aa8830 Mon Sep 17 00:00:00 2001
From: Steffen Trumtrar <s.trumtrar@pengutronix.de>
Date: Wed, 14 Nov 2012 11:22:52 +0100
Subject: drm_modes: add videomode helpers

Add conversion from videomode to drm_display_mode

Signed-off-by: Steffen Trumtrar <s.trumtrar@pengutronix.de>
Reviewed-by: Thierry Reding <thierry.reding@avionic-design.de>
Acked-by: Thierry Reding <thierry.reding@avionic-design.de>
Tested-by: Thierry Reding <thierry.reding@avionic-design.de>
Tested-by: Philipp Zabel <p.zabel@pengutronix.de>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-by: Afzal Mohammed <Afzal@ti.com>
Tested-by: Rob Clark <robclark@gmail.com>
Tested-by: Leela Krishna Amudala <leelakrishna.a@gmail.com>

diff --git a/drivers/gpu/drm/drm_modes.c b/drivers/gpu/drm/drm_modes.c
index d8da30e..9f3f20b 100644
--- a/drivers/gpu/drm/drm_modes.c
+++ b/drivers/gpu/drm/drm_modes.c
@@ -35,6 +35,7 @@
 #include <linux/export.h>
 #include <drm/drmP.h>
 #include <drm/drm_crtc.h>
+#include <video/videomode.h>
 
 /**
  * drm_mode_debug_printmodeline - debug print a mode
@@ -504,6 +505,42 @@ drm_gtf_mode(struct drm_device *dev, int hdisplay, int vdisplay, int vrefresh,
 }
 EXPORT_SYMBOL(drm_gtf_mode);
 
+#if IS_ENABLED(CONFIG_VIDEOMODE)
+int drm_display_mode_from_videomode(const struct videomode *vm,
+				    struct drm_display_mode *dmode)
+{
+	dmode->hdisplay = vm->hactive;
+	dmode->hsync_start = dmode->hdisplay + vm->hfront_porch;
+	dmode->hsync_end = dmode->hsync_start + vm->hsync_len;
+	dmode->htotal = dmode->hsync_end + vm->hback_porch;
+
+	dmode->vdisplay = vm->vactive;
+	dmode->vsync_start = dmode->vdisplay + vm->vfront_porch;
+	dmode->vsync_end = dmode->vsync_start + vm->vsync_len;
+	dmode->vtotal = dmode->vsync_end + vm->vback_porch;
+
+	dmode->clock = vm->pixelclock / 1000;
+
+	dmode->flags = 0;
+	if (vm->dmt_flags & VESA_DMT_HSYNC_HIGH)
+		dmode->flags |= DRM_MODE_FLAG_PHSYNC;
+	else if (vm->dmt_flags & VESA_DMT_HSYNC_LOW)
+		dmode->flags |= DRM_MODE_FLAG_NHSYNC;
+	if (vm->dmt_flags & VESA_DMT_VSYNC_HIGH)
+		dmode->flags |= DRM_MODE_FLAG_PVSYNC;
+	else if (vm->dmt_flags & VESA_DMT_VSYNC_LOW)
+		dmode->flags |= DRM_MODE_FLAG_NVSYNC;
+	if (vm->data_flags & DISPLAY_FLAGS_INTERLACED)
+		dmode->flags |= DRM_MODE_FLAG_INTERLACE;
+	if (vm->data_flags & DISPLAY_FLAGS_DOUBLESCAN)
+		dmode->flags |= DRM_MODE_FLAG_DBLSCAN;
+	drm_mode_set_name(dmode);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(drm_display_mode_from_videomode);
+#endif
+
 /**
  * drm_mode_set_name - set the name on a mode
  * @mode: name will be set in this mode
diff --git a/include/drm/drmP.h b/include/drm/drmP.h
index fad21c9..d5c06ff 100644
--- a/include/drm/drmP.h
+++ b/include/drm/drmP.h
@@ -85,6 +85,8 @@ struct module;
 struct drm_file;
 struct drm_device;
 
+struct videomode;
+
 #include <drm/drm_os_linux.h>
 #include <drm/drm_hashtab.h>
 #include <drm/drm_mm.h>
@@ -1456,6 +1458,9 @@ extern struct drm_display_mode *
 drm_mode_create_from_cmdline_mode(struct drm_device *dev,
 				  struct drm_cmdline_mode *cmd);
 
+extern int drm_display_mode_from_videomode(const struct videomode *vm,
+					   struct drm_display_mode *dmode);
+
 /* Modesetting support */
 extern void drm_vblank_pre_modeset(struct drm_device *dev, int crtc);
 extern void drm_vblank_post_modeset(struct drm_device *dev, int crtc);
-- 
cgit v0.10.2


From edb37a95c58147f89713e6c5cd220fa8fdfb4833 Mon Sep 17 00:00:00 2001
From: Steffen Trumtrar <s.trumtrar@pengutronix.de>
Date: Sun, 28 Oct 2012 18:28:06 +0100
Subject: drm_modes: add of_videomode helpers

Add helper to get drm_display_mode from devicetree.

Signed-off-by: Steffen Trumtrar <s.trumtrar@pengutronix.de>
Reviewed-by: Thierry Reding <thierry.reding@avionic-design.de>
Acked-by: Thierry Reding <thierry.reding@avionic-design.de>
Tested-by: Thierry Reding <thierry.reding@avionic-design.de>
Tested-by: Philipp Zabel <p.zabel@pengutronix.de>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-by: Afzal Mohammed <Afzal@ti.com>
Tested-by: Rob Clark <robclark@gmail.com>
Tested-by: Leela Krishna Amudala <leelakrishna.a@gmail.com>

diff --git a/drivers/gpu/drm/drm_modes.c b/drivers/gpu/drm/drm_modes.c
index 9f3f20b..04fa6f1 100644
--- a/drivers/gpu/drm/drm_modes.c
+++ b/drivers/gpu/drm/drm_modes.c
@@ -35,6 +35,7 @@
 #include <linux/export.h>
 #include <drm/drmP.h>
 #include <drm/drm_crtc.h>
+#include <video/of_videomode.h>
 #include <video/videomode.h>
 
 /**
@@ -541,6 +542,38 @@ int drm_display_mode_from_videomode(const struct videomode *vm,
 EXPORT_SYMBOL_GPL(drm_display_mode_from_videomode);
 #endif
 
+#if IS_ENABLED(CONFIG_OF_VIDEOMODE)
+/**
+ * of_get_drm_display_mode - get a drm_display_mode from devicetree
+ * @np: device_node with the timing specification
+ * @dmode: will be set to the return value
+ * @index: index into the list of display timings in devicetree
+ *
+ * This function is expensive and should only be used, if only one mode is to be
+ * read from DT. To get multiple modes start with of_get_display_timings and
+ * work with that instead.
+ */
+int of_get_drm_display_mode(struct device_node *np,
+			    struct drm_display_mode *dmode, int index)
+{
+	struct videomode vm;
+	int ret;
+
+	ret = of_get_videomode(np, &vm, index);
+	if (ret)
+		return ret;
+
+	drm_display_mode_from_videomode(&vm, dmode);
+
+	pr_debug("%s: got %dx%d display mode from %s\n",
+		of_node_full_name(np), vm.hactive, vm.vactive, np->name);
+	drm_mode_debug_printmodeline(dmode);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(of_get_drm_display_mode);
+#endif
+
 /**
  * drm_mode_set_name - set the name on a mode
  * @mode: name will be set in this mode
diff --git a/include/drm/drmP.h b/include/drm/drmP.h
index d5c06ff..fcc9d23 100644
--- a/include/drm/drmP.h
+++ b/include/drm/drmP.h
@@ -85,6 +85,7 @@ struct module;
 struct drm_file;
 struct drm_device;
 
+struct device_node;
 struct videomode;
 
 #include <drm/drm_os_linux.h>
@@ -1460,6 +1461,9 @@ drm_mode_create_from_cmdline_mode(struct drm_device *dev,
 
 extern int drm_display_mode_from_videomode(const struct videomode *vm,
 					   struct drm_display_mode *dmode);
+extern int of_get_drm_display_mode(struct device_node *np,
+				   struct drm_display_mode *dmode,
+				   int index);
 
 /* Modesetting support */
 extern void drm_vblank_pre_modeset(struct drm_device *dev, int crtc);
-- 
cgit v0.10.2


From 8de0add7231f5747f192c1ed35fd792a866c885b Mon Sep 17 00:00:00 2001
From: Paulo Zanoni <paulo.r.zanoni@intel.com>
Date: Fri, 18 Jan 2013 18:29:03 -0200
Subject: drm/i915: don't save/restore DSPARB on gen5+

Because the register does not exist in gen5+.

This patch solves "unclaimed register" messages on Haswell after
suspend/resume.

Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_suspend.c b/drivers/gpu/drm/i915/i915_suspend.c
index 63d4d30..95c0582 100644
--- a/drivers/gpu/drm/i915/i915_suspend.c
+++ b/drivers/gpu/drm/i915/i915_suspend.c
@@ -620,7 +620,8 @@ static void i915_save_display(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
 	/* Display arbitration control */
-	dev_priv->regfile.saveDSPARB = I915_READ(DSPARB);
+	if (INTEL_INFO(dev)->gen <= 4)
+		dev_priv->regfile.saveDSPARB = I915_READ(DSPARB);
 
 	/* This is only meaningful in non-KMS mode */
 	/* Don't regfile.save them in KMS mode */
@@ -707,7 +708,8 @@ static void i915_restore_display(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
 	/* Display arbitration */
-	I915_WRITE(DSPARB, dev_priv->regfile.saveDSPARB);
+	if (INTEL_INFO(dev)->gen <= 4)
+		I915_WRITE(DSPARB, dev_priv->regfile.saveDSPARB);
 
 	if (!drm_core_check_feature(dev, DRIVER_MODESET)) {
 		/* Display port ratios (must be done before clock is set) */
-- 
cgit v0.10.2


From 69fde0a6109eb04d4a32b6a205cff48e9058f16e Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 24 Jan 2013 15:29:26 +0200
Subject: drm/i915: Convert intel_hdmi to enum port
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Use intel_dig_port->port rather than intel_hdmi->sdvox_erg.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c
index d53b731..5b4efd6 100644
--- a/drivers/gpu/drm/i915/intel_hdmi.c
+++ b/drivers/gpu/drm/i915/intel_hdmi.c
@@ -372,7 +372,8 @@ static void g4x_set_infoframes(struct drm_encoder *encoder,
 			       struct drm_display_mode *adjusted_mode)
 {
 	struct drm_i915_private *dev_priv = encoder->dev->dev_private;
-	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
+	struct intel_digital_port *intel_dig_port = enc_to_dig_port(encoder);
+	struct intel_hdmi *intel_hdmi = &intel_dig_port->hdmi;
 	u32 reg = VIDEO_DIP_CTL;
 	u32 val = I915_READ(reg);
 	u32 port;
@@ -399,11 +400,11 @@ static void g4x_set_infoframes(struct drm_encoder *encoder,
 		return;
 	}
 
-	switch (intel_hdmi->sdvox_reg) {
-	case SDVOB:
+	switch (intel_dig_port->port) {
+	case PORT_B:
 		port = VIDEO_DIP_PORT_B;
 		break;
-	case SDVOC:
+	case PORT_C:
 		port = VIDEO_DIP_PORT_C;
 		break;
 	default:
@@ -436,7 +437,8 @@ static void ibx_set_infoframes(struct drm_encoder *encoder,
 {
 	struct drm_i915_private *dev_priv = encoder->dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->crtc);
-	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
+	struct intel_digital_port *intel_dig_port = enc_to_dig_port(encoder);
+	struct intel_hdmi *intel_hdmi = &intel_dig_port->hdmi;
 	u32 reg = TVIDEO_DIP_CTL(intel_crtc->pipe);
 	u32 val = I915_READ(reg);
 	u32 port;
@@ -455,14 +457,14 @@ static void ibx_set_infoframes(struct drm_encoder *encoder,
 		return;
 	}
 
-	switch (intel_hdmi->sdvox_reg) {
-	case HDMIB:
+	switch (intel_dig_port->port) {
+	case PORT_B:
 		port = VIDEO_DIP_PORT_B;
 		break;
-	case HDMIC:
+	case PORT_C:
 		port = VIDEO_DIP_PORT_C;
 		break;
-	case HDMID:
+	case PORT_D:
 		port = VIDEO_DIP_PORT_D;
 		break;
 	default:
@@ -795,13 +797,14 @@ static bool g4x_hdmi_connected(struct intel_hdmi *intel_hdmi)
 {
 	struct drm_device *dev = intel_hdmi_to_dev(intel_hdmi);
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_digital_port *intel_dig_port = hdmi_to_dig_port(intel_hdmi);
 	uint32_t bit;
 
-	switch (intel_hdmi->sdvox_reg) {
-	case SDVOB:
+	switch (intel_dig_port->port) {
+	case PORT_B:
 		bit = HDMIB_HOTPLUG_LIVE_STATUS;
 		break;
-	case SDVOC:
+	case PORT_C:
 		bit = HDMIC_HOTPLUG_LIVE_STATUS;
 		break;
 	default:
-- 
cgit v0.10.2


From 34f2be46c4a4db47aafa1c9e79320723dda7b5e4 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 24 Jan 2013 15:29:27 +0200
Subject: drm/i915: Convert intel_dp to enum port
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Use intel_dig_port->port rather than intel_dp->output_reg.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 1492706..51fd797 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -2291,16 +2291,17 @@ g4x_dp_detect(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	uint32_t bit;
 
-	switch (intel_dp->output_reg) {
-	case DP_B:
+	switch (intel_dig_port->port) {
+	case PORT_B:
 		bit = DPB_HOTPLUG_LIVE_STATUS;
 		break;
-	case DP_C:
+	case PORT_C:
 		bit = DPC_HOTPLUG_LIVE_STATUS;
 		break;
-	case DP_D:
+	case PORT_D:
 		bit = DPD_HOTPLUG_LIVE_STATUS;
 		break;
 	default:
-- 
cgit v0.10.2


From 10fce67a971775cfa906eb176e2772938c2aaaf3 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 24 Jan 2013 15:29:28 +0200
Subject: drm/i915: Add display_display_mmio_offset to intel_device_info
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Add an optional offset to intel_device_info, which will added
to most display register offsets.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 56ece50..09753e5 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -337,6 +337,7 @@ struct drm_i915_gt_funcs {
 	DEV_INFO_FLAG(has_llc)
 
 struct intel_device_info {
+	u32 display_mmio_offset;
 	u8 gen;
 	u8 is_mobile:1;
 	u8 is_i85x:1;
-- 
cgit v0.10.2


From f4ba9f81715ed0ba000750bf8586d097077ce910 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 24 Jan 2013 15:29:29 +0200
Subject: drm/i915: AUD_VID_DID needs an offset on VLV
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 2521617..7dd7adb 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -4298,7 +4298,7 @@
 #define GEN7_ROW_CHICKEN2_GT2		0xf4f4
 #define   DOP_CLOCK_GATING_DISABLE	(1<<0)
 
-#define G4X_AUD_VID_DID			0x62020
+#define G4X_AUD_VID_DID			(dev_priv->info->display_mmio_offset + 0x62020)
 #define INTEL_AUDIO_DEVCL		0x808629FB
 #define INTEL_AUDIO_DEVBLC		0x80862801
 #define INTEL_AUDIO_DEVCTG		0x80862802
-- 
cgit v0.10.2


From f12c47b279ad851f86026b2b142a81f6a63772b9 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 24 Jan 2013 15:29:30 +0200
Subject: drm/i915: Per-pipe PP registers are for VLV only
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 7dd7adb..425ea99 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -3996,17 +3996,17 @@
 #define  LVDS_DETECTED	(1 << 1)
 
 /* vlv has 2 sets of panel control regs. */
-#define PIPEA_PP_STATUS         0x61200
-#define PIPEA_PP_CONTROL        0x61204
-#define PIPEA_PP_ON_DELAYS      0x61208
-#define PIPEA_PP_OFF_DELAYS     0x6120c
-#define PIPEA_PP_DIVISOR        0x61210
-
-#define PIPEB_PP_STATUS         0x61300
-#define PIPEB_PP_CONTROL        0x61304
-#define PIPEB_PP_ON_DELAYS      0x61308
-#define PIPEB_PP_OFF_DELAYS     0x6130c
-#define PIPEB_PP_DIVISOR        0x61310
+#define PIPEA_PP_STATUS         (VLV_DISPLAY_BASE + 0x61200)
+#define PIPEA_PP_CONTROL        (VLV_DISPLAY_BASE + 0x61204)
+#define PIPEA_PP_ON_DELAYS      (VLV_DISPLAY_BASE + 0x61208)
+#define PIPEA_PP_OFF_DELAYS     (VLV_DISPLAY_BASE + 0x6120c)
+#define PIPEA_PP_DIVISOR        (VLV_DISPLAY_BASE + 0x61210)
+
+#define PIPEB_PP_STATUS         (VLV_DISPLAY_BASE + 0x61300)
+#define PIPEB_PP_CONTROL        (VLV_DISPLAY_BASE + 0x61304)
+#define PIPEB_PP_ON_DELAYS      (VLV_DISPLAY_BASE + 0x61308)
+#define PIPEB_PP_OFF_DELAYS     (VLV_DISPLAY_BASE + 0x6130c)
+#define PIPEB_PP_DIVISOR        (VLV_DISPLAY_BASE + 0x61310)
 
 #define PCH_PP_STATUS		0xc7200
 #define PCH_PP_CONTROL		0xc7204
-- 
cgit v0.10.2


From b906487c515711d79e31ceb9fe75738574b5c3a2 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 24 Jan 2013 15:29:31 +0200
Subject: drm/i915: VLV_VIDEO_DIP_CTL is for VLV only
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 425ea99..b97c67e 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -3700,13 +3700,13 @@
 #define TVIDEO_DIP_DATA(pipe) _PIPE(pipe, _VIDEO_DIP_DATA_A, _VIDEO_DIP_DATA_B)
 #define TVIDEO_DIP_GCP(pipe) _PIPE(pipe, _VIDEO_DIP_GCP_A, _VIDEO_DIP_GCP_B)
 
-#define VLV_VIDEO_DIP_CTL_A		0x60200
-#define VLV_VIDEO_DIP_DATA_A		0x60208
-#define VLV_VIDEO_DIP_GDCP_PAYLOAD_A	0x60210
+#define VLV_VIDEO_DIP_CTL_A		(VLV_DISPLAY_BASE + 0x60200)
+#define VLV_VIDEO_DIP_DATA_A		(VLV_DISPLAY_BASE + 0x60208)
+#define VLV_VIDEO_DIP_GDCP_PAYLOAD_A	(VLV_DISPLAY_BASE + 0x60210)
 
-#define VLV_VIDEO_DIP_CTL_B		0x61170
-#define VLV_VIDEO_DIP_DATA_B		0x61174
-#define VLV_VIDEO_DIP_GDCP_PAYLOAD_B	0x61178
+#define VLV_VIDEO_DIP_CTL_B		(VLV_DISPLAY_BASE + 0x61170)
+#define VLV_VIDEO_DIP_DATA_B		(VLV_DISPLAY_BASE + 0x61174)
+#define VLV_VIDEO_DIP_GDCP_PAYLOAD_B	(VLV_DISPLAY_BASE + 0x61178)
 
 #define VLV_TVIDEO_DIP_CTL(pipe) \
 	 _PIPE(pipe, VLV_VIDEO_DIP_CTL_A, VLV_VIDEO_DIP_CTL_B)
-- 
cgit v0.10.2


From aab17139a0eec686bad80579f7a5d6969c998392 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 24 Jan 2013 15:29:32 +0200
Subject: drm/i915: PIPE M/N registers need an offset on VLV
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index b97c67e..0132447 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -3269,41 +3269,41 @@
 #define  FDI_PLL_FREQ_DISABLE_COUNT_LIMIT_MASK  0xff
 
 
-#define _PIPEA_DATA_M1           0x60030
+#define _PIPEA_DATA_M1           (dev_priv->info->display_mmio_offset + 0x60030)
 #define  TU_SIZE(x)             (((x)-1) << 25) /* default size 64 */
 #define  TU_SIZE_MASK           0x7e000000
 #define  PIPE_DATA_M1_OFFSET    0
-#define _PIPEA_DATA_N1           0x60034
+#define _PIPEA_DATA_N1           (dev_priv->info->display_mmio_offset + 0x60034)
 #define  PIPE_DATA_N1_OFFSET    0
 
-#define _PIPEA_DATA_M2           0x60038
+#define _PIPEA_DATA_M2           (dev_priv->info->display_mmio_offset + 0x60038)
 #define  PIPE_DATA_M2_OFFSET    0
-#define _PIPEA_DATA_N2           0x6003c
+#define _PIPEA_DATA_N2           (dev_priv->info->display_mmio_offset + 0x6003c)
 #define  PIPE_DATA_N2_OFFSET    0
 
-#define _PIPEA_LINK_M1           0x60040
+#define _PIPEA_LINK_M1           (dev_priv->info->display_mmio_offset + 0x60040)
 #define  PIPE_LINK_M1_OFFSET    0
-#define _PIPEA_LINK_N1           0x60044
+#define _PIPEA_LINK_N1           (dev_priv->info->display_mmio_offset + 0x60044)
 #define  PIPE_LINK_N1_OFFSET    0
 
-#define _PIPEA_LINK_M2           0x60048
+#define _PIPEA_LINK_M2           (dev_priv->info->display_mmio_offset + 0x60048)
 #define  PIPE_LINK_M2_OFFSET    0
-#define _PIPEA_LINK_N2           0x6004c
+#define _PIPEA_LINK_N2           (dev_priv->info->display_mmio_offset + 0x6004c)
 #define  PIPE_LINK_N2_OFFSET    0
 
 /* PIPEB timing regs are same start from 0x61000 */
 
-#define _PIPEB_DATA_M1           0x61030
-#define _PIPEB_DATA_N1           0x61034
+#define _PIPEB_DATA_M1           (dev_priv->info->display_mmio_offset + 0x61030)
+#define _PIPEB_DATA_N1           (dev_priv->info->display_mmio_offset + 0x61034)
 
-#define _PIPEB_DATA_M2           0x61038
-#define _PIPEB_DATA_N2           0x6103c
+#define _PIPEB_DATA_M2           (dev_priv->info->display_mmio_offset + 0x61038)
+#define _PIPEB_DATA_N2           (dev_priv->info->display_mmio_offset + 0x6103c)
 
-#define _PIPEB_LINK_M1           0x61040
-#define _PIPEB_LINK_N1           0x61044
+#define _PIPEB_LINK_M1           (dev_priv->info->display_mmio_offset + 0x61040)
+#define _PIPEB_LINK_N1           (dev_priv->info->display_mmio_offset + 0x61044)
 
-#define _PIPEB_LINK_M2           0x61048
-#define _PIPEB_LINK_N2           0x6104c
+#define _PIPEB_LINK_M2           (dev_priv->info->display_mmio_offset + 0x61048)
+#define _PIPEB_LINK_N2           (dev_priv->info->display_mmio_offset + 0x6104c)
 
 #define PIPE_DATA_M1(tran) _TRANSCODER(tran, _PIPEA_DATA_M1, _PIPEB_DATA_M1)
 #define PIPE_DATA_N1(tran) _TRANSCODER(tran, _PIPEA_DATA_N1, _PIPEB_DATA_N1)
-- 
cgit v0.10.2


From 895abf0c3c6923c4203630d2288bb1fbe39c39db Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 24 Jan 2013 15:29:35 +0200
Subject: drm/i915: Primary plane registers need an offset on VLV
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 0132447..37589cd 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -2982,7 +2982,7 @@
 #define CURPOS_IVB(pipe) _PIPE(pipe, _CURAPOS, _CURBPOS_IVB)
 
 /* Display A control */
-#define _DSPACNTR                0x70180
+#define _DSPACNTR                (dev_priv->info->display_mmio_offset + 0x70180)
 #define   DISPLAY_PLANE_ENABLE			(1<<31)
 #define   DISPLAY_PLANE_DISABLE			0
 #define   DISPPLANE_GAMMA_ENABLE		(1<<30)
@@ -3015,14 +3015,14 @@
 #define   DISPPLANE_STEREO_POLARITY_SECOND	(1<<18)
 #define   DISPPLANE_TRICKLE_FEED_DISABLE	(1<<14) /* Ironlake */
 #define   DISPPLANE_TILED			(1<<10)
-#define _DSPAADDR		0x70184
-#define _DSPASTRIDE		0x70188
-#define _DSPAPOS			0x7018C /* reserved */
-#define _DSPASIZE		0x70190
-#define _DSPASURF		0x7019C /* 965+ only */
-#define _DSPATILEOFF		0x701A4 /* 965+ only */
-#define _DSPAOFFSET		0x701A4 /* HSW */
-#define _DSPASURFLIVE		0x701AC
+#define _DSPAADDR		(dev_priv->info->display_mmio_offset + 0x70184)
+#define _DSPASTRIDE		(dev_priv->info->display_mmio_offset + 0x70188)
+#define _DSPAPOS		(dev_priv->info->display_mmio_offset + 0x7018C) /* reserved */
+#define _DSPASIZE		(dev_priv->info->display_mmio_offset + 0x70190)
+#define _DSPASURF		(dev_priv->info->display_mmio_offset + 0x7019C) /* 965+ only */
+#define _DSPATILEOFF		(dev_priv->info->display_mmio_offset + 0x701A4) /* 965+ only */
+#define _DSPAOFFSET		(dev_priv->info->display_mmio_offset + 0x701A4) /* HSW */
+#define _DSPASURFLIVE		(dev_priv->info->display_mmio_offset + 0x701AC)
 
 #define DSPCNTR(plane) _PIPE(plane, _DSPACNTR, _DSPBCNTR)
 #define DSPADDR(plane) _PIPE(plane, _DSPAADDR, _DSPBADDR)
@@ -3068,19 +3068,19 @@
 
 
 /* Display B control */
-#define _DSPBCNTR		0x71180
+#define _DSPBCNTR		(dev_priv->info->display_mmio_offset + 0x71180)
 #define   DISPPLANE_ALPHA_TRANS_ENABLE		(1<<15)
 #define   DISPPLANE_ALPHA_TRANS_DISABLE		0
 #define   DISPPLANE_SPRITE_ABOVE_DISPLAY	0
 #define   DISPPLANE_SPRITE_ABOVE_OVERLAY	(1)
-#define _DSPBADDR		0x71184
-#define _DSPBSTRIDE		0x71188
-#define _DSPBPOS			0x7118C
-#define _DSPBSIZE		0x71190
-#define _DSPBSURF		0x7119C
-#define _DSPBTILEOFF		0x711A4
-#define _DSPBOFFSET		0x711A4
-#define _DSPBSURFLIVE		0x711AC
+#define _DSPBADDR		(dev_priv->info->display_mmio_offset + 0x71184)
+#define _DSPBSTRIDE		(dev_priv->info->display_mmio_offset + 0x71188)
+#define _DSPBPOS		(dev_priv->info->display_mmio_offset + 0x7118C)
+#define _DSPBSIZE		(dev_priv->info->display_mmio_offset + 0x71190)
+#define _DSPBSURF		(dev_priv->info->display_mmio_offset + 0x7119C)
+#define _DSPBTILEOFF		(dev_priv->info->display_mmio_offset + 0x711A4)
+#define _DSPBOFFSET		(dev_priv->info->display_mmio_offset + 0x711A4)
+#define _DSPBSURFLIVE		(dev_priv->info->display_mmio_offset + 0x711AC)
 
 /* Sprite A control */
 #define _DVSACNTR		0x72180
-- 
cgit v0.10.2


From 0c3870ee585f94650e36918cb29c7dd6b3f3c779 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 24 Jan 2013 15:29:36 +0200
Subject: drm/i915: Pipe registers need an offset on VLV
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 37589cd..899c5ad 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -2618,10 +2618,10 @@
 /* Display & cursor control */
 
 /* Pipe A */
-#define _PIPEADSL		0x70000
+#define _PIPEADSL		(dev_priv->info->display_mmio_offset + 0x70000)
 #define   DSL_LINEMASK_GEN2	0x00000fff
 #define   DSL_LINEMASK_GEN3	0x00001fff
-#define _PIPEACONF		0x70008
+#define _PIPEACONF		(dev_priv->info->display_mmio_offset + 0x70008)
 #define   PIPECONF_ENABLE	(1<<31)
 #define   PIPECONF_DISABLE	0
 #define   PIPECONF_DOUBLE_WIDE	(1<<30)
@@ -2662,7 +2662,7 @@
 #define   PIPECONF_DITHER_TYPE_ST1 (1<<2)
 #define   PIPECONF_DITHER_TYPE_ST2 (2<<2)
 #define   PIPECONF_DITHER_TYPE_TEMP (3<<2)
-#define _PIPEASTAT		0x70024
+#define _PIPEASTAT		(dev_priv->info->display_mmio_offset + 0x70024)
 #define   PIPE_FIFO_UNDERRUN_STATUS		(1UL<<31)
 #define   SPRITE1_FLIPDONE_INT_EN_VLV		(1UL<<30)
 #define   PIPE_CRC_ERROR_ENABLE			(1UL<<29)
@@ -2923,10 +2923,10 @@
  *  } while (high1 != high2);
  *  frame = (high1 << 8) | low1;
  */
-#define _PIPEAFRAMEHIGH          0x70040
+#define _PIPEAFRAMEHIGH          (dev_priv->info->display_mmio_offset + 0x70040)
 #define   PIPE_FRAME_HIGH_MASK    0x0000ffff
 #define   PIPE_FRAME_HIGH_SHIFT   0
-#define _PIPEAFRAMEPIXEL         0x70044
+#define _PIPEAFRAMEPIXEL         (dev_priv->info->display_mmio_offset + 0x70044)
 #define   PIPE_FRAME_LOW_MASK     0xff000000
 #define   PIPE_FRAME_LOW_SHIFT    24
 #define   PIPE_PIXEL_MASK         0x00ffffff
@@ -3058,11 +3058,11 @@
 #define SWF32			0x7241c
 
 /* Pipe B */
-#define _PIPEBDSL		0x71000
-#define _PIPEBCONF		0x71008
-#define _PIPEBSTAT		0x71024
-#define _PIPEBFRAMEHIGH		0x71040
-#define _PIPEBFRAMEPIXEL		0x71044
+#define _PIPEBDSL		(dev_priv->info->display_mmio_offset + 0x71000)
+#define _PIPEBCONF		(dev_priv->info->display_mmio_offset + 0x71008)
+#define _PIPEBSTAT		(dev_priv->info->display_mmio_offset + 0x71024)
+#define _PIPEBFRAMEHIGH		(dev_priv->info->display_mmio_offset + 0x71040)
+#define _PIPEBFRAMEPIXEL	(dev_priv->info->display_mmio_offset + 0x71044)
 #define _PIPEB_FRMCOUNT_GM45	0x71040
 #define _PIPEB_FLIPCOUNT_GM45	0x71044
 
-- 
cgit v0.10.2


From 9dc33f31f2ed609eb77ddebb82f93c7cdf348879 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 24 Jan 2013 15:29:37 +0200
Subject: drm/i915: Cursor registers need an offset on VLV
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

CURSIZE is not present on VLV, so it was left out, as were the IVB
specific cursor B registers.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 899c5ad..8969468 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -2937,7 +2937,7 @@
 #define PIPE_FRMCOUNT_GM45(pipe) _PIPE(pipe, _PIPEA_FRMCOUNT_GM45, _PIPEB_FRMCOUNT_GM45)
 
 /* Cursor A & B regs */
-#define _CURACNTR		0x70080
+#define _CURACNTR		(dev_priv->info->display_mmio_offset + 0x70080)
 /* Old style CUR*CNTR flags (desktop 8xx) */
 #define   CURSOR_ENABLE		0x80000000
 #define   CURSOR_GAMMA_ENABLE	0x40000000
@@ -2958,16 +2958,16 @@
 #define   MCURSOR_PIPE_A	0x00
 #define   MCURSOR_PIPE_B	(1 << 28)
 #define   MCURSOR_GAMMA_ENABLE  (1 << 26)
-#define _CURABASE		0x70084
-#define _CURAPOS			0x70088
+#define _CURABASE		(dev_priv->info->display_mmio_offset + 0x70084)
+#define _CURAPOS		(dev_priv->info->display_mmio_offset + 0x70088)
 #define   CURSOR_POS_MASK       0x007FF
 #define   CURSOR_POS_SIGN       0x8000
 #define   CURSOR_X_SHIFT        0
 #define   CURSOR_Y_SHIFT        16
 #define CURSIZE			0x700a0
-#define _CURBCNTR		0x700c0
-#define _CURBBASE		0x700c4
-#define _CURBPOS			0x700c8
+#define _CURBCNTR		(dev_priv->info->display_mmio_offset + 0x700c0)
+#define _CURBBASE		(dev_priv->info->display_mmio_offset + 0x700c4)
+#define _CURBPOS		(dev_priv->info->display_mmio_offset + 0x700c8)
 
 #define _CURBCNTR_IVB		0x71080
 #define _CURBBASE_IVB		0x71084
-- 
cgit v0.10.2


From 8f6d8ee9f6884b8fea026b614d0475177975d066 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 24 Jan 2013 15:29:38 +0200
Subject: drm/i915: VLV_DDL is VLV only and needs an offset
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 8969468..3540a06 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -2773,13 +2773,13 @@
 /* drain latency register values*/
 #define DRAIN_LATENCY_PRECISION_32	32
 #define DRAIN_LATENCY_PRECISION_16	16
-#define VLV_DDL1			0x70050
+#define VLV_DDL1			(VLV_DISPLAY_BASE + 0x70050)
 #define DDL_CURSORA_PRECISION_32	(1<<31)
 #define DDL_CURSORA_PRECISION_16	(0<<31)
 #define DDL_CURSORA_SHIFT		24
 #define DDL_PLANEA_PRECISION_32		(1<<7)
 #define DDL_PLANEA_PRECISION_16		(0<<7)
-#define VLV_DDL2			0x70054
+#define VLV_DDL2			(VLV_DISPLAY_BASE + 0x70054)
 #define DDL_CURSORB_PRECISION_32	(1<<31)
 #define DDL_CURSORB_PRECISION_16	(0<<31)
 #define DDL_CURSORB_SHIFT		24
-- 
cgit v0.10.2


From 90f7da3fb5c9988d2ce9e87a35446f8bc9aa5997 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 24 Jan 2013 15:29:39 +0200
Subject: drm/i915: DSPFW registers need an offset on VLV
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 3540a06..4f4b989 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -2749,7 +2749,7 @@
 #define   DSPARB_BEND_SHIFT	9 /* on 855 */
 #define   DSPARB_AEND_SHIFT	0
 
-#define DSPFW1			0x70034
+#define DSPFW1			(dev_priv->info->display_mmio_offset + 0x70034)
 #define   DSPFW_SR_SHIFT	23
 #define   DSPFW_SR_MASK		(0x1ff<<23)
 #define   DSPFW_CURSORB_SHIFT	16
@@ -2757,11 +2757,11 @@
 #define   DSPFW_PLANEB_SHIFT	8
 #define   DSPFW_PLANEB_MASK	(0x7f<<8)
 #define   DSPFW_PLANEA_MASK	(0x7f)
-#define DSPFW2			0x70038
+#define DSPFW2			(dev_priv->info->display_mmio_offset + 0x70038)
 #define   DSPFW_CURSORA_MASK	0x00003f00
 #define   DSPFW_CURSORA_SHIFT	8
 #define   DSPFW_PLANEC_MASK	(0x7f)
-#define DSPFW3			0x7003c
+#define DSPFW3			(dev_priv->info->display_mmio_offset + 0x7003c)
 #define   DSPFW_HPLL_SR_EN	(1<<31)
 #define   DSPFW_CURSOR_SR_SHIFT	24
 #define   PINEVIEW_SELF_REFRESH_EN	(1<<30)
-- 
cgit v0.10.2


From b41fbda15184e0b4c4e0c0c21737e4abcbaff955 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 24 Jan 2013 15:29:41 +0200
Subject: drm/i915: DPFLIPSTAT and DPINVGTT registers are VLV only and need an
 offset
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 4f4b989..e7b13c2 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -2707,7 +2707,7 @@
 #define PIPEFRAMEPIXEL(pipe)  _PIPE(pipe, _PIPEAFRAMEPIXEL, _PIPEBFRAMEPIXEL)
 #define PIPESTAT(pipe) _PIPE(pipe, _PIPEASTAT, _PIPEBSTAT)
 
-#define VLV_DPFLIPSTAT				0x70028
+#define VLV_DPFLIPSTAT				(VLV_DISPLAY_BASE + 0x70028)
 #define   PIPEB_LINE_COMPARE_INT_EN		(1<<29)
 #define   PIPEB_HLINE_INT_EN			(1<<28)
 #define   PIPEB_VBLANK_INT_EN			(1<<27)
@@ -2721,7 +2721,7 @@
 #define   SPRITEA_FLIPDONE_INT_EN		(1<<17)
 #define   PLANEA_FLIPDONE_INT_EN		(1<<16)
 
-#define DPINVGTT				0x7002c /* VLV only */
+#define DPINVGTT				(VLV_DISPLAY_BASE + 0x7002c) /* VLV only */
 #define   CURSORB_INVALID_GTT_INT_EN		(1<<23)
 #define   CURSORA_INVALID_GTT_INT_EN		(1<<22)
 #define   SPRITED_INVALID_GTT_INT_EN		(1<<21)
-- 
cgit v0.10.2


From 7e470abf5488c17b3ca5b5abbd7db12f338cec20 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 24 Jan 2013 15:29:43 +0200
Subject: drm/i915: Panel fitter registers need an offset on VLV
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index e7b13c2..0bf70ae 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -1858,7 +1858,7 @@
 #define PP_DIVISOR	0x61210
 
 /* Panel fitting */
-#define PFIT_CONTROL	0x61230
+#define PFIT_CONTROL	(dev_priv->info->display_mmio_offset + 0x61230)
 #define   PFIT_ENABLE		(1 << 31)
 #define   PFIT_PIPE_MASK	(3 << 29)
 #define   PFIT_PIPE_SHIFT	29
@@ -1876,7 +1876,7 @@
 #define   PFIT_SCALING_PROGRAMMED (1 << 26)
 #define   PFIT_SCALING_PILLAR	(2 << 26)
 #define   PFIT_SCALING_LETTER	(3 << 26)
-#define PFIT_PGM_RATIOS	0x61234
+#define PFIT_PGM_RATIOS	(dev_priv->info->display_mmio_offset + 0x61234)
 /* Pre-965 */
 #define		PFIT_VERT_SCALE_SHIFT		20
 #define		PFIT_VERT_SCALE_MASK		0xfff00000
@@ -1888,7 +1888,7 @@
 #define		PFIT_HORIZ_SCALE_SHIFT_965	0
 #define		PFIT_HORIZ_SCALE_MASK_965	0x00001fff
 
-#define PFIT_AUTO_RATIOS 0x61238
+#define PFIT_AUTO_RATIOS (dev_priv->info->display_mmio_offset + 0x61238)
 
 /* Backlight control */
 #define BLC_PWM_CTL2		0x61250 /* 965+ only */
-- 
cgit v0.10.2


From 67d62c57465e5da7647cb13ef567b80f6deb9a3c Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 24 Jan 2013 15:29:44 +0200
Subject: drm/i915: PORT_HOTPLUG registers need an offset on VLV
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 0bf70ae..4807db6 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -1612,7 +1612,7 @@
 
 
 /* Hotplug control (945+ only) */
-#define PORT_HOTPLUG_EN		0x61110
+#define PORT_HOTPLUG_EN		(dev_priv->info->display_mmio_offset + 0x61110)
 #define   HDMIB_HOTPLUG_INT_EN			(1 << 29)
 #define   DPB_HOTPLUG_INT_EN			(1 << 29)
 #define   HDMIC_HOTPLUG_INT_EN			(1 << 28)
@@ -1639,7 +1639,7 @@
 #define CRT_HOTPLUG_DETECT_VOLTAGE_325MV	(0 << 2)
 #define CRT_HOTPLUG_DETECT_VOLTAGE_475MV	(1 << 2)
 
-#define PORT_HOTPLUG_STAT	0x61114
+#define PORT_HOTPLUG_STAT	(dev_priv->info->display_mmio_offset + 0x61114)
 /* HDMI/DP bits are gen4+ */
 #define   DPB_HOTPLUG_LIVE_STATUS               (1 << 29)
 #define   DPC_HOTPLUG_LIVE_STATUS               (1 << 28)
-- 
cgit v0.10.2


From 4e8e7eb70388c90a2d0ea2ccf951b11c3ec24b3e Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 24 Jan 2013 15:29:46 +0200
Subject: drm/i915: Pipe timing registers need an offset on VLV
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 4807db6..2982a3b 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -1532,26 +1532,26 @@
  */
 
 /* Pipe A timing regs */
-#define _HTOTAL_A	0x60000
-#define _HBLANK_A	0x60004
-#define _HSYNC_A		0x60008
-#define _VTOTAL_A	0x6000c
-#define _VBLANK_A	0x60010
-#define _VSYNC_A		0x60014
-#define _PIPEASRC	0x6001c
-#define _BCLRPAT_A	0x60020
-#define _VSYNCSHIFT_A	0x60028
+#define _HTOTAL_A	(dev_priv->info->display_mmio_offset + 0x60000)
+#define _HBLANK_A	(dev_priv->info->display_mmio_offset + 0x60004)
+#define _HSYNC_A	(dev_priv->info->display_mmio_offset + 0x60008)
+#define _VTOTAL_A	(dev_priv->info->display_mmio_offset + 0x6000c)
+#define _VBLANK_A	(dev_priv->info->display_mmio_offset + 0x60010)
+#define _VSYNC_A	(dev_priv->info->display_mmio_offset + 0x60014)
+#define _PIPEASRC	(dev_priv->info->display_mmio_offset + 0x6001c)
+#define _BCLRPAT_A	(dev_priv->info->display_mmio_offset + 0x60020)
+#define _VSYNCSHIFT_A	(dev_priv->info->display_mmio_offset + 0x60028)
 
 /* Pipe B timing regs */
-#define _HTOTAL_B	0x61000
-#define _HBLANK_B	0x61004
-#define _HSYNC_B		0x61008
-#define _VTOTAL_B	0x6100c
-#define _VBLANK_B	0x61010
-#define _VSYNC_B		0x61014
-#define _PIPEBSRC	0x6101c
-#define _BCLRPAT_B	0x61020
-#define _VSYNCSHIFT_B	0x61028
+#define _HTOTAL_B	(dev_priv->info->display_mmio_offset + 0x61000)
+#define _HBLANK_B	(dev_priv->info->display_mmio_offset + 0x61004)
+#define _HSYNC_B	(dev_priv->info->display_mmio_offset + 0x61008)
+#define _VTOTAL_B	(dev_priv->info->display_mmio_offset + 0x6100c)
+#define _VBLANK_B	(dev_priv->info->display_mmio_offset + 0x61010)
+#define _VSYNC_B	(dev_priv->info->display_mmio_offset + 0x61014)
+#define _PIPEBSRC	(dev_priv->info->display_mmio_offset + 0x6101c)
+#define _BCLRPAT_B	(dev_priv->info->display_mmio_offset + 0x61020)
+#define _VSYNCSHIFT_B	(dev_priv->info->display_mmio_offset + 0x61028)
 
 
 #define HTOTAL(trans) _TRANSCODER(trans, _HTOTAL_A, _HTOTAL_B)
-- 
cgit v0.10.2


From 4b0599854b5669fbbb6c8b9a9463b550718bda11 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 24 Jan 2013 15:29:47 +0200
Subject: drm/i915: Pipe palette registers need an offset on VLV
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 2982a3b..693baf9 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -1166,8 +1166,8 @@
  * Palette regs
  */
 
-#define _PALETTE_A		0x0a000
-#define _PALETTE_B		0x0a800
+#define _PALETTE_A		(dev_priv->info->display_mmio_offset + 0xa000)
+#define _PALETTE_B		(dev_priv->info->display_mmio_offset + 0xa800)
 #define PALETTE(pipe) _PIPE(pipe, _PALETTE_A, _PALETTE_B)
 
 /* MCH MMIO space */
-- 
cgit v0.10.2


From d88b2270862087544dd5234d95976db51d8ee5bf Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 24 Jan 2013 15:29:48 +0200
Subject: drm/i915: FB_BLC_SELF_VLV is VLV only and needs an offset
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 693baf9..86c4ffa 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -1159,7 +1159,7 @@
 #define RAMCLK_GATE_D		0x6210		/* CRL only */
 #define DEUC			0x6214          /* CRL only */
 
-#define FW_BLC_SELF_VLV		0x6500
+#define FW_BLC_SELF_VLV		(VLV_DISPLAY_BASE + 0x6500)
 #define  FW_CSPWRDWNEN		(1<<15)
 
 /*
-- 
cgit v0.10.2


From 07ec7ec55baec26a3b77d310b0399d2e20f0f67b Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 24 Jan 2013 15:29:51 +0200
Subject: drm/i915: Make VLV_GUNIT_CLOCK_GATE register value more readable
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Instead of 0x18xxxx use (VLV_DISPLAY_BASE + xxxx).

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 86c4ffa..06a37a0 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -554,7 +554,7 @@
 #define IIR		0x020a4
 #define IMR		0x020a8
 #define ISR		0x020ac
-#define VLV_GUNIT_CLOCK_GATE	0x182060
+#define VLV_GUNIT_CLOCK_GATE	(VLV_DISPLAY_BASE + 0x2060)
 #define   GCFG_DIS		(1<<8)
 #define VLV_IIR_RW	0x182084
 #define VLV_IER		0x1820a0
-- 
cgit v0.10.2


From ff7630109978a800481a97d98acab70677c05259 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 24 Jan 2013 15:29:52 +0200
Subject: drm/i915: Spell out VLV_DISPLAY_BASE for interrupt registers
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Instead of 0x18xxxx use (VLV_DISPLAY_BASE + xxxx).

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 06a37a0..f842dd9 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -556,11 +556,11 @@
 #define ISR		0x020ac
 #define VLV_GUNIT_CLOCK_GATE	(VLV_DISPLAY_BASE + 0x2060)
 #define   GCFG_DIS		(1<<8)
-#define VLV_IIR_RW	0x182084
-#define VLV_IER		0x1820a0
-#define VLV_IIR		0x1820a4
-#define VLV_IMR		0x1820a8
-#define VLV_ISR		0x1820ac
+#define VLV_IIR_RW	(VLV_DISPLAY_BASE + 0x2084)
+#define VLV_IER		(VLV_DISPLAY_BASE + 0x20a0)
+#define VLV_IIR		(VLV_DISPLAY_BASE + 0x20a4)
+#define VLV_IMR		(VLV_DISPLAY_BASE + 0x20a8)
+#define VLV_ISR		(VLV_DISPLAY_BASE + 0x20ac)
 #define   I915_PIPE_CONTROL_NOTIFY_INTERRUPT		(1<<18)
 #define   I915_DISPLAY_PORT_INTERRUPT			(1<<17)
 #define   I915_RENDER_COMMAND_PARSER_ERROR_INTERRUPT	(1<<15)
-- 
cgit v0.10.2


From 54d9d493ce08a7b106336c7393a7f7d3f2b2d9c0 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 24 Jan 2013 15:29:53 +0200
Subject: drm/i915: DPIO registers are VLV only and need an offset
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index f842dd9..94a08b9 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -336,17 +336,19 @@
  *  0x801c/3c: core clock bits
  *  0x8048/68: low pass filter coefficients
  *  0x8100: fast clock controls
+ *
+ * DPIO is VLV only.
  */
-#define DPIO_PKT			0x2100
+#define DPIO_PKT			(VLV_DISPLAY_BASE + 0x2100)
 #define  DPIO_RID			(0<<24)
 #define  DPIO_OP_WRITE			(1<<16)
 #define  DPIO_OP_READ			(0<<16)
 #define  DPIO_PORTID			(0x12<<8)
 #define  DPIO_BYTE			(0xf<<4)
 #define  DPIO_BUSY			(1<<0) /* status only */
-#define DPIO_DATA			0x2104
-#define DPIO_REG			0x2108
-#define DPIO_CTL			0x2110
+#define DPIO_DATA			(VLV_DISPLAY_BASE + 0x2104)
+#define DPIO_REG			(VLV_DISPLAY_BASE + 0x2108)
+#define DPIO_CTL			(VLV_DISPLAY_BASE + 0x2110)
 #define  DPIO_MODSEL1			(1<<3) /* if ref clk b == 27 */
 #define  DPIO_MODSEL0			(1<<2) /* if ref clk a == 27 */
 #define  DPIO_SFR_BYPASS		(1<<1)
-- 
cgit v0.10.2


From d811215004dbcc213e70eef94541aa10a2258aef Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 24 Jan 2013 15:29:55 +0200
Subject: drm/i915: GPIO/GMBUS registers need an offset on VLV
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

GPIO/GMBUS registers must be offset on VLV, so simply
adjust gpio_mmio_base to include the correct offset.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_i2c.c b/drivers/gpu/drm/i915/intel_i2c.c
index 7f09041..acf8aec 100644
--- a/drivers/gpu/drm/i915/intel_i2c.c
+++ b/drivers/gpu/drm/i915/intel_i2c.c
@@ -515,6 +515,8 @@ int intel_setup_gmbus(struct drm_device *dev)
 
 	if (HAS_PCH_SPLIT(dev))
 		dev_priv->gpio_mmio_base = PCH_GPIOA - GPIOA;
+	else if (IS_VALLEYVIEW(dev))
+		dev_priv->gpio_mmio_base = VLV_DISPLAY_BASE;
 	else
 		dev_priv->gpio_mmio_base = 0;
 
-- 
cgit v0.10.2


From fba5d532d16db812dabaa80fb7570820daa2707b Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 24 Jan 2013 15:29:56 +0200
Subject: drm/i915: Set display_mmio_offset for VLV
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

This will cause display registers to include the correct
offset on VLV.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index 3ff8e73..521a253 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -276,6 +276,7 @@ static const struct intel_device_info intel_valleyview_m_info = {
 	.has_bsd_ring = 1,
 	.has_blt_ring = 1,
 	.is_valleyview = 1,
+	.display_mmio_offset = VLV_DISPLAY_BASE,
 };
 
 static const struct intel_device_info intel_valleyview_d_info = {
@@ -285,6 +286,7 @@ static const struct intel_device_info intel_valleyview_d_info = {
 	.has_bsd_ring = 1,
 	.has_blt_ring = 1,
 	.is_valleyview = 1,
+	.display_mmio_offset = VLV_DISPLAY_BASE,
 };
 
 static const struct intel_device_info intel_haswell_d_info = {
-- 
cgit v0.10.2


From fc2de40986f5a35c02f06dea4221113b3a7a7c3c Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Fri, 25 Jan 2013 21:44:41 +0200
Subject: drm/i915: PLL registers need an offset on VLV
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

v2: Dropped the clock gating registers

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 94a08b9..452cb18 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -921,8 +921,8 @@
 #define   VGA1_PD_P1_DIV_2	(1 << 13)
 #define   VGA1_PD_P1_SHIFT	8
 #define   VGA1_PD_P1_MASK	(0x1f << 8)
-#define _DPLL_A	0x06014
-#define _DPLL_B	0x06018
+#define _DPLL_A	(dev_priv->info->display_mmio_offset + 0x6014)
+#define _DPLL_B	(dev_priv->info->display_mmio_offset + 0x6018)
 #define DPLL(pipe) _PIPE(pipe, _DPLL_A, _DPLL_B)
 #define   DPLL_VCO_ENABLE		(1 << 31)
 #define   DPLL_DVO_HIGH_SPEED		(1 << 30)
@@ -981,7 +981,7 @@
 #define   SDVO_MULTIPLIER_MASK			0x000000ff
 #define   SDVO_MULTIPLIER_SHIFT_HIRES		4
 #define   SDVO_MULTIPLIER_SHIFT_VGA		0
-#define _DPLL_A_MD 0x0601c /* 965+ only */
+#define _DPLL_A_MD (dev_priv->info->display_mmio_offset + 0x601c) /* 965+ only */
 /*
  * UDI pixel divider, controlling how many pixels are stuffed into a packet.
  *
@@ -1018,7 +1018,7 @@
  */
 #define   DPLL_MD_VGA_UDI_MULTIPLIER_MASK	0x0000003f
 #define   DPLL_MD_VGA_UDI_MULTIPLIER_SHIFT	0
-#define _DPLL_B_MD 0x06020 /* 965+ only */
+#define _DPLL_B_MD (dev_priv->info->display_mmio_offset + 0x6020) /* 965+ only */
 #define DPLL_MD(pipe) _PIPE(pipe, _DPLL_A_MD, _DPLL_B_MD)
 
 #define _FPA0	0x06040
-- 
cgit v0.10.2


From ca54b8107f6ad8c161fa102920dd70dfef19cd2e Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Fri, 25 Jan 2013 21:44:42 +0200
Subject: drm/i915: Always use adpa_reg
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Instead of using ADPA/VLV_ADPA/PCH_ADPA in various parts of
intel_crt code, just use adpa_reg which always contains the
correct value for the platform.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_crt.c b/drivers/gpu/drm/i915/intel_crt.c
index 71a5eba..68e79f3 100644
--- a/drivers/gpu/drm/i915/intel_crt.c
+++ b/drivers/gpu/drm/i915/intel_crt.c
@@ -267,27 +267,27 @@ static bool intel_ironlake_crt_detect_hotplug(struct drm_connector *connector)
 
 		crt->force_hotplug_required = 0;
 
-		save_adpa = adpa = I915_READ(PCH_ADPA);
+		save_adpa = adpa = I915_READ(crt->adpa_reg);
 		DRM_DEBUG_KMS("trigger hotplug detect cycle: adpa=0x%x\n", adpa);
 
 		adpa |= ADPA_CRT_HOTPLUG_FORCE_TRIGGER;
 		if (turn_off_dac)
 			adpa &= ~ADPA_DAC_ENABLE;
 
-		I915_WRITE(PCH_ADPA, adpa);
+		I915_WRITE(crt->adpa_reg, adpa);
 
-		if (wait_for((I915_READ(PCH_ADPA) & ADPA_CRT_HOTPLUG_FORCE_TRIGGER) == 0,
+		if (wait_for((I915_READ(crt->adpa_reg) & ADPA_CRT_HOTPLUG_FORCE_TRIGGER) == 0,
 			     1000))
 			DRM_DEBUG_KMS("timed out waiting for FORCE_TRIGGER");
 
 		if (turn_off_dac) {
-			I915_WRITE(PCH_ADPA, save_adpa);
-			POSTING_READ(PCH_ADPA);
+			I915_WRITE(crt->adpa_reg, save_adpa);
+			POSTING_READ(crt->adpa_reg);
 		}
 	}
 
 	/* Check the status to see if both blue and green are on now */
-	adpa = I915_READ(PCH_ADPA);
+	adpa = I915_READ(crt->adpa_reg);
 	if ((adpa & ADPA_CRT_HOTPLUG_MONITOR_MASK) != 0)
 		ret = true;
 	else
@@ -300,26 +300,27 @@ static bool intel_ironlake_crt_detect_hotplug(struct drm_connector *connector)
 static bool valleyview_crt_detect_hotplug(struct drm_connector *connector)
 {
 	struct drm_device *dev = connector->dev;
+	struct intel_crt *crt = intel_attached_crt(connector);
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 adpa;
 	bool ret;
 	u32 save_adpa;
 
-	save_adpa = adpa = I915_READ(ADPA);
+	save_adpa = adpa = I915_READ(crt->adpa_reg);
 	DRM_DEBUG_KMS("trigger hotplug detect cycle: adpa=0x%x\n", adpa);
 
 	adpa |= ADPA_CRT_HOTPLUG_FORCE_TRIGGER;
 
-	I915_WRITE(ADPA, adpa);
+	I915_WRITE(crt->adpa_reg, adpa);
 
-	if (wait_for((I915_READ(ADPA) & ADPA_CRT_HOTPLUG_FORCE_TRIGGER) == 0,
+	if (wait_for((I915_READ(crt->adpa_reg) & ADPA_CRT_HOTPLUG_FORCE_TRIGGER) == 0,
 		     1000)) {
 		DRM_DEBUG_KMS("timed out waiting for FORCE_TRIGGER");
-		I915_WRITE(ADPA, save_adpa);
+		I915_WRITE(crt->adpa_reg, save_adpa);
 	}
 
 	/* Check the status to see if both blue and green are on now */
-	adpa = I915_READ(ADPA);
+	adpa = I915_READ(crt->adpa_reg);
 	if ((adpa & ADPA_CRT_HOTPLUG_MONITOR_MASK) != 0)
 		ret = true;
 	else
@@ -665,11 +666,11 @@ static void intel_crt_reset(struct drm_connector *connector)
 	if (HAS_PCH_SPLIT(dev)) {
 		u32 adpa;
 
-		adpa = I915_READ(PCH_ADPA);
+		adpa = I915_READ(crt->adpa_reg);
 		adpa &= ~ADPA_CRT_HOTPLUG_MASK;
 		adpa |= ADPA_HOTPLUG_BITS;
-		I915_WRITE(PCH_ADPA, adpa);
-		POSTING_READ(PCH_ADPA);
+		I915_WRITE(crt->adpa_reg, adpa);
+		POSTING_READ(crt->adpa_reg);
 
 		DRM_DEBUG_KMS("pch crt adpa set to 0x%x\n", adpa);
 		crt->force_hotplug_required = 1;
-- 
cgit v0.10.2


From 9d5f78fbbf924492001ac7b2915537968906e517 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Fri, 25 Jan 2013 21:44:43 +0200
Subject: drm/i915: VLV doesn't have SDVO
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Don't call intel_sdvo_init() for VLV.

Preserve the same behaviour as when intel_sdvo_init() would
have returned false.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 886124a..4c4fbc5 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -8227,18 +8227,13 @@ static void intel_setup_outputs(struct drm_device *dev)
 		if (I915_READ(PCH_DP_D) & DP_DETECTED)
 			intel_dp_init(dev, PCH_DP_D, PORT_D);
 	} else if (IS_VALLEYVIEW(dev)) {
-		int found;
-
 		/* Check for built-in panel first. Shares lanes with HDMI on SDVOC */
 		if (I915_READ(DP_C) & DP_DETECTED)
 			intel_dp_init(dev, DP_C, PORT_C);
 
 		if (I915_READ(SDVOB) & PORT_DETECTED) {
-			/* SDVOB multiplex with HDMIB */
-			found = intel_sdvo_init(dev, SDVOB, true);
-			if (!found)
-				intel_hdmi_init(dev, SDVOB, PORT_B);
-			if (!found && (I915_READ(DP_B) & DP_DETECTED))
+			intel_hdmi_init(dev, SDVOB, PORT_B);
+			if (I915_READ(DP_B) & DP_DETECTED)
 				intel_dp_init(dev, DP_B, PORT_B);
 		}
 
-- 
cgit v0.10.2


From 67cfc2032b516e44b972abe30209234343e1f145 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Fri, 25 Jan 2013 21:44:44 +0200
Subject: drm/i915: Pass VLV_DISPLAY_BASE + reg to intel_{hdmi, dp}_init on VLV
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

When passing the DP/HDMI/SDVO registers to the encoder init functions,
include the VLV specific offset in the value.

v2: Resolved conflicts w/ VLV SDVO elimination

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 4c4fbc5..c5997d2 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -8228,17 +8228,17 @@ static void intel_setup_outputs(struct drm_device *dev)
 			intel_dp_init(dev, PCH_DP_D, PORT_D);
 	} else if (IS_VALLEYVIEW(dev)) {
 		/* Check for built-in panel first. Shares lanes with HDMI on SDVOC */
-		if (I915_READ(DP_C) & DP_DETECTED)
-			intel_dp_init(dev, DP_C, PORT_C);
+		if (I915_READ(VLV_DISPLAY_BASE + DP_C) & DP_DETECTED)
+			intel_dp_init(dev, VLV_DISPLAY_BASE + DP_C, PORT_C);
 
-		if (I915_READ(SDVOB) & PORT_DETECTED) {
-			intel_hdmi_init(dev, SDVOB, PORT_B);
-			if (I915_READ(DP_B) & DP_DETECTED)
-				intel_dp_init(dev, DP_B, PORT_B);
+		if (I915_READ(VLV_DISPLAY_BASE + SDVOB) & PORT_DETECTED) {
+			intel_hdmi_init(dev, VLV_DISPLAY_BASE + SDVOB, PORT_B);
+			if (I915_READ(VLV_DISPLAY_BASE + DP_B) & DP_DETECTED)
+				intel_dp_init(dev, VLV_DISPLAY_BASE + DP_B, PORT_B);
 		}
 
-		if (I915_READ(SDVOC) & PORT_DETECTED)
-			intel_hdmi_init(dev, SDVOC, PORT_C);
+		if (I915_READ(VLV_DISPLAY_BASE + SDVOC) & PORT_DETECTED)
+			intel_hdmi_init(dev, VLV_DISPLAY_BASE + SDVOC, PORT_C);
 
 	} else if (SUPPORTS_DIGITAL_OUTPUTS(dev)) {
 		bool found = false;
-- 
cgit v0.10.2


From 56a12a509296c87d6f149be86c6694d312b21d35 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Fri, 25 Jan 2013 21:44:45 +0200
Subject: drm/i915: Include display_mmio_offset in sequencer index/data
 registers
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

SR01 needs to be touched to disable VGA on non-UMS setups too.
So the sequencer registers need to include the appripriate offset
on VLV.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 452cb18..5944b4b6 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -141,9 +141,15 @@
 #define   VGA_MSR_MEM_EN (1<<1)
 #define   VGA_MSR_CGA_MODE (1<<0)
 
-#define VGA_SR_INDEX 0x3c4
+/*
+ * SR01 is the only VGA register touched on non-UMS setups.
+ * VLV doesn't do UMS, so the sequencer index/data registers
+ * are the only VGA registers which need to include
+ * display_mmio_offset.
+ */
+#define VGA_SR_INDEX (dev_priv->info->display_mmio_offset + 0x3c4)
 #define SR01			1
-#define VGA_SR_DATA 0x3c5
+#define VGA_SR_DATA (dev_priv->info->display_mmio_offset + 0x3c5)
 
 #define VGA_AR_INDEX 0x3c0
 #define   VGA_AR_VID_EN (1<<5)
-- 
cgit v0.10.2


From 80a75f7c44c8fd8cf6b271dc20801846bd630349 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 24 Jan 2013 15:29:33 +0200
Subject: drm/i915: SWF screatch registers need an offset on VLV
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 5944b4b6..215443c 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -3051,19 +3051,19 @@
 		(I915_WRITE((reg), (gfx_addr) | I915_LO_DISPBASE(I915_READ(reg))))
 
 /* VBIOS flags */
-#define SWF00			0x71410
-#define SWF01			0x71414
-#define SWF02			0x71418
-#define SWF03			0x7141c
-#define SWF04			0x71420
-#define SWF05			0x71424
-#define SWF06			0x71428
-#define SWF10			0x70410
-#define SWF11			0x70414
-#define SWF14			0x71420
-#define SWF30			0x72414
-#define SWF31			0x72418
-#define SWF32			0x7241c
+#define SWF00			(dev_priv->info->display_mmio_offset + 0x71410)
+#define SWF01			(dev_priv->info->display_mmio_offset + 0x71414)
+#define SWF02			(dev_priv->info->display_mmio_offset + 0x71418)
+#define SWF03			(dev_priv->info->display_mmio_offset + 0x7141c)
+#define SWF04			(dev_priv->info->display_mmio_offset + 0x71420)
+#define SWF05			(dev_priv->info->display_mmio_offset + 0x71424)
+#define SWF06			(dev_priv->info->display_mmio_offset + 0x71428)
+#define SWF10			(dev_priv->info->display_mmio_offset + 0x70410)
+#define SWF11			(dev_priv->info->display_mmio_offset + 0x70414)
+#define SWF14			(dev_priv->info->display_mmio_offset + 0x71420)
+#define SWF30			(dev_priv->info->display_mmio_offset + 0x72414)
+#define SWF31			(dev_priv->info->display_mmio_offset + 0x72418)
+#define SWF32			(dev_priv->info->display_mmio_offset + 0x7241c)
 
 /* Pipe B */
 #define _PIPEBDSL		(dev_priv->info->display_mmio_offset + 0x71000)
-- 
cgit v0.10.2


From fa42e23c1055a4549c433ac56ff869e441c973c8 Mon Sep 17 00:00:00 2001
From: Paulo Zanoni <paulo.r.zanoni@intel.com>
Date: Fri, 25 Jan 2013 16:59:11 -0200
Subject: drm/i915: fix intel_init_power_wells

The current code was wrong in many different ways, so this is a full
rewrite. We don't have "different power wells for different parts of
the GPU", we have a single power well, but we have multiple registers
that can be used to request enabling/disabling the power well. So
let's be a good citizen and only use the register we're suppose to
use, except when we're loading the driver, where we clear the request
made by the BIOS.

If any of the registers is requesting the power well to be enabled, it
will be enabled. If none of the registers is requesting the power well
to be enabled, it will be disabled.

For now we're just forcing the power well to be enabled, but in the
next commits we'll change this.

V2:
  - Remove debug messages that could be misleading due to possible
    race conditions with KVMr, Debug and BIOS.
  - Don't wait on disabling: after a conversaion with a hardware
    engineer we discovered that the "restriction" on bit 31 is just
    for the "enable" case, and we don't even need to wait on the
    "disable" case.

Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 215443c..a18218c 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -4422,10 +4422,10 @@
 #define   AUDIO_CP_READY_C		(1<<9)
 
 /* HSW Power Wells */
-#define HSW_PWR_WELL_CTL1			0x45400 /* BIOS */
-#define HSW_PWR_WELL_CTL2			0x45404 /* Driver */
-#define HSW_PWR_WELL_CTL3			0x45408 /* KVMR */
-#define HSW_PWR_WELL_CTL4			0x4540C /* Debug */
+#define HSW_PWR_WELL_BIOS			0x45400 /* CTL1 */
+#define HSW_PWR_WELL_DRIVER			0x45404 /* CTL2 */
+#define HSW_PWR_WELL_KVMR			0x45408 /* CTL3 */
+#define HSW_PWR_WELL_DEBUG			0x4540C /* CTL4 */
 #define   HSW_PWR_WELL_ENABLE			(1<<31)
 #define   HSW_PWR_WELL_STATE			(1<<30)
 #define HSW_PWR_WELL_CTL5			0x45410
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index c5997d2..19ab75f 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -8659,10 +8659,7 @@ static void i915_disable_vga(struct drm_device *dev)
 
 void intel_modeset_init_hw(struct drm_device *dev)
 {
-	/* We attempt to init the necessary power wells early in the initialization
-	 * time, so the subsystems that expect power to be enabled can work.
-	 */
-	intel_init_power_wells(dev);
+	intel_init_power_well(dev);
 
 	intel_prepare_ddi(dev);
 
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 66619d8..32c3042 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -667,7 +667,7 @@ extern void intel_update_fbc(struct drm_device *dev);
 extern void intel_gpu_ips_init(struct drm_i915_private *dev_priv);
 extern void intel_gpu_ips_teardown(void);
 
-extern void intel_init_power_wells(struct drm_device *dev);
+extern void intel_init_power_well(struct drm_device *dev);
 extern void intel_enable_gt_powersave(struct drm_device *dev);
 extern void intel_disable_gt_powersave(struct drm_device *dev);
 extern void gen6_gt_check_fifodbg(struct drm_i915_private *dev_priv);
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index 5a8a72c..b8cf16c 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -4043,33 +4043,56 @@ void intel_init_clock_gating(struct drm_device *dev)
 	dev_priv->display.init_clock_gating(dev);
 }
 
-/* Starting with Haswell, we have different power wells for
- * different parts of the GPU. This attempts to enable them all.
+static void intel_set_power_well(struct drm_device *dev, bool enable)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	bool is_enabled, enable_requested;
+	uint32_t tmp;
+
+	tmp = I915_READ(HSW_PWR_WELL_DRIVER);
+	is_enabled = tmp & HSW_PWR_WELL_STATE;
+	enable_requested = tmp & HSW_PWR_WELL_ENABLE;
+
+	if (enable) {
+		if (!enable_requested)
+			I915_WRITE(HSW_PWR_WELL_DRIVER, HSW_PWR_WELL_ENABLE);
+
+		if (!is_enabled) {
+			DRM_DEBUG_KMS("Enabling power well\n");
+			if (wait_for((I915_READ(HSW_PWR_WELL_DRIVER) &
+				      HSW_PWR_WELL_STATE), 20))
+				DRM_ERROR("Timeout enabling power well\n");
+		}
+	} else {
+		if (enable_requested) {
+			I915_WRITE(HSW_PWR_WELL_DRIVER, 0);
+			DRM_DEBUG_KMS("Requesting to disable the power well\n");
+		}
+	}
+}
+
+/*
+ * Starting with Haswell, we have a "Power Down Well" that can be turned off
+ * when not needed anymore. We have 4 registers that can request the power well
+ * to be enabled, and it will only be disabled if none of the registers is
+ * requesting it to be enabled.
  */
-void intel_init_power_wells(struct drm_device *dev)
+void intel_init_power_well(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	unsigned long power_wells[] = {
-		HSW_PWR_WELL_CTL1,
-		HSW_PWR_WELL_CTL2,
-		HSW_PWR_WELL_CTL4
-	};
-	int i;
 
 	if (!IS_HASWELL(dev))
 		return;
 
 	mutex_lock(&dev->struct_mutex);
 
-	for (i = 0; i < ARRAY_SIZE(power_wells); i++) {
-		int well = I915_READ(power_wells[i]);
+	/* For now, we need the power well to be always enabled. */
+	intel_set_power_well(dev, true);
 
-		if ((well & HSW_PWR_WELL_STATE) == 0) {
-			I915_WRITE(power_wells[i], well & HSW_PWR_WELL_ENABLE);
-			if (wait_for((I915_READ(power_wells[i]) & HSW_PWR_WELL_STATE), 20))
-				DRM_ERROR("Error enabling power well %lx\n", power_wells[i]);
-		}
-	}
+	/* We're taking over the BIOS, so clear any requests made by it since
+	 * the driver is in charge now. */
+	if (I915_READ(HSW_PWR_WELL_BIOS) & HSW_PWR_WELL_ENABLE)
+		I915_WRITE(HSW_PWR_WELL_BIOS, 0);
 
 	mutex_unlock(&dev->struct_mutex);
 }
-- 
cgit v0.10.2


From 56e5a3f043f0440cc0cce288e71134d1cb1b44a7 Mon Sep 17 00:00:00 2001
From: Paulo Zanoni <paulo.r.zanoni@intel.com>
Date: Fri, 25 Jan 2013 16:59:13 -0200
Subject: drm/i915: only disable enabled planes on intel_fb_restore_mode

We should avoid touching registers that are on the power down well
when we don't need to, because if we touch these registers when the
power well is disabled we'll get tons of "unclaimed register"
messages. This commit fixes some of these messages.

Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_fb.c b/drivers/gpu/drm/i915/intel_fb.c
index ce5f544..6591029 100644
--- a/drivers/gpu/drm/i915/intel_fb.c
+++ b/drivers/gpu/drm/i915/intel_fb.c
@@ -304,7 +304,8 @@ void intel_fb_restore_mode(struct drm_device *dev)
 
 	/* Be sure to shut off any planes that may be active */
 	list_for_each_entry(plane, &config->plane_list, head)
-		plane->funcs->disable_plane(plane);
+		if (plane->enabled)
+			plane->funcs->disable_plane(plane);
 
 	mutex_unlock(&dev->mode_config.mutex);
 }
-- 
cgit v0.10.2


From 67c964000236497e00c646472cd6b70b5c5109c8 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Wed, 23 Jan 2013 16:25:09 +0000
Subject: drm/i915: fixup per-crtc locking in intel_release_load_detect_pipe

One of the early return cases missed the mutex unlocking. Hilarity
ensued.

This regression has been introduced in

commit 7b24056be6db7ce907baffdd4cf142ab774ea60c
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Wed Dec 12 00:35:33 2012 +0100

    drm: don't hold crtc mutexes for connector ->detect callbacks

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=59750
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Tested-by: Cancan Feng <cancan.feng@intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 8cda2ad..53ec6cb 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -6513,6 +6513,7 @@ void intel_release_load_detect_pipe(struct drm_connector *connector,
 			drm_framebuffer_unreference(old->release_fb);
 		}
 
+		mutex_unlock(&crtc->mutex);
 		return;
 	}
 
-- 
cgit v0.10.2


From cc464b2a17c59adedbdc02cc54341d630354edc3 Mon Sep 17 00:00:00 2001
From: Paulo Zanoni <paulo.r.zanoni@intel.com>
Date: Fri, 25 Jan 2013 16:59:16 -0200
Subject: drm/i915: set TRANSCODER_EDP even earlier

Instead of setting it at the beginning of haswell_crtc_mode_set, let's
set it at the beginning of intel_crtc_mode_set. When
intel_crt_mode_set calls drm_vblank_pre_modeset we already need to
have the transcoder_edp correctly set, because eventually
drm_vblank_pre_modeset calls functions that call i915_pipe_enabled
from i915_irq.c, which will read PIPECONF(cpu_transcoder).

This is a bug that affects us since we added support for
TRANSCODER_EDP, but I was only able to see the problem after
suspending a machine with the power well disabled (got an "unclaimed
register" error.

Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 19ab75f..c0fadf7 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -5633,11 +5633,6 @@ static int haswell_crtc_mode_set(struct drm_crtc *crtc,
 		num_connectors++;
 	}
 
-	if (is_cpu_edp)
-		intel_crtc->cpu_transcoder = TRANSCODER_EDP;
-	else
-		intel_crtc->cpu_transcoder = pipe;
-
 	/* We are not sure yet this won't happen. */
 	WARN(!HAS_PCH_LPT(dev), "Unexpected PCH type %d\n",
 	     INTEL_PCH_TYPE(dev));
@@ -5702,6 +5697,11 @@ static int intel_crtc_mode_set(struct drm_crtc *crtc,
 	int pipe = intel_crtc->pipe;
 	int ret;
 
+	if (IS_HASWELL(dev) && intel_pipe_has_type(crtc, INTEL_OUTPUT_EDP))
+		intel_crtc->cpu_transcoder = TRANSCODER_EDP;
+	else
+		intel_crtc->cpu_transcoder = pipe;
+
 	drm_vblank_pre_modeset(dev, pipe);
 
 	ret = dev_priv->display.crtc_mode_set(crtc, mode, adjusted_mode,
-- 
cgit v0.10.2


From cb10799c194369633b183262edc75ec4b3e4e346 Mon Sep 17 00:00:00 2001
From: Paulo Zanoni <paulo.r.zanoni@intel.com>
Date: Fri, 25 Jan 2013 16:59:15 -0200
Subject: drm/i915: turn on the power well before suspending

Our suspend code touches a lot of registers all over the place, so we
need to enable the power well before suspending.

Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
[danvet: Fixup compilation by stealing the header decl from the
dynamic power wells patch.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index 521a253..9cc8f87 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -470,6 +470,8 @@ static int i915_drm_freeze(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
+	intel_set_power_well(dev, true);
+
 	drm_kms_helper_poll_disable(dev);
 
 	pci_save_state(dev->pdev);
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 32c3042..fcdfe42 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -668,6 +668,7 @@ extern void intel_gpu_ips_init(struct drm_i915_private *dev_priv);
 extern void intel_gpu_ips_teardown(void);
 
 extern void intel_init_power_well(struct drm_device *dev);
+extern void intel_set_power_well(struct drm_device *dev, bool enable);
 extern void intel_enable_gt_powersave(struct drm_device *dev);
 extern void intel_disable_gt_powersave(struct drm_device *dev);
 extern void gen6_gt_check_fifodbg(struct drm_i915_private *dev_priv);
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index b8cf16c..c97714e 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -4043,7 +4043,7 @@ void intel_init_clock_gating(struct drm_device *dev)
 	dev_priv->display.init_clock_gating(dev);
 }
 
-static void intel_set_power_well(struct drm_device *dev, bool enable)
+void intel_set_power_well(struct drm_device *dev, bool enable)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	bool is_enabled, enable_requested;
-- 
cgit v0.10.2


From 41c0b3a88c7bae96d8e2ee60c7ed91f57fd152d7 Mon Sep 17 00:00:00 2001
From: Ben Widawsky <ben@bwidawsk.net>
Date: Sat, 26 Jan 2013 11:52:00 -0800
Subject: drm/i915: Implement WaVSRefCountFullforceMissDisable
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Implements WaVSRefCountFullforceMissDisable as documented in the BSpec
3D workarounds chapter.

Cc: Paulo Zanoni <przanoni@gmail.com>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index a18218c..0c89cf5 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -741,6 +741,7 @@
 #define   GEN7_FF_TS_SCHED_HS0		(0x3<<16)
 #define   GEN7_FF_TS_SCHED_LOAD_BALANCE	(0x1<<16)
 #define   GEN7_FF_TS_SCHED_HW		(0x0<<16) /* Default */
+#define   GEN7_FF_VS_REF_CNT_FFME	(1 << 15)
 #define   GEN7_FF_VS_SCHED_HS1		(0x5<<12)
 #define   GEN7_FF_VS_SCHED_HS0		(0x3<<12)
 #define   GEN7_FF_VS_SCHED_LOAD_BALANCE	(0x1<<12) /* Default */
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index c97714e..64d65f5 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -3680,6 +3680,10 @@ static void gen7_setup_fixed_func_scheduler(struct drm_i915_private *dev_priv)
 	reg |= GEN7_FF_VS_SCHED_HW;
 	reg |= GEN7_FF_DS_SCHED_HW;
 
+	/* WaVSRefCountFullforceMissDisable */
+	if (IS_HASWELL(dev_priv->dev))
+		reg &= ~GEN7_FF_VS_REF_CNT_FFME;
+
 	I915_WRITE(GEN7_FF_THREAD_MODE, reg);
 }
 
-- 
cgit v0.10.2


From 2e9723a3e76a050f80064249dac644f86541d3b4 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Fri, 25 Jan 2013 17:53:19 +0100
Subject: drm/i915: move modeset checks out of save/restore_modeset_reg

That way the control flow is clearer, and it prepares the stage
to extract these ums functions and hide them somewhere.

There's still tons of display stuff outside of these, but that
requires more work.

Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_suspend.c b/drivers/gpu/drm/i915/i915_suspend.c
index 95c0582..bcde494 100644
--- a/drivers/gpu/drm/i915/i915_suspend.c
+++ b/drivers/gpu/drm/i915/i915_suspend.c
@@ -240,9 +240,6 @@ static void i915_save_modeset_reg(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int i;
 
-	if (drm_core_check_feature(dev, DRIVER_MODESET))
-		return;
-
 	/* Cursor state */
 	dev_priv->regfile.saveCURACNTR = I915_READ(_CURACNTR);
 	dev_priv->regfile.saveCURAPOS = I915_READ(_CURAPOS);
@@ -410,8 +407,17 @@ static void i915_restore_modeset_reg(struct drm_device *dev)
 	int dpll_b_reg, fpb0_reg, fpb1_reg;
 	int i;
 
-	if (drm_core_check_feature(dev, DRIVER_MODESET))
-		return;
+	/* Display port ratios (must be done before clock is set) */
+	if (SUPPORTS_INTEGRATED_DP(dev)) {
+		I915_WRITE(_PIPEA_GMCH_DATA_M, dev_priv->regfile.savePIPEA_GMCH_DATA_M);
+		I915_WRITE(_PIPEB_GMCH_DATA_M, dev_priv->regfile.savePIPEB_GMCH_DATA_M);
+		I915_WRITE(_PIPEA_GMCH_DATA_N, dev_priv->regfile.savePIPEA_GMCH_DATA_N);
+		I915_WRITE(_PIPEB_GMCH_DATA_N, dev_priv->regfile.savePIPEB_GMCH_DATA_N);
+		I915_WRITE(_PIPEA_DP_LINK_M, dev_priv->regfile.savePIPEA_DP_LINK_M);
+		I915_WRITE(_PIPEB_DP_LINK_M, dev_priv->regfile.savePIPEB_DP_LINK_M);
+		I915_WRITE(_PIPEA_DP_LINK_N, dev_priv->regfile.savePIPEA_DP_LINK_N);
+		I915_WRITE(_PIPEB_DP_LINK_N, dev_priv->regfile.savePIPEB_DP_LINK_N);
+	}
 
 	/* Fences */
 	switch (INTEL_INFO(dev)->gen) {
@@ -625,7 +631,8 @@ static void i915_save_display(struct drm_device *dev)
 
 	/* This is only meaningful in non-KMS mode */
 	/* Don't regfile.save them in KMS mode */
-	i915_save_modeset_reg(dev);
+	if (!drm_core_check_feature(dev, DRIVER_MODESET))
+		i915_save_modeset_reg(dev);
 
 	/* LVDS state */
 	if (HAS_PCH_SPLIT(dev)) {
@@ -711,23 +718,8 @@ static void i915_restore_display(struct drm_device *dev)
 	if (INTEL_INFO(dev)->gen <= 4)
 		I915_WRITE(DSPARB, dev_priv->regfile.saveDSPARB);
 
-	if (!drm_core_check_feature(dev, DRIVER_MODESET)) {
-		/* Display port ratios (must be done before clock is set) */
-		if (SUPPORTS_INTEGRATED_DP(dev)) {
-			I915_WRITE(_PIPEA_GMCH_DATA_M, dev_priv->regfile.savePIPEA_GMCH_DATA_M);
-			I915_WRITE(_PIPEB_GMCH_DATA_M, dev_priv->regfile.savePIPEB_GMCH_DATA_M);
-			I915_WRITE(_PIPEA_GMCH_DATA_N, dev_priv->regfile.savePIPEA_GMCH_DATA_N);
-			I915_WRITE(_PIPEB_GMCH_DATA_N, dev_priv->regfile.savePIPEB_GMCH_DATA_N);
-			I915_WRITE(_PIPEA_DP_LINK_M, dev_priv->regfile.savePIPEA_DP_LINK_M);
-			I915_WRITE(_PIPEB_DP_LINK_M, dev_priv->regfile.savePIPEB_DP_LINK_M);
-			I915_WRITE(_PIPEA_DP_LINK_N, dev_priv->regfile.savePIPEA_DP_LINK_N);
-			I915_WRITE(_PIPEB_DP_LINK_N, dev_priv->regfile.savePIPEB_DP_LINK_N);
-		}
-	}
-
-	/* This is only meaningful in non-KMS mode */
-	/* Don't restore them in KMS mode */
-	i915_restore_modeset_reg(dev);
+	if (!drm_core_check_feature(dev, DRIVER_MODESET))
+		i915_restore_modeset_reg(dev);
 
 	/* LVDS state */
 	if (INTEL_INFO(dev)->gen >= 4 && !HAS_PCH_SPLIT(dev))
-- 
cgit v0.10.2


From d8157a3687323ae5a76f1ecd0fc592b1d87be81b Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Fri, 25 Jan 2013 17:53:20 +0100
Subject: drm/i915: extract ums suspend/resume into i915_ums.c

Similarly to how i915_dma.c is shaping up to be the dungeon hole for
all things supporting dri1, create a new one to hide all the crazy
things which are only really useful for ums support. Biggest part is
the register suspend/resume support.

Unfortunately a lot of it is still intermingled with bits and pieces
we might still need, so needs more analysis and needs to stay in
i915_suspend.c for now.

Reviewed-by: Imre Deak <imre.deak@intel.com>

v2: s/modeset_reg/display_reg/ as suggested by Imre, to avoid
confusion between the kernel modeset code and display save/restore to
support ums.

v3: Fixup alphabetical order in the Makefile, spotted by Chris Wilson.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile
index 0f2c549..91f3ac6 100644
--- a/drivers/gpu/drm/i915/Makefile
+++ b/drivers/gpu/drm/i915/Makefile
@@ -16,6 +16,7 @@ i915-y := i915_drv.o i915_dma.o i915_irq.o \
 	  i915_gem_tiling.o \
 	  i915_sysfs.o \
 	  i915_trace_points.o \
+	  i915_ums.o \
 	  intel_display.o \
 	  intel_crt.o \
 	  intel_lvds.o \
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 09753e5..a263a0c 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -1730,9 +1730,9 @@ void i915_debugfs_cleanup(struct drm_minor *minor);
 extern int i915_save_state(struct drm_device *dev);
 extern int i915_restore_state(struct drm_device *dev);
 
-/* i915_suspend.c */
-extern int i915_save_state(struct drm_device *dev);
-extern int i915_restore_state(struct drm_device *dev);
+/* i915_ums.c */
+void i915_save_display_reg(struct drm_device *dev);
+void i915_restore_display_reg(struct drm_device *dev);
 
 /* i915_sysfs.c */
 void i915_setup_sysfs(struct drm_device *dev_priv);
diff --git a/drivers/gpu/drm/i915/i915_suspend.c b/drivers/gpu/drm/i915/i915_suspend.c
index bcde494..1f581fc 100644
--- a/drivers/gpu/drm/i915/i915_suspend.c
+++ b/drivers/gpu/drm/i915/i915_suspend.c
@@ -29,67 +29,6 @@
 #include "intel_drv.h"
 #include "i915_reg.h"
 
-static bool i915_pipe_enabled(struct drm_device *dev, enum pipe pipe)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32	dpll_reg;
-
-	/* On IVB, 3rd pipe shares PLL with another one */
-	if (pipe > 1)
-		return false;
-
-	if (HAS_PCH_SPLIT(dev))
-		dpll_reg = _PCH_DPLL(pipe);
-	else
-		dpll_reg = (pipe == PIPE_A) ? _DPLL_A : _DPLL_B;
-
-	return (I915_READ(dpll_reg) & DPLL_VCO_ENABLE);
-}
-
-static void i915_save_palette(struct drm_device *dev, enum pipe pipe)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	unsigned long reg = (pipe == PIPE_A ? _PALETTE_A : _PALETTE_B);
-	u32 *array;
-	int i;
-
-	if (!i915_pipe_enabled(dev, pipe))
-		return;
-
-	if (HAS_PCH_SPLIT(dev))
-		reg = (pipe == PIPE_A) ? _LGC_PALETTE_A : _LGC_PALETTE_B;
-
-	if (pipe == PIPE_A)
-		array = dev_priv->regfile.save_palette_a;
-	else
-		array = dev_priv->regfile.save_palette_b;
-
-	for (i = 0; i < 256; i++)
-		array[i] = I915_READ(reg + (i << 2));
-}
-
-static void i915_restore_palette(struct drm_device *dev, enum pipe pipe)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	unsigned long reg = (pipe == PIPE_A ? _PALETTE_A : _PALETTE_B);
-	u32 *array;
-	int i;
-
-	if (!i915_pipe_enabled(dev, pipe))
-		return;
-
-	if (HAS_PCH_SPLIT(dev))
-		reg = (pipe == PIPE_A) ? _LGC_PALETTE_A : _LGC_PALETTE_B;
-
-	if (pipe == PIPE_A)
-		array = dev_priv->regfile.save_palette_a;
-	else
-		array = dev_priv->regfile.save_palette_b;
-
-	for (i = 0; i < 256; i++)
-		I915_WRITE(reg + (i << 2), array[i]);
-}
-
 static u8 i915_read_indexed(struct drm_device *dev, u16 index_port, u16 data_port, u8 reg)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -235,392 +174,6 @@ static void i915_restore_vga(struct drm_device *dev)
 	I915_WRITE8(VGA_DACMASK, dev_priv->regfile.saveDACMASK);
 }
 
-static void i915_save_modeset_reg(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int i;
-
-	/* Cursor state */
-	dev_priv->regfile.saveCURACNTR = I915_READ(_CURACNTR);
-	dev_priv->regfile.saveCURAPOS = I915_READ(_CURAPOS);
-	dev_priv->regfile.saveCURABASE = I915_READ(_CURABASE);
-	dev_priv->regfile.saveCURBCNTR = I915_READ(_CURBCNTR);
-	dev_priv->regfile.saveCURBPOS = I915_READ(_CURBPOS);
-	dev_priv->regfile.saveCURBBASE = I915_READ(_CURBBASE);
-	if (IS_GEN2(dev))
-		dev_priv->regfile.saveCURSIZE = I915_READ(CURSIZE);
-
-	if (HAS_PCH_SPLIT(dev)) {
-		dev_priv->regfile.savePCH_DREF_CONTROL = I915_READ(PCH_DREF_CONTROL);
-		dev_priv->regfile.saveDISP_ARB_CTL = I915_READ(DISP_ARB_CTL);
-	}
-
-	/* Pipe & plane A info */
-	dev_priv->regfile.savePIPEACONF = I915_READ(_PIPEACONF);
-	dev_priv->regfile.savePIPEASRC = I915_READ(_PIPEASRC);
-	if (HAS_PCH_SPLIT(dev)) {
-		dev_priv->regfile.saveFPA0 = I915_READ(_PCH_FPA0);
-		dev_priv->regfile.saveFPA1 = I915_READ(_PCH_FPA1);
-		dev_priv->regfile.saveDPLL_A = I915_READ(_PCH_DPLL_A);
-	} else {
-		dev_priv->regfile.saveFPA0 = I915_READ(_FPA0);
-		dev_priv->regfile.saveFPA1 = I915_READ(_FPA1);
-		dev_priv->regfile.saveDPLL_A = I915_READ(_DPLL_A);
-	}
-	if (INTEL_INFO(dev)->gen >= 4 && !HAS_PCH_SPLIT(dev))
-		dev_priv->regfile.saveDPLL_A_MD = I915_READ(_DPLL_A_MD);
-	dev_priv->regfile.saveHTOTAL_A = I915_READ(_HTOTAL_A);
-	dev_priv->regfile.saveHBLANK_A = I915_READ(_HBLANK_A);
-	dev_priv->regfile.saveHSYNC_A = I915_READ(_HSYNC_A);
-	dev_priv->regfile.saveVTOTAL_A = I915_READ(_VTOTAL_A);
-	dev_priv->regfile.saveVBLANK_A = I915_READ(_VBLANK_A);
-	dev_priv->regfile.saveVSYNC_A = I915_READ(_VSYNC_A);
-	if (!HAS_PCH_SPLIT(dev))
-		dev_priv->regfile.saveBCLRPAT_A = I915_READ(_BCLRPAT_A);
-
-	if (HAS_PCH_SPLIT(dev)) {
-		dev_priv->regfile.savePIPEA_DATA_M1 = I915_READ(_PIPEA_DATA_M1);
-		dev_priv->regfile.savePIPEA_DATA_N1 = I915_READ(_PIPEA_DATA_N1);
-		dev_priv->regfile.savePIPEA_LINK_M1 = I915_READ(_PIPEA_LINK_M1);
-		dev_priv->regfile.savePIPEA_LINK_N1 = I915_READ(_PIPEA_LINK_N1);
-
-		dev_priv->regfile.saveFDI_TXA_CTL = I915_READ(_FDI_TXA_CTL);
-		dev_priv->regfile.saveFDI_RXA_CTL = I915_READ(_FDI_RXA_CTL);
-
-		dev_priv->regfile.savePFA_CTL_1 = I915_READ(_PFA_CTL_1);
-		dev_priv->regfile.savePFA_WIN_SZ = I915_READ(_PFA_WIN_SZ);
-		dev_priv->regfile.savePFA_WIN_POS = I915_READ(_PFA_WIN_POS);
-
-		dev_priv->regfile.saveTRANSACONF = I915_READ(_TRANSACONF);
-		dev_priv->regfile.saveTRANS_HTOTAL_A = I915_READ(_TRANS_HTOTAL_A);
-		dev_priv->regfile.saveTRANS_HBLANK_A = I915_READ(_TRANS_HBLANK_A);
-		dev_priv->regfile.saveTRANS_HSYNC_A = I915_READ(_TRANS_HSYNC_A);
-		dev_priv->regfile.saveTRANS_VTOTAL_A = I915_READ(_TRANS_VTOTAL_A);
-		dev_priv->regfile.saveTRANS_VBLANK_A = I915_READ(_TRANS_VBLANK_A);
-		dev_priv->regfile.saveTRANS_VSYNC_A = I915_READ(_TRANS_VSYNC_A);
-	}
-
-	dev_priv->regfile.saveDSPACNTR = I915_READ(_DSPACNTR);
-	dev_priv->regfile.saveDSPASTRIDE = I915_READ(_DSPASTRIDE);
-	dev_priv->regfile.saveDSPASIZE = I915_READ(_DSPASIZE);
-	dev_priv->regfile.saveDSPAPOS = I915_READ(_DSPAPOS);
-	dev_priv->regfile.saveDSPAADDR = I915_READ(_DSPAADDR);
-	if (INTEL_INFO(dev)->gen >= 4) {
-		dev_priv->regfile.saveDSPASURF = I915_READ(_DSPASURF);
-		dev_priv->regfile.saveDSPATILEOFF = I915_READ(_DSPATILEOFF);
-	}
-	i915_save_palette(dev, PIPE_A);
-	dev_priv->regfile.savePIPEASTAT = I915_READ(_PIPEASTAT);
-
-	/* Pipe & plane B info */
-	dev_priv->regfile.savePIPEBCONF = I915_READ(_PIPEBCONF);
-	dev_priv->regfile.savePIPEBSRC = I915_READ(_PIPEBSRC);
-	if (HAS_PCH_SPLIT(dev)) {
-		dev_priv->regfile.saveFPB0 = I915_READ(_PCH_FPB0);
-		dev_priv->regfile.saveFPB1 = I915_READ(_PCH_FPB1);
-		dev_priv->regfile.saveDPLL_B = I915_READ(_PCH_DPLL_B);
-	} else {
-		dev_priv->regfile.saveFPB0 = I915_READ(_FPB0);
-		dev_priv->regfile.saveFPB1 = I915_READ(_FPB1);
-		dev_priv->regfile.saveDPLL_B = I915_READ(_DPLL_B);
-	}
-	if (INTEL_INFO(dev)->gen >= 4 && !HAS_PCH_SPLIT(dev))
-		dev_priv->regfile.saveDPLL_B_MD = I915_READ(_DPLL_B_MD);
-	dev_priv->regfile.saveHTOTAL_B = I915_READ(_HTOTAL_B);
-	dev_priv->regfile.saveHBLANK_B = I915_READ(_HBLANK_B);
-	dev_priv->regfile.saveHSYNC_B = I915_READ(_HSYNC_B);
-	dev_priv->regfile.saveVTOTAL_B = I915_READ(_VTOTAL_B);
-	dev_priv->regfile.saveVBLANK_B = I915_READ(_VBLANK_B);
-	dev_priv->regfile.saveVSYNC_B = I915_READ(_VSYNC_B);
-	if (!HAS_PCH_SPLIT(dev))
-		dev_priv->regfile.saveBCLRPAT_B = I915_READ(_BCLRPAT_B);
-
-	if (HAS_PCH_SPLIT(dev)) {
-		dev_priv->regfile.savePIPEB_DATA_M1 = I915_READ(_PIPEB_DATA_M1);
-		dev_priv->regfile.savePIPEB_DATA_N1 = I915_READ(_PIPEB_DATA_N1);
-		dev_priv->regfile.savePIPEB_LINK_M1 = I915_READ(_PIPEB_LINK_M1);
-		dev_priv->regfile.savePIPEB_LINK_N1 = I915_READ(_PIPEB_LINK_N1);
-
-		dev_priv->regfile.saveFDI_TXB_CTL = I915_READ(_FDI_TXB_CTL);
-		dev_priv->regfile.saveFDI_RXB_CTL = I915_READ(_FDI_RXB_CTL);
-
-		dev_priv->regfile.savePFB_CTL_1 = I915_READ(_PFB_CTL_1);
-		dev_priv->regfile.savePFB_WIN_SZ = I915_READ(_PFB_WIN_SZ);
-		dev_priv->regfile.savePFB_WIN_POS = I915_READ(_PFB_WIN_POS);
-
-		dev_priv->regfile.saveTRANSBCONF = I915_READ(_TRANSBCONF);
-		dev_priv->regfile.saveTRANS_HTOTAL_B = I915_READ(_TRANS_HTOTAL_B);
-		dev_priv->regfile.saveTRANS_HBLANK_B = I915_READ(_TRANS_HBLANK_B);
-		dev_priv->regfile.saveTRANS_HSYNC_B = I915_READ(_TRANS_HSYNC_B);
-		dev_priv->regfile.saveTRANS_VTOTAL_B = I915_READ(_TRANS_VTOTAL_B);
-		dev_priv->regfile.saveTRANS_VBLANK_B = I915_READ(_TRANS_VBLANK_B);
-		dev_priv->regfile.saveTRANS_VSYNC_B = I915_READ(_TRANS_VSYNC_B);
-	}
-
-	dev_priv->regfile.saveDSPBCNTR = I915_READ(_DSPBCNTR);
-	dev_priv->regfile.saveDSPBSTRIDE = I915_READ(_DSPBSTRIDE);
-	dev_priv->regfile.saveDSPBSIZE = I915_READ(_DSPBSIZE);
-	dev_priv->regfile.saveDSPBPOS = I915_READ(_DSPBPOS);
-	dev_priv->regfile.saveDSPBADDR = I915_READ(_DSPBADDR);
-	if (INTEL_INFO(dev)->gen >= 4) {
-		dev_priv->regfile.saveDSPBSURF = I915_READ(_DSPBSURF);
-		dev_priv->regfile.saveDSPBTILEOFF = I915_READ(_DSPBTILEOFF);
-	}
-	i915_save_palette(dev, PIPE_B);
-	dev_priv->regfile.savePIPEBSTAT = I915_READ(_PIPEBSTAT);
-
-	/* Fences */
-	switch (INTEL_INFO(dev)->gen) {
-	case 7:
-	case 6:
-		for (i = 0; i < 16; i++)
-			dev_priv->regfile.saveFENCE[i] = I915_READ64(FENCE_REG_SANDYBRIDGE_0 + (i * 8));
-		break;
-	case 5:
-	case 4:
-		for (i = 0; i < 16; i++)
-			dev_priv->regfile.saveFENCE[i] = I915_READ64(FENCE_REG_965_0 + (i * 8));
-		break;
-	case 3:
-		if (IS_I945G(dev) || IS_I945GM(dev) || IS_G33(dev))
-			for (i = 0; i < 8; i++)
-				dev_priv->regfile.saveFENCE[i+8] = I915_READ(FENCE_REG_945_8 + (i * 4));
-	case 2:
-		for (i = 0; i < 8; i++)
-			dev_priv->regfile.saveFENCE[i] = I915_READ(FENCE_REG_830_0 + (i * 4));
-		break;
-	}
-
-	/* CRT state */
-	if (HAS_PCH_SPLIT(dev))
-		dev_priv->regfile.saveADPA = I915_READ(PCH_ADPA);
-	else
-		dev_priv->regfile.saveADPA = I915_READ(ADPA);
-
-	return;
-}
-
-static void i915_restore_modeset_reg(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int dpll_a_reg, fpa0_reg, fpa1_reg;
-	int dpll_b_reg, fpb0_reg, fpb1_reg;
-	int i;
-
-	/* Display port ratios (must be done before clock is set) */
-	if (SUPPORTS_INTEGRATED_DP(dev)) {
-		I915_WRITE(_PIPEA_GMCH_DATA_M, dev_priv->regfile.savePIPEA_GMCH_DATA_M);
-		I915_WRITE(_PIPEB_GMCH_DATA_M, dev_priv->regfile.savePIPEB_GMCH_DATA_M);
-		I915_WRITE(_PIPEA_GMCH_DATA_N, dev_priv->regfile.savePIPEA_GMCH_DATA_N);
-		I915_WRITE(_PIPEB_GMCH_DATA_N, dev_priv->regfile.savePIPEB_GMCH_DATA_N);
-		I915_WRITE(_PIPEA_DP_LINK_M, dev_priv->regfile.savePIPEA_DP_LINK_M);
-		I915_WRITE(_PIPEB_DP_LINK_M, dev_priv->regfile.savePIPEB_DP_LINK_M);
-		I915_WRITE(_PIPEA_DP_LINK_N, dev_priv->regfile.savePIPEA_DP_LINK_N);
-		I915_WRITE(_PIPEB_DP_LINK_N, dev_priv->regfile.savePIPEB_DP_LINK_N);
-	}
-
-	/* Fences */
-	switch (INTEL_INFO(dev)->gen) {
-	case 7:
-	case 6:
-		for (i = 0; i < 16; i++)
-			I915_WRITE64(FENCE_REG_SANDYBRIDGE_0 + (i * 8), dev_priv->regfile.saveFENCE[i]);
-		break;
-	case 5:
-	case 4:
-		for (i = 0; i < 16; i++)
-			I915_WRITE64(FENCE_REG_965_0 + (i * 8), dev_priv->regfile.saveFENCE[i]);
-		break;
-	case 3:
-	case 2:
-		if (IS_I945G(dev) || IS_I945GM(dev) || IS_G33(dev))
-			for (i = 0; i < 8; i++)
-				I915_WRITE(FENCE_REG_945_8 + (i * 4), dev_priv->regfile.saveFENCE[i+8]);
-		for (i = 0; i < 8; i++)
-			I915_WRITE(FENCE_REG_830_0 + (i * 4), dev_priv->regfile.saveFENCE[i]);
-		break;
-	}
-
-
-	if (HAS_PCH_SPLIT(dev)) {
-		dpll_a_reg = _PCH_DPLL_A;
-		dpll_b_reg = _PCH_DPLL_B;
-		fpa0_reg = _PCH_FPA0;
-		fpb0_reg = _PCH_FPB0;
-		fpa1_reg = _PCH_FPA1;
-		fpb1_reg = _PCH_FPB1;
-	} else {
-		dpll_a_reg = _DPLL_A;
-		dpll_b_reg = _DPLL_B;
-		fpa0_reg = _FPA0;
-		fpb0_reg = _FPB0;
-		fpa1_reg = _FPA1;
-		fpb1_reg = _FPB1;
-	}
-
-	if (HAS_PCH_SPLIT(dev)) {
-		I915_WRITE(PCH_DREF_CONTROL, dev_priv->regfile.savePCH_DREF_CONTROL);
-		I915_WRITE(DISP_ARB_CTL, dev_priv->regfile.saveDISP_ARB_CTL);
-	}
-
-	/* Pipe & plane A info */
-	/* Prime the clock */
-	if (dev_priv->regfile.saveDPLL_A & DPLL_VCO_ENABLE) {
-		I915_WRITE(dpll_a_reg, dev_priv->regfile.saveDPLL_A &
-			   ~DPLL_VCO_ENABLE);
-		POSTING_READ(dpll_a_reg);
-		udelay(150);
-	}
-	I915_WRITE(fpa0_reg, dev_priv->regfile.saveFPA0);
-	I915_WRITE(fpa1_reg, dev_priv->regfile.saveFPA1);
-	/* Actually enable it */
-	I915_WRITE(dpll_a_reg, dev_priv->regfile.saveDPLL_A);
-	POSTING_READ(dpll_a_reg);
-	udelay(150);
-	if (INTEL_INFO(dev)->gen >= 4 && !HAS_PCH_SPLIT(dev)) {
-		I915_WRITE(_DPLL_A_MD, dev_priv->regfile.saveDPLL_A_MD);
-		POSTING_READ(_DPLL_A_MD);
-	}
-	udelay(150);
-
-	/* Restore mode */
-	I915_WRITE(_HTOTAL_A, dev_priv->regfile.saveHTOTAL_A);
-	I915_WRITE(_HBLANK_A, dev_priv->regfile.saveHBLANK_A);
-	I915_WRITE(_HSYNC_A, dev_priv->regfile.saveHSYNC_A);
-	I915_WRITE(_VTOTAL_A, dev_priv->regfile.saveVTOTAL_A);
-	I915_WRITE(_VBLANK_A, dev_priv->regfile.saveVBLANK_A);
-	I915_WRITE(_VSYNC_A, dev_priv->regfile.saveVSYNC_A);
-	if (!HAS_PCH_SPLIT(dev))
-		I915_WRITE(_BCLRPAT_A, dev_priv->regfile.saveBCLRPAT_A);
-
-	if (HAS_PCH_SPLIT(dev)) {
-		I915_WRITE(_PIPEA_DATA_M1, dev_priv->regfile.savePIPEA_DATA_M1);
-		I915_WRITE(_PIPEA_DATA_N1, dev_priv->regfile.savePIPEA_DATA_N1);
-		I915_WRITE(_PIPEA_LINK_M1, dev_priv->regfile.savePIPEA_LINK_M1);
-		I915_WRITE(_PIPEA_LINK_N1, dev_priv->regfile.savePIPEA_LINK_N1);
-
-		I915_WRITE(_FDI_RXA_CTL, dev_priv->regfile.saveFDI_RXA_CTL);
-		I915_WRITE(_FDI_TXA_CTL, dev_priv->regfile.saveFDI_TXA_CTL);
-
-		I915_WRITE(_PFA_CTL_1, dev_priv->regfile.savePFA_CTL_1);
-		I915_WRITE(_PFA_WIN_SZ, dev_priv->regfile.savePFA_WIN_SZ);
-		I915_WRITE(_PFA_WIN_POS, dev_priv->regfile.savePFA_WIN_POS);
-
-		I915_WRITE(_TRANSACONF, dev_priv->regfile.saveTRANSACONF);
-		I915_WRITE(_TRANS_HTOTAL_A, dev_priv->regfile.saveTRANS_HTOTAL_A);
-		I915_WRITE(_TRANS_HBLANK_A, dev_priv->regfile.saveTRANS_HBLANK_A);
-		I915_WRITE(_TRANS_HSYNC_A, dev_priv->regfile.saveTRANS_HSYNC_A);
-		I915_WRITE(_TRANS_VTOTAL_A, dev_priv->regfile.saveTRANS_VTOTAL_A);
-		I915_WRITE(_TRANS_VBLANK_A, dev_priv->regfile.saveTRANS_VBLANK_A);
-		I915_WRITE(_TRANS_VSYNC_A, dev_priv->regfile.saveTRANS_VSYNC_A);
-	}
-
-	/* Restore plane info */
-	I915_WRITE(_DSPASIZE, dev_priv->regfile.saveDSPASIZE);
-	I915_WRITE(_DSPAPOS, dev_priv->regfile.saveDSPAPOS);
-	I915_WRITE(_PIPEASRC, dev_priv->regfile.savePIPEASRC);
-	I915_WRITE(_DSPAADDR, dev_priv->regfile.saveDSPAADDR);
-	I915_WRITE(_DSPASTRIDE, dev_priv->regfile.saveDSPASTRIDE);
-	if (INTEL_INFO(dev)->gen >= 4) {
-		I915_WRITE(_DSPASURF, dev_priv->regfile.saveDSPASURF);
-		I915_WRITE(_DSPATILEOFF, dev_priv->regfile.saveDSPATILEOFF);
-	}
-
-	I915_WRITE(_PIPEACONF, dev_priv->regfile.savePIPEACONF);
-
-	i915_restore_palette(dev, PIPE_A);
-	/* Enable the plane */
-	I915_WRITE(_DSPACNTR, dev_priv->regfile.saveDSPACNTR);
-	I915_WRITE(_DSPAADDR, I915_READ(_DSPAADDR));
-
-	/* Pipe & plane B info */
-	if (dev_priv->regfile.saveDPLL_B & DPLL_VCO_ENABLE) {
-		I915_WRITE(dpll_b_reg, dev_priv->regfile.saveDPLL_B &
-			   ~DPLL_VCO_ENABLE);
-		POSTING_READ(dpll_b_reg);
-		udelay(150);
-	}
-	I915_WRITE(fpb0_reg, dev_priv->regfile.saveFPB0);
-	I915_WRITE(fpb1_reg, dev_priv->regfile.saveFPB1);
-	/* Actually enable it */
-	I915_WRITE(dpll_b_reg, dev_priv->regfile.saveDPLL_B);
-	POSTING_READ(dpll_b_reg);
-	udelay(150);
-	if (INTEL_INFO(dev)->gen >= 4 && !HAS_PCH_SPLIT(dev)) {
-		I915_WRITE(_DPLL_B_MD, dev_priv->regfile.saveDPLL_B_MD);
-		POSTING_READ(_DPLL_B_MD);
-	}
-	udelay(150);
-
-	/* Restore mode */
-	I915_WRITE(_HTOTAL_B, dev_priv->regfile.saveHTOTAL_B);
-	I915_WRITE(_HBLANK_B, dev_priv->regfile.saveHBLANK_B);
-	I915_WRITE(_HSYNC_B, dev_priv->regfile.saveHSYNC_B);
-	I915_WRITE(_VTOTAL_B, dev_priv->regfile.saveVTOTAL_B);
-	I915_WRITE(_VBLANK_B, dev_priv->regfile.saveVBLANK_B);
-	I915_WRITE(_VSYNC_B, dev_priv->regfile.saveVSYNC_B);
-	if (!HAS_PCH_SPLIT(dev))
-		I915_WRITE(_BCLRPAT_B, dev_priv->regfile.saveBCLRPAT_B);
-
-	if (HAS_PCH_SPLIT(dev)) {
-		I915_WRITE(_PIPEB_DATA_M1, dev_priv->regfile.savePIPEB_DATA_M1);
-		I915_WRITE(_PIPEB_DATA_N1, dev_priv->regfile.savePIPEB_DATA_N1);
-		I915_WRITE(_PIPEB_LINK_M1, dev_priv->regfile.savePIPEB_LINK_M1);
-		I915_WRITE(_PIPEB_LINK_N1, dev_priv->regfile.savePIPEB_LINK_N1);
-
-		I915_WRITE(_FDI_RXB_CTL, dev_priv->regfile.saveFDI_RXB_CTL);
-		I915_WRITE(_FDI_TXB_CTL, dev_priv->regfile.saveFDI_TXB_CTL);
-
-		I915_WRITE(_PFB_CTL_1, dev_priv->regfile.savePFB_CTL_1);
-		I915_WRITE(_PFB_WIN_SZ, dev_priv->regfile.savePFB_WIN_SZ);
-		I915_WRITE(_PFB_WIN_POS, dev_priv->regfile.savePFB_WIN_POS);
-
-		I915_WRITE(_TRANSBCONF, dev_priv->regfile.saveTRANSBCONF);
-		I915_WRITE(_TRANS_HTOTAL_B, dev_priv->regfile.saveTRANS_HTOTAL_B);
-		I915_WRITE(_TRANS_HBLANK_B, dev_priv->regfile.saveTRANS_HBLANK_B);
-		I915_WRITE(_TRANS_HSYNC_B, dev_priv->regfile.saveTRANS_HSYNC_B);
-		I915_WRITE(_TRANS_VTOTAL_B, dev_priv->regfile.saveTRANS_VTOTAL_B);
-		I915_WRITE(_TRANS_VBLANK_B, dev_priv->regfile.saveTRANS_VBLANK_B);
-		I915_WRITE(_TRANS_VSYNC_B, dev_priv->regfile.saveTRANS_VSYNC_B);
-	}
-
-	/* Restore plane info */
-	I915_WRITE(_DSPBSIZE, dev_priv->regfile.saveDSPBSIZE);
-	I915_WRITE(_DSPBPOS, dev_priv->regfile.saveDSPBPOS);
-	I915_WRITE(_PIPEBSRC, dev_priv->regfile.savePIPEBSRC);
-	I915_WRITE(_DSPBADDR, dev_priv->regfile.saveDSPBADDR);
-	I915_WRITE(_DSPBSTRIDE, dev_priv->regfile.saveDSPBSTRIDE);
-	if (INTEL_INFO(dev)->gen >= 4) {
-		I915_WRITE(_DSPBSURF, dev_priv->regfile.saveDSPBSURF);
-		I915_WRITE(_DSPBTILEOFF, dev_priv->regfile.saveDSPBTILEOFF);
-	}
-
-	I915_WRITE(_PIPEBCONF, dev_priv->regfile.savePIPEBCONF);
-
-	i915_restore_palette(dev, PIPE_B);
-	/* Enable the plane */
-	I915_WRITE(_DSPBCNTR, dev_priv->regfile.saveDSPBCNTR);
-	I915_WRITE(_DSPBADDR, I915_READ(_DSPBADDR));
-
-	/* Cursor state */
-	I915_WRITE(_CURAPOS, dev_priv->regfile.saveCURAPOS);
-	I915_WRITE(_CURACNTR, dev_priv->regfile.saveCURACNTR);
-	I915_WRITE(_CURABASE, dev_priv->regfile.saveCURABASE);
-	I915_WRITE(_CURBPOS, dev_priv->regfile.saveCURBPOS);
-	I915_WRITE(_CURBCNTR, dev_priv->regfile.saveCURBCNTR);
-	I915_WRITE(_CURBBASE, dev_priv->regfile.saveCURBBASE);
-	if (IS_GEN2(dev))
-		I915_WRITE(CURSIZE, dev_priv->regfile.saveCURSIZE);
-
-	/* CRT state */
-	if (HAS_PCH_SPLIT(dev))
-		I915_WRITE(PCH_ADPA, dev_priv->regfile.saveADPA);
-	else
-		I915_WRITE(ADPA, dev_priv->regfile.saveADPA);
-
-	return;
-}
-
 static void i915_save_display(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -632,7 +185,7 @@ static void i915_save_display(struct drm_device *dev)
 	/* This is only meaningful in non-KMS mode */
 	/* Don't regfile.save them in KMS mode */
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
-		i915_save_modeset_reg(dev);
+		i915_save_display_reg(dev);
 
 	/* LVDS state */
 	if (HAS_PCH_SPLIT(dev)) {
@@ -719,7 +272,7 @@ static void i915_restore_display(struct drm_device *dev)
 		I915_WRITE(DSPARB, dev_priv->regfile.saveDSPARB);
 
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
-		i915_restore_modeset_reg(dev);
+		i915_restore_display_reg(dev);
 
 	/* LVDS state */
 	if (INTEL_INFO(dev)->gen >= 4 && !HAS_PCH_SPLIT(dev))
diff --git a/drivers/gpu/drm/i915/i915_ums.c b/drivers/gpu/drm/i915/i915_ums.c
new file mode 100644
index 0000000..359c77b
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_ums.c
@@ -0,0 +1,479 @@
+/*
+ *
+ * Copyright 2008 (c) Intel Corporation
+ *   Jesse Barnes <jbarnes@virtuousgeek.org>
+ * Copyright 2013 (c) Intel Corporation
+ *   Daniel Vetter <daniel.vetter@ffwll.ch>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include <drm/drmP.h>
+#include <drm/i915_drm.h>
+#include "intel_drv.h"
+#include "i915_reg.h"
+
+static bool i915_pipe_enabled(struct drm_device *dev, enum pipe pipe)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32	dpll_reg;
+
+	/* On IVB, 3rd pipe shares PLL with another one */
+	if (pipe > 1)
+		return false;
+
+	if (HAS_PCH_SPLIT(dev))
+		dpll_reg = _PCH_DPLL(pipe);
+	else
+		dpll_reg = (pipe == PIPE_A) ? _DPLL_A : _DPLL_B;
+
+	return (I915_READ(dpll_reg) & DPLL_VCO_ENABLE);
+}
+
+static void i915_save_palette(struct drm_device *dev, enum pipe pipe)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	unsigned long reg = (pipe == PIPE_A ? _PALETTE_A : _PALETTE_B);
+	u32 *array;
+	int i;
+
+	if (!i915_pipe_enabled(dev, pipe))
+		return;
+
+	if (HAS_PCH_SPLIT(dev))
+		reg = (pipe == PIPE_A) ? _LGC_PALETTE_A : _LGC_PALETTE_B;
+
+	if (pipe == PIPE_A)
+		array = dev_priv->regfile.save_palette_a;
+	else
+		array = dev_priv->regfile.save_palette_b;
+
+	for (i = 0; i < 256; i++)
+		array[i] = I915_READ(reg + (i << 2));
+}
+
+static void i915_restore_palette(struct drm_device *dev, enum pipe pipe)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	unsigned long reg = (pipe == PIPE_A ? _PALETTE_A : _PALETTE_B);
+	u32 *array;
+	int i;
+
+	if (!i915_pipe_enabled(dev, pipe))
+		return;
+
+	if (HAS_PCH_SPLIT(dev))
+		reg = (pipe == PIPE_A) ? _LGC_PALETTE_A : _LGC_PALETTE_B;
+
+	if (pipe == PIPE_A)
+		array = dev_priv->regfile.save_palette_a;
+	else
+		array = dev_priv->regfile.save_palette_b;
+
+	for (i = 0; i < 256; i++)
+		I915_WRITE(reg + (i << 2), array[i]);
+}
+
+void i915_save_display_reg(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int i;
+
+	/* Cursor state */
+	dev_priv->regfile.saveCURACNTR = I915_READ(_CURACNTR);
+	dev_priv->regfile.saveCURAPOS = I915_READ(_CURAPOS);
+	dev_priv->regfile.saveCURABASE = I915_READ(_CURABASE);
+	dev_priv->regfile.saveCURBCNTR = I915_READ(_CURBCNTR);
+	dev_priv->regfile.saveCURBPOS = I915_READ(_CURBPOS);
+	dev_priv->regfile.saveCURBBASE = I915_READ(_CURBBASE);
+	if (IS_GEN2(dev))
+		dev_priv->regfile.saveCURSIZE = I915_READ(CURSIZE);
+
+	if (HAS_PCH_SPLIT(dev)) {
+		dev_priv->regfile.savePCH_DREF_CONTROL = I915_READ(PCH_DREF_CONTROL);
+		dev_priv->regfile.saveDISP_ARB_CTL = I915_READ(DISP_ARB_CTL);
+	}
+
+	/* Pipe & plane A info */
+	dev_priv->regfile.savePIPEACONF = I915_READ(_PIPEACONF);
+	dev_priv->regfile.savePIPEASRC = I915_READ(_PIPEASRC);
+	if (HAS_PCH_SPLIT(dev)) {
+		dev_priv->regfile.saveFPA0 = I915_READ(_PCH_FPA0);
+		dev_priv->regfile.saveFPA1 = I915_READ(_PCH_FPA1);
+		dev_priv->regfile.saveDPLL_A = I915_READ(_PCH_DPLL_A);
+	} else {
+		dev_priv->regfile.saveFPA0 = I915_READ(_FPA0);
+		dev_priv->regfile.saveFPA1 = I915_READ(_FPA1);
+		dev_priv->regfile.saveDPLL_A = I915_READ(_DPLL_A);
+	}
+	if (INTEL_INFO(dev)->gen >= 4 && !HAS_PCH_SPLIT(dev))
+		dev_priv->regfile.saveDPLL_A_MD = I915_READ(_DPLL_A_MD);
+	dev_priv->regfile.saveHTOTAL_A = I915_READ(_HTOTAL_A);
+	dev_priv->regfile.saveHBLANK_A = I915_READ(_HBLANK_A);
+	dev_priv->regfile.saveHSYNC_A = I915_READ(_HSYNC_A);
+	dev_priv->regfile.saveVTOTAL_A = I915_READ(_VTOTAL_A);
+	dev_priv->regfile.saveVBLANK_A = I915_READ(_VBLANK_A);
+	dev_priv->regfile.saveVSYNC_A = I915_READ(_VSYNC_A);
+	if (!HAS_PCH_SPLIT(dev))
+		dev_priv->regfile.saveBCLRPAT_A = I915_READ(_BCLRPAT_A);
+
+	if (HAS_PCH_SPLIT(dev)) {
+		dev_priv->regfile.savePIPEA_DATA_M1 = I915_READ(_PIPEA_DATA_M1);
+		dev_priv->regfile.savePIPEA_DATA_N1 = I915_READ(_PIPEA_DATA_N1);
+		dev_priv->regfile.savePIPEA_LINK_M1 = I915_READ(_PIPEA_LINK_M1);
+		dev_priv->regfile.savePIPEA_LINK_N1 = I915_READ(_PIPEA_LINK_N1);
+
+		dev_priv->regfile.saveFDI_TXA_CTL = I915_READ(_FDI_TXA_CTL);
+		dev_priv->regfile.saveFDI_RXA_CTL = I915_READ(_FDI_RXA_CTL);
+
+		dev_priv->regfile.savePFA_CTL_1 = I915_READ(_PFA_CTL_1);
+		dev_priv->regfile.savePFA_WIN_SZ = I915_READ(_PFA_WIN_SZ);
+		dev_priv->regfile.savePFA_WIN_POS = I915_READ(_PFA_WIN_POS);
+
+		dev_priv->regfile.saveTRANSACONF = I915_READ(_TRANSACONF);
+		dev_priv->regfile.saveTRANS_HTOTAL_A = I915_READ(_TRANS_HTOTAL_A);
+		dev_priv->regfile.saveTRANS_HBLANK_A = I915_READ(_TRANS_HBLANK_A);
+		dev_priv->regfile.saveTRANS_HSYNC_A = I915_READ(_TRANS_HSYNC_A);
+		dev_priv->regfile.saveTRANS_VTOTAL_A = I915_READ(_TRANS_VTOTAL_A);
+		dev_priv->regfile.saveTRANS_VBLANK_A = I915_READ(_TRANS_VBLANK_A);
+		dev_priv->regfile.saveTRANS_VSYNC_A = I915_READ(_TRANS_VSYNC_A);
+	}
+
+	dev_priv->regfile.saveDSPACNTR = I915_READ(_DSPACNTR);
+	dev_priv->regfile.saveDSPASTRIDE = I915_READ(_DSPASTRIDE);
+	dev_priv->regfile.saveDSPASIZE = I915_READ(_DSPASIZE);
+	dev_priv->regfile.saveDSPAPOS = I915_READ(_DSPAPOS);
+	dev_priv->regfile.saveDSPAADDR = I915_READ(_DSPAADDR);
+	if (INTEL_INFO(dev)->gen >= 4) {
+		dev_priv->regfile.saveDSPASURF = I915_READ(_DSPASURF);
+		dev_priv->regfile.saveDSPATILEOFF = I915_READ(_DSPATILEOFF);
+	}
+	i915_save_palette(dev, PIPE_A);
+	dev_priv->regfile.savePIPEASTAT = I915_READ(_PIPEASTAT);
+
+	/* Pipe & plane B info */
+	dev_priv->regfile.savePIPEBCONF = I915_READ(_PIPEBCONF);
+	dev_priv->regfile.savePIPEBSRC = I915_READ(_PIPEBSRC);
+	if (HAS_PCH_SPLIT(dev)) {
+		dev_priv->regfile.saveFPB0 = I915_READ(_PCH_FPB0);
+		dev_priv->regfile.saveFPB1 = I915_READ(_PCH_FPB1);
+		dev_priv->regfile.saveDPLL_B = I915_READ(_PCH_DPLL_B);
+	} else {
+		dev_priv->regfile.saveFPB0 = I915_READ(_FPB0);
+		dev_priv->regfile.saveFPB1 = I915_READ(_FPB1);
+		dev_priv->regfile.saveDPLL_B = I915_READ(_DPLL_B);
+	}
+	if (INTEL_INFO(dev)->gen >= 4 && !HAS_PCH_SPLIT(dev))
+		dev_priv->regfile.saveDPLL_B_MD = I915_READ(_DPLL_B_MD);
+	dev_priv->regfile.saveHTOTAL_B = I915_READ(_HTOTAL_B);
+	dev_priv->regfile.saveHBLANK_B = I915_READ(_HBLANK_B);
+	dev_priv->regfile.saveHSYNC_B = I915_READ(_HSYNC_B);
+	dev_priv->regfile.saveVTOTAL_B = I915_READ(_VTOTAL_B);
+	dev_priv->regfile.saveVBLANK_B = I915_READ(_VBLANK_B);
+	dev_priv->regfile.saveVSYNC_B = I915_READ(_VSYNC_B);
+	if (!HAS_PCH_SPLIT(dev))
+		dev_priv->regfile.saveBCLRPAT_B = I915_READ(_BCLRPAT_B);
+
+	if (HAS_PCH_SPLIT(dev)) {
+		dev_priv->regfile.savePIPEB_DATA_M1 = I915_READ(_PIPEB_DATA_M1);
+		dev_priv->regfile.savePIPEB_DATA_N1 = I915_READ(_PIPEB_DATA_N1);
+		dev_priv->regfile.savePIPEB_LINK_M1 = I915_READ(_PIPEB_LINK_M1);
+		dev_priv->regfile.savePIPEB_LINK_N1 = I915_READ(_PIPEB_LINK_N1);
+
+		dev_priv->regfile.saveFDI_TXB_CTL = I915_READ(_FDI_TXB_CTL);
+		dev_priv->regfile.saveFDI_RXB_CTL = I915_READ(_FDI_RXB_CTL);
+
+		dev_priv->regfile.savePFB_CTL_1 = I915_READ(_PFB_CTL_1);
+		dev_priv->regfile.savePFB_WIN_SZ = I915_READ(_PFB_WIN_SZ);
+		dev_priv->regfile.savePFB_WIN_POS = I915_READ(_PFB_WIN_POS);
+
+		dev_priv->regfile.saveTRANSBCONF = I915_READ(_TRANSBCONF);
+		dev_priv->regfile.saveTRANS_HTOTAL_B = I915_READ(_TRANS_HTOTAL_B);
+		dev_priv->regfile.saveTRANS_HBLANK_B = I915_READ(_TRANS_HBLANK_B);
+		dev_priv->regfile.saveTRANS_HSYNC_B = I915_READ(_TRANS_HSYNC_B);
+		dev_priv->regfile.saveTRANS_VTOTAL_B = I915_READ(_TRANS_VTOTAL_B);
+		dev_priv->regfile.saveTRANS_VBLANK_B = I915_READ(_TRANS_VBLANK_B);
+		dev_priv->regfile.saveTRANS_VSYNC_B = I915_READ(_TRANS_VSYNC_B);
+	}
+
+	dev_priv->regfile.saveDSPBCNTR = I915_READ(_DSPBCNTR);
+	dev_priv->regfile.saveDSPBSTRIDE = I915_READ(_DSPBSTRIDE);
+	dev_priv->regfile.saveDSPBSIZE = I915_READ(_DSPBSIZE);
+	dev_priv->regfile.saveDSPBPOS = I915_READ(_DSPBPOS);
+	dev_priv->regfile.saveDSPBADDR = I915_READ(_DSPBADDR);
+	if (INTEL_INFO(dev)->gen >= 4) {
+		dev_priv->regfile.saveDSPBSURF = I915_READ(_DSPBSURF);
+		dev_priv->regfile.saveDSPBTILEOFF = I915_READ(_DSPBTILEOFF);
+	}
+	i915_save_palette(dev, PIPE_B);
+	dev_priv->regfile.savePIPEBSTAT = I915_READ(_PIPEBSTAT);
+
+	/* Fences */
+	switch (INTEL_INFO(dev)->gen) {
+	case 7:
+	case 6:
+		for (i = 0; i < 16; i++)
+			dev_priv->regfile.saveFENCE[i] = I915_READ64(FENCE_REG_SANDYBRIDGE_0 + (i * 8));
+		break;
+	case 5:
+	case 4:
+		for (i = 0; i < 16; i++)
+			dev_priv->regfile.saveFENCE[i] = I915_READ64(FENCE_REG_965_0 + (i * 8));
+		break;
+	case 3:
+		if (IS_I945G(dev) || IS_I945GM(dev) || IS_G33(dev))
+			for (i = 0; i < 8; i++)
+				dev_priv->regfile.saveFENCE[i+8] = I915_READ(FENCE_REG_945_8 + (i * 4));
+	case 2:
+		for (i = 0; i < 8; i++)
+			dev_priv->regfile.saveFENCE[i] = I915_READ(FENCE_REG_830_0 + (i * 4));
+		break;
+	}
+
+	/* CRT state */
+	if (HAS_PCH_SPLIT(dev))
+		dev_priv->regfile.saveADPA = I915_READ(PCH_ADPA);
+	else
+		dev_priv->regfile.saveADPA = I915_READ(ADPA);
+
+	return;
+}
+
+void i915_restore_display_reg(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int dpll_a_reg, fpa0_reg, fpa1_reg;
+	int dpll_b_reg, fpb0_reg, fpb1_reg;
+	int i;
+
+	/* Display port ratios (must be done before clock is set) */
+	if (SUPPORTS_INTEGRATED_DP(dev)) {
+		I915_WRITE(_PIPEA_GMCH_DATA_M, dev_priv->regfile.savePIPEA_GMCH_DATA_M);
+		I915_WRITE(_PIPEB_GMCH_DATA_M, dev_priv->regfile.savePIPEB_GMCH_DATA_M);
+		I915_WRITE(_PIPEA_GMCH_DATA_N, dev_priv->regfile.savePIPEA_GMCH_DATA_N);
+		I915_WRITE(_PIPEB_GMCH_DATA_N, dev_priv->regfile.savePIPEB_GMCH_DATA_N);
+		I915_WRITE(_PIPEA_DP_LINK_M, dev_priv->regfile.savePIPEA_DP_LINK_M);
+		I915_WRITE(_PIPEB_DP_LINK_M, dev_priv->regfile.savePIPEB_DP_LINK_M);
+		I915_WRITE(_PIPEA_DP_LINK_N, dev_priv->regfile.savePIPEA_DP_LINK_N);
+		I915_WRITE(_PIPEB_DP_LINK_N, dev_priv->regfile.savePIPEB_DP_LINK_N);
+	}
+
+	/* Fences */
+	switch (INTEL_INFO(dev)->gen) {
+	case 7:
+	case 6:
+		for (i = 0; i < 16; i++)
+			I915_WRITE64(FENCE_REG_SANDYBRIDGE_0 + (i * 8), dev_priv->regfile.saveFENCE[i]);
+		break;
+	case 5:
+	case 4:
+		for (i = 0; i < 16; i++)
+			I915_WRITE64(FENCE_REG_965_0 + (i * 8), dev_priv->regfile.saveFENCE[i]);
+		break;
+	case 3:
+	case 2:
+		if (IS_I945G(dev) || IS_I945GM(dev) || IS_G33(dev))
+			for (i = 0; i < 8; i++)
+				I915_WRITE(FENCE_REG_945_8 + (i * 4), dev_priv->regfile.saveFENCE[i+8]);
+		for (i = 0; i < 8; i++)
+			I915_WRITE(FENCE_REG_830_0 + (i * 4), dev_priv->regfile.saveFENCE[i]);
+		break;
+	}
+
+
+	if (HAS_PCH_SPLIT(dev)) {
+		dpll_a_reg = _PCH_DPLL_A;
+		dpll_b_reg = _PCH_DPLL_B;
+		fpa0_reg = _PCH_FPA0;
+		fpb0_reg = _PCH_FPB0;
+		fpa1_reg = _PCH_FPA1;
+		fpb1_reg = _PCH_FPB1;
+	} else {
+		dpll_a_reg = _DPLL_A;
+		dpll_b_reg = _DPLL_B;
+		fpa0_reg = _FPA0;
+		fpb0_reg = _FPB0;
+		fpa1_reg = _FPA1;
+		fpb1_reg = _FPB1;
+	}
+
+	if (HAS_PCH_SPLIT(dev)) {
+		I915_WRITE(PCH_DREF_CONTROL, dev_priv->regfile.savePCH_DREF_CONTROL);
+		I915_WRITE(DISP_ARB_CTL, dev_priv->regfile.saveDISP_ARB_CTL);
+	}
+
+	/* Pipe & plane A info */
+	/* Prime the clock */
+	if (dev_priv->regfile.saveDPLL_A & DPLL_VCO_ENABLE) {
+		I915_WRITE(dpll_a_reg, dev_priv->regfile.saveDPLL_A &
+			   ~DPLL_VCO_ENABLE);
+		POSTING_READ(dpll_a_reg);
+		udelay(150);
+	}
+	I915_WRITE(fpa0_reg, dev_priv->regfile.saveFPA0);
+	I915_WRITE(fpa1_reg, dev_priv->regfile.saveFPA1);
+	/* Actually enable it */
+	I915_WRITE(dpll_a_reg, dev_priv->regfile.saveDPLL_A);
+	POSTING_READ(dpll_a_reg);
+	udelay(150);
+	if (INTEL_INFO(dev)->gen >= 4 && !HAS_PCH_SPLIT(dev)) {
+		I915_WRITE(_DPLL_A_MD, dev_priv->regfile.saveDPLL_A_MD);
+		POSTING_READ(_DPLL_A_MD);
+	}
+	udelay(150);
+
+	/* Restore mode */
+	I915_WRITE(_HTOTAL_A, dev_priv->regfile.saveHTOTAL_A);
+	I915_WRITE(_HBLANK_A, dev_priv->regfile.saveHBLANK_A);
+	I915_WRITE(_HSYNC_A, dev_priv->regfile.saveHSYNC_A);
+	I915_WRITE(_VTOTAL_A, dev_priv->regfile.saveVTOTAL_A);
+	I915_WRITE(_VBLANK_A, dev_priv->regfile.saveVBLANK_A);
+	I915_WRITE(_VSYNC_A, dev_priv->regfile.saveVSYNC_A);
+	if (!HAS_PCH_SPLIT(dev))
+		I915_WRITE(_BCLRPAT_A, dev_priv->regfile.saveBCLRPAT_A);
+
+	if (HAS_PCH_SPLIT(dev)) {
+		I915_WRITE(_PIPEA_DATA_M1, dev_priv->regfile.savePIPEA_DATA_M1);
+		I915_WRITE(_PIPEA_DATA_N1, dev_priv->regfile.savePIPEA_DATA_N1);
+		I915_WRITE(_PIPEA_LINK_M1, dev_priv->regfile.savePIPEA_LINK_M1);
+		I915_WRITE(_PIPEA_LINK_N1, dev_priv->regfile.savePIPEA_LINK_N1);
+
+		I915_WRITE(_FDI_RXA_CTL, dev_priv->regfile.saveFDI_RXA_CTL);
+		I915_WRITE(_FDI_TXA_CTL, dev_priv->regfile.saveFDI_TXA_CTL);
+
+		I915_WRITE(_PFA_CTL_1, dev_priv->regfile.savePFA_CTL_1);
+		I915_WRITE(_PFA_WIN_SZ, dev_priv->regfile.savePFA_WIN_SZ);
+		I915_WRITE(_PFA_WIN_POS, dev_priv->regfile.savePFA_WIN_POS);
+
+		I915_WRITE(_TRANSACONF, dev_priv->regfile.saveTRANSACONF);
+		I915_WRITE(_TRANS_HTOTAL_A, dev_priv->regfile.saveTRANS_HTOTAL_A);
+		I915_WRITE(_TRANS_HBLANK_A, dev_priv->regfile.saveTRANS_HBLANK_A);
+		I915_WRITE(_TRANS_HSYNC_A, dev_priv->regfile.saveTRANS_HSYNC_A);
+		I915_WRITE(_TRANS_VTOTAL_A, dev_priv->regfile.saveTRANS_VTOTAL_A);
+		I915_WRITE(_TRANS_VBLANK_A, dev_priv->regfile.saveTRANS_VBLANK_A);
+		I915_WRITE(_TRANS_VSYNC_A, dev_priv->regfile.saveTRANS_VSYNC_A);
+	}
+
+	/* Restore plane info */
+	I915_WRITE(_DSPASIZE, dev_priv->regfile.saveDSPASIZE);
+	I915_WRITE(_DSPAPOS, dev_priv->regfile.saveDSPAPOS);
+	I915_WRITE(_PIPEASRC, dev_priv->regfile.savePIPEASRC);
+	I915_WRITE(_DSPAADDR, dev_priv->regfile.saveDSPAADDR);
+	I915_WRITE(_DSPASTRIDE, dev_priv->regfile.saveDSPASTRIDE);
+	if (INTEL_INFO(dev)->gen >= 4) {
+		I915_WRITE(_DSPASURF, dev_priv->regfile.saveDSPASURF);
+		I915_WRITE(_DSPATILEOFF, dev_priv->regfile.saveDSPATILEOFF);
+	}
+
+	I915_WRITE(_PIPEACONF, dev_priv->regfile.savePIPEACONF);
+
+	i915_restore_palette(dev, PIPE_A);
+	/* Enable the plane */
+	I915_WRITE(_DSPACNTR, dev_priv->regfile.saveDSPACNTR);
+	I915_WRITE(_DSPAADDR, I915_READ(_DSPAADDR));
+
+	/* Pipe & plane B info */
+	if (dev_priv->regfile.saveDPLL_B & DPLL_VCO_ENABLE) {
+		I915_WRITE(dpll_b_reg, dev_priv->regfile.saveDPLL_B &
+			   ~DPLL_VCO_ENABLE);
+		POSTING_READ(dpll_b_reg);
+		udelay(150);
+	}
+	I915_WRITE(fpb0_reg, dev_priv->regfile.saveFPB0);
+	I915_WRITE(fpb1_reg, dev_priv->regfile.saveFPB1);
+	/* Actually enable it */
+	I915_WRITE(dpll_b_reg, dev_priv->regfile.saveDPLL_B);
+	POSTING_READ(dpll_b_reg);
+	udelay(150);
+	if (INTEL_INFO(dev)->gen >= 4 && !HAS_PCH_SPLIT(dev)) {
+		I915_WRITE(_DPLL_B_MD, dev_priv->regfile.saveDPLL_B_MD);
+		POSTING_READ(_DPLL_B_MD);
+	}
+	udelay(150);
+
+	/* Restore mode */
+	I915_WRITE(_HTOTAL_B, dev_priv->regfile.saveHTOTAL_B);
+	I915_WRITE(_HBLANK_B, dev_priv->regfile.saveHBLANK_B);
+	I915_WRITE(_HSYNC_B, dev_priv->regfile.saveHSYNC_B);
+	I915_WRITE(_VTOTAL_B, dev_priv->regfile.saveVTOTAL_B);
+	I915_WRITE(_VBLANK_B, dev_priv->regfile.saveVBLANK_B);
+	I915_WRITE(_VSYNC_B, dev_priv->regfile.saveVSYNC_B);
+	if (!HAS_PCH_SPLIT(dev))
+		I915_WRITE(_BCLRPAT_B, dev_priv->regfile.saveBCLRPAT_B);
+
+	if (HAS_PCH_SPLIT(dev)) {
+		I915_WRITE(_PIPEB_DATA_M1, dev_priv->regfile.savePIPEB_DATA_M1);
+		I915_WRITE(_PIPEB_DATA_N1, dev_priv->regfile.savePIPEB_DATA_N1);
+		I915_WRITE(_PIPEB_LINK_M1, dev_priv->regfile.savePIPEB_LINK_M1);
+		I915_WRITE(_PIPEB_LINK_N1, dev_priv->regfile.savePIPEB_LINK_N1);
+
+		I915_WRITE(_FDI_RXB_CTL, dev_priv->regfile.saveFDI_RXB_CTL);
+		I915_WRITE(_FDI_TXB_CTL, dev_priv->regfile.saveFDI_TXB_CTL);
+
+		I915_WRITE(_PFB_CTL_1, dev_priv->regfile.savePFB_CTL_1);
+		I915_WRITE(_PFB_WIN_SZ, dev_priv->regfile.savePFB_WIN_SZ);
+		I915_WRITE(_PFB_WIN_POS, dev_priv->regfile.savePFB_WIN_POS);
+
+		I915_WRITE(_TRANSBCONF, dev_priv->regfile.saveTRANSBCONF);
+		I915_WRITE(_TRANS_HTOTAL_B, dev_priv->regfile.saveTRANS_HTOTAL_B);
+		I915_WRITE(_TRANS_HBLANK_B, dev_priv->regfile.saveTRANS_HBLANK_B);
+		I915_WRITE(_TRANS_HSYNC_B, dev_priv->regfile.saveTRANS_HSYNC_B);
+		I915_WRITE(_TRANS_VTOTAL_B, dev_priv->regfile.saveTRANS_VTOTAL_B);
+		I915_WRITE(_TRANS_VBLANK_B, dev_priv->regfile.saveTRANS_VBLANK_B);
+		I915_WRITE(_TRANS_VSYNC_B, dev_priv->regfile.saveTRANS_VSYNC_B);
+	}
+
+	/* Restore plane info */
+	I915_WRITE(_DSPBSIZE, dev_priv->regfile.saveDSPBSIZE);
+	I915_WRITE(_DSPBPOS, dev_priv->regfile.saveDSPBPOS);
+	I915_WRITE(_PIPEBSRC, dev_priv->regfile.savePIPEBSRC);
+	I915_WRITE(_DSPBADDR, dev_priv->regfile.saveDSPBADDR);
+	I915_WRITE(_DSPBSTRIDE, dev_priv->regfile.saveDSPBSTRIDE);
+	if (INTEL_INFO(dev)->gen >= 4) {
+		I915_WRITE(_DSPBSURF, dev_priv->regfile.saveDSPBSURF);
+		I915_WRITE(_DSPBTILEOFF, dev_priv->regfile.saveDSPBTILEOFF);
+	}
+
+	I915_WRITE(_PIPEBCONF, dev_priv->regfile.savePIPEBCONF);
+
+	i915_restore_palette(dev, PIPE_B);
+	/* Enable the plane */
+	I915_WRITE(_DSPBCNTR, dev_priv->regfile.saveDSPBCNTR);
+	I915_WRITE(_DSPBADDR, I915_READ(_DSPBADDR));
+
+	/* Cursor state */
+	I915_WRITE(_CURAPOS, dev_priv->regfile.saveCURAPOS);
+	I915_WRITE(_CURACNTR, dev_priv->regfile.saveCURACNTR);
+	I915_WRITE(_CURABASE, dev_priv->regfile.saveCURABASE);
+	I915_WRITE(_CURBPOS, dev_priv->regfile.saveCURBPOS);
+	I915_WRITE(_CURBCNTR, dev_priv->regfile.saveCURBCNTR);
+	I915_WRITE(_CURBBASE, dev_priv->regfile.saveCURBBASE);
+	if (IS_GEN2(dev))
+		I915_WRITE(CURSIZE, dev_priv->regfile.saveCURSIZE);
+
+	/* CRT state */
+	if (HAS_PCH_SPLIT(dev))
+		I915_WRITE(PCH_ADPA, dev_priv->regfile.saveADPA);
+	else
+		I915_WRITE(ADPA, dev_priv->regfile.saveADPA);
+
+	return;
+}
-- 
cgit v0.10.2


From 44cec74040564cba2ace8c3756d2fc908bc7a373 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Fri, 25 Jan 2013 17:53:21 +0100
Subject: drm/i915: dont save/restore VGA state for kms

The only thing we really care about that it is off. To do so, reuse
the recently created i915_redisable_vga function, which is already
used to put obnoxious firmware into check on lid reopening.

Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index a263a0c..bffe222 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -1789,6 +1789,7 @@ extern void intel_modeset_cleanup(struct drm_device *dev);
 extern int intel_modeset_vga_set_state(struct drm_device *dev, bool state);
 extern void intel_modeset_setup_hw_state(struct drm_device *dev,
 					 bool force_restore);
+extern void i915_redisable_vga(struct drm_device *dev);
 extern bool intel_fbc_enabled(struct drm_device *dev);
 extern void intel_disable_fbc(struct drm_device *dev);
 extern bool ironlake_set_drps(struct drm_device *dev, u8 val);
diff --git a/drivers/gpu/drm/i915/i915_suspend.c b/drivers/gpu/drm/i915/i915_suspend.c
index 1f581fc..3b9baea 100644
--- a/drivers/gpu/drm/i915/i915_suspend.c
+++ b/drivers/gpu/drm/i915/i915_suspend.c
@@ -69,6 +69,15 @@ static void i915_save_vga(struct drm_device *dev)
 	int i;
 	u16 cr_index, cr_data, st01;
 
+	/* VGA state */
+	dev_priv->regfile.saveVGA0 = I915_READ(VGA0);
+	dev_priv->regfile.saveVGA1 = I915_READ(VGA1);
+	dev_priv->regfile.saveVGA_PD = I915_READ(VGA_PD);
+	if (HAS_PCH_SPLIT(dev))
+		dev_priv->regfile.saveVGACNTRL = I915_READ(CPU_VGACNTRL);
+	else
+		dev_priv->regfile.saveVGACNTRL = I915_READ(VGACNTRL);
+
 	/* VGA color palette registers */
 	dev_priv->regfile.saveDACMASK = I915_READ8(VGA_DACMASK);
 
@@ -127,6 +136,18 @@ static void i915_restore_vga(struct drm_device *dev)
 	int i;
 	u16 cr_index, cr_data, st01;
 
+	/* VGA state */
+	if (HAS_PCH_SPLIT(dev))
+		I915_WRITE(CPU_VGACNTRL, dev_priv->regfile.saveVGACNTRL);
+	else
+		I915_WRITE(VGACNTRL, dev_priv->regfile.saveVGACNTRL);
+
+	I915_WRITE(VGA0, dev_priv->regfile.saveVGA0);
+	I915_WRITE(VGA1, dev_priv->regfile.saveVGA1);
+	I915_WRITE(VGA_PD, dev_priv->regfile.saveVGA_PD);
+	POSTING_READ(VGA_PD);
+	udelay(150);
+
 	/* MSR bits */
 	I915_WRITE8(VGA_MSR_WRITE, dev_priv->regfile.saveMSR);
 	if (dev_priv->regfile.saveMSR & VGA_MSR_CGA_MODE) {
@@ -251,16 +272,8 @@ static void i915_save_display(struct drm_device *dev)
 		}
 	}
 
-	/* VGA state */
-	dev_priv->regfile.saveVGA0 = I915_READ(VGA0);
-	dev_priv->regfile.saveVGA1 = I915_READ(VGA1);
-	dev_priv->regfile.saveVGA_PD = I915_READ(VGA_PD);
-	if (HAS_PCH_SPLIT(dev))
-		dev_priv->regfile.saveVGACNTRL = I915_READ(CPU_VGACNTRL);
-	else
-		dev_priv->regfile.saveVGACNTRL = I915_READ(VGACNTRL);
-
-	i915_save_vga(dev);
+	if (!drm_core_check_feature(dev, DRIVER_MODESET))
+		i915_save_vga(dev);
 }
 
 static void i915_restore_display(struct drm_device *dev)
@@ -334,19 +347,10 @@ static void i915_restore_display(struct drm_device *dev)
 			I915_WRITE(FBC_CONTROL, dev_priv->regfile.saveFBC_CONTROL);
 		}
 	}
-	/* VGA state */
-	if (HAS_PCH_SPLIT(dev))
-		I915_WRITE(CPU_VGACNTRL, dev_priv->regfile.saveVGACNTRL);
+	if (!drm_core_check_feature(dev, DRIVER_MODESET))
+		i915_restore_vga(dev);
 	else
-		I915_WRITE(VGACNTRL, dev_priv->regfile.saveVGACNTRL);
-
-	I915_WRITE(VGA0, dev_priv->regfile.saveVGA0);
-	I915_WRITE(VGA1, dev_priv->regfile.saveVGA1);
-	I915_WRITE(VGA_PD, dev_priv->regfile.saveVGA_PD);
-	POSTING_READ(VGA_PD);
-	udelay(150);
-
-	i915_restore_vga(dev);
+		i915_redisable_vga(dev);
 }
 
 int i915_save_state(struct drm_device *dev)
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index c0fadf7..56c51dd 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -8903,7 +8903,7 @@ static void intel_sanitize_encoder(struct intel_encoder *encoder)
 	 * the crtc fixup. */
 }
 
-static void i915_redisable_vga(struct drm_device *dev)
+void i915_redisable_vga(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 vga_reg;
-- 
cgit v0.10.2


From a65e827dd57b2fd48a698209dd7701eb13e72095 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Fri, 25 Jan 2013 17:53:22 +0100
Subject: drm/i915: move DP save/restore into i915_ums.c

Note that this slightly changes the order, but we only move it within
the block of registers that restore encoder state. Specifically LVDS
is now restored after DP, whereas previously it was done before.

Legacy vga is still restored afterwards, which seems to be the
important thing (if there's anything important in this restore
ordering at all).

Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_suspend.c b/drivers/gpu/drm/i915/i915_suspend.c
index 3b9baea..3911e04 100644
--- a/drivers/gpu/drm/i915/i915_suspend.c
+++ b/drivers/gpu/drm/i915/i915_suspend.c
@@ -240,24 +240,6 @@ static void i915_save_display(struct drm_device *dev)
 		dev_priv->regfile.savePP_DIVISOR = I915_READ(PP_DIVISOR);
 	}
 
-	if (!drm_core_check_feature(dev, DRIVER_MODESET)) {
-		/* Display Port state */
-		if (SUPPORTS_INTEGRATED_DP(dev)) {
-			dev_priv->regfile.saveDP_B = I915_READ(DP_B);
-			dev_priv->regfile.saveDP_C = I915_READ(DP_C);
-			dev_priv->regfile.saveDP_D = I915_READ(DP_D);
-			dev_priv->regfile.savePIPEA_GMCH_DATA_M = I915_READ(_PIPEA_GMCH_DATA_M);
-			dev_priv->regfile.savePIPEB_GMCH_DATA_M = I915_READ(_PIPEB_GMCH_DATA_M);
-			dev_priv->regfile.savePIPEA_GMCH_DATA_N = I915_READ(_PIPEA_GMCH_DATA_N);
-			dev_priv->regfile.savePIPEB_GMCH_DATA_N = I915_READ(_PIPEB_GMCH_DATA_N);
-			dev_priv->regfile.savePIPEA_DP_LINK_M = I915_READ(_PIPEA_DP_LINK_M);
-			dev_priv->regfile.savePIPEB_DP_LINK_M = I915_READ(_PIPEB_DP_LINK_M);
-			dev_priv->regfile.savePIPEA_DP_LINK_N = I915_READ(_PIPEA_DP_LINK_N);
-			dev_priv->regfile.savePIPEB_DP_LINK_N = I915_READ(_PIPEB_DP_LINK_N);
-		}
-		/* FIXME: regfile.save TV & SDVO state */
-	}
-
 	/* Only regfile.save FBC state on the platform that supports FBC */
 	if (I915_HAS_FBC(dev)) {
 		if (HAS_PCH_SPLIT(dev)) {
@@ -323,16 +305,6 @@ static void i915_restore_display(struct drm_device *dev)
 		I915_WRITE(PP_CONTROL, dev_priv->regfile.savePP_CONTROL);
 	}
 
-	if (!drm_core_check_feature(dev, DRIVER_MODESET)) {
-		/* Display Port state */
-		if (SUPPORTS_INTEGRATED_DP(dev)) {
-			I915_WRITE(DP_B, dev_priv->regfile.saveDP_B);
-			I915_WRITE(DP_C, dev_priv->regfile.saveDP_C);
-			I915_WRITE(DP_D, dev_priv->regfile.saveDP_D);
-		}
-		/* FIXME: restore TV & SDVO state */
-	}
-
 	/* only restore FBC info on the platform that supports FBC*/
 	intel_disable_fbc(dev);
 	if (I915_HAS_FBC(dev)) {
@@ -347,6 +319,7 @@ static void i915_restore_display(struct drm_device *dev)
 			I915_WRITE(FBC_CONTROL, dev_priv->regfile.saveFBC_CONTROL);
 		}
 	}
+
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		i915_restore_vga(dev);
 	else
diff --git a/drivers/gpu/drm/i915/i915_ums.c b/drivers/gpu/drm/i915/i915_ums.c
index 359c77b..985a097 100644
--- a/drivers/gpu/drm/i915/i915_ums.c
+++ b/drivers/gpu/drm/i915/i915_ums.c
@@ -254,6 +254,22 @@ void i915_save_display_reg(struct drm_device *dev)
 	else
 		dev_priv->regfile.saveADPA = I915_READ(ADPA);
 
+	/* Display Port state */
+	if (SUPPORTS_INTEGRATED_DP(dev)) {
+		dev_priv->regfile.saveDP_B = I915_READ(DP_B);
+		dev_priv->regfile.saveDP_C = I915_READ(DP_C);
+		dev_priv->regfile.saveDP_D = I915_READ(DP_D);
+		dev_priv->regfile.savePIPEA_GMCH_DATA_M = I915_READ(_PIPEA_GMCH_DATA_M);
+		dev_priv->regfile.savePIPEB_GMCH_DATA_M = I915_READ(_PIPEB_GMCH_DATA_M);
+		dev_priv->regfile.savePIPEA_GMCH_DATA_N = I915_READ(_PIPEA_GMCH_DATA_N);
+		dev_priv->regfile.savePIPEB_GMCH_DATA_N = I915_READ(_PIPEB_GMCH_DATA_N);
+		dev_priv->regfile.savePIPEA_DP_LINK_M = I915_READ(_PIPEA_DP_LINK_M);
+		dev_priv->regfile.savePIPEB_DP_LINK_M = I915_READ(_PIPEB_DP_LINK_M);
+		dev_priv->regfile.savePIPEA_DP_LINK_N = I915_READ(_PIPEA_DP_LINK_N);
+		dev_priv->regfile.savePIPEB_DP_LINK_N = I915_READ(_PIPEB_DP_LINK_N);
+	}
+	/* FIXME: regfile.save TV & SDVO state */
+
 	return;
 }
 
@@ -475,5 +491,13 @@ void i915_restore_display_reg(struct drm_device *dev)
 	else
 		I915_WRITE(ADPA, dev_priv->regfile.saveADPA);
 
+	/* Display Port state */
+	if (SUPPORTS_INTEGRATED_DP(dev)) {
+		I915_WRITE(DP_B, dev_priv->regfile.saveDP_B);
+		I915_WRITE(DP_C, dev_priv->regfile.saveDP_C);
+		I915_WRITE(DP_D, dev_priv->regfile.saveDP_D);
+	}
+	/* FIXME: restore TV & SDVO state */
+
 	return;
 }
-- 
cgit v0.10.2


From 2f86f1916504525a6fdd6b412374b4ebf1102cbe Mon Sep 17 00:00:00 2001
From: Ben Widawsky <ben@bwidawsk.net>
Date: Mon, 28 Jan 2013 15:32:15 -0800
Subject: drm/i915: Error state should print /sys/kernel/debug

/sys/kernel/debug has more or less been the standard location of debugfs
for several years now. Other parts of DRM already use this location, so
we should as well.

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Carl Worth <cworth@cworth.org>
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
[danvet: split up long line.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 943db10..5648d84 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -1305,7 +1305,8 @@ static void i915_capture_error_state(struct drm_device *dev)
 		return;
 	}
 
-	DRM_INFO("capturing error event; look for more information in /debug/dri/%d/i915_error_state\n",
+	DRM_INFO("capturing error event; look for more information in"
+		 "/sys/kernel/debug/dri/%d/i915_error_state\n",
 		 dev->primary->index);
 
 	kref_init(&error->ref);
-- 
cgit v0.10.2


From 7faf1ab2ff816d6ea9cde099e794a4e18cf13bbc Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 24 Jan 2013 14:44:55 -0800
Subject: drm/i915: vfuncs for gtt_clear_range/insert_entries

We have a few too many differences here, so finally take the prepared
abstraction and run with it. A few smaller changes are required to get
things into shape:

- move i915_cache_level up since we need it in the gt funcs
- split up i915_ggtt_clear_range and move the two functions down to
  where the relevant insert_entries functions are
- adjustments to a few function parameter lists

Now we have 2 functions which deal with the gen6+ global gtt
(gen6_ggtt_ prefix) and 2 functions which deal with the legacy gtt
code in the intel-gtt.c fake agp driver (i915_ggtt_ prefix).

Init is still a bit a mess, but honestly I don't care about that.

One thing I've thought about while deciding on the exact interfaces is
a flag parameter for ->clear_range: We could use that to decide
between writing invalid pte entries or scratch pte entries. In case we
ever get around to fixing all our bugs which currently prevent us from
filling the gtt with empty ptes for the truly unused ranges ...

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
[bwidawsk: Moved functions to the gtt struct]
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index bffe222..fb0815b 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -365,6 +365,12 @@ struct intel_device_info {
 	u8 has_llc:1;
 };
 
+enum i915_cache_level {
+	I915_CACHE_NONE = 0,
+	I915_CACHE_LLC,
+	I915_CACHE_LLC_MLC, /* gen6+, in docs at least! */
+};
+
 /* The Graphics Translation Table is the way in which GEN hardware translates a
  * Graphics Virtual Address into a Physical Address. In addition to the normal
  * collateral associated with any va->pa translations GEN hardware also has a
@@ -386,6 +392,15 @@ struct i915_gtt {
 	bool do_idle_maps;
 	dma_addr_t scratch_page_dma;
 	struct page *scratch_page;
+
+	/* global gtt ops */
+	void (*gtt_clear_range)(struct drm_device *dev,
+				unsigned int first_entry,
+				unsigned int num_entries);
+	void (*gtt_insert_entries)(struct drm_device *dev,
+				   struct sg_table *st,
+				   unsigned int pg_start,
+				   enum i915_cache_level cache_level);
 };
 
 #define I915_PPGTT_PD_ENTRIES 512
@@ -1023,12 +1038,6 @@ enum hdmi_force_audio {
 	HDMI_AUDIO_ON,			/* force turn on HDMI audio */
 };
 
-enum i915_cache_level {
-	I915_CACHE_NONE = 0,
-	I915_CACHE_LLC,
-	I915_CACHE_LLC_MLC, /* gen6+, in docs at least! */
-};
-
 #define I915_GTT_RESERVED ((struct drm_mm_node *)0x1)
 
 struct drm_i915_gem_object_ops {
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index a0ba4a9..4712626 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -367,40 +367,14 @@ static void undo_idling(struct drm_i915_private *dev_priv, bool interruptible)
 		dev_priv->mm.interruptible = interruptible;
 }
 
-static void i915_ggtt_clear_range(struct drm_device *dev,
-				 unsigned first_entry,
-				 unsigned num_entries)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	gtt_pte_t scratch_pte;
-	gtt_pte_t __iomem *gtt_base = (gtt_pte_t __iomem *) dev_priv->gtt.gsm + first_entry;
-	const int max_entries = dev_priv->mm.gtt->gtt_total_entries - first_entry;
-	int i;
-
-	if (INTEL_INFO(dev)->gen < 6) {
-		intel_gtt_clear_range(first_entry, num_entries);
-		return;
-	}
-
-	if (WARN(num_entries > max_entries,
-		 "First entry = %d; Num entries = %d (max=%d)\n",
-		 first_entry, num_entries, max_entries))
-		num_entries = max_entries;
-
-	scratch_pte = pte_encode(dev, dev_priv->gtt.scratch_page_dma, I915_CACHE_LLC);
-	for (i = 0; i < num_entries; i++)
-		iowrite32(scratch_pte, &gtt_base[i]);
-	readl(gtt_base);
-}
-
 void i915_gem_restore_gtt_mappings(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_i915_gem_object *obj;
 
 	/* First fill our portion of the GTT with scratch pages */
-	i915_ggtt_clear_range(dev, dev_priv->gtt.start / PAGE_SIZE,
-			      dev_priv->gtt.total / PAGE_SIZE);
+	dev_priv->gtt.gtt_clear_range(dev, dev_priv->gtt.start / PAGE_SIZE,
+				      dev_priv->gtt.total / PAGE_SIZE);
 
 	list_for_each_entry(obj, &dev_priv->mm.bound_list, gtt_list) {
 		i915_gem_clflush_object(obj);
@@ -429,15 +403,13 @@ int i915_gem_gtt_prepare_object(struct drm_i915_gem_object *obj)
  * within the global GTT as well as accessible by the GPU through the GMADR
  * mapped BAR (dev_priv->mm.gtt->gtt).
  */
-static void gen6_ggtt_bind_object(struct drm_i915_gem_object *obj,
-				  enum i915_cache_level level)
+static void gen6_ggtt_insert_entries(struct drm_device *dev,
+				     struct sg_table *st,
+				     unsigned int first_entry,
+				     enum i915_cache_level level)
 {
-	struct drm_device *dev = obj->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct sg_table *st = obj->pages;
 	struct scatterlist *sg = st->sgl;
-	const int first_entry = obj->gtt_space->start >> PAGE_SHIFT;
-	const int max_entries = dev_priv->mm.gtt->gtt_total_entries - first_entry;
 	gtt_pte_t __iomem *gtt_entries =
 		(gtt_pte_t __iomem *)dev_priv->gtt.gsm + first_entry;
 	int unused, i = 0;
@@ -453,9 +425,6 @@ static void gen6_ggtt_bind_object(struct drm_i915_gem_object *obj,
 		}
 	}
 
-	BUG_ON(i > max_entries);
-	BUG_ON(i != obj->base.size / PAGE_SIZE);
-
 	/* XXX: This serves as a posting read to make sure that the PTE has
 	 * actually been updated. There is some concern that even though
 	 * registers and PTEs are within the same BAR that they are potentially
@@ -473,28 +442,69 @@ static void gen6_ggtt_bind_object(struct drm_i915_gem_object *obj,
 	POSTING_READ(GFX_FLSH_CNTL_GEN6);
 }
 
+static void gen6_ggtt_clear_range(struct drm_device *dev,
+				  unsigned int first_entry,
+				  unsigned int num_entries)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	gtt_pte_t scratch_pte;
+	gtt_pte_t __iomem *gtt_base = (gtt_pte_t __iomem *) dev_priv->gtt.gsm + first_entry;
+	const int max_entries = dev_priv->mm.gtt->gtt_total_entries - first_entry;
+	int i;
+
+	if (WARN(num_entries > max_entries,
+		 "First entry = %d; Num entries = %d (max=%d)\n",
+		 first_entry, num_entries, max_entries))
+		num_entries = max_entries;
+
+	scratch_pte = pte_encode(dev, dev_priv->gtt.scratch_page_dma, I915_CACHE_LLC);
+	for (i = 0; i < num_entries; i++)
+		iowrite32(scratch_pte, &gtt_base[i]);
+	readl(gtt_base);
+}
+
+
+static void i915_ggtt_insert_entries(struct drm_device *dev,
+				     struct sg_table *st,
+				     unsigned int pg_start,
+				     enum i915_cache_level cache_level)
+{
+	unsigned int flags = (cache_level == I915_CACHE_NONE) ?
+		AGP_USER_MEMORY : AGP_USER_CACHED_MEMORY;
+
+	intel_gtt_insert_sg_entries(st, pg_start, flags);
+
+}
+
+static void i915_ggtt_clear_range(struct drm_device *dev,
+				  unsigned int first_entry,
+				  unsigned int num_entries)
+{
+	intel_gtt_clear_range(first_entry, num_entries);
+}
+
+
 void i915_gem_gtt_bind_object(struct drm_i915_gem_object *obj,
 			      enum i915_cache_level cache_level)
 {
 	struct drm_device *dev = obj->base.dev;
-	if (INTEL_INFO(dev)->gen < 6) {
-		unsigned int flags = (cache_level == I915_CACHE_NONE) ?
-			AGP_USER_MEMORY : AGP_USER_CACHED_MEMORY;
-		intel_gtt_insert_sg_entries(obj->pages,
-					    obj->gtt_space->start >> PAGE_SHIFT,
-					    flags);
-	} else {
-		gen6_ggtt_bind_object(obj, cache_level);
-	}
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	dev_priv->gtt.gtt_insert_entries(dev, obj->pages,
+					 obj->gtt_space->start >> PAGE_SHIFT,
+					 cache_level);
 
 	obj->has_global_gtt_mapping = 1;
 }
 
 void i915_gem_gtt_unbind_object(struct drm_i915_gem_object *obj)
 {
-	i915_ggtt_clear_range(obj->base.dev,
-			      obj->gtt_space->start >> PAGE_SHIFT,
-			      obj->base.size >> PAGE_SHIFT);
+	struct drm_device *dev = obj->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	dev_priv->gtt.gtt_clear_range(obj->base.dev,
+				      obj->gtt_space->start >> PAGE_SHIFT,
+				      obj->base.size >> PAGE_SHIFT);
 
 	obj->has_global_gtt_mapping = 0;
 }
@@ -570,13 +580,12 @@ void i915_gem_setup_global_gtt(struct drm_device *dev,
 			     hole_start, hole_end) {
 		DRM_DEBUG_KMS("clearing unused GTT space: [%lx, %lx]\n",
 			      hole_start, hole_end);
-		i915_ggtt_clear_range(dev,
-				      hole_start / PAGE_SIZE,
-				      (hole_end-hole_start) / PAGE_SIZE);
+		dev_priv->gtt.gtt_clear_range(dev, hole_start / PAGE_SIZE,
+					      (hole_end-hole_start) / PAGE_SIZE);
 	}
 
 	/* And finally clear the reserved guard page */
-	i915_ggtt_clear_range(dev, end / PAGE_SIZE - 1, 1);
+	dev_priv->gtt.gtt_clear_range(dev, end / PAGE_SIZE - 1, 1);
 }
 
 static bool
@@ -718,6 +727,9 @@ int i915_gem_gtt_init(struct drm_device *dev)
 
 		dev_priv->gtt.do_idle_maps = needs_idle_maps(dev);
 
+		dev_priv->gtt.gtt_clear_range = i915_ggtt_clear_range;
+		dev_priv->gtt.gtt_insert_entries = i915_ggtt_insert_entries;
+
 		return 0;
 	}
 
@@ -771,6 +783,9 @@ int i915_gem_gtt_init(struct drm_device *dev)
 	DRM_DEBUG_DRIVER("GMADR size = %ldM\n", dev_priv->gtt.mappable_end >> 20);
 	DRM_DEBUG_DRIVER("GTT stolen size = %dM\n", dev_priv->mm.gtt->stolen_size >> 20);
 
+	dev_priv->gtt.gtt_clear_range = gen6_ggtt_clear_range;
+	dev_priv->gtt.gtt_insert_entries = gen6_ggtt_insert_entries;
+
 	return 0;
 
 err_out:
-- 
cgit v0.10.2


From def886c3768d24c4e0aa56ff98b5a468c2b5c9bf Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 24 Jan 2013 14:44:56 -0800
Subject: drm/i915: vfuncs for ppgtt

Like for the global gtt we want a notch more flexibility here. Only
big change (besides a few tiny function parameter adjustments) was to
move gen6_ppgtt_insert_entries up (and remove _sg_ from its name, we
only have one kind of insert_entries since the last gtt cleanup).

We could also extract the platform ppgtt setup/teardown code a bit
better, but I don't care that much.

With this we have the hw details of pte writing nicely hidden away
behind a bit of abstraction. Which should pave the way for
different/multiple ppgtts (e.g. what we need for real ppgtt support).

Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index fb0815b..4be7c68 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -412,6 +412,15 @@ struct i915_hw_ppgtt {
 	uint32_t pd_offset;
 	dma_addr_t *pt_dma_addr;
 	dma_addr_t scratch_page_dma_addr;
+
+	/* pte functions, mirroring the interface of the global gtt. */
+	void (*clear_range)(struct i915_hw_ppgtt *ppgtt,
+			    unsigned int first_entry,
+			    unsigned int num_entries);
+	void (*insert_entries)(struct i915_hw_ppgtt *ppgtt,
+			       struct sg_table *st,
+			       unsigned int pg_start,
+			       enum i915_cache_level cache_level);
 };
 
 
@@ -1662,7 +1671,6 @@ int i915_gem_context_destroy_ioctl(struct drm_device *dev, void *data,
 				   struct drm_file *file);
 
 /* i915_gem_gtt.c */
-int __must_check i915_gem_init_aliasing_ppgtt(struct drm_device *dev);
 void i915_gem_cleanup_aliasing_ppgtt(struct drm_device *dev);
 void i915_ppgtt_bind_object(struct i915_hw_ppgtt *ppgtt,
 			    struct drm_i915_gem_object *obj,
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 4712626..f63dbc7 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -77,7 +77,7 @@ static inline gtt_pte_t pte_encode(struct drm_device *dev,
 }
 
 /* PPGTT support for Sandybdrige/Gen6 and later */
-static void i915_ppgtt_clear_range(struct i915_hw_ppgtt *ppgtt,
+static void gen6_ppgtt_clear_range(struct i915_hw_ppgtt *ppgtt,
 				   unsigned first_entry,
 				   unsigned num_entries)
 {
@@ -108,7 +108,51 @@ static void i915_ppgtt_clear_range(struct i915_hw_ppgtt *ppgtt,
 	}
 }
 
-int i915_gem_init_aliasing_ppgtt(struct drm_device *dev)
+static void gen6_ppgtt_insert_entries(struct i915_hw_ppgtt *ppgtt,
+				      struct sg_table *pages,
+				      unsigned first_entry,
+				      enum i915_cache_level cache_level)
+{
+	gtt_pte_t *pt_vaddr;
+	unsigned act_pd = first_entry / I915_PPGTT_PT_ENTRIES;
+	unsigned first_pte = first_entry % I915_PPGTT_PT_ENTRIES;
+	unsigned i, j, m, segment_len;
+	dma_addr_t page_addr;
+	struct scatterlist *sg;
+
+	/* init sg walking */
+	sg = pages->sgl;
+	i = 0;
+	segment_len = sg_dma_len(sg) >> PAGE_SHIFT;
+	m = 0;
+
+	while (i < pages->nents) {
+		pt_vaddr = kmap_atomic(ppgtt->pt_pages[act_pd]);
+
+		for (j = first_pte; j < I915_PPGTT_PT_ENTRIES; j++) {
+			page_addr = sg_dma_address(sg) + (m << PAGE_SHIFT);
+			pt_vaddr[j] = pte_encode(ppgtt->dev, page_addr,
+						 cache_level);
+
+			/* grab the next page */
+			if (++m == segment_len) {
+				if (++i == pages->nents)
+					break;
+
+				sg = sg_next(sg);
+				segment_len = sg_dma_len(sg) >> PAGE_SHIFT;
+				m = 0;
+			}
+		}
+
+		kunmap_atomic(pt_vaddr);
+
+		first_pte = 0;
+		act_pd++;
+	}
+}
+
+static int i915_gem_init_aliasing_ppgtt(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct i915_hw_ppgtt *ppgtt;
@@ -127,6 +171,8 @@ int i915_gem_init_aliasing_ppgtt(struct drm_device *dev)
 
 	ppgtt->dev = dev;
 	ppgtt->num_pd_entries = I915_PPGTT_PD_ENTRIES;
+	ppgtt->clear_range = gen6_ppgtt_clear_range;
+	ppgtt->insert_entries = gen6_ppgtt_insert_entries;
 	ppgtt->pt_pages = kzalloc(sizeof(struct page *)*ppgtt->num_pd_entries,
 				  GFP_KERNEL);
 	if (!ppgtt->pt_pages)
@@ -159,8 +205,8 @@ int i915_gem_init_aliasing_ppgtt(struct drm_device *dev)
 
 	ppgtt->scratch_page_dma_addr = dev_priv->gtt.scratch_page_dma;
 
-	i915_ppgtt_clear_range(ppgtt, 0,
-			       ppgtt->num_pd_entries*I915_PPGTT_PT_ENTRIES);
+	ppgtt->clear_range(ppgtt, 0,
+			   ppgtt->num_pd_entries*I915_PPGTT_PT_ENTRIES);
 
 	ppgtt->pd_offset = (first_pd_entry_in_global_pt)*sizeof(gtt_pte_t);
 
@@ -209,66 +255,21 @@ void i915_gem_cleanup_aliasing_ppgtt(struct drm_device *dev)
 	kfree(ppgtt);
 }
 
-static void i915_ppgtt_insert_sg_entries(struct i915_hw_ppgtt *ppgtt,
-					 const struct sg_table *pages,
-					 unsigned first_entry,
-					 enum i915_cache_level cache_level)
-{
-	gtt_pte_t *pt_vaddr;
-	unsigned act_pd = first_entry / I915_PPGTT_PT_ENTRIES;
-	unsigned first_pte = first_entry % I915_PPGTT_PT_ENTRIES;
-	unsigned i, j, m, segment_len;
-	dma_addr_t page_addr;
-	struct scatterlist *sg;
-
-	/* init sg walking */
-	sg = pages->sgl;
-	i = 0;
-	segment_len = sg_dma_len(sg) >> PAGE_SHIFT;
-	m = 0;
-
-	while (i < pages->nents) {
-		pt_vaddr = kmap_atomic(ppgtt->pt_pages[act_pd]);
-
-		for (j = first_pte; j < I915_PPGTT_PT_ENTRIES; j++) {
-			page_addr = sg_dma_address(sg) + (m << PAGE_SHIFT);
-			pt_vaddr[j] = pte_encode(ppgtt->dev, page_addr,
-						 cache_level);
-
-			/* grab the next page */
-			if (++m == segment_len) {
-				if (++i == pages->nents)
-					break;
-
-				sg = sg_next(sg);
-				segment_len = sg_dma_len(sg) >> PAGE_SHIFT;
-				m = 0;
-			}
-		}
-
-		kunmap_atomic(pt_vaddr);
-
-		first_pte = 0;
-		act_pd++;
-	}
-}
-
 void i915_ppgtt_bind_object(struct i915_hw_ppgtt *ppgtt,
 			    struct drm_i915_gem_object *obj,
 			    enum i915_cache_level cache_level)
 {
-	i915_ppgtt_insert_sg_entries(ppgtt,
-				     obj->pages,
-				     obj->gtt_space->start >> PAGE_SHIFT,
-				     cache_level);
+	ppgtt->insert_entries(ppgtt, obj->pages,
+			      obj->gtt_space->start >> PAGE_SHIFT,
+			      cache_level);
 }
 
 void i915_ppgtt_unbind_object(struct i915_hw_ppgtt *ppgtt,
 			      struct drm_i915_gem_object *obj)
 {
-	i915_ppgtt_clear_range(ppgtt,
-			       obj->gtt_space->start >> PAGE_SHIFT,
-			       obj->base.size >> PAGE_SHIFT);
+	ppgtt->clear_range(ppgtt,
+			   obj->gtt_space->start >> PAGE_SHIFT,
+			   obj->base.size >> PAGE_SHIFT);
 }
 
 void i915_gem_init_ppgtt(struct drm_device *dev)
-- 
cgit v0.10.2


From 960e3e429f0c7d9e27e60cf8fa2f51ada71e717e Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 24 Jan 2013 14:44:57 -0800
Subject: drm/i915: pte_encode is gen6+

All the other gen6+ hw code has the gen6_ prefix, so be consistent
about it.

Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index f63dbc7..d171982 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -44,9 +44,9 @@ typedef uint32_t gtt_pte_t;
 #define GEN6_PTE_CACHE_LLC_MLC		(3 << 1)
 #define GEN6_PTE_ADDR_ENCODE(addr)	GEN6_GTT_ADDR_ENCODE(addr)
 
-static inline gtt_pte_t pte_encode(struct drm_device *dev,
-				   dma_addr_t addr,
-				   enum i915_cache_level level)
+static inline gtt_pte_t gen6_pte_encode(struct drm_device *dev,
+					dma_addr_t addr,
+					enum i915_cache_level level)
 {
 	gtt_pte_t pte = GEN6_PTE_VALID;
 	pte |= GEN6_PTE_ADDR_ENCODE(addr);
@@ -87,8 +87,9 @@ static void gen6_ppgtt_clear_range(struct i915_hw_ppgtt *ppgtt,
 	unsigned first_pte = first_entry % I915_PPGTT_PT_ENTRIES;
 	unsigned last_pte, i;
 
-	scratch_pte = pte_encode(ppgtt->dev, ppgtt->scratch_page_dma_addr,
-				 I915_CACHE_LLC);
+	scratch_pte = gen6_pte_encode(ppgtt->dev,
+				      ppgtt->scratch_page_dma_addr,
+				      I915_CACHE_LLC);
 
 	while (num_entries) {
 		last_pte = first_pte + num_entries;
@@ -131,8 +132,8 @@ static void gen6_ppgtt_insert_entries(struct i915_hw_ppgtt *ppgtt,
 
 		for (j = first_pte; j < I915_PPGTT_PT_ENTRIES; j++) {
 			page_addr = sg_dma_address(sg) + (m << PAGE_SHIFT);
-			pt_vaddr[j] = pte_encode(ppgtt->dev, page_addr,
-						 cache_level);
+			pt_vaddr[j] = gen6_pte_encode(ppgtt->dev, page_addr,
+						      cache_level);
 
 			/* grab the next page */
 			if (++m == segment_len) {
@@ -421,7 +422,8 @@ static void gen6_ggtt_insert_entries(struct drm_device *dev,
 		len = sg_dma_len(sg) >> PAGE_SHIFT;
 		for (m = 0; m < len; m++) {
 			addr = sg_dma_address(sg) + (m << PAGE_SHIFT);
-			iowrite32(pte_encode(dev, addr, level), &gtt_entries[i]);
+			iowrite32(gen6_pte_encode(dev, addr, level),
+				  &gtt_entries[i]);
 			i++;
 		}
 	}
@@ -433,7 +435,8 @@ static void gen6_ggtt_insert_entries(struct drm_device *dev,
 	 * hardware should work, we must keep this posting read for paranoia.
 	 */
 	if (i != 0)
-		WARN_ON(readl(&gtt_entries[i-1]) != pte_encode(dev, addr, level));
+		WARN_ON(readl(&gtt_entries[i-1])
+			!= gen6_pte_encode(dev, addr, level));
 
 	/* This next bit makes the above posting read even more important. We
 	 * want to flush the TLBs only after we're certain all the PTE updates
@@ -458,7 +461,8 @@ static void gen6_ggtt_clear_range(struct drm_device *dev,
 		 first_entry, num_entries, max_entries))
 		num_entries = max_entries;
 
-	scratch_pte = pte_encode(dev, dev_priv->gtt.scratch_page_dma, I915_CACHE_LLC);
+	scratch_pte = gen6_pte_encode(dev, dev_priv->gtt.scratch_page_dma,
+				      I915_CACHE_LLC);
 	for (i = 0; i < num_entries; i++)
 		iowrite32(scratch_pte, &gtt_base[i]);
 	readl(gtt_base);
-- 
cgit v0.10.2


From 3440d265857b411a13ed8f67814a29fa2011cdb3 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 24 Jan 2013 13:49:56 -0800
Subject: drm/i915: extract hw ppgtt setup/cleanup code

At the moment only cosmetics, but being able to initialize/cleanup
arbitrary ppgtt address spaces paves the way to have more than one of
them ... Just in case we ever get around to implementing real
per-process address spaces. Note that in that case another vfunc for
ppgtt would be beneficial though. But that can wait until the code
grows a second place which initializes ppgtts.

Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 4be7c68..cfae2ce 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -421,6 +421,7 @@ struct i915_hw_ppgtt {
 			       struct sg_table *st,
 			       unsigned int pg_start,
 			       enum i915_cache_level cache_level);
+	void (*cleanup)(struct i915_hw_ppgtt *ppgtt);
 };
 
 
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index d171982..5502067 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -153,10 +153,28 @@ static void gen6_ppgtt_insert_entries(struct i915_hw_ppgtt *ppgtt,
 	}
 }
 
-static int i915_gem_init_aliasing_ppgtt(struct drm_device *dev)
+static void gen6_ppgtt_cleanup(struct i915_hw_ppgtt *ppgtt)
 {
+	int i;
+
+	if (ppgtt->pt_dma_addr) {
+		for (i = 0; i < ppgtt->num_pd_entries; i++)
+			pci_unmap_page(ppgtt->dev->pdev,
+				       ppgtt->pt_dma_addr[i],
+				       4096, PCI_DMA_BIDIRECTIONAL);
+	}
+
+	kfree(ppgtt->pt_dma_addr);
+	for (i = 0; i < ppgtt->num_pd_entries; i++)
+		__free_page(ppgtt->pt_pages[i]);
+	kfree(ppgtt->pt_pages);
+	kfree(ppgtt);
+}
+
+static int gen6_ppgtt_init(struct i915_hw_ppgtt *ppgtt)
+{
+	struct drm_device *dev = ppgtt->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct i915_hw_ppgtt *ppgtt;
 	unsigned first_pd_entry_in_global_pt;
 	int i;
 	int ret = -ENOMEM;
@@ -166,18 +184,14 @@ static int i915_gem_init_aliasing_ppgtt(struct drm_device *dev)
 	 * now. */
 	first_pd_entry_in_global_pt = dev_priv->mm.gtt->gtt_total_entries - I915_PPGTT_PD_ENTRIES;
 
-	ppgtt = kzalloc(sizeof(*ppgtt), GFP_KERNEL);
-	if (!ppgtt)
-		return ret;
-
-	ppgtt->dev = dev;
 	ppgtt->num_pd_entries = I915_PPGTT_PD_ENTRIES;
 	ppgtt->clear_range = gen6_ppgtt_clear_range;
 	ppgtt->insert_entries = gen6_ppgtt_insert_entries;
+	ppgtt->cleanup = gen6_ppgtt_cleanup;
 	ppgtt->pt_pages = kzalloc(sizeof(struct page *)*ppgtt->num_pd_entries,
 				  GFP_KERNEL);
 	if (!ppgtt->pt_pages)
-		goto err_ppgtt;
+		return -ENOMEM;
 
 	for (i = 0; i < ppgtt->num_pd_entries; i++) {
 		ppgtt->pt_pages[i] = alloc_page(GFP_KERNEL);
@@ -211,8 +225,6 @@ static int i915_gem_init_aliasing_ppgtt(struct drm_device *dev)
 
 	ppgtt->pd_offset = (first_pd_entry_in_global_pt)*sizeof(gtt_pte_t);
 
-	dev_priv->mm.aliasing_ppgtt = ppgtt;
-
 	return 0;
 
 err_pd_pin:
@@ -228,8 +240,27 @@ err_pt_alloc:
 			__free_page(ppgtt->pt_pages[i]);
 	}
 	kfree(ppgtt->pt_pages);
-err_ppgtt:
-	kfree(ppgtt);
+
+	return ret;
+}
+
+static int i915_gem_init_aliasing_ppgtt(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct i915_hw_ppgtt *ppgtt;
+	int ret;
+
+	ppgtt = kzalloc(sizeof(*ppgtt), GFP_KERNEL);
+	if (!ppgtt)
+		return -ENOMEM;
+
+	ppgtt->dev = dev;
+
+	ret = gen6_ppgtt_init(ppgtt);
+	if (ret)
+		kfree(ppgtt);
+	else
+		dev_priv->mm.aliasing_ppgtt = ppgtt;
 
 	return ret;
 }
@@ -238,22 +269,11 @@ void i915_gem_cleanup_aliasing_ppgtt(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct i915_hw_ppgtt *ppgtt = dev_priv->mm.aliasing_ppgtt;
-	int i;
 
 	if (!ppgtt)
 		return;
 
-	if (ppgtt->pt_dma_addr) {
-		for (i = 0; i < ppgtt->num_pd_entries; i++)
-			pci_unmap_page(dev->pdev, ppgtt->pt_dma_addr[i],
-				       4096, PCI_DMA_BIDIRECTIONAL);
-	}
-
-	kfree(ppgtt->pt_dma_addr);
-	for (i = 0; i < ppgtt->num_pd_entries; i++)
-		__free_page(ppgtt->pt_pages[i]);
-	kfree(ppgtt->pt_pages);
-	kfree(ppgtt);
+	ppgtt->cleanup(ppgtt);
 }
 
 void i915_ppgtt_bind_object(struct i915_hw_ppgtt *ppgtt,
-- 
cgit v0.10.2


From baa09f5fd8a6d033ec075355dda99a65b7f6a0f3 Mon Sep 17 00:00:00 2001
From: Ben Widawsky <ben@bwidawsk.net>
Date: Thu, 24 Jan 2013 13:49:57 -0800
Subject: drm/i915: Add probe and remove to the gtt ops

The idea, and much of the code came originally from:

commit 0712f0249c3148d8cf42a3703403c278590d4de5
Author: Ben Widawsky <ben@bwidawsk.net>
Date:   Fri Jan 18 17:23:16 2013 -0800

    drm/i915: Create a vtable for i915 gtt

Daniel didn't like the color of that patch series, and so I asked him to
start something which appealed to his sense of color. The preceding
patches are those, and now this is going on top of that.

[extracted from the original commit message]

One immediately obvious thing to implement is our gmch probing. The init
function was getting massively bloated. Fundamentally, all that's needed
from GMCH probing is the GTT size, and the stolen size. It makes design
sense to put the mappable calculation in there as well, but the code
turns out a bit nicer without it (IMO)

The intel_gtt bridge thing is still here, but the subsequent patches
will finish ripping that out.

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
[danvet: Bikeshedded one comment (GMADR is just the PCI aperture, we
use it for other things than just accessing tiled surfaces through a
linear view) and cut the newly added long lines a bit. Also one
checkpatch error.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index 11c7aa8..cf06103 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -1665,7 +1665,7 @@ out_mtrrfree:
 out_rmmap:
 	pci_iounmap(dev->pdev, dev_priv->regs);
 put_gmch:
-	i915_gem_gtt_fini(dev);
+	dev_priv->gtt.gtt_remove(dev);
 put_bridge:
 	pci_dev_put(dev_priv->bridge_dev);
 free_priv:
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index cfae2ce..dee8d5c 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -381,6 +381,7 @@ enum i915_cache_level {
 struct i915_gtt {
 	unsigned long start;		/* Start offset of used GTT */
 	size_t total;			/* Total size GTT can map */
+	size_t stolen_size;		/* Total size of stolen memory */
 
 	unsigned long mappable_end;	/* End offset that we can CPU map */
 	struct io_mapping *mappable;	/* Mapping to our CPU mappable region */
@@ -394,6 +395,9 @@ struct i915_gtt {
 	struct page *scratch_page;
 
 	/* global gtt ops */
+	int (*gtt_probe)(struct drm_device *dev, size_t *gtt_total,
+			  size_t *stolen);
+	void (*gtt_remove)(struct drm_device *dev);
 	void (*gtt_clear_range)(struct drm_device *dev,
 				unsigned int first_entry,
 				unsigned int num_entries);
@@ -1689,7 +1693,6 @@ void i915_gem_init_global_gtt(struct drm_device *dev);
 void i915_gem_setup_global_gtt(struct drm_device *dev, unsigned long start,
 			       unsigned long mappable_end, unsigned long end);
 int i915_gem_gtt_init(struct drm_device *dev);
-void i915_gem_gtt_fini(struct drm_device *dev);
 static inline void i915_gem_chipset_flush(struct drm_device *dev)
 {
 	if (INTEL_INFO(dev)->gen < 6)
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 5502067..87249e8 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -707,14 +707,14 @@ static inline unsigned int gen6_get_total_gtt_size(u16 snb_gmch_ctl)
 	return snb_gmch_ctl << 20;
 }
 
-static inline unsigned int gen6_get_stolen_size(u16 snb_gmch_ctl)
+static inline size_t gen6_get_stolen_size(u16 snb_gmch_ctl)
 {
 	snb_gmch_ctl >>= SNB_GMCH_GMS_SHIFT;
 	snb_gmch_ctl &= SNB_GMCH_GMS_MASK;
 	return snb_gmch_ctl << 25; /* 32 MB units */
 }
 
-static inline unsigned int gen7_get_stolen_size(u16 snb_gmch_ctl)
+static inline size_t gen7_get_stolen_size(u16 snb_gmch_ctl)
 {
 	static const int stolen_decoder[] = {
 		0, 0, 0, 0, 0, 32, 48, 64, 128, 256, 96, 160, 224, 352};
@@ -723,109 +723,141 @@ static inline unsigned int gen7_get_stolen_size(u16 snb_gmch_ctl)
 	return stolen_decoder[snb_gmch_ctl] << 20;
 }
 
-int i915_gem_gtt_init(struct drm_device *dev)
+static int gen6_gmch_probe(struct drm_device *dev,
+			   size_t *gtt_total,
+			   size_t *stolen)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	phys_addr_t gtt_bus_addr;
+	unsigned int gtt_size;
 	u16 snb_gmch_ctl;
 	int ret;
 
-	dev_priv->gtt.mappable_base = pci_resource_start(dev->pdev, 2);
-	dev_priv->gtt.mappable_end = pci_resource_len(dev->pdev, 2);
-
-	/* On modern platforms we need not worry ourself with the legacy
-	 * hostbridge query stuff. Skip it entirely
+	/* 64/512MB is the current min/max we actually know of, but this is just
+	 * a coarse sanity check.
 	 */
-	if (INTEL_INFO(dev)->gen < 6) {
-		ret = intel_gmch_probe(dev_priv->bridge_dev, dev->pdev, NULL);
-		if (!ret) {
-			DRM_ERROR("failed to set up gmch\n");
-			return -EIO;
-		}
-
-		dev_priv->mm.gtt = intel_gtt_get();
-		if (!dev_priv->mm.gtt) {
-			DRM_ERROR("Failed to initialize GTT\n");
-			intel_gmch_remove();
-			return -ENODEV;
-		}
-
-		dev_priv->gtt.do_idle_maps = needs_idle_maps(dev);
-
-		dev_priv->gtt.gtt_clear_range = i915_ggtt_clear_range;
-		dev_priv->gtt.gtt_insert_entries = i915_ggtt_insert_entries;
-
-		return 0;
+	if ((dev_priv->gtt.mappable_end < (64<<20) ||
+	     (dev_priv->gtt.mappable_end > (512<<20)))) {
+		DRM_ERROR("Unknown GMADR size (%lx)\n",
+			  dev_priv->gtt.mappable_end);
+		return -ENXIO;
 	}
 
 	if (!pci_set_dma_mask(dev->pdev, DMA_BIT_MASK(40)))
 		pci_set_consistent_dma_mask(dev->pdev, DMA_BIT_MASK(40));
-
-	dev_priv->mm.gtt = kzalloc(sizeof(*dev_priv->mm.gtt), GFP_KERNEL);
-	if (!dev_priv->mm.gtt)
-		return -ENOMEM;
-
-	/* For GEN6+ the PTEs for the ggtt live at 2MB + BAR0 */
-	gtt_bus_addr = pci_resource_start(dev->pdev, 0) + (2<<20);
-
-	/* i9xx_setup */
 	pci_read_config_word(dev->pdev, SNB_GMCH_CTRL, &snb_gmch_ctl);
-	dev_priv->mm.gtt->gtt_total_entries =
-		gen6_get_total_gtt_size(snb_gmch_ctl) / sizeof(gtt_pte_t);
-	if (INTEL_INFO(dev)->gen < 7)
-		dev_priv->mm.gtt->stolen_size = gen6_get_stolen_size(snb_gmch_ctl);
-	else
-		dev_priv->mm.gtt->stolen_size = gen7_get_stolen_size(snb_gmch_ctl);
+	gtt_size = gen6_get_total_gtt_size(snb_gmch_ctl);
 
-	/* 64/512MB is the current min/max we actually know of, but this is just a
-	 * coarse sanity check.
-	 */
-	if ((dev_priv->gtt.mappable_end < (64<<20) ||
-	    (dev_priv->gtt.mappable_end > (512<<20)))) {
-		DRM_ERROR("Unknown GMADR size (%lx)\n",
-			  dev_priv->gtt.mappable_end);
-		ret = -ENXIO;
-		goto err_out;
-	}
+	if (IS_GEN7(dev))
+		*stolen = gen7_get_stolen_size(snb_gmch_ctl);
+	else
+		*stolen = gen6_get_stolen_size(snb_gmch_ctl);
 
-	ret = setup_scratch_page(dev);
-	if (ret) {
-		DRM_ERROR("Scratch setup failed\n");
-		goto err_out;
-	}
+	*gtt_total = (gtt_size / sizeof(gtt_pte_t)) << PAGE_SHIFT;
 
-	dev_priv->gtt.gsm = ioremap_wc(gtt_bus_addr,
-				       dev_priv->mm.gtt->gtt_total_entries * sizeof(gtt_pte_t));
+	/* For GEN6+ the PTEs for the ggtt live at 2MB + BAR0 */
+	gtt_bus_addr = pci_resource_start(dev->pdev, 0) + (2<<20);
+	dev_priv->gtt.gsm = ioremap_wc(gtt_bus_addr, gtt_size);
 	if (!dev_priv->gtt.gsm) {
 		DRM_ERROR("Failed to map the gtt page table\n");
-		teardown_scratch_page(dev);
-		ret = -ENOMEM;
-		goto err_out;
+		return -ENOMEM;
 	}
 
-	/* GMADR is the PCI aperture used by SW to access tiled GFX surfaces in a linear fashion. */
-	DRM_INFO("Memory usable by graphics device = %dM\n", dev_priv->mm.gtt->gtt_total_entries >> 8);
-	DRM_DEBUG_DRIVER("GMADR size = %ldM\n", dev_priv->gtt.mappable_end >> 20);
-	DRM_DEBUG_DRIVER("GTT stolen size = %dM\n", dev_priv->mm.gtt->stolen_size >> 20);
+	ret = setup_scratch_page(dev);
+	if (ret)
+		DRM_ERROR("Scratch setup failed\n");
 
 	dev_priv->gtt.gtt_clear_range = gen6_ggtt_clear_range;
 	dev_priv->gtt.gtt_insert_entries = gen6_ggtt_insert_entries;
 
-	return 0;
-
-err_out:
-	kfree(dev_priv->mm.gtt);
-	if (INTEL_INFO(dev)->gen < 6)
-		intel_gmch_remove();
 	return ret;
 }
 
-void i915_gem_gtt_fini(struct drm_device *dev)
+void gen6_gmch_remove(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	iounmap(dev_priv->gtt.gsm);
-	teardown_scratch_page(dev);
-	if (INTEL_INFO(dev)->gen < 6)
-		intel_gmch_remove();
+	teardown_scratch_page(dev_priv->dev);
 	kfree(dev_priv->mm.gtt);
 }
+
+static int i915_gmch_probe(struct drm_device *dev,
+			   size_t *gtt_total,
+			   size_t *stolen)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int ret;
+
+	/* This is a temporary hack to make the code cleaner in
+	 * i915_gem_gtt_init. I promise it will go away very shortly. */
+	kfree(dev_priv->mm.gtt);
+
+	ret = intel_gmch_probe(dev_priv->bridge_dev, dev_priv->dev->pdev, NULL);
+	if (!ret) {
+		DRM_ERROR("failed to set up gmch\n");
+		return -EIO;
+	}
+
+	dev_priv->mm.gtt = intel_gtt_get();
+	if (!dev_priv->mm.gtt) {
+		DRM_ERROR("Failed to initialize GTT\n");
+		intel_gmch_remove();
+		return -ENODEV;
+	}
+
+	dev_priv->gtt.do_idle_maps = needs_idle_maps(dev_priv->dev);
+	dev_priv->gtt.gtt_clear_range = i915_ggtt_clear_range;
+	dev_priv->gtt.gtt_insert_entries = i915_ggtt_insert_entries;
+
+	return 0;
+}
+
+static void i915_gmch_remove(struct drm_device *dev)
+{
+	intel_gmch_remove();
+}
+
+int i915_gem_gtt_init(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct i915_gtt *gtt = &dev_priv->gtt;
+	unsigned long gtt_size;
+	int ret;
+
+	dev_priv->mm.gtt = kzalloc(sizeof(*dev_priv->mm.gtt), GFP_KERNEL);
+	if (!dev_priv->mm.gtt)
+		return -ENOMEM;
+
+	gtt->mappable_base = pci_resource_start(dev->pdev, 2);
+	gtt->mappable_end = pci_resource_len(dev->pdev, 2);
+
+	if (INTEL_INFO(dev)->gen <= 5) {
+		dev_priv->gtt.gtt_probe = i915_gmch_probe;
+		dev_priv->gtt.gtt_remove = i915_gmch_remove;
+	} else {
+		dev_priv->gtt.gtt_probe = gen6_gmch_probe;
+		dev_priv->gtt.gtt_remove = gen6_gmch_remove;
+	}
+
+	ret = dev_priv->gtt.gtt_probe(dev, &dev_priv->gtt.total,
+				     &dev_priv->gtt.stolen_size);
+	if (ret) {
+		kfree(dev_priv->mm.gtt);
+		return ret;
+	}
+
+	dev_priv->mm.gtt->gtt_total_entries = dev_priv->gtt.total >> PAGE_SHIFT;
+	dev_priv->mm.gtt->stolen_size = dev_priv->gtt.stolen_size;
+
+	gtt_size = (dev_priv->gtt.total >> PAGE_SHIFT) * sizeof(gtt_pte_t);
+
+	/* GMADR is the PCI mmio aperture into the global GTT. */
+	DRM_INFO("Memory usable by graphics device = %zdM\n",
+		 dev_priv->gtt.total >> 20);
+	DRM_DEBUG_DRIVER("GMADR size = %ldM\n",
+			 dev_priv->gtt.mappable_end >> 20);
+	DRM_DEBUG_DRIVER("GTT stolen size = %zdM\n",
+			 dev_priv->gtt.stolen_size >> 20);
+
+	return 0;
+}
-- 
cgit v0.10.2


From a54c0c279f3864171fe53c66e769d5a137c5c651 Mon Sep 17 00:00:00 2001
From: Ben Widawsky <ben@bwidawsk.net>
Date: Thu, 24 Jan 2013 14:45:00 -0800
Subject: drm/i915: remove intel_gtt structure

With the probe call in our dispatch table, we can now cut away the
last three remaining members in the intel_gtt shared struct and so
remove it completely.

v2: Rebased on top of Daniel's series

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
[danvet: bikeshed commit message a bit.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/char/agp/intel-gtt.c b/drivers/char/agp/intel-gtt.c
index ff5f348..d8e7e6c 100644
--- a/drivers/char/agp/intel-gtt.c
+++ b/drivers/char/agp/intel-gtt.c
@@ -60,7 +60,6 @@ struct intel_gtt_driver {
 };
 
 static struct _intel_private {
-	struct intel_gtt base;
 	const struct intel_gtt_driver *driver;
 	struct pci_dev *pcidev;	/* device one */
 	struct pci_dev *bridge_dev;
@@ -80,6 +79,13 @@ static struct _intel_private {
 	/* Whether i915 needs to use the dmar apis or not. */
 	unsigned int needs_dmar : 1;
 	phys_addr_t gma_bus_addr;
+	/*  Size of memory reserved for graphics by the BIOS */
+	unsigned int stolen_size;
+	/* Total number of gtt entries. */
+	unsigned int gtt_total_entries;
+	/* Part of the gtt that is mappable by the cpu, for those chips where
+	 * this is not the full gtt. */
+	unsigned int gtt_mappable_entries;
 } intel_private;
 
 #define INTEL_GTT_GEN	intel_private.driver->gen
@@ -510,7 +516,7 @@ static unsigned int intel_gtt_total_entries(void)
 		/* On previous hardware, the GTT size was just what was
 		 * required to map the aperture.
 		 */
-		return intel_private.base.gtt_mappable_entries;
+		return intel_private.gtt_mappable_entries;
 	}
 }
 
@@ -576,8 +582,8 @@ static int intel_gtt_init(void)
 	if (ret != 0)
 		return ret;
 
-	intel_private.base.gtt_mappable_entries = intel_gtt_mappable_entries();
-	intel_private.base.gtt_total_entries = intel_gtt_total_entries();
+	intel_private.gtt_mappable_entries = intel_gtt_mappable_entries();
+	intel_private.gtt_total_entries = intel_gtt_total_entries();
 
 	/* save the PGETBL reg for resume */
 	intel_private.PGETBL_save =
@@ -589,10 +595,10 @@ static int intel_gtt_init(void)
 
 	dev_info(&intel_private.bridge_dev->dev,
 			"detected gtt size: %dK total, %dK mappable\n",
-			intel_private.base.gtt_total_entries * 4,
-			intel_private.base.gtt_mappable_entries * 4);
+			intel_private.gtt_total_entries * 4,
+			intel_private.gtt_mappable_entries * 4);
 
-	gtt_map_size = intel_private.base.gtt_total_entries * 4;
+	gtt_map_size = intel_private.gtt_total_entries * 4;
 
 	intel_private.gtt = NULL;
 	if (INTEL_GTT_GEN < 6 && INTEL_GTT_GEN > 2)
@@ -609,7 +615,7 @@ static int intel_gtt_init(void)
 
 	global_cache_flush();   /* FIXME: ? */
 
-	intel_private.base.stolen_size = intel_gtt_stolen_size();
+	intel_private.stolen_size = intel_gtt_stolen_size();
 
 	intel_private.needs_dmar = USE_PCI_DMA_API && INTEL_GTT_GEN > 2;
 
@@ -637,8 +643,7 @@ static int intel_fake_agp_fetch_size(void)
 	unsigned int aper_size;
 	int i;
 
-	aper_size = (intel_private.base.gtt_mappable_entries << PAGE_SHIFT)
-		    / MB(1);
+	aper_size = (intel_private.gtt_mappable_entries << PAGE_SHIFT) / MB(1);
 
 	for (i = 0; i < num_sizes; i++) {
 		if (aper_size == intel_fake_agp_sizes[i].size) {
@@ -845,8 +850,8 @@ static int intel_fake_agp_insert_entries(struct agp_memory *mem,
 	int ret = -EINVAL;
 
 	if (intel_private.clear_fake_agp) {
-		int start = intel_private.base.stolen_size / PAGE_SIZE;
-		int end = intel_private.base.gtt_mappable_entries;
+		int start = intel_private.stolen_size / PAGE_SIZE;
+		int end = intel_private.gtt_mappable_entries;
 		intel_gtt_clear_range(start, end - start);
 		intel_private.clear_fake_agp = false;
 	}
@@ -857,7 +862,7 @@ static int intel_fake_agp_insert_entries(struct agp_memory *mem,
 	if (mem->page_count == 0)
 		goto out;
 
-	if (pg_start + mem->page_count > intel_private.base.gtt_total_entries)
+	if (pg_start + mem->page_count > intel_private.gtt_total_entries)
 		goto out_err;
 
 	if (type != mem->type)
@@ -1366,9 +1371,10 @@ int intel_gmch_probe(struct pci_dev *bridge_pdev, struct pci_dev *gpu_pdev,
 }
 EXPORT_SYMBOL(intel_gmch_probe);
 
-struct intel_gtt *intel_gtt_get(void)
+void intel_gtt_get(size_t *gtt_total, size_t *stolen_size)
 {
-	return &intel_private.base;
+	*gtt_total = intel_private.gtt_total_entries << PAGE_SHIFT;
+	*stolen_size = intel_private.stolen_size;
 }
 EXPORT_SYMBOL(intel_gtt_get);
 
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index dee8d5c..d327416 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -406,6 +406,7 @@ struct i915_gtt {
 				   unsigned int pg_start,
 				   enum i915_cache_level cache_level);
 };
+#define gtt_total_entries(gtt) ((gtt).total >> PAGE_SHIFT)
 
 #define I915_PPGTT_PD_ENTRIES 512
 #define I915_PPGTT_PT_ENTRIES 1024
@@ -696,8 +697,6 @@ struct intel_l3_parity {
 };
 
 struct i915_gem_mm {
-	/** Bridge to intel-gtt-ko */
-	struct intel_gtt *gtt;
 	/** Memory allocator for GTT stolen memory */
 	struct drm_mm stolen;
 	/** Memory allocator for GTT */
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 87249e8..f7e0018 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -182,7 +182,8 @@ static int gen6_ppgtt_init(struct i915_hw_ppgtt *ppgtt)
 	/* ppgtt PDEs reside in the global gtt pagetable, which has 512*1024
 	 * entries. For aliasing ppgtt support we just steal them at the end for
 	 * now. */
-	first_pd_entry_in_global_pt = dev_priv->mm.gtt->gtt_total_entries - I915_PPGTT_PD_ENTRIES;
+	first_pd_entry_in_global_pt =
+		gtt_total_entries(dev_priv->gtt) - I915_PPGTT_PD_ENTRIES;
 
 	ppgtt->num_pd_entries = I915_PPGTT_PD_ENTRIES;
 	ppgtt->clear_range = gen6_ppgtt_clear_range;
@@ -473,7 +474,7 @@ static void gen6_ggtt_clear_range(struct drm_device *dev,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	gtt_pte_t scratch_pte;
 	gtt_pte_t __iomem *gtt_base = (gtt_pte_t __iomem *) dev_priv->gtt.gsm + first_entry;
-	const int max_entries = dev_priv->mm.gtt->gtt_total_entries - first_entry;
+	const int max_entries = gtt_total_entries(dev_priv->gtt) - first_entry;
 	int i;
 
 	if (WARN(num_entries > max_entries,
@@ -634,7 +635,7 @@ void i915_gem_init_global_gtt(struct drm_device *dev)
 	unsigned long gtt_size, mappable_size;
 	int ret;
 
-	gtt_size = dev_priv->mm.gtt->gtt_total_entries << PAGE_SHIFT;
+	gtt_size = dev_priv->gtt.total;
 	mappable_size = dev_priv->gtt.mappable_end;
 
 	if (intel_enable_ppgtt(dev) && HAS_ALIASING_PPGTT(dev)) {
@@ -778,7 +779,6 @@ void gen6_gmch_remove(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	iounmap(dev_priv->gtt.gsm);
 	teardown_scratch_page(dev_priv->dev);
-	kfree(dev_priv->mm.gtt);
 }
 
 static int i915_gmch_probe(struct drm_device *dev,
@@ -788,22 +788,13 @@ static int i915_gmch_probe(struct drm_device *dev,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int ret;
 
-	/* This is a temporary hack to make the code cleaner in
-	 * i915_gem_gtt_init. I promise it will go away very shortly. */
-	kfree(dev_priv->mm.gtt);
-
 	ret = intel_gmch_probe(dev_priv->bridge_dev, dev_priv->dev->pdev, NULL);
 	if (!ret) {
 		DRM_ERROR("failed to set up gmch\n");
 		return -EIO;
 	}
 
-	dev_priv->mm.gtt = intel_gtt_get();
-	if (!dev_priv->mm.gtt) {
-		DRM_ERROR("Failed to initialize GTT\n");
-		intel_gmch_remove();
-		return -ENODEV;
-	}
+	intel_gtt_get(gtt_total, stolen);
 
 	dev_priv->gtt.do_idle_maps = needs_idle_maps(dev_priv->dev);
 	dev_priv->gtt.gtt_clear_range = i915_ggtt_clear_range;
@@ -824,10 +815,6 @@ int i915_gem_gtt_init(struct drm_device *dev)
 	unsigned long gtt_size;
 	int ret;
 
-	dev_priv->mm.gtt = kzalloc(sizeof(*dev_priv->mm.gtt), GFP_KERNEL);
-	if (!dev_priv->mm.gtt)
-		return -ENOMEM;
-
 	gtt->mappable_base = pci_resource_start(dev->pdev, 2);
 	gtt->mappable_end = pci_resource_len(dev->pdev, 2);
 
@@ -841,13 +828,8 @@ int i915_gem_gtt_init(struct drm_device *dev)
 
 	ret = dev_priv->gtt.gtt_probe(dev, &dev_priv->gtt.total,
 				     &dev_priv->gtt.stolen_size);
-	if (ret) {
-		kfree(dev_priv->mm.gtt);
+	if (ret)
 		return ret;
-	}
-
-	dev_priv->mm.gtt->gtt_total_entries = dev_priv->gtt.total >> PAGE_SHIFT;
-	dev_priv->mm.gtt->stolen_size = dev_priv->gtt.stolen_size;
 
 	gtt_size = (dev_priv->gtt.total >> PAGE_SHIFT) * sizeof(gtt_pte_t);
 
diff --git a/drivers/gpu/drm/i915/i915_gem_stolen.c b/drivers/gpu/drm/i915/i915_gem_stolen.c
index f21ae17..69d97cb 100644
--- a/drivers/gpu/drm/i915/i915_gem_stolen.c
+++ b/drivers/gpu/drm/i915/i915_gem_stolen.c
@@ -187,11 +187,11 @@ int i915_gem_init_stolen(struct drm_device *dev)
 	if (dev_priv->mm.stolen_base == 0)
 		return 0;
 
-	DRM_DEBUG_KMS("found %d bytes of stolen memory at %08lx\n",
-		      dev_priv->mm.gtt->stolen_size, dev_priv->mm.stolen_base);
+	DRM_DEBUG_KMS("found %zd bytes of stolen memory at %08lx\n",
+		      dev_priv->gtt.stolen_size, dev_priv->mm.stolen_base);
 
 	/* Basic memrange allocator for stolen space */
-	drm_mm_init(&dev_priv->mm.stolen, 0, dev_priv->mm.gtt->stolen_size);
+	drm_mm_init(&dev_priv->mm.stolen, 0, dev_priv->gtt.stolen_size);
 
 	return 0;
 }
@@ -205,7 +205,7 @@ i915_pages_create_for_stolen(struct drm_device *dev,
 	struct scatterlist *sg;
 
 	DRM_DEBUG_DRIVER("offset=0x%x, size=%d\n", offset, size);
-	BUG_ON(offset > dev_priv->mm.gtt->stolen_size - size);
+	BUG_ON(offset > dev_priv->gtt.stolen_size - size);
 
 	/* We hide that we have no struct page backing our stolen object
 	 * by wrapping the contiguous physical allocation with a fake
diff --git a/include/drm/intel-gtt.h b/include/drm/intel-gtt.h
index 769b6c7..cf10555 100644
--- a/include/drm/intel-gtt.h
+++ b/include/drm/intel-gtt.h
@@ -3,15 +3,7 @@
 #ifndef _DRM_INTEL_GTT_H
 #define	_DRM_INTEL_GTT_H
 
-struct intel_gtt {
-	/* Size of memory reserved for graphics by the BIOS */
-	unsigned int stolen_size;
-	/* Total number of gtt entries. */
-	unsigned int gtt_total_entries;
-	/* Part of the gtt that is mappable by the cpu, for those chips where
-	 * this is not the full gtt. */
-	unsigned int gtt_mappable_entries;
-} *intel_gtt_get(void);
+void intel_gtt_get(size_t *gtt_total, size_t *stolen_size);
 
 int intel_gmch_probe(struct pci_dev *bridge_pdev, struct pci_dev *gpu_pdev,
 		     struct agp_bridge_data *bridge);
-- 
cgit v0.10.2


From e78891ca7648a14dda5760be7b03eba7c628a804 Mon Sep 17 00:00:00 2001
From: Ben Widawsky <ben@bwidawsk.net>
Date: Fri, 25 Jan 2013 16:41:04 -0800
Subject: drm/i915: Reclaim GTT space for failed PPGTT

When the PPGTT init fails, we may as well reuse the space that we were
reserving for the PPGTT PDEs.

This also fixes an extraneous mutex_unlock.

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index f7e0018..71f09a8 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -567,12 +567,20 @@ static void i915_gtt_color_adjust(struct drm_mm_node *node,
 			*end -= 4096;
 	}
 }
-
 void i915_gem_setup_global_gtt(struct drm_device *dev,
 			       unsigned long start,
 			       unsigned long mappable_end,
 			       unsigned long end)
 {
+	/* Let GEM Manage all of the aperture.
+	 *
+	 * However, leave one page at the end still bound to the scratch page.
+	 * There are a number of places where the hardware apparently prefetches
+	 * past the end of the object, and we've seen multiple hangs with the
+	 * GPU head pointer stuck in a batchbuffer bound at the last page of the
+	 * aperture.  One page should be enough to keep any prefetching inside
+	 * of the aperture.
+	 */
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	struct drm_mm_node *entry;
 	struct drm_i915_gem_object *obj;
@@ -633,12 +641,12 @@ void i915_gem_init_global_gtt(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	unsigned long gtt_size, mappable_size;
-	int ret;
 
 	gtt_size = dev_priv->gtt.total;
 	mappable_size = dev_priv->gtt.mappable_end;
 
 	if (intel_enable_ppgtt(dev) && HAS_ALIASING_PPGTT(dev)) {
+		int ret;
 		/* PPGTT pdes are stolen from global gtt ptes, so shrink the
 		 * aperture accordingly when using aliasing ppgtt. */
 		gtt_size -= I915_PPGTT_PD_ENTRIES*PAGE_SIZE;
@@ -646,23 +654,14 @@ void i915_gem_init_global_gtt(struct drm_device *dev)
 		i915_gem_setup_global_gtt(dev, 0, mappable_size, gtt_size);
 
 		ret = i915_gem_init_aliasing_ppgtt(dev);
-		if (ret) {
-			mutex_unlock(&dev->struct_mutex);
+		if (!ret)
 			return;
-		}
-	} else {
-		/* Let GEM Manage all of the aperture.
-		 *
-		 * However, leave one page at the end still bound to the scratch
-		 * page.  There are a number of places where the hardware
-		 * apparently prefetches past the end of the object, and we've
-		 * seen multiple hangs with the GPU head pointer stuck in a
-		 * batchbuffer bound at the last page of the aperture.  One page
-		 * should be enough to keep any prefetching inside of the
-		 * aperture.
-		 */
-		i915_gem_setup_global_gtt(dev, 0, mappable_size, gtt_size);
+
+		DRM_ERROR("Aliased PPGTT setup failed %d\n", ret);
+		drm_mm_takedown(&dev_priv->mm.gtt_space);
+		gtt_size += I915_PPGTT_PD_ENTRIES*PAGE_SIZE;
 	}
+	i915_gem_setup_global_gtt(dev, 0, mappable_size, gtt_size);
 }
 
 static int setup_scratch_page(struct drm_device *dev)
-- 
cgit v0.10.2


From f82855d342b6c8483c56e6d2e200a71731509a39 Mon Sep 17 00:00:00 2001
From: Ben Widawsky <ben@bwidawsk.net>
Date: Tue, 29 Jan 2013 12:00:15 -0800
Subject: drm/i915: Fix CAGF for HSW

The shift changed, hurray.

Reported-by: Kenneth Graunke <kenneth@whitecape.org>
Cc: Paulo Zanoni <przanoni@gmail.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Tested-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Cc: stable@vger.kernel.org
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index 384f193..7491148 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -957,7 +957,7 @@ static int i915_cur_delayinfo(struct seq_file *m, void *unused)
 		u32 gt_perf_status = I915_READ(GEN6_GT_PERF_STATUS);
 		u32 rp_state_limits = I915_READ(GEN6_RP_STATE_LIMITS);
 		u32 rp_state_cap = I915_READ(GEN6_RP_STATE_CAP);
-		u32 rpstat;
+		u32 rpstat, cagf;
 		u32 rpupei, rpcurup, rpprevup;
 		u32 rpdownei, rpcurdown, rpprevdown;
 		int max_freq;
@@ -976,6 +976,11 @@ static int i915_cur_delayinfo(struct seq_file *m, void *unused)
 		rpdownei = I915_READ(GEN6_RP_CUR_DOWN_EI);
 		rpcurdown = I915_READ(GEN6_RP_CUR_DOWN);
 		rpprevdown = I915_READ(GEN6_RP_PREV_DOWN);
+		if (IS_HASWELL(dev))
+			cagf = (rpstat & HSW_CAGF_MASK) >> HSW_CAGF_SHIFT;
+		else
+			cagf = (rpstat & GEN6_CAGF_MASK) >> GEN6_CAGF_SHIFT;
+		cagf *= GT_FREQUENCY_MULTIPLIER;
 
 		gen6_gt_force_wake_put(dev_priv);
 		mutex_unlock(&dev->struct_mutex);
@@ -988,8 +993,7 @@ static int i915_cur_delayinfo(struct seq_file *m, void *unused)
 			   gt_perf_status & 0xff);
 		seq_printf(m, "Render p-state limit: %d\n",
 			   rp_state_limits & 0xff);
-		seq_printf(m, "CAGF: %dMHz\n", ((rpstat & GEN6_CAGF_MASK) >>
-						GEN6_CAGF_SHIFT) * GT_FREQUENCY_MULTIPLIER);
+		seq_printf(m, "CAGF: %dMHz\n", cagf);
 		seq_printf(m, "RP CUR UP EI: %dus\n", rpupei &
 			   GEN6_CURICONT_MASK);
 		seq_printf(m, "RP CUR UP: %dus\n", rpcurup &
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 0c89cf5..9a3cd04 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -4196,7 +4196,9 @@
 #define GEN6_RP_INTERRUPT_LIMITS		0xA014
 #define GEN6_RPSTAT1				0xA01C
 #define   GEN6_CAGF_SHIFT			8
+#define   HSW_CAGF_SHIFT			7
 #define   GEN6_CAGF_MASK			(0x7f << GEN6_CAGF_SHIFT)
+#define   HSW_CAGF_MASK				(0x7f << HSW_CAGF_SHIFT)
 #define GEN6_RP_CONTROL				0xA024
 #define   GEN6_RP_MEDIA_TURBO			(1<<11)
 #define   GEN6_RP_MEDIA_MODE_MASK		(3<<9)
-- 
cgit v0.10.2


From 725a5b54028916cd2511a251c5b5b13d1715addc Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Tue, 8 Jan 2013 11:02:57 +0000
Subject: drm/i915: Only run idle processing from
 i915_gem_retire_requests_worker

When adding the fb idle detection to mark-inactive, it was forgotten
that userspace can drive the processing of retire-requests. We assumed
that it would be principally driven by the retire requests worker,
running once every second whilst active and so we would get the deferred
timer for free. Instead we spend too many CPU cycles reclocking the LVDS
preventing real work from being done.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reported-and-tested-by: Alexander Lam <lambchop468@gmail.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=58843
Cc: stable@vger.kernel.org
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 801f77e..d746177 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1899,9 +1899,6 @@ i915_gem_object_move_to_inactive(struct drm_i915_gem_object *obj)
 	BUG_ON(obj->base.write_domain & ~I915_GEM_GPU_DOMAINS);
 	BUG_ON(!obj->active);
 
-	if (obj->pin_count) /* are we a framebuffer? */
-		intel_mark_fb_idle(obj);
-
 	list_move_tail(&obj->mm_list, &dev_priv->mm.inactive_list);
 
 	list_del_init(&obj->ring_list);
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 56c51dd..0d45487 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -6725,11 +6725,6 @@ void intel_mark_busy(struct drm_device *dev)
 
 void intel_mark_idle(struct drm_device *dev)
 {
-}
-
-void intel_mark_fb_busy(struct drm_i915_gem_object *obj)
-{
-	struct drm_device *dev = obj->base.dev;
 	struct drm_crtc *crtc;
 
 	if (!i915_powersave)
@@ -6739,12 +6734,11 @@ void intel_mark_fb_busy(struct drm_i915_gem_object *obj)
 		if (!crtc->fb)
 			continue;
 
-		if (to_intel_framebuffer(crtc->fb)->obj == obj)
-			intel_increase_pllclock(crtc);
+		intel_decrease_pllclock(crtc);
 	}
 }
 
-void intel_mark_fb_idle(struct drm_i915_gem_object *obj)
+void intel_mark_fb_busy(struct drm_i915_gem_object *obj)
 {
 	struct drm_device *dev = obj->base.dev;
 	struct drm_crtc *crtc;
@@ -6757,7 +6751,7 @@ void intel_mark_fb_idle(struct drm_i915_gem_object *obj)
 			continue;
 
 		if (to_intel_framebuffer(crtc->fb)->obj == obj)
-			intel_decrease_pllclock(crtc);
+			intel_increase_pllclock(crtc);
 	}
 }
 
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index fcdfe42..13afb37d 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -452,9 +452,8 @@ extern bool intel_sdvo_init(struct drm_device *dev, uint32_t sdvo_reg,
 extern void intel_dvo_init(struct drm_device *dev);
 extern void intel_tv_init(struct drm_device *dev);
 extern void intel_mark_busy(struct drm_device *dev);
-extern void intel_mark_idle(struct drm_device *dev);
 extern void intel_mark_fb_busy(struct drm_i915_gem_object *obj);
-extern void intel_mark_fb_idle(struct drm_i915_gem_object *obj);
+extern void intel_mark_idle(struct drm_device *dev);
 extern bool intel_lvds_init(struct drm_device *dev);
 extern bool intel_is_dual_link_lvds(struct drm_device *dev);
 extern void intel_dp_init(struct drm_device *dev, int output_reg,
-- 
cgit v0.10.2


From 6b25a88752e8e9dc33c700712059b094d3da3219 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Wed, 30 Jan 2013 15:59:57 +0100
Subject: drm/i915: kill cargo-culted locking from power well code

We may not concurrently change the power wells code. Which
is already guaranteed since modesets aren't concurrent. That
leaves races against setup/teardown/suspend/resume, and for
those we already (try) rather hard not to hit concurrent
modesets.

No debug WARN_ON added since that would require us to grab the
modeset locks in init/suspend code. Which is again just cargo
culting since just grabbing the locks in those paths isn't good
enough, we need the right order of operations, too.

Cc: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index 64d65f5..bec1f9f 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -4088,8 +4088,6 @@ void intel_init_power_well(struct drm_device *dev)
 	if (!IS_HASWELL(dev))
 		return;
 
-	mutex_lock(&dev->struct_mutex);
-
 	/* For now, we need the power well to be always enabled. */
 	intel_set_power_well(dev, true);
 
@@ -4097,8 +4095,6 @@ void intel_init_power_well(struct drm_device *dev)
 	 * the driver is in charge now. */
 	if (I915_READ(HSW_PWR_WELL_BIOS) & HSW_PWR_WELL_ENABLE)
 		I915_WRITE(HSW_PWR_WELL_BIOS, 0);
-
-	mutex_unlock(&dev->struct_mutex);
 }
 
 /* Set up chip specific power management-related functions */
-- 
cgit v0.10.2


From d5f21e4072645d0174ed4cc7197e0e30f84f96a2 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Wed, 30 Jan 2013 15:59:56 +0100
Subject: drm/i915: don't run hsw power well code on !hsw

Dumps annoying noise into the dmesg:

[drm:intel_set_power_well] *ERROR* Timeout enabling power well

Reported-by: Sedat Dilek <sedat.dilek@gmail.com>
Cc: Sedat Dilek <sedat.dilek@gmail.com>
Cc: Paulo Zanoni <paulo.r.zanoni@intel.com>
Tested-by: Sedat Dilek <sedat.dilek@gmail.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index bec1f9f..f024e7d 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -4053,6 +4053,9 @@ void intel_set_power_well(struct drm_device *dev, bool enable)
 	bool is_enabled, enable_requested;
 	uint32_t tmp;
 
+	if (!IS_HASWELL(dev))
+		return;
+
 	tmp = I915_READ(HSW_PWR_WELL_DRIVER);
 	is_enabled = tmp & HSW_PWR_WELL_STATE;
 	enable_requested = tmp & HSW_PWR_WELL_ENABLE;
-- 
cgit v0.10.2


From 10aa17c86f0de1fd902f7b8ee487a3682d7401df Mon Sep 17 00:00:00 2001
From: Paulo Zanoni <paulo.r.zanoni@intel.com>
Date: Tue, 29 Jan 2013 16:35:18 -0200
Subject: drm/i915: don't send DP "idle" pattern before "normal" on HSW PORT_A

The DP_TP_STATUS register for PORT_A doesn't exist. Our documentation
will be fixed soon, so the code does not match it for now.

This solves "Timed out waiting for DP idle patterns" and "unclaimed
register" messages on eDP.

V1: Was called "drm/i915: don't read DP_TP_STATUS(PORT_A)"
V2: Was called "drm/i915: don't send DP idle pattern before normal
pattern on HSW"
V3: Only change the code that touches PORT_A.

Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 51fd797..1b76b04 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -1785,14 +1785,18 @@ intel_dp_set_link_train(struct intel_dp *intel_dp,
 		temp &= ~DP_TP_CTL_LINK_TRAIN_MASK;
 		switch (dp_train_pat & DP_TRAINING_PATTERN_MASK) {
 		case DP_TRAINING_PATTERN_DISABLE:
-			temp |= DP_TP_CTL_LINK_TRAIN_IDLE;
-			I915_WRITE(DP_TP_CTL(port), temp);
 
-			if (wait_for((I915_READ(DP_TP_STATUS(port)) &
-				      DP_TP_STATUS_IDLE_DONE), 1))
-				DRM_ERROR("Timed out waiting for DP idle patterns\n");
+			if (port != PORT_A) {
+				temp |= DP_TP_CTL_LINK_TRAIN_IDLE;
+				I915_WRITE(DP_TP_CTL(port), temp);
+
+				if (wait_for((I915_READ(DP_TP_STATUS(port)) &
+					      DP_TP_STATUS_IDLE_DONE), 1))
+					DRM_ERROR("Timed out waiting for DP idle patterns\n");
+
+				temp &= ~DP_TP_CTL_LINK_TRAIN_MASK;
+			}
 
-			temp &= ~DP_TP_CTL_LINK_TRAIN_MASK;
 			temp |= DP_TP_CTL_LINK_TRAIN_NORMAL;
 
 			break;
-- 
cgit v0.10.2


From 693101618a4beedf1b3d291860aee56c5fe44a1c Mon Sep 17 00:00:00 2001
From: Paulo Zanoni <paulo.r.zanoni@intel.com>
Date: Tue, 29 Jan 2013 16:35:19 -0200
Subject: drm/i915: check the power down well on assert_pipe()

If the power well is disabled, we should not try to read its
registers, otherwise we'll get "unclaimed register" messages.

V2: Don't check whether the power well is enabled or not, just check
whether we asked it to be enabled or not: if we asked to disable the
power well, don't use the registers on it, even if it's still enabled.

V3: Fix bug that breaks all non-Haswell machines.

Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 0d45487..429b66b 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -1214,9 +1214,15 @@ void assert_pipe(struct drm_i915_private *dev_priv,
 	if (pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE)
 		state = true;
 
-	reg = PIPECONF(cpu_transcoder);
-	val = I915_READ(reg);
-	cur_state = !!(val & PIPECONF_ENABLE);
+	if (IS_HASWELL(dev_priv->dev) && cpu_transcoder != TRANSCODER_EDP &&
+	    !(I915_READ(HSW_PWR_WELL_DRIVER) & HSW_PWR_WELL_ENABLE)) {
+		cur_state = false;
+	} else {
+		reg = PIPECONF(cpu_transcoder);
+		val = I915_READ(reg);
+		cur_state = !!(val & PIPECONF_ENABLE);
+	}
+
 	WARN(cur_state != state,
 	     "pipe %c assertion failure (expected %s, current %s)\n",
 	     pipe_name(pipe), state_string(state), state_string(cur_state));
-- 
cgit v0.10.2


From d6dd9eb1d96d2b7345fe4664066c2b7ed86da898 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Tue, 29 Jan 2013 16:35:20 -0200
Subject: drm/i915: dynamic Haswell display power well support

We can disable (almost) all the display hw if we only use pipe A, with
the integrated edp transcoder on port A. Because we don't set the cpu
transcoder that early (yet), we need to help us with a trick to simply
check for any edp encoders.

v2: Paulo Zanoni pointed out that we also need to configure the eDP
cpu transcoder correctly.

v3: Made by Paulo Zanoni
  - Rebase patch to be on top of "fix intel_init_power_wells" patch
  - Fix typos
  - Fix a small bug by adding a "connectors_active" check
  - Restore the initial code that unconditionally enables the power
    well when taking over from the BIOS

v4: Made by Paulo Zanoni
  - One more typo spotted by Jani Nikula

v5: Made by Paulo Zanoni
  - Rebase

Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c
index 33b9112..cedf4ab 100644
--- a/drivers/gpu/drm/i915/intel_ddi.c
+++ b/drivers/gpu/drm/i915/intel_ddi.c
@@ -988,7 +988,13 @@ void intel_ddi_enable_pipe_func(struct drm_crtc *crtc)
 	if (cpu_transcoder == TRANSCODER_EDP) {
 		switch (pipe) {
 		case PIPE_A:
-			temp |= TRANS_DDI_EDP_INPUT_A_ONOFF;
+			/* Can only use the always-on power well for eDP when
+			 * not using the panel fitter, and when not using motion
+			  * blur mitigation (which we don't support). */
+			if (dev_priv->pch_pf_size)
+				temp |= TRANS_DDI_EDP_INPUT_A_ONOFF;
+			else
+				temp |= TRANS_DDI_EDP_INPUT_A_ON;
 			break;
 		case PIPE_B:
 			temp |= TRANS_DDI_EDP_INPUT_B_ONOFF;
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 429b66b..1f1d045 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -5607,6 +5607,35 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 	return fdi_config_ok ? ret : -EINVAL;
 }
 
+static void haswell_modeset_global_resources(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	bool enable = false;
+	struct intel_crtc *crtc;
+	struct intel_encoder *encoder;
+
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, base.head) {
+		if (crtc->pipe != PIPE_A && crtc->base.enabled)
+			enable = true;
+		/* XXX: Should check for edp transcoder here, but thanks to init
+		 * sequence that's not yet available. Just in case desktop eDP
+		 * on PORT D is possible on haswell, too. */
+	}
+
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
+			    base.head) {
+		if (encoder->type != INTEL_OUTPUT_EDP &&
+		    encoder->connectors_active)
+			enable = true;
+	}
+
+	/* Even the eDP panel fitter is outside the always-on well. */
+	if (dev_priv->pch_pf_size)
+		enable = true;
+
+	intel_set_power_well(dev, enable);
+}
+
 static int haswell_crtc_mode_set(struct drm_crtc *crtc,
 				 struct drm_display_mode *mode,
 				 struct drm_display_mode *adjusted_mode,
@@ -8480,6 +8509,8 @@ static void intel_init_display(struct drm_device *dev)
 		} else if (IS_HASWELL(dev)) {
 			dev_priv->display.fdi_link_train = hsw_fdi_link_train;
 			dev_priv->display.write_eld = haswell_write_eld;
+			dev_priv->display.modeset_global_resources =
+				haswell_modeset_global_resources;
 		}
 	} else if (IS_G4X(dev)) {
 		dev_priv->display.write_eld = g4x_write_eld;
-- 
cgit v0.10.2


From d93c6233543cdab83eca4d0590737c3e99314fd6 Mon Sep 17 00:00:00 2001
From: Changlong Xie <changlongx.xie@intel.com>
Date: Thu, 31 Jan 2013 11:32:50 +0800
Subject: drm/i915: gen6_gmch_remove can be static

Signed-off-by: Changlong Xie <changlongx.xie@intel.com>
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 71f09a8..bdaca3f 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -773,7 +773,7 @@ static int gen6_gmch_probe(struct drm_device *dev,
 	return ret;
 }
 
-void gen6_gmch_remove(struct drm_device *dev)
+static void gen6_gmch_remove(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	iounmap(dev_priv->gtt.gsm);
-- 
cgit v0.10.2


From 766aa1c42362eafc53f5939cd3f83113e9d062ed Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Fri, 25 Jan 2013 21:44:46 +0200
Subject: drm/i915: Introduce i915_vgacntrl_reg()
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The VGACNTRL register has moved around between different platforms.
To handle the differences add i915_vgacntrl_reg() which returns the
correct offset for the VGACNTRL register.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index d327416..984523d 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -1887,4 +1887,14 @@ __i915_write(64, q)
 #define INTEL_BROADCAST_RGB_FULL 1
 #define INTEL_BROADCAST_RGB_LIMITED 2
 
+static inline uint32_t i915_vgacntrl_reg(struct drm_device *dev)
+{
+	if (HAS_PCH_SPLIT(dev))
+		return CPU_VGACNTRL;
+	else if (IS_VALLEYVIEW(dev))
+		return VLV_VGACNTRL;
+	else
+		return VGACNTRL;
+}
+
 #endif
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 9a3cd04..e2b592a 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -3238,6 +3238,8 @@
 # define VGA_2X_MODE				(1 << 30)
 # define VGA_PIPE_B_SELECT			(1 << 29)
 
+#define VLV_VGACNTRL		(VLV_DISPLAY_BASE + 0x71400)
+
 /* Ironlake */
 
 #define CPU_VGACNTRL	0x41000
diff --git a/drivers/gpu/drm/i915/i915_suspend.c b/drivers/gpu/drm/i915/i915_suspend.c
index 3911e04..2135f21 100644
--- a/drivers/gpu/drm/i915/i915_suspend.c
+++ b/drivers/gpu/drm/i915/i915_suspend.c
@@ -73,10 +73,7 @@ static void i915_save_vga(struct drm_device *dev)
 	dev_priv->regfile.saveVGA0 = I915_READ(VGA0);
 	dev_priv->regfile.saveVGA1 = I915_READ(VGA1);
 	dev_priv->regfile.saveVGA_PD = I915_READ(VGA_PD);
-	if (HAS_PCH_SPLIT(dev))
-		dev_priv->regfile.saveVGACNTRL = I915_READ(CPU_VGACNTRL);
-	else
-		dev_priv->regfile.saveVGACNTRL = I915_READ(VGACNTRL);
+	dev_priv->regfile.saveVGACNTRL = I915_READ(i915_vgacntrl_reg(dev));
 
 	/* VGA color palette registers */
 	dev_priv->regfile.saveDACMASK = I915_READ8(VGA_DACMASK);
@@ -137,10 +134,7 @@ static void i915_restore_vga(struct drm_device *dev)
 	u16 cr_index, cr_data, st01;
 
 	/* VGA state */
-	if (HAS_PCH_SPLIT(dev))
-		I915_WRITE(CPU_VGACNTRL, dev_priv->regfile.saveVGACNTRL);
-	else
-		I915_WRITE(VGACNTRL, dev_priv->regfile.saveVGACNTRL);
+	I915_WRITE(i915_vgacntrl_reg(dev), dev_priv->regfile.saveVGACNTRL);
 
 	I915_WRITE(VGA0, dev_priv->regfile.saveVGA0);
 	I915_WRITE(VGA1, dev_priv->regfile.saveVGA1);
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 1f1d045..62f4590 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -8670,12 +8670,7 @@ static void i915_disable_vga(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u8 sr1;
-	u32 vga_reg;
-
-	if (HAS_PCH_SPLIT(dev))
-		vga_reg = CPU_VGACNTRL;
-	else
-		vga_reg = VGACNTRL;
+	u32 vga_reg = i915_vgacntrl_reg(dev);
 
 	vga_get_uninterruptible(dev->pdev, VGA_RSRC_LEGACY_IO);
 	outb(SR01, VGA_SR_INDEX);
@@ -8937,12 +8932,7 @@ static void intel_sanitize_encoder(struct intel_encoder *encoder)
 void i915_redisable_vga(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 vga_reg;
-
-	if (HAS_PCH_SPLIT(dev))
-		vga_reg = CPU_VGACNTRL;
-	else
-		vga_reg = VGACNTRL;
+	u32 vga_reg = i915_vgacntrl_reg(dev);
 
 	if (I915_READ(vga_reg) != VGA_DISP_DISABLE) {
 		DRM_DEBUG_KMS("Something enabled VGA plane, disabling it\n");
-- 
cgit v0.10.2


From fe31b574fb2563f3aaa24dec302ddd0a033cda0f Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Fri, 25 Jan 2013 21:44:47 +0200
Subject: drm/i915: Kill IS_DISPLAYREG()
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

All display registers should now include the proper offset on VLV.
That means IS_DISPLAYREG() is now useless, and we can eliminate it.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index 9cc8f87..d159d7a 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -1119,102 +1119,6 @@ MODULE_LICENSE("GPL and additional rights");
 	((HAS_FORCE_WAKE((dev_priv)->dev)) && \
 	 ((reg) < 0x40000) &&            \
 	 ((reg) != FORCEWAKE))
-
-static bool IS_DISPLAYREG(u32 reg)
-{
-	/*
-	 * This should make it easier to transition modules over to the
-	 * new register block scheme, since we can do it incrementally.
-	 */
-	if (reg >= VLV_DISPLAY_BASE)
-		return false;
-
-	if (reg >= RENDER_RING_BASE &&
-	    reg < RENDER_RING_BASE + 0xff)
-		return false;
-	if (reg >= GEN6_BSD_RING_BASE &&
-	    reg < GEN6_BSD_RING_BASE + 0xff)
-		return false;
-	if (reg >= BLT_RING_BASE &&
-	    reg < BLT_RING_BASE + 0xff)
-		return false;
-
-	if (reg == PGTBL_ER)
-		return false;
-
-	if (reg >= IPEIR_I965 &&
-	    reg < HWSTAM)
-		return false;
-
-	if (reg == MI_MODE)
-		return false;
-
-	if (reg == GFX_MODE_GEN7)
-		return false;
-
-	if (reg == RENDER_HWS_PGA_GEN7 ||
-	    reg == BSD_HWS_PGA_GEN7 ||
-	    reg == BLT_HWS_PGA_GEN7)
-		return false;
-
-	if (reg == GEN6_BSD_SLEEP_PSMI_CONTROL ||
-	    reg == GEN6_BSD_RNCID)
-		return false;
-
-	if (reg == GEN6_BLITTER_ECOSKPD)
-		return false;
-
-	if (reg >= 0x4000c &&
-	    reg <= 0x4002c)
-		return false;
-
-	if (reg >= 0x4f000 &&
-	    reg <= 0x4f08f)
-		return false;
-
-	if (reg >= 0x4f100 &&
-	    reg <= 0x4f11f)
-		return false;
-
-	if (reg >= VLV_MASTER_IER &&
-	    reg <= GEN6_PMIER)
-		return false;
-
-	if (reg >= FENCE_REG_SANDYBRIDGE_0 &&
-	    reg < (FENCE_REG_SANDYBRIDGE_0 + (16*8)))
-		return false;
-
-	if (reg >= VLV_IIR_RW &&
-	    reg <= VLV_ISR)
-		return false;
-
-	if (reg == FORCEWAKE_VLV ||
-	    reg == FORCEWAKE_ACK_VLV)
-		return false;
-
-	if (reg == GEN6_GDRST)
-		return false;
-
-	switch (reg) {
-	case _3D_CHICKEN3:
-	case IVB_CHICKEN3:
-	case GEN7_COMMON_SLICE_CHICKEN1:
-	case GEN7_L3CNTLREG1:
-	case GEN7_L3_CHICKEN_MODE_REGISTER:
-	case GEN7_ROW_CHICKEN2:
-	case GEN7_L3SQCREG4:
-	case GEN7_SQ_CHICKEN_MBCUNIT_CONFIG:
-	case GEN7_HALF_SLICE_CHICKEN1:
-	case GEN6_MBCTL:
-	case GEN6_UCGCTL2:
-		return false;
-	default:
-		break;
-	}
-
-	return true;
-}
-
 static void
 ilk_dummy_write(struct drm_i915_private *dev_priv)
 {
@@ -1238,8 +1142,6 @@ u##x i915_read##x(struct drm_i915_private *dev_priv, u32 reg) { \
 		if (dev_priv->forcewake_count == 0) \
 			dev_priv->gt.force_wake_put(dev_priv); \
 		spin_unlock_irqrestore(&dev_priv->gt_lock, irqflags); \
-	} else if (IS_VALLEYVIEW(dev_priv->dev) && IS_DISPLAYREG(reg)) { \
-		val = read##y(dev_priv->regs + reg + 0x180000);		\
 	} else { \
 		val = read##y(dev_priv->regs + reg); \
 	} \
@@ -1266,11 +1168,7 @@ void i915_write##x(struct drm_i915_private *dev_priv, u32 reg, u##x val) { \
 		DRM_ERROR("Unknown unclaimed register before writing to %x\n", reg); \
 		I915_WRITE_NOTRACE(GEN7_ERR_INT, ERR_INT_MMIO_UNCLAIMED); \
 	} \
-	if (IS_VALLEYVIEW(dev_priv->dev) && IS_DISPLAYREG(reg)) { \
-		write##y(val, dev_priv->regs + reg + 0x180000);		\
-	} else {							\
-		write##y(val, dev_priv->regs + reg);			\
-	}								\
+	write##y(val, dev_priv->regs + reg); \
 	if (unlikely(__fifo_ret)) { \
 		gen6_gt_check_fifodbg(dev_priv); \
 	} \
-- 
cgit v0.10.2


From 209d52110a32c2069b5d870504e73fdb0e30fc51 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Fri, 25 Jan 2013 21:44:48 +0200
Subject: drm/i915: Set the SR01 "screen off" bit in i915_redisable_vga() too
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

From BSpec / SR01 - Clocking Mode:
"The following sequence must be used when disabling the VGA plane.
 Write SR01 to set bit 5 = 1 to disable video output.
 Wait for 100us.
 Disable the VGA plane via Bit 31 of the MMIO VGA control."

So simply call i915_disable_vga() from i915_redisable_vga().

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 62f4590..d75c6a0 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -8936,8 +8936,7 @@ void i915_redisable_vga(struct drm_device *dev)
 
 	if (I915_READ(vga_reg) != VGA_DISP_DISABLE) {
 		DRM_DEBUG_KMS("Something enabled VGA plane, disabling it\n");
-		I915_WRITE(vga_reg, VGA_DISP_DISABLE);
-		POSTING_READ(vga_reg);
+		i915_disable_vga(dev);
 	}
 }
 
-- 
cgit v0.10.2


From 580f839892d659097459eaf3b52fb1af1a83f358 Mon Sep 17 00:00:00 2001
From: Ilija Hadzic <ihadzic@research.bell-labs.com>
Date: Wed, 2 Jan 2013 18:27:37 -0500
Subject: drm/radeon: remove unecessary assignment
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

length_dw field was assigned twice. While at it, move user_ptr
assignment together with all other assignments to p->chunks[i]
structure to make the code more readable.

Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
Reviewed-by: Marek Olšák <maraeo@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c
index 5407459..56f9282 100644
--- a/drivers/gpu/drm/radeon/radeon_cs.c
+++ b/drivers/gpu/drm/radeon/radeon_cs.c
@@ -203,7 +203,7 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 		p->chunks[i].length_dw = user_chunk.length_dw;
 		p->chunks[i].kdata = NULL;
 		p->chunks[i].chunk_id = user_chunk.chunk_id;
-
+		p->chunks[i].user_ptr = (void __user *)(unsigned long)user_chunk.chunk_data;
 		if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_RELOCS) {
 			p->chunk_relocs_idx = i;
 		}
@@ -226,9 +226,6 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 				return -EINVAL;
 		}
 
-		p->chunks[i].length_dw = user_chunk.length_dw;
-		p->chunks[i].user_ptr = (void __user *)(unsigned long)user_chunk.chunk_data;
-
 		cdata = (uint32_t *)(unsigned long)user_chunk.chunk_data;
 		if ((p->chunks[i].chunk_id == RADEON_CHUNK_ID_RELOCS) ||
 		    (p->chunks[i].chunk_id == RADEON_CHUNK_ID_FLAGS)) {
-- 
cgit v0.10.2


From d20145043524c4f28d090f5176cf9bf5cc7306c6 Mon Sep 17 00:00:00 2001
From: Ilija Hadzic <ihadzic@research.bell-labs.com>
Date: Wed, 2 Jan 2013 18:27:38 -0500
Subject: drm/radeon: remove unused prototype from radeon_cs
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
Reviewed-by: Marek Olšák <maraeo@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c
index 56f9282..0a23589 100644
--- a/drivers/gpu/drm/radeon/radeon_cs.c
+++ b/drivers/gpu/drm/radeon/radeon_cs.c
@@ -29,9 +29,6 @@
 #include "radeon_reg.h"
 #include "radeon.h"
 
-void r100_cs_dump_packet(struct radeon_cs_parser *p,
-			 struct radeon_cs_packet *pkt);
-
 static int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
 {
 	struct drm_device *ddev = p->rdev->ddev;
-- 
cgit v0.10.2


From 66b3543ef38216bdaf529a207fb495c500eb98be Mon Sep 17 00:00:00 2001
From: Ilija Hadzic <ihadzic@research.bell-labs.com>
Date: Wed, 2 Jan 2013 18:27:39 -0500
Subject: drm/radeon: fix formatting
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Preparatory patch: patches to follow will touch a piece of code
that had broken indentication, so fix it before touching it.

Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
Reviewed-by: Marek Olšák <maraeo@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c
index 8ff7cac..40c0318 100644
--- a/drivers/gpu/drm/radeon/r100.c
+++ b/drivers/gpu/drm/radeon/r100.c
@@ -2106,31 +2106,30 @@ int r100_cs_parse(struct radeon_cs_parser *p)
 		}
 		p->idx += pkt.count + 2;
 		switch (pkt.type) {
-			case PACKET_TYPE0:
-				if (p->rdev->family >= CHIP_R200)
-					r = r100_cs_parse_packet0(p, &pkt,
-								  p->rdev->config.r100.reg_safe_bm,
-								  p->rdev->config.r100.reg_safe_bm_size,
-								  &r200_packet0_check);
-				else
-					r = r100_cs_parse_packet0(p, &pkt,
-								  p->rdev->config.r100.reg_safe_bm,
-								  p->rdev->config.r100.reg_safe_bm_size,
-								  &r100_packet0_check);
-				break;
-			case PACKET_TYPE2:
-				break;
-			case PACKET_TYPE3:
-				r = r100_packet3_check(p, &pkt);
-				break;
-			default:
-				DRM_ERROR("Unknown packet type %d !\n",
-					  pkt.type);
-				return -EINVAL;
+		case PACKET_TYPE0:
+			if (p->rdev->family >= CHIP_R200)
+				r = r100_cs_parse_packet0(p, &pkt,
+					p->rdev->config.r100.reg_safe_bm,
+					p->rdev->config.r100.reg_safe_bm_size,
+					&r200_packet0_check);
+			else
+				r = r100_cs_parse_packet0(p, &pkt,
+					p->rdev->config.r100.reg_safe_bm,
+					p->rdev->config.r100.reg_safe_bm_size,
+					&r100_packet0_check);
+			break;
+		case PACKET_TYPE2:
+			break;
+		case PACKET_TYPE3:
+			r = r100_packet3_check(p, &pkt);
+			break;
+		default:
+			DRM_ERROR("Unknown packet type %d !\n",
+				  pkt.type);
+			return -EINVAL;
 		}
-		if (r) {
+		if (r)
 			return r;
-		}
 	} while (p->idx < p->chunks[p->chunk_ib_idx].length_dw);
 	return 0;
 }
-- 
cgit v0.10.2


From 4db013110cd3da05e4cf7e1119468817709cb9db Mon Sep 17 00:00:00 2001
From: Ilija Hadzic <ihadzic@research.bell-labs.com>
Date: Wed, 2 Jan 2013 18:27:40 -0500
Subject: drm/radeon: implement common cs packet parse function
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

CS packet parse functions have a lot of in common across
all ASICs. Implement a common function and take care of
small differences between families inside the function.

This patch is a prep for major refactoring and consolidation
of CS parsing code.

Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
Reviewed-by: Marek Olšák <maraeo@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c
index 0a23589..7355a36 100644
--- a/drivers/gpu/drm/radeon/radeon_cs.c
+++ b/drivers/gpu/drm/radeon/radeon_cs.c
@@ -642,3 +642,56 @@ u32 radeon_get_ib_value(struct radeon_cs_parser *p, int idx)
 	idx_value = ibc->kpage[new_page][pg_offset/4];
 	return idx_value;
 }
+
+/**
+ * radeon_cs_packet_parse() - parse cp packet and point ib index to next packet
+ * @parser:	parser structure holding parsing context.
+ * @pkt:	where to store packet information
+ *
+ * Assume that chunk_ib_index is properly set. Will return -EINVAL
+ * if packet is bigger than remaining ib size. or if packets is unknown.
+ **/
+int radeon_cs_packet_parse(struct radeon_cs_parser *p,
+			   struct radeon_cs_packet *pkt,
+			   unsigned idx)
+{
+	struct radeon_cs_chunk *ib_chunk = &p->chunks[p->chunk_ib_idx];
+	struct radeon_device *rdev = p->rdev;
+	uint32_t header;
+
+	if (idx >= ib_chunk->length_dw) {
+		DRM_ERROR("Can not parse packet at %d after CS end %d !\n",
+			  idx, ib_chunk->length_dw);
+		return -EINVAL;
+	}
+	header = radeon_get_ib_value(p, idx);
+	pkt->idx = idx;
+	pkt->type = RADEON_CP_PACKET_GET_TYPE(header);
+	pkt->count = RADEON_CP_PACKET_GET_COUNT(header);
+	pkt->one_reg_wr = 0;
+	switch (pkt->type) {
+	case RADEON_PACKET_TYPE0:
+		if (rdev->family < CHIP_R600) {
+			pkt->reg = R100_CP_PACKET0_GET_REG(header);
+			pkt->one_reg_wr =
+				RADEON_CP_PACKET0_GET_ONE_REG_WR(header);
+		} else
+			pkt->reg = R600_CP_PACKET0_GET_REG(header);
+		break;
+	case RADEON_PACKET_TYPE3:
+		pkt->opcode = RADEON_CP_PACKET3_GET_OPCODE(header);
+		break;
+	case RADEON_PACKET_TYPE2:
+		pkt->count = -1;
+		break;
+	default:
+		DRM_ERROR("Unknown packet type %d at %d !\n", pkt->type, idx);
+		return -EINVAL;
+	}
+	if ((pkt->count + 1 + pkt->idx) >= ib_chunk->length_dw) {
+		DRM_ERROR("Packet (%d:%d:%d) end after CS buffer (%d) !\n",
+			  pkt->idx, pkt->type, pkt->count, ib_chunk->length_dw);
+		return -EINVAL;
+	}
+	return 0;
+}
diff --git a/drivers/gpu/drm/radeon/radeon_reg.h b/drivers/gpu/drm/radeon/radeon_reg.h
index 5d8f735..d9e4204 100644
--- a/drivers/gpu/drm/radeon/radeon_reg.h
+++ b/drivers/gpu/drm/radeon/radeon_reg.h
@@ -3706,4 +3706,15 @@
 
 #define RV530_GB_PIPE_SELECT2           0x4124
 
+#define RADEON_CP_PACKET_GET_TYPE(h) (((h) >> 30) & 3)
+#define RADEON_CP_PACKET_GET_COUNT(h) (((h) >> 16) & 0x3FFF)
+#define RADEON_CP_PACKET0_GET_ONE_REG_WR(h) (((h) >> 15) & 1)
+#define RADEON_CP_PACKET3_GET_OPCODE(h) (((h) >> 8) & 0xFF)
+#define R100_CP_PACKET0_GET_REG(h) (((h) & 0x1FFF) << 2)
+#define R600_CP_PACKET0_GET_REG(h) (((h) & 0xFFFF) << 2)
+#define RADEON_PACKET_TYPE0 0
+#define RADEON_PACKET_TYPE1 1
+#define RADEON_PACKET_TYPE2 2
+#define RADEON_PACKET_TYPE3 3
+
 #endif
-- 
cgit v0.10.2


From c38f34b53e74dec4e58fef0c895d9e6df7da1190 Mon Sep 17 00:00:00 2001
From: Ilija Hadzic <ihadzic@research.bell-labs.com>
Date: Wed, 2 Jan 2013 18:27:41 -0500
Subject: drm/radeon: use common cs packet parse function
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

We now have a common radeon_cs_packet_parse function
that is good for all ASICs. Hook it up and eliminate
ASIC-specific versions.

Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
Reviewed-by: Marek Olšák <maraeo@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/evergreen_cs.c b/drivers/gpu/drm/radeon/evergreen_cs.c
index 7a44566..1ba4ca3 100644
--- a/drivers/gpu/drm/radeon/evergreen_cs.c
+++ b/drivers/gpu/drm/radeon/evergreen_cs.c
@@ -1009,53 +1009,6 @@ static int evergreen_cs_track_check(struct radeon_cs_parser *p)
 }
 
 /**
- * evergreen_cs_packet_parse() - parse cp packet and point ib index to next packet
- * @parser:	parser structure holding parsing context.
- * @pkt:	where to store packet informations
- *
- * Assume that chunk_ib_index is properly set. Will return -EINVAL
- * if packet is bigger than remaining ib size. or if packets is unknown.
- **/
-static int evergreen_cs_packet_parse(struct radeon_cs_parser *p,
-			      struct radeon_cs_packet *pkt,
-			      unsigned idx)
-{
-	struct radeon_cs_chunk *ib_chunk = &p->chunks[p->chunk_ib_idx];
-	uint32_t header;
-
-	if (idx >= ib_chunk->length_dw) {
-		DRM_ERROR("Can not parse packet at %d after CS end %d !\n",
-			  idx, ib_chunk->length_dw);
-		return -EINVAL;
-	}
-	header = radeon_get_ib_value(p, idx);
-	pkt->idx = idx;
-	pkt->type = CP_PACKET_GET_TYPE(header);
-	pkt->count = CP_PACKET_GET_COUNT(header);
-	pkt->one_reg_wr = 0;
-	switch (pkt->type) {
-	case PACKET_TYPE0:
-		pkt->reg = CP_PACKET0_GET_REG(header);
-		break;
-	case PACKET_TYPE3:
-		pkt->opcode = CP_PACKET3_GET_OPCODE(header);
-		break;
-	case PACKET_TYPE2:
-		pkt->count = -1;
-		break;
-	default:
-		DRM_ERROR("Unknown packet type %d at %d !\n", pkt->type, idx);
-		return -EINVAL;
-	}
-	if ((pkt->count + 1 + pkt->idx) >= ib_chunk->length_dw) {
-		DRM_ERROR("Packet (%d:%d:%d) end after CS buffer (%d) !\n",
-			  pkt->idx, pkt->type, pkt->count, ib_chunk->length_dw);
-		return -EINVAL;
-	}
-	return 0;
-}
-
-/**
  * evergreen_cs_packet_next_reloc() - parse next packet which should be reloc packet3
  * @parser:		parser structure holding parsing context.
  * @data:		pointer to relocation data
@@ -1080,7 +1033,7 @@ static int evergreen_cs_packet_next_reloc(struct radeon_cs_parser *p,
 	}
 	*cs_reloc = NULL;
 	relocs_chunk = &p->chunks[p->chunk_relocs_idx];
-	r = evergreen_cs_packet_parse(p, &p3reloc, p->idx);
+	r = radeon_cs_packet_parse(p, &p3reloc, p->idx);
 	if (r) {
 		return r;
 	}
@@ -1112,7 +1065,7 @@ static bool evergreen_cs_packet_next_is_pkt3_nop(struct radeon_cs_parser *p)
 	struct radeon_cs_packet p3reloc;
 	int r;
 
-	r = evergreen_cs_packet_parse(p, &p3reloc, p->idx);
+	r = radeon_cs_packet_parse(p, &p3reloc, p->idx);
 	if (r) {
 		return false;
 	}
@@ -1150,7 +1103,7 @@ static int evergreen_cs_packet_parse_vline(struct radeon_cs_parser *p)
 	ib = p->ib.ptr;
 
 	/* parse the WAIT_REG_MEM */
-	r = evergreen_cs_packet_parse(p, &wait_reg_mem, p->idx);
+	r = radeon_cs_packet_parse(p, &wait_reg_mem, p->idx);
 	if (r)
 		return r;
 
@@ -1183,7 +1136,7 @@ static int evergreen_cs_packet_parse_vline(struct radeon_cs_parser *p)
 	}
 
 	/* jump over the NOP */
-	r = evergreen_cs_packet_parse(p, &p3reloc, p->idx + wait_reg_mem.count + 2);
+	r = radeon_cs_packet_parse(p, &p3reloc, p->idx + wait_reg_mem.count + 2);
 	if (r)
 		return r;
 
@@ -2819,7 +2772,7 @@ int evergreen_cs_parse(struct radeon_cs_parser *p)
 		p->track = track;
 	}
 	do {
-		r = evergreen_cs_packet_parse(p, &pkt, p->idx);
+		r = radeon_cs_packet_parse(p, &pkt, p->idx);
 		if (r) {
 			kfree(p->track);
 			p->track = NULL;
diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c
index 40c0318..7842447 100644
--- a/drivers/gpu/drm/radeon/r100.c
+++ b/drivers/gpu/drm/radeon/r100.c
@@ -1370,53 +1370,6 @@ void r100_cs_dump_packet(struct radeon_cs_parser *p,
 }
 
 /**
- * r100_cs_packet_parse() - parse cp packet and point ib index to next packet
- * @parser:	parser structure holding parsing context.
- * @pkt:	where to store packet informations
- *
- * Assume that chunk_ib_index is properly set. Will return -EINVAL
- * if packet is bigger than remaining ib size. or if packets is unknown.
- **/
-int r100_cs_packet_parse(struct radeon_cs_parser *p,
-			 struct radeon_cs_packet *pkt,
-			 unsigned idx)
-{
-	struct radeon_cs_chunk *ib_chunk = &p->chunks[p->chunk_ib_idx];
-	uint32_t header;
-
-	if (idx >= ib_chunk->length_dw) {
-		DRM_ERROR("Can not parse packet at %d after CS end %d !\n",
-			  idx, ib_chunk->length_dw);
-		return -EINVAL;
-	}
-	header = radeon_get_ib_value(p, idx);
-	pkt->idx = idx;
-	pkt->type = CP_PACKET_GET_TYPE(header);
-	pkt->count = CP_PACKET_GET_COUNT(header);
-	switch (pkt->type) {
-	case PACKET_TYPE0:
-		pkt->reg = CP_PACKET0_GET_REG(header);
-		pkt->one_reg_wr = CP_PACKET0_GET_ONE_REG_WR(header);
-		break;
-	case PACKET_TYPE3:
-		pkt->opcode = CP_PACKET3_GET_OPCODE(header);
-		break;
-	case PACKET_TYPE2:
-		pkt->count = -1;
-		break;
-	default:
-		DRM_ERROR("Unknown packet type %d at %d !\n", pkt->type, idx);
-		return -EINVAL;
-	}
-	if ((pkt->count + 1 + pkt->idx) >= ib_chunk->length_dw) {
-		DRM_ERROR("Packet (%d:%d:%d) end after CS buffer (%d) !\n",
-			  pkt->idx, pkt->type, pkt->count, ib_chunk->length_dw);
-		return -EINVAL;
-	}
-	return 0;
-}
-
-/**
  * r100_cs_packet_next_vline() - parse userspace VLINE packet
  * @parser:		parser structure holding parsing context.
  *
@@ -1444,7 +1397,7 @@ int r100_cs_packet_parse_vline(struct radeon_cs_parser *p)
 	ib = p->ib.ptr;
 
 	/* parse the wait until */
-	r = r100_cs_packet_parse(p, &waitreloc, p->idx);
+	r = radeon_cs_packet_parse(p, &waitreloc, p->idx);
 	if (r)
 		return r;
 
@@ -1461,7 +1414,7 @@ int r100_cs_packet_parse_vline(struct radeon_cs_parser *p)
 	}
 
 	/* jump over the NOP */
-	r = r100_cs_packet_parse(p, &p3reloc, p->idx + waitreloc.count + 2);
+	r = radeon_cs_packet_parse(p, &p3reloc, p->idx + waitreloc.count + 2);
 	if (r)
 		return r;
 
@@ -1531,7 +1484,7 @@ int r100_cs_packet_next_reloc(struct radeon_cs_parser *p,
 	}
 	*cs_reloc = NULL;
 	relocs_chunk = &p->chunks[p->chunk_relocs_idx];
-	r = r100_cs_packet_parse(p, &p3reloc, p->idx);
+	r = radeon_cs_packet_parse(p, &p3reloc, p->idx);
 	if (r) {
 		return r;
 	}
@@ -2100,7 +2053,7 @@ int r100_cs_parse(struct radeon_cs_parser *p)
 	r100_cs_track_clear(p->rdev, track);
 	p->track = track;
 	do {
-		r = r100_cs_packet_parse(p, &pkt, p->idx);
+		r = radeon_cs_packet_parse(p, &pkt, p->idx);
 		if (r) {
 			return r;
 		}
diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c
index d0ba602..34ca46e 100644
--- a/drivers/gpu/drm/radeon/r300.c
+++ b/drivers/gpu/drm/radeon/r300.c
@@ -1257,7 +1257,7 @@ int r300_cs_parse(struct radeon_cs_parser *p)
 	r100_cs_track_clear(p->rdev, track);
 	p->track = track;
 	do {
-		r = r100_cs_packet_parse(p, &pkt, p->idx);
+		r = radeon_cs_packet_parse(p, &pkt, p->idx);
 		if (r) {
 			return r;
 		}
diff --git a/drivers/gpu/drm/radeon/r600_cs.c b/drivers/gpu/drm/radeon/r600_cs.c
index 69ec24a..3870bf8 100644
--- a/drivers/gpu/drm/radeon/r600_cs.c
+++ b/drivers/gpu/drm/radeon/r600_cs.c
@@ -784,53 +784,6 @@ static int r600_cs_track_check(struct radeon_cs_parser *p)
 }
 
 /**
- * r600_cs_packet_parse() - parse cp packet and point ib index to next packet
- * @parser:	parser structure holding parsing context.
- * @pkt:	where to store packet informations
- *
- * Assume that chunk_ib_index is properly set. Will return -EINVAL
- * if packet is bigger than remaining ib size. or if packets is unknown.
- **/
-static int r600_cs_packet_parse(struct radeon_cs_parser *p,
-			struct radeon_cs_packet *pkt,
-			unsigned idx)
-{
-	struct radeon_cs_chunk *ib_chunk = &p->chunks[p->chunk_ib_idx];
-	uint32_t header;
-
-	if (idx >= ib_chunk->length_dw) {
-		DRM_ERROR("Can not parse packet at %d after CS end %d !\n",
-			  idx, ib_chunk->length_dw);
-		return -EINVAL;
-	}
-	header = radeon_get_ib_value(p, idx);
-	pkt->idx = idx;
-	pkt->type = CP_PACKET_GET_TYPE(header);
-	pkt->count = CP_PACKET_GET_COUNT(header);
-	pkt->one_reg_wr = 0;
-	switch (pkt->type) {
-	case PACKET_TYPE0:
-		pkt->reg = CP_PACKET0_GET_REG(header);
-		break;
-	case PACKET_TYPE3:
-		pkt->opcode = CP_PACKET3_GET_OPCODE(header);
-		break;
-	case PACKET_TYPE2:
-		pkt->count = -1;
-		break;
-	default:
-		DRM_ERROR("Unknown packet type %d at %d !\n", pkt->type, idx);
-		return -EINVAL;
-	}
-	if ((pkt->count + 1 + pkt->idx) >= ib_chunk->length_dw) {
-		DRM_ERROR("Packet (%d:%d:%d) end after CS buffer (%d) !\n",
-			  pkt->idx, pkt->type, pkt->count, ib_chunk->length_dw);
-		return -EINVAL;
-	}
-	return 0;
-}
-
-/**
  * r600_cs_packet_next_reloc_mm() - parse next packet which should be reloc packet3
  * @parser:		parser structure holding parsing context.
  * @data:		pointer to relocation data
@@ -855,7 +808,7 @@ static int r600_cs_packet_next_reloc_mm(struct radeon_cs_parser *p,
 	}
 	*cs_reloc = NULL;
 	relocs_chunk = &p->chunks[p->chunk_relocs_idx];
-	r = r600_cs_packet_parse(p, &p3reloc, p->idx);
+	r = radeon_cs_packet_parse(p, &p3reloc, p->idx);
 	if (r) {
 		return r;
 	}
@@ -901,7 +854,7 @@ static int r600_cs_packet_next_reloc_nomm(struct radeon_cs_parser *p,
 	}
 	*cs_reloc = NULL;
 	relocs_chunk = &p->chunks[p->chunk_relocs_idx];
-	r = r600_cs_packet_parse(p, &p3reloc, p->idx);
+	r = radeon_cs_packet_parse(p, &p3reloc, p->idx);
 	if (r) {
 		return r;
 	}
@@ -935,7 +888,7 @@ static int r600_cs_packet_next_is_pkt3_nop(struct radeon_cs_parser *p)
 	struct radeon_cs_packet p3reloc;
 	int r;
 
-	r = r600_cs_packet_parse(p, &p3reloc, p->idx);
+	r = radeon_cs_packet_parse(p, &p3reloc, p->idx);
 	if (r) {
 		return 0;
 	}
@@ -973,7 +926,7 @@ static int r600_cs_packet_parse_vline(struct radeon_cs_parser *p)
 	ib = p->ib.ptr;
 
 	/* parse the WAIT_REG_MEM */
-	r = r600_cs_packet_parse(p, &wait_reg_mem, p->idx);
+	r = radeon_cs_packet_parse(p, &wait_reg_mem, p->idx);
 	if (r)
 		return r;
 
@@ -1006,7 +959,7 @@ static int r600_cs_packet_parse_vline(struct radeon_cs_parser *p)
 	}
 
 	/* jump over the NOP */
-	r = r600_cs_packet_parse(p, &p3reloc, p->idx + wait_reg_mem.count + 2);
+	r = radeon_cs_packet_parse(p, &p3reloc, p->idx + wait_reg_mem.count + 2);
 	if (r)
 		return r;
 
@@ -2410,7 +2363,7 @@ int r600_cs_parse(struct radeon_cs_parser *p)
 		p->track = track;
 	}
 	do {
-		r = r600_cs_packet_parse(p, &pkt, p->idx);
+		r = radeon_cs_packet_parse(p, &pkt, p->idx);
 		if (r) {
 			kfree(p->track);
 			p->track = NULL;
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index a08f657..b8d1c5c 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -1972,6 +1972,10 @@ static inline int radeon_acpi_init(struct radeon_device *rdev) { return 0; }
 static inline void radeon_acpi_fini(struct radeon_device *rdev) { }
 #endif
 
+int radeon_cs_packet_parse(struct radeon_cs_parser *p,
+			   struct radeon_cs_packet *pkt,
+			   unsigned idx);
+
 #include "radeon_object.h"
 
 #endif
-- 
cgit v0.10.2


From 9ffb7a6dca4fd260db91c808efd4d5c56057600c Mon Sep 17 00:00:00 2001
From: Ilija Hadzic <ihadzic@research.bell-labs.com>
Date: Wed, 2 Jan 2013 18:27:42 -0500
Subject: drm/radeon: factor out cs_next_is_pkt3_nop function
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Once we factored out radeon_cs_packet_parse function,
evergreen_cs_next_is_pkt3_nop and r600_cs_next_is_pkt3_nop
functions became identical, so they can be factored out
into a common function.

Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
Reviewed-by: Marek Olšák <maraeo@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/evergreen_cs.c b/drivers/gpu/drm/radeon/evergreen_cs.c
index 1ba4ca3..883b9f7 100644
--- a/drivers/gpu/drm/radeon/evergreen_cs.c
+++ b/drivers/gpu/drm/radeon/evergreen_cs.c
@@ -1055,27 +1055,6 @@ static int evergreen_cs_packet_next_reloc(struct radeon_cs_parser *p,
 }
 
 /**
- * evergreen_cs_packet_next_is_pkt3_nop() - test if the next packet is NOP
- * @p:		structure holding the parser context.
- *
- * Check if the next packet is a relocation packet3.
- **/
-static bool evergreen_cs_packet_next_is_pkt3_nop(struct radeon_cs_parser *p)
-{
-	struct radeon_cs_packet p3reloc;
-	int r;
-
-	r = radeon_cs_packet_parse(p, &p3reloc, p->idx);
-	if (r) {
-		return false;
-	}
-	if (p3reloc.type != PACKET_TYPE3 || p3reloc.opcode != PACKET3_NOP) {
-		return false;
-	}
-	return true;
-}
-
-/**
  * evergreen_cs_packet_next_vline() - parse userspace VLINE packet
  * @parser:		parser structure holding parsing context.
  *
@@ -2464,7 +2443,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 
 				if ((tex_dim == SQ_TEX_DIM_2D_MSAA || tex_dim == SQ_TEX_DIM_2D_ARRAY_MSAA) &&
 				    !mip_address &&
-				    !evergreen_cs_packet_next_is_pkt3_nop(p)) {
+				    !radeon_cs_packet_next_is_pkt3_nop(p)) {
 					/* MIP_ADDRESS should point to FMASK for an MSAA texture.
 					 * It should be 0 if FMASK is disabled. */
 					moffset = 0;
diff --git a/drivers/gpu/drm/radeon/r600_cs.c b/drivers/gpu/drm/radeon/r600_cs.c
index 3870bf8..6e6fa97 100644
--- a/drivers/gpu/drm/radeon/r600_cs.c
+++ b/drivers/gpu/drm/radeon/r600_cs.c
@@ -877,28 +877,6 @@ static int r600_cs_packet_next_reloc_nomm(struct radeon_cs_parser *p,
 }
 
 /**
- * r600_cs_packet_next_is_pkt3_nop() - test if next packet is packet3 nop for reloc
- * @parser:		parser structure holding parsing context.
- *
- * Check next packet is relocation packet3, do bo validation and compute
- * GPU offset using the provided start.
- **/
-static int r600_cs_packet_next_is_pkt3_nop(struct radeon_cs_parser *p)
-{
-	struct radeon_cs_packet p3reloc;
-	int r;
-
-	r = radeon_cs_packet_parse(p, &p3reloc, p->idx);
-	if (r) {
-		return 0;
-	}
-	if (p3reloc.type != PACKET_TYPE3 || p3reloc.opcode != PACKET3_NOP) {
-		return 0;
-	}
-	return 1;
-}
-
-/**
  * r600_cs_packet_next_vline() - parse userspace VLINE packet
  * @parser:		parser structure holding parsing context.
  *
@@ -1108,7 +1086,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		break;
 	case R_028010_DB_DEPTH_INFO:
 		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS) &&
-		    r600_cs_packet_next_is_pkt3_nop(p)) {
+		    radeon_cs_packet_next_is_pkt3_nop(p)) {
 			r = r600_cs_packet_next_reloc(p, &reloc);
 			if (r) {
 				dev_warn(p->dev, "bad SET_CONTEXT_REG "
@@ -1209,7 +1187,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case R_0280B8_CB_COLOR6_INFO:
 	case R_0280BC_CB_COLOR7_INFO:
 		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS) &&
-		     r600_cs_packet_next_is_pkt3_nop(p)) {
+		     radeon_cs_packet_next_is_pkt3_nop(p)) {
 			r = r600_cs_packet_next_reloc(p, &reloc);
 			if (r) {
 				dev_err(p->dev, "bad SET_CONTEXT_REG 0x%04X\n", reg);
@@ -1273,7 +1251,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case R_0280F8_CB_COLOR6_FRAG:
 	case R_0280FC_CB_COLOR7_FRAG:
 		tmp = (reg - R_0280E0_CB_COLOR0_FRAG) / 4;
-		if (!r600_cs_packet_next_is_pkt3_nop(p)) {
+		if (!radeon_cs_packet_next_is_pkt3_nop(p)) {
 			if (!track->cb_color_base_last[tmp]) {
 				dev_err(p->dev, "Broken old userspace ? no cb_color0_base supplied before trying to write 0x%08X\n", reg);
 				return -EINVAL;
@@ -1304,7 +1282,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case R_0280D8_CB_COLOR6_TILE:
 	case R_0280DC_CB_COLOR7_TILE:
 		tmp = (reg - R_0280C0_CB_COLOR0_TILE) / 4;
-		if (!r600_cs_packet_next_is_pkt3_nop(p)) {
+		if (!radeon_cs_packet_next_is_pkt3_nop(p)) {
 			if (!track->cb_color_base_last[tmp]) {
 				dev_err(p->dev, "Broken old userspace ? no cb_color0_base supplied before trying to write 0x%08X\n", reg);
 				return -EINVAL;
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index b8d1c5c..72c4836 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -1975,6 +1975,8 @@ static inline void radeon_acpi_fini(struct radeon_device *rdev) { }
 int radeon_cs_packet_parse(struct radeon_cs_parser *p,
 			   struct radeon_cs_packet *pkt,
 			   unsigned idx);
+bool radeon_cs_packet_next_is_pkt3_nop(struct radeon_cs_parser *p);
+
 
 #include "radeon_object.h"
 
diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c
index 7355a36..56cdb44 100644
--- a/drivers/gpu/drm/radeon/radeon_cs.c
+++ b/drivers/gpu/drm/radeon/radeon_cs.c
@@ -695,3 +695,24 @@ int radeon_cs_packet_parse(struct radeon_cs_parser *p,
 	}
 	return 0;
 }
+
+/**
+ * radeon_cs_packet_next_is_pkt3_nop() - test if the next packet is P3 NOP
+ * @p:		structure holding the parser context.
+ *
+ * Check if the next packet is NOP relocation packet3.
+ **/
+bool radeon_cs_packet_next_is_pkt3_nop(struct radeon_cs_parser *p)
+{
+	struct radeon_cs_packet p3reloc;
+	int r;
+
+	r = radeon_cs_packet_parse(p, &p3reloc, p->idx);
+	if (r)
+		return false;
+	if (p3reloc.type != RADEON_PACKET_TYPE3)
+		return false;
+	if (p3reloc.opcode != RADEON_PACKET3_NOP)
+		return false;
+	return true;
+}
diff --git a/drivers/gpu/drm/radeon/radeon_reg.h b/drivers/gpu/drm/radeon/radeon_reg.h
index d9e4204..ce7e9f2 100644
--- a/drivers/gpu/drm/radeon/radeon_reg.h
+++ b/drivers/gpu/drm/radeon/radeon_reg.h
@@ -3717,4 +3717,6 @@
 #define RADEON_PACKET_TYPE2 2
 #define RADEON_PACKET_TYPE3 3
 
+#define RADEON_PACKET3_NOP 0x10
+
 #endif
-- 
cgit v0.10.2


From 40592a17b8747903be95338f461573916a71d739 Mon Sep 17 00:00:00 2001
From: Ilija Hadzic <ihadzic@research.bell-labs.com>
Date: Wed, 2 Jan 2013 18:27:43 -0500
Subject: drm/radeon: refactor vline packet parsing function
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

vline packet parsing function for R600 and Evergreen+ are
the same, except that they use different registers. Factor
out the algorithm into a common function that uses register
table passed from ASIC-specific caller.

This reduces ASIC-specific function to (trivial) setup
of register table and call into the common function.

Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
Reviewed-by: Marek Olšák <maraeo@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/evergreen_cs.c b/drivers/gpu/drm/radeon/evergreen_cs.c
index 883b9f7..2690532 100644
--- a/drivers/gpu/drm/radeon/evergreen_cs.c
+++ b/drivers/gpu/drm/radeon/evergreen_cs.c
@@ -1055,109 +1055,35 @@ static int evergreen_cs_packet_next_reloc(struct radeon_cs_parser *p,
 }
 
 /**
- * evergreen_cs_packet_next_vline() - parse userspace VLINE packet
+ * evergreen_cs_packet_parse_vline() - parse userspace VLINE packet
  * @parser:		parser structure holding parsing context.
  *
- * Userspace sends a special sequence for VLINE waits.
- * PACKET0 - VLINE_START_END + value
- * PACKET3 - WAIT_REG_MEM poll vline status reg
- * RELOC (P3) - crtc_id in reloc.
- *
- * This function parses this and relocates the VLINE START END
- * and WAIT_REG_MEM packets to the correct crtc.
- * It also detects a switched off crtc and nulls out the
- * wait in that case.
+ * This is an Evergreen(+)-specific function for parsing VLINE packets.
+ * Real work is done by r600_cs_common_vline_parse function.
+ * Here we just set up ASIC-specific register table and call
+ * the common implementation function.
  */
 static int evergreen_cs_packet_parse_vline(struct radeon_cs_parser *p)
 {
-	struct drm_mode_object *obj;
-	struct drm_crtc *crtc;
-	struct radeon_crtc *radeon_crtc;
-	struct radeon_cs_packet p3reloc, wait_reg_mem;
-	int crtc_id;
-	int r;
-	uint32_t header, h_idx, reg, wait_reg_mem_info;
-	volatile uint32_t *ib;
-
-	ib = p->ib.ptr;
-
-	/* parse the WAIT_REG_MEM */
-	r = radeon_cs_packet_parse(p, &wait_reg_mem, p->idx);
-	if (r)
-		return r;
-
-	/* check its a WAIT_REG_MEM */
-	if (wait_reg_mem.type != PACKET_TYPE3 ||
-	    wait_reg_mem.opcode != PACKET3_WAIT_REG_MEM) {
-		DRM_ERROR("vline wait missing WAIT_REG_MEM segment\n");
-		return -EINVAL;
-	}
-
-	wait_reg_mem_info = radeon_get_ib_value(p, wait_reg_mem.idx + 1);
-	/* bit 4 is reg (0) or mem (1) */
-	if (wait_reg_mem_info & 0x10) {
-		DRM_ERROR("vline WAIT_REG_MEM waiting on MEM rather than REG\n");
-		return -EINVAL;
-	}
-	/* waiting for value to be equal */
-	if ((wait_reg_mem_info & 0x7) != 0x3) {
-		DRM_ERROR("vline WAIT_REG_MEM function not equal\n");
-		return -EINVAL;
-	}
-	if ((radeon_get_ib_value(p, wait_reg_mem.idx + 2) << 2) != EVERGREEN_VLINE_STATUS) {
-		DRM_ERROR("vline WAIT_REG_MEM bad reg\n");
-		return -EINVAL;
-	}
 
-	if (radeon_get_ib_value(p, wait_reg_mem.idx + 5) != EVERGREEN_VLINE_STAT) {
-		DRM_ERROR("vline WAIT_REG_MEM bad bit mask\n");
-		return -EINVAL;
-	}
-
-	/* jump over the NOP */
-	r = radeon_cs_packet_parse(p, &p3reloc, p->idx + wait_reg_mem.count + 2);
-	if (r)
-		return r;
-
-	h_idx = p->idx - 2;
-	p->idx += wait_reg_mem.count + 2;
-	p->idx += p3reloc.count + 2;
-
-	header = radeon_get_ib_value(p, h_idx);
-	crtc_id = radeon_get_ib_value(p, h_idx + 2 + 7 + 1);
-	reg = CP_PACKET0_GET_REG(header);
-	obj = drm_mode_object_find(p->rdev->ddev, crtc_id, DRM_MODE_OBJECT_CRTC);
-	if (!obj) {
-		DRM_ERROR("cannot find crtc %d\n", crtc_id);
-		return -EINVAL;
-	}
-	crtc = obj_to_crtc(obj);
-	radeon_crtc = to_radeon_crtc(crtc);
-	crtc_id = radeon_crtc->crtc_id;
-
-	if (!crtc->enabled) {
-		/* if the CRTC isn't enabled - we need to nop out the WAIT_REG_MEM */
-		ib[h_idx + 2] = PACKET2(0);
-		ib[h_idx + 3] = PACKET2(0);
-		ib[h_idx + 4] = PACKET2(0);
-		ib[h_idx + 5] = PACKET2(0);
-		ib[h_idx + 6] = PACKET2(0);
-		ib[h_idx + 7] = PACKET2(0);
-		ib[h_idx + 8] = PACKET2(0);
-	} else {
-		switch (reg) {
-		case EVERGREEN_VLINE_START_END:
-			header &= ~R600_CP_PACKET0_REG_MASK;
-			header |= (EVERGREEN_VLINE_START_END + radeon_crtc->crtc_offset) >> 2;
-			ib[h_idx] = header;
-			ib[h_idx + 4] = (EVERGREEN_VLINE_STATUS + radeon_crtc->crtc_offset) >> 2;
-			break;
-		default:
-			DRM_ERROR("unknown crtc reloc\n");
-			return -EINVAL;
-		}
-	}
-	return 0;
+	static uint32_t vline_start_end[6] = {
+		EVERGREEN_VLINE_START_END + EVERGREEN_CRTC0_REGISTER_OFFSET,
+		EVERGREEN_VLINE_START_END + EVERGREEN_CRTC1_REGISTER_OFFSET,
+		EVERGREEN_VLINE_START_END + EVERGREEN_CRTC2_REGISTER_OFFSET,
+		EVERGREEN_VLINE_START_END + EVERGREEN_CRTC3_REGISTER_OFFSET,
+		EVERGREEN_VLINE_START_END + EVERGREEN_CRTC4_REGISTER_OFFSET,
+		EVERGREEN_VLINE_START_END + EVERGREEN_CRTC5_REGISTER_OFFSET
+	};
+	static uint32_t vline_status[6] = {
+		EVERGREEN_VLINE_STATUS + EVERGREEN_CRTC0_REGISTER_OFFSET,
+		EVERGREEN_VLINE_STATUS + EVERGREEN_CRTC1_REGISTER_OFFSET,
+		EVERGREEN_VLINE_STATUS + EVERGREEN_CRTC2_REGISTER_OFFSET,
+		EVERGREEN_VLINE_STATUS + EVERGREEN_CRTC3_REGISTER_OFFSET,
+		EVERGREEN_VLINE_STATUS + EVERGREEN_CRTC4_REGISTER_OFFSET,
+		EVERGREEN_VLINE_STATUS + EVERGREEN_CRTC5_REGISTER_OFFSET
+	};
+
+	return r600_cs_common_vline_parse(p, vline_start_end, vline_status);
 }
 
 static int evergreen_packet0_check(struct radeon_cs_parser *p,
diff --git a/drivers/gpu/drm/radeon/r600_cs.c b/drivers/gpu/drm/radeon/r600_cs.c
index 6e6fa97..e61a013 100644
--- a/drivers/gpu/drm/radeon/r600_cs.c
+++ b/drivers/gpu/drm/radeon/r600_cs.c
@@ -877,9 +877,30 @@ static int r600_cs_packet_next_reloc_nomm(struct radeon_cs_parser *p,
 }
 
 /**
- * r600_cs_packet_next_vline() - parse userspace VLINE packet
+ * r600_cs_packet_parse_vline() - parse userspace VLINE packet
  * @parser:		parser structure holding parsing context.
  *
+ * This is an R600-specific function for parsing VLINE packets.
+ * Real work is done by r600_cs_common_vline_parse function.
+ * Here we just set up ASIC-specific register table and call
+ * the common implementation function.
+ */
+static int r600_cs_packet_parse_vline(struct radeon_cs_parser *p)
+{
+	static uint32_t vline_start_end[2] = {AVIVO_D1MODE_VLINE_START_END,
+					      AVIVO_D2MODE_VLINE_START_END};
+	static uint32_t vline_status[2] = {AVIVO_D1MODE_VLINE_STATUS,
+					   AVIVO_D2MODE_VLINE_STATUS};
+
+	return r600_cs_common_vline_parse(p, vline_start_end, vline_status);
+}
+
+/**
+ * r600_cs_common_vline_parse() - common vline parser
+ * @parser:		parser structure holding parsing context.
+ * @vline_start_end:    table of vline_start_end registers
+ * @vline_status:       table of vline_status registers
+ *
  * Userspace sends a special sequence for VLINE waits.
  * PACKET0 - VLINE_START_END + value
  * PACKET3 - WAIT_REG_MEM poll vline status reg
@@ -888,9 +909,16 @@ static int r600_cs_packet_next_reloc_nomm(struct radeon_cs_parser *p,
  * This function parses this and relocates the VLINE START END
  * and WAIT_REG_MEM packets to the correct crtc.
  * It also detects a switched off crtc and nulls out the
- * wait in that case.
+ * wait in that case. This function is common for all ASICs that
+ * are R600 and newer. The parsing algorithm is the same, and only
+ * differs in which registers are used.
+ *
+ * Caller is the ASIC-specific function which passes the parser
+ * context and ASIC-specific register table
  */
-static int r600_cs_packet_parse_vline(struct radeon_cs_parser *p)
+int r600_cs_common_vline_parse(struct radeon_cs_parser *p,
+			       uint32_t *vline_start_end,
+			       uint32_t *vline_status)
 {
 	struct drm_mode_object *obj;
 	struct drm_crtc *crtc;
@@ -918,7 +946,7 @@ static int r600_cs_packet_parse_vline(struct radeon_cs_parser *p)
 	wait_reg_mem_info = radeon_get_ib_value(p, wait_reg_mem.idx + 1);
 	/* bit 4 is reg (0) or mem (1) */
 	if (wait_reg_mem_info & 0x10) {
-		DRM_ERROR("vline WAIT_REG_MEM waiting on MEM rather than REG\n");
+		DRM_ERROR("vline WAIT_REG_MEM waiting on MEM instead of REG\n");
 		return -EINVAL;
 	}
 	/* waiting for value to be equal */
@@ -926,12 +954,12 @@ static int r600_cs_packet_parse_vline(struct radeon_cs_parser *p)
 		DRM_ERROR("vline WAIT_REG_MEM function not equal\n");
 		return -EINVAL;
 	}
-	if ((radeon_get_ib_value(p, wait_reg_mem.idx + 2) << 2) != AVIVO_D1MODE_VLINE_STATUS) {
+	if ((radeon_get_ib_value(p, wait_reg_mem.idx + 2) << 2) != vline_status[0]) {
 		DRM_ERROR("vline WAIT_REG_MEM bad reg\n");
 		return -EINVAL;
 	}
 
-	if (radeon_get_ib_value(p, wait_reg_mem.idx + 5) != AVIVO_D1MODE_VLINE_STAT) {
+	if (radeon_get_ib_value(p, wait_reg_mem.idx + 5) != RADEON_VLINE_STAT) {
 		DRM_ERROR("vline WAIT_REG_MEM bad bit mask\n");
 		return -EINVAL;
 	}
@@ -959,7 +987,7 @@ static int r600_cs_packet_parse_vline(struct radeon_cs_parser *p)
 	crtc_id = radeon_crtc->crtc_id;
 
 	if (!crtc->enabled) {
-		/* if the CRTC isn't enabled - we need to nop out the WAIT_REG_MEM */
+		/* CRTC isn't enabled - we need to nop out the WAIT_REG_MEM */
 		ib[h_idx + 2] = PACKET2(0);
 		ib[h_idx + 3] = PACKET2(0);
 		ib[h_idx + 4] = PACKET2(0);
@@ -967,20 +995,15 @@ static int r600_cs_packet_parse_vline(struct radeon_cs_parser *p)
 		ib[h_idx + 6] = PACKET2(0);
 		ib[h_idx + 7] = PACKET2(0);
 		ib[h_idx + 8] = PACKET2(0);
-	} else if (crtc_id == 1) {
-		switch (reg) {
-		case AVIVO_D1MODE_VLINE_START_END:
-			header &= ~R600_CP_PACKET0_REG_MASK;
-			header |= AVIVO_D2MODE_VLINE_START_END >> 2;
-			break;
-		default:
-			DRM_ERROR("unknown crtc reloc\n");
-			return -EINVAL;
-		}
+	} else if (reg == vline_start_end[0]) {
+		header &= ~R600_CP_PACKET0_REG_MASK;
+		header |= vline_start_end[crtc_id] >> 2;
 		ib[h_idx] = header;
-		ib[h_idx + 4] = AVIVO_D2MODE_VLINE_STATUS >> 2;
+		ib[h_idx + 4] = vline_status[crtc_id] >> 2;
+	} else {
+		DRM_ERROR("unknown crtc reloc\n");
+		return -EINVAL;
 	}
-
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 72c4836..5caeca3 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -1976,7 +1976,9 @@ int radeon_cs_packet_parse(struct radeon_cs_parser *p,
 			   struct radeon_cs_packet *pkt,
 			   unsigned idx);
 bool radeon_cs_packet_next_is_pkt3_nop(struct radeon_cs_parser *p);
-
+int r600_cs_common_vline_parse(struct radeon_cs_parser *p,
+			       uint32_t *vline_start_end,
+			       uint32_t *vline_status);
 
 #include "radeon_object.h"
 
diff --git a/drivers/gpu/drm/radeon/radeon_reg.h b/drivers/gpu/drm/radeon/radeon_reg.h
index ce7e9f2..7e2c2b7 100644
--- a/drivers/gpu/drm/radeon/radeon_reg.h
+++ b/drivers/gpu/drm/radeon/radeon_reg.h
@@ -3719,4 +3719,6 @@
 
 #define RADEON_PACKET3_NOP 0x10
 
+#define RADEON_VLINE_STAT (1 << 12)
+
 #endif
-- 
cgit v0.10.2


From d6e18a3406d401edeb96a01c7bb9d1689454c41b Mon Sep 17 00:00:00 2001
From: Ilija Hadzic <ihadzic@research.bell-labs.com>
Date: Wed, 2 Jan 2013 18:27:44 -0500
Subject: drm/radeon: add a check to wait_reg_mem command
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

WAIT_REG_MEM on register does not allow the use of PFP.
Enforce this restriction when checking packets sent from
userland.

Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
Reviewed-by: Marek Olšák <maraeo@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/evergreen_cs.c b/drivers/gpu/drm/radeon/evergreen_cs.c
index 2690532..02aeb7f 100644
--- a/drivers/gpu/drm/radeon/evergreen_cs.c
+++ b/drivers/gpu/drm/radeon/evergreen_cs.c
@@ -2101,6 +2101,9 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 
 			ib[idx+1] = (ib[idx+1] & 0x3) | (offset & 0xfffffffc);
 			ib[idx+2] = upper_32_bits(offset) & 0xff;
+		} else if (idx_value & 0x100) {
+			DRM_ERROR("cannot use PFP on REG wait\n");
+			return -EINVAL;
 		}
 		break;
 	case PACKET3_CP_DMA:
diff --git a/drivers/gpu/drm/radeon/r600_cs.c b/drivers/gpu/drm/radeon/r600_cs.c
index e61a013..9d83372 100644
--- a/drivers/gpu/drm/radeon/r600_cs.c
+++ b/drivers/gpu/drm/radeon/r600_cs.c
@@ -949,6 +949,11 @@ int r600_cs_common_vline_parse(struct radeon_cs_parser *p,
 		DRM_ERROR("vline WAIT_REG_MEM waiting on MEM instead of REG\n");
 		return -EINVAL;
 	}
+	/* bit 8 is me (0) or pfp (1) */
+	if (wait_reg_mem_info & 0x100) {
+		DRM_ERROR("vline WAIT_REG_MEM waiting on PFP instead of ME\n");
+		return -EINVAL;
+	}
 	/* waiting for value to be equal */
 	if ((wait_reg_mem_info & 0x7) != 0x3) {
 		DRM_ERROR("vline WAIT_REG_MEM function not equal\n");
@@ -1847,6 +1852,9 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 
 			ib[idx+1] = (ib[idx+1] & 0x3) | (offset & 0xfffffff0);
 			ib[idx+2] = upper_32_bits(offset) & 0xff;
+		} else if (idx_value & 0x100) {
+			DRM_ERROR("cannot use PFP on REG wait\n");
+			return -EINVAL;
 		}
 		break;
 	case PACKET3_CP_DMA:
-- 
cgit v0.10.2


From c3ad63afcdb931159690aa7ba2906079c3b38a13 Mon Sep 17 00:00:00 2001
From: Ilija Hadzic <ihadzic@research.bell-labs.com>
Date: Wed, 2 Jan 2013 18:27:45 -0500
Subject: drm/radeon: rename r100_cs_dump_packet to radeon_cs_dump_packet
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

This function is not limited to r100, but it can dump a
(raw) packet for any ASIC. Rename it accordingly and move
its declaration to radeon.h

Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
Reviewed-by: Marek Olšák <maraeo@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c
index 7842447..cf7e359 100644
--- a/drivers/gpu/drm/radeon/r100.c
+++ b/drivers/gpu/drm/radeon/r100.c
@@ -1219,7 +1219,7 @@ int r100_reloc_pitch_offset(struct radeon_cs_parser *p,
 	if (r) {
 		DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 			  idx, reg);
-		r100_cs_dump_packet(p, pkt);
+		radeon_cs_dump_packet(p, pkt);
 		return r;
 	}
 
@@ -1233,7 +1233,7 @@ int r100_reloc_pitch_offset(struct radeon_cs_parser *p,
 		if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) {
 			if (reg == RADEON_SRC_PITCH_OFFSET) {
 				DRM_ERROR("Cannot src blit from microtiled surface\n");
-				r100_cs_dump_packet(p, pkt);
+				radeon_cs_dump_packet(p, pkt);
 				return -EINVAL;
 			}
 			tile_flags |= RADEON_DST_TILE_MICRO;
@@ -1263,7 +1263,7 @@ int r100_packet3_load_vbpntr(struct radeon_cs_parser *p,
 	if (c > 16) {
 	    DRM_ERROR("Only 16 vertex buffers are allowed %d\n",
 		      pkt->opcode);
-	    r100_cs_dump_packet(p, pkt);
+	    radeon_cs_dump_packet(p, pkt);
 	    return -EINVAL;
 	}
 	track->num_arrays = c;
@@ -1272,7 +1272,7 @@ int r100_packet3_load_vbpntr(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for packet3 %d\n",
 				  pkt->opcode);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		idx_value = radeon_get_ib_value(p, idx);
@@ -1285,7 +1285,7 @@ int r100_packet3_load_vbpntr(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for packet3 %d\n",
 				  pkt->opcode);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		ib[idx+2] = radeon_get_ib_value(p, idx + 2) + ((u32)reloc->lobj.gpu_offset);
@@ -1298,7 +1298,7 @@ int r100_packet3_load_vbpntr(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for packet3 %d\n",
 					  pkt->opcode);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		idx_value = radeon_get_ib_value(p, idx);
@@ -1355,20 +1355,6 @@ int r100_cs_parse_packet0(struct radeon_cs_parser *p,
 	return 0;
 }
 
-void r100_cs_dump_packet(struct radeon_cs_parser *p,
-			 struct radeon_cs_packet *pkt)
-{
-	volatile uint32_t *ib;
-	unsigned i;
-	unsigned idx;
-
-	ib = p->ib.ptr;
-	idx = pkt->idx;
-	for (i = 0; i <= (pkt->count + 1); i++, idx++) {
-		DRM_INFO("ib[%d]=0x%08X\n", idx, ib[idx]);
-	}
-}
-
 /**
  * r100_cs_packet_next_vline() - parse userspace VLINE packet
  * @parser:		parser structure holding parsing context.
@@ -1492,14 +1478,14 @@ int r100_cs_packet_next_reloc(struct radeon_cs_parser *p,
 	if (p3reloc.type != PACKET_TYPE3 || p3reloc.opcode != PACKET3_NOP) {
 		DRM_ERROR("No packet3 for relocation for packet at %d.\n",
 			  p3reloc.idx);
-		r100_cs_dump_packet(p, &p3reloc);
+		radeon_cs_dump_packet(p, &p3reloc);
 		return -EINVAL;
 	}
 	idx = radeon_get_ib_value(p, p3reloc.idx + 1);
 	if (idx >= relocs_chunk->length_dw) {
 		DRM_ERROR("Relocs at %d after relocations chunk end %d !\n",
 			  idx, relocs_chunk->length_dw);
-		r100_cs_dump_packet(p, &p3reloc);
+		radeon_cs_dump_packet(p, &p3reloc);
 		return -EINVAL;
 	}
 	/* FIXME: we assume reloc size is 4 dwords */
@@ -1584,7 +1570,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		break;
@@ -1601,7 +1587,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		track->zb.robj = reloc->robj;
@@ -1614,7 +1600,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		track->cb[0].robj = reloc->robj;
@@ -1630,7 +1616,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
@@ -1657,7 +1643,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		track->textures[0].cube_info[i].offset = idx_value;
@@ -1675,7 +1661,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		track->textures[1].cube_info[i].offset = idx_value;
@@ -1693,7 +1679,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		track->textures[2].cube_info[i].offset = idx_value;
@@ -1711,7 +1697,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
@@ -1782,7 +1768,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset);
@@ -1942,7 +1928,7 @@ static int r100_packet3_check(struct radeon_cs_parser *p,
 		r = r100_cs_packet_next_reloc(p, &reloc);
 		if (r) {
 			DRM_ERROR("No reloc for packet3 %d\n", pkt->opcode);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		ib[idx+1] = radeon_get_ib_value(p, idx+1) + ((u32)reloc->lobj.gpu_offset);
@@ -1956,7 +1942,7 @@ static int r100_packet3_check(struct radeon_cs_parser *p,
 		r = r100_cs_packet_next_reloc(p, &reloc);
 		if (r) {
 			DRM_ERROR("No reloc for packet3 %d\n", pkt->opcode);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		ib[idx] = radeon_get_ib_value(p, idx) + ((u32)reloc->lobj.gpu_offset);
diff --git a/drivers/gpu/drm/radeon/r100_track.h b/drivers/gpu/drm/radeon/r100_track.h
index 6a603b3..c81b174 100644
--- a/drivers/gpu/drm/radeon/r100_track.h
+++ b/drivers/gpu/drm/radeon/r100_track.h
@@ -83,8 +83,6 @@ int r100_cs_track_check(struct radeon_device *rdev, struct r100_cs_track *track)
 void r100_cs_track_clear(struct radeon_device *rdev, struct r100_cs_track *track);
 int r100_cs_packet_next_reloc(struct radeon_cs_parser *p,
 			      struct radeon_cs_reloc **cs_reloc);
-void r100_cs_dump_packet(struct radeon_cs_parser *p,
-			 struct radeon_cs_packet *pkt);
 
 int r100_cs_packet_parse_vline(struct radeon_cs_parser *p);
 
diff --git a/drivers/gpu/drm/radeon/r200.c b/drivers/gpu/drm/radeon/r200.c
index 98143a5..07ea6f4 100644
--- a/drivers/gpu/drm/radeon/r200.c
+++ b/drivers/gpu/drm/radeon/r200.c
@@ -162,7 +162,7 @@ int r200_packet0_check(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		break;
@@ -179,7 +179,7 @@ int r200_packet0_check(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		track->zb.robj = reloc->robj;
@@ -192,7 +192,7 @@ int r200_packet0_check(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		track->cb[0].robj = reloc->robj;
@@ -211,7 +211,7 @@ int r200_packet0_check(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
@@ -264,7 +264,7 @@ int r200_packet0_check(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		track->textures[i].cube_info[face - 1].offset = idx_value;
@@ -282,7 +282,7 @@ int r200_packet0_check(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 
@@ -359,7 +359,7 @@ int r200_packet0_check(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset);
diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c
index 34ca46e..47bfca1 100644
--- a/drivers/gpu/drm/radeon/r300.c
+++ b/drivers/gpu/drm/radeon/r300.c
@@ -615,7 +615,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 					idx, reg);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		break;
@@ -634,7 +634,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 					idx, reg);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		track->cb[i].robj = reloc->robj;
@@ -647,7 +647,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 					idx, reg);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		track->zb.robj = reloc->robj;
@@ -676,7 +676,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 					idx, reg);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 
@@ -749,7 +749,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 			if (r) {
 				DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 					  idx, reg);
-				r100_cs_dump_packet(p, pkt);
+				radeon_cs_dump_packet(p, pkt);
 				return r;
 			}
 
@@ -834,7 +834,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 			if (r) {
 				DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 					  idx, reg);
-				r100_cs_dump_packet(p, pkt);
+				radeon_cs_dump_packet(p, pkt);
 				return r;
 			}
 
@@ -1049,7 +1049,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 					idx, reg);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset);
@@ -1091,7 +1091,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		track->aa.robj = reloc->robj;
@@ -1159,7 +1159,7 @@ static int r300_packet3_check(struct radeon_cs_parser *p,
 		r = r100_cs_packet_next_reloc(p, &reloc);
 		if (r) {
 			DRM_ERROR("No reloc for packet3 %d\n", pkt->opcode);
-			r100_cs_dump_packet(p, pkt);
+			radeon_cs_dump_packet(p, pkt);
 			return r;
 		}
 		ib[idx+1] = radeon_get_ib_value(p, idx + 1) + ((u32)reloc->lobj.gpu_offset);
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 5caeca3..5f96430 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -1976,6 +1976,8 @@ int radeon_cs_packet_parse(struct radeon_cs_parser *p,
 			   struct radeon_cs_packet *pkt,
 			   unsigned idx);
 bool radeon_cs_packet_next_is_pkt3_nop(struct radeon_cs_parser *p);
+void radeon_cs_dump_packet(struct radeon_cs_parser *p,
+			   struct radeon_cs_packet *pkt);
 int r600_cs_common_vline_parse(struct radeon_cs_parser *p,
 			       uint32_t *vline_start_end,
 			       uint32_t *vline_status);
diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c
index 56cdb44..7f48d46 100644
--- a/drivers/gpu/drm/radeon/radeon_cs.c
+++ b/drivers/gpu/drm/radeon/radeon_cs.c
@@ -716,3 +716,24 @@ bool radeon_cs_packet_next_is_pkt3_nop(struct radeon_cs_parser *p)
 		return false;
 	return true;
 }
+
+/**
+ * radeon_cs_dump_packet() - dump raw packet context
+ * @p:		structure holding the parser context.
+ * @pkt:	structure holding the packet.
+ *
+ * Used mostly for debugging and error reporting.
+ **/
+void radeon_cs_dump_packet(struct radeon_cs_parser *p,
+			   struct radeon_cs_packet *pkt)
+{
+	volatile uint32_t *ib;
+	unsigned i;
+	unsigned idx;
+
+	ib = p->ib.ptr;
+	idx = pkt->idx;
+	for (i = 0; i <= (pkt->count + 1); i++, idx++)
+		DRM_INFO("ib[%d]=0x%08X\n", idx, ib[idx]);
+}
+
-- 
cgit v0.10.2


From e97169930941c4326dd563578369590a52aec707 Mon Sep 17 00:00:00 2001
From: Ilija Hadzic <ihadzic@research.bell-labs.com>
Date: Wed, 2 Jan 2013 18:27:46 -0500
Subject: drm/radeon: pull out common next_reloc function
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

next_reloc function does the same thing in all ASICs with
the exception of R600 which has a special case in legacy mode.
Pull out the common function in preparation for refactoring.

Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
Reviewed-by: Marek Olšák <maraeo@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 5f96430..fa17ce0 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -1978,6 +1978,9 @@ int radeon_cs_packet_parse(struct radeon_cs_parser *p,
 bool radeon_cs_packet_next_is_pkt3_nop(struct radeon_cs_parser *p);
 void radeon_cs_dump_packet(struct radeon_cs_parser *p,
 			   struct radeon_cs_packet *pkt);
+int radeon_cs_packet_next_reloc(struct radeon_cs_parser *p,
+				struct radeon_cs_reloc **cs_reloc,
+				int nomm);
 int r600_cs_common_vline_parse(struct radeon_cs_parser *p,
 			       uint32_t *vline_start_end,
 			       uint32_t *vline_status);
diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c
index 7f48d46..1d214b6 100644
--- a/drivers/gpu/drm/radeon/radeon_cs.c
+++ b/drivers/gpu/drm/radeon/radeon_cs.c
@@ -737,3 +737,57 @@ void radeon_cs_dump_packet(struct radeon_cs_parser *p,
 		DRM_INFO("ib[%d]=0x%08X\n", idx, ib[idx]);
 }
 
+/**
+ * radeon_cs_packet_next_reloc() - parse next (should be reloc) packet
+ * @parser:		parser structure holding parsing context.
+ * @data:		pointer to relocation data
+ * @offset_start:	starting offset
+ * @offset_mask:	offset mask (to align start offset on)
+ * @reloc:		reloc informations
+ *
+ * Check if next packet is relocation packet3, do bo validation and compute
+ * GPU offset using the provided start.
+ **/
+int radeon_cs_packet_next_reloc(struct radeon_cs_parser *p,
+				struct radeon_cs_reloc **cs_reloc,
+				int nomm)
+{
+	struct radeon_cs_chunk *relocs_chunk;
+	struct radeon_cs_packet p3reloc;
+	unsigned idx;
+	int r;
+
+	if (p->chunk_relocs_idx == -1) {
+		DRM_ERROR("No relocation chunk !\n");
+		return -EINVAL;
+	}
+	*cs_reloc = NULL;
+	relocs_chunk = &p->chunks[p->chunk_relocs_idx];
+	r = radeon_cs_packet_parse(p, &p3reloc, p->idx);
+	if (r)
+		return r;
+	p->idx += p3reloc.count + 2;
+	if (p3reloc.type != RADEON_PACKET_TYPE3 ||
+	    p3reloc.opcode != RADEON_PACKET3_NOP) {
+		DRM_ERROR("No packet3 for relocation for packet at %d.\n",
+			  p3reloc.idx);
+		radeon_cs_dump_packet(p, &p3reloc);
+		return -EINVAL;
+	}
+	idx = radeon_get_ib_value(p, p3reloc.idx + 1);
+	if (idx >= relocs_chunk->length_dw) {
+		DRM_ERROR("Relocs at %d after relocations chunk end %d !\n",
+			  idx, relocs_chunk->length_dw);
+		radeon_cs_dump_packet(p, &p3reloc);
+		return -EINVAL;
+	}
+	/* FIXME: we assume reloc size is 4 dwords */
+	if (nomm) {
+		*cs_reloc = p->relocs;
+		(*cs_reloc)->lobj.gpu_offset =
+			(u64)relocs_chunk->kdata[idx + 3] << 32;
+		(*cs_reloc)->lobj.gpu_offset |= relocs_chunk->kdata[idx + 0];
+	} else
+		*cs_reloc = p->relocs_ptr[(idx / 4)];
+	return 0;
+}
-- 
cgit v0.10.2


From 012e976d42d2819c79bdd4ef2843515bdd44e408 Mon Sep 17 00:00:00 2001
From: Ilija Hadzic <ihadzic@research.bell-labs.com>
Date: Wed, 2 Jan 2013 18:27:47 -0500
Subject: drm/radeon: use common next_reloc function
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

This patch eliminates ASIC-specific ***_cs_packet_next_reloc
functions and hooks up the new common function.

Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
Reviewed-by: Marek Olšák <maraeo@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/evergreen_cs.c b/drivers/gpu/drm/radeon/evergreen_cs.c
index 02aeb7f..211b509 100644
--- a/drivers/gpu/drm/radeon/evergreen_cs.c
+++ b/drivers/gpu/drm/radeon/evergreen_cs.c
@@ -36,9 +36,6 @@
 
 int r600_dma_cs_next_reloc(struct radeon_cs_parser *p,
 			   struct radeon_cs_reloc **cs_reloc);
-static int evergreen_cs_packet_next_reloc(struct radeon_cs_parser *p,
-					  struct radeon_cs_reloc **cs_reloc);
-
 struct evergreen_cs_track {
 	u32			group_size;
 	u32			nbanks;
@@ -1009,52 +1006,6 @@ static int evergreen_cs_track_check(struct radeon_cs_parser *p)
 }
 
 /**
- * evergreen_cs_packet_next_reloc() - parse next packet which should be reloc packet3
- * @parser:		parser structure holding parsing context.
- * @data:		pointer to relocation data
- * @offset_start:	starting offset
- * @offset_mask:	offset mask (to align start offset on)
- * @reloc:		reloc informations
- *
- * Check next packet is relocation packet3, do bo validation and compute
- * GPU offset using the provided start.
- **/
-static int evergreen_cs_packet_next_reloc(struct radeon_cs_parser *p,
-					  struct radeon_cs_reloc **cs_reloc)
-{
-	struct radeon_cs_chunk *relocs_chunk;
-	struct radeon_cs_packet p3reloc;
-	unsigned idx;
-	int r;
-
-	if (p->chunk_relocs_idx == -1) {
-		DRM_ERROR("No relocation chunk !\n");
-		return -EINVAL;
-	}
-	*cs_reloc = NULL;
-	relocs_chunk = &p->chunks[p->chunk_relocs_idx];
-	r = radeon_cs_packet_parse(p, &p3reloc, p->idx);
-	if (r) {
-		return r;
-	}
-	p->idx += p3reloc.count + 2;
-	if (p3reloc.type != PACKET_TYPE3 || p3reloc.opcode != PACKET3_NOP) {
-		DRM_ERROR("No packet3 for relocation for packet at %d.\n",
-			  p3reloc.idx);
-		return -EINVAL;
-	}
-	idx = radeon_get_ib_value(p, p3reloc.idx + 1);
-	if (idx >= relocs_chunk->length_dw) {
-		DRM_ERROR("Relocs at %d after relocations chunk end %d !\n",
-			  idx, relocs_chunk->length_dw);
-		return -EINVAL;
-	}
-	/* FIXME: we assume reloc size is 4 dwords */
-	*cs_reloc = p->relocs_ptr[(idx / 4)];
-	return 0;
-}
-
-/**
  * evergreen_cs_packet_parse_vline() - parse userspace VLINE packet
  * @parser:		parser structure holding parsing context.
  *
@@ -1205,7 +1156,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case SQ_LSTMP_RING_BASE:
 	case SQ_PSTMP_RING_BASE:
 	case SQ_VSTMP_RING_BASE:
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			dev_warn(p->dev, "bad SET_CONTEXT_REG "
 					"0x%04X\n", reg);
@@ -1234,7 +1185,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case DB_Z_INFO:
 		track->db_z_info = radeon_get_ib_value(p, idx);
 		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
-			r = evergreen_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 			if (r) {
 				dev_warn(p->dev, "bad SET_CONTEXT_REG "
 						"0x%04X\n", reg);
@@ -1276,7 +1227,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		track->db_dirty = true;
 		break;
 	case DB_Z_READ_BASE:
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			dev_warn(p->dev, "bad SET_CONTEXT_REG "
 					"0x%04X\n", reg);
@@ -1288,7 +1239,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		track->db_dirty = true;
 		break;
 	case DB_Z_WRITE_BASE:
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			dev_warn(p->dev, "bad SET_CONTEXT_REG "
 					"0x%04X\n", reg);
@@ -1300,7 +1251,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		track->db_dirty = true;
 		break;
 	case DB_STENCIL_READ_BASE:
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			dev_warn(p->dev, "bad SET_CONTEXT_REG "
 					"0x%04X\n", reg);
@@ -1312,7 +1263,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		track->db_dirty = true;
 		break;
 	case DB_STENCIL_WRITE_BASE:
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			dev_warn(p->dev, "bad SET_CONTEXT_REG "
 					"0x%04X\n", reg);
@@ -1335,7 +1286,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case VGT_STRMOUT_BUFFER_BASE_1:
 	case VGT_STRMOUT_BUFFER_BASE_2:
 	case VGT_STRMOUT_BUFFER_BASE_3:
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			dev_warn(p->dev, "bad SET_CONTEXT_REG "
 					"0x%04X\n", reg);
@@ -1357,7 +1308,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		track->streamout_dirty = true;
 		break;
 	case CP_COHER_BASE:
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			dev_warn(p->dev, "missing reloc for CP_COHER_BASE "
 					"0x%04X\n", reg);
@@ -1421,7 +1372,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		tmp = (reg - CB_COLOR0_INFO) / 0x3c;
 		track->cb_color_info[tmp] = radeon_get_ib_value(p, idx);
 		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
-			r = evergreen_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 			if (r) {
 				dev_warn(p->dev, "bad SET_CONTEXT_REG "
 						"0x%04X\n", reg);
@@ -1439,7 +1390,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		tmp = ((reg - CB_COLOR8_INFO) / 0x1c) + 8;
 		track->cb_color_info[tmp] = radeon_get_ib_value(p, idx);
 		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
-			r = evergreen_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 			if (r) {
 				dev_warn(p->dev, "bad SET_CONTEXT_REG "
 						"0x%04X\n", reg);
@@ -1500,7 +1451,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case CB_COLOR5_ATTRIB:
 	case CB_COLOR6_ATTRIB:
 	case CB_COLOR7_ATTRIB:
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			dev_warn(p->dev, "bad SET_CONTEXT_REG "
 					"0x%04X\n", reg);
@@ -1528,7 +1479,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case CB_COLOR9_ATTRIB:
 	case CB_COLOR10_ATTRIB:
 	case CB_COLOR11_ATTRIB:
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			dev_warn(p->dev, "bad SET_CONTEXT_REG "
 					"0x%04X\n", reg);
@@ -1561,7 +1512,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case CB_COLOR6_FMASK:
 	case CB_COLOR7_FMASK:
 		tmp = (reg - CB_COLOR0_FMASK) / 0x3c;
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			dev_err(p->dev, "bad SET_CONTEXT_REG 0x%04X\n", reg);
 			return -EINVAL;
@@ -1578,7 +1529,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case CB_COLOR6_CMASK:
 	case CB_COLOR7_CMASK:
 		tmp = (reg - CB_COLOR0_CMASK) / 0x3c;
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			dev_err(p->dev, "bad SET_CONTEXT_REG 0x%04X\n", reg);
 			return -EINVAL;
@@ -1616,7 +1567,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case CB_COLOR5_BASE:
 	case CB_COLOR6_BASE:
 	case CB_COLOR7_BASE:
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			dev_warn(p->dev, "bad SET_CONTEXT_REG "
 					"0x%04X\n", reg);
@@ -1632,7 +1583,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case CB_COLOR9_BASE:
 	case CB_COLOR10_BASE:
 	case CB_COLOR11_BASE:
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			dev_warn(p->dev, "bad SET_CONTEXT_REG "
 					"0x%04X\n", reg);
@@ -1645,7 +1596,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		track->cb_dirty = true;
 		break;
 	case DB_HTILE_DATA_BASE:
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			dev_warn(p->dev, "bad SET_CONTEXT_REG "
 					"0x%04X\n", reg);
@@ -1763,7 +1714,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case SQ_ALU_CONST_CACHE_LS_13:
 	case SQ_ALU_CONST_CACHE_LS_14:
 	case SQ_ALU_CONST_CACHE_LS_15:
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			dev_warn(p->dev, "bad SET_CONTEXT_REG "
 					"0x%04X\n", reg);
@@ -1777,7 +1728,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 				 "0x%04X\n", reg);
 			return -EINVAL;
 		}
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			dev_warn(p->dev, "bad SET_CONFIG_REG "
 					"0x%04X\n", reg);
@@ -1791,7 +1742,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 				 "0x%04X\n", reg);
 			return -EINVAL;
 		}
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			dev_warn(p->dev, "bad SET_CONTEXT_REG "
 					"0x%04X\n", reg);
@@ -1876,7 +1827,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			return -EINVAL;
 		}
 
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("bad SET PREDICATION\n");
 			return -EINVAL;
@@ -1922,7 +1873,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			DRM_ERROR("bad INDEX_BASE\n");
 			return -EINVAL;
 		}
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("bad INDEX_BASE\n");
 			return -EINVAL;
@@ -1949,7 +1900,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			DRM_ERROR("bad DRAW_INDEX\n");
 			return -EINVAL;
 		}
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("bad DRAW_INDEX\n");
 			return -EINVAL;
@@ -1977,7 +1928,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			DRM_ERROR("bad DRAW_INDEX_2\n");
 			return -EINVAL;
 		}
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("bad DRAW_INDEX_2\n");
 			return -EINVAL;
@@ -2068,7 +2019,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			DRM_ERROR("bad DISPATCH_INDIRECT\n");
 			return -EINVAL;
 		}
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("bad DISPATCH_INDIRECT\n");
 			return -EINVAL;
@@ -2089,7 +2040,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 		if (idx_value & 0x10) {
 			uint64_t offset;
 
-			r = evergreen_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 			if (r) {
 				DRM_ERROR("bad WAIT_REG_MEM\n");
 				return -EINVAL;
@@ -2143,7 +2094,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			}
 			/* src address space is memory */
 			if (((info & 0x60000000) >> 29) == 0) {
-				r = evergreen_cs_packet_next_reloc(p, &reloc);
+				r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 				if (r) {
 					DRM_ERROR("bad CP DMA SRC\n");
 					return -EINVAL;
@@ -2181,7 +2132,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 				return -EINVAL;
 			}
 			if (((info & 0x00300000) >> 20) == 0) {
-				r = evergreen_cs_packet_next_reloc(p, &reloc);
+				r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 				if (r) {
 					DRM_ERROR("bad CP DMA DST\n");
 					return -EINVAL;
@@ -2215,7 +2166,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 		/* 0xffffffff/0x0 is flush all cache flag */
 		if (radeon_get_ib_value(p, idx + 1) != 0xffffffff ||
 		    radeon_get_ib_value(p, idx + 2) != 0) {
-			r = evergreen_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 			if (r) {
 				DRM_ERROR("bad SURFACE_SYNC\n");
 				return -EINVAL;
@@ -2231,7 +2182,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 		if (pkt->count) {
 			uint64_t offset;
 
-			r = evergreen_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 			if (r) {
 				DRM_ERROR("bad EVENT_WRITE\n");
 				return -EINVAL;
@@ -2252,7 +2203,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			DRM_ERROR("bad EVENT_WRITE_EOP\n");
 			return -EINVAL;
 		}
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("bad EVENT_WRITE_EOP\n");
 			return -EINVAL;
@@ -2274,7 +2225,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			DRM_ERROR("bad EVENT_WRITE_EOS\n");
 			return -EINVAL;
 		}
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("bad EVENT_WRITE_EOS\n");
 			return -EINVAL;
@@ -2341,7 +2292,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			switch (G__SQ_CONSTANT_TYPE(radeon_get_ib_value(p, idx+1+(i*8)+7))) {
 			case SQ_TEX_VTX_VALID_TEXTURE:
 				/* tex base */
-				r = evergreen_cs_packet_next_reloc(p, &reloc);
+				r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 				if (r) {
 					DRM_ERROR("bad SET_RESOURCE (tex)\n");
 					return -EINVAL;
@@ -2378,7 +2329,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 					moffset = 0;
 					mipmap = NULL;
 				} else {
-					r = evergreen_cs_packet_next_reloc(p, &reloc);
+					r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 					if (r) {
 						DRM_ERROR("bad SET_RESOURCE (tex)\n");
 						return -EINVAL;
@@ -2397,7 +2348,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			{
 				uint64_t offset64;
 				/* vtx base */
-				r = evergreen_cs_packet_next_reloc(p, &reloc);
+				r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 				if (r) {
 					DRM_ERROR("bad SET_RESOURCE (vtx)\n");
 					return -EINVAL;
@@ -2479,7 +2430,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 		/* Updating memory at DST_ADDRESS. */
 		if (idx_value & 0x1) {
 			u64 offset;
-			r = evergreen_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 			if (r) {
 				DRM_ERROR("bad STRMOUT_BUFFER_UPDATE (missing dst reloc)\n");
 				return -EINVAL;
@@ -2498,7 +2449,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 		/* Reading data from SRC_ADDRESS. */
 		if (((idx_value >> 1) & 0x3) == 2) {
 			u64 offset;
-			r = evergreen_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 			if (r) {
 				DRM_ERROR("bad STRMOUT_BUFFER_UPDATE (missing src reloc)\n");
 				return -EINVAL;
@@ -2523,7 +2474,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			DRM_ERROR("bad MEM_WRITE (invalid count)\n");
 			return -EINVAL;
 		}
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("bad MEM_WRITE (missing reloc)\n");
 			return -EINVAL;
@@ -2552,7 +2503,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 		if (idx_value & 0x1) {
 			u64 offset;
 			/* SRC is memory. */
-			r = evergreen_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 			if (r) {
 				DRM_ERROR("bad COPY_DW (missing src reloc)\n");
 				return -EINVAL;
@@ -2576,7 +2527,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 		if (idx_value & 0x2) {
 			u64 offset;
 			/* DST is memory. */
-			r = evergreen_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 			if (r) {
 				DRM_ERROR("bad COPY_DW (missing dst reloc)\n");
 				return -EINVAL;
diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c
index cf7e359..8469607 100644
--- a/drivers/gpu/drm/radeon/r100.c
+++ b/drivers/gpu/drm/radeon/r100.c
@@ -1215,7 +1215,7 @@ int r100_reloc_pitch_offset(struct radeon_cs_parser *p,
 	struct radeon_cs_reloc *reloc;
 	u32 value;
 
-	r = r100_cs_packet_next_reloc(p, &reloc);
+	r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 	if (r) {
 		DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 			  idx, reg);
@@ -1268,7 +1268,7 @@ int r100_packet3_load_vbpntr(struct radeon_cs_parser *p,
 	}
 	track->num_arrays = c;
 	for (i = 0; i < (c - 1); i+=2, idx+=3) {
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for packet3 %d\n",
 				  pkt->opcode);
@@ -1281,7 +1281,7 @@ int r100_packet3_load_vbpntr(struct radeon_cs_parser *p,
 		track->arrays[i + 0].esize = idx_value >> 8;
 		track->arrays[i + 0].robj = reloc->robj;
 		track->arrays[i + 0].esize &= 0x7F;
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for packet3 %d\n",
 				  pkt->opcode);
@@ -1294,7 +1294,7 @@ int r100_packet3_load_vbpntr(struct radeon_cs_parser *p,
 		track->arrays[i + 1].esize &= 0x7F;
 	}
 	if (c & 1) {
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for packet3 %d\n",
 					  pkt->opcode);
@@ -1445,54 +1445,6 @@ int r100_cs_packet_parse_vline(struct radeon_cs_parser *p)
 	return 0;
 }
 
-/**
- * r100_cs_packet_next_reloc() - parse next packet which should be reloc packet3
- * @parser:		parser structure holding parsing context.
- * @data:		pointer to relocation data
- * @offset_start:	starting offset
- * @offset_mask:	offset mask (to align start offset on)
- * @reloc:		reloc informations
- *
- * Check next packet is relocation packet3, do bo validation and compute
- * GPU offset using the provided start.
- **/
-int r100_cs_packet_next_reloc(struct radeon_cs_parser *p,
-			      struct radeon_cs_reloc **cs_reloc)
-{
-	struct radeon_cs_chunk *relocs_chunk;
-	struct radeon_cs_packet p3reloc;
-	unsigned idx;
-	int r;
-
-	if (p->chunk_relocs_idx == -1) {
-		DRM_ERROR("No relocation chunk !\n");
-		return -EINVAL;
-	}
-	*cs_reloc = NULL;
-	relocs_chunk = &p->chunks[p->chunk_relocs_idx];
-	r = radeon_cs_packet_parse(p, &p3reloc, p->idx);
-	if (r) {
-		return r;
-	}
-	p->idx += p3reloc.count + 2;
-	if (p3reloc.type != PACKET_TYPE3 || p3reloc.opcode != PACKET3_NOP) {
-		DRM_ERROR("No packet3 for relocation for packet at %d.\n",
-			  p3reloc.idx);
-		radeon_cs_dump_packet(p, &p3reloc);
-		return -EINVAL;
-	}
-	idx = radeon_get_ib_value(p, p3reloc.idx + 1);
-	if (idx >= relocs_chunk->length_dw) {
-		DRM_ERROR("Relocs at %d after relocations chunk end %d !\n",
-			  idx, relocs_chunk->length_dw);
-		radeon_cs_dump_packet(p, &p3reloc);
-		return -EINVAL;
-	}
-	/* FIXME: we assume reloc size is 4 dwords */
-	*cs_reloc = p->relocs_ptr[(idx / 4)];
-	return 0;
-}
-
 static int r100_get_vtx_size(uint32_t vtx_fmt)
 {
 	int vtx_size;
@@ -1583,7 +1535,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 			return r;
 		break;
 	case RADEON_RB3D_DEPTHOFFSET:
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
@@ -1596,7 +1548,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 		ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset);
 		break;
 	case RADEON_RB3D_COLOROFFSET:
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
@@ -1612,7 +1564,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 	case RADEON_PP_TXOFFSET_1:
 	case RADEON_PP_TXOFFSET_2:
 		i = (reg - RADEON_PP_TXOFFSET_0) / 24;
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
@@ -1639,7 +1591,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 	case RADEON_PP_CUBIC_OFFSET_T0_3:
 	case RADEON_PP_CUBIC_OFFSET_T0_4:
 		i = (reg - RADEON_PP_CUBIC_OFFSET_T0_0) / 4;
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
@@ -1657,7 +1609,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 	case RADEON_PP_CUBIC_OFFSET_T1_3:
 	case RADEON_PP_CUBIC_OFFSET_T1_4:
 		i = (reg - RADEON_PP_CUBIC_OFFSET_T1_0) / 4;
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
@@ -1675,7 +1627,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 	case RADEON_PP_CUBIC_OFFSET_T2_3:
 	case RADEON_PP_CUBIC_OFFSET_T2_4:
 		i = (reg - RADEON_PP_CUBIC_OFFSET_T2_0) / 4;
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
@@ -1693,7 +1645,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 		track->zb_dirty = true;
 		break;
 	case RADEON_RB3D_COLORPITCH:
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
@@ -1764,7 +1716,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 		track->zb_dirty = true;
 		break;
 	case RADEON_RB3D_ZPASS_ADDR:
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
@@ -1925,7 +1877,7 @@ static int r100_packet3_check(struct radeon_cs_parser *p,
 			return r;
 		break;
 	case PACKET3_INDX_BUFFER:
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for packet3 %d\n", pkt->opcode);
 			radeon_cs_dump_packet(p, pkt);
@@ -1939,7 +1891,7 @@ static int r100_packet3_check(struct radeon_cs_parser *p,
 		break;
 	case 0x23:
 		/* 3D_RNDR_GEN_INDX_PRIM on r100/r200 */
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for packet3 %d\n", pkt->opcode);
 			radeon_cs_dump_packet(p, pkt);
diff --git a/drivers/gpu/drm/radeon/r100_track.h b/drivers/gpu/drm/radeon/r100_track.h
index c81b174..eb40888 100644
--- a/drivers/gpu/drm/radeon/r100_track.h
+++ b/drivers/gpu/drm/radeon/r100_track.h
@@ -81,8 +81,6 @@ struct r100_cs_track {
 
 int r100_cs_track_check(struct radeon_device *rdev, struct r100_cs_track *track);
 void r100_cs_track_clear(struct radeon_device *rdev, struct r100_cs_track *track);
-int r100_cs_packet_next_reloc(struct radeon_cs_parser *p,
-			      struct radeon_cs_reloc **cs_reloc);
 
 int r100_cs_packet_parse_vline(struct radeon_cs_parser *p);
 
diff --git a/drivers/gpu/drm/radeon/r200.c b/drivers/gpu/drm/radeon/r200.c
index 07ea6f4..b3807ed 100644
--- a/drivers/gpu/drm/radeon/r200.c
+++ b/drivers/gpu/drm/radeon/r200.c
@@ -175,7 +175,7 @@ int r200_packet0_check(struct radeon_cs_parser *p,
 			return r;
 		break;
 	case RADEON_RB3D_DEPTHOFFSET:
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
@@ -188,7 +188,7 @@ int r200_packet0_check(struct radeon_cs_parser *p,
 		ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset);
 		break;
 	case RADEON_RB3D_COLOROFFSET:
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
@@ -207,7 +207,7 @@ int r200_packet0_check(struct radeon_cs_parser *p,
 	case R200_PP_TXOFFSET_4:
 	case R200_PP_TXOFFSET_5:
 		i = (reg - R200_PP_TXOFFSET_0) / 24;
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
@@ -260,7 +260,7 @@ int r200_packet0_check(struct radeon_cs_parser *p,
 	case R200_PP_CUBIC_OFFSET_F5_5:
 		i = (reg - R200_PP_TXOFFSET_0) / 24;
 		face = (reg - ((i * 24) + R200_PP_TXOFFSET_0)) / 4;
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
@@ -278,7 +278,7 @@ int r200_packet0_check(struct radeon_cs_parser *p,
 		track->zb_dirty = true;
 		break;
 	case RADEON_RB3D_COLORPITCH:
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
@@ -355,7 +355,7 @@ int r200_packet0_check(struct radeon_cs_parser *p,
 		track->zb_dirty = true;
 		break;
 	case RADEON_RB3D_ZPASS_ADDR:
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c
index 47bfca1..98f723c7 100644
--- a/drivers/gpu/drm/radeon/r300.c
+++ b/drivers/gpu/drm/radeon/r300.c
@@ -630,7 +630,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 	case R300_RB3D_COLOROFFSET2:
 	case R300_RB3D_COLOROFFSET3:
 		i = (reg - R300_RB3D_COLOROFFSET0) >> 2;
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 					idx, reg);
@@ -643,7 +643,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 		ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset);
 		break;
 	case R300_ZB_DEPTHOFFSET:
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 					idx, reg);
@@ -672,7 +672,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 	case R300_TX_OFFSET_0+56:
 	case R300_TX_OFFSET_0+60:
 		i = (reg - R300_TX_OFFSET_0) >> 2;
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 					idx, reg);
@@ -745,7 +745,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 		/* RB3D_COLORPITCH2 */
 		/* RB3D_COLORPITCH3 */
 		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
-			r = r100_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 			if (r) {
 				DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 					  idx, reg);
@@ -830,7 +830,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 	case 0x4F24:
 		/* ZB_DEPTHPITCH */
 		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
-			r = r100_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 			if (r) {
 				DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 					  idx, reg);
@@ -1045,7 +1045,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 		track->tex_dirty = true;
 		break;
 	case R300_ZB_ZPASS_ADDR:
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 					idx, reg);
@@ -1087,7 +1087,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 		track->cb_dirty = true;
 		break;
 	case R300_RB3D_AARESOLVE_OFFSET:
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
 				  idx, reg);
@@ -1156,7 +1156,7 @@ static int r300_packet3_check(struct radeon_cs_parser *p,
 			return r;
 		break;
 	case PACKET3_INDX_BUFFER:
-		r = r100_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, 0);
 		if (r) {
 			DRM_ERROR("No reloc for packet3 %d\n", pkt->opcode);
 			radeon_cs_dump_packet(p, pkt);
diff --git a/drivers/gpu/drm/radeon/r600_cs.c b/drivers/gpu/drm/radeon/r600_cs.c
index 9d83372..f8fe50c 100644
--- a/drivers/gpu/drm/radeon/r600_cs.c
+++ b/drivers/gpu/drm/radeon/r600_cs.c
@@ -31,12 +31,7 @@
 #include "r600d.h"
 #include "r600_reg_safe.h"
 
-static int r600_cs_packet_next_reloc_mm(struct radeon_cs_parser *p,
-					struct radeon_cs_reloc **cs_reloc);
-static int r600_cs_packet_next_reloc_nomm(struct radeon_cs_parser *p,
-					struct radeon_cs_reloc **cs_reloc);
-typedef int (*next_reloc_t)(struct radeon_cs_parser*, struct radeon_cs_reloc**);
-static next_reloc_t r600_cs_packet_next_reloc = &r600_cs_packet_next_reloc_mm;
+static int r600_nomm;
 extern void r600_cs_legacy_get_tiling_conf(struct drm_device *dev, u32 *npipes, u32 *nbanks, u32 *group_size);
 
 
@@ -784,99 +779,6 @@ static int r600_cs_track_check(struct radeon_cs_parser *p)
 }
 
 /**
- * r600_cs_packet_next_reloc_mm() - parse next packet which should be reloc packet3
- * @parser:		parser structure holding parsing context.
- * @data:		pointer to relocation data
- * @offset_start:	starting offset
- * @offset_mask:	offset mask (to align start offset on)
- * @reloc:		reloc informations
- *
- * Check next packet is relocation packet3, do bo validation and compute
- * GPU offset using the provided start.
- **/
-static int r600_cs_packet_next_reloc_mm(struct radeon_cs_parser *p,
-					struct radeon_cs_reloc **cs_reloc)
-{
-	struct radeon_cs_chunk *relocs_chunk;
-	struct radeon_cs_packet p3reloc;
-	unsigned idx;
-	int r;
-
-	if (p->chunk_relocs_idx == -1) {
-		DRM_ERROR("No relocation chunk !\n");
-		return -EINVAL;
-	}
-	*cs_reloc = NULL;
-	relocs_chunk = &p->chunks[p->chunk_relocs_idx];
-	r = radeon_cs_packet_parse(p, &p3reloc, p->idx);
-	if (r) {
-		return r;
-	}
-	p->idx += p3reloc.count + 2;
-	if (p3reloc.type != PACKET_TYPE3 || p3reloc.opcode != PACKET3_NOP) {
-		DRM_ERROR("No packet3 for relocation for packet at %d.\n",
-			  p3reloc.idx);
-		return -EINVAL;
-	}
-	idx = radeon_get_ib_value(p, p3reloc.idx + 1);
-	if (idx >= relocs_chunk->length_dw) {
-		DRM_ERROR("Relocs at %d after relocations chunk end %d !\n",
-			  idx, relocs_chunk->length_dw);
-		return -EINVAL;
-	}
-	/* FIXME: we assume reloc size is 4 dwords */
-	*cs_reloc = p->relocs_ptr[(idx / 4)];
-	return 0;
-}
-
-/**
- * r600_cs_packet_next_reloc_nomm() - parse next packet which should be reloc packet3
- * @parser:		parser structure holding parsing context.
- * @data:		pointer to relocation data
- * @offset_start:	starting offset
- * @offset_mask:	offset mask (to align start offset on)
- * @reloc:		reloc informations
- *
- * Check next packet is relocation packet3, do bo validation and compute
- * GPU offset using the provided start.
- **/
-static int r600_cs_packet_next_reloc_nomm(struct radeon_cs_parser *p,
-					struct radeon_cs_reloc **cs_reloc)
-{
-	struct radeon_cs_chunk *relocs_chunk;
-	struct radeon_cs_packet p3reloc;
-	unsigned idx;
-	int r;
-
-	if (p->chunk_relocs_idx == -1) {
-		DRM_ERROR("No relocation chunk !\n");
-		return -EINVAL;
-	}
-	*cs_reloc = NULL;
-	relocs_chunk = &p->chunks[p->chunk_relocs_idx];
-	r = radeon_cs_packet_parse(p, &p3reloc, p->idx);
-	if (r) {
-		return r;
-	}
-	p->idx += p3reloc.count + 2;
-	if (p3reloc.type != PACKET_TYPE3 || p3reloc.opcode != PACKET3_NOP) {
-		DRM_ERROR("No packet3 for relocation for packet at %d.\n",
-			  p3reloc.idx);
-		return -EINVAL;
-	}
-	idx = radeon_get_ib_value(p, p3reloc.idx + 1);
-	if (idx >= relocs_chunk->length_dw) {
-		DRM_ERROR("Relocs at %d after relocations chunk end %d !\n",
-			  idx, relocs_chunk->length_dw);
-		return -EINVAL;
-	}
-	*cs_reloc = p->relocs;
-	(*cs_reloc)->lobj.gpu_offset = (u64)relocs_chunk->kdata[idx + 3] << 32;
-	(*cs_reloc)->lobj.gpu_offset |= relocs_chunk->kdata[idx + 0];
-	return 0;
-}
-
-/**
  * r600_cs_packet_parse_vline() - parse userspace VLINE packet
  * @parser:		parser structure holding parsing context.
  *
@@ -1115,7 +1017,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case R_028010_DB_DEPTH_INFO:
 		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS) &&
 		    radeon_cs_packet_next_is_pkt3_nop(p)) {
-			r = r600_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 			if (r) {
 				dev_warn(p->dev, "bad SET_CONTEXT_REG "
 					 "0x%04X\n", reg);
@@ -1157,7 +1059,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case VGT_STRMOUT_BUFFER_BASE_1:
 	case VGT_STRMOUT_BUFFER_BASE_2:
 	case VGT_STRMOUT_BUFFER_BASE_3:
-		r = r600_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 		if (r) {
 			dev_warn(p->dev, "bad SET_CONTEXT_REG "
 					"0x%04X\n", reg);
@@ -1180,7 +1082,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		track->streamout_dirty = true;
 		break;
 	case CP_COHER_BASE:
-		r = r600_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 		if (r) {
 			dev_warn(p->dev, "missing reloc for CP_COHER_BASE "
 					"0x%04X\n", reg);
@@ -1216,7 +1118,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case R_0280BC_CB_COLOR7_INFO:
 		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS) &&
 		     radeon_cs_packet_next_is_pkt3_nop(p)) {
-			r = r600_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 			if (r) {
 				dev_err(p->dev, "bad SET_CONTEXT_REG 0x%04X\n", reg);
 				return -EINVAL;
@@ -1288,7 +1190,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 			track->cb_color_frag_offset[tmp] = track->cb_color_bo_offset[tmp];
 			ib[idx] = track->cb_color_base_last[tmp];
 		} else {
-			r = r600_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 			if (r) {
 				dev_err(p->dev, "bad SET_CONTEXT_REG 0x%04X\n", reg);
 				return -EINVAL;
@@ -1319,7 +1221,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 			track->cb_color_tile_offset[tmp] = track->cb_color_bo_offset[tmp];
 			ib[idx] = track->cb_color_base_last[tmp];
 		} else {
-			r = r600_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 			if (r) {
 				dev_err(p->dev, "bad SET_CONTEXT_REG 0x%04X\n", reg);
 				return -EINVAL;
@@ -1354,7 +1256,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case CB_COLOR5_BASE:
 	case CB_COLOR6_BASE:
 	case CB_COLOR7_BASE:
-		r = r600_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 		if (r) {
 			dev_warn(p->dev, "bad SET_CONTEXT_REG "
 					"0x%04X\n", reg);
@@ -1369,7 +1271,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		track->cb_dirty = true;
 		break;
 	case DB_DEPTH_BASE:
-		r = r600_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 		if (r) {
 			dev_warn(p->dev, "bad SET_CONTEXT_REG "
 					"0x%04X\n", reg);
@@ -1382,7 +1284,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		track->db_dirty = true;
 		break;
 	case DB_HTILE_DATA_BASE:
-		r = r600_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 		if (r) {
 			dev_warn(p->dev, "bad SET_CONTEXT_REG "
 					"0x%04X\n", reg);
@@ -1452,7 +1354,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case SQ_ALU_CONST_CACHE_VS_13:
 	case SQ_ALU_CONST_CACHE_VS_14:
 	case SQ_ALU_CONST_CACHE_VS_15:
-		r = r600_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 		if (r) {
 			dev_warn(p->dev, "bad SET_CONTEXT_REG "
 					"0x%04X\n", reg);
@@ -1461,7 +1363,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
 		break;
 	case SX_MEMORY_EXPORT_BASE:
-		r = r600_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 		if (r) {
 			dev_warn(p->dev, "bad SET_CONFIG_REG "
 					"0x%04X\n", reg);
@@ -1747,7 +1649,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 			return -EINVAL;
 		}
 
-		r = r600_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 		if (r) {
 			DRM_ERROR("bad SET PREDICATION\n");
 			return -EINVAL;
@@ -1788,7 +1690,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 			DRM_ERROR("bad DRAW_INDEX\n");
 			return -EINVAL;
 		}
-		r = r600_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 		if (r) {
 			DRM_ERROR("bad DRAW_INDEX\n");
 			return -EINVAL;
@@ -1840,7 +1742,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 		if (idx_value & 0x10) {
 			uint64_t offset;
 
-			r = r600_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 			if (r) {
 				DRM_ERROR("bad WAIT_REG_MEM\n");
 				return -EINVAL;
@@ -1877,7 +1779,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 				return -EINVAL;
 			}
 			/* src address space is memory */
-			r = r600_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 			if (r) {
 				DRM_ERROR("bad CP DMA SRC\n");
 				return -EINVAL;
@@ -1907,7 +1809,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 				DRM_ERROR("CP DMA DAIC only supported for registers\n");
 				return -EINVAL;
 			}
-			r = r600_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 			if (r) {
 				DRM_ERROR("bad CP DMA DST\n");
 				return -EINVAL;
@@ -1937,7 +1839,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 		/* 0xffffffff/0x0 is flush all cache flag */
 		if (radeon_get_ib_value(p, idx + 1) != 0xffffffff ||
 		    radeon_get_ib_value(p, idx + 2) != 0) {
-			r = r600_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 			if (r) {
 				DRM_ERROR("bad SURFACE_SYNC\n");
 				return -EINVAL;
@@ -1953,7 +1855,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 		if (pkt->count) {
 			uint64_t offset;
 
-			r = r600_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 			if (r) {
 				DRM_ERROR("bad EVENT_WRITE\n");
 				return -EINVAL;
@@ -1974,7 +1876,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 			DRM_ERROR("bad EVENT_WRITE_EOP\n");
 			return -EINVAL;
 		}
-		r = r600_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 		if (r) {
 			DRM_ERROR("bad EVENT_WRITE\n");
 			return -EINVAL;
@@ -2040,7 +1942,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 			switch (G__SQ_VTX_CONSTANT_TYPE(radeon_get_ib_value(p, idx+(i*7)+6+1))) {
 			case SQ_TEX_VTX_VALID_TEXTURE:
 				/* tex base */
-				r = r600_cs_packet_next_reloc(p, &reloc);
+				r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 				if (r) {
 					DRM_ERROR("bad SET_RESOURCE\n");
 					return -EINVAL;
@@ -2054,7 +1956,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 				}
 				texture = reloc->robj;
 				/* tex mip base */
-				r = r600_cs_packet_next_reloc(p, &reloc);
+				r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 				if (r) {
 					DRM_ERROR("bad SET_RESOURCE\n");
 					return -EINVAL;
@@ -2075,7 +1977,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 			{
 				uint64_t offset64;
 				/* vtx base */
-				r = r600_cs_packet_next_reloc(p, &reloc);
+				r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 				if (r) {
 					DRM_ERROR("bad SET_RESOURCE\n");
 					return -EINVAL;
@@ -2176,7 +2078,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 		{
 			u64 offset;
 
-			r = r600_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 			if (r) {
 				DRM_ERROR("bad STRMOUT_BASE_UPDATE reloc\n");
 				return -EINVAL;
@@ -2220,7 +2122,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 		/* Updating memory at DST_ADDRESS. */
 		if (idx_value & 0x1) {
 			u64 offset;
-			r = r600_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 			if (r) {
 				DRM_ERROR("bad STRMOUT_BUFFER_UPDATE (missing dst reloc)\n");
 				return -EINVAL;
@@ -2239,7 +2141,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 		/* Reading data from SRC_ADDRESS. */
 		if (((idx_value >> 1) & 0x3) == 2) {
 			u64 offset;
-			r = r600_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 			if (r) {
 				DRM_ERROR("bad STRMOUT_BUFFER_UPDATE (missing src reloc)\n");
 				return -EINVAL;
@@ -2264,7 +2166,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 			DRM_ERROR("bad MEM_WRITE (invalid count)\n");
 			return -EINVAL;
 		}
-		r = r600_cs_packet_next_reloc(p, &reloc);
+		r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 		if (r) {
 			DRM_ERROR("bad MEM_WRITE (missing reloc)\n");
 			return -EINVAL;
@@ -2293,7 +2195,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 		if (idx_value & 0x1) {
 			u64 offset;
 			/* SRC is memory. */
-			r = r600_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 			if (r) {
 				DRM_ERROR("bad COPY_DW (missing src reloc)\n");
 				return -EINVAL;
@@ -2317,7 +2219,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 		if (idx_value & 0x2) {
 			u64 offset;
 			/* DST is memory. */
-			r = r600_cs_packet_next_reloc(p, &reloc);
+			r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm);
 			if (r) {
 				DRM_ERROR("bad COPY_DW (missing dst reloc)\n");
 				return -EINVAL;
@@ -2505,7 +2407,7 @@ int r600_cs_legacy(struct drm_device *dev, void *data, struct drm_file *filp,
 
 void r600_cs_legacy_init(void)
 {
-	r600_cs_packet_next_reloc = &r600_cs_packet_next_reloc_nomm;
+	r600_nomm = 1;
 }
 
 /*
-- 
cgit v0.10.2


From 4e872ae2bbb9de838212680c52dcbb85be1e4419 Mon Sep 17 00:00:00 2001
From: Ilija Hadzic <ihadzic@research.bell-labs.com>
Date: Wed, 2 Jan 2013 18:27:48 -0500
Subject: drm/radeon: consolidate redundant macros and constants
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

After refactoring the _cs logic, we ended up with many
macros and constants that #define the same thing.
Clean'em up.

Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
Reviewed-by: Marek Olšák <maraeo@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/evergreen_cs.c b/drivers/gpu/drm/radeon/evergreen_cs.c
index 211b509..4a9760a 100644
--- a/drivers/gpu/drm/radeon/evergreen_cs.c
+++ b/drivers/gpu/drm/radeon/evergreen_cs.c
@@ -2639,12 +2639,12 @@ int evergreen_cs_parse(struct radeon_cs_parser *p)
 		}
 		p->idx += pkt.count + 2;
 		switch (pkt.type) {
-		case PACKET_TYPE0:
+		case RADEON_PACKET_TYPE0:
 			r = evergreen_cs_parse_packet0(p, &pkt);
 			break;
-		case PACKET_TYPE2:
+		case RADEON_PACKET_TYPE2:
 			break;
-		case PACKET_TYPE3:
+		case RADEON_PACKET_TYPE3:
 			r = evergreen_packet3_check(p, &pkt);
 			break;
 		default:
@@ -3395,19 +3395,19 @@ int evergreen_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib)
 
 	do {
 		pkt.idx = idx;
-		pkt.type = CP_PACKET_GET_TYPE(ib->ptr[idx]);
-		pkt.count = CP_PACKET_GET_COUNT(ib->ptr[idx]);
+		pkt.type = RADEON_CP_PACKET_GET_TYPE(ib->ptr[idx]);
+		pkt.count = RADEON_CP_PACKET_GET_COUNT(ib->ptr[idx]);
 		pkt.one_reg_wr = 0;
 		switch (pkt.type) {
-		case PACKET_TYPE0:
+		case RADEON_PACKET_TYPE0:
 			dev_err(rdev->dev, "Packet0 not allowed!\n");
 			ret = -EINVAL;
 			break;
-		case PACKET_TYPE2:
+		case RADEON_PACKET_TYPE2:
 			idx += 1;
 			break;
-		case PACKET_TYPE3:
-			pkt.opcode = CP_PACKET3_GET_OPCODE(ib->ptr[idx]);
+		case RADEON_PACKET_TYPE3:
+			pkt.opcode = RADEON_CP_PACKET3_GET_OPCODE(ib->ptr[idx]);
 			ret = evergreen_vm_packet3_check(rdev, ib->ptr, &pkt);
 			idx += pkt.count + 2;
 			break;
diff --git a/drivers/gpu/drm/radeon/evergreend.h b/drivers/gpu/drm/radeon/evergreend.h
index 0bfd0e9e..70388fe 100644
--- a/drivers/gpu/drm/radeon/evergreend.h
+++ b/drivers/gpu/drm/radeon/evergreend.h
@@ -980,16 +980,7 @@
 /*
  * PM4
  */
-#define	PACKET_TYPE0	0
-#define	PACKET_TYPE1	1
-#define	PACKET_TYPE2	2
-#define	PACKET_TYPE3	3
-
-#define CP_PACKET_GET_TYPE(h) (((h) >> 30) & 3)
-#define CP_PACKET_GET_COUNT(h) (((h) >> 16) & 0x3FFF)
-#define CP_PACKET0_GET_REG(h) (((h) & 0xFFFF) << 2)
-#define CP_PACKET3_GET_OPCODE(h) (((h) >> 8) & 0xFF)
-#define PACKET0(reg, n)	((PACKET_TYPE0 << 30) |				\
+#define PACKET0(reg, n)	((RADEON_PACKET_TYPE0 << 30) |			\
 			 (((reg) >> 2) & 0xFFFF) |			\
 			 ((n) & 0x3FFF) << 16)
 #define CP_PACKET2			0x80000000
@@ -998,7 +989,7 @@
 
 #define PACKET2(v)	(CP_PACKET2 | REG_SET(PACKET2_PAD, (v)))
 
-#define PACKET3(op, n)	((PACKET_TYPE3 << 30) |				\
+#define PACKET3(op, n)	((RADEON_PACKET_TYPE3 << 30) |			\
 			 (((op) & 0xFF) << 8) |				\
 			 ((n) & 0x3FFF) << 16)
 
diff --git a/drivers/gpu/drm/radeon/nid.h b/drivers/gpu/drm/radeon/nid.h
index 48e5022..c4b84c5 100644
--- a/drivers/gpu/drm/radeon/nid.h
+++ b/drivers/gpu/drm/radeon/nid.h
@@ -474,16 +474,7 @@
 /*
  * PM4
  */
-#define	PACKET_TYPE0	0
-#define	PACKET_TYPE1	1
-#define	PACKET_TYPE2	2
-#define	PACKET_TYPE3	3
-
-#define CP_PACKET_GET_TYPE(h) (((h) >> 30) & 3)
-#define CP_PACKET_GET_COUNT(h) (((h) >> 16) & 0x3FFF)
-#define CP_PACKET0_GET_REG(h) (((h) & 0xFFFF) << 2)
-#define CP_PACKET3_GET_OPCODE(h) (((h) >> 8) & 0xFF)
-#define PACKET0(reg, n)	((PACKET_TYPE0 << 30) |				\
+#define PACKET0(reg, n)	((RADEON_PACKET_TYPE0 << 30) |			\
 			 (((reg) >> 2) & 0xFFFF) |			\
 			 ((n) & 0x3FFF) << 16)
 #define CP_PACKET2			0x80000000
@@ -492,7 +483,7 @@
 
 #define PACKET2(v)	(CP_PACKET2 | REG_SET(PACKET2_PAD, (v)))
 
-#define PACKET3(op, n)	((PACKET_TYPE3 << 30) |				\
+#define PACKET3(op, n)	((RADEON_PACKET_TYPE3 << 30) |			\
 			 (((op) & 0xFF) << 8) |				\
 			 ((n) & 0x3FFF) << 16)
 
diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c
index 8469607..9db5853 100644
--- a/drivers/gpu/drm/radeon/r100.c
+++ b/drivers/gpu/drm/radeon/r100.c
@@ -1410,7 +1410,7 @@ int r100_cs_packet_parse_vline(struct radeon_cs_parser *p)
 
 	header = radeon_get_ib_value(p, h_idx);
 	crtc_id = radeon_get_ib_value(p, h_idx + 5);
-	reg = CP_PACKET0_GET_REG(header);
+	reg = R100_CP_PACKET0_GET_REG(header);
 	obj = drm_mode_object_find(p->rdev->ddev, crtc_id, DRM_MODE_OBJECT_CRTC);
 	if (!obj) {
 		DRM_ERROR("cannot find crtc %d\n", crtc_id);
@@ -1997,7 +1997,7 @@ int r100_cs_parse(struct radeon_cs_parser *p)
 		}
 		p->idx += pkt.count + 2;
 		switch (pkt.type) {
-		case PACKET_TYPE0:
+		case RADEON_PACKET_TYPE0:
 			if (p->rdev->family >= CHIP_R200)
 				r = r100_cs_parse_packet0(p, &pkt,
 					p->rdev->config.r100.reg_safe_bm,
@@ -2009,9 +2009,9 @@ int r100_cs_parse(struct radeon_cs_parser *p)
 					p->rdev->config.r100.reg_safe_bm_size,
 					&r100_packet0_check);
 			break;
-		case PACKET_TYPE2:
+		case RADEON_PACKET_TYPE2:
 			break;
-		case PACKET_TYPE3:
+		case RADEON_PACKET_TYPE3:
 			r = r100_packet3_check(p, &pkt);
 			break;
 		default:
diff --git a/drivers/gpu/drm/radeon/r100d.h b/drivers/gpu/drm/radeon/r100d.h
index eab9176..f0f8ee6 100644
--- a/drivers/gpu/drm/radeon/r100d.h
+++ b/drivers/gpu/drm/radeon/r100d.h
@@ -64,17 +64,6 @@
 			 REG_SET(PACKET3_IT_OPCODE, (op)) |		\
 			 REG_SET(PACKET3_COUNT, (n)))
 
-#define	PACKET_TYPE0	0
-#define	PACKET_TYPE1	1
-#define	PACKET_TYPE2	2
-#define	PACKET_TYPE3	3
-
-#define CP_PACKET_GET_TYPE(h) (((h) >> 30) & 3)
-#define CP_PACKET_GET_COUNT(h) (((h) >> 16) & 0x3FFF)
-#define CP_PACKET0_GET_REG(h) (((h) & 0x1FFF) << 2)
-#define CP_PACKET0_GET_ONE_REG_WR(h) (((h) >> 15) & 1)
-#define CP_PACKET3_GET_OPCODE(h) (((h) >> 8) & 0xFF)
-
 /* Registers */
 #define R_0000F0_RBBM_SOFT_RESET                     0x0000F0
 #define   S_0000F0_SOFT_RESET_CP(x)                    (((x) & 0x1) << 0)
diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c
index 98f723c7..c60350e 100644
--- a/drivers/gpu/drm/radeon/r300.c
+++ b/drivers/gpu/drm/radeon/r300.c
@@ -1263,15 +1263,15 @@ int r300_cs_parse(struct radeon_cs_parser *p)
 		}
 		p->idx += pkt.count + 2;
 		switch (pkt.type) {
-		case PACKET_TYPE0:
+		case RADEON_PACKET_TYPE0:
 			r = r100_cs_parse_packet0(p, &pkt,
 						  p->rdev->config.r300.reg_safe_bm,
 						  p->rdev->config.r300.reg_safe_bm_size,
 						  &r300_packet0_check);
 			break;
-		case PACKET_TYPE2:
+		case RADEON_PACKET_TYPE2:
 			break;
-		case PACKET_TYPE3:
+		case RADEON_PACKET_TYPE3:
 			r = r300_packet3_check(p, &pkt);
 			break;
 		default:
diff --git a/drivers/gpu/drm/radeon/r300d.h b/drivers/gpu/drm/radeon/r300d.h
index 1f519a5..ff229a0 100644
--- a/drivers/gpu/drm/radeon/r300d.h
+++ b/drivers/gpu/drm/radeon/r300d.h
@@ -65,17 +65,6 @@
 			 REG_SET(PACKET3_IT_OPCODE, (op)) |		\
 			 REG_SET(PACKET3_COUNT, (n)))
 
-#define	PACKET_TYPE0	0
-#define	PACKET_TYPE1	1
-#define	PACKET_TYPE2	2
-#define	PACKET_TYPE3	3
-
-#define CP_PACKET_GET_TYPE(h) (((h) >> 30) & 3)
-#define CP_PACKET_GET_COUNT(h) (((h) >> 16) & 0x3FFF)
-#define CP_PACKET0_GET_REG(h) (((h) & 0x1FFF) << 2)
-#define CP_PACKET0_GET_ONE_REG_WR(h) (((h) >> 15) & 1)
-#define CP_PACKET3_GET_OPCODE(h) (((h) >> 8) & 0xFF)
-
 /* Registers */
 #define R_000148_MC_FB_LOCATION                      0x000148
 #define   S_000148_MC_FB_START(x)                      (((x) & 0xFFFF) << 0)
diff --git a/drivers/gpu/drm/radeon/r600_cs.c b/drivers/gpu/drm/radeon/r600_cs.c
index f8fe50c..0d21016 100644
--- a/drivers/gpu/drm/radeon/r600_cs.c
+++ b/drivers/gpu/drm/radeon/r600_cs.c
@@ -839,7 +839,7 @@ int r600_cs_common_vline_parse(struct radeon_cs_parser *p,
 		return r;
 
 	/* check its a WAIT_REG_MEM */
-	if (wait_reg_mem.type != PACKET_TYPE3 ||
+	if (wait_reg_mem.type != RADEON_PACKET_TYPE3 ||
 	    wait_reg_mem.opcode != PACKET3_WAIT_REG_MEM) {
 		DRM_ERROR("vline wait missing WAIT_REG_MEM segment\n");
 		return -EINVAL;
@@ -882,7 +882,7 @@ int r600_cs_common_vline_parse(struct radeon_cs_parser *p,
 
 	header = radeon_get_ib_value(p, h_idx);
 	crtc_id = radeon_get_ib_value(p, h_idx + 2 + 7 + 1);
-	reg = CP_PACKET0_GET_REG(header);
+	reg = R600_CP_PACKET0_GET_REG(header);
 
 	obj = drm_mode_object_find(p->rdev->ddev, crtc_id, DRM_MODE_OBJECT_CRTC);
 	if (!obj) {
@@ -2282,12 +2282,12 @@ int r600_cs_parse(struct radeon_cs_parser *p)
 		}
 		p->idx += pkt.count + 2;
 		switch (pkt.type) {
-		case PACKET_TYPE0:
+		case RADEON_PACKET_TYPE0:
 			r = r600_cs_parse_packet0(p, &pkt);
 			break;
-		case PACKET_TYPE2:
+		case RADEON_PACKET_TYPE2:
 			break;
-		case PACKET_TYPE3:
+		case RADEON_PACKET_TYPE3:
 			r = r600_packet3_check(p, &pkt);
 			break;
 		default:
diff --git a/drivers/gpu/drm/radeon/r600d.h b/drivers/gpu/drm/radeon/r600d.h
index 4a53402..62423b0 100644
--- a/drivers/gpu/drm/radeon/r600d.h
+++ b/drivers/gpu/drm/radeon/r600d.h
@@ -1143,19 +1143,10 @@
 /*
  * PM4
  */
-#define	PACKET_TYPE0	0
-#define	PACKET_TYPE1	1
-#define	PACKET_TYPE2	2
-#define	PACKET_TYPE3	3
-
-#define CP_PACKET_GET_TYPE(h) (((h) >> 30) & 3)
-#define CP_PACKET_GET_COUNT(h) (((h) >> 16) & 0x3FFF)
-#define CP_PACKET0_GET_REG(h) (((h) & 0xFFFF) << 2)
-#define CP_PACKET3_GET_OPCODE(h) (((h) >> 8) & 0xFF)
-#define PACKET0(reg, n)	((PACKET_TYPE0 << 30) |				\
+#define PACKET0(reg, n)	((RADEON_PACKET_TYPE0 << 30) |			\
 			 (((reg) >> 2) & 0xFFFF) |			\
 			 ((n) & 0x3FFF) << 16)
-#define PACKET3(op, n)	((PACKET_TYPE3 << 30) |				\
+#define PACKET3(op, n)	((RADEON_PACKET_TYPE3 << 30) |			\
 			 (((op) & 0xFF) << 8) |				\
 			 ((n) & 0x3FFF) << 16)
 
diff --git a/drivers/gpu/drm/radeon/rv515d.h b/drivers/gpu/drm/radeon/rv515d.h
index 590309a..6927a20 100644
--- a/drivers/gpu/drm/radeon/rv515d.h
+++ b/drivers/gpu/drm/radeon/rv515d.h
@@ -205,17 +205,6 @@
 			 REG_SET(PACKET3_IT_OPCODE, (op)) |		\
 			 REG_SET(PACKET3_COUNT, (n)))
 
-#define	PACKET_TYPE0	0
-#define	PACKET_TYPE1	1
-#define	PACKET_TYPE2	2
-#define	PACKET_TYPE3	3
-
-#define CP_PACKET_GET_TYPE(h) (((h) >> 30) & 3)
-#define CP_PACKET_GET_COUNT(h) (((h) >> 16) & 0x3FFF)
-#define CP_PACKET0_GET_REG(h) (((h) & 0x1FFF) << 2)
-#define CP_PACKET0_GET_ONE_REG_WR(h) (((h) >> 15) & 1)
-#define CP_PACKET3_GET_OPCODE(h) (((h) >> 8) & 0xFF)
-
 /* Registers */
 #define R_0000F0_RBBM_SOFT_RESET                     0x0000F0
 #define   S_0000F0_SOFT_RESET_CP(x)                    (((x) & 0x1) << 0)
diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
index ae8b482..f06072f 100644
--- a/drivers/gpu/drm/radeon/si.c
+++ b/drivers/gpu/drm/radeon/si.c
@@ -2855,19 +2855,19 @@ int si_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib)
 
 	do {
 		pkt.idx = idx;
-		pkt.type = CP_PACKET_GET_TYPE(ib->ptr[idx]);
-		pkt.count = CP_PACKET_GET_COUNT(ib->ptr[idx]);
+		pkt.type = RADEON_CP_PACKET_GET_TYPE(ib->ptr[idx]);
+		pkt.count = RADEON_CP_PACKET_GET_COUNT(ib->ptr[idx]);
 		pkt.one_reg_wr = 0;
 		switch (pkt.type) {
-		case PACKET_TYPE0:
+		case RADEON_PACKET_TYPE0:
 			dev_err(rdev->dev, "Packet0 not allowed!\n");
 			ret = -EINVAL;
 			break;
-		case PACKET_TYPE2:
+		case RADEON_PACKET_TYPE2:
 			idx += 1;
 			break;
-		case PACKET_TYPE3:
-			pkt.opcode = CP_PACKET3_GET_OPCODE(ib->ptr[idx]);
+		case RADEON_PACKET_TYPE3:
+			pkt.opcode = RADEON_CP_PACKET3_GET_OPCODE(ib->ptr[idx]);
 			if (ib->is_const_ib)
 				ret = si_vm_packet3_ce_check(rdev, ib->ptr, &pkt);
 			else {
diff --git a/drivers/gpu/drm/radeon/sid.h b/drivers/gpu/drm/radeon/sid.h
index c056aae..76b5911 100644
--- a/drivers/gpu/drm/radeon/sid.h
+++ b/drivers/gpu/drm/radeon/sid.h
@@ -783,16 +783,7 @@
 /*
  * PM4
  */
-#define	PACKET_TYPE0	0
-#define	PACKET_TYPE1	1
-#define	PACKET_TYPE2	2
-#define	PACKET_TYPE3	3
-
-#define CP_PACKET_GET_TYPE(h) (((h) >> 30) & 3)
-#define CP_PACKET_GET_COUNT(h) (((h) >> 16) & 0x3FFF)
-#define CP_PACKET0_GET_REG(h) (((h) & 0xFFFF) << 2)
-#define CP_PACKET3_GET_OPCODE(h) (((h) >> 8) & 0xFF)
-#define PACKET0(reg, n)	((PACKET_TYPE0 << 30) |				\
+#define PACKET0(reg, n)	((RADEON_PACKET_TYPE0 << 30) |			\
 			 (((reg) >> 2) & 0xFFFF) |			\
 			 ((n) & 0x3FFF) << 16)
 #define CP_PACKET2			0x80000000
@@ -801,7 +792,7 @@
 
 #define PACKET2(v)	(CP_PACKET2 | REG_SET(PACKET2_PAD, (v)))
 
-#define PACKET3(op, n)	((PACKET_TYPE3 << 30) |				\
+#define PACKET3(op, n)	((RADEON_PACKET_TYPE3 << 30) |			\
 			 (((op) & 0xFF) << 8) |				\
 			 ((n) & 0x3FFF) << 16)
 
-- 
cgit v0.10.2


From 0fcb6155cb5c146b57a8c50e35afb9b7cfbde394 Mon Sep 17 00:00:00 2001
From: Jerome Glisse <jglisse@redhat.com>
Date: Mon, 14 Jan 2013 11:32:27 -0500
Subject: radeon/kms: cleanup async dma packet checking

This simplify and cleanup the async dma checking.

Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c
index a2d478e..a1e12bf2 100644
--- a/drivers/gpu/drm/radeon/evergreen.c
+++ b/drivers/gpu/drm/radeon/evergreen.c
@@ -3280,14 +3280,14 @@ void evergreen_dma_fence_ring_emit(struct radeon_device *rdev,
 	struct radeon_ring *ring = &rdev->ring[fence->ring];
 	u64 addr = rdev->fence_drv[fence->ring].gpu_addr;
 	/* write the fence */
-	radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_FENCE, 0, 0, 0));
+	radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_FENCE, 0, 0));
 	radeon_ring_write(ring, addr & 0xfffffffc);
 	radeon_ring_write(ring, (upper_32_bits(addr) & 0xff));
 	radeon_ring_write(ring, fence->seq);
 	/* generate an interrupt */
-	radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_TRAP, 0, 0, 0));
+	radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_TRAP, 0, 0));
 	/* flush HDP */
-	radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0));
+	radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0));
 	radeon_ring_write(ring, (0xf << 16) | (HDP_MEM_COHERENCY_FLUSH_CNTL >> 2));
 	radeon_ring_write(ring, 1);
 }
@@ -3310,7 +3310,7 @@ void evergreen_dma_ring_ib_execute(struct radeon_device *rdev,
 		while ((next_rptr & 7) != 5)
 			next_rptr++;
 		next_rptr += 3;
-		radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 1));
+		radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_WRITE, 0, 1));
 		radeon_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc);
 		radeon_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xff);
 		radeon_ring_write(ring, next_rptr);
@@ -3320,8 +3320,8 @@ void evergreen_dma_ring_ib_execute(struct radeon_device *rdev,
 	 * Pad as necessary with NOPs.
 	 */
 	while ((ring->wptr & 7) != 5)
-		radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0));
-	radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_INDIRECT_BUFFER, 0, 0, 0));
+		radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_NOP, 0, 0));
+	radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_INDIRECT_BUFFER, 0, 0));
 	radeon_ring_write(ring, (ib->gpu_addr & 0xFFFFFFE0));
 	radeon_ring_write(ring, (ib->length_dw << 12) | (upper_32_bits(ib->gpu_addr) & 0xFF));
 
@@ -3380,7 +3380,7 @@ int evergreen_copy_dma(struct radeon_device *rdev,
 		if (cur_size_in_dw > 0xFFFFF)
 			cur_size_in_dw = 0xFFFFF;
 		size_in_dw -= cur_size_in_dw;
-		radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_COPY, 0, 0, cur_size_in_dw));
+		radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_COPY, 0, cur_size_in_dw));
 		radeon_ring_write(ring, dst_offset & 0xfffffffc);
 		radeon_ring_write(ring, src_offset & 0xfffffffc);
 		radeon_ring_write(ring, upper_32_bits(dst_offset) & 0xff);
@@ -3488,7 +3488,7 @@ static int evergreen_startup(struct radeon_device *rdev)
 	ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX];
 	r = radeon_ring_init(rdev, ring, ring->ring_size, R600_WB_DMA_RPTR_OFFSET,
 			     DMA_RB_RPTR, DMA_RB_WPTR,
-			     2, 0x3fffc, DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0));
+			     2, 0x3fffc, DMA_PACKET(DMA_PACKET_NOP, 0, 0));
 	if (r)
 		return r;
 
diff --git a/drivers/gpu/drm/radeon/evergreen_cs.c b/drivers/gpu/drm/radeon/evergreen_cs.c
index 4a9760a..d8f5d5f 100644
--- a/drivers/gpu/drm/radeon/evergreen_cs.c
+++ b/drivers/gpu/drm/radeon/evergreen_cs.c
@@ -2670,16 +2670,6 @@ int evergreen_cs_parse(struct radeon_cs_parser *p)
 	return 0;
 }
 
-/*
- *  DMA
- */
-
-#define GET_DMA_CMD(h) (((h) & 0xf0000000) >> 28)
-#define GET_DMA_COUNT(h) ((h) & 0x000fffff)
-#define GET_DMA_T(h) (((h) & 0x00800000) >> 23)
-#define GET_DMA_NEW(h) (((h) & 0x04000000) >> 26)
-#define GET_DMA_MISC(h) (((h) & 0x0700000) >> 20)
-
 /**
  * evergreen_dma_cs_parse() - parse the DMA IB
  * @p:		parser structure holding parsing context.
@@ -2693,9 +2683,9 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p)
 {
 	struct radeon_cs_chunk *ib_chunk = &p->chunks[p->chunk_ib_idx];
 	struct radeon_cs_reloc *src_reloc, *dst_reloc, *dst2_reloc;
-	u32 header, cmd, count, tiled, new_cmd, misc;
+	u32 header, cmd, count, sub_cmd;
 	volatile u32 *ib = p->ib.ptr;
-	u32 idx, idx_value;
+	u32 idx;
 	u64 src_offset, dst_offset, dst2_offset;
 	int r;
 
@@ -2709,9 +2699,7 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p)
 		header = radeon_get_ib_value(p, idx);
 		cmd = GET_DMA_CMD(header);
 		count = GET_DMA_COUNT(header);
-		tiled = GET_DMA_T(header);
-		new_cmd = GET_DMA_NEW(header);
-		misc = GET_DMA_MISC(header);
+		sub_cmd = GET_DMA_SUB_CMD(header);
 
 		switch (cmd) {
 		case DMA_PACKET_WRITE:
@@ -2720,19 +2708,27 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p)
 				DRM_ERROR("bad DMA_PACKET_WRITE\n");
 				return -EINVAL;
 			}
-			if (tiled) {
+			switch (sub_cmd) {
+			/* tiled */
+			case 8:
 				dst_offset = ib[idx+1];
 				dst_offset <<= 8;
 
 				ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8);
 				p->idx += count + 7;
-			} else {
+				break;
+			/* linear */
+			case 0:
 				dst_offset = ib[idx+1];
 				dst_offset |= ((u64)(ib[idx+2] & 0xff)) << 32;
 
 				ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc);
 				ib[idx+2] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff;
 				p->idx += count + 3;
+				break;
+			default:
+				DRM_ERROR("bad DMA_PACKET_WRITE [%6d] 0x%08x sub cmd is not 0 or 8\n", idx, ib[idx+0]);
+				return -EINVAL;
 			}
 			if ((dst_offset + (count * 4)) > radeon_bo_size(dst_reloc->robj)) {
 				dev_warn(p->dev, "DMA write buffer too small (%llu %lu)\n",
@@ -2751,338 +2747,330 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p)
 				DRM_ERROR("bad DMA_PACKET_COPY\n");
 				return -EINVAL;
 			}
-			if (tiled) {
-				idx_value = radeon_get_ib_value(p, idx + 2);
-				if (new_cmd) {
-					switch (misc) {
-					case 0:
-						/* L2T, frame to fields */
-						if (idx_value & (1 << 31)) {
-							DRM_ERROR("bad L2T, frame to fields DMA_PACKET_COPY\n");
-							return -EINVAL;
-						}
-						r = r600_dma_cs_next_reloc(p, &dst2_reloc);
-						if (r) {
-							DRM_ERROR("bad L2T, frame to fields DMA_PACKET_COPY\n");
-							return -EINVAL;
-						}
-						dst_offset = ib[idx+1];
-						dst_offset <<= 8;
-						dst2_offset = ib[idx+2];
-						dst2_offset <<= 8;
-						src_offset = ib[idx+8];
-						src_offset |= ((u64)(ib[idx+9] & 0xff)) << 32;
-						if ((src_offset + (count * 4)) > radeon_bo_size(src_reloc->robj)) {
-							dev_warn(p->dev, "DMA L2T, frame to fields src buffer too small (%llu %lu)\n",
-								 src_offset + (count * 4), radeon_bo_size(src_reloc->robj));
-							return -EINVAL;
-						}
-						if ((dst_offset + (count * 4)) > radeon_bo_size(dst_reloc->robj)) {
-							dev_warn(p->dev, "DMA L2T, frame to fields buffer too small (%llu %lu)\n",
-								 dst_offset + (count * 4), radeon_bo_size(dst_reloc->robj));
-							return -EINVAL;
-						}
-						if ((dst2_offset + (count * 4)) > radeon_bo_size(dst2_reloc->robj)) {
-							dev_warn(p->dev, "DMA L2T, frame to fields buffer too small (%llu %lu)\n",
-								 dst2_offset + (count * 4), radeon_bo_size(dst2_reloc->robj));
-							return -EINVAL;
-						}
-						ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8);
-						ib[idx+2] += (u32)(dst2_reloc->lobj.gpu_offset >> 8);
-						ib[idx+8] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc);
-						ib[idx+9] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff;
-						p->idx += 10;
-						break;
-					case 1:
-						/* L2T, T2L partial */
-						if (p->family < CHIP_CAYMAN) {
-							DRM_ERROR("L2T, T2L Partial is cayman only !\n");
-							return -EINVAL;
-						}
-						/* detile bit */
-						if (idx_value & (1 << 31)) {
-							/* tiled src, linear dst */
-							ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset >> 8);
-
-							ib[idx+7] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc);
-							ib[idx+8] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff;
-						} else {
-							/* linear src, tiled dst */
-							ib[idx+7] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc);
-							ib[idx+8] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff;
-
-							ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8);
-						}
-						p->idx += 12;
-						break;
-					case 3:
-						/* L2T, broadcast */
-						if (idx_value & (1 << 31)) {
-							DRM_ERROR("bad L2T, broadcast DMA_PACKET_COPY\n");
-							return -EINVAL;
-						}
-						r = r600_dma_cs_next_reloc(p, &dst2_reloc);
-						if (r) {
-							DRM_ERROR("bad L2T, broadcast DMA_PACKET_COPY\n");
-							return -EINVAL;
-						}
-						dst_offset = ib[idx+1];
-						dst_offset <<= 8;
-						dst2_offset = ib[idx+2];
-						dst2_offset <<= 8;
-						src_offset = ib[idx+8];
-						src_offset |= ((u64)(ib[idx+9] & 0xff)) << 32;
-						if ((src_offset + (count * 4)) > radeon_bo_size(src_reloc->robj)) {
-							dev_warn(p->dev, "DMA L2T, broadcast src buffer too small (%llu %lu)\n",
-								 src_offset + (count * 4), radeon_bo_size(src_reloc->robj));
-							return -EINVAL;
-						}
-						if ((dst_offset + (count * 4)) > radeon_bo_size(dst_reloc->robj)) {
-							dev_warn(p->dev, "DMA L2T, broadcast dst buffer too small (%llu %lu)\n",
-								 dst_offset + (count * 4), radeon_bo_size(dst_reloc->robj));
-							return -EINVAL;
-						}
-						if ((dst2_offset + (count * 4)) > radeon_bo_size(dst2_reloc->robj)) {
-							dev_warn(p->dev, "DMA L2T, broadcast dst2 buffer too small (%llu %lu)\n",
-								 dst2_offset + (count * 4), radeon_bo_size(dst2_reloc->robj));
-							return -EINVAL;
-						}
-						ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8);
-						ib[idx+2] += (u32)(dst2_reloc->lobj.gpu_offset >> 8);
-						ib[idx+8] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc);
-						ib[idx+9] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff;
-						p->idx += 10;
-						break;
-					case 4:
-						/* L2T, T2L */
-						/* detile bit */
-						if (idx_value & (1 << 31)) {
-							/* tiled src, linear dst */
-							src_offset = ib[idx+1];
-							src_offset <<= 8;
-							ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset >> 8);
-
-							dst_offset = ib[idx+7];
-							dst_offset |= ((u64)(ib[idx+8] & 0xff)) << 32;
-							ib[idx+7] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc);
-							ib[idx+8] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff;
-						} else {
-							/* linear src, tiled dst */
-							src_offset = ib[idx+7];
-							src_offset |= ((u64)(ib[idx+8] & 0xff)) << 32;
-							ib[idx+7] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc);
-							ib[idx+8] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff;
-
-							dst_offset = ib[idx+1];
-							dst_offset <<= 8;
-							ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8);
-						}
-						if ((src_offset + (count * 4)) > radeon_bo_size(src_reloc->robj)) {
-							dev_warn(p->dev, "DMA L2T, T2L src buffer too small (%llu %lu)\n",
-								 src_offset + (count * 4), radeon_bo_size(src_reloc->robj));
-							return -EINVAL;
-						}
-						if ((dst_offset + (count * 4)) > radeon_bo_size(dst_reloc->robj)) {
-							dev_warn(p->dev, "DMA L2T, T2L dst buffer too small (%llu %lu)\n",
-								 dst_offset + (count * 4), radeon_bo_size(dst_reloc->robj));
-							return -EINVAL;
-						}
-						p->idx += 9;
-						break;
-					case 5:
-						/* T2T partial */
-						if (p->family < CHIP_CAYMAN) {
-							DRM_ERROR("L2T, T2L Partial is cayman only !\n");
-							return -EINVAL;
-						}
-						ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset >> 8);
-						ib[idx+4] += (u32)(dst_reloc->lobj.gpu_offset >> 8);
-						p->idx += 13;
-						break;
-					case 7:
-						/* L2T, broadcast */
-						if (idx_value & (1 << 31)) {
-							DRM_ERROR("bad L2T, broadcast DMA_PACKET_COPY\n");
-							return -EINVAL;
-						}
-						r = r600_dma_cs_next_reloc(p, &dst2_reloc);
-						if (r) {
-							DRM_ERROR("bad L2T, broadcast DMA_PACKET_COPY\n");
-							return -EINVAL;
-						}
-						dst_offset = ib[idx+1];
-						dst_offset <<= 8;
-						dst2_offset = ib[idx+2];
-						dst2_offset <<= 8;
-						src_offset = ib[idx+8];
-						src_offset |= ((u64)(ib[idx+9] & 0xff)) << 32;
-						if ((src_offset + (count * 4)) > radeon_bo_size(src_reloc->robj)) {
-							dev_warn(p->dev, "DMA L2T, broadcast src buffer too small (%llu %lu)\n",
-								 src_offset + (count * 4), radeon_bo_size(src_reloc->robj));
-							return -EINVAL;
-						}
-						if ((dst_offset + (count * 4)) > radeon_bo_size(dst_reloc->robj)) {
-							dev_warn(p->dev, "DMA L2T, broadcast dst buffer too small (%llu %lu)\n",
-								 dst_offset + (count * 4), radeon_bo_size(dst_reloc->robj));
-							return -EINVAL;
-						}
-						if ((dst2_offset + (count * 4)) > radeon_bo_size(dst2_reloc->robj)) {
-							dev_warn(p->dev, "DMA L2T, broadcast dst2 buffer too small (%llu %lu)\n",
-								 dst2_offset + (count * 4), radeon_bo_size(dst2_reloc->robj));
-							return -EINVAL;
-						}
-						ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8);
-						ib[idx+2] += (u32)(dst2_reloc->lobj.gpu_offset >> 8);
-						ib[idx+8] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc);
-						ib[idx+9] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff;
-						p->idx += 10;
-						break;
-					default:
-						DRM_ERROR("bad DMA_PACKET_COPY misc %u\n", misc);
-						return -EINVAL;
-					}
+			switch (sub_cmd) {
+			/* Copy L2L, DW aligned */
+			case 0x00:
+				/* L2L, dw */
+				src_offset = ib[idx+2];
+				src_offset |= ((u64)(ib[idx+4] & 0xff)) << 32;
+				dst_offset = ib[idx+1];
+				dst_offset |= ((u64)(ib[idx+3] & 0xff)) << 32;
+				if ((src_offset + (count * 4)) > radeon_bo_size(src_reloc->robj)) {
+					dev_warn(p->dev, "DMA L2L, dw src buffer too small (%llu %lu)\n",
+							src_offset + (count * 4), radeon_bo_size(src_reloc->robj));
+					return -EINVAL;
+				}
+				if ((dst_offset + (count * 4)) > radeon_bo_size(dst_reloc->robj)) {
+					dev_warn(p->dev, "DMA L2L, dw dst buffer too small (%llu %lu)\n",
+							dst_offset + (count * 4), radeon_bo_size(dst_reloc->robj));
+					return -EINVAL;
+				}
+				ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc);
+				ib[idx+2] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc);
+				ib[idx+3] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff;
+				ib[idx+4] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff;
+				p->idx += 5;
+				break;
+			/* Copy L2T/T2L */
+			case 0x08:
+				/* detile bit */
+				if (ib[idx + 2] & (1 << 31)) {
+					/* tiled src, linear dst */
+					src_offset = ib[idx+1];
+					src_offset <<= 8;
+					ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset >> 8);
+
+					dst_offset = radeon_get_ib_value(p, idx + 7);
+					dst_offset |= ((u64)(ib[idx+8] & 0xff)) << 32;
+					ib[idx+7] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc);
+					ib[idx+8] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff;
 				} else {
-					switch (misc) {
-					case 0:
-						/* detile bit */
-						if (idx_value & (1 << 31)) {
-							/* tiled src, linear dst */
-							src_offset = ib[idx+1];
-							src_offset <<= 8;
-							ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset >> 8);
-
-							dst_offset = ib[idx+7];
-							dst_offset |= ((u64)(ib[idx+8] & 0xff)) << 32;
-							ib[idx+7] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc);
-							ib[idx+8] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff;
-						} else {
-							/* linear src, tiled dst */
-							src_offset = ib[idx+7];
-							src_offset |= ((u64)(ib[idx+8] & 0xff)) << 32;
-							ib[idx+7] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc);
-							ib[idx+8] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff;
-
-							dst_offset = ib[idx+1];
-							dst_offset <<= 8;
-							ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8);
-						}
-						if ((src_offset + (count * 4)) > radeon_bo_size(src_reloc->robj)) {
-							dev_warn(p->dev, "DMA L2T, broadcast src buffer too small (%llu %lu)\n",
-								 src_offset + (count * 4), radeon_bo_size(src_reloc->robj));
-							return -EINVAL;
-						}
-						if ((dst_offset + (count * 4)) > radeon_bo_size(dst_reloc->robj)) {
-							dev_warn(p->dev, "DMA L2T, broadcast dst buffer too small (%llu %lu)\n",
-								 dst_offset + (count * 4), radeon_bo_size(dst_reloc->robj));
-							return -EINVAL;
-						}
-						p->idx += 9;
-						break;
-					default:
-						DRM_ERROR("bad DMA_PACKET_COPY misc %u\n", misc);
-						return -EINVAL;
-					}
+					/* linear src, tiled dst */
+					src_offset = ib[idx+7];
+					src_offset |= ((u64)(ib[idx+8] & 0xff)) << 32;
+					ib[idx+7] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc);
+					ib[idx+8] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff;
+
+					dst_offset = ib[idx+1];
+					dst_offset <<= 8;
+					ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8);
 				}
-			} else {
-				if (new_cmd) {
-					switch (misc) {
-					case 0:
-						/* L2L, byte */
-						src_offset = ib[idx+2];
-						src_offset |= ((u64)(ib[idx+4] & 0xff)) << 32;
-						dst_offset = ib[idx+1];
-						dst_offset |= ((u64)(ib[idx+3] & 0xff)) << 32;
-						if ((src_offset + count) > radeon_bo_size(src_reloc->robj)) {
-							dev_warn(p->dev, "DMA L2L, byte src buffer too small (%llu %lu)\n",
-								 src_offset + count, radeon_bo_size(src_reloc->robj));
-							return -EINVAL;
-						}
-						if ((dst_offset + count) > radeon_bo_size(dst_reloc->robj)) {
-							dev_warn(p->dev, "DMA L2L, byte dst buffer too small (%llu %lu)\n",
-								 dst_offset + count, radeon_bo_size(dst_reloc->robj));
-							return -EINVAL;
-						}
-						ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xffffffff);
-						ib[idx+2] += (u32)(src_reloc->lobj.gpu_offset & 0xffffffff);
-						ib[idx+3] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff;
-						ib[idx+4] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff;
-						p->idx += 5;
-						break;
-					case 1:
-						/* L2L, partial */
-						if (p->family < CHIP_CAYMAN) {
-							DRM_ERROR("L2L Partial is cayman only !\n");
-							return -EINVAL;
-						}
-						ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset & 0xffffffff);
-						ib[idx+2] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff;
-						ib[idx+4] += (u32)(dst_reloc->lobj.gpu_offset & 0xffffffff);
-						ib[idx+5] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff;
-
-						p->idx += 9;
-						break;
-					case 4:
-						/* L2L, dw, broadcast */
-						r = r600_dma_cs_next_reloc(p, &dst2_reloc);
-						if (r) {
-							DRM_ERROR("bad L2L, dw, broadcast DMA_PACKET_COPY\n");
-							return -EINVAL;
-						}
-						dst_offset = ib[idx+1];
-						dst_offset |= ((u64)(ib[idx+4] & 0xff)) << 32;
-						dst2_offset = ib[idx+2];
-						dst2_offset |= ((u64)(ib[idx+5] & 0xff)) << 32;
-						src_offset = ib[idx+3];
-						src_offset |= ((u64)(ib[idx+6] & 0xff)) << 32;
-						if ((src_offset + (count * 4)) > radeon_bo_size(src_reloc->robj)) {
-							dev_warn(p->dev, "DMA L2L, dw, broadcast src buffer too small (%llu %lu)\n",
-								 src_offset + (count * 4), radeon_bo_size(src_reloc->robj));
-							return -EINVAL;
-						}
-						if ((dst_offset + (count * 4)) > radeon_bo_size(dst_reloc->robj)) {
-							dev_warn(p->dev, "DMA L2L, dw, broadcast dst buffer too small (%llu %lu)\n",
-								 dst_offset + (count * 4), radeon_bo_size(dst_reloc->robj));
-							return -EINVAL;
-						}
-						if ((dst2_offset + (count * 4)) > radeon_bo_size(dst2_reloc->robj)) {
-							dev_warn(p->dev, "DMA L2L, dw, broadcast dst2 buffer too small (%llu %lu)\n",
-								 dst2_offset + (count * 4), radeon_bo_size(dst2_reloc->robj));
-							return -EINVAL;
-						}
-						ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc);
-						ib[idx+2] += (u32)(dst2_reloc->lobj.gpu_offset & 0xfffffffc);
-						ib[idx+3] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc);
-						ib[idx+4] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff;
-						ib[idx+5] += upper_32_bits(dst2_reloc->lobj.gpu_offset) & 0xff;
-						ib[idx+6] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff;
-						p->idx += 7;
-						break;
-					default:
-						DRM_ERROR("bad DMA_PACKET_COPY misc %u\n", misc);
-						return -EINVAL;
-					}
+				if ((src_offset + (count * 4)) > radeon_bo_size(src_reloc->robj)) {
+					dev_warn(p->dev, "DMA L2T, src buffer too small (%llu %lu)\n",
+							src_offset + (count * 4), radeon_bo_size(src_reloc->robj));
+					return -EINVAL;
+				}
+				if ((dst_offset + (count * 4)) > radeon_bo_size(dst_reloc->robj)) {
+					dev_warn(p->dev, "DMA L2T, dst buffer too small (%llu %lu)\n",
+							dst_offset + (count * 4), radeon_bo_size(dst_reloc->robj));
+					return -EINVAL;
+				}
+				p->idx += 9;
+				break;
+			/* Copy L2L, byte aligned */
+			case 0x40:
+				/* L2L, byte */
+				src_offset = ib[idx+2];
+				src_offset |= ((u64)(ib[idx+4] & 0xff)) << 32;
+				dst_offset = ib[idx+1];
+				dst_offset |= ((u64)(ib[idx+3] & 0xff)) << 32;
+				if ((src_offset + count) > radeon_bo_size(src_reloc->robj)) {
+					dev_warn(p->dev, "DMA L2L, byte src buffer too small (%llu %lu)\n",
+							src_offset + count, radeon_bo_size(src_reloc->robj));
+					return -EINVAL;
+				}
+				if ((dst_offset + count) > radeon_bo_size(dst_reloc->robj)) {
+					dev_warn(p->dev, "DMA L2L, byte dst buffer too small (%llu %lu)\n",
+							dst_offset + count, radeon_bo_size(dst_reloc->robj));
+					return -EINVAL;
+				}
+				ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xffffffff);
+				ib[idx+2] += (u32)(src_reloc->lobj.gpu_offset & 0xffffffff);
+				ib[idx+3] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff;
+				ib[idx+4] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff;
+				p->idx += 5;
+				break;
+			/* Copy L2L, partial */
+			case 0x41:
+				/* L2L, partial */
+				if (p->family < CHIP_CAYMAN) {
+					DRM_ERROR("L2L Partial is cayman only !\n");
+					return -EINVAL;
+				}
+				ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset & 0xffffffff);
+				ib[idx+2] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff;
+				ib[idx+4] += (u32)(dst_reloc->lobj.gpu_offset & 0xffffffff);
+				ib[idx+5] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff;
+
+				p->idx += 9;
+				break;
+			/* Copy L2L, DW aligned, broadcast */
+			case 0x44:
+				/* L2L, dw, broadcast */
+				r = r600_dma_cs_next_reloc(p, &dst2_reloc);
+				if (r) {
+					DRM_ERROR("bad L2L, dw, broadcast DMA_PACKET_COPY\n");
+					return -EINVAL;
+				}
+				dst_offset = ib[idx+1];
+				dst_offset |= ((u64)(ib[idx+4] & 0xff)) << 32;
+				dst2_offset = ib[idx+2];
+				dst2_offset |= ((u64)(ib[idx+5] & 0xff)) << 32;
+				src_offset = ib[idx+3];
+				src_offset |= ((u64)(ib[idx+6] & 0xff)) << 32;
+				if ((src_offset + (count * 4)) > radeon_bo_size(src_reloc->robj)) {
+					dev_warn(p->dev, "DMA L2L, dw, broadcast src buffer too small (%llu %lu)\n",
+							src_offset + (count * 4), radeon_bo_size(src_reloc->robj));
+					return -EINVAL;
+				}
+				if ((dst_offset + (count * 4)) > radeon_bo_size(dst_reloc->robj)) {
+					dev_warn(p->dev, "DMA L2L, dw, broadcast dst buffer too small (%llu %lu)\n",
+							dst_offset + (count * 4), radeon_bo_size(dst_reloc->robj));
+					return -EINVAL;
+				}
+				if ((dst2_offset + (count * 4)) > radeon_bo_size(dst2_reloc->robj)) {
+					dev_warn(p->dev, "DMA L2L, dw, broadcast dst2 buffer too small (%llu %lu)\n",
+							dst2_offset + (count * 4), radeon_bo_size(dst2_reloc->robj));
+					return -EINVAL;
+				}
+				ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc);
+				ib[idx+2] += (u32)(dst2_reloc->lobj.gpu_offset & 0xfffffffc);
+				ib[idx+3] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc);
+				ib[idx+4] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff;
+				ib[idx+5] += upper_32_bits(dst2_reloc->lobj.gpu_offset) & 0xff;
+				ib[idx+6] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff;
+				p->idx += 7;
+				break;
+			/* Copy L2T Frame to Field */
+			case 0x48:
+				if (ib[idx + 2] & (1 << 31)) {
+					DRM_ERROR("bad L2T, frame to fields DMA_PACKET_COPY\n");
+					return -EINVAL;
+				}
+				r = r600_dma_cs_next_reloc(p, &dst2_reloc);
+				if (r) {
+					DRM_ERROR("bad L2T, frame to fields DMA_PACKET_COPY\n");
+					return -EINVAL;
+				}
+				dst_offset = ib[idx+1];
+				dst_offset <<= 8;
+				dst2_offset = ib[idx+2];
+				dst2_offset <<= 8;
+				src_offset = ib[idx+8];
+				src_offset |= ((u64)(ib[idx+9] & 0xff)) << 32;
+				if ((src_offset + (count * 4)) > radeon_bo_size(src_reloc->robj)) {
+					dev_warn(p->dev, "DMA L2T, frame to fields src buffer too small (%llu %lu)\n",
+							src_offset + (count * 4), radeon_bo_size(src_reloc->robj));
+					return -EINVAL;
+				}
+				if ((dst_offset + (count * 4)) > radeon_bo_size(dst_reloc->robj)) {
+					dev_warn(p->dev, "DMA L2T, frame to fields buffer too small (%llu %lu)\n",
+							dst_offset + (count * 4), radeon_bo_size(dst_reloc->robj));
+					return -EINVAL;
+				}
+				if ((dst2_offset + (count * 4)) > radeon_bo_size(dst2_reloc->robj)) {
+					dev_warn(p->dev, "DMA L2T, frame to fields buffer too small (%llu %lu)\n",
+							dst2_offset + (count * 4), radeon_bo_size(dst2_reloc->robj));
+					return -EINVAL;
+				}
+				ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8);
+				ib[idx+2] += (u32)(dst2_reloc->lobj.gpu_offset >> 8);
+				ib[idx+8] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc);
+				ib[idx+9] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff;
+				p->idx += 10;
+				break;
+			/* Copy L2T/T2L, partial */
+			case 0x49:
+				/* L2T, T2L partial */
+				if (p->family < CHIP_CAYMAN) {
+					DRM_ERROR("L2T, T2L Partial is cayman only !\n");
+					return -EINVAL;
+				}
+				/* detile bit */
+				if (ib[idx + 2 ] & (1 << 31)) {
+					/* tiled src, linear dst */
+					ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset >> 8);
+
+					ib[idx+7] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc);
+					ib[idx+8] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff;
+				} else {
+					/* linear src, tiled dst */
+					ib[idx+7] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc);
+					ib[idx+8] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff;
+
+					ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8);
+				}
+				p->idx += 12;
+				break;
+			/* Copy L2T broadcast */
+			case 0x4b:
+				/* L2T, broadcast */
+				if (ib[idx + 2] & (1 << 31)) {
+					DRM_ERROR("bad L2T, broadcast DMA_PACKET_COPY\n");
+					return -EINVAL;
+				}
+				r = r600_dma_cs_next_reloc(p, &dst2_reloc);
+				if (r) {
+					DRM_ERROR("bad L2T, broadcast DMA_PACKET_COPY\n");
+					return -EINVAL;
+				}
+				dst_offset = ib[idx+1];
+				dst_offset <<= 8;
+				dst2_offset = ib[idx+2];
+				dst2_offset <<= 8;
+				src_offset = ib[idx+8];
+				src_offset |= ((u64)(ib[idx+9] & 0xff)) << 32;
+				if ((src_offset + (count * 4)) > radeon_bo_size(src_reloc->robj)) {
+					dev_warn(p->dev, "DMA L2T, broadcast src buffer too small (%llu %lu)\n",
+							src_offset + (count * 4), radeon_bo_size(src_reloc->robj));
+					return -EINVAL;
+				}
+				if ((dst_offset + (count * 4)) > radeon_bo_size(dst_reloc->robj)) {
+					dev_warn(p->dev, "DMA L2T, broadcast dst buffer too small (%llu %lu)\n",
+							dst_offset + (count * 4), radeon_bo_size(dst_reloc->robj));
+					return -EINVAL;
+				}
+				if ((dst2_offset + (count * 4)) > radeon_bo_size(dst2_reloc->robj)) {
+					dev_warn(p->dev, "DMA L2T, broadcast dst2 buffer too small (%llu %lu)\n",
+							dst2_offset + (count * 4), radeon_bo_size(dst2_reloc->robj));
+					return -EINVAL;
+				}
+				ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8);
+				ib[idx+2] += (u32)(dst2_reloc->lobj.gpu_offset >> 8);
+				ib[idx+8] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc);
+				ib[idx+9] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff;
+				p->idx += 10;
+				break;
+			/* Copy L2T/T2L (tile units) */
+			case 0x4c:
+				/* L2T, T2L */
+				/* detile bit */
+				if (ib[idx + 2] & (1 << 31)) {
+					/* tiled src, linear dst */
+					src_offset = ib[idx+1];
+					src_offset <<= 8;
+					ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset >> 8);
+
+					dst_offset = ib[idx+7];
+					dst_offset |= ((u64)(ib[idx+8] & 0xff)) << 32;
+					ib[idx+7] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc);
+					ib[idx+8] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff;
 				} else {
-					/* L2L, dw */
-					src_offset = ib[idx+2];
-					src_offset |= ((u64)(ib[idx+4] & 0xff)) << 32;
+					/* linear src, tiled dst */
+					src_offset = ib[idx+7];
+					src_offset |= ((u64)(ib[idx+8] & 0xff)) << 32;
+					ib[idx+7] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc);
+					ib[idx+8] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff;
+
 					dst_offset = ib[idx+1];
-					dst_offset |= ((u64)(ib[idx+3] & 0xff)) << 32;
-					if ((src_offset + (count * 4)) > radeon_bo_size(src_reloc->robj)) {
-						dev_warn(p->dev, "DMA L2L, dw src buffer too small (%llu %lu)\n",
-							 src_offset + (count * 4), radeon_bo_size(src_reloc->robj));
-						return -EINVAL;
-					}
-					if ((dst_offset + (count * 4)) > radeon_bo_size(dst_reloc->robj)) {
-						dev_warn(p->dev, "DMA L2L, dw dst buffer too small (%llu %lu)\n",
-							 dst_offset + (count * 4), radeon_bo_size(dst_reloc->robj));
-						return -EINVAL;
-					}
-					ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc);
-					ib[idx+2] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc);
-					ib[idx+3] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff;
-					ib[idx+4] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff;
-					p->idx += 5;
+					dst_offset <<= 8;
+					ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8);
 				}
+				if ((src_offset + (count * 4)) > radeon_bo_size(src_reloc->robj)) {
+					dev_warn(p->dev, "DMA L2T, T2L src buffer too small (%llu %lu)\n",
+							src_offset + (count * 4), radeon_bo_size(src_reloc->robj));
+					return -EINVAL;
+				}
+				if ((dst_offset + (count * 4)) > radeon_bo_size(dst_reloc->robj)) {
+					dev_warn(p->dev, "DMA L2T, T2L dst buffer too small (%llu %lu)\n",
+							dst_offset + (count * 4), radeon_bo_size(dst_reloc->robj));
+					return -EINVAL;
+				}
+				p->idx += 9;
+				break;
+			/* Copy T2T, partial (tile units) */
+			case 0x4d:
+				/* T2T partial */
+				if (p->family < CHIP_CAYMAN) {
+					DRM_ERROR("L2T, T2L Partial is cayman only !\n");
+					return -EINVAL;
+				}
+				ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset >> 8);
+				ib[idx+4] += (u32)(dst_reloc->lobj.gpu_offset >> 8);
+				p->idx += 13;
+				break;
+			/* Copy L2T broadcast (tile units) */
+			case 0x4f:
+				/* L2T, broadcast */
+				if (ib[idx + 2] & (1 << 31)) {
+					DRM_ERROR("bad L2T, broadcast DMA_PACKET_COPY\n");
+					return -EINVAL;
+				}
+				r = r600_dma_cs_next_reloc(p, &dst2_reloc);
+				if (r) {
+					DRM_ERROR("bad L2T, broadcast DMA_PACKET_COPY\n");
+					return -EINVAL;
+				}
+				dst_offset = ib[idx+1];
+				dst_offset <<= 8;
+				dst2_offset = ib[idx+2];
+				dst2_offset <<= 8;
+				src_offset = ib[idx+8];
+				src_offset |= ((u64)(ib[idx+9] & 0xff)) << 32;
+				if ((src_offset + (count * 4)) > radeon_bo_size(src_reloc->robj)) {
+					dev_warn(p->dev, "DMA L2T, broadcast src buffer too small (%llu %lu)\n",
+							src_offset + (count * 4), radeon_bo_size(src_reloc->robj));
+					return -EINVAL;
+				}
+				if ((dst_offset + (count * 4)) > radeon_bo_size(dst_reloc->robj)) {
+					dev_warn(p->dev, "DMA L2T, broadcast dst buffer too small (%llu %lu)\n",
+							dst_offset + (count * 4), radeon_bo_size(dst_reloc->robj));
+					return -EINVAL;
+				}
+				if ((dst2_offset + (count * 4)) > radeon_bo_size(dst2_reloc->robj)) {
+					dev_warn(p->dev, "DMA L2T, broadcast dst2 buffer too small (%llu %lu)\n",
+							dst2_offset + (count * 4), radeon_bo_size(dst2_reloc->robj));
+					return -EINVAL;
+				}
+				ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8);
+				ib[idx+2] += (u32)(dst2_reloc->lobj.gpu_offset >> 8);
+				ib[idx+8] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc);
+				ib[idx+9] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff;
+				p->idx += 10;
+				break;
+			default:
+				DRM_ERROR("bad DMA_PACKET_COPY [%6d] 0x%08x invalid sub cmd\n", idx, ib[idx+0]);
+				return -EINVAL;
 			}
 			break;
 		case DMA_PACKET_CONSTANT_FILL:
@@ -3435,88 +3423,79 @@ int evergreen_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib)
 int evergreen_dma_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib)
 {
 	u32 idx = 0;
-	u32 header, cmd, count, tiled, new_cmd, misc;
+	u32 header, cmd, count, sub_cmd;
 
 	do {
 		header = ib->ptr[idx];
 		cmd = GET_DMA_CMD(header);
 		count = GET_DMA_COUNT(header);
-		tiled = GET_DMA_T(header);
-		new_cmd = GET_DMA_NEW(header);
-		misc = GET_DMA_MISC(header);
+		sub_cmd = GET_DMA_SUB_CMD(header);
 
 		switch (cmd) {
 		case DMA_PACKET_WRITE:
-			if (tiled)
+			switch (sub_cmd) {
+			/* tiled */
+			case 8:
 				idx += count + 7;
-			else
+				break;
+			/* linear */
+			case 0:
 				idx += count + 3;
+				break;
+			default:
+				DRM_ERROR("bad DMA_PACKET_WRITE [%6d] 0x%08x sub cmd is not 0 or 8\n", idx, ib->ptr[idx]);
+				return -EINVAL;
+			}
 			break;
 		case DMA_PACKET_COPY:
-			if (tiled) {
-				if (new_cmd) {
-					switch (misc) {
-					case 0:
-						/* L2T, frame to fields */
-						idx += 10;
-						break;
-					case 1:
-						/* L2T, T2L partial */
-						idx += 12;
-						break;
-					case 3:
-						/* L2T, broadcast */
-						idx += 10;
-						break;
-					case 4:
-						/* L2T, T2L */
-						idx += 9;
-						break;
-					case 5:
-						/* T2T partial */
-						idx += 13;
-						break;
-					case 7:
-						/* L2T, broadcast */
-						idx += 10;
-						break;
-					default:
-						DRM_ERROR("bad DMA_PACKET_COPY misc %u\n", misc);
-						return -EINVAL;
-					}
-				} else {
-					switch (misc) {
-					case 0:
-						idx += 9;
-						break;
-					default:
-						DRM_ERROR("bad DMA_PACKET_COPY misc %u\n", misc);
-						return -EINVAL;
-					}
-				}
-			} else {
-				if (new_cmd) {
-					switch (misc) {
-					case 0:
-						/* L2L, byte */
-						idx += 5;
-						break;
-					case 1:
-						/* L2L, partial */
-						idx += 9;
-						break;
-					case 4:
-						/* L2L, dw, broadcast */
-						idx += 7;
-						break;
-					default:
-						DRM_ERROR("bad DMA_PACKET_COPY misc %u\n", misc);
-						return -EINVAL;
-					}
-				} else {
-					/* L2L, dw */
-					idx += 5;
-				}
+			switch (sub_cmd) {
+			/* Copy L2L, DW aligned */
+			case 0x00:
+				idx += 5;
+				break;
+			/* Copy L2T/T2L */
+			case 0x08:
+				idx += 9;
+				break;
+			/* Copy L2L, byte aligned */
+			case 0x40:
+				idx += 5;
+				break;
+			/* Copy L2L, partial */
+			case 0x41:
+				idx += 9;
+				break;
+			/* Copy L2L, DW aligned, broadcast */
+			case 0x44:
+				idx += 7;
+				break;
+			/* Copy L2T Frame to Field */
+			case 0x48:
+				idx += 10;
+				break;
+			/* Copy L2T/T2L, partial */
+			case 0x49:
+				idx += 12;
+				break;
+			/* Copy L2T broadcast */
+			case 0x4b:
+				idx += 10;
+				break;
+			/* Copy L2T/T2L (tile units) */
+			case 0x4c:
+				idx += 9;
+				break;
+			/* Copy T2T, partial (tile units) */
+			case 0x4d:
+				idx += 13;
+				break;
+			/* Copy L2T broadcast (tile units) */
+			case 0x4f:
+				idx += 10;
+				break;
+			default:
+				DRM_ERROR("bad DMA_PACKET_COPY [%6d] 0x%08x invalid sub cmd\n", idx, ib->ptr[idx]);
+				return -EINVAL;
 			}
 			break;
 		case DMA_PACKET_CONSTANT_FILL:
diff --git a/drivers/gpu/drm/radeon/evergreend.h b/drivers/gpu/drm/radeon/evergreend.h
index 70388fe..5ad29f3 100644
--- a/drivers/gpu/drm/radeon/evergreend.h
+++ b/drivers/gpu/drm/radeon/evergreend.h
@@ -924,20 +924,23 @@
 #define CAYMAN_DMA1_CNTL                                  0xd82c
 
 /* async DMA packets */
-#define DMA_PACKET(cmd, t, s, n)	((((cmd) & 0xF) << 28) |	\
-					 (((t) & 0x1) << 23) |		\
-					 (((s) & 0x1) << 22) |		\
-					 (((n) & 0xFFFFF) << 0))
+#define DMA_PACKET(cmd, sub_cmd, n) ((((cmd) & 0xF) << 28) |    \
+                                    (((sub_cmd) & 0xFF) << 20) |\
+                                    (((n) & 0xFFFFF) << 0))
+#define GET_DMA_CMD(h) (((h) & 0xf0000000) >> 28)
+#define GET_DMA_COUNT(h) ((h) & 0x000fffff)
+#define GET_DMA_SUB_CMD(h) (((h) & 0x0ff00000) >> 20)
+
 /* async DMA Packet types */
-#define	DMA_PACKET_WRITE				  0x2
-#define	DMA_PACKET_COPY					  0x3
-#define	DMA_PACKET_INDIRECT_BUFFER			  0x4
-#define	DMA_PACKET_SEMAPHORE				  0x5
-#define	DMA_PACKET_FENCE				  0x6
-#define	DMA_PACKET_TRAP					  0x7
-#define	DMA_PACKET_SRBM_WRITE				  0x9
-#define	DMA_PACKET_CONSTANT_FILL			  0xd
-#define	DMA_PACKET_NOP					  0xf
+#define	DMA_PACKET_WRITE                        0x2
+#define	DMA_PACKET_COPY                         0x3
+#define	DMA_PACKET_INDIRECT_BUFFER              0x4
+#define	DMA_PACKET_SEMAPHORE                    0x5
+#define	DMA_PACKET_FENCE                        0x6
+#define	DMA_PACKET_TRAP                         0x7
+#define	DMA_PACKET_SRBM_WRITE                   0x9
+#define	DMA_PACKET_CONSTANT_FILL                0xd
+#define	DMA_PACKET_NOP                          0xf
 
 /* PCIE link stuff */
 #define PCIE_LC_TRAINING_CNTL                             0xa1 /* PCIE_P */
-- 
cgit v0.10.2


From 14adc89298f894816ea2f3aef4d8d2d6ede18575 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Christian=20K=C3=B6nig?= <christian.koenig@amd.com>
Date: Mon, 21 Jan 2013 13:58:46 +0100
Subject: drm/radeon: Deprecate UMS support v2
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

KMS support is out and stable for a couple of years now and
the userspace code has deprecated or abandoned the old UMS interface.

So make the KMS interface the default and deprecate the UMS interface
in the kernel as well.

v2: rebased on alex/drm-next-3.9-wip

Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
index 983201b..3399b20 100644
--- a/drivers/gpu/drm/Kconfig
+++ b/drivers/gpu/drm/Kconfig
@@ -96,6 +96,7 @@ config DRM_RADEON
         select DRM_TTM
 	select POWER_SUPPLY
 	select HWMON
+	select BACKLIGHT_CLASS_DEVICE
 	help
 	  Choose this option if you have an ATI Radeon graphics card.  There
 	  are both PCI and AGP versions.  You don't need to choose this to
diff --git a/drivers/gpu/drm/radeon/Kconfig b/drivers/gpu/drm/radeon/Kconfig
index ea92bbe..970f8e9 100644
--- a/drivers/gpu/drm/radeon/Kconfig
+++ b/drivers/gpu/drm/radeon/Kconfig
@@ -1,31 +1,8 @@
-config DRM_RADEON_KMS
-	bool "Enable modesetting on radeon by default - NEW DRIVER"
+config DRM_RADEON_UMS
+	bool "Enable userspace modesetting on radeon (DEPRECATED)"
 	depends on DRM_RADEON
-	select BACKLIGHT_CLASS_DEVICE
 	help
-	  Choose this option if you want kernel modesetting enabled by default.
+	  Choose this option if you still need userspace modesetting.
 
-	  This is a completely new driver. It's only part of the existing drm
-	  for compatibility reasons. It requires an entirely different graphics
-	  stack above it and works very differently from the old drm stack.
-	  i.e. don't enable this unless you know what you are doing it may
-	  cause issues or bugs compared to the previous userspace driver stack.
-
-	  When kernel modesetting is enabled the IOCTL of radeon/drm
-	  driver are considered as invalid and an error message is printed
-	  in the log and they return failure.
-
-	  KMS enabled userspace will use new API to talk with the radeon/drm
-	  driver. The new API provide functions to create/destroy/share/mmap
-	  buffer object which are then managed by the kernel memory manager
-	  (here TTM). In order to submit command to the GPU the userspace
-	  provide a buffer holding the command stream, along this buffer
-	  userspace have to provide a list of buffer object used by the
-	  command stream. The kernel radeon driver will then place buffer
-	  in GPU accessible memory and will update command stream to reflect
-	  the position of the different buffers.
-
-	  The kernel will also perform security check on command stream
-	  provided by the user, we want to catch and forbid any illegal use
-	  of the GPU such as DMA into random system memory or into memory
-	  not owned by the process supplying the command stream.
+	  Userspace modesetting is deprecated for quite some time now, so
+	  enable this only if you have ancient versions of the DDX drivers.
diff --git a/drivers/gpu/drm/radeon/Makefile b/drivers/gpu/drm/radeon/Makefile
index a6598fd..bf17252 100644
--- a/drivers/gpu/drm/radeon/Makefile
+++ b/drivers/gpu/drm/radeon/Makefile
@@ -56,8 +56,12 @@ $(obj)/r600_cs.o: $(obj)/r600_reg_safe.h
 
 $(obj)/evergreen_cs.o: $(obj)/evergreen_reg_safe.h $(obj)/cayman_reg_safe.h
 
-radeon-y := radeon_drv.o radeon_cp.o radeon_state.o radeon_mem.o \
-	radeon_irq.o r300_cmdbuf.o r600_cp.o
+radeon-y := radeon_drv.o
+
+# add UMS driver
+radeon-$(CONFIG_DRM_RADEON_UMS)+= radeon_cp.o radeon_state.o radeon_mem.o \
+	radeon_irq.o r300_cmdbuf.o r600_cp.o r600_blit.o
+
 # add KMS driver
 radeon-y += radeon_device.o radeon_asic.o radeon_kms.o \
 	radeon_atombios.o radeon_agp.o atombios_crtc.o radeon_combios.o \
@@ -67,7 +71,7 @@ radeon-y += radeon_device.o radeon_asic.o radeon_kms.o \
 	radeon_clocks.o radeon_fb.o radeon_gem.o radeon_ring.o radeon_irq_kms.o \
 	radeon_cs.o radeon_bios.o radeon_benchmark.o r100.o r300.o r420.o \
 	rs400.o rs600.o rs690.o rv515.o r520.o r600.o rv770.o radeon_test.o \
-	r200.o radeon_legacy_tv.o r600_cs.o r600_blit.o r600_blit_shaders.o \
+	r200.o radeon_legacy_tv.o r600_cs.o r600_blit_shaders.o \
 	r600_blit_kms.o radeon_pm.o atombios_dp.o r600_audio.o r600_hdmi.o \
 	evergreen.o evergreen_cs.o evergreen_blit_shaders.o evergreen_blit_kms.o \
 	evergreen_hdmi.o radeon_trace_points.o ni.o cayman_blit_shaders.o \
diff --git a/drivers/gpu/drm/radeon/r300_cmdbuf.c b/drivers/gpu/drm/radeon/r300_cmdbuf.c
index 002ab03..865e2c9 100644
--- a/drivers/gpu/drm/radeon/r300_cmdbuf.c
+++ b/drivers/gpu/drm/radeon/r300_cmdbuf.c
@@ -29,6 +29,8 @@
  *
  * Authors:
  *    Nicolai Haehnle <prefect_@gmx.net>
+ *
+ * ------------------------ This file is DEPRECATED! -------------------------
  */
 
 #include <drm/drmP.h>
diff --git a/drivers/gpu/drm/radeon/r600_blit.c b/drivers/gpu/drm/radeon/r600_blit.c
index 77da1f9..f651881 100644
--- a/drivers/gpu/drm/radeon/r600_blit.c
+++ b/drivers/gpu/drm/radeon/r600_blit.c
@@ -22,6 +22,8 @@
  *
  * Authors:
  *     Alex Deucher <alexander.deucher@amd.com>
+ *
+ * ------------------------ This file is DEPRECATED! -------------------------
  */
 #include <drm/drmP.h>
 #include <drm/radeon_drm.h>
@@ -488,37 +490,6 @@ set_default_state(drm_radeon_private_t *dev_priv)
 	ADVANCE_RING();
 }
 
-/* 23 bits of float fractional data */
-#define I2F_FRAC_BITS  23
-#define I2F_MASK ((1 << I2F_FRAC_BITS) - 1)
-
-/*
- * Converts unsigned integer into 32-bit IEEE floating point representation.
- * Will be exact from 0 to 2^24.  Above that, we round towards zero
- * as the fractional bits will not fit in a float.  (It would be better to
- * round towards even as the fpu does, but that is slower.)
- */
-__pure uint32_t int2float(uint32_t x)
-{
-	uint32_t msb, exponent, fraction;
-
-	/* Zero is special */
-	if (!x) return 0;
-
-	/* Get location of the most significant bit */
-	msb = __fls(x);
-
-	/*
-	 * Use a rotate instead of a shift because that works both leftwards
-	 * and rightwards due to the mod(32) behaviour.  This means we don't
-	 * need to check to see if we are above 2^24 or not.
-	 */
-	fraction = ror32(x, (msb - I2F_FRAC_BITS) & 0x1f) & I2F_MASK;
-	exponent = (127 + msb) << I2F_FRAC_BITS;
-
-	return fraction + exponent;
-}
-
 static int r600_nomm_get_vb(struct drm_device *dev)
 {
 	drm_radeon_private_t *dev_priv = dev->dev_private;
diff --git a/drivers/gpu/drm/radeon/r600_blit_kms.c b/drivers/gpu/drm/radeon/r600_blit_kms.c
index e082dca..9fb5780 100644
--- a/drivers/gpu/drm/radeon/r600_blit_kms.c
+++ b/drivers/gpu/drm/radeon/r600_blit_kms.c
@@ -31,6 +31,37 @@
 #include "r600_blit_shaders.h"
 #include "radeon_blit_common.h"
 
+/* 23 bits of float fractional data */
+#define I2F_FRAC_BITS  23
+#define I2F_MASK ((1 << I2F_FRAC_BITS) - 1)
+
+/*
+ * Converts unsigned integer into 32-bit IEEE floating point representation.
+ * Will be exact from 0 to 2^24.  Above that, we round towards zero
+ * as the fractional bits will not fit in a float.  (It would be better to
+ * round towards even as the fpu does, but that is slower.)
+ */
+__pure uint32_t int2float(uint32_t x)
+{
+	uint32_t msb, exponent, fraction;
+
+	/* Zero is special */
+	if (!x) return 0;
+
+	/* Get location of the most significant bit */
+	msb = __fls(x);
+
+	/*
+	 * Use a rotate instead of a shift because that works both leftwards
+	 * and rightwards due to the mod(32) behaviour.  This means we don't
+	 * need to check to see if we are above 2^24 or not.
+	 */
+	fraction = ror32(x, (msb - I2F_FRAC_BITS) & 0x1f) & I2F_MASK;
+	exponent = (127 + msb) << I2F_FRAC_BITS;
+
+	return fraction + exponent;
+}
+
 /* emits 21 on rv770+, 23 on r600 */
 static void
 set_render_target(struct radeon_device *rdev, int format,
diff --git a/drivers/gpu/drm/radeon/r600_cp.c b/drivers/gpu/drm/radeon/r600_cp.c
index be85f75..1c51c08 100644
--- a/drivers/gpu/drm/radeon/r600_cp.c
+++ b/drivers/gpu/drm/radeon/r600_cp.c
@@ -24,6 +24,8 @@
  * Authors:
  *     Dave Airlie <airlied@redhat.com>
  *     Alex Deucher <alexander.deucher@amd.com>
+ *
+ * ------------------------ This file is DEPRECATED! -------------------------
  */
 
 #include <linux/module.h>
diff --git a/drivers/gpu/drm/radeon/r600_cs.c b/drivers/gpu/drm/radeon/r600_cs.c
index 0d21016..931a702 100644
--- a/drivers/gpu/drm/radeon/r600_cs.c
+++ b/drivers/gpu/drm/radeon/r600_cs.c
@@ -2313,17 +2313,7 @@ int r600_cs_parse(struct radeon_cs_parser *p)
 	return 0;
 }
 
-static int r600_cs_parser_relocs_legacy(struct radeon_cs_parser *p)
-{
-	if (p->chunk_relocs_idx == -1) {
-		return 0;
-	}
-	p->relocs = kzalloc(sizeof(struct radeon_cs_reloc), GFP_KERNEL);
-	if (p->relocs == NULL) {
-		return -ENOMEM;
-	}
-	return 0;
-}
+#ifdef CONFIG_DRM_RADEON_UMS
 
 /**
  * cs_parser_fini() - clean parser states
@@ -2349,6 +2339,18 @@ static void r600_cs_parser_fini(struct radeon_cs_parser *parser, int error)
 	kfree(parser->chunks_array);
 }
 
+static int r600_cs_parser_relocs_legacy(struct radeon_cs_parser *p)
+{
+	if (p->chunk_relocs_idx == -1) {
+		return 0;
+	}
+	p->relocs = kzalloc(sizeof(struct radeon_cs_reloc), GFP_KERNEL);
+	if (p->relocs == NULL) {
+		return -ENOMEM;
+	}
+	return 0;
+}
+
 int r600_cs_legacy(struct drm_device *dev, void *data, struct drm_file *filp,
 			unsigned family, u32 *ib, int *l)
 {
@@ -2410,6 +2412,8 @@ void r600_cs_legacy_init(void)
 	r600_nomm = 1;
 }
 
+#endif
+
 /*
  *  DMA
  */
diff --git a/drivers/gpu/drm/radeon/radeon_cp.c b/drivers/gpu/drm/radeon/radeon_cp.c
index 9143fc4..efc4f64 100644
--- a/drivers/gpu/drm/radeon/radeon_cp.c
+++ b/drivers/gpu/drm/radeon/radeon_cp.c
@@ -27,6 +27,8 @@
  * Authors:
  *    Kevin E. Martin <martin@valinux.com>
  *    Gareth Hughes <gareth@valinux.com>
+ *
+ * ------------------------ This file is DEPRECATED! -------------------------
  */
 
 #include <linux/module.h>
diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c
index d9bf96e..833484d 100644
--- a/drivers/gpu/drm/radeon/radeon_drv.c
+++ b/drivers/gpu/drm/radeon/radeon_drv.c
@@ -123,15 +123,25 @@ struct dma_buf *radeon_gem_prime_export(struct drm_device *dev,
 					int flags);
 struct drm_gem_object *radeon_gem_prime_import(struct drm_device *dev,
 					       struct dma_buf *dma_buf);
+extern long radeon_kms_compat_ioctl(struct file *filp, unsigned int cmd,
+				    unsigned long arg);
 
 #if defined(CONFIG_DEBUG_FS)
 int radeon_debugfs_init(struct drm_minor *minor);
 void radeon_debugfs_cleanup(struct drm_minor *minor);
 #endif
 
+/* atpx handler */
+#if defined(CONFIG_VGA_SWITCHEROO)
+void radeon_register_atpx_handler(void);
+void radeon_unregister_atpx_handler(void);
+#else
+static inline void radeon_register_atpx_handler(void) {}
+static inline void radeon_unregister_atpx_handler(void) {}
+#endif
 
 int radeon_no_wb;
-int radeon_modeset = -1;
+int radeon_modeset = 1;
 int radeon_dynclks = -1;
 int radeon_r4xx_atom = 0;
 int radeon_agpmode = 0;
@@ -199,6 +209,14 @@ module_param_named(msi, radeon_msi, int, 0444);
 MODULE_PARM_DESC(lockup_timeout, "GPU lockup timeout in ms (defaul 10000 = 10 seconds, 0 = disable)");
 module_param_named(lockup_timeout, radeon_lockup_timeout, int, 0444);
 
+static struct pci_device_id pciidlist[] = {
+	radeon_PCI_IDS
+};
+
+MODULE_DEVICE_TABLE(pci, pciidlist);
+
+#ifdef CONFIG_DRM_RADEON_UMS
+
 static int radeon_suspend(struct drm_device *dev, pm_message_t state)
 {
 	drm_radeon_private_t *dev_priv = dev->dev_private;
@@ -227,14 +245,6 @@ static int radeon_resume(struct drm_device *dev)
 	return 0;
 }
 
-static struct pci_device_id pciidlist[] = {
-	radeon_PCI_IDS
-};
-
-#if defined(CONFIG_DRM_RADEON_KMS)
-MODULE_DEVICE_TABLE(pci, pciidlist);
-#endif
-
 static const struct file_operations radeon_driver_old_fops = {
 	.owner = THIS_MODULE,
 	.open = drm_open,
@@ -284,6 +294,8 @@ static struct drm_driver driver_old = {
 	.patchlevel = DRIVER_PATCHLEVEL,
 };
 
+#endif
+
 static struct drm_driver kms_driver;
 
 static int radeon_kick_out_firmware_fb(struct pci_dev *pdev)
@@ -411,10 +423,12 @@ static struct drm_driver kms_driver = {
 static struct drm_driver *driver;
 static struct pci_driver *pdriver;
 
+#ifdef CONFIG_DRM_RADEON_UMS
 static struct pci_driver radeon_pci_driver = {
 	.name = DRIVER_NAME,
 	.id_table = pciidlist,
 };
+#endif
 
 static struct pci_driver radeon_kms_pci_driver = {
 	.name = DRIVER_NAME,
@@ -427,28 +441,6 @@ static struct pci_driver radeon_kms_pci_driver = {
 
 static int __init radeon_init(void)
 {
-	driver = &driver_old;
-	pdriver = &radeon_pci_driver;
-	driver->num_ioctls = radeon_max_ioctl;
-#ifdef CONFIG_VGA_CONSOLE
-	if (vgacon_text_force() && radeon_modeset == -1) {
-		DRM_INFO("VGACON disable radeon kernel modesetting.\n");
-		driver = &driver_old;
-		pdriver = &radeon_pci_driver;
-		driver->driver_features &= ~DRIVER_MODESET;
-		radeon_modeset = 0;
-	}
-#endif
-	/* if enabled by default */
-	if (radeon_modeset == -1) {
-#ifdef CONFIG_DRM_RADEON_KMS
-		DRM_INFO("radeon defaulting to kernel modesetting.\n");
-		radeon_modeset = 1;
-#else
-		DRM_INFO("radeon defaulting to userspace modesetting.\n");
-		radeon_modeset = 0;
-#endif
-	}
 	if (radeon_modeset == 1) {
 		DRM_INFO("radeon kernel modesetting enabled.\n");
 		driver = &kms_driver;
@@ -456,9 +448,21 @@ static int __init radeon_init(void)
 		driver->driver_features |= DRIVER_MODESET;
 		driver->num_ioctls = radeon_max_kms_ioctl;
 		radeon_register_atpx_handler();
+
+	} else {
+#ifdef CONFIG_DRM_RADEON_UMS
+		DRM_INFO("radeon userspace modesetting enabled.\n");
+		driver = &driver_old;
+		pdriver = &radeon_pci_driver;
+		driver->driver_features &= ~DRIVER_MODESET;
+		driver->num_ioctls = radeon_max_ioctl;
+#else
+		DRM_ERROR("No UMS support in radeon module!\n");
+		return -EINVAL;
+#endif
 	}
-	/* if the vga console setting is enabled still
-	 * let modprobe override it */
+
+	/* let modprobe override vga console setting */
 	return drm_pci_init(driver, pdriver);
 }
 
diff --git a/drivers/gpu/drm/radeon/radeon_drv.h b/drivers/gpu/drm/radeon/radeon_drv.h
index e7fdf16..b369d42 100644
--- a/drivers/gpu/drm/radeon/radeon_drv.h
+++ b/drivers/gpu/drm/radeon/radeon_drv.h
@@ -113,6 +113,9 @@
 #define DRIVER_MINOR		33
 #define DRIVER_PATCHLEVEL	0
 
+/* The rest of the file is DEPRECATED! */
+#ifdef CONFIG_DRM_RADEON_UMS
+
 enum radeon_cp_microcode_version {
 	UCODE_R100,
 	UCODE_R200,
@@ -418,8 +421,6 @@ extern int radeon_driver_open(struct drm_device *dev,
 			      struct drm_file *file_priv);
 extern long radeon_compat_ioctl(struct file *filp, unsigned int cmd,
 				unsigned long arg);
-extern long radeon_kms_compat_ioctl(struct file *filp, unsigned int cmd,
-				    unsigned long arg);
 
 extern int radeon_master_create(struct drm_device *dev, struct drm_master *master);
 extern void radeon_master_destroy(struct drm_device *dev, struct drm_master *master);
@@ -462,15 +463,6 @@ extern void r600_blit_swap(struct drm_device *dev,
 			   int sx, int sy, int dx, int dy,
 			   int w, int h, int src_pitch, int dst_pitch, int cpp);
 
-/* atpx handler */
-#if defined(CONFIG_VGA_SWITCHEROO)
-void radeon_register_atpx_handler(void);
-void radeon_unregister_atpx_handler(void);
-#else
-static inline void radeon_register_atpx_handler(void) {}
-static inline void radeon_unregister_atpx_handler(void) {}
-#endif
-
 /* Flags for stats.boxes
  */
 #define RADEON_BOX_DMA_IDLE      0x1
@@ -2167,4 +2159,6 @@ extern void radeon_commit_ring(drm_radeon_private_t *dev_priv);
 } while (0)
 
 
+#endif				/* CONFIG_DRM_RADEON_UMS */
+
 #endif				/* __RADEON_DRV_H__ */
diff --git a/drivers/gpu/drm/radeon/radeon_irq.c b/drivers/gpu/drm/radeon/radeon_irq.c
index e771033..8d68e97 100644
--- a/drivers/gpu/drm/radeon/radeon_irq.c
+++ b/drivers/gpu/drm/radeon/radeon_irq.c
@@ -28,6 +28,8 @@
  * Authors:
  *    Keith Whitwell <keith@tungstengraphics.com>
  *    Michel D�zer <michel@daenzer.net>
+ *
+ * ------------------------ This file is DEPRECATED! -------------------------
  */
 
 #include <drm/drmP.h>
diff --git a/drivers/gpu/drm/radeon/radeon_mem.c b/drivers/gpu/drm/radeon/radeon_mem.c
index b9f0672..d54d2d7 100644
--- a/drivers/gpu/drm/radeon/radeon_mem.c
+++ b/drivers/gpu/drm/radeon/radeon_mem.c
@@ -27,6 +27,8 @@
  *
  * Authors:
  *    Keith Whitwell <keith@tungstengraphics.com>
+ *
+ * ------------------------ This file is DEPRECATED! -------------------------
  */
 
 #include <drm/drmP.h>
diff --git a/drivers/gpu/drm/radeon/radeon_state.c b/drivers/gpu/drm/radeon/radeon_state.c
index 8e9057b..4d20910 100644
--- a/drivers/gpu/drm/radeon/radeon_state.c
+++ b/drivers/gpu/drm/radeon/radeon_state.c
@@ -25,6 +25,8 @@
  * Authors:
  *    Gareth Hughes <gareth@valinux.com>
  *    Kevin E. Martin <martin@valinux.com>
+ *
+ * ------------------------ This file is DEPRECATED! -------------------------
  */
 
 #include <drm/drmP.h>
-- 
cgit v0.10.2


From 9ff0744c6d7745c331a484a937160f4c2c056923 Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Fri, 18 Jan 2013 12:18:17 -0500
Subject: drm/radeon: add additional reset flags

This adds further flags for fine grained reset.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index fa17ce0..ec3b739f 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -136,6 +136,15 @@ extern int radeon_lockup_timeout;
 #define RADEON_RESET_GFX			(1 << 0)
 #define RADEON_RESET_COMPUTE			(1 << 1)
 #define RADEON_RESET_DMA			(1 << 2)
+#define RADEON_RESET_CP				(1 << 3)
+#define RADEON_RESET_GRBM			(1 << 4)
+#define RADEON_RESET_DMA1			(1 << 5)
+#define RADEON_RESET_RLC			(1 << 6)
+#define RADEON_RESET_SEM			(1 << 7)
+#define RADEON_RESET_IH				(1 << 8)
+#define RADEON_RESET_VMC			(1 << 9)
+#define RADEON_RESET_MC				(1 << 10)
+#define RADEON_RESET_DISPLAY			(1 << 11)
 
 /*
  * Errata workarounds.
-- 
cgit v0.10.2


From 410a3418a88cc1273a281d347687f736fc39dd86 Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Fri, 18 Jan 2013 13:05:39 -0500
Subject: drm/radeon: add a bios scratch asic hung helper

Used by all asic families from r600+.
Flag for the vbios and later instances of the driver
that the GPU is hung.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c
index a1e12bf2..dd1d188 100644
--- a/drivers/gpu/drm/radeon/evergreen.c
+++ b/drivers/gpu/drm/radeon/evergreen.c
@@ -2433,6 +2433,8 @@ static int evergreen_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 
 	dev_info(rdev->dev, "GPU softreset: 0x%08X\n", reset_mask);
 
+	r600_set_bios_scratch_engine_hung(rdev, true);
+
 	evergreen_mc_stop(rdev, &save);
 	if (evergreen_mc_wait_for_idle(rdev)) {
 		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
@@ -2448,6 +2450,9 @@ static int evergreen_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 	udelay(50);
 
 	evergreen_mc_resume(rdev, &save);
+
+	r600_set_bios_scratch_engine_hung(rdev, false);
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c
index 835992d..af9e86d 100644
--- a/drivers/gpu/drm/radeon/ni.c
+++ b/drivers/gpu/drm/radeon/ni.c
@@ -1422,6 +1422,8 @@ static int cayman_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 	if (reset_mask == 0)
 		return 0;
 
+	r600_set_bios_scratch_engine_hung(rdev, true);
+
 	dev_info(rdev->dev, "GPU softreset: 0x%08X\n", reset_mask);
 
 	dev_info(rdev->dev, "  VM_CONTEXT0_PROTECTION_FAULT_ADDR   0x%08X\n",
@@ -1448,6 +1450,9 @@ static int cayman_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 	udelay(50);
 
 	evergreen_mc_resume(rdev, &save);
+
+	r600_set_bios_scratch_engine_hung(rdev, false);
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c
index becb03e..a620817 100644
--- a/drivers/gpu/drm/radeon/r600.c
+++ b/drivers/gpu/drm/radeon/r600.c
@@ -1254,6 +1254,18 @@ void r600_vram_scratch_fini(struct radeon_device *rdev)
 	radeon_bo_unref(&rdev->vram_scratch.robj);
 }
 
+void r600_set_bios_scratch_engine_hung(struct radeon_device *rdev, bool hung)
+{
+	u32 tmp = RREG32(R600_BIOS_3_SCRATCH);
+
+	if (hung)
+		tmp |= ATOM_S3_ASIC_GUI_ENGINE_HUNG;
+	else
+		tmp &= ~ATOM_S3_ASIC_GUI_ENGINE_HUNG;
+
+	WREG32(R600_BIOS_3_SCRATCH, tmp);
+}
+
 /* We doesn't check that the GPU really needs a reset we simply do the
  * reset, it's up to the caller to determine if the GPU needs one. We
  * might add an helper function to check that.
@@ -1389,6 +1401,8 @@ static int r600_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 
 	dev_info(rdev->dev, "GPU softreset: 0x%08X\n", reset_mask);
 
+	r600_set_bios_scratch_engine_hung(rdev, true);
+
 	rv515_mc_stop(rdev, &save);
 	if (r600_mc_wait_for_idle(rdev)) {
 		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
@@ -1404,6 +1418,9 @@ static int r600_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 	mdelay(1);
 
 	rv515_mc_resume(rdev, &save);
+
+	r600_set_bios_scratch_engine_hung(rdev, false);
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index ec3b739f..6539d6c 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -1860,6 +1860,7 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v);
 /* Common functions */
 /* AGP */
 extern int radeon_gpu_reset(struct radeon_device *rdev);
+extern void r600_set_bios_scratch_engine_hung(struct radeon_device *rdev, bool hung);
 extern void radeon_agp_disable(struct radeon_device *rdev);
 extern int radeon_modeset_init(struct radeon_device *rdev);
 extern void radeon_modeset_fini(struct radeon_device *rdev);
diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
index f06072f..1fc0e55 100644
--- a/drivers/gpu/drm/radeon/si.c
+++ b/drivers/gpu/drm/radeon/si.c
@@ -2231,6 +2231,8 @@ static int si_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 	dev_info(rdev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n",
 		 RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS));
 
+	r600_set_bios_scratch_engine_hung(rdev, true);
+
 	evergreen_mc_stop(rdev, &save);
 	if (radeon_mc_wait_for_idle(rdev)) {
 		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
@@ -2246,6 +2248,9 @@ static int si_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 	udelay(50);
 
 	evergreen_mc_resume(rdev, &save);
+
+	r600_set_bios_scratch_engine_hung(rdev, false);
+
 	return 0;
 }
 
-- 
cgit v0.10.2


From d3cb781e83b39561f717358e95d357e53a0da720 Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Fri, 18 Jan 2013 13:53:37 -0500
Subject: drm/radeon: rework GPU reset on r6xx/r7xx

Update the code to better match the recommended
programming sequence for soft reset.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c
index a620817..083eeb0 100644
--- a/drivers/gpu/drm/radeon/r600.c
+++ b/drivers/gpu/drm/radeon/r600.c
@@ -1266,122 +1266,22 @@ void r600_set_bios_scratch_engine_hung(struct radeon_device *rdev, bool hung)
 	WREG32(R600_BIOS_3_SCRATCH, tmp);
 }
 
-/* We doesn't check that the GPU really needs a reset we simply do the
- * reset, it's up to the caller to determine if the GPU needs one. We
- * might add an helper function to check that.
- */
-static void r600_gpu_soft_reset_gfx(struct radeon_device *rdev)
+static void r600_print_gpu_status_regs(struct radeon_device *rdev)
 {
-	u32 grbm_busy_mask = S_008010_VC_BUSY(1) | S_008010_VGT_BUSY_NO_DMA(1) |
-				S_008010_VGT_BUSY(1) | S_008010_TA03_BUSY(1) |
-				S_008010_TC_BUSY(1) | S_008010_SX_BUSY(1) |
-				S_008010_SH_BUSY(1) | S_008010_SPI03_BUSY(1) |
-				S_008010_SMX_BUSY(1) | S_008010_SC_BUSY(1) |
-				S_008010_PA_BUSY(1) | S_008010_DB03_BUSY(1) |
-				S_008010_CR_BUSY(1) | S_008010_CB03_BUSY(1) |
-				S_008010_GUI_ACTIVE(1);
-	u32 grbm2_busy_mask = S_008014_SPI0_BUSY(1) | S_008014_SPI1_BUSY(1) |
-			S_008014_SPI2_BUSY(1) | S_008014_SPI3_BUSY(1) |
-			S_008014_TA0_BUSY(1) | S_008014_TA1_BUSY(1) |
-			S_008014_TA2_BUSY(1) | S_008014_TA3_BUSY(1) |
-			S_008014_DB0_BUSY(1) | S_008014_DB1_BUSY(1) |
-			S_008014_DB2_BUSY(1) | S_008014_DB3_BUSY(1) |
-			S_008014_CB0_BUSY(1) | S_008014_CB1_BUSY(1) |
-			S_008014_CB2_BUSY(1) | S_008014_CB3_BUSY(1);
-	u32 tmp;
-
-	if (!(RREG32(GRBM_STATUS) & GUI_ACTIVE))
-		return;
-
-	dev_info(rdev->dev, "  R_008010_GRBM_STATUS      = 0x%08X\n",
-		RREG32(R_008010_GRBM_STATUS));
-	dev_info(rdev->dev, "  R_008014_GRBM_STATUS2     = 0x%08X\n",
-		RREG32(R_008014_GRBM_STATUS2));
-	dev_info(rdev->dev, "  R_000E50_SRBM_STATUS      = 0x%08X\n",
-		RREG32(R_000E50_SRBM_STATUS));
-	dev_info(rdev->dev, "  R_008674_CP_STALLED_STAT1 = 0x%08X\n",
-		RREG32(CP_STALLED_STAT1));
-	dev_info(rdev->dev, "  R_008678_CP_STALLED_STAT2 = 0x%08X\n",
-		RREG32(CP_STALLED_STAT2));
-	dev_info(rdev->dev, "  R_00867C_CP_BUSY_STAT     = 0x%08X\n",
-		RREG32(CP_BUSY_STAT));
-	dev_info(rdev->dev, "  R_008680_CP_STAT          = 0x%08X\n",
-		RREG32(CP_STAT));
-
-	/* Disable CP parsing/prefetching */
-	WREG32(R_0086D8_CP_ME_CNTL, S_0086D8_CP_ME_HALT(1));
-
-	/* Check if any of the rendering block is busy and reset it */
-	if ((RREG32(R_008010_GRBM_STATUS) & grbm_busy_mask) ||
-	    (RREG32(R_008014_GRBM_STATUS2) & grbm2_busy_mask)) {
-		tmp = S_008020_SOFT_RESET_CR(1) |
-			S_008020_SOFT_RESET_DB(1) |
-			S_008020_SOFT_RESET_CB(1) |
-			S_008020_SOFT_RESET_PA(1) |
-			S_008020_SOFT_RESET_SC(1) |
-			S_008020_SOFT_RESET_SMX(1) |
-			S_008020_SOFT_RESET_SPI(1) |
-			S_008020_SOFT_RESET_SX(1) |
-			S_008020_SOFT_RESET_SH(1) |
-			S_008020_SOFT_RESET_TC(1) |
-			S_008020_SOFT_RESET_TA(1) |
-			S_008020_SOFT_RESET_VC(1) |
-			S_008020_SOFT_RESET_VGT(1);
-		dev_info(rdev->dev, "  R_008020_GRBM_SOFT_RESET=0x%08X\n", tmp);
-		WREG32(R_008020_GRBM_SOFT_RESET, tmp);
-		RREG32(R_008020_GRBM_SOFT_RESET);
-		mdelay(15);
-		WREG32(R_008020_GRBM_SOFT_RESET, 0);
-	}
-	/* Reset CP (we always reset CP) */
-	tmp = S_008020_SOFT_RESET_CP(1);
-	dev_info(rdev->dev, "R_008020_GRBM_SOFT_RESET=0x%08X\n", tmp);
-	WREG32(R_008020_GRBM_SOFT_RESET, tmp);
-	RREG32(R_008020_GRBM_SOFT_RESET);
-	mdelay(15);
-	WREG32(R_008020_GRBM_SOFT_RESET, 0);
-
 	dev_info(rdev->dev, "  R_008010_GRBM_STATUS      = 0x%08X\n",
-		RREG32(R_008010_GRBM_STATUS));
+		 RREG32(R_008010_GRBM_STATUS));
 	dev_info(rdev->dev, "  R_008014_GRBM_STATUS2     = 0x%08X\n",
-		RREG32(R_008014_GRBM_STATUS2));
+		 RREG32(R_008014_GRBM_STATUS2));
 	dev_info(rdev->dev, "  R_000E50_SRBM_STATUS      = 0x%08X\n",
-		RREG32(R_000E50_SRBM_STATUS));
+		 RREG32(R_000E50_SRBM_STATUS));
 	dev_info(rdev->dev, "  R_008674_CP_STALLED_STAT1 = 0x%08X\n",
-		RREG32(CP_STALLED_STAT1));
+		 RREG32(CP_STALLED_STAT1));
 	dev_info(rdev->dev, "  R_008678_CP_STALLED_STAT2 = 0x%08X\n",
-		RREG32(CP_STALLED_STAT2));
+		 RREG32(CP_STALLED_STAT2));
 	dev_info(rdev->dev, "  R_00867C_CP_BUSY_STAT     = 0x%08X\n",
-		RREG32(CP_BUSY_STAT));
+		 RREG32(CP_BUSY_STAT));
 	dev_info(rdev->dev, "  R_008680_CP_STAT          = 0x%08X\n",
-		RREG32(CP_STAT));
-
-}
-
-static void r600_gpu_soft_reset_dma(struct radeon_device *rdev)
-{
-	u32 tmp;
-
-	if (RREG32(DMA_STATUS_REG) & DMA_IDLE)
-		return;
-
-	dev_info(rdev->dev, "  R_00D034_DMA_STATUS_REG   = 0x%08X\n",
-		RREG32(DMA_STATUS_REG));
-
-	/* Disable DMA */
-	tmp = RREG32(DMA_RB_CNTL);
-	tmp &= ~DMA_RB_ENABLE;
-	WREG32(DMA_RB_CNTL, tmp);
-
-	/* Reset dma */
-	if (rdev->family >= CHIP_RV770)
-		WREG32(SRBM_SOFT_RESET, RV770_SOFT_RESET_DMA);
-	else
-		WREG32(SRBM_SOFT_RESET, SOFT_RESET_DMA);
-	RREG32(SRBM_SOFT_RESET);
-	udelay(50);
-	WREG32(SRBM_SOFT_RESET, 0);
-
+		 RREG32(CP_STAT));
 	dev_info(rdev->dev, "  R_00D034_DMA_STATUS_REG   = 0x%08X\n",
 		RREG32(DMA_STATUS_REG));
 }
@@ -1389,9 +1289,12 @@ static void r600_gpu_soft_reset_dma(struct radeon_device *rdev)
 static int r600_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 {
 	struct rv515_mc_save save;
+	u32 grbm_soft_reset = 0, srbm_soft_reset = 0;
+	u32 tmp;
+	int ret = 0;
 
 	if (!(RREG32(GRBM_STATUS) & GUI_ACTIVE))
-		reset_mask &= ~(RADEON_RESET_GFX | RADEON_RESET_COMPUTE);
+		reset_mask &= ~(RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP);
 
 	if (RREG32(DMA_STATUS_REG) & DMA_IDLE)
 		reset_mask &= ~RADEON_RESET_DMA;
@@ -1401,6 +1304,8 @@ static int r600_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 
 	dev_info(rdev->dev, "GPU softreset: 0x%08X\n", reset_mask);
 
+	r600_print_gpu_status_regs(rdev);
+
 	r600_set_bios_scratch_engine_hung(rdev, true);
 
 	rv515_mc_stop(rdev, &save);
@@ -1408,20 +1313,127 @@ static int r600_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
 	}
 
-	if (reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE))
-		r600_gpu_soft_reset_gfx(rdev);
+	/* Disable CP parsing/prefetching */
+	if (rdev->family >= CHIP_RV770)
+		WREG32(R_0086D8_CP_ME_CNTL, S_0086D8_CP_ME_HALT(1) | S_0086D8_CP_PFP_HALT(1));
+	else
+		WREG32(R_0086D8_CP_ME_CNTL, S_0086D8_CP_ME_HALT(1));
+
+	/* disable the RLC */
+	WREG32(RLC_CNTL, 0);
+
+	if (reset_mask & RADEON_RESET_DMA) {
+		/* Disable DMA */
+		tmp = RREG32(DMA_RB_CNTL);
+		tmp &= ~DMA_RB_ENABLE;
+		WREG32(DMA_RB_CNTL, tmp);
+	}
+
+	mdelay(50);
+
+	if (reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE)) {
+		if (rdev->family >= CHIP_RV770)
+			grbm_soft_reset |= S_008020_SOFT_RESET_DB(1) |
+				S_008020_SOFT_RESET_CB(1) |
+				S_008020_SOFT_RESET_PA(1) |
+				S_008020_SOFT_RESET_SC(1) |
+				S_008020_SOFT_RESET_SPI(1) |
+				S_008020_SOFT_RESET_SX(1) |
+				S_008020_SOFT_RESET_SH(1) |
+				S_008020_SOFT_RESET_TC(1) |
+				S_008020_SOFT_RESET_TA(1) |
+				S_008020_SOFT_RESET_VC(1) |
+				S_008020_SOFT_RESET_VGT(1);
+		else
+			grbm_soft_reset |= S_008020_SOFT_RESET_CR(1) |
+				S_008020_SOFT_RESET_DB(1) |
+				S_008020_SOFT_RESET_CB(1) |
+				S_008020_SOFT_RESET_PA(1) |
+				S_008020_SOFT_RESET_SC(1) |
+				S_008020_SOFT_RESET_SMX(1) |
+				S_008020_SOFT_RESET_SPI(1) |
+				S_008020_SOFT_RESET_SX(1) |
+				S_008020_SOFT_RESET_SH(1) |
+				S_008020_SOFT_RESET_TC(1) |
+				S_008020_SOFT_RESET_TA(1) |
+				S_008020_SOFT_RESET_VC(1) |
+				S_008020_SOFT_RESET_VGT(1);
+	}
+
+	if (reset_mask & RADEON_RESET_CP) {
+		grbm_soft_reset |= S_008020_SOFT_RESET_CP(1) |
+			S_008020_SOFT_RESET_VGT(1);
 
-	if (reset_mask & RADEON_RESET_DMA)
-		r600_gpu_soft_reset_dma(rdev);
+		srbm_soft_reset |= S_000E60_SOFT_RESET_GRBM(1);
+	}
+
+	if (reset_mask & RADEON_RESET_DMA) {
+		if (rdev->family >= CHIP_RV770)
+			srbm_soft_reset |= RV770_SOFT_RESET_DMA;
+		else
+			srbm_soft_reset |= SOFT_RESET_DMA;
+	}
+
+	if (grbm_soft_reset) {
+		tmp = RREG32(R_008020_GRBM_SOFT_RESET);
+		tmp |= grbm_soft_reset;
+		dev_info(rdev->dev, "R_008020_GRBM_SOFT_RESET=0x%08X\n", tmp);
+		WREG32(R_008020_GRBM_SOFT_RESET, tmp);
+		tmp = RREG32(R_008020_GRBM_SOFT_RESET);
+
+		udelay(50);
+
+		tmp &= ~grbm_soft_reset;
+		WREG32(R_008020_GRBM_SOFT_RESET, tmp);
+		tmp = RREG32(R_008020_GRBM_SOFT_RESET);
+	}
+
+	if (srbm_soft_reset) {
+		tmp = RREG32(SRBM_SOFT_RESET);
+		tmp |= srbm_soft_reset;
+		dev_info(rdev->dev, "SRBM_SOFT_RESET=0x%08X\n", tmp);
+		WREG32(SRBM_SOFT_RESET, tmp);
+		tmp = RREG32(SRBM_SOFT_RESET);
+
+		udelay(50);
+
+		tmp &= ~srbm_soft_reset;
+		WREG32(SRBM_SOFT_RESET, tmp);
+		tmp = RREG32(SRBM_SOFT_RESET);
+	}
 
 	/* Wait a little for things to settle down */
 	mdelay(1);
 
 	rv515_mc_resume(rdev, &save);
+	udelay(50);
 
-	r600_set_bios_scratch_engine_hung(rdev, false);
+#if 0
+	if (reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP)) {
+		if (RREG32(GRBM_STATUS) & GUI_ACTIVE)
+			ret = -EAGAIN;
+	}
 
-	return 0;
+	if (reset_mask & RADEON_RESET_DMA) {
+		if (!(RREG32(DMA_STATUS_REG) & DMA_IDLE))
+			ret = -EAGAIN;
+	}
+#endif
+
+	if (!ret)
+		r600_set_bios_scratch_engine_hung(rdev, false);
+
+	r600_print_gpu_status_regs(rdev);
+
+	return ret;
+}
+
+int r600_asic_reset(struct radeon_device *rdev)
+{
+	return r600_gpu_soft_reset(rdev, (RADEON_RESET_GFX |
+					  RADEON_RESET_COMPUTE |
+					  RADEON_RESET_DMA |
+					  RADEON_RESET_CP));
 }
 
 bool r600_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
@@ -1465,13 +1477,6 @@ bool r600_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
 	return radeon_ring_test_lockup(rdev, ring);
 }
 
-int r600_asic_reset(struct radeon_device *rdev)
-{
-	return r600_gpu_soft_reset(rdev, (RADEON_RESET_GFX |
-					  RADEON_RESET_COMPUTE |
-					  RADEON_RESET_DMA));
-}
-
 u32 r6xx_remap_render_backend(struct radeon_device *rdev,
 			      u32 tiling_pipe_num,
 			      u32 max_rb_num,
diff --git a/drivers/gpu/drm/radeon/r600d.h b/drivers/gpu/drm/radeon/r600d.h
index 62423b0..494ef76 100644
--- a/drivers/gpu/drm/radeon/r600d.h
+++ b/drivers/gpu/drm/radeon/r600d.h
@@ -182,6 +182,8 @@
 #define	CP_COHER_BASE					0x85F8
 #define	CP_DEBUG					0xC1FC
 #define	R_0086D8_CP_ME_CNTL			0x86D8
+#define		S_0086D8_CP_PFP_HALT(x)			(((x) & 1)<<26)
+#define		C_0086D8_CP_PFP_HALT(x)			((x) & 0xFBFFFFFF)
 #define		S_0086D8_CP_ME_HALT(x)			(((x) & 1)<<28)
 #define		C_0086D8_CP_ME_HALT(x)			((x) & 0xEFFFFFFF)
 #define	CP_ME_RAM_DATA					0xC160
-- 
cgit v0.10.2


From b7630473def7dad61728c68b4f23d0f2294405bc Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Fri, 18 Jan 2013 14:28:41 -0500
Subject: drm/radeon: rework GPU reset on evergreen

Update the code to better match the recommended
programming sequence for soft reset.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c
index dd1d188..903cc54 100644
--- a/drivers/gpu/drm/radeon/evergreen.c
+++ b/drivers/gpu/drm/radeon/evergreen.c
@@ -2327,54 +2327,8 @@ bool evergreen_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *rin
 	return radeon_ring_test_lockup(rdev, ring);
 }
 
-static void evergreen_gpu_soft_reset_gfx(struct radeon_device *rdev)
+static void evergreen_print_gpu_status_regs(struct radeon_device *rdev)
 {
-	u32 grbm_reset = 0;
-
-	if (!(RREG32(GRBM_STATUS) & GUI_ACTIVE))
-		return;
-
-	dev_info(rdev->dev, "  GRBM_STATUS               = 0x%08X\n",
-		RREG32(GRBM_STATUS));
-	dev_info(rdev->dev, "  GRBM_STATUS_SE0           = 0x%08X\n",
-		RREG32(GRBM_STATUS_SE0));
-	dev_info(rdev->dev, "  GRBM_STATUS_SE1           = 0x%08X\n",
-		RREG32(GRBM_STATUS_SE1));
-	dev_info(rdev->dev, "  SRBM_STATUS               = 0x%08X\n",
-		RREG32(SRBM_STATUS));
-	dev_info(rdev->dev, "  R_008674_CP_STALLED_STAT1 = 0x%08X\n",
-		RREG32(CP_STALLED_STAT1));
-	dev_info(rdev->dev, "  R_008678_CP_STALLED_STAT2 = 0x%08X\n",
-		RREG32(CP_STALLED_STAT2));
-	dev_info(rdev->dev, "  R_00867C_CP_BUSY_STAT     = 0x%08X\n",
-		RREG32(CP_BUSY_STAT));
-	dev_info(rdev->dev, "  R_008680_CP_STAT          = 0x%08X\n",
-		RREG32(CP_STAT));
-
-	/* Disable CP parsing/prefetching */
-	WREG32(CP_ME_CNTL, CP_ME_HALT | CP_PFP_HALT);
-
-	/* reset all the gfx blocks */
-	grbm_reset = (SOFT_RESET_CP |
-		      SOFT_RESET_CB |
-		      SOFT_RESET_DB |
-		      SOFT_RESET_PA |
-		      SOFT_RESET_SC |
-		      SOFT_RESET_SPI |
-		      SOFT_RESET_SH |
-		      SOFT_RESET_SX |
-		      SOFT_RESET_TC |
-		      SOFT_RESET_TA |
-		      SOFT_RESET_VC |
-		      SOFT_RESET_VGT);
-
-	dev_info(rdev->dev, "  GRBM_SOFT_RESET=0x%08X\n", grbm_reset);
-	WREG32(GRBM_SOFT_RESET, grbm_reset);
-	(void)RREG32(GRBM_SOFT_RESET);
-	udelay(50);
-	WREG32(GRBM_SOFT_RESET, 0);
-	(void)RREG32(GRBM_SOFT_RESET);
-
 	dev_info(rdev->dev, "  GRBM_STATUS               = 0x%08X\n",
 		RREG32(GRBM_STATUS));
 	dev_info(rdev->dev, "  GRBM_STATUS_SE0           = 0x%08X\n",
@@ -2391,29 +2345,6 @@ static void evergreen_gpu_soft_reset_gfx(struct radeon_device *rdev)
 		RREG32(CP_BUSY_STAT));
 	dev_info(rdev->dev, "  R_008680_CP_STAT          = 0x%08X\n",
 		RREG32(CP_STAT));
-}
-
-static void evergreen_gpu_soft_reset_dma(struct radeon_device *rdev)
-{
-	u32 tmp;
-
-	if (RREG32(DMA_STATUS_REG) & DMA_IDLE)
-		return;
-
-	dev_info(rdev->dev, "  R_00D034_DMA_STATUS_REG   = 0x%08X\n",
-		RREG32(DMA_STATUS_REG));
-
-	/* Disable DMA */
-	tmp = RREG32(DMA_RB_CNTL);
-	tmp &= ~DMA_RB_ENABLE;
-	WREG32(DMA_RB_CNTL, tmp);
-
-	/* Reset dma */
-	WREG32(SRBM_SOFT_RESET, SOFT_RESET_DMA);
-	RREG32(SRBM_SOFT_RESET);
-	udelay(50);
-	WREG32(SRBM_SOFT_RESET, 0);
-
 	dev_info(rdev->dev, "  R_00D034_DMA_STATUS_REG   = 0x%08X\n",
 		RREG32(DMA_STATUS_REG));
 }
@@ -2421,9 +2352,12 @@ static void evergreen_gpu_soft_reset_dma(struct radeon_device *rdev)
 static int evergreen_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 {
 	struct evergreen_mc_save save;
+	u32 grbm_soft_reset = 0, srbm_soft_reset = 0;
+	u32 tmp;
+	int ret = 0;
 
 	if (!(RREG32(GRBM_STATUS) & GUI_ACTIVE))
-		reset_mask &= ~(RADEON_RESET_GFX | RADEON_RESET_COMPUTE);
+		reset_mask &= ~(RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP);
 
 	if (RREG32(DMA_STATUS_REG) & DMA_IDLE)
 		reset_mask &= ~RADEON_RESET_DMA;
@@ -2433,6 +2367,8 @@ static int evergreen_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 
 	dev_info(rdev->dev, "GPU softreset: 0x%08X\n", reset_mask);
 
+	evergreen_print_gpu_status_regs(rdev);
+
 	r600_set_bios_scratch_engine_hung(rdev, true);
 
 	evergreen_mc_stop(rdev, &save);
@@ -2440,18 +2376,90 @@ static int evergreen_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
 	}
 
-	if (reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE))
-		evergreen_gpu_soft_reset_gfx(rdev);
+	/* Disable CP parsing/prefetching */
+	WREG32(CP_ME_CNTL, CP_ME_HALT | CP_PFP_HALT);
+
+	if (reset_mask & RADEON_RESET_DMA) {
+		/* Disable DMA */
+		tmp = RREG32(DMA_RB_CNTL);
+		tmp &= ~DMA_RB_ENABLE;
+		WREG32(DMA_RB_CNTL, tmp);
+	}
+
+	if (reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE)) {
+		grbm_soft_reset |= SOFT_RESET_DB |
+			SOFT_RESET_CB |
+			SOFT_RESET_PA |
+			SOFT_RESET_SC |
+			SOFT_RESET_SPI |
+			SOFT_RESET_SX |
+			SOFT_RESET_SH |
+			SOFT_RESET_TC |
+			SOFT_RESET_TA |
+			SOFT_RESET_VC |
+			SOFT_RESET_VGT;
+	}
+
+	if (reset_mask & RADEON_RESET_CP) {
+		grbm_soft_reset |= SOFT_RESET_CP |
+			SOFT_RESET_VGT;
+
+		srbm_soft_reset |= SOFT_RESET_GRBM;
+	}
 
 	if (reset_mask & RADEON_RESET_DMA)
-		evergreen_gpu_soft_reset_dma(rdev);
+		srbm_soft_reset |= SOFT_RESET_DMA;
+
+	if (grbm_soft_reset) {
+		tmp = RREG32(GRBM_SOFT_RESET);
+		tmp |= grbm_soft_reset;
+		dev_info(rdev->dev, "GRBM_SOFT_RESET=0x%08X\n", tmp);
+		WREG32(GRBM_SOFT_RESET, tmp);
+		tmp = RREG32(GRBM_SOFT_RESET);
+
+		udelay(50);
+
+		tmp &= ~grbm_soft_reset;
+		WREG32(GRBM_SOFT_RESET, tmp);
+		tmp = RREG32(GRBM_SOFT_RESET);
+	}
+
+	if (srbm_soft_reset) {
+		tmp = RREG32(SRBM_SOFT_RESET);
+		tmp |= srbm_soft_reset;
+		dev_info(rdev->dev, "SRBM_SOFT_RESET=0x%08X\n", tmp);
+		WREG32(SRBM_SOFT_RESET, tmp);
+		tmp = RREG32(SRBM_SOFT_RESET);
+
+		udelay(50);
+
+		tmp &= ~srbm_soft_reset;
+		WREG32(SRBM_SOFT_RESET, tmp);
+		tmp = RREG32(SRBM_SOFT_RESET);
+	}
 
 	/* Wait a little for things to settle down */
 	udelay(50);
 
 	evergreen_mc_resume(rdev, &save);
+	udelay(50);
+
+#if 0
+	if (reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP)) {
+		if (RREG32(GRBM_STATUS) & GUI_ACTIVE)
+			ret = -EAGAIN;
+	}
+
+	if (reset_mask & RADEON_RESET_DMA) {
+		if (!(RREG32(DMA_STATUS_REG) & DMA_IDLE))
+			ret = -EAGAIN;
+	}
+#endif
+
+	if (!ret)
+		r600_set_bios_scratch_engine_hung(rdev, false);
 
-	r600_set_bios_scratch_engine_hung(rdev, false);
+	evergreen_print_gpu_status_regs(rdev);
 
 	return 0;
 }
@@ -2460,7 +2468,8 @@ int evergreen_asic_reset(struct radeon_device *rdev)
 {
 	return evergreen_gpu_soft_reset(rdev, (RADEON_RESET_GFX |
 					       RADEON_RESET_COMPUTE |
-					       RADEON_RESET_DMA));
+					       RADEON_RESET_DMA |
+						RADEON_RESET_CP));
 }
 
 /* Interrupts */
-- 
cgit v0.10.2


From 187e359311ac101c3a0c5078e161d64789ef16e0 Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Fri, 18 Jan 2013 14:51:38 -0500
Subject: drm/radeon: rework GPU reset on cayman/TN

Update the code to better match the recommended
programming sequence for soft reset.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c
index 903cc54..48967a9 100644
--- a/drivers/gpu/drm/radeon/evergreen.c
+++ b/drivers/gpu/drm/radeon/evergreen.c
@@ -2327,7 +2327,7 @@ bool evergreen_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *rin
 	return radeon_ring_test_lockup(rdev, ring);
 }
 
-static void evergreen_print_gpu_status_regs(struct radeon_device *rdev)
+void evergreen_print_gpu_status_regs(struct radeon_device *rdev)
 {
 	dev_info(rdev->dev, "  GRBM_STATUS               = 0x%08X\n",
 		RREG32(GRBM_STATUS));
diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c
index af9e86d..88f6fbb 100644
--- a/drivers/gpu/drm/radeon/ni.c
+++ b/drivers/gpu/drm/radeon/ni.c
@@ -34,6 +34,7 @@
 #include "ni_reg.h"
 #include "cayman_blit_shaders.h"
 
+extern void evergreen_print_gpu_status_regs(struct radeon_device *rdev);
 extern void evergreen_mc_stop(struct radeon_device *rdev, struct evergreen_mc_save *save);
 extern void evergreen_mc_resume(struct radeon_device *rdev, struct evergreen_mc_save *save);
 extern int evergreen_mc_wait_for_idle(struct radeon_device *rdev);
@@ -1310,111 +1311,15 @@ void cayman_dma_fini(struct radeon_device *rdev)
 	radeon_ring_fini(rdev, &rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX]);
 }
 
-static void cayman_gpu_soft_reset_gfx(struct radeon_device *rdev)
-{
-	u32 grbm_reset = 0;
-
-	if (!(RREG32(GRBM_STATUS) & GUI_ACTIVE))
-		return;
-
-	dev_info(rdev->dev, "  GRBM_STATUS               = 0x%08X\n",
-		RREG32(GRBM_STATUS));
-	dev_info(rdev->dev, "  GRBM_STATUS_SE0           = 0x%08X\n",
-		RREG32(GRBM_STATUS_SE0));
-	dev_info(rdev->dev, "  GRBM_STATUS_SE1           = 0x%08X\n",
-		RREG32(GRBM_STATUS_SE1));
-	dev_info(rdev->dev, "  SRBM_STATUS               = 0x%08X\n",
-		RREG32(SRBM_STATUS));
-	dev_info(rdev->dev, "  R_008674_CP_STALLED_STAT1 = 0x%08X\n",
-		RREG32(CP_STALLED_STAT1));
-	dev_info(rdev->dev, "  R_008678_CP_STALLED_STAT2 = 0x%08X\n",
-		RREG32(CP_STALLED_STAT2));
-	dev_info(rdev->dev, "  R_00867C_CP_BUSY_STAT     = 0x%08X\n",
-		RREG32(CP_BUSY_STAT));
-	dev_info(rdev->dev, "  R_008680_CP_STAT          = 0x%08X\n",
-		RREG32(CP_STAT));
-
-	/* Disable CP parsing/prefetching */
-	WREG32(CP_ME_CNTL, CP_ME_HALT | CP_PFP_HALT);
-
-	/* reset all the gfx blocks */
-	grbm_reset = (SOFT_RESET_CP |
-		      SOFT_RESET_CB |
-		      SOFT_RESET_DB |
-		      SOFT_RESET_GDS |
-		      SOFT_RESET_PA |
-		      SOFT_RESET_SC |
-		      SOFT_RESET_SPI |
-		      SOFT_RESET_SH |
-		      SOFT_RESET_SX |
-		      SOFT_RESET_TC |
-		      SOFT_RESET_TA |
-		      SOFT_RESET_VGT |
-		      SOFT_RESET_IA);
-
-	dev_info(rdev->dev, "  GRBM_SOFT_RESET=0x%08X\n", grbm_reset);
-	WREG32(GRBM_SOFT_RESET, grbm_reset);
-	(void)RREG32(GRBM_SOFT_RESET);
-	udelay(50);
-	WREG32(GRBM_SOFT_RESET, 0);
-	(void)RREG32(GRBM_SOFT_RESET);
-
-	dev_info(rdev->dev, "  GRBM_STATUS               = 0x%08X\n",
-		RREG32(GRBM_STATUS));
-	dev_info(rdev->dev, "  GRBM_STATUS_SE0           = 0x%08X\n",
-		RREG32(GRBM_STATUS_SE0));
-	dev_info(rdev->dev, "  GRBM_STATUS_SE1           = 0x%08X\n",
-		RREG32(GRBM_STATUS_SE1));
-	dev_info(rdev->dev, "  SRBM_STATUS               = 0x%08X\n",
-		RREG32(SRBM_STATUS));
-	dev_info(rdev->dev, "  R_008674_CP_STALLED_STAT1 = 0x%08X\n",
-		RREG32(CP_STALLED_STAT1));
-	dev_info(rdev->dev, "  R_008678_CP_STALLED_STAT2 = 0x%08X\n",
-		RREG32(CP_STALLED_STAT2));
-	dev_info(rdev->dev, "  R_00867C_CP_BUSY_STAT     = 0x%08X\n",
-		RREG32(CP_BUSY_STAT));
-	dev_info(rdev->dev, "  R_008680_CP_STAT          = 0x%08X\n",
-		RREG32(CP_STAT));
-
-}
-
-static void cayman_gpu_soft_reset_dma(struct radeon_device *rdev)
-{
-	u32 tmp;
-
-	if (RREG32(DMA_STATUS_REG) & DMA_IDLE)
-		return;
-
-	dev_info(rdev->dev, "  R_00D034_DMA_STATUS_REG   = 0x%08X\n",
-		RREG32(DMA_STATUS_REG));
-
-	/* dma0 */
-	tmp = RREG32(DMA_RB_CNTL + DMA0_REGISTER_OFFSET);
-	tmp &= ~DMA_RB_ENABLE;
-	WREG32(DMA_RB_CNTL + DMA0_REGISTER_OFFSET, tmp);
-
-	/* dma1 */
-	tmp = RREG32(DMA_RB_CNTL + DMA1_REGISTER_OFFSET);
-	tmp &= ~DMA_RB_ENABLE;
-	WREG32(DMA_RB_CNTL + DMA1_REGISTER_OFFSET, tmp);
-
-	/* Reset dma */
-	WREG32(SRBM_SOFT_RESET, SOFT_RESET_DMA | SOFT_RESET_DMA1);
-	RREG32(SRBM_SOFT_RESET);
-	udelay(50);
-	WREG32(SRBM_SOFT_RESET, 0);
-
-	dev_info(rdev->dev, "  R_00D034_DMA_STATUS_REG   = 0x%08X\n",
-		RREG32(DMA_STATUS_REG));
-
-}
-
 static int cayman_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 {
 	struct evergreen_mc_save save;
+	u32 grbm_soft_reset = 0, srbm_soft_reset = 0;
+	u32 tmp;
+	int ret = 0;
 
 	if (!(RREG32(GRBM_STATUS) & GUI_ACTIVE))
-		reset_mask &= ~(RADEON_RESET_GFX | RADEON_RESET_COMPUTE);
+		reset_mask &= ~(RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP);
 
 	if (RREG32(DMA_STATUS_REG) & DMA_IDLE)
 		reset_mask &= ~RADEON_RESET_DMA;
@@ -1422,10 +1327,9 @@ static int cayman_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 	if (reset_mask == 0)
 		return 0;
 
-	r600_set_bios_scratch_engine_hung(rdev, true);
-
 	dev_info(rdev->dev, "GPU softreset: 0x%08X\n", reset_mask);
 
+	evergreen_print_gpu_status_regs(rdev);
 	dev_info(rdev->dev, "  VM_CONTEXT0_PROTECTION_FAULT_ADDR   0x%08X\n",
 		 RREG32(0x14F8));
 	dev_info(rdev->dev, "  VM_CONTEXT0_PROTECTION_FAULT_STATUS 0x%08X\n",
@@ -1435,23 +1339,102 @@ static int cayman_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 	dev_info(rdev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n",
 		 RREG32(0x14DC));
 
+	r600_set_bios_scratch_engine_hung(rdev, true);
+
 	evergreen_mc_stop(rdev, &save);
 	if (evergreen_mc_wait_for_idle(rdev)) {
 		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
 	}
 
-	if (reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE))
-		cayman_gpu_soft_reset_gfx(rdev);
+	/* Disable CP parsing/prefetching */
+	WREG32(CP_ME_CNTL, CP_ME_HALT | CP_PFP_HALT);
+
+	if (reset_mask & RADEON_RESET_DMA) {
+		/* dma0 */
+		tmp = RREG32(DMA_RB_CNTL + DMA0_REGISTER_OFFSET);
+		tmp &= ~DMA_RB_ENABLE;
+		WREG32(DMA_RB_CNTL + DMA0_REGISTER_OFFSET, tmp);
+
+		/* dma1 */
+		tmp = RREG32(DMA_RB_CNTL + DMA1_REGISTER_OFFSET);
+		tmp &= ~DMA_RB_ENABLE;
+		WREG32(DMA_RB_CNTL + DMA1_REGISTER_OFFSET, tmp);
+	}
+
+	if (reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE)) {
+		grbm_soft_reset = SOFT_RESET_CB |
+			SOFT_RESET_DB |
+			SOFT_RESET_GDS |
+			SOFT_RESET_PA |
+			SOFT_RESET_SC |
+			SOFT_RESET_SPI |
+			SOFT_RESET_SH |
+			SOFT_RESET_SX |
+			SOFT_RESET_TC |
+			SOFT_RESET_TA |
+			SOFT_RESET_VGT |
+			SOFT_RESET_IA;
+	}
+
+	if (reset_mask & RADEON_RESET_CP) {
+		grbm_soft_reset |= SOFT_RESET_CP | SOFT_RESET_VGT;
+
+		srbm_soft_reset |= SOFT_RESET_GRBM;
+	}
 
 	if (reset_mask & RADEON_RESET_DMA)
-		cayman_gpu_soft_reset_dma(rdev);
+		srbm_soft_reset |= SOFT_RESET_DMA | SOFT_RESET_DMA1;
+
+	if (grbm_soft_reset) {
+		tmp = RREG32(GRBM_SOFT_RESET);
+		tmp |= grbm_soft_reset;
+		dev_info(rdev->dev, "GRBM_SOFT_RESET=0x%08X\n", tmp);
+		WREG32(GRBM_SOFT_RESET, tmp);
+		tmp = RREG32(GRBM_SOFT_RESET);
+
+		udelay(50);
+
+		tmp &= ~grbm_soft_reset;
+		WREG32(GRBM_SOFT_RESET, tmp);
+		tmp = RREG32(GRBM_SOFT_RESET);
+	}
+
+	if (srbm_soft_reset) {
+		tmp = RREG32(SRBM_SOFT_RESET);
+		tmp |= srbm_soft_reset;
+		dev_info(rdev->dev, "SRBM_SOFT_RESET=0x%08X\n", tmp);
+		WREG32(SRBM_SOFT_RESET, tmp);
+		tmp = RREG32(SRBM_SOFT_RESET);
+
+		udelay(50);
+
+		tmp &= ~srbm_soft_reset;
+		WREG32(SRBM_SOFT_RESET, tmp);
+		tmp = RREG32(SRBM_SOFT_RESET);
+	}
 
 	/* Wait a little for things to settle down */
 	udelay(50);
 
 	evergreen_mc_resume(rdev, &save);
+	udelay(50);
+
+#if 0
+	if (reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP)) {
+		if (RREG32(GRBM_STATUS) & GUI_ACTIVE)
+			ret = -EAGAIN;
+	}
+
+	if (reset_mask & RADEON_RESET_DMA) {
+		if (!(RREG32(DMA_STATUS_REG) & DMA_IDLE))
+			ret = -EAGAIN;
+	}
+#endif
+
+	if (!ret)
+		r600_set_bios_scratch_engine_hung(rdev, false);
 
-	r600_set_bios_scratch_engine_hung(rdev, false);
+	evergreen_print_gpu_status_regs(rdev);
 
 	return 0;
 }
@@ -1460,7 +1443,8 @@ int cayman_asic_reset(struct radeon_device *rdev)
 {
 	return cayman_gpu_soft_reset(rdev, (RADEON_RESET_GFX |
 					    RADEON_RESET_COMPUTE |
-					    RADEON_RESET_DMA));
+					    RADEON_RESET_DMA |
+					    RADEON_RESET_CP));
 }
 
 /**
-- 
cgit v0.10.2


From 1c53467144c3af58f5dbccb13ff48466c2de0496 Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Fri, 18 Jan 2013 15:08:38 -0500
Subject: drm/radeon: rework GPU reset on cayman/TN

Update the code to better match the recommended
programming sequence for soft reset.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
index 1fc0e55..6484155 100644
--- a/drivers/gpu/drm/radeon/si.c
+++ b/drivers/gpu/drm/radeon/si.c
@@ -61,6 +61,7 @@ extern void evergreen_fix_pci_max_read_req_size(struct radeon_device *rdev);
 extern void evergreen_mc_stop(struct radeon_device *rdev, struct evergreen_mc_save *save);
 extern void evergreen_mc_resume(struct radeon_device *rdev, struct evergreen_mc_save *save);
 extern u32 evergreen_get_number_of_dram_channels(struct radeon_device *rdev);
+extern void evergreen_print_gpu_status_regs(struct radeon_device *rdev);
 
 /* get temperature in millidegrees */
 int si_get_temp(struct radeon_device *rdev)
@@ -2126,97 +2127,15 @@ bool si_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
 	return radeon_ring_test_lockup(rdev, ring);
 }
 
-static void si_gpu_soft_reset_gfx(struct radeon_device *rdev)
-{
-	u32 grbm_reset = 0;
-
-	if (!(RREG32(GRBM_STATUS) & GUI_ACTIVE))
-		return;
-
-	dev_info(rdev->dev, "  GRBM_STATUS=0x%08X\n",
-		RREG32(GRBM_STATUS));
-	dev_info(rdev->dev, "  GRBM_STATUS2=0x%08X\n",
-		RREG32(GRBM_STATUS2));
-	dev_info(rdev->dev, "  GRBM_STATUS_SE0=0x%08X\n",
-		RREG32(GRBM_STATUS_SE0));
-	dev_info(rdev->dev, "  GRBM_STATUS_SE1=0x%08X\n",
-		RREG32(GRBM_STATUS_SE1));
-	dev_info(rdev->dev, "  SRBM_STATUS=0x%08X\n",
-		RREG32(SRBM_STATUS));
-
-	/* Disable CP parsing/prefetching */
-	WREG32(CP_ME_CNTL, CP_ME_HALT | CP_PFP_HALT | CP_CE_HALT);
-
-	/* reset all the gfx blocks */
-	grbm_reset = (SOFT_RESET_CP |
-		      SOFT_RESET_CB |
-		      SOFT_RESET_DB |
-		      SOFT_RESET_GDS |
-		      SOFT_RESET_PA |
-		      SOFT_RESET_SC |
-		      SOFT_RESET_BCI |
-		      SOFT_RESET_SPI |
-		      SOFT_RESET_SX |
-		      SOFT_RESET_TC |
-		      SOFT_RESET_TA |
-		      SOFT_RESET_VGT |
-		      SOFT_RESET_IA);
-
-	dev_info(rdev->dev, "  GRBM_SOFT_RESET=0x%08X\n", grbm_reset);
-	WREG32(GRBM_SOFT_RESET, grbm_reset);
-	(void)RREG32(GRBM_SOFT_RESET);
-	udelay(50);
-	WREG32(GRBM_SOFT_RESET, 0);
-	(void)RREG32(GRBM_SOFT_RESET);
-
-	dev_info(rdev->dev, "  GRBM_STATUS=0x%08X\n",
-		RREG32(GRBM_STATUS));
-	dev_info(rdev->dev, "  GRBM_STATUS2=0x%08X\n",
-		RREG32(GRBM_STATUS2));
-	dev_info(rdev->dev, "  GRBM_STATUS_SE0=0x%08X\n",
-		RREG32(GRBM_STATUS_SE0));
-	dev_info(rdev->dev, "  GRBM_STATUS_SE1=0x%08X\n",
-		RREG32(GRBM_STATUS_SE1));
-	dev_info(rdev->dev, "  SRBM_STATUS=0x%08X\n",
-		RREG32(SRBM_STATUS));
-}
-
-static void si_gpu_soft_reset_dma(struct radeon_device *rdev)
-{
-	u32 tmp;
-
-	if (RREG32(DMA_STATUS_REG) & DMA_IDLE)
-		return;
-
-	dev_info(rdev->dev, "  DMA_STATUS_REG   = 0x%08X\n",
-		RREG32(DMA_STATUS_REG));
-
-	/* dma0 */
-	tmp = RREG32(DMA_RB_CNTL + DMA0_REGISTER_OFFSET);
-	tmp &= ~DMA_RB_ENABLE;
-	WREG32(DMA_RB_CNTL + DMA0_REGISTER_OFFSET, tmp);
-
-	/* dma1 */
-	tmp = RREG32(DMA_RB_CNTL + DMA1_REGISTER_OFFSET);
-	tmp &= ~DMA_RB_ENABLE;
-	WREG32(DMA_RB_CNTL + DMA1_REGISTER_OFFSET, tmp);
-
-	/* Reset dma */
-	WREG32(SRBM_SOFT_RESET, SOFT_RESET_DMA | SOFT_RESET_DMA1);
-	RREG32(SRBM_SOFT_RESET);
-	udelay(50);
-	WREG32(SRBM_SOFT_RESET, 0);
-
-	dev_info(rdev->dev, "  DMA_STATUS_REG   = 0x%08X\n",
-		RREG32(DMA_STATUS_REG));
-}
-
 static int si_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 {
 	struct evergreen_mc_save save;
+	u32 grbm_soft_reset = 0, srbm_soft_reset = 0;
+	u32 tmp;
+	int ret = 0;
 
 	if (!(RREG32(GRBM_STATUS) & GUI_ACTIVE))
-		reset_mask &= ~(RADEON_RESET_GFX | RADEON_RESET_COMPUTE);
+		reset_mask &= ~(RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP);
 
 	if (RREG32(DMA_STATUS_REG) & DMA_IDLE)
 		reset_mask &= ~RADEON_RESET_DMA;
@@ -2226,6 +2145,7 @@ static int si_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 
 	dev_info(rdev->dev, "GPU softreset: 0x%08X\n", reset_mask);
 
+	evergreen_print_gpu_status_regs(rdev);
 	dev_info(rdev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x%08X\n",
 		 RREG32(VM_CONTEXT1_PROTECTION_FAULT_ADDR));
 	dev_info(rdev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n",
@@ -2234,22 +2154,99 @@ static int si_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 	r600_set_bios_scratch_engine_hung(rdev, true);
 
 	evergreen_mc_stop(rdev, &save);
-	if (radeon_mc_wait_for_idle(rdev)) {
+	if (evergreen_mc_wait_for_idle(rdev)) {
 		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
 	}
 
-	if (reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE))
-		si_gpu_soft_reset_gfx(rdev);
+	/* Disable CP parsing/prefetching */
+	WREG32(CP_ME_CNTL, CP_ME_HALT | CP_PFP_HALT | CP_CE_HALT);
+
+	if (reset_mask & RADEON_RESET_DMA) {
+		/* dma0 */
+		tmp = RREG32(DMA_RB_CNTL + DMA0_REGISTER_OFFSET);
+		tmp &= ~DMA_RB_ENABLE;
+		WREG32(DMA_RB_CNTL + DMA0_REGISTER_OFFSET, tmp);
+
+		/* dma1 */
+		tmp = RREG32(DMA_RB_CNTL + DMA1_REGISTER_OFFSET);
+		tmp &= ~DMA_RB_ENABLE;
+		WREG32(DMA_RB_CNTL + DMA1_REGISTER_OFFSET, tmp);
+	}
+
+	if (reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP)) {
+		grbm_soft_reset = SOFT_RESET_CB |
+			SOFT_RESET_DB |
+			SOFT_RESET_GDS |
+			SOFT_RESET_PA |
+			SOFT_RESET_SC |
+			SOFT_RESET_BCI |
+			SOFT_RESET_SPI |
+			SOFT_RESET_SX |
+			SOFT_RESET_TC |
+			SOFT_RESET_TA |
+			SOFT_RESET_VGT |
+			SOFT_RESET_IA;
+	}
+
+	if (reset_mask & RADEON_RESET_CP) {
+		grbm_soft_reset |= SOFT_RESET_CP | SOFT_RESET_VGT;
+
+		srbm_soft_reset |= SOFT_RESET_GRBM;
+	}
 
 	if (reset_mask & RADEON_RESET_DMA)
-		si_gpu_soft_reset_dma(rdev);
+		srbm_soft_reset |= SOFT_RESET_DMA | SOFT_RESET_DMA1;
+
+	if (grbm_soft_reset) {
+		tmp = RREG32(GRBM_SOFT_RESET);
+		tmp |= grbm_soft_reset;
+		dev_info(rdev->dev, "GRBM_SOFT_RESET=0x%08X\n", tmp);
+		WREG32(GRBM_SOFT_RESET, tmp);
+		tmp = RREG32(GRBM_SOFT_RESET);
+
+		udelay(50);
+
+		tmp &= ~grbm_soft_reset;
+		WREG32(GRBM_SOFT_RESET, tmp);
+		tmp = RREG32(GRBM_SOFT_RESET);
+	}
+
+	if (srbm_soft_reset) {
+		tmp = RREG32(SRBM_SOFT_RESET);
+		tmp |= srbm_soft_reset;
+		dev_info(rdev->dev, "SRBM_SOFT_RESET=0x%08X\n", tmp);
+		WREG32(SRBM_SOFT_RESET, tmp);
+		tmp = RREG32(SRBM_SOFT_RESET);
+
+		udelay(50);
+
+		tmp &= ~srbm_soft_reset;
+		WREG32(SRBM_SOFT_RESET, tmp);
+		tmp = RREG32(SRBM_SOFT_RESET);
+	}
 
 	/* Wait a little for things to settle down */
 	udelay(50);
 
 	evergreen_mc_resume(rdev, &save);
+	udelay(50);
+
+#if 0
+	if (reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP)) {
+		if (RREG32(GRBM_STATUS) & GUI_ACTIVE)
+			ret = -EAGAIN;
+	}
+
+	if (reset_mask & RADEON_RESET_DMA) {
+		if (!(RREG32(DMA_STATUS_REG) & DMA_IDLE))
+			ret = -EAGAIN;
+	}
+#endif
+
+	if (!ret)
+		r600_set_bios_scratch_engine_hung(rdev, false);
 
-	r600_set_bios_scratch_engine_hung(rdev, false);
+	evergreen_print_gpu_status_regs(rdev);
 
 	return 0;
 }
@@ -2258,7 +2255,8 @@ int si_asic_reset(struct radeon_device *rdev)
 {
 	return si_gpu_soft_reset(rdev, (RADEON_RESET_GFX |
 					RADEON_RESET_COMPUTE |
-					RADEON_RESET_DMA));
+					RADEON_RESET_DMA |
+					RADEON_RESET_CP));
 }
 
 /* MC */
-- 
cgit v0.10.2


From f13f7731a20ad2ceee8fa8a940b5a68abc733e06 Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Fri, 18 Jan 2013 18:12:22 -0500
Subject: drm/radeon: use status regs to determine what to reset (6xx/7xx)

When we attempt the reset the GPU, look at the status registers
to determine what blocks need to be reset.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/r500_reg.h b/drivers/gpu/drm/radeon/r500_reg.h
index ec576aa..c0dc8d3 100644
--- a/drivers/gpu/drm/radeon/r500_reg.h
+++ b/drivers/gpu/drm/radeon/r500_reg.h
@@ -355,6 +355,7 @@
 #       define AVIVO_D1CRTC_V_BLANK                             (1 << 0)
 #define AVIVO_D1CRTC_STATUS_POSITION                            0x60a0
 #define AVIVO_D1CRTC_FRAME_COUNT                                0x60a4
+#define AVIVO_D1CRTC_STATUS_HV_COUNT                            0x60ac
 #define AVIVO_D1CRTC_STEREO_CONTROL                             0x60c4
 
 #define AVIVO_D1MODE_MASTER_UPDATE_MODE                         0x60e4
diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c
index 083eeb0..997707b 100644
--- a/drivers/gpu/drm/radeon/r600.c
+++ b/drivers/gpu/drm/radeon/r600.c
@@ -94,6 +94,12 @@ MODULE_FIRMWARE("radeon/SUMO_me.bin");
 MODULE_FIRMWARE("radeon/SUMO2_pfp.bin");
 MODULE_FIRMWARE("radeon/SUMO2_me.bin");
 
+static const u32 crtc_offsets[2] =
+{
+	0,
+	AVIVO_D2CRTC_H_TOTAL - AVIVO_D1CRTC_H_TOTAL
+};
+
 int r600_debugfs_mc_info_init(struct radeon_device *rdev);
 
 /* r600,rv610,rv630,rv620,rv635,rv670 */
@@ -1286,28 +1292,111 @@ static void r600_print_gpu_status_regs(struct radeon_device *rdev)
 		RREG32(DMA_STATUS_REG));
 }
 
-static int r600_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
+static bool r600_is_display_hung(struct radeon_device *rdev)
 {
-	struct rv515_mc_save save;
-	u32 grbm_soft_reset = 0, srbm_soft_reset = 0;
+	u32 crtc_hung = 0;
+	u32 crtc_status[2];
+	u32 i, j, tmp;
+
+	for (i = 0; i < rdev->num_crtc; i++) {
+		if (RREG32(AVIVO_D1CRTC_CONTROL + crtc_offsets[i]) & AVIVO_CRTC_EN) {
+			crtc_status[i] = RREG32(AVIVO_D1CRTC_STATUS_HV_COUNT + crtc_offsets[i]);
+			crtc_hung |= (1 << i);
+		}
+	}
+
+	for (j = 0; j < 10; j++) {
+		for (i = 0; i < rdev->num_crtc; i++) {
+			if (crtc_hung & (1 << i)) {
+				tmp = RREG32(AVIVO_D1CRTC_STATUS_HV_COUNT + crtc_offsets[i]);
+				if (tmp != crtc_status[i])
+					crtc_hung &= ~(1 << i);
+			}
+		}
+		if (crtc_hung == 0)
+			return false;
+		udelay(100);
+	}
+
+	return true;
+}
+
+static u32 r600_gpu_check_soft_reset(struct radeon_device *rdev)
+{
+	u32 reset_mask = 0;
 	u32 tmp;
-	int ret = 0;
 
-	if (!(RREG32(GRBM_STATUS) & GUI_ACTIVE))
-		reset_mask &= ~(RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP);
+	/* GRBM_STATUS */
+	tmp = RREG32(R_008010_GRBM_STATUS);
+	if (rdev->family >= CHIP_RV770) {
+		if (G_008010_PA_BUSY(tmp) | G_008010_SC_BUSY(tmp) |
+		    G_008010_SH_BUSY(tmp) | G_008010_SX_BUSY(tmp) |
+		    G_008010_TA_BUSY(tmp) | G_008010_VGT_BUSY(tmp) |
+		    G_008010_DB03_BUSY(tmp) | G_008010_CB03_BUSY(tmp) |
+		    G_008010_SPI03_BUSY(tmp) | G_008010_VGT_BUSY_NO_DMA(tmp))
+			reset_mask |= RADEON_RESET_GFX;
+	} else {
+		if (G_008010_PA_BUSY(tmp) | G_008010_SC_BUSY(tmp) |
+		    G_008010_SH_BUSY(tmp) | G_008010_SX_BUSY(tmp) |
+		    G_008010_TA03_BUSY(tmp) | G_008010_VGT_BUSY(tmp) |
+		    G_008010_DB03_BUSY(tmp) | G_008010_CB03_BUSY(tmp) |
+		    G_008010_SPI03_BUSY(tmp) | G_008010_VGT_BUSY_NO_DMA(tmp))
+			reset_mask |= RADEON_RESET_GFX;
+	}
+
+	if (G_008010_CF_RQ_PENDING(tmp) | G_008010_PF_RQ_PENDING(tmp) |
+	    G_008010_CP_BUSY(tmp) | G_008010_CP_COHERENCY_BUSY(tmp))
+		reset_mask |= RADEON_RESET_CP;
+
+	if (G_008010_GRBM_EE_BUSY(tmp))
+		reset_mask |= RADEON_RESET_GRBM | RADEON_RESET_GFX | RADEON_RESET_CP;
+
+	/* DMA_STATUS_REG */
+	tmp = RREG32(DMA_STATUS_REG);
+	if (!(tmp & DMA_IDLE))
+		reset_mask |= RADEON_RESET_DMA;
+
+	/* SRBM_STATUS */
+	tmp = RREG32(R_000E50_SRBM_STATUS);
+	if (G_000E50_RLC_RQ_PENDING(tmp) | G_000E50_RLC_BUSY(tmp))
+		reset_mask |= RADEON_RESET_RLC;
+
+	if (G_000E50_IH_BUSY(tmp))
+		reset_mask |= RADEON_RESET_IH;
 
-	if (RREG32(DMA_STATUS_REG) & DMA_IDLE)
-		reset_mask &= ~RADEON_RESET_DMA;
+	if (G_000E50_SEM_BUSY(tmp))
+		reset_mask |= RADEON_RESET_SEM;
+
+	if (G_000E50_GRBM_RQ_PENDING(tmp))
+		reset_mask |= RADEON_RESET_GRBM;
+
+	if (G_000E50_VMC_BUSY(tmp))
+		reset_mask |= RADEON_RESET_VMC;
+
+	if (G_000E50_MCB_BUSY(tmp) | G_000E50_MCDZ_BUSY(tmp) |
+	    G_000E50_MCDY_BUSY(tmp) | G_000E50_MCDX_BUSY(tmp) |
+	    G_000E50_MCDW_BUSY(tmp))
+		reset_mask |= RADEON_RESET_MC;
+
+	if (r600_is_display_hung(rdev))
+		reset_mask |= RADEON_RESET_DISPLAY;
+
+	return reset_mask;
+}
+
+static void r600_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
+{
+	struct rv515_mc_save save;
+	u32 grbm_soft_reset = 0, srbm_soft_reset = 0;
+	u32 tmp;
 
 	if (reset_mask == 0)
-		return 0;
+		return;
 
 	dev_info(rdev->dev, "GPU softreset: 0x%08X\n", reset_mask);
 
 	r600_print_gpu_status_regs(rdev);
 
-	r600_set_bios_scratch_engine_hung(rdev, true);
-
 	rv515_mc_stop(rdev, &save);
 	if (r600_mc_wait_for_idle(rdev)) {
 		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
@@ -1374,6 +1463,24 @@ static int r600_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 			srbm_soft_reset |= SOFT_RESET_DMA;
 	}
 
+	if (reset_mask & RADEON_RESET_RLC)
+		srbm_soft_reset |= S_000E60_SOFT_RESET_RLC(1);
+
+	if (reset_mask & RADEON_RESET_SEM)
+		srbm_soft_reset |= S_000E60_SOFT_RESET_SEM(1);
+
+	if (reset_mask & RADEON_RESET_IH)
+		srbm_soft_reset |= S_000E60_SOFT_RESET_IH(1);
+
+	if (reset_mask & RADEON_RESET_GRBM)
+		srbm_soft_reset |= S_000E60_SOFT_RESET_GRBM(1);
+
+	if (reset_mask & RADEON_RESET_MC)
+		srbm_soft_reset |= S_000E60_SOFT_RESET_MC(1);
+
+	if (reset_mask & RADEON_RESET_VMC)
+		srbm_soft_reset |= S_000E60_SOFT_RESET_VMC(1);
+
 	if (grbm_soft_reset) {
 		tmp = RREG32(R_008020_GRBM_SOFT_RESET);
 		tmp |= grbm_soft_reset;
@@ -1408,32 +1515,26 @@ static int r600_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 	rv515_mc_resume(rdev, &save);
 	udelay(50);
 
-#if 0
-	if (reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP)) {
-		if (RREG32(GRBM_STATUS) & GUI_ACTIVE)
-			ret = -EAGAIN;
-	}
-
-	if (reset_mask & RADEON_RESET_DMA) {
-		if (!(RREG32(DMA_STATUS_REG) & DMA_IDLE))
-			ret = -EAGAIN;
-	}
-#endif
-
-	if (!ret)
-		r600_set_bios_scratch_engine_hung(rdev, false);
-
 	r600_print_gpu_status_regs(rdev);
-
-	return ret;
 }
 
 int r600_asic_reset(struct radeon_device *rdev)
 {
-	return r600_gpu_soft_reset(rdev, (RADEON_RESET_GFX |
-					  RADEON_RESET_COMPUTE |
-					  RADEON_RESET_DMA |
-					  RADEON_RESET_CP));
+	u32 reset_mask;
+
+	reset_mask = r600_gpu_check_soft_reset(rdev);
+
+	if (reset_mask)
+		r600_set_bios_scratch_engine_hung(rdev, true);
+
+	r600_gpu_soft_reset(rdev, reset_mask);
+
+	reset_mask = r600_gpu_check_soft_reset(rdev);
+
+	if (!reset_mask)
+		r600_set_bios_scratch_engine_hung(rdev, false);
+
+	return 0;
 }
 
 bool r600_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
diff --git a/drivers/gpu/drm/radeon/r600d.h b/drivers/gpu/drm/radeon/r600d.h
index 494ef76..a42ba11 100644
--- a/drivers/gpu/drm/radeon/r600d.h
+++ b/drivers/gpu/drm/radeon/r600d.h
@@ -1321,6 +1321,7 @@
 #define		G_008010_VC_BUSY(x)			(((x) >> 11) & 1)
 #define		G_008010_DB03_CLEAN(x)			(((x) >> 12) & 1)
 #define		G_008010_CB03_CLEAN(x)			(((x) >> 13) & 1)
+#define		G_008010_TA_BUSY(x)			(((x) >> 14) & 1)
 #define		G_008010_VGT_BUSY_NO_DMA(x)		(((x) >> 16) & 1)
 #define		G_008010_VGT_BUSY(x)			(((x) >> 17) & 1)
 #define		G_008010_TA03_BUSY(x)			(((x) >> 18) & 1)
@@ -1388,6 +1389,7 @@
 #define		G_000E50_MCDW_BUSY(x)			(((x) >> 13) & 1)
 #define		G_000E50_SEM_BUSY(x)			(((x) >> 14) & 1)
 #define		G_000E50_RLC_BUSY(x)			(((x) >> 15) & 1)
+#define		G_000E50_IH_BUSY(x)			(((x) >> 17) & 1)
 #define		G_000E50_BIF_BUSY(x)			(((x) >> 29) & 1)
 #define	R_000E60_SRBM_SOFT_RESET			0x0E60
 #define		S_000E60_SOFT_RESET_BIF(x)		(((x) & 1) << 1)
-- 
cgit v0.10.2


From a65a4369f7a69281b65ed42b7fa2a8e8d6ff585f Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Fri, 18 Jan 2013 18:55:54 -0500
Subject: drm/radeon: use status regs to determine what to reset (evergreen)

When we attempt the reset the GPU, look at the status registers
to determine what blocks need to be reset.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c
index 48967a9..8b73743 100644
--- a/drivers/gpu/drm/radeon/evergreen.c
+++ b/drivers/gpu/drm/radeon/evergreen.c
@@ -2337,6 +2337,8 @@ void evergreen_print_gpu_status_regs(struct radeon_device *rdev)
 		RREG32(GRBM_STATUS_SE1));
 	dev_info(rdev->dev, "  SRBM_STATUS               = 0x%08X\n",
 		RREG32(SRBM_STATUS));
+	dev_info(rdev->dev, "  SRBM_STATUS2              = 0x%08X\n",
+		RREG32(SRBM_STATUS2));
 	dev_info(rdev->dev, "  R_008674_CP_STALLED_STAT1 = 0x%08X\n",
 		RREG32(CP_STALLED_STAT1));
 	dev_info(rdev->dev, "  R_008678_CP_STALLED_STAT2 = 0x%08X\n",
@@ -2349,28 +2351,111 @@ void evergreen_print_gpu_status_regs(struct radeon_device *rdev)
 		RREG32(DMA_STATUS_REG));
 }
 
-static int evergreen_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
+static bool evergreen_is_display_hung(struct radeon_device *rdev)
 {
-	struct evergreen_mc_save save;
-	u32 grbm_soft_reset = 0, srbm_soft_reset = 0;
+	u32 crtc_hung = 0;
+	u32 crtc_status[6];
+	u32 i, j, tmp;
+
+	for (i = 0; i < rdev->num_crtc; i++) {
+		if (RREG32(EVERGREEN_CRTC_CONTROL + crtc_offsets[i]) & EVERGREEN_CRTC_MASTER_EN) {
+			crtc_status[i] = RREG32(EVERGREEN_CRTC_STATUS_HV_COUNT + crtc_offsets[i]);
+			crtc_hung |= (1 << i);
+		}
+	}
+
+	for (j = 0; j < 10; j++) {
+		for (i = 0; i < rdev->num_crtc; i++) {
+			if (crtc_hung & (1 << i)) {
+				tmp = RREG32(EVERGREEN_CRTC_STATUS_HV_COUNT + crtc_offsets[i]);
+				if (tmp != crtc_status[i])
+					crtc_hung &= ~(1 << i);
+			}
+		}
+		if (crtc_hung == 0)
+			return false;
+		udelay(100);
+	}
+
+	return true;
+}
+
+static u32 evergreen_gpu_check_soft_reset(struct radeon_device *rdev)
+{
+	u32 reset_mask = 0;
 	u32 tmp;
-	int ret = 0;
 
-	if (!(RREG32(GRBM_STATUS) & GUI_ACTIVE))
-		reset_mask &= ~(RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP);
+	/* GRBM_STATUS */
+	tmp = RREG32(GRBM_STATUS);
+	if (tmp & (PA_BUSY | SC_BUSY |
+		   SH_BUSY | SX_BUSY |
+		   TA_BUSY | VGT_BUSY |
+		   DB_BUSY | CB_BUSY |
+		   SPI_BUSY | VGT_BUSY_NO_DMA))
+		reset_mask |= RADEON_RESET_GFX;
+
+	if (tmp & (CF_RQ_PENDING | PF_RQ_PENDING |
+		   CP_BUSY | CP_COHERENCY_BUSY))
+		reset_mask |= RADEON_RESET_CP;
+
+	if (tmp & GRBM_EE_BUSY)
+		reset_mask |= RADEON_RESET_GRBM | RADEON_RESET_GFX | RADEON_RESET_CP;
 
-	if (RREG32(DMA_STATUS_REG) & DMA_IDLE)
-		reset_mask &= ~RADEON_RESET_DMA;
+	/* DMA_STATUS_REG */
+	tmp = RREG32(DMA_STATUS_REG);
+	if (!(tmp & DMA_IDLE))
+		reset_mask |= RADEON_RESET_DMA;
+
+	/* SRBM_STATUS2 */
+	tmp = RREG32(SRBM_STATUS2);
+	if (tmp & DMA_BUSY)
+		reset_mask |= RADEON_RESET_DMA;
+
+	/* SRBM_STATUS */
+	tmp = RREG32(SRBM_STATUS);
+	if (tmp & (RLC_RQ_PENDING | RLC_BUSY))
+		reset_mask |= RADEON_RESET_RLC;
+
+	if (tmp & IH_BUSY)
+		reset_mask |= RADEON_RESET_IH;
+
+	if (tmp & SEM_BUSY)
+		reset_mask |= RADEON_RESET_SEM;
+
+	if (tmp & GRBM_RQ_PENDING)
+		reset_mask |= RADEON_RESET_GRBM;
+
+	if (tmp & VMC_BUSY)
+		reset_mask |= RADEON_RESET_VMC;
+
+	if (tmp & (MCB_BUSY | MCB_NON_DISPLAY_BUSY |
+		   MCC_BUSY | MCD_BUSY))
+		reset_mask |= RADEON_RESET_MC;
+
+	if (evergreen_is_display_hung(rdev))
+		reset_mask |= RADEON_RESET_DISPLAY;
+
+	/* VM_L2_STATUS */
+	tmp = RREG32(VM_L2_STATUS);
+	if (tmp & L2_BUSY)
+		reset_mask |= RADEON_RESET_VMC;
+
+	return reset_mask;
+}
+
+static void evergreen_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
+{
+	struct evergreen_mc_save save;
+	u32 grbm_soft_reset = 0, srbm_soft_reset = 0;
+	u32 tmp;
 
 	if (reset_mask == 0)
-		return 0;
+		return;
 
 	dev_info(rdev->dev, "GPU softreset: 0x%08X\n", reset_mask);
 
 	evergreen_print_gpu_status_regs(rdev);
 
-	r600_set_bios_scratch_engine_hung(rdev, true);
-
 	evergreen_mc_stop(rdev, &save);
 	if (evergreen_mc_wait_for_idle(rdev)) {
 		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
@@ -2410,6 +2495,27 @@ static int evergreen_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 	if (reset_mask & RADEON_RESET_DMA)
 		srbm_soft_reset |= SOFT_RESET_DMA;
 
+	if (reset_mask & RADEON_RESET_DISPLAY)
+		srbm_soft_reset |= SOFT_RESET_DC;
+
+	if (reset_mask & RADEON_RESET_RLC)
+		srbm_soft_reset |= SOFT_RESET_RLC;
+
+	if (reset_mask & RADEON_RESET_SEM)
+		srbm_soft_reset |= SOFT_RESET_SEM;
+
+	if (reset_mask & RADEON_RESET_IH)
+		srbm_soft_reset |= SOFT_RESET_IH;
+
+	if (reset_mask & RADEON_RESET_GRBM)
+		srbm_soft_reset |= SOFT_RESET_GRBM;
+
+	if (reset_mask & RADEON_RESET_VMC)
+		srbm_soft_reset |= SOFT_RESET_VMC;
+
+	if (reset_mask & RADEON_RESET_MC)
+		srbm_soft_reset |= SOFT_RESET_MC;
+
 	if (grbm_soft_reset) {
 		tmp = RREG32(GRBM_SOFT_RESET);
 		tmp |= grbm_soft_reset;
@@ -2444,32 +2550,26 @@ static int evergreen_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 	evergreen_mc_resume(rdev, &save);
 	udelay(50);
 
-#if 0
-	if (reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP)) {
-		if (RREG32(GRBM_STATUS) & GUI_ACTIVE)
-			ret = -EAGAIN;
-	}
-
-	if (reset_mask & RADEON_RESET_DMA) {
-		if (!(RREG32(DMA_STATUS_REG) & DMA_IDLE))
-			ret = -EAGAIN;
-	}
-#endif
-
-	if (!ret)
-		r600_set_bios_scratch_engine_hung(rdev, false);
-
 	evergreen_print_gpu_status_regs(rdev);
-
-	return 0;
 }
 
 int evergreen_asic_reset(struct radeon_device *rdev)
 {
-	return evergreen_gpu_soft_reset(rdev, (RADEON_RESET_GFX |
-					       RADEON_RESET_COMPUTE |
-					       RADEON_RESET_DMA |
-						RADEON_RESET_CP));
+	u32 reset_mask;
+
+	reset_mask = evergreen_gpu_check_soft_reset(rdev);
+
+	if (reset_mask)
+		r600_set_bios_scratch_engine_hung(rdev, true);
+
+	evergreen_gpu_soft_reset(rdev, reset_mask);
+
+	reset_mask = evergreen_gpu_check_soft_reset(rdev);
+
+	if (!reset_mask)
+		r600_set_bios_scratch_engine_hung(rdev, false);
+
+	return 0;
 }
 
 /* Interrupts */
diff --git a/drivers/gpu/drm/radeon/evergreen_reg.h b/drivers/gpu/drm/radeon/evergreen_reg.h
index 034f4c2..f585be1 100644
--- a/drivers/gpu/drm/radeon/evergreen_reg.h
+++ b/drivers/gpu/drm/radeon/evergreen_reg.h
@@ -223,6 +223,7 @@
 #define EVERGREEN_CRTC_STATUS                           0x6e8c
 #       define EVERGREEN_CRTC_V_BLANK                   (1 << 0)
 #define EVERGREEN_CRTC_STATUS_POSITION                  0x6e90
+#define EVERGREEN_CRTC_STATUS_HV_COUNT                  0x6ea0
 #define EVERGREEN_MASTER_UPDATE_MODE                    0x6ef8
 #define EVERGREEN_CRTC_UPDATE_LOCK                      0x6ed4
 
diff --git a/drivers/gpu/drm/radeon/evergreend.h b/drivers/gpu/drm/radeon/evergreend.h
index 5ad29f3..982d25a 100644
--- a/drivers/gpu/drm/radeon/evergreend.h
+++ b/drivers/gpu/drm/radeon/evergreend.h
@@ -729,6 +729,18 @@
 #define	WAIT_UNTIL					0x8040
 
 #define	SRBM_STATUS				        0x0E50
+#define		RLC_RQ_PENDING 				(1 << 3)
+#define		GRBM_RQ_PENDING 			(1 << 5)
+#define		VMC_BUSY 				(1 << 8)
+#define		MCB_BUSY 				(1 << 9)
+#define		MCB_NON_DISPLAY_BUSY 			(1 << 10)
+#define		MCC_BUSY 				(1 << 11)
+#define		MCD_BUSY 				(1 << 12)
+#define		SEM_BUSY 				(1 << 14)
+#define		RLC_BUSY 				(1 << 15)
+#define		IH_BUSY 				(1 << 17)
+#define	SRBM_STATUS2				        0x0EC4
+#define		DMA_BUSY 				(1 << 5)
 #define	SRBM_SOFT_RESET				        0x0E60
 #define		SRBM_SOFT_RESET_ALL_MASK    	       	0x00FEEFA6
 #define		SOFT_RESET_BIF				(1 << 1)
-- 
cgit v0.10.2


From 168757ea854b7dd8eac3962753ed75a542ebca92 Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Fri, 18 Jan 2013 19:17:22 -0500
Subject: drm/radeon: use status regs to determine what to reset (cayman)

When we attempt the reset the GPU, look at the status registers
to determine what blocks need to be reset.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c
index 8b73743..fd5e002 100644
--- a/drivers/gpu/drm/radeon/evergreen.c
+++ b/drivers/gpu/drm/radeon/evergreen.c
@@ -2349,9 +2349,13 @@ void evergreen_print_gpu_status_regs(struct radeon_device *rdev)
 		RREG32(CP_STAT));
 	dev_info(rdev->dev, "  R_00D034_DMA_STATUS_REG   = 0x%08X\n",
 		RREG32(DMA_STATUS_REG));
+	if (rdev->family >= CHIP_CAYMAN) {
+		dev_info(rdev->dev, "  R_00D834_DMA_STATUS_REG   = 0x%08X\n",
+			 RREG32(DMA_STATUS_REG + 0x800));
+	}
 }
 
-static bool evergreen_is_display_hung(struct radeon_device *rdev)
+bool evergreen_is_display_hung(struct radeon_device *rdev)
 {
 	u32 crtc_hung = 0;
 	u32 crtc_status[6];
diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c
index 88f6fbb..eee013f 100644
--- a/drivers/gpu/drm/radeon/ni.c
+++ b/drivers/gpu/drm/radeon/ni.c
@@ -34,6 +34,7 @@
 #include "ni_reg.h"
 #include "cayman_blit_shaders.h"
 
+extern bool evergreen_is_display_hung(struct radeon_device *rdev);
 extern void evergreen_print_gpu_status_regs(struct radeon_device *rdev);
 extern void evergreen_mc_stop(struct radeon_device *rdev, struct evergreen_mc_save *save);
 extern void evergreen_mc_resume(struct radeon_device *rdev, struct evergreen_mc_save *save);
@@ -1311,21 +1312,86 @@ void cayman_dma_fini(struct radeon_device *rdev)
 	radeon_ring_fini(rdev, &rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX]);
 }
 
-static int cayman_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
+static u32 cayman_gpu_check_soft_reset(struct radeon_device *rdev)
 {
-	struct evergreen_mc_save save;
-	u32 grbm_soft_reset = 0, srbm_soft_reset = 0;
+	u32 reset_mask = 0;
 	u32 tmp;
-	int ret = 0;
 
-	if (!(RREG32(GRBM_STATUS) & GUI_ACTIVE))
-		reset_mask &= ~(RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP);
+	/* GRBM_STATUS */
+	tmp = RREG32(GRBM_STATUS);
+	if (tmp & (PA_BUSY | SC_BUSY |
+		   SH_BUSY | SX_BUSY |
+		   TA_BUSY | VGT_BUSY |
+		   DB_BUSY | CB_BUSY |
+		   GDS_BUSY | SPI_BUSY |
+		   IA_BUSY | IA_BUSY_NO_DMA))
+		reset_mask |= RADEON_RESET_GFX;
+
+	if (tmp & (CF_RQ_PENDING | PF_RQ_PENDING |
+		   CP_BUSY | CP_COHERENCY_BUSY))
+		reset_mask |= RADEON_RESET_CP;
+
+	if (tmp & GRBM_EE_BUSY)
+		reset_mask |= RADEON_RESET_GRBM | RADEON_RESET_GFX | RADEON_RESET_CP;
+
+	/* DMA_STATUS_REG 0 */
+	tmp = RREG32(DMA_STATUS_REG + DMA0_REGISTER_OFFSET);
+	if (!(tmp & DMA_IDLE))
+		reset_mask |= RADEON_RESET_DMA;
+
+	/* DMA_STATUS_REG 1 */
+	tmp = RREG32(DMA_STATUS_REG + DMA1_REGISTER_OFFSET);
+	if (!(tmp & DMA_IDLE))
+		reset_mask |= RADEON_RESET_DMA1;
+
+	/* SRBM_STATUS2 */
+	tmp = RREG32(SRBM_STATUS2);
+	if (tmp & DMA_BUSY)
+		reset_mask |= RADEON_RESET_DMA;
+
+	if (tmp & DMA1_BUSY)
+		reset_mask |= RADEON_RESET_DMA1;
+
+	/* SRBM_STATUS */
+	tmp = RREG32(SRBM_STATUS);
+	if (tmp & (RLC_RQ_PENDING | RLC_BUSY))
+		reset_mask |= RADEON_RESET_RLC;
+
+	if (tmp & IH_BUSY)
+		reset_mask |= RADEON_RESET_IH;
+
+	if (tmp & SEM_BUSY)
+		reset_mask |= RADEON_RESET_SEM;
+
+	if (tmp & GRBM_RQ_PENDING)
+		reset_mask |= RADEON_RESET_GRBM;
+
+	if (tmp & VMC_BUSY)
+		reset_mask |= RADEON_RESET_VMC;
 
-	if (RREG32(DMA_STATUS_REG) & DMA_IDLE)
-		reset_mask &= ~RADEON_RESET_DMA;
+	if (tmp & (MCB_BUSY | MCB_NON_DISPLAY_BUSY |
+		   MCC_BUSY | MCD_BUSY))
+		reset_mask |= RADEON_RESET_MC;
+
+	if (evergreen_is_display_hung(rdev))
+		reset_mask |= RADEON_RESET_DISPLAY;
+
+	/* VM_L2_STATUS */
+	tmp = RREG32(VM_L2_STATUS);
+	if (tmp & L2_BUSY)
+		reset_mask |= RADEON_RESET_VMC;
+
+	return reset_mask;
+}
+
+static void cayman_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
+{
+	struct evergreen_mc_save save;
+	u32 grbm_soft_reset = 0, srbm_soft_reset = 0;
+	u32 tmp;
 
 	if (reset_mask == 0)
-		return 0;
+		return;
 
 	dev_info(rdev->dev, "GPU softreset: 0x%08X\n", reset_mask);
 
@@ -1354,7 +1420,9 @@ static int cayman_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 		tmp = RREG32(DMA_RB_CNTL + DMA0_REGISTER_OFFSET);
 		tmp &= ~DMA_RB_ENABLE;
 		WREG32(DMA_RB_CNTL + DMA0_REGISTER_OFFSET, tmp);
+	}
 
+	if (reset_mask & RADEON_RESET_DMA1) {
 		/* dma1 */
 		tmp = RREG32(DMA_RB_CNTL + DMA1_REGISTER_OFFSET);
 		tmp &= ~DMA_RB_ENABLE;
@@ -1383,7 +1451,31 @@ static int cayman_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 	}
 
 	if (reset_mask & RADEON_RESET_DMA)
-		srbm_soft_reset |= SOFT_RESET_DMA | SOFT_RESET_DMA1;
+		srbm_soft_reset |= SOFT_RESET_DMA;
+
+	if (reset_mask & RADEON_RESET_DMA1)
+		srbm_soft_reset |= SOFT_RESET_DMA1;
+
+	if (reset_mask & RADEON_RESET_DISPLAY)
+		srbm_soft_reset |= SOFT_RESET_DC;
+
+	if (reset_mask & RADEON_RESET_RLC)
+		srbm_soft_reset |= SOFT_RESET_RLC;
+
+	if (reset_mask & RADEON_RESET_SEM)
+		srbm_soft_reset |= SOFT_RESET_SEM;
+
+	if (reset_mask & RADEON_RESET_IH)
+		srbm_soft_reset |= SOFT_RESET_IH;
+
+	if (reset_mask & RADEON_RESET_GRBM)
+		srbm_soft_reset |= SOFT_RESET_GRBM;
+
+	if (reset_mask & RADEON_RESET_VMC)
+		srbm_soft_reset |= SOFT_RESET_VMC;
+
+	if (reset_mask & RADEON_RESET_MC)
+		srbm_soft_reset |= SOFT_RESET_MC;
 
 	if (grbm_soft_reset) {
 		tmp = RREG32(GRBM_SOFT_RESET);
@@ -1419,32 +1511,26 @@ static int cayman_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 	evergreen_mc_resume(rdev, &save);
 	udelay(50);
 
-#if 0
-	if (reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP)) {
-		if (RREG32(GRBM_STATUS) & GUI_ACTIVE)
-			ret = -EAGAIN;
-	}
-
-	if (reset_mask & RADEON_RESET_DMA) {
-		if (!(RREG32(DMA_STATUS_REG) & DMA_IDLE))
-			ret = -EAGAIN;
-	}
-#endif
-
-	if (!ret)
-		r600_set_bios_scratch_engine_hung(rdev, false);
-
 	evergreen_print_gpu_status_regs(rdev);
-
-	return 0;
 }
 
 int cayman_asic_reset(struct radeon_device *rdev)
 {
-	return cayman_gpu_soft_reset(rdev, (RADEON_RESET_GFX |
-					    RADEON_RESET_COMPUTE |
-					    RADEON_RESET_DMA |
-					    RADEON_RESET_CP));
+	u32 reset_mask;
+
+	reset_mask = cayman_gpu_check_soft_reset(rdev);
+
+	if (reset_mask)
+		r600_set_bios_scratch_engine_hung(rdev, true);
+
+	cayman_gpu_soft_reset(rdev, reset_mask);
+
+	reset_mask = cayman_gpu_check_soft_reset(rdev);
+
+	if (!reset_mask)
+		r600_set_bios_scratch_engine_hung(rdev, false);
+
+	return 0;
 }
 
 /**
diff --git a/drivers/gpu/drm/radeon/nid.h b/drivers/gpu/drm/radeon/nid.h
index c4b84c5..079dee2 100644
--- a/drivers/gpu/drm/radeon/nid.h
+++ b/drivers/gpu/drm/radeon/nid.h
@@ -49,6 +49,16 @@
 #define		RINGID(x)					(((x) & 0x3) << 0)
 #define		VMID(x)						(((x) & 0x7) << 0)
 #define	SRBM_STATUS				        0x0E50
+#define		RLC_RQ_PENDING 				(1 << 3)
+#define		GRBM_RQ_PENDING 			(1 << 5)
+#define		VMC_BUSY 				(1 << 8)
+#define		MCB_BUSY 				(1 << 9)
+#define		MCB_NON_DISPLAY_BUSY 			(1 << 10)
+#define		MCC_BUSY 				(1 << 11)
+#define		MCD_BUSY 				(1 << 12)
+#define		SEM_BUSY 				(1 << 14)
+#define		RLC_BUSY 				(1 << 15)
+#define		IH_BUSY 				(1 << 17)
 
 #define	SRBM_SOFT_RESET				        0x0E60
 #define		SOFT_RESET_BIF				(1 << 1)
@@ -68,6 +78,10 @@
 #define		SOFT_RESET_REGBB			(1 << 22)
 #define		SOFT_RESET_ORB				(1 << 23)
 
+#define	SRBM_STATUS2				        0x0EC4
+#define		DMA_BUSY 				(1 << 5)
+#define		DMA1_BUSY 				(1 << 6)
+
 #define VM_CONTEXT0_REQUEST_RESPONSE			0x1470
 #define		REQUEST_TYPE(x)					(((x) & 0xf) << 0)
 #define		RESPONSE_TYPE_MASK				0x000000F0
-- 
cgit v0.10.2


From 014bb209214d8dce9698efea71c68d20ba477abc Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Fri, 18 Jan 2013 19:36:20 -0500
Subject: drm/radeon: use status regs to determine what to reset (si)

When we attempt the reset the GPU, look at the status registers
to determine what blocks need to be reset.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
index 6484155..7a8ca72 100644
--- a/drivers/gpu/drm/radeon/si.c
+++ b/drivers/gpu/drm/radeon/si.c
@@ -62,6 +62,7 @@ extern void evergreen_mc_stop(struct radeon_device *rdev, struct evergreen_mc_sa
 extern void evergreen_mc_resume(struct radeon_device *rdev, struct evergreen_mc_save *save);
 extern u32 evergreen_get_number_of_dram_channels(struct radeon_device *rdev);
 extern void evergreen_print_gpu_status_regs(struct radeon_device *rdev);
+extern bool evergreen_is_display_hung(struct radeon_device *rdev);
 
 /* get temperature in millidegrees */
 int si_get_temp(struct radeon_device *rdev)
@@ -2127,21 +2128,89 @@ bool si_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
 	return radeon_ring_test_lockup(rdev, ring);
 }
 
-static int si_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
+static u32 si_gpu_check_soft_reset(struct radeon_device *rdev)
 {
-	struct evergreen_mc_save save;
-	u32 grbm_soft_reset = 0, srbm_soft_reset = 0;
+	u32 reset_mask = 0;
 	u32 tmp;
-	int ret = 0;
 
-	if (!(RREG32(GRBM_STATUS) & GUI_ACTIVE))
-		reset_mask &= ~(RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP);
+	/* GRBM_STATUS */
+	tmp = RREG32(GRBM_STATUS);
+	if (tmp & (PA_BUSY | SC_BUSY |
+		   BCI_BUSY | SX_BUSY |
+		   TA_BUSY | VGT_BUSY |
+		   DB_BUSY | CB_BUSY |
+		   GDS_BUSY | SPI_BUSY |
+		   IA_BUSY | IA_BUSY_NO_DMA))
+		reset_mask |= RADEON_RESET_GFX;
+
+	if (tmp & (CF_RQ_PENDING | PF_RQ_PENDING |
+		   CP_BUSY | CP_COHERENCY_BUSY))
+		reset_mask |= RADEON_RESET_CP;
+
+	if (tmp & GRBM_EE_BUSY)
+		reset_mask |= RADEON_RESET_GRBM | RADEON_RESET_GFX | RADEON_RESET_CP;
+
+	/* GRBM_STATUS2 */
+	tmp = RREG32(GRBM_STATUS2);
+	if (tmp & (RLC_RQ_PENDING | RLC_BUSY))
+		reset_mask |= RADEON_RESET_RLC;
+
+	/* DMA_STATUS_REG 0 */
+	tmp = RREG32(DMA_STATUS_REG + DMA0_REGISTER_OFFSET);
+	if (!(tmp & DMA_IDLE))
+		reset_mask |= RADEON_RESET_DMA;
+
+	/* DMA_STATUS_REG 1 */
+	tmp = RREG32(DMA_STATUS_REG + DMA1_REGISTER_OFFSET);
+	if (!(tmp & DMA_IDLE))
+		reset_mask |= RADEON_RESET_DMA1;
+
+	/* SRBM_STATUS2 */
+	tmp = RREG32(SRBM_STATUS2);
+	if (tmp & DMA_BUSY)
+		reset_mask |= RADEON_RESET_DMA;
+
+	if (tmp & DMA1_BUSY)
+		reset_mask |= RADEON_RESET_DMA1;
+
+	/* SRBM_STATUS */
+	tmp = RREG32(SRBM_STATUS);
+
+	if (tmp & IH_BUSY)
+		reset_mask |= RADEON_RESET_IH;
+
+	if (tmp & SEM_BUSY)
+		reset_mask |= RADEON_RESET_SEM;
+
+	if (tmp & GRBM_RQ_PENDING)
+		reset_mask |= RADEON_RESET_GRBM;
+
+	if (tmp & VMC_BUSY)
+		reset_mask |= RADEON_RESET_VMC;
 
-	if (RREG32(DMA_STATUS_REG) & DMA_IDLE)
-		reset_mask &= ~RADEON_RESET_DMA;
+	if (tmp & (MCB_BUSY | MCB_NON_DISPLAY_BUSY |
+		   MCC_BUSY | MCD_BUSY))
+		reset_mask |= RADEON_RESET_MC;
+
+	if (evergreen_is_display_hung(rdev))
+		reset_mask |= RADEON_RESET_DISPLAY;
+
+	/* VM_L2_STATUS */
+	tmp = RREG32(VM_L2_STATUS);
+	if (tmp & L2_BUSY)
+		reset_mask |= RADEON_RESET_VMC;
+
+	return reset_mask;
+}
+
+static void si_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
+{
+	struct evergreen_mc_save save;
+	u32 grbm_soft_reset = 0, srbm_soft_reset = 0;
+	u32 tmp;
 
 	if (reset_mask == 0)
-		return 0;
+		return;
 
 	dev_info(rdev->dev, "GPU softreset: 0x%08X\n", reset_mask);
 
@@ -2151,8 +2220,6 @@ static int si_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 	dev_info(rdev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n",
 		 RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS));
 
-	r600_set_bios_scratch_engine_hung(rdev, true);
-
 	evergreen_mc_stop(rdev, &save);
 	if (evergreen_mc_wait_for_idle(rdev)) {
 		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
@@ -2166,7 +2233,8 @@ static int si_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 		tmp = RREG32(DMA_RB_CNTL + DMA0_REGISTER_OFFSET);
 		tmp &= ~DMA_RB_ENABLE;
 		WREG32(DMA_RB_CNTL + DMA0_REGISTER_OFFSET, tmp);
-
+	}
+	if (reset_mask & RADEON_RESET_DMA1) {
 		/* dma1 */
 		tmp = RREG32(DMA_RB_CNTL + DMA1_REGISTER_OFFSET);
 		tmp &= ~DMA_RB_ENABLE;
@@ -2195,7 +2263,31 @@ static int si_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 	}
 
 	if (reset_mask & RADEON_RESET_DMA)
-		srbm_soft_reset |= SOFT_RESET_DMA | SOFT_RESET_DMA1;
+		srbm_soft_reset |= SOFT_RESET_DMA;
+
+	if (reset_mask & RADEON_RESET_DMA1)
+		srbm_soft_reset |= SOFT_RESET_DMA1;
+
+	if (reset_mask & RADEON_RESET_DISPLAY)
+		srbm_soft_reset |= SOFT_RESET_DC;
+
+	if (reset_mask & RADEON_RESET_RLC)
+		grbm_soft_reset |= SOFT_RESET_RLC;
+
+	if (reset_mask & RADEON_RESET_SEM)
+		srbm_soft_reset |= SOFT_RESET_SEM;
+
+	if (reset_mask & RADEON_RESET_IH)
+		srbm_soft_reset |= SOFT_RESET_IH;
+
+	if (reset_mask & RADEON_RESET_GRBM)
+		srbm_soft_reset |= SOFT_RESET_GRBM;
+
+	if (reset_mask & RADEON_RESET_VMC)
+		srbm_soft_reset |= SOFT_RESET_VMC;
+
+	if (reset_mask & RADEON_RESET_MC)
+		srbm_soft_reset |= SOFT_RESET_MC;
 
 	if (grbm_soft_reset) {
 		tmp = RREG32(GRBM_SOFT_RESET);
@@ -2231,32 +2323,26 @@ static int si_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 	evergreen_mc_resume(rdev, &save);
 	udelay(50);
 
-#if 0
-	if (reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP)) {
-		if (RREG32(GRBM_STATUS) & GUI_ACTIVE)
-			ret = -EAGAIN;
-	}
-
-	if (reset_mask & RADEON_RESET_DMA) {
-		if (!(RREG32(DMA_STATUS_REG) & DMA_IDLE))
-			ret = -EAGAIN;
-	}
-#endif
-
-	if (!ret)
-		r600_set_bios_scratch_engine_hung(rdev, false);
-
 	evergreen_print_gpu_status_regs(rdev);
-
-	return 0;
 }
 
 int si_asic_reset(struct radeon_device *rdev)
 {
-	return si_gpu_soft_reset(rdev, (RADEON_RESET_GFX |
-					RADEON_RESET_COMPUTE |
-					RADEON_RESET_DMA |
-					RADEON_RESET_CP));
+	u32 reset_mask;
+
+	reset_mask = si_gpu_check_soft_reset(rdev);
+
+	if (reset_mask)
+		r600_set_bios_scratch_engine_hung(rdev, true);
+
+	si_gpu_soft_reset(rdev, reset_mask);
+
+	reset_mask = si_gpu_check_soft_reset(rdev);
+
+	if (!reset_mask)
+		r600_set_bios_scratch_engine_hung(rdev, false);
+
+	return 0;
 }
 
 /* MC */
diff --git a/drivers/gpu/drm/radeon/sid.h b/drivers/gpu/drm/radeon/sid.h
index 76b5911..07fc455 100644
--- a/drivers/gpu/drm/radeon/sid.h
+++ b/drivers/gpu/drm/radeon/sid.h
@@ -61,6 +61,14 @@
 #define DMIF_ADDR_CONFIG  				0xBD4
 
 #define	SRBM_STATUS				        0xE50
+#define		GRBM_RQ_PENDING 			(1 << 5)
+#define		VMC_BUSY 				(1 << 8)
+#define		MCB_BUSY 				(1 << 9)
+#define		MCB_NON_DISPLAY_BUSY 			(1 << 10)
+#define		MCC_BUSY 				(1 << 11)
+#define		MCD_BUSY 				(1 << 12)
+#define		SEM_BUSY 				(1 << 14)
+#define		IH_BUSY 				(1 << 17)
 
 #define	SRBM_SOFT_RESET				        0x0E60
 #define		SOFT_RESET_BIF				(1 << 1)
@@ -81,6 +89,10 @@
 #define	CC_SYS_RB_BACKEND_DISABLE			0xe80
 #define	GC_USER_SYS_RB_BACKEND_DISABLE			0xe84
 
+#define	SRBM_STATUS2				        0x0EC4
+#define		DMA_BUSY 				(1 << 5)
+#define		DMA1_BUSY 				(1 << 6)
+
 #define VM_L2_CNTL					0x1400
 #define		ENABLE_L2_CACHE					(1 << 0)
 #define		ENABLE_L2_FRAGMENT_PROCESSING			(1 << 1)
-- 
cgit v0.10.2


From ca57802e521de54341efc8a56f70571f79ffac72 Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Wed, 23 Jan 2013 18:56:08 -0500
Subject: drm/radeon: halt engines before disabling MC (6xx/7xx)

It's better to halt the engines before we disable the MC.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c
index 997707b..abb143c 100644
--- a/drivers/gpu/drm/radeon/r600.c
+++ b/drivers/gpu/drm/radeon/r600.c
@@ -1397,11 +1397,6 @@ static void r600_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 
 	r600_print_gpu_status_regs(rdev);
 
-	rv515_mc_stop(rdev, &save);
-	if (r600_mc_wait_for_idle(rdev)) {
-		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
-	}
-
 	/* Disable CP parsing/prefetching */
 	if (rdev->family >= CHIP_RV770)
 		WREG32(R_0086D8_CP_ME_CNTL, S_0086D8_CP_ME_HALT(1) | S_0086D8_CP_PFP_HALT(1));
@@ -1420,6 +1415,11 @@ static void r600_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 
 	mdelay(50);
 
+	rv515_mc_stop(rdev, &save);
+	if (r600_mc_wait_for_idle(rdev)) {
+		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
+	}
+
 	if (reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE)) {
 		if (rdev->family >= CHIP_RV770)
 			grbm_soft_reset |= S_008020_SOFT_RESET_DB(1) |
-- 
cgit v0.10.2


From b21b6e7aab9d0ee97f6d3cf3654fda4348b3b7ea Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Wed, 23 Jan 2013 18:57:56 -0500
Subject: drm/radeon: halt engines before disabling MC (evergreen)

It's better to halt the engines before we disable the MC.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c
index fd5e002..f77e332 100644
--- a/drivers/gpu/drm/radeon/evergreen.c
+++ b/drivers/gpu/drm/radeon/evergreen.c
@@ -2460,11 +2460,6 @@ static void evergreen_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 
 	evergreen_print_gpu_status_regs(rdev);
 
-	evergreen_mc_stop(rdev, &save);
-	if (evergreen_mc_wait_for_idle(rdev)) {
-		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
-	}
-
 	/* Disable CP parsing/prefetching */
 	WREG32(CP_ME_CNTL, CP_ME_HALT | CP_PFP_HALT);
 
@@ -2475,6 +2470,13 @@ static void evergreen_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 		WREG32(DMA_RB_CNTL, tmp);
 	}
 
+	udelay(50);
+
+	evergreen_mc_stop(rdev, &save);
+	if (evergreen_mc_wait_for_idle(rdev)) {
+		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
+	}
+
 	if (reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE)) {
 		grbm_soft_reset |= SOFT_RESET_DB |
 			SOFT_RESET_CB |
-- 
cgit v0.10.2


From 90fb87791a698ae16af374aaaa9540fde37f6195 Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Wed, 23 Jan 2013 18:59:17 -0500
Subject: drm/radeon: halt engines before disabling MC (cayman/TN)

It's better to halt the engines before we disable the MC.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c
index eee013f..4784c4e 100644
--- a/drivers/gpu/drm/radeon/ni.c
+++ b/drivers/gpu/drm/radeon/ni.c
@@ -1405,13 +1405,6 @@ static void cayman_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 	dev_info(rdev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n",
 		 RREG32(0x14DC));
 
-	r600_set_bios_scratch_engine_hung(rdev, true);
-
-	evergreen_mc_stop(rdev, &save);
-	if (evergreen_mc_wait_for_idle(rdev)) {
-		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
-	}
-
 	/* Disable CP parsing/prefetching */
 	WREG32(CP_ME_CNTL, CP_ME_HALT | CP_PFP_HALT);
 
@@ -1429,6 +1422,13 @@ static void cayman_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 		WREG32(DMA_RB_CNTL + DMA1_REGISTER_OFFSET, tmp);
 	}
 
+	udelay(50);
+
+	evergreen_mc_stop(rdev, &save);
+	if (evergreen_mc_wait_for_idle(rdev)) {
+		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
+	}
+
 	if (reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE)) {
 		grbm_soft_reset = SOFT_RESET_CB |
 			SOFT_RESET_DB |
-- 
cgit v0.10.2


From f770d78ac159a96071e3c4e4ab97c262e79506d3 Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Wed, 23 Jan 2013 19:00:25 -0500
Subject: drm/radeon: halt engines before disabling MC (si)

It's better to halt the engines before we disable the MC.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
index 7a8ca72..89b564e 100644
--- a/drivers/gpu/drm/radeon/si.c
+++ b/drivers/gpu/drm/radeon/si.c
@@ -2220,11 +2220,6 @@ static void si_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 	dev_info(rdev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n",
 		 RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS));
 
-	evergreen_mc_stop(rdev, &save);
-	if (evergreen_mc_wait_for_idle(rdev)) {
-		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
-	}
-
 	/* Disable CP parsing/prefetching */
 	WREG32(CP_ME_CNTL, CP_ME_HALT | CP_PFP_HALT | CP_CE_HALT);
 
@@ -2241,6 +2236,13 @@ static void si_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 		WREG32(DMA_RB_CNTL + DMA1_REGISTER_OFFSET, tmp);
 	}
 
+	udelay(50);
+
+	evergreen_mc_stop(rdev, &save);
+	if (evergreen_mc_wait_for_idle(rdev)) {
+		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
+	}
+
 	if (reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP)) {
 		grbm_soft_reset = SOFT_RESET_CB |
 			SOFT_RESET_DB |
-- 
cgit v0.10.2


From 123bc1832c33218dfa677a88c2c54bc1a48a9e72 Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Thu, 24 Jan 2013 11:37:19 -0500
Subject: drm/radeon: use the reset mask to determine if rings are hung

fetch the reset mask and check if the relevant ring flags
are set to determine whether the ring is hung or not.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c
index f77e332..045955d 100644
--- a/drivers/gpu/drm/radeon/evergreen.c
+++ b/drivers/gpu/drm/radeon/evergreen.c
@@ -2308,25 +2308,6 @@ int evergreen_mc_init(struct radeon_device *rdev)
 	return 0;
 }
 
-bool evergreen_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
-{
-	u32 srbm_status;
-	u32 grbm_status;
-	u32 grbm_status_se0, grbm_status_se1;
-
-	srbm_status = RREG32(SRBM_STATUS);
-	grbm_status = RREG32(GRBM_STATUS);
-	grbm_status_se0 = RREG32(GRBM_STATUS_SE0);
-	grbm_status_se1 = RREG32(GRBM_STATUS_SE1);
-	if (!(grbm_status & GUI_ACTIVE)) {
-		radeon_ring_lockup_update(ring);
-		return false;
-	}
-	/* force CP activities */
-	radeon_ring_force_activity(rdev, ring);
-	return radeon_ring_test_lockup(rdev, ring);
-}
-
 void evergreen_print_gpu_status_regs(struct radeon_device *rdev)
 {
 	dev_info(rdev->dev, "  GRBM_STATUS               = 0x%08X\n",
@@ -2578,6 +2559,52 @@ int evergreen_asic_reset(struct radeon_device *rdev)
 	return 0;
 }
 
+/**
+ * evergreen_gfx_is_lockup - Check if the GFX engine is locked up
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon_ring structure holding ring information
+ *
+ * Check if the GFX engine is locked up.
+ * Returns true if the engine appears to be locked up, false if not.
+ */
+bool evergreen_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
+{
+	u32 reset_mask = evergreen_gpu_check_soft_reset(rdev);
+
+	if (!(reset_mask & (RADEON_RESET_GFX |
+			    RADEON_RESET_COMPUTE |
+			    RADEON_RESET_CP))) {
+		radeon_ring_lockup_update(ring);
+		return false;
+	}
+	/* force CP activities */
+	radeon_ring_force_activity(rdev, ring);
+	return radeon_ring_test_lockup(rdev, ring);
+}
+
+/**
+ * evergreen_dma_is_lockup - Check if the DMA engine is locked up
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon_ring structure holding ring information
+ *
+ * Check if the async DMA engine is locked up.
+ * Returns true if the engine appears to be locked up, false if not.
+ */
+bool evergreen_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
+{
+	u32 reset_mask = evergreen_gpu_check_soft_reset(rdev);
+
+	if (!(reset_mask & RADEON_RESET_DMA)) {
+		radeon_ring_lockup_update(ring);
+		return false;
+	}
+	/* force ring activities */
+	radeon_ring_force_activity(rdev, ring);
+	return radeon_ring_test_lockup(rdev, ring);
+}
+
 /* Interrupts */
 
 u32 evergreen_get_vblank_counter(struct radeon_device *rdev, int crtc)
diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c
index 4784c4e..b6e8055 100644
--- a/drivers/gpu/drm/radeon/ni.c
+++ b/drivers/gpu/drm/radeon/ni.c
@@ -1534,23 +1534,49 @@ int cayman_asic_reset(struct radeon_device *rdev)
 }
 
 /**
+ * cayman_gfx_is_lockup - Check if the GFX engine is locked up
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon_ring structure holding ring information
+ *
+ * Check if the GFX engine is locked up.
+ * Returns true if the engine appears to be locked up, false if not.
+ */
+bool cayman_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
+{
+	u32 reset_mask = cayman_gpu_check_soft_reset(rdev);
+
+	if (!(reset_mask & (RADEON_RESET_GFX |
+			    RADEON_RESET_COMPUTE |
+			    RADEON_RESET_CP))) {
+		radeon_ring_lockup_update(ring);
+		return false;
+	}
+	/* force CP activities */
+	radeon_ring_force_activity(rdev, ring);
+	return radeon_ring_test_lockup(rdev, ring);
+}
+
+/**
  * cayman_dma_is_lockup - Check if the DMA engine is locked up
  *
  * @rdev: radeon_device pointer
  * @ring: radeon_ring structure holding ring information
  *
- * Check if the async DMA engine is locked up (cayman-SI).
+ * Check if the async DMA engine is locked up.
  * Returns true if the engine appears to be locked up, false if not.
  */
 bool cayman_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
 {
-	u32 dma_status_reg;
+	u32 reset_mask = cayman_gpu_check_soft_reset(rdev);
+	u32 mask;
 
 	if (ring->idx == R600_RING_TYPE_DMA_INDEX)
-		dma_status_reg = RREG32(DMA_STATUS_REG + DMA0_REGISTER_OFFSET);
+		mask = RADEON_RESET_DMA;
 	else
-		dma_status_reg = RREG32(DMA_STATUS_REG + DMA1_REGISTER_OFFSET);
-	if (dma_status_reg & DMA_IDLE) {
+		mask = RADEON_RESET_DMA1;
+
+	if (!(reset_mask & mask)) {
 		radeon_ring_lockup_update(ring);
 		return false;
 	}
diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c
index abb143c..3f29276 100644
--- a/drivers/gpu/drm/radeon/r600.c
+++ b/drivers/gpu/drm/radeon/r600.c
@@ -1537,16 +1537,22 @@ int r600_asic_reset(struct radeon_device *rdev)
 	return 0;
 }
 
-bool r600_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
+/**
+ * r600_gfx_is_lockup - Check if the GFX engine is locked up
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon_ring structure holding ring information
+ *
+ * Check if the GFX engine is locked up.
+ * Returns true if the engine appears to be locked up, false if not.
+ */
+bool r600_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
 {
-	u32 srbm_status;
-	u32 grbm_status;
-	u32 grbm_status2;
-
-	srbm_status = RREG32(R_000E50_SRBM_STATUS);
-	grbm_status = RREG32(R_008010_GRBM_STATUS);
-	grbm_status2 = RREG32(R_008014_GRBM_STATUS2);
-	if (!G_008010_GUI_ACTIVE(grbm_status)) {
+	u32 reset_mask = r600_gpu_check_soft_reset(rdev);
+
+	if (!(reset_mask & (RADEON_RESET_GFX |
+			    RADEON_RESET_COMPUTE |
+			    RADEON_RESET_CP))) {
 		radeon_ring_lockup_update(ring);
 		return false;
 	}
@@ -1561,15 +1567,14 @@ bool r600_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
  * @rdev: radeon_device pointer
  * @ring: radeon_ring structure holding ring information
  *
- * Check if the async DMA engine is locked up (r6xx-evergreen).
+ * Check if the async DMA engine is locked up.
  * Returns true if the engine appears to be locked up, false if not.
  */
 bool r600_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
 {
-	u32 dma_status_reg;
+	u32 reset_mask = r600_gpu_check_soft_reset(rdev);
 
-	dma_status_reg = RREG32(DMA_STATUS_REG);
-	if (dma_status_reg & DMA_IDLE) {
+	if (!(reset_mask & RADEON_RESET_DMA)) {
 		radeon_ring_lockup_update(ring);
 		return false;
 	}
diff --git a/drivers/gpu/drm/radeon/radeon_asic.c b/drivers/gpu/drm/radeon/radeon_asic.c
index 0b202c0..82b5ef0 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.c
+++ b/drivers/gpu/drm/radeon/radeon_asic.c
@@ -946,7 +946,7 @@ static struct radeon_asic r600_asic = {
 			.cs_parse = &r600_cs_parse,
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
-			.is_lockup = &r600_gpu_is_lockup,
+			.is_lockup = &r600_gfx_is_lockup,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
 			.ib_execute = &r600_dma_ring_ib_execute,
@@ -1030,7 +1030,7 @@ static struct radeon_asic rs780_asic = {
 			.cs_parse = &r600_cs_parse,
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
-			.is_lockup = &r600_gpu_is_lockup,
+			.is_lockup = &r600_gfx_is_lockup,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
 			.ib_execute = &r600_dma_ring_ib_execute,
@@ -1114,7 +1114,7 @@ static struct radeon_asic rv770_asic = {
 			.cs_parse = &r600_cs_parse,
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
-			.is_lockup = &r600_gpu_is_lockup,
+			.is_lockup = &r600_gfx_is_lockup,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
 			.ib_execute = &r600_dma_ring_ib_execute,
@@ -1198,7 +1198,7 @@ static struct radeon_asic evergreen_asic = {
 			.cs_parse = &evergreen_cs_parse,
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
-			.is_lockup = &evergreen_gpu_is_lockup,
+			.is_lockup = &evergreen_gfx_is_lockup,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
 			.ib_execute = &evergreen_dma_ring_ib_execute,
@@ -1207,7 +1207,7 @@ static struct radeon_asic evergreen_asic = {
 			.cs_parse = &evergreen_dma_cs_parse,
 			.ring_test = &r600_dma_ring_test,
 			.ib_test = &r600_dma_ib_test,
-			.is_lockup = &r600_dma_is_lockup,
+			.is_lockup = &evergreen_dma_is_lockup,
 		}
 	},
 	.irq = {
@@ -1282,7 +1282,7 @@ static struct radeon_asic sumo_asic = {
 			.cs_parse = &evergreen_cs_parse,
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
-			.is_lockup = &evergreen_gpu_is_lockup,
+			.is_lockup = &evergreen_gfx_is_lockup,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
 			.ib_execute = &evergreen_dma_ring_ib_execute,
@@ -1291,7 +1291,7 @@ static struct radeon_asic sumo_asic = {
 			.cs_parse = &evergreen_dma_cs_parse,
 			.ring_test = &r600_dma_ring_test,
 			.ib_test = &r600_dma_ib_test,
-			.is_lockup = &r600_dma_is_lockup,
+			.is_lockup = &evergreen_dma_is_lockup,
 		}
 	},
 	.irq = {
@@ -1366,7 +1366,7 @@ static struct radeon_asic btc_asic = {
 			.cs_parse = &evergreen_cs_parse,
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
-			.is_lockup = &evergreen_gpu_is_lockup,
+			.is_lockup = &evergreen_gfx_is_lockup,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
 			.ib_execute = &evergreen_dma_ring_ib_execute,
@@ -1375,7 +1375,7 @@ static struct radeon_asic btc_asic = {
 			.cs_parse = &evergreen_dma_cs_parse,
 			.ring_test = &r600_dma_ring_test,
 			.ib_test = &r600_dma_ib_test,
-			.is_lockup = &r600_dma_is_lockup,
+			.is_lockup = &evergreen_dma_is_lockup,
 		}
 	},
 	.irq = {
@@ -1457,7 +1457,7 @@ static struct radeon_asic cayman_asic = {
 			.cs_parse = &evergreen_cs_parse,
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
-			.is_lockup = &evergreen_gpu_is_lockup,
+			.is_lockup = &cayman_gfx_is_lockup,
 			.vm_flush = &cayman_vm_flush,
 		},
 		[CAYMAN_RING_TYPE_CP1_INDEX] = {
@@ -1468,7 +1468,7 @@ static struct radeon_asic cayman_asic = {
 			.cs_parse = &evergreen_cs_parse,
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
-			.is_lockup = &evergreen_gpu_is_lockup,
+			.is_lockup = &cayman_gfx_is_lockup,
 			.vm_flush = &cayman_vm_flush,
 		},
 		[CAYMAN_RING_TYPE_CP2_INDEX] = {
@@ -1479,7 +1479,7 @@ static struct radeon_asic cayman_asic = {
 			.cs_parse = &evergreen_cs_parse,
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
-			.is_lockup = &evergreen_gpu_is_lockup,
+			.is_lockup = &cayman_gfx_is_lockup,
 			.vm_flush = &cayman_vm_flush,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
@@ -1584,7 +1584,7 @@ static struct radeon_asic trinity_asic = {
 			.cs_parse = &evergreen_cs_parse,
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
-			.is_lockup = &evergreen_gpu_is_lockup,
+			.is_lockup = &cayman_gfx_is_lockup,
 			.vm_flush = &cayman_vm_flush,
 		},
 		[CAYMAN_RING_TYPE_CP1_INDEX] = {
@@ -1595,7 +1595,7 @@ static struct radeon_asic trinity_asic = {
 			.cs_parse = &evergreen_cs_parse,
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
-			.is_lockup = &evergreen_gpu_is_lockup,
+			.is_lockup = &cayman_gfx_is_lockup,
 			.vm_flush = &cayman_vm_flush,
 		},
 		[CAYMAN_RING_TYPE_CP2_INDEX] = {
@@ -1606,7 +1606,7 @@ static struct radeon_asic trinity_asic = {
 			.cs_parse = &evergreen_cs_parse,
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
-			.is_lockup = &evergreen_gpu_is_lockup,
+			.is_lockup = &cayman_gfx_is_lockup,
 			.vm_flush = &cayman_vm_flush,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
@@ -1711,7 +1711,7 @@ static struct radeon_asic si_asic = {
 			.cs_parse = NULL,
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
-			.is_lockup = &si_gpu_is_lockup,
+			.is_lockup = &si_gfx_is_lockup,
 			.vm_flush = &si_vm_flush,
 		},
 		[CAYMAN_RING_TYPE_CP1_INDEX] = {
@@ -1722,7 +1722,7 @@ static struct radeon_asic si_asic = {
 			.cs_parse = NULL,
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
-			.is_lockup = &si_gpu_is_lockup,
+			.is_lockup = &si_gfx_is_lockup,
 			.vm_flush = &si_vm_flush,
 		},
 		[CAYMAN_RING_TYPE_CP2_INDEX] = {
@@ -1733,7 +1733,7 @@ static struct radeon_asic si_asic = {
 			.cs_parse = NULL,
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
-			.is_lockup = &si_gpu_is_lockup,
+			.is_lockup = &si_gfx_is_lockup,
 			.vm_flush = &si_vm_flush,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
@@ -1744,7 +1744,7 @@ static struct radeon_asic si_asic = {
 			.cs_parse = NULL,
 			.ring_test = &r600_dma_ring_test,
 			.ib_test = &r600_dma_ib_test,
-			.is_lockup = &cayman_dma_is_lockup,
+			.is_lockup = &si_dma_is_lockup,
 			.vm_flush = &si_dma_vm_flush,
 		},
 		[CAYMAN_RING_TYPE_DMA1_INDEX] = {
@@ -1755,7 +1755,7 @@ static struct radeon_asic si_asic = {
 			.cs_parse = NULL,
 			.ring_test = &r600_dma_ring_test,
 			.ib_test = &r600_dma_ib_test,
-			.is_lockup = &cayman_dma_is_lockup,
+			.is_lockup = &si_dma_is_lockup,
 			.vm_flush = &si_dma_vm_flush,
 		}
 	},
diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h
index 15d70e6..e429e25 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.h
+++ b/drivers/gpu/drm/radeon/radeon_asic.h
@@ -319,7 +319,7 @@ void r600_dma_semaphore_ring_emit(struct radeon_device *rdev,
 				  bool emit_wait);
 void r600_dma_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib);
 bool r600_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring);
-bool r600_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp);
+bool r600_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp);
 int r600_asic_reset(struct radeon_device *rdev);
 int r600_set_surface_reg(struct radeon_device *rdev, int reg,
 			 uint32_t tiling_flags, uint32_t pitch,
@@ -422,7 +422,8 @@ int evergreen_init(struct radeon_device *rdev);
 void evergreen_fini(struct radeon_device *rdev);
 int evergreen_suspend(struct radeon_device *rdev);
 int evergreen_resume(struct radeon_device *rdev);
-bool evergreen_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp);
+bool evergreen_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp);
+bool evergreen_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp);
 int evergreen_asic_reset(struct radeon_device *rdev);
 void evergreen_bandwidth_update(struct radeon_device *rdev);
 void evergreen_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib);
@@ -480,6 +481,7 @@ int evergreen_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib);
 int evergreen_dma_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib);
 void cayman_dma_ring_ib_execute(struct radeon_device *rdev,
 				struct radeon_ib *ib);
+bool cayman_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring);
 bool cayman_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring);
 void cayman_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
 
@@ -496,7 +498,8 @@ int si_init(struct radeon_device *rdev);
 void si_fini(struct radeon_device *rdev);
 int si_suspend(struct radeon_device *rdev);
 int si_resume(struct radeon_device *rdev);
-bool si_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp);
+bool si_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp);
+bool si_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp);
 int si_asic_reset(struct radeon_device *rdev);
 void si_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib);
 int si_irq_set(struct radeon_device *rdev);
diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
index 89b564e..cd83bc5 100644
--- a/drivers/gpu/drm/radeon/si.c
+++ b/drivers/gpu/drm/radeon/si.c
@@ -2108,26 +2108,6 @@ static int si_cp_resume(struct radeon_device *rdev)
 	return 0;
 }
 
-bool si_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
-{
-	u32 srbm_status;
-	u32 grbm_status, grbm_status2;
-	u32 grbm_status_se0, grbm_status_se1;
-
-	srbm_status = RREG32(SRBM_STATUS);
-	grbm_status = RREG32(GRBM_STATUS);
-	grbm_status2 = RREG32(GRBM_STATUS2);
-	grbm_status_se0 = RREG32(GRBM_STATUS_SE0);
-	grbm_status_se1 = RREG32(GRBM_STATUS_SE1);
-	if (!(grbm_status & GUI_ACTIVE)) {
-		radeon_ring_lockup_update(ring);
-		return false;
-	}
-	/* force CP activities */
-	radeon_ring_force_activity(rdev, ring);
-	return radeon_ring_test_lockup(rdev, ring);
-}
-
 static u32 si_gpu_check_soft_reset(struct radeon_device *rdev)
 {
 	u32 reset_mask = 0;
@@ -2347,6 +2327,58 @@ int si_asic_reset(struct radeon_device *rdev)
 	return 0;
 }
 
+/**
+ * si_gfx_is_lockup - Check if the GFX engine is locked up
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon_ring structure holding ring information
+ *
+ * Check if the GFX engine is locked up.
+ * Returns true if the engine appears to be locked up, false if not.
+ */
+bool si_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
+{
+	u32 reset_mask = si_gpu_check_soft_reset(rdev);
+
+	if (!(reset_mask & (RADEON_RESET_GFX |
+			    RADEON_RESET_COMPUTE |
+			    RADEON_RESET_CP))) {
+		radeon_ring_lockup_update(ring);
+		return false;
+	}
+	/* force CP activities */
+	radeon_ring_force_activity(rdev, ring);
+	return radeon_ring_test_lockup(rdev, ring);
+}
+
+/**
+ * si_dma_is_lockup - Check if the DMA engine is locked up
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon_ring structure holding ring information
+ *
+ * Check if the async DMA engine is locked up.
+ * Returns true if the engine appears to be locked up, false if not.
+ */
+bool si_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
+{
+	u32 reset_mask = si_gpu_check_soft_reset(rdev);
+	u32 mask;
+
+	if (ring->idx == R600_RING_TYPE_DMA_INDEX)
+		mask = RADEON_RESET_DMA;
+	else
+		mask = RADEON_RESET_DMA1;
+
+	if (!(reset_mask & mask)) {
+		radeon_ring_lockup_update(ring);
+		return false;
+	}
+	/* force ring activities */
+	radeon_ring_force_activity(rdev, ring);
+	return radeon_ring_test_lockup(rdev, ring);
+}
+
 /* MC */
 static void si_mc_program(struct radeon_device *rdev)
 {
-- 
cgit v0.10.2


From 24178ec42b0985d485886bc43b97e54ff173627e Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Thu, 24 Jan 2013 15:00:17 -0500
Subject: drm/radeon: don't reset the MC on IGPs/APUs

The MC isn't part of the GPU per se.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c
index 045955d..2916de8 100644
--- a/drivers/gpu/drm/radeon/evergreen.c
+++ b/drivers/gpu/drm/radeon/evergreen.c
@@ -2500,8 +2500,10 @@ static void evergreen_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 	if (reset_mask & RADEON_RESET_VMC)
 		srbm_soft_reset |= SOFT_RESET_VMC;
 
-	if (reset_mask & RADEON_RESET_MC)
-		srbm_soft_reset |= SOFT_RESET_MC;
+	if (!(rdev->flags & RADEON_IS_IGP)) {
+		if (reset_mask & RADEON_RESET_MC)
+			srbm_soft_reset |= SOFT_RESET_MC;
+	}
 
 	if (grbm_soft_reset) {
 		tmp = RREG32(GRBM_SOFT_RESET);
diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c
index b6e8055..170bd03 100644
--- a/drivers/gpu/drm/radeon/ni.c
+++ b/drivers/gpu/drm/radeon/ni.c
@@ -1474,8 +1474,10 @@ static void cayman_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 	if (reset_mask & RADEON_RESET_VMC)
 		srbm_soft_reset |= SOFT_RESET_VMC;
 
-	if (reset_mask & RADEON_RESET_MC)
-		srbm_soft_reset |= SOFT_RESET_MC;
+	if (!(rdev->flags & RADEON_IS_IGP)) {
+		if (reset_mask & RADEON_RESET_MC)
+			srbm_soft_reset |= SOFT_RESET_MC;
+	}
 
 	if (grbm_soft_reset) {
 		tmp = RREG32(GRBM_SOFT_RESET);
diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c
index 3f29276..dbcb075 100644
--- a/drivers/gpu/drm/radeon/r600.c
+++ b/drivers/gpu/drm/radeon/r600.c
@@ -1475,8 +1475,10 @@ static void r600_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
 	if (reset_mask & RADEON_RESET_GRBM)
 		srbm_soft_reset |= S_000E60_SOFT_RESET_GRBM(1);
 
-	if (reset_mask & RADEON_RESET_MC)
-		srbm_soft_reset |= S_000E60_SOFT_RESET_MC(1);
+	if (!(rdev->flags & RADEON_IS_IGP)) {
+		if (reset_mask & RADEON_RESET_MC)
+			srbm_soft_reset |= S_000E60_SOFT_RESET_MC(1);
+	}
 
 	if (reset_mask & RADEON_RESET_VMC)
 		srbm_soft_reset |= S_000E60_SOFT_RESET_VMC(1);
-- 
cgit v0.10.2


From 7d37beaaf3dbc6ff16f4d32a4dd6f8c557c6ab50 Mon Sep 17 00:00:00 2001
From: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Date: Fri, 1 Feb 2013 10:14:20 +0900
Subject: GPU/i915: Fix acpi_bus_get_device() check in
 drivers/gpu/drm/i915/intel_opregion.c

acpi_bus_get_device() returns int not acpi_status.

The patch change not to apply ACPI_FAILURE() to the return value of
acpi_bus_get_device().

Signed-off-by: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_opregion.c b/drivers/gpu/drm/i915/intel_opregion.c
index 7741c22..4d33874 100644
--- a/drivers/gpu/drm/i915/intel_opregion.c
+++ b/drivers/gpu/drm/i915/intel_opregion.c
@@ -347,7 +347,7 @@ static void intel_didl_outputs(struct drm_device *dev)
 	int i = 0;
 
 	handle = DEVICE_ACPI_HANDLE(&dev->pdev->dev);
-	if (!handle || ACPI_FAILURE(acpi_bus_get_device(handle, &acpi_dev)))
+	if (!handle || acpi_bus_get_device(handle, &acpi_dev))
 		return;
 
 	if (acpi_is_video_device(acpi_dev))
-- 
cgit v0.10.2


From 43f1214aa094e46efdfc0255d9601be0e5ea0f62 Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Fri, 1 Feb 2013 17:32:42 +0100
Subject: drm/radeon: use IBs for VM page table updates v2
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

For very large page table updates, we can exceed the
size of the ring.  To avoid this, use an IB to perform
the page table update.

v2(ck): cleanup the IB infrastructure and the use it instead
        of filling the struct ourself.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Christian König <christian.koenig@amd.com>

diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c
index 170bd03..7cead76 100644
--- a/drivers/gpu/drm/radeon/ni.c
+++ b/drivers/gpu/drm/radeon/ni.c
@@ -1946,19 +1946,21 @@ uint32_t cayman_vm_page_flags(struct radeon_device *rdev, uint32_t flags)
  * cayman_vm_set_page - update the page tables using the CP
  *
  * @rdev: radeon_device pointer
+ * @ib: indirect buffer to fill with commands
  * @pe: addr of the page entry
  * @addr: dst addr to write into pe
  * @count: number of page entries to update
  * @incr: increase next addr by incr bytes
  * @flags: access flags
  *
- * Update the page tables using the CP (cayman-si).
+ * Update the page tables using the CP (cayman/TN).
  */
-void cayman_vm_set_page(struct radeon_device *rdev, uint64_t pe,
+void cayman_vm_set_page(struct radeon_device *rdev,
+			struct radeon_ib *ib,
+			uint64_t pe,
 			uint64_t addr, unsigned count,
 			uint32_t incr, uint32_t flags)
 {
-	struct radeon_ring *ring = &rdev->ring[rdev->asic->vm.pt_ring_index];
 	uint32_t r600_flags = cayman_vm_page_flags(rdev, flags);
 	uint64_t value;
 	unsigned ndw;
@@ -1969,9 +1971,9 @@ void cayman_vm_set_page(struct radeon_device *rdev, uint64_t pe,
 			if (ndw > 0x3FFF)
 				ndw = 0x3FFF;
 
-			radeon_ring_write(ring, PACKET3(PACKET3_ME_WRITE, ndw));
-			radeon_ring_write(ring, pe);
-			radeon_ring_write(ring, upper_32_bits(pe) & 0xff);
+			ib->ptr[ib->length_dw++] = PACKET3(PACKET3_ME_WRITE, ndw);
+			ib->ptr[ib->length_dw++] = pe;
+			ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
 			for (; ndw > 1; ndw -= 2, --count, pe += 8) {
 				if (flags & RADEON_VM_PAGE_SYSTEM) {
 					value = radeon_vm_map_gart(rdev, addr);
@@ -1983,8 +1985,8 @@ void cayman_vm_set_page(struct radeon_device *rdev, uint64_t pe,
 				}
 				addr += incr;
 				value |= r600_flags;
-				radeon_ring_write(ring, value);
-				radeon_ring_write(ring, upper_32_bits(value));
+				ib->ptr[ib->length_dw++] = value;
+				ib->ptr[ib->length_dw++] = upper_32_bits(value);
 			}
 		}
 	} else {
@@ -1994,9 +1996,9 @@ void cayman_vm_set_page(struct radeon_device *rdev, uint64_t pe,
 				ndw = 0xFFFFE;
 
 			/* for non-physically contiguous pages (system) */
-			radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_WRITE, 0, 0, ndw));
-			radeon_ring_write(ring, pe);
-			radeon_ring_write(ring, upper_32_bits(pe) & 0xff);
+			ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_WRITE, 0, 0, ndw);
+			ib->ptr[ib->length_dw++] = pe;
+			ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
 			for (; ndw > 0; ndw -= 2, --count, pe += 8) {
 				if (flags & RADEON_VM_PAGE_SYSTEM) {
 					value = radeon_vm_map_gart(rdev, addr);
@@ -2008,10 +2010,12 @@ void cayman_vm_set_page(struct radeon_device *rdev, uint64_t pe,
 				}
 				addr += incr;
 				value |= r600_flags;
-				radeon_ring_write(ring, value);
-				radeon_ring_write(ring, upper_32_bits(value));
+				ib->ptr[ib->length_dw++] = value;
+				ib->ptr[ib->length_dw++] = upper_32_bits(value);
 			}
 		}
+		while (ib->length_dw & 0x7)
+			ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0);
 	}
 }
 
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 6539d6c..307d681 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -780,6 +780,7 @@ int radeon_ib_get(struct radeon_device *rdev, int ring,
 		  struct radeon_ib *ib, struct radeon_vm *vm,
 		  unsigned size);
 void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib *ib);
+void radeon_ib_sync_to(struct radeon_ib *ib, struct radeon_fence *fence);
 int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib,
 		       struct radeon_ib *const_ib);
 int radeon_ib_pool_init(struct radeon_device *rdev);
@@ -1188,7 +1189,9 @@ struct radeon_asic {
 		void (*fini)(struct radeon_device *rdev);
 
 		u32 pt_ring_index;
-		void (*set_page)(struct radeon_device *rdev, uint64_t pe,
+		void (*set_page)(struct radeon_device *rdev,
+				 struct radeon_ib *ib,
+				 uint64_t pe,
 				 uint64_t addr, unsigned count,
 				 uint32_t incr, uint32_t flags);
 	} vm;
@@ -1810,7 +1813,7 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v);
 #define radeon_gart_set_page(rdev, i, p) (rdev)->asic->gart.set_page((rdev), (i), (p))
 #define radeon_asic_vm_init(rdev) (rdev)->asic->vm.init((rdev))
 #define radeon_asic_vm_fini(rdev) (rdev)->asic->vm.fini((rdev))
-#define radeon_asic_vm_set_page(rdev, pe, addr, count, incr, flags) ((rdev)->asic->vm.set_page((rdev), (pe), (addr), (count), (incr), (flags)))
+#define radeon_asic_vm_set_page(rdev, ib, pe, addr, count, incr, flags) ((rdev)->asic->vm.set_page((rdev), (ib), (pe), (addr), (count), (incr), (flags)))
 #define radeon_ring_start(rdev, r, cp) (rdev)->asic->ring[(r)].ring_start((rdev), (cp))
 #define radeon_ring_test(rdev, r, cp) (rdev)->asic->ring[(r)].ring_test((rdev), (cp))
 #define radeon_ib_test(rdev, r, cp) (rdev)->asic->ring[(r)].ib_test((rdev), (cp))
diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h
index e429e25..f4134a8 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.h
+++ b/drivers/gpu/drm/radeon/radeon_asic.h
@@ -474,7 +474,9 @@ int cayman_vm_init(struct radeon_device *rdev);
 void cayman_vm_fini(struct radeon_device *rdev);
 void cayman_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
 uint32_t cayman_vm_page_flags(struct radeon_device *rdev, uint32_t flags);
-void cayman_vm_set_page(struct radeon_device *rdev, uint64_t pe,
+void cayman_vm_set_page(struct radeon_device *rdev,
+			struct radeon_ib *ib,
+			uint64_t pe,
 			uint64_t addr, unsigned count,
 			uint32_t incr, uint32_t flags);
 int evergreen_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib);
@@ -506,7 +508,9 @@ int si_irq_set(struct radeon_device *rdev);
 int si_irq_process(struct radeon_device *rdev);
 int si_vm_init(struct radeon_device *rdev);
 void si_vm_fini(struct radeon_device *rdev);
-void si_vm_set_page(struct radeon_device *rdev, uint64_t pe,
+void si_vm_set_page(struct radeon_device *rdev,
+		    struct radeon_ib *ib,
+		    uint64_t pe,
 		    uint64_t addr, unsigned count,
 		    uint32_t incr, uint32_t flags);
 void si_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c
index 1d214b6..70d3824 100644
--- a/drivers/gpu/drm/radeon/radeon_cs.c
+++ b/drivers/gpu/drm/radeon/radeon_cs.c
@@ -125,18 +125,6 @@ static int radeon_cs_get_ring(struct radeon_cs_parser *p, u32 ring, s32 priority
 	return 0;
 }
 
-static void radeon_cs_sync_to(struct radeon_cs_parser *p,
-			      struct radeon_fence *fence)
-{
-	struct radeon_fence *other;
-
-	if (!fence)
-		return;
-
-	other = p->ib.sync_to[fence->ring];
-	p->ib.sync_to[fence->ring] = radeon_fence_later(fence, other);
-}
-
 static void radeon_cs_sync_rings(struct radeon_cs_parser *p)
 {
 	int i;
@@ -145,7 +133,7 @@ static void radeon_cs_sync_rings(struct radeon_cs_parser *p)
 		if (!p->relocs[i].robj)
 			continue;
 
-		radeon_cs_sync_to(p, p->relocs[i].robj->tbo.sync_obj);
+		radeon_ib_sync_to(&p->ib, p->relocs[i].robj->tbo.sync_obj);
 	}
 }
 
@@ -472,8 +460,9 @@ static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev,
 		goto out;
 	}
 	radeon_cs_sync_rings(parser);
-	radeon_cs_sync_to(parser, vm->fence);
-	radeon_cs_sync_to(parser, radeon_vm_grab_id(rdev, vm, parser->ring));
+	radeon_ib_sync_to(&parser->ib, vm->fence);
+	radeon_ib_sync_to(&parser->ib, radeon_vm_grab_id(
+		rdev, vm, parser->ring));
 
 	if ((rdev->family >= CHIP_TAHITI) &&
 	    (parser->chunk_const_ib_idx != -1)) {
diff --git a/drivers/gpu/drm/radeon/radeon_gart.c b/drivers/gpu/drm/radeon/radeon_gart.c
index 6e24f84..2c1341f 100644
--- a/drivers/gpu/drm/radeon/radeon_gart.c
+++ b/drivers/gpu/drm/radeon/radeon_gart.c
@@ -929,6 +929,7 @@ uint64_t radeon_vm_map_gart(struct radeon_device *rdev, uint64_t addr)
  */
 static int radeon_vm_update_pdes(struct radeon_device *rdev,
 				 struct radeon_vm *vm,
+				 struct radeon_ib *ib,
 				 uint64_t start, uint64_t end)
 {
 	static const uint32_t incr = RADEON_VM_PTE_COUNT * 8;
@@ -971,7 +972,7 @@ retry:
 		    ((last_pt + incr * count) != pt)) {
 
 			if (count) {
-				radeon_asic_vm_set_page(rdev, last_pde,
+				radeon_asic_vm_set_page(rdev, ib, last_pde,
 							last_pt, count, incr,
 							RADEON_VM_PAGE_VALID);
 			}
@@ -985,7 +986,7 @@ retry:
 	}
 
 	if (count) {
-		radeon_asic_vm_set_page(rdev, last_pde, last_pt, count,
+		radeon_asic_vm_set_page(rdev, ib, last_pde, last_pt, count,
 					incr, RADEON_VM_PAGE_VALID);
 
 	}
@@ -1009,6 +1010,7 @@ retry:
  */
 static void radeon_vm_update_ptes(struct radeon_device *rdev,
 				  struct radeon_vm *vm,
+				  struct radeon_ib *ib,
 				  uint64_t start, uint64_t end,
 				  uint64_t dst, uint32_t flags)
 {
@@ -1038,7 +1040,7 @@ static void radeon_vm_update_ptes(struct radeon_device *rdev,
 		if ((last_pte + 8 * count) != pte) {
 
 			if (count) {
-				radeon_asic_vm_set_page(rdev, last_pte,
+				radeon_asic_vm_set_page(rdev, ib, last_pte,
 							last_dst, count,
 							RADEON_GPU_PAGE_SIZE,
 							flags);
@@ -1056,7 +1058,8 @@ static void radeon_vm_update_ptes(struct radeon_device *rdev,
 	}
 
 	if (count) {
-		radeon_asic_vm_set_page(rdev, last_pte,	last_dst, count,
+		radeon_asic_vm_set_page(rdev, ib, last_pte,
+					last_dst, count,
 					RADEON_GPU_PAGE_SIZE, flags);
 	}
 }
@@ -1080,8 +1083,7 @@ int radeon_vm_bo_update_pte(struct radeon_device *rdev,
 			    struct ttm_mem_reg *mem)
 {
 	unsigned ridx = rdev->asic->vm.pt_ring_index;
-	struct radeon_ring *ring = &rdev->ring[ridx];
-	struct radeon_semaphore *sem = NULL;
+	struct radeon_ib ib;
 	struct radeon_bo_va *bo_va;
 	unsigned nptes, npdes, ndw;
 	uint64_t addr;
@@ -1124,25 +1126,13 @@ int radeon_vm_bo_update_pte(struct radeon_device *rdev,
 		bo_va->valid = false;
 	}
 
-	if (vm->fence && radeon_fence_signaled(vm->fence)) {
-		radeon_fence_unref(&vm->fence);
-	}
-
-	if (vm->fence && vm->fence->ring != ridx) {
-		r = radeon_semaphore_create(rdev, &sem);
-		if (r) {
-			return r;
-		}
-	}
-
 	nptes = radeon_bo_ngpu_pages(bo);
 
 	/* assume two extra pdes in case the mapping overlaps the borders */
 	npdes = (nptes >> RADEON_VM_BLOCK_SIZE) + 2;
 
-	/* estimate number of dw needed */
-	/* semaphore, fence and padding */
-	ndw = 32;
+	/* padding, etc. */
+	ndw = 64;
 
 	if (RADEON_VM_BLOCK_SIZE > 11)
 		/* reserve space for one header for every 2k dwords */
@@ -1161,33 +1151,31 @@ int radeon_vm_bo_update_pte(struct radeon_device *rdev,
 	/* reserve space for pde addresses */
 	ndw += npdes * 2;
 
-	r = radeon_ring_lock(rdev, ring, ndw);
-	if (r) {
-		return r;
-	}
+	/* update too big for an IB */
+	if (ndw > 0xfffff)
+		return -ENOMEM;
 
-	if (sem && radeon_fence_need_sync(vm->fence, ridx)) {
-		radeon_semaphore_sync_rings(rdev, sem, vm->fence->ring, ridx);
-		radeon_fence_note_sync(vm->fence, ridx);
-	}
+	r = radeon_ib_get(rdev, ridx, &ib, NULL, ndw * 4);
+	ib.length_dw = 0;
 
-	r = radeon_vm_update_pdes(rdev, vm, bo_va->soffset, bo_va->eoffset);
+	r = radeon_vm_update_pdes(rdev, vm, &ib, bo_va->soffset, bo_va->eoffset);
 	if (r) {
-		radeon_ring_unlock_undo(rdev, ring);
+		radeon_ib_free(rdev, &ib);
 		return r;
 	}
 
-	radeon_vm_update_ptes(rdev, vm, bo_va->soffset, bo_va->eoffset,
+	radeon_vm_update_ptes(rdev, vm, &ib, bo_va->soffset, bo_va->eoffset,
 			      addr, bo_va->flags);
 
-	radeon_fence_unref(&vm->fence);
-	r = radeon_fence_emit(rdev, &vm->fence, ridx);
+	radeon_ib_sync_to(&ib, vm->fence);
+	r = radeon_ib_schedule(rdev, &ib, NULL);
 	if (r) {
-		radeon_ring_unlock_undo(rdev, ring);
+		radeon_ib_free(rdev, &ib);
 		return r;
 	}
-	radeon_ring_unlock_commit(rdev, ring);
-	radeon_semaphore_free(rdev, &sem, vm->fence);
+	radeon_fence_unref(&vm->fence);
+	vm->fence = radeon_fence_ref(ib.fence);
+	radeon_ib_free(rdev, &ib);
 	radeon_fence_unref(&vm->last_flush);
 
 	return 0;
diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c
index cd72062..8d58e26 100644
--- a/drivers/gpu/drm/radeon/radeon_ring.c
+++ b/drivers/gpu/drm/radeon/radeon_ring.c
@@ -109,6 +109,25 @@ void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib *ib)
 }
 
 /**
+ * radeon_ib_sync_to - sync to fence before executing the IB
+ *
+ * @ib: IB object to add fence to
+ * @fence: fence to sync to
+ *
+ * Sync to the fence before executing the IB
+ */
+void radeon_ib_sync_to(struct radeon_ib *ib, struct radeon_fence *fence)
+{
+	struct radeon_fence *other;
+
+	if (!fence)
+		return;
+
+	other = ib->sync_to[fence->ring];
+	ib->sync_to[fence->ring] = radeon_fence_later(fence, other);
+}
+
+/**
  * radeon_ib_schedule - schedule an IB (Indirect Buffer) on the ring
  *
  * @rdev: radeon_device pointer
diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
index cd83bc5..a910cb9 100644
--- a/drivers/gpu/drm/radeon/si.c
+++ b/drivers/gpu/drm/radeon/si.c
@@ -3043,19 +3043,21 @@ void si_vm_fini(struct radeon_device *rdev)
  * si_vm_set_page - update the page tables using the CP
  *
  * @rdev: radeon_device pointer
+ * @ib: indirect buffer to fill with commands
  * @pe: addr of the page entry
  * @addr: dst addr to write into pe
  * @count: number of page entries to update
  * @incr: increase next addr by incr bytes
  * @flags: access flags
  *
- * Update the page tables using the CP (cayman-si).
+ * Update the page tables using the CP (SI).
  */
-void si_vm_set_page(struct radeon_device *rdev, uint64_t pe,
+void si_vm_set_page(struct radeon_device *rdev,
+		    struct radeon_ib *ib,
+		    uint64_t pe,
 		    uint64_t addr, unsigned count,
 		    uint32_t incr, uint32_t flags)
 {
-	struct radeon_ring *ring = &rdev->ring[rdev->asic->vm.pt_ring_index];
 	uint32_t r600_flags = cayman_vm_page_flags(rdev, flags);
 	uint64_t value;
 	unsigned ndw;
@@ -3066,11 +3068,11 @@ void si_vm_set_page(struct radeon_device *rdev, uint64_t pe,
 			if (ndw > 0x3FFE)
 				ndw = 0x3FFE;
 
-			radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, ndw));
-			radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
-						 WRITE_DATA_DST_SEL(1)));
-			radeon_ring_write(ring, pe);
-			radeon_ring_write(ring, upper_32_bits(pe));
+			ib->ptr[ib->length_dw++] = PACKET3(PACKET3_WRITE_DATA, ndw);
+			ib->ptr[ib->length_dw++] = (WRITE_DATA_ENGINE_SEL(0) |
+					WRITE_DATA_DST_SEL(1));
+			ib->ptr[ib->length_dw++] = pe;
+			ib->ptr[ib->length_dw++] = upper_32_bits(pe);
 			for (; ndw > 2; ndw -= 2, --count, pe += 8) {
 				if (flags & RADEON_VM_PAGE_SYSTEM) {
 					value = radeon_vm_map_gart(rdev, addr);
@@ -3082,8 +3084,8 @@ void si_vm_set_page(struct radeon_device *rdev, uint64_t pe,
 				}
 				addr += incr;
 				value |= r600_flags;
-				radeon_ring_write(ring, value);
-				radeon_ring_write(ring, upper_32_bits(value));
+				ib->ptr[ib->length_dw++] = value;
+				ib->ptr[ib->length_dw++] = upper_32_bits(value);
 			}
 		}
 	} else {
@@ -3095,9 +3097,9 @@ void si_vm_set_page(struct radeon_device *rdev, uint64_t pe,
 					ndw = 0xFFFFE;
 
 				/* for non-physically contiguous pages (system) */
-				radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 0, ndw));
-				radeon_ring_write(ring, pe);
-				radeon_ring_write(ring, upper_32_bits(pe) & 0xff);
+				ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 0, ndw);
+				ib->ptr[ib->length_dw++] = pe;
+				ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
 				for (; ndw > 0; ndw -= 2, --count, pe += 8) {
 					if (flags & RADEON_VM_PAGE_SYSTEM) {
 						value = radeon_vm_map_gart(rdev, addr);
@@ -3109,8 +3111,8 @@ void si_vm_set_page(struct radeon_device *rdev, uint64_t pe,
 					}
 					addr += incr;
 					value |= r600_flags;
-					radeon_ring_write(ring, value);
-					radeon_ring_write(ring, upper_32_bits(value));
+					ib->ptr[ib->length_dw++] = value;
+					ib->ptr[ib->length_dw++] = upper_32_bits(value);
 				}
 			}
 		} else {
@@ -3124,20 +3126,22 @@ void si_vm_set_page(struct radeon_device *rdev, uint64_t pe,
 				else
 					value = 0;
 				/* for physically contiguous pages (vram) */
-				radeon_ring_write(ring, DMA_PTE_PDE_PACKET(ndw));
-				radeon_ring_write(ring, pe); /* dst addr */
-				radeon_ring_write(ring, upper_32_bits(pe) & 0xff);
-				radeon_ring_write(ring, r600_flags); /* mask */
-				radeon_ring_write(ring, 0);
-				radeon_ring_write(ring, value); /* value */
-				radeon_ring_write(ring, upper_32_bits(value));
-				radeon_ring_write(ring, incr); /* increment size */
-				radeon_ring_write(ring, 0);
+				ib->ptr[ib->length_dw++] = DMA_PTE_PDE_PACKET(ndw);
+				ib->ptr[ib->length_dw++] = pe; /* dst addr */
+				ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
+				ib->ptr[ib->length_dw++] = r600_flags; /* mask */
+				ib->ptr[ib->length_dw++] = 0;
+				ib->ptr[ib->length_dw++] = value; /* value */
+				ib->ptr[ib->length_dw++] = upper_32_bits(value);
+				ib->ptr[ib->length_dw++] = incr; /* increment size */
+				ib->ptr[ib->length_dw++] = 0;
 				pe += ndw * 4;
 				addr += (ndw / 2) * incr;
 				count -= ndw / 2;
 			}
 		}
+		while (ib->length_dw & 0x7)
+			ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0, 0);
 	}
 }
 
-- 
cgit v0.10.2


From df16004447fe55dcd6f05f6bb47bf8eaeddb160c Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Thu, 31 Jan 2013 16:26:02 -0500
Subject: drm/radeon: switch back to using the DMA ring for VM PT updates

Now that we have switched to using IBs for page table updates,
we can switch back the using the DMA ring.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/radeon_asic.c b/drivers/gpu/drm/radeon/radeon_asic.c
index 82b5ef0..60edad7 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.c
+++ b/drivers/gpu/drm/radeon/radeon_asic.c
@@ -1445,7 +1445,7 @@ static struct radeon_asic cayman_asic = {
 	.vm = {
 		.init = &cayman_vm_init,
 		.fini = &cayman_vm_fini,
-		.pt_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.pt_ring_index = R600_RING_TYPE_DMA_INDEX,
 		.set_page = &cayman_vm_set_page,
 	},
 	.ring = {
@@ -1572,7 +1572,7 @@ static struct radeon_asic trinity_asic = {
 	.vm = {
 		.init = &cayman_vm_init,
 		.fini = &cayman_vm_fini,
-		.pt_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.pt_ring_index = R600_RING_TYPE_DMA_INDEX,
 		.set_page = &cayman_vm_set_page,
 	},
 	.ring = {
@@ -1699,7 +1699,7 @@ static struct radeon_asic si_asic = {
 	.vm = {
 		.init = &si_vm_init,
 		.fini = &si_vm_fini,
-		.pt_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.pt_ring_index = R600_RING_TYPE_DMA_INDEX,
 		.set_page = &si_vm_set_page,
 	},
 	.ring = {
-- 
cgit v0.10.2


From 624d35242a90cb508064c11c1ff3beb54f08a6a5 Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Tue, 18 Dec 2012 17:01:35 -0500
Subject: drm/radeon: add Oland chip family

Oland is a new asic in the SI family.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 307d681..59bfbd3 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -1769,6 +1769,7 @@ void r100_pll_errata_after_index(struct radeon_device *rdev);
 #define ASIC_IS_DCE6(rdev) ((rdev->family >= CHIP_ARUBA))
 #define ASIC_IS_DCE61(rdev) ((rdev->family >= CHIP_ARUBA) && \
 			     (rdev->flags & RADEON_IS_IGP))
+#define ASIC_IS_DCE64(rdev) ((rdev->family == CHIP_OLAND))
 
 /*
  * BIOS helpers.
diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c
index edfc54e..6a36f4b 100644
--- a/drivers/gpu/drm/radeon/radeon_device.c
+++ b/drivers/gpu/drm/radeon/radeon_device.c
@@ -93,6 +93,7 @@ static const char radeon_family_name[][16] = {
 	"TAHITI",
 	"PITCAIRN",
 	"VERDE",
+	"OLAND",
 	"LAST",
 };
 
diff --git a/drivers/gpu/drm/radeon/radeon_family.h b/drivers/gpu/drm/radeon/radeon_family.h
index d1fafea..2d91123 100644
--- a/drivers/gpu/drm/radeon/radeon_family.h
+++ b/drivers/gpu/drm/radeon/radeon_family.h
@@ -91,6 +91,7 @@ enum radeon_family {
 	CHIP_TAHITI,
 	CHIP_PITCAIRN,
 	CHIP_VERDE,
+	CHIP_OLAND,
 	CHIP_LAST,
 };
 
-- 
cgit v0.10.2


From d0ae7fccb3bae2f3749aa70636c537aa16f37bb7 Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Thu, 26 Jul 2012 17:42:25 -0400
Subject: drm/radeon: fill in gpu init for Oland

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
index a910cb9..ec62110 100644
--- a/drivers/gpu/drm/radeon/si.c
+++ b/drivers/gpu/drm/radeon/si.c
@@ -1127,7 +1127,8 @@ static void si_tiling_mode_table_init(struct radeon_device *rdev)
 			}
 			WREG32(GB_TILE_MODE0 + (reg_offset * 4), gb_tile_moden);
 		}
-	} else if (rdev->family == CHIP_VERDE) {
+	} else if ((rdev->family == CHIP_VERDE) ||
+		   (rdev->family == CHIP_OLAND)) {
 		for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) {
 			switch (reg_offset) {
 			case 0:  /* non-AA compressed depth or any compressed stencil */
@@ -1572,6 +1573,23 @@ static void si_gpu_init(struct radeon_device *rdev)
 		rdev->config.si.sc_earlyz_tile_fifo_size = 0x130;
 		gb_addr_config = VERDE_GB_ADDR_CONFIG_GOLDEN;
 		break;
+	case CHIP_OLAND:
+		rdev->config.si.max_shader_engines = 1;
+		rdev->config.si.max_tile_pipes = 4;
+		rdev->config.si.max_cu_per_sh = 6;
+		rdev->config.si.max_sh_per_se = 1;
+		rdev->config.si.max_backends_per_se = 2;
+		rdev->config.si.max_texture_channel_caches = 4;
+		rdev->config.si.max_gprs = 256;
+		rdev->config.si.max_gs_threads = 16;
+		rdev->config.si.max_hw_contexts = 8;
+
+		rdev->config.si.sc_prim_fifo_size_frontend = 0x20;
+		rdev->config.si.sc_prim_fifo_size_backend = 0x40;
+		rdev->config.si.sc_hiz_tile_fifo_size = 0x30;
+		rdev->config.si.sc_earlyz_tile_fifo_size = 0x130;
+		gb_addr_config = VERDE_GB_ADDR_CONFIG_GOLDEN;
+		break;
 	}
 
 	/* Initialize HDP */
-- 
cgit v0.10.2


From bcc7f5d24dddfbe75893e25d84ebf9d7579c34c1 Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Thu, 26 Jul 2012 18:36:28 -0400
Subject: drm/radeon: add ucode loading support for Oland

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
index ec62110..719f03e 100644
--- a/drivers/gpu/drm/radeon/si.c
+++ b/drivers/gpu/drm/radeon/si.c
@@ -38,6 +38,7 @@
 #define SI_CE_UCODE_SIZE 2144
 #define SI_RLC_UCODE_SIZE 2048
 #define SI_MC_UCODE_SIZE 7769
+#define OLAND_MC_UCODE_SIZE 7863
 
 MODULE_FIRMWARE("radeon/TAHITI_pfp.bin");
 MODULE_FIRMWARE("radeon/TAHITI_me.bin");
@@ -54,6 +55,11 @@ MODULE_FIRMWARE("radeon/VERDE_me.bin");
 MODULE_FIRMWARE("radeon/VERDE_ce.bin");
 MODULE_FIRMWARE("radeon/VERDE_mc.bin");
 MODULE_FIRMWARE("radeon/VERDE_rlc.bin");
+MODULE_FIRMWARE("radeon/OLAND_pfp.bin");
+MODULE_FIRMWARE("radeon/OLAND_me.bin");
+MODULE_FIRMWARE("radeon/OLAND_ce.bin");
+MODULE_FIRMWARE("radeon/OLAND_mc.bin");
+MODULE_FIRMWARE("radeon/OLAND_rlc.bin");
 
 extern int r600_ih_ring_alloc(struct radeon_device *rdev);
 extern void r600_ih_ring_fini(struct radeon_device *rdev);
@@ -202,6 +208,45 @@ static const u32 verde_io_mc_regs[TAHITI_IO_MC_REGS_SIZE][2] = {
 	{0x0000009f, 0x00a37400}
 };
 
+static const u32 oland_io_mc_regs[TAHITI_IO_MC_REGS_SIZE][2] = {
+	{0x0000006f, 0x03044000},
+	{0x00000070, 0x0480c018},
+	{0x00000071, 0x00000040},
+	{0x00000072, 0x01000000},
+	{0x00000074, 0x000000ff},
+	{0x00000075, 0x00143400},
+	{0x00000076, 0x08ec0800},
+	{0x00000077, 0x040000cc},
+	{0x00000079, 0x00000000},
+	{0x0000007a, 0x21000409},
+	{0x0000007c, 0x00000000},
+	{0x0000007d, 0xe8000000},
+	{0x0000007e, 0x044408a8},
+	{0x0000007f, 0x00000003},
+	{0x00000080, 0x00000000},
+	{0x00000081, 0x01000000},
+	{0x00000082, 0x02000000},
+	{0x00000083, 0x00000000},
+	{0x00000084, 0xe3f3e4f4},
+	{0x00000085, 0x00052024},
+	{0x00000087, 0x00000000},
+	{0x00000088, 0x66036603},
+	{0x00000089, 0x01000000},
+	{0x0000008b, 0x1c0a0000},
+	{0x0000008c, 0xff010000},
+	{0x0000008e, 0xffffefff},
+	{0x0000008f, 0xfff3efff},
+	{0x00000090, 0xfff3efbf},
+	{0x00000094, 0x00101101},
+	{0x00000095, 0x00000fff},
+	{0x00000096, 0x00116fff},
+	{0x00000097, 0x60010000},
+	{0x00000098, 0x10010000},
+	{0x00000099, 0x00006000},
+	{0x0000009a, 0x00001000},
+	{0x0000009f, 0x00a17730}
+};
+
 /* ucode loading */
 static int si_mc_load_microcode(struct radeon_device *rdev)
 {
@@ -230,6 +275,11 @@ static int si_mc_load_microcode(struct radeon_device *rdev)
 		ucode_size = SI_MC_UCODE_SIZE;
 		regs_size = TAHITI_IO_MC_REGS_SIZE;
 		break;
+	case CHIP_OLAND:
+		io_mc_regs = (u32 *)&oland_io_mc_regs;
+		ucode_size = OLAND_MC_UCODE_SIZE;
+		regs_size = TAHITI_IO_MC_REGS_SIZE;
+		break;
 	}
 
 	running = RREG32(MC_SEQ_SUP_CNTL) & RUN_MASK;
@@ -324,6 +374,15 @@ static int si_init_microcode(struct radeon_device *rdev)
 		rlc_req_size = SI_RLC_UCODE_SIZE * 4;
 		mc_req_size = SI_MC_UCODE_SIZE * 4;
 		break;
+	case CHIP_OLAND:
+		chip_name = "OLAND";
+		rlc_chip_name = "OLAND";
+		pfp_req_size = SI_PFP_UCODE_SIZE * 4;
+		me_req_size = SI_PM4_UCODE_SIZE * 4;
+		ce_req_size = SI_CE_UCODE_SIZE * 4;
+		rlc_req_size = SI_RLC_UCODE_SIZE * 4;
+		mc_req_size = OLAND_MC_UCODE_SIZE * 4;
+		break;
 	default: BUG();
 	}
 
-- 
cgit v0.10.2


From e737a14cd1f30a0bca753ae32ebada10adf93c3b Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Thu, 30 Aug 2012 14:00:03 -0400
Subject: drm/radeon: radeon-asic updates for Oland

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/radeon_asic.c b/drivers/gpu/drm/radeon/radeon_asic.c
index 60edad7..67f008f 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.c
+++ b/drivers/gpu/drm/radeon/radeon_asic.c
@@ -1944,9 +1944,13 @@ int radeon_asic_init(struct radeon_device *rdev)
 	case CHIP_TAHITI:
 	case CHIP_PITCAIRN:
 	case CHIP_VERDE:
+	case CHIP_OLAND:
 		rdev->asic = &si_asic;
 		/* set num crtcs */
-		rdev->num_crtc = 6;
+		if (rdev->family == CHIP_OLAND)
+			rdev->num_crtc = 2;
+		else
+			rdev->num_crtc = 6;
 		break;
 	default:
 		/* FIXME: not supported yet */
-- 
cgit v0.10.2


From 7271c7f999609787c4617c51750d39e447859e7d Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Thu, 26 Jul 2012 18:48:01 -0400
Subject: drm/radeon: add Oland pci ids

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/include/drm/drm_pciids.h b/include/drm/drm_pciids.h
index c5c35e6..a386b0b 100644
--- a/include/drm/drm_pciids.h
+++ b/include/drm/drm_pciids.h
@@ -139,6 +139,19 @@
 	{0x1002, 0x5e4c, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RV410|RADEON_NEW_MEMMAP}, \
 	{0x1002, 0x5e4d, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RV410|RADEON_NEW_MEMMAP}, \
 	{0x1002, 0x5e4f, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RV410|RADEON_NEW_MEMMAP}, \
+	{0x1002, 0x6600, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \
+	{0x1002, 0x6601, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \
+	{0x1002, 0x6602, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \
+	{0x1002, 0x6603, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \
+	{0x1002, 0x6606, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \
+	{0x1002, 0x6607, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \
+	{0x1002, 0x6610, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|RADEON_NEW_MEMMAP}, \
+	{0x1002, 0x6611, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|RADEON_NEW_MEMMAP}, \
+	{0x1002, 0x6613, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|RADEON_NEW_MEMMAP}, \
+	{0x1002, 0x6620, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \
+	{0x1002, 0x6621, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \
+	{0x1002, 0x6623, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \
+	{0x1002, 0x6631, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|RADEON_NEW_MEMMAP}, \
 	{0x1002, 0x6700, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_CAYMAN|RADEON_NEW_MEMMAP}, \
 	{0x1002, 0x6701, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_CAYMAN|RADEON_NEW_MEMMAP}, \
 	{0x1002, 0x6702, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_CAYMAN|RADEON_NEW_MEMMAP}, \
-- 
cgit v0.10.2


From 0e3d50bfcbd338254795a700dcff429a96cba1a6 Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Tue, 5 Feb 2013 11:47:09 -0500
Subject: drm/radeon/dce6: fix display powergating

Only enable it when we disable the display rather than
at DPMS time since enabling it requires a full modeset
to restore the display state.  Fixes blank screens in
certain cases.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

diff --git a/drivers/gpu/drm/radeon/atombios_crtc.c b/drivers/gpu/drm/radeon/atombios_crtc.c
index 9175615..21a892c 100644
--- a/drivers/gpu/drm/radeon/atombios_crtc.c
+++ b/drivers/gpu/drm/radeon/atombios_crtc.c
@@ -252,8 +252,6 @@ void atombios_crtc_dpms(struct drm_crtc *crtc, int mode)
 		radeon_crtc->enabled = true;
 		/* adjust pm to dpms changes BEFORE enabling crtcs */
 		radeon_pm_compute_clocks(rdev);
-		if (ASIC_IS_DCE6(rdev) && !radeon_crtc->in_mode_set)
-			atombios_powergate_crtc(crtc, ATOM_DISABLE);
 		atombios_enable_crtc(crtc, ATOM_ENABLE);
 		if (ASIC_IS_DCE3(rdev) && !ASIC_IS_DCE6(rdev))
 			atombios_enable_crtc_memreq(crtc, ATOM_ENABLE);
@@ -271,8 +269,6 @@ void atombios_crtc_dpms(struct drm_crtc *crtc, int mode)
 			atombios_enable_crtc_memreq(crtc, ATOM_DISABLE);
 		atombios_enable_crtc(crtc, ATOM_DISABLE);
 		radeon_crtc->enabled = false;
-		if (ASIC_IS_DCE6(rdev) && !radeon_crtc->in_mode_set)
-			atombios_powergate_crtc(crtc, ATOM_ENABLE);
 		/* adjust pm to dpms changes AFTER disabling crtcs */
 		radeon_pm_compute_clocks(rdev);
 		break;
@@ -1844,6 +1840,8 @@ static void atombios_crtc_disable(struct drm_crtc *crtc)
 	int i;
 
 	atombios_crtc_dpms(crtc, DRM_MODE_DPMS_OFF);
+	if (ASIC_IS_DCE6(rdev))
+		atombios_powergate_crtc(crtc, ATOM_ENABLE);
 
 	for (i = 0; i < rdev->num_crtc; i++) {
 		if (rdev->mode_info.crtcs[i] &&
-- 
cgit v0.10.2


From 2a2483072393b27f4336ab068a1f48ca19ff1c1e Mon Sep 17 00:00:00 2001
From: Dave Airlie <airlied@redhat.com>
Date: Thu, 24 Jan 2013 14:14:19 +1000
Subject: vgacon/vt: clear buffer attributes when we load a 512 character font
 (v2)

When we switch from 256->512 byte font rendering mode, it means the
current contents of the screen is being reinterpreted. The bit that holds
the high bit of the 9-bit font, may have been previously set, and thus
the new font misrenders.

The problem case we see is grub2 writes spaces with the bit set, so it
ends up with data like 0x820, which gets reinterpreted into 0x120 char
which the font translates into G with a circumflex. This flashes up on
screen at boot and is quite ugly.

A current side effect of this patch though is that any rendering on the
screen changes color to a slightly darker color, but at least the screen
no longer corrupts.

v2: as suggested by hpa, always clear the attribute space, whether we
are are going to or from 512 chars.

Cc: stable@vger.kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/tty/vt/vt.c b/drivers/tty/vt/vt.c
index 8fd8968..c8067ae 100644
--- a/drivers/tty/vt/vt.c
+++ b/drivers/tty/vt/vt.c
@@ -638,7 +638,7 @@ static inline void save_screen(struct vc_data *vc)
  *	Redrawing of screen
  */
 
-static void clear_buffer_attributes(struct vc_data *vc)
+void clear_buffer_attributes(struct vc_data *vc)
 {
 	unsigned short *p = (unsigned short *)vc->vc_origin;
 	int count = vc->vc_screenbuf_size / 2;
diff --git a/drivers/video/console/vgacon.c b/drivers/video/console/vgacon.c
index d449a74..5855d17 100644
--- a/drivers/video/console/vgacon.c
+++ b/drivers/video/console/vgacon.c
@@ -1064,7 +1064,7 @@ static int vgacon_do_font_op(struct vgastate *state,char *arg,int set,int ch512)
 	unsigned short video_port_status = vga_video_port_reg + 6;
 	int font_select = 0x00, beg, i;
 	char *charmap;
-	
+	bool clear_attribs = false;
 	if (vga_video_type != VIDEO_TYPE_EGAM) {
 		charmap = (char *) VGA_MAP_MEM(colourmap, 0);
 		beg = 0x0e;
@@ -1169,12 +1169,6 @@ static int vgacon_do_font_op(struct vgastate *state,char *arg,int set,int ch512)
 
 	/* if 512 char mode is already enabled don't re-enable it. */
 	if ((set) && (ch512 != vga_512_chars)) {
-		/* attribute controller */
-		for (i = 0; i < MAX_NR_CONSOLES; i++) {
-			struct vc_data *c = vc_cons[i].d;
-			if (c && c->vc_sw == &vga_con)
-				c->vc_hi_font_mask = ch512 ? 0x0800 : 0;
-		}
 		vga_512_chars = ch512;
 		/* 256-char: enable intensity bit
 		   512-char: disable intensity bit */
@@ -1185,8 +1179,22 @@ static int vgacon_do_font_op(struct vgastate *state,char *arg,int set,int ch512)
 		   it means, but it works, and it appears necessary */
 		inb_p(video_port_status);
 		vga_wattr(state->vgabase, VGA_AR_ENABLE_DISPLAY, 0);	
+		clear_attribs = true;
 	}
 	raw_spin_unlock_irq(&vga_lock);
+
+	if (clear_attribs) {
+		for (i = 0; i < MAX_NR_CONSOLES; i++) {
+			struct vc_data *c = vc_cons[i].d;
+			if (c && c->vc_sw == &vga_con) {
+				/* force hi font mask to 0, so we always clear
+				   the bit on either transition */
+				c->vc_hi_font_mask = 0x00;
+				clear_buffer_attributes(c);
+				c->vc_hi_font_mask = ch512 ? 0x0800 : 0;
+			}
+		}
+	}
 	return 0;
 }
 
diff --git a/include/linux/vt_kern.h b/include/linux/vt_kern.h
index 50ae7d0..1f55665 100644
--- a/include/linux/vt_kern.h
+++ b/include/linux/vt_kern.h
@@ -47,6 +47,7 @@ int con_set_cmap(unsigned char __user *cmap);
 int con_get_cmap(unsigned char __user *cmap);
 void scrollback(struct vc_data *vc, int lines);
 void scrollfront(struct vc_data *vc, int lines);
+void clear_buffer_attributes(struct vc_data *vc);
 void update_region(struct vc_data *vc, unsigned long start, int count);
 void redraw_screen(struct vc_data *vc, int is_switch);
 #define update_screen(x) redraw_screen(x, 0)
-- 
cgit v0.10.2


From ae1287865f5361fa138d4d3b1b6277908b54eac9 Mon Sep 17 00:00:00 2001
From: Dave Airlie <airlied@redhat.com>
Date: Thu, 24 Jan 2013 16:12:41 +1000
Subject: fbcon: don't lose the console font across generic->chip driver switch

If grub2 loads efifb/vesafb, then when systemd starts it can set the console
font on that framebuffer device, however when we then load the native KMS
driver, the first thing it does is tear down the generic framebuffer driver.

The thing is the generic code is doing the right thing, it frees the font
because otherwise it would leak memory. However we can assume that if you
are removing the generic firmware driver (vesa/efi/offb), that a new driver
*should* be loading soon after, so we effectively leak the font.

However the old code left a dangling pointer in vc->vc_font.data and we
can now reuse that dangling pointer to load the font into the new
driver, now that we aren't freeing it.

Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=892340

Cc: Kay Sievers <kay.sievers@vrfy.org>
Cc: stable@vger.kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/video/console/fbcon.c b/drivers/video/console/fbcon.c
index fdefa8f..f8dbe76 100644
--- a/drivers/video/console/fbcon.c
+++ b/drivers/video/console/fbcon.c
@@ -990,7 +990,7 @@ static const char *fbcon_startup(void)
 	}
 
 	/* Setup default font */
-	if (!p->fontdata) {
+	if (!p->fontdata && !vc->vc_font.data) {
 		if (!fontname[0] || !(font = find_font(fontname)))
 			font = get_default_font(info->var.xres,
 						info->var.yres,
@@ -1000,6 +1000,8 @@ static const char *fbcon_startup(void)
 		vc->vc_font.height = font->height;
 		vc->vc_font.data = (void *)(p->fontdata = font->data);
 		vc->vc_font.charcount = 256; /* FIXME  Need to support more fonts */
+	} else {
+		p->fontdata = vc->vc_font.data;
 	}
 
 	cols = FBCON_SWAP(ops->rotate, info->var.xres, info->var.yres);
@@ -1159,9 +1161,9 @@ static void fbcon_init(struct vc_data *vc, int init)
 	ops->p = &fb_display[fg_console];
 }
 
-static void fbcon_free_font(struct display *p)
+static void fbcon_free_font(struct display *p, bool freefont)
 {
-	if (p->userfont && p->fontdata && (--REFCOUNT(p->fontdata) == 0))
+	if (freefont && p->userfont && p->fontdata && (--REFCOUNT(p->fontdata) == 0))
 		kfree(p->fontdata - FONT_EXTRA_WORDS * sizeof(int));
 	p->fontdata = NULL;
 	p->userfont = 0;
@@ -1173,8 +1175,8 @@ static void fbcon_deinit(struct vc_data *vc)
 	struct fb_info *info;
 	struct fbcon_ops *ops;
 	int idx;
+	bool free_font = true;
 
-	fbcon_free_font(p);
 	idx = con2fb_map[vc->vc_num];
 
 	if (idx == -1)
@@ -1185,6 +1187,8 @@ static void fbcon_deinit(struct vc_data *vc)
 	if (!info)
 		goto finished;
 
+	if (info->flags & FBINFO_MISC_FIRMWARE)
+		free_font = false;
 	ops = info->fbcon_par;
 
 	if (!ops)
@@ -1196,6 +1200,8 @@ static void fbcon_deinit(struct vc_data *vc)
 	ops->flags &= ~FBCON_FLAGS_INIT;
 finished:
 
+	fbcon_free_font(p, free_font);
+
 	if (!con_is_bound(&fb_con))
 		fbcon_exit();
 
-- 
cgit v0.10.2


From 9f23de52b64f7fb801fd76f3dd8651a0dc89187b Mon Sep 17 00:00:00 2001
From: Dave Airlie <airlied@redhat.com>
Date: Thu, 7 Feb 2013 10:10:04 +1000
Subject: drm/usb: bind driver to correct device

While looking at plymouth on udl I noticed that plymouth was trying
to use its fb plugin not its drm one, it was trying to drmOpen a driver called
usb not udl, noticed that we actually had out driver pointing at the wrong
device.

Cc: stable@vger.kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/gpu/drm/drm_usb.c b/drivers/gpu/drm/drm_usb.c
index 3cec306..34a156f 100644
--- a/drivers/gpu/drm/drm_usb.c
+++ b/drivers/gpu/drm/drm_usb.c
@@ -18,7 +18,7 @@ int drm_get_usb_dev(struct usb_interface *interface,
 
 	usbdev = interface_to_usbdev(interface);
 	dev->usbdev = usbdev;
-	dev->dev = &usbdev->dev;
+	dev->dev = &interface->dev;
 
 	mutex_lock(&drm_global_mutex);
 
-- 
cgit v0.10.2


From bcb39af4486be07e896fc374a2336bad3104ae0a Mon Sep 17 00:00:00 2001
From: Dave Airlie <airlied@redhat.com>
Date: Thu, 7 Feb 2013 11:19:15 +1000
Subject: drm/udl: make usage as a console safer

Okay you don't really want to use udl devices as your console, but if
you are unlucky enough to do so, you run into a lot of schedule while atomic
due to printk being called from all sorts of funky places. So check if we
are in an atomic context, and queue the damage for later, the next printk
should cause it to appear. This isn't ideal, but it is simple, and seems to
work okay in my testing here.

(dirty area idea came from xenfb)

fixes a bunch of sleeping while atomic issues running fbcon on udl devices.

Cc: stable@vger.kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/gpu/drm/udl/udl_drv.h b/drivers/gpu/drm/udl/udl_drv.h
index 87aa5f5..cc6d90f 100644
--- a/drivers/gpu/drm/udl/udl_drv.h
+++ b/drivers/gpu/drm/udl/udl_drv.h
@@ -75,6 +75,8 @@ struct udl_framebuffer {
 	struct drm_framebuffer base;
 	struct udl_gem_object *obj;
 	bool active_16; /* active on the 16-bit channel */
+	int x1, y1, x2, y2; /* dirty rect */
+	spinlock_t dirty_lock;
 };
 
 #define to_udl_fb(x) container_of(x, struct udl_framebuffer, base)
diff --git a/drivers/gpu/drm/udl/udl_fb.c b/drivers/gpu/drm/udl/udl_fb.c
index d4ab3be..c35880f 100644
--- a/drivers/gpu/drm/udl/udl_fb.c
+++ b/drivers/gpu/drm/udl/udl_fb.c
@@ -153,6 +153,9 @@ int udl_handle_damage(struct udl_framebuffer *fb, int x, int y,
 	struct urb *urb;
 	int aligned_x;
 	int bpp = (fb->base.bits_per_pixel / 8);
+	int x2, y2;
+	bool store_for_later = false;
+	unsigned long flags;
 
 	if (!fb->active_16)
 		return 0;
@@ -169,8 +172,6 @@ int udl_handle_damage(struct udl_framebuffer *fb, int x, int y,
 		}
 	}
 
-	start_cycles = get_cycles();
-
 	aligned_x = DL_ALIGN_DOWN(x, sizeof(unsigned long));
 	width = DL_ALIGN_UP(width + (x-aligned_x), sizeof(unsigned long));
 	x = aligned_x;
@@ -180,19 +181,53 @@ int udl_handle_damage(struct udl_framebuffer *fb, int x, int y,
 	    (y + height > fb->base.height))
 		return -EINVAL;
 
+	/* if we are in atomic just store the info
+	   can't test inside spin lock */
+	if (in_atomic())
+		store_for_later = true;
+
+	x2 = x + width - 1;
+	y2 = y + height - 1;
+
+	spin_lock_irqsave(&fb->dirty_lock, flags);
+
+	if (fb->y1 < y)
+		y = fb->y1;
+	if (fb->y2 > y2)
+		y2 = fb->y2;
+	if (fb->x1 < x)
+		x = fb->x1;
+	if (fb->x2 > x2)
+		x2 = fb->x2;
+
+	if (store_for_later) {
+		fb->x1 = x;
+		fb->x2 = x2;
+		fb->y1 = y;
+		fb->y2 = y2;
+		spin_unlock_irqrestore(&fb->dirty_lock, flags);
+		return 0;
+	}
+
+	fb->x1 = fb->y1 = INT_MAX;
+	fb->x2 = fb->y2 = 0;
+
+	spin_unlock_irqrestore(&fb->dirty_lock, flags);
+	start_cycles = get_cycles();
+
 	urb = udl_get_urb(dev);
 	if (!urb)
 		return 0;
 	cmd = urb->transfer_buffer;
 
-	for (i = y; i < y + height ; i++) {
+	for (i = y; i <= y2 ; i++) {
 		const int line_offset = fb->base.pitches[0] * i;
 		const int byte_offset = line_offset + (x * bpp);
 		const int dev_byte_offset = (fb->base.width * bpp * i) + (x * bpp);
 		if (udl_render_hline(dev, bpp, &urb,
 				     (char *) fb->obj->vmapping,
 				     &cmd, byte_offset, dev_byte_offset,
-				     width * bpp,
+				     (x2 - x + 1) * bpp,
 				     &bytes_identical, &bytes_sent))
 			goto error;
 	}
@@ -434,6 +469,7 @@ udl_framebuffer_init(struct drm_device *dev,
 {
 	int ret;
 
+	spin_lock_init(&ufb->dirty_lock);
 	ufb->obj = obj;
 	ret = drm_framebuffer_init(dev, &ufb->base, &udlfb_funcs);
 	drm_helper_mode_fill_fb_struct(&ufb->base, mode_cmd);
-- 
cgit v0.10.2


From e90a4ea534b110a43df87a05587c53cd78569467 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Fri, 18 Jan 2013 16:31:14 +0000
Subject: drm/udl: Inline memcmp() for RLE compression of xfer

As we use a variable length the compiler does not realise that it is a
fixed value of either 2 or 4 bytes. Instead of performing the inline
comparison itself, the compiler inserts a function call to the generic
memcmp routine which is optimised for long comparisons of variable
length. That turns out to be quite expensive...

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/gpu/drm/udl/udl_transfer.c b/drivers/gpu/drm/udl/udl_transfer.c
index 142fee5..f343db7 100644
--- a/drivers/gpu/drm/udl/udl_transfer.c
+++ b/drivers/gpu/drm/udl/udl_transfer.c
@@ -75,15 +75,19 @@ static int udl_trim_hline(const u8 *bback, const u8 **bfront, int *width_bytes)
 }
 #endif
 
-static inline u16 pixel32_to_be16p(const uint8_t *pixel)
+static inline u16 pixel32_to_be16(const uint32_t pixel)
 {
-	uint32_t pix = *(uint32_t *)pixel;
-	u16 retval;
+	return (((pixel >> 3) & 0x001f) |
+		((pixel >> 5) & 0x07e0) |
+		((pixel >> 8) & 0xf800));
+}
 
-	retval =  (((pix >> 3) & 0x001f) |
-		   ((pix >> 5) & 0x07e0) |
-		   ((pix >> 8) & 0xf800));
-	return retval;
+static bool pixel_repeats(const void *pixel, const uint32_t repeat, int bpp)
+{
+	if (bpp == 2)
+		return *(const uint16_t *)pixel == repeat;
+	else
+		return *(const uint32_t *)pixel == repeat;
 }
 
 /*
@@ -152,29 +156,33 @@ static void udl_compress_hline16(
 		prefetch_range((void *) pixel, (cmd_pixel_end - pixel) * bpp);
 
 		while (pixel < cmd_pixel_end) {
-			const u8 * const repeating_pixel = pixel;
-
-			if (bpp == 2)
-				*(uint16_t *)cmd = cpu_to_be16p((uint16_t *)pixel);
-			else if (bpp == 4)
-				*(uint16_t *)cmd = cpu_to_be16(pixel32_to_be16p(pixel));
+			const u8 *const start = pixel;
+			u32 repeating_pixel;
+
+			if (bpp == 2) {
+				repeating_pixel = *(uint16_t *)pixel;
+				*(uint16_t *)cmd = cpu_to_be16(repeating_pixel);
+			} else {
+				repeating_pixel = *(uint32_t *)pixel;
+				*(uint16_t *)cmd = cpu_to_be16(pixel32_to_be16(repeating_pixel));
+			}
 
 			cmd += 2;
 			pixel += bpp;
 
 			if (unlikely((pixel < cmd_pixel_end) &&
-				     (!memcmp(pixel, repeating_pixel, bpp)))) {
+				     (pixel_repeats(pixel, repeating_pixel, bpp)))) {
 				/* go back and fill in raw pixel count */
-				*raw_pixels_count_byte = (((repeating_pixel -
+				*raw_pixels_count_byte = (((start -
 						raw_pixel_start) / bpp) + 1) & 0xFF;
 
-				while ((pixel < cmd_pixel_end)
-				       && (!memcmp(pixel, repeating_pixel, bpp))) {
+				while ((pixel < cmd_pixel_end) &&
+				       (pixel_repeats(pixel, repeating_pixel, bpp))) {
 					pixel += bpp;
 				}
 
 				/* immediately after raw data is repeat byte */
-				*cmd++ = (((pixel - repeating_pixel) / bpp) - 1) & 0xFF;
+				*cmd++ = (((pixel - start) / bpp) - 1) & 0xFF;
 
 				/* Then start another raw pixel span */
 				raw_pixel_start = pixel;
@@ -223,6 +231,8 @@ int udl_render_hline(struct drm_device *dev, int bpp, struct urb **urb_ptr,
 	u8 *cmd = *urb_buf_ptr;
 	u8 *cmd_end = (u8 *) urb->transfer_buffer + urb->transfer_buffer_length;
 
+	BUG_ON(!(bpp == 2 || bpp == 4));
+
 	line_start = (u8 *) (front + byte_offset);
 	next_pixel = line_start;
 	line_end = next_pixel + byte_width;
-- 
cgit v0.10.2


From 677d23b70bf949f75746c80cbae92c233c6b5e2a Mon Sep 17 00:00:00 2001
From: Dave Airlie <airlied@redhat.com>
Date: Thu, 7 Feb 2013 12:30:25 +1000
Subject: drm/udl: disable fb_defio by default

There seems to be a bad interaction between gem/shmem and defio on top,
I get list corruption on the page lru in the shmem code.

Turn it off for now until we get some more digging done.

Cc: stable@vger.kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/gpu/drm/udl/udl_fb.c b/drivers/gpu/drm/udl/udl_fb.c
index c35880f..1eb060c 100644
--- a/drivers/gpu/drm/udl/udl_fb.c
+++ b/drivers/gpu/drm/udl/udl_fb.c
@@ -22,9 +22,9 @@
 
 #include <drm/drm_fb_helper.h>
 
-#define DL_DEFIO_WRITE_DELAY    5 /* fb_deferred_io.delay in jiffies */
+#define DL_DEFIO_WRITE_DELAY    (HZ/20) /* fb_deferred_io.delay in jiffies */
 
-static int fb_defio = 1;  /* Optionally enable experimental fb_defio mmap support */
+static int fb_defio = 0;  /* Optionally enable experimental fb_defio mmap support */
 static int fb_bpp = 16;
 
 module_param(fb_bpp, int, S_IWUSR | S_IRUSR | S_IWGRP | S_IRGRP);
-- 
cgit v0.10.2


From 313751fa7bb9364f320727c2af0735dc9104d188 Mon Sep 17 00:00:00 2001
From: Jan Beulich <JBeulich@suse.com>
Date: Wed, 30 Jan 2013 07:55:53 +0000
Subject: x86-64: Replace left over sti/cli in ia32 audit exit code

For some reason they didn't get replaced so far by their
paravirt equivalents, resulting in code to be run with
interrupts disabled that doesn't expect so (causing, in the
observed case, a BUG_ON() to trigger) when syscall auditing is
enabled.

David (Cc-ed) came up with an identical fix, so likely this can
be taken to count as an ack from him.

Reported-by: Peter Moody <pmoody@google.com>
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Cc: David Vrabel <david.vrabel@citrix.com>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Link: http://lkml.kernel.org/r/5108E01902000078000BA9C5@nat28.tlf.novell.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: stable@vger.kernel.org
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: David Vrabel <david.vrabel@citrix.com>
Tested-by: Peter Moody <pmoody@google.com>

diff --git a/arch/x86/ia32/ia32entry.S b/arch/x86/ia32/ia32entry.S
index 102ff7c..142c4ce 100644
--- a/arch/x86/ia32/ia32entry.S
+++ b/arch/x86/ia32/ia32entry.S
@@ -207,7 +207,7 @@ sysexit_from_sys_call:
 	testl $(_TIF_ALLWORK_MASK & ~_TIF_SYSCALL_AUDIT),TI_flags+THREAD_INFO(%rsp,RIP-ARGOFFSET)
 	jnz ia32_ret_from_sys_call
 	TRACE_IRQS_ON
-	sti
+	ENABLE_INTERRUPTS(CLBR_NONE)
 	movl %eax,%esi		/* second arg, syscall return value */
 	cmpl $-MAX_ERRNO,%eax	/* is it an error ? */
 	jbe 1f
@@ -217,7 +217,7 @@ sysexit_from_sys_call:
 	call __audit_syscall_exit
 	movq RAX-ARGOFFSET(%rsp),%rax	/* reload syscall return value */
 	movl $(_TIF_ALLWORK_MASK & ~_TIF_SYSCALL_AUDIT),%edi
-	cli
+	DISABLE_INTERRUPTS(CLBR_NONE)
 	TRACE_IRQS_OFF
 	testl %edi,TI_flags+THREAD_INFO(%rsp,RIP-ARGOFFSET)
 	jz \exit
-- 
cgit v0.10.2


From 521dfda999fd4cfe801c49d28ccc09350c0389f5 Mon Sep 17 00:00:00 2001
From: "H. Peter Anvin" <hpa@linux.intel.com>
Date: Thu, 31 Jan 2013 20:23:49 -0800
Subject: x86, doc: Boot protocol 2.12 is in 3.8

The boot protocol 2.12 changes were pulled for 3.8, so update the
documentation accordingly.

Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>

diff --git a/Documentation/x86/boot.txt b/Documentation/x86/boot.txt
index 3edb4c2..e540fd6 100644
--- a/Documentation/x86/boot.txt
+++ b/Documentation/x86/boot.txt
@@ -57,7 +57,7 @@ Protocol 2.10:	(Kernel 2.6.31) Added a protocol for relaxed alignment
 Protocol 2.11:	(Kernel 3.6) Added a field for offset of EFI handover
 		protocol entry point.
 
-Protocol 2.12:	(Kernel 3.9) Added the xloadflags field and extension fields
+Protocol 2.12:	(Kernel 3.8) Added the xloadflags field and extension fields
 	 	to struct boot_params for for loading bzImage and ramdisk
 		above 4G in 64bit.
 
-- 
cgit v0.10.2


From 84b603abd23300df8ee5fb1c23c83634c2aaedea Mon Sep 17 00:00:00 2001
From: Borislav Petkov <bp@suse.de>
Date: Mon, 4 Feb 2013 10:13:15 +0100
Subject: x86/intel/cacheinfo: Shut up annoying warning
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

I've been getting the following warning when doing randbuilds
since forever. Now it finally pissed me off just the perfect
amount so that I can fix it.

  arch/x86/kernel/cpu/intel_cacheinfo.c:489:27: warning: ‘cache_disable_0’ defined but not used [-Wunused-variable]
  arch/x86/kernel/cpu/intel_cacheinfo.c:491:27: warning: ‘cache_disable_1’ defined but not used [-Wunused-variable] arch/x86/kernel/cpu/intel_cacheinfo.c:524:27: warning: ‘subcaches’ defined but not used [-Wunused-variable]

It happens because in randconfigs where CONFIG_SYSFS is not set,
the whole sysfs-interface to L3 cache index disabling is
remaining unused and gcc correctly warns about it. Make it
optional, depending on CONFIG_SYSFS too, as is the case with
other sysfs-related machinery in this file.

Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: Andreas Herrmann <andreas.herrmann3@amd.com>
Link: http://lkml.kernel.org/r/1359969195-27362-1-git-send-email-bp@alien8.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>

diff --git a/arch/x86/kernel/cpu/intel_cacheinfo.c b/arch/x86/kernel/cpu/intel_cacheinfo.c
index fe9edec..84c1309c 100644
--- a/arch/x86/kernel/cpu/intel_cacheinfo.c
+++ b/arch/x86/kernel/cpu/intel_cacheinfo.c
@@ -298,8 +298,7 @@ struct _cache_attr {
 			 unsigned int);
 };
 
-#ifdef CONFIG_AMD_NB
-
+#if defined(CONFIG_AMD_NB) && defined(CONFIG_SYSFS)
 /*
  * L3 cache descriptors
  */
@@ -524,9 +523,9 @@ store_subcaches(struct _cpuid4_info *this_leaf, const char *buf, size_t count,
 static struct _cache_attr subcaches =
 	__ATTR(subcaches, 0644, show_subcaches, store_subcaches);
 
-#else	/* CONFIG_AMD_NB */
+#else
 #define amd_init_l3_cache(x, y)
-#endif /* CONFIG_AMD_NB */
+#endif  /* CONFIG_AMD_NB && CONFIG_SYSFS */
 
 static int
 __cpuinit cpuid4_cache_lookup_regs(int index,
-- 
cgit v0.10.2


From 50e244cc793d511b86adea24972f3a7264cae114 Mon Sep 17 00:00:00 2001
From: Alan Cox <alan@linux.intel.com>
Date: Fri, 25 Jan 2013 10:28:15 +1000
Subject: fb: rework locking to fix lock ordering on takeover

Adjust the console layer to allow a take over call where the caller
already holds the locks.  Make the fb layer lock in order.

This is partly a band aid, the fb layer is terminally confused about the
locking rules it uses for its notifiers it seems.

[akpm@linux-foundation.org: remove stray non-ascii char, tidy comment]
[akpm@linux-foundation.org: export do_take_over_console()]
[airlied: cleanup another non-ascii char]
Signed-off-by: Alan Cox <alan@linux.intel.com>
Cc: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Jiri Kosina <jkosina@suse.cz>
Cc: stable <stable@vger.kernel.org>
Tested-by: Sedat Dilek <sedat.dilek@gmail.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/tty/vt/vt.c b/drivers/tty/vt/vt.c
index 8fd8968..c076af0 100644
--- a/drivers/tty/vt/vt.c
+++ b/drivers/tty/vt/vt.c
@@ -2987,7 +2987,7 @@ int __init vty_init(const struct file_operations *console_fops)
 
 static struct class *vtconsole_class;
 
-static int bind_con_driver(const struct consw *csw, int first, int last,
+static int do_bind_con_driver(const struct consw *csw, int first, int last,
 			   int deflt)
 {
 	struct module *owner = csw->owner;
@@ -2998,7 +2998,7 @@ static int bind_con_driver(const struct consw *csw, int first, int last,
 	if (!try_module_get(owner))
 		return -ENODEV;
 
-	console_lock();
+	WARN_CONSOLE_UNLOCKED();
 
 	/* check if driver is registered */
 	for (i = 0; i < MAX_NR_CON_DRIVER; i++) {
@@ -3083,11 +3083,22 @@ static int bind_con_driver(const struct consw *csw, int first, int last,
 
 	retval = 0;
 err:
-	console_unlock();
 	module_put(owner);
 	return retval;
 };
 
+
+static int bind_con_driver(const struct consw *csw, int first, int last,
+			   int deflt)
+{
+	int ret;
+
+	console_lock();
+	ret = do_bind_con_driver(csw, first, last, deflt);
+	console_unlock();
+	return ret;
+}
+
 #ifdef CONFIG_VT_HW_CONSOLE_BINDING
 static int con_is_graphics(const struct consw *csw, int first, int last)
 {
@@ -3199,9 +3210,9 @@ int unbind_con_driver(const struct consw *csw, int first, int last, int deflt)
 	if (!con_is_bound(csw))
 		con_driver->flag &= ~CON_DRIVER_FLAG_INIT;
 
-	console_unlock();
 	/* ignore return value, binding should not fail */
-	bind_con_driver(defcsw, first, last, deflt);
+	do_bind_con_driver(defcsw, first, last, deflt);
+	console_unlock();
 err:
 	module_put(owner);
 	return retval;
@@ -3492,28 +3503,18 @@ int con_debug_leave(void)
 }
 EXPORT_SYMBOL_GPL(con_debug_leave);
 
-/**
- * register_con_driver - register console driver to console layer
- * @csw: console driver
- * @first: the first console to take over, minimum value is 0
- * @last: the last console to take over, maximum value is MAX_NR_CONSOLES -1
- *
- * DESCRIPTION: This function registers a console driver which can later
- * bind to a range of consoles specified by @first and @last. It will
- * also initialize the console driver by calling con_startup().
- */
-int register_con_driver(const struct consw *csw, int first, int last)
+static int do_register_con_driver(const struct consw *csw, int first, int last)
 {
 	struct module *owner = csw->owner;
 	struct con_driver *con_driver;
 	const char *desc;
 	int i, retval = 0;
 
+	WARN_CONSOLE_UNLOCKED();
+
 	if (!try_module_get(owner))
 		return -ENODEV;
 
-	console_lock();
-
 	for (i = 0; i < MAX_NR_CON_DRIVER; i++) {
 		con_driver = &registered_con_driver[i];
 
@@ -3566,10 +3567,29 @@ int register_con_driver(const struct consw *csw, int first, int last)
 	}
 
 err:
-	console_unlock();
 	module_put(owner);
 	return retval;
 }
+
+/**
+ * register_con_driver - register console driver to console layer
+ * @csw: console driver
+ * @first: the first console to take over, minimum value is 0
+ * @last: the last console to take over, maximum value is MAX_NR_CONSOLES -1
+ *
+ * DESCRIPTION: This function registers a console driver which can later
+ * bind to a range of consoles specified by @first and @last. It will
+ * also initialize the console driver by calling con_startup().
+ */
+int register_con_driver(const struct consw *csw, int first, int last)
+{
+	int retval;
+
+	console_lock();
+	retval = do_register_con_driver(csw, first, last);
+	console_unlock();
+	return retval;
+}
 EXPORT_SYMBOL(register_con_driver);
 
 /**
@@ -3623,17 +3643,44 @@ EXPORT_SYMBOL(unregister_con_driver);
  *	when a driver wants to take over some existing consoles
  *	and become default driver for newly opened ones.
  *
- *      take_over_console is basically a register followed by unbind
+ *	take_over_console is basically a register followed by unbind
+ */
+int do_take_over_console(const struct consw *csw, int first, int last, int deflt)
+{
+	int err;
+
+	err = do_register_con_driver(csw, first, last);
+	/*
+	 * If we get an busy error we still want to bind the console driver
+	 * and return success, as we may have unbound the console driver
+	 * but not unregistered it.
+	 */
+	if (err == -EBUSY)
+		err = 0;
+	if (!err)
+		do_bind_con_driver(csw, first, last, deflt);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(do_take_over_console);
+
+/*
+ *	If we support more console drivers, this function is used
+ *	when a driver wants to take over some existing consoles
+ *	and become default driver for newly opened ones.
+ *
+ *	take_over_console is basically a register followed by unbind
  */
 int take_over_console(const struct consw *csw, int first, int last, int deflt)
 {
 	int err;
 
 	err = register_con_driver(csw, first, last);
-	/* if we get an busy error we still want to bind the console driver
+	/*
+	 * If we get an busy error we still want to bind the console driver
 	 * and return success, as we may have unbound the console driver
-	 * but not unregistered it.
-	*/
+	 * but not unregistered it.
+	 */
 	if (err == -EBUSY)
 		err = 0;
 	if (!err)
diff --git a/drivers/video/console/fbcon.c b/drivers/video/console/fbcon.c
index fdefa8f..4bd7820 100644
--- a/drivers/video/console/fbcon.c
+++ b/drivers/video/console/fbcon.c
@@ -529,6 +529,33 @@ static int search_for_mapped_con(void)
 	return retval;
 }
 
+static int do_fbcon_takeover(int show_logo)
+{
+	int err, i;
+
+	if (!num_registered_fb)
+		return -ENODEV;
+
+	if (!show_logo)
+		logo_shown = FBCON_LOGO_DONTSHOW;
+
+	for (i = first_fb_vc; i <= last_fb_vc; i++)
+		con2fb_map[i] = info_idx;
+
+	err = do_take_over_console(&fb_con, first_fb_vc, last_fb_vc,
+				fbcon_is_default);
+
+	if (err) {
+		for (i = first_fb_vc; i <= last_fb_vc; i++)
+			con2fb_map[i] = -1;
+		info_idx = -1;
+	} else {
+		fbcon_has_console_bind = 1;
+	}
+
+	return err;
+}
+
 static int fbcon_takeover(int show_logo)
 {
 	int err, i;
@@ -3115,7 +3142,7 @@ static int fbcon_fb_registered(struct fb_info *info)
 		}
 
 		if (info_idx != -1)
-			ret = fbcon_takeover(1);
+			ret = do_fbcon_takeover(1);
 	} else {
 		for (i = first_fb_vc; i <= last_fb_vc; i++) {
 			if (con2fb_map_boot[i] == idx)
diff --git a/drivers/video/fbmem.c b/drivers/video/fbmem.c
index 3ff0105..d8d9831 100644
--- a/drivers/video/fbmem.c
+++ b/drivers/video/fbmem.c
@@ -1650,7 +1650,9 @@ static int do_register_framebuffer(struct fb_info *fb_info)
 	event.info = fb_info;
 	if (!lock_fb_info(fb_info))
 		return -ENODEV;
+	console_lock();
 	fb_notifier_call_chain(FB_EVENT_FB_REGISTERED, &event);
+	console_unlock();
 	unlock_fb_info(fb_info);
 	return 0;
 }
@@ -1853,11 +1855,8 @@ int fb_new_modelist(struct fb_info *info)
 	err = 1;
 
 	if (!list_empty(&info->modelist)) {
-		if (!lock_fb_info(info))
-			return -ENODEV;
 		event.info = info;
 		err = fb_notifier_call_chain(FB_EVENT_NEW_MODELIST, &event);
-		unlock_fb_info(info);
 	}
 
 	return err;
diff --git a/drivers/video/fbsysfs.c b/drivers/video/fbsysfs.c
index a55e366..ef476b0 100644
--- a/drivers/video/fbsysfs.c
+++ b/drivers/video/fbsysfs.c
@@ -177,6 +177,8 @@ static ssize_t store_modes(struct device *device,
 	if (i * sizeof(struct fb_videomode) != count)
 		return -EINVAL;
 
+	if (!lock_fb_info(fb_info))
+		return -ENODEV;
 	console_lock();
 	list_splice(&fb_info->modelist, &old_list);
 	fb_videomode_to_modelist((const struct fb_videomode *)buf, i,
@@ -188,6 +190,7 @@ static ssize_t store_modes(struct device *device,
 		fb_destroy_modelist(&old_list);
 
 	console_unlock();
+	unlock_fb_info(fb_info);
 
 	return 0;
 }
diff --git a/include/linux/console.h b/include/linux/console.h
index dedb082..4ef4307 100644
--- a/include/linux/console.h
+++ b/include/linux/console.h
@@ -78,6 +78,7 @@ int con_is_bound(const struct consw *csw);
 int register_con_driver(const struct consw *csw, int first, int last);
 int unregister_con_driver(const struct consw *csw);
 int take_over_console(const struct consw *sw, int first, int last, int deflt);
+int do_take_over_console(const struct consw *sw, int first, int last, int deflt);
 void give_up_console(const struct consw *sw);
 #ifdef CONFIG_HW_CONSOLE
 int con_debug_enter(struct vc_data *vc);
-- 
cgit v0.10.2


From e93a9a868792ad71cdd09d75e5a02d8067473c4e Mon Sep 17 00:00:00 2001
From: Takashi Iwai <tiwai@suse.de>
Date: Fri, 25 Jan 2013 10:28:18 +1000
Subject: fb: Yet another band-aid for fixing lockdep mess

I've still got lockdep warnings even after Alan's patch, and it seems that
yet more band aids are required to paper over similar paths for
unbind_con_driver() and unregister_con_driver().  After this hack, lockdep
warnings are finally gone.

Signed-off-by: Takashi Iwai <tiwai@suse.de>
Cc: Alan Cox <alan@linux.intel.com>
Cc: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Cc: Jiri Kosina <jkosina@suse.cz>
Cc: stable <stable@vger.kernel.org>
Tested-by: Sedat Dilek <sedat.dilek@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/tty/vt/vt.c b/drivers/tty/vt/vt.c
index c076af0..457c41f 100644
--- a/drivers/tty/vt/vt.c
+++ b/drivers/tty/vt/vt.c
@@ -3135,6 +3135,18 @@ static int con_is_graphics(const struct consw *csw, int first, int last)
  */
 int unbind_con_driver(const struct consw *csw, int first, int last, int deflt)
 {
+	int retval;
+
+	console_lock();
+	retval = do_unbind_con_driver(csw, first, last, deflt);
+	console_unlock();
+	return retval;
+}
+EXPORT_SYMBOL(unbind_con_driver);
+
+/* unlocked version of unbind_con_driver() */
+int do_unbind_con_driver(const struct consw *csw, int first, int last, int deflt)
+{
 	struct module *owner = csw->owner;
 	const struct consw *defcsw = NULL;
 	struct con_driver *con_driver = NULL, *con_back = NULL;
@@ -3143,7 +3155,7 @@ int unbind_con_driver(const struct consw *csw, int first, int last, int deflt)
 	if (!try_module_get(owner))
 		return -ENODEV;
 
-	console_lock();
+	WARN_CONSOLE_UNLOCKED();
 
 	/* check if driver is registered and if it is unbindable */
 	for (i = 0; i < MAX_NR_CON_DRIVER; i++) {
@@ -3156,10 +3168,8 @@ int unbind_con_driver(const struct consw *csw, int first, int last, int deflt)
 		}
 	}
 
-	if (retval) {
-		console_unlock();
+	if (retval)
 		goto err;
-	}
 
 	retval = -ENODEV;
 
@@ -3175,15 +3185,11 @@ int unbind_con_driver(const struct consw *csw, int first, int last, int deflt)
 		}
 	}
 
-	if (retval) {
-		console_unlock();
+	if (retval)
 		goto err;
-	}
 
-	if (!con_is_bound(csw)) {
-		console_unlock();
+	if (!con_is_bound(csw))
 		goto err;
-	}
 
 	first = max(first, con_driver->first);
 	last = min(last, con_driver->last);
@@ -3212,13 +3218,12 @@ int unbind_con_driver(const struct consw *csw, int first, int last, int deflt)
 
 	/* ignore return value, binding should not fail */
 	do_bind_con_driver(defcsw, first, last, deflt);
-	console_unlock();
 err:
 	module_put(owner);
 	return retval;
 
 }
-EXPORT_SYMBOL(unbind_con_driver);
+EXPORT_SYMBOL_GPL(do_unbind_con_driver);
 
 static int vt_bind(struct con_driver *con)
 {
@@ -3605,9 +3610,18 @@ EXPORT_SYMBOL(register_con_driver);
  */
 int unregister_con_driver(const struct consw *csw)
 {
-	int i, retval = -ENODEV;
+	int retval;
 
 	console_lock();
+	retval = do_unregister_con_driver(csw);
+	console_unlock();
+	return retval;
+}
+EXPORT_SYMBOL(unregister_con_driver);
+
+int do_unregister_con_driver(const struct consw *csw)
+{
+	int i, retval = -ENODEV;
 
 	/* cannot unregister a bound driver */
 	if (con_is_bound(csw))
@@ -3633,10 +3647,9 @@ int unregister_con_driver(const struct consw *csw)
 		}
 	}
 err:
-	console_unlock();
 	return retval;
 }
-EXPORT_SYMBOL(unregister_con_driver);
+EXPORT_SYMBOL_GPL(do_unregister_con_driver);
 
 /*
  *	If we support more console drivers, this function is used
diff --git a/drivers/video/console/fbcon.c b/drivers/video/console/fbcon.c
index 4bd7820..2aef9ca 100644
--- a/drivers/video/console/fbcon.c
+++ b/drivers/video/console/fbcon.c
@@ -3004,7 +3004,7 @@ static int fbcon_unbind(void)
 {
 	int ret;
 
-	ret = unbind_con_driver(&fb_con, first_fb_vc, last_fb_vc,
+	ret = do_unbind_con_driver(&fb_con, first_fb_vc, last_fb_vc,
 				fbcon_is_default);
 
 	if (!ret)
@@ -3077,7 +3077,7 @@ static int fbcon_fb_unregistered(struct fb_info *info)
 		primary_device = -1;
 
 	if (!num_registered_fb)
-		unregister_con_driver(&fb_con);
+		do_unregister_con_driver(&fb_con);
 
 	return 0;
 }
diff --git a/drivers/video/fbmem.c b/drivers/video/fbmem.c
index d8d9831..070b9a1 100644
--- a/drivers/video/fbmem.c
+++ b/drivers/video/fbmem.c
@@ -1668,8 +1668,10 @@ static int do_unregister_framebuffer(struct fb_info *fb_info)
 
 	if (!lock_fb_info(fb_info))
 		return -ENODEV;
+	console_lock();
 	event.info = fb_info;
 	ret = fb_notifier_call_chain(FB_EVENT_FB_UNBIND, &event);
+	console_unlock();
 	unlock_fb_info(fb_info);
 
 	if (ret)
@@ -1684,7 +1686,9 @@ static int do_unregister_framebuffer(struct fb_info *fb_info)
 	num_registered_fb--;
 	fb_cleanup_device(fb_info);
 	event.info = fb_info;
+	console_lock();
 	fb_notifier_call_chain(FB_EVENT_FB_UNREGISTERED, &event);
+	console_unlock();
 
 	/* this may free fb info */
 	put_fb_info(fb_info);
diff --git a/include/linux/console.h b/include/linux/console.h
index 4ef4307..47b858c 100644
--- a/include/linux/console.h
+++ b/include/linux/console.h
@@ -77,6 +77,7 @@ extern const struct consw prom_con;	/* SPARC PROM console */
 int con_is_bound(const struct consw *csw);
 int register_con_driver(const struct consw *csw, int first, int last);
 int unregister_con_driver(const struct consw *csw);
+int do_unregister_con_driver(const struct consw *csw);
 int take_over_console(const struct consw *sw, int first, int last, int deflt);
 int do_take_over_console(const struct consw *sw, int first, int last, int deflt);
 void give_up_console(const struct consw *sw);
diff --git a/include/linux/vt_kern.h b/include/linux/vt_kern.h
index 50ae7d0..dbbc6bf 100644
--- a/include/linux/vt_kern.h
+++ b/include/linux/vt_kern.h
@@ -130,6 +130,8 @@ void vt_event_post(unsigned int event, unsigned int old, unsigned int new);
 int vt_waitactive(int n);
 void change_console(struct vc_data *new_vc);
 void reset_vc(struct vc_data *vc);
+extern int do_unbind_con_driver(const struct consw *csw, int first, int last,
+			     int deflt);
 extern int unbind_con_driver(const struct consw *csw, int first, int last,
 			     int deflt);
 int vty_init(const struct file_operations *console_fops);
-- 
cgit v0.10.2


From 054430e773c9a1e26f38e30156eff02dedfffc17 Mon Sep 17 00:00:00 2001
From: Dave Airlie <airlied@gmail.com>
Date: Fri, 25 Jan 2013 11:38:56 +1000
Subject: fbcon: fix locking harder

Okay so Alan's patch handled the case where there was no registered fbcon,
however the other path entered in set_con2fb_map pit.

In there we called fbcon_takeover, but we also took the console lock in a couple
of places. So push the console lock out to the callers of set_con2fb_map,

this means fbmem and switcheroo needed to take the lock around the fb notifier
entry points that lead to this.

This should fix the efifb regression seen by Maarten.

Tested-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Tested-by: Lu Hua <huax.lu@intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/gpu/vga/vga_switcheroo.c b/drivers/gpu/vga/vga_switcheroo.c
index fa60add..cf787e1 100644
--- a/drivers/gpu/vga/vga_switcheroo.c
+++ b/drivers/gpu/vga/vga_switcheroo.c
@@ -25,6 +25,7 @@
 #include <linux/fb.h>
 
 #include <linux/pci.h>
+#include <linux/console.h>
 #include <linux/vga_switcheroo.h>
 
 #include <linux/vgaarb.h>
@@ -337,8 +338,10 @@ static int vga_switchto_stage2(struct vga_switcheroo_client *new_client)
 
 	if (new_client->fb_info) {
 		struct fb_event event;
+		console_lock();
 		event.info = new_client->fb_info;
 		fb_notifier_call_chain(FB_EVENT_REMAP_ALL_CONSOLE, &event);
+		console_unlock();
 	}
 
 	ret = vgasr_priv.handler->switchto(new_client->id);
diff --git a/drivers/video/console/fbcon.c b/drivers/video/console/fbcon.c
index 2aef9ca..2e2d959 100644
--- a/drivers/video/console/fbcon.c
+++ b/drivers/video/console/fbcon.c
@@ -842,6 +842,8 @@ static void con2fb_init_display(struct vc_data *vc, struct fb_info *info,
  *
  *	Maps a virtual console @unit to a frame buffer device
  *	@newidx.
+ *
+ *	This should be called with the console lock held.
  */
 static int set_con2fb_map(int unit, int newidx, int user)
 {
@@ -859,7 +861,7 @@ static int set_con2fb_map(int unit, int newidx, int user)
 
 	if (!search_for_mapped_con() || !con_is_bound(&fb_con)) {
 		info_idx = newidx;
-		return fbcon_takeover(0);
+		return do_fbcon_takeover(0);
 	}
 
 	if (oldidx != -1)
@@ -867,7 +869,6 @@ static int set_con2fb_map(int unit, int newidx, int user)
 
 	found = search_fb_in_map(newidx);
 
-	console_lock();
 	con2fb_map[unit] = newidx;
 	if (!err && !found)
  		err = con2fb_acquire_newinfo(vc, info, unit, oldidx);
@@ -894,7 +895,6 @@ static int set_con2fb_map(int unit, int newidx, int user)
 	if (!search_fb_in_map(info_idx))
 		info_idx = newidx;
 
-	console_unlock();
  	return err;
 }
 
@@ -3019,6 +3019,7 @@ static inline int fbcon_unbind(void)
 }
 #endif /* CONFIG_VT_HW_CONSOLE_BINDING */
 
+/* called with console_lock held */
 static int fbcon_fb_unbind(int idx)
 {
 	int i, new_idx = -1, ret = 0;
@@ -3045,6 +3046,7 @@ static int fbcon_fb_unbind(int idx)
 	return ret;
 }
 
+/* called with console_lock held */
 static int fbcon_fb_unregistered(struct fb_info *info)
 {
 	int i, idx;
@@ -3082,6 +3084,7 @@ static int fbcon_fb_unregistered(struct fb_info *info)
 	return 0;
 }
 
+/* called with console_lock held */
 static void fbcon_remap_all(int idx)
 {
 	int i;
@@ -3126,6 +3129,7 @@ static inline void fbcon_select_primary(struct fb_info *info)
 }
 #endif /* CONFIG_FRAMEBUFFER_DETECT_PRIMARY */
 
+/* called with console_lock held */
 static int fbcon_fb_registered(struct fb_info *info)
 {
 	int ret = 0, i, idx;
@@ -3278,6 +3282,7 @@ static int fbcon_event_notify(struct notifier_block *self,
 		ret = fbcon_fb_unregistered(info);
 		break;
 	case FB_EVENT_SET_CONSOLE_MAP:
+		/* called with console lock held */
 		con2fb = event->data;
 		ret = set_con2fb_map(con2fb->console - 1,
 				     con2fb->framebuffer, 1);
diff --git a/drivers/video/fbmem.c b/drivers/video/fbmem.c
index 070b9a1..dc61c12 100644
--- a/drivers/video/fbmem.c
+++ b/drivers/video/fbmem.c
@@ -1177,8 +1177,10 @@ static long do_fb_ioctl(struct fb_info *info, unsigned int cmd,
 		event.data = &con2fb;
 		if (!lock_fb_info(info))
 			return -ENODEV;
+		console_lock();
 		event.info = info;
 		ret = fb_notifier_call_chain(FB_EVENT_SET_CONSOLE_MAP, &event);
+		console_unlock();
 		unlock_fb_info(info);
 		break;
 	case FBIOBLANK:
-- 
cgit v0.10.2


From 5845b81bdad374f98f809a658ec747d92c9595c4 Mon Sep 17 00:00:00 2001
From: Dave Airlie <airlied@redhat.com>
Date: Fri, 8 Feb 2013 12:07:01 +1000
Subject: Revert "Revert "console: implement lockdep support for console_lock""

This reverts commit ff0d05bf73620eb7dc8aee7423e992ef87870bdf.

Now that we have all the locking fixes in place, we can revert the
revert. This re-enables lockdep tracking for the console lock,
daee779718a319ff9f83e1ba3339334ac650bb22.

Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/kernel/printk.c b/kernel/printk.c
index 267ce78..357f714 100644
--- a/kernel/printk.c
+++ b/kernel/printk.c
@@ -87,6 +87,12 @@ static DEFINE_SEMAPHORE(console_sem);
 struct console *console_drivers;
 EXPORT_SYMBOL_GPL(console_drivers);
 
+#ifdef CONFIG_LOCKDEP
+static struct lockdep_map console_lock_dep_map = {
+	.name = "console_lock"
+};
+#endif
+
 /*
  * This is used for debugging the mess that is the VT code by
  * keeping track if we have the console semaphore held. It's
@@ -1918,6 +1924,7 @@ void console_lock(void)
 		return;
 	console_locked = 1;
 	console_may_schedule = 1;
+	mutex_acquire(&console_lock_dep_map, 0, 0, _RET_IP_);
 }
 EXPORT_SYMBOL(console_lock);
 
@@ -1939,6 +1946,7 @@ int console_trylock(void)
 	}
 	console_locked = 1;
 	console_may_schedule = 0;
+	mutex_acquire(&console_lock_dep_map, 0, 1, _RET_IP_);
 	return 1;
 }
 EXPORT_SYMBOL(console_trylock);
@@ -2099,6 +2107,7 @@ skip:
 		local_irq_restore(flags);
 	}
 	console_locked = 0;
+	mutex_release(&console_lock_dep_map, 1, _RET_IP_);
 
 	/* Release the exclusive_console once it is used */
 	if (unlikely(exclusive_console))
-- 
cgit v0.10.2


From 6504d0d9900a2c05ea1fbab2ec008bf442993d94 Mon Sep 17 00:00:00 2001
From: Rob Clark <robdclark@gmail.com>
Date: Thu, 24 Jan 2013 21:46:07 +0000
Subject: drm/i2c: give i2c it's own Kconfig

Move this out of nouveau directory.  As we start to add more encoder
slaves used by other drivers, it makes sense to put the Kconfig bits in
one place.

Signed-off-by: Rob Clark <robdclark@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
index 3399b20..ed9e3af 100644
--- a/drivers/gpu/drm/Kconfig
+++ b/drivers/gpu/drm/Kconfig
@@ -69,6 +69,8 @@ config DRM_KMS_CMA_HELPER
 	help
 	  Choose this if you need the KMS CMA helper functions
 
+source "drivers/gpu/drm/i2c/Kconfig"
+
 config DRM_TDFX
 	tristate "3dfx Banshee/Voodoo3+"
 	depends on DRM && PCI
diff --git a/drivers/gpu/drm/i2c/Kconfig b/drivers/gpu/drm/i2c/Kconfig
new file mode 100644
index 0000000..1611836
--- /dev/null
+++ b/drivers/gpu/drm/i2c/Kconfig
@@ -0,0 +1,22 @@
+menu "I2C encoder or helper chips"
+     depends on DRM && DRM_KMS_HELPER && I2C
+
+config DRM_I2C_CH7006
+	tristate "Chrontel ch7006 TV encoder"
+	default m if DRM_NOUVEAU
+	help
+	  Support for Chrontel ch7006 and similar TV encoders, found
+	  on some nVidia video cards.
+
+	  This driver is currently only useful if you're also using
+	  the nouveau driver.
+
+config DRM_I2C_SIL164
+	tristate "Silicon Image sil164 TMDS transmitter"
+	default m if DRM_NOUVEAU
+	help
+	  Support for sil164 and similar single-link (or dual-link
+	  when used in pairs) TMDS transmitters, used in some nVidia
+	  video cards.
+
+endmenu
diff --git a/drivers/gpu/drm/nouveau/Kconfig b/drivers/gpu/drm/nouveau/Kconfig
index 8a55bee..47ccc1a 100644
--- a/drivers/gpu/drm/nouveau/Kconfig
+++ b/drivers/gpu/drm/nouveau/Kconfig
@@ -52,26 +52,3 @@ config DRM_NOUVEAU_BACKLIGHT
 	help
 	  Say Y here if you want to control the backlight of your display
 	  (e.g. a laptop panel).
-
-menu "I2C encoder or helper chips"
-     depends on DRM && DRM_KMS_HELPER && I2C
-
-config DRM_I2C_CH7006
-	tristate "Chrontel ch7006 TV encoder"
-	default m if DRM_NOUVEAU
-	help
-	  Support for Chrontel ch7006 and similar TV encoders, found
-	  on some nVidia video cards.
-
-	  This driver is currently only useful if you're also using
-	  the nouveau driver.
-
-config DRM_I2C_SIL164
-	tristate "Silicon Image sil164 TMDS transmitter"
-	default m if DRM_NOUVEAU
-	help
-	  Support for sil164 and similar single-link (or dual-link
-	  when used in pairs) TMDS transmitters, used in some nVidia
-	  video cards.
-
-endmenu
-- 
cgit v0.10.2


From 89177644a7b6306e6084a89eab7e290f4bfef397 Mon Sep 17 00:00:00 2001
From: Aaron Plattner <aplattner@nvidia.com>
Date: Tue, 15 Jan 2013 20:47:42 +0000
Subject: drm: add prime helpers

Instead of reimplementing all of the dma_buf functionality in every driver,
create helpers drm_prime_import and drm_prime_export that implement them in
terms of new, lower-level hook functions:

  gem_prime_pin: callback when a buffer is created, used to pin buffers into GTT
  gem_prime_get_sg_table: convert a drm_gem_object to an sg_table for export
  gem_prime_import_sg_table: convert an sg_table into a drm_gem_object
  gem_prime_vmap, gem_prime_vunmap: map and unmap an object

These hooks are optional; drivers can opt in by using drm_gem_prime_import and
drm_gem_prime_export as the .gem_prime_import and .gem_prime_export fields of
struct drm_driver.

v2:
- Drop .begin_cpu_access.  None of the drivers this code replaces implemented
  it.  Having it here was a leftover from when I was trying to include i915 in
  this rework.
- Use mutex_lock instead of mutex_lock_interruptible, as these three drivers
  did.  This patch series shouldn't change that behavior.
- Rename helpers to gem_prime_get_sg_table and gem_prime_import_sg_table.
  Rename struct sg_table* variables to 'sgt' for clarity.
- Update drm.tmpl for these new hooks.

v3:
- Pass the vaddr down to the driver.  This lets drivers that just call vunmap on
  the pointer avoid having to store the pointer in their GEM private structures.
- Move documentation into a /** DOC */ comment in drm_prime.c and include it in
  drm.tmpl with a !P line.  I tried to use !F lines to include documentation of
  the individual functions from drmP.h, but the docproc / kernel-doc scripts
  barf on that file, so hopefully this is good enough for now.
- apply refcount fix from commit be8a42ae60addd8b6092535c11b42d099d6470ec
  ("drm/prime: drop reference on imported dma-buf come from gem")

Signed-off-by: Aaron Plattner <aplattner@nvidia.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: David Airlie <airlied@linux.ie>
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl
index 6c11d77..b26de52 100644
--- a/Documentation/DocBook/drm.tmpl
+++ b/Documentation/DocBook/drm.tmpl
@@ -743,6 +743,10 @@ char *date;</synopsis>
           These two operations are mandatory for GEM drivers that support DRM
           PRIME.
         </para>
+        <sect4>
+          <title>DRM PRIME Helper Functions Reference</title>
+!Pdrivers/gpu/drm/drm_prime.c PRIME Helpers
+        </sect4>
       </sect3>
       <sect3 id="drm-gem-objects-mapping">
         <title>GEM Objects Mapping</title>
diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c
index 7f12573..366910d 100644
--- a/drivers/gpu/drm/drm_prime.c
+++ b/drivers/gpu/drm/drm_prime.c
@@ -53,7 +53,8 @@
  * Self-importing: if userspace is using PRIME as a replacement for flink
  * then it will get a fd->handle request for a GEM object that it created.
  * Drivers should detect this situation and return back the gem object
- * from the dma-buf private.
+ * from the dma-buf private.  Prime will do this automatically for drivers that
+ * use the drm_gem_prime_{import,export} helpers.
  */
 
 struct drm_prime_member {
@@ -62,6 +63,137 @@ struct drm_prime_member {
 	uint32_t handle;
 };
 
+static struct sg_table *drm_gem_map_dma_buf(struct dma_buf_attachment *attach,
+		enum dma_data_direction dir)
+{
+	struct drm_gem_object *obj = attach->dmabuf->priv;
+	struct sg_table *sgt;
+
+	mutex_lock(&obj->dev->struct_mutex);
+
+	sgt = obj->dev->driver->gem_prime_get_sg_table(obj);
+
+	if (!IS_ERR_OR_NULL(sgt))
+		dma_map_sg(attach->dev, sgt->sgl, sgt->nents, dir);
+
+	mutex_unlock(&obj->dev->struct_mutex);
+	return sgt;
+}
+
+static void drm_gem_unmap_dma_buf(struct dma_buf_attachment *attach,
+		struct sg_table *sgt, enum dma_data_direction dir)
+{
+	dma_unmap_sg(attach->dev, sgt->sgl, sgt->nents, dir);
+	sg_free_table(sgt);
+	kfree(sgt);
+}
+
+static void drm_gem_dmabuf_release(struct dma_buf *dma_buf)
+{
+	struct drm_gem_object *obj = dma_buf->priv;
+
+	if (obj->export_dma_buf == dma_buf) {
+		/* drop the reference on the export fd holds */
+		obj->export_dma_buf = NULL;
+		drm_gem_object_unreference_unlocked(obj);
+	}
+}
+
+static void *drm_gem_dmabuf_vmap(struct dma_buf *dma_buf)
+{
+	struct drm_gem_object *obj = dma_buf->priv;
+	struct drm_device *dev = obj->dev;
+
+	return dev->driver->gem_prime_vmap(obj);
+}
+
+static void drm_gem_dmabuf_vunmap(struct dma_buf *dma_buf, void *vaddr)
+{
+	struct drm_gem_object *obj = dma_buf->priv;
+	struct drm_device *dev = obj->dev;
+
+	dev->driver->gem_prime_vunmap(obj, vaddr);
+}
+
+static void *drm_gem_dmabuf_kmap_atomic(struct dma_buf *dma_buf,
+		unsigned long page_num)
+{
+	return NULL;
+}
+
+static void drm_gem_dmabuf_kunmap_atomic(struct dma_buf *dma_buf,
+		unsigned long page_num, void *addr)
+{
+
+}
+static void *drm_gem_dmabuf_kmap(struct dma_buf *dma_buf,
+		unsigned long page_num)
+{
+	return NULL;
+}
+
+static void drm_gem_dmabuf_kunmap(struct dma_buf *dma_buf,
+		unsigned long page_num, void *addr)
+{
+
+}
+
+static int drm_gem_dmabuf_mmap(struct dma_buf *dma_buf,
+		struct vm_area_struct *vma)
+{
+	return -EINVAL;
+}
+
+static const struct dma_buf_ops drm_gem_prime_dmabuf_ops =  {
+	.map_dma_buf = drm_gem_map_dma_buf,
+	.unmap_dma_buf = drm_gem_unmap_dma_buf,
+	.release = drm_gem_dmabuf_release,
+	.kmap = drm_gem_dmabuf_kmap,
+	.kmap_atomic = drm_gem_dmabuf_kmap_atomic,
+	.kunmap = drm_gem_dmabuf_kunmap,
+	.kunmap_atomic = drm_gem_dmabuf_kunmap_atomic,
+	.mmap = drm_gem_dmabuf_mmap,
+	.vmap = drm_gem_dmabuf_vmap,
+	.vunmap = drm_gem_dmabuf_vunmap,
+};
+
+/**
+ * DOC: PRIME Helpers
+ *
+ * Drivers can implement @gem_prime_export and @gem_prime_import in terms of
+ * simpler APIs by using the helper functions @drm_gem_prime_export and
+ * @drm_gem_prime_import.  These functions implement dma-buf support in terms of
+ * five lower-level driver callbacks:
+ *
+ * Export callbacks:
+ *
+ *  - @gem_prime_pin (optional): prepare a GEM object for exporting
+ *
+ *  - @gem_prime_get_sg_table: provide a scatter/gather table of pinned pages
+ *
+ *  - @gem_prime_vmap: vmap a buffer exported by your driver
+ *
+ *  - @gem_prime_vunmap: vunmap a buffer exported by your driver
+ *
+ * Import callback:
+ *
+ *  - @gem_prime_import_sg_table (import): produce a GEM object from another
+ *    driver's scatter/gather table
+ */
+
+struct dma_buf *drm_gem_prime_export(struct drm_device *dev,
+				     struct drm_gem_object *obj, int flags)
+{
+	if (dev->driver->gem_prime_pin) {
+		int ret = dev->driver->gem_prime_pin(obj);
+		if (ret)
+			return ERR_PTR(ret);
+	}
+	return dma_buf_export(obj, &drm_gem_prime_dmabuf_ops, obj->size,
+			      0600);
+}
+EXPORT_SYMBOL(drm_gem_prime_export);
+
 int drm_gem_prime_handle_to_fd(struct drm_device *dev,
 		struct drm_file *file_priv, uint32_t handle, uint32_t flags,
 		int *prime_fd)
@@ -117,6 +249,58 @@ int drm_gem_prime_handle_to_fd(struct drm_device *dev,
 }
 EXPORT_SYMBOL(drm_gem_prime_handle_to_fd);
 
+struct drm_gem_object *drm_gem_prime_import(struct drm_device *dev,
+					    struct dma_buf *dma_buf)
+{
+	struct dma_buf_attachment *attach;
+	struct sg_table *sgt;
+	struct drm_gem_object *obj;
+	int ret;
+
+	if (!dev->driver->gem_prime_import_sg_table)
+		return ERR_PTR(-EINVAL);
+
+	if (dma_buf->ops == &drm_gem_prime_dmabuf_ops) {
+		obj = dma_buf->priv;
+		if (obj->dev == dev) {
+			/*
+			 * Importing dmabuf exported from out own gem increases
+			 * refcount on gem itself instead of f_count of dmabuf.
+			 */
+			drm_gem_object_reference(obj);
+			dma_buf_put(dma_buf);
+			return obj;
+		}
+	}
+
+	attach = dma_buf_attach(dma_buf, dev->dev);
+	if (IS_ERR(attach))
+		return ERR_PTR(PTR_ERR(attach));
+
+	sgt = dma_buf_map_attachment(attach, DMA_BIDIRECTIONAL);
+	if (IS_ERR_OR_NULL(sgt)) {
+		ret = PTR_ERR(sgt);
+		goto fail_detach;
+	}
+
+	obj = dev->driver->gem_prime_import_sg_table(dev, dma_buf->size, sgt);
+	if (IS_ERR(obj)) {
+		ret = PTR_ERR(obj);
+		goto fail_unmap;
+	}
+
+	obj->import_attach = attach;
+
+	return obj;
+
+fail_unmap:
+	dma_buf_unmap_attachment(attach, sgt, DMA_BIDIRECTIONAL);
+fail_detach:
+	dma_buf_detach(dma_buf, attach);
+	return ERR_PTR(ret);
+}
+EXPORT_SYMBOL(drm_gem_prime_import);
+
 int drm_gem_prime_fd_to_handle(struct drm_device *dev,
 		struct drm_file *file_priv, int prime_fd, uint32_t *handle)
 {
diff --git a/include/drm/drmP.h b/include/drm/drmP.h
index 3aaa50d..2d94d74 100644
--- a/include/drm/drmP.h
+++ b/include/drm/drmP.h
@@ -930,6 +930,14 @@ struct drm_driver {
 	/* import dmabuf -> GEM */
 	struct drm_gem_object * (*gem_prime_import)(struct drm_device *dev,
 				struct dma_buf *dma_buf);
+	/* low-level interface used by drm_gem_prime_{import,export} */
+	int (*gem_prime_pin)(struct drm_gem_object *obj);
+	struct sg_table *(*gem_prime_get_sg_table)(struct drm_gem_object *obj);
+	struct drm_gem_object *(*gem_prime_import_sg_table)(
+				struct drm_device *dev, size_t size,
+				struct sg_table *sgt);
+	void *(*gem_prime_vmap)(struct drm_gem_object *obj);
+	void (*gem_prime_vunmap)(struct drm_gem_object *obj, void *vaddr);
 
 	/* vga arb irq handler */
 	void (*vgaarb_irq)(struct drm_device *dev, bool state);
@@ -1562,9 +1570,13 @@ extern int drm_clients_info(struct seq_file *m, void* data);
 extern int drm_gem_name_info(struct seq_file *m, void *data);
 
 
+extern struct dma_buf *drm_gem_prime_export(struct drm_device *dev,
+		struct drm_gem_object *obj, int flags);
 extern int drm_gem_prime_handle_to_fd(struct drm_device *dev,
 		struct drm_file *file_priv, uint32_t handle, uint32_t flags,
 		int *prime_fd);
+extern struct drm_gem_object *drm_gem_prime_import(struct drm_device *dev,
+		struct dma_buf *dma_buf);
 extern int drm_gem_prime_fd_to_handle(struct drm_device *dev,
 		struct drm_file *file_priv, int prime_fd, uint32_t *handle);
 
-- 
cgit v0.10.2


From ab9ccb96a6e6f95bcde6b8b2a524370efdbfdcd6 Mon Sep 17 00:00:00 2001
From: Aaron Plattner <aplattner@nvidia.com>
Date: Tue, 15 Jan 2013 20:47:43 +0000
Subject: drm/nouveau: use prime helpers

Simplify the Nouveau prime implementation by using the default behavior provided
by drm_gem_prime_import and drm_gem_prime_export.

v2: Rename functions to nouveau_gem_prime_get_sg_table and
nouveau_gem_prime_import_sg_table.

Signed-off-by: Aaron Plattner <aplattner@nvidia.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: David Airlie <airlied@linux.ie>
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.h b/drivers/gpu/drm/nouveau/nouveau_bo.h
index 81d00fe..653dbbb 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.h
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.h
@@ -33,7 +33,6 @@ struct nouveau_bo {
 	int pin_refcnt;
 
 	struct ttm_bo_kmap_obj dma_buf_vmap;
-	int vmapping_count;
 };
 
 static inline struct nouveau_bo *
diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c
index 8b090f1..8e8e8ce 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_drm.c
@@ -650,8 +650,13 @@ driver = {
 
 	.prime_handle_to_fd = drm_gem_prime_handle_to_fd,
 	.prime_fd_to_handle = drm_gem_prime_fd_to_handle,
-	.gem_prime_export = nouveau_gem_prime_export,
-	.gem_prime_import = nouveau_gem_prime_import,
+	.gem_prime_export = drm_gem_prime_export,
+	.gem_prime_import = drm_gem_prime_import,
+	.gem_prime_pin = nouveau_gem_prime_pin,
+	.gem_prime_get_sg_table = nouveau_gem_prime_get_sg_table,
+	.gem_prime_import_sg_table = nouveau_gem_prime_import_sg_table,
+	.gem_prime_vmap = nouveau_gem_prime_vmap,
+	.gem_prime_vunmap = nouveau_gem_prime_vunmap,
 
 	.gem_init_object = nouveau_gem_object_new,
 	.gem_free_object = nouveau_gem_object_del,
diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c b/drivers/gpu/drm/nouveau/nouveau_gem.c
index 8bf695c..24e0aab 100644
--- a/drivers/gpu/drm/nouveau/nouveau_gem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_gem.c
@@ -24,8 +24,6 @@
  *
  */
 
-#include <linux/dma-buf.h>
-
 #include <subdev/fb.h>
 
 #include "nouveau_drm.h"
diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.h b/drivers/gpu/drm/nouveau/nouveau_gem.h
index 5c10492..8d7a3f0 100644
--- a/drivers/gpu/drm/nouveau/nouveau_gem.h
+++ b/drivers/gpu/drm/nouveau/nouveau_gem.h
@@ -35,9 +35,11 @@ extern int nouveau_gem_ioctl_cpu_fini(struct drm_device *, void *,
 extern int nouveau_gem_ioctl_info(struct drm_device *, void *,
 				  struct drm_file *);
 
-extern struct dma_buf *nouveau_gem_prime_export(struct drm_device *dev,
-				struct drm_gem_object *obj, int flags);
-extern struct drm_gem_object *nouveau_gem_prime_import(struct drm_device *dev,
-				struct dma_buf *dma_buf);
+extern int nouveau_gem_prime_pin(struct drm_gem_object *);
+extern struct sg_table *nouveau_gem_prime_get_sg_table(struct drm_gem_object *);
+extern struct drm_gem_object *nouveau_gem_prime_import_sg_table(
+	struct drm_device *, size_t size, struct sg_table *);
+extern void *nouveau_gem_prime_vmap(struct drm_gem_object *);
+extern void nouveau_gem_prime_vunmap(struct drm_gem_object *, void *);
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/nouveau_prime.c b/drivers/gpu/drm/nouveau/nouveau_prime.c
index b8e05ae..f53e108 100644
--- a/drivers/gpu/drm/nouveau/nouveau_prime.c
+++ b/drivers/gpu/drm/nouveau/nouveau_prime.c
@@ -22,126 +22,42 @@
  * Authors: Dave Airlie
  */
 
-#include <linux/dma-buf.h>
-
 #include <drm/drmP.h>
 
 #include "nouveau_drm.h"
 #include "nouveau_gem.h"
 
-static struct sg_table *nouveau_gem_map_dma_buf(struct dma_buf_attachment *attachment,
-					  enum dma_data_direction dir)
+struct sg_table *nouveau_gem_prime_get_sg_table(struct drm_gem_object *obj)
 {
-	struct nouveau_bo *nvbo = attachment->dmabuf->priv;
-	struct drm_device *dev = nvbo->gem->dev;
+	struct nouveau_bo *nvbo = nouveau_gem_object(obj);
 	int npages = nvbo->bo.num_pages;
-	struct sg_table *sg;
-	int nents;
-
-	mutex_lock(&dev->struct_mutex);
-	sg = drm_prime_pages_to_sg(nvbo->bo.ttm->pages, npages);
-	nents = dma_map_sg(attachment->dev, sg->sgl, sg->nents, dir);
-	mutex_unlock(&dev->struct_mutex);
-	return sg;
-}
-
-static void nouveau_gem_unmap_dma_buf(struct dma_buf_attachment *attachment,
-				      struct sg_table *sg, enum dma_data_direction dir)
-{
-	dma_unmap_sg(attachment->dev, sg->sgl, sg->nents, dir);
-	sg_free_table(sg);
-	kfree(sg);
-}
-
-static void nouveau_gem_dmabuf_release(struct dma_buf *dma_buf)
-{
-	struct nouveau_bo *nvbo = dma_buf->priv;
-
-	if (nvbo->gem->export_dma_buf == dma_buf) {
-		nvbo->gem->export_dma_buf = NULL;
-		drm_gem_object_unreference_unlocked(nvbo->gem);
-	}
-}
-
-static void *nouveau_gem_kmap_atomic(struct dma_buf *dma_buf, unsigned long page_num)
-{
-	return NULL;
-}
-
-static void nouveau_gem_kunmap_atomic(struct dma_buf *dma_buf, unsigned long page_num, void *addr)
-{
 
-}
-static void *nouveau_gem_kmap(struct dma_buf *dma_buf, unsigned long page_num)
-{
-	return NULL;
+	return drm_prime_pages_to_sg(nvbo->bo.ttm->pages, npages);
 }
 
-static void nouveau_gem_kunmap(struct dma_buf *dma_buf, unsigned long page_num, void *addr)
+void *nouveau_gem_prime_vmap(struct drm_gem_object *obj)
 {
-
-}
-
-static int nouveau_gem_prime_mmap(struct dma_buf *dma_buf, struct vm_area_struct *vma)
-{
-	return -EINVAL;
-}
-
-static void *nouveau_gem_prime_vmap(struct dma_buf *dma_buf)
-{
-	struct nouveau_bo *nvbo = dma_buf->priv;
-	struct drm_device *dev = nvbo->gem->dev;
+	struct nouveau_bo *nvbo = nouveau_gem_object(obj);
 	int ret;
 
-	mutex_lock(&dev->struct_mutex);
-	if (nvbo->vmapping_count) {
-		nvbo->vmapping_count++;
-		goto out_unlock;
-	}
-
 	ret = ttm_bo_kmap(&nvbo->bo, 0, nvbo->bo.num_pages,
 			  &nvbo->dma_buf_vmap);
-	if (ret) {
-		mutex_unlock(&dev->struct_mutex);
+	if (ret)
 		return ERR_PTR(ret);
-	}
-	nvbo->vmapping_count = 1;
-out_unlock:
-	mutex_unlock(&dev->struct_mutex);
+
 	return nvbo->dma_buf_vmap.virtual;
 }
 
-static void nouveau_gem_prime_vunmap(struct dma_buf *dma_buf, void *vaddr)
+void nouveau_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr)
 {
-	struct nouveau_bo *nvbo = dma_buf->priv;
-	struct drm_device *dev = nvbo->gem->dev;
+	struct nouveau_bo *nvbo = nouveau_gem_object(obj);
 
-	mutex_lock(&dev->struct_mutex);
-	nvbo->vmapping_count--;
-	if (nvbo->vmapping_count == 0) {
-		ttm_bo_kunmap(&nvbo->dma_buf_vmap);
-	}
-	mutex_unlock(&dev->struct_mutex);
+	ttm_bo_kunmap(&nvbo->dma_buf_vmap);
 }
 
-static const struct dma_buf_ops nouveau_dmabuf_ops =  {
-	.map_dma_buf = nouveau_gem_map_dma_buf,
-	.unmap_dma_buf = nouveau_gem_unmap_dma_buf,
-	.release = nouveau_gem_dmabuf_release,
-	.kmap = nouveau_gem_kmap,
-	.kmap_atomic = nouveau_gem_kmap_atomic,
-	.kunmap = nouveau_gem_kunmap,
-	.kunmap_atomic = nouveau_gem_kunmap_atomic,
-	.mmap = nouveau_gem_prime_mmap,
-	.vmap = nouveau_gem_prime_vmap,
-	.vunmap = nouveau_gem_prime_vunmap,
-};
-
-static int
-nouveau_prime_new(struct drm_device *dev,
-		  size_t size,
-		  struct sg_table *sg,
-		  struct nouveau_bo **pnvbo)
+struct drm_gem_object *nouveau_gem_prime_import_sg_table(struct drm_device *dev,
+							 size_t size,
+							 struct sg_table *sg)
 {
 	struct nouveau_bo *nvbo;
 	u32 flags = 0;
@@ -150,24 +66,22 @@ nouveau_prime_new(struct drm_device *dev,
 	flags = TTM_PL_FLAG_TT;
 
 	ret = nouveau_bo_new(dev, size, 0, flags, 0, 0,
-			     sg, pnvbo);
+			     sg, &nvbo);
 	if (ret)
-		return ret;
-	nvbo = *pnvbo;
+		return ERR_PTR(ret);
 
 	nvbo->valid_domains = NOUVEAU_GEM_DOMAIN_GART;
 	nvbo->gem = drm_gem_object_alloc(dev, nvbo->bo.mem.size);
 	if (!nvbo->gem) {
-		nouveau_bo_ref(NULL, pnvbo);
-		return -ENOMEM;
+		nouveau_bo_ref(NULL, &nvbo);
+		return ERR_PTR(-ENOMEM);
 	}
 
 	nvbo->gem->driver_private = nvbo;
-	return 0;
+	return nvbo->gem;
 }
 
-struct dma_buf *nouveau_gem_prime_export(struct drm_device *dev,
-				struct drm_gem_object *obj, int flags)
+int nouveau_gem_prime_pin(struct drm_gem_object *obj)
 {
 	struct nouveau_bo *nvbo = nouveau_gem_object(obj);
 	int ret = 0;
@@ -175,52 +89,7 @@ struct dma_buf *nouveau_gem_prime_export(struct drm_device *dev,
 	/* pin buffer into GTT */
 	ret = nouveau_bo_pin(nvbo, TTM_PL_FLAG_TT);
 	if (ret)
-		return ERR_PTR(-EINVAL);
-
-	return dma_buf_export(nvbo, &nouveau_dmabuf_ops, obj->size, flags);
-}
-
-struct drm_gem_object *nouveau_gem_prime_import(struct drm_device *dev,
-				struct dma_buf *dma_buf)
-{
-	struct dma_buf_attachment *attach;
-	struct sg_table *sg;
-	struct nouveau_bo *nvbo;
-	int ret;
-
-	if (dma_buf->ops == &nouveau_dmabuf_ops) {
-		nvbo = dma_buf->priv;
-		if (nvbo->gem) {
-			if (nvbo->gem->dev == dev) {
-				drm_gem_object_reference(nvbo->gem);
-				dma_buf_put(dma_buf);
-				return nvbo->gem;
-			}
-		}
-	}
-	/* need to attach */
-	attach = dma_buf_attach(dma_buf, dev->dev);
-	if (IS_ERR(attach))
-		return ERR_PTR(PTR_ERR(attach));
-
-	sg = dma_buf_map_attachment(attach, DMA_BIDIRECTIONAL);
-	if (IS_ERR(sg)) {
-		ret = PTR_ERR(sg);
-		goto fail_detach;
-	}
-
-	ret = nouveau_prime_new(dev, dma_buf->size, sg, &nvbo);
-	if (ret)
-		goto fail_unmap;
-
-	nvbo->gem->import_attach = attach;
-
-	return nvbo->gem;
+		return -EINVAL;
 
-fail_unmap:
-	dma_buf_unmap_attachment(attach, sg, DMA_BIDIRECTIONAL);
-fail_detach:
-	dma_buf_detach(dma_buf, attach);
-	return ERR_PTR(ret);
+	return 0;
 }
-
-- 
cgit v0.10.2


From 1e6d17a5df848cf8e483b689c6295776e9d6d997 Mon Sep 17 00:00:00 2001
From: Aaron Plattner <aplattner@nvidia.com>
Date: Tue, 15 Jan 2013 20:47:44 +0000
Subject: drm/radeon: use prime helpers

Simplify the Radeon prime implementation by using the default behavior provided
by drm_gem_prime_import and drm_gem_prime_export.

v2:
- Rename functions to radeon_gem_prime_get_sg_table and
  radeon_gem_prime_import_sg_table.
- Delete the now-unused vmapping_count variable.

Signed-off-by: Aaron Plattner <aplattner@nvidia.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: David Airlie <airlied@linux.ie>
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 59bfbd3..bb43a84 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -350,7 +350,6 @@ struct radeon_bo {
 	struct drm_gem_object		gem_base;
 
 	struct ttm_bo_kmap_obj dma_buf_vmap;
-	int vmapping_count;
 };
 #define gem_to_radeon_bo(gobj) container_of((gobj), struct radeon_bo, gem_base)
 
diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c
index 833484d..1677584 100644
--- a/drivers/gpu/drm/radeon/radeon_drv.c
+++ b/drivers/gpu/drm/radeon/radeon_drv.c
@@ -118,11 +118,13 @@ int radeon_mode_dumb_create(struct drm_file *file_priv,
 int radeon_mode_dumb_destroy(struct drm_file *file_priv,
 			     struct drm_device *dev,
 			     uint32_t handle);
-struct dma_buf *radeon_gem_prime_export(struct drm_device *dev,
-					struct drm_gem_object *obj,
-					int flags);
-struct drm_gem_object *radeon_gem_prime_import(struct drm_device *dev,
-					       struct dma_buf *dma_buf);
+struct sg_table *radeon_gem_prime_get_sg_table(struct drm_gem_object *obj);
+struct drm_gem_object *radeon_gem_prime_import_sg_table(struct drm_device *dev,
+							size_t size,
+							struct sg_table *sg);
+int radeon_gem_prime_pin(struct drm_gem_object *obj);
+void *radeon_gem_prime_vmap(struct drm_gem_object *obj);
+void radeon_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr);
 extern long radeon_kms_compat_ioctl(struct file *filp, unsigned int cmd,
 				    unsigned long arg);
 
@@ -409,8 +411,13 @@ static struct drm_driver kms_driver = {
 
 	.prime_handle_to_fd = drm_gem_prime_handle_to_fd,
 	.prime_fd_to_handle = drm_gem_prime_fd_to_handle,
-	.gem_prime_export = radeon_gem_prime_export,
-	.gem_prime_import = radeon_gem_prime_import,
+	.gem_prime_export = drm_gem_prime_export,
+	.gem_prime_import = drm_gem_prime_import,
+	.gem_prime_pin = radeon_gem_prime_pin,
+	.gem_prime_get_sg_table = radeon_gem_prime_get_sg_table,
+	.gem_prime_import_sg_table = radeon_gem_prime_import_sg_table,
+	.gem_prime_vmap = radeon_gem_prime_vmap,
+	.gem_prime_vunmap = radeon_gem_prime_vunmap,
 
 	.name = DRIVER_NAME,
 	.desc = DRIVER_DESC,
diff --git a/drivers/gpu/drm/radeon/radeon_prime.c b/drivers/gpu/drm/radeon/radeon_prime.c
index 26c23bb..4940af7 100644
--- a/drivers/gpu/drm/radeon/radeon_prime.c
+++ b/drivers/gpu/drm/radeon/radeon_prime.c
@@ -28,199 +28,71 @@
 #include "radeon.h"
 #include <drm/radeon_drm.h>
 
-#include <linux/dma-buf.h>
-
-static struct sg_table *radeon_gem_map_dma_buf(struct dma_buf_attachment *attachment,
-					       enum dma_data_direction dir)
+struct sg_table *radeon_gem_prime_get_sg_table(struct drm_gem_object *obj)
 {
-	struct radeon_bo *bo = attachment->dmabuf->priv;
-	struct drm_device *dev = bo->rdev->ddev;
+	struct radeon_bo *bo = gem_to_radeon_bo(obj);
 	int npages = bo->tbo.num_pages;
-	struct sg_table *sg;
-	int nents;
-
-	mutex_lock(&dev->struct_mutex);
-	sg = drm_prime_pages_to_sg(bo->tbo.ttm->pages, npages);
-	nents = dma_map_sg(attachment->dev, sg->sgl, sg->nents, dir);
-	mutex_unlock(&dev->struct_mutex);
-	return sg;
-}
-
-static void radeon_gem_unmap_dma_buf(struct dma_buf_attachment *attachment,
-				     struct sg_table *sg, enum dma_data_direction dir)
-{
-	dma_unmap_sg(attachment->dev, sg->sgl, sg->nents, dir);
-	sg_free_table(sg);
-	kfree(sg);
-}
-
-static void radeon_gem_dmabuf_release(struct dma_buf *dma_buf)
-{
-	struct radeon_bo *bo = dma_buf->priv;
-
-	if (bo->gem_base.export_dma_buf == dma_buf) {
-		DRM_ERROR("unreference dmabuf %p\n", &bo->gem_base);
-		bo->gem_base.export_dma_buf = NULL;
-		drm_gem_object_unreference_unlocked(&bo->gem_base);
-	}
-}
-
-static void *radeon_gem_kmap_atomic(struct dma_buf *dma_buf, unsigned long page_num)
-{
-	return NULL;
-}
-
-static void radeon_gem_kunmap_atomic(struct dma_buf *dma_buf, unsigned long page_num, void *addr)
-{
-
-}
-static void *radeon_gem_kmap(struct dma_buf *dma_buf, unsigned long page_num)
-{
-	return NULL;
-}
-
-static void radeon_gem_kunmap(struct dma_buf *dma_buf, unsigned long page_num, void *addr)
-{
 
+	return drm_prime_pages_to_sg(bo->tbo.ttm->pages, npages);
 }
 
-static int radeon_gem_prime_mmap(struct dma_buf *dma_buf, struct vm_area_struct *vma)
+void *radeon_gem_prime_vmap(struct drm_gem_object *obj)
 {
-	return -EINVAL;
-}
-
-static void *radeon_gem_prime_vmap(struct dma_buf *dma_buf)
-{
-	struct radeon_bo *bo = dma_buf->priv;
-	struct drm_device *dev = bo->rdev->ddev;
+	struct radeon_bo *bo = gem_to_radeon_bo(obj);
 	int ret;
 
-	mutex_lock(&dev->struct_mutex);
-	if (bo->vmapping_count) {
-		bo->vmapping_count++;
-		goto out_unlock;
-	}
-
 	ret = ttm_bo_kmap(&bo->tbo, 0, bo->tbo.num_pages,
 			  &bo->dma_buf_vmap);
-	if (ret) {
-		mutex_unlock(&dev->struct_mutex);
+	if (ret)
 		return ERR_PTR(ret);
-	}
-	bo->vmapping_count = 1;
-out_unlock:
-	mutex_unlock(&dev->struct_mutex);
+
 	return bo->dma_buf_vmap.virtual;
 }
 
-static void radeon_gem_prime_vunmap(struct dma_buf *dma_buf, void *vaddr)
+void radeon_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr)
 {
-	struct radeon_bo *bo = dma_buf->priv;
-	struct drm_device *dev = bo->rdev->ddev;
+	struct radeon_bo *bo = gem_to_radeon_bo(obj);
 
-	mutex_lock(&dev->struct_mutex);
-	bo->vmapping_count--;
-	if (bo->vmapping_count == 0) {
-		ttm_bo_kunmap(&bo->dma_buf_vmap);
-	}
-	mutex_unlock(&dev->struct_mutex);
+	ttm_bo_kunmap(&bo->dma_buf_vmap);
 }
-const static struct dma_buf_ops radeon_dmabuf_ops =  {
-	.map_dma_buf = radeon_gem_map_dma_buf,
-	.unmap_dma_buf = radeon_gem_unmap_dma_buf,
-	.release = radeon_gem_dmabuf_release,
-	.kmap = radeon_gem_kmap,
-	.kmap_atomic = radeon_gem_kmap_atomic,
-	.kunmap = radeon_gem_kunmap,
-	.kunmap_atomic = radeon_gem_kunmap_atomic,
-	.mmap = radeon_gem_prime_mmap,
-	.vmap = radeon_gem_prime_vmap,
-	.vunmap = radeon_gem_prime_vunmap,
-};
-
-static int radeon_prime_create(struct drm_device *dev,
-			       size_t size,
-			       struct sg_table *sg,
-			       struct radeon_bo **pbo)
+
+struct drm_gem_object *radeon_gem_prime_import_sg_table(struct drm_device *dev,
+							size_t size,
+							struct sg_table *sg)
 {
 	struct radeon_device *rdev = dev->dev_private;
 	struct radeon_bo *bo;
 	int ret;
 
 	ret = radeon_bo_create(rdev, size, PAGE_SIZE, false,
-			       RADEON_GEM_DOMAIN_GTT, sg, pbo);
+			       RADEON_GEM_DOMAIN_GTT, sg, &bo);
 	if (ret)
-		return ret;
-	bo = *pbo;
+		return ERR_PTR(ret);
 	bo->gem_base.driver_private = bo;
 
 	mutex_lock(&rdev->gem.mutex);
 	list_add_tail(&bo->list, &rdev->gem.objects);
 	mutex_unlock(&rdev->gem.mutex);
 
-	return 0;
+	return &bo->gem_base;
 }
 
-struct dma_buf *radeon_gem_prime_export(struct drm_device *dev,
-					struct drm_gem_object *obj,
-					int flags)
+int radeon_gem_prime_pin(struct drm_gem_object *obj)
 {
 	struct radeon_bo *bo = gem_to_radeon_bo(obj);
 	int ret = 0;
 
 	ret = radeon_bo_reserve(bo, false);
 	if (unlikely(ret != 0))
-		return ERR_PTR(ret);
+		return ret;
 
 	/* pin buffer into GTT */
 	ret = radeon_bo_pin(bo, RADEON_GEM_DOMAIN_GTT, NULL);
 	if (ret) {
 		radeon_bo_unreserve(bo);
-		return ERR_PTR(ret);
+		return ret;
 	}
 	radeon_bo_unreserve(bo);
-	return dma_buf_export(bo, &radeon_dmabuf_ops, obj->size, flags);
-}
 
-struct drm_gem_object *radeon_gem_prime_import(struct drm_device *dev,
-					       struct dma_buf *dma_buf)
-{
-	struct dma_buf_attachment *attach;
-	struct sg_table *sg;
-	struct radeon_bo *bo;
-	int ret;
-
-	if (dma_buf->ops == &radeon_dmabuf_ops) {
-		bo = dma_buf->priv;
-		if (bo->gem_base.dev == dev) {
-			drm_gem_object_reference(&bo->gem_base);
-			dma_buf_put(dma_buf);
-			return &bo->gem_base;
-		}
-	}
-
-	/* need to attach */
-	attach = dma_buf_attach(dma_buf, dev->dev);
-	if (IS_ERR(attach))
-		return ERR_CAST(attach);
-
-	sg = dma_buf_map_attachment(attach, DMA_BIDIRECTIONAL);
-	if (IS_ERR(sg)) {
-		ret = PTR_ERR(sg);
-		goto fail_detach;
-	}
-
-	ret = radeon_prime_create(dev, dma_buf->size, sg, &bo);
-	if (ret)
-		goto fail_unmap;
-
-	bo->gem_base.import_attach = attach;
-
-	return &bo->gem_base;
-
-fail_unmap:
-	dma_buf_unmap_attachment(attach, sg, DMA_BIDIRECTIONAL);
-fail_detach:
-	dma_buf_detach(dma_buf, attach);
-	return ERR_PTR(ret);
+	return 0;
 }
-- 
cgit v0.10.2


From 03f6509df9218c760ae74f41a609233220d33f19 Mon Sep 17 00:00:00 2001
From: Thierry Reding <thierry.reding@avionic-design.de>
Date: Mon, 14 Jan 2013 16:05:56 +0000
Subject: drm: Allow vblank support without DRIVER_HAVE_IRQ

Drivers that register interrupt handlers without the DRM core helpers
don't initialize the .irq_enabled field and drm_dev_to_irq() may fail
when called on them. This shouldn't preclude them from implementing
the vblank IOCTL.

Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/gpu/drm/drm_irq.c b/drivers/gpu/drm/drm_irq.c
index 19c01ca..71f8205 100644
--- a/drivers/gpu/drm/drm_irq.c
+++ b/drivers/gpu/drm/drm_irq.c
@@ -1218,8 +1218,9 @@ int drm_wait_vblank(struct drm_device *dev, void *data,
 	int ret;
 	unsigned int flags, seq, crtc, high_crtc;
 
-	if ((!drm_dev_to_irq(dev)) || (!dev->irq_enabled))
-		return -EINVAL;
+	if (drm_core_check_feature(dev, DRIVER_HAVE_IRQ))
+		if ((!drm_dev_to_irq(dev)) || (!dev->irq_enabled))
+			return -EINVAL;
 
 	if (vblwait->request.type & _DRM_VBLANK_SIGNAL)
 		return -EINVAL;
-- 
cgit v0.10.2


From 9fe0423eac2c0781a3ece8779b086acd3e76f2c8 Mon Sep 17 00:00:00 2001
From: Bjorn Helgaas <bhelgaas@google.com>
Date: Fri, 4 Jan 2013 19:10:32 +0000
Subject: drm/pci: Use the standard #defines for PCIe Link Capability bits

Use the standard #defines rather than bare numbers for the PCIe Link
Capabilities speed bits.

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/gpu/drm/drm_pci.c b/drivers/gpu/drm/drm_pci.c
index 754bc96..11c8add 100644
--- a/drivers/gpu/drm/drm_pci.c
+++ b/drivers/gpu/drm/drm_pci.c
@@ -504,9 +504,9 @@ int drm_pcie_get_speed_cap_mask(struct drm_device *dev, u32 *mask)
 		if (lnkcap2 & PCI_EXP_LNKCAP2_SLS_8_0GB)
 			*mask |= DRM_PCIE_SPEED_80;
 	} else {
-		if (lnkcap & 1)
+		if (lnkcap & PCI_EXP_LNKCAP_SLS_2_5GB)
 			*mask |= DRM_PCIE_SPEED_25;
-		if (lnkcap & 2)
+		if (lnkcap & PCI_EXP_LNKCAP_SLS_5_0GB)
 			*mask |= DRM_PCIE_SPEED_50;
 	}
 
-- 
cgit v0.10.2


From f8acf6f4c8fe1fd4de1f669ac6a3c71e89f13523 Mon Sep 17 00:00:00 2001
From: Bjorn Helgaas <bhelgaas@google.com>
Date: Fri, 4 Jan 2013 19:10:37 +0000
Subject: drm/pci: Set all supported speeds in speed cap mask for pre-3.0
 devices

For devices that conform to PCIe r3.0 and have a Link Capabilities 2
register, we test and report every bit in the Supported Link Speeds Vector
field.  For a device that supports both 2.5GT/s and 5.0GT/s, we set both
DRM_PCIE_SPEED_25 and DRM_PCIE_SPEED_50 in the returned mask.

For pre-r3.0 devices, the Link Capabilities 0010b encoding
(PCI_EXP_LNKCAP_SLS_5_0GB) means that both 5.0GT/s and 2.5GT/s are
supported, so set both DRM_PCIE_SPEED_25 and DRM_PCIE_SPEED_50 in this
case as well.

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/gpu/drm/drm_pci.c b/drivers/gpu/drm/drm_pci.c
index 11c8add..50e26f2 100644
--- a/drivers/gpu/drm/drm_pci.c
+++ b/drivers/gpu/drm/drm_pci.c
@@ -507,7 +507,7 @@ int drm_pcie_get_speed_cap_mask(struct drm_device *dev, u32 *mask)
 		if (lnkcap & PCI_EXP_LNKCAP_SLS_2_5GB)
 			*mask |= DRM_PCIE_SPEED_25;
 		if (lnkcap & PCI_EXP_LNKCAP_SLS_5_0GB)
-			*mask |= DRM_PCIE_SPEED_50;
+			*mask |= (DRM_PCIE_SPEED_25 | DRM_PCIE_SPEED_50);
 	}
 
 	DRM_INFO("probing gen 2 caps for device %x:%x = %x/%x\n", root->vendor, root->device, lnkcap, lnkcap2);
-- 
cgit v0.10.2


From dd66cc2e1f4765d0e6f39eb1e7d8d64d3f1cc522 Mon Sep 17 00:00:00 2001
From: Bjorn Helgaas <bhelgaas@google.com>
Date: Fri, 4 Jan 2013 19:10:42 +0000
Subject: drm/pci: Use PCI Express Capability accessors

Use PCI Express Capability access functions to simplify this code a bit.
For non-PCIe devices or pre-PCIe 3.0 devices that don't implement the Link
Capabilities 2 register, pcie_capability_read_dword() reads a zero.

Since we're only testing whether the bits we care about are set, there's no
need to mask out the other bits we *don't* care about.

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/gpu/drm/drm_pci.c b/drivers/gpu/drm/drm_pci.c
index 50e26f2..86102a0 100644
--- a/drivers/gpu/drm/drm_pci.c
+++ b/drivers/gpu/drm/drm_pci.c
@@ -469,41 +469,30 @@ EXPORT_SYMBOL(drm_pci_exit);
 int drm_pcie_get_speed_cap_mask(struct drm_device *dev, u32 *mask)
 {
 	struct pci_dev *root;
-	int pos;
-	u32 lnkcap = 0, lnkcap2 = 0;
+	u32 lnkcap, lnkcap2;
 
 	*mask = 0;
 	if (!dev->pdev)
 		return -EINVAL;
 
-	if (!pci_is_pcie(dev->pdev))
-		return -EINVAL;
-
 	root = dev->pdev->bus->self;
 
-	pos = pci_pcie_cap(root);
-	if (!pos)
-		return -EINVAL;
-
 	/* we've been informed via and serverworks don't make the cut */
 	if (root->vendor == PCI_VENDOR_ID_VIA ||
 	    root->vendor == PCI_VENDOR_ID_SERVERWORKS)
 		return -EINVAL;
 
-	pci_read_config_dword(root, pos + PCI_EXP_LNKCAP, &lnkcap);
-	pci_read_config_dword(root, pos + PCI_EXP_LNKCAP2, &lnkcap2);
-
-	lnkcap &= PCI_EXP_LNKCAP_SLS;
-	lnkcap2 &= 0xfe;
+	pcie_capability_read_dword(root, PCI_EXP_LNKCAP, &lnkcap);
+	pcie_capability_read_dword(root, PCI_EXP_LNKCAP2, &lnkcap2);
 
-	if (lnkcap2) { /* PCIE GEN 3.0 */
+	if (lnkcap2) {	/* PCIe r3.0-compliant */
 		if (lnkcap2 & PCI_EXP_LNKCAP2_SLS_2_5GB)
 			*mask |= DRM_PCIE_SPEED_25;
 		if (lnkcap2 & PCI_EXP_LNKCAP2_SLS_5_0GB)
 			*mask |= DRM_PCIE_SPEED_50;
 		if (lnkcap2 & PCI_EXP_LNKCAP2_SLS_8_0GB)
 			*mask |= DRM_PCIE_SPEED_80;
-	} else {
+	} else {	/* pre-r3.0 */
 		if (lnkcap & PCI_EXP_LNKCAP_SLS_2_5GB)
 			*mask |= DRM_PCIE_SPEED_25;
 		if (lnkcap & PCI_EXP_LNKCAP_SLS_5_0GB)
-- 
cgit v0.10.2


From 85a7ce67f3ebfd5975ffd1febcabfe4999ca911d Mon Sep 17 00:00:00 2001
From: Daniel Kurtz <djkurtz@chromium.org>
Date: Thu, 27 Dec 2012 01:01:46 +0000
Subject: drm: make frame duration time calculation more precise

It is a bit more precise to compute the total number of pixels first and
then divide, rather than multiplying the line pixel count by the
already-rounded line duration.

Signed-off-by: Daniel Kurtz <djkurtz@chromium.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/gpu/drm/drm_irq.c b/drivers/gpu/drm/drm_irq.c
index 71f8205..38e7992 100644
--- a/drivers/gpu/drm/drm_irq.c
+++ b/drivers/gpu/drm/drm_irq.c
@@ -505,6 +505,7 @@ void drm_calc_timestamping_constants(struct drm_crtc *crtc)
 
 	/* Valid dotclock? */
 	if (dotclock > 0) {
+		int frame_size;
 		/* Convert scanline length in pixels and video dot clock to
 		 * line duration, frame duration and pixel duration in
 		 * nanoseconds:
@@ -512,7 +513,10 @@ void drm_calc_timestamping_constants(struct drm_crtc *crtc)
 		pixeldur_ns = (s64) div64_u64(1000000000, dotclock);
 		linedur_ns  = (s64) div64_u64(((u64) crtc->hwmode.crtc_htotal *
 					      1000000000), dotclock);
-		framedur_ns = (s64) crtc->hwmode.crtc_vtotal * linedur_ns;
+		frame_size = crtc->hwmode.crtc_htotal *
+				crtc->hwmode.crtc_vtotal;
+		framedur_ns = (s64) div64_u64((u64) frame_size * 1000000000,
+					      dotclock);
 	} else
 		DRM_ERROR("crtc %d: Can't calculate constants, dotclock = 0!\n",
 			  crtc->base.id);
-- 
cgit v0.10.2


From db3985e5ca8f50fc17606855ba394783d11683a5 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Wed, 13 Feb 2013 20:37:48 +1000
Subject: Revert "drm: Add EDID_QUIRK_FORCE_REDUCED_BLANKING for ASUS VW222S"

This reverts commit 6f33814bd4d9cfe76033a31b1c0c76c960cd8e4b.

The quirk cause a regression, and it looks like the original bug was
simply a lack of FIFO bandwidth on the i915G of the reporter. Which
should eventually be fixed as soon as we get around to implemented
DSPARB FIFO reassignment on gen 3.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=52281
Cc: stable@vger.kernel.org
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
index a3a3b61..5132425 100644
--- a/drivers/gpu/drm/drm_edid.c
+++ b/drivers/gpu/drm/drm_edid.c
@@ -87,9 +87,6 @@ static struct edid_quirk {
 	int product_id;
 	u32 quirks;
 } edid_quirk_list[] = {
-	/* ASUS VW222S */
-	{ "ACI", 0x22a2, EDID_QUIRK_FORCE_REDUCED_BLANKING },
-
 	/* Acer AL1706 */
 	{ "ACR", 44358, EDID_QUIRK_PREFER_LARGE_60 },
 	/* Acer F51 */
-- 
cgit v0.10.2


From 6aed8ec3f76a22217c9ae183d32b1aa990bed069 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sun, 20 Jan 2013 17:32:21 +0100
Subject: drm: review locking for drm_fb_helper_restore_fbdev_mode

... it's required. Fix up exynos and the cma helper, and add a
corresponding WARN_ON to drm_fb_helper_restore_fbdev_mode.

Note that tegra calls the fbdev cma helper restore function also from
it's driver-load callback. Which is a bit against current practice,
since usually the call is only from ->lastclose, and initial setup is
done by drm_fb_helper_initial_config.

Also add the relevant drm DocBook entry.

v2: Add promised WARN to restore_fbdev_mode.

Reviewed-by: Rob Clark <robdclark@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index 9c797f6..f170773 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -68,9 +68,23 @@ void drm_modeset_unlock_all(struct drm_device *dev)
 
 	mutex_unlock(&dev->mode_config.mutex);
 }
-
 EXPORT_SYMBOL(drm_modeset_unlock_all);
 
+/**
+ * drm_warn_on_modeset_not_all_locked - check that all modeset locks are locked
+ * @dev: device
+ */
+void drm_warn_on_modeset_not_all_locked(struct drm_device *dev)
+{
+	struct drm_crtc *crtc;
+
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head)
+		WARN_ON(!mutex_is_locked(&crtc->mutex));
+
+	WARN_ON(!mutex_is_locked(&dev->mode_config.mutex));
+}
+EXPORT_SYMBOL(drm_warn_on_modeset_not_all_locked);
+
 /* Avoid boilerplate.  I'm tired of typing. */
 #define DRM_ENUM_NAME_FN(fnname, list)				\
 	char *fnname(int val)					\
diff --git a/drivers/gpu/drm/drm_fb_cma_helper.c b/drivers/gpu/drm/drm_fb_cma_helper.c
index 3742bc9..1b6ba2d 100644
--- a/drivers/gpu/drm/drm_fb_cma_helper.c
+++ b/drivers/gpu/drm/drm_fb_cma_helper.c
@@ -389,8 +389,10 @@ EXPORT_SYMBOL_GPL(drm_fbdev_cma_fini);
  */
 void drm_fbdev_cma_restore_mode(struct drm_fbdev_cma *fbdev_cma)
 {
+	drm_modeset_lock_all(dev);
 	if (fbdev_cma)
 		drm_fb_helper_restore_fbdev_mode(&fbdev_cma->fb_helper);
+	drm_modeset_unlock_all(dev);
 }
 EXPORT_SYMBOL_GPL(drm_fbdev_cma_restore_mode);
 
diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index 0c6e25e..77378a1 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -239,10 +239,21 @@ int drm_fb_helper_debug_leave(struct fb_info *info)
 }
 EXPORT_SYMBOL(drm_fb_helper_debug_leave);
 
+/**
+ * drm_fb_helper_restore_fbdev_mode - restore fbdev configuration
+ * @fb_helper: fbcon to restore
+ *
+ * This should be called from driver's drm->lastclose callback when implementing
+ * an fbcon on top of kms using this helper. This ensures that the user isn't
+ * greeted with a black screen when e.g. X dies.
+ */
 bool drm_fb_helper_restore_fbdev_mode(struct drm_fb_helper *fb_helper)
 {
 	bool error = false;
 	int i, ret;
+
+	drm_warn_on_modeset_not_all_locked(fb_helper->dev);
+
 	for (i = 0; i < fb_helper->crtc_count; i++) {
 		struct drm_mode_set *mode_set = &fb_helper->crtc_info[i].mode_set;
 		ret = drm_mode_set_config_internal(mode_set);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
index 90d335c..086d0f7 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
@@ -376,5 +376,7 @@ void exynos_drm_fbdev_restore_mode(struct drm_device *dev)
 	if (!private || !private->fb_helper)
 		return;
 
+	drm_modeset_lock_all(dev);
 	drm_fb_helper_restore_fbdev_mode(private->fb_helper);
+	drm_modeset_unlock_all(dev);
 }
diff --git a/include/drm/drm_crtc.h b/include/drm/drm_crtc.h
index e90c8dc..8b77627 100644
--- a/include/drm/drm_crtc.h
+++ b/include/drm/drm_crtc.h
@@ -867,6 +867,7 @@ struct drm_prop_enum_list {
 
 extern void drm_modeset_lock_all(struct drm_device *dev);
 extern void drm_modeset_unlock_all(struct drm_device *dev);
+extern void drm_warn_on_modeset_not_all_locked(struct drm_device *dev);
 
 extern int drm_crtc_init(struct drm_device *dev,
 			 struct drm_crtc *crtc,
-- 
cgit v0.10.2


From d21bf469d5301d025cd82997bb1529bcdc7086af Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sun, 20 Jan 2013 18:09:52 +0100
Subject: drm/fb-helper: kill drm_fb_helper_restore

It's only used internally for the sysrq and panic handlers provided by
the drm fb helper implementation. Hence just inline it, kill the
export and remove the confusing kerneldoc. Driver's are supposed to
call drm_fb_helper_restore_fbdev_mode on lastclose.

Note that locking is totally fubar - the sysrq case doesn't take any
locks at all. The panic handler probably shouldn't take any locks
since it'll only make things worse. Otoh it's probably better to
switch things over to the atomic modeset callbacks (and disable the
panic handler for those drivers which don't implement it).

But that's both better done in separate patches.

Reviewed-by: Rob Clark <robdclark@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index 77378a1..14a81fc 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -264,6 +264,10 @@ bool drm_fb_helper_restore_fbdev_mode(struct drm_fb_helper *fb_helper)
 }
 EXPORT_SYMBOL(drm_fb_helper_restore_fbdev_mode);
 
+/*
+ * restore fbcon display for all kms driver's using this helper, used for sysrq
+ * and panic handling.
+ */
 static bool drm_fb_helper_force_kernel_mode(void)
 {
 	bool ret, error = false;
@@ -302,20 +306,6 @@ static struct notifier_block paniced = {
 	.notifier_call = drm_fb_helper_panic,
 };
 
-/**
- * drm_fb_helper_restore - restore the framebuffer console (kernel) config
- *
- * Restore's the kernel's fbcon mode, used for lastclose & panic paths.
- */
-void drm_fb_helper_restore(void)
-{
-	bool ret;
-	ret = drm_fb_helper_force_kernel_mode();
-	if (ret == true)
-		DRM_ERROR("Failed to restore crtc configuration\n");
-}
-EXPORT_SYMBOL(drm_fb_helper_restore);
-
 static bool drm_fb_helper_is_bound(struct drm_fb_helper *fb_helper)
 {
 	struct drm_device *dev = fb_helper->dev;
@@ -337,7 +327,10 @@ static bool drm_fb_helper_is_bound(struct drm_fb_helper *fb_helper)
 #ifdef CONFIG_MAGIC_SYSRQ
 static void drm_fb_helper_restore_work_fn(struct work_struct *ignored)
 {
-	drm_fb_helper_restore();
+	bool ret;
+	ret = drm_fb_helper_force_kernel_mode();
+	if (ret == true)
+		DRM_ERROR("Failed to restore crtc configuration\n");
 }
 static DECLARE_WORK(drm_fb_helper_restore_work, drm_fb_helper_restore_work_fn);
 
diff --git a/include/drm/drm_fb_helper.h b/include/drm/drm_fb_helper.h
index 5120b01..ba32505 100644
--- a/include/drm/drm_fb_helper.h
+++ b/include/drm/drm_fb_helper.h
@@ -103,7 +103,6 @@ int drm_fb_helper_setcolreg(unsigned regno,
 			    struct fb_info *info);
 
 bool drm_fb_helper_restore_fbdev_mode(struct drm_fb_helper *fb_helper);
-void drm_fb_helper_restore(void);
 void drm_fb_helper_fill_var(struct fb_info *info, struct drm_fb_helper *fb_helper,
 			    uint32_t fb_width, uint32_t fb_height);
 void drm_fb_helper_fill_fix(struct fb_info *info, uint32_t pitch,
-- 
cgit v0.10.2


From 43c8a849a1e92fba6ea4493ed950d2a2e31ac87c Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sun, 20 Jan 2013 18:18:07 +0100
Subject: drm/fb-helper: unexport drm_fb_helper_panic

It doesn't even show up in any header files and only used iternally.
Originally it was (ab)used to restore the fbcon on lastclose, but that
died with

commit e8e7a2b8ccfdae0d4cb6bd25824bbedcd42da316
Author: Dave Airlie <airlied@redhat.com>
Date:   Thu Apr 21 22:18:32 2011 +0100

    drm/i915: restore only the mode of this driver on lastclose (v2)

Reviewed-by: Rob Clark <robdclark@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index 14a81fc..eaab092 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -287,7 +287,7 @@ static bool drm_fb_helper_force_kernel_mode(void)
 	return error;
 }
 
-int drm_fb_helper_panic(struct notifier_block *n, unsigned long ununsed,
+static int drm_fb_helper_panic(struct notifier_block *n, unsigned long ununsed,
 			void *panic_str)
 {
 	/*
@@ -300,7 +300,6 @@ int drm_fb_helper_panic(struct notifier_block *n, unsigned long ununsed,
 	pr_err("panic occurred, switching back to text console\n");
 	return drm_fb_helper_force_kernel_mode();
 }
-EXPORT_SYMBOL(drm_fb_helper_panic);
 
 static struct notifier_block paniced = {
 	.notifier_call = drm_fb_helper_panic,
-- 
cgit v0.10.2


From de1ace5b56d11db26abe69fd19754f91c326763f Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sun, 20 Jan 2013 21:50:49 +0100
Subject: drm/fb-helper: unexport drm_fb_helper_single_fb_probe

Not called by anyone, and really, shouldn't be. Drivers are supposed
either drm_fb_helper_initial_config or drm_fb_helper_hotplug_event.
Originally this was done differently, but is now consolidated in the
helper functions and no longer done by drivers directly.

Reviewed-by: Rob Clark <robdclark@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index eaab092..f5d3626 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -754,8 +754,8 @@ int drm_fb_helper_pan_display(struct fb_var_screeninfo *var,
 }
 EXPORT_SYMBOL(drm_fb_helper_pan_display);
 
-int drm_fb_helper_single_fb_probe(struct drm_fb_helper *fb_helper,
-				  int preferred_bpp)
+static int drm_fb_helper_single_fb_probe(struct drm_fb_helper *fb_helper,
+					 int preferred_bpp)
 {
 	int new_fb = 0;
 	int crtc_count = 0;
@@ -870,7 +870,6 @@ int drm_fb_helper_single_fb_probe(struct drm_fb_helper *fb_helper,
 
 	return 0;
 }
-EXPORT_SYMBOL(drm_fb_helper_single_fb_probe);
 
 void drm_fb_helper_fill_fix(struct fb_info *info, uint32_t pitch,
 			    uint32_t depth)
diff --git a/include/drm/drm_fb_helper.h b/include/drm/drm_fb_helper.h
index ba32505..973402d 100644
--- a/include/drm/drm_fb_helper.h
+++ b/include/drm/drm_fb_helper.h
@@ -82,9 +82,6 @@ struct drm_fb_helper {
 	bool delayed_hotplug;
 };
 
-int drm_fb_helper_single_fb_probe(struct drm_fb_helper *helper,
-				  int preferred_bpp);
-
 int drm_fb_helper_init(struct drm_device *dev,
 		       struct drm_fb_helper *helper, int crtc_count,
 		       int max_conn);
-- 
cgit v0.10.2


From 203cb50143029b2bc95736ce10f7defcf59aca44 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sun, 20 Jan 2013 21:56:20 +0100
Subject: drm/tegra: don't set up initial fbcon config twice

drm_fbdev_cma_init does the inital fbcon setup by calling down into
drm_fb_helper_initial_config, so no need at all to restore the just
set up configuration right away ...

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/tegra/fb.c b/drivers/gpu/drm/tegra/fb.c
index 97993c6..0391495 100644
--- a/drivers/gpu/drm/tegra/fb.c
+++ b/drivers/gpu/drm/tegra/fb.c
@@ -39,10 +39,6 @@ int tegra_drm_fb_init(struct drm_device *drm)
 	if (IS_ERR(fbdev))
 		return PTR_ERR(fbdev);
 
-#ifndef CONFIG_FRAMEBUFFER_CONSOLE
-	drm_fbdev_cma_restore_mode(fbdev);
-#endif
-
 	host1x->fbdev = fbdev;
 
 	return 0;
-- 
cgit v0.10.2


From 76a39dbfb2d1bc45219839e5a95d4ceaf6ca114f Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sun, 20 Jan 2013 23:12:54 +0100
Subject: drm/fb-helper: don't disable everything in initial_config

This should be done in the drivers for two reasons:
- it gets in the way of fastboot efforts
- it links the fb helpers with the crtc helpers instead of going
  through the real interface vfuncs, forcing i915 to fake all the
  ->disable callbacks used by the crtc helper to avoid ugly Oopsen

v2: Resolve conflicts since drivers still call
drm_fb_helper_single_add_all_connectors.

Reviewed-by: Rob Clark <robdclark@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/ast/ast_fb.c b/drivers/gpu/drm/ast/ast_fb.c
index 3e6584b..81763ca 100644
--- a/drivers/gpu/drm/ast/ast_fb.c
+++ b/drivers/gpu/drm/ast/ast_fb.c
@@ -40,6 +40,7 @@
 #include <drm/drmP.h>
 #include <drm/drm_crtc.h>
 #include <drm/drm_fb_helper.h>
+#include <drm/drm_crtc_helper.h>
 #include "ast_drv.h"
 
 static void ast_dirty_update(struct ast_fbdev *afbdev,
@@ -314,6 +315,10 @@ int ast_fbdev_init(struct drm_device *dev)
 	}
 
 	drm_fb_helper_single_add_all_connectors(&afbdev->helper);
+
+	/* disable all the possible outputs/crtcs before entering KMS mode */
+	drm_helper_disable_unused_functions(dev);
+
 	drm_fb_helper_initial_config(&afbdev->helper, 32);
 	return 0;
 }
diff --git a/drivers/gpu/drm/cirrus/cirrus_fbdev.c b/drivers/gpu/drm/cirrus/cirrus_fbdev.c
index 3daea0f..b96605c 100644
--- a/drivers/gpu/drm/cirrus/cirrus_fbdev.c
+++ b/drivers/gpu/drm/cirrus/cirrus_fbdev.c
@@ -11,6 +11,7 @@
 #include <linux/module.h>
 #include <drm/drmP.h>
 #include <drm/drm_fb_helper.h>
+#include <drm/drm_crtc_helper.h>
 
 #include <linux/fb.h>
 
@@ -291,6 +292,9 @@ int cirrus_fbdev_init(struct cirrus_device *cdev)
 		return ret;
 	}
 	drm_fb_helper_single_add_all_connectors(&gfbdev->helper);
+
+	/* disable all the possible outputs/crtcs before entering KMS mode */
+	drm_helper_disable_unused_functions(cdev->dev);
 	drm_fb_helper_initial_config(&gfbdev->helper, bpp_sel);
 
 	return 0;
diff --git a/drivers/gpu/drm/drm_fb_cma_helper.c b/drivers/gpu/drm/drm_fb_cma_helper.c
index 1b6ba2d..ef3d33a 100644
--- a/drivers/gpu/drm/drm_fb_cma_helper.c
+++ b/drivers/gpu/drm/drm_fb_cma_helper.c
@@ -333,6 +333,9 @@ struct drm_fbdev_cma *drm_fbdev_cma_init(struct drm_device *dev,
 
 	}
 
+	/* disable all the possible outputs/crtcs before entering KMS mode */
+	drm_helper_disable_unused_functions(dev);
+
 	ret = drm_fb_helper_initial_config(helper, preferred_bpp);
 	if (ret < 0) {
 		dev_err(dev->dev, "Failed to set inital hw configuration.\n");
diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index f5d3626..d841b68 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -1360,9 +1360,6 @@ bool drm_fb_helper_initial_config(struct drm_fb_helper *fb_helper, int bpp_sel)
 	struct drm_device *dev = fb_helper->dev;
 	int count = 0;
 
-	/* disable all the possible outputs/crtcs before entering KMS mode */
-	drm_helper_disable_unused_functions(fb_helper->dev);
-
 	drm_fb_helper_parse_command_line(fb_helper);
 
 	count = drm_fb_helper_probe_connector_modes(fb_helper,
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
index 086d0f7..fe2a0f0 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
@@ -295,6 +295,9 @@ int exynos_drm_fbdev_init(struct drm_device *dev)
 
 	}
 
+	/* disable all the possible outputs/crtcs before entering KMS mode */
+	drm_helper_disable_unused_functions(dev);
+
 	ret = drm_fb_helper_initial_config(helper, PREFERRED_BPP);
 	if (ret < 0) {
 		DRM_ERROR("failed to set up hw configuration.\n");
diff --git a/drivers/gpu/drm/gma500/framebuffer.c b/drivers/gpu/drm/gma500/framebuffer.c
index c1ef37e..fee3bf8 100644
--- a/drivers/gpu/drm/gma500/framebuffer.c
+++ b/drivers/gpu/drm/gma500/framebuffer.c
@@ -616,6 +616,10 @@ int psb_fbdev_init(struct drm_device *dev)
 							INTELFB_CONN_LIMIT);
 
 	drm_fb_helper_single_add_all_connectors(&fbdev->psb_fb_helper);
+
+	/* disable all the possible outputs/crtcs before entering KMS mode */
+	drm_helper_disable_unused_functions(dev);
+
 	drm_fb_helper_initial_config(&fbdev->psb_fb_helper, 32);
 	return 0;
 }
diff --git a/drivers/gpu/drm/i915/intel_fb.c b/drivers/gpu/drm/i915/intel_fb.c
index 1c510da..e670612 100644
--- a/drivers/gpu/drm/i915/intel_fb.c
+++ b/drivers/gpu/drm/i915/intel_fb.c
@@ -258,6 +258,9 @@ void intel_fbdev_initial_config(struct drm_device *dev)
 {
 	drm_i915_private_t *dev_priv = dev->dev_private;
 
+	/* disable all the possible outputs/crtcs before entering KMS mode */
+	drm_helper_disable_unused_functions(dev);
+
 	/* Due to peculiar init order wrt to hpd handling this is separate. */
 	drm_fb_helper_initial_config(&dev_priv->fbdev->helper, 32);
 }
diff --git a/drivers/gpu/drm/mgag200/mgag200_fb.c b/drivers/gpu/drm/mgag200/mgag200_fb.c
index 5c69b43..5bded5b 100644
--- a/drivers/gpu/drm/mgag200/mgag200_fb.c
+++ b/drivers/gpu/drm/mgag200/mgag200_fb.c
@@ -13,6 +13,7 @@
 #include <linux/module.h>
 #include <drm/drmP.h>
 #include <drm/drm_fb_helper.h>
+#include <drm/drm_crtc_helper.h>
 
 #include <linux/fb.h>
 
@@ -278,6 +279,10 @@ int mgag200_fbdev_init(struct mga_device *mdev)
 		return ret;
 	}
 	drm_fb_helper_single_add_all_connectors(&mfbdev->helper);
+
+	/* disable all the possible outputs/crtcs before entering KMS mode */
+	drm_helper_disable_unused_functions(mdev->dev);
+
 	drm_fb_helper_initial_config(&mfbdev->helper, 32);
 
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/nouveau_fbcon.c b/drivers/gpu/drm/nouveau/nouveau_fbcon.c
index d4ecb4d..b1ebfe3 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fbcon.c
+++ b/drivers/gpu/drm/nouveau/nouveau_fbcon.c
@@ -491,6 +491,9 @@ nouveau_fbcon_init(struct drm_device *dev)
 	else
 		preferred_bpp = 32;
 
+	/* disable all the possible outputs/crtcs before entering KMS mode */
+	drm_helper_disable_unused_functions(dev);
+
 	drm_fb_helper_initial_config(&fbcon->helper, preferred_bpp);
 	return 0;
 }
diff --git a/drivers/gpu/drm/radeon/radeon_fb.c b/drivers/gpu/drm/radeon/radeon_fb.c
index 515e5ee..b48d1c8 100644
--- a/drivers/gpu/drm/radeon/radeon_fb.c
+++ b/drivers/gpu/drm/radeon/radeon_fb.c
@@ -379,6 +379,10 @@ int radeon_fbdev_init(struct radeon_device *rdev)
 	}
 
 	drm_fb_helper_single_add_all_connectors(&rfbdev->helper);
+
+	/* disable all the possible outputs/crtcs before entering KMS mode */
+	drm_helper_disable_unused_functions(rdev->ddev);
+
 	drm_fb_helper_initial_config(&rfbdev->helper, bpp_sel);
 	return 0;
 }
diff --git a/drivers/gpu/drm/udl/udl_fb.c b/drivers/gpu/drm/udl/udl_fb.c
index b9feec9..cf5d05a 100644
--- a/drivers/gpu/drm/udl/udl_fb.c
+++ b/drivers/gpu/drm/udl/udl_fb.c
@@ -619,6 +619,10 @@ int udl_fbdev_init(struct drm_device *dev)
 	}
 
 	drm_fb_helper_single_add_all_connectors(&ufbdev->helper);
+
+	/* disable all the possible outputs/crtcs before entering KMS mode */
+	drm_helper_disable_unused_functions(dev);
+
 	drm_fb_helper_initial_config(&ufbdev->helper, bpp_sel);
 	return 0;
 }
diff --git a/drivers/staging/omapdrm/omap_fbdev.c b/drivers/staging/omapdrm/omap_fbdev.c
index 2728e37..7e66eb1 100644
--- a/drivers/staging/omapdrm/omap_fbdev.c
+++ b/drivers/staging/omapdrm/omap_fbdev.c
@@ -369,6 +369,10 @@ struct drm_fb_helper *omap_fbdev_init(struct drm_device *dev)
 	}
 
 	drm_fb_helper_single_add_all_connectors(helper);
+
+	/* disable all the possible outputs/crtcs before entering KMS mode */
+	drm_helper_disable_unused_functions(dev);
+
 	drm_fb_helper_initial_config(helper, 32);
 
 	priv->fbdev = helper;
-- 
cgit v0.10.2


From af5676f1f91585cabe811b8f697e32015e2be826 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sun, 20 Jan 2013 23:28:14 +0100
Subject: drm/i915: rip out helper->disable noop functions

Now that the driver is in control of whether it needs to disable
everything at take-over or not, we can rip this all out.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_crt.c b/drivers/gpu/drm/i915/intel_crt.c
index 68e79f3..729dd1a 100644
--- a/drivers/gpu/drm/i915/intel_crt.c
+++ b/drivers/gpu/drm/i915/intel_crt.c
@@ -685,7 +685,6 @@ static void intel_crt_reset(struct drm_connector *connector)
 static const struct drm_encoder_helper_funcs crt_encoder_funcs = {
 	.mode_fixup = intel_crt_mode_fixup,
 	.mode_set = intel_crt_mode_set,
-	.disable = intel_encoder_noop,
 };
 
 static const struct drm_connector_funcs intel_crt_connector_funcs = {
diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c
index cedf4ab..a259e09 100644
--- a/drivers/gpu/drm/i915/intel_ddi.c
+++ b/drivers/gpu/drm/i915/intel_ddi.c
@@ -1481,7 +1481,6 @@ static const struct drm_encoder_funcs intel_ddi_funcs = {
 static const struct drm_encoder_helper_funcs intel_ddi_helper_funcs = {
 	.mode_fixup = intel_ddi_mode_fixup,
 	.mode_set = intel_ddi_mode_set,
-	.disable = intel_encoder_noop,
 };
 
 void intel_ddi_init(struct drm_device *dev, enum port port)
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 0dfecaf..24f2654 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -3718,10 +3718,6 @@ void intel_crtc_update_dpms(struct drm_crtc *crtc)
 	intel_crtc_update_sarea(crtc, enable);
 }
 
-static void intel_crtc_noop(struct drm_crtc *crtc)
-{
-}
-
 static void intel_crtc_disable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
@@ -3770,10 +3766,6 @@ void intel_modeset_disable(struct drm_device *dev)
 	}
 }
 
-void intel_encoder_noop(struct drm_encoder *encoder)
-{
-}
-
 void intel_encoder_destroy(struct drm_encoder *encoder)
 {
 	struct intel_encoder *intel_encoder = to_intel_encoder(encoder);
@@ -7277,7 +7269,6 @@ free_work:
 static struct drm_crtc_helper_funcs intel_helper_funcs = {
 	.mode_set_base_atomic = intel_pipe_set_base_atomic,
 	.load_lut = intel_crtc_load_lut,
-	.disable = intel_crtc_noop,
 };
 
 bool intel_encoder_check_is_cloned(struct intel_encoder *encoder)
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 56e408b..13c1536 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -2561,7 +2561,6 @@ void intel_dp_encoder_destroy(struct drm_encoder *encoder)
 static const struct drm_encoder_helper_funcs intel_dp_helper_funcs = {
 	.mode_fixup = intel_dp_mode_fixup,
 	.mode_set = intel_dp_mode_set,
-	.disable = intel_encoder_noop,
 };
 
 static const struct drm_connector_funcs intel_dp_connector_funcs = {
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 13afb37d..a2ec01c 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -521,7 +521,6 @@ extern void intel_modeset_disable(struct drm_device *dev);
 extern void intel_crtc_restore_mode(struct drm_crtc *crtc);
 extern void intel_crtc_load_lut(struct drm_crtc *crtc);
 extern void intel_crtc_update_dpms(struct drm_crtc *crtc);
-extern void intel_encoder_noop(struct drm_encoder *encoder);
 extern void intel_encoder_destroy(struct drm_encoder *encoder);
 extern void intel_encoder_dpms(struct intel_encoder *encoder, int mode);
 extern bool intel_encoder_check_is_cloned(struct intel_encoder *encoder);
diff --git a/drivers/gpu/drm/i915/intel_dvo.c b/drivers/gpu/drm/i915/intel_dvo.c
index 15da995..00e70db 100644
--- a/drivers/gpu/drm/i915/intel_dvo.c
+++ b/drivers/gpu/drm/i915/intel_dvo.c
@@ -345,7 +345,6 @@ static void intel_dvo_destroy(struct drm_connector *connector)
 static const struct drm_encoder_helper_funcs intel_dvo_helper_funcs = {
 	.mode_fixup = intel_dvo_mode_fixup,
 	.mode_set = intel_dvo_mode_set,
-	.disable = intel_encoder_noop,
 };
 
 static const struct drm_connector_funcs intel_dvo_connector_funcs = {
diff --git a/drivers/gpu/drm/i915/intel_fb.c b/drivers/gpu/drm/i915/intel_fb.c
index e670612..1c510da 100644
--- a/drivers/gpu/drm/i915/intel_fb.c
+++ b/drivers/gpu/drm/i915/intel_fb.c
@@ -258,9 +258,6 @@ void intel_fbdev_initial_config(struct drm_device *dev)
 {
 	drm_i915_private_t *dev_priv = dev->dev_private;
 
-	/* disable all the possible outputs/crtcs before entering KMS mode */
-	drm_helper_disable_unused_functions(dev);
-
 	/* Due to peculiar init order wrt to hpd handling this is separate. */
 	drm_fb_helper_initial_config(&dev_priv->fbdev->helper, 32);
 }
diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c
index 5b4efd6..3883bed 100644
--- a/drivers/gpu/drm/i915/intel_hdmi.c
+++ b/drivers/gpu/drm/i915/intel_hdmi.c
@@ -972,7 +972,6 @@ static void intel_hdmi_destroy(struct drm_connector *connector)
 static const struct drm_encoder_helper_funcs intel_hdmi_helper_funcs = {
 	.mode_fixup = intel_hdmi_mode_fixup,
 	.mode_set = intel_hdmi_mode_set,
-	.disable = intel_encoder_noop,
 };
 
 static const struct drm_connector_funcs intel_hdmi_connector_funcs = {
diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c
index 5e3f08e..feb43fd 100644
--- a/drivers/gpu/drm/i915/intel_lvds.c
+++ b/drivers/gpu/drm/i915/intel_lvds.c
@@ -656,7 +656,6 @@ static int intel_lvds_set_property(struct drm_connector *connector,
 static const struct drm_encoder_helper_funcs intel_lvds_helper_funcs = {
 	.mode_fixup = intel_lvds_mode_fixup,
 	.mode_set = intel_lvds_mode_set,
-	.disable = intel_encoder_noop,
 };
 
 static const struct drm_connector_helper_funcs intel_lvds_connector_helper_funcs = {
diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c
index f01063a..33b46d9 100644
--- a/drivers/gpu/drm/i915/intel_sdvo.c
+++ b/drivers/gpu/drm/i915/intel_sdvo.c
@@ -2043,7 +2043,6 @@ done:
 static const struct drm_encoder_helper_funcs intel_sdvo_helper_funcs = {
 	.mode_fixup = intel_sdvo_mode_fixup,
 	.mode_set = intel_sdvo_mode_set,
-	.disable = intel_encoder_noop,
 };
 
 static const struct drm_connector_funcs intel_sdvo_connector_funcs = {
diff --git a/drivers/gpu/drm/i915/intel_tv.c b/drivers/gpu/drm/i915/intel_tv.c
index 984a113..d808421 100644
--- a/drivers/gpu/drm/i915/intel_tv.c
+++ b/drivers/gpu/drm/i915/intel_tv.c
@@ -1487,7 +1487,6 @@ out:
 static const struct drm_encoder_helper_funcs intel_tv_helper_funcs = {
 	.mode_fixup = intel_tv_mode_fixup,
 	.mode_set = intel_tv_mode_set,
-	.disable = intel_encoder_noop,
 };
 
 static const struct drm_connector_funcs intel_tv_connector_funcs = {
-- 
cgit v0.10.2


From 7e53f3a423146745a4e4bb93362d488dfad502a8 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Mon, 21 Jan 2013 10:52:17 +0100
Subject: drm/fb-helper: fixup set_config semantics

While doing the modeset rework for drm/i915 I've noticed that the fb
helper is very liberal with the semantics of the ->set_config
interface:
- It doesn't bother clearing stale modes (e.g. when unplugging a
  screen).
- It unconditionally sets the fb, even if no mode will be set on a
  given crtc.
- The initial setup is a bit fun since we need to pick crtcs to decide
  the desired fb size, but also should set the modeset->fb pointer.
  Explain what's going on in the fixup code after the fb is allocated.

The crtc helper didn't really care, but the new i915 modeset
infrastructure did, so I've had to add a bunch of special-cases to
catch this.

Fix this all up and enforce the interface by converting the checks in
drm/i915/intel_display.c to BUG_ONs.

v2: Fix commit message spell fail spotted by Rob Clark.

Reviewed-by: Rob Clark <robdclark@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index d841b68..809ef99 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -689,7 +689,6 @@ int drm_fb_helper_set_par(struct fb_info *info)
 	struct drm_fb_helper *fb_helper = info->par;
 	struct drm_device *dev = fb_helper->dev;
 	struct fb_var_screeninfo *var = &info->var;
-	struct drm_crtc *crtc;
 	int ret;
 	int i;
 
@@ -700,7 +699,6 @@ int drm_fb_helper_set_par(struct fb_info *info)
 
 	drm_modeset_lock_all(dev);
 	for (i = 0; i < fb_helper->crtc_count; i++) {
-		crtc = fb_helper->crtc_info[i].mode_set.crtc;
 		ret = drm_mode_set_config_internal(&fb_helper->crtc_info[i].mode_set);
 		if (ret) {
 			drm_modeset_unlock_all(dev);
@@ -841,9 +839,17 @@ static int drm_fb_helper_single_fb_probe(struct drm_fb_helper *fb_helper,
 
 	info = fb_helper->fbdev;
 
-	/* set the fb pointer */
+	/*
+	 * Set the fb pointer - usually drm_setup_crtcs does this for hotplug
+	 * events, but at init time drm_setup_crtcs needs to be called before
+	 * the fb is allocated (since we need to figure out the desired size of
+	 * the fb before we can allocate it ...). Hence we need to fix things up
+	 * here again.
+	 */
 	for (i = 0; i < fb_helper->crtc_count; i++)
-		fb_helper->crtc_info[i].mode_set.fb = fb_helper->fb;
+		if (fb_helper->crtc_info[i].mode_set.num_connectors)
+			fb_helper->crtc_info[i].mode_set.fb = fb_helper->fb;
+
 
 	if (new_fb) {
 		info->var.pixclock = 0;
@@ -1314,6 +1320,7 @@ static void drm_setup_crtcs(struct drm_fb_helper *fb_helper)
 	for (i = 0; i < fb_helper->crtc_count; i++) {
 		modeset = &fb_helper->crtc_info[i].mode_set;
 		modeset->num_connectors = 0;
+		modeset->fb = NULL;
 	}
 
 	for (i = 0; i < fb_helper->connector_count; i++) {
@@ -1330,9 +1337,21 @@ static void drm_setup_crtcs(struct drm_fb_helper *fb_helper)
 			modeset->mode = drm_mode_duplicate(dev,
 							   fb_crtc->desired_mode);
 			modeset->connectors[modeset->num_connectors++] = fb_helper->connector_info[i]->connector;
+			modeset->fb = fb_helper->fb;
 		}
 	}
 
+	/* Clear out any old modes if there are no more connected outputs. */
+	for (i = 0; i < fb_helper->crtc_count; i++) {
+		modeset = &fb_helper->crtc_info[i].mode_set;
+		if (modeset->num_connectors == 0) {
+			BUG_ON(modeset->fb);
+			BUG_ON(modeset->num_connectors);
+			if (modeset->mode)
+				drm_mode_destroy(dev, modeset->mode);
+			modeset->mode = NULL;
+		}
+	}
 out:
 	kfree(crtcs);
 	kfree(modes);
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 24f2654..ca8d592 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -7978,14 +7978,9 @@ static int intel_crtc_set_config(struct drm_mode_set *set)
 	BUG_ON(!set->crtc);
 	BUG_ON(!set->crtc->helper_private);
 
-	if (!set->mode)
-		set->fb = NULL;
-
-	/* The fb helper likes to play gross jokes with ->mode_set_config.
-	 * Unfortunately the crtc helper doesn't do much at all for this case,
-	 * so we have to cope with this madness until the fb helper is fixed up. */
-	if (set->fb && set->num_connectors == 0)
-		return 0;
+	/* Enforce sane interface api - has been abused by the fb helper. */
+	BUG_ON(!set->mode && set->fb);
+	BUG_ON(set->fb && set->num_connectors == 0);
 
 	if (set->fb) {
 		DRM_DEBUG_KMS("[CRTC:%d] [FB:%d] #connectors=%d (x y) (%i %i)\n",
-- 
cgit v0.10.2


From 2180c3c7e76536ce8ff0fd957a11dbfaa1e93403 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Mon, 21 Jan 2013 23:12:36 +0100
Subject: drm/fb-helper: directly call set_par from the hotplug handler

The idea behind calling down into the driver's ->fb_probe function on each
hotplug seems to be able to reallocate the backing storage (if e.g. a screen
with higher resolution gets added). But that requires quite a bit of work in the
fb helper itself, since currently we limit new screens to the currently
allocated fb. An no kms driver supports fbdev fb resizing.

So don't bother and start to simplify the code by calling drm_fb_helper_set_par
directly from the fbdev hotplug function, since that's the only thing left in
drm_fb_helper_single_fb_probe which does not concern itself with fb allocation
and initial setup. Follow-on patches will streamline the initial setup
code.

Reviewed-by: Rob Clark <robdclark@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index 809ef99..535c2ca 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -859,8 +859,6 @@ static int drm_fb_helper_single_fb_probe(struct drm_fb_helper *fb_helper,
 		dev_info(fb_helper->dev->dev, "fb%d: %s frame buffer device\n",
 				info->node, info->fix.id);
 
-	} else {
-		drm_fb_helper_set_par(info);
 	}
 
 	/* Switch back to kernel console on panic */
@@ -1436,7 +1434,9 @@ int drm_fb_helper_hotplug_event(struct drm_fb_helper *fb_helper)
 	drm_setup_crtcs(fb_helper);
 	drm_modeset_unlock_all(dev);
 
-	return drm_fb_helper_single_fb_probe(fb_helper, bpp_sel);
+	drm_fb_helper_set_par(fb_helper->fbdev);
+
+	return 0;
 }
 EXPORT_SYMBOL(drm_fb_helper_hotplug_event);
 
-- 
cgit v0.10.2


From 8acf658ad6246be7913647e6e76da6a539bff425 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Mon, 21 Jan 2013 23:38:37 +0100
Subject: drm/fb-helper: streamline drm_fb_helper_single_fb_probe

No need to check whether we've allocated a new fb since we're not
always doing that. Also, we always need to register the fbdev and add
it to the panic notifier.

Reviewed-by: Rob Clark <robdclark@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index 535c2ca..e0d4b04 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -752,10 +752,14 @@ int drm_fb_helper_pan_display(struct fb_var_screeninfo *var,
 }
 EXPORT_SYMBOL(drm_fb_helper_pan_display);
 
+/*
+ * Allocates the backing storage through the ->fb_probe callback and then
+ * registers the fbdev and sets up the panic notifier.
+ */
 static int drm_fb_helper_single_fb_probe(struct drm_fb_helper *fb_helper,
 					 int preferred_bpp)
 {
-	int new_fb = 0;
+	int ret = 0;
 	int crtc_count = 0;
 	int i;
 	struct fb_info *info;
@@ -833,9 +837,9 @@ static int drm_fb_helper_single_fb_probe(struct drm_fb_helper *fb_helper,
 	}
 
 	/* push down into drivers */
-	new_fb = (*fb_helper->funcs->fb_probe)(fb_helper, &sizes);
-	if (new_fb < 0)
-		return new_fb;
+	ret = (*fb_helper->funcs->fb_probe)(fb_helper, &sizes);
+	if (ret < 0)
+		return ret;
 
 	info = fb_helper->fbdev;
 
@@ -851,15 +855,12 @@ static int drm_fb_helper_single_fb_probe(struct drm_fb_helper *fb_helper,
 			fb_helper->crtc_info[i].mode_set.fb = fb_helper->fb;
 
 
-	if (new_fb) {
-		info->var.pixclock = 0;
-		if (register_framebuffer(info) < 0)
-			return -EINVAL;
-
-		dev_info(fb_helper->dev->dev, "fb%d: %s frame buffer device\n",
-				info->node, info->fix.id);
+	info->var.pixclock = 0;
+	if (register_framebuffer(info) < 0)
+		return -EINVAL;
 
-	}
+	dev_info(fb_helper->dev->dev, "fb%d: %s frame buffer device\n",
+			info->node, info->fix.id);
 
 	/* Switch back to kernel console on panic */
 	/* multi card linked list maybe */
@@ -869,8 +870,8 @@ static int drm_fb_helper_single_fb_probe(struct drm_fb_helper *fb_helper,
 					       &paniced);
 		register_sysrq_key('v', &sysrq_drm_fb_helper_restore_op);
 	}
-	if (new_fb)
-		list_add(&fb_helper->kernel_fb_list, &kernel_fb_helper_list);
+
+	list_add(&fb_helper->kernel_fb_list, &kernel_fb_helper_list);
 
 	return 0;
 }
-- 
cgit v0.10.2


From cd5428a5447cc6ca77ec6547d6f86834b205eac7 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Mon, 21 Jan 2013 23:42:49 +0100
Subject: drm/<drivers>: simplify ->fb_probe callback

The fb helper lost its support for reallocating an fb completely, so
no need to return special success values any more.

Reviewed-by: Rob Clark <robdclark@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/ast/ast_fb.c b/drivers/gpu/drm/ast/ast_fb.c
index 81763ca..34931fe 100644
--- a/drivers/gpu/drm/ast/ast_fb.c
+++ b/drivers/gpu/drm/ast/ast_fb.c
@@ -146,9 +146,10 @@ static int astfb_create_object(struct ast_fbdev *afbdev,
 	return ret;
 }
 
-static int astfb_create(struct ast_fbdev *afbdev,
+static int astfb_create(struct drm_fb_helper *helper,
 			struct drm_fb_helper_surface_size *sizes)
 {
+	struct ast_fbdev *afbdev = (struct ast_fbdev *)helper;
 	struct drm_device *dev = afbdev->helper.dev;
 	struct drm_mode_fb_cmd2 mode_cmd;
 	struct drm_framebuffer *fb;
@@ -249,26 +250,10 @@ static void ast_fb_gamma_get(struct drm_crtc *crtc, u16 *red, u16 *green,
 	*blue = ast_crtc->lut_b[regno] << 8;
 }
 
-static int ast_find_or_create_single(struct drm_fb_helper *helper,
-					  struct drm_fb_helper_surface_size *sizes)
-{
-	struct ast_fbdev *afbdev = (struct ast_fbdev *)helper;
-	int new_fb = 0;
-	int ret;
-
-	if (!helper->fb) {
-		ret = astfb_create(afbdev, sizes);
-		if (ret)
-			return ret;
-		new_fb = 1;
-	}
-	return new_fb;
-}
-
 static struct drm_fb_helper_funcs ast_fb_helper_funcs = {
 	.gamma_set = ast_fb_gamma_set,
 	.gamma_get = ast_fb_gamma_get,
-	.fb_probe = ast_find_or_create_single,
+	.fb_probe = astfb_create,
 };
 
 static void ast_fbdev_destroy(struct drm_device *dev,
diff --git a/drivers/gpu/drm/cirrus/cirrus_fbdev.c b/drivers/gpu/drm/cirrus/cirrus_fbdev.c
index b96605c..e25afcc 100644
--- a/drivers/gpu/drm/cirrus/cirrus_fbdev.c
+++ b/drivers/gpu/drm/cirrus/cirrus_fbdev.c
@@ -121,9 +121,10 @@ static int cirrusfb_create_object(struct cirrus_fbdev *afbdev,
 	return ret;
 }
 
-static int cirrusfb_create(struct cirrus_fbdev *gfbdev,
+static int cirrusfb_create(struct drm_fb_helper *helper,
 			   struct drm_fb_helper_surface_size *sizes)
 {
+	struct cirrus_fbdev *gfbdev = (struct cirrus_fbdev *)helper;
 	struct drm_device *dev = gfbdev->helper.dev;
 	struct cirrus_device *cdev = gfbdev->helper.dev->dev_private;
 	struct fb_info *info;
@@ -220,23 +221,6 @@ out_iounmap:
 	return ret;
 }
 
-static int cirrus_fb_find_or_create_single(struct drm_fb_helper *helper,
-					   struct drm_fb_helper_surface_size
-					   *sizes)
-{
-	struct cirrus_fbdev *gfbdev = (struct cirrus_fbdev *)helper;
-	int new_fb = 0;
-	int ret;
-
-	if (!helper->fb) {
-		ret = cirrusfb_create(gfbdev, sizes);
-		if (ret)
-			return ret;
-		new_fb = 1;
-	}
-	return new_fb;
-}
-
 static int cirrus_fbdev_destroy(struct drm_device *dev,
 				struct cirrus_fbdev *gfbdev)
 {
@@ -268,7 +252,7 @@ static int cirrus_fbdev_destroy(struct drm_device *dev,
 static struct drm_fb_helper_funcs cirrus_fb_helper_funcs = {
 	.gamma_set = cirrus_crtc_fb_gamma_set,
 	.gamma_get = cirrus_crtc_fb_gamma_get,
-	.fb_probe = cirrus_fb_find_or_create_single,
+	.fb_probe = cirrusfb_create,
 };
 
 int cirrus_fbdev_init(struct cirrus_device *cdev)
diff --git a/drivers/gpu/drm/drm_fb_cma_helper.c b/drivers/gpu/drm/drm_fb_cma_helper.c
index ef3d33a..e851658 100644
--- a/drivers/gpu/drm/drm_fb_cma_helper.c
+++ b/drivers/gpu/drm/drm_fb_cma_helper.c
@@ -275,23 +275,8 @@ err_drm_gem_cma_free_object:
 	return ret;
 }
 
-static int drm_fbdev_cma_probe(struct drm_fb_helper *helper,
-	struct drm_fb_helper_surface_size *sizes)
-{
-	int ret = 0;
-
-	if (!helper->fb) {
-		ret = drm_fbdev_cma_create(helper, sizes);
-		if (ret < 0)
-			return ret;
-		ret = 1;
-	}
-
-	return ret;
-}
-
 static struct drm_fb_helper_funcs drm_fb_cma_helper_funcs = {
-	.fb_probe = drm_fbdev_cma_probe,
+	.fb_probe = drm_fbdev_cma_create,
 };
 
 /**
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
index fe2a0f0..68f0045 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
@@ -226,36 +226,8 @@ out:
 	return ret;
 }
 
-static int exynos_drm_fbdev_probe(struct drm_fb_helper *helper,
-				   struct drm_fb_helper_surface_size *sizes)
-{
-	int ret = 0;
-
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
-	/*
-	 * with !helper->fb, it means that this funcion is called first time
-	 * and after that, the helper->fb would be used as clone mode.
-	 */
-	if (!helper->fb) {
-		ret = exynos_drm_fbdev_create(helper, sizes);
-		if (ret < 0) {
-			DRM_ERROR("failed to create fbdev.\n");
-			return ret;
-		}
-
-		/*
-		 * fb_helper expects a value more than 1 if succeed
-		 * because register_framebuffer() should be called.
-		 */
-		ret = 1;
-	}
-
-	return ret;
-}
-
 static struct drm_fb_helper_funcs exynos_drm_fb_helper_funcs = {
-	.fb_probe =	exynos_drm_fbdev_probe,
+	.fb_probe =	exynos_drm_fbdev_create,
 };
 
 int exynos_drm_fbdev_init(struct drm_device *dev)
diff --git a/drivers/gpu/drm/gma500/framebuffer.c b/drivers/gpu/drm/gma500/framebuffer.c
index fee3bf8..2590cac 100644
--- a/drivers/gpu/drm/gma500/framebuffer.c
+++ b/drivers/gpu/drm/gma500/framebuffer.c
@@ -545,9 +545,7 @@ static int psbfb_probe(struct drm_fb_helper *helper,
 	struct psb_fbdev *psb_fbdev = (struct psb_fbdev *)helper;
 	struct drm_device *dev = psb_fbdev->psb_fb_helper.dev;
 	struct drm_psb_private *dev_priv = dev->dev_private;
-	int new_fb = 0;
 	int bytespp;
-	int ret;
 
 	bytespp = sizes->surface_bpp / 8;
 	if (bytespp == 3)	/* no 24bit packed */
@@ -562,13 +560,7 @@ static int psbfb_probe(struct drm_fb_helper *helper,
                 sizes->surface_depth = 16;
         }
 
-	if (!helper->fb) {
-		ret = psbfb_create(psb_fbdev, sizes);
-		if (ret)
-			return ret;
-		new_fb = 1;
-	}
-	return new_fb;
+	return psbfb_create(psb_fbdev, sizes);
 }
 
 static struct drm_fb_helper_funcs psb_fb_helper_funcs = {
diff --git a/drivers/gpu/drm/i915/intel_fb.c b/drivers/gpu/drm/i915/intel_fb.c
index 1c510da..981bdce 100644
--- a/drivers/gpu/drm/i915/intel_fb.c
+++ b/drivers/gpu/drm/i915/intel_fb.c
@@ -57,9 +57,10 @@ static struct fb_ops intelfb_ops = {
 	.fb_debug_leave = drm_fb_helper_debug_leave,
 };
 
-static int intelfb_create(struct intel_fbdev *ifbdev,
+static int intelfb_create(struct drm_fb_helper *helper,
 			  struct drm_fb_helper_surface_size *sizes)
 {
+	struct intel_fbdev *ifbdev = (struct intel_fbdev *)helper;
 	struct drm_device *dev = ifbdev->helper.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct fb_info *info;
@@ -181,26 +182,10 @@ out:
 	return ret;
 }
 
-static int intel_fb_find_or_create_single(struct drm_fb_helper *helper,
-					  struct drm_fb_helper_surface_size *sizes)
-{
-	struct intel_fbdev *ifbdev = (struct intel_fbdev *)helper;
-	int new_fb = 0;
-	int ret;
-
-	if (!helper->fb) {
-		ret = intelfb_create(ifbdev, sizes);
-		if (ret)
-			return ret;
-		new_fb = 1;
-	}
-	return new_fb;
-}
-
 static struct drm_fb_helper_funcs intel_fb_helper_funcs = {
 	.gamma_set = intel_crtc_fb_gamma_set,
 	.gamma_get = intel_crtc_fb_gamma_get,
-	.fb_probe = intel_fb_find_or_create_single,
+	.fb_probe = intelfb_create,
 };
 
 static void intel_fbdev_destroy(struct drm_device *dev,
diff --git a/drivers/gpu/drm/mgag200/mgag200_fb.c b/drivers/gpu/drm/mgag200/mgag200_fb.c
index 5bded5b..d2253f6 100644
--- a/drivers/gpu/drm/mgag200/mgag200_fb.c
+++ b/drivers/gpu/drm/mgag200/mgag200_fb.c
@@ -121,9 +121,10 @@ static int mgag200fb_create_object(struct mga_fbdev *afbdev,
 	return ret;
 }
 
-static int mgag200fb_create(struct mga_fbdev *mfbdev,
+static int mgag200fb_create(struct drm_fb_helper *helper,
 			   struct drm_fb_helper_surface_size *sizes)
 {
+	struct mga_fbdev *mfbdev = (struct mga_fbdev *)helper;
 	struct drm_device *dev = mfbdev->helper.dev;
 	struct drm_mode_fb_cmd2 mode_cmd;
 	struct mga_device *mdev = dev->dev_private;
@@ -210,23 +211,6 @@ out:
 	return ret;
 }
 
-static int mga_fb_find_or_create_single(struct drm_fb_helper *helper,
-					   struct drm_fb_helper_surface_size
-					   *sizes)
-{
-	struct mga_fbdev *mfbdev = (struct mga_fbdev *)helper;
-	int new_fb = 0;
-	int ret;
-
-	if (!helper->fb) {
-		ret = mgag200fb_create(mfbdev, sizes);
-		if (ret)
-			return ret;
-		new_fb = 1;
-	}
-	return new_fb;
-}
-
 static int mga_fbdev_destroy(struct drm_device *dev,
 				struct mga_fbdev *mfbdev)
 {
@@ -257,7 +241,7 @@ static int mga_fbdev_destroy(struct drm_device *dev,
 static struct drm_fb_helper_funcs mga_fb_helper_funcs = {
 	.gamma_set = mga_crtc_fb_gamma_set,
 	.gamma_get = mga_crtc_fb_gamma_get,
-	.fb_probe = mga_fb_find_or_create_single,
+	.fb_probe = mgag200fb_create,
 };
 
 int mgag200_fbdev_init(struct mga_device *mdev)
diff --git a/drivers/gpu/drm/nouveau/nouveau_fbcon.c b/drivers/gpu/drm/nouveau/nouveau_fbcon.c
index b1ebfe3..b035317 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fbcon.c
+++ b/drivers/gpu/drm/nouveau/nouveau_fbcon.c
@@ -251,9 +251,10 @@ nouveau_fbcon_zfill(struct drm_device *dev, struct nouveau_fbdev *fbcon)
 }
 
 static int
-nouveau_fbcon_create(struct nouveau_fbdev *fbcon,
+nouveau_fbcon_create(struct drm_fb_helper *helper,
 		     struct drm_fb_helper_surface_size *sizes)
 {
+	struct nouveau_fbdev *fbcon = (struct nouveau_fbdev *)helper;
 	struct drm_device *dev = fbcon->dev;
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	struct nouveau_device *device = nv_device(drm->device);
@@ -388,23 +389,6 @@ out:
 	return ret;
 }
 
-static int
-nouveau_fbcon_find_or_create_single(struct drm_fb_helper *helper,
-				    struct drm_fb_helper_surface_size *sizes)
-{
-	struct nouveau_fbdev *fbcon = (struct nouveau_fbdev *)helper;
-	int new_fb = 0;
-	int ret;
-
-	if (!helper->fb) {
-		ret = nouveau_fbcon_create(fbcon, sizes);
-		if (ret)
-			return ret;
-		new_fb = 1;
-	}
-	return new_fb;
-}
-
 void
 nouveau_fbcon_output_poll_changed(struct drm_device *dev)
 {
@@ -450,7 +434,7 @@ void nouveau_fbcon_gpu_lockup(struct fb_info *info)
 static struct drm_fb_helper_funcs nouveau_fbcon_helper_funcs = {
 	.gamma_set = nouveau_fbcon_gamma_set,
 	.gamma_get = nouveau_fbcon_gamma_get,
-	.fb_probe = nouveau_fbcon_find_or_create_single,
+	.fb_probe = nouveau_fbcon_create,
 };
 
 
diff --git a/drivers/gpu/drm/radeon/radeon_fb.c b/drivers/gpu/drm/radeon/radeon_fb.c
index b48d1c8..b174674 100644
--- a/drivers/gpu/drm/radeon/radeon_fb.c
+++ b/drivers/gpu/drm/radeon/radeon_fb.c
@@ -187,9 +187,10 @@ out_unref:
 	return ret;
 }
 
-static int radeonfb_create(struct radeon_fbdev *rfbdev,
+static int radeonfb_create(struct drm_fb_helper *helper,
 			   struct drm_fb_helper_surface_size *sizes)
 {
+	struct radeon_fbdev *rfbdev = (struct radeon_fbdev *)helper;
 	struct radeon_device *rdev = rfbdev->rdev;
 	struct fb_info *info;
 	struct drm_framebuffer *fb = NULL;
@@ -300,22 +301,6 @@ out_unref:
 	return ret;
 }
 
-static int radeon_fb_find_or_create_single(struct drm_fb_helper *helper,
-					   struct drm_fb_helper_surface_size *sizes)
-{
-	struct radeon_fbdev *rfbdev = (struct radeon_fbdev *)helper;
-	int new_fb = 0;
-	int ret;
-
-	if (!helper->fb) {
-		ret = radeonfb_create(rfbdev, sizes);
-		if (ret)
-			return ret;
-		new_fb = 1;
-	}
-	return new_fb;
-}
-
 void radeon_fb_output_poll_changed(struct radeon_device *rdev)
 {
 	drm_fb_helper_hotplug_event(&rdev->mode_info.rfbdev->helper);
@@ -349,7 +334,7 @@ static int radeon_fbdev_destroy(struct drm_device *dev, struct radeon_fbdev *rfb
 static struct drm_fb_helper_funcs radeon_fb_helper_funcs = {
 	.gamma_set = radeon_crtc_fb_gamma_set,
 	.gamma_get = radeon_crtc_fb_gamma_get,
-	.fb_probe = radeon_fb_find_or_create_single,
+	.fb_probe = radeonfb_create,
 };
 
 int radeon_fbdev_init(struct radeon_device *rdev)
diff --git a/drivers/gpu/drm/udl/udl_fb.c b/drivers/gpu/drm/udl/udl_fb.c
index cf5d05a..9f4be3d 100644
--- a/drivers/gpu/drm/udl/udl_fb.c
+++ b/drivers/gpu/drm/udl/udl_fb.c
@@ -476,9 +476,10 @@ udl_framebuffer_init(struct drm_device *dev,
 }
 
 
-static int udlfb_create(struct udl_fbdev *ufbdev,
+static int udlfb_create(struct drm_fb_helper *helper,
 			struct drm_fb_helper_surface_size *sizes)
 {
+	struct udl_fbdev *ufbdev = (struct udl_fbdev *)helper;
 	struct drm_device *dev = ufbdev->helper.dev;
 	struct fb_info *info;
 	struct device *device = &dev->usbdev->dev;
@@ -556,27 +557,10 @@ out:
 	return ret;
 }
 
-static int udl_fb_find_or_create_single(struct drm_fb_helper *helper,
-					struct drm_fb_helper_surface_size *sizes)
-{
-	struct udl_fbdev *ufbdev = (struct udl_fbdev *)helper;
-	int new_fb = 0;
-	int ret;
-
-	if (!helper->fb) {
-		ret = udlfb_create(ufbdev, sizes);
-		if (ret)
-			return ret;
-
-		new_fb = 1;
-	}
-	return new_fb;
-}
-
 static struct drm_fb_helper_funcs udl_fb_helper_funcs = {
 	.gamma_set = udl_crtc_fb_gamma_set,
 	.gamma_get = udl_crtc_fb_gamma_get,
-	.fb_probe = udl_fb_find_or_create_single,
+	.fb_probe = udlfb_create,
 };
 
 static void udl_fbdev_destroy(struct drm_device *dev,
diff --git a/drivers/staging/omapdrm/omap_fbdev.c b/drivers/staging/omapdrm/omap_fbdev.c
index 7e66eb1..caefdf9 100644
--- a/drivers/staging/omapdrm/omap_fbdev.c
+++ b/drivers/staging/omapdrm/omap_fbdev.c
@@ -296,25 +296,10 @@ static void omap_crtc_fb_gamma_get(struct drm_crtc *crtc,
 	DBG("fbdev: get gamma");
 }
 
-static int omap_fbdev_probe(struct drm_fb_helper *helper,
-		struct drm_fb_helper_surface_size *sizes)
-{
-	int new_fb = 0;
-	int ret;
-
-	if (!helper->fb) {
-		ret = omap_fbdev_create(helper, sizes);
-		if (ret)
-			return ret;
-		new_fb = 1;
-	}
-	return new_fb;
-}
-
 static struct drm_fb_helper_funcs omap_fb_helper_funcs = {
 	.gamma_set = omap_crtc_fb_gamma_set,
 	.gamma_get = omap_crtc_fb_gamma_get,
-	.fb_probe = omap_fbdev_probe,
+	.fb_probe = omap_fbdev_create,
 };
 
 static struct drm_fb_helper *get_fb(struct fb_info *fbi)
-- 
cgit v0.10.2


From 207fd32970b1def91b11ae28f6bebffc792db714 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sun, 20 Jan 2013 22:13:14 +0100
Subject: drm/fb-helper: improve kerneldoc

Now that the fbdev helper interface for drivers is trimmed down,
update the kerneldoc for all the remaining exported functions.

I've tried to beat the DocBook a bit by reordering the function
references a bit into a more sensible ordering. But that didn't work
out at all. Hence just extend the in-code DOC: section a bit.

Also remove the LOCKING: sections - especially for the setup functions
they're totally bogus. But that's not a documentation problem, but
simply an artifact of the current rather hazardous locking around drm
init and even more so around fbdev setup ...

v2: Some further improvements:
- Also add documentation for drm_fb_helper_single_add_all_connectors,
  Dave Airlie didn't want me to kill this one from the fb helper
  interface.
- Update docs for drm_fb_helper_fill_var/fix - they should be used
  from the driver's ->fb_probe callback to setup the fbdev info
  structure.
- Clarify what the ->fb_probe callback should all do - it needs to
  setup both the fbdev info and allocate the drm framebuffer used as
  backing storage.
- Add basic documentaation for the drm_fb_helper_funcs driver callback
  vfunc.

v3: Implement clarifications Laurent Pinchart suggested in his review.

v4: Fix another mispelling Laurent spotted.

Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl
index b26de52..51e1904 100644
--- a/Documentation/DocBook/drm.tmpl
+++ b/Documentation/DocBook/drm.tmpl
@@ -2143,6 +2143,7 @@ void intel_crt_init(struct drm_device *dev)
       <title>fbdev Helper Functions Reference</title>
 !Pdrivers/gpu/drm/drm_fb_helper.c fbdev helpers
 !Edrivers/gpu/drm/drm_fb_helper.c
+!Iinclude/drm/drm_fb_helper.h
     </sect2>
     <sect2>
       <title>Display Port Helper Functions Reference</title>
diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index e0d4b04..94f304d 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -52,9 +52,36 @@ static LIST_HEAD(kernel_fb_helper_list);
  * mode setting driver. They can be used mostly independantely from the crtc
  * helper functions used by many drivers to implement the kernel mode setting
  * interfaces.
+ *
+ * Initialization is done as a three-step process with drm_fb_helper_init(),
+ * drm_fb_helper_single_add_all_connectors() and drm_fb_helper_initial_config().
+ * Drivers with fancier requirements than the default beheviour can override the
+ * second step with their own code.  Teardown is done with drm_fb_helper_fini().
+ *
+ * At runtime drivers should restore the fbdev console by calling
+ * drm_fb_helper_restore_fbdev_mode() from their ->lastclose callback. They
+ * should also notify the fb helper code from updates to the output
+ * configuration by calling drm_fb_helper_hotplug_event(). For easier
+ * integration with the output polling code in drm_crtc_helper.c the modeset
+ * code proves a ->output_poll_changed callback.
+ *
+ * All other functions exported by the fb helper library can be used to
+ * implement the fbdev driver interface by the driver.
  */
 
-/* simple single crtc case helper function */
+/**
+ * drm_fb_helper_single_add_all_connectors() - add all connectors to fbdev
+ * 					       emulation helper
+ * @fb_helper: fbdev initialized with drm_fb_helper_init
+ *
+ * This functions adds all the available connectors for use with the given
+ * fb_helper. This is a separate step to allow drivers to freely assign
+ * connectors to the fbdev, e.g. if some are reserved for special purposes or
+ * not adequate to be used for the fbcon.
+ *
+ * Since this is part of the initial setup before the fbdev is published, no
+ * locking is required.
+ */
 int drm_fb_helper_single_add_all_connectors(struct drm_fb_helper *fb_helper)
 {
 	struct drm_device *dev = fb_helper->dev;
@@ -163,6 +190,10 @@ static void drm_fb_helper_restore_lut_atomic(struct drm_crtc *crtc)
 	crtc->funcs->gamma_set(crtc, r_base, g_base, b_base, 0, crtc->gamma_size);
 }
 
+/**
+ * drm_fb_helper_debug_enter - implementation for ->fb_debug_enter
+ * @info: fbdev registered by the helper
+ */
 int drm_fb_helper_debug_enter(struct fb_info *info)
 {
 	struct drm_fb_helper *helper = info->par;
@@ -208,6 +239,10 @@ static struct drm_framebuffer *drm_mode_config_fb(struct drm_crtc *crtc)
 	return NULL;
 }
 
+/**
+ * drm_fb_helper_debug_leave - implementation for ->fb_debug_leave
+ * @info: fbdev registered by the helper
+ */
 int drm_fb_helper_debug_leave(struct fb_info *info)
 {
 	struct drm_fb_helper *helper = info->par;
@@ -243,9 +278,9 @@ EXPORT_SYMBOL(drm_fb_helper_debug_leave);
  * drm_fb_helper_restore_fbdev_mode - restore fbdev configuration
  * @fb_helper: fbcon to restore
  *
- * This should be called from driver's drm->lastclose callback when implementing
- * an fbcon on top of kms using this helper. This ensures that the user isn't
- * greeted with a black screen when e.g. X dies.
+ * This should be called from driver's drm ->lastclose callback
+ * when implementing an fbcon on top of kms using this helper. This ensures that
+ * the user isn't greeted with a black screen when e.g. X dies.
  */
 bool drm_fb_helper_restore_fbdev_mode(struct drm_fb_helper *fb_helper)
 {
@@ -381,6 +416,11 @@ static void drm_fb_helper_dpms(struct fb_info *info, int dpms_mode)
 	drm_modeset_unlock_all(dev);
 }
 
+/**
+ * drm_fb_helper_blank - implementation for ->fb_blank
+ * @blank: desired blanking state
+ * @info: fbdev registered by the helper
+ */
 int drm_fb_helper_blank(int blank, struct fb_info *info)
 {
 	switch (blank) {
@@ -424,6 +464,24 @@ static void drm_fb_helper_crtc_free(struct drm_fb_helper *helper)
 	kfree(helper->crtc_info);
 }
 
+/**
+ * drm_fb_helper_init - initialize a drm_fb_helper structure
+ * @dev: drm device
+ * @fb_helper: driver-allocated fbdev helper structure to initialize
+ * @crtc_count: maximum number of crtcs to support in this fbdev emulation
+ * @max_conn_count: max connector count
+ *
+ * This allocates the structures for the fbdev helper with the given limits.
+ * Note that this won't yet touch the hardware (through the driver interfaces)
+ * nor register the fbdev. This is only done in drm_fb_helper_initial_config()
+ * to allow driver writes more control over the exact init sequence.
+ *
+ * Drivers must set fb_helper->funcs before calling
+ * drm_fb_helper_initial_config().
+ *
+ * RETURNS:
+ * Zero if everything went ok, nonzero otherwise.
+ */
 int drm_fb_helper_init(struct drm_device *dev,
 		       struct drm_fb_helper *fb_helper,
 		       int crtc_count, int max_conn_count)
@@ -552,6 +610,11 @@ static int setcolreg(struct drm_crtc *crtc, u16 red, u16 green,
 	return 0;
 }
 
+/**
+ * drm_fb_helper_setcmap - implementation for ->fb_setcmap
+ * @cmap: cmap to set
+ * @info: fbdev registered by the helper
+ */
 int drm_fb_helper_setcmap(struct fb_cmap *cmap, struct fb_info *info)
 {
 	struct drm_fb_helper *fb_helper = info->par;
@@ -591,6 +654,11 @@ int drm_fb_helper_setcmap(struct fb_cmap *cmap, struct fb_info *info)
 }
 EXPORT_SYMBOL(drm_fb_helper_setcmap);
 
+/**
+ * drm_fb_helper_check_var - implementation for ->fb_check_var
+ * @var: screeninfo to check
+ * @info: fbdev registered by the helper
+ */
 int drm_fb_helper_check_var(struct fb_var_screeninfo *var,
 			    struct fb_info *info)
 {
@@ -683,7 +751,14 @@ int drm_fb_helper_check_var(struct fb_var_screeninfo *var,
 }
 EXPORT_SYMBOL(drm_fb_helper_check_var);
 
-/* this will let fbcon do the mode init */
+/**
+ * drm_fb_helper_set_par - implementation for ->fb_set_par
+ * @info: fbdev registered by the helper
+ *
+ * This will let fbcon do the mode init and is called at initialization time by
+ * the fbdev core when registering the driver, and later on through the hotplug
+ * callback.
+ */
 int drm_fb_helper_set_par(struct fb_info *info)
 {
 	struct drm_fb_helper *fb_helper = info->par;
@@ -715,6 +790,11 @@ int drm_fb_helper_set_par(struct fb_info *info)
 }
 EXPORT_SYMBOL(drm_fb_helper_set_par);
 
+/**
+ * drm_fb_helper_pan_display - implementation for ->fb_pan_display
+ * @var: updated screen information
+ * @info: fbdev registered by the helper
+ */
 int drm_fb_helper_pan_display(struct fb_var_screeninfo *var,
 			      struct fb_info *info)
 {
@@ -753,8 +833,9 @@ int drm_fb_helper_pan_display(struct fb_var_screeninfo *var,
 EXPORT_SYMBOL(drm_fb_helper_pan_display);
 
 /*
- * Allocates the backing storage through the ->fb_probe callback and then
- * registers the fbdev and sets up the panic notifier.
+ * Allocates the backing storage and sets up the fbdev info structure through
+ * the ->fb_probe callback and then registers the fbdev and sets up the panic
+ * notifier.
  */
 static int drm_fb_helper_single_fb_probe(struct drm_fb_helper *fb_helper,
 					 int preferred_bpp)
@@ -876,6 +957,19 @@ static int drm_fb_helper_single_fb_probe(struct drm_fb_helper *fb_helper,
 	return 0;
 }
 
+/**
+ * drm_fb_helper_fill_fix - initializes fixed fbdev information
+ * @info: fbdev registered by the helper
+ * @pitch: desired pitch
+ * @depth: desired depth
+ *
+ * Helper to fill in the fixed fbdev information useful for a non-accelerated
+ * fbdev emulations. Drivers which support acceleration methods which impose
+ * additional constraints need to set up their own limits.
+ *
+ * Drivers should call this (or their equivalent setup code) from their
+ * ->fb_probe callback.
+ */
 void drm_fb_helper_fill_fix(struct fb_info *info, uint32_t pitch,
 			    uint32_t depth)
 {
@@ -896,6 +990,20 @@ void drm_fb_helper_fill_fix(struct fb_info *info, uint32_t pitch,
 }
 EXPORT_SYMBOL(drm_fb_helper_fill_fix);
 
+/**
+ * drm_fb_helper_fill_var - initalizes variable fbdev information
+ * @info: fbdev instance to set up
+ * @fb_helper: fb helper instance to use as template
+ * @fb_width: desired fb width
+ * @fb_height: desired fb height
+ *
+ * Sets up the variable fbdev metainformation from the given fb helper instance
+ * and the drm framebuffer allocated in fb_helper->fb.
+ *
+ * Drivers should call this (or their equivalent setup code) from their
+ * ->fb_probe callback after having allocated the fbdev backing
+ * storage framebuffer.
+ */
 void drm_fb_helper_fill_var(struct fb_info *info, struct drm_fb_helper *fb_helper,
 			    uint32_t fb_width, uint32_t fb_height)
 {
@@ -1358,18 +1466,23 @@ out:
 }
 
 /**
- * drm_helper_initial_config - setup a sane initial connector configuration
+ * drm_fb_helper_initial_config - setup a sane initial connector configuration
  * @fb_helper: fb_helper device struct
  * @bpp_sel: bpp value to use for the framebuffer configuration
  *
- * LOCKING:
- * Called at init time by the driver to set up the @fb_helper initial
- * configuration, must take the mode config lock.
- *
  * Scans the CRTCs and connectors and tries to put together an initial setup.
  * At the moment, this is a cloned configuration across all heads with
  * a new framebuffer object as the backing store.
  *
+ * Note that this also registers the fbdev and so allows userspace to call into
+ * the driver through the fbdev interfaces.
+ *
+ * This function will call down into the ->fb_probe callback to let
+ * the driver allocate and initialize the fbdev info structure and the drm
+ * framebuffer used to back the fbdev. drm_fb_helper_fill_var() and
+ * drm_fb_helper_fill_fix() are provided as helpers to setup simple default
+ * values for the fbdev info structure.
+ *
  * RETURNS:
  * Zero if everything went ok, nonzero otherwise.
  */
@@ -1400,12 +1513,17 @@ EXPORT_SYMBOL(drm_fb_helper_initial_config);
  *                               probing all the outputs attached to the fb
  * @fb_helper: the drm_fb_helper
  *
- * LOCKING:
- * Called at runtime, must take mode config lock.
- *
  * Scan the connectors attached to the fb_helper and try to put together a
  * setup after *notification of a change in output configuration.
  *
+ * Called at runtime, takes the mode config locks to be able to check/change the
+ * modeset configuration. Must be run from process context (which usually means
+ * either the output polling work or a work item launched from the driver's
+ * hotplug interrupt).
+ *
+ * Note that the driver must ensure that this is only called _after_ the fb has
+ * been fully set up, i.e. after the call to drm_fb_helper_initial_config.
+ *
  * RETURNS:
  * 0 on success and a non-zero error code otherwise.
  */
diff --git a/include/drm/drm_fb_helper.h b/include/drm/drm_fb_helper.h
index 973402d..31e5b97 100644
--- a/include/drm/drm_fb_helper.h
+++ b/include/drm/drm_fb_helper.h
@@ -48,6 +48,18 @@ struct drm_fb_helper_surface_size {
 	u32 surface_depth;
 };
 
+/**
+ * struct drm_fb_helper_funcs - driver callbacks for the fbdev emulation library
+ * @gamma_set: - Set the given gamma lut register on the given crtc.
+ * @gamma_get: - Read the given gamma lut register on the given crtc, used to
+ * 		 save the current lut when force-restoring the fbdev for e.g.
+ * 		 kdbg.
+ * @fb_probe: - Driver callback to allocate and initialize the fbdev info
+ * 		structure. Futhermore it also needs to allocate the drm
+ * 		framebuffer used to back the fbdev.
+ *
+ * Driver callbacks used by the fbdev emulation helper library.
+ */
 struct drm_fb_helper_funcs {
 	void (*gamma_set)(struct drm_crtc *crtc, u16 red, u16 green,
 			  u16 blue, int regno);
-- 
cgit v0.10.2


From 1b1d5397058f06bc5bd87d43ed93f34b28546ea4 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 24 Jan 2013 16:42:07 +0100
Subject: drm/fb-helper: don't sleep for screen unblank when an oopps is in
 progress

Otherwise the system will burn even brighter and worse, leave the user
wondering what's going on exactly.

Since we already have a panic handler which will (try) to restore the
entire fbdev console mode, we can just bail out. Inspired by a patch
from Konstantin Khlebnikov. The callchain leading to this, cut&pasted
from Konstantin's original patch:

callstack:
panic()
bust_spinlocks(1)
unblank_screen()
vc->vc_sw->con_blank()
fbcon_blank()
fb_blank()
info->fbops->fb_blank()
drm_fb_helper_blank()
drm_fb_helper_dpms()
drm_modeset_lock_all()
mutex_lock(&dev->mode_config.mutex)

Note that the entire locking in the fb helper around panic/sysrq and
kdbg is ... non-existant. So we have a decent change of blowing up
everything. But since reworking this ties in with funny concepts like
the fbdev notifier chain or the impressive things which happen around
console_lock while oopsing, I'll leave that as an exercise for braver
souls than me.

v2: Drop the -EBUSY return value I've copied, we don't need it since
the we'll take care of things ourselves anyway.

Cc: Konstantin Khlebnikov <khlebnikov@openvz.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
References: https://patchwork.kernel.org/patch/1878181/
Reviewed-by: Rob Clark <robdclark@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index 94f304d..59d6b9b 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -391,6 +391,14 @@ static void drm_fb_helper_dpms(struct fb_info *info, int dpms_mode)
 	int i, j;
 
 	/*
+	 * fbdev->blank can be called from irq context in case of a panic.
+	 * Since we already have our own special panic handler which will
+	 * restore the fbdev console mode completely, just bail out early.
+	 */
+	if (oops_in_progress)
+		return;
+
+	/*
 	 * For each CRTC in this fb, turn the connectors on/off.
 	 */
 	drm_modeset_lock_all(dev);
-- 
cgit v0.10.2


From a065b46a01b25d7d364e01e75f7ec2bd9ed5d9cb Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Tue, 12 Feb 2013 00:17:09 +0100
Subject: drm/fb-helper: remove unused members of struct drm_fb_helper

Spotted by Rob Clark.

Reviewed-by: Rob Clark <robdclark@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/include/drm/drm_fb_helper.h b/include/drm/drm_fb_helper.h
index 31e5b97..c095116 100644
--- a/include/drm/drm_fb_helper.h
+++ b/include/drm/drm_fb_helper.h
@@ -77,9 +77,7 @@ struct drm_fb_helper_connector {
 
 struct drm_fb_helper {
 	struct drm_framebuffer *fb;
-	struct drm_framebuffer *saved_fb;
 	struct drm_device *dev;
-	struct drm_display_mode *mode;
 	int crtc_count;
 	struct drm_fb_helper_crtc *crtc_info;
 	int connector_count;
-- 
cgit v0.10.2


From 7083e05072b88d503d257b6f012ce56367f3ac97 Mon Sep 17 00:00:00 2001
From: Ben Widawsky <ben@bwidawsk.net>
Date: Fri, 1 Feb 2013 16:41:14 -0800
Subject: drm/i915: Fix RC6VIDS encode/decode

The RC6 VIDS has a linear ramp starting at 250mv, which means any values
below 250 are invalid. The old buggy macros tried to adjust for this to
be more flexible, but there is no need. As Dan pointed out the ENCODE
only ever has one value. The only invalid value for decode is an input
of 0 which means something is really wonky, and the cases where DECODE
are used either don't matter (debug values), or would be implicitly
correct (the check for less than 450).

This patch makes simpler, easier to read macros which are actually
correct. Maybe this patch can actually fix some bugs now.

Thanks to Dan for catching this. /me hides

Cc: stable@kernel.org
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 15f5e7f..8754f91 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -4272,8 +4272,8 @@
 #define   GEN6_PCODE_READ_MIN_FREQ_TABLE	0x9
 #define	  GEN6_PCODE_WRITE_RC6VIDS		0x4
 #define	  GEN6_PCODE_READ_RC6VIDS		0x5
-#define   GEN6_ENCODE_RC6_VID(mv)		(((mv) / 5) - 245) < 0 ?: 0
-#define   GEN6_DECODE_RC6_VID(vids)		(((vids) * 5) > 0 ? ((vids) * 5) + 245 : 0)
+#define   GEN6_ENCODE_RC6_VID(mv)		(((mv) - 245) / 5)
+#define   GEN6_DECODE_RC6_VID(vids)		(((vids) * 5) + 245)
 #define GEN6_PCODE_DATA				0x138128
 #define   GEN6_PCODE_FREQ_IA_RATIO_SHIFT	8
 
-- 
cgit v0.10.2


From b8efb17b3d687695b81485f606fc4e6c35a50f9a Mon Sep 17 00:00:00 2001
From: Zhang Rui <rui.zhang@intel.com>
Date: Tue, 5 Feb 2013 15:41:53 +0800
Subject: i915: ignore lid open event when resuming

i915 driver needs to do modeset when
1. system resumes from sleep
2. lid is opened

In PM_SUSPEND_MEM state, all the GPEs are cleared when system resumes,
thus it is the i915_resume code does the modeset rather than intel_lid_notify().

But in PM_SUSPEND_FREEZE state, this will be broken because
system is still responsive to the lid events.
1. When we close the lid in Freeze state, intel_lid_notify() sets modeset_on_lid.
2. When we reopen the lid, intel_lid_notify() will do a modeset,
   before the system is resumed.
here is the error log,

[92146.548074] WARNING: at drivers/gpu/drm/i915/intel_display.c:1028 intel_wait_for_pipe_off+0x184/0x190 [i915]()
[92146.548076] Hardware name: VGN-Z540N
[92146.548078] pipe_off wait timed out
[92146.548167] Modules linked in: hid_generic usbhid hid snd_hda_codec_realtek snd_hda_intel snd_hda_codec parport_pc snd_hwdep ppdev snd_pcm_oss i915 snd_mixer_oss snd_pcm arc4 iwldvm snd_seq_dummy mac80211 snd_seq_oss snd_seq_midi fbcon tileblit font bitblit softcursor drm_kms_helper snd_rawmidi snd_seq_midi_event coretemp drm snd_seq kvm btusb bluetooth snd_timer iwlwifi pcmcia tpm_infineon i2c_algo_bit joydev snd_seq_device intel_agp cfg80211 snd intel_gtt yenta_socket pcmcia_rsrc sony_laptop agpgart microcode psmouse tpm_tis serio_raw mxm_wmi soundcore snd_page_alloc tpm acpi_cpufreq lpc_ich pcmcia_core tpm_bios mperf processor lp parport firewire_ohci firewire_core crc_itu_t sdhci_pci sdhci thermal e1000e
[92146.548173] Pid: 4304, comm: kworker/0:0 Tainted: G        W    3.8.0-rc3-s0i3-v3-test+ #9
[92146.548175] Call Trace:
[92146.548189]  [<c10378e2>] warn_slowpath_common+0x72/0xa0
[92146.548227]  [<f86398b4>] ? intel_wait_for_pipe_off+0x184/0x190 [i915]
[92146.548263]  [<f86398b4>] ? intel_wait_for_pipe_off+0x184/0x190 [i915]
[92146.548270]  [<c10379b3>] warn_slowpath_fmt+0x33/0x40
[92146.548307]  [<f86398b4>] intel_wait_for_pipe_off+0x184/0x190 [i915]
[92146.548344]  [<f86399c2>] intel_disable_pipe+0x102/0x190 [i915]
[92146.548380]  [<f8639ea4>] ? intel_disable_plane+0x64/0x80 [i915]
[92146.548417]  [<f8639f7c>] i9xx_crtc_disable+0xbc/0x150 [i915]
[92146.548456]  [<f863ebee>] intel_crtc_update_dpms+0x5e/0x90 [i915]
[92146.548493]  [<f86437cf>] intel_modeset_setup_hw_state+0x42f/0x8f0 [i915]
[92146.548535]  [<f8645b0b>] intel_lid_notify+0x9b/0xc0 [i915]
[92146.548543]  [<c15610d3>] notifier_call_chain+0x43/0x60
[92146.548550]  [<c105d1e1>] __blocking_notifier_call_chain+0x41/0x80
[92146.548556]  [<c105d23f>] blocking_notifier_call_chain+0x1f/0x30
[92146.548563]  [<c131a684>] acpi_lid_send_state+0x78/0xa4
[92146.548569]  [<c131aa9e>] acpi_button_notify+0x3b/0xf1
[92146.548577]  [<c12df56a>] ? acpi_os_execute+0x17/0x19
[92146.548582]  [<c12e591a>] ? acpi_ec_sync_query+0xa5/0xbc
[92146.548589]  [<c12e2b82>] acpi_device_notify+0x16/0x18
[92146.548595]  [<c12f4904>] acpi_ev_notify_dispatch+0x38/0x4f
[92146.548600]  [<c12df0e8>] acpi_os_execute_deferred+0x20/0x2b
[92146.548607]  [<c1051208>] process_one_work+0x128/0x3f0
[92146.548613]  [<c1564f73>] ? common_interrupt+0x33/0x38
[92146.548618]  [<c104f8c0>] ? wake_up_worker+0x30/0x30
[92146.548624]  [<c12df0c8>] ? acpi_os_wait_events_complete+0x1e/0x1e
[92146.548629]  [<c10524f9>] worker_thread+0x119/0x3b0
[92146.548634]  [<c10523e0>] ? manage_workers+0x240/0x240
[92146.548640]  [<c1056e84>] kthread+0x94/0xa0
[92146.548647]  [<c1060000>] ? ftrace_raw_output_sched_stat_runtime+0x70/0xf0
[92146.548652]  [<c15649b7>] ret_from_kernel_thread+0x1b/0x28
[92146.548658]  [<c1056df0>] ? kthread_create_on_node+0xc0/0xc0

three different modeset flags are introduced in this patch
MODESET_ON_LID_OPEN: do modeset on next lid open event
MODESET_DONE:  modeset already done
MODESET_SUSPENDED:  suspended, only do modeset when system is resumed

In this way,
1. when lid is closed, MODESET_ON_LID_OPEN is set so that
   we'll do modeset on next lid open event.
2. when lid is opened, MODESET_DONE is set
   so that duplicate lid open events will be ignored.
3. when system suspends, MODESET_SUSPENDED is set.
   In this case, we will not do modeset on any lid events.

Plus, locking mechanism is also introduced to avoid racing.

Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index cf06103..4fa6beb 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -1610,6 +1610,7 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 	mutex_init(&dev_priv->dpio_lock);
 
 	mutex_init(&dev_priv->rps.hw_lock);
+	mutex_init(&dev_priv->modeset_restore_lock);
 
 	if (IS_IVYBRIDGE(dev) || IS_HASWELL(dev))
 		dev_priv->num_pipe = 3;
diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index d159d7a..c5b8c81 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -470,6 +470,11 @@ static int i915_drm_freeze(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
+	/* ignore lid events during suspend */
+	mutex_lock(&dev_priv->modeset_restore_lock);
+	dev_priv->modeset_restore = MODESET_SUSPENDED;
+	mutex_unlock(&dev_priv->modeset_restore_lock);
+
 	intel_set_power_well(dev, true);
 
 	drm_kms_helper_poll_disable(dev);
@@ -496,9 +501,6 @@ static int i915_drm_freeze(struct drm_device *dev)
 
 	intel_opregion_fini(dev);
 
-	/* Modeset on resume, not lid events */
-	dev_priv->modeset_on_lid = 0;
-
 	console_lock();
 	intel_fbdev_set_suspend(dev, 1);
 	console_unlock();
@@ -574,8 +576,6 @@ static int __i915_drm_thaw(struct drm_device *dev)
 
 	intel_opregion_init(dev);
 
-	dev_priv->modeset_on_lid = 0;
-
 	/*
 	 * The console lock can be pretty contented on resume due
 	 * to all the printk activity.  Try to keep it out of the hot
@@ -588,6 +588,9 @@ static int __i915_drm_thaw(struct drm_device *dev)
 		schedule_work(&dev_priv->console_resume_work);
 	}
 
+	mutex_lock(&dev_priv->modeset_restore_lock);
+	dev_priv->modeset_restore = MODESET_DONE;
+	mutex_unlock(&dev_priv->modeset_restore_lock);
 	return error;
 }
 
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index c338b44..4e5a377 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -846,6 +846,12 @@ struct i915_gpu_error {
 	unsigned int stop_rings;
 };
 
+enum modeset_restore {
+	MODESET_ON_LID_OPEN,
+	MODESET_DONE,
+	MODESET_SUSPENDED,
+};
+
 typedef struct drm_i915_private {
 	struct drm_device *dev;
 	struct kmem_cache *slab;
@@ -967,8 +973,8 @@ typedef struct drm_i915_private {
 
 	unsigned long quirks;
 
-	/* Register state */
-	bool modeset_on_lid;
+	enum modeset_restore modeset_restore;
+	struct mutex modeset_restore_lock;
 
 	struct i915_gtt gtt;
 
diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c
index 5e3f08e..bff78a3 100644
--- a/drivers/gpu/drm/i915/intel_lvds.c
+++ b/drivers/gpu/drm/i915/intel_lvds.c
@@ -547,13 +547,14 @@ static const struct dmi_system_id intel_no_modeset_on_lid[] = {
 };
 
 /*
- * Lid events. Note the use of 'modeset_on_lid':
- *  - we set it on lid close, and reset it on open
+ * Lid events. Note the use of 'modeset':
+ *  - we set it to MODESET_ON_LID_OPEN on lid close,
+ *    and set it to MODESET_DONE on open
  *  - we use it as a "only once" bit (ie we ignore
- *    duplicate events where it was already properly
- *    set/reset)
- *  - the suspend/resume paths will also set it to
- *    zero, since they restore the mode ("lid open").
+ *    duplicate events where it was already properly set)
+ *  - the suspend/resume paths will set it to
+ *    MODESET_SUSPENDED and ignore the lid open event,
+ *    because they restore the mode ("lid open").
  */
 static int intel_lid_notify(struct notifier_block *nb, unsigned long val,
 			    void *unused)
@@ -567,6 +568,9 @@ static int intel_lid_notify(struct notifier_block *nb, unsigned long val,
 	if (dev->switch_power_state != DRM_SWITCH_POWER_ON)
 		return NOTIFY_OK;
 
+	mutex_lock(&dev_priv->modeset_restore_lock);
+	if (dev_priv->modeset_restore == MODESET_SUSPENDED)
+		goto exit;
 	/*
 	 * check and update the status of LVDS connector after receiving
 	 * the LID nofication event.
@@ -575,21 +579,24 @@ static int intel_lid_notify(struct notifier_block *nb, unsigned long val,
 
 	/* Don't force modeset on machines where it causes a GPU lockup */
 	if (dmi_check_system(intel_no_modeset_on_lid))
-		return NOTIFY_OK;
+		goto exit;
 	if (!acpi_lid_open()) {
-		dev_priv->modeset_on_lid = 1;
-		return NOTIFY_OK;
+		/* do modeset on next lid open event */
+		dev_priv->modeset_restore = MODESET_ON_LID_OPEN;
+		goto exit;
 	}
 
-	if (!dev_priv->modeset_on_lid)
-		return NOTIFY_OK;
-
-	dev_priv->modeset_on_lid = 0;
+	if (dev_priv->modeset_restore == MODESET_DONE)
+		goto exit;
 
 	drm_modeset_lock_all(dev);
 	intel_modeset_setup_hw_state(dev, true);
 	drm_modeset_unlock_all(dev);
 
+	dev_priv->modeset_restore = MODESET_DONE;
+
+exit:
+	mutex_unlock(&dev_priv->modeset_restore_lock);
 	return NOTIFY_OK;
 }
 
-- 
cgit v0.10.2


From 93711d8becca550154da3b9e422be77083ab51ad Mon Sep 17 00:00:00 2001
From: Bjorn Helgaas <bhelgaas@google.com>
Date: Fri, 8 Feb 2013 15:27:01 -0700
Subject: drm/pci: define drm_pcie_get_speed_cap_mask() only when CONFIG_PCI=y

Move drm_pcie_get_speed_cap_mask() under #ifdef CONFIG_PCI because it
it used only for PCI devices (evergreen, r600, r770), and it uses
PCI interfaces that only exist when CONFIG_PCI=y.

Previously, we tried to compile drm_pcie_get_speed_cap_mask() even when
CONFIG_PCI=n, which fails.

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/gpu/drm/drm_pci.c b/drivers/gpu/drm/drm_pci.c
index 86102a0..bd719e9 100644
--- a/drivers/gpu/drm/drm_pci.c
+++ b/drivers/gpu/drm/drm_pci.c
@@ -439,33 +439,6 @@ int drm_pci_init(struct drm_driver *driver, struct pci_driver *pdriver)
 	return 0;
 }
 
-#else
-
-int drm_pci_init(struct drm_driver *driver, struct pci_driver *pdriver)
-{
-	return -1;
-}
-
-#endif
-
-EXPORT_SYMBOL(drm_pci_init);
-
-/*@}*/
-void drm_pci_exit(struct drm_driver *driver, struct pci_driver *pdriver)
-{
-	struct drm_device *dev, *tmp;
-	DRM_DEBUG("\n");
-
-	if (driver->driver_features & DRIVER_MODESET) {
-		pci_unregister_driver(pdriver);
-	} else {
-		list_for_each_entry_safe(dev, tmp, &driver->device_list, driver_item)
-			drm_put_dev(dev);
-	}
-	DRM_INFO("Module unloaded\n");
-}
-EXPORT_SYMBOL(drm_pci_exit);
-
 int drm_pcie_get_speed_cap_mask(struct drm_device *dev, u32 *mask)
 {
 	struct pci_dev *root;
@@ -503,3 +476,30 @@ int drm_pcie_get_speed_cap_mask(struct drm_device *dev, u32 *mask)
 	return 0;
 }
 EXPORT_SYMBOL(drm_pcie_get_speed_cap_mask);
+
+#else
+
+int drm_pci_init(struct drm_driver *driver, struct pci_driver *pdriver)
+{
+	return -1;
+}
+
+#endif
+
+EXPORT_SYMBOL(drm_pci_init);
+
+/*@}*/
+void drm_pci_exit(struct drm_driver *driver, struct pci_driver *pdriver)
+{
+	struct drm_device *dev, *tmp;
+	DRM_DEBUG("\n");
+
+	if (driver->driver_features & DRIVER_MODESET) {
+		pci_unregister_driver(pdriver);
+	} else {
+		list_for_each_entry_safe(dev, tmp, &driver->device_list, driver_item)
+			drm_put_dev(dev);
+	}
+	DRM_INFO("Module unloaded\n");
+}
+EXPORT_SYMBOL(drm_pci_exit);
-- 
cgit v0.10.2


From 3e2b756ba330343c960c332695608b9c5881a173 Mon Sep 17 00:00:00 2001
From: Chris Metcalf <cmetcalf@tilera.com>
Date: Fri, 1 Feb 2013 13:44:33 -0500
Subject: drm: fix compile failure by including <linux/swiotlb.h>

On tile architecture (with "make allyesconfig") including
<linux/swiotlb.h> is required to call swiotlb_nr_tbl().

Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>
Acked-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 64d6e304..2f27414 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -28,6 +28,7 @@
  */
 
 #include <core/engine.h>
+#include <linux/swiotlb.h>
 
 #include <subdev/fb.h>
 #include <subdev/vm.h>
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index 1d8ff2f..93f760e 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -38,6 +38,7 @@
 #include <drm/radeon_drm.h>
 #include <linux/seq_file.h>
 #include <linux/slab.h>
+#include <linux/swiotlb.h>
 #include "radeon_reg.h"
 #include "radeon.h"
 
-- 
cgit v0.10.2


From f934ec8c34b9dcefb5a4f35b0bda33bca289cbe6 Mon Sep 17 00:00:00 2001
From: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Date: Tue, 29 Jan 2013 14:27:39 +0100
Subject: drm: shut up invalid edid messages

My cheapo monitor has an invalid block 1, resulting in a lot of dmesg spam every few seconds.

I get it the first time that the entire block is all 0xff..

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Cc: stable@vger.kernel.org [v3.7]
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
index 5132425..e1aca7b 100644
--- a/drivers/gpu/drm/drm_edid.c
+++ b/drivers/gpu/drm/drm_edid.c
@@ -354,10 +354,14 @@ drm_do_get_edid(struct drm_connector *connector, struct i2c_adapter *adapter)
 				break;
 			}
 		}
-		if (i == 4)
+
+		if (i == 4 && print_bad_edid) {
 			dev_warn(connector->dev->dev,
 			 "%s: Ignoring invalid EDID block %d.\n",
 			 drm_get_connector_name(connector), j);
+
+			connector->bad_edid_counter++;
+		}
 	}
 
 	if (valid_extensions != block[0x7e]) {
-- 
cgit v0.10.2


From cf0a6584aa6d382f802f2c3cacac23ccbccde0cd Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Wed, 6 Feb 2013 11:24:41 +0100
Subject: drm/i915: write backlight harder

770c12312ad617172b1a65b911d3e6564fc5aca8 is the first bad commit
commit 770c12312ad617172b1a65b911d3e6564fc5aca8
Author: Takashi Iwai <tiwai@suse.de>
Date:   Sat Aug 11 08:56:42 2012 +0200

    drm/i915: Fix blank panel at reopening lid

changed the register write sequence for restoring the backlight, which
helped prevent non-working backlights on some machines. Turns out that
the original sequence was the right thing to do for a different set of
machines. Worse, setting the backlight level _after_ enabling it seems
to reset it somehow. So we need to make that one conditional upon the
backlight having been reset to zero, and add the old one back.

Cargo-culting at it's best, but it seems to work.

Cc: stable@vger.kernel.org
Cc: Takashi Iwai <tiwai@suse.de>
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=47941
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Acked-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_panel.c b/drivers/gpu/drm/i915/intel_panel.c
index bee8cb6..a3730e0 100644
--- a/drivers/gpu/drm/i915/intel_panel.c
+++ b/drivers/gpu/drm/i915/intel_panel.c
@@ -321,6 +321,9 @@ void intel_panel_enable_backlight(struct drm_device *dev,
 	if (dev_priv->backlight_level == 0)
 		dev_priv->backlight_level = intel_panel_get_max_backlight(dev);
 
+	dev_priv->backlight_enabled = true;
+	intel_panel_actually_set_backlight(dev, dev_priv->backlight_level);
+
 	if (INTEL_INFO(dev)->gen >= 4) {
 		uint32_t reg, tmp;
 
@@ -356,12 +359,12 @@ void intel_panel_enable_backlight(struct drm_device *dev,
 	}
 
 set_level:
-	/* Call below after setting BLC_PWM_CPU_CTL2 and BLC_PWM_PCH_CTL1.
-	 * BLC_PWM_CPU_CTL may be cleared to zero automatically when these
-	 * registers are set.
+	/* Check the current backlight level and try to set again if it's zero.
+	 * On some machines, BLC_PWM_CPU_CTL is cleared to zero automatically
+	 * when BLC_PWM_CPU_CTL2 and BLC_PWM_PCH_CTL1 are written.
 	 */
-	dev_priv->backlight_enabled = true;
-	intel_panel_actually_set_backlight(dev, dev_priv->backlight_level);
+	if (!intel_panel_get_backlight(dev))
+		intel_panel_actually_set_backlight(dev, dev_priv->backlight_level);
 }
 
 static void intel_panel_init_backlight(struct drm_device *dev)
-- 
cgit v0.10.2


From 07ea0d85ac8adb87b817913d9720e3c76171b1f6 Mon Sep 17 00:00:00 2001
From: Ben Widawsky <ben@bwidawsk.net>
Date: Thu, 7 Feb 2013 13:34:19 -0800
Subject: drm/i915: Clarify HW context size logic

This was a rebase error from when the patches originally landed. Since
the context size is unsigned, there is also no use in checking if it's
less than 0.

The existing code is not really wrong, but it's not simple as it should
be.

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c
index a3f06bc..e7ecafa 100644
--- a/drivers/gpu/drm/i915/i915_gem_context.c
+++ b/drivers/gpu/drm/i915/i915_gem_context.c
@@ -242,7 +242,6 @@ err_destroy:
 void i915_gem_context_init(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	uint32_t ctx_size;
 
 	if (!HAS_HW_CONTEXTS(dev)) {
 		dev_priv->hw_contexts_disabled = true;
@@ -254,11 +253,9 @@ void i915_gem_context_init(struct drm_device *dev)
 	    dev_priv->ring[RCS].default_context)
 		return;
 
-	ctx_size = get_context_size(dev);
-	dev_priv->hw_context_size = get_context_size(dev);
-	dev_priv->hw_context_size = round_up(dev_priv->hw_context_size, 4096);
+	dev_priv->hw_context_size = round_up(get_context_size(dev), 4096);
 
-	if (ctx_size <= 0 || ctx_size > (1<<20)) {
+	if (dev_priv->hw_context_size > (1<<20)) {
 		dev_priv->hw_contexts_disabled = true;
 		return;
 	}
-- 
cgit v0.10.2


From 26739f12cf210cb8df35969258a1f064e8e12b63 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 7 Feb 2013 12:42:32 +0100
Subject: drm/i915: unify HDMI/DP hpd definitions

They're physically the same pins and also the same bits, duplicating
only confuses the reader. This also makes it a bit obvious that we
have quite some code duplication going on here. Squashing that is for
a larger rework in our hpd handling though.

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 13bb8d3..b27ffdb 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -2137,12 +2137,12 @@ static void valleyview_hpd_irq_setup(struct drm_device *dev)
 	u32 hotplug_en = I915_READ(PORT_HOTPLUG_EN);
 
 	/* Note HDMI and DP share bits */
-	if (dev_priv->hotplug_supported_mask & HDMIB_HOTPLUG_INT_STATUS)
-		hotplug_en |= HDMIB_HOTPLUG_INT_EN;
-	if (dev_priv->hotplug_supported_mask & HDMIC_HOTPLUG_INT_STATUS)
-		hotplug_en |= HDMIC_HOTPLUG_INT_EN;
-	if (dev_priv->hotplug_supported_mask & HDMID_HOTPLUG_INT_STATUS)
-		hotplug_en |= HDMID_HOTPLUG_INT_EN;
+	if (dev_priv->hotplug_supported_mask & PORTB_HOTPLUG_INT_STATUS)
+		hotplug_en |= PORTB_HOTPLUG_INT_EN;
+	if (dev_priv->hotplug_supported_mask & PORTC_HOTPLUG_INT_STATUS)
+		hotplug_en |= PORTC_HOTPLUG_INT_EN;
+	if (dev_priv->hotplug_supported_mask & PORTD_HOTPLUG_INT_STATUS)
+		hotplug_en |= PORTD_HOTPLUG_INT_EN;
 	if (dev_priv->hotplug_supported_mask & SDVOC_HOTPLUG_INT_STATUS_I915)
 		hotplug_en |= SDVOC_HOTPLUG_INT_EN;
 	if (dev_priv->hotplug_supported_mask & SDVOB_HOTPLUG_INT_STATUS_I915)
@@ -2408,12 +2408,12 @@ static void i915_hpd_irq_setup(struct drm_device *dev)
 	if (I915_HAS_HOTPLUG(dev)) {
 		hotplug_en = I915_READ(PORT_HOTPLUG_EN);
 
-		if (dev_priv->hotplug_supported_mask & HDMIB_HOTPLUG_INT_STATUS)
-			hotplug_en |= HDMIB_HOTPLUG_INT_EN;
-		if (dev_priv->hotplug_supported_mask & HDMIC_HOTPLUG_INT_STATUS)
-			hotplug_en |= HDMIC_HOTPLUG_INT_EN;
-		if (dev_priv->hotplug_supported_mask & HDMID_HOTPLUG_INT_STATUS)
-			hotplug_en |= HDMID_HOTPLUG_INT_EN;
+		if (dev_priv->hotplug_supported_mask & PORTB_HOTPLUG_INT_STATUS)
+			hotplug_en |= PORTB_HOTPLUG_INT_EN;
+		if (dev_priv->hotplug_supported_mask & PORTC_HOTPLUG_INT_STATUS)
+			hotplug_en |= PORTC_HOTPLUG_INT_EN;
+		if (dev_priv->hotplug_supported_mask & PORTD_HOTPLUG_INT_STATUS)
+			hotplug_en |= PORTD_HOTPLUG_INT_EN;
 		if (dev_priv->hotplug_supported_mask & SDVOC_HOTPLUG_INT_STATUS_I915)
 			hotplug_en |= SDVOC_HOTPLUG_INT_EN;
 		if (dev_priv->hotplug_supported_mask & SDVOB_HOTPLUG_INT_STATUS_I915)
@@ -2642,12 +2642,12 @@ static void i965_hpd_irq_setup(struct drm_device *dev)
 
 	/* Note HDMI and DP share hotplug bits */
 	hotplug_en = 0;
-	if (dev_priv->hotplug_supported_mask & HDMIB_HOTPLUG_INT_STATUS)
-		hotplug_en |= HDMIB_HOTPLUG_INT_EN;
-	if (dev_priv->hotplug_supported_mask & HDMIC_HOTPLUG_INT_STATUS)
-		hotplug_en |= HDMIC_HOTPLUG_INT_EN;
-	if (dev_priv->hotplug_supported_mask & HDMID_HOTPLUG_INT_STATUS)
-		hotplug_en |= HDMID_HOTPLUG_INT_EN;
+	if (dev_priv->hotplug_supported_mask & PORTB_HOTPLUG_INT_STATUS)
+		hotplug_en |= PORTB_HOTPLUG_INT_EN;
+	if (dev_priv->hotplug_supported_mask & PORTC_HOTPLUG_INT_STATUS)
+		hotplug_en |= PORTC_HOTPLUG_INT_EN;
+	if (dev_priv->hotplug_supported_mask & PORTD_HOTPLUG_INT_STATUS)
+		hotplug_en |= PORTD_HOTPLUG_INT_EN;
 	if (IS_G4X(dev)) {
 		if (dev_priv->hotplug_supported_mask & SDVOC_HOTPLUG_INT_STATUS_G4X)
 			hotplug_en |= SDVOC_HOTPLUG_INT_EN;
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 8754f91..b9fdc99 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -1625,12 +1625,9 @@
 
 /* Hotplug control (945+ only) */
 #define PORT_HOTPLUG_EN		(dev_priv->info->display_mmio_offset + 0x61110)
-#define   HDMIB_HOTPLUG_INT_EN			(1 << 29)
-#define   DPB_HOTPLUG_INT_EN			(1 << 29)
-#define   HDMIC_HOTPLUG_INT_EN			(1 << 28)
-#define   DPC_HOTPLUG_INT_EN			(1 << 28)
-#define   HDMID_HOTPLUG_INT_EN			(1 << 27)
-#define   DPD_HOTPLUG_INT_EN			(1 << 27)
+#define   PORTB_HOTPLUG_INT_EN			(1 << 29)
+#define   PORTC_HOTPLUG_INT_EN			(1 << 28)
+#define   PORTD_HOTPLUG_INT_EN			(1 << 27)
 #define   SDVOB_HOTPLUG_INT_EN			(1 << 26)
 #define   SDVOC_HOTPLUG_INT_EN			(1 << 25)
 #define   TV_HOTPLUG_INT_EN			(1 << 18)
@@ -1653,19 +1650,12 @@
 
 #define PORT_HOTPLUG_STAT	(dev_priv->info->display_mmio_offset + 0x61114)
 /* HDMI/DP bits are gen4+ */
-#define   DPB_HOTPLUG_LIVE_STATUS               (1 << 29)
-#define   DPC_HOTPLUG_LIVE_STATUS               (1 << 28)
-#define   DPD_HOTPLUG_LIVE_STATUS               (1 << 27)
-#define   DPD_HOTPLUG_INT_STATUS		(3 << 21)
-#define   DPC_HOTPLUG_INT_STATUS		(3 << 19)
-#define   DPB_HOTPLUG_INT_STATUS		(3 << 17)
-/* HDMI bits are shared with the DP bits */
-#define   HDMIB_HOTPLUG_LIVE_STATUS             (1 << 29)
-#define   HDMIC_HOTPLUG_LIVE_STATUS             (1 << 28)
-#define   HDMID_HOTPLUG_LIVE_STATUS             (1 << 27)
-#define   HDMID_HOTPLUG_INT_STATUS		(3 << 21)
-#define   HDMIC_HOTPLUG_INT_STATUS		(3 << 19)
-#define   HDMIB_HOTPLUG_INT_STATUS		(3 << 17)
+#define   PORTB_HOTPLUG_LIVE_STATUS               (1 << 29)
+#define   PORTC_HOTPLUG_LIVE_STATUS               (1 << 28)
+#define   PORTD_HOTPLUG_LIVE_STATUS               (1 << 27)
+#define   PORTD_HOTPLUG_INT_STATUS		(3 << 21)
+#define   PORTC_HOTPLUG_INT_STATUS		(3 << 19)
+#define   PORTB_HOTPLUG_INT_STATUS		(3 << 17)
 /* CRT/TV common between gen3+ */
 #define   CRT_HOTPLUG_INT_STATUS		(1 << 11)
 #define   TV_HOTPLUG_INT_STATUS			(1 << 10)
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 56e408b..7b8bfe8 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -2302,13 +2302,13 @@ g4x_dp_detect(struct intel_dp *intel_dp)
 
 	switch (intel_dig_port->port) {
 	case PORT_B:
-		bit = DPB_HOTPLUG_LIVE_STATUS;
+		bit = PORTB_HOTPLUG_LIVE_STATUS;
 		break;
 	case PORT_C:
-		bit = DPC_HOTPLUG_LIVE_STATUS;
+		bit = PORTC_HOTPLUG_LIVE_STATUS;
 		break;
 	case PORT_D:
-		bit = DPD_HOTPLUG_LIVE_STATUS;
+		bit = PORTD_HOTPLUG_LIVE_STATUS;
 		break;
 	default:
 		return connector_status_unknown;
@@ -2838,15 +2838,15 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 		name = "DPDDC-A";
 		break;
 	case PORT_B:
-		dev_priv->hotplug_supported_mask |= DPB_HOTPLUG_INT_STATUS;
+		dev_priv->hotplug_supported_mask |= PORTB_HOTPLUG_INT_STATUS;
 		name = "DPDDC-B";
 		break;
 	case PORT_C:
-		dev_priv->hotplug_supported_mask |= DPC_HOTPLUG_INT_STATUS;
+		dev_priv->hotplug_supported_mask |= PORTC_HOTPLUG_INT_STATUS;
 		name = "DPDDC-C";
 		break;
 	case PORT_D:
-		dev_priv->hotplug_supported_mask |= DPD_HOTPLUG_INT_STATUS;
+		dev_priv->hotplug_supported_mask |= PORTD_HOTPLUG_INT_STATUS;
 		name = "DPDDC-D";
 		break;
 	default:
diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c
index 5b4efd6..5a6138c 100644
--- a/drivers/gpu/drm/i915/intel_hdmi.c
+++ b/drivers/gpu/drm/i915/intel_hdmi.c
@@ -802,10 +802,10 @@ static bool g4x_hdmi_connected(struct intel_hdmi *intel_hdmi)
 
 	switch (intel_dig_port->port) {
 	case PORT_B:
-		bit = HDMIB_HOTPLUG_LIVE_STATUS;
+		bit = PORTB_HOTPLUG_LIVE_STATUS;
 		break;
 	case PORT_C:
-		bit = HDMIC_HOTPLUG_LIVE_STATUS;
+		bit = PORTC_HOTPLUG_LIVE_STATUS;
 		break;
 	default:
 		bit = 0;
@@ -1022,15 +1022,15 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
 	switch (port) {
 	case PORT_B:
 		intel_hdmi->ddc_bus = GMBUS_PORT_DPB;
-		dev_priv->hotplug_supported_mask |= HDMIB_HOTPLUG_INT_STATUS;
+		dev_priv->hotplug_supported_mask |= PORTB_HOTPLUG_INT_STATUS;
 		break;
 	case PORT_C:
 		intel_hdmi->ddc_bus = GMBUS_PORT_DPC;
-		dev_priv->hotplug_supported_mask |= HDMIC_HOTPLUG_INT_STATUS;
+		dev_priv->hotplug_supported_mask |= PORTC_HOTPLUG_INT_STATUS;
 		break;
 	case PORT_D:
 		intel_hdmi->ddc_bus = GMBUS_PORT_DPD;
-		dev_priv->hotplug_supported_mask |= HDMID_HOTPLUG_INT_STATUS;
+		dev_priv->hotplug_supported_mask |= PORTD_HOTPLUG_INT_STATUS;
 		break;
 	case PORT_A:
 		/* Internal port only for eDP. */
-- 
cgit v0.10.2


From 2c6602df1a7ff79c9e489602445a6d7eb728744e Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Fri, 8 Feb 2013 23:13:35 +0200
Subject: drm/i915: Fix sprite_scaling_enabled for multiple sprites
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

We have more than one sprite, so a boolean simply won't cut it.
Turn sprite_scaling_enabled into a bitmask and track the state
of sprite scaler for each sprite independently.

Also don't re-enable LP watermarks until the sprite registers
have actually been written, and thus sprite scaling has really
been disabled.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 4e5a377..a7860ea 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -925,7 +925,7 @@ typedef struct drm_i915_private {
 
 	/* overlay */
 	struct intel_overlay *overlay;
-	bool sprite_scaling_enabled;
+	unsigned int sprite_scaling_enabled;
 
 	/* LVDS info */
 	int backlight_level;  /* restore backlight to this value */
diff --git a/drivers/gpu/drm/i915/intel_sprite.c b/drivers/gpu/drm/i915/intel_sprite.c
index f829306..03cfd62 100644
--- a/drivers/gpu/drm/i915/intel_sprite.c
+++ b/drivers/gpu/drm/i915/intel_sprite.c
@@ -50,6 +50,7 @@ ivb_update_plane(struct drm_plane *plane, struct drm_framebuffer *fb,
 	u32 sprctl, sprscale = 0;
 	unsigned long sprsurf_offset, linear_offset;
 	int pixel_size = drm_format_plane_cpp(fb->pixel_format, 0);
+	bool scaling_was_enabled = dev_priv->sprite_scaling_enabled;
 
 	sprctl = I915_READ(SPRCTL(pipe));
 
@@ -103,19 +104,15 @@ ivb_update_plane(struct drm_plane *plane, struct drm_framebuffer *fb,
 	 * when scaling is disabled.
 	 */
 	if (crtc_w != src_w || crtc_h != src_h) {
-		if (!dev_priv->sprite_scaling_enabled) {
-			dev_priv->sprite_scaling_enabled = true;
+		dev_priv->sprite_scaling_enabled |= 1 << pipe;
+
+		if (!scaling_was_enabled) {
 			intel_update_watermarks(dev);
 			intel_wait_for_vblank(dev, pipe);
 		}
 		sprscale = SPRITE_SCALE_ENABLE | (src_w << 16) | src_h;
-	} else {
-		if (dev_priv->sprite_scaling_enabled) {
-			dev_priv->sprite_scaling_enabled = false;
-			/* potentially re-enable LP watermarks */
-			intel_update_watermarks(dev);
-		}
-	}
+	} else
+		dev_priv->sprite_scaling_enabled &= ~(1 << pipe);
 
 	I915_WRITE(SPRSTRIDE(pipe), fb->pitches[0]);
 	I915_WRITE(SPRPOS(pipe), (crtc_y << 16) | crtc_x);
@@ -141,6 +138,10 @@ ivb_update_plane(struct drm_plane *plane, struct drm_framebuffer *fb,
 	I915_WRITE(SPRCTL(pipe), sprctl);
 	I915_MODIFY_DISPBASE(SPRSURF(pipe), obj->gtt_offset + sprsurf_offset);
 	POSTING_READ(SPRSURF(pipe));
+
+	/* potentially re-enable LP watermarks */
+	if (scaling_was_enabled && !dev_priv->sprite_scaling_enabled)
+		intel_update_watermarks(dev);
 }
 
 static void
@@ -150,6 +151,7 @@ ivb_disable_plane(struct drm_plane *plane)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_plane *intel_plane = to_intel_plane(plane);
 	int pipe = intel_plane->pipe;
+	bool scaling_was_enabled = dev_priv->sprite_scaling_enabled;
 
 	I915_WRITE(SPRCTL(pipe), I915_READ(SPRCTL(pipe)) & ~SPRITE_ENABLE);
 	/* Can't leave the scaler enabled... */
@@ -159,8 +161,11 @@ ivb_disable_plane(struct drm_plane *plane)
 	I915_MODIFY_DISPBASE(SPRSURF(pipe), 0);
 	POSTING_READ(SPRSURF(pipe));
 
-	dev_priv->sprite_scaling_enabled = false;
-	intel_update_watermarks(dev);
+	dev_priv->sprite_scaling_enabled &= ~(1 << pipe);
+
+	/* potentially re-enable LP watermarks */
+	if (scaling_was_enabled && !dev_priv->sprite_scaling_enabled)
+		intel_update_watermarks(dev);
 }
 
 static int
-- 
cgit v0.10.2


From d46da4377689bd938795e53c4e2fb54dbcaeea44 Mon Sep 17 00:00:00 2001
From: Paulo Zanoni <paulo.r.zanoni@intel.com>
Date: Fri, 8 Feb 2013 17:35:15 -0200
Subject: drm/i915: add ibx_irq_postinstall

So we can remove duplicated code. Note that this function is used not
only on IBX, but also CPT and LPT.

Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
[danvet: Also bikeshed s/ironlake_enable_pch_hotplug/ibx_enable_hotplug
to keep consistent with our ibx for pch naming scheme.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index b27ffdb..2cd97d1 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -1924,7 +1924,7 @@ static void valleyview_irq_preinstall(struct drm_device *dev)
  * This register is the same on all known PCH chips.
  */
 
-static void ironlake_enable_pch_hotplug(struct drm_device *dev)
+static void ibx_enable_hotplug(struct drm_device *dev)
 {
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
 	u32	hotplug;
@@ -1937,6 +1937,28 @@ static void ironlake_enable_pch_hotplug(struct drm_device *dev)
 	I915_WRITE(PCH_PORT_HOTPLUG, hotplug);
 }
 
+static void ibx_irq_postinstall(struct drm_device *dev)
+{
+	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
+	u32 mask;
+
+	if (HAS_PCH_IBX(dev))
+		mask = SDE_HOTPLUG_MASK |
+		       SDE_GMBUS |
+		       SDE_AUX_MASK;
+	else
+		mask = SDE_HOTPLUG_MASK_CPT |
+		       SDE_GMBUS_CPT |
+		       SDE_AUX_MASK_CPT;
+
+	I915_WRITE(SDEIIR, I915_READ(SDEIIR));
+	I915_WRITE(SDEIMR, ~mask);
+	I915_WRITE(SDEIER, mask);
+	POSTING_READ(SDEIER);
+
+	ibx_enable_hotplug(dev);
+}
+
 static int ironlake_irq_postinstall(struct drm_device *dev)
 {
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
@@ -1945,8 +1967,6 @@ static int ironlake_irq_postinstall(struct drm_device *dev)
 			   DE_PLANEA_FLIP_DONE | DE_PLANEB_FLIP_DONE |
 			   DE_AUX_CHANNEL_A;
 	u32 render_irqs;
-	u32 hotplug_mask;
-	u32 pch_irq_mask;
 
 	dev_priv->irq_mask = ~display_mask;
 
@@ -1974,30 +1994,7 @@ static int ironlake_irq_postinstall(struct drm_device *dev)
 	I915_WRITE(GTIER, render_irqs);
 	POSTING_READ(GTIER);
 
-	if (HAS_PCH_CPT(dev)) {
-		hotplug_mask = (SDE_CRT_HOTPLUG_CPT |
-				SDE_PORTB_HOTPLUG_CPT |
-				SDE_PORTC_HOTPLUG_CPT |
-				SDE_PORTD_HOTPLUG_CPT |
-				SDE_GMBUS_CPT |
-				SDE_AUX_MASK_CPT);
-	} else {
-		hotplug_mask = (SDE_CRT_HOTPLUG |
-				SDE_PORTB_HOTPLUG |
-				SDE_PORTC_HOTPLUG |
-				SDE_PORTD_HOTPLUG |
-				SDE_GMBUS |
-				SDE_AUX_MASK);
-	}
-
-	pch_irq_mask = ~hotplug_mask;
-
-	I915_WRITE(SDEIIR, I915_READ(SDEIIR));
-	I915_WRITE(SDEIMR, pch_irq_mask);
-	I915_WRITE(SDEIER, hotplug_mask);
-	POSTING_READ(SDEIER);
-
-	ironlake_enable_pch_hotplug(dev);
+	ibx_irq_postinstall(dev);
 
 	if (IS_IRONLAKE_M(dev)) {
 		/* Clear & enable PCU event interrupts */
@@ -2020,8 +2017,6 @@ static int ivybridge_irq_postinstall(struct drm_device *dev)
 		DE_PLANEA_FLIP_DONE_IVB |
 		DE_AUX_CHANNEL_A_IVB;
 	u32 render_irqs;
-	u32 hotplug_mask;
-	u32 pch_irq_mask;
 
 	dev_priv->irq_mask = ~display_mask;
 
@@ -2045,20 +2040,7 @@ static int ivybridge_irq_postinstall(struct drm_device *dev)
 	I915_WRITE(GTIER, render_irqs);
 	POSTING_READ(GTIER);
 
-	hotplug_mask = (SDE_CRT_HOTPLUG_CPT |
-			SDE_PORTB_HOTPLUG_CPT |
-			SDE_PORTC_HOTPLUG_CPT |
-			SDE_PORTD_HOTPLUG_CPT |
-			SDE_GMBUS_CPT |
-			SDE_AUX_MASK_CPT);
-	pch_irq_mask = ~hotplug_mask;
-
-	I915_WRITE(SDEIIR, I915_READ(SDEIIR));
-	I915_WRITE(SDEIMR, pch_irq_mask);
-	I915_WRITE(SDEIER, hotplug_mask);
-	POSTING_READ(SDEIER);
-
-	ironlake_enable_pch_hotplug(dev);
+	ibx_irq_postinstall(dev);
 
 	return 0;
 }
-- 
cgit v0.10.2


From 41907ddc1b71aaa4ef5290f46f0ec49d581d6aac Mon Sep 17 00:00:00 2001
From: Ben Widawsky <ben@bwidawsk.net>
Date: Fri, 8 Feb 2013 11:32:47 -0800
Subject: drm/i915: Fix gen2 mappable calculations

When I refactored the code initially, I forgot that gen2 uses a
different bar for the CPU mappable aperture. The agp-less code knows
nothing of generations less than 5, so we have to expand the gtt_probe
function to include the mappable base and end.

It was originally broken by me:
commit baa09f5fd8a6d033ec075355dda99a65b7f6a0f3
Author: Ben Widawsky <ben@bwidawsk.net>
Date:   Thu Jan 24 13:49:57 2013 -0800

    drm/i915: Add probe and remove to the gtt ops

Reported-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/char/agp/intel-gtt.c b/drivers/char/agp/intel-gtt.c
index d8e7e6c..207e5c3 100644
--- a/drivers/char/agp/intel-gtt.c
+++ b/drivers/char/agp/intel-gtt.c
@@ -1371,10 +1371,13 @@ int intel_gmch_probe(struct pci_dev *bridge_pdev, struct pci_dev *gpu_pdev,
 }
 EXPORT_SYMBOL(intel_gmch_probe);
 
-void intel_gtt_get(size_t *gtt_total, size_t *stolen_size)
+void intel_gtt_get(size_t *gtt_total, size_t *stolen_size,
+		   phys_addr_t *mappable_base, unsigned long *mappable_end)
 {
 	*gtt_total = intel_private.gtt_total_entries << PAGE_SHIFT;
 	*stolen_size = intel_private.stolen_size;
+	*mappable_base = intel_private.gma_bus_addr;
+	*mappable_end = intel_private.gtt_mappable_entries << PAGE_SHIFT;
 }
 EXPORT_SYMBOL(intel_gtt_get);
 
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index a7860ea..0f1d15b 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -399,7 +399,8 @@ struct i915_gtt {
 
 	/* global gtt ops */
 	int (*gtt_probe)(struct drm_device *dev, size_t *gtt_total,
-			  size_t *stolen);
+			  size_t *stolen, phys_addr_t *mappable_base,
+			  unsigned long *mappable_end);
 	void (*gtt_remove)(struct drm_device *dev);
 	void (*gtt_clear_range)(struct drm_device *dev,
 				unsigned int first_entry,
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index bdaca3f..926a1e2 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -725,7 +725,9 @@ static inline size_t gen7_get_stolen_size(u16 snb_gmch_ctl)
 
 static int gen6_gmch_probe(struct drm_device *dev,
 			   size_t *gtt_total,
-			   size_t *stolen)
+			   size_t *stolen,
+			   phys_addr_t *mappable_base,
+			   unsigned long *mappable_end)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	phys_addr_t gtt_bus_addr;
@@ -733,11 +735,13 @@ static int gen6_gmch_probe(struct drm_device *dev,
 	u16 snb_gmch_ctl;
 	int ret;
 
+	*mappable_base = pci_resource_start(dev->pdev, 2);
+	*mappable_end = pci_resource_len(dev->pdev, 2);
+
 	/* 64/512MB is the current min/max we actually know of, but this is just
 	 * a coarse sanity check.
 	 */
-	if ((dev_priv->gtt.mappable_end < (64<<20) ||
-	     (dev_priv->gtt.mappable_end > (512<<20)))) {
+	if ((*mappable_end < (64<<20) || (*mappable_end > (512<<20)))) {
 		DRM_ERROR("Unknown GMADR size (%lx)\n",
 			  dev_priv->gtt.mappable_end);
 		return -ENXIO;
@@ -782,7 +786,9 @@ static void gen6_gmch_remove(struct drm_device *dev)
 
 static int i915_gmch_probe(struct drm_device *dev,
 			   size_t *gtt_total,
-			   size_t *stolen)
+			   size_t *stolen,
+			   phys_addr_t *mappable_base,
+			   unsigned long *mappable_end)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int ret;
@@ -793,7 +799,7 @@ static int i915_gmch_probe(struct drm_device *dev,
 		return -EIO;
 	}
 
-	intel_gtt_get(gtt_total, stolen);
+	intel_gtt_get(gtt_total, stolen, mappable_base, mappable_end);
 
 	dev_priv->gtt.do_idle_maps = needs_idle_maps(dev_priv->dev);
 	dev_priv->gtt.gtt_clear_range = i915_ggtt_clear_range;
@@ -814,9 +820,6 @@ int i915_gem_gtt_init(struct drm_device *dev)
 	unsigned long gtt_size;
 	int ret;
 
-	gtt->mappable_base = pci_resource_start(dev->pdev, 2);
-	gtt->mappable_end = pci_resource_len(dev->pdev, 2);
-
 	if (INTEL_INFO(dev)->gen <= 5) {
 		dev_priv->gtt.gtt_probe = i915_gmch_probe;
 		dev_priv->gtt.gtt_remove = i915_gmch_remove;
@@ -826,7 +829,9 @@ int i915_gem_gtt_init(struct drm_device *dev)
 	}
 
 	ret = dev_priv->gtt.gtt_probe(dev, &dev_priv->gtt.total,
-				     &dev_priv->gtt.stolen_size);
+				     &dev_priv->gtt.stolen_size,
+				     &gtt->mappable_base,
+				     &gtt->mappable_end);
 	if (ret)
 		return ret;
 
diff --git a/include/drm/intel-gtt.h b/include/drm/intel-gtt.h
index cf10555..b08bdad 100644
--- a/include/drm/intel-gtt.h
+++ b/include/drm/intel-gtt.h
@@ -3,7 +3,8 @@
 #ifndef _DRM_INTEL_GTT_H
 #define	_DRM_INTEL_GTT_H
 
-void intel_gtt_get(size_t *gtt_total, size_t *stolen_size);
+void intel_gtt_get(size_t *gtt_total, size_t *stolen_size,
+		   phys_addr_t *mappable_base, unsigned long *mappable_end);
 
 int intel_gmch_probe(struct pci_dev *bridge_pdev, struct pci_dev *gpu_pdev,
 		     struct agp_bridge_data *bridge);
-- 
cgit v0.10.2


From 4fc7c971c3aedf937f824c063d698779d25c3330 Mon Sep 17 00:00:00 2001
From: Ben Widawsky <ben@bwidawsk.net>
Date: Fri, 8 Feb 2013 11:49:24 -0800
Subject: drm/i915: Extract ring init from hw_init

The ring initialization will differ a bit in upcoming generations, and
this split will prepare the code for what's needed.

This patch also fixes a bug introduced in:
commit 99433931950f33039d9e1a52b4ed9af3f1b58e84
Author: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Date:   Tue Jan 22 14:12:17 2013 +0200

    drm/i915: use gem_set_seqno() on hardware init

After doing the extraction, the bad error handling became obvious.  I
acknowledge that this should be two patches, but it's a pretty
small/trivial patch. If requested, I can certainly do the fix as a
distinct patch.

v2: Should be cleanup blt, not init blt on failure (Chris)

v3: Forgot to git add on v2

Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 62be748..d8618f3 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -3930,22 +3930,11 @@ intel_enable_blt(struct drm_device *dev)
 	return true;
 }
 
-int
-i915_gem_init_hw(struct drm_device *dev)
+static int i915_gem_init_rings(struct drm_device *dev)
 {
-	drm_i915_private_t *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	int ret;
 
-	if (INTEL_INFO(dev)->gen < 6 && !intel_enable_gtt())
-		return -EIO;
-
-	if (IS_HASWELL(dev) && (I915_READ(0x120010) == 1))
-		I915_WRITE(0x9008, I915_READ(0x9008) | 0xf0000);
-
-	i915_gem_l3_remap(dev);
-
-	i915_gem_init_swizzling(dev);
-
 	ret = intel_init_render_ring_buffer(dev);
 	if (ret)
 		return ret;
@@ -3964,6 +3953,38 @@ i915_gem_init_hw(struct drm_device *dev)
 
 	ret = i915_gem_set_seqno(dev, ((u32)~0 - 0x1000));
 	if (ret)
+		goto cleanup_blt_ring;
+
+	return 0;
+
+cleanup_blt_ring:
+	intel_cleanup_ring_buffer(&dev_priv->ring[BCS]);
+cleanup_bsd_ring:
+	intel_cleanup_ring_buffer(&dev_priv->ring[VCS]);
+cleanup_render_ring:
+	intel_cleanup_ring_buffer(&dev_priv->ring[RCS]);
+
+	return ret;
+}
+
+int
+i915_gem_init_hw(struct drm_device *dev)
+{
+	drm_i915_private_t *dev_priv = dev->dev_private;
+	int ret;
+
+	if (INTEL_INFO(dev)->gen < 6 && !intel_enable_gtt())
+		return -EIO;
+
+	if (IS_HASWELL(dev) && (I915_READ(0x120010) == 1))
+		I915_WRITE(0x9008, I915_READ(0x9008) | 0xf0000);
+
+	i915_gem_l3_remap(dev);
+
+	i915_gem_init_swizzling(dev);
+
+	ret = i915_gem_init_rings(dev);
+	if (ret)
 		return ret;
 
 	/*
@@ -3974,12 +3995,6 @@ i915_gem_init_hw(struct drm_device *dev)
 	i915_gem_init_ppgtt(dev);
 
 	return 0;
-
-cleanup_bsd_ring:
-	intel_cleanup_ring_buffer(&dev_priv->ring[VCS]);
-cleanup_render_ring:
-	intel_cleanup_ring_buffer(&dev_priv->ring[RCS]);
-	return ret;
 }
 
 int i915_gem_init(struct drm_device *dev)
-- 
cgit v0.10.2


From f73f760725636b9d0c3786273e185b053516d1eb Mon Sep 17 00:00:00 2001
From: Ben Widawsky <ben@bwidawsk.net>
Date: Mon, 11 Feb 2013 13:31:27 -0800
Subject: drm/i915/ctx: Remove bad invariant

It's not that the assertion is incorrect, but rather that we can call
do_destroy early in loading, and we will falsely BUG().

Since contexts have been in for a while now, and in the internal APIs
are pretty stable, it should be fairly safe to remove this.

v2: Remove unused dev_priv, and dev

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c
index e7ecafa..21177d9 100644
--- a/drivers/gpu/drm/i915/i915_gem_context.c
+++ b/drivers/gpu/drm/i915/i915_gem_context.c
@@ -126,13 +126,8 @@ static int get_context_size(struct drm_device *dev)
 
 static void do_destroy(struct i915_hw_context *ctx)
 {
-	struct drm_device *dev = ctx->obj->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
 	if (ctx->file_priv)
 		idr_remove(&ctx->file_priv->context_idr, ctx->id);
-	else
-		BUG_ON(ctx != dev_priv->ring[RCS].default_context);
 
 	drm_gem_object_unreference(&ctx->obj->base);
 	kfree(ctx);
-- 
cgit v0.10.2


From 4836d15789c85224662d3e037694d15ec9901631 Mon Sep 17 00:00:00 2001
From: Andy Gross <andy.gross@ti.com>
Date: Wed, 19 Dec 2012 14:53:37 -0600
Subject: drm/omap: Add PM capabilities

Added power management capabilities into the omapdrm and DMM drivers.
During suspend, we don't need to do anything to maintain the state of
the LUT.  We have all the necessary information to recreate the mappings
of the GEM object list maintained by the omapdrm driver.

On resume, the DMM resume handler will first reprogram the LUT to point
to the dummy page.  The subsequent resume handler in the omapdrm will call
into the DMM and reprogram each of the buffer objects.  This will ensure
that all of the necessary objects will be pinned into the DMM properly.

Order of suspend/resume handlers is done by device creation.  We create
the DMM device before the omapdrm, so the correct order is maintained.

Signed-off-by: Andy Gross <andy.gross@ti.com>
Signed-off-by: Rob Clark <rob@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

diff --git a/drivers/staging/omapdrm/omap_dmm_tiler.c b/drivers/staging/omapdrm/omap_dmm_tiler.c
index 59bf438..c42c5e5 100644
--- a/drivers/staging/omapdrm/omap_dmm_tiler.c
+++ b/drivers/staging/omapdrm/omap_dmm_tiler.c
@@ -903,12 +903,46 @@ error:
 }
 #endif
 
+#ifdef CONFIG_PM
+static int omap_dmm_resume(struct device *dev)
+{
+	struct tcm_area area;
+	int i;
+
+	if (!omap_dmm)
+		return -ENODEV;
+
+	area = (struct tcm_area) {
+		.is2d = true,
+		.tcm = NULL,
+		.p1.x = omap_dmm->container_width - 1,
+		.p1.y = omap_dmm->container_height - 1,
+	};
+
+	/* initialize all LUTs to dummy page entries */
+	for (i = 0; i < omap_dmm->num_lut; i++) {
+		area.tcm = omap_dmm->tcm[i];
+		if (fill(&area, NULL, 0, 0, true))
+			dev_err(dev, "refill failed");
+	}
+
+	return 0;
+}
+
+static const struct dev_pm_ops omap_dmm_pm_ops = {
+	.resume = omap_dmm_resume,
+};
+#endif
+
 struct platform_driver omap_dmm_driver = {
 	.probe = omap_dmm_probe,
 	.remove = omap_dmm_remove,
 	.driver = {
 		.owner = THIS_MODULE,
 		.name = DMM_DRIVER_NAME,
+#ifdef CONFIG_PM
+		.pm = &omap_dmm_pm_ops,
+#endif
 	},
 };
 
diff --git a/drivers/staging/omapdrm/omap_drv.c b/drivers/staging/omapdrm/omap_drv.c
index dfdb4ba..9d6a584 100644
--- a/drivers/staging/omapdrm/omap_drv.c
+++ b/drivers/staging/omapdrm/omap_drv.c
@@ -368,6 +368,9 @@ static int dev_load(struct drm_device *dev, unsigned long flags)
 		/* well, limp along without an fbdev.. maybe X11 will work? */
 	}
 
+	/* store off drm_device for use in pm ops */
+	dev_set_drvdata(dev->dev, dev);
+
 	drm_kms_helper_poll_init(dev);
 
 	return 0;
@@ -393,6 +396,8 @@ static int dev_unload(struct drm_device *dev)
 	kfree(dev->dev_private);
 	dev->dev_private = NULL;
 
+	dev_set_drvdata(dev->dev, NULL);
+
 	return 0;
 }
 
@@ -558,10 +563,19 @@ static int pdev_remove(struct platform_device *device)
 	return 0;
 }
 
+#ifdef CONFIG_PM
+static const struct dev_pm_ops omapdrm_pm_ops = {
+	.resume = omap_gem_resume,
+};
+#endif
+
 struct platform_driver pdev = {
 		.driver = {
 			.name = DRIVER_NAME,
 			.owner = THIS_MODULE,
+#ifdef CONFIG_PM
+			.pm = &omapdrm_pm_ops,
+#endif
 		},
 		.probe = pdev_probe,
 		.remove = pdev_remove,
diff --git a/drivers/staging/omapdrm/omap_drv.h b/drivers/staging/omapdrm/omap_drv.h
index cd1f22b..f921027 100644
--- a/drivers/staging/omapdrm/omap_drv.h
+++ b/drivers/staging/omapdrm/omap_drv.h
@@ -135,6 +135,10 @@ void omap_gem_describe(struct drm_gem_object *obj, struct seq_file *m);
 void omap_gem_describe_objects(struct list_head *list, struct seq_file *m);
 #endif
 
+#ifdef CONFIG_PM
+int omap_gem_resume(struct device *dev);
+#endif
+
 int omap_irq_enable_vblank(struct drm_device *dev, int crtc);
 void omap_irq_disable_vblank(struct drm_device *dev, int crtc);
 irqreturn_t omap_irq_handler(DRM_IRQ_ARGS);
diff --git a/drivers/staging/omapdrm/omap_gem.c b/drivers/staging/omapdrm/omap_gem.c
index c38992b..08f1e292 100644
--- a/drivers/staging/omapdrm/omap_gem.c
+++ b/drivers/staging/omapdrm/omap_gem.c
@@ -964,6 +964,34 @@ void *omap_gem_vaddr(struct drm_gem_object *obj)
 	return omap_obj->vaddr;
 }
 
+#ifdef CONFIG_PM
+/* re-pin objects in DMM in resume path: */
+int omap_gem_resume(struct device *dev)
+{
+	struct drm_device *drm_dev = dev_get_drvdata(dev);
+	struct omap_drm_private *priv = drm_dev->dev_private;
+	struct omap_gem_object *omap_obj;
+	int ret = 0;
+
+	list_for_each_entry(omap_obj, &priv->obj_list, mm_list) {
+		if (omap_obj->block) {
+			struct drm_gem_object *obj = &omap_obj->base;
+			uint32_t npages = obj->size >> PAGE_SHIFT;
+			WARN_ON(!omap_obj->pages);  /* this can't happen */
+			ret = tiler_pin(omap_obj->block,
+					omap_obj->pages, npages,
+					omap_obj->roll, true);
+			if (ret) {
+				dev_err(dev, "could not repin: %d\n", ret);
+				return ret;
+			}
+		}
+	}
+
+	return 0;
+}
+#endif
+
 #ifdef CONFIG_DEBUG_FS
 void omap_gem_describe(struct drm_gem_object *obj, struct seq_file *m)
 {
-- 
cgit v0.10.2


From 238083ad5d67d1c9ba6e7158a027e834a48bb23b Mon Sep 17 00:00:00 2001
From: Andy Gross <andy.gross@ti.com>
Date: Wed, 19 Dec 2012 14:53:38 -0600
Subject: drm/omap: Add OMAP5 support

Add support for OMAP5 processor.  The main differences are that the OMAP5
has 2 containers, one for 1D and one for 2D.  Each container is 128MiB in
size.

Signed-off-by: Andy Gross <andy.gross@ti.com>
Signed-off-by: Rob Clark <rob@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

diff --git a/drivers/staging/omapdrm/omap_dmm_priv.h b/drivers/staging/omapdrm/omap_dmm_priv.h
index 273ec12..58bcd6a 100644
--- a/drivers/staging/omapdrm/omap_dmm_priv.h
+++ b/drivers/staging/omapdrm/omap_dmm_priv.h
@@ -118,6 +118,11 @@ struct pat {
 #define DESCR_SIZE 128
 #define REFILL_BUFFER_SIZE ((4 * 128 * 256) + (3 * DESCR_SIZE))
 
+/* For OMAP5, a fixed offset is added to all Y coordinates for 1D buffers.
+ * This is used in programming to address the upper portion of the LUT
+*/
+#define OMAP5_LUT_OFFSET       128
+
 struct dmm;
 
 struct dmm_txn {
diff --git a/drivers/staging/omapdrm/omap_dmm_tiler.c b/drivers/staging/omapdrm/omap_dmm_tiler.c
index c42c5e5..3910215 100644
--- a/drivers/staging/omapdrm/omap_dmm_tiler.c
+++ b/drivers/staging/omapdrm/omap_dmm_tiler.c
@@ -213,6 +213,11 @@ static void dmm_txn_append(struct dmm_txn *txn, struct pat_area *area,
 		txn->last_pat->next_pa = (uint32_t)pat_pa;
 
 	pat->area = *area;
+
+	/* adjust Y coordinates based off of container parameters */
+	pat->area.y0 += engine->tcm->y_offset;
+	pat->area.y1 += engine->tcm->y_offset;
+
 	pat->ctrl = (struct pat_ctrl){
 			.start = 1,
 			.lut_id = engine->tcm->lut_id,
@@ -622,6 +627,11 @@ static int omap_dmm_probe(struct platform_device *dev)
 	omap_dmm->lut_width = ((pat_geom >> 16) & 0xF) << 5;
 	omap_dmm->lut_height = ((pat_geom >> 24) & 0xF) << 5;
 
+	/* increment LUT by one if on OMAP5 */
+	/* LUT has twice the height, and is split into a separate container */
+	if (omap_dmm->lut_height != omap_dmm->container_height)
+		omap_dmm->num_lut++;
+
 	/* initialize DMM registers */
 	writel(0x88888888, omap_dmm->base + DMM_PAT_VIEW__0);
 	writel(0x88888888, omap_dmm->base + DMM_PAT_VIEW__1);
@@ -701,6 +711,9 @@ static int omap_dmm_probe(struct platform_device *dev)
 	}
 
 	/* init containers */
+	/* Each LUT is associated with a TCM (container manager).  We use the
+	   lut_id to denote the lut_id used to identify the correct LUT for
+	   programming during reill operations */
 	for (i = 0; i < omap_dmm->num_lut; i++) {
 		omap_dmm->tcm[i] = sita_init(omap_dmm->container_width,
 						omap_dmm->container_height,
@@ -717,13 +730,23 @@ static int omap_dmm_probe(struct platform_device *dev)
 
 	/* assign access mode containers to applicable tcm container */
 	/* OMAP 4 has 1 container for all 4 views */
+	/* OMAP 5 has 2 containers, 1 for 2D and 1 for 1D */
 	containers[TILFMT_8BIT] = omap_dmm->tcm[0];
 	containers[TILFMT_16BIT] = omap_dmm->tcm[0];
 	containers[TILFMT_32BIT] = omap_dmm->tcm[0];
-	containers[TILFMT_PAGE] = omap_dmm->tcm[0];
+
+	if (omap_dmm->container_height != omap_dmm->lut_height) {
+		/* second LUT is used for PAGE mode.  Programming must use
+		   y offset that is added to all y coordinates.  LUT id is still
+		   0, because it is the same LUT, just the upper 128 lines */
+		containers[TILFMT_PAGE] = omap_dmm->tcm[1];
+		omap_dmm->tcm[1]->y_offset = OMAP5_LUT_OFFSET;
+		omap_dmm->tcm[1]->lut_id = 0;
+	} else {
+		containers[TILFMT_PAGE] = omap_dmm->tcm[0];
+	}
 
 	area = (struct tcm_area) {
-		.is2d = true,
 		.tcm = NULL,
 		.p1.x = omap_dmm->container_width - 1,
 		.p1.y = omap_dmm->container_height - 1,
@@ -835,64 +858,81 @@ int tiler_map_show(struct seq_file *s, void *arg)
 	int h_adj;
 	int w_adj;
 	unsigned long flags;
+	int lut_idx;
+
 
 	if (!omap_dmm) {
 		/* early return if dmm/tiler device is not initialized */
 		return 0;
 	}
 
-	h_adj = omap_dmm->lut_height / ydiv;
-	w_adj = omap_dmm->lut_width / xdiv;
+	h_adj = omap_dmm->container_height / ydiv;
+	w_adj = omap_dmm->container_width / xdiv;
 
-	map = kzalloc(h_adj * sizeof(*map), GFP_KERNEL);
-	global_map = kzalloc((w_adj + 1) * h_adj, GFP_KERNEL);
+	map = kmalloc(h_adj * sizeof(*map), GFP_KERNEL);
+	global_map = kmalloc((w_adj + 1) * h_adj, GFP_KERNEL);
 
 	if (!map || !global_map)
 		goto error;
 
-	memset(global_map, ' ', (w_adj + 1) * h_adj);
-	for (i = 0; i < omap_dmm->lut_height; i++) {
-		map[i] = global_map + i * (w_adj + 1);
-		map[i][w_adj] = 0;
-	}
-	spin_lock_irqsave(&list_lock, flags);
+	for (lut_idx = 0; lut_idx < omap_dmm->num_lut; lut_idx++) {
+		memset(map, 0, sizeof(h_adj * sizeof(*map)));
+		memset(global_map, ' ', (w_adj + 1) * h_adj);
 
-	list_for_each_entry(block, &omap_dmm->alloc_head, alloc_node) {
-		if (block->fmt != TILFMT_PAGE) {
-			fill_map(map, xdiv, ydiv, &block->area, *m2dp, true);
-			if (!*++a2dp)
-				a2dp = a2d;
-			if (!*++m2dp)
-				m2dp = m2d;
-			map_2d_info(map, xdiv, ydiv, nice, &block->area);
-		} else {
-			bool start = read_map_pt(map, xdiv, ydiv,
-							&block->area.p0)
-									== ' ';
-			bool end = read_map_pt(map, xdiv, ydiv, &block->area.p1)
-									== ' ';
-			tcm_for_each_slice(a, block->area, p)
-				fill_map(map, xdiv, ydiv, &a, '=', true);
-			fill_map_pt(map, xdiv, ydiv, &block->area.p0,
+		for (i = 0; i < omap_dmm->container_height; i++) {
+			map[i] = global_map + i * (w_adj + 1);
+			map[i][w_adj] = 0;
+		}
+
+		spin_lock_irqsave(&list_lock, flags);
+
+		list_for_each_entry(block, &omap_dmm->alloc_head, alloc_node) {
+			if (block->area.tcm == omap_dmm->tcm[lut_idx]) {
+				if (block->fmt != TILFMT_PAGE) {
+					fill_map(map, xdiv, ydiv, &block->area,
+						*m2dp, true);
+					if (!*++a2dp)
+						a2dp = a2d;
+					if (!*++m2dp)
+						m2dp = m2d;
+					map_2d_info(map, xdiv, ydiv, nice,
+							&block->area);
+				} else {
+					bool start = read_map_pt(map, xdiv,
+						ydiv, &block->area.p0) == ' ';
+					bool end = read_map_pt(map, xdiv, ydiv,
+							&block->area.p1) == ' ';
+
+					tcm_for_each_slice(a, block->area, p)
+						fill_map(map, xdiv, ydiv, &a,
+							'=', true);
+					fill_map_pt(map, xdiv, ydiv,
+							&block->area.p0,
 							start ? '<' : 'X');
-			fill_map_pt(map, xdiv, ydiv, &block->area.p1,
+					fill_map_pt(map, xdiv, ydiv,
+							&block->area.p1,
 							end ? '>' : 'X');
-			map_1d_info(map, xdiv, ydiv, nice, &block->area);
+					map_1d_info(map, xdiv, ydiv, nice,
+							&block->area);
+				}
+			}
 		}
-	}
 
-	spin_unlock_irqrestore(&list_lock, flags);
+		spin_unlock_irqrestore(&list_lock, flags);
 
-	if (s) {
-		seq_printf(s, "BEGIN DMM TILER MAP\n");
-		for (i = 0; i < 128; i++)
-			seq_printf(s, "%03d:%s\n", i, map[i]);
-		seq_printf(s, "END TILER MAP\n");
-	} else {
-		dev_dbg(omap_dmm->dev, "BEGIN DMM TILER MAP\n");
-		for (i = 0; i < 128; i++)
-			dev_dbg(omap_dmm->dev, "%03d:%s\n", i, map[i]);
-		dev_dbg(omap_dmm->dev, "END TILER MAP\n");
+		if (s) {
+			seq_printf(s, "CONTAINER %d DUMP BEGIN\n", lut_idx);
+			for (i = 0; i < 128; i++)
+				seq_printf(s, "%03d:%s\n", i, map[i]);
+			seq_printf(s, "CONTAINER %d DUMP END\n", lut_idx);
+		} else {
+			dev_dbg(omap_dmm->dev, "CONTAINER %d DUMP BEGIN\n",
+				lut_idx);
+			for (i = 0; i < 128; i++)
+				dev_dbg(omap_dmm->dev, "%03d:%s\n", i, map[i]);
+			dev_dbg(omap_dmm->dev, "CONTAINER %d DUMP END\n",
+				lut_idx);
+		}
 	}
 
 error:
@@ -913,7 +953,6 @@ static int omap_dmm_resume(struct device *dev)
 		return -ENODEV;
 
 	area = (struct tcm_area) {
-		.is2d = true,
 		.tcm = NULL,
 		.p1.x = omap_dmm->container_width - 1,
 		.p1.y = omap_dmm->container_height - 1,
diff --git a/drivers/staging/omapdrm/tcm.h b/drivers/staging/omapdrm/tcm.h
index d273e3e..a8d5ce4 100644
--- a/drivers/staging/omapdrm/tcm.h
+++ b/drivers/staging/omapdrm/tcm.h
@@ -59,6 +59,8 @@ struct tcm {
 	u16 width, height;	/* container dimensions */
 	int lut_id;		/* Lookup table identifier */
 
+	unsigned int y_offset;	/* offset to use for y coordinates */
+
 	/* 'pvt' structure shall contain any tcm details (attr) along with
 	linked list of allocated areas and mutex for mutually exclusive access
 	to the list.  It may also contain copies of width and height to notice
-- 
cgit v0.10.2


From 32ac1a5286d56b3c822f4a1a8d0f7fc5d586de38 Mon Sep 17 00:00:00 2001
From: Cong Ding <dinggnu@gmail.com>
Date: Tue, 15 Jan 2013 20:46:50 +0100
Subject: staging: omapdrm/omap_gem_dmabuf.c: fix memory leakage

There is a memory leakage in variable sg if it goes to error.

Signed-off-by: Cong Ding <dinggnu@gmail.com>
Signed-off-by: Rob Clark <rob@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

diff --git a/drivers/staging/omapdrm/omap_gem_dmabuf.c b/drivers/staging/omapdrm/omap_gem_dmabuf.c
index b6c5b5c..a3236ab 100644
--- a/drivers/staging/omapdrm/omap_gem_dmabuf.c
+++ b/drivers/staging/omapdrm/omap_gem_dmabuf.c
@@ -53,10 +53,10 @@ static struct sg_table *omap_gem_map_dma_buf(
 	/* this should be after _get_paddr() to ensure we have pages attached */
 	omap_gem_dma_sync(obj, dir);
 
-out:
-	if (ret)
-		return ERR_PTR(ret);
 	return sg;
+out:
+	kfree(sg);
+	return ERR_PTR(ret);
 }
 
 static void omap_gem_unmap_dma_buf(struct dma_buf_attachment *attachment,
-- 
cgit v0.10.2


From a4462f246c8821f625f45bce52c7ca7e0207dffe Mon Sep 17 00:00:00 2001
From: Peter Huewe <peterhuewe@gmx.de>
Date: Sat, 26 Jan 2013 00:40:13 +0100
Subject: staging/omapdrm: Use kmemdup rather than duplicating its
 implementation

Found with coccicheck.
The semantic patch that makes this change is available
in scripts/coccinelle/api/memdup.cocci.

Signed-off-by: Peter Huewe <peterhuewe@gmx.de>
Signed-off-by: Rob Clark <rob@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

diff --git a/drivers/staging/omapdrm/omap_gem.c b/drivers/staging/omapdrm/omap_gem.c
index 08f1e292..f9297eb 100644
--- a/drivers/staging/omapdrm/omap_gem.c
+++ b/drivers/staging/omapdrm/omap_gem.c
@@ -1267,12 +1267,12 @@ int omap_gem_set_sync_object(struct drm_gem_object *obj, void *syncobj)
 
 	if ((omap_obj->flags & OMAP_BO_EXT_SYNC) && !syncobj) {
 		/* clearing a previously set syncobj */
-		syncobj = kzalloc(sizeof(*omap_obj->sync), GFP_ATOMIC);
+		syncobj = kmemdup(omap_obj->sync, sizeof(*omap_obj->sync),
+				  GFP_ATOMIC);
 		if (!syncobj) {
 			ret = -ENOMEM;
 			goto unlock;
 		}
-		memcpy(syncobj, omap_obj->sync, sizeof(*omap_obj->sync));
 		omap_obj->flags &= ~OMAP_BO_EXT_SYNC;
 		omap_obj->sync = syncobj;
 	} else if (syncobj && !(omap_obj->flags & OMAP_BO_EXT_SYNC)) {
-- 
cgit v0.10.2


From 8bb0daffb0b8e45188066255b4203446eae181f1 Mon Sep 17 00:00:00 2001
From: Rob Clark <robdclark@gmail.com>
Date: Mon, 11 Feb 2013 12:43:09 -0500
Subject: drm/omap: move out of staging

Now that the omapdss interface has been reworked so that omapdrm can use
dispc directly, we have been able to fix the remaining functional kms
issues with omapdrm.  And in the mean time the PM sequencing and many
other of that open issues have been solved.  So I think it makes sense
to finally move omapdrm out of staging.

Signed-off-by: Rob Clark <robdclark@gmail.com>

diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
index ed9e3af..0ce5f52 100644
--- a/drivers/gpu/drm/Kconfig
+++ b/drivers/gpu/drm/Kconfig
@@ -215,3 +215,5 @@ source "drivers/gpu/drm/cirrus/Kconfig"
 source "drivers/gpu/drm/shmobile/Kconfig"
 
 source "drivers/gpu/drm/tegra/Kconfig"
+
+source "drivers/gpu/drm/omapdrm/Kconfig"
diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile
index 6f58c81..b6b43cb 100644
--- a/drivers/gpu/drm/Makefile
+++ b/drivers/gpu/drm/Makefile
@@ -50,4 +50,5 @@ obj-$(CONFIG_DRM_UDL) += udl/
 obj-$(CONFIG_DRM_AST) += ast/
 obj-$(CONFIG_DRM_SHMOBILE) +=shmobile/
 obj-$(CONFIG_DRM_TEGRA) += tegra/
+obj-$(CONFIG_DRM_OMAP)	+= omapdrm/
 obj-y			+= i2c/
diff --git a/drivers/gpu/drm/omapdrm/Kconfig b/drivers/gpu/drm/omapdrm/Kconfig
new file mode 100644
index 0000000..b724a41
--- /dev/null
+++ b/drivers/gpu/drm/omapdrm/Kconfig
@@ -0,0 +1,25 @@
+
+config DRM_OMAP
+	tristate "OMAP DRM"
+	depends on DRM && !CONFIG_FB_OMAP2
+	depends on ARCH_OMAP2PLUS || ARCH_MULTIPLATFORM
+	select DRM_KMS_HELPER
+	select OMAP2_DSS
+	select FB_SYS_FILLRECT
+	select FB_SYS_COPYAREA
+	select FB_SYS_IMAGEBLIT
+	select FB_SYS_FOPS
+	default n
+	help
+	  DRM display driver for OMAP2/3/4 based boards.
+
+config DRM_OMAP_NUM_CRTCS
+	int "Number of CRTCs"
+	range 1 10
+	default 1  if ARCH_OMAP2 || ARCH_OMAP3
+	default 2  if ARCH_OMAP4
+	depends on DRM_OMAP
+	help
+	  Select the number of video overlays which can be used as framebuffers.
+	  The remaining overlays are reserved for video.
+
diff --git a/drivers/gpu/drm/omapdrm/Makefile b/drivers/gpu/drm/omapdrm/Makefile
new file mode 100644
index 0000000..d85e058
--- /dev/null
+++ b/drivers/gpu/drm/omapdrm/Makefile
@@ -0,0 +1,24 @@
+#
+# Makefile for the drm device driver.  This driver provides support for the
+# Direct Rendering Infrastructure (DRI)
+#
+
+ccflags-y := -Iinclude/drm -Werror
+omapdrm-y := omap_drv.o \
+	omap_irq.o \
+	omap_debugfs.o \
+	omap_crtc.o \
+	omap_plane.o \
+	omap_encoder.o \
+	omap_connector.o \
+	omap_fb.o \
+	omap_fbdev.o \
+	omap_gem.o \
+	omap_gem_dmabuf.o \
+	omap_dmm_tiler.o \
+	tcm-sita.o
+
+# temporary:
+omapdrm-y += omap_gem_helpers.o
+
+obj-$(CONFIG_DRM_OMAP)	+= omapdrm.o
diff --git a/drivers/gpu/drm/omapdrm/TODO b/drivers/gpu/drm/omapdrm/TODO
new file mode 100644
index 0000000..4d8c18a
--- /dev/null
+++ b/drivers/gpu/drm/omapdrm/TODO
@@ -0,0 +1,23 @@
+TODO
+. Where should we do eviction (detatch_pages())?  We aren't necessarily
+  accessing the pages via a GART, so maybe we need some other threshold
+  to put a cap on the # of pages that can be pin'd.
+  . Use mm_shrinker to trigger unpinning pages.
+  . This is mainly theoretical since most of these devices don't actually
+    have swap or harddrive.
+. GEM/shmem backed pages can have existing mappings (kernel linear map,
+  etc..), which isn't really ideal.
+. Revisit GEM sync object infrastructure.. TTM has some framework for this
+  already.  Possibly this could be refactored out and made more common?
+  There should be some way to do this with less wheel-reinvention.
+  . This can be handled by the dma-buf fence/reservation stuff when it
+    lands
+
+Userspace:
+. git://anongit.freedesktop.org/xorg/driver/xf86-video-omap
+
+Currently tested on
+. OMAP3530 beagleboard
+. OMAP4430 pandaboard
+. OMAP4460 pandaboard
+. OMAP5432 uEVM
diff --git a/drivers/gpu/drm/omapdrm/omap_connector.c b/drivers/gpu/drm/omapdrm/omap_connector.c
new file mode 100644
index 0000000..44284fd
--- /dev/null
+++ b/drivers/gpu/drm/omapdrm/omap_connector.c
@@ -0,0 +1,298 @@
+/*
+ * drivers/gpu/drm/omapdrm/omap_connector.c
+ *
+ * Copyright (C) 2011 Texas Instruments
+ * Author: Rob Clark <rob@ti.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "omap_drv.h"
+
+#include "drm_crtc.h"
+#include "drm_crtc_helper.h"
+
+/*
+ * connector funcs
+ */
+
+#define to_omap_connector(x) container_of(x, struct omap_connector, base)
+
+struct omap_connector {
+	struct drm_connector base;
+	struct omap_dss_device *dssdev;
+	struct drm_encoder *encoder;
+};
+
+void copy_timings_omap_to_drm(struct drm_display_mode *mode,
+		struct omap_video_timings *timings)
+{
+	mode->clock = timings->pixel_clock;
+
+	mode->hdisplay = timings->x_res;
+	mode->hsync_start = mode->hdisplay + timings->hfp;
+	mode->hsync_end = mode->hsync_start + timings->hsw;
+	mode->htotal = mode->hsync_end + timings->hbp;
+
+	mode->vdisplay = timings->y_res;
+	mode->vsync_start = mode->vdisplay + timings->vfp;
+	mode->vsync_end = mode->vsync_start + timings->vsw;
+	mode->vtotal = mode->vsync_end + timings->vbp;
+
+	mode->flags = 0;
+
+	if (timings->interlace)
+		mode->flags |= DRM_MODE_FLAG_INTERLACE;
+
+	if (timings->hsync_level == OMAPDSS_SIG_ACTIVE_HIGH)
+		mode->flags |= DRM_MODE_FLAG_PHSYNC;
+	else
+		mode->flags |= DRM_MODE_FLAG_NHSYNC;
+
+	if (timings->vsync_level == OMAPDSS_SIG_ACTIVE_HIGH)
+		mode->flags |= DRM_MODE_FLAG_PVSYNC;
+	else
+		mode->flags |= DRM_MODE_FLAG_NVSYNC;
+}
+
+void copy_timings_drm_to_omap(struct omap_video_timings *timings,
+		struct drm_display_mode *mode)
+{
+	timings->pixel_clock = mode->clock;
+
+	timings->x_res = mode->hdisplay;
+	timings->hfp = mode->hsync_start - mode->hdisplay;
+	timings->hsw = mode->hsync_end - mode->hsync_start;
+	timings->hbp = mode->htotal - mode->hsync_end;
+
+	timings->y_res = mode->vdisplay;
+	timings->vfp = mode->vsync_start - mode->vdisplay;
+	timings->vsw = mode->vsync_end - mode->vsync_start;
+	timings->vbp = mode->vtotal - mode->vsync_end;
+
+	timings->interlace = !!(mode->flags & DRM_MODE_FLAG_INTERLACE);
+
+	if (mode->flags & DRM_MODE_FLAG_PHSYNC)
+		timings->hsync_level = OMAPDSS_SIG_ACTIVE_HIGH;
+	else
+		timings->hsync_level = OMAPDSS_SIG_ACTIVE_LOW;
+
+	if (mode->flags & DRM_MODE_FLAG_PVSYNC)
+		timings->vsync_level = OMAPDSS_SIG_ACTIVE_HIGH;
+	else
+		timings->vsync_level = OMAPDSS_SIG_ACTIVE_LOW;
+
+	timings->data_pclk_edge = OMAPDSS_DRIVE_SIG_RISING_EDGE;
+	timings->de_level = OMAPDSS_SIG_ACTIVE_HIGH;
+	timings->sync_pclk_edge = OMAPDSS_DRIVE_SIG_OPPOSITE_EDGES;
+}
+
+static enum drm_connector_status omap_connector_detect(
+		struct drm_connector *connector, bool force)
+{
+	struct omap_connector *omap_connector = to_omap_connector(connector);
+	struct omap_dss_device *dssdev = omap_connector->dssdev;
+	struct omap_dss_driver *dssdrv = dssdev->driver;
+	enum drm_connector_status ret;
+
+	if (dssdrv->detect) {
+		if (dssdrv->detect(dssdev))
+			ret = connector_status_connected;
+		else
+			ret = connector_status_disconnected;
+	} else {
+		ret = connector_status_unknown;
+	}
+
+	VERB("%s: %d (force=%d)", omap_connector->dssdev->name, ret, force);
+
+	return ret;
+}
+
+static void omap_connector_destroy(struct drm_connector *connector)
+{
+	struct omap_connector *omap_connector = to_omap_connector(connector);
+	struct omap_dss_device *dssdev = omap_connector->dssdev;
+
+	DBG("%s", omap_connector->dssdev->name);
+	drm_sysfs_connector_remove(connector);
+	drm_connector_cleanup(connector);
+	kfree(omap_connector);
+
+	omap_dss_put_device(dssdev);
+}
+
+#define MAX_EDID  512
+
+static int omap_connector_get_modes(struct drm_connector *connector)
+{
+	struct omap_connector *omap_connector = to_omap_connector(connector);
+	struct omap_dss_device *dssdev = omap_connector->dssdev;
+	struct omap_dss_driver *dssdrv = dssdev->driver;
+	struct drm_device *dev = connector->dev;
+	int n = 0;
+
+	DBG("%s", omap_connector->dssdev->name);
+
+	/* if display exposes EDID, then we parse that in the normal way to
+	 * build table of supported modes.. otherwise (ie. fixed resolution
+	 * LCD panels) we just return a single mode corresponding to the
+	 * currently configured timings:
+	 */
+	if (dssdrv->read_edid) {
+		void *edid = kzalloc(MAX_EDID, GFP_KERNEL);
+
+		if ((dssdrv->read_edid(dssdev, edid, MAX_EDID) > 0) &&
+				drm_edid_is_valid(edid)) {
+			drm_mode_connector_update_edid_property(
+					connector, edid);
+			n = drm_add_edid_modes(connector, edid);
+		} else {
+			drm_mode_connector_update_edid_property(
+					connector, NULL);
+		}
+		kfree(edid);
+	} else {
+		struct drm_display_mode *mode = drm_mode_create(dev);
+		struct omap_video_timings timings = {0};
+
+		dssdrv->get_timings(dssdev, &timings);
+
+		copy_timings_omap_to_drm(mode, &timings);
+
+		mode->type = DRM_MODE_TYPE_DRIVER | DRM_MODE_TYPE_PREFERRED;
+		drm_mode_set_name(mode);
+		drm_mode_probed_add(connector, mode);
+
+		n = 1;
+	}
+
+	return n;
+}
+
+static int omap_connector_mode_valid(struct drm_connector *connector,
+				 struct drm_display_mode *mode)
+{
+	struct omap_connector *omap_connector = to_omap_connector(connector);
+	struct omap_dss_device *dssdev = omap_connector->dssdev;
+	struct omap_dss_driver *dssdrv = dssdev->driver;
+	struct omap_video_timings timings = {0};
+	struct drm_device *dev = connector->dev;
+	struct drm_display_mode *new_mode;
+	int ret = MODE_BAD;
+
+	copy_timings_drm_to_omap(&timings, mode);
+	mode->vrefresh = drm_mode_vrefresh(mode);
+
+	if (!dssdrv->check_timings(dssdev, &timings)) {
+		/* check if vrefresh is still valid */
+		new_mode = drm_mode_duplicate(dev, mode);
+		new_mode->clock = timings.pixel_clock;
+		new_mode->vrefresh = 0;
+		if (mode->vrefresh == drm_mode_vrefresh(new_mode))
+			ret = MODE_OK;
+		drm_mode_destroy(dev, new_mode);
+	}
+
+	DBG("connector: mode %s: "
+			"%d:\"%s\" %d %d %d %d %d %d %d %d %d %d 0x%x 0x%x",
+			(ret == MODE_OK) ? "valid" : "invalid",
+			mode->base.id, mode->name, mode->vrefresh, mode->clock,
+			mode->hdisplay, mode->hsync_start,
+			mode->hsync_end, mode->htotal,
+			mode->vdisplay, mode->vsync_start,
+			mode->vsync_end, mode->vtotal, mode->type, mode->flags);
+
+	return ret;
+}
+
+struct drm_encoder *omap_connector_attached_encoder(
+		struct drm_connector *connector)
+{
+	struct omap_connector *omap_connector = to_omap_connector(connector);
+	return omap_connector->encoder;
+}
+
+static const struct drm_connector_funcs omap_connector_funcs = {
+	.dpms = drm_helper_connector_dpms,
+	.detect = omap_connector_detect,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.destroy = omap_connector_destroy,
+};
+
+static const struct drm_connector_helper_funcs omap_connector_helper_funcs = {
+	.get_modes = omap_connector_get_modes,
+	.mode_valid = omap_connector_mode_valid,
+	.best_encoder = omap_connector_attached_encoder,
+};
+
+/* flush an area of the framebuffer (in case of manual update display that
+ * is not automatically flushed)
+ */
+void omap_connector_flush(struct drm_connector *connector,
+		int x, int y, int w, int h)
+{
+	struct omap_connector *omap_connector = to_omap_connector(connector);
+
+	/* TODO: enable when supported in dss */
+	VERB("%s: %d,%d, %dx%d", omap_connector->dssdev->name, x, y, w, h);
+}
+
+/* initialize connector */
+struct drm_connector *omap_connector_init(struct drm_device *dev,
+		int connector_type, struct omap_dss_device *dssdev,
+		struct drm_encoder *encoder)
+{
+	struct drm_connector *connector = NULL;
+	struct omap_connector *omap_connector;
+
+	DBG("%s", dssdev->name);
+
+	omap_dss_get_device(dssdev);
+
+	omap_connector = kzalloc(sizeof(struct omap_connector), GFP_KERNEL);
+	if (!omap_connector) {
+		dev_err(dev->dev, "could not allocate connector\n");
+		goto fail;
+	}
+
+	omap_connector->dssdev = dssdev;
+	omap_connector->encoder = encoder;
+
+	connector = &omap_connector->base;
+
+	drm_connector_init(dev, connector, &omap_connector_funcs,
+				connector_type);
+	drm_connector_helper_add(connector, &omap_connector_helper_funcs);
+
+#if 0 /* enable when dss2 supports hotplug */
+	if (dssdev->caps & OMAP_DSS_DISPLAY_CAP_HPD)
+		connector->polled = 0;
+	else
+#endif
+		connector->polled = DRM_CONNECTOR_POLL_CONNECT |
+				DRM_CONNECTOR_POLL_DISCONNECT;
+
+	connector->interlace_allowed = 1;
+	connector->doublescan_allowed = 0;
+
+	drm_sysfs_connector_add(connector);
+
+	return connector;
+
+fail:
+	if (connector)
+		omap_connector_destroy(connector);
+
+	return NULL;
+}
diff --git a/drivers/gpu/drm/omapdrm/omap_crtc.c b/drivers/gpu/drm/omapdrm/omap_crtc.c
new file mode 100644
index 0000000..2b97cf9
--- /dev/null
+++ b/drivers/gpu/drm/omapdrm/omap_crtc.c
@@ -0,0 +1,659 @@
+/*
+ * drivers/gpu/drm/omapdrm/omap_crtc.c
+ *
+ * Copyright (C) 2011 Texas Instruments
+ * Author: Rob Clark <rob@ti.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "omap_drv.h"
+
+#include <drm/drm_mode.h>
+#include "drm_crtc.h"
+#include "drm_crtc_helper.h"
+
+#define to_omap_crtc(x) container_of(x, struct omap_crtc, base)
+
+struct omap_crtc {
+	struct drm_crtc base;
+	struct drm_plane *plane;
+
+	const char *name;
+	int pipe;
+	enum omap_channel channel;
+	struct omap_overlay_manager_info info;
+
+	/*
+	 * Temporary: eventually this will go away, but it is needed
+	 * for now to keep the output's happy.  (They only need
+	 * mgr->id.)  Eventually this will be replaced w/ something
+	 * more common-panel-framework-y
+	 */
+	struct omap_overlay_manager mgr;
+
+	struct omap_video_timings timings;
+	bool enabled;
+	bool full_update;
+
+	struct omap_drm_apply apply;
+
+	struct omap_drm_irq apply_irq;
+	struct omap_drm_irq error_irq;
+
+	/* list of in-progress apply's: */
+	struct list_head pending_applies;
+
+	/* list of queued apply's: */
+	struct list_head queued_applies;
+
+	/* for handling queued and in-progress applies: */
+	struct work_struct apply_work;
+
+	/* if there is a pending flip, these will be non-null: */
+	struct drm_pending_vblank_event *event;
+	struct drm_framebuffer *old_fb;
+
+	/* for handling page flips without caring about what
+	 * the callback is called from.  Possibly we should just
+	 * make omap_gem always call the cb from the worker so
+	 * we don't have to care about this..
+	 *
+	 * XXX maybe fold into apply_work??
+	 */
+	struct work_struct page_flip_work;
+};
+
+/*
+ * Manager-ops, callbacks from output when they need to configure
+ * the upstream part of the video pipe.
+ *
+ * Most of these we can ignore until we add support for command-mode
+ * panels.. for video-mode the crtc-helpers already do an adequate
+ * job of sequencing the setup of the video pipe in the proper order
+ */
+
+/* we can probably ignore these until we support command-mode panels: */
+static void omap_crtc_start_update(struct omap_overlay_manager *mgr)
+{
+}
+
+static int omap_crtc_enable(struct omap_overlay_manager *mgr)
+{
+	return 0;
+}
+
+static void omap_crtc_disable(struct omap_overlay_manager *mgr)
+{
+}
+
+static void omap_crtc_set_timings(struct omap_overlay_manager *mgr,
+		const struct omap_video_timings *timings)
+{
+	struct omap_crtc *omap_crtc = container_of(mgr, struct omap_crtc, mgr);
+	DBG("%s", omap_crtc->name);
+	omap_crtc->timings = *timings;
+	omap_crtc->full_update = true;
+}
+
+static void omap_crtc_set_lcd_config(struct omap_overlay_manager *mgr,
+		const struct dss_lcd_mgr_config *config)
+{
+	struct omap_crtc *omap_crtc = container_of(mgr, struct omap_crtc, mgr);
+	DBG("%s", omap_crtc->name);
+	dispc_mgr_set_lcd_config(omap_crtc->channel, config);
+}
+
+static int omap_crtc_register_framedone_handler(
+		struct omap_overlay_manager *mgr,
+		void (*handler)(void *), void *data)
+{
+	return 0;
+}
+
+static void omap_crtc_unregister_framedone_handler(
+		struct omap_overlay_manager *mgr,
+		void (*handler)(void *), void *data)
+{
+}
+
+static const struct dss_mgr_ops mgr_ops = {
+		.start_update = omap_crtc_start_update,
+		.enable = omap_crtc_enable,
+		.disable = omap_crtc_disable,
+		.set_timings = omap_crtc_set_timings,
+		.set_lcd_config = omap_crtc_set_lcd_config,
+		.register_framedone_handler = omap_crtc_register_framedone_handler,
+		.unregister_framedone_handler = omap_crtc_unregister_framedone_handler,
+};
+
+/*
+ * CRTC funcs:
+ */
+
+static void omap_crtc_destroy(struct drm_crtc *crtc)
+{
+	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
+
+	DBG("%s", omap_crtc->name);
+
+	WARN_ON(omap_crtc->apply_irq.registered);
+	omap_irq_unregister(crtc->dev, &omap_crtc->error_irq);
+
+	omap_crtc->plane->funcs->destroy(omap_crtc->plane);
+	drm_crtc_cleanup(crtc);
+
+	kfree(omap_crtc);
+}
+
+static void omap_crtc_dpms(struct drm_crtc *crtc, int mode)
+{
+	struct omap_drm_private *priv = crtc->dev->dev_private;
+	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
+	bool enabled = (mode == DRM_MODE_DPMS_ON);
+	int i;
+
+	DBG("%s: %d", omap_crtc->name, mode);
+
+	if (enabled != omap_crtc->enabled) {
+		omap_crtc->enabled = enabled;
+		omap_crtc->full_update = true;
+		omap_crtc_apply(crtc, &omap_crtc->apply);
+
+		/* also enable our private plane: */
+		WARN_ON(omap_plane_dpms(omap_crtc->plane, mode));
+
+		/* and any attached overlay planes: */
+		for (i = 0; i < priv->num_planes; i++) {
+			struct drm_plane *plane = priv->planes[i];
+			if (plane->crtc == crtc)
+				WARN_ON(omap_plane_dpms(plane, mode));
+		}
+	}
+}
+
+static bool omap_crtc_mode_fixup(struct drm_crtc *crtc,
+		const struct drm_display_mode *mode,
+		struct drm_display_mode *adjusted_mode)
+{
+	return true;
+}
+
+static int omap_crtc_mode_set(struct drm_crtc *crtc,
+		struct drm_display_mode *mode,
+		struct drm_display_mode *adjusted_mode,
+		int x, int y,
+		struct drm_framebuffer *old_fb)
+{
+	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
+
+	mode = adjusted_mode;
+
+	DBG("%s: set mode: %d:\"%s\" %d %d %d %d %d %d %d %d %d %d 0x%x 0x%x",
+			omap_crtc->name, mode->base.id, mode->name,
+			mode->vrefresh, mode->clock,
+			mode->hdisplay, mode->hsync_start,
+			mode->hsync_end, mode->htotal,
+			mode->vdisplay, mode->vsync_start,
+			mode->vsync_end, mode->vtotal,
+			mode->type, mode->flags);
+
+	copy_timings_drm_to_omap(&omap_crtc->timings, mode);
+	omap_crtc->full_update = true;
+
+	return omap_plane_mode_set(omap_crtc->plane, crtc, crtc->fb,
+			0, 0, mode->hdisplay, mode->vdisplay,
+			x << 16, y << 16,
+			mode->hdisplay << 16, mode->vdisplay << 16,
+			NULL, NULL);
+}
+
+static void omap_crtc_prepare(struct drm_crtc *crtc)
+{
+	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
+	DBG("%s", omap_crtc->name);
+	omap_crtc_dpms(crtc, DRM_MODE_DPMS_OFF);
+}
+
+static void omap_crtc_commit(struct drm_crtc *crtc)
+{
+	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
+	DBG("%s", omap_crtc->name);
+	omap_crtc_dpms(crtc, DRM_MODE_DPMS_ON);
+}
+
+static int omap_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
+		struct drm_framebuffer *old_fb)
+{
+	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
+	struct drm_plane *plane = omap_crtc->plane;
+	struct drm_display_mode *mode = &crtc->mode;
+
+	return omap_plane_mode_set(plane, crtc, crtc->fb,
+			0, 0, mode->hdisplay, mode->vdisplay,
+			x << 16, y << 16,
+			mode->hdisplay << 16, mode->vdisplay << 16,
+			NULL, NULL);
+}
+
+static void omap_crtc_load_lut(struct drm_crtc *crtc)
+{
+}
+
+static void vblank_cb(void *arg)
+{
+	struct drm_crtc *crtc = arg;
+	struct drm_device *dev = crtc->dev;
+	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
+	unsigned long flags;
+
+	spin_lock_irqsave(&dev->event_lock, flags);
+
+	/* wakeup userspace */
+	if (omap_crtc->event)
+		drm_send_vblank_event(dev, omap_crtc->pipe, omap_crtc->event);
+
+	omap_crtc->event = NULL;
+	omap_crtc->old_fb = NULL;
+
+	spin_unlock_irqrestore(&dev->event_lock, flags);
+}
+
+static void page_flip_worker(struct work_struct *work)
+{
+	struct omap_crtc *omap_crtc =
+			container_of(work, struct omap_crtc, page_flip_work);
+	struct drm_crtc *crtc = &omap_crtc->base;
+	struct drm_device *dev = crtc->dev;
+	struct drm_display_mode *mode = &crtc->mode;
+	struct drm_gem_object *bo;
+
+	drm_modeset_lock_all(dev);
+	omap_plane_mode_set(omap_crtc->plane, crtc, crtc->fb,
+			0, 0, mode->hdisplay, mode->vdisplay,
+			crtc->x << 16, crtc->y << 16,
+			mode->hdisplay << 16, mode->vdisplay << 16,
+			vblank_cb, crtc);
+	drm_modeset_unlock_all(dev);
+
+	bo = omap_framebuffer_bo(crtc->fb, 0);
+	drm_gem_object_unreference_unlocked(bo);
+}
+
+static void page_flip_cb(void *arg)
+{
+	struct drm_crtc *crtc = arg;
+	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
+	struct omap_drm_private *priv = crtc->dev->dev_private;
+
+	/* avoid assumptions about what ctxt we are called from: */
+	queue_work(priv->wq, &omap_crtc->page_flip_work);
+}
+
+static int omap_crtc_page_flip_locked(struct drm_crtc *crtc,
+		 struct drm_framebuffer *fb,
+		 struct drm_pending_vblank_event *event)
+{
+	struct drm_device *dev = crtc->dev;
+	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
+	struct drm_gem_object *bo;
+
+	DBG("%d -> %d (event=%p)", crtc->fb ? crtc->fb->base.id : -1,
+			fb->base.id, event);
+
+	if (omap_crtc->old_fb) {
+		dev_err(dev->dev, "already a pending flip\n");
+		return -EINVAL;
+	}
+
+	omap_crtc->event = event;
+	crtc->fb = fb;
+
+	/*
+	 * Hold a reference temporarily until the crtc is updated
+	 * and takes the reference to the bo.  This avoids it
+	 * getting freed from under us:
+	 */
+	bo = omap_framebuffer_bo(fb, 0);
+	drm_gem_object_reference(bo);
+
+	omap_gem_op_async(bo, OMAP_GEM_READ, page_flip_cb, crtc);
+
+	return 0;
+}
+
+static int omap_crtc_set_property(struct drm_crtc *crtc,
+		struct drm_property *property, uint64_t val)
+{
+	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
+	struct omap_drm_private *priv = crtc->dev->dev_private;
+
+	if (property == priv->rotation_prop) {
+		crtc->invert_dimensions =
+				!!(val & ((1LL << DRM_ROTATE_90) | (1LL << DRM_ROTATE_270)));
+	}
+
+	return omap_plane_set_property(omap_crtc->plane, property, val);
+}
+
+static const struct drm_crtc_funcs omap_crtc_funcs = {
+	.set_config = drm_crtc_helper_set_config,
+	.destroy = omap_crtc_destroy,
+	.page_flip = omap_crtc_page_flip_locked,
+	.set_property = omap_crtc_set_property,
+};
+
+static const struct drm_crtc_helper_funcs omap_crtc_helper_funcs = {
+	.dpms = omap_crtc_dpms,
+	.mode_fixup = omap_crtc_mode_fixup,
+	.mode_set = omap_crtc_mode_set,
+	.prepare = omap_crtc_prepare,
+	.commit = omap_crtc_commit,
+	.mode_set_base = omap_crtc_mode_set_base,
+	.load_lut = omap_crtc_load_lut,
+};
+
+const struct omap_video_timings *omap_crtc_timings(struct drm_crtc *crtc)
+{
+	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
+	return &omap_crtc->timings;
+}
+
+enum omap_channel omap_crtc_channel(struct drm_crtc *crtc)
+{
+	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
+	return omap_crtc->channel;
+}
+
+static void omap_crtc_error_irq(struct omap_drm_irq *irq, uint32_t irqstatus)
+{
+	struct omap_crtc *omap_crtc =
+			container_of(irq, struct omap_crtc, error_irq);
+	struct drm_crtc *crtc = &omap_crtc->base;
+	DRM_ERROR("%s: errors: %08x\n", omap_crtc->name, irqstatus);
+	/* avoid getting in a flood, unregister the irq until next vblank */
+	omap_irq_unregister(crtc->dev, &omap_crtc->error_irq);
+}
+
+static void omap_crtc_apply_irq(struct omap_drm_irq *irq, uint32_t irqstatus)
+{
+	struct omap_crtc *omap_crtc =
+			container_of(irq, struct omap_crtc, apply_irq);
+	struct drm_crtc *crtc = &omap_crtc->base;
+
+	if (!omap_crtc->error_irq.registered)
+		omap_irq_register(crtc->dev, &omap_crtc->error_irq);
+
+	if (!dispc_mgr_go_busy(omap_crtc->channel)) {
+		struct omap_drm_private *priv =
+				crtc->dev->dev_private;
+		DBG("%s: apply done", omap_crtc->name);
+		omap_irq_unregister(crtc->dev, &omap_crtc->apply_irq);
+		queue_work(priv->wq, &omap_crtc->apply_work);
+	}
+}
+
+static void apply_worker(struct work_struct *work)
+{
+	struct omap_crtc *omap_crtc =
+			container_of(work, struct omap_crtc, apply_work);
+	struct drm_crtc *crtc = &omap_crtc->base;
+	struct drm_device *dev = crtc->dev;
+	struct omap_drm_apply *apply, *n;
+	bool need_apply;
+
+	/*
+	 * Synchronize everything on mode_config.mutex, to keep
+	 * the callbacks and list modification all serialized
+	 * with respect to modesetting ioctls from userspace.
+	 */
+	drm_modeset_lock_all(dev);
+	dispc_runtime_get();
+
+	/*
+	 * If we are still pending a previous update, wait.. when the
+	 * pending update completes, we get kicked again.
+	 */
+	if (omap_crtc->apply_irq.registered)
+		goto out;
+
+	/* finish up previous apply's: */
+	list_for_each_entry_safe(apply, n,
+			&omap_crtc->pending_applies, pending_node) {
+		apply->post_apply(apply);
+		list_del(&apply->pending_node);
+	}
+
+	need_apply = !list_empty(&omap_crtc->queued_applies);
+
+	/* then handle the next round of of queued apply's: */
+	list_for_each_entry_safe(apply, n,
+			&omap_crtc->queued_applies, queued_node) {
+		apply->pre_apply(apply);
+		list_del(&apply->queued_node);
+		apply->queued = false;
+		list_add_tail(&apply->pending_node,
+				&omap_crtc->pending_applies);
+	}
+
+	if (need_apply) {
+		enum omap_channel channel = omap_crtc->channel;
+
+		DBG("%s: GO", omap_crtc->name);
+
+		if (dispc_mgr_is_enabled(channel)) {
+			omap_irq_register(dev, &omap_crtc->apply_irq);
+			dispc_mgr_go(channel);
+		} else {
+			struct omap_drm_private *priv = dev->dev_private;
+			queue_work(priv->wq, &omap_crtc->apply_work);
+		}
+	}
+
+out:
+	dispc_runtime_put();
+	drm_modeset_unlock_all(dev);
+}
+
+int omap_crtc_apply(struct drm_crtc *crtc,
+		struct omap_drm_apply *apply)
+{
+	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
+	struct drm_device *dev = crtc->dev;
+
+	WARN_ON(!mutex_is_locked(&dev->mode_config.mutex));
+
+	/* no need to queue it again if it is already queued: */
+	if (apply->queued)
+		return 0;
+
+	apply->queued = true;
+	list_add_tail(&apply->queued_node, &omap_crtc->queued_applies);
+
+	/*
+	 * If there are no currently pending updates, then go ahead and
+	 * kick the worker immediately, otherwise it will run again when
+	 * the current update finishes.
+	 */
+	if (list_empty(&omap_crtc->pending_applies)) {
+		struct omap_drm_private *priv = crtc->dev->dev_private;
+		queue_work(priv->wq, &omap_crtc->apply_work);
+	}
+
+	return 0;
+}
+
+/* called only from apply */
+static void set_enabled(struct drm_crtc *crtc, bool enable)
+{
+	struct drm_device *dev = crtc->dev;
+	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
+	enum omap_channel channel = omap_crtc->channel;
+	struct omap_irq_wait *wait = NULL;
+
+	if (dispc_mgr_is_enabled(channel) == enable)
+		return;
+
+	/* ignore sync-lost irqs during enable/disable */
+	omap_irq_unregister(crtc->dev, &omap_crtc->error_irq);
+
+	if (dispc_mgr_get_framedone_irq(channel)) {
+		if (!enable) {
+			wait = omap_irq_wait_init(dev,
+					dispc_mgr_get_framedone_irq(channel), 1);
+		}
+	} else {
+		/*
+		 * When we disable digit output, we need to wait until fields
+		 * are done.  Otherwise the DSS is still working, and turning
+		 * off the clocks prevents DSS from going to OFF mode. And when
+		 * enabling, we need to wait for the extra sync losts
+		 */
+		wait = omap_irq_wait_init(dev,
+				dispc_mgr_get_vsync_irq(channel), 2);
+	}
+
+	dispc_mgr_enable(channel, enable);
+
+	if (wait) {
+		int ret = omap_irq_wait(dev, wait, msecs_to_jiffies(100));
+		if (ret) {
+			dev_err(dev->dev, "%s: timeout waiting for %s\n",
+					omap_crtc->name, enable ? "enable" : "disable");
+		}
+	}
+
+	omap_irq_register(crtc->dev, &omap_crtc->error_irq);
+}
+
+static void omap_crtc_pre_apply(struct omap_drm_apply *apply)
+{
+	struct omap_crtc *omap_crtc =
+			container_of(apply, struct omap_crtc, apply);
+	struct drm_crtc *crtc = &omap_crtc->base;
+	struct drm_encoder *encoder = NULL;
+
+	DBG("%s: enabled=%d, full=%d", omap_crtc->name,
+			omap_crtc->enabled, omap_crtc->full_update);
+
+	if (omap_crtc->full_update) {
+		struct omap_drm_private *priv = crtc->dev->dev_private;
+		int i;
+		for (i = 0; i < priv->num_encoders; i++) {
+			if (priv->encoders[i]->crtc == crtc) {
+				encoder = priv->encoders[i];
+				break;
+			}
+		}
+	}
+
+	if (!omap_crtc->enabled) {
+		set_enabled(&omap_crtc->base, false);
+		if (encoder)
+			omap_encoder_set_enabled(encoder, false);
+	} else {
+		if (encoder) {
+			omap_encoder_set_enabled(encoder, false);
+			omap_encoder_update(encoder, &omap_crtc->mgr,
+					&omap_crtc->timings);
+			omap_encoder_set_enabled(encoder, true);
+			omap_crtc->full_update = false;
+		}
+
+		dispc_mgr_setup(omap_crtc->channel, &omap_crtc->info);
+		dispc_mgr_set_timings(omap_crtc->channel,
+				&omap_crtc->timings);
+		set_enabled(&omap_crtc->base, true);
+	}
+
+	omap_crtc->full_update = false;
+}
+
+static void omap_crtc_post_apply(struct omap_drm_apply *apply)
+{
+	/* nothing needed for post-apply */
+}
+
+static const char *channel_names[] = {
+		[OMAP_DSS_CHANNEL_LCD] = "lcd",
+		[OMAP_DSS_CHANNEL_DIGIT] = "tv",
+		[OMAP_DSS_CHANNEL_LCD2] = "lcd2",
+};
+
+/* initialize crtc */
+struct drm_crtc *omap_crtc_init(struct drm_device *dev,
+		struct drm_plane *plane, enum omap_channel channel, int id)
+{
+	struct drm_crtc *crtc = NULL;
+	struct omap_crtc *omap_crtc;
+	struct omap_overlay_manager_info *info;
+
+	DBG("%s", channel_names[channel]);
+
+	omap_crtc = kzalloc(sizeof(*omap_crtc), GFP_KERNEL);
+
+	if (!omap_crtc) {
+		dev_err(dev->dev, "could not allocate CRTC\n");
+		goto fail;
+	}
+
+	crtc = &omap_crtc->base;
+
+	INIT_WORK(&omap_crtc->page_flip_work, page_flip_worker);
+	INIT_WORK(&omap_crtc->apply_work, apply_worker);
+
+	INIT_LIST_HEAD(&omap_crtc->pending_applies);
+	INIT_LIST_HEAD(&omap_crtc->queued_applies);
+
+	omap_crtc->apply.pre_apply  = omap_crtc_pre_apply;
+	omap_crtc->apply.post_apply = omap_crtc_post_apply;
+
+	omap_crtc->apply_irq.irqmask = pipe2vbl(id);
+	omap_crtc->apply_irq.irq = omap_crtc_apply_irq;
+
+	omap_crtc->error_irq.irqmask =
+			dispc_mgr_get_sync_lost_irq(channel);
+	omap_crtc->error_irq.irq = omap_crtc_error_irq;
+	omap_irq_register(dev, &omap_crtc->error_irq);
+
+	omap_crtc->channel = channel;
+	omap_crtc->plane = plane;
+	omap_crtc->plane->crtc = crtc;
+	omap_crtc->name = channel_names[channel];
+	omap_crtc->pipe = id;
+
+	/* temporary: */
+	omap_crtc->mgr.id = channel;
+
+	dss_install_mgr_ops(&mgr_ops);
+
+	/* TODO: fix hard-coded setup.. add properties! */
+	info = &omap_crtc->info;
+	info->default_color = 0x00000000;
+	info->trans_key = 0x00000000;
+	info->trans_key_type = OMAP_DSS_COLOR_KEY_GFX_DST;
+	info->trans_enabled = false;
+
+	drm_crtc_init(dev, crtc, &omap_crtc_funcs);
+	drm_crtc_helper_add(crtc, &omap_crtc_helper_funcs);
+
+	omap_plane_install_properties(omap_crtc->plane, &crtc->base);
+
+	return crtc;
+
+fail:
+	if (crtc)
+		omap_crtc_destroy(crtc);
+
+	return NULL;
+}
diff --git a/drivers/gpu/drm/omapdrm/omap_debugfs.c b/drivers/gpu/drm/omapdrm/omap_debugfs.c
new file mode 100644
index 0000000..c0aa40f
--- /dev/null
+++ b/drivers/gpu/drm/omapdrm/omap_debugfs.c
@@ -0,0 +1,139 @@
+/*
+ * drivers/gpu/drm/omapdrm/omap_debugfs.c
+ *
+ * Copyright (C) 2011 Texas Instruments
+ * Author: Rob Clark <rob.clark@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "omap_drv.h"
+#include "omap_dmm_tiler.h"
+
+#include "drm_fb_helper.h"
+
+
+#ifdef CONFIG_DEBUG_FS
+
+static int gem_show(struct seq_file *m, void *arg)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct omap_drm_private *priv = dev->dev_private;
+	int ret;
+
+	ret = mutex_lock_interruptible(&dev->struct_mutex);
+	if (ret)
+		return ret;
+
+	seq_printf(m, "All Objects:\n");
+	omap_gem_describe_objects(&priv->obj_list, m);
+
+	mutex_unlock(&dev->struct_mutex);
+
+	return 0;
+}
+
+static int mm_show(struct seq_file *m, void *arg)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	return drm_mm_dump_table(m, dev->mm_private);
+}
+
+static int fb_show(struct seq_file *m, void *arg)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct omap_drm_private *priv = dev->dev_private;
+	struct drm_framebuffer *fb;
+	int ret;
+
+	ret = mutex_lock_interruptible(&dev->mode_config.mutex);
+	if (ret)
+		return ret;
+
+	ret = mutex_lock_interruptible(&dev->struct_mutex);
+	if (ret) {
+		mutex_unlock(&dev->mode_config.mutex);
+		return ret;
+	}
+
+	seq_printf(m, "fbcon ");
+	omap_framebuffer_describe(priv->fbdev->fb, m);
+
+	mutex_lock(&dev->mode_config.fb_lock);
+	list_for_each_entry(fb, &dev->mode_config.fb_list, head) {
+		if (fb == priv->fbdev->fb)
+			continue;
+
+		seq_printf(m, "user ");
+		omap_framebuffer_describe(fb, m);
+	}
+	mutex_unlock(&dev->mode_config.fb_lock);
+
+	mutex_unlock(&dev->struct_mutex);
+	mutex_unlock(&dev->mode_config.mutex);
+
+	return 0;
+}
+
+/* list of debufs files that are applicable to all devices */
+static struct drm_info_list omap_debugfs_list[] = {
+	{"gem", gem_show, 0},
+	{"mm", mm_show, 0},
+	{"fb", fb_show, 0},
+};
+
+/* list of debugfs files that are specific to devices with dmm/tiler */
+static struct drm_info_list omap_dmm_debugfs_list[] = {
+	{"tiler_map", tiler_map_show, 0},
+};
+
+int omap_debugfs_init(struct drm_minor *minor)
+{
+	struct drm_device *dev = minor->dev;
+	int ret;
+
+	ret = drm_debugfs_create_files(omap_debugfs_list,
+			ARRAY_SIZE(omap_debugfs_list),
+			minor->debugfs_root, minor);
+
+	if (ret) {
+		dev_err(dev->dev, "could not install omap_debugfs_list\n");
+		return ret;
+	}
+
+	if (dmm_is_available())
+		ret = drm_debugfs_create_files(omap_dmm_debugfs_list,
+				ARRAY_SIZE(omap_dmm_debugfs_list),
+				minor->debugfs_root, minor);
+
+	if (ret) {
+		dev_err(dev->dev, "could not install omap_dmm_debugfs_list\n");
+		return ret;
+	}
+
+	return ret;
+}
+
+void omap_debugfs_cleanup(struct drm_minor *minor)
+{
+	drm_debugfs_remove_files(omap_debugfs_list,
+			ARRAY_SIZE(omap_debugfs_list), minor);
+	if (dmm_is_available())
+		drm_debugfs_remove_files(omap_dmm_debugfs_list,
+				ARRAY_SIZE(omap_dmm_debugfs_list), minor);
+}
+
+#endif
diff --git a/drivers/gpu/drm/omapdrm/omap_dmm_priv.h b/drivers/gpu/drm/omapdrm/omap_dmm_priv.h
new file mode 100644
index 0000000..58bcd6a
--- /dev/null
+++ b/drivers/gpu/drm/omapdrm/omap_dmm_priv.h
@@ -0,0 +1,188 @@
+/*
+ *
+ * Copyright (C) 2011 Texas Instruments Incorporated - http://www.ti.com/
+ * Author: Rob Clark <rob@ti.com>
+ *         Andy Gross <andy.gross@ti.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation version 2.
+ *
+ * This program is distributed "as is" WITHOUT ANY WARRANTY of any
+ * kind, whether express or implied; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+#ifndef OMAP_DMM_PRIV_H
+#define OMAP_DMM_PRIV_H
+
+#define DMM_REVISION          0x000
+#define DMM_HWINFO            0x004
+#define DMM_LISA_HWINFO       0x008
+#define DMM_DMM_SYSCONFIG     0x010
+#define DMM_LISA_LOCK         0x01C
+#define DMM_LISA_MAP__0       0x040
+#define DMM_LISA_MAP__1       0x044
+#define DMM_TILER_HWINFO      0x208
+#define DMM_TILER_OR__0       0x220
+#define DMM_TILER_OR__1       0x224
+#define DMM_PAT_HWINFO        0x408
+#define DMM_PAT_GEOMETRY      0x40C
+#define DMM_PAT_CONFIG        0x410
+#define DMM_PAT_VIEW__0       0x420
+#define DMM_PAT_VIEW__1       0x424
+#define DMM_PAT_VIEW_MAP__0   0x440
+#define DMM_PAT_VIEW_MAP_BASE 0x460
+#define DMM_PAT_IRQ_EOI       0x478
+#define DMM_PAT_IRQSTATUS_RAW 0x480
+#define DMM_PAT_IRQSTATUS     0x490
+#define DMM_PAT_IRQENABLE_SET 0x4A0
+#define DMM_PAT_IRQENABLE_CLR 0x4B0
+#define DMM_PAT_STATUS__0     0x4C0
+#define DMM_PAT_STATUS__1     0x4C4
+#define DMM_PAT_STATUS__2     0x4C8
+#define DMM_PAT_STATUS__3     0x4CC
+#define DMM_PAT_DESCR__0      0x500
+#define DMM_PAT_DESCR__1      0x510
+#define DMM_PAT_DESCR__2      0x520
+#define DMM_PAT_DESCR__3      0x530
+#define DMM_PEG_HWINFO        0x608
+#define DMM_PEG_PRIO          0x620
+#define DMM_PEG_PRIO_PAT      0x640
+
+#define DMM_IRQSTAT_DST			(1<<0)
+#define DMM_IRQSTAT_LST			(1<<1)
+#define DMM_IRQSTAT_ERR_INV_DSC		(1<<2)
+#define DMM_IRQSTAT_ERR_INV_DATA	(1<<3)
+#define DMM_IRQSTAT_ERR_UPD_AREA	(1<<4)
+#define DMM_IRQSTAT_ERR_UPD_CTRL	(1<<5)
+#define DMM_IRQSTAT_ERR_UPD_DATA	(1<<6)
+#define DMM_IRQSTAT_ERR_LUT_MISS	(1<<7)
+
+#define DMM_IRQSTAT_ERR_MASK	(DMM_IRQ_STAT_ERR_INV_DSC | \
+				DMM_IRQ_STAT_ERR_INV_DATA | \
+				DMM_IRQ_STAT_ERR_UPD_AREA | \
+				DMM_IRQ_STAT_ERR_UPD_CTRL | \
+				DMM_IRQ_STAT_ERR_UPD_DATA | \
+				DMM_IRQ_STAT_ERR_LUT_MISS)
+
+#define DMM_PATSTATUS_READY		(1<<0)
+#define DMM_PATSTATUS_VALID		(1<<1)
+#define DMM_PATSTATUS_RUN		(1<<2)
+#define DMM_PATSTATUS_DONE		(1<<3)
+#define DMM_PATSTATUS_LINKED		(1<<4)
+#define DMM_PATSTATUS_BYPASSED		(1<<7)
+#define DMM_PATSTATUS_ERR_INV_DESCR	(1<<10)
+#define DMM_PATSTATUS_ERR_INV_DATA	(1<<11)
+#define DMM_PATSTATUS_ERR_UPD_AREA	(1<<12)
+#define DMM_PATSTATUS_ERR_UPD_CTRL	(1<<13)
+#define DMM_PATSTATUS_ERR_UPD_DATA	(1<<14)
+#define DMM_PATSTATUS_ERR_ACCESS	(1<<15)
+
+/* note: don't treat DMM_PATSTATUS_ERR_ACCESS as an error */
+#define DMM_PATSTATUS_ERR	(DMM_PATSTATUS_ERR_INV_DESCR | \
+				DMM_PATSTATUS_ERR_INV_DATA | \
+				DMM_PATSTATUS_ERR_UPD_AREA | \
+				DMM_PATSTATUS_ERR_UPD_CTRL | \
+				DMM_PATSTATUS_ERR_UPD_DATA)
+
+
+
+enum {
+	PAT_STATUS,
+	PAT_DESCR
+};
+
+struct pat_ctrl {
+	u32 start:4;
+	u32 dir:4;
+	u32 lut_id:8;
+	u32 sync:12;
+	u32 ini:4;
+};
+
+struct pat {
+	uint32_t next_pa;
+	struct pat_area area;
+	struct pat_ctrl ctrl;
+	uint32_t data_pa;
+};
+
+#define DMM_FIXED_RETRY_COUNT 1000
+
+/* create refill buffer big enough to refill all slots, plus 3 descriptors..
+ * 3 descriptors is probably the worst-case for # of 2d-slices in a 1d area,
+ * but I guess you don't hit that worst case at the same time as full area
+ * refill
+ */
+#define DESCR_SIZE 128
+#define REFILL_BUFFER_SIZE ((4 * 128 * 256) + (3 * DESCR_SIZE))
+
+/* For OMAP5, a fixed offset is added to all Y coordinates for 1D buffers.
+ * This is used in programming to address the upper portion of the LUT
+*/
+#define OMAP5_LUT_OFFSET       128
+
+struct dmm;
+
+struct dmm_txn {
+	void *engine_handle;
+	struct tcm *tcm;
+
+	uint8_t *current_va;
+	dma_addr_t current_pa;
+
+	struct pat *last_pat;
+};
+
+struct refill_engine {
+	int id;
+	struct dmm *dmm;
+	struct tcm *tcm;
+
+	uint8_t *refill_va;
+	dma_addr_t refill_pa;
+
+	/* only one trans per engine for now */
+	struct dmm_txn txn;
+
+	bool async;
+
+	wait_queue_head_t wait_for_refill;
+
+	struct list_head idle_node;
+};
+
+struct dmm {
+	struct device *dev;
+	void __iomem *base;
+	int irq;
+
+	struct page *dummy_page;
+	dma_addr_t dummy_pa;
+
+	void *refill_va;
+	dma_addr_t refill_pa;
+
+	/* refill engines */
+	wait_queue_head_t engine_queue;
+	struct list_head idle_head;
+	struct refill_engine *engines;
+	int num_engines;
+	atomic_t engine_counter;
+
+	/* container information */
+	int container_width;
+	int container_height;
+	int lut_width;
+	int lut_height;
+	int num_lut;
+
+	/* array of LUT - TCM containers */
+	struct tcm **tcm;
+
+	/* allocation list and lock */
+	struct list_head alloc_head;
+};
+
+#endif
diff --git a/drivers/gpu/drm/omapdrm/omap_dmm_tiler.c b/drivers/gpu/drm/omapdrm/omap_dmm_tiler.c
new file mode 100644
index 0000000..3910215
--- /dev/null
+++ b/drivers/gpu/drm/omapdrm/omap_dmm_tiler.c
@@ -0,0 +1,991 @@
+/*
+ * DMM IOMMU driver support functions for TI OMAP processors.
+ *
+ * Author: Rob Clark <rob@ti.com>
+ *         Andy Gross <andy.gross@ti.com>
+ *
+ * Copyright (C) 2011 Texas Instruments Incorporated - http://www.ti.com/
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation version 2.
+ *
+ * This program is distributed "as is" WITHOUT ANY WARRANTY of any
+ * kind, whether express or implied; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/platform_device.h> /* platform_device() */
+#include <linux/errno.h>
+#include <linux/sched.h>
+#include <linux/wait.h>
+#include <linux/interrupt.h>
+#include <linux/dma-mapping.h>
+#include <linux/slab.h>
+#include <linux/vmalloc.h>
+#include <linux/delay.h>
+#include <linux/mm.h>
+#include <linux/time.h>
+#include <linux/list.h>
+
+#include "omap_dmm_tiler.h"
+#include "omap_dmm_priv.h"
+
+#define DMM_DRIVER_NAME "dmm"
+
+/* mappings for associating views to luts */
+static struct tcm *containers[TILFMT_NFORMATS];
+static struct dmm *omap_dmm;
+
+/* global spinlock for protecting lists */
+static DEFINE_SPINLOCK(list_lock);
+
+/* Geometry table */
+#define GEOM(xshift, yshift, bytes_per_pixel) { \
+		.x_shft = (xshift), \
+		.y_shft = (yshift), \
+		.cpp    = (bytes_per_pixel), \
+		.slot_w = 1 << (SLOT_WIDTH_BITS - (xshift)), \
+		.slot_h = 1 << (SLOT_HEIGHT_BITS - (yshift)), \
+	}
+
+static const struct {
+	uint32_t x_shft;	/* unused X-bits (as part of bpp) */
+	uint32_t y_shft;	/* unused Y-bits (as part of bpp) */
+	uint32_t cpp;		/* bytes/chars per pixel */
+	uint32_t slot_w;	/* width of each slot (in pixels) */
+	uint32_t slot_h;	/* height of each slot (in pixels) */
+} geom[TILFMT_NFORMATS] = {
+		[TILFMT_8BIT]  = GEOM(0, 0, 1),
+		[TILFMT_16BIT] = GEOM(0, 1, 2),
+		[TILFMT_32BIT] = GEOM(1, 1, 4),
+		[TILFMT_PAGE]  = GEOM(SLOT_WIDTH_BITS, SLOT_HEIGHT_BITS, 1),
+};
+
+
+/* lookup table for registers w/ per-engine instances */
+static const uint32_t reg[][4] = {
+		[PAT_STATUS] = {DMM_PAT_STATUS__0, DMM_PAT_STATUS__1,
+				DMM_PAT_STATUS__2, DMM_PAT_STATUS__3},
+		[PAT_DESCR]  = {DMM_PAT_DESCR__0, DMM_PAT_DESCR__1,
+				DMM_PAT_DESCR__2, DMM_PAT_DESCR__3},
+};
+
+/* simple allocator to grab next 16 byte aligned memory from txn */
+static void *alloc_dma(struct dmm_txn *txn, size_t sz, dma_addr_t *pa)
+{
+	void *ptr;
+	struct refill_engine *engine = txn->engine_handle;
+
+	/* dmm programming requires 16 byte aligned addresses */
+	txn->current_pa = round_up(txn->current_pa, 16);
+	txn->current_va = (void *)round_up((long)txn->current_va, 16);
+
+	ptr = txn->current_va;
+	*pa = txn->current_pa;
+
+	txn->current_pa += sz;
+	txn->current_va += sz;
+
+	BUG_ON((txn->current_va - engine->refill_va) > REFILL_BUFFER_SIZE);
+
+	return ptr;
+}
+
+/* check status and spin until wait_mask comes true */
+static int wait_status(struct refill_engine *engine, uint32_t wait_mask)
+{
+	struct dmm *dmm = engine->dmm;
+	uint32_t r = 0, err, i;
+
+	i = DMM_FIXED_RETRY_COUNT;
+	while (true) {
+		r = readl(dmm->base + reg[PAT_STATUS][engine->id]);
+		err = r & DMM_PATSTATUS_ERR;
+		if (err)
+			return -EFAULT;
+
+		if ((r & wait_mask) == wait_mask)
+			break;
+
+		if (--i == 0)
+			return -ETIMEDOUT;
+
+		udelay(1);
+	}
+
+	return 0;
+}
+
+static void release_engine(struct refill_engine *engine)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&list_lock, flags);
+	list_add(&engine->idle_node, &omap_dmm->idle_head);
+	spin_unlock_irqrestore(&list_lock, flags);
+
+	atomic_inc(&omap_dmm->engine_counter);
+	wake_up_interruptible(&omap_dmm->engine_queue);
+}
+
+static irqreturn_t omap_dmm_irq_handler(int irq, void *arg)
+{
+	struct dmm *dmm = arg;
+	uint32_t status = readl(dmm->base + DMM_PAT_IRQSTATUS);
+	int i;
+
+	/* ack IRQ */
+	writel(status, dmm->base + DMM_PAT_IRQSTATUS);
+
+	for (i = 0; i < dmm->num_engines; i++) {
+		if (status & DMM_IRQSTAT_LST) {
+			wake_up_interruptible(&dmm->engines[i].wait_for_refill);
+
+			if (dmm->engines[i].async)
+				release_engine(&dmm->engines[i]);
+		}
+
+		status >>= 8;
+	}
+
+	return IRQ_HANDLED;
+}
+
+/**
+ * Get a handle for a DMM transaction
+ */
+static struct dmm_txn *dmm_txn_init(struct dmm *dmm, struct tcm *tcm)
+{
+	struct dmm_txn *txn = NULL;
+	struct refill_engine *engine = NULL;
+	int ret;
+	unsigned long flags;
+
+
+	/* wait until an engine is available */
+	ret = wait_event_interruptible(omap_dmm->engine_queue,
+		atomic_add_unless(&omap_dmm->engine_counter, -1, 0));
+	if (ret)
+		return ERR_PTR(ret);
+
+	/* grab an idle engine */
+	spin_lock_irqsave(&list_lock, flags);
+	if (!list_empty(&dmm->idle_head)) {
+		engine = list_entry(dmm->idle_head.next, struct refill_engine,
+					idle_node);
+		list_del(&engine->idle_node);
+	}
+	spin_unlock_irqrestore(&list_lock, flags);
+
+	BUG_ON(!engine);
+
+	txn = &engine->txn;
+	engine->tcm = tcm;
+	txn->engine_handle = engine;
+	txn->last_pat = NULL;
+	txn->current_va = engine->refill_va;
+	txn->current_pa = engine->refill_pa;
+
+	return txn;
+}
+
+/**
+ * Add region to DMM transaction.  If pages or pages[i] is NULL, then the
+ * corresponding slot is cleared (ie. dummy_pa is programmed)
+ */
+static void dmm_txn_append(struct dmm_txn *txn, struct pat_area *area,
+		struct page **pages, uint32_t npages, uint32_t roll)
+{
+	dma_addr_t pat_pa = 0;
+	uint32_t *data;
+	struct pat *pat;
+	struct refill_engine *engine = txn->engine_handle;
+	int columns = (1 + area->x1 - area->x0);
+	int rows = (1 + area->y1 - area->y0);
+	int i = columns*rows;
+
+	pat = alloc_dma(txn, sizeof(struct pat), &pat_pa);
+
+	if (txn->last_pat)
+		txn->last_pat->next_pa = (uint32_t)pat_pa;
+
+	pat->area = *area;
+
+	/* adjust Y coordinates based off of container parameters */
+	pat->area.y0 += engine->tcm->y_offset;
+	pat->area.y1 += engine->tcm->y_offset;
+
+	pat->ctrl = (struct pat_ctrl){
+			.start = 1,
+			.lut_id = engine->tcm->lut_id,
+		};
+
+	data = alloc_dma(txn, 4*i, &pat->data_pa);
+
+	while (i--) {
+		int n = i + roll;
+		if (n >= npages)
+			n -= npages;
+		data[i] = (pages && pages[n]) ?
+			page_to_phys(pages[n]) : engine->dmm->dummy_pa;
+	}
+
+	txn->last_pat = pat;
+
+	return;
+}
+
+/**
+ * Commit the DMM transaction.
+ */
+static int dmm_txn_commit(struct dmm_txn *txn, bool wait)
+{
+	int ret = 0;
+	struct refill_engine *engine = txn->engine_handle;
+	struct dmm *dmm = engine->dmm;
+
+	if (!txn->last_pat) {
+		dev_err(engine->dmm->dev, "need at least one txn\n");
+		ret = -EINVAL;
+		goto cleanup;
+	}
+
+	txn->last_pat->next_pa = 0;
+
+	/* write to PAT_DESCR to clear out any pending transaction */
+	writel(0x0, dmm->base + reg[PAT_DESCR][engine->id]);
+
+	/* wait for engine ready: */
+	ret = wait_status(engine, DMM_PATSTATUS_READY);
+	if (ret) {
+		ret = -EFAULT;
+		goto cleanup;
+	}
+
+	/* mark whether it is async to denote list management in IRQ handler */
+	engine->async = wait ? false : true;
+
+	/* kick reload */
+	writel(engine->refill_pa,
+		dmm->base + reg[PAT_DESCR][engine->id]);
+
+	if (wait) {
+		if (wait_event_interruptible_timeout(engine->wait_for_refill,
+				wait_status(engine, DMM_PATSTATUS_READY) == 0,
+				msecs_to_jiffies(1)) <= 0) {
+			dev_err(dmm->dev, "timed out waiting for done\n");
+			ret = -ETIMEDOUT;
+		}
+	}
+
+cleanup:
+	/* only place engine back on list if we are done with it */
+	if (ret || wait)
+		release_engine(engine);
+
+	return ret;
+}
+
+/*
+ * DMM programming
+ */
+static int fill(struct tcm_area *area, struct page **pages,
+		uint32_t npages, uint32_t roll, bool wait)
+{
+	int ret = 0;
+	struct tcm_area slice, area_s;
+	struct dmm_txn *txn;
+
+	txn = dmm_txn_init(omap_dmm, area->tcm);
+	if (IS_ERR_OR_NULL(txn))
+		return -ENOMEM;
+
+	tcm_for_each_slice(slice, *area, area_s) {
+		struct pat_area p_area = {
+				.x0 = slice.p0.x,  .y0 = slice.p0.y,
+				.x1 = slice.p1.x,  .y1 = slice.p1.y,
+		};
+
+		dmm_txn_append(txn, &p_area, pages, npages, roll);
+
+		roll += tcm_sizeof(slice);
+	}
+
+	ret = dmm_txn_commit(txn, wait);
+
+	return ret;
+}
+
+/*
+ * Pin/unpin
+ */
+
+/* note: slots for which pages[i] == NULL are filled w/ dummy page
+ */
+int tiler_pin(struct tiler_block *block, struct page **pages,
+		uint32_t npages, uint32_t roll, bool wait)
+{
+	int ret;
+
+	ret = fill(&block->area, pages, npages, roll, wait);
+
+	if (ret)
+		tiler_unpin(block);
+
+	return ret;
+}
+
+int tiler_unpin(struct tiler_block *block)
+{
+	return fill(&block->area, NULL, 0, 0, false);
+}
+
+/*
+ * Reserve/release
+ */
+struct tiler_block *tiler_reserve_2d(enum tiler_fmt fmt, uint16_t w,
+		uint16_t h, uint16_t align)
+{
+	struct tiler_block *block = kzalloc(sizeof(*block), GFP_KERNEL);
+	u32 min_align = 128;
+	int ret;
+	unsigned long flags;
+
+	BUG_ON(!validfmt(fmt));
+
+	/* convert width/height to slots */
+	w = DIV_ROUND_UP(w, geom[fmt].slot_w);
+	h = DIV_ROUND_UP(h, geom[fmt].slot_h);
+
+	/* convert alignment to slots */
+	min_align = max(min_align, (geom[fmt].slot_w * geom[fmt].cpp));
+	align = ALIGN(align, min_align);
+	align /= geom[fmt].slot_w * geom[fmt].cpp;
+
+	block->fmt = fmt;
+
+	ret = tcm_reserve_2d(containers[fmt], w, h, align, &block->area);
+	if (ret) {
+		kfree(block);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	/* add to allocation list */
+	spin_lock_irqsave(&list_lock, flags);
+	list_add(&block->alloc_node, &omap_dmm->alloc_head);
+	spin_unlock_irqrestore(&list_lock, flags);
+
+	return block;
+}
+
+struct tiler_block *tiler_reserve_1d(size_t size)
+{
+	struct tiler_block *block = kzalloc(sizeof(*block), GFP_KERNEL);
+	int num_pages = (size + PAGE_SIZE - 1) >> PAGE_SHIFT;
+	unsigned long flags;
+
+	if (!block)
+		return ERR_PTR(-ENOMEM);
+
+	block->fmt = TILFMT_PAGE;
+
+	if (tcm_reserve_1d(containers[TILFMT_PAGE], num_pages,
+				&block->area)) {
+		kfree(block);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	spin_lock_irqsave(&list_lock, flags);
+	list_add(&block->alloc_node, &omap_dmm->alloc_head);
+	spin_unlock_irqrestore(&list_lock, flags);
+
+	return block;
+}
+
+/* note: if you have pin'd pages, you should have already unpin'd first! */
+int tiler_release(struct tiler_block *block)
+{
+	int ret = tcm_free(&block->area);
+	unsigned long flags;
+
+	if (block->area.tcm)
+		dev_err(omap_dmm->dev, "failed to release block\n");
+
+	spin_lock_irqsave(&list_lock, flags);
+	list_del(&block->alloc_node);
+	spin_unlock_irqrestore(&list_lock, flags);
+
+	kfree(block);
+	return ret;
+}
+
+/*
+ * Utils
+ */
+
+/* calculate the tiler space address of a pixel in a view orientation...
+ * below description copied from the display subsystem section of TRM:
+ *
+ * When the TILER is addressed, the bits:
+ *   [28:27] = 0x0 for 8-bit tiled
+ *             0x1 for 16-bit tiled
+ *             0x2 for 32-bit tiled
+ *             0x3 for page mode
+ *   [31:29] = 0x0 for 0-degree view
+ *             0x1 for 180-degree view + mirroring
+ *             0x2 for 0-degree view + mirroring
+ *             0x3 for 180-degree view
+ *             0x4 for 270-degree view + mirroring
+ *             0x5 for 270-degree view
+ *             0x6 for 90-degree view
+ *             0x7 for 90-degree view + mirroring
+ * Otherwise the bits indicated the corresponding bit address to access
+ * the SDRAM.
+ */
+static u32 tiler_get_address(enum tiler_fmt fmt, u32 orient, u32 x, u32 y)
+{
+	u32 x_bits, y_bits, tmp, x_mask, y_mask, alignment;
+
+	x_bits = CONT_WIDTH_BITS - geom[fmt].x_shft;
+	y_bits = CONT_HEIGHT_BITS - geom[fmt].y_shft;
+	alignment = geom[fmt].x_shft + geom[fmt].y_shft;
+
+	/* validate coordinate */
+	x_mask = MASK(x_bits);
+	y_mask = MASK(y_bits);
+
+	if (x < 0 || x > x_mask || y < 0 || y > y_mask) {
+		DBG("invalid coords: %u < 0 || %u > %u || %u < 0 || %u > %u",
+				x, x, x_mask, y, y, y_mask);
+		return 0;
+	}
+
+	/* account for mirroring */
+	if (orient & MASK_X_INVERT)
+		x ^= x_mask;
+	if (orient & MASK_Y_INVERT)
+		y ^= y_mask;
+
+	/* get coordinate address */
+	if (orient & MASK_XY_FLIP)
+		tmp = ((x << y_bits) + y);
+	else
+		tmp = ((y << x_bits) + x);
+
+	return TIL_ADDR((tmp << alignment), orient, fmt);
+}
+
+dma_addr_t tiler_ssptr(struct tiler_block *block)
+{
+	BUG_ON(!validfmt(block->fmt));
+
+	return TILVIEW_8BIT + tiler_get_address(block->fmt, 0,
+			block->area.p0.x * geom[block->fmt].slot_w,
+			block->area.p0.y * geom[block->fmt].slot_h);
+}
+
+dma_addr_t tiler_tsptr(struct tiler_block *block, uint32_t orient,
+		uint32_t x, uint32_t y)
+{
+	struct tcm_pt *p = &block->area.p0;
+	BUG_ON(!validfmt(block->fmt));
+
+	return tiler_get_address(block->fmt, orient,
+			(p->x * geom[block->fmt].slot_w) + x,
+			(p->y * geom[block->fmt].slot_h) + y);
+}
+
+void tiler_align(enum tiler_fmt fmt, uint16_t *w, uint16_t *h)
+{
+	BUG_ON(!validfmt(fmt));
+	*w = round_up(*w, geom[fmt].slot_w);
+	*h = round_up(*h, geom[fmt].slot_h);
+}
+
+uint32_t tiler_stride(enum tiler_fmt fmt, uint32_t orient)
+{
+	BUG_ON(!validfmt(fmt));
+
+	if (orient & MASK_XY_FLIP)
+		return 1 << (CONT_HEIGHT_BITS + geom[fmt].x_shft);
+	else
+		return 1 << (CONT_WIDTH_BITS + geom[fmt].y_shft);
+}
+
+size_t tiler_size(enum tiler_fmt fmt, uint16_t w, uint16_t h)
+{
+	tiler_align(fmt, &w, &h);
+	return geom[fmt].cpp * w * h;
+}
+
+size_t tiler_vsize(enum tiler_fmt fmt, uint16_t w, uint16_t h)
+{
+	BUG_ON(!validfmt(fmt));
+	return round_up(geom[fmt].cpp * w, PAGE_SIZE) * h;
+}
+
+bool dmm_is_available(void)
+{
+	return omap_dmm ? true : false;
+}
+
+static int omap_dmm_remove(struct platform_device *dev)
+{
+	struct tiler_block *block, *_block;
+	int i;
+	unsigned long flags;
+
+	if (omap_dmm) {
+		/* free all area regions */
+		spin_lock_irqsave(&list_lock, flags);
+		list_for_each_entry_safe(block, _block, &omap_dmm->alloc_head,
+					alloc_node) {
+			list_del(&block->alloc_node);
+			kfree(block);
+		}
+		spin_unlock_irqrestore(&list_lock, flags);
+
+		for (i = 0; i < omap_dmm->num_lut; i++)
+			if (omap_dmm->tcm && omap_dmm->tcm[i])
+				omap_dmm->tcm[i]->deinit(omap_dmm->tcm[i]);
+		kfree(omap_dmm->tcm);
+
+		kfree(omap_dmm->engines);
+		if (omap_dmm->refill_va)
+			dma_free_writecombine(omap_dmm->dev,
+				REFILL_BUFFER_SIZE * omap_dmm->num_engines,
+				omap_dmm->refill_va,
+				omap_dmm->refill_pa);
+		if (omap_dmm->dummy_page)
+			__free_page(omap_dmm->dummy_page);
+
+		if (omap_dmm->irq > 0)
+			free_irq(omap_dmm->irq, omap_dmm);
+
+		iounmap(omap_dmm->base);
+		kfree(omap_dmm);
+		omap_dmm = NULL;
+	}
+
+	return 0;
+}
+
+static int omap_dmm_probe(struct platform_device *dev)
+{
+	int ret = -EFAULT, i;
+	struct tcm_area area = {0};
+	u32 hwinfo, pat_geom;
+	struct resource *mem;
+
+	omap_dmm = kzalloc(sizeof(*omap_dmm), GFP_KERNEL);
+	if (!omap_dmm) {
+		dev_err(&dev->dev, "failed to allocate driver data section\n");
+		goto fail;
+	}
+
+	/* initialize lists */
+	INIT_LIST_HEAD(&omap_dmm->alloc_head);
+	INIT_LIST_HEAD(&omap_dmm->idle_head);
+
+	init_waitqueue_head(&omap_dmm->engine_queue);
+
+	/* lookup hwmod data - base address and irq */
+	mem = platform_get_resource(dev, IORESOURCE_MEM, 0);
+	if (!mem) {
+		dev_err(&dev->dev, "failed to get base address resource\n");
+		goto fail;
+	}
+
+	omap_dmm->base = ioremap(mem->start, SZ_2K);
+
+	if (!omap_dmm->base) {
+		dev_err(&dev->dev, "failed to get dmm base address\n");
+		goto fail;
+	}
+
+	omap_dmm->irq = platform_get_irq(dev, 0);
+	if (omap_dmm->irq < 0) {
+		dev_err(&dev->dev, "failed to get IRQ resource\n");
+		goto fail;
+	}
+
+	omap_dmm->dev = &dev->dev;
+
+	hwinfo = readl(omap_dmm->base + DMM_PAT_HWINFO);
+	omap_dmm->num_engines = (hwinfo >> 24) & 0x1F;
+	omap_dmm->num_lut = (hwinfo >> 16) & 0x1F;
+	omap_dmm->container_width = 256;
+	omap_dmm->container_height = 128;
+
+	atomic_set(&omap_dmm->engine_counter, omap_dmm->num_engines);
+
+	/* read out actual LUT width and height */
+	pat_geom = readl(omap_dmm->base + DMM_PAT_GEOMETRY);
+	omap_dmm->lut_width = ((pat_geom >> 16) & 0xF) << 5;
+	omap_dmm->lut_height = ((pat_geom >> 24) & 0xF) << 5;
+
+	/* increment LUT by one if on OMAP5 */
+	/* LUT has twice the height, and is split into a separate container */
+	if (omap_dmm->lut_height != omap_dmm->container_height)
+		omap_dmm->num_lut++;
+
+	/* initialize DMM registers */
+	writel(0x88888888, omap_dmm->base + DMM_PAT_VIEW__0);
+	writel(0x88888888, omap_dmm->base + DMM_PAT_VIEW__1);
+	writel(0x80808080, omap_dmm->base + DMM_PAT_VIEW_MAP__0);
+	writel(0x80000000, omap_dmm->base + DMM_PAT_VIEW_MAP_BASE);
+	writel(0x88888888, omap_dmm->base + DMM_TILER_OR__0);
+	writel(0x88888888, omap_dmm->base + DMM_TILER_OR__1);
+
+	ret = request_irq(omap_dmm->irq, omap_dmm_irq_handler, IRQF_SHARED,
+				"omap_dmm_irq_handler", omap_dmm);
+
+	if (ret) {
+		dev_err(&dev->dev, "couldn't register IRQ %d, error %d\n",
+			omap_dmm->irq, ret);
+		omap_dmm->irq = -1;
+		goto fail;
+	}
+
+	/* Enable all interrupts for each refill engine except
+	 * ERR_LUT_MISS<n> (which is just advisory, and we don't care
+	 * about because we want to be able to refill live scanout
+	 * buffers for accelerated pan/scroll) and FILL_DSC<n> which
+	 * we just generally don't care about.
+	 */
+	writel(0x7e7e7e7e, omap_dmm->base + DMM_PAT_IRQENABLE_SET);
+
+	omap_dmm->dummy_page = alloc_page(GFP_KERNEL | __GFP_DMA32);
+	if (!omap_dmm->dummy_page) {
+		dev_err(&dev->dev, "could not allocate dummy page\n");
+		ret = -ENOMEM;
+		goto fail;
+	}
+
+	/* set dma mask for device */
+	/* NOTE: this is a workaround for the hwmod not initializing properly */
+	dev->dev.coherent_dma_mask = DMA_BIT_MASK(32);
+
+	omap_dmm->dummy_pa = page_to_phys(omap_dmm->dummy_page);
+
+	/* alloc refill memory */
+	omap_dmm->refill_va = dma_alloc_writecombine(&dev->dev,
+				REFILL_BUFFER_SIZE * omap_dmm->num_engines,
+				&omap_dmm->refill_pa, GFP_KERNEL);
+	if (!omap_dmm->refill_va) {
+		dev_err(&dev->dev, "could not allocate refill memory\n");
+		goto fail;
+	}
+
+	/* alloc engines */
+	omap_dmm->engines = kzalloc(
+			omap_dmm->num_engines * sizeof(struct refill_engine),
+			GFP_KERNEL);
+	if (!omap_dmm->engines) {
+		dev_err(&dev->dev, "could not allocate engines\n");
+		ret = -ENOMEM;
+		goto fail;
+	}
+
+	for (i = 0; i < omap_dmm->num_engines; i++) {
+		omap_dmm->engines[i].id = i;
+		omap_dmm->engines[i].dmm = omap_dmm;
+		omap_dmm->engines[i].refill_va = omap_dmm->refill_va +
+						(REFILL_BUFFER_SIZE * i);
+		omap_dmm->engines[i].refill_pa = omap_dmm->refill_pa +
+						(REFILL_BUFFER_SIZE * i);
+		init_waitqueue_head(&omap_dmm->engines[i].wait_for_refill);
+
+		list_add(&omap_dmm->engines[i].idle_node, &omap_dmm->idle_head);
+	}
+
+	omap_dmm->tcm = kzalloc(omap_dmm->num_lut * sizeof(*omap_dmm->tcm),
+				GFP_KERNEL);
+	if (!omap_dmm->tcm) {
+		dev_err(&dev->dev, "failed to allocate lut ptrs\n");
+		ret = -ENOMEM;
+		goto fail;
+	}
+
+	/* init containers */
+	/* Each LUT is associated with a TCM (container manager).  We use the
+	   lut_id to denote the lut_id used to identify the correct LUT for
+	   programming during reill operations */
+	for (i = 0; i < omap_dmm->num_lut; i++) {
+		omap_dmm->tcm[i] = sita_init(omap_dmm->container_width,
+						omap_dmm->container_height,
+						NULL);
+
+		if (!omap_dmm->tcm[i]) {
+			dev_err(&dev->dev, "failed to allocate container\n");
+			ret = -ENOMEM;
+			goto fail;
+		}
+
+		omap_dmm->tcm[i]->lut_id = i;
+	}
+
+	/* assign access mode containers to applicable tcm container */
+	/* OMAP 4 has 1 container for all 4 views */
+	/* OMAP 5 has 2 containers, 1 for 2D and 1 for 1D */
+	containers[TILFMT_8BIT] = omap_dmm->tcm[0];
+	containers[TILFMT_16BIT] = omap_dmm->tcm[0];
+	containers[TILFMT_32BIT] = omap_dmm->tcm[0];
+
+	if (omap_dmm->container_height != omap_dmm->lut_height) {
+		/* second LUT is used for PAGE mode.  Programming must use
+		   y offset that is added to all y coordinates.  LUT id is still
+		   0, because it is the same LUT, just the upper 128 lines */
+		containers[TILFMT_PAGE] = omap_dmm->tcm[1];
+		omap_dmm->tcm[1]->y_offset = OMAP5_LUT_OFFSET;
+		omap_dmm->tcm[1]->lut_id = 0;
+	} else {
+		containers[TILFMT_PAGE] = omap_dmm->tcm[0];
+	}
+
+	area = (struct tcm_area) {
+		.tcm = NULL,
+		.p1.x = omap_dmm->container_width - 1,
+		.p1.y = omap_dmm->container_height - 1,
+	};
+
+	/* initialize all LUTs to dummy page entries */
+	for (i = 0; i < omap_dmm->num_lut; i++) {
+		area.tcm = omap_dmm->tcm[i];
+		if (fill(&area, NULL, 0, 0, true))
+			dev_err(omap_dmm->dev, "refill failed");
+	}
+
+	dev_info(omap_dmm->dev, "initialized all PAT entries\n");
+
+	return 0;
+
+fail:
+	if (omap_dmm_remove(dev))
+		dev_err(&dev->dev, "cleanup failed\n");
+	return ret;
+}
+
+/*
+ * debugfs support
+ */
+
+#ifdef CONFIG_DEBUG_FS
+
+static const char *alphabet = "abcdefghijklmnopqrstuvwxyz"
+				"ABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789";
+static const char *special = ".,:;'\"`~!^-+";
+
+static void fill_map(char **map, int xdiv, int ydiv, struct tcm_area *a,
+							char c, bool ovw)
+{
+	int x, y;
+	for (y = a->p0.y / ydiv; y <= a->p1.y / ydiv; y++)
+		for (x = a->p0.x / xdiv; x <= a->p1.x / xdiv; x++)
+			if (map[y][x] == ' ' || ovw)
+				map[y][x] = c;
+}
+
+static void fill_map_pt(char **map, int xdiv, int ydiv, struct tcm_pt *p,
+									char c)
+{
+	map[p->y / ydiv][p->x / xdiv] = c;
+}
+
+static char read_map_pt(char **map, int xdiv, int ydiv, struct tcm_pt *p)
+{
+	return map[p->y / ydiv][p->x / xdiv];
+}
+
+static int map_width(int xdiv, int x0, int x1)
+{
+	return (x1 / xdiv) - (x0 / xdiv) + 1;
+}
+
+static void text_map(char **map, int xdiv, char *nice, int yd, int x0, int x1)
+{
+	char *p = map[yd] + (x0 / xdiv);
+	int w = (map_width(xdiv, x0, x1) - strlen(nice)) / 2;
+	if (w >= 0) {
+		p += w;
+		while (*nice)
+			*p++ = *nice++;
+	}
+}
+
+static void map_1d_info(char **map, int xdiv, int ydiv, char *nice,
+							struct tcm_area *a)
+{
+	sprintf(nice, "%dK", tcm_sizeof(*a) * 4);
+	if (a->p0.y + 1 < a->p1.y) {
+		text_map(map, xdiv, nice, (a->p0.y + a->p1.y) / 2 / ydiv, 0,
+							256 - 1);
+	} else if (a->p0.y < a->p1.y) {
+		if (strlen(nice) < map_width(xdiv, a->p0.x, 256 - 1))
+			text_map(map, xdiv, nice, a->p0.y / ydiv,
+					a->p0.x + xdiv,	256 - 1);
+		else if (strlen(nice) < map_width(xdiv, 0, a->p1.x))
+			text_map(map, xdiv, nice, a->p1.y / ydiv,
+					0, a->p1.y - xdiv);
+	} else if (strlen(nice) + 1 < map_width(xdiv, a->p0.x, a->p1.x)) {
+		text_map(map, xdiv, nice, a->p0.y / ydiv, a->p0.x, a->p1.x);
+	}
+}
+
+static void map_2d_info(char **map, int xdiv, int ydiv, char *nice,
+							struct tcm_area *a)
+{
+	sprintf(nice, "(%d*%d)", tcm_awidth(*a), tcm_aheight(*a));
+	if (strlen(nice) + 1 < map_width(xdiv, a->p0.x, a->p1.x))
+		text_map(map, xdiv, nice, (a->p0.y + a->p1.y) / 2 / ydiv,
+							a->p0.x, a->p1.x);
+}
+
+int tiler_map_show(struct seq_file *s, void *arg)
+{
+	int xdiv = 2, ydiv = 1;
+	char **map = NULL, *global_map;
+	struct tiler_block *block;
+	struct tcm_area a, p;
+	int i;
+	const char *m2d = alphabet;
+	const char *a2d = special;
+	const char *m2dp = m2d, *a2dp = a2d;
+	char nice[128];
+	int h_adj;
+	int w_adj;
+	unsigned long flags;
+	int lut_idx;
+
+
+	if (!omap_dmm) {
+		/* early return if dmm/tiler device is not initialized */
+		return 0;
+	}
+
+	h_adj = omap_dmm->container_height / ydiv;
+	w_adj = omap_dmm->container_width / xdiv;
+
+	map = kmalloc(h_adj * sizeof(*map), GFP_KERNEL);
+	global_map = kmalloc((w_adj + 1) * h_adj, GFP_KERNEL);
+
+	if (!map || !global_map)
+		goto error;
+
+	for (lut_idx = 0; lut_idx < omap_dmm->num_lut; lut_idx++) {
+		memset(map, 0, sizeof(h_adj * sizeof(*map)));
+		memset(global_map, ' ', (w_adj + 1) * h_adj);
+
+		for (i = 0; i < omap_dmm->container_height; i++) {
+			map[i] = global_map + i * (w_adj + 1);
+			map[i][w_adj] = 0;
+		}
+
+		spin_lock_irqsave(&list_lock, flags);
+
+		list_for_each_entry(block, &omap_dmm->alloc_head, alloc_node) {
+			if (block->area.tcm == omap_dmm->tcm[lut_idx]) {
+				if (block->fmt != TILFMT_PAGE) {
+					fill_map(map, xdiv, ydiv, &block->area,
+						*m2dp, true);
+					if (!*++a2dp)
+						a2dp = a2d;
+					if (!*++m2dp)
+						m2dp = m2d;
+					map_2d_info(map, xdiv, ydiv, nice,
+							&block->area);
+				} else {
+					bool start = read_map_pt(map, xdiv,
+						ydiv, &block->area.p0) == ' ';
+					bool end = read_map_pt(map, xdiv, ydiv,
+							&block->area.p1) == ' ';
+
+					tcm_for_each_slice(a, block->area, p)
+						fill_map(map, xdiv, ydiv, &a,
+							'=', true);
+					fill_map_pt(map, xdiv, ydiv,
+							&block->area.p0,
+							start ? '<' : 'X');
+					fill_map_pt(map, xdiv, ydiv,
+							&block->area.p1,
+							end ? '>' : 'X');
+					map_1d_info(map, xdiv, ydiv, nice,
+							&block->area);
+				}
+			}
+		}
+
+		spin_unlock_irqrestore(&list_lock, flags);
+
+		if (s) {
+			seq_printf(s, "CONTAINER %d DUMP BEGIN\n", lut_idx);
+			for (i = 0; i < 128; i++)
+				seq_printf(s, "%03d:%s\n", i, map[i]);
+			seq_printf(s, "CONTAINER %d DUMP END\n", lut_idx);
+		} else {
+			dev_dbg(omap_dmm->dev, "CONTAINER %d DUMP BEGIN\n",
+				lut_idx);
+			for (i = 0; i < 128; i++)
+				dev_dbg(omap_dmm->dev, "%03d:%s\n", i, map[i]);
+			dev_dbg(omap_dmm->dev, "CONTAINER %d DUMP END\n",
+				lut_idx);
+		}
+	}
+
+error:
+	kfree(map);
+	kfree(global_map);
+
+	return 0;
+}
+#endif
+
+#ifdef CONFIG_PM
+static int omap_dmm_resume(struct device *dev)
+{
+	struct tcm_area area;
+	int i;
+
+	if (!omap_dmm)
+		return -ENODEV;
+
+	area = (struct tcm_area) {
+		.tcm = NULL,
+		.p1.x = omap_dmm->container_width - 1,
+		.p1.y = omap_dmm->container_height - 1,
+	};
+
+	/* initialize all LUTs to dummy page entries */
+	for (i = 0; i < omap_dmm->num_lut; i++) {
+		area.tcm = omap_dmm->tcm[i];
+		if (fill(&area, NULL, 0, 0, true))
+			dev_err(dev, "refill failed");
+	}
+
+	return 0;
+}
+
+static const struct dev_pm_ops omap_dmm_pm_ops = {
+	.resume = omap_dmm_resume,
+};
+#endif
+
+struct platform_driver omap_dmm_driver = {
+	.probe = omap_dmm_probe,
+	.remove = omap_dmm_remove,
+	.driver = {
+		.owner = THIS_MODULE,
+		.name = DMM_DRIVER_NAME,
+#ifdef CONFIG_PM
+		.pm = &omap_dmm_pm_ops,
+#endif
+	},
+};
+
+MODULE_LICENSE("GPL v2");
+MODULE_AUTHOR("Andy Gross <andy.gross@ti.com>");
+MODULE_DESCRIPTION("OMAP DMM/Tiler Driver");
+MODULE_ALIAS("platform:" DMM_DRIVER_NAME);
diff --git a/drivers/gpu/drm/omapdrm/omap_dmm_tiler.h b/drivers/gpu/drm/omapdrm/omap_dmm_tiler.h
new file mode 100644
index 0000000..4fdd61e
--- /dev/null
+++ b/drivers/gpu/drm/omapdrm/omap_dmm_tiler.h
@@ -0,0 +1,141 @@
+/*
+ *
+ * Copyright (C) 2011 Texas Instruments Incorporated - http://www.ti.com/
+ * Author: Rob Clark <rob@ti.com>
+ *         Andy Gross <andy.gross@ti.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation version 2.
+ *
+ * This program is distributed "as is" WITHOUT ANY WARRANTY of any
+ * kind, whether express or implied; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+#ifndef OMAP_DMM_TILER_H
+#define OMAP_DMM_TILER_H
+
+#include "omap_drv.h"
+#include "tcm.h"
+
+enum tiler_fmt {
+	TILFMT_8BIT = 0,
+	TILFMT_16BIT,
+	TILFMT_32BIT,
+	TILFMT_PAGE,
+	TILFMT_NFORMATS
+};
+
+struct pat_area {
+	u32 x0:8;
+	u32 y0:8;
+	u32 x1:8;
+	u32 y1:8;
+};
+
+struct tiler_block {
+	struct list_head alloc_node;	/* node for global block list */
+	struct tcm_area area;		/* area */
+	enum tiler_fmt fmt;		/* format */
+};
+
+/* bits representing the same slot in DMM-TILER hw-block */
+#define SLOT_WIDTH_BITS         6
+#define SLOT_HEIGHT_BITS        6
+
+/* bits reserved to describe coordinates in DMM-TILER hw-block */
+#define CONT_WIDTH_BITS         14
+#define CONT_HEIGHT_BITS        13
+
+/* calculated constants */
+#define TILER_PAGE              (1 << (SLOT_WIDTH_BITS + SLOT_HEIGHT_BITS))
+#define TILER_WIDTH             (1 << (CONT_WIDTH_BITS - SLOT_WIDTH_BITS))
+#define TILER_HEIGHT            (1 << (CONT_HEIGHT_BITS - SLOT_HEIGHT_BITS))
+
+/*
+Table 15-11. Coding and Description of TILER Orientations
+S Y X	Description				Alternate description
+0 0 0	0-degree view				Natural view
+0 0 1	0-degree view with vertical mirror 	180-degree view with horizontal mirror
+0 1 0	0-degree view with horizontal mirror 	180-degree view with vertical mirror
+0 1 1	180-degree view
+1 0 0	90-degree view with vertical mirror	270-degree view with horizontal mirror
+1 0 1	270-degree view
+1 1 0	90-degree view
+1 1 1	90-degree view with horizontal mirror	270-degree view with vertical mirror
+ */
+#define MASK_XY_FLIP		(1 << 31)
+#define MASK_Y_INVERT		(1 << 30)
+#define MASK_X_INVERT		(1 << 29)
+#define SHIFT_ACC_MODE		27
+#define MASK_ACC_MODE		3
+
+#define MASK(bits) ((1 << (bits)) - 1)
+
+#define TILVIEW_8BIT    0x60000000u
+#define TILVIEW_16BIT   (TILVIEW_8BIT  + VIEW_SIZE)
+#define TILVIEW_32BIT   (TILVIEW_16BIT + VIEW_SIZE)
+#define TILVIEW_PAGE    (TILVIEW_32BIT + VIEW_SIZE)
+#define TILVIEW_END     (TILVIEW_PAGE  + VIEW_SIZE)
+
+/* create tsptr by adding view orientation and access mode */
+#define TIL_ADDR(x, orient, a)\
+	((u32) (x) | (orient) | ((a) << SHIFT_ACC_MODE))
+
+#ifdef CONFIG_DEBUG_FS
+int tiler_map_show(struct seq_file *s, void *arg);
+#endif
+
+/* pin/unpin */
+int tiler_pin(struct tiler_block *block, struct page **pages,
+		uint32_t npages, uint32_t roll, bool wait);
+int tiler_unpin(struct tiler_block *block);
+
+/* reserve/release */
+struct tiler_block *tiler_reserve_2d(enum tiler_fmt fmt, uint16_t w, uint16_t h,
+				uint16_t align);
+struct tiler_block *tiler_reserve_1d(size_t size);
+int tiler_release(struct tiler_block *block);
+
+/* utilities */
+dma_addr_t tiler_ssptr(struct tiler_block *block);
+dma_addr_t tiler_tsptr(struct tiler_block *block, uint32_t orient,
+		uint32_t x, uint32_t y);
+uint32_t tiler_stride(enum tiler_fmt fmt, uint32_t orient);
+size_t tiler_size(enum tiler_fmt fmt, uint16_t w, uint16_t h);
+size_t tiler_vsize(enum tiler_fmt fmt, uint16_t w, uint16_t h);
+void tiler_align(enum tiler_fmt fmt, uint16_t *w, uint16_t *h);
+bool dmm_is_available(void);
+
+extern struct platform_driver omap_dmm_driver;
+
+/* GEM bo flags -> tiler fmt */
+static inline enum tiler_fmt gem2fmt(uint32_t flags)
+{
+	switch (flags & OMAP_BO_TILED) {
+	case OMAP_BO_TILED_8:
+		return TILFMT_8BIT;
+	case OMAP_BO_TILED_16:
+		return TILFMT_16BIT;
+	case OMAP_BO_TILED_32:
+		return TILFMT_32BIT;
+	default:
+		return TILFMT_PAGE;
+	}
+}
+
+static inline bool validfmt(enum tiler_fmt fmt)
+{
+	switch (fmt) {
+	case TILFMT_8BIT:
+	case TILFMT_16BIT:
+	case TILFMT_32BIT:
+	case TILFMT_PAGE:
+		return true;
+	default:
+		return false;
+	}
+}
+
+#endif
diff --git a/drivers/gpu/drm/omapdrm/omap_drv.c b/drivers/gpu/drm/omapdrm/omap_drv.c
new file mode 100644
index 0000000..9083538
--- /dev/null
+++ b/drivers/gpu/drm/omapdrm/omap_drv.c
@@ -0,0 +1,610 @@
+/*
+ * drivers/gpu/drm/omapdrm/omap_drv.c
+ *
+ * Copyright (C) 2011 Texas Instruments
+ * Author: Rob Clark <rob@ti.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "omap_drv.h"
+
+#include "drm_crtc_helper.h"
+#include "drm_fb_helper.h"
+#include "omap_dmm_tiler.h"
+
+#define DRIVER_NAME		MODULE_NAME
+#define DRIVER_DESC		"OMAP DRM"
+#define DRIVER_DATE		"20110917"
+#define DRIVER_MAJOR		1
+#define DRIVER_MINOR		0
+#define DRIVER_PATCHLEVEL	0
+
+static int num_crtc = CONFIG_DRM_OMAP_NUM_CRTCS;
+
+MODULE_PARM_DESC(num_crtc, "Number of overlays to use as CRTCs");
+module_param(num_crtc, int, 0600);
+
+/*
+ * mode config funcs
+ */
+
+/* Notes about mapping DSS and DRM entities:
+ *    CRTC:        overlay
+ *    encoder:     manager.. with some extension to allow one primary CRTC
+ *                 and zero or more video CRTC's to be mapped to one encoder?
+ *    connector:   dssdev.. manager can be attached/detached from different
+ *                 devices
+ */
+
+static void omap_fb_output_poll_changed(struct drm_device *dev)
+{
+	struct omap_drm_private *priv = dev->dev_private;
+	DBG("dev=%p", dev);
+	if (priv->fbdev)
+		drm_fb_helper_hotplug_event(priv->fbdev);
+}
+
+static const struct drm_mode_config_funcs omap_mode_config_funcs = {
+	.fb_create = omap_framebuffer_create,
+	.output_poll_changed = omap_fb_output_poll_changed,
+};
+
+static int get_connector_type(struct omap_dss_device *dssdev)
+{
+	switch (dssdev->type) {
+	case OMAP_DISPLAY_TYPE_HDMI:
+		return DRM_MODE_CONNECTOR_HDMIA;
+	case OMAP_DISPLAY_TYPE_DPI:
+		if (!strcmp(dssdev->name, "dvi"))
+			return DRM_MODE_CONNECTOR_DVID;
+		/* fallthrough */
+	default:
+		return DRM_MODE_CONNECTOR_Unknown;
+	}
+}
+
+static int omap_modeset_init(struct drm_device *dev)
+{
+	struct omap_drm_private *priv = dev->dev_private;
+	struct omap_dss_device *dssdev = NULL;
+	int num_ovls = dss_feat_get_num_ovls();
+	int id;
+
+	drm_mode_config_init(dev);
+
+	omap_drm_irq_install(dev);
+
+	/*
+	 * Create private planes and CRTCs for the last NUM_CRTCs overlay
+	 * plus manager:
+	 */
+	for (id = 0; id < min(num_crtc, num_ovls); id++) {
+		struct drm_plane *plane;
+		struct drm_crtc *crtc;
+
+		plane = omap_plane_init(dev, id, true);
+		crtc = omap_crtc_init(dev, plane, pipe2chan(id), id);
+
+		BUG_ON(priv->num_crtcs >= ARRAY_SIZE(priv->crtcs));
+		priv->crtcs[id] = crtc;
+		priv->num_crtcs++;
+
+		priv->planes[id] = plane;
+		priv->num_planes++;
+	}
+
+	/*
+	 * Create normal planes for the remaining overlays:
+	 */
+	for (; id < num_ovls; id++) {
+		struct drm_plane *plane = omap_plane_init(dev, id, false);
+
+		BUG_ON(priv->num_planes >= ARRAY_SIZE(priv->planes));
+		priv->planes[priv->num_planes++] = plane;
+	}
+
+	for_each_dss_dev(dssdev) {
+		struct drm_connector *connector;
+		struct drm_encoder *encoder;
+
+		if (!dssdev->driver) {
+			dev_warn(dev->dev, "%s has no driver.. skipping it\n",
+					dssdev->name);
+			return 0;
+		}
+
+		if (!(dssdev->driver->get_timings ||
+					dssdev->driver->read_edid)) {
+			dev_warn(dev->dev, "%s driver does not support "
+				"get_timings or read_edid.. skipping it!\n",
+				dssdev->name);
+			return 0;
+		}
+
+		encoder = omap_encoder_init(dev, dssdev);
+
+		if (!encoder) {
+			dev_err(dev->dev, "could not create encoder: %s\n",
+					dssdev->name);
+			return -ENOMEM;
+		}
+
+		connector = omap_connector_init(dev,
+				get_connector_type(dssdev), dssdev, encoder);
+
+		if (!connector) {
+			dev_err(dev->dev, "could not create connector: %s\n",
+					dssdev->name);
+			return -ENOMEM;
+		}
+
+		BUG_ON(priv->num_encoders >= ARRAY_SIZE(priv->encoders));
+		BUG_ON(priv->num_connectors >= ARRAY_SIZE(priv->connectors));
+
+		priv->encoders[priv->num_encoders++] = encoder;
+		priv->connectors[priv->num_connectors++] = connector;
+
+		drm_mode_connector_attach_encoder(connector, encoder);
+
+		/* figure out which crtc's we can connect the encoder to: */
+		encoder->possible_crtcs = 0;
+		for (id = 0; id < priv->num_crtcs; id++) {
+			enum omap_dss_output_id supported_outputs =
+					dss_feat_get_supported_outputs(pipe2chan(id));
+			if (supported_outputs & dssdev->output->id)
+				encoder->possible_crtcs |= (1 << id);
+		}
+	}
+
+	dev->mode_config.min_width = 32;
+	dev->mode_config.min_height = 32;
+
+	/* note: eventually will need some cpu_is_omapXYZ() type stuff here
+	 * to fill in these limits properly on different OMAP generations..
+	 */
+	dev->mode_config.max_width = 2048;
+	dev->mode_config.max_height = 2048;
+
+	dev->mode_config.funcs = &omap_mode_config_funcs;
+
+	return 0;
+}
+
+static void omap_modeset_free(struct drm_device *dev)
+{
+	drm_mode_config_cleanup(dev);
+}
+
+/*
+ * drm ioctl funcs
+ */
+
+
+static int ioctl_get_param(struct drm_device *dev, void *data,
+		struct drm_file *file_priv)
+{
+	struct omap_drm_private *priv = dev->dev_private;
+	struct drm_omap_param *args = data;
+
+	DBG("%p: param=%llu", dev, args->param);
+
+	switch (args->param) {
+	case OMAP_PARAM_CHIPSET_ID:
+		args->value = priv->omaprev;
+		break;
+	default:
+		DBG("unknown parameter %lld", args->param);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int ioctl_set_param(struct drm_device *dev, void *data,
+		struct drm_file *file_priv)
+{
+	struct drm_omap_param *args = data;
+
+	switch (args->param) {
+	default:
+		DBG("unknown parameter %lld", args->param);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int ioctl_gem_new(struct drm_device *dev, void *data,
+		struct drm_file *file_priv)
+{
+	struct drm_omap_gem_new *args = data;
+	VERB("%p:%p: size=0x%08x, flags=%08x", dev, file_priv,
+			args->size.bytes, args->flags);
+	return omap_gem_new_handle(dev, file_priv, args->size,
+			args->flags, &args->handle);
+}
+
+static int ioctl_gem_cpu_prep(struct drm_device *dev, void *data,
+		struct drm_file *file_priv)
+{
+	struct drm_omap_gem_cpu_prep *args = data;
+	struct drm_gem_object *obj;
+	int ret;
+
+	VERB("%p:%p: handle=%d, op=%x", dev, file_priv, args->handle, args->op);
+
+	obj = drm_gem_object_lookup(dev, file_priv, args->handle);
+	if (!obj)
+		return -ENOENT;
+
+	ret = omap_gem_op_sync(obj, args->op);
+
+	if (!ret)
+		ret = omap_gem_op_start(obj, args->op);
+
+	drm_gem_object_unreference_unlocked(obj);
+
+	return ret;
+}
+
+static int ioctl_gem_cpu_fini(struct drm_device *dev, void *data,
+		struct drm_file *file_priv)
+{
+	struct drm_omap_gem_cpu_fini *args = data;
+	struct drm_gem_object *obj;
+	int ret;
+
+	VERB("%p:%p: handle=%d", dev, file_priv, args->handle);
+
+	obj = drm_gem_object_lookup(dev, file_priv, args->handle);
+	if (!obj)
+		return -ENOENT;
+
+	/* XXX flushy, flushy */
+	ret = 0;
+
+	if (!ret)
+		ret = omap_gem_op_finish(obj, args->op);
+
+	drm_gem_object_unreference_unlocked(obj);
+
+	return ret;
+}
+
+static int ioctl_gem_info(struct drm_device *dev, void *data,
+		struct drm_file *file_priv)
+{
+	struct drm_omap_gem_info *args = data;
+	struct drm_gem_object *obj;
+	int ret = 0;
+
+	VERB("%p:%p: handle=%d", dev, file_priv, args->handle);
+
+	obj = drm_gem_object_lookup(dev, file_priv, args->handle);
+	if (!obj)
+		return -ENOENT;
+
+	args->size = omap_gem_mmap_size(obj);
+	args->offset = omap_gem_mmap_offset(obj);
+
+	drm_gem_object_unreference_unlocked(obj);
+
+	return ret;
+}
+
+struct drm_ioctl_desc ioctls[DRM_COMMAND_END - DRM_COMMAND_BASE] = {
+	DRM_IOCTL_DEF_DRV(OMAP_GET_PARAM, ioctl_get_param, DRM_UNLOCKED|DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(OMAP_SET_PARAM, ioctl_set_param, DRM_UNLOCKED|DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(OMAP_GEM_NEW, ioctl_gem_new, DRM_UNLOCKED|DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(OMAP_GEM_CPU_PREP, ioctl_gem_cpu_prep, DRM_UNLOCKED|DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(OMAP_GEM_CPU_FINI, ioctl_gem_cpu_fini, DRM_UNLOCKED|DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(OMAP_GEM_INFO, ioctl_gem_info, DRM_UNLOCKED|DRM_AUTH),
+};
+
+/*
+ * drm driver funcs
+ */
+
+/**
+ * load - setup chip and create an initial config
+ * @dev: DRM device
+ * @flags: startup flags
+ *
+ * The driver load routine has to do several things:
+ *   - initialize the memory manager
+ *   - allocate initial config memory
+ *   - setup the DRM framebuffer with the allocated memory
+ */
+static int dev_load(struct drm_device *dev, unsigned long flags)
+{
+	struct omap_drm_platform_data *pdata = dev->dev->platform_data;
+	struct omap_drm_private *priv;
+	int ret;
+
+	DBG("load: dev=%p", dev);
+
+	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
+	if (!priv) {
+		dev_err(dev->dev, "could not allocate priv\n");
+		return -ENOMEM;
+	}
+
+	priv->omaprev = pdata->omaprev;
+
+	dev->dev_private = priv;
+
+	priv->wq = alloc_ordered_workqueue("omapdrm", 0);
+
+	INIT_LIST_HEAD(&priv->obj_list);
+
+	omap_gem_init(dev);
+
+	ret = omap_modeset_init(dev);
+	if (ret) {
+		dev_err(dev->dev, "omap_modeset_init failed: ret=%d\n", ret);
+		dev->dev_private = NULL;
+		kfree(priv);
+		return ret;
+	}
+
+	ret = drm_vblank_init(dev, priv->num_crtcs);
+	if (ret)
+		dev_warn(dev->dev, "could not init vblank\n");
+
+	priv->fbdev = omap_fbdev_init(dev);
+	if (!priv->fbdev) {
+		dev_warn(dev->dev, "omap_fbdev_init failed\n");
+		/* well, limp along without an fbdev.. maybe X11 will work? */
+	}
+
+	/* store off drm_device for use in pm ops */
+	dev_set_drvdata(dev->dev, dev);
+
+	drm_kms_helper_poll_init(dev);
+
+	return 0;
+}
+
+static int dev_unload(struct drm_device *dev)
+{
+	struct omap_drm_private *priv = dev->dev_private;
+
+	DBG("unload: dev=%p", dev);
+
+	drm_kms_helper_poll_fini(dev);
+	drm_vblank_cleanup(dev);
+	omap_drm_irq_uninstall(dev);
+
+	omap_fbdev_free(dev);
+	omap_modeset_free(dev);
+	omap_gem_deinit(dev);
+
+	flush_workqueue(priv->wq);
+	destroy_workqueue(priv->wq);
+
+	kfree(dev->dev_private);
+	dev->dev_private = NULL;
+
+	dev_set_drvdata(dev->dev, NULL);
+
+	return 0;
+}
+
+static int dev_open(struct drm_device *dev, struct drm_file *file)
+{
+	file->driver_priv = NULL;
+
+	DBG("open: dev=%p, file=%p", dev, file);
+
+	return 0;
+}
+
+static int dev_firstopen(struct drm_device *dev)
+{
+	DBG("firstopen: dev=%p", dev);
+	return 0;
+}
+
+/**
+ * lastclose - clean up after all DRM clients have exited
+ * @dev: DRM device
+ *
+ * Take care of cleaning up after all DRM clients have exited.  In the
+ * mode setting case, we want to restore the kernel's initial mode (just
+ * in case the last client left us in a bad state).
+ */
+static void dev_lastclose(struct drm_device *dev)
+{
+	int i;
+
+	/* we don't support vga-switcheroo.. so just make sure the fbdev
+	 * mode is active
+	 */
+	struct omap_drm_private *priv = dev->dev_private;
+	int ret;
+
+	DBG("lastclose: dev=%p", dev);
+
+	if (priv->rotation_prop) {
+		/* need to restore default rotation state.. not sure
+		 * if there is a cleaner way to restore properties to
+		 * default state?  Maybe a flag that properties should
+		 * automatically be restored to default state on
+		 * lastclose?
+		 */
+		for (i = 0; i < priv->num_crtcs; i++) {
+			drm_object_property_set_value(&priv->crtcs[i]->base,
+					priv->rotation_prop, 0);
+		}
+
+		for (i = 0; i < priv->num_planes; i++) {
+			drm_object_property_set_value(&priv->planes[i]->base,
+					priv->rotation_prop, 0);
+		}
+	}
+
+	drm_modeset_lock_all(dev);
+	ret = drm_fb_helper_restore_fbdev_mode(priv->fbdev);
+	drm_modeset_unlock_all(dev);
+	if (ret)
+		DBG("failed to restore crtc mode");
+}
+
+static void dev_preclose(struct drm_device *dev, struct drm_file *file)
+{
+	DBG("preclose: dev=%p", dev);
+}
+
+static void dev_postclose(struct drm_device *dev, struct drm_file *file)
+{
+	DBG("postclose: dev=%p, file=%p", dev, file);
+}
+
+static const struct vm_operations_struct omap_gem_vm_ops = {
+	.fault = omap_gem_fault,
+	.open = drm_gem_vm_open,
+	.close = drm_gem_vm_close,
+};
+
+static const struct file_operations omapdriver_fops = {
+		.owner = THIS_MODULE,
+		.open = drm_open,
+		.unlocked_ioctl = drm_ioctl,
+		.release = drm_release,
+		.mmap = omap_gem_mmap,
+		.poll = drm_poll,
+		.fasync = drm_fasync,
+		.read = drm_read,
+		.llseek = noop_llseek,
+};
+
+static struct drm_driver omap_drm_driver = {
+		.driver_features =
+				DRIVER_HAVE_IRQ | DRIVER_MODESET | DRIVER_GEM | DRIVER_PRIME,
+		.load = dev_load,
+		.unload = dev_unload,
+		.open = dev_open,
+		.firstopen = dev_firstopen,
+		.lastclose = dev_lastclose,
+		.preclose = dev_preclose,
+		.postclose = dev_postclose,
+		.get_vblank_counter = drm_vblank_count,
+		.enable_vblank = omap_irq_enable_vblank,
+		.disable_vblank = omap_irq_disable_vblank,
+		.irq_preinstall = omap_irq_preinstall,
+		.irq_postinstall = omap_irq_postinstall,
+		.irq_uninstall = omap_irq_uninstall,
+		.irq_handler = omap_irq_handler,
+#ifdef CONFIG_DEBUG_FS
+		.debugfs_init = omap_debugfs_init,
+		.debugfs_cleanup = omap_debugfs_cleanup,
+#endif
+		.prime_handle_to_fd = drm_gem_prime_handle_to_fd,
+		.prime_fd_to_handle = drm_gem_prime_fd_to_handle,
+		.gem_prime_export = omap_gem_prime_export,
+		.gem_prime_import = omap_gem_prime_import,
+		.gem_init_object = omap_gem_init_object,
+		.gem_free_object = omap_gem_free_object,
+		.gem_vm_ops = &omap_gem_vm_ops,
+		.dumb_create = omap_gem_dumb_create,
+		.dumb_map_offset = omap_gem_dumb_map_offset,
+		.dumb_destroy = omap_gem_dumb_destroy,
+		.ioctls = ioctls,
+		.num_ioctls = DRM_OMAP_NUM_IOCTLS,
+		.fops = &omapdriver_fops,
+		.name = DRIVER_NAME,
+		.desc = DRIVER_DESC,
+		.date = DRIVER_DATE,
+		.major = DRIVER_MAJOR,
+		.minor = DRIVER_MINOR,
+		.patchlevel = DRIVER_PATCHLEVEL,
+};
+
+static int pdev_suspend(struct platform_device *pDevice, pm_message_t state)
+{
+	DBG("");
+	return 0;
+}
+
+static int pdev_resume(struct platform_device *device)
+{
+	DBG("");
+	return 0;
+}
+
+static void pdev_shutdown(struct platform_device *device)
+{
+	DBG("");
+}
+
+static int pdev_probe(struct platform_device *device)
+{
+	DBG("%s", device->name);
+	return drm_platform_init(&omap_drm_driver, device);
+}
+
+static int pdev_remove(struct platform_device *device)
+{
+	DBG("");
+	drm_platform_exit(&omap_drm_driver, device);
+
+	platform_driver_unregister(&omap_dmm_driver);
+	return 0;
+}
+
+#ifdef CONFIG_PM
+static const struct dev_pm_ops omapdrm_pm_ops = {
+	.resume = omap_gem_resume,
+};
+#endif
+
+struct platform_driver pdev = {
+		.driver = {
+			.name = DRIVER_NAME,
+			.owner = THIS_MODULE,
+#ifdef CONFIG_PM
+			.pm = &omapdrm_pm_ops,
+#endif
+		},
+		.probe = pdev_probe,
+		.remove = pdev_remove,
+		.suspend = pdev_suspend,
+		.resume = pdev_resume,
+		.shutdown = pdev_shutdown,
+};
+
+static int __init omap_drm_init(void)
+{
+	DBG("init");
+	if (platform_driver_register(&omap_dmm_driver)) {
+		/* we can continue on without DMM.. so not fatal */
+		dev_err(NULL, "DMM registration failed\n");
+	}
+	return platform_driver_register(&pdev);
+}
+
+static void __exit omap_drm_fini(void)
+{
+	DBG("fini");
+	platform_driver_unregister(&pdev);
+}
+
+/* need late_initcall() so we load after dss_driver's are loaded */
+late_initcall(omap_drm_init);
+module_exit(omap_drm_fini);
+
+MODULE_AUTHOR("Rob Clark <rob@ti.com>");
+MODULE_DESCRIPTION("OMAP DRM Display Driver");
+MODULE_ALIAS("platform:" DRIVER_NAME);
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/gpu/drm/omapdrm/omap_drv.h b/drivers/gpu/drm/omapdrm/omap_drv.h
new file mode 100644
index 0000000..d4f997b
--- /dev/null
+++ b/drivers/gpu/drm/omapdrm/omap_drv.h
@@ -0,0 +1,333 @@
+/*
+ * drivers/gpu/drm/omapdrm/omap_drv.h
+ *
+ * Copyright (C) 2011 Texas Instruments
+ * Author: Rob Clark <rob@ti.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef __OMAP_DRV_H__
+#define __OMAP_DRV_H__
+
+#include <video/omapdss.h>
+#include <linux/module.h>
+#include <linux/types.h>
+#include <drm/drmP.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/omap_drm.h>
+#include <linux/platform_data/omap_drm.h>
+
+
+#define DBG(fmt, ...) DRM_DEBUG(fmt"\n", ##__VA_ARGS__)
+#define VERB(fmt, ...) if (0) DRM_DEBUG(fmt, ##__VA_ARGS__) /* verbose debug */
+
+#define MODULE_NAME     "omapdrm"
+
+/* max # of mapper-id's that can be assigned.. todo, come up with a better
+ * (but still inexpensive) way to store/access per-buffer mapper private
+ * data..
+ */
+#define MAX_MAPPERS 2
+
+/* parameters which describe (unrotated) coordinates of scanout within a fb: */
+struct omap_drm_window {
+	uint32_t rotation;
+	int32_t  crtc_x, crtc_y;		/* signed because can be offscreen */
+	uint32_t crtc_w, crtc_h;
+	uint32_t src_x, src_y;
+	uint32_t src_w, src_h;
+};
+
+/* Once GO bit is set, we can't make further updates to shadowed registers
+ * until the GO bit is cleared.  So various parts in the kms code that need
+ * to update shadowed registers queue up a pair of callbacks, pre_apply
+ * which is called before setting GO bit, and post_apply that is called
+ * after GO bit is cleared.  The crtc manages the queuing, and everyone
+ * else goes thru omap_crtc_apply() using these callbacks so that the
+ * code which has to deal w/ GO bit state is centralized.
+ */
+struct omap_drm_apply {
+	struct list_head pending_node, queued_node;
+	bool queued;
+	void (*pre_apply)(struct omap_drm_apply *apply);
+	void (*post_apply)(struct omap_drm_apply *apply);
+};
+
+/* For transiently registering for different DSS irqs that various parts
+ * of the KMS code need during setup/configuration.  We these are not
+ * necessarily the same as what drm_vblank_get/put() are requesting, and
+ * the hysteresis in drm_vblank_put() is not necessarily desirable for
+ * internal housekeeping related irq usage.
+ */
+struct omap_drm_irq {
+	struct list_head node;
+	uint32_t irqmask;
+	bool registered;
+	void (*irq)(struct omap_drm_irq *irq, uint32_t irqstatus);
+};
+
+/* For KMS code that needs to wait for a certain # of IRQs:
+ */
+struct omap_irq_wait;
+struct omap_irq_wait * omap_irq_wait_init(struct drm_device *dev,
+		uint32_t irqmask, int count);
+int omap_irq_wait(struct drm_device *dev, struct omap_irq_wait *wait,
+		unsigned long timeout);
+
+struct omap_drm_private {
+	uint32_t omaprev;
+
+	unsigned int num_crtcs;
+	struct drm_crtc *crtcs[8];
+
+	unsigned int num_planes;
+	struct drm_plane *planes[8];
+
+	unsigned int num_encoders;
+	struct drm_encoder *encoders[8];
+
+	unsigned int num_connectors;
+	struct drm_connector *connectors[8];
+
+	struct drm_fb_helper *fbdev;
+
+	struct workqueue_struct *wq;
+
+	/* list of GEM objects: */
+	struct list_head obj_list;
+
+	bool has_dmm;
+
+	/* properties: */
+	struct drm_property *rotation_prop;
+	struct drm_property *zorder_prop;
+
+	/* irq handling: */
+	struct list_head irq_list;    /* list of omap_drm_irq */
+	uint32_t vblank_mask;         /* irq bits set for userspace vblank */
+	struct omap_drm_irq error_handler;
+};
+
+/* this should probably be in drm-core to standardize amongst drivers */
+#define DRM_ROTATE_0	0
+#define DRM_ROTATE_90	1
+#define DRM_ROTATE_180	2
+#define DRM_ROTATE_270	3
+#define DRM_REFLECT_X	4
+#define DRM_REFLECT_Y	5
+
+#ifdef CONFIG_DEBUG_FS
+int omap_debugfs_init(struct drm_minor *minor);
+void omap_debugfs_cleanup(struct drm_minor *minor);
+void omap_framebuffer_describe(struct drm_framebuffer *fb, struct seq_file *m);
+void omap_gem_describe(struct drm_gem_object *obj, struct seq_file *m);
+void omap_gem_describe_objects(struct list_head *list, struct seq_file *m);
+#endif
+
+#ifdef CONFIG_PM
+int omap_gem_resume(struct device *dev);
+#endif
+
+int omap_irq_enable_vblank(struct drm_device *dev, int crtc);
+void omap_irq_disable_vblank(struct drm_device *dev, int crtc);
+irqreturn_t omap_irq_handler(DRM_IRQ_ARGS);
+void omap_irq_preinstall(struct drm_device *dev);
+int omap_irq_postinstall(struct drm_device *dev);
+void omap_irq_uninstall(struct drm_device *dev);
+void omap_irq_register(struct drm_device *dev, struct omap_drm_irq *irq);
+void omap_irq_unregister(struct drm_device *dev, struct omap_drm_irq *irq);
+int omap_drm_irq_uninstall(struct drm_device *dev);
+int omap_drm_irq_install(struct drm_device *dev);
+
+struct drm_fb_helper *omap_fbdev_init(struct drm_device *dev);
+void omap_fbdev_free(struct drm_device *dev);
+
+const struct omap_video_timings *omap_crtc_timings(struct drm_crtc *crtc);
+enum omap_channel omap_crtc_channel(struct drm_crtc *crtc);
+int omap_crtc_apply(struct drm_crtc *crtc,
+		struct omap_drm_apply *apply);
+struct drm_crtc *omap_crtc_init(struct drm_device *dev,
+		struct drm_plane *plane, enum omap_channel channel, int id);
+
+struct drm_plane *omap_plane_init(struct drm_device *dev,
+		int plane_id, bool private_plane);
+int omap_plane_dpms(struct drm_plane *plane, int mode);
+int omap_plane_mode_set(struct drm_plane *plane,
+		struct drm_crtc *crtc, struct drm_framebuffer *fb,
+		int crtc_x, int crtc_y,
+		unsigned int crtc_w, unsigned int crtc_h,
+		uint32_t src_x, uint32_t src_y,
+		uint32_t src_w, uint32_t src_h,
+		void (*fxn)(void *), void *arg);
+void omap_plane_install_properties(struct drm_plane *plane,
+		struct drm_mode_object *obj);
+int omap_plane_set_property(struct drm_plane *plane,
+		struct drm_property *property, uint64_t val);
+
+struct drm_encoder *omap_encoder_init(struct drm_device *dev,
+		struct omap_dss_device *dssdev);
+int omap_encoder_set_enabled(struct drm_encoder *encoder, bool enabled);
+int omap_encoder_update(struct drm_encoder *encoder,
+		struct omap_overlay_manager *mgr,
+		struct omap_video_timings *timings);
+
+struct drm_connector *omap_connector_init(struct drm_device *dev,
+		int connector_type, struct omap_dss_device *dssdev,
+		struct drm_encoder *encoder);
+struct drm_encoder *omap_connector_attached_encoder(
+		struct drm_connector *connector);
+void omap_connector_flush(struct drm_connector *connector,
+		int x, int y, int w, int h);
+
+void copy_timings_omap_to_drm(struct drm_display_mode *mode,
+		struct omap_video_timings *timings);
+void copy_timings_drm_to_omap(struct omap_video_timings *timings,
+		struct drm_display_mode *mode);
+
+uint32_t omap_framebuffer_get_formats(uint32_t *pixel_formats,
+		uint32_t max_formats, enum omap_color_mode supported_modes);
+struct drm_framebuffer *omap_framebuffer_create(struct drm_device *dev,
+		struct drm_file *file, struct drm_mode_fb_cmd2 *mode_cmd);
+struct drm_framebuffer *omap_framebuffer_init(struct drm_device *dev,
+		struct drm_mode_fb_cmd2 *mode_cmd, struct drm_gem_object **bos);
+struct drm_gem_object *omap_framebuffer_bo(struct drm_framebuffer *fb, int p);
+int omap_framebuffer_replace(struct drm_framebuffer *a,
+		struct drm_framebuffer *b, void *arg,
+		void (*unpin)(void *arg, struct drm_gem_object *bo));
+void omap_framebuffer_update_scanout(struct drm_framebuffer *fb,
+		struct omap_drm_window *win, struct omap_overlay_info *info);
+struct drm_connector *omap_framebuffer_get_next_connector(
+		struct drm_framebuffer *fb, struct drm_connector *from);
+void omap_framebuffer_flush(struct drm_framebuffer *fb,
+		int x, int y, int w, int h);
+
+void omap_gem_init(struct drm_device *dev);
+void omap_gem_deinit(struct drm_device *dev);
+
+struct drm_gem_object *omap_gem_new(struct drm_device *dev,
+		union omap_gem_size gsize, uint32_t flags);
+int omap_gem_new_handle(struct drm_device *dev, struct drm_file *file,
+		union omap_gem_size gsize, uint32_t flags, uint32_t *handle);
+void omap_gem_free_object(struct drm_gem_object *obj);
+int omap_gem_init_object(struct drm_gem_object *obj);
+void *omap_gem_vaddr(struct drm_gem_object *obj);
+int omap_gem_dumb_map_offset(struct drm_file *file, struct drm_device *dev,
+		uint32_t handle, uint64_t *offset);
+int omap_gem_dumb_destroy(struct drm_file *file, struct drm_device *dev,
+		uint32_t handle);
+int omap_gem_dumb_create(struct drm_file *file, struct drm_device *dev,
+		struct drm_mode_create_dumb *args);
+int omap_gem_mmap(struct file *filp, struct vm_area_struct *vma);
+int omap_gem_mmap_obj(struct drm_gem_object *obj,
+		struct vm_area_struct *vma);
+int omap_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf);
+int omap_gem_op_start(struct drm_gem_object *obj, enum omap_gem_op op);
+int omap_gem_op_finish(struct drm_gem_object *obj, enum omap_gem_op op);
+int omap_gem_op_sync(struct drm_gem_object *obj, enum omap_gem_op op);
+int omap_gem_op_async(struct drm_gem_object *obj, enum omap_gem_op op,
+		void (*fxn)(void *arg), void *arg);
+int omap_gem_roll(struct drm_gem_object *obj, uint32_t roll);
+void omap_gem_cpu_sync(struct drm_gem_object *obj, int pgoff);
+void omap_gem_dma_sync(struct drm_gem_object *obj,
+		enum dma_data_direction dir);
+int omap_gem_get_paddr(struct drm_gem_object *obj,
+		dma_addr_t *paddr, bool remap);
+int omap_gem_put_paddr(struct drm_gem_object *obj);
+int omap_gem_get_pages(struct drm_gem_object *obj, struct page ***pages,
+		bool remap);
+int omap_gem_put_pages(struct drm_gem_object *obj);
+uint32_t omap_gem_flags(struct drm_gem_object *obj);
+int omap_gem_rotated_paddr(struct drm_gem_object *obj, uint32_t orient,
+		int x, int y, dma_addr_t *paddr);
+uint64_t omap_gem_mmap_offset(struct drm_gem_object *obj);
+size_t omap_gem_mmap_size(struct drm_gem_object *obj);
+int omap_gem_tiled_size(struct drm_gem_object *obj, uint16_t *w, uint16_t *h);
+int omap_gem_tiled_stride(struct drm_gem_object *obj, uint32_t orient);
+
+struct dma_buf *omap_gem_prime_export(struct drm_device *dev,
+		struct drm_gem_object *obj, int flags);
+struct drm_gem_object *omap_gem_prime_import(struct drm_device *dev,
+		struct dma_buf *buffer);
+
+static inline int align_pitch(int pitch, int width, int bpp)
+{
+	int bytespp = (bpp + 7) / 8;
+	/* in case someone tries to feed us a completely bogus stride: */
+	pitch = max(pitch, width * bytespp);
+	/* PVR needs alignment to 8 pixels.. right now that is the most
+	 * restrictive stride requirement..
+	 */
+	return ALIGN(pitch, 8 * bytespp);
+}
+
+static inline enum omap_channel pipe2chan(int pipe)
+{
+	int num_mgrs = dss_feat_get_num_mgrs();
+
+	/*
+	 * We usually don't want to create a CRTC for each manager,
+	 * at least not until we have a way to expose private planes
+	 * to userspace.  Otherwise there would not be enough video
+	 * pipes left for drm planes.  The higher #'d managers tend
+	 * to have more features so start in reverse order.
+	 */
+	return num_mgrs - pipe - 1;
+}
+
+/* map crtc to vblank mask */
+static inline uint32_t pipe2vbl(int crtc)
+{
+	enum omap_channel channel = pipe2chan(crtc);
+	return dispc_mgr_get_vsync_irq(channel);
+}
+
+static inline int crtc2pipe(struct drm_device *dev, struct drm_crtc *crtc)
+{
+	struct omap_drm_private *priv = dev->dev_private;
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(priv->crtcs); i++)
+		if (priv->crtcs[i] == crtc)
+			return i;
+
+	BUG();  /* bogus CRTC ptr */
+	return -1;
+}
+
+/* should these be made into common util helpers?
+ */
+
+static inline int objects_lookup(struct drm_device *dev,
+		struct drm_file *filp, uint32_t pixel_format,
+		struct drm_gem_object **bos, uint32_t *handles)
+{
+	int i, n = drm_format_num_planes(pixel_format);
+
+	for (i = 0; i < n; i++) {
+		bos[i] = drm_gem_object_lookup(dev, filp, handles[i]);
+		if (!bos[i])
+			goto fail;
+
+	}
+
+	return 0;
+
+fail:
+	while (--i > 0)
+		drm_gem_object_unreference_unlocked(bos[i]);
+
+	return -ENOENT;
+}
+
+#endif /* __OMAP_DRV_H__ */
diff --git a/drivers/gpu/drm/omapdrm/omap_encoder.c b/drivers/gpu/drm/omapdrm/omap_encoder.c
new file mode 100644
index 0000000..7e1f2ab
--- /dev/null
+++ b/drivers/gpu/drm/omapdrm/omap_encoder.c
@@ -0,0 +1,170 @@
+/*
+ * drivers/gpu/drm/omapdrm/omap_encoder.c
+ *
+ * Copyright (C) 2011 Texas Instruments
+ * Author: Rob Clark <rob@ti.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "omap_drv.h"
+
+#include "drm_crtc.h"
+#include "drm_crtc_helper.h"
+
+#include <linux/list.h>
+
+
+/*
+ * encoder funcs
+ */
+
+#define to_omap_encoder(x) container_of(x, struct omap_encoder, base)
+
+/* The encoder and connector both map to same dssdev.. the encoder
+ * handles the 'active' parts, ie. anything the modifies the state
+ * of the hw, and the connector handles the 'read-only' parts, like
+ * detecting connection and reading edid.
+ */
+struct omap_encoder {
+	struct drm_encoder base;
+	struct omap_dss_device *dssdev;
+};
+
+static void omap_encoder_destroy(struct drm_encoder *encoder)
+{
+	struct omap_encoder *omap_encoder = to_omap_encoder(encoder);
+	drm_encoder_cleanup(encoder);
+	kfree(omap_encoder);
+}
+
+static const struct drm_encoder_funcs omap_encoder_funcs = {
+	.destroy = omap_encoder_destroy,
+};
+
+/*
+ * The CRTC drm_crtc_helper_set_mode() doesn't really give us the right
+ * order.. the easiest way to work around this for now is to make all
+ * the encoder-helper's no-op's and have the omap_crtc code take care
+ * of the sequencing and call us in the right points.
+ *
+ * Eventually to handle connecting CRTCs to different encoders properly,
+ * either the CRTC helpers need to change or we need to replace
+ * drm_crtc_helper_set_mode(), but lets wait until atomic-modeset for
+ * that.
+ */
+
+static void omap_encoder_dpms(struct drm_encoder *encoder, int mode)
+{
+}
+
+static bool omap_encoder_mode_fixup(struct drm_encoder *encoder,
+				  const struct drm_display_mode *mode,
+				  struct drm_display_mode *adjusted_mode)
+{
+	return true;
+}
+
+static void omap_encoder_mode_set(struct drm_encoder *encoder,
+				struct drm_display_mode *mode,
+				struct drm_display_mode *adjusted_mode)
+{
+}
+
+static void omap_encoder_prepare(struct drm_encoder *encoder)
+{
+}
+
+static void omap_encoder_commit(struct drm_encoder *encoder)
+{
+}
+
+static const struct drm_encoder_helper_funcs omap_encoder_helper_funcs = {
+	.dpms = omap_encoder_dpms,
+	.mode_fixup = omap_encoder_mode_fixup,
+	.mode_set = omap_encoder_mode_set,
+	.prepare = omap_encoder_prepare,
+	.commit = omap_encoder_commit,
+};
+
+/*
+ * Instead of relying on the helpers for modeset, the omap_crtc code
+ * calls these functions in the proper sequence.
+ */
+
+int omap_encoder_set_enabled(struct drm_encoder *encoder, bool enabled)
+{
+	struct omap_encoder *omap_encoder = to_omap_encoder(encoder);
+	struct omap_dss_device *dssdev = omap_encoder->dssdev;
+	struct omap_dss_driver *dssdrv = dssdev->driver;
+
+	if (enabled) {
+		return dssdrv->enable(dssdev);
+	} else {
+		dssdrv->disable(dssdev);
+		return 0;
+	}
+}
+
+int omap_encoder_update(struct drm_encoder *encoder,
+		struct omap_overlay_manager *mgr,
+		struct omap_video_timings *timings)
+{
+	struct drm_device *dev = encoder->dev;
+	struct omap_encoder *omap_encoder = to_omap_encoder(encoder);
+	struct omap_dss_device *dssdev = omap_encoder->dssdev;
+	struct omap_dss_driver *dssdrv = dssdev->driver;
+	int ret;
+
+	dssdev->output->manager = mgr;
+
+	ret = dssdrv->check_timings(dssdev, timings);
+	if (ret) {
+		dev_err(dev->dev, "could not set timings: %d\n", ret);
+		return ret;
+	}
+
+	dssdrv->set_timings(dssdev, timings);
+
+	return 0;
+}
+
+/* initialize encoder */
+struct drm_encoder *omap_encoder_init(struct drm_device *dev,
+		struct omap_dss_device *dssdev)
+{
+	struct drm_encoder *encoder = NULL;
+	struct omap_encoder *omap_encoder;
+
+	omap_encoder = kzalloc(sizeof(*omap_encoder), GFP_KERNEL);
+	if (!omap_encoder) {
+		dev_err(dev->dev, "could not allocate encoder\n");
+		goto fail;
+	}
+
+	omap_encoder->dssdev = dssdev;
+
+	encoder = &omap_encoder->base;
+
+	drm_encoder_init(dev, encoder, &omap_encoder_funcs,
+			 DRM_MODE_ENCODER_TMDS);
+	drm_encoder_helper_add(encoder, &omap_encoder_helper_funcs);
+
+	return encoder;
+
+fail:
+	if (encoder)
+		omap_encoder_destroy(encoder);
+
+	return NULL;
+}
diff --git a/drivers/gpu/drm/omapdrm/omap_fb.c b/drivers/gpu/drm/omapdrm/omap_fb.c
new file mode 100644
index 0000000..9d5f6f6
--- /dev/null
+++ b/drivers/gpu/drm/omapdrm/omap_fb.c
@@ -0,0 +1,472 @@
+/*
+ * drivers/gpu/drm/omapdrm/omap_fb.c
+ *
+ * Copyright (C) 2011 Texas Instruments
+ * Author: Rob Clark <rob@ti.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "omap_drv.h"
+#include "omap_dmm_tiler.h"
+
+#include "drm_crtc.h"
+#include "drm_crtc_helper.h"
+
+/*
+ * framebuffer funcs
+ */
+
+/* per-format info: */
+struct format {
+	enum omap_color_mode dss_format;
+	uint32_t pixel_format;
+	struct {
+		int stride_bpp;           /* this times width is stride */
+		int sub_y;                /* sub-sample in y dimension */
+	} planes[4];
+	bool yuv;
+};
+
+static const struct format formats[] = {
+	/* 16bpp [A]RGB: */
+	{ OMAP_DSS_COLOR_RGB16,       DRM_FORMAT_RGB565,   {{2, 1}}, false }, /* RGB16-565 */
+	{ OMAP_DSS_COLOR_RGB12U,      DRM_FORMAT_RGBX4444, {{2, 1}}, false }, /* RGB12x-4444 */
+	{ OMAP_DSS_COLOR_RGBX16,      DRM_FORMAT_XRGB4444, {{2, 1}}, false }, /* xRGB12-4444 */
+	{ OMAP_DSS_COLOR_RGBA16,      DRM_FORMAT_RGBA4444, {{2, 1}}, false }, /* RGBA12-4444 */
+	{ OMAP_DSS_COLOR_ARGB16,      DRM_FORMAT_ARGB4444, {{2, 1}}, false }, /* ARGB16-4444 */
+	{ OMAP_DSS_COLOR_XRGB16_1555, DRM_FORMAT_XRGB1555, {{2, 1}}, false }, /* xRGB15-1555 */
+	{ OMAP_DSS_COLOR_ARGB16_1555, DRM_FORMAT_ARGB1555, {{2, 1}}, false }, /* ARGB16-1555 */
+	/* 24bpp RGB: */
+	{ OMAP_DSS_COLOR_RGB24P,      DRM_FORMAT_RGB888,   {{3, 1}}, false }, /* RGB24-888 */
+	/* 32bpp [A]RGB: */
+	{ OMAP_DSS_COLOR_RGBX32,      DRM_FORMAT_RGBX8888, {{4, 1}}, false }, /* RGBx24-8888 */
+	{ OMAP_DSS_COLOR_RGB24U,      DRM_FORMAT_XRGB8888, {{4, 1}}, false }, /* xRGB24-8888 */
+	{ OMAP_DSS_COLOR_RGBA32,      DRM_FORMAT_RGBA8888, {{4, 1}}, false }, /* RGBA32-8888 */
+	{ OMAP_DSS_COLOR_ARGB32,      DRM_FORMAT_ARGB8888, {{4, 1}}, false }, /* ARGB32-8888 */
+	/* YUV: */
+	{ OMAP_DSS_COLOR_NV12,        DRM_FORMAT_NV12,     {{1, 1}, {1, 2}}, true },
+	{ OMAP_DSS_COLOR_YUV2,        DRM_FORMAT_YUYV,     {{2, 1}}, true },
+	{ OMAP_DSS_COLOR_UYVY,        DRM_FORMAT_UYVY,     {{2, 1}}, true },
+};
+
+/* convert from overlay's pixel formats bitmask to an array of fourcc's */
+uint32_t omap_framebuffer_get_formats(uint32_t *pixel_formats,
+		uint32_t max_formats, enum omap_color_mode supported_modes)
+{
+	uint32_t nformats = 0;
+	int i = 0;
+
+	for (i = 0; i < ARRAY_SIZE(formats) && nformats < max_formats; i++)
+		if (formats[i].dss_format & supported_modes)
+			pixel_formats[nformats++] = formats[i].pixel_format;
+
+	return nformats;
+}
+
+/* per-plane info for the fb: */
+struct plane {
+	struct drm_gem_object *bo;
+	uint32_t pitch;
+	uint32_t offset;
+	dma_addr_t paddr;
+};
+
+#define to_omap_framebuffer(x) container_of(x, struct omap_framebuffer, base)
+
+struct omap_framebuffer {
+	struct drm_framebuffer base;
+	const struct format *format;
+	struct plane planes[4];
+};
+
+static int omap_framebuffer_create_handle(struct drm_framebuffer *fb,
+		struct drm_file *file_priv,
+		unsigned int *handle)
+{
+	struct omap_framebuffer *omap_fb = to_omap_framebuffer(fb);
+	return drm_gem_handle_create(file_priv,
+			omap_fb->planes[0].bo, handle);
+}
+
+static void omap_framebuffer_destroy(struct drm_framebuffer *fb)
+{
+	struct omap_framebuffer *omap_fb = to_omap_framebuffer(fb);
+	int i, n = drm_format_num_planes(fb->pixel_format);
+
+	DBG("destroy: FB ID: %d (%p)", fb->base.id, fb);
+
+	drm_framebuffer_cleanup(fb);
+
+	for (i = 0; i < n; i++) {
+		struct plane *plane = &omap_fb->planes[i];
+		if (plane->bo)
+			drm_gem_object_unreference_unlocked(plane->bo);
+	}
+
+	kfree(omap_fb);
+}
+
+static int omap_framebuffer_dirty(struct drm_framebuffer *fb,
+		struct drm_file *file_priv, unsigned flags, unsigned color,
+		struct drm_clip_rect *clips, unsigned num_clips)
+{
+	int i;
+
+	for (i = 0; i < num_clips; i++) {
+		omap_framebuffer_flush(fb, clips[i].x1, clips[i].y1,
+					clips[i].x2 - clips[i].x1,
+					clips[i].y2 - clips[i].y1);
+	}
+
+	return 0;
+}
+
+static const struct drm_framebuffer_funcs omap_framebuffer_funcs = {
+	.create_handle = omap_framebuffer_create_handle,
+	.destroy = omap_framebuffer_destroy,
+	.dirty = omap_framebuffer_dirty,
+};
+
+static uint32_t get_linear_addr(struct plane *plane,
+		const struct format *format, int n, int x, int y)
+{
+	uint32_t offset;
+
+	offset = plane->offset +
+			(x * format->planes[n].stride_bpp) +
+			(y * plane->pitch / format->planes[n].sub_y);
+
+	return plane->paddr + offset;
+}
+
+/* update ovl info for scanout, handles cases of multi-planar fb's, etc.
+ */
+void omap_framebuffer_update_scanout(struct drm_framebuffer *fb,
+		struct omap_drm_window *win, struct omap_overlay_info *info)
+{
+	struct omap_framebuffer *omap_fb = to_omap_framebuffer(fb);
+	const struct format *format = omap_fb->format;
+	struct plane *plane = &omap_fb->planes[0];
+	uint32_t x, y, orient = 0;
+
+	info->color_mode = format->dss_format;
+
+	info->pos_x      = win->crtc_x;
+	info->pos_y      = win->crtc_y;
+	info->out_width  = win->crtc_w;
+	info->out_height = win->crtc_h;
+	info->width      = win->src_w;
+	info->height     = win->src_h;
+
+	x = win->src_x;
+	y = win->src_y;
+
+	if (omap_gem_flags(plane->bo) & OMAP_BO_TILED) {
+		uint32_t w = win->src_w;
+		uint32_t h = win->src_h;
+
+		switch (win->rotation & 0xf) {
+		default:
+			dev_err(fb->dev->dev, "invalid rotation: %02x",
+					(uint32_t)win->rotation);
+			/* fallthru to default to no rotation */
+		case 0:
+		case BIT(DRM_ROTATE_0):
+			orient = 0;
+			break;
+		case BIT(DRM_ROTATE_90):
+			orient = MASK_XY_FLIP | MASK_X_INVERT;
+			break;
+		case BIT(DRM_ROTATE_180):
+			orient = MASK_X_INVERT | MASK_Y_INVERT;
+			break;
+		case BIT(DRM_ROTATE_270):
+			orient = MASK_XY_FLIP | MASK_Y_INVERT;
+			break;
+		}
+
+		if (win->rotation & BIT(DRM_REFLECT_X))
+			orient ^= MASK_X_INVERT;
+
+		if (win->rotation & BIT(DRM_REFLECT_Y))
+			orient ^= MASK_Y_INVERT;
+
+		/* adjust x,y offset for flip/invert: */
+		if (orient & MASK_XY_FLIP)
+			swap(w, h);
+		if (orient & MASK_Y_INVERT)
+			y += h - 1;
+		if (orient & MASK_X_INVERT)
+			x += w - 1;
+
+		omap_gem_rotated_paddr(plane->bo, orient, x, y, &info->paddr);
+		info->rotation_type = OMAP_DSS_ROT_TILER;
+		info->screen_width  = omap_gem_tiled_stride(plane->bo, orient);
+	} else {
+		info->paddr         = get_linear_addr(plane, format, 0, x, y);
+		info->rotation_type = OMAP_DSS_ROT_DMA;
+		info->screen_width  = plane->pitch;
+	}
+
+	/* convert to pixels: */
+	info->screen_width /= format->planes[0].stride_bpp;
+
+	if (format->dss_format == OMAP_DSS_COLOR_NV12) {
+		plane = &omap_fb->planes[1];
+
+		if (info->rotation_type == OMAP_DSS_ROT_TILER) {
+			WARN_ON(!(omap_gem_flags(plane->bo) & OMAP_BO_TILED));
+			omap_gem_rotated_paddr(plane->bo, orient,
+					x/2, y/2, &info->p_uv_addr);
+		} else {
+			info->p_uv_addr = get_linear_addr(plane, format, 1, x, y);
+		}
+	} else {
+		info->p_uv_addr = 0;
+	}
+}
+
+/* Call for unpin 'a' (if not NULL), and pin 'b' (if not NULL).  Although
+ * buffers to unpin are just pushed to the unpin fifo so that the
+ * caller can defer unpin until vblank.
+ *
+ * Note if this fails (ie. something went very wrong!), all buffers are
+ * unpinned, and the caller disables the overlay.  We could have tried
+ * to revert back to the previous set of pinned buffers but if things are
+ * hosed there is no guarantee that would succeed.
+ */
+int omap_framebuffer_replace(struct drm_framebuffer *a,
+		struct drm_framebuffer *b, void *arg,
+		void (*unpin)(void *arg, struct drm_gem_object *bo))
+{
+	int ret = 0, i, na, nb;
+	struct omap_framebuffer *ofba = to_omap_framebuffer(a);
+	struct omap_framebuffer *ofbb = to_omap_framebuffer(b);
+	uint32_t pinned_mask = 0;
+
+	na = a ? drm_format_num_planes(a->pixel_format) : 0;
+	nb = b ? drm_format_num_planes(b->pixel_format) : 0;
+
+	for (i = 0; i < max(na, nb); i++) {
+		struct plane *pa, *pb;
+
+		pa = (i < na) ? &ofba->planes[i] : NULL;
+		pb = (i < nb) ? &ofbb->planes[i] : NULL;
+
+		if (pa)
+			unpin(arg, pa->bo);
+
+		if (pb && !ret) {
+			ret = omap_gem_get_paddr(pb->bo, &pb->paddr, true);
+			if (!ret) {
+				omap_gem_dma_sync(pb->bo, DMA_TO_DEVICE);
+				pinned_mask |= (1 << i);
+			}
+		}
+	}
+
+	if (ret) {
+		/* something went wrong.. unpin what has been pinned */
+		for (i = 0; i < nb; i++) {
+			if (pinned_mask & (1 << i)) {
+				struct plane *pb = &ofba->planes[i];
+				unpin(arg, pb->bo);
+			}
+		}
+	}
+
+	return ret;
+}
+
+struct drm_gem_object *omap_framebuffer_bo(struct drm_framebuffer *fb, int p)
+{
+	struct omap_framebuffer *omap_fb = to_omap_framebuffer(fb);
+	if (p >= drm_format_num_planes(fb->pixel_format))
+		return NULL;
+	return omap_fb->planes[p].bo;
+}
+
+/* iterate thru all the connectors, returning ones that are attached
+ * to the same fb..
+ */
+struct drm_connector *omap_framebuffer_get_next_connector(
+		struct drm_framebuffer *fb, struct drm_connector *from)
+{
+	struct drm_device *dev = fb->dev;
+	struct list_head *connector_list = &dev->mode_config.connector_list;
+	struct drm_connector *connector = from;
+
+	if (!from)
+		return list_first_entry(connector_list, typeof(*from), head);
+
+	list_for_each_entry_from(connector, connector_list, head) {
+		if (connector != from) {
+			struct drm_encoder *encoder = connector->encoder;
+			struct drm_crtc *crtc = encoder ? encoder->crtc : NULL;
+			if (crtc && crtc->fb == fb)
+				return connector;
+
+		}
+	}
+
+	return NULL;
+}
+
+/* flush an area of the framebuffer (in case of manual update display that
+ * is not automatically flushed)
+ */
+void omap_framebuffer_flush(struct drm_framebuffer *fb,
+		int x, int y, int w, int h)
+{
+	struct drm_connector *connector = NULL;
+
+	VERB("flush: %d,%d %dx%d, fb=%p", x, y, w, h, fb);
+
+	while ((connector = omap_framebuffer_get_next_connector(fb, connector))) {
+		/* only consider connectors that are part of a chain */
+		if (connector->encoder && connector->encoder->crtc) {
+			/* TODO: maybe this should propagate thru the crtc who
+			 * could do the coordinate translation..
+			 */
+			struct drm_crtc *crtc = connector->encoder->crtc;
+			int cx = max(0, x - crtc->x);
+			int cy = max(0, y - crtc->y);
+			int cw = w + (x - crtc->x) - cx;
+			int ch = h + (y - crtc->y) - cy;
+
+			omap_connector_flush(connector, cx, cy, cw, ch);
+		}
+	}
+}
+
+#ifdef CONFIG_DEBUG_FS
+void omap_framebuffer_describe(struct drm_framebuffer *fb, struct seq_file *m)
+{
+	struct omap_framebuffer *omap_fb = to_omap_framebuffer(fb);
+	int i, n = drm_format_num_planes(fb->pixel_format);
+
+	seq_printf(m, "fb: %dx%d@%4.4s\n", fb->width, fb->height,
+			(char *)&fb->pixel_format);
+
+	for (i = 0; i < n; i++) {
+		struct plane *plane = &omap_fb->planes[i];
+		seq_printf(m, "   %d: offset=%d pitch=%d, obj: ",
+				i, plane->offset, plane->pitch);
+		omap_gem_describe(plane->bo, m);
+	}
+}
+#endif
+
+struct drm_framebuffer *omap_framebuffer_create(struct drm_device *dev,
+		struct drm_file *file, struct drm_mode_fb_cmd2 *mode_cmd)
+{
+	struct drm_gem_object *bos[4];
+	struct drm_framebuffer *fb;
+	int ret;
+
+	ret = objects_lookup(dev, file, mode_cmd->pixel_format,
+			bos, mode_cmd->handles);
+	if (ret)
+		return ERR_PTR(ret);
+
+	fb = omap_framebuffer_init(dev, mode_cmd, bos);
+	if (IS_ERR(fb)) {
+		int i, n = drm_format_num_planes(mode_cmd->pixel_format);
+		for (i = 0; i < n; i++)
+			drm_gem_object_unreference_unlocked(bos[i]);
+		return fb;
+	}
+	return fb;
+}
+
+struct drm_framebuffer *omap_framebuffer_init(struct drm_device *dev,
+		struct drm_mode_fb_cmd2 *mode_cmd, struct drm_gem_object **bos)
+{
+	struct omap_framebuffer *omap_fb;
+	struct drm_framebuffer *fb = NULL;
+	const struct format *format = NULL;
+	int ret, i, n = drm_format_num_planes(mode_cmd->pixel_format);
+
+	DBG("create framebuffer: dev=%p, mode_cmd=%p (%dx%d@%4.4s)",
+			dev, mode_cmd, mode_cmd->width, mode_cmd->height,
+			(char *)&mode_cmd->pixel_format);
+
+	for (i = 0; i < ARRAY_SIZE(formats); i++) {
+		if (formats[i].pixel_format == mode_cmd->pixel_format) {
+			format = &formats[i];
+			break;
+		}
+	}
+
+	if (!format) {
+		dev_err(dev->dev, "unsupported pixel format: %4.4s\n",
+				(char *)&mode_cmd->pixel_format);
+		ret = -EINVAL;
+		goto fail;
+	}
+
+	omap_fb = kzalloc(sizeof(*omap_fb), GFP_KERNEL);
+	if (!omap_fb) {
+		dev_err(dev->dev, "could not allocate fb\n");
+		ret = -ENOMEM;
+		goto fail;
+	}
+
+	fb = &omap_fb->base;
+	omap_fb->format = format;
+
+	for (i = 0; i < n; i++) {
+		struct plane *plane = &omap_fb->planes[i];
+		int size, pitch = mode_cmd->pitches[i];
+
+		if (pitch < (mode_cmd->width * format->planes[i].stride_bpp)) {
+			dev_err(dev->dev, "provided buffer pitch is too small! %d < %d\n",
+					pitch, mode_cmd->width * format->planes[i].stride_bpp);
+			ret = -EINVAL;
+			goto fail;
+		}
+
+		size = pitch * mode_cmd->height / format->planes[i].sub_y;
+
+		if (size > (omap_gem_mmap_size(bos[i]) - mode_cmd->offsets[i])) {
+			dev_err(dev->dev, "provided buffer object is too small! %d < %d\n",
+					bos[i]->size - mode_cmd->offsets[i], size);
+			ret = -EINVAL;
+			goto fail;
+		}
+
+		plane->bo     = bos[i];
+		plane->offset = mode_cmd->offsets[i];
+		plane->pitch  = pitch;
+		plane->paddr  = 0;
+	}
+
+	drm_helper_mode_fill_fb_struct(fb, mode_cmd);
+
+	ret = drm_framebuffer_init(dev, fb, &omap_framebuffer_funcs);
+	if (ret) {
+		dev_err(dev->dev, "framebuffer init failed: %d\n", ret);
+		goto fail;
+	}
+
+	DBG("create: FB ID: %d (%p)", fb->base.id, fb);
+
+	return fb;
+
+fail:
+	if (fb)
+		omap_framebuffer_destroy(fb);
+
+	return ERR_PTR(ret);
+}
diff --git a/drivers/gpu/drm/omapdrm/omap_fbdev.c b/drivers/gpu/drm/omapdrm/omap_fbdev.c
new file mode 100644
index 0000000..11eed30
--- /dev/null
+++ b/drivers/gpu/drm/omapdrm/omap_fbdev.c
@@ -0,0 +1,402 @@
+/*
+ * drivers/gpu/drm/omapdrm/omap_fbdev.c
+ *
+ * Copyright (C) 2011 Texas Instruments
+ * Author: Rob Clark <rob@ti.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "omap_drv.h"
+
+#include "drm_crtc.h"
+#include "drm_fb_helper.h"
+
+MODULE_PARM_DESC(ywrap, "Enable ywrap scrolling (omap44xx and later, default 'y')");
+static bool ywrap_enabled = true;
+module_param_named(ywrap, ywrap_enabled, bool, 0644);
+
+/*
+ * fbdev funcs, to implement legacy fbdev interface on top of drm driver
+ */
+
+#define to_omap_fbdev(x) container_of(x, struct omap_fbdev, base)
+
+struct omap_fbdev {
+	struct drm_fb_helper base;
+	struct drm_framebuffer *fb;
+	struct drm_gem_object *bo;
+	bool ywrap_enabled;
+
+	/* for deferred dmm roll when getting called in atomic ctx */
+	struct work_struct work;
+};
+
+static void omap_fbdev_flush(struct fb_info *fbi, int x, int y, int w, int h);
+static struct drm_fb_helper *get_fb(struct fb_info *fbi);
+
+static ssize_t omap_fbdev_write(struct fb_info *fbi, const char __user *buf,
+		size_t count, loff_t *ppos)
+{
+	ssize_t res;
+
+	res = fb_sys_write(fbi, buf, count, ppos);
+	omap_fbdev_flush(fbi, 0, 0, fbi->var.xres, fbi->var.yres);
+
+	return res;
+}
+
+static void omap_fbdev_fillrect(struct fb_info *fbi,
+		const struct fb_fillrect *rect)
+{
+	sys_fillrect(fbi, rect);
+	omap_fbdev_flush(fbi, rect->dx, rect->dy, rect->width, rect->height);
+}
+
+static void omap_fbdev_copyarea(struct fb_info *fbi,
+		const struct fb_copyarea *area)
+{
+	sys_copyarea(fbi, area);
+	omap_fbdev_flush(fbi, area->dx, area->dy, area->width, area->height);
+}
+
+static void omap_fbdev_imageblit(struct fb_info *fbi,
+		const struct fb_image *image)
+{
+	sys_imageblit(fbi, image);
+	omap_fbdev_flush(fbi, image->dx, image->dy,
+				image->width, image->height);
+}
+
+static void pan_worker(struct work_struct *work)
+{
+	struct omap_fbdev *fbdev = container_of(work, struct omap_fbdev, work);
+	struct fb_info *fbi = fbdev->base.fbdev;
+	int npages;
+
+	/* DMM roll shifts in 4K pages: */
+	npages = fbi->fix.line_length >> PAGE_SHIFT;
+	omap_gem_roll(fbdev->bo, fbi->var.yoffset * npages);
+}
+
+static int omap_fbdev_pan_display(struct fb_var_screeninfo *var,
+		struct fb_info *fbi)
+{
+	struct drm_fb_helper *helper = get_fb(fbi);
+	struct omap_fbdev *fbdev = to_omap_fbdev(helper);
+
+	if (!helper)
+		goto fallback;
+
+	if (!fbdev->ywrap_enabled)
+		goto fallback;
+
+	if (drm_can_sleep()) {
+		pan_worker(&fbdev->work);
+	} else {
+		struct omap_drm_private *priv = helper->dev->dev_private;
+		queue_work(priv->wq, &fbdev->work);
+	}
+
+	return 0;
+
+fallback:
+	return drm_fb_helper_pan_display(var, fbi);
+}
+
+static struct fb_ops omap_fb_ops = {
+	.owner = THIS_MODULE,
+
+	/* Note: to properly handle manual update displays, we wrap the
+	 * basic fbdev ops which write to the framebuffer
+	 */
+	.fb_read = fb_sys_read,
+	.fb_write = omap_fbdev_write,
+	.fb_fillrect = omap_fbdev_fillrect,
+	.fb_copyarea = omap_fbdev_copyarea,
+	.fb_imageblit = omap_fbdev_imageblit,
+
+	.fb_check_var = drm_fb_helper_check_var,
+	.fb_set_par = drm_fb_helper_set_par,
+	.fb_pan_display = omap_fbdev_pan_display,
+	.fb_blank = drm_fb_helper_blank,
+	.fb_setcmap = drm_fb_helper_setcmap,
+
+	.fb_debug_enter = drm_fb_helper_debug_enter,
+	.fb_debug_leave = drm_fb_helper_debug_leave,
+};
+
+static int omap_fbdev_create(struct drm_fb_helper *helper,
+		struct drm_fb_helper_surface_size *sizes)
+{
+	struct omap_fbdev *fbdev = to_omap_fbdev(helper);
+	struct drm_device *dev = helper->dev;
+	struct omap_drm_private *priv = dev->dev_private;
+	struct drm_framebuffer *fb = NULL;
+	union omap_gem_size gsize;
+	struct fb_info *fbi = NULL;
+	struct drm_mode_fb_cmd2 mode_cmd = {0};
+	dma_addr_t paddr;
+	int ret;
+
+	/* only doing ARGB32 since this is what is needed to alpha-blend
+	 * with video overlays:
+	 */
+	sizes->surface_bpp = 32;
+	sizes->surface_depth = 32;
+
+	DBG("create fbdev: %dx%d@%d (%dx%d)", sizes->surface_width,
+			sizes->surface_height, sizes->surface_bpp,
+			sizes->fb_width, sizes->fb_height);
+
+	mode_cmd.pixel_format = drm_mode_legacy_fb_format(sizes->surface_bpp,
+			sizes->surface_depth);
+
+	mode_cmd.width = sizes->surface_width;
+	mode_cmd.height = sizes->surface_height;
+
+	mode_cmd.pitches[0] = align_pitch(
+			mode_cmd.width * ((sizes->surface_bpp + 7) / 8),
+			mode_cmd.width, sizes->surface_bpp);
+
+	fbdev->ywrap_enabled = priv->has_dmm && ywrap_enabled;
+	if (fbdev->ywrap_enabled) {
+		/* need to align pitch to page size if using DMM scrolling */
+		mode_cmd.pitches[0] = ALIGN(mode_cmd.pitches[0], PAGE_SIZE);
+	}
+
+	/* allocate backing bo */
+	gsize = (union omap_gem_size){
+		.bytes = PAGE_ALIGN(mode_cmd.pitches[0] * mode_cmd.height),
+	};
+	DBG("allocating %d bytes for fb %d", gsize.bytes, dev->primary->index);
+	fbdev->bo = omap_gem_new(dev, gsize, OMAP_BO_SCANOUT | OMAP_BO_WC);
+	if (!fbdev->bo) {
+		dev_err(dev->dev, "failed to allocate buffer object\n");
+		ret = -ENOMEM;
+		goto fail;
+	}
+
+	fb = omap_framebuffer_init(dev, &mode_cmd, &fbdev->bo);
+	if (IS_ERR(fb)) {
+		dev_err(dev->dev, "failed to allocate fb\n");
+		/* note: if fb creation failed, we can't rely on fb destroy
+		 * to unref the bo:
+		 */
+		drm_gem_object_unreference(fbdev->bo);
+		ret = PTR_ERR(fb);
+		goto fail;
+	}
+
+	/* note: this keeps the bo pinned.. which is perhaps not ideal,
+	 * but is needed as long as we use fb_mmap() to mmap to userspace
+	 * (since this happens using fix.smem_start).  Possibly we could
+	 * implement our own mmap using GEM mmap support to avoid this
+	 * (non-tiled buffer doesn't need to be pinned for fbcon to write
+	 * to it).  Then we just need to be sure that we are able to re-
+	 * pin it in case of an opps.
+	 */
+	ret = omap_gem_get_paddr(fbdev->bo, &paddr, true);
+	if (ret) {
+		dev_err(dev->dev,
+			"could not map (paddr)!  Skipping framebuffer alloc\n");
+		ret = -ENOMEM;
+		goto fail;
+	}
+
+	mutex_lock(&dev->struct_mutex);
+
+	fbi = framebuffer_alloc(0, dev->dev);
+	if (!fbi) {
+		dev_err(dev->dev, "failed to allocate fb info\n");
+		ret = -ENOMEM;
+		goto fail_unlock;
+	}
+
+	DBG("fbi=%p, dev=%p", fbi, dev);
+
+	fbdev->fb = fb;
+	helper->fb = fb;
+	helper->fbdev = fbi;
+
+	fbi->par = helper;
+	fbi->flags = FBINFO_DEFAULT;
+	fbi->fbops = &omap_fb_ops;
+
+	strcpy(fbi->fix.id, MODULE_NAME);
+
+	ret = fb_alloc_cmap(&fbi->cmap, 256, 0);
+	if (ret) {
+		ret = -ENOMEM;
+		goto fail_unlock;
+	}
+
+	drm_fb_helper_fill_fix(fbi, fb->pitches[0], fb->depth);
+	drm_fb_helper_fill_var(fbi, helper, sizes->fb_width, sizes->fb_height);
+
+	dev->mode_config.fb_base = paddr;
+
+	fbi->screen_base = omap_gem_vaddr(fbdev->bo);
+	fbi->screen_size = fbdev->bo->size;
+	fbi->fix.smem_start = paddr;
+	fbi->fix.smem_len = fbdev->bo->size;
+
+	/* if we have DMM, then we can use it for scrolling by just
+	 * shuffling pages around in DMM rather than doing sw blit.
+	 */
+	if (fbdev->ywrap_enabled) {
+		DRM_INFO("Enabling DMM ywrap scrolling\n");
+		fbi->flags |= FBINFO_HWACCEL_YWRAP | FBINFO_READS_FAST;
+		fbi->fix.ywrapstep = 1;
+	}
+
+
+	DBG("par=%p, %dx%d", fbi->par, fbi->var.xres, fbi->var.yres);
+	DBG("allocated %dx%d fb", fbdev->fb->width, fbdev->fb->height);
+
+	mutex_unlock(&dev->struct_mutex);
+
+	return 0;
+
+fail_unlock:
+	mutex_unlock(&dev->struct_mutex);
+fail:
+
+	if (ret) {
+		if (fbi)
+			framebuffer_release(fbi);
+		if (fb) {
+			drm_framebuffer_unregister_private(fb);
+			drm_framebuffer_remove(fb);
+		}
+	}
+
+	return ret;
+}
+
+static void omap_crtc_fb_gamma_set(struct drm_crtc *crtc,
+		u16 red, u16 green, u16 blue, int regno)
+{
+	DBG("fbdev: set gamma");
+}
+
+static void omap_crtc_fb_gamma_get(struct drm_crtc *crtc,
+		u16 *red, u16 *green, u16 *blue, int regno)
+{
+	DBG("fbdev: get gamma");
+}
+
+static struct drm_fb_helper_funcs omap_fb_helper_funcs = {
+	.gamma_set = omap_crtc_fb_gamma_set,
+	.gamma_get = omap_crtc_fb_gamma_get,
+	.fb_probe = omap_fbdev_create,
+};
+
+static struct drm_fb_helper *get_fb(struct fb_info *fbi)
+{
+	if (!fbi || strcmp(fbi->fix.id, MODULE_NAME)) {
+		/* these are not the fb's you're looking for */
+		return NULL;
+	}
+	return fbi->par;
+}
+
+/* flush an area of the framebuffer (in case of manual update display that
+ * is not automatically flushed)
+ */
+static void omap_fbdev_flush(struct fb_info *fbi, int x, int y, int w, int h)
+{
+	struct drm_fb_helper *helper = get_fb(fbi);
+
+	if (!helper)
+		return;
+
+	VERB("flush fbdev: %d,%d %dx%d, fbi=%p", x, y, w, h, fbi);
+
+	omap_framebuffer_flush(helper->fb, x, y, w, h);
+}
+
+/* initialize fbdev helper */
+struct drm_fb_helper *omap_fbdev_init(struct drm_device *dev)
+{
+	struct omap_drm_private *priv = dev->dev_private;
+	struct omap_fbdev *fbdev = NULL;
+	struct drm_fb_helper *helper;
+	int ret = 0;
+
+	fbdev = kzalloc(sizeof(*fbdev), GFP_KERNEL);
+	if (!fbdev) {
+		dev_err(dev->dev, "could not allocate fbdev\n");
+		goto fail;
+	}
+
+	INIT_WORK(&fbdev->work, pan_worker);
+
+	helper = &fbdev->base;
+
+	helper->funcs = &omap_fb_helper_funcs;
+
+	ret = drm_fb_helper_init(dev, helper,
+			priv->num_crtcs, priv->num_connectors);
+	if (ret) {
+		dev_err(dev->dev, "could not init fbdev: ret=%d\n", ret);
+		goto fail;
+	}
+
+	drm_fb_helper_single_add_all_connectors(helper);
+
+	/* disable all the possible outputs/crtcs before entering KMS mode */
+	drm_helper_disable_unused_functions(dev);
+
+	drm_fb_helper_initial_config(helper, 32);
+
+	priv->fbdev = helper;
+
+	return helper;
+
+fail:
+	kfree(fbdev);
+	return NULL;
+}
+
+void omap_fbdev_free(struct drm_device *dev)
+{
+	struct omap_drm_private *priv = dev->dev_private;
+	struct drm_fb_helper *helper = priv->fbdev;
+	struct omap_fbdev *fbdev;
+	struct fb_info *fbi;
+
+	DBG();
+
+	fbi = helper->fbdev;
+
+	/* only cleanup framebuffer if it is present */
+	if (fbi) {
+		unregister_framebuffer(fbi);
+		framebuffer_release(fbi);
+	}
+
+	drm_fb_helper_fini(helper);
+
+	fbdev = to_omap_fbdev(priv->fbdev);
+
+	/* this will free the backing object */
+	if (fbdev->fb) {
+		drm_framebuffer_unregister_private(fbdev->fb);
+		drm_framebuffer_remove(fbdev->fb);
+	}
+
+	kfree(fbdev);
+
+	priv->fbdev = NULL;
+}
diff --git a/drivers/gpu/drm/omapdrm/omap_gem.c b/drivers/gpu/drm/omapdrm/omap_gem.c
new file mode 100644
index 0000000..e8302b0
--- /dev/null
+++ b/drivers/gpu/drm/omapdrm/omap_gem.c
@@ -0,0 +1,1511 @@
+/*
+ * drivers/gpu/drm/omapdrm/omap_gem.c
+ *
+ * Copyright (C) 2011 Texas Instruments
+ * Author: Rob Clark <rob.clark@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+
+#include <linux/spinlock.h>
+#include <linux/shmem_fs.h>
+
+#include "omap_drv.h"
+#include "omap_dmm_tiler.h"
+
+/* remove these once drm core helpers are merged */
+struct page **_drm_gem_get_pages(struct drm_gem_object *obj, gfp_t gfpmask);
+void _drm_gem_put_pages(struct drm_gem_object *obj, struct page **pages,
+		bool dirty, bool accessed);
+int _drm_gem_create_mmap_offset_size(struct drm_gem_object *obj, size_t size);
+
+/*
+ * GEM buffer object implementation.
+ */
+
+#define to_omap_bo(x) container_of(x, struct omap_gem_object, base)
+
+/* note: we use upper 8 bits of flags for driver-internal flags: */
+#define OMAP_BO_DMA			0x01000000	/* actually is physically contiguous */
+#define OMAP_BO_EXT_SYNC	0x02000000	/* externally allocated sync object */
+#define OMAP_BO_EXT_MEM		0x04000000	/* externally allocated memory */
+
+
+struct omap_gem_object {
+	struct drm_gem_object base;
+
+	struct list_head mm_list;
+
+	uint32_t flags;
+
+	/** width/height for tiled formats (rounded up to slot boundaries) */
+	uint16_t width, height;
+
+	/** roll applied when mapping to DMM */
+	uint32_t roll;
+
+	/**
+	 * If buffer is allocated physically contiguous, the OMAP_BO_DMA flag
+	 * is set and the paddr is valid.  Also if the buffer is remapped in
+	 * TILER and paddr_cnt > 0, then paddr is valid.  But if you are using
+	 * the physical address and OMAP_BO_DMA is not set, then you should
+	 * be going thru omap_gem_{get,put}_paddr() to ensure the mapping is
+	 * not removed from under your feet.
+	 *
+	 * Note that OMAP_BO_SCANOUT is a hint from userspace that DMA capable
+	 * buffer is requested, but doesn't mean that it is.  Use the
+	 * OMAP_BO_DMA flag to determine if the buffer has a DMA capable
+	 * physical address.
+	 */
+	dma_addr_t paddr;
+
+	/**
+	 * # of users of paddr
+	 */
+	uint32_t paddr_cnt;
+
+	/**
+	 * tiler block used when buffer is remapped in DMM/TILER.
+	 */
+	struct tiler_block *block;
+
+	/**
+	 * Array of backing pages, if allocated.  Note that pages are never
+	 * allocated for buffers originally allocated from contiguous memory
+	 */
+	struct page **pages;
+
+	/** addresses corresponding to pages in above array */
+	dma_addr_t *addrs;
+
+	/**
+	 * Virtual address, if mapped.
+	 */
+	void *vaddr;
+
+	/**
+	 * sync-object allocated on demand (if needed)
+	 *
+	 * Per-buffer sync-object for tracking pending and completed hw/dma
+	 * read and write operations.  The layout in memory is dictated by
+	 * the SGX firmware, which uses this information to stall the command
+	 * stream if a surface is not ready yet.
+	 *
+	 * Note that when buffer is used by SGX, the sync-object needs to be
+	 * allocated from a special heap of sync-objects.  This way many sync
+	 * objects can be packed in a page, and not waste GPU virtual address
+	 * space.  Because of this we have to have a omap_gem_set_sync_object()
+	 * API to allow replacement of the syncobj after it has (potentially)
+	 * already been allocated.  A bit ugly but I haven't thought of a
+	 * better alternative.
+	 */
+	struct {
+		uint32_t write_pending;
+		uint32_t write_complete;
+		uint32_t read_pending;
+		uint32_t read_complete;
+	} *sync;
+};
+
+static int get_pages(struct drm_gem_object *obj, struct page ***pages);
+static uint64_t mmap_offset(struct drm_gem_object *obj);
+
+/* To deal with userspace mmap'ings of 2d tiled buffers, which (a) are
+ * not necessarily pinned in TILER all the time, and (b) when they are
+ * they are not necessarily page aligned, we reserve one or more small
+ * regions in each of the 2d containers to use as a user-GART where we
+ * can create a second page-aligned mapping of parts of the buffer
+ * being accessed from userspace.
+ *
+ * Note that we could optimize slightly when we know that multiple
+ * tiler containers are backed by the same PAT.. but I'll leave that
+ * for later..
+ */
+#define NUM_USERGART_ENTRIES 2
+struct usergart_entry {
+	struct tiler_block *block;	/* the reserved tiler block */
+	dma_addr_t paddr;
+	struct drm_gem_object *obj;	/* the current pinned obj */
+	pgoff_t obj_pgoff;		/* page offset of obj currently
+					   mapped in */
+};
+static struct {
+	struct usergart_entry entry[NUM_USERGART_ENTRIES];
+	int height;				/* height in rows */
+	int height_shift;		/* ilog2(height in rows) */
+	int slot_shift;			/* ilog2(width per slot) */
+	int stride_pfn;			/* stride in pages */
+	int last;				/* index of last used entry */
+} *usergart;
+
+static void evict_entry(struct drm_gem_object *obj,
+		enum tiler_fmt fmt, struct usergart_entry *entry)
+{
+	if (obj->dev->dev_mapping) {
+		struct omap_gem_object *omap_obj = to_omap_bo(obj);
+		int n = usergart[fmt].height;
+		size_t size = PAGE_SIZE * n;
+		loff_t off = mmap_offset(obj) +
+				(entry->obj_pgoff << PAGE_SHIFT);
+		const int m = 1 + ((omap_obj->width << fmt) / PAGE_SIZE);
+		if (m > 1) {
+			int i;
+			/* if stride > than PAGE_SIZE then sparse mapping: */
+			for (i = n; i > 0; i--) {
+				unmap_mapping_range(obj->dev->dev_mapping,
+						off, PAGE_SIZE, 1);
+				off += PAGE_SIZE * m;
+			}
+		} else {
+			unmap_mapping_range(obj->dev->dev_mapping, off, size, 1);
+		}
+	}
+
+	entry->obj = NULL;
+}
+
+/* Evict a buffer from usergart, if it is mapped there */
+static void evict(struct drm_gem_object *obj)
+{
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+
+	if (omap_obj->flags & OMAP_BO_TILED) {
+		enum tiler_fmt fmt = gem2fmt(omap_obj->flags);
+		int i;
+
+		if (!usergart)
+			return;
+
+		for (i = 0; i < NUM_USERGART_ENTRIES; i++) {
+			struct usergart_entry *entry = &usergart[fmt].entry[i];
+			if (entry->obj == obj)
+				evict_entry(obj, fmt, entry);
+		}
+	}
+}
+
+/* GEM objects can either be allocated from contiguous memory (in which
+ * case obj->filp==NULL), or w/ shmem backing (obj->filp!=NULL).  But non
+ * contiguous buffers can be remapped in TILER/DMM if they need to be
+ * contiguous... but we don't do this all the time to reduce pressure
+ * on TILER/DMM space when we know at allocation time that the buffer
+ * will need to be scanned out.
+ */
+static inline bool is_shmem(struct drm_gem_object *obj)
+{
+	return obj->filp != NULL;
+}
+
+/**
+ * shmem buffers that are mapped cached can simulate coherency via using
+ * page faulting to keep track of dirty pages
+ */
+static inline bool is_cached_coherent(struct drm_gem_object *obj)
+{
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+	return is_shmem(obj) &&
+		((omap_obj->flags & OMAP_BO_CACHE_MASK) == OMAP_BO_CACHED);
+}
+
+static DEFINE_SPINLOCK(sync_lock);
+
+/** ensure backing pages are allocated */
+static int omap_gem_attach_pages(struct drm_gem_object *obj)
+{
+	struct drm_device *dev = obj->dev;
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+	struct page **pages;
+	int npages = obj->size >> PAGE_SHIFT;
+	int i, ret;
+	dma_addr_t *addrs;
+
+	WARN_ON(omap_obj->pages);
+
+	/* TODO: __GFP_DMA32 .. but somehow GFP_HIGHMEM is coming from the
+	 * mapping_gfp_mask(mapping) which conflicts w/ GFP_DMA32.. probably
+	 * we actually want CMA memory for it all anyways..
+	 */
+	pages = _drm_gem_get_pages(obj, GFP_KERNEL);
+	if (IS_ERR(pages)) {
+		dev_err(obj->dev->dev, "could not get pages: %ld\n", PTR_ERR(pages));
+		return PTR_ERR(pages);
+	}
+
+	/* for non-cached buffers, ensure the new pages are clean because
+	 * DSS, GPU, etc. are not cache coherent:
+	 */
+	if (omap_obj->flags & (OMAP_BO_WC|OMAP_BO_UNCACHED)) {
+		addrs = kmalloc(npages * sizeof(*addrs), GFP_KERNEL);
+		if (!addrs) {
+			ret = -ENOMEM;
+			goto free_pages;
+		}
+
+		for (i = 0; i < npages; i++) {
+			addrs[i] = dma_map_page(dev->dev, pages[i],
+					0, PAGE_SIZE, DMA_BIDIRECTIONAL);
+		}
+	} else {
+		addrs = kzalloc(npages * sizeof(*addrs), GFP_KERNEL);
+		if (!addrs) {
+			ret = -ENOMEM;
+			goto free_pages;
+		}
+	}
+
+	omap_obj->addrs = addrs;
+	omap_obj->pages = pages;
+
+	return 0;
+
+free_pages:
+	_drm_gem_put_pages(obj, pages, true, false);
+
+	return ret;
+}
+
+/** release backing pages */
+static void omap_gem_detach_pages(struct drm_gem_object *obj)
+{
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+
+	/* for non-cached buffers, ensure the new pages are clean because
+	 * DSS, GPU, etc. are not cache coherent:
+	 */
+	if (omap_obj->flags & (OMAP_BO_WC|OMAP_BO_UNCACHED)) {
+		int i, npages = obj->size >> PAGE_SHIFT;
+		for (i = 0; i < npages; i++) {
+			dma_unmap_page(obj->dev->dev, omap_obj->addrs[i],
+					PAGE_SIZE, DMA_BIDIRECTIONAL);
+		}
+	}
+
+	kfree(omap_obj->addrs);
+	omap_obj->addrs = NULL;
+
+	_drm_gem_put_pages(obj, omap_obj->pages, true, false);
+	omap_obj->pages = NULL;
+}
+
+/* get buffer flags */
+uint32_t omap_gem_flags(struct drm_gem_object *obj)
+{
+	return to_omap_bo(obj)->flags;
+}
+
+/** get mmap offset */
+static uint64_t mmap_offset(struct drm_gem_object *obj)
+{
+	struct drm_device *dev = obj->dev;
+
+	WARN_ON(!mutex_is_locked(&dev->struct_mutex));
+
+	if (!obj->map_list.map) {
+		/* Make it mmapable */
+		size_t size = omap_gem_mmap_size(obj);
+		int ret = _drm_gem_create_mmap_offset_size(obj, size);
+
+		if (ret) {
+			dev_err(dev->dev, "could not allocate mmap offset\n");
+			return 0;
+		}
+	}
+
+	return (uint64_t)obj->map_list.hash.key << PAGE_SHIFT;
+}
+
+uint64_t omap_gem_mmap_offset(struct drm_gem_object *obj)
+{
+	uint64_t offset;
+	mutex_lock(&obj->dev->struct_mutex);
+	offset = mmap_offset(obj);
+	mutex_unlock(&obj->dev->struct_mutex);
+	return offset;
+}
+
+/** get mmap size */
+size_t omap_gem_mmap_size(struct drm_gem_object *obj)
+{
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+	size_t size = obj->size;
+
+	if (omap_obj->flags & OMAP_BO_TILED) {
+		/* for tiled buffers, the virtual size has stride rounded up
+		 * to 4kb.. (to hide the fact that row n+1 might start 16kb or
+		 * 32kb later!).  But we don't back the entire buffer with
+		 * pages, only the valid picture part.. so need to adjust for
+		 * this in the size used to mmap and generate mmap offset
+		 */
+		size = tiler_vsize(gem2fmt(omap_obj->flags),
+				omap_obj->width, omap_obj->height);
+	}
+
+	return size;
+}
+
+/* get tiled size, returns -EINVAL if not tiled buffer */
+int omap_gem_tiled_size(struct drm_gem_object *obj, uint16_t *w, uint16_t *h)
+{
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+	if (omap_obj->flags & OMAP_BO_TILED) {
+		*w = omap_obj->width;
+		*h = omap_obj->height;
+		return 0;
+	}
+	return -EINVAL;
+}
+
+/* Normal handling for the case of faulting in non-tiled buffers */
+static int fault_1d(struct drm_gem_object *obj,
+		struct vm_area_struct *vma, struct vm_fault *vmf)
+{
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+	unsigned long pfn;
+	pgoff_t pgoff;
+
+	/* We don't use vmf->pgoff since that has the fake offset: */
+	pgoff = ((unsigned long)vmf->virtual_address -
+			vma->vm_start) >> PAGE_SHIFT;
+
+	if (omap_obj->pages) {
+		omap_gem_cpu_sync(obj, pgoff);
+		pfn = page_to_pfn(omap_obj->pages[pgoff]);
+	} else {
+		BUG_ON(!(omap_obj->flags & OMAP_BO_DMA));
+		pfn = (omap_obj->paddr >> PAGE_SHIFT) + pgoff;
+	}
+
+	VERB("Inserting %p pfn %lx, pa %lx", vmf->virtual_address,
+			pfn, pfn << PAGE_SHIFT);
+
+	return vm_insert_mixed(vma, (unsigned long)vmf->virtual_address, pfn);
+}
+
+/* Special handling for the case of faulting in 2d tiled buffers */
+static int fault_2d(struct drm_gem_object *obj,
+		struct vm_area_struct *vma, struct vm_fault *vmf)
+{
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+	struct usergart_entry *entry;
+	enum tiler_fmt fmt = gem2fmt(omap_obj->flags);
+	struct page *pages[64];  /* XXX is this too much to have on stack? */
+	unsigned long pfn;
+	pgoff_t pgoff, base_pgoff;
+	void __user *vaddr;
+	int i, ret, slots;
+
+	/*
+	 * Note the height of the slot is also equal to the number of pages
+	 * that need to be mapped in to fill 4kb wide CPU page.  If the slot
+	 * height is 64, then 64 pages fill a 4kb wide by 64 row region.
+	 */
+	const int n = usergart[fmt].height;
+	const int n_shift = usergart[fmt].height_shift;
+
+	/*
+	 * If buffer width in bytes > PAGE_SIZE then the virtual stride is
+	 * rounded up to next multiple of PAGE_SIZE.. this need to be taken
+	 * into account in some of the math, so figure out virtual stride
+	 * in pages
+	 */
+	const int m = 1 + ((omap_obj->width << fmt) / PAGE_SIZE);
+
+	/* We don't use vmf->pgoff since that has the fake offset: */
+	pgoff = ((unsigned long)vmf->virtual_address -
+			vma->vm_start) >> PAGE_SHIFT;
+
+	/*
+	 * Actual address we start mapping at is rounded down to previous slot
+	 * boundary in the y direction:
+	 */
+	base_pgoff = round_down(pgoff, m << n_shift);
+
+	/* figure out buffer width in slots */
+	slots = omap_obj->width >> usergart[fmt].slot_shift;
+
+	vaddr = vmf->virtual_address - ((pgoff - base_pgoff) << PAGE_SHIFT);
+
+	entry = &usergart[fmt].entry[usergart[fmt].last];
+
+	/* evict previous buffer using this usergart entry, if any: */
+	if (entry->obj)
+		evict_entry(entry->obj, fmt, entry);
+
+	entry->obj = obj;
+	entry->obj_pgoff = base_pgoff;
+
+	/* now convert base_pgoff to phys offset from virt offset: */
+	base_pgoff = (base_pgoff >> n_shift) * slots;
+
+	/* for wider-than 4k.. figure out which part of the slot-row we want: */
+	if (m > 1) {
+		int off = pgoff % m;
+		entry->obj_pgoff += off;
+		base_pgoff /= m;
+		slots = min(slots - (off << n_shift), n);
+		base_pgoff += off << n_shift;
+		vaddr += off << PAGE_SHIFT;
+	}
+
+	/*
+	 * Map in pages. Beyond the valid pixel part of the buffer, we set
+	 * pages[i] to NULL to get a dummy page mapped in.. if someone
+	 * reads/writes it they will get random/undefined content, but at
+	 * least it won't be corrupting whatever other random page used to
+	 * be mapped in, or other undefined behavior.
+	 */
+	memcpy(pages, &omap_obj->pages[base_pgoff],
+			sizeof(struct page *) * slots);
+	memset(pages + slots, 0,
+			sizeof(struct page *) * (n - slots));
+
+	ret = tiler_pin(entry->block, pages, ARRAY_SIZE(pages), 0, true);
+	if (ret) {
+		dev_err(obj->dev->dev, "failed to pin: %d\n", ret);
+		return ret;
+	}
+
+	pfn = entry->paddr >> PAGE_SHIFT;
+
+	VERB("Inserting %p pfn %lx, pa %lx", vmf->virtual_address,
+			pfn, pfn << PAGE_SHIFT);
+
+	for (i = n; i > 0; i--) {
+		vm_insert_mixed(vma, (unsigned long)vaddr, pfn);
+		pfn += usergart[fmt].stride_pfn;
+		vaddr += PAGE_SIZE * m;
+	}
+
+	/* simple round-robin: */
+	usergart[fmt].last = (usergart[fmt].last + 1) % NUM_USERGART_ENTRIES;
+
+	return 0;
+}
+
+/**
+ * omap_gem_fault		-	pagefault handler for GEM objects
+ * @vma: the VMA of the GEM object
+ * @vmf: fault detail
+ *
+ * Invoked when a fault occurs on an mmap of a GEM managed area. GEM
+ * does most of the work for us including the actual map/unmap calls
+ * but we need to do the actual page work.
+ *
+ * The VMA was set up by GEM. In doing so it also ensured that the
+ * vma->vm_private_data points to the GEM object that is backing this
+ * mapping.
+ */
+int omap_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
+{
+	struct drm_gem_object *obj = vma->vm_private_data;
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+	struct drm_device *dev = obj->dev;
+	struct page **pages;
+	int ret;
+
+	/* Make sure we don't parallel update on a fault, nor move or remove
+	 * something from beneath our feet
+	 */
+	mutex_lock(&dev->struct_mutex);
+
+	/* if a shmem backed object, make sure we have pages attached now */
+	ret = get_pages(obj, &pages);
+	if (ret)
+		goto fail;
+
+	/* where should we do corresponding put_pages().. we are mapping
+	 * the original page, rather than thru a GART, so we can't rely
+	 * on eviction to trigger this.  But munmap() or all mappings should
+	 * probably trigger put_pages()?
+	 */
+
+	if (omap_obj->flags & OMAP_BO_TILED)
+		ret = fault_2d(obj, vma, vmf);
+	else
+		ret = fault_1d(obj, vma, vmf);
+
+
+fail:
+	mutex_unlock(&dev->struct_mutex);
+	switch (ret) {
+	case 0:
+	case -ERESTARTSYS:
+	case -EINTR:
+		return VM_FAULT_NOPAGE;
+	case -ENOMEM:
+		return VM_FAULT_OOM;
+	default:
+		return VM_FAULT_SIGBUS;
+	}
+}
+
+/** We override mainly to fix up some of the vm mapping flags.. */
+int omap_gem_mmap(struct file *filp, struct vm_area_struct *vma)
+{
+	int ret;
+
+	ret = drm_gem_mmap(filp, vma);
+	if (ret) {
+		DBG("mmap failed: %d", ret);
+		return ret;
+	}
+
+	return omap_gem_mmap_obj(vma->vm_private_data, vma);
+}
+
+int omap_gem_mmap_obj(struct drm_gem_object *obj,
+		struct vm_area_struct *vma)
+{
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+
+	vma->vm_flags &= ~VM_PFNMAP;
+	vma->vm_flags |= VM_MIXEDMAP;
+
+	if (omap_obj->flags & OMAP_BO_WC) {
+		vma->vm_page_prot = pgprot_writecombine(vm_get_page_prot(vma->vm_flags));
+	} else if (omap_obj->flags & OMAP_BO_UNCACHED) {
+		vma->vm_page_prot = pgprot_noncached(vm_get_page_prot(vma->vm_flags));
+	} else {
+		/*
+		 * We do have some private objects, at least for scanout buffers
+		 * on hardware without DMM/TILER.  But these are allocated write-
+		 * combine
+		 */
+		if (WARN_ON(!obj->filp))
+			return -EINVAL;
+
+		/*
+		 * Shunt off cached objs to shmem file so they have their own
+		 * address_space (so unmap_mapping_range does what we want,
+		 * in particular in the case of mmap'd dmabufs)
+		 */
+		fput(vma->vm_file);
+		vma->vm_pgoff = 0;
+		vma->vm_file  = get_file(obj->filp);
+
+		vma->vm_page_prot = vm_get_page_prot(vma->vm_flags);
+	}
+
+	return 0;
+}
+
+
+/**
+ * omap_gem_dumb_create	-	create a dumb buffer
+ * @drm_file: our client file
+ * @dev: our device
+ * @args: the requested arguments copied from userspace
+ *
+ * Allocate a buffer suitable for use for a frame buffer of the
+ * form described by user space. Give userspace a handle by which
+ * to reference it.
+ */
+int omap_gem_dumb_create(struct drm_file *file, struct drm_device *dev,
+		struct drm_mode_create_dumb *args)
+{
+	union omap_gem_size gsize;
+
+	/* in case someone tries to feed us a completely bogus stride: */
+	args->pitch = align_pitch(args->pitch, args->width, args->bpp);
+	args->size = PAGE_ALIGN(args->pitch * args->height);
+
+	gsize = (union omap_gem_size){
+		.bytes = args->size,
+	};
+
+	return omap_gem_new_handle(dev, file, gsize,
+			OMAP_BO_SCANOUT | OMAP_BO_WC, &args->handle);
+}
+
+/**
+ * omap_gem_dumb_destroy	-	destroy a dumb buffer
+ * @file: client file
+ * @dev: our DRM device
+ * @handle: the object handle
+ *
+ * Destroy a handle that was created via omap_gem_dumb_create.
+ */
+int omap_gem_dumb_destroy(struct drm_file *file, struct drm_device *dev,
+		uint32_t handle)
+{
+	/* No special work needed, drop the reference and see what falls out */
+	return drm_gem_handle_delete(file, handle);
+}
+
+/**
+ * omap_gem_dumb_map	-	buffer mapping for dumb interface
+ * @file: our drm client file
+ * @dev: drm device
+ * @handle: GEM handle to the object (from dumb_create)
+ *
+ * Do the necessary setup to allow the mapping of the frame buffer
+ * into user memory. We don't have to do much here at the moment.
+ */
+int omap_gem_dumb_map_offset(struct drm_file *file, struct drm_device *dev,
+		uint32_t handle, uint64_t *offset)
+{
+	struct drm_gem_object *obj;
+	int ret = 0;
+
+	/* GEM does all our handle to object mapping */
+	obj = drm_gem_object_lookup(dev, file, handle);
+	if (obj == NULL) {
+		ret = -ENOENT;
+		goto fail;
+	}
+
+	*offset = omap_gem_mmap_offset(obj);
+
+	drm_gem_object_unreference_unlocked(obj);
+
+fail:
+	return ret;
+}
+
+/* Set scrolling position.  This allows us to implement fast scrolling
+ * for console.
+ *
+ * Call only from non-atomic contexts.
+ */
+int omap_gem_roll(struct drm_gem_object *obj, uint32_t roll)
+{
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+	uint32_t npages = obj->size >> PAGE_SHIFT;
+	int ret = 0;
+
+	if (roll > npages) {
+		dev_err(obj->dev->dev, "invalid roll: %d\n", roll);
+		return -EINVAL;
+	}
+
+	omap_obj->roll = roll;
+
+	mutex_lock(&obj->dev->struct_mutex);
+
+	/* if we aren't mapped yet, we don't need to do anything */
+	if (omap_obj->block) {
+		struct page **pages;
+		ret = get_pages(obj, &pages);
+		if (ret)
+			goto fail;
+		ret = tiler_pin(omap_obj->block, pages, npages, roll, true);
+		if (ret)
+			dev_err(obj->dev->dev, "could not repin: %d\n", ret);
+	}
+
+fail:
+	mutex_unlock(&obj->dev->struct_mutex);
+
+	return ret;
+}
+
+/* Sync the buffer for CPU access.. note pages should already be
+ * attached, ie. omap_gem_get_pages()
+ */
+void omap_gem_cpu_sync(struct drm_gem_object *obj, int pgoff)
+{
+	struct drm_device *dev = obj->dev;
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+
+	if (is_cached_coherent(obj) && omap_obj->addrs[pgoff]) {
+		dma_unmap_page(dev->dev, omap_obj->addrs[pgoff],
+				PAGE_SIZE, DMA_BIDIRECTIONAL);
+		omap_obj->addrs[pgoff] = 0;
+	}
+}
+
+/* sync the buffer for DMA access */
+void omap_gem_dma_sync(struct drm_gem_object *obj,
+		enum dma_data_direction dir)
+{
+	struct drm_device *dev = obj->dev;
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+
+	if (is_cached_coherent(obj)) {
+		int i, npages = obj->size >> PAGE_SHIFT;
+		struct page **pages = omap_obj->pages;
+		bool dirty = false;
+
+		for (i = 0; i < npages; i++) {
+			if (!omap_obj->addrs[i]) {
+				omap_obj->addrs[i] = dma_map_page(dev->dev, pages[i], 0,
+						PAGE_SIZE, DMA_BIDIRECTIONAL);
+				dirty = true;
+			}
+		}
+
+		if (dirty) {
+			unmap_mapping_range(obj->filp->f_mapping, 0,
+					omap_gem_mmap_size(obj), 1);
+		}
+	}
+}
+
+/* Get physical address for DMA.. if 'remap' is true, and the buffer is not
+ * already contiguous, remap it to pin in physically contiguous memory.. (ie.
+ * map in TILER)
+ */
+int omap_gem_get_paddr(struct drm_gem_object *obj,
+		dma_addr_t *paddr, bool remap)
+{
+	struct omap_drm_private *priv = obj->dev->dev_private;
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+	int ret = 0;
+
+	mutex_lock(&obj->dev->struct_mutex);
+
+	if (remap && is_shmem(obj) && priv->has_dmm) {
+		if (omap_obj->paddr_cnt == 0) {
+			struct page **pages;
+			uint32_t npages = obj->size >> PAGE_SHIFT;
+			enum tiler_fmt fmt = gem2fmt(omap_obj->flags);
+			struct tiler_block *block;
+
+			BUG_ON(omap_obj->block);
+
+			ret = get_pages(obj, &pages);
+			if (ret)
+				goto fail;
+
+			if (omap_obj->flags & OMAP_BO_TILED) {
+				block = tiler_reserve_2d(fmt,
+						omap_obj->width,
+						omap_obj->height, 0);
+			} else {
+				block = tiler_reserve_1d(obj->size);
+			}
+
+			if (IS_ERR(block)) {
+				ret = PTR_ERR(block);
+				dev_err(obj->dev->dev,
+					"could not remap: %d (%d)\n", ret, fmt);
+				goto fail;
+			}
+
+			/* TODO: enable async refill.. */
+			ret = tiler_pin(block, pages, npages,
+					omap_obj->roll, true);
+			if (ret) {
+				tiler_release(block);
+				dev_err(obj->dev->dev,
+						"could not pin: %d\n", ret);
+				goto fail;
+			}
+
+			omap_obj->paddr = tiler_ssptr(block);
+			omap_obj->block = block;
+
+			DBG("got paddr: %08x", omap_obj->paddr);
+		}
+
+		omap_obj->paddr_cnt++;
+
+		*paddr = omap_obj->paddr;
+	} else if (omap_obj->flags & OMAP_BO_DMA) {
+		*paddr = omap_obj->paddr;
+	} else {
+		ret = -EINVAL;
+		goto fail;
+	}
+
+fail:
+	mutex_unlock(&obj->dev->struct_mutex);
+
+	return ret;
+}
+
+/* Release physical address, when DMA is no longer being performed.. this
+ * could potentially unpin and unmap buffers from TILER
+ */
+int omap_gem_put_paddr(struct drm_gem_object *obj)
+{
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+	int ret = 0;
+
+	mutex_lock(&obj->dev->struct_mutex);
+	if (omap_obj->paddr_cnt > 0) {
+		omap_obj->paddr_cnt--;
+		if (omap_obj->paddr_cnt == 0) {
+			ret = tiler_unpin(omap_obj->block);
+			if (ret) {
+				dev_err(obj->dev->dev,
+					"could not unpin pages: %d\n", ret);
+				goto fail;
+			}
+			ret = tiler_release(omap_obj->block);
+			if (ret) {
+				dev_err(obj->dev->dev,
+					"could not release unmap: %d\n", ret);
+			}
+			omap_obj->block = NULL;
+		}
+	}
+fail:
+	mutex_unlock(&obj->dev->struct_mutex);
+	return ret;
+}
+
+/* Get rotated scanout address (only valid if already pinned), at the
+ * specified orientation and x,y offset from top-left corner of buffer
+ * (only valid for tiled 2d buffers)
+ */
+int omap_gem_rotated_paddr(struct drm_gem_object *obj, uint32_t orient,
+		int x, int y, dma_addr_t *paddr)
+{
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+	int ret = -EINVAL;
+
+	mutex_lock(&obj->dev->struct_mutex);
+	if ((omap_obj->paddr_cnt > 0) && omap_obj->block &&
+			(omap_obj->flags & OMAP_BO_TILED)) {
+		*paddr = tiler_tsptr(omap_obj->block, orient, x, y);
+		ret = 0;
+	}
+	mutex_unlock(&obj->dev->struct_mutex);
+	return ret;
+}
+
+/* Get tiler stride for the buffer (only valid for 2d tiled buffers) */
+int omap_gem_tiled_stride(struct drm_gem_object *obj, uint32_t orient)
+{
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+	int ret = -EINVAL;
+	if (omap_obj->flags & OMAP_BO_TILED)
+		ret = tiler_stride(gem2fmt(omap_obj->flags), orient);
+	return ret;
+}
+
+/* acquire pages when needed (for example, for DMA where physically
+ * contiguous buffer is not required
+ */
+static int get_pages(struct drm_gem_object *obj, struct page ***pages)
+{
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+	int ret = 0;
+
+	if (is_shmem(obj) && !omap_obj->pages) {
+		ret = omap_gem_attach_pages(obj);
+		if (ret) {
+			dev_err(obj->dev->dev, "could not attach pages\n");
+			return ret;
+		}
+	}
+
+	/* TODO: even phys-contig.. we should have a list of pages? */
+	*pages = omap_obj->pages;
+
+	return 0;
+}
+
+/* if !remap, and we don't have pages backing, then fail, rather than
+ * increasing the pin count (which we don't really do yet anyways,
+ * because we don't support swapping pages back out).  And 'remap'
+ * might not be quite the right name, but I wanted to keep it working
+ * similarly to omap_gem_get_paddr().  Note though that mutex is not
+ * aquired if !remap (because this can be called in atomic ctxt),
+ * but probably omap_gem_get_paddr() should be changed to work in the
+ * same way.  If !remap, a matching omap_gem_put_pages() call is not
+ * required (and should not be made).
+ */
+int omap_gem_get_pages(struct drm_gem_object *obj, struct page ***pages,
+		bool remap)
+{
+	int ret;
+	if (!remap) {
+		struct omap_gem_object *omap_obj = to_omap_bo(obj);
+		if (!omap_obj->pages)
+			return -ENOMEM;
+		*pages = omap_obj->pages;
+		return 0;
+	}
+	mutex_lock(&obj->dev->struct_mutex);
+	ret = get_pages(obj, pages);
+	mutex_unlock(&obj->dev->struct_mutex);
+	return ret;
+}
+
+/* release pages when DMA no longer being performed */
+int omap_gem_put_pages(struct drm_gem_object *obj)
+{
+	/* do something here if we dynamically attach/detach pages.. at
+	 * least they would no longer need to be pinned if everyone has
+	 * released the pages..
+	 */
+	return 0;
+}
+
+/* Get kernel virtual address for CPU access.. this more or less only
+ * exists for omap_fbdev.  This should be called with struct_mutex
+ * held.
+ */
+void *omap_gem_vaddr(struct drm_gem_object *obj)
+{
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+	WARN_ON(!mutex_is_locked(&obj->dev->struct_mutex));
+	if (!omap_obj->vaddr) {
+		struct page **pages;
+		int ret = get_pages(obj, &pages);
+		if (ret)
+			return ERR_PTR(ret);
+		omap_obj->vaddr = vmap(pages, obj->size >> PAGE_SHIFT,
+				VM_MAP, pgprot_writecombine(PAGE_KERNEL));
+	}
+	return omap_obj->vaddr;
+}
+
+#ifdef CONFIG_PM
+/* re-pin objects in DMM in resume path: */
+int omap_gem_resume(struct device *dev)
+{
+	struct drm_device *drm_dev = dev_get_drvdata(dev);
+	struct omap_drm_private *priv = drm_dev->dev_private;
+	struct omap_gem_object *omap_obj;
+	int ret = 0;
+
+	list_for_each_entry(omap_obj, &priv->obj_list, mm_list) {
+		if (omap_obj->block) {
+			struct drm_gem_object *obj = &omap_obj->base;
+			uint32_t npages = obj->size >> PAGE_SHIFT;
+			WARN_ON(!omap_obj->pages);  /* this can't happen */
+			ret = tiler_pin(omap_obj->block,
+					omap_obj->pages, npages,
+					omap_obj->roll, true);
+			if (ret) {
+				dev_err(dev, "could not repin: %d\n", ret);
+				return ret;
+			}
+		}
+	}
+
+	return 0;
+}
+#endif
+
+#ifdef CONFIG_DEBUG_FS
+void omap_gem_describe(struct drm_gem_object *obj, struct seq_file *m)
+{
+	struct drm_device *dev = obj->dev;
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+	uint64_t off = 0;
+
+	WARN_ON(!mutex_is_locked(&dev->struct_mutex));
+
+	if (obj->map_list.map)
+		off = (uint64_t)obj->map_list.hash.key;
+
+	seq_printf(m, "%08x: %2d (%2d) %08llx %08Zx (%2d) %p %4d",
+			omap_obj->flags, obj->name, obj->refcount.refcount.counter,
+			off, omap_obj->paddr, omap_obj->paddr_cnt,
+			omap_obj->vaddr, omap_obj->roll);
+
+	if (omap_obj->flags & OMAP_BO_TILED) {
+		seq_printf(m, " %dx%d", omap_obj->width, omap_obj->height);
+		if (omap_obj->block) {
+			struct tcm_area *area = &omap_obj->block->area;
+			seq_printf(m, " (%dx%d, %dx%d)",
+					area->p0.x, area->p0.y,
+					area->p1.x, area->p1.y);
+		}
+	} else {
+		seq_printf(m, " %d", obj->size);
+	}
+
+	seq_printf(m, "\n");
+}
+
+void omap_gem_describe_objects(struct list_head *list, struct seq_file *m)
+{
+	struct omap_gem_object *omap_obj;
+	int count = 0;
+	size_t size = 0;
+
+	list_for_each_entry(omap_obj, list, mm_list) {
+		struct drm_gem_object *obj = &omap_obj->base;
+		seq_printf(m, "   ");
+		omap_gem_describe(obj, m);
+		count++;
+		size += obj->size;
+	}
+
+	seq_printf(m, "Total %d objects, %zu bytes\n", count, size);
+}
+#endif
+
+/* Buffer Synchronization:
+ */
+
+struct omap_gem_sync_waiter {
+	struct list_head list;
+	struct omap_gem_object *omap_obj;
+	enum omap_gem_op op;
+	uint32_t read_target, write_target;
+	/* notify called w/ sync_lock held */
+	void (*notify)(void *arg);
+	void *arg;
+};
+
+/* list of omap_gem_sync_waiter.. the notify fxn gets called back when
+ * the read and/or write target count is achieved which can call a user
+ * callback (ex. to kick 3d and/or 2d), wakeup blocked task (prep for
+ * cpu access), etc.
+ */
+static LIST_HEAD(waiters);
+
+static inline bool is_waiting(struct omap_gem_sync_waiter *waiter)
+{
+	struct omap_gem_object *omap_obj = waiter->omap_obj;
+	if ((waiter->op & OMAP_GEM_READ) &&
+			(omap_obj->sync->read_complete < waiter->read_target))
+		return true;
+	if ((waiter->op & OMAP_GEM_WRITE) &&
+			(omap_obj->sync->write_complete < waiter->write_target))
+		return true;
+	return false;
+}
+
+/* macro for sync debug.. */
+#define SYNCDBG 0
+#define SYNC(fmt, ...) do { if (SYNCDBG) \
+		printk(KERN_ERR "%s:%d: "fmt"\n", \
+				__func__, __LINE__, ##__VA_ARGS__); \
+	} while (0)
+
+
+static void sync_op_update(void)
+{
+	struct omap_gem_sync_waiter *waiter, *n;
+	list_for_each_entry_safe(waiter, n, &waiters, list) {
+		if (!is_waiting(waiter)) {
+			list_del(&waiter->list);
+			SYNC("notify: %p", waiter);
+			waiter->notify(waiter->arg);
+			kfree(waiter);
+		}
+	}
+}
+
+static inline int sync_op(struct drm_gem_object *obj,
+		enum omap_gem_op op, bool start)
+{
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+	int ret = 0;
+
+	spin_lock(&sync_lock);
+
+	if (!omap_obj->sync) {
+		omap_obj->sync = kzalloc(sizeof(*omap_obj->sync), GFP_ATOMIC);
+		if (!omap_obj->sync) {
+			ret = -ENOMEM;
+			goto unlock;
+		}
+	}
+
+	if (start) {
+		if (op & OMAP_GEM_READ)
+			omap_obj->sync->read_pending++;
+		if (op & OMAP_GEM_WRITE)
+			omap_obj->sync->write_pending++;
+	} else {
+		if (op & OMAP_GEM_READ)
+			omap_obj->sync->read_complete++;
+		if (op & OMAP_GEM_WRITE)
+			omap_obj->sync->write_complete++;
+		sync_op_update();
+	}
+
+unlock:
+	spin_unlock(&sync_lock);
+
+	return ret;
+}
+
+/* it is a bit lame to handle updates in this sort of polling way, but
+ * in case of PVR, the GPU can directly update read/write complete
+ * values, and not really tell us which ones it updated.. this also
+ * means that sync_lock is not quite sufficient.  So we'll need to
+ * do something a bit better when it comes time to add support for
+ * separate 2d hw..
+ */
+void omap_gem_op_update(void)
+{
+	spin_lock(&sync_lock);
+	sync_op_update();
+	spin_unlock(&sync_lock);
+}
+
+/* mark the start of read and/or write operation */
+int omap_gem_op_start(struct drm_gem_object *obj, enum omap_gem_op op)
+{
+	return sync_op(obj, op, true);
+}
+
+int omap_gem_op_finish(struct drm_gem_object *obj, enum omap_gem_op op)
+{
+	return sync_op(obj, op, false);
+}
+
+static DECLARE_WAIT_QUEUE_HEAD(sync_event);
+
+static void sync_notify(void *arg)
+{
+	struct task_struct **waiter_task = arg;
+	*waiter_task = NULL;
+	wake_up_all(&sync_event);
+}
+
+int omap_gem_op_sync(struct drm_gem_object *obj, enum omap_gem_op op)
+{
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+	int ret = 0;
+	if (omap_obj->sync) {
+		struct task_struct *waiter_task = current;
+		struct omap_gem_sync_waiter *waiter =
+				kzalloc(sizeof(*waiter), GFP_KERNEL);
+
+		if (!waiter)
+			return -ENOMEM;
+
+		waiter->omap_obj = omap_obj;
+		waiter->op = op;
+		waiter->read_target = omap_obj->sync->read_pending;
+		waiter->write_target = omap_obj->sync->write_pending;
+		waiter->notify = sync_notify;
+		waiter->arg = &waiter_task;
+
+		spin_lock(&sync_lock);
+		if (is_waiting(waiter)) {
+			SYNC("waited: %p", waiter);
+			list_add_tail(&waiter->list, &waiters);
+			spin_unlock(&sync_lock);
+			ret = wait_event_interruptible(sync_event,
+					(waiter_task == NULL));
+			spin_lock(&sync_lock);
+			if (waiter_task) {
+				SYNC("interrupted: %p", waiter);
+				/* we were interrupted */
+				list_del(&waiter->list);
+				waiter_task = NULL;
+			} else {
+				/* freed in sync_op_update() */
+				waiter = NULL;
+			}
+		}
+		spin_unlock(&sync_lock);
+
+		if (waiter)
+			kfree(waiter);
+	}
+	return ret;
+}
+
+/* call fxn(arg), either synchronously or asynchronously if the op
+ * is currently blocked..  fxn() can be called from any context
+ *
+ * (TODO for now fxn is called back from whichever context calls
+ * omap_gem_op_update().. but this could be better defined later
+ * if needed)
+ *
+ * TODO more code in common w/ _sync()..
+ */
+int omap_gem_op_async(struct drm_gem_object *obj, enum omap_gem_op op,
+		void (*fxn)(void *arg), void *arg)
+{
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+	if (omap_obj->sync) {
+		struct omap_gem_sync_waiter *waiter =
+				kzalloc(sizeof(*waiter), GFP_ATOMIC);
+
+		if (!waiter)
+			return -ENOMEM;
+
+		waiter->omap_obj = omap_obj;
+		waiter->op = op;
+		waiter->read_target = omap_obj->sync->read_pending;
+		waiter->write_target = omap_obj->sync->write_pending;
+		waiter->notify = fxn;
+		waiter->arg = arg;
+
+		spin_lock(&sync_lock);
+		if (is_waiting(waiter)) {
+			SYNC("waited: %p", waiter);
+			list_add_tail(&waiter->list, &waiters);
+			spin_unlock(&sync_lock);
+			return 0;
+		}
+
+		spin_unlock(&sync_lock);
+	}
+
+	/* no waiting.. */
+	fxn(arg);
+
+	return 0;
+}
+
+/* special API so PVR can update the buffer to use a sync-object allocated
+ * from it's sync-obj heap.  Only used for a newly allocated (from PVR's
+ * perspective) sync-object, so we overwrite the new syncobj w/ values
+ * from the already allocated syncobj (if there is one)
+ */
+int omap_gem_set_sync_object(struct drm_gem_object *obj, void *syncobj)
+{
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+	int ret = 0;
+
+	spin_lock(&sync_lock);
+
+	if ((omap_obj->flags & OMAP_BO_EXT_SYNC) && !syncobj) {
+		/* clearing a previously set syncobj */
+		syncobj = kmemdup(omap_obj->sync, sizeof(*omap_obj->sync),
+				  GFP_ATOMIC);
+		if (!syncobj) {
+			ret = -ENOMEM;
+			goto unlock;
+		}
+		omap_obj->flags &= ~OMAP_BO_EXT_SYNC;
+		omap_obj->sync = syncobj;
+	} else if (syncobj && !(omap_obj->flags & OMAP_BO_EXT_SYNC)) {
+		/* replacing an existing syncobj */
+		if (omap_obj->sync) {
+			memcpy(syncobj, omap_obj->sync, sizeof(*omap_obj->sync));
+			kfree(omap_obj->sync);
+		}
+		omap_obj->flags |= OMAP_BO_EXT_SYNC;
+		omap_obj->sync = syncobj;
+	}
+
+unlock:
+	spin_unlock(&sync_lock);
+	return ret;
+}
+
+int omap_gem_init_object(struct drm_gem_object *obj)
+{
+	return -EINVAL;          /* unused */
+}
+
+/* don't call directly.. called from GEM core when it is time to actually
+ * free the object..
+ */
+void omap_gem_free_object(struct drm_gem_object *obj)
+{
+	struct drm_device *dev = obj->dev;
+	struct omap_gem_object *omap_obj = to_omap_bo(obj);
+
+	evict(obj);
+
+	WARN_ON(!mutex_is_locked(&dev->struct_mutex));
+
+	list_del(&omap_obj->mm_list);
+
+	if (obj->map_list.map)
+		drm_gem_free_mmap_offset(obj);
+
+	/* this means the object is still pinned.. which really should
+	 * not happen.  I think..
+	 */
+	WARN_ON(omap_obj->paddr_cnt > 0);
+
+	/* don't free externally allocated backing memory */
+	if (!(omap_obj->flags & OMAP_BO_EXT_MEM)) {
+		if (omap_obj->pages)
+			omap_gem_detach_pages(obj);
+
+		if (!is_shmem(obj)) {
+			dma_free_writecombine(dev->dev, obj->size,
+					omap_obj->vaddr, omap_obj->paddr);
+		} else if (omap_obj->vaddr) {
+			vunmap(omap_obj->vaddr);
+		}
+	}
+
+	/* don't free externally allocated syncobj */
+	if (!(omap_obj->flags & OMAP_BO_EXT_SYNC))
+		kfree(omap_obj->sync);
+
+	drm_gem_object_release(obj);
+
+	kfree(obj);
+}
+
+/* convenience method to construct a GEM buffer object, and userspace handle */
+int omap_gem_new_handle(struct drm_device *dev, struct drm_file *file,
+		union omap_gem_size gsize, uint32_t flags, uint32_t *handle)
+{
+	struct drm_gem_object *obj;
+	int ret;
+
+	obj = omap_gem_new(dev, gsize, flags);
+	if (!obj)
+		return -ENOMEM;
+
+	ret = drm_gem_handle_create(file, obj, handle);
+	if (ret) {
+		drm_gem_object_release(obj);
+		kfree(obj); /* TODO isn't there a dtor to call? just copying i915 */
+		return ret;
+	}
+
+	/* drop reference from allocate - handle holds it now */
+	drm_gem_object_unreference_unlocked(obj);
+
+	return 0;
+}
+
+/* GEM buffer object constructor */
+struct drm_gem_object *omap_gem_new(struct drm_device *dev,
+		union omap_gem_size gsize, uint32_t flags)
+{
+	struct omap_drm_private *priv = dev->dev_private;
+	struct omap_gem_object *omap_obj;
+	struct drm_gem_object *obj = NULL;
+	size_t size;
+	int ret;
+
+	if (flags & OMAP_BO_TILED) {
+		if (!usergart) {
+			dev_err(dev->dev, "Tiled buffers require DMM\n");
+			goto fail;
+		}
+
+		/* tiled buffers are always shmem paged backed.. when they are
+		 * scanned out, they are remapped into DMM/TILER
+		 */
+		flags &= ~OMAP_BO_SCANOUT;
+
+		/* currently don't allow cached buffers.. there is some caching
+		 * stuff that needs to be handled better
+		 */
+		flags &= ~(OMAP_BO_CACHED|OMAP_BO_UNCACHED);
+		flags |= OMAP_BO_WC;
+
+		/* align dimensions to slot boundaries... */
+		tiler_align(gem2fmt(flags),
+				&gsize.tiled.width, &gsize.tiled.height);
+
+		/* ...and calculate size based on aligned dimensions */
+		size = tiler_size(gem2fmt(flags),
+				gsize.tiled.width, gsize.tiled.height);
+	} else {
+		size = PAGE_ALIGN(gsize.bytes);
+	}
+
+	omap_obj = kzalloc(sizeof(*omap_obj), GFP_KERNEL);
+	if (!omap_obj) {
+		dev_err(dev->dev, "could not allocate GEM object\n");
+		goto fail;
+	}
+
+	list_add(&omap_obj->mm_list, &priv->obj_list);
+
+	obj = &omap_obj->base;
+
+	if ((flags & OMAP_BO_SCANOUT) && !priv->has_dmm) {
+		/* attempt to allocate contiguous memory if we don't
+		 * have DMM for remappign discontiguous buffers
+		 */
+		omap_obj->vaddr =  dma_alloc_writecombine(dev->dev, size,
+				&omap_obj->paddr, GFP_KERNEL);
+		if (omap_obj->vaddr)
+			flags |= OMAP_BO_DMA;
+
+	}
+
+	omap_obj->flags = flags;
+
+	if (flags & OMAP_BO_TILED) {
+		omap_obj->width = gsize.tiled.width;
+		omap_obj->height = gsize.tiled.height;
+	}
+
+	if (flags & (OMAP_BO_DMA|OMAP_BO_EXT_MEM))
+		ret = drm_gem_private_object_init(dev, obj, size);
+	else
+		ret = drm_gem_object_init(dev, obj, size);
+
+	if (ret)
+		goto fail;
+
+	return obj;
+
+fail:
+	if (obj)
+		omap_gem_free_object(obj);
+
+	return NULL;
+}
+
+/* init/cleanup.. if DMM is used, we need to set some stuff up.. */
+void omap_gem_init(struct drm_device *dev)
+{
+	struct omap_drm_private *priv = dev->dev_private;
+	const enum tiler_fmt fmts[] = {
+			TILFMT_8BIT, TILFMT_16BIT, TILFMT_32BIT
+	};
+	int i, j;
+
+	if (!dmm_is_available()) {
+		/* DMM only supported on OMAP4 and later, so this isn't fatal */
+		dev_warn(dev->dev, "DMM not available, disable DMM support\n");
+		return;
+	}
+
+	usergart = kzalloc(3 * sizeof(*usergart), GFP_KERNEL);
+	if (!usergart) {
+		dev_warn(dev->dev, "could not allocate usergart\n");
+		return;
+	}
+
+	/* reserve 4k aligned/wide regions for userspace mappings: */
+	for (i = 0; i < ARRAY_SIZE(fmts); i++) {
+		uint16_t h = 1, w = PAGE_SIZE >> i;
+		tiler_align(fmts[i], &w, &h);
+		/* note: since each region is 1 4kb page wide, and minimum
+		 * number of rows, the height ends up being the same as the
+		 * # of pages in the region
+		 */
+		usergart[i].height = h;
+		usergart[i].height_shift = ilog2(h);
+		usergart[i].stride_pfn = tiler_stride(fmts[i], 0) >> PAGE_SHIFT;
+		usergart[i].slot_shift = ilog2((PAGE_SIZE / h) >> i);
+		for (j = 0; j < NUM_USERGART_ENTRIES; j++) {
+			struct usergart_entry *entry = &usergart[i].entry[j];
+			struct tiler_block *block =
+					tiler_reserve_2d(fmts[i], w, h,
+							PAGE_SIZE);
+			if (IS_ERR(block)) {
+				dev_err(dev->dev,
+						"reserve failed: %d, %d, %ld\n",
+						i, j, PTR_ERR(block));
+				return;
+			}
+			entry->paddr = tiler_ssptr(block);
+			entry->block = block;
+
+			DBG("%d:%d: %dx%d: paddr=%08x stride=%d", i, j, w, h,
+					entry->paddr,
+					usergart[i].stride_pfn << PAGE_SHIFT);
+		}
+	}
+
+	priv->has_dmm = true;
+}
+
+void omap_gem_deinit(struct drm_device *dev)
+{
+	/* I believe we can rely on there being no more outstanding GEM
+	 * objects which could depend on usergart/dmm at this point.
+	 */
+	kfree(usergart);
+}
diff --git a/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c b/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c
new file mode 100644
index 0000000..ac74d1b
--- /dev/null
+++ b/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c
@@ -0,0 +1,225 @@
+/*
+ * drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c
+ *
+ * Copyright (C) 2011 Texas Instruments
+ * Author: Rob Clark <rob.clark@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "omap_drv.h"
+
+#include <linux/dma-buf.h>
+
+static struct sg_table *omap_gem_map_dma_buf(
+		struct dma_buf_attachment *attachment,
+		enum dma_data_direction dir)
+{
+	struct drm_gem_object *obj = attachment->dmabuf->priv;
+	struct sg_table *sg;
+	dma_addr_t paddr;
+	int ret;
+
+	sg = kzalloc(sizeof(*sg), GFP_KERNEL);
+	if (!sg)
+		return ERR_PTR(-ENOMEM);
+
+	/* camera, etc, need physically contiguous.. but we need a
+	 * better way to know this..
+	 */
+	ret = omap_gem_get_paddr(obj, &paddr, true);
+	if (ret)
+		goto out;
+
+	ret = sg_alloc_table(sg, 1, GFP_KERNEL);
+	if (ret)
+		goto out;
+
+	sg_init_table(sg->sgl, 1);
+	sg_dma_len(sg->sgl) = obj->size;
+	sg_set_page(sg->sgl, pfn_to_page(PFN_DOWN(paddr)), obj->size, 0);
+	sg_dma_address(sg->sgl) = paddr;
+
+	/* this should be after _get_paddr() to ensure we have pages attached */
+	omap_gem_dma_sync(obj, dir);
+
+	return sg;
+out:
+	kfree(sg);
+	return ERR_PTR(ret);
+}
+
+static void omap_gem_unmap_dma_buf(struct dma_buf_attachment *attachment,
+		struct sg_table *sg, enum dma_data_direction dir)
+{
+	struct drm_gem_object *obj = attachment->dmabuf->priv;
+	omap_gem_put_paddr(obj);
+	sg_free_table(sg);
+	kfree(sg);
+}
+
+static void omap_gem_dmabuf_release(struct dma_buf *buffer)
+{
+	struct drm_gem_object *obj = buffer->priv;
+	/* release reference that was taken when dmabuf was exported
+	 * in omap_gem_prime_set()..
+	 */
+	drm_gem_object_unreference_unlocked(obj);
+}
+
+
+static int omap_gem_dmabuf_begin_cpu_access(struct dma_buf *buffer,
+		size_t start, size_t len, enum dma_data_direction dir)
+{
+	struct drm_gem_object *obj = buffer->priv;
+	struct page **pages;
+	if (omap_gem_flags(obj) & OMAP_BO_TILED) {
+		/* TODO we would need to pin at least part of the buffer to
+		 * get de-tiled view.  For now just reject it.
+		 */
+		return -ENOMEM;
+	}
+	/* make sure we have the pages: */
+	return omap_gem_get_pages(obj, &pages, true);
+}
+
+static void omap_gem_dmabuf_end_cpu_access(struct dma_buf *buffer,
+		size_t start, size_t len, enum dma_data_direction dir)
+{
+	struct drm_gem_object *obj = buffer->priv;
+	omap_gem_put_pages(obj);
+}
+
+
+static void *omap_gem_dmabuf_kmap_atomic(struct dma_buf *buffer,
+		unsigned long page_num)
+{
+	struct drm_gem_object *obj = buffer->priv;
+	struct page **pages;
+	omap_gem_get_pages(obj, &pages, false);
+	omap_gem_cpu_sync(obj, page_num);
+	return kmap_atomic(pages[page_num]);
+}
+
+static void omap_gem_dmabuf_kunmap_atomic(struct dma_buf *buffer,
+		unsigned long page_num, void *addr)
+{
+	kunmap_atomic(addr);
+}
+
+static void *omap_gem_dmabuf_kmap(struct dma_buf *buffer,
+		unsigned long page_num)
+{
+	struct drm_gem_object *obj = buffer->priv;
+	struct page **pages;
+	omap_gem_get_pages(obj, &pages, false);
+	omap_gem_cpu_sync(obj, page_num);
+	return kmap(pages[page_num]);
+}
+
+static void omap_gem_dmabuf_kunmap(struct dma_buf *buffer,
+		unsigned long page_num, void *addr)
+{
+	struct drm_gem_object *obj = buffer->priv;
+	struct page **pages;
+	omap_gem_get_pages(obj, &pages, false);
+	kunmap(pages[page_num]);
+}
+
+/*
+ * TODO maybe we can split up drm_gem_mmap to avoid duplicating
+ * some here.. or at least have a drm_dmabuf_mmap helper.
+ */
+static int omap_gem_dmabuf_mmap(struct dma_buf *buffer,
+		struct vm_area_struct *vma)
+{
+	struct drm_gem_object *obj = buffer->priv;
+	int ret = 0;
+
+	if (WARN_ON(!obj->filp))
+		return -EINVAL;
+
+	/* Check for valid size. */
+	if (omap_gem_mmap_size(obj) < vma->vm_end - vma->vm_start) {
+		ret = -EINVAL;
+		goto out_unlock;
+	}
+
+	if (!obj->dev->driver->gem_vm_ops) {
+		ret = -EINVAL;
+		goto out_unlock;
+	}
+
+	vma->vm_flags |= VM_IO | VM_PFNMAP | VM_DONTEXPAND | VM_DONTDUMP;
+	vma->vm_ops = obj->dev->driver->gem_vm_ops;
+	vma->vm_private_data = obj;
+	vma->vm_page_prot =  pgprot_writecombine(vm_get_page_prot(vma->vm_flags));
+
+	/* Take a ref for this mapping of the object, so that the fault
+	 * handler can dereference the mmap offset's pointer to the object.
+	 * This reference is cleaned up by the corresponding vm_close
+	 * (which should happen whether the vma was created by this call, or
+	 * by a vm_open due to mremap or partial unmap or whatever).
+	 */
+	vma->vm_ops->open(vma);
+
+out_unlock:
+
+	return omap_gem_mmap_obj(obj, vma);
+}
+
+struct dma_buf_ops omap_dmabuf_ops = {
+		.map_dma_buf = omap_gem_map_dma_buf,
+		.unmap_dma_buf = omap_gem_unmap_dma_buf,
+		.release = omap_gem_dmabuf_release,
+		.begin_cpu_access = omap_gem_dmabuf_begin_cpu_access,
+		.end_cpu_access = omap_gem_dmabuf_end_cpu_access,
+		.kmap_atomic = omap_gem_dmabuf_kmap_atomic,
+		.kunmap_atomic = omap_gem_dmabuf_kunmap_atomic,
+		.kmap = omap_gem_dmabuf_kmap,
+		.kunmap = omap_gem_dmabuf_kunmap,
+		.mmap = omap_gem_dmabuf_mmap,
+};
+
+struct dma_buf *omap_gem_prime_export(struct drm_device *dev,
+		struct drm_gem_object *obj, int flags)
+{
+	return dma_buf_export(obj, &omap_dmabuf_ops, obj->size, flags);
+}
+
+struct drm_gem_object *omap_gem_prime_import(struct drm_device *dev,
+		struct dma_buf *buffer)
+{
+	struct drm_gem_object *obj;
+
+	/* is this one of own objects? */
+	if (buffer->ops == &omap_dmabuf_ops) {
+		obj = buffer->priv;
+		/* is it from our device? */
+		if (obj->dev == dev) {
+			/*
+			 * Importing dmabuf exported from out own gem increases
+			 * refcount on gem itself instead of f_count of dmabuf.
+			 */
+			drm_gem_object_reference(obj);
+			dma_buf_put(buffer);
+			return obj;
+		}
+	}
+
+	/*
+	 * TODO add support for importing buffers from other devices..
+	 * for now we don't need this but would be nice to add eventually
+	 */
+	return ERR_PTR(-EINVAL);
+}
diff --git a/drivers/gpu/drm/omapdrm/omap_gem_helpers.c b/drivers/gpu/drm/omapdrm/omap_gem_helpers.c
new file mode 100644
index 0000000..e4a66a3
--- /dev/null
+++ b/drivers/gpu/drm/omapdrm/omap_gem_helpers.c
@@ -0,0 +1,169 @@
+/*
+ * drivers/gpu/drm/omapdrm/omap_gem_helpers.c
+ *
+ * Copyright (C) 2011 Texas Instruments
+ * Author: Rob Clark <rob.clark@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+/* temporary copy of drm_gem_{get,put}_pages() until the
+ * "drm/gem: add functions to get/put pages" patch is merged..
+ */
+
+#include <linux/module.h>
+#include <linux/types.h>
+#include <linux/shmem_fs.h>
+
+#include <drm/drmP.h>
+
+/**
+ * drm_gem_get_pages - helper to allocate backing pages for a GEM object
+ * @obj: obj in question
+ * @gfpmask: gfp mask of requested pages
+ */
+struct page **_drm_gem_get_pages(struct drm_gem_object *obj, gfp_t gfpmask)
+{
+	struct inode *inode;
+	struct address_space *mapping;
+	struct page *p, **pages;
+	int i, npages;
+
+	/* This is the shared memory object that backs the GEM resource */
+	inode = obj->filp->f_path.dentry->d_inode;
+	mapping = inode->i_mapping;
+
+	npages = obj->size >> PAGE_SHIFT;
+
+	pages = drm_malloc_ab(npages, sizeof(struct page *));
+	if (pages == NULL)
+		return ERR_PTR(-ENOMEM);
+
+	gfpmask |= mapping_gfp_mask(mapping);
+
+	for (i = 0; i < npages; i++) {
+		p = shmem_read_mapping_page_gfp(mapping, i, gfpmask);
+		if (IS_ERR(p))
+			goto fail;
+		pages[i] = p;
+
+		/* There is a hypothetical issue w/ drivers that require
+		 * buffer memory in the low 4GB.. if the pages are un-
+		 * pinned, and swapped out, they can end up swapped back
+		 * in above 4GB.  If pages are already in memory, then
+		 * shmem_read_mapping_page_gfp will ignore the gfpmask,
+		 * even if the already in-memory page disobeys the mask.
+		 *
+		 * It is only a theoretical issue today, because none of
+		 * the devices with this limitation can be populated with
+		 * enough memory to trigger the issue.  But this BUG_ON()
+		 * is here as a reminder in case the problem with
+		 * shmem_read_mapping_page_gfp() isn't solved by the time
+		 * it does become a real issue.
+		 *
+		 * See this thread: http://lkml.org/lkml/2011/7/11/238
+		 */
+		BUG_ON((gfpmask & __GFP_DMA32) &&
+				(page_to_pfn(p) >= 0x00100000UL));
+	}
+
+	return pages;
+
+fail:
+	while (i--)
+		page_cache_release(pages[i]);
+
+	drm_free_large(pages);
+	return ERR_CAST(p);
+}
+
+/**
+ * drm_gem_put_pages - helper to free backing pages for a GEM object
+ * @obj: obj in question
+ * @pages: pages to free
+ */
+void _drm_gem_put_pages(struct drm_gem_object *obj, struct page **pages,
+		bool dirty, bool accessed)
+{
+	int i, npages;
+
+	npages = obj->size >> PAGE_SHIFT;
+
+	for (i = 0; i < npages; i++) {
+		if (dirty)
+			set_page_dirty(pages[i]);
+
+		if (accessed)
+			mark_page_accessed(pages[i]);
+
+		/* Undo the reference we took when populating the table */
+		page_cache_release(pages[i]);
+	}
+
+	drm_free_large(pages);
+}
+
+int
+_drm_gem_create_mmap_offset_size(struct drm_gem_object *obj, size_t size)
+{
+	struct drm_device *dev = obj->dev;
+	struct drm_gem_mm *mm = dev->mm_private;
+	struct drm_map_list *list;
+	struct drm_local_map *map;
+	int ret = 0;
+
+	/* Set the object up for mmap'ing */
+	list = &obj->map_list;
+	list->map = kzalloc(sizeof(struct drm_map_list), GFP_KERNEL);
+	if (!list->map)
+		return -ENOMEM;
+
+	map = list->map;
+	map->type = _DRM_GEM;
+	map->size = size;
+	map->handle = obj;
+
+	/* Get a DRM GEM mmap offset allocated... */
+	list->file_offset_node = drm_mm_search_free(&mm->offset_manager,
+			size / PAGE_SIZE, 0, 0);
+
+	if (!list->file_offset_node) {
+		DRM_ERROR("failed to allocate offset for bo %d\n", obj->name);
+		ret = -ENOSPC;
+		goto out_free_list;
+	}
+
+	list->file_offset_node = drm_mm_get_block(list->file_offset_node,
+			size / PAGE_SIZE, 0);
+	if (!list->file_offset_node) {
+		ret = -ENOMEM;
+		goto out_free_list;
+	}
+
+	list->hash.key = list->file_offset_node->start;
+	ret = drm_ht_insert_item(&mm->offset_hash, &list->hash);
+	if (ret) {
+		DRM_ERROR("failed to add to map hash\n");
+		goto out_free_mm;
+	}
+
+	return 0;
+
+out_free_mm:
+	drm_mm_put_block(list->file_offset_node);
+out_free_list:
+	kfree(list->map);
+	list->map = NULL;
+
+	return ret;
+}
diff --git a/drivers/gpu/drm/omapdrm/omap_irq.c b/drivers/gpu/drm/omapdrm/omap_irq.c
new file mode 100644
index 0000000..e01303e
--- /dev/null
+++ b/drivers/gpu/drm/omapdrm/omap_irq.c
@@ -0,0 +1,322 @@
+/*
+ * drivers/gpu/drm/omapdrm/omap_irq.c
+ *
+ * Copyright (C) 2012 Texas Instruments
+ * Author: Rob Clark <rob.clark@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "omap_drv.h"
+
+static DEFINE_SPINLOCK(list_lock);
+
+static void omap_irq_error_handler(struct omap_drm_irq *irq,
+		uint32_t irqstatus)
+{
+	DRM_ERROR("errors: %08x\n", irqstatus);
+}
+
+/* call with list_lock and dispc runtime held */
+static void omap_irq_update(struct drm_device *dev)
+{
+	struct omap_drm_private *priv = dev->dev_private;
+	struct omap_drm_irq *irq;
+	uint32_t irqmask = priv->vblank_mask;
+
+	BUG_ON(!spin_is_locked(&list_lock));
+
+	list_for_each_entry(irq, &priv->irq_list, node)
+		irqmask |= irq->irqmask;
+
+	DBG("irqmask=%08x", irqmask);
+
+	dispc_write_irqenable(irqmask);
+	dispc_read_irqenable();        /* flush posted write */
+}
+
+void omap_irq_register(struct drm_device *dev, struct omap_drm_irq *irq)
+{
+	struct omap_drm_private *priv = dev->dev_private;
+	unsigned long flags;
+
+	dispc_runtime_get();
+	spin_lock_irqsave(&list_lock, flags);
+
+	if (!WARN_ON(irq->registered)) {
+		irq->registered = true;
+		list_add(&irq->node, &priv->irq_list);
+		omap_irq_update(dev);
+	}
+
+	spin_unlock_irqrestore(&list_lock, flags);
+	dispc_runtime_put();
+}
+
+void omap_irq_unregister(struct drm_device *dev, struct omap_drm_irq *irq)
+{
+	unsigned long flags;
+
+	dispc_runtime_get();
+	spin_lock_irqsave(&list_lock, flags);
+
+	if (!WARN_ON(!irq->registered)) {
+		irq->registered = false;
+		list_del(&irq->node);
+		omap_irq_update(dev);
+	}
+
+	spin_unlock_irqrestore(&list_lock, flags);
+	dispc_runtime_put();
+}
+
+struct omap_irq_wait {
+	struct omap_drm_irq irq;
+	int count;
+};
+
+static DECLARE_WAIT_QUEUE_HEAD(wait_event);
+
+static void wait_irq(struct omap_drm_irq *irq, uint32_t irqstatus)
+{
+	struct omap_irq_wait *wait =
+			container_of(irq, struct omap_irq_wait, irq);
+	wait->count--;
+	wake_up_all(&wait_event);
+}
+
+struct omap_irq_wait * omap_irq_wait_init(struct drm_device *dev,
+		uint32_t irqmask, int count)
+{
+	struct omap_irq_wait *wait = kzalloc(sizeof(*wait), GFP_KERNEL);
+	wait->irq.irq = wait_irq;
+	wait->irq.irqmask = irqmask;
+	wait->count = count;
+	omap_irq_register(dev, &wait->irq);
+	return wait;
+}
+
+int omap_irq_wait(struct drm_device *dev, struct omap_irq_wait *wait,
+		unsigned long timeout)
+{
+	int ret = wait_event_timeout(wait_event, (wait->count <= 0), timeout);
+	omap_irq_unregister(dev, &wait->irq);
+	kfree(wait);
+	if (ret == 0)
+		return -1;
+	return 0;
+}
+
+/**
+ * enable_vblank - enable vblank interrupt events
+ * @dev: DRM device
+ * @crtc: which irq to enable
+ *
+ * Enable vblank interrupts for @crtc.  If the device doesn't have
+ * a hardware vblank counter, this routine should be a no-op, since
+ * interrupts will have to stay on to keep the count accurate.
+ *
+ * RETURNS
+ * Zero on success, appropriate errno if the given @crtc's vblank
+ * interrupt cannot be enabled.
+ */
+int omap_irq_enable_vblank(struct drm_device *dev, int crtc)
+{
+	struct omap_drm_private *priv = dev->dev_private;
+	unsigned long flags;
+
+	DBG("dev=%p, crtc=%d", dev, crtc);
+
+	dispc_runtime_get();
+	spin_lock_irqsave(&list_lock, flags);
+	priv->vblank_mask |= pipe2vbl(crtc);
+	omap_irq_update(dev);
+	spin_unlock_irqrestore(&list_lock, flags);
+	dispc_runtime_put();
+
+	return 0;
+}
+
+/**
+ * disable_vblank - disable vblank interrupt events
+ * @dev: DRM device
+ * @crtc: which irq to enable
+ *
+ * Disable vblank interrupts for @crtc.  If the device doesn't have
+ * a hardware vblank counter, this routine should be a no-op, since
+ * interrupts will have to stay on to keep the count accurate.
+ */
+void omap_irq_disable_vblank(struct drm_device *dev, int crtc)
+{
+	struct omap_drm_private *priv = dev->dev_private;
+	unsigned long flags;
+
+	DBG("dev=%p, crtc=%d", dev, crtc);
+
+	dispc_runtime_get();
+	spin_lock_irqsave(&list_lock, flags);
+	priv->vblank_mask &= ~pipe2vbl(crtc);
+	omap_irq_update(dev);
+	spin_unlock_irqrestore(&list_lock, flags);
+	dispc_runtime_put();
+}
+
+irqreturn_t omap_irq_handler(DRM_IRQ_ARGS)
+{
+	struct drm_device *dev = (struct drm_device *) arg;
+	struct omap_drm_private *priv = dev->dev_private;
+	struct omap_drm_irq *handler, *n;
+	unsigned long flags;
+	unsigned int id;
+	u32 irqstatus;
+
+	irqstatus = dispc_read_irqstatus();
+	dispc_clear_irqstatus(irqstatus);
+	dispc_read_irqstatus();        /* flush posted write */
+
+	VERB("irqs: %08x", irqstatus);
+
+	for (id = 0; id < priv->num_crtcs; id++)
+		if (irqstatus & pipe2vbl(id))
+			drm_handle_vblank(dev, id);
+
+	spin_lock_irqsave(&list_lock, flags);
+	list_for_each_entry_safe(handler, n, &priv->irq_list, node) {
+		if (handler->irqmask & irqstatus) {
+			spin_unlock_irqrestore(&list_lock, flags);
+			handler->irq(handler, handler->irqmask & irqstatus);
+			spin_lock_irqsave(&list_lock, flags);
+		}
+	}
+	spin_unlock_irqrestore(&list_lock, flags);
+
+	return IRQ_HANDLED;
+}
+
+void omap_irq_preinstall(struct drm_device *dev)
+{
+	DBG("dev=%p", dev);
+	dispc_runtime_get();
+	dispc_clear_irqstatus(0xffffffff);
+	dispc_runtime_put();
+}
+
+int omap_irq_postinstall(struct drm_device *dev)
+{
+	struct omap_drm_private *priv = dev->dev_private;
+	struct omap_drm_irq *error_handler = &priv->error_handler;
+
+	DBG("dev=%p", dev);
+
+	INIT_LIST_HEAD(&priv->irq_list);
+
+	error_handler->irq = omap_irq_error_handler;
+	error_handler->irqmask = DISPC_IRQ_OCP_ERR;
+
+	/* for now ignore DISPC_IRQ_SYNC_LOST_DIGIT.. really I think
+	 * we just need to ignore it while enabling tv-out
+	 */
+	error_handler->irqmask &= ~DISPC_IRQ_SYNC_LOST_DIGIT;
+
+	omap_irq_register(dev, error_handler);
+
+	return 0;
+}
+
+void omap_irq_uninstall(struct drm_device *dev)
+{
+	DBG("dev=%p", dev);
+	// TODO prolly need to call drm_irq_uninstall() somewhere too
+}
+
+/*
+ * We need a special version, instead of just using drm_irq_install(),
+ * because we need to register the irq via omapdss.  Once omapdss and
+ * omapdrm are merged together we can assign the dispc hwmod data to
+ * ourselves and drop these and just use drm_irq_{install,uninstall}()
+ */
+
+int omap_drm_irq_install(struct drm_device *dev)
+{
+	int ret;
+
+	mutex_lock(&dev->struct_mutex);
+
+	if (dev->irq_enabled) {
+		mutex_unlock(&dev->struct_mutex);
+		return -EBUSY;
+	}
+	dev->irq_enabled = 1;
+	mutex_unlock(&dev->struct_mutex);
+
+	/* Before installing handler */
+	if (dev->driver->irq_preinstall)
+		dev->driver->irq_preinstall(dev);
+
+	ret = dispc_request_irq(dev->driver->irq_handler, dev);
+
+	if (ret < 0) {
+		mutex_lock(&dev->struct_mutex);
+		dev->irq_enabled = 0;
+		mutex_unlock(&dev->struct_mutex);
+		return ret;
+	}
+
+	/* After installing handler */
+	if (dev->driver->irq_postinstall)
+		ret = dev->driver->irq_postinstall(dev);
+
+	if (ret < 0) {
+		mutex_lock(&dev->struct_mutex);
+		dev->irq_enabled = 0;
+		mutex_unlock(&dev->struct_mutex);
+		dispc_free_irq(dev);
+	}
+
+	return ret;
+}
+
+int omap_drm_irq_uninstall(struct drm_device *dev)
+{
+	unsigned long irqflags;
+	int irq_enabled, i;
+
+	mutex_lock(&dev->struct_mutex);
+	irq_enabled = dev->irq_enabled;
+	dev->irq_enabled = 0;
+	mutex_unlock(&dev->struct_mutex);
+
+	/*
+	 * Wake up any waiters so they don't hang.
+	 */
+	if (dev->num_crtcs) {
+		spin_lock_irqsave(&dev->vbl_lock, irqflags);
+		for (i = 0; i < dev->num_crtcs; i++) {
+			DRM_WAKEUP(&dev->vbl_queue[i]);
+			dev->vblank_enabled[i] = 0;
+			dev->last_vblank[i] =
+				dev->driver->get_vblank_counter(dev, i);
+		}
+		spin_unlock_irqrestore(&dev->vbl_lock, irqflags);
+	}
+
+	if (!irq_enabled)
+		return -EINVAL;
+
+	if (dev->driver->irq_uninstall)
+		dev->driver->irq_uninstall(dev);
+
+	dispc_free_irq(dev);
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/omapdrm/omap_plane.c b/drivers/gpu/drm/omapdrm/omap_plane.c
new file mode 100644
index 0000000..dd68d14
--- /dev/null
+++ b/drivers/gpu/drm/omapdrm/omap_plane.c
@@ -0,0 +1,450 @@
+/*
+ * drivers/gpu/drm/omapdrm/omap_plane.c
+ *
+ * Copyright (C) 2011 Texas Instruments
+ * Author: Rob Clark <rob.clark@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/kfifo.h>
+
+#include "omap_drv.h"
+#include "omap_dmm_tiler.h"
+
+/* some hackery because omapdss has an 'enum omap_plane' (which would be
+ * better named omap_plane_id).. and compiler seems unhappy about having
+ * both a 'struct omap_plane' and 'enum omap_plane'
+ */
+#define omap_plane _omap_plane
+
+/*
+ * plane funcs
+ */
+
+struct callback {
+	void (*fxn)(void *);
+	void *arg;
+};
+
+#define to_omap_plane(x) container_of(x, struct omap_plane, base)
+
+struct omap_plane {
+	struct drm_plane base;
+	int id;  /* TODO rename omap_plane -> omap_plane_id in omapdss so I can use the enum */
+	const char *name;
+	struct omap_overlay_info info;
+	struct omap_drm_apply apply;
+
+	/* position/orientation of scanout within the fb: */
+	struct omap_drm_window win;
+	bool enabled;
+
+	/* last fb that we pinned: */
+	struct drm_framebuffer *pinned_fb;
+
+	uint32_t nformats;
+	uint32_t formats[32];
+
+	struct omap_drm_irq error_irq;
+
+	/* set of bo's pending unpin until next post_apply() */
+	DECLARE_KFIFO_PTR(unpin_fifo, struct drm_gem_object *);
+
+	// XXX maybe get rid of this and handle vblank in crtc too?
+	struct callback apply_done_cb;
+};
+
+static void unpin(void *arg, struct drm_gem_object *bo)
+{
+	struct drm_plane *plane = arg;
+	struct omap_plane *omap_plane = to_omap_plane(plane);
+
+	if (kfifo_put(&omap_plane->unpin_fifo,
+			(const struct drm_gem_object **)&bo)) {
+		/* also hold a ref so it isn't free'd while pinned */
+		drm_gem_object_reference(bo);
+	} else {
+		dev_err(plane->dev->dev, "unpin fifo full!\n");
+		omap_gem_put_paddr(bo);
+	}
+}
+
+/* update which fb (if any) is pinned for scanout */
+static int update_pin(struct drm_plane *plane, struct drm_framebuffer *fb)
+{
+	struct omap_plane *omap_plane = to_omap_plane(plane);
+	struct drm_framebuffer *pinned_fb = omap_plane->pinned_fb;
+
+	if (pinned_fb != fb) {
+		int ret;
+
+		DBG("%p -> %p", pinned_fb, fb);
+
+		if (fb)
+			drm_framebuffer_reference(fb);
+
+		ret = omap_framebuffer_replace(pinned_fb, fb, plane, unpin);
+
+		if (pinned_fb)
+			drm_framebuffer_unreference(pinned_fb);
+
+		if (ret) {
+			dev_err(plane->dev->dev, "could not swap %p -> %p\n",
+					omap_plane->pinned_fb, fb);
+			if (fb)
+				drm_framebuffer_unreference(fb);
+			omap_plane->pinned_fb = NULL;
+			return ret;
+		}
+
+		omap_plane->pinned_fb = fb;
+	}
+
+	return 0;
+}
+
+static void omap_plane_pre_apply(struct omap_drm_apply *apply)
+{
+	struct omap_plane *omap_plane =
+			container_of(apply, struct omap_plane, apply);
+	struct omap_drm_window *win = &omap_plane->win;
+	struct drm_plane *plane = &omap_plane->base;
+	struct drm_device *dev = plane->dev;
+	struct omap_overlay_info *info = &omap_plane->info;
+	struct drm_crtc *crtc = plane->crtc;
+	enum omap_channel channel;
+	bool enabled = omap_plane->enabled && crtc;
+	bool ilace, replication;
+	int ret;
+
+	DBG("%s, enabled=%d", omap_plane->name, enabled);
+
+	/* if fb has changed, pin new fb: */
+	update_pin(plane, enabled ? plane->fb : NULL);
+
+	if (!enabled) {
+		dispc_ovl_enable(omap_plane->id, false);
+		return;
+	}
+
+	channel = omap_crtc_channel(crtc);
+
+	/* update scanout: */
+	omap_framebuffer_update_scanout(plane->fb, win, info);
+
+	DBG("%dx%d -> %dx%d (%d)", info->width, info->height,
+			info->out_width, info->out_height,
+			info->screen_width);
+	DBG("%d,%d %08x %08x", info->pos_x, info->pos_y,
+			info->paddr, info->p_uv_addr);
+
+	/* TODO: */
+	ilace = false;
+	replication = false;
+
+	/* and finally, update omapdss: */
+	ret = dispc_ovl_setup(omap_plane->id, info,
+			replication, omap_crtc_timings(crtc), false);
+	if (ret) {
+		dev_err(dev->dev, "dispc_ovl_setup failed: %d\n", ret);
+		return;
+	}
+
+	dispc_ovl_enable(omap_plane->id, true);
+	dispc_ovl_set_channel_out(omap_plane->id, channel);
+}
+
+static void omap_plane_post_apply(struct omap_drm_apply *apply)
+{
+	struct omap_plane *omap_plane =
+			container_of(apply, struct omap_plane, apply);
+	struct drm_plane *plane = &omap_plane->base;
+	struct omap_overlay_info *info = &omap_plane->info;
+	struct drm_gem_object *bo = NULL;
+	struct callback cb;
+
+	cb = omap_plane->apply_done_cb;
+	omap_plane->apply_done_cb.fxn = NULL;
+
+	while (kfifo_get(&omap_plane->unpin_fifo, &bo)) {
+		omap_gem_put_paddr(bo);
+		drm_gem_object_unreference_unlocked(bo);
+	}
+
+	if (cb.fxn)
+		cb.fxn(cb.arg);
+
+	if (omap_plane->enabled) {
+		omap_framebuffer_flush(plane->fb, info->pos_x, info->pos_y,
+				info->out_width, info->out_height);
+	}
+}
+
+static int apply(struct drm_plane *plane)
+{
+	if (plane->crtc) {
+		struct omap_plane *omap_plane = to_omap_plane(plane);
+		return omap_crtc_apply(plane->crtc, &omap_plane->apply);
+	}
+	return 0;
+}
+
+int omap_plane_mode_set(struct drm_plane *plane,
+		struct drm_crtc *crtc, struct drm_framebuffer *fb,
+		int crtc_x, int crtc_y,
+		unsigned int crtc_w, unsigned int crtc_h,
+		uint32_t src_x, uint32_t src_y,
+		uint32_t src_w, uint32_t src_h,
+		void (*fxn)(void *), void *arg)
+{
+	struct omap_plane *omap_plane = to_omap_plane(plane);
+	struct omap_drm_window *win = &omap_plane->win;
+
+	win->crtc_x = crtc_x;
+	win->crtc_y = crtc_y;
+	win->crtc_w = crtc_w;
+	win->crtc_h = crtc_h;
+
+	/* src values are in Q16 fixed point, convert to integer: */
+	win->src_x = src_x >> 16;
+	win->src_y = src_y >> 16;
+	win->src_w = src_w >> 16;
+	win->src_h = src_h >> 16;
+
+	if (fxn) {
+		/* omap_crtc should ensure that a new page flip
+		 * isn't permitted while there is one pending:
+		 */
+		BUG_ON(omap_plane->apply_done_cb.fxn);
+
+		omap_plane->apply_done_cb.fxn = fxn;
+		omap_plane->apply_done_cb.arg = arg;
+	}
+
+	plane->fb = fb;
+	plane->crtc = crtc;
+
+	return apply(plane);
+}
+
+static int omap_plane_update(struct drm_plane *plane,
+		struct drm_crtc *crtc, struct drm_framebuffer *fb,
+		int crtc_x, int crtc_y,
+		unsigned int crtc_w, unsigned int crtc_h,
+		uint32_t src_x, uint32_t src_y,
+		uint32_t src_w, uint32_t src_h)
+{
+	struct omap_plane *omap_plane = to_omap_plane(plane);
+	omap_plane->enabled = true;
+	return omap_plane_mode_set(plane, crtc, fb,
+			crtc_x, crtc_y, crtc_w, crtc_h,
+			src_x, src_y, src_w, src_h,
+			NULL, NULL);
+}
+
+static int omap_plane_disable(struct drm_plane *plane)
+{
+	struct omap_plane *omap_plane = to_omap_plane(plane);
+	omap_plane->win.rotation = BIT(DRM_ROTATE_0);
+	return omap_plane_dpms(plane, DRM_MODE_DPMS_OFF);
+}
+
+static void omap_plane_destroy(struct drm_plane *plane)
+{
+	struct omap_plane *omap_plane = to_omap_plane(plane);
+
+	DBG("%s", omap_plane->name);
+
+	omap_irq_unregister(plane->dev, &omap_plane->error_irq);
+
+	omap_plane_disable(plane);
+	drm_plane_cleanup(plane);
+
+	WARN_ON(!kfifo_is_empty(&omap_plane->unpin_fifo));
+	kfifo_free(&omap_plane->unpin_fifo);
+
+	kfree(omap_plane);
+}
+
+int omap_plane_dpms(struct drm_plane *plane, int mode)
+{
+	struct omap_plane *omap_plane = to_omap_plane(plane);
+	bool enabled = (mode == DRM_MODE_DPMS_ON);
+	int ret = 0;
+
+	if (enabled != omap_plane->enabled) {
+		omap_plane->enabled = enabled;
+		ret = apply(plane);
+	}
+
+	return ret;
+}
+
+/* helper to install properties which are common to planes and crtcs */
+void omap_plane_install_properties(struct drm_plane *plane,
+		struct drm_mode_object *obj)
+{
+	struct drm_device *dev = plane->dev;
+	struct omap_drm_private *priv = dev->dev_private;
+	struct drm_property *prop;
+
+	if (priv->has_dmm) {
+		prop = priv->rotation_prop;
+		if (!prop) {
+			const struct drm_prop_enum_list props[] = {
+					{ DRM_ROTATE_0,   "rotate-0" },
+					{ DRM_ROTATE_90,  "rotate-90" },
+					{ DRM_ROTATE_180, "rotate-180" },
+					{ DRM_ROTATE_270, "rotate-270" },
+					{ DRM_REFLECT_X,  "reflect-x" },
+					{ DRM_REFLECT_Y,  "reflect-y" },
+			};
+			prop = drm_property_create_bitmask(dev, 0, "rotation",
+					props, ARRAY_SIZE(props));
+			if (prop == NULL)
+				return;
+			priv->rotation_prop = prop;
+		}
+		drm_object_attach_property(obj, prop, 0);
+	}
+
+	prop = priv->zorder_prop;
+	if (!prop) {
+		prop = drm_property_create_range(dev, 0, "zorder", 0, 3);
+		if (prop == NULL)
+			return;
+		priv->zorder_prop = prop;
+	}
+	drm_object_attach_property(obj, prop, 0);
+}
+
+int omap_plane_set_property(struct drm_plane *plane,
+		struct drm_property *property, uint64_t val)
+{
+	struct omap_plane *omap_plane = to_omap_plane(plane);
+	struct omap_drm_private *priv = plane->dev->dev_private;
+	int ret = -EINVAL;
+
+	if (property == priv->rotation_prop) {
+		DBG("%s: rotation: %02x", omap_plane->name, (uint32_t)val);
+		omap_plane->win.rotation = val;
+		ret = apply(plane);
+	} else if (property == priv->zorder_prop) {
+		DBG("%s: zorder: %02x", omap_plane->name, (uint32_t)val);
+		omap_plane->info.zorder = val;
+		ret = apply(plane);
+	}
+
+	return ret;
+}
+
+static const struct drm_plane_funcs omap_plane_funcs = {
+		.update_plane = omap_plane_update,
+		.disable_plane = omap_plane_disable,
+		.destroy = omap_plane_destroy,
+		.set_property = omap_plane_set_property,
+};
+
+static void omap_plane_error_irq(struct omap_drm_irq *irq, uint32_t irqstatus)
+{
+	struct omap_plane *omap_plane =
+			container_of(irq, struct omap_plane, error_irq);
+	DRM_ERROR("%s: errors: %08x\n", omap_plane->name, irqstatus);
+}
+
+static const char *plane_names[] = {
+		[OMAP_DSS_GFX] = "gfx",
+		[OMAP_DSS_VIDEO1] = "vid1",
+		[OMAP_DSS_VIDEO2] = "vid2",
+		[OMAP_DSS_VIDEO3] = "vid3",
+};
+
+static const uint32_t error_irqs[] = {
+		[OMAP_DSS_GFX] = DISPC_IRQ_GFX_FIFO_UNDERFLOW,
+		[OMAP_DSS_VIDEO1] = DISPC_IRQ_VID1_FIFO_UNDERFLOW,
+		[OMAP_DSS_VIDEO2] = DISPC_IRQ_VID2_FIFO_UNDERFLOW,
+		[OMAP_DSS_VIDEO3] = DISPC_IRQ_VID3_FIFO_UNDERFLOW,
+};
+
+/* initialize plane */
+struct drm_plane *omap_plane_init(struct drm_device *dev,
+		int id, bool private_plane)
+{
+	struct omap_drm_private *priv = dev->dev_private;
+	struct drm_plane *plane = NULL;
+	struct omap_plane *omap_plane;
+	struct omap_overlay_info *info;
+	int ret;
+
+	DBG("%s: priv=%d", plane_names[id], private_plane);
+
+	omap_plane = kzalloc(sizeof(*omap_plane), GFP_KERNEL);
+	if (!omap_plane) {
+		dev_err(dev->dev, "could not allocate plane\n");
+		goto fail;
+	}
+
+	ret = kfifo_alloc(&omap_plane->unpin_fifo, 16, GFP_KERNEL);
+	if (ret) {
+		dev_err(dev->dev, "could not allocate unpin FIFO\n");
+		goto fail;
+	}
+
+	omap_plane->nformats = omap_framebuffer_get_formats(
+			omap_plane->formats, ARRAY_SIZE(omap_plane->formats),
+			dss_feat_get_supported_color_modes(id));
+	omap_plane->id = id;
+	omap_plane->name = plane_names[id];
+
+	plane = &omap_plane->base;
+
+	omap_plane->apply.pre_apply  = omap_plane_pre_apply;
+	omap_plane->apply.post_apply = omap_plane_post_apply;
+
+	omap_plane->error_irq.irqmask = error_irqs[id];
+	omap_plane->error_irq.irq = omap_plane_error_irq;
+	omap_irq_register(dev, &omap_plane->error_irq);
+
+	drm_plane_init(dev, plane, (1 << priv->num_crtcs) - 1, &omap_plane_funcs,
+			omap_plane->formats, omap_plane->nformats, private_plane);
+
+	omap_plane_install_properties(plane, &plane->base);
+
+	/* get our starting configuration, set defaults for parameters
+	 * we don't currently use, etc:
+	 */
+	info = &omap_plane->info;
+	info->rotation_type = OMAP_DSS_ROT_DMA;
+	info->rotation = OMAP_DSS_ROT_0;
+	info->global_alpha = 0xff;
+	info->mirror = 0;
+
+	/* Set defaults depending on whether we are a CRTC or overlay
+	 * layer.
+	 * TODO add ioctl to give userspace an API to change this.. this
+	 * will come in a subsequent patch.
+	 */
+	if (private_plane)
+		omap_plane->info.zorder = 0;
+	else
+		omap_plane->info.zorder = id;
+
+	return plane;
+
+fail:
+	if (plane)
+		omap_plane_destroy(plane);
+
+	return NULL;
+}
diff --git a/drivers/gpu/drm/omapdrm/tcm-sita.c b/drivers/gpu/drm/omapdrm/tcm-sita.c
new file mode 100644
index 0000000..efb6095
--- /dev/null
+++ b/drivers/gpu/drm/omapdrm/tcm-sita.c
@@ -0,0 +1,703 @@
+/*
+ * tcm-sita.c
+ *
+ * SImple Tiler Allocator (SiTA): 2D and 1D allocation(reservation) algorithm
+ *
+ * Authors: Ravi Ramachandra <r.ramachandra@ti.com>,
+ *          Lajos Molnar <molnar@ti.com>
+ *
+ * Copyright (C) 2009-2010 Texas Instruments, Inc.
+ *
+ * This package is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * THIS PACKAGE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED
+ * WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE.
+ *
+ */
+#include <linux/slab.h>
+#include <linux/spinlock.h>
+
+#include "tcm-sita.h"
+
+#define ALIGN_DOWN(value, align) ((value) & ~((align) - 1))
+
+/* Individual selection criteria for different scan areas */
+static s32 CR_L2R_T2B = CR_BIAS_HORIZONTAL;
+static s32 CR_R2L_T2B = CR_DIAGONAL_BALANCE;
+
+/*********************************************
+ *	TCM API - Sita Implementation
+ *********************************************/
+static s32 sita_reserve_2d(struct tcm *tcm, u16 h, u16 w, u8 align,
+			   struct tcm_area *area);
+static s32 sita_reserve_1d(struct tcm *tcm, u32 slots, struct tcm_area *area);
+static s32 sita_free(struct tcm *tcm, struct tcm_area *area);
+static void sita_deinit(struct tcm *tcm);
+
+/*********************************************
+ *	Main Scanner functions
+ *********************************************/
+static s32 scan_areas_and_find_fit(struct tcm *tcm, u16 w, u16 h, u16 align,
+				   struct tcm_area *area);
+
+static s32 scan_l2r_t2b(struct tcm *tcm, u16 w, u16 h, u16 align,
+			struct tcm_area *field, struct tcm_area *area);
+
+static s32 scan_r2l_t2b(struct tcm *tcm, u16 w, u16 h, u16 align,
+			struct tcm_area *field, struct tcm_area *area);
+
+static s32 scan_r2l_b2t_one_dim(struct tcm *tcm, u32 num_slots,
+			struct tcm_area *field, struct tcm_area *area);
+
+/*********************************************
+ *	Support Infrastructure Methods
+ *********************************************/
+static s32 is_area_free(struct tcm_area ***map, u16 x0, u16 y0, u16 w, u16 h);
+
+static s32 update_candidate(struct tcm *tcm, u16 x0, u16 y0, u16 w, u16 h,
+			    struct tcm_area *field, s32 criteria,
+			    struct score *best);
+
+static void get_nearness_factor(struct tcm_area *field,
+				struct tcm_area *candidate,
+				struct nearness_factor *nf);
+
+static void get_neighbor_stats(struct tcm *tcm, struct tcm_area *area,
+			       struct neighbor_stats *stat);
+
+static void fill_area(struct tcm *tcm,
+				struct tcm_area *area, struct tcm_area *parent);
+
+
+/*********************************************/
+
+/*********************************************
+ *	Utility Methods
+ *********************************************/
+struct tcm *sita_init(u16 width, u16 height, struct tcm_pt *attr)
+{
+	struct tcm *tcm;
+	struct sita_pvt *pvt;
+	struct tcm_area area = {0};
+	s32 i;
+
+	if (width == 0 || height == 0)
+		return NULL;
+
+	tcm = kmalloc(sizeof(*tcm), GFP_KERNEL);
+	pvt = kmalloc(sizeof(*pvt), GFP_KERNEL);
+	if (!tcm || !pvt)
+		goto error;
+
+	memset(tcm, 0, sizeof(*tcm));
+	memset(pvt, 0, sizeof(*pvt));
+
+	/* Updating the pointers to SiTA implementation APIs */
+	tcm->height = height;
+	tcm->width = width;
+	tcm->reserve_2d = sita_reserve_2d;
+	tcm->reserve_1d = sita_reserve_1d;
+	tcm->free = sita_free;
+	tcm->deinit = sita_deinit;
+	tcm->pvt = (void *)pvt;
+
+	spin_lock_init(&(pvt->lock));
+
+	/* Creating tam map */
+	pvt->map = kmalloc(sizeof(*pvt->map) * tcm->width, GFP_KERNEL);
+	if (!pvt->map)
+		goto error;
+
+	for (i = 0; i < tcm->width; i++) {
+		pvt->map[i] =
+			kmalloc(sizeof(**pvt->map) * tcm->height,
+								GFP_KERNEL);
+		if (pvt->map[i] == NULL) {
+			while (i--)
+				kfree(pvt->map[i]);
+			kfree(pvt->map);
+			goto error;
+		}
+	}
+
+	if (attr && attr->x <= tcm->width && attr->y <= tcm->height) {
+		pvt->div_pt.x = attr->x;
+		pvt->div_pt.y = attr->y;
+
+	} else {
+		/* Defaulting to 3:1 ratio on width for 2D area split */
+		/* Defaulting to 3:1 ratio on height for 2D and 1D split */
+		pvt->div_pt.x = (tcm->width * 3) / 4;
+		pvt->div_pt.y = (tcm->height * 3) / 4;
+	}
+
+	spin_lock(&(pvt->lock));
+	assign(&area, 0, 0, width - 1, height - 1);
+	fill_area(tcm, &area, NULL);
+	spin_unlock(&(pvt->lock));
+	return tcm;
+
+error:
+	kfree(tcm);
+	kfree(pvt);
+	return NULL;
+}
+
+static void sita_deinit(struct tcm *tcm)
+{
+	struct sita_pvt *pvt = (struct sita_pvt *)tcm->pvt;
+	struct tcm_area area = {0};
+	s32 i;
+
+	area.p1.x = tcm->width - 1;
+	area.p1.y = tcm->height - 1;
+
+	spin_lock(&(pvt->lock));
+	fill_area(tcm, &area, NULL);
+	spin_unlock(&(pvt->lock));
+
+	for (i = 0; i < tcm->height; i++)
+		kfree(pvt->map[i]);
+	kfree(pvt->map);
+	kfree(pvt);
+}
+
+/**
+ * Reserve a 1D area in the container
+ *
+ * @param num_slots	size of 1D area
+ * @param area		pointer to the area that will be populated with the
+ *			reserved area
+ *
+ * @return 0 on success, non-0 error value on failure.
+ */
+static s32 sita_reserve_1d(struct tcm *tcm, u32 num_slots,
+			   struct tcm_area *area)
+{
+	s32 ret;
+	struct tcm_area field = {0};
+	struct sita_pvt *pvt = (struct sita_pvt *)tcm->pvt;
+
+	spin_lock(&(pvt->lock));
+
+	/* Scanning entire container */
+	assign(&field, tcm->width - 1, tcm->height - 1, 0, 0);
+
+	ret = scan_r2l_b2t_one_dim(tcm, num_slots, &field, area);
+	if (!ret)
+		/* update map */
+		fill_area(tcm, area, area);
+
+	spin_unlock(&(pvt->lock));
+	return ret;
+}
+
+/**
+ * Reserve a 2D area in the container
+ *
+ * @param w	width
+ * @param h	height
+ * @param area	pointer to the area that will be populated with the reserved
+ *		area
+ *
+ * @return 0 on success, non-0 error value on failure.
+ */
+static s32 sita_reserve_2d(struct tcm *tcm, u16 h, u16 w, u8 align,
+			   struct tcm_area *area)
+{
+	s32 ret;
+	struct sita_pvt *pvt = (struct sita_pvt *)tcm->pvt;
+
+	/* not supporting more than 64 as alignment */
+	if (align > 64)
+		return -EINVAL;
+
+	/* we prefer 1, 32 and 64 as alignment */
+	align = align <= 1 ? 1 : align <= 32 ? 32 : 64;
+
+	spin_lock(&(pvt->lock));
+	ret = scan_areas_and_find_fit(tcm, w, h, align, area);
+	if (!ret)
+		/* update map */
+		fill_area(tcm, area, area);
+
+	spin_unlock(&(pvt->lock));
+	return ret;
+}
+
+/**
+ * Unreserve a previously allocated 2D or 1D area
+ * @param area	area to be freed
+ * @return 0 - success
+ */
+static s32 sita_free(struct tcm *tcm, struct tcm_area *area)
+{
+	struct sita_pvt *pvt = (struct sita_pvt *)tcm->pvt;
+
+	spin_lock(&(pvt->lock));
+
+	/* check that this is in fact an existing area */
+	WARN_ON(pvt->map[area->p0.x][area->p0.y] != area ||
+		pvt->map[area->p1.x][area->p1.y] != area);
+
+	/* Clear the contents of the associated tiles in the map */
+	fill_area(tcm, area, NULL);
+
+	spin_unlock(&(pvt->lock));
+
+	return 0;
+}
+
+/**
+ * Note: In general the cordinates in the scan field area relevant to the can
+ * sweep directions. The scan origin (e.g. top-left corner) will always be
+ * the p0 member of the field.  Therfore, for a scan from top-left p0.x <= p1.x
+ * and p0.y <= p1.y; whereas, for a scan from bottom-right p1.x <= p0.x and p1.y
+ * <= p0.y
+ */
+
+/**
+ * Raster scan horizontally right to left from top to bottom to find a place for
+ * a 2D area of given size inside a scan field.
+ *
+ * @param w	width of desired area
+ * @param h	height of desired area
+ * @param align	desired area alignment
+ * @param area	pointer to the area that will be set to the best position
+ * @param field	area to scan (inclusive)
+ *
+ * @return 0 on success, non-0 error value on failure.
+ */
+static s32 scan_r2l_t2b(struct tcm *tcm, u16 w, u16 h, u16 align,
+			struct tcm_area *field, struct tcm_area *area)
+{
+	s32 x, y;
+	s16 start_x, end_x, start_y, end_y, found_x = -1;
+	struct tcm_area ***map = ((struct sita_pvt *)tcm->pvt)->map;
+	struct score best = {{0}, {0}, {0}, 0};
+
+	start_x = field->p0.x;
+	end_x = field->p1.x;
+	start_y = field->p0.y;
+	end_y = field->p1.y;
+
+	/* check scan area co-ordinates */
+	if (field->p0.x < field->p1.x ||
+	    field->p1.y < field->p0.y)
+		return -EINVAL;
+
+	/* check if allocation would fit in scan area */
+	if (w > LEN(start_x, end_x) || h > LEN(end_y, start_y))
+		return -ENOSPC;
+
+	/* adjust start_x and end_y, as allocation would not fit beyond */
+	start_x = ALIGN_DOWN(start_x - w + 1, align); /* - 1 to be inclusive */
+	end_y = end_y - h + 1;
+
+	/* check if allocation would still fit in scan area */
+	if (start_x < end_x)
+		return -ENOSPC;
+
+	/* scan field top-to-bottom, right-to-left */
+	for (y = start_y; y <= end_y; y++) {
+		for (x = start_x; x >= end_x; x -= align) {
+			if (is_area_free(map, x, y, w, h)) {
+				found_x = x;
+
+				/* update best candidate */
+				if (update_candidate(tcm, x, y, w, h, field,
+							CR_R2L_T2B, &best))
+					goto done;
+
+				/* change upper x bound */
+				end_x = x + 1;
+				break;
+			} else if (map[x][y] && map[x][y]->is2d) {
+				/* step over 2D areas */
+				x = ALIGN(map[x][y]->p0.x - w + 1, align);
+			}
+		}
+
+		/* break if you find a free area shouldering the scan field */
+		if (found_x == start_x)
+			break;
+	}
+
+	if (!best.a.tcm)
+		return -ENOSPC;
+done:
+	assign(area, best.a.p0.x, best.a.p0.y, best.a.p1.x, best.a.p1.y);
+	return 0;
+}
+
+/**
+ * Raster scan horizontally left to right from top to bottom to find a place for
+ * a 2D area of given size inside a scan field.
+ *
+ * @param w	width of desired area
+ * @param h	height of desired area
+ * @param align	desired area alignment
+ * @param area	pointer to the area that will be set to the best position
+ * @param field	area to scan (inclusive)
+ *
+ * @return 0 on success, non-0 error value on failure.
+ */
+static s32 scan_l2r_t2b(struct tcm *tcm, u16 w, u16 h, u16 align,
+			struct tcm_area *field, struct tcm_area *area)
+{
+	s32 x, y;
+	s16 start_x, end_x, start_y, end_y, found_x = -1;
+	struct tcm_area ***map = ((struct sita_pvt *)tcm->pvt)->map;
+	struct score best = {{0}, {0}, {0}, 0};
+
+	start_x = field->p0.x;
+	end_x = field->p1.x;
+	start_y = field->p0.y;
+	end_y = field->p1.y;
+
+	/* check scan area co-ordinates */
+	if (field->p1.x < field->p0.x ||
+	    field->p1.y < field->p0.y)
+		return -EINVAL;
+
+	/* check if allocation would fit in scan area */
+	if (w > LEN(end_x, start_x) || h > LEN(end_y, start_y))
+		return -ENOSPC;
+
+	start_x = ALIGN(start_x, align);
+
+	/* check if allocation would still fit in scan area */
+	if (w > LEN(end_x, start_x))
+		return -ENOSPC;
+
+	/* adjust end_x and end_y, as allocation would not fit beyond */
+	end_x = end_x - w + 1; /* + 1 to be inclusive */
+	end_y = end_y - h + 1;
+
+	/* scan field top-to-bottom, left-to-right */
+	for (y = start_y; y <= end_y; y++) {
+		for (x = start_x; x <= end_x; x += align) {
+			if (is_area_free(map, x, y, w, h)) {
+				found_x = x;
+
+				/* update best candidate */
+				if (update_candidate(tcm, x, y, w, h, field,
+							CR_L2R_T2B, &best))
+					goto done;
+				/* change upper x bound */
+				end_x = x - 1;
+
+				break;
+			} else if (map[x][y] && map[x][y]->is2d) {
+				/* step over 2D areas */
+				x = ALIGN_DOWN(map[x][y]->p1.x, align);
+			}
+		}
+
+		/* break if you find a free area shouldering the scan field */
+		if (found_x == start_x)
+			break;
+	}
+
+	if (!best.a.tcm)
+		return -ENOSPC;
+done:
+	assign(area, best.a.p0.x, best.a.p0.y, best.a.p1.x, best.a.p1.y);
+	return 0;
+}
+
+/**
+ * Raster scan horizontally right to left from bottom to top to find a place
+ * for a 1D area of given size inside a scan field.
+ *
+ * @param num_slots	size of desired area
+ * @param align		desired area alignment
+ * @param area		pointer to the area that will be set to the best
+ *			position
+ * @param field		area to scan (inclusive)
+ *
+ * @return 0 on success, non-0 error value on failure.
+ */
+static s32 scan_r2l_b2t_one_dim(struct tcm *tcm, u32 num_slots,
+				struct tcm_area *field, struct tcm_area *area)
+{
+	s32 found = 0;
+	s16 x, y;
+	struct sita_pvt *pvt = (struct sita_pvt *)tcm->pvt;
+	struct tcm_area *p;
+
+	/* check scan area co-ordinates */
+	if (field->p0.y < field->p1.y)
+		return -EINVAL;
+
+	/**
+	 * Currently we only support full width 1D scan field, which makes sense
+	 * since 1D slot-ordering spans the full container width.
+	 */
+	if (tcm->width != field->p0.x - field->p1.x + 1)
+		return -EINVAL;
+
+	/* check if allocation would fit in scan area */
+	if (num_slots > tcm->width * LEN(field->p0.y, field->p1.y))
+		return -ENOSPC;
+
+	x = field->p0.x;
+	y = field->p0.y;
+
+	/* find num_slots consecutive free slots to the left */
+	while (found < num_slots) {
+		if (y < 0)
+			return -ENOSPC;
+
+		/* remember bottom-right corner */
+		if (found == 0) {
+			area->p1.x = x;
+			area->p1.y = y;
+		}
+
+		/* skip busy regions */
+		p = pvt->map[x][y];
+		if (p) {
+			/* move to left of 2D areas, top left of 1D */
+			x = p->p0.x;
+			if (!p->is2d)
+				y = p->p0.y;
+
+			/* start over */
+			found = 0;
+		} else {
+			/* count consecutive free slots */
+			found++;
+			if (found == num_slots)
+				break;
+		}
+
+		/* move to the left */
+		if (x == 0)
+			y--;
+		x = (x ? : tcm->width) - 1;
+
+	}
+
+	/* set top-left corner */
+	area->p0.x = x;
+	area->p0.y = y;
+	return 0;
+}
+
+/**
+ * Find a place for a 2D area of given size inside a scan field based on its
+ * alignment needs.
+ *
+ * @param w	width of desired area
+ * @param h	height of desired area
+ * @param align	desired area alignment
+ * @param area	pointer to the area that will be set to the best position
+ *
+ * @return 0 on success, non-0 error value on failure.
+ */
+static s32 scan_areas_and_find_fit(struct tcm *tcm, u16 w, u16 h, u16 align,
+				   struct tcm_area *area)
+{
+	s32 ret = 0;
+	struct tcm_area field = {0};
+	u16 boundary_x, boundary_y;
+	struct sita_pvt *pvt = (struct sita_pvt *)tcm->pvt;
+
+	if (align > 1) {
+		/* prefer top-left corner */
+		boundary_x = pvt->div_pt.x - 1;
+		boundary_y = pvt->div_pt.y - 1;
+
+		/* expand width and height if needed */
+		if (w > pvt->div_pt.x)
+			boundary_x = tcm->width - 1;
+		if (h > pvt->div_pt.y)
+			boundary_y = tcm->height - 1;
+
+		assign(&field, 0, 0, boundary_x, boundary_y);
+		ret = scan_l2r_t2b(tcm, w, h, align, &field, area);
+
+		/* scan whole container if failed, but do not scan 2x */
+		if (ret != 0 && (boundary_x != tcm->width - 1 ||
+				 boundary_y != tcm->height - 1)) {
+			/* scan the entire container if nothing found */
+			assign(&field, 0, 0, tcm->width - 1, tcm->height - 1);
+			ret = scan_l2r_t2b(tcm, w, h, align, &field, area);
+		}
+	} else if (align == 1) {
+		/* prefer top-right corner */
+		boundary_x = pvt->div_pt.x;
+		boundary_y = pvt->div_pt.y - 1;
+
+		/* expand width and height if needed */
+		if (w > (tcm->width - pvt->div_pt.x))
+			boundary_x = 0;
+		if (h > pvt->div_pt.y)
+			boundary_y = tcm->height - 1;
+
+		assign(&field, tcm->width - 1, 0, boundary_x, boundary_y);
+		ret = scan_r2l_t2b(tcm, w, h, align, &field, area);
+
+		/* scan whole container if failed, but do not scan 2x */
+		if (ret != 0 && (boundary_x != 0 ||
+				 boundary_y != tcm->height - 1)) {
+			/* scan the entire container if nothing found */
+			assign(&field, tcm->width - 1, 0, 0, tcm->height - 1);
+			ret = scan_r2l_t2b(tcm, w, h, align, &field,
+					   area);
+		}
+	}
+
+	return ret;
+}
+
+/* check if an entire area is free */
+static s32 is_area_free(struct tcm_area ***map, u16 x0, u16 y0, u16 w, u16 h)
+{
+	u16 x = 0, y = 0;
+	for (y = y0; y < y0 + h; y++) {
+		for (x = x0; x < x0 + w; x++) {
+			if (map[x][y])
+				return false;
+		}
+	}
+	return true;
+}
+
+/* fills an area with a parent tcm_area */
+static void fill_area(struct tcm *tcm, struct tcm_area *area,
+			struct tcm_area *parent)
+{
+	s32 x, y;
+	struct sita_pvt *pvt = (struct sita_pvt *)tcm->pvt;
+	struct tcm_area a, a_;
+
+	/* set area's tcm; otherwise, enumerator considers it invalid */
+	area->tcm = tcm;
+
+	tcm_for_each_slice(a, *area, a_) {
+		for (x = a.p0.x; x <= a.p1.x; ++x)
+			for (y = a.p0.y; y <= a.p1.y; ++y)
+				pvt->map[x][y] = parent;
+
+	}
+}
+
+/**
+ * Compares a candidate area to the current best area, and if it is a better
+ * fit, it updates the best to this one.
+ *
+ * @param x0, y0, w, h		top, left, width, height of candidate area
+ * @param field			scan field
+ * @param criteria		scan criteria
+ * @param best			best candidate and its scores
+ *
+ * @return 1 (true) if the candidate area is known to be the final best, so no
+ * more searching should be performed
+ */
+static s32 update_candidate(struct tcm *tcm, u16 x0, u16 y0, u16 w, u16 h,
+			    struct tcm_area *field, s32 criteria,
+			    struct score *best)
+{
+	struct score me;	/* score for area */
+
+	/*
+	 * NOTE: For horizontal bias we always give the first found, because our
+	 * scan is horizontal-raster-based and the first candidate will always
+	 * have the horizontal bias.
+	 */
+	bool first = criteria & CR_BIAS_HORIZONTAL;
+
+	assign(&me.a, x0, y0, x0 + w - 1, y0 + h - 1);
+
+	/* calculate score for current candidate */
+	if (!first) {
+		get_neighbor_stats(tcm, &me.a, &me.n);
+		me.neighs = me.n.edge + me.n.busy;
+		get_nearness_factor(field, &me.a, &me.f);
+	}
+
+	/* the 1st candidate is always the best */
+	if (!best->a.tcm)
+		goto better;
+
+	BUG_ON(first);
+
+	/* diagonal balance check */
+	if ((criteria & CR_DIAGONAL_BALANCE) &&
+		best->neighs <= me.neighs &&
+		(best->neighs < me.neighs ||
+		 /* this implies that neighs and occupied match */
+		 best->n.busy < me.n.busy ||
+		 (best->n.busy == me.n.busy &&
+		  /* check the nearness factor */
+		  best->f.x + best->f.y > me.f.x + me.f.y)))
+		goto better;
+
+	/* not better, keep going */
+	return 0;
+
+better:
+	/* save current area as best */
+	memcpy(best, &me, sizeof(me));
+	best->a.tcm = tcm;
+	return first;
+}
+
+/**
+ * Calculate the nearness factor of an area in a search field.  The nearness
+ * factor is smaller if the area is closer to the search origin.
+ */
+static void get_nearness_factor(struct tcm_area *field, struct tcm_area *area,
+				struct nearness_factor *nf)
+{
+	/**
+	 * Using signed math as field coordinates may be reversed if
+	 * search direction is right-to-left or bottom-to-top.
+	 */
+	nf->x = (s32)(area->p0.x - field->p0.x) * 1000 /
+		(field->p1.x - field->p0.x);
+	nf->y = (s32)(area->p0.y - field->p0.y) * 1000 /
+		(field->p1.y - field->p0.y);
+}
+
+/* get neighbor statistics */
+static void get_neighbor_stats(struct tcm *tcm, struct tcm_area *area,
+			 struct neighbor_stats *stat)
+{
+	s16 x = 0, y = 0;
+	struct sita_pvt *pvt = (struct sita_pvt *)tcm->pvt;
+
+	/* Clearing any exisiting values */
+	memset(stat, 0, sizeof(*stat));
+
+	/* process top & bottom edges */
+	for (x = area->p0.x; x <= area->p1.x; x++) {
+		if (area->p0.y == 0)
+			stat->edge++;
+		else if (pvt->map[x][area->p0.y - 1])
+			stat->busy++;
+
+		if (area->p1.y == tcm->height - 1)
+			stat->edge++;
+		else if (pvt->map[x][area->p1.y + 1])
+			stat->busy++;
+	}
+
+	/* process left & right edges */
+	for (y = area->p0.y; y <= area->p1.y; ++y) {
+		if (area->p0.x == 0)
+			stat->edge++;
+		else if (pvt->map[area->p0.x - 1][y])
+			stat->busy++;
+
+		if (area->p1.x == tcm->width - 1)
+			stat->edge++;
+		else if (pvt->map[area->p1.x + 1][y])
+			stat->busy++;
+	}
+}
diff --git a/drivers/gpu/drm/omapdrm/tcm-sita.h b/drivers/gpu/drm/omapdrm/tcm-sita.h
new file mode 100644
index 0000000..0444f86
--- /dev/null
+++ b/drivers/gpu/drm/omapdrm/tcm-sita.h
@@ -0,0 +1,95 @@
+/*
+ * tcm_sita.h
+ *
+ * SImple Tiler Allocator (SiTA) private structures.
+ *
+ * Author: Ravi Ramachandra <r.ramachandra@ti.com>
+ *
+ * Copyright (C) 2009-2011 Texas Instruments, Inc.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * * Redistributions of source code must retain the above copyright
+ *   notice, this list of conditions and the following disclaimer.
+ *
+ * * Redistributions in binary form must reproduce the above copyright
+ *   notice, this list of conditions and the following disclaimer in the
+ *   documentation and/or other materials provided with the distribution.
+ *
+ * * Neither the name of Texas Instruments Incorporated nor the names of
+ *   its contributors may be used to endorse or promote products derived
+ *   from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO,
+ * THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
+ * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
+ * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+ * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
+ * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
+ * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
+ * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
+ * EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#ifndef _TCM_SITA_H
+#define _TCM_SITA_H
+
+#include "tcm.h"
+
+/* length between two coordinates */
+#define LEN(a, b) ((a) > (b) ? (a) - (b) + 1 : (b) - (a) + 1)
+
+enum criteria {
+	CR_MAX_NEIGHS		= 0x01,
+	CR_FIRST_FOUND		= 0x10,
+	CR_BIAS_HORIZONTAL	= 0x20,
+	CR_BIAS_VERTICAL	= 0x40,
+	CR_DIAGONAL_BALANCE	= 0x80
+};
+
+/* nearness to the beginning of the search field from 0 to 1000 */
+struct nearness_factor {
+	s32 x;
+	s32 y;
+};
+
+/*
+ * Statistics on immediately neighboring slots.  Edge is the number of
+ * border segments that are also border segments of the scan field.  Busy
+ * refers to the number of neighbors that are occupied.
+ */
+struct neighbor_stats {
+	u16 edge;
+	u16 busy;
+};
+
+/* structure to keep the score of a potential allocation */
+struct score {
+	struct nearness_factor	f;
+	struct neighbor_stats	n;
+	struct tcm_area		a;
+	u16    neighs;		/* number of busy neighbors */
+};
+
+struct sita_pvt {
+	spinlock_t lock;	/* spinlock to protect access */
+	struct tcm_pt div_pt;	/* divider point splitting container */
+	struct tcm_area ***map;	/* pointers to the parent area for each slot */
+};
+
+/* assign coordinates to area */
+static inline
+void assign(struct tcm_area *a, u16 x0, u16 y0, u16 x1, u16 y1)
+{
+	a->p0.x = x0;
+	a->p0.y = y0;
+	a->p1.x = x1;
+	a->p1.y = y1;
+}
+
+#endif
diff --git a/drivers/gpu/drm/omapdrm/tcm.h b/drivers/gpu/drm/omapdrm/tcm.h
new file mode 100644
index 0000000..a8d5ce4
--- /dev/null
+++ b/drivers/gpu/drm/omapdrm/tcm.h
@@ -0,0 +1,328 @@
+/*
+ * tcm.h
+ *
+ * TILER container manager specification and support functions for TI
+ * TILER driver.
+ *
+ * Author: Lajos Molnar <molnar@ti.com>
+ *
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * * Redistributions of source code must retain the above copyright
+ *   notice, this list of conditions and the following disclaimer.
+ *
+ * * Redistributions in binary form must reproduce the above copyright
+ *   notice, this list of conditions and the following disclaimer in the
+ *   documentation and/or other materials provided with the distribution.
+ *
+ * * Neither the name of Texas Instruments Incorporated nor the names of
+ *   its contributors may be used to endorse or promote products derived
+ *   from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO,
+ * THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
+ * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
+ * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+ * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
+ * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
+ * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
+ * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
+ * EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#ifndef TCM_H
+#define TCM_H
+
+struct tcm;
+
+/* point */
+struct tcm_pt {
+	u16 x;
+	u16 y;
+};
+
+/* 1d or 2d area */
+struct tcm_area {
+	bool is2d;		/* whether area is 1d or 2d */
+	struct tcm    *tcm;	/* parent */
+	struct tcm_pt  p0;
+	struct tcm_pt  p1;
+};
+
+struct tcm {
+	u16 width, height;	/* container dimensions */
+	int lut_id;		/* Lookup table identifier */
+
+	unsigned int y_offset;	/* offset to use for y coordinates */
+
+	/* 'pvt' structure shall contain any tcm details (attr) along with
+	linked list of allocated areas and mutex for mutually exclusive access
+	to the list.  It may also contain copies of width and height to notice
+	any changes to the publicly available width and height fields. */
+	void *pvt;
+
+	/* function table */
+	s32 (*reserve_2d)(struct tcm *tcm, u16 height, u16 width, u8 align,
+			  struct tcm_area *area);
+	s32 (*reserve_1d)(struct tcm *tcm, u32 slots, struct tcm_area *area);
+	s32 (*free)      (struct tcm *tcm, struct tcm_area *area);
+	void (*deinit)   (struct tcm *tcm);
+};
+
+/*=============================================================================
+    BASIC TILER CONTAINER MANAGER INTERFACE
+=============================================================================*/
+
+/*
+ * NOTE:
+ *
+ * Since some basic parameter checking is done outside the TCM algorithms,
+ * TCM implementation do NOT have to check the following:
+ *
+ *   area pointer is NULL
+ *   width and height fits within container
+ *   number of pages is more than the size of the container
+ *
+ */
+
+struct tcm *sita_init(u16 width, u16 height, struct tcm_pt *attr);
+
+
+/**
+ * Deinitialize tiler container manager.
+ *
+ * @param tcm	Pointer to container manager.
+ *
+ * @return 0 on success, non-0 error value on error.  The call
+ *	   should free as much memory as possible and meaningful
+ *	   even on failure.  Some error codes: -ENODEV: invalid
+ *	   manager.
+ */
+static inline void tcm_deinit(struct tcm *tcm)
+{
+	if (tcm)
+		tcm->deinit(tcm);
+}
+
+/**
+ * Reserves a 2D area in the container.
+ *
+ * @param tcm		Pointer to container manager.
+ * @param height	Height(in pages) of area to be reserved.
+ * @param width		Width(in pages) of area to be reserved.
+ * @param align		Alignment requirement for top-left corner of area. Not
+ *			all values may be supported by the container manager,
+ *			but it must support 0 (1), 32 and 64.
+ *			0 value is equivalent to 1.
+ * @param area		Pointer to where the reserved area should be stored.
+ *
+ * @return 0 on success.  Non-0 error code on failure.  Also,
+ *	   the tcm field of the area will be set to NULL on
+ *	   failure.  Some error codes: -ENODEV: invalid manager,
+ *	   -EINVAL: invalid area, -ENOMEM: not enough space for
+ *	    allocation.
+ */
+static inline s32 tcm_reserve_2d(struct tcm *tcm, u16 width, u16 height,
+				 u16 align, struct tcm_area *area)
+{
+	/* perform rudimentary error checking */
+	s32 res = tcm  == NULL ? -ENODEV :
+		(area == NULL || width == 0 || height == 0 ||
+		 /* align must be a 2 power */
+		 (align & (align - 1))) ? -EINVAL :
+		(height > tcm->height || width > tcm->width) ? -ENOMEM : 0;
+
+	if (!res) {
+		area->is2d = true;
+		res = tcm->reserve_2d(tcm, height, width, align, area);
+		area->tcm = res ? NULL : tcm;
+	}
+
+	return res;
+}
+
+/**
+ * Reserves a 1D area in the container.
+ *
+ * @param tcm		Pointer to container manager.
+ * @param slots		Number of (contiguous) slots to reserve.
+ * @param area		Pointer to where the reserved area should be stored.
+ *
+ * @return 0 on success.  Non-0 error code on failure.  Also,
+ *	   the tcm field of the area will be set to NULL on
+ *	   failure.  Some error codes: -ENODEV: invalid manager,
+ *	   -EINVAL: invalid area, -ENOMEM: not enough space for
+ *	    allocation.
+ */
+static inline s32 tcm_reserve_1d(struct tcm *tcm, u32 slots,
+				 struct tcm_area *area)
+{
+	/* perform rudimentary error checking */
+	s32 res = tcm  == NULL ? -ENODEV :
+		(area == NULL || slots == 0) ? -EINVAL :
+		slots > (tcm->width * (u32) tcm->height) ? -ENOMEM : 0;
+
+	if (!res) {
+		area->is2d = false;
+		res = tcm->reserve_1d(tcm, slots, area);
+		area->tcm = res ? NULL : tcm;
+	}
+
+	return res;
+}
+
+/**
+ * Free a previously reserved area from the container.
+ *
+ * @param area	Pointer to area reserved by a prior call to
+ *		tcm_reserve_1d or tcm_reserve_2d call, whether
+ *		it was successful or not. (Note: all fields of
+ *		the structure must match.)
+ *
+ * @return 0 on success.  Non-0 error code on failure.  Also, the tcm
+ *	   field of the area is set to NULL on success to avoid subsequent
+ *	   freeing.  This call will succeed even if supplying
+ *	   the area from a failed reserved call.
+ */
+static inline s32 tcm_free(struct tcm_area *area)
+{
+	s32 res = 0; /* free succeeds by default */
+
+	if (area && area->tcm) {
+		res = area->tcm->free(area->tcm, area);
+		if (res == 0)
+			area->tcm = NULL;
+	}
+
+	return res;
+}
+
+/*=============================================================================
+    HELPER FUNCTION FOR ANY TILER CONTAINER MANAGER
+=============================================================================*/
+
+/**
+ * This method slices off the topmost 2D slice from the parent area, and stores
+ * it in the 'slice' parameter.  The 'parent' parameter will get modified to
+ * contain the remaining portion of the area.  If the whole parent area can
+ * fit in a 2D slice, its tcm pointer is set to NULL to mark that it is no
+ * longer a valid area.
+ *
+ * @param parent	Pointer to a VALID parent area that will get modified
+ * @param slice		Pointer to the slice area that will get modified
+ */
+static inline void tcm_slice(struct tcm_area *parent, struct tcm_area *slice)
+{
+	*slice = *parent;
+
+	/* check if we need to slice */
+	if (slice->tcm && !slice->is2d &&
+		slice->p0.y != slice->p1.y &&
+		(slice->p0.x || (slice->p1.x != slice->tcm->width - 1))) {
+		/* set end point of slice (start always remains) */
+		slice->p1.x = slice->tcm->width - 1;
+		slice->p1.y = (slice->p0.x) ? slice->p0.y : slice->p1.y - 1;
+		/* adjust remaining area */
+		parent->p0.x = 0;
+		parent->p0.y = slice->p1.y + 1;
+	} else {
+		/* mark this as the last slice */
+		parent->tcm = NULL;
+	}
+}
+
+/* Verify if a tcm area is logically valid */
+static inline bool tcm_area_is_valid(struct tcm_area *area)
+{
+	return area && area->tcm &&
+		/* coordinate bounds */
+		area->p1.x < area->tcm->width &&
+		area->p1.y < area->tcm->height &&
+		area->p0.y <= area->p1.y &&
+		/* 1D coordinate relationship + p0.x check */
+		((!area->is2d &&
+		  area->p0.x < area->tcm->width &&
+		  area->p0.x + area->p0.y * area->tcm->width <=
+		  area->p1.x + area->p1.y * area->tcm->width) ||
+		 /* 2D coordinate relationship */
+		 (area->is2d &&
+		  area->p0.x <= area->p1.x));
+}
+
+/* see if a coordinate is within an area */
+static inline bool __tcm_is_in(struct tcm_pt *p, struct tcm_area *a)
+{
+	u16 i;
+
+	if (a->is2d) {
+		return p->x >= a->p0.x && p->x <= a->p1.x &&
+		       p->y >= a->p0.y && p->y <= a->p1.y;
+	} else {
+		i = p->x + p->y * a->tcm->width;
+		return i >= a->p0.x + a->p0.y * a->tcm->width &&
+		       i <= a->p1.x + a->p1.y * a->tcm->width;
+	}
+}
+
+/* calculate area width */
+static inline u16 __tcm_area_width(struct tcm_area *area)
+{
+	return area->p1.x - area->p0.x + 1;
+}
+
+/* calculate area height */
+static inline u16 __tcm_area_height(struct tcm_area *area)
+{
+	return area->p1.y - area->p0.y + 1;
+}
+
+/* calculate number of slots in an area */
+static inline u16 __tcm_sizeof(struct tcm_area *area)
+{
+	return area->is2d ?
+		__tcm_area_width(area) * __tcm_area_height(area) :
+		(area->p1.x - area->p0.x + 1) + (area->p1.y - area->p0.y) *
+							area->tcm->width;
+}
+#define tcm_sizeof(area) __tcm_sizeof(&(area))
+#define tcm_awidth(area) __tcm_area_width(&(area))
+#define tcm_aheight(area) __tcm_area_height(&(area))
+#define tcm_is_in(pt, area) __tcm_is_in(&(pt), &(area))
+
+/* limit a 1D area to the first N pages */
+static inline s32 tcm_1d_limit(struct tcm_area *a, u32 num_pg)
+{
+	if (__tcm_sizeof(a) < num_pg)
+		return -ENOMEM;
+	if (!num_pg)
+		return -EINVAL;
+
+	a->p1.x = (a->p0.x + num_pg - 1) % a->tcm->width;
+	a->p1.y = a->p0.y + ((a->p0.x + num_pg - 1) / a->tcm->width);
+	return 0;
+}
+
+/**
+ * Iterate through 2D slices of a valid area. Behaves
+ * syntactically as a for(;;) statement.
+ *
+ * @param var		Name of a local variable of type 'struct
+ *			tcm_area *' that will get modified to
+ *			contain each slice.
+ * @param area		Pointer to the VALID parent area. This
+ *			structure will not get modified
+ *			throughout the loop.
+ *
+ */
+#define tcm_for_each_slice(var, area, safe) \
+	for (safe = area, \
+	     tcm_slice(&safe, &var); \
+	     var.tcm; tcm_slice(&safe, &var))
+
+#endif
diff --git a/drivers/staging/Kconfig b/drivers/staging/Kconfig
index 329bdb4..eca907b 100644
--- a/drivers/staging/Kconfig
+++ b/drivers/staging/Kconfig
@@ -114,8 +114,6 @@ source "drivers/staging/media/Kconfig"
 
 source "drivers/staging/net/Kconfig"
 
-source "drivers/staging/omapdrm/Kconfig"
-
 source "drivers/staging/android/Kconfig"
 
 source "drivers/staging/ozwpan/Kconfig"
diff --git a/drivers/staging/Makefile b/drivers/staging/Makefile
index c7ec486..d810ed7 100644
--- a/drivers/staging/Makefile
+++ b/drivers/staging/Makefile
@@ -49,7 +49,6 @@ obj-$(CONFIG_SPEAKUP)		+= speakup/
 obj-$(CONFIG_TOUCHSCREEN_CLEARPAD_TM1217)	+= cptm1217/
 obj-$(CONFIG_TOUCHSCREEN_SYNAPTICS_I2C_RMI4)	+= ste_rmi4/
 obj-$(CONFIG_MFD_NVEC)		+= nvec/
-obj-$(CONFIG_DRM_OMAP)		+= omapdrm/
 obj-$(CONFIG_ANDROID)		+= android/
 obj-$(CONFIG_USB_WPAN_HCD)	+= ozwpan/
 obj-$(CONFIG_USB_G_CCG)		+= ccg/
diff --git a/drivers/staging/omapdrm/Kconfig b/drivers/staging/omapdrm/Kconfig
deleted file mode 100644
index b724a41..0000000
--- a/drivers/staging/omapdrm/Kconfig
+++ /dev/null
@@ -1,25 +0,0 @@
-
-config DRM_OMAP
-	tristate "OMAP DRM"
-	depends on DRM && !CONFIG_FB_OMAP2
-	depends on ARCH_OMAP2PLUS || ARCH_MULTIPLATFORM
-	select DRM_KMS_HELPER
-	select OMAP2_DSS
-	select FB_SYS_FILLRECT
-	select FB_SYS_COPYAREA
-	select FB_SYS_IMAGEBLIT
-	select FB_SYS_FOPS
-	default n
-	help
-	  DRM display driver for OMAP2/3/4 based boards.
-
-config DRM_OMAP_NUM_CRTCS
-	int "Number of CRTCs"
-	range 1 10
-	default 1  if ARCH_OMAP2 || ARCH_OMAP3
-	default 2  if ARCH_OMAP4
-	depends on DRM_OMAP
-	help
-	  Select the number of video overlays which can be used as framebuffers.
-	  The remaining overlays are reserved for video.
-
diff --git a/drivers/staging/omapdrm/Makefile b/drivers/staging/omapdrm/Makefile
deleted file mode 100644
index d85e058..0000000
--- a/drivers/staging/omapdrm/Makefile
+++ /dev/null
@@ -1,24 +0,0 @@
-#
-# Makefile for the drm device driver.  This driver provides support for the
-# Direct Rendering Infrastructure (DRI)
-#
-
-ccflags-y := -Iinclude/drm -Werror
-omapdrm-y := omap_drv.o \
-	omap_irq.o \
-	omap_debugfs.o \
-	omap_crtc.o \
-	omap_plane.o \
-	omap_encoder.o \
-	omap_connector.o \
-	omap_fb.o \
-	omap_fbdev.o \
-	omap_gem.o \
-	omap_gem_dmabuf.o \
-	omap_dmm_tiler.o \
-	tcm-sita.o
-
-# temporary:
-omapdrm-y += omap_gem_helpers.o
-
-obj-$(CONFIG_DRM_OMAP)	+= omapdrm.o
diff --git a/drivers/staging/omapdrm/TODO b/drivers/staging/omapdrm/TODO
deleted file mode 100644
index abeeb00..0000000
--- a/drivers/staging/omapdrm/TODO
+++ /dev/null
@@ -1,32 +0,0 @@
-TODO
-. add video decode/encode support (via syslink3 + codec-engine)
-  . NOTE: with dmabuf this probably could be split into different driver
-    so perhaps this TODO doesn't belong here
-. where should we do eviction (detatch_pages())?  We aren't necessarily
-  accessing the pages via a GART, so maybe we need some other threshold
-  to put a cap on the # of pages that can be pin'd.  (It is mostly only
-  of interest in case you have a swap partition/file.. which a lot of
-  these devices do not.. but it doesn't hurt for the driver to do the
-  right thing anyways.)
-  . Use mm_shrinker to trigger unpinning pages.  Need to figure out how
-    to handle next issue first (I think?)
-  . Note TTM already has some mm_shrinker stuff..  maybe an argument to
-    move to TTM?  Or maybe something that could be factored out in common?
-. GEM/shmem backed pages can have existing mappings (kernel linear map,
-  etc..), which isn't really ideal.
-. Revisit GEM sync object infrastructure.. TTM has some framework for this
-  already.  Possibly this could be refactored out and made more common?
-  There should be some way to do this with less wheel-reinvention.
-. Solve PM sequencing on resume.  DMM/TILER must be reloaded before any
-  access is made from any component in the system.  Which means on suspend
-  CRTC's should be disabled, and on resume the LUT should be reprogrammed
-  before CRTC's are re-enabled, to prevent DSS from trying to DMA from a
-  buffer mapped in DMM/TILER before LUT is reloaded.
-
-Userspace:
-. git://github.com/robclark/xf86-video-omap.git
-
-Currently tested on
-. OMAP3530 beagleboard
-. OMAP4430 pandaboard
-. OMAP4460 pandaboard
diff --git a/drivers/staging/omapdrm/omap_connector.c b/drivers/staging/omapdrm/omap_connector.c
deleted file mode 100644
index 4cc9ee7..0000000
--- a/drivers/staging/omapdrm/omap_connector.c
+++ /dev/null
@@ -1,298 +0,0 @@
-/*
- * drivers/staging/omapdrm/omap_connector.c
- *
- * Copyright (C) 2011 Texas Instruments
- * Author: Rob Clark <rob@ti.com>
- *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of the GNU General Public License version 2 as published by
- * the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but WITHOUT
- * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
- * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
- * more details.
- *
- * You should have received a copy of the GNU General Public License along with
- * this program.  If not, see <http://www.gnu.org/licenses/>.
- */
-
-#include "omap_drv.h"
-
-#include "drm_crtc.h"
-#include "drm_crtc_helper.h"
-
-/*
- * connector funcs
- */
-
-#define to_omap_connector(x) container_of(x, struct omap_connector, base)
-
-struct omap_connector {
-	struct drm_connector base;
-	struct omap_dss_device *dssdev;
-	struct drm_encoder *encoder;
-};
-
-void copy_timings_omap_to_drm(struct drm_display_mode *mode,
-		struct omap_video_timings *timings)
-{
-	mode->clock = timings->pixel_clock;
-
-	mode->hdisplay = timings->x_res;
-	mode->hsync_start = mode->hdisplay + timings->hfp;
-	mode->hsync_end = mode->hsync_start + timings->hsw;
-	mode->htotal = mode->hsync_end + timings->hbp;
-
-	mode->vdisplay = timings->y_res;
-	mode->vsync_start = mode->vdisplay + timings->vfp;
-	mode->vsync_end = mode->vsync_start + timings->vsw;
-	mode->vtotal = mode->vsync_end + timings->vbp;
-
-	mode->flags = 0;
-
-	if (timings->interlace)
-		mode->flags |= DRM_MODE_FLAG_INTERLACE;
-
-	if (timings->hsync_level == OMAPDSS_SIG_ACTIVE_HIGH)
-		mode->flags |= DRM_MODE_FLAG_PHSYNC;
-	else
-		mode->flags |= DRM_MODE_FLAG_NHSYNC;
-
-	if (timings->vsync_level == OMAPDSS_SIG_ACTIVE_HIGH)
-		mode->flags |= DRM_MODE_FLAG_PVSYNC;
-	else
-		mode->flags |= DRM_MODE_FLAG_NVSYNC;
-}
-
-void copy_timings_drm_to_omap(struct omap_video_timings *timings,
-		struct drm_display_mode *mode)
-{
-	timings->pixel_clock = mode->clock;
-
-	timings->x_res = mode->hdisplay;
-	timings->hfp = mode->hsync_start - mode->hdisplay;
-	timings->hsw = mode->hsync_end - mode->hsync_start;
-	timings->hbp = mode->htotal - mode->hsync_end;
-
-	timings->y_res = mode->vdisplay;
-	timings->vfp = mode->vsync_start - mode->vdisplay;
-	timings->vsw = mode->vsync_end - mode->vsync_start;
-	timings->vbp = mode->vtotal - mode->vsync_end;
-
-	timings->interlace = !!(mode->flags & DRM_MODE_FLAG_INTERLACE);
-
-	if (mode->flags & DRM_MODE_FLAG_PHSYNC)
-		timings->hsync_level = OMAPDSS_SIG_ACTIVE_HIGH;
-	else
-		timings->hsync_level = OMAPDSS_SIG_ACTIVE_LOW;
-
-	if (mode->flags & DRM_MODE_FLAG_PVSYNC)
-		timings->vsync_level = OMAPDSS_SIG_ACTIVE_HIGH;
-	else
-		timings->vsync_level = OMAPDSS_SIG_ACTIVE_LOW;
-
-	timings->data_pclk_edge = OMAPDSS_DRIVE_SIG_RISING_EDGE;
-	timings->de_level = OMAPDSS_SIG_ACTIVE_HIGH;
-	timings->sync_pclk_edge = OMAPDSS_DRIVE_SIG_OPPOSITE_EDGES;
-}
-
-static enum drm_connector_status omap_connector_detect(
-		struct drm_connector *connector, bool force)
-{
-	struct omap_connector *omap_connector = to_omap_connector(connector);
-	struct omap_dss_device *dssdev = omap_connector->dssdev;
-	struct omap_dss_driver *dssdrv = dssdev->driver;
-	enum drm_connector_status ret;
-
-	if (dssdrv->detect) {
-		if (dssdrv->detect(dssdev))
-			ret = connector_status_connected;
-		else
-			ret = connector_status_disconnected;
-	} else {
-		ret = connector_status_unknown;
-	}
-
-	VERB("%s: %d (force=%d)", omap_connector->dssdev->name, ret, force);
-
-	return ret;
-}
-
-static void omap_connector_destroy(struct drm_connector *connector)
-{
-	struct omap_connector *omap_connector = to_omap_connector(connector);
-	struct omap_dss_device *dssdev = omap_connector->dssdev;
-
-	DBG("%s", omap_connector->dssdev->name);
-	drm_sysfs_connector_remove(connector);
-	drm_connector_cleanup(connector);
-	kfree(omap_connector);
-
-	omap_dss_put_device(dssdev);
-}
-
-#define MAX_EDID  512
-
-static int omap_connector_get_modes(struct drm_connector *connector)
-{
-	struct omap_connector *omap_connector = to_omap_connector(connector);
-	struct omap_dss_device *dssdev = omap_connector->dssdev;
-	struct omap_dss_driver *dssdrv = dssdev->driver;
-	struct drm_device *dev = connector->dev;
-	int n = 0;
-
-	DBG("%s", omap_connector->dssdev->name);
-
-	/* if display exposes EDID, then we parse that in the normal way to
-	 * build table of supported modes.. otherwise (ie. fixed resolution
-	 * LCD panels) we just return a single mode corresponding to the
-	 * currently configured timings:
-	 */
-	if (dssdrv->read_edid) {
-		void *edid = kzalloc(MAX_EDID, GFP_KERNEL);
-
-		if ((dssdrv->read_edid(dssdev, edid, MAX_EDID) > 0) &&
-				drm_edid_is_valid(edid)) {
-			drm_mode_connector_update_edid_property(
-					connector, edid);
-			n = drm_add_edid_modes(connector, edid);
-		} else {
-			drm_mode_connector_update_edid_property(
-					connector, NULL);
-		}
-		kfree(edid);
-	} else {
-		struct drm_display_mode *mode = drm_mode_create(dev);
-		struct omap_video_timings timings = {0};
-
-		dssdrv->get_timings(dssdev, &timings);
-
-		copy_timings_omap_to_drm(mode, &timings);
-
-		mode->type = DRM_MODE_TYPE_DRIVER | DRM_MODE_TYPE_PREFERRED;
-		drm_mode_set_name(mode);
-		drm_mode_probed_add(connector, mode);
-
-		n = 1;
-	}
-
-	return n;
-}
-
-static int omap_connector_mode_valid(struct drm_connector *connector,
-				 struct drm_display_mode *mode)
-{
-	struct omap_connector *omap_connector = to_omap_connector(connector);
-	struct omap_dss_device *dssdev = omap_connector->dssdev;
-	struct omap_dss_driver *dssdrv = dssdev->driver;
-	struct omap_video_timings timings = {0};
-	struct drm_device *dev = connector->dev;
-	struct drm_display_mode *new_mode;
-	int ret = MODE_BAD;
-
-	copy_timings_drm_to_omap(&timings, mode);
-	mode->vrefresh = drm_mode_vrefresh(mode);
-
-	if (!dssdrv->check_timings(dssdev, &timings)) {
-		/* check if vrefresh is still valid */
-		new_mode = drm_mode_duplicate(dev, mode);
-		new_mode->clock = timings.pixel_clock;
-		new_mode->vrefresh = 0;
-		if (mode->vrefresh == drm_mode_vrefresh(new_mode))
-			ret = MODE_OK;
-		drm_mode_destroy(dev, new_mode);
-	}
-
-	DBG("connector: mode %s: "
-			"%d:\"%s\" %d %d %d %d %d %d %d %d %d %d 0x%x 0x%x",
-			(ret == MODE_OK) ? "valid" : "invalid",
-			mode->base.id, mode->name, mode->vrefresh, mode->clock,
-			mode->hdisplay, mode->hsync_start,
-			mode->hsync_end, mode->htotal,
-			mode->vdisplay, mode->vsync_start,
-			mode->vsync_end, mode->vtotal, mode->type, mode->flags);
-
-	return ret;
-}
-
-struct drm_encoder *omap_connector_attached_encoder(
-		struct drm_connector *connector)
-{
-	struct omap_connector *omap_connector = to_omap_connector(connector);
-	return omap_connector->encoder;
-}
-
-static const struct drm_connector_funcs omap_connector_funcs = {
-	.dpms = drm_helper_connector_dpms,
-	.detect = omap_connector_detect,
-	.fill_modes = drm_helper_probe_single_connector_modes,
-	.destroy = omap_connector_destroy,
-};
-
-static const struct drm_connector_helper_funcs omap_connector_helper_funcs = {
-	.get_modes = omap_connector_get_modes,
-	.mode_valid = omap_connector_mode_valid,
-	.best_encoder = omap_connector_attached_encoder,
-};
-
-/* flush an area of the framebuffer (in case of manual update display that
- * is not automatically flushed)
- */
-void omap_connector_flush(struct drm_connector *connector,
-		int x, int y, int w, int h)
-{
-	struct omap_connector *omap_connector = to_omap_connector(connector);
-
-	/* TODO: enable when supported in dss */
-	VERB("%s: %d,%d, %dx%d", omap_connector->dssdev->name, x, y, w, h);
-}
-
-/* initialize connector */
-struct drm_connector *omap_connector_init(struct drm_device *dev,
-		int connector_type, struct omap_dss_device *dssdev,
-		struct drm_encoder *encoder)
-{
-	struct drm_connector *connector = NULL;
-	struct omap_connector *omap_connector;
-
-	DBG("%s", dssdev->name);
-
-	omap_dss_get_device(dssdev);
-
-	omap_connector = kzalloc(sizeof(struct omap_connector), GFP_KERNEL);
-	if (!omap_connector) {
-		dev_err(dev->dev, "could not allocate connector\n");
-		goto fail;
-	}
-
-	omap_connector->dssdev = dssdev;
-	omap_connector->encoder = encoder;
-
-	connector = &omap_connector->base;
-
-	drm_connector_init(dev, connector, &omap_connector_funcs,
-				connector_type);
-	drm_connector_helper_add(connector, &omap_connector_helper_funcs);
-
-#if 0 /* enable when dss2 supports hotplug */
-	if (dssdev->caps & OMAP_DSS_DISPLAY_CAP_HPD)
-		connector->polled = 0;
-	else
-#endif
-		connector->polled = DRM_CONNECTOR_POLL_CONNECT |
-				DRM_CONNECTOR_POLL_DISCONNECT;
-
-	connector->interlace_allowed = 1;
-	connector->doublescan_allowed = 0;
-
-	drm_sysfs_connector_add(connector);
-
-	return connector;
-
-fail:
-	if (connector)
-		omap_connector_destroy(connector);
-
-	return NULL;
-}
diff --git a/drivers/staging/omapdrm/omap_crtc.c b/drivers/staging/omapdrm/omap_crtc.c
deleted file mode 100644
index 510942e..0000000
--- a/drivers/staging/omapdrm/omap_crtc.c
+++ /dev/null
@@ -1,659 +0,0 @@
-/*
- * drivers/staging/omapdrm/omap_crtc.c
- *
- * Copyright (C) 2011 Texas Instruments
- * Author: Rob Clark <rob@ti.com>
- *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of the GNU General Public License version 2 as published by
- * the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but WITHOUT
- * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
- * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
- * more details.
- *
- * You should have received a copy of the GNU General Public License along with
- * this program.  If not, see <http://www.gnu.org/licenses/>.
- */
-
-#include "omap_drv.h"
-
-#include <drm/drm_mode.h>
-#include "drm_crtc.h"
-#include "drm_crtc_helper.h"
-
-#define to_omap_crtc(x) container_of(x, struct omap_crtc, base)
-
-struct omap_crtc {
-	struct drm_crtc base;
-	struct drm_plane *plane;
-
-	const char *name;
-	int pipe;
-	enum omap_channel channel;
-	struct omap_overlay_manager_info info;
-
-	/*
-	 * Temporary: eventually this will go away, but it is needed
-	 * for now to keep the output's happy.  (They only need
-	 * mgr->id.)  Eventually this will be replaced w/ something
-	 * more common-panel-framework-y
-	 */
-	struct omap_overlay_manager mgr;
-
-	struct omap_video_timings timings;
-	bool enabled;
-	bool full_update;
-
-	struct omap_drm_apply apply;
-
-	struct omap_drm_irq apply_irq;
-	struct omap_drm_irq error_irq;
-
-	/* list of in-progress apply's: */
-	struct list_head pending_applies;
-
-	/* list of queued apply's: */
-	struct list_head queued_applies;
-
-	/* for handling queued and in-progress applies: */
-	struct work_struct apply_work;
-
-	/* if there is a pending flip, these will be non-null: */
-	struct drm_pending_vblank_event *event;
-	struct drm_framebuffer *old_fb;
-
-	/* for handling page flips without caring about what
-	 * the callback is called from.  Possibly we should just
-	 * make omap_gem always call the cb from the worker so
-	 * we don't have to care about this..
-	 *
-	 * XXX maybe fold into apply_work??
-	 */
-	struct work_struct page_flip_work;
-};
-
-/*
- * Manager-ops, callbacks from output when they need to configure
- * the upstream part of the video pipe.
- *
- * Most of these we can ignore until we add support for command-mode
- * panels.. for video-mode the crtc-helpers already do an adequate
- * job of sequencing the setup of the video pipe in the proper order
- */
-
-/* we can probably ignore these until we support command-mode panels: */
-static void omap_crtc_start_update(struct omap_overlay_manager *mgr)
-{
-}
-
-static int omap_crtc_enable(struct omap_overlay_manager *mgr)
-{
-	return 0;
-}
-
-static void omap_crtc_disable(struct omap_overlay_manager *mgr)
-{
-}
-
-static void omap_crtc_set_timings(struct omap_overlay_manager *mgr,
-		const struct omap_video_timings *timings)
-{
-	struct omap_crtc *omap_crtc = container_of(mgr, struct omap_crtc, mgr);
-	DBG("%s", omap_crtc->name);
-	omap_crtc->timings = *timings;
-	omap_crtc->full_update = true;
-}
-
-static void omap_crtc_set_lcd_config(struct omap_overlay_manager *mgr,
-		const struct dss_lcd_mgr_config *config)
-{
-	struct omap_crtc *omap_crtc = container_of(mgr, struct omap_crtc, mgr);
-	DBG("%s", omap_crtc->name);
-	dispc_mgr_set_lcd_config(omap_crtc->channel, config);
-}
-
-static int omap_crtc_register_framedone_handler(
-		struct omap_overlay_manager *mgr,
-		void (*handler)(void *), void *data)
-{
-	return 0;
-}
-
-static void omap_crtc_unregister_framedone_handler(
-		struct omap_overlay_manager *mgr,
-		void (*handler)(void *), void *data)
-{
-}
-
-static const struct dss_mgr_ops mgr_ops = {
-		.start_update = omap_crtc_start_update,
-		.enable = omap_crtc_enable,
-		.disable = omap_crtc_disable,
-		.set_timings = omap_crtc_set_timings,
-		.set_lcd_config = omap_crtc_set_lcd_config,
-		.register_framedone_handler = omap_crtc_register_framedone_handler,
-		.unregister_framedone_handler = omap_crtc_unregister_framedone_handler,
-};
-
-/*
- * CRTC funcs:
- */
-
-static void omap_crtc_destroy(struct drm_crtc *crtc)
-{
-	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
-
-	DBG("%s", omap_crtc->name);
-
-	WARN_ON(omap_crtc->apply_irq.registered);
-	omap_irq_unregister(crtc->dev, &omap_crtc->error_irq);
-
-	omap_crtc->plane->funcs->destroy(omap_crtc->plane);
-	drm_crtc_cleanup(crtc);
-
-	kfree(omap_crtc);
-}
-
-static void omap_crtc_dpms(struct drm_crtc *crtc, int mode)
-{
-	struct omap_drm_private *priv = crtc->dev->dev_private;
-	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
-	bool enabled = (mode == DRM_MODE_DPMS_ON);
-	int i;
-
-	DBG("%s: %d", omap_crtc->name, mode);
-
-	if (enabled != omap_crtc->enabled) {
-		omap_crtc->enabled = enabled;
-		omap_crtc->full_update = true;
-		omap_crtc_apply(crtc, &omap_crtc->apply);
-
-		/* also enable our private plane: */
-		WARN_ON(omap_plane_dpms(omap_crtc->plane, mode));
-
-		/* and any attached overlay planes: */
-		for (i = 0; i < priv->num_planes; i++) {
-			struct drm_plane *plane = priv->planes[i];
-			if (plane->crtc == crtc)
-				WARN_ON(omap_plane_dpms(plane, mode));
-		}
-	}
-}
-
-static bool omap_crtc_mode_fixup(struct drm_crtc *crtc,
-		const struct drm_display_mode *mode,
-		struct drm_display_mode *adjusted_mode)
-{
-	return true;
-}
-
-static int omap_crtc_mode_set(struct drm_crtc *crtc,
-		struct drm_display_mode *mode,
-		struct drm_display_mode *adjusted_mode,
-		int x, int y,
-		struct drm_framebuffer *old_fb)
-{
-	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
-
-	mode = adjusted_mode;
-
-	DBG("%s: set mode: %d:\"%s\" %d %d %d %d %d %d %d %d %d %d 0x%x 0x%x",
-			omap_crtc->name, mode->base.id, mode->name,
-			mode->vrefresh, mode->clock,
-			mode->hdisplay, mode->hsync_start,
-			mode->hsync_end, mode->htotal,
-			mode->vdisplay, mode->vsync_start,
-			mode->vsync_end, mode->vtotal,
-			mode->type, mode->flags);
-
-	copy_timings_drm_to_omap(&omap_crtc->timings, mode);
-	omap_crtc->full_update = true;
-
-	return omap_plane_mode_set(omap_crtc->plane, crtc, crtc->fb,
-			0, 0, mode->hdisplay, mode->vdisplay,
-			x << 16, y << 16,
-			mode->hdisplay << 16, mode->vdisplay << 16,
-			NULL, NULL);
-}
-
-static void omap_crtc_prepare(struct drm_crtc *crtc)
-{
-	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
-	DBG("%s", omap_crtc->name);
-	omap_crtc_dpms(crtc, DRM_MODE_DPMS_OFF);
-}
-
-static void omap_crtc_commit(struct drm_crtc *crtc)
-{
-	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
-	DBG("%s", omap_crtc->name);
-	omap_crtc_dpms(crtc, DRM_MODE_DPMS_ON);
-}
-
-static int omap_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
-		struct drm_framebuffer *old_fb)
-{
-	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
-	struct drm_plane *plane = omap_crtc->plane;
-	struct drm_display_mode *mode = &crtc->mode;
-
-	return omap_plane_mode_set(plane, crtc, crtc->fb,
-			0, 0, mode->hdisplay, mode->vdisplay,
-			x << 16, y << 16,
-			mode->hdisplay << 16, mode->vdisplay << 16,
-			NULL, NULL);
-}
-
-static void omap_crtc_load_lut(struct drm_crtc *crtc)
-{
-}
-
-static void vblank_cb(void *arg)
-{
-	struct drm_crtc *crtc = arg;
-	struct drm_device *dev = crtc->dev;
-	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
-	unsigned long flags;
-
-	spin_lock_irqsave(&dev->event_lock, flags);
-
-	/* wakeup userspace */
-	if (omap_crtc->event)
-		drm_send_vblank_event(dev, omap_crtc->pipe, omap_crtc->event);
-
-	omap_crtc->event = NULL;
-	omap_crtc->old_fb = NULL;
-
-	spin_unlock_irqrestore(&dev->event_lock, flags);
-}
-
-static void page_flip_worker(struct work_struct *work)
-{
-	struct omap_crtc *omap_crtc =
-			container_of(work, struct omap_crtc, page_flip_work);
-	struct drm_crtc *crtc = &omap_crtc->base;
-	struct drm_device *dev = crtc->dev;
-	struct drm_display_mode *mode = &crtc->mode;
-	struct drm_gem_object *bo;
-
-	drm_modeset_lock_all(dev);
-	omap_plane_mode_set(omap_crtc->plane, crtc, crtc->fb,
-			0, 0, mode->hdisplay, mode->vdisplay,
-			crtc->x << 16, crtc->y << 16,
-			mode->hdisplay << 16, mode->vdisplay << 16,
-			vblank_cb, crtc);
-	drm_modeset_unlock_all(dev);
-
-	bo = omap_framebuffer_bo(crtc->fb, 0);
-	drm_gem_object_unreference_unlocked(bo);
-}
-
-static void page_flip_cb(void *arg)
-{
-	struct drm_crtc *crtc = arg;
-	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
-	struct omap_drm_private *priv = crtc->dev->dev_private;
-
-	/* avoid assumptions about what ctxt we are called from: */
-	queue_work(priv->wq, &omap_crtc->page_flip_work);
-}
-
-static int omap_crtc_page_flip_locked(struct drm_crtc *crtc,
-		 struct drm_framebuffer *fb,
-		 struct drm_pending_vblank_event *event)
-{
-	struct drm_device *dev = crtc->dev;
-	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
-	struct drm_gem_object *bo;
-
-	DBG("%d -> %d (event=%p)", crtc->fb ? crtc->fb->base.id : -1,
-			fb->base.id, event);
-
-	if (omap_crtc->old_fb) {
-		dev_err(dev->dev, "already a pending flip\n");
-		return -EINVAL;
-	}
-
-	omap_crtc->event = event;
-	crtc->fb = fb;
-
-	/*
-	 * Hold a reference temporarily until the crtc is updated
-	 * and takes the reference to the bo.  This avoids it
-	 * getting freed from under us:
-	 */
-	bo = omap_framebuffer_bo(fb, 0);
-	drm_gem_object_reference(bo);
-
-	omap_gem_op_async(bo, OMAP_GEM_READ, page_flip_cb, crtc);
-
-	return 0;
-}
-
-static int omap_crtc_set_property(struct drm_crtc *crtc,
-		struct drm_property *property, uint64_t val)
-{
-	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
-	struct omap_drm_private *priv = crtc->dev->dev_private;
-
-	if (property == priv->rotation_prop) {
-		crtc->invert_dimensions =
-				!!(val & ((1LL << DRM_ROTATE_90) | (1LL << DRM_ROTATE_270)));
-	}
-
-	return omap_plane_set_property(omap_crtc->plane, property, val);
-}
-
-static const struct drm_crtc_funcs omap_crtc_funcs = {
-	.set_config = drm_crtc_helper_set_config,
-	.destroy = omap_crtc_destroy,
-	.page_flip = omap_crtc_page_flip_locked,
-	.set_property = omap_crtc_set_property,
-};
-
-static const struct drm_crtc_helper_funcs omap_crtc_helper_funcs = {
-	.dpms = omap_crtc_dpms,
-	.mode_fixup = omap_crtc_mode_fixup,
-	.mode_set = omap_crtc_mode_set,
-	.prepare = omap_crtc_prepare,
-	.commit = omap_crtc_commit,
-	.mode_set_base = omap_crtc_mode_set_base,
-	.load_lut = omap_crtc_load_lut,
-};
-
-const struct omap_video_timings *omap_crtc_timings(struct drm_crtc *crtc)
-{
-	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
-	return &omap_crtc->timings;
-}
-
-enum omap_channel omap_crtc_channel(struct drm_crtc *crtc)
-{
-	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
-	return omap_crtc->channel;
-}
-
-static void omap_crtc_error_irq(struct omap_drm_irq *irq, uint32_t irqstatus)
-{
-	struct omap_crtc *omap_crtc =
-			container_of(irq, struct omap_crtc, error_irq);
-	struct drm_crtc *crtc = &omap_crtc->base;
-	DRM_ERROR("%s: errors: %08x\n", omap_crtc->name, irqstatus);
-	/* avoid getting in a flood, unregister the irq until next vblank */
-	omap_irq_unregister(crtc->dev, &omap_crtc->error_irq);
-}
-
-static void omap_crtc_apply_irq(struct omap_drm_irq *irq, uint32_t irqstatus)
-{
-	struct omap_crtc *omap_crtc =
-			container_of(irq, struct omap_crtc, apply_irq);
-	struct drm_crtc *crtc = &omap_crtc->base;
-
-	if (!omap_crtc->error_irq.registered)
-		omap_irq_register(crtc->dev, &omap_crtc->error_irq);
-
-	if (!dispc_mgr_go_busy(omap_crtc->channel)) {
-		struct omap_drm_private *priv =
-				crtc->dev->dev_private;
-		DBG("%s: apply done", omap_crtc->name);
-		omap_irq_unregister(crtc->dev, &omap_crtc->apply_irq);
-		queue_work(priv->wq, &omap_crtc->apply_work);
-	}
-}
-
-static void apply_worker(struct work_struct *work)
-{
-	struct omap_crtc *omap_crtc =
-			container_of(work, struct omap_crtc, apply_work);
-	struct drm_crtc *crtc = &omap_crtc->base;
-	struct drm_device *dev = crtc->dev;
-	struct omap_drm_apply *apply, *n;
-	bool need_apply;
-
-	/*
-	 * Synchronize everything on mode_config.mutex, to keep
-	 * the callbacks and list modification all serialized
-	 * with respect to modesetting ioctls from userspace.
-	 */
-	drm_modeset_lock_all(dev);
-	dispc_runtime_get();
-
-	/*
-	 * If we are still pending a previous update, wait.. when the
-	 * pending update completes, we get kicked again.
-	 */
-	if (omap_crtc->apply_irq.registered)
-		goto out;
-
-	/* finish up previous apply's: */
-	list_for_each_entry_safe(apply, n,
-			&omap_crtc->pending_applies, pending_node) {
-		apply->post_apply(apply);
-		list_del(&apply->pending_node);
-	}
-
-	need_apply = !list_empty(&omap_crtc->queued_applies);
-
-	/* then handle the next round of of queued apply's: */
-	list_for_each_entry_safe(apply, n,
-			&omap_crtc->queued_applies, queued_node) {
-		apply->pre_apply(apply);
-		list_del(&apply->queued_node);
-		apply->queued = false;
-		list_add_tail(&apply->pending_node,
-				&omap_crtc->pending_applies);
-	}
-
-	if (need_apply) {
-		enum omap_channel channel = omap_crtc->channel;
-
-		DBG("%s: GO", omap_crtc->name);
-
-		if (dispc_mgr_is_enabled(channel)) {
-			omap_irq_register(dev, &omap_crtc->apply_irq);
-			dispc_mgr_go(channel);
-		} else {
-			struct omap_drm_private *priv = dev->dev_private;
-			queue_work(priv->wq, &omap_crtc->apply_work);
-		}
-	}
-
-out:
-	dispc_runtime_put();
-	drm_modeset_unlock_all(dev);
-}
-
-int omap_crtc_apply(struct drm_crtc *crtc,
-		struct omap_drm_apply *apply)
-{
-	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
-	struct drm_device *dev = crtc->dev;
-
-	WARN_ON(!mutex_is_locked(&dev->mode_config.mutex));
-
-	/* no need to queue it again if it is already queued: */
-	if (apply->queued)
-		return 0;
-
-	apply->queued = true;
-	list_add_tail(&apply->queued_node, &omap_crtc->queued_applies);
-
-	/*
-	 * If there are no currently pending updates, then go ahead and
-	 * kick the worker immediately, otherwise it will run again when
-	 * the current update finishes.
-	 */
-	if (list_empty(&omap_crtc->pending_applies)) {
-		struct omap_drm_private *priv = crtc->dev->dev_private;
-		queue_work(priv->wq, &omap_crtc->apply_work);
-	}
-
-	return 0;
-}
-
-/* called only from apply */
-static void set_enabled(struct drm_crtc *crtc, bool enable)
-{
-	struct drm_device *dev = crtc->dev;
-	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
-	enum omap_channel channel = omap_crtc->channel;
-	struct omap_irq_wait *wait = NULL;
-
-	if (dispc_mgr_is_enabled(channel) == enable)
-		return;
-
-	/* ignore sync-lost irqs during enable/disable */
-	omap_irq_unregister(crtc->dev, &omap_crtc->error_irq);
-
-	if (dispc_mgr_get_framedone_irq(channel)) {
-		if (!enable) {
-			wait = omap_irq_wait_init(dev,
-					dispc_mgr_get_framedone_irq(channel), 1);
-		}
-	} else {
-		/*
-		 * When we disable digit output, we need to wait until fields
-		 * are done.  Otherwise the DSS is still working, and turning
-		 * off the clocks prevents DSS from going to OFF mode. And when
-		 * enabling, we need to wait for the extra sync losts
-		 */
-		wait = omap_irq_wait_init(dev,
-				dispc_mgr_get_vsync_irq(channel), 2);
-	}
-
-	dispc_mgr_enable(channel, enable);
-
-	if (wait) {
-		int ret = omap_irq_wait(dev, wait, msecs_to_jiffies(100));
-		if (ret) {
-			dev_err(dev->dev, "%s: timeout waiting for %s\n",
-					omap_crtc->name, enable ? "enable" : "disable");
-		}
-	}
-
-	omap_irq_register(crtc->dev, &omap_crtc->error_irq);
-}
-
-static void omap_crtc_pre_apply(struct omap_drm_apply *apply)
-{
-	struct omap_crtc *omap_crtc =
-			container_of(apply, struct omap_crtc, apply);
-	struct drm_crtc *crtc = &omap_crtc->base;
-	struct drm_encoder *encoder = NULL;
-
-	DBG("%s: enabled=%d, full=%d", omap_crtc->name,
-			omap_crtc->enabled, omap_crtc->full_update);
-
-	if (omap_crtc->full_update) {
-		struct omap_drm_private *priv = crtc->dev->dev_private;
-		int i;
-		for (i = 0; i < priv->num_encoders; i++) {
-			if (priv->encoders[i]->crtc == crtc) {
-				encoder = priv->encoders[i];
-				break;
-			}
-		}
-	}
-
-	if (!omap_crtc->enabled) {
-		set_enabled(&omap_crtc->base, false);
-		if (encoder)
-			omap_encoder_set_enabled(encoder, false);
-	} else {
-		if (encoder) {
-			omap_encoder_set_enabled(encoder, false);
-			omap_encoder_update(encoder, &omap_crtc->mgr,
-					&omap_crtc->timings);
-			omap_encoder_set_enabled(encoder, true);
-			omap_crtc->full_update = false;
-		}
-
-		dispc_mgr_setup(omap_crtc->channel, &omap_crtc->info);
-		dispc_mgr_set_timings(omap_crtc->channel,
-				&omap_crtc->timings);
-		set_enabled(&omap_crtc->base, true);
-	}
-
-	omap_crtc->full_update = false;
-}
-
-static void omap_crtc_post_apply(struct omap_drm_apply *apply)
-{
-	/* nothing needed for post-apply */
-}
-
-static const char *channel_names[] = {
-		[OMAP_DSS_CHANNEL_LCD] = "lcd",
-		[OMAP_DSS_CHANNEL_DIGIT] = "tv",
-		[OMAP_DSS_CHANNEL_LCD2] = "lcd2",
-};
-
-/* initialize crtc */
-struct drm_crtc *omap_crtc_init(struct drm_device *dev,
-		struct drm_plane *plane, enum omap_channel channel, int id)
-{
-	struct drm_crtc *crtc = NULL;
-	struct omap_crtc *omap_crtc;
-	struct omap_overlay_manager_info *info;
-
-	DBG("%s", channel_names[channel]);
-
-	omap_crtc = kzalloc(sizeof(*omap_crtc), GFP_KERNEL);
-
-	if (!omap_crtc) {
-		dev_err(dev->dev, "could not allocate CRTC\n");
-		goto fail;
-	}
-
-	crtc = &omap_crtc->base;
-
-	INIT_WORK(&omap_crtc->page_flip_work, page_flip_worker);
-	INIT_WORK(&omap_crtc->apply_work, apply_worker);
-
-	INIT_LIST_HEAD(&omap_crtc->pending_applies);
-	INIT_LIST_HEAD(&omap_crtc->queued_applies);
-
-	omap_crtc->apply.pre_apply  = omap_crtc_pre_apply;
-	omap_crtc->apply.post_apply = omap_crtc_post_apply;
-
-	omap_crtc->apply_irq.irqmask = pipe2vbl(id);
-	omap_crtc->apply_irq.irq = omap_crtc_apply_irq;
-
-	omap_crtc->error_irq.irqmask =
-			dispc_mgr_get_sync_lost_irq(channel);
-	omap_crtc->error_irq.irq = omap_crtc_error_irq;
-	omap_irq_register(dev, &omap_crtc->error_irq);
-
-	omap_crtc->channel = channel;
-	omap_crtc->plane = plane;
-	omap_crtc->plane->crtc = crtc;
-	omap_crtc->name = channel_names[channel];
-	omap_crtc->pipe = id;
-
-	/* temporary: */
-	omap_crtc->mgr.id = channel;
-
-	dss_install_mgr_ops(&mgr_ops);
-
-	/* TODO: fix hard-coded setup.. add properties! */
-	info = &omap_crtc->info;
-	info->default_color = 0x00000000;
-	info->trans_key = 0x00000000;
-	info->trans_key_type = OMAP_DSS_COLOR_KEY_GFX_DST;
-	info->trans_enabled = false;
-
-	drm_crtc_init(dev, crtc, &omap_crtc_funcs);
-	drm_crtc_helper_add(crtc, &omap_crtc_helper_funcs);
-
-	omap_plane_install_properties(omap_crtc->plane, &crtc->base);
-
-	return crtc;
-
-fail:
-	if (crtc)
-		omap_crtc_destroy(crtc);
-
-	return NULL;
-}
diff --git a/drivers/staging/omapdrm/omap_debugfs.c b/drivers/staging/omapdrm/omap_debugfs.c
deleted file mode 100644
index e95540b..0000000
--- a/drivers/staging/omapdrm/omap_debugfs.c
+++ /dev/null
@@ -1,139 +0,0 @@
-/*
- * drivers/staging/omapdrm/omap_debugfs.c
- *
- * Copyright (C) 2011 Texas Instruments
- * Author: Rob Clark <rob.clark@linaro.org>
- *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of the GNU General Public License version 2 as published by
- * the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but WITHOUT
- * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
- * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
- * more details.
- *
- * You should have received a copy of the GNU General Public License along with
- * this program.  If not, see <http://www.gnu.org/licenses/>.
- */
-
-#include "omap_drv.h"
-#include "omap_dmm_tiler.h"
-
-#include "drm_fb_helper.h"
-
-
-#ifdef CONFIG_DEBUG_FS
-
-static int gem_show(struct seq_file *m, void *arg)
-{
-	struct drm_info_node *node = (struct drm_info_node *) m->private;
-	struct drm_device *dev = node->minor->dev;
-	struct omap_drm_private *priv = dev->dev_private;
-	int ret;
-
-	ret = mutex_lock_interruptible(&dev->struct_mutex);
-	if (ret)
-		return ret;
-
-	seq_printf(m, "All Objects:\n");
-	omap_gem_describe_objects(&priv->obj_list, m);
-
-	mutex_unlock(&dev->struct_mutex);
-
-	return 0;
-}
-
-static int mm_show(struct seq_file *m, void *arg)
-{
-	struct drm_info_node *node = (struct drm_info_node *) m->private;
-	struct drm_device *dev = node->minor->dev;
-	return drm_mm_dump_table(m, dev->mm_private);
-}
-
-static int fb_show(struct seq_file *m, void *arg)
-{
-	struct drm_info_node *node = (struct drm_info_node *) m->private;
-	struct drm_device *dev = node->minor->dev;
-	struct omap_drm_private *priv = dev->dev_private;
-	struct drm_framebuffer *fb;
-	int ret;
-
-	ret = mutex_lock_interruptible(&dev->mode_config.mutex);
-	if (ret)
-		return ret;
-
-	ret = mutex_lock_interruptible(&dev->struct_mutex);
-	if (ret) {
-		mutex_unlock(&dev->mode_config.mutex);
-		return ret;
-	}
-
-	seq_printf(m, "fbcon ");
-	omap_framebuffer_describe(priv->fbdev->fb, m);
-
-	mutex_lock(&dev->mode_config.fb_lock);
-	list_for_each_entry(fb, &dev->mode_config.fb_list, head) {
-		if (fb == priv->fbdev->fb)
-			continue;
-
-		seq_printf(m, "user ");
-		omap_framebuffer_describe(fb, m);
-	}
-	mutex_unlock(&dev->mode_config.fb_lock);
-
-	mutex_unlock(&dev->struct_mutex);
-	mutex_unlock(&dev->mode_config.mutex);
-
-	return 0;
-}
-
-/* list of debufs files that are applicable to all devices */
-static struct drm_info_list omap_debugfs_list[] = {
-	{"gem", gem_show, 0},
-	{"mm", mm_show, 0},
-	{"fb", fb_show, 0},
-};
-
-/* list of debugfs files that are specific to devices with dmm/tiler */
-static struct drm_info_list omap_dmm_debugfs_list[] = {
-	{"tiler_map", tiler_map_show, 0},
-};
-
-int omap_debugfs_init(struct drm_minor *minor)
-{
-	struct drm_device *dev = minor->dev;
-	int ret;
-
-	ret = drm_debugfs_create_files(omap_debugfs_list,
-			ARRAY_SIZE(omap_debugfs_list),
-			minor->debugfs_root, minor);
-
-	if (ret) {
-		dev_err(dev->dev, "could not install omap_debugfs_list\n");
-		return ret;
-	}
-
-	if (dmm_is_available())
-		ret = drm_debugfs_create_files(omap_dmm_debugfs_list,
-				ARRAY_SIZE(omap_dmm_debugfs_list),
-				minor->debugfs_root, minor);
-
-	if (ret) {
-		dev_err(dev->dev, "could not install omap_dmm_debugfs_list\n");
-		return ret;
-	}
-
-	return ret;
-}
-
-void omap_debugfs_cleanup(struct drm_minor *minor)
-{
-	drm_debugfs_remove_files(omap_debugfs_list,
-			ARRAY_SIZE(omap_debugfs_list), minor);
-	if (dmm_is_available())
-		drm_debugfs_remove_files(omap_dmm_debugfs_list,
-				ARRAY_SIZE(omap_dmm_debugfs_list), minor);
-}
-
-#endif
diff --git a/drivers/staging/omapdrm/omap_dmm_priv.h b/drivers/staging/omapdrm/omap_dmm_priv.h
deleted file mode 100644
index 58bcd6a..0000000
--- a/drivers/staging/omapdrm/omap_dmm_priv.h
+++ /dev/null
@@ -1,188 +0,0 @@
-/*
- *
- * Copyright (C) 2011 Texas Instruments Incorporated - http://www.ti.com/
- * Author: Rob Clark <rob@ti.com>
- *         Andy Gross <andy.gross@ti.com>
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation version 2.
- *
- * This program is distributed "as is" WITHOUT ANY WARRANTY of any
- * kind, whether express or implied; without even the implied warranty
- * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- */
-#ifndef OMAP_DMM_PRIV_H
-#define OMAP_DMM_PRIV_H
-
-#define DMM_REVISION          0x000
-#define DMM_HWINFO            0x004
-#define DMM_LISA_HWINFO       0x008
-#define DMM_DMM_SYSCONFIG     0x010
-#define DMM_LISA_LOCK         0x01C
-#define DMM_LISA_MAP__0       0x040
-#define DMM_LISA_MAP__1       0x044
-#define DMM_TILER_HWINFO      0x208
-#define DMM_TILER_OR__0       0x220
-#define DMM_TILER_OR__1       0x224
-#define DMM_PAT_HWINFO        0x408
-#define DMM_PAT_GEOMETRY      0x40C
-#define DMM_PAT_CONFIG        0x410
-#define DMM_PAT_VIEW__0       0x420
-#define DMM_PAT_VIEW__1       0x424
-#define DMM_PAT_VIEW_MAP__0   0x440
-#define DMM_PAT_VIEW_MAP_BASE 0x460
-#define DMM_PAT_IRQ_EOI       0x478
-#define DMM_PAT_IRQSTATUS_RAW 0x480
-#define DMM_PAT_IRQSTATUS     0x490
-#define DMM_PAT_IRQENABLE_SET 0x4A0
-#define DMM_PAT_IRQENABLE_CLR 0x4B0
-#define DMM_PAT_STATUS__0     0x4C0
-#define DMM_PAT_STATUS__1     0x4C4
-#define DMM_PAT_STATUS__2     0x4C8
-#define DMM_PAT_STATUS__3     0x4CC
-#define DMM_PAT_DESCR__0      0x500
-#define DMM_PAT_DESCR__1      0x510
-#define DMM_PAT_DESCR__2      0x520
-#define DMM_PAT_DESCR__3      0x530
-#define DMM_PEG_HWINFO        0x608
-#define DMM_PEG_PRIO          0x620
-#define DMM_PEG_PRIO_PAT      0x640
-
-#define DMM_IRQSTAT_DST			(1<<0)
-#define DMM_IRQSTAT_LST			(1<<1)
-#define DMM_IRQSTAT_ERR_INV_DSC		(1<<2)
-#define DMM_IRQSTAT_ERR_INV_DATA	(1<<3)
-#define DMM_IRQSTAT_ERR_UPD_AREA	(1<<4)
-#define DMM_IRQSTAT_ERR_UPD_CTRL	(1<<5)
-#define DMM_IRQSTAT_ERR_UPD_DATA	(1<<6)
-#define DMM_IRQSTAT_ERR_LUT_MISS	(1<<7)
-
-#define DMM_IRQSTAT_ERR_MASK	(DMM_IRQ_STAT_ERR_INV_DSC | \
-				DMM_IRQ_STAT_ERR_INV_DATA | \
-				DMM_IRQ_STAT_ERR_UPD_AREA | \
-				DMM_IRQ_STAT_ERR_UPD_CTRL | \
-				DMM_IRQ_STAT_ERR_UPD_DATA | \
-				DMM_IRQ_STAT_ERR_LUT_MISS)
-
-#define DMM_PATSTATUS_READY		(1<<0)
-#define DMM_PATSTATUS_VALID		(1<<1)
-#define DMM_PATSTATUS_RUN		(1<<2)
-#define DMM_PATSTATUS_DONE		(1<<3)
-#define DMM_PATSTATUS_LINKED		(1<<4)
-#define DMM_PATSTATUS_BYPASSED		(1<<7)
-#define DMM_PATSTATUS_ERR_INV_DESCR	(1<<10)
-#define DMM_PATSTATUS_ERR_INV_DATA	(1<<11)
-#define DMM_PATSTATUS_ERR_UPD_AREA	(1<<12)
-#define DMM_PATSTATUS_ERR_UPD_CTRL	(1<<13)
-#define DMM_PATSTATUS_ERR_UPD_DATA	(1<<14)
-#define DMM_PATSTATUS_ERR_ACCESS	(1<<15)
-
-/* note: don't treat DMM_PATSTATUS_ERR_ACCESS as an error */
-#define DMM_PATSTATUS_ERR	(DMM_PATSTATUS_ERR_INV_DESCR | \
-				DMM_PATSTATUS_ERR_INV_DATA | \
-				DMM_PATSTATUS_ERR_UPD_AREA | \
-				DMM_PATSTATUS_ERR_UPD_CTRL | \
-				DMM_PATSTATUS_ERR_UPD_DATA)
-
-
-
-enum {
-	PAT_STATUS,
-	PAT_DESCR
-};
-
-struct pat_ctrl {
-	u32 start:4;
-	u32 dir:4;
-	u32 lut_id:8;
-	u32 sync:12;
-	u32 ini:4;
-};
-
-struct pat {
-	uint32_t next_pa;
-	struct pat_area area;
-	struct pat_ctrl ctrl;
-	uint32_t data_pa;
-};
-
-#define DMM_FIXED_RETRY_COUNT 1000
-
-/* create refill buffer big enough to refill all slots, plus 3 descriptors..
- * 3 descriptors is probably the worst-case for # of 2d-slices in a 1d area,
- * but I guess you don't hit that worst case at the same time as full area
- * refill
- */
-#define DESCR_SIZE 128
-#define REFILL_BUFFER_SIZE ((4 * 128 * 256) + (3 * DESCR_SIZE))
-
-/* For OMAP5, a fixed offset is added to all Y coordinates for 1D buffers.
- * This is used in programming to address the upper portion of the LUT
-*/
-#define OMAP5_LUT_OFFSET       128
-
-struct dmm;
-
-struct dmm_txn {
-	void *engine_handle;
-	struct tcm *tcm;
-
-	uint8_t *current_va;
-	dma_addr_t current_pa;
-
-	struct pat *last_pat;
-};
-
-struct refill_engine {
-	int id;
-	struct dmm *dmm;
-	struct tcm *tcm;
-
-	uint8_t *refill_va;
-	dma_addr_t refill_pa;
-
-	/* only one trans per engine for now */
-	struct dmm_txn txn;
-
-	bool async;
-
-	wait_queue_head_t wait_for_refill;
-
-	struct list_head idle_node;
-};
-
-struct dmm {
-	struct device *dev;
-	void __iomem *base;
-	int irq;
-
-	struct page *dummy_page;
-	dma_addr_t dummy_pa;
-
-	void *refill_va;
-	dma_addr_t refill_pa;
-
-	/* refill engines */
-	wait_queue_head_t engine_queue;
-	struct list_head idle_head;
-	struct refill_engine *engines;
-	int num_engines;
-	atomic_t engine_counter;
-
-	/* container information */
-	int container_width;
-	int container_height;
-	int lut_width;
-	int lut_height;
-	int num_lut;
-
-	/* array of LUT - TCM containers */
-	struct tcm **tcm;
-
-	/* allocation list and lock */
-	struct list_head alloc_head;
-};
-
-#endif
diff --git a/drivers/staging/omapdrm/omap_dmm_tiler.c b/drivers/staging/omapdrm/omap_dmm_tiler.c
deleted file mode 100644
index 3910215..0000000
--- a/drivers/staging/omapdrm/omap_dmm_tiler.c
+++ /dev/null
@@ -1,991 +0,0 @@
-/*
- * DMM IOMMU driver support functions for TI OMAP processors.
- *
- * Author: Rob Clark <rob@ti.com>
- *         Andy Gross <andy.gross@ti.com>
- *
- * Copyright (C) 2011 Texas Instruments Incorporated - http://www.ti.com/
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation version 2.
- *
- * This program is distributed "as is" WITHOUT ANY WARRANTY of any
- * kind, whether express or implied; without even the implied warranty
- * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- */
-#include <linux/init.h>
-#include <linux/module.h>
-#include <linux/platform_device.h> /* platform_device() */
-#include <linux/errno.h>
-#include <linux/sched.h>
-#include <linux/wait.h>
-#include <linux/interrupt.h>
-#include <linux/dma-mapping.h>
-#include <linux/slab.h>
-#include <linux/vmalloc.h>
-#include <linux/delay.h>
-#include <linux/mm.h>
-#include <linux/time.h>
-#include <linux/list.h>
-
-#include "omap_dmm_tiler.h"
-#include "omap_dmm_priv.h"
-
-#define DMM_DRIVER_NAME "dmm"
-
-/* mappings for associating views to luts */
-static struct tcm *containers[TILFMT_NFORMATS];
-static struct dmm *omap_dmm;
-
-/* global spinlock for protecting lists */
-static DEFINE_SPINLOCK(list_lock);
-
-/* Geometry table */
-#define GEOM(xshift, yshift, bytes_per_pixel) { \
-		.x_shft = (xshift), \
-		.y_shft = (yshift), \
-		.cpp    = (bytes_per_pixel), \
-		.slot_w = 1 << (SLOT_WIDTH_BITS - (xshift)), \
-		.slot_h = 1 << (SLOT_HEIGHT_BITS - (yshift)), \
-	}
-
-static const struct {
-	uint32_t x_shft;	/* unused X-bits (as part of bpp) */
-	uint32_t y_shft;	/* unused Y-bits (as part of bpp) */
-	uint32_t cpp;		/* bytes/chars per pixel */
-	uint32_t slot_w;	/* width of each slot (in pixels) */
-	uint32_t slot_h;	/* height of each slot (in pixels) */
-} geom[TILFMT_NFORMATS] = {
-		[TILFMT_8BIT]  = GEOM(0, 0, 1),
-		[TILFMT_16BIT] = GEOM(0, 1, 2),
-		[TILFMT_32BIT] = GEOM(1, 1, 4),
-		[TILFMT_PAGE]  = GEOM(SLOT_WIDTH_BITS, SLOT_HEIGHT_BITS, 1),
-};
-
-
-/* lookup table for registers w/ per-engine instances */
-static const uint32_t reg[][4] = {
-		[PAT_STATUS] = {DMM_PAT_STATUS__0, DMM_PAT_STATUS__1,
-				DMM_PAT_STATUS__2, DMM_PAT_STATUS__3},
-		[PAT_DESCR]  = {DMM_PAT_DESCR__0, DMM_PAT_DESCR__1,
-				DMM_PAT_DESCR__2, DMM_PAT_DESCR__3},
-};
-
-/* simple allocator to grab next 16 byte aligned memory from txn */
-static void *alloc_dma(struct dmm_txn *txn, size_t sz, dma_addr_t *pa)
-{
-	void *ptr;
-	struct refill_engine *engine = txn->engine_handle;
-
-	/* dmm programming requires 16 byte aligned addresses */
-	txn->current_pa = round_up(txn->current_pa, 16);
-	txn->current_va = (void *)round_up((long)txn->current_va, 16);
-
-	ptr = txn->current_va;
-	*pa = txn->current_pa;
-
-	txn->current_pa += sz;
-	txn->current_va += sz;
-
-	BUG_ON((txn->current_va - engine->refill_va) > REFILL_BUFFER_SIZE);
-
-	return ptr;
-}
-
-/* check status and spin until wait_mask comes true */
-static int wait_status(struct refill_engine *engine, uint32_t wait_mask)
-{
-	struct dmm *dmm = engine->dmm;
-	uint32_t r = 0, err, i;
-
-	i = DMM_FIXED_RETRY_COUNT;
-	while (true) {
-		r = readl(dmm->base + reg[PAT_STATUS][engine->id]);
-		err = r & DMM_PATSTATUS_ERR;
-		if (err)
-			return -EFAULT;
-
-		if ((r & wait_mask) == wait_mask)
-			break;
-
-		if (--i == 0)
-			return -ETIMEDOUT;
-
-		udelay(1);
-	}
-
-	return 0;
-}
-
-static void release_engine(struct refill_engine *engine)
-{
-	unsigned long flags;
-
-	spin_lock_irqsave(&list_lock, flags);
-	list_add(&engine->idle_node, &omap_dmm->idle_head);
-	spin_unlock_irqrestore(&list_lock, flags);
-
-	atomic_inc(&omap_dmm->engine_counter);
-	wake_up_interruptible(&omap_dmm->engine_queue);
-}
-
-static irqreturn_t omap_dmm_irq_handler(int irq, void *arg)
-{
-	struct dmm *dmm = arg;
-	uint32_t status = readl(dmm->base + DMM_PAT_IRQSTATUS);
-	int i;
-
-	/* ack IRQ */
-	writel(status, dmm->base + DMM_PAT_IRQSTATUS);
-
-	for (i = 0; i < dmm->num_engines; i++) {
-		if (status & DMM_IRQSTAT_LST) {
-			wake_up_interruptible(&dmm->engines[i].wait_for_refill);
-
-			if (dmm->engines[i].async)
-				release_engine(&dmm->engines[i]);
-		}
-
-		status >>= 8;
-	}
-
-	return IRQ_HANDLED;
-}
-
-/**
- * Get a handle for a DMM transaction
- */
-static struct dmm_txn *dmm_txn_init(struct dmm *dmm, struct tcm *tcm)
-{
-	struct dmm_txn *txn = NULL;
-	struct refill_engine *engine = NULL;
-	int ret;
-	unsigned long flags;
-
-
-	/* wait until an engine is available */
-	ret = wait_event_interruptible(omap_dmm->engine_queue,
-		atomic_add_unless(&omap_dmm->engine_counter, -1, 0));
-	if (ret)
-		return ERR_PTR(ret);
-
-	/* grab an idle engine */
-	spin_lock_irqsave(&list_lock, flags);
-	if (!list_empty(&dmm->idle_head)) {
-		engine = list_entry(dmm->idle_head.next, struct refill_engine,
-					idle_node);
-		list_del(&engine->idle_node);
-	}
-	spin_unlock_irqrestore(&list_lock, flags);
-
-	BUG_ON(!engine);
-
-	txn = &engine->txn;
-	engine->tcm = tcm;
-	txn->engine_handle = engine;
-	txn->last_pat = NULL;
-	txn->current_va = engine->refill_va;
-	txn->current_pa = engine->refill_pa;
-
-	return txn;
-}
-
-/**
- * Add region to DMM transaction.  If pages or pages[i] is NULL, then the
- * corresponding slot is cleared (ie. dummy_pa is programmed)
- */
-static void dmm_txn_append(struct dmm_txn *txn, struct pat_area *area,
-		struct page **pages, uint32_t npages, uint32_t roll)
-{
-	dma_addr_t pat_pa = 0;
-	uint32_t *data;
-	struct pat *pat;
-	struct refill_engine *engine = txn->engine_handle;
-	int columns = (1 + area->x1 - area->x0);
-	int rows = (1 + area->y1 - area->y0);
-	int i = columns*rows;
-
-	pat = alloc_dma(txn, sizeof(struct pat), &pat_pa);
-
-	if (txn->last_pat)
-		txn->last_pat->next_pa = (uint32_t)pat_pa;
-
-	pat->area = *area;
-
-	/* adjust Y coordinates based off of container parameters */
-	pat->area.y0 += engine->tcm->y_offset;
-	pat->area.y1 += engine->tcm->y_offset;
-
-	pat->ctrl = (struct pat_ctrl){
-			.start = 1,
-			.lut_id = engine->tcm->lut_id,
-		};
-
-	data = alloc_dma(txn, 4*i, &pat->data_pa);
-
-	while (i--) {
-		int n = i + roll;
-		if (n >= npages)
-			n -= npages;
-		data[i] = (pages && pages[n]) ?
-			page_to_phys(pages[n]) : engine->dmm->dummy_pa;
-	}
-
-	txn->last_pat = pat;
-
-	return;
-}
-
-/**
- * Commit the DMM transaction.
- */
-static int dmm_txn_commit(struct dmm_txn *txn, bool wait)
-{
-	int ret = 0;
-	struct refill_engine *engine = txn->engine_handle;
-	struct dmm *dmm = engine->dmm;
-
-	if (!txn->last_pat) {
-		dev_err(engine->dmm->dev, "need at least one txn\n");
-		ret = -EINVAL;
-		goto cleanup;
-	}
-
-	txn->last_pat->next_pa = 0;
-
-	/* write to PAT_DESCR to clear out any pending transaction */
-	writel(0x0, dmm->base + reg[PAT_DESCR][engine->id]);
-
-	/* wait for engine ready: */
-	ret = wait_status(engine, DMM_PATSTATUS_READY);
-	if (ret) {
-		ret = -EFAULT;
-		goto cleanup;
-	}
-
-	/* mark whether it is async to denote list management in IRQ handler */
-	engine->async = wait ? false : true;
-
-	/* kick reload */
-	writel(engine->refill_pa,
-		dmm->base + reg[PAT_DESCR][engine->id]);
-
-	if (wait) {
-		if (wait_event_interruptible_timeout(engine->wait_for_refill,
-				wait_status(engine, DMM_PATSTATUS_READY) == 0,
-				msecs_to_jiffies(1)) <= 0) {
-			dev_err(dmm->dev, "timed out waiting for done\n");
-			ret = -ETIMEDOUT;
-		}
-	}
-
-cleanup:
-	/* only place engine back on list if we are done with it */
-	if (ret || wait)
-		release_engine(engine);
-
-	return ret;
-}
-
-/*
- * DMM programming
- */
-static int fill(struct tcm_area *area, struct page **pages,
-		uint32_t npages, uint32_t roll, bool wait)
-{
-	int ret = 0;
-	struct tcm_area slice, area_s;
-	struct dmm_txn *txn;
-
-	txn = dmm_txn_init(omap_dmm, area->tcm);
-	if (IS_ERR_OR_NULL(txn))
-		return -ENOMEM;
-
-	tcm_for_each_slice(slice, *area, area_s) {
-		struct pat_area p_area = {
-				.x0 = slice.p0.x,  .y0 = slice.p0.y,
-				.x1 = slice.p1.x,  .y1 = slice.p1.y,
-		};
-
-		dmm_txn_append(txn, &p_area, pages, npages, roll);
-
-		roll += tcm_sizeof(slice);
-	}
-
-	ret = dmm_txn_commit(txn, wait);
-
-	return ret;
-}
-
-/*
- * Pin/unpin
- */
-
-/* note: slots for which pages[i] == NULL are filled w/ dummy page
- */
-int tiler_pin(struct tiler_block *block, struct page **pages,
-		uint32_t npages, uint32_t roll, bool wait)
-{
-	int ret;
-
-	ret = fill(&block->area, pages, npages, roll, wait);
-
-	if (ret)
-		tiler_unpin(block);
-
-	return ret;
-}
-
-int tiler_unpin(struct tiler_block *block)
-{
-	return fill(&block->area, NULL, 0, 0, false);
-}
-
-/*
- * Reserve/release
- */
-struct tiler_block *tiler_reserve_2d(enum tiler_fmt fmt, uint16_t w,
-		uint16_t h, uint16_t align)
-{
-	struct tiler_block *block = kzalloc(sizeof(*block), GFP_KERNEL);
-	u32 min_align = 128;
-	int ret;
-	unsigned long flags;
-
-	BUG_ON(!validfmt(fmt));
-
-	/* convert width/height to slots */
-	w = DIV_ROUND_UP(w, geom[fmt].slot_w);
-	h = DIV_ROUND_UP(h, geom[fmt].slot_h);
-
-	/* convert alignment to slots */
-	min_align = max(min_align, (geom[fmt].slot_w * geom[fmt].cpp));
-	align = ALIGN(align, min_align);
-	align /= geom[fmt].slot_w * geom[fmt].cpp;
-
-	block->fmt = fmt;
-
-	ret = tcm_reserve_2d(containers[fmt], w, h, align, &block->area);
-	if (ret) {
-		kfree(block);
-		return ERR_PTR(-ENOMEM);
-	}
-
-	/* add to allocation list */
-	spin_lock_irqsave(&list_lock, flags);
-	list_add(&block->alloc_node, &omap_dmm->alloc_head);
-	spin_unlock_irqrestore(&list_lock, flags);
-
-	return block;
-}
-
-struct tiler_block *tiler_reserve_1d(size_t size)
-{
-	struct tiler_block *block = kzalloc(sizeof(*block), GFP_KERNEL);
-	int num_pages = (size + PAGE_SIZE - 1) >> PAGE_SHIFT;
-	unsigned long flags;
-
-	if (!block)
-		return ERR_PTR(-ENOMEM);
-
-	block->fmt = TILFMT_PAGE;
-
-	if (tcm_reserve_1d(containers[TILFMT_PAGE], num_pages,
-				&block->area)) {
-		kfree(block);
-		return ERR_PTR(-ENOMEM);
-	}
-
-	spin_lock_irqsave(&list_lock, flags);
-	list_add(&block->alloc_node, &omap_dmm->alloc_head);
-	spin_unlock_irqrestore(&list_lock, flags);
-
-	return block;
-}
-
-/* note: if you have pin'd pages, you should have already unpin'd first! */
-int tiler_release(struct tiler_block *block)
-{
-	int ret = tcm_free(&block->area);
-	unsigned long flags;
-
-	if (block->area.tcm)
-		dev_err(omap_dmm->dev, "failed to release block\n");
-
-	spin_lock_irqsave(&list_lock, flags);
-	list_del(&block->alloc_node);
-	spin_unlock_irqrestore(&list_lock, flags);
-
-	kfree(block);
-	return ret;
-}
-
-/*
- * Utils
- */
-
-/* calculate the tiler space address of a pixel in a view orientation...
- * below description copied from the display subsystem section of TRM:
- *
- * When the TILER is addressed, the bits:
- *   [28:27] = 0x0 for 8-bit tiled
- *             0x1 for 16-bit tiled
- *             0x2 for 32-bit tiled
- *             0x3 for page mode
- *   [31:29] = 0x0 for 0-degree view
- *             0x1 for 180-degree view + mirroring
- *             0x2 for 0-degree view + mirroring
- *             0x3 for 180-degree view
- *             0x4 for 270-degree view + mirroring
- *             0x5 for 270-degree view
- *             0x6 for 90-degree view
- *             0x7 for 90-degree view + mirroring
- * Otherwise the bits indicated the corresponding bit address to access
- * the SDRAM.
- */
-static u32 tiler_get_address(enum tiler_fmt fmt, u32 orient, u32 x, u32 y)
-{
-	u32 x_bits, y_bits, tmp, x_mask, y_mask, alignment;
-
-	x_bits = CONT_WIDTH_BITS - geom[fmt].x_shft;
-	y_bits = CONT_HEIGHT_BITS - geom[fmt].y_shft;
-	alignment = geom[fmt].x_shft + geom[fmt].y_shft;
-
-	/* validate coordinate */
-	x_mask = MASK(x_bits);
-	y_mask = MASK(y_bits);
-
-	if (x < 0 || x > x_mask || y < 0 || y > y_mask) {
-		DBG("invalid coords: %u < 0 || %u > %u || %u < 0 || %u > %u",
-				x, x, x_mask, y, y, y_mask);
-		return 0;
-	}
-
-	/* account for mirroring */
-	if (orient & MASK_X_INVERT)
-		x ^= x_mask;
-	if (orient & MASK_Y_INVERT)
-		y ^= y_mask;
-
-	/* get coordinate address */
-	if (orient & MASK_XY_FLIP)
-		tmp = ((x << y_bits) + y);
-	else
-		tmp = ((y << x_bits) + x);
-
-	return TIL_ADDR((tmp << alignment), orient, fmt);
-}
-
-dma_addr_t tiler_ssptr(struct tiler_block *block)
-{
-	BUG_ON(!validfmt(block->fmt));
-
-	return TILVIEW_8BIT + tiler_get_address(block->fmt, 0,
-			block->area.p0.x * geom[block->fmt].slot_w,
-			block->area.p0.y * geom[block->fmt].slot_h);
-}
-
-dma_addr_t tiler_tsptr(struct tiler_block *block, uint32_t orient,
-		uint32_t x, uint32_t y)
-{
-	struct tcm_pt *p = &block->area.p0;
-	BUG_ON(!validfmt(block->fmt));
-
-	return tiler_get_address(block->fmt, orient,
-			(p->x * geom[block->fmt].slot_w) + x,
-			(p->y * geom[block->fmt].slot_h) + y);
-}
-
-void tiler_align(enum tiler_fmt fmt, uint16_t *w, uint16_t *h)
-{
-	BUG_ON(!validfmt(fmt));
-	*w = round_up(*w, geom[fmt].slot_w);
-	*h = round_up(*h, geom[fmt].slot_h);
-}
-
-uint32_t tiler_stride(enum tiler_fmt fmt, uint32_t orient)
-{
-	BUG_ON(!validfmt(fmt));
-
-	if (orient & MASK_XY_FLIP)
-		return 1 << (CONT_HEIGHT_BITS + geom[fmt].x_shft);
-	else
-		return 1 << (CONT_WIDTH_BITS + geom[fmt].y_shft);
-}
-
-size_t tiler_size(enum tiler_fmt fmt, uint16_t w, uint16_t h)
-{
-	tiler_align(fmt, &w, &h);
-	return geom[fmt].cpp * w * h;
-}
-
-size_t tiler_vsize(enum tiler_fmt fmt, uint16_t w, uint16_t h)
-{
-	BUG_ON(!validfmt(fmt));
-	return round_up(geom[fmt].cpp * w, PAGE_SIZE) * h;
-}
-
-bool dmm_is_available(void)
-{
-	return omap_dmm ? true : false;
-}
-
-static int omap_dmm_remove(struct platform_device *dev)
-{
-	struct tiler_block *block, *_block;
-	int i;
-	unsigned long flags;
-
-	if (omap_dmm) {
-		/* free all area regions */
-		spin_lock_irqsave(&list_lock, flags);
-		list_for_each_entry_safe(block, _block, &omap_dmm->alloc_head,
-					alloc_node) {
-			list_del(&block->alloc_node);
-			kfree(block);
-		}
-		spin_unlock_irqrestore(&list_lock, flags);
-
-		for (i = 0; i < omap_dmm->num_lut; i++)
-			if (omap_dmm->tcm && omap_dmm->tcm[i])
-				omap_dmm->tcm[i]->deinit(omap_dmm->tcm[i]);
-		kfree(omap_dmm->tcm);
-
-		kfree(omap_dmm->engines);
-		if (omap_dmm->refill_va)
-			dma_free_writecombine(omap_dmm->dev,
-				REFILL_BUFFER_SIZE * omap_dmm->num_engines,
-				omap_dmm->refill_va,
-				omap_dmm->refill_pa);
-		if (omap_dmm->dummy_page)
-			__free_page(omap_dmm->dummy_page);
-
-		if (omap_dmm->irq > 0)
-			free_irq(omap_dmm->irq, omap_dmm);
-
-		iounmap(omap_dmm->base);
-		kfree(omap_dmm);
-		omap_dmm = NULL;
-	}
-
-	return 0;
-}
-
-static int omap_dmm_probe(struct platform_device *dev)
-{
-	int ret = -EFAULT, i;
-	struct tcm_area area = {0};
-	u32 hwinfo, pat_geom;
-	struct resource *mem;
-
-	omap_dmm = kzalloc(sizeof(*omap_dmm), GFP_KERNEL);
-	if (!omap_dmm) {
-		dev_err(&dev->dev, "failed to allocate driver data section\n");
-		goto fail;
-	}
-
-	/* initialize lists */
-	INIT_LIST_HEAD(&omap_dmm->alloc_head);
-	INIT_LIST_HEAD(&omap_dmm->idle_head);
-
-	init_waitqueue_head(&omap_dmm->engine_queue);
-
-	/* lookup hwmod data - base address and irq */
-	mem = platform_get_resource(dev, IORESOURCE_MEM, 0);
-	if (!mem) {
-		dev_err(&dev->dev, "failed to get base address resource\n");
-		goto fail;
-	}
-
-	omap_dmm->base = ioremap(mem->start, SZ_2K);
-
-	if (!omap_dmm->base) {
-		dev_err(&dev->dev, "failed to get dmm base address\n");
-		goto fail;
-	}
-
-	omap_dmm->irq = platform_get_irq(dev, 0);
-	if (omap_dmm->irq < 0) {
-		dev_err(&dev->dev, "failed to get IRQ resource\n");
-		goto fail;
-	}
-
-	omap_dmm->dev = &dev->dev;
-
-	hwinfo = readl(omap_dmm->base + DMM_PAT_HWINFO);
-	omap_dmm->num_engines = (hwinfo >> 24) & 0x1F;
-	omap_dmm->num_lut = (hwinfo >> 16) & 0x1F;
-	omap_dmm->container_width = 256;
-	omap_dmm->container_height = 128;
-
-	atomic_set(&omap_dmm->engine_counter, omap_dmm->num_engines);
-
-	/* read out actual LUT width and height */
-	pat_geom = readl(omap_dmm->base + DMM_PAT_GEOMETRY);
-	omap_dmm->lut_width = ((pat_geom >> 16) & 0xF) << 5;
-	omap_dmm->lut_height = ((pat_geom >> 24) & 0xF) << 5;
-
-	/* increment LUT by one if on OMAP5 */
-	/* LUT has twice the height, and is split into a separate container */
-	if (omap_dmm->lut_height != omap_dmm->container_height)
-		omap_dmm->num_lut++;
-
-	/* initialize DMM registers */
-	writel(0x88888888, omap_dmm->base + DMM_PAT_VIEW__0);
-	writel(0x88888888, omap_dmm->base + DMM_PAT_VIEW__1);
-	writel(0x80808080, omap_dmm->base + DMM_PAT_VIEW_MAP__0);
-	writel(0x80000000, omap_dmm->base + DMM_PAT_VIEW_MAP_BASE);
-	writel(0x88888888, omap_dmm->base + DMM_TILER_OR__0);
-	writel(0x88888888, omap_dmm->base + DMM_TILER_OR__1);
-
-	ret = request_irq(omap_dmm->irq, omap_dmm_irq_handler, IRQF_SHARED,
-				"omap_dmm_irq_handler", omap_dmm);
-
-	if (ret) {
-		dev_err(&dev->dev, "couldn't register IRQ %d, error %d\n",
-			omap_dmm->irq, ret);
-		omap_dmm->irq = -1;
-		goto fail;
-	}
-
-	/* Enable all interrupts for each refill engine except
-	 * ERR_LUT_MISS<n> (which is just advisory, and we don't care
-	 * about because we want to be able to refill live scanout
-	 * buffers for accelerated pan/scroll) and FILL_DSC<n> which
-	 * we just generally don't care about.
-	 */
-	writel(0x7e7e7e7e, omap_dmm->base + DMM_PAT_IRQENABLE_SET);
-
-	omap_dmm->dummy_page = alloc_page(GFP_KERNEL | __GFP_DMA32);
-	if (!omap_dmm->dummy_page) {
-		dev_err(&dev->dev, "could not allocate dummy page\n");
-		ret = -ENOMEM;
-		goto fail;
-	}
-
-	/* set dma mask for device */
-	/* NOTE: this is a workaround for the hwmod not initializing properly */
-	dev->dev.coherent_dma_mask = DMA_BIT_MASK(32);
-
-	omap_dmm->dummy_pa = page_to_phys(omap_dmm->dummy_page);
-
-	/* alloc refill memory */
-	omap_dmm->refill_va = dma_alloc_writecombine(&dev->dev,
-				REFILL_BUFFER_SIZE * omap_dmm->num_engines,
-				&omap_dmm->refill_pa, GFP_KERNEL);
-	if (!omap_dmm->refill_va) {
-		dev_err(&dev->dev, "could not allocate refill memory\n");
-		goto fail;
-	}
-
-	/* alloc engines */
-	omap_dmm->engines = kzalloc(
-			omap_dmm->num_engines * sizeof(struct refill_engine),
-			GFP_KERNEL);
-	if (!omap_dmm->engines) {
-		dev_err(&dev->dev, "could not allocate engines\n");
-		ret = -ENOMEM;
-		goto fail;
-	}
-
-	for (i = 0; i < omap_dmm->num_engines; i++) {
-		omap_dmm->engines[i].id = i;
-		omap_dmm->engines[i].dmm = omap_dmm;
-		omap_dmm->engines[i].refill_va = omap_dmm->refill_va +
-						(REFILL_BUFFER_SIZE * i);
-		omap_dmm->engines[i].refill_pa = omap_dmm->refill_pa +
-						(REFILL_BUFFER_SIZE * i);
-		init_waitqueue_head(&omap_dmm->engines[i].wait_for_refill);
-
-		list_add(&omap_dmm->engines[i].idle_node, &omap_dmm->idle_head);
-	}
-
-	omap_dmm->tcm = kzalloc(omap_dmm->num_lut * sizeof(*omap_dmm->tcm),
-				GFP_KERNEL);
-	if (!omap_dmm->tcm) {
-		dev_err(&dev->dev, "failed to allocate lut ptrs\n");
-		ret = -ENOMEM;
-		goto fail;
-	}
-
-	/* init containers */
-	/* Each LUT is associated with a TCM (container manager).  We use the
-	   lut_id to denote the lut_id used to identify the correct LUT for
-	   programming during reill operations */
-	for (i = 0; i < omap_dmm->num_lut; i++) {
-		omap_dmm->tcm[i] = sita_init(omap_dmm->container_width,
-						omap_dmm->container_height,
-						NULL);
-
-		if (!omap_dmm->tcm[i]) {
-			dev_err(&dev->dev, "failed to allocate container\n");
-			ret = -ENOMEM;
-			goto fail;
-		}
-
-		omap_dmm->tcm[i]->lut_id = i;
-	}
-
-	/* assign access mode containers to applicable tcm container */
-	/* OMAP 4 has 1 container for all 4 views */
-	/* OMAP 5 has 2 containers, 1 for 2D and 1 for 1D */
-	containers[TILFMT_8BIT] = omap_dmm->tcm[0];
-	containers[TILFMT_16BIT] = omap_dmm->tcm[0];
-	containers[TILFMT_32BIT] = omap_dmm->tcm[0];
-
-	if (omap_dmm->container_height != omap_dmm->lut_height) {
-		/* second LUT is used for PAGE mode.  Programming must use
-		   y offset that is added to all y coordinates.  LUT id is still
-		   0, because it is the same LUT, just the upper 128 lines */
-		containers[TILFMT_PAGE] = omap_dmm->tcm[1];
-		omap_dmm->tcm[1]->y_offset = OMAP5_LUT_OFFSET;
-		omap_dmm->tcm[1]->lut_id = 0;
-	} else {
-		containers[TILFMT_PAGE] = omap_dmm->tcm[0];
-	}
-
-	area = (struct tcm_area) {
-		.tcm = NULL,
-		.p1.x = omap_dmm->container_width - 1,
-		.p1.y = omap_dmm->container_height - 1,
-	};
-
-	/* initialize all LUTs to dummy page entries */
-	for (i = 0; i < omap_dmm->num_lut; i++) {
-		area.tcm = omap_dmm->tcm[i];
-		if (fill(&area, NULL, 0, 0, true))
-			dev_err(omap_dmm->dev, "refill failed");
-	}
-
-	dev_info(omap_dmm->dev, "initialized all PAT entries\n");
-
-	return 0;
-
-fail:
-	if (omap_dmm_remove(dev))
-		dev_err(&dev->dev, "cleanup failed\n");
-	return ret;
-}
-
-/*
- * debugfs support
- */
-
-#ifdef CONFIG_DEBUG_FS
-
-static const char *alphabet = "abcdefghijklmnopqrstuvwxyz"
-				"ABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789";
-static const char *special = ".,:;'\"`~!^-+";
-
-static void fill_map(char **map, int xdiv, int ydiv, struct tcm_area *a,
-							char c, bool ovw)
-{
-	int x, y;
-	for (y = a->p0.y / ydiv; y <= a->p1.y / ydiv; y++)
-		for (x = a->p0.x / xdiv; x <= a->p1.x / xdiv; x++)
-			if (map[y][x] == ' ' || ovw)
-				map[y][x] = c;
-}
-
-static void fill_map_pt(char **map, int xdiv, int ydiv, struct tcm_pt *p,
-									char c)
-{
-	map[p->y / ydiv][p->x / xdiv] = c;
-}
-
-static char read_map_pt(char **map, int xdiv, int ydiv, struct tcm_pt *p)
-{
-	return map[p->y / ydiv][p->x / xdiv];
-}
-
-static int map_width(int xdiv, int x0, int x1)
-{
-	return (x1 / xdiv) - (x0 / xdiv) + 1;
-}
-
-static void text_map(char **map, int xdiv, char *nice, int yd, int x0, int x1)
-{
-	char *p = map[yd] + (x0 / xdiv);
-	int w = (map_width(xdiv, x0, x1) - strlen(nice)) / 2;
-	if (w >= 0) {
-		p += w;
-		while (*nice)
-			*p++ = *nice++;
-	}
-}
-
-static void map_1d_info(char **map, int xdiv, int ydiv, char *nice,
-							struct tcm_area *a)
-{
-	sprintf(nice, "%dK", tcm_sizeof(*a) * 4);
-	if (a->p0.y + 1 < a->p1.y) {
-		text_map(map, xdiv, nice, (a->p0.y + a->p1.y) / 2 / ydiv, 0,
-							256 - 1);
-	} else if (a->p0.y < a->p1.y) {
-		if (strlen(nice) < map_width(xdiv, a->p0.x, 256 - 1))
-			text_map(map, xdiv, nice, a->p0.y / ydiv,
-					a->p0.x + xdiv,	256 - 1);
-		else if (strlen(nice) < map_width(xdiv, 0, a->p1.x))
-			text_map(map, xdiv, nice, a->p1.y / ydiv,
-					0, a->p1.y - xdiv);
-	} else if (strlen(nice) + 1 < map_width(xdiv, a->p0.x, a->p1.x)) {
-		text_map(map, xdiv, nice, a->p0.y / ydiv, a->p0.x, a->p1.x);
-	}
-}
-
-static void map_2d_info(char **map, int xdiv, int ydiv, char *nice,
-							struct tcm_area *a)
-{
-	sprintf(nice, "(%d*%d)", tcm_awidth(*a), tcm_aheight(*a));
-	if (strlen(nice) + 1 < map_width(xdiv, a->p0.x, a->p1.x))
-		text_map(map, xdiv, nice, (a->p0.y + a->p1.y) / 2 / ydiv,
-							a->p0.x, a->p1.x);
-}
-
-int tiler_map_show(struct seq_file *s, void *arg)
-{
-	int xdiv = 2, ydiv = 1;
-	char **map = NULL, *global_map;
-	struct tiler_block *block;
-	struct tcm_area a, p;
-	int i;
-	const char *m2d = alphabet;
-	const char *a2d = special;
-	const char *m2dp = m2d, *a2dp = a2d;
-	char nice[128];
-	int h_adj;
-	int w_adj;
-	unsigned long flags;
-	int lut_idx;
-
-
-	if (!omap_dmm) {
-		/* early return if dmm/tiler device is not initialized */
-		return 0;
-	}
-
-	h_adj = omap_dmm->container_height / ydiv;
-	w_adj = omap_dmm->container_width / xdiv;
-
-	map = kmalloc(h_adj * sizeof(*map), GFP_KERNEL);
-	global_map = kmalloc((w_adj + 1) * h_adj, GFP_KERNEL);
-
-	if (!map || !global_map)
-		goto error;
-
-	for (lut_idx = 0; lut_idx < omap_dmm->num_lut; lut_idx++) {
-		memset(map, 0, sizeof(h_adj * sizeof(*map)));
-		memset(global_map, ' ', (w_adj + 1) * h_adj);
-
-		for (i = 0; i < omap_dmm->container_height; i++) {
-			map[i] = global_map + i * (w_adj + 1);
-			map[i][w_adj] = 0;
-		}
-
-		spin_lock_irqsave(&list_lock, flags);
-
-		list_for_each_entry(block, &omap_dmm->alloc_head, alloc_node) {
-			if (block->area.tcm == omap_dmm->tcm[lut_idx]) {
-				if (block->fmt != TILFMT_PAGE) {
-					fill_map(map, xdiv, ydiv, &block->area,
-						*m2dp, true);
-					if (!*++a2dp)
-						a2dp = a2d;
-					if (!*++m2dp)
-						m2dp = m2d;
-					map_2d_info(map, xdiv, ydiv, nice,
-							&block->area);
-				} else {
-					bool start = read_map_pt(map, xdiv,
-						ydiv, &block->area.p0) == ' ';
-					bool end = read_map_pt(map, xdiv, ydiv,
-							&block->area.p1) == ' ';
-
-					tcm_for_each_slice(a, block->area, p)
-						fill_map(map, xdiv, ydiv, &a,
-							'=', true);
-					fill_map_pt(map, xdiv, ydiv,
-							&block->area.p0,
-							start ? '<' : 'X');
-					fill_map_pt(map, xdiv, ydiv,
-							&block->area.p1,
-							end ? '>' : 'X');
-					map_1d_info(map, xdiv, ydiv, nice,
-							&block->area);
-				}
-			}
-		}
-
-		spin_unlock_irqrestore(&list_lock, flags);
-
-		if (s) {
-			seq_printf(s, "CONTAINER %d DUMP BEGIN\n", lut_idx);
-			for (i = 0; i < 128; i++)
-				seq_printf(s, "%03d:%s\n", i, map[i]);
-			seq_printf(s, "CONTAINER %d DUMP END\n", lut_idx);
-		} else {
-			dev_dbg(omap_dmm->dev, "CONTAINER %d DUMP BEGIN\n",
-				lut_idx);
-			for (i = 0; i < 128; i++)
-				dev_dbg(omap_dmm->dev, "%03d:%s\n", i, map[i]);
-			dev_dbg(omap_dmm->dev, "CONTAINER %d DUMP END\n",
-				lut_idx);
-		}
-	}
-
-error:
-	kfree(map);
-	kfree(global_map);
-
-	return 0;
-}
-#endif
-
-#ifdef CONFIG_PM
-static int omap_dmm_resume(struct device *dev)
-{
-	struct tcm_area area;
-	int i;
-
-	if (!omap_dmm)
-		return -ENODEV;
-
-	area = (struct tcm_area) {
-		.tcm = NULL,
-		.p1.x = omap_dmm->container_width - 1,
-		.p1.y = omap_dmm->container_height - 1,
-	};
-
-	/* initialize all LUTs to dummy page entries */
-	for (i = 0; i < omap_dmm->num_lut; i++) {
-		area.tcm = omap_dmm->tcm[i];
-		if (fill(&area, NULL, 0, 0, true))
-			dev_err(dev, "refill failed");
-	}
-
-	return 0;
-}
-
-static const struct dev_pm_ops omap_dmm_pm_ops = {
-	.resume = omap_dmm_resume,
-};
-#endif
-
-struct platform_driver omap_dmm_driver = {
-	.probe = omap_dmm_probe,
-	.remove = omap_dmm_remove,
-	.driver = {
-		.owner = THIS_MODULE,
-		.name = DMM_DRIVER_NAME,
-#ifdef CONFIG_PM
-		.pm = &omap_dmm_pm_ops,
-#endif
-	},
-};
-
-MODULE_LICENSE("GPL v2");
-MODULE_AUTHOR("Andy Gross <andy.gross@ti.com>");
-MODULE_DESCRIPTION("OMAP DMM/Tiler Driver");
-MODULE_ALIAS("platform:" DMM_DRIVER_NAME);
diff --git a/drivers/staging/omapdrm/omap_dmm_tiler.h b/drivers/staging/omapdrm/omap_dmm_tiler.h
deleted file mode 100644
index 4fdd61e..0000000
--- a/drivers/staging/omapdrm/omap_dmm_tiler.h
+++ /dev/null
@@ -1,141 +0,0 @@
-/*
- *
- * Copyright (C) 2011 Texas Instruments Incorporated - http://www.ti.com/
- * Author: Rob Clark <rob@ti.com>
- *         Andy Gross <andy.gross@ti.com>
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation version 2.
- *
- * This program is distributed "as is" WITHOUT ANY WARRANTY of any
- * kind, whether express or implied; without even the implied warranty
- * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- */
-#ifndef OMAP_DMM_TILER_H
-#define OMAP_DMM_TILER_H
-
-#include "omap_drv.h"
-#include "tcm.h"
-
-enum tiler_fmt {
-	TILFMT_8BIT = 0,
-	TILFMT_16BIT,
-	TILFMT_32BIT,
-	TILFMT_PAGE,
-	TILFMT_NFORMATS
-};
-
-struct pat_area {
-	u32 x0:8;
-	u32 y0:8;
-	u32 x1:8;
-	u32 y1:8;
-};
-
-struct tiler_block {
-	struct list_head alloc_node;	/* node for global block list */
-	struct tcm_area area;		/* area */
-	enum tiler_fmt fmt;		/* format */
-};
-
-/* bits representing the same slot in DMM-TILER hw-block */
-#define SLOT_WIDTH_BITS         6
-#define SLOT_HEIGHT_BITS        6
-
-/* bits reserved to describe coordinates in DMM-TILER hw-block */
-#define CONT_WIDTH_BITS         14
-#define CONT_HEIGHT_BITS        13
-
-/* calculated constants */
-#define TILER_PAGE              (1 << (SLOT_WIDTH_BITS + SLOT_HEIGHT_BITS))
-#define TILER_WIDTH             (1 << (CONT_WIDTH_BITS - SLOT_WIDTH_BITS))
-#define TILER_HEIGHT            (1 << (CONT_HEIGHT_BITS - SLOT_HEIGHT_BITS))
-
-/*
-Table 15-11. Coding and Description of TILER Orientations
-S Y X	Description				Alternate description
-0 0 0	0-degree view				Natural view
-0 0 1	0-degree view with vertical mirror 	180-degree view with horizontal mirror
-0 1 0	0-degree view with horizontal mirror 	180-degree view with vertical mirror
-0 1 1	180-degree view
-1 0 0	90-degree view with vertical mirror	270-degree view with horizontal mirror
-1 0 1	270-degree view
-1 1 0	90-degree view
-1 1 1	90-degree view with horizontal mirror	270-degree view with vertical mirror
- */
-#define MASK_XY_FLIP		(1 << 31)
-#define MASK_Y_INVERT		(1 << 30)
-#define MASK_X_INVERT		(1 << 29)
-#define SHIFT_ACC_MODE		27
-#define MASK_ACC_MODE		3
-
-#define MASK(bits) ((1 << (bits)) - 1)
-
-#define TILVIEW_8BIT    0x60000000u
-#define TILVIEW_16BIT   (TILVIEW_8BIT  + VIEW_SIZE)
-#define TILVIEW_32BIT   (TILVIEW_16BIT + VIEW_SIZE)
-#define TILVIEW_PAGE    (TILVIEW_32BIT + VIEW_SIZE)
-#define TILVIEW_END     (TILVIEW_PAGE  + VIEW_SIZE)
-
-/* create tsptr by adding view orientation and access mode */
-#define TIL_ADDR(x, orient, a)\
-	((u32) (x) | (orient) | ((a) << SHIFT_ACC_MODE))
-
-#ifdef CONFIG_DEBUG_FS
-int tiler_map_show(struct seq_file *s, void *arg);
-#endif
-
-/* pin/unpin */
-int tiler_pin(struct tiler_block *block, struct page **pages,
-		uint32_t npages, uint32_t roll, bool wait);
-int tiler_unpin(struct tiler_block *block);
-
-/* reserve/release */
-struct tiler_block *tiler_reserve_2d(enum tiler_fmt fmt, uint16_t w, uint16_t h,
-				uint16_t align);
-struct tiler_block *tiler_reserve_1d(size_t size);
-int tiler_release(struct tiler_block *block);
-
-/* utilities */
-dma_addr_t tiler_ssptr(struct tiler_block *block);
-dma_addr_t tiler_tsptr(struct tiler_block *block, uint32_t orient,
-		uint32_t x, uint32_t y);
-uint32_t tiler_stride(enum tiler_fmt fmt, uint32_t orient);
-size_t tiler_size(enum tiler_fmt fmt, uint16_t w, uint16_t h);
-size_t tiler_vsize(enum tiler_fmt fmt, uint16_t w, uint16_t h);
-void tiler_align(enum tiler_fmt fmt, uint16_t *w, uint16_t *h);
-bool dmm_is_available(void);
-
-extern struct platform_driver omap_dmm_driver;
-
-/* GEM bo flags -> tiler fmt */
-static inline enum tiler_fmt gem2fmt(uint32_t flags)
-{
-	switch (flags & OMAP_BO_TILED) {
-	case OMAP_BO_TILED_8:
-		return TILFMT_8BIT;
-	case OMAP_BO_TILED_16:
-		return TILFMT_16BIT;
-	case OMAP_BO_TILED_32:
-		return TILFMT_32BIT;
-	default:
-		return TILFMT_PAGE;
-	}
-}
-
-static inline bool validfmt(enum tiler_fmt fmt)
-{
-	switch (fmt) {
-	case TILFMT_8BIT:
-	case TILFMT_16BIT:
-	case TILFMT_32BIT:
-	case TILFMT_PAGE:
-		return true;
-	default:
-		return false;
-	}
-}
-
-#endif
diff --git a/drivers/staging/omapdrm/omap_drm.h b/drivers/staging/omapdrm/omap_drm.h
deleted file mode 100644
index f0ac34a..0000000
--- a/drivers/staging/omapdrm/omap_drm.h
+++ /dev/null
@@ -1,123 +0,0 @@
-/*
- * include/drm/omap_drm.h
- *
- * Copyright (C) 2011 Texas Instruments
- * Author: Rob Clark <rob@ti.com>
- *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of the GNU General Public License version 2 as published by
- * the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but WITHOUT
- * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
- * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
- * more details.
- *
- * You should have received a copy of the GNU General Public License along with
- * this program.  If not, see <http://www.gnu.org/licenses/>.
- */
-
-#ifndef __OMAP_DRM_H__
-#define __OMAP_DRM_H__
-
-#include <drm/drm.h>
-
-/* Please note that modifications to all structs defined here are
- * subject to backwards-compatibility constraints.
- */
-
-#define OMAP_PARAM_CHIPSET_ID	1	/* ie. 0x3430, 0x4430, etc */
-
-struct drm_omap_param {
-	uint64_t param;			/* in */
-	uint64_t value;			/* in (set_param), out (get_param) */
-};
-
-#define OMAP_BO_SCANOUT		0x00000001	/* scanout capable (phys contiguous) */
-#define OMAP_BO_CACHE_MASK	0x00000006	/* cache type mask, see cache modes */
-#define OMAP_BO_TILED_MASK	0x00000f00	/* tiled mapping mask, see tiled modes */
-
-/* cache modes */
-#define OMAP_BO_CACHED		0x00000000	/* default */
-#define OMAP_BO_WC		0x00000002	/* write-combine */
-#define OMAP_BO_UNCACHED	0x00000004	/* strongly-ordered (uncached) */
-
-/* tiled modes */
-#define OMAP_BO_TILED_8		0x00000100
-#define OMAP_BO_TILED_16	0x00000200
-#define OMAP_BO_TILED_32	0x00000300
-#define OMAP_BO_TILED		(OMAP_BO_TILED_8 | OMAP_BO_TILED_16 | OMAP_BO_TILED_32)
-
-union omap_gem_size {
-	uint32_t bytes;		/* (for non-tiled formats) */
-	struct {
-		uint16_t width;
-		uint16_t height;
-	} tiled;		/* (for tiled formats) */
-};
-
-struct drm_omap_gem_new {
-	union omap_gem_size size;	/* in */
-	uint32_t flags;			/* in */
-	uint32_t handle;		/* out */
-	uint32_t __pad;
-};
-
-/* mask of operations: */
-enum omap_gem_op {
-	OMAP_GEM_READ = 0x01,
-	OMAP_GEM_WRITE = 0x02,
-};
-
-struct drm_omap_gem_cpu_prep {
-	uint32_t handle;		/* buffer handle (in) */
-	uint32_t op;			/* mask of omap_gem_op (in) */
-};
-
-struct drm_omap_gem_cpu_fini {
-	uint32_t handle;		/* buffer handle (in) */
-	uint32_t op;			/* mask of omap_gem_op (in) */
-	/* TODO maybe here we pass down info about what regions are touched
-	 * by sw so we can be clever about cache ops?  For now a placeholder,
-	 * set to zero and we just do full buffer flush..
-	 */
-	uint32_t nregions;
-	uint32_t __pad;
-};
-
-struct drm_omap_gem_info {
-	uint32_t handle;		/* buffer handle (in) */
-	uint32_t pad;
-	uint64_t offset;		/* mmap offset (out) */
-	/* note: in case of tiled buffers, the user virtual size can be
-	 * different from the physical size (ie. how many pages are needed
-	 * to back the object) which is returned in DRM_IOCTL_GEM_OPEN..
-	 * This size here is the one that should be used if you want to
-	 * mmap() the buffer:
-	 */
-	uint32_t size;			/* virtual size for mmap'ing (out) */
-	uint32_t __pad;
-};
-
-#define DRM_OMAP_GET_PARAM		0x00
-#define DRM_OMAP_SET_PARAM		0x01
-/* placeholder for plugin-api
-#define DRM_OMAP_GET_BASE		0x02
-*/
-#define DRM_OMAP_GEM_NEW		0x03
-#define DRM_OMAP_GEM_CPU_PREP		0x04
-#define DRM_OMAP_GEM_CPU_FINI		0x05
-#define DRM_OMAP_GEM_INFO		0x06
-#define DRM_OMAP_NUM_IOCTLS		0x07
-
-#define DRM_IOCTL_OMAP_GET_PARAM	DRM_IOWR(DRM_COMMAND_BASE + DRM_OMAP_GET_PARAM, struct drm_omap_param)
-#define DRM_IOCTL_OMAP_SET_PARAM	DRM_IOW (DRM_COMMAND_BASE + DRM_OMAP_SET_PARAM, struct drm_omap_param)
-/* placeholder for plugin-api
-#define DRM_IOCTL_OMAP_GET_BASE		DRM_IOWR(DRM_COMMAND_BASE + DRM_OMAP_GET_BASE, struct drm_omap_get_base)
-*/
-#define DRM_IOCTL_OMAP_GEM_NEW		DRM_IOWR(DRM_COMMAND_BASE + DRM_OMAP_GEM_NEW, struct drm_omap_gem_new)
-#define DRM_IOCTL_OMAP_GEM_CPU_PREP	DRM_IOW (DRM_COMMAND_BASE + DRM_OMAP_GEM_CPU_PREP, struct drm_omap_gem_cpu_prep)
-#define DRM_IOCTL_OMAP_GEM_CPU_FINI	DRM_IOW (DRM_COMMAND_BASE + DRM_OMAP_GEM_CPU_FINI, struct drm_omap_gem_cpu_fini)
-#define DRM_IOCTL_OMAP_GEM_INFO		DRM_IOWR(DRM_COMMAND_BASE + DRM_OMAP_GEM_INFO, struct drm_omap_gem_info)
-
-#endif /* __OMAP_DRM_H__ */
diff --git a/drivers/staging/omapdrm/omap_drv.c b/drivers/staging/omapdrm/omap_drv.c
deleted file mode 100644
index 9d6a584..0000000
--- a/drivers/staging/omapdrm/omap_drv.c
+++ /dev/null
@@ -1,610 +0,0 @@
-/*
- * drivers/staging/omapdrm/omap_drv.c
- *
- * Copyright (C) 2011 Texas Instruments
- * Author: Rob Clark <rob@ti.com>
- *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of the GNU General Public License version 2 as published by
- * the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but WITHOUT
- * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
- * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
- * more details.
- *
- * You should have received a copy of the GNU General Public License along with
- * this program.  If not, see <http://www.gnu.org/licenses/>.
- */
-
-#include "omap_drv.h"
-
-#include "drm_crtc_helper.h"
-#include "drm_fb_helper.h"
-#include "omap_dmm_tiler.h"
-
-#define DRIVER_NAME		MODULE_NAME
-#define DRIVER_DESC		"OMAP DRM"
-#define DRIVER_DATE		"20110917"
-#define DRIVER_MAJOR		1
-#define DRIVER_MINOR		0
-#define DRIVER_PATCHLEVEL	0
-
-static int num_crtc = CONFIG_DRM_OMAP_NUM_CRTCS;
-
-MODULE_PARM_DESC(num_crtc, "Number of overlays to use as CRTCs");
-module_param(num_crtc, int, 0600);
-
-/*
- * mode config funcs
- */
-
-/* Notes about mapping DSS and DRM entities:
- *    CRTC:        overlay
- *    encoder:     manager.. with some extension to allow one primary CRTC
- *                 and zero or more video CRTC's to be mapped to one encoder?
- *    connector:   dssdev.. manager can be attached/detached from different
- *                 devices
- */
-
-static void omap_fb_output_poll_changed(struct drm_device *dev)
-{
-	struct omap_drm_private *priv = dev->dev_private;
-	DBG("dev=%p", dev);
-	if (priv->fbdev)
-		drm_fb_helper_hotplug_event(priv->fbdev);
-}
-
-static const struct drm_mode_config_funcs omap_mode_config_funcs = {
-	.fb_create = omap_framebuffer_create,
-	.output_poll_changed = omap_fb_output_poll_changed,
-};
-
-static int get_connector_type(struct omap_dss_device *dssdev)
-{
-	switch (dssdev->type) {
-	case OMAP_DISPLAY_TYPE_HDMI:
-		return DRM_MODE_CONNECTOR_HDMIA;
-	case OMAP_DISPLAY_TYPE_DPI:
-		if (!strcmp(dssdev->name, "dvi"))
-			return DRM_MODE_CONNECTOR_DVID;
-		/* fallthrough */
-	default:
-		return DRM_MODE_CONNECTOR_Unknown;
-	}
-}
-
-static int omap_modeset_init(struct drm_device *dev)
-{
-	struct omap_drm_private *priv = dev->dev_private;
-	struct omap_dss_device *dssdev = NULL;
-	int num_ovls = dss_feat_get_num_ovls();
-	int id;
-
-	drm_mode_config_init(dev);
-
-	omap_drm_irq_install(dev);
-
-	/*
-	 * Create private planes and CRTCs for the last NUM_CRTCs overlay
-	 * plus manager:
-	 */
-	for (id = 0; id < min(num_crtc, num_ovls); id++) {
-		struct drm_plane *plane;
-		struct drm_crtc *crtc;
-
-		plane = omap_plane_init(dev, id, true);
-		crtc = omap_crtc_init(dev, plane, pipe2chan(id), id);
-
-		BUG_ON(priv->num_crtcs >= ARRAY_SIZE(priv->crtcs));
-		priv->crtcs[id] = crtc;
-		priv->num_crtcs++;
-
-		priv->planes[id] = plane;
-		priv->num_planes++;
-	}
-
-	/*
-	 * Create normal planes for the remaining overlays:
-	 */
-	for (; id < num_ovls; id++) {
-		struct drm_plane *plane = omap_plane_init(dev, id, false);
-
-		BUG_ON(priv->num_planes >= ARRAY_SIZE(priv->planes));
-		priv->planes[priv->num_planes++] = plane;
-	}
-
-	for_each_dss_dev(dssdev) {
-		struct drm_connector *connector;
-		struct drm_encoder *encoder;
-
-		if (!dssdev->driver) {
-			dev_warn(dev->dev, "%s has no driver.. skipping it\n",
-					dssdev->name);
-			return 0;
-		}
-
-		if (!(dssdev->driver->get_timings ||
-					dssdev->driver->read_edid)) {
-			dev_warn(dev->dev, "%s driver does not support "
-				"get_timings or read_edid.. skipping it!\n",
-				dssdev->name);
-			return 0;
-		}
-
-		encoder = omap_encoder_init(dev, dssdev);
-
-		if (!encoder) {
-			dev_err(dev->dev, "could not create encoder: %s\n",
-					dssdev->name);
-			return -ENOMEM;
-		}
-
-		connector = omap_connector_init(dev,
-				get_connector_type(dssdev), dssdev, encoder);
-
-		if (!connector) {
-			dev_err(dev->dev, "could not create connector: %s\n",
-					dssdev->name);
-			return -ENOMEM;
-		}
-
-		BUG_ON(priv->num_encoders >= ARRAY_SIZE(priv->encoders));
-		BUG_ON(priv->num_connectors >= ARRAY_SIZE(priv->connectors));
-
-		priv->encoders[priv->num_encoders++] = encoder;
-		priv->connectors[priv->num_connectors++] = connector;
-
-		drm_mode_connector_attach_encoder(connector, encoder);
-
-		/* figure out which crtc's we can connect the encoder to: */
-		encoder->possible_crtcs = 0;
-		for (id = 0; id < priv->num_crtcs; id++) {
-			enum omap_dss_output_id supported_outputs =
-					dss_feat_get_supported_outputs(pipe2chan(id));
-			if (supported_outputs & dssdev->output->id)
-				encoder->possible_crtcs |= (1 << id);
-		}
-	}
-
-	dev->mode_config.min_width = 32;
-	dev->mode_config.min_height = 32;
-
-	/* note: eventually will need some cpu_is_omapXYZ() type stuff here
-	 * to fill in these limits properly on different OMAP generations..
-	 */
-	dev->mode_config.max_width = 2048;
-	dev->mode_config.max_height = 2048;
-
-	dev->mode_config.funcs = &omap_mode_config_funcs;
-
-	return 0;
-}
-
-static void omap_modeset_free(struct drm_device *dev)
-{
-	drm_mode_config_cleanup(dev);
-}
-
-/*
- * drm ioctl funcs
- */
-
-
-static int ioctl_get_param(struct drm_device *dev, void *data,
-		struct drm_file *file_priv)
-{
-	struct omap_drm_private *priv = dev->dev_private;
-	struct drm_omap_param *args = data;
-
-	DBG("%p: param=%llu", dev, args->param);
-
-	switch (args->param) {
-	case OMAP_PARAM_CHIPSET_ID:
-		args->value = priv->omaprev;
-		break;
-	default:
-		DBG("unknown parameter %lld", args->param);
-		return -EINVAL;
-	}
-
-	return 0;
-}
-
-static int ioctl_set_param(struct drm_device *dev, void *data,
-		struct drm_file *file_priv)
-{
-	struct drm_omap_param *args = data;
-
-	switch (args->param) {
-	default:
-		DBG("unknown parameter %lld", args->param);
-		return -EINVAL;
-	}
-
-	return 0;
-}
-
-static int ioctl_gem_new(struct drm_device *dev, void *data,
-		struct drm_file *file_priv)
-{
-	struct drm_omap_gem_new *args = data;
-	VERB("%p:%p: size=0x%08x, flags=%08x", dev, file_priv,
-			args->size.bytes, args->flags);
-	return omap_gem_new_handle(dev, file_priv, args->size,
-			args->flags, &args->handle);
-}
-
-static int ioctl_gem_cpu_prep(struct drm_device *dev, void *data,
-		struct drm_file *file_priv)
-{
-	struct drm_omap_gem_cpu_prep *args = data;
-	struct drm_gem_object *obj;
-	int ret;
-
-	VERB("%p:%p: handle=%d, op=%x", dev, file_priv, args->handle, args->op);
-
-	obj = drm_gem_object_lookup(dev, file_priv, args->handle);
-	if (!obj)
-		return -ENOENT;
-
-	ret = omap_gem_op_sync(obj, args->op);
-
-	if (!ret)
-		ret = omap_gem_op_start(obj, args->op);
-
-	drm_gem_object_unreference_unlocked(obj);
-
-	return ret;
-}
-
-static int ioctl_gem_cpu_fini(struct drm_device *dev, void *data,
-		struct drm_file *file_priv)
-{
-	struct drm_omap_gem_cpu_fini *args = data;
-	struct drm_gem_object *obj;
-	int ret;
-
-	VERB("%p:%p: handle=%d", dev, file_priv, args->handle);
-
-	obj = drm_gem_object_lookup(dev, file_priv, args->handle);
-	if (!obj)
-		return -ENOENT;
-
-	/* XXX flushy, flushy */
-	ret = 0;
-
-	if (!ret)
-		ret = omap_gem_op_finish(obj, args->op);
-
-	drm_gem_object_unreference_unlocked(obj);
-
-	return ret;
-}
-
-static int ioctl_gem_info(struct drm_device *dev, void *data,
-		struct drm_file *file_priv)
-{
-	struct drm_omap_gem_info *args = data;
-	struct drm_gem_object *obj;
-	int ret = 0;
-
-	VERB("%p:%p: handle=%d", dev, file_priv, args->handle);
-
-	obj = drm_gem_object_lookup(dev, file_priv, args->handle);
-	if (!obj)
-		return -ENOENT;
-
-	args->size = omap_gem_mmap_size(obj);
-	args->offset = omap_gem_mmap_offset(obj);
-
-	drm_gem_object_unreference_unlocked(obj);
-
-	return ret;
-}
-
-struct drm_ioctl_desc ioctls[DRM_COMMAND_END - DRM_COMMAND_BASE] = {
-	DRM_IOCTL_DEF_DRV(OMAP_GET_PARAM, ioctl_get_param, DRM_UNLOCKED|DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(OMAP_SET_PARAM, ioctl_set_param, DRM_UNLOCKED|DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF_DRV(OMAP_GEM_NEW, ioctl_gem_new, DRM_UNLOCKED|DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(OMAP_GEM_CPU_PREP, ioctl_gem_cpu_prep, DRM_UNLOCKED|DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(OMAP_GEM_CPU_FINI, ioctl_gem_cpu_fini, DRM_UNLOCKED|DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(OMAP_GEM_INFO, ioctl_gem_info, DRM_UNLOCKED|DRM_AUTH),
-};
-
-/*
- * drm driver funcs
- */
-
-/**
- * load - setup chip and create an initial config
- * @dev: DRM device
- * @flags: startup flags
- *
- * The driver load routine has to do several things:
- *   - initialize the memory manager
- *   - allocate initial config memory
- *   - setup the DRM framebuffer with the allocated memory
- */
-static int dev_load(struct drm_device *dev, unsigned long flags)
-{
-	struct omap_drm_platform_data *pdata = dev->dev->platform_data;
-	struct omap_drm_private *priv;
-	int ret;
-
-	DBG("load: dev=%p", dev);
-
-	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
-	if (!priv) {
-		dev_err(dev->dev, "could not allocate priv\n");
-		return -ENOMEM;
-	}
-
-	priv->omaprev = pdata->omaprev;
-
-	dev->dev_private = priv;
-
-	priv->wq = alloc_ordered_workqueue("omapdrm", 0);
-
-	INIT_LIST_HEAD(&priv->obj_list);
-
-	omap_gem_init(dev);
-
-	ret = omap_modeset_init(dev);
-	if (ret) {
-		dev_err(dev->dev, "omap_modeset_init failed: ret=%d\n", ret);
-		dev->dev_private = NULL;
-		kfree(priv);
-		return ret;
-	}
-
-	ret = drm_vblank_init(dev, priv->num_crtcs);
-	if (ret)
-		dev_warn(dev->dev, "could not init vblank\n");
-
-	priv->fbdev = omap_fbdev_init(dev);
-	if (!priv->fbdev) {
-		dev_warn(dev->dev, "omap_fbdev_init failed\n");
-		/* well, limp along without an fbdev.. maybe X11 will work? */
-	}
-
-	/* store off drm_device for use in pm ops */
-	dev_set_drvdata(dev->dev, dev);
-
-	drm_kms_helper_poll_init(dev);
-
-	return 0;
-}
-
-static int dev_unload(struct drm_device *dev)
-{
-	struct omap_drm_private *priv = dev->dev_private;
-
-	DBG("unload: dev=%p", dev);
-
-	drm_kms_helper_poll_fini(dev);
-	drm_vblank_cleanup(dev);
-	omap_drm_irq_uninstall(dev);
-
-	omap_fbdev_free(dev);
-	omap_modeset_free(dev);
-	omap_gem_deinit(dev);
-
-	flush_workqueue(priv->wq);
-	destroy_workqueue(priv->wq);
-
-	kfree(dev->dev_private);
-	dev->dev_private = NULL;
-
-	dev_set_drvdata(dev->dev, NULL);
-
-	return 0;
-}
-
-static int dev_open(struct drm_device *dev, struct drm_file *file)
-{
-	file->driver_priv = NULL;
-
-	DBG("open: dev=%p, file=%p", dev, file);
-
-	return 0;
-}
-
-static int dev_firstopen(struct drm_device *dev)
-{
-	DBG("firstopen: dev=%p", dev);
-	return 0;
-}
-
-/**
- * lastclose - clean up after all DRM clients have exited
- * @dev: DRM device
- *
- * Take care of cleaning up after all DRM clients have exited.  In the
- * mode setting case, we want to restore the kernel's initial mode (just
- * in case the last client left us in a bad state).
- */
-static void dev_lastclose(struct drm_device *dev)
-{
-	int i;
-
-	/* we don't support vga-switcheroo.. so just make sure the fbdev
-	 * mode is active
-	 */
-	struct omap_drm_private *priv = dev->dev_private;
-	int ret;
-
-	DBG("lastclose: dev=%p", dev);
-
-	if (priv->rotation_prop) {
-		/* need to restore default rotation state.. not sure
-		 * if there is a cleaner way to restore properties to
-		 * default state?  Maybe a flag that properties should
-		 * automatically be restored to default state on
-		 * lastclose?
-		 */
-		for (i = 0; i < priv->num_crtcs; i++) {
-			drm_object_property_set_value(&priv->crtcs[i]->base,
-					priv->rotation_prop, 0);
-		}
-
-		for (i = 0; i < priv->num_planes; i++) {
-			drm_object_property_set_value(&priv->planes[i]->base,
-					priv->rotation_prop, 0);
-		}
-	}
-
-	drm_modeset_lock_all(dev);
-	ret = drm_fb_helper_restore_fbdev_mode(priv->fbdev);
-	drm_modeset_unlock_all(dev);
-	if (ret)
-		DBG("failed to restore crtc mode");
-}
-
-static void dev_preclose(struct drm_device *dev, struct drm_file *file)
-{
-	DBG("preclose: dev=%p", dev);
-}
-
-static void dev_postclose(struct drm_device *dev, struct drm_file *file)
-{
-	DBG("postclose: dev=%p, file=%p", dev, file);
-}
-
-static const struct vm_operations_struct omap_gem_vm_ops = {
-	.fault = omap_gem_fault,
-	.open = drm_gem_vm_open,
-	.close = drm_gem_vm_close,
-};
-
-static const struct file_operations omapdriver_fops = {
-		.owner = THIS_MODULE,
-		.open = drm_open,
-		.unlocked_ioctl = drm_ioctl,
-		.release = drm_release,
-		.mmap = omap_gem_mmap,
-		.poll = drm_poll,
-		.fasync = drm_fasync,
-		.read = drm_read,
-		.llseek = noop_llseek,
-};
-
-static struct drm_driver omap_drm_driver = {
-		.driver_features =
-				DRIVER_HAVE_IRQ | DRIVER_MODESET | DRIVER_GEM | DRIVER_PRIME,
-		.load = dev_load,
-		.unload = dev_unload,
-		.open = dev_open,
-		.firstopen = dev_firstopen,
-		.lastclose = dev_lastclose,
-		.preclose = dev_preclose,
-		.postclose = dev_postclose,
-		.get_vblank_counter = drm_vblank_count,
-		.enable_vblank = omap_irq_enable_vblank,
-		.disable_vblank = omap_irq_disable_vblank,
-		.irq_preinstall = omap_irq_preinstall,
-		.irq_postinstall = omap_irq_postinstall,
-		.irq_uninstall = omap_irq_uninstall,
-		.irq_handler = omap_irq_handler,
-#ifdef CONFIG_DEBUG_FS
-		.debugfs_init = omap_debugfs_init,
-		.debugfs_cleanup = omap_debugfs_cleanup,
-#endif
-		.prime_handle_to_fd = drm_gem_prime_handle_to_fd,
-		.prime_fd_to_handle = drm_gem_prime_fd_to_handle,
-		.gem_prime_export = omap_gem_prime_export,
-		.gem_prime_import = omap_gem_prime_import,
-		.gem_init_object = omap_gem_init_object,
-		.gem_free_object = omap_gem_free_object,
-		.gem_vm_ops = &omap_gem_vm_ops,
-		.dumb_create = omap_gem_dumb_create,
-		.dumb_map_offset = omap_gem_dumb_map_offset,
-		.dumb_destroy = omap_gem_dumb_destroy,
-		.ioctls = ioctls,
-		.num_ioctls = DRM_OMAP_NUM_IOCTLS,
-		.fops = &omapdriver_fops,
-		.name = DRIVER_NAME,
-		.desc = DRIVER_DESC,
-		.date = DRIVER_DATE,
-		.major = DRIVER_MAJOR,
-		.minor = DRIVER_MINOR,
-		.patchlevel = DRIVER_PATCHLEVEL,
-};
-
-static int pdev_suspend(struct platform_device *pDevice, pm_message_t state)
-{
-	DBG("");
-	return 0;
-}
-
-static int pdev_resume(struct platform_device *device)
-{
-	DBG("");
-	return 0;
-}
-
-static void pdev_shutdown(struct platform_device *device)
-{
-	DBG("");
-}
-
-static int pdev_probe(struct platform_device *device)
-{
-	DBG("%s", device->name);
-	return drm_platform_init(&omap_drm_driver, device);
-}
-
-static int pdev_remove(struct platform_device *device)
-{
-	DBG("");
-	drm_platform_exit(&omap_drm_driver, device);
-
-	platform_driver_unregister(&omap_dmm_driver);
-	return 0;
-}
-
-#ifdef CONFIG_PM
-static const struct dev_pm_ops omapdrm_pm_ops = {
-	.resume = omap_gem_resume,
-};
-#endif
-
-struct platform_driver pdev = {
-		.driver = {
-			.name = DRIVER_NAME,
-			.owner = THIS_MODULE,
-#ifdef CONFIG_PM
-			.pm = &omapdrm_pm_ops,
-#endif
-		},
-		.probe = pdev_probe,
-		.remove = pdev_remove,
-		.suspend = pdev_suspend,
-		.resume = pdev_resume,
-		.shutdown = pdev_shutdown,
-};
-
-static int __init omap_drm_init(void)
-{
-	DBG("init");
-	if (platform_driver_register(&omap_dmm_driver)) {
-		/* we can continue on without DMM.. so not fatal */
-		dev_err(NULL, "DMM registration failed\n");
-	}
-	return platform_driver_register(&pdev);
-}
-
-static void __exit omap_drm_fini(void)
-{
-	DBG("fini");
-	platform_driver_unregister(&pdev);
-}
-
-/* need late_initcall() so we load after dss_driver's are loaded */
-late_initcall(omap_drm_init);
-module_exit(omap_drm_fini);
-
-MODULE_AUTHOR("Rob Clark <rob@ti.com>");
-MODULE_DESCRIPTION("OMAP DRM Display Driver");
-MODULE_ALIAS("platform:" DRIVER_NAME);
-MODULE_LICENSE("GPL v2");
diff --git a/drivers/staging/omapdrm/omap_drv.h b/drivers/staging/omapdrm/omap_drv.h
deleted file mode 100644
index f921027..0000000
--- a/drivers/staging/omapdrm/omap_drv.h
+++ /dev/null
@@ -1,333 +0,0 @@
-/*
- * drivers/staging/omapdrm/omap_drv.h
- *
- * Copyright (C) 2011 Texas Instruments
- * Author: Rob Clark <rob@ti.com>
- *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of the GNU General Public License version 2 as published by
- * the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but WITHOUT
- * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
- * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
- * more details.
- *
- * You should have received a copy of the GNU General Public License along with
- * this program.  If not, see <http://www.gnu.org/licenses/>.
- */
-
-#ifndef __OMAP_DRV_H__
-#define __OMAP_DRV_H__
-
-#include <video/omapdss.h>
-#include <linux/module.h>
-#include <linux/types.h>
-#include <drm/drmP.h>
-#include <drm/drm_crtc_helper.h>
-#include <linux/platform_data/omap_drm.h>
-#include "omap_drm.h"
-
-
-#define DBG(fmt, ...) DRM_DEBUG(fmt"\n", ##__VA_ARGS__)
-#define VERB(fmt, ...) if (0) DRM_DEBUG(fmt, ##__VA_ARGS__) /* verbose debug */
-
-#define MODULE_NAME     "omapdrm"
-
-/* max # of mapper-id's that can be assigned.. todo, come up with a better
- * (but still inexpensive) way to store/access per-buffer mapper private
- * data..
- */
-#define MAX_MAPPERS 2
-
-/* parameters which describe (unrotated) coordinates of scanout within a fb: */
-struct omap_drm_window {
-	uint32_t rotation;
-	int32_t  crtc_x, crtc_y;		/* signed because can be offscreen */
-	uint32_t crtc_w, crtc_h;
-	uint32_t src_x, src_y;
-	uint32_t src_w, src_h;
-};
-
-/* Once GO bit is set, we can't make further updates to shadowed registers
- * until the GO bit is cleared.  So various parts in the kms code that need
- * to update shadowed registers queue up a pair of callbacks, pre_apply
- * which is called before setting GO bit, and post_apply that is called
- * after GO bit is cleared.  The crtc manages the queuing, and everyone
- * else goes thru omap_crtc_apply() using these callbacks so that the
- * code which has to deal w/ GO bit state is centralized.
- */
-struct omap_drm_apply {
-	struct list_head pending_node, queued_node;
-	bool queued;
-	void (*pre_apply)(struct omap_drm_apply *apply);
-	void (*post_apply)(struct omap_drm_apply *apply);
-};
-
-/* For transiently registering for different DSS irqs that various parts
- * of the KMS code need during setup/configuration.  We these are not
- * necessarily the same as what drm_vblank_get/put() are requesting, and
- * the hysteresis in drm_vblank_put() is not necessarily desirable for
- * internal housekeeping related irq usage.
- */
-struct omap_drm_irq {
-	struct list_head node;
-	uint32_t irqmask;
-	bool registered;
-	void (*irq)(struct omap_drm_irq *irq, uint32_t irqstatus);
-};
-
-/* For KMS code that needs to wait for a certain # of IRQs:
- */
-struct omap_irq_wait;
-struct omap_irq_wait * omap_irq_wait_init(struct drm_device *dev,
-		uint32_t irqmask, int count);
-int omap_irq_wait(struct drm_device *dev, struct omap_irq_wait *wait,
-		unsigned long timeout);
-
-struct omap_drm_private {
-	uint32_t omaprev;
-
-	unsigned int num_crtcs;
-	struct drm_crtc *crtcs[8];
-
-	unsigned int num_planes;
-	struct drm_plane *planes[8];
-
-	unsigned int num_encoders;
-	struct drm_encoder *encoders[8];
-
-	unsigned int num_connectors;
-	struct drm_connector *connectors[8];
-
-	struct drm_fb_helper *fbdev;
-
-	struct workqueue_struct *wq;
-
-	/* list of GEM objects: */
-	struct list_head obj_list;
-
-	bool has_dmm;
-
-	/* properties: */
-	struct drm_property *rotation_prop;
-	struct drm_property *zorder_prop;
-
-	/* irq handling: */
-	struct list_head irq_list;    /* list of omap_drm_irq */
-	uint32_t vblank_mask;         /* irq bits set for userspace vblank */
-	struct omap_drm_irq error_handler;
-};
-
-/* this should probably be in drm-core to standardize amongst drivers */
-#define DRM_ROTATE_0	0
-#define DRM_ROTATE_90	1
-#define DRM_ROTATE_180	2
-#define DRM_ROTATE_270	3
-#define DRM_REFLECT_X	4
-#define DRM_REFLECT_Y	5
-
-#ifdef CONFIG_DEBUG_FS
-int omap_debugfs_init(struct drm_minor *minor);
-void omap_debugfs_cleanup(struct drm_minor *minor);
-void omap_framebuffer_describe(struct drm_framebuffer *fb, struct seq_file *m);
-void omap_gem_describe(struct drm_gem_object *obj, struct seq_file *m);
-void omap_gem_describe_objects(struct list_head *list, struct seq_file *m);
-#endif
-
-#ifdef CONFIG_PM
-int omap_gem_resume(struct device *dev);
-#endif
-
-int omap_irq_enable_vblank(struct drm_device *dev, int crtc);
-void omap_irq_disable_vblank(struct drm_device *dev, int crtc);
-irqreturn_t omap_irq_handler(DRM_IRQ_ARGS);
-void omap_irq_preinstall(struct drm_device *dev);
-int omap_irq_postinstall(struct drm_device *dev);
-void omap_irq_uninstall(struct drm_device *dev);
-void omap_irq_register(struct drm_device *dev, struct omap_drm_irq *irq);
-void omap_irq_unregister(struct drm_device *dev, struct omap_drm_irq *irq);
-int omap_drm_irq_uninstall(struct drm_device *dev);
-int omap_drm_irq_install(struct drm_device *dev);
-
-struct drm_fb_helper *omap_fbdev_init(struct drm_device *dev);
-void omap_fbdev_free(struct drm_device *dev);
-
-const struct omap_video_timings *omap_crtc_timings(struct drm_crtc *crtc);
-enum omap_channel omap_crtc_channel(struct drm_crtc *crtc);
-int omap_crtc_apply(struct drm_crtc *crtc,
-		struct omap_drm_apply *apply);
-struct drm_crtc *omap_crtc_init(struct drm_device *dev,
-		struct drm_plane *plane, enum omap_channel channel, int id);
-
-struct drm_plane *omap_plane_init(struct drm_device *dev,
-		int plane_id, bool private_plane);
-int omap_plane_dpms(struct drm_plane *plane, int mode);
-int omap_plane_mode_set(struct drm_plane *plane,
-		struct drm_crtc *crtc, struct drm_framebuffer *fb,
-		int crtc_x, int crtc_y,
-		unsigned int crtc_w, unsigned int crtc_h,
-		uint32_t src_x, uint32_t src_y,
-		uint32_t src_w, uint32_t src_h,
-		void (*fxn)(void *), void *arg);
-void omap_plane_install_properties(struct drm_plane *plane,
-		struct drm_mode_object *obj);
-int omap_plane_set_property(struct drm_plane *plane,
-		struct drm_property *property, uint64_t val);
-
-struct drm_encoder *omap_encoder_init(struct drm_device *dev,
-		struct omap_dss_device *dssdev);
-int omap_encoder_set_enabled(struct drm_encoder *encoder, bool enabled);
-int omap_encoder_update(struct drm_encoder *encoder,
-		struct omap_overlay_manager *mgr,
-		struct omap_video_timings *timings);
-
-struct drm_connector *omap_connector_init(struct drm_device *dev,
-		int connector_type, struct omap_dss_device *dssdev,
-		struct drm_encoder *encoder);
-struct drm_encoder *omap_connector_attached_encoder(
-		struct drm_connector *connector);
-void omap_connector_flush(struct drm_connector *connector,
-		int x, int y, int w, int h);
-
-void copy_timings_omap_to_drm(struct drm_display_mode *mode,
-		struct omap_video_timings *timings);
-void copy_timings_drm_to_omap(struct omap_video_timings *timings,
-		struct drm_display_mode *mode);
-
-uint32_t omap_framebuffer_get_formats(uint32_t *pixel_formats,
-		uint32_t max_formats, enum omap_color_mode supported_modes);
-struct drm_framebuffer *omap_framebuffer_create(struct drm_device *dev,
-		struct drm_file *file, struct drm_mode_fb_cmd2 *mode_cmd);
-struct drm_framebuffer *omap_framebuffer_init(struct drm_device *dev,
-		struct drm_mode_fb_cmd2 *mode_cmd, struct drm_gem_object **bos);
-struct drm_gem_object *omap_framebuffer_bo(struct drm_framebuffer *fb, int p);
-int omap_framebuffer_replace(struct drm_framebuffer *a,
-		struct drm_framebuffer *b, void *arg,
-		void (*unpin)(void *arg, struct drm_gem_object *bo));
-void omap_framebuffer_update_scanout(struct drm_framebuffer *fb,
-		struct omap_drm_window *win, struct omap_overlay_info *info);
-struct drm_connector *omap_framebuffer_get_next_connector(
-		struct drm_framebuffer *fb, struct drm_connector *from);
-void omap_framebuffer_flush(struct drm_framebuffer *fb,
-		int x, int y, int w, int h);
-
-void omap_gem_init(struct drm_device *dev);
-void omap_gem_deinit(struct drm_device *dev);
-
-struct drm_gem_object *omap_gem_new(struct drm_device *dev,
-		union omap_gem_size gsize, uint32_t flags);
-int omap_gem_new_handle(struct drm_device *dev, struct drm_file *file,
-		union omap_gem_size gsize, uint32_t flags, uint32_t *handle);
-void omap_gem_free_object(struct drm_gem_object *obj);
-int omap_gem_init_object(struct drm_gem_object *obj);
-void *omap_gem_vaddr(struct drm_gem_object *obj);
-int omap_gem_dumb_map_offset(struct drm_file *file, struct drm_device *dev,
-		uint32_t handle, uint64_t *offset);
-int omap_gem_dumb_destroy(struct drm_file *file, struct drm_device *dev,
-		uint32_t handle);
-int omap_gem_dumb_create(struct drm_file *file, struct drm_device *dev,
-		struct drm_mode_create_dumb *args);
-int omap_gem_mmap(struct file *filp, struct vm_area_struct *vma);
-int omap_gem_mmap_obj(struct drm_gem_object *obj,
-		struct vm_area_struct *vma);
-int omap_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf);
-int omap_gem_op_start(struct drm_gem_object *obj, enum omap_gem_op op);
-int omap_gem_op_finish(struct drm_gem_object *obj, enum omap_gem_op op);
-int omap_gem_op_sync(struct drm_gem_object *obj, enum omap_gem_op op);
-int omap_gem_op_async(struct drm_gem_object *obj, enum omap_gem_op op,
-		void (*fxn)(void *arg), void *arg);
-int omap_gem_roll(struct drm_gem_object *obj, uint32_t roll);
-void omap_gem_cpu_sync(struct drm_gem_object *obj, int pgoff);
-void omap_gem_dma_sync(struct drm_gem_object *obj,
-		enum dma_data_direction dir);
-int omap_gem_get_paddr(struct drm_gem_object *obj,
-		dma_addr_t *paddr, bool remap);
-int omap_gem_put_paddr(struct drm_gem_object *obj);
-int omap_gem_get_pages(struct drm_gem_object *obj, struct page ***pages,
-		bool remap);
-int omap_gem_put_pages(struct drm_gem_object *obj);
-uint32_t omap_gem_flags(struct drm_gem_object *obj);
-int omap_gem_rotated_paddr(struct drm_gem_object *obj, uint32_t orient,
-		int x, int y, dma_addr_t *paddr);
-uint64_t omap_gem_mmap_offset(struct drm_gem_object *obj);
-size_t omap_gem_mmap_size(struct drm_gem_object *obj);
-int omap_gem_tiled_size(struct drm_gem_object *obj, uint16_t *w, uint16_t *h);
-int omap_gem_tiled_stride(struct drm_gem_object *obj, uint32_t orient);
-
-struct dma_buf *omap_gem_prime_export(struct drm_device *dev,
-		struct drm_gem_object *obj, int flags);
-struct drm_gem_object *omap_gem_prime_import(struct drm_device *dev,
-		struct dma_buf *buffer);
-
-static inline int align_pitch(int pitch, int width, int bpp)
-{
-	int bytespp = (bpp + 7) / 8;
-	/* in case someone tries to feed us a completely bogus stride: */
-	pitch = max(pitch, width * bytespp);
-	/* PVR needs alignment to 8 pixels.. right now that is the most
-	 * restrictive stride requirement..
-	 */
-	return ALIGN(pitch, 8 * bytespp);
-}
-
-static inline enum omap_channel pipe2chan(int pipe)
-{
-	int num_mgrs = dss_feat_get_num_mgrs();
-
-	/*
-	 * We usually don't want to create a CRTC for each manager,
-	 * at least not until we have a way to expose private planes
-	 * to userspace.  Otherwise there would not be enough video
-	 * pipes left for drm planes.  The higher #'d managers tend
-	 * to have more features so start in reverse order.
-	 */
-	return num_mgrs - pipe - 1;
-}
-
-/* map crtc to vblank mask */
-static inline uint32_t pipe2vbl(int crtc)
-{
-	enum omap_channel channel = pipe2chan(crtc);
-	return dispc_mgr_get_vsync_irq(channel);
-}
-
-static inline int crtc2pipe(struct drm_device *dev, struct drm_crtc *crtc)
-{
-	struct omap_drm_private *priv = dev->dev_private;
-	int i;
-
-	for (i = 0; i < ARRAY_SIZE(priv->crtcs); i++)
-		if (priv->crtcs[i] == crtc)
-			return i;
-
-	BUG();  /* bogus CRTC ptr */
-	return -1;
-}
-
-/* should these be made into common util helpers?
- */
-
-static inline int objects_lookup(struct drm_device *dev,
-		struct drm_file *filp, uint32_t pixel_format,
-		struct drm_gem_object **bos, uint32_t *handles)
-{
-	int i, n = drm_format_num_planes(pixel_format);
-
-	for (i = 0; i < n; i++) {
-		bos[i] = drm_gem_object_lookup(dev, filp, handles[i]);
-		if (!bos[i])
-			goto fail;
-
-	}
-
-	return 0;
-
-fail:
-	while (--i > 0)
-		drm_gem_object_unreference_unlocked(bos[i]);
-
-	return -ENOENT;
-}
-
-#endif /* __OMAP_DRV_H__ */
diff --git a/drivers/staging/omapdrm/omap_encoder.c b/drivers/staging/omapdrm/omap_encoder.c
deleted file mode 100644
index e053160..0000000
--- a/drivers/staging/omapdrm/omap_encoder.c
+++ /dev/null
@@ -1,170 +0,0 @@
-/*
- * drivers/staging/omapdrm/omap_encoder.c
- *
- * Copyright (C) 2011 Texas Instruments
- * Author: Rob Clark <rob@ti.com>
- *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of the GNU General Public License version 2 as published by
- * the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but WITHOUT
- * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
- * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
- * more details.
- *
- * You should have received a copy of the GNU General Public License along with
- * this program.  If not, see <http://www.gnu.org/licenses/>.
- */
-
-#include "omap_drv.h"
-
-#include "drm_crtc.h"
-#include "drm_crtc_helper.h"
-
-#include <linux/list.h>
-
-
-/*
- * encoder funcs
- */
-
-#define to_omap_encoder(x) container_of(x, struct omap_encoder, base)
-
-/* The encoder and connector both map to same dssdev.. the encoder
- * handles the 'active' parts, ie. anything the modifies the state
- * of the hw, and the connector handles the 'read-only' parts, like
- * detecting connection and reading edid.
- */
-struct omap_encoder {
-	struct drm_encoder base;
-	struct omap_dss_device *dssdev;
-};
-
-static void omap_encoder_destroy(struct drm_encoder *encoder)
-{
-	struct omap_encoder *omap_encoder = to_omap_encoder(encoder);
-	drm_encoder_cleanup(encoder);
-	kfree(omap_encoder);
-}
-
-static const struct drm_encoder_funcs omap_encoder_funcs = {
-	.destroy = omap_encoder_destroy,
-};
-
-/*
- * The CRTC drm_crtc_helper_set_mode() doesn't really give us the right
- * order.. the easiest way to work around this for now is to make all
- * the encoder-helper's no-op's and have the omap_crtc code take care
- * of the sequencing and call us in the right points.
- *
- * Eventually to handle connecting CRTCs to different encoders properly,
- * either the CRTC helpers need to change or we need to replace
- * drm_crtc_helper_set_mode(), but lets wait until atomic-modeset for
- * that.
- */
-
-static void omap_encoder_dpms(struct drm_encoder *encoder, int mode)
-{
-}
-
-static bool omap_encoder_mode_fixup(struct drm_encoder *encoder,
-				  const struct drm_display_mode *mode,
-				  struct drm_display_mode *adjusted_mode)
-{
-	return true;
-}
-
-static void omap_encoder_mode_set(struct drm_encoder *encoder,
-				struct drm_display_mode *mode,
-				struct drm_display_mode *adjusted_mode)
-{
-}
-
-static void omap_encoder_prepare(struct drm_encoder *encoder)
-{
-}
-
-static void omap_encoder_commit(struct drm_encoder *encoder)
-{
-}
-
-static const struct drm_encoder_helper_funcs omap_encoder_helper_funcs = {
-	.dpms = omap_encoder_dpms,
-	.mode_fixup = omap_encoder_mode_fixup,
-	.mode_set = omap_encoder_mode_set,
-	.prepare = omap_encoder_prepare,
-	.commit = omap_encoder_commit,
-};
-
-/*
- * Instead of relying on the helpers for modeset, the omap_crtc code
- * calls these functions in the proper sequence.
- */
-
-int omap_encoder_set_enabled(struct drm_encoder *encoder, bool enabled)
-{
-	struct omap_encoder *omap_encoder = to_omap_encoder(encoder);
-	struct omap_dss_device *dssdev = omap_encoder->dssdev;
-	struct omap_dss_driver *dssdrv = dssdev->driver;
-
-	if (enabled) {
-		return dssdrv->enable(dssdev);
-	} else {
-		dssdrv->disable(dssdev);
-		return 0;
-	}
-}
-
-int omap_encoder_update(struct drm_encoder *encoder,
-		struct omap_overlay_manager *mgr,
-		struct omap_video_timings *timings)
-{
-	struct drm_device *dev = encoder->dev;
-	struct omap_encoder *omap_encoder = to_omap_encoder(encoder);
-	struct omap_dss_device *dssdev = omap_encoder->dssdev;
-	struct omap_dss_driver *dssdrv = dssdev->driver;
-	int ret;
-
-	dssdev->output->manager = mgr;
-
-	ret = dssdrv->check_timings(dssdev, timings);
-	if (ret) {
-		dev_err(dev->dev, "could not set timings: %d\n", ret);
-		return ret;
-	}
-
-	dssdrv->set_timings(dssdev, timings);
-
-	return 0;
-}
-
-/* initialize encoder */
-struct drm_encoder *omap_encoder_init(struct drm_device *dev,
-		struct omap_dss_device *dssdev)
-{
-	struct drm_encoder *encoder = NULL;
-	struct omap_encoder *omap_encoder;
-
-	omap_encoder = kzalloc(sizeof(*omap_encoder), GFP_KERNEL);
-	if (!omap_encoder) {
-		dev_err(dev->dev, "could not allocate encoder\n");
-		goto fail;
-	}
-
-	omap_encoder->dssdev = dssdev;
-
-	encoder = &omap_encoder->base;
-
-	drm_encoder_init(dev, encoder, &omap_encoder_funcs,
-			 DRM_MODE_ENCODER_TMDS);
-	drm_encoder_helper_add(encoder, &omap_encoder_helper_funcs);
-
-	return encoder;
-
-fail:
-	if (encoder)
-		omap_encoder_destroy(encoder);
-
-	return NULL;
-}
diff --git a/drivers/staging/omapdrm/omap_fb.c b/drivers/staging/omapdrm/omap_fb.c
deleted file mode 100644
index bf6421f..0000000
--- a/drivers/staging/omapdrm/omap_fb.c
+++ /dev/null
@@ -1,472 +0,0 @@
-/*
- * drivers/staging/omapdrm/omap_fb.c
- *
- * Copyright (C) 2011 Texas Instruments
- * Author: Rob Clark <rob@ti.com>
- *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of the GNU General Public License version 2 as published by
- * the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but WITHOUT
- * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
- * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
- * more details.
- *
- * You should have received a copy of the GNU General Public License along with
- * this program.  If not, see <http://www.gnu.org/licenses/>.
- */
-
-#include "omap_drv.h"
-#include "omap_dmm_tiler.h"
-
-#include "drm_crtc.h"
-#include "drm_crtc_helper.h"
-
-/*
- * framebuffer funcs
- */
-
-/* per-format info: */
-struct format {
-	enum omap_color_mode dss_format;
-	uint32_t pixel_format;
-	struct {
-		int stride_bpp;           /* this times width is stride */
-		int sub_y;                /* sub-sample in y dimension */
-	} planes[4];
-	bool yuv;
-};
-
-static const struct format formats[] = {
-	/* 16bpp [A]RGB: */
-	{ OMAP_DSS_COLOR_RGB16,       DRM_FORMAT_RGB565,   {{2, 1}}, false }, /* RGB16-565 */
-	{ OMAP_DSS_COLOR_RGB12U,      DRM_FORMAT_RGBX4444, {{2, 1}}, false }, /* RGB12x-4444 */
-	{ OMAP_DSS_COLOR_RGBX16,      DRM_FORMAT_XRGB4444, {{2, 1}}, false }, /* xRGB12-4444 */
-	{ OMAP_DSS_COLOR_RGBA16,      DRM_FORMAT_RGBA4444, {{2, 1}}, false }, /* RGBA12-4444 */
-	{ OMAP_DSS_COLOR_ARGB16,      DRM_FORMAT_ARGB4444, {{2, 1}}, false }, /* ARGB16-4444 */
-	{ OMAP_DSS_COLOR_XRGB16_1555, DRM_FORMAT_XRGB1555, {{2, 1}}, false }, /* xRGB15-1555 */
-	{ OMAP_DSS_COLOR_ARGB16_1555, DRM_FORMAT_ARGB1555, {{2, 1}}, false }, /* ARGB16-1555 */
-	/* 24bpp RGB: */
-	{ OMAP_DSS_COLOR_RGB24P,      DRM_FORMAT_RGB888,   {{3, 1}}, false }, /* RGB24-888 */
-	/* 32bpp [A]RGB: */
-	{ OMAP_DSS_COLOR_RGBX32,      DRM_FORMAT_RGBX8888, {{4, 1}}, false }, /* RGBx24-8888 */
-	{ OMAP_DSS_COLOR_RGB24U,      DRM_FORMAT_XRGB8888, {{4, 1}}, false }, /* xRGB24-8888 */
-	{ OMAP_DSS_COLOR_RGBA32,      DRM_FORMAT_RGBA8888, {{4, 1}}, false }, /* RGBA32-8888 */
-	{ OMAP_DSS_COLOR_ARGB32,      DRM_FORMAT_ARGB8888, {{4, 1}}, false }, /* ARGB32-8888 */
-	/* YUV: */
-	{ OMAP_DSS_COLOR_NV12,        DRM_FORMAT_NV12,     {{1, 1}, {1, 2}}, true },
-	{ OMAP_DSS_COLOR_YUV2,        DRM_FORMAT_YUYV,     {{2, 1}}, true },
-	{ OMAP_DSS_COLOR_UYVY,        DRM_FORMAT_UYVY,     {{2, 1}}, true },
-};
-
-/* convert from overlay's pixel formats bitmask to an array of fourcc's */
-uint32_t omap_framebuffer_get_formats(uint32_t *pixel_formats,
-		uint32_t max_formats, enum omap_color_mode supported_modes)
-{
-	uint32_t nformats = 0;
-	int i = 0;
-
-	for (i = 0; i < ARRAY_SIZE(formats) && nformats < max_formats; i++)
-		if (formats[i].dss_format & supported_modes)
-			pixel_formats[nformats++] = formats[i].pixel_format;
-
-	return nformats;
-}
-
-/* per-plane info for the fb: */
-struct plane {
-	struct drm_gem_object *bo;
-	uint32_t pitch;
-	uint32_t offset;
-	dma_addr_t paddr;
-};
-
-#define to_omap_framebuffer(x) container_of(x, struct omap_framebuffer, base)
-
-struct omap_framebuffer {
-	struct drm_framebuffer base;
-	const struct format *format;
-	struct plane planes[4];
-};
-
-static int omap_framebuffer_create_handle(struct drm_framebuffer *fb,
-		struct drm_file *file_priv,
-		unsigned int *handle)
-{
-	struct omap_framebuffer *omap_fb = to_omap_framebuffer(fb);
-	return drm_gem_handle_create(file_priv,
-			omap_fb->planes[0].bo, handle);
-}
-
-static void omap_framebuffer_destroy(struct drm_framebuffer *fb)
-{
-	struct omap_framebuffer *omap_fb = to_omap_framebuffer(fb);
-	int i, n = drm_format_num_planes(fb->pixel_format);
-
-	DBG("destroy: FB ID: %d (%p)", fb->base.id, fb);
-
-	drm_framebuffer_cleanup(fb);
-
-	for (i = 0; i < n; i++) {
-		struct plane *plane = &omap_fb->planes[i];
-		if (plane->bo)
-			drm_gem_object_unreference_unlocked(plane->bo);
-	}
-
-	kfree(omap_fb);
-}
-
-static int omap_framebuffer_dirty(struct drm_framebuffer *fb,
-		struct drm_file *file_priv, unsigned flags, unsigned color,
-		struct drm_clip_rect *clips, unsigned num_clips)
-{
-	int i;
-
-	for (i = 0; i < num_clips; i++) {
-		omap_framebuffer_flush(fb, clips[i].x1, clips[i].y1,
-					clips[i].x2 - clips[i].x1,
-					clips[i].y2 - clips[i].y1);
-	}
-
-	return 0;
-}
-
-static const struct drm_framebuffer_funcs omap_framebuffer_funcs = {
-	.create_handle = omap_framebuffer_create_handle,
-	.destroy = omap_framebuffer_destroy,
-	.dirty = omap_framebuffer_dirty,
-};
-
-static uint32_t get_linear_addr(struct plane *plane,
-		const struct format *format, int n, int x, int y)
-{
-	uint32_t offset;
-
-	offset = plane->offset +
-			(x * format->planes[n].stride_bpp) +
-			(y * plane->pitch / format->planes[n].sub_y);
-
-	return plane->paddr + offset;
-}
-
-/* update ovl info for scanout, handles cases of multi-planar fb's, etc.
- */
-void omap_framebuffer_update_scanout(struct drm_framebuffer *fb,
-		struct omap_drm_window *win, struct omap_overlay_info *info)
-{
-	struct omap_framebuffer *omap_fb = to_omap_framebuffer(fb);
-	const struct format *format = omap_fb->format;
-	struct plane *plane = &omap_fb->planes[0];
-	uint32_t x, y, orient = 0;
-
-	info->color_mode = format->dss_format;
-
-	info->pos_x      = win->crtc_x;
-	info->pos_y      = win->crtc_y;
-	info->out_width  = win->crtc_w;
-	info->out_height = win->crtc_h;
-	info->width      = win->src_w;
-	info->height     = win->src_h;
-
-	x = win->src_x;
-	y = win->src_y;
-
-	if (omap_gem_flags(plane->bo) & OMAP_BO_TILED) {
-		uint32_t w = win->src_w;
-		uint32_t h = win->src_h;
-
-		switch (win->rotation & 0xf) {
-		default:
-			dev_err(fb->dev->dev, "invalid rotation: %02x",
-					(uint32_t)win->rotation);
-			/* fallthru to default to no rotation */
-		case 0:
-		case BIT(DRM_ROTATE_0):
-			orient = 0;
-			break;
-		case BIT(DRM_ROTATE_90):
-			orient = MASK_XY_FLIP | MASK_X_INVERT;
-			break;
-		case BIT(DRM_ROTATE_180):
-			orient = MASK_X_INVERT | MASK_Y_INVERT;
-			break;
-		case BIT(DRM_ROTATE_270):
-			orient = MASK_XY_FLIP | MASK_Y_INVERT;
-			break;
-		}
-
-		if (win->rotation & BIT(DRM_REFLECT_X))
-			orient ^= MASK_X_INVERT;
-
-		if (win->rotation & BIT(DRM_REFLECT_Y))
-			orient ^= MASK_Y_INVERT;
-
-		/* adjust x,y offset for flip/invert: */
-		if (orient & MASK_XY_FLIP)
-			swap(w, h);
-		if (orient & MASK_Y_INVERT)
-			y += h - 1;
-		if (orient & MASK_X_INVERT)
-			x += w - 1;
-
-		omap_gem_rotated_paddr(plane->bo, orient, x, y, &info->paddr);
-		info->rotation_type = OMAP_DSS_ROT_TILER;
-		info->screen_width  = omap_gem_tiled_stride(plane->bo, orient);
-	} else {
-		info->paddr         = get_linear_addr(plane, format, 0, x, y);
-		info->rotation_type = OMAP_DSS_ROT_DMA;
-		info->screen_width  = plane->pitch;
-	}
-
-	/* convert to pixels: */
-	info->screen_width /= format->planes[0].stride_bpp;
-
-	if (format->dss_format == OMAP_DSS_COLOR_NV12) {
-		plane = &omap_fb->planes[1];
-
-		if (info->rotation_type == OMAP_DSS_ROT_TILER) {
-			WARN_ON(!(omap_gem_flags(plane->bo) & OMAP_BO_TILED));
-			omap_gem_rotated_paddr(plane->bo, orient,
-					x/2, y/2, &info->p_uv_addr);
-		} else {
-			info->p_uv_addr = get_linear_addr(plane, format, 1, x, y);
-		}
-	} else {
-		info->p_uv_addr = 0;
-	}
-}
-
-/* Call for unpin 'a' (if not NULL), and pin 'b' (if not NULL).  Although
- * buffers to unpin are just pushed to the unpin fifo so that the
- * caller can defer unpin until vblank.
- *
- * Note if this fails (ie. something went very wrong!), all buffers are
- * unpinned, and the caller disables the overlay.  We could have tried
- * to revert back to the previous set of pinned buffers but if things are
- * hosed there is no guarantee that would succeed.
- */
-int omap_framebuffer_replace(struct drm_framebuffer *a,
-		struct drm_framebuffer *b, void *arg,
-		void (*unpin)(void *arg, struct drm_gem_object *bo))
-{
-	int ret = 0, i, na, nb;
-	struct omap_framebuffer *ofba = to_omap_framebuffer(a);
-	struct omap_framebuffer *ofbb = to_omap_framebuffer(b);
-	uint32_t pinned_mask = 0;
-
-	na = a ? drm_format_num_planes(a->pixel_format) : 0;
-	nb = b ? drm_format_num_planes(b->pixel_format) : 0;
-
-	for (i = 0; i < max(na, nb); i++) {
-		struct plane *pa, *pb;
-
-		pa = (i < na) ? &ofba->planes[i] : NULL;
-		pb = (i < nb) ? &ofbb->planes[i] : NULL;
-
-		if (pa)
-			unpin(arg, pa->bo);
-
-		if (pb && !ret) {
-			ret = omap_gem_get_paddr(pb->bo, &pb->paddr, true);
-			if (!ret) {
-				omap_gem_dma_sync(pb->bo, DMA_TO_DEVICE);
-				pinned_mask |= (1 << i);
-			}
-		}
-	}
-
-	if (ret) {
-		/* something went wrong.. unpin what has been pinned */
-		for (i = 0; i < nb; i++) {
-			if (pinned_mask & (1 << i)) {
-				struct plane *pb = &ofba->planes[i];
-				unpin(arg, pb->bo);
-			}
-		}
-	}
-
-	return ret;
-}
-
-struct drm_gem_object *omap_framebuffer_bo(struct drm_framebuffer *fb, int p)
-{
-	struct omap_framebuffer *omap_fb = to_omap_framebuffer(fb);
-	if (p >= drm_format_num_planes(fb->pixel_format))
-		return NULL;
-	return omap_fb->planes[p].bo;
-}
-
-/* iterate thru all the connectors, returning ones that are attached
- * to the same fb..
- */
-struct drm_connector *omap_framebuffer_get_next_connector(
-		struct drm_framebuffer *fb, struct drm_connector *from)
-{
-	struct drm_device *dev = fb->dev;
-	struct list_head *connector_list = &dev->mode_config.connector_list;
-	struct drm_connector *connector = from;
-
-	if (!from)
-		return list_first_entry(connector_list, typeof(*from), head);
-
-	list_for_each_entry_from(connector, connector_list, head) {
-		if (connector != from) {
-			struct drm_encoder *encoder = connector->encoder;
-			struct drm_crtc *crtc = encoder ? encoder->crtc : NULL;
-			if (crtc && crtc->fb == fb)
-				return connector;
-
-		}
-	}
-
-	return NULL;
-}
-
-/* flush an area of the framebuffer (in case of manual update display that
- * is not automatically flushed)
- */
-void omap_framebuffer_flush(struct drm_framebuffer *fb,
-		int x, int y, int w, int h)
-{
-	struct drm_connector *connector = NULL;
-
-	VERB("flush: %d,%d %dx%d, fb=%p", x, y, w, h, fb);
-
-	while ((connector = omap_framebuffer_get_next_connector(fb, connector))) {
-		/* only consider connectors that are part of a chain */
-		if (connector->encoder && connector->encoder->crtc) {
-			/* TODO: maybe this should propagate thru the crtc who
-			 * could do the coordinate translation..
-			 */
-			struct drm_crtc *crtc = connector->encoder->crtc;
-			int cx = max(0, x - crtc->x);
-			int cy = max(0, y - crtc->y);
-			int cw = w + (x - crtc->x) - cx;
-			int ch = h + (y - crtc->y) - cy;
-
-			omap_connector_flush(connector, cx, cy, cw, ch);
-		}
-	}
-}
-
-#ifdef CONFIG_DEBUG_FS
-void omap_framebuffer_describe(struct drm_framebuffer *fb, struct seq_file *m)
-{
-	struct omap_framebuffer *omap_fb = to_omap_framebuffer(fb);
-	int i, n = drm_format_num_planes(fb->pixel_format);
-
-	seq_printf(m, "fb: %dx%d@%4.4s\n", fb->width, fb->height,
-			(char *)&fb->pixel_format);
-
-	for (i = 0; i < n; i++) {
-		struct plane *plane = &omap_fb->planes[i];
-		seq_printf(m, "   %d: offset=%d pitch=%d, obj: ",
-				i, plane->offset, plane->pitch);
-		omap_gem_describe(plane->bo, m);
-	}
-}
-#endif
-
-struct drm_framebuffer *omap_framebuffer_create(struct drm_device *dev,
-		struct drm_file *file, struct drm_mode_fb_cmd2 *mode_cmd)
-{
-	struct drm_gem_object *bos[4];
-	struct drm_framebuffer *fb;
-	int ret;
-
-	ret = objects_lookup(dev, file, mode_cmd->pixel_format,
-			bos, mode_cmd->handles);
-	if (ret)
-		return ERR_PTR(ret);
-
-	fb = omap_framebuffer_init(dev, mode_cmd, bos);
-	if (IS_ERR(fb)) {
-		int i, n = drm_format_num_planes(mode_cmd->pixel_format);
-		for (i = 0; i < n; i++)
-			drm_gem_object_unreference_unlocked(bos[i]);
-		return fb;
-	}
-	return fb;
-}
-
-struct drm_framebuffer *omap_framebuffer_init(struct drm_device *dev,
-		struct drm_mode_fb_cmd2 *mode_cmd, struct drm_gem_object **bos)
-{
-	struct omap_framebuffer *omap_fb;
-	struct drm_framebuffer *fb = NULL;
-	const struct format *format = NULL;
-	int ret, i, n = drm_format_num_planes(mode_cmd->pixel_format);
-
-	DBG("create framebuffer: dev=%p, mode_cmd=%p (%dx%d@%4.4s)",
-			dev, mode_cmd, mode_cmd->width, mode_cmd->height,
-			(char *)&mode_cmd->pixel_format);
-
-	for (i = 0; i < ARRAY_SIZE(formats); i++) {
-		if (formats[i].pixel_format == mode_cmd->pixel_format) {
-			format = &formats[i];
-			break;
-		}
-	}
-
-	if (!format) {
-		dev_err(dev->dev, "unsupported pixel format: %4.4s\n",
-				(char *)&mode_cmd->pixel_format);
-		ret = -EINVAL;
-		goto fail;
-	}
-
-	omap_fb = kzalloc(sizeof(*omap_fb), GFP_KERNEL);
-	if (!omap_fb) {
-		dev_err(dev->dev, "could not allocate fb\n");
-		ret = -ENOMEM;
-		goto fail;
-	}
-
-	fb = &omap_fb->base;
-	omap_fb->format = format;
-
-	for (i = 0; i < n; i++) {
-		struct plane *plane = &omap_fb->planes[i];
-		int size, pitch = mode_cmd->pitches[i];
-
-		if (pitch < (mode_cmd->width * format->planes[i].stride_bpp)) {
-			dev_err(dev->dev, "provided buffer pitch is too small! %d < %d\n",
-					pitch, mode_cmd->width * format->planes[i].stride_bpp);
-			ret = -EINVAL;
-			goto fail;
-		}
-
-		size = pitch * mode_cmd->height / format->planes[i].sub_y;
-
-		if (size > (omap_gem_mmap_size(bos[i]) - mode_cmd->offsets[i])) {
-			dev_err(dev->dev, "provided buffer object is too small! %d < %d\n",
-					bos[i]->size - mode_cmd->offsets[i], size);
-			ret = -EINVAL;
-			goto fail;
-		}
-
-		plane->bo     = bos[i];
-		plane->offset = mode_cmd->offsets[i];
-		plane->pitch  = pitch;
-		plane->paddr  = 0;
-	}
-
-	drm_helper_mode_fill_fb_struct(fb, mode_cmd);
-
-	ret = drm_framebuffer_init(dev, fb, &omap_framebuffer_funcs);
-	if (ret) {
-		dev_err(dev->dev, "framebuffer init failed: %d\n", ret);
-		goto fail;
-	}
-
-	DBG("create: FB ID: %d (%p)", fb->base.id, fb);
-
-	return fb;
-
-fail:
-	if (fb)
-		omap_framebuffer_destroy(fb);
-
-	return ERR_PTR(ret);
-}
diff --git a/drivers/staging/omapdrm/omap_fbdev.c b/drivers/staging/omapdrm/omap_fbdev.c
deleted file mode 100644
index caefdf9..0000000
--- a/drivers/staging/omapdrm/omap_fbdev.c
+++ /dev/null
@@ -1,402 +0,0 @@
-/*
- * drivers/staging/omapdrm/omap_fbdev.c
- *
- * Copyright (C) 2011 Texas Instruments
- * Author: Rob Clark <rob@ti.com>
- *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of the GNU General Public License version 2 as published by
- * the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but WITHOUT
- * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
- * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
- * more details.
- *
- * You should have received a copy of the GNU General Public License along with
- * this program.  If not, see <http://www.gnu.org/licenses/>.
- */
-
-#include "omap_drv.h"
-
-#include "drm_crtc.h"
-#include "drm_fb_helper.h"
-
-MODULE_PARM_DESC(ywrap, "Enable ywrap scrolling (omap44xx and later, default 'y')");
-static bool ywrap_enabled = true;
-module_param_named(ywrap, ywrap_enabled, bool, 0644);
-
-/*
- * fbdev funcs, to implement legacy fbdev interface on top of drm driver
- */
-
-#define to_omap_fbdev(x) container_of(x, struct omap_fbdev, base)
-
-struct omap_fbdev {
-	struct drm_fb_helper base;
-	struct drm_framebuffer *fb;
-	struct drm_gem_object *bo;
-	bool ywrap_enabled;
-
-	/* for deferred dmm roll when getting called in atomic ctx */
-	struct work_struct work;
-};
-
-static void omap_fbdev_flush(struct fb_info *fbi, int x, int y, int w, int h);
-static struct drm_fb_helper *get_fb(struct fb_info *fbi);
-
-static ssize_t omap_fbdev_write(struct fb_info *fbi, const char __user *buf,
-		size_t count, loff_t *ppos)
-{
-	ssize_t res;
-
-	res = fb_sys_write(fbi, buf, count, ppos);
-	omap_fbdev_flush(fbi, 0, 0, fbi->var.xres, fbi->var.yres);
-
-	return res;
-}
-
-static void omap_fbdev_fillrect(struct fb_info *fbi,
-		const struct fb_fillrect *rect)
-{
-	sys_fillrect(fbi, rect);
-	omap_fbdev_flush(fbi, rect->dx, rect->dy, rect->width, rect->height);
-}
-
-static void omap_fbdev_copyarea(struct fb_info *fbi,
-		const struct fb_copyarea *area)
-{
-	sys_copyarea(fbi, area);
-	omap_fbdev_flush(fbi, area->dx, area->dy, area->width, area->height);
-}
-
-static void omap_fbdev_imageblit(struct fb_info *fbi,
-		const struct fb_image *image)
-{
-	sys_imageblit(fbi, image);
-	omap_fbdev_flush(fbi, image->dx, image->dy,
-				image->width, image->height);
-}
-
-static void pan_worker(struct work_struct *work)
-{
-	struct omap_fbdev *fbdev = container_of(work, struct omap_fbdev, work);
-	struct fb_info *fbi = fbdev->base.fbdev;
-	int npages;
-
-	/* DMM roll shifts in 4K pages: */
-	npages = fbi->fix.line_length >> PAGE_SHIFT;
-	omap_gem_roll(fbdev->bo, fbi->var.yoffset * npages);
-}
-
-static int omap_fbdev_pan_display(struct fb_var_screeninfo *var,
-		struct fb_info *fbi)
-{
-	struct drm_fb_helper *helper = get_fb(fbi);
-	struct omap_fbdev *fbdev = to_omap_fbdev(helper);
-
-	if (!helper)
-		goto fallback;
-
-	if (!fbdev->ywrap_enabled)
-		goto fallback;
-
-	if (drm_can_sleep()) {
-		pan_worker(&fbdev->work);
-	} else {
-		struct omap_drm_private *priv = helper->dev->dev_private;
-		queue_work(priv->wq, &fbdev->work);
-	}
-
-	return 0;
-
-fallback:
-	return drm_fb_helper_pan_display(var, fbi);
-}
-
-static struct fb_ops omap_fb_ops = {
-	.owner = THIS_MODULE,
-
-	/* Note: to properly handle manual update displays, we wrap the
-	 * basic fbdev ops which write to the framebuffer
-	 */
-	.fb_read = fb_sys_read,
-	.fb_write = omap_fbdev_write,
-	.fb_fillrect = omap_fbdev_fillrect,
-	.fb_copyarea = omap_fbdev_copyarea,
-	.fb_imageblit = omap_fbdev_imageblit,
-
-	.fb_check_var = drm_fb_helper_check_var,
-	.fb_set_par = drm_fb_helper_set_par,
-	.fb_pan_display = omap_fbdev_pan_display,
-	.fb_blank = drm_fb_helper_blank,
-	.fb_setcmap = drm_fb_helper_setcmap,
-
-	.fb_debug_enter = drm_fb_helper_debug_enter,
-	.fb_debug_leave = drm_fb_helper_debug_leave,
-};
-
-static int omap_fbdev_create(struct drm_fb_helper *helper,
-		struct drm_fb_helper_surface_size *sizes)
-{
-	struct omap_fbdev *fbdev = to_omap_fbdev(helper);
-	struct drm_device *dev = helper->dev;
-	struct omap_drm_private *priv = dev->dev_private;
-	struct drm_framebuffer *fb = NULL;
-	union omap_gem_size gsize;
-	struct fb_info *fbi = NULL;
-	struct drm_mode_fb_cmd2 mode_cmd = {0};
-	dma_addr_t paddr;
-	int ret;
-
-	/* only doing ARGB32 since this is what is needed to alpha-blend
-	 * with video overlays:
-	 */
-	sizes->surface_bpp = 32;
-	sizes->surface_depth = 32;
-
-	DBG("create fbdev: %dx%d@%d (%dx%d)", sizes->surface_width,
-			sizes->surface_height, sizes->surface_bpp,
-			sizes->fb_width, sizes->fb_height);
-
-	mode_cmd.pixel_format = drm_mode_legacy_fb_format(sizes->surface_bpp,
-			sizes->surface_depth);
-
-	mode_cmd.width = sizes->surface_width;
-	mode_cmd.height = sizes->surface_height;
-
-	mode_cmd.pitches[0] = align_pitch(
-			mode_cmd.width * ((sizes->surface_bpp + 7) / 8),
-			mode_cmd.width, sizes->surface_bpp);
-
-	fbdev->ywrap_enabled = priv->has_dmm && ywrap_enabled;
-	if (fbdev->ywrap_enabled) {
-		/* need to align pitch to page size if using DMM scrolling */
-		mode_cmd.pitches[0] = ALIGN(mode_cmd.pitches[0], PAGE_SIZE);
-	}
-
-	/* allocate backing bo */
-	gsize = (union omap_gem_size){
-		.bytes = PAGE_ALIGN(mode_cmd.pitches[0] * mode_cmd.height),
-	};
-	DBG("allocating %d bytes for fb %d", gsize.bytes, dev->primary->index);
-	fbdev->bo = omap_gem_new(dev, gsize, OMAP_BO_SCANOUT | OMAP_BO_WC);
-	if (!fbdev->bo) {
-		dev_err(dev->dev, "failed to allocate buffer object\n");
-		ret = -ENOMEM;
-		goto fail;
-	}
-
-	fb = omap_framebuffer_init(dev, &mode_cmd, &fbdev->bo);
-	if (IS_ERR(fb)) {
-		dev_err(dev->dev, "failed to allocate fb\n");
-		/* note: if fb creation failed, we can't rely on fb destroy
-		 * to unref the bo:
-		 */
-		drm_gem_object_unreference(fbdev->bo);
-		ret = PTR_ERR(fb);
-		goto fail;
-	}
-
-	/* note: this keeps the bo pinned.. which is perhaps not ideal,
-	 * but is needed as long as we use fb_mmap() to mmap to userspace
-	 * (since this happens using fix.smem_start).  Possibly we could
-	 * implement our own mmap using GEM mmap support to avoid this
-	 * (non-tiled buffer doesn't need to be pinned for fbcon to write
-	 * to it).  Then we just need to be sure that we are able to re-
-	 * pin it in case of an opps.
-	 */
-	ret = omap_gem_get_paddr(fbdev->bo, &paddr, true);
-	if (ret) {
-		dev_err(dev->dev,
-			"could not map (paddr)!  Skipping framebuffer alloc\n");
-		ret = -ENOMEM;
-		goto fail;
-	}
-
-	mutex_lock(&dev->struct_mutex);
-
-	fbi = framebuffer_alloc(0, dev->dev);
-	if (!fbi) {
-		dev_err(dev->dev, "failed to allocate fb info\n");
-		ret = -ENOMEM;
-		goto fail_unlock;
-	}
-
-	DBG("fbi=%p, dev=%p", fbi, dev);
-
-	fbdev->fb = fb;
-	helper->fb = fb;
-	helper->fbdev = fbi;
-
-	fbi->par = helper;
-	fbi->flags = FBINFO_DEFAULT;
-	fbi->fbops = &omap_fb_ops;
-
-	strcpy(fbi->fix.id, MODULE_NAME);
-
-	ret = fb_alloc_cmap(&fbi->cmap, 256, 0);
-	if (ret) {
-		ret = -ENOMEM;
-		goto fail_unlock;
-	}
-
-	drm_fb_helper_fill_fix(fbi, fb->pitches[0], fb->depth);
-	drm_fb_helper_fill_var(fbi, helper, sizes->fb_width, sizes->fb_height);
-
-	dev->mode_config.fb_base = paddr;
-
-	fbi->screen_base = omap_gem_vaddr(fbdev->bo);
-	fbi->screen_size = fbdev->bo->size;
-	fbi->fix.smem_start = paddr;
-	fbi->fix.smem_len = fbdev->bo->size;
-
-	/* if we have DMM, then we can use it for scrolling by just
-	 * shuffling pages around in DMM rather than doing sw blit.
-	 */
-	if (fbdev->ywrap_enabled) {
-		DRM_INFO("Enabling DMM ywrap scrolling\n");
-		fbi->flags |= FBINFO_HWACCEL_YWRAP | FBINFO_READS_FAST;
-		fbi->fix.ywrapstep = 1;
-	}
-
-
-	DBG("par=%p, %dx%d", fbi->par, fbi->var.xres, fbi->var.yres);
-	DBG("allocated %dx%d fb", fbdev->fb->width, fbdev->fb->height);
-
-	mutex_unlock(&dev->struct_mutex);
-
-	return 0;
-
-fail_unlock:
-	mutex_unlock(&dev->struct_mutex);
-fail:
-
-	if (ret) {
-		if (fbi)
-			framebuffer_release(fbi);
-		if (fb) {
-			drm_framebuffer_unregister_private(fb);
-			drm_framebuffer_remove(fb);
-		}
-	}
-
-	return ret;
-}
-
-static void omap_crtc_fb_gamma_set(struct drm_crtc *crtc,
-		u16 red, u16 green, u16 blue, int regno)
-{
-	DBG("fbdev: set gamma");
-}
-
-static void omap_crtc_fb_gamma_get(struct drm_crtc *crtc,
-		u16 *red, u16 *green, u16 *blue, int regno)
-{
-	DBG("fbdev: get gamma");
-}
-
-static struct drm_fb_helper_funcs omap_fb_helper_funcs = {
-	.gamma_set = omap_crtc_fb_gamma_set,
-	.gamma_get = omap_crtc_fb_gamma_get,
-	.fb_probe = omap_fbdev_create,
-};
-
-static struct drm_fb_helper *get_fb(struct fb_info *fbi)
-{
-	if (!fbi || strcmp(fbi->fix.id, MODULE_NAME)) {
-		/* these are not the fb's you're looking for */
-		return NULL;
-	}
-	return fbi->par;
-}
-
-/* flush an area of the framebuffer (in case of manual update display that
- * is not automatically flushed)
- */
-static void omap_fbdev_flush(struct fb_info *fbi, int x, int y, int w, int h)
-{
-	struct drm_fb_helper *helper = get_fb(fbi);
-
-	if (!helper)
-		return;
-
-	VERB("flush fbdev: %d,%d %dx%d, fbi=%p", x, y, w, h, fbi);
-
-	omap_framebuffer_flush(helper->fb, x, y, w, h);
-}
-
-/* initialize fbdev helper */
-struct drm_fb_helper *omap_fbdev_init(struct drm_device *dev)
-{
-	struct omap_drm_private *priv = dev->dev_private;
-	struct omap_fbdev *fbdev = NULL;
-	struct drm_fb_helper *helper;
-	int ret = 0;
-
-	fbdev = kzalloc(sizeof(*fbdev), GFP_KERNEL);
-	if (!fbdev) {
-		dev_err(dev->dev, "could not allocate fbdev\n");
-		goto fail;
-	}
-
-	INIT_WORK(&fbdev->work, pan_worker);
-
-	helper = &fbdev->base;
-
-	helper->funcs = &omap_fb_helper_funcs;
-
-	ret = drm_fb_helper_init(dev, helper,
-			priv->num_crtcs, priv->num_connectors);
-	if (ret) {
-		dev_err(dev->dev, "could not init fbdev: ret=%d\n", ret);
-		goto fail;
-	}
-
-	drm_fb_helper_single_add_all_connectors(helper);
-
-	/* disable all the possible outputs/crtcs before entering KMS mode */
-	drm_helper_disable_unused_functions(dev);
-
-	drm_fb_helper_initial_config(helper, 32);
-
-	priv->fbdev = helper;
-
-	return helper;
-
-fail:
-	kfree(fbdev);
-	return NULL;
-}
-
-void omap_fbdev_free(struct drm_device *dev)
-{
-	struct omap_drm_private *priv = dev->dev_private;
-	struct drm_fb_helper *helper = priv->fbdev;
-	struct omap_fbdev *fbdev;
-	struct fb_info *fbi;
-
-	DBG();
-
-	fbi = helper->fbdev;
-
-	/* only cleanup framebuffer if it is present */
-	if (fbi) {
-		unregister_framebuffer(fbi);
-		framebuffer_release(fbi);
-	}
-
-	drm_fb_helper_fini(helper);
-
-	fbdev = to_omap_fbdev(priv->fbdev);
-
-	/* this will free the backing object */
-	if (fbdev->fb) {
-		drm_framebuffer_unregister_private(fbdev->fb);
-		drm_framebuffer_remove(fbdev->fb);
-	}
-
-	kfree(fbdev);
-
-	priv->fbdev = NULL;
-}
diff --git a/drivers/staging/omapdrm/omap_gem.c b/drivers/staging/omapdrm/omap_gem.c
deleted file mode 100644
index f9297eb..0000000
--- a/drivers/staging/omapdrm/omap_gem.c
+++ /dev/null
@@ -1,1511 +0,0 @@
-/*
- * drivers/staging/omapdrm/omap_gem.c
- *
- * Copyright (C) 2011 Texas Instruments
- * Author: Rob Clark <rob.clark@linaro.org>
- *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of the GNU General Public License version 2 as published by
- * the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but WITHOUT
- * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
- * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
- * more details.
- *
- * You should have received a copy of the GNU General Public License along with
- * this program.  If not, see <http://www.gnu.org/licenses/>.
- */
-
-
-#include <linux/spinlock.h>
-#include <linux/shmem_fs.h>
-
-#include "omap_drv.h"
-#include "omap_dmm_tiler.h"
-
-/* remove these once drm core helpers are merged */
-struct page **_drm_gem_get_pages(struct drm_gem_object *obj, gfp_t gfpmask);
-void _drm_gem_put_pages(struct drm_gem_object *obj, struct page **pages,
-		bool dirty, bool accessed);
-int _drm_gem_create_mmap_offset_size(struct drm_gem_object *obj, size_t size);
-
-/*
- * GEM buffer object implementation.
- */
-
-#define to_omap_bo(x) container_of(x, struct omap_gem_object, base)
-
-/* note: we use upper 8 bits of flags for driver-internal flags: */
-#define OMAP_BO_DMA			0x01000000	/* actually is physically contiguous */
-#define OMAP_BO_EXT_SYNC	0x02000000	/* externally allocated sync object */
-#define OMAP_BO_EXT_MEM		0x04000000	/* externally allocated memory */
-
-
-struct omap_gem_object {
-	struct drm_gem_object base;
-
-	struct list_head mm_list;
-
-	uint32_t flags;
-
-	/** width/height for tiled formats (rounded up to slot boundaries) */
-	uint16_t width, height;
-
-	/** roll applied when mapping to DMM */
-	uint32_t roll;
-
-	/**
-	 * If buffer is allocated physically contiguous, the OMAP_BO_DMA flag
-	 * is set and the paddr is valid.  Also if the buffer is remapped in
-	 * TILER and paddr_cnt > 0, then paddr is valid.  But if you are using
-	 * the physical address and OMAP_BO_DMA is not set, then you should
-	 * be going thru omap_gem_{get,put}_paddr() to ensure the mapping is
-	 * not removed from under your feet.
-	 *
-	 * Note that OMAP_BO_SCANOUT is a hint from userspace that DMA capable
-	 * buffer is requested, but doesn't mean that it is.  Use the
-	 * OMAP_BO_DMA flag to determine if the buffer has a DMA capable
-	 * physical address.
-	 */
-	dma_addr_t paddr;
-
-	/**
-	 * # of users of paddr
-	 */
-	uint32_t paddr_cnt;
-
-	/**
-	 * tiler block used when buffer is remapped in DMM/TILER.
-	 */
-	struct tiler_block *block;
-
-	/**
-	 * Array of backing pages, if allocated.  Note that pages are never
-	 * allocated for buffers originally allocated from contiguous memory
-	 */
-	struct page **pages;
-
-	/** addresses corresponding to pages in above array */
-	dma_addr_t *addrs;
-
-	/**
-	 * Virtual address, if mapped.
-	 */
-	void *vaddr;
-
-	/**
-	 * sync-object allocated on demand (if needed)
-	 *
-	 * Per-buffer sync-object for tracking pending and completed hw/dma
-	 * read and write operations.  The layout in memory is dictated by
-	 * the SGX firmware, which uses this information to stall the command
-	 * stream if a surface is not ready yet.
-	 *
-	 * Note that when buffer is used by SGX, the sync-object needs to be
-	 * allocated from a special heap of sync-objects.  This way many sync
-	 * objects can be packed in a page, and not waste GPU virtual address
-	 * space.  Because of this we have to have a omap_gem_set_sync_object()
-	 * API to allow replacement of the syncobj after it has (potentially)
-	 * already been allocated.  A bit ugly but I haven't thought of a
-	 * better alternative.
-	 */
-	struct {
-		uint32_t write_pending;
-		uint32_t write_complete;
-		uint32_t read_pending;
-		uint32_t read_complete;
-	} *sync;
-};
-
-static int get_pages(struct drm_gem_object *obj, struct page ***pages);
-static uint64_t mmap_offset(struct drm_gem_object *obj);
-
-/* To deal with userspace mmap'ings of 2d tiled buffers, which (a) are
- * not necessarily pinned in TILER all the time, and (b) when they are
- * they are not necessarily page aligned, we reserve one or more small
- * regions in each of the 2d containers to use as a user-GART where we
- * can create a second page-aligned mapping of parts of the buffer
- * being accessed from userspace.
- *
- * Note that we could optimize slightly when we know that multiple
- * tiler containers are backed by the same PAT.. but I'll leave that
- * for later..
- */
-#define NUM_USERGART_ENTRIES 2
-struct usergart_entry {
-	struct tiler_block *block;	/* the reserved tiler block */
-	dma_addr_t paddr;
-	struct drm_gem_object *obj;	/* the current pinned obj */
-	pgoff_t obj_pgoff;		/* page offset of obj currently
-					   mapped in */
-};
-static struct {
-	struct usergart_entry entry[NUM_USERGART_ENTRIES];
-	int height;				/* height in rows */
-	int height_shift;		/* ilog2(height in rows) */
-	int slot_shift;			/* ilog2(width per slot) */
-	int stride_pfn;			/* stride in pages */
-	int last;				/* index of last used entry */
-} *usergart;
-
-static void evict_entry(struct drm_gem_object *obj,
-		enum tiler_fmt fmt, struct usergart_entry *entry)
-{
-	if (obj->dev->dev_mapping) {
-		struct omap_gem_object *omap_obj = to_omap_bo(obj);
-		int n = usergart[fmt].height;
-		size_t size = PAGE_SIZE * n;
-		loff_t off = mmap_offset(obj) +
-				(entry->obj_pgoff << PAGE_SHIFT);
-		const int m = 1 + ((omap_obj->width << fmt) / PAGE_SIZE);
-		if (m > 1) {
-			int i;
-			/* if stride > than PAGE_SIZE then sparse mapping: */
-			for (i = n; i > 0; i--) {
-				unmap_mapping_range(obj->dev->dev_mapping,
-						off, PAGE_SIZE, 1);
-				off += PAGE_SIZE * m;
-			}
-		} else {
-			unmap_mapping_range(obj->dev->dev_mapping, off, size, 1);
-		}
-	}
-
-	entry->obj = NULL;
-}
-
-/* Evict a buffer from usergart, if it is mapped there */
-static void evict(struct drm_gem_object *obj)
-{
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-
-	if (omap_obj->flags & OMAP_BO_TILED) {
-		enum tiler_fmt fmt = gem2fmt(omap_obj->flags);
-		int i;
-
-		if (!usergart)
-			return;
-
-		for (i = 0; i < NUM_USERGART_ENTRIES; i++) {
-			struct usergart_entry *entry = &usergart[fmt].entry[i];
-			if (entry->obj == obj)
-				evict_entry(obj, fmt, entry);
-		}
-	}
-}
-
-/* GEM objects can either be allocated from contiguous memory (in which
- * case obj->filp==NULL), or w/ shmem backing (obj->filp!=NULL).  But non
- * contiguous buffers can be remapped in TILER/DMM if they need to be
- * contiguous... but we don't do this all the time to reduce pressure
- * on TILER/DMM space when we know at allocation time that the buffer
- * will need to be scanned out.
- */
-static inline bool is_shmem(struct drm_gem_object *obj)
-{
-	return obj->filp != NULL;
-}
-
-/**
- * shmem buffers that are mapped cached can simulate coherency via using
- * page faulting to keep track of dirty pages
- */
-static inline bool is_cached_coherent(struct drm_gem_object *obj)
-{
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-	return is_shmem(obj) &&
-		((omap_obj->flags & OMAP_BO_CACHE_MASK) == OMAP_BO_CACHED);
-}
-
-static DEFINE_SPINLOCK(sync_lock);
-
-/** ensure backing pages are allocated */
-static int omap_gem_attach_pages(struct drm_gem_object *obj)
-{
-	struct drm_device *dev = obj->dev;
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-	struct page **pages;
-	int npages = obj->size >> PAGE_SHIFT;
-	int i, ret;
-	dma_addr_t *addrs;
-
-	WARN_ON(omap_obj->pages);
-
-	/* TODO: __GFP_DMA32 .. but somehow GFP_HIGHMEM is coming from the
-	 * mapping_gfp_mask(mapping) which conflicts w/ GFP_DMA32.. probably
-	 * we actually want CMA memory for it all anyways..
-	 */
-	pages = _drm_gem_get_pages(obj, GFP_KERNEL);
-	if (IS_ERR(pages)) {
-		dev_err(obj->dev->dev, "could not get pages: %ld\n", PTR_ERR(pages));
-		return PTR_ERR(pages);
-	}
-
-	/* for non-cached buffers, ensure the new pages are clean because
-	 * DSS, GPU, etc. are not cache coherent:
-	 */
-	if (omap_obj->flags & (OMAP_BO_WC|OMAP_BO_UNCACHED)) {
-		addrs = kmalloc(npages * sizeof(*addrs), GFP_KERNEL);
-		if (!addrs) {
-			ret = -ENOMEM;
-			goto free_pages;
-		}
-
-		for (i = 0; i < npages; i++) {
-			addrs[i] = dma_map_page(dev->dev, pages[i],
-					0, PAGE_SIZE, DMA_BIDIRECTIONAL);
-		}
-	} else {
-		addrs = kzalloc(npages * sizeof(*addrs), GFP_KERNEL);
-		if (!addrs) {
-			ret = -ENOMEM;
-			goto free_pages;
-		}
-	}
-
-	omap_obj->addrs = addrs;
-	omap_obj->pages = pages;
-
-	return 0;
-
-free_pages:
-	_drm_gem_put_pages(obj, pages, true, false);
-
-	return ret;
-}
-
-/** release backing pages */
-static void omap_gem_detach_pages(struct drm_gem_object *obj)
-{
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-
-	/* for non-cached buffers, ensure the new pages are clean because
-	 * DSS, GPU, etc. are not cache coherent:
-	 */
-	if (omap_obj->flags & (OMAP_BO_WC|OMAP_BO_UNCACHED)) {
-		int i, npages = obj->size >> PAGE_SHIFT;
-		for (i = 0; i < npages; i++) {
-			dma_unmap_page(obj->dev->dev, omap_obj->addrs[i],
-					PAGE_SIZE, DMA_BIDIRECTIONAL);
-		}
-	}
-
-	kfree(omap_obj->addrs);
-	omap_obj->addrs = NULL;
-
-	_drm_gem_put_pages(obj, omap_obj->pages, true, false);
-	omap_obj->pages = NULL;
-}
-
-/* get buffer flags */
-uint32_t omap_gem_flags(struct drm_gem_object *obj)
-{
-	return to_omap_bo(obj)->flags;
-}
-
-/** get mmap offset */
-static uint64_t mmap_offset(struct drm_gem_object *obj)
-{
-	struct drm_device *dev = obj->dev;
-
-	WARN_ON(!mutex_is_locked(&dev->struct_mutex));
-
-	if (!obj->map_list.map) {
-		/* Make it mmapable */
-		size_t size = omap_gem_mmap_size(obj);
-		int ret = _drm_gem_create_mmap_offset_size(obj, size);
-
-		if (ret) {
-			dev_err(dev->dev, "could not allocate mmap offset\n");
-			return 0;
-		}
-	}
-
-	return (uint64_t)obj->map_list.hash.key << PAGE_SHIFT;
-}
-
-uint64_t omap_gem_mmap_offset(struct drm_gem_object *obj)
-{
-	uint64_t offset;
-	mutex_lock(&obj->dev->struct_mutex);
-	offset = mmap_offset(obj);
-	mutex_unlock(&obj->dev->struct_mutex);
-	return offset;
-}
-
-/** get mmap size */
-size_t omap_gem_mmap_size(struct drm_gem_object *obj)
-{
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-	size_t size = obj->size;
-
-	if (omap_obj->flags & OMAP_BO_TILED) {
-		/* for tiled buffers, the virtual size has stride rounded up
-		 * to 4kb.. (to hide the fact that row n+1 might start 16kb or
-		 * 32kb later!).  But we don't back the entire buffer with
-		 * pages, only the valid picture part.. so need to adjust for
-		 * this in the size used to mmap and generate mmap offset
-		 */
-		size = tiler_vsize(gem2fmt(omap_obj->flags),
-				omap_obj->width, omap_obj->height);
-	}
-
-	return size;
-}
-
-/* get tiled size, returns -EINVAL if not tiled buffer */
-int omap_gem_tiled_size(struct drm_gem_object *obj, uint16_t *w, uint16_t *h)
-{
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-	if (omap_obj->flags & OMAP_BO_TILED) {
-		*w = omap_obj->width;
-		*h = omap_obj->height;
-		return 0;
-	}
-	return -EINVAL;
-}
-
-/* Normal handling for the case of faulting in non-tiled buffers */
-static int fault_1d(struct drm_gem_object *obj,
-		struct vm_area_struct *vma, struct vm_fault *vmf)
-{
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-	unsigned long pfn;
-	pgoff_t pgoff;
-
-	/* We don't use vmf->pgoff since that has the fake offset: */
-	pgoff = ((unsigned long)vmf->virtual_address -
-			vma->vm_start) >> PAGE_SHIFT;
-
-	if (omap_obj->pages) {
-		omap_gem_cpu_sync(obj, pgoff);
-		pfn = page_to_pfn(omap_obj->pages[pgoff]);
-	} else {
-		BUG_ON(!(omap_obj->flags & OMAP_BO_DMA));
-		pfn = (omap_obj->paddr >> PAGE_SHIFT) + pgoff;
-	}
-
-	VERB("Inserting %p pfn %lx, pa %lx", vmf->virtual_address,
-			pfn, pfn << PAGE_SHIFT);
-
-	return vm_insert_mixed(vma, (unsigned long)vmf->virtual_address, pfn);
-}
-
-/* Special handling for the case of faulting in 2d tiled buffers */
-static int fault_2d(struct drm_gem_object *obj,
-		struct vm_area_struct *vma, struct vm_fault *vmf)
-{
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-	struct usergart_entry *entry;
-	enum tiler_fmt fmt = gem2fmt(omap_obj->flags);
-	struct page *pages[64];  /* XXX is this too much to have on stack? */
-	unsigned long pfn;
-	pgoff_t pgoff, base_pgoff;
-	void __user *vaddr;
-	int i, ret, slots;
-
-	/*
-	 * Note the height of the slot is also equal to the number of pages
-	 * that need to be mapped in to fill 4kb wide CPU page.  If the slot
-	 * height is 64, then 64 pages fill a 4kb wide by 64 row region.
-	 */
-	const int n = usergart[fmt].height;
-	const int n_shift = usergart[fmt].height_shift;
-
-	/*
-	 * If buffer width in bytes > PAGE_SIZE then the virtual stride is
-	 * rounded up to next multiple of PAGE_SIZE.. this need to be taken
-	 * into account in some of the math, so figure out virtual stride
-	 * in pages
-	 */
-	const int m = 1 + ((omap_obj->width << fmt) / PAGE_SIZE);
-
-	/* We don't use vmf->pgoff since that has the fake offset: */
-	pgoff = ((unsigned long)vmf->virtual_address -
-			vma->vm_start) >> PAGE_SHIFT;
-
-	/*
-	 * Actual address we start mapping at is rounded down to previous slot
-	 * boundary in the y direction:
-	 */
-	base_pgoff = round_down(pgoff, m << n_shift);
-
-	/* figure out buffer width in slots */
-	slots = omap_obj->width >> usergart[fmt].slot_shift;
-
-	vaddr = vmf->virtual_address - ((pgoff - base_pgoff) << PAGE_SHIFT);
-
-	entry = &usergart[fmt].entry[usergart[fmt].last];
-
-	/* evict previous buffer using this usergart entry, if any: */
-	if (entry->obj)
-		evict_entry(entry->obj, fmt, entry);
-
-	entry->obj = obj;
-	entry->obj_pgoff = base_pgoff;
-
-	/* now convert base_pgoff to phys offset from virt offset: */
-	base_pgoff = (base_pgoff >> n_shift) * slots;
-
-	/* for wider-than 4k.. figure out which part of the slot-row we want: */
-	if (m > 1) {
-		int off = pgoff % m;
-		entry->obj_pgoff += off;
-		base_pgoff /= m;
-		slots = min(slots - (off << n_shift), n);
-		base_pgoff += off << n_shift;
-		vaddr += off << PAGE_SHIFT;
-	}
-
-	/*
-	 * Map in pages. Beyond the valid pixel part of the buffer, we set
-	 * pages[i] to NULL to get a dummy page mapped in.. if someone
-	 * reads/writes it they will get random/undefined content, but at
-	 * least it won't be corrupting whatever other random page used to
-	 * be mapped in, or other undefined behavior.
-	 */
-	memcpy(pages, &omap_obj->pages[base_pgoff],
-			sizeof(struct page *) * slots);
-	memset(pages + slots, 0,
-			sizeof(struct page *) * (n - slots));
-
-	ret = tiler_pin(entry->block, pages, ARRAY_SIZE(pages), 0, true);
-	if (ret) {
-		dev_err(obj->dev->dev, "failed to pin: %d\n", ret);
-		return ret;
-	}
-
-	pfn = entry->paddr >> PAGE_SHIFT;
-
-	VERB("Inserting %p pfn %lx, pa %lx", vmf->virtual_address,
-			pfn, pfn << PAGE_SHIFT);
-
-	for (i = n; i > 0; i--) {
-		vm_insert_mixed(vma, (unsigned long)vaddr, pfn);
-		pfn += usergart[fmt].stride_pfn;
-		vaddr += PAGE_SIZE * m;
-	}
-
-	/* simple round-robin: */
-	usergart[fmt].last = (usergart[fmt].last + 1) % NUM_USERGART_ENTRIES;
-
-	return 0;
-}
-
-/**
- * omap_gem_fault		-	pagefault handler for GEM objects
- * @vma: the VMA of the GEM object
- * @vmf: fault detail
- *
- * Invoked when a fault occurs on an mmap of a GEM managed area. GEM
- * does most of the work for us including the actual map/unmap calls
- * but we need to do the actual page work.
- *
- * The VMA was set up by GEM. In doing so it also ensured that the
- * vma->vm_private_data points to the GEM object that is backing this
- * mapping.
- */
-int omap_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
-{
-	struct drm_gem_object *obj = vma->vm_private_data;
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-	struct drm_device *dev = obj->dev;
-	struct page **pages;
-	int ret;
-
-	/* Make sure we don't parallel update on a fault, nor move or remove
-	 * something from beneath our feet
-	 */
-	mutex_lock(&dev->struct_mutex);
-
-	/* if a shmem backed object, make sure we have pages attached now */
-	ret = get_pages(obj, &pages);
-	if (ret)
-		goto fail;
-
-	/* where should we do corresponding put_pages().. we are mapping
-	 * the original page, rather than thru a GART, so we can't rely
-	 * on eviction to trigger this.  But munmap() or all mappings should
-	 * probably trigger put_pages()?
-	 */
-
-	if (omap_obj->flags & OMAP_BO_TILED)
-		ret = fault_2d(obj, vma, vmf);
-	else
-		ret = fault_1d(obj, vma, vmf);
-
-
-fail:
-	mutex_unlock(&dev->struct_mutex);
-	switch (ret) {
-	case 0:
-	case -ERESTARTSYS:
-	case -EINTR:
-		return VM_FAULT_NOPAGE;
-	case -ENOMEM:
-		return VM_FAULT_OOM;
-	default:
-		return VM_FAULT_SIGBUS;
-	}
-}
-
-/** We override mainly to fix up some of the vm mapping flags.. */
-int omap_gem_mmap(struct file *filp, struct vm_area_struct *vma)
-{
-	int ret;
-
-	ret = drm_gem_mmap(filp, vma);
-	if (ret) {
-		DBG("mmap failed: %d", ret);
-		return ret;
-	}
-
-	return omap_gem_mmap_obj(vma->vm_private_data, vma);
-}
-
-int omap_gem_mmap_obj(struct drm_gem_object *obj,
-		struct vm_area_struct *vma)
-{
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-
-	vma->vm_flags &= ~VM_PFNMAP;
-	vma->vm_flags |= VM_MIXEDMAP;
-
-	if (omap_obj->flags & OMAP_BO_WC) {
-		vma->vm_page_prot = pgprot_writecombine(vm_get_page_prot(vma->vm_flags));
-	} else if (omap_obj->flags & OMAP_BO_UNCACHED) {
-		vma->vm_page_prot = pgprot_noncached(vm_get_page_prot(vma->vm_flags));
-	} else {
-		/*
-		 * We do have some private objects, at least for scanout buffers
-		 * on hardware without DMM/TILER.  But these are allocated write-
-		 * combine
-		 */
-		if (WARN_ON(!obj->filp))
-			return -EINVAL;
-
-		/*
-		 * Shunt off cached objs to shmem file so they have their own
-		 * address_space (so unmap_mapping_range does what we want,
-		 * in particular in the case of mmap'd dmabufs)
-		 */
-		fput(vma->vm_file);
-		vma->vm_pgoff = 0;
-		vma->vm_file  = get_file(obj->filp);
-
-		vma->vm_page_prot = vm_get_page_prot(vma->vm_flags);
-	}
-
-	return 0;
-}
-
-
-/**
- * omap_gem_dumb_create	-	create a dumb buffer
- * @drm_file: our client file
- * @dev: our device
- * @args: the requested arguments copied from userspace
- *
- * Allocate a buffer suitable for use for a frame buffer of the
- * form described by user space. Give userspace a handle by which
- * to reference it.
- */
-int omap_gem_dumb_create(struct drm_file *file, struct drm_device *dev,
-		struct drm_mode_create_dumb *args)
-{
-	union omap_gem_size gsize;
-
-	/* in case someone tries to feed us a completely bogus stride: */
-	args->pitch = align_pitch(args->pitch, args->width, args->bpp);
-	args->size = PAGE_ALIGN(args->pitch * args->height);
-
-	gsize = (union omap_gem_size){
-		.bytes = args->size,
-	};
-
-	return omap_gem_new_handle(dev, file, gsize,
-			OMAP_BO_SCANOUT | OMAP_BO_WC, &args->handle);
-}
-
-/**
- * omap_gem_dumb_destroy	-	destroy a dumb buffer
- * @file: client file
- * @dev: our DRM device
- * @handle: the object handle
- *
- * Destroy a handle that was created via omap_gem_dumb_create.
- */
-int omap_gem_dumb_destroy(struct drm_file *file, struct drm_device *dev,
-		uint32_t handle)
-{
-	/* No special work needed, drop the reference and see what falls out */
-	return drm_gem_handle_delete(file, handle);
-}
-
-/**
- * omap_gem_dumb_map	-	buffer mapping for dumb interface
- * @file: our drm client file
- * @dev: drm device
- * @handle: GEM handle to the object (from dumb_create)
- *
- * Do the necessary setup to allow the mapping of the frame buffer
- * into user memory. We don't have to do much here at the moment.
- */
-int omap_gem_dumb_map_offset(struct drm_file *file, struct drm_device *dev,
-		uint32_t handle, uint64_t *offset)
-{
-	struct drm_gem_object *obj;
-	int ret = 0;
-
-	/* GEM does all our handle to object mapping */
-	obj = drm_gem_object_lookup(dev, file, handle);
-	if (obj == NULL) {
-		ret = -ENOENT;
-		goto fail;
-	}
-
-	*offset = omap_gem_mmap_offset(obj);
-
-	drm_gem_object_unreference_unlocked(obj);
-
-fail:
-	return ret;
-}
-
-/* Set scrolling position.  This allows us to implement fast scrolling
- * for console.
- *
- * Call only from non-atomic contexts.
- */
-int omap_gem_roll(struct drm_gem_object *obj, uint32_t roll)
-{
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-	uint32_t npages = obj->size >> PAGE_SHIFT;
-	int ret = 0;
-
-	if (roll > npages) {
-		dev_err(obj->dev->dev, "invalid roll: %d\n", roll);
-		return -EINVAL;
-	}
-
-	omap_obj->roll = roll;
-
-	mutex_lock(&obj->dev->struct_mutex);
-
-	/* if we aren't mapped yet, we don't need to do anything */
-	if (omap_obj->block) {
-		struct page **pages;
-		ret = get_pages(obj, &pages);
-		if (ret)
-			goto fail;
-		ret = tiler_pin(omap_obj->block, pages, npages, roll, true);
-		if (ret)
-			dev_err(obj->dev->dev, "could not repin: %d\n", ret);
-	}
-
-fail:
-	mutex_unlock(&obj->dev->struct_mutex);
-
-	return ret;
-}
-
-/* Sync the buffer for CPU access.. note pages should already be
- * attached, ie. omap_gem_get_pages()
- */
-void omap_gem_cpu_sync(struct drm_gem_object *obj, int pgoff)
-{
-	struct drm_device *dev = obj->dev;
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-
-	if (is_cached_coherent(obj) && omap_obj->addrs[pgoff]) {
-		dma_unmap_page(dev->dev, omap_obj->addrs[pgoff],
-				PAGE_SIZE, DMA_BIDIRECTIONAL);
-		omap_obj->addrs[pgoff] = 0;
-	}
-}
-
-/* sync the buffer for DMA access */
-void omap_gem_dma_sync(struct drm_gem_object *obj,
-		enum dma_data_direction dir)
-{
-	struct drm_device *dev = obj->dev;
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-
-	if (is_cached_coherent(obj)) {
-		int i, npages = obj->size >> PAGE_SHIFT;
-		struct page **pages = omap_obj->pages;
-		bool dirty = false;
-
-		for (i = 0; i < npages; i++) {
-			if (!omap_obj->addrs[i]) {
-				omap_obj->addrs[i] = dma_map_page(dev->dev, pages[i], 0,
-						PAGE_SIZE, DMA_BIDIRECTIONAL);
-				dirty = true;
-			}
-		}
-
-		if (dirty) {
-			unmap_mapping_range(obj->filp->f_mapping, 0,
-					omap_gem_mmap_size(obj), 1);
-		}
-	}
-}
-
-/* Get physical address for DMA.. if 'remap' is true, and the buffer is not
- * already contiguous, remap it to pin in physically contiguous memory.. (ie.
- * map in TILER)
- */
-int omap_gem_get_paddr(struct drm_gem_object *obj,
-		dma_addr_t *paddr, bool remap)
-{
-	struct omap_drm_private *priv = obj->dev->dev_private;
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-	int ret = 0;
-
-	mutex_lock(&obj->dev->struct_mutex);
-
-	if (remap && is_shmem(obj) && priv->has_dmm) {
-		if (omap_obj->paddr_cnt == 0) {
-			struct page **pages;
-			uint32_t npages = obj->size >> PAGE_SHIFT;
-			enum tiler_fmt fmt = gem2fmt(omap_obj->flags);
-			struct tiler_block *block;
-
-			BUG_ON(omap_obj->block);
-
-			ret = get_pages(obj, &pages);
-			if (ret)
-				goto fail;
-
-			if (omap_obj->flags & OMAP_BO_TILED) {
-				block = tiler_reserve_2d(fmt,
-						omap_obj->width,
-						omap_obj->height, 0);
-			} else {
-				block = tiler_reserve_1d(obj->size);
-			}
-
-			if (IS_ERR(block)) {
-				ret = PTR_ERR(block);
-				dev_err(obj->dev->dev,
-					"could not remap: %d (%d)\n", ret, fmt);
-				goto fail;
-			}
-
-			/* TODO: enable async refill.. */
-			ret = tiler_pin(block, pages, npages,
-					omap_obj->roll, true);
-			if (ret) {
-				tiler_release(block);
-				dev_err(obj->dev->dev,
-						"could not pin: %d\n", ret);
-				goto fail;
-			}
-
-			omap_obj->paddr = tiler_ssptr(block);
-			omap_obj->block = block;
-
-			DBG("got paddr: %08x", omap_obj->paddr);
-		}
-
-		omap_obj->paddr_cnt++;
-
-		*paddr = omap_obj->paddr;
-	} else if (omap_obj->flags & OMAP_BO_DMA) {
-		*paddr = omap_obj->paddr;
-	} else {
-		ret = -EINVAL;
-		goto fail;
-	}
-
-fail:
-	mutex_unlock(&obj->dev->struct_mutex);
-
-	return ret;
-}
-
-/* Release physical address, when DMA is no longer being performed.. this
- * could potentially unpin and unmap buffers from TILER
- */
-int omap_gem_put_paddr(struct drm_gem_object *obj)
-{
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-	int ret = 0;
-
-	mutex_lock(&obj->dev->struct_mutex);
-	if (omap_obj->paddr_cnt > 0) {
-		omap_obj->paddr_cnt--;
-		if (omap_obj->paddr_cnt == 0) {
-			ret = tiler_unpin(omap_obj->block);
-			if (ret) {
-				dev_err(obj->dev->dev,
-					"could not unpin pages: %d\n", ret);
-				goto fail;
-			}
-			ret = tiler_release(omap_obj->block);
-			if (ret) {
-				dev_err(obj->dev->dev,
-					"could not release unmap: %d\n", ret);
-			}
-			omap_obj->block = NULL;
-		}
-	}
-fail:
-	mutex_unlock(&obj->dev->struct_mutex);
-	return ret;
-}
-
-/* Get rotated scanout address (only valid if already pinned), at the
- * specified orientation and x,y offset from top-left corner of buffer
- * (only valid for tiled 2d buffers)
- */
-int omap_gem_rotated_paddr(struct drm_gem_object *obj, uint32_t orient,
-		int x, int y, dma_addr_t *paddr)
-{
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-	int ret = -EINVAL;
-
-	mutex_lock(&obj->dev->struct_mutex);
-	if ((omap_obj->paddr_cnt > 0) && omap_obj->block &&
-			(omap_obj->flags & OMAP_BO_TILED)) {
-		*paddr = tiler_tsptr(omap_obj->block, orient, x, y);
-		ret = 0;
-	}
-	mutex_unlock(&obj->dev->struct_mutex);
-	return ret;
-}
-
-/* Get tiler stride for the buffer (only valid for 2d tiled buffers) */
-int omap_gem_tiled_stride(struct drm_gem_object *obj, uint32_t orient)
-{
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-	int ret = -EINVAL;
-	if (omap_obj->flags & OMAP_BO_TILED)
-		ret = tiler_stride(gem2fmt(omap_obj->flags), orient);
-	return ret;
-}
-
-/* acquire pages when needed (for example, for DMA where physically
- * contiguous buffer is not required
- */
-static int get_pages(struct drm_gem_object *obj, struct page ***pages)
-{
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-	int ret = 0;
-
-	if (is_shmem(obj) && !omap_obj->pages) {
-		ret = omap_gem_attach_pages(obj);
-		if (ret) {
-			dev_err(obj->dev->dev, "could not attach pages\n");
-			return ret;
-		}
-	}
-
-	/* TODO: even phys-contig.. we should have a list of pages? */
-	*pages = omap_obj->pages;
-
-	return 0;
-}
-
-/* if !remap, and we don't have pages backing, then fail, rather than
- * increasing the pin count (which we don't really do yet anyways,
- * because we don't support swapping pages back out).  And 'remap'
- * might not be quite the right name, but I wanted to keep it working
- * similarly to omap_gem_get_paddr().  Note though that mutex is not
- * aquired if !remap (because this can be called in atomic ctxt),
- * but probably omap_gem_get_paddr() should be changed to work in the
- * same way.  If !remap, a matching omap_gem_put_pages() call is not
- * required (and should not be made).
- */
-int omap_gem_get_pages(struct drm_gem_object *obj, struct page ***pages,
-		bool remap)
-{
-	int ret;
-	if (!remap) {
-		struct omap_gem_object *omap_obj = to_omap_bo(obj);
-		if (!omap_obj->pages)
-			return -ENOMEM;
-		*pages = omap_obj->pages;
-		return 0;
-	}
-	mutex_lock(&obj->dev->struct_mutex);
-	ret = get_pages(obj, pages);
-	mutex_unlock(&obj->dev->struct_mutex);
-	return ret;
-}
-
-/* release pages when DMA no longer being performed */
-int omap_gem_put_pages(struct drm_gem_object *obj)
-{
-	/* do something here if we dynamically attach/detach pages.. at
-	 * least they would no longer need to be pinned if everyone has
-	 * released the pages..
-	 */
-	return 0;
-}
-
-/* Get kernel virtual address for CPU access.. this more or less only
- * exists for omap_fbdev.  This should be called with struct_mutex
- * held.
- */
-void *omap_gem_vaddr(struct drm_gem_object *obj)
-{
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-	WARN_ON(!mutex_is_locked(&obj->dev->struct_mutex));
-	if (!omap_obj->vaddr) {
-		struct page **pages;
-		int ret = get_pages(obj, &pages);
-		if (ret)
-			return ERR_PTR(ret);
-		omap_obj->vaddr = vmap(pages, obj->size >> PAGE_SHIFT,
-				VM_MAP, pgprot_writecombine(PAGE_KERNEL));
-	}
-	return omap_obj->vaddr;
-}
-
-#ifdef CONFIG_PM
-/* re-pin objects in DMM in resume path: */
-int omap_gem_resume(struct device *dev)
-{
-	struct drm_device *drm_dev = dev_get_drvdata(dev);
-	struct omap_drm_private *priv = drm_dev->dev_private;
-	struct omap_gem_object *omap_obj;
-	int ret = 0;
-
-	list_for_each_entry(omap_obj, &priv->obj_list, mm_list) {
-		if (omap_obj->block) {
-			struct drm_gem_object *obj = &omap_obj->base;
-			uint32_t npages = obj->size >> PAGE_SHIFT;
-			WARN_ON(!omap_obj->pages);  /* this can't happen */
-			ret = tiler_pin(omap_obj->block,
-					omap_obj->pages, npages,
-					omap_obj->roll, true);
-			if (ret) {
-				dev_err(dev, "could not repin: %d\n", ret);
-				return ret;
-			}
-		}
-	}
-
-	return 0;
-}
-#endif
-
-#ifdef CONFIG_DEBUG_FS
-void omap_gem_describe(struct drm_gem_object *obj, struct seq_file *m)
-{
-	struct drm_device *dev = obj->dev;
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-	uint64_t off = 0;
-
-	WARN_ON(!mutex_is_locked(&dev->struct_mutex));
-
-	if (obj->map_list.map)
-		off = (uint64_t)obj->map_list.hash.key;
-
-	seq_printf(m, "%08x: %2d (%2d) %08llx %08Zx (%2d) %p %4d",
-			omap_obj->flags, obj->name, obj->refcount.refcount.counter,
-			off, omap_obj->paddr, omap_obj->paddr_cnt,
-			omap_obj->vaddr, omap_obj->roll);
-
-	if (omap_obj->flags & OMAP_BO_TILED) {
-		seq_printf(m, " %dx%d", omap_obj->width, omap_obj->height);
-		if (omap_obj->block) {
-			struct tcm_area *area = &omap_obj->block->area;
-			seq_printf(m, " (%dx%d, %dx%d)",
-					area->p0.x, area->p0.y,
-					area->p1.x, area->p1.y);
-		}
-	} else {
-		seq_printf(m, " %d", obj->size);
-	}
-
-	seq_printf(m, "\n");
-}
-
-void omap_gem_describe_objects(struct list_head *list, struct seq_file *m)
-{
-	struct omap_gem_object *omap_obj;
-	int count = 0;
-	size_t size = 0;
-
-	list_for_each_entry(omap_obj, list, mm_list) {
-		struct drm_gem_object *obj = &omap_obj->base;
-		seq_printf(m, "   ");
-		omap_gem_describe(obj, m);
-		count++;
-		size += obj->size;
-	}
-
-	seq_printf(m, "Total %d objects, %zu bytes\n", count, size);
-}
-#endif
-
-/* Buffer Synchronization:
- */
-
-struct omap_gem_sync_waiter {
-	struct list_head list;
-	struct omap_gem_object *omap_obj;
-	enum omap_gem_op op;
-	uint32_t read_target, write_target;
-	/* notify called w/ sync_lock held */
-	void (*notify)(void *arg);
-	void *arg;
-};
-
-/* list of omap_gem_sync_waiter.. the notify fxn gets called back when
- * the read and/or write target count is achieved which can call a user
- * callback (ex. to kick 3d and/or 2d), wakeup blocked task (prep for
- * cpu access), etc.
- */
-static LIST_HEAD(waiters);
-
-static inline bool is_waiting(struct omap_gem_sync_waiter *waiter)
-{
-	struct omap_gem_object *omap_obj = waiter->omap_obj;
-	if ((waiter->op & OMAP_GEM_READ) &&
-			(omap_obj->sync->read_complete < waiter->read_target))
-		return true;
-	if ((waiter->op & OMAP_GEM_WRITE) &&
-			(omap_obj->sync->write_complete < waiter->write_target))
-		return true;
-	return false;
-}
-
-/* macro for sync debug.. */
-#define SYNCDBG 0
-#define SYNC(fmt, ...) do { if (SYNCDBG) \
-		printk(KERN_ERR "%s:%d: "fmt"\n", \
-				__func__, __LINE__, ##__VA_ARGS__); \
-	} while (0)
-
-
-static void sync_op_update(void)
-{
-	struct omap_gem_sync_waiter *waiter, *n;
-	list_for_each_entry_safe(waiter, n, &waiters, list) {
-		if (!is_waiting(waiter)) {
-			list_del(&waiter->list);
-			SYNC("notify: %p", waiter);
-			waiter->notify(waiter->arg);
-			kfree(waiter);
-		}
-	}
-}
-
-static inline int sync_op(struct drm_gem_object *obj,
-		enum omap_gem_op op, bool start)
-{
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-	int ret = 0;
-
-	spin_lock(&sync_lock);
-
-	if (!omap_obj->sync) {
-		omap_obj->sync = kzalloc(sizeof(*omap_obj->sync), GFP_ATOMIC);
-		if (!omap_obj->sync) {
-			ret = -ENOMEM;
-			goto unlock;
-		}
-	}
-
-	if (start) {
-		if (op & OMAP_GEM_READ)
-			omap_obj->sync->read_pending++;
-		if (op & OMAP_GEM_WRITE)
-			omap_obj->sync->write_pending++;
-	} else {
-		if (op & OMAP_GEM_READ)
-			omap_obj->sync->read_complete++;
-		if (op & OMAP_GEM_WRITE)
-			omap_obj->sync->write_complete++;
-		sync_op_update();
-	}
-
-unlock:
-	spin_unlock(&sync_lock);
-
-	return ret;
-}
-
-/* it is a bit lame to handle updates in this sort of polling way, but
- * in case of PVR, the GPU can directly update read/write complete
- * values, and not really tell us which ones it updated.. this also
- * means that sync_lock is not quite sufficient.  So we'll need to
- * do something a bit better when it comes time to add support for
- * separate 2d hw..
- */
-void omap_gem_op_update(void)
-{
-	spin_lock(&sync_lock);
-	sync_op_update();
-	spin_unlock(&sync_lock);
-}
-
-/* mark the start of read and/or write operation */
-int omap_gem_op_start(struct drm_gem_object *obj, enum omap_gem_op op)
-{
-	return sync_op(obj, op, true);
-}
-
-int omap_gem_op_finish(struct drm_gem_object *obj, enum omap_gem_op op)
-{
-	return sync_op(obj, op, false);
-}
-
-static DECLARE_WAIT_QUEUE_HEAD(sync_event);
-
-static void sync_notify(void *arg)
-{
-	struct task_struct **waiter_task = arg;
-	*waiter_task = NULL;
-	wake_up_all(&sync_event);
-}
-
-int omap_gem_op_sync(struct drm_gem_object *obj, enum omap_gem_op op)
-{
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-	int ret = 0;
-	if (omap_obj->sync) {
-		struct task_struct *waiter_task = current;
-		struct omap_gem_sync_waiter *waiter =
-				kzalloc(sizeof(*waiter), GFP_KERNEL);
-
-		if (!waiter)
-			return -ENOMEM;
-
-		waiter->omap_obj = omap_obj;
-		waiter->op = op;
-		waiter->read_target = omap_obj->sync->read_pending;
-		waiter->write_target = omap_obj->sync->write_pending;
-		waiter->notify = sync_notify;
-		waiter->arg = &waiter_task;
-
-		spin_lock(&sync_lock);
-		if (is_waiting(waiter)) {
-			SYNC("waited: %p", waiter);
-			list_add_tail(&waiter->list, &waiters);
-			spin_unlock(&sync_lock);
-			ret = wait_event_interruptible(sync_event,
-					(waiter_task == NULL));
-			spin_lock(&sync_lock);
-			if (waiter_task) {
-				SYNC("interrupted: %p", waiter);
-				/* we were interrupted */
-				list_del(&waiter->list);
-				waiter_task = NULL;
-			} else {
-				/* freed in sync_op_update() */
-				waiter = NULL;
-			}
-		}
-		spin_unlock(&sync_lock);
-
-		if (waiter)
-			kfree(waiter);
-	}
-	return ret;
-}
-
-/* call fxn(arg), either synchronously or asynchronously if the op
- * is currently blocked..  fxn() can be called from any context
- *
- * (TODO for now fxn is called back from whichever context calls
- * omap_gem_op_update().. but this could be better defined later
- * if needed)
- *
- * TODO more code in common w/ _sync()..
- */
-int omap_gem_op_async(struct drm_gem_object *obj, enum omap_gem_op op,
-		void (*fxn)(void *arg), void *arg)
-{
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-	if (omap_obj->sync) {
-		struct omap_gem_sync_waiter *waiter =
-				kzalloc(sizeof(*waiter), GFP_ATOMIC);
-
-		if (!waiter)
-			return -ENOMEM;
-
-		waiter->omap_obj = omap_obj;
-		waiter->op = op;
-		waiter->read_target = omap_obj->sync->read_pending;
-		waiter->write_target = omap_obj->sync->write_pending;
-		waiter->notify = fxn;
-		waiter->arg = arg;
-
-		spin_lock(&sync_lock);
-		if (is_waiting(waiter)) {
-			SYNC("waited: %p", waiter);
-			list_add_tail(&waiter->list, &waiters);
-			spin_unlock(&sync_lock);
-			return 0;
-		}
-
-		spin_unlock(&sync_lock);
-	}
-
-	/* no waiting.. */
-	fxn(arg);
-
-	return 0;
-}
-
-/* special API so PVR can update the buffer to use a sync-object allocated
- * from it's sync-obj heap.  Only used for a newly allocated (from PVR's
- * perspective) sync-object, so we overwrite the new syncobj w/ values
- * from the already allocated syncobj (if there is one)
- */
-int omap_gem_set_sync_object(struct drm_gem_object *obj, void *syncobj)
-{
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-	int ret = 0;
-
-	spin_lock(&sync_lock);
-
-	if ((omap_obj->flags & OMAP_BO_EXT_SYNC) && !syncobj) {
-		/* clearing a previously set syncobj */
-		syncobj = kmemdup(omap_obj->sync, sizeof(*omap_obj->sync),
-				  GFP_ATOMIC);
-		if (!syncobj) {
-			ret = -ENOMEM;
-			goto unlock;
-		}
-		omap_obj->flags &= ~OMAP_BO_EXT_SYNC;
-		omap_obj->sync = syncobj;
-	} else if (syncobj && !(omap_obj->flags & OMAP_BO_EXT_SYNC)) {
-		/* replacing an existing syncobj */
-		if (omap_obj->sync) {
-			memcpy(syncobj, omap_obj->sync, sizeof(*omap_obj->sync));
-			kfree(omap_obj->sync);
-		}
-		omap_obj->flags |= OMAP_BO_EXT_SYNC;
-		omap_obj->sync = syncobj;
-	}
-
-unlock:
-	spin_unlock(&sync_lock);
-	return ret;
-}
-
-int omap_gem_init_object(struct drm_gem_object *obj)
-{
-	return -EINVAL;          /* unused */
-}
-
-/* don't call directly.. called from GEM core when it is time to actually
- * free the object..
- */
-void omap_gem_free_object(struct drm_gem_object *obj)
-{
-	struct drm_device *dev = obj->dev;
-	struct omap_gem_object *omap_obj = to_omap_bo(obj);
-
-	evict(obj);
-
-	WARN_ON(!mutex_is_locked(&dev->struct_mutex));
-
-	list_del(&omap_obj->mm_list);
-
-	if (obj->map_list.map)
-		drm_gem_free_mmap_offset(obj);
-
-	/* this means the object is still pinned.. which really should
-	 * not happen.  I think..
-	 */
-	WARN_ON(omap_obj->paddr_cnt > 0);
-
-	/* don't free externally allocated backing memory */
-	if (!(omap_obj->flags & OMAP_BO_EXT_MEM)) {
-		if (omap_obj->pages)
-			omap_gem_detach_pages(obj);
-
-		if (!is_shmem(obj)) {
-			dma_free_writecombine(dev->dev, obj->size,
-					omap_obj->vaddr, omap_obj->paddr);
-		} else if (omap_obj->vaddr) {
-			vunmap(omap_obj->vaddr);
-		}
-	}
-
-	/* don't free externally allocated syncobj */
-	if (!(omap_obj->flags & OMAP_BO_EXT_SYNC))
-		kfree(omap_obj->sync);
-
-	drm_gem_object_release(obj);
-
-	kfree(obj);
-}
-
-/* convenience method to construct a GEM buffer object, and userspace handle */
-int omap_gem_new_handle(struct drm_device *dev, struct drm_file *file,
-		union omap_gem_size gsize, uint32_t flags, uint32_t *handle)
-{
-	struct drm_gem_object *obj;
-	int ret;
-
-	obj = omap_gem_new(dev, gsize, flags);
-	if (!obj)
-		return -ENOMEM;
-
-	ret = drm_gem_handle_create(file, obj, handle);
-	if (ret) {
-		drm_gem_object_release(obj);
-		kfree(obj); /* TODO isn't there a dtor to call? just copying i915 */
-		return ret;
-	}
-
-	/* drop reference from allocate - handle holds it now */
-	drm_gem_object_unreference_unlocked(obj);
-
-	return 0;
-}
-
-/* GEM buffer object constructor */
-struct drm_gem_object *omap_gem_new(struct drm_device *dev,
-		union omap_gem_size gsize, uint32_t flags)
-{
-	struct omap_drm_private *priv = dev->dev_private;
-	struct omap_gem_object *omap_obj;
-	struct drm_gem_object *obj = NULL;
-	size_t size;
-	int ret;
-
-	if (flags & OMAP_BO_TILED) {
-		if (!usergart) {
-			dev_err(dev->dev, "Tiled buffers require DMM\n");
-			goto fail;
-		}
-
-		/* tiled buffers are always shmem paged backed.. when they are
-		 * scanned out, they are remapped into DMM/TILER
-		 */
-		flags &= ~OMAP_BO_SCANOUT;
-
-		/* currently don't allow cached buffers.. there is some caching
-		 * stuff that needs to be handled better
-		 */
-		flags &= ~(OMAP_BO_CACHED|OMAP_BO_UNCACHED);
-		flags |= OMAP_BO_WC;
-
-		/* align dimensions to slot boundaries... */
-		tiler_align(gem2fmt(flags),
-				&gsize.tiled.width, &gsize.tiled.height);
-
-		/* ...and calculate size based on aligned dimensions */
-		size = tiler_size(gem2fmt(flags),
-				gsize.tiled.width, gsize.tiled.height);
-	} else {
-		size = PAGE_ALIGN(gsize.bytes);
-	}
-
-	omap_obj = kzalloc(sizeof(*omap_obj), GFP_KERNEL);
-	if (!omap_obj) {
-		dev_err(dev->dev, "could not allocate GEM object\n");
-		goto fail;
-	}
-
-	list_add(&omap_obj->mm_list, &priv->obj_list);
-
-	obj = &omap_obj->base;
-
-	if ((flags & OMAP_BO_SCANOUT) && !priv->has_dmm) {
-		/* attempt to allocate contiguous memory if we don't
-		 * have DMM for remappign discontiguous buffers
-		 */
-		omap_obj->vaddr =  dma_alloc_writecombine(dev->dev, size,
-				&omap_obj->paddr, GFP_KERNEL);
-		if (omap_obj->vaddr)
-			flags |= OMAP_BO_DMA;
-
-	}
-
-	omap_obj->flags = flags;
-
-	if (flags & OMAP_BO_TILED) {
-		omap_obj->width = gsize.tiled.width;
-		omap_obj->height = gsize.tiled.height;
-	}
-
-	if (flags & (OMAP_BO_DMA|OMAP_BO_EXT_MEM))
-		ret = drm_gem_private_object_init(dev, obj, size);
-	else
-		ret = drm_gem_object_init(dev, obj, size);
-
-	if (ret)
-		goto fail;
-
-	return obj;
-
-fail:
-	if (obj)
-		omap_gem_free_object(obj);
-
-	return NULL;
-}
-
-/* init/cleanup.. if DMM is used, we need to set some stuff up.. */
-void omap_gem_init(struct drm_device *dev)
-{
-	struct omap_drm_private *priv = dev->dev_private;
-	const enum tiler_fmt fmts[] = {
-			TILFMT_8BIT, TILFMT_16BIT, TILFMT_32BIT
-	};
-	int i, j;
-
-	if (!dmm_is_available()) {
-		/* DMM only supported on OMAP4 and later, so this isn't fatal */
-		dev_warn(dev->dev, "DMM not available, disable DMM support\n");
-		return;
-	}
-
-	usergart = kzalloc(3 * sizeof(*usergart), GFP_KERNEL);
-	if (!usergart) {
-		dev_warn(dev->dev, "could not allocate usergart\n");
-		return;
-	}
-
-	/* reserve 4k aligned/wide regions for userspace mappings: */
-	for (i = 0; i < ARRAY_SIZE(fmts); i++) {
-		uint16_t h = 1, w = PAGE_SIZE >> i;
-		tiler_align(fmts[i], &w, &h);
-		/* note: since each region is 1 4kb page wide, and minimum
-		 * number of rows, the height ends up being the same as the
-		 * # of pages in the region
-		 */
-		usergart[i].height = h;
-		usergart[i].height_shift = ilog2(h);
-		usergart[i].stride_pfn = tiler_stride(fmts[i], 0) >> PAGE_SHIFT;
-		usergart[i].slot_shift = ilog2((PAGE_SIZE / h) >> i);
-		for (j = 0; j < NUM_USERGART_ENTRIES; j++) {
-			struct usergart_entry *entry = &usergart[i].entry[j];
-			struct tiler_block *block =
-					tiler_reserve_2d(fmts[i], w, h,
-							PAGE_SIZE);
-			if (IS_ERR(block)) {
-				dev_err(dev->dev,
-						"reserve failed: %d, %d, %ld\n",
-						i, j, PTR_ERR(block));
-				return;
-			}
-			entry->paddr = tiler_ssptr(block);
-			entry->block = block;
-
-			DBG("%d:%d: %dx%d: paddr=%08x stride=%d", i, j, w, h,
-					entry->paddr,
-					usergart[i].stride_pfn << PAGE_SHIFT);
-		}
-	}
-
-	priv->has_dmm = true;
-}
-
-void omap_gem_deinit(struct drm_device *dev)
-{
-	/* I believe we can rely on there being no more outstanding GEM
-	 * objects which could depend on usergart/dmm at this point.
-	 */
-	kfree(usergart);
-}
diff --git a/drivers/staging/omapdrm/omap_gem_dmabuf.c b/drivers/staging/omapdrm/omap_gem_dmabuf.c
deleted file mode 100644
index a3236ab..0000000
--- a/drivers/staging/omapdrm/omap_gem_dmabuf.c
+++ /dev/null
@@ -1,225 +0,0 @@
-/*
- * drivers/staging/omapdrm/omap_gem_dmabuf.c
- *
- * Copyright (C) 2011 Texas Instruments
- * Author: Rob Clark <rob.clark@linaro.org>
- *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of the GNU General Public License version 2 as published by
- * the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but WITHOUT
- * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
- * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
- * more details.
- *
- * You should have received a copy of the GNU General Public License along with
- * this program.  If not, see <http://www.gnu.org/licenses/>.
- */
-
-#include "omap_drv.h"
-
-#include <linux/dma-buf.h>
-
-static struct sg_table *omap_gem_map_dma_buf(
-		struct dma_buf_attachment *attachment,
-		enum dma_data_direction dir)
-{
-	struct drm_gem_object *obj = attachment->dmabuf->priv;
-	struct sg_table *sg;
-	dma_addr_t paddr;
-	int ret;
-
-	sg = kzalloc(sizeof(*sg), GFP_KERNEL);
-	if (!sg)
-		return ERR_PTR(-ENOMEM);
-
-	/* camera, etc, need physically contiguous.. but we need a
-	 * better way to know this..
-	 */
-	ret = omap_gem_get_paddr(obj, &paddr, true);
-	if (ret)
-		goto out;
-
-	ret = sg_alloc_table(sg, 1, GFP_KERNEL);
-	if (ret)
-		goto out;
-
-	sg_init_table(sg->sgl, 1);
-	sg_dma_len(sg->sgl) = obj->size;
-	sg_set_page(sg->sgl, pfn_to_page(PFN_DOWN(paddr)), obj->size, 0);
-	sg_dma_address(sg->sgl) = paddr;
-
-	/* this should be after _get_paddr() to ensure we have pages attached */
-	omap_gem_dma_sync(obj, dir);
-
-	return sg;
-out:
-	kfree(sg);
-	return ERR_PTR(ret);
-}
-
-static void omap_gem_unmap_dma_buf(struct dma_buf_attachment *attachment,
-		struct sg_table *sg, enum dma_data_direction dir)
-{
-	struct drm_gem_object *obj = attachment->dmabuf->priv;
-	omap_gem_put_paddr(obj);
-	sg_free_table(sg);
-	kfree(sg);
-}
-
-static void omap_gem_dmabuf_release(struct dma_buf *buffer)
-{
-	struct drm_gem_object *obj = buffer->priv;
-	/* release reference that was taken when dmabuf was exported
-	 * in omap_gem_prime_set()..
-	 */
-	drm_gem_object_unreference_unlocked(obj);
-}
-
-
-static int omap_gem_dmabuf_begin_cpu_access(struct dma_buf *buffer,
-		size_t start, size_t len, enum dma_data_direction dir)
-{
-	struct drm_gem_object *obj = buffer->priv;
-	struct page **pages;
-	if (omap_gem_flags(obj) & OMAP_BO_TILED) {
-		/* TODO we would need to pin at least part of the buffer to
-		 * get de-tiled view.  For now just reject it.
-		 */
-		return -ENOMEM;
-	}
-	/* make sure we have the pages: */
-	return omap_gem_get_pages(obj, &pages, true);
-}
-
-static void omap_gem_dmabuf_end_cpu_access(struct dma_buf *buffer,
-		size_t start, size_t len, enum dma_data_direction dir)
-{
-	struct drm_gem_object *obj = buffer->priv;
-	omap_gem_put_pages(obj);
-}
-
-
-static void *omap_gem_dmabuf_kmap_atomic(struct dma_buf *buffer,
-		unsigned long page_num)
-{
-	struct drm_gem_object *obj = buffer->priv;
-	struct page **pages;
-	omap_gem_get_pages(obj, &pages, false);
-	omap_gem_cpu_sync(obj, page_num);
-	return kmap_atomic(pages[page_num]);
-}
-
-static void omap_gem_dmabuf_kunmap_atomic(struct dma_buf *buffer,
-		unsigned long page_num, void *addr)
-{
-	kunmap_atomic(addr);
-}
-
-static void *omap_gem_dmabuf_kmap(struct dma_buf *buffer,
-		unsigned long page_num)
-{
-	struct drm_gem_object *obj = buffer->priv;
-	struct page **pages;
-	omap_gem_get_pages(obj, &pages, false);
-	omap_gem_cpu_sync(obj, page_num);
-	return kmap(pages[page_num]);
-}
-
-static void omap_gem_dmabuf_kunmap(struct dma_buf *buffer,
-		unsigned long page_num, void *addr)
-{
-	struct drm_gem_object *obj = buffer->priv;
-	struct page **pages;
-	omap_gem_get_pages(obj, &pages, false);
-	kunmap(pages[page_num]);
-}
-
-/*
- * TODO maybe we can split up drm_gem_mmap to avoid duplicating
- * some here.. or at least have a drm_dmabuf_mmap helper.
- */
-static int omap_gem_dmabuf_mmap(struct dma_buf *buffer,
-		struct vm_area_struct *vma)
-{
-	struct drm_gem_object *obj = buffer->priv;
-	int ret = 0;
-
-	if (WARN_ON(!obj->filp))
-		return -EINVAL;
-
-	/* Check for valid size. */
-	if (omap_gem_mmap_size(obj) < vma->vm_end - vma->vm_start) {
-		ret = -EINVAL;
-		goto out_unlock;
-	}
-
-	if (!obj->dev->driver->gem_vm_ops) {
-		ret = -EINVAL;
-		goto out_unlock;
-	}
-
-	vma->vm_flags |= VM_IO | VM_PFNMAP | VM_DONTEXPAND | VM_DONTDUMP;
-	vma->vm_ops = obj->dev->driver->gem_vm_ops;
-	vma->vm_private_data = obj;
-	vma->vm_page_prot =  pgprot_writecombine(vm_get_page_prot(vma->vm_flags));
-
-	/* Take a ref for this mapping of the object, so that the fault
-	 * handler can dereference the mmap offset's pointer to the object.
-	 * This reference is cleaned up by the corresponding vm_close
-	 * (which should happen whether the vma was created by this call, or
-	 * by a vm_open due to mremap or partial unmap or whatever).
-	 */
-	vma->vm_ops->open(vma);
-
-out_unlock:
-
-	return omap_gem_mmap_obj(obj, vma);
-}
-
-struct dma_buf_ops omap_dmabuf_ops = {
-		.map_dma_buf = omap_gem_map_dma_buf,
-		.unmap_dma_buf = omap_gem_unmap_dma_buf,
-		.release = omap_gem_dmabuf_release,
-		.begin_cpu_access = omap_gem_dmabuf_begin_cpu_access,
-		.end_cpu_access = omap_gem_dmabuf_end_cpu_access,
-		.kmap_atomic = omap_gem_dmabuf_kmap_atomic,
-		.kunmap_atomic = omap_gem_dmabuf_kunmap_atomic,
-		.kmap = omap_gem_dmabuf_kmap,
-		.kunmap = omap_gem_dmabuf_kunmap,
-		.mmap = omap_gem_dmabuf_mmap,
-};
-
-struct dma_buf *omap_gem_prime_export(struct drm_device *dev,
-		struct drm_gem_object *obj, int flags)
-{
-	return dma_buf_export(obj, &omap_dmabuf_ops, obj->size, flags);
-}
-
-struct drm_gem_object *omap_gem_prime_import(struct drm_device *dev,
-		struct dma_buf *buffer)
-{
-	struct drm_gem_object *obj;
-
-	/* is this one of own objects? */
-	if (buffer->ops == &omap_dmabuf_ops) {
-		obj = buffer->priv;
-		/* is it from our device? */
-		if (obj->dev == dev) {
-			/*
-			 * Importing dmabuf exported from out own gem increases
-			 * refcount on gem itself instead of f_count of dmabuf.
-			 */
-			drm_gem_object_reference(obj);
-			dma_buf_put(buffer);
-			return obj;
-		}
-	}
-
-	/*
-	 * TODO add support for importing buffers from other devices..
-	 * for now we don't need this but would be nice to add eventually
-	 */
-	return ERR_PTR(-EINVAL);
-}
diff --git a/drivers/staging/omapdrm/omap_gem_helpers.c b/drivers/staging/omapdrm/omap_gem_helpers.c
deleted file mode 100644
index ffb8cce..0000000
--- a/drivers/staging/omapdrm/omap_gem_helpers.c
+++ /dev/null
@@ -1,169 +0,0 @@
-/*
- * drivers/staging/omapdrm/omap_gem_helpers.c
- *
- * Copyright (C) 2011 Texas Instruments
- * Author: Rob Clark <rob.clark@linaro.org>
- *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of the GNU General Public License version 2 as published by
- * the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but WITHOUT
- * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
- * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
- * more details.
- *
- * You should have received a copy of the GNU General Public License along with
- * this program.  If not, see <http://www.gnu.org/licenses/>.
- */
-
-/* temporary copy of drm_gem_{get,put}_pages() until the
- * "drm/gem: add functions to get/put pages" patch is merged..
- */
-
-#include <linux/module.h>
-#include <linux/types.h>
-#include <linux/shmem_fs.h>
-
-#include <drm/drmP.h>
-
-/**
- * drm_gem_get_pages - helper to allocate backing pages for a GEM object
- * @obj: obj in question
- * @gfpmask: gfp mask of requested pages
- */
-struct page **_drm_gem_get_pages(struct drm_gem_object *obj, gfp_t gfpmask)
-{
-	struct inode *inode;
-	struct address_space *mapping;
-	struct page *p, **pages;
-	int i, npages;
-
-	/* This is the shared memory object that backs the GEM resource */
-	inode = obj->filp->f_path.dentry->d_inode;
-	mapping = inode->i_mapping;
-
-	npages = obj->size >> PAGE_SHIFT;
-
-	pages = drm_malloc_ab(npages, sizeof(struct page *));
-	if (pages == NULL)
-		return ERR_PTR(-ENOMEM);
-
-	gfpmask |= mapping_gfp_mask(mapping);
-
-	for (i = 0; i < npages; i++) {
-		p = shmem_read_mapping_page_gfp(mapping, i, gfpmask);
-		if (IS_ERR(p))
-			goto fail;
-		pages[i] = p;
-
-		/* There is a hypothetical issue w/ drivers that require
-		 * buffer memory in the low 4GB.. if the pages are un-
-		 * pinned, and swapped out, they can end up swapped back
-		 * in above 4GB.  If pages are already in memory, then
-		 * shmem_read_mapping_page_gfp will ignore the gfpmask,
-		 * even if the already in-memory page disobeys the mask.
-		 *
-		 * It is only a theoretical issue today, because none of
-		 * the devices with this limitation can be populated with
-		 * enough memory to trigger the issue.  But this BUG_ON()
-		 * is here as a reminder in case the problem with
-		 * shmem_read_mapping_page_gfp() isn't solved by the time
-		 * it does become a real issue.
-		 *
-		 * See this thread: http://lkml.org/lkml/2011/7/11/238
-		 */
-		BUG_ON((gfpmask & __GFP_DMA32) &&
-				(page_to_pfn(p) >= 0x00100000UL));
-	}
-
-	return pages;
-
-fail:
-	while (i--)
-		page_cache_release(pages[i]);
-
-	drm_free_large(pages);
-	return ERR_CAST(p);
-}
-
-/**
- * drm_gem_put_pages - helper to free backing pages for a GEM object
- * @obj: obj in question
- * @pages: pages to free
- */
-void _drm_gem_put_pages(struct drm_gem_object *obj, struct page **pages,
-		bool dirty, bool accessed)
-{
-	int i, npages;
-
-	npages = obj->size >> PAGE_SHIFT;
-
-	for (i = 0; i < npages; i++) {
-		if (dirty)
-			set_page_dirty(pages[i]);
-
-		if (accessed)
-			mark_page_accessed(pages[i]);
-
-		/* Undo the reference we took when populating the table */
-		page_cache_release(pages[i]);
-	}
-
-	drm_free_large(pages);
-}
-
-int
-_drm_gem_create_mmap_offset_size(struct drm_gem_object *obj, size_t size)
-{
-	struct drm_device *dev = obj->dev;
-	struct drm_gem_mm *mm = dev->mm_private;
-	struct drm_map_list *list;
-	struct drm_local_map *map;
-	int ret = 0;
-
-	/* Set the object up for mmap'ing */
-	list = &obj->map_list;
-	list->map = kzalloc(sizeof(struct drm_map_list), GFP_KERNEL);
-	if (!list->map)
-		return -ENOMEM;
-
-	map = list->map;
-	map->type = _DRM_GEM;
-	map->size = size;
-	map->handle = obj;
-
-	/* Get a DRM GEM mmap offset allocated... */
-	list->file_offset_node = drm_mm_search_free(&mm->offset_manager,
-			size / PAGE_SIZE, 0, 0);
-
-	if (!list->file_offset_node) {
-		DRM_ERROR("failed to allocate offset for bo %d\n", obj->name);
-		ret = -ENOSPC;
-		goto out_free_list;
-	}
-
-	list->file_offset_node = drm_mm_get_block(list->file_offset_node,
-			size / PAGE_SIZE, 0);
-	if (!list->file_offset_node) {
-		ret = -ENOMEM;
-		goto out_free_list;
-	}
-
-	list->hash.key = list->file_offset_node->start;
-	ret = drm_ht_insert_item(&mm->offset_hash, &list->hash);
-	if (ret) {
-		DRM_ERROR("failed to add to map hash\n");
-		goto out_free_mm;
-	}
-
-	return 0;
-
-out_free_mm:
-	drm_mm_put_block(list->file_offset_node);
-out_free_list:
-	kfree(list->map);
-	list->map = NULL;
-
-	return ret;
-}
diff --git a/drivers/staging/omapdrm/omap_irq.c b/drivers/staging/omapdrm/omap_irq.c
deleted file mode 100644
index 2629ba7..0000000
--- a/drivers/staging/omapdrm/omap_irq.c
+++ /dev/null
@@ -1,322 +0,0 @@
-/*
- * drivers/staging/omapdrm/omap_irq.c
- *
- * Copyright (C) 2012 Texas Instruments
- * Author: Rob Clark <rob.clark@linaro.org>
- *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of the GNU General Public License version 2 as published by
- * the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but WITHOUT
- * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
- * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
- * more details.
- *
- * You should have received a copy of the GNU General Public License along with
- * this program.  If not, see <http://www.gnu.org/licenses/>.
- */
-
-#include "omap_drv.h"
-
-static DEFINE_SPINLOCK(list_lock);
-
-static void omap_irq_error_handler(struct omap_drm_irq *irq,
-		uint32_t irqstatus)
-{
-	DRM_ERROR("errors: %08x\n", irqstatus);
-}
-
-/* call with list_lock and dispc runtime held */
-static void omap_irq_update(struct drm_device *dev)
-{
-	struct omap_drm_private *priv = dev->dev_private;
-	struct omap_drm_irq *irq;
-	uint32_t irqmask = priv->vblank_mask;
-
-	BUG_ON(!spin_is_locked(&list_lock));
-
-	list_for_each_entry(irq, &priv->irq_list, node)
-		irqmask |= irq->irqmask;
-
-	DBG("irqmask=%08x", irqmask);
-
-	dispc_write_irqenable(irqmask);
-	dispc_read_irqenable();        /* flush posted write */
-}
-
-void omap_irq_register(struct drm_device *dev, struct omap_drm_irq *irq)
-{
-	struct omap_drm_private *priv = dev->dev_private;
-	unsigned long flags;
-
-	dispc_runtime_get();
-	spin_lock_irqsave(&list_lock, flags);
-
-	if (!WARN_ON(irq->registered)) {
-		irq->registered = true;
-		list_add(&irq->node, &priv->irq_list);
-		omap_irq_update(dev);
-	}
-
-	spin_unlock_irqrestore(&list_lock, flags);
-	dispc_runtime_put();
-}
-
-void omap_irq_unregister(struct drm_device *dev, struct omap_drm_irq *irq)
-{
-	unsigned long flags;
-
-	dispc_runtime_get();
-	spin_lock_irqsave(&list_lock, flags);
-
-	if (!WARN_ON(!irq->registered)) {
-		irq->registered = false;
-		list_del(&irq->node);
-		omap_irq_update(dev);
-	}
-
-	spin_unlock_irqrestore(&list_lock, flags);
-	dispc_runtime_put();
-}
-
-struct omap_irq_wait {
-	struct omap_drm_irq irq;
-	int count;
-};
-
-static DECLARE_WAIT_QUEUE_HEAD(wait_event);
-
-static void wait_irq(struct omap_drm_irq *irq, uint32_t irqstatus)
-{
-	struct omap_irq_wait *wait =
-			container_of(irq, struct omap_irq_wait, irq);
-	wait->count--;
-	wake_up_all(&wait_event);
-}
-
-struct omap_irq_wait * omap_irq_wait_init(struct drm_device *dev,
-		uint32_t irqmask, int count)
-{
-	struct omap_irq_wait *wait = kzalloc(sizeof(*wait), GFP_KERNEL);
-	wait->irq.irq = wait_irq;
-	wait->irq.irqmask = irqmask;
-	wait->count = count;
-	omap_irq_register(dev, &wait->irq);
-	return wait;
-}
-
-int omap_irq_wait(struct drm_device *dev, struct omap_irq_wait *wait,
-		unsigned long timeout)
-{
-	int ret = wait_event_timeout(wait_event, (wait->count <= 0), timeout);
-	omap_irq_unregister(dev, &wait->irq);
-	kfree(wait);
-	if (ret == 0)
-		return -1;
-	return 0;
-}
-
-/**
- * enable_vblank - enable vblank interrupt events
- * @dev: DRM device
- * @crtc: which irq to enable
- *
- * Enable vblank interrupts for @crtc.  If the device doesn't have
- * a hardware vblank counter, this routine should be a no-op, since
- * interrupts will have to stay on to keep the count accurate.
- *
- * RETURNS
- * Zero on success, appropriate errno if the given @crtc's vblank
- * interrupt cannot be enabled.
- */
-int omap_irq_enable_vblank(struct drm_device *dev, int crtc)
-{
-	struct omap_drm_private *priv = dev->dev_private;
-	unsigned long flags;
-
-	DBG("dev=%p, crtc=%d", dev, crtc);
-
-	dispc_runtime_get();
-	spin_lock_irqsave(&list_lock, flags);
-	priv->vblank_mask |= pipe2vbl(crtc);
-	omap_irq_update(dev);
-	spin_unlock_irqrestore(&list_lock, flags);
-	dispc_runtime_put();
-
-	return 0;
-}
-
-/**
- * disable_vblank - disable vblank interrupt events
- * @dev: DRM device
- * @crtc: which irq to enable
- *
- * Disable vblank interrupts for @crtc.  If the device doesn't have
- * a hardware vblank counter, this routine should be a no-op, since
- * interrupts will have to stay on to keep the count accurate.
- */
-void omap_irq_disable_vblank(struct drm_device *dev, int crtc)
-{
-	struct omap_drm_private *priv = dev->dev_private;
-	unsigned long flags;
-
-	DBG("dev=%p, crtc=%d", dev, crtc);
-
-	dispc_runtime_get();
-	spin_lock_irqsave(&list_lock, flags);
-	priv->vblank_mask &= ~pipe2vbl(crtc);
-	omap_irq_update(dev);
-	spin_unlock_irqrestore(&list_lock, flags);
-	dispc_runtime_put();
-}
-
-irqreturn_t omap_irq_handler(DRM_IRQ_ARGS)
-{
-	struct drm_device *dev = (struct drm_device *) arg;
-	struct omap_drm_private *priv = dev->dev_private;
-	struct omap_drm_irq *handler, *n;
-	unsigned long flags;
-	unsigned int id;
-	u32 irqstatus;
-
-	irqstatus = dispc_read_irqstatus();
-	dispc_clear_irqstatus(irqstatus);
-	dispc_read_irqstatus();        /* flush posted write */
-
-	VERB("irqs: %08x", irqstatus);
-
-	for (id = 0; id < priv->num_crtcs; id++)
-		if (irqstatus & pipe2vbl(id))
-			drm_handle_vblank(dev, id);
-
-	spin_lock_irqsave(&list_lock, flags);
-	list_for_each_entry_safe(handler, n, &priv->irq_list, node) {
-		if (handler->irqmask & irqstatus) {
-			spin_unlock_irqrestore(&list_lock, flags);
-			handler->irq(handler, handler->irqmask & irqstatus);
-			spin_lock_irqsave(&list_lock, flags);
-		}
-	}
-	spin_unlock_irqrestore(&list_lock, flags);
-
-	return IRQ_HANDLED;
-}
-
-void omap_irq_preinstall(struct drm_device *dev)
-{
-	DBG("dev=%p", dev);
-	dispc_runtime_get();
-	dispc_clear_irqstatus(0xffffffff);
-	dispc_runtime_put();
-}
-
-int omap_irq_postinstall(struct drm_device *dev)
-{
-	struct omap_drm_private *priv = dev->dev_private;
-	struct omap_drm_irq *error_handler = &priv->error_handler;
-
-	DBG("dev=%p", dev);
-
-	INIT_LIST_HEAD(&priv->irq_list);
-
-	error_handler->irq = omap_irq_error_handler;
-	error_handler->irqmask = DISPC_IRQ_OCP_ERR;
-
-	/* for now ignore DISPC_IRQ_SYNC_LOST_DIGIT.. really I think
-	 * we just need to ignore it while enabling tv-out
-	 */
-	error_handler->irqmask &= ~DISPC_IRQ_SYNC_LOST_DIGIT;
-
-	omap_irq_register(dev, error_handler);
-
-	return 0;
-}
-
-void omap_irq_uninstall(struct drm_device *dev)
-{
-	DBG("dev=%p", dev);
-	// TODO prolly need to call drm_irq_uninstall() somewhere too
-}
-
-/*
- * We need a special version, instead of just using drm_irq_install(),
- * because we need to register the irq via omapdss.  Once omapdss and
- * omapdrm are merged together we can assign the dispc hwmod data to
- * ourselves and drop these and just use drm_irq_{install,uninstall}()
- */
-
-int omap_drm_irq_install(struct drm_device *dev)
-{
-	int ret;
-
-	mutex_lock(&dev->struct_mutex);
-
-	if (dev->irq_enabled) {
-		mutex_unlock(&dev->struct_mutex);
-		return -EBUSY;
-	}
-	dev->irq_enabled = 1;
-	mutex_unlock(&dev->struct_mutex);
-
-	/* Before installing handler */
-	if (dev->driver->irq_preinstall)
-		dev->driver->irq_preinstall(dev);
-
-	ret = dispc_request_irq(dev->driver->irq_handler, dev);
-
-	if (ret < 0) {
-		mutex_lock(&dev->struct_mutex);
-		dev->irq_enabled = 0;
-		mutex_unlock(&dev->struct_mutex);
-		return ret;
-	}
-
-	/* After installing handler */
-	if (dev->driver->irq_postinstall)
-		ret = dev->driver->irq_postinstall(dev);
-
-	if (ret < 0) {
-		mutex_lock(&dev->struct_mutex);
-		dev->irq_enabled = 0;
-		mutex_unlock(&dev->struct_mutex);
-		dispc_free_irq(dev);
-	}
-
-	return ret;
-}
-
-int omap_drm_irq_uninstall(struct drm_device *dev)
-{
-	unsigned long irqflags;
-	int irq_enabled, i;
-
-	mutex_lock(&dev->struct_mutex);
-	irq_enabled = dev->irq_enabled;
-	dev->irq_enabled = 0;
-	mutex_unlock(&dev->struct_mutex);
-
-	/*
-	 * Wake up any waiters so they don't hang.
-	 */
-	if (dev->num_crtcs) {
-		spin_lock_irqsave(&dev->vbl_lock, irqflags);
-		for (i = 0; i < dev->num_crtcs; i++) {
-			DRM_WAKEUP(&dev->vbl_queue[i]);
-			dev->vblank_enabled[i] = 0;
-			dev->last_vblank[i] =
-				dev->driver->get_vblank_counter(dev, i);
-		}
-		spin_unlock_irqrestore(&dev->vbl_lock, irqflags);
-	}
-
-	if (!irq_enabled)
-		return -EINVAL;
-
-	if (dev->driver->irq_uninstall)
-		dev->driver->irq_uninstall(dev);
-
-	dispc_free_irq(dev);
-
-	return 0;
-}
diff --git a/drivers/staging/omapdrm/omap_plane.c b/drivers/staging/omapdrm/omap_plane.c
deleted file mode 100644
index bb989d7..0000000
--- a/drivers/staging/omapdrm/omap_plane.c
+++ /dev/null
@@ -1,450 +0,0 @@
-/*
- * drivers/staging/omapdrm/omap_plane.c
- *
- * Copyright (C) 2011 Texas Instruments
- * Author: Rob Clark <rob.clark@linaro.org>
- *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of the GNU General Public License version 2 as published by
- * the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but WITHOUT
- * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
- * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
- * more details.
- *
- * You should have received a copy of the GNU General Public License along with
- * this program.  If not, see <http://www.gnu.org/licenses/>.
- */
-
-#include <linux/kfifo.h>
-
-#include "omap_drv.h"
-#include "omap_dmm_tiler.h"
-
-/* some hackery because omapdss has an 'enum omap_plane' (which would be
- * better named omap_plane_id).. and compiler seems unhappy about having
- * both a 'struct omap_plane' and 'enum omap_plane'
- */
-#define omap_plane _omap_plane
-
-/*
- * plane funcs
- */
-
-struct callback {
-	void (*fxn)(void *);
-	void *arg;
-};
-
-#define to_omap_plane(x) container_of(x, struct omap_plane, base)
-
-struct omap_plane {
-	struct drm_plane base;
-	int id;  /* TODO rename omap_plane -> omap_plane_id in omapdss so I can use the enum */
-	const char *name;
-	struct omap_overlay_info info;
-	struct omap_drm_apply apply;
-
-	/* position/orientation of scanout within the fb: */
-	struct omap_drm_window win;
-	bool enabled;
-
-	/* last fb that we pinned: */
-	struct drm_framebuffer *pinned_fb;
-
-	uint32_t nformats;
-	uint32_t formats[32];
-
-	struct omap_drm_irq error_irq;
-
-	/* set of bo's pending unpin until next post_apply() */
-	DECLARE_KFIFO_PTR(unpin_fifo, struct drm_gem_object *);
-
-	// XXX maybe get rid of this and handle vblank in crtc too?
-	struct callback apply_done_cb;
-};
-
-static void unpin(void *arg, struct drm_gem_object *bo)
-{
-	struct drm_plane *plane = arg;
-	struct omap_plane *omap_plane = to_omap_plane(plane);
-
-	if (kfifo_put(&omap_plane->unpin_fifo,
-			(const struct drm_gem_object **)&bo)) {
-		/* also hold a ref so it isn't free'd while pinned */
-		drm_gem_object_reference(bo);
-	} else {
-		dev_err(plane->dev->dev, "unpin fifo full!\n");
-		omap_gem_put_paddr(bo);
-	}
-}
-
-/* update which fb (if any) is pinned for scanout */
-static int update_pin(struct drm_plane *plane, struct drm_framebuffer *fb)
-{
-	struct omap_plane *omap_plane = to_omap_plane(plane);
-	struct drm_framebuffer *pinned_fb = omap_plane->pinned_fb;
-
-	if (pinned_fb != fb) {
-		int ret;
-
-		DBG("%p -> %p", pinned_fb, fb);
-
-		if (fb)
-			drm_framebuffer_reference(fb);
-
-		ret = omap_framebuffer_replace(pinned_fb, fb, plane, unpin);
-
-		if (pinned_fb)
-			drm_framebuffer_unreference(pinned_fb);
-
-		if (ret) {
-			dev_err(plane->dev->dev, "could not swap %p -> %p\n",
-					omap_plane->pinned_fb, fb);
-			if (fb)
-				drm_framebuffer_unreference(fb);
-			omap_plane->pinned_fb = NULL;
-			return ret;
-		}
-
-		omap_plane->pinned_fb = fb;
-	}
-
-	return 0;
-}
-
-static void omap_plane_pre_apply(struct omap_drm_apply *apply)
-{
-	struct omap_plane *omap_plane =
-			container_of(apply, struct omap_plane, apply);
-	struct omap_drm_window *win = &omap_plane->win;
-	struct drm_plane *plane = &omap_plane->base;
-	struct drm_device *dev = plane->dev;
-	struct omap_overlay_info *info = &omap_plane->info;
-	struct drm_crtc *crtc = plane->crtc;
-	enum omap_channel channel;
-	bool enabled = omap_plane->enabled && crtc;
-	bool ilace, replication;
-	int ret;
-
-	DBG("%s, enabled=%d", omap_plane->name, enabled);
-
-	/* if fb has changed, pin new fb: */
-	update_pin(plane, enabled ? plane->fb : NULL);
-
-	if (!enabled) {
-		dispc_ovl_enable(omap_plane->id, false);
-		return;
-	}
-
-	channel = omap_crtc_channel(crtc);
-
-	/* update scanout: */
-	omap_framebuffer_update_scanout(plane->fb, win, info);
-
-	DBG("%dx%d -> %dx%d (%d)", info->width, info->height,
-			info->out_width, info->out_height,
-			info->screen_width);
-	DBG("%d,%d %08x %08x", info->pos_x, info->pos_y,
-			info->paddr, info->p_uv_addr);
-
-	/* TODO: */
-	ilace = false;
-	replication = false;
-
-	/* and finally, update omapdss: */
-	ret = dispc_ovl_setup(omap_plane->id, info,
-			replication, omap_crtc_timings(crtc), false);
-	if (ret) {
-		dev_err(dev->dev, "dispc_ovl_setup failed: %d\n", ret);
-		return;
-	}
-
-	dispc_ovl_enable(omap_plane->id, true);
-	dispc_ovl_set_channel_out(omap_plane->id, channel);
-}
-
-static void omap_plane_post_apply(struct omap_drm_apply *apply)
-{
-	struct omap_plane *omap_plane =
-			container_of(apply, struct omap_plane, apply);
-	struct drm_plane *plane = &omap_plane->base;
-	struct omap_overlay_info *info = &omap_plane->info;
-	struct drm_gem_object *bo = NULL;
-	struct callback cb;
-
-	cb = omap_plane->apply_done_cb;
-	omap_plane->apply_done_cb.fxn = NULL;
-
-	while (kfifo_get(&omap_plane->unpin_fifo, &bo)) {
-		omap_gem_put_paddr(bo);
-		drm_gem_object_unreference_unlocked(bo);
-	}
-
-	if (cb.fxn)
-		cb.fxn(cb.arg);
-
-	if (omap_plane->enabled) {
-		omap_framebuffer_flush(plane->fb, info->pos_x, info->pos_y,
-				info->out_width, info->out_height);
-	}
-}
-
-static int apply(struct drm_plane *plane)
-{
-	if (plane->crtc) {
-		struct omap_plane *omap_plane = to_omap_plane(plane);
-		return omap_crtc_apply(plane->crtc, &omap_plane->apply);
-	}
-	return 0;
-}
-
-int omap_plane_mode_set(struct drm_plane *plane,
-		struct drm_crtc *crtc, struct drm_framebuffer *fb,
-		int crtc_x, int crtc_y,
-		unsigned int crtc_w, unsigned int crtc_h,
-		uint32_t src_x, uint32_t src_y,
-		uint32_t src_w, uint32_t src_h,
-		void (*fxn)(void *), void *arg)
-{
-	struct omap_plane *omap_plane = to_omap_plane(plane);
-	struct omap_drm_window *win = &omap_plane->win;
-
-	win->crtc_x = crtc_x;
-	win->crtc_y = crtc_y;
-	win->crtc_w = crtc_w;
-	win->crtc_h = crtc_h;
-
-	/* src values are in Q16 fixed point, convert to integer: */
-	win->src_x = src_x >> 16;
-	win->src_y = src_y >> 16;
-	win->src_w = src_w >> 16;
-	win->src_h = src_h >> 16;
-
-	if (fxn) {
-		/* omap_crtc should ensure that a new page flip
-		 * isn't permitted while there is one pending:
-		 */
-		BUG_ON(omap_plane->apply_done_cb.fxn);
-
-		omap_plane->apply_done_cb.fxn = fxn;
-		omap_plane->apply_done_cb.arg = arg;
-	}
-
-	plane->fb = fb;
-	plane->crtc = crtc;
-
-	return apply(plane);
-}
-
-static int omap_plane_update(struct drm_plane *plane,
-		struct drm_crtc *crtc, struct drm_framebuffer *fb,
-		int crtc_x, int crtc_y,
-		unsigned int crtc_w, unsigned int crtc_h,
-		uint32_t src_x, uint32_t src_y,
-		uint32_t src_w, uint32_t src_h)
-{
-	struct omap_plane *omap_plane = to_omap_plane(plane);
-	omap_plane->enabled = true;
-	return omap_plane_mode_set(plane, crtc, fb,
-			crtc_x, crtc_y, crtc_w, crtc_h,
-			src_x, src_y, src_w, src_h,
-			NULL, NULL);
-}
-
-static int omap_plane_disable(struct drm_plane *plane)
-{
-	struct omap_plane *omap_plane = to_omap_plane(plane);
-	omap_plane->win.rotation = BIT(DRM_ROTATE_0);
-	return omap_plane_dpms(plane, DRM_MODE_DPMS_OFF);
-}
-
-static void omap_plane_destroy(struct drm_plane *plane)
-{
-	struct omap_plane *omap_plane = to_omap_plane(plane);
-
-	DBG("%s", omap_plane->name);
-
-	omap_irq_unregister(plane->dev, &omap_plane->error_irq);
-
-	omap_plane_disable(plane);
-	drm_plane_cleanup(plane);
-
-	WARN_ON(!kfifo_is_empty(&omap_plane->unpin_fifo));
-	kfifo_free(&omap_plane->unpin_fifo);
-
-	kfree(omap_plane);
-}
-
-int omap_plane_dpms(struct drm_plane *plane, int mode)
-{
-	struct omap_plane *omap_plane = to_omap_plane(plane);
-	bool enabled = (mode == DRM_MODE_DPMS_ON);
-	int ret = 0;
-
-	if (enabled != omap_plane->enabled) {
-		omap_plane->enabled = enabled;
-		ret = apply(plane);
-	}
-
-	return ret;
-}
-
-/* helper to install properties which are common to planes and crtcs */
-void omap_plane_install_properties(struct drm_plane *plane,
-		struct drm_mode_object *obj)
-{
-	struct drm_device *dev = plane->dev;
-	struct omap_drm_private *priv = dev->dev_private;
-	struct drm_property *prop;
-
-	if (priv->has_dmm) {
-		prop = priv->rotation_prop;
-		if (!prop) {
-			const struct drm_prop_enum_list props[] = {
-					{ DRM_ROTATE_0,   "rotate-0" },
-					{ DRM_ROTATE_90,  "rotate-90" },
-					{ DRM_ROTATE_180, "rotate-180" },
-					{ DRM_ROTATE_270, "rotate-270" },
-					{ DRM_REFLECT_X,  "reflect-x" },
-					{ DRM_REFLECT_Y,  "reflect-y" },
-			};
-			prop = drm_property_create_bitmask(dev, 0, "rotation",
-					props, ARRAY_SIZE(props));
-			if (prop == NULL)
-				return;
-			priv->rotation_prop = prop;
-		}
-		drm_object_attach_property(obj, prop, 0);
-	}
-
-	prop = priv->zorder_prop;
-	if (!prop) {
-		prop = drm_property_create_range(dev, 0, "zorder", 0, 3);
-		if (prop == NULL)
-			return;
-		priv->zorder_prop = prop;
-	}
-	drm_object_attach_property(obj, prop, 0);
-}
-
-int omap_plane_set_property(struct drm_plane *plane,
-		struct drm_property *property, uint64_t val)
-{
-	struct omap_plane *omap_plane = to_omap_plane(plane);
-	struct omap_drm_private *priv = plane->dev->dev_private;
-	int ret = -EINVAL;
-
-	if (property == priv->rotation_prop) {
-		DBG("%s: rotation: %02x", omap_plane->name, (uint32_t)val);
-		omap_plane->win.rotation = val;
-		ret = apply(plane);
-	} else if (property == priv->zorder_prop) {
-		DBG("%s: zorder: %02x", omap_plane->name, (uint32_t)val);
-		omap_plane->info.zorder = val;
-		ret = apply(plane);
-	}
-
-	return ret;
-}
-
-static const struct drm_plane_funcs omap_plane_funcs = {
-		.update_plane = omap_plane_update,
-		.disable_plane = omap_plane_disable,
-		.destroy = omap_plane_destroy,
-		.set_property = omap_plane_set_property,
-};
-
-static void omap_plane_error_irq(struct omap_drm_irq *irq, uint32_t irqstatus)
-{
-	struct omap_plane *omap_plane =
-			container_of(irq, struct omap_plane, error_irq);
-	DRM_ERROR("%s: errors: %08x\n", omap_plane->name, irqstatus);
-}
-
-static const char *plane_names[] = {
-		[OMAP_DSS_GFX] = "gfx",
-		[OMAP_DSS_VIDEO1] = "vid1",
-		[OMAP_DSS_VIDEO2] = "vid2",
-		[OMAP_DSS_VIDEO3] = "vid3",
-};
-
-static const uint32_t error_irqs[] = {
-		[OMAP_DSS_GFX] = DISPC_IRQ_GFX_FIFO_UNDERFLOW,
-		[OMAP_DSS_VIDEO1] = DISPC_IRQ_VID1_FIFO_UNDERFLOW,
-		[OMAP_DSS_VIDEO2] = DISPC_IRQ_VID2_FIFO_UNDERFLOW,
-		[OMAP_DSS_VIDEO3] = DISPC_IRQ_VID3_FIFO_UNDERFLOW,
-};
-
-/* initialize plane */
-struct drm_plane *omap_plane_init(struct drm_device *dev,
-		int id, bool private_plane)
-{
-	struct omap_drm_private *priv = dev->dev_private;
-	struct drm_plane *plane = NULL;
-	struct omap_plane *omap_plane;
-	struct omap_overlay_info *info;
-	int ret;
-
-	DBG("%s: priv=%d", plane_names[id], private_plane);
-
-	omap_plane = kzalloc(sizeof(*omap_plane), GFP_KERNEL);
-	if (!omap_plane) {
-		dev_err(dev->dev, "could not allocate plane\n");
-		goto fail;
-	}
-
-	ret = kfifo_alloc(&omap_plane->unpin_fifo, 16, GFP_KERNEL);
-	if (ret) {
-		dev_err(dev->dev, "could not allocate unpin FIFO\n");
-		goto fail;
-	}
-
-	omap_plane->nformats = omap_framebuffer_get_formats(
-			omap_plane->formats, ARRAY_SIZE(omap_plane->formats),
-			dss_feat_get_supported_color_modes(id));
-	omap_plane->id = id;
-	omap_plane->name = plane_names[id];
-
-	plane = &omap_plane->base;
-
-	omap_plane->apply.pre_apply  = omap_plane_pre_apply;
-	omap_plane->apply.post_apply = omap_plane_post_apply;
-
-	omap_plane->error_irq.irqmask = error_irqs[id];
-	omap_plane->error_irq.irq = omap_plane_error_irq;
-	omap_irq_register(dev, &omap_plane->error_irq);
-
-	drm_plane_init(dev, plane, (1 << priv->num_crtcs) - 1, &omap_plane_funcs,
-			omap_plane->formats, omap_plane->nformats, private_plane);
-
-	omap_plane_install_properties(plane, &plane->base);
-
-	/* get our starting configuration, set defaults for parameters
-	 * we don't currently use, etc:
-	 */
-	info = &omap_plane->info;
-	info->rotation_type = OMAP_DSS_ROT_DMA;
-	info->rotation = OMAP_DSS_ROT_0;
-	info->global_alpha = 0xff;
-	info->mirror = 0;
-
-	/* Set defaults depending on whether we are a CRTC or overlay
-	 * layer.
-	 * TODO add ioctl to give userspace an API to change this.. this
-	 * will come in a subsequent patch.
-	 */
-	if (private_plane)
-		omap_plane->info.zorder = 0;
-	else
-		omap_plane->info.zorder = id;
-
-	return plane;
-
-fail:
-	if (plane)
-		omap_plane_destroy(plane);
-
-	return NULL;
-}
diff --git a/drivers/staging/omapdrm/tcm-sita.c b/drivers/staging/omapdrm/tcm-sita.c
deleted file mode 100644
index efb6095..0000000
--- a/drivers/staging/omapdrm/tcm-sita.c
+++ /dev/null
@@ -1,703 +0,0 @@
-/*
- * tcm-sita.c
- *
- * SImple Tiler Allocator (SiTA): 2D and 1D allocation(reservation) algorithm
- *
- * Authors: Ravi Ramachandra <r.ramachandra@ti.com>,
- *          Lajos Molnar <molnar@ti.com>
- *
- * Copyright (C) 2009-2010 Texas Instruments, Inc.
- *
- * This package is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 as
- * published by the Free Software Foundation.
- *
- * THIS PACKAGE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR
- * IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED
- * WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE.
- *
- */
-#include <linux/slab.h>
-#include <linux/spinlock.h>
-
-#include "tcm-sita.h"
-
-#define ALIGN_DOWN(value, align) ((value) & ~((align) - 1))
-
-/* Individual selection criteria for different scan areas */
-static s32 CR_L2R_T2B = CR_BIAS_HORIZONTAL;
-static s32 CR_R2L_T2B = CR_DIAGONAL_BALANCE;
-
-/*********************************************
- *	TCM API - Sita Implementation
- *********************************************/
-static s32 sita_reserve_2d(struct tcm *tcm, u16 h, u16 w, u8 align,
-			   struct tcm_area *area);
-static s32 sita_reserve_1d(struct tcm *tcm, u32 slots, struct tcm_area *area);
-static s32 sita_free(struct tcm *tcm, struct tcm_area *area);
-static void sita_deinit(struct tcm *tcm);
-
-/*********************************************
- *	Main Scanner functions
- *********************************************/
-static s32 scan_areas_and_find_fit(struct tcm *tcm, u16 w, u16 h, u16 align,
-				   struct tcm_area *area);
-
-static s32 scan_l2r_t2b(struct tcm *tcm, u16 w, u16 h, u16 align,
-			struct tcm_area *field, struct tcm_area *area);
-
-static s32 scan_r2l_t2b(struct tcm *tcm, u16 w, u16 h, u16 align,
-			struct tcm_area *field, struct tcm_area *area);
-
-static s32 scan_r2l_b2t_one_dim(struct tcm *tcm, u32 num_slots,
-			struct tcm_area *field, struct tcm_area *area);
-
-/*********************************************
- *	Support Infrastructure Methods
- *********************************************/
-static s32 is_area_free(struct tcm_area ***map, u16 x0, u16 y0, u16 w, u16 h);
-
-static s32 update_candidate(struct tcm *tcm, u16 x0, u16 y0, u16 w, u16 h,
-			    struct tcm_area *field, s32 criteria,
-			    struct score *best);
-
-static void get_nearness_factor(struct tcm_area *field,
-				struct tcm_area *candidate,
-				struct nearness_factor *nf);
-
-static void get_neighbor_stats(struct tcm *tcm, struct tcm_area *area,
-			       struct neighbor_stats *stat);
-
-static void fill_area(struct tcm *tcm,
-				struct tcm_area *area, struct tcm_area *parent);
-
-
-/*********************************************/
-
-/*********************************************
- *	Utility Methods
- *********************************************/
-struct tcm *sita_init(u16 width, u16 height, struct tcm_pt *attr)
-{
-	struct tcm *tcm;
-	struct sita_pvt *pvt;
-	struct tcm_area area = {0};
-	s32 i;
-
-	if (width == 0 || height == 0)
-		return NULL;
-
-	tcm = kmalloc(sizeof(*tcm), GFP_KERNEL);
-	pvt = kmalloc(sizeof(*pvt), GFP_KERNEL);
-	if (!tcm || !pvt)
-		goto error;
-
-	memset(tcm, 0, sizeof(*tcm));
-	memset(pvt, 0, sizeof(*pvt));
-
-	/* Updating the pointers to SiTA implementation APIs */
-	tcm->height = height;
-	tcm->width = width;
-	tcm->reserve_2d = sita_reserve_2d;
-	tcm->reserve_1d = sita_reserve_1d;
-	tcm->free = sita_free;
-	tcm->deinit = sita_deinit;
-	tcm->pvt = (void *)pvt;
-
-	spin_lock_init(&(pvt->lock));
-
-	/* Creating tam map */
-	pvt->map = kmalloc(sizeof(*pvt->map) * tcm->width, GFP_KERNEL);
-	if (!pvt->map)
-		goto error;
-
-	for (i = 0; i < tcm->width; i++) {
-		pvt->map[i] =
-			kmalloc(sizeof(**pvt->map) * tcm->height,
-								GFP_KERNEL);
-		if (pvt->map[i] == NULL) {
-			while (i--)
-				kfree(pvt->map[i]);
-			kfree(pvt->map);
-			goto error;
-		}
-	}
-
-	if (attr && attr->x <= tcm->width && attr->y <= tcm->height) {
-		pvt->div_pt.x = attr->x;
-		pvt->div_pt.y = attr->y;
-
-	} else {
-		/* Defaulting to 3:1 ratio on width for 2D area split */
-		/* Defaulting to 3:1 ratio on height for 2D and 1D split */
-		pvt->div_pt.x = (tcm->width * 3) / 4;
-		pvt->div_pt.y = (tcm->height * 3) / 4;
-	}
-
-	spin_lock(&(pvt->lock));
-	assign(&area, 0, 0, width - 1, height - 1);
-	fill_area(tcm, &area, NULL);
-	spin_unlock(&(pvt->lock));
-	return tcm;
-
-error:
-	kfree(tcm);
-	kfree(pvt);
-	return NULL;
-}
-
-static void sita_deinit(struct tcm *tcm)
-{
-	struct sita_pvt *pvt = (struct sita_pvt *)tcm->pvt;
-	struct tcm_area area = {0};
-	s32 i;
-
-	area.p1.x = tcm->width - 1;
-	area.p1.y = tcm->height - 1;
-
-	spin_lock(&(pvt->lock));
-	fill_area(tcm, &area, NULL);
-	spin_unlock(&(pvt->lock));
-
-	for (i = 0; i < tcm->height; i++)
-		kfree(pvt->map[i]);
-	kfree(pvt->map);
-	kfree(pvt);
-}
-
-/**
- * Reserve a 1D area in the container
- *
- * @param num_slots	size of 1D area
- * @param area		pointer to the area that will be populated with the
- *			reserved area
- *
- * @return 0 on success, non-0 error value on failure.
- */
-static s32 sita_reserve_1d(struct tcm *tcm, u32 num_slots,
-			   struct tcm_area *area)
-{
-	s32 ret;
-	struct tcm_area field = {0};
-	struct sita_pvt *pvt = (struct sita_pvt *)tcm->pvt;
-
-	spin_lock(&(pvt->lock));
-
-	/* Scanning entire container */
-	assign(&field, tcm->width - 1, tcm->height - 1, 0, 0);
-
-	ret = scan_r2l_b2t_one_dim(tcm, num_slots, &field, area);
-	if (!ret)
-		/* update map */
-		fill_area(tcm, area, area);
-
-	spin_unlock(&(pvt->lock));
-	return ret;
-}
-
-/**
- * Reserve a 2D area in the container
- *
- * @param w	width
- * @param h	height
- * @param area	pointer to the area that will be populated with the reserved
- *		area
- *
- * @return 0 on success, non-0 error value on failure.
- */
-static s32 sita_reserve_2d(struct tcm *tcm, u16 h, u16 w, u8 align,
-			   struct tcm_area *area)
-{
-	s32 ret;
-	struct sita_pvt *pvt = (struct sita_pvt *)tcm->pvt;
-
-	/* not supporting more than 64 as alignment */
-	if (align > 64)
-		return -EINVAL;
-
-	/* we prefer 1, 32 and 64 as alignment */
-	align = align <= 1 ? 1 : align <= 32 ? 32 : 64;
-
-	spin_lock(&(pvt->lock));
-	ret = scan_areas_and_find_fit(tcm, w, h, align, area);
-	if (!ret)
-		/* update map */
-		fill_area(tcm, area, area);
-
-	spin_unlock(&(pvt->lock));
-	return ret;
-}
-
-/**
- * Unreserve a previously allocated 2D or 1D area
- * @param area	area to be freed
- * @return 0 - success
- */
-static s32 sita_free(struct tcm *tcm, struct tcm_area *area)
-{
-	struct sita_pvt *pvt = (struct sita_pvt *)tcm->pvt;
-
-	spin_lock(&(pvt->lock));
-
-	/* check that this is in fact an existing area */
-	WARN_ON(pvt->map[area->p0.x][area->p0.y] != area ||
-		pvt->map[area->p1.x][area->p1.y] != area);
-
-	/* Clear the contents of the associated tiles in the map */
-	fill_area(tcm, area, NULL);
-
-	spin_unlock(&(pvt->lock));
-
-	return 0;
-}
-
-/**
- * Note: In general the cordinates in the scan field area relevant to the can
- * sweep directions. The scan origin (e.g. top-left corner) will always be
- * the p0 member of the field.  Therfore, for a scan from top-left p0.x <= p1.x
- * and p0.y <= p1.y; whereas, for a scan from bottom-right p1.x <= p0.x and p1.y
- * <= p0.y
- */
-
-/**
- * Raster scan horizontally right to left from top to bottom to find a place for
- * a 2D area of given size inside a scan field.
- *
- * @param w	width of desired area
- * @param h	height of desired area
- * @param align	desired area alignment
- * @param area	pointer to the area that will be set to the best position
- * @param field	area to scan (inclusive)
- *
- * @return 0 on success, non-0 error value on failure.
- */
-static s32 scan_r2l_t2b(struct tcm *tcm, u16 w, u16 h, u16 align,
-			struct tcm_area *field, struct tcm_area *area)
-{
-	s32 x, y;
-	s16 start_x, end_x, start_y, end_y, found_x = -1;
-	struct tcm_area ***map = ((struct sita_pvt *)tcm->pvt)->map;
-	struct score best = {{0}, {0}, {0}, 0};
-
-	start_x = field->p0.x;
-	end_x = field->p1.x;
-	start_y = field->p0.y;
-	end_y = field->p1.y;
-
-	/* check scan area co-ordinates */
-	if (field->p0.x < field->p1.x ||
-	    field->p1.y < field->p0.y)
-		return -EINVAL;
-
-	/* check if allocation would fit in scan area */
-	if (w > LEN(start_x, end_x) || h > LEN(end_y, start_y))
-		return -ENOSPC;
-
-	/* adjust start_x and end_y, as allocation would not fit beyond */
-	start_x = ALIGN_DOWN(start_x - w + 1, align); /* - 1 to be inclusive */
-	end_y = end_y - h + 1;
-
-	/* check if allocation would still fit in scan area */
-	if (start_x < end_x)
-		return -ENOSPC;
-
-	/* scan field top-to-bottom, right-to-left */
-	for (y = start_y; y <= end_y; y++) {
-		for (x = start_x; x >= end_x; x -= align) {
-			if (is_area_free(map, x, y, w, h)) {
-				found_x = x;
-
-				/* update best candidate */
-				if (update_candidate(tcm, x, y, w, h, field,
-							CR_R2L_T2B, &best))
-					goto done;
-
-				/* change upper x bound */
-				end_x = x + 1;
-				break;
-			} else if (map[x][y] && map[x][y]->is2d) {
-				/* step over 2D areas */
-				x = ALIGN(map[x][y]->p0.x - w + 1, align);
-			}
-		}
-
-		/* break if you find a free area shouldering the scan field */
-		if (found_x == start_x)
-			break;
-	}
-
-	if (!best.a.tcm)
-		return -ENOSPC;
-done:
-	assign(area, best.a.p0.x, best.a.p0.y, best.a.p1.x, best.a.p1.y);
-	return 0;
-}
-
-/**
- * Raster scan horizontally left to right from top to bottom to find a place for
- * a 2D area of given size inside a scan field.
- *
- * @param w	width of desired area
- * @param h	height of desired area
- * @param align	desired area alignment
- * @param area	pointer to the area that will be set to the best position
- * @param field	area to scan (inclusive)
- *
- * @return 0 on success, non-0 error value on failure.
- */
-static s32 scan_l2r_t2b(struct tcm *tcm, u16 w, u16 h, u16 align,
-			struct tcm_area *field, struct tcm_area *area)
-{
-	s32 x, y;
-	s16 start_x, end_x, start_y, end_y, found_x = -1;
-	struct tcm_area ***map = ((struct sita_pvt *)tcm->pvt)->map;
-	struct score best = {{0}, {0}, {0}, 0};
-
-	start_x = field->p0.x;
-	end_x = field->p1.x;
-	start_y = field->p0.y;
-	end_y = field->p1.y;
-
-	/* check scan area co-ordinates */
-	if (field->p1.x < field->p0.x ||
-	    field->p1.y < field->p0.y)
-		return -EINVAL;
-
-	/* check if allocation would fit in scan area */
-	if (w > LEN(end_x, start_x) || h > LEN(end_y, start_y))
-		return -ENOSPC;
-
-	start_x = ALIGN(start_x, align);
-
-	/* check if allocation would still fit in scan area */
-	if (w > LEN(end_x, start_x))
-		return -ENOSPC;
-
-	/* adjust end_x and end_y, as allocation would not fit beyond */
-	end_x = end_x - w + 1; /* + 1 to be inclusive */
-	end_y = end_y - h + 1;
-
-	/* scan field top-to-bottom, left-to-right */
-	for (y = start_y; y <= end_y; y++) {
-		for (x = start_x; x <= end_x; x += align) {
-			if (is_area_free(map, x, y, w, h)) {
-				found_x = x;
-
-				/* update best candidate */
-				if (update_candidate(tcm, x, y, w, h, field,
-							CR_L2R_T2B, &best))
-					goto done;
-				/* change upper x bound */
-				end_x = x - 1;
-
-				break;
-			} else if (map[x][y] && map[x][y]->is2d) {
-				/* step over 2D areas */
-				x = ALIGN_DOWN(map[x][y]->p1.x, align);
-			}
-		}
-
-		/* break if you find a free area shouldering the scan field */
-		if (found_x == start_x)
-			break;
-	}
-
-	if (!best.a.tcm)
-		return -ENOSPC;
-done:
-	assign(area, best.a.p0.x, best.a.p0.y, best.a.p1.x, best.a.p1.y);
-	return 0;
-}
-
-/**
- * Raster scan horizontally right to left from bottom to top to find a place
- * for a 1D area of given size inside a scan field.
- *
- * @param num_slots	size of desired area
- * @param align		desired area alignment
- * @param area		pointer to the area that will be set to the best
- *			position
- * @param field		area to scan (inclusive)
- *
- * @return 0 on success, non-0 error value on failure.
- */
-static s32 scan_r2l_b2t_one_dim(struct tcm *tcm, u32 num_slots,
-				struct tcm_area *field, struct tcm_area *area)
-{
-	s32 found = 0;
-	s16 x, y;
-	struct sita_pvt *pvt = (struct sita_pvt *)tcm->pvt;
-	struct tcm_area *p;
-
-	/* check scan area co-ordinates */
-	if (field->p0.y < field->p1.y)
-		return -EINVAL;
-
-	/**
-	 * Currently we only support full width 1D scan field, which makes sense
-	 * since 1D slot-ordering spans the full container width.
-	 */
-	if (tcm->width != field->p0.x - field->p1.x + 1)
-		return -EINVAL;
-
-	/* check if allocation would fit in scan area */
-	if (num_slots > tcm->width * LEN(field->p0.y, field->p1.y))
-		return -ENOSPC;
-
-	x = field->p0.x;
-	y = field->p0.y;
-
-	/* find num_slots consecutive free slots to the left */
-	while (found < num_slots) {
-		if (y < 0)
-			return -ENOSPC;
-
-		/* remember bottom-right corner */
-		if (found == 0) {
-			area->p1.x = x;
-			area->p1.y = y;
-		}
-
-		/* skip busy regions */
-		p = pvt->map[x][y];
-		if (p) {
-			/* move to left of 2D areas, top left of 1D */
-			x = p->p0.x;
-			if (!p->is2d)
-				y = p->p0.y;
-
-			/* start over */
-			found = 0;
-		} else {
-			/* count consecutive free slots */
-			found++;
-			if (found == num_slots)
-				break;
-		}
-
-		/* move to the left */
-		if (x == 0)
-			y--;
-		x = (x ? : tcm->width) - 1;
-
-	}
-
-	/* set top-left corner */
-	area->p0.x = x;
-	area->p0.y = y;
-	return 0;
-}
-
-/**
- * Find a place for a 2D area of given size inside a scan field based on its
- * alignment needs.
- *
- * @param w	width of desired area
- * @param h	height of desired area
- * @param align	desired area alignment
- * @param area	pointer to the area that will be set to the best position
- *
- * @return 0 on success, non-0 error value on failure.
- */
-static s32 scan_areas_and_find_fit(struct tcm *tcm, u16 w, u16 h, u16 align,
-				   struct tcm_area *area)
-{
-	s32 ret = 0;
-	struct tcm_area field = {0};
-	u16 boundary_x, boundary_y;
-	struct sita_pvt *pvt = (struct sita_pvt *)tcm->pvt;
-
-	if (align > 1) {
-		/* prefer top-left corner */
-		boundary_x = pvt->div_pt.x - 1;
-		boundary_y = pvt->div_pt.y - 1;
-
-		/* expand width and height if needed */
-		if (w > pvt->div_pt.x)
-			boundary_x = tcm->width - 1;
-		if (h > pvt->div_pt.y)
-			boundary_y = tcm->height - 1;
-
-		assign(&field, 0, 0, boundary_x, boundary_y);
-		ret = scan_l2r_t2b(tcm, w, h, align, &field, area);
-
-		/* scan whole container if failed, but do not scan 2x */
-		if (ret != 0 && (boundary_x != tcm->width - 1 ||
-				 boundary_y != tcm->height - 1)) {
-			/* scan the entire container if nothing found */
-			assign(&field, 0, 0, tcm->width - 1, tcm->height - 1);
-			ret = scan_l2r_t2b(tcm, w, h, align, &field, area);
-		}
-	} else if (align == 1) {
-		/* prefer top-right corner */
-		boundary_x = pvt->div_pt.x;
-		boundary_y = pvt->div_pt.y - 1;
-
-		/* expand width and height if needed */
-		if (w > (tcm->width - pvt->div_pt.x))
-			boundary_x = 0;
-		if (h > pvt->div_pt.y)
-			boundary_y = tcm->height - 1;
-
-		assign(&field, tcm->width - 1, 0, boundary_x, boundary_y);
-		ret = scan_r2l_t2b(tcm, w, h, align, &field, area);
-
-		/* scan whole container if failed, but do not scan 2x */
-		if (ret != 0 && (boundary_x != 0 ||
-				 boundary_y != tcm->height - 1)) {
-			/* scan the entire container if nothing found */
-			assign(&field, tcm->width - 1, 0, 0, tcm->height - 1);
-			ret = scan_r2l_t2b(tcm, w, h, align, &field,
-					   area);
-		}
-	}
-
-	return ret;
-}
-
-/* check if an entire area is free */
-static s32 is_area_free(struct tcm_area ***map, u16 x0, u16 y0, u16 w, u16 h)
-{
-	u16 x = 0, y = 0;
-	for (y = y0; y < y0 + h; y++) {
-		for (x = x0; x < x0 + w; x++) {
-			if (map[x][y])
-				return false;
-		}
-	}
-	return true;
-}
-
-/* fills an area with a parent tcm_area */
-static void fill_area(struct tcm *tcm, struct tcm_area *area,
-			struct tcm_area *parent)
-{
-	s32 x, y;
-	struct sita_pvt *pvt = (struct sita_pvt *)tcm->pvt;
-	struct tcm_area a, a_;
-
-	/* set area's tcm; otherwise, enumerator considers it invalid */
-	area->tcm = tcm;
-
-	tcm_for_each_slice(a, *area, a_) {
-		for (x = a.p0.x; x <= a.p1.x; ++x)
-			for (y = a.p0.y; y <= a.p1.y; ++y)
-				pvt->map[x][y] = parent;
-
-	}
-}
-
-/**
- * Compares a candidate area to the current best area, and if it is a better
- * fit, it updates the best to this one.
- *
- * @param x0, y0, w, h		top, left, width, height of candidate area
- * @param field			scan field
- * @param criteria		scan criteria
- * @param best			best candidate and its scores
- *
- * @return 1 (true) if the candidate area is known to be the final best, so no
- * more searching should be performed
- */
-static s32 update_candidate(struct tcm *tcm, u16 x0, u16 y0, u16 w, u16 h,
-			    struct tcm_area *field, s32 criteria,
-			    struct score *best)
-{
-	struct score me;	/* score for area */
-
-	/*
-	 * NOTE: For horizontal bias we always give the first found, because our
-	 * scan is horizontal-raster-based and the first candidate will always
-	 * have the horizontal bias.
-	 */
-	bool first = criteria & CR_BIAS_HORIZONTAL;
-
-	assign(&me.a, x0, y0, x0 + w - 1, y0 + h - 1);
-
-	/* calculate score for current candidate */
-	if (!first) {
-		get_neighbor_stats(tcm, &me.a, &me.n);
-		me.neighs = me.n.edge + me.n.busy;
-		get_nearness_factor(field, &me.a, &me.f);
-	}
-
-	/* the 1st candidate is always the best */
-	if (!best->a.tcm)
-		goto better;
-
-	BUG_ON(first);
-
-	/* diagonal balance check */
-	if ((criteria & CR_DIAGONAL_BALANCE) &&
-		best->neighs <= me.neighs &&
-		(best->neighs < me.neighs ||
-		 /* this implies that neighs and occupied match */
-		 best->n.busy < me.n.busy ||
-		 (best->n.busy == me.n.busy &&
-		  /* check the nearness factor */
-		  best->f.x + best->f.y > me.f.x + me.f.y)))
-		goto better;
-
-	/* not better, keep going */
-	return 0;
-
-better:
-	/* save current area as best */
-	memcpy(best, &me, sizeof(me));
-	best->a.tcm = tcm;
-	return first;
-}
-
-/**
- * Calculate the nearness factor of an area in a search field.  The nearness
- * factor is smaller if the area is closer to the search origin.
- */
-static void get_nearness_factor(struct tcm_area *field, struct tcm_area *area,
-				struct nearness_factor *nf)
-{
-	/**
-	 * Using signed math as field coordinates may be reversed if
-	 * search direction is right-to-left or bottom-to-top.
-	 */
-	nf->x = (s32)(area->p0.x - field->p0.x) * 1000 /
-		(field->p1.x - field->p0.x);
-	nf->y = (s32)(area->p0.y - field->p0.y) * 1000 /
-		(field->p1.y - field->p0.y);
-}
-
-/* get neighbor statistics */
-static void get_neighbor_stats(struct tcm *tcm, struct tcm_area *area,
-			 struct neighbor_stats *stat)
-{
-	s16 x = 0, y = 0;
-	struct sita_pvt *pvt = (struct sita_pvt *)tcm->pvt;
-
-	/* Clearing any exisiting values */
-	memset(stat, 0, sizeof(*stat));
-
-	/* process top & bottom edges */
-	for (x = area->p0.x; x <= area->p1.x; x++) {
-		if (area->p0.y == 0)
-			stat->edge++;
-		else if (pvt->map[x][area->p0.y - 1])
-			stat->busy++;
-
-		if (area->p1.y == tcm->height - 1)
-			stat->edge++;
-		else if (pvt->map[x][area->p1.y + 1])
-			stat->busy++;
-	}
-
-	/* process left & right edges */
-	for (y = area->p0.y; y <= area->p1.y; ++y) {
-		if (area->p0.x == 0)
-			stat->edge++;
-		else if (pvt->map[area->p0.x - 1][y])
-			stat->busy++;
-
-		if (area->p1.x == tcm->width - 1)
-			stat->edge++;
-		else if (pvt->map[area->p1.x + 1][y])
-			stat->busy++;
-	}
-}
diff --git a/drivers/staging/omapdrm/tcm-sita.h b/drivers/staging/omapdrm/tcm-sita.h
deleted file mode 100644
index 0444f86..0000000
--- a/drivers/staging/omapdrm/tcm-sita.h
+++ /dev/null
@@ -1,95 +0,0 @@
-/*
- * tcm_sita.h
- *
- * SImple Tiler Allocator (SiTA) private structures.
- *
- * Author: Ravi Ramachandra <r.ramachandra@ti.com>
- *
- * Copyright (C) 2009-2011 Texas Instruments, Inc.
- * All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- * * Redistributions of source code must retain the above copyright
- *   notice, this list of conditions and the following disclaimer.
- *
- * * Redistributions in binary form must reproduce the above copyright
- *   notice, this list of conditions and the following disclaimer in the
- *   documentation and/or other materials provided with the distribution.
- *
- * * Neither the name of Texas Instruments Incorporated nor the names of
- *   its contributors may be used to endorse or promote products derived
- *   from this software without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
- * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO,
- * THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
- * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
- * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
- * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
- * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
- * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
- * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
- * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
- * EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- */
-
-#ifndef _TCM_SITA_H
-#define _TCM_SITA_H
-
-#include "tcm.h"
-
-/* length between two coordinates */
-#define LEN(a, b) ((a) > (b) ? (a) - (b) + 1 : (b) - (a) + 1)
-
-enum criteria {
-	CR_MAX_NEIGHS		= 0x01,
-	CR_FIRST_FOUND		= 0x10,
-	CR_BIAS_HORIZONTAL	= 0x20,
-	CR_BIAS_VERTICAL	= 0x40,
-	CR_DIAGONAL_BALANCE	= 0x80
-};
-
-/* nearness to the beginning of the search field from 0 to 1000 */
-struct nearness_factor {
-	s32 x;
-	s32 y;
-};
-
-/*
- * Statistics on immediately neighboring slots.  Edge is the number of
- * border segments that are also border segments of the scan field.  Busy
- * refers to the number of neighbors that are occupied.
- */
-struct neighbor_stats {
-	u16 edge;
-	u16 busy;
-};
-
-/* structure to keep the score of a potential allocation */
-struct score {
-	struct nearness_factor	f;
-	struct neighbor_stats	n;
-	struct tcm_area		a;
-	u16    neighs;		/* number of busy neighbors */
-};
-
-struct sita_pvt {
-	spinlock_t lock;	/* spinlock to protect access */
-	struct tcm_pt div_pt;	/* divider point splitting container */
-	struct tcm_area ***map;	/* pointers to the parent area for each slot */
-};
-
-/* assign coordinates to area */
-static inline
-void assign(struct tcm_area *a, u16 x0, u16 y0, u16 x1, u16 y1)
-{
-	a->p0.x = x0;
-	a->p0.y = y0;
-	a->p1.x = x1;
-	a->p1.y = y1;
-}
-
-#endif
diff --git a/drivers/staging/omapdrm/tcm.h b/drivers/staging/omapdrm/tcm.h
deleted file mode 100644
index a8d5ce4..0000000
--- a/drivers/staging/omapdrm/tcm.h
+++ /dev/null
@@ -1,328 +0,0 @@
-/*
- * tcm.h
- *
- * TILER container manager specification and support functions for TI
- * TILER driver.
- *
- * Author: Lajos Molnar <molnar@ti.com>
- *
- * All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- * * Redistributions of source code must retain the above copyright
- *   notice, this list of conditions and the following disclaimer.
- *
- * * Redistributions in binary form must reproduce the above copyright
- *   notice, this list of conditions and the following disclaimer in the
- *   documentation and/or other materials provided with the distribution.
- *
- * * Neither the name of Texas Instruments Incorporated nor the names of
- *   its contributors may be used to endorse or promote products derived
- *   from this software without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
- * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO,
- * THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
- * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
- * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
- * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
- * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
- * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
- * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
- * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
- * EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- */
-
-#ifndef TCM_H
-#define TCM_H
-
-struct tcm;
-
-/* point */
-struct tcm_pt {
-	u16 x;
-	u16 y;
-};
-
-/* 1d or 2d area */
-struct tcm_area {
-	bool is2d;		/* whether area is 1d or 2d */
-	struct tcm    *tcm;	/* parent */
-	struct tcm_pt  p0;
-	struct tcm_pt  p1;
-};
-
-struct tcm {
-	u16 width, height;	/* container dimensions */
-	int lut_id;		/* Lookup table identifier */
-
-	unsigned int y_offset;	/* offset to use for y coordinates */
-
-	/* 'pvt' structure shall contain any tcm details (attr) along with
-	linked list of allocated areas and mutex for mutually exclusive access
-	to the list.  It may also contain copies of width and height to notice
-	any changes to the publicly available width and height fields. */
-	void *pvt;
-
-	/* function table */
-	s32 (*reserve_2d)(struct tcm *tcm, u16 height, u16 width, u8 align,
-			  struct tcm_area *area);
-	s32 (*reserve_1d)(struct tcm *tcm, u32 slots, struct tcm_area *area);
-	s32 (*free)      (struct tcm *tcm, struct tcm_area *area);
-	void (*deinit)   (struct tcm *tcm);
-};
-
-/*=============================================================================
-    BASIC TILER CONTAINER MANAGER INTERFACE
-=============================================================================*/
-
-/*
- * NOTE:
- *
- * Since some basic parameter checking is done outside the TCM algorithms,
- * TCM implementation do NOT have to check the following:
- *
- *   area pointer is NULL
- *   width and height fits within container
- *   number of pages is more than the size of the container
- *
- */
-
-struct tcm *sita_init(u16 width, u16 height, struct tcm_pt *attr);
-
-
-/**
- * Deinitialize tiler container manager.
- *
- * @param tcm	Pointer to container manager.
- *
- * @return 0 on success, non-0 error value on error.  The call
- *	   should free as much memory as possible and meaningful
- *	   even on failure.  Some error codes: -ENODEV: invalid
- *	   manager.
- */
-static inline void tcm_deinit(struct tcm *tcm)
-{
-	if (tcm)
-		tcm->deinit(tcm);
-}
-
-/**
- * Reserves a 2D area in the container.
- *
- * @param tcm		Pointer to container manager.
- * @param height	Height(in pages) of area to be reserved.
- * @param width		Width(in pages) of area to be reserved.
- * @param align		Alignment requirement for top-left corner of area. Not
- *			all values may be supported by the container manager,
- *			but it must support 0 (1), 32 and 64.
- *			0 value is equivalent to 1.
- * @param area		Pointer to where the reserved area should be stored.
- *
- * @return 0 on success.  Non-0 error code on failure.  Also,
- *	   the tcm field of the area will be set to NULL on
- *	   failure.  Some error codes: -ENODEV: invalid manager,
- *	   -EINVAL: invalid area, -ENOMEM: not enough space for
- *	    allocation.
- */
-static inline s32 tcm_reserve_2d(struct tcm *tcm, u16 width, u16 height,
-				 u16 align, struct tcm_area *area)
-{
-	/* perform rudimentary error checking */
-	s32 res = tcm  == NULL ? -ENODEV :
-		(area == NULL || width == 0 || height == 0 ||
-		 /* align must be a 2 power */
-		 (align & (align - 1))) ? -EINVAL :
-		(height > tcm->height || width > tcm->width) ? -ENOMEM : 0;
-
-	if (!res) {
-		area->is2d = true;
-		res = tcm->reserve_2d(tcm, height, width, align, area);
-		area->tcm = res ? NULL : tcm;
-	}
-
-	return res;
-}
-
-/**
- * Reserves a 1D area in the container.
- *
- * @param tcm		Pointer to container manager.
- * @param slots		Number of (contiguous) slots to reserve.
- * @param area		Pointer to where the reserved area should be stored.
- *
- * @return 0 on success.  Non-0 error code on failure.  Also,
- *	   the tcm field of the area will be set to NULL on
- *	   failure.  Some error codes: -ENODEV: invalid manager,
- *	   -EINVAL: invalid area, -ENOMEM: not enough space for
- *	    allocation.
- */
-static inline s32 tcm_reserve_1d(struct tcm *tcm, u32 slots,
-				 struct tcm_area *area)
-{
-	/* perform rudimentary error checking */
-	s32 res = tcm  == NULL ? -ENODEV :
-		(area == NULL || slots == 0) ? -EINVAL :
-		slots > (tcm->width * (u32) tcm->height) ? -ENOMEM : 0;
-
-	if (!res) {
-		area->is2d = false;
-		res = tcm->reserve_1d(tcm, slots, area);
-		area->tcm = res ? NULL : tcm;
-	}
-
-	return res;
-}
-
-/**
- * Free a previously reserved area from the container.
- *
- * @param area	Pointer to area reserved by a prior call to
- *		tcm_reserve_1d or tcm_reserve_2d call, whether
- *		it was successful or not. (Note: all fields of
- *		the structure must match.)
- *
- * @return 0 on success.  Non-0 error code on failure.  Also, the tcm
- *	   field of the area is set to NULL on success to avoid subsequent
- *	   freeing.  This call will succeed even if supplying
- *	   the area from a failed reserved call.
- */
-static inline s32 tcm_free(struct tcm_area *area)
-{
-	s32 res = 0; /* free succeeds by default */
-
-	if (area && area->tcm) {
-		res = area->tcm->free(area->tcm, area);
-		if (res == 0)
-			area->tcm = NULL;
-	}
-
-	return res;
-}
-
-/*=============================================================================
-    HELPER FUNCTION FOR ANY TILER CONTAINER MANAGER
-=============================================================================*/
-
-/**
- * This method slices off the topmost 2D slice from the parent area, and stores
- * it in the 'slice' parameter.  The 'parent' parameter will get modified to
- * contain the remaining portion of the area.  If the whole parent area can
- * fit in a 2D slice, its tcm pointer is set to NULL to mark that it is no
- * longer a valid area.
- *
- * @param parent	Pointer to a VALID parent area that will get modified
- * @param slice		Pointer to the slice area that will get modified
- */
-static inline void tcm_slice(struct tcm_area *parent, struct tcm_area *slice)
-{
-	*slice = *parent;
-
-	/* check if we need to slice */
-	if (slice->tcm && !slice->is2d &&
-		slice->p0.y != slice->p1.y &&
-		(slice->p0.x || (slice->p1.x != slice->tcm->width - 1))) {
-		/* set end point of slice (start always remains) */
-		slice->p1.x = slice->tcm->width - 1;
-		slice->p1.y = (slice->p0.x) ? slice->p0.y : slice->p1.y - 1;
-		/* adjust remaining area */
-		parent->p0.x = 0;
-		parent->p0.y = slice->p1.y + 1;
-	} else {
-		/* mark this as the last slice */
-		parent->tcm = NULL;
-	}
-}
-
-/* Verify if a tcm area is logically valid */
-static inline bool tcm_area_is_valid(struct tcm_area *area)
-{
-	return area && area->tcm &&
-		/* coordinate bounds */
-		area->p1.x < area->tcm->width &&
-		area->p1.y < area->tcm->height &&
-		area->p0.y <= area->p1.y &&
-		/* 1D coordinate relationship + p0.x check */
-		((!area->is2d &&
-		  area->p0.x < area->tcm->width &&
-		  area->p0.x + area->p0.y * area->tcm->width <=
-		  area->p1.x + area->p1.y * area->tcm->width) ||
-		 /* 2D coordinate relationship */
-		 (area->is2d &&
-		  area->p0.x <= area->p1.x));
-}
-
-/* see if a coordinate is within an area */
-static inline bool __tcm_is_in(struct tcm_pt *p, struct tcm_area *a)
-{
-	u16 i;
-
-	if (a->is2d) {
-		return p->x >= a->p0.x && p->x <= a->p1.x &&
-		       p->y >= a->p0.y && p->y <= a->p1.y;
-	} else {
-		i = p->x + p->y * a->tcm->width;
-		return i >= a->p0.x + a->p0.y * a->tcm->width &&
-		       i <= a->p1.x + a->p1.y * a->tcm->width;
-	}
-}
-
-/* calculate area width */
-static inline u16 __tcm_area_width(struct tcm_area *area)
-{
-	return area->p1.x - area->p0.x + 1;
-}
-
-/* calculate area height */
-static inline u16 __tcm_area_height(struct tcm_area *area)
-{
-	return area->p1.y - area->p0.y + 1;
-}
-
-/* calculate number of slots in an area */
-static inline u16 __tcm_sizeof(struct tcm_area *area)
-{
-	return area->is2d ?
-		__tcm_area_width(area) * __tcm_area_height(area) :
-		(area->p1.x - area->p0.x + 1) + (area->p1.y - area->p0.y) *
-							area->tcm->width;
-}
-#define tcm_sizeof(area) __tcm_sizeof(&(area))
-#define tcm_awidth(area) __tcm_area_width(&(area))
-#define tcm_aheight(area) __tcm_area_height(&(area))
-#define tcm_is_in(pt, area) __tcm_is_in(&(pt), &(area))
-
-/* limit a 1D area to the first N pages */
-static inline s32 tcm_1d_limit(struct tcm_area *a, u32 num_pg)
-{
-	if (__tcm_sizeof(a) < num_pg)
-		return -ENOMEM;
-	if (!num_pg)
-		return -EINVAL;
-
-	a->p1.x = (a->p0.x + num_pg - 1) % a->tcm->width;
-	a->p1.y = a->p0.y + ((a->p0.x + num_pg - 1) / a->tcm->width);
-	return 0;
-}
-
-/**
- * Iterate through 2D slices of a valid area. Behaves
- * syntactically as a for(;;) statement.
- *
- * @param var		Name of a local variable of type 'struct
- *			tcm_area *' that will get modified to
- *			contain each slice.
- * @param area		Pointer to the VALID parent area. This
- *			structure will not get modified
- *			throughout the loop.
- *
- */
-#define tcm_for_each_slice(var, area, safe) \
-	for (safe = area, \
-	     tcm_slice(&safe, &var); \
-	     var.tcm; tcm_slice(&safe, &var))
-
-#endif
diff --git a/include/uapi/drm/omap_drm.h b/include/uapi/drm/omap_drm.h
new file mode 100644
index 0000000..1d0b117
--- /dev/null
+++ b/include/uapi/drm/omap_drm.h
@@ -0,0 +1,123 @@
+/*
+ * include/uapi/drm/omap_drm.h
+ *
+ * Copyright (C) 2011 Texas Instruments
+ * Author: Rob Clark <rob@ti.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef __OMAP_DRM_H__
+#define __OMAP_DRM_H__
+
+#include <drm/drm.h>
+
+/* Please note that modifications to all structs defined here are
+ * subject to backwards-compatibility constraints.
+ */
+
+#define OMAP_PARAM_CHIPSET_ID	1	/* ie. 0x3430, 0x4430, etc */
+
+struct drm_omap_param {
+	uint64_t param;			/* in */
+	uint64_t value;			/* in (set_param), out (get_param) */
+};
+
+#define OMAP_BO_SCANOUT		0x00000001	/* scanout capable (phys contiguous) */
+#define OMAP_BO_CACHE_MASK	0x00000006	/* cache type mask, see cache modes */
+#define OMAP_BO_TILED_MASK	0x00000f00	/* tiled mapping mask, see tiled modes */
+
+/* cache modes */
+#define OMAP_BO_CACHED		0x00000000	/* default */
+#define OMAP_BO_WC		0x00000002	/* write-combine */
+#define OMAP_BO_UNCACHED	0x00000004	/* strongly-ordered (uncached) */
+
+/* tiled modes */
+#define OMAP_BO_TILED_8		0x00000100
+#define OMAP_BO_TILED_16	0x00000200
+#define OMAP_BO_TILED_32	0x00000300
+#define OMAP_BO_TILED		(OMAP_BO_TILED_8 | OMAP_BO_TILED_16 | OMAP_BO_TILED_32)
+
+union omap_gem_size {
+	uint32_t bytes;		/* (for non-tiled formats) */
+	struct {
+		uint16_t width;
+		uint16_t height;
+	} tiled;		/* (for tiled formats) */
+};
+
+struct drm_omap_gem_new {
+	union omap_gem_size size;	/* in */
+	uint32_t flags;			/* in */
+	uint32_t handle;		/* out */
+	uint32_t __pad;
+};
+
+/* mask of operations: */
+enum omap_gem_op {
+	OMAP_GEM_READ = 0x01,
+	OMAP_GEM_WRITE = 0x02,
+};
+
+struct drm_omap_gem_cpu_prep {
+	uint32_t handle;		/* buffer handle (in) */
+	uint32_t op;			/* mask of omap_gem_op (in) */
+};
+
+struct drm_omap_gem_cpu_fini {
+	uint32_t handle;		/* buffer handle (in) */
+	uint32_t op;			/* mask of omap_gem_op (in) */
+	/* TODO maybe here we pass down info about what regions are touched
+	 * by sw so we can be clever about cache ops?  For now a placeholder,
+	 * set to zero and we just do full buffer flush..
+	 */
+	uint32_t nregions;
+	uint32_t __pad;
+};
+
+struct drm_omap_gem_info {
+	uint32_t handle;		/* buffer handle (in) */
+	uint32_t pad;
+	uint64_t offset;		/* mmap offset (out) */
+	/* note: in case of tiled buffers, the user virtual size can be
+	 * different from the physical size (ie. how many pages are needed
+	 * to back the object) which is returned in DRM_IOCTL_GEM_OPEN..
+	 * This size here is the one that should be used if you want to
+	 * mmap() the buffer:
+	 */
+	uint32_t size;			/* virtual size for mmap'ing (out) */
+	uint32_t __pad;
+};
+
+#define DRM_OMAP_GET_PARAM		0x00
+#define DRM_OMAP_SET_PARAM		0x01
+/* placeholder for plugin-api
+#define DRM_OMAP_GET_BASE		0x02
+*/
+#define DRM_OMAP_GEM_NEW		0x03
+#define DRM_OMAP_GEM_CPU_PREP		0x04
+#define DRM_OMAP_GEM_CPU_FINI		0x05
+#define DRM_OMAP_GEM_INFO		0x06
+#define DRM_OMAP_NUM_IOCTLS		0x07
+
+#define DRM_IOCTL_OMAP_GET_PARAM	DRM_IOWR(DRM_COMMAND_BASE + DRM_OMAP_GET_PARAM, struct drm_omap_param)
+#define DRM_IOCTL_OMAP_SET_PARAM	DRM_IOW (DRM_COMMAND_BASE + DRM_OMAP_SET_PARAM, struct drm_omap_param)
+/* placeholder for plugin-api
+#define DRM_IOCTL_OMAP_GET_BASE		DRM_IOWR(DRM_COMMAND_BASE + DRM_OMAP_GET_BASE, struct drm_omap_get_base)
+*/
+#define DRM_IOCTL_OMAP_GEM_NEW		DRM_IOWR(DRM_COMMAND_BASE + DRM_OMAP_GEM_NEW, struct drm_omap_gem_new)
+#define DRM_IOCTL_OMAP_GEM_CPU_PREP	DRM_IOW (DRM_COMMAND_BASE + DRM_OMAP_GEM_CPU_PREP, struct drm_omap_gem_cpu_prep)
+#define DRM_IOCTL_OMAP_GEM_CPU_FINI	DRM_IOW (DRM_COMMAND_BASE + DRM_OMAP_GEM_CPU_FINI, struct drm_omap_gem_cpu_fini)
+#define DRM_IOCTL_OMAP_GEM_INFO		DRM_IOWR(DRM_COMMAND_BASE + DRM_OMAP_GEM_INFO, struct drm_omap_gem_info)
+
+#endif /* __OMAP_DRM_H__ */
-- 
cgit v0.10.2


From 16ef3dfe460616f14120973f78fe640e79862654 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 24 Jan 2013 17:20:33 +0100
Subject: omapdrm: only take crtc->mutex in crtc callbacks

Omapdrm doesn't do anything nefarious with crtc load detection or has
any shared resources, so this is enough. We also need to adjust the
WARN_ON.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Rob Clark <robdclark@gmail.com>

diff --git a/drivers/gpu/drm/omapdrm/omap_crtc.c b/drivers/gpu/drm/omapdrm/omap_crtc.c
index 2b97cf9..ac2258f 100644
--- a/drivers/gpu/drm/omapdrm/omap_crtc.c
+++ b/drivers/gpu/drm/omapdrm/omap_crtc.c
@@ -274,17 +274,16 @@ static void page_flip_worker(struct work_struct *work)
 	struct omap_crtc *omap_crtc =
 			container_of(work, struct omap_crtc, page_flip_work);
 	struct drm_crtc *crtc = &omap_crtc->base;
-	struct drm_device *dev = crtc->dev;
 	struct drm_display_mode *mode = &crtc->mode;
 	struct drm_gem_object *bo;
 
-	drm_modeset_lock_all(dev);
+	mutex_lock(&crtc->mutex);
 	omap_plane_mode_set(omap_crtc->plane, crtc, crtc->fb,
 			0, 0, mode->hdisplay, mode->vdisplay,
 			crtc->x << 16, crtc->y << 16,
 			mode->hdisplay << 16, mode->vdisplay << 16,
 			vblank_cb, crtc);
-	drm_modeset_unlock_all(dev);
+	mutex_unlock(&crtc->mutex);
 
 	bo = omap_framebuffer_bo(crtc->fb, 0);
 	drm_gem_object_unreference_unlocked(bo);
@@ -417,7 +416,7 @@ static void apply_worker(struct work_struct *work)
 	 * the callbacks and list modification all serialized
 	 * with respect to modesetting ioctls from userspace.
 	 */
-	drm_modeset_lock_all(dev);
+	mutex_lock(&crtc->mutex);
 	dispc_runtime_get();
 
 	/*
@@ -462,16 +461,15 @@ static void apply_worker(struct work_struct *work)
 
 out:
 	dispc_runtime_put();
-	drm_modeset_unlock_all(dev);
+	mutex_unlock(&crtc->mutex);
 }
 
 int omap_crtc_apply(struct drm_crtc *crtc,
 		struct omap_drm_apply *apply)
 {
 	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
-	struct drm_device *dev = crtc->dev;
 
-	WARN_ON(!mutex_is_locked(&dev->mode_config.mutex));
+	WARN_ON(!mutex_is_locked(&crtc->mutex));
 
 	/* no need to queue it again if it is already queued: */
 	if (apply->queued)
-- 
cgit v0.10.2


From dfe96ddcfa22b44100814b9435770f6ff1309d37 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 24 Jan 2013 17:20:34 +0100
Subject: omapdrm: simplify locking in the fb debugfs file

We don't need to hold onto mode_config.mutex any more to keep the fb
objects around. And locking dev->struct_mutex is also not required,
since omap_gem_describe only reads data anyway. And for a debug
interface it's better to grab fewer locks in case the driver is
deadlocked already ...

The only thing we need is to hold onto mode_config.fb_lock to ensure
the user-created fbs don't disappear. The fbcon fb doesn't need any
protection, since it lives as long as the driver (and so the debugfs
files) itself. And if the teardown/setup isn't following the right
sequence grabbing locks won't prevent a NULL deref on priv->fbdev if
the fb is not yet (or no longer) there.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Rob Clark <robdclark@gmail.com>

diff --git a/drivers/gpu/drm/omapdrm/omap_debugfs.c b/drivers/gpu/drm/omapdrm/omap_debugfs.c
index c0aa40f..c27f59d 100644
--- a/drivers/gpu/drm/omapdrm/omap_debugfs.c
+++ b/drivers/gpu/drm/omapdrm/omap_debugfs.c
@@ -57,17 +57,6 @@ static int fb_show(struct seq_file *m, void *arg)
 	struct drm_device *dev = node->minor->dev;
 	struct omap_drm_private *priv = dev->dev_private;
 	struct drm_framebuffer *fb;
-	int ret;
-
-	ret = mutex_lock_interruptible(&dev->mode_config.mutex);
-	if (ret)
-		return ret;
-
-	ret = mutex_lock_interruptible(&dev->struct_mutex);
-	if (ret) {
-		mutex_unlock(&dev->mode_config.mutex);
-		return ret;
-	}
 
 	seq_printf(m, "fbcon ");
 	omap_framebuffer_describe(priv->fbdev->fb, m);
@@ -82,9 +71,6 @@ static int fb_show(struct seq_file *m, void *arg)
 	}
 	mutex_unlock(&dev->mode_config.fb_lock);
 
-	mutex_unlock(&dev->struct_mutex);
-	mutex_unlock(&dev->mode_config.mutex);
-
 	return 0;
 }
 
-- 
cgit v0.10.2


From 8e44770f093c003d11784060fea4ffc82ad576eb Mon Sep 17 00:00:00 2001
From: Rob Clark <robdclark@gmail.com>
Date: Sat, 16 Feb 2013 16:40:36 -0500
Subject: drm/omap: remove fbdev debug enter/leave hooks

This will result in badness for drivers that do not implement
mode_set_base_atomic().  So don't pretend like we can support this.

Signed-off-by: Rob Clark <robdclark@gmail.com>

diff --git a/drivers/gpu/drm/omapdrm/omap_fbdev.c b/drivers/gpu/drm/omapdrm/omap_fbdev.c
index 11eed30..f0033bd 100644
--- a/drivers/gpu/drm/omapdrm/omap_fbdev.c
+++ b/drivers/gpu/drm/omapdrm/omap_fbdev.c
@@ -131,9 +131,6 @@ static struct fb_ops omap_fb_ops = {
 	.fb_pan_display = omap_fbdev_pan_display,
 	.fb_blank = drm_fb_helper_blank,
 	.fb_setcmap = drm_fb_helper_setcmap,
-
-	.fb_debug_enter = drm_fb_helper_debug_enter,
-	.fb_debug_leave = drm_fb_helper_debug_leave,
 };
 
 static int omap_fbdev_create(struct drm_fb_helper *helper,
-- 
cgit v0.10.2


From 50fdaae79abbfc4f04b5380ac5abf00fd448c5a5 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Fri, 15 Feb 2013 10:24:35 +0000
Subject: drm/cma-helper: fixup compilation

/me grabs a few brown paper bags

So it looks like I've broken compilation in

commit 6aed8ec3f76a22217c9ae183d32b1aa990bed069
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Sun Jan 20 17:32:21 2013 +0100

    drm: review locking for drm_fb_helper_restore_fbdev_mode

Fix it up again.

Reported-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_fb_cma_helper.c b/drivers/gpu/drm/drm_fb_cma_helper.c
index e851658..ef68e34 100644
--- a/drivers/gpu/drm/drm_fb_cma_helper.c
+++ b/drivers/gpu/drm/drm_fb_cma_helper.c
@@ -377,6 +377,8 @@ EXPORT_SYMBOL_GPL(drm_fbdev_cma_fini);
  */
 void drm_fbdev_cma_restore_mode(struct drm_fbdev_cma *fbdev_cma)
 {
+	struct drm_device *dev = fbdev_cma->fb_helper.dev;
+
 	drm_modeset_lock_all(dev);
 	if (fbdev_cma)
 		drm_fb_helper_restore_fbdev_mode(&fbdev_cma->fb_helper);
-- 
cgit v0.10.2


From 6e488c00457dad0d433af7f937211fba8d76fd04 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Fri, 15 Feb 2013 20:21:37 +0000
Subject: drm: Don't set the plane->fb to NULL on successfull set_plane
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

We need to clear the local variable to get the refcounting right
(since the reference drm_mode_setplane holds is transferred to the
plane->fb pointer). But should be done _after_ we update the pointer.

Breakage introduced in

commit 6c2a75325c800de286166c693e0cd33c3a1c5ec8
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Tue Dec 11 00:59:24 2012 +0100

    drm: refcounting for sprite framebuffers

Reported-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Cc: Rob Clark <rob@ti.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index f170773..e7471b0 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -1996,9 +1996,9 @@ int drm_mode_setplane(struct drm_device *dev, void *data,
 					 plane_req->src_w, plane_req->src_h);
 	if (!ret) {
 		old_fb = plane->fb;
-		fb = NULL;
 		plane->crtc = crtc;
 		plane->fb = fb;
+		fb = NULL;
 	}
 	drm_modeset_unlock_all(dev);
 
-- 
cgit v0.10.2


From 21a245d2d62ef617978316203af032d499805cd2 Mon Sep 17 00:00:00 2001
From: Rob Clark <robdclark@gmail.com>
Date: Mon, 10 Dec 2012 10:49:46 -0600
Subject: drm: small fix in drm_send_vblank_event()

Initialize e->pipe.. some drivers set this themselves, others do not.
Setting it in drm_send_vblank_event() should help ensure more consistent
behavior with the different drivers.

Signed-off-by: Rob Clark <robdclark@gmail.com>

diff --git a/drivers/gpu/drm/drm_irq.c b/drivers/gpu/drm/drm_irq.c
index 38e7992..a6a8643 100644
--- a/drivers/gpu/drm/drm_irq.c
+++ b/drivers/gpu/drm/drm_irq.c
@@ -867,6 +867,7 @@ void drm_send_vblank_event(struct drm_device *dev, int crtc,
 
 		now = get_drm_timestamp();
 	}
+	e->pipe = crtc;
 	send_vblank_event(dev, e, seq, &now);
 }
 EXPORT_SYMBOL(drm_send_vblank_event);
-- 
cgit v0.10.2


From 6f646095ec2f5d38a6429dd896cb2d5cbbb776dc Mon Sep 17 00:00:00 2001
From: Rob Clark <robdclark@gmail.com>
Date: Mon, 10 Dec 2012 10:46:43 -0600
Subject: drm/cma: add debugfs helpers

Add helper to display fb's which can be used directly in drm_info_list:

static struct drm_info_list foo_debugfs_list[] = {
		...
		{ "fb",   drm_fb_cma_debugfs_show, 0 },
};

to display information about CMA fb objects, as well as a
drm_gem_cma_describe() which can be used if the driver bothers to keep
a list of CMA GEM objects.

Signed-off-by: Rob Clark <robdclark@gmail.com>

diff --git a/drivers/gpu/drm/drm_fb_cma_helper.c b/drivers/gpu/drm/drm_fb_cma_helper.c
index ef68e34..8044c5e 100644
--- a/drivers/gpu/drm/drm_fb_cma_helper.c
+++ b/drivers/gpu/drm/drm_fb_cma_helper.c
@@ -180,6 +180,59 @@ struct drm_gem_cma_object *drm_fb_cma_get_gem_obj(struct drm_framebuffer *fb,
 }
 EXPORT_SYMBOL_GPL(drm_fb_cma_get_gem_obj);
 
+#ifdef CONFIG_DEBUG_FS
+/**
+ * drm_fb_cma_describe() - Helper to dump information about a single
+ * CMA framebuffer object
+ */
+void drm_fb_cma_describe(struct drm_framebuffer *fb, struct seq_file *m)
+{
+	struct drm_fb_cma *fb_cma = to_fb_cma(fb);
+	int i, n = drm_format_num_planes(fb->pixel_format);
+
+	seq_printf(m, "fb: %dx%d@%4.4s\n", fb->width, fb->height,
+			(char *)&fb->pixel_format);
+
+	for (i = 0; i < n; i++) {
+		seq_printf(m, "   %d: offset=%d pitch=%d, obj: ",
+				i, fb->offsets[i], fb->pitches[i]);
+		drm_gem_cma_describe(fb_cma->obj[i], m);
+	}
+}
+EXPORT_SYMBOL_GPL(drm_fb_cma_describe);
+
+/**
+ * drm_fb_cma_debugfs_show() - Helper to list CMA framebuffer objects
+ * in debugfs.
+ */
+int drm_fb_cma_debugfs_show(struct seq_file *m, void *arg)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct drm_framebuffer *fb;
+	int ret;
+
+	ret = mutex_lock_interruptible(&dev->mode_config.mutex);
+	if (ret)
+		return ret;
+
+	ret = mutex_lock_interruptible(&dev->struct_mutex);
+	if (ret) {
+		mutex_unlock(&dev->mode_config.mutex);
+		return ret;
+	}
+
+	list_for_each_entry(fb, &dev->mode_config.fb_list, head)
+		drm_fb_cma_describe(fb, m);
+
+	mutex_unlock(&dev->struct_mutex);
+	mutex_unlock(&dev->mode_config.mutex);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(drm_fb_cma_debugfs_show);
+#endif
+
 static struct fb_ops drm_fbdev_cma_ops = {
 	.owner		= THIS_MODULE,
 	.fb_fillrect	= sys_fillrect,
diff --git a/drivers/gpu/drm/drm_gem_cma_helper.c b/drivers/gpu/drm/drm_gem_cma_helper.c
index 1aa8fee..0a7e011 100644
--- a/drivers/gpu/drm/drm_gem_cma_helper.c
+++ b/drivers/gpu/drm/drm_gem_cma_helper.c
@@ -249,3 +249,24 @@ int drm_gem_cma_dumb_destroy(struct drm_file *file_priv,
 	return drm_gem_handle_delete(file_priv, handle);
 }
 EXPORT_SYMBOL_GPL(drm_gem_cma_dumb_destroy);
+
+#ifdef CONFIG_DEBUG_FS
+void drm_gem_cma_describe(struct drm_gem_cma_object *cma_obj, struct seq_file *m)
+{
+	struct drm_gem_object *obj = &cma_obj->base;
+	struct drm_device *dev = obj->dev;
+	uint64_t off = 0;
+
+	WARN_ON(!mutex_is_locked(&dev->struct_mutex));
+
+	if (obj->map_list.map)
+		off = (uint64_t)obj->map_list.hash.key;
+
+	seq_printf(m, "%2d (%2d) %08llx %08Zx %p %d",
+			obj->name, obj->refcount.refcount.counter,
+			off, cma_obj->paddr, cma_obj->vaddr, obj->size);
+
+	seq_printf(m, "\n");
+}
+EXPORT_SYMBOL_GPL(drm_gem_cma_describe);
+#endif
diff --git a/include/drm/drm_fb_cma_helper.h b/include/drm/drm_fb_cma_helper.h
index 76c7098..4a3fc24 100644
--- a/include/drm/drm_fb_cma_helper.h
+++ b/include/drm/drm_fb_cma_helper.h
@@ -23,5 +23,10 @@ struct drm_framebuffer *drm_fb_cma_create(struct drm_device *dev,
 struct drm_gem_cma_object *drm_fb_cma_get_gem_obj(struct drm_framebuffer *fb,
 	unsigned int plane);
 
+#ifdef CONFIG_DEBUG_FS
+void drm_fb_cma_describe(struct drm_framebuffer *fb, struct seq_file *m);
+int drm_fb_cma_debugfs_show(struct seq_file *m, void *arg);
+#endif
+
 #endif
 
diff --git a/include/drm/drm_gem_cma_helper.h b/include/drm/drm_gem_cma_helper.h
index f0f6b1a..63397ce 100644
--- a/include/drm/drm_gem_cma_helper.h
+++ b/include/drm/drm_gem_cma_helper.h
@@ -41,4 +41,8 @@ struct drm_gem_cma_object *drm_gem_cma_create(struct drm_device *drm,
 
 extern const struct vm_operations_struct drm_gem_cma_vm_ops;
 
+#ifdef CONFIG_DEBUG_FS
+void drm_gem_cma_describe(struct drm_gem_cma_object *obj, struct seq_file *m);
+#endif
+
 #endif /* __DRM_GEM_CMA_HELPER_H__ */
-- 
cgit v0.10.2


From a7c47d6dc4a201b811e847e5449c8cffdc556deb Mon Sep 17 00:00:00 2001
From: Rob Clark <robdclark@gmail.com>
Date: Tue, 8 Jan 2013 17:50:48 -0600
Subject: drm: i2c encoder helper wrappers

Simplify life for drivers using an encoder-slave, so that they can make
their drm_encoder_helper_funcs const, rather than needing to dynamically
allocate and populate them.

Signed-off-by: Rob Clark <robdclark@gmail.com>

diff --git a/drivers/gpu/drm/drm_encoder_slave.c b/drivers/gpu/drm/drm_encoder_slave.c
index 63e7334..48c52f7 100644
--- a/drivers/gpu/drm/drm_encoder_slave.c
+++ b/drivers/gpu/drm/drm_encoder_slave.c
@@ -123,3 +123,66 @@ void drm_i2c_encoder_destroy(struct drm_encoder *drm_encoder)
 	module_put(module);
 }
 EXPORT_SYMBOL(drm_i2c_encoder_destroy);
+
+/*
+ * Wrapper fxns which can be plugged in to drm_encoder_helper_funcs:
+ */
+
+static inline struct drm_encoder_slave_funcs *
+get_slave_funcs(struct drm_encoder *enc)
+{
+	return to_encoder_slave(enc)->slave_funcs;
+}
+
+void drm_i2c_encoder_dpms(struct drm_encoder *encoder, int mode)
+{
+	get_slave_funcs(encoder)->dpms(encoder, mode);
+}
+EXPORT_SYMBOL(drm_i2c_encoder_dpms);
+
+bool drm_i2c_encoder_mode_fixup(struct drm_encoder *encoder,
+		const struct drm_display_mode *mode,
+		struct drm_display_mode *adjusted_mode)
+{
+	return get_slave_funcs(encoder)->mode_fixup(encoder, mode, adjusted_mode);
+}
+EXPORT_SYMBOL(drm_i2c_encoder_mode_fixup);
+
+void drm_i2c_encoder_prepare(struct drm_encoder *encoder)
+{
+	drm_i2c_encoder_dpms(encoder, DRM_MODE_DPMS_OFF);
+}
+EXPORT_SYMBOL(drm_i2c_encoder_prepare);
+
+void drm_i2c_encoder_commit(struct drm_encoder *encoder)
+{
+	drm_i2c_encoder_dpms(encoder, DRM_MODE_DPMS_ON);
+}
+EXPORT_SYMBOL(drm_i2c_encoder_commit);
+
+void drm_i2c_encoder_mode_set(struct drm_encoder *encoder,
+		struct drm_display_mode *mode,
+		struct drm_display_mode *adjusted_mode)
+{
+	get_slave_funcs(encoder)->mode_set(encoder, mode, adjusted_mode);
+}
+EXPORT_SYMBOL(drm_i2c_encoder_mode_set);
+
+enum drm_connector_status drm_i2c_encoder_detect(struct drm_encoder *encoder,
+	    struct drm_connector *connector)
+{
+	return get_slave_funcs(encoder)->detect(encoder, connector);
+}
+EXPORT_SYMBOL(drm_i2c_encoder_detect);
+
+void drm_i2c_encoder_save(struct drm_encoder *encoder)
+{
+	get_slave_funcs(encoder)->save(encoder);
+}
+EXPORT_SYMBOL(drm_i2c_encoder_save);
+
+void drm_i2c_encoder_restore(struct drm_encoder *encoder)
+{
+	get_slave_funcs(encoder)->restore(encoder);
+}
+EXPORT_SYMBOL(drm_i2c_encoder_restore);
diff --git a/include/drm/drm_encoder_slave.h b/include/drm/drm_encoder_slave.h
index b0c11a7..8b9cc36 100644
--- a/include/drm/drm_encoder_slave.h
+++ b/include/drm/drm_encoder_slave.h
@@ -159,4 +159,24 @@ static inline void drm_i2c_encoder_unregister(struct drm_i2c_encoder_driver *dri
 
 void drm_i2c_encoder_destroy(struct drm_encoder *encoder);
 
+
+/*
+ * Wrapper fxns which can be plugged in to drm_encoder_helper_funcs:
+ */
+
+void drm_i2c_encoder_dpms(struct drm_encoder *encoder, int mode);
+bool drm_i2c_encoder_mode_fixup(struct drm_encoder *encoder,
+		const struct drm_display_mode *mode,
+		struct drm_display_mode *adjusted_mode);
+void drm_i2c_encoder_prepare(struct drm_encoder *encoder);
+void drm_i2c_encoder_commit(struct drm_encoder *encoder);
+void drm_i2c_encoder_mode_set(struct drm_encoder *encoder,
+		struct drm_display_mode *mode,
+		struct drm_display_mode *adjusted_mode);
+enum drm_connector_status drm_i2c_encoder_detect(struct drm_encoder *encoder,
+	    struct drm_connector *connector);
+void drm_i2c_encoder_save(struct drm_encoder *encoder);
+void drm_i2c_encoder_restore(struct drm_encoder *encoder);
+
+
 #endif
-- 
cgit v0.10.2


From b3d3de80133d32723224329f87edecaff230fefc Mon Sep 17 00:00:00 2001
From: Rob Clark <robdclark@gmail.com>
Date: Tue, 8 Jan 2013 19:19:13 -0600
Subject: drm/nouveau: use i2c encoder helper wrappers

Signed-off-by: Rob Clark <robdclark@gmail.com>

diff --git a/drivers/gpu/drm/nouveau/nv04_tv.c b/drivers/gpu/drm/nouveau/nv04_tv.c
index 62e826a..4a69ccd 100644
--- a/drivers/gpu/drm/nouveau/nv04_tv.c
+++ b/drivers/gpu/drm/nouveau/nv04_tv.c
@@ -184,14 +184,23 @@ static const struct drm_encoder_funcs nv04_tv_funcs = {
 	.destroy = nv04_tv_destroy,
 };
 
+static const struct drm_encoder_helper_funcs nv04_tv_helper_funcs = {
+	.dpms = nv04_tv_dpms,
+	.save = drm_i2c_encoder_save,
+	.restore = drm_i2c_encoder_restore,
+	.mode_fixup = drm_i2c_encoder_mode_fixup,
+	.prepare = nv04_tv_prepare,
+	.commit = nv04_tv_commit,
+	.mode_set = nv04_tv_mode_set,
+	.detect = drm_i2c_encoder_detect,
+};
+
 int
 nv04_tv_create(struct drm_connector *connector, struct dcb_output *entry)
 {
 	struct nouveau_encoder *nv_encoder;
 	struct drm_encoder *encoder;
 	struct drm_device *dev = connector->dev;
-	struct drm_encoder_helper_funcs *hfuncs;
-	struct drm_encoder_slave_funcs *sfuncs;
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	struct nouveau_i2c *i2c = nouveau_i2c(drm->device);
 	struct nouveau_i2c_port *port = i2c->find(i2c, entry->i2c_index);
@@ -207,17 +216,11 @@ nv04_tv_create(struct drm_connector *connector, struct dcb_output *entry)
 	if (!nv_encoder)
 		return -ENOMEM;
 
-	hfuncs = kzalloc(sizeof(*hfuncs), GFP_KERNEL);
-	if (!hfuncs) {
-		ret = -ENOMEM;
-		goto fail_free;
-	}
-
 	/* Initialize the common members */
 	encoder = to_drm_encoder(nv_encoder);
 
 	drm_encoder_init(dev, encoder, &nv04_tv_funcs, DRM_MODE_ENCODER_TVDAC);
-	drm_encoder_helper_add(encoder, hfuncs);
+	drm_encoder_helper_add(encoder, &nv04_tv_helper_funcs);
 
 	encoder->possible_crtcs = entry->heads;
 	encoder->possible_clones = 0;
@@ -230,30 +233,14 @@ nv04_tv_create(struct drm_connector *connector, struct dcb_output *entry)
 	if (ret < 0)
 		goto fail_cleanup;
 
-	/* Fill the function pointers */
-	sfuncs = get_slave_funcs(encoder);
-
-	*hfuncs = (struct drm_encoder_helper_funcs) {
-		.dpms = nv04_tv_dpms,
-		.save = sfuncs->save,
-		.restore = sfuncs->restore,
-		.mode_fixup = sfuncs->mode_fixup,
-		.prepare = nv04_tv_prepare,
-		.commit = nv04_tv_commit,
-		.mode_set = nv04_tv_mode_set,
-		.detect = sfuncs->detect,
-	};
-
 	/* Attach it to the specified connector. */
-	sfuncs->create_resources(encoder, connector);
+	get_slave_funcs(encoder)->create_resources(encoder, connector);
 	drm_mode_connector_attach_encoder(connector, encoder);
 
 	return 0;
 
 fail_cleanup:
 	drm_encoder_cleanup(encoder);
-	kfree(hfuncs);
-fail_free:
 	kfree(nv_encoder);
 	return ret;
 }
-- 
cgit v0.10.2


From 8fcb6c4748822f8f594536e47b9b0701b6c33e7a Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Tue, 19 Feb 2013 11:18:04 +0100
Subject: drm/cma-helper: fixup compilation

/me grabs a few brown paper bags

So it looks like I've broken compilation in

commit 6aed8ec3f76a22217c9ae183d32b1aa990bed069
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Sun Jan 20 17:32:21 2013 +0100

    drm: review locking for drm_fb_helper_restore_fbdev_mode

Fix it up again.

v2: Only deref fbdev_cma once we're sure it's non-NULL, noticed by
Thierry Reding.

Reported-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: Thierry Reding <thierry.reding@avionic-design.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/gpu/drm/drm_fb_cma_helper.c b/drivers/gpu/drm/drm_fb_cma_helper.c
index e851658..1c8549d 100644
--- a/drivers/gpu/drm/drm_fb_cma_helper.c
+++ b/drivers/gpu/drm/drm_fb_cma_helper.c
@@ -377,10 +377,13 @@ EXPORT_SYMBOL_GPL(drm_fbdev_cma_fini);
  */
 void drm_fbdev_cma_restore_mode(struct drm_fbdev_cma *fbdev_cma)
 {
-	drm_modeset_lock_all(dev);
-	if (fbdev_cma)
+	if (fbdev_cma) {
+		struct drm_device *dev = fbdev_cma->fb_helper.dev;
+
+		drm_modeset_lock_all(dev);
 		drm_fb_helper_restore_fbdev_mode(&fbdev_cma->fb_helper);
-	drm_modeset_unlock_all(dev);
+		drm_modeset_unlock_all(dev);
+	}
 }
 EXPORT_SYMBOL_GPL(drm_fbdev_cma_restore_mode);
 
-- 
cgit v0.10.2


From 196e077dc165a307efbd9e7569f81bbdbcf18f65 Mon Sep 17 00:00:00 2001
From: Paulo Zanoni <paulo.r.zanoni@intel.com>
Date: Fri, 15 Feb 2013 13:36:27 -0200
Subject: drm: don't add inferred modes for monitors that don't support them

If bit 0 of the features byte (0x18) is set to 0, then, according to
the EDID spec, "the display is non-continuous frequency (multi-mode)
and is only specified to accept the video timing formats that are
listed in Base EDID and certain Extension Blocks".

For more information, please see the EDID spec, check the notes of the
table that explains the "Feature Support" byte (18h) and also the
notes on the tables of the section that explains "Display Range Limits
& Additional Timing Description Definition (tag #FDh)".

Cc: stable@vger.kernel.org
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=45729
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
index e1aca7b..67aa0dd 100644
--- a/drivers/gpu/drm/drm_edid.c
+++ b/drivers/gpu/drm/drm_edid.c
@@ -2054,7 +2054,8 @@ int drm_add_edid_modes(struct drm_connector *connector, struct edid *edid)
 	num_modes += add_cvt_modes(connector, edid);
 	num_modes += add_standard_modes(connector, edid);
 	num_modes += add_established_modes(connector, edid);
-	num_modes += add_inferred_modes(connector, edid);
+	if (edid->features & DRM_EDID_FEATURE_DEFAULT_GTF)
+		num_modes += add_inferred_modes(connector, edid);
 	num_modes += add_cea_modes(connector, edid);
 
 	if (quirks & (EDID_QUIRK_PREFER_LARGE_60 | EDID_QUIRK_PREFER_LARGE_75))
-- 
cgit v0.10.2


From 35f8badc1cf652381fa3f82c1fbea39f4dbe87fd Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Fri, 15 Feb 2013 21:21:37 +0100
Subject: drm: Don't set the plane->fb to NULL on successfull set_plane
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

We need to clear the local variable to get the refcounting right
(since the reference drm_mode_setplane holds is transferred to the
plane->fb pointer). But should be done _after_ we update the pointer.

Breakage introduced in

commit 6c2a75325c800de286166c693e0cd33c3a1c5ec8
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Tue Dec 11 00:59:24 2012 +0100

    drm: refcounting for sprite framebuffers

Reported-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Cc: Rob Clark <rob@ti.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Thierry Reding <thierry.reding@avionic-design.de>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index f170773..e7471b0 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -1996,9 +1996,9 @@ int drm_mode_setplane(struct drm_device *dev, void *data,
 					 plane_req->src_w, plane_req->src_h);
 	if (!ret) {
 		old_fb = plane->fb;
-		fb = NULL;
 		plane->crtc = crtc;
 		plane->fb = fb;
+		fb = NULL;
 	}
 	drm_modeset_unlock_all(dev);
 
-- 
cgit v0.10.2


From 16ea975eac671fa40a78594a116a44fef8e3f4a9 Mon Sep 17 00:00:00 2001
From: Rob Clark <robdclark@gmail.com>
Date: Tue, 8 Jan 2013 15:04:28 -0600
Subject: drm/tilcdc: add TI LCD Controller DRM driver (v4)

A simple DRM/KMS driver for the TI LCD Controller found in various
smaller TI parts (AM33xx, OMAPL138, etc).  This driver uses the
CMA helpers.  Currently only the TFP410 DVI encoder is supported
(tested with beaglebone + DVI cape).  There are also various LCD
displays, for which support can be added (as I get hw to test on),
and an external i2c HDMI encoder found on some boards.

The display controller supports a single CRTC.  And the encoder+
connector are split out into sub-devices.  Depending on which LCD
or external encoder is actually present, the appropriate output
module(s) will be loaded.

v1: original
v2: fix fb refcnting and few other cleanups
v3: get +/- vsync/hsync from timings rather than panel-info, add
    option DT max-bandwidth field so driver doesn't attempt to
    pick a display mode with too high memory bandwidth, and other
    small cleanups
v4: remove some unneeded stuff from panel-info struct, properly
    set high bits for hfp/hsw/hbp for rev 2, add DT bindings docs

Signed-off-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Tested-by: Koen Kooi <koen@dominion.thruhere.net>

diff --git a/Documentation/devicetree/bindings/drm/tilcdc/tfp410.txt b/Documentation/devicetree/bindings/drm/tilcdc/tfp410.txt
new file mode 100644
index 0000000..a58ae77
--- /dev/null
+++ b/Documentation/devicetree/bindings/drm/tilcdc/tfp410.txt
@@ -0,0 +1,21 @@
+Device-Tree bindings for tilcdc DRM TFP410 output driver
+
+Required properties:
+ - compatible: value should be "ti,tilcdc,tfp410".
+ - i2c: the phandle for the i2c device to use for DDC
+
+Recommended properties:
+ - pinctrl-names, pinctrl-0: the pincontrol settings to configure
+   muxing properly for pins that connect to TFP410 device
+ - powerdn-gpio: the powerdown GPIO, pulled low to power down the
+   TFP410 device (for DPMS_OFF)
+
+Example:
+
+	dvicape {
+		compatible = "ti,tilcdc,tfp410";
+		i2c = <&i2c2>;
+		pinctrl-names = "default";
+		pinctrl-0 = <&bone_dvi_cape_dvi_00A1_pins>;
+		powerdn-gpio = <&gpio2 31 0>;
+	};
diff --git a/Documentation/devicetree/bindings/drm/tilcdc/tilcdc.txt b/Documentation/devicetree/bindings/drm/tilcdc/tilcdc.txt
new file mode 100644
index 0000000..e5f1301
--- /dev/null
+++ b/Documentation/devicetree/bindings/drm/tilcdc/tilcdc.txt
@@ -0,0 +1,21 @@
+Device-Tree bindings for tilcdc DRM driver
+
+Required properties:
+ - compatible: value should be "ti,am33xx-tilcdc".
+ - interrupts: the interrupt number
+ - reg: base address and size of the LCDC device
+
+Recommended properties:
+ - interrupt-parent: the phandle for the interrupt controller that
+   services interrupts for this device.
+ - ti,hwmods: Name of the hwmod associated to the LCDC
+
+Example:
+
+	fb: fb@4830e000 {
+		compatible = "ti,am33xx-tilcdc";
+		reg = <0x4830e000 0x1000>;
+		interrupt-parent = <&intc>;
+		interrupts = <36>;
+		ti,hwmods = "lcdc";
+	};
diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
index ed9e3af..ba74e75 100644
--- a/drivers/gpu/drm/Kconfig
+++ b/drivers/gpu/drm/Kconfig
@@ -215,3 +215,5 @@ source "drivers/gpu/drm/cirrus/Kconfig"
 source "drivers/gpu/drm/shmobile/Kconfig"
 
 source "drivers/gpu/drm/tegra/Kconfig"
+
+source "drivers/gpu/drm/tilcdc/Kconfig"
diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile
index 6f58c81..3af934d 100644
--- a/drivers/gpu/drm/Makefile
+++ b/drivers/gpu/drm/Makefile
@@ -50,4 +50,5 @@ obj-$(CONFIG_DRM_UDL) += udl/
 obj-$(CONFIG_DRM_AST) += ast/
 obj-$(CONFIG_DRM_SHMOBILE) +=shmobile/
 obj-$(CONFIG_DRM_TEGRA) += tegra/
+obj-$(CONFIG_DRM_TILCDC)	+= tilcdc/
 obj-y			+= i2c/
diff --git a/drivers/gpu/drm/tilcdc/Kconfig b/drivers/gpu/drm/tilcdc/Kconfig
new file mode 100644
index 0000000..ee9b592
--- /dev/null
+++ b/drivers/gpu/drm/tilcdc/Kconfig
@@ -0,0 +1,10 @@
+config DRM_TILCDC
+	tristate "DRM Support for TI LCDC Display Controller"
+	depends on DRM && OF
+	select DRM_KMS_HELPER
+	select DRM_KMS_CMA_HELPER
+	select DRM_GEM_CMA_HELPER
+	help
+	  Choose this option if you have an TI SoC with LCDC display
+	  controller, for example AM33xx in beagle-bone, DA8xx, or
+	  OMAP-L1xx.  This driver replaces the FB_DA8XX fbdev driver.
diff --git a/drivers/gpu/drm/tilcdc/Makefile b/drivers/gpu/drm/tilcdc/Makefile
new file mode 100644
index 0000000..1359cc2
--- /dev/null
+++ b/drivers/gpu/drm/tilcdc/Makefile
@@ -0,0 +1,8 @@
+ccflags-y := -Iinclude/drm -Werror
+
+tilcdc-y := \
+	tilcdc_crtc.o \
+	tilcdc_tfp410.o \
+	tilcdc_drv.o
+
+obj-$(CONFIG_DRM_TILCDC)	+= tilcdc.o
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_crtc.c b/drivers/gpu/drm/tilcdc/tilcdc_crtc.c
new file mode 100644
index 0000000..5dd3c7d
--- /dev/null
+++ b/drivers/gpu/drm/tilcdc/tilcdc_crtc.c
@@ -0,0 +1,602 @@
+/*
+ * Copyright (C) 2012 Texas Instruments
+ * Author: Rob Clark <robdclark@gmail.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/kfifo.h>
+
+#include "tilcdc_drv.h"
+#include "tilcdc_regs.h"
+
+struct tilcdc_crtc {
+	struct drm_crtc base;
+
+	const struct tilcdc_panel_info *info;
+	uint32_t dirty;
+	dma_addr_t start, end;
+	struct drm_pending_vblank_event *event;
+	int dpms;
+	wait_queue_head_t frame_done_wq;
+	bool frame_done;
+
+	/* fb currently set to scanout 0/1: */
+	struct drm_framebuffer *scanout[2];
+
+	/* for deferred fb unref's: */
+	DECLARE_KFIFO_PTR(unref_fifo, struct drm_framebuffer *);
+	struct work_struct work;
+};
+#define to_tilcdc_crtc(x) container_of(x, struct tilcdc_crtc, base)
+
+static void unref_worker(struct work_struct *work)
+{
+	struct tilcdc_crtc *tilcdc_crtc = container_of(work, struct tilcdc_crtc, work);
+	struct drm_device *dev = tilcdc_crtc->base.dev;
+	struct drm_framebuffer *fb;
+
+	mutex_lock(&dev->mode_config.mutex);
+	while (kfifo_get(&tilcdc_crtc->unref_fifo, &fb))
+		drm_framebuffer_unreference(fb);
+	mutex_unlock(&dev->mode_config.mutex);
+}
+
+static void set_scanout(struct drm_crtc *crtc, int n)
+{
+	static const uint32_t base_reg[] = {
+			LCDC_DMA_FB_BASE_ADDR_0_REG, LCDC_DMA_FB_BASE_ADDR_1_REG,
+	};
+	static const uint32_t ceil_reg[] = {
+			LCDC_DMA_FB_CEILING_ADDR_0_REG, LCDC_DMA_FB_CEILING_ADDR_1_REG,
+	};
+	static const uint32_t stat[] = {
+			LCDC_END_OF_FRAME0, LCDC_END_OF_FRAME1,
+	};
+	struct tilcdc_crtc *tilcdc_crtc = to_tilcdc_crtc(crtc);
+	struct drm_device *dev = crtc->dev;
+
+	pm_runtime_get_sync(dev->dev);
+	tilcdc_write(dev, base_reg[n], tilcdc_crtc->start);
+	tilcdc_write(dev, ceil_reg[n], tilcdc_crtc->end);
+	if (tilcdc_crtc->scanout[n]) {
+		if (kfifo_put(&tilcdc_crtc->unref_fifo,
+				(const struct drm_framebuffer **)&tilcdc_crtc->scanout[n])) {
+			struct tilcdc_drm_private *priv = dev->dev_private;
+			queue_work(priv->wq, &tilcdc_crtc->work);
+		} else {
+			dev_err(dev->dev, "unref fifo full!\n");
+			drm_framebuffer_unreference(tilcdc_crtc->scanout[n]);
+		}
+	}
+	tilcdc_crtc->scanout[n] = crtc->fb;
+	drm_framebuffer_reference(tilcdc_crtc->scanout[n]);
+	tilcdc_crtc->dirty &= ~stat[n];
+	pm_runtime_put_sync(dev->dev);
+}
+
+static void update_scanout(struct drm_crtc *crtc)
+{
+	struct tilcdc_crtc *tilcdc_crtc = to_tilcdc_crtc(crtc);
+	struct drm_device *dev = crtc->dev;
+	struct drm_framebuffer *fb = crtc->fb;
+	struct drm_gem_cma_object *gem;
+	unsigned int depth, bpp;
+
+	drm_fb_get_bpp_depth(fb->pixel_format, &depth, &bpp);
+	gem = drm_fb_cma_get_gem_obj(fb, 0);
+
+	tilcdc_crtc->start = gem->paddr + fb->offsets[0] +
+			(crtc->y * fb->pitches[0]) + (crtc->x * bpp/8);
+
+	tilcdc_crtc->end = tilcdc_crtc->start +
+			(crtc->mode.vdisplay * fb->pitches[0]);
+
+	if (tilcdc_crtc->dpms == DRM_MODE_DPMS_ON) {
+		/* already enabled, so just mark the frames that need
+		 * updating and they will be updated on vblank:
+		 */
+		tilcdc_crtc->dirty |= LCDC_END_OF_FRAME0 | LCDC_END_OF_FRAME1;
+		drm_vblank_get(dev, 0);
+	} else {
+		/* not enabled yet, so update registers immediately: */
+		set_scanout(crtc, 0);
+		set_scanout(crtc, 1);
+	}
+}
+
+static void start(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	struct tilcdc_drm_private *priv = dev->dev_private;
+
+	if (priv->rev == 2) {
+		tilcdc_set(dev, LCDC_CLK_RESET_REG, LCDC_CLK_MAIN_RESET);
+		msleep(1);
+		tilcdc_clear(dev, LCDC_CLK_RESET_REG, LCDC_CLK_MAIN_RESET);
+		msleep(1);
+	}
+
+	tilcdc_set(dev, LCDC_DMA_CTRL_REG, LCDC_DUAL_FRAME_BUFFER_ENABLE);
+	tilcdc_set(dev, LCDC_RASTER_CTRL_REG, LCDC_PALETTE_LOAD_MODE(DATA_ONLY));
+	tilcdc_set(dev, LCDC_RASTER_CTRL_REG, LCDC_RASTER_ENABLE);
+}
+
+static void stop(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+
+	tilcdc_clear(dev, LCDC_RASTER_CTRL_REG, LCDC_RASTER_ENABLE);
+}
+
+static void tilcdc_crtc_destroy(struct drm_crtc *crtc)
+{
+	struct tilcdc_crtc *tilcdc_crtc = to_tilcdc_crtc(crtc);
+
+	WARN_ON(tilcdc_crtc->dpms == DRM_MODE_DPMS_ON);
+
+	drm_crtc_cleanup(crtc);
+	WARN_ON(!kfifo_is_empty(&tilcdc_crtc->unref_fifo));
+	kfifo_free(&tilcdc_crtc->unref_fifo);
+	kfree(tilcdc_crtc);
+}
+
+static int tilcdc_crtc_page_flip(struct drm_crtc *crtc,
+		struct drm_framebuffer *fb,
+		struct drm_pending_vblank_event *event)
+{
+	struct tilcdc_crtc *tilcdc_crtc = to_tilcdc_crtc(crtc);
+	struct drm_device *dev = crtc->dev;
+
+	if (tilcdc_crtc->event) {
+		dev_err(dev->dev, "already pending page flip!\n");
+		return -EBUSY;
+	}
+
+	crtc->fb = fb;
+	tilcdc_crtc->event = event;
+	update_scanout(crtc);
+
+	return 0;
+}
+
+static void tilcdc_crtc_dpms(struct drm_crtc *crtc, int mode)
+{
+	struct tilcdc_crtc *tilcdc_crtc = to_tilcdc_crtc(crtc);
+	struct drm_device *dev = crtc->dev;
+	struct tilcdc_drm_private *priv = dev->dev_private;
+
+	/* we really only care about on or off: */
+	if (mode != DRM_MODE_DPMS_ON)
+		mode = DRM_MODE_DPMS_OFF;
+
+	if (tilcdc_crtc->dpms == mode)
+		return;
+
+	tilcdc_crtc->dpms = mode;
+
+	pm_runtime_get_sync(dev->dev);
+
+	if (mode == DRM_MODE_DPMS_ON) {
+		pm_runtime_forbid(dev->dev);
+		start(crtc);
+	} else {
+		tilcdc_crtc->frame_done = false;
+		stop(crtc);
+
+		/* if necessary wait for framedone irq which will still come
+		 * before putting things to sleep..
+		 */
+		if (priv->rev == 2) {
+			int ret = wait_event_timeout(
+					tilcdc_crtc->frame_done_wq,
+					tilcdc_crtc->frame_done,
+					msecs_to_jiffies(50));
+			if (ret == 0)
+				dev_err(dev->dev, "timeout waiting for framedone\n");
+		}
+		pm_runtime_allow(dev->dev);
+	}
+
+	pm_runtime_put_sync(dev->dev);
+}
+
+static bool tilcdc_crtc_mode_fixup(struct drm_crtc *crtc,
+		const struct drm_display_mode *mode,
+		struct drm_display_mode *adjusted_mode)
+{
+	return true;
+}
+
+static void tilcdc_crtc_prepare(struct drm_crtc *crtc)
+{
+	tilcdc_crtc_dpms(crtc, DRM_MODE_DPMS_OFF);
+}
+
+static void tilcdc_crtc_commit(struct drm_crtc *crtc)
+{
+	tilcdc_crtc_dpms(crtc, DRM_MODE_DPMS_ON);
+}
+
+static int tilcdc_crtc_mode_set(struct drm_crtc *crtc,
+		struct drm_display_mode *mode,
+		struct drm_display_mode *adjusted_mode,
+		int x, int y,
+		struct drm_framebuffer *old_fb)
+{
+	struct tilcdc_crtc *tilcdc_crtc = to_tilcdc_crtc(crtc);
+	struct drm_device *dev = crtc->dev;
+	struct tilcdc_drm_private *priv = dev->dev_private;
+	const struct tilcdc_panel_info *info = tilcdc_crtc->info;
+	uint32_t reg, hbp, hfp, hsw, vbp, vfp, vsw;
+	int ret;
+
+	ret = tilcdc_crtc_mode_valid(crtc, mode);
+	if (WARN_ON(ret))
+		return ret;
+
+	if (WARN_ON(!info))
+		return -EINVAL;
+
+	pm_runtime_get_sync(dev->dev);
+
+	/* Configure the Burst Size and fifo threshold of DMA: */
+	reg = tilcdc_read(dev, LCDC_DMA_CTRL_REG) & ~0x00000770;
+	switch (info->dma_burst_sz) {
+	case 1:
+		reg |= LCDC_DMA_BURST_SIZE(LCDC_DMA_BURST_1);
+		break;
+	case 2:
+		reg |= LCDC_DMA_BURST_SIZE(LCDC_DMA_BURST_2);
+		break;
+	case 4:
+		reg |= LCDC_DMA_BURST_SIZE(LCDC_DMA_BURST_4);
+		break;
+	case 8:
+		reg |= LCDC_DMA_BURST_SIZE(LCDC_DMA_BURST_8);
+		break;
+	case 16:
+		reg |= LCDC_DMA_BURST_SIZE(LCDC_DMA_BURST_16);
+		break;
+	default:
+		return -EINVAL;
+	}
+	reg |= (info->fifo_th << 8);
+	tilcdc_write(dev, LCDC_DMA_CTRL_REG, reg);
+
+	/* Configure timings: */
+	hbp = mode->htotal - mode->hsync_end;
+	hfp = mode->hsync_start - mode->hdisplay;
+	hsw = mode->hsync_end - mode->hsync_start;
+	vbp = mode->vtotal - mode->vsync_end;
+	vfp = mode->vsync_start - mode->vdisplay;
+	vsw = mode->vsync_end - mode->vsync_start;
+
+	DBG("%dx%d, hbp=%u, hfp=%u, hsw=%u, vbp=%u, vfp=%u, vsw=%u",
+			mode->hdisplay, mode->vdisplay, hbp, hfp, hsw, vbp, vfp, vsw);
+
+	/* Configure the AC Bias Period and Number of Transitions per Interrupt: */
+	reg = tilcdc_read(dev, LCDC_RASTER_TIMING_2_REG) & ~0x000fff00;
+	reg |= LCDC_AC_BIAS_FREQUENCY(info->ac_bias) |
+		LCDC_AC_BIAS_TRANSITIONS_PER_INT(info->ac_bias_intrpt);
+	if (priv->rev == 2) {
+		reg |= (hfp & 0x300) >> 8;
+		reg |= (hbp & 0x300) >> 4;
+		reg |= (hsw & 0x3c0) << 21;
+	}
+	tilcdc_write(dev, LCDC_RASTER_TIMING_2_REG, reg);
+
+	reg = (((mode->hdisplay >> 4) - 1) << 4) |
+		((hbp & 0xff) << 24) |
+		((hfp & 0xff) << 16) |
+		((hsw & 0x3f) << 10);
+	if (priv->rev == 2)
+		reg |= (((mode->hdisplay >> 4) - 1) & 0x40) >> 3;
+	tilcdc_write(dev, LCDC_RASTER_TIMING_0_REG, reg);
+
+	reg = ((mode->vdisplay - 1) & 0x3ff) |
+		((vbp & 0xff) << 24) |
+		((vfp & 0xff) << 16) |
+		((vsw & 0x3f) << 10);
+	tilcdc_write(dev, LCDC_RASTER_TIMING_1_REG, reg);
+
+	/* Configure display type: */
+	reg = tilcdc_read(dev, LCDC_RASTER_CTRL_REG) &
+		~(LCDC_TFT_MODE | LCDC_MONO_8BIT_MODE | LCDC_MONOCHROME_MODE |
+			LCDC_V2_TFT_24BPP_MODE | LCDC_V2_TFT_24BPP_UNPACK | 0x000ff000);
+	reg |= LCDC_TFT_MODE; /* no monochrome/passive support */
+	if (info->tft_alt_mode)
+		reg |= LCDC_TFT_ALT_ENABLE;
+	if (priv->rev == 2) {
+		unsigned int depth, bpp;
+
+		drm_fb_get_bpp_depth(crtc->fb->pixel_format, &depth, &bpp);
+		switch (bpp) {
+		case 16:
+			break;
+		case 32:
+			reg |= LCDC_V2_TFT_24BPP_UNPACK;
+			/* fallthrough */
+		case 24:
+			reg |= LCDC_V2_TFT_24BPP_MODE;
+			break;
+		default:
+			dev_err(dev->dev, "invalid pixel format\n");
+			return -EINVAL;
+		}
+	}
+	reg |= info->fdd < 12;
+	tilcdc_write(dev, LCDC_RASTER_CTRL_REG, reg);
+
+	if (info->invert_pxl_clk)
+		tilcdc_set(dev, LCDC_RASTER_TIMING_2_REG, LCDC_INVERT_PIXEL_CLOCK);
+	else
+		tilcdc_clear(dev, LCDC_RASTER_TIMING_2_REG, LCDC_INVERT_PIXEL_CLOCK);
+
+	if (info->sync_ctrl)
+		tilcdc_set(dev, LCDC_RASTER_TIMING_2_REG, LCDC_SYNC_CTRL);
+	else
+		tilcdc_clear(dev, LCDC_RASTER_TIMING_2_REG, LCDC_SYNC_CTRL);
+
+	if (info->sync_edge)
+		tilcdc_set(dev, LCDC_RASTER_TIMING_2_REG, LCDC_SYNC_EDGE);
+	else
+		tilcdc_clear(dev, LCDC_RASTER_TIMING_2_REG, LCDC_SYNC_EDGE);
+
+	if (mode->flags & DRM_MODE_FLAG_NHSYNC)
+		tilcdc_set(dev, LCDC_RASTER_TIMING_2_REG, LCDC_INVERT_HSYNC);
+	else
+		tilcdc_clear(dev, LCDC_RASTER_TIMING_2_REG, LCDC_INVERT_HSYNC);
+
+	if (mode->flags & DRM_MODE_FLAG_NVSYNC)
+		tilcdc_set(dev, LCDC_RASTER_TIMING_2_REG, LCDC_INVERT_VSYNC);
+	else
+		tilcdc_clear(dev, LCDC_RASTER_TIMING_2_REG, LCDC_INVERT_VSYNC);
+
+	if (info->raster_order)
+		tilcdc_set(dev, LCDC_RASTER_CTRL_REG, LCDC_RASTER_ORDER);
+	else
+		tilcdc_clear(dev, LCDC_RASTER_CTRL_REG, LCDC_RASTER_ORDER);
+
+
+	update_scanout(crtc);
+	tilcdc_crtc_update_clk(crtc);
+
+	pm_runtime_put_sync(dev->dev);
+
+	return 0;
+}
+
+static int tilcdc_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
+		struct drm_framebuffer *old_fb)
+{
+	update_scanout(crtc);
+	return 0;
+}
+
+static void tilcdc_crtc_load_lut(struct drm_crtc *crtc)
+{
+}
+
+static const struct drm_crtc_funcs tilcdc_crtc_funcs = {
+		.destroy        = tilcdc_crtc_destroy,
+		.set_config     = drm_crtc_helper_set_config,
+		.page_flip      = tilcdc_crtc_page_flip,
+};
+
+static const struct drm_crtc_helper_funcs tilcdc_crtc_helper_funcs = {
+		.dpms           = tilcdc_crtc_dpms,
+		.mode_fixup     = tilcdc_crtc_mode_fixup,
+		.prepare        = tilcdc_crtc_prepare,
+		.commit         = tilcdc_crtc_commit,
+		.mode_set       = tilcdc_crtc_mode_set,
+		.mode_set_base  = tilcdc_crtc_mode_set_base,
+		.load_lut       = tilcdc_crtc_load_lut,
+};
+
+int tilcdc_crtc_max_width(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	struct tilcdc_drm_private *priv = dev->dev_private;
+	int max_width = 0;
+
+	if (priv->rev == 1)
+		max_width = 1024;
+	else if (priv->rev == 2)
+		max_width = 2048;
+
+	return max_width;
+}
+
+int tilcdc_crtc_mode_valid(struct drm_crtc *crtc, struct drm_display_mode *mode)
+{
+	struct tilcdc_drm_private *priv = crtc->dev->dev_private;
+	unsigned int bandwidth;
+
+	if (mode->hdisplay > tilcdc_crtc_max_width(crtc))
+		return MODE_VIRTUAL_X;
+
+	/* width must be multiple of 16 */
+	if (mode->hdisplay & 0xf)
+		return MODE_VIRTUAL_X;
+
+	if (mode->vdisplay > 2048)
+		return MODE_VIRTUAL_Y;
+
+	/* filter out modes that would require too much memory bandwidth: */
+	bandwidth = mode->hdisplay * mode->vdisplay * drm_mode_vrefresh(mode);
+	if (bandwidth > priv->max_bandwidth)
+		return MODE_BAD;
+
+	return MODE_OK;
+}
+
+void tilcdc_crtc_set_panel_info(struct drm_crtc *crtc,
+		const struct tilcdc_panel_info *info)
+{
+	struct tilcdc_crtc *tilcdc_crtc = to_tilcdc_crtc(crtc);
+	tilcdc_crtc->info = info;
+}
+
+void tilcdc_crtc_update_clk(struct drm_crtc *crtc)
+{
+	struct tilcdc_crtc *tilcdc_crtc = to_tilcdc_crtc(crtc);
+	struct drm_device *dev = crtc->dev;
+	struct tilcdc_drm_private *priv = dev->dev_private;
+	int dpms = tilcdc_crtc->dpms;
+	unsigned int lcd_clk, div;
+	int ret;
+
+	pm_runtime_get_sync(dev->dev);
+
+	if (dpms == DRM_MODE_DPMS_ON)
+		tilcdc_crtc_dpms(crtc, DRM_MODE_DPMS_OFF);
+
+	/* in raster mode, minimum divisor is 2: */
+	ret = clk_set_rate(priv->disp_clk, crtc->mode.clock * 1000 * 2);
+	if (ret) {
+		dev_err(dev->dev, "failed to set display clock rate to: %d\n",
+				crtc->mode.clock);
+		goto out;
+	}
+
+	lcd_clk = clk_get_rate(priv->clk);
+	div = lcd_clk / (crtc->mode.clock * 1000);
+
+	DBG("lcd_clk=%u, mode clock=%d, div=%u", lcd_clk, crtc->mode.clock, div);
+	DBG("fck=%lu, dpll_disp_ck=%lu", clk_get_rate(priv->clk), clk_get_rate(priv->disp_clk));
+
+	/* Configure the LCD clock divisor. */
+	tilcdc_write(dev, LCDC_CTRL_REG, LCDC_CLK_DIVISOR(div) |
+			LCDC_RASTER_MODE);
+
+	if (priv->rev == 2)
+		tilcdc_set(dev, LCDC_CLK_ENABLE_REG,
+				LCDC_V2_DMA_CLK_EN | LCDC_V2_LIDD_CLK_EN |
+				LCDC_V2_CORE_CLK_EN);
+
+	if (dpms == DRM_MODE_DPMS_ON)
+		tilcdc_crtc_dpms(crtc, DRM_MODE_DPMS_ON);
+
+out:
+	pm_runtime_put_sync(dev->dev);
+}
+
+irqreturn_t tilcdc_crtc_irq(struct drm_crtc *crtc)
+{
+	struct tilcdc_crtc *tilcdc_crtc = to_tilcdc_crtc(crtc);
+	struct drm_device *dev = crtc->dev;
+	struct tilcdc_drm_private *priv = dev->dev_private;
+	uint32_t stat = tilcdc_read_irqstatus(dev);
+
+	if ((stat & LCDC_SYNC_LOST) && (stat & LCDC_FIFO_UNDERFLOW)) {
+		stop(crtc);
+		dev_err(dev->dev, "error: %08x\n", stat);
+		tilcdc_clear_irqstatus(dev, stat);
+		start(crtc);
+	} else if (stat & LCDC_PL_LOAD_DONE) {
+		tilcdc_clear_irqstatus(dev, stat);
+	} else {
+		struct drm_pending_vblank_event *event;
+		unsigned long flags;
+		uint32_t dirty = tilcdc_crtc->dirty & stat;
+
+		tilcdc_clear_irqstatus(dev, stat);
+
+		if (dirty & LCDC_END_OF_FRAME0)
+			set_scanout(crtc, 0);
+
+		if (dirty & LCDC_END_OF_FRAME1)
+			set_scanout(crtc, 1);
+
+		drm_handle_vblank(dev, 0);
+
+		spin_lock_irqsave(&dev->event_lock, flags);
+		event = tilcdc_crtc->event;
+		tilcdc_crtc->event = NULL;
+		if (event)
+			drm_send_vblank_event(dev, 0, event);
+		spin_unlock_irqrestore(&dev->event_lock, flags);
+
+		if (dirty && !tilcdc_crtc->dirty)
+			drm_vblank_put(dev, 0);
+	}
+
+	if (priv->rev == 2) {
+		if (stat & LCDC_FRAME_DONE) {
+			tilcdc_crtc->frame_done = true;
+			wake_up(&tilcdc_crtc->frame_done_wq);
+		}
+		tilcdc_write(dev, LCDC_END_OF_INT_IND_REG, 0);
+	}
+
+	return IRQ_HANDLED;
+}
+
+void tilcdc_crtc_cancel_page_flip(struct drm_crtc *crtc, struct drm_file *file)
+{
+	struct tilcdc_crtc *tilcdc_crtc = to_tilcdc_crtc(crtc);
+	struct drm_pending_vblank_event *event;
+	struct drm_device *dev = crtc->dev;
+	unsigned long flags;
+
+	/* Destroy the pending vertical blanking event associated with the
+	 * pending page flip, if any, and disable vertical blanking interrupts.
+	 */
+	spin_lock_irqsave(&dev->event_lock, flags);
+	event = tilcdc_crtc->event;
+	if (event && event->base.file_priv == file) {
+		tilcdc_crtc->event = NULL;
+		event->base.destroy(&event->base);
+		drm_vblank_put(dev, 0);
+	}
+	spin_unlock_irqrestore(&dev->event_lock, flags);
+}
+
+struct drm_crtc *tilcdc_crtc_create(struct drm_device *dev)
+{
+	struct tilcdc_crtc *tilcdc_crtc;
+	struct drm_crtc *crtc;
+	int ret;
+
+	tilcdc_crtc = kzalloc(sizeof(*tilcdc_crtc), GFP_KERNEL);
+	if (!tilcdc_crtc) {
+		dev_err(dev->dev, "allocation failed\n");
+		return NULL;
+	}
+
+	crtc = &tilcdc_crtc->base;
+
+	tilcdc_crtc->dpms = DRM_MODE_DPMS_OFF;
+	init_waitqueue_head(&tilcdc_crtc->frame_done_wq);
+
+	ret = kfifo_alloc(&tilcdc_crtc->unref_fifo, 16, GFP_KERNEL);
+	if (ret) {
+		dev_err(dev->dev, "could not allocate unref FIFO\n");
+		goto fail;
+	}
+
+	INIT_WORK(&tilcdc_crtc->work, unref_worker);
+
+	ret = drm_crtc_init(dev, crtc, &tilcdc_crtc_funcs);
+	if (ret < 0)
+		goto fail;
+
+	drm_crtc_helper_add(crtc, &tilcdc_crtc_helper_funcs);
+
+	return crtc;
+
+fail:
+	tilcdc_crtc_destroy(crtc);
+	return NULL;
+}
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_drv.c b/drivers/gpu/drm/tilcdc/tilcdc_drv.c
new file mode 100644
index 0000000..f6defbf
--- /dev/null
+++ b/drivers/gpu/drm/tilcdc/tilcdc_drv.c
@@ -0,0 +1,605 @@
+/*
+ * Copyright (C) 2012 Texas Instruments
+ * Author: Rob Clark <robdclark@gmail.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+/* LCDC DRM driver, based on da8xx-fb */
+
+#include "tilcdc_drv.h"
+#include "tilcdc_regs.h"
+#include "tilcdc_tfp410.h"
+
+#include "drm_fb_helper.h"
+
+static LIST_HEAD(module_list);
+
+void tilcdc_module_init(struct tilcdc_module *mod, const char *name,
+		const struct tilcdc_module_ops *funcs)
+{
+	mod->name = name;
+	mod->funcs = funcs;
+	INIT_LIST_HEAD(&mod->list);
+	list_add(&mod->list, &module_list);
+}
+
+void tilcdc_module_cleanup(struct tilcdc_module *mod)
+{
+	list_del(&mod->list);
+}
+
+static struct of_device_id tilcdc_of_match[];
+
+static struct drm_framebuffer *tilcdc_fb_create(struct drm_device *dev,
+		struct drm_file *file_priv, struct drm_mode_fb_cmd2 *mode_cmd)
+{
+	return drm_fb_cma_create(dev, file_priv, mode_cmd);
+}
+
+static void tilcdc_fb_output_poll_changed(struct drm_device *dev)
+{
+	struct tilcdc_drm_private *priv = dev->dev_private;
+	if (priv->fbdev)
+		drm_fbdev_cma_hotplug_event(priv->fbdev);
+}
+
+static const struct drm_mode_config_funcs mode_config_funcs = {
+	.fb_create = tilcdc_fb_create,
+	.output_poll_changed = tilcdc_fb_output_poll_changed,
+};
+
+static int modeset_init(struct drm_device *dev)
+{
+	struct tilcdc_drm_private *priv = dev->dev_private;
+	struct tilcdc_module *mod;
+
+	drm_mode_config_init(dev);
+
+	priv->crtc = tilcdc_crtc_create(dev);
+
+	list_for_each_entry(mod, &module_list, list) {
+		DBG("loading module: %s", mod->name);
+		mod->funcs->modeset_init(mod, dev);
+	}
+
+	if ((priv->num_encoders = 0) || (priv->num_connectors == 0)) {
+		/* oh nos! */
+		dev_err(dev->dev, "no encoders/connectors found\n");
+		return -ENXIO;
+	}
+
+	dev->mode_config.min_width = 0;
+	dev->mode_config.min_height = 0;
+	dev->mode_config.max_width = tilcdc_crtc_max_width(priv->crtc);
+	dev->mode_config.max_height = 2048;
+	dev->mode_config.funcs = &mode_config_funcs;
+
+	return 0;
+}
+
+#ifdef CONFIG_CPU_FREQ
+static int cpufreq_transition(struct notifier_block *nb,
+				     unsigned long val, void *data)
+{
+	struct tilcdc_drm_private *priv = container_of(nb,
+			struct tilcdc_drm_private, freq_transition);
+	if (val == CPUFREQ_POSTCHANGE) {
+		if (priv->lcd_fck_rate != clk_get_rate(priv->clk)) {
+			priv->lcd_fck_rate = clk_get_rate(priv->clk);
+			tilcdc_crtc_update_clk(priv->crtc);
+		}
+	}
+
+	return 0;
+}
+#endif
+
+/*
+ * DRM operations:
+ */
+
+static int tilcdc_unload(struct drm_device *dev)
+{
+	struct tilcdc_drm_private *priv = dev->dev_private;
+	struct tilcdc_module *mod, *cur;
+
+	drm_kms_helper_poll_fini(dev);
+	drm_mode_config_cleanup(dev);
+	drm_vblank_cleanup(dev);
+
+	pm_runtime_get_sync(dev->dev);
+	drm_irq_uninstall(dev);
+	pm_runtime_put_sync(dev->dev);
+
+#ifdef CONFIG_CPU_FREQ
+	cpufreq_unregister_notifier(&priv->freq_transition,
+			CPUFREQ_TRANSITION_NOTIFIER);
+#endif
+
+	if (priv->clk)
+		clk_put(priv->clk);
+
+	if (priv->mmio)
+		iounmap(priv->mmio);
+
+	flush_workqueue(priv->wq);
+	destroy_workqueue(priv->wq);
+
+	dev->dev_private = NULL;
+
+	pm_runtime_disable(dev->dev);
+
+	list_for_each_entry_safe(mod, cur, &module_list, list) {
+		DBG("destroying module: %s", mod->name);
+		mod->funcs->destroy(mod);
+	}
+
+	kfree(priv);
+
+	return 0;
+}
+
+static int tilcdc_load(struct drm_device *dev, unsigned long flags)
+{
+	struct platform_device *pdev = dev->platformdev;
+	struct device_node *node = pdev->dev.of_node;
+	struct tilcdc_drm_private *priv;
+	struct resource *res;
+	int ret;
+
+	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
+	if (!priv) {
+		dev_err(dev->dev, "failed to allocate private data\n");
+		return -ENOMEM;
+	}
+
+	dev->dev_private = priv;
+
+	priv->wq = alloc_ordered_workqueue("tilcdc", 0);
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!res) {
+		dev_err(dev->dev, "failed to get memory resource\n");
+		ret = -EINVAL;
+		goto fail;
+	}
+
+	priv->mmio = ioremap_nocache(res->start, resource_size(res));
+	if (!priv->mmio) {
+		dev_err(dev->dev, "failed to ioremap\n");
+		ret = -ENOMEM;
+		goto fail;
+	}
+
+	priv->clk = clk_get(dev->dev, "fck");
+	if (IS_ERR(priv->clk)) {
+		dev_err(dev->dev, "failed to get functional clock\n");
+		ret = -ENODEV;
+		goto fail;
+	}
+
+	priv->disp_clk = clk_get(dev->dev, "dpll_disp_ck");
+	if (IS_ERR(priv->clk)) {
+		dev_err(dev->dev, "failed to get display clock\n");
+		ret = -ENODEV;
+		goto fail;
+	}
+
+#ifdef CONFIG_CPU_FREQ
+	priv->lcd_fck_rate = clk_get_rate(priv->clk);
+	priv->freq_transition.notifier_call = cpufreq_transition;
+	ret = cpufreq_register_notifier(&priv->freq_transition,
+			CPUFREQ_TRANSITION_NOTIFIER);
+	if (ret) {
+		dev_err(dev->dev, "failed to register cpufreq notifier\n");
+		goto fail;
+	}
+#endif
+
+	if (of_property_read_u32(node, "max-bandwidth", &priv->max_bandwidth))
+		priv->max_bandwidth = 1280 * 1024 * 60;
+
+	pm_runtime_enable(dev->dev);
+
+	/* Determine LCD IP Version */
+	pm_runtime_get_sync(dev->dev);
+	switch (tilcdc_read(dev, LCDC_PID_REG)) {
+	case 0x4c100102:
+		priv->rev = 1;
+		break;
+	case 0x4f200800:
+	case 0x4f201000:
+		priv->rev = 2;
+		break;
+	default:
+		dev_warn(dev->dev, "Unknown PID Reg value 0x%08x, "
+				"defaulting to LCD revision 1\n",
+				tilcdc_read(dev, LCDC_PID_REG));
+		priv->rev = 1;
+		break;
+	}
+
+	pm_runtime_put_sync(dev->dev);
+
+	ret = modeset_init(dev);
+	if (ret < 0) {
+		dev_err(dev->dev, "failed to initialize mode setting\n");
+		goto fail;
+	}
+
+	ret = drm_vblank_init(dev, 1);
+	if (ret < 0) {
+		dev_err(dev->dev, "failed to initialize vblank\n");
+		goto fail;
+	}
+
+	pm_runtime_get_sync(dev->dev);
+	ret = drm_irq_install(dev);
+	pm_runtime_put_sync(dev->dev);
+	if (ret < 0) {
+		dev_err(dev->dev, "failed to install IRQ handler\n");
+		goto fail;
+	}
+
+	platform_set_drvdata(pdev, dev);
+
+	priv->fbdev = drm_fbdev_cma_init(dev, 16,
+			dev->mode_config.num_crtc,
+			dev->mode_config.num_connector);
+
+	drm_kms_helper_poll_init(dev);
+
+	return 0;
+
+fail:
+	tilcdc_unload(dev);
+	return ret;
+}
+
+static void tilcdc_preclose(struct drm_device *dev, struct drm_file *file)
+{
+	struct tilcdc_drm_private *priv = dev->dev_private;
+
+	tilcdc_crtc_cancel_page_flip(priv->crtc, file);
+}
+
+static void tilcdc_lastclose(struct drm_device *dev)
+{
+	struct tilcdc_drm_private *priv = dev->dev_private;
+	drm_fbdev_cma_restore_mode(priv->fbdev);
+}
+
+static irqreturn_t tilcdc_irq(DRM_IRQ_ARGS)
+{
+	struct drm_device *dev = arg;
+	struct tilcdc_drm_private *priv = dev->dev_private;
+	return tilcdc_crtc_irq(priv->crtc);
+}
+
+static void tilcdc_irq_preinstall(struct drm_device *dev)
+{
+	tilcdc_clear_irqstatus(dev, 0xffffffff);
+}
+
+static int tilcdc_irq_postinstall(struct drm_device *dev)
+{
+	struct tilcdc_drm_private *priv = dev->dev_private;
+
+	/* enable FIFO underflow irq: */
+	if (priv->rev == 1) {
+		tilcdc_set(dev, LCDC_RASTER_CTRL_REG, LCDC_V1_UNDERFLOW_INT_ENA);
+	} else {
+		tilcdc_set(dev, LCDC_INT_ENABLE_SET_REG, LCDC_V2_UNDERFLOW_INT_ENA);
+	}
+
+	return 0;
+}
+
+static void tilcdc_irq_uninstall(struct drm_device *dev)
+{
+	struct tilcdc_drm_private *priv = dev->dev_private;
+
+	/* disable irqs that we might have enabled: */
+	if (priv->rev == 1) {
+		tilcdc_clear(dev, LCDC_RASTER_CTRL_REG,
+				LCDC_V1_UNDERFLOW_INT_ENA | LCDC_V1_PL_INT_ENA);
+		tilcdc_clear(dev, LCDC_DMA_CTRL_REG, LCDC_V1_END_OF_FRAME_INT_ENA);
+	} else {
+		tilcdc_clear(dev, LCDC_INT_ENABLE_SET_REG,
+			LCDC_V2_UNDERFLOW_INT_ENA | LCDC_V2_PL_INT_ENA |
+			LCDC_V2_END_OF_FRAME0_INT_ENA | LCDC_V2_END_OF_FRAME1_INT_ENA |
+			LCDC_FRAME_DONE);
+	}
+
+}
+
+static void enable_vblank(struct drm_device *dev, bool enable)
+{
+	struct tilcdc_drm_private *priv = dev->dev_private;
+	u32 reg, mask;
+
+	if (priv->rev == 1) {
+		reg = LCDC_DMA_CTRL_REG;
+		mask = LCDC_V1_END_OF_FRAME_INT_ENA;
+	} else {
+		reg = LCDC_INT_ENABLE_SET_REG;
+		mask = LCDC_V2_END_OF_FRAME0_INT_ENA |
+			LCDC_V2_END_OF_FRAME1_INT_ENA | LCDC_FRAME_DONE;
+	}
+
+	if (enable)
+		tilcdc_set(dev, reg, mask);
+	else
+		tilcdc_clear(dev, reg, mask);
+}
+
+static int tilcdc_enable_vblank(struct drm_device *dev, int crtc)
+{
+	enable_vblank(dev, true);
+	return 0;
+}
+
+static void tilcdc_disable_vblank(struct drm_device *dev, int crtc)
+{
+	enable_vblank(dev, false);
+}
+
+#if defined(CONFIG_DEBUG_FS) || defined(CONFIG_PM_SLEEP)
+static const struct {
+	const char *name;
+	uint8_t  rev;
+	uint8_t  save;
+	uint32_t reg;
+} registers[] = 	{
+#define REG(rev, save, reg) { #reg, rev, save, reg }
+		/* exists in revision 1: */
+		REG(1, false, LCDC_PID_REG),
+		REG(1, true,  LCDC_CTRL_REG),
+		REG(1, false, LCDC_STAT_REG),
+		REG(1, true,  LCDC_RASTER_CTRL_REG),
+		REG(1, true,  LCDC_RASTER_TIMING_0_REG),
+		REG(1, true,  LCDC_RASTER_TIMING_1_REG),
+		REG(1, true,  LCDC_RASTER_TIMING_2_REG),
+		REG(1, true,  LCDC_DMA_CTRL_REG),
+		REG(1, true,  LCDC_DMA_FB_BASE_ADDR_0_REG),
+		REG(1, true,  LCDC_DMA_FB_CEILING_ADDR_0_REG),
+		REG(1, true,  LCDC_DMA_FB_BASE_ADDR_1_REG),
+		REG(1, true,  LCDC_DMA_FB_CEILING_ADDR_1_REG),
+		/* new in revision 2: */
+		REG(2, false, LCDC_RAW_STAT_REG),
+		REG(2, false, LCDC_MASKED_STAT_REG),
+		REG(2, false, LCDC_INT_ENABLE_SET_REG),
+		REG(2, false, LCDC_INT_ENABLE_CLR_REG),
+		REG(2, false, LCDC_END_OF_INT_IND_REG),
+		REG(2, true,  LCDC_CLK_ENABLE_REG),
+		REG(2, true,  LCDC_INT_ENABLE_SET_REG),
+#undef REG
+};
+#endif
+
+#ifdef CONFIG_DEBUG_FS
+static int tilcdc_regs_show(struct seq_file *m, void *arg)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct tilcdc_drm_private *priv = dev->dev_private;
+	unsigned i;
+
+	pm_runtime_get_sync(dev->dev);
+
+	seq_printf(m, "revision: %d\n", priv->rev);
+
+	for (i = 0; i < ARRAY_SIZE(registers); i++)
+		if (priv->rev >= registers[i].rev)
+			seq_printf(m, "%s:\t %08x\n", registers[i].name,
+					tilcdc_read(dev, registers[i].reg));
+
+	pm_runtime_put_sync(dev->dev);
+
+	return 0;
+}
+
+static int tilcdc_mm_show(struct seq_file *m, void *arg)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	return drm_mm_dump_table(m, dev->mm_private);
+}
+
+static struct drm_info_list tilcdc_debugfs_list[] = {
+		{ "regs", tilcdc_regs_show, 0 },
+		{ "mm",   tilcdc_mm_show,   0 },
+		{ "fb",   drm_fb_cma_debugfs_show, 0 },
+};
+
+static int tilcdc_debugfs_init(struct drm_minor *minor)
+{
+	struct drm_device *dev = minor->dev;
+	struct tilcdc_module *mod;
+	int ret;
+
+	ret = drm_debugfs_create_files(tilcdc_debugfs_list,
+			ARRAY_SIZE(tilcdc_debugfs_list),
+			minor->debugfs_root, minor);
+
+	list_for_each_entry(mod, &module_list, list)
+		if (mod->funcs->debugfs_init)
+			mod->funcs->debugfs_init(mod, minor);
+
+	if (ret) {
+		dev_err(dev->dev, "could not install tilcdc_debugfs_list\n");
+		return ret;
+	}
+
+	return ret;
+}
+
+static void tilcdc_debugfs_cleanup(struct drm_minor *minor)
+{
+	struct tilcdc_module *mod;
+	drm_debugfs_remove_files(tilcdc_debugfs_list,
+			ARRAY_SIZE(tilcdc_debugfs_list), minor);
+
+	list_for_each_entry(mod, &module_list, list)
+		if (mod->funcs->debugfs_cleanup)
+			mod->funcs->debugfs_cleanup(mod, minor);
+}
+#endif
+
+static const struct file_operations fops = {
+	.owner              = THIS_MODULE,
+	.open               = drm_open,
+	.release            = drm_release,
+	.unlocked_ioctl     = drm_ioctl,
+#ifdef CONFIG_COMPAT
+	.compat_ioctl       = drm_compat_ioctl,
+#endif
+	.poll               = drm_poll,
+	.read               = drm_read,
+	.fasync             = drm_fasync,
+	.llseek             = no_llseek,
+	.mmap               = drm_gem_cma_mmap,
+};
+
+static struct drm_driver tilcdc_driver = {
+	.driver_features    = DRIVER_HAVE_IRQ | DRIVER_GEM | DRIVER_MODESET,
+	.load               = tilcdc_load,
+	.unload             = tilcdc_unload,
+	.preclose           = tilcdc_preclose,
+	.lastclose          = tilcdc_lastclose,
+	.irq_handler        = tilcdc_irq,
+	.irq_preinstall     = tilcdc_irq_preinstall,
+	.irq_postinstall    = tilcdc_irq_postinstall,
+	.irq_uninstall      = tilcdc_irq_uninstall,
+	.get_vblank_counter = drm_vblank_count,
+	.enable_vblank      = tilcdc_enable_vblank,
+	.disable_vblank     = tilcdc_disable_vblank,
+	.gem_free_object    = drm_gem_cma_free_object,
+	.gem_vm_ops         = &drm_gem_cma_vm_ops,
+	.dumb_create        = drm_gem_cma_dumb_create,
+	.dumb_map_offset    = drm_gem_cma_dumb_map_offset,
+	.dumb_destroy       = drm_gem_cma_dumb_destroy,
+#ifdef CONFIG_DEBUG_FS
+	.debugfs_init       = tilcdc_debugfs_init,
+	.debugfs_cleanup    = tilcdc_debugfs_cleanup,
+#endif
+	.fops               = &fops,
+	.name               = "tilcdc",
+	.desc               = "TI LCD Controller DRM",
+	.date               = "20121205",
+	.major              = 1,
+	.minor              = 0,
+};
+
+/*
+ * Power management:
+ */
+
+#ifdef CONFIG_PM_SLEEP
+static int tilcdc_pm_suspend(struct device *dev)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct tilcdc_drm_private *priv = ddev->dev_private;
+	unsigned i, n = 0;
+
+	drm_kms_helper_poll_disable(ddev);
+
+	/* Save register state: */
+	for (i = 0; i < ARRAY_SIZE(registers); i++)
+		if (registers[i].save && (priv->rev >= registers[i].rev))
+			priv->saved_register[n++] = tilcdc_read(ddev, registers[i].reg);
+
+	return 0;
+}
+
+static int tilcdc_pm_resume(struct device *dev)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct tilcdc_drm_private *priv = ddev->dev_private;
+	unsigned i, n = 0;
+
+	/* Restore register state: */
+	for (i = 0; i < ARRAY_SIZE(registers); i++)
+		if (registers[i].save && (priv->rev >= registers[i].rev))
+			tilcdc_write(ddev, registers[i].reg, priv->saved_register[n++]);
+
+	drm_kms_helper_poll_enable(ddev);
+
+	return 0;
+}
+#endif
+
+static const struct dev_pm_ops tilcdc_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(tilcdc_pm_suspend, tilcdc_pm_resume)
+};
+
+/*
+ * Platform driver:
+ */
+
+static int tilcdc_pdev_probe(struct platform_device *pdev)
+{
+	/* bail out early if no DT data: */
+	if (!pdev->dev.of_node) {
+		dev_err(&pdev->dev, "device-tree data is missing\n");
+		return -ENXIO;
+	}
+
+	return drm_platform_init(&tilcdc_driver, pdev);
+}
+
+static int tilcdc_pdev_remove(struct platform_device *pdev)
+{
+	drm_platform_exit(&tilcdc_driver, pdev);
+
+	return 0;
+}
+
+static struct of_device_id tilcdc_of_match[] = {
+		{ .compatible = "ti,am33xx-tilcdc", },
+		{ },
+};
+MODULE_DEVICE_TABLE(of, tilcdc_of_match);
+
+static struct platform_driver tilcdc_platform_driver = {
+	.probe      = tilcdc_pdev_probe,
+	.remove     = tilcdc_pdev_remove,
+	.driver     = {
+		.owner  = THIS_MODULE,
+		.name   = "tilcdc",
+		.pm     = &tilcdc_pm_ops,
+		.of_match_table = tilcdc_of_match,
+	},
+};
+
+static int __init tilcdc_drm_init(void)
+{
+	DBG("init");
+	tilcdc_tfp410_init();
+	return platform_driver_register(&tilcdc_platform_driver);
+}
+
+static void __exit tilcdc_drm_fini(void)
+{
+	DBG("fini");
+	tilcdc_tfp410_fini();
+	platform_driver_unregister(&tilcdc_platform_driver);
+}
+
+module_init(tilcdc_drm_init);
+module_exit(tilcdc_drm_fini);
+
+MODULE_AUTHOR("Rob Clark <robdclark@gmail.com");
+MODULE_DESCRIPTION("TI LCD Controller DRM Driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_drv.h b/drivers/gpu/drm/tilcdc/tilcdc_drv.h
new file mode 100644
index 0000000..8242b5a
--- /dev/null
+++ b/drivers/gpu/drm/tilcdc/tilcdc_drv.h
@@ -0,0 +1,150 @@
+/*
+ * Copyright (C) 2012 Texas Instruments
+ * Author: Rob Clark <robdclark@gmail.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef __TILCDC_DRV_H__
+#define __TILCDC_DRV_H__
+
+#include <linux/clk.h>
+#include <linux/cpufreq.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/pm.h>
+#include <linux/pm_runtime.h>
+#include <linux/slab.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/list.h>
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_gem_cma_helper.h>
+#include <drm/drm_fb_cma_helper.h>
+
+struct tilcdc_drm_private {
+	void __iomem *mmio;
+
+	struct clk *disp_clk;    /* display dpll */
+	struct clk *clk;         /* functional clock */
+	int rev;                 /* IP revision */
+
+	/* don't attempt resolutions w/ higher W * H * Hz: */
+	uint32_t max_bandwidth;
+
+	/* register contents saved across suspend/resume: */
+	u32 saved_register[12];
+
+#ifdef CONFIG_CPU_FREQ
+	struct notifier_block freq_transition;
+	unsigned int lcd_fck_rate;
+#endif
+
+	struct workqueue_struct *wq;
+
+	struct drm_fbdev_cma *fbdev;
+
+	struct drm_crtc *crtc;
+
+	unsigned int num_encoders;
+	struct drm_encoder *encoders[8];
+
+	unsigned int num_connectors;
+	struct drm_connector *connectors[8];
+};
+
+/* Sub-module for display.  Since we don't know at compile time what panels
+ * or display adapter(s) might be present (for ex, off chip dvi/tfp410,
+ * hdmi encoder, various lcd panels), the connector/encoder(s) are split into
+ * separate drivers.  If they are probed and found to be present, they
+ * register themselves with tilcdc_register_module().
+ */
+struct tilcdc_module;
+
+struct tilcdc_module_ops {
+	/* create appropriate encoders/connectors: */
+	int (*modeset_init)(struct tilcdc_module *mod, struct drm_device *dev);
+	void (*destroy)(struct tilcdc_module *mod);
+#ifdef CONFIG_DEBUG_FS
+	/* create debugfs nodes (can be NULL): */
+	int (*debugfs_init)(struct tilcdc_module *mod, struct drm_minor *minor);
+	/* cleanup debugfs nodes (can be NULL): */
+	void (*debugfs_cleanup)(struct tilcdc_module *mod, struct drm_minor *minor);
+#endif
+};
+
+struct tilcdc_module {
+	const char *name;
+	struct list_head list;
+	const struct tilcdc_module_ops *funcs;
+};
+
+void tilcdc_module_init(struct tilcdc_module *mod, const char *name,
+		const struct tilcdc_module_ops *funcs);
+void tilcdc_module_cleanup(struct tilcdc_module *mod);
+
+
+/* Panel config that needs to be set in the crtc, but is not coming from
+ * the mode timings.  The display module is expected to call
+ * tilcdc_crtc_set_panel_info() to set this during modeset.
+ */
+struct tilcdc_panel_info {
+
+	/* AC Bias Pin Frequency */
+	uint32_t ac_bias;
+
+	/* AC Bias Pin Transitions per Interrupt */
+	uint32_t ac_bias_intrpt;
+
+	/* DMA burst size */
+	uint32_t dma_burst_sz;
+
+	/* Bits per pixel */
+	uint32_t bpp;
+
+	/* FIFO DMA Request Delay */
+	uint32_t fdd;
+
+	/* TFT Alternative Signal Mapping (Only for active) */
+	bool tft_alt_mode;
+
+	/* Invert pixel clock */
+	bool invert_pxl_clk;
+
+	/* Horizontal and Vertical Sync Edge: 0=rising 1=falling */
+	uint32_t sync_edge;
+
+	/* Horizontal and Vertical Sync: Control: 0=ignore */
+	uint32_t sync_ctrl;
+
+	/* Raster Data Order Select: 1=Most-to-least 0=Least-to-most */
+	uint32_t raster_order;
+
+	/* DMA FIFO threshold */
+	uint32_t fifo_th;
+};
+
+#define DBG(fmt, ...) DRM_DEBUG(fmt"\n", ##__VA_ARGS__)
+
+struct drm_crtc *tilcdc_crtc_create(struct drm_device *dev);
+void tilcdc_crtc_cancel_page_flip(struct drm_crtc *crtc, struct drm_file *file);
+irqreturn_t tilcdc_crtc_irq(struct drm_crtc *crtc);
+void tilcdc_crtc_update_clk(struct drm_crtc *crtc);
+void tilcdc_crtc_set_panel_info(struct drm_crtc *crtc,
+		const struct tilcdc_panel_info *info);
+int tilcdc_crtc_mode_valid(struct drm_crtc *crtc, struct drm_display_mode *mode);
+int tilcdc_crtc_max_width(struct drm_crtc *crtc);
+
+#endif /* __TILCDC_DRV_H__ */
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_regs.h b/drivers/gpu/drm/tilcdc/tilcdc_regs.h
new file mode 100644
index 0000000..17fd1b4
--- /dev/null
+++ b/drivers/gpu/drm/tilcdc/tilcdc_regs.h
@@ -0,0 +1,154 @@
+/*
+ * Copyright (C) 2012 Texas Instruments
+ * Author: Rob Clark <robdclark@gmail.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef __TILCDC_REGS_H__
+#define __TILCDC_REGS_H__
+
+/* LCDC register definitions, based on da8xx-fb */
+
+#include <linux/bitops.h>
+
+#include "tilcdc_drv.h"
+
+/* LCDC Status Register */
+#define LCDC_END_OF_FRAME1                       BIT(9)
+#define LCDC_END_OF_FRAME0                       BIT(8)
+#define LCDC_PL_LOAD_DONE                        BIT(6)
+#define LCDC_FIFO_UNDERFLOW                      BIT(5)
+#define LCDC_SYNC_LOST                           BIT(2)
+#define LCDC_FRAME_DONE                          BIT(0)
+
+/* LCDC DMA Control Register */
+#define LCDC_DMA_BURST_SIZE(x)                   ((x) << 4)
+#define LCDC_DMA_BURST_1                         0x0
+#define LCDC_DMA_BURST_2                         0x1
+#define LCDC_DMA_BURST_4                         0x2
+#define LCDC_DMA_BURST_8                         0x3
+#define LCDC_DMA_BURST_16                        0x4
+#define LCDC_V1_END_OF_FRAME_INT_ENA             BIT(2)
+#define LCDC_V2_END_OF_FRAME0_INT_ENA            BIT(8)
+#define LCDC_V2_END_OF_FRAME1_INT_ENA            BIT(9)
+#define LCDC_DUAL_FRAME_BUFFER_ENABLE            BIT(0)
+
+/* LCDC Control Register */
+#define LCDC_CLK_DIVISOR(x)                      ((x) << 8)
+#define LCDC_RASTER_MODE                         0x01
+
+/* LCDC Raster Control Register */
+#define LCDC_PALETTE_LOAD_MODE(x)                ((x) << 20)
+#define PALETTE_AND_DATA                         0x00
+#define PALETTE_ONLY                             0x01
+#define DATA_ONLY                                0x02
+
+#define LCDC_MONO_8BIT_MODE                      BIT(9)
+#define LCDC_RASTER_ORDER                        BIT(8)
+#define LCDC_TFT_MODE                            BIT(7)
+#define LCDC_V1_UNDERFLOW_INT_ENA                BIT(6)
+#define LCDC_V2_UNDERFLOW_INT_ENA                BIT(5)
+#define LCDC_V1_PL_INT_ENA                       BIT(4)
+#define LCDC_V2_PL_INT_ENA                       BIT(6)
+#define LCDC_MONOCHROME_MODE                     BIT(1)
+#define LCDC_RASTER_ENABLE                       BIT(0)
+#define LCDC_TFT_ALT_ENABLE                      BIT(23)
+#define LCDC_STN_565_ENABLE                      BIT(24)
+#define LCDC_V2_DMA_CLK_EN                       BIT(2)
+#define LCDC_V2_LIDD_CLK_EN                      BIT(1)
+#define LCDC_V2_CORE_CLK_EN                      BIT(0)
+#define LCDC_V2_LPP_B10                          26
+#define LCDC_V2_TFT_24BPP_MODE                   BIT(25)
+#define LCDC_V2_TFT_24BPP_UNPACK                 BIT(26)
+
+/* LCDC Raster Timing 2 Register */
+#define LCDC_AC_BIAS_TRANSITIONS_PER_INT(x)      ((x) << 16)
+#define LCDC_AC_BIAS_FREQUENCY(x)                ((x) << 8)
+#define LCDC_SYNC_CTRL                           BIT(25)
+#define LCDC_SYNC_EDGE                           BIT(24)
+#define LCDC_INVERT_PIXEL_CLOCK                  BIT(22)
+#define LCDC_INVERT_HSYNC                        BIT(21)
+#define LCDC_INVERT_VSYNC                        BIT(20)
+
+/* LCDC Block */
+#define LCDC_PID_REG                             0x0
+#define LCDC_CTRL_REG                            0x4
+#define LCDC_STAT_REG                            0x8
+#define LCDC_RASTER_CTRL_REG                     0x28
+#define LCDC_RASTER_TIMING_0_REG                 0x2c
+#define LCDC_RASTER_TIMING_1_REG                 0x30
+#define LCDC_RASTER_TIMING_2_REG                 0x34
+#define LCDC_DMA_CTRL_REG                        0x40
+#define LCDC_DMA_FB_BASE_ADDR_0_REG              0x44
+#define LCDC_DMA_FB_CEILING_ADDR_0_REG           0x48
+#define LCDC_DMA_FB_BASE_ADDR_1_REG              0x4c
+#define LCDC_DMA_FB_CEILING_ADDR_1_REG           0x50
+
+/* Interrupt Registers available only in Version 2 */
+#define LCDC_RAW_STAT_REG                        0x58
+#define LCDC_MASKED_STAT_REG                     0x5c
+#define LCDC_INT_ENABLE_SET_REG                  0x60
+#define LCDC_INT_ENABLE_CLR_REG                  0x64
+#define LCDC_END_OF_INT_IND_REG                  0x68
+
+/* Clock registers available only on Version 2 */
+#define LCDC_CLK_ENABLE_REG                      0x6c
+#define LCDC_CLK_RESET_REG                       0x70
+#define LCDC_CLK_MAIN_RESET                      BIT(3)
+
+
+/*
+ * Helpers:
+ */
+
+static inline void tilcdc_write(struct drm_device *dev, u32 reg, u32 data)
+{
+	struct tilcdc_drm_private *priv = dev->dev_private;
+	iowrite32(data, priv->mmio + reg);
+}
+
+static inline u32 tilcdc_read(struct drm_device *dev, u32 reg)
+{
+	struct tilcdc_drm_private *priv = dev->dev_private;
+	return ioread32(priv->mmio + reg);
+}
+
+static inline void tilcdc_set(struct drm_device *dev, u32 reg, u32 mask)
+{
+	tilcdc_write(dev, reg, tilcdc_read(dev, reg) | mask);
+}
+
+static inline void tilcdc_clear(struct drm_device *dev, u32 reg, u32 mask)
+{
+	tilcdc_write(dev, reg, tilcdc_read(dev, reg) & ~mask);
+}
+
+/* the register to read/clear irqstatus differs between v1 and v2 of the IP */
+static inline u32 tilcdc_irqstatus_reg(struct drm_device *dev)
+{
+	struct tilcdc_drm_private *priv = dev->dev_private;
+	return (priv->rev == 2) ? LCDC_MASKED_STAT_REG : LCDC_STAT_REG;
+}
+
+static inline u32 tilcdc_read_irqstatus(struct drm_device *dev)
+{
+	return tilcdc_read(dev, tilcdc_irqstatus_reg(dev));
+}
+
+static inline void tilcdc_clear_irqstatus(struct drm_device *dev, u32 mask)
+{
+	tilcdc_write(dev, tilcdc_irqstatus_reg(dev), mask);
+}
+
+#endif /* __TILCDC_REGS_H__ */
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c b/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c
new file mode 100644
index 0000000..58d487b
--- /dev/null
+++ b/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c
@@ -0,0 +1,419 @@
+/*
+ * Copyright (C) 2012 Texas Instruments
+ * Author: Rob Clark <robdclark@gmail.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/i2c.h>
+#include <linux/of_i2c.h>
+#include <linux/gpio.h>
+#include <linux/of_gpio.h>
+#include <linux/pinctrl/pinmux.h>
+#include <linux/pinctrl/consumer.h>
+
+#include "tilcdc_drv.h"
+
+struct tfp410_module {
+	struct tilcdc_module base;
+	struct i2c_adapter *i2c;
+	int gpio;
+};
+#define to_tfp410_module(x) container_of(x, struct tfp410_module, base)
+
+
+static const struct tilcdc_panel_info dvi_info = {
+		.ac_bias                = 255,
+		.ac_bias_intrpt         = 0,
+		.dma_burst_sz           = 16,
+		.bpp                    = 16,
+		.fdd                    = 0x80,
+		.tft_alt_mode           = 0,
+		.sync_edge              = 0,
+		.sync_ctrl              = 1,
+		.raster_order           = 0,
+};
+
+/*
+ * Encoder:
+ */
+
+struct tfp410_encoder {
+	struct drm_encoder base;
+	struct tfp410_module *mod;
+	int dpms;
+};
+#define to_tfp410_encoder(x) container_of(x, struct tfp410_encoder, base)
+
+
+static void tfp410_encoder_destroy(struct drm_encoder *encoder)
+{
+	struct tfp410_encoder *tfp410_encoder = to_tfp410_encoder(encoder);
+	drm_encoder_cleanup(encoder);
+	kfree(tfp410_encoder);
+}
+
+static void tfp410_encoder_dpms(struct drm_encoder *encoder, int mode)
+{
+	struct tfp410_encoder *tfp410_encoder = to_tfp410_encoder(encoder);
+
+	if (tfp410_encoder->dpms == mode)
+		return;
+
+	if (mode == DRM_MODE_DPMS_ON) {
+		DBG("Power on");
+		gpio_direction_output(tfp410_encoder->mod->gpio, 1);
+	} else {
+		DBG("Power off");
+		gpio_direction_output(tfp410_encoder->mod->gpio, 0);
+	}
+
+	tfp410_encoder->dpms = mode;
+}
+
+static bool tfp410_encoder_mode_fixup(struct drm_encoder *encoder,
+		const struct drm_display_mode *mode,
+		struct drm_display_mode *adjusted_mode)
+{
+	/* nothing needed */
+	return true;
+}
+
+static void tfp410_encoder_prepare(struct drm_encoder *encoder)
+{
+	tfp410_encoder_dpms(encoder, DRM_MODE_DPMS_OFF);
+	tilcdc_crtc_set_panel_info(encoder->crtc, &dvi_info);
+}
+
+static void tfp410_encoder_commit(struct drm_encoder *encoder)
+{
+	tfp410_encoder_dpms(encoder, DRM_MODE_DPMS_ON);
+}
+
+static void tfp410_encoder_mode_set(struct drm_encoder *encoder,
+		struct drm_display_mode *mode,
+		struct drm_display_mode *adjusted_mode)
+{
+	/* nothing needed */
+}
+
+static const struct drm_encoder_funcs tfp410_encoder_funcs = {
+		.destroy        = tfp410_encoder_destroy,
+};
+
+static const struct drm_encoder_helper_funcs tfp410_encoder_helper_funcs = {
+		.dpms           = tfp410_encoder_dpms,
+		.mode_fixup     = tfp410_encoder_mode_fixup,
+		.prepare        = tfp410_encoder_prepare,
+		.commit         = tfp410_encoder_commit,
+		.mode_set       = tfp410_encoder_mode_set,
+};
+
+static struct drm_encoder *tfp410_encoder_create(struct drm_device *dev,
+		struct tfp410_module *mod)
+{
+	struct tfp410_encoder *tfp410_encoder;
+	struct drm_encoder *encoder;
+	int ret;
+
+	tfp410_encoder = kzalloc(sizeof(*tfp410_encoder), GFP_KERNEL);
+	if (!tfp410_encoder) {
+		dev_err(dev->dev, "allocation failed\n");
+		return NULL;
+	}
+
+	tfp410_encoder->dpms = DRM_MODE_DPMS_OFF;
+	tfp410_encoder->mod = mod;
+
+	encoder = &tfp410_encoder->base;
+	encoder->possible_crtcs = 1;
+
+	ret = drm_encoder_init(dev, encoder, &tfp410_encoder_funcs,
+			DRM_MODE_ENCODER_TMDS);
+	if (ret < 0)
+		goto fail;
+
+	drm_encoder_helper_add(encoder, &tfp410_encoder_helper_funcs);
+
+	return encoder;
+
+fail:
+	tfp410_encoder_destroy(encoder);
+	return NULL;
+}
+
+/*
+ * Connector:
+ */
+
+struct tfp410_connector {
+	struct drm_connector base;
+
+	struct drm_encoder *encoder;  /* our connected encoder */
+	struct tfp410_module *mod;
+};
+#define to_tfp410_connector(x) container_of(x, struct tfp410_connector, base)
+
+
+static void tfp410_connector_destroy(struct drm_connector *connector)
+{
+	struct tfp410_connector *tfp410_connector = to_tfp410_connector(connector);
+	drm_connector_cleanup(connector);
+	kfree(tfp410_connector);
+}
+
+static enum drm_connector_status tfp410_connector_detect(
+		struct drm_connector *connector,
+		bool force)
+{
+	struct tfp410_connector *tfp410_connector = to_tfp410_connector(connector);
+
+	if (drm_probe_ddc(tfp410_connector->mod->i2c))
+		return connector_status_connected;
+
+	return connector_status_unknown;
+}
+
+static int tfp410_connector_get_modes(struct drm_connector *connector)
+{
+	struct tfp410_connector *tfp410_connector = to_tfp410_connector(connector);
+	struct edid *edid;
+	int ret = 0;
+
+	edid = drm_get_edid(connector, tfp410_connector->mod->i2c);
+
+	drm_mode_connector_update_edid_property(connector, edid);
+
+	if (edid) {
+		ret = drm_add_edid_modes(connector, edid);
+		kfree(edid);
+	}
+
+	return ret;
+}
+
+static int tfp410_connector_mode_valid(struct drm_connector *connector,
+		  struct drm_display_mode *mode)
+{
+	struct tilcdc_drm_private *priv = connector->dev->dev_private;
+	/* our only constraints are what the crtc can generate: */
+	return tilcdc_crtc_mode_valid(priv->crtc, mode);
+}
+
+static struct drm_encoder *tfp410_connector_best_encoder(
+		struct drm_connector *connector)
+{
+	struct tfp410_connector *tfp410_connector = to_tfp410_connector(connector);
+	return tfp410_connector->encoder;
+}
+
+static const struct drm_connector_funcs tfp410_connector_funcs = {
+	.destroy            = tfp410_connector_destroy,
+	.dpms               = drm_helper_connector_dpms,
+	.detect             = tfp410_connector_detect,
+	.fill_modes         = drm_helper_probe_single_connector_modes,
+};
+
+static const struct drm_connector_helper_funcs tfp410_connector_helper_funcs = {
+	.get_modes          = tfp410_connector_get_modes,
+	.mode_valid         = tfp410_connector_mode_valid,
+	.best_encoder       = tfp410_connector_best_encoder,
+};
+
+static struct drm_connector *tfp410_connector_create(struct drm_device *dev,
+		struct tfp410_module *mod, struct drm_encoder *encoder)
+{
+	struct tfp410_connector *tfp410_connector;
+	struct drm_connector *connector;
+	int ret;
+
+	tfp410_connector = kzalloc(sizeof(*tfp410_connector), GFP_KERNEL);
+	if (!tfp410_connector) {
+		dev_err(dev->dev, "allocation failed\n");
+		return NULL;
+	}
+
+	tfp410_connector->encoder = encoder;
+	tfp410_connector->mod = mod;
+
+	connector = &tfp410_connector->base;
+
+	drm_connector_init(dev, connector, &tfp410_connector_funcs,
+			DRM_MODE_CONNECTOR_DVID);
+	drm_connector_helper_add(connector, &tfp410_connector_helper_funcs);
+
+	connector->polled = DRM_CONNECTOR_POLL_CONNECT |
+			DRM_CONNECTOR_POLL_DISCONNECT;
+
+	connector->interlace_allowed = 0;
+	connector->doublescan_allowed = 0;
+
+	ret = drm_mode_connector_attach_encoder(connector, encoder);
+	if (ret)
+		goto fail;
+
+	drm_sysfs_connector_add(connector);
+
+	return connector;
+
+fail:
+	tfp410_connector_destroy(connector);
+	return NULL;
+}
+
+/*
+ * Module:
+ */
+
+static int tfp410_modeset_init(struct tilcdc_module *mod, struct drm_device *dev)
+{
+	struct tfp410_module *tfp410_mod = to_tfp410_module(mod);
+	struct tilcdc_drm_private *priv = dev->dev_private;
+	struct drm_encoder *encoder;
+	struct drm_connector *connector;
+
+	encoder = tfp410_encoder_create(dev, tfp410_mod);
+	if (!encoder)
+		return -ENOMEM;
+
+	connector = tfp410_connector_create(dev, tfp410_mod, encoder);
+	if (!connector)
+		return -ENOMEM;
+
+	priv->encoders[priv->num_encoders++] = encoder;
+	priv->connectors[priv->num_connectors++] = connector;
+
+	return 0;
+}
+
+static void tfp410_destroy(struct tilcdc_module *mod)
+{
+	struct tfp410_module *tfp410_mod = to_tfp410_module(mod);
+
+	if (tfp410_mod->i2c)
+		i2c_put_adapter(tfp410_mod->i2c);
+
+	if (!IS_ERR_VALUE(tfp410_mod->gpio))
+		gpio_free(tfp410_mod->gpio);
+
+	tilcdc_module_cleanup(mod);
+	kfree(tfp410_mod);
+}
+
+static const struct tilcdc_module_ops tfp410_module_ops = {
+		.modeset_init = tfp410_modeset_init,
+		.destroy = tfp410_destroy,
+};
+
+/*
+ * Device:
+ */
+
+static struct of_device_id tfp410_of_match[];
+
+static int tfp410_probe(struct platform_device *pdev)
+{
+	struct device_node *node = pdev->dev.of_node;
+	struct device_node *i2c_node;
+	struct tfp410_module *tfp410_mod;
+	struct tilcdc_module *mod;
+	struct pinctrl *pinctrl;
+	uint32_t i2c_phandle;
+	int ret = -EINVAL;
+
+	/* bail out early if no DT data: */
+	if (!node) {
+		dev_err(&pdev->dev, "device-tree data is missing\n");
+		return -ENXIO;
+	}
+
+	tfp410_mod = kzalloc(sizeof(*tfp410_mod), GFP_KERNEL);
+	if (!tfp410_mod)
+		return -ENOMEM;
+
+	mod = &tfp410_mod->base;
+
+	tilcdc_module_init(mod, "tfp410", &tfp410_module_ops);
+
+	pinctrl = devm_pinctrl_get_select_default(&pdev->dev);
+	if (IS_ERR(pinctrl))
+		dev_warn(&pdev->dev, "pins are not configured\n");
+
+	if (of_property_read_u32(node, "i2c", &i2c_phandle)) {
+		dev_err(&pdev->dev, "could not get i2c bus phandle\n");
+		goto fail;
+	}
+
+	i2c_node = of_find_node_by_phandle(i2c_phandle);
+	if (!i2c_node) {
+		dev_err(&pdev->dev, "could not get i2c bus node\n");
+		goto fail;
+	}
+
+	tfp410_mod->i2c = of_find_i2c_adapter_by_node(i2c_node);
+	if (!tfp410_mod->i2c) {
+		dev_err(&pdev->dev, "could not get i2c\n");
+		goto fail;
+	}
+
+	of_node_put(i2c_node);
+
+	tfp410_mod->gpio = of_get_named_gpio_flags(node, "powerdn-gpio",
+			0, NULL);
+	if (IS_ERR_VALUE(tfp410_mod->gpio)) {
+		dev_warn(&pdev->dev, "No power down GPIO\n");
+	} else {
+		ret = gpio_request(tfp410_mod->gpio, "DVI_PDn");
+		if (ret) {
+			dev_err(&pdev->dev, "could not get DVI_PDn gpio\n");
+			goto fail;
+		}
+	}
+
+	return 0;
+
+fail:
+	tfp410_destroy(mod);
+	return ret;
+}
+
+static int tfp410_remove(struct platform_device *pdev)
+{
+	return 0;
+}
+
+static struct of_device_id tfp410_of_match[] = {
+		{ .compatible = "ti,tilcdc,tfp410", },
+		{ },
+};
+MODULE_DEVICE_TABLE(of, tfp410_of_match);
+
+struct platform_driver tfp410_driver = {
+	.probe = tfp410_probe,
+	.remove = tfp410_remove,
+	.driver = {
+		.owner = THIS_MODULE,
+		.name = "tfp410",
+		.of_match_table = tfp410_of_match,
+	},
+};
+
+int __init tilcdc_tfp410_init(void)
+{
+	return platform_driver_register(&tfp410_driver);
+}
+
+void __exit tilcdc_tfp410_fini(void)
+{
+	platform_driver_unregister(&tfp410_driver);
+}
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_tfp410.h b/drivers/gpu/drm/tilcdc/tilcdc_tfp410.h
new file mode 100644
index 0000000..5b800f1
--- /dev/null
+++ b/drivers/gpu/drm/tilcdc/tilcdc_tfp410.h
@@ -0,0 +1,26 @@
+/*
+ * Copyright (C) 2012 Texas Instruments
+ * Author: Rob Clark <robdclark@gmail.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef __TILCDC_TFP410_H__
+#define __TILCDC_TFP410_H__
+
+/* sub-module for tfp410 dvi adaptor */
+
+int tilcdc_tfp410_init(void);
+void tilcdc_tfp410_fini(void);
+
+#endif /* __TILCDC_TFP410_H__ */
-- 
cgit v0.10.2


From e7792ce2da5ded80861db787ace9b57ecf7bc96c Mon Sep 17 00:00:00 2001
From: Rob Clark <robdclark@gmail.com>
Date: Tue, 8 Jan 2013 19:21:02 -0600
Subject: drm/i2c: nxp-tda998x (v3)

Driver for the NXP TDA998X i2c hdmi encoder slave.

v1: original
v2: fix npix/nline programming
v3: add Kconfig, fix dup'd MODULE_DESCRIPTION

Signed-off-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Tested-by: Koen Kooi <koen@dominion.thruhere.net>

diff --git a/drivers/gpu/drm/i2c/Kconfig b/drivers/gpu/drm/i2c/Kconfig
index 1611836..4d341db 100644
--- a/drivers/gpu/drm/i2c/Kconfig
+++ b/drivers/gpu/drm/i2c/Kconfig
@@ -19,4 +19,10 @@ config DRM_I2C_SIL164
 	  when used in pairs) TMDS transmitters, used in some nVidia
 	  video cards.
 
+config DRM_I2C_NXP_TDA998X
+	tristate "NXP Semiconductors TDA998X HDMI encoder"
+	default m if DRM_TILCDC
+	help
+	  Support for NXP Semiconductors TDA998X HDMI encoders.
+
 endmenu
diff --git a/drivers/gpu/drm/i2c/Makefile b/drivers/gpu/drm/i2c/Makefile
index 9286256..43aa33b 100644
--- a/drivers/gpu/drm/i2c/Makefile
+++ b/drivers/gpu/drm/i2c/Makefile
@@ -5,3 +5,6 @@ obj-$(CONFIG_DRM_I2C_CH7006) += ch7006.o
 
 sil164-y := sil164_drv.o
 obj-$(CONFIG_DRM_I2C_SIL164) += sil164.o
+
+tda998x-y := tda998x_drv.o
+obj-$(CONFIG_DRM_I2C_NXP_TDA998X) += tda998x.o
diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c
new file mode 100644
index 0000000..e68b58a
--- /dev/null
+++ b/drivers/gpu/drm/i2c/tda998x_drv.c
@@ -0,0 +1,906 @@
+/*
+ * Copyright (C) 2012 Texas Instruments
+ * Author: Rob Clark <robdclark@gmail.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+
+
+#include <linux/module.h>
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_encoder_slave.h>
+#include <drm/drm_edid.h>
+
+
+#define DBG(fmt, ...) DRM_DEBUG(fmt"\n", ##__VA_ARGS__)
+
+struct tda998x_priv {
+	struct i2c_client *cec;
+	uint16_t rev;
+	uint8_t current_page;
+	int dpms;
+};
+
+#define to_tda998x_priv(x)  ((struct tda998x_priv *)to_encoder_slave(x)->slave_priv)
+
+/* The TDA9988 series of devices use a paged register scheme.. to simplify
+ * things we encode the page # in upper bits of the register #.  To read/
+ * write a given register, we need to make sure CURPAGE register is set
+ * appropriately.  Which implies reads/writes are not atomic.  Fun!
+ */
+
+#define REG(page, addr) (((page) << 8) | (addr))
+#define REG2ADDR(reg)   ((reg) & 0xff)
+#define REG2PAGE(reg)   (((reg) >> 8) & 0xff)
+
+#define REG_CURPAGE               0xff                /* write */
+
+
+/* Page 00h: General Control */
+#define REG_VERSION_LSB           REG(0x00, 0x00)     /* read */
+#define REG_MAIN_CNTRL0           REG(0x00, 0x01)     /* read/write */
+# define MAIN_CNTRL0_SR           (1 << 0)
+# define MAIN_CNTRL0_DECS         (1 << 1)
+# define MAIN_CNTRL0_DEHS         (1 << 2)
+# define MAIN_CNTRL0_CECS         (1 << 3)
+# define MAIN_CNTRL0_CEHS         (1 << 4)
+# define MAIN_CNTRL0_SCALER       (1 << 7)
+#define REG_VERSION_MSB           REG(0x00, 0x02)     /* read */
+#define REG_SOFTRESET             REG(0x00, 0x0a)     /* write */
+# define SOFTRESET_AUDIO          (1 << 0)
+# define SOFTRESET_I2C_MASTER     (1 << 1)
+#define REG_DDC_DISABLE           REG(0x00, 0x0b)     /* read/write */
+#define REG_CCLK_ON               REG(0x00, 0x0c)     /* read/write */
+#define REG_I2C_MASTER            REG(0x00, 0x0d)     /* read/write */
+# define I2C_MASTER_DIS_MM        (1 << 0)
+# define I2C_MASTER_DIS_FILT      (1 << 1)
+# define I2C_MASTER_APP_STRT_LAT  (1 << 2)
+#define REG_INT_FLAGS_0           REG(0x00, 0x0f)     /* read/write */
+#define REG_INT_FLAGS_1           REG(0x00, 0x10)     /* read/write */
+#define REG_INT_FLAGS_2           REG(0x00, 0x11)     /* read/write */
+# define INT_FLAGS_2_EDID_BLK_RD  (1 << 1)
+#define REG_ENA_VP_0              REG(0x00, 0x18)     /* read/write */
+#define REG_ENA_VP_1              REG(0x00, 0x19)     /* read/write */
+#define REG_ENA_VP_2              REG(0x00, 0x1a)     /* read/write */
+#define REG_ENA_AP                REG(0x00, 0x1e)     /* read/write */
+#define REG_VIP_CNTRL_0           REG(0x00, 0x20)     /* write */
+# define VIP_CNTRL_0_MIRR_A       (1 << 7)
+# define VIP_CNTRL_0_SWAP_A(x)    (((x) & 7) << 4)
+# define VIP_CNTRL_0_MIRR_B       (1 << 3)
+# define VIP_CNTRL_0_SWAP_B(x)    (((x) & 7) << 0)
+#define REG_VIP_CNTRL_1           REG(0x00, 0x21)     /* write */
+# define VIP_CNTRL_1_MIRR_C       (1 << 7)
+# define VIP_CNTRL_1_SWAP_C(x)    (((x) & 7) << 4)
+# define VIP_CNTRL_1_MIRR_D       (1 << 3)
+# define VIP_CNTRL_1_SWAP_D(x)    (((x) & 7) << 0)
+#define REG_VIP_CNTRL_2           REG(0x00, 0x22)     /* write */
+# define VIP_CNTRL_2_MIRR_E       (1 << 7)
+# define VIP_CNTRL_2_SWAP_E(x)    (((x) & 7) << 4)
+# define VIP_CNTRL_2_MIRR_F       (1 << 3)
+# define VIP_CNTRL_2_SWAP_F(x)    (((x) & 7) << 0)
+#define REG_VIP_CNTRL_3           REG(0x00, 0x23)     /* write */
+# define VIP_CNTRL_3_X_TGL        (1 << 0)
+# define VIP_CNTRL_3_H_TGL        (1 << 1)
+# define VIP_CNTRL_3_V_TGL        (1 << 2)
+# define VIP_CNTRL_3_EMB          (1 << 3)
+# define VIP_CNTRL_3_SYNC_DE      (1 << 4)
+# define VIP_CNTRL_3_SYNC_HS      (1 << 5)
+# define VIP_CNTRL_3_DE_INT       (1 << 6)
+# define VIP_CNTRL_3_EDGE         (1 << 7)
+#define REG_VIP_CNTRL_4           REG(0x00, 0x24)     /* write */
+# define VIP_CNTRL_4_BLC(x)       (((x) & 3) << 0)
+# define VIP_CNTRL_4_BLANKIT(x)   (((x) & 3) << 2)
+# define VIP_CNTRL_4_CCIR656      (1 << 4)
+# define VIP_CNTRL_4_656_ALT      (1 << 5)
+# define VIP_CNTRL_4_TST_656      (1 << 6)
+# define VIP_CNTRL_4_TST_PAT      (1 << 7)
+#define REG_VIP_CNTRL_5           REG(0x00, 0x25)     /* write */
+# define VIP_CNTRL_5_CKCASE       (1 << 0)
+# define VIP_CNTRL_5_SP_CNT(x)    (((x) & 3) << 1)
+#define REG_MAT_CONTRL            REG(0x00, 0x80)     /* write */
+# define MAT_CONTRL_MAT_SC(x)     (((x) & 3) << 0)
+# define MAT_CONTRL_MAT_BP        (1 << 2)
+#define REG_VIDFORMAT             REG(0x00, 0xa0)     /* write */
+#define REG_REFPIX_MSB            REG(0x00, 0xa1)     /* write */
+#define REG_REFPIX_LSB            REG(0x00, 0xa2)     /* write */
+#define REG_REFLINE_MSB           REG(0x00, 0xa3)     /* write */
+#define REG_REFLINE_LSB           REG(0x00, 0xa4)     /* write */
+#define REG_NPIX_MSB              REG(0x00, 0xa5)     /* write */
+#define REG_NPIX_LSB              REG(0x00, 0xa6)     /* write */
+#define REG_NLINE_MSB             REG(0x00, 0xa7)     /* write */
+#define REG_NLINE_LSB             REG(0x00, 0xa8)     /* write */
+#define REG_VS_LINE_STRT_1_MSB    REG(0x00, 0xa9)     /* write */
+#define REG_VS_LINE_STRT_1_LSB    REG(0x00, 0xaa)     /* write */
+#define REG_VS_PIX_STRT_1_MSB     REG(0x00, 0xab)     /* write */
+#define REG_VS_PIX_STRT_1_LSB     REG(0x00, 0xac)     /* write */
+#define REG_VS_LINE_END_1_MSB     REG(0x00, 0xad)     /* write */
+#define REG_VS_LINE_END_1_LSB     REG(0x00, 0xae)     /* write */
+#define REG_VS_PIX_END_1_MSB      REG(0x00, 0xaf)     /* write */
+#define REG_VS_PIX_END_1_LSB      REG(0x00, 0xb0)     /* write */
+#define REG_VS_PIX_STRT_2_MSB     REG(0x00, 0xb3)     /* write */
+#define REG_VS_PIX_STRT_2_LSB     REG(0x00, 0xb4)     /* write */
+#define REG_VS_PIX_END_2_MSB      REG(0x00, 0xb7)     /* write */
+#define REG_VS_PIX_END_2_LSB      REG(0x00, 0xb8)     /* write */
+#define REG_HS_PIX_START_MSB      REG(0x00, 0xb9)     /* write */
+#define REG_HS_PIX_START_LSB      REG(0x00, 0xba)     /* write */
+#define REG_HS_PIX_STOP_MSB       REG(0x00, 0xbb)     /* write */
+#define REG_HS_PIX_STOP_LSB       REG(0x00, 0xbc)     /* write */
+#define REG_VWIN_START_1_MSB      REG(0x00, 0xbd)     /* write */
+#define REG_VWIN_START_1_LSB      REG(0x00, 0xbe)     /* write */
+#define REG_VWIN_END_1_MSB        REG(0x00, 0xbf)     /* write */
+#define REG_VWIN_END_1_LSB        REG(0x00, 0xc0)     /* write */
+#define REG_DE_START_MSB          REG(0x00, 0xc5)     /* write */
+#define REG_DE_START_LSB          REG(0x00, 0xc6)     /* write */
+#define REG_DE_STOP_MSB           REG(0x00, 0xc7)     /* write */
+#define REG_DE_STOP_LSB           REG(0x00, 0xc8)     /* write */
+#define REG_TBG_CNTRL_0           REG(0x00, 0xca)     /* write */
+# define TBG_CNTRL_0_FRAME_DIS    (1 << 5)
+# define TBG_CNTRL_0_SYNC_MTHD    (1 << 6)
+# define TBG_CNTRL_0_SYNC_ONCE    (1 << 7)
+#define REG_TBG_CNTRL_1           REG(0x00, 0xcb)     /* write */
+# define TBG_CNTRL_1_VH_TGL_0     (1 << 0)
+# define TBG_CNTRL_1_VH_TGL_1     (1 << 1)
+# define TBG_CNTRL_1_VH_TGL_2     (1 << 2)
+# define TBG_CNTRL_1_VHX_EXT_DE   (1 << 3)
+# define TBG_CNTRL_1_VHX_EXT_HS   (1 << 4)
+# define TBG_CNTRL_1_VHX_EXT_VS   (1 << 5)
+# define TBG_CNTRL_1_DWIN_DIS     (1 << 6)
+#define REG_ENABLE_SPACE          REG(0x00, 0xd6)     /* write */
+#define REG_HVF_CNTRL_0           REG(0x00, 0xe4)     /* write */
+# define HVF_CNTRL_0_SM           (1 << 7)
+# define HVF_CNTRL_0_RWB          (1 << 6)
+# define HVF_CNTRL_0_PREFIL(x)    (((x) & 3) << 2)
+# define HVF_CNTRL_0_INTPOL(x)    (((x) & 3) << 0)
+#define REG_HVF_CNTRL_1           REG(0x00, 0xe5)     /* write */
+# define HVF_CNTRL_1_FOR          (1 << 0)
+# define HVF_CNTRL_1_YUVBLK       (1 << 1)
+# define HVF_CNTRL_1_VQR(x)       (((x) & 3) << 2)
+# define HVF_CNTRL_1_PAD(x)       (((x) & 3) << 4)
+# define HVF_CNTRL_1_SEMI_PLANAR  (1 << 6)
+#define REG_RPT_CNTRL             REG(0x00, 0xf0)     /* write */
+
+
+/* Page 02h: PLL settings */
+#define REG_PLL_SERIAL_1          REG(0x02, 0x00)     /* read/write */
+# define PLL_SERIAL_1_SRL_FDN     (1 << 0)
+# define PLL_SERIAL_1_SRL_IZ(x)   (((x) & 3) << 1)
+# define PLL_SERIAL_1_SRL_MAN_IZ  (1 << 6)
+#define REG_PLL_SERIAL_2          REG(0x02, 0x01)     /* read/write */
+# define PLL_SERIAL_2_SRL_NOSC(x) (((x) & 3) << 0)
+# define PLL_SERIAL_2_SRL_PR(x)   (((x) & 0xf) << 4)
+#define REG_PLL_SERIAL_3          REG(0x02, 0x02)     /* read/write */
+# define PLL_SERIAL_3_SRL_CCIR    (1 << 0)
+# define PLL_SERIAL_3_SRL_DE      (1 << 2)
+# define PLL_SERIAL_3_SRL_PXIN_SEL (1 << 4)
+#define REG_SERIALIZER            REG(0x02, 0x03)     /* read/write */
+#define REG_BUFFER_OUT            REG(0x02, 0x04)     /* read/write */
+#define REG_PLL_SCG1              REG(0x02, 0x05)     /* read/write */
+#define REG_PLL_SCG2              REG(0x02, 0x06)     /* read/write */
+#define REG_PLL_SCGN1             REG(0x02, 0x07)     /* read/write */
+#define REG_PLL_SCGN2             REG(0x02, 0x08)     /* read/write */
+#define REG_PLL_SCGR1             REG(0x02, 0x09)     /* read/write */
+#define REG_PLL_SCGR2             REG(0x02, 0x0a)     /* read/write */
+#define REG_AUDIO_DIV             REG(0x02, 0x0e)     /* read/write */
+#define REG_SEL_CLK               REG(0x02, 0x11)     /* read/write */
+# define SEL_CLK_SEL_CLK1         (1 << 0)
+# define SEL_CLK_SEL_VRF_CLK(x)   (((x) & 3) << 1)
+# define SEL_CLK_ENA_SC_CLK       (1 << 3)
+#define REG_ANA_GENERAL           REG(0x02, 0x12)     /* read/write */
+
+
+/* Page 09h: EDID Control */
+#define REG_EDID_DATA_0           REG(0x09, 0x00)     /* read */
+/* next 127 successive registers are the EDID block */
+#define REG_EDID_CTRL             REG(0x09, 0xfa)     /* read/write */
+#define REG_DDC_ADDR              REG(0x09, 0xfb)     /* read/write */
+#define REG_DDC_OFFS              REG(0x09, 0xfc)     /* read/write */
+#define REG_DDC_SEGM_ADDR         REG(0x09, 0xfd)     /* read/write */
+#define REG_DDC_SEGM              REG(0x09, 0xfe)     /* read/write */
+
+
+/* Page 10h: information frames and packets */
+
+
+/* Page 11h: audio settings and content info packets */
+#define REG_AIP_CNTRL_0           REG(0x11, 0x00)     /* read/write */
+# define AIP_CNTRL_0_RST_FIFO     (1 << 0)
+# define AIP_CNTRL_0_SWAP         (1 << 1)
+# define AIP_CNTRL_0_LAYOUT       (1 << 2)
+# define AIP_CNTRL_0_ACR_MAN      (1 << 5)
+# define AIP_CNTRL_0_RST_CTS      (1 << 6)
+#define REG_ENC_CNTRL             REG(0x11, 0x0d)     /* read/write */
+# define ENC_CNTRL_RST_ENC        (1 << 0)
+# define ENC_CNTRL_RST_SEL        (1 << 1)
+# define ENC_CNTRL_CTL_CODE(x)    (((x) & 3) << 2)
+
+
+/* Page 12h: HDCP and OTP */
+#define REG_TX3                   REG(0x12, 0x9a)     /* read/write */
+#define REG_TX33                  REG(0x12, 0xb8)     /* read/write */
+# define TX33_HDMI                (1 << 1)
+
+
+/* Page 13h: Gamut related metadata packets */
+
+
+
+/* CEC registers: (not paged)
+ */
+#define REG_CEC_FRO_IM_CLK_CTRL   0xfb                /* read/write */
+# define CEC_FRO_IM_CLK_CTRL_GHOST_DIS (1 << 7)
+# define CEC_FRO_IM_CLK_CTRL_ENA_OTP   (1 << 6)
+# define CEC_FRO_IM_CLK_CTRL_IMCLK_SEL (1 << 1)
+# define CEC_FRO_IM_CLK_CTRL_FRO_DIV   (1 << 0)
+#define REG_CEC_RXSHPDLEV         0xfe                /* read */
+# define CEC_RXSHPDLEV_RXSENS     (1 << 0)
+# define CEC_RXSHPDLEV_HPD        (1 << 1)
+
+#define REG_CEC_ENAMODS           0xff                /* read/write */
+# define CEC_ENAMODS_DIS_FRO      (1 << 6)
+# define CEC_ENAMODS_DIS_CCLK     (1 << 5)
+# define CEC_ENAMODS_EN_RXSENS    (1 << 2)
+# define CEC_ENAMODS_EN_HDMI      (1 << 1)
+# define CEC_ENAMODS_EN_CEC       (1 << 0)
+
+
+/* Device versions: */
+#define TDA9989N2                 0x0101
+#define TDA19989                  0x0201
+#define TDA19989N2                0x0202
+#define TDA19988                  0x0301
+
+static void
+cec_write(struct drm_encoder *encoder, uint16_t addr, uint8_t val)
+{
+	struct i2c_client *client = to_tda998x_priv(encoder)->cec;
+	uint8_t buf[] = {addr, val};
+	int ret;
+
+	ret = i2c_master_send(client, buf, ARRAY_SIZE(buf));
+	if (ret < 0)
+		dev_err(&client->dev, "Error %d writing to cec:0x%x\n", ret, addr);
+}
+
+static uint8_t
+cec_read(struct drm_encoder *encoder, uint8_t addr)
+{
+	struct i2c_client *client = to_tda998x_priv(encoder)->cec;
+	uint8_t val;
+	int ret;
+
+	ret = i2c_master_send(client, &addr, sizeof(addr));
+	if (ret < 0)
+		goto fail;
+
+	ret = i2c_master_recv(client, &val, sizeof(val));
+	if (ret < 0)
+		goto fail;
+
+	return val;
+
+fail:
+	dev_err(&client->dev, "Error %d reading from cec:0x%x\n", ret, addr);
+	return 0;
+}
+
+static void
+set_page(struct drm_encoder *encoder, uint16_t reg)
+{
+	struct tda998x_priv *priv = to_tda998x_priv(encoder);
+
+	if (REG2PAGE(reg) != priv->current_page) {
+		struct i2c_client *client = drm_i2c_encoder_get_client(encoder);
+		uint8_t buf[] = {
+				REG_CURPAGE, REG2PAGE(reg)
+		};
+		int ret = i2c_master_send(client, buf, sizeof(buf));
+		if (ret < 0)
+			dev_err(&client->dev, "Error %d writing to REG_CURPAGE\n", ret);
+
+		priv->current_page = REG2PAGE(reg);
+	}
+}
+
+static int
+reg_read_range(struct drm_encoder *encoder, uint16_t reg, char *buf, int cnt)
+{
+	struct i2c_client *client = drm_i2c_encoder_get_client(encoder);
+	uint8_t addr = REG2ADDR(reg);
+	int ret;
+
+	set_page(encoder, reg);
+
+	ret = i2c_master_send(client, &addr, sizeof(addr));
+	if (ret < 0)
+		goto fail;
+
+	ret = i2c_master_recv(client, buf, cnt);
+	if (ret < 0)
+		goto fail;
+
+	return ret;
+
+fail:
+	dev_err(&client->dev, "Error %d reading from 0x%x\n", ret, reg);
+	return ret;
+}
+
+static uint8_t
+reg_read(struct drm_encoder *encoder, uint16_t reg)
+{
+	uint8_t val = 0;
+	reg_read_range(encoder, reg, &val, sizeof(val));
+	return val;
+}
+
+static void
+reg_write(struct drm_encoder *encoder, uint16_t reg, uint8_t val)
+{
+	struct i2c_client *client = drm_i2c_encoder_get_client(encoder);
+	uint8_t buf[] = {REG2ADDR(reg), val};
+	int ret;
+
+	set_page(encoder, reg);
+
+	ret = i2c_master_send(client, buf, ARRAY_SIZE(buf));
+	if (ret < 0)
+		dev_err(&client->dev, "Error %d writing to 0x%x\n", ret, reg);
+}
+
+static void
+reg_write16(struct drm_encoder *encoder, uint16_t reg, uint16_t val)
+{
+	struct i2c_client *client = drm_i2c_encoder_get_client(encoder);
+	uint8_t buf[] = {REG2ADDR(reg), val >> 8, val};
+	int ret;
+
+	set_page(encoder, reg);
+
+	ret = i2c_master_send(client, buf, ARRAY_SIZE(buf));
+	if (ret < 0)
+		dev_err(&client->dev, "Error %d writing to 0x%x\n", ret, reg);
+}
+
+static void
+reg_set(struct drm_encoder *encoder, uint16_t reg, uint8_t val)
+{
+	reg_write(encoder, reg, reg_read(encoder, reg) | val);
+}
+
+static void
+reg_clear(struct drm_encoder *encoder, uint16_t reg, uint8_t val)
+{
+	reg_write(encoder, reg, reg_read(encoder, reg) & ~val);
+}
+
+static void
+tda998x_reset(struct drm_encoder *encoder)
+{
+	/* reset audio and i2c master: */
+	reg_set(encoder, REG_SOFTRESET, SOFTRESET_AUDIO | SOFTRESET_I2C_MASTER);
+	msleep(50);
+	reg_clear(encoder, REG_SOFTRESET, SOFTRESET_AUDIO | SOFTRESET_I2C_MASTER);
+	msleep(50);
+
+	/* reset transmitter: */
+	reg_set(encoder, REG_MAIN_CNTRL0, MAIN_CNTRL0_SR);
+	reg_clear(encoder, REG_MAIN_CNTRL0, MAIN_CNTRL0_SR);
+
+	/* PLL registers common configuration */
+	reg_write(encoder, REG_PLL_SERIAL_1, 0x00);
+	reg_write(encoder, REG_PLL_SERIAL_2, PLL_SERIAL_2_SRL_NOSC(1));
+	reg_write(encoder, REG_PLL_SERIAL_3, 0x00);
+	reg_write(encoder, REG_SERIALIZER,   0x00);
+	reg_write(encoder, REG_BUFFER_OUT,   0x00);
+	reg_write(encoder, REG_PLL_SCG1,     0x00);
+	reg_write(encoder, REG_AUDIO_DIV,    0x03);
+	reg_write(encoder, REG_SEL_CLK,      SEL_CLK_SEL_CLK1 | SEL_CLK_ENA_SC_CLK);
+	reg_write(encoder, REG_PLL_SCGN1,    0xfa);
+	reg_write(encoder, REG_PLL_SCGN2,    0x00);
+	reg_write(encoder, REG_PLL_SCGR1,    0x5b);
+	reg_write(encoder, REG_PLL_SCGR2,    0x00);
+	reg_write(encoder, REG_PLL_SCG2,     0x10);
+}
+
+/* DRM encoder functions */
+
+static void
+tda998x_encoder_set_config(struct drm_encoder *encoder, void *params)
+{
+}
+
+static void
+tda998x_encoder_dpms(struct drm_encoder *encoder, int mode)
+{
+	struct tda998x_priv *priv = to_tda998x_priv(encoder);
+
+	/* we only care about on or off: */
+	if (mode != DRM_MODE_DPMS_ON)
+		mode = DRM_MODE_DPMS_OFF;
+
+	if (mode == priv->dpms)
+		return;
+
+	switch (mode) {
+	case DRM_MODE_DPMS_ON:
+		/* enable audio and video ports */
+		reg_write(encoder, REG_ENA_AP, 0xff);
+		reg_write(encoder, REG_ENA_VP_0, 0xff);
+		reg_write(encoder, REG_ENA_VP_1, 0xff);
+		reg_write(encoder, REG_ENA_VP_2, 0xff);
+		/* set muxing after enabling ports: */
+		reg_write(encoder, REG_VIP_CNTRL_0,
+				VIP_CNTRL_0_SWAP_A(2) | VIP_CNTRL_0_SWAP_B(3));
+		reg_write(encoder, REG_VIP_CNTRL_1,
+				VIP_CNTRL_1_SWAP_C(0) | VIP_CNTRL_1_SWAP_D(1));
+		reg_write(encoder, REG_VIP_CNTRL_2,
+				VIP_CNTRL_2_SWAP_E(4) | VIP_CNTRL_2_SWAP_F(5));
+		break;
+	case DRM_MODE_DPMS_OFF:
+		/* disable audio and video ports */
+		reg_write(encoder, REG_ENA_AP, 0x00);
+		reg_write(encoder, REG_ENA_VP_0, 0x00);
+		reg_write(encoder, REG_ENA_VP_1, 0x00);
+		reg_write(encoder, REG_ENA_VP_2, 0x00);
+		break;
+	}
+
+	priv->dpms = mode;
+}
+
+static void
+tda998x_encoder_save(struct drm_encoder *encoder)
+{
+	DBG("");
+}
+
+static void
+tda998x_encoder_restore(struct drm_encoder *encoder)
+{
+	DBG("");
+}
+
+static bool
+tda998x_encoder_mode_fixup(struct drm_encoder *encoder,
+			  const struct drm_display_mode *mode,
+			  struct drm_display_mode *adjusted_mode)
+{
+	return true;
+}
+
+static int
+tda998x_encoder_mode_valid(struct drm_encoder *encoder,
+			  struct drm_display_mode *mode)
+{
+	return MODE_OK;
+}
+
+static void
+tda998x_encoder_mode_set(struct drm_encoder *encoder,
+			struct drm_display_mode *mode,
+			struct drm_display_mode *adjusted_mode)
+{
+	struct tda998x_priv *priv = to_tda998x_priv(encoder);
+	uint16_t hs_start, hs_end, line_start, line_end;
+	uint16_t vwin_start, vwin_end, de_start, de_end;
+	uint16_t ref_pix, ref_line, pix_start2;
+	uint8_t reg, div, rep;
+
+	hs_start   = mode->hsync_start - mode->hdisplay;
+	hs_end     = mode->hsync_end - mode->hdisplay;
+	line_start = 1;
+	line_end   = 1 + mode->vsync_end - mode->vsync_start;
+	vwin_start = mode->vtotal - mode->vsync_start;
+	vwin_end   = vwin_start + mode->vdisplay;
+	de_start   = mode->htotal - mode->hdisplay;
+	de_end     = mode->htotal;
+
+	pix_start2 = 0;
+	if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+		pix_start2 = (mode->htotal / 2) + hs_start;
+
+	/* TODO how is this value calculated?  It is 2 for all common
+	 * formats in the tables in out of tree nxp driver (assuming
+	 * I've properly deciphered their byzantine table system)
+	 */
+	ref_line = 2;
+
+	/* this might changes for other color formats from the CRTC: */
+	ref_pix = 3 + hs_start;
+
+	div = 148500 / mode->clock;
+
+	DBG("clock=%d, div=%u", mode->clock, div);
+	DBG("hs_start=%u, hs_end=%u, line_start=%u, line_end=%u",
+			hs_start, hs_end, line_start, line_end);
+	DBG("vwin_start=%u, vwin_end=%u, de_start=%u, de_end=%u",
+			vwin_start, vwin_end, de_start, de_end);
+	DBG("ref_line=%u, ref_pix=%u, pix_start2=%u",
+			ref_line, ref_pix, pix_start2);
+
+	/* mute the audio FIFO: */
+	reg_set(encoder, REG_AIP_CNTRL_0, AIP_CNTRL_0_RST_FIFO);
+
+	/* set HDMI HDCP mode off: */
+	reg_set(encoder, REG_TBG_CNTRL_1, TBG_CNTRL_1_DWIN_DIS);
+	reg_clear(encoder, REG_TX33, TX33_HDMI);
+
+	reg_write(encoder, REG_ENC_CNTRL, ENC_CNTRL_CTL_CODE(0));
+	/* no pre-filter or interpolator: */
+	reg_write(encoder, REG_HVF_CNTRL_0, HVF_CNTRL_0_PREFIL(0) |
+			HVF_CNTRL_0_INTPOL(0));
+	reg_write(encoder, REG_VIP_CNTRL_5, VIP_CNTRL_5_SP_CNT(0));
+	reg_write(encoder, REG_VIP_CNTRL_4, VIP_CNTRL_4_BLANKIT(0) |
+			VIP_CNTRL_4_BLC(0));
+	reg_clear(encoder, REG_PLL_SERIAL_3, PLL_SERIAL_3_SRL_CCIR);
+
+	reg_clear(encoder, REG_PLL_SERIAL_1, PLL_SERIAL_1_SRL_MAN_IZ);
+	reg_clear(encoder, REG_PLL_SERIAL_3, PLL_SERIAL_3_SRL_DE);
+	reg_write(encoder, REG_SERIALIZER, 0);
+	reg_write(encoder, REG_HVF_CNTRL_1, HVF_CNTRL_1_VQR(0));
+
+	/* TODO enable pixel repeat for pixel rates less than 25Msamp/s */
+	rep = 0;
+	reg_write(encoder, REG_RPT_CNTRL, 0);
+	reg_write(encoder, REG_SEL_CLK, SEL_CLK_SEL_VRF_CLK(0) |
+			SEL_CLK_SEL_CLK1 | SEL_CLK_ENA_SC_CLK);
+
+	reg_write(encoder, REG_PLL_SERIAL_2, PLL_SERIAL_2_SRL_NOSC(div) |
+			PLL_SERIAL_2_SRL_PR(rep));
+
+	reg_write16(encoder, REG_VS_PIX_STRT_2_MSB, pix_start2);
+	reg_write16(encoder, REG_VS_PIX_END_2_MSB, pix_start2);
+
+	/* set color matrix bypass flag: */
+	reg_set(encoder, REG_MAT_CONTRL, MAT_CONTRL_MAT_BP);
+
+	/* set BIAS tmds value: */
+	reg_write(encoder, REG_ANA_GENERAL, 0x09);
+
+	reg_clear(encoder, REG_TBG_CNTRL_0, TBG_CNTRL_0_SYNC_MTHD);
+
+	reg_write(encoder, REG_VIP_CNTRL_3, 0);
+	reg_set(encoder, REG_VIP_CNTRL_3, VIP_CNTRL_3_SYNC_HS);
+	if (mode->flags & DRM_MODE_FLAG_NVSYNC)
+		reg_set(encoder, REG_VIP_CNTRL_3, VIP_CNTRL_3_V_TGL);
+
+	if (mode->flags & DRM_MODE_FLAG_NHSYNC)
+		reg_set(encoder, REG_VIP_CNTRL_3, VIP_CNTRL_3_H_TGL);
+
+	reg_write(encoder, REG_VIDFORMAT, 0x00);
+	reg_write16(encoder, REG_NPIX_MSB, mode->hdisplay - 1);
+	reg_write16(encoder, REG_NLINE_MSB, mode->vdisplay - 1);
+	reg_write16(encoder, REG_VS_LINE_STRT_1_MSB, line_start);
+	reg_write16(encoder, REG_VS_LINE_END_1_MSB, line_end);
+	reg_write16(encoder, REG_VS_PIX_STRT_1_MSB, hs_start);
+	reg_write16(encoder, REG_VS_PIX_END_1_MSB, hs_start);
+	reg_write16(encoder, REG_HS_PIX_START_MSB, hs_start);
+	reg_write16(encoder, REG_HS_PIX_STOP_MSB, hs_end);
+	reg_write16(encoder, REG_VWIN_START_1_MSB, vwin_start);
+	reg_write16(encoder, REG_VWIN_END_1_MSB, vwin_end);
+	reg_write16(encoder, REG_DE_START_MSB, de_start);
+	reg_write16(encoder, REG_DE_STOP_MSB, de_end);
+
+	if (priv->rev == TDA19988) {
+		/* let incoming pixels fill the active space (if any) */
+		reg_write(encoder, REG_ENABLE_SPACE, 0x01);
+	}
+
+	reg_write16(encoder, REG_REFPIX_MSB, ref_pix);
+	reg_write16(encoder, REG_REFLINE_MSB, ref_line);
+
+	reg = TBG_CNTRL_1_VHX_EXT_DE |
+			TBG_CNTRL_1_VHX_EXT_HS |
+			TBG_CNTRL_1_VHX_EXT_VS |
+			TBG_CNTRL_1_DWIN_DIS | /* HDCP off */
+			TBG_CNTRL_1_VH_TGL_2;
+	if (mode->flags & (DRM_MODE_FLAG_NVSYNC | DRM_MODE_FLAG_NHSYNC))
+		reg |= TBG_CNTRL_1_VH_TGL_0;
+	reg_set(encoder, REG_TBG_CNTRL_1, reg);
+
+	/* must be last register set: */
+	reg_clear(encoder, REG_TBG_CNTRL_0, TBG_CNTRL_0_SYNC_ONCE);
+}
+
+static enum drm_connector_status
+tda998x_encoder_detect(struct drm_encoder *encoder,
+		      struct drm_connector *connector)
+{
+	uint8_t val = cec_read(encoder, REG_CEC_RXSHPDLEV);
+	return (val & CEC_RXSHPDLEV_HPD) ? connector_status_connected :
+			connector_status_disconnected;
+}
+
+static int
+read_edid_block(struct drm_encoder *encoder, uint8_t *buf, int blk)
+{
+	uint8_t offset, segptr;
+	int ret, i;
+
+	/* enable EDID read irq: */
+	reg_set(encoder, REG_INT_FLAGS_2, INT_FLAGS_2_EDID_BLK_RD);
+
+	offset = (blk & 1) ? 128 : 0;
+	segptr = blk / 2;
+
+	reg_write(encoder, REG_DDC_ADDR, 0xa0);
+	reg_write(encoder, REG_DDC_OFFS, offset);
+	reg_write(encoder, REG_DDC_SEGM_ADDR, 0x60);
+	reg_write(encoder, REG_DDC_SEGM, segptr);
+
+	/* enable reading EDID: */
+	reg_write(encoder, REG_EDID_CTRL, 0x1);
+
+	/* flag must be cleared by sw: */
+	reg_write(encoder, REG_EDID_CTRL, 0x0);
+
+	/* wait for block read to complete: */
+	for (i = 100; i > 0; i--) {
+		uint8_t val = reg_read(encoder, REG_INT_FLAGS_2);
+		if (val & INT_FLAGS_2_EDID_BLK_RD)
+			break;
+		msleep(1);
+	}
+
+	if (i == 0)
+		return -ETIMEDOUT;
+
+	ret = reg_read_range(encoder, REG_EDID_DATA_0, buf, EDID_LENGTH);
+	if (ret != EDID_LENGTH) {
+		dev_err(encoder->dev->dev, "failed to read edid block %d: %d",
+				blk, ret);
+		return ret;
+	}
+
+	reg_clear(encoder, REG_INT_FLAGS_2, INT_FLAGS_2_EDID_BLK_RD);
+
+	return 0;
+}
+
+static uint8_t *
+do_get_edid(struct drm_encoder *encoder)
+{
+	int j = 0, valid_extensions = 0;
+	uint8_t *block, *new;
+	bool print_bad_edid = drm_debug & DRM_UT_KMS;
+
+	if ((block = kmalloc(EDID_LENGTH, GFP_KERNEL)) == NULL)
+		return NULL;
+
+	/* base block fetch */
+	if (read_edid_block(encoder, block, 0))
+		goto fail;
+
+	if (!drm_edid_block_valid(block, 0, print_bad_edid))
+		goto fail;
+
+	/* if there's no extensions, we're done */
+	if (block[0x7e] == 0)
+		return block;
+
+	new = krealloc(block, (block[0x7e] + 1) * EDID_LENGTH, GFP_KERNEL);
+	if (!new)
+		goto fail;
+	block = new;
+
+	for (j = 1; j <= block[0x7e]; j++) {
+		uint8_t *ext_block = block + (valid_extensions + 1) * EDID_LENGTH;
+		if (read_edid_block(encoder, ext_block, j))
+			goto fail;
+
+		if (!drm_edid_block_valid(ext_block, j, print_bad_edid))
+			goto fail;
+
+		valid_extensions++;
+	}
+
+	if (valid_extensions != block[0x7e]) {
+		block[EDID_LENGTH-1] += block[0x7e] - valid_extensions;
+		block[0x7e] = valid_extensions;
+		new = krealloc(block, (valid_extensions + 1) * EDID_LENGTH, GFP_KERNEL);
+		if (!new)
+			goto fail;
+		block = new;
+	}
+
+	return block;
+
+fail:
+	dev_warn(encoder->dev->dev, "failed to read EDID\n");
+	kfree(block);
+	return NULL;
+}
+
+static int
+tda998x_encoder_get_modes(struct drm_encoder *encoder,
+			 struct drm_connector *connector)
+{
+	struct edid *edid = (struct edid *)do_get_edid(encoder);
+	int n = 0;
+
+	if (edid) {
+		drm_mode_connector_update_edid_property(connector, edid);
+		n = drm_add_edid_modes(connector, edid);
+		kfree(edid);
+	}
+
+	return n;
+}
+
+static int
+tda998x_encoder_create_resources(struct drm_encoder *encoder,
+				struct drm_connector *connector)
+{
+	DBG("");
+	return 0;
+}
+
+static int
+tda998x_encoder_set_property(struct drm_encoder *encoder,
+			    struct drm_connector *connector,
+			    struct drm_property *property,
+			    uint64_t val)
+{
+	DBG("");
+	return 0;
+}
+
+static void
+tda998x_encoder_destroy(struct drm_encoder *encoder)
+{
+	struct tda998x_priv *priv = to_tda998x_priv(encoder);
+	drm_i2c_encoder_destroy(encoder);
+	kfree(priv);
+}
+
+static struct drm_encoder_slave_funcs tda998x_encoder_funcs = {
+	.set_config = tda998x_encoder_set_config,
+	.destroy = tda998x_encoder_destroy,
+	.dpms = tda998x_encoder_dpms,
+	.save = tda998x_encoder_save,
+	.restore = tda998x_encoder_restore,
+	.mode_fixup = tda998x_encoder_mode_fixup,
+	.mode_valid = tda998x_encoder_mode_valid,
+	.mode_set = tda998x_encoder_mode_set,
+	.detect = tda998x_encoder_detect,
+	.get_modes = tda998x_encoder_get_modes,
+	.create_resources = tda998x_encoder_create_resources,
+	.set_property = tda998x_encoder_set_property,
+};
+
+/* I2C driver functions */
+
+static int
+tda998x_probe(struct i2c_client *client, const struct i2c_device_id *id)
+{
+	return 0;
+}
+
+static int
+tda998x_remove(struct i2c_client *client)
+{
+	return 0;
+}
+
+static int
+tda998x_encoder_init(struct i2c_client *client,
+		    struct drm_device *dev,
+		    struct drm_encoder_slave *encoder_slave)
+{
+	struct drm_encoder *encoder = &encoder_slave->base;
+	struct tda998x_priv *priv;
+
+	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	priv->current_page = 0;
+	priv->cec = i2c_new_dummy(client->adapter, 0x34);
+	priv->dpms = DRM_MODE_DPMS_OFF;
+
+	encoder_slave->slave_priv = priv;
+	encoder_slave->slave_funcs = &tda998x_encoder_funcs;
+
+	/* wake up the device: */
+	cec_write(encoder, REG_CEC_ENAMODS,
+			CEC_ENAMODS_EN_RXSENS | CEC_ENAMODS_EN_HDMI);
+
+	tda998x_reset(encoder);
+
+	/* read version: */
+	priv->rev = reg_read(encoder, REG_VERSION_LSB) |
+			reg_read(encoder, REG_VERSION_MSB) << 8;
+
+	/* mask off feature bits: */
+	priv->rev &= ~0x30; /* not-hdcp and not-scalar bit */
+
+	switch (priv->rev) {
+	case TDA9989N2:  dev_info(dev->dev, "found TDA9989 n2");  break;
+	case TDA19989:   dev_info(dev->dev, "found TDA19989");    break;
+	case TDA19989N2: dev_info(dev->dev, "found TDA19989 n2"); break;
+	case TDA19988:   dev_info(dev->dev, "found TDA19988");    break;
+	default:
+		DBG("found unsupported device: %04x", priv->rev);
+		goto fail;
+	}
+
+	/* after reset, enable DDC: */
+	reg_write(encoder, REG_DDC_DISABLE, 0x00);
+
+	/* set clock on DDC channel: */
+	reg_write(encoder, REG_TX3, 39);
+
+	/* if necessary, disable multi-master: */
+	if (priv->rev == TDA19989)
+		reg_set(encoder, REG_I2C_MASTER, I2C_MASTER_DIS_MM);
+
+	cec_write(encoder, REG_CEC_FRO_IM_CLK_CTRL,
+			CEC_FRO_IM_CLK_CTRL_GHOST_DIS | CEC_FRO_IM_CLK_CTRL_IMCLK_SEL);
+
+	return 0;
+
+fail:
+	/* if encoder_init fails, the encoder slave is never registered,
+	 * so cleanup here:
+	 */
+	if (priv->cec)
+		i2c_unregister_device(priv->cec);
+	kfree(priv);
+	encoder_slave->slave_priv = NULL;
+	encoder_slave->slave_funcs = NULL;
+	return -ENXIO;
+}
+
+static struct i2c_device_id tda998x_ids[] = {
+	{ "tda998x", 0 },
+	{ }
+};
+MODULE_DEVICE_TABLE(i2c, tda998x_ids);
+
+static struct drm_i2c_encoder_driver tda998x_driver = {
+	.i2c_driver = {
+		.probe = tda998x_probe,
+		.remove = tda998x_remove,
+		.driver = {
+			.name = "tda998x",
+		},
+		.id_table = tda998x_ids,
+	},
+	.encoder_init = tda998x_encoder_init,
+};
+
+/* Module initialization */
+
+static int __init
+tda998x_init(void)
+{
+	DBG("");
+	return drm_i2c_encoder_register(THIS_MODULE, &tda998x_driver);
+}
+
+static void __exit
+tda998x_exit(void)
+{
+	DBG("");
+	drm_i2c_encoder_unregister(&tda998x_driver);
+}
+
+MODULE_AUTHOR("Rob Clark <robdclark@gmail.com");
+MODULE_DESCRIPTION("NXP Semiconductors TDA998X HDMI Encoder");
+MODULE_LICENSE("GPL");
+
+module_init(tda998x_init);
+module_exit(tda998x_exit);
-- 
cgit v0.10.2


From 6e8de0bd6a51fdeebd5d975c4fcc426f730b339b Mon Sep 17 00:00:00 2001
From: Rob Clark <robdclark@gmail.com>
Date: Tue, 22 Jan 2013 16:02:21 -0600
Subject: drm/tilcdc: add encoder slave (v2)

Add output panel driver for i2c encoder slaves.

v1: original
v2: add DT bindings docs, and minor updates for review comments

Signed-off-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Tested-by: Koen Kooi <koen@dominion.thruhere.net>

diff --git a/Documentation/devicetree/bindings/drm/tilcdc/slave.txt b/Documentation/devicetree/bindings/drm/tilcdc/slave.txt
new file mode 100644
index 0000000..3d2c524
--- /dev/null
+++ b/Documentation/devicetree/bindings/drm/tilcdc/slave.txt
@@ -0,0 +1,18 @@
+Device-Tree bindings for tilcdc DRM encoder slave output driver
+
+Required properties:
+ - compatible: value should be "ti,tilcdc,slave".
+ - i2c: the phandle for the i2c device the encoder slave is connected to
+
+Recommended properties:
+ - pinctrl-names, pinctrl-0: the pincontrol settings to configure
+   muxing properly for pins that connect to TFP410 device
+
+Example:
+
+	hdmi {
+		compatible = "ti,tilcdc,slave";
+		i2c = <&i2c0>;
+		pinctrl-names = "default";
+		pinctrl-0 = <&nxp_hdmi_bonelt_pins>;
+	};
diff --git a/drivers/gpu/drm/tilcdc/Makefile b/drivers/gpu/drm/tilcdc/Makefile
index 1359cc2..aa9097e 100644
--- a/drivers/gpu/drm/tilcdc/Makefile
+++ b/drivers/gpu/drm/tilcdc/Makefile
@@ -3,6 +3,7 @@ ccflags-y := -Iinclude/drm -Werror
 tilcdc-y := \
 	tilcdc_crtc.o \
 	tilcdc_tfp410.o \
+	tilcdc_slave.o \
 	tilcdc_drv.o
 
 obj-$(CONFIG_DRM_TILCDC)	+= tilcdc.o
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_drv.c b/drivers/gpu/drm/tilcdc/tilcdc_drv.c
index f6defbf..25f3add 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_drv.c
+++ b/drivers/gpu/drm/tilcdc/tilcdc_drv.c
@@ -20,6 +20,7 @@
 #include "tilcdc_drv.h"
 #include "tilcdc_regs.h"
 #include "tilcdc_tfp410.h"
+#include "tilcdc_slave.h"
 
 #include "drm_fb_helper.h"
 
@@ -587,6 +588,7 @@ static int __init tilcdc_drm_init(void)
 {
 	DBG("init");
 	tilcdc_tfp410_init();
+	tilcdc_slave_init();
 	return platform_driver_register(&tilcdc_platform_driver);
 }
 
@@ -594,10 +596,11 @@ static void __exit tilcdc_drm_fini(void)
 {
 	DBG("fini");
 	tilcdc_tfp410_fini();
+	tilcdc_slave_fini();
 	platform_driver_unregister(&tilcdc_platform_driver);
 }
 
-module_init(tilcdc_drm_init);
+late_initcall(tilcdc_drm_init);
 module_exit(tilcdc_drm_fini);
 
 MODULE_AUTHOR("Rob Clark <robdclark@gmail.com");
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_slave.c b/drivers/gpu/drm/tilcdc/tilcdc_slave.c
new file mode 100644
index 0000000..568dc1c
--- /dev/null
+++ b/drivers/gpu/drm/tilcdc/tilcdc_slave.c
@@ -0,0 +1,376 @@
+/*
+ * Copyright (C) 2012 Texas Instruments
+ * Author: Rob Clark <robdclark@gmail.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/i2c.h>
+#include <linux/of_i2c.h>
+#include <linux/pinctrl/pinmux.h>
+#include <linux/pinctrl/consumer.h>
+#include <drm/drm_encoder_slave.h>
+
+#include "tilcdc_drv.h"
+
+struct slave_module {
+	struct tilcdc_module base;
+	struct i2c_adapter *i2c;
+};
+#define to_slave_module(x) container_of(x, struct slave_module, base)
+
+static const struct tilcdc_panel_info slave_info = {
+		.bpp                    = 16,
+		.ac_bias                = 255,
+		.ac_bias_intrpt         = 0,
+		.dma_burst_sz           = 16,
+		.fdd                    = 0x80,
+		.tft_alt_mode           = 0,
+		.sync_edge              = 0,
+		.sync_ctrl              = 1,
+		.raster_order           = 0,
+};
+
+
+/*
+ * Encoder:
+ */
+
+struct slave_encoder {
+	struct drm_encoder_slave base;
+	struct slave_module *mod;
+};
+#define to_slave_encoder(x) container_of(to_encoder_slave(x), struct slave_encoder, base)
+
+static inline struct drm_encoder_slave_funcs *
+get_slave_funcs(struct drm_encoder *enc)
+{
+	return to_encoder_slave(enc)->slave_funcs;
+}
+
+static void slave_encoder_destroy(struct drm_encoder *encoder)
+{
+	struct slave_encoder *slave_encoder = to_slave_encoder(encoder);
+	if (get_slave_funcs(encoder))
+		get_slave_funcs(encoder)->destroy(encoder);
+	drm_encoder_cleanup(encoder);
+	kfree(slave_encoder);
+}
+
+static void slave_encoder_prepare(struct drm_encoder *encoder)
+{
+	drm_i2c_encoder_prepare(encoder);
+	tilcdc_crtc_set_panel_info(encoder->crtc, &slave_info);
+}
+
+static const struct drm_encoder_funcs slave_encoder_funcs = {
+		.destroy        = slave_encoder_destroy,
+};
+
+static const struct drm_encoder_helper_funcs slave_encoder_helper_funcs = {
+		.dpms           = drm_i2c_encoder_dpms,
+		.mode_fixup     = drm_i2c_encoder_mode_fixup,
+		.prepare        = slave_encoder_prepare,
+		.commit         = drm_i2c_encoder_commit,
+		.mode_set       = drm_i2c_encoder_mode_set,
+		.save           = drm_i2c_encoder_save,
+		.restore        = drm_i2c_encoder_restore,
+};
+
+static const struct i2c_board_info info = {
+		I2C_BOARD_INFO("tda998x", 0x70)
+};
+
+static struct drm_encoder *slave_encoder_create(struct drm_device *dev,
+		struct slave_module *mod)
+{
+	struct slave_encoder *slave_encoder;
+	struct drm_encoder *encoder;
+	int ret;
+
+	slave_encoder = kzalloc(sizeof(*slave_encoder), GFP_KERNEL);
+	if (!slave_encoder) {
+		dev_err(dev->dev, "allocation failed\n");
+		return NULL;
+	}
+
+	slave_encoder->mod = mod;
+
+	encoder = &slave_encoder->base.base;
+	encoder->possible_crtcs = 1;
+
+	ret = drm_encoder_init(dev, encoder, &slave_encoder_funcs,
+			DRM_MODE_ENCODER_TMDS);
+	if (ret)
+		goto fail;
+
+	drm_encoder_helper_add(encoder, &slave_encoder_helper_funcs);
+
+	ret = drm_i2c_encoder_init(dev, to_encoder_slave(encoder), mod->i2c, &info);
+	if (ret)
+		goto fail;
+
+	return encoder;
+
+fail:
+	slave_encoder_destroy(encoder);
+	return NULL;
+}
+
+/*
+ * Connector:
+ */
+
+struct slave_connector {
+	struct drm_connector base;
+
+	struct drm_encoder *encoder;  /* our connected encoder */
+	struct slave_module *mod;
+};
+#define to_slave_connector(x) container_of(x, struct slave_connector, base)
+
+static void slave_connector_destroy(struct drm_connector *connector)
+{
+	struct slave_connector *slave_connector = to_slave_connector(connector);
+	drm_connector_cleanup(connector);
+	kfree(slave_connector);
+}
+
+static enum drm_connector_status slave_connector_detect(
+		struct drm_connector *connector,
+		bool force)
+{
+	struct drm_encoder *encoder = to_slave_connector(connector)->encoder;
+	return get_slave_funcs(encoder)->detect(encoder, connector);
+}
+
+static int slave_connector_get_modes(struct drm_connector *connector)
+{
+	struct drm_encoder *encoder = to_slave_connector(connector)->encoder;
+	return get_slave_funcs(encoder)->get_modes(encoder, connector);
+}
+
+static int slave_connector_mode_valid(struct drm_connector *connector,
+		  struct drm_display_mode *mode)
+{
+	struct drm_encoder *encoder = to_slave_connector(connector)->encoder;
+	struct tilcdc_drm_private *priv = connector->dev->dev_private;
+	int ret;
+
+	ret = tilcdc_crtc_mode_valid(priv->crtc, mode);
+	if (ret != MODE_OK)
+		return ret;
+
+	return get_slave_funcs(encoder)->mode_valid(encoder, mode);
+}
+
+static struct drm_encoder *slave_connector_best_encoder(
+		struct drm_connector *connector)
+{
+	struct slave_connector *slave_connector = to_slave_connector(connector);
+	return slave_connector->encoder;
+}
+
+static int slave_connector_set_property(struct drm_connector *connector,
+		struct drm_property *property, uint64_t value)
+{
+	struct drm_encoder *encoder = to_slave_connector(connector)->encoder;
+	return get_slave_funcs(encoder)->set_property(encoder,
+			connector, property, value);
+}
+
+static const struct drm_connector_funcs slave_connector_funcs = {
+	.destroy            = slave_connector_destroy,
+	.dpms               = drm_helper_connector_dpms,
+	.detect             = slave_connector_detect,
+	.fill_modes         = drm_helper_probe_single_connector_modes,
+	.set_property       = slave_connector_set_property,
+};
+
+static const struct drm_connector_helper_funcs slave_connector_helper_funcs = {
+	.get_modes          = slave_connector_get_modes,
+	.mode_valid         = slave_connector_mode_valid,
+	.best_encoder       = slave_connector_best_encoder,
+};
+
+static struct drm_connector *slave_connector_create(struct drm_device *dev,
+		struct slave_module *mod, struct drm_encoder *encoder)
+{
+	struct slave_connector *slave_connector;
+	struct drm_connector *connector;
+	int ret;
+
+	slave_connector = kzalloc(sizeof(*slave_connector), GFP_KERNEL);
+	if (!slave_connector) {
+		dev_err(dev->dev, "allocation failed\n");
+		return NULL;
+	}
+
+	slave_connector->encoder = encoder;
+	slave_connector->mod = mod;
+
+	connector = &slave_connector->base;
+
+	drm_connector_init(dev, connector, &slave_connector_funcs,
+			DRM_MODE_CONNECTOR_HDMIA);
+	drm_connector_helper_add(connector, &slave_connector_helper_funcs);
+
+	connector->polled = DRM_CONNECTOR_POLL_CONNECT |
+			DRM_CONNECTOR_POLL_DISCONNECT;
+
+	connector->interlace_allowed = 0;
+	connector->doublescan_allowed = 0;
+
+	get_slave_funcs(encoder)->create_resources(encoder, connector);
+
+	ret = drm_mode_connector_attach_encoder(connector, encoder);
+	if (ret)
+		goto fail;
+
+	drm_sysfs_connector_add(connector);
+
+	return connector;
+
+fail:
+	slave_connector_destroy(connector);
+	return NULL;
+}
+
+/*
+ * Module:
+ */
+
+static int slave_modeset_init(struct tilcdc_module *mod, struct drm_device *dev)
+{
+	struct slave_module *slave_mod = to_slave_module(mod);
+	struct tilcdc_drm_private *priv = dev->dev_private;
+	struct drm_encoder *encoder;
+	struct drm_connector *connector;
+
+	encoder = slave_encoder_create(dev, slave_mod);
+	if (!encoder)
+		return -ENOMEM;
+
+	connector = slave_connector_create(dev, slave_mod, encoder);
+	if (!connector)
+		return -ENOMEM;
+
+	priv->encoders[priv->num_encoders++] = encoder;
+	priv->connectors[priv->num_connectors++] = connector;
+
+	return 0;
+}
+
+static void slave_destroy(struct tilcdc_module *mod)
+{
+	struct slave_module *slave_mod = to_slave_module(mod);
+
+	tilcdc_module_cleanup(mod);
+	kfree(slave_mod);
+}
+
+static const struct tilcdc_module_ops slave_module_ops = {
+		.modeset_init = slave_modeset_init,
+		.destroy = slave_destroy,
+};
+
+/*
+ * Device:
+ */
+
+static struct of_device_id slave_of_match[];
+
+static int slave_probe(struct platform_device *pdev)
+{
+	struct device_node *node = pdev->dev.of_node;
+	struct device_node *i2c_node;
+	struct slave_module *slave_mod;
+	struct tilcdc_module *mod;
+	struct pinctrl *pinctrl;
+	uint32_t i2c_phandle;
+	int ret = -EINVAL;
+
+	/* bail out early if no DT data: */
+	if (!node) {
+		dev_err(&pdev->dev, "device-tree data is missing\n");
+		return -ENXIO;
+	}
+
+	slave_mod = kzalloc(sizeof(*slave_mod), GFP_KERNEL);
+	if (!slave_mod)
+		return -ENOMEM;
+
+	mod = &slave_mod->base;
+
+	tilcdc_module_init(mod, "slave", &slave_module_ops);
+
+	pinctrl = devm_pinctrl_get_select_default(&pdev->dev);
+	if (IS_ERR(pinctrl))
+		dev_warn(&pdev->dev, "pins are not configured\n");
+
+	if (of_property_read_u32(node, "i2c", &i2c_phandle)) {
+		dev_err(&pdev->dev, "could not get i2c bus phandle\n");
+		goto fail;
+	}
+
+	i2c_node = of_find_node_by_phandle(i2c_phandle);
+	if (!i2c_node) {
+		dev_err(&pdev->dev, "could not get i2c bus node\n");
+		goto fail;
+	}
+
+	slave_mod->i2c = of_find_i2c_adapter_by_node(i2c_node);
+	if (!slave_mod->i2c) {
+		dev_err(&pdev->dev, "could not get i2c\n");
+		goto fail;
+	}
+
+	of_node_put(i2c_node);
+
+	return 0;
+
+fail:
+	slave_destroy(mod);
+	return ret;
+}
+
+static int slave_remove(struct platform_device *pdev)
+{
+	return 0;
+}
+
+static struct of_device_id slave_of_match[] = {
+		{ .compatible = "ti,tilcdc,slave", },
+		{ },
+};
+MODULE_DEVICE_TABLE(of, slave_of_match);
+
+struct platform_driver slave_driver = {
+	.probe = slave_probe,
+	.remove = slave_remove,
+	.driver = {
+		.owner = THIS_MODULE,
+		.name = "slave",
+		.of_match_table = slave_of_match,
+	},
+};
+
+int __init tilcdc_slave_init(void)
+{
+	return platform_driver_register(&slave_driver);
+}
+
+void __exit tilcdc_slave_fini(void)
+{
+	platform_driver_unregister(&slave_driver);
+}
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_slave.h b/drivers/gpu/drm/tilcdc/tilcdc_slave.h
new file mode 100644
index 0000000..2f85048
--- /dev/null
+++ b/drivers/gpu/drm/tilcdc/tilcdc_slave.h
@@ -0,0 +1,26 @@
+/*
+ * Copyright (C) 2012 Texas Instruments
+ * Author: Rob Clark <robdclark@gmail.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef __TILCDC_SLAVE_H__
+#define __TILCDC_SLAVE_H__
+
+/* sub-module for i2c slave encoder output */
+
+int tilcdc_slave_init(void);
+void tilcdc_slave_fini(void);
+
+#endif /* __TILCDC_SLAVE_H__ */
-- 
cgit v0.10.2


From 0d4bbaf9f3e5b9f52150ddc5a4ee8b0ab83a440b Mon Sep 17 00:00:00 2001
From: Rob Clark <robdclark@gmail.com>
Date: Tue, 18 Dec 2012 17:34:16 -0600
Subject: drm/tilcdc: add support for LCD panels (v5)

Add an output panel driver for LCD panels.  Tested with LCD3 cape on
beaglebone.

v1: original
v2: s/of_find_node_by_name()/of_get_child_by_name()/ from Pantelis
    Antoniou
v3: add backlight support
v4: rebase to latest of video timing helpers
v5: remove some unneeded fields from panel-info struct, add DT bindings
    docs

Signed-off-by: Rob Clark <robdclark@gmail.com>
Tested-by: Koen Kooi <koen@dominion.thruhere.net>

diff --git a/Documentation/devicetree/bindings/drm/tilcdc/panel.txt b/Documentation/devicetree/bindings/drm/tilcdc/panel.txt
new file mode 100644
index 0000000..9301c33
--- /dev/null
+++ b/Documentation/devicetree/bindings/drm/tilcdc/panel.txt
@@ -0,0 +1,59 @@
+Device-Tree bindings for tilcdc DRM generic panel output driver
+
+Required properties:
+ - compatible: value should be "ti,tilcdc,panel".
+ - panel-info: configuration info to configure LCDC correctly for the panel
+   - ac-bias: AC Bias Pin Frequency
+   - ac-bias-intrpt: AC Bias Pin Transitions per Interrupt
+   - dma-burst-sz: DMA burst size
+   - bpp: Bits per pixel
+   - fdd: FIFO DMA Request Delay
+   - sync-edge: Horizontal and Vertical Sync Edge: 0=rising 1=falling
+   - sync-ctrl: Horizontal and Vertical Sync: Control: 0=ignore
+   - raster-order: Raster Data Order Select: 1=Most-to-least 0=Least-to-most
+   - fifo-th: DMA FIFO threshold
+ - display-timings: typical videomode of lcd panel.  Multiple video modes
+   can be listed if the panel supports multiple timings, but the 'native-mode'
+   should be the preferred/default resolution.  Refer to
+   Documentation/devicetree/bindings/video/display-timing.txt for display
+   timing binding details.
+
+Recommended properties:
+ - pinctrl-names, pinctrl-0: the pincontrol settings to configure
+   muxing properly for pins that connect to TFP410 device
+
+Example:
+
+	/* Settings for CDTech_S035Q01 / LCD3 cape: */
+	lcd3 {
+		compatible = "ti,tilcdc,panel";
+		pinctrl-names = "default";
+		pinctrl-0 = <&bone_lcd3_cape_lcd_pins>;
+		panel-info {
+			ac-bias           = <255>;
+			ac-bias-intrpt    = <0>;
+			dma-burst-sz      = <16>;
+			bpp               = <16>;
+			fdd               = <0x80>;
+			sync-edge         = <0>;
+			sync-ctrl         = <1>;
+			raster-order      = <0>;
+			fifo-th           = <0>;
+		};
+		display-timings {
+			native-mode = <&timing0>;
+			timing0: 320x240 {
+				hactive         = <320>;
+				vactive         = <240>;
+				hback-porch     = <21>;
+				hfront-porch    = <58>;
+				hsync-len       = <47>;
+				vback-porch     = <11>;
+				vfront-porch    = <23>;
+				vsync-len       = <2>;
+				clock-frequency = <8000000>;
+				hsync-active    = <0>;
+				vsync-active    = <0>;
+			};
+		};
+	};
diff --git a/drivers/gpu/drm/tilcdc/Kconfig b/drivers/gpu/drm/tilcdc/Kconfig
index ee9b592..ae14fd6 100644
--- a/drivers/gpu/drm/tilcdc/Kconfig
+++ b/drivers/gpu/drm/tilcdc/Kconfig
@@ -4,6 +4,9 @@ config DRM_TILCDC
 	select DRM_KMS_HELPER
 	select DRM_KMS_CMA_HELPER
 	select DRM_GEM_CMA_HELPER
+	select OF_VIDEOMODE
+	select OF_DISPLAY_TIMING
+	select BACKLIGHT_CLASS_DEVICE
 	help
 	  Choose this option if you have an TI SoC with LCDC display
 	  controller, for example AM33xx in beagle-bone, DA8xx, or
diff --git a/drivers/gpu/drm/tilcdc/Makefile b/drivers/gpu/drm/tilcdc/Makefile
index aa9097e..deda656 100644
--- a/drivers/gpu/drm/tilcdc/Makefile
+++ b/drivers/gpu/drm/tilcdc/Makefile
@@ -4,6 +4,7 @@ tilcdc-y := \
 	tilcdc_crtc.o \
 	tilcdc_tfp410.o \
 	tilcdc_slave.o \
+	tilcdc_panel.o \
 	tilcdc_drv.o
 
 obj-$(CONFIG_DRM_TILCDC)	+= tilcdc.o
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_drv.c b/drivers/gpu/drm/tilcdc/tilcdc_drv.c
index 25f3add..c5b592d 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_drv.c
+++ b/drivers/gpu/drm/tilcdc/tilcdc_drv.c
@@ -21,6 +21,7 @@
 #include "tilcdc_regs.h"
 #include "tilcdc_tfp410.h"
 #include "tilcdc_slave.h"
+#include "tilcdc_panel.h"
 
 #include "drm_fb_helper.h"
 
@@ -589,6 +590,7 @@ static int __init tilcdc_drm_init(void)
 	DBG("init");
 	tilcdc_tfp410_init();
 	tilcdc_slave_init();
+	tilcdc_panel_init();
 	return platform_driver_register(&tilcdc_platform_driver);
 }
 
@@ -597,6 +599,7 @@ static void __exit tilcdc_drm_fini(void)
 	DBG("fini");
 	tilcdc_tfp410_fini();
 	tilcdc_slave_fini();
+	tilcdc_panel_fini();
 	platform_driver_unregister(&tilcdc_platform_driver);
 }
 
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_panel.c b/drivers/gpu/drm/tilcdc/tilcdc_panel.c
new file mode 100644
index 0000000..580b74e
--- /dev/null
+++ b/drivers/gpu/drm/tilcdc/tilcdc_panel.c
@@ -0,0 +1,436 @@
+/*
+ * Copyright (C) 2012 Texas Instruments
+ * Author: Rob Clark <robdclark@gmail.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/pinctrl/pinmux.h>
+#include <linux/pinctrl/consumer.h>
+#include <linux/backlight.h>
+#include <video/display_timing.h>
+#include <video/of_display_timing.h>
+#include <video/videomode.h>
+
+#include "tilcdc_drv.h"
+
+struct panel_module {
+	struct tilcdc_module base;
+	struct tilcdc_panel_info *info;
+	struct display_timings *timings;
+	struct backlight_device *backlight;
+};
+#define to_panel_module(x) container_of(x, struct panel_module, base)
+
+
+/*
+ * Encoder:
+ */
+
+struct panel_encoder {
+	struct drm_encoder base;
+	struct panel_module *mod;
+};
+#define to_panel_encoder(x) container_of(x, struct panel_encoder, base)
+
+
+static void panel_encoder_destroy(struct drm_encoder *encoder)
+{
+	struct panel_encoder *panel_encoder = to_panel_encoder(encoder);
+	drm_encoder_cleanup(encoder);
+	kfree(panel_encoder);
+}
+
+static void panel_encoder_dpms(struct drm_encoder *encoder, int mode)
+{
+	struct panel_encoder *panel_encoder = to_panel_encoder(encoder);
+	struct backlight_device *backlight = panel_encoder->mod->backlight;
+
+	if (!backlight)
+		return;
+
+	backlight->props.power = mode == DRM_MODE_DPMS_ON
+				     ? FB_BLANK_UNBLANK : FB_BLANK_POWERDOWN;
+	backlight_update_status(backlight);
+}
+
+static bool panel_encoder_mode_fixup(struct drm_encoder *encoder,
+		const struct drm_display_mode *mode,
+		struct drm_display_mode *adjusted_mode)
+{
+	/* nothing needed */
+	return true;
+}
+
+static void panel_encoder_prepare(struct drm_encoder *encoder)
+{
+	struct panel_encoder *panel_encoder = to_panel_encoder(encoder);
+	panel_encoder_dpms(encoder, DRM_MODE_DPMS_OFF);
+	tilcdc_crtc_set_panel_info(encoder->crtc, panel_encoder->mod->info);
+}
+
+static void panel_encoder_commit(struct drm_encoder *encoder)
+{
+	panel_encoder_dpms(encoder, DRM_MODE_DPMS_ON);
+}
+
+static void panel_encoder_mode_set(struct drm_encoder *encoder,
+		struct drm_display_mode *mode,
+		struct drm_display_mode *adjusted_mode)
+{
+	/* nothing needed */
+}
+
+static const struct drm_encoder_funcs panel_encoder_funcs = {
+		.destroy        = panel_encoder_destroy,
+};
+
+static const struct drm_encoder_helper_funcs panel_encoder_helper_funcs = {
+		.dpms           = panel_encoder_dpms,
+		.mode_fixup     = panel_encoder_mode_fixup,
+		.prepare        = panel_encoder_prepare,
+		.commit         = panel_encoder_commit,
+		.mode_set       = panel_encoder_mode_set,
+};
+
+static struct drm_encoder *panel_encoder_create(struct drm_device *dev,
+		struct panel_module *mod)
+{
+	struct panel_encoder *panel_encoder;
+	struct drm_encoder *encoder;
+	int ret;
+
+	panel_encoder = kzalloc(sizeof(*panel_encoder), GFP_KERNEL);
+	if (!panel_encoder) {
+		dev_err(dev->dev, "allocation failed\n");
+		return NULL;
+	}
+
+	panel_encoder->mod = mod;
+
+	encoder = &panel_encoder->base;
+	encoder->possible_crtcs = 1;
+
+	ret = drm_encoder_init(dev, encoder, &panel_encoder_funcs,
+			DRM_MODE_ENCODER_LVDS);
+	if (ret < 0)
+		goto fail;
+
+	drm_encoder_helper_add(encoder, &panel_encoder_helper_funcs);
+
+	return encoder;
+
+fail:
+	panel_encoder_destroy(encoder);
+	return NULL;
+}
+
+/*
+ * Connector:
+ */
+
+struct panel_connector {
+	struct drm_connector base;
+
+	struct drm_encoder *encoder;  /* our connected encoder */
+	struct panel_module *mod;
+};
+#define to_panel_connector(x) container_of(x, struct panel_connector, base)
+
+
+static void panel_connector_destroy(struct drm_connector *connector)
+{
+	struct panel_connector *panel_connector = to_panel_connector(connector);
+	drm_connector_cleanup(connector);
+	kfree(panel_connector);
+}
+
+static enum drm_connector_status panel_connector_detect(
+		struct drm_connector *connector,
+		bool force)
+{
+	return connector_status_connected;
+}
+
+static int panel_connector_get_modes(struct drm_connector *connector)
+{
+	struct drm_device *dev = connector->dev;
+	struct panel_connector *panel_connector = to_panel_connector(connector);
+	struct display_timings *timings = panel_connector->mod->timings;
+	int i;
+
+	for (i = 0; i < timings->num_timings; i++) {
+		struct drm_display_mode *mode = drm_mode_create(dev);
+		struct videomode vm;
+
+		if (videomode_from_timing(timings, &vm, i))
+			break;
+
+		drm_display_mode_from_videomode(&vm, mode);
+
+		mode->type = DRM_MODE_TYPE_DRIVER;
+
+		if (timings->native_mode == i)
+			mode->type |= DRM_MODE_TYPE_PREFERRED;
+
+		drm_mode_set_name(mode);
+		drm_mode_probed_add(connector, mode);
+	}
+
+	return i;
+}
+
+static int panel_connector_mode_valid(struct drm_connector *connector,
+		  struct drm_display_mode *mode)
+{
+	struct tilcdc_drm_private *priv = connector->dev->dev_private;
+	/* our only constraints are what the crtc can generate: */
+	return tilcdc_crtc_mode_valid(priv->crtc, mode);
+}
+
+static struct drm_encoder *panel_connector_best_encoder(
+		struct drm_connector *connector)
+{
+	struct panel_connector *panel_connector = to_panel_connector(connector);
+	return panel_connector->encoder;
+}
+
+static const struct drm_connector_funcs panel_connector_funcs = {
+	.destroy            = panel_connector_destroy,
+	.dpms               = drm_helper_connector_dpms,
+	.detect             = panel_connector_detect,
+	.fill_modes         = drm_helper_probe_single_connector_modes,
+};
+
+static const struct drm_connector_helper_funcs panel_connector_helper_funcs = {
+	.get_modes          = panel_connector_get_modes,
+	.mode_valid         = panel_connector_mode_valid,
+	.best_encoder       = panel_connector_best_encoder,
+};
+
+static struct drm_connector *panel_connector_create(struct drm_device *dev,
+		struct panel_module *mod, struct drm_encoder *encoder)
+{
+	struct panel_connector *panel_connector;
+	struct drm_connector *connector;
+	int ret;
+
+	panel_connector = kzalloc(sizeof(*panel_connector), GFP_KERNEL);
+	if (!panel_connector) {
+		dev_err(dev->dev, "allocation failed\n");
+		return NULL;
+	}
+
+	panel_connector->encoder = encoder;
+	panel_connector->mod = mod;
+
+	connector = &panel_connector->base;
+
+	drm_connector_init(dev, connector, &panel_connector_funcs,
+			DRM_MODE_CONNECTOR_LVDS);
+	drm_connector_helper_add(connector, &panel_connector_helper_funcs);
+
+	connector->interlace_allowed = 0;
+	connector->doublescan_allowed = 0;
+
+	ret = drm_mode_connector_attach_encoder(connector, encoder);
+	if (ret)
+		goto fail;
+
+	drm_sysfs_connector_add(connector);
+
+	return connector;
+
+fail:
+	panel_connector_destroy(connector);
+	return NULL;
+}
+
+/*
+ * Module:
+ */
+
+static int panel_modeset_init(struct tilcdc_module *mod, struct drm_device *dev)
+{
+	struct panel_module *panel_mod = to_panel_module(mod);
+	struct tilcdc_drm_private *priv = dev->dev_private;
+	struct drm_encoder *encoder;
+	struct drm_connector *connector;
+
+	encoder = panel_encoder_create(dev, panel_mod);
+	if (!encoder)
+		return -ENOMEM;
+
+	connector = panel_connector_create(dev, panel_mod, encoder);
+	if (!connector)
+		return -ENOMEM;
+
+	priv->encoders[priv->num_encoders++] = encoder;
+	priv->connectors[priv->num_connectors++] = connector;
+
+	return 0;
+}
+
+static void panel_destroy(struct tilcdc_module *mod)
+{
+	struct panel_module *panel_mod = to_panel_module(mod);
+
+	if (panel_mod->timings) {
+		display_timings_release(panel_mod->timings);
+		kfree(panel_mod->timings);
+	}
+
+	tilcdc_module_cleanup(mod);
+	kfree(panel_mod->info);
+	kfree(panel_mod);
+}
+
+static const struct tilcdc_module_ops panel_module_ops = {
+		.modeset_init = panel_modeset_init,
+		.destroy = panel_destroy,
+};
+
+/*
+ * Device:
+ */
+
+/* maybe move this somewhere common if it is needed by other outputs? */
+static struct tilcdc_panel_info * of_get_panel_info(struct device_node *np)
+{
+	struct device_node *info_np;
+	struct tilcdc_panel_info *info;
+	int ret = 0;
+
+	if (!np) {
+		pr_err("%s: no devicenode given\n", __func__);
+		return NULL;
+	}
+
+	info_np = of_get_child_by_name(np, "panel-info");
+	if (!info_np) {
+		pr_err("%s: could not find panel-info node\n", __func__);
+		return NULL;
+	}
+
+	info = kzalloc(sizeof(*info), GFP_KERNEL);
+	if (!info) {
+		pr_err("%s: allocation failed\n", __func__);
+		return NULL;
+	}
+
+	ret |= of_property_read_u32(info_np, "ac-bias", &info->ac_bias);
+	ret |= of_property_read_u32(info_np, "ac-bias-intrpt", &info->ac_bias_intrpt);
+	ret |= of_property_read_u32(info_np, "dma-burst-sz", &info->dma_burst_sz);
+	ret |= of_property_read_u32(info_np, "bpp", &info->bpp);
+	ret |= of_property_read_u32(info_np, "fdd", &info->fdd);
+	ret |= of_property_read_u32(info_np, "sync-edge", &info->sync_edge);
+	ret |= of_property_read_u32(info_np, "sync-ctrl", &info->sync_ctrl);
+	ret |= of_property_read_u32(info_np, "raster-order", &info->raster_order);
+	ret |= of_property_read_u32(info_np, "fifo-th", &info->fifo_th);
+
+	/* optional: */
+	info->tft_alt_mode      = of_property_read_bool(info_np, "tft-alt-mode");
+	info->invert_pxl_clk    = of_property_read_bool(info_np, "invert-pxl-clk");
+
+	if (ret) {
+		pr_err("%s: error reading panel-info properties\n", __func__);
+		kfree(info);
+		return NULL;
+	}
+
+	return info;
+}
+
+static struct of_device_id panel_of_match[];
+
+static int panel_probe(struct platform_device *pdev)
+{
+	struct device_node *node = pdev->dev.of_node;
+	struct panel_module *panel_mod;
+	struct tilcdc_module *mod;
+	struct pinctrl *pinctrl;
+	int ret = -EINVAL;
+
+
+	/* bail out early if no DT data: */
+	if (!node) {
+		dev_err(&pdev->dev, "device-tree data is missing\n");
+		return -ENXIO;
+	}
+
+	panel_mod = kzalloc(sizeof(*panel_mod), GFP_KERNEL);
+	if (!panel_mod)
+		return -ENOMEM;
+
+	mod = &panel_mod->base;
+
+	tilcdc_module_init(mod, "panel", &panel_module_ops);
+
+	pinctrl = devm_pinctrl_get_select_default(&pdev->dev);
+	if (IS_ERR(pinctrl))
+		dev_warn(&pdev->dev, "pins are not configured\n");
+
+
+	panel_mod->timings = of_get_display_timings(node);
+	if (!panel_mod->timings) {
+		dev_err(&pdev->dev, "could not get panel timings\n");
+		goto fail;
+	}
+
+	panel_mod->info = of_get_panel_info(node);
+	if (!panel_mod->info) {
+		dev_err(&pdev->dev, "could not get panel info\n");
+		goto fail;
+	}
+
+	panel_mod->backlight = of_find_backlight_by_node(node);
+	if (panel_mod->backlight)
+		dev_info(&pdev->dev, "found backlight\n");
+
+	return 0;
+
+fail:
+	panel_destroy(mod);
+	return ret;
+}
+
+static int panel_remove(struct platform_device *pdev)
+{
+	return 0;
+}
+
+static struct of_device_id panel_of_match[] = {
+		{ .compatible = "ti,tilcdc,panel", },
+		{ },
+};
+MODULE_DEVICE_TABLE(of, panel_of_match);
+
+struct platform_driver panel_driver = {
+	.probe = panel_probe,
+	.remove = panel_remove,
+	.driver = {
+		.owner = THIS_MODULE,
+		.name = "panel",
+		.of_match_table = panel_of_match,
+	},
+};
+
+int __init tilcdc_panel_init(void)
+{
+	return platform_driver_register(&panel_driver);
+}
+
+void __exit tilcdc_panel_fini(void)
+{
+	platform_driver_unregister(&panel_driver);
+}
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_panel.h b/drivers/gpu/drm/tilcdc/tilcdc_panel.h
new file mode 100644
index 0000000..7db40aa
--- /dev/null
+++ b/drivers/gpu/drm/tilcdc/tilcdc_panel.h
@@ -0,0 +1,26 @@
+/*
+ * Copyright (C) 2012 Texas Instruments
+ * Author: Rob Clark <robdclark@gmail.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef __TILCDC_PANEL_H__
+#define __TILCDC_PANEL_H__
+
+/* sub-module for generic lcd panel output */
+
+int tilcdc_panel_init(void);
+void tilcdc_panel_fini(void);
+
+#endif /* __TILCDC_PANEL_H__ */
-- 
cgit v0.10.2


From 1d7aaa0cfe3d1b5ee23958fce71973e3ae9e52e4 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sat, 9 Feb 2013 21:03:42 +0100
Subject: drm/i915: detect wrong MCH watermark values

Some early bios versions seem to ship with the wrong tuning values for
the MCH, possible resulting in pipe underruns under load. Especially
on DP outputs this can lead to black screen, since DP really doesn't
like an occasional whack from an underrun.

Unfortunately the registers seem to be locked after boot, so the only
thing we can do is politely point out issues and suggest a BIOS
upgrade.

Arthur Runyan pointed us at this issue while discussion DP bugs - thus
far no confirmation from a bug report yet that it helps. But at least
some of my machines here have wrong values, so this might be useful in
understanding bug reports.

v2: After a bit more discussion with Art and Ben we've decided to only
the check the watermark values, since the OREF ones could be be a
notch more aggressive on certain machines.

Cc: Ben Widawsky <ben@bwidawsk.net>
Cc: Runyan, Arthur J <arthur.j.runyan@intel.com>
Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index b9fdc99..d7542cd 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -1235,6 +1235,10 @@
 #define   MAD_DIMM_A_SIZE_SHIFT		0
 #define   MAD_DIMM_A_SIZE_MASK		(0xff << MAD_DIMM_A_SIZE_SHIFT)
 
+/** snb MCH registers for priority tuning */
+#define MCH_SSKPD			(MCHBAR_MIRROR_BASE_SNB + 0x5d10)
+#define   MCH_SSKPD_WM0_MASK		0x3f
+#define   MCH_SSKPD_WM0_VAL		0xc
 
 /* Clocking configuration register */
 #define CLKCFG			0x10c00
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index 7c9a6d1..3bcc745 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -3584,6 +3584,19 @@ static void cpt_init_clock_gating(struct drm_device *dev)
 	}
 }
 
+static void gen6_check_mch_setup(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t tmp;
+
+	tmp = I915_READ(MCH_SSKPD);
+	if ((tmp & MCH_SSKPD_WM0_MASK) != MCH_SSKPD_WM0_VAL) {
+		DRM_INFO("Wrong MCH_SSKPD value: 0x%08x\n", tmp);
+		DRM_INFO("This can cause pipe underruns and display issues.\n");
+		DRM_INFO("Please upgrade your BIOS to fix this.\n");
+	}
+}
+
 static void gen6_init_clock_gating(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -3676,6 +3689,8 @@ static void gen6_init_clock_gating(struct drm_device *dev)
 	I915_WRITE(GEN6_GT_MODE, _MASKED_BIT_ENABLE(GEN6_GT_MODE_HI));
 
 	cpt_init_clock_gating(dev);
+
+	gen6_check_mch_setup(dev);
 }
 
 static void gen7_setup_fixed_func_scheduler(struct drm_i915_private *dev_priv)
@@ -3861,6 +3876,8 @@ static void ivybridge_init_clock_gating(struct drm_device *dev)
 	I915_WRITE(GEN6_MBCUNIT_SNPCR, snpcr);
 
 	cpt_init_clock_gating(dev);
+
+	gen6_check_mch_setup(dev);
 }
 
 static void valleyview_init_clock_gating(struct drm_device *dev)
-- 
cgit v0.10.2


From 24a1f16de97c4cf0029d9acd04be06db32208726 Mon Sep 17 00:00:00 2001
From: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Date: Fri, 8 Feb 2013 16:35:37 +0200
Subject: drm/i915: disable shared panel fitter for pipe

If encoder is switched off by BIOS, but the panel fitter is left on,
we never try to turn off the panel fitter and leave it still attached
to the pipe - which can cause blurry output elsewhere.

Based on work by Chris Wilson <chris@chris-wilson.co.uk>

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=58867
Cc: stable@vger.kernel.org
Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Tested-by: Andreas Sturmlechner <andreas.sturmlechner@gmail.com>
[danvet: Remove the redundant HAS_PCH_SPLIT check and add a tiny
comment.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 0dfecaf..fc9b987 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -3637,6 +3637,7 @@ static void i9xx_crtc_disable(struct drm_crtc *crtc)
 	struct intel_encoder *encoder;
 	int pipe = intel_crtc->pipe;
 	int plane = intel_crtc->plane;
+	u32 pctl;
 
 
 	if (!intel_crtc->active)
@@ -3656,6 +3657,13 @@ static void i9xx_crtc_disable(struct drm_crtc *crtc)
 
 	intel_disable_plane(dev_priv, plane, pipe);
 	intel_disable_pipe(dev_priv, pipe);
+
+	/* Disable pannel fitter if it is on this pipe. */
+	pctl = I915_READ(PFIT_CONTROL);
+	if ((pctl & PFIT_ENABLE) &&
+	    ((pctl & PFIT_PIPE_MASK) >> PFIT_PIPE_SHIFT) == pipe)
+		I915_WRITE(PFIT_CONTROL, 0);
+
 	intel_disable_pll(dev_priv, pipe);
 
 	intel_crtc->active = false;
-- 
cgit v0.10.2


From 9d6d9f19e8146fa24903cb561e204a22232740e3 Mon Sep 17 00:00:00 2001
From: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Date: Fri, 8 Feb 2013 16:35:38 +0200
Subject: drm/i915: clean up panel fitter handling in lvds

With the previous patch "drm/i915: disable shared panel fitter for
pipe" we now disable the panel fitter at the right spot in the modeset
sequence in the crtc functions on all platforms. Hence the disabling
in intel_disable_lvds is no longer required and potentially harmful
(since the plane is still enabled at this point).

Similarly on the enabling side we enable the panel fitter in the lvds
callback only once the plane is enabled. Which is too late. Hence move
this into a new intel_pre_enable_lvds callback.

Finally we can ditch lvds_encoder->pfit_dirty - this was required to
work around the crtc helper semantics, but with the new i915 modeset
infrastructure we should enable/disable the pfit only when enabling or
disabling the entire output pipeline. So separate state tracking for
the pfit is no longer required.

Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
[danvet: Bikeshed the commit message a bit to stress that now we
enable/disable the pfit on i9xx platforms at the right point of time
compared to the old code.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index fc9b987..6fe4d07 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -3615,6 +3615,11 @@ static void i9xx_crtc_enable(struct drm_crtc *crtc)
 	intel_update_watermarks(dev);
 
 	intel_enable_pll(dev_priv, pipe);
+
+	for_each_encoder_on_crtc(dev, crtc, encoder)
+		if (encoder->pre_enable)
+			encoder->pre_enable(encoder);
+
 	intel_enable_pipe(dev_priv, pipe, false);
 	intel_enable_plane(dev_priv, plane, pipe);
 
diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c
index bff78a3..c7154bf 100644
--- a/drivers/gpu/drm/i915/intel_lvds.c
+++ b/drivers/gpu/drm/i915/intel_lvds.c
@@ -51,7 +51,6 @@ struct intel_lvds_encoder {
 
 	u32 pfit_control;
 	u32 pfit_pgm_ratios;
-	bool pfit_dirty;
 	bool is_dual_link;
 	u32 reg;
 
@@ -151,6 +150,29 @@ static void intel_pre_pll_enable_lvds(struct intel_encoder *encoder)
 	I915_WRITE(lvds_encoder->reg, temp);
 }
 
+static void intel_pre_enable_lvds(struct intel_encoder *encoder)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct intel_lvds_encoder *enc = to_lvds_encoder(&encoder->base);
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (HAS_PCH_SPLIT(dev) || !enc->pfit_control)
+		return;
+
+	/*
+	 * Enable automatic panel scaling so that non-native modes
+	 * fill the screen.  The panel fitter should only be
+	 * adjusted whilst the pipe is disabled, according to
+	 * register description and PRM.
+	 */
+	DRM_DEBUG_KMS("applying panel-fitter: %x, %x\n",
+		      enc->pfit_control,
+		      enc->pfit_pgm_ratios);
+
+	I915_WRITE(PFIT_PGM_RATIOS, enc->pfit_pgm_ratios);
+	I915_WRITE(PFIT_CONTROL, enc->pfit_control);
+}
+
 /**
  * Sets the power state for the panel.
  */
@@ -172,22 +194,6 @@ static void intel_enable_lvds(struct intel_encoder *encoder)
 
 	I915_WRITE(lvds_encoder->reg, I915_READ(lvds_encoder->reg) | LVDS_PORT_EN);
 
-	if (lvds_encoder->pfit_dirty) {
-		/*
-		 * Enable automatic panel scaling so that non-native modes
-		 * fill the screen.  The panel fitter should only be
-		 * adjusted whilst the pipe is disabled, according to
-		 * register description and PRM.
-		 */
-		DRM_DEBUG_KMS("applying panel-fitter: %x, %x\n",
-			      lvds_encoder->pfit_control,
-			      lvds_encoder->pfit_pgm_ratios);
-
-		I915_WRITE(PFIT_PGM_RATIOS, lvds_encoder->pfit_pgm_ratios);
-		I915_WRITE(PFIT_CONTROL, lvds_encoder->pfit_control);
-		lvds_encoder->pfit_dirty = false;
-	}
-
 	I915_WRITE(ctl_reg, I915_READ(ctl_reg) | POWER_TARGET_ON);
 	POSTING_READ(lvds_encoder->reg);
 	if (wait_for((I915_READ(stat_reg) & PP_ON) != 0, 1000))
@@ -217,11 +223,6 @@ static void intel_disable_lvds(struct intel_encoder *encoder)
 	if (wait_for((I915_READ(stat_reg) & PP_ON) == 0, 1000))
 		DRM_ERROR("timed out waiting for panel to power off\n");
 
-	if (lvds_encoder->pfit_control) {
-		I915_WRITE(PFIT_CONTROL, 0);
-		lvds_encoder->pfit_dirty = true;
-	}
-
 	I915_WRITE(lvds_encoder->reg, I915_READ(lvds_encoder->reg) & ~LVDS_PORT_EN);
 	POSTING_READ(lvds_encoder->reg);
 }
@@ -461,7 +462,6 @@ out:
 	    pfit_pgm_ratios != lvds_encoder->pfit_pgm_ratios) {
 		lvds_encoder->pfit_control = pfit_control;
 		lvds_encoder->pfit_pgm_ratios = pfit_pgm_ratios;
-		lvds_encoder->pfit_dirty = true;
 	}
 	dev_priv->lvds_border_bits = border;
 
@@ -1101,6 +1101,7 @@ bool intel_lvds_init(struct drm_device *dev)
 			 DRM_MODE_ENCODER_LVDS);
 
 	intel_encoder->enable = intel_enable_lvds;
+	intel_encoder->pre_enable = intel_pre_enable_lvds;
 	intel_encoder->pre_pll_enable = intel_pre_pll_enable_lvds;
 	intel_encoder->disable = intel_disable_lvds;
 	intel_encoder->get_hw_state = intel_lvds_get_hw_state;
-- 
cgit v0.10.2


From 2b1086cc584286838ec55de24c1fb749f97bc3db Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Tue, 12 Feb 2013 22:01:38 +0200
Subject: drm/i915: Print the pipe control page GTT address
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

We already print the HWS addresses during init, so do the same for the
pipe control page. Reduces guesswork when looking at hex addresses
later.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index 00525ff..9b8b058 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -467,6 +467,9 @@ init_pipe_control(struct intel_ring_buffer *ring)
 	if (pc->cpu_page == NULL)
 		goto err_unpin;
 
+	DRM_DEBUG_DRIVER("%s pipe control offset: 0x%08x\n",
+			 ring->name, pc->gtt_offset);
+
 	pc->obj = obj;
 	ring->private = pc;
 	return 0;
-- 
cgit v0.10.2


From 002d71f2627ebeecb102d3cfe74644c3d3d06539 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Tue, 12 Feb 2013 15:36:03 +0100
Subject: drm/i915: remove bogus mutex_unlock from error-path

This has been lost in the locking rework for intel_alloc_context_page:

commit 2c34b850ee1e9f86b41706149d0954eee58757a3
Author: Ben Widawsky <ben@bwidawsk.net>
Date:   Sat Mar 19 18:14:26 2011 -0700

    drm/i915: fix ilk rc6 teardown locking

Cc: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index 3bcc745..61fee7f 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -2289,7 +2289,6 @@ err_unpin:
 	i915_gem_object_unpin(ctx);
 err_unref:
 	drm_gem_object_unreference(&ctx->base);
-	mutex_unlock(&dev->struct_mutex);
 	return NULL;
 }
 
-- 
cgit v0.10.2


From 9abdda74acb0027561747fa40a1e8d387a3e08c2 Mon Sep 17 00:00:00 2001
From: Damien Lespiau <damien.lespiau@intel.com>
Date: Wed, 13 Feb 2013 13:29:23 +0000
Subject: drm/i915: Fix a typo in a intel_modeset_stage_output_state() comment

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 6fe4d07..d28f43f 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -7898,7 +7898,7 @@ intel_modeset_stage_output_state(struct drm_device *dev,
 	struct intel_encoder *encoder;
 	int count, ro;
 
-	/* The upper layers ensure that we either disabl a crtc or have a list
+	/* The upper layers ensure that we either disable a crtc or have a list
 	 * of connectors. For paranoia, double-check this. */
 	WARN_ON(!set->fb && (set->num_connectors != 0));
 	WARN_ON(set->fb && (set->num_connectors == 0));
-- 
cgit v0.10.2


From 3d56e2d62b333aab434cba5a40e2353cac61154f Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Tue, 29 Jan 2013 18:13:33 +0200
Subject: drm/i915: Kill obj->pending_flip
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The pending flip mask no longer set anywhere, so trying to wait for
while it's non-zero is a no-op. Remove it completely.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 0f1d15b..b1f6a09 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -1215,13 +1215,6 @@ struct drm_i915_gem_object {
 
 	/** for phy allocated objects */
 	struct drm_i915_gem_phys_object *phys_obj;
-
-	/**
-	 * Number of crtcs where this object is currently the fb, but
-	 * will be page flipped away on the next vblank.  When it
-	 * reaches 0, dev_priv->pending_flip_queue will be woken up.
-	 */
-	atomic_t pending_flip;
 };
 #define to_gem_object(obj) (&((struct drm_i915_gem_object *)(obj))->base)
 
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index d28f43f..6dfb49c 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -2226,12 +2226,6 @@ intel_finish_fb(struct drm_framebuffer *old_fb)
 	bool was_interruptible = dev_priv->mm.interruptible;
 	int ret;
 
-	WARN_ON(waitqueue_active(&dev_priv->pending_flip_queue));
-
-	wait_event(dev_priv->pending_flip_queue,
-		   i915_reset_in_progress(&dev_priv->gpu_error) ||
-		   atomic_read(&obj->pending_flip) == 0);
-
 	/* Big Hammer, we also need to ensure that any pending
 	 * MI_WAIT_FOR_EVENT inside a user batch buffer on the
 	 * current scanout is retired before unpinning the old
-- 
cgit v0.10.2


From 10d8373064762b21e16e73bc7dfde49d68f4e31f Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Tue, 29 Jan 2013 18:13:34 +0200
Subject: drm/i915: Don't wait for page flips if there was GPU reset
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

If a GPU reset occurs while a page flip has been submitted to the ring,
the flip will never complete once the ring has been reset.

The GPU reset can be detected by sampling the reset_counter before the
flip is submitted, and then while waiting for the flip, the sampled
counter is compared with the current reset_counter value.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
[danvet: Move the reset_counter assignment to an earlier place in
common code as discussed on the mailing list.]
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=60140
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 6dfb49c..593c668 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -2868,10 +2868,12 @@ static bool intel_crtc_has_pending_flip(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	unsigned long flags;
 	bool pending;
 
-	if (i915_reset_in_progress(&dev_priv->gpu_error))
+	if (i915_reset_in_progress(&dev_priv->gpu_error) ||
+	    intel_crtc->reset_counter != atomic_read(&dev_priv->gpu_error.reset_counter))
 		return false;
 
 	spin_lock_irqsave(&dev->event_lock, flags);
@@ -7250,6 +7252,7 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc,
 	work->enable_stall_check = true;
 
 	atomic_inc(&intel_crtc->unpin_work_count);
+	intel_crtc->reset_counter = atomic_read(&dev_priv->gpu_error.reset_counter);
 
 	ret = dev_priv->display.queue_flip(dev, crtc, fb, obj);
 	if (ret)
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 13afb37d..006b5aa 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -235,6 +235,9 @@ struct intel_crtc {
 	/* We can share PLLs across outputs if the timings match */
 	struct intel_pch_pll *pch_pll;
 	uint32_t ddi_pll_sel;
+
+	/* reset counter value when the last flip was submitted */
+	unsigned int reset_counter;
 };
 
 struct intel_plane {
-- 
cgit v0.10.2


From 769ce4643b775496166d09c227b070cc38f9e205 Mon Sep 17 00:00:00 2001
From: Imre Deak <imre.deak@intel.com>
Date: Wed, 13 Feb 2013 21:56:05 +0200
Subject: drm/i915: don't clflush gem objects in stolen memory

As explained by Chris Wilson gem objects in stolen memory are always
coherent with the GPU so we don't need to ever flush the CPU caches for
these.

This fixes a breakage - at least with the compact sg patches applied -
during the resume/restore gtt mappings path, when we tried to clflush an
FB object in stolen memory, but since stolen objects don't have backing
pages we passed an invalid page pointer to drm_clflush_page().

Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index d8618f3..a367801 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -3021,6 +3021,13 @@ i915_gem_clflush_object(struct drm_i915_gem_object *obj)
 	if (obj->pages == NULL)
 		return;
 
+	/*
+	 * Stolen memory is always coherent with the GPU as it is explicitly
+	 * marked as wc by the system, or the system is cache-coherent.
+	 */
+	if (obj->stolen)
+		return;
+
 	/* If the GPU is snooping the contents of the CPU cache,
 	 * we do not need to manually clear the CPU cache lines.  However,
 	 * the caches are only snooped when the render cache is
-- 
cgit v0.10.2


From c51a6bc5f6d328926a9a4a1247c5030faf190a80 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 31 Jan 2013 19:43:37 +0200
Subject: drm: Fill depth/bits_per_pixel for C8 format
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Set depth/bits_per_pixel to 8 for C8 format.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Acked-by: Dave Airlie <airlied@gmail.com>
Cc: stable@vger.kernel.org
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index 9c797f6..ff7344c 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -3856,6 +3856,7 @@ void drm_fb_get_bpp_depth(uint32_t format, unsigned int *depth,
 			  int *bpp)
 {
 	switch (format) {
+	case DRM_FORMAT_C8:
 	case DRM_FORMAT_RGB332:
 	case DRM_FORMAT_BGR233:
 		*depth = 8;
-- 
cgit v0.10.2


From d84f031bd230fdf9c3b7734940c859bf28b90219 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 31 Jan 2013 19:43:38 +0200
Subject: drm: Use C8 instead of RGB332 when determining the format from
 depth/bpp
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Support for real RGB332 is a rarity, most hardware only really support
C8. So use C8 instead of RGB332 when determining the format based on
depth/bpp.

This fixes 8bpp fbcon on i915, since i915 will only accept C8 and not
RGB332.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=59572
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Acked-by: Dave Airlie <airlied@gmail.com>
Tested-by: mlsemon35@gmail.com
Cc: stable@vger.kernel.org
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index ff7344c..826a5ca 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -2253,7 +2253,7 @@ uint32_t drm_mode_legacy_fb_format(uint32_t bpp, uint32_t depth)
 
 	switch (bpp) {
 	case 8:
-		fmt = DRM_FORMAT_RGB332;
+		fmt = DRM_FORMAT_C8;
 		break;
 	case 16:
 		if (depth == 15)
-- 
cgit v0.10.2


From fdfa175d0a9cfa2082ce24e67e284e5acbba452a Mon Sep 17 00:00:00 2001
From: Jani Nikula <jani.nikula@intel.com>
Date: Thu, 14 Feb 2013 11:23:35 +0200
Subject: drm/i915: add missing \n to UTS_RELEASE in the error_state

Amending
commit 4518f611ba21ba165ea3714055938a8984a44ff9
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Wed Jan 23 16:16:35 2013 +0100

    drm/i915: dump UTS_RELEASE into the error_state

CC: stable@vger.kernel.org
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index e0e77b8..41c65ca 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -694,7 +694,7 @@ static int i915_error_state(struct seq_file *m, void *unused)
 
 	seq_printf(m, "Time: %ld s %ld us\n", error->time.tv_sec,
 		   error->time.tv_usec);
-	seq_printf(m, "Kernel: " UTS_RELEASE);
+	seq_printf(m, "Kernel: " UTS_RELEASE "\n");
 	seq_printf(m, "PCI ID: 0x%04x\n", dev->pci_device);
 	seq_printf(m, "EIR: 0x%08x\n", error->eir);
 	seq_printf(m, "IER: 0x%08x\n", error->ier);
-- 
cgit v0.10.2


From 3e68320ef84528604d971afc3cadfbe208bef001 Mon Sep 17 00:00:00 2001
From: Damien Lespiau <damien.lespiau@intel.com>
Date: Tue, 11 Dec 2012 18:48:29 +0000
Subject: drm/i915: Preserve the FDI line reversal override bit on CPT

The FDI link has supported link reversal to make the PCB layout
engineer's life easier for quite a while and we have always presered
this bit as we programmed FDI_RX_CTL with a read/modify/write sequence.

We're trying to take a bit more control over what the BIOS leaves in
various register and with the introduction of DDI, started to program
FDI_RX_CTL fully.

There's a fused bit to indicate DMI link reversal and FDI defaults to
mirroring that configuration. We have a bit to override that behaviour
that we need to preserve from the BIOS.

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@gmail.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Cc: stable@vger.kernel.org
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index b1f6a09..e95337c 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -1040,7 +1040,7 @@ typedef struct drm_i915_private {
 	bool hw_contexts_disabled;
 	uint32_t hw_context_size;
 
-	bool fdi_rx_polarity_reversed;
+	u32 fdi_rx_config;
 
 	struct i915_suspend_saved_registers regfile;
 
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index d7542cd..64919ee 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -3911,7 +3911,7 @@
 #define  FDI_10BPC                      (1<<16)
 #define  FDI_6BPC                       (2<<16)
 #define  FDI_12BPC                      (3<<16)
-#define  FDI_LINK_REVERSE_OVERWRITE     (1<<15)
+#define  FDI_RX_LINK_REVERSAL_OVERRIDE  (1<<15)
 #define  FDI_DMI_LINK_REVERSE_MASK      (1<<14)
 #define  FDI_RX_PLL_ENABLE              (1<<13)
 #define  FDI_FS_ERR_CORRECT_ENABLE      (1<<11)
diff --git a/drivers/gpu/drm/i915/intel_crt.c b/drivers/gpu/drm/i915/intel_crt.c
index 68e79f3..cfc9687 100644
--- a/drivers/gpu/drm/i915/intel_crt.c
+++ b/drivers/gpu/drm/i915/intel_crt.c
@@ -801,10 +801,14 @@ void intel_crt_init(struct drm_device *dev)
 	dev_priv->hotplug_supported_mask |= CRT_HOTPLUG_INT_STATUS;
 
 	/*
-	 * TODO: find a proper way to discover whether we need to set the
-	 * polarity reversal bit or not, instead of relying on the BIOS.
+	 * TODO: find a proper way to discover whether we need to set the the
+	 * polarity and link reversal bits or not, instead of relying on the
+	 * BIOS.
 	 */
-	if (HAS_PCH_LPT(dev))
-		dev_priv->fdi_rx_polarity_reversed =
-		     !!(I915_READ(_FDI_RXA_CTL) & FDI_RX_POLARITY_REVERSED_LPT);
+	if (HAS_PCH_LPT(dev)) {
+		u32 fdi_config = FDI_RX_POLARITY_REVERSED_LPT |
+				 FDI_RX_LINK_REVERSAL_OVERRIDE;
+
+		dev_priv->fdi_rx_config = I915_READ(_FDI_RXA_CTL) & fdi_config;
+	}
 }
diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c
index cedf4ab..89bdb1b 100644
--- a/drivers/gpu/drm/i915/intel_ddi.c
+++ b/drivers/gpu/drm/i915/intel_ddi.c
@@ -180,10 +180,8 @@ void hsw_fdi_link_train(struct drm_crtc *crtc)
 				  FDI_RX_TP1_TO_TP2_48 | FDI_RX_FDI_DELAY_90);
 
 	/* Enable the PCH Receiver FDI PLL */
-	rx_ctl_val = FDI_RX_PLL_ENABLE | FDI_RX_ENHANCE_FRAME_ENABLE |
-		     ((intel_crtc->fdi_lanes - 1) << 19);
-	if (dev_priv->fdi_rx_polarity_reversed)
-		rx_ctl_val |= FDI_RX_POLARITY_REVERSED_LPT;
+	rx_ctl_val = dev_priv->fdi_rx_config | FDI_RX_ENHANCE_FRAME_ENABLE |
+		     FDI_RX_PLL_ENABLE | ((intel_crtc->fdi_lanes - 1) << 19);
 	I915_WRITE(_FDI_RXA_CTL, rx_ctl_val);
 	POSTING_READ(_FDI_RXA_CTL);
 	udelay(220);
-- 
cgit v0.10.2


From 876a8cdf92b23d268275cdce4397df0c37dac3fe Mon Sep 17 00:00:00 2001
From: Damien Lespiau <damien.lespiau@intel.com>
Date: Tue, 11 Dec 2012 18:48:30 +0000
Subject: drm/i915: Preserve the DDI link reversal configuration

Similarly to:

  commit 6a0d1df3d3a0d2370541164eb0595fe35dcd6de3
  Author: Damien Lespiau <damien.lespiau@intel.com>
  Date:   Tue Dec 11 15:18:28 2012 +0000

      drm/i915: Preserve the FDI line reversal override bit on CPT

DDI port support lane reversal to easy the PCB layouting work. Let's
preserve the bit configured by the BIOS (until we find how to correctly
retrieve the information from the VBT, but this does sound more fragile
then just relying on the BIOS that has, hopefully, been validated
already.

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Cc: stable@vger.kernel.org
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 64919ee..2593fef 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -4510,6 +4510,7 @@
 #define  DDI_BUF_EMP_800MV_0DB_HSW		(7<<24)   /* Sel7 */
 #define  DDI_BUF_EMP_800MV_3_5DB_HSW		(8<<24)   /* Sel8 */
 #define  DDI_BUF_EMP_MASK			(0xf<<24)
+#define  DDI_BUF_PORT_REVERSAL			(1<<16)
 #define  DDI_BUF_IS_IDLE			(1<<7)
 #define  DDI_A_4_LANES				(1<<4)
 #define  DDI_PORT_WIDTH_X1			(0<<1)
diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c
index 89bdb1b..816c45c 100644
--- a/drivers/gpu/drm/i915/intel_ddi.c
+++ b/drivers/gpu/drm/i915/intel_ddi.c
@@ -203,7 +203,10 @@ void hsw_fdi_link_train(struct drm_crtc *crtc)
 					DP_TP_CTL_LINK_TRAIN_PAT1 |
 					DP_TP_CTL_ENABLE);
 
-		/* Configure and enable DDI_BUF_CTL for DDI E with next voltage */
+		/* Configure and enable DDI_BUF_CTL for DDI E with next voltage.
+		 * DDI E does not support port reversal, the functionality is
+		 * achieved on the PCH side in FDI_RX_CTL, so no need to set the
+		 * port reversal bit */
 		I915_WRITE(DDI_BUF_CTL(PORT_E),
 			   DDI_BUF_CTL_ENABLE |
 			   ((intel_crtc->fdi_lanes - 1) << 1) |
@@ -678,8 +681,11 @@ static void intel_ddi_mode_set(struct drm_encoder *encoder,
 	intel_crtc->eld_vld = false;
 	if (type == INTEL_OUTPUT_DISPLAYPORT || type == INTEL_OUTPUT_EDP) {
 		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+		struct intel_digital_port *intel_dig_port =
+			enc_to_dig_port(encoder);
 
-		intel_dp->DP = DDI_BUF_CTL_ENABLE | DDI_BUF_EMP_400MV_0DB_HSW;
+		intel_dp->DP = intel_dig_port->port_reversal |
+			       DDI_BUF_CTL_ENABLE | DDI_BUF_EMP_400MV_0DB_HSW;
 		switch (intel_dp->lane_count) {
 		case 1:
 			intel_dp->DP |= DDI_PORT_WIDTH_X1;
@@ -1302,11 +1308,15 @@ static void intel_enable_ddi(struct intel_encoder *intel_encoder)
 	uint32_t tmp;
 
 	if (type == INTEL_OUTPUT_HDMI) {
+		struct intel_digital_port *intel_dig_port =
+			enc_to_dig_port(encoder);
+
 		/* In HDMI/DVI mode, the port width, and swing/emphasis values
 		 * are ignored so nothing special needs to be done besides
 		 * enabling the port.
 		 */
-		I915_WRITE(DDI_BUF_CTL(port), DDI_BUF_CTL_ENABLE);
+		I915_WRITE(DDI_BUF_CTL(port),
+			   intel_dig_port->port_reversal | DDI_BUF_CTL_ENABLE);
 	} else if (type == INTEL_OUTPUT_EDP) {
 		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
 
@@ -1484,6 +1494,7 @@ static const struct drm_encoder_helper_funcs intel_ddi_helper_funcs = {
 
 void intel_ddi_init(struct drm_device *dev, enum port port)
 {
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_digital_port *intel_dig_port;
 	struct intel_encoder *intel_encoder;
 	struct drm_encoder *encoder;
@@ -1524,6 +1535,8 @@ void intel_ddi_init(struct drm_device *dev, enum port port)
 	intel_encoder->get_hw_state = intel_ddi_get_hw_state;
 
 	intel_dig_port->port = port;
+	intel_dig_port->port_reversal = I915_READ(DDI_BUF_CTL(port)) &
+					DDI_BUF_PORT_REVERSAL;
 	if (hdmi_connector)
 		intel_dig_port->hdmi.sdvox_reg = DDI_BUF_CTL(port);
 	else
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 006b5aa..005a91f 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -393,6 +393,7 @@ struct intel_dp {
 struct intel_digital_port {
 	struct intel_encoder base;
 	enum port port;
+	u32 port_reversal;
 	struct intel_dp dp;
 	struct intel_hdmi hdmi;
 };
-- 
cgit v0.10.2


From 53a7d2d15ef45fb892defaf624ad6db7d528d8ac Mon Sep 17 00:00:00 2001
From: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Date: Wed, 13 Feb 2013 22:20:21 +0100
Subject: drm/i915: Set i9xx lvds clock limits according to specifications

The Intel PRM says the M1 and M2 divisors must be in the range of 10-20 and 5-9.
Since we do all calculations based on them being register values (which are
subtracted by 2) we need to specify them accordingly.

Signed-off-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 593c668..aa31b96 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -168,8 +168,8 @@ static const intel_limit_t intel_limits_i9xx_lvds = {
 	.vco = { .min = 1400000, .max = 2800000 },
 	.n = { .min = 1, .max = 6 },
 	.m = { .min = 70, .max = 120 },
-	.m1 = { .min = 10, .max = 22 },
-	.m2 = { .min = 5, .max = 9 },
+	.m1 = { .min = 8, .max = 18 },
+	.m2 = { .min = 3, .max = 7 },
 	.p = { .min = 7, .max = 98 },
 	.p1 = { .min = 1, .max = 8 },
 	.p2 = { .dot_limit = 112000,
-- 
cgit v0.10.2


From 4f7dfb6788dd022446847fbbfbe45e13bedb5be2 Mon Sep 17 00:00:00 2001
From: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Date: Wed, 13 Feb 2013 22:20:22 +0100
Subject: drm/i915: Set i9xx sdvo clock limits according to specifications

The Intel PRM says the M1 and M2 divisors must be in the range of 10-20 and 5-9.
Since we do all calculations based on them being register values (which are
subtracted by 2) we need to specify them accordingly.

Signed-off-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=56359
Cc: stable@vger.kernel.org
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index aa31b96..3e6dadf 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -154,8 +154,8 @@ static const intel_limit_t intel_limits_i9xx_sdvo = {
 	.vco = { .min = 1400000, .max = 2800000 },
 	.n = { .min = 1, .max = 6 },
 	.m = { .min = 70, .max = 120 },
-	.m1 = { .min = 10, .max = 22 },
-	.m2 = { .min = 5, .max = 9 },
+	.m1 = { .min = 8, .max = 18 },
+	.m2 = { .min = 3, .max = 7 },
 	.p = { .min = 5, .max = 80 },
 	.p1 = { .min = 1, .max = 8 },
 	.p2 = { .dot_limit = 200000,
-- 
cgit v0.10.2


From b9e1faa7634e68bfcdff00a8e9378fcb662a7f30 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Thu, 14 Feb 2013 21:53:51 +0200
Subject: drm/i915: Fix PIPE_CONTROL DW/QW write through global GTT on IVB+
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The bit controlling whether PIPE_CONTROL DW/QW write targets
the global GTT or PPGTT moved moved from DW 2 bit 2 to
DW 1 bit 24 on IVB.

I verified on IVB that the fix is in fact effective. Without the fix
none of the scratch writes actually landed in the pipe control page.
With the fix the writes show up correctly.

v2: move PIPE_CONTROL_GLOBAL_GTT_IVB setup to where other flags are set

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 2593fef..7e77396 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -308,6 +308,7 @@
 #define   DISPLAY_PLANE_A           (0<<20)
 #define   DISPLAY_PLANE_B           (1<<20)
 #define GFX_OP_PIPE_CONTROL(len)	((0x3<<29)|(0x3<<27)|(0x2<<24)|(len-2))
+#define   PIPE_CONTROL_GLOBAL_GTT_IVB			(1<<24) /* gen7+ */
 #define   PIPE_CONTROL_CS_STALL				(1<<20)
 #define   PIPE_CONTROL_TLB_INVALIDATE			(1<<18)
 #define   PIPE_CONTROL_QW_WRITE				(1<<14)
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index 9b8b058..1d5d613 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -318,6 +318,7 @@ gen7_render_ring_flush(struct intel_ring_buffer *ring,
 		 * TLB invalidate requires a post-sync write.
 		 */
 		flags |= PIPE_CONTROL_QW_WRITE;
+		flags |= PIPE_CONTROL_GLOBAL_GTT_IVB;
 
 		/* Workaround: we must issue a pipe_control with CS-stall bit
 		 * set before a pipe_control command that has the state cache
@@ -331,7 +332,7 @@ gen7_render_ring_flush(struct intel_ring_buffer *ring,
 
 	intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(4));
 	intel_ring_emit(ring, flags);
-	intel_ring_emit(ring, scratch_addr | PIPE_CONTROL_GLOBAL_GTT);
+	intel_ring_emit(ring, scratch_addr);
 	intel_ring_emit(ring, 0);
 	intel_ring_advance(ring);
 
-- 
cgit v0.10.2


From eb32e4584d8e9d6cbec20550d4f91396de2cdb55 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Thu, 14 Feb 2013 19:46:07 +0100
Subject: drm/i915: Use HAS_L3_GPU_CACHE in i915_gem_l3_remap
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Yet another remnant ... this might explain why l3 remapping didn't
really work on HSW.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=57441
Spotted-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: stable@vger.kernel.org
Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index a367801..8413ffc 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -3872,7 +3872,7 @@ void i915_gem_l3_remap(struct drm_device *dev)
 	u32 misccpctl;
 	int i;
 
-	if (!IS_IVYBRIDGE(dev))
+	if (!HAS_L3_GPU_CACHE(dev))
 		return;
 
 	if (!dev_priv->l3_parity.remap_info)
-- 
cgit v0.10.2


From a168c293d5f4f89806dbe58055036e9933115ab3 Mon Sep 17 00:00:00 2001
From: Ben Widawsky <ben@bwidawsk.net>
Date: Thu, 14 Feb 2013 15:05:12 -0800
Subject: drm/i915: Print the hw context status is debugfs

Print out the HW context object information per ring. Even though the
existing code only utilizes the render ring, it's simple enough to
support future expansion.

I had this in a patch somewhere in a rev of the original implementation,
but I must have lost it.

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
[danvet: s/context/default context/ bikeshed applied.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index 41c65ca..7c65ab8 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -1484,7 +1484,8 @@ static int i915_context_status(struct seq_file *m, void *unused)
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
 	struct drm_device *dev = node->minor->dev;
 	drm_i915_private_t *dev_priv = dev->dev_private;
-	int ret;
+	struct intel_ring_buffer *ring;
+	int ret, i;
 
 	ret = mutex_lock_interruptible(&dev->mode_config.mutex);
 	if (ret)
@@ -1502,6 +1503,14 @@ static int i915_context_status(struct seq_file *m, void *unused)
 		seq_printf(m, "\n");
 	}
 
+	for_each_ring(ring, dev_priv, i) {
+		if (ring->default_context) {
+			seq_printf(m, "HW default context %s ring ", ring->name);
+			describe_obj(m, ring->default_context->obj);
+			seq_printf(m, "\n");
+		}
+	}
+
 	mutex_unlock(&dev->mode_config.mutex);
 
 	return 0;
-- 
cgit v0.10.2


From ac4199e0f047546aa40172785e26c82b54bbe811 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Fri, 15 Feb 2013 18:35:30 +0100
Subject: drm/i915: inverted brightness quirk for Acer Aspire 4736Z

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=53881
Cc: stable@vger.kernel.org
Cc: Jani Nikula <jani.nikula@intel.com>
Tested-by: Jani Monoses <jani@ubuntu.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 3e6dadf..ba1d913 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -8679,6 +8679,9 @@ static struct intel_quirk intel_quirks[] = {
 
 	/* Acer/Packard Bell NCL20 */
 	{ 0x2a42, 0x1025, 0x034b, quirk_invert_brightness },
+
+	/* Acer Aspire 4736Z */
+	{ 0x2a42, 0x1025, 0x0260, quirk_invert_brightness },
 };
 
 static void intel_init_quirks(struct drm_device *dev)
-- 
cgit v0.10.2


From 86d3efce2c37d3fb98f75f56f21e6ab75c745bb6 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <ville.syrjala@linux.intel.com>
Date: Fri, 18 Jan 2013 19:11:38 +0200
Subject: drm/i915: Implement pipe CSC based limited range RGB output
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

HSW no longer has the PIPECONF bit for limited range RGB output.
Instead the pipe CSC unit must be used to perform that task.

The CSC pre offset are set to 0, since the incoming data is full
[0:255] range RGB, the coefficients are programmed to compress the
data into [0:219] range, and then we use either the CSC_MODE black
screen offset bit, or the CSC post offsets to shift the data to
the correct [16:235] range.

Also have to change the confiuration of all planes so that the
data is sent through the pipe CSC unit. For simplicity send the
plane data through the pipe CSC unit always, and in case full
range output is requested, the pipe CSC unit is set up with an
identity transform to pass the plane data through unchanged.

I've been told by some hardware people that the use of the pipe
CSC unit shouldn't result in any measurable increase in power
consumption numbers.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 7e77396..527b664 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -2949,6 +2949,7 @@
 #define   CURSOR_ENABLE		0x80000000
 #define   CURSOR_GAMMA_ENABLE	0x40000000
 #define   CURSOR_STRIDE_MASK	0x30000000
+#define   CURSOR_PIPE_CSC_ENABLE (1<<24)
 #define   CURSOR_FORMAT_SHIFT	24
 #define   CURSOR_FORMAT_MASK	(0x07 << CURSOR_FORMAT_SHIFT)
 #define   CURSOR_FORMAT_2C	(0x00 << CURSOR_FORMAT_SHIFT)
@@ -3010,6 +3011,7 @@
 #define   DISPPLANE_RGBA888			(0xf<<26)
 #define   DISPPLANE_STEREO_ENABLE		(1<<25)
 #define   DISPPLANE_STEREO_DISABLE		0
+#define   DISPPLANE_PIPE_CSC_ENABLE		(1<<24)
 #define   DISPPLANE_SEL_PIPE_SHIFT		24
 #define   DISPPLANE_SEL_PIPE_MASK		(3<<DISPPLANE_SEL_PIPE_SHIFT)
 #define   DISPPLANE_SEL_PIPE_A			0
@@ -3098,6 +3100,7 @@
 #define   DVS_FORMAT_RGBX101010	(1<<25)
 #define   DVS_FORMAT_RGBX888	(2<<25)
 #define   DVS_FORMAT_RGBX161616	(3<<25)
+#define   DVS_PIPE_CSC_ENABLE   (1<<24)
 #define   DVS_SOURCE_KEY	(1<<22)
 #define   DVS_RGB_ORDER_XBGR	(1<<20)
 #define   DVS_YUV_BYTE_ORDER_MASK (3<<16)
@@ -3165,7 +3168,7 @@
 #define   SPRITE_FORMAT_RGBX161616	(3<<25)
 #define   SPRITE_FORMAT_YUV444		(4<<25)
 #define   SPRITE_FORMAT_XR_BGR101010	(5<<25) /* Extended range */
-#define   SPRITE_CSC_ENABLE		(1<<24)
+#define   SPRITE_PIPE_CSC_ENABLE	(1<<24)
 #define   SPRITE_SOURCE_KEY		(1<<22)
 #define   SPRITE_RGB_ORDER_RGBX		(1<<20) /* only for 888 and 161616 */
 #define   SPRITE_YUV_TO_RGB_CSC_DISABLE	(1<<19)
@@ -4645,4 +4648,51 @@
 #define  WM_DBG_DISALLOW_MAXFIFO	(1<<1)
 #define  WM_DBG_DISALLOW_SPRITE		(1<<2)
 
+/* pipe CSC */
+#define _PIPE_A_CSC_COEFF_RY_GY	0x49010
+#define _PIPE_A_CSC_COEFF_BY	0x49014
+#define _PIPE_A_CSC_COEFF_RU_GU	0x49018
+#define _PIPE_A_CSC_COEFF_BU	0x4901c
+#define _PIPE_A_CSC_COEFF_RV_GV	0x49020
+#define _PIPE_A_CSC_COEFF_BV	0x49024
+#define _PIPE_A_CSC_MODE	0x49028
+#define _PIPE_A_CSC_PREOFF_HI	0x49030
+#define _PIPE_A_CSC_PREOFF_ME	0x49034
+#define _PIPE_A_CSC_PREOFF_LO	0x49038
+#define _PIPE_A_CSC_POSTOFF_HI	0x49040
+#define _PIPE_A_CSC_POSTOFF_ME	0x49044
+#define _PIPE_A_CSC_POSTOFF_LO	0x49048
+
+#define _PIPE_B_CSC_COEFF_RY_GY	0x49110
+#define _PIPE_B_CSC_COEFF_BY	0x49114
+#define _PIPE_B_CSC_COEFF_RU_GU	0x49118
+#define _PIPE_B_CSC_COEFF_BU	0x4911c
+#define _PIPE_B_CSC_COEFF_RV_GV	0x49120
+#define _PIPE_B_CSC_COEFF_BV	0x49124
+#define _PIPE_B_CSC_MODE	0x49128
+#define _PIPE_B_CSC_PREOFF_HI	0x49130
+#define _PIPE_B_CSC_PREOFF_ME	0x49134
+#define _PIPE_B_CSC_PREOFF_LO	0x49138
+#define _PIPE_B_CSC_POSTOFF_HI	0x49140
+#define _PIPE_B_CSC_POSTOFF_ME	0x49144
+#define _PIPE_B_CSC_POSTOFF_LO	0x49148
+
+#define CSC_BLACK_SCREEN_OFFSET (1 << 2)
+#define CSC_POSITION_BEFORE_GAMMA (1 << 1)
+#define CSC_MODE_YUV_TO_RGB (1 << 0)
+
+#define PIPE_CSC_COEFF_RY_GY(pipe) _PIPE(pipe, _PIPE_A_CSC_COEFF_RY_GY, _PIPE_B_CSC_COEFF_RY_GY)
+#define PIPE_CSC_COEFF_BY(pipe) _PIPE(pipe, _PIPE_A_CSC_COEFF_BY, _PIPE_B_CSC_COEFF_BY)
+#define PIPE_CSC_COEFF_RU_GU(pipe) _PIPE(pipe, _PIPE_A_CSC_COEFF_RU_GU, _PIPE_B_CSC_COEFF_RU_GU)
+#define PIPE_CSC_COEFF_BU(pipe) _PIPE(pipe, _PIPE_A_CSC_COEFF_BU, _PIPE_B_CSC_COEFF_BU)
+#define PIPE_CSC_COEFF_RV_GV(pipe) _PIPE(pipe, _PIPE_A_CSC_COEFF_RV_GV, _PIPE_B_CSC_COEFF_RV_GV)
+#define PIPE_CSC_COEFF_BV(pipe) _PIPE(pipe, _PIPE_A_CSC_COEFF_BV, _PIPE_B_CSC_COEFF_BV)
+#define PIPE_CSC_MODE(pipe) _PIPE(pipe, _PIPE_A_CSC_MODE, _PIPE_B_CSC_MODE)
+#define PIPE_CSC_PREOFF_HI(pipe) _PIPE(pipe, _PIPE_A_CSC_PREOFF_HI, _PIPE_B_CSC_PREOFF_HI)
+#define PIPE_CSC_PREOFF_ME(pipe) _PIPE(pipe, _PIPE_A_CSC_PREOFF_ME, _PIPE_B_CSC_PREOFF_ME)
+#define PIPE_CSC_PREOFF_LO(pipe) _PIPE(pipe, _PIPE_A_CSC_PREOFF_LO, _PIPE_B_CSC_PREOFF_LO)
+#define PIPE_CSC_POSTOFF_HI(pipe) _PIPE(pipe, _PIPE_A_CSC_POSTOFF_HI, _PIPE_B_CSC_POSTOFF_HI)
+#define PIPE_CSC_POSTOFF_ME(pipe) _PIPE(pipe, _PIPE_A_CSC_POSTOFF_ME, _PIPE_B_CSC_POSTOFF_ME)
+#define PIPE_CSC_POSTOFF_LO(pipe) _PIPE(pipe, _PIPE_A_CSC_POSTOFF_LO, _PIPE_B_CSC_POSTOFF_LO)
+
 #endif /* _I915_REG_H_ */
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index ba1d913..6eb3882 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -5126,6 +5126,71 @@ static void ironlake_set_pipeconf(struct drm_crtc *crtc,
 	POSTING_READ(PIPECONF(pipe));
 }
 
+/*
+ * Set up the pipe CSC unit.
+ *
+ * Currently only full range RGB to limited range RGB conversion
+ * is supported, but eventually this should handle various
+ * RGB<->YCbCr scenarios as well.
+ */
+static void intel_set_pipe_csc(struct drm_crtc *crtc,
+			       const struct drm_display_mode *adjusted_mode)
+{
+	struct drm_device *dev = crtc->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	int pipe = intel_crtc->pipe;
+	uint16_t coeff = 0x7800; /* 1.0 */
+
+	/*
+	 * TODO: Check what kind of values actually come out of the pipe
+	 * with these coeff/postoff values and adjust to get the best
+	 * accuracy. Perhaps we even need to take the bpc value into
+	 * consideration.
+	 */
+
+	if (adjusted_mode->private_flags & INTEL_MODE_LIMITED_COLOR_RANGE)
+		coeff = ((235 - 16) * (1 << 12) / 255) & 0xff8; /* 0.xxx... */
+
+	/*
+	 * GY/GU and RY/RU should be the other way around according
+	 * to BSpec, but reality doesn't agree. Just set them up in
+	 * a way that results in the correct picture.
+	 */
+	I915_WRITE(PIPE_CSC_COEFF_RY_GY(pipe), coeff << 16);
+	I915_WRITE(PIPE_CSC_COEFF_BY(pipe), 0);
+
+	I915_WRITE(PIPE_CSC_COEFF_RU_GU(pipe), coeff);
+	I915_WRITE(PIPE_CSC_COEFF_BU(pipe), 0);
+
+	I915_WRITE(PIPE_CSC_COEFF_RV_GV(pipe), 0);
+	I915_WRITE(PIPE_CSC_COEFF_BV(pipe), coeff << 16);
+
+	I915_WRITE(PIPE_CSC_PREOFF_HI(pipe), 0);
+	I915_WRITE(PIPE_CSC_PREOFF_ME(pipe), 0);
+	I915_WRITE(PIPE_CSC_PREOFF_LO(pipe), 0);
+
+	if (INTEL_INFO(dev)->gen > 6) {
+		uint16_t postoff = 0;
+
+		if (adjusted_mode->private_flags & INTEL_MODE_LIMITED_COLOR_RANGE)
+			postoff = (16 * (1 << 13) / 255) & 0x1fff;
+
+		I915_WRITE(PIPE_CSC_POSTOFF_HI(pipe), postoff);
+		I915_WRITE(PIPE_CSC_POSTOFF_ME(pipe), postoff);
+		I915_WRITE(PIPE_CSC_POSTOFF_LO(pipe), postoff);
+
+		I915_WRITE(PIPE_CSC_MODE(pipe), 0);
+	} else {
+		uint32_t mode = CSC_MODE_YUV_TO_RGB;
+
+		if (adjusted_mode->private_flags & INTEL_MODE_LIMITED_COLOR_RANGE)
+			mode |= CSC_BLACK_SCREEN_OFFSET;
+
+		I915_WRITE(PIPE_CSC_MODE(pipe), mode);
+	}
+}
+
 static void haswell_set_pipeconf(struct drm_crtc *crtc,
 				 struct drm_display_mode *adjusted_mode,
 				 bool dither)
@@ -5714,8 +5779,10 @@ static int haswell_crtc_mode_set(struct drm_crtc *crtc,
 
 	haswell_set_pipeconf(crtc, adjusted_mode, dither);
 
+	intel_set_pipe_csc(crtc, adjusted_mode);
+
 	/* Set up the display plane register */
-	I915_WRITE(DSPCNTR(plane), DISPPLANE_GAMMA_ENABLE);
+	I915_WRITE(DSPCNTR(plane), DISPPLANE_GAMMA_ENABLE | DISPPLANE_PIPE_CSC_ENABLE);
 	POSTING_READ(DSPCNTR(plane));
 
 	ret = intel_pipe_set_base(crtc, x, y, fb);
@@ -6120,6 +6187,8 @@ static void ivb_update_cursor(struct drm_crtc *crtc, u32 base)
 			cntl &= ~(CURSOR_MODE | MCURSOR_GAMMA_ENABLE);
 			cntl |= CURSOR_MODE_DISABLE;
 		}
+		if (IS_HASWELL(dev))
+			cntl |= CURSOR_PIPE_CSC_ENABLE;
 		I915_WRITE(CURCNTR_IVB(pipe), cntl);
 
 		intel_crtc->cursor_visible = visible;
diff --git a/drivers/gpu/drm/i915/intel_sprite.c b/drivers/gpu/drm/i915/intel_sprite.c
index 03cfd62..d086e48 100644
--- a/drivers/gpu/drm/i915/intel_sprite.c
+++ b/drivers/gpu/drm/i915/intel_sprite.c
@@ -90,6 +90,9 @@ ivb_update_plane(struct drm_plane *plane, struct drm_framebuffer *fb,
 	sprctl |= SPRITE_TRICKLE_FEED_DISABLE;
 	sprctl |= SPRITE_ENABLE;
 
+	if (IS_HASWELL(dev))
+		sprctl |= SPRITE_PIPE_CSC_ENABLE;
+
 	/* Sizes are 0 based */
 	src_w--;
 	src_h--;
-- 
cgit v0.10.2


From da88a5f7f7d434e2cde1b3e19d952e6d84533662 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Wed, 13 Feb 2013 09:31:53 +0000
Subject: drm/i915: Disable WC PTE updates to w/a buggy IOMMU on ILK

Whilst IOMMU is enabled for the Intel GPU on Ironlake, it appears that
using WC writes to update the PTE on the GPU fails miserably. The
result looks like the majority of the writes do not land leading to
lots of screen corruption and a hard system hang.

v2: s/</<=/ to preserve the current exclusion of Sandybridge

Reported-by: Nathan Myers <ncm@cantrip.org>
Bugzilla: https://bugzilla.freedesktop.org/show_bug.cgi?id=60391
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Tested-by: Nathan Myers <ncm@cantrip.org>
[danvet: Remove cc: stable and add tested-by.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/char/agp/intel-gtt.c b/drivers/char/agp/intel-gtt.c
index 207e5c3..b8e2014 100644
--- a/drivers/char/agp/intel-gtt.c
+++ b/drivers/char/agp/intel-gtt.c
@@ -572,6 +572,40 @@ static void intel_gtt_cleanup(void)
 	intel_gtt_teardown_scratch_page();
 }
 
+/* Certain Gen5 chipsets require require idling the GPU before
+ * unmapping anything from the GTT when VT-d is enabled.
+ */
+static inline int needs_ilk_vtd_wa(void)
+{
+#ifdef CONFIG_INTEL_IOMMU
+	const unsigned short gpu_devid = intel_private.pcidev->device;
+
+	/* Query intel_iommu to see if we need the workaround. Presumably that
+	 * was loaded first.
+	 */
+	if ((gpu_devid == PCI_DEVICE_ID_INTEL_IRONLAKE_M_HB ||
+	     gpu_devid == PCI_DEVICE_ID_INTEL_IRONLAKE_M_IG) &&
+	     intel_iommu_gfx_mapped)
+		return 1;
+#endif
+	return 0;
+}
+
+static bool intel_gtt_can_wc(void)
+{
+	if (INTEL_GTT_GEN <= 2)
+		return false;
+
+	if (INTEL_GTT_GEN >= 6)
+		return false;
+
+	/* Reports of major corruption with ILK vt'd enabled */
+	if (needs_ilk_vtd_wa())
+		return false;
+
+	return true;
+}
+
 static int intel_gtt_init(void)
 {
 	u32 gma_addr;
@@ -601,7 +635,7 @@ static int intel_gtt_init(void)
 	gtt_map_size = intel_private.gtt_total_entries * 4;
 
 	intel_private.gtt = NULL;
-	if (INTEL_GTT_GEN < 6 && INTEL_GTT_GEN > 2)
+	if (intel_gtt_can_wc())
 		intel_private.gtt = ioremap_wc(intel_private.gtt_bus_addr,
 					       gtt_map_size);
 	if (intel_private.gtt == NULL)
@@ -1072,7 +1106,6 @@ static void i965_write_entry(dma_addr_t addr,
 	writel(addr | pte_flags, intel_private.gtt + entry);
 }
 
-
 static int i9xx_setup(void)
 {
 	u32 reg_addr, gtt_addr;
-- 
cgit v0.10.2


From 210561ffd72d00eccf12c0131b8024d5436bae95 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Mon, 21 Jan 2013 19:48:59 +0100
Subject: intel/iommu: force writebuffer-flush quirk on Gen 4 Chipsets

We already have the quirk entry for the mobile platform, but also
reports on some desktop versions. So be paranoid and set it
everywhere.

References: http://www.mail-archive.com/dri-devel@lists.freedesktop.org/msg33138.html
Cc: stable@vger.kernel.org
Cc: David Woodhouse <dwmw2@infradead.org>
Cc: "Sankaran, Rajesh" <rajesh.sankaran@intel.com>
Reported-and-tested-by: Mihai Moldovan <ionic@ionic.de>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/iommu/intel-iommu.c b/drivers/iommu/intel-iommu.c
index eca2801..f1e7b86 100644
--- a/drivers/iommu/intel-iommu.c
+++ b/drivers/iommu/intel-iommu.c
@@ -4253,13 +4253,19 @@ static void quirk_iommu_rwbf(struct pci_dev *dev)
 {
 	/*
 	 * Mobile 4 Series Chipset neglects to set RWBF capability,
-	 * but needs it:
+	 * but needs it. Same seems to hold for the desktop versions.
 	 */
 	printk(KERN_INFO "DMAR: Forcing write-buffer flush capability\n");
 	rwbf_quirk = 1;
 }
 
 DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, 0x2a40, quirk_iommu_rwbf);
+DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, 0x2e00, quirk_iommu_rwbf);
+DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, 0x2e10, quirk_iommu_rwbf);
+DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, 0x2e20, quirk_iommu_rwbf);
+DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, 0x2e30, quirk_iommu_rwbf);
+DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, 0x2e40, quirk_iommu_rwbf);
+DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, 0x2e90, quirk_iommu_rwbf);
 
 #define GGC 0x52
 #define GGC_MEMORY_SIZE_MASK	(0xf << 8)
-- 
cgit v0.10.2


From bc79202f4c8ff6977133255dc2bb7d4dac3936f0 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Tue, 4 Dec 2012 09:50:33 +1000
Subject: nvd0/therm: implement more appropriate pwm fan control functions

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index ab25752..4fc39ec 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -109,6 +109,7 @@ nouveau-y += core/subdev/therm/fan.o
 nouveau-y += core/subdev/therm/ic.o
 nouveau-y += core/subdev/therm/nv40.o
 nouveau-y += core/subdev/therm/nv50.o
+nouveau-y += core/subdev/therm/nvd0.o
 nouveau-y += core/subdev/therm/temp.o
 nouveau-y += core/subdev/timer/base.o
 nouveau-y += core/subdev/timer/nv04.o
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/therm.h b/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
index faee569..6cf9f93 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
@@ -54,5 +54,6 @@ nouveau_therm(void *obj)
 
 extern struct nouveau_oclass nv40_therm_oclass;
 extern struct nouveau_oclass nv50_therm_oclass;
+extern struct nouveau_oclass nvd0_therm_oclass;
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/device/nvc0.c
index f046168..c189a12 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/device/nvc0.c
@@ -255,7 +255,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nvd0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nve0.c b/drivers/gpu/drm/nouveau/core/subdev/device/nve0.c
index 03a6528..7bef3b1 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/device/nve0.c
@@ -59,7 +59,7 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nvd0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
@@ -87,7 +87,7 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nvd0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
@@ -115,7 +115,7 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nvd0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
new file mode 100644
index 0000000..a87338c
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
@@ -0,0 +1,116 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+static int
+pwm_info(struct nouveau_therm *therm, int line)
+{
+	u32 gpio = nv_rd32(therm, 0x00d610 + (line * 0x04));
+	if (gpio & 0x00000040) {
+		switch (gpio & 0x0000001f) {
+		case 0x19: return 1;
+		case 0x1c: return 0;
+		default:
+			break;
+		}
+	}
+
+	nv_error(therm, "GPIO %d unknown PWM: 0x%08x\n", line, gpio);
+	return -EINVAL;
+}
+
+static int
+nvd0_fan_pwm_get(struct nouveau_therm *therm, int line, u32 *divs, u32 *duty)
+{
+	int indx = pwm_info(therm, line);
+	if (indx < 0)
+		return indx;
+
+	*divs = nv_rd32(therm, 0x00e114 + (indx * 8));
+	*duty = nv_rd32(therm, 0x00e118 + (indx * 8));
+	return 0;
+}
+
+static int
+nvd0_fan_pwm_set(struct nouveau_therm *therm, int line, u32 divs, u32 duty)
+{
+	int indx = pwm_info(therm, line);
+	if (indx < 0)
+		return indx;
+
+	nv_wr32(therm, 0x00e114 + (indx * 8), divs);
+	nv_wr32(therm, 0x00e118 + (indx * 8), duty | 0x80000000);
+	return 0;
+}
+
+static int
+nvd0_fan_pwm_clock(struct nouveau_therm *therm)
+{
+	return (nv_device(therm)->crystal * 1000) / 20;
+}
+
+static int
+nvd0_therm_ctor(struct nouveau_object *parent,
+		   struct nouveau_object *engine,
+		   struct nouveau_oclass *oclass, void *data, u32 size,
+		   struct nouveau_object **pobject)
+{
+	struct nouveau_therm_priv *priv;
+	struct nouveau_therm *therm;
+	int ret;
+
+	ret = nouveau_therm_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
+	therm = (void *) priv;
+	if (ret)
+		return ret;
+
+	nouveau_therm_ic_ctor(therm);
+	nouveau_therm_sensor_ctor(therm);
+	nouveau_therm_fan_ctor(therm);
+
+	priv->fan.pwm_get = nvd0_fan_pwm_get;
+	priv->fan.pwm_set = nvd0_fan_pwm_set;
+	priv->fan.pwm_clock = nvd0_fan_pwm_clock;
+
+	therm->temp_get = nv50_temp_get;
+	therm->fan_get = nouveau_therm_fan_user_get;
+	therm->fan_set = nouveau_therm_fan_user_set;
+	therm->fan_sense = nouveau_therm_fan_sense;
+	therm->attr_get = nouveau_therm_attr_get;
+	therm->attr_set = nouveau_therm_attr_set;
+	return 0;
+}
+
+struct nouveau_oclass
+nvd0_therm_oclass = {
+	.handle = NV_SUBDEV(THERM, 0xd0),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvd0_therm_ctor,
+		.dtor = _nouveau_therm_dtor,
+		.init = nouveau_therm_init,
+		.fini = nouveau_therm_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
index 1c3cd6a..ac26ef5 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
@@ -1,3 +1,6 @@
+#ifndef __NVTHERM_PRIV_H__
+#define __NVTHERM_PRIV_H__
+
 /*
  * Copyright 2012 The Nouveau community
  *
@@ -71,3 +74,7 @@ int nouveau_therm_fan_set_mode(struct nouveau_therm *therm,
 
 
 int nouveau_therm_fan_sense(struct nouveau_therm *therm);
+
+int nv50_temp_get(struct nouveau_therm *therm);
+
+#endif
-- 
cgit v0.10.2


From 5f066c32a50eac584c0c8c19d8d5e38714f7c574 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Tue, 4 Dec 2012 11:35:40 +1000
Subject: drm/nouveau/therm: fix various style issues, make more consistent

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Martin Peres <martin.peres@labri.fr>

diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index 4fc39ec..26e0d56 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -107,10 +107,10 @@ nouveau-y += core/subdev/mxm/nv50.o
 nouveau-y += core/subdev/therm/base.o
 nouveau-y += core/subdev/therm/fan.o
 nouveau-y += core/subdev/therm/ic.o
+nouveau-y += core/subdev/therm/temp.o
 nouveau-y += core/subdev/therm/nv40.o
 nouveau-y += core/subdev/therm/nv50.o
 nouveau-y += core/subdev/therm/nvd0.o
-nouveau-y += core/subdev/therm/temp.o
 nouveau-y += core/subdev/timer/base.o
 nouveau-y += core/subdev/timer/nv04.o
 nouveau-y += core/subdev/vm/base.o
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/therm.h b/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
index 6cf9f93..2a48bff 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
@@ -46,11 +46,24 @@ nouveau_therm(void *obj)
 }
 
 #define nouveau_therm_create(p,e,o,d)                                          \
-	nouveau_subdev_create((p), (e), (o), 0, "THERM", "therm", d)
+	nouveau_therm_create_((p), (e), (o), sizeof(**d), (void **)d)
 #define nouveau_therm_destroy(p)                                               \
 	nouveau_subdev_destroy(&(p)->base)
+#define nouveau_therm_init(p) ({                                               \
+	struct nouveau_therm *therm = (p);                                     \
+        _nouveau_therm_init(nv_object(therm));                                 \
+})
+#define nouveau_therm_fini(p,s) ({                                             \
+	struct nouveau_therm *therm = (p);                                     \
+        _nouveau_therm_init(nv_object(therm), (s));                            \
+})
+
+int nouveau_therm_create_(struct nouveau_object *, struct nouveau_object *,
+			  struct nouveau_oclass *, int, void **);
 
 #define _nouveau_therm_dtor _nouveau_subdev_dtor
+int _nouveau_therm_init(struct nouveau_object *);
+int _nouveau_therm_fini(struct nouveau_object *, bool);
 
 extern struct nouveau_oclass nv40_therm_oclass;
 extern struct nouveau_oclass nv50_therm_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/base.c b/drivers/gpu/drm/nouveau/core/subdev/therm/base.c
index 1674c74..6260eac 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/base.c
@@ -116,7 +116,7 @@ nouveau_therm_attr_set(struct nouveau_therm *therm,
 }
 
 int
-nouveau_therm_init(struct nouveau_object *object)
+_nouveau_therm_init(struct nouveau_object *object)
 {
 	struct nouveau_therm *therm = (void *)object;
 	struct nouveau_therm_priv *priv = (void *)therm;
@@ -133,7 +133,7 @@ nouveau_therm_init(struct nouveau_object *object)
 }
 
 int
-nouveau_therm_fini(struct nouveau_object *object, bool suspend)
+_nouveau_therm_fini(struct nouveau_object *object, bool suspend)
 {
 	struct nouveau_therm *therm = (void *)object;
 	struct nouveau_therm_priv *priv = (void *)therm;
@@ -142,3 +142,30 @@ nouveau_therm_fini(struct nouveau_object *object, bool suspend)
 
 	return nouveau_subdev_fini(&therm->base, suspend);
 }
+
+int
+nouveau_therm_create_(struct nouveau_object *parent,
+		      struct nouveau_object *engine,
+		      struct nouveau_oclass *oclass,
+		      int length, void **pobject)
+{
+	struct nouveau_therm_priv *priv;
+	int ret;
+
+	ret = nouveau_subdev_create_(parent, engine, oclass, 0, "PTHERM",
+				     "therm", length, pobject);
+	priv = *pobject;
+	if (ret)
+		return ret;
+
+	nouveau_therm_ic_ctor(&priv->base);
+	nouveau_therm_sensor_ctor(&priv->base);
+	nouveau_therm_fan_ctor(&priv->base);
+
+	priv->base.fan_get = nouveau_therm_fan_user_get;
+	priv->base.fan_set = nouveau_therm_fan_user_set;
+	priv->base.fan_sense = nouveau_therm_fan_sense;
+	priv->base.attr_get = nouveau_therm_attr_get;
+	priv->base.attr_set = nouveau_therm_attr_set;
+	return 0;
+}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c
index fcf2cfe..f2092af 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c
@@ -25,6 +25,10 @@
 
 #include "priv.h"
 
+struct nv40_therm_priv {
+	struct nouveau_therm_priv base;
+};
+
 static int
 nv40_sensor_setup(struct nouveau_therm *therm)
 {
@@ -120,34 +124,21 @@ nv40_fan_pwm_set(struct nouveau_therm *therm, int line, u32 divs, u32 duty)
 
 static int
 nv40_therm_ctor(struct nouveau_object *parent,
-		   struct nouveau_object *engine,
-		   struct nouveau_oclass *oclass, void *data, u32 size,
-		   struct nouveau_object **pobject)
+		struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
 {
-	struct nouveau_therm_priv *priv;
-	struct nouveau_therm *therm;
+	struct nv40_therm_priv *priv;
 	int ret;
 
 	ret = nouveau_therm_create(parent, engine, oclass, &priv);
 	*pobject = nv_object(priv);
-	therm = (void *) priv;
 	if (ret)
 		return ret;
 
-	nouveau_therm_ic_ctor(therm);
-	nouveau_therm_sensor_ctor(therm);
-	nouveau_therm_fan_ctor(therm);
-
-	priv->fan.pwm_get = nv40_fan_pwm_get;
-	priv->fan.pwm_set = nv40_fan_pwm_set;
-
-	therm->temp_get = nv40_temp_get;
-	therm->fan_get = nouveau_therm_fan_user_get;
-	therm->fan_set = nouveau_therm_fan_user_set;
-	therm->fan_sense = nouveau_therm_fan_sense;
-	therm->attr_get = nouveau_therm_attr_get;
-	therm->attr_set = nouveau_therm_attr_set;
-
+	priv->base.fan.pwm_get = nv40_fan_pwm_get;
+	priv->base.fan.pwm_set = nv40_fan_pwm_set;
+	priv->base.base.temp_get = nv40_temp_get;
 	return 0;
 }
 
@@ -157,7 +148,7 @@ nv40_therm_oclass = {
 	.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = nv40_therm_ctor,
 		.dtor = _nouveau_therm_dtor,
-		.init = nouveau_therm_init,
-		.fini = nouveau_therm_fini,
+		.init = _nouveau_therm_init,
+		.fini = _nouveau_therm_fini,
 	},
-};
\ No newline at end of file
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nv50.c
index 9360ddd..b7e1ecf 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nv50.c
@@ -25,6 +25,10 @@
 
 #include "priv.h"
 
+struct nv50_therm_priv {
+	struct nouveau_therm_priv base;
+};
+
 static int
 pwm_info(struct nouveau_therm *therm, int *line, int *ctrl, int *indx)
 {
@@ -113,35 +117,22 @@ nv50_temp_get(struct nouveau_therm *therm)
 
 static int
 nv50_therm_ctor(struct nouveau_object *parent,
-		   struct nouveau_object *engine,
-		   struct nouveau_oclass *oclass, void *data, u32 size,
-		   struct nouveau_object **pobject)
+		struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
 {
-	struct nouveau_therm_priv *priv;
-	struct nouveau_therm *therm;
+	struct nv50_therm_priv *priv;
 	int ret;
 
 	ret = nouveau_therm_create(parent, engine, oclass, &priv);
 	*pobject = nv_object(priv);
-	therm = (void *) priv;
 	if (ret)
 		return ret;
 
-	nouveau_therm_ic_ctor(therm);
-	nouveau_therm_sensor_ctor(therm);
-	nouveau_therm_fan_ctor(therm);
-
-	priv->fan.pwm_get = nv50_fan_pwm_get;
-	priv->fan.pwm_set = nv50_fan_pwm_set;
-	priv->fan.pwm_clock = nv50_fan_pwm_clock;
-
-	therm->temp_get = nv50_temp_get;
-	therm->fan_get = nouveau_therm_fan_user_get;
-	therm->fan_set = nouveau_therm_fan_user_set;
-	therm->fan_sense = nouveau_therm_fan_sense;
-	therm->attr_get = nouveau_therm_attr_get;
-	therm->attr_set = nouveau_therm_attr_set;
-
+	priv->base.fan.pwm_get = nv50_fan_pwm_get;
+	priv->base.fan.pwm_set = nv50_fan_pwm_set;
+	priv->base.fan.pwm_clock = nv50_fan_pwm_clock;
+	priv->base.base.temp_get = nv50_temp_get;
 	return 0;
 }
 
@@ -151,7 +142,7 @@ nv50_therm_oclass = {
 	.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = nv50_therm_ctor,
 		.dtor = _nouveau_therm_dtor,
-		.init = nouveau_therm_init,
-		.fini = nouveau_therm_fini,
+		.init = _nouveau_therm_init,
+		.fini = _nouveau_therm_fini,
 	},
 };
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
index a87338c..b001cd3 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
@@ -24,6 +24,10 @@
 
 #include "priv.h"
 
+struct nvd0_therm_priv {
+	struct nouveau_therm_priv base;
+};
+
 static int
 pwm_info(struct nouveau_therm *therm, int line)
 {
@@ -73,34 +77,22 @@ nvd0_fan_pwm_clock(struct nouveau_therm *therm)
 
 static int
 nvd0_therm_ctor(struct nouveau_object *parent,
-		   struct nouveau_object *engine,
-		   struct nouveau_oclass *oclass, void *data, u32 size,
-		   struct nouveau_object **pobject)
+		struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
 {
-	struct nouveau_therm_priv *priv;
-	struct nouveau_therm *therm;
+	struct nvd0_therm_priv *priv;
 	int ret;
 
 	ret = nouveau_therm_create(parent, engine, oclass, &priv);
 	*pobject = nv_object(priv);
-	therm = (void *) priv;
 	if (ret)
 		return ret;
 
-	nouveau_therm_ic_ctor(therm);
-	nouveau_therm_sensor_ctor(therm);
-	nouveau_therm_fan_ctor(therm);
-
-	priv->fan.pwm_get = nvd0_fan_pwm_get;
-	priv->fan.pwm_set = nvd0_fan_pwm_set;
-	priv->fan.pwm_clock = nvd0_fan_pwm_clock;
-
-	therm->temp_get = nv50_temp_get;
-	therm->fan_get = nouveau_therm_fan_user_get;
-	therm->fan_set = nouveau_therm_fan_user_set;
-	therm->fan_sense = nouveau_therm_fan_sense;
-	therm->attr_get = nouveau_therm_attr_get;
-	therm->attr_set = nouveau_therm_attr_set;
+	priv->base.fan.pwm_get = nvd0_fan_pwm_get;
+	priv->base.fan.pwm_set = nvd0_fan_pwm_set;
+	priv->base.fan.pwm_clock = nvd0_fan_pwm_clock;
+	priv->base.base.temp_get = nv50_temp_get;
 	return 0;
 }
 
@@ -110,7 +102,7 @@ nvd0_therm_oclass = {
 	.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = nvd0_therm_ctor,
 		.dtor = _nouveau_therm_dtor,
-		.init = nouveau_therm_init,
-		.fini = nouveau_therm_fini,
+		.init = _nouveau_therm_init,
+		.fini = _nouveau_therm_fini,
 	},
 };
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
index ac26ef5..64f4a4a 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
@@ -53,8 +53,6 @@ struct nouveau_therm_priv {
 	struct i2c_client *ic;
 };
 
-int nouveau_therm_init(struct nouveau_object *object);
-int nouveau_therm_fini(struct nouveau_object *object, bool suspend);
 int nouveau_therm_attr_get(struct nouveau_therm *therm,
 		       enum nouveau_therm_attr_type type);
 int nouveau_therm_attr_set(struct nouveau_therm *therm,
-- 
cgit v0.10.2


From d639b4f5ba5f1bbdfbba920e1deb035a1bed83ae Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Tue, 4 Dec 2012 13:25:26 +1000
Subject: drm/nouveau/therm: collect fan tach info in common fan constructor

This info will be used by two more implementations in upcoming commits.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Martin Peres <martin.peres@labri.fr>

diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
index 5231786..4822733 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
@@ -27,6 +27,7 @@
 
 #include <core/object.h>
 #include <core/device.h>
+
 #include <subdev/gpio.h>
 #include <subdev/timer.h>
 
@@ -104,13 +105,13 @@ nouveau_therm_fan_set(struct nouveau_therm *therm, int percent)
 int
 nouveau_therm_fan_sense(struct nouveau_therm *therm)
 {
+	struct nouveau_therm_priv *priv = (void *)therm;
 	struct nouveau_timer *ptimer = nouveau_timer(therm);
 	struct nouveau_gpio *gpio = nouveau_gpio(therm);
-	struct dcb_gpio_func func;
 	u32 cycles, cur, prev;
 	u64 start, end, tach;
 
-	if (gpio->find(gpio, 0, DCB_GPIO_FAN_SENSE, 0xff, &func))
+	if (priv->fan.tach.func == DCB_GPIO_UNUSED)
 		return -ENODEV;
 
 	/* Time a complete rotation and extrapolate to RPM:
@@ -118,12 +119,12 @@ nouveau_therm_fan_sense(struct nouveau_therm *therm)
 	 * We get 4 changes (0 -> 1 -> 0 -> 1) per complete rotation.
 	 */
 	start = ptimer->read(ptimer);
-	prev = gpio->get(gpio, 0, func.func, func.line);
+	prev = gpio->get(gpio, 0, priv->fan.tach.func, priv->fan.tach.line);
 	cycles = 0;
 	do {
 		usleep_range(500, 1000); /* supports 0 < rpm < 7500 */
 
-		cur = gpio->get(gpio, 0, func.func, func.line);
+		cur = gpio->get(gpio, 0, priv->fan.tach.func, priv->fan.tach.line);
 		if (prev != cur) {
 			if (!start)
 				start = ptimer->read(ptimer);
@@ -220,7 +221,13 @@ int
 nouveau_therm_fan_ctor(struct nouveau_therm *therm)
 {
 	struct nouveau_therm_priv *priv = (void *)therm;
+	struct nouveau_gpio *gpio = nouveau_gpio(therm);
 	struct nouveau_bios *bios = nouveau_bios(therm);
+	int ret;
+
+	ret = gpio->find(gpio, 0, DCB_GPIO_FAN_SENSE, 0xff, &priv->fan.tach);
+	if (ret)
+		priv->fan.tach.func = DCB_GPIO_UNUSED;
 
 	nouveau_therm_fan_set_defaults(therm);
 	nvbios_perf_fan_parse(bios, &priv->bios_perf_fan);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
index 64f4a4a..33fbc59 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
@@ -28,6 +28,7 @@
 #include <subdev/therm.h>
 
 #include <subdev/bios/extdev.h>
+#include <subdev/bios/gpio.h>
 #include <subdev/bios/perf.h>
 #include <subdev/bios/therm.h>
 
@@ -44,6 +45,8 @@ struct nouveau_therm_priv {
 		enum nouveau_therm_fan_mode mode;
 		int percent;
 
+		struct dcb_gpio_func tach;
+
 		int (*pwm_get)(struct nouveau_therm *, int line, u32*, u32*);
 		int (*pwm_set)(struct nouveau_therm *, int line, u32, u32);
 		int (*pwm_clock)(struct nouveau_therm *);
@@ -70,9 +73,11 @@ int nouveau_therm_fan_user_set(struct nouveau_therm *therm, int percent);
 int nouveau_therm_fan_set_mode(struct nouveau_therm *therm,
 			   enum nouveau_therm_fan_mode mode);
 
-
 int nouveau_therm_fan_sense(struct nouveau_therm *therm);
 
+int nv50_fan_pwm_get(struct nouveau_therm *, int, u32 *, u32 *);
+int nv50_fan_pwm_set(struct nouveau_therm *, int, u32, u32);
+int nv50_fan_pwm_clock(struct nouveau_therm *);
 int nv50_temp_get(struct nouveau_therm *therm);
 
 #endif
-- 
cgit v0.10.2


From 7b49bd684cea3916405e1de2f8173ac42cebc87b Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Tue, 4 Dec 2012 12:18:59 +1000
Subject: drm/nva3/therm: add support for hardware fan tachometer

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index 26e0d56..7938d1a 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -110,6 +110,7 @@ nouveau-y += core/subdev/therm/ic.o
 nouveau-y += core/subdev/therm/temp.o
 nouveau-y += core/subdev/therm/nv40.o
 nouveau-y += core/subdev/therm/nv50.o
+nouveau-y += core/subdev/therm/nva3.o
 nouveau-y += core/subdev/therm/nvd0.o
 nouveau-y += core/subdev/timer/base.o
 nouveau-y += core/subdev/timer/nv04.o
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/therm.h b/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
index 2a48bff..a96d43a 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
@@ -67,6 +67,7 @@ int _nouveau_therm_fini(struct nouveau_object *, bool);
 
 extern struct nouveau_oclass nv40_therm_oclass;
 extern struct nouveau_oclass nv50_therm_oclass;
+extern struct nouveau_oclass nva3_therm_oclass;
 extern struct nouveau_oclass nvd0_therm_oclass;
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/device/nv50.c
index 6ccfd85..940c401 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/device/nv50.c
@@ -306,7 +306,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nva3_clock_oclass;
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv98_mc_oclass;
@@ -332,7 +332,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nva3_clock_oclass;
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv98_mc_oclass;
@@ -357,7 +357,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nva3_clock_oclass;
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv98_mc_oclass;
@@ -382,7 +382,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nva3_clock_oclass;
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv98_mc_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/device/nvc0.c
index c189a12..005d0b6 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/device/nvc0.c
@@ -59,7 +59,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
@@ -87,7 +87,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
@@ -115,7 +115,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
@@ -143,7 +143,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
@@ -171,7 +171,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
@@ -199,7 +199,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
@@ -227,7 +227,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
@@ -282,4 +282,4 @@ nvc0_identify(struct nouveau_device *device)
 	}
 
 	return 0;
-}
+	}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c
new file mode 100644
index 0000000..f473808
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c
@@ -0,0 +1,97 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <subdev/gpio.h>
+
+#include "priv.h"
+
+struct nva3_therm_priv {
+	struct nouveau_therm_priv base;
+};
+
+static int
+nva3_therm_fan_sense(struct nouveau_therm *therm)
+{
+	u32 tach = nv_rd32(therm, 0x00e728) & 0x0000ffff;
+	u32 ctrl = nv_rd32(therm, 0x00e720);
+	if (ctrl & 0x00000001)
+		return tach * 60;
+	return -ENODEV;
+}
+
+static int
+nva3_therm_init(struct nouveau_object *object)
+{
+	struct nva3_therm_priv *priv = (void *)object;
+	struct dcb_gpio_func *tach = &priv->base.fan.tach;
+	int ret;
+
+	ret = nouveau_therm_init(&priv->base.base);
+	if (ret)
+		return ret;
+
+	/* enable fan tach, count revolutions per-second */
+	nv_mask(priv, 0x00e720, 0x00000003, 0x00000002);
+	if (tach->func != DCB_GPIO_UNUSED) {
+		nv_wr32(priv, 0x00e724, nv_device(priv)->crystal * 1000);
+		nv_mask(priv, 0x00e720, 0x001f0000, tach->line << 16);
+		nv_mask(priv, 0x00e720, 0x00000001, 0x00000001);
+	}
+	nv_mask(priv, 0x00e720, 0x00000002, 0x00000000);
+
+	return 0;
+}
+
+static int
+nva3_therm_ctor(struct nouveau_object *parent,
+		struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nva3_therm_priv *priv;
+	int ret;
+
+	ret = nouveau_therm_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	priv->base.fan.pwm_get = nv50_fan_pwm_get;
+	priv->base.fan.pwm_set = nv50_fan_pwm_set;
+	priv->base.fan.pwm_clock = nv50_fan_pwm_clock;
+	priv->base.base.temp_get = nv50_temp_get;
+	priv->base.base.fan_sense = nva3_therm_fan_sense;
+	return 0;
+}
+
+struct nouveau_oclass
+nva3_therm_oclass = {
+	.handle = NV_SUBDEV(THERM, 0xa3),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nva3_therm_ctor,
+		.dtor = _nouveau_therm_dtor,
+		.init = nva3_therm_init,
+		.fini = _nouveau_therm_fini,
+	},
+};
-- 
cgit v0.10.2


From 9cbcd3374a1db6a2faeb660bb49feb9af25c84ec Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Tue, 4 Dec 2012 12:10:19 +1000
Subject: drm/nvd0/therm: add support for hardware fan tachometer

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c
index f473808..9a8daf4 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c
@@ -30,7 +30,7 @@ struct nva3_therm_priv {
 	struct nouveau_therm_priv base;
 };
 
-static int
+int
 nva3_therm_fan_sense(struct nouveau_therm *therm)
 {
 	u32 tach = nv_rd32(therm, 0x00e728) & 0x0000ffff;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
index b001cd3..f0336a0 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
@@ -76,6 +76,28 @@ nvd0_fan_pwm_clock(struct nouveau_therm *therm)
 }
 
 static int
+nvd0_therm_init(struct nouveau_object *object)
+{
+	struct nvd0_therm_priv *priv = (void *)object;
+	int ret;
+
+	ret = nouveau_therm_init(&priv->base.base);
+	if (ret)
+		return ret;
+
+	/* enable fan tach, count revolutions per-second */
+	nv_mask(priv, 0x00e720, 0x00000003, 0x00000002);
+	if (priv->base.fan.tach.func != DCB_GPIO_UNUSED) {
+		nv_mask(priv, 0x00d79c, 0x000000ff, priv->base.fan.tach.line);
+		nv_wr32(priv, 0x00e724, nv_device(priv)->crystal * 1000);
+		nv_mask(priv, 0x00e720, 0x00000001, 0x00000001);
+	}
+	nv_mask(priv, 0x00e720, 0x00000002, 0x00000000);
+
+	return 0;
+}
+
+static int
 nvd0_therm_ctor(struct nouveau_object *parent,
 		struct nouveau_object *engine,
 		struct nouveau_oclass *oclass, void *data, u32 size,
@@ -93,6 +115,7 @@ nvd0_therm_ctor(struct nouveau_object *parent,
 	priv->base.fan.pwm_set = nvd0_fan_pwm_set;
 	priv->base.fan.pwm_clock = nvd0_fan_pwm_clock;
 	priv->base.base.temp_get = nv50_temp_get;
+	priv->base.base.fan_sense = nva3_therm_fan_sense;
 	return 0;
 }
 
@@ -102,7 +125,7 @@ nvd0_therm_oclass = {
 	.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = nvd0_therm_ctor,
 		.dtor = _nouveau_therm_dtor,
-		.init = _nouveau_therm_init,
+		.init = nvd0_therm_init,
 		.fini = _nouveau_therm_fini,
 	},
 };
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
index 33fbc59..bb940e9 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
@@ -80,4 +80,6 @@ int nv50_fan_pwm_set(struct nouveau_therm *, int, u32, u32);
 int nv50_fan_pwm_clock(struct nouveau_therm *);
 int nv50_temp_get(struct nouveau_therm *therm);
 
+int nva3_therm_fan_sense(struct nouveau_therm *);
+
 #endif
-- 
cgit v0.10.2


From 68197b4ba0a3c90190795b97248b840a3e506545 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Wed, 5 Dec 2012 14:56:37 +1000
Subject: drm/nouveau/therm: add interfaces to allow forcing off pwm fan
 control

Mostly to allow for the possibility of testing 'toggle' fan control easily.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Martin Peres <martin.peres@labri.fr>

diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
index 4822733..b4d1449 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
@@ -96,6 +96,8 @@ nouveau_therm_fan_set(struct nouveau_therm *therm, int percent)
 			duty = divs - duty;
 
 		ret = priv->fan.pwm_set(therm, func.line, divs, duty);
+		if (ret == 0)
+			ret = priv->fan.pwm_ctrl(therm, func.line, true);
 		return ret;
 	}
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c
index f2092af..47a15ec 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c
@@ -80,6 +80,19 @@ nv40_temp_get(struct nouveau_therm *therm)
 }
 
 int
+nv40_fan_pwm_ctrl(struct nouveau_therm *therm, int line, bool enable)
+{
+	u32 mask = enable ? 0x80000000 : 0x0000000;
+	if      (line == 2) nv_mask(therm, 0x0010f0, 0x80000000, mask);
+	else if (line == 9) nv_mask(therm, 0x0015f4, 0x80000000, mask);
+	else {
+		nv_error(therm, "unknown pwm ctrl for gpio %d\n", line);
+		return -ENODEV;
+	}
+	return 0;
+}
+
+int
 nv40_fan_pwm_get(struct nouveau_therm *therm, int line, u32 *divs, u32 *duty)
 {
 	if (line == 2) {
@@ -109,11 +122,11 @@ int
 nv40_fan_pwm_set(struct nouveau_therm *therm, int line, u32 divs, u32 duty)
 {
 	if (line == 2) {
-		nv_wr32(therm, 0x0010f0, 0x80000000 | (duty << 16) | divs);
+		nv_mask(therm, 0x0010f0, 0x7fff7fff, (duty << 16) | divs);
 	} else
 	if (line == 9) {
 		nv_wr32(therm, 0x0015f8, divs);
-		nv_wr32(therm, 0x0015f4, duty | 0x80000000);
+		nv_mask(therm, 0x0015f4, 0x7fffffff, duty);
 	} else {
 		nv_error(therm, "unknown pwm ctrl for gpio %d\n", line);
 		return -ENODEV;
@@ -136,6 +149,7 @@ nv40_therm_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
+	priv->base.fan.pwm_ctrl = nv40_fan_pwm_ctrl;
 	priv->base.fan.pwm_get = nv40_fan_pwm_get;
 	priv->base.fan.pwm_set = nv40_fan_pwm_set;
 	priv->base.base.temp_get = nv40_temp_get;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nv50.c
index b7e1ecf..6ebad82 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nv50.c
@@ -55,6 +55,16 @@ pwm_info(struct nouveau_therm *therm, int *line, int *ctrl, int *indx)
 }
 
 int
+nv50_fan_pwm_ctrl(struct nouveau_therm *therm, int line, bool enable)
+{
+	u32 data = enable ? 0x00000001 : 0x00000000;
+	int ctrl, id, ret = pwm_info(therm, &line, &ctrl, &id);
+	if (ret == 0)
+		nv_mask(therm, ctrl, 0x00010001 << line, data << line);
+	return ret;
+}
+
+int
 nv50_fan_pwm_get(struct nouveau_therm *therm, int line, u32 *divs, u32 *duty)
 {
 	int ctrl, id, ret = pwm_info(therm, &line, &ctrl, &id);
@@ -77,7 +87,6 @@ nv50_fan_pwm_set(struct nouveau_therm *therm, int line, u32 divs, u32 duty)
 	if (ret)
 		return ret;
 
-	nv_mask(therm, ctrl, 0x00010001 << line, 0x00000001 << line);
 	nv_wr32(therm, 0x00e114 + (id * 8), divs);
 	nv_wr32(therm, 0x00e118 + (id * 8), duty | 0x80000000);
 	return 0;
@@ -129,6 +138,7 @@ nv50_therm_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
+	priv->base.fan.pwm_ctrl = nv50_fan_pwm_ctrl;
 	priv->base.fan.pwm_get = nv50_fan_pwm_get;
 	priv->base.fan.pwm_set = nv50_fan_pwm_set;
 	priv->base.fan.pwm_clock = nv50_fan_pwm_clock;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
index f0336a0..ec06e89 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
@@ -32,32 +32,52 @@ static int
 pwm_info(struct nouveau_therm *therm, int line)
 {
 	u32 gpio = nv_rd32(therm, 0x00d610 + (line * 0x04));
-	if (gpio & 0x00000040) {
+	switch (gpio & 0x000000c0) {
+	case 0x00000000: /* normal mode, possibly pwm forced off by us */
+	case 0x00000040: /* nvio special */
 		switch (gpio & 0x0000001f) {
 		case 0x19: return 1;
 		case 0x1c: return 0;
 		default:
 			break;
 		}
+	default:
+		break;
 	}
 
 	nv_error(therm, "GPIO %d unknown PWM: 0x%08x\n", line, gpio);
-	return -EINVAL;
+	return -ENODEV;
 }
 
 static int
-nvd0_fan_pwm_get(struct nouveau_therm *therm, int line, u32 *divs, u32 *duty)
+nvd0_fan_pwm_ctrl(struct nouveau_therm *therm, int line, bool enable)
 {
+	u32 data = enable ? 0x00000040 : 0x00000000;
 	int indx = pwm_info(therm, line);
 	if (indx < 0)
 		return indx;
 
-	*divs = nv_rd32(therm, 0x00e114 + (indx * 8));
-	*duty = nv_rd32(therm, 0x00e118 + (indx * 8));
+	nv_mask(therm, 0x00d610 + (line * 0x04), 0x000000c0, data);
 	return 0;
 }
 
 static int
+nvd0_fan_pwm_get(struct nouveau_therm *therm, int line, u32 *divs, u32 *duty)
+{
+	int indx = pwm_info(therm, line);
+	if (indx < 0)
+		return indx;
+
+	if (nv_rd32(therm, 0x00d610 + (line * 0x04)) & 0x00000040) {
+		*divs = nv_rd32(therm, 0x00e114 + (indx * 8));
+		*duty = nv_rd32(therm, 0x00e118 + (indx * 8));
+		return 0;
+	}
+
+	return -EINVAL;
+}
+
+static int
 nvd0_fan_pwm_set(struct nouveau_therm *therm, int line, u32 divs, u32 duty)
 {
 	int indx = pwm_info(therm, line);
@@ -111,6 +131,7 @@ nvd0_therm_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
+	priv->base.fan.pwm_ctrl = nvd0_fan_pwm_ctrl;
 	priv->base.fan.pwm_get = nvd0_fan_pwm_get;
 	priv->base.fan.pwm_set = nvd0_fan_pwm_set;
 	priv->base.fan.pwm_clock = nvd0_fan_pwm_clock;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
index bb940e9..a08ca85 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
@@ -47,6 +47,7 @@ struct nouveau_therm_priv {
 
 		struct dcb_gpio_func tach;
 
+		int (*pwm_ctrl)(struct nouveau_therm *, int line, bool);
 		int (*pwm_get)(struct nouveau_therm *, int line, u32*, u32*);
 		int (*pwm_set)(struct nouveau_therm *, int line, u32, u32);
 		int (*pwm_clock)(struct nouveau_therm *);
-- 
cgit v0.10.2


From 9c3bd3a53129639f10e129b007862340dba16a09 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Wed, 5 Dec 2012 16:21:59 +1000
Subject: drm/nouveau/therm: cleanly separate pwm control logic from therm

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Martin Peres <martin.peres@labri.fr>

diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index 7938d1a..c58d66c 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -106,6 +106,8 @@ nouveau-y += core/subdev/mxm/mxms.o
 nouveau-y += core/subdev/mxm/nv50.o
 nouveau-y += core/subdev/therm/base.o
 nouveau-y += core/subdev/therm/fan.o
+nouveau-y += core/subdev/therm/fannil.o
+nouveau-y += core/subdev/therm/fanpwm.o
 nouveau-y += core/subdev/therm/ic.o
 nouveau-y += core/subdev/therm/temp.o
 nouveau-y += core/subdev/therm/nv40.o
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/therm.h b/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
index a96d43a..ad3aea0 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
@@ -28,6 +28,11 @@ enum nouveau_therm_attr_type {
 struct nouveau_therm {
 	struct nouveau_subdev base;
 
+	int (*pwm_ctrl)(struct nouveau_therm *, int line, bool);
+	int (*pwm_get)(struct nouveau_therm *, int line, u32 *, u32 *);
+	int (*pwm_set)(struct nouveau_therm *, int line, u32, u32);
+	int (*pwm_clock)(struct nouveau_therm *);
+
 	int (*fan_get)(struct nouveau_therm *);
 	int (*fan_set)(struct nouveau_therm *, int);
 	int (*fan_sense)(struct nouveau_therm *);
@@ -47,8 +52,10 @@ nouveau_therm(void *obj)
 
 #define nouveau_therm_create(p,e,o,d)                                          \
 	nouveau_therm_create_((p), (e), (o), sizeof(**d), (void **)d)
-#define nouveau_therm_destroy(p)                                               \
-	nouveau_subdev_destroy(&(p)->base)
+#define nouveau_therm_destroy(p) ({                                            \
+	struct nouveau_therm *therm = (p);                                     \
+        _nouveau_therm_dtor(nv_object(therm));                                 \
+})
 #define nouveau_therm_init(p) ({                                               \
 	struct nouveau_therm *therm = (p);                                     \
         _nouveau_therm_init(nv_object(therm));                                 \
@@ -58,12 +65,11 @@ nouveau_therm(void *obj)
         _nouveau_therm_init(nv_object(therm), (s));                            \
 })
 
-int nouveau_therm_create_(struct nouveau_object *, struct nouveau_object *,
-			  struct nouveau_oclass *, int, void **);
-
-#define _nouveau_therm_dtor _nouveau_subdev_dtor
-int _nouveau_therm_init(struct nouveau_object *);
-int _nouveau_therm_fini(struct nouveau_object *, bool);
+int  nouveau_therm_create_(struct nouveau_object *, struct nouveau_object *,
+			   struct nouveau_oclass *, int, void **);
+void _nouveau_therm_dtor(struct nouveau_object *);
+int  _nouveau_therm_init(struct nouveau_object *);
+int  _nouveau_therm_fini(struct nouveau_object *, bool);
 
 extern struct nouveau_oclass nv40_therm_oclass;
 extern struct nouveau_oclass nv50_therm_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/base.c b/drivers/gpu/drm/nouveau/core/subdev/therm/base.c
index 6260eac..4ebd3c8 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/base.c
@@ -37,11 +37,11 @@ nouveau_therm_attr_get(struct nouveau_therm *therm,
 
 	switch (type) {
 	case NOUVEAU_THERM_ATTR_FAN_MIN_DUTY:
-		return priv->bios_fan.min_duty;
+		return priv->fan->bios.min_duty;
 	case NOUVEAU_THERM_ATTR_FAN_MAX_DUTY:
-		return priv->bios_fan.max_duty;
+		return priv->fan->bios.max_duty;
 	case NOUVEAU_THERM_ATTR_FAN_MODE:
-		return priv->fan.mode;
+		return priv->fan->mode;
 	case NOUVEAU_THERM_ATTR_THRS_FAN_BOOST:
 		return priv->bios_sensor.thrs_fan_boost.temp;
 	case NOUVEAU_THERM_ATTR_THRS_FAN_BOOST_HYST:
@@ -73,16 +73,16 @@ nouveau_therm_attr_set(struct nouveau_therm *therm,
 	case NOUVEAU_THERM_ATTR_FAN_MIN_DUTY:
 		if (value < 0)
 			value = 0;
-		if (value > priv->bios_fan.max_duty)
-			value = priv->bios_fan.max_duty;
-		priv->bios_fan.min_duty = value;
+		if (value > priv->fan->bios.max_duty)
+			value = priv->fan->bios.max_duty;
+		priv->fan->bios.min_duty = value;
 		return 0;
 	case NOUVEAU_THERM_ATTR_FAN_MAX_DUTY:
 		if (value < 0)
 			value = 0;
-		if (value < priv->bios_fan.min_duty)
-			value = priv->bios_fan.min_duty;
-		priv->bios_fan.max_duty = value;
+		if (value < priv->fan->bios.min_duty)
+			value = priv->fan->bios.min_duty;
+		priv->fan->bios.max_duty = value;
 		return 0;
 	case NOUVEAU_THERM_ATTR_FAN_MODE:
 		return nouveau_therm_fan_set_mode(therm, value);
@@ -126,8 +126,8 @@ _nouveau_therm_init(struct nouveau_object *object)
 	if (ret)
 		return ret;
 
-	if (priv->fan.percent >= 0)
-		therm->fan_set(therm, priv->fan.percent);
+	if (priv->fan->percent >= 0)
+		therm->fan_set(therm, priv->fan->percent);
 
 	return 0;
 }
@@ -138,7 +138,7 @@ _nouveau_therm_fini(struct nouveau_object *object, bool suspend)
 	struct nouveau_therm *therm = (void *)object;
 	struct nouveau_therm_priv *priv = (void *)therm;
 
-	priv->fan.percent = therm->fan_get(therm);
+	priv->fan->percent = therm->fan_get(therm);
 
 	return nouveau_subdev_fini(&therm->base, suspend);
 }
@@ -158,10 +158,6 @@ nouveau_therm_create_(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	nouveau_therm_ic_ctor(&priv->base);
-	nouveau_therm_sensor_ctor(&priv->base);
-	nouveau_therm_fan_ctor(&priv->base);
-
 	priv->base.fan_get = nouveau_therm_fan_user_get;
 	priv->base.fan_set = nouveau_therm_fan_user_set;
 	priv->base.fan_sense = nouveau_therm_fan_sense;
@@ -169,3 +165,20 @@ nouveau_therm_create_(struct nouveau_object *parent,
 	priv->base.attr_set = nouveau_therm_attr_set;
 	return 0;
 }
+
+int
+nouveau_therm_preinit(struct nouveau_therm *therm)
+{
+	nouveau_therm_ic_ctor(therm);
+	nouveau_therm_sensor_ctor(therm);
+	nouveau_therm_fan_ctor(therm);
+	return 0;
+}
+
+void
+_nouveau_therm_dtor(struct nouveau_object *object)
+{
+	struct nouveau_therm_priv *priv = (void *)object;
+	kfree(priv->fan);
+	nouveau_subdev_destroy(&priv->base.base);
+}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
index b4d1449..e2fbddd 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
@@ -35,73 +35,23 @@ int
 nouveau_therm_fan_get(struct nouveau_therm *therm)
 {
 	struct nouveau_therm_priv *priv = (void *)therm;
-	struct nouveau_gpio *gpio = nouveau_gpio(therm);
-	struct dcb_gpio_func func;
-	int card_type = nv_device(therm)->card_type;
-	u32 divs, duty;
-	int ret;
-
-	if (!priv->fan.pwm_get)
-		return -ENODEV;
-
-	ret = gpio->find(gpio, 0, DCB_GPIO_PWM_FAN, 0xff, &func);
-	if (ret == 0) {
-		ret = priv->fan.pwm_get(therm, func.line, &divs, &duty);
-		if (ret == 0 && divs) {
-			divs = max(divs, duty);
-			if (card_type <= NV_40 || (func.log[0] & 1))
-				duty = divs - duty;
-			return (duty * 100) / divs;
-		}
-
-		return gpio->get(gpio, 0, func.func, func.line) * 100;
-	}
-
-	return -ENODEV;
+	return priv->fan->get(therm);
 }
 
 int
 nouveau_therm_fan_set(struct nouveau_therm *therm, int percent)
 {
 	struct nouveau_therm_priv *priv = (void *)therm;
-	struct nouveau_gpio *gpio = nouveau_gpio(therm);
-	struct dcb_gpio_func func;
-	int card_type = nv_device(therm)->card_type;
-	u32 divs, duty;
-	int ret;
-
-	if (priv->fan.mode == FAN_CONTROL_NONE)
-		return -EINVAL;
 
-	if (!priv->fan.pwm_set)
-		return -ENODEV;
-
-	if (percent < priv->bios_fan.min_duty)
-		percent = priv->bios_fan.min_duty;
-	if (percent > priv->bios_fan.max_duty)
-		percent = priv->bios_fan.max_duty;
-
-	ret = gpio->find(gpio, 0, DCB_GPIO_PWM_FAN, 0xff, &func);
-	if (ret == 0) {
-		divs = priv->bios_perf_fan.pwm_divisor;
-		if (priv->bios_fan.pwm_freq) {
-			divs = 1;
-			if (priv->fan.pwm_clock)
-				divs = priv->fan.pwm_clock(therm);
-			divs /= priv->bios_fan.pwm_freq;
-		}
-
-		duty = ((divs * percent) + 99) / 100;
-		if (card_type <= NV_40 || (func.log[0] & 1))
-			duty = divs - duty;
+	if (percent < priv->fan->bios.min_duty)
+		percent = priv->fan->bios.min_duty;
+	if (percent > priv->fan->bios.max_duty)
+		percent = priv->fan->bios.max_duty;
 
-		ret = priv->fan.pwm_set(therm, func.line, divs, duty);
-		if (ret == 0)
-			ret = priv->fan.pwm_ctrl(therm, func.line, true);
-		return ret;
-	}
+	if (priv->fan->mode == FAN_CONTROL_NONE)
+		return -EINVAL;
 
-	return -ENODEV;
+	return priv->fan->set(therm, percent);
 }
 
 int
@@ -113,7 +63,7 @@ nouveau_therm_fan_sense(struct nouveau_therm *therm)
 	u32 cycles, cur, prev;
 	u64 start, end, tach;
 
-	if (priv->fan.tach.func == DCB_GPIO_UNUSED)
+	if (priv->fan->tach.func == DCB_GPIO_UNUSED)
 		return -ENODEV;
 
 	/* Time a complete rotation and extrapolate to RPM:
@@ -121,12 +71,12 @@ nouveau_therm_fan_sense(struct nouveau_therm *therm)
 	 * We get 4 changes (0 -> 1 -> 0 -> 1) per complete rotation.
 	 */
 	start = ptimer->read(ptimer);
-	prev = gpio->get(gpio, 0, priv->fan.tach.func, priv->fan.tach.line);
+	prev = gpio->get(gpio, 0, priv->fan->tach.func, priv->fan->tach.line);
 	cycles = 0;
 	do {
 		usleep_range(500, 1000); /* supports 0 < rpm < 7500 */
 
-		cur = gpio->get(gpio, 0, priv->fan.tach.func, priv->fan.tach.line);
+		cur = gpio->get(gpio, 0, priv->fan->tach.func, priv->fan->tach.line);
 		if (prev != cur) {
 			if (!start)
 				start = ptimer->read(ptimer);
@@ -150,7 +100,7 @@ nouveau_therm_fan_set_mode(struct nouveau_therm *therm,
 {
 	struct nouveau_therm_priv *priv = (void *)therm;
 
-	if (priv->fan.mode == mode)
+	if (priv->fan->mode == mode)
 		return 0;
 
 	if (mode < FAN_CONTROL_NONE || mode >= FAN_CONTROL_NR)
@@ -168,7 +118,7 @@ nouveau_therm_fan_set_mode(struct nouveau_therm *therm,
 		break;
 	}
 
-	priv->fan.mode = mode;
+	priv->fan->mode = mode;
 	return 0;
 }
 
@@ -183,7 +133,7 @@ nouveau_therm_fan_user_set(struct nouveau_therm *therm, int percent)
 {
 	struct nouveau_therm_priv *priv = (void *)therm;
 
-	if (priv->fan.mode != FAN_CONTROL_MANUAL)
+	if (priv->fan->mode != FAN_CONTROL_MANUAL)
 		return -EINVAL;
 
 	return nouveau_therm_fan_set(therm, percent);
@@ -194,9 +144,9 @@ nouveau_therm_fan_set_defaults(struct nouveau_therm *therm)
 {
 	struct nouveau_therm_priv *priv = (void *)therm;
 
-	priv->bios_fan.pwm_freq = 0;
-	priv->bios_fan.min_duty = 0;
-	priv->bios_fan.max_duty = 100;
+	priv->fan->bios.pwm_freq = 0;
+	priv->fan->bios.min_duty = 0;
+	priv->fan->bios.max_duty = 100;
 }
 
 
@@ -205,13 +155,13 @@ nouveau_therm_fan_safety_checks(struct nouveau_therm *therm)
 {
 	struct nouveau_therm_priv *priv = (void *)therm;
 
-	if (priv->bios_fan.min_duty > 100)
-		priv->bios_fan.min_duty = 100;
-	if (priv->bios_fan.max_duty > 100)
-		priv->bios_fan.max_duty = 100;
+	if (priv->fan->bios.min_duty > 100)
+		priv->fan->bios.min_duty = 100;
+	if (priv->fan->bios.max_duty > 100)
+		priv->fan->bios.max_duty = 100;
 
-	if (priv->bios_fan.min_duty > priv->bios_fan.max_duty)
-		priv->bios_fan.min_duty = priv->bios_fan.max_duty;
+	if (priv->fan->bios.min_duty > priv->fan->bios.max_duty)
+		priv->fan->bios.min_duty = priv->fan->bios.max_duty;
 }
 
 int nouveau_fan_pwm_clock_dummy(struct nouveau_therm *therm)
@@ -225,15 +175,29 @@ nouveau_therm_fan_ctor(struct nouveau_therm *therm)
 	struct nouveau_therm_priv *priv = (void *)therm;
 	struct nouveau_gpio *gpio = nouveau_gpio(therm);
 	struct nouveau_bios *bios = nouveau_bios(therm);
+	struct dcb_gpio_func func;
 	int ret;
 
-	ret = gpio->find(gpio, 0, DCB_GPIO_FAN_SENSE, 0xff, &priv->fan.tach);
+	/* attempt to locate a drivable fan, and determine control method */
+	ret = gpio->find(gpio, 0, DCB_GPIO_PWM_FAN, 0xff, &func);
+	if (ret == 0)
+		ret = nouveau_fanpwm_create(therm, &func);
+	if (ret)
+		ret = nouveau_fannil_create(therm);
+	if (ret)
+		return ret;
+
+	nv_info(therm, "FAN control: %s\n", priv->fan->type);
+
+	/* attempt to detect a tachometer connection */
+	ret = gpio->find(gpio, 0, DCB_GPIO_FAN_SENSE, 0xff, &priv->fan->tach);
 	if (ret)
-		priv->fan.tach.func = DCB_GPIO_UNUSED;
+		priv->fan->tach.func = DCB_GPIO_UNUSED;
 
+	/* other random init... */
 	nouveau_therm_fan_set_defaults(therm);
-	nvbios_perf_fan_parse(bios, &priv->bios_perf_fan);
-	if (nvbios_therm_fan_parse(bios, &priv->bios_fan))
+	nvbios_perf_fan_parse(bios, &priv->fan->perf);
+	if (nvbios_therm_fan_parse(bios, &priv->fan->bios))
 		nv_error(therm, "parsing the thermal table failed\n");
 	nouveau_therm_fan_safety_checks(therm);
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/fannil.c b/drivers/gpu/drm/nouveau/core/subdev/therm/fannil.c
new file mode 100644
index 0000000..82f0432
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/fannil.c
@@ -0,0 +1,54 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+static int
+nouveau_fannil_get(struct nouveau_therm *therm)
+{
+	return -ENODEV;
+}
+
+int
+nouveau_fannil_set(struct nouveau_therm *therm, int percent)
+{
+	return -ENODEV;
+}
+
+int
+nouveau_fannil_create(struct nouveau_therm *therm)
+{
+	struct nouveau_therm_priv *tpriv = (void *)therm;
+	struct nouveau_fan *priv;
+
+	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
+	tpriv->fan = priv;
+	if (!priv)
+		return -ENOMEM;
+
+	priv->type = "none / external";
+	priv->get = nouveau_fannil_get;
+	priv->set = nouveau_fannil_set;
+	return 0;
+}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/fanpwm.c b/drivers/gpu/drm/nouveau/core/subdev/therm/fanpwm.c
new file mode 100644
index 0000000..5f71db8
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/fanpwm.c
@@ -0,0 +1,107 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ * 	    Martin Peres
+ */
+
+#include <core/option.h>
+#include <subdev/gpio.h>
+
+#include "priv.h"
+
+struct nouveau_fanpwm_priv {
+	struct nouveau_fan base;
+	struct dcb_gpio_func func;
+};
+
+static int
+nouveau_fanpwm_get(struct nouveau_therm *therm)
+{
+	struct nouveau_therm_priv *tpriv = (void *)therm;
+	struct nouveau_fanpwm_priv *priv = (void *)tpriv->fan;
+	struct nouveau_gpio *gpio = nouveau_gpio(therm);
+	int card_type = nv_device(therm)->card_type;
+	u32 divs, duty;
+	int ret;
+
+	ret = therm->pwm_get(therm, priv->func.line, &divs, &duty);
+	if (ret == 0 && divs) {
+		divs = max(divs, duty);
+		if (card_type <= NV_40 || (priv->func.log[0] & 1))
+			duty = divs - duty;
+		return (duty * 100) / divs;
+	}
+
+	return gpio->get(gpio, 0, priv->func.func, priv->func.line) * 100;
+}
+
+static int
+nouveau_fanpwm_set(struct nouveau_therm *therm, int percent)
+{
+	struct nouveau_therm_priv *tpriv = (void *)therm;
+	struct nouveau_fanpwm_priv *priv = (void *)tpriv->fan;
+	int card_type = nv_device(therm)->card_type;
+	u32 divs, duty;
+	int ret;
+
+	divs = priv->base.perf.pwm_divisor;
+	if (priv->base.bios.pwm_freq) {
+		divs = 1;
+		if (therm->pwm_clock)
+			divs = therm->pwm_clock(therm);
+		divs /= priv->base.bios.pwm_freq;
+	}
+
+	duty = ((divs * percent) + 99) / 100;
+	if (card_type <= NV_40 || (priv->func.log[0] & 1))
+		duty = divs - duty;
+
+	ret = therm->pwm_set(therm, priv->func.line, divs, duty);
+	if (ret == 0)
+		ret = therm->pwm_ctrl(therm, priv->func.line, true);
+	return ret;
+}
+
+int
+nouveau_fanpwm_create(struct nouveau_therm *therm, struct dcb_gpio_func *func)
+{
+	struct nouveau_device *device = nv_device(therm);
+	struct nouveau_therm_priv *tpriv = (void *)therm;
+	struct nouveau_fanpwm_priv *priv;
+	u32 divs, duty;
+
+	if (!nouveau_boolopt(device->cfgopt, "NvFanPWM", func->param) ||
+	    !therm->pwm_ctrl ||
+	     therm->pwm_get(therm, func->line, &divs, &duty) == -ENODEV)
+		return -ENODEV;
+
+	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
+	tpriv->fan = &priv->base;
+	if (!priv)
+		return -ENOMEM;
+
+	priv->base.type = "PWM";
+	priv->base.get = nouveau_fanpwm_get;
+	priv->base.set = nouveau_fanpwm_set;
+	priv->func = *func;
+	return 0;
+}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c
index 47a15ec..3f7c115 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c
@@ -149,11 +149,11 @@ nv40_therm_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	priv->base.fan.pwm_ctrl = nv40_fan_pwm_ctrl;
-	priv->base.fan.pwm_get = nv40_fan_pwm_get;
-	priv->base.fan.pwm_set = nv40_fan_pwm_set;
+	priv->base.base.pwm_ctrl = nv40_fan_pwm_ctrl;
+	priv->base.base.pwm_get = nv40_fan_pwm_get;
+	priv->base.base.pwm_set = nv40_fan_pwm_set;
 	priv->base.base.temp_get = nv40_temp_get;
-	return 0;
+	return nouveau_therm_preinit(&priv->base.base);
 }
 
 struct nouveau_oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nv50.c
index 6ebad82..8bb4c2a 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nv50.c
@@ -138,12 +138,12 @@ nv50_therm_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	priv->base.fan.pwm_ctrl = nv50_fan_pwm_ctrl;
-	priv->base.fan.pwm_get = nv50_fan_pwm_get;
-	priv->base.fan.pwm_set = nv50_fan_pwm_set;
-	priv->base.fan.pwm_clock = nv50_fan_pwm_clock;
+	priv->base.base.pwm_ctrl = nv50_fan_pwm_ctrl;
+	priv->base.base.pwm_get = nv50_fan_pwm_get;
+	priv->base.base.pwm_set = nv50_fan_pwm_set;
+	priv->base.base.pwm_clock = nv50_fan_pwm_clock;
 	priv->base.base.temp_get = nv50_temp_get;
-	return 0;
+	return nouveau_therm_preinit(&priv->base.base);
 }
 
 struct nouveau_oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c
index 9a8daf4..8625d21 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c
@@ -44,7 +44,7 @@ static int
 nva3_therm_init(struct nouveau_object *object)
 {
 	struct nva3_therm_priv *priv = (void *)object;
-	struct dcb_gpio_func *tach = &priv->base.fan.tach;
+	struct dcb_gpio_func *tach = &priv->base.fan->tach;
 	int ret;
 
 	ret = nouveau_therm_init(&priv->base.base);
@@ -77,12 +77,13 @@ nva3_therm_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	priv->base.fan.pwm_get = nv50_fan_pwm_get;
-	priv->base.fan.pwm_set = nv50_fan_pwm_set;
-	priv->base.fan.pwm_clock = nv50_fan_pwm_clock;
+	priv->base.base.pwm_ctrl = nv50_fan_pwm_ctrl;
+	priv->base.base.pwm_get = nv50_fan_pwm_get;
+	priv->base.base.pwm_set = nv50_fan_pwm_set;
+	priv->base.base.pwm_clock = nv50_fan_pwm_clock;
 	priv->base.base.temp_get = nv50_temp_get;
 	priv->base.base.fan_sense = nva3_therm_fan_sense;
-	return 0;
+	return nouveau_therm_preinit(&priv->base.base);
 }
 
 struct nouveau_oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
index ec06e89..e722dcd 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
@@ -107,8 +107,8 @@ nvd0_therm_init(struct nouveau_object *object)
 
 	/* enable fan tach, count revolutions per-second */
 	nv_mask(priv, 0x00e720, 0x00000003, 0x00000002);
-	if (priv->base.fan.tach.func != DCB_GPIO_UNUSED) {
-		nv_mask(priv, 0x00d79c, 0x000000ff, priv->base.fan.tach.line);
+	if (priv->base.fan->tach.func != DCB_GPIO_UNUSED) {
+		nv_mask(priv, 0x00d79c, 0x000000ff, priv->base.fan->tach.line);
 		nv_wr32(priv, 0x00e724, nv_device(priv)->crystal * 1000);
 		nv_mask(priv, 0x00e720, 0x00000001, 0x00000001);
 	}
@@ -131,13 +131,13 @@ nvd0_therm_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	priv->base.fan.pwm_ctrl = nvd0_fan_pwm_ctrl;
-	priv->base.fan.pwm_get = nvd0_fan_pwm_get;
-	priv->base.fan.pwm_set = nvd0_fan_pwm_set;
-	priv->base.fan.pwm_clock = nvd0_fan_pwm_clock;
+	priv->base.base.pwm_ctrl = nvd0_fan_pwm_ctrl;
+	priv->base.base.pwm_get = nvd0_fan_pwm_get;
+	priv->base.base.pwm_set = nvd0_fan_pwm_set;
+	priv->base.base.pwm_clock = nvd0_fan_pwm_clock;
 	priv->base.base.temp_get = nv50_temp_get;
 	priv->base.base.fan_sense = nva3_therm_fan_sense;
-	return 0;
+	return nouveau_therm_preinit(&priv->base.base);
 }
 
 struct nouveau_oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
index a08ca85..fa93932 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
@@ -32,26 +32,28 @@
 #include <subdev/bios/perf.h>
 #include <subdev/bios/therm.h>
 
+struct nouveau_fan {
+	const char *type;
+	enum nouveau_therm_fan_mode mode;
+	int percent;
+
+	struct nvbios_therm_fan bios;
+	struct nvbios_perf_fan perf;
+
+	int (*get)(struct nouveau_therm *therm);
+	int (*set)(struct nouveau_therm *therm, int percent);
+
+	struct dcb_gpio_func tach;
+};
+
 struct nouveau_therm_priv {
 	struct nouveau_therm base;
 
 	/* bios */
 	struct nvbios_therm_sensor bios_sensor;
-	struct nvbios_therm_fan bios_fan;
-	struct nvbios_perf_fan bios_perf_fan;
 
 	/* fan priv */
-	struct {
-		enum nouveau_therm_fan_mode mode;
-		int percent;
-
-		struct dcb_gpio_func tach;
-
-		int (*pwm_ctrl)(struct nouveau_therm *, int line, bool);
-		int (*pwm_get)(struct nouveau_therm *, int line, u32*, u32*);
-		int (*pwm_set)(struct nouveau_therm *, int line, u32, u32);
-		int (*pwm_clock)(struct nouveau_therm *);
-	} fan;
+	struct nouveau_fan *fan;
 
 	/* ic */
 	struct i2c_client *ic;
@@ -76,6 +78,9 @@ int nouveau_therm_fan_set_mode(struct nouveau_therm *therm,
 
 int nouveau_therm_fan_sense(struct nouveau_therm *therm);
 
+int nouveau_therm_preinit(struct nouveau_therm *);
+
+int nv50_fan_pwm_ctrl(struct nouveau_therm *, int, bool);
 int nv50_fan_pwm_get(struct nouveau_therm *, int, u32 *, u32 *);
 int nv50_fan_pwm_set(struct nouveau_therm *, int, u32, u32);
 int nv50_fan_pwm_clock(struct nouveau_therm *);
@@ -83,4 +88,7 @@ int nv50_temp_get(struct nouveau_therm *therm);
 
 int nva3_therm_fan_sense(struct nouveau_therm *);
 
+int nouveau_fanpwm_create(struct nouveau_therm *, struct dcb_gpio_func *);
+int nouveau_fannil_create(struct nouveau_therm *);
+
 #endif
-- 
cgit v0.10.2


From 0cbf83bbe524d6014daf11005b99084d50489b80 Mon Sep 17 00:00:00 2001
From: Martin Peres <martin.peres@labri.fr>
Date: Mon, 3 Sep 2012 01:37:02 +0200
Subject: drm/nouveau/fan: add toggle fan support

v2: change percent from int to atomic_t
v3: random fixes
v4 (Ben Skeggs):
- adapted for split-out fan-control "protocol" structure
- removed need for timer resched
- support for forcing 'toggle' control on PWM boards

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Martin Peres <martin.peres@labri.fr>

diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index c58d66c..21c5d9a 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -108,6 +108,7 @@ nouveau-y += core/subdev/therm/base.o
 nouveau-y += core/subdev/therm/fan.o
 nouveau-y += core/subdev/therm/fannil.o
 nouveau-y += core/subdev/therm/fanpwm.o
+nouveau-y += core/subdev/therm/fantog.o
 nouveau-y += core/subdev/therm/ic.o
 nouveau-y += core/subdev/therm/temp.o
 nouveau-y += core/subdev/therm/nv40.o
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/timer.h b/drivers/gpu/drm/nouveau/core/include/subdev/timer.h
index c24ec8a..e465d15 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/timer.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/timer.h
@@ -10,6 +10,14 @@ struct nouveau_alarm {
 	void (*func)(struct nouveau_alarm *);
 };
 
+static inline void
+nouveau_alarm_init(struct nouveau_alarm *alarm,
+		   void (*func)(struct nouveau_alarm *))
+{
+	INIT_LIST_HEAD(&alarm->head);
+	alarm->func = func;
+}
+
 bool nouveau_timer_wait_eq(void *, u64 nsec, u32 addr, u32 mask, u32 data);
 bool nouveau_timer_wait_ne(void *, u64 nsec, u32 addr, u32 mask, u32 data);
 bool nouveau_timer_wait_cb(void *, u64 nsec, bool (*func)(void *), void *data);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
index e2fbddd..aed72a2 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
@@ -149,7 +149,6 @@ nouveau_therm_fan_set_defaults(struct nouveau_therm *therm)
 	priv->fan->bios.max_duty = 100;
 }
 
-
 static void
 nouveau_therm_fan_safety_checks(struct nouveau_therm *therm)
 {
@@ -164,11 +163,6 @@ nouveau_therm_fan_safety_checks(struct nouveau_therm *therm)
 		priv->fan->bios.min_duty = priv->fan->bios.max_duty;
 }
 
-int nouveau_fan_pwm_clock_dummy(struct nouveau_therm *therm)
-{
-	return 1;
-}
-
 int
 nouveau_therm_fan_ctor(struct nouveau_therm *therm)
 {
@@ -182,7 +176,9 @@ nouveau_therm_fan_ctor(struct nouveau_therm *therm)
 	ret = gpio->find(gpio, 0, DCB_GPIO_PWM_FAN, 0xff, &func);
 	if (ret == 0)
 		ret = nouveau_fanpwm_create(therm, &func);
-	if (ret)
+	if (ret != 0)
+		ret = nouveau_fantog_create(therm, &func);
+	if (ret != 0)
 		ret = nouveau_fannil_create(therm);
 	if (ret)
 		return ret;
@@ -202,6 +198,5 @@ nouveau_therm_fan_ctor(struct nouveau_therm *therm)
 	nouveau_therm_fan_safety_checks(therm);
 
 	nouveau_therm_fan_set_mode(therm, FAN_CONTROL_NONE);
-
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/fantog.c b/drivers/gpu/drm/nouveau/core/subdev/therm/fantog.c
new file mode 100644
index 0000000..b08e4e2
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/fantog.c
@@ -0,0 +1,117 @@
+/*
+ * Copyright 2012 The Nouveau community
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Martin Peres
+ */
+
+#include "priv.h"
+
+#include <core/object.h>
+#include <core/device.h>
+
+#include <subdev/gpio.h>
+#include <subdev/timer.h>
+
+struct nouveau_fantog_priv {
+	struct nouveau_fan base;
+	struct nouveau_therm *parent;
+	struct nouveau_alarm alarm;
+	spinlock_t lock;
+	u32 period_us;
+	u32 percent;
+	struct dcb_gpio_func func;
+};
+
+static void
+nouveau_fantog_update(struct nouveau_fantog_priv *priv, int percent)
+{
+	struct nouveau_therm_priv *tpriv = (void *)priv->parent;
+	struct nouveau_timer *ptimer = nouveau_timer(tpriv);
+	struct nouveau_gpio *gpio = nouveau_gpio(tpriv);
+	unsigned long flags;
+	int duty;
+
+	spin_lock_irqsave(&priv->lock, flags);
+	if (percent < 0)
+		percent = priv->percent;
+	priv->percent = percent;
+
+	duty = !gpio->get(gpio, 0, DCB_GPIO_PWM_FAN, 0xff);
+	gpio->set(gpio, 0, DCB_GPIO_PWM_FAN, 0xff, duty);
+
+	if (list_empty(&priv->alarm.head) && percent != (duty * 100)) {
+		u64 next_change = (percent * priv->period_us) / 100;
+		if (!duty)
+			next_change = priv->period_us - next_change;
+		ptimer->alarm(ptimer, next_change * 1000, &priv->alarm);
+	}
+	spin_unlock_irqrestore(&priv->lock, flags);
+}
+
+static void
+nouveau_fantog_alarm(struct nouveau_alarm *alarm)
+{
+	struct nouveau_fantog_priv *priv =
+	       container_of(alarm, struct nouveau_fantog_priv, alarm);
+	nouveau_fantog_update(priv, -1);
+}
+
+static int
+nouveau_fantog_get(struct nouveau_therm *therm)
+{
+	struct nouveau_therm_priv *tpriv = (void *)therm;
+	struct nouveau_fantog_priv *priv = (void *)tpriv->fan;
+	return priv->percent;
+}
+
+int
+nouveau_fantog_set(struct nouveau_therm *therm, int percent)
+{
+	struct nouveau_therm_priv *tpriv = (void *)therm;
+	struct nouveau_fantog_priv *priv = (void *)tpriv->fan;
+	if (therm->pwm_ctrl)
+		therm->pwm_ctrl(therm, priv->func.line, false);
+	nouveau_fantog_update(priv, percent);
+	return 0;
+}
+
+int
+nouveau_fantog_create(struct nouveau_therm *therm, struct dcb_gpio_func *func)
+{
+	struct nouveau_therm_priv *tpriv = (void *)therm;
+	struct nouveau_fantog_priv *priv;
+
+	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
+	tpriv->fan = &priv->base;
+	if (!priv)
+		return -ENOMEM;
+
+	priv->parent = therm;
+	priv->base.type = "toggle";
+	priv->base.get = nouveau_fantog_get;
+	priv->base.set = nouveau_fantog_set;
+	nouveau_alarm_init(&priv->alarm, nouveau_fantog_alarm);
+	priv->period_us = 100000; /* 10Hz */
+	priv->percent = 100;
+	priv->func = *func;
+	spin_lock_init(&priv->lock);
+	return 0;
+}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
index fa93932..34c85e6 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
@@ -31,6 +31,7 @@
 #include <subdev/bios/gpio.h>
 #include <subdev/bios/perf.h>
 #include <subdev/bios/therm.h>
+#include <subdev/timer.h>
 
 struct nouveau_fan {
 	const char *type;
@@ -89,6 +90,7 @@ int nv50_temp_get(struct nouveau_therm *therm);
 int nva3_therm_fan_sense(struct nouveau_therm *);
 
 int nouveau_fanpwm_create(struct nouveau_therm *, struct dcb_gpio_func *);
+int nouveau_fantog_create(struct nouveau_therm *, struct dcb_gpio_func *);
 int nouveau_fannil_create(struct nouveau_therm *);
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/timer/nv04.c b/drivers/gpu/drm/nouveau/core/subdev/timer/nv04.c
index c26ca9b..8e1bae4 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/timer/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/timer/nv04.c
@@ -79,7 +79,7 @@ nv04_timer_alarm_trigger(struct nouveau_timer *ptimer)
 
 	/* execute any pending alarm handlers */
 	list_for_each_entry_safe(alarm, atemp, &exec, head) {
-		list_del(&alarm->head);
+		list_del_init(&alarm->head);
 		alarm->func(alarm);
 	}
 }
-- 
cgit v0.10.2


From 06afd4e83b83907b735279bb0f08d74aeb8f2e3b Mon Sep 17 00:00:00 2001
From: Martin Peres <martin.peres@labri.fr>
Date: Wed, 5 Dec 2012 18:42:00 +1000
Subject: drm/nouveau/bios: parse fan bump/slow periods, and trip points

Signed-off-by: Martin Peres <martin.peres@labri.fr>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/therm.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/therm.h
index a2c4296..083541d 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/bios/therm.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/therm.h
@@ -23,11 +23,27 @@ struct nvbios_therm_sensor {
 	struct nvbios_therm_threshold thrs_shutdown;
 };
 
+/* no vbios have more than 6 */
+#define NOUVEAU_TEMP_FAN_TRIP_MAX 10
+struct nouveau_therm_trip_point {
+	int fan_duty;
+	int temp;
+	int hysteresis;
+};
+
 struct nvbios_therm_fan {
 	u16 pwm_freq;
 
 	u8 min_duty;
 	u8 max_duty;
+
+	u16 bump_period;
+	u16 slow_down_period;
+
+	struct nouveau_therm_trip_point trip[NOUVEAU_TEMP_FAN_TRIP_MAX];
+	u8 nr_fan_trip;
+	u8 linear_min_temp;
+	u8 linear_max_temp;
 };
 
 enum nvbios_therm_domain {
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/therm.c b/drivers/gpu/drm/nouveau/core/subdev/bios/therm.c
index 862a08a..b7916a5 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/therm.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/therm.c
@@ -155,10 +155,15 @@ int
 nvbios_therm_fan_parse(struct nouveau_bios *bios,
 			  struct nvbios_therm_fan *fan)
 {
+	struct nouveau_therm_trip_point *cur_trip = NULL;
 	u8 ver, len, i;
 	u16 entry;
 
+	uint8_t duty_lut[] = { 0, 0, 25, 0, 40, 0, 50, 0,
+				75, 0, 85, 0, 100, 0, 100, 0 };
+
 	i = 0;
+	fan->nr_fan_trip = 0;
 	while ((entry = nvbios_therm_entry(bios, i++, &ver, &len))) {
 		s16 value = nv_ro16(bios, entry + 1);
 
@@ -167,9 +172,30 @@ nvbios_therm_fan_parse(struct nouveau_bios *bios,
 			fan->min_duty = value & 0xff;
 			fan->max_duty = (value & 0xff00) >> 8;
 			break;
+		case 0x24:
+			fan->nr_fan_trip++;
+			cur_trip = &fan->trip[fan->nr_fan_trip - 1];
+			cur_trip->hysteresis = value & 0xf;
+			cur_trip->temp = (value & 0xff0) >> 4;
+			cur_trip->fan_duty = duty_lut[(value & 0xf000) >> 12];
+			break;
+		case 0x25:
+			cur_trip = &fan->trip[fan->nr_fan_trip - 1];
+			cur_trip->fan_duty = value;
+			break;
 		case 0x26:
 			fan->pwm_freq = value;
 			break;
+		case 0x3b:
+			fan->bump_period = value;
+			break;
+		case 0x3c:
+			fan->slow_down_period = value;
+			break;
+		case 0x46:
+			fan->linear_min_temp = nv_ro08(bios, entry + 1);
+			fan->linear_max_temp = nv_ro08(bios, entry + 2);
+			break;
 		}
 	}
 
-- 
cgit v0.10.2


From fa37e8dda2617d48fbc6b17dd6e986e7f4c2bc8b Mon Sep 17 00:00:00 2001
From: Martin Peres <martin.peres@labri.fr>
Date: Wed, 5 Dec 2012 19:46:35 +1000
Subject: drm/nouveau/fan: obey fan bump/slow periods as defined by vbios

v2 (Ben Skeggs):
- split from larger patch
- fixed to not require alarm resched patch

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Martin Peres <martin.peres@labri.fr>

diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
index aed72a2..1a0f863 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
@@ -31,6 +31,71 @@
 #include <subdev/gpio.h>
 #include <subdev/timer.h>
 
+static int
+nouveau_fan_update(struct nouveau_fan *fan, bool immediate, int target)
+{
+	struct nouveau_therm *therm = fan->parent;
+	struct nouveau_therm_priv *priv = (void *)therm;
+	struct nouveau_timer *ptimer = nouveau_timer(priv);
+	unsigned long flags;
+	int ret = 0;
+	u32 duty;
+
+	/* update target fan speed, restricting to allowed range */
+	spin_lock_irqsave(&fan->lock, flags);
+	if (target < 0)
+		target = fan->percent;
+	target = max_t(u8, target, fan->bios.min_duty);
+	target = min_t(u8, target, fan->bios.max_duty);
+	fan->percent = target;
+
+	/* smooth out the fanspeed increase/decrease */
+	duty = fan->get(therm);
+	if (!immediate && duty >= 0) {
+		/* the constant "3" is a rough approximation taken from
+		 * nvidia's behaviour.
+		 * it is meant to bump the fan speed more incrementally
+		 */
+		if (duty < target)
+			duty = min(duty + 3, (u32) target);
+		else if (duty > target)
+			duty = max(duty - 3, (u32) target);
+	} else {
+		duty = target;
+	}
+
+	ret = fan->set(therm, duty);
+	if (ret)
+		goto done;
+
+	/* schedule next fan update, if not at target speed already */
+	if (list_empty(&fan->alarm.head) && target != duty) {
+		u16 bump_period = fan->bios.bump_period;
+		u16 slow_down_period = fan->bios.slow_down_period;
+		u64 delay;
+
+		if (duty > target)
+			delay = slow_down_period;
+		else if (duty == target)
+			delay = min(bump_period, slow_down_period) ;
+		else
+			delay = bump_period;
+
+		ptimer->alarm(ptimer, delay * 1000 * 1000, &fan->alarm);
+	}
+
+done:
+	spin_unlock_irqrestore(&fan->lock, flags);
+	return ret;
+}
+
+static void
+nouveau_fan_alarm(struct nouveau_alarm *alarm)
+{
+	struct nouveau_fan *fan = container_of(alarm, struct nouveau_fan, alarm);
+	nouveau_fan_update(fan, false, -1);
+}
+
 int
 nouveau_therm_fan_get(struct nouveau_therm *therm)
 {
@@ -39,19 +104,14 @@ nouveau_therm_fan_get(struct nouveau_therm *therm)
 }
 
 int
-nouveau_therm_fan_set(struct nouveau_therm *therm, int percent)
+nouveau_therm_fan_set(struct nouveau_therm *therm, bool immediate, int percent)
 {
 	struct nouveau_therm_priv *priv = (void *)therm;
 
-	if (percent < priv->fan->bios.min_duty)
-		percent = priv->fan->bios.min_duty;
-	if (percent > priv->fan->bios.max_duty)
-		percent = priv->fan->bios.max_duty;
-
 	if (priv->fan->mode == FAN_CONTROL_NONE)
 		return -EINVAL;
 
-	return priv->fan->set(therm, percent);
+	return nouveau_fan_update(priv->fan, immediate, percent);
 }
 
 int
@@ -136,7 +196,7 @@ nouveau_therm_fan_user_set(struct nouveau_therm *therm, int percent)
 	if (priv->fan->mode != FAN_CONTROL_MANUAL)
 		return -EINVAL;
 
-	return nouveau_therm_fan_set(therm, percent);
+	return nouveau_therm_fan_set(therm, true, percent);
 }
 
 void
@@ -147,6 +207,8 @@ nouveau_therm_fan_set_defaults(struct nouveau_therm *therm)
 	priv->fan->bios.pwm_freq = 0;
 	priv->fan->bios.min_duty = 0;
 	priv->fan->bios.max_duty = 100;
+	priv->fan->bios.bump_period = 500;
+	priv->fan->bios.slow_down_period = 2000;
 }
 
 static void
@@ -190,6 +252,11 @@ nouveau_therm_fan_ctor(struct nouveau_therm *therm)
 	if (ret)
 		priv->fan->tach.func = DCB_GPIO_UNUSED;
 
+	/* initialise fan bump/slow update handling */
+	priv->fan->parent = therm;
+	nouveau_alarm_init(&priv->fan->alarm, nouveau_fan_alarm);
+	spin_lock_init(&priv->fan->lock);
+
 	/* other random init... */
 	nouveau_therm_fan_set_defaults(therm);
 	nvbios_perf_fan_parse(bios, &priv->fan->perf);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/fantog.c b/drivers/gpu/drm/nouveau/core/subdev/therm/fantog.c
index b08e4e2..7e50f14 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/fantog.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/fantog.c
@@ -32,7 +32,6 @@
 
 struct nouveau_fantog_priv {
 	struct nouveau_fan base;
-	struct nouveau_therm *parent;
 	struct nouveau_alarm alarm;
 	spinlock_t lock;
 	u32 period_us;
@@ -43,7 +42,7 @@ struct nouveau_fantog_priv {
 static void
 nouveau_fantog_update(struct nouveau_fantog_priv *priv, int percent)
 {
-	struct nouveau_therm_priv *tpriv = (void *)priv->parent;
+	struct nouveau_therm_priv *tpriv = (void *)priv->base.parent;
 	struct nouveau_timer *ptimer = nouveau_timer(tpriv);
 	struct nouveau_gpio *gpio = nouveau_gpio(tpriv);
 	unsigned long flags;
@@ -104,7 +103,6 @@ nouveau_fantog_create(struct nouveau_therm *therm, struct dcb_gpio_func *func)
 	if (!priv)
 		return -ENOMEM;
 
-	priv->parent = therm;
 	priv->base.type = "toggle";
 	priv->base.get = nouveau_fantog_get;
 	priv->base.set = nouveau_fantog_set;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
index 34c85e6..fca580f 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
@@ -34,13 +34,17 @@
 #include <subdev/timer.h>
 
 struct nouveau_fan {
+	struct nouveau_therm *parent;
 	const char *type;
 	enum nouveau_therm_fan_mode mode;
-	int percent;
 
 	struct nvbios_therm_fan bios;
 	struct nvbios_perf_fan perf;
 
+	struct nouveau_alarm alarm;
+	spinlock_t lock;
+	int percent;
+
 	int (*get)(struct nouveau_therm *therm);
 	int (*set)(struct nouveau_therm *therm, int percent);
 
@@ -71,7 +75,7 @@ int nouveau_therm_sensor_ctor(struct nouveau_therm *therm);
 
 int nouveau_therm_fan_ctor(struct nouveau_therm *therm);
 int nouveau_therm_fan_get(struct nouveau_therm *therm);
-int nouveau_therm_fan_set(struct nouveau_therm *therm, int percent);
+int nouveau_therm_fan_set(struct nouveau_therm *therm, bool now, int percent);
 int nouveau_therm_fan_user_get(struct nouveau_therm *therm);
 int nouveau_therm_fan_user_set(struct nouveau_therm *therm, int percent);
 int nouveau_therm_fan_set_mode(struct nouveau_therm *therm,
-- 
cgit v0.10.2


From 694472f4170c52a18893b0db8c8e3b865a85a457 Mon Sep 17 00:00:00 2001
From: Martin Peres <martin.peres@labri.fr>
Date: Wed, 5 Dec 2012 20:28:09 +1000
Subject: drm/nouveau/therm: implement automatic fan management

v2: improved design but drops safety monitoring (to be in a later patch)
v3: fix locking and mode management
v4: gently fallback to the no-control mode when temperature cannot be got
    and use kernel-provided min/max macros
v5 (Ben Skeggs):
- rebased on my previous patches
v6: fix hysterisis management in trip-based auto fan management

This commit also forbids access to fan management to nvc0+ chipsets as
fan management is already taken care of my PDAEMON's default fw.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Martin Peres <martin.peres@labri.fr>

diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/therm.h b/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
index ad3aea0..7c83ecc 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
@@ -7,6 +7,7 @@
 enum nouveau_therm_fan_mode {
 	FAN_CONTROL_NONE = 0,
 	FAN_CONTROL_MANUAL = 1,
+	FAN_CONTROL_AUTO = 2,
 	FAN_CONTROL_NR,
 };
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/base.c b/drivers/gpu/drm/nouveau/core/subdev/therm/base.c
index 4ebd3c8..dee764c 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/base.c
@@ -29,6 +29,130 @@
 
 #include "priv.h"
 
+static int
+nouveau_therm_update_trip(struct nouveau_therm *therm)
+{
+	struct nouveau_therm_priv *priv = (void *)therm;
+	struct nouveau_therm_trip_point *trip = priv->fan->bios.trip,
+					*cur_trip = NULL,
+					*last_trip = priv->last_trip;
+	u8  temp = therm->temp_get(therm);
+	u16 duty, i;
+
+	/* look for the trip point corresponding to the current temperature */
+	cur_trip = NULL;
+	for (i = 0; i < priv->fan->bios.nr_fan_trip; i++) {
+		if (temp >= trip[i].temp)
+			cur_trip = &trip[i];
+	}
+
+	/* account for the hysteresis cycle */
+	if (last_trip && temp <= (last_trip->temp) &&
+	    temp > (last_trip->temp - last_trip->hysteresis))
+		cur_trip = last_trip;
+
+	if (cur_trip) {
+		duty = cur_trip->fan_duty;
+		priv->last_trip = cur_trip;
+	} else {
+		duty = 0;
+		priv->last_trip = NULL;
+	}
+
+	return duty;
+}
+
+static int
+nouveau_therm_update_linear(struct nouveau_therm *therm)
+{
+	struct nouveau_therm_priv *priv = (void *)therm;
+	u8  linear_min_temp = priv->fan->bios.linear_min_temp;
+	u8  linear_max_temp = priv->fan->bios.linear_max_temp;
+	u8  temp = therm->temp_get(therm);
+	u16 duty;
+
+	duty  = (temp - linear_min_temp);
+	duty *= (priv->fan->bios.max_duty - priv->fan->bios.min_duty);
+	duty /= (linear_max_temp - linear_min_temp);
+	duty += priv->fan->bios.min_duty;
+
+	return duty;
+}
+
+static void
+nouveau_therm_update(struct nouveau_therm *therm, int mode)
+{
+	struct nouveau_timer *ptimer = nouveau_timer(therm);
+	struct nouveau_therm_priv *priv = (void *)therm;
+	unsigned long flags;
+	int duty;
+
+	spin_lock_irqsave(&priv->lock, flags);
+	if (mode < 0)
+		mode = priv->mode;
+	priv->mode = mode;
+
+	switch (mode) {
+	case FAN_CONTROL_MANUAL:
+		duty = priv->fan->percent;
+		break;
+	case FAN_CONTROL_AUTO:
+		if (priv->fan->bios.nr_fan_trip)
+			duty = nouveau_therm_update_trip(therm);
+		else
+			duty = nouveau_therm_update_linear(therm);
+		break;
+	case FAN_CONTROL_NONE:
+	default:
+		goto done;
+	}
+
+	nouveau_therm_fan_set(therm, (mode != FAN_CONTROL_AUTO), duty);
+
+done:
+	if (list_empty(&priv->alarm.head) && (mode == FAN_CONTROL_AUTO))
+		ptimer->alarm(ptimer, 1000000000ULL, &priv->alarm);
+	spin_unlock_irqrestore(&priv->lock, flags);
+}
+
+static void
+nouveau_therm_alarm(struct nouveau_alarm *alarm)
+{
+	struct nouveau_therm_priv *priv =
+	       container_of(alarm, struct nouveau_therm_priv, alarm);
+	nouveau_therm_update(&priv->base, -1);
+}
+
+static int
+nouveau_therm_mode(struct nouveau_therm *therm, int mode)
+{
+	struct nouveau_therm_priv *priv = (void *)therm;
+
+	if (priv->mode == mode)
+		return 0;
+
+	/* The default PDAEMON ucode interferes with fan management */
+	if (nv_device(therm)->card_type >= NV_C0)
+		return -EINVAL;
+
+	switch (mode) {
+	case FAN_CONTROL_NONE:
+		nv_info(therm, "switch to no-control mode\n");
+		break;
+	case FAN_CONTROL_MANUAL:
+		nv_info(therm, "switch to manual mode\n");
+		break;
+	case FAN_CONTROL_AUTO:
+		nv_info(therm, "switch to automatic mode\n");
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	nouveau_therm_update(therm, mode);
+	return 0;
+}
+
 int
 nouveau_therm_attr_get(struct nouveau_therm *therm,
 		       enum nouveau_therm_attr_type type)
@@ -41,7 +165,7 @@ nouveau_therm_attr_get(struct nouveau_therm *therm,
 	case NOUVEAU_THERM_ATTR_FAN_MAX_DUTY:
 		return priv->fan->bios.max_duty;
 	case NOUVEAU_THERM_ATTR_FAN_MODE:
-		return priv->fan->mode;
+		return priv->mode;
 	case NOUVEAU_THERM_ATTR_THRS_FAN_BOOST:
 		return priv->bios_sensor.thrs_fan_boost.temp;
 	case NOUVEAU_THERM_ATTR_THRS_FAN_BOOST_HYST:
@@ -85,7 +209,7 @@ nouveau_therm_attr_set(struct nouveau_therm *therm,
 		priv->fan->bios.max_duty = value;
 		return 0;
 	case NOUVEAU_THERM_ATTR_FAN_MODE:
-		return nouveau_therm_fan_set_mode(therm, value);
+		return nouveau_therm_mode(therm, value);
 	case NOUVEAU_THERM_ATTR_THRS_FAN_BOOST:
 		priv->bios_sensor.thrs_fan_boost.temp = value;
 		return 0;
@@ -158,6 +282,9 @@ nouveau_therm_create_(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
+	nouveau_alarm_init(&priv->alarm, nouveau_therm_alarm);
+	spin_lock_init(&priv->lock);
+
 	priv->base.fan_get = nouveau_therm_fan_user_get;
 	priv->base.fan_set = nouveau_therm_fan_user_set;
 	priv->base.fan_sense = nouveau_therm_fan_sense;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
index 1a0f863..0f210f0 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
@@ -107,10 +107,6 @@ int
 nouveau_therm_fan_set(struct nouveau_therm *therm, bool immediate, int percent)
 {
 	struct nouveau_therm_priv *priv = (void *)therm;
-
-	if (priv->fan->mode == FAN_CONTROL_NONE)
-		return -EINVAL;
-
 	return nouveau_fan_update(priv->fan, immediate, percent);
 }
 
@@ -155,34 +151,6 @@ nouveau_therm_fan_sense(struct nouveau_therm *therm)
 }
 
 int
-nouveau_therm_fan_set_mode(struct nouveau_therm *therm,
-			   enum nouveau_therm_fan_mode mode)
-{
-	struct nouveau_therm_priv *priv = (void *)therm;
-
-	if (priv->fan->mode == mode)
-		return 0;
-
-	if (mode < FAN_CONTROL_NONE || mode >= FAN_CONTROL_NR)
-		return -EINVAL;
-
-	switch (mode)
-	{
-	case FAN_CONTROL_NONE:
-		nv_info(therm, "switch fan to no-control mode\n");
-		break;
-	case FAN_CONTROL_MANUAL:
-		nv_info(therm, "switch fan to manual mode\n");
-		break;
-	case FAN_CONTROL_NR:
-		break;
-	}
-
-	priv->fan->mode = mode;
-	return 0;
-}
-
-int
 nouveau_therm_fan_user_get(struct nouveau_therm *therm)
 {
 	return nouveau_therm_fan_get(therm);
@@ -193,7 +161,7 @@ nouveau_therm_fan_user_set(struct nouveau_therm *therm, int percent)
 {
 	struct nouveau_therm_priv *priv = (void *)therm;
 
-	if (priv->fan->mode != FAN_CONTROL_MANUAL)
+	if (priv->mode != FAN_CONTROL_MANUAL)
 		return -EINVAL;
 
 	return nouveau_therm_fan_set(therm, true, percent);
@@ -209,6 +177,8 @@ nouveau_therm_fan_set_defaults(struct nouveau_therm *therm)
 	priv->fan->bios.max_duty = 100;
 	priv->fan->bios.bump_period = 500;
 	priv->fan->bios.slow_down_period = 2000;
+	priv->fan->bios.linear_min_temp = 40;
+	priv->fan->bios.linear_max_temp = 85;
 }
 
 static void
@@ -263,7 +233,5 @@ nouveau_therm_fan_ctor(struct nouveau_therm *therm)
 	if (nvbios_therm_fan_parse(bios, &priv->fan->bios))
 		nv_error(therm, "parsing the thermal table failed\n");
 	nouveau_therm_fan_safety_checks(therm);
-
-	nouveau_therm_fan_set_mode(therm, FAN_CONTROL_NONE);
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
index fca580f..5c628b5 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
@@ -36,7 +36,6 @@
 struct nouveau_fan {
 	struct nouveau_therm *parent;
 	const char *type;
-	enum nouveau_therm_fan_mode mode;
 
 	struct nvbios_therm_fan bios;
 	struct nvbios_perf_fan perf;
@@ -54,6 +53,12 @@ struct nouveau_fan {
 struct nouveau_therm_priv {
 	struct nouveau_therm base;
 
+	/* automatic thermal management */
+	struct nouveau_alarm alarm;
+	spinlock_t lock;
+	struct nouveau_therm_trip_point *last_trip;
+	int mode;
+
 	/* bios */
 	struct nvbios_therm_sensor bios_sensor;
 
@@ -78,8 +83,6 @@ int nouveau_therm_fan_get(struct nouveau_therm *therm);
 int nouveau_therm_fan_set(struct nouveau_therm *therm, bool now, int percent);
 int nouveau_therm_fan_user_get(struct nouveau_therm *therm);
 int nouveau_therm_fan_user_set(struct nouveau_therm *therm, int percent);
-int nouveau_therm_fan_set_mode(struct nouveau_therm *therm,
-			   enum nouveau_therm_fan_mode mode);
 
 int nouveau_therm_fan_sense(struct nouveau_therm *therm);
 
-- 
cgit v0.10.2


From a10220bbf883931fcf18eb0e850b87d82cfaa412 Mon Sep 17 00:00:00 2001
From: Martin Peres <martin.peres@labri.fr>
Date: Sun, 4 Nov 2012 01:01:53 +0100
Subject: drm/nouveau/pbus: add a PBUS subdev that hands IRQs to the right
 subdevs

We are going to use PTHERM's IRQs for thermal monitoring but we need to route
them first.

On nv31-50, PBUS's IRQ line is shared with GPIOs IRQs.

It seems like nv10-31 GPIO interruptions aren't well handled. I kept the
original behaviour but it is wrong and may lead to an IRQ storm.

Since we enable all PBUS IRQs, we need a way to avoid being stormed if we
don't handle them. The solution I used was to mask the IRQs that have not been
handled. This will also print one message in the logs to let us know.

v2: drop the shared intr handler because of was racy
v3: style fixes
v4: drop a useless construct in the chipset-dependent INTR
v5: add BUS to the disable mask
v6 (Ben Skeggs):
- general tidy to match the rest of the driver's style
- nva3->nvc0, nva3 can be serviced just fine with nv50.c, rnndb even notes
  that the THERM_ALARM bit got left in the hw until fermi anyway.. so, it's
  not going to conflict
- removed the peephole and user stuff, for the moment.. will handle them
  later if we find a good reason to actually care..
- limited INTR_EN to just what we can handle for now, mostly to prevent
  spam of unknown status bits (seen on at least nv4x)

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Martin Peres <martin.peres@labri.fr>

diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index 21c5d9a..ad586f0 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -40,6 +40,10 @@ nouveau-y += core/subdev/bios/mxm.o
 nouveau-y += core/subdev/bios/perf.o
 nouveau-y += core/subdev/bios/pll.o
 nouveau-y += core/subdev/bios/therm.o
+nouveau-y += core/subdev/bus/nv04.o
+nouveau-y += core/subdev/bus/nv31.o
+nouveau-y += core/subdev/bus/nv50.o
+nouveau-y += core/subdev/bus/nvc0.o
 nouveau-y += core/subdev/clock/nv04.o
 nouveau-y += core/subdev/clock/nv40.o
 nouveau-y += core/subdev/clock/nv50.o
diff --git a/drivers/gpu/drm/nouveau/core/include/core/device.h b/drivers/gpu/drm/nouveau/core/include/core/device.h
index e58b6f0..d351a4e 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/device.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/device.h
@@ -26,6 +26,7 @@ enum nv_subdev_type {
 	 */
 	NVDEV_SUBDEV_MXM,
 	NVDEV_SUBDEV_MC,
+	NVDEV_SUBDEV_BUS,
 	NVDEV_SUBDEV_TIMER,
 	NVDEV_SUBDEV_FB,
 	NVDEV_SUBDEV_LTCG,
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bus.h b/drivers/gpu/drm/nouveau/core/include/subdev/bus.h
new file mode 100644
index 0000000..7d88ec4
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bus.h
@@ -0,0 +1,41 @@
+#ifndef __NOUVEAU_BUS_H__
+#define __NOUVEAU_BUS_H__
+
+#include <core/subdev.h>
+#include <core/device.h>
+
+struct nouveau_bus_intr {
+	u32 stat;
+	u32 unit;
+};
+
+struct nouveau_bus {
+	struct nouveau_subdev base;
+};
+
+static inline struct nouveau_bus *
+nouveau_bus(void *obj)
+{
+	return (void *)nv_device(obj)->subdev[NVDEV_SUBDEV_BUS];
+}
+
+#define nouveau_bus_create(p, e, o, d)                                         \
+	nouveau_subdev_create_((p), (e), (o), 0, "PBUS", "master",             \
+			       sizeof(**d), (void **)d)
+#define nouveau_bus_destroy(p)                                                 \
+	nouveau_subdev_destroy(&(p)->base)
+#define nouveau_bus_init(p)                                                    \
+	nouveau_subdev_init(&(p)->base)
+#define nouveau_bus_fini(p, s)                                                 \
+	nouveau_subdev_fini(&(p)->base, (s))
+
+#define _nouveau_bus_dtor _nouveau_subdev_dtor
+#define _nouveau_bus_init _nouveau_subdev_init
+#define _nouveau_bus_fini _nouveau_subdev_fini
+
+extern struct nouveau_oclass nv04_bus_oclass;
+extern struct nouveau_oclass nv31_bus_oclass;
+extern struct nouveau_oclass nv50_bus_oclass;
+extern struct nouveau_oclass nvc0_bus_oclass;
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bus/nv04.c b/drivers/gpu/drm/nouveau/core/subdev/bus/nv04.c
new file mode 100644
index 0000000..0c224e3
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/bus/nv04.c
@@ -0,0 +1,95 @@
+/*
+ * Copyright 2012 Nouveau Community
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Martin Peres <martin.peres@labri.fr>
+ *          Ben Skeggs
+ */
+
+#include <subdev/bus.h>
+
+struct nv04_bus_priv {
+	struct nouveau_bus base;
+};
+
+void
+nv04_bus_intr(struct nouveau_subdev *subdev)
+{
+	struct nouveau_bus *pbus = nouveau_bus(subdev);
+	u32 stat = nv_rd32(pbus, 0x001100) & nv_rd32(pbus, 0x001140);
+
+	if (stat & 0x00000001) {
+		nv_error(pbus, "BUS ERROR\n");
+		stat &= ~0x00000001;
+		nv_wr32(pbus, 0x001100, 0x00000001);
+	}
+
+	if (stat & 0x00000110) {
+		subdev = nouveau_subdev(subdev, NVDEV_SUBDEV_GPIO);
+		if (subdev && subdev->intr)
+			subdev->intr(subdev);
+		stat &= ~0x00000110;
+		nv_wr32(pbus, 0x001100, 0x00000110);
+	}
+
+	if (stat) {
+		nv_error(pbus, "unknown intr 0x%08x\n", stat);
+		nv_mask(pbus, 0x001140, stat, 0x00000000);
+	}
+}
+
+static int
+nv04_bus_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+	      struct nouveau_oclass *oclass, void *data, u32 size,
+	      struct nouveau_object **pobject)
+{
+	struct nv04_bus_priv *priv;
+	int ret;
+
+	ret = nouveau_bus_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	nv_subdev(priv)->intr = nv04_bus_intr;
+	return 0;
+}
+
+static int
+nv04_bus_init(struct nouveau_object *object)
+{
+	struct nv04_bus_priv *priv = (void *)object;
+
+	nv_wr32(priv, 0x001100, 0xffffffff);
+	nv_wr32(priv, 0x001140, 0x00000111);
+
+	return nouveau_bus_init(&priv->base);
+}
+
+struct nouveau_oclass
+nv04_bus_oclass = {
+	.handle = NV_SUBDEV(BUS, 0x04),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv04_bus_ctor,
+		.dtor = _nouveau_bus_dtor,
+		.init = nv04_bus_init,
+		.fini = _nouveau_bus_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bus/nv31.c b/drivers/gpu/drm/nouveau/core/subdev/bus/nv31.c
new file mode 100644
index 0000000..d996b5d
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/bus/nv31.c
@@ -0,0 +1,106 @@
+/*
+ * Copyright 2012 Nouveau Community
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Martin Peres <martin.peres@labri.fr>
+ *          Ben Skeggs
+ */
+
+#include <subdev/bus.h>
+
+struct nv31_bus_priv {
+	struct nouveau_bus base;
+};
+
+static void
+nv31_bus_intr(struct nouveau_subdev *subdev)
+{
+	struct nouveau_bus *pbus = nouveau_bus(subdev);
+	u32 stat = nv_rd32(pbus, 0x001100) & nv_rd32(pbus, 0x001140);
+	u32 gpio = nv_rd32(pbus, 0x001104) & nv_rd32(pbus, 0x001144);
+
+	if (gpio) {
+		subdev = nouveau_subdev(pbus, NVDEV_SUBDEV_GPIO);
+		if (subdev && subdev->intr)
+			subdev->intr(subdev);
+	}
+
+	if (stat & 0x00000008) {
+		nv_error(pbus, "MMIO FAULT\n"); /* NV41- */
+		stat &= ~0x00000008;
+		nv_wr32(pbus, 0x001100, 0x00000008);
+	}
+
+	if (stat & 0x00070000) {
+		subdev = nouveau_subdev(pbus, NVDEV_SUBDEV_THERM);
+		if (subdev && subdev->intr)
+			subdev->intr(subdev);
+		stat &= ~0x00070000;
+		nv_wr32(pbus, 0x001100, 0x00070000);
+	}
+
+	if (stat) {
+		nv_error(pbus, "unknown intr 0x%08x\n", stat);
+		nv_mask(pbus, 0x001140, stat, 0x00000000);
+	}
+}
+
+static int
+nv31_bus_init(struct nouveau_object *object)
+{
+	struct nv31_bus_priv *priv = (void *)object;
+	int ret;
+
+	ret = nouveau_bus_init(&priv->base);
+	if (ret)
+		return ret;
+
+	nv_wr32(priv, 0x001100, 0xffffffff);
+	nv_wr32(priv, 0x001140, 0x00070008);
+	return 0;
+}
+
+static int
+nv31_bus_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+	      struct nouveau_oclass *oclass, void *data, u32 size,
+	      struct nouveau_object **pobject)
+{
+	struct nv31_bus_priv *priv;
+	int ret;
+
+	ret = nouveau_bus_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	nv_subdev(priv)->intr = nv31_bus_intr;
+	return 0;
+}
+
+struct nouveau_oclass
+nv31_bus_oclass = {
+	.handle = NV_SUBDEV(BUS, 0x31),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv31_bus_ctor,
+		.dtor = _nouveau_bus_dtor,
+		.init = nv31_bus_init,
+		.fini = _nouveau_bus_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bus/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/bus/nv50.c
new file mode 100644
index 0000000..0eed34f
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/bus/nv50.c
@@ -0,0 +1,99 @@
+/*
+ * Copyright 2012 Nouveau Community
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Martin Peres <martin.peres@labri.fr>
+ *          Ben Skeggs
+ */
+
+#include <subdev/bus.h>
+
+struct nv50_bus_priv {
+	struct nouveau_bus base;
+};
+
+static void
+nv50_bus_intr(struct nouveau_subdev *subdev)
+{
+	struct nouveau_bus *pbus = nouveau_bus(subdev);
+	u32 stat = nv_rd32(pbus, 0x001100) & nv_rd32(pbus, 0x001140);
+
+	if (stat & 0x00000008) {
+		nv_error(pbus, "MMIO FAULT\n");
+		stat &= ~0x00000008;
+		nv_wr32(pbus, 0x001100, 0x00000008);
+	}
+
+	if (stat & 0x00010000) {
+		subdev = nouveau_subdev(pbus, NVDEV_SUBDEV_THERM);
+		if (subdev && subdev->intr)
+			subdev->intr(subdev);
+		stat &= ~0x00010000;
+		nv_wr32(pbus, 0x001100, 0x00010000);
+	}
+
+	if (stat) {
+		nv_error(pbus, "unknown intr 0x%08x\n", stat);
+		nv_mask(pbus, 0x001140, stat, 0);
+	}
+}
+
+static int
+nv50_bus_init(struct nouveau_object *object)
+{
+	struct nv50_bus_priv *priv = (void *)object;
+	int ret;
+
+	ret = nouveau_bus_init(&priv->base);
+	if (ret)
+		return ret;
+
+	nv_wr32(priv, 0x001100, 0xffffffff);
+	nv_wr32(priv, 0x001140, 0x00010008);
+	return 0;
+}
+
+static int
+nv50_bus_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+	      struct nouveau_oclass *oclass, void *data, u32 size,
+	      struct nouveau_object **pobject)
+{
+	struct nv50_bus_priv *priv;
+	int ret;
+
+	ret = nouveau_bus_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	nv_subdev(priv)->intr = nv50_bus_intr;
+	return 0;
+}
+
+struct nouveau_oclass
+nv50_bus_oclass = {
+	.handle = NV_SUBDEV(BUS, 0x50),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv50_bus_ctor,
+		.dtor = _nouveau_bus_dtor,
+		.init = nv50_bus_init,
+		.fini = _nouveau_bus_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bus/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/bus/nvc0.c
new file mode 100644
index 0000000..10d95f5
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/bus/nvc0.c
@@ -0,0 +1,96 @@
+/*
+ * Copyright 2012 Nouveau Community
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Martin Peres <martin.peres@labri.fr>
+ *          Ben Skeggs
+ */
+
+#include <subdev/bus.h>
+
+struct nvc0_bus_priv {
+	struct nouveau_bus base;
+};
+
+static void
+nvc0_bus_intr(struct nouveau_subdev *subdev)
+{
+	struct nouveau_bus *pbus = nouveau_bus(subdev);
+	u32 stat = nv_rd32(pbus, 0x001100) & nv_rd32(pbus, 0x001140);
+
+	if (stat & 0x0000000e) {
+		nv_error(pbus, "MMIO FAULT [");
+		if (stat & 0x00000002) pr_cont(" !ENGINE");
+		if (stat & 0x00000004) pr_cont(" IBUS");
+		if (stat & 0x00000004) pr_cont(" TIMEOUT");
+		pr_cont("]\n");
+
+		nv_wr32(pbus, 0x001100, (stat & 0x0000000e));
+		stat &= ~0x0000000e;
+	}
+
+	if (stat) {
+		nv_error(pbus, "unknown intr 0x%08x\n", stat);
+		nv_mask(pbus, 0x001140, stat, 0x00000000);
+	}
+}
+
+static int
+nvc0_bus_init(struct nouveau_object *object)
+{
+	struct nvc0_bus_priv *priv = (void *)object;
+	int ret;
+
+	ret = nouveau_bus_init(&priv->base);
+	if (ret)
+		return ret;
+
+	nv_wr32(priv, 0x001100, 0xffffffff);
+	nv_wr32(priv, 0x001140, 0x0000000e);
+	return 0;
+}
+
+static int
+nvc0_bus_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+	      struct nouveau_oclass *oclass, void *data, u32 size,
+	      struct nouveau_object **pobject)
+{
+	struct nvc0_bus_priv *priv;
+	int ret;
+
+	ret = nouveau_bus_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	nv_subdev(priv)->intr = nvc0_bus_intr;
+	return 0;
+}
+
+struct nouveau_oclass
+nvc0_bus_oclass = {
+	.handle = NV_SUBDEV(BUS, 0xc0),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_bus_ctor,
+		.dtor = _nouveau_bus_dtor,
+		.init = nvc0_bus_init,
+		.fini = _nouveau_bus_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/base.c b/drivers/gpu/drm/nouveau/core/subdev/device/base.c
index f8a7ed4..6837213 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/device/base.c
@@ -66,6 +66,7 @@ static const u64 disable_map[] = {
 	[NVDEV_SUBDEV_CLOCK]	= NV_DEVICE_DISABLE_CORE,
 	[NVDEV_SUBDEV_MXM]	= NV_DEVICE_DISABLE_CORE,
 	[NVDEV_SUBDEV_MC]	= NV_DEVICE_DISABLE_CORE,
+	[NVDEV_SUBDEV_BUS]	= NV_DEVICE_DISABLE_CORE,
 	[NVDEV_SUBDEV_TIMER]	= NV_DEVICE_DISABLE_CORE,
 	[NVDEV_SUBDEV_FB]	= NV_DEVICE_DISABLE_CORE,
 	[NVDEV_SUBDEV_LTCG]	= NV_DEVICE_DISABLE_CORE,
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nv04.c b/drivers/gpu/drm/nouveau/core/subdev/device/nv04.c
index 8626d0d..f76bd4e 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/device/nv04.c
@@ -24,6 +24,7 @@
 
 #include <subdev/device.h>
 #include <subdev/bios.h>
+#include <subdev/bus.h>
 #include <subdev/i2c.h>
 #include <subdev/clock.h>
 #include <subdev/devinit.h>
@@ -50,6 +51,7 @@ nv04_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv04_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv04_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv04_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv04_instmem_oclass;
@@ -67,6 +69,7 @@ nv04_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv05_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv04_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv04_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv04_instmem_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nv10.c b/drivers/gpu/drm/nouveau/core/subdev/device/nv10.c
index 9c40b0f..46d55ea 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nv10.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/device/nv10.c
@@ -24,6 +24,7 @@
 
 #include <subdev/device.h>
 #include <subdev/bios.h>
+#include <subdev/bus.h>
 #include <subdev/gpio.h>
 #include <subdev/i2c.h>
 #include <subdev/clock.h>
@@ -52,6 +53,7 @@ nv10_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv10_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv04_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv10_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv04_instmem_oclass;
@@ -68,6 +70,7 @@ nv10_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv10_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv04_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv10_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv04_instmem_oclass;
@@ -86,6 +89,7 @@ nv10_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv10_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv04_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv10_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv04_instmem_oclass;
@@ -104,6 +108,7 @@ nv10_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv04_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv1a_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv04_instmem_oclass;
@@ -122,6 +127,7 @@ nv10_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv10_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv04_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv10_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv04_instmem_oclass;
@@ -140,6 +146,7 @@ nv10_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv10_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv04_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv10_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv04_instmem_oclass;
@@ -158,6 +165,7 @@ nv10_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv04_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv1a_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv04_instmem_oclass;
@@ -176,6 +184,7 @@ nv10_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv10_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv04_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv10_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv04_instmem_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nv20.c b/drivers/gpu/drm/nouveau/core/subdev/device/nv20.c
index 74f88f4..ad09854 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nv20.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/device/nv20.c
@@ -24,6 +24,7 @@
 
 #include <subdev/device.h>
 #include <subdev/bios.h>
+#include <subdev/bus.h>
 #include <subdev/gpio.h>
 #include <subdev/i2c.h>
 #include <subdev/clock.h>
@@ -53,6 +54,7 @@ nv20_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv20_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv04_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv20_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv04_instmem_oclass;
@@ -71,6 +73,7 @@ nv20_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv20_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv04_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv25_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv04_instmem_oclass;
@@ -89,6 +92,7 @@ nv20_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv20_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv04_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv25_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv04_instmem_oclass;
@@ -107,6 +111,7 @@ nv20_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv20_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv04_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv25_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv04_instmem_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nv30.c b/drivers/gpu/drm/nouveau/core/subdev/device/nv30.c
index 0ac1b2c..2a98223 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nv30.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/device/nv30.c
@@ -24,6 +24,7 @@
 
 #include <subdev/device.h>
 #include <subdev/bios.h>
+#include <subdev/bus.h>
 #include <subdev/gpio.h>
 #include <subdev/i2c.h>
 #include <subdev/clock.h>
@@ -53,6 +54,7 @@ nv30_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv20_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv04_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv30_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv04_instmem_oclass;
@@ -71,6 +73,7 @@ nv30_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv20_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv04_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv35_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv04_instmem_oclass;
@@ -89,6 +92,7 @@ nv30_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv20_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv31_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv30_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv04_instmem_oclass;
@@ -108,6 +112,7 @@ nv30_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv20_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv31_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv36_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv04_instmem_oclass;
@@ -127,6 +132,7 @@ nv30_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv10_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv31_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv10_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv04_instmem_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nv40.c b/drivers/gpu/drm/nouveau/core/subdev/device/nv40.c
index 41d5968..3e43460 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/device/nv40.c
@@ -24,6 +24,8 @@
 
 #include <subdev/device.h>
 #include <subdev/bios.h>
+#include <subdev/bus.h>
+#include <subdev/vm.h>
 #include <subdev/gpio.h>
 #include <subdev/i2c.h>
 #include <subdev/clock.h>
@@ -55,6 +57,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv31_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv40_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv40_instmem_oclass;
@@ -75,6 +78,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv31_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv41_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv40_instmem_oclass;
@@ -95,6 +99,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv31_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv41_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv40_instmem_oclass;
@@ -115,6 +120,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv31_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv41_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv40_instmem_oclass;
@@ -135,6 +141,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv31_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv40_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv40_instmem_oclass;
@@ -155,6 +162,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv31_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv47_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv40_instmem_oclass;
@@ -175,6 +183,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv31_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv49_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv40_instmem_oclass;
@@ -195,6 +204,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv31_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv49_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv40_instmem_oclass;
@@ -215,6 +225,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv44_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv31_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv44_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv40_instmem_oclass;
@@ -235,6 +246,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv44_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv31_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv46_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv40_instmem_oclass;
@@ -255,6 +267,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv44_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv31_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv44_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv40_instmem_oclass;
@@ -275,6 +288,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv44_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv31_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv46_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv40_instmem_oclass;
@@ -295,6 +309,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv44_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv31_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv4e_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv40_instmem_oclass;
@@ -315,6 +330,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv44_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv31_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv46_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv40_instmem_oclass;
@@ -335,6 +351,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv44_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv31_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv46_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv40_instmem_oclass;
@@ -355,6 +372,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv44_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv31_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv46_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv40_instmem_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/device/nv50.c
index 940c401..0444f46 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/device/nv50.c
@@ -24,6 +24,7 @@
 
 #include <subdev/device.h>
 #include <subdev/bios.h>
+#include <subdev/bus.h>
 #include <subdev/gpio.h>
 #include <subdev/i2c.h>
 #include <subdev/clock.h>
@@ -63,6 +64,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv50_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv50_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv50_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv50_instmem_oclass;
@@ -85,6 +87,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv50_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv50_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv50_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv50_instmem_oclass;
@@ -110,6 +113,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv50_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv50_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv50_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv50_instmem_oclass;
@@ -135,6 +139,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv50_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv50_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv50_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv50_instmem_oclass;
@@ -160,6 +165,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv50_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv50_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv50_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv50_instmem_oclass;
@@ -185,6 +191,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv50_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv50_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv50_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv50_instmem_oclass;
@@ -210,6 +217,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv98_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv50_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv50_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv50_instmem_oclass;
@@ -235,6 +243,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv98_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv50_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv50_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv50_instmem_oclass;
@@ -260,6 +269,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv98_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv50_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv50_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv50_instmem_oclass;
@@ -285,6 +295,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv98_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv50_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv50_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv50_instmem_oclass;
@@ -310,6 +321,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv98_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv50_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv50_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv50_instmem_oclass;
@@ -336,6 +348,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv98_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv50_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv50_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv50_instmem_oclass;
@@ -361,6 +374,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv98_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv50_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv50_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv50_instmem_oclass;
@@ -386,6 +400,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv98_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv50_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nv50_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv50_instmem_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/device/nvc0.c
index 005d0b6..a3e0e5d 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/device/nvc0.c
@@ -24,6 +24,7 @@
 
 #include <subdev/device.h>
 #include <subdev/bios.h>
+#include <subdev/bus.h>
 #include <subdev/gpio.h>
 #include <subdev/i2c.h>
 #include <subdev/clock.h>
@@ -63,6 +64,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nvc0_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_LTCG   ] = &nvc0_ltcg_oclass;
@@ -91,6 +93,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nvc0_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_LTCG   ] = &nvc0_ltcg_oclass;
@@ -119,6 +122,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nvc0_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_LTCG   ] = &nvc0_ltcg_oclass;
@@ -147,6 +151,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nvc0_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_LTCG   ] = &nvc0_ltcg_oclass;
@@ -175,6 +180,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nvc0_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_LTCG   ] = &nvc0_ltcg_oclass;
@@ -203,6 +209,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nvc0_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_LTCG   ] = &nvc0_ltcg_oclass;
@@ -231,6 +238,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nvc0_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_LTCG   ] = &nvc0_ltcg_oclass;
@@ -259,6 +267,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nvc0_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_LTCG   ] = &nvc0_ltcg_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nve0.c b/drivers/gpu/drm/nouveau/core/subdev/device/nve0.c
index 7bef3b1..59d0d0f 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/device/nve0.c
@@ -24,6 +24,7 @@
 
 #include <subdev/device.h>
 #include <subdev/bios.h>
+#include <subdev/bus.h>
 #include <subdev/gpio.h>
 #include <subdev/i2c.h>
 #include <subdev/clock.h>
@@ -63,6 +64,7 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nvc0_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_LTCG   ] = &nvc0_ltcg_oclass;
@@ -91,6 +93,7 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nvc0_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_LTCG   ] = &nvc0_ltcg_oclass;
@@ -119,6 +122,7 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] = &nvc0_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_LTCG   ] = &nvc0_ltcg_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/gpio/nv10.c b/drivers/gpu/drm/nouveau/core/subdev/gpio/nv10.c
index 168d16a..3bed67a 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/gpio/nv10.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/gpio/nv10.c
@@ -141,8 +141,6 @@ nv10_gpio_init(struct nouveau_object *object)
 	if (ret)
 		return ret;
 
-	nv_wr32(priv, 0x001140, 0x00000000);
-	nv_wr32(priv, 0x001100, 0xffffffff);
 	nv_wr32(priv, 0x001144, 0x00000000);
 	nv_wr32(priv, 0x001104, 0xffffffff);
 	return 0;
@@ -152,7 +150,6 @@ static int
 nv10_gpio_fini(struct nouveau_object *object, bool suspend)
 {
 	struct nv10_gpio_priv *priv = (void *)object;
-	nv_wr32(priv, 0x001140, 0x00000000);
 	nv_wr32(priv, 0x001144, 0x00000000);
 	return nouveau_gpio_fini(&priv->base, suspend);
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/mc/nv04.c b/drivers/gpu/drm/nouveau/core/subdev/mc/nv04.c
index 23ebe47..89da8fa 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/mc/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/mc/nv04.c
@@ -37,7 +37,7 @@ nv04_mc_intr[] = {
 	{ 0x00100000, NVDEV_SUBDEV_TIMER },
 	{ 0x01000000, NVDEV_ENGINE_DISP },	/* NV04- PCRTC0 */
 	{ 0x02000000, NVDEV_ENGINE_DISP },	/* NV11- PCRTC1 */
-	{ 0x10000000, NVDEV_SUBDEV_GPIO },	/* PBUS */
+	{ 0x10000000, NVDEV_SUBDEV_BUS },
 	{ 0x80000000, NVDEV_ENGINE_SW },
 	{}
 };
diff --git a/drivers/gpu/drm/nouveau/core/subdev/mc/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/mc/nv50.c
index 8d759f83..5965add 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/mc/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/mc/nv50.c
@@ -38,6 +38,7 @@ nv50_mc_intr[] = {
 	{ 0x00100000, NVDEV_SUBDEV_TIMER },
 	{ 0x00200000, NVDEV_SUBDEV_GPIO },
 	{ 0x04000000, NVDEV_ENGINE_DISP },
+	{ 0x10000000, NVDEV_SUBDEV_BUS },
 	{ 0x80000000, NVDEV_ENGINE_SW },
 	{ 0x0000d101, NVDEV_SUBDEV_FB },
 	{},
diff --git a/drivers/gpu/drm/nouveau/core/subdev/mc/nv98.c b/drivers/gpu/drm/nouveau/core/subdev/mc/nv98.c
index ceb5c83..b93fd09 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/mc/nv98.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/mc/nv98.c
@@ -39,6 +39,7 @@ nv98_mc_intr[] = {
 	{ 0x00200000, NVDEV_SUBDEV_GPIO },
 	{ 0x00400000, NVDEV_ENGINE_COPY0 },	/* NVA3-     */
 	{ 0x04000000, NVDEV_ENGINE_DISP },
+	{ 0x10000000, NVDEV_SUBDEV_BUS },
 	{ 0x80000000, NVDEV_ENGINE_SW },
 	{ 0x0040d101, NVDEV_SUBDEV_FB },
 	{},
diff --git a/drivers/gpu/drm/nouveau/core/subdev/mc/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/mc/nvc0.c
index 9279668..42bbf72 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/mc/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/mc/nvc0.c
@@ -36,11 +36,13 @@ nvc0_mc_intr[] = {
 	{ 0x00000100, NVDEV_ENGINE_FIFO },
 	{ 0x00001000, NVDEV_ENGINE_GR },
 	{ 0x00008000, NVDEV_ENGINE_BSP },
+	{ 0x00040000, NVDEV_SUBDEV_THERM },
 	{ 0x00020000, NVDEV_ENGINE_VP },
 	{ 0x00100000, NVDEV_SUBDEV_TIMER },
 	{ 0x00200000, NVDEV_SUBDEV_GPIO },
 	{ 0x02000000, NVDEV_SUBDEV_LTCG },
 	{ 0x04000000, NVDEV_ENGINE_DISP },
+	{ 0x10000000, NVDEV_SUBDEV_BUS },
 	{ 0x40000000, NVDEV_SUBDEV_IBUS },
 	{ 0x80000000, NVDEV_ENGINE_SW },
 	{},
-- 
cgit v0.10.2


From 382b5bbb4a1ef96be863b71593d23968308b6d24 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Tue, 4 Dec 2012 15:25:28 +1000
Subject: drm/nvc0/bus: report useful data on mmio fault

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/subdev/bus/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/bus/nvc0.c
index 10d95f5..b192d62 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bus/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bus/nvc0.c
@@ -36,12 +36,17 @@ nvc0_bus_intr(struct nouveau_subdev *subdev)
 	u32 stat = nv_rd32(pbus, 0x001100) & nv_rd32(pbus, 0x001140);
 
 	if (stat & 0x0000000e) {
-		nv_error(pbus, "MMIO FAULT [");
-		if (stat & 0x00000002) pr_cont(" !ENGINE");
-		if (stat & 0x00000004) pr_cont(" IBUS");
-		if (stat & 0x00000004) pr_cont(" TIMEOUT");
-		pr_cont("]\n");
+		u32 addr = nv_rd32(pbus, 0x009084);
+		u32 data = nv_rd32(pbus, 0x009088);
 
+		nv_error(pbus, "MMIO %s of 0x%08x FAULT at 0x%06x [ %s%s%s]\n",
+			 (addr & 0x00000002) ? "write" : "read", data,
+			 (addr & 0x00fffffc),
+			 (stat & 0x00000002) ? "!ENGINE " : "",
+			 (stat & 0x00000004) ? "IBUS " : "",
+			 (stat & 0x00000008) ? "TIMEOUT " : "");
+
+		nv_wr32(pbus, 0x009084, 0x00000000);
 		nv_wr32(pbus, 0x001100, (stat & 0x0000000e));
 		stat &= ~0x0000000e;
 	}
-- 
cgit v0.10.2


From 9d7175c808793b3e30db455da7529d3c05b00712 Mon Sep 17 00:00:00 2001
From: Martin Peres <martin.peres@labri.fr>
Date: Fri, 7 Dec 2012 02:26:02 +0100
Subject: drm/nv41/bus: report useful data on mmio fault

Based on Ben Skeggs's nvc0 patch. Tested on my nv4b, 84 and 92.

Signed-off-by: Martin Peres <martin.peres@labri.fr>

diff --git a/drivers/gpu/drm/nouveau/core/subdev/bus/nv31.c b/drivers/gpu/drm/nouveau/core/subdev/bus/nv31.c
index d996b5d..34132ae 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bus/nv31.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bus/nv31.c
@@ -42,8 +42,14 @@ nv31_bus_intr(struct nouveau_subdev *subdev)
 			subdev->intr(subdev);
 	}
 
-	if (stat & 0x00000008) {
-		nv_error(pbus, "MMIO FAULT\n"); /* NV41- */
+	if (stat & 0x00000008) {  /* NV41- */
+		u32 addr = nv_rd32(pbus, 0x009084);
+		u32 data = nv_rd32(pbus, 0x009088);
+
+		nv_error(pbus, "MMIO %s of 0x%08x FAULT at 0x%06x\n",
+			 (addr & 0x00000002) ? "write" : "read", data,
+			 (addr & 0x00fffffc));
+
 		stat &= ~0x00000008;
 		nv_wr32(pbus, 0x001100, 0x00000008);
 	}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bus/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/bus/nv50.c
index 0eed34f..f5b2117 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bus/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bus/nv50.c
@@ -36,7 +36,13 @@ nv50_bus_intr(struct nouveau_subdev *subdev)
 	u32 stat = nv_rd32(pbus, 0x001100) & nv_rd32(pbus, 0x001140);
 
 	if (stat & 0x00000008) {
-		nv_error(pbus, "MMIO FAULT\n");
+		u32 addr = nv_rd32(pbus, 0x009084);
+		u32 data = nv_rd32(pbus, 0x009088);
+
+		nv_error(pbus, "MMIO %s of 0x%08x FAULT at 0x%06x\n",
+			 (addr & 0x00000002) ? "write" : "read", data,
+			 (addr & 0x00fffffc));
+
 		stat &= ~0x00000008;
 		nv_wr32(pbus, 0x001100, 0x00000008);
 	}
-- 
cgit v0.10.2


From 0083b91dac482c70eeb96745d9ef604f904da3e5 Mon Sep 17 00:00:00 2001
From: Martin Peres <martin.peres@labri.fr>
Date: Mon, 5 Nov 2012 00:18:49 +0100
Subject: drm/nouveau/therm: implement support for temperature alarms

For now, we only boost the fan speed to the maximum and auto-mode
when hitting the FAN_BOOST threshold and halt the computer when it
reaches the shutdown temperature. The downclock and critical thresholds
do nothing.

On nv43:50 and nva3+, temperature is polled because of the limited hardware.
I'll improve the nva3+ situation by implementing alarm management in PDAEMON
whenever I can but polling once every second shouldn't be such a problem.

v2 (Ben Skeggs):
- rebased

v3: fixed false-detections and threshold reprogrammation handling on nv50:nvc0

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Martin Peres <martin.peres@labri.fr>

diff --git a/drivers/gpu/drm/nouveau/core/os.h b/drivers/gpu/drm/nouveau/core/os.h
index cfe3b9c..eb49603 100644
--- a/drivers/gpu/drm/nouveau/core/os.h
+++ b/drivers/gpu/drm/nouveau/core/os.h
@@ -16,6 +16,7 @@
 #include <linux/vmalloc.h>
 #include <linux/acpi.h>
 #include <linux/dmi.h>
+#include <linux/reboot.h>
 
 #include <asm/unaligned.h>
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/mc/nv98.c b/drivers/gpu/drm/nouveau/core/subdev/mc/nv98.c
index b93fd09..3a80b29 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/mc/nv98.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/mc/nv98.c
@@ -35,6 +35,7 @@ nv98_mc_intr[] = {
 	{ 0x00001000, NVDEV_ENGINE_GR },
 	{ 0x00004000, NVDEV_ENGINE_CRYPT },	/* NV84:NVA3 */
 	{ 0x00008000, NVDEV_ENGINE_BSP },
+	{ 0x00080000, NVDEV_SUBDEV_THERM },	/* NVA3:NVC0 */
 	{ 0x00100000, NVDEV_SUBDEV_TIMER },
 	{ 0x00200000, NVDEV_SUBDEV_GPIO },
 	{ 0x00400000, NVDEV_ENGINE_COPY0 },	/* NVA3-     */
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/base.c b/drivers/gpu/drm/nouveau/core/subdev/therm/base.c
index dee764c..f54495a 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/base.c
@@ -123,7 +123,7 @@ nouveau_therm_alarm(struct nouveau_alarm *alarm)
 	nouveau_therm_update(&priv->base, -1);
 }
 
-static int
+int
 nouveau_therm_mode(struct nouveau_therm *therm, int mode)
 {
 	struct nouveau_therm_priv *priv = (void *)therm;
@@ -212,27 +212,35 @@ nouveau_therm_attr_set(struct nouveau_therm *therm,
 		return nouveau_therm_mode(therm, value);
 	case NOUVEAU_THERM_ATTR_THRS_FAN_BOOST:
 		priv->bios_sensor.thrs_fan_boost.temp = value;
+		priv->sensor.program_alarms(therm);
 		return 0;
 	case NOUVEAU_THERM_ATTR_THRS_FAN_BOOST_HYST:
 		priv->bios_sensor.thrs_fan_boost.hysteresis = value;
+		priv->sensor.program_alarms(therm);
 		return 0;
 	case NOUVEAU_THERM_ATTR_THRS_DOWN_CLK:
 		priv->bios_sensor.thrs_down_clock.temp = value;
+		priv->sensor.program_alarms(therm);
 		return 0;
 	case NOUVEAU_THERM_ATTR_THRS_DOWN_CLK_HYST:
 		priv->bios_sensor.thrs_down_clock.hysteresis = value;
+		priv->sensor.program_alarms(therm);
 		return 0;
 	case NOUVEAU_THERM_ATTR_THRS_CRITICAL:
 		priv->bios_sensor.thrs_critical.temp = value;
+		priv->sensor.program_alarms(therm);
 		return 0;
 	case NOUVEAU_THERM_ATTR_THRS_CRITICAL_HYST:
 		priv->bios_sensor.thrs_critical.hysteresis = value;
+		priv->sensor.program_alarms(therm);
 		return 0;
 	case NOUVEAU_THERM_ATTR_THRS_SHUTDOWN:
 		priv->bios_sensor.thrs_shutdown.temp = value;
+		priv->sensor.program_alarms(therm);
 		return 0;
 	case NOUVEAU_THERM_ATTR_THRS_SHUTDOWN_HYST:
 		priv->bios_sensor.thrs_shutdown.hysteresis = value;
+		priv->sensor.program_alarms(therm);
 		return 0;
 	}
 
@@ -253,6 +261,7 @@ _nouveau_therm_init(struct nouveau_object *object)
 	if (priv->fan->percent >= 0)
 		therm->fan_set(therm, priv->fan->percent);
 
+	priv->sensor.program_alarms(therm);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c
index 3f7c115..5c17577 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c
@@ -38,6 +38,7 @@ nv40_sensor_setup(struct nouveau_therm *therm)
 	if (device->chipset >= 0x46) {
 		nv_mask(therm, 0x15b8, 0x80000000, 0);
 		nv_wr32(therm, 0x15b0, 0x80003fff);
+		mdelay(10); /* wait for the temperature to stabilize */
 		return nv_rd32(therm, 0x15b4) & 0x3fff;
 	} else {
 		nv_wr32(therm, 0x15b0, 0xff);
@@ -135,6 +136,20 @@ nv40_fan_pwm_set(struct nouveau_therm *therm, int line, u32 divs, u32 duty)
 	return 0;
 }
 
+static void
+nv40_therm_intr(struct nouveau_subdev *subdev)
+{
+	struct nouveau_therm *therm = nouveau_therm(subdev);
+	uint32_t stat = nv_rd32(therm, 0x1100);
+
+	/* traitement */
+
+	/* ack all IRQs */
+	nv_wr32(therm, 0x1100, 0x70000);
+
+	nv_error(therm, "THERM received an IRQ: stat = %x\n", stat);
+}
+
 static int
 nv40_therm_ctor(struct nouveau_object *parent,
 		struct nouveau_object *engine,
@@ -153,6 +168,8 @@ nv40_therm_ctor(struct nouveau_object *parent,
 	priv->base.base.pwm_get = nv40_fan_pwm_get;
 	priv->base.base.pwm_set = nv40_fan_pwm_set;
 	priv->base.base.temp_get = nv40_temp_get;
+	priv->base.sensor.program_alarms = nouveau_therm_program_alarms_polling;
+	nv_subdev(priv)->intr = nv40_therm_intr;
 	return nouveau_therm_preinit(&priv->base.base);
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nv50.c
index 8bb4c2a..4b7fe24 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nv50.c
@@ -124,6 +124,141 @@ nv50_temp_get(struct nouveau_therm *therm)
 	return nv_rd32(therm, 0x20400);
 }
 
+static void
+nv50_therm_program_alarms(struct nouveau_therm *therm)
+{
+	struct nouveau_therm_priv *priv = (void *)therm;
+	struct nvbios_therm_sensor *sensor = &priv->bios_sensor;
+	unsigned long flags;
+
+	spin_lock_irqsave(&priv->sensor.alarm_program_lock, flags);
+
+	/* enable RISING and FALLING IRQs for shutdown, THRS 0, 1, 2 and 4 */
+	nv_wr32(therm, 0x20000, 0x000003ff);
+
+	/* shutdown: The computer should be shutdown when reached */
+	nv_wr32(therm, 0x20484, sensor->thrs_shutdown.hysteresis);
+	nv_wr32(therm, 0x20480, sensor->thrs_shutdown.temp);
+
+	/* THRS_1 : fan boost*/
+	nv_wr32(therm, 0x204c4, sensor->thrs_fan_boost.temp);
+
+	/* THRS_2 : critical */
+	nv_wr32(therm, 0x204c0, sensor->thrs_critical.temp);
+
+	/* THRS_4 : down clock */
+	nv_wr32(therm, 0x20414, sensor->thrs_down_clock.temp);
+	spin_unlock_irqrestore(&priv->sensor.alarm_program_lock, flags);
+
+	nv_info(therm,
+		"Programmed thresholds [ %d(%d), %d(%d), %d(%d), %d(%d) ]\n",
+		sensor->thrs_fan_boost.temp, sensor->thrs_fan_boost.hysteresis,
+		sensor->thrs_down_clock.temp,
+		sensor->thrs_down_clock.hysteresis,
+		sensor->thrs_critical.temp, sensor->thrs_critical.hysteresis,
+		sensor->thrs_shutdown.temp, sensor->thrs_shutdown.hysteresis);
+
+}
+
+/* must be called with alarm_program_lock taken ! */
+static void
+nv50_therm_threshold_hyst_emulation(struct nouveau_therm *therm,
+				   uint32_t thrs_reg, u8 status_bit,
+				   const struct nvbios_therm_threshold *thrs,
+				   enum nouveau_therm_thrs thrs_name)
+{
+	enum nouveau_therm_thrs_direction direction;
+	enum nouveau_therm_thrs_state prev_state, new_state;
+	int temp, cur;
+
+	prev_state = nouveau_therm_sensor_get_threshold_state(therm, thrs_name);
+	temp = nv_rd32(therm, thrs_reg);
+
+	/* program the next threshold */
+	if (temp == thrs->temp) {
+		nv_wr32(therm, thrs_reg, thrs->temp - thrs->hysteresis);
+		new_state = NOUVEAU_THERM_THRS_HIGHER;
+	} else {
+		nv_wr32(therm, thrs_reg, thrs->temp);
+		new_state = NOUVEAU_THERM_THRS_LOWER;
+	}
+
+	/* fix the state (in case someone reprogrammed the alarms) */
+	cur = therm->temp_get(therm);
+	if (new_state == NOUVEAU_THERM_THRS_LOWER && cur > thrs->temp)
+		new_state = NOUVEAU_THERM_THRS_HIGHER;
+	else if (new_state == NOUVEAU_THERM_THRS_HIGHER &&
+		cur < thrs->temp - thrs->hysteresis)
+		new_state = NOUVEAU_THERM_THRS_LOWER;
+	nouveau_therm_sensor_set_threshold_state(therm, thrs_name, new_state);
+
+	/* find the direction */
+	if (prev_state < new_state)
+		direction = NOUVEAU_THERM_THRS_RISING;
+	else if (prev_state > new_state)
+		direction = NOUVEAU_THERM_THRS_FALLING;
+	else
+		return;
+
+	/* advertise a change in direction */
+	nouveau_therm_sensor_event(therm, thrs_name, direction);
+}
+
+static void
+nv50_therm_intr(struct nouveau_subdev *subdev)
+{
+	struct nouveau_therm *therm = nouveau_therm(subdev);
+	struct nouveau_therm_priv *priv = (void *)therm;
+	struct nvbios_therm_sensor *sensor = &priv->bios_sensor;
+	unsigned long flags;
+	uint32_t intr;
+
+	spin_lock_irqsave(&priv->sensor.alarm_program_lock, flags);
+
+	intr = nv_rd32(therm, 0x20100);
+
+	/* THRS_4: downclock */
+	if (intr & 0x002) {
+		nv50_therm_threshold_hyst_emulation(therm, 0x20414, 24,
+						  &sensor->thrs_down_clock,
+						  NOUVEAU_THERM_THRS_DOWNCLOCK);
+		intr &= ~0x002;
+	}
+
+	/* shutdown */
+	if (intr & 0x004) {
+		nv50_therm_threshold_hyst_emulation(therm, 0x20480, 20,
+						   &sensor->thrs_shutdown,
+						   NOUVEAU_THERM_THRS_SHUTDOWN);
+		intr &= ~0x004;
+	}
+
+	/* THRS_1 : fan boost */
+	if (intr & 0x008) {
+		nv50_therm_threshold_hyst_emulation(therm, 0x204c4, 21,
+						   &sensor->thrs_fan_boost,
+						   NOUVEAU_THERM_THRS_FANBOOST);
+		intr &= ~0x008;
+	}
+
+	/* THRS_2 : critical */
+	if (intr & 0x010) {
+		nv50_therm_threshold_hyst_emulation(therm, 0x204c0, 22,
+						   &sensor->thrs_critical,
+						   NOUVEAU_THERM_THRS_CRITICAL);
+		intr &= ~0x010;
+	}
+
+	if (intr)
+		nv_error(therm, "unhandled intr 0x%08x\n", intr);
+
+	/* ACK everything */
+	nv_wr32(therm, 0x20100, 0xffffffff);
+	nv_wr32(therm, 0x1100, 0x10000); /* PBUS */
+
+	spin_unlock_irqrestore(&priv->sensor.alarm_program_lock, flags);
+}
+
 static int
 nv50_therm_ctor(struct nouveau_object *parent,
 		struct nouveau_object *engine,
@@ -143,6 +278,24 @@ nv50_therm_ctor(struct nouveau_object *parent,
 	priv->base.base.pwm_set = nv50_fan_pwm_set;
 	priv->base.base.pwm_clock = nv50_fan_pwm_clock;
 	priv->base.base.temp_get = nv50_temp_get;
+	priv->base.sensor.program_alarms = nv50_therm_program_alarms;
+	spin_lock_init(&priv->base.sensor.alarm_program_lock);
+	nv_subdev(priv)->intr = nv50_therm_intr;
+
+	/* init the thresholds */
+	nouveau_therm_sensor_set_threshold_state(&priv->base.base,
+						 NOUVEAU_THERM_THRS_SHUTDOWN,
+						 NOUVEAU_THERM_THRS_LOWER);
+	nouveau_therm_sensor_set_threshold_state(&priv->base.base,
+						 NOUVEAU_THERM_THRS_FANBOOST,
+						 NOUVEAU_THERM_THRS_LOWER);
+	nouveau_therm_sensor_set_threshold_state(&priv->base.base,
+						 NOUVEAU_THERM_THRS_CRITICAL,
+						 NOUVEAU_THERM_THRS_LOWER);
+	nouveau_therm_sensor_set_threshold_state(&priv->base.base,
+						 NOUVEAU_THERM_THRS_DOWNCLOCK,
+						 NOUVEAU_THERM_THRS_LOWER);
+
 	return nouveau_therm_preinit(&priv->base.base);
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c
index 8625d21..2dcc543 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c
@@ -83,6 +83,7 @@ nva3_therm_ctor(struct nouveau_object *parent,
 	priv->base.base.pwm_clock = nv50_fan_pwm_clock;
 	priv->base.base.temp_get = nv50_temp_get;
 	priv->base.base.fan_sense = nva3_therm_fan_sense;
+	priv->base.sensor.program_alarms = nouveau_therm_program_alarms_polling;
 	return nouveau_therm_preinit(&priv->base.base);
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
index e722dcd..d7d30ee 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
@@ -137,6 +137,7 @@ nvd0_therm_ctor(struct nouveau_object *parent,
 	priv->base.base.pwm_clock = nvd0_fan_pwm_clock;
 	priv->base.base.temp_get = nv50_temp_get;
 	priv->base.base.fan_sense = nva3_therm_fan_sense;
+	priv->base.sensor.program_alarms = nouveau_therm_program_alarms_polling;
 	return nouveau_therm_preinit(&priv->base.base);
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
index 5c628b5..c2413e6 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
@@ -50,6 +50,24 @@ struct nouveau_fan {
 	struct dcb_gpio_func tach;
 };
 
+enum nouveau_therm_thrs_direction {
+	NOUVEAU_THERM_THRS_FALLING = 0,
+	NOUVEAU_THERM_THRS_RISING = 1
+};
+
+enum nouveau_therm_thrs_state {
+	NOUVEAU_THERM_THRS_LOWER = 0,
+	NOUVEAU_THERM_THRS_HIGHER = 1
+};
+
+enum nouveau_therm_thrs {
+	NOUVEAU_THERM_THRS_FANBOOST = 0,
+	NOUVEAU_THERM_THRS_DOWNCLOCK = 1,
+	NOUVEAU_THERM_THRS_CRITICAL = 2,
+	NOUVEAU_THERM_THRS_SHUTDOWN = 3,
+	NOUVEAU_THERM_THRS_NR
+};
+
 struct nouveau_therm_priv {
 	struct nouveau_therm base;
 
@@ -65,10 +83,25 @@ struct nouveau_therm_priv {
 	/* fan priv */
 	struct nouveau_fan *fan;
 
+	/* alarms priv */
+	struct {
+		spinlock_t alarm_program_lock;
+		struct nouveau_alarm therm_poll_alarm;
+		enum nouveau_therm_thrs_state alarm_state[NOUVEAU_THERM_THRS_NR];
+		void (*program_alarms)(struct nouveau_therm *);
+	} sensor;
+
+	/* what should be done if the card overheats */
+	struct {
+		void (*downclock)(struct nouveau_therm *, bool active);
+		void (*pause)(struct nouveau_therm *, bool active);
+	} emergency;
+
 	/* ic */
 	struct i2c_client *ic;
 };
 
+int nouveau_therm_mode(struct nouveau_therm *therm, int mode);
 int nouveau_therm_attr_get(struct nouveau_therm *therm,
 		       enum nouveau_therm_attr_type type);
 int nouveau_therm_attr_set(struct nouveau_therm *therm,
@@ -88,6 +121,17 @@ int nouveau_therm_fan_sense(struct nouveau_therm *therm);
 
 int nouveau_therm_preinit(struct nouveau_therm *);
 
+void nouveau_therm_sensor_set_threshold_state(struct nouveau_therm *therm,
+					     enum nouveau_therm_thrs thrs,
+					     enum nouveau_therm_thrs_state st);
+enum nouveau_therm_thrs_state
+nouveau_therm_sensor_get_threshold_state(struct nouveau_therm *therm,
+					 enum nouveau_therm_thrs thrs);
+void nouveau_therm_sensor_event(struct nouveau_therm *therm,
+			        enum nouveau_therm_thrs thrs,
+			        enum nouveau_therm_thrs_direction dir);
+void nouveau_therm_program_alarms_polling(struct nouveau_therm *therm);
+
 int nv50_fan_pwm_ctrl(struct nouveau_therm *, int, bool);
 int nv50_fan_pwm_get(struct nouveau_therm *, int, u32 *, u32 *);
 int nv50_fan_pwm_set(struct nouveau_therm *, int, u32, u32);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/temp.c b/drivers/gpu/drm/nouveau/core/subdev/therm/temp.c
index 2042823..d80c572 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/temp.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/temp.c
@@ -65,12 +65,150 @@ nouveau_therm_temp_safety_checks(struct nouveau_therm *therm)
 		priv->bios_sensor.offset_den = 1;
 }
 
+/* must be called with alarm_program_lock taken ! */
+void nouveau_therm_sensor_set_threshold_state(struct nouveau_therm *therm,
+					     enum nouveau_therm_thrs thrs,
+					     enum nouveau_therm_thrs_state st)
+{
+	struct nouveau_therm_priv *priv = (void *)therm;
+	priv->sensor.alarm_state[thrs] = st;
+}
+
+/* must be called with alarm_program_lock taken ! */
+enum nouveau_therm_thrs_state
+nouveau_therm_sensor_get_threshold_state(struct nouveau_therm *therm,
+					 enum nouveau_therm_thrs thrs)
+{
+	struct nouveau_therm_priv *priv = (void *)therm;
+	return priv->sensor.alarm_state[thrs];
+}
+
+void nouveau_therm_sensor_event(struct nouveau_therm *therm,
+			        enum nouveau_therm_thrs thrs,
+			        enum nouveau_therm_thrs_direction dir)
+{
+	struct nouveau_therm_priv *priv = (void *)therm;
+	bool active;
+	const char *thresolds[] = {
+		"fanboost", "downclock", "critical", "shutdown"
+	};
+	uint8_t temperature = therm->temp_get(therm);
+
+	if (thrs < 0 || thrs > 3)
+		return;
+
+	if (dir == NOUVEAU_THERM_THRS_FALLING)
+		nv_info(therm, "temperature (%u C) went bellow the '%s' threshold\n",
+			temperature, thresolds[thrs]);
+	else
+		nv_info(therm, "temperature (%u C) hit the '%s' threshold\n",
+			temperature, thresolds[thrs]);
+
+	active = (dir == NOUVEAU_THERM_THRS_RISING);
+	switch (thrs) {
+	case NOUVEAU_THERM_THRS_FANBOOST:
+		nouveau_therm_fan_set(therm, true, 100);
+		nouveau_therm_mode(therm, FAN_CONTROL_AUTO);
+		break;
+	case NOUVEAU_THERM_THRS_DOWNCLOCK:
+		if (priv->emergency.downclock)
+			priv->emergency.downclock(therm, active);
+		break;
+	case NOUVEAU_THERM_THRS_CRITICAL:
+		if (priv->emergency.pause)
+			priv->emergency.pause(therm, active);
+		break;
+	case NOUVEAU_THERM_THRS_SHUTDOWN:
+		orderly_poweroff(true);
+		break;
+	case NOUVEAU_THERM_THRS_NR:
+		break;
+	}
+
+}
+
+/* must be called with alarm_program_lock taken ! */
+static void
+nouveau_therm_threshold_hyst_polling(struct nouveau_therm *therm,
+				   const struct nvbios_therm_threshold *thrs,
+				   enum nouveau_therm_thrs thrs_name)
+{
+	enum nouveau_therm_thrs_direction direction;
+	enum nouveau_therm_thrs_state prev_state, new_state;
+	int temp = therm->temp_get(therm);
+
+	prev_state = nouveau_therm_sensor_get_threshold_state(therm, thrs_name);
+
+	if (temp >= thrs->temp && prev_state == NOUVEAU_THERM_THRS_LOWER) {
+		direction = NOUVEAU_THERM_THRS_RISING;
+		new_state = NOUVEAU_THERM_THRS_HIGHER;
+	} else if (temp <= thrs->temp - thrs->hysteresis &&
+			prev_state == NOUVEAU_THERM_THRS_HIGHER) {
+		direction = NOUVEAU_THERM_THRS_FALLING;
+		new_state = NOUVEAU_THERM_THRS_LOWER;
+	} else
+		return; /* nothing to do */
+
+	nouveau_therm_sensor_set_threshold_state(therm, thrs_name, new_state);
+	nouveau_therm_sensor_event(therm, thrs_name, direction);
+}
+
+static void
+alarm_timer_callback(struct nouveau_alarm *alarm)
+{
+	struct nouveau_therm_priv *priv =
+	container_of(alarm, struct nouveau_therm_priv, sensor.therm_poll_alarm);
+	struct nvbios_therm_sensor *sensor = &priv->bios_sensor;
+	struct nouveau_timer *ptimer = nouveau_timer(priv);
+	struct nouveau_therm *therm = &priv->base;
+	unsigned long flags;
+
+	spin_lock_irqsave(&priv->sensor.alarm_program_lock, flags);
+
+	nouveau_therm_threshold_hyst_polling(therm, &sensor->thrs_fan_boost,
+					     NOUVEAU_THERM_THRS_FANBOOST);
+
+	nouveau_therm_threshold_hyst_polling(therm, &sensor->thrs_down_clock,
+					     NOUVEAU_THERM_THRS_DOWNCLOCK);
+
+	nouveau_therm_threshold_hyst_polling(therm, &sensor->thrs_critical,
+					     NOUVEAU_THERM_THRS_CRITICAL);
+
+	nouveau_therm_threshold_hyst_polling(therm, &sensor->thrs_shutdown,
+					     NOUVEAU_THERM_THRS_SHUTDOWN);
+
+	/* schedule the next poll in one second */
+	if (list_empty(&alarm->head))
+		ptimer->alarm(ptimer, 1000 * 1000 * 1000, alarm);
+
+	spin_unlock_irqrestore(&priv->sensor.alarm_program_lock, flags);
+}
+
+void
+nouveau_therm_program_alarms_polling(struct nouveau_therm *therm)
+{
+	struct nouveau_therm_priv *priv = (void *)therm;
+	struct nvbios_therm_sensor *sensor = &priv->bios_sensor;
+
+	nv_info(therm,
+		"programmed thresholds [ %d(%d), %d(%d), %d(%d), %d(%d) ]\n",
+		sensor->thrs_fan_boost.temp, sensor->thrs_fan_boost.hysteresis,
+		sensor->thrs_down_clock.temp,
+		sensor->thrs_down_clock.hysteresis,
+		sensor->thrs_critical.temp, sensor->thrs_critical.hysteresis,
+		sensor->thrs_shutdown.temp, sensor->thrs_shutdown.hysteresis);
+
+	alarm_timer_callback(&priv->sensor.therm_poll_alarm);
+}
+
 int
 nouveau_therm_sensor_ctor(struct nouveau_therm *therm)
 {
 	struct nouveau_therm_priv *priv = (void *)therm;
 	struct nouveau_bios *bios = nouveau_bios(therm);
 
+	nouveau_alarm_init(&priv->sensor.therm_poll_alarm, alarm_timer_callback);
+
 	nouveau_therm_temp_set_defaults(therm);
 	if (nvbios_therm_sensor_parse(bios, NVBIOS_THERM_DOMAIN_CORE,
 				      &priv->bios_sensor))
-- 
cgit v0.10.2


From 12e32896f796ea772d4443c789f9358d46461835 Mon Sep 17 00:00:00 2001
From: Martin Peres <martin.peres@labri.fr>
Date: Tue, 20 Nov 2012 01:14:13 +0100
Subject: drm/nouveau/hwmon: add missing alarm thresholds

Expose all the hysteresis parameters + shutdown (emergency) +
fan_boost (fixed pwm trip point).

Signed-off-by: Martin Peres <martin.peres@labri.fr>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nouveau_pm.c b/drivers/gpu/drm/nouveau/nouveau_pm.c
index a701ff5..d456e10 100644
--- a/drivers/gpu/drm/nouveau/nouveau_pm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_pm.c
@@ -409,6 +409,81 @@ static SENSOR_DEVICE_ATTR(temp1_input, S_IRUGO, nouveau_hwmon_show_temp,
 						  NULL, 0);
 
 static ssize_t
+nouveau_hwmon_show_temp1_auto_point1_pwm(struct device *d,
+					 struct device_attribute *a, char *buf)
+{
+	return snprintf(buf, PAGE_SIZE, "%d\n", 100);
+}
+static SENSOR_DEVICE_ATTR(temp1_auto_point1_pwm, S_IRUGO,
+			  nouveau_hwmon_show_temp1_auto_point1_pwm, NULL, 0);
+
+static ssize_t
+nouveau_hwmon_temp1_auto_point1_temp(struct device *d,
+				     struct device_attribute *a, char *buf)
+{
+	struct drm_device *dev = dev_get_drvdata(d);
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nouveau_therm *therm = nouveau_therm(drm->device);
+
+	return snprintf(buf, PAGE_SIZE, "%d\n",
+	      therm->attr_get(therm, NOUVEAU_THERM_ATTR_THRS_FAN_BOOST) * 1000);
+}
+static ssize_t
+nouveau_hwmon_set_temp1_auto_point1_temp(struct device *d,
+					 struct device_attribute *a,
+					 const char *buf, size_t count)
+{
+	struct drm_device *dev = dev_get_drvdata(d);
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	long value;
+
+	if (kstrtol(buf, 10, &value) == -EINVAL)
+		return count;
+
+	therm->attr_set(therm, NOUVEAU_THERM_ATTR_THRS_FAN_BOOST,
+			value / 1000);
+
+	return count;
+}
+static SENSOR_DEVICE_ATTR(temp1_auto_point1_temp, S_IRUGO | S_IWUSR,
+			  nouveau_hwmon_temp1_auto_point1_temp,
+			  nouveau_hwmon_set_temp1_auto_point1_temp, 0);
+
+static ssize_t
+nouveau_hwmon_temp1_auto_point1_temp_hyst(struct device *d,
+					  struct device_attribute *a, char *buf)
+{
+	struct drm_device *dev = dev_get_drvdata(d);
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nouveau_therm *therm = nouveau_therm(drm->device);
+
+	return snprintf(buf, PAGE_SIZE, "%d\n",
+	 therm->attr_get(therm, NOUVEAU_THERM_ATTR_THRS_FAN_BOOST_HYST) * 1000);
+}
+static ssize_t
+nouveau_hwmon_set_temp1_auto_point1_temp_hyst(struct device *d,
+					      struct device_attribute *a,
+					      const char *buf, size_t count)
+{
+	struct drm_device *dev = dev_get_drvdata(d);
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	long value;
+
+	if (kstrtol(buf, 10, &value) == -EINVAL)
+		return count;
+
+	therm->attr_set(therm, NOUVEAU_THERM_ATTR_THRS_FAN_BOOST_HYST,
+			value / 1000);
+
+	return count;
+}
+static SENSOR_DEVICE_ATTR(temp1_auto_point1_temp_hyst, S_IRUGO | S_IWUSR,
+			  nouveau_hwmon_temp1_auto_point1_temp_hyst,
+			  nouveau_hwmon_set_temp1_auto_point1_temp_hyst, 0);
+
+static ssize_t
 nouveau_hwmon_max_temp(struct device *d, struct device_attribute *a, char *buf)
 {
 	struct drm_device *dev = dev_get_drvdata(d);
@@ -439,6 +514,38 @@ static SENSOR_DEVICE_ATTR(temp1_max, S_IRUGO | S_IWUSR, nouveau_hwmon_max_temp,
 						  0);
 
 static ssize_t
+nouveau_hwmon_max_temp_hyst(struct device *d, struct device_attribute *a,
+			    char *buf)
+{
+	struct drm_device *dev = dev_get_drvdata(d);
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nouveau_therm *therm = nouveau_therm(drm->device);
+
+	return snprintf(buf, PAGE_SIZE, "%d\n",
+	  therm->attr_get(therm, NOUVEAU_THERM_ATTR_THRS_DOWN_CLK_HYST) * 1000);
+}
+static ssize_t
+nouveau_hwmon_set_max_temp_hyst(struct device *d, struct device_attribute *a,
+						const char *buf, size_t count)
+{
+	struct drm_device *dev = dev_get_drvdata(d);
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	long value;
+
+	if (kstrtol(buf, 10, &value) == -EINVAL)
+		return count;
+
+	therm->attr_set(therm, NOUVEAU_THERM_ATTR_THRS_DOWN_CLK_HYST,
+			value / 1000);
+
+	return count;
+}
+static SENSOR_DEVICE_ATTR(temp1_max_hyst, S_IRUGO | S_IWUSR,
+			  nouveau_hwmon_max_temp_hyst,
+			  nouveau_hwmon_set_max_temp_hyst, 0);
+
+static ssize_t
 nouveau_hwmon_critical_temp(struct device *d, struct device_attribute *a,
 							char *buf)
 {
@@ -471,6 +578,107 @@ static SENSOR_DEVICE_ATTR(temp1_crit, S_IRUGO | S_IWUSR,
 						nouveau_hwmon_set_critical_temp,
 						0);
 
+static ssize_t
+nouveau_hwmon_critical_temp_hyst(struct device *d, struct device_attribute *a,
+							char *buf)
+{
+	struct drm_device *dev = dev_get_drvdata(d);
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nouveau_therm *therm = nouveau_therm(drm->device);
+
+	return snprintf(buf, PAGE_SIZE, "%d\n",
+	  therm->attr_get(therm, NOUVEAU_THERM_ATTR_THRS_CRITICAL_HYST) * 1000);
+}
+static ssize_t
+nouveau_hwmon_set_critical_temp_hyst(struct device *d,
+				     struct device_attribute *a,
+				     const char *buf,
+				     size_t count)
+{
+	struct drm_device *dev = dev_get_drvdata(d);
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	long value;
+
+	if (kstrtol(buf, 10, &value) == -EINVAL)
+		return count;
+
+	therm->attr_set(therm, NOUVEAU_THERM_ATTR_THRS_CRITICAL_HYST,
+			value / 1000);
+
+	return count;
+}
+static SENSOR_DEVICE_ATTR(temp1_crit_hyst, S_IRUGO | S_IWUSR,
+			  nouveau_hwmon_critical_temp_hyst,
+			  nouveau_hwmon_set_critical_temp_hyst, 0);
+static ssize_t
+nouveau_hwmon_emergency_temp(struct device *d, struct device_attribute *a,
+							char *buf)
+{
+	struct drm_device *dev = dev_get_drvdata(d);
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nouveau_therm *therm = nouveau_therm(drm->device);
+
+	return snprintf(buf, PAGE_SIZE, "%d\n",
+	       therm->attr_get(therm, NOUVEAU_THERM_ATTR_THRS_SHUTDOWN) * 1000);
+}
+static ssize_t
+nouveau_hwmon_set_emergency_temp(struct device *d, struct device_attribute *a,
+							    const char *buf,
+								size_t count)
+{
+	struct drm_device *dev = dev_get_drvdata(d);
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	long value;
+
+	if (kstrtol(buf, 10, &value) == -EINVAL)
+		return count;
+
+	therm->attr_set(therm, NOUVEAU_THERM_ATTR_THRS_SHUTDOWN, value / 1000);
+
+	return count;
+}
+static SENSOR_DEVICE_ATTR(temp1_emergency, S_IRUGO | S_IWUSR,
+					nouveau_hwmon_emergency_temp,
+					nouveau_hwmon_set_emergency_temp,
+					0);
+
+static ssize_t
+nouveau_hwmon_emergency_temp_hyst(struct device *d, struct device_attribute *a,
+							char *buf)
+{
+	struct drm_device *dev = dev_get_drvdata(d);
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nouveau_therm *therm = nouveau_therm(drm->device);
+
+	return snprintf(buf, PAGE_SIZE, "%d\n",
+	  therm->attr_get(therm, NOUVEAU_THERM_ATTR_THRS_SHUTDOWN_HYST) * 1000);
+}
+static ssize_t
+nouveau_hwmon_set_emergency_temp_hyst(struct device *d,
+				      struct device_attribute *a,
+				      const char *buf,
+				      size_t count)
+{
+	struct drm_device *dev = dev_get_drvdata(d);
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	long value;
+
+	if (kstrtol(buf, 10, &value) == -EINVAL)
+		return count;
+
+	therm->attr_set(therm, NOUVEAU_THERM_ATTR_THRS_SHUTDOWN_HYST,
+			value / 1000);
+
+	return count;
+}
+static SENSOR_DEVICE_ATTR(temp1_emergency_hyst, S_IRUGO | S_IWUSR,
+					nouveau_hwmon_emergency_temp_hyst,
+					nouveau_hwmon_set_emergency_temp_hyst,
+					0);
+
 static ssize_t nouveau_hwmon_show_name(struct device *dev,
 				      struct device_attribute *attr,
 				      char *buf)
@@ -665,8 +873,15 @@ static SENSOR_DEVICE_ATTR(pwm1_max, S_IRUGO | S_IWUSR,
 
 static struct attribute *hwmon_attributes[] = {
 	&sensor_dev_attr_temp1_input.dev_attr.attr,
+	&sensor_dev_attr_temp1_auto_point1_pwm.dev_attr.attr,
+	&sensor_dev_attr_temp1_auto_point1_temp.dev_attr.attr,
+	&sensor_dev_attr_temp1_auto_point1_temp_hyst.dev_attr.attr,
 	&sensor_dev_attr_temp1_max.dev_attr.attr,
+	&sensor_dev_attr_temp1_max_hyst.dev_attr.attr,
 	&sensor_dev_attr_temp1_crit.dev_attr.attr,
+	&sensor_dev_attr_temp1_crit_hyst.dev_attr.attr,
+	&sensor_dev_attr_temp1_emergency.dev_attr.attr,
+	&sensor_dev_attr_temp1_emergency_hyst.dev_attr.attr,
 	&sensor_dev_attr_name.dev_attr.attr,
 	&sensor_dev_attr_update_rate.dev_attr.attr,
 	NULL
-- 
cgit v0.10.2


From 041d62469f1efff280bbb9b2239e4e246da6e0d1 Mon Sep 17 00:00:00 2001
From: Martin Peres <martin.peres@labri.fr>
Date: Tue, 20 Nov 2012 02:29:22 +0100
Subject: drm/nouveau/doc: document the sysfs thermal management interface

Signed-off-by: Martin Peres <martin.peres@labri.fr>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/Documentation/thermal/nouveau_thermal b/Documentation/thermal/nouveau_thermal
new file mode 100644
index 0000000..ec854fd
--- /dev/null
+++ b/Documentation/thermal/nouveau_thermal
@@ -0,0 +1,81 @@
+Kernel driver nouveau
+===================
+
+Supported chips:
+* NV43+
+
+Authors: Martin Peres (mupuf) <martin.peres@labri.fr>
+
+Description
+---------
+
+This driver allows to read the GPU core temperature, drive the GPU fan and
+set temperature alarms.
+
+Currently, due to the absence of in-kernel API to access HWMON drivers, Nouveau
+cannot access any of the i2c external monitoring chips it may find. If you
+have one of those, temperature and/or fan management through Nouveau's HWMON
+interface is likely not to work. This document may then not cover your situation
+entirely.
+
+Temperature management
+--------------------
+
+Temperature is exposed under as a read-only HWMON attribute temp1_input.
+
+In order to protect the GPU from overheating, Nouveau supports 4 configurable
+temperature thresholds:
+
+ * Fan_boost: Fan speed is set to 100% when reaching this temperature;
+ * Downclock: The GPU will be downclocked to reduce its power dissipation;
+ * Critical: The GPU is put on hold to further lower power dissipation;
+ * Shutdown: Shut the computer down to protect your GPU.
+
+WARNING: Some of these thresholds may not be used by Nouveau depending
+on your chipset.
+
+The default value for these thresholds comes from the GPU's vbios. These
+thresholds can be configured thanks to the following HWMON attributes:
+
+ * Fan_boost: temp1_auto_point1_temp and temp1_auto_point1_temp_hyst;
+ * Downclock: temp1_max and temp1_max_hyst;
+ * Critical: temp1_crit and temp1_crit_hyst;
+ * Shutdown: temp1_emergency and temp1_emergency_hyst.
+
+NOTE: Remember that the values are stored as milli degrees Celcius. Don't forget
+to multiply!
+
+Fan management
+------------
+
+Not all cards have a drivable fan. If you do, then the following HWMON
+attributes should be available:
+
+ * pwm1_enable: Current fan management mode (NONE, MANUAL or AUTO);
+ * pwm1: Current PWM value (power percentage);
+ * pwm1_min: The minimum PWM speed allowed;
+ * pwm1_max: The maximum PWM speed allowed (bypassed when hitting Fan_boost);
+
+You may also have the following attribute:
+
+ * fan0_input: Speed in RPM of your fan.
+
+Your fan can be driven in different modes:
+
+ * 0: The fan is left untouched;
+ * 1: The fan can be driven in manual (use pwm1 to change the speed);
+ * 2; The fan is driven automatically depending on the temperature.
+
+NOTE: Be sure to use the manual mode if you want to drive the fan speed manually
+
+NOTE2: Not all fan management modes may be supported on all chipsets. We are
+working on it.
+
+Bug reports
+---------
+
+Thermal management on Nouveau is new and may not work on all cards. If you have
+inquiries, please ping mupuf on IRC (#nouveau, freenode).
+
+Bug reports should be filled on Freedesktop's bug tracker. Please follow
+http://nouveau.freedesktop.org/wiki/Bugs
-- 
cgit v0.10.2


From 1a22274b28fd47514f1463051d93028ad5f54af6 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Thu, 6 Dec 2012 10:28:34 +1000
Subject: drm/nouveau/therm: better transitions and debug logging

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Martin Peres <martin.peres@labri.fr>

diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/therm.h b/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
index 7c83ecc..6b17b61 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
@@ -4,11 +4,10 @@
 #include <core/device.h>
 #include <core/subdev.h>
 
-enum nouveau_therm_fan_mode {
-	FAN_CONTROL_NONE = 0,
-	FAN_CONTROL_MANUAL = 1,
-	FAN_CONTROL_AUTO = 2,
-	FAN_CONTROL_NR,
+enum nouveau_therm_mode {
+	NOUVEAU_THERM_CTRL_NONE = 0,
+	NOUVEAU_THERM_CTRL_MANUAL = 1,
+	NOUVEAU_THERM_CTRL_AUTO = 2,
 };
 
 enum nouveau_therm_attr_type {
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/base.c b/drivers/gpu/drm/nouveau/core/subdev/therm/base.c
index f54495a..b35b4a2 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/base.c
@@ -93,24 +93,27 @@ nouveau_therm_update(struct nouveau_therm *therm, int mode)
 	priv->mode = mode;
 
 	switch (mode) {
-	case FAN_CONTROL_MANUAL:
-		duty = priv->fan->percent;
+	case NOUVEAU_THERM_CTRL_MANUAL:
+		duty = nouveau_therm_fan_get(therm);
+		if (duty < 0)
+			duty = 100;
 		break;
-	case FAN_CONTROL_AUTO:
+	case NOUVEAU_THERM_CTRL_AUTO:
 		if (priv->fan->bios.nr_fan_trip)
 			duty = nouveau_therm_update_trip(therm);
 		else
 			duty = nouveau_therm_update_linear(therm);
 		break;
-	case FAN_CONTROL_NONE:
+	case NOUVEAU_THERM_CTRL_NONE:
 	default:
 		goto done;
 	}
 
-	nouveau_therm_fan_set(therm, (mode != FAN_CONTROL_AUTO), duty);
+	nv_debug(therm, "FAN target request: %d%%\n", duty);
+	nouveau_therm_fan_set(therm, (mode != NOUVEAU_THERM_CTRL_AUTO), duty);
 
 done:
-	if (list_empty(&priv->alarm.head) && (mode == FAN_CONTROL_AUTO))
+	if (list_empty(&priv->alarm.head) && (mode == NOUVEAU_THERM_CTRL_AUTO))
 		ptimer->alarm(ptimer, 1000000000ULL, &priv->alarm);
 	spin_unlock_irqrestore(&priv->lock, flags);
 }
@@ -127,28 +130,22 @@ int
 nouveau_therm_mode(struct nouveau_therm *therm, int mode)
 {
 	struct nouveau_therm_priv *priv = (void *)therm;
-
-	if (priv->mode == mode)
-		return 0;
+	struct nouveau_device *device = nv_device(therm);
+	static const char *name[] = {
+		"disabled",
+		"manual",
+		"automatic"
+	};
 
 	/* The default PDAEMON ucode interferes with fan management */
-	if (nv_device(therm)->card_type >= NV_C0)
+	if ((mode >= ARRAY_SIZE(name)) ||
+	    (mode != NOUVEAU_THERM_CTRL_NONE && device->card_type >= NV_C0))
 		return -EINVAL;
 
-	switch (mode) {
-	case FAN_CONTROL_NONE:
-		nv_info(therm, "switch to no-control mode\n");
-		break;
-	case FAN_CONTROL_MANUAL:
-		nv_info(therm, "switch to manual mode\n");
-		break;
-	case FAN_CONTROL_AUTO:
-		nv_info(therm, "switch to automatic mode\n");
-		break;
-	default:
-		return -EINVAL;
-	}
+	if (priv->mode == mode)
+		return 0;
 
+	nv_info(therm, "Thermal management: %s\n", name[mode]);
 	nouveau_therm_update(therm, mode);
 	return 0;
 }
@@ -258,9 +255,8 @@ _nouveau_therm_init(struct nouveau_object *object)
 	if (ret)
 		return ret;
 
-	if (priv->fan->percent >= 0)
-		therm->fan_set(therm, priv->fan->percent);
-
+	if (priv->suspend >= 0)
+		nouveau_therm_mode(therm, priv->mode);
 	priv->sensor.program_alarms(therm);
 	return 0;
 }
@@ -271,7 +267,10 @@ _nouveau_therm_fini(struct nouveau_object *object, bool suspend)
 	struct nouveau_therm *therm = (void *)object;
 	struct nouveau_therm_priv *priv = (void *)therm;
 
-	priv->fan->percent = therm->fan_get(therm);
+	if (suspend) {
+		priv->suspend = priv->mode;
+		priv->mode = NOUVEAU_THERM_CTRL_NONE;
+	}
 
 	return nouveau_subdev_fini(&therm->base, suspend);
 }
@@ -299,6 +298,7 @@ nouveau_therm_create_(struct nouveau_object *parent,
 	priv->base.fan_sense = nouveau_therm_fan_sense;
 	priv->base.attr_get = nouveau_therm_attr_get;
 	priv->base.attr_set = nouveau_therm_attr_set;
+	priv->mode = priv->suspend = -1; /* undefined */
 	return 0;
 }
 
@@ -308,6 +308,8 @@ nouveau_therm_preinit(struct nouveau_therm *therm)
 	nouveau_therm_ic_ctor(therm);
 	nouveau_therm_sensor_ctor(therm);
 	nouveau_therm_fan_ctor(therm);
+
+	nouveau_therm_mode(therm, NOUVEAU_THERM_CTRL_NONE);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
index 0f210f0..f7d32f4 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
@@ -47,10 +47,17 @@ nouveau_fan_update(struct nouveau_fan *fan, bool immediate, int target)
 		target = fan->percent;
 	target = max_t(u8, target, fan->bios.min_duty);
 	target = min_t(u8, target, fan->bios.max_duty);
-	fan->percent = target;
+	if (fan->percent != target) {
+		nv_debug(therm, "FAN target: %d\n", target);
+		fan->percent = target;
+	}
 
-	/* smooth out the fanspeed increase/decrease */
+	/* check that we're not already at the target duty cycle */
 	duty = fan->get(therm);
+	if (duty == target)
+		goto done;
+
+	/* smooth out the fanspeed increase/decrease */
 	if (!immediate && duty >= 0) {
 		/* the constant "3" is a rough approximation taken from
 		 * nvidia's behaviour.
@@ -64,6 +71,7 @@ nouveau_fan_update(struct nouveau_fan *fan, bool immediate, int target)
 		duty = target;
 	}
 
+	nv_debug(therm, "FAN update: %d\n", duty);
 	ret = fan->set(therm, duty);
 	if (ret)
 		goto done;
@@ -161,7 +169,7 @@ nouveau_therm_fan_user_set(struct nouveau_therm *therm, int percent)
 {
 	struct nouveau_therm_priv *priv = (void *)therm;
 
-	if (priv->mode != FAN_CONTROL_MANUAL)
+	if (priv->mode != NOUVEAU_THERM_CTRL_MANUAL)
 		return -EINVAL;
 
 	return nouveau_therm_fan_set(therm, true, percent);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
index c2413e6..06b9870 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
@@ -76,6 +76,7 @@ struct nouveau_therm_priv {
 	spinlock_t lock;
 	struct nouveau_therm_trip_point *last_trip;
 	int mode;
+	int suspend;
 
 	/* bios */
 	struct nvbios_therm_sensor bios_sensor;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/temp.c b/drivers/gpu/drm/nouveau/core/subdev/therm/temp.c
index d80c572..820e097 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/temp.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/temp.c
@@ -108,7 +108,7 @@ void nouveau_therm_sensor_event(struct nouveau_therm *therm,
 	switch (thrs) {
 	case NOUVEAU_THERM_THRS_FANBOOST:
 		nouveau_therm_fan_set(therm, true, 100);
-		nouveau_therm_mode(therm, FAN_CONTROL_AUTO);
+		nouveau_therm_mode(therm, NOUVEAU_THERM_CTRL_AUTO);
 		break;
 	case NOUVEAU_THERM_THRS_DOWNCLOCK:
 		if (priv->emergency.downclock)
-- 
cgit v0.10.2


From b2c36312c9532491ca38275c3862e5c5dfac6033 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Thu, 6 Dec 2012 15:13:06 +1000
Subject: drm/nouveau/hwmon: s/fan0/fan1/

Fan speed info now shown by sensors.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/Documentation/thermal/nouveau_thermal b/Documentation/thermal/nouveau_thermal
index ec854fd..efceb78 100644
--- a/Documentation/thermal/nouveau_thermal
+++ b/Documentation/thermal/nouveau_thermal
@@ -58,7 +58,7 @@ attributes should be available:
 
 You may also have the following attribute:
 
- * fan0_input: Speed in RPM of your fan.
+ * fan1_input: Speed in RPM of your fan.
 
 Your fan can be driven in different modes:
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_pm.c b/drivers/gpu/drm/nouveau/nouveau_pm.c
index d456e10..3a6ee7e 100644
--- a/drivers/gpu/drm/nouveau/nouveau_pm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_pm.c
@@ -698,7 +698,7 @@ static SENSOR_DEVICE_ATTR(update_rate, S_IRUGO,
 						NULL, 0);
 
 static ssize_t
-nouveau_hwmon_show_fan0_input(struct device *d, struct device_attribute *attr,
+nouveau_hwmon_show_fan1_input(struct device *d, struct device_attribute *attr,
 			      char *buf)
 {
 	struct drm_device *dev = dev_get_drvdata(d);
@@ -707,7 +707,7 @@ nouveau_hwmon_show_fan0_input(struct device *d, struct device_attribute *attr,
 
 	return snprintf(buf, PAGE_SIZE, "%d\n", therm->fan_sense(therm));
 }
-static SENSOR_DEVICE_ATTR(fan0_input, S_IRUGO, nouveau_hwmon_show_fan0_input,
+static SENSOR_DEVICE_ATTR(fan1_input, S_IRUGO, nouveau_hwmon_show_fan1_input,
 			  NULL, 0);
 
  static ssize_t
@@ -887,7 +887,7 @@ static struct attribute *hwmon_attributes[] = {
 	NULL
 };
 static struct attribute *hwmon_fan_rpm_attributes[] = {
-	&sensor_dev_attr_fan0_input.dev_attr.attr,
+	&sensor_dev_attr_fan1_input.dev_attr.attr,
 	NULL
 };
 static struct attribute *hwmon_pwm_fan_attributes[] = {
-- 
cgit v0.10.2


From 5e90a88cb37b665cc0609f1236d6099421139449 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Thu, 6 Dec 2012 15:17:12 +1000
Subject: drm/nouveau/hwmon: create hwmon attributes under hwmon device in
 sysfs

From browsing my /sys, a few other things seem to do this, and it looks
cleaner this way too :)

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nouveau_pm.c b/drivers/gpu/drm/nouveau/nouveau_pm.c
index 3a6ee7e..bb54098 100644
--- a/drivers/gpu/drm/nouveau/nouveau_pm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_pm.c
@@ -932,7 +932,7 @@ nouveau_hwmon_init(struct drm_device *dev)
 	dev_set_drvdata(hwmon_dev, dev);
 
 	/* default sysfs entries */
-	ret = sysfs_create_group(&dev->pdev->dev.kobj, &hwmon_attrgroup);
+	ret = sysfs_create_group(&hwmon_dev->kobj, &hwmon_attrgroup);
 	if (ret) {
 		if (ret)
 			goto error;
@@ -943,7 +943,7 @@ nouveau_hwmon_init(struct drm_device *dev)
 	 *     the gpio entries for pwm fan control even when there's no
 	 *     actual fan connected to it... therm table? */
 	if (therm->fan_get && therm->fan_get(therm) >= 0) {
-		ret = sysfs_create_group(&dev->pdev->dev.kobj,
+		ret = sysfs_create_group(&hwmon_dev->kobj,
 					 &hwmon_pwm_fan_attrgroup);
 		if (ret)
 			goto error;
@@ -951,7 +951,7 @@ nouveau_hwmon_init(struct drm_device *dev)
 
 	/* if the card can read the fan rpm */
 	if (therm->fan_sense(therm) >= 0) {
-		ret = sysfs_create_group(&dev->pdev->dev.kobj,
+		ret = sysfs_create_group(&hwmon_dev->kobj,
 					 &hwmon_fan_rpm_attrgroup);
 		if (ret)
 			goto error;
@@ -979,10 +979,10 @@ nouveau_hwmon_fini(struct drm_device *dev)
 	struct nouveau_pm *pm = nouveau_pm(dev);
 
 	if (pm->hwmon) {
-		sysfs_remove_group(&dev->pdev->dev.kobj, &hwmon_attrgroup);
-		sysfs_remove_group(&dev->pdev->dev.kobj,
+		sysfs_remove_group(&pm->hwmon->kobj, &hwmon_attrgroup);
+		sysfs_remove_group(&pm->hwmon->kobj,
 				   &hwmon_pwm_fan_attrgroup);
-		sysfs_remove_group(&dev->pdev->dev.kobj,
+		sysfs_remove_group(&pm->hwmon->kobj,
 				   &hwmon_fan_rpm_attrgroup);
 
 		hwmon_device_unregister(pm->hwmon);
-- 
cgit v0.10.2


From e84e4b67ddd6fcbfe5ab39c7aed1d490a8c6283c Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Thu, 6 Dec 2012 15:32:56 +1000
Subject: drm/nouveau: remove legacy vbios type detection

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.c b/drivers/gpu/drm/nouveau/nouveau_bios.c
index 865eddf..9a69e0b 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bios.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bios.c
@@ -2052,45 +2052,27 @@ uint8_t *nouveau_bios_embedded_edid(struct drm_device *dev)
 static bool NVInitVBIOS(struct drm_device *dev)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nvbios *bios = &drm->vbios;
-
-	memset(bios, 0, sizeof(struct nvbios));
-	spin_lock_init(&bios->lock);
-	bios->dev = dev;
-
-	bios->data = nouveau_bios(drm->device)->data;
-	bios->length = nouveau_bios(drm->device)->size;
-	return true;
-}
+	struct nouveau_bios *bios = nouveau_bios(drm->device);
+	struct nvbios *legacy = &drm->vbios;
+
+	memset(legacy, 0, sizeof(struct nvbios));
+	spin_lock_init(&legacy->lock);
+	legacy->dev = dev;
+
+	legacy->data = bios->data;
+	legacy->length = bios->size;
+	if (bios->bit_offset) {
+		legacy->type = NVBIOS_BIT;
+		legacy->offset = bios->bit_offset;
+		return !parse_bit_structure(legacy, legacy->offset + 6);
+	} else
+	if (bios->bmp_offset) {
+		legacy->type = NVBIOS_BMP;
+		legacy->offset = bios->bmp_offset;
+		return !parse_bmp_structure(dev, legacy, legacy->offset);
+	}
 
-static int nouveau_parse_vbios_struct(struct drm_device *dev)
-{
-	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nvbios *bios = &drm->vbios;
-	const uint8_t bit_signature[] = { 0xff, 0xb8, 'B', 'I', 'T' };
-	const uint8_t bmp_signature[] = { 0xff, 0x7f, 'N', 'V', 0x0 };
-	int offset;
-
-	offset = findstr(bios->data, bios->length,
-					bit_signature, sizeof(bit_signature));
-	if (offset) {
-		NV_INFO(drm, "BIT BIOS found\n");
-		bios->type = NVBIOS_BIT;
-		bios->offset = offset;
-		return parse_bit_structure(bios, offset + 6);
-	}
-
-	offset = findstr(bios->data, bios->length,
-					bmp_signature, sizeof(bmp_signature));
-	if (offset) {
-		NV_INFO(drm, "BMP BIOS found\n");
-		bios->type = NVBIOS_BMP;
-		bios->offset = offset;
-		return parse_bmp_structure(dev, bios, offset);
-	}
-
-	NV_ERROR(drm, "No known BIOS signature found\n");
-	return -ENODEV;
+	return false;
 }
 
 int
@@ -2146,10 +2128,6 @@ nouveau_bios_init(struct drm_device *dev)
 	if (!NVInitVBIOS(dev))
 		return -ENODEV;
 
-	ret = nouveau_parse_vbios_struct(dev);
-	if (ret)
-		return ret;
-
 	ret = parse_dcb_table(dev, bios);
 	if (ret)
 		return ret;
-- 
cgit v0.10.2


From 230881821a4467727f4100b5ee5c3ef0b67f3128 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Thu, 6 Dec 2012 15:44:23 +1000
Subject: drm/nouveau: remove some more unnecessary legacy bios code

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.c b/drivers/gpu/drm/nouveau/nouveau_bios.c
index 9a69e0b..5e1cfca 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bios.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bios.c
@@ -678,23 +678,6 @@ int run_tmds_table(struct drm_device *dev, struct dcb_output *dcbent, int head,
 	return 0;
 }
 
-static void parse_bios_version(struct drm_device *dev, struct nvbios *bios, uint16_t offset)
-{
-	/*
-	 * offset + 0  (8 bits): Micro version
-	 * offset + 1  (8 bits): Minor version
-	 * offset + 2  (8 bits): Chip version
-	 * offset + 3  (8 bits): Major version
-	 */
-	struct nouveau_drm *drm = nouveau_drm(dev);
-
-	bios->major_version = bios->data[offset + 3];
-	bios->chip_version = bios->data[offset + 2];
-	NV_INFO(drm, "Bios version %02x.%02x.%02x.%02x\n",
-		 bios->data[offset + 3], bios->data[offset + 2],
-		 bios->data[offset + 1], bios->data[offset]);
-}
-
 static void parse_script_table_pointers(struct nvbios *bios, uint16_t offset)
 {
 	/*
@@ -710,12 +693,6 @@ static void parse_script_table_pointers(struct nvbios *bios, uint16_t offset)
 	 */
 
 	bios->init_script_tbls_ptr = ROM16(bios->data[offset]);
-	bios->macro_index_tbl_ptr = ROM16(bios->data[offset + 2]);
-	bios->macro_tbl_ptr = ROM16(bios->data[offset + 4]);
-	bios->condition_tbl_ptr = ROM16(bios->data[offset + 6]);
-	bios->io_condition_tbl_ptr = ROM16(bios->data[offset + 8]);
-	bios->io_flag_condition_tbl_ptr = ROM16(bios->data[offset + 10]);
-	bios->init_function_tbl_ptr = ROM16(bios->data[offset + 12]);
 }
 
 static int parse_bit_A_tbl_entry(struct drm_device *dev, struct nvbios *bios, struct bit_entry *bitentry)
@@ -765,25 +742,6 @@ static int parse_bit_A_tbl_entry(struct drm_device *dev, struct nvbios *bios, st
 	return 0;
 }
 
-static int parse_bit_C_tbl_entry(struct drm_device *dev, struct nvbios *bios, struct bit_entry *bitentry)
-{
-	/*
-	 * offset + 8  (16 bits): PLL limits table pointer
-	 *
-	 * There's more in here, but that's unknown.
-	 */
-	struct nouveau_drm *drm = nouveau_drm(dev);
-
-	if (bitentry->length < 10) {
-		NV_ERROR(drm, "Do not understand BIT C table\n");
-		return -EINVAL;
-	}
-
-	bios->pll_limit_tbl_ptr = ROM16(bios->data[bitentry->offset + 8]);
-
-	return 0;
-}
-
 static int parse_bit_display_tbl_entry(struct drm_device *dev, struct nvbios *bios, struct bit_entry *bitentry)
 {
 	/*
@@ -821,12 +779,6 @@ static int parse_bit_init_tbl_entry(struct drm_device *dev, struct nvbios *bios,
 	}
 
 	parse_script_table_pointers(bios, bitentry->offset);
-
-	if (bitentry->length >= 16)
-		bios->some_script_ptr = ROM16(bios->data[bitentry->offset + 14]);
-	if (bitentry->length >= 18)
-		bios->init96_tbl_ptr = ROM16(bios->data[bitentry->offset + 16]);
-
 	return 0;
 }
 
@@ -852,8 +804,6 @@ static int parse_bit_i_tbl_entry(struct drm_device *dev, struct nvbios *bios, st
 		return -EINVAL;
 	}
 
-	parse_bios_version(dev, bios, bitentry->offset);
-
 	/*
 	 * bit 4 seems to indicate a mobile bios (doesn't suffer from BMP's
 	 * Quadro identity crisis), other bits possibly as for BMP feature byte
@@ -1078,9 +1028,6 @@ parse_bit_structure(struct nvbios *bios, const uint16_t bitoffset)
 		return ret;
 	if (bios->major_version >= 0x60) /* g80+ */
 		parse_bit_table(bios, bitoffset, &BIT_TABLE('A', A));
-	ret = parse_bit_table(bios, bitoffset, &BIT_TABLE('C', C));
-	if (ret)
-		return ret;
 	parse_bit_table(bios, bitoffset, &BIT_TABLE('D', display));
 	ret = parse_bit_table(bios, bitoffset, &BIT_TABLE('I', init));
 	if (ret)
@@ -1228,8 +1175,6 @@ static int parse_bmp_structure(struct drm_device *dev, struct nvbios *bios, unsi
 	 */
 	bios->feature_byte = bmp[9];
 
-	parse_bios_version(dev, bios, offset + 10);
-
 	if (bmp_version_major < 5 || bmp_version_minor < 0x10)
 		bios->old_style_init = true;
 	legacy_scripts_offset = 18;
@@ -1276,8 +1221,10 @@ static int parse_bmp_structure(struct drm_device *dev, struct nvbios *bios, unsi
 		bios->fp.lvdsmanufacturerpointer = ROM16(bmp[117]);
 		bios->fp.fpxlatemanufacturertableptr = ROM16(bmp[119]);
 	}
+#if 0
 	if (bmplength > 143)
 		bios->pll_limit_tbl_ptr = ROM16(bmp[142]);
+#endif
 
 	if (bmplength > 157)
 		bios->fp.duallink_transition_clk = ROM16(bmp[156]) * 10;
@@ -2072,6 +2019,8 @@ static bool NVInitVBIOS(struct drm_device *dev)
 		return !parse_bmp_structure(dev, legacy, legacy->offset);
 	}
 
+	legacy->major_version = bios->version.major;
+	legacy->chip_version = bios->version.chip;
 	return false;
 }
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.h b/drivers/gpu/drm/nouveau/nouveau_bios.h
index f68c54c..7ccd28f 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bios.h
+++ b/drivers/gpu/drm/nouveau/nouveau_bios.h
@@ -107,20 +107,10 @@ struct nvbios {
 	bool old_style_init;
 	uint16_t init_script_tbls_ptr;
 	uint16_t extra_init_script_tbl_ptr;
-	uint16_t macro_index_tbl_ptr;
-	uint16_t macro_tbl_ptr;
-	uint16_t condition_tbl_ptr;
-	uint16_t io_condition_tbl_ptr;
-	uint16_t io_flag_condition_tbl_ptr;
-	uint16_t init_function_tbl_ptr;
-
-	uint16_t pll_limit_tbl_ptr;
+
 	uint16_t ram_restrict_tbl_ptr;
 	uint8_t ram_restrict_group_count;
 
-	uint16_t some_script_ptr; /* BIT I + 14 */
-	uint16_t init96_tbl_ptr; /* BIT I + 16 */
-
 	struct dcb_table dcb;
 
 	struct {
-- 
cgit v0.10.2


From 112a12aab8a5c1137a140c341f98a160e27b8c49 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Fri, 7 Dec 2012 10:31:09 +1000
Subject: drm/nouveau/bios: add support for parsing xpio table data

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index ad586f0..6985740 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -40,6 +40,7 @@ nouveau-y += core/subdev/bios/mxm.o
 nouveau-y += core/subdev/bios/perf.o
 nouveau-y += core/subdev/bios/pll.o
 nouveau-y += core/subdev/bios/therm.o
+nouveau-y += core/subdev/bios/xpio.o
 nouveau-y += core/subdev/bus/nv04.o
 nouveau-y += core/subdev/bus/nv31.o
 nouveau-y += core/subdev/bus/nv50.o
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/gpio.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/gpio.h
index e6563b5..94e540e 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/bios/gpio.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/gpio.h
@@ -1,8 +1,6 @@
 #ifndef __NVBIOS_GPIO_H__
 #define __NVBIOS_GPIO_H__
 
-struct nouveau_bios;
-
 enum dcb_gpio_func_name {
 	DCB_GPIO_PANEL_POWER = 0x01,
 	DCB_GPIO_TVDAC0 = 0x0c,
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/xpio.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/xpio.h
new file mode 100644
index 0000000..360baab
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/xpio.h
@@ -0,0 +1,19 @@
+#ifndef __NVBIOS_XPIO_H__
+#define __NVBIOS_XPIO_H__
+
+#define NVBIOS_XPIO_FLAG_AUX  0x10
+#define NVBIOS_XPIO_FLAG_AUX0 0x00
+#define NVBIOS_XPIO_FLAG_AUX1 0x10
+
+struct nvbios_xpio {
+	u8 type;
+	u8 addr;
+	u8 flags;
+};
+
+u16 dcb_xpio_table(struct nouveau_bios *, u8 idx,
+		   u8 *ver, u8 *hdr, u8 *cnt, u8 *len);
+u16 dcb_xpio_parse(struct nouveau_bios *, u8 idx,
+		   u8 *ver, u8 *hdr, u8 *cnt, u8 *len, struct nvbios_xpio *);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/gpio.c b/drivers/gpu/drm/nouveau/core/subdev/bios/gpio.c
index c84e93f..172a4f9 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/gpio.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/gpio.c
@@ -25,6 +25,7 @@
 #include <subdev/bios.h>
 #include <subdev/bios/dcb.h>
 #include <subdev/bios/gpio.h>
+#include <subdev/bios/xpio.h>
 
 u16
 dcb_gpio_table(struct nouveau_bios *bios, u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
@@ -60,8 +61,14 @@ dcb_gpio_table(struct nouveau_bios *bios, u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
 u16
 dcb_gpio_entry(struct nouveau_bios *bios, int idx, int ent, u8 *ver, u8 *len)
 {
-	u8  hdr, cnt;
-	u16 gpio = !idx ? dcb_gpio_table(bios, ver, &hdr, &cnt, len) : 0x0000;
+	u8  hdr, cnt, xver; /* use gpio version for xpio entry parsing */
+	u16 gpio;
+
+	if (!idx--)
+		gpio = dcb_gpio_table(bios, ver, &hdr, &cnt, len);
+	else
+		gpio = dcb_xpio_table(bios, idx, &xver, &hdr, &cnt, len);
+
 	if (gpio && ent < cnt)
 		return gpio + hdr + (ent * *len);
 	return 0x0000;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/xpio.c b/drivers/gpu/drm/nouveau/core/subdev/bios/xpio.c
new file mode 100644
index 0000000..e9b8e5d
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/xpio.c
@@ -0,0 +1,76 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <subdev/bios.h>
+#include <subdev/bios/gpio.h>
+#include <subdev/bios/xpio.h>
+
+static u16
+dcb_xpiod_table(struct nouveau_bios *bios, u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
+{
+	u16 data = dcb_gpio_table(bios, ver, hdr, cnt, len);
+	if (data && *ver >= 0x40 && *hdr >= 0x06) {
+		u16 xpio = nv_ro16(bios, data + 0x04);
+		if (xpio) {
+			*ver = nv_ro08(bios, data + 0x00);
+			*hdr = nv_ro08(bios, data + 0x01);
+			*cnt = nv_ro08(bios, data + 0x02);
+			*len = nv_ro08(bios, data + 0x03);
+			return xpio;
+		}
+	}
+	return 0x0000;
+}
+
+u16
+dcb_xpio_table(struct nouveau_bios *bios, u8 idx,
+	       u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
+{
+	u16 data = dcb_xpiod_table(bios, ver, hdr, cnt, len);
+	if (data && idx < *cnt) {
+		u16 xpio = nv_ro16(bios, data + *hdr + (idx * *len));
+		if (xpio) {
+			*ver = nv_ro08(bios, data + 0x00);
+			*hdr = nv_ro08(bios, data + 0x01);
+			*cnt = nv_ro08(bios, data + 0x02);
+			*len = nv_ro08(bios, data + 0x03);
+			return xpio;
+		}
+	}
+	return 0x0000;
+}
+
+u16
+dcb_xpio_parse(struct nouveau_bios *bios, u8 idx,
+	       u8 *ver, u8 *hdr, u8 *cnt, u8 *len,
+	       struct nvbios_xpio *info)
+{
+	u16 data = dcb_xpio_table(bios, idx, ver, hdr, cnt, len);
+	if (data && *len >= 6) {
+		info->type = nv_ro08(bios, data + 0x04);
+		info->addr = nv_ro08(bios, data + 0x05);
+		info->flags = nv_ro08(bios, data + 0x06);
+	}
+	return 0x0000;
+}
-- 
cgit v0.10.2


From e8c3cc0803b7f70f2f96d3bae3c4a4ac5238ed10 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Fri, 7 Dec 2012 11:01:25 +1000
Subject: drm/nouveau/bios: rename DCB_GPIO_PWM_FAN to DCB_GPIO_FAN

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/gpio.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/gpio.h
index 94e540e..cb06057 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/bios/gpio.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/gpio.h
@@ -5,7 +5,7 @@ enum dcb_gpio_func_name {
 	DCB_GPIO_PANEL_POWER = 0x01,
 	DCB_GPIO_TVDAC0 = 0x0c,
 	DCB_GPIO_TVDAC1 = 0x2d,
-	DCB_GPIO_PWM_FAN = 0x09,
+	DCB_GPIO_FAN = 0x09,
 	DCB_GPIO_FAN_SENSE = 0x3d,
 	DCB_GPIO_UNUSED = 0xff
 };
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
index f7d32f4..edc1664 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
@@ -213,7 +213,7 @@ nouveau_therm_fan_ctor(struct nouveau_therm *therm)
 	int ret;
 
 	/* attempt to locate a drivable fan, and determine control method */
-	ret = gpio->find(gpio, 0, DCB_GPIO_PWM_FAN, 0xff, &func);
+	ret = gpio->find(gpio, 0, DCB_GPIO_FAN, 0xff, &func);
 	if (ret == 0)
 		ret = nouveau_fanpwm_create(therm, &func);
 	if (ret != 0)
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/fantog.c b/drivers/gpu/drm/nouveau/core/subdev/therm/fantog.c
index 7e50f14..e0693e0 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/fantog.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/fantog.c
@@ -53,8 +53,8 @@ nouveau_fantog_update(struct nouveau_fantog_priv *priv, int percent)
 		percent = priv->percent;
 	priv->percent = percent;
 
-	duty = !gpio->get(gpio, 0, DCB_GPIO_PWM_FAN, 0xff);
-	gpio->set(gpio, 0, DCB_GPIO_PWM_FAN, 0xff, duty);
+	duty = !gpio->get(gpio, 0, DCB_GPIO_FAN, 0xff);
+	gpio->set(gpio, 0, DCB_GPIO_FAN, 0xff, duty);
 
 	if (list_empty(&priv->alarm.head) && percent != (duty * 100)) {
 		u64 next_change = (percent * priv->period_us) / 100;
-- 
cgit v0.10.2


From 708ff04b137f1ec233b9e82cb4852ee53ad26f9f Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Fri, 7 Dec 2012 11:37:28 +1000
Subject: drm/nouveau/therm: don't try pwm/toggle control if GPIO_FAN is input

My GTX660 has the GPIO_FAN function, but it's configured in input-mode;
presumably to monitor the frequency set by an I2C fan controller?

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/gpio.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/gpio.h
index cb06057..96d3364 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/bios/gpio.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/gpio.h
@@ -10,6 +10,13 @@ enum dcb_gpio_func_name {
 	DCB_GPIO_UNUSED = 0xff
 };
 
+#define DCB_GPIO_LOG_DIR     0x02
+#define DCB_GPIO_LOG_DIR_OUT 0x00
+#define DCB_GPIO_LOG_DIR_IN  0x02
+#define DCB_GPIO_LOG_VAL     0x01
+#define DCB_GPIO_LOG_VAL_LO  0x00
+#define DCB_GPIO_LOG_VAL_HI  0x01
+
 struct dcb_gpio_func {
 	u8 func;
 	u8 line;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
index edc1664..ddb23dc 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
@@ -214,14 +214,23 @@ nouveau_therm_fan_ctor(struct nouveau_therm *therm)
 
 	/* attempt to locate a drivable fan, and determine control method */
 	ret = gpio->find(gpio, 0, DCB_GPIO_FAN, 0xff, &func);
-	if (ret == 0)
-		ret = nouveau_fanpwm_create(therm, &func);
-	if (ret != 0)
-		ret = nouveau_fantog_create(therm, &func);
-	if (ret != 0)
+	if (ret == 0) {
+		if (func.log[0] & DCB_GPIO_LOG_DIR_IN) {
+			nv_debug(therm, "GPIO_FAN is in input mode\n");
+			ret = -EINVAL;
+		} else {
+			ret = nouveau_fanpwm_create(therm, &func);
+			if (ret != 0)
+				ret = nouveau_fantog_create(therm, &func);
+		}
+	}
+
+	/* no controllable fan found, create a dummy fan module */
+	if (ret != 0) {
 		ret = nouveau_fannil_create(therm);
-	if (ret)
-		return ret;
+		if (ret)
+			return ret;
+	}
 
 	nv_info(therm, "FAN control: %s\n", priv->fan->type);
 
-- 
cgit v0.10.2


From f1aa4a1de5e13b794a7e31efc0b3ddcda6926c56 Mon Sep 17 00:00:00 2001
From: Emil Velikov <emil.l.velikov@gmail.com>
Date: Fri, 7 Dec 2012 17:26:44 +0000
Subject: drm/nouveau: set legacy bios data before parsing the structure

Commit 767baf82 drm/nouveau: remove some more unnecessary legacy bios code

has introduced a regression my misplacing the code that sets the major/chip
versions, which are used whist parsing the bmp/bit structure in vbios

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.c b/drivers/gpu/drm/nouveau/nouveau_bios.c
index 5e1cfca..235de99 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bios.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bios.c
@@ -2008,6 +2008,8 @@ static bool NVInitVBIOS(struct drm_device *dev)
 
 	legacy->data = bios->data;
 	legacy->length = bios->size;
+	legacy->major_version = bios->version.major;
+	legacy->chip_version = bios->version.chip;
 	if (bios->bit_offset) {
 		legacy->type = NVBIOS_BIT;
 		legacy->offset = bios->bit_offset;
@@ -2019,8 +2021,6 @@ static bool NVInitVBIOS(struct drm_device *dev)
 		return !parse_bmp_structure(dev, legacy, legacy->offset);
 	}
 
-	legacy->major_version = bios->version.major;
-	legacy->chip_version = bios->version.chip;
 	return false;
 }
 
-- 
cgit v0.10.2


From 3600ad5e1b343f2c259901f736c717f5bc4deb50 Mon Sep 17 00:00:00 2001
From: Martin Peres <martin.peres@labri.fr>
Date: Sat, 8 Dec 2012 14:00:38 +0100
Subject: drm/nouveau/therm: force a minimum hysteresis on temperature alarm
 thresholds
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

This should avoid the situation where a user gets its kernel logs flooded when
temperature oscillates around a threshold with 0°C hysteresis.

This patch is just meant to fix broken vbios (as reported on a nv4e on
sysfs hwmon interface.

Signed-off-by: Martin Peres <martin.peres@labri.fr>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/temp.c b/drivers/gpu/drm/nouveau/core/subdev/therm/temp.c
index 820e097..bf9b3ce 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/temp.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/temp.c
@@ -58,11 +58,18 @@ static void
 nouveau_therm_temp_safety_checks(struct nouveau_therm *therm)
 {
 	struct nouveau_therm_priv *priv = (void *)therm;
+	struct nvbios_therm_sensor *s = &priv->bios_sensor;
 
 	if (!priv->bios_sensor.slope_div)
 		priv->bios_sensor.slope_div = 1;
 	if (!priv->bios_sensor.offset_den)
 		priv->bios_sensor.offset_den = 1;
+
+	/* enforce a minimum hysteresis on thresholds */
+	s->thrs_fan_boost.hysteresis = max_t(u8, s->thrs_fan_boost.hysteresis, 2);
+	s->thrs_down_clock.hysteresis = max_t(u8, s->thrs_down_clock.hysteresis, 2);
+	s->thrs_critical.hysteresis = max_t(u8, s->thrs_critical.hysteresis, 2);
+	s->thrs_shutdown.hysteresis = max_t(u8, s->thrs_shutdown.hysteresis, 2);
 }
 
 /* must be called with alarm_program_lock taken ! */
-- 
cgit v0.10.2


From fc10199ee3ef98fa899b6d629b310da181f0abfd Mon Sep 17 00:00:00 2001
From: Marcin Slusarz <marcin.slusarz@gmail.com>
Date: Sun, 9 Dec 2012 15:45:19 +0100
Subject: drm/nouveau: split fifo interrupt handler

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nv04.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nv04.c
index a47a854..f6c4e71 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nv04.c
@@ -398,6 +398,92 @@ out:
 	return handled;
 }
 
+static void
+nv04_fifo_cache_error(struct nouveau_device *device,
+		struct nv04_fifo_priv *priv, u32 chid, u32 get)
+{
+	u32 mthd, data;
+	int ptr;
+
+	/* NV_PFIFO_CACHE1_GET actually goes to 0xffc before wrapping on my
+	 * G80 chips, but CACHE1 isn't big enough for this much data.. Tests
+	 * show that it wraps around to the start at GET=0x800.. No clue as to
+	 * why..
+	 */
+	ptr = (get & 0x7ff) >> 2;
+
+	if (device->card_type < NV_40) {
+		mthd = nv_rd32(priv, NV04_PFIFO_CACHE1_METHOD(ptr));
+		data = nv_rd32(priv, NV04_PFIFO_CACHE1_DATA(ptr));
+	} else {
+		mthd = nv_rd32(priv, NV40_PFIFO_CACHE1_METHOD(ptr));
+		data = nv_rd32(priv, NV40_PFIFO_CACHE1_DATA(ptr));
+	}
+
+	if (!nv04_fifo_swmthd(priv, chid, mthd, data)) {
+		nv_error(priv,
+			 "CACHE_ERROR - Ch %d/%d Mthd 0x%04x Data 0x%08x\n",
+			 chid, (mthd >> 13) & 7, mthd & 0x1ffc, data);
+	}
+
+	nv_wr32(priv, NV04_PFIFO_CACHE1_DMA_PUSH, 0);
+	nv_wr32(priv, NV03_PFIFO_INTR_0, NV_PFIFO_INTR_CACHE_ERROR);
+
+	nv_wr32(priv, NV03_PFIFO_CACHE1_PUSH0,
+		nv_rd32(priv, NV03_PFIFO_CACHE1_PUSH0) & ~1);
+	nv_wr32(priv, NV03_PFIFO_CACHE1_GET, get + 4);
+	nv_wr32(priv, NV03_PFIFO_CACHE1_PUSH0,
+		nv_rd32(priv, NV03_PFIFO_CACHE1_PUSH0) | 1);
+	nv_wr32(priv, NV04_PFIFO_CACHE1_HASH, 0);
+
+	nv_wr32(priv, NV04_PFIFO_CACHE1_DMA_PUSH,
+		nv_rd32(priv, NV04_PFIFO_CACHE1_DMA_PUSH) | 1);
+	nv_wr32(priv, NV04_PFIFO_CACHE1_PULL0, 1);
+}
+
+static void
+nv04_fifo_dma_pusher(struct nouveau_device *device, struct nv04_fifo_priv *priv,
+		u32 chid)
+{
+	u32 dma_get = nv_rd32(priv, 0x003244);
+	u32 dma_put = nv_rd32(priv, 0x003240);
+	u32 push = nv_rd32(priv, 0x003220);
+	u32 state = nv_rd32(priv, 0x003228);
+
+	if (device->card_type == NV_50) {
+		u32 ho_get = nv_rd32(priv, 0x003328);
+		u32 ho_put = nv_rd32(priv, 0x003320);
+		u32 ib_get = nv_rd32(priv, 0x003334);
+		u32 ib_put = nv_rd32(priv, 0x003330);
+
+		nv_error(priv,
+			 "DMA_PUSHER - Ch %d Get 0x%02x%08x Put 0x%02x%08x IbGet 0x%08x IbPut 0x%08x State 0x%08x (err: %s) Push 0x%08x\n",
+			 chid, ho_get, dma_get, ho_put, dma_put, ib_get, ib_put,
+			 state, nv_dma_state_err(state), push);
+
+		/* METHOD_COUNT, in DMA_STATE on earlier chipsets */
+		nv_wr32(priv, 0x003364, 0x00000000);
+		if (dma_get != dma_put || ho_get != ho_put) {
+			nv_wr32(priv, 0x003244, dma_put);
+			nv_wr32(priv, 0x003328, ho_put);
+		} else
+		if (ib_get != ib_put)
+			nv_wr32(priv, 0x003334, ib_put);
+	} else {
+		nv_error(priv,
+			 "DMA_PUSHER - Ch %d Get 0x%08x Put 0x%08x State 0x%08x (err: %s) Push 0x%08x\n",
+			 chid, dma_get, dma_put, state, nv_dma_state_err(state),
+			 push);
+
+		if (dma_get != dma_put)
+			nv_wr32(priv, 0x003244, dma_put);
+	}
+
+	nv_wr32(priv, 0x003228, 0x00000000);
+	nv_wr32(priv, 0x003220, 0x00000001);
+	nv_wr32(priv, 0x002100, NV_PFIFO_INTR_DMA_PUSHER);
+}
+
 void
 nv04_fifo_intr(struct nouveau_subdev *subdev)
 {
@@ -416,96 +502,12 @@ nv04_fifo_intr(struct nouveau_subdev *subdev)
 		get  = nv_rd32(priv, NV03_PFIFO_CACHE1_GET);
 
 		if (status & NV_PFIFO_INTR_CACHE_ERROR) {
-			uint32_t mthd, data;
-			int ptr;
-
-			/* NV_PFIFO_CACHE1_GET actually goes to 0xffc before
-			 * wrapping on my G80 chips, but CACHE1 isn't big
-			 * enough for this much data.. Tests show that it
-			 * wraps around to the start at GET=0x800.. No clue
-			 * as to why..
-			 */
-			ptr = (get & 0x7ff) >> 2;
-
-			if (device->card_type < NV_40) {
-				mthd = nv_rd32(priv,
-					NV04_PFIFO_CACHE1_METHOD(ptr));
-				data = nv_rd32(priv,
-					NV04_PFIFO_CACHE1_DATA(ptr));
-			} else {
-				mthd = nv_rd32(priv,
-					NV40_PFIFO_CACHE1_METHOD(ptr));
-				data = nv_rd32(priv,
-					NV40_PFIFO_CACHE1_DATA(ptr));
-			}
-
-			if (!nv04_fifo_swmthd(priv, chid, mthd, data)) {
-				nv_error(priv, "CACHE_ERROR - Ch %d/%d "
-					      "Mthd 0x%04x Data 0x%08x\n",
-					chid, (mthd >> 13) & 7, mthd & 0x1ffc,
-					data);
-			}
-
-			nv_wr32(priv, NV04_PFIFO_CACHE1_DMA_PUSH, 0);
-			nv_wr32(priv, NV03_PFIFO_INTR_0,
-						NV_PFIFO_INTR_CACHE_ERROR);
-
-			nv_wr32(priv, NV03_PFIFO_CACHE1_PUSH0,
-				nv_rd32(priv, NV03_PFIFO_CACHE1_PUSH0) & ~1);
-			nv_wr32(priv, NV03_PFIFO_CACHE1_GET, get + 4);
-			nv_wr32(priv, NV03_PFIFO_CACHE1_PUSH0,
-				nv_rd32(priv, NV03_PFIFO_CACHE1_PUSH0) | 1);
-			nv_wr32(priv, NV04_PFIFO_CACHE1_HASH, 0);
-
-			nv_wr32(priv, NV04_PFIFO_CACHE1_DMA_PUSH,
-				nv_rd32(priv, NV04_PFIFO_CACHE1_DMA_PUSH) | 1);
-			nv_wr32(priv, NV04_PFIFO_CACHE1_PULL0, 1);
-
+			nv04_fifo_cache_error(device, priv, chid, get);
 			status &= ~NV_PFIFO_INTR_CACHE_ERROR;
 		}
 
 		if (status & NV_PFIFO_INTR_DMA_PUSHER) {
-			u32 dma_get = nv_rd32(priv, 0x003244);
-			u32 dma_put = nv_rd32(priv, 0x003240);
-			u32 push = nv_rd32(priv, 0x003220);
-			u32 state = nv_rd32(priv, 0x003228);
-
-			if (device->card_type == NV_50) {
-				u32 ho_get = nv_rd32(priv, 0x003328);
-				u32 ho_put = nv_rd32(priv, 0x003320);
-				u32 ib_get = nv_rd32(priv, 0x003334);
-				u32 ib_put = nv_rd32(priv, 0x003330);
-
-				nv_error(priv, "DMA_PUSHER - Ch %d Get 0x%02x%08x "
-				     "Put 0x%02x%08x IbGet 0x%08x IbPut 0x%08x "
-				     "State 0x%08x (err: %s) Push 0x%08x\n",
-					chid, ho_get, dma_get, ho_put,
-					dma_put, ib_get, ib_put, state,
-					nv_dma_state_err(state),
-					push);
-
-				/* METHOD_COUNT, in DMA_STATE on earlier chipsets */
-				nv_wr32(priv, 0x003364, 0x00000000);
-				if (dma_get != dma_put || ho_get != ho_put) {
-					nv_wr32(priv, 0x003244, dma_put);
-					nv_wr32(priv, 0x003328, ho_put);
-				} else
-				if (ib_get != ib_put) {
-					nv_wr32(priv, 0x003334, ib_put);
-				}
-			} else {
-				nv_error(priv, "DMA_PUSHER - Ch %d Get 0x%08x "
-					     "Put 0x%08x State 0x%08x (err: %s) Push 0x%08x\n",
-					chid, dma_get, dma_put, state,
-					nv_dma_state_err(state), push);
-
-				if (dma_get != dma_put)
-					nv_wr32(priv, 0x003244, dma_put);
-			}
-
-			nv_wr32(priv, 0x003228, 0x00000000);
-			nv_wr32(priv, 0x003220, 0x00000001);
-			nv_wr32(priv, 0x002100, NV_PFIFO_INTR_DMA_PUSHER);
+			nv04_fifo_dma_pusher(device, priv, chid);
 			status &= ~NV_PFIFO_INTR_DMA_PUSHER;
 		}
 
-- 
cgit v0.10.2


From f533da1027e294a2d6e0b455bf8ee915609ff216 Mon Sep 17 00:00:00 2001
From: Marcin Slusarz <marcin.slusarz@gmail.com>
Date: Sun, 9 Dec 2012 15:45:20 +0100
Subject: drm/nouveau: use pr_cont

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/core/enum.c b/drivers/gpu/drm/nouveau/core/core/enum.c
index 7cc7133..148f91f 100644
--- a/drivers/gpu/drm/nouveau/core/core/enum.c
+++ b/drivers/gpu/drm/nouveau/core/core/enum.c
@@ -45,9 +45,9 @@ nouveau_enum_print(const struct nouveau_enum *en, u32 value)
 {
 	en = nouveau_enum_find(en, value);
 	if (en)
-		printk("%s", en->name);
+		pr_cont("%s", en->name);
 	else
-		printk("(unknown enum 0x%08x)", value);
+		pr_cont("(unknown enum 0x%08x)", value);
 }
 
 void
@@ -55,7 +55,7 @@ nouveau_bitfield_print(const struct nouveau_bitfield *bf, u32 value)
 {
 	while (bf->name) {
 		if (value & bf->mask) {
-			printk(" %s", bf->name);
+			pr_cont(" %s", bf->name);
 			value &= ~bf->mask;
 		}
 
@@ -63,5 +63,5 @@ nouveau_bitfield_print(const struct nouveau_bitfield *bf, u32 value)
 	}
 
 	if (value)
-		printk(" (unknown bits 0x%08x)", value);
+		pr_cont(" (unknown bits 0x%08x)", value);
 }
diff --git a/drivers/gpu/drm/nouveau/core/engine/copy/nva3.c b/drivers/gpu/drm/nouveau/core/engine/copy/nva3.c
index 283248c..6068f5f 100644
--- a/drivers/gpu/drm/nouveau/core/engine/copy/nva3.c
+++ b/drivers/gpu/drm/nouveau/core/engine/copy/nva3.c
@@ -100,7 +100,7 @@ nva3_copy_intr(struct nouveau_subdev *subdev)
 	if (stat & 0x00000040) {
 		nv_error(falcon, "DISPATCH_ERROR [");
 		nouveau_enum_print(nva3_copy_isr_error_name, ssta);
-		printk("] ch %d [0x%010llx] subc %d mthd 0x%04x data 0x%08x\n",
+		pr_cont("] ch %d [0x%010llx] subc %d mthd 0x%04x data 0x%08x\n",
 		       chid, inst << 12, subc, mthd, data);
 		nv_wo32(falcon, 0x004, 0x00000040);
 		stat &= ~0x00000040;
diff --git a/drivers/gpu/drm/nouveau/core/engine/crypt/nv84.c b/drivers/gpu/drm/nouveau/core/engine/crypt/nv84.c
index b974905..8d0a440 100644
--- a/drivers/gpu/drm/nouveau/core/engine/crypt/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/crypt/nv84.c
@@ -128,7 +128,7 @@ nv84_crypt_intr(struct nouveau_subdev *subdev)
 	if (stat) {
 		nv_error(priv, "");
 		nouveau_bitfield_print(nv84_crypt_intr_mask, stat);
-		printk(" ch %d [0x%010llx] mthd 0x%04x data 0x%08x\n",
+		pr_cont(" ch %d [0x%010llx] mthd 0x%04x data 0x%08x\n",
 		       chid, (u64)inst << 12, mthd, data);
 	}
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/crypt/nv98.c b/drivers/gpu/drm/nouveau/core/engine/crypt/nv98.c
index 21986f3..1b33c20 100644
--- a/drivers/gpu/drm/nouveau/core/engine/crypt/nv98.c
+++ b/drivers/gpu/drm/nouveau/core/engine/crypt/nv98.c
@@ -102,7 +102,7 @@ nv98_crypt_intr(struct nouveau_subdev *subdev)
 	if (stat & 0x00000040) {
 		nv_error(priv, "DISPATCH_ERROR [");
 		nouveau_enum_print(nv98_crypt_isr_error_name, ssta);
-		printk("] ch %d [0x%010llx] subc %d mthd 0x%04x data 0x%08x\n",
+		pr_cont("] ch %d [0x%010llx] subc %d mthd 0x%04x data 0x%08x\n",
 		       chid, (u64)inst << 12, subc, mthd, data);
 		nv_wr32(priv, 0x087004, 0x00000040);
 		stat &= ~0x00000040;
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
index b4365dd..d341475 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
@@ -420,16 +420,16 @@ nvc0_fifo_isr_vm_fault(struct nvc0_fifo_priv *priv, int unit)
 	nv_error(priv, "%s fault at 0x%010llx [", (stat & 0x00000080) ?
 		 "write" : "read", (u64)vahi << 32 | valo);
 	nouveau_enum_print(nvc0_fifo_fault_reason, stat & 0x0000000f);
-	printk("] from ");
+	pr_cont("] from ");
 	nouveau_enum_print(nvc0_fifo_fault_unit, unit);
 	if (stat & 0x00000040) {
-		printk("/");
+		pr_cont("/");
 		nouveau_enum_print(nvc0_fifo_fault_hubclient, client);
 	} else {
-		printk("/GPC%d/", (stat & 0x1f000000) >> 24);
+		pr_cont("/GPC%d/", (stat & 0x1f000000) >> 24);
 		nouveau_enum_print(nvc0_fifo_fault_gpcclient, client);
 	}
-	printk(" on channel 0x%010llx\n", (u64)inst << 12);
+	pr_cont(" on channel 0x%010llx\n", (u64)inst << 12);
 }
 
 static int
@@ -484,7 +484,7 @@ nvc0_fifo_isr_subfifo_intr(struct nvc0_fifo_priv *priv, int unit)
 	if (show) {
 		nv_error(priv, "SUBFIFO%d:", unit);
 		nouveau_bitfield_print(nvc0_fifo_subfifo_intr, show);
-		printk("\n");
+		pr_cont("\n");
 		nv_error(priv, "SUBFIFO%d: ch %d subc %d mthd 0x%04x "
 			       "data 0x%08x\n",
 			 unit, chid, subc, mthd, data);
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
index c930da9..c874c24 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
@@ -416,16 +416,16 @@ nve0_fifo_isr_vm_fault(struct nve0_fifo_priv *priv, int unit)
 	nv_error(priv, "PFIFO: %s fault at 0x%010llx [", (stat & 0x00000080) ?
 		       "write" : "read", (u64)vahi << 32 | valo);
 	nouveau_enum_print(nve0_fifo_fault_reason, stat & 0x0000000f);
-	printk("] from ");
+	pr_cont("] from ");
 	nouveau_enum_print(nve0_fifo_fault_unit, unit);
 	if (stat & 0x00000040) {
-		printk("/");
+		pr_cont("/");
 		nouveau_enum_print(nve0_fifo_fault_hubclient, client);
 	} else {
-		printk("/GPC%d/", (stat & 0x1f000000) >> 24);
+		pr_cont("/GPC%d/", (stat & 0x1f000000) >> 24);
 		nouveau_enum_print(nve0_fifo_fault_gpcclient, client);
 	}
-	printk(" on channel 0x%010llx\n", (u64)inst << 12);
+	pr_cont(" on channel 0x%010llx\n", (u64)inst << 12);
 }
 
 static int
@@ -480,7 +480,7 @@ nve0_fifo_isr_subfifo_intr(struct nve0_fifo_priv *priv, int unit)
 	if (show) {
 		nv_error(priv, "SUBFIFO%d:", unit);
 		nouveau_bitfield_print(nve0_fifo_subfifo_intr, show);
-		printk("\n");
+		pr_cont("\n");
 		nv_error(priv, "SUBFIFO%d: ch %d subc %d mthd 0x%04x "
 			       "data 0x%08x\n",
 			 unit, chid, subc, mthd, data);
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv04.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv04.c
index e30a9c5..f4fb844 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv04.c
@@ -1299,11 +1299,11 @@ nv04_graph_intr(struct nouveau_subdev *subdev)
 	if (show) {
 		nv_error(priv, "");
 		nouveau_bitfield_print(nv04_graph_intr_name, show);
-		printk(" nsource:");
+		pr_cont(" nsource:");
 		nouveau_bitfield_print(nv04_graph_nsource, nsource);
-		printk(" nstatus:");
+		pr_cont(" nstatus:");
 		nouveau_bitfield_print(nv04_graph_nstatus, nstatus);
-		printk("\n");
+		pr_cont("\n");
 		nv_error(priv, "ch %d/%d class 0x%04x "
 			       "mthd 0x%04x data 0x%08x\n",
 			 chid, subc, class, mthd, data);
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv10.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv10.c
index 5c0f843..deea80f 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv10.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv10.c
@@ -1195,11 +1195,11 @@ nv10_graph_intr(struct nouveau_subdev *subdev)
 	if (show) {
 		nv_error(priv, "");
 		nouveau_bitfield_print(nv10_graph_intr_name, show);
-		printk(" nsource:");
+		pr_cont(" nsource:");
 		nouveau_bitfield_print(nv04_graph_nsource, nsource);
-		printk(" nstatus:");
+		pr_cont(" nstatus:");
 		nouveau_bitfield_print(nv10_graph_nstatus, nstatus);
-		printk("\n");
+		pr_cont("\n");
 		nv_error(priv, "ch %d/%d class 0x%04x "
 			       "mthd 0x%04x data 0x%08x\n",
 			 chid, subc, class, mthd, data);
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv20.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv20.c
index 5b20401..4c0d79c5 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv20.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv20.c
@@ -226,11 +226,11 @@ nv20_graph_intr(struct nouveau_subdev *subdev)
 	if (show) {
 		nv_error(priv, "");
 		nouveau_bitfield_print(nv10_graph_intr_name, show);
-		printk(" nsource:");
+		pr_cont(" nsource:");
 		nouveau_bitfield_print(nv04_graph_nsource, nsource);
-		printk(" nstatus:");
+		pr_cont(" nstatus:");
 		nouveau_bitfield_print(nv10_graph_nstatus, nstatus);
-		printk("\n");
+		pr_cont("\n");
 		nv_error(priv, "ch %d/%d class 0x%04x mthd 0x%04x data 0x%08x\n",
 			chid, subc, class, mthd, data);
 	}
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv40.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv40.c
index 0b36dd3..961a073 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv40.c
@@ -323,11 +323,11 @@ nv40_graph_intr(struct nouveau_subdev *subdev)
 	if (show) {
 		nv_error(priv, "");
 		nouveau_bitfield_print(nv10_graph_intr_name, show);
-		printk(" nsource:");
+		pr_cont(" nsource:");
 		nouveau_bitfield_print(nv04_graph_nsource, nsource);
-		printk(" nstatus:");
+		pr_cont(" nstatus:");
 		nouveau_bitfield_print(nv10_graph_nstatus, nstatus);
-		printk("\n");
+		pr_cont("\n");
 		nv_error(priv, "ch %d [0x%08x] subc %d class 0x%04x "
 			       "mthd 0x%04x data 0x%08x\n",
 			 chid, inst << 4, subc, class, mthd, data);
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c
index b1c3d83..2c9e9ad 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c
@@ -418,7 +418,7 @@ nv50_priv_mp_trap(struct nv50_graph_priv *priv, int tpid, int display)
 			nv_error(priv, "TRAP_MP_EXEC - "
 					"TP %d MP %d: ", tpid, i);
 			nouveau_enum_print(nv50_mp_exec_error_names, status);
-			printk(" at %06x warp %d, opcode %08x %08x\n",
+			pr_cont(" at %06x warp %d, opcode %08x %08x\n",
 					pc&0xffffff, pc >> 24,
 					oplow, ophigh);
 		}
@@ -623,7 +623,7 @@ nv50_graph_trap_handler(struct nv50_graph_priv *priv, u32 display,
 		if (display) {
 			nv_error(priv, "TRAP_M2MF");
 			nouveau_bitfield_print(nv50_graph_trap_m2mf, ustatus);
-			printk("\n");
+			pr_cont("\n");
 			nv_error(priv, "TRAP_M2MF %08x %08x %08x %08x\n",
 				nv_rd32(priv, 0x406804), nv_rd32(priv, 0x406808),
 				nv_rd32(priv, 0x40680c), nv_rd32(priv, 0x406810));
@@ -644,7 +644,7 @@ nv50_graph_trap_handler(struct nv50_graph_priv *priv, u32 display,
 		if (display) {
 			nv_error(priv, "TRAP_VFETCH");
 			nouveau_bitfield_print(nv50_graph_trap_vfetch, ustatus);
-			printk("\n");
+			pr_cont("\n");
 			nv_error(priv, "TRAP_VFETCH %08x %08x %08x %08x\n",
 				nv_rd32(priv, 0x400c00), nv_rd32(priv, 0x400c08),
 				nv_rd32(priv, 0x400c0c), nv_rd32(priv, 0x400c10));
@@ -661,7 +661,7 @@ nv50_graph_trap_handler(struct nv50_graph_priv *priv, u32 display,
 		if (display) {
 			nv_error(priv, "TRAP_STRMOUT");
 			nouveau_bitfield_print(nv50_graph_trap_strmout, ustatus);
-			printk("\n");
+			pr_cont("\n");
 			nv_error(priv, "TRAP_STRMOUT %08x %08x %08x %08x\n",
 				nv_rd32(priv, 0x401804), nv_rd32(priv, 0x401808),
 				nv_rd32(priv, 0x40180c), nv_rd32(priv, 0x401810));
@@ -682,7 +682,7 @@ nv50_graph_trap_handler(struct nv50_graph_priv *priv, u32 display,
 		if (display) {
 			nv_error(priv, "TRAP_CCACHE");
 			nouveau_bitfield_print(nv50_graph_trap_ccache, ustatus);
-			printk("\n");
+			pr_cont("\n");
 			nv_error(priv, "TRAP_CCACHE %08x %08x %08x %08x"
 				     " %08x %08x %08x\n",
 				nv_rd32(priv, 0x405000), nv_rd32(priv, 0x405004),
@@ -774,7 +774,7 @@ nv50_graph_intr(struct nouveau_subdev *subdev)
 		u32 ecode = nv_rd32(priv, 0x400110);
 		nv_error(priv, "DATA_ERROR ");
 		nouveau_enum_print(nv50_data_error_names, ecode);
-		printk("\n");
+		pr_cont("\n");
 	}
 
 	if (stat & 0x00200000) {
@@ -788,7 +788,7 @@ nv50_graph_intr(struct nouveau_subdev *subdev)
 	if (show) {
 		nv_error(priv, "");
 		nouveau_bitfield_print(nv50_graph_intr_name, show);
-		printk("\n");
+		pr_cont("\n");
 		nv_error(priv, "ch %d [0x%010llx] subc %d class 0x%04x "
 			       "mthd 0x%04x data 0x%08x\n",
 			 chid, (u64)inst << 12, subc, class, mthd, data);
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c
index 45aff5f..e644298 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c
@@ -454,7 +454,7 @@ nvc0_graph_intr(struct nouveau_subdev *subdev)
 	if (stat & 0x00100000) {
 		nv_error(priv, "DATA_ERROR [");
 		nouveau_enum_print(nv50_data_error_names, code);
-		printk("] ch %d [0x%010llx] subc %d class 0x%04x "
+		pr_cont("] ch %d [0x%010llx] subc %d class 0x%04x "
 		       "mthd 0x%04x data 0x%08x\n",
 		       chid, inst << 12, subc, class, mthd, data);
 		nv_wr32(priv, 0x400100, 0x00100000);
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nve0.c b/drivers/gpu/drm/nouveau/core/engine/graph/nve0.c
index 9f82e97..0f35a34 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nve0.c
@@ -166,7 +166,7 @@ nve0_graph_intr(struct nouveau_subdev *subdev)
 	if (stat & 0x00100000) {
 		nv_error(priv, "DATA_ERROR [");
 		nouveau_enum_print(nv50_data_error_names, code);
-		printk("] ch %d [0x%010llx] subc %d class 0x%04x "
+		pr_cont("] ch %d [0x%010llx] subc %d class 0x%04x "
 		       "mthd 0x%04x data 0x%08x\n",
 		       chid, inst, subc, class, mthd, data);
 		nv_wr32(priv, 0x400100, 0x00100000);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c
index 487cb8c..88bdb02 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c
@@ -372,30 +372,30 @@ nv50_fb_intr(struct nouveau_subdev *subdev)
 
 	en = nouveau_enum_find(vm_engine, st0);
 	if (en)
-		printk("%s/", en->name);
+		pr_cont("%s/", en->name);
 	else
-		printk("%02x/", st0);
+		pr_cont("%02x/", st0);
 
 	cl = nouveau_enum_find(vm_client, st2);
 	if (cl)
-		printk("%s/", cl->name);
+		pr_cont("%s/", cl->name);
 	else
-		printk("%02x/", st2);
+		pr_cont("%02x/", st2);
 
 	if      (cl && cl->data) cl = nouveau_enum_find(cl->data, st3);
 	else if (en && en->data) cl = nouveau_enum_find(en->data, st3);
 	else                     cl = NULL;
 	if (cl)
-		printk("%s", cl->name);
+		pr_cont("%s", cl->name);
 	else
-		printk("%02x", st3);
+		pr_cont("%02x", st3);
 
-	printk(" reason: ");
+	pr_cont(" reason: ");
 	en = nouveau_enum_find(vm_fault, st1);
 	if (en)
-		printk("%s\n", en->name);
+		pr_cont("%s\n", en->name);
 	else
-		printk("0x%08x\n", st1);
+		pr_cont("0x%08x\n", st1);
 }
 
 static int
diff --git a/drivers/gpu/drm/nouveau/core/subdev/mxm/mxms.c b/drivers/gpu/drm/nouveau/core/subdev/mxm/mxms.c
index 839ca1e..4bde7f7 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/mxm/mxms.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/mxm/mxms.c
@@ -156,15 +156,15 @@ mxms_foreach(struct nouveau_mxm *mxm, u8 types,
 
 			nv_debug(mxm, "%4s: ", mxms_desc_name[type]);
 			for (j = headerlen - 1; j >= 0; j--)
-				printk("%02x", dump[j]);
-			printk("\n");
+				pr_cont("%02x", dump[j]);
+			pr_cont("\n");
 			dump += headerlen;
 
 			for (i = 0; i < entries; i++, dump += recordlen) {
 				nv_debug(mxm, "      ");
 				for (j = recordlen - 1; j >= 0; j--)
-					printk("%02x", dump[j]);
-				printk("\n");
+					pr_cont("%02x", dump[j]);
+				pr_cont("\n");
 			}
 		}
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.c b/drivers/gpu/drm/nouveau/nouveau_bios.c
index 235de99..b640657 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bios.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bios.c
@@ -1884,9 +1884,9 @@ parse_dcb_table(struct drm_device *dev, struct nvbios *bios)
 		if (conn[0] != 0xff) {
 			NV_INFO(drm, "DCB conn %02d: ", idx);
 			if (olddcb_conntab(dev)[3] < 4)
-				printk("%04x\n", ROM16(conn[0]));
+				pr_cont("%04x\n", ROM16(conn[0]));
 			else
-				printk("%08x\n", ROM32(conn[0]));
+				pr_cont("%08x\n", ROM32(conn[0]));
 		}
 	}
 	dcb_fake_connectors(bios);
-- 
cgit v0.10.2


From a2896cede08ddeed4f2c4fca616943a63dff5fa8 Mon Sep 17 00:00:00 2001
From: Marcin Slusarz <marcin.slusarz@gmail.com>
Date: Sun, 9 Dec 2012 15:45:21 +0100
Subject: drm/nouveau: prepare for reporting channel owner

- record channel owner process name
- add some helpers for accessing this information
- let nouveau_enum hold additional value (will be needed in the next patch)

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/core/client.c b/drivers/gpu/drm/nouveau/core/core/client.c
index 8bbb58f..295c221 100644
--- a/drivers/gpu/drm/nouveau/core/core/client.c
+++ b/drivers/gpu/drm/nouveau/core/core/client.c
@@ -99,3 +99,13 @@ nouveau_client_fini(struct nouveau_client *client, bool suspend)
 	nv_debug(client, "%s completed with %d\n", name[suspend], ret);
 	return ret;
 }
+
+const char *
+nouveau_client_name(void *obj)
+{
+	const char *client_name = "unknown";
+	struct nouveau_client *client = nouveau_client(obj);
+	if (client)
+		client_name = client->name;
+	return client_name;
+}
diff --git a/drivers/gpu/drm/nouveau/core/core/enum.c b/drivers/gpu/drm/nouveau/core/core/enum.c
index 148f91f..dd43479 100644
--- a/drivers/gpu/drm/nouveau/core/core/enum.c
+++ b/drivers/gpu/drm/nouveau/core/core/enum.c
@@ -40,7 +40,7 @@ nouveau_enum_find(const struct nouveau_enum *en, u32 value)
 	return NULL;
 }
 
-void
+const struct nouveau_enum *
 nouveau_enum_print(const struct nouveau_enum *en, u32 value)
 {
 	en = nouveau_enum_find(en, value);
@@ -48,6 +48,7 @@ nouveau_enum_print(const struct nouveau_enum *en, u32 value)
 		pr_cont("%s", en->name);
 	else
 		pr_cont("(unknown enum 0x%08x)", value);
+	return en;
 }
 
 void
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/base.c b/drivers/gpu/drm/nouveau/core/engine/fifo/base.c
index c2b9db3..ca1057a 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/base.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/base.c
@@ -22,6 +22,7 @@
  * Authors: Ben Skeggs
  */
 
+#include <core/client.h>
 #include <core/object.h>
 #include <core/handle.h>
 #include <core/class.h>
@@ -146,6 +147,20 @@ nouveau_fifo_chid(struct nouveau_fifo *priv, struct nouveau_object *object)
 	return -1;
 }
 
+const char *
+nouveau_client_name_for_fifo_chid(struct nouveau_fifo *fifo, u32 chid)
+{
+	struct nouveau_fifo_chan *chan = NULL;
+	unsigned long flags;
+
+	spin_lock_irqsave(&fifo->lock, flags);
+	if (chid >= fifo->min && chid <= fifo->max)
+		chan = (void *)fifo->channel[chid];
+	spin_unlock_irqrestore(&fifo->lock, flags);
+
+	return nouveau_client_name(chan);
+}
+
 void
 nouveau_fifo_destroy(struct nouveau_fifo *priv)
 {
diff --git a/drivers/gpu/drm/nouveau/core/include/core/client.h b/drivers/gpu/drm/nouveau/core/include/core/client.h
index 63acc03..c66eac5 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/client.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/client.h
@@ -7,7 +7,7 @@ struct nouveau_client {
 	struct nouveau_namedb base;
 	struct nouveau_handle *root;
 	struct nouveau_object *device;
-	char name[16];
+	char name[32];
 	u32 debug;
 	struct nouveau_vm *vm;
 };
@@ -41,5 +41,6 @@ int  nouveau_client_create_(const char *name, u64 device, const char *cfg,
 
 int  nouveau_client_init(struct nouveau_client *);
 int  nouveau_client_fini(struct nouveau_client *, bool suspend);
+const char *nouveau_client_name(void *obj);
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/include/core/enum.h b/drivers/gpu/drm/nouveau/core/include/core/enum.h
index e7b1e18..4fc62bb 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/enum.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/enum.h
@@ -5,12 +5,13 @@ struct nouveau_enum {
 	u32 value;
 	const char *name;
 	const void *data;
+	u32 data2;
 };
 
 const struct nouveau_enum *
 nouveau_enum_find(const struct nouveau_enum *, u32 value);
 
-void
+const struct nouveau_enum *
 nouveau_enum_print(const struct nouveau_enum *en, u32 value);
 
 struct nouveau_bitfield {
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/fifo.h b/drivers/gpu/drm/nouveau/core/include/engine/fifo.h
index f18846c..543e4ef 100644
--- a/drivers/gpu/drm/nouveau/core/include/engine/fifo.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/fifo.h
@@ -92,6 +92,8 @@ int nouveau_fifo_create_(struct nouveau_object *, struct nouveau_object *,
 			 struct nouveau_oclass *, int min, int max,
 			 int size, void **);
 void nouveau_fifo_destroy(struct nouveau_fifo *);
+const char *
+nouveau_client_name_for_fifo_chid(struct nouveau_fifo *fifo, u32 chid);
 
 #define _nouveau_fifo_init _nouveau_engine_init
 #define _nouveau_fifo_fini _nouveau_engine_fini
diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c
index 8b090f1..2cd66cd 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_drm.c
@@ -543,10 +543,11 @@ nouveau_drm_open(struct drm_device *dev, struct drm_file *fpriv)
 	struct pci_dev *pdev = dev->pdev;
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	struct nouveau_cli *cli;
-	char name[16];
+	char name[32], tmpname[TASK_COMM_LEN];
 	int ret;
 
-	snprintf(name, sizeof(name), "%d", pid_nr(fpriv->pid));
+	get_task_comm(tmpname, current);
+	snprintf(name, sizeof(name), "%s[%d]", tmpname, pid_nr(fpriv->pid));
 
 	ret = nouveau_cli_create(pdev, name, sizeof(*cli), (void **)&cli);
 	if (ret)
-- 
cgit v0.10.2


From 93260d3c026b539931d909a4d68490c32b6d73ce Mon Sep 17 00:00:00 2001
From: Marcin Slusarz <marcin.slusarz@gmail.com>
Date: Sun, 9 Dec 2012 23:00:34 +0100
Subject: drm/nouveau: report channel owner in error messages

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/engine/copy/nva3.c b/drivers/gpu/drm/nouveau/core/engine/copy/nva3.c
index 6068f5f..d6dc2a6 100644
--- a/drivers/gpu/drm/nouveau/core/engine/copy/nva3.c
+++ b/drivers/gpu/drm/nouveau/core/engine/copy/nva3.c
@@ -22,6 +22,7 @@
  * Authors: Ben Skeggs
  */
 
+#include <core/client.h>
 #include <core/falcon.h>
 #include <core/class.h>
 #include <core/enum.h>
@@ -100,8 +101,9 @@ nva3_copy_intr(struct nouveau_subdev *subdev)
 	if (stat & 0x00000040) {
 		nv_error(falcon, "DISPATCH_ERROR [");
 		nouveau_enum_print(nva3_copy_isr_error_name, ssta);
-		pr_cont("] ch %d [0x%010llx] subc %d mthd 0x%04x data 0x%08x\n",
-		       chid, inst << 12, subc, mthd, data);
+		pr_cont("] ch %d [0x%010llx %s] subc %d mthd 0x%04x data 0x%08x\n",
+		       chid, inst << 12, nouveau_client_name(engctx), subc,
+		       mthd, data);
 		nv_wo32(falcon, 0x004, 0x00000040);
 		stat &= ~0x00000040;
 	}
diff --git a/drivers/gpu/drm/nouveau/core/engine/crypt/nv84.c b/drivers/gpu/drm/nouveau/core/engine/crypt/nv84.c
index 8d0a440..f431759 100644
--- a/drivers/gpu/drm/nouveau/core/engine/crypt/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/crypt/nv84.c
@@ -22,6 +22,7 @@
  * Authors: Ben Skeggs
  */
 
+#include <core/client.h>
 #include <core/os.h>
 #include <core/enum.h>
 #include <core/class.h>
@@ -128,8 +129,9 @@ nv84_crypt_intr(struct nouveau_subdev *subdev)
 	if (stat) {
 		nv_error(priv, "");
 		nouveau_bitfield_print(nv84_crypt_intr_mask, stat);
-		pr_cont(" ch %d [0x%010llx] mthd 0x%04x data 0x%08x\n",
-		       chid, (u64)inst << 12, mthd, data);
+		pr_cont(" ch %d [0x%010llx %s] mthd 0x%04x data 0x%08x\n",
+		       chid, (u64)inst << 12, nouveau_client_name(engctx),
+		       mthd, data);
 	}
 
 	nv_wr32(priv, 0x102130, stat);
diff --git a/drivers/gpu/drm/nouveau/core/engine/crypt/nv98.c b/drivers/gpu/drm/nouveau/core/engine/crypt/nv98.c
index 1b33c20..8bf8955 100644
--- a/drivers/gpu/drm/nouveau/core/engine/crypt/nv98.c
+++ b/drivers/gpu/drm/nouveau/core/engine/crypt/nv98.c
@@ -22,6 +22,7 @@
  * Authors: Ben Skeggs
  */
 
+#include <core/client.h>
 #include <core/os.h>
 #include <core/enum.h>
 #include <core/class.h>
@@ -102,8 +103,9 @@ nv98_crypt_intr(struct nouveau_subdev *subdev)
 	if (stat & 0x00000040) {
 		nv_error(priv, "DISPATCH_ERROR [");
 		nouveau_enum_print(nv98_crypt_isr_error_name, ssta);
-		pr_cont("] ch %d [0x%010llx] subc %d mthd 0x%04x data 0x%08x\n",
-		       chid, (u64)inst << 12, subc, mthd, data);
+		pr_cont("] ch %d [0x%010llx %s] subc %d mthd 0x%04x data 0x%08x\n",
+		       chid, (u64)inst << 12, nouveau_client_name(engctx),
+		       subc, mthd, data);
 		nv_wr32(priv, 0x087004, 0x00000040);
 		stat &= ~0x00000040;
 	}
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nv04.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nv04.c
index f6c4e71..e34ab40 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nv04.c
@@ -421,9 +421,12 @@ nv04_fifo_cache_error(struct nouveau_device *device,
 	}
 
 	if (!nv04_fifo_swmthd(priv, chid, mthd, data)) {
+		const char *client_name =
+			nouveau_client_name_for_fifo_chid(&priv->base, chid);
 		nv_error(priv,
-			 "CACHE_ERROR - Ch %d/%d Mthd 0x%04x Data 0x%08x\n",
-			 chid, (mthd >> 13) & 7, mthd & 0x1ffc, data);
+			 "CACHE_ERROR - ch %d [%s] subc %d mthd 0x%04x data 0x%08x\n",
+			 chid, client_name, (mthd >> 13) & 7, mthd & 0x1ffc,
+			 data);
 	}
 
 	nv_wr32(priv, NV04_PFIFO_CACHE1_DMA_PUSH, 0);
@@ -445,11 +448,14 @@ static void
 nv04_fifo_dma_pusher(struct nouveau_device *device, struct nv04_fifo_priv *priv,
 		u32 chid)
 {
+	const char *client_name;
 	u32 dma_get = nv_rd32(priv, 0x003244);
 	u32 dma_put = nv_rd32(priv, 0x003240);
 	u32 push = nv_rd32(priv, 0x003220);
 	u32 state = nv_rd32(priv, 0x003228);
 
+	client_name = nouveau_client_name_for_fifo_chid(&priv->base, chid);
+
 	if (device->card_type == NV_50) {
 		u32 ho_get = nv_rd32(priv, 0x003328);
 		u32 ho_put = nv_rd32(priv, 0x003320);
@@ -457,9 +463,9 @@ nv04_fifo_dma_pusher(struct nouveau_device *device, struct nv04_fifo_priv *priv,
 		u32 ib_put = nv_rd32(priv, 0x003330);
 
 		nv_error(priv,
-			 "DMA_PUSHER - Ch %d Get 0x%02x%08x Put 0x%02x%08x IbGet 0x%08x IbPut 0x%08x State 0x%08x (err: %s) Push 0x%08x\n",
-			 chid, ho_get, dma_get, ho_put, dma_put, ib_get, ib_put,
-			 state, nv_dma_state_err(state), push);
+			 "DMA_PUSHER - ch %d [%s] get 0x%02x%08x put 0x%02x%08x ib_get 0x%08x ib_put 0x%08x state 0x%08x (err: %s) push 0x%08x\n",
+			 chid, client_name, ho_get, dma_get, ho_put, dma_put,
+			 ib_get, ib_put, state, nv_dma_state_err(state), push);
 
 		/* METHOD_COUNT, in DMA_STATE on earlier chipsets */
 		nv_wr32(priv, 0x003364, 0x00000000);
@@ -471,9 +477,9 @@ nv04_fifo_dma_pusher(struct nouveau_device *device, struct nv04_fifo_priv *priv,
 			nv_wr32(priv, 0x003334, ib_put);
 	} else {
 		nv_error(priv,
-			 "DMA_PUSHER - Ch %d Get 0x%08x Put 0x%08x State 0x%08x (err: %s) Push 0x%08x\n",
-			 chid, dma_get, dma_put, state, nv_dma_state_err(state),
-			 push);
+			 "DMA_PUSHER - ch %d [%s] get 0x%08x put 0x%08x state 0x%08x (err: %s) push 0x%08x\n",
+			 chid, client_name, dma_get, dma_put, state,
+			 nv_dma_state_err(state), push);
 
 		if (dma_get != dma_put)
 			nv_wr32(priv, 0x003244, dma_put);
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nv50.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nv50.c
index bd09636..cfea451 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nv50.c
@@ -129,7 +129,8 @@ nv50_fifo_context_detach(struct nouveau_object *parent, bool suspend,
 	/* do the kickoff... */
 	nv_wr32(priv, 0x0032fc, nv_gpuobj(base)->addr >> 12);
 	if (!nv_wait_ne(priv, 0x0032fc, 0xffffffff, 0xffffffff)) {
-		nv_error(priv, "channel %d unload timeout\n", chan->base.chid);
+		nv_error(priv, "channel %d [%s] unload timeout\n",
+			 chan->base.chid, nouveau_client_name(chan));
 		if (suspend)
 			ret = -EBUSY;
 	}
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nv84.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nv84.c
index 1eb1c51..8ae745b 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nv84.c
@@ -100,7 +100,8 @@ nv84_fifo_context_detach(struct nouveau_object *parent, bool suspend,
 	done = nv_wait_ne(priv, 0x0032fc, 0xffffffff, 0xffffffff);
 	nv_wr32(priv, 0x002520, save);
 	if (!done) {
-		nv_error(priv, "channel %d unload timeout\n", chan->base.chid);
+		nv_error(priv, "channel %d [%s] unload timeout\n",
+			 chan->base.chid, nouveau_client_name(chan));
 		if (suspend)
 			return -EBUSY;
 	}
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
index d341475..1580aae 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
@@ -149,7 +149,8 @@ nvc0_fifo_context_detach(struct nouveau_object *parent, bool suspend,
 
 	nv_wr32(priv, 0x002634, chan->base.chid);
 	if (!nv_wait(priv, 0x002634, 0xffffffff, chan->base.chid)) {
-		nv_error(priv, "channel %d kick timeout\n", chan->base.chid);
+		nv_error(priv, "channel %d [%s] kick timeout\n",
+			 chan->base.chid, nouveau_client_name(chan));
 		if (suspend)
 			return -EBUSY;
 	}
@@ -333,17 +334,17 @@ nvc0_fifo_cclass = {
  ******************************************************************************/
 
 static const struct nouveau_enum nvc0_fifo_fault_unit[] = {
-	{ 0x00, "PGRAPH" },
+	{ 0x00, "PGRAPH", NULL, NVDEV_ENGINE_GR },
 	{ 0x03, "PEEPHOLE" },
 	{ 0x04, "BAR1" },
 	{ 0x05, "BAR3" },
-	{ 0x07, "PFIFO" },
-	{ 0x10, "PBSP" },
-	{ 0x11, "PPPP" },
+	{ 0x07, "PFIFO", NULL, NVDEV_ENGINE_FIFO },
+	{ 0x10, "PBSP", NULL, NVDEV_ENGINE_BSP },
+	{ 0x11, "PPPP", NULL, NVDEV_ENGINE_PPP },
 	{ 0x13, "PCOUNTER" },
-	{ 0x14, "PVP" },
-	{ 0x15, "PCOPY0" },
-	{ 0x16, "PCOPY1" },
+	{ 0x14, "PVP", NULL, NVDEV_ENGINE_VP },
+	{ 0x15, "PCOPY0", NULL, NVDEV_ENGINE_COPY0 },
+	{ 0x16, "PCOPY1", NULL, NVDEV_ENGINE_COPY1 },
 	{ 0x17, "PDAEMON" },
 	{}
 };
@@ -402,6 +403,9 @@ nvc0_fifo_isr_vm_fault(struct nvc0_fifo_priv *priv, int unit)
 	u32 vahi = nv_rd32(priv, 0x002808 + (unit * 0x10));
 	u32 stat = nv_rd32(priv, 0x00280c + (unit * 0x10));
 	u32 client = (stat & 0x00001f00) >> 8;
+	const struct nouveau_enum *en;
+	struct nouveau_engine *engine;
+	struct nouveau_object *engctx = NULL;
 
 	switch (unit) {
 	case 3: /* PEEPHOLE */
@@ -421,7 +425,7 @@ nvc0_fifo_isr_vm_fault(struct nvc0_fifo_priv *priv, int unit)
 		 "write" : "read", (u64)vahi << 32 | valo);
 	nouveau_enum_print(nvc0_fifo_fault_reason, stat & 0x0000000f);
 	pr_cont("] from ");
-	nouveau_enum_print(nvc0_fifo_fault_unit, unit);
+	en = nouveau_enum_print(nvc0_fifo_fault_unit, unit);
 	if (stat & 0x00000040) {
 		pr_cont("/");
 		nouveau_enum_print(nvc0_fifo_fault_hubclient, client);
@@ -429,7 +433,17 @@ nvc0_fifo_isr_vm_fault(struct nvc0_fifo_priv *priv, int unit)
 		pr_cont("/GPC%d/", (stat & 0x1f000000) >> 24);
 		nouveau_enum_print(nvc0_fifo_fault_gpcclient, client);
 	}
-	pr_cont(" on channel 0x%010llx\n", (u64)inst << 12);
+
+	if (en && en->data2) {
+		engine = nouveau_engine(priv, en->data2);
+		if (engine)
+			engctx = nouveau_engctx_get(engine, inst);
+
+	}
+	pr_cont(" on channel 0x%010llx [%s]\n", (u64)inst << 12,
+			nouveau_client_name(engctx));
+
+	nouveau_engctx_put(engctx);
 }
 
 static int
@@ -485,9 +499,11 @@ nvc0_fifo_isr_subfifo_intr(struct nvc0_fifo_priv *priv, int unit)
 		nv_error(priv, "SUBFIFO%d:", unit);
 		nouveau_bitfield_print(nvc0_fifo_subfifo_intr, show);
 		pr_cont("\n");
-		nv_error(priv, "SUBFIFO%d: ch %d subc %d mthd 0x%04x "
-			       "data 0x%08x\n",
-			 unit, chid, subc, mthd, data);
+		nv_error(priv,
+			 "SUBFIFO%d: ch %d [%s] subc %d mthd 0x%04x data 0x%08x\n",
+			 unit, chid,
+			 nouveau_client_name_for_fifo_chid(&priv->base, chid),
+			 subc, mthd, data);
 	}
 
 	nv_wr32(priv, 0x0400c0 + (unit * 0x2000), 0x80600008);
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
index c874c24..410c6e2 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
@@ -184,7 +184,8 @@ nve0_fifo_context_detach(struct nouveau_object *parent, bool suspend,
 
 	nv_wr32(priv, 0x002634, chan->base.chid);
 	if (!nv_wait(priv, 0x002634, 0xffffffff, chan->base.chid)) {
-		nv_error(priv, "channel %d kick timeout\n", chan->base.chid);
+		nv_error(priv, "channel %d [%s] kick timeout\n",
+			 chan->base.chid, nouveau_client_name(chan));
 		if (suspend)
 			return -EBUSY;
 	}
@@ -412,12 +413,15 @@ nve0_fifo_isr_vm_fault(struct nve0_fifo_priv *priv, int unit)
 	u32 vahi = nv_rd32(priv, 0x2808 + (unit * 0x10));
 	u32 stat = nv_rd32(priv, 0x280c + (unit * 0x10));
 	u32 client = (stat & 0x00001f00) >> 8;
+	const struct nouveau_enum *en;
+	struct nouveau_engine *engine;
+	struct nouveau_object *engctx = NULL;
 
 	nv_error(priv, "PFIFO: %s fault at 0x%010llx [", (stat & 0x00000080) ?
 		       "write" : "read", (u64)vahi << 32 | valo);
 	nouveau_enum_print(nve0_fifo_fault_reason, stat & 0x0000000f);
 	pr_cont("] from ");
-	nouveau_enum_print(nve0_fifo_fault_unit, unit);
+	en = nouveau_enum_print(nve0_fifo_fault_unit, unit);
 	if (stat & 0x00000040) {
 		pr_cont("/");
 		nouveau_enum_print(nve0_fifo_fault_hubclient, client);
@@ -425,7 +429,18 @@ nve0_fifo_isr_vm_fault(struct nve0_fifo_priv *priv, int unit)
 		pr_cont("/GPC%d/", (stat & 0x1f000000) >> 24);
 		nouveau_enum_print(nve0_fifo_fault_gpcclient, client);
 	}
-	pr_cont(" on channel 0x%010llx\n", (u64)inst << 12);
+
+	if (en && en->data2) {
+		engine = nouveau_engine(priv, en->data2);
+		if (engine)
+			engctx = nouveau_engctx_get(engine, inst);
+
+	}
+
+	pr_cont(" on channel 0x%010llx [%s]\n", (u64)inst << 12,
+			nouveau_client_name(engctx));
+
+	nouveau_engctx_put(engctx);
 }
 
 static int
@@ -481,9 +496,11 @@ nve0_fifo_isr_subfifo_intr(struct nve0_fifo_priv *priv, int unit)
 		nv_error(priv, "SUBFIFO%d:", unit);
 		nouveau_bitfield_print(nve0_fifo_subfifo_intr, show);
 		pr_cont("\n");
-		nv_error(priv, "SUBFIFO%d: ch %d subc %d mthd 0x%04x "
-			       "data 0x%08x\n",
-			 unit, chid, subc, mthd, data);
+		nv_error(priv,
+			 "SUBFIFO%d: ch %d [%s] subc %d mthd 0x%04x data 0x%08x\n",
+			 unit, chid,
+			 nouveau_client_name_for_fifo_chid(&priv->base, chid),
+			 subc, mthd, data);
 	}
 
 	nv_wr32(priv, 0x0400c0 + (unit * 0x2000), 0x80600008);
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv04.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv04.c
index f4fb844..2ea3d14 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv04.c
@@ -22,6 +22,7 @@
  * DEALINGS IN THE SOFTWARE.
  */
 
+#include <core/client.h>
 #include <core/os.h>
 #include <core/class.h>
 #include <core/handle.h>
@@ -1304,9 +1305,10 @@ nv04_graph_intr(struct nouveau_subdev *subdev)
 		pr_cont(" nstatus:");
 		nouveau_bitfield_print(nv04_graph_nstatus, nstatus);
 		pr_cont("\n");
-		nv_error(priv, "ch %d/%d class 0x%04x "
-			       "mthd 0x%04x data 0x%08x\n",
-			 chid, subc, class, mthd, data);
+		nv_error(priv,
+			 "ch %d [%s] subc %d class 0x%04x mthd 0x%04x data 0x%08x\n",
+			 chid, nouveau_client_name(chan), subc, class, mthd,
+			 data);
 	}
 
 	nouveau_namedb_put(handle);
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv10.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv10.c
index deea80f..e5afc8f 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv10.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv10.c
@@ -22,6 +22,7 @@
  * DEALINGS IN THE SOFTWARE.
  */
 
+#include <core/client.h>
 #include <core/os.h>
 #include <core/class.h>
 #include <core/handle.h>
@@ -1200,9 +1201,10 @@ nv10_graph_intr(struct nouveau_subdev *subdev)
 		pr_cont(" nstatus:");
 		nouveau_bitfield_print(nv10_graph_nstatus, nstatus);
 		pr_cont("\n");
-		nv_error(priv, "ch %d/%d class 0x%04x "
-			       "mthd 0x%04x data 0x%08x\n",
-			 chid, subc, class, mthd, data);
+		nv_error(priv,
+			 "ch %d [%s] subc %d class 0x%04x mthd 0x%04x data 0x%08x\n",
+			 chid, nouveau_client_name(chan), subc, class, mthd,
+			 data);
 	}
 
 	nouveau_namedb_put(handle);
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv20.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv20.c
index 4c0d79c5..8cc01c1 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv20.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv20.c
@@ -1,3 +1,4 @@
+#include <core/client.h>
 #include <core/os.h>
 #include <core/class.h>
 #include <core/engctx.h>
@@ -231,8 +232,10 @@ nv20_graph_intr(struct nouveau_subdev *subdev)
 		pr_cont(" nstatus:");
 		nouveau_bitfield_print(nv10_graph_nstatus, nstatus);
 		pr_cont("\n");
-		nv_error(priv, "ch %d/%d class 0x%04x mthd 0x%04x data 0x%08x\n",
-			chid, subc, class, mthd, data);
+		nv_error(priv,
+			 "ch %d [%s] subc %d class 0x%04x mthd 0x%04x data 0x%08x\n",
+			 chid, nouveau_client_name(engctx), subc, class, mthd,
+			 data);
 	}
 
 	nouveau_engctx_put(engctx);
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv40.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv40.c
index 961a073..81f0f5d 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv40.c
@@ -22,6 +22,7 @@
  * Authors: Ben Skeggs
  */
 
+#include <core/client.h>
 #include <core/os.h>
 #include <core/class.h>
 #include <core/handle.h>
@@ -328,9 +329,10 @@ nv40_graph_intr(struct nouveau_subdev *subdev)
 		pr_cont(" nstatus:");
 		nouveau_bitfield_print(nv10_graph_nstatus, nstatus);
 		pr_cont("\n");
-		nv_error(priv, "ch %d [0x%08x] subc %d class 0x%04x "
-			       "mthd 0x%04x data 0x%08x\n",
-			 chid, inst << 4, subc, class, mthd, data);
+		nv_error(priv,
+			 "ch %d [0x%08x %s] subc %d class 0x%04x mthd 0x%04x data 0x%08x\n",
+			 chid, inst << 4, nouveau_client_name(engctx), subc,
+			 class, mthd, data);
 	}
 
 	nouveau_engctx_put(engctx);
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c
index 2c9e9ad..c1c14d8 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c
@@ -24,6 +24,7 @@
 
 #include <core/os.h>
 #include <core/class.h>
+#include <core/client.h>
 #include <core/handle.h>
 #include <core/engctx.h>
 #include <core/enum.h>
@@ -532,7 +533,7 @@ nv50_priv_tp_trap(struct nv50_graph_priv *priv, int type, u32 ustatus_old,
 
 static int
 nv50_graph_trap_handler(struct nv50_graph_priv *priv, u32 display,
-			int chid, u64 inst)
+			int chid, u64 inst, struct nouveau_object *engctx)
 {
 	u32 status = nv_rd32(priv, 0x400108);
 	u32 ustatus;
@@ -565,12 +566,11 @@ nv50_graph_trap_handler(struct nv50_graph_priv *priv, u32 display,
 
 			nv_error(priv, "TRAP DISPATCH_FAULT\n");
 			if (display && (addr & 0x80000000)) {
-				nv_error(priv, "ch %d [0x%010llx] "
-					     "subc %d class 0x%04x mthd 0x%04x "
-					     "data 0x%08x%08x "
-					     "400808 0x%08x 400848 0x%08x\n",
-					chid, inst, subc, class, mthd, datah,
-					datal, addr, r848);
+				nv_error(priv,
+					 "ch %d [0x%010llx %s] subc %d class 0x%04x mthd 0x%04x data 0x%08x%08x 400808 0x%08x 400848 0x%08x\n",
+					 chid, inst,
+					 nouveau_client_name(engctx), subc,
+					 class, mthd, datah, datal, addr, r848);
 			} else
 			if (display) {
 				nv_error(priv, "no stuck command?\n");
@@ -591,11 +591,11 @@ nv50_graph_trap_handler(struct nv50_graph_priv *priv, u32 display,
 
 			nv_error(priv, "TRAP DISPATCH_QUERY\n");
 			if (display && (addr & 0x80000000)) {
-				nv_error(priv, "ch %d [0x%010llx] "
-					     "subc %d class 0x%04x mthd 0x%04x "
-					     "data 0x%08x 40084c 0x%08x\n",
-					chid, inst, subc, class, mthd,
-					data, addr);
+				nv_error(priv,
+					 "ch %d [0x%010llx %s] subc %d class 0x%04x mthd 0x%04x data 0x%08x 40084c 0x%08x\n",
+					 chid, inst,
+					 nouveau_client_name(engctx), subc,
+					 class, mthd, data, addr);
 			} else
 			if (display) {
 				nv_error(priv, "no stuck command?\n");
@@ -778,7 +778,8 @@ nv50_graph_intr(struct nouveau_subdev *subdev)
 	}
 
 	if (stat & 0x00200000) {
-		if (!nv50_graph_trap_handler(priv, show, chid, (u64)inst << 12))
+		if (!nv50_graph_trap_handler(priv, show, chid, (u64)inst << 12,
+				engctx))
 			show &= ~0x00200000;
 	}
 
@@ -789,9 +790,10 @@ nv50_graph_intr(struct nouveau_subdev *subdev)
 		nv_error(priv, "");
 		nouveau_bitfield_print(nv50_graph_intr_name, show);
 		pr_cont("\n");
-		nv_error(priv, "ch %d [0x%010llx] subc %d class 0x%04x "
-			       "mthd 0x%04x data 0x%08x\n",
-			 chid, (u64)inst << 12, subc, class, mthd, data);
+		nv_error(priv,
+			 "ch %d [0x%010llx %s] subc %d class 0x%04x mthd 0x%04x data 0x%08x\n",
+			 chid, (u64)inst << 12, nouveau_client_name(engctx),
+			 subc, class, mthd, data);
 	}
 
 	if (nv_rd32(priv, 0x400824) & (1 << 31))
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c
index e644298..15bfe55 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c
@@ -433,10 +433,10 @@ nvc0_graph_intr(struct nouveau_subdev *subdev)
 	if (stat & 0x00000010) {
 		handle = nouveau_handle_get_class(engctx, class);
 		if (!handle || nv_call(handle->object, mthd, data)) {
-			nv_error(priv, "ILLEGAL_MTHD ch %d [0x%010llx] "
-				     "subc %d class 0x%04x mthd 0x%04x "
-				     "data 0x%08x\n",
-				 chid, inst << 12, subc, class, mthd, data);
+			nv_error(priv,
+				 "ILLEGAL_MTHD ch %d [0x%010llx %s] subc %d class 0x%04x mthd 0x%04x data 0x%08x\n",
+				 chid, inst << 12, nouveau_client_name(engctx),
+				 subc, class, mthd, data);
 		}
 		nouveau_handle_put(handle);
 		nv_wr32(priv, 0x400100, 0x00000010);
@@ -444,9 +444,10 @@ nvc0_graph_intr(struct nouveau_subdev *subdev)
 	}
 
 	if (stat & 0x00000020) {
-		nv_error(priv, "ILLEGAL_CLASS ch %d [0x%010llx] subc %d "
-			     "class 0x%04x mthd 0x%04x data 0x%08x\n",
-			chid, inst << 12, subc, class, mthd, data);
+		nv_error(priv,
+			 "ILLEGAL_CLASS ch %d [0x%010llx %s] subc %d class 0x%04x mthd 0x%04x data 0x%08x\n",
+			 chid, inst << 12, nouveau_client_name(engctx), subc,
+			 class, mthd, data);
 		nv_wr32(priv, 0x400100, 0x00000020);
 		stat &= ~0x00000020;
 	}
@@ -454,15 +455,16 @@ nvc0_graph_intr(struct nouveau_subdev *subdev)
 	if (stat & 0x00100000) {
 		nv_error(priv, "DATA_ERROR [");
 		nouveau_enum_print(nv50_data_error_names, code);
-		pr_cont("] ch %d [0x%010llx] subc %d class 0x%04x "
-		       "mthd 0x%04x data 0x%08x\n",
-		       chid, inst << 12, subc, class, mthd, data);
+		pr_cont("] ch %d [0x%010llx %s] subc %d class 0x%04x mthd 0x%04x data 0x%08x\n",
+			chid, inst << 12, nouveau_client_name(engctx), subc,
+			class, mthd, data);
 		nv_wr32(priv, 0x400100, 0x00100000);
 		stat &= ~0x00100000;
 	}
 
 	if (stat & 0x00200000) {
-		nv_error(priv, "TRAP ch %d [0x%010llx]\n", chid, inst << 12);
+		nv_error(priv, "TRAP ch %d [0x%010llx %s]\n", chid, inst << 12,
+			 nouveau_client_name(engctx));
 		nvc0_graph_trap_intr(priv);
 		nv_wr32(priv, 0x400100, 0x00200000);
 		stat &= ~0x00200000;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nve0.c b/drivers/gpu/drm/nouveau/core/engine/graph/nve0.c
index 0f35a34..61cec0f 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nve0.c
@@ -78,15 +78,16 @@ nve0_graph_ctxctl_isr(struct nvc0_graph_priv *priv)
 }
 
 static void
-nve0_graph_trap_isr(struct nvc0_graph_priv *priv, int chid, u64 inst)
+nve0_graph_trap_isr(struct nvc0_graph_priv *priv, int chid, u64 inst,
+		struct nouveau_object *engctx)
 {
 	u32 trap = nv_rd32(priv, 0x400108);
 	int rop;
 
 	if (trap & 0x00000001) {
 		u32 stat = nv_rd32(priv, 0x404000);
-		nv_error(priv, "DISPATCH ch %d [0x%010llx] 0x%08x\n",
-			 chid, inst, stat);
+		nv_error(priv, "DISPATCH ch %d [0x%010llx %s] 0x%08x\n",
+			 chid, inst, nouveau_client_name(engctx), stat);
 		nv_wr32(priv, 0x404000, 0xc0000000);
 		nv_wr32(priv, 0x400108, 0x00000001);
 		trap &= ~0x00000001;
@@ -94,8 +95,8 @@ nve0_graph_trap_isr(struct nvc0_graph_priv *priv, int chid, u64 inst)
 
 	if (trap & 0x00000010) {
 		u32 stat = nv_rd32(priv, 0x405840);
-		nv_error(priv, "SHADER ch %d [0x%010llx] 0x%08x\n",
-			 chid, inst, stat);
+		nv_error(priv, "SHADER ch %d [0x%010llx %s] 0x%08x\n",
+			 chid, inst, nouveau_client_name(engctx), stat);
 		nv_wr32(priv, 0x405840, 0xc0000000);
 		nv_wr32(priv, 0x400108, 0x00000010);
 		trap &= ~0x00000010;
@@ -105,8 +106,10 @@ nve0_graph_trap_isr(struct nvc0_graph_priv *priv, int chid, u64 inst)
 		for (rop = 0; rop < priv->rop_nr; rop++) {
 			u32 statz = nv_rd32(priv, ROP_UNIT(rop, 0x070));
 			u32 statc = nv_rd32(priv, ROP_UNIT(rop, 0x144));
-			nv_error(priv, "ROP%d ch %d [0x%010llx] 0x%08x 0x%08x\n",
-				 rop, chid, inst, statz, statc);
+			nv_error(priv,
+				 "ROP%d ch %d [0x%010llx %s] 0x%08x 0x%08x\n",
+				 rop, chid, inst, nouveau_client_name(engctx),
+				 statz, statc);
 			nv_wr32(priv, ROP_UNIT(rop, 0x070), 0xc0000000);
 			nv_wr32(priv, ROP_UNIT(rop, 0x144), 0xc0000000);
 		}
@@ -115,8 +118,8 @@ nve0_graph_trap_isr(struct nvc0_graph_priv *priv, int chid, u64 inst)
 	}
 
 	if (trap) {
-		nv_error(priv, "TRAP ch %d [0x%010llx] 0x%08x\n",
-			 chid, inst, trap);
+		nv_error(priv, "TRAP ch %d [0x%010llx %s] 0x%08x\n",
+			 chid, inst, nouveau_client_name(engctx), trap);
 		nv_wr32(priv, 0x400108, trap);
 	}
 }
@@ -145,10 +148,10 @@ nve0_graph_intr(struct nouveau_subdev *subdev)
 	if (stat & 0x00000010) {
 		handle = nouveau_handle_get_class(engctx, class);
 		if (!handle || nv_call(handle->object, mthd, data)) {
-			nv_error(priv, "ILLEGAL_MTHD ch %d [0x%010llx] "
-				     "subc %d class 0x%04x mthd 0x%04x "
-				     "data 0x%08x\n",
-				 chid, inst, subc, class, mthd, data);
+			nv_error(priv,
+				 "ILLEGAL_MTHD ch %d [0x%010llx %s] subc %d class 0x%04x mthd 0x%04x data 0x%08x\n",
+				 chid, inst, nouveau_client_name(engctx), subc,
+				 class, mthd, data);
 		}
 		nouveau_handle_put(handle);
 		nv_wr32(priv, 0x400100, 0x00000010);
@@ -156,9 +159,10 @@ nve0_graph_intr(struct nouveau_subdev *subdev)
 	}
 
 	if (stat & 0x00000020) {
-		nv_error(priv, "ILLEGAL_CLASS ch %d [0x%010llx] subc %d "
-			     "class 0x%04x mthd 0x%04x data 0x%08x\n",
-			 chid, inst, subc, class, mthd, data);
+		nv_error(priv,
+			 "ILLEGAL_CLASS ch %d [0x%010llx %s] subc %d class 0x%04x mthd 0x%04x data 0x%08x\n",
+			 chid, inst, nouveau_client_name(engctx), subc, class,
+			 mthd, data);
 		nv_wr32(priv, 0x400100, 0x00000020);
 		stat &= ~0x00000020;
 	}
@@ -166,15 +170,15 @@ nve0_graph_intr(struct nouveau_subdev *subdev)
 	if (stat & 0x00100000) {
 		nv_error(priv, "DATA_ERROR [");
 		nouveau_enum_print(nv50_data_error_names, code);
-		pr_cont("] ch %d [0x%010llx] subc %d class 0x%04x "
-		       "mthd 0x%04x data 0x%08x\n",
-		       chid, inst, subc, class, mthd, data);
+		pr_cont("] ch %d [0x%010llx %s] subc %d class 0x%04x mthd 0x%04x data 0x%08x\n",
+			chid, inst, nouveau_client_name(engctx), subc, class,
+			mthd, data);
 		nv_wr32(priv, 0x400100, 0x00100000);
 		stat &= ~0x00100000;
 	}
 
 	if (stat & 0x00200000) {
-		nve0_graph_trap_isr(priv, chid, inst);
+		nve0_graph_trap_isr(priv, chid, inst, engctx);
 		nv_wr32(priv, 0x400100, 0x00200000);
 		stat &= ~0x00200000;
 	}
diff --git a/drivers/gpu/drm/nouveau/core/engine/mpeg/nv31.c b/drivers/gpu/drm/nouveau/core/engine/mpeg/nv31.c
index 9fd8637..49ecbb8 100644
--- a/drivers/gpu/drm/nouveau/core/engine/mpeg/nv31.c
+++ b/drivers/gpu/drm/nouveau/core/engine/mpeg/nv31.c
@@ -22,6 +22,7 @@
  * Authors: Ben Skeggs
  */
 
+#include <core/client.h>
 #include <core/os.h>
 #include <core/class.h>
 #include <core/engctx.h>
@@ -231,8 +232,10 @@ nv31_mpeg_intr(struct nouveau_subdev *subdev)
 	nv_wr32(priv, 0x00b230, 0x00000001);
 
 	if (show) {
-		nv_error(priv, "ch %d [0x%08x] 0x%08x 0x%08x 0x%08x 0x%08x\n",
-			 chid, inst << 4, stat, type, mthd, data);
+		nv_error(priv,
+			 "ch %d [0x%08x %s] 0x%08x 0x%08x 0x%08x 0x%08x\n",
+			 chid, inst << 4, nouveau_client_name(engctx), stat,
+			 type, mthd, data);
 	}
 
 	nouveau_engctx_put(engctx);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c
index 88bdb02..a4338d9 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c
@@ -22,8 +22,10 @@
  * Authors: Ben Skeggs
  */
 
-#include <core/object.h>
+#include <core/client.h>
 #include <core/enum.h>
+#include <core/engctx.h>
+#include <core/object.h>
 
 #include <subdev/fb.h>
 #include <subdev/bios.h>
@@ -302,17 +304,18 @@ static const struct nouveau_enum vm_client[] = {
 };
 
 static const struct nouveau_enum vm_engine[] = {
-	{ 0x00000000, "PGRAPH", NULL },
-	{ 0x00000001, "PVP", NULL },
+	{ 0x00000000, "PGRAPH", NULL, NVDEV_ENGINE_GR },
+	{ 0x00000001, "PVP", NULL, NVDEV_ENGINE_VP },
 	{ 0x00000004, "PEEPHOLE", NULL },
-	{ 0x00000005, "PFIFO", vm_pfifo_subclients },
+	{ 0x00000005, "PFIFO", vm_pfifo_subclients, NVDEV_ENGINE_FIFO },
 	{ 0x00000006, "BAR", vm_bar_subclients },
-	{ 0x00000008, "PPPP", NULL },
-	{ 0x00000009, "PBSP", NULL },
-	{ 0x0000000a, "PCRYPT", NULL },
+	{ 0x00000008, "PPPP", NULL, NVDEV_ENGINE_PPP },
+	{ 0x00000008, "PMPEG", NULL, NVDEV_ENGINE_MPEG },
+	{ 0x00000009, "PBSP", NULL, NVDEV_ENGINE_BSP },
+	{ 0x0000000a, "PCRYPT", NULL, NVDEV_ENGINE_CRYPT },
 	{ 0x0000000b, "PCOUNTER", NULL },
 	{ 0x0000000c, "SEMAPHORE_BG", NULL },
-	{ 0x0000000d, "PCOPY", NULL },
+	{ 0x0000000d, "PCOPY", NULL, NVDEV_ENGINE_COPY0 },
 	{ 0x0000000e, "PDAEMON", NULL },
 	{}
 };
@@ -334,8 +337,10 @@ static void
 nv50_fb_intr(struct nouveau_subdev *subdev)
 {
 	struct nouveau_device *device = nv_device(subdev);
+	struct nouveau_engine *engine;
 	struct nv50_fb_priv *priv = (void *)subdev;
 	const struct nouveau_enum *en, *cl;
+	struct nouveau_object *engctx = NULL;
 	u32 trap[6], idx, chan;
 	u8 st0, st1, st2, st3;
 	int i;
@@ -366,11 +371,30 @@ nv50_fb_intr(struct nouveau_subdev *subdev)
 	}
 	chan = (trap[2] << 16) | trap[1];
 
-	nv_error(priv, "trapped %s at 0x%02x%04x%04x on channel 0x%08x ",
+	en = nouveau_enum_find(vm_engine, st0);
+
+	if (en && en->data2) {
+		const struct nouveau_enum *orig_en = en;
+		while (en->name && en->value == st0 && en->data2) {
+			engine = nouveau_engine(subdev, en->data2);
+			if (engine) {
+				engctx = nouveau_engctx_get(engine, chan);
+				if (engctx)
+					break;
+			}
+			en++;
+		}
+		if (!engctx)
+			en = orig_en;
+	}
+
+	nv_error(priv, "trapped %s at 0x%02x%04x%04x on channel 0x%08x [%s] ",
 		 (trap[5] & 0x00000100) ? "read" : "write",
-		 trap[5] & 0xff, trap[4] & 0xffff, trap[3] & 0xffff, chan);
+		 trap[5] & 0xff, trap[4] & 0xffff, trap[3] & 0xffff, chan,
+		 nouveau_client_name(engctx));
+
+	nouveau_engctx_put(engctx);
 
-	en = nouveau_enum_find(vm_engine, st0);
 	if (en)
 		pr_cont("%s/", en->name);
 	else
diff --git a/drivers/gpu/drm/nouveau/nouveau_chan.c b/drivers/gpu/drm/nouveau/nouveau_chan.c
index 174300b..525a517 100644
--- a/drivers/gpu/drm/nouveau/nouveau_chan.c
+++ b/drivers/gpu/drm/nouveau/nouveau_chan.c
@@ -58,7 +58,8 @@ nouveau_channel_idle(struct nouveau_channel *chan)
 	}
 
 	if (ret)
-		NV_ERROR(cli, "failed to idle channel 0x%08x\n", chan->handle);
+		NV_ERROR(cli, "failed to idle channel 0x%08x [%s]\n",
+			 chan->handle, cli->base.name);
 	return ret;
 }
 
-- 
cgit v0.10.2


From a624bafbf11c098a859dba4118a33605bfd25788 Mon Sep 17 00:00:00 2001
From: Martin Peres <martin.peres@labri.fr>
Date: Thu, 20 Dec 2012 01:32:09 +0100
Subject: drm/nouveau/fan: handle the cases where we are outside of the linear
 zone
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

This fixes a bug where, when temperature is outside of the linear range, fan
pwm would be outside of the allowed range ([0, 100]) and could get negative in
some cases.

It seems like a regression that happened when we re-worked the fan management
logic before merging.

Tested-by: Ozan Çağlayan <ozancag@gmail.com>
Signed-off-by: Martin Peres <martin.peres@labri.fr>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/base.c b/drivers/gpu/drm/nouveau/core/subdev/therm/base.c
index b35b4a2..25b7f6a 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/base.c
@@ -71,6 +71,13 @@ nouveau_therm_update_linear(struct nouveau_therm *therm)
 	u8  temp = therm->temp_get(therm);
 	u16 duty;
 
+	/* handle the non-linear part first */
+	if (temp < linear_min_temp)
+		return priv->fan->bios.min_duty;
+	else if (temp > linear_max_temp)
+		return priv->fan->bios.max_duty;
+
+	/* we are in the linear zone */
 	duty  = (temp - linear_min_temp);
 	duty *= (priv->fan->bios.max_duty - priv->fan->bios.min_duty);
 	duty /= (linear_max_temp - linear_min_temp);
-- 
cgit v0.10.2


From a4cea27b6950885a743a4a000d5f2ea3fd445d25 Mon Sep 17 00:00:00 2001
From: Marcin Slusarz <marcin.slusarz@gmail.com>
Date: Tue, 25 Dec 2012 17:50:43 +0100
Subject: drm/nouveau: share fence structures between nv10+ and nv50
 implementations

We already rely on them having the same fields and layout.

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nv10_fence.c b/drivers/gpu/drm/nouveau/nv10_fence.c
index 03017f2..a8ea4af 100644
--- a/drivers/gpu/drm/nouveau/nv10_fence.c
+++ b/drivers/gpu/drm/nouveau/nv10_fence.c
@@ -27,18 +27,7 @@
 
 #include "nouveau_drm.h"
 #include "nouveau_dma.h"
-#include "nouveau_fence.h"
-
-struct nv10_fence_chan {
-	struct nouveau_fence_chan base;
-};
-
-struct nv10_fence_priv {
-	struct nouveau_fence_priv base;
-	struct nouveau_bo *bo;
-	spinlock_t lock;
-	u32 sequence;
-};
+#include "nv10_fence.h"
 
 int
 nv10_fence_emit(struct nouveau_fence *fence)
diff --git a/drivers/gpu/drm/nouveau/nv10_fence.h b/drivers/gpu/drm/nouveau/nv10_fence.h
new file mode 100644
index 0000000..e5d9204
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nv10_fence.h
@@ -0,0 +1,19 @@
+#ifndef __NV10_FENCE_H_
+#define __NV10_FENCE_H_
+
+#include <core/os.h>
+#include "nouveau_fence.h"
+#include "nouveau_bo.h"
+
+struct nv10_fence_chan {
+	struct nouveau_fence_chan base;
+};
+
+struct nv10_fence_priv {
+	struct nouveau_fence_priv base;
+	struct nouveau_bo *bo;
+	spinlock_t lock;
+	u32 sequence;
+};
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/nv50_fence.c b/drivers/gpu/drm/nouveau/nv50_fence.c
index d889f3a..72791d6 100644
--- a/drivers/gpu/drm/nouveau/nv50_fence.c
+++ b/drivers/gpu/drm/nouveau/nv50_fence.c
@@ -27,27 +27,16 @@
 
 #include "nouveau_drm.h"
 #include "nouveau_dma.h"
-#include "nouveau_fence.h"
+#include "nv10_fence.h"
 
 #include "nv50_display.h"
 
-struct nv50_fence_chan {
-	struct nouveau_fence_chan base;
-};
-
-struct nv50_fence_priv {
-	struct nouveau_fence_priv base;
-	struct nouveau_bo *bo;
-	spinlock_t lock;
-	u32 sequence;
-};
-
 static int
 nv50_fence_context_new(struct nouveau_channel *chan)
 {
 	struct drm_device *dev = chan->drm->dev;
-	struct nv50_fence_priv *priv = chan->drm->fence;
-	struct nv50_fence_chan *fctx;
+	struct nv10_fence_priv *priv = chan->drm->fence;
+	struct nv10_fence_chan *fctx;
 	struct ttm_mem_reg *mem = &priv->bo->bo.mem;
 	struct nouveau_object *object;
 	int ret, i;
@@ -91,7 +80,7 @@ nv50_fence_context_new(struct nouveau_channel *chan)
 int
 nv50_fence_create(struct nouveau_drm *drm)
 {
-	struct nv50_fence_priv *priv;
+	struct nv10_fence_priv *priv;
 	int ret = 0;
 
 	priv = drm->fence = kzalloc(sizeof(*priv), GFP_KERNEL);
-- 
cgit v0.10.2


From 950fbfab425be54aaac76648ba28a4b6018762d8 Mon Sep 17 00:00:00 2001
From: Marcin Slusarz <marcin.slusarz@gmail.com>
Date: Sat, 29 Dec 2012 16:24:37 +0100
Subject: drm/nouveau: mark nv_printk_ as printf-like function

...and fix all warnings

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/engine/crypt/nv84.c b/drivers/gpu/drm/nouveau/core/engine/crypt/nv84.c
index f431759..5bc021f 100644
--- a/drivers/gpu/drm/nouveau/core/engine/crypt/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/crypt/nv84.c
@@ -127,7 +127,7 @@ nv84_crypt_intr(struct nouveau_subdev *subdev)
 	chid   = pfifo->chid(pfifo, engctx);
 
 	if (stat) {
-		nv_error(priv, "");
+		nv_error(priv, "%s", "");
 		nouveau_bitfield_print(nv84_crypt_intr_mask, stat);
 		pr_cont(" ch %d [0x%010llx %s] mthd 0x%04x data 0x%08x\n",
 		       chid, (u64)inst << 12, nouveau_client_name(engctx),
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv04.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv04.c
index 2ea3d14..ad13dcd 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv04.c
@@ -1298,7 +1298,7 @@ nv04_graph_intr(struct nouveau_subdev *subdev)
 	nv_wr32(priv, NV04_PGRAPH_FIFO, 0x00000001);
 
 	if (show) {
-		nv_error(priv, "");
+		nv_error(priv, "%s", "");
 		nouveau_bitfield_print(nv04_graph_intr_name, show);
 		pr_cont(" nsource:");
 		nouveau_bitfield_print(nv04_graph_nsource, nsource);
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv10.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv10.c
index e5afc8f..23c143a 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv10.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv10.c
@@ -1194,7 +1194,7 @@ nv10_graph_intr(struct nouveau_subdev *subdev)
 	nv_wr32(priv, NV04_PGRAPH_FIFO, 0x00000001);
 
 	if (show) {
-		nv_error(priv, "");
+		nv_error(priv, "%s", "");
 		nouveau_bitfield_print(nv10_graph_intr_name, show);
 		pr_cont(" nsource:");
 		nouveau_bitfield_print(nv04_graph_nsource, nsource);
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv20.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv20.c
index 8cc01c1..0607b98 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv20.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv20.c
@@ -225,7 +225,7 @@ nv20_graph_intr(struct nouveau_subdev *subdev)
 	nv_wr32(priv, NV04_PGRAPH_FIFO, 0x00000001);
 
 	if (show) {
-		nv_error(priv, "");
+		nv_error(priv, "%s", "");
 		nouveau_bitfield_print(nv10_graph_intr_name, show);
 		pr_cont(" nsource:");
 		nouveau_bitfield_print(nv04_graph_nsource, nsource);
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv40.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv40.c
index 81f0f5d..17049d5 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv40.c
@@ -322,7 +322,7 @@ nv40_graph_intr(struct nouveau_subdev *subdev)
 	nv_wr32(priv, NV04_PGRAPH_FIFO, 0x00000001);
 
 	if (show) {
-		nv_error(priv, "");
+		nv_error(priv, "%s", "");
 		nouveau_bitfield_print(nv10_graph_intr_name, show);
 		pr_cont(" nsource:");
 		nouveau_bitfield_print(nv04_graph_nsource, nsource);
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c
index c1c14d8..d51f29d 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c
@@ -787,7 +787,7 @@ nv50_graph_intr(struct nouveau_subdev *subdev)
 	nv_wr32(priv, 0x400500, 0x00010001);
 
 	if (show) {
-		nv_error(priv, "");
+		nv_error(priv, "%s", "");
 		nouveau_bitfield_print(nv50_graph_intr_name, show);
 		pr_cont("\n");
 		nv_error(priv,
diff --git a/drivers/gpu/drm/nouveau/core/include/core/object.h b/drivers/gpu/drm/nouveau/core/include/core/object.h
index 5982935..6a90267 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/object.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/object.h
@@ -133,7 +133,7 @@ static inline u8
 nv_ro08(void *obj, u64 addr)
 {
 	u8 data = nv_ofuncs(obj)->rd08(obj, addr);
-	nv_spam(obj, "nv_ro08 0x%08x 0x%02x\n", addr, data);
+	nv_spam(obj, "nv_ro08 0x%08llx 0x%02x\n", addr, data);
 	return data;
 }
 
@@ -141,7 +141,7 @@ static inline u16
 nv_ro16(void *obj, u64 addr)
 {
 	u16 data = nv_ofuncs(obj)->rd16(obj, addr);
-	nv_spam(obj, "nv_ro16 0x%08x 0x%04x\n", addr, data);
+	nv_spam(obj, "nv_ro16 0x%08llx 0x%04x\n", addr, data);
 	return data;
 }
 
@@ -149,28 +149,28 @@ static inline u32
 nv_ro32(void *obj, u64 addr)
 {
 	u32 data = nv_ofuncs(obj)->rd32(obj, addr);
-	nv_spam(obj, "nv_ro32 0x%08x 0x%08x\n", addr, data);
+	nv_spam(obj, "nv_ro32 0x%08llx 0x%08x\n", addr, data);
 	return data;
 }
 
 static inline void
 nv_wo08(void *obj, u64 addr, u8 data)
 {
-	nv_spam(obj, "nv_wo08 0x%08x 0x%02x\n", addr, data);
+	nv_spam(obj, "nv_wo08 0x%08llx 0x%02x\n", addr, data);
 	nv_ofuncs(obj)->wr08(obj, addr, data);
 }
 
 static inline void
 nv_wo16(void *obj, u64 addr, u16 data)
 {
-	nv_spam(obj, "nv_wo16 0x%08x 0x%04x\n", addr, data);
+	nv_spam(obj, "nv_wo16 0x%08llx 0x%04x\n", addr, data);
 	nv_ofuncs(obj)->wr16(obj, addr, data);
 }
 
 static inline void
 nv_wo32(void *obj, u64 addr, u32 data)
 {
-	nv_spam(obj, "nv_wo32 0x%08x 0x%08x\n", addr, data);
+	nv_spam(obj, "nv_wo32 0x%08llx 0x%08x\n", addr, data);
 	nv_ofuncs(obj)->wr32(obj, addr, data);
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/include/core/printk.h b/drivers/gpu/drm/nouveau/core/include/core/printk.h
index 1d62966..febed2e 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/printk.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/printk.h
@@ -15,7 +15,8 @@ struct nouveau_object;
 #define NV_PRINTK_TRACE    KERN_DEBUG
 #define NV_PRINTK_SPAM     KERN_DEBUG
 
-void nv_printk_(struct nouveau_object *, const char *, int, const char *, ...);
+void __printf(4, 5)
+nv_printk_(struct nouveau_object *, const char *, int, const char *, ...);
 
 #define nv_printk(o,l,f,a...) do {                                             \
 	if (NV_DBG_##l <= CONFIG_NOUVEAU_DEBUG)                                \
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/init.c b/drivers/gpu/drm/nouveau/core/subdev/bios/init.c
index 690ed43..c97c972 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/init.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/init.c
@@ -1866,7 +1866,7 @@ init_zm_reg_group(struct nvbios_init *init)
 	u32 addr = nv_ro32(bios, init->offset + 1);
 	u8 count = nv_ro08(bios, init->offset + 5);
 
-	trace("ZM_REG_GROUP\tR[0x%06x] =\n");
+	trace("ZM_REG_GROUP\tR[0x%06x] =\n", addr);
 	init->offset += 6;
 
 	while (count--) {
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/base.c b/drivers/gpu/drm/nouveau/core/subdev/device/base.c
index 6837213..3937ced 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/device/base.c
@@ -104,8 +104,8 @@ nouveau_devobj_ctor(struct nouveau_object *parent,
 	struct nouveau_device *device;
 	struct nouveau_devobj *devobj;
 	struct nv_device_class *args = data;
-	u64 disable, boot0, strap;
-	u64 mmio_base, mmio_size;
+	u32 boot0, strap;
+	u64 disable, mmio_base, mmio_size;
 	void __iomem *map;
 	int ret, i, c;
 
-- 
cgit v0.10.2


From bfd8bd1f81dfe6110af3fbd0ba11f8d5940b87dd Mon Sep 17 00:00:00 2001
From: Marcin Slusarz <marcin.slusarz@gmail.com>
Date: Sun, 30 Dec 2012 16:35:24 +0100
Subject: drm/nouveau/bios: tiny debugging messages fixes

COPY_ZM_REG: destination and source addresses were swapped
RAM_RESTRICT_ZM_REG_GROUP: missing 0x prefix for register address

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/init.c b/drivers/gpu/drm/nouveau/core/subdev/bios/init.c
index c97c972..7d2ce11 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/init.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/init.c
@@ -1816,7 +1816,7 @@ init_ram_restrict_zm_reg_group(struct nvbios_init *init)
 	u8 i, j;
 
 	trace("RAM_RESTRICT_ZM_REG_GROUP\t"
-	      "R[%08x] 0x%02x 0x%02x\n", addr, incr, num);
+	      "R[0x%08x] 0x%02x 0x%02x\n", addr, incr, num);
 	init->offset += 7;
 
 	for (i = 0; i < num; i++) {
@@ -1849,7 +1849,7 @@ init_copy_zm_reg(struct nvbios_init *init)
 	u32 sreg = nv_ro32(bios, init->offset + 1);
 	u32 dreg = nv_ro32(bios, init->offset + 5);
 
-	trace("COPY_ZM_REG\tR[0x%06x] = R[0x%06x]\n", sreg, dreg);
+	trace("COPY_ZM_REG\tR[0x%06x] = R[0x%06x]\n", dreg, sreg);
 	init->offset += 9;
 
 	init_wr32(init, dreg, init_rd32(init, sreg));
-- 
cgit v0.10.2


From cd897837eacc6ce0b883b5e6c9000cb2e5f11c39 Mon Sep 17 00:00:00 2001
From: Marcin Slusarz <marcin.slusarz@gmail.com>
Date: Sun, 27 Jan 2013 15:01:55 +0100
Subject: drm/nouveau: quiet static-related sparse noise

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/extdev.c b/drivers/gpu/drm/nouveau/core/subdev/bios/extdev.c
index 5afb568..b2a676e 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/extdev.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/extdev.c
@@ -48,7 +48,7 @@ extdev_table(struct nouveau_bios *bios, u8 *ver, u8 *hdr, u8 *len, u8 *cnt)
 	return extdev + *hdr;
 }
 
-u16
+static u16
 nvbios_extdev_entry(struct nouveau_bios *bios, int idx, u8 *ver, u8 *len)
 {
 	u8 hdr, cnt;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/therm.c b/drivers/gpu/drm/nouveau/core/subdev/bios/therm.c
index b7916a5..22a2057 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/therm.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/therm.c
@@ -55,7 +55,7 @@ therm_table(struct nouveau_bios *bios, u8 *ver, u8 *hdr, u8 *len, u8 *cnt)
 	return therm + nv_ro08(bios, therm + 1);
 }
 
-u16
+static u16
 nvbios_therm_entry(struct nouveau_bios *bios, int idx, u8 *ver, u8 *len)
 {
 	u8 hdr, cnt;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bus/nv04.c b/drivers/gpu/drm/nouveau/core/subdev/bus/nv04.c
index 0c224e3..8c7f805 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bus/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bus/nv04.c
@@ -29,7 +29,7 @@ struct nv04_bus_priv {
 	struct nouveau_bus base;
 };
 
-void
+static void
 nv04_bus_intr(struct nouveau_subdev *subdev)
 {
 	struct nouveau_bus *pbus = nouveau_bus(subdev);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
index ddb23dc..b179655 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
@@ -175,7 +175,7 @@ nouveau_therm_fan_user_set(struct nouveau_therm *therm, int percent)
 	return nouveau_therm_fan_set(therm, true, percent);
 }
 
-void
+static void
 nouveau_therm_fan_set_defaults(struct nouveau_therm *therm)
 {
 	struct nouveau_therm_priv *priv = (void *)therm;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/fannil.c b/drivers/gpu/drm/nouveau/core/subdev/therm/fannil.c
index 82f0432..b78c182 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/fannil.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/fannil.c
@@ -30,7 +30,7 @@ nouveau_fannil_get(struct nouveau_therm *therm)
 	return -ENODEV;
 }
 
-int
+static int
 nouveau_fannil_set(struct nouveau_therm *therm, int percent)
 {
 	return -ENODEV;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/fantog.c b/drivers/gpu/drm/nouveau/core/subdev/therm/fantog.c
index e0693e0..e601773 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/fantog.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/fantog.c
@@ -81,7 +81,7 @@ nouveau_fantog_get(struct nouveau_therm *therm)
 	return priv->percent;
 }
 
-int
+static int
 nouveau_fantog_set(struct nouveau_therm *therm, int percent)
 {
 	struct nouveau_therm_priv *tpriv = (void *)therm;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c
index 5c17577..accc628 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c
@@ -80,7 +80,7 @@ nv40_temp_get(struct nouveau_therm *therm)
 	return core_temp;
 }
 
-int
+static int
 nv40_fan_pwm_ctrl(struct nouveau_therm *therm, int line, bool enable)
 {
 	u32 mask = enable ? 0x80000000 : 0x0000000;
@@ -93,7 +93,7 @@ nv40_fan_pwm_ctrl(struct nouveau_therm *therm, int line, bool enable)
 	return 0;
 }
 
-int
+static int
 nv40_fan_pwm_get(struct nouveau_therm *therm, int line, u32 *divs, u32 *duty)
 {
 	if (line == 2) {
@@ -119,7 +119,7 @@ nv40_fan_pwm_get(struct nouveau_therm *therm, int line, u32 *divs, u32 *duty)
 	return -EINVAL;
 }
 
-int
+static int
 nv40_fan_pwm_set(struct nouveau_therm *therm, int line, u32 divs, u32 duty)
 {
 	if (line == 2) {
diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c
index 2cd66cd..ef1ad21 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_drm.c
@@ -398,7 +398,7 @@ nouveau_drm_remove(struct pci_dev *pdev)
 	nouveau_object_debug();
 }
 
-int
+static int
 nouveau_do_suspend(struct drm_device *dev)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
@@ -469,7 +469,7 @@ int nouveau_pmops_suspend(struct device *dev)
 	return 0;
 }
 
-int
+static int
 nouveau_do_resume(struct drm_device *dev)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
-- 
cgit v0.10.2


From fc3109a2cb294335bcf3c5db5e16ae5fe68849f2 Mon Sep 17 00:00:00 2001
From: Marcin Slusarz <marcin.slusarz@gmail.com>
Date: Sun, 27 Jan 2013 15:51:09 +0100
Subject: drm/nouveau/fan: fix selection of fan speed when fan->get returns an
 error

fan->get returns int, but we write it to unsigned variable, and then check
whether it's >= 0 (it always is)

Found by smatch:
drivers/gpu/drm/nouveau/core/subdev/therm/fan.c:61 nouveau_fan_update() warn: always true condition '(duty >= 0) => (0-u32max >= 0)'

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
index b179655..c728380 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
@@ -39,7 +39,7 @@ nouveau_fan_update(struct nouveau_fan *fan, bool immediate, int target)
 	struct nouveau_timer *ptimer = nouveau_timer(priv);
 	unsigned long flags;
 	int ret = 0;
-	u32 duty;
+	int duty;
 
 	/* update target fan speed, restricting to allowed range */
 	spin_lock_irqsave(&fan->lock, flags);
@@ -64,9 +64,9 @@ nouveau_fan_update(struct nouveau_fan *fan, bool immediate, int target)
 		 * it is meant to bump the fan speed more incrementally
 		 */
 		if (duty < target)
-			duty = min(duty + 3, (u32) target);
+			duty = min(duty + 3, target);
 		else if (duty > target)
-			duty = max(duty - 3, (u32) target);
+			duty = max(duty - 3, target);
 	} else {
 		duty = target;
 	}
-- 
cgit v0.10.2


From fd69aee4271ead4bf5c759aeaf4f16ab3d63322f Mon Sep 17 00:00:00 2001
From: Marcin Slusarz <marcin.slusarz@gmail.com>
Date: Sun, 27 Jan 2013 16:11:18 +0100
Subject: drm/nvc0/graph: remove redundant null checks

It's safe to call kfree(NULL).

Found by smatch.

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c
index 15bfe55..0de0dd7 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c
@@ -613,10 +613,8 @@ nvc0_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 static void
 nvc0_graph_dtor_fw(struct nvc0_graph_fuc *fuc)
 {
-	if (fuc->data) {
-		kfree(fuc->data);
-		fuc->data = NULL;
-	}
+	kfree(fuc->data);
+	fuc->data = NULL;
 }
 
 void
@@ -624,8 +622,7 @@ nvc0_graph_dtor(struct nouveau_object *object)
 {
 	struct nvc0_graph_priv *priv = (void *)object;
 
-	if (priv->data)
-		kfree(priv->data);
+	kfree(priv->data);
 
 	nvc0_graph_dtor_fw(&priv->fuc409c);
 	nvc0_graph_dtor_fw(&priv->fuc409d);
-- 
cgit v0.10.2


From b5d8f05204da22ec896145729eea1f875085126b Mon Sep 17 00:00:00 2001
From: Cong Ding <dinggnu@gmail.com>
Date: Tue, 15 Jan 2013 18:19:49 +0100
Subject: drm/nouveau: remove unnecessary null pointer check from
 nouveau_fence_new

the variable chan is dereferenced in line 190, so it is no reason to check
null again in line 198.

Signed-off-by: Cong Ding <dinggnu@gmail.com>
Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.c b/drivers/gpu/drm/nouveau/nouveau_fence.c
index 1d049be..b6b8e49 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fence.c
+++ b/drivers/gpu/drm/nouveau/nouveau_fence.c
@@ -195,11 +195,9 @@ nouveau_fence_new(struct nouveau_channel *chan, struct nouveau_fence **pfence)
 		return -ENOMEM;
 	kref_init(&fence->kref);
 
-	if (chan) {
-		ret = nouveau_fence_emit(fence, chan);
-		if (ret)
-			nouveau_fence_unref(&fence);
-	}
+	ret = nouveau_fence_emit(fence, chan);
+	if (ret)
+		nouveau_fence_unref(&fence);
 
 	*pfence = fence;
 	return ret;
-- 
cgit v0.10.2


From 03e9a04050f5333c779ed37c027a6aae5644cf6c Mon Sep 17 00:00:00 2001
From: Marcin Slusarz <marcin.slusarz@gmail.com>
Date: Sun, 27 Jan 2013 16:43:00 +0100
Subject: drm/nouveau: use drm_property_create_range helper

Avoids potential null pointer dereference.

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nouveau_display.c b/drivers/gpu/drm/nouveau/nouveau_display.c
index 508b00a..e2fdd75 100644
--- a/drivers/gpu/drm/nouveau/nouveau_display.c
+++ b/drivers/gpu/drm/nouveau/nouveau_display.c
@@ -316,17 +316,13 @@ nouveau_display_create(struct drm_device *dev)
 		drm_property_create_range(dev, 0, "underscan vborder", 0, 128);
 
 	if (gen >= 1) {
+		/* -90..+90 */
 		disp->vibrant_hue_property =
-			drm_property_create(dev, DRM_MODE_PROP_RANGE,
-					    "vibrant hue", 2);
-		disp->vibrant_hue_property->values[0] = 0;
-		disp->vibrant_hue_property->values[1] = 180; /* -90..+90 */
+			drm_property_create_range(dev, 0, "vibrant hue", 0, 180);
 
+		/* -100..+100 */
 		disp->color_vibrance_property =
-			drm_property_create(dev, DRM_MODE_PROP_RANGE,
-					    "color vibrance", 2);
-		disp->color_vibrance_property->values[0] = 0;
-		disp->color_vibrance_property->values[1] = 200; /* -100..+100 */
+			drm_property_create_range(dev, 0, "color vibrance", 0, 200);
 	}
 
 	dev->mode_config.funcs = &nouveau_mode_config_funcs;
-- 
cgit v0.10.2


From a441dbb1d674b5696e3a27a95e72988f35236c9f Mon Sep 17 00:00:00 2001
From: Marcin Slusarz <marcin.slusarz@gmail.com>
Date: Sun, 27 Jan 2013 17:04:48 +0100
Subject: drm/nouveau: use kmemdup for edid allocation/copying

Avoids potential null pointer dereference.

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.c b/drivers/gpu/drm/nouveau/nouveau_connector.c
index e620ba8..2f3e40e 100644
--- a/drivers/gpu/drm/nouveau/nouveau_connector.c
+++ b/drivers/gpu/drm/nouveau/nouveau_connector.c
@@ -399,9 +399,10 @@ nouveau_connector_detect_lvds(struct drm_connector *connector, bool force)
 		struct edid *edid =
 			(struct edid *)nouveau_bios_embedded_edid(dev);
 		if (edid) {
-			nv_connector->edid = kmalloc(EDID_LENGTH, GFP_KERNEL);
-			*(nv_connector->edid) = *edid;
-			status = connector_status_connected;
+			nv_connector->edid =
+					kmemdup(edid, EDID_LENGTH, GFP_KERNEL);
+			if (nv_connector->edid)
+				status = connector_status_connected;
 		}
 	}
 
-- 
cgit v0.10.2


From b795016cd2c6aebfa40cbccb3bad4e0325815fb4 Mon Sep 17 00:00:00 2001
From: Marcin Slusarz <marcin.slusarz@gmail.com>
Date: Sun, 27 Jan 2013 17:13:52 +0100
Subject: drm/nouveau: handle backlight_device_register failure

Found by smatch.

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nouveau_backlight.c b/drivers/gpu/drm/nouveau/nouveau_backlight.c
index f65b20a..5d94030 100644
--- a/drivers/gpu/drm/nouveau/nouveau_backlight.c
+++ b/drivers/gpu/drm/nouveau/nouveau_backlight.c
@@ -84,6 +84,8 @@ nv40_backlight_init(struct drm_connector *connector)
 	props.max_brightness = 31;
 	bd = backlight_device_register("nv_backlight", &connector->kdev, drm,
 				       &nv40_bl_ops, &props);
+	if (IS_ERR(bd))
+		return PTR_ERR(bd);
 	drm->backlight = bd;
 	bd->props.brightness = nv40_get_intensity(bd);
 	backlight_update_status(bd);
-- 
cgit v0.10.2


From 3969f05bb8d0d72ba974ab5f74a7bafb61dd16d0 Mon Sep 17 00:00:00 2001
From: Marcin Slusarz <marcin.slusarz@gmail.com>
Date: Wed, 30 Jan 2013 22:21:31 +0100
Subject: drm/nouveau/therm: always initialize alarm_program_lock

Fixes "BUG: spinlock bad magic" on module load for nva3+ cards.

Introduced in commit "drm/nouveau/therm: implement support for temperature
alarms".

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/base.c b/drivers/gpu/drm/nouveau/core/subdev/therm/base.c
index 25b7f6a..f794dc8 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/base.c
@@ -299,6 +299,7 @@ nouveau_therm_create_(struct nouveau_object *parent,
 
 	nouveau_alarm_init(&priv->alarm, nouveau_therm_alarm);
 	spin_lock_init(&priv->lock);
+	spin_lock_init(&priv->sensor.alarm_program_lock);
 
 	priv->base.fan_get = nouveau_therm_fan_user_get;
 	priv->base.fan_set = nouveau_therm_fan_user_set;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nv50.c
index 4b7fe24..86632cb 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nv50.c
@@ -279,7 +279,6 @@ nv50_therm_ctor(struct nouveau_object *parent,
 	priv->base.base.pwm_clock = nv50_fan_pwm_clock;
 	priv->base.base.temp_get = nv50_temp_get;
 	priv->base.sensor.program_alarms = nv50_therm_program_alarms;
-	spin_lock_init(&priv->base.sensor.alarm_program_lock);
 	nv_subdev(priv)->intr = nv50_therm_intr;
 
 	/* init the thresholds */
-- 
cgit v0.10.2


From a84fa1a32e4c224edd16c825a571fe27ab366913 Mon Sep 17 00:00:00 2001
From: Marcin Slusarz <marcin.slusarz@gmail.com>
Date: Sun, 3 Feb 2013 22:02:47 +0100
Subject: drm/nouveau: report channel owner in ioctl error paths

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c b/drivers/gpu/drm/nouveau/nouveau_gem.c
index 8bf695c..7061138 100644
--- a/drivers/gpu/drm/nouveau/nouveau_gem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_gem.c
@@ -205,6 +205,7 @@ nouveau_gem_ioctl_new(struct drm_device *dev, void *data,
 		      struct drm_file *file_priv)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nouveau_cli *cli = nouveau_cli(file_priv);
 	struct nouveau_fb *pfb = nouveau_fb(drm->device);
 	struct drm_nouveau_gem_new *req = data;
 	struct nouveau_bo *nvbo = NULL;
@@ -213,7 +214,7 @@ nouveau_gem_ioctl_new(struct drm_device *dev, void *data,
 	drm->ttm.bdev.dev_mapping = drm->dev->dev_mapping;
 
 	if (!pfb->memtype_valid(pfb, req->info.tile_flags)) {
-		NV_ERROR(drm, "bad page flags: 0x%08x\n", req->info.tile_flags);
+		NV_ERROR(cli, "bad page flags: 0x%08x\n", req->info.tile_flags);
 		return -EINVAL;
 	}
 
@@ -315,6 +316,7 @@ validate_init(struct nouveau_channel *chan, struct drm_file *file_priv,
 	      struct drm_nouveau_gem_pushbuf_bo *pbbo,
 	      int nr_buffers, struct validate_op *op)
 {
+	struct nouveau_cli *cli = nouveau_cli(file_priv);
 	struct drm_device *dev = chan->drm->dev;
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	uint32_t sequence;
@@ -324,7 +326,7 @@ validate_init(struct nouveau_channel *chan, struct drm_file *file_priv,
 	sequence = atomic_add_return(1, &drm->ttm.validate_sequence);
 retry:
 	if (++trycnt > 100000) {
-		NV_ERROR(drm, "%s failed and gave up.\n", __func__);
+		NV_ERROR(cli, "%s failed and gave up.\n", __func__);
 		return -EINVAL;
 	}
 
@@ -335,14 +337,14 @@ retry:
 
 		gem = drm_gem_object_lookup(dev, file_priv, b->handle);
 		if (!gem) {
-			NV_ERROR(drm, "Unknown handle 0x%08x\n", b->handle);
+			NV_ERROR(cli, "Unknown handle 0x%08x\n", b->handle);
 			validate_fini(op, NULL);
 			return -ENOENT;
 		}
 		nvbo = gem->driver_private;
 
 		if (nvbo->reserved_by && nvbo->reserved_by == file_priv) {
-			NV_ERROR(drm, "multiple instances of buffer %d on "
+			NV_ERROR(cli, "multiple instances of buffer %d on "
 				      "validation list\n", b->handle);
 			drm_gem_object_unreference_unlocked(gem);
 			validate_fini(op, NULL);
@@ -357,7 +359,7 @@ retry:
 			drm_gem_object_unreference_unlocked(gem);
 			if (unlikely(ret)) {
 				if (ret != -ERESTARTSYS)
-					NV_ERROR(drm, "fail reserve\n");
+					NV_ERROR(cli, "fail reserve\n");
 				return ret;
 			}
 			goto retry;
@@ -376,7 +378,7 @@ retry:
 		if (b->valid_domains & NOUVEAU_GEM_DOMAIN_GART)
 			list_add_tail(&nvbo->entry, &op->gart_list);
 		else {
-			NV_ERROR(drm, "invalid valid domains: 0x%08x\n",
+			NV_ERROR(cli, "invalid valid domains: 0x%08x\n",
 				 b->valid_domains);
 			list_add_tail(&nvbo->entry, &op->both_list);
 			validate_fini(op, NULL);
@@ -407,8 +409,9 @@ validate_sync(struct nouveau_channel *chan, struct nouveau_bo *nvbo)
 }
 
 static int
-validate_list(struct nouveau_channel *chan, struct list_head *list,
-	      struct drm_nouveau_gem_pushbuf_bo *pbbo, uint64_t user_pbbo_ptr)
+validate_list(struct nouveau_channel *chan, struct nouveau_cli *cli,
+	      struct list_head *list, struct drm_nouveau_gem_pushbuf_bo *pbbo,
+	      uint64_t user_pbbo_ptr)
 {
 	struct nouveau_drm *drm = chan->drm;
 	struct drm_nouveau_gem_pushbuf_bo __user *upbbo =
@@ -421,7 +424,7 @@ validate_list(struct nouveau_channel *chan, struct list_head *list,
 
 		ret = validate_sync(chan, nvbo);
 		if (unlikely(ret)) {
-			NV_ERROR(drm, "fail pre-validate sync\n");
+			NV_ERROR(cli, "fail pre-validate sync\n");
 			return ret;
 		}
 
@@ -429,20 +432,20 @@ validate_list(struct nouveau_channel *chan, struct list_head *list,
 					     b->write_domains,
 					     b->valid_domains);
 		if (unlikely(ret)) {
-			NV_ERROR(drm, "fail set_domain\n");
+			NV_ERROR(cli, "fail set_domain\n");
 			return ret;
 		}
 
 		ret = nouveau_bo_validate(nvbo, true, false);
 		if (unlikely(ret)) {
 			if (ret != -ERESTARTSYS)
-				NV_ERROR(drm, "fail ttm_validate\n");
+				NV_ERROR(cli, "fail ttm_validate\n");
 			return ret;
 		}
 
 		ret = validate_sync(chan, nvbo);
 		if (unlikely(ret)) {
-			NV_ERROR(drm, "fail post-validate sync\n");
+			NV_ERROR(cli, "fail post-validate sync\n");
 			return ret;
 		}
 
@@ -478,7 +481,7 @@ nouveau_gem_pushbuf_validate(struct nouveau_channel *chan,
 			     uint64_t user_buffers, int nr_buffers,
 			     struct validate_op *op, int *apply_relocs)
 {
-	struct nouveau_drm *drm = chan->drm;
+	struct nouveau_cli *cli = nouveau_cli(file_priv);
 	int ret, relocs = 0;
 
 	INIT_LIST_HEAD(&op->vram_list);
@@ -491,32 +494,32 @@ nouveau_gem_pushbuf_validate(struct nouveau_channel *chan,
 	ret = validate_init(chan, file_priv, pbbo, nr_buffers, op);
 	if (unlikely(ret)) {
 		if (ret != -ERESTARTSYS)
-			NV_ERROR(drm, "validate_init\n");
+			NV_ERROR(cli, "validate_init\n");
 		return ret;
 	}
 
-	ret = validate_list(chan, &op->vram_list, pbbo, user_buffers);
+	ret = validate_list(chan, cli, &op->vram_list, pbbo, user_buffers);
 	if (unlikely(ret < 0)) {
 		if (ret != -ERESTARTSYS)
-			NV_ERROR(drm, "validate vram_list\n");
+			NV_ERROR(cli, "validate vram_list\n");
 		validate_fini(op, NULL);
 		return ret;
 	}
 	relocs += ret;
 
-	ret = validate_list(chan, &op->gart_list, pbbo, user_buffers);
+	ret = validate_list(chan, cli, &op->gart_list, pbbo, user_buffers);
 	if (unlikely(ret < 0)) {
 		if (ret != -ERESTARTSYS)
-			NV_ERROR(drm, "validate gart_list\n");
+			NV_ERROR(cli, "validate gart_list\n");
 		validate_fini(op, NULL);
 		return ret;
 	}
 	relocs += ret;
 
-	ret = validate_list(chan, &op->both_list, pbbo, user_buffers);
+	ret = validate_list(chan, cli, &op->both_list, pbbo, user_buffers);
 	if (unlikely(ret < 0)) {
 		if (ret != -ERESTARTSYS)
-			NV_ERROR(drm, "validate both_list\n");
+			NV_ERROR(cli, "validate both_list\n");
 		validate_fini(op, NULL);
 		return ret;
 	}
@@ -545,11 +548,10 @@ u_memcpya(uint64_t user, unsigned nmemb, unsigned size)
 }
 
 static int
-nouveau_gem_pushbuf_reloc_apply(struct drm_device *dev,
+nouveau_gem_pushbuf_reloc_apply(struct nouveau_cli *cli,
 				struct drm_nouveau_gem_pushbuf *req,
 				struct drm_nouveau_gem_pushbuf_bo *bo)
 {
-	struct nouveau_drm *drm = nouveau_drm(dev);
 	struct drm_nouveau_gem_pushbuf_reloc *reloc = NULL;
 	int ret = 0;
 	unsigned i;
@@ -565,7 +567,7 @@ nouveau_gem_pushbuf_reloc_apply(struct drm_device *dev,
 		uint32_t data;
 
 		if (unlikely(r->bo_index > req->nr_buffers)) {
-			NV_ERROR(drm, "reloc bo index invalid\n");
+			NV_ERROR(cli, "reloc bo index invalid\n");
 			ret = -EINVAL;
 			break;
 		}
@@ -575,7 +577,7 @@ nouveau_gem_pushbuf_reloc_apply(struct drm_device *dev,
 			continue;
 
 		if (unlikely(r->reloc_bo_index > req->nr_buffers)) {
-			NV_ERROR(drm, "reloc container bo index invalid\n");
+			NV_ERROR(cli, "reloc container bo index invalid\n");
 			ret = -EINVAL;
 			break;
 		}
@@ -583,7 +585,7 @@ nouveau_gem_pushbuf_reloc_apply(struct drm_device *dev,
 
 		if (unlikely(r->reloc_bo_offset + 4 >
 			     nvbo->bo.mem.num_pages << PAGE_SHIFT)) {
-			NV_ERROR(drm, "reloc outside of bo\n");
+			NV_ERROR(cli, "reloc outside of bo\n");
 			ret = -EINVAL;
 			break;
 		}
@@ -592,7 +594,7 @@ nouveau_gem_pushbuf_reloc_apply(struct drm_device *dev,
 			ret = ttm_bo_kmap(&nvbo->bo, 0, nvbo->bo.mem.num_pages,
 					  &nvbo->kmap);
 			if (ret) {
-				NV_ERROR(drm, "failed kmap for reloc\n");
+				NV_ERROR(cli, "failed kmap for reloc\n");
 				break;
 			}
 			nvbo->validate_mapped = true;
@@ -617,7 +619,7 @@ nouveau_gem_pushbuf_reloc_apply(struct drm_device *dev,
 		ret = ttm_bo_wait(&nvbo->bo, false, false, false);
 		spin_unlock(&nvbo->bo.bdev->fence_lock);
 		if (ret) {
-			NV_ERROR(drm, "reloc wait_idle failed: %d\n", ret);
+			NV_ERROR(cli, "reloc wait_idle failed: %d\n", ret);
 			break;
 		}
 
@@ -633,6 +635,7 @@ nouveau_gem_ioctl_pushbuf(struct drm_device *dev, void *data,
 			  struct drm_file *file_priv)
 {
 	struct nouveau_abi16 *abi16 = nouveau_abi16_get(file_priv, dev);
+	struct nouveau_cli *cli = nouveau_cli(file_priv);
 	struct nouveau_abi16_chan *temp;
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	struct drm_nouveau_gem_pushbuf *req = data;
@@ -662,19 +665,19 @@ nouveau_gem_ioctl_pushbuf(struct drm_device *dev, void *data,
 		goto out_next;
 
 	if (unlikely(req->nr_push > NOUVEAU_GEM_MAX_PUSH)) {
-		NV_ERROR(drm, "pushbuf push count exceeds limit: %d max %d\n",
+		NV_ERROR(cli, "pushbuf push count exceeds limit: %d max %d\n",
 			 req->nr_push, NOUVEAU_GEM_MAX_PUSH);
 		return nouveau_abi16_put(abi16, -EINVAL);
 	}
 
 	if (unlikely(req->nr_buffers > NOUVEAU_GEM_MAX_BUFFERS)) {
-		NV_ERROR(drm, "pushbuf bo count exceeds limit: %d max %d\n",
+		NV_ERROR(cli, "pushbuf bo count exceeds limit: %d max %d\n",
 			 req->nr_buffers, NOUVEAU_GEM_MAX_BUFFERS);
 		return nouveau_abi16_put(abi16, -EINVAL);
 	}
 
 	if (unlikely(req->nr_relocs > NOUVEAU_GEM_MAX_RELOCS)) {
-		NV_ERROR(drm, "pushbuf reloc count exceeds limit: %d max %d\n",
+		NV_ERROR(cli, "pushbuf reloc count exceeds limit: %d max %d\n",
 			 req->nr_relocs, NOUVEAU_GEM_MAX_RELOCS);
 		return nouveau_abi16_put(abi16, -EINVAL);
 	}
@@ -692,7 +695,7 @@ nouveau_gem_ioctl_pushbuf(struct drm_device *dev, void *data,
 	/* Ensure all push buffers are on validate list */
 	for (i = 0; i < req->nr_push; i++) {
 		if (push[i].bo_index >= req->nr_buffers) {
-			NV_ERROR(drm, "push %d buffer not in list\n", i);
+			NV_ERROR(cli, "push %d buffer not in list\n", i);
 			ret = -EINVAL;
 			goto out_prevalid;
 		}
@@ -703,15 +706,15 @@ nouveau_gem_ioctl_pushbuf(struct drm_device *dev, void *data,
 					   req->nr_buffers, &op, &do_reloc);
 	if (ret) {
 		if (ret != -ERESTARTSYS)
-			NV_ERROR(drm, "validate: %d\n", ret);
+			NV_ERROR(cli, "validate: %d\n", ret);
 		goto out_prevalid;
 	}
 
 	/* Apply any relocations that are required */
 	if (do_reloc) {
-		ret = nouveau_gem_pushbuf_reloc_apply(dev, req, bo);
+		ret = nouveau_gem_pushbuf_reloc_apply(cli, req, bo);
 		if (ret) {
-			NV_ERROR(drm, "reloc apply: %d\n", ret);
+			NV_ERROR(cli, "reloc apply: %d\n", ret);
 			goto out;
 		}
 	}
@@ -719,7 +722,7 @@ nouveau_gem_ioctl_pushbuf(struct drm_device *dev, void *data,
 	if (chan->dma.ib_max) {
 		ret = nouveau_dma_wait(chan, req->nr_push + 1, 16);
 		if (ret) {
-			NV_ERROR(drm, "nv50cal_space: %d\n", ret);
+			NV_ERROR(cli, "nv50cal_space: %d\n", ret);
 			goto out;
 		}
 
@@ -734,7 +737,7 @@ nouveau_gem_ioctl_pushbuf(struct drm_device *dev, void *data,
 	if (nv_device(drm->device)->chipset >= 0x25) {
 		ret = RING_SPACE(chan, req->nr_push * 2);
 		if (ret) {
-			NV_ERROR(drm, "cal_space: %d\n", ret);
+			NV_ERROR(cli, "cal_space: %d\n", ret);
 			goto out;
 		}
 
@@ -748,7 +751,7 @@ nouveau_gem_ioctl_pushbuf(struct drm_device *dev, void *data,
 	} else {
 		ret = RING_SPACE(chan, req->nr_push * (2 + NOUVEAU_DMA_SKIPS));
 		if (ret) {
-			NV_ERROR(drm, "jmp_space: %d\n", ret);
+			NV_ERROR(cli, "jmp_space: %d\n", ret);
 			goto out;
 		}
 
@@ -786,7 +789,7 @@ nouveau_gem_ioctl_pushbuf(struct drm_device *dev, void *data,
 
 	ret = nouveau_fence_new(chan, &fence);
 	if (ret) {
-		NV_ERROR(drm, "error fencing pushbuf: %d\n", ret);
+		NV_ERROR(cli, "error fencing pushbuf: %d\n", ret);
 		WIND_RING(chan);
 		goto out;
 	}
-- 
cgit v0.10.2


From 134fc0327552f3ebee2cdcb50b956a31029661c0 Mon Sep 17 00:00:00 2001
From: Marcin Slusarz <marcin.slusarz@gmail.com>
Date: Sun, 3 Feb 2013 18:17:41 +0100
Subject: drm/nouveau/therm: turn on fan only when threshold hit in positive
 direction

+ the same for shutdown threshold - seems impossible, but shutdown can fail.

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Acked-by: Martin Peres <martin.peres@labri.fr>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/temp.c b/drivers/gpu/drm/nouveau/core/subdev/therm/temp.c
index bf9b3ce..8f27b44 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/temp.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/temp.c
@@ -105,7 +105,7 @@ void nouveau_therm_sensor_event(struct nouveau_therm *therm,
 		return;
 
 	if (dir == NOUVEAU_THERM_THRS_FALLING)
-		nv_info(therm, "temperature (%u C) went bellow the '%s' threshold\n",
+		nv_info(therm, "temperature (%u C) went below the '%s' threshold\n",
 			temperature, thresolds[thrs]);
 	else
 		nv_info(therm, "temperature (%u C) hit the '%s' threshold\n",
@@ -114,8 +114,10 @@ void nouveau_therm_sensor_event(struct nouveau_therm *therm,
 	active = (dir == NOUVEAU_THERM_THRS_RISING);
 	switch (thrs) {
 	case NOUVEAU_THERM_THRS_FANBOOST:
-		nouveau_therm_fan_set(therm, true, 100);
-		nouveau_therm_mode(therm, NOUVEAU_THERM_CTRL_AUTO);
+		if (active) {
+			nouveau_therm_fan_set(therm, true, 100);
+			nouveau_therm_mode(therm, NOUVEAU_THERM_CTRL_AUTO);
+		}
 		break;
 	case NOUVEAU_THERM_THRS_DOWNCLOCK:
 		if (priv->emergency.downclock)
@@ -126,7 +128,8 @@ void nouveau_therm_sensor_event(struct nouveau_therm *therm,
 			priv->emergency.pause(therm, active);
 		break;
 	case NOUVEAU_THERM_THRS_SHUTDOWN:
-		orderly_poweroff(true);
+		if (active)
+			orderly_poweroff(true);
 		break;
 	case NOUVEAU_THERM_THRS_NR:
 		break;
-- 
cgit v0.10.2


From d0a5191dcf5f37cf0942c4b65fa4a2c975631c0b Mon Sep 17 00:00:00 2001
From: Marcin Slusarz <marcin.slusarz@gmail.com>
Date: Sun, 3 Feb 2013 19:12:49 +0100
Subject: drm/nv40/therm: reset temperature sensor on init
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Current uninitialized sensor detection does not work for me on nv4b and
sensor returns crazy values (>190°C). It stabilises later, but it's too
late - therm code shutdowns the machine...

Let's just reset it on init.

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Acked-by: Martin Peres <martin.peres@labri.fr>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c
index accc628..0f5363e 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c
@@ -173,13 +173,23 @@ nv40_therm_ctor(struct nouveau_object *parent,
 	return nouveau_therm_preinit(&priv->base.base);
 }
 
+static int
+nv40_therm_init(struct nouveau_object *object)
+{
+	struct nouveau_therm *therm = (void *)object;
+
+	nv40_sensor_setup(therm);
+
+	return _nouveau_therm_init(object);
+}
+
 struct nouveau_oclass
 nv40_therm_oclass = {
 	.handle = NV_SUBDEV(THERM, 0x40),
 	.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = nv40_therm_ctor,
 		.dtor = _nouveau_therm_dtor,
-		.init = _nouveau_therm_init,
+		.init = nv40_therm_init,
 		.fini = _nouveau_therm_fini,
 	},
 };
-- 
cgit v0.10.2


From 07df30435eecce98961dabbac16787969975a080 Mon Sep 17 00:00:00 2001
From: Marcin Slusarz <marcin.slusarz@gmail.com>
Date: Sun, 3 Feb 2013 19:28:14 +0100
Subject: drm/nouveau/therm: use workqueue to shutdown the machine

orderly_poweroff cannot be called from atomic context.

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Acked-by: Martin Peres <martin.peres@labri.fr>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/temp.c b/drivers/gpu/drm/nouveau/core/subdev/therm/temp.c
index 8f27b44..b37624a 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/temp.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/temp.c
@@ -90,6 +90,13 @@ nouveau_therm_sensor_get_threshold_state(struct nouveau_therm *therm,
 	return priv->sensor.alarm_state[thrs];
 }
 
+static void
+nv_poweroff_work(struct work_struct *work)
+{
+	orderly_poweroff(true);
+	kfree(work);
+}
+
 void nouveau_therm_sensor_event(struct nouveau_therm *therm,
 			        enum nouveau_therm_thrs thrs,
 			        enum nouveau_therm_thrs_direction dir)
@@ -128,8 +135,15 @@ void nouveau_therm_sensor_event(struct nouveau_therm *therm,
 			priv->emergency.pause(therm, active);
 		break;
 	case NOUVEAU_THERM_THRS_SHUTDOWN:
-		if (active)
-			orderly_poweroff(true);
+		if (active) {
+			struct work_struct *work;
+
+			work = kmalloc(sizeof(*work), GFP_ATOMIC);
+			if (work) {
+				INIT_WORK(work, nv_poweroff_work);
+				schedule_work(work);
+			}
+		}
 		break;
 	case NOUVEAU_THERM_THRS_NR:
 		break;
-- 
cgit v0.10.2


From ec49b5c2fb75338a733f4b1d435977a460a8cb7a Mon Sep 17 00:00:00 2001
From: Marcin Slusarz <marcin.slusarz@gmail.com>
Date: Tue, 5 Feb 2013 20:44:19 +0100
Subject: drm/nouveau/therm: reduce stack usage of nouveau_therm_ic_ctor

Before: 1496 bytes
After:   152 bytes

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/ic.c b/drivers/gpu/drm/nouveau/core/subdev/therm/ic.c
index e512ff0..95a63e1 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/ic.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/ic.c
@@ -53,6 +53,31 @@ probe_monitoring_device(struct nouveau_i2c_port *i2c,
 	return true;
 }
 
+static struct i2c_board_info
+nv_board_infos[] = {
+	{ I2C_BOARD_INFO("w83l785ts", 0x2d) },
+	{ I2C_BOARD_INFO("w83781d", 0x2d) },
+	{ I2C_BOARD_INFO("adt7473", 0x2e) },
+	{ I2C_BOARD_INFO("adt7473", 0x2d) },
+	{ I2C_BOARD_INFO("adt7473", 0x2c) },
+	{ I2C_BOARD_INFO("f75375", 0x2e) },
+	{ I2C_BOARD_INFO("lm99", 0x4c) },
+	{ I2C_BOARD_INFO("lm90", 0x4c) },
+	{ I2C_BOARD_INFO("lm90", 0x4d) },
+	{ I2C_BOARD_INFO("adm1021", 0x18) },
+	{ I2C_BOARD_INFO("adm1021", 0x19) },
+	{ I2C_BOARD_INFO("adm1021", 0x1a) },
+	{ I2C_BOARD_INFO("adm1021", 0x29) },
+	{ I2C_BOARD_INFO("adm1021", 0x2a) },
+	{ I2C_BOARD_INFO("adm1021", 0x2b) },
+	{ I2C_BOARD_INFO("adm1021", 0x4c) },
+	{ I2C_BOARD_INFO("adm1021", 0x4d) },
+	{ I2C_BOARD_INFO("adm1021", 0x4e) },
+	{ I2C_BOARD_INFO("lm63", 0x18) },
+	{ I2C_BOARD_INFO("lm63", 0x4e) },
+	{ }
+};
+
 void
 nouveau_therm_ic_ctor(struct nouveau_therm *therm)
 {
@@ -60,29 +85,6 @@ nouveau_therm_ic_ctor(struct nouveau_therm *therm)
 	struct nouveau_bios *bios = nouveau_bios(therm);
 	struct nouveau_i2c *i2c = nouveau_i2c(therm);
 	struct nvbios_extdev_func extdev_entry;
-	struct i2c_board_info info[] = {
-		{ I2C_BOARD_INFO("w83l785ts", 0x2d) },
-		{ I2C_BOARD_INFO("w83781d", 0x2d) },
-		{ I2C_BOARD_INFO("adt7473", 0x2e) },
-		{ I2C_BOARD_INFO("adt7473", 0x2d) },
-		{ I2C_BOARD_INFO("adt7473", 0x2c) },
-		{ I2C_BOARD_INFO("f75375", 0x2e) },
-		{ I2C_BOARD_INFO("lm99", 0x4c) },
-		{ I2C_BOARD_INFO("lm90", 0x4c) },
-		{ I2C_BOARD_INFO("lm90", 0x4d) },
-		{ I2C_BOARD_INFO("adm1021", 0x18) },
-		{ I2C_BOARD_INFO("adm1021", 0x19) },
-		{ I2C_BOARD_INFO("adm1021", 0x1a) },
-		{ I2C_BOARD_INFO("adm1021", 0x29) },
-		{ I2C_BOARD_INFO("adm1021", 0x2a) },
-		{ I2C_BOARD_INFO("adm1021", 0x2b) },
-		{ I2C_BOARD_INFO("adm1021", 0x4c) },
-		{ I2C_BOARD_INFO("adm1021", 0x4d) },
-		{ I2C_BOARD_INFO("adm1021", 0x4e) },
-		{ I2C_BOARD_INFO("lm63", 0x18) },
-		{ I2C_BOARD_INFO("lm63", 0x4e) },
-		{ }
-	};
 
 	if (!nvbios_extdev_find(bios, NVBIOS_EXTDEV_LM89, &extdev_entry)) {
 		struct i2c_board_info board[] = {
@@ -111,6 +113,6 @@ nouveau_therm_ic_ctor(struct nouveau_therm *therm)
 	/* The vbios doesn't provide the address of an exisiting monitoring
 	   device. Let's try our static list.
 	 */
-	i2c->identify(i2c, NV_I2C_DEFAULT(0), "monitoring device", info,
-		      probe_monitoring_device);
+	i2c->identify(i2c, NV_I2C_DEFAULT(0), "monitoring device",
+		      nv_board_infos, probe_monitoring_device);
 }
-- 
cgit v0.10.2


From e2de179458fe681a1a7c63b5911a8dc86b00ca44 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Mon, 11 Feb 2013 13:56:31 +1000
Subject: drm/nv50/disp: fix missing sor modectrl sync flags

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nv50_display.c b/drivers/gpu/drm/nouveau/nv50_display.c
index 3587408..c9da4f1 100644
--- a/drivers/gpu/drm/nouveau/nv50_display.c
+++ b/drivers/gpu/drm/nouveau/nv50_display.c
@@ -1825,8 +1825,13 @@ nv50_sor_mode_set(struct drm_encoder *encoder, struct drm_display_mode *umode,
 	push = evo_wait(nv50_mast(dev), 8);
 	if (push) {
 		if (nv50_vers(mast) < NVD0_DISP_CLASS) {
+			u32 ctrl = (depth << 16) | (proto << 8) | owner;
+			if (mode->flags & DRM_MODE_FLAG_NHSYNC)
+				ctrl |= 0x00001000;
+			if (mode->flags & DRM_MODE_FLAG_NVSYNC)
+				ctrl |= 0x00002000;
 			evo_mthd(push, 0x0600 + (nv_encoder->or * 0x040), 1);
-			evo_data(push, (depth << 16) | (proto << 8) | owner);
+			evo_data(push, ctrl);
 		} else {
 			u32 magic = 0x31ec6000 | (nv_crtc->index << 25);
 			u32 syncs = 0x00000001;
-- 
cgit v0.10.2


From 32256c87ead3edec86bed5023a0ff96a6d907931 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Thu, 31 Jan 2013 19:49:33 -0500
Subject: drm/nouveau/fifo/nvc0: improve interrupt handler somewhat

Logs extra info for interrupts that have a sub-status register, and
handles the "special" ack from INTR bit 31.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
index 1580aae..2a9919b 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
@@ -517,12 +517,34 @@ nvc0_fifo_intr(struct nouveau_subdev *subdev)
 	u32 mask = nv_rd32(priv, 0x002140);
 	u32 stat = nv_rd32(priv, 0x002100) & mask;
 
+	if (stat & 0x00000001) {
+		u32 intr = nv_rd32(priv, 0x00252c);
+		nv_warn(priv, "INTR 0x00000001: 0x%08x\n", intr);
+		nv_wr32(priv, 0x002100, 0x00000001);
+		stat &= ~0x00000001;
+	}
+
 	if (stat & 0x00000100) {
-		nv_warn(priv, "unknown status 0x00000100\n");
+		u32 intr = nv_rd32(priv, 0x00254c);
+		nv_warn(priv, "INTR 0x00000100: 0x%08x\n", intr);
 		nv_wr32(priv, 0x002100, 0x00000100);
 		stat &= ~0x00000100;
 	}
 
+	if (stat & 0x00010000) {
+		u32 intr = nv_rd32(priv, 0x00256c);
+		nv_warn(priv, "INTR 0x00010000: 0x%08x\n", intr);
+		nv_wr32(priv, 0x002100, 0x00010000);
+		stat &= ~0x00010000;
+	}
+
+	if (stat & 0x01000000) {
+		u32 intr = nv_rd32(priv, 0x00258c);
+		nv_warn(priv, "INTR 0x01000000: 0x%08x\n", intr);
+		nv_wr32(priv, 0x002100, 0x01000000);
+		stat &= ~0x01000000;
+	}
+
 	if (stat & 0x10000000) {
 		u32 units = nv_rd32(priv, 0x00259c);
 		u32 u = units;
@@ -552,11 +574,19 @@ nvc0_fifo_intr(struct nouveau_subdev *subdev)
 	}
 
 	if (stat & 0x40000000) {
-		nv_warn(priv, "unknown status 0x40000000\n");
-		nv_mask(priv, 0x002a00, 0x00000000, 0x00000000);
+		u32 intr0 = nv_rd32(priv, 0x0025a4);
+		u32 intr1 = nv_mask(priv, 0x002a00, 0x00000000, 0x00000);
+		nv_debug(priv, "INTR 0x40000000: 0x%08x 0x%08x\n",
+			       intr0, intr1);
 		stat &= ~0x40000000;
 	}
 
+	if (stat & 0x80000000) {
+		u32 intr = nv_mask(priv, 0x0025a8, 0x00000000, 0x00000000);
+		nv_warn(priv, "INTR 0x80000000: 0x%08x\n", intr);
+		stat &= ~0x80000000;
+	}
+
 	if (stat) {
 		nv_fatal(priv, "unhandled status 0x%08x\n", stat);
 		nv_wr32(priv, 0x002100, stat);
-- 
cgit v0.10.2


From 51fa0253fbc5cdf26b85f620bf1a1034e2eda868 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Thu, 31 Jan 2013 00:38:22 +1000
Subject: drm/nouveau/core: basic event interface between core and drm

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index 6985740..6d4cb8a 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -11,6 +11,7 @@ nouveau-y := core/core/client.o
 nouveau-y += core/core/engctx.o
 nouveau-y += core/core/engine.o
 nouveau-y += core/core/enum.o
+nouveau-y += core/core/event.o
 nouveau-y += core/core/falcon.o
 nouveau-y += core/core/gpuobj.o
 nouveau-y += core/core/handle.o
diff --git a/drivers/gpu/drm/nouveau/core/core/event.c b/drivers/gpu/drm/nouveau/core/core/event.c
new file mode 100644
index 0000000..6d01e0f
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/core/event.c
@@ -0,0 +1,106 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include <core/os.h>
+#include <core/event.h>
+
+static void
+nouveau_event_put_locked(struct nouveau_event *event, int index,
+			 struct nouveau_eventh *handler)
+{
+	if (!--event->index[index].refs)
+		event->disable(event, index);
+	list_del(&handler->head);
+}
+
+void
+nouveau_event_put(struct nouveau_event *event, int index,
+		  struct nouveau_eventh *handler)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&event->lock, flags);
+	if (index < event->index_nr)
+		nouveau_event_put_locked(event, index, handler);
+	spin_unlock_irqrestore(&event->lock, flags);
+}
+
+void
+nouveau_event_get(struct nouveau_event *event, int index,
+		  struct nouveau_eventh *handler)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&event->lock, flags);
+	if (index < event->index_nr) {
+		list_add(&handler->head, &event->index[index].list);
+		if (!event->index[index].refs++)
+			event->enable(event, index);
+	}
+	spin_unlock_irqrestore(&event->lock, flags);
+}
+
+void
+nouveau_event_trigger(struct nouveau_event *event, int index)
+{
+	struct nouveau_eventh *handler, *temp;
+	unsigned long flags;
+
+	if (index >= event->index_nr)
+		return;
+
+	spin_lock_irqsave(&event->lock, flags);
+	list_for_each_entry_safe(handler, temp, &event->index[index].list, head) {
+		if (handler->func(handler, index) == NVKM_EVENT_DROP) {
+			nouveau_event_put_locked(event, index, handler);
+		}
+	}
+	spin_unlock_irqrestore(&event->lock, flags);
+}
+
+void
+nouveau_event_destroy(struct nouveau_event **pevent)
+{
+	struct nouveau_event *event = *pevent;
+	if (event) {
+		kfree(event);
+		*pevent = NULL;
+	}
+}
+
+int
+nouveau_event_create(int index_nr, struct nouveau_event **pevent)
+{
+	struct nouveau_event *event;
+	int i;
+
+	event = *pevent = kzalloc(sizeof(*event) + index_nr *
+				  sizeof(event->index[0]), GFP_KERNEL);
+	if (!event)
+		return -ENOMEM;
+
+	spin_lock_init(&event->lock);
+	for (i = 0; i < index_nr; i++)
+		INIT_LIST_HEAD(&event->index[i].list);
+	event->index_nr = index_nr;
+	return 0;
+}
diff --git a/drivers/gpu/drm/nouveau/core/include/core/event.h b/drivers/gpu/drm/nouveau/core/include/core/event.h
new file mode 100644
index 0000000..9e09440
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/core/event.h
@@ -0,0 +1,36 @@
+#ifndef __NVKM_EVENT_H__
+#define __NVKM_EVENT_H__
+
+/* return codes from event handlers */
+#define NVKM_EVENT_DROP 0
+#define NVKM_EVENT_KEEP 1
+
+struct nouveau_eventh {
+	struct list_head head;
+	int (*func)(struct nouveau_eventh *, int index);
+};
+
+struct nouveau_event {
+	spinlock_t lock;
+
+	void *priv;
+	void (*enable)(struct nouveau_event *, int index);
+	void (*disable)(struct nouveau_event *, int index);
+
+	int index_nr;
+	struct {
+		struct list_head list;
+		int refs;
+	} index[];
+};
+
+int  nouveau_event_create(int index_nr, struct nouveau_event **);
+void nouveau_event_destroy(struct nouveau_event **);
+void nouveau_event_trigger(struct nouveau_event *, int index);
+
+void nouveau_event_get(struct nouveau_event *, int index,
+		       struct nouveau_eventh *);
+void nouveau_event_put(struct nouveau_event *, int index,
+		       struct nouveau_eventh *);
+
+#endif
-- 
cgit v0.10.2


From 21a5ace0bfb737d65e6d345ccf3d63fdee141f98 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Thu, 31 Jan 2013 09:04:48 +1000
Subject: drm/nouveau/disp/nv04: implement a base display object class

Will be used for upcoming vblank event interfaces.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv04.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv04.c
index 1c919f2..6eaf725 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv04.c
@@ -23,6 +23,7 @@
  */
 
 #include <engine/disp.h>
+#include <core/class.h>
 
 struct nv04_disp_priv {
 	struct nouveau_disp base;
@@ -30,6 +31,7 @@ struct nv04_disp_priv {
 
 static struct nouveau_oclass
 nv04_disp_sclass[] = {
+	{ NV04_DISP_CLASS, &nouveau_object_ofuncs },
 	{},
 };
 
diff --git a/drivers/gpu/drm/nouveau/core/include/core/class.h b/drivers/gpu/drm/nouveau/core/include/core/class.h
index 47c4b3a..994a280 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/class.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/class.h
@@ -154,6 +154,14 @@ struct nve0_channel_ind_class {
 	u32 engine;
 };
 
+/* 0046: NV04_DISP
+ */
+
+#define NV04_DISP_CLASS                                              0x00000046
+
+struct nv04_display_class {
+};
+
 /* 5070: NV50_DISP
  * 8270: NV84_DISP
  * 8370: NVA0_DISP
diff --git a/drivers/gpu/drm/nouveau/nv04_display.c b/drivers/gpu/drm/nouveau/nv04_display.c
index 2cd6fb8..edc3156 100644
--- a/drivers/gpu/drm/nouveau/nv04_display.c
+++ b/drivers/gpu/drm/nouveau/nv04_display.c
@@ -22,6 +22,9 @@
  * Author: Ben Skeggs
  */
 
+#include <core/object.h>
+#include <core/class.h>
+
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
 
@@ -71,6 +74,11 @@ nv04_display_create(struct drm_device *dev)
 
 	nouveau_hw_save_vga_fonts(dev, 1);
 
+	ret = nouveau_object_new(nv_object(drm), NVDRM_DEVICE, 0xd1500000,
+				 NV04_DISP_CLASS, NULL, 0, &disp->core);
+	if (ret)
+		return ret;
+
 	nv04_crtc_create(dev, 0);
 	if (nv_two_heads(dev))
 		nv04_crtc_create(dev, 1);
diff --git a/drivers/gpu/drm/nouveau/nv04_display.h b/drivers/gpu/drm/nouveau/nv04_display.h
index 4532280..a0a031d 100644
--- a/drivers/gpu/drm/nouveau/nv04_display.h
+++ b/drivers/gpu/drm/nouveau/nv04_display.h
@@ -80,6 +80,7 @@ struct nv04_display {
 	struct nv04_mode_state saved_reg;
 	uint32_t saved_vga_font[4][16384];
 	uint32_t dac_users[4];
+	struct nouveau_object *core;
 };
 
 static inline struct nv04_display *
-- 
cgit v0.10.2


From 1d7c71a3e2f77336df536855b0efd2dc5bdeb41b Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Thu, 31 Jan 2013 09:23:34 +1000
Subject: drm/nouveau/disp: port vblank handling to event interface

This removes the nastiness with the interactions between display and
software engines when handling vblank semaphore release interrupts.

Now, all the semantics are handled in one place (sw) \o/.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index 6d4cb8a..ebe37fe 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -143,6 +143,7 @@ nouveau-y += core/engine/copy/nvc0.o
 nouveau-y += core/engine/copy/nve0.o
 nouveau-y += core/engine/crypt/nv84.o
 nouveau-y += core/engine/crypt/nv98.o
+nouveau-y += core/engine/disp/base.o
 nouveau-y += core/engine/disp/nv04.o
 nouveau-y += core/engine/disp/nv50.o
 nouveau-y += core/engine/disp/nv84.o
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/base.c b/drivers/gpu/drm/nouveau/core/engine/disp/base.c
new file mode 100644
index 0000000..7a5cae4
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/base.c
@@ -0,0 +1,52 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <engine/disp.h>
+
+void
+_nouveau_disp_dtor(struct nouveau_object *object)
+{
+	struct nouveau_disp *disp = (void *)object;
+	nouveau_event_destroy(&disp->vblank);
+	nouveau_engine_destroy(&disp->base);
+}
+
+int
+nouveau_disp_create_(struct nouveau_object *parent,
+		     struct nouveau_object *engine,
+		     struct nouveau_oclass *oclass, int heads,
+		     const char *intname, const char *extname,
+		     int length, void **pobject)
+{
+	struct nouveau_disp *disp;
+	int ret;
+
+	ret = nouveau_engine_create_(parent, engine, oclass, true,
+				     intname, extname, length, pobject);
+	disp = *pobject;
+	if (ret)
+		return ret;
+
+	return nouveau_event_create(heads, &disp->vblank);
+}
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv04.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv04.c
index 6eaf725..05e903f 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv04.c
@@ -23,6 +23,8 @@
  */
 
 #include <engine/disp.h>
+
+#include <core/event.h>
 #include <core/class.h>
 
 struct nv04_disp_priv {
@@ -35,12 +37,20 @@ nv04_disp_sclass[] = {
 	{},
 };
 
+/*******************************************************************************
+ * Display engine implementation
+ ******************************************************************************/
+
+static void
+nv04_disp_vblank_enable(struct nouveau_event *event, int head)
+{
+	nv_wr32(event->priv, 0x600140 + (head * 0x2000) , 0x00000001);
+}
+
 static void
-nv04_disp_intr_vblank(struct nv04_disp_priv *priv, int crtc)
+nv04_disp_vblank_disable(struct nouveau_event *event, int head)
 {
-	struct nouveau_disp *disp = &priv->base;
-	if (disp->vblank.notify)
-		disp->vblank.notify(disp->vblank.data, crtc);
+	nv_wr32(event->priv, 0x600140 + (head * 0x2000) , 0x00000000);
 }
 
 static void
@@ -51,25 +61,25 @@ nv04_disp_intr(struct nouveau_subdev *subdev)
 	u32 crtc1 = nv_rd32(priv, 0x602100);
 
 	if (crtc0 & 0x00000001) {
-		nv04_disp_intr_vblank(priv, 0);
+		nouveau_event_trigger(priv->base.vblank, 0);
 		nv_wr32(priv, 0x600100, 0x00000001);
 	}
 
 	if (crtc1 & 0x00000001) {
-		nv04_disp_intr_vblank(priv, 1);
+		nouveau_event_trigger(priv->base.vblank, 1);
 		nv_wr32(priv, 0x602100, 0x00000001);
 	}
 }
 
 static int
 nv04_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
-		  struct nouveau_oclass *oclass, void *data, u32 size,
-		  struct nouveau_object **pobject)
+	       struct nouveau_oclass *oclass, void *data, u32 size,
+	       struct nouveau_object **pobject)
 {
 	struct nv04_disp_priv *priv;
 	int ret;
 
-	ret = nouveau_disp_create(parent, engine, oclass, "DISPLAY",
+	ret = nouveau_disp_create(parent, engine, oclass, 2, "DISPLAY",
 				  "display", &priv);
 	*pobject = nv_object(priv);
 	if (ret)
@@ -77,6 +87,9 @@ nv04_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 
 	nv_engine(priv)->sclass = nv04_disp_sclass;
 	nv_subdev(priv)->intr = nv04_disp_intr;
+	priv->base.vblank->priv = priv;
+	priv->base.vblank->enable = nv04_disp_vblank_enable;
+	priv->base.vblank->disable = nv04_disp_vblank_disable;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
index ca1a7d7..0d26f00 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
@@ -27,7 +27,6 @@
 #include <core/handle.h>
 #include <core/class.h>
 
-#include <engine/software.h>
 #include <engine/disp.h>
 
 #include <subdev/bios.h>
@@ -37,7 +36,6 @@
 #include <subdev/bios/pll.h>
 #include <subdev/timer.h>
 #include <subdev/fb.h>
-#include <subdev/bar.h>
 #include <subdev/clock.h>
 
 #include "nv50.h"
@@ -543,6 +541,18 @@ nv50_disp_curs_ofuncs = {
  * Base display object
  ******************************************************************************/
 
+static void
+nv50_disp_base_vblank_enable(struct nouveau_event *event, int head)
+{
+	nv_mask(event->priv, 0x61002c, (1 << head), (1 << head));
+}
+
+static void
+nv50_disp_base_vblank_disable(struct nouveau_event *event, int head)
+{
+	nv_mask(event->priv, 0x61002c, (1 << head), (0 << head));
+}
+
 static int
 nv50_disp_base_ctor(struct nouveau_object *parent,
 		    struct nouveau_object *engine,
@@ -559,6 +569,9 @@ nv50_disp_base_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
+	priv->base.vblank->priv = priv;
+	priv->base.vblank->enable = nv50_disp_base_vblank_enable;
+	priv->base.vblank->disable = nv50_disp_base_vblank_disable;
 	return nouveau_ramht_new(parent, parent, 0x1000, 0, &base->ramht);
 }
 
@@ -756,50 +769,6 @@ nv50_disp_intr_error(struct nv50_disp_priv *priv)
 	}
 }
 
-static void
-nv50_disp_intr_vblank(struct nv50_disp_priv *priv, int crtc)
-{
-	struct nouveau_bar *bar = nouveau_bar(priv);
-	struct nouveau_disp *disp = &priv->base;
-	struct nouveau_software_chan *chan, *temp;
-	unsigned long flags;
-
-	spin_lock_irqsave(&disp->vblank.lock, flags);
-	list_for_each_entry_safe(chan, temp, &disp->vblank.list, vblank.head) {
-		if (chan->vblank.crtc != crtc)
-			continue;
-
-		if (nv_device(priv)->chipset >= 0xc0) {
-			nv_wr32(priv, 0x001718, 0x80000000 | chan->vblank.channel);
-			bar->flush(bar);
-			nv_wr32(priv, 0x06000c,
-				upper_32_bits(chan->vblank.offset));
-			nv_wr32(priv, 0x060010,
-				lower_32_bits(chan->vblank.offset));
-			nv_wr32(priv, 0x060014, chan->vblank.value);
-		} else {
-			nv_wr32(priv, 0x001704, chan->vblank.channel);
-			nv_wr32(priv, 0x001710, 0x80000000 | chan->vblank.ctxdma);
-			bar->flush(bar);
-			if (nv_device(priv)->chipset == 0x50) {
-				nv_wr32(priv, 0x001570, chan->vblank.offset);
-				nv_wr32(priv, 0x001574, chan->vblank.value);
-			} else {
-				nv_wr32(priv, 0x060010, chan->vblank.offset);
-				nv_wr32(priv, 0x060014, chan->vblank.value);
-			}
-		}
-
-		list_del(&chan->vblank.head);
-		if (disp->vblank.put)
-			disp->vblank.put(disp->vblank.data, crtc);
-	}
-	spin_unlock_irqrestore(&disp->vblank.lock, flags);
-
-	if (disp->vblank.notify)
-		disp->vblank.notify(disp->vblank.data, crtc);
-}
-
 static u16
 exec_lookup(struct nv50_disp_priv *priv, int head, int outp, u32 ctrl,
 	    struct dcb_output *dcb, u8 *ver, u8 *hdr, u8 *cnt, u8 *len,
@@ -1201,13 +1170,13 @@ nv50_disp_intr(struct nouveau_subdev *subdev)
 	}
 
 	if (intr1 & 0x00000004) {
-		nv50_disp_intr_vblank(priv, 0);
+		nouveau_event_trigger(priv->base.vblank, 0);
 		nv_wr32(priv, 0x610024, 0x00000004);
 		intr1 &= ~0x00000004;
 	}
 
 	if (intr1 & 0x00000008) {
-		nv50_disp_intr_vblank(priv, 1);
+		nouveau_event_trigger(priv->base.vblank, 1);
 		nv_wr32(priv, 0x610024, 0x00000008);
 		intr1 &= ~0x00000008;
 	}
@@ -1226,7 +1195,7 @@ nv50_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv50_disp_priv *priv;
 	int ret;
 
-	ret = nouveau_disp_create(parent, engine, oclass, "PDISP",
+	ret = nouveau_disp_create(parent, engine, oclass, 2, "PDISP",
 				  "display", &priv);
 	*pobject = nv_object(priv);
 	if (ret)
@@ -1242,9 +1211,6 @@ nv50_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->dac.power = nv50_dac_power;
 	priv->dac.sense = nv50_dac_sense;
 	priv->sor.power = nv50_sor_power;
-
-	INIT_LIST_HEAD(&priv->base.vblank.list);
-	spin_lock_init(&priv->base.vblank.lock);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.h b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.h
index a6bb9314..fc89718 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.h
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.h
@@ -3,7 +3,9 @@
 
 #include <core/parent.h>
 #include <core/namedb.h>
+#include <core/engctx.h>
 #include <core/ramht.h>
+#include <core/event.h>
 
 #include <engine/dmaobj.h>
 #include <engine/disp.h>
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv84.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv84.c
index fc84eac..a215342 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv84.c
@@ -63,7 +63,7 @@ nv84_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv50_disp_priv *priv;
 	int ret;
 
-	ret = nouveau_disp_create(parent, engine, oclass, "PDISP",
+	ret = nouveau_disp_create(parent, engine, oclass, 2, "PDISP",
 				  "display", &priv);
 	*pobject = nv_object(priv);
 	if (ret)
@@ -80,9 +80,6 @@ nv84_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->dac.sense = nv50_dac_sense;
 	priv->sor.power = nv50_sor_power;
 	priv->sor.hdmi = nv84_hdmi_ctrl;
-
-	INIT_LIST_HEAD(&priv->base.vblank.list);
-	spin_lock_init(&priv->base.vblank.lock);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv94.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv94.c
index ba9dfd4..a315e28 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv94.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv94.c
@@ -69,7 +69,7 @@ nv94_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv50_disp_priv *priv;
 	int ret;
 
-	ret = nouveau_disp_create(parent, engine, oclass, "PDISP",
+	ret = nouveau_disp_create(parent, engine, oclass, 2, "PDISP",
 				  "display", &priv);
 	*pobject = nv_object(priv);
 	if (ret)
@@ -91,9 +91,6 @@ nv94_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->sor.dp_train_fini = nv94_sor_dp_train_fini;
 	priv->sor.dp_lnkctl = nv94_sor_dp_lnkctl;
 	priv->sor.dp_drvctl = nv94_sor_dp_drvctl;
-
-	INIT_LIST_HEAD(&priv->base.vblank.list);
-	spin_lock_init(&priv->base.vblank.lock);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nva0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nva0.c
index 5d63902..480e2de 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nva0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nva0.c
@@ -53,7 +53,7 @@ nva0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv50_disp_priv *priv;
 	int ret;
 
-	ret = nouveau_disp_create(parent, engine, oclass, "PDISP",
+	ret = nouveau_disp_create(parent, engine, oclass, 2, "PDISP",
 				  "display", &priv);
 	*pobject = nv_object(priv);
 	if (ret)
@@ -70,9 +70,6 @@ nva0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->dac.sense = nv50_dac_sense;
 	priv->sor.power = nv50_sor_power;
 	priv->sor.hdmi = nv84_hdmi_ctrl;
-
-	INIT_LIST_HEAD(&priv->base.vblank.list);
-	spin_lock_init(&priv->base.vblank.lock);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nva3.c b/drivers/gpu/drm/nouveau/core/engine/disp/nva3.c
index e9192ca..718b4f6 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nva3.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nva3.c
@@ -70,7 +70,7 @@ nva3_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv50_disp_priv *priv;
 	int ret;
 
-	ret = nouveau_disp_create(parent, engine, oclass, "PDISP",
+	ret = nouveau_disp_create(parent, engine, oclass, 2, "PDISP",
 				  "display", &priv);
 	*pobject = nv_object(priv);
 	if (ret)
@@ -93,9 +93,6 @@ nva3_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->sor.dp_train_fini = nv94_sor_dp_train_fini;
 	priv->sor.dp_lnkctl = nv94_sor_dp_lnkctl;
 	priv->sor.dp_drvctl = nv94_sor_dp_drvctl;
-
-	INIT_LIST_HEAD(&priv->base.vblank.list);
-	spin_lock_init(&priv->base.vblank.lock);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
index 9e38ebf..74626e8 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
@@ -27,12 +27,10 @@
 #include <core/handle.h>
 #include <core/class.h>
 
-#include <engine/software.h>
 #include <engine/disp.h>
 
 #include <subdev/timer.h>
 #include <subdev/fb.h>
-#include <subdev/bar.h>
 #include <subdev/clock.h>
 
 #include <subdev/bios.h>
@@ -443,6 +441,18 @@ nvd0_disp_curs_ofuncs = {
  * Base display object
  ******************************************************************************/
 
+static void
+nvd0_disp_base_vblank_enable(struct nouveau_event *event, int head)
+{
+	nv_mask(event->priv, 0x6100c0 + (head * 0x800), 0x00000001, 0x00000001);
+}
+
+static void
+nvd0_disp_base_vblank_disable(struct nouveau_event *event, int head)
+{
+	nv_mask(event->priv, 0x6100c0 + (head * 0x800), 0x00000001, 0x00000000);
+}
+
 static int
 nvd0_disp_base_ctor(struct nouveau_object *parent,
 		    struct nouveau_object *engine,
@@ -459,6 +469,10 @@ nvd0_disp_base_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
+	priv->base.vblank->priv = priv;
+	priv->base.vblank->enable = nvd0_disp_base_vblank_enable;
+	priv->base.vblank->disable = nvd0_disp_base_vblank_disable;
+
 	return nouveau_ramht_new(parent, parent, 0x1000, 0, &base->ramht);
 }
 
@@ -822,35 +836,6 @@ nvd0_display_unk4_handler(struct nv50_disp_priv *priv, u32 head, u32 mask)
 	nv_wr32(priv, 0x6101d0, 0x80000000);
 }
 
-static void
-nvd0_disp_intr_vblank(struct nv50_disp_priv *priv, int crtc)
-{
-	struct nouveau_bar *bar = nouveau_bar(priv);
-	struct nouveau_disp *disp = &priv->base;
-	struct nouveau_software_chan *chan, *temp;
-	unsigned long flags;
-
-	spin_lock_irqsave(&disp->vblank.lock, flags);
-	list_for_each_entry_safe(chan, temp, &disp->vblank.list, vblank.head) {
-		if (chan->vblank.crtc != crtc)
-			continue;
-
-		nv_wr32(priv, 0x001718, 0x80000000 | chan->vblank.channel);
-		bar->flush(bar);
-		nv_wr32(priv, 0x06000c, upper_32_bits(chan->vblank.offset));
-		nv_wr32(priv, 0x060010, lower_32_bits(chan->vblank.offset));
-		nv_wr32(priv, 0x060014, chan->vblank.value);
-
-		list_del(&chan->vblank.head);
-		if (disp->vblank.put)
-			disp->vblank.put(disp->vblank.data, crtc);
-	}
-	spin_unlock_irqrestore(&disp->vblank.lock, flags);
-
-	if (disp->vblank.notify)
-		disp->vblank.notify(disp->vblank.data, crtc);
-}
-
 void
 nvd0_disp_intr(struct nouveau_subdev *subdev)
 {
@@ -920,7 +905,7 @@ nvd0_disp_intr(struct nouveau_subdev *subdev)
 		if (mask & intr) {
 			u32 stat = nv_rd32(priv, 0x6100bc + (i * 0x800));
 			if (stat & 0x00000001)
-				nvd0_disp_intr_vblank(priv, i);
+				nouveau_event_trigger(priv->base.vblank, i);
 			nv_mask(priv, 0x6100bc + (i * 0x800), 0, 0);
 			nv_rd32(priv, 0x6100c0 + (i * 0x800));
 		}
@@ -933,10 +918,11 @@ nvd0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	       struct nouveau_object **pobject)
 {
 	struct nv50_disp_priv *priv;
+	int heads = nv_rd32(parent, 0x022448);
 	int ret;
 
-	ret = nouveau_disp_create(parent, engine, oclass, "PDISP",
-				  "display", &priv);
+	ret = nouveau_disp_create(parent, engine, oclass, heads,
+				  "PDISP", "display", &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
@@ -945,7 +931,7 @@ nvd0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nvd0_disp_intr;
 	priv->sclass = nvd0_disp_sclass;
-	priv->head.nr = nv_rd32(priv, 0x022448);
+	priv->head.nr = heads;
 	priv->dac.nr = 3;
 	priv->sor.nr = 4;
 	priv->dac.power = nv50_dac_power;
@@ -958,9 +944,6 @@ nvd0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->sor.dp_train_fini = nv94_sor_dp_train_fini;
 	priv->sor.dp_lnkctl = nvd0_sor_dp_lnkctl;
 	priv->sor.dp_drvctl = nvd0_sor_dp_drvctl;
-
-	INIT_LIST_HEAD(&priv->base.vblank.list);
-	spin_lock_init(&priv->base.vblank.lock);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c
index 259537c..5512296 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c
@@ -51,10 +51,11 @@ nve0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	       struct nouveau_object **pobject)
 {
 	struct nv50_disp_priv *priv;
+	int heads = nv_rd32(parent, 0x022448);
 	int ret;
 
-	ret = nouveau_disp_create(parent, engine, oclass, "PDISP",
-				  "display", &priv);
+	ret = nouveau_disp_create(parent, engine, oclass, heads,
+				  "PDISP", "display", &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
@@ -63,7 +64,7 @@ nve0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nvd0_disp_intr;
 	priv->sclass = nve0_disp_sclass;
-	priv->head.nr = nv_rd32(priv, 0x022448);
+	priv->head.nr = heads;
 	priv->dac.nr = 3;
 	priv->sor.nr = 4;
 	priv->dac.power = nv50_dac_power;
@@ -76,9 +77,6 @@ nve0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->sor.dp_train_fini = nv94_sor_dp_train_fini;
 	priv->sor.dp_lnkctl = nvd0_sor_dp_lnkctl;
 	priv->sor.dp_drvctl = nvd0_sor_dp_drvctl;
-
-	INIT_LIST_HEAD(&priv->base.vblank.list);
-	spin_lock_init(&priv->base.vblank.lock);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/software/nv50.c b/drivers/gpu/drm/nouveau/core/engine/software/nv50.c
index b0e7e1c..c48e749 100644
--- a/drivers/gpu/drm/nouveau/core/engine/software/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/software/nv50.c
@@ -28,6 +28,9 @@
 #include <core/namedb.h>
 #include <core/handle.h>
 #include <core/gpuobj.h>
+#include <core/event.h>
+
+#include <subdev/bar.h>
 
 #include <engine/software.h>
 #include <engine/disp.h>
@@ -90,18 +93,11 @@ nv50_software_mthd_vblsem_release(struct nouveau_object *object, u32 mthd,
 {
 	struct nv50_software_chan *chan = (void *)nv_engctx(object->parent);
 	struct nouveau_disp *disp = nouveau_disp(object);
-	unsigned long flags;
 	u32 crtc = *(u32 *)args;
-
 	if (crtc > 1)
 		return -EINVAL;
 
-	disp->vblank.get(disp->vblank.data, crtc);
-
-	spin_lock_irqsave(&disp->vblank.lock, flags);
-	list_add(&chan->base.vblank.head, &disp->vblank.list);
-	chan->base.vblank.crtc = crtc;
-	spin_unlock_irqrestore(&disp->vblank.lock, flags);
+	nouveau_event_get(disp->vblank, crtc, &chan->base.vblank.event);
 	return 0;
 }
 
@@ -136,6 +132,29 @@ nv50_software_sclass[] = {
  ******************************************************************************/
 
 static int
+nv50_software_vblsem_release(struct nouveau_eventh *event, int head)
+{
+	struct nouveau_software_chan *chan =
+		container_of(event, struct nouveau_software_chan, vblank.event);
+	struct nv50_software_priv *priv = (void *)nv_object(chan)->engine;
+	struct nouveau_bar *bar = nouveau_bar(priv);
+
+	nv_wr32(priv, 0x001704, chan->vblank.channel);
+	nv_wr32(priv, 0x001710, 0x80000000 | chan->vblank.ctxdma);
+	bar->flush(bar);
+
+	if (nv_device(priv)->chipset == 0x50) {
+		nv_wr32(priv, 0x001570, chan->vblank.offset);
+		nv_wr32(priv, 0x001574, chan->vblank.value);
+	} else {
+		nv_wr32(priv, 0x060010, chan->vblank.offset);
+		nv_wr32(priv, 0x060014, chan->vblank.value);
+	}
+
+	return NVKM_EVENT_DROP;
+}
+
+static int
 nv50_software_context_ctor(struct nouveau_object *parent,
 			   struct nouveau_object *engine,
 			   struct nouveau_oclass *oclass, void *data, u32 size,
@@ -150,6 +169,7 @@ nv50_software_context_ctor(struct nouveau_object *parent,
 		return ret;
 
 	chan->base.vblank.channel = nv_gpuobj(parent->parent)->addr >> 12;
+	chan->base.vblank.event.func = nv50_software_vblsem_release;
 	return 0;
 }
 
@@ -170,8 +190,8 @@ nv50_software_cclass = {
 
 static int
 nv50_software_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
-	      struct nouveau_oclass *oclass, void *data, u32 size,
-	      struct nouveau_object **pobject)
+		   struct nouveau_oclass *oclass, void *data, u32 size,
+		   struct nouveau_object **pobject)
 {
 	struct nv50_software_priv *priv;
 	int ret;
diff --git a/drivers/gpu/drm/nouveau/core/engine/software/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/software/nvc0.c
index 282a1cd..a523eaa 100644
--- a/drivers/gpu/drm/nouveau/core/engine/software/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/software/nvc0.c
@@ -25,6 +25,9 @@
 #include <core/os.h>
 #include <core/class.h>
 #include <core/engctx.h>
+#include <core/event.h>
+
+#include <subdev/bar.h>
 
 #include <engine/software.h>
 #include <engine/disp.h>
@@ -72,18 +75,12 @@ nvc0_software_mthd_vblsem_release(struct nouveau_object *object, u32 mthd,
 {
 	struct nvc0_software_chan *chan = (void *)nv_engctx(object->parent);
 	struct nouveau_disp *disp = nouveau_disp(object);
-	unsigned long flags;
 	u32 crtc = *(u32 *)args;
 
 	if ((nv_device(object)->card_type < NV_E0 && crtc > 1) || crtc > 3)
 		return -EINVAL;
 
-	disp->vblank.get(disp->vblank.data, crtc);
-
-	spin_lock_irqsave(&disp->vblank.lock, flags);
-	list_add(&chan->base.vblank.head, &disp->vblank.list);
-	chan->base.vblank.crtc = crtc;
-	spin_unlock_irqrestore(&disp->vblank.lock, flags);
+	nouveau_event_get(disp->vblank, crtc, &chan->base.vblank.event);
 	return 0;
 }
 
@@ -118,6 +115,23 @@ nvc0_software_sclass[] = {
  ******************************************************************************/
 
 static int
+nvc0_software_vblsem_release(struct nouveau_eventh *event, int head)
+{
+	struct nouveau_software_chan *chan =
+		container_of(event, struct nouveau_software_chan, vblank.event);
+	struct nvc0_software_priv *priv = (void *)nv_object(chan)->engine;
+	struct nouveau_bar *bar = nouveau_bar(priv);
+
+	nv_wr32(priv, 0x001718, 0x80000000 | chan->vblank.channel);
+	bar->flush(bar);
+	nv_wr32(priv, 0x06000c, upper_32_bits(chan->vblank.offset));
+	nv_wr32(priv, 0x060010, lower_32_bits(chan->vblank.offset));
+	nv_wr32(priv, 0x060014, chan->vblank.value);
+
+	return NVKM_EVENT_DROP;
+}
+
+static int
 nvc0_software_context_ctor(struct nouveau_object *parent,
 			   struct nouveau_object *engine,
 			   struct nouveau_oclass *oclass, void *data, u32 size,
@@ -132,6 +146,7 @@ nvc0_software_context_ctor(struct nouveau_object *parent,
 		return ret;
 
 	chan->base.vblank.channel = nv_gpuobj(parent->parent)->addr >> 12;
+	chan->base.vblank.event.func = nvc0_software_vblsem_release;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/disp.h b/drivers/gpu/drm/nouveau/core/include/engine/disp.h
index 4694828..28da677 100644
--- a/drivers/gpu/drm/nouveau/core/include/engine/disp.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/disp.h
@@ -4,18 +4,11 @@
 #include <core/object.h>
 #include <core/engine.h>
 #include <core/device.h>
+#include <core/event.h>
 
 struct nouveau_disp {
 	struct nouveau_engine base;
-
-	struct {
-		struct list_head list;
-		spinlock_t lock;
-		void (*notify)(void *, int);
-		void (*get)(void *, int);
-		void (*put)(void *, int);
-		void *data;
-	} vblank;
+	struct nouveau_event *vblank;
 };
 
 static inline struct nouveau_disp *
@@ -24,16 +17,22 @@ nouveau_disp(void *obj)
 	return (void *)nv_device(obj)->subdev[NVDEV_ENGINE_DISP];
 }
 
-#define nouveau_disp_create(p,e,c,i,x,d)                                       \
-	nouveau_engine_create((p), (e), (c), true, (i), (x), (d))
-#define nouveau_disp_destroy(d)                                                \
-	nouveau_engine_destroy(&(d)->base)
+#define nouveau_disp_create(p,e,c,h,i,x,d)                                     \
+	nouveau_disp_create_((p), (e), (c), (h), (i), (x),                     \
+			     sizeof(**d), (void **)d)
+#define nouveau_disp_destroy(d) ({                                             \
+	struct nouveau_disp *disp = (d);                                       \
+	_nouveau_disp_dtor(nv_object(disp));                                   \
+})
 #define nouveau_disp_init(d)                                                   \
 	nouveau_engine_init(&(d)->base)
 #define nouveau_disp_fini(d,s)                                                 \
 	nouveau_engine_fini(&(d)->base, (s))
 
-#define _nouveau_disp_dtor _nouveau_engine_dtor
+int  nouveau_disp_create_(struct nouveau_object *, struct nouveau_object *,
+			  struct nouveau_oclass *, int heads,
+			  const char *, const char *, int, void **);
+void _nouveau_disp_dtor(struct nouveau_object *);
 #define _nouveau_disp_init _nouveau_engine_init
 #define _nouveau_disp_fini _nouveau_engine_fini
 
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/software.h b/drivers/gpu/drm/nouveau/core/include/engine/software.h
index c945691..4579948 100644
--- a/drivers/gpu/drm/nouveau/core/include/engine/software.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/software.h
@@ -3,17 +3,17 @@
 
 #include <core/engine.h>
 #include <core/engctx.h>
+#include <core/event.h>
 
 struct nouveau_software_chan {
 	struct nouveau_engctx base;
 
 	struct {
-		struct list_head head;
+		struct nouveau_eventh event;
 		u32 channel;
 		u32 ctxdma;
 		u64 offset;
 		u32 value;
-		u32 crtc;
 	} vblank;
 
 	int (*flip)(void *);
diff --git a/drivers/gpu/drm/nouveau/nouveau_display.c b/drivers/gpu/drm/nouveau/nouveau_display.c
index e2fdd75..9f84803 100644
--- a/drivers/gpu/drm/nouveau/nouveau_display.c
+++ b/drivers/gpu/drm/nouveau/nouveau_display.c
@@ -41,6 +41,8 @@
 #include <subdev/gpio.h>
 #include <engine/disp.h>
 
+#include <core/class.h>
+
 static void
 nouveau_user_framebuffer_destroy(struct drm_framebuffer *drm_fb)
 {
@@ -257,29 +259,10 @@ nouveau_display_fini(struct drm_device *dev)
 	disp->fini(dev);
 }
 
-static void
-nouveau_display_vblank_notify(void *data, int crtc)
-{
-	drm_handle_vblank(data, crtc);
-}
-
-static void
-nouveau_display_vblank_get(void *data, int crtc)
-{
-	drm_vblank_get(data, crtc);
-}
-
-static void
-nouveau_display_vblank_put(void *data, int crtc)
-{
-	drm_vblank_put(data, crtc);
-}
-
 int
 nouveau_display_create(struct drm_device *dev)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_disp *pdisp = nouveau_disp(drm->device);
 	struct nouveau_display *disp;
 	u32 pclass = dev->pdev->class >> 8;
 	int ret, gen;
@@ -288,11 +271,6 @@ nouveau_display_create(struct drm_device *dev)
 	if (!disp)
 		return -ENOMEM;
 
-	pdisp->vblank.data = dev;
-	pdisp->vblank.notify = nouveau_display_vblank_notify;
-	pdisp->vblank.get = nouveau_display_vblank_get;
-	pdisp->vblank.put = nouveau_display_vblank_put;
-
 	drm_mode_config_init(dev);
 	drm_mode_create_scaling_mode_property(dev);
 	drm_mode_create_dvi_i_properties(dev);
@@ -474,39 +452,6 @@ nouveau_display_resume(struct drm_device *dev)
 	}
 }
 
-int
-nouveau_vblank_enable(struct drm_device *dev, int crtc)
-{
-	struct nouveau_device *device = nouveau_dev(dev);
-
-	if (device->card_type >= NV_D0)
-		nv_mask(device, 0x6100c0 + (crtc * 0x800), 1, 1);
-	else
-	if (device->card_type >= NV_50)
-		nv_mask(device, NV50_PDISPLAY_INTR_EN_1, 0,
-			NV50_PDISPLAY_INTR_EN_1_VBLANK_CRTC_(crtc));
-	else
-		NVWriteCRTC(dev, crtc, NV_PCRTC_INTR_EN_0,
-			    NV_PCRTC_INTR_0_VBLANK);
-
-	return 0;
-}
-
-void
-nouveau_vblank_disable(struct drm_device *dev, int crtc)
-{
-	struct nouveau_device *device = nouveau_dev(dev);
-
-	if (device->card_type >= NV_D0)
-		nv_mask(device, 0x6100c0 + (crtc * 0x800), 1, 0);
-	else
-	if (device->card_type >= NV_50)
-		nv_mask(device, NV50_PDISPLAY_INTR_EN_1,
-			NV50_PDISPLAY_INTR_EN_1_VBLANK_CRTC_(crtc), 0);
-	else
-		NVWriteCRTC(dev, crtc, NV_PCRTC_INTR_EN_0, 0);
-}
-
 static int
 nouveau_page_flip_reserve(struct nouveau_bo *old_bo,
 			  struct nouveau_bo *new_bo)
diff --git a/drivers/gpu/drm/nouveau/nouveau_display.h b/drivers/gpu/drm/nouveau/nouveau_display.h
index 722548b..1ea3e47 100644
--- a/drivers/gpu/drm/nouveau/nouveau_display.h
+++ b/drivers/gpu/drm/nouveau/nouveau_display.h
@@ -59,9 +59,6 @@ void nouveau_display_fini(struct drm_device *dev);
 int  nouveau_display_suspend(struct drm_device *dev);
 void nouveau_display_resume(struct drm_device *dev);
 
-int  nouveau_vblank_enable(struct drm_device *dev, int crtc);
-void nouveau_vblank_disable(struct drm_device *dev, int crtc);
-
 int  nouveau_crtc_page_flip(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 			    struct drm_pending_vblank_event *event);
 int  nouveau_finish_page_flip(struct nouveau_channel *,
diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c
index ef1ad21..ce91c8d 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_drm.c
@@ -34,6 +34,8 @@
 #include <subdev/device.h>
 #include <subdev/vm.h>
 
+#include <engine/disp.h>
+
 #include "nouveau_drm.h"
 #include "nouveau_irq.h"
 #include "nouveau_dma.h"
@@ -68,6 +70,32 @@ module_param_named(modeset, nouveau_modeset, int, 0400);
 
 static struct drm_driver driver;
 
+static int
+nouveau_drm_vblank_enable(struct drm_device *dev, int head)
+{
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nouveau_disp *pdisp = nouveau_disp(drm->device);
+	nouveau_event_get(pdisp->vblank, head, &drm->vblank);
+	return 0;
+}
+
+static void
+nouveau_drm_vblank_disable(struct drm_device *dev, int head)
+{
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nouveau_disp *pdisp = nouveau_disp(drm->device);
+	nouveau_event_put(pdisp->vblank, head, &drm->vblank);
+}
+
+static int
+nouveau_drm_vblank_handler(struct nouveau_eventh *event, int head)
+{
+	struct nouveau_drm *drm =
+		container_of(event, struct nouveau_drm, vblank);
+	drm_handle_vblank(drm->dev, head);
+	return NVKM_EVENT_KEEP;
+}
+
 static u64
 nouveau_name(struct pci_dev *pdev)
 {
@@ -259,6 +287,7 @@ nouveau_drm_load(struct drm_device *dev, unsigned long flags)
 
 	dev->dev_private = drm;
 	drm->dev = dev;
+	drm->vblank.func = nouveau_drm_vblank_handler;
 
 	INIT_LIST_HEAD(&drm->clients);
 	spin_lock_init(&drm->tile.lock);
@@ -643,8 +672,8 @@ driver = {
 	.irq_handler = nouveau_irq_handler,
 
 	.get_vblank_counter = drm_vblank_count,
-	.enable_vblank = nouveau_vblank_enable,
-	.disable_vblank = nouveau_vblank_disable,
+	.enable_vblank = nouveau_drm_vblank_enable,
+	.disable_vblank = nouveau_drm_vblank_disable,
 
 	.ioctls = nouveau_ioctls,
 	.fops = &nouveau_driver_fops,
diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.h b/drivers/gpu/drm/nouveau/nouveau_drm.h
index aa89eb9..b25df37 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drm.h
+++ b/drivers/gpu/drm/nouveau/nouveau_drm.h
@@ -13,6 +13,7 @@
 #define DRIVER_PATCHLEVEL	0
 
 #include <core/client.h>
+#include <core/event.h>
 
 #include <subdev/vm.h>
 
@@ -112,6 +113,7 @@ struct nouveau_drm {
 	struct nvbios vbios;
 	struct nouveau_display *display;
 	struct backlight_device *backlight;
+	struct nouveau_eventh vblank;
 
 	/* power management */
 	struct nouveau_pm *pm;
-- 
cgit v0.10.2


From 9bd2ddbaa241274cd11191838d080fc308ecf6c7 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Thu, 31 Jan 2013 13:51:20 +1000
Subject: drm/nouveau/fifo/nvc0-: use interrupt 31 as an event trigger

Generated if you try and use fifo method 0x20 on any subchannel, appears
that it can be safely masked off without stalling the whole GPU.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/base.c b/drivers/gpu/drm/nouveau/core/engine/fifo/base.c
index ca1057a..7341ebe 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/base.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/base.c
@@ -25,6 +25,7 @@
 #include <core/client.h>
 #include <core/object.h>
 #include <core/handle.h>
+#include <core/event.h>
 #include <core/class.h>
 
 #include <engine/dmaobj.h>
@@ -165,6 +166,7 @@ void
 nouveau_fifo_destroy(struct nouveau_fifo *priv)
 {
 	kfree(priv->channel);
+	nouveau_event_destroy(&priv->uevent);
 	nouveau_engine_destroy(&priv->base);
 }
 
@@ -189,6 +191,10 @@ nouveau_fifo_create_(struct nouveau_object *parent,
 	if (!priv->channel)
 		return -ENOMEM;
 
+	ret = nouveau_event_create(1, &priv->uevent);
+	if (ret)
+		return ret;
+
 	priv->chid = nouveau_fifo_chid;
 	spin_lock_init(&priv->lock);
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
index 2a9919b..bf86f40 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
@@ -27,6 +27,7 @@
 #include <core/namedb.h>
 #include <core/gpuobj.h>
 #include <core/engctx.h>
+#include <core/event.h>
 #include <core/class.h>
 #include <core/math.h>
 #include <core/enum.h>
@@ -583,7 +584,8 @@ nvc0_fifo_intr(struct nouveau_subdev *subdev)
 
 	if (stat & 0x80000000) {
 		u32 intr = nv_mask(priv, 0x0025a8, 0x00000000, 0x00000000);
-		nv_warn(priv, "INTR 0x80000000: 0x%08x\n", intr);
+		nouveau_event_trigger(priv->base.uevent, 0);
+		nv_debug(priv, "INTR 0x80000000: 0x%08x\n", intr);
 		stat &= ~0x80000000;
 	}
 
@@ -594,6 +596,20 @@ nvc0_fifo_intr(struct nouveau_subdev *subdev)
 	}
 }
 
+static void
+nvc0_fifo_uevent_enable(struct nouveau_event *event, int index)
+{
+	struct nvc0_fifo_priv *priv = event->priv;
+	nv_mask(priv, 0x002140, 0x80000000, 0x80000000);
+}
+
+static void
+nvc0_fifo_uevent_disable(struct nouveau_event *event, int index)
+{
+	struct nvc0_fifo_priv *priv = event->priv;
+	nv_mask(priv, 0x002140, 0x80000000, 0x00000000);
+}
+
 static int
 nvc0_fifo_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	       struct nouveau_oclass *oclass, void *data, u32 size,
@@ -627,6 +643,10 @@ nvc0_fifo_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	priv->base.uevent->enable = nvc0_fifo_uevent_enable;
+	priv->base.uevent->disable = nvc0_fifo_uevent_disable;
+	priv->base.uevent->priv = priv;
+
 	nv_subdev(priv)->unit = 0x00000100;
 	nv_subdev(priv)->intr = nvc0_fifo_intr;
 	nv_engine(priv)->cclass = &nvc0_fifo_cclass;
@@ -685,7 +705,7 @@ nvc0_fifo_init(struct nouveau_object *object)
 
 	nv_wr32(priv, 0x002a00, 0xffffffff); /* clears PFIFO.INTR bit 30 */
 	nv_wr32(priv, 0x002100, 0xffffffff);
-	nv_wr32(priv, 0x002140, 0xbfffffff);
+	nv_wr32(priv, 0x002140, 0x3fffffff);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
index 410c6e2..4419e40 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
@@ -27,6 +27,7 @@
 #include <core/namedb.h>
 #include <core/gpuobj.h>
 #include <core/engctx.h>
+#include <core/event.h>
 #include <core/class.h>
 #include <core/math.h>
 #include <core/enum.h>
@@ -554,6 +555,12 @@ nve0_fifo_intr(struct nouveau_subdev *subdev)
 		stat &= ~0x40000000;
 	}
 
+	if (stat & 0x80000000) {
+		nouveau_event_trigger(priv->base.uevent, 0);
+		nv_wr32(priv, 0x002100, 0x80000000);
+		stat &= ~0x80000000;
+	}
+
 	if (stat) {
 		nv_fatal(priv, "unhandled status 0x%08x\n", stat);
 		nv_wr32(priv, 0x002100, stat);
@@ -561,6 +568,20 @@ nve0_fifo_intr(struct nouveau_subdev *subdev)
 	}
 }
 
+static void
+nve0_fifo_uevent_enable(struct nouveau_event *event, int index)
+{
+	struct nve0_fifo_priv *priv = event->priv;
+	nv_mask(priv, 0x002140, 0x80000000, 0x80000000);
+}
+
+static void
+nve0_fifo_uevent_disable(struct nouveau_event *event, int index)
+{
+	struct nve0_fifo_priv *priv = event->priv;
+	nv_mask(priv, 0x002140, 0x80000000, 0x00000000);
+}
+
 static int
 nve0_fifo_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	       struct nouveau_oclass *oclass, void *data, u32 size,
@@ -584,6 +605,10 @@ nve0_fifo_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	priv->base.uevent->enable = nve0_fifo_uevent_enable;
+	priv->base.uevent->disable = nve0_fifo_uevent_disable;
+	priv->base.uevent->priv = priv;
+
 	nv_subdev(priv)->unit = 0x00000100;
 	nv_subdev(priv)->intr = nve0_fifo_intr;
 	nv_engine(priv)->cclass = &nve0_fifo_cclass;
@@ -634,7 +659,7 @@ nve0_fifo_init(struct nouveau_object *object)
 
 	nv_wr32(priv, 0x002a00, 0xffffffff);
 	nv_wr32(priv, 0x002100, 0xffffffff);
-	nv_wr32(priv, 0x002140, 0xbfffffff);
+	nv_wr32(priv, 0x002140, 0x3fffffff);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/fifo.h b/drivers/gpu/drm/nouveau/core/include/engine/fifo.h
index 543e4ef..b46c197 100644
--- a/drivers/gpu/drm/nouveau/core/include/engine/fifo.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/fifo.h
@@ -65,6 +65,8 @@ struct nouveau_fifo_base {
 struct nouveau_fifo {
 	struct nouveau_engine base;
 
+	struct nouveau_event *uevent;
+
 	struct nouveau_object **channel;
 	spinlock_t lock;
 	u16 min;
-- 
cgit v0.10.2


From 750087f1245e5a5d0692de4e3e7dd76d0f5b5b0a Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Thu, 31 Jan 2013 22:41:07 +1000
Subject: drm/nouveau/fifo/nv84: support user event trigger

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nv04.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nv04.c
index e34ab40..f877bd5 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nv04.c
@@ -28,6 +28,7 @@
 #include <core/namedb.h>
 #include <core/handle.h>
 #include <core/ramht.h>
+#include <core/event.h>
 
 #include <subdev/instmem.h>
 #include <subdev/instmem/nv04.h>
@@ -536,6 +537,12 @@ nv04_fifo_intr(struct nouveau_subdev *subdev)
 				status &= ~0x00000010;
 				nv_wr32(priv, 0x002100, 0x00000010);
 			}
+
+			if (status & 0x40000000) {
+				nouveau_event_trigger(priv->base.uevent, 0);
+				nv_wr32(priv, 0x002100, 0x40000000);
+				status &= ~0x40000000;
+			}
 		}
 
 		if (status) {
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nv50.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nv50.c
index cfea451..840af61 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nv50.c
@@ -481,7 +481,7 @@ nv50_fifo_init(struct nouveau_object *object)
 	nv_wr32(priv, 0x002044, 0x01003fff);
 
 	nv_wr32(priv, 0x002100, 0xffffffff);
-	nv_wr32(priv, 0x002140, 0xffffffff);
+	nv_wr32(priv, 0x002140, 0xbfffffff);
 
 	for (i = 0; i < 128; i++)
 		nv_wr32(priv, 0x002600 + (i * 4), 0x00000000);
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nv84.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nv84.c
index 8ae745b..094000e 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nv84.c
@@ -26,6 +26,7 @@
 #include <core/client.h>
 #include <core/engctx.h>
 #include <core/ramht.h>
+#include <core/event.h>
 #include <core/class.h>
 #include <core/math.h>
 
@@ -379,6 +380,20 @@ nv84_fifo_cclass = {
  * PFIFO engine
  ******************************************************************************/
 
+static void
+nv84_fifo_uevent_enable(struct nouveau_event *event, int index)
+{
+	struct nv84_fifo_priv *priv = event->priv;
+	nv_mask(priv, 0x002140, 0x40000000, 0x40000000);
+}
+
+static void
+nv84_fifo_uevent_disable(struct nouveau_event *event, int index)
+{
+	struct nv84_fifo_priv *priv = event->priv;
+	nv_mask(priv, 0x002140, 0x40000000, 0x00000000);
+}
+
 static int
 nv84_fifo_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	       struct nouveau_oclass *oclass, void *data, u32 size,
@@ -402,6 +417,10 @@ nv84_fifo_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	priv->base.uevent->enable = nv84_fifo_uevent_enable;
+	priv->base.uevent->disable = nv84_fifo_uevent_disable;
+	priv->base.uevent->priv = priv;
+
 	nv_subdev(priv)->unit = 0x00000100;
 	nv_subdev(priv)->intr = nv04_fifo_intr;
 	nv_engine(priv)->cclass = &nv84_fifo_cclass;
-- 
cgit v0.10.2


From a2fa297378c54e9b8b8ad355e34c9fbed730250b Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Thu, 31 Jan 2013 17:43:55 -0500
Subject: drm/nouveau/fifo/nvc0: bash some magic reg to make uevent interrupt
 work

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
index bf86f40..4f226af 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
@@ -706,6 +706,7 @@ nvc0_fifo_init(struct nouveau_object *object)
 	nv_wr32(priv, 0x002a00, 0xffffffff); /* clears PFIFO.INTR bit 30 */
 	nv_wr32(priv, 0x002100, 0xffffffff);
 	nv_wr32(priv, 0x002140, 0x3fffffff);
+	nv_wr32(priv, 0x002628, 0x00000001); /* makes mthd 0x20 work */
 	return 0;
 }
 
-- 
cgit v0.10.2


From e18c080fb8695d038f69c26c248f5ecbd9e8aa77 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Thu, 31 Jan 2013 14:57:33 +1000
Subject: drm/nouveau/fence/nv84-: put processes to sleep while waiting on
 fences

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nouveau_dma.h b/drivers/gpu/drm/nouveau/nouveau_dma.h
index 5c2e229..690d593 100644
--- a/drivers/gpu/drm/nouveau/nouveau_dma.h
+++ b/drivers/gpu/drm/nouveau/nouveau_dma.h
@@ -191,7 +191,7 @@ WIND_RING(struct nouveau_channel *chan)
 #define NV84_SUBCHAN_SEMAPHORE_TRIGGER_WRITE_LONG                    0x00000002
 #define NV84_SUBCHAN_SEMAPHORE_TRIGGER_ACQUIRE_GEQUAL                0x00000004
 #define NVC0_SUBCHAN_SEMAPHORE_TRIGGER_YIELD                         0x00001000
-#define NV84_SUBCHAN_NOTIFY_INTR                                     0x00000020
+#define NV84_SUBCHAN_UEVENT                                          0x00000020
 #define NV84_SUBCHAN_WRCACHE_FLUSH                                   0x00000024
 #define NV10_SUBCHAN_REF_CNT                                         0x00000050
 #define NVSW_SUBCHAN_PAGE_FLIP                                       0x00000054
diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.c b/drivers/gpu/drm/nouveau/nouveau_fence.c
index b6b8e49..b7349a6 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fence.c
+++ b/drivers/gpu/drm/nouveau/nouveau_fence.c
@@ -33,6 +33,8 @@
 #include "nouveau_dma.h"
 #include "nouveau_fence.h"
 
+#include <engine/fifo.h>
+
 void
 nouveau_fence_context_del(struct nouveau_fence_chan *fctx)
 {
@@ -107,13 +109,87 @@ nouveau_fence_done(struct nouveau_fence *fence)
 	return !fence->channel;
 }
 
+struct nouveau_fence_uevent {
+	struct nouveau_eventh handler;
+	struct nouveau_fence_priv *priv;
+};
+
+static int
+nouveau_fence_wait_uevent_handler(struct nouveau_eventh *event, int index)
+{
+	struct nouveau_fence_uevent *uevent =
+		container_of(event, struct nouveau_fence_uevent, handler);
+	wake_up_all(&uevent->priv->waiting);
+	return NVKM_EVENT_KEEP;
+}
+
+static int
+nouveau_fence_wait_uevent(struct nouveau_fence *fence, bool intr)
+
+{
+	struct nouveau_channel *chan = fence->channel;
+	struct nouveau_fifo *pfifo = nouveau_fifo(chan->drm->device);
+	struct nouveau_fence_priv *priv = chan->drm->fence;
+	struct nouveau_fence_uevent uevent = {
+		.handler.func = nouveau_fence_wait_uevent_handler,
+		.priv = priv,
+	};
+	int ret = 0;
+
+	nouveau_event_get(pfifo->uevent, 0, &uevent.handler);
+
+	if (fence->timeout) {
+		unsigned long timeout = fence->timeout - jiffies;
+
+		if (time_before(jiffies, fence->timeout)) {
+			if (intr) {
+				ret = wait_event_interruptible_timeout(
+						priv->waiting,
+						nouveau_fence_done(fence),
+						timeout);
+			} else {
+				ret = wait_event_timeout(priv->waiting,
+						nouveau_fence_done(fence),
+						timeout);
+			}
+		}
+
+		if (ret >= 0) {
+			fence->timeout = jiffies + ret;
+			if (time_after_eq(jiffies, fence->timeout))
+				ret = -EBUSY;
+		}
+	} else {
+		if (intr) {
+			ret = wait_event_interruptible(priv->waiting,
+					nouveau_fence_done(fence));
+		} else {
+			wait_event(priv->waiting, nouveau_fence_done(fence));
+		}
+	}
+
+	nouveau_event_put(pfifo->uevent, 0, &uevent.handler);
+	if (unlikely(ret < 0))
+		return ret;
+
+	return 0;
+}
+
 int
 nouveau_fence_wait(struct nouveau_fence *fence, bool lazy, bool intr)
 {
+	struct nouveau_channel *chan = fence->channel;
+	struct nouveau_fence_priv *priv = chan ? chan->drm->fence : NULL;
 	unsigned long sleep_time = NSEC_PER_MSEC / 1000;
 	ktime_t t;
 	int ret = 0;
 
+	while (priv && priv->uevent && lazy && !nouveau_fence_done(fence)) {
+		ret = nouveau_fence_wait_uevent(fence, intr);
+		if (ret < 0)
+			return ret;
+	}
+
 	while (!nouveau_fence_done(fence)) {
 		if (fence->timeout && time_after_eq(jiffies, fence->timeout)) {
 			ret = -EBUSY;
diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.h b/drivers/gpu/drm/nouveau/nouveau_fence.h
index cdb83ac..be6166e 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fence.h
+++ b/drivers/gpu/drm/nouveau/nouveau_fence.h
@@ -43,6 +43,9 @@ struct nouveau_fence_priv {
 	int  (*sync)(struct nouveau_fence *, struct nouveau_channel *,
 		     struct nouveau_channel *);
 	u32  (*read)(struct nouveau_channel *);
+
+	wait_queue_head_t waiting;
+	bool uevent;
 };
 
 #define nouveau_fence(drm) ((struct nouveau_fence_priv *)(drm)->fence)
diff --git a/drivers/gpu/drm/nouveau/nv84_fence.c b/drivers/gpu/drm/nouveau/nv84_fence.c
index c686650..e64e815 100644
--- a/drivers/gpu/drm/nouveau/nv84_fence.c
+++ b/drivers/gpu/drm/nouveau/nv84_fence.c
@@ -47,15 +47,16 @@ nv84_fence_emit(struct nouveau_fence *fence)
 {
 	struct nouveau_channel *chan = fence->channel;
 	struct nouveau_fifo_chan *fifo = (void *)chan->object;
-	int ret = RING_SPACE(chan, 7);
+	int ret = RING_SPACE(chan, 8);
 	if (ret == 0) {
 		BEGIN_NV04(chan, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 1);
 		OUT_RING  (chan, NvSema);
-		BEGIN_NV04(chan, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4);
+		BEGIN_NV04(chan, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 5);
 		OUT_RING  (chan, upper_32_bits(fifo->chid * 16));
 		OUT_RING  (chan, lower_32_bits(fifo->chid * 16));
 		OUT_RING  (chan, fence->sequence);
 		OUT_RING  (chan, NV84_SUBCHAN_SEMAPHORE_TRIGGER_WRITE_LONG);
+		OUT_RING  (chan, 0x00000000);
 		FIRE_RING (chan);
 	}
 	return ret;
@@ -174,6 +175,9 @@ nv84_fence_create(struct nouveau_drm *drm)
 	priv->base.sync = nv84_fence_sync;
 	priv->base.read = nv84_fence_read;
 
+	init_waitqueue_head(&priv->base.waiting);
+	priv->base.uevent = true;
+
 	ret = nouveau_gpuobj_new(drm->device, NULL, chan * 16, 0x1000, 0,
 				&priv->mem);
 	if (ret)
diff --git a/drivers/gpu/drm/nouveau/nvc0_fence.c b/drivers/gpu/drm/nouveau/nvc0_fence.c
index 2a56b1b..d8ed2c5 100644
--- a/drivers/gpu/drm/nouveau/nvc0_fence.c
+++ b/drivers/gpu/drm/nouveau/nvc0_fence.c
@@ -62,13 +62,14 @@ nvc0_fence_emit(struct nouveau_fence *fence)
 	u64 addr = fctx->vma.offset + fifo->chid * 16;
 	int ret;
 
-	ret = RING_SPACE(chan, 5);
+	ret = RING_SPACE(chan, 6);
 	if (ret == 0) {
-		BEGIN_NVC0(chan, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4);
+		BEGIN_NVC0(chan, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 5);
 		OUT_RING  (chan, upper_32_bits(addr));
 		OUT_RING  (chan, lower_32_bits(addr));
 		OUT_RING  (chan, fence->sequence);
 		OUT_RING  (chan, NV84_SUBCHAN_SEMAPHORE_TRIGGER_WRITE_LONG);
+		OUT_RING  (chan, 0x00000000);
 		FIRE_RING (chan);
 	}
 
@@ -217,6 +218,9 @@ nvc0_fence_create(struct nouveau_drm *drm)
 	priv->base.sync = nvc0_fence_sync;
 	priv->base.read = nvc0_fence_read;
 
+	init_waitqueue_head(&priv->base.waiting);
+	priv->base.uevent = true;
+
 	ret = nouveau_bo_new(drm->dev, 16 * (pfifo->max + 1), 0,
 			     TTM_PL_FLAG_VRAM, 0, 0, NULL, &priv->bo);
 	if (ret == 0) {
-- 
cgit v0.10.2


From 23fc09ee09c3b8f904a2220c7f71b2ff04e91219 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Sun, 3 Feb 2013 20:29:53 +1000
Subject: drm/nouveau/drm: store full dcb gpio function data in connector

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.c b/drivers/gpu/drm/nouveau/nouveau_connector.c
index 2f3e40e..a64e877 100644
--- a/drivers/gpu/drm/nouveau/nouveau_connector.c
+++ b/drivers/gpu/drm/nouveau/nouveau_connector.c
@@ -111,8 +111,8 @@ nouveau_connector_destroy(struct drm_connector *connector)
 	drm  = nouveau_drm(dev);
 	gpio = nouveau_gpio(drm->device);
 
-	if (gpio && nv_connector->hpd != DCB_GPIO_UNUSED) {
-		gpio->isr_del(gpio, 0, nv_connector->hpd, 0xff,
+	if (gpio && nv_connector->hpd.func != DCB_GPIO_UNUSED) {
+		gpio->isr_del(gpio, 0, nv_connector->hpd.func, 0xff,
 			      nouveau_connector_hotplug, connector);
 	}
 
@@ -976,8 +976,10 @@ nouveau_connector_create(struct drm_device *dev, int index)
 		if (olddcb_conntab(dev)[3] >= 4)
 			entry |= (u32)ROM16(nv_connector->dcb[2]) << 16;
 
-		nv_connector->hpd = ffs((entry & 0x07033000) >> 12);
-		nv_connector->hpd = hpd[nv_connector->hpd];
+		ret = gpio->find(gpio, 0, hpd[ffs((entry & 0x07033000) >> 12)],
+				 DCB_GPIO_UNUSED, &nv_connector->hpd);
+		if (ret)
+			nv_connector->hpd.func = DCB_GPIO_UNUSED;
 
 		nv_connector->type = nv_connector->dcb[0];
 		if (drm_conntype_from_dcb(nv_connector->type) ==
@@ -1000,7 +1002,7 @@ nouveau_connector_create(struct drm_device *dev, int index)
 		}
 	} else {
 		nv_connector->type = DCB_CONNECTOR_NONE;
-		nv_connector->hpd = DCB_GPIO_UNUSED;
+		nv_connector->hpd.func = DCB_GPIO_UNUSED;
 	}
 
 	/* no vbios data, or an unknown dcb connector type - attempt to
@@ -1127,8 +1129,8 @@ nouveau_connector_create(struct drm_device *dev, int index)
 	}
 
 	connector->polled = DRM_CONNECTOR_POLL_CONNECT;
-	if (gpio && nv_connector->hpd != DCB_GPIO_UNUSED) {
-		ret = gpio->isr_add(gpio, 0, nv_connector->hpd, 0xff,
+	if (gpio && nv_connector->hpd.func != DCB_GPIO_UNUSED) {
+		ret = gpio->isr_add(gpio, 0, nv_connector->hpd.func, 0xff,
 				    nouveau_connector_hotplug, connector);
 		if (ret == 0)
 			connector->polled = DRM_CONNECTOR_POLL_HPD;
diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.h b/drivers/gpu/drm/nouveau/nouveau_connector.h
index 20eb84c..f37250c 100644
--- a/drivers/gpu/drm/nouveau/nouveau_connector.h
+++ b/drivers/gpu/drm/nouveau/nouveau_connector.h
@@ -30,6 +30,9 @@
 #include <drm/drm_edid.h>
 #include "nouveau_crtc.h"
 
+#include <subdev/bios.h>
+#include <subdev/bios/gpio.h>
+
 struct nouveau_i2c_port;
 
 enum nouveau_underscan_type {
@@ -59,9 +62,9 @@ enum nouveau_dithering_depth {
 struct nouveau_connector {
 	struct drm_connector base;
 	enum dcb_connector_type type;
+	struct dcb_gpio_func hpd;
 	u8 index;
 	u8 *dcb;
-	u8 hpd;
 
 	int dithering_mode;
 	int dithering_depth;
diff --git a/drivers/gpu/drm/nouveau/nouveau_display.c b/drivers/gpu/drm/nouveau/nouveau_display.c
index 9f84803..a698e79 100644
--- a/drivers/gpu/drm/nouveau/nouveau_display.c
+++ b/drivers/gpu/drm/nouveau/nouveau_display.c
@@ -234,7 +234,7 @@ nouveau_display_init(struct drm_device *dev)
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
 		struct nouveau_connector *conn = nouveau_connector(connector);
 		if (gpio)
-			gpio->irq(gpio, 0, conn->hpd, 0xff, true);
+			gpio->irq(gpio, 0, conn->hpd.func, 0xff, true);
 	}
 
 	return ret;
@@ -252,7 +252,7 @@ nouveau_display_fini(struct drm_device *dev)
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
 		struct nouveau_connector *conn = nouveau_connector(connector);
 		if (gpio)
-			gpio->irq(gpio, 0, conn->hpd, 0xff, false);
+			gpio->irq(gpio, 0, conn->hpd.func, 0xff, false);
 	}
 
 	drm_kms_helper_poll_disable(dev);
diff --git a/drivers/gpu/drm/nouveau/nouveau_dp.c b/drivers/gpu/drm/nouveau/nouveau_dp.c
index 5983865..c273c2a 100644
--- a/drivers/gpu/drm/nouveau/nouveau_dp.c
+++ b/drivers/gpu/drm/nouveau/nouveau_dp.c
@@ -260,7 +260,7 @@ nouveau_dp_link_train(struct drm_encoder *encoder, u32 datarate,
 	 * we take during link training (DP_SET_POWER is one), we need
 	 * to ignore them for the moment to avoid races.
 	 */
-	gpio->irq(gpio, 0, nv_connector->hpd, 0xff, false);
+	gpio->irq(gpio, 0, nv_connector->hpd.func, 0xff, false);
 
 	/* enable down-spreading and execute pre-train script from vbios */
 	dp_link_train_init(dev, &dp, nv_encoder->dp.dpcd[3] & 1);
@@ -300,7 +300,7 @@ nouveau_dp_link_train(struct drm_encoder *encoder, u32 datarate,
 	dp_link_train_fini(dev, &dp);
 
 	/* re-enable hotplug detect */
-	gpio->irq(gpio, 0, nv_connector->hpd, 0xff, true);
+	gpio->irq(gpio, 0, nv_connector->hpd.func, 0xff, true);
 	return true;
 }
 
-- 
cgit v0.10.2


From 0f0800661a125ddb038462570c869fe6f8ab5737 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Mon, 4 Feb 2013 07:08:20 +1000
Subject: drm/nouveau/gpio: pass number of on-die gpio lines to base

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/gpio.h b/drivers/gpu/drm/nouveau/core/include/subdev/gpio.h
index b75e8f1..43cd205 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/gpio.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/gpio.h
@@ -40,15 +40,15 @@ nouveau_gpio(void *obj)
 	return (void *)nv_device(obj)->subdev[NVDEV_SUBDEV_GPIO];
 }
 
-#define nouveau_gpio_create(p,e,o,d)                                           \
-	nouveau_gpio_create_((p), (e), (o), sizeof(**d), (void **)d)
+#define nouveau_gpio_create(p,e,o,l,d)                                         \
+	nouveau_gpio_create_((p), (e), (o), (l), sizeof(**d), (void **)d)
 #define nouveau_gpio_destroy(p)                                                \
 	nouveau_subdev_destroy(&(p)->base)
 #define nouveau_gpio_fini(p,s)                                                 \
 	nouveau_subdev_fini(&(p)->base, (s))
 
 int nouveau_gpio_create_(struct nouveau_object *, struct nouveau_object *,
-			 struct nouveau_oclass *, int, void **);
+			 struct nouveau_oclass *, int, int, void **);
 int nouveau_gpio_init(struct nouveau_gpio *);
 
 extern struct nouveau_oclass nv10_gpio_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/gpio/base.c b/drivers/gpu/drm/nouveau/core/subdev/gpio/base.c
index 9fb0f9b..6f574fd 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/gpio/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/gpio/base.c
@@ -230,7 +230,8 @@ nouveau_gpio_isr_del(struct nouveau_gpio *gpio, int idx, u8 tag, u8 line,
 int
 nouveau_gpio_create_(struct nouveau_object *parent,
 		     struct nouveau_object *engine,
-		     struct nouveau_oclass *oclass, int length, void **pobject)
+		     struct nouveau_oclass *oclass, int lines,
+		     int length, void **pobject)
 {
 	struct nouveau_gpio *gpio;
 	int ret;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/gpio/nv10.c b/drivers/gpu/drm/nouveau/core/subdev/gpio/nv10.c
index 3bed67a..cf38d2a 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/gpio/nv10.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/gpio/nv10.c
@@ -112,7 +112,7 @@ nv10_gpio_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv10_gpio_priv *priv;
 	int ret;
 
-	ret = nouveau_gpio_create(parent, engine, oclass, &priv);
+	ret = nouveau_gpio_create(parent, engine, oclass, 16, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/gpio/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/gpio/nv50.c
index bf13a12..dd022a5 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/gpio/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/gpio/nv50.c
@@ -132,7 +132,9 @@ nv50_gpio_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv50_gpio_priv *priv;
 	int ret;
 
-	ret = nouveau_gpio_create(parent, engine, oclass, &priv);
+	ret = nouveau_gpio_create(parent, engine, oclass,
+				  nv_device(parent)->chipset >= 0x90 ? 32 : 16,
+				  &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/gpio/nvd0.c b/drivers/gpu/drm/nouveau/core/subdev/gpio/nvd0.c
index 83e8b8f..bc74199 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/gpio/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/gpio/nvd0.c
@@ -80,7 +80,7 @@ nvd0_gpio_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nvd0_gpio_priv *priv;
 	int ret;
 
-	ret = nouveau_gpio_create(parent, engine, oclass, &priv);
+	ret = nouveau_gpio_create(parent, engine, oclass, 32, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
-- 
cgit v0.10.2


From 4f47643dbb4c345c5beebe53588682a7ff2c872a Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Sun, 3 Feb 2013 12:56:16 +1000
Subject: drm/nouveau/gpio: use event interfaces for interrupt signalling

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/gpio.h b/drivers/gpu/drm/nouveau/core/include/subdev/gpio.h
index 43cd205..b0007b5 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/gpio.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/gpio.h
@@ -3,6 +3,7 @@
 
 #include <core/subdev.h>
 #include <core/device.h>
+#include <core/event.h>
 
 #include <subdev/bios.h>
 #include <subdev/bios/gpio.h>
@@ -10,28 +11,18 @@
 struct nouveau_gpio {
 	struct nouveau_subdev base;
 
+	struct nouveau_event *events;
+
 	/* hardware interfaces */
 	void (*reset)(struct nouveau_gpio *, u8 func);
 	int  (*drive)(struct nouveau_gpio *, int line, int dir, int out);
 	int  (*sense)(struct nouveau_gpio *, int line);
-	void (*irq_enable)(struct nouveau_gpio *, int line, bool);
 
 	/* software interfaces */
 	int  (*find)(struct nouveau_gpio *, int idx, u8 tag, u8 line,
 		     struct dcb_gpio_func *);
 	int  (*set)(struct nouveau_gpio *, int idx, u8 tag, u8 line, int state);
 	int  (*get)(struct nouveau_gpio *, int idx, u8 tag, u8 line);
-	int  (*irq)(struct nouveau_gpio *, int idx, u8 tag, u8 line, bool on);
-
-	/* interrupt handling */
-	struct list_head isr;
-	spinlock_t lock;
-
-	void (*isr_run)(struct nouveau_gpio *, int idx, u32 mask);
-	int  (*isr_add)(struct nouveau_gpio *, int idx, u8 tag, u8 line,
-			void (*)(void *, int state), void *data);
-	void (*isr_del)(struct nouveau_gpio *, int idx, u8 tag, u8 line,
-			void (*)(void *, int state), void *data);
 };
 
 static inline struct nouveau_gpio *
@@ -42,14 +33,17 @@ nouveau_gpio(void *obj)
 
 #define nouveau_gpio_create(p,e,o,l,d)                                         \
 	nouveau_gpio_create_((p), (e), (o), (l), sizeof(**d), (void **)d)
-#define nouveau_gpio_destroy(p)                                                \
-	nouveau_subdev_destroy(&(p)->base)
+#define nouveau_gpio_destroy(p) ({                                             \
+	struct nouveau_gpio *gpio = (p);                                       \
+	_nouveau_gpio_dtor(nv_object(gpio));                                   \
+})
 #define nouveau_gpio_fini(p,s)                                                 \
 	nouveau_subdev_fini(&(p)->base, (s))
 
-int nouveau_gpio_create_(struct nouveau_object *, struct nouveau_object *,
-			 struct nouveau_oclass *, int, int, void **);
-int nouveau_gpio_init(struct nouveau_gpio *);
+int  nouveau_gpio_create_(struct nouveau_object *, struct nouveau_object *,
+			  struct nouveau_oclass *, int, int, void **);
+void _nouveau_gpio_dtor(struct nouveau_object *);
+int  nouveau_gpio_init(struct nouveau_gpio *);
 
 extern struct nouveau_oclass nv10_gpio_oclass;
 extern struct nouveau_oclass nv50_gpio_oclass;
@@ -59,6 +53,7 @@ void nv50_gpio_dtor(struct nouveau_object *);
 int  nv50_gpio_init(struct nouveau_object *);
 int  nv50_gpio_fini(struct nouveau_object *, bool);
 void nv50_gpio_intr(struct nouveau_subdev *);
-void nv50_gpio_irq_enable(struct nouveau_gpio *, int line, bool);
+void nv50_gpio_intr_enable(struct nouveau_event *, int line);
+void nv50_gpio_intr_disable(struct nouveau_event *, int line);
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/gpio/base.c b/drivers/gpu/drm/nouveau/core/subdev/gpio/base.c
index 6f574fd..d422acc 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/gpio/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/gpio/base.c
@@ -102,129 +102,12 @@ nouveau_gpio_get(struct nouveau_gpio *gpio, int idx, u8 tag, u8 line)
 	return ret;
 }
 
-static int
-nouveau_gpio_irq(struct nouveau_gpio *gpio, int idx, u8 tag, u8 line, bool on)
-{
-	struct dcb_gpio_func func;
-	int ret;
-
-	ret = nouveau_gpio_find(gpio, idx, tag, line, &func);
-	if (ret == 0) {
-		if (idx == 0 && gpio->irq_enable)
-			gpio->irq_enable(gpio, func.line, on);
-		else
-			ret = -ENODEV;
-	}
-
-	return ret;
-}
-
-struct gpio_isr {
-	struct nouveau_gpio *gpio;
-	struct list_head head;
-	struct work_struct work;
-	int idx;
-	struct dcb_gpio_func func;
-	void (*handler)(void *, int);
-	void *data;
-	bool inhibit;
-};
-
-static void
-nouveau_gpio_isr_bh(struct work_struct *work)
-{
-	struct gpio_isr *isr = container_of(work, struct gpio_isr, work);
-	struct nouveau_gpio *gpio = isr->gpio;
-	unsigned long flags;
-	int state;
-
-	state = nouveau_gpio_get(gpio, isr->idx, isr->func.func,
-						 isr->func.line);
-	if (state >= 0)
-		isr->handler(isr->data, state);
-
-	spin_lock_irqsave(&gpio->lock, flags);
-	isr->inhibit = false;
-	spin_unlock_irqrestore(&gpio->lock, flags);
-}
-
-static void
-nouveau_gpio_isr_run(struct nouveau_gpio *gpio, int idx, u32 line_mask)
-{
-	struct gpio_isr *isr;
-
-	if (idx != 0)
-		return;
-
-	spin_lock(&gpio->lock);
-	list_for_each_entry(isr, &gpio->isr, head) {
-		if (line_mask & (1 << isr->func.line)) {
-			if (isr->inhibit)
-				continue;
-			isr->inhibit = true;
-			schedule_work(&isr->work);
-		}
-	}
-	spin_unlock(&gpio->lock);
-}
-
-static int
-nouveau_gpio_isr_add(struct nouveau_gpio *gpio, int idx, u8 tag, u8 line,
-		     void (*handler)(void *, int), void *data)
-{
-	struct gpio_isr *isr;
-	unsigned long flags;
-	int ret;
-
-	isr = kzalloc(sizeof(*isr), GFP_KERNEL);
-	if (!isr)
-		return -ENOMEM;
-
-	ret = nouveau_gpio_find(gpio, idx, tag, line, &isr->func);
-	if (ret) {
-		kfree(isr);
-		return ret;
-	}
-
-	INIT_WORK(&isr->work, nouveau_gpio_isr_bh);
-	isr->gpio = gpio;
-	isr->handler = handler;
-	isr->data = data;
-	isr->idx = idx;
-
-	spin_lock_irqsave(&gpio->lock, flags);
-	list_add(&isr->head, &gpio->isr);
-	spin_unlock_irqrestore(&gpio->lock, flags);
-	return 0;
-}
-
-static void
-nouveau_gpio_isr_del(struct nouveau_gpio *gpio, int idx, u8 tag, u8 line,
-		     void (*handler)(void *, int), void *data)
+void
+_nouveau_gpio_dtor(struct nouveau_object *object)
 {
-	struct gpio_isr *isr, *tmp;
-	struct dcb_gpio_func func;
-	unsigned long flags;
-	LIST_HEAD(tofree);
-	int ret;
-
-	ret = nouveau_gpio_find(gpio, idx, tag, line, &func);
-	if (ret == 0) {
-		spin_lock_irqsave(&gpio->lock, flags);
-		list_for_each_entry_safe(isr, tmp, &gpio->isr, head) {
-			if (memcmp(&isr->func, &func, sizeof(func)) ||
-			    isr->idx != idx ||
-			    isr->handler != handler || isr->data != data)
-				continue;
-			list_move_tail(&isr->head, &tofree);
-		}
-		spin_unlock_irqrestore(&gpio->lock, flags);
-
-		list_for_each_entry_safe(isr, tmp, &tofree, head) {
-			flush_work(&isr->work);
-			kfree(isr);
-		}
-	}
+	struct nouveau_gpio *gpio = (void *)object;
+	nouveau_event_destroy(&gpio->events);
+	nouveau_subdev_destroy(&gpio->base);
 }
 
 int
@@ -242,15 +125,13 @@ nouveau_gpio_create_(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
+	ret = nouveau_event_create(lines, &gpio->events);
+	if (ret)
+		return ret;
+
 	gpio->find = nouveau_gpio_find;
 	gpio->set  = nouveau_gpio_set;
 	gpio->get  = nouveau_gpio_get;
-	gpio->irq  = nouveau_gpio_irq;
-	gpio->isr_run = nouveau_gpio_isr_run;
-	gpio->isr_add = nouveau_gpio_isr_add;
-	gpio->isr_del = nouveau_gpio_isr_del;
-	INIT_LIST_HEAD(&gpio->isr);
-	spin_lock_init(&gpio->lock);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/gpio/nv10.c b/drivers/gpu/drm/nouveau/core/subdev/gpio/nv10.c
index cf38d2a..9665f5f 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/gpio/nv10.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/gpio/nv10.c
@@ -83,27 +83,36 @@ nv10_gpio_drive(struct nouveau_gpio *gpio, int line, int dir, int out)
 }
 
 static void
-nv10_gpio_irq_enable(struct nouveau_gpio *gpio, int line, bool on)
-{
-	u32 mask = 0x00010001 << line;
-
-	nv_wr32(gpio, 0x001104, mask);
-	nv_mask(gpio, 0x001144, mask, on ? mask : 0);
-}
-
-static void
 nv10_gpio_intr(struct nouveau_subdev *subdev)
 {
 	struct nv10_gpio_priv *priv = (void *)subdev;
 	u32 intr = nv_rd32(priv, 0x001104);
 	u32 hi = (intr & 0x0000ffff) >> 0;
 	u32 lo = (intr & 0xffff0000) >> 16;
+	int i;
 
-	priv->base.isr_run(&priv->base, 0, hi | lo);
+	for (i = 0; (hi | lo) && i < 32; i++) {
+		if ((hi | lo) & (1 << i))
+			nouveau_event_trigger(priv->base.events, i);
+	}
 
 	nv_wr32(priv, 0x001104, intr);
 }
 
+static void
+nv10_gpio_intr_enable(struct nouveau_event *event, int line)
+{
+	nv_wr32(event->priv, 0x001104, 0x00010001 << line);
+	nv_mask(event->priv, 0x001144, 0x00010001 << line, 0x00010001 << line);
+}
+
+static void
+nv10_gpio_intr_disable(struct nouveau_event *event, int line)
+{
+	nv_wr32(event->priv, 0x001104, 0x00010001 << line);
+	nv_mask(event->priv, 0x001144, 0x00010001 << line, 0x00000000);
+}
+
 static int
 nv10_gpio_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	       struct nouveau_oclass *oclass, void *data, u32 size,
@@ -119,7 +128,9 @@ nv10_gpio_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 
 	priv->base.drive = nv10_gpio_drive;
 	priv->base.sense = nv10_gpio_sense;
-	priv->base.irq_enable = nv10_gpio_irq_enable;
+	priv->base.events->priv = priv;
+	priv->base.events->enable = nv10_gpio_intr_enable;
+	priv->base.events->disable = nv10_gpio_intr_disable;
 	nv_subdev(priv)->intr = nv10_gpio_intr;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/gpio/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/gpio/nv50.c
index dd022a5..cbe609a 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/gpio/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/gpio/nv50.c
@@ -95,21 +95,12 @@ nv50_gpio_sense(struct nouveau_gpio *gpio, int line)
 }
 
 void
-nv50_gpio_irq_enable(struct nouveau_gpio *gpio, int line, bool on)
-{
-	u32 reg  = line < 16 ? 0xe050 : 0xe070;
-	u32 mask = 0x00010001 << (line & 0xf);
-
-	nv_wr32(gpio, reg + 4, mask);
-	nv_mask(gpio, reg + 0, mask, on ? mask : 0);
-}
-
-void
 nv50_gpio_intr(struct nouveau_subdev *subdev)
 {
 	struct nv50_gpio_priv *priv = (void *)subdev;
 	u32 intr0, intr1 = 0;
 	u32 hi, lo;
+	int i;
 
 	intr0 = nv_rd32(priv, 0xe054) & nv_rd32(priv, 0xe050);
 	if (nv_device(priv)->chipset >= 0x90)
@@ -117,13 +108,35 @@ nv50_gpio_intr(struct nouveau_subdev *subdev)
 
 	hi = (intr0 & 0x0000ffff) | (intr1 << 16);
 	lo = (intr0 >> 16) | (intr1 & 0xffff0000);
-	priv->base.isr_run(&priv->base, 0, hi | lo);
+
+	for (i = 0; (hi | lo) && i < 32; i++) {
+		if ((hi | lo) & (1 << i))
+			nouveau_event_trigger(priv->base.events, i);
+	}
 
 	nv_wr32(priv, 0xe054, intr0);
 	if (nv_device(priv)->chipset >= 0x90)
 		nv_wr32(priv, 0xe074, intr1);
 }
 
+void
+nv50_gpio_intr_enable(struct nouveau_event *event, int line)
+{
+	const u32 addr = line < 16 ? 0xe050 : 0xe070;
+	const u32 mask = 0x00010001 << (line & 0xf);
+	nv_wr32(event->priv, addr + 0x04, mask);
+	nv_mask(event->priv, addr + 0x00, mask, mask);
+}
+
+void
+nv50_gpio_intr_disable(struct nouveau_event *event, int line)
+{
+	const u32 addr = line < 16 ? 0xe050 : 0xe070;
+	const u32 mask = 0x00010001 << (line & 0xf);
+	nv_wr32(event->priv, addr + 0x04, mask);
+	nv_mask(event->priv, addr + 0x00, mask, 0x00000000);
+}
+
 static int
 nv50_gpio_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	       struct nouveau_oclass *oclass, void *data, u32 size,
@@ -142,7 +155,9 @@ nv50_gpio_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->base.reset = nv50_gpio_reset;
 	priv->base.drive = nv50_gpio_drive;
 	priv->base.sense = nv50_gpio_sense;
-	priv->base.irq_enable = nv50_gpio_irq_enable;
+	priv->base.events->priv = priv;
+	priv->base.events->enable = nv50_gpio_intr_enable;
+	priv->base.events->disable = nv50_gpio_intr_disable;
 	nv_subdev(priv)->intr = nv50_gpio_intr;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/gpio/nvd0.c b/drivers/gpu/drm/nouveau/core/subdev/gpio/nvd0.c
index bc74199..5ef16a2 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/gpio/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/gpio/nvd0.c
@@ -88,7 +88,9 @@ nvd0_gpio_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->base.reset = nvd0_gpio_reset;
 	priv->base.drive = nvd0_gpio_drive;
 	priv->base.sense = nvd0_gpio_sense;
-	priv->base.irq_enable = nv50_gpio_irq_enable;
+	priv->base.events->priv = priv;
+	priv->base.events->enable = nv50_gpio_intr_enable;
+	priv->base.events->disable = nv50_gpio_intr_disable;
 	nv_subdev(priv)->intr = nv50_gpio_intr;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.c b/drivers/gpu/drm/nouveau/nouveau_connector.c
index a64e877..9c4b3f5 100644
--- a/drivers/gpu/drm/nouveau/nouveau_connector.c
+++ b/drivers/gpu/drm/nouveau/nouveau_connector.c
@@ -55,8 +55,6 @@ MODULE_PARM_DESC(duallink, "Allow dual-link TMDS (default: enabled)");
 static int nouveau_duallink = 1;
 module_param_named(duallink, nouveau_duallink, int, 0400);
 
-static void nouveau_connector_hotplug(void *, int);
-
 struct nouveau_encoder *
 find_encoder(struct drm_connector *connector, int type)
 {
@@ -100,22 +98,6 @@ static void
 nouveau_connector_destroy(struct drm_connector *connector)
 {
 	struct nouveau_connector *nv_connector = nouveau_connector(connector);
-	struct nouveau_gpio *gpio;
-	struct nouveau_drm *drm;
-	struct drm_device *dev;
-
-	if (!nv_connector)
-		return;
-
-	dev  = nv_connector->base.dev;
-	drm  = nouveau_drm(dev);
-	gpio = nouveau_gpio(drm->device);
-
-	if (gpio && nv_connector->hpd.func != DCB_GPIO_UNUSED) {
-		gpio->isr_del(gpio, 0, nv_connector->hpd.func, 0xff,
-			      nouveau_connector_hotplug, connector);
-	}
-
 	kfree(nv_connector->edid);
 	drm_sysfs_connector_remove(connector);
 	drm_connector_cleanup(connector);
@@ -912,6 +894,37 @@ nouveau_connector_funcs_lvds = {
 	.force = nouveau_connector_force
 };
 
+static void
+nouveau_connector_hotplug_work(struct work_struct *work)
+{
+	struct nouveau_connector *nv_connector =
+		container_of(work, struct nouveau_connector, hpd_work);
+	struct drm_connector *connector = &nv_connector->base;
+	struct drm_device *dev = connector->dev;
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nouveau_gpio *gpio = nouveau_gpio(drm->device);
+	bool plugged = gpio->get(gpio, 0, nv_connector->hpd.func, 0xff);
+
+	NV_DEBUG(drm, "%splugged %s\n", plugged ? "" : "un",
+		 drm_get_connector_name(connector));
+
+	if (plugged)
+		drm_helper_connector_dpms(connector, DRM_MODE_DPMS_ON);
+	else
+		drm_helper_connector_dpms(connector, DRM_MODE_DPMS_OFF);
+
+	drm_helper_hpd_irq_event(dev);
+}
+
+static int
+nouveau_connector_hotplug(struct nouveau_eventh *event, int index)
+{
+	struct nouveau_connector *nv_connector =
+		container_of(event, struct nouveau_connector, hpd_func);
+	schedule_work(&nv_connector->hpd_work);
+	return NVKM_EVENT_KEEP;
+}
+
 static int
 drm_conntype_from_dcb(enum dcb_connector_type dcb)
 {
@@ -962,6 +975,7 @@ nouveau_connector_create(struct drm_device *dev, int index)
 		return ERR_PTR(-ENOMEM);
 
 	connector = &nv_connector->base;
+	INIT_WORK(&nv_connector->hpd_work, nouveau_connector_hotplug_work);
 	nv_connector->index = index;
 
 	/* attempt to parse vbios connector type and hotplug gpio */
@@ -978,6 +992,7 @@ nouveau_connector_create(struct drm_device *dev, int index)
 
 		ret = gpio->find(gpio, 0, hpd[ffs((entry & 0x07033000) >> 12)],
 				 DCB_GPIO_UNUSED, &nv_connector->hpd);
+		nv_connector->hpd_func.func = nouveau_connector_hotplug;
 		if (ret)
 			nv_connector->hpd.func = DCB_GPIO_UNUSED;
 
@@ -1129,31 +1144,9 @@ nouveau_connector_create(struct drm_device *dev, int index)
 	}
 
 	connector->polled = DRM_CONNECTOR_POLL_CONNECT;
-	if (gpio && nv_connector->hpd.func != DCB_GPIO_UNUSED) {
-		ret = gpio->isr_add(gpio, 0, nv_connector->hpd.func, 0xff,
-				    nouveau_connector_hotplug, connector);
-		if (ret == 0)
-			connector->polled = DRM_CONNECTOR_POLL_HPD;
-	}
+	if (nv_connector->hpd.func != DCB_GPIO_UNUSED)
+		connector->polled = DRM_CONNECTOR_POLL_HPD;
 
 	drm_sysfs_connector_add(connector);
 	return connector;
 }
-
-static void
-nouveau_connector_hotplug(void *data, int plugged)
-{
-	struct drm_connector *connector = data;
-	struct drm_device *dev = connector->dev;
-	struct nouveau_drm *drm = nouveau_drm(dev);
-
-	NV_DEBUG(drm, "%splugged %s\n", plugged ? "" : "un",
-		 drm_get_connector_name(connector));
-
-	if (plugged)
-		drm_helper_connector_dpms(connector, DRM_MODE_DPMS_ON);
-	else
-		drm_helper_connector_dpms(connector, DRM_MODE_DPMS_OFF);
-
-	drm_helper_hpd_irq_event(dev);
-}
diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.h b/drivers/gpu/drm/nouveau/nouveau_connector.h
index f37250c..6e399aa 100644
--- a/drivers/gpu/drm/nouveau/nouveau_connector.h
+++ b/drivers/gpu/drm/nouveau/nouveau_connector.h
@@ -30,6 +30,8 @@
 #include <drm/drm_edid.h>
 #include "nouveau_crtc.h"
 
+#include <core/event.h>
+
 #include <subdev/bios.h>
 #include <subdev/bios/gpio.h>
 
@@ -62,10 +64,13 @@ enum nouveau_dithering_depth {
 struct nouveau_connector {
 	struct drm_connector base;
 	enum dcb_connector_type type;
-	struct dcb_gpio_func hpd;
 	u8 index;
 	u8 *dcb;
 
+	struct dcb_gpio_func hpd;
+	struct work_struct hpd_work;
+	struct nouveau_eventh hpd_func;
+
 	int dithering_mode;
 	int dithering_depth;
 	int scaling_mode;
diff --git a/drivers/gpu/drm/nouveau/nouveau_display.c b/drivers/gpu/drm/nouveau/nouveau_display.c
index a698e79..78fc5aa 100644
--- a/drivers/gpu/drm/nouveau/nouveau_display.c
+++ b/drivers/gpu/drm/nouveau/nouveau_display.c
@@ -233,8 +233,10 @@ nouveau_display_init(struct drm_device *dev)
 	/* enable hotplug interrupts */
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
 		struct nouveau_connector *conn = nouveau_connector(connector);
-		if (gpio)
-			gpio->irq(gpio, 0, conn->hpd.func, 0xff, true);
+		if (gpio && conn->hpd.func != DCB_GPIO_UNUSED) {
+			nouveau_event_get(gpio->events, conn->hpd.line,
+					 &conn->hpd_func);
+		}
 	}
 
 	return ret;
@@ -251,8 +253,10 @@ nouveau_display_fini(struct drm_device *dev)
 	/* disable hotplug interrupts */
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
 		struct nouveau_connector *conn = nouveau_connector(connector);
-		if (gpio)
-			gpio->irq(gpio, 0, conn->hpd.func, 0xff, false);
+		if (gpio && conn->hpd.func != DCB_GPIO_UNUSED) {
+			nouveau_event_put(gpio->events, conn->hpd.line,
+					 &conn->hpd_func);
+		}
 	}
 
 	drm_kms_helper_poll_disable(dev);
diff --git a/drivers/gpu/drm/nouveau/nouveau_dp.c b/drivers/gpu/drm/nouveau/nouveau_dp.c
index c273c2a..a87c367 100644
--- a/drivers/gpu/drm/nouveau/nouveau_dp.c
+++ b/drivers/gpu/drm/nouveau/nouveau_dp.c
@@ -260,7 +260,8 @@ nouveau_dp_link_train(struct drm_encoder *encoder, u32 datarate,
 	 * we take during link training (DP_SET_POWER is one), we need
 	 * to ignore them for the moment to avoid races.
 	 */
-	gpio->irq(gpio, 0, nv_connector->hpd.func, 0xff, false);
+	nouveau_event_put(gpio->events, nv_connector->hpd.line,
+			 &nv_connector->hpd_func);
 
 	/* enable down-spreading and execute pre-train script from vbios */
 	dp_link_train_init(dev, &dp, nv_encoder->dp.dpcd[3] & 1);
@@ -300,7 +301,8 @@ nouveau_dp_link_train(struct drm_encoder *encoder, u32 datarate,
 	dp_link_train_fini(dev, &dp);
 
 	/* re-enable hotplug detect */
-	gpio->irq(gpio, 0, nv_connector->hpd.func, 0xff, true);
+	nouveau_event_get(gpio->events, nv_connector->hpd.line,
+			 &nv_connector->hpd_func);
 	return true;
 }
 
-- 
cgit v0.10.2


From fa531bc8b4278010fd11819c089f6679890addee Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Wed, 13 Feb 2013 13:34:39 +1000
Subject: drm/nouveau/gpio/nve0: interrupt regs moved on kepler apparently

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index ebe37fe..e0bd21a 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -91,6 +91,7 @@ nouveau-y += core/subdev/gpio/base.o
 nouveau-y += core/subdev/gpio/nv10.o
 nouveau-y += core/subdev/gpio/nv50.o
 nouveau-y += core/subdev/gpio/nvd0.o
+nouveau-y += core/subdev/gpio/nve0.o
 nouveau-y += core/subdev/i2c/base.o
 nouveau-y += core/subdev/i2c/aux.o
 nouveau-y += core/subdev/i2c/bit.o
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/gpio.h b/drivers/gpu/drm/nouveau/core/include/subdev/gpio.h
index b0007b5..c85b9f1 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/gpio.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/gpio.h
@@ -48,12 +48,6 @@ int  nouveau_gpio_init(struct nouveau_gpio *);
 extern struct nouveau_oclass nv10_gpio_oclass;
 extern struct nouveau_oclass nv50_gpio_oclass;
 extern struct nouveau_oclass nvd0_gpio_oclass;
-
-void nv50_gpio_dtor(struct nouveau_object *);
-int  nv50_gpio_init(struct nouveau_object *);
-int  nv50_gpio_fini(struct nouveau_object *, bool);
-void nv50_gpio_intr(struct nouveau_subdev *);
-void nv50_gpio_intr_enable(struct nouveau_event *, int line);
-void nv50_gpio_intr_disable(struct nouveau_event *, int line);
+extern struct nouveau_oclass nve0_gpio_oclass;
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nve0.c b/drivers/gpu/drm/nouveau/core/subdev/device/nve0.c
index 59d0d0f..1e02bec 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/device/nve0.c
@@ -57,7 +57,7 @@ nve0_identify(struct nouveau_device *device)
 	case 0xe4:
 		device->cname = "GK104";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nvd0_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nve0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
@@ -86,7 +86,7 @@ nve0_identify(struct nouveau_device *device)
 	case 0xe7:
 		device->cname = "GK107";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nvd0_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nve0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
@@ -115,7 +115,7 @@ nve0_identify(struct nouveau_device *device)
 	case 0xe6:
 		device->cname = "GK106";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nvd0_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nve0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/gpio/nv10.c b/drivers/gpu/drm/nouveau/core/subdev/gpio/nv10.c
index 9665f5f..76d5d54 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/gpio/nv10.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/gpio/nv10.c
@@ -24,7 +24,7 @@
  *
  */
 
-#include <subdev/gpio.h>
+#include "priv.h"
 
 struct nv10_gpio_priv {
 	struct nouveau_gpio base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/gpio/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/gpio/nv50.c
index cbe609a..bf489dc 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/gpio/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/gpio/nv50.c
@@ -22,7 +22,7 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/gpio.h>
+#include "priv.h"
 
 struct nv50_gpio_priv {
 	struct nouveau_gpio base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/gpio/nvd0.c b/drivers/gpu/drm/nouveau/core/subdev/gpio/nvd0.c
index 5ef16a2..010431e 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/gpio/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/gpio/nvd0.c
@@ -22,13 +22,13 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/gpio.h>
+#include "priv.h"
 
 struct nvd0_gpio_priv {
 	struct nouveau_gpio base;
 };
 
-static void
+void
 nvd0_gpio_reset(struct nouveau_gpio *gpio, u8 match)
 {
 	struct nouveau_bios *bios = nouveau_bios(gpio);
@@ -57,7 +57,7 @@ nvd0_gpio_reset(struct nouveau_gpio *gpio, u8 match)
 	}
 }
 
-static int
+int
 nvd0_gpio_drive(struct nouveau_gpio *gpio, int line, int dir, int out)
 {
 	u32 data = ((dir ^ 1) << 13) | (out << 12);
@@ -66,7 +66,7 @@ nvd0_gpio_drive(struct nouveau_gpio *gpio, int line, int dir, int out)
 	return 0;
 }
 
-static int
+int
 nvd0_gpio_sense(struct nouveau_gpio *gpio, int line)
 {
 	return !!(nv_rd32(gpio, 0x00d610 + (line * 4)) & 0x00004000);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/gpio/nve0.c b/drivers/gpu/drm/nouveau/core/subdev/gpio/nve0.c
new file mode 100644
index 0000000..16b8c5b
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/gpio/nve0.c
@@ -0,0 +1,131 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+struct nve0_gpio_priv {
+	struct nouveau_gpio base;
+};
+
+void
+nve0_gpio_intr(struct nouveau_subdev *subdev)
+{
+	struct nve0_gpio_priv *priv = (void *)subdev;
+	u32 intr0 = nv_rd32(priv, 0xdc00) & nv_rd32(priv, 0xdc08);
+	u32 intr1 = nv_rd32(priv, 0xdc80) & nv_rd32(priv, 0xdc88);
+	u32 hi = (intr0 & 0x0000ffff) | (intr1 << 16);
+	u32 lo = (intr0 >> 16) | (intr1 & 0xffff0000);
+	int i;
+
+	for (i = 0; (hi | lo) && i < 32; i++) {
+		if ((hi | lo) & (1 << i))
+			nouveau_event_trigger(priv->base.events, i);
+	}
+
+	nv_wr32(priv, 0xdc00, intr0);
+	nv_wr32(priv, 0xdc88, intr1);
+}
+
+void
+nve0_gpio_intr_enable(struct nouveau_event *event, int line)
+{
+	const u32 addr = line < 16 ? 0xdc00 : 0xdc80;
+	const u32 mask = 0x00010001 << (line & 0xf);
+	nv_wr32(event->priv, addr + 0x08, mask);
+	nv_mask(event->priv, addr + 0x00, mask, mask);
+}
+
+void
+nve0_gpio_intr_disable(struct nouveau_event *event, int line)
+{
+	const u32 addr = line < 16 ? 0xdc00 : 0xdc80;
+	const u32 mask = 0x00010001 << (line & 0xf);
+	nv_wr32(event->priv, addr + 0x08, mask);
+	nv_mask(event->priv, addr + 0x00, mask, 0x00000000);
+}
+
+int
+nve0_gpio_fini(struct nouveau_object *object, bool suspend)
+{
+	struct nve0_gpio_priv *priv = (void *)object;
+	nv_wr32(priv, 0xdc08, 0x00000000);
+	nv_wr32(priv, 0xdc88, 0x00000000);
+	return nouveau_gpio_fini(&priv->base, suspend);
+}
+
+int
+nve0_gpio_init(struct nouveau_object *object)
+{
+	struct nve0_gpio_priv *priv = (void *)object;
+	int ret;
+
+	ret = nouveau_gpio_init(&priv->base);
+	if (ret)
+		return ret;
+
+	nv_wr32(priv, 0xdc00, 0xffffffff);
+	nv_wr32(priv, 0xdc80, 0xffffffff);
+	return 0;
+}
+
+void
+nve0_gpio_dtor(struct nouveau_object *object)
+{
+	struct nve0_gpio_priv *priv = (void *)object;
+	nouveau_gpio_destroy(&priv->base);
+}
+
+static int
+nve0_gpio_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+	       struct nouveau_oclass *oclass, void *data, u32 size,
+	       struct nouveau_object **pobject)
+{
+	struct nve0_gpio_priv *priv;
+	int ret;
+
+	ret = nouveau_gpio_create(parent, engine, oclass, 32, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	priv->base.reset = nvd0_gpio_reset;
+	priv->base.drive = nvd0_gpio_drive;
+	priv->base.sense = nvd0_gpio_sense;
+	priv->base.events->priv = priv;
+	priv->base.events->enable = nve0_gpio_intr_enable;
+	priv->base.events->disable = nve0_gpio_intr_disable;
+	nv_subdev(priv)->intr = nve0_gpio_intr;
+	return 0;
+}
+
+struct nouveau_oclass
+nve0_gpio_oclass = {
+	.handle = NV_SUBDEV(GPIO, 0xe0),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nve0_gpio_ctor,
+		.dtor = nv50_gpio_dtor,
+		.init = nve0_gpio_init,
+		.fini = nve0_gpio_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/gpio/priv.h b/drivers/gpu/drm/nouveau/core/subdev/gpio/priv.h
new file mode 100644
index 0000000..2ee1c89
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/gpio/priv.h
@@ -0,0 +1,17 @@
+#ifndef __NVKM_GPIO_H__
+#define __NVKM_GPIO_H__
+
+#include <subdev/gpio.h>
+
+void nv50_gpio_dtor(struct nouveau_object *);
+int  nv50_gpio_init(struct nouveau_object *);
+int  nv50_gpio_fini(struct nouveau_object *, bool);
+void nv50_gpio_intr(struct nouveau_subdev *);
+void nv50_gpio_intr_enable(struct nouveau_event *, int line);
+void nv50_gpio_intr_disable(struct nouveau_event *, int line);
+
+void nvd0_gpio_reset(struct nouveau_gpio *, u8);
+int  nvd0_gpio_drive(struct nouveau_gpio *, int, int, int);
+int  nvd0_gpio_sense(struct nouveau_gpio *, int);
+
+#endif
-- 
cgit v0.10.2


From a34caf78f26bda63869471cb3f46f354f4658758 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Thu, 14 Feb 2013 09:28:37 +1000
Subject: drm/nv84/fence: access fences with full virtual address, not offset

Allows most of the code to be shared between nv84/nvc0 implementations,
and paves the way for doing emit/sync on non-VRAM buffers (multi-gpu,
dma-buf).

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.h b/drivers/gpu/drm/nouveau/nouveau_fence.h
index be6166e..2324911 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fence.h
+++ b/drivers/gpu/drm/nouveau/nouveau_fence.h
@@ -68,8 +68,27 @@ void nv17_fence_resume(struct nouveau_drm *drm);
 int nv50_fence_create(struct nouveau_drm *);
 int nv84_fence_create(struct nouveau_drm *);
 int nvc0_fence_create(struct nouveau_drm *);
-u64 nvc0_fence_crtc(struct nouveau_channel *, int crtc);
 
 int nouveau_flip_complete(void *chan);
 
+struct nv84_fence_chan {
+	struct nouveau_fence_chan base;
+	struct nouveau_vma vma;
+	struct nouveau_vma dispc_vma[4];
+};
+
+struct nv84_fence_priv {
+	struct nouveau_fence_priv base;
+	struct nouveau_bo *bo;
+	u32 *suspend;
+};
+
+u64  nv84_fence_crtc(struct nouveau_channel *, int);
+u32  nv84_fence_read(struct nouveau_channel *);
+int  nv84_fence_context_new(struct nouveau_channel *);
+void nv84_fence_context_del(struct nouveau_channel *);
+bool nv84_fence_suspend(struct nouveau_drm *);
+void nv84_fence_resume(struct nouveau_drm *);
+void nv84_fence_destroy(struct nouveau_drm *);
+
 #endif
diff --git a/drivers/gpu/drm/nouveau/nv50_display.c b/drivers/gpu/drm/nouveau/nv50_display.c
index c9da4f1..102a873 100644
--- a/drivers/gpu/drm/nouveau/nv50_display.c
+++ b/drivers/gpu/drm/nouveau/nv50_display.c
@@ -502,7 +502,7 @@ nv50_display_flip_next(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 		if (ret)
 			return ret;
 
-		if (nv_mclass(chan->object) < NVC0_CHANNEL_IND_CLASS) {
+		if (nv_mclass(chan->object) < NV84_CHANNEL_IND_CLASS) {
 			BEGIN_NV04(chan, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 2);
 			OUT_RING  (chan, NvEvoSema0 + nv_crtc->index);
 			OUT_RING  (chan, sync->sem.offset);
@@ -512,24 +512,36 @@ nv50_display_flip_next(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 			OUT_RING  (chan, sync->sem.offset ^ 0x10);
 			OUT_RING  (chan, 0x74b1e000);
 			BEGIN_NV04(chan, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 1);
-			if (nv_mclass(chan->object) < NV84_CHANNEL_DMA_CLASS)
-				OUT_RING  (chan, NvSema);
-			else
-				OUT_RING  (chan, chan->vram);
+			OUT_RING  (chan, NvSema);
+		} else
+		if (nv_mclass(chan->object) < NVC0_CHANNEL_IND_CLASS) {
+			u64 offset = nv84_fence_crtc(chan, nv_crtc->index);
+			offset += sync->sem.offset;
+
+			BEGIN_NV04(chan, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4);
+			OUT_RING  (chan, upper_32_bits(offset));
+			OUT_RING  (chan, lower_32_bits(offset));
+			OUT_RING  (chan, 0xf00d0000 | sync->sem.value);
+			OUT_RING  (chan, 0x00000002);
+			BEGIN_NV04(chan, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4);
+			OUT_RING  (chan, upper_32_bits(offset));
+			OUT_RING  (chan, lower_32_bits(offset ^ 0x10));
+			OUT_RING  (chan, 0x74b1e000);
+			OUT_RING  (chan, 0x00000001);
 		} else {
-			u64 offset = nvc0_fence_crtc(chan, nv_crtc->index);
+			u64 offset = nv84_fence_crtc(chan, nv_crtc->index);
 			offset += sync->sem.offset;
 
 			BEGIN_NVC0(chan, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4);
 			OUT_RING  (chan, upper_32_bits(offset));
 			OUT_RING  (chan, lower_32_bits(offset));
 			OUT_RING  (chan, 0xf00d0000 | sync->sem.value);
-			OUT_RING  (chan, 0x1002);
+			OUT_RING  (chan, 0x00001002);
 			BEGIN_NVC0(chan, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4);
 			OUT_RING  (chan, upper_32_bits(offset));
 			OUT_RING  (chan, lower_32_bits(offset ^ 0x10));
 			OUT_RING  (chan, 0x74b1e000);
-			OUT_RING  (chan, 0x1001);
+			OUT_RING  (chan, 0x00001001);
 		}
 
 		FIRE_RING (chan);
diff --git a/drivers/gpu/drm/nouveau/nv84_fence.c b/drivers/gpu/drm/nouveau/nv84_fence.c
index e64e815..58c2401 100644
--- a/drivers/gpu/drm/nouveau/nv84_fence.c
+++ b/drivers/gpu/drm/nouveau/nv84_fence.c
@@ -23,6 +23,7 @@
  */
 
 #include <core/object.h>
+#include <core/client.h>
 #include <core/class.h>
 
 #include <engine/fifo.h>
@@ -33,80 +34,96 @@
 
 #include "nv50_display.h"
 
-struct nv84_fence_chan {
-	struct nouveau_fence_chan base;
-};
-
-struct nv84_fence_priv {
-	struct nouveau_fence_priv base;
-	struct nouveau_gpuobj *mem;
-};
+u64
+nv84_fence_crtc(struct nouveau_channel *chan, int crtc)
+{
+	struct nv84_fence_chan *fctx = chan->fence;
+	return fctx->dispc_vma[crtc].offset;
+}
 
 static int
 nv84_fence_emit(struct nouveau_fence *fence)
 {
 	struct nouveau_channel *chan = fence->channel;
+	struct nv84_fence_chan *fctx = chan->fence;
 	struct nouveau_fifo_chan *fifo = (void *)chan->object;
-	int ret = RING_SPACE(chan, 8);
+	u64 addr = fctx->vma.offset + fifo->chid * 16;
+	int ret;
+
+	ret = RING_SPACE(chan, 8);
 	if (ret == 0) {
 		BEGIN_NV04(chan, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 1);
-		OUT_RING  (chan, NvSema);
+		OUT_RING  (chan, chan->vram);
 		BEGIN_NV04(chan, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 5);
-		OUT_RING  (chan, upper_32_bits(fifo->chid * 16));
-		OUT_RING  (chan, lower_32_bits(fifo->chid * 16));
+		OUT_RING  (chan, upper_32_bits(addr));
+		OUT_RING  (chan, lower_32_bits(addr));
 		OUT_RING  (chan, fence->sequence);
 		OUT_RING  (chan, NV84_SUBCHAN_SEMAPHORE_TRIGGER_WRITE_LONG);
 		OUT_RING  (chan, 0x00000000);
 		FIRE_RING (chan);
 	}
+
 	return ret;
 }
 
-
 static int
 nv84_fence_sync(struct nouveau_fence *fence,
 		struct nouveau_channel *prev, struct nouveau_channel *chan)
 {
+	struct nv84_fence_chan *fctx = chan->fence;
 	struct nouveau_fifo_chan *fifo = (void *)prev->object;
-	int ret = RING_SPACE(chan, 7);
+	u64 addr = fctx->vma.offset + fifo->chid * 16;
+	int ret;
+
+	ret = RING_SPACE(chan, 7);
 	if (ret == 0) {
 		BEGIN_NV04(chan, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 1);
-		OUT_RING  (chan, NvSema);
+		OUT_RING  (chan, chan->vram);
 		BEGIN_NV04(chan, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4);
-		OUT_RING  (chan, upper_32_bits(fifo->chid * 16));
-		OUT_RING  (chan, lower_32_bits(fifo->chid * 16));
+		OUT_RING  (chan, upper_32_bits(addr));
+		OUT_RING  (chan, lower_32_bits(addr));
 		OUT_RING  (chan, fence->sequence);
 		OUT_RING  (chan, NV84_SUBCHAN_SEMAPHORE_TRIGGER_ACQUIRE_GEQUAL);
 		FIRE_RING (chan);
 	}
+
 	return ret;
 }
 
-static u32
+u32
 nv84_fence_read(struct nouveau_channel *chan)
 {
 	struct nouveau_fifo_chan *fifo = (void *)chan->object;
 	struct nv84_fence_priv *priv = chan->drm->fence;
-	return nv_ro32(priv->mem, fifo->chid * 16);
+	return nouveau_bo_rd32(priv->bo, fifo->chid * 16/4);
 }
 
-static void
+void
 nv84_fence_context_del(struct nouveau_channel *chan)
 {
+	struct drm_device *dev = chan->drm->dev;
+	struct nv84_fence_priv *priv = chan->drm->fence;
 	struct nv84_fence_chan *fctx = chan->fence;
+	int i;
+
+	for (i = 0; i < dev->mode_config.num_crtc; i++) {
+		struct nouveau_bo *bo = nv50_display_crtc_sema(dev, i);
+		nouveau_bo_vma_del(bo, &fctx->dispc_vma[i]);
+	}
+
+	nouveau_bo_vma_del(priv->bo, &fctx->vma);
 	nouveau_fence_context_del(&fctx->base);
 	chan->fence = NULL;
 	kfree(fctx);
 }
 
-static int
+int
 nv84_fence_context_new(struct nouveau_channel *chan)
 {
-	struct drm_device *dev = chan->drm->dev;
 	struct nouveau_fifo_chan *fifo = (void *)chan->object;
+	struct nouveau_client *client = nouveau_client(fifo);
 	struct nv84_fence_priv *priv = chan->drm->fence;
 	struct nv84_fence_chan *fctx;
-	struct nouveau_object *object;
 	int ret, i;
 
 	fctx = chan->fence = kzalloc(sizeof(*fctx), GFP_KERNEL);
@@ -115,43 +132,59 @@ nv84_fence_context_new(struct nouveau_channel *chan)
 
 	nouveau_fence_context_new(&fctx->base);
 
-	ret = nouveau_object_new(nv_object(chan->cli), chan->handle,
-				 NvSema, 0x0002,
-				 &(struct nv_dma_class) {
-					.flags = NV_DMA_TARGET_VRAM |
-						 NV_DMA_ACCESS_RDWR,
-					.start = priv->mem->addr,
-					.limit = priv->mem->addr +
-						 priv->mem->size - 1,
-				 }, sizeof(struct nv_dma_class),
-				 &object);
-
-	/* dma objects for display sync channel semaphore blocks */
-	for (i = 0; !ret && i < dev->mode_config.num_crtc; i++) {
-		struct nouveau_bo *bo = nv50_display_crtc_sema(dev, i);
+	ret = nouveau_bo_vma_add(priv->bo, client->vm, &fctx->vma);
+	if (ret)
+		nv84_fence_context_del(chan);
 
-		ret = nouveau_object_new(nv_object(chan->cli), chan->handle,
-					 NvEvoSema0 + i, 0x003d,
-					 &(struct nv_dma_class) {
-						.flags = NV_DMA_TARGET_VRAM |
-							 NV_DMA_ACCESS_RDWR,
-						.start = bo->bo.offset,
-						.limit = bo->bo.offset + 0xfff,
-					 }, sizeof(struct nv_dma_class),
-					 &object);
+	/* map display semaphore buffers into channel's vm */
+	for (i = 0; !ret && i < chan->drm->dev->mode_config.num_crtc; i++) {
+		struct nouveau_bo *bo = nv50_display_crtc_sema(chan->drm->dev, i);
+		ret = nouveau_bo_vma_add(bo, client->vm, &fctx->dispc_vma[i]);
 	}
 
-	if (ret)
-		nv84_fence_context_del(chan);
-	nv_wo32(priv->mem, fifo->chid * 16, 0x00000000);
+	nouveau_bo_wr32(priv->bo, fifo->chid * 16/4, 0x00000000);
 	return ret;
 }
 
-static void
+bool
+nv84_fence_suspend(struct nouveau_drm *drm)
+{
+	struct nouveau_fifo *pfifo = nouveau_fifo(drm->device);
+	struct nv84_fence_priv *priv = drm->fence;
+	int i;
+
+	priv->suspend = vmalloc((pfifo->max + 1) * sizeof(u32));
+	if (priv->suspend) {
+		for (i = 0; i <= pfifo->max; i++)
+			priv->suspend[i] = nouveau_bo_rd32(priv->bo, i*4);
+	}
+
+	return priv->suspend != NULL;
+}
+
+void
+nv84_fence_resume(struct nouveau_drm *drm)
+{
+	struct nouveau_fifo *pfifo = nouveau_fifo(drm->device);
+	struct nv84_fence_priv *priv = drm->fence;
+	int i;
+
+	if (priv->suspend) {
+		for (i = 0; i <= pfifo->max; i++)
+			nouveau_bo_wr32(priv->bo, i*4, priv->suspend[i]);
+		vfree(priv->suspend);
+		priv->suspend = NULL;
+	}
+}
+
+void
 nv84_fence_destroy(struct nouveau_drm *drm)
 {
 	struct nv84_fence_priv *priv = drm->fence;
-	nouveau_gpuobj_ref(NULL, &priv->mem);
+	nouveau_bo_unmap(priv->bo);
+	if (priv->bo)
+		nouveau_bo_unpin(priv->bo);
+	nouveau_bo_ref(NULL, &priv->bo);
 	drm->fence = NULL;
 	kfree(priv);
 }
@@ -161,7 +194,6 @@ nv84_fence_create(struct nouveau_drm *drm)
 {
 	struct nouveau_fifo *pfifo = nouveau_fifo(drm->device);
 	struct nv84_fence_priv *priv;
-	u32 chan = pfifo->max + 1;
 	int ret;
 
 	priv = drm->fence = kzalloc(sizeof(*priv), GFP_KERNEL);
@@ -169,6 +201,8 @@ nv84_fence_create(struct nouveau_drm *drm)
 		return -ENOMEM;
 
 	priv->base.dtor = nv84_fence_destroy;
+	priv->base.suspend = nv84_fence_suspend;
+	priv->base.resume = nv84_fence_resume;
 	priv->base.context_new = nv84_fence_context_new;
 	priv->base.context_del = nv84_fence_context_del;
 	priv->base.emit = nv84_fence_emit;
@@ -178,8 +212,19 @@ nv84_fence_create(struct nouveau_drm *drm)
 	init_waitqueue_head(&priv->base.waiting);
 	priv->base.uevent = true;
 
-	ret = nouveau_gpuobj_new(drm->device, NULL, chan * 16, 0x1000, 0,
-				&priv->mem);
+	ret = nouveau_bo_new(drm->dev, 16 * (pfifo->max + 1), 0,
+			     TTM_PL_FLAG_VRAM, 0, 0, NULL, &priv->bo);
+	if (ret == 0) {
+		ret = nouveau_bo_pin(priv->bo, TTM_PL_FLAG_VRAM);
+		if (ret == 0) {
+			ret = nouveau_bo_map(priv->bo);
+			if (ret)
+				nouveau_bo_unpin(priv->bo);
+		}
+		if (ret)
+			nouveau_bo_ref(NULL, &priv->bo);
+	}
+
 	if (ret)
 		nv84_fence_destroy(drm);
 	return ret;
diff --git a/drivers/gpu/drm/nouveau/nvc0_fence.c b/drivers/gpu/drm/nouveau/nvc0_fence.c
index d8ed2c5..e4c4ead 100644
--- a/drivers/gpu/drm/nouveau/nvc0_fence.c
+++ b/drivers/gpu/drm/nouveau/nvc0_fence.c
@@ -34,30 +34,11 @@
 
 #include "nv50_display.h"
 
-struct nvc0_fence_priv {
-	struct nouveau_fence_priv base;
-	struct nouveau_bo *bo;
-	u32 *suspend;
-};
-
-struct nvc0_fence_chan {
-	struct nouveau_fence_chan base;
-	struct nouveau_vma vma;
-	struct nouveau_vma dispc_vma[4];
-};
-
-u64
-nvc0_fence_crtc(struct nouveau_channel *chan, int crtc)
-{
-	struct nvc0_fence_chan *fctx = chan->fence;
-	return fctx->dispc_vma[crtc].offset;
-}
-
 static int
 nvc0_fence_emit(struct nouveau_fence *fence)
 {
 	struct nouveau_channel *chan = fence->channel;
-	struct nvc0_fence_chan *fctx = chan->fence;
+	struct nv84_fence_chan *fctx = chan->fence;
 	struct nouveau_fifo_chan *fifo = (void *)chan->object;
 	u64 addr = fctx->vma.offset + fifo->chid * 16;
 	int ret;
@@ -80,7 +61,7 @@ static int
 nvc0_fence_sync(struct nouveau_fence *fence,
 		struct nouveau_channel *prev, struct nouveau_channel *chan)
 {
-	struct nvc0_fence_chan *fctx = chan->fence;
+	struct nv84_fence_chan *fctx = chan->fence;
 	struct nouveau_fifo_chan *fifo = (void *)prev->object;
 	u64 addr = fctx->vma.offset + fifo->chid * 16;
 	int ret;
@@ -99,124 +80,25 @@ nvc0_fence_sync(struct nouveau_fence *fence,
 	return ret;
 }
 
-static u32
-nvc0_fence_read(struct nouveau_channel *chan)
-{
-	struct nouveau_fifo_chan *fifo = (void *)chan->object;
-	struct nvc0_fence_priv *priv = chan->drm->fence;
-	return nouveau_bo_rd32(priv->bo, fifo->chid * 16/4);
-}
-
-static void
-nvc0_fence_context_del(struct nouveau_channel *chan)
-{
-	struct drm_device *dev = chan->drm->dev;
-	struct nvc0_fence_priv *priv = chan->drm->fence;
-	struct nvc0_fence_chan *fctx = chan->fence;
-	int i;
-
-	for (i = 0; i < dev->mode_config.num_crtc; i++) {
-		struct nouveau_bo *bo = nv50_display_crtc_sema(dev, i);
-		nouveau_bo_vma_del(bo, &fctx->dispc_vma[i]);
-	}
-
-	nouveau_bo_vma_del(priv->bo, &fctx->vma);
-	nouveau_fence_context_del(&fctx->base);
-	chan->fence = NULL;
-	kfree(fctx);
-}
-
-static int
-nvc0_fence_context_new(struct nouveau_channel *chan)
-{
-	struct nouveau_fifo_chan *fifo = (void *)chan->object;
-	struct nouveau_client *client = nouveau_client(fifo);
-	struct nvc0_fence_priv *priv = chan->drm->fence;
-	struct nvc0_fence_chan *fctx;
-	int ret, i;
-
-	fctx = chan->fence = kzalloc(sizeof(*fctx), GFP_KERNEL);
-	if (!fctx)
-		return -ENOMEM;
-
-	nouveau_fence_context_new(&fctx->base);
-
-	ret = nouveau_bo_vma_add(priv->bo, client->vm, &fctx->vma);
-	if (ret)
-		nvc0_fence_context_del(chan);
-
-	/* map display semaphore buffers into channel's vm */
-	for (i = 0; !ret && i < chan->drm->dev->mode_config.num_crtc; i++) {
-		struct nouveau_bo *bo = nv50_display_crtc_sema(chan->drm->dev, i);
-		ret = nouveau_bo_vma_add(bo, client->vm, &fctx->dispc_vma[i]);
-	}
-
-	nouveau_bo_wr32(priv->bo, fifo->chid * 16/4, 0x00000000);
-	return ret;
-}
-
-static bool
-nvc0_fence_suspend(struct nouveau_drm *drm)
-{
-	struct nouveau_fifo *pfifo = nouveau_fifo(drm->device);
-	struct nvc0_fence_priv *priv = drm->fence;
-	int i;
-
-	priv->suspend = vmalloc((pfifo->max + 1) * sizeof(u32));
-	if (priv->suspend) {
-		for (i = 0; i <= pfifo->max; i++)
-			priv->suspend[i] = nouveau_bo_rd32(priv->bo, i);
-	}
-
-	return priv->suspend != NULL;
-}
-
-static void
-nvc0_fence_resume(struct nouveau_drm *drm)
-{
-	struct nouveau_fifo *pfifo = nouveau_fifo(drm->device);
-	struct nvc0_fence_priv *priv = drm->fence;
-	int i;
-
-	if (priv->suspend) {
-		for (i = 0; i <= pfifo->max; i++)
-			nouveau_bo_wr32(priv->bo, i, priv->suspend[i]);
-		vfree(priv->suspend);
-		priv->suspend = NULL;
-	}
-}
-
-static void
-nvc0_fence_destroy(struct nouveau_drm *drm)
-{
-	struct nvc0_fence_priv *priv = drm->fence;
-	nouveau_bo_unmap(priv->bo);
-	if (priv->bo)
-		nouveau_bo_unpin(priv->bo);
-	nouveau_bo_ref(NULL, &priv->bo);
-	drm->fence = NULL;
-	kfree(priv);
-}
-
 int
 nvc0_fence_create(struct nouveau_drm *drm)
 {
 	struct nouveau_fifo *pfifo = nouveau_fifo(drm->device);
-	struct nvc0_fence_priv *priv;
+	struct nv84_fence_priv *priv;
 	int ret;
 
 	priv = drm->fence = kzalloc(sizeof(*priv), GFP_KERNEL);
 	if (!priv)
 		return -ENOMEM;
 
-	priv->base.dtor = nvc0_fence_destroy;
-	priv->base.suspend = nvc0_fence_suspend;
-	priv->base.resume = nvc0_fence_resume;
-	priv->base.context_new = nvc0_fence_context_new;
-	priv->base.context_del = nvc0_fence_context_del;
+	priv->base.dtor = nv84_fence_destroy;
+	priv->base.suspend = nv84_fence_suspend;
+	priv->base.resume = nv84_fence_resume;
+	priv->base.context_new = nv84_fence_context_new;
+	priv->base.context_del = nv84_fence_context_del;
 	priv->base.emit = nvc0_fence_emit;
 	priv->base.sync = nvc0_fence_sync;
-	priv->base.read = nvc0_fence_read;
+	priv->base.read = nv84_fence_read;
 
 	init_waitqueue_head(&priv->base.waiting);
 	priv->base.uevent = true;
@@ -235,6 +117,6 @@ nvc0_fence_create(struct nouveau_drm *drm)
 	}
 
 	if (ret)
-		nvc0_fence_destroy(drm);
+		nv84_fence_destroy(drm);
 	return ret;
 }
-- 
cgit v0.10.2


From bba9852feedf3d38f963278e07bdd3db622090b9 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Thu, 14 Feb 2013 09:37:35 +1000
Subject: drm/nv84-/fence: abstract class emit/sync functions to virt+sequence

Now can be used to operate on any buffer mapped into the GPU virtual
address and not just the main inter-channel sync buffer.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.h b/drivers/gpu/drm/nouveau/nouveau_fence.h
index 2324911..f1f0c6d 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fence.h
+++ b/drivers/gpu/drm/nouveau/nouveau_fence.h
@@ -39,7 +39,9 @@ struct nouveau_fence_priv {
 	void (*resume)(struct nouveau_drm *);
 	int  (*context_new)(struct nouveau_channel *);
 	void (*context_del)(struct nouveau_channel *);
+	int  (*emit32)(struct nouveau_channel *, u64, u32);
 	int  (*emit)(struct nouveau_fence *);
+	int  (*sync32)(struct nouveau_channel *, u64, u32);
 	int  (*sync)(struct nouveau_fence *, struct nouveau_channel *,
 		     struct nouveau_channel *);
 	u32  (*read)(struct nouveau_channel *);
@@ -84,6 +86,9 @@ struct nv84_fence_priv {
 };
 
 u64  nv84_fence_crtc(struct nouveau_channel *, int);
+int  nv84_fence_emit(struct nouveau_fence *);
+int  nv84_fence_sync(struct nouveau_fence *, struct nouveau_channel *,
+		     struct nouveau_channel *);
 u32  nv84_fence_read(struct nouveau_channel *);
 int  nv84_fence_context_new(struct nouveau_channel *);
 void nv84_fence_context_del(struct nouveau_channel *);
diff --git a/drivers/gpu/drm/nouveau/nv84_fence.c b/drivers/gpu/drm/nouveau/nv84_fence.c
index 58c2401..8a80ad7 100644
--- a/drivers/gpu/drm/nouveau/nv84_fence.c
+++ b/drivers/gpu/drm/nouveau/nv84_fence.c
@@ -42,54 +42,62 @@ nv84_fence_crtc(struct nouveau_channel *chan, int crtc)
 }
 
 static int
-nv84_fence_emit(struct nouveau_fence *fence)
+nv84_fence_emit32(struct nouveau_channel *chan, u64 virtual, u32 sequence)
 {
-	struct nouveau_channel *chan = fence->channel;
-	struct nv84_fence_chan *fctx = chan->fence;
-	struct nouveau_fifo_chan *fifo = (void *)chan->object;
-	u64 addr = fctx->vma.offset + fifo->chid * 16;
-	int ret;
-
-	ret = RING_SPACE(chan, 8);
+	int ret = RING_SPACE(chan, 8);
 	if (ret == 0) {
 		BEGIN_NV04(chan, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 1);
 		OUT_RING  (chan, chan->vram);
 		BEGIN_NV04(chan, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 5);
-		OUT_RING  (chan, upper_32_bits(addr));
-		OUT_RING  (chan, lower_32_bits(addr));
-		OUT_RING  (chan, fence->sequence);
+		OUT_RING  (chan, upper_32_bits(virtual));
+		OUT_RING  (chan, lower_32_bits(virtual));
+		OUT_RING  (chan, sequence);
 		OUT_RING  (chan, NV84_SUBCHAN_SEMAPHORE_TRIGGER_WRITE_LONG);
 		OUT_RING  (chan, 0x00000000);
 		FIRE_RING (chan);
 	}
-
 	return ret;
 }
 
 static int
-nv84_fence_sync(struct nouveau_fence *fence,
-		struct nouveau_channel *prev, struct nouveau_channel *chan)
+nv84_fence_sync32(struct nouveau_channel *chan, u64 virtual, u32 sequence)
 {
-	struct nv84_fence_chan *fctx = chan->fence;
-	struct nouveau_fifo_chan *fifo = (void *)prev->object;
-	u64 addr = fctx->vma.offset + fifo->chid * 16;
-	int ret;
-
-	ret = RING_SPACE(chan, 7);
+	int ret = RING_SPACE(chan, 7);
 	if (ret == 0) {
 		BEGIN_NV04(chan, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 1);
 		OUT_RING  (chan, chan->vram);
 		BEGIN_NV04(chan, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4);
-		OUT_RING  (chan, upper_32_bits(addr));
-		OUT_RING  (chan, lower_32_bits(addr));
-		OUT_RING  (chan, fence->sequence);
+		OUT_RING  (chan, upper_32_bits(virtual));
+		OUT_RING  (chan, lower_32_bits(virtual));
+		OUT_RING  (chan, sequence);
 		OUT_RING  (chan, NV84_SUBCHAN_SEMAPHORE_TRIGGER_ACQUIRE_GEQUAL);
 		FIRE_RING (chan);
 	}
-
 	return ret;
 }
 
+int
+nv84_fence_emit(struct nouveau_fence *fence)
+{
+	struct nouveau_channel *chan = fence->channel;
+	struct nv84_fence_priv *priv = chan->drm->fence;
+	struct nv84_fence_chan *fctx = chan->fence;
+	struct nouveau_fifo_chan *fifo = (void *)chan->object;
+	u64 addr = fctx->vma.offset + fifo->chid * 16;
+	return priv->base.emit32(chan, addr, fence->sequence);
+}
+
+int
+nv84_fence_sync(struct nouveau_fence *fence,
+		struct nouveau_channel *prev, struct nouveau_channel *chan)
+{
+	struct nv84_fence_priv *priv = chan->drm->fence;
+	struct nv84_fence_chan *fctx = chan->fence;
+	struct nouveau_fifo_chan *fifo = (void *)prev->object;
+	u64 addr = fctx->vma.offset + fifo->chid * 16;
+	return priv->base.sync32(chan, addr, fence->sequence);
+}
+
 u32
 nv84_fence_read(struct nouveau_channel *chan)
 {
@@ -205,7 +213,9 @@ nv84_fence_create(struct nouveau_drm *drm)
 	priv->base.resume = nv84_fence_resume;
 	priv->base.context_new = nv84_fence_context_new;
 	priv->base.context_del = nv84_fence_context_del;
+	priv->base.emit32 = nv84_fence_emit32;
 	priv->base.emit = nv84_fence_emit;
+	priv->base.sync32 = nv84_fence_sync32;
 	priv->base.sync = nv84_fence_sync;
 	priv->base.read = nv84_fence_read;
 
diff --git a/drivers/gpu/drm/nouveau/nvc0_fence.c b/drivers/gpu/drm/nouveau/nvc0_fence.c
index e4c4ead..8213f7d 100644
--- a/drivers/gpu/drm/nouveau/nvc0_fence.c
+++ b/drivers/gpu/drm/nouveau/nvc0_fence.c
@@ -35,48 +35,34 @@
 #include "nv50_display.h"
 
 static int
-nvc0_fence_emit(struct nouveau_fence *fence)
+nvc0_fence_emit32(struct nouveau_channel *chan, u64 virtual, u32 sequence)
 {
-	struct nouveau_channel *chan = fence->channel;
-	struct nv84_fence_chan *fctx = chan->fence;
-	struct nouveau_fifo_chan *fifo = (void *)chan->object;
-	u64 addr = fctx->vma.offset + fifo->chid * 16;
-	int ret;
-
-	ret = RING_SPACE(chan, 6);
+	int ret = RING_SPACE(chan, 6);
 	if (ret == 0) {
 		BEGIN_NVC0(chan, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 5);
-		OUT_RING  (chan, upper_32_bits(addr));
-		OUT_RING  (chan, lower_32_bits(addr));
-		OUT_RING  (chan, fence->sequence);
+		OUT_RING  (chan, upper_32_bits(virtual));
+		OUT_RING  (chan, lower_32_bits(virtual));
+		OUT_RING  (chan, sequence);
 		OUT_RING  (chan, NV84_SUBCHAN_SEMAPHORE_TRIGGER_WRITE_LONG);
 		OUT_RING  (chan, 0x00000000);
 		FIRE_RING (chan);
 	}
-
 	return ret;
 }
 
 static int
-nvc0_fence_sync(struct nouveau_fence *fence,
-		struct nouveau_channel *prev, struct nouveau_channel *chan)
+nvc0_fence_sync32(struct nouveau_channel *chan, u64 virtual, u32 sequence)
 {
-	struct nv84_fence_chan *fctx = chan->fence;
-	struct nouveau_fifo_chan *fifo = (void *)prev->object;
-	u64 addr = fctx->vma.offset + fifo->chid * 16;
-	int ret;
-
-	ret = RING_SPACE(chan, 5);
+	int ret = RING_SPACE(chan, 5);
 	if (ret == 0) {
 		BEGIN_NVC0(chan, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4);
-		OUT_RING  (chan, upper_32_bits(addr));
-		OUT_RING  (chan, lower_32_bits(addr));
-		OUT_RING  (chan, fence->sequence);
+		OUT_RING  (chan, upper_32_bits(virtual));
+		OUT_RING  (chan, lower_32_bits(virtual));
+		OUT_RING  (chan, sequence);
 		OUT_RING  (chan, NV84_SUBCHAN_SEMAPHORE_TRIGGER_ACQUIRE_GEQUAL |
 				 NVC0_SUBCHAN_SEMAPHORE_TRIGGER_YIELD);
 		FIRE_RING (chan);
 	}
-
 	return ret;
 }
 
@@ -96,8 +82,10 @@ nvc0_fence_create(struct nouveau_drm *drm)
 	priv->base.resume = nv84_fence_resume;
 	priv->base.context_new = nv84_fence_context_new;
 	priv->base.context_del = nv84_fence_context_del;
-	priv->base.emit = nvc0_fence_emit;
-	priv->base.sync = nvc0_fence_sync;
+	priv->base.emit32 = nvc0_fence_emit32;
+	priv->base.emit = nv84_fence_emit;
+	priv->base.sync32 = nvc0_fence_sync32;
+	priv->base.sync = nv84_fence_sync;
 	priv->base.read = nv84_fence_read;
 
 	init_waitqueue_head(&priv->base.waiting);
-- 
cgit v0.10.2


From 60e5cb79cbd27a36836fc04177d7c323ee873563 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Thu, 14 Feb 2013 12:59:36 +1000
Subject: drm/nv17/fence: split from nv10 code

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index e0bd21a..4208db9 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -207,7 +207,8 @@ nouveau-y += nouveau_drm.o nouveau_chan.o nouveau_dma.o nouveau_fence.o
 nouveau-y += nouveau_irq.o nouveau_vga.o nouveau_agp.o
 nouveau-y += nouveau_ttm.o nouveau_sgdma.o nouveau_bo.o nouveau_gem.o
 nouveau-y += nouveau_prime.o nouveau_abi16.o
-nouveau-y += nv04_fence.o nv10_fence.o nv50_fence.o nv84_fence.o nvc0_fence.o
+nouveau-y += nv04_fence.o nv10_fence.o nv17_fence.o
+nouveau-y += nv50_fence.o nv84_fence.o nvc0_fence.o
 
 # drm/kms
 nouveau-y += nouveau_bios.o nouveau_fbcon.o nouveau_display.o
diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c
index ce91c8d..f914491 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_drm.c
@@ -160,7 +160,8 @@ nouveau_accel_init(struct nouveau_drm *drm)
 
 	/* initialise synchronisation routines */
 	if      (device->card_type < NV_10) ret = nv04_fence_create(drm);
-	else if (device->card_type < NV_50) ret = nv10_fence_create(drm);
+	else if (device->chipset   <  0x17) ret = nv10_fence_create(drm);
+	else if (device->card_type < NV_50) ret = nv17_fence_create(drm);
 	else if (device->chipset   <  0x84) ret = nv50_fence_create(drm);
 	else if (device->card_type < NV_C0) ret = nv84_fence_create(drm);
 	else                                ret = nvc0_fence_create(drm);
diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.h b/drivers/gpu/drm/nouveau/nouveau_fence.h
index f1f0c6d..fb0993c 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fence.h
+++ b/drivers/gpu/drm/nouveau/nouveau_fence.h
@@ -65,6 +65,8 @@ u32  nv10_fence_read(struct nouveau_channel *);
 void nv10_fence_context_del(struct nouveau_channel *);
 void nv10_fence_destroy(struct nouveau_drm *);
 int  nv10_fence_create(struct nouveau_drm *);
+
+int  nv17_fence_create(struct nouveau_drm *);
 void nv17_fence_resume(struct nouveau_drm *drm);
 
 int nv50_fence_create(struct nouveau_drm *);
diff --git a/drivers/gpu/drm/nouveau/nv10_fence.c b/drivers/gpu/drm/nouveau/nv10_fence.c
index a8ea4af..e4f124a 100644
--- a/drivers/gpu/drm/nouveau/nv10_fence.c
+++ b/drivers/gpu/drm/nouveau/nv10_fence.c
@@ -50,45 +50,6 @@ nv10_fence_sync(struct nouveau_fence *fence,
 	return -ENODEV;
 }
 
-int
-nv17_fence_sync(struct nouveau_fence *fence,
-		struct nouveau_channel *prev, struct nouveau_channel *chan)
-{
-	struct nv10_fence_priv *priv = chan->drm->fence;
-	u32 value;
-	int ret;
-
-	if (!mutex_trylock(&prev->cli->mutex))
-		return -EBUSY;
-
-	spin_lock(&priv->lock);
-	value = priv->sequence;
-	priv->sequence += 2;
-	spin_unlock(&priv->lock);
-
-	ret = RING_SPACE(prev, 5);
-	if (!ret) {
-		BEGIN_NV04(prev, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 4);
-		OUT_RING  (prev, NvSema);
-		OUT_RING  (prev, 0);
-		OUT_RING  (prev, value + 0);
-		OUT_RING  (prev, value + 1);
-		FIRE_RING (prev);
-	}
-
-	if (!ret && !(ret = RING_SPACE(chan, 5))) {
-		BEGIN_NV04(chan, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 4);
-		OUT_RING  (chan, NvSema);
-		OUT_RING  (chan, 0);
-		OUT_RING  (chan, value + 1);
-		OUT_RING  (chan, value + 2);
-		FIRE_RING (chan);
-	}
-
-	mutex_unlock(&prev->cli->mutex);
-	return 0;
-}
-
 u32
 nv10_fence_read(struct nouveau_channel *chan)
 {
@@ -104,39 +65,17 @@ nv10_fence_context_del(struct nouveau_channel *chan)
 	kfree(fctx);
 }
 
-static int
+int
 nv10_fence_context_new(struct nouveau_channel *chan)
 {
-	struct nv10_fence_priv *priv = chan->drm->fence;
 	struct nv10_fence_chan *fctx;
-	int ret = 0;
 
 	fctx = chan->fence = kzalloc(sizeof(*fctx), GFP_KERNEL);
 	if (!fctx)
 		return -ENOMEM;
 
 	nouveau_fence_context_new(&fctx->base);
-
-	if (priv->bo) {
-		struct ttm_mem_reg *mem = &priv->bo->bo.mem;
-		struct nouveau_object *object;
-		u32 start = mem->start * PAGE_SIZE;
-		u32 limit = mem->start + mem->size - 1;
-
-		ret = nouveau_object_new(nv_object(chan->cli), chan->handle,
-					 NvSema, 0x0002,
-					 &(struct nv_dma_class) {
-						.flags = NV_DMA_TARGET_VRAM |
-							 NV_DMA_ACCESS_RDWR,
-						.start = start,
-						.limit = limit,
-					 }, sizeof(struct nv_dma_class),
-					 &object);
-	}
-
-	if (ret)
-		nv10_fence_context_del(chan);
-	return ret;
+	return 0;
 }
 
 void
@@ -151,18 +90,10 @@ nv10_fence_destroy(struct nouveau_drm *drm)
 	kfree(priv);
 }
 
-void nv17_fence_resume(struct nouveau_drm *drm)
-{
-	struct nv10_fence_priv *priv = drm->fence;
-
-	nouveau_bo_wr32(priv->bo, 0, priv->sequence);
-}
-
 int
 nv10_fence_create(struct nouveau_drm *drm)
 {
 	struct nv10_fence_priv *priv;
-	int ret = 0;
 
 	priv = drm->fence = kzalloc(sizeof(*priv), GFP_KERNEL);
 	if (!priv)
@@ -175,29 +106,5 @@ nv10_fence_create(struct nouveau_drm *drm)
 	priv->base.read = nv10_fence_read;
 	priv->base.sync = nv10_fence_sync;
 	spin_lock_init(&priv->lock);
-
-	if (nv_device(drm->device)->chipset >= 0x17) {
-		ret = nouveau_bo_new(drm->dev, 4096, 0x1000, TTM_PL_FLAG_VRAM,
-				     0, 0x0000, NULL, &priv->bo);
-		if (!ret) {
-			ret = nouveau_bo_pin(priv->bo, TTM_PL_FLAG_VRAM);
-			if (!ret) {
-				ret = nouveau_bo_map(priv->bo);
-				if (ret)
-					nouveau_bo_unpin(priv->bo);
-			}
-			if (ret)
-				nouveau_bo_ref(NULL, &priv->bo);
-		}
-
-		if (ret == 0) {
-			nouveau_bo_wr32(priv->bo, 0x000, 0x00000000);
-			priv->base.sync = nv17_fence_sync;
-			priv->base.resume = nv17_fence_resume;
-		}
-	}
-
-	if (ret)
-		nv10_fence_destroy(drm);
-	return ret;
+	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nv17_fence.c b/drivers/gpu/drm/nouveau/nv17_fence.c
new file mode 100644
index 0000000..fe19445
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nv17_fence.c
@@ -0,0 +1,149 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include <core/object.h>
+#include <core/class.h>
+
+#include "nouveau_drm.h"
+#include "nouveau_dma.h"
+#include "nv10_fence.h"
+
+int
+nv17_fence_sync(struct nouveau_fence *fence,
+		struct nouveau_channel *prev, struct nouveau_channel *chan)
+{
+	struct nv10_fence_priv *priv = chan->drm->fence;
+	u32 value;
+	int ret;
+
+	if (!mutex_trylock(&prev->cli->mutex))
+		return -EBUSY;
+
+	spin_lock(&priv->lock);
+	value = priv->sequence;
+	priv->sequence += 2;
+	spin_unlock(&priv->lock);
+
+	ret = RING_SPACE(prev, 5);
+	if (!ret) {
+		BEGIN_NV04(prev, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 4);
+		OUT_RING  (prev, NvSema);
+		OUT_RING  (prev, 0);
+		OUT_RING  (prev, value + 0);
+		OUT_RING  (prev, value + 1);
+		FIRE_RING (prev);
+	}
+
+	if (!ret && !(ret = RING_SPACE(chan, 5))) {
+		BEGIN_NV04(chan, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 4);
+		OUT_RING  (chan, NvSema);
+		OUT_RING  (chan, 0);
+		OUT_RING  (chan, value + 1);
+		OUT_RING  (chan, value + 2);
+		FIRE_RING (chan);
+	}
+
+	mutex_unlock(&prev->cli->mutex);
+	return 0;
+}
+
+static int
+nv17_fence_context_new(struct nouveau_channel *chan)
+{
+	struct nv10_fence_priv *priv = chan->drm->fence;
+	struct nv10_fence_chan *fctx;
+	struct ttm_mem_reg *mem = &priv->bo->bo.mem;
+	struct nouveau_object *object;
+	u32 start = mem->start * PAGE_SIZE;
+	u32 limit = mem->start + mem->size - 1;
+	int ret = 0;
+
+	fctx = chan->fence = kzalloc(sizeof(*fctx), GFP_KERNEL);
+	if (!fctx)
+		return -ENOMEM;
+
+	nouveau_fence_context_new(&fctx->base);
+
+	ret = nouveau_object_new(nv_object(chan->cli), chan->handle,
+				 NvSema, 0x0002,
+				 &(struct nv_dma_class) {
+					.flags = NV_DMA_TARGET_VRAM |
+						 NV_DMA_ACCESS_RDWR,
+					.start = start,
+					.limit = limit,
+				 }, sizeof(struct nv_dma_class),
+				 &object);
+	if (ret)
+		nv10_fence_context_del(chan);
+	return ret;
+}
+
+void
+nv17_fence_resume(struct nouveau_drm *drm)
+{
+	struct nv10_fence_priv *priv = drm->fence;
+
+	nouveau_bo_wr32(priv->bo, 0, priv->sequence);
+}
+
+int
+nv17_fence_create(struct nouveau_drm *drm)
+{
+	struct nv10_fence_priv *priv;
+	int ret = 0;
+
+	priv = drm->fence = kzalloc(sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	priv->base.dtor = nv10_fence_destroy;
+	priv->base.resume = nv17_fence_resume;
+	priv->base.context_new = nv17_fence_context_new;
+	priv->base.context_del = nv10_fence_context_del;
+	priv->base.emit = nv10_fence_emit;
+	priv->base.read = nv10_fence_read;
+	priv->base.sync = nv17_fence_sync;
+	spin_lock_init(&priv->lock);
+
+	ret = nouveau_bo_new(drm->dev, 4096, 0x1000, TTM_PL_FLAG_VRAM,
+			     0, 0x0000, NULL, &priv->bo);
+	if (!ret) {
+		ret = nouveau_bo_pin(priv->bo, TTM_PL_FLAG_VRAM);
+		if (!ret) {
+			ret = nouveau_bo_map(priv->bo);
+			if (ret)
+				nouveau_bo_unpin(priv->bo);
+		}
+		if (ret)
+			nouveau_bo_ref(NULL, &priv->bo);
+	}
+
+	if (ret) {
+		nv10_fence_destroy(drm);
+		return ret;
+	}
+
+	nouveau_bo_wr32(priv->bo, 0x000, 0x00000000);
+	return ret;
+}
-- 
cgit v0.10.2


From 827520ce06568f699dad275dcca61647cce08757 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Thu, 14 Feb 2013 13:20:17 +1000
Subject: drm/nouveau/fence: make internal hooks part of the context

A step towards being able to provide fences from other engines not
connected to PFIFO.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.c b/drivers/gpu/drm/nouveau/nouveau_fence.c
index b7349a6..6a7a5b5 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fence.c
+++ b/drivers/gpu/drm/nouveau/nouveau_fence.c
@@ -61,13 +61,12 @@ nouveau_fence_context_new(struct nouveau_fence_chan *fctx)
 static void
 nouveau_fence_update(struct nouveau_channel *chan)
 {
-	struct nouveau_fence_priv *priv = chan->drm->fence;
 	struct nouveau_fence_chan *fctx = chan->fence;
 	struct nouveau_fence *fence, *fnext;
 
 	spin_lock(&fctx->lock);
 	list_for_each_entry_safe(fence, fnext, &fctx->pending, head) {
-		if (priv->read(chan) < fence->sequence)
+		if (fctx->read(chan) < fence->sequence)
 			break;
 
 		if (fence->work)
@@ -82,7 +81,6 @@ nouveau_fence_update(struct nouveau_channel *chan)
 int
 nouveau_fence_emit(struct nouveau_fence *fence, struct nouveau_channel *chan)
 {
-	struct nouveau_fence_priv *priv = chan->drm->fence;
 	struct nouveau_fence_chan *fctx = chan->fence;
 	int ret;
 
@@ -90,7 +88,7 @@ nouveau_fence_emit(struct nouveau_fence *fence, struct nouveau_channel *chan)
 	fence->timeout  = jiffies + (3 * DRM_HZ);
 	fence->sequence = ++fctx->sequence;
 
-	ret = priv->emit(fence);
+	ret = fctx->emit(fence);
 	if (!ret) {
 		kref_get(&fence->kref);
 		spin_lock(&fctx->lock);
@@ -219,14 +217,14 @@ nouveau_fence_wait(struct nouveau_fence *fence, bool lazy, bool intr)
 int
 nouveau_fence_sync(struct nouveau_fence *fence, struct nouveau_channel *chan)
 {
-	struct nouveau_fence_priv *priv = chan->drm->fence;
+	struct nouveau_fence_chan *fctx = chan->fence;
 	struct nouveau_channel *prev;
 	int ret = 0;
 
 	prev = fence ? fence->channel : NULL;
 	if (prev) {
 		if (unlikely(prev != chan && !nouveau_fence_done(fence))) {
-			ret = priv->sync(fence, prev, chan);
+			ret = fctx->sync(fence, prev, chan);
 			if (unlikely(ret))
 				ret = nouveau_fence_wait(fence, true, false);
 		}
diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.h b/drivers/gpu/drm/nouveau/nouveau_fence.h
index fb0993c..a5c47e3 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fence.h
+++ b/drivers/gpu/drm/nouveau/nouveau_fence.h
@@ -29,6 +29,13 @@ struct nouveau_fence_chan {
 	struct list_head pending;
 	struct list_head flip;
 
+	int  (*emit)(struct nouveau_fence *);
+	int  (*sync)(struct nouveau_fence *, struct nouveau_channel *,
+		     struct nouveau_channel *);
+	u32  (*read)(struct nouveau_channel *);
+	int  (*emit32)(struct nouveau_channel *, u64, u32);
+	int  (*sync32)(struct nouveau_channel *, u64, u32);
+
 	spinlock_t lock;
 	u32 sequence;
 };
@@ -39,12 +46,6 @@ struct nouveau_fence_priv {
 	void (*resume)(struct nouveau_drm *);
 	int  (*context_new)(struct nouveau_channel *);
 	void (*context_del)(struct nouveau_channel *);
-	int  (*emit32)(struct nouveau_channel *, u64, u32);
-	int  (*emit)(struct nouveau_fence *);
-	int  (*sync32)(struct nouveau_channel *, u64, u32);
-	int  (*sync)(struct nouveau_fence *, struct nouveau_channel *,
-		     struct nouveau_channel *);
-	u32  (*read)(struct nouveau_channel *);
 
 	wait_queue_head_t waiting;
 	bool uevent;
diff --git a/drivers/gpu/drm/nouveau/nv04_fence.c b/drivers/gpu/drm/nouveau/nv04_fence.c
index a220b94..94eadd1 100644
--- a/drivers/gpu/drm/nouveau/nv04_fence.c
+++ b/drivers/gpu/drm/nouveau/nv04_fence.c
@@ -78,6 +78,9 @@ nv04_fence_context_new(struct nouveau_channel *chan)
 	struct nv04_fence_chan *fctx = kzalloc(sizeof(*fctx), GFP_KERNEL);
 	if (fctx) {
 		nouveau_fence_context_new(&fctx->base);
+		fctx->base.emit = nv04_fence_emit;
+		fctx->base.sync = nv04_fence_sync;
+		fctx->base.read = nv04_fence_read;
 		chan->fence = fctx;
 		return 0;
 	}
@@ -104,8 +107,5 @@ nv04_fence_create(struct nouveau_drm *drm)
 	priv->base.dtor = nv04_fence_destroy;
 	priv->base.context_new = nv04_fence_context_new;
 	priv->base.context_del = nv04_fence_context_del;
-	priv->base.emit = nv04_fence_emit;
-	priv->base.sync = nv04_fence_sync;
-	priv->base.read = nv04_fence_read;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nv10_fence.c b/drivers/gpu/drm/nouveau/nv10_fence.c
index e4f124a..06f434f 100644
--- a/drivers/gpu/drm/nouveau/nv10_fence.c
+++ b/drivers/gpu/drm/nouveau/nv10_fence.c
@@ -75,6 +75,9 @@ nv10_fence_context_new(struct nouveau_channel *chan)
 		return -ENOMEM;
 
 	nouveau_fence_context_new(&fctx->base);
+	fctx->base.emit = nv10_fence_emit;
+	fctx->base.read = nv10_fence_read;
+	fctx->base.sync = nv10_fence_sync;
 	return 0;
 }
 
@@ -102,9 +105,6 @@ nv10_fence_create(struct nouveau_drm *drm)
 	priv->base.dtor = nv10_fence_destroy;
 	priv->base.context_new = nv10_fence_context_new;
 	priv->base.context_del = nv10_fence_context_del;
-	priv->base.emit = nv10_fence_emit;
-	priv->base.read = nv10_fence_read;
-	priv->base.sync = nv10_fence_sync;
 	spin_lock_init(&priv->lock);
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nv17_fence.c b/drivers/gpu/drm/nouveau/nv17_fence.c
index fe19445..8e47a9b 100644
--- a/drivers/gpu/drm/nouveau/nv17_fence.c
+++ b/drivers/gpu/drm/nouveau/nv17_fence.c
@@ -84,6 +84,9 @@ nv17_fence_context_new(struct nouveau_channel *chan)
 		return -ENOMEM;
 
 	nouveau_fence_context_new(&fctx->base);
+	fctx->base.emit = nv10_fence_emit;
+	fctx->base.read = nv10_fence_read;
+	fctx->base.sync = nv17_fence_sync;
 
 	ret = nouveau_object_new(nv_object(chan->cli), chan->handle,
 				 NvSema, 0x0002,
@@ -121,9 +124,6 @@ nv17_fence_create(struct nouveau_drm *drm)
 	priv->base.resume = nv17_fence_resume;
 	priv->base.context_new = nv17_fence_context_new;
 	priv->base.context_del = nv10_fence_context_del;
-	priv->base.emit = nv10_fence_emit;
-	priv->base.read = nv10_fence_read;
-	priv->base.sync = nv17_fence_sync;
 	spin_lock_init(&priv->lock);
 
 	ret = nouveau_bo_new(drm->dev, 4096, 0x1000, TTM_PL_FLAG_VRAM,
diff --git a/drivers/gpu/drm/nouveau/nv50_fence.c b/drivers/gpu/drm/nouveau/nv50_fence.c
index 72791d6..f9701e5 100644
--- a/drivers/gpu/drm/nouveau/nv50_fence.c
+++ b/drivers/gpu/drm/nouveau/nv50_fence.c
@@ -46,6 +46,9 @@ nv50_fence_context_new(struct nouveau_channel *chan)
 		return -ENOMEM;
 
 	nouveau_fence_context_new(&fctx->base);
+	fctx->base.emit = nv10_fence_emit;
+	fctx->base.read = nv10_fence_read;
+	fctx->base.sync = nv17_fence_sync;
 
 	ret = nouveau_object_new(nv_object(chan->cli), chan->handle,
 				 NvSema, 0x0002,
@@ -88,11 +91,9 @@ nv50_fence_create(struct nouveau_drm *drm)
 		return -ENOMEM;
 
 	priv->base.dtor = nv10_fence_destroy;
+	priv->base.resume = nv17_fence_resume;
 	priv->base.context_new = nv50_fence_context_new;
 	priv->base.context_del = nv10_fence_context_del;
-	priv->base.emit = nv10_fence_emit;
-	priv->base.read = nv10_fence_read;
-	priv->base.sync = nv17_fence_sync;
 	spin_lock_init(&priv->lock);
 
 	ret = nouveau_bo_new(drm->dev, 4096, 0x1000, TTM_PL_FLAG_VRAM,
@@ -108,13 +109,11 @@ nv50_fence_create(struct nouveau_drm *drm)
 			nouveau_bo_ref(NULL, &priv->bo);
 	}
 
-	if (ret == 0) {
-		nouveau_bo_wr32(priv->bo, 0x000, 0x00000000);
-		priv->base.sync = nv17_fence_sync;
-		priv->base.resume = nv17_fence_resume;
+	if (ret) {
+		nv10_fence_destroy(drm);
+		return ret;
 	}
 
-	if (ret)
-		nv10_fence_destroy(drm);
+	nouveau_bo_wr32(priv->bo, 0x000, 0x00000000);
 	return ret;
 }
diff --git a/drivers/gpu/drm/nouveau/nv84_fence.c b/drivers/gpu/drm/nouveau/nv84_fence.c
index 8a80ad7..bc6493c 100644
--- a/drivers/gpu/drm/nouveau/nv84_fence.c
+++ b/drivers/gpu/drm/nouveau/nv84_fence.c
@@ -80,22 +80,20 @@ int
 nv84_fence_emit(struct nouveau_fence *fence)
 {
 	struct nouveau_channel *chan = fence->channel;
-	struct nv84_fence_priv *priv = chan->drm->fence;
 	struct nv84_fence_chan *fctx = chan->fence;
 	struct nouveau_fifo_chan *fifo = (void *)chan->object;
 	u64 addr = fctx->vma.offset + fifo->chid * 16;
-	return priv->base.emit32(chan, addr, fence->sequence);
+	return fctx->base.emit32(chan, addr, fence->sequence);
 }
 
 int
 nv84_fence_sync(struct nouveau_fence *fence,
 		struct nouveau_channel *prev, struct nouveau_channel *chan)
 {
-	struct nv84_fence_priv *priv = chan->drm->fence;
 	struct nv84_fence_chan *fctx = chan->fence;
 	struct nouveau_fifo_chan *fifo = (void *)prev->object;
 	u64 addr = fctx->vma.offset + fifo->chid * 16;
-	return priv->base.sync32(chan, addr, fence->sequence);
+	return fctx->base.sync32(chan, addr, fence->sequence);
 }
 
 u32
@@ -139,6 +137,11 @@ nv84_fence_context_new(struct nouveau_channel *chan)
 		return -ENOMEM;
 
 	nouveau_fence_context_new(&fctx->base);
+	fctx->base.emit = nv84_fence_emit;
+	fctx->base.sync = nv84_fence_sync;
+	fctx->base.read = nv84_fence_read;
+	fctx->base.emit32 = nv84_fence_emit32;
+	fctx->base.sync32 = nv84_fence_sync32;
 
 	ret = nouveau_bo_vma_add(priv->bo, client->vm, &fctx->vma);
 	if (ret)
@@ -213,11 +216,6 @@ nv84_fence_create(struct nouveau_drm *drm)
 	priv->base.resume = nv84_fence_resume;
 	priv->base.context_new = nv84_fence_context_new;
 	priv->base.context_del = nv84_fence_context_del;
-	priv->base.emit32 = nv84_fence_emit32;
-	priv->base.emit = nv84_fence_emit;
-	priv->base.sync32 = nv84_fence_sync32;
-	priv->base.sync = nv84_fence_sync;
-	priv->base.read = nv84_fence_read;
 
 	init_waitqueue_head(&priv->base.waiting);
 	priv->base.uevent = true;
diff --git a/drivers/gpu/drm/nouveau/nvc0_fence.c b/drivers/gpu/drm/nouveau/nvc0_fence.c
index 8213f7d..b7def39 100644
--- a/drivers/gpu/drm/nouveau/nvc0_fence.c
+++ b/drivers/gpu/drm/nouveau/nvc0_fence.c
@@ -66,6 +66,18 @@ nvc0_fence_sync32(struct nouveau_channel *chan, u64 virtual, u32 sequence)
 	return ret;
 }
 
+static int
+nvc0_fence_context_new(struct nouveau_channel *chan)
+{
+	int ret = nv84_fence_context_new(chan);
+	if (ret == 0) {
+		struct nv84_fence_chan *fctx = chan->fence;
+		fctx->base.emit32 = nvc0_fence_emit32;
+		fctx->base.sync32 = nvc0_fence_sync32;
+	}
+	return ret;
+}
+
 int
 nvc0_fence_create(struct nouveau_drm *drm)
 {
@@ -80,13 +92,8 @@ nvc0_fence_create(struct nouveau_drm *drm)
 	priv->base.dtor = nv84_fence_destroy;
 	priv->base.suspend = nv84_fence_suspend;
 	priv->base.resume = nv84_fence_resume;
-	priv->base.context_new = nv84_fence_context_new;
+	priv->base.context_new = nvc0_fence_context_new;
 	priv->base.context_del = nv84_fence_context_del;
-	priv->base.emit32 = nvc0_fence_emit32;
-	priv->base.emit = nv84_fence_emit;
-	priv->base.sync32 = nvc0_fence_sync32;
-	priv->base.sync = nv84_fence_sync;
-	priv->base.read = nv84_fence_read;
 
 	init_waitqueue_head(&priv->base.waiting);
 	priv->base.uevent = true;
-- 
cgit v0.10.2


From 264ce192b3e7f45d0adb37bfbab2b01a3fbe6c30 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Thu, 14 Feb 2013 13:43:21 +1000
Subject: drm/nv84-/fence: prepare for emit/sync support of sysram sequences

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 69d7b1d..fce944c 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -561,7 +561,7 @@ nouveau_bo_move_accel_cleanup(struct nouveau_channel *chan,
 	struct nouveau_fence *fence = NULL;
 	int ret;
 
-	ret = nouveau_fence_new(chan, &fence);
+	ret = nouveau_fence_new(chan, false, &fence);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_chan.c b/drivers/gpu/drm/nouveau/nouveau_chan.c
index 525a517..eaa80a2 100644
--- a/drivers/gpu/drm/nouveau/nouveau_chan.c
+++ b/drivers/gpu/drm/nouveau/nouveau_chan.c
@@ -51,7 +51,7 @@ nouveau_channel_idle(struct nouveau_channel *chan)
 	struct nouveau_fence *fence = NULL;
 	int ret;
 
-	ret = nouveau_fence_new(chan, &fence);
+	ret = nouveau_fence_new(chan, false, &fence);
 	if (!ret) {
 		ret = nouveau_fence_wait(fence, false, false);
 		nouveau_fence_unref(&fence);
diff --git a/drivers/gpu/drm/nouveau/nouveau_display.c b/drivers/gpu/drm/nouveau/nouveau_display.c
index 78fc5aa..de87417 100644
--- a/drivers/gpu/drm/nouveau/nouveau_display.c
+++ b/drivers/gpu/drm/nouveau/nouveau_display.c
@@ -540,7 +540,7 @@ nouveau_page_flip_emit(struct nouveau_channel *chan,
 	}
 	FIRE_RING (chan);
 
-	ret = nouveau_fence_new(chan, pfence);
+	ret = nouveau_fence_new(chan, false, pfence);
 	if (ret)
 		goto fail;
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.c b/drivers/gpu/drm/nouveau/nouveau_fence.c
index 6a7a5b5..6c94683 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fence.c
+++ b/drivers/gpu/drm/nouveau/nouveau_fence.c
@@ -41,8 +41,6 @@ nouveau_fence_context_del(struct nouveau_fence_chan *fctx)
 	struct nouveau_fence *fence, *fnext;
 	spin_lock(&fctx->lock);
 	list_for_each_entry_safe(fence, fnext, &fctx->pending, head) {
-		if (fence->work)
-			fence->work(fence->priv, false);
 		fence->channel = NULL;
 		list_del(&fence->head);
 		nouveau_fence_unref(&fence);
@@ -69,8 +67,6 @@ nouveau_fence_update(struct nouveau_channel *chan)
 		if (fctx->read(chan) < fence->sequence)
 			break;
 
-		if (fence->work)
-			fence->work(fence->priv, true);
 		fence->channel = NULL;
 		list_del(&fence->head);
 		nouveau_fence_unref(&fence);
@@ -256,7 +252,8 @@ nouveau_fence_ref(struct nouveau_fence *fence)
 }
 
 int
-nouveau_fence_new(struct nouveau_channel *chan, struct nouveau_fence **pfence)
+nouveau_fence_new(struct nouveau_channel *chan, bool sysmem,
+		  struct nouveau_fence **pfence)
 {
 	struct nouveau_fence *fence;
 	int ret = 0;
@@ -267,6 +264,8 @@ nouveau_fence_new(struct nouveau_channel *chan, struct nouveau_fence **pfence)
 	fence = kzalloc(sizeof(*fence), GFP_KERNEL);
 	if (!fence)
 		return -ENOMEM;
+
+	fence->sysmem = sysmem;
 	kref_init(&fence->kref);
 
 	ret = nouveau_fence_emit(fence, chan);
diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.h b/drivers/gpu/drm/nouveau/nouveau_fence.h
index a5c47e3..c899434 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fence.h
+++ b/drivers/gpu/drm/nouveau/nouveau_fence.h
@@ -7,15 +7,15 @@ struct nouveau_fence {
 	struct list_head head;
 	struct kref kref;
 
+	bool sysmem;
+
 	struct nouveau_channel *channel;
 	unsigned long timeout;
 	u32 sequence;
-
-	void (*work)(void *priv, bool signalled);
-	void *priv;
 };
 
-int  nouveau_fence_new(struct nouveau_channel *, struct nouveau_fence **);
+int  nouveau_fence_new(struct nouveau_channel *, bool sysmem,
+		       struct nouveau_fence **);
 struct nouveau_fence *
 nouveau_fence_ref(struct nouveau_fence *);
 void nouveau_fence_unref(struct nouveau_fence **);
@@ -79,24 +79,18 @@ int nouveau_flip_complete(void *chan);
 struct nv84_fence_chan {
 	struct nouveau_fence_chan base;
 	struct nouveau_vma vma;
+	struct nouveau_vma vma_gart;
 	struct nouveau_vma dispc_vma[4];
 };
 
 struct nv84_fence_priv {
 	struct nouveau_fence_priv base;
 	struct nouveau_bo *bo;
+	struct nouveau_bo *bo_gart;
 	u32 *suspend;
 };
 
 u64  nv84_fence_crtc(struct nouveau_channel *, int);
-int  nv84_fence_emit(struct nouveau_fence *);
-int  nv84_fence_sync(struct nouveau_fence *, struct nouveau_channel *,
-		     struct nouveau_channel *);
-u32  nv84_fence_read(struct nouveau_channel *);
 int  nv84_fence_context_new(struct nouveau_channel *);
-void nv84_fence_context_del(struct nouveau_channel *);
-bool nv84_fence_suspend(struct nouveau_drm *);
-void nv84_fence_resume(struct nouveau_drm *);
-void nv84_fence_destroy(struct nouveau_drm *);
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c b/drivers/gpu/drm/nouveau/nouveau_gem.c
index 7061138..6c45ddb 100644
--- a/drivers/gpu/drm/nouveau/nouveau_gem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_gem.c
@@ -787,7 +787,7 @@ nouveau_gem_ioctl_pushbuf(struct drm_device *dev, void *data,
 		}
 	}
 
-	ret = nouveau_fence_new(chan, &fence);
+	ret = nouveau_fence_new(chan, false, &fence);
 	if (ret) {
 		NV_ERROR(cli, "error fencing pushbuf: %d\n", ret);
 		WIND_RING(chan);
diff --git a/drivers/gpu/drm/nouveau/nv84_fence.c b/drivers/gpu/drm/nouveau/nv84_fence.c
index bc6493c..9fd475c 100644
--- a/drivers/gpu/drm/nouveau/nv84_fence.c
+++ b/drivers/gpu/drm/nouveau/nv84_fence.c
@@ -76,27 +76,39 @@ nv84_fence_sync32(struct nouveau_channel *chan, u64 virtual, u32 sequence)
 	return ret;
 }
 
-int
+static int
 nv84_fence_emit(struct nouveau_fence *fence)
 {
 	struct nouveau_channel *chan = fence->channel;
 	struct nv84_fence_chan *fctx = chan->fence;
 	struct nouveau_fifo_chan *fifo = (void *)chan->object;
-	u64 addr = fctx->vma.offset + fifo->chid * 16;
+	u64 addr = fifo->chid * 16;
+
+	if (fence->sysmem)
+		addr += fctx->vma_gart.offset;
+	else
+		addr += fctx->vma.offset;
+
 	return fctx->base.emit32(chan, addr, fence->sequence);
 }
 
-int
+static int
 nv84_fence_sync(struct nouveau_fence *fence,
 		struct nouveau_channel *prev, struct nouveau_channel *chan)
 {
 	struct nv84_fence_chan *fctx = chan->fence;
 	struct nouveau_fifo_chan *fifo = (void *)prev->object;
-	u64 addr = fctx->vma.offset + fifo->chid * 16;
+	u64 addr = fifo->chid * 16;
+
+	if (fence->sysmem)
+		addr += fctx->vma_gart.offset;
+	else
+		addr += fctx->vma.offset;
+
 	return fctx->base.sync32(chan, addr, fence->sequence);
 }
 
-u32
+static u32
 nv84_fence_read(struct nouveau_channel *chan)
 {
 	struct nouveau_fifo_chan *fifo = (void *)chan->object;
@@ -104,7 +116,7 @@ nv84_fence_read(struct nouveau_channel *chan)
 	return nouveau_bo_rd32(priv->bo, fifo->chid * 16/4);
 }
 
-void
+static void
 nv84_fence_context_del(struct nouveau_channel *chan)
 {
 	struct drm_device *dev = chan->drm->dev;
@@ -117,6 +129,7 @@ nv84_fence_context_del(struct nouveau_channel *chan)
 		nouveau_bo_vma_del(bo, &fctx->dispc_vma[i]);
 	}
 
+	nouveau_bo_vma_del(priv->bo, &fctx->vma_gart);
 	nouveau_bo_vma_del(priv->bo, &fctx->vma);
 	nouveau_fence_context_del(&fctx->base);
 	chan->fence = NULL;
@@ -144,8 +157,10 @@ nv84_fence_context_new(struct nouveau_channel *chan)
 	fctx->base.sync32 = nv84_fence_sync32;
 
 	ret = nouveau_bo_vma_add(priv->bo, client->vm, &fctx->vma);
-	if (ret)
-		nv84_fence_context_del(chan);
+	if (ret == 0) {
+		ret = nouveau_bo_vma_add(priv->bo_gart, client->vm,
+					&fctx->vma_gart);
+	}
 
 	/* map display semaphore buffers into channel's vm */
 	for (i = 0; !ret && i < chan->drm->dev->mode_config.num_crtc; i++) {
@@ -154,10 +169,13 @@ nv84_fence_context_new(struct nouveau_channel *chan)
 	}
 
 	nouveau_bo_wr32(priv->bo, fifo->chid * 16/4, 0x00000000);
+
+	if (ret)
+		nv84_fence_context_del(chan);
 	return ret;
 }
 
-bool
+static bool
 nv84_fence_suspend(struct nouveau_drm *drm)
 {
 	struct nouveau_fifo *pfifo = nouveau_fifo(drm->device);
@@ -173,7 +191,7 @@ nv84_fence_suspend(struct nouveau_drm *drm)
 	return priv->suspend != NULL;
 }
 
-void
+static void
 nv84_fence_resume(struct nouveau_drm *drm)
 {
 	struct nouveau_fifo *pfifo = nouveau_fifo(drm->device);
@@ -188,10 +206,14 @@ nv84_fence_resume(struct nouveau_drm *drm)
 	}
 }
 
-void
+static void
 nv84_fence_destroy(struct nouveau_drm *drm)
 {
 	struct nv84_fence_priv *priv = drm->fence;
+	nouveau_bo_unmap(priv->bo_gart);
+	if (priv->bo_gart)
+		nouveau_bo_unpin(priv->bo_gart);
+	nouveau_bo_ref(NULL, &priv->bo_gart);
 	nouveau_bo_unmap(priv->bo);
 	if (priv->bo)
 		nouveau_bo_unpin(priv->bo);
@@ -233,6 +255,21 @@ nv84_fence_create(struct nouveau_drm *drm)
 			nouveau_bo_ref(NULL, &priv->bo);
 	}
 
+	if (ret == 0)
+		ret = nouveau_bo_new(drm->dev, 16 * (pfifo->max + 1), 0,
+				     TTM_PL_FLAG_TT, 0, 0, NULL,
+				     &priv->bo_gart);
+	if (ret == 0) {
+		ret = nouveau_bo_pin(priv->bo_gart, TTM_PL_FLAG_TT);
+		if (ret == 0) {
+			ret = nouveau_bo_map(priv->bo_gart);
+			if (ret)
+				nouveau_bo_unpin(priv->bo_gart);
+		}
+		if (ret)
+			nouveau_bo_ref(NULL, &priv->bo_gart);
+	}
+
 	if (ret)
 		nv84_fence_destroy(drm);
 	return ret;
diff --git a/drivers/gpu/drm/nouveau/nvc0_fence.c b/drivers/gpu/drm/nouveau/nvc0_fence.c
index b7def39..9566267 100644
--- a/drivers/gpu/drm/nouveau/nvc0_fence.c
+++ b/drivers/gpu/drm/nouveau/nvc0_fence.c
@@ -81,37 +81,10 @@ nvc0_fence_context_new(struct nouveau_channel *chan)
 int
 nvc0_fence_create(struct nouveau_drm *drm)
 {
-	struct nouveau_fifo *pfifo = nouveau_fifo(drm->device);
-	struct nv84_fence_priv *priv;
-	int ret;
-
-	priv = drm->fence = kzalloc(sizeof(*priv), GFP_KERNEL);
-	if (!priv)
-		return -ENOMEM;
-
-	priv->base.dtor = nv84_fence_destroy;
-	priv->base.suspend = nv84_fence_suspend;
-	priv->base.resume = nv84_fence_resume;
-	priv->base.context_new = nvc0_fence_context_new;
-	priv->base.context_del = nv84_fence_context_del;
-
-	init_waitqueue_head(&priv->base.waiting);
-	priv->base.uevent = true;
-
-	ret = nouveau_bo_new(drm->dev, 16 * (pfifo->max + 1), 0,
-			     TTM_PL_FLAG_VRAM, 0, 0, NULL, &priv->bo);
+	int ret = nv84_fence_create(drm);
 	if (ret == 0) {
-		ret = nouveau_bo_pin(priv->bo, TTM_PL_FLAG_VRAM);
-		if (ret == 0) {
-			ret = nouveau_bo_map(priv->bo);
-			if (ret)
-				nouveau_bo_unpin(priv->bo);
-		}
-		if (ret)
-			nouveau_bo_ref(NULL, &priv->bo);
+		struct nv84_fence_priv *priv = drm->fence;
+		priv->base.context_new = nvc0_fence_context_new;
 	}
-
-	if (ret)
-		nv84_fence_destroy(drm);
 	return ret;
 }
-- 
cgit v0.10.2


From af1ac18a19b80f922046fd383a3a3e780ee94915 Mon Sep 17 00:00:00 2001
From: Dan Carpenter <dan.carpenter@oracle.com>
Date: Wed, 23 Jan 2013 11:27:56 +0300
Subject: drm/nouveau/disp: sizeof() wrong pointer

"data" is a void pointer and "args" is "data" after we have casted it to
a struct.  We care about the size of the struct here.  Btw,
sizeof(*data) is 1.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
index 0d26f00..78f09c3 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
@@ -333,7 +333,7 @@ nv50_disp_sync_ctor(struct nouveau_object *parent,
 	struct nv50_disp_dmac *dmac;
 	int ret;
 
-	if (size < sizeof(*data) || args->head > 1)
+	if (size < sizeof(*args) || args->head > 1)
 		return -EINVAL;
 
 	ret = nv50_disp_dmac_create_(parent, engine, oclass, args->pushbuf,
@@ -372,7 +372,7 @@ nv50_disp_ovly_ctor(struct nouveau_object *parent,
 	struct nv50_disp_dmac *dmac;
 	int ret;
 
-	if (size < sizeof(*data) || args->head > 1)
+	if (size < sizeof(*args) || args->head > 1)
 		return -EINVAL;
 
 	ret = nv50_disp_dmac_create_(parent, engine, oclass, args->pushbuf,
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
index 74626e8..4461899 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
@@ -228,7 +228,7 @@ nvd0_disp_sync_ctor(struct nouveau_object *parent,
 	struct nv50_disp_dmac *dmac;
 	int ret;
 
-	if (size < sizeof(*data) || args->head >= priv->head.nr)
+	if (size < sizeof(*args) || args->head >= priv->head.nr)
 		return -EINVAL;
 
 	ret = nv50_disp_dmac_create_(parent, engine, oclass, args->pushbuf,
@@ -268,7 +268,7 @@ nvd0_disp_ovly_ctor(struct nouveau_object *parent,
 	struct nv50_disp_dmac *dmac;
 	int ret;
 
-	if (size < sizeof(*data) || args->head >= priv->head.nr)
+	if (size < sizeof(*args) || args->head >= priv->head.nr)
 		return -EINVAL;
 
 	ret = nv50_disp_dmac_create_(parent, engine, oclass, args->pushbuf,
-- 
cgit v0.10.2


From 89e728ce07630a5584875b8a811a8bd02844b1d3 Mon Sep 17 00:00:00 2001
From: Stefan de Konink <stefan@konink.de>
Date: Sat, 22 Dec 2012 18:04:37 +0100
Subject: drm/nouveau: Fix DPMS 1 on G4 Snowball, from snow white to coal
 black.

Fixes https://bugs.freedesktop.org/show_bug.cgi?id=40275.

Signed-off-by: Stefan de Konink <stefan@konink.de>
Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nv04_dfp.c b/drivers/gpu/drm/nouveau/nv04_dfp.c
index 39ffc07..7e24cdf 100644
--- a/drivers/gpu/drm/nouveau/nv04_dfp.c
+++ b/drivers/gpu/drm/nouveau/nv04_dfp.c
@@ -490,8 +490,8 @@ static void nv04_dfp_update_backlight(struct drm_encoder *encoder, int mode)
 	/* BIOS scripts usually take care of the backlight, thanks
 	 * Apple for your consistency.
 	 */
-	if (dev->pci_device == 0x0179 || dev->pci_device == 0x0189 ||
-	    dev->pci_device == 0x0329) {
+	if (dev->pci_device == 0x0174 || dev->pci_device == 0x0179 ||
+	    dev->pci_device == 0x0189 || dev->pci_device == 0x0329) {
 		if (mode == DRM_MODE_DPMS_ON) {
 			nv_mask(device, NV_PBUS_DEBUG_DUALHEAD_CTL, 0, 1 << 31);
 			nv_mask(device, NV_PCRTC_GPIO_EXT, 3, 1);
-- 
cgit v0.10.2


From f63740fd580e0645c5123897891b72ec25b396ef Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Sun, 17 Feb 2013 11:31:17 +1000
Subject: drm/nv50/graph: avoid touching 400724, it doesn't exist

Harmless, but we now get MMIO fault reports, so silence it.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c
index d51f29d..f2b1a7a 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c
@@ -909,9 +909,8 @@ nv50_graph_init(struct nouveau_object *object)
 	nv_wr32(priv, 0x400828, 0x00000000);
 	nv_wr32(priv, 0x40082c, 0x00000000);
 	nv_wr32(priv, 0x400830, 0x00000000);
-	nv_wr32(priv, 0x400724, 0x00000000);
 	nv_wr32(priv, 0x40032c, 0x00000000);
-	nv_wr32(priv, 0x400320, 4);	/* CTXCTL_CMD = NEWCTXDMA */
+	nv_wr32(priv, 0x400330, 0x00000000);
 
 	/* some unknown zcull magic */
 	switch (nv_device(priv)->chipset & 0xf0) {
-- 
cgit v0.10.2


From df3ef6a1091fbdfb57306b0205edef33a1f1dcb4 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Mon, 11 Feb 2013 20:06:04 +1000
Subject: drm/nouveau/i2c: handle i2c/aux mux outside of port lookup function

Not quite how I want it yet, but, I'll fix that at some point.  For
right now, it's needed because find() won't necessarily be used right
before a transaction anymore.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/aux.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/aux.c
index dc27e79..1a42874 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/i2c/aux.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/aux.c
@@ -150,15 +150,29 @@ out:
 	return ret;
 }
 
+static void
+auxch_mux(struct nouveau_i2c_port *port)
+{
+	if (port->dcb & 0x00000100) {
+		u32 reg = 0x00e500 + (port->drive * 0x50);
+		/* nfi, but neither auxch or i2c work if it's 1 */
+		nv_mask(port->i2c, reg + 0x0c, 0x00000001, 0x00000000);
+		/* nfi, but switches auxch vs normal i2c */
+		nv_mask(port->i2c, reg + 0x00, 0x0000f003, 0x00002002);
+	}
+}
+
 int
 nv_rdaux(struct nouveau_i2c_port *auxch, u32 addr, u8 *data, u8 size)
 {
+	auxch_mux(auxch);
 	return auxch_tx(auxch->i2c, auxch->drive, 9, addr, data, size);
 }
 
 int
 nv_wraux(struct nouveau_i2c_port *auxch, u32 addr, u8 *data, u8 size)
 {
+	auxch_mux(auxch);
 	return auxch_tx(auxch->i2c, auxch->drive, 8, addr, data, size);
 }
 
@@ -169,6 +183,8 @@ aux_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, int num)
 	struct i2c_msg *msg = msgs;
 	int ret, mcnt = num;
 
+	auxch_mux(auxch);
+
 	while (mcnt--) {
 		u8 remaining = msg->len;
 		u8 *ptr = msg->buf;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
index dbfc2abf..1ee53d3 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
@@ -103,29 +103,10 @@ nouveau_i2c_find(struct nouveau_i2c *i2c, u8 index)
 
 	list_for_each_entry(port, &i2c->ports, head) {
 		if (port->index == index)
-			break;
-	}
-
-	if (&port->head == &i2c->ports)
-		return NULL;
-
-	if (nv_device(i2c)->card_type >= NV_50 && (port->dcb & 0x00000100)) {
-		u32 reg = 0x00e500, val;
-		if (port->type == 6) {
-			reg += port->drive * 0x50;
-			val  = 0x2002;
-		} else {
-			reg += ((port->dcb & 0x1e00) >> 9) * 0x50;
-			val  = 0xe001;
-		}
-
-		/* nfi, but neither auxch or i2c work if it's 1 */
-		nv_mask(i2c, reg + 0x0c, 0x00000001, 0x00000000);
-		/* nfi, but switches auxch vs normal i2c */
-		nv_mask(i2c, reg + 0x00, 0x0000f003, val);
+			return port;
 	}
 
-	return port;
+	return NULL;
 }
 
 static int
@@ -241,6 +222,23 @@ nouveau_i2c_sense_sda(void *data)
 	return 0;
 }
 
+static int
+nouveau_i2c_pre_xfer(struct i2c_adapter *adap)
+{
+	struct nouveau_i2c_port *port = (void *)adap;
+	struct nouveau_i2c *i2c = port->i2c;
+
+	if (nv_device(i2c)->card_type >= NV_50 && (port->dcb & 0x00000100)) {
+		u32 reg = 0x00e500 + ((port->dcb & 0x1e00) >> 9) * 0x50;
+		/* nfi, but neither auxch or i2c work if it's 1 */
+		nv_mask(i2c, reg + 0x0c, 0x00000001, 0x00000000);
+		/* nfi, but switches auxch vs normal i2c */
+		nv_mask(i2c, reg + 0x00, 0x0000f003, 0x0000e001);
+	}
+
+	return 0;
+}
+
 static const u32 nv50_i2c_port[] = {
 	0x00e138, 0x00e150, 0x00e168, 0x00e180,
 	0x00e254, 0x00e274, 0x00e764, 0x00e780,
@@ -347,6 +345,7 @@ nouveau_i2c_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 				port->bit.setscl = nouveau_i2c_drive_scl;
 				port->bit.getsda = nouveau_i2c_sense_sda;
 				port->bit.getscl = nouveau_i2c_sense_scl;
+				port->bit.pre_xfer = nouveau_i2c_pre_xfer;
 				ret = i2c_bit_add_bus(&port->adapter);
 			}
 		} else {
-- 
cgit v0.10.2


From 5ed502096f698b978c12a435f04be5afb195b485 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Mon, 11 Feb 2013 20:15:03 +1000
Subject: drm/nouveau: store i2c port pointer directly in nouveau_encoder

This is about to become somewhat more complicated to determine in a
number of cases, so store the "common" case (DDC/AUX) directly inside
the encoder structure.

Pre-nv50 code not touched except to fill the pointer, don't care.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.c b/drivers/gpu/drm/nouveau/nouveau_connector.c
index 9c4b3f5..4dd7ae2 100644
--- a/drivers/gpu/drm/nouveau/nouveau_connector.c
+++ b/drivers/gpu/drm/nouveau/nouveau_connector.c
@@ -112,7 +112,6 @@ nouveau_connector_ddc_detect(struct drm_connector *connector,
 	struct nouveau_connector *nv_connector = nouveau_connector(connector);
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	struct nouveau_gpio *gpio = nouveau_gpio(drm->device);
-	struct nouveau_i2c *i2c = nouveau_i2c(drm->device);
 	struct nouveau_i2c_port *port = NULL;
 	int i, panel = -ENODEV;
 
@@ -142,8 +141,7 @@ nouveau_connector_ddc_detect(struct drm_connector *connector,
 			continue;
 		nv_encoder = nouveau_encoder(obj_to_encoder(obj));
 
-		if (nv_encoder->dcb->i2c_index < 0xf)
-			port = i2c->find(i2c, nv_encoder->dcb->i2c_index);
+		port = nv_encoder->i2c;
 		if (port && nv_probe_i2c(port, 0x50)) {
 			*pnv_encoder = nv_encoder;
 			break;
diff --git a/drivers/gpu/drm/nouveau/nouveau_dp.c b/drivers/gpu/drm/nouveau/nouveau_dp.c
index a87c367..d4da57f 100644
--- a/drivers/gpu/drm/nouveau/nouveau_dp.c
+++ b/drivers/gpu/drm/nouveau/nouveau_dp.c
@@ -238,13 +238,12 @@ nouveau_dp_link_train(struct drm_encoder *encoder, u32 datarate,
 		nouveau_encoder_connector_get(nv_encoder);
 	struct drm_device *dev = encoder->dev;
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_i2c *i2c = nouveau_i2c(drm->device);
 	struct nouveau_gpio *gpio = nouveau_gpio(drm->device);
 	const u32 bw_list[] = { 270000, 162000, 0 };
 	const u32 *link_bw = bw_list;
 	struct dp_state dp;
 
-	dp.auxch = i2c->find(i2c, nv_encoder->dcb->i2c_index);
+	dp.auxch = nv_encoder->i2c;
 	if (!dp.auxch)
 		return false;
 
@@ -311,12 +310,10 @@ nouveau_dp_dpms(struct drm_encoder *encoder, int mode, u32 datarate,
 		struct nouveau_object *core)
 {
 	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct nouveau_drm *drm = nouveau_drm(encoder->dev);
-	struct nouveau_i2c *i2c = nouveau_i2c(drm->device);
 	struct nouveau_i2c_port *auxch;
 	u8 status;
 
-	auxch = i2c->find(i2c, nv_encoder->dcb->i2c_index);
+	auxch = nv_encoder->i2c;
 	if (!auxch)
 		return;
 
@@ -357,12 +354,11 @@ nouveau_dp_detect(struct drm_encoder *encoder)
 	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
 	struct drm_device *dev = encoder->dev;
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_i2c *i2c = nouveau_i2c(drm->device);
 	struct nouveau_i2c_port *auxch;
 	u8 *dpcd = nv_encoder->dp.dpcd;
 	int ret;
 
-	auxch = i2c->find(i2c, nv_encoder->dcb->i2c_index);
+	auxch = nv_encoder->i2c;
 	if (!auxch)
 		return false;
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_encoder.h b/drivers/gpu/drm/nouveau/nouveau_encoder.h
index d0d95bd..e243412 100644
--- a/drivers/gpu/drm/nouveau/nouveau_encoder.h
+++ b/drivers/gpu/drm/nouveau/nouveau_encoder.h
@@ -36,19 +36,12 @@
 
 struct nouveau_i2c_port;
 
-struct dp_train_func {
-	void (*link_set)(struct drm_device *, struct dcb_output *, int crtc,
-			 int nr, u32 bw, bool enhframe);
-	void (*train_set)(struct drm_device *, struct dcb_output *, u8 pattern);
-	void (*train_adj)(struct drm_device *, struct dcb_output *,
-			  u8 lane, u8 swing, u8 preem);
-};
-
 struct nouveau_encoder {
 	struct drm_encoder_slave base;
 
 	struct dcb_output *dcb;
 	int or;
+	struct nouveau_i2c_port *i2c;
 
 	/* different to drm_encoder.crtc, this reflects what's
 	 * actually programmed on the hw, not the proposed crtc */
diff --git a/drivers/gpu/drm/nouveau/nv04_display.c b/drivers/gpu/drm/nouveau/nv04_display.c
index edc3156..e61f740 100644
--- a/drivers/gpu/drm/nouveau/nv04_display.c
+++ b/drivers/gpu/drm/nouveau/nv04_display.c
@@ -34,6 +34,8 @@
 #include "nouveau_encoder.h"
 #include "nouveau_connector.h"
 
+#include <subdev/i2c.h>
+
 int
 nv04_display_early_init(struct drm_device *dev)
 {
@@ -56,6 +58,7 @@ int
 nv04_display_create(struct drm_device *dev)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nouveau_i2c *i2c = nouveau_i2c(drm->device);
 	struct dcb_table *dcb = &drm->vbios.dcb;
 	struct drm_connector *connector, *ct;
 	struct drm_encoder *encoder;
@@ -122,6 +125,11 @@ nv04_display_create(struct drm_device *dev)
 		}
 	}
 
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
+		struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
+		nv_encoder->i2c = i2c->find(i2c, nv_encoder->dcb->i2c_index);
+	}
+
 	/* Save previous state */
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head)
 		crtc->funcs->save(crtc);
diff --git a/drivers/gpu/drm/nouveau/nv50_display.c b/drivers/gpu/drm/nouveau/nv50_display.c
index 102a873..75d1161 100644
--- a/drivers/gpu/drm/nouveau/nv50_display.c
+++ b/drivers/gpu/drm/nouveau/nv50_display.c
@@ -43,6 +43,7 @@
 #include <subdev/timer.h>
 #include <subdev/bar.h>
 #include <subdev/fb.h>
+#include <subdev/i2c.h>
 
 #define EVO_DMA_NR 9
 
@@ -1554,20 +1555,23 @@ static const struct drm_encoder_funcs nv50_dac_func = {
 static int
 nv50_dac_create(struct drm_connector *connector, struct dcb_output *dcbe)
 {
-	struct drm_device *dev = connector->dev;
+	struct nouveau_drm *drm = nouveau_drm(connector->dev);
+	struct nouveau_i2c *i2c = nouveau_i2c(drm->device);
 	struct nouveau_encoder *nv_encoder;
 	struct drm_encoder *encoder;
+	int type = DRM_MODE_ENCODER_DAC;
 
 	nv_encoder = kzalloc(sizeof(*nv_encoder), GFP_KERNEL);
 	if (!nv_encoder)
 		return -ENOMEM;
 	nv_encoder->dcb = dcbe;
 	nv_encoder->or = ffs(dcbe->or) - 1;
+	nv_encoder->i2c = i2c->find(i2c, dcbe->i2c_index);
 
 	encoder = to_drm_encoder(nv_encoder);
 	encoder->possible_crtcs = dcbe->heads;
 	encoder->possible_clones = 0;
-	drm_encoder_init(dev, encoder, &nv50_dac_func, DRM_MODE_ENCODER_DAC);
+	drm_encoder_init(connector->dev, encoder, &nv50_dac_func, type);
 	drm_encoder_helper_add(encoder, &nv50_dac_hfunc);
 
 	drm_mode_connector_attach_encoder(connector, encoder);
@@ -1893,21 +1897,33 @@ static const struct drm_encoder_funcs nv50_sor_func = {
 static int
 nv50_sor_create(struct drm_connector *connector, struct dcb_output *dcbe)
 {
-	struct drm_device *dev = connector->dev;
+	struct nouveau_drm *drm = nouveau_drm(connector->dev);
+	struct nouveau_i2c *i2c = nouveau_i2c(drm->device);
 	struct nouveau_encoder *nv_encoder;
 	struct drm_encoder *encoder;
+	int type;
+
+	switch (dcbe->type) {
+	case DCB_OUTPUT_LVDS: type = DRM_MODE_ENCODER_LVDS; break;
+	case DCB_OUTPUT_TMDS:
+	case DCB_OUTPUT_DP:
+	default:
+		type = DRM_MODE_ENCODER_TMDS;
+		break;
+	}
 
 	nv_encoder = kzalloc(sizeof(*nv_encoder), GFP_KERNEL);
 	if (!nv_encoder)
 		return -ENOMEM;
 	nv_encoder->dcb = dcbe;
 	nv_encoder->or = ffs(dcbe->or) - 1;
+	nv_encoder->i2c = i2c->find(i2c, dcbe->i2c_index);
 	nv_encoder->last_dpms = DRM_MODE_DPMS_OFF;
 
 	encoder = to_drm_encoder(nv_encoder);
 	encoder->possible_crtcs = dcbe->heads;
 	encoder->possible_clones = 0;
-	drm_encoder_init(dev, encoder, &nv50_sor_func, DRM_MODE_ENCODER_TMDS);
+	drm_encoder_init(connector->dev, encoder, &nv50_sor_func, type);
 	drm_encoder_helper_add(encoder, &nv50_sor_hfunc);
 
 	drm_mode_connector_attach_encoder(connector, encoder);
-- 
cgit v0.10.2


From f3ed1048715f2edc10c4dda6148b60e93f6282ed Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Tue, 12 Feb 2013 10:16:31 +1000
Subject: drm/nouveau/bios: parse external transmitter type if off-chip

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/dcb.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/dcb.h
index b79025d..9e54678 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/bios/dcb.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/dcb.h
@@ -25,6 +25,7 @@ struct dcb_output {
 	uint8_t or;
 	uint8_t link;
 	bool duallink_possible;
+	uint8_t extdev;
 	union {
 		struct sor_conf {
 			int link;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/dcb.c b/drivers/gpu/drm/nouveau/core/subdev/bios/dcb.c
index 0fd87df..73a29da 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/dcb.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/dcb.c
@@ -135,6 +135,9 @@ dcb_outp_parse(struct nouveau_bios *bios, u8 idx, u8 *ver, u8 *len,
 			case DCB_OUTPUT_DP:
 				outp->link = (conf & 0x00000030) >> 4;
 				outp->sorconf.link = outp->link; /*XXX*/
+				outp->extdev = 0x00;
+				if (outp->location != 0)
+					outp->extdev = (conf & 0x0000ff00) >> 8;
 				break;
 			default:
 				break;
@@ -147,7 +150,7 @@ dcb_outp_parse(struct nouveau_bios *bios, u8 idx, u8 *ver, u8 *len,
 static inline u16
 dcb_outp_hasht(struct dcb_output *outp)
 {
-	return outp->type;
+	return (outp->location << 4) | outp->type;
 }
 
 static inline u16
-- 
cgit v0.10.2


From eaa8e7ab99d1b33db9362f35c1d65df8df39dea9 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Thu, 14 Feb 2013 19:54:42 -0500
Subject: drm/nouveau/i2c: fix a bit of a thinko in nv_wri2cr helper functions

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
index 1ee53d3..e09af32 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
@@ -46,13 +46,13 @@ nv_rdi2cr(struct nouveau_i2c_port *port, u8 addr, u8 reg)
 int
 nv_wri2cr(struct nouveau_i2c_port *port, u8 addr, u8 reg, u8 val)
 {
+	u8 buf[2] = { reg, val };
 	struct i2c_msg msgs[] = {
-		{ .addr = addr, .flags = 0, .len = 1, .buf = &reg },
-		{ .addr = addr, .flags = 0, .len = 1, .buf = &val },
+		{ .addr = addr, .flags = 0, .len = 2, .buf = buf },
 	};
 
-	int ret = i2c_transfer(&port->adapter, msgs, 2);
-	if (ret != 2)
+	int ret = i2c_transfer(&port->adapter, msgs, 1);
+	if (ret != 1)
 		return -EIO;
 
 	return 0;
-- 
cgit v0.10.2


From 31a34aa421032cfe3b2b892c929e7539e747a7ac Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Thu, 14 Feb 2013 20:59:41 -0500
Subject: drm/nouveau/i2c: aux channels not necessarily on nvio

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index 4208db9..29f8cc5 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -95,6 +95,7 @@ nouveau-y += core/subdev/gpio/nve0.o
 nouveau-y += core/subdev/i2c/base.o
 nouveau-y += core/subdev/i2c/aux.o
 nouveau-y += core/subdev/i2c/bit.o
+nouveau-y += core/subdev/i2c/nv94.o
 nouveau-y += core/subdev/ibus/nvc0.o
 nouveau-y += core/subdev/ibus/nve0.o
 nouveau-y += core/subdev/instmem/base.o
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/i2c.h b/drivers/gpu/drm/nouveau/core/include/subdev/i2c.h
index b93ab01..d2c067a 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/i2c.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/i2c.h
@@ -21,6 +21,8 @@ struct nouveau_i2c_port {
 	u32 drive;
 	u32 sense;
 	u32 state;
+	void (*aux_mux)(struct nouveau_i2c_port *);
+	int  (*aux)(struct nouveau_i2c_port *, u8, u32, u8 *, u8);
 };
 
 struct nouveau_i2c {
@@ -57,4 +59,7 @@ int nv_wraux(struct nouveau_i2c_port *, u32 addr, u8 *data, u8 size);
 extern const struct i2c_algorithm nouveau_i2c_bit_algo;
 extern const struct i2c_algorithm nouveau_i2c_aux_algo;
 
+void nv94_aux_mux(struct nouveau_i2c_port *);
+int  nv94_aux(struct nouveau_i2c_port *, u8, u32, u8 *, u8);
+
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/aux.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/aux.c
index 1a42874..7ad0d94 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/i2c/aux.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/aux.c
@@ -24,166 +24,39 @@
 
 #include <subdev/i2c.h>
 
-/******************************************************************************
- * aux channel util functions
- *****************************************************************************/
-#define AUX_DBG(fmt, args...) nv_debug(aux, "AUXCH(%d): " fmt, ch, ##args)
-#define AUX_ERR(fmt, args...) nv_error(aux, "AUXCH(%d): " fmt, ch, ##args)
-
-static void
-auxch_fini(struct nouveau_i2c *aux, int ch)
-{
-	nv_mask(aux, 0x00e4e4 + (ch * 0x50), 0x00310000, 0x00000000);
-}
-
-static int
-auxch_init(struct nouveau_i2c *aux, int ch)
-{
-	const u32 unksel = 1; /* nfi which to use, or if it matters.. */
-	const u32 ureq = unksel ? 0x00100000 : 0x00200000;
-	const u32 urep = unksel ? 0x01000000 : 0x02000000;
-	u32 ctrl, timeout;
-
-	/* wait up to 1ms for any previous transaction to be done... */
-	timeout = 1000;
-	do {
-		ctrl = nv_rd32(aux, 0x00e4e4 + (ch * 0x50));
-		udelay(1);
-		if (!timeout--) {
-			AUX_ERR("begin idle timeout 0x%08x\n", ctrl);
-			return -EBUSY;
-		}
-	} while (ctrl & 0x03010000);
-
-	/* set some magic, and wait up to 1ms for it to appear */
-	nv_mask(aux, 0x00e4e4 + (ch * 0x50), 0x00300000, ureq);
-	timeout = 1000;
-	do {
-		ctrl = nv_rd32(aux, 0x00e4e4 + (ch * 0x50));
-		udelay(1);
-		if (!timeout--) {
-			AUX_ERR("magic wait 0x%08x\n", ctrl);
-			auxch_fini(aux, ch);
-			return -EBUSY;
-		}
-	} while ((ctrl & 0x03000000) != urep);
-
-	return 0;
-}
-
-static int
-auxch_tx(struct nouveau_i2c *aux, int ch, u8 type, u32 addr, u8 *data, u8 size)
-{
-	u32 ctrl, stat, timeout, retries;
-	u32 xbuf[4] = {};
-	int ret, i;
-
-	AUX_DBG("%d: 0x%08x %d\n", type, addr, size);
-
-	ret = auxch_init(aux, ch);
-	if (ret)
-		goto out;
-
-	stat = nv_rd32(aux, 0x00e4e8 + (ch * 0x50));
-	if (!(stat & 0x10000000)) {
-		AUX_DBG("sink not detected\n");
-		ret = -ENXIO;
-		goto out;
-	}
-
-	if (!(type & 1)) {
-		memcpy(xbuf, data, size);
-		for (i = 0; i < 16; i += 4) {
-			AUX_DBG("wr 0x%08x\n", xbuf[i / 4]);
-			nv_wr32(aux, 0x00e4c0 + (ch * 0x50) + i, xbuf[i / 4]);
-		}
-	}
-
-	ctrl  = nv_rd32(aux, 0x00e4e4 + (ch * 0x50));
-	ctrl &= ~0x0001f0ff;
-	ctrl |= type << 12;
-	ctrl |= size - 1;
-	nv_wr32(aux, 0x00e4e0 + (ch * 0x50), addr);
-
-	/* retry transaction a number of times on failure... */
-	ret = -EREMOTEIO;
-	for (retries = 0; retries < 32; retries++) {
-		/* reset, and delay a while if this is a retry */
-		nv_wr32(aux, 0x00e4e4 + (ch * 0x50), 0x80000000 | ctrl);
-		nv_wr32(aux, 0x00e4e4 + (ch * 0x50), 0x00000000 | ctrl);
-		if (retries)
-			udelay(400);
-
-		/* transaction request, wait up to 1ms for it to complete */
-		nv_wr32(aux, 0x00e4e4 + (ch * 0x50), 0x00010000 | ctrl);
-
-		timeout = 1000;
-		do {
-			ctrl = nv_rd32(aux, 0x00e4e4 + (ch * 0x50));
-			udelay(1);
-			if (!timeout--) {
-				AUX_ERR("tx req timeout 0x%08x\n", ctrl);
-				goto out;
-			}
-		} while (ctrl & 0x00010000);
-
-		/* read status, and check if transaction completed ok */
-		stat = nv_mask(aux, 0x00e4e8 + (ch * 0x50), 0, 0);
-		if (!(stat & 0x000f0f00)) {
-			ret = 0;
-			break;
-		}
-
-		AUX_DBG("%02d 0x%08x 0x%08x\n", retries, ctrl, stat);
-	}
-
-	if (type & 1) {
-		for (i = 0; i < 16; i += 4) {
-			xbuf[i / 4] = nv_rd32(aux, 0x00e4d0 + (ch * 0x50) + i);
-			AUX_DBG("rd 0x%08x\n", xbuf[i / 4]);
-		}
-		memcpy(data, xbuf, size);
-	}
-
-out:
-	auxch_fini(aux, ch);
-	return ret;
-}
-
-static void
-auxch_mux(struct nouveau_i2c_port *port)
-{
-	if (port->dcb & 0x00000100) {
-		u32 reg = 0x00e500 + (port->drive * 0x50);
-		/* nfi, but neither auxch or i2c work if it's 1 */
-		nv_mask(port->i2c, reg + 0x0c, 0x00000001, 0x00000000);
-		/* nfi, but switches auxch vs normal i2c */
-		nv_mask(port->i2c, reg + 0x00, 0x0000f003, 0x00002002);
-	}
-}
-
 int
-nv_rdaux(struct nouveau_i2c_port *auxch, u32 addr, u8 *data, u8 size)
+nv_rdaux(struct nouveau_i2c_port *port, u32 addr, u8 *data, u8 size)
 {
-	auxch_mux(auxch);
-	return auxch_tx(auxch->i2c, auxch->drive, 9, addr, data, size);
+	if (port->aux) {
+		if (port->aux_mux)
+			port->aux_mux(port);
+		return port->aux(port, 9, addr, data, size);
+	}
+	return -ENODEV;
 }
 
 int
-nv_wraux(struct nouveau_i2c_port *auxch, u32 addr, u8 *data, u8 size)
+nv_wraux(struct nouveau_i2c_port *port, u32 addr, u8 *data, u8 size)
 {
-	auxch_mux(auxch);
-	return auxch_tx(auxch->i2c, auxch->drive, 8, addr, data, size);
+	if (port->aux) {
+		if (port->aux_mux)
+			port->aux_mux(port);
+		return port->aux(port, 8, addr, data, size);
+	}
+	return -ENODEV;
 }
 
 static int
 aux_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, int num)
 {
-	struct nouveau_i2c_port *auxch = (struct nouveau_i2c_port *)adap;
+	struct nouveau_i2c_port *port = (struct nouveau_i2c_port *)adap;
 	struct i2c_msg *msg = msgs;
 	int ret, mcnt = num;
 
-	auxch_mux(auxch);
+	if (!port->aux)
+		return -ENODEV;
+	if ( port->aux_mux)
+		port->aux_mux(port);
 
 	while (mcnt--) {
 		u8 remaining = msg->len;
@@ -201,8 +74,7 @@ aux_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, int num)
 			if (mcnt || remaining > 16)
 				cmd |= 4; /* MOT */
 
-			ret = auxch_tx(auxch->i2c, auxch->drive, cmd,
-				       msg->addr, ptr, cnt);
+			ret = port->aux(port, cmd, msg->addr, ptr, cnt);
 			if (ret < 0)
 				return ret;
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
index e09af32..bc8ec7a 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
@@ -303,6 +303,8 @@ nouveau_i2c_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 			port->drive = info.drive & 0x0f;
 			port->sense = port->drive;
 			port->adapter.algo = &nouveau_i2c_aux_algo;
+			port->aux_mux = nv94_aux_mux;
+			port->aux = nv94_aux;
 			break;
 		default:
 			break;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/nv94.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/nv94.c
new file mode 100644
index 0000000..4c498532
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/nv94.c
@@ -0,0 +1,165 @@
+/*
+ * Copyright 2009 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <subdev/i2c.h>
+
+/******************************************************************************
+ * aux channel util functions
+ *****************************************************************************/
+#define AUX_DBG(fmt, args...) nv_debug(aux, "AUXCH(%d): " fmt, ch, ##args)
+#define AUX_ERR(fmt, args...) nv_error(aux, "AUXCH(%d): " fmt, ch, ##args)
+
+static void
+auxch_fini(struct nouveau_i2c *aux, int ch)
+{
+	nv_mask(aux, 0x00e4e4 + (ch * 0x50), 0x00310000, 0x00000000);
+}
+
+static int
+auxch_init(struct nouveau_i2c *aux, int ch)
+{
+	const u32 unksel = 1; /* nfi which to use, or if it matters.. */
+	const u32 ureq = unksel ? 0x00100000 : 0x00200000;
+	const u32 urep = unksel ? 0x01000000 : 0x02000000;
+	u32 ctrl, timeout;
+
+	/* wait up to 1ms for any previous transaction to be done... */
+	timeout = 1000;
+	do {
+		ctrl = nv_rd32(aux, 0x00e4e4 + (ch * 0x50));
+		udelay(1);
+		if (!timeout--) {
+			AUX_ERR("begin idle timeout 0x%08x\n", ctrl);
+			return -EBUSY;
+		}
+	} while (ctrl & 0x03010000);
+
+	/* set some magic, and wait up to 1ms for it to appear */
+	nv_mask(aux, 0x00e4e4 + (ch * 0x50), 0x00300000, ureq);
+	timeout = 1000;
+	do {
+		ctrl = nv_rd32(aux, 0x00e4e4 + (ch * 0x50));
+		udelay(1);
+		if (!timeout--) {
+			AUX_ERR("magic wait 0x%08x\n", ctrl);
+			auxch_fini(aux, ch);
+			return -EBUSY;
+		}
+	} while ((ctrl & 0x03000000) != urep);
+
+	return 0;
+}
+
+int
+nv94_aux(struct nouveau_i2c_port *port, u8 type, u32 addr, u8 *data, u8 size)
+{
+	struct nouveau_i2c *aux = port->i2c;
+	u32 ctrl, stat, timeout, retries;
+	u32 xbuf[4] = {};
+	int ch = port->drive;
+	int ret, i;
+
+	AUX_DBG("%d: 0x%08x %d\n", type, addr, size);
+
+	ret = auxch_init(aux, ch);
+	if (ret)
+		goto out;
+
+	stat = nv_rd32(aux, 0x00e4e8 + (ch * 0x50));
+	if (!(stat & 0x10000000)) {
+		AUX_DBG("sink not detected\n");
+		ret = -ENXIO;
+		goto out;
+	}
+
+	if (!(type & 1)) {
+		memcpy(xbuf, data, size);
+		for (i = 0; i < 16; i += 4) {
+			AUX_DBG("wr 0x%08x\n", xbuf[i / 4]);
+			nv_wr32(aux, 0x00e4c0 + (ch * 0x50) + i, xbuf[i / 4]);
+		}
+	}
+
+	ctrl  = nv_rd32(aux, 0x00e4e4 + (ch * 0x50));
+	ctrl &= ~0x0001f0ff;
+	ctrl |= type << 12;
+	ctrl |= size - 1;
+	nv_wr32(aux, 0x00e4e0 + (ch * 0x50), addr);
+
+	/* retry transaction a number of times on failure... */
+	ret = -EREMOTEIO;
+	for (retries = 0; retries < 32; retries++) {
+		/* reset, and delay a while if this is a retry */
+		nv_wr32(aux, 0x00e4e4 + (ch * 0x50), 0x80000000 | ctrl);
+		nv_wr32(aux, 0x00e4e4 + (ch * 0x50), 0x00000000 | ctrl);
+		if (retries)
+			udelay(400);
+
+		/* transaction request, wait up to 1ms for it to complete */
+		nv_wr32(aux, 0x00e4e4 + (ch * 0x50), 0x00010000 | ctrl);
+
+		timeout = 1000;
+		do {
+			ctrl = nv_rd32(aux, 0x00e4e4 + (ch * 0x50));
+			udelay(1);
+			if (!timeout--) {
+				AUX_ERR("tx req timeout 0x%08x\n", ctrl);
+				goto out;
+			}
+		} while (ctrl & 0x00010000);
+
+		/* read status, and check if transaction completed ok */
+		stat = nv_mask(aux, 0x00e4e8 + (ch * 0x50), 0, 0);
+		if (!(stat & 0x000f0f00)) {
+			ret = 0;
+			break;
+		}
+
+		AUX_DBG("%02d 0x%08x 0x%08x\n", retries, ctrl, stat);
+	}
+
+	if (type & 1) {
+		for (i = 0; i < 16; i += 4) {
+			xbuf[i / 4] = nv_rd32(aux, 0x00e4d0 + (ch * 0x50) + i);
+			AUX_DBG("rd 0x%08x\n", xbuf[i / 4]);
+		}
+		memcpy(data, xbuf, size);
+	}
+
+out:
+	auxch_fini(aux, ch);
+	return ret;
+}
+
+void
+nv94_aux_mux(struct nouveau_i2c_port *port)
+{
+	if (port->dcb & 0x00000100) {
+		u32 reg = 0x00e500 + (port->drive * 0x50);
+		/* nfi, but neither auxch or i2c work if it's 1 */
+		nv_mask(port->i2c, reg + 0x0c, 0x00000001, 0x00000000);
+		/* nfi, but switches auxch vs normal i2c */
+		nv_mask(port->i2c, reg + 0x00, 0x0000f003, 0x00002002);
+	}
+}
-- 
cgit v0.10.2


From 548ddb6dec9964fc9c0812409f3e105b07324c4f Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Sat, 16 Feb 2013 13:19:18 +1000
Subject: drm/nouveau/i2c: extend type to 16-bits, add lookup-by-type function

For off-chip transmitters we won't necessarily have an i2c table entry
to lookup, but we can do it instead by encoding the type to include
the extdev type and looking that up instead.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/i2c.h b/drivers/gpu/drm/nouveau/core/include/subdev/i2c.h
index d2c067a..a804317 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/i2c.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/i2c.h
@@ -16,7 +16,7 @@ struct nouveau_i2c_port {
 	struct i2c_algo_bit_data bit;
 	struct list_head head;
 	u8  index;
-	u8  type;
+	u16 type;
 	u32 dcb;
 	u32 drive;
 	u32 sense;
@@ -29,6 +29,7 @@ struct nouveau_i2c {
 	struct nouveau_subdev base;
 
 	struct nouveau_i2c_port *(*find)(struct nouveau_i2c *, u8 index);
+	struct nouveau_i2c_port *(*find_type)(struct nouveau_i2c *, u16 type);
 	int (*identify)(struct nouveau_i2c *, int index,
 			const char *what, struct i2c_board_info *,
 			bool (*match)(struct nouveau_i2c_port *,
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
index bc8ec7a..6ab51d3 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
@@ -109,6 +109,19 @@ nouveau_i2c_find(struct nouveau_i2c *i2c, u8 index)
 	return NULL;
 }
 
+static struct nouveau_i2c_port *
+nouveau_i2c_find_type(struct nouveau_i2c *i2c, u16 type)
+{
+	struct nouveau_i2c_port *port;
+
+	list_for_each_entry(port, &i2c->ports, head) {
+		if (port->type == type)
+			return port;
+	}
+
+	return NULL;
+}
+
 static int
 nouveau_i2c_identify(struct nouveau_i2c *i2c, int index, const char *what,
 		     struct i2c_board_info *info,
@@ -264,6 +277,7 @@ nouveau_i2c_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		return ret;
 
 	i2c->find = nouveau_i2c_find;
+	i2c->find_type = nouveau_i2c_find_type;
 	i2c->identify = nouveau_i2c_identify;
 	INIT_LIST_HEAD(&i2c->ports);
 
-- 
cgit v0.10.2


From 8e992c8d9eebc2bd3246252ee5c0422dbbbce7ae Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Sat, 16 Feb 2013 12:01:59 +1000
Subject: drm/nouveau/bios: store a type/mask hash in parsed dcb data

Matches format used by a couple of other vbios tables, useful
to have laying around already calculated.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/dcb.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/dcb.h
index 9e54678..123270e9 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/bios/dcb.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/dcb.h
@@ -16,6 +16,8 @@ enum dcb_output_type {
 
 struct dcb_output {
 	int index;	/* may not be raw dcb index if merging has happened */
+	u16 hasht;
+	u16 hashm;
 	enum dcb_output_type type;
 	uint8_t i2c_index;
 	uint8_t heads;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/dcb.c b/drivers/gpu/drm/nouveau/core/subdev/bios/dcb.c
index 73a29da..2d9b9d7 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/dcb.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/dcb.c
@@ -107,6 +107,18 @@ dcb_outp(struct nouveau_bios *bios, u8 idx, u8 *ver, u8 *len)
 	return 0x0000;
 }
 
+static inline u16
+dcb_outp_hasht(struct dcb_output *outp)
+{
+	return (outp->extdev << 8) | (outp->location << 4) | outp->type;
+}
+
+static inline u16
+dcb_outp_hashm(struct dcb_output *outp)
+{
+	return (outp->heads << 8) | (outp->link << 6) | outp->or;
+}
+
 u16
 dcb_outp_parse(struct nouveau_bios *bios, u8 idx, u8 *ver, u8 *len,
 	       struct dcb_output *outp)
@@ -143,29 +155,20 @@ dcb_outp_parse(struct nouveau_bios *bios, u8 idx, u8 *ver, u8 *len,
 				break;
 			}
 		}
+
+		outp->hasht = dcb_outp_hasht(outp);
+		outp->hashm = dcb_outp_hashm(outp);
 	}
 	return dcb;
 }
 
-static inline u16
-dcb_outp_hasht(struct dcb_output *outp)
-{
-	return (outp->location << 4) | outp->type;
-}
-
-static inline u16
-dcb_outp_hashm(struct dcb_output *outp)
-{
-	return (outp->heads << 8) | (outp->link << 6) | outp->or;
-}
-
 u16
 dcb_outp_match(struct nouveau_bios *bios, u16 type, u16 mask,
 	       u8 *ver, u8 *len, struct dcb_output *outp)
 {
 	u16 dcb, idx = 0;
 	while ((dcb = dcb_outp_parse(bios, idx++, ver, len, outp))) {
-		if (dcb_outp_hasht(outp) == type) {
+		if ((dcb_outp_hasht(outp) & 0x00ff) == (type & 0x00ff)) {
 			if ((dcb_outp_hashm(outp) & mask) == mask)
 				break;
 		}
-- 
cgit v0.10.2


From ac8cc241a81941932da44993242e68c62e115ec7 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Sat, 16 Feb 2013 12:10:38 +1000
Subject: drm/nv50/devinit: reverse the logic for running encoder init scripts

A single U encoder table can match multiple DCB entries, whereas the
reverse is not true and can lead to us not matching a DCB entry at
all, and fail to initialise some encoders.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv50.c
index ae7249b..4a85778 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv50.c
@@ -78,12 +78,13 @@ nv50_devinit_init(struct nouveau_object *object)
 	if (ret)
 		return ret;
 
-	/* if we ran the init tables, execute first script pointer for each
-	 * display table output entry that has a matching dcb entry.
+	/* if we ran the init tables, we have to execute the first script
+	 * pointer of each dcb entry's display encoder table in order
+	 * to properly initialise each encoder.
 	 */
-	while (priv->base.post && ver) {
-		u16 data = nvbios_outp_parse(bios, i++, &ver, &hdr, &cnt, &len, &info);
-		if (data && dcb_outp_match(bios, info.type, info.mask, &ver, &len, &outp)) {
+	while (priv->base.post && dcb_outp_parse(bios, i, &ver, &hdr, &outp)) {
+		if (nvbios_outp_match(bios, outp.hasht, outp.hashm,
+				     &ver, &hdr, &cnt, &len, &info)) {
 			struct nvbios_init init = {
 				.subdev = nv_subdev(priv),
 				.bios = bios,
@@ -95,7 +96,8 @@ nv50_devinit_init(struct nouveau_object *object)
 
 			nvbios_exec(&init);
 		}
-	};
+		i++;
+	}
 
 	return 0;
 }
-- 
cgit v0.10.2


From 46c13c131d3b73080aa0f50f45e834a9ab3c0e71 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Sat, 16 Feb 2013 13:49:21 +1000
Subject: drm/nv50-/disp: 0x0000 is a valid udisp config value

Return an out-of-range value instead to signal a failure from
exec_clkcmp().

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
index 78f09c3..c3806122d 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
@@ -863,8 +863,8 @@ exec_clkcmp(struct nv50_disp_priv *priv, int head, int id, u32 pclk,
 	struct nvbios_outp info1;
 	struct nvbios_ocfg info2;
 	u8  ver, hdr, cnt, len;
-	u16 data, conf;
 	u32 ctrl = 0x00000000;
+	u32 data, conf = ~0;
 	int i;
 
 	for (i = 0; !(ctrl & (1 << head)) && i < 3; i++)
@@ -885,12 +885,12 @@ exec_clkcmp(struct nv50_disp_priv *priv, int head, int id, u32 pclk,
 	}
 
 	if (!(ctrl & (1 << head)))
-		return 0x0000;
+		return conf;
 	i--;
 
 	data = exec_lookup(priv, head, i, ctrl, outp, &ver, &hdr, &cnt, &len, &info1);
 	if (!data)
-		return 0x0000;
+		return conf;
 
 	switch (outp->type) {
 	case DCB_OUTPUT_TMDS:
@@ -923,13 +923,11 @@ exec_clkcmp(struct nv50_disp_priv *priv, int head, int id, u32 pclk,
 				.execute = 1,
 			};
 
-			if (nvbios_exec(&init))
-				return 0x0000;
-			return conf;
+			nvbios_exec(&init);
 		}
 	}
 
-	return 0x0000;
+	return conf;
 }
 
 static void
@@ -1082,7 +1080,7 @@ nv50_disp_intr_unk20(struct nv50_disp_priv *priv, u32 super)
 	if (head >= 0) {
 		u32 pclk = nv_rd32(priv, 0x610ad0 + (head * 0x540)) & 0x3fffff;
 		u32 conf = exec_clkcmp(priv, head, 0, pclk, &outp);
-		if (conf) {
+		if (conf != ~0) {
 			if (outp.type == DCB_OUTPUT_ANALOG) {
 				addr = 0x614280 + (ffs(outp.or) - 1) * 0x800;
 				mask = 0xffffffff;
@@ -1132,7 +1130,7 @@ nv50_disp_intr_unk40(struct nv50_disp_priv *priv, u32 super)
 	if (head >= 0) {
 		struct dcb_output outp;
 		u32 pclk = nv_rd32(priv, 0x610ad0 + (head * 0x540)) & 0x3fffff;
-		if (pclk && exec_clkcmp(priv, head, 1, pclk, &outp)) {
+		if (exec_clkcmp(priv, head, 1, pclk, &outp) != ~0) {
 			if (outp.type == DCB_OUTPUT_TMDS)
 				nv50_disp_intr_unk40_tmds(priv, &outp);
 		}
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
index 4461899..77cc730 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
@@ -657,11 +657,11 @@ exec_clkcmp(struct nv50_disp_priv *priv, int head, int outp,
 	struct nvbios_ocfg info2;
 	struct dcb_output dcb;
 	u8  ver, hdr, cnt, len;
-	u16 data, conf;
+	u32 data, conf = ~0;
 
 	data = exec_lookup(priv, head, outp, ctrl, &dcb, &ver, &hdr, &cnt, &len, &info1);
 	if (data == 0x0000)
-		return false;
+		return conf;
 
 	switch (dcb.type) {
 	case DCB_OUTPUT_TMDS:
@@ -694,13 +694,11 @@ exec_clkcmp(struct nv50_disp_priv *priv, int head, int outp,
 				.execute = 1,
 			};
 
-			if (nvbios_exec(&init))
-				return 0x0000;
-			return conf;
+			nvbios_exec(&init);
 		}
 	}
 
-	return 0x0000;
+	return conf;
 }
 
 static void
@@ -785,9 +783,10 @@ nvd0_display_unk2_handler(struct nv50_disp_priv *priv, u32 head, u32 mask)
 	nv_wr32(priv, 0x612200 + (head * 0x800), 0x00000000);
 
 	for (i = 0; mask && i < 8; i++) {
-		u32 mcp = nv_rd32(priv, 0x660180 + (i * 0x20)), cfg;
+		u32 mcp = nv_rd32(priv, 0x660180 + (i * 0x20));
 		if (mcp & (1 << head)) {
-			if ((cfg = exec_clkcmp(priv, head, i, mcp, 0, pclk))) {
+			u32 cfg = exec_clkcmp(priv, head, i, mcp, 0, pclk);
+			if (cfg != ~0) {
 				u32 addr, mask, data = 0x00000000;
 				if (i < 4) {
 					addr = 0x612280 + ((i - 0) * 0x800);
-- 
cgit v0.10.2


From 7dcd060c0e6e4f76ce7eea94c66437ce8f546530 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Sat, 16 Feb 2013 15:21:58 +1000
Subject: drm/nouveau/i2c: create proper chipset-specific class implementations

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index 29f8cc5..d8374ec 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -95,7 +95,11 @@ nouveau-y += core/subdev/gpio/nve0.o
 nouveau-y += core/subdev/i2c/base.o
 nouveau-y += core/subdev/i2c/aux.o
 nouveau-y += core/subdev/i2c/bit.o
+nouveau-y += core/subdev/i2c/nv04.o
+nouveau-y += core/subdev/i2c/nv4e.o
+nouveau-y += core/subdev/i2c/nv50.o
 nouveau-y += core/subdev/i2c/nv94.o
+nouveau-y += core/subdev/i2c/nvd0.o
 nouveau-y += core/subdev/ibus/nvc0.o
 nouveau-y += core/subdev/ibus/nve0.o
 nouveau-y += core/subdev/instmem/base.o
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/i2c.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/i2c.h
index 5079bedf..10b57a1 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/bios/i2c.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/i2c.h
@@ -15,7 +15,7 @@ struct dcb_i2c_entry {
 	enum dcb_i2c_type type;
 	u8 drive;
 	u8 sense;
-	u32 data;
+	u8 share;
 };
 
 u16 dcb_i2c_table(struct nouveau_bios *, u8 *ver, u8 *hdr, u8 *cnt, u8 *len);
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/i2c.h b/drivers/gpu/drm/nouveau/core/include/subdev/i2c.h
index a804317..c5868cd 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/i2c.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/i2c.h
@@ -10,21 +10,54 @@
 #define NV_I2C_PORT(n)    (0x00 + (n))
 #define NV_I2C_DEFAULT(n) (0x80 + (n))
 
+#define NV_I2C_TYPE_DCBI2C(n) (0x0000 | (n))
+#define NV_I2C_TYPE_EXTDDC(e) (0x0005 | (e) << 8)
+#define NV_I2C_TYPE_EXTAUX(e) (0x0006 | (e) << 8)
+
 struct nouveau_i2c_port {
+	struct nouveau_object base;
 	struct i2c_adapter adapter;
-	struct nouveau_i2c *i2c;
-	struct i2c_algo_bit_data bit;
+
 	struct list_head head;
 	u8  index;
-	u16 type;
-	u32 dcb;
-	u32 drive;
-	u32 sense;
-	u32 state;
-	void (*aux_mux)(struct nouveau_i2c_port *);
+
+	const struct nouveau_i2c_func *func;
+};
+
+struct nouveau_i2c_func {
+	void (*acquire)(struct nouveau_i2c_port *);
+	void (*release)(struct nouveau_i2c_port *);
+
+	void (*drive_scl)(struct nouveau_i2c_port *, int);
+	void (*drive_sda)(struct nouveau_i2c_port *, int);
+	int  (*sense_scl)(struct nouveau_i2c_port *);
+	int  (*sense_sda)(struct nouveau_i2c_port *);
+
 	int  (*aux)(struct nouveau_i2c_port *, u8, u32, u8 *, u8);
+	int  (*pattern)(struct nouveau_i2c_port *, int pattern);
+	int  (*lnk_ctl)(struct nouveau_i2c_port *, int nr, int bw, bool enh);
+	int  (*drv_ctl)(struct nouveau_i2c_port *, int lane, int sw, int pe);
 };
 
+#define nouveau_i2c_port_create(p,e,o,i,a,d)                                   \
+	nouveau_i2c_port_create_((p), (e), (o), (i), (a),                      \
+				 sizeof(**d), (void **)d)
+#define nouveau_i2c_port_destroy(p) ({                                         \
+	struct nouveau_i2c_port *port = (p);                                   \
+	_nouveau_i2c_port_dtor(nv_object(i2c));                                \
+})
+#define nouveau_i2c_port_init(p)                                               \
+	nouveau_object_init(&(p)->base)
+#define nouveau_i2c_port_fini(p,s)                                             \
+	nouveau_object_fini(&(p)->base, (s))
+
+int nouveau_i2c_port_create_(struct nouveau_object *, struct nouveau_object *,
+			     struct nouveau_oclass *, u8,
+			     const struct i2c_algorithm *, int, void **);
+void _nouveau_i2c_port_dtor(struct nouveau_object *);
+#define _nouveau_i2c_port_init nouveau_object_init
+#define _nouveau_i2c_port_fini nouveau_object_fini
+
 struct nouveau_i2c {
 	struct nouveau_subdev base;
 
@@ -43,24 +76,75 @@ nouveau_i2c(void *obj)
 	return (void *)nv_device(obj)->subdev[NVDEV_SUBDEV_I2C];
 }
 
-extern struct nouveau_oclass nouveau_i2c_oclass;
+#define nouveau_i2c_create(p,e,o,s,d)                                          \
+	nouveau_i2c_create_((p), (e), (o), (s), sizeof(**d), (void **)d)
+#define nouveau_i2c_destroy(p) ({                                              \
+	struct nouveau_i2c *i2c = (p);                                         \
+	_nouveau_i2c_dtor(nv_object(i2c));                                     \
+})
+#define nouveau_i2c_init(p) ({                                                 \
+	struct nouveau_i2c *i2c = (p);                                         \
+	_nouveau_i2c_init(nv_object(i2c));                                     \
+})
+#define nouveau_i2c_fini(p,s) ({                                               \
+	struct nouveau_i2c *i2c = (p);                                         \
+	_nouveau_i2c_fini(nv_object(i2c), (s));                                \
+})
 
-void nouveau_i2c_drive_scl(void *, int);
-void nouveau_i2c_drive_sda(void *, int);
-int  nouveau_i2c_sense_scl(void *);
-int  nouveau_i2c_sense_sda(void *);
+int nouveau_i2c_create_(struct nouveau_object *, struct nouveau_object *,
+			struct nouveau_oclass *, struct nouveau_oclass *,
+			int, void **);
+void _nouveau_i2c_dtor(struct nouveau_object *);
+int  _nouveau_i2c_init(struct nouveau_object *);
+int  _nouveau_i2c_fini(struct nouveau_object *, bool);
 
-int  nv_rdi2cr(struct nouveau_i2c_port *, u8 addr, u8 reg);
-int  nv_wri2cr(struct nouveau_i2c_port *, u8 addr, u8 reg, u8 val);
-bool nv_probe_i2c(struct nouveau_i2c_port *, u8 addr);
-
-int nv_rdaux(struct nouveau_i2c_port *, u32 addr, u8 *data, u8 size);
-int nv_wraux(struct nouveau_i2c_port *, u32 addr, u8 *data, u8 size);
+extern struct nouveau_oclass nv04_i2c_oclass;
+extern struct nouveau_oclass nv4e_i2c_oclass;
+extern struct nouveau_oclass nv50_i2c_oclass;
+extern struct nouveau_oclass nv94_i2c_oclass;
+extern struct nouveau_oclass nvd0_i2c_oclass;
 
 extern const struct i2c_algorithm nouveau_i2c_bit_algo;
 extern const struct i2c_algorithm nouveau_i2c_aux_algo;
 
-void nv94_aux_mux(struct nouveau_i2c_port *);
-int  nv94_aux(struct nouveau_i2c_port *, u8, u32, u8 *, u8);
+static inline int
+nv_rdi2cr(struct nouveau_i2c_port *port, u8 addr, u8 reg)
+{
+	u8 val;
+	struct i2c_msg msgs[] = {
+		{ .addr = addr, .flags = 0, .len = 1, .buf = &reg },
+		{ .addr = addr, .flags = I2C_M_RD, .len = 1, .buf = &val },
+	};
+
+	int ret = i2c_transfer(&port->adapter, msgs, 2);
+	if (ret != 2)
+		return -EIO;
+
+	return val;
+}
+
+static inline int
+nv_wri2cr(struct nouveau_i2c_port *port, u8 addr, u8 reg, u8 val)
+{
+	u8 buf[2] = { reg, val };
+	struct i2c_msg msgs[] = {
+		{ .addr = addr, .flags = 0, .len = 2, .buf = buf },
+	};
+
+	int ret = i2c_transfer(&port->adapter, msgs, 1);
+	if (ret != 1)
+		return -EIO;
+
+	return 0;
+}
+
+static inline bool
+nv_probe_i2c(struct nouveau_i2c_port *port, u8 addr)
+{
+	return nv_rdi2cr(port, addr, 0) >= 0;
+}
+
+int nv_rdaux(struct nouveau_i2c_port *, u32 addr, u8 *data, u8 size);
+int nv_wraux(struct nouveau_i2c_port *, u32 addr, u8 *data, u8 size);
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/i2c.c b/drivers/gpu/drm/nouveau/core/subdev/bios/i2c.c
index ad577db..cfb9288 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/i2c.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/i2c.c
@@ -70,12 +70,12 @@ dcb_i2c_parse(struct nouveau_bios *bios, u8 idx, struct dcb_i2c_entry *info)
 	u8  ver, len;
 	u16 ent = dcb_i2c_entry(bios, idx, &ver, &len);
 	if (ent) {
-		info->data = nv_ro32(bios, ent + 0);
-		info->type = nv_ro08(bios, ent + 3);
+		info->type  = nv_ro08(bios, ent + 3);
+		info->share = DCB_I2C_UNUSED;
 		if (ver < 0x30) {
 			info->type &= 0x07;
 			if (info->type == 0x07)
-				info->type = 0xff;
+				info->type = DCB_I2C_UNUSED;
 		}
 
 		switch (info->type) {
@@ -88,7 +88,11 @@ dcb_i2c_parse(struct nouveau_bios *bios, u8 idx, struct dcb_i2c_entry *info)
 			return 0;
 		case DCB_I2C_NVIO_BIT:
 		case DCB_I2C_NVIO_AUX:
-			info->drive = nv_ro08(bios, ent + 0);
+			info->drive = nv_ro08(bios, ent + 0) & 0x0f;
+			if (nv_ro08(bios, ent + 1) & 0x01) {
+				info->share  = nv_ro08(bios, ent + 1) >> 1;
+				info->share &= 0x0f;
+			}
 			return 0;
 		case DCB_I2C_UNUSED:
 			return 0;
@@ -121,7 +125,8 @@ dcb_i2c_parse(struct nouveau_bios *bios, u8 idx, struct dcb_i2c_entry *info)
 			if (!info->sense) info->sense = 0x36;
 		}
 
-		info->type = DCB_I2C_NV04_BIT;
+		info->type  = DCB_I2C_NV04_BIT;
+		info->share = DCB_I2C_UNUSED;
 		return 0;
 	}
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nv04.c b/drivers/gpu/drm/nouveau/core/subdev/device/nv04.c
index f76bd4e..473c5c0 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/device/nv04.c
@@ -47,7 +47,7 @@ nv04_identify(struct nouveau_device *device)
 	case 0x04:
 		device->cname = "NV04";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv04_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
@@ -65,7 +65,7 @@ nv04_identify(struct nouveau_device *device)
 	case 0x05:
 		device->cname = "NV05";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv05_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nv10.c b/drivers/gpu/drm/nouveau/core/subdev/device/nv10.c
index 46d55ea..d0774f5 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nv10.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/device/nv10.c
@@ -49,7 +49,7 @@ nv10_identify(struct nouveau_device *device)
 		device->cname = "NV10";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv10_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
@@ -66,7 +66,7 @@ nv10_identify(struct nouveau_device *device)
 		device->cname = "NV15";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv10_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
@@ -85,7 +85,7 @@ nv10_identify(struct nouveau_device *device)
 		device->cname = "NV16";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv10_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
@@ -104,7 +104,7 @@ nv10_identify(struct nouveau_device *device)
 		device->cname = "nForce";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
@@ -123,7 +123,7 @@ nv10_identify(struct nouveau_device *device)
 		device->cname = "NV11";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv10_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
@@ -142,7 +142,7 @@ nv10_identify(struct nouveau_device *device)
 		device->cname = "NV17";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv10_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
@@ -161,7 +161,7 @@ nv10_identify(struct nouveau_device *device)
 		device->cname = "nForce2";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
@@ -180,7 +180,7 @@ nv10_identify(struct nouveau_device *device)
 		device->cname = "NV18";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv10_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nv20.c b/drivers/gpu/drm/nouveau/core/subdev/device/nv20.c
index ad09854..ab920e0 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nv20.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/device/nv20.c
@@ -50,7 +50,7 @@ nv20_identify(struct nouveau_device *device)
 		device->cname = "NV20";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv20_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
@@ -69,7 +69,7 @@ nv20_identify(struct nouveau_device *device)
 		device->cname = "NV25";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv20_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
@@ -88,7 +88,7 @@ nv20_identify(struct nouveau_device *device)
 		device->cname = "NV28";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv20_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
@@ -107,7 +107,7 @@ nv20_identify(struct nouveau_device *device)
 		device->cname = "NV2A";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv20_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nv30.c b/drivers/gpu/drm/nouveau/core/subdev/device/nv30.c
index 2a98223..5f21102 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nv30.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/device/nv30.c
@@ -50,7 +50,7 @@ nv30_identify(struct nouveau_device *device)
 		device->cname = "NV30";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv20_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
@@ -69,7 +69,7 @@ nv30_identify(struct nouveau_device *device)
 		device->cname = "NV35";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv20_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
@@ -88,7 +88,7 @@ nv30_identify(struct nouveau_device *device)
 		device->cname = "NV31";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv20_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
@@ -108,7 +108,7 @@ nv30_identify(struct nouveau_device *device)
 		device->cname = "NV36";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv20_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
@@ -128,7 +128,7 @@ nv30_identify(struct nouveau_device *device)
 		device->cname = "NV34";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv04_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv10_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv04_mc_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nv40.c b/drivers/gpu/drm/nouveau/core/subdev/device/nv40.c
index 3e43460..f3d55ef 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/device/nv40.c
@@ -52,7 +52,7 @@ nv40_identify(struct nouveau_device *device)
 		device->cname = "NV40";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv40_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
@@ -73,7 +73,7 @@ nv40_identify(struct nouveau_device *device)
 		device->cname = "NV41";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv40_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
@@ -94,7 +94,7 @@ nv40_identify(struct nouveau_device *device)
 		device->cname = "NV42";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv40_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
@@ -115,7 +115,7 @@ nv40_identify(struct nouveau_device *device)
 		device->cname = "NV43";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv40_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
@@ -136,7 +136,7 @@ nv40_identify(struct nouveau_device *device)
 		device->cname = "NV45";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv40_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
@@ -157,7 +157,7 @@ nv40_identify(struct nouveau_device *device)
 		device->cname = "G70";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv40_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
@@ -178,7 +178,7 @@ nv40_identify(struct nouveau_device *device)
 		device->cname = "G71";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv40_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
@@ -199,7 +199,7 @@ nv40_identify(struct nouveau_device *device)
 		device->cname = "G73";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv40_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
@@ -220,7 +220,7 @@ nv40_identify(struct nouveau_device *device)
 		device->cname = "NV44";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv40_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
@@ -241,7 +241,7 @@ nv40_identify(struct nouveau_device *device)
 		device->cname = "G72";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv40_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
@@ -262,7 +262,7 @@ nv40_identify(struct nouveau_device *device)
 		device->cname = "NV44A";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv40_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
@@ -283,7 +283,7 @@ nv40_identify(struct nouveau_device *device)
 		device->cname = "C61";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv40_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
@@ -304,7 +304,7 @@ nv40_identify(struct nouveau_device *device)
 		device->cname = "C51";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv4e_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv40_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
@@ -325,7 +325,7 @@ nv40_identify(struct nouveau_device *device)
 		device->cname = "C73";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv40_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
@@ -346,7 +346,7 @@ nv40_identify(struct nouveau_device *device)
 		device->cname = "C67";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv40_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
@@ -367,7 +367,7 @@ nv40_identify(struct nouveau_device *device)
 		device->cname = "C68";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv10_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv04_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv40_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv40_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv1a_devinit_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/device/nv50.c
index 0444f46..5ed2fa5 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/device/nv50.c
@@ -58,7 +58,7 @@ nv50_identify(struct nouveau_device *device)
 		device->cname = "G80";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv50_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv50_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -81,7 +81,7 @@ nv50_identify(struct nouveau_device *device)
 		device->cname = "G84";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv50_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv50_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -107,7 +107,7 @@ nv50_identify(struct nouveau_device *device)
 		device->cname = "G86";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv50_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv50_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -133,7 +133,7 @@ nv50_identify(struct nouveau_device *device)
 		device->cname = "G92";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv50_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv50_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -159,7 +159,7 @@ nv50_identify(struct nouveau_device *device)
 		device->cname = "G94";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv94_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv50_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -185,7 +185,7 @@ nv50_identify(struct nouveau_device *device)
 		device->cname = "G96";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv94_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv50_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -211,7 +211,7 @@ nv50_identify(struct nouveau_device *device)
 		device->cname = "G98";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv94_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv50_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -237,7 +237,7 @@ nv50_identify(struct nouveau_device *device)
 		device->cname = "G200";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv50_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv50_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -263,7 +263,7 @@ nv50_identify(struct nouveau_device *device)
 		device->cname = "MCP77/MCP78";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv94_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv50_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -289,7 +289,7 @@ nv50_identify(struct nouveau_device *device)
 		device->cname = "MCP79/MCP7A";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv94_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv50_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -315,7 +315,7 @@ nv50_identify(struct nouveau_device *device)
 		device->cname = "GT215";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv94_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nva3_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -342,7 +342,7 @@ nv50_identify(struct nouveau_device *device)
 		device->cname = "GT216";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv94_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nva3_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -368,7 +368,7 @@ nv50_identify(struct nouveau_device *device)
 		device->cname = "GT218";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv94_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nva3_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -394,7 +394,7 @@ nv50_identify(struct nouveau_device *device)
 		device->cname = "MCP89";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv94_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nva3_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/device/nvc0.c
index a3e0e5d..4393eb4 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/device/nvc0.c
@@ -58,7 +58,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->cname = "GF100";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv94_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -87,7 +87,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->cname = "GF104";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv94_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -116,7 +116,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->cname = "GF106";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv94_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -145,7 +145,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->cname = "GF114";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv94_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -174,7 +174,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->cname = "GF116";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv94_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -203,7 +203,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->cname = "GF108";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv94_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -232,7 +232,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->cname = "GF110";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv94_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -261,7 +261,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->cname = "GF119";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nvd0_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nvd0_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nve0.c b/drivers/gpu/drm/nouveau/core/subdev/device/nve0.c
index 1e02bec..5c12391 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/device/nve0.c
@@ -58,7 +58,7 @@ nve0_identify(struct nouveau_device *device)
 		device->cname = "GK104";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nve0_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nvd0_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -87,7 +87,7 @@ nve0_identify(struct nouveau_device *device)
 		device->cname = "GK107";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nve0_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nvd0_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -116,7 +116,7 @@ nve0_identify(struct nouveau_device *device)
 		device->cname = "GK106";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nve0_gpio_oclass;
-		device->oclass[NVDEV_SUBDEV_I2C    ] = &nouveau_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nvd0_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/aux.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/aux.c
index 7ad0d94..5de074a 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/i2c/aux.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/aux.c
@@ -27,10 +27,10 @@
 int
 nv_rdaux(struct nouveau_i2c_port *port, u32 addr, u8 *data, u8 size)
 {
-	if (port->aux) {
-		if (port->aux_mux)
-			port->aux_mux(port);
-		return port->aux(port, 9, addr, data, size);
+	if (port->func->aux) {
+		if (port->func->acquire)
+			port->func->acquire(port);
+		return port->func->aux(port, 9, addr, data, size);
 	}
 	return -ENODEV;
 }
@@ -38,10 +38,10 @@ nv_rdaux(struct nouveau_i2c_port *port, u32 addr, u8 *data, u8 size)
 int
 nv_wraux(struct nouveau_i2c_port *port, u32 addr, u8 *data, u8 size)
 {
-	if (port->aux) {
-		if (port->aux_mux)
-			port->aux_mux(port);
-		return port->aux(port, 8, addr, data, size);
+	if (port->func->aux) {
+		if (port->func->acquire)
+			port->func->acquire(port);
+		return port->func->aux(port, 8, addr, data, size);
 	}
 	return -ENODEV;
 }
@@ -49,14 +49,14 @@ nv_wraux(struct nouveau_i2c_port *port, u32 addr, u8 *data, u8 size)
 static int
 aux_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, int num)
 {
-	struct nouveau_i2c_port *port = (struct nouveau_i2c_port *)adap;
+	struct nouveau_i2c_port *port = adap->algo_data;
 	struct i2c_msg *msg = msgs;
 	int ret, mcnt = num;
 
-	if (!port->aux)
+	if (!port->func->aux)
 		return -ENODEV;
-	if ( port->aux_mux)
-		port->aux_mux(port);
+	if ( port->func->acquire)
+		port->func->acquire(port);
 
 	while (mcnt--) {
 		u8 remaining = msg->len;
@@ -74,7 +74,7 @@ aux_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, int num)
 			if (mcnt || remaining > 16)
 				cmd |= 4; /* MOT */
 
-			ret = port->aux(port, cmd, msg->addr, ptr, cnt);
+			ret = port->func->aux(port, cmd, msg->addr, ptr, cnt);
 			if (ret < 0)
 				return ret;
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
index 6ab51d3..34a0523 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
@@ -1,5 +1,5 @@
 /*
- * Copyright 2012 Red Hat Inc.
+ * Copyright 2013 Red Hat Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -22,64 +22,133 @@
  * Authors: Ben Skeggs
  */
 
-#include "core/option.h"
+#include <core/option.h>
 
-#include "subdev/i2c.h"
-#include "subdev/vga.h"
+#include <subdev/i2c.h>
+#include <subdev/vga.h>
 
-int
-nv_rdi2cr(struct nouveau_i2c_port *port, u8 addr, u8 reg)
+/******************************************************************************
+ * interface to linux i2c bit-banging algorithm
+ *****************************************************************************/
+
+#ifdef CONFIG_NOUVEAU_I2C_INTERNAL_DEFAULT
+#define CSTMSEL true
+#else
+#define CSTMSEL false
+#endif
+
+static int
+nouveau_i2c_pre_xfer(struct i2c_adapter *adap)
 {
-	u8 val;
-	struct i2c_msg msgs[] = {
-		{ .addr = addr, .flags = 0, .len = 1, .buf = &reg },
-		{ .addr = addr, .flags = I2C_M_RD, .len = 1, .buf = &val },
-	};
+	struct i2c_algo_bit_data *bit = adap->algo_data;
+	struct nouveau_i2c_port *port = bit->data;
+	if (port->func->acquire)
+		port->func->acquire(port);
+	return 0;
+}
 
-	int ret = i2c_transfer(&port->adapter, msgs, 2);
-	if (ret != 2)
-		return -EIO;
+static void
+nouveau_i2c_setscl(void *data, int state)
+{
+	struct nouveau_i2c_port *port = data;
+	port->func->drive_scl(port, state);
+}
 
-	return val;
+static void
+nouveau_i2c_setsda(void *data, int state)
+{
+	struct nouveau_i2c_port *port = data;
+	port->func->drive_sda(port, state);
 }
 
-int
-nv_wri2cr(struct nouveau_i2c_port *port, u8 addr, u8 reg, u8 val)
+static int
+nouveau_i2c_getscl(void *data)
 {
-	u8 buf[2] = { reg, val };
-	struct i2c_msg msgs[] = {
-		{ .addr = addr, .flags = 0, .len = 2, .buf = buf },
-	};
+	struct nouveau_i2c_port *port = data;
+	return port->func->sense_scl(port);
+}
 
-	int ret = i2c_transfer(&port->adapter, msgs, 1);
-	if (ret != 1)
-		return -EIO;
+static int
+nouveau_i2c_getsda(void *data)
+{
+	struct nouveau_i2c_port *port = data;
+	return port->func->sense_sda(port);
+}
 
-	return 0;
+/******************************************************************************
+ * base i2c "port" class implementation
+ *****************************************************************************/
+
+void
+_nouveau_i2c_port_dtor(struct nouveau_object *object)
+{
+	struct nouveau_i2c_port *port = (void *)object;
+	i2c_del_adapter(&port->adapter);
+	nouveau_object_destroy(&port->base);
 }
 
-bool
-nv_probe_i2c(struct nouveau_i2c_port *port, u8 addr)
+int
+nouveau_i2c_port_create_(struct nouveau_object *parent,
+			 struct nouveau_object *engine,
+			 struct nouveau_oclass *oclass, u8 index,
+			 const struct i2c_algorithm *algo,
+			 int size, void **pobject)
 {
-	u8 buf[] = { 0 };
-	struct i2c_msg msgs[] = {
-		{
-			.addr = addr,
-			.flags = 0,
-			.len = 1,
-			.buf = buf,
-		},
-		{
-			.addr = addr,
-			.flags = I2C_M_RD,
-			.len = 1,
-			.buf = buf,
-		}
-	};
+	struct nouveau_device *device = nv_device(parent);
+	struct nouveau_i2c *i2c = (void *)engine;
+	struct nouveau_i2c_port *port;
+	int ret;
+
+	ret = nouveau_object_create_(parent, engine, oclass, 0, size, pobject);
+	port = *pobject;
+	if (ret)
+		return ret;
+
+	snprintf(port->adapter.name, sizeof(port->adapter.name),
+		 "nouveau-%s-%d", device->name, index);
+	port->adapter.owner = THIS_MODULE;
+	port->adapter.dev.parent = &device->pdev->dev;
+	port->index = index;
+	i2c_set_adapdata(&port->adapter, i2c);
+
+	if ( algo == &nouveau_i2c_bit_algo &&
+	    !nouveau_boolopt(device->cfgopt, "NvI2C", CSTMSEL)) {
+		struct i2c_algo_bit_data *bit;
+
+		bit = kzalloc(sizeof(*bit), GFP_KERNEL);
+		if (!bit)
+			return -ENOMEM;
+
+		bit->udelay = 10;
+		bit->timeout = usecs_to_jiffies(2200);
+		bit->data = port;
+		bit->pre_xfer = nouveau_i2c_pre_xfer;
+		bit->setsda = nouveau_i2c_setsda;
+		bit->setscl = nouveau_i2c_setscl;
+		bit->getsda = nouveau_i2c_getsda;
+		bit->getscl = nouveau_i2c_getscl;
+
+		port->adapter.algo_data = bit;
+		ret = i2c_bit_add_bus(&port->adapter);
+	} else {
+		port->adapter.algo_data = port;
+		port->adapter.algo = algo;
+		ret = i2c_add_adapter(&port->adapter);
+	}
+
+	/* drop port's i2c subdev refcount, i2c handles this itself */
+	if (ret == 0) {
+		list_add_tail(&port->head, &i2c->ports);
+		atomic_dec(&engine->refcount);
+	}
 
-	return i2c_transfer(&port->adapter, msgs, 2) == 2;
+	return ret;
 }
 
+/******************************************************************************
+ * base i2c subdev class implementation
+ *****************************************************************************/
+
 static struct nouveau_i2c_port *
 nouveau_i2c_find(struct nouveau_i2c *i2c, u8 index)
 {
@@ -115,7 +184,7 @@ nouveau_i2c_find_type(struct nouveau_i2c *i2c, u16 type)
 	struct nouveau_i2c_port *port;
 
 	list_for_each_entry(port, &i2c->ports, head) {
-		if (port->type == type)
+		if (nv_hclass(port) == type)
 			return port;
 	}
 
@@ -149,126 +218,78 @@ nouveau_i2c_identify(struct nouveau_i2c *i2c, int index, const char *what,
 	return -ENODEV;
 }
 
-void
-nouveau_i2c_drive_scl(void *data, int state)
+int
+_nouveau_i2c_fini(struct nouveau_object *object, bool suspend)
 {
-	struct nouveau_i2c_port *port = data;
+	struct nouveau_i2c *i2c = (void *)object;
+	struct nouveau_i2c_port *port;
+	int ret;
 
-	if (port->type == DCB_I2C_NV04_BIT) {
-		u8 val = nv_rdvgac(port->i2c, 0, port->drive);
-		if (state) val |= 0x20;
-		else	   val &= 0xdf;
-		nv_wrvgac(port->i2c, 0, port->drive, val | 0x01);
-	} else
-	if (port->type == DCB_I2C_NV4E_BIT) {
-		nv_mask(port->i2c, port->drive, 0x2f, state ? 0x21 : 0x01);
-	} else
-	if (port->type == DCB_I2C_NVIO_BIT) {
-		if (state) port->state |= 0x01;
-		else	   port->state &= 0xfe;
-		nv_wr32(port->i2c, port->drive, 4 | port->state);
+	list_for_each_entry(port, &i2c->ports, head) {
+		ret = nv_ofuncs(port)->fini(nv_object(port), suspend);
+		if (ret && suspend)
+			goto fail;
 	}
-}
-
-void
-nouveau_i2c_drive_sda(void *data, int state)
-{
-	struct nouveau_i2c_port *port = data;
 
-	if (port->type == DCB_I2C_NV04_BIT) {
-		u8 val = nv_rdvgac(port->i2c, 0, port->drive);
-		if (state) val |= 0x10;
-		else	   val &= 0xef;
-		nv_wrvgac(port->i2c, 0, port->drive, val | 0x01);
-	} else
-	if (port->type == DCB_I2C_NV4E_BIT) {
-		nv_mask(port->i2c, port->drive, 0x1f, state ? 0x11 : 0x01);
-	} else
-	if (port->type == DCB_I2C_NVIO_BIT) {
-		if (state) port->state |= 0x02;
-		else	   port->state &= 0xfd;
-		nv_wr32(port->i2c, port->drive, 4 | port->state);
+	return nouveau_subdev_fini(&i2c->base, suspend);
+fail:
+	list_for_each_entry_continue_reverse(port, &i2c->ports, head) {
+		nv_ofuncs(port)->init(nv_object(port));
 	}
+
+	return ret;
 }
 
 int
-nouveau_i2c_sense_scl(void *data)
+_nouveau_i2c_init(struct nouveau_object *object)
 {
-	struct nouveau_i2c_port *port = data;
-	struct nouveau_device *device = nv_device(port->i2c);
-
-	if (port->type == DCB_I2C_NV04_BIT) {
-		return !!(nv_rdvgac(port->i2c, 0, port->sense) & 0x04);
-	} else
-	if (port->type == DCB_I2C_NV4E_BIT) {
-		return !!(nv_rd32(port->i2c, port->sense) & 0x00040000);
-	} else
-	if (port->type == DCB_I2C_NVIO_BIT) {
-		if (device->card_type < NV_D0)
-			return !!(nv_rd32(port->i2c, port->sense) & 0x01);
-		else
-			return !!(nv_rd32(port->i2c, port->sense) & 0x10);
+	struct nouveau_i2c *i2c = (void *)object;
+	struct nouveau_i2c_port *port;
+	int ret;
+
+	ret = nouveau_subdev_init(&i2c->base);
+	if (ret == 0) {
+		list_for_each_entry(port, &i2c->ports, head) {
+			ret = nv_ofuncs(port)->init(nv_object(port));
+			if (ret)
+				goto fail;
+		}
 	}
 
-	return 0;
-}
-
-int
-nouveau_i2c_sense_sda(void *data)
-{
-	struct nouveau_i2c_port *port = data;
-	struct nouveau_device *device = nv_device(port->i2c);
-
-	if (port->type == DCB_I2C_NV04_BIT) {
-		return !!(nv_rdvgac(port->i2c, 0, port->sense) & 0x08);
-	} else
-	if (port->type == DCB_I2C_NV4E_BIT) {
-		return !!(nv_rd32(port->i2c, port->sense) & 0x00080000);
-	} else
-	if (port->type == DCB_I2C_NVIO_BIT) {
-		if (device->card_type < NV_D0)
-			return !!(nv_rd32(port->i2c, port->sense) & 0x02);
-		else
-			return !!(nv_rd32(port->i2c, port->sense) & 0x20);
+	return ret;
+fail:
+	list_for_each_entry_continue_reverse(port, &i2c->ports, head) {
+		nv_ofuncs(port)->fini(nv_object(port), false);
 	}
 
-	return 0;
+	return ret;
 }
 
-static int
-nouveau_i2c_pre_xfer(struct i2c_adapter *adap)
+void
+_nouveau_i2c_dtor(struct nouveau_object *object)
 {
-	struct nouveau_i2c_port *port = (void *)adap;
-	struct nouveau_i2c *i2c = port->i2c;
-
-	if (nv_device(i2c)->card_type >= NV_50 && (port->dcb & 0x00000100)) {
-		u32 reg = 0x00e500 + ((port->dcb & 0x1e00) >> 9) * 0x50;
-		/* nfi, but neither auxch or i2c work if it's 1 */
-		nv_mask(i2c, reg + 0x0c, 0x00000001, 0x00000000);
-		/* nfi, but switches auxch vs normal i2c */
-		nv_mask(i2c, reg + 0x00, 0x0000f003, 0x0000e001);
+	struct nouveau_i2c *i2c = (void *)object;
+	struct nouveau_i2c_port *port, *temp;
+
+	list_for_each_entry_safe(port, temp, &i2c->ports, head) {
+		nouveau_object_ref(NULL, (struct nouveau_object **)&port);
 	}
 
-	return 0;
+	nouveau_subdev_destroy(&i2c->base);
 }
 
-static const u32 nv50_i2c_port[] = {
-	0x00e138, 0x00e150, 0x00e168, 0x00e180,
-	0x00e254, 0x00e274, 0x00e764, 0x00e780,
-	0x00e79c, 0x00e7b8
-};
-
-static int
-nouveau_i2c_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
-		 struct nouveau_oclass *oclass, void *data, u32 size,
-		 struct nouveau_object **pobject)
+int
+nouveau_i2c_create_(struct nouveau_object *parent,
+		    struct nouveau_object *engine,
+		    struct nouveau_oclass *oclass,
+		    struct nouveau_oclass *sclass,
+		    int length, void **pobject)
 {
-	struct nouveau_device *device = nv_device(parent);
 	struct nouveau_bios *bios = nouveau_bios(parent);
-	struct nouveau_i2c_port *port;
 	struct nouveau_i2c *i2c;
+	struct nouveau_object *object;
 	struct dcb_i2c_entry info;
-	int ret, i = -1;
+	int ret, index = -1;
 
 	ret = nouveau_subdev_create(parent, engine, oclass, 0,
 				    "I2C", "i2c", &i2c);
@@ -281,142 +302,20 @@ nouveau_i2c_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	i2c->identify = nouveau_i2c_identify;
 	INIT_LIST_HEAD(&i2c->ports);
 
-	while (!dcb_i2c_parse(bios, ++i, &info)) {
+	while (!dcb_i2c_parse(bios, ++index, &info)) {
 		if (info.type == DCB_I2C_UNUSED)
 			continue;
 
-		port = kzalloc(sizeof(*port), GFP_KERNEL);
-		if (!port) {
-			nv_error(i2c, "failed port memory alloc at %d\n", i);
-			break;
-		}
-
-		port->type = info.type;
-		switch (port->type) {
-		case DCB_I2C_NV04_BIT:
-			port->drive = info.drive;
-			port->sense = info.sense;
-			break;
-		case DCB_I2C_NV4E_BIT:
-			port->drive = 0x600800 + info.drive;
-			port->sense = port->drive;
-			break;
-		case DCB_I2C_NVIO_BIT:
-			port->drive = info.drive & 0x0f;
-			if (device->card_type < NV_D0) {
-				if (port->drive >= ARRAY_SIZE(nv50_i2c_port))
-					break;
-				port->drive = nv50_i2c_port[port->drive];
-				port->sense = port->drive;
-			} else {
-				port->drive = 0x00d014 + (port->drive * 0x20);
-				port->sense = port->drive;
-			}
-			break;
-		case DCB_I2C_NVIO_AUX:
-			port->drive = info.drive & 0x0f;
-			port->sense = port->drive;
-			port->adapter.algo = &nouveau_i2c_aux_algo;
-			port->aux_mux = nv94_aux_mux;
-			port->aux = nv94_aux;
-			break;
-		default:
-			break;
-		}
-
-		if (!port->adapter.algo && !port->drive) {
-			nv_error(i2c, "I2C%d: type %d index %x/%x unknown\n",
-				 i, port->type, port->drive, port->sense);
-			kfree(port);
-			continue;
-		}
-
-		snprintf(port->adapter.name, sizeof(port->adapter.name),
-			 "nouveau-%s-%d", device->name, i);
-		port->adapter.owner = THIS_MODULE;
-		port->adapter.dev.parent = &device->pdev->dev;
-		port->i2c = i2c;
-		port->index = i;
-		port->dcb = info.data;
-		i2c_set_adapdata(&port->adapter, i2c);
-
-		if (port->adapter.algo != &nouveau_i2c_aux_algo) {
-			nouveau_i2c_drive_scl(port, 0);
-			nouveau_i2c_drive_sda(port, 1);
-			nouveau_i2c_drive_scl(port, 1);
-
-#ifdef CONFIG_NOUVEAU_I2C_INTERNAL_DEFAULT
-			if (nouveau_boolopt(device->cfgopt, "NvI2C", true)) {
-#else
-			if (nouveau_boolopt(device->cfgopt, "NvI2C", false)) {
-#endif
-				port->adapter.algo = &nouveau_i2c_bit_algo;
-				ret = i2c_add_adapter(&port->adapter);
-			} else {
-				port->adapter.algo_data = &port->bit;
-				port->bit.udelay = 10;
-				port->bit.timeout = usecs_to_jiffies(2200);
-				port->bit.data = port;
-				port->bit.setsda = nouveau_i2c_drive_sda;
-				port->bit.setscl = nouveau_i2c_drive_scl;
-				port->bit.getsda = nouveau_i2c_sense_sda;
-				port->bit.getscl = nouveau_i2c_sense_scl;
-				port->bit.pre_xfer = nouveau_i2c_pre_xfer;
-				ret = i2c_bit_add_bus(&port->adapter);
+		oclass = sclass;
+		do {
+			ret = -EINVAL;
+			if (oclass->handle == info.type) {
+				ret = nouveau_object_ctor(*pobject, *pobject,
+							  oclass, &info,
+							  index, &object);
 			}
-		} else {
-			port->adapter.algo = &nouveau_i2c_aux_algo;
-			ret = i2c_add_adapter(&port->adapter);
-		}
-
-		if (ret) {
-			nv_error(i2c, "I2C%d: failed register: %d\n", i, ret);
-			kfree(port);
-			continue;
-		}
-
-		list_add_tail(&port->head, &i2c->ports);
+		} while (ret && (++oclass)->handle);
 	}
 
 	return 0;
 }
-
-static void
-nouveau_i2c_dtor(struct nouveau_object *object)
-{
-	struct nouveau_i2c *i2c = (void *)object;
-	struct nouveau_i2c_port *port, *temp;
-
-	list_for_each_entry_safe(port, temp, &i2c->ports, head) {
-		i2c_del_adapter(&port->adapter);
-		list_del(&port->head);
-		kfree(port);
-	}
-
-	nouveau_subdev_destroy(&i2c->base);
-}
-
-static int
-nouveau_i2c_init(struct nouveau_object *object)
-{
-	struct nouveau_i2c *i2c = (void *)object;
-	return nouveau_subdev_init(&i2c->base);
-}
-
-static int
-nouveau_i2c_fini(struct nouveau_object *object, bool suspend)
-{
-	struct nouveau_i2c *i2c = (void *)object;
-	return nouveau_subdev_fini(&i2c->base, suspend);
-}
-
-struct nouveau_oclass
-nouveau_i2c_oclass = {
-	.handle = NV_SUBDEV(I2C, 0x00),
-	.ofuncs = &(struct nouveau_ofuncs) {
-		.ctor = nouveau_i2c_ctor,
-		.dtor = nouveau_i2c_dtor,
-		.init = nouveau_i2c_init,
-		.fini = nouveau_i2c_fini,
-	},
-};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/bit.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/bit.c
index 1c4c9a5..a6e72d3 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/i2c/bit.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/bit.c
@@ -32,25 +32,25 @@
 static inline void
 i2c_drive_scl(struct nouveau_i2c_port *port, int state)
 {
-	nouveau_i2c_drive_scl(port, state);
+	port->func->drive_scl(port, state);
 }
 
 static inline void
 i2c_drive_sda(struct nouveau_i2c_port *port, int state)
 {
-	nouveau_i2c_drive_sda(port, state);
+	port->func->drive_sda(port, state);
 }
 
 static inline int
 i2c_sense_scl(struct nouveau_i2c_port *port)
 {
-	return nouveau_i2c_sense_scl(port);
+	return port->func->sense_scl(port);
 }
 
 static inline int
 i2c_sense_sda(struct nouveau_i2c_port *port)
 {
-	return nouveau_i2c_sense_sda(port);
+	return port->func->sense_sda(port);
 }
 
 static void
@@ -77,9 +77,8 @@ i2c_start(struct nouveau_i2c_port *port)
 {
 	int ret = 0;
 
-	port->state  = i2c_sense_scl(port);
-	port->state |= i2c_sense_sda(port) << 1;
-	if (port->state != 3) {
+	if (!i2c_sense_scl(port) ||
+	    !i2c_sense_sda(port)) {
 		i2c_drive_scl(port, 0);
 		i2c_drive_sda(port, 1);
 		if (!i2c_raise_scl(port))
@@ -184,10 +183,13 @@ i2c_addr(struct nouveau_i2c_port *port, struct i2c_msg *msg)
 static int
 i2c_bit_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, int num)
 {
-	struct nouveau_i2c_port *port = (struct nouveau_i2c_port *)adap;
+	struct nouveau_i2c_port *port = adap->algo_data;
 	struct i2c_msg *msg = msgs;
 	int ret = 0, mcnt = num;
 
+	if (port->func->acquire)
+		port->func->acquire(port);
+
 	while (!ret && mcnt--) {
 		u8 remaining = msg->len;
 		u8 *ptr = msg->buf;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/nv04.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/nv04.c
new file mode 100644
index 0000000..2ad1884
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/nv04.c
@@ -0,0 +1,143 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <subdev/i2c.h>
+#include <subdev/vga.h>
+
+struct nv04_i2c_priv {
+	struct nouveau_i2c base;
+};
+
+struct nv04_i2c_port {
+	struct nouveau_i2c_port base;
+	u8 drive;
+	u8 sense;
+};
+
+static void
+nv04_i2c_drive_scl(struct nouveau_i2c_port *base, int state)
+{
+	struct nv04_i2c_priv *priv = (void *)nv_object(base)->engine;
+	struct nv04_i2c_port *port = (void *)base;
+	u8 val = nv_rdvgac(priv, 0, port->drive);
+	if (state) val |= 0x20;
+	else	   val &= 0xdf;
+	nv_wrvgac(priv, 0, port->drive, val | 0x01);
+}
+
+static void
+nv04_i2c_drive_sda(struct nouveau_i2c_port *base, int state)
+{
+	struct nv04_i2c_priv *priv = (void *)nv_object(base)->engine;
+	struct nv04_i2c_port *port = (void *)base;
+	u8 val = nv_rdvgac(priv, 0, port->drive);
+	if (state) val |= 0x10;
+	else	   val &= 0xef;
+	nv_wrvgac(priv, 0, port->drive, val | 0x01);
+}
+
+static int
+nv04_i2c_sense_scl(struct nouveau_i2c_port *base)
+{
+	struct nv04_i2c_priv *priv = (void *)nv_object(base)->engine;
+	struct nv04_i2c_port *port = (void *)base;
+	return !!(nv_rdvgac(priv, 0, port->sense) & 0x04);
+}
+
+static int
+nv04_i2c_sense_sda(struct nouveau_i2c_port *base)
+{
+	struct nv04_i2c_priv *priv = (void *)nv_object(base)->engine;
+	struct nv04_i2c_port *port = (void *)base;
+	return !!(nv_rdvgac(priv, 0, port->sense) & 0x08);
+}
+
+static const struct nouveau_i2c_func
+nv04_i2c_func = {
+	.drive_scl = nv04_i2c_drive_scl,
+	.drive_sda = nv04_i2c_drive_sda,
+	.sense_scl = nv04_i2c_sense_scl,
+	.sense_sda = nv04_i2c_sense_sda,
+};
+
+static int
+nv04_i2c_port_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+		   struct nouveau_oclass *oclass, void *data, u32 index,
+		   struct nouveau_object **pobject)
+{
+	struct dcb_i2c_entry *info = data;
+	struct nv04_i2c_port *port;
+	int ret;
+
+	ret = nouveau_i2c_port_create(parent, engine, oclass, index,
+				     &nouveau_i2c_bit_algo, &port);
+	*pobject = nv_object(port);
+	if (ret)
+		return ret;
+
+	port->base.func = &nv04_i2c_func;
+	port->drive = info->drive;
+	port->sense = info->sense;
+	return 0;
+}
+
+static struct nouveau_oclass
+nv04_i2c_sclass[] = {
+	{ .handle = NV_I2C_TYPE_DCBI2C(DCB_I2C_NV04_BIT),
+	  .ofuncs = &(struct nouveau_ofuncs) {
+		  .ctor = nv04_i2c_port_ctor,
+		  .dtor = _nouveau_i2c_port_dtor,
+		  .init = _nouveau_i2c_port_init,
+		  .fini = _nouveau_i2c_port_fini,
+	  },
+	},
+	{}
+};
+
+static int
+nv04_i2c_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+	      struct nouveau_oclass *oclass, void *data, u32 size,
+	      struct nouveau_object **pobject)
+{
+	struct nv04_i2c_priv *priv;
+	int ret;
+
+	ret = nouveau_i2c_create(parent, engine, oclass, nv04_i2c_sclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+struct nouveau_oclass
+nv04_i2c_oclass = {
+	.handle = NV_SUBDEV(I2C, 0x04),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv04_i2c_ctor,
+		.dtor = _nouveau_i2c_dtor,
+		.init = _nouveau_i2c_init,
+		.fini = _nouveau_i2c_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/nv4e.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/nv4e.c
new file mode 100644
index 0000000..f501ae2
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/nv4e.c
@@ -0,0 +1,135 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <subdev/i2c.h>
+#include <subdev/vga.h>
+
+struct nv4e_i2c_priv {
+	struct nouveau_i2c base;
+};
+
+struct nv4e_i2c_port {
+	struct nouveau_i2c_port base;
+	u32 addr;
+};
+
+static void
+nv4e_i2c_drive_scl(struct nouveau_i2c_port *base, int state)
+{
+	struct nv4e_i2c_priv *priv = (void *)nv_object(base)->engine;
+	struct nv4e_i2c_port *port = (void *)base;
+	nv_mask(priv, port->addr, 0x2f, state ? 0x21 : 0x01);
+}
+
+static void
+nv4e_i2c_drive_sda(struct nouveau_i2c_port *base, int state)
+{
+	struct nv4e_i2c_priv *priv = (void *)nv_object(base)->engine;
+	struct nv4e_i2c_port *port = (void *)base;
+	nv_mask(priv, port->addr, 0x1f, state ? 0x11 : 0x01);
+}
+
+static int
+nv4e_i2c_sense_scl(struct nouveau_i2c_port *base)
+{
+	struct nv4e_i2c_priv *priv = (void *)nv_object(base)->engine;
+	struct nv4e_i2c_port *port = (void *)base;
+	return !!(nv_rd32(priv, port->addr) & 0x00040000);
+}
+
+static int
+nv4e_i2c_sense_sda(struct nouveau_i2c_port *base)
+{
+	struct nv4e_i2c_priv *priv = (void *)nv_object(base)->engine;
+	struct nv4e_i2c_port *port = (void *)base;
+	return !!(nv_rd32(priv, port->addr) & 0x00080000);
+}
+
+static const struct nouveau_i2c_func
+nv4e_i2c_func = {
+	.drive_scl = nv4e_i2c_drive_scl,
+	.drive_sda = nv4e_i2c_drive_sda,
+	.sense_scl = nv4e_i2c_sense_scl,
+	.sense_sda = nv4e_i2c_sense_sda,
+};
+
+static int
+nv4e_i2c_port_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+		   struct nouveau_oclass *oclass, void *data, u32 index,
+		   struct nouveau_object **pobject)
+{
+	struct dcb_i2c_entry *info = data;
+	struct nv4e_i2c_port *port;
+	int ret;
+
+	ret = nouveau_i2c_port_create(parent, engine, oclass, index,
+				     &nouveau_i2c_bit_algo, &port);
+	*pobject = nv_object(port);
+	if (ret)
+		return ret;
+
+	port->base.func = &nv4e_i2c_func;
+	port->addr = 0x600800 + info->drive;
+	return 0;
+}
+
+static struct nouveau_oclass
+nv4e_i2c_sclass[] = {
+	{ .handle = NV_I2C_TYPE_DCBI2C(DCB_I2C_NV4E_BIT),
+	  .ofuncs = &(struct nouveau_ofuncs) {
+		  .ctor = nv4e_i2c_port_ctor,
+		  .dtor = _nouveau_i2c_port_dtor,
+		  .init = _nouveau_i2c_port_init,
+		  .fini = _nouveau_i2c_port_fini,
+	  },
+	},
+	{}
+};
+
+static int
+nv4e_i2c_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+	      struct nouveau_oclass *oclass, void *data, u32 size,
+	      struct nouveau_object **pobject)
+{
+	struct nv4e_i2c_priv *priv;
+	int ret;
+
+	ret = nouveau_i2c_create(parent, engine, oclass, nv4e_i2c_sclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+struct nouveau_oclass
+nv4e_i2c_oclass = {
+	.handle = NV_SUBDEV(I2C, 0x4e),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv4e_i2c_ctor,
+		.dtor = _nouveau_i2c_dtor,
+		.init = _nouveau_i2c_init,
+		.fini = _nouveau_i2c_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/nv50.c
new file mode 100644
index 0000000..378dfa3
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/nv50.c
@@ -0,0 +1,149 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "nv50.h"
+
+void
+nv50_i2c_drive_scl(struct nouveau_i2c_port *base, int state)
+{
+	struct nv50_i2c_priv *priv = (void *)nv_object(base)->engine;
+	struct nv50_i2c_port *port = (void *)base;
+	if (state) port->state |= 0x01;
+	else	   port->state &= 0xfe;
+	nv_wr32(priv, port->addr, port->state);
+}
+
+void
+nv50_i2c_drive_sda(struct nouveau_i2c_port *base, int state)
+{
+	struct nv50_i2c_priv *priv = (void *)nv_object(base)->engine;
+	struct nv50_i2c_port *port = (void *)base;
+	if (state) port->state |= 0x02;
+	else	   port->state &= 0xfd;
+	nv_wr32(priv, port->addr, port->state);
+}
+
+int
+nv50_i2c_sense_scl(struct nouveau_i2c_port *base)
+{
+	struct nv50_i2c_priv *priv = (void *)nv_object(base)->engine;
+	struct nv50_i2c_port *port = (void *)base;
+	return !!(nv_rd32(priv, port->addr) & 0x00000001);
+}
+
+int
+nv50_i2c_sense_sda(struct nouveau_i2c_port *base)
+{
+	struct nv50_i2c_priv *priv = (void *)nv_object(base)->engine;
+	struct nv50_i2c_port *port = (void *)base;
+	return !!(nv_rd32(priv, port->addr) & 0x00000002);
+}
+
+static const struct nouveau_i2c_func
+nv50_i2c_func = {
+	.drive_scl = nv50_i2c_drive_scl,
+	.drive_sda = nv50_i2c_drive_sda,
+	.sense_scl = nv50_i2c_sense_scl,
+	.sense_sda = nv50_i2c_sense_sda,
+};
+
+const u32 nv50_i2c_addr[] = {
+	0x00e138, 0x00e150, 0x00e168, 0x00e180,
+	0x00e254, 0x00e274, 0x00e764, 0x00e780,
+	0x00e79c, 0x00e7b8
+};
+const int nv50_i2c_addr_nr = ARRAY_SIZE(nv50_i2c_addr);
+
+static int
+nv50_i2c_port_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+		   struct nouveau_oclass *oclass, void *data, u32 index,
+		   struct nouveau_object **pobject)
+{
+	struct dcb_i2c_entry *info = data;
+	struct nv50_i2c_port *port;
+	int ret;
+
+	ret = nouveau_i2c_port_create(parent, engine, oclass, index,
+				     &nouveau_i2c_bit_algo, &port);
+	*pobject = nv_object(port);
+	if (ret)
+		return ret;
+
+	if (info->drive >= nv50_i2c_addr_nr)
+		return -EINVAL;
+
+	port->base.func = &nv50_i2c_func;
+	port->state = 0x00000007;
+	port->addr = nv50_i2c_addr[info->drive];
+	return 0;
+}
+
+int
+nv50_i2c_port_init(struct nouveau_object *object)
+{
+	struct nv50_i2c_priv *priv = (void *)object->engine;
+	struct nv50_i2c_port *port = (void *)object;
+	nv_wr32(priv, port->addr, port->state);
+	return nouveau_i2c_port_init(&port->base);
+}
+
+static struct nouveau_oclass
+nv50_i2c_sclass[] = {
+	{ .handle = NV_I2C_TYPE_DCBI2C(DCB_I2C_NVIO_BIT),
+	  .ofuncs = &(struct nouveau_ofuncs) {
+		  .ctor = nv50_i2c_port_ctor,
+		  .dtor = _nouveau_i2c_port_dtor,
+		  .init = nv50_i2c_port_init,
+		  .fini = _nouveau_i2c_port_fini,
+	  },
+	},
+	{}
+};
+
+static int
+nv50_i2c_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+	      struct nouveau_oclass *oclass, void *data, u32 size,
+	      struct nouveau_object **pobject)
+{
+	struct nv50_i2c_priv *priv;
+	int ret;
+
+	ret = nouveau_i2c_create(parent, engine, oclass, nv50_i2c_sclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+struct nouveau_oclass
+nv50_i2c_oclass = {
+	.handle = NV_SUBDEV(I2C, 0x50),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv50_i2c_ctor,
+		.dtor = _nouveau_i2c_dtor,
+		.init = _nouveau_i2c_init,
+		.fini = _nouveau_i2c_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/nv50.h b/drivers/gpu/drm/nouveau/core/subdev/i2c/nv50.h
new file mode 100644
index 0000000..4e5ba48
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/nv50.h
@@ -0,0 +1,32 @@
+#ifndef __NV50_I2C_H__
+#define __NV50_I2C_H__
+
+#include <subdev/i2c.h>
+
+struct nv50_i2c_priv {
+	struct nouveau_i2c base;
+};
+
+struct nv50_i2c_port {
+	struct nouveau_i2c_port base;
+	u32 addr;
+	u32 ctrl;
+	u32 data;
+	u32 state;
+};
+
+extern const u32 nv50_i2c_addr[];
+extern const int nv50_i2c_addr_nr;
+int  nv50_i2c_port_init(struct nouveau_object *);
+int  nv50_i2c_sense_scl(struct nouveau_i2c_port *);
+int  nv50_i2c_sense_sda(struct nouveau_i2c_port *);
+void nv50_i2c_drive_scl(struct nouveau_i2c_port *, int state);
+void nv50_i2c_drive_sda(struct nouveau_i2c_port *, int state);
+
+int  nv94_aux_port_ctor(struct nouveau_object *, struct nouveau_object *,
+			struct nouveau_oclass *, void *, u32,
+			struct nouveau_object **);
+void nv94_i2c_acquire(struct nouveau_i2c_port *);
+void nv94_i2c_release(struct nouveau_i2c_port *);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/nv94.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/nv94.c
index 4c498532..61b7716 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/i2c/nv94.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/nv94.c
@@ -1,5 +1,5 @@
 /*
- * Copyright 2009 Red Hat Inc.
+ * Copyright 2012 Red Hat Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -22,11 +22,8 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/i2c.h>
+#include "nv50.h"
 
-/******************************************************************************
- * aux channel util functions
- *****************************************************************************/
 #define AUX_DBG(fmt, args...) nv_debug(aux, "AUXCH(%d): " fmt, ch, ##args)
 #define AUX_ERR(fmt, args...) nv_error(aux, "AUXCH(%d): " fmt, ch, ##args)
 
@@ -72,12 +69,13 @@ auxch_init(struct nouveau_i2c *aux, int ch)
 }
 
 int
-nv94_aux(struct nouveau_i2c_port *port, u8 type, u32 addr, u8 *data, u8 size)
+nv94_aux(struct nouveau_i2c_port *base, u8 type, u32 addr, u8 *data, u8 size)
 {
-	struct nouveau_i2c *aux = port->i2c;
+	struct nouveau_i2c *aux = nouveau_i2c(base);
+	struct nv50_i2c_port *port = (void *)base;
 	u32 ctrl, stat, timeout, retries;
 	u32 xbuf[4] = {};
-	int ch = port->drive;
+	int ch = port->addr;
 	int ret, i;
 
 	AUX_DBG("%d: 0x%08x %d\n", type, addr, size);
@@ -153,13 +151,135 @@ out:
 }
 
 void
-nv94_aux_mux(struct nouveau_i2c_port *port)
+nv94_i2c_acquire(struct nouveau_i2c_port *base)
 {
-	if (port->dcb & 0x00000100) {
-		u32 reg = 0x00e500 + (port->drive * 0x50);
-		/* nfi, but neither auxch or i2c work if it's 1 */
-		nv_mask(port->i2c, reg + 0x0c, 0x00000001, 0x00000000);
-		/* nfi, but switches auxch vs normal i2c */
-		nv_mask(port->i2c, reg + 0x00, 0x0000f003, 0x00002002);
+	struct nv50_i2c_priv *priv = (void *)nv_object(base)->engine;
+	struct nv50_i2c_port *port = (void *)base;
+	if (port->ctrl) {
+		nv_mask(priv, port->ctrl + 0x0c, 0x00000001, 0x00000000);
+		nv_mask(priv, port->ctrl + 0x00, 0x0000f003, port->data);
 	}
 }
+
+void
+nv94_i2c_release(struct nouveau_i2c_port *base)
+{
+}
+
+static const struct nouveau_i2c_func
+nv94_i2c_func = {
+	.acquire   = nv94_i2c_acquire,
+	.release   = nv94_i2c_release,
+	.drive_scl = nv50_i2c_drive_scl,
+	.drive_sda = nv50_i2c_drive_sda,
+	.sense_scl = nv50_i2c_sense_scl,
+	.sense_sda = nv50_i2c_sense_sda,
+};
+
+static int
+nv94_i2c_port_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+		   struct nouveau_oclass *oclass, void *data, u32 index,
+		   struct nouveau_object **pobject)
+{
+	struct dcb_i2c_entry *info = data;
+	struct nv50_i2c_port *port;
+	int ret;
+
+	ret = nouveau_i2c_port_create(parent, engine, oclass, index,
+				     &nouveau_i2c_bit_algo, &port);
+	*pobject = nv_object(port);
+	if (ret)
+		return ret;
+
+	if (info->drive >= nv50_i2c_addr_nr)
+		return -EINVAL;
+
+	port->base.func = &nv94_i2c_func;
+	port->state = 7;
+	port->addr = nv50_i2c_addr[info->drive];
+	if (info->share != DCB_I2C_UNUSED) {
+		port->ctrl = 0x00e500 + (info->share * 0x50);
+		port->data = 0x0000e001;
+	}
+	return 0;
+}
+
+static const struct nouveau_i2c_func
+nv94_aux_func = {
+	.acquire   = nv94_i2c_acquire,
+	.release   = nv94_i2c_release,
+	.aux       = nv94_aux,
+};
+
+int
+nv94_aux_port_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+		   struct nouveau_oclass *oclass, void *data, u32 index,
+		   struct nouveau_object **pobject)
+{
+	struct dcb_i2c_entry *info = data;
+	struct nv50_i2c_port *port;
+	int ret;
+
+	ret = nouveau_i2c_port_create(parent, engine, oclass, index,
+				     &nouveau_i2c_aux_algo, &port);
+	*pobject = nv_object(port);
+	if (ret)
+		return ret;
+
+	port->base.func = &nv94_aux_func;
+	port->addr = info->drive;
+	if (info->share != DCB_I2C_UNUSED) {
+		port->ctrl = 0x00e500 + (info->drive * 0x50);
+		port->data = 0x00002002;
+	}
+
+	return 0;
+}
+
+static struct nouveau_oclass
+nv94_i2c_sclass[] = {
+	{ .handle = NV_I2C_TYPE_DCBI2C(DCB_I2C_NVIO_BIT),
+	  .ofuncs = &(struct nouveau_ofuncs) {
+		  .ctor = nv94_i2c_port_ctor,
+		  .dtor = _nouveau_i2c_port_dtor,
+		  .init = nv50_i2c_port_init,
+		  .fini = _nouveau_i2c_port_fini,
+	  },
+	},
+	{ .handle = NV_I2C_TYPE_DCBI2C(DCB_I2C_NVIO_AUX),
+	  .ofuncs = &(struct nouveau_ofuncs) {
+		  .ctor = nv94_aux_port_ctor,
+		  .dtor = _nouveau_i2c_port_dtor,
+		  .init = _nouveau_i2c_port_init,
+		  .fini = _nouveau_i2c_port_fini,
+	  },
+	},
+	{}
+};
+
+static int
+nv94_i2c_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+	      struct nouveau_oclass *oclass, void *data, u32 size,
+	      struct nouveau_object **pobject)
+{
+	struct nv50_i2c_priv *priv;
+	int ret;
+
+	ret = nouveau_i2c_create(parent, engine, oclass, nv94_i2c_sclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+struct nouveau_oclass
+nv94_i2c_oclass = {
+	.handle = NV_SUBDEV(I2C, 0x94),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv94_i2c_ctor,
+		.dtor = _nouveau_i2c_dtor,
+		.init = _nouveau_i2c_init,
+		.fini = _nouveau_i2c_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/nvd0.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/nvd0.c
new file mode 100644
index 0000000..f761b8a
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/nvd0.c
@@ -0,0 +1,124 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "nv50.h"
+
+static int
+nvd0_i2c_sense_scl(struct nouveau_i2c_port *base)
+{
+	struct nv50_i2c_priv *priv = (void *)nv_object(base)->engine;
+	struct nv50_i2c_port *port = (void *)base;
+	return !!(nv_rd32(priv, port->addr) & 0x00000010);
+}
+
+static int
+nvd0_i2c_sense_sda(struct nouveau_i2c_port *base)
+{
+	struct nv50_i2c_priv *priv = (void *)nv_object(base)->engine;
+	struct nv50_i2c_port *port = (void *)base;
+	return !!(nv_rd32(priv, port->addr) & 0x00000020);
+}
+
+static const struct nouveau_i2c_func
+nvd0_i2c_func = {
+	.acquire   = nv94_i2c_acquire,
+	.release   = nv94_i2c_release,
+	.drive_scl = nv50_i2c_drive_scl,
+	.drive_sda = nv50_i2c_drive_sda,
+	.sense_scl = nvd0_i2c_sense_scl,
+	.sense_sda = nvd0_i2c_sense_sda,
+};
+
+static int
+nvd0_i2c_port_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+		   struct nouveau_oclass *oclass, void *data, u32 index,
+		   struct nouveau_object **pobject)
+{
+	struct dcb_i2c_entry *info = data;
+	struct nv50_i2c_port *port;
+	int ret;
+
+	ret = nouveau_i2c_port_create(parent, engine, oclass, index,
+				     &nouveau_i2c_bit_algo, &port);
+	*pobject = nv_object(port);
+	if (ret)
+		return ret;
+
+	port->base.func = &nvd0_i2c_func;
+	port->state = 0x00000007;
+	port->addr = 0x00d014 + (info->drive * 0x20);
+	if (info->share != DCB_I2C_UNUSED) {
+		port->ctrl = 0x00e500 + (info->share * 0x50);
+		port->data = 0x0000e001;
+	}
+	return 0;
+}
+
+static struct nouveau_oclass
+nvd0_i2c_sclass[] = {
+	{ .handle = NV_I2C_TYPE_DCBI2C(DCB_I2C_NVIO_BIT),
+	  .ofuncs = &(struct nouveau_ofuncs) {
+		  .ctor = nvd0_i2c_port_ctor,
+		  .dtor = _nouveau_i2c_port_dtor,
+		  .init = nv50_i2c_port_init,
+		  .fini = _nouveau_i2c_port_fini,
+	  },
+	},
+	{ .handle = NV_I2C_TYPE_DCBI2C(DCB_I2C_NVIO_AUX),
+	  .ofuncs = &(struct nouveau_ofuncs) {
+		  .ctor = nv94_aux_port_ctor,
+		  .dtor = _nouveau_i2c_port_dtor,
+		  .init = _nouveau_i2c_port_init,
+		  .fini = _nouveau_i2c_port_fini,
+	  },
+	},
+	{}
+};
+
+static int
+nvd0_i2c_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+	      struct nouveau_oclass *oclass, void *data, u32 size,
+	      struct nouveau_object **pobject)
+{
+	struct nv50_i2c_priv *priv;
+	int ret;
+
+	ret = nouveau_i2c_create(parent, engine, oclass, nvd0_i2c_sclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+struct nouveau_oclass
+nvd0_i2c_oclass = {
+	.handle = NV_SUBDEV(I2C, 0xd0),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvd0_i2c_ctor,
+		.dtor = _nouveau_i2c_dtor,
+		.init = _nouveau_i2c_init,
+		.fini = _nouveau_i2c_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/ic.c b/drivers/gpu/drm/nouveau/core/subdev/therm/ic.c
index 95a63e1..e24090b 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/ic.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/ic.c
@@ -31,7 +31,7 @@ static bool
 probe_monitoring_device(struct nouveau_i2c_port *i2c,
 			struct i2c_board_info *info)
 {
-	struct nouveau_therm_priv *priv = (void *)nouveau_therm(i2c->i2c);
+	struct nouveau_therm_priv *priv = (void *)nouveau_therm(i2c);
 	struct i2c_client *client;
 
 	request_module("%s%s", I2C_MODULE_PREFIX, info->type);
-- 
cgit v0.10.2


From 5cc027f6b1ec651c18a4322ed3e30c6e9cf01e96 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Mon, 18 Feb 2013 17:50:51 -0500
Subject: drm/nv50-/disp: handle supervisor tasks from workqueue

i2c_algo_bit sleeps...

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
index c3806122d..d22f656 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
@@ -940,7 +940,6 @@ nv50_disp_intr_unk10(struct nv50_disp_priv *priv, u32 super)
 			exec_script(priv, head, 1);
 	}
 
-	nv_wr32(priv, 0x610024, 0x00000010);
 	nv_wr32(priv, 0x610030, 0x80000000);
 }
 
@@ -1097,7 +1096,6 @@ nv50_disp_intr_unk20(struct nv50_disp_priv *priv, u32 super)
 		}
 	}
 
-	nv_wr32(priv, 0x610024, 0x00000020);
 	nv_wr32(priv, 0x610030, 0x80000000);
 }
 
@@ -1136,22 +1134,23 @@ nv50_disp_intr_unk40(struct nv50_disp_priv *priv, u32 super)
 		}
 	}
 
-	nv_wr32(priv, 0x610024, 0x00000040);
 	nv_wr32(priv, 0x610030, 0x80000000);
 }
 
-static void
-nv50_disp_intr_super(struct nv50_disp_priv *priv, u32 intr1)
+void
+nv50_disp_intr_supervisor(struct work_struct *work)
 {
+	struct nv50_disp_priv *priv =
+		container_of(work, struct nv50_disp_priv, supervisor);
 	u32 super = nv_rd32(priv, 0x610030);
 
-	nv_debug(priv, "supervisor 0x%08x 0x%08x\n", intr1, super);
+	nv_debug(priv, "supervisor 0x%08x 0x%08x\n", priv->super, super);
 
-	if (intr1 & 0x00000010)
+	if (priv->super & 0x00000010)
 		nv50_disp_intr_unk10(priv, super);
-	if (intr1 & 0x00000020)
+	if (priv->super & 0x00000020)
 		nv50_disp_intr_unk20(priv, super);
-	if (intr1 & 0x00000040)
+	if (priv->super & 0x00000040)
 		nv50_disp_intr_unk40(priv, super);
 }
 
@@ -1180,7 +1179,9 @@ nv50_disp_intr(struct nouveau_subdev *subdev)
 	}
 
 	if (intr1 & 0x00000070) {
-		nv50_disp_intr_super(priv, intr1);
+		priv->super = (intr1 & 0x00000070);
+		schedule_work(&priv->supervisor);
+		nv_wr32(priv, 0x610024, priv->super);
 		intr1 &= ~0x00000070;
 	}
 }
@@ -1202,6 +1203,7 @@ nv50_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_engine(priv)->sclass = nv50_disp_base_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nv50_disp_intr;
+	INIT_WORK(&priv->supervisor, nv50_disp_intr_supervisor);
 	priv->sclass = nv50_disp_sclass;
 	priv->head.nr = 2;
 	priv->dac.nr = 3;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.h b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.h
index fc89718..b88a621 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.h
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.h
@@ -15,6 +15,10 @@ struct dcb_output;
 struct nv50_disp_priv {
 	struct nouveau_disp base;
 	struct nouveau_oclass *sclass;
+
+	struct work_struct supervisor;
+	u32 super;
+
 	struct {
 		int nr;
 	} head;
@@ -126,6 +130,7 @@ extern struct nouveau_ofuncs nv50_disp_oimm_ofuncs;
 extern struct nouveau_ofuncs nv50_disp_curs_ofuncs;
 extern struct nouveau_ofuncs nv50_disp_base_ofuncs;
 extern struct nouveau_oclass nv50_disp_cclass;
+void nv50_disp_intr_supervisor(struct work_struct *);
 void nv50_disp_intr(struct nouveau_subdev *);
 
 extern struct nouveau_omthds nv84_disp_base_omthds[];
@@ -139,6 +144,7 @@ extern struct nouveau_ofuncs nvd0_disp_oimm_ofuncs;
 extern struct nouveau_ofuncs nvd0_disp_curs_ofuncs;
 extern struct nouveau_ofuncs nvd0_disp_base_ofuncs;
 extern struct nouveau_oclass nvd0_disp_cclass;
+void nvd0_disp_intr_supervisor(struct work_struct *);
 void nvd0_disp_intr(struct nouveau_subdev *);
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv84.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv84.c
index a215342..9ec942a 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv84.c
@@ -72,6 +72,7 @@ nv84_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_engine(priv)->sclass = nv84_disp_base_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nv50_disp_intr;
+	INIT_WORK(&priv->supervisor, nv50_disp_intr_supervisor);
 	priv->sclass = nv84_disp_sclass;
 	priv->head.nr = 2;
 	priv->dac.nr = 3;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv94.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv94.c
index a315e28..a838bdd 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv94.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv94.c
@@ -78,6 +78,7 @@ nv94_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_engine(priv)->sclass = nv94_disp_base_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nv50_disp_intr;
+	INIT_WORK(&priv->supervisor, nv50_disp_intr_supervisor);
 	priv->sclass = nv94_disp_sclass;
 	priv->head.nr = 2;
 	priv->dac.nr = 3;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nva0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nva0.c
index 480e2de..ce539ca 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nva0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nva0.c
@@ -62,6 +62,7 @@ nva0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_engine(priv)->sclass = nva0_disp_base_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nv50_disp_intr;
+	INIT_WORK(&priv->supervisor, nv50_disp_intr_supervisor);
 	priv->sclass = nva0_disp_sclass;
 	priv->head.nr = 2;
 	priv->dac.nr = 3;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nva3.c b/drivers/gpu/drm/nouveau/core/engine/disp/nva3.c
index 718b4f6..bd2bb47 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nva3.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nva3.c
@@ -79,6 +79,7 @@ nva3_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_engine(priv)->sclass = nva3_disp_base_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nv50_disp_intr;
+	INIT_WORK(&priv->supervisor, nv50_disp_intr_supervisor);
 	priv->sclass = nva3_disp_sclass;
 	priv->head.nr = 2;
 	priv->dac.nr = 3;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
index 77cc730..1c91eb1 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
@@ -836,6 +836,26 @@ nvd0_display_unk4_handler(struct nv50_disp_priv *priv, u32 head, u32 mask)
 }
 
 void
+nvd0_disp_intr_supervisor(struct work_struct *work)
+{
+	struct nv50_disp_priv *priv =
+		container_of(work, struct nv50_disp_priv, supervisor);
+	u32 mask = 0, head = ~0;
+
+	while (!mask && ++head < priv->head.nr)
+		mask = nv_rd32(priv, 0x6101d4 + (head * 0x800));
+
+	nv_debug(priv, "supervisor %08x %08x %d\n", priv->super, mask, head);
+
+	if (priv->super & 0x00000001)
+		nvd0_display_unk1_handler(priv, head, mask);
+	if (priv->super & 0x00000002)
+		nvd0_display_unk2_handler(priv, head, mask);
+	if (priv->super & 0x00000004)
+		nvd0_display_unk4_handler(priv, head, mask);
+}
+
+void
 nvd0_disp_intr(struct nouveau_subdev *subdev)
 {
 	struct nv50_disp_priv *priv = (void *)subdev;
@@ -868,27 +888,11 @@ nvd0_disp_intr(struct nouveau_subdev *subdev)
 
 	if (intr & 0x00100000) {
 		u32 stat = nv_rd32(priv, 0x6100ac);
-		u32 mask = 0, crtc = ~0;
-
-		while (!mask && ++crtc < priv->head.nr)
-			mask = nv_rd32(priv, 0x6101d4 + (crtc * 0x800));
-
-		if (stat & 0x00000001) {
-			nv_wr32(priv, 0x6100ac, 0x00000001);
-			nvd0_display_unk1_handler(priv, crtc, mask);
-			stat &= ~0x00000001;
-		}
-
-		if (stat & 0x00000002) {
-			nv_wr32(priv, 0x6100ac, 0x00000002);
-			nvd0_display_unk2_handler(priv, crtc, mask);
-			stat &= ~0x00000002;
-		}
-
-		if (stat & 0x00000004) {
-			nv_wr32(priv, 0x6100ac, 0x00000004);
-			nvd0_display_unk4_handler(priv, crtc, mask);
-			stat &= ~0x00000004;
+		if (stat & 0x00000007) {
+			priv->super = (stat & 0x00000007);
+			schedule_work(&priv->supervisor);
+			nv_wr32(priv, 0x6100ac, priv->super);
+			stat &= ~0x00000007;
 		}
 
 		if (stat) {
@@ -929,6 +933,7 @@ nvd0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_engine(priv)->sclass = nvd0_disp_base_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nvd0_disp_intr;
+	INIT_WORK(&priv->supervisor, nvd0_disp_intr_supervisor);
 	priv->sclass = nvd0_disp_sclass;
 	priv->head.nr = heads;
 	priv->dac.nr = 3;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c
index 5512296..65c19ca 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c
@@ -63,6 +63,7 @@ nve0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_engine(priv)->sclass = nve0_disp_base_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nvd0_disp_intr;
+	INIT_WORK(&priv->supervisor, nvd0_disp_intr_supervisor);
 	priv->sclass = nve0_disp_sclass;
 	priv->head.nr = heads;
 	priv->dac.nr = 3;
diff --git a/drivers/gpu/drm/nouveau/nv50_display.c b/drivers/gpu/drm/nouveau/nv50_display.c
index 75d1161..2cd3797 100644
--- a/drivers/gpu/drm/nouveau/nv50_display.c
+++ b/drivers/gpu/drm/nouveau/nv50_display.c
@@ -424,7 +424,10 @@ evo_kick(u32 *push, void *evoc)
 static bool
 evo_sync_wait(void *data)
 {
-	return nouveau_bo_rd32(data, EVO_MAST_NTFY) != 0x00000000;
+	if (nouveau_bo_rd32(data, EVO_MAST_NTFY) != 0x00000000)
+		return true;
+	usleep_range(1, 2);
+	return false;
 }
 
 static int
-- 
cgit v0.10.2


From 0a0afd282fd715dd63d64b243299a64da14f8e8d Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Mon, 18 Feb 2013 23:17:53 -0500
Subject: drm/nv50-/disp: move DP link training to core and train from
 supervisor

We need to be able to do link training for PIOR-connected ANX9805 from
the third supervisor handler (due to script ordering in the bios, can't
have the "user" call train because some settings are overwritten from
the modesetting bios scripts).

This moves link training for SOR-connected DP encoders to the second
supervisor interrupt, *before* we call the modesetting scripts (yes,
different ordering from PIOR is necessary).  This is useful since we
should now be able to remove some hacks to workaround races between
the supervisor and link training paths.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index d8374ec..1452507 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -159,6 +159,7 @@ nouveau-y += core/engine/disp/nva3.o
 nouveau-y += core/engine/disp/nvd0.o
 nouveau-y += core/engine/disp/nve0.o
 nouveau-y += core/engine/disp/dacnv50.o
+nouveau-y += core/engine/disp/dport.o
 nouveau-y += core/engine/disp/hdanva3.o
 nouveau-y += core/engine/disp/hdanvd0.o
 nouveau-y += core/engine/disp/hdminv84.o
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/dport.c b/drivers/gpu/drm/nouveau/core/engine/disp/dport.c
new file mode 100644
index 0000000..fa27b02
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/dport.c
@@ -0,0 +1,346 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <subdev/bios.h>
+#include <subdev/bios/dcb.h>
+#include <subdev/bios/dp.h>
+#include <subdev/bios/init.h>
+#include <subdev/i2c.h>
+
+#include <engine/disp.h>
+
+#include "dport.h"
+
+#define DBG(fmt, args...) nv_debug(dp->disp, "DP:%04x:%04x: " fmt,             \
+				   dp->outp->hasht, dp->outp->hashm, ##args)
+#define ERR(fmt, args...) nv_error(dp->disp, "DP:%04x:%04x: " fmt,             \
+				   dp->outp->hasht, dp->outp->hashm, ##args)
+
+/******************************************************************************
+ * link training
+ *****************************************************************************/
+struct dp_state {
+	const struct nouveau_dp_func *func;
+	struct nouveau_disp *disp;
+	struct dcb_output *outp;
+	struct nvbios_dpout info;
+	u8 version;
+	struct nouveau_i2c_port *aux;
+	int head;
+	u8  dpcd[4];
+	int link_nr;
+	u32 link_bw;
+	u8  stat[6];
+	u8  conf[4];
+};
+
+static int
+dp_set_link_config(struct dp_state *dp)
+{
+	struct nouveau_disp *disp = dp->disp;
+	struct nouveau_bios *bios = nouveau_bios(disp);
+	struct nvbios_init init = {
+		.subdev = nv_subdev(dp->disp),
+		.bios = bios,
+		.offset = 0x0000,
+		.outp = dp->outp,
+		.crtc = dp->head,
+		.execute = 1,
+	};
+	u32 lnkcmp;
+	u8 sink[2];
+
+	DBG("%d lanes at %d KB/s\n", dp->link_nr, dp->link_bw);
+
+	/* set desired link configuration on the sink */
+	sink[0] = dp->link_bw / 27000;
+	sink[1] = dp->link_nr;
+	if (dp->dpcd[DPCD_RC02] & DPCD_RC02_ENHANCED_FRAME_CAP)
+		sink[1] |= DPCD_LC01_ENHANCED_FRAME_EN;
+
+	nv_wraux(dp->aux, DPCD_LC00, sink, 2);
+
+	/* set desired link configuration on the source */
+	if ((lnkcmp = dp->info.lnkcmp)) {
+		if (dp->version < 0x30) {
+			while ((dp->link_bw / 10) < nv_ro16(bios, lnkcmp))
+				lnkcmp += 4;
+			init.offset = nv_ro16(bios, lnkcmp + 2);
+		} else {
+			while ((dp->link_bw / 27000) < nv_ro08(bios, lnkcmp))
+				lnkcmp += 3;
+			init.offset = nv_ro16(bios, lnkcmp + 1);
+		}
+
+		nvbios_exec(&init);
+	}
+
+	return dp->func->lnk_ctl(dp->disp, dp->outp, dp->head,
+				 dp->link_nr, dp->link_bw / 27000,
+				 dp->dpcd[DPCD_RC02] &
+					  DPCD_RC02_ENHANCED_FRAME_CAP);
+}
+
+static void
+dp_set_training_pattern(struct dp_state *dp, u8 pattern)
+{
+	u8 sink_tp;
+
+	DBG("training pattern %d\n", pattern);
+	dp->func->pattern(dp->disp, dp->outp, dp->head, pattern);
+
+	nv_rdaux(dp->aux, DPCD_LC02, &sink_tp, 1);
+	sink_tp &= ~DPCD_LC02_TRAINING_PATTERN_SET;
+	sink_tp |= pattern;
+	nv_wraux(dp->aux, DPCD_LC02, &sink_tp, 1);
+}
+
+static int
+dp_link_train_commit(struct dp_state *dp)
+{
+	int i;
+
+	for (i = 0; i < dp->link_nr; i++) {
+		u8 lane = (dp->stat[4 + (i >> 1)] >> ((i & 1) * 4)) & 0xf;
+		u8 lpre = (lane & 0x0c) >> 2;
+		u8 lvsw = (lane & 0x03) >> 0;
+
+		dp->conf[i] = (lpre << 3) | lvsw;
+		if (lvsw == 3)
+			dp->conf[i] |= DPCD_LC03_MAX_SWING_REACHED;
+		if (lpre == 3)
+			dp->conf[i] |= DPCD_LC03_MAX_PRE_EMPHASIS_REACHED;
+
+		DBG("config lane %d %02x\n", i, dp->conf[i]);
+		dp->func->drv_ctl(dp->disp, dp->outp, dp->head, i, lvsw, lpre);
+	}
+
+	return nv_wraux(dp->aux, DPCD_LC03(0), dp->conf, 4);
+}
+
+static int
+dp_link_train_update(struct dp_state *dp, u32 delay)
+{
+	int ret;
+
+	udelay(delay);
+
+	ret = nv_rdaux(dp->aux, DPCD_LS02, dp->stat, 6);
+	if (ret)
+		return ret;
+
+	DBG("status %*ph\n", 6, dp->stat);
+	return 0;
+}
+
+static int
+dp_link_train_cr(struct dp_state *dp)
+{
+	bool cr_done = false, abort = false;
+	int voltage = dp->conf[0] & DPCD_LC03_VOLTAGE_SWING_SET;
+	int tries = 0, i;
+
+	dp_set_training_pattern(dp, 1);
+
+	do {
+		if (dp_link_train_commit(dp) ||
+		    dp_link_train_update(dp, 100))
+			break;
+
+		cr_done = true;
+		for (i = 0; i < dp->link_nr; i++) {
+			u8 lane = (dp->stat[i >> 1] >> ((i & 1) * 4)) & 0xf;
+			if (!(lane & DPCD_LS02_LANE0_CR_DONE)) {
+				cr_done = false;
+				if (dp->conf[i] & DPCD_LC03_MAX_SWING_REACHED)
+					abort = true;
+				break;
+			}
+		}
+
+		if ((dp->conf[0] & DPCD_LC03_VOLTAGE_SWING_SET) != voltage) {
+			voltage = dp->conf[0] & DPCD_LC03_VOLTAGE_SWING_SET;
+			tries = 0;
+		}
+	} while (!cr_done && !abort && ++tries < 5);
+
+	return cr_done ? 0 : -1;
+}
+
+static int
+dp_link_train_eq(struct dp_state *dp)
+{
+	bool eq_done, cr_done = true;
+	int tries = 0, i;
+
+	dp_set_training_pattern(dp, 2);
+
+	do {
+		if (dp_link_train_update(dp, 400))
+			break;
+
+		eq_done = !!(dp->stat[2] & DPCD_LS04_INTERLANE_ALIGN_DONE);
+		for (i = 0; i < dp->link_nr && eq_done; i++) {
+			u8 lane = (dp->stat[i >> 1] >> ((i & 1) * 4)) & 0xf;
+			if (!(lane & DPCD_LS02_LANE0_CR_DONE))
+				cr_done = false;
+			if (!(lane & DPCD_LS02_LANE0_CHANNEL_EQ_DONE) ||
+			    !(lane & DPCD_LS02_LANE0_SYMBOL_LOCKED))
+				eq_done = false;
+		}
+
+		if (dp_link_train_commit(dp))
+			break;
+	} while (!eq_done && cr_done && ++tries <= 5);
+
+	return eq_done ? 0 : -1;
+}
+
+static void
+dp_link_train_init(struct dp_state *dp, bool spread)
+{
+	struct nvbios_init init = {
+		.subdev = nv_subdev(dp->disp),
+		.bios = nouveau_bios(dp->disp),
+		.outp = dp->outp,
+		.crtc = dp->head,
+		.execute = 1,
+	};
+
+	/* set desired spread */
+	if (spread)
+		init.offset = dp->info.script[2];
+	else
+		init.offset = dp->info.script[3];
+	nvbios_exec(&init);
+
+	/* pre-train script */
+	init.offset = dp->info.script[0];
+	nvbios_exec(&init);
+}
+
+static void
+dp_link_train_fini(struct dp_state *dp)
+{
+	struct nvbios_init init = {
+		.subdev = nv_subdev(dp->disp),
+		.bios = nouveau_bios(dp->disp),
+		.outp = dp->outp,
+		.crtc = dp->head,
+		.execute = 1,
+	};
+
+	/* post-train script */
+	init.offset = dp->info.script[1],
+	nvbios_exec(&init);
+}
+
+int
+nouveau_dp_train(struct nouveau_disp *disp, const struct nouveau_dp_func *func,
+		 struct dcb_output *outp, int head, u32 datarate)
+{
+	struct nouveau_bios *bios = nouveau_bios(disp);
+	struct nouveau_i2c *i2c = nouveau_i2c(disp);
+	struct dp_state _dp = {
+		.disp = disp,
+		.func = func,
+		.outp = outp,
+		.head = head,
+	}, *dp = &_dp;
+	const u32 bw_list[] = { 270000, 162000, 0 };
+	const u32 *link_bw = bw_list;
+	u8  hdr, cnt, len;
+	u32 data;
+	int ret;
+
+	/* find the bios displayport data relevant to this output */
+	data = nvbios_dpout_match(bios, outp->hasht, outp->hashm, &dp->version,
+				 &hdr, &cnt, &len, &dp->info);
+	if (!data) {
+		ERR("bios data not found\n");
+		return -EINVAL;
+	}
+
+	/* acquire the aux channel and fetch some info about the display */
+	if (outp->location)
+		dp->aux = i2c->find_type(i2c, NV_I2C_TYPE_EXTAUX(outp->extdev));
+	else
+		dp->aux = i2c->find(i2c, NV_I2C_TYPE_DCBI2C(outp->i2c_index));
+	if (!dp->aux) {
+		ERR("no aux channel?!\n");
+		return -ENODEV;
+	}
+
+	ret = nv_rdaux(dp->aux, 0x00000, dp->dpcd, sizeof(dp->dpcd));
+	if (ret) {
+		ERR("failed to read DPCD\n");
+		return ret;
+	}
+
+	/* adjust required bandwidth for 8B/10B coding overhead */
+	datarate = (datarate / 8) * 10;
+
+	/* enable down-spreading and execute pre-train script from vbios */
+	dp_link_train_init(dp, dp->dpcd[3] & 0x01);
+
+	/* start off at highest link rate supported by encoder and display */
+	while (*link_bw > (dp->dpcd[1] * 27000))
+		link_bw++;
+
+	while (link_bw[0]) {
+		/* find minimum required lane count at this link rate */
+		dp->link_nr = dp->dpcd[2] & DPCD_RC02_MAX_LANE_COUNT;
+		while ((dp->link_nr >> 1) * link_bw[0] > datarate)
+			dp->link_nr >>= 1;
+
+		/* drop link rate to minimum with this lane count */
+		while ((link_bw[1] * dp->link_nr) > datarate)
+			link_bw++;
+		dp->link_bw = link_bw[0];
+
+		/* program selected link configuration */
+		ret = dp_set_link_config(dp);
+		if (ret == 0) {
+			/* attempt to train the link at this configuration */
+			memset(dp->stat, 0x00, sizeof(dp->stat));
+			if (!dp_link_train_cr(dp) &&
+			    !dp_link_train_eq(dp))
+				break;
+		} else
+		if (ret >= 1) {
+			/* dp_set_link_config() handled training */
+			break;
+		}
+
+		/* retry at lower rate */
+		link_bw++;
+	}
+
+	/* finish link training */
+	dp_set_training_pattern(dp, 0);
+
+	/* execute post-train script from vbios */
+	dp_link_train_fini(dp);
+	return true;
+}
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/dport.h b/drivers/gpu/drm/nouveau/core/engine/disp/dport.h
new file mode 100644
index 0000000..d0acd01
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/dport.h
@@ -0,0 +1,77 @@
+#ifndef __NVKM_DISP_DPORT_H__
+#define __NVKM_DISP_DPORT_H__
+
+/* DPCD Receiver Capabilities */
+#define DPCD_RC00                                                       0x00000
+#define DPCD_RC00_DPCD_REV                                                 0xff
+#define DPCD_RC01                                                       0x00001
+#define DPCD_RC01_MAX_LINK_RATE                                            0xff
+#define DPCD_RC02                                                       0x00002
+#define DPCD_RC02_ENHANCED_FRAME_CAP                                       0x80
+#define DPCD_RC02_MAX_LANE_COUNT                                           0x1f
+#define DPCD_RC03                                                       0x00003
+#define DPCD_RC03_MAX_DOWNSPREAD                                           0x01
+
+/* DPCD Link Configuration */
+#define DPCD_LC00                                                       0x00100
+#define DPCD_LC00_LINK_BW_SET                                              0xff
+#define DPCD_LC01                                                       0x00101
+#define DPCD_LC01_ENHANCED_FRAME_EN                                        0x80
+#define DPCD_LC01_LANE_COUNT_SET                                           0x1f
+#define DPCD_LC02                                                       0x00102
+#define DPCD_LC02_TRAINING_PATTERN_SET                                     0x03
+#define DPCD_LC03(l)                                            ((l) +  0x00103)
+#define DPCD_LC03_MAX_PRE_EMPHASIS_REACHED                                 0x20
+#define DPCD_LC03_PRE_EMPHASIS_SET                                         0x18
+#define DPCD_LC03_MAX_SWING_REACHED                                        0x04
+#define DPCD_LC03_VOLTAGE_SWING_SET                                        0x03
+
+/* DPCD Link/Sink Status */
+#define DPCD_LS02                                                       0x00202
+#define DPCD_LS02_LANE1_SYMBOL_LOCKED                                      0x40
+#define DPCD_LS02_LANE1_CHANNEL_EQ_DONE                                    0x20
+#define DPCD_LS02_LANE1_CR_DONE                                            0x10
+#define DPCD_LS02_LANE0_SYMBOL_LOCKED                                      0x04
+#define DPCD_LS02_LANE0_CHANNEL_EQ_DONE                                    0x02
+#define DPCD_LS02_LANE0_CR_DONE                                            0x01
+#define DPCD_LS03                                                       0x00203
+#define DPCD_LS03_LANE3_SYMBOL_LOCKED                                      0x40
+#define DPCD_LS03_LANE3_CHANNEL_EQ_DONE                                    0x20
+#define DPCD_LS03_LANE3_CR_DONE                                            0x10
+#define DPCD_LS03_LANE2_SYMBOL_LOCKED                                      0x04
+#define DPCD_LS03_LANE2_CHANNEL_EQ_DONE                                    0x02
+#define DPCD_LS03_LANE2_CR_DONE                                            0x01
+#define DPCD_LS04                                                       0x00204
+#define DPCD_LS04_LINK_STATUS_UPDATED                                      0x80
+#define DPCD_LS04_DOWNSTREAM_PORT_STATUS_CHANGED                           0x40
+#define DPCD_LS04_INTERLANE_ALIGN_DONE                                     0x01
+#define DPCD_LS06                                                       0x00206
+#define DPCD_LS06_LANE1_PRE_EMPHASIS                                       0xc0
+#define DPCD_LS06_LANE1_VOLTAGE_SWING                                      0x30
+#define DPCD_LS06_LANE0_PRE_EMPHASIS                                       0x0c
+#define DPCD_LS06_LANE0_VOLTAGE_SWING                                      0x03
+#define DPCD_LS07                                                       0x00207
+#define DPCD_LS07_LANE3_PRE_EMPHASIS                                       0xc0
+#define DPCD_LS07_LANE3_VOLTAGE_SWING                                      0x30
+#define DPCD_LS07_LANE2_PRE_EMPHASIS                                       0x0c
+#define DPCD_LS07_LANE2_VOLTAGE_SWING                                      0x03
+
+struct nouveau_disp;
+struct dcb_output;
+
+struct nouveau_dp_func {
+	int (*pattern)(struct nouveau_disp *, struct dcb_output *,
+		       int head, int pattern);
+	int (*lnk_ctl)(struct nouveau_disp *, struct dcb_output *, int head,
+		       int link_nr, int link_bw, bool enh_frame);
+	int (*drv_ctl)(struct nouveau_disp *, struct dcb_output *, int head,
+		       int lane, int swing, int preem);
+};
+
+extern const struct nouveau_dp_func nv94_sor_dp_func;
+extern const struct nouveau_dp_func nvd0_sor_dp_func;
+
+int nouveau_dp_train(struct nouveau_disp *, const struct nouveau_dp_func *,
+		     struct dcb_output *, int, u32);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
index d22f656..a4e2112 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
@@ -911,7 +911,7 @@ exec_clkcmp(struct nv50_disp_priv *priv, int head, int id, u32 pclk,
 	}
 
 	data = nvbios_ocfg_match(bios, data, conf, &ver, &hdr, &cnt, &len, &info2);
-	if (data) {
+	if (data && id < 0xff) {
 		data = nvbios_oclk_match(bios, info2.clkcmp[id], pclk);
 		if (data) {
 			struct nvbios_init init = {
@@ -1078,8 +1078,28 @@ nv50_disp_intr_unk20(struct nv50_disp_priv *priv, u32 super)
 	head = ffs((super & 0x00000180) >> 7) - 1;
 	if (head >= 0) {
 		u32 pclk = nv_rd32(priv, 0x610ad0 + (head * 0x540)) & 0x3fffff;
-		u32 conf = exec_clkcmp(priv, head, 0, pclk, &outp);
+		u32 conf = exec_clkcmp(priv, head, 0xff, pclk, &outp);
 		if (conf != ~0) {
+			if (outp.location == 0 && outp.type == DCB_OUTPUT_DP) {
+				u32 soff = (ffs(outp.or) - 1) * 0x08;
+				u32 ctrl = nv_rd32(priv, 0x610798 + soff);
+				u32 datarate;
+
+				switch ((ctrl & 0x000f0000) >> 16) {
+				case 6: datarate = pclk * 30 / 8; break;
+				case 5: datarate = pclk * 24 / 8; break;
+				case 2:
+				default:
+					datarate = pclk * 18 / 8;
+					break;
+				}
+
+				nouveau_dp_train(&priv->base, priv->sor.dp,
+						 &outp, head, datarate);
+			}
+
+			exec_clkcmp(priv, head, 0, pclk, &outp);
+
 			if (outp.type == DCB_OUTPUT_ANALOG) {
 				addr = 0x614280 + (ffs(outp.or) - 1) * 0x800;
 				mask = 0xffffffff;
@@ -1128,10 +1148,9 @@ nv50_disp_intr_unk40(struct nv50_disp_priv *priv, u32 super)
 	if (head >= 0) {
 		struct dcb_output outp;
 		u32 pclk = nv_rd32(priv, 0x610ad0 + (head * 0x540)) & 0x3fffff;
-		if (exec_clkcmp(priv, head, 1, pclk, &outp) != ~0) {
-			if (outp.type == DCB_OUTPUT_TMDS)
+		if (exec_clkcmp(priv, head, 1, pclk, &outp) != ~0)
+			if (outp.location == 0 && outp.type == DCB_OUTPUT_TMDS)
 				nv50_disp_intr_unk40_tmds(priv, &outp);
-		}
 	}
 
 	nv_wr32(priv, 0x610030, 0x80000000);
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.h b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.h
index b88a621..781a816 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.h
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.h
@@ -10,7 +10,7 @@
 #include <engine/dmaobj.h>
 #include <engine/disp.h>
 
-struct dcb_output;
+#include "dport.h"
 
 struct nv50_disp_priv {
 	struct nouveau_disp base;
@@ -32,22 +32,8 @@ struct nv50_disp_priv {
 		int (*power)(struct nv50_disp_priv *, int sor, u32 data);
 		int (*hda_eld)(struct nv50_disp_priv *, int sor, u8 *, u32);
 		int (*hdmi)(struct nv50_disp_priv *, int head, int sor, u32);
-		int (*dp_train_init)(struct nv50_disp_priv *, int sor, int link,
-				     int head, u16 type, u16 mask, u32 data,
-				     struct dcb_output *);
-		int (*dp_train_fini)(struct nv50_disp_priv *, int sor, int link,
-				     int head, u16 type, u16 mask, u32 data,
-				     struct dcb_output *);
-		int (*dp_train)(struct nv50_disp_priv *, int sor, int link,
-				u16 type, u16 mask, u32 data,
-				struct dcb_output *);
-		int (*dp_lnkctl)(struct nv50_disp_priv *, int sor, int link,
-				 int head, u16 type, u16 mask, u32 data,
-				 struct dcb_output *);
-		int (*dp_drvctl)(struct nv50_disp_priv *, int sor, int link,
-				 int lane, u16 type, u16 mask, u32 data,
-				 struct dcb_output *);
 		u32 lvdsconf;
+		const struct nouveau_dp_func *dp;
 	} sor;
 };
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv94.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv94.c
index a838bdd..a449890 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv94.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv94.c
@@ -44,12 +44,6 @@ nv94_disp_base_omthds[] = {
 	{ SOR_MTHD(NV50_DISP_SOR_PWR)         , nv50_sor_mthd },
 	{ SOR_MTHD(NV84_DISP_SOR_HDMI_PWR)    , nv50_sor_mthd },
 	{ SOR_MTHD(NV50_DISP_SOR_LVDS_SCRIPT) , nv50_sor_mthd },
-	{ SOR_MTHD(NV94_DISP_SOR_DP_TRAIN)    , nv50_sor_mthd },
-	{ SOR_MTHD(NV94_DISP_SOR_DP_LNKCTL)   , nv50_sor_mthd },
-	{ SOR_MTHD(NV94_DISP_SOR_DP_DRVCTL(0)), nv50_sor_mthd },
-	{ SOR_MTHD(NV94_DISP_SOR_DP_DRVCTL(1)), nv50_sor_mthd },
-	{ SOR_MTHD(NV94_DISP_SOR_DP_DRVCTL(2)), nv50_sor_mthd },
-	{ SOR_MTHD(NV94_DISP_SOR_DP_DRVCTL(3)), nv50_sor_mthd },
 	{ DAC_MTHD(NV50_DISP_DAC_PWR)         , nv50_dac_mthd },
 	{ DAC_MTHD(NV50_DISP_DAC_LOAD)        , nv50_dac_mthd },
 	{},
@@ -87,11 +81,7 @@ nv94_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->dac.sense = nv50_dac_sense;
 	priv->sor.power = nv50_sor_power;
 	priv->sor.hdmi = nv84_hdmi_ctrl;
-	priv->sor.dp_train = nv94_sor_dp_train;
-	priv->sor.dp_train_init = nv94_sor_dp_train_init;
-	priv->sor.dp_train_fini = nv94_sor_dp_train_fini;
-	priv->sor.dp_lnkctl = nv94_sor_dp_lnkctl;
-	priv->sor.dp_drvctl = nv94_sor_dp_drvctl;
+	priv->sor.dp = &nv94_sor_dp_func;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nva3.c b/drivers/gpu/drm/nouveau/core/engine/disp/nva3.c
index bd2bb47..2f78c94 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nva3.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nva3.c
@@ -45,12 +45,6 @@ nva3_disp_base_omthds[] = {
 	{ SOR_MTHD(NVA3_DISP_SOR_HDA_ELD)     , nv50_sor_mthd },
 	{ SOR_MTHD(NV84_DISP_SOR_HDMI_PWR)    , nv50_sor_mthd },
 	{ SOR_MTHD(NV50_DISP_SOR_LVDS_SCRIPT) , nv50_sor_mthd },
-	{ SOR_MTHD(NV94_DISP_SOR_DP_TRAIN)    , nv50_sor_mthd },
-	{ SOR_MTHD(NV94_DISP_SOR_DP_LNKCTL)   , nv50_sor_mthd },
-	{ SOR_MTHD(NV94_DISP_SOR_DP_DRVCTL(0)), nv50_sor_mthd },
-	{ SOR_MTHD(NV94_DISP_SOR_DP_DRVCTL(1)), nv50_sor_mthd },
-	{ SOR_MTHD(NV94_DISP_SOR_DP_DRVCTL(2)), nv50_sor_mthd },
-	{ SOR_MTHD(NV94_DISP_SOR_DP_DRVCTL(3)), nv50_sor_mthd },
 	{ DAC_MTHD(NV50_DISP_DAC_PWR)         , nv50_dac_mthd },
 	{ DAC_MTHD(NV50_DISP_DAC_LOAD)        , nv50_dac_mthd },
 	{},
@@ -89,11 +83,7 @@ nva3_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->sor.power = nv50_sor_power;
 	priv->sor.hda_eld = nva3_hda_eld;
 	priv->sor.hdmi = nva3_hdmi_ctrl;
-	priv->sor.dp_train = nv94_sor_dp_train;
-	priv->sor.dp_train_init = nv94_sor_dp_train_init;
-	priv->sor.dp_train_fini = nv94_sor_dp_train_fini;
-	priv->sor.dp_lnkctl = nv94_sor_dp_lnkctl;
-	priv->sor.dp_drvctl = nv94_sor_dp_drvctl;
+	priv->sor.dp = &nv94_sor_dp_func;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
index 1c91eb1..7106a72 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
@@ -650,20 +650,19 @@ exec_script(struct nv50_disp_priv *priv, int head, int outp, u32 ctrl, int id)
 
 static u32
 exec_clkcmp(struct nv50_disp_priv *priv, int head, int outp,
-	    u32 ctrl, int id, u32 pclk)
+	    u32 ctrl, int id, u32 pclk, struct dcb_output *dcb)
 {
 	struct nouveau_bios *bios = nouveau_bios(priv);
 	struct nvbios_outp info1;
 	struct nvbios_ocfg info2;
-	struct dcb_output dcb;
 	u8  ver, hdr, cnt, len;
 	u32 data, conf = ~0;
 
-	data = exec_lookup(priv, head, outp, ctrl, &dcb, &ver, &hdr, &cnt, &len, &info1);
+	data = exec_lookup(priv, head, outp, ctrl, dcb, &ver, &hdr, &cnt, &len, &info1);
 	if (data == 0x0000)
 		return conf;
 
-	switch (dcb.type) {
+	switch (dcb->type) {
 	case DCB_OUTPUT_TMDS:
 		conf = (ctrl & 0x00000f00) >> 8;
 		if (pclk >= 165000)
@@ -682,14 +681,14 @@ exec_clkcmp(struct nv50_disp_priv *priv, int head, int outp,
 	}
 
 	data = nvbios_ocfg_match(bios, data, conf, &ver, &hdr, &cnt, &len, &info2);
-	if (data) {
+	if (data && id < 0xff) {
 		data = nvbios_oclk_match(bios, info2.clkcmp[id], pclk);
 		if (data) {
 			struct nvbios_init init = {
 				.subdev = nv_subdev(priv),
 				.bios = bios,
 				.offset = data,
-				.outp = &dcb,
+				.outp = dcb,
 				.crtc = head,
 				.execute = 1,
 			};
@@ -764,6 +763,7 @@ nvd0_display_unk2_calc_tu(struct nv50_disp_priv *priv, int head, int or)
 static void
 nvd0_display_unk2_handler(struct nv50_disp_priv *priv, u32 head, u32 mask)
 {
+	struct dcb_output outp;
 	u32 pclk;
 	int i;
 
@@ -785,9 +785,27 @@ nvd0_display_unk2_handler(struct nv50_disp_priv *priv, u32 head, u32 mask)
 	for (i = 0; mask && i < 8; i++) {
 		u32 mcp = nv_rd32(priv, 0x660180 + (i * 0x20));
 		if (mcp & (1 << head)) {
-			u32 cfg = exec_clkcmp(priv, head, i, mcp, 0, pclk);
+			u32 cfg = exec_clkcmp(priv, head, i, mcp, 0xff, pclk, &outp);
 			if (cfg != ~0) {
 				u32 addr, mask, data = 0x00000000;
+
+				if (outp.type == DCB_OUTPUT_DP) {
+					switch ((mcp & 0x000f0000) >> 16) {
+					case 6: pclk = pclk * 30 / 8; break;
+					case 5: pclk = pclk * 24 / 8; break;
+					case 2:
+					default:
+						pclk = pclk * 18 / 8;
+						break;
+					}
+
+					nouveau_dp_train(&priv->base,
+							  priv->sor.dp,
+							 &outp, head, pclk);
+				}
+
+				exec_clkcmp(priv, head, i, mcp, 0, pclk, &outp);
+
 				if (i < 4) {
 					addr = 0x612280 + ((i - 0) * 0x800);
 					mask = 0xffffffff;
@@ -820,6 +838,7 @@ nvd0_display_unk2_handler(struct nv50_disp_priv *priv, u32 head, u32 mask)
 static void
 nvd0_display_unk4_handler(struct nv50_disp_priv *priv, u32 head, u32 mask)
 {
+	struct dcb_output outp;
 	int pclk, i;
 
 	pclk = nv_rd32(priv, 0x660450 + (head * 0x300)) / 1000;
@@ -827,7 +846,7 @@ nvd0_display_unk4_handler(struct nv50_disp_priv *priv, u32 head, u32 mask)
 	for (i = 0; mask && i < 8; i++) {
 		u32 mcp = nv_rd32(priv, 0x660180 + (i * 0x20));
 		if (mcp & (1 << head))
-			exec_clkcmp(priv, head, i, mcp, 1, pclk);
+			exec_clkcmp(priv, head, i, mcp, 1, pclk, &outp);
 	}
 
 	nv_wr32(priv, 0x6101d4, 0x00000000);
@@ -943,11 +962,7 @@ nvd0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->sor.power = nv50_sor_power;
 	priv->sor.hda_eld = nvd0_hda_eld;
 	priv->sor.hdmi = nvd0_hdmi_ctrl;
-	priv->sor.dp_train = nvd0_sor_dp_train;
-	priv->sor.dp_train_init = nv94_sor_dp_train_init;
-	priv->sor.dp_train_fini = nv94_sor_dp_train_fini;
-	priv->sor.dp_lnkctl = nvd0_sor_dp_lnkctl;
-	priv->sor.dp_drvctl = nvd0_sor_dp_drvctl;
+	priv->sor.dp = &nvd0_sor_dp_func;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c
index 65c19ca..20725b3 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c
@@ -73,11 +73,7 @@ nve0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->sor.power = nv50_sor_power;
 	priv->sor.hda_eld = nvd0_hda_eld;
 	priv->sor.hdmi = nvd0_hdmi_ctrl;
-	priv->sor.dp_train = nvd0_sor_dp_train;
-	priv->sor.dp_train_init = nv94_sor_dp_train_init;
-	priv->sor.dp_train_fini = nv94_sor_dp_train_fini;
-	priv->sor.dp_lnkctl = nvd0_sor_dp_lnkctl;
-	priv->sor.dp_drvctl = nvd0_sor_dp_drvctl;
+	priv->sor.dp = &nvd0_sor_dp_func;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/sornv50.c b/drivers/gpu/drm/nouveau/core/engine/disp/sornv50.c
index 39b6b677..ab1e918 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/sornv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/sornv50.c
@@ -79,31 +79,6 @@ nv50_sor_mthd(struct nouveau_object *object, u32 mthd, void *args, u32 size)
 		priv->sor.lvdsconf = data & NV50_DISP_SOR_LVDS_SCRIPT_ID;
 		ret = 0;
 		break;
-	case NV94_DISP_SOR_DP_TRAIN:
-		switch (data & NV94_DISP_SOR_DP_TRAIN_OP) {
-		case NV94_DISP_SOR_DP_TRAIN_OP_PATTERN:
-			ret = priv->sor.dp_train(priv, or, link, type, mask, data, &outp);
-			break;
-		case NV94_DISP_SOR_DP_TRAIN_OP_INIT:
-			ret = priv->sor.dp_train_init(priv, or, link, head, type, mask, data, &outp);
-			break;
-		case NV94_DISP_SOR_DP_TRAIN_OP_FINI:
-			ret = priv->sor.dp_train_fini(priv, or, link, head, type, mask, data, &outp);
-			break;
-		default:
-			break;
-		}
-		break;
-	case NV94_DISP_SOR_DP_LNKCTL:
-		ret = priv->sor.dp_lnkctl(priv, or, link, head, type, mask, data, &outp);
-		break;
-	case NV94_DISP_SOR_DP_DRVCTL(0):
-	case NV94_DISP_SOR_DP_DRVCTL(1):
-	case NV94_DISP_SOR_DP_DRVCTL(2):
-	case NV94_DISP_SOR_DP_DRVCTL(3):
-		ret = priv->sor.dp_drvctl(priv, or, link, (mthd & 0xc0) >> 6,
-				          type, mask, data, &outp);
-		break;
 	default:
 		BUG_ON(1);
 	}
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/sornv94.c b/drivers/gpu/drm/nouveau/core/engine/disp/sornv94.c
index f6edd00..7ec4ee83 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/sornv94.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/sornv94.c
@@ -33,124 +33,53 @@
 #include "nv50.h"
 
 static inline u32
-nv94_sor_dp_lane_map(struct nv50_disp_priv *priv, u8 lane)
+nv94_sor_soff(struct dcb_output *outp)
 {
-	static const u8 nvaf[] = { 24, 16, 8, 0 }; /* thanks, apple.. */
-	static const u8 nv94[] = { 16, 8, 0, 24 };
-	if (nv_device(priv)->chipset == 0xaf)
-		return nvaf[lane];
-	return nv94[lane];
+	return (ffs(outp->or) - 1) * 0x800;
 }
 
-int
-nv94_sor_dp_train_init(struct nv50_disp_priv *priv, int or, int link, int head,
-		       u16 type, u16 mask, u32 data, struct dcb_output *dcbo)
+static inline u32
+nv94_sor_loff(struct dcb_output *outp)
 {
-	struct nouveau_bios *bios = nouveau_bios(priv);
-	struct nvbios_dpout info;
-	u8  ver, hdr, cnt, len;
-	u16 outp;
-
-	outp = nvbios_dpout_match(bios, type, mask, &ver, &hdr, &cnt, &len, &info);
-	if (outp) {
-		struct nvbios_init init = {
-			.subdev = nv_subdev(priv),
-			.bios = bios,
-			.outp = dcbo,
-			.crtc = head,
-			.execute = 1,
-		};
-
-		if (data & NV94_DISP_SOR_DP_TRAIN_INIT_SPREAD_ON)
-			init.offset = info.script[2];
-		else
-			init.offset = info.script[3];
-		nvbios_exec(&init);
-
-		init.offset = info.script[0];
-		nvbios_exec(&init);
-	}
-
-	return 0;
+	return nv94_sor_soff(outp) + !(outp->sorconf.link & 1) * 0x80;
 }
 
-int
-nv94_sor_dp_train_fini(struct nv50_disp_priv *priv, int or, int link, int head,
-		       u16 type, u16 mask, u32 data, struct dcb_output *dcbo)
+static inline u32
+nv94_sor_dp_lane_map(struct nv50_disp_priv *priv, u8 lane)
 {
-	struct nouveau_bios *bios = nouveau_bios(priv);
-	struct nvbios_dpout info;
-	u8  ver, hdr, cnt, len;
-	u16 outp;
-
-	outp = nvbios_dpout_match(bios, type, mask, &ver, &hdr, &cnt, &len, &info);
-	if (outp) {
-		struct nvbios_init init = {
-			.subdev = nv_subdev(priv),
-			.bios = bios,
-			.offset = info.script[1],
-			.outp = dcbo,
-			.crtc = head,
-			.execute = 1,
-		};
-
-		nvbios_exec(&init);
-	}
-
-	return 0;
+	static const u8 nvaf[] = { 24, 16, 8, 0 }; /* thanks, apple.. */
+	static const u8 nv94[] = { 16, 8, 0, 24 };
+	if (nv_device(priv)->chipset == 0xaf)
+		return nvaf[lane];
+	return nv94[lane];
 }
 
-int
-nv94_sor_dp_train(struct nv50_disp_priv *priv, int or, int link,
-		  u16 type, u16 mask, u32 data, struct dcb_output *info)
+static int
+nv94_sor_dp_pattern(struct nouveau_disp *disp, struct dcb_output *outp,
+		    int head, int pattern)
 {
-	const u32 loff = (or * 0x800) + (link * 0x80);
-	const u32 patt = (data & NV94_DISP_SOR_DP_TRAIN_PATTERN);
-	nv_mask(priv, 0x61c10c + loff, 0x0f000000, patt << 24);
+	struct nv50_disp_priv *priv = (void *)disp;
+	const u32 loff = nv94_sor_loff(outp);
+	nv_mask(priv, 0x61c10c + loff, 0x0f000000, pattern << 24);
 	return 0;
 }
 
-int
-nv94_sor_dp_lnkctl(struct nv50_disp_priv *priv, int or, int link, int head,
-		   u16 type, u16 mask, u32 data, struct dcb_output *dcbo)
+static int
+nv94_sor_dp_lnk_ctl(struct nouveau_disp *disp, struct dcb_output *outp,
+		    int head, int link_nr, int link_bw, bool enh_frame)
 {
-	struct nouveau_bios *bios = nouveau_bios(priv);
-	const u32 loff = (or * 0x800) + (link * 0x80);
-	const u32 soff = (or * 0x800);
-	u16 link_bw = (data & NV94_DISP_SOR_DP_LNKCTL_WIDTH) >> 8;
-	u8  link_nr = (data & NV94_DISP_SOR_DP_LNKCTL_COUNT);
+	struct nv50_disp_priv *priv = (void *)disp;
+	const u32 soff = nv94_sor_soff(outp);
+	const u32 loff = nv94_sor_loff(outp);
 	u32 dpctrl = 0x00000000;
 	u32 clksor = 0x00000000;
-	u32 outp, lane = 0;
-	u8  ver, hdr, cnt, len;
-	struct nvbios_dpout info;
+	u32 lane = 0;
 	int i;
 
-	/* -> 10Khz units */
-	link_bw *= 2700;
-
-	outp = nvbios_dpout_match(bios, type, mask, &ver, &hdr, &cnt, &len, &info);
-	if (outp && info.lnkcmp) {
-		struct nvbios_init init = {
-			.subdev = nv_subdev(priv),
-			.bios = bios,
-			.offset = 0x0000,
-			.outp = dcbo,
-			.crtc = head,
-			.execute = 1,
-		};
-
-		while (link_bw < nv_ro16(bios, info.lnkcmp))
-			info.lnkcmp += 4;
-		init.offset = nv_ro16(bios, info.lnkcmp + 2);
-
-		nvbios_exec(&init);
-	}
-
 	dpctrl |= ((1 << link_nr) - 1) << 16;
-	if (data & NV94_DISP_SOR_DP_LNKCTL_FRAME_ENH)
+	if (enh_frame)
 		dpctrl |= 0x00004000;
-	if (link_bw > 16200)
+	if (link_bw > 0x06)
 		clksor |= 0x00040000;
 
 	for (i = 0; i < link_nr; i++)
@@ -162,24 +91,25 @@ nv94_sor_dp_lnkctl(struct nv50_disp_priv *priv, int or, int link, int head,
 	return 0;
 }
 
-int
-nv94_sor_dp_drvctl(struct nv50_disp_priv *priv, int or, int link, int lane,
-		   u16 type, u16 mask, u32 data, struct dcb_output *dcbo)
+static int
+nv94_sor_dp_drv_ctl(struct nouveau_disp *disp, struct dcb_output *outp,
+		    int head, int lane, int swing, int preem)
 {
-	struct nouveau_bios *bios = nouveau_bios(priv);
-	const u32 loff = (or * 0x800) + (link * 0x80);
-	const u8 swing = (data & NV94_DISP_SOR_DP_DRVCTL_VS) >> 8;
-	const u8 preem = (data & NV94_DISP_SOR_DP_DRVCTL_PE);
+	struct nouveau_bios *bios = nouveau_bios(disp);
+	struct nv50_disp_priv *priv = (void *)disp;
+	const u32 loff = nv94_sor_loff(outp);
 	u32 addr, shift = nv94_sor_dp_lane_map(priv, lane);
 	u8  ver, hdr, cnt, len;
-	struct nvbios_dpout outp;
+	struct nvbios_dpout info;
 	struct nvbios_dpcfg ocfg;
 
-	addr = nvbios_dpout_match(bios, type, mask, &ver, &hdr, &cnt, &len, &outp);
+	addr = nvbios_dpout_match(bios, outp->hasht, outp->hashm,
+				 &ver, &hdr, &cnt, &len, &info);
 	if (!addr)
 		return -ENODEV;
 
-	addr = nvbios_dpcfg_match(bios, addr, 0, swing, preem, &ver, &hdr, &cnt, &len, &ocfg);
+	addr = nvbios_dpcfg_match(bios, addr, 0, swing, preem,
+				 &ver, &hdr, &cnt, &len, &ocfg);
 	if (!addr)
 		return -EINVAL;
 
@@ -188,3 +118,10 @@ nv94_sor_dp_drvctl(struct nv50_disp_priv *priv, int or, int link, int lane,
 	nv_mask(priv, 0x61c130 + loff, 0x0000ff00, ocfg.unk << 8);
 	return 0;
 }
+
+const struct nouveau_dp_func
+nv94_sor_dp_func = {
+	.pattern = nv94_sor_dp_pattern,
+	.lnk_ctl = nv94_sor_dp_lnk_ctl,
+	.drv_ctl = nv94_sor_dp_drv_ctl,
+};
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/sornvd0.c b/drivers/gpu/drm/nouveau/core/engine/disp/sornvd0.c
index c37ce7e..9e1d435 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/sornvd0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/sornvd0.c
@@ -33,59 +33,49 @@
 #include "nv50.h"
 
 static inline u32
+nvd0_sor_soff(struct dcb_output *outp)
+{
+	return (ffs(outp->or) - 1) * 0x800;
+}
+
+static inline u32
+nvd0_sor_loff(struct dcb_output *outp)
+{
+	return nvd0_sor_soff(outp) + !(outp->sorconf.link & 1) * 0x80;
+}
+
+static inline u32
 nvd0_sor_dp_lane_map(struct nv50_disp_priv *priv, u8 lane)
 {
 	static const u8 nvd0[] = { 16, 8, 0, 24 };
 	return nvd0[lane];
 }
 
-int
-nvd0_sor_dp_train(struct nv50_disp_priv *priv, int or, int link,
-		  u16 type, u16 mask, u32 data, struct dcb_output *info)
+static int
+nvd0_sor_dp_pattern(struct nouveau_disp *disp, struct dcb_output *outp,
+		    int head, int pattern)
 {
-	const u32 loff = (or * 0x800) + (link * 0x80);
-	const u32 patt = (data & NV94_DISP_SOR_DP_TRAIN_PATTERN);
-	nv_mask(priv, 0x61c110 + loff, 0x0f0f0f0f, 0x01010101 * patt);
+	struct nv50_disp_priv *priv = (void *)disp;
+	const u32 loff = nvd0_sor_loff(outp);
+	nv_mask(priv, 0x61c110 + loff, 0x0f0f0f0f, 0x01010101 * pattern);
 	return 0;
 }
 
-int
-nvd0_sor_dp_lnkctl(struct nv50_disp_priv *priv, int or, int link, int head,
-		   u16 type, u16 mask, u32 data, struct dcb_output *dcbo)
+static int
+nvd0_sor_dp_lnk_ctl(struct nouveau_disp *disp, struct dcb_output *outp,
+		    int head, int link_nr, int link_bw, bool enh_frame)
 {
-	struct nouveau_bios *bios = nouveau_bios(priv);
-	const u32 loff = (or * 0x800) + (link * 0x80);
-	const u32 soff = (or * 0x800);
-	const u8  link_bw = (data & NV94_DISP_SOR_DP_LNKCTL_WIDTH) >> 8;
-	const u8  link_nr = (data & NV94_DISP_SOR_DP_LNKCTL_COUNT);
+	struct nv50_disp_priv *priv = (void *)disp;
+	const u32 soff = nvd0_sor_soff(outp);
+	const u32 loff = nvd0_sor_loff(outp);
 	u32 dpctrl = 0x00000000;
 	u32 clksor = 0x00000000;
-	u32 outp, lane = 0;
-	u8  ver, hdr, cnt, len;
-	struct nvbios_dpout info;
+	u32 lane = 0;
 	int i;
 
-	outp = nvbios_dpout_match(bios, type, mask, &ver, &hdr, &cnt, &len, &info);
-	if (outp && info.lnkcmp) {
-		struct nvbios_init init = {
-			.subdev = nv_subdev(priv),
-			.bios = bios,
-			.offset = 0x0000,
-			.outp = dcbo,
-			.crtc = head,
-			.execute = 1,
-		};
-
-		while (nv_ro08(bios, info.lnkcmp) < link_bw)
-			info.lnkcmp += 3;
-		init.offset = nv_ro16(bios, info.lnkcmp + 1);
-
-		nvbios_exec(&init);
-	}
-
 	clksor |= link_bw << 18;
 	dpctrl |= ((1 << link_nr) - 1) << 16;
-	if (data & NV94_DISP_SOR_DP_LNKCTL_FRAME_ENH)
+	if (enh_frame)
 		dpctrl |= 0x00004000;
 
 	for (i = 0; i < link_nr; i++)
@@ -97,24 +87,25 @@ nvd0_sor_dp_lnkctl(struct nv50_disp_priv *priv, int or, int link, int head,
 	return 0;
 }
 
-int
-nvd0_sor_dp_drvctl(struct nv50_disp_priv *priv, int or, int link, int lane,
-		   u16 type, u16 mask, u32 data, struct dcb_output *dcbo)
+static int
+nvd0_sor_dp_drv_ctl(struct nouveau_disp *disp, struct dcb_output *outp,
+		    int head, int lane, int swing, int preem)
 {
-	struct nouveau_bios *bios = nouveau_bios(priv);
-	const u32 loff = (or * 0x800) + (link * 0x80);
-	const u8 swing = (data & NV94_DISP_SOR_DP_DRVCTL_VS) >> 8;
-	const u8 preem = (data & NV94_DISP_SOR_DP_DRVCTL_PE);
+	struct nouveau_bios *bios = nouveau_bios(disp);
+	struct nv50_disp_priv *priv = (void *)disp;
+	const u32 loff = nvd0_sor_loff(outp);
 	u32 addr, shift = nvd0_sor_dp_lane_map(priv, lane);
 	u8  ver, hdr, cnt, len;
-	struct nvbios_dpout outp;
+	struct nvbios_dpout info;
 	struct nvbios_dpcfg ocfg;
 
-	addr = nvbios_dpout_match(bios, type, mask, &ver, &hdr, &cnt, &len, &outp);
+	addr = nvbios_dpout_match(bios, outp->hasht, outp->hashm,
+				 &ver, &hdr, &cnt, &len, &info);
 	if (!addr)
 		return -ENODEV;
 
-	addr = nvbios_dpcfg_match(bios, addr, 0, swing, preem, &ver, &hdr, &cnt, &len, &ocfg);
+	addr = nvbios_dpcfg_match(bios, addr, 0, swing, preem,
+				 &ver, &hdr, &cnt, &len, &ocfg);
 	if (!addr)
 		return -EINVAL;
 
@@ -124,3 +115,10 @@ nvd0_sor_dp_drvctl(struct nv50_disp_priv *priv, int or, int link, int lane,
 	nv_mask(priv, 0x61c13c + loff, 0x00000000, 0x00000000);
 	return 0;
 }
+
+const struct nouveau_dp_func
+nvd0_sor_dp_func = {
+	.pattern = nvd0_sor_dp_pattern,
+	.lnk_ctl = nvd0_sor_dp_lnk_ctl,
+	.drv_ctl = nvd0_sor_dp_drv_ctl,
+};
diff --git a/drivers/gpu/drm/nouveau/core/include/core/class.h b/drivers/gpu/drm/nouveau/core/include/core/class.h
index 994a280..86515cc 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/class.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/class.h
@@ -198,25 +198,6 @@ struct nv04_display_class {
 #define NV84_DISP_SOR_HDMI_PWR_REKEY                                 0x0000007f
 #define NV50_DISP_SOR_LVDS_SCRIPT                                    0x00013000
 #define NV50_DISP_SOR_LVDS_SCRIPT_ID                                 0x0000ffff
-#define NV94_DISP_SOR_DP_TRAIN                                       0x00016000
-#define NV94_DISP_SOR_DP_TRAIN_OP                                    0xf0000000
-#define NV94_DISP_SOR_DP_TRAIN_OP_PATTERN                            0x00000000
-#define NV94_DISP_SOR_DP_TRAIN_OP_INIT                               0x10000000
-#define NV94_DISP_SOR_DP_TRAIN_OP_FINI                               0x20000000
-#define NV94_DISP_SOR_DP_TRAIN_INIT_SPREAD                           0x00000001
-#define NV94_DISP_SOR_DP_TRAIN_INIT_SPREAD_OFF                       0x00000000
-#define NV94_DISP_SOR_DP_TRAIN_INIT_SPREAD_ON                        0x00000001
-#define NV94_DISP_SOR_DP_TRAIN_PATTERN                               0x00000003
-#define NV94_DISP_SOR_DP_TRAIN_PATTERN_DISABLED                      0x00000000
-#define NV94_DISP_SOR_DP_LNKCTL                                      0x00016040
-#define NV94_DISP_SOR_DP_LNKCTL_FRAME                                0x80000000
-#define NV94_DISP_SOR_DP_LNKCTL_FRAME_STD                            0x00000000
-#define NV94_DISP_SOR_DP_LNKCTL_FRAME_ENH                            0x80000000
-#define NV94_DISP_SOR_DP_LNKCTL_WIDTH                                0x00001f00
-#define NV94_DISP_SOR_DP_LNKCTL_COUNT                                0x00000007
-#define NV94_DISP_SOR_DP_DRVCTL(l)                     ((l) * 0x40 + 0x00016100)
-#define NV94_DISP_SOR_DP_DRVCTL_VS                                   0x00000300
-#define NV94_DISP_SOR_DP_DRVCTL_PE                                   0x00000003
 
 #define NV50_DISP_DAC_MTHD                                           0x00020000
 #define NV50_DISP_DAC_MTHD_TYPE                                      0x0000f000
diff --git a/drivers/gpu/drm/nouveau/nouveau_dp.c b/drivers/gpu/drm/nouveau/nouveau_dp.c
index d4da57f..36fd225 100644
--- a/drivers/gpu/drm/nouveau/nouveau_dp.c
+++ b/drivers/gpu/drm/nouveau/nouveau_dp.c
@@ -35,299 +35,6 @@
 #include <subdev/gpio.h>
 #include <subdev/i2c.h>
 
-/******************************************************************************
- * link training
- *****************************************************************************/
-struct dp_state {
-	struct nouveau_i2c_port *auxch;
-	struct nouveau_object *core;
-	struct dcb_output *dcb;
-	int crtc;
-	u8 *dpcd;
-	int link_nr;
-	u32 link_bw;
-	u8  stat[6];
-	u8  conf[4];
-};
-
-static void
-dp_set_link_config(struct drm_device *dev, struct dp_state *dp)
-{
-	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct dcb_output *dcb = dp->dcb;
-	const u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1);
-	const u32 moff = (dp->crtc << 3) | (link << 2) | or;
-	u8 sink[2];
-	u32 data;
-
-	NV_DEBUG(drm, "%d lanes at %d KB/s\n", dp->link_nr, dp->link_bw);
-
-	/* set desired link configuration on the source */
-	data = ((dp->link_bw / 27000) << 8) | dp->link_nr;
-	if (dp->dpcd[2] & DP_ENHANCED_FRAME_CAP)
-		data |= NV94_DISP_SOR_DP_LNKCTL_FRAME_ENH;
-
-	nv_call(dp->core, NV94_DISP_SOR_DP_LNKCTL + moff, data);
-
-	/* inform the sink of the new configuration */
-	sink[0] = dp->link_bw / 27000;
-	sink[1] = dp->link_nr;
-	if (dp->dpcd[2] & DP_ENHANCED_FRAME_CAP)
-		sink[1] |= DP_LANE_COUNT_ENHANCED_FRAME_EN;
-
-	nv_wraux(dp->auxch, DP_LINK_BW_SET, sink, 2);
-}
-
-static void
-dp_set_training_pattern(struct drm_device *dev, struct dp_state *dp, u8 pattern)
-{
-	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct dcb_output *dcb = dp->dcb;
-	const u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1);
-	const u32 moff = (dp->crtc << 3) | (link << 2) | or;
-	u8 sink_tp;
-
-	NV_DEBUG(drm, "training pattern %d\n", pattern);
-
-	nv_call(dp->core, NV94_DISP_SOR_DP_TRAIN + moff, pattern);
-
-	nv_rdaux(dp->auxch, DP_TRAINING_PATTERN_SET, &sink_tp, 1);
-	sink_tp &= ~DP_TRAINING_PATTERN_MASK;
-	sink_tp |= pattern;
-	nv_wraux(dp->auxch, DP_TRAINING_PATTERN_SET, &sink_tp, 1);
-}
-
-static int
-dp_link_train_commit(struct drm_device *dev, struct dp_state *dp)
-{
-	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct dcb_output *dcb = dp->dcb;
-	const u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1);
-	const u32 moff = (dp->crtc << 3) | (link << 2) | or;
-	int i;
-
-	for (i = 0; i < dp->link_nr; i++) {
-		u8 lane = (dp->stat[4 + (i >> 1)] >> ((i & 1) * 4)) & 0xf;
-		u8 lpre = (lane & 0x0c) >> 2;
-		u8 lvsw = (lane & 0x03) >> 0;
-
-		dp->conf[i] = (lpre << 3) | lvsw;
-		if (lvsw == DP_TRAIN_VOLTAGE_SWING_1200)
-			dp->conf[i] |= DP_TRAIN_MAX_SWING_REACHED;
-		if ((lpre << 3) == DP_TRAIN_PRE_EMPHASIS_9_5)
-			dp->conf[i] |= DP_TRAIN_MAX_PRE_EMPHASIS_REACHED;
-
-		NV_DEBUG(drm, "config lane %d %02x\n", i, dp->conf[i]);
-
-		nv_call(dp->core, NV94_DISP_SOR_DP_DRVCTL(i) + moff, (lvsw << 8) | lpre);
-	}
-
-	return nv_wraux(dp->auxch, DP_TRAINING_LANE0_SET, dp->conf, 4);
-}
-
-static int
-dp_link_train_update(struct drm_device *dev, struct dp_state *dp, u32 delay)
-{
-	struct nouveau_drm *drm = nouveau_drm(dev);
-	int ret;
-
-	udelay(delay);
-
-	ret = nv_rdaux(dp->auxch, DP_LANE0_1_STATUS, dp->stat, 6);
-	if (ret)
-		return ret;
-
-	NV_DEBUG(drm, "status %*ph\n", 6, dp->stat);
-	return 0;
-}
-
-static int
-dp_link_train_cr(struct drm_device *dev, struct dp_state *dp)
-{
-	bool cr_done = false, abort = false;
-	int voltage = dp->conf[0] & DP_TRAIN_VOLTAGE_SWING_MASK;
-	int tries = 0, i;
-
-	dp_set_training_pattern(dev, dp, DP_TRAINING_PATTERN_1);
-
-	do {
-		if (dp_link_train_commit(dev, dp) ||
-		    dp_link_train_update(dev, dp, 100))
-			break;
-
-		cr_done = true;
-		for (i = 0; i < dp->link_nr; i++) {
-			u8 lane = (dp->stat[i >> 1] >> ((i & 1) * 4)) & 0xf;
-			if (!(lane & DP_LANE_CR_DONE)) {
-				cr_done = false;
-				if (dp->conf[i] & DP_TRAIN_MAX_SWING_REACHED)
-					abort = true;
-				break;
-			}
-		}
-
-		if ((dp->conf[0] & DP_TRAIN_VOLTAGE_SWING_MASK) != voltage) {
-			voltage = dp->conf[0] & DP_TRAIN_VOLTAGE_SWING_MASK;
-			tries = 0;
-		}
-	} while (!cr_done && !abort && ++tries < 5);
-
-	return cr_done ? 0 : -1;
-}
-
-static int
-dp_link_train_eq(struct drm_device *dev, struct dp_state *dp)
-{
-	bool eq_done, cr_done = true;
-	int tries = 0, i;
-
-	dp_set_training_pattern(dev, dp, DP_TRAINING_PATTERN_2);
-
-	do {
-		if (dp_link_train_update(dev, dp, 400))
-			break;
-
-		eq_done = !!(dp->stat[2] & DP_INTERLANE_ALIGN_DONE);
-		for (i = 0; i < dp->link_nr && eq_done; i++) {
-			u8 lane = (dp->stat[i >> 1] >> ((i & 1) * 4)) & 0xf;
-			if (!(lane & DP_LANE_CR_DONE))
-				cr_done = false;
-			if (!(lane & DP_LANE_CHANNEL_EQ_DONE) ||
-			    !(lane & DP_LANE_SYMBOL_LOCKED))
-				eq_done = false;
-		}
-
-		if (dp_link_train_commit(dev, dp))
-			break;
-	} while (!eq_done && cr_done && ++tries <= 5);
-
-	return eq_done ? 0 : -1;
-}
-
-static void
-dp_link_train_init(struct drm_device *dev, struct dp_state *dp, bool spread)
-{
-	struct dcb_output *dcb = dp->dcb;
-	const u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1);
-	const u32 moff = (dp->crtc << 3) | (link << 2) | or;
-
-	nv_call(dp->core, NV94_DISP_SOR_DP_TRAIN + moff, (spread ?
-			  NV94_DISP_SOR_DP_TRAIN_INIT_SPREAD_ON :
-			  NV94_DISP_SOR_DP_TRAIN_INIT_SPREAD_OFF) |
-			  NV94_DISP_SOR_DP_TRAIN_OP_INIT);
-}
-
-static void
-dp_link_train_fini(struct drm_device *dev, struct dp_state *dp)
-{
-	struct dcb_output *dcb = dp->dcb;
-	const u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1);
-	const u32 moff = (dp->crtc << 3) | (link << 2) | or;
-
-	nv_call(dp->core, NV94_DISP_SOR_DP_TRAIN + moff,
-			  NV94_DISP_SOR_DP_TRAIN_OP_FINI);
-}
-
-static bool
-nouveau_dp_link_train(struct drm_encoder *encoder, u32 datarate,
-		      struct nouveau_object *core)
-{
-	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(encoder->crtc);
-	struct nouveau_connector *nv_connector =
-		nouveau_encoder_connector_get(nv_encoder);
-	struct drm_device *dev = encoder->dev;
-	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_gpio *gpio = nouveau_gpio(drm->device);
-	const u32 bw_list[] = { 270000, 162000, 0 };
-	const u32 *link_bw = bw_list;
-	struct dp_state dp;
-
-	dp.auxch = nv_encoder->i2c;
-	if (!dp.auxch)
-		return false;
-
-	dp.core = core;
-	dp.dcb = nv_encoder->dcb;
-	dp.crtc = nv_crtc->index;
-	dp.dpcd = nv_encoder->dp.dpcd;
-
-	/* adjust required bandwidth for 8B/10B coding overhead */
-	datarate = (datarate / 8) * 10;
-
-	/* some sinks toggle hotplug in response to some of the actions
-	 * we take during link training (DP_SET_POWER is one), we need
-	 * to ignore them for the moment to avoid races.
-	 */
-	nouveau_event_put(gpio->events, nv_connector->hpd.line,
-			 &nv_connector->hpd_func);
-
-	/* enable down-spreading and execute pre-train script from vbios */
-	dp_link_train_init(dev, &dp, nv_encoder->dp.dpcd[3] & 1);
-
-	/* start off at highest link rate supported by encoder and display */
-	while (*link_bw > nv_encoder->dp.link_bw)
-		link_bw++;
-
-	while (link_bw[0]) {
-		/* find minimum required lane count at this link rate */
-		dp.link_nr = nv_encoder->dp.link_nr;
-		while ((dp.link_nr >> 1) * link_bw[0] > datarate)
-			dp.link_nr >>= 1;
-
-		/* drop link rate to minimum with this lane count */
-		while ((link_bw[1] * dp.link_nr) > datarate)
-			link_bw++;
-		dp.link_bw = link_bw[0];
-
-		/* program selected link configuration */
-		dp_set_link_config(dev, &dp);
-
-		/* attempt to train the link at this configuration */
-		memset(dp.stat, 0x00, sizeof(dp.stat));
-		if (!dp_link_train_cr(dev, &dp) &&
-		    !dp_link_train_eq(dev, &dp))
-			break;
-
-		/* retry at lower rate */
-		link_bw++;
-	}
-
-	/* finish link training */
-	dp_set_training_pattern(dev, &dp, DP_TRAINING_PATTERN_DISABLE);
-
-	/* execute post-train script from vbios */
-	dp_link_train_fini(dev, &dp);
-
-	/* re-enable hotplug detect */
-	nouveau_event_get(gpio->events, nv_connector->hpd.line,
-			 &nv_connector->hpd_func);
-	return true;
-}
-
-void
-nouveau_dp_dpms(struct drm_encoder *encoder, int mode, u32 datarate,
-		struct nouveau_object *core)
-{
-	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct nouveau_i2c_port *auxch;
-	u8 status;
-
-	auxch = nv_encoder->i2c;
-	if (!auxch)
-		return;
-
-	if (mode == DRM_MODE_DPMS_ON)
-		status = DP_SET_POWER_D0;
-	else
-		status = DP_SET_POWER_D3;
-
-	nv_wraux(auxch, DP_SET_POWER, &status, 1);
-
-	if (mode == DRM_MODE_DPMS_ON)
-		nouveau_dp_link_train(encoder, datarate, core);
-}
-
 static void
 nouveau_dp_probe_oui(struct drm_device *dev, struct nouveau_i2c_port *auxch,
 		     u8 *dpcd)
diff --git a/drivers/gpu/drm/nouveau/nv50_display.c b/drivers/gpu/drm/nouveau/nv50_display.c
index 2cd3797..2c00a5f 100644
--- a/drivers/gpu/drm/nouveau/nv50_display.c
+++ b/drivers/gpu/drm/nouveau/nv50_display.c
@@ -1683,9 +1683,6 @@ nv50_sor_dpms(struct drm_encoder *encoder, int mode)
 	}
 
 	nv_call(disp->core, NV50_DISP_SOR_PWR + or, (mode == DRM_MODE_DPMS_ON));
-
-	if (nv_encoder->dcb->type == DCB_OUTPUT_DP)
-		nouveau_dp_dpms(encoder, mode, nv_encoder->dp.datarate, disp->core);
 }
 
 static bool
-- 
cgit v0.10.2


From 5a885f0b757ba4483d790c40813d8a66278bdda7 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Wed, 20 Feb 2013 14:34:18 +1000
Subject: drm/nv50-/kms: remove unnecessary wait-for-completion points

DP link training is now handled as part of the supervisor processing,
and can no longer race with it.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nv50_display.c b/drivers/gpu/drm/nouveau/nv50_display.c
index 2c00a5f..5c56575 100644
--- a/drivers/gpu/drm/nouveau/nv50_display.c
+++ b/drivers/gpu/drm/nouveau/nv50_display.c
@@ -1739,14 +1739,6 @@ nv50_sor_disconnect(struct drm_encoder *encoder)
 }
 
 static void
-nv50_sor_prepare(struct drm_encoder *encoder)
-{
-	nv50_sor_disconnect(encoder);
-	if (nouveau_encoder(encoder)->dcb->type == DCB_OUTPUT_DP)
-		evo_sync(encoder->dev);
-}
-
-static void
 nv50_sor_commit(struct drm_encoder *encoder)
 {
 }
@@ -1883,7 +1875,7 @@ nv50_sor_destroy(struct drm_encoder *encoder)
 static const struct drm_encoder_helper_funcs nv50_sor_hfunc = {
 	.dpms = nv50_sor_dpms,
 	.mode_fixup = nv50_sor_mode_fixup,
-	.prepare = nv50_sor_prepare,
+	.prepare = nv50_sor_disconnect,
 	.commit = nv50_sor_commit,
 	.mode_set = nv50_sor_mode_set,
 	.disable = nv50_sor_disconnect,
@@ -1946,7 +1938,7 @@ nv50_display_init(struct drm_device *dev)
 		evo_mthd(push, 0x0088, 1);
 		evo_data(push, NvEvoSync);
 		evo_kick(push, nv50_mast(dev));
-		return evo_sync(dev);
+		return 0;
 	}
 
 	return -EBUSY;
-- 
cgit v0.10.2


From a2bc283f3905389ba53962a2bbb05ede0c16193d Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Mon, 11 Feb 2013 09:11:08 +1000
Subject: drm/nv50-/disp: initial work towards supporting external encoders

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index 1452507..bda661e 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -165,6 +165,7 @@ nouveau-y += core/engine/disp/hdanvd0.o
 nouveau-y += core/engine/disp/hdminv84.o
 nouveau-y += core/engine/disp/hdminva3.o
 nouveau-y += core/engine/disp/hdminvd0.o
+nouveau-y += core/engine/disp/piornv50.o
 nouveau-y += core/engine/disp/sornv50.o
 nouveau-y += core/engine/disp/sornv94.o
 nouveau-y += core/engine/disp/sornvd0.o
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/dport.h b/drivers/gpu/drm/nouveau/core/engine/disp/dport.h
index d0acd01..0e1bbd1 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/dport.h
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/dport.h
@@ -70,6 +70,7 @@ struct nouveau_dp_func {
 
 extern const struct nouveau_dp_func nv94_sor_dp_func;
 extern const struct nouveau_dp_func nvd0_sor_dp_func;
+extern const struct nouveau_dp_func nv50_pior_dp_func;
 
 int nouveau_dp_train(struct nouveau_disp *, const struct nouveau_dp_func *,
 		     struct dcb_output *, int, u32);
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
index a4e2112..1298153 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
@@ -678,6 +678,9 @@ nv50_disp_base_omthds[] = {
 	{ SOR_MTHD(NV50_DISP_SOR_LVDS_SCRIPT) , nv50_sor_mthd },
 	{ DAC_MTHD(NV50_DISP_DAC_PWR)         , nv50_dac_mthd },
 	{ DAC_MTHD(NV50_DISP_DAC_LOAD)        , nv50_dac_mthd },
+	{ PIOR_MTHD(NV50_DISP_PIOR_PWR)       , nv50_pior_mthd },
+	{ PIOR_MTHD(NV50_DISP_PIOR_TMDS_PWR)  , nv50_pior_mthd },
+	{ PIOR_MTHD(NV50_DISP_PIOR_DP_PWR)    , nv50_pior_mthd },
 	{},
 };
 
@@ -1227,9 +1230,12 @@ nv50_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->head.nr = 2;
 	priv->dac.nr = 3;
 	priv->sor.nr = 2;
+	priv->pior.nr = 3;
 	priv->dac.power = nv50_dac_power;
 	priv->dac.sense = nv50_dac_sense;
 	priv->sor.power = nv50_sor_power;
+	priv->pior.power = nv50_pior_power;
+	priv->pior.dp = &nv50_pior_dp_func;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.h b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.h
index 781a816..1ae6ceb 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.h
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.h
@@ -35,6 +35,12 @@ struct nv50_disp_priv {
 		u32 lvdsconf;
 		const struct nouveau_dp_func *dp;
 	} sor;
+	struct {
+		int nr;
+		int (*power)(struct nv50_disp_priv *, int ext, u32 data);
+		u8 type[3];
+		const struct nouveau_dp_func *dp;
+	} pior;
 };
 
 #define DAC_MTHD(n) (n), (n) + 0x03
@@ -73,6 +79,11 @@ int nvd0_sor_dp_lnkctl(struct nv50_disp_priv *, int, int, int, u16, u16, u32,
 int nvd0_sor_dp_drvctl(struct nv50_disp_priv *, int, int, int, u16, u16, u32,
 		       struct dcb_output *);
 
+#define PIOR_MTHD(n) (n), (n) + 0x03
+
+int nv50_pior_mthd(struct nouveau_object *, u32, void *, u32);
+int nv50_pior_power(struct nv50_disp_priv *, int, u32);
+
 struct nv50_disp_base {
 	struct nouveau_parent base;
 	struct nouveau_ramht *ramht;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv84.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv84.c
index 9ec942a..d8c74c0 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv84.c
@@ -46,6 +46,9 @@ nv84_disp_base_omthds[] = {
 	{ SOR_MTHD(NV50_DISP_SOR_LVDS_SCRIPT) , nv50_sor_mthd },
 	{ DAC_MTHD(NV50_DISP_DAC_PWR)         , nv50_dac_mthd },
 	{ DAC_MTHD(NV50_DISP_DAC_LOAD)        , nv50_dac_mthd },
+	{ PIOR_MTHD(NV50_DISP_PIOR_PWR)       , nv50_pior_mthd },
+	{ PIOR_MTHD(NV50_DISP_PIOR_TMDS_PWR)  , nv50_pior_mthd },
+	{ PIOR_MTHD(NV50_DISP_PIOR_DP_PWR)    , nv50_pior_mthd },
 	{},
 };
 
@@ -77,10 +80,13 @@ nv84_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->head.nr = 2;
 	priv->dac.nr = 3;
 	priv->sor.nr = 2;
+	priv->pior.nr = 3;
 	priv->dac.power = nv50_dac_power;
 	priv->dac.sense = nv50_dac_sense;
 	priv->sor.power = nv50_sor_power;
 	priv->sor.hdmi = nv84_hdmi_ctrl;
+	priv->pior.power = nv50_pior_power;
+	priv->pior.dp = &nv50_pior_dp_func;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv94.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv94.c
index a449890..a66f949 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv94.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv94.c
@@ -46,6 +46,9 @@ nv94_disp_base_omthds[] = {
 	{ SOR_MTHD(NV50_DISP_SOR_LVDS_SCRIPT) , nv50_sor_mthd },
 	{ DAC_MTHD(NV50_DISP_DAC_PWR)         , nv50_dac_mthd },
 	{ DAC_MTHD(NV50_DISP_DAC_LOAD)        , nv50_dac_mthd },
+	{ PIOR_MTHD(NV50_DISP_PIOR_PWR)       , nv50_pior_mthd },
+	{ PIOR_MTHD(NV50_DISP_PIOR_TMDS_PWR)  , nv50_pior_mthd },
+	{ PIOR_MTHD(NV50_DISP_PIOR_DP_PWR)    , nv50_pior_mthd },
 	{},
 };
 
@@ -77,11 +80,14 @@ nv94_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->head.nr = 2;
 	priv->dac.nr = 3;
 	priv->sor.nr = 4;
+	priv->pior.nr = 3;
 	priv->dac.power = nv50_dac_power;
 	priv->dac.sense = nv50_dac_sense;
 	priv->sor.power = nv50_sor_power;
 	priv->sor.hdmi = nv84_hdmi_ctrl;
 	priv->sor.dp = &nv94_sor_dp_func;
+	priv->pior.power = nv50_pior_power;
+	priv->pior.dp = &nv50_pior_dp_func;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nva0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nva0.c
index ce539ca..6cf8eef 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nva0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nva0.c
@@ -67,10 +67,13 @@ nva0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->head.nr = 2;
 	priv->dac.nr = 3;
 	priv->sor.nr = 2;
+	priv->pior.nr = 3;
 	priv->dac.power = nv50_dac_power;
 	priv->dac.sense = nv50_dac_sense;
 	priv->sor.power = nv50_sor_power;
 	priv->sor.hdmi = nv84_hdmi_ctrl;
+	priv->pior.power = nv50_pior_power;
+	priv->pior.dp = &nv50_pior_dp_func;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nva3.c b/drivers/gpu/drm/nouveau/core/engine/disp/nva3.c
index 2f78c94..b754131 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nva3.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nva3.c
@@ -47,6 +47,9 @@ nva3_disp_base_omthds[] = {
 	{ SOR_MTHD(NV50_DISP_SOR_LVDS_SCRIPT) , nv50_sor_mthd },
 	{ DAC_MTHD(NV50_DISP_DAC_PWR)         , nv50_dac_mthd },
 	{ DAC_MTHD(NV50_DISP_DAC_LOAD)        , nv50_dac_mthd },
+	{ PIOR_MTHD(NV50_DISP_PIOR_PWR)       , nv50_pior_mthd },
+	{ PIOR_MTHD(NV50_DISP_PIOR_TMDS_PWR)  , nv50_pior_mthd },
+	{ PIOR_MTHD(NV50_DISP_PIOR_DP_PWR)    , nv50_pior_mthd },
 	{},
 };
 
@@ -78,12 +81,15 @@ nva3_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->head.nr = 2;
 	priv->dac.nr = 3;
 	priv->sor.nr = 4;
+	priv->pior.nr = 3;
 	priv->dac.power = nv50_dac_power;
 	priv->dac.sense = nv50_dac_sense;
 	priv->sor.power = nv50_sor_power;
 	priv->sor.hda_eld = nva3_hda_eld;
 	priv->sor.hdmi = nva3_hdmi_ctrl;
 	priv->sor.dp = &nv94_sor_dp_func;
+	priv->pior.power = nv50_pior_power;
+	priv->pior.dp = &nv50_pior_dp_func;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/piornv50.c b/drivers/gpu/drm/nouveau/core/engine/disp/piornv50.c
new file mode 100644
index 0000000..2c8ce35
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/piornv50.c
@@ -0,0 +1,140 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <core/os.h>
+#include <core/class.h>
+
+#include <subdev/bios.h>
+#include <subdev/bios/dcb.h>
+#include <subdev/timer.h>
+#include <subdev/i2c.h>
+
+#include "nv50.h"
+
+/******************************************************************************
+ * DisplayPort
+ *****************************************************************************/
+static struct nouveau_i2c_port *
+nv50_pior_dp_find(struct nouveau_disp *disp, struct dcb_output *outp)
+{
+	struct nouveau_i2c *i2c = nouveau_i2c(disp);
+	return i2c->find_type(i2c, NV_I2C_TYPE_EXTAUX(outp->extdev));
+}
+
+static int
+nv50_pior_dp_pattern(struct nouveau_disp *disp, struct dcb_output *outp,
+		     int head, int pattern)
+{
+	struct nouveau_i2c_port *port;
+	int ret = -EINVAL;
+
+	port = nv50_pior_dp_find(disp, outp);
+	if (port) {
+		if (port->func->pattern)
+			ret = port->func->pattern(port, pattern);
+		else
+			ret = 0;
+	}
+
+	return ret;
+}
+
+static int
+nv50_pior_dp_lnk_ctl(struct nouveau_disp *disp, struct dcb_output *outp,
+		     int head, int lane_nr, int link_bw, bool enh)
+{
+	struct nouveau_i2c_port *port;
+	int ret = -EINVAL;
+
+	port = nv50_pior_dp_find(disp, outp);
+	if (port && port->func->lnk_ctl)
+		ret = port->func->lnk_ctl(port, lane_nr, link_bw, enh);
+
+	return ret;
+}
+
+static int
+nv50_pior_dp_drv_ctl(struct nouveau_disp *disp, struct dcb_output *outp,
+		     int head, int lane, int vsw, int pre)
+{
+	struct nouveau_i2c_port *port;
+	int ret = -EINVAL;
+
+	port = nv50_pior_dp_find(disp, outp);
+	if (port) {
+		if (port->func->drv_ctl)
+			ret = port->func->drv_ctl(port, lane, vsw, pre);
+		else
+			ret = 0;
+	}
+
+	return ret;
+}
+
+const struct nouveau_dp_func
+nv50_pior_dp_func = {
+	.pattern = nv50_pior_dp_pattern,
+	.lnk_ctl = nv50_pior_dp_lnk_ctl,
+	.drv_ctl = nv50_pior_dp_drv_ctl,
+};
+
+/******************************************************************************
+ * General PIOR handling
+ *****************************************************************************/
+int
+nv50_pior_power(struct nv50_disp_priv *priv, int or, u32 data)
+{
+	const u32 stat = data & NV50_DISP_PIOR_PWR_STATE;
+	const u32 soff = (or * 0x800);
+	nv_wait(priv, 0x61e004 + soff, 0x80000000, 0x00000000);
+	nv_mask(priv, 0x61e004 + soff, 0x80000101, 0x80000000 | stat);
+	nv_wait(priv, 0x61e004 + soff, 0x80000000, 0x00000000);
+	return 0;
+}
+
+int
+nv50_pior_mthd(struct nouveau_object *object, u32 mthd, void *args, u32 size)
+{
+	struct nv50_disp_priv *priv = (void *)object->engine;
+	const u8 type = (mthd & NV50_DISP_PIOR_MTHD_TYPE) >> 12;
+	const u8 or   = (mthd & NV50_DISP_PIOR_MTHD_OR);
+	u32 *data = args;
+	int ret;
+
+	if (size < sizeof(u32))
+		return -EINVAL;
+
+	mthd &= ~NV50_DISP_PIOR_MTHD_TYPE;
+	mthd &= ~NV50_DISP_PIOR_MTHD_OR;
+	switch (mthd) {
+	case NV50_DISP_PIOR_PWR:
+		ret = priv->pior.power(priv, or, data[0]);
+		priv->pior.type[or] = type;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return ret;
+}
diff --git a/drivers/gpu/drm/nouveau/core/include/core/class.h b/drivers/gpu/drm/nouveau/core/include/core/class.h
index 86515cc..92d3ab1 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/class.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/class.h
@@ -219,6 +219,23 @@ struct nv04_display_class {
 #define NV50_DISP_DAC_LOAD                                           0x0002000c
 #define NV50_DISP_DAC_LOAD_VALUE                                     0x00000007
 
+#define NV50_DISP_PIOR_MTHD                                          0x00030000
+#define NV50_DISP_PIOR_MTHD_TYPE                                     0x0000f000
+#define NV50_DISP_PIOR_MTHD_OR                                       0x00000003
+
+#define NV50_DISP_PIOR_PWR                                           0x00030000
+#define NV50_DISP_PIOR_PWR_STATE                                     0x00000001
+#define NV50_DISP_PIOR_PWR_STATE_ON                                  0x00000001
+#define NV50_DISP_PIOR_PWR_STATE_OFF                                 0x00000000
+#define NV50_DISP_PIOR_TMDS_PWR                                      0x00032000
+#define NV50_DISP_PIOR_TMDS_PWR_STATE                                0x00000001
+#define NV50_DISP_PIOR_TMDS_PWR_STATE_ON                             0x00000001
+#define NV50_DISP_PIOR_TMDS_PWR_STATE_OFF                            0x00000000
+#define NV50_DISP_PIOR_DP_PWR                                        0x00036000
+#define NV50_DISP_PIOR_DP_PWR_STATE                                  0x00000001
+#define NV50_DISP_PIOR_DP_PWR_STATE_ON                               0x00000001
+#define NV50_DISP_PIOR_DP_PWR_STATE_OFF                              0x00000000
+
 struct nv50_display_class {
 };
 
-- 
cgit v0.10.2


From 476e84e126171d809f9c0b5d97137f5055f95ca8 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Mon, 11 Feb 2013 09:24:23 +1000
Subject: drm/nv50-/disp: initial supervisor support for off-chip encoders

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
index 1298153..314dda6 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
@@ -626,7 +626,7 @@ nv50_disp_base_init(struct nouveau_object *object)
 		nv_wr32(priv, 0x6101e0 + (i * 0x04), tmp);
 	}
 
-	/* ... EXT caps */
+	/* ... PIOR caps */
 	for (i = 0; i < 3; i++) {
 		tmp = nv_rd32(priv, 0x61e000 + (i * 0x800));
 		nv_wr32(priv, 0x6101f0 + (i * 0x04), tmp);
@@ -783,8 +783,8 @@ exec_lookup(struct nv50_disp_priv *priv, int head, int outp, u32 ctrl,
 	if (outp < 4) {
 		type = DCB_OUTPUT_ANALOG;
 		mask = 0;
-	} else {
-		outp -= 4;
+	} else
+	if (outp < 8) {
 		switch (ctrl & 0x00000f00) {
 		case 0x00000000: type = DCB_OUTPUT_LVDS; mask = 1; break;
 		case 0x00000100: type = DCB_OUTPUT_TMDS; mask = 1; break;
@@ -796,6 +796,17 @@ exec_lookup(struct nv50_disp_priv *priv, int head, int outp, u32 ctrl,
 			nv_error(priv, "unknown SOR mc 0x%08x\n", ctrl);
 			return 0x0000;
 		}
+		outp -= 4;
+	} else {
+		outp = outp - 8;
+		type = 0x0010;
+		mask = 0;
+		switch (ctrl & 0x00000f00) {
+		case 0x00000000: type |= priv->pior.type[outp]; break;
+		default:
+			nv_error(priv, "unknown PIOR mc 0x%08x\n", ctrl);
+			return 0x0000;
+		}
 	}
 
 	mask  = 0x00c0 & (mask << 6);
@@ -806,6 +817,10 @@ exec_lookup(struct nv50_disp_priv *priv, int head, int outp, u32 ctrl,
 	if (!data)
 		return 0x0000;
 
+	/* off-chip encoders require matching the exact encoder type */
+	if (dcb->location != 0)
+		type |= dcb->extdev << 8;
+
 	return nvbios_outp_match(bios, type, mask, ver, hdr, cnt, len, info);
 }
 
@@ -820,9 +835,11 @@ exec_script(struct nv50_disp_priv *priv, int head, int id)
 	u32 ctrl = 0x00000000;
 	int i;
 
+	/* DAC */
 	for (i = 0; !(ctrl & (1 << head)) && i < 3; i++)
 		ctrl = nv_rd32(priv, 0x610b5c + (i * 8));
 
+	/* SOR */
 	if (!(ctrl & (1 << head))) {
 		if (nv_device(priv)->chipset  < 0x90 ||
 		    nv_device(priv)->chipset == 0x92 ||
@@ -837,6 +854,13 @@ exec_script(struct nv50_disp_priv *priv, int head, int id)
 		}
 	}
 
+	/* PIOR */
+	if (!(ctrl & (1 << head))) {
+		for (i = 0; !(ctrl & (1 << head)) && i < 3; i++)
+			ctrl = nv_rd32(priv, 0x610b84 + (i * 8));
+		i += 8;
+	}
+
 	if (!(ctrl & (1 << head)))
 		return false;
 	i--;
@@ -870,9 +894,11 @@ exec_clkcmp(struct nv50_disp_priv *priv, int head, int id, u32 pclk,
 	u32 data, conf = ~0;
 	int i;
 
+	/* DAC */
 	for (i = 0; !(ctrl & (1 << head)) && i < 3; i++)
 		ctrl = nv_rd32(priv, 0x610b58 + (i * 8));
 
+	/* SOR */
 	if (!(ctrl & (1 << head))) {
 		if (nv_device(priv)->chipset  < 0x90 ||
 		    nv_device(priv)->chipset == 0x92 ||
@@ -887,6 +913,13 @@ exec_clkcmp(struct nv50_disp_priv *priv, int head, int id, u32 pclk,
 		}
 	}
 
+	/* PIOR */
+	if (!(ctrl & (1 << head))) {
+		for (i = 0; !(ctrl & (1 << head)) && i < 3; i++)
+			ctrl = nv_rd32(priv, 0x610b80 + (i * 8));
+		i += 8;
+	}
+
 	if (!(ctrl & (1 << head)))
 		return conf;
 	i--;
@@ -895,22 +928,27 @@ exec_clkcmp(struct nv50_disp_priv *priv, int head, int id, u32 pclk,
 	if (!data)
 		return conf;
 
-	switch (outp->type) {
-	case DCB_OUTPUT_TMDS:
-		conf = (ctrl & 0x00000f00) >> 8;
-		if (pclk >= 165000)
-			conf |= 0x0100;
-		break;
-	case DCB_OUTPUT_LVDS:
-		conf = priv->sor.lvdsconf;
-		break;
-	case DCB_OUTPUT_DP:
+	if (outp->location == 0) {
+		switch (outp->type) {
+		case DCB_OUTPUT_TMDS:
+			conf = (ctrl & 0x00000f00) >> 8;
+			if (pclk >= 165000)
+				conf |= 0x0100;
+			break;
+		case DCB_OUTPUT_LVDS:
+			conf = priv->sor.lvdsconf;
+			break;
+		case DCB_OUTPUT_DP:
+			conf = (ctrl & 0x00000f00) >> 8;
+			break;
+		case DCB_OUTPUT_ANALOG:
+		default:
+			conf = 0x00ff;
+			break;
+		}
+	} else {
 		conf = (ctrl & 0x00000f00) >> 8;
-		break;
-	case DCB_OUTPUT_ANALOG:
-	default:
-		conf = 0x00ff;
-		break;
+		pclk = pclk / 2;
 	}
 
 	data = nvbios_ocfg_match(bios, data, conf, &ver, &hdr, &cnt, &len, &info2);
@@ -1057,7 +1095,6 @@ static void
 nv50_disp_intr_unk20(struct nv50_disp_priv *priv, u32 super)
 {
 	struct dcb_output outp;
-	u32 addr, mask, data;
 	int head;
 
 	/* finish detaching encoder? */
@@ -1073,14 +1110,14 @@ nv50_disp_intr_unk20(struct nv50_disp_priv *priv, u32 super)
 			struct nouveau_clock *clk = nouveau_clock(priv);
 			clk->pll_set(clk, PLL_VPLL0 + head, pclk);
 		}
-
-		nv_mask(priv, 0x614200 + head * 0x800, 0x0000000f, 0x00000000);
 	}
 
 	/* (re)attach the relevant OR to the head */
 	head = ffs((super & 0x00000180) >> 7) - 1;
 	if (head >= 0) {
 		u32 pclk = nv_rd32(priv, 0x610ad0 + (head * 0x540)) & 0x3fffff;
+		u32 hval, hreg = 0x614200 + (head * 0x800);
+		u32 oval, oreg;
 		u32 conf = exec_clkcmp(priv, head, 0xff, pclk, &outp);
 		if (conf != ~0) {
 			if (outp.location == 0 && outp.type == DCB_OUTPUT_DP) {
@@ -1103,19 +1140,25 @@ nv50_disp_intr_unk20(struct nv50_disp_priv *priv, u32 super)
 
 			exec_clkcmp(priv, head, 0, pclk, &outp);
 
-			if (outp.type == DCB_OUTPUT_ANALOG) {
-				addr = 0x614280 + (ffs(outp.or) - 1) * 0x800;
-				mask = 0xffffffff;
-				data = 0x00000000;
-			} else {
+			if (!outp.location && outp.type == DCB_OUTPUT_ANALOG) {
+				oreg = 0x614280 + (ffs(outp.or) - 1) * 0x800;
+				oval = 0x00000000;
+				hval = 0x00000000;
+			} else
+			if (!outp.location) {
 				if (outp.type == DCB_OUTPUT_DP)
 					nv50_disp_intr_unk20_dp(priv, &outp, pclk);
-				addr = 0x614300 + (ffs(outp.or) - 1) * 0x800;
-				mask = 0x00000707;
-				data = (conf & 0x0100) ? 0x0101 : 0x0000;
+				oreg = 0x614300 + (ffs(outp.or) - 1) * 0x800;
+				oval = (conf & 0x0100) ? 0x0101 : 0x0000;
+				hval = 0x00000000;
+			} else {
+				oreg = 0x614380 + (ffs(outp.or) - 1) * 0x800;
+				oval = 0x00000001;
+				hval = 0x00000001;
 			}
 
-			nv_mask(priv, addr, mask, data);
+			nv_mask(priv, hreg, 0x0000000f, hval);
+			nv_mask(priv, oreg, 0x00000707, oval);
 		}
 	}
 
@@ -1151,9 +1194,28 @@ nv50_disp_intr_unk40(struct nv50_disp_priv *priv, u32 super)
 	if (head >= 0) {
 		struct dcb_output outp;
 		u32 pclk = nv_rd32(priv, 0x610ad0 + (head * 0x540)) & 0x3fffff;
-		if (exec_clkcmp(priv, head, 1, pclk, &outp) != ~0)
+		if (exec_clkcmp(priv, head, 1, pclk, &outp) != ~0) {
 			if (outp.location == 0 && outp.type == DCB_OUTPUT_TMDS)
 				nv50_disp_intr_unk40_tmds(priv, &outp);
+			else
+			if (outp.location == 1 && outp.type == DCB_OUTPUT_DP) {
+				u32 soff = (ffs(outp.or) - 1) * 0x08;
+				u32 ctrl = nv_rd32(priv, 0x610b84 + soff);
+				u32 datarate;
+
+				switch ((ctrl & 0x000f0000) >> 16) {
+				case 6: datarate = pclk * 30 / 8; break;
+				case 5: datarate = pclk * 24 / 8; break;
+				case 2:
+				default:
+					datarate = pclk * 18 / 8;
+					break;
+				}
+
+				nouveau_dp_train(&priv->base, priv->pior.dp,
+						 &outp, head, datarate);
+			}
+		}
 	}
 
 	nv_wr32(priv, 0x610030, 0x80000000);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/init.c b/drivers/gpu/drm/nouveau/core/subdev/bios/init.c
index 7d2ce11..2cc1e6a 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/init.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/init.c
@@ -231,6 +231,11 @@ init_i2c(struct nvbios_init *init, int index)
 			return NULL;
 		}
 
+		if (index == -2 && init->outp->location) {
+			index = NV_I2C_TYPE_EXTAUX(init->outp->extdev);
+			return i2c->find_type(i2c, index);
+		}
+
 		index = init->outp->i2c_index;
 	}
 
@@ -258,7 +263,7 @@ init_wri2cr(struct nvbios_init *init, u8 index, u8 addr, u8 reg, u8 val)
 static int
 init_rdauxr(struct nvbios_init *init, u32 addr)
 {
-	struct nouveau_i2c_port *port = init_i2c(init, -1);
+	struct nouveau_i2c_port *port = init_i2c(init, -2);
 	u8 data;
 
 	if (port && init_exec(init)) {
@@ -274,7 +279,7 @@ init_rdauxr(struct nvbios_init *init, u32 addr)
 static int
 init_wrauxr(struct nvbios_init *init, u32 addr, u8 data)
 {
-	struct nouveau_i2c_port *port = init_i2c(init, -1);
+	struct nouveau_i2c_port *port = init_i2c(init, -2);
 	if (port && init_exec(init))
 		return nv_wraux(port, addr, &data, 1);
 	return -ENODEV;
-- 
cgit v0.10.2


From eb6313add6dddf07ea3e50c4caa33a9c3b2379f1 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Mon, 11 Feb 2013 09:52:58 +1000
Subject: drm/nv50: initial kms support for off-chip TMDS/DP encoders

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.c b/drivers/gpu/drm/nouveau/nouveau_bios.c
index b640657..50a6dd0 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bios.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bios.c
@@ -1469,6 +1469,7 @@ parse_dcb20_entry(struct drm_device *dev, struct dcb_table *dcb,
 	}
 	case DCB_OUTPUT_DP:
 		entry->dpconf.sor.link = (conf & 0x00000030) >> 4;
+		entry->extdev = (conf & 0x0000ff00) >> 8;
 		switch ((conf & 0x00e00000) >> 21) {
 		case 0:
 			entry->dpconf.link_bw = 162000;
@@ -1490,8 +1491,10 @@ parse_dcb20_entry(struct drm_device *dev, struct dcb_table *dcb,
 		}
 		break;
 	case DCB_OUTPUT_TMDS:
-		if (dcb->version >= 0x40)
+		if (dcb->version >= 0x40) {
 			entry->tmdsconf.sor.link = (conf & 0x00000030) >> 4;
+			entry->extdev = (conf & 0x0000ff00) >> 8;
+		}
 		else if (dcb->version >= 0x30)
 			entry->tmdsconf.slave_addr = (conf & 0x00000700) >> 8;
 		else if (dcb->version >= 0x22)
diff --git a/drivers/gpu/drm/nouveau/nv50_display.c b/drivers/gpu/drm/nouveau/nv50_display.c
index 5c56575..49f26cd 100644
--- a/drivers/gpu/drm/nouveau/nv50_display.c
+++ b/drivers/gpu/drm/nouveau/nv50_display.c
@@ -1923,6 +1923,184 @@ nv50_sor_create(struct drm_connector *connector, struct dcb_output *dcbe)
 }
 
 /******************************************************************************
+ * PIOR
+ *****************************************************************************/
+
+static void
+nv50_pior_dpms(struct drm_encoder *encoder, int mode)
+{
+	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
+	struct nv50_disp *disp = nv50_disp(encoder->dev);
+	u32 mthd = (nv_encoder->dcb->type << 12) | nv_encoder->or;
+	u32 ctrl = (mode == DRM_MODE_DPMS_ON);
+	nv_call(disp->core, NV50_DISP_PIOR_PWR + mthd, ctrl);
+}
+
+static bool
+nv50_pior_mode_fixup(struct drm_encoder *encoder,
+		     const struct drm_display_mode *mode,
+		     struct drm_display_mode *adjusted_mode)
+{
+	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
+	struct nouveau_connector *nv_connector;
+
+	nv_connector = nouveau_encoder_connector_get(nv_encoder);
+	if (nv_connector && nv_connector->native_mode) {
+		if (nv_connector->scaling_mode != DRM_MODE_SCALE_NONE) {
+			int id = adjusted_mode->base.id;
+			*adjusted_mode = *nv_connector->native_mode;
+			adjusted_mode->base.id = id;
+		}
+	}
+
+	adjusted_mode->clock *= 2;
+	return true;
+}
+
+static void
+nv50_pior_commit(struct drm_encoder *encoder)
+{
+}
+
+static void
+nv50_pior_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
+		   struct drm_display_mode *adjusted_mode)
+{
+	struct nv50_mast *mast = nv50_mast(encoder->dev);
+	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
+	struct nouveau_crtc *nv_crtc = nouveau_crtc(encoder->crtc);
+	struct nouveau_connector *nv_connector;
+	u8 owner = 1 << nv_crtc->index;
+	u8 proto, depth;
+	u32 *push;
+
+	nv_connector = nouveau_encoder_connector_get(nv_encoder);
+	switch (nv_connector->base.display_info.bpc) {
+	case 10: depth = 0x6; break;
+	case  8: depth = 0x5; break;
+	case  6: depth = 0x2; break;
+	default: depth = 0x0; break;
+	}
+
+	switch (nv_encoder->dcb->type) {
+	case DCB_OUTPUT_TMDS:
+	case DCB_OUTPUT_DP:
+		proto = 0x0;
+		break;
+	default:
+		BUG_ON(1);
+		break;
+	}
+
+	nv50_pior_dpms(encoder, DRM_MODE_DPMS_ON);
+
+	push = evo_wait(mast, 8);
+	if (push) {
+		if (nv50_vers(mast) < NVD0_DISP_MAST_CLASS) {
+			u32 ctrl = (depth << 16) | (proto << 8) | owner;
+			if (mode->flags & DRM_MODE_FLAG_NHSYNC)
+				ctrl |= 0x00001000;
+			if (mode->flags & DRM_MODE_FLAG_NVSYNC)
+				ctrl |= 0x00002000;
+			evo_mthd(push, 0x0700 + (nv_encoder->or * 0x040), 1);
+			evo_data(push, ctrl);
+		}
+
+		evo_kick(push, mast);
+	}
+
+	nv_encoder->crtc = encoder->crtc;
+}
+
+static void
+nv50_pior_disconnect(struct drm_encoder *encoder)
+{
+	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
+	struct nv50_mast *mast = nv50_mast(encoder->dev);
+	const int or = nv_encoder->or;
+	u32 *push;
+
+	if (nv_encoder->crtc) {
+		nv50_crtc_prepare(nv_encoder->crtc);
+
+		push = evo_wait(mast, 4);
+		if (push) {
+			if (nv50_vers(mast) < NVD0_DISP_MAST_CLASS) {
+				evo_mthd(push, 0x0700 + (or * 0x040), 1);
+				evo_data(push, 0x00000000);
+			}
+
+			evo_mthd(push, 0x0080, 1);
+			evo_data(push, 0x00000000);
+			evo_kick(push, mast);
+		}
+	}
+
+	nv_encoder->crtc = NULL;
+}
+
+static void
+nv50_pior_destroy(struct drm_encoder *encoder)
+{
+	drm_encoder_cleanup(encoder);
+	kfree(encoder);
+}
+
+static const struct drm_encoder_helper_funcs nv50_pior_hfunc = {
+	.dpms = nv50_pior_dpms,
+	.mode_fixup = nv50_pior_mode_fixup,
+	.prepare = nv50_pior_disconnect,
+	.commit = nv50_pior_commit,
+	.mode_set = nv50_pior_mode_set,
+	.disable = nv50_pior_disconnect,
+	.get_crtc = nv50_display_crtc_get,
+};
+
+static const struct drm_encoder_funcs nv50_pior_func = {
+	.destroy = nv50_pior_destroy,
+};
+
+static int
+nv50_pior_create(struct drm_connector *connector, struct dcb_output *dcbe)
+{
+	struct nouveau_drm *drm = nouveau_drm(connector->dev);
+	struct nouveau_i2c *i2c = nouveau_i2c(drm->device);
+	struct nouveau_i2c_port *ddc = NULL;
+	struct nouveau_encoder *nv_encoder;
+	struct drm_encoder *encoder;
+	int type;
+
+	switch (dcbe->type) {
+	case DCB_OUTPUT_TMDS:
+		ddc  = i2c->find_type(i2c, NV_I2C_TYPE_EXTDDC(dcbe->extdev));
+		type = DRM_MODE_ENCODER_TMDS;
+		break;
+	case DCB_OUTPUT_DP:
+		ddc  = i2c->find_type(i2c, NV_I2C_TYPE_EXTAUX(dcbe->extdev));
+		type = DRM_MODE_ENCODER_TMDS;
+		break;
+	default:
+		return -ENODEV;
+	}
+
+	nv_encoder = kzalloc(sizeof(*nv_encoder), GFP_KERNEL);
+	if (!nv_encoder)
+		return -ENOMEM;
+	nv_encoder->dcb = dcbe;
+	nv_encoder->or = ffs(dcbe->or) - 1;
+	nv_encoder->i2c = ddc;
+
+	encoder = to_drm_encoder(nv_encoder);
+	encoder->possible_crtcs = dcbe->heads;
+	encoder->possible_clones = 0;
+	drm_encoder_init(connector->dev, encoder, &nv50_pior_func, type);
+	drm_encoder_helper_add(encoder, &nv50_pior_hfunc);
+
+	drm_mode_connector_attach_encoder(connector, encoder);
+	return 0;
+}
+
+/******************************************************************************
  * Init
  *****************************************************************************/
 void
@@ -2044,25 +2222,28 @@ nv50_display_create(struct drm_device *dev)
 		if (IS_ERR(connector))
 			continue;
 
-		if (dcbe->location != DCB_LOC_ON_CHIP) {
-			NV_WARN(drm, "skipping off-chip encoder %d/%d\n",
-				dcbe->type, ffs(dcbe->or) - 1);
-			continue;
+		if (dcbe->location == DCB_LOC_ON_CHIP) {
+			switch (dcbe->type) {
+			case DCB_OUTPUT_TMDS:
+			case DCB_OUTPUT_LVDS:
+			case DCB_OUTPUT_DP:
+				ret = nv50_sor_create(connector, dcbe);
+				break;
+			case DCB_OUTPUT_ANALOG:
+				ret = nv50_dac_create(connector, dcbe);
+				break;
+			default:
+				ret = -ENODEV;
+				break;
+			}
+		} else {
+			ret = nv50_pior_create(connector, dcbe);
 		}
 
-		switch (dcbe->type) {
-		case DCB_OUTPUT_TMDS:
-		case DCB_OUTPUT_LVDS:
-		case DCB_OUTPUT_DP:
-			nv50_sor_create(connector, dcbe);
-			break;
-		case DCB_OUTPUT_ANALOG:
-			nv50_dac_create(connector, dcbe);
-			break;
-		default:
-			NV_WARN(drm, "skipping unsupported encoder %d/%d\n",
-				dcbe->type, ffs(dcbe->or) - 1);
-			continue;
+		if (ret) {
+			NV_WARN(drm, "failed to create encoder %d/%d/%d: %d\n",
+				     dcbe->location, dcbe->type,
+				     ffs(dcbe->or) - 1, ret);
 		}
 	}
 
-- 
cgit v0.10.2


From 5effecd4f8ff0200c4224060275a6c4ecbeed3eb Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Thu, 14 Feb 2013 17:48:58 -0500
Subject: drm/nouveau/i2c: add support for ddc/aux, and dp link training on
 anx9805

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index bda661e..c892988 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -93,6 +93,7 @@ nouveau-y += core/subdev/gpio/nv50.o
 nouveau-y += core/subdev/gpio/nvd0.o
 nouveau-y += core/subdev/gpio/nve0.o
 nouveau-y += core/subdev/i2c/base.o
+nouveau-y += core/subdev/i2c/anx9805.o
 nouveau-y += core/subdev/i2c/aux.o
 nouveau-y += core/subdev/i2c/bit.o
 nouveau-y += core/subdev/i2c/nv04.o
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/i2c.h b/drivers/gpu/drm/nouveau/core/include/subdev/i2c.h
index c5868cd..888384c 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/i2c.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/i2c.h
@@ -103,6 +103,7 @@ extern struct nouveau_oclass nv4e_i2c_oclass;
 extern struct nouveau_oclass nv50_i2c_oclass;
 extern struct nouveau_oclass nv94_i2c_oclass;
 extern struct nouveau_oclass nvd0_i2c_oclass;
+extern struct nouveau_oclass nouveau_anx9805_sclass[];
 
 extern const struct i2c_algorithm nouveau_i2c_bit_algo;
 extern const struct i2c_algorithm nouveau_i2c_aux_algo;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/anx9805.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/anx9805.c
new file mode 100644
index 0000000..dec94e9
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/anx9805.c
@@ -0,0 +1,279 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include <subdev/i2c.h>
+
+struct anx9805_i2c_port {
+	struct nouveau_i2c_port base;
+	u32 addr;
+	u32 ctrl;
+};
+
+static int
+anx9805_train(struct nouveau_i2c_port *port, int link_nr, int link_bw, bool enh)
+{
+	struct anx9805_i2c_port *chan = (void *)port;
+	struct nouveau_i2c_port *mast = (void *)nv_object(chan)->parent;
+	u8 tmp, i;
+
+	nv_wri2cr(mast, chan->addr, 0xa0, link_bw);
+	nv_wri2cr(mast, chan->addr, 0xa1, link_nr | (enh ? 0x80 : 0x00));
+	nv_wri2cr(mast, chan->addr, 0xa2, 0x01);
+	nv_wri2cr(mast, chan->addr, 0xa8, 0x01);
+
+	i = 0;
+	while ((tmp = nv_rdi2cr(mast, chan->addr, 0xa8)) & 0x01) {
+		mdelay(5);
+		if (i++ == 100) {
+			nv_error(port, "link training timed out\n");
+			return -ETIMEDOUT;
+		}
+	}
+
+	if (tmp & 0x70) {
+		nv_error(port, "link training failed: 0x%02x\n", tmp);
+		return -EIO;
+	}
+
+	return 1;
+}
+
+static int
+anx9805_aux(struct nouveau_i2c_port *port, u8 type, u32 addr, u8 *data, u8 size)
+{
+	struct anx9805_i2c_port *chan = (void *)port;
+	struct nouveau_i2c_port *mast = (void *)nv_object(chan)->parent;
+	int i, ret = -ETIMEDOUT;
+	u8 tmp;
+
+	tmp = nv_rdi2cr(mast, chan->ctrl, 0x07) & ~0x04;
+	nv_wri2cr(mast, chan->ctrl, 0x07, tmp | 0x04);
+	nv_wri2cr(mast, chan->ctrl, 0x07, tmp);
+	nv_wri2cr(mast, chan->ctrl, 0xf7, 0x01);
+
+	nv_wri2cr(mast, chan->addr, 0xe4, 0x80);
+	for (i = 0; !(type & 1) && i < size; i++)
+		nv_wri2cr(mast, chan->addr, 0xf0 + i, data[i]);
+	nv_wri2cr(mast, chan->addr, 0xe5, ((size - 1) << 4) | type);
+	nv_wri2cr(mast, chan->addr, 0xe6, (addr & 0x000ff) >>  0);
+	nv_wri2cr(mast, chan->addr, 0xe7, (addr & 0x0ff00) >>  8);
+	nv_wri2cr(mast, chan->addr, 0xe8, (addr & 0xf0000) >> 16);
+	nv_wri2cr(mast, chan->addr, 0xe9, 0x01);
+
+	i = 0;
+	while ((tmp = nv_rdi2cr(mast, chan->addr, 0xe9)) & 0x01) {
+		mdelay(5);
+		if (i++ == 32)
+			goto done;
+	}
+
+	if ((tmp = nv_rdi2cr(mast, chan->ctrl, 0xf7)) & 0x01) {
+		ret = -EIO;
+		goto done;
+	}
+
+	for (i = 0; (type & 1) && i < size; i++)
+		data[i] = nv_rdi2cr(mast, chan->addr, 0xf0 + i);
+	ret = 0;
+done:
+	nv_wri2cr(mast, chan->ctrl, 0xf7, 0x01);
+	return ret;
+}
+
+static const struct nouveau_i2c_func
+anx9805_aux_func = {
+	.aux = anx9805_aux,
+	.lnk_ctl = anx9805_train,
+};
+
+static int
+anx9805_aux_chan_ctor(struct nouveau_object *parent,
+		      struct nouveau_object *engine,
+		      struct nouveau_oclass *oclass, void *data, u32 index,
+		      struct nouveau_object **pobject)
+{
+	struct nouveau_i2c_port *mast = (void *)parent;
+	struct anx9805_i2c_port *chan;
+	int ret;
+
+	ret = nouveau_i2c_port_create(parent, engine, oclass, index,
+				     &nouveau_i2c_aux_algo, &chan);
+	*pobject = nv_object(chan);
+	if (ret)
+		return ret;
+
+	switch ((oclass->handle & 0xff00) >> 8) {
+	case 0x0d:
+		chan->addr = 0x38;
+		chan->ctrl = 0x39;
+		break;
+	case 0x0e:
+		chan->addr = 0x3c;
+		chan->ctrl = 0x3b;
+		break;
+	default:
+		BUG_ON(1);
+	}
+
+	if (mast->adapter.algo == &i2c_bit_algo) {
+		struct i2c_algo_bit_data *algo = mast->adapter.algo_data;
+		algo->udelay = max(algo->udelay, 40);
+	}
+
+	chan->base.func = &anx9805_aux_func;
+	return 0;
+}
+
+static struct nouveau_ofuncs
+anx9805_aux_ofuncs = {
+	.ctor =  anx9805_aux_chan_ctor,
+	.dtor = _nouveau_i2c_port_dtor,
+	.init = _nouveau_i2c_port_init,
+	.fini = _nouveau_i2c_port_fini,
+};
+
+static int
+anx9805_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, int num)
+{
+	struct anx9805_i2c_port *port = adap->algo_data;
+	struct nouveau_i2c_port *mast = (void *)nv_object(port)->parent;
+	struct i2c_msg *msg = msgs;
+	int ret = -ETIMEDOUT;
+	int i, j, cnt = num;
+	u8 seg = 0x00, off = 0x00, tmp;
+
+	tmp = nv_rdi2cr(mast, port->ctrl, 0x07) & ~0x10;
+	nv_wri2cr(mast, port->ctrl, 0x07, tmp | 0x10);
+	nv_wri2cr(mast, port->ctrl, 0x07, tmp);
+	nv_wri2cr(mast, port->addr, 0x43, 0x05);
+	mdelay(5);
+
+	while (cnt--) {
+		if ( (msg->flags & I2C_M_RD) && msg->addr == 0x50) {
+			nv_wri2cr(mast, port->addr, 0x40, msg->addr << 1);
+			nv_wri2cr(mast, port->addr, 0x41, seg);
+			nv_wri2cr(mast, port->addr, 0x42, off);
+			nv_wri2cr(mast, port->addr, 0x44, msg->len);
+			nv_wri2cr(mast, port->addr, 0x45, 0x00);
+			nv_wri2cr(mast, port->addr, 0x43, 0x01);
+			for (i = 0; i < msg->len; i++) {
+				j = 0;
+				while (nv_rdi2cr(mast, port->addr, 0x46) & 0x10) {
+					mdelay(5);
+					if (j++ == 32)
+						goto done;
+				}
+				msg->buf[i] = nv_rdi2cr(mast, port->addr, 0x47);
+			}
+		} else
+		if (!(msg->flags & I2C_M_RD)) {
+			if (msg->addr == 0x50 && msg->len == 0x01) {
+				off = msg->buf[0];
+			} else
+			if (msg->addr == 0x30 && msg->len == 0x01) {
+				seg = msg->buf[0];
+			} else
+				goto done;
+		} else {
+			goto done;
+		}
+		msg++;
+	}
+
+	ret = num;
+done:
+	nv_wri2cr(mast, port->addr, 0x43, 0x00);
+	return ret;
+}
+
+static u32
+anx9805_func(struct i2c_adapter *adap)
+{
+	return I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL;
+}
+
+static const struct i2c_algorithm
+anx9805_i2c_algo = {
+	.master_xfer = anx9805_xfer,
+	.functionality = anx9805_func
+};
+
+static const struct nouveau_i2c_func
+anx9805_i2c_func = {
+};
+
+static int
+anx9805_ddc_port_ctor(struct nouveau_object *parent,
+		      struct nouveau_object *engine,
+		      struct nouveau_oclass *oclass, void *data, u32 index,
+		      struct nouveau_object **pobject)
+{
+	struct nouveau_i2c_port *mast = (void *)parent;
+	struct anx9805_i2c_port *port;
+	int ret;
+
+	ret = nouveau_i2c_port_create(parent, engine, oclass, index,
+				     &anx9805_i2c_algo, &port);
+	*pobject = nv_object(port);
+	if (ret)
+		return ret;
+
+	switch ((oclass->handle & 0xff00) >> 8) {
+	case 0x0d:
+		port->addr = 0x3d;
+		port->ctrl = 0x39;
+		break;
+	case 0x0e:
+		port->addr = 0x3f;
+		port->ctrl = 0x3b;
+		break;
+	default:
+		BUG_ON(1);
+	}
+
+	if (mast->adapter.algo == &i2c_bit_algo) {
+		struct i2c_algo_bit_data *algo = mast->adapter.algo_data;
+		algo->udelay = max(algo->udelay, 40);
+	}
+
+	port->base.func = &anx9805_i2c_func;
+	return 0;
+}
+
+static struct nouveau_ofuncs
+anx9805_ddc_ofuncs = {
+	.ctor =  anx9805_ddc_port_ctor,
+	.dtor = _nouveau_i2c_port_dtor,
+	.init = _nouveau_i2c_port_init,
+	.fini = _nouveau_i2c_port_fini,
+};
+
+struct nouveau_oclass
+nouveau_anx9805_sclass[] = {
+	{ .handle = NV_I2C_TYPE_EXTDDC(0x0d), .ofuncs = &anx9805_ddc_ofuncs },
+	{ .handle = NV_I2C_TYPE_EXTAUX(0x0d), .ofuncs = &anx9805_aux_ofuncs },
+	{ .handle = NV_I2C_TYPE_EXTDDC(0x0e), .ofuncs = &anx9805_ddc_ofuncs },
+	{ .handle = NV_I2C_TYPE_EXTAUX(0x0e), .ofuncs = &anx9805_aux_ofuncs },
+	{}
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
index 34a0523..a114a0e 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
@@ -24,6 +24,9 @@
 
 #include <core/option.h>
 
+#include <subdev/bios.h>
+#include <subdev/bios/dcb.h>
+#include <subdev/bios/i2c.h>
 #include <subdev/i2c.h>
 #include <subdev/vga.h>
 
@@ -278,6 +281,11 @@ _nouveau_i2c_dtor(struct nouveau_object *object)
 	nouveau_subdev_destroy(&i2c->base);
 }
 
+static struct nouveau_oclass *
+nouveau_i2c_extdev_sclass[] = {
+	nouveau_anx9805_sclass,
+};
+
 int
 nouveau_i2c_create_(struct nouveau_object *parent,
 		    struct nouveau_object *engine,
@@ -289,7 +297,10 @@ nouveau_i2c_create_(struct nouveau_object *parent,
 	struct nouveau_i2c *i2c;
 	struct nouveau_object *object;
 	struct dcb_i2c_entry info;
-	int ret, index = -1;
+	int ret, i, j, index = -1;
+	struct dcb_output outp;
+	u8  ver, hdr;
+	u32 data;
 
 	ret = nouveau_subdev_create(parent, engine, oclass, 0,
 				    "I2C", "i2c", &i2c);
@@ -317,5 +328,41 @@ nouveau_i2c_create_(struct nouveau_object *parent,
 		} while (ret && (++oclass)->handle);
 	}
 
+	/* in addition to the busses specified in the i2c table, there
+	 * may be ddc/aux channels hiding behind external tmds/dp/etc
+	 * transmitters.
+	 */
+	index = ((index + 0x0f) / 0x10) * 0x10;
+	i = -1;
+	while ((data = dcb_outp_parse(bios, ++i, &ver, &hdr, &outp))) {
+		if (!outp.location || !outp.extdev)
+			continue;
+
+		switch (outp.type) {
+		case DCB_OUTPUT_TMDS:
+			info.type = NV_I2C_TYPE_EXTDDC(outp.extdev);
+			break;
+		case DCB_OUTPUT_DP:
+			info.type = NV_I2C_TYPE_EXTAUX(outp.extdev);
+			break;
+		default:
+			continue;
+		}
+
+		ret = -ENODEV;
+		j = -1;
+		while (ret && ++j < ARRAY_SIZE(nouveau_i2c_extdev_sclass)) {
+			parent = nv_object(i2c->find(i2c, outp.i2c_index));
+			oclass = nouveau_i2c_extdev_sclass[j];
+			do {
+				if (oclass->handle != info.type)
+					continue;
+				ret = nouveau_object_ctor(parent, *pobject,
+							  oclass, NULL,
+							  index++, &object);
+			} while (ret && (++oclass)->handle);
+		}
+	}
+
 	return 0;
 }
-- 
cgit v0.10.2


From a91ed42de25e7e81159c0dd59faf8cac9dfa1d32 Mon Sep 17 00:00:00 2001
From: Ben Hutchings <ben@decadent.org.uk>
Date: Wed, 20 Feb 2013 02:57:32 +0000
Subject: nouveau: ACPI support depends on X86 and X86_PLATFORM_DEVICES

If I build nouveau on ia64, Kconfig warns:

warning: (DRM_NOUVEAU) selects ACPI_WMI which has unmet direct dependencies (X86 && X86_PLATFORM_DEVICES && ACPI)
warning: (DRM_NOUVEAU) selects MXM_WMI which has unmet direct dependencies (X86 && X86_PLATFORM_DEVICES && ACPI_WMI)

Make all the ACPI support depend on X86 and select
X86_PLATFORM_DEVICES.

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/Kconfig b/drivers/gpu/drm/nouveau/Kconfig
index 8a55bee..120e7d1 100644
--- a/drivers/gpu/drm/nouveau/Kconfig
+++ b/drivers/gpu/drm/nouveau/Kconfig
@@ -11,8 +11,9 @@ config DRM_NOUVEAU
 	select FRAMEBUFFER_CONSOLE if !EXPERT
 	select FB_BACKLIGHT if DRM_NOUVEAU_BACKLIGHT
 	select ACPI_VIDEO if ACPI && X86 && BACKLIGHT_CLASS_DEVICE && VIDEO_OUTPUT_CONTROL && INPUT
-	select ACPI_WMI if ACPI
-	select MXM_WMI if ACPI
+	select X86_PLATFORM_DEVICES if ACPI && X86
+	select ACPI_WMI if ACPI && X86
+	select MXM_WMI if ACPI && X86
 	select POWER_SUPPLY
 	help
 	  Choose this option for open-source nVidia support.
diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index c892988..51d0972 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -238,7 +238,9 @@ nouveau-y += nouveau_mem.o
 
 # other random bits
 nouveau-$(CONFIG_COMPAT) += nouveau_ioc32.o
+ifdef CONFIG_X86
 nouveau-$(CONFIG_ACPI) += nouveau_acpi.o
+endif
 nouveau-$(CONFIG_DRM_NOUVEAU_BACKLIGHT) += nouveau_backlight.o
 
 obj-$(CONFIG_DRM_NOUVEAU)+= nouveau.o
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/base.c b/drivers/gpu/drm/nouveau/core/subdev/bios/base.c
index f621f69..e816f06 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/base.c
@@ -172,7 +172,7 @@ out:
 	nv_wr32(bios, pcireg, access);
 }
 
-#if defined(CONFIG_ACPI)
+#if defined(CONFIG_ACPI) && defined(CONFIG_X86)
 int nouveau_acpi_get_bios_chunk(uint8_t *bios, int offset, int len);
 bool nouveau_acpi_rom_supported(struct pci_dev *pdev);
 #else
diff --git a/drivers/gpu/drm/nouveau/nouveau_acpi.h b/drivers/gpu/drm/nouveau/nouveau_acpi.h
index d0da230..74acf0f 100644
--- a/drivers/gpu/drm/nouveau/nouveau_acpi.h
+++ b/drivers/gpu/drm/nouveau/nouveau_acpi.h
@@ -3,7 +3,7 @@
 
 #define ROM_BIOS_PAGE 4096
 
-#if defined(CONFIG_ACPI)
+#if defined(CONFIG_ACPI) && defined(CONFIG_X86)
 bool nouveau_is_optimus(void);
 bool nouveau_is_v1_dsm(void);
 void nouveau_register_dsm_handler(void);
-- 
cgit v0.10.2


From 2e82b5dd47d826baf0f6e34334fe6b899aa6ad1e Mon Sep 17 00:00:00 2001
From: "Lee, Chun-Yi" <joeyli.kernel@gmail.com>
Date: Wed, 20 Feb 2013 14:32:01 +0800
Subject: gpu: remove gma500 stub driver

In v3.3, the gma500 drm driver moved from staging to drm group by
Alan Cox's 3abcf41fb patch. the gma500 drm driver should control
brightness well and don't need gma500 stub driver anymore.

Reference:
http://lists.freedesktop.org/archives/dri-devel/2012-May/023426.html
http://lists.freedesktop.org/archives/dri-devel/2012-May/023467.html

Acked-by: Matthew Garrett <mjg59@srcf.ucam.org>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Lee, Chun-Yi <jlee@suse.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/gpu/Makefile b/drivers/gpu/Makefile
index cc92778..30879df 100644
--- a/drivers/gpu/Makefile
+++ b/drivers/gpu/Makefile
@@ -1 +1 @@
-obj-y			+= drm/ vga/ stub/
+obj-y			+= drm/ vga/
diff --git a/drivers/gpu/stub/Kconfig b/drivers/gpu/stub/Kconfig
deleted file mode 100644
index 4199179..0000000
--- a/drivers/gpu/stub/Kconfig
+++ /dev/null
@@ -1,18 +0,0 @@
-config STUB_POULSBO
-	tristate "Intel GMA500 Stub Driver"
-	depends on PCI
-	depends on NET # for THERMAL
-	# Poulsbo stub depends on ACPI_VIDEO when ACPI is enabled
-	# but for select to work, need to select ACPI_VIDEO's dependencies, ick
-	select BACKLIGHT_CLASS_DEVICE if ACPI
-	select VIDEO_OUTPUT_CONTROL if ACPI
-	select INPUT if ACPI
-	select ACPI_VIDEO if ACPI
-	select THERMAL if ACPI
-	help
-	  Choose this option if you have a system that has Intel GMA500
-	  (Poulsbo) integrated graphics. If M is selected, the module will
-	  be called Poulsbo. This driver is a stub driver for Poulsbo that
-	  will call poulsbo.ko to enable the acpi backlight control sysfs
-	  entry file because there have no poulsbo native driver can support
-	  intel opregion.
diff --git a/drivers/gpu/stub/Makefile b/drivers/gpu/stub/Makefile
deleted file mode 100644
index cd940cc..0000000
--- a/drivers/gpu/stub/Makefile
+++ /dev/null
@@ -1 +0,0 @@
-obj-$(CONFIG_STUB_POULSBO) += poulsbo.o
diff --git a/drivers/gpu/stub/poulsbo.c b/drivers/gpu/stub/poulsbo.c
deleted file mode 100644
index 7edfd27..0000000
--- a/drivers/gpu/stub/poulsbo.c
+++ /dev/null
@@ -1,64 +0,0 @@
-/*
- * Intel Poulsbo Stub driver
- *
- * Copyright (C) 2010 Novell <jlee@novell.com>
- *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of the GNU General Public License version 2 as published by
- * the Free Software Foundation.
- *
- */
-
-#include <linux/module.h>
-#include <linux/pci.h>
-#include <linux/acpi.h>
-#include <acpi/video.h>
-
-#define DRIVER_NAME "poulsbo"
-
-enum {
-	CHIP_PSB_8108 = 0,
-	CHIP_PSB_8109 = 1,
-};
-
-static DEFINE_PCI_DEVICE_TABLE(pciidlist) = {
-	{0x8086, 0x8108, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PSB_8108}, \
-	{0x8086, 0x8109, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PSB_8109}, \
-	{0, 0, 0}
-};
-
-static int poulsbo_probe(struct pci_dev *pdev, const struct pci_device_id *id)
-{
-	return acpi_video_register();
-}
-
-static void poulsbo_remove(struct pci_dev *pdev)
-{
-	acpi_video_unregister();
-}
-
-static struct pci_driver poulsbo_driver = {
-	.name = DRIVER_NAME,
-	.id_table = pciidlist,
-	.probe = poulsbo_probe,
-	.remove = poulsbo_remove,
-};
-
-static int __init poulsbo_init(void)
-{
-	return pci_register_driver(&poulsbo_driver);
-}
-
-static void __exit poulsbo_exit(void)
-{
-	pci_unregister_driver(&poulsbo_driver);
-}
-
-module_init(poulsbo_init);
-module_exit(poulsbo_exit);
-
-MODULE_AUTHOR("Lee, Chun-Yi <jlee@novell.com>");
-MODULE_DESCRIPTION("Poulsbo Stub Driver");
-MODULE_LICENSE("GPL");
-
-MODULE_DEVICE_TABLE(pci, pciidlist);
diff --git a/drivers/video/Kconfig b/drivers/video/Kconfig
index 807c7fa..09f1a18 100644
--- a/drivers/video/Kconfig
+++ b/drivers/video/Kconfig
@@ -21,8 +21,6 @@ source "drivers/gpu/vga/Kconfig"
 
 source "drivers/gpu/drm/Kconfig"
 
-source "drivers/gpu/stub/Kconfig"
-
 config VGASTATE
        tristate
        default n
-- 
cgit v0.10.2


From 16d4c031dd3befa1a85ff4645e28bb173e16ef43 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Wed, 20 Feb 2013 18:56:33 +1000
Subject: drm/nv50/disp: handle multiple actions from one set of supervisor
 intrs

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
index 314dda6..5fa1326 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
@@ -972,21 +972,29 @@ exec_clkcmp(struct nv50_disp_priv *priv, int head, int id, u32 pclk,
 }
 
 static void
-nv50_disp_intr_unk10(struct nv50_disp_priv *priv, u32 super)
+nv50_disp_intr_unk10_0(struct nv50_disp_priv *priv, int head)
 {
-	int head = ffs((super & 0x00000060) >> 5) - 1;
-	if (head >= 0) {
-		head = ffs((super & 0x00000180) >> 7) - 1;
-		if (head >= 0)
-			exec_script(priv, head, 1);
-	}
+	exec_script(priv, head, 1);
+}
 
-	nv_wr32(priv, 0x610030, 0x80000000);
+static void
+nv50_disp_intr_unk20_0(struct nv50_disp_priv *priv, int head)
+{
+	exec_script(priv, head, 2);
 }
 
 static void
-nv50_disp_intr_unk20_dp(struct nv50_disp_priv *priv,
-		        struct dcb_output *outp, u32 pclk)
+nv50_disp_intr_unk20_1(struct nv50_disp_priv *priv, int head)
+{
+	struct nouveau_clock *clk = nouveau_clock(priv);
+	u32 pclk = nv_rd32(priv, 0x610ad0 + (head * 0x540)) & 0x3fffff;
+	if (pclk)
+		clk->pll_set(clk, PLL_VPLL0 + head, pclk);
+}
+
+static void
+nv50_disp_intr_unk20_2_dp(struct nv50_disp_priv *priv,
+			  struct dcb_output *outp, u32 pclk)
 {
 	const int link = !(outp->sorconf.link & 1);
 	const int   or = ffs(outp->or) - 1;
@@ -1092,77 +1100,54 @@ nv50_disp_intr_unk20_dp(struct nv50_disp_priv *priv,
 }
 
 static void
-nv50_disp_intr_unk20(struct nv50_disp_priv *priv, u32 super)
+nv50_disp_intr_unk20_2(struct nv50_disp_priv *priv, int head)
 {
 	struct dcb_output outp;
-	int head;
-
-	/* finish detaching encoder? */
-	head = ffs((super & 0x00000180) >> 7) - 1;
-	if (head >= 0)
-		exec_script(priv, head, 2);
-
-	/* check whether a vpll change is required */
-	head = ffs((super & 0x00000600) >> 9) - 1;
-	if (head >= 0) {
-		u32 pclk = nv_rd32(priv, 0x610ad0 + (head * 0x540)) & 0x3fffff;
-		if (pclk) {
-			struct nouveau_clock *clk = nouveau_clock(priv);
-			clk->pll_set(clk, PLL_VPLL0 + head, pclk);
-		}
-	}
-
-	/* (re)attach the relevant OR to the head */
-	head = ffs((super & 0x00000180) >> 7) - 1;
-	if (head >= 0) {
-		u32 pclk = nv_rd32(priv, 0x610ad0 + (head * 0x540)) & 0x3fffff;
-		u32 hval, hreg = 0x614200 + (head * 0x800);
-		u32 oval, oreg;
-		u32 conf = exec_clkcmp(priv, head, 0xff, pclk, &outp);
-		if (conf != ~0) {
-			if (outp.location == 0 && outp.type == DCB_OUTPUT_DP) {
-				u32 soff = (ffs(outp.or) - 1) * 0x08;
-				u32 ctrl = nv_rd32(priv, 0x610798 + soff);
-				u32 datarate;
-
-				switch ((ctrl & 0x000f0000) >> 16) {
-				case 6: datarate = pclk * 30 / 8; break;
-				case 5: datarate = pclk * 24 / 8; break;
-				case 2:
-				default:
-					datarate = pclk * 18 / 8;
-					break;
-				}
-
-				nouveau_dp_train(&priv->base, priv->sor.dp,
-						 &outp, head, datarate);
+	u32 pclk = nv_rd32(priv, 0x610ad0 + (head * 0x540)) & 0x3fffff;
+	u32 hval, hreg = 0x614200 + (head * 0x800);
+	u32 oval, oreg;
+	u32 conf = exec_clkcmp(priv, head, 0xff, pclk, &outp);
+	if (conf != ~0) {
+		if (outp.location == 0 && outp.type == DCB_OUTPUT_DP) {
+			u32 soff = (ffs(outp.or) - 1) * 0x08;
+			u32 ctrl = nv_rd32(priv, 0x610798 + soff);
+			u32 datarate;
+
+			switch ((ctrl & 0x000f0000) >> 16) {
+			case 6: datarate = pclk * 30 / 8; break;
+			case 5: datarate = pclk * 24 / 8; break;
+			case 2:
+			default:
+				datarate = pclk * 18 / 8;
+				break;
 			}
 
-			exec_clkcmp(priv, head, 0, pclk, &outp);
-
-			if (!outp.location && outp.type == DCB_OUTPUT_ANALOG) {
-				oreg = 0x614280 + (ffs(outp.or) - 1) * 0x800;
-				oval = 0x00000000;
-				hval = 0x00000000;
-			} else
-			if (!outp.location) {
-				if (outp.type == DCB_OUTPUT_DP)
-					nv50_disp_intr_unk20_dp(priv, &outp, pclk);
-				oreg = 0x614300 + (ffs(outp.or) - 1) * 0x800;
-				oval = (conf & 0x0100) ? 0x0101 : 0x0000;
-				hval = 0x00000000;
-			} else {
-				oreg = 0x614380 + (ffs(outp.or) - 1) * 0x800;
-				oval = 0x00000001;
-				hval = 0x00000001;
-			}
+			nouveau_dp_train(&priv->base, priv->sor.dp,
+					 &outp, head, datarate);
+		}
 
-			nv_mask(priv, hreg, 0x0000000f, hval);
-			nv_mask(priv, oreg, 0x00000707, oval);
+		exec_clkcmp(priv, head, 0, pclk, &outp);
+
+		if (!outp.location && outp.type == DCB_OUTPUT_ANALOG) {
+			oreg = 0x614280 + (ffs(outp.or) - 1) * 0x800;
+			oval = 0x00000000;
+			hval = 0x00000000;
+		} else
+		if (!outp.location) {
+			if (outp.type == DCB_OUTPUT_DP)
+				nv50_disp_intr_unk20_2_dp(priv, &outp, pclk);
+			oreg = 0x614300 + (ffs(outp.or) - 1) * 0x800;
+			oval = (conf & 0x0100) ? 0x00000101 : 0x00000000;
+			hval = 0x00000000;
+		} else {
+			oreg = 0x614380 + (ffs(outp.or) - 1) * 0x800;
+			oval = 0x00000001;
+			hval = 0x00000001;
 		}
-	}
 
-	nv_wr32(priv, 0x610030, 0x80000000);
+		nv_mask(priv, hreg, 0x0000000f, hval);
+		nv_mask(priv, oreg, 0x00000707, oval);
+	}
 }
 
 /* If programming a TMDS output on a SOR that can also be configured for
@@ -1174,7 +1159,7 @@ nv50_disp_intr_unk20(struct nv50_disp_priv *priv, u32 super)
  * programmed for DisplayPort.
  */
 static void
-nv50_disp_intr_unk40_tmds(struct nv50_disp_priv *priv, struct dcb_output *outp)
+nv50_disp_intr_unk40_0_tmds(struct nv50_disp_priv *priv, struct dcb_output *outp)
 {
 	struct nouveau_bios *bios = nouveau_bios(priv);
 	const int link = !(outp->sorconf.link & 1);
@@ -1188,37 +1173,32 @@ nv50_disp_intr_unk40_tmds(struct nv50_disp_priv *priv, struct dcb_output *outp)
 }
 
 static void
-nv50_disp_intr_unk40(struct nv50_disp_priv *priv, u32 super)
+nv50_disp_intr_unk40_0(struct nv50_disp_priv *priv, int head)
 {
-	int head = ffs((super & 0x00000180) >> 7) - 1;
-	if (head >= 0) {
-		struct dcb_output outp;
-		u32 pclk = nv_rd32(priv, 0x610ad0 + (head * 0x540)) & 0x3fffff;
-		if (exec_clkcmp(priv, head, 1, pclk, &outp) != ~0) {
-			if (outp.location == 0 && outp.type == DCB_OUTPUT_TMDS)
-				nv50_disp_intr_unk40_tmds(priv, &outp);
-			else
-			if (outp.location == 1 && outp.type == DCB_OUTPUT_DP) {
-				u32 soff = (ffs(outp.or) - 1) * 0x08;
-				u32 ctrl = nv_rd32(priv, 0x610b84 + soff);
-				u32 datarate;
-
-				switch ((ctrl & 0x000f0000) >> 16) {
-				case 6: datarate = pclk * 30 / 8; break;
-				case 5: datarate = pclk * 24 / 8; break;
-				case 2:
-				default:
-					datarate = pclk * 18 / 8;
-					break;
-				}
-
-				nouveau_dp_train(&priv->base, priv->pior.dp,
-						 &outp, head, datarate);
+	struct dcb_output outp;
+	u32 pclk = nv_rd32(priv, 0x610ad0 + (head * 0x540)) & 0x3fffff;
+	if (exec_clkcmp(priv, head, 1, pclk, &outp) != ~0) {
+		if (outp.location == 0 && outp.type == DCB_OUTPUT_TMDS)
+			nv50_disp_intr_unk40_0_tmds(priv, &outp);
+		else
+		if (outp.location == 1 && outp.type == DCB_OUTPUT_DP) {
+			u32 soff = (ffs(outp.or) - 1) * 0x08;
+			u32 ctrl = nv_rd32(priv, 0x610b84 + soff);
+			u32 datarate;
+
+			switch ((ctrl & 0x000f0000) >> 16) {
+			case 6: datarate = pclk * 30 / 8; break;
+			case 5: datarate = pclk * 24 / 8; break;
+			case 2:
+			default:
+				datarate = pclk * 18 / 8;
+				break;
 			}
+
+			nouveau_dp_train(&priv->base, priv->pior.dp,
+					 &outp, head, datarate);
 		}
 	}
-
-	nv_wr32(priv, 0x610030, 0x80000000);
 }
 
 void
@@ -1227,15 +1207,45 @@ nv50_disp_intr_supervisor(struct work_struct *work)
 	struct nv50_disp_priv *priv =
 		container_of(work, struct nv50_disp_priv, supervisor);
 	u32 super = nv_rd32(priv, 0x610030);
+	int head;
 
 	nv_debug(priv, "supervisor 0x%08x 0x%08x\n", priv->super, super);
 
-	if (priv->super & 0x00000010)
-		nv50_disp_intr_unk10(priv, super);
-	if (priv->super & 0x00000020)
-		nv50_disp_intr_unk20(priv, super);
-	if (priv->super & 0x00000040)
-		nv50_disp_intr_unk40(priv, super);
+	if (priv->super & 0x00000010) {
+		for (head = 0; head < priv->head.nr; head++) {
+			if (!(super & (0x00000020 << head)))
+				continue;
+			if (!(super & (0x00000080 << head)))
+				continue;
+			nv50_disp_intr_unk10_0(priv, head);
+		}
+	} else
+	if (priv->super & 0x00000020) {
+		for (head = 0; head < priv->head.nr; head++) {
+			if (!(super & (0x00000080 << head)))
+				continue;
+			nv50_disp_intr_unk20_0(priv, head);
+		}
+		for (head = 0; head < priv->head.nr; head++) {
+			if (!(super & (0x00000200 << head)))
+				continue;
+			nv50_disp_intr_unk20_1(priv, head);
+		}
+		for (head = 0; head < priv->head.nr; head++) {
+			if (!(super & (0x00000080 << head)))
+				continue;
+			nv50_disp_intr_unk20_2(priv, head);
+		}
+	} else
+	if (priv->super & 0x00000040) {
+		for (head = 0; head < priv->head.nr; head++) {
+			if (!(super & (0x00000080 << head)))
+				continue;
+			nv50_disp_intr_unk40_0(priv, head);
+		}
+	}
+
+	nv_wr32(priv, 0x610030, 0x80000000);
 }
 
 void
-- 
cgit v0.10.2


From 4ea253adf0307dfd9eed3f744a1501dafcfa4de8 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Wed, 20 Feb 2013 19:21:08 +1000
Subject: drm/nvd0/disp: handle multiple actions from one set of supervisor
 intrs

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
index 7106a72..788dd34c 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
@@ -623,13 +623,24 @@ exec_lookup(struct nv50_disp_priv *priv, int head, int outp, u32 ctrl,
 }
 
 static bool
-exec_script(struct nv50_disp_priv *priv, int head, int outp, u32 ctrl, int id)
+exec_script(struct nv50_disp_priv *priv, int head, int id)
 {
 	struct nouveau_bios *bios = nouveau_bios(priv);
 	struct nvbios_outp info;
 	struct dcb_output dcb;
 	u8  ver, hdr, cnt, len;
+	u32 ctrl = 0x00000000;
 	u16 data;
+	int outp;
+
+	for (outp = 0; !(ctrl & (1 << head)) && outp < 8; outp++) {
+		ctrl = nv_rd32(priv, 0x640180 + (outp * 0x20));
+		if (ctrl & (1 << head))
+			break;
+	}
+
+	if (outp == 8)
+		return false;
 
 	data = exec_lookup(priv, head, outp, ctrl, &dcb, &ver, &hdr, &cnt, &len, &info);
 	if (data) {
@@ -649,14 +660,25 @@ exec_script(struct nv50_disp_priv *priv, int head, int outp, u32 ctrl, int id)
 }
 
 static u32
-exec_clkcmp(struct nv50_disp_priv *priv, int head, int outp,
-	    u32 ctrl, int id, u32 pclk, struct dcb_output *dcb)
+exec_clkcmp(struct nv50_disp_priv *priv, int head, int id,
+	    u32 pclk, struct dcb_output *dcb)
 {
 	struct nouveau_bios *bios = nouveau_bios(priv);
 	struct nvbios_outp info1;
 	struct nvbios_ocfg info2;
 	u8  ver, hdr, cnt, len;
+	u32 ctrl = 0x00000000;
 	u32 data, conf = ~0;
+	int outp;
+
+	for (outp = 0; !(ctrl & (1 << head)) && outp < 8; outp++) {
+		ctrl = nv_rd32(priv, 0x660180 + (outp * 0x20));
+		if (ctrl & (1 << head))
+			break;
+	}
+
+	if (outp == 8)
+		return false;
 
 	data = exec_lookup(priv, head, outp, ctrl, dcb, &ver, &hdr, &cnt, &len, &info1);
 	if (data == 0x0000)
@@ -701,24 +723,32 @@ exec_clkcmp(struct nv50_disp_priv *priv, int head, int outp,
 }
 
 static void
-nvd0_display_unk1_handler(struct nv50_disp_priv *priv, u32 head, u32 mask)
+nvd0_disp_intr_unk1_0(struct nv50_disp_priv *priv, int head)
 {
-	int i;
+	exec_script(priv, head, 1);
+}
 
-	for (i = 0; mask && i < 8; i++) {
-		u32 mcc = nv_rd32(priv, 0x640180 + (i * 0x20));
-		if (mcc & (1 << head))
-			exec_script(priv, head, i, mcc, 1);
-	}
+static void
+nvd0_disp_intr_unk2_0(struct nv50_disp_priv *priv, int head)
+{
+	exec_script(priv, head, 2);
+}
 
-	nv_wr32(priv, 0x6101d4, 0x00000000);
-	nv_wr32(priv, 0x6109d4, 0x00000000);
-	nv_wr32(priv, 0x6101d0, 0x80000000);
+static void
+nvd0_disp_intr_unk2_1(struct nv50_disp_priv *priv, int head)
+{
+	struct nouveau_clock *clk = nouveau_clock(priv);
+	u32 pclk = nv_rd32(priv, 0x660450 + (head * 0x300)) / 1000;
+	if (pclk)
+		clk->pll_set(clk, PLL_VPLL0 + head, pclk);
+	nv_wr32(priv, 0x612200 + (head * 0x800), 0x00000000);
 }
 
 static void
-nvd0_display_unk2_calc_tu(struct nv50_disp_priv *priv, int head, int or)
+nvd0_disp_intr_unk2_2_tu(struct nv50_disp_priv *priv, int head,
+			 struct dcb_output *outp)
 {
+	const int or = ffs(outp->or) - 1;
 	const u32 ctrl = nv_rd32(priv, 0x660200 + (or   * 0x020));
 	const u32 conf = nv_rd32(priv, 0x660404 + (head * 0x300));
 	const u32 pclk = nv_rd32(priv, 0x660450 + (head * 0x300)) / 1000;
@@ -761,97 +791,51 @@ nvd0_display_unk2_calc_tu(struct nv50_disp_priv *priv, int head, int or)
 }
 
 static void
-nvd0_display_unk2_handler(struct nv50_disp_priv *priv, u32 head, u32 mask)
+nvd0_disp_intr_unk2_2(struct nv50_disp_priv *priv, int head)
 {
 	struct dcb_output outp;
-	u32 pclk;
-	int i;
-
-	for (i = 0; mask && i < 8; i++) {
-		u32 mcc = nv_rd32(priv, 0x640180 + (i * 0x20));
-		if (mcc & (1 << head))
-			exec_script(priv, head, i, mcc, 2);
-	}
+	u32 pclk = nv_rd32(priv, 0x660450 + (head * 0x300)) / 1000;
+	u32 conf = exec_clkcmp(priv, head, 0xff, pclk, &outp);
+	if (conf != ~0) {
+		u32 addr, data;
+
+		if (outp.type == DCB_OUTPUT_DP) {
+			u32 sync = nv_rd32(priv, 0x660404 + (head * 0x300));
+			switch ((sync & 0x000003c0) >> 6) {
+			case 6: pclk = pclk * 30 / 8; break;
+			case 5: pclk = pclk * 24 / 8; break;
+			case 2:
+			default:
+				pclk = pclk * 18 / 8;
+				break;
+			}
 
-	pclk = nv_rd32(priv, 0x660450 + (head * 0x300)) / 1000;
-	nv_debug(priv, "head %d pclk %d mask 0x%08x\n", head, pclk, mask);
-	if (pclk && (mask & 0x00010000)) {
-		struct nouveau_clock *clk = nouveau_clock(priv);
-		clk->pll_set(clk, PLL_VPLL0 + head, pclk);
-	}
+			nouveau_dp_train(&priv->base, priv->sor.dp,
+					 &outp, head, pclk);
+		}
 
-	nv_wr32(priv, 0x612200 + (head * 0x800), 0x00000000);
+		exec_clkcmp(priv, head, 0, pclk, &outp);
 
-	for (i = 0; mask && i < 8; i++) {
-		u32 mcp = nv_rd32(priv, 0x660180 + (i * 0x20));
-		if (mcp & (1 << head)) {
-			u32 cfg = exec_clkcmp(priv, head, i, mcp, 0xff, pclk, &outp);
-			if (cfg != ~0) {
-				u32 addr, mask, data = 0x00000000;
-
-				if (outp.type == DCB_OUTPUT_DP) {
-					switch ((mcp & 0x000f0000) >> 16) {
-					case 6: pclk = pclk * 30 / 8; break;
-					case 5: pclk = pclk * 24 / 8; break;
-					case 2:
-					default:
-						pclk = pclk * 18 / 8;
-						break;
-					}
-
-					nouveau_dp_train(&priv->base,
-							  priv->sor.dp,
-							 &outp, head, pclk);
-				}
-
-				exec_clkcmp(priv, head, i, mcp, 0, pclk, &outp);
-
-				if (i < 4) {
-					addr = 0x612280 + ((i - 0) * 0x800);
-					mask = 0xffffffff;
-				} else {
-					switch (mcp & 0x00000f00) {
-					case 0x00000800:
-					case 0x00000900:
-						nvd0_display_unk2_calc_tu(priv, head, i - 4);
-						break;
-					default:
-						break;
-					}
-
-					addr = 0x612300 + ((i - 4) * 0x800);
-					mask = 0x00000707;
-					if (cfg & 0x00000100)
-						data = 0x00000101;
-				}
-				nv_mask(priv, addr, mask, data);
-			}
-			break;
+		if (outp.type == DCB_OUTPUT_ANALOG) {
+			addr = 0x612280 + (ffs(outp.or) - 1) * 0x800;
+			data = 0x00000000;
+		} else {
+			if (outp.type == DCB_OUTPUT_DP)
+				nvd0_disp_intr_unk2_2_tu(priv, head, &outp);
+			addr = 0x612300 + (ffs(outp.or) - 1) * 0x800;
+			data = (conf & 0x0100) ? 0x00000101 : 0x00000000;
 		}
-	}
 
-	nv_wr32(priv, 0x6101d4, 0x00000000);
-	nv_wr32(priv, 0x6109d4, 0x00000000);
-	nv_wr32(priv, 0x6101d0, 0x80000000);
+		nv_mask(priv, addr, 0x00000707, data);
+	}
 }
 
 static void
-nvd0_display_unk4_handler(struct nv50_disp_priv *priv, u32 head, u32 mask)
+nvd0_disp_intr_unk4_0(struct nv50_disp_priv *priv, int head)
 {
 	struct dcb_output outp;
-	int pclk, i;
-
-	pclk = nv_rd32(priv, 0x660450 + (head * 0x300)) / 1000;
-
-	for (i = 0; mask && i < 8; i++) {
-		u32 mcp = nv_rd32(priv, 0x660180 + (i * 0x20));
-		if (mcp & (1 << head))
-			exec_clkcmp(priv, head, i, mcp, 1, pclk, &outp);
-	}
-
-	nv_wr32(priv, 0x6101d4, 0x00000000);
-	nv_wr32(priv, 0x6109d4, 0x00000000);
-	nv_wr32(priv, 0x6101d0, 0x80000000);
+	u32 pclk = nv_rd32(priv, 0x660450 + (head * 0x300)) / 1000;
+	exec_clkcmp(priv, head, 1, pclk, &outp);
 }
 
 void
@@ -859,19 +843,50 @@ nvd0_disp_intr_supervisor(struct work_struct *work)
 {
 	struct nv50_disp_priv *priv =
 		container_of(work, struct nv50_disp_priv, supervisor);
-	u32 mask = 0, head = ~0;
+	u32 mask[4];
+	int head;
 
-	while (!mask && ++head < priv->head.nr)
-		mask = nv_rd32(priv, 0x6101d4 + (head * 0x800));
+	nv_debug(priv, "supervisor %08x\n", priv->super);
+	for (head = 0; head < priv->head.nr; head++) {
+		mask[head] = nv_rd32(priv, 0x6101d4 + (head * 0x800));
+		nv_debug(priv, "head %d: 0x%08x\n", head, mask[head]);
+	}
 
-	nv_debug(priv, "supervisor %08x %08x %d\n", priv->super, mask, head);
+	if (priv->super & 0x00000001) {
+		for (head = 0; head < priv->head.nr; head++) {
+			if (!(mask[head] & 0x00001000))
+				continue;
+			nvd0_disp_intr_unk1_0(priv, head);
+		}
+	} else
+	if (priv->super & 0x00000002) {
+		for (head = 0; head < priv->head.nr; head++) {
+			if (!(mask[head] & 0x00001000))
+				continue;
+			nvd0_disp_intr_unk2_0(priv, head);
+		}
+		for (head = 0; head < priv->head.nr; head++) {
+			if (!(mask[head] & 0x00010000))
+				continue;
+			nvd0_disp_intr_unk2_1(priv, head);
+		}
+		for (head = 0; head < priv->head.nr; head++) {
+			if (!(mask[head] & 0x00001000))
+				continue;
+			nvd0_disp_intr_unk2_2(priv, head);
+		}
+	} else
+	if (priv->super & 0x00000004) {
+		for (head = 0; head < priv->head.nr; head++) {
+			if (!(mask[head] & 0x00001000))
+				continue;
+			nvd0_disp_intr_unk4_0(priv, head);
+		}
+	}
 
-	if (priv->super & 0x00000001)
-		nvd0_display_unk1_handler(priv, head, mask);
-	if (priv->super & 0x00000002)
-		nvd0_display_unk2_handler(priv, head, mask);
-	if (priv->super & 0x00000004)
-		nvd0_display_unk4_handler(priv, head, mask);
+	for (head = 0; head < priv->head.nr; head++)
+		nv_wr32(priv, 0x6101d4 + (head * 0x800), 0x00000000);
+	nv_wr32(priv, 0x6101d0, 0x80000000);
 }
 
 void
-- 
cgit v0.10.2


From eff76ed7881842ffcd2d77cf217eabcb1134b38b Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Wed, 20 Feb 2013 20:37:15 +1000
Subject: drm/nv50-/kms: remove UPDATE methods after each encoder disconnect

Supervisor can now handle more than one operation within a single
series of interrupts.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

diff --git a/drivers/gpu/drm/nouveau/nv50_display.c b/drivers/gpu/drm/nouveau/nv50_display.c
index 49f26cd..d903cc0 100644
--- a/drivers/gpu/drm/nouveau/nv50_display.c
+++ b/drivers/gpu/drm/nouveau/nv50_display.c
@@ -1509,9 +1509,6 @@ nv50_dac_disconnect(struct drm_encoder *encoder)
 				evo_mthd(push, 0x0180 + (or * 0x020), 1);
 				evo_data(push, 0x00000000);
 			}
-
-			evo_mthd(push, 0x0080, 1);
-			evo_data(push, 0x00000000);
 			evo_kick(push, mast);
 		}
 	}
@@ -1725,9 +1722,6 @@ nv50_sor_disconnect(struct drm_encoder *encoder)
 				evo_mthd(push, 0x0200 + (or * 0x20), 1);
 				evo_data(push, 0x00000000);
 			}
-
-			evo_mthd(push, 0x0080, 1);
-			evo_data(push, 0x00000000);
 			evo_kick(push, mast);
 		}
 
@@ -2029,9 +2023,6 @@ nv50_pior_disconnect(struct drm_encoder *encoder)
 				evo_mthd(push, 0x0700 + (or * 0x040), 1);
 				evo_data(push, 0x00000000);
 			}
-
-			evo_mthd(push, 0x0080, 1);
-			evo_data(push, 0x00000000);
 			evo_kick(push, mast);
 		}
 	}
-- 
cgit v0.10.2


From 33b903e82873ba9233bd89c44d016b87347ed158 Mon Sep 17 00:00:00 2001
From: Marcin Slusarz <marcin.slusarz@gmail.com>
Date: Fri, 8 Feb 2013 21:42:13 +0100
Subject: drm/nouveau: restore debugfs/vbios.rom support

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>

diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index 51d0972..90f9140 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -242,5 +242,6 @@ ifdef CONFIG_X86
 nouveau-$(CONFIG_ACPI) += nouveau_acpi.o
 endif
 nouveau-$(CONFIG_DRM_NOUVEAU_BACKLIGHT) += nouveau_backlight.o
+nouveau-$(CONFIG_DEBUG_FS) += nouveau_debugfs.o
 
 obj-$(CONFIG_DRM_NOUVEAU)+= nouveau.o
diff --git a/drivers/gpu/drm/nouveau/nouveau_debugfs.c b/drivers/gpu/drm/nouveau/nouveau_debugfs.c
new file mode 100644
index 0000000..5392e07
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nouveau_debugfs.c
@@ -0,0 +1,64 @@
+/*
+ * Copyright (C) 2009 Red Hat <bskeggs@redhat.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining
+ * a copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sublicense, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial
+ * portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ * IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ * LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ * OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ * WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+/*
+ * Authors:
+ *  Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nouveau_debugfs.h"
+#include "nouveau_drm.h"
+
+static int
+nouveau_debugfs_vbios_image(struct seq_file *m, void *data)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct nouveau_drm *drm = nouveau_drm(node->minor->dev);
+	int i;
+
+	for (i = 0; i < drm->vbios.length; i++)
+		seq_printf(m, "%c", drm->vbios.data[i]);
+	return 0;
+}
+
+static struct drm_info_list nouveau_debugfs_list[] = {
+	{ "vbios.rom", nouveau_debugfs_vbios_image, 0, NULL },
+};
+#define NOUVEAU_DEBUGFS_ENTRIES ARRAY_SIZE(nouveau_debugfs_list)
+
+int
+nouveau_debugfs_init(struct drm_minor *minor)
+{
+	drm_debugfs_create_files(nouveau_debugfs_list, NOUVEAU_DEBUGFS_ENTRIES,
+				 minor->debugfs_root, minor);
+	return 0;
+}
+
+void
+nouveau_debugfs_takedown(struct drm_minor *minor)
+{
+	drm_debugfs_remove_files(nouveau_debugfs_list, NOUVEAU_DEBUGFS_ENTRIES,
+				 minor);
+}
diff --git a/drivers/gpu/drm/nouveau/nouveau_debugfs.h b/drivers/gpu/drm/nouveau/nouveau_debugfs.h
new file mode 100644
index 0000000..a62af6f
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nouveau_debugfs.h
@@ -0,0 +1,22 @@
+#ifndef __NOUVEAU_DEBUGFS_H__
+#define __NOUVEAU_DEBUGFS_H__
+
+#include <drm/drmP.h>
+
+#if defined(CONFIG_DEBUG_FS)
+extern int  nouveau_debugfs_init(struct drm_minor *);
+extern void nouveau_debugfs_takedown(struct drm_minor *);
+#else
+static inline int
+nouveau_debugfs_init(struct drm_minor *minor)
+{
+       return 0;
+}
+
+static inline void nouveau_debugfs_takedown(struct drm_minor *minor)
+{
+}
+
+#endif
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c
index f914491..b1c4ce1 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_drm.c
@@ -50,6 +50,7 @@
 #include "nouveau_abi16.h"
 #include "nouveau_fbcon.h"
 #include "nouveau_fence.h"
+#include "nouveau_debugfs.h"
 
 MODULE_PARM_DESC(config, "option string to pass to driver core");
 static char *nouveau_config;
@@ -667,6 +668,11 @@ driver = {
 	.postclose = nouveau_drm_postclose,
 	.lastclose = nouveau_vga_lastclose,
 
+#if defined(CONFIG_DEBUG_FS)
+	.debugfs_init = nouveau_debugfs_init,
+	.debugfs_cleanup = nouveau_debugfs_takedown,
+#endif
+
 	.irq_preinstall = nouveau_irq_preinstall,
 	.irq_postinstall = nouveau_irq_postinstall,
 	.irq_uninstall = nouveau_irq_uninstall,
-- 
cgit v0.10.2


From 7ae764b11ed63279e9dcf25be972ff4ca21a9875 Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Mon, 11 Feb 2013 08:44:48 -0500
Subject: drm/radeon: fix multi-head power profile stability on BTC+ asics

vddci needs to track mclk for multi-head.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c
index 2916de8..3c38ea4 100644
--- a/drivers/gpu/drm/radeon/evergreen.c
+++ b/drivers/gpu/drm/radeon/evergreen.c
@@ -403,6 +403,19 @@ void evergreen_pm_misc(struct radeon_device *rdev)
 			rdev->pm.current_vddc = voltage->voltage;
 			DRM_DEBUG("Setting: vddc: %d\n", voltage->voltage);
 		}
+
+		/* starting with BTC, there is one state that is used for both
+		 * MH and SH.  Difference is that we always use the high clock index for
+		 * mclk and vddci.
+		 */
+		if ((rdev->pm.pm_method == PM_METHOD_PROFILE) &&
+		    (rdev->family >= CHIP_BARTS) &&
+		    rdev->pm.active_crtc_count &&
+		    ((rdev->pm.profile_index == PM_PROFILE_MID_MH_IDX) ||
+		     (rdev->pm.profile_index == PM_PROFILE_LOW_MH_IDX)))
+			voltage = &rdev->pm.power_state[req_ps_idx].
+				clock_info[rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_on_cm_idx].voltage;
+
 		/* 0xff01 is a flag rather then an actual voltage */
 		if (voltage->vddci == 0xff01)
 			return;
diff --git a/drivers/gpu/drm/radeon/radeon_pm.c b/drivers/gpu/drm/radeon/radeon_pm.c
index 0bfa656..338fd6a 100644
--- a/drivers/gpu/drm/radeon/radeon_pm.c
+++ b/drivers/gpu/drm/radeon/radeon_pm.c
@@ -169,7 +169,7 @@ static void radeon_set_power_state(struct radeon_device *rdev)
 
 		/* starting with BTC, there is one state that is used for both
 		 * MH and SH.  Difference is that we always use the high clock index for
-		 * mclk.
+		 * mclk and vddci.
 		 */
 		if ((rdev->pm.pm_method == PM_METHOD_PROFILE) &&
 		    (rdev->family >= CHIP_BARTS) &&
-- 
cgit v0.10.2


From c944b2abb067130542055666f23409fd5e1afc8e Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Tue, 12 Feb 2013 08:39:10 -0500
Subject: drm/radeon: remove overzealous warning in hdmi handling

hdmi audio works fine.  The warning just confuses users.

fixes:
https://bugzilla.kernel.org/show_bug.cgi?id=44341

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Cc: stable@vger.kernel.org

diff --git a/drivers/gpu/drm/radeon/r600_hdmi.c b/drivers/gpu/drm/radeon/r600_hdmi.c
index ff80efe..95970ec 100644
--- a/drivers/gpu/drm/radeon/r600_hdmi.c
+++ b/drivers/gpu/drm/radeon/r600_hdmi.c
@@ -544,7 +544,6 @@ void r600_hdmi_disable(struct drm_encoder *encoder)
 
 	/* Called for ATOM_ENCODER_MODE_HDMI only */
 	if (!dig || !dig->afmt) {
-		WARN_ON(1);
 		return;
 	}
 	if (!dig->afmt->enabled)
-- 
cgit v0.10.2


From 0e34d0945ed16aeb72c863be4efe4cd0c62f9c53 Mon Sep 17 00:00:00 2001
From: Tim Gardner <tim.gardner@canonical.com>
Date: Mon, 11 Feb 2013 14:34:32 -0700
Subject: drm/radeon: Avoid NULL pointer dereference from atom_index_iio()
 allocation failure
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Smatch anlysis:

drivers/gpu/drm/radeon/atom.c:1242 atom_index_iio() error: potential null
 dereference 'ctx->iio'.  (kzalloc returns null)

Also cleaned up some checks before calls to kfree(). kfree(NULL) is OK.

Cc: David Airlie <airlied@linux.ie>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: "Michel Dänzer" <michel.daenzer@amd.com>
Cc: Dave Airlie <airlied@redhat.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: Jerome Glisse <jglisse@redhat.com>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/atom.c b/drivers/gpu/drm/radeon/atom.c
index 5ce9bf5..46a9c37 100644
--- a/drivers/gpu/drm/radeon/atom.c
+++ b/drivers/gpu/drm/radeon/atom.c
@@ -1238,6 +1238,8 @@ static int atom_iio_len[] = { 1, 2, 3, 3, 3, 3, 4, 4, 4, 3 };
 static void atom_index_iio(struct atom_context *ctx, int base)
 {
 	ctx->iio = kzalloc(2 * 256, GFP_KERNEL);
+	if (!ctx->iio)
+		return;
 	while (CU8(base) == ATOM_IIO_START) {
 		ctx->iio[CU8(base + 1)] = base + 2;
 		base += 2;
@@ -1287,6 +1289,10 @@ struct atom_context *atom_parse(struct card_info *card, void *bios)
 	ctx->cmd_table = CU16(base + ATOM_ROM_CMD_PTR);
 	ctx->data_table = CU16(base + ATOM_ROM_DATA_PTR);
 	atom_index_iio(ctx, CU16(ctx->data_table + ATOM_DATA_IIO_PTR) + 4);
+	if (!ctx->iio) {
+		atom_destroy(ctx);
+		return NULL;
+	}
 
 	str = CSTR(CU16(base + ATOM_ROM_MSG_PTR));
 	while (*str && ((*str == '\n') || (*str == '\r')))
@@ -1335,8 +1341,7 @@ int atom_asic_init(struct atom_context *ctx)
 
 void atom_destroy(struct atom_context *ctx)
 {
-	if (ctx->iio)
-		kfree(ctx->iio);
+	kfree(ctx->iio);
 	kfree(ctx);
 }
 
diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c
index 6a36f4b..9f02fdd 100644
--- a/drivers/gpu/drm/radeon/radeon_device.c
+++ b/drivers/gpu/drm/radeon/radeon_device.c
@@ -758,6 +758,11 @@ int radeon_atombios_init(struct radeon_device *rdev)
 	atom_card_info->pll_write = cail_pll_write;
 
 	rdev->mode_info.atom_context = atom_parse(atom_card_info, rdev->bios);
+	if (!rdev->mode_info.atom_context) {
+		radeon_atombios_fini(rdev);
+		return -ENOMEM;
+	}
+
 	mutex_init(&rdev->mode_info.atom_context->mutex);
 	radeon_atom_initialize_bios_scratch_regs(rdev->ddev);
 	atom_allocate_fb_scratch(rdev->mode_info.atom_context);
@@ -777,9 +782,11 @@ void radeon_atombios_fini(struct radeon_device *rdev)
 {
 	if (rdev->mode_info.atom_context) {
 		kfree(rdev->mode_info.atom_context->scratch);
-		kfree(rdev->mode_info.atom_context);
 	}
+	kfree(rdev->mode_info.atom_context);
+	rdev->mode_info.atom_context = NULL;
 	kfree(rdev->mode_info.atom_card_info);
+	rdev->mode_info.atom_card_info = NULL;
 }
 
 /* COMBIOS */
-- 
cgit v0.10.2


From 454d2e2a32f02a4f8e0364ba0e8ce4f8dbe9810c Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Thu, 14 Feb 2013 10:04:02 -0500
Subject: drm/radeon: add a asic callback to get the xclk

This is required to get the reference clock used
by the gfx engine for things like timestamps. Fixes
support for GL extensions the use timestamps on
certain boards.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c
index dbcb075..37d81fd 100644
--- a/drivers/gpu/drm/radeon/r600.c
+++ b/drivers/gpu/drm/radeon/r600.c
@@ -109,6 +109,19 @@ void r600_fini(struct radeon_device *rdev);
 void r600_irq_disable(struct radeon_device *rdev);
 static void r600_pcie_gen2_enable(struct radeon_device *rdev);
 
+/**
+ * r600_get_xclk - get the xclk
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Returns the reference clock used by the gfx engine
+ * (r6xx, IGPs, APUs).
+ */
+u32 r600_get_xclk(struct radeon_device *rdev)
+{
+	return rdev->clock.spll.reference_freq;
+}
+
 /* get temperature in millidegrees */
 int rv6xx_get_temp(struct radeon_device *rdev)
 {
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 59bfbd3..e425b41 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -1179,6 +1179,8 @@ struct radeon_asic {
 	bool (*gui_idle)(struct radeon_device *rdev);
 	/* wait for mc_idle */
 	int (*mc_wait_for_idle)(struct radeon_device *rdev);
+	/* get the reference clock */
+	u32 (*get_xclk)(struct radeon_device *rdev);
 	/* gart */
 	struct {
 		void (*tlb_flush)(struct radeon_device *rdev);
@@ -1860,6 +1862,7 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v);
 #define radeon_post_page_flip(rdev, crtc) (rdev)->asic->pflip.post_page_flip((rdev), (crtc))
 #define radeon_wait_for_vblank(rdev, crtc) (rdev)->asic->display.wait_for_vblank((rdev), (crtc))
 #define radeon_mc_wait_for_idle(rdev) (rdev)->asic->mc_wait_for_idle((rdev))
+#define radeon_get_xclk(rdev) (rdev)->asic->get_xclk((rdev))
 
 /* Common functions */
 /* AGP */
diff --git a/drivers/gpu/drm/radeon/radeon_asic.c b/drivers/gpu/drm/radeon/radeon_asic.c
index 67f008f..e1b4a68 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.c
+++ b/drivers/gpu/drm/radeon/radeon_asic.c
@@ -934,6 +934,7 @@ static struct radeon_asic r600_asic = {
 	.ioctl_wait_idle = r600_ioctl_wait_idle,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &r600_mc_wait_for_idle,
+	.get_xclk = &r600_get_xclk,
 	.gart = {
 		.tlb_flush = &r600_pcie_gart_tlb_flush,
 		.set_page = &rs600_gart_set_page,
@@ -1018,6 +1019,7 @@ static struct radeon_asic rs780_asic = {
 	.ioctl_wait_idle = r600_ioctl_wait_idle,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &r600_mc_wait_for_idle,
+	.get_xclk = &r600_get_xclk,
 	.gart = {
 		.tlb_flush = &r600_pcie_gart_tlb_flush,
 		.set_page = &rs600_gart_set_page,
@@ -1102,6 +1104,7 @@ static struct radeon_asic rv770_asic = {
 	.ioctl_wait_idle = r600_ioctl_wait_idle,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &r600_mc_wait_for_idle,
+	.get_xclk = &rv770_get_xclk,
 	.gart = {
 		.tlb_flush = &r600_pcie_gart_tlb_flush,
 		.set_page = &rs600_gart_set_page,
@@ -1186,6 +1189,7 @@ static struct radeon_asic evergreen_asic = {
 	.ioctl_wait_idle = r600_ioctl_wait_idle,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
+	.get_xclk = &rv770_get_xclk,
 	.gart = {
 		.tlb_flush = &evergreen_pcie_gart_tlb_flush,
 		.set_page = &rs600_gart_set_page,
@@ -1270,6 +1274,7 @@ static struct radeon_asic sumo_asic = {
 	.ioctl_wait_idle = r600_ioctl_wait_idle,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
+	.get_xclk = &r600_get_xclk,
 	.gart = {
 		.tlb_flush = &evergreen_pcie_gart_tlb_flush,
 		.set_page = &rs600_gart_set_page,
@@ -1354,6 +1359,7 @@ static struct radeon_asic btc_asic = {
 	.ioctl_wait_idle = r600_ioctl_wait_idle,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
+	.get_xclk = &rv770_get_xclk,
 	.gart = {
 		.tlb_flush = &evergreen_pcie_gart_tlb_flush,
 		.set_page = &rs600_gart_set_page,
@@ -1438,6 +1444,7 @@ static struct radeon_asic cayman_asic = {
 	.ioctl_wait_idle = r600_ioctl_wait_idle,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
+	.get_xclk = &rv770_get_xclk,
 	.gart = {
 		.tlb_flush = &cayman_pcie_gart_tlb_flush,
 		.set_page = &rs600_gart_set_page,
@@ -1565,6 +1572,7 @@ static struct radeon_asic trinity_asic = {
 	.ioctl_wait_idle = r600_ioctl_wait_idle,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
+	.get_xclk = &r600_get_xclk,
 	.gart = {
 		.tlb_flush = &cayman_pcie_gart_tlb_flush,
 		.set_page = &rs600_gart_set_page,
@@ -1692,6 +1700,7 @@ static struct radeon_asic si_asic = {
 	.ioctl_wait_idle = r600_ioctl_wait_idle,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
+	.get_xclk = &si_get_xclk,
 	.gart = {
 		.tlb_flush = &si_pcie_gart_tlb_flush,
 		.set_page = &rs600_gart_set_page,
diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h
index f4134a8..f15758c 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.h
+++ b/drivers/gpu/drm/radeon/radeon_asic.h
@@ -390,6 +390,7 @@ void r600_kms_blit_copy(struct radeon_device *rdev,
 			struct radeon_sa_bo *vb);
 int r600_mc_wait_for_idle(struct radeon_device *rdev);
 uint64_t r600_get_gpu_clock(struct radeon_device *rdev);
+u32 r600_get_xclk(struct radeon_device *rdev);
 
 /*
  * rv770,rv730,rv710,rv740
@@ -407,6 +408,7 @@ int rv770_copy_dma(struct radeon_device *rdev,
 		  uint64_t src_offset, uint64_t dst_offset,
 		  unsigned num_gpu_pages,
 		   struct radeon_fence **fence);
+u32 rv770_get_xclk(struct radeon_device *rdev);
 
 /*
  * evergreen
@@ -521,5 +523,6 @@ int si_copy_dma(struct radeon_device *rdev,
 		unsigned num_gpu_pages,
 		struct radeon_fence **fence);
 void si_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
+u32 si_get_xclk(struct radeon_device *rdev);
 
 #endif
diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index 9c312f9..96f05cd 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -282,7 +282,10 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		break;
 	case RADEON_INFO_CLOCK_CRYSTAL_FREQ:
 		/* return clock value in KHz */
-		value = rdev->clock.spll.reference_freq * 10;
+		if (rdev->asic->get_xclk)
+			value = radeon_get_xclk(rdev) * 10;
+		else
+			value = rdev->clock.spll.reference_freq * 10;
 		break;
 	case RADEON_INFO_NUM_BACKENDS:
 		if (rdev->family >= CHIP_TAHITI)
diff --git a/drivers/gpu/drm/radeon/rv770.c b/drivers/gpu/drm/radeon/rv770.c
index 1b2444f..d63fe1d 100644
--- a/drivers/gpu/drm/radeon/rv770.c
+++ b/drivers/gpu/drm/radeon/rv770.c
@@ -43,6 +43,31 @@ static void rv770_gpu_init(struct radeon_device *rdev);
 void rv770_fini(struct radeon_device *rdev);
 static void rv770_pcie_gen2_enable(struct radeon_device *rdev);
 
+#define PCIE_BUS_CLK                10000
+#define TCLK                        (PCIE_BUS_CLK / 10)
+
+/**
+ * rv770_get_xclk - get the xclk
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Returns the reference clock used by the gfx engine
+ * (r7xx-cayman).
+ */
+u32 rv770_get_xclk(struct radeon_device *rdev)
+{
+	u32 reference_clock = rdev->clock.spll.reference_freq;
+	u32 tmp = RREG32(CG_CLKPIN_CNTL);
+
+	if (tmp & MUX_TCLK_TO_XCLK)
+		return TCLK;
+
+	if (tmp & XTALIN_DIVIDE)
+		return reference_clock / 4;
+
+	return reference_clock;
+}
+
 u32 rv770_page_flip(struct radeon_device *rdev, int crtc_id, u64 crtc_base)
 {
 	struct radeon_crtc *radeon_crtc = rdev->mode_info.crtcs[crtc_id];
diff --git a/drivers/gpu/drm/radeon/rv770d.h b/drivers/gpu/drm/radeon/rv770d.h
index 20e29d2..c55f950 100644
--- a/drivers/gpu/drm/radeon/rv770d.h
+++ b/drivers/gpu/drm/radeon/rv770d.h
@@ -128,6 +128,10 @@
 #define		GUI_ACTIVE					(1<<31)
 #define	GRBM_STATUS2					0x8014
 
+#define CG_CLKPIN_CNTL                                    0x660
+#       define MUX_TCLK_TO_XCLK                           (1 << 8)
+#       define XTALIN_DIVIDE                              (1 << 9)
+
 #define	CG_MULT_THERMAL_STATUS				0x740
 #define		ASIC_T(x)			        ((x) << 16)
 #define		ASIC_T_MASK			        0x3FF0000
diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
index 719f03e..b5064fa 100644
--- a/drivers/gpu/drm/radeon/si.c
+++ b/drivers/gpu/drm/radeon/si.c
@@ -70,6 +70,33 @@ extern u32 evergreen_get_number_of_dram_channels(struct radeon_device *rdev);
 extern void evergreen_print_gpu_status_regs(struct radeon_device *rdev);
 extern bool evergreen_is_display_hung(struct radeon_device *rdev);
 
+#define PCIE_BUS_CLK                10000
+#define TCLK                        (PCIE_BUS_CLK / 10)
+
+/**
+ * si_get_xclk - get the xclk
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Returns the reference clock used by the gfx engine
+ * (SI).
+ */
+u32 si_get_xclk(struct radeon_device *rdev)
+{
+        u32 reference_clock = rdev->clock.spll.reference_freq;
+	u32 tmp;
+
+	tmp = RREG32(CG_CLKPIN_CNTL_2);
+	if (tmp & MUX_TCLK_TO_XCLK)
+		return TCLK;
+
+	tmp = RREG32(CG_CLKPIN_CNTL);
+	if (tmp & XTALIN_DIVIDE)
+		return reference_clock / 4;
+
+	return reference_clock;
+}
+
 /* get temperature in millidegrees */
 int si_get_temp(struct radeon_device *rdev)
 {
diff --git a/drivers/gpu/drm/radeon/sid.h b/drivers/gpu/drm/radeon/sid.h
index 07fc455..23fc08f 100644
--- a/drivers/gpu/drm/radeon/sid.h
+++ b/drivers/gpu/drm/radeon/sid.h
@@ -58,6 +58,11 @@
 #define VGA_HDP_CONTROL  				0x328
 #define		VGA_MEMORY_DISABLE				(1 << 4)
 
+#define CG_CLKPIN_CNTL                                    0x660
+#       define XTALIN_DIVIDE                              (1 << 1)
+#define CG_CLKPIN_CNTL_2                                  0x664
+#       define MUX_TCLK_TO_XCLK                           (1 << 8)
+
 #define DMIF_ADDR_CONFIG  				0xBD4
 
 #define	SRBM_STATUS				        0xE50
-- 
cgit v0.10.2


From d0418894142f88041d2b7b5aa4f8bf3178d89514 Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Thu, 24 Jan 2013 10:35:23 -0500
Subject: drm/radeon: switch get_gpu_clock() to a callback (v2)

Cleans up the code for future asics

v2: rebase, fix some missing radeon_asic updates

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c
index 37d81fd..6d4b561 100644
--- a/drivers/gpu/drm/radeon/r600.c
+++ b/drivers/gpu/drm/radeon/r600.c
@@ -4461,14 +4461,14 @@ static void r600_pcie_gen2_enable(struct radeon_device *rdev)
 }
 
 /**
- * r600_get_gpu_clock - return GPU clock counter snapshot
+ * r600_get_gpu_clock_counter - return GPU clock counter snapshot
  *
  * @rdev: radeon_device pointer
  *
  * Fetches a GPU clock counter snapshot (R6xx-cayman).
  * Returns the 64 bit clock counter snapshot.
  */
-uint64_t r600_get_gpu_clock(struct radeon_device *rdev)
+uint64_t r600_get_gpu_clock_counter(struct radeon_device *rdev)
 {
 	uint64_t clock;
 
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index e425b41..081e55f 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -1181,6 +1181,8 @@ struct radeon_asic {
 	int (*mc_wait_for_idle)(struct radeon_device *rdev);
 	/* get the reference clock */
 	u32 (*get_xclk)(struct radeon_device *rdev);
+	/* get the gpu clock counter */
+	uint64_t (*get_gpu_clock_counter)(struct radeon_device *rdev);
 	/* gart */
 	struct {
 		void (*tlb_flush)(struct radeon_device *rdev);
@@ -1863,6 +1865,7 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v);
 #define radeon_wait_for_vblank(rdev, crtc) (rdev)->asic->display.wait_for_vblank((rdev), (crtc))
 #define radeon_mc_wait_for_idle(rdev) (rdev)->asic->mc_wait_for_idle((rdev))
 #define radeon_get_xclk(rdev) (rdev)->asic->get_xclk((rdev))
+#define radeon_get_gpu_clock_counter(rdev) (rdev)->asic->get_gpu_clock_counter((rdev))
 
 /* Common functions */
 /* AGP */
diff --git a/drivers/gpu/drm/radeon/radeon_asic.c b/drivers/gpu/drm/radeon/radeon_asic.c
index e1b4a68..aba0a89 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.c
+++ b/drivers/gpu/drm/radeon/radeon_asic.c
@@ -935,6 +935,7 @@ static struct radeon_asic r600_asic = {
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &r600_mc_wait_for_idle,
 	.get_xclk = &r600_get_xclk,
+	.get_gpu_clock_counter = &r600_get_gpu_clock_counter,
 	.gart = {
 		.tlb_flush = &r600_pcie_gart_tlb_flush,
 		.set_page = &rs600_gart_set_page,
@@ -1020,6 +1021,7 @@ static struct radeon_asic rs780_asic = {
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &r600_mc_wait_for_idle,
 	.get_xclk = &r600_get_xclk,
+	.get_gpu_clock_counter = &r600_get_gpu_clock_counter,
 	.gart = {
 		.tlb_flush = &r600_pcie_gart_tlb_flush,
 		.set_page = &rs600_gart_set_page,
@@ -1105,6 +1107,7 @@ static struct radeon_asic rv770_asic = {
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &r600_mc_wait_for_idle,
 	.get_xclk = &rv770_get_xclk,
+	.get_gpu_clock_counter = &r600_get_gpu_clock_counter,
 	.gart = {
 		.tlb_flush = &r600_pcie_gart_tlb_flush,
 		.set_page = &rs600_gart_set_page,
@@ -1190,6 +1193,7 @@ static struct radeon_asic evergreen_asic = {
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
 	.get_xclk = &rv770_get_xclk,
+	.get_gpu_clock_counter = &r600_get_gpu_clock_counter,
 	.gart = {
 		.tlb_flush = &evergreen_pcie_gart_tlb_flush,
 		.set_page = &rs600_gart_set_page,
@@ -1275,6 +1279,7 @@ static struct radeon_asic sumo_asic = {
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
 	.get_xclk = &r600_get_xclk,
+	.get_gpu_clock_counter = &r600_get_gpu_clock_counter,
 	.gart = {
 		.tlb_flush = &evergreen_pcie_gart_tlb_flush,
 		.set_page = &rs600_gart_set_page,
@@ -1360,6 +1365,7 @@ static struct radeon_asic btc_asic = {
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
 	.get_xclk = &rv770_get_xclk,
+	.get_gpu_clock_counter = &r600_get_gpu_clock_counter,
 	.gart = {
 		.tlb_flush = &evergreen_pcie_gart_tlb_flush,
 		.set_page = &rs600_gart_set_page,
@@ -1445,6 +1451,7 @@ static struct radeon_asic cayman_asic = {
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
 	.get_xclk = &rv770_get_xclk,
+	.get_gpu_clock_counter = &r600_get_gpu_clock_counter,
 	.gart = {
 		.tlb_flush = &cayman_pcie_gart_tlb_flush,
 		.set_page = &rs600_gart_set_page,
@@ -1573,6 +1580,7 @@ static struct radeon_asic trinity_asic = {
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
 	.get_xclk = &r600_get_xclk,
+	.get_gpu_clock_counter = &r600_get_gpu_clock_counter,
 	.gart = {
 		.tlb_flush = &cayman_pcie_gart_tlb_flush,
 		.set_page = &rs600_gart_set_page,
@@ -1701,6 +1709,7 @@ static struct radeon_asic si_asic = {
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
 	.get_xclk = &si_get_xclk,
+	.get_gpu_clock_counter = &si_get_gpu_clock_counter,
 	.gart = {
 		.tlb_flush = &si_pcie_gart_tlb_flush,
 		.set_page = &rs600_gart_set_page,
diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h
index f15758c..3535f73 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.h
+++ b/drivers/gpu/drm/radeon/radeon_asic.h
@@ -389,8 +389,8 @@ void r600_kms_blit_copy(struct radeon_device *rdev,
 			unsigned num_gpu_pages,
 			struct radeon_sa_bo *vb);
 int r600_mc_wait_for_idle(struct radeon_device *rdev);
-uint64_t r600_get_gpu_clock(struct radeon_device *rdev);
 u32 r600_get_xclk(struct radeon_device *rdev);
+uint64_t r600_get_gpu_clock_counter(struct radeon_device *rdev);
 
 /*
  * rv770,rv730,rv710,rv740
@@ -517,12 +517,12 @@ void si_vm_set_page(struct radeon_device *rdev,
 		    uint32_t incr, uint32_t flags);
 void si_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
 int si_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib);
-uint64_t si_get_gpu_clock(struct radeon_device *rdev);
 int si_copy_dma(struct radeon_device *rdev,
 		uint64_t src_offset, uint64_t dst_offset,
 		unsigned num_gpu_pages,
 		struct radeon_fence **fence);
 void si_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
 u32 si_get_xclk(struct radeon_device *rdev);
+uint64_t si_get_gpu_clock_counter(struct radeon_device *rdev);
 
 #endif
diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index 96f05cd..c75cb2c 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -185,11 +185,7 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 	if (info->request == RADEON_INFO_TIMESTAMP) {
 		if (rdev->family >= CHIP_R600) {
 			value_ptr64 = (uint64_t*)((unsigned long)info->value);
-			if (rdev->family >= CHIP_TAHITI) {
-				value64 = si_get_gpu_clock(rdev);
-			} else {
-				value64 = r600_get_gpu_clock(rdev);
-			}
+			value64 = radeon_get_gpu_clock_counter(rdev);
 
 			if (DRM_COPY_TO_USER(value_ptr64, &value64, sizeof(value64))) {
 				DRM_ERROR("copy_to_user %s:%u\n", __func__, __LINE__);
diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
index b5064fa..80979ed 100644
--- a/drivers/gpu/drm/radeon/si.c
+++ b/drivers/gpu/drm/radeon/si.c
@@ -4609,14 +4609,14 @@ void si_fini(struct radeon_device *rdev)
 }
 
 /**
- * si_get_gpu_clock - return GPU clock counter snapshot
+ * si_get_gpu_clock_counter - return GPU clock counter snapshot
  *
  * @rdev: radeon_device pointer
  *
  * Fetches a GPU clock counter snapshot (SI).
  * Returns the 64 bit clock counter snapshot.
  */
-uint64_t si_get_gpu_clock(struct radeon_device *rdev)
+uint64_t si_get_gpu_clock_counter(struct radeon_device *rdev)
 {
 	uint64_t clock;
 
-- 
cgit v0.10.2


From 43a23aa450cc19fe8996caf09e7e21ae5f6e56e8 Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexander.deucher@amd.com>
Date: Tue, 19 Feb 2013 12:55:52 -0500
Subject: drm/radeon: properly validate the atpx interface

Some bioses don't set the function mask correctly
which caused required functions to be disabled.

Fixes:
https://bugzilla.kernel.org/show_bug.cgi?id=53111

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

diff --git a/drivers/gpu/drm/radeon/radeon_atpx_handler.c b/drivers/gpu/drm/radeon/radeon_atpx_handler.c
index 15f5ded..d96070b 100644
--- a/drivers/gpu/drm/radeon/radeon_atpx_handler.c
+++ b/drivers/gpu/drm/radeon/radeon_atpx_handler.c
@@ -43,6 +43,12 @@ struct atpx_verify_interface {
 	u32 function_bits;	/* supported functions bit vector */
 } __packed;
 
+struct atpx_px_params {
+	u16 size;		/* structure size in bytes (includes size field) */
+	u32 valid_flags;	/* which flags are valid */
+	u32 flags;		/* flags */
+} __packed;
+
 struct atpx_power_control {
 	u16 size;
 	u8 dgpu_state;
@@ -123,9 +129,61 @@ static void radeon_atpx_parse_functions(struct radeon_atpx_functions *f, u32 mas
 }
 
 /**
+ * radeon_atpx_validate_functions - validate ATPX functions
+ *
+ * @atpx: radeon atpx struct
+ *
+ * Validate that required functions are enabled (all asics).
+ * returns 0 on success, error on failure.
+ */
+static int radeon_atpx_validate(struct radeon_atpx *atpx)
+{
+	/* make sure required functions are enabled */
+	/* dGPU power control is required */
+	atpx->functions.power_cntl = true;
+
+	if (atpx->functions.px_params) {
+		union acpi_object *info;
+		struct atpx_px_params output;
+		size_t size;
+		u32 valid_bits;
+
+		info = radeon_atpx_call(atpx->handle, ATPX_FUNCTION_GET_PX_PARAMETERS, NULL);
+		if (!info)
+			return -EIO;
+
+		memset(&output, 0, sizeof(output));
+
+		size = *(u16 *) info->buffer.pointer;
+		if (size < 10) {
+			printk("ATPX buffer is too small: %zu\n", size);
+			kfree(info);
+			return -EINVAL;
+		}
+		size = min(sizeof(output), size);
+
+		memcpy(&output, info->buffer.pointer, size);
+
+		valid_bits = output.flags & output.valid_flags;
+		/* if separate mux flag is set, mux controls are required */
+		if (valid_bits & ATPX_SEPARATE_MUX_FOR_I2C) {
+			atpx->functions.i2c_mux_cntl = true;
+			atpx->functions.disp_mux_cntl = true;
+		}
+		/* if any outputs are muxed, mux controls are required */
+		if (valid_bits & (ATPX_CRT1_RGB_SIGNAL_MUXED |
+				  ATPX_TV_SIGNAL_MUXED |
+				  ATPX_DFP_SIGNAL_MUXED))
+			atpx->functions.disp_mux_cntl = true;
+
+		kfree(info);
+	}
+	return 0;
+}
+
+/**
  * radeon_atpx_verify_interface - verify ATPX
  *
- * @handle: acpi handle
  * @atpx: radeon atpx struct
  *
  * Execute the ATPX_FUNCTION_VERIFY_INTERFACE ATPX function
@@ -406,8 +464,19 @@ static bool radeon_atpx_pci_probe_handle(struct pci_dev *pdev)
  */
 static int radeon_atpx_init(void)
 {
+	int r;
+
 	/* set up the ATPX handle */
-	return radeon_atpx_verify_interface(&radeon_atpx_priv.atpx);
+	r = radeon_atpx_verify_interface(&radeon_atpx_priv.atpx);
+	if (r)
+		return r;
+
+	/* validate the atpx setup */
+	r = radeon_atpx_validate(&radeon_atpx_priv.atpx);
+	if (r)
+		return r;
+
+	return 0;
 }
 
 /**
-- 
cgit v0.10.2


From fe9e3137cffc880b5162f2cc039df48712c496bb Mon Sep 17 00:00:00 2001
From: Inki Dae <inki.dae@samsung.com>
Date: Tue, 29 Jan 2013 16:28:30 +0900
Subject: drm/exynos: fix iommu address allocation order

This patch modifies iommu address allocation order from 64k
to 4k. 64k order causes waste of the io space and this was
our mistakes.

Signed-off-by: Inki Dae <inki.dae@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>

diff --git a/drivers/gpu/drm/exynos/exynos_drm_iommu.h b/drivers/gpu/drm/exynos/exynos_drm_iommu.h
index 53b7dee..598e60f 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_iommu.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_iommu.h
@@ -14,7 +14,7 @@
 
 #define EXYNOS_DEV_ADDR_START	0x20000000
 #define EXYNOS_DEV_ADDR_SIZE	0x40000000
-#define EXYNOS_DEV_ADDR_ORDER	0x4
+#define EXYNOS_DEV_ADDR_ORDER	0x0
 
 #ifdef CONFIG_DRM_EXYNOS_IOMMU
 
-- 
cgit v0.10.2


From b9ede277e15916a9ec3c6c1932c390e4768e71a9 Mon Sep 17 00:00:00 2001
From: Inki Dae <inki.dae@samsung.com>
Date: Tue, 29 Jan 2013 17:51:09 +0900
Subject: drm/exynos: consider exception case to fb handle creation

GETFB ioctl request creates a new handle to only one gem object
so it should check if the given fb has one gem object.

Signed-off-by: Inki Dae <inki.dae@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>

diff --git a/drivers/gpu/drm/exynos/exynos_drm_fb.c b/drivers/gpu/drm/exynos/exynos_drm_fb.c
index 294c051..31d4cb1 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fb.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fb.c
@@ -99,6 +99,10 @@ static int exynos_drm_fb_create_handle(struct drm_framebuffer *fb,
 
 	DRM_DEBUG_KMS("%s\n", __FILE__);
 
+	/* This fb should have only one gem object. */
+	if (WARN_ON(exynos_fb->buf_cnt != 1))
+		return -EINVAL;
+
 	return drm_gem_handle_create(file_priv,
 			&exynos_fb->exynos_gem_obj[0]->base, handle);
 }
-- 
cgit v0.10.2


From e44a5c00169df634c152bdba65c7b735eb3a7a81 Mon Sep 17 00:00:00 2001
From: Sachin Kamat <sachin.kamat@linaro.org>
Date: Fri, 25 Jan 2013 14:45:42 +0530
Subject: drm/exynos: Add missing braces around sizeof

Fixes the following checkpatch warning:
WARNING: sizeof *sgt should be sizeof(*sgt)

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

diff --git a/drivers/gpu/drm/exynos/exynos_drm_g2d.c b/drivers/gpu/drm/exynos/exynos_drm_g2d.c
index 9a4c08e..ddcfb5d 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_g2d.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_g2d.c
@@ -429,7 +429,7 @@ static dma_addr_t *g2d_userptr_get_dma_addr(struct drm_device *drm_dev,
 
 	g2d_userptr->pages = pages;
 
-	sgt = kzalloc(sizeof *sgt, GFP_KERNEL);
+	sgt = kzalloc(sizeof(*sgt), GFP_KERNEL);
 	if (!sgt) {
 		DRM_ERROR("failed to allocate sg table.\n");
 		ret = -ENOMEM;
-- 
cgit v0.10.2


From 5af9f116e6a0e1de675da979a19f95a74ce2aa2e Mon Sep 17 00:00:00 2001
From: YoungJun Cho <yj44.cho@samsung.com>
Date: Thu, 7 Feb 2013 16:17:54 +0900
Subject: drm/exynos: fix wrong pointer access at vm close.

This patch fixes wrong pointer access issue to filp->f_op and
filp->private_data.

The exynos_drm_gem_mmap_ioctl() changes filp->f_op and
filp->private_data temporarily and restore them to use
original ones in exynos_drm_gem_mmap_buffer() but there
was no lock between the changing and the restoring so
wrong pointer access to filp->f_op and filp->private_data
was induced by vm close callback.

So this patch uses mutex lock properly to resolve this issue.

Signed-off-by: YoungJun Cho <yj44.cho@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>

diff --git a/drivers/gpu/drm/exynos/exynos_drm_gem.c b/drivers/gpu/drm/exynos/exynos_drm_gem.c
index 4731807..67e17ce 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_gem.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_gem.c
@@ -329,17 +329,11 @@ static struct drm_file *exynos_drm_find_drm_file(struct drm_device *drm_dev,
 {
 	struct drm_file *file_priv;
 
-	mutex_lock(&drm_dev->struct_mutex);
-
 	/* find current process's drm_file from filelist. */
-	list_for_each_entry(file_priv, &drm_dev->filelist, lhead) {
-		if (file_priv->filp == filp) {
-			mutex_unlock(&drm_dev->struct_mutex);
+	list_for_each_entry(file_priv, &drm_dev->filelist, lhead)
+		if (file_priv->filp == filp)
 			return file_priv;
-		}
-	}
 
-	mutex_unlock(&drm_dev->struct_mutex);
 	WARN_ON(1);
 
 	return ERR_PTR(-EFAULT);
@@ -400,9 +394,7 @@ static int exynos_drm_gem_mmap_buffer(struct file *filp,
 	 */
 	drm_gem_object_reference(obj);
 
-	mutex_lock(&drm_dev->struct_mutex);
 	drm_vm_open_locked(drm_dev, vma);
-	mutex_unlock(&drm_dev->struct_mutex);
 
 	return 0;
 }
@@ -432,6 +424,16 @@ int exynos_drm_gem_mmap_ioctl(struct drm_device *dev, void *data,
 	}
 
 	/*
+	 * We have to use gem object and its fops for specific mmaper,
+	 * but vm_mmap() can deliver only filp. So we have to change
+	 * filp->f_op and filp->private_data temporarily, then restore
+	 * again. So it is important to keep lock until restoration the
+	 * settings to prevent others from misuse of filp->f_op or
+	 * filp->private_data.
+	 */
+	mutex_lock(&dev->struct_mutex);
+
+	/*
 	 * Set specific mmper's fops. And it will be restored by
 	 * exynos_drm_gem_mmap_buffer to dev->driver->fops.
 	 * This is used to call specific mapper temporarily.
@@ -448,13 +450,20 @@ int exynos_drm_gem_mmap_ioctl(struct drm_device *dev, void *data,
 	addr = vm_mmap(file_priv->filp, 0, args->size,
 			PROT_READ | PROT_WRITE, MAP_SHARED, 0);
 
-	drm_gem_object_unreference_unlocked(obj);
+	drm_gem_object_unreference(obj);
 
 	if (IS_ERR((void *)addr)) {
-		file_priv->filp->private_data = file_priv;
+		/* check filp->f_op, filp->private_data are restored */
+		if (file_priv->filp->f_op == &exynos_drm_gem_fops) {
+			file_priv->filp->f_op = fops_get(dev->driver->fops);
+			file_priv->filp->private_data = file_priv;
+		}
+		mutex_unlock(&dev->struct_mutex);
 		return PTR_ERR((void *)addr);
 	}
 
+	mutex_unlock(&dev->struct_mutex);
+
 	args->mapped = addr;
 
 	DRM_DEBUG_KMS("mapped = 0x%lx\n", (unsigned long)args->mapped);
-- 
cgit v0.10.2


From 979c0c7eb0fb969f0621c7017a6a87e6a0e36a4c Mon Sep 17 00:00:00 2001
From: YoungJun Cho <yj44.cho@samsung.com>
Date: Tue, 12 Feb 2013 21:23:54 +0900
Subject: drm/exynos: release resources properly when fb creation is failed.

This patch releases allocated resources properly when
exynos_user_fb_create() is failed.

Signed-off-by: YoungJun Cho <yj44.cho@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>

diff --git a/drivers/gpu/drm/exynos/exynos_drm_fb.c b/drivers/gpu/drm/exynos/exynos_drm_fb.c
index 31d4cb1..0e04f4e 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fb.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fb.c
@@ -221,23 +221,25 @@ exynos_user_fb_create(struct drm_device *dev, struct drm_file *file_priv,
 		      struct drm_mode_fb_cmd2 *mode_cmd)
 {
 	struct drm_gem_object *obj;
+	struct exynos_drm_gem_obj *exynos_gem_obj;
 	struct exynos_drm_fb *exynos_fb;
 	int i, ret;
 
 	DRM_DEBUG_KMS("%s\n", __FILE__);
 
-	obj = drm_gem_object_lookup(dev, file_priv, mode_cmd->handles[0]);
-	if (!obj) {
-		DRM_ERROR("failed to lookup gem object\n");
-		return ERR_PTR(-ENOENT);
-	}
-
 	exynos_fb = kzalloc(sizeof(*exynos_fb), GFP_KERNEL);
 	if (!exynos_fb) {
 		DRM_ERROR("failed to allocate exynos drm framebuffer\n");
 		return ERR_PTR(-ENOMEM);
 	}
 
+	obj = drm_gem_object_lookup(dev, file_priv, mode_cmd->handles[0]);
+	if (!obj) {
+		DRM_ERROR("failed to lookup gem object\n");
+		ret = -ENOENT;
+		goto err_free;
+	}
+
 	drm_helper_mode_fill_fb_struct(&exynos_fb->fb, mode_cmd);
 	exynos_fb->exynos_gem_obj[0] = to_exynos_gem_obj(obj);
 	exynos_fb->buf_cnt = exynos_drm_format_num_buffers(mode_cmd);
@@ -245,43 +247,44 @@ exynos_user_fb_create(struct drm_device *dev, struct drm_file *file_priv,
 	DRM_DEBUG_KMS("buf_cnt = %d\n", exynos_fb->buf_cnt);
 
 	for (i = 1; i < exynos_fb->buf_cnt; i++) {
-		struct exynos_drm_gem_obj *exynos_gem_obj;
-		int ret;
-
 		obj = drm_gem_object_lookup(dev, file_priv,
 				mode_cmd->handles[i]);
 		if (!obj) {
 			DRM_ERROR("failed to lookup gem object\n");
-			kfree(exynos_fb);
-			return ERR_PTR(-ENOENT);
+			ret = -ENOENT;
+			exynos_fb->buf_cnt = i;
+			goto err_unreference;
 		}
 
 		exynos_gem_obj = to_exynos_gem_obj(obj);
+		exynos_fb->exynos_gem_obj[i] = exynos_gem_obj;
 
 		ret = check_fb_gem_memory_type(dev, exynos_gem_obj);
 		if (ret < 0) {
 			DRM_ERROR("cannot use this gem memory type for fb.\n");
-			kfree(exynos_fb);
-			return ERR_PTR(ret);
+			goto err_unreference;
 		}
-
-		exynos_fb->exynos_gem_obj[i] = to_exynos_gem_obj(obj);
 	}
 
 	ret = drm_framebuffer_init(dev, &exynos_fb->fb, &exynos_drm_fb_funcs);
 	if (ret) {
-		for (i = 0; i < exynos_fb->buf_cnt; i++) {
-			struct exynos_drm_gem_obj *gem_obj;
-
-			gem_obj = exynos_fb->exynos_gem_obj[i];
-			drm_gem_object_unreference_unlocked(&gem_obj->base);
-		}
-
-		kfree(exynos_fb);
-		return ERR_PTR(ret);
+		DRM_ERROR("failed to init framebuffer.\n");
+		goto err_unreference;
 	}
 
 	return &exynos_fb->fb;
+
+err_unreference:
+	for (i = 0; i < exynos_fb->buf_cnt; i++) {
+		struct drm_gem_object *obj;
+
+		obj = &exynos_fb->exynos_gem_obj[i]->base;
+		if (obj)
+			drm_gem_object_unreference_unlocked(obj);
+	}
+err_free:
+	kfree(exynos_fb);
+	return ERR_PTR(ret);
 }
 
 struct exynos_drm_gem_buf *exynos_drm_fb_buffer(struct drm_framebuffer *fb,
-- 
cgit v0.10.2


From 438c0f85cbbe6e7025e2375f97f5223b6eac5cd7 Mon Sep 17 00:00:00 2001
From: Rahul Sharma <rahul.sharma@samsung.com>
Date: Tue, 15 Jan 2013 08:11:05 -0500
Subject: drm/exynos: add display-mode-check operation to exynos_mixer_ops
 struct

This patch adds the display mode check operation to exynos_mixer_ops
in drm-common-hdmi. In Exynos SoCs, mixer IP can put certain restrictions
on the proposed display modes. These restriction needs to be considered
during mode negotiation, which happens immediately after edid parsing.

Both, mixer check-mode and hdmi check-timing callbacks are called one after
another and ANDed result is returned back.

Signed-off-by: Rahul Sharma <rahul.sharma@samsung.com>
Reviewed-by: Sean Paul <seanpaul@chromium.org>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

diff --git a/drivers/gpu/drm/exynos/exynos_drm_hdmi.c b/drivers/gpu/drm/exynos/exynos_drm_hdmi.c
index 2864453..7c27df0 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_hdmi.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_hdmi.c
@@ -124,9 +124,21 @@ static struct edid *drm_hdmi_get_edid(struct device *dev,
 static int drm_hdmi_check_timing(struct device *dev, void *timing)
 {
 	struct drm_hdmi_context *ctx = to_context(dev);
+	int ret = 0;
 
 	DRM_DEBUG_KMS("%s\n", __FILE__);
 
+	/*
+	* Both, mixer and hdmi should be able to handle the requested mode.
+	* If any of the two fails, return mode as BAD.
+	*/
+
+	if (mixer_ops && mixer_ops->check_timing)
+		ret = mixer_ops->check_timing(ctx->mixer_ctx->ctx, timing);
+
+	if (ret)
+		return ret;
+
 	if (hdmi_ops && hdmi_ops->check_timing)
 		return hdmi_ops->check_timing(ctx->hdmi_ctx->ctx, timing);
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_hdmi.h b/drivers/gpu/drm/exynos/exynos_drm_hdmi.h
index d80516f..b7faa36 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_hdmi.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_hdmi.h
@@ -32,7 +32,7 @@ struct exynos_hdmi_ops {
 	bool (*is_connected)(void *ctx);
 	struct edid *(*get_edid)(void *ctx,
 			struct drm_connector *connector);
-	int (*check_timing)(void *ctx, void *timing);
+	int (*check_timing)(void *ctx, struct fb_videomode *timing);
 	int (*power_on)(void *ctx, int mode);
 
 	/* manager */
@@ -58,6 +58,9 @@ struct exynos_mixer_ops {
 	void (*win_mode_set)(void *ctx, struct exynos_drm_overlay *overlay);
 	void (*win_commit)(void *ctx, int zpos);
 	void (*win_disable)(void *ctx, int zpos);
+
+	/* display */
+	int (*check_timing)(void *ctx, struct fb_videomode *timing);
 };
 
 void exynos_hdmi_drv_attach(struct exynos_drm_hdmi_context *ctx);
diff --git a/drivers/gpu/drm/exynos/exynos_hdmi.c b/drivers/gpu/drm/exynos/exynos_hdmi.c
index fbab3c4..ba33015 100644
--- a/drivers/gpu/drm/exynos/exynos_hdmi.c
+++ b/drivers/gpu/drm/exynos/exynos_hdmi.c
@@ -1457,21 +1457,20 @@ static int hdmi_v14_check_timing(struct fb_videomode *check_timing)
 	return -EINVAL;
 }
 
-static int hdmi_check_timing(void *ctx, void *timing)
+static int hdmi_check_timing(void *ctx, struct fb_videomode *timing)
 {
 	struct hdmi_context *hdata = ctx;
-	struct fb_videomode *check_timing = timing;
 
 	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
 
-	DRM_DEBUG_KMS("[%d]x[%d] [%d]Hz [%x]\n", check_timing->xres,
-			check_timing->yres, check_timing->refresh,
-			check_timing->vmode);
+	DRM_DEBUG_KMS("[%d]x[%d] [%d]Hz [%x]\n", timing->xres,
+			timing->yres, timing->refresh,
+			timing->vmode);
 
 	if (hdata->type == HDMI_TYPE13)
-		return hdmi_v13_check_timing(check_timing);
+		return hdmi_v13_check_timing(timing);
 	else
-		return hdmi_v14_check_timing(check_timing);
+		return hdmi_v14_check_timing(timing);
 }
 
 static void hdmi_set_acr(u32 freq, u8 *acr)
-- 
cgit v0.10.2


From 0ea6822f5215b1771949e8b278f66ed787ba83f1 Mon Sep 17 00:00:00 2001
From: Rahul Sharma <rahul.sharma@samsung.com>
Date: Tue, 15 Jan 2013 08:11:06 -0500
Subject: drm/exynos: implement display-mode-check callback in mixer driver

This patch adds the implementation of check_timing callback in the mixer
driver. Based on the mixer version, correct set of restrictions will be
exposed by the mixer driver. A resolution will be acceptable only if passes
the criteria set by mixer and hdmi IPs.

Signed-off-by: Rahul Sharma <rahul.sharma@samsung.com>
Signed-off-by: Sean Paul <seanpaul@chromium.org>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

diff --git a/drivers/gpu/drm/exynos/exynos_mixer.c b/drivers/gpu/drm/exynos/exynos_mixer.c
index c414584..cc809ca 100644
--- a/drivers/gpu/drm/exynos/exynos_mixer.c
+++ b/drivers/gpu/drm/exynos/exynos_mixer.c
@@ -818,6 +818,29 @@ static void mixer_win_disable(void *ctx, int win)
 	mixer_ctx->win_data[win].enabled = false;
 }
 
+int mixer_check_timing(void *ctx, struct fb_videomode *timing)
+{
+	struct mixer_context *mixer_ctx = ctx;
+	u32 w, h;
+
+	w = timing->xres;
+	h = timing->yres;
+
+	DRM_DEBUG_KMS("%s : xres=%d, yres=%d, refresh=%d, intl=%d\n",
+		__func__, timing->xres, timing->yres,
+		timing->refresh, (timing->vmode &
+		FB_VMODE_INTERLACED) ? true : false);
+
+	if (mixer_ctx->mxr_ver == MXR_VER_0_0_0_16)
+		return 0;
+
+	if ((w >= 464 && w <= 720 && h >= 261 && h <= 576) ||
+		(w >= 1024 && w <= 1280 && h >= 576 && h <= 720) ||
+		(w >= 1664 && w <= 1920 && h >= 936 && h <= 1080))
+		return 0;
+
+	return -EINVAL;
+}
 static void mixer_wait_for_vblank(void *ctx)
 {
 	struct mixer_context *mixer_ctx = ctx;
@@ -955,6 +978,9 @@ static struct exynos_mixer_ops mixer_ops = {
 	.win_mode_set		= mixer_win_mode_set,
 	.win_commit		= mixer_win_commit,
 	.win_disable		= mixer_win_disable,
+
+	/* display */
+	.check_timing		= mixer_check_timing,
 };
 
 static irqreturn_t mixer_irq_handler(int irq, void *arg)
-- 
cgit v0.10.2


From 29630743c97cc4c395c308ea9d0c12bf55c55a92 Mon Sep 17 00:00:00 2001
From: Rahul Sharma <rahul.sharma@samsung.com>
Date: Tue, 15 Jan 2013 08:11:07 -0500
Subject: drm/exynos: mixer: set correct mode for range of resolutions

With this patch, mixer driver find the correct resolution mode for
the range of resolutions, upto 1080 vertical lines. Resolution will
be categorized to NTSC SD, PAL SD or HD and the correct mode is
set to the mixer configuration register.

Signed-off-by: Rahul Sharma <rahul.sharma@samsung.com>
Signed-off-by: Sean Paul <seanpaul@chromium.org>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

diff --git a/drivers/gpu/drm/exynos/exynos_mixer.c b/drivers/gpu/drm/exynos/exynos_mixer.c
index cc809ca..e919aba 100644
--- a/drivers/gpu/drm/exynos/exynos_mixer.c
+++ b/drivers/gpu/drm/exynos/exynos_mixer.c
@@ -284,13 +284,13 @@ static void mixer_cfg_scan(struct mixer_context *ctx, unsigned int height)
 				MXR_CFG_SCAN_PROGRASSIVE);
 
 	/* choosing between porper HD and SD mode */
-	if (height == 480)
+	if (height <= 480)
 		val |= MXR_CFG_SCAN_NTSC | MXR_CFG_SCAN_SD;
-	else if (height == 576)
+	else if (height <= 576)
 		val |= MXR_CFG_SCAN_PAL | MXR_CFG_SCAN_SD;
-	else if (height == 720)
+	else if (height <= 720)
 		val |= MXR_CFG_SCAN_HD_720 | MXR_CFG_SCAN_HD;
-	else if (height == 1080)
+	else if (height <= 1080)
 		val |= MXR_CFG_SCAN_HD_1080 | MXR_CFG_SCAN_HD;
 	else
 		val |= MXR_CFG_SCAN_HD_720 | MXR_CFG_SCAN_HD;
-- 
cgit v0.10.2


From 2f7e2ed073f629405a8054f14bd001dc68b0e17a Mon Sep 17 00:00:00 2001
From: Sean Paul <seanpaul@chromium.org>
Date: Tue, 15 Jan 2013 08:11:08 -0500
Subject: drm/exynos: hdmi: support extra resolutions using drm_display_mode
 timings

This patch programs the core and timing generator registers using the
timing data provided in drm_display_mode and not using hard-coded
configurations.

Additional PHY configs has been added. This allows us to support more
permissible resolutions and refresh rates.

Signed-off-by: Rahul Sharma <rahul.sharma@samsung.com>
Signed-off-by: Sean Paul <seanpaul@chromium.org>
Signed-off-by: Shirish S <s.shirish@samsung.com>
Signed-off-by: Akshay Saraswat <Akshay.s@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

diff --git a/drivers/gpu/drm/exynos/exynos_hdmi.c b/drivers/gpu/drm/exynos/exynos_hdmi.c
index ba33015..6d63f90 100644
--- a/drivers/gpu/drm/exynos/exynos_hdmi.c
+++ b/drivers/gpu/drm/exynos/exynos_hdmi.c
@@ -87,6 +87,73 @@ struct hdmi_resources {
 	int				regul_count;
 };
 
+struct hdmi_tg_regs {
+	u8 cmd[1];
+	u8 h_fsz[2];
+	u8 hact_st[2];
+	u8 hact_sz[2];
+	u8 v_fsz[2];
+	u8 vsync[2];
+	u8 vsync2[2];
+	u8 vact_st[2];
+	u8 vact_sz[2];
+	u8 field_chg[2];
+	u8 vact_st2[2];
+	u8 vact_st3[2];
+	u8 vact_st4[2];
+	u8 vsync_top_hdmi[2];
+	u8 vsync_bot_hdmi[2];
+	u8 field_top_hdmi[2];
+	u8 field_bot_hdmi[2];
+	u8 tg_3d[1];
+};
+
+struct hdmi_core_regs {
+	u8 h_blank[2];
+	u8 v2_blank[2];
+	u8 v1_blank[2];
+	u8 v_line[2];
+	u8 h_line[2];
+	u8 hsync_pol[1];
+	u8 vsync_pol[1];
+	u8 int_pro_mode[1];
+	u8 v_blank_f0[2];
+	u8 v_blank_f1[2];
+	u8 h_sync_start[2];
+	u8 h_sync_end[2];
+	u8 v_sync_line_bef_2[2];
+	u8 v_sync_line_bef_1[2];
+	u8 v_sync_line_aft_2[2];
+	u8 v_sync_line_aft_1[2];
+	u8 v_sync_line_aft_pxl_2[2];
+	u8 v_sync_line_aft_pxl_1[2];
+	u8 v_blank_f2[2]; /* for 3D mode */
+	u8 v_blank_f3[2]; /* for 3D mode */
+	u8 v_blank_f4[2]; /* for 3D mode */
+	u8 v_blank_f5[2]; /* for 3D mode */
+	u8 v_sync_line_aft_3[2];
+	u8 v_sync_line_aft_4[2];
+	u8 v_sync_line_aft_5[2];
+	u8 v_sync_line_aft_6[2];
+	u8 v_sync_line_aft_pxl_3[2];
+	u8 v_sync_line_aft_pxl_4[2];
+	u8 v_sync_line_aft_pxl_5[2];
+	u8 v_sync_line_aft_pxl_6[2];
+	u8 vact_space_1[2];
+	u8 vact_space_2[2];
+	u8 vact_space_3[2];
+	u8 vact_space_4[2];
+	u8 vact_space_5[2];
+	u8 vact_space_6[2];
+};
+
+struct hdmi_v14_conf {
+	int pixel_clock;
+	struct hdmi_core_regs core;
+	struct hdmi_tg_regs tg;
+	int cea_video_id;
+};
+
 struct hdmi_context {
 	struct device			*dev;
 	struct drm_device		*drm_dev;
@@ -104,6 +171,7 @@ struct hdmi_context {
 
 	/* current hdmiphy conf index */
 	int cur_conf;
+	struct hdmi_v14_conf		mode_conf;
 
 	struct hdmi_resources		res;
 
@@ -392,586 +460,132 @@ static const struct hdmi_v13_conf hdmi_v13_confs[] = {
 };
 
 /* HDMI Version 1.4 */
-static const u8 hdmiphy_conf27_027[32] = {
-	0x01, 0xd1, 0x2d, 0x72, 0x40, 0x64, 0x12, 0x08,
-	0x43, 0xa0, 0x0e, 0xd9, 0x45, 0xa0, 0xac, 0x80,
-	0x08, 0x80, 0x11, 0x04, 0x02, 0x22, 0x44, 0x86,
-	0x54, 0xe3, 0x24, 0x00, 0x00, 0x00, 0x01, 0x00,
-};
-
-static const u8 hdmiphy_conf74_176[32] = {
-	0x01, 0xd1, 0x1f, 0x10, 0x40, 0x5b, 0xef, 0x08,
-	0x81, 0xa0, 0xb9, 0xd8, 0x45, 0xa0, 0xac, 0x80,
-	0x5a, 0x80, 0x11, 0x04, 0x02, 0x22, 0x44, 0x86,
-	0x54, 0xa6, 0x24, 0x01, 0x00, 0x00, 0x01, 0x00,
-};
-
-static const u8 hdmiphy_conf74_25[32] = {
-	0x01, 0xd1, 0x1f, 0x10, 0x40, 0x40, 0xf8, 0x08,
-	0x81, 0xa0, 0xba, 0xd8, 0x45, 0xa0, 0xac, 0x80,
-	0x3c, 0x80, 0x11, 0x04, 0x02, 0x22, 0x44, 0x86,
-	0x54, 0xa5, 0x24, 0x01, 0x00, 0x00, 0x01, 0x00,
-};
-
-static const u8 hdmiphy_conf148_5[32] = {
-	0x01, 0xd1, 0x1f, 0x00, 0x40, 0x40, 0xf8, 0x08,
-	0x81, 0xa0, 0xba, 0xd8, 0x45, 0xa0, 0xac, 0x80,
-	0x3c, 0x80, 0x11, 0x04, 0x02, 0x22, 0x44, 0x86,
-	0x54, 0x4b, 0x25, 0x03, 0x00, 0x00, 0x01, 0x00,
-};
-
-struct hdmi_tg_regs {
-	u8 cmd;
-	u8 h_fsz_l;
-	u8 h_fsz_h;
-	u8 hact_st_l;
-	u8 hact_st_h;
-	u8 hact_sz_l;
-	u8 hact_sz_h;
-	u8 v_fsz_l;
-	u8 v_fsz_h;
-	u8 vsync_l;
-	u8 vsync_h;
-	u8 vsync2_l;
-	u8 vsync2_h;
-	u8 vact_st_l;
-	u8 vact_st_h;
-	u8 vact_sz_l;
-	u8 vact_sz_h;
-	u8 field_chg_l;
-	u8 field_chg_h;
-	u8 vact_st2_l;
-	u8 vact_st2_h;
-	u8 vact_st3_l;
-	u8 vact_st3_h;
-	u8 vact_st4_l;
-	u8 vact_st4_h;
-	u8 vsync_top_hdmi_l;
-	u8 vsync_top_hdmi_h;
-	u8 vsync_bot_hdmi_l;
-	u8 vsync_bot_hdmi_h;
-	u8 field_top_hdmi_l;
-	u8 field_top_hdmi_h;
-	u8 field_bot_hdmi_l;
-	u8 field_bot_hdmi_h;
-	u8 tg_3d;
-};
-
-struct hdmi_core_regs {
-	u8 h_blank[2];
-	u8 v2_blank[2];
-	u8 v1_blank[2];
-	u8 v_line[2];
-	u8 h_line[2];
-	u8 hsync_pol[1];
-	u8 vsync_pol[1];
-	u8 int_pro_mode[1];
-	u8 v_blank_f0[2];
-	u8 v_blank_f1[2];
-	u8 h_sync_start[2];
-	u8 h_sync_end[2];
-	u8 v_sync_line_bef_2[2];
-	u8 v_sync_line_bef_1[2];
-	u8 v_sync_line_aft_2[2];
-	u8 v_sync_line_aft_1[2];
-	u8 v_sync_line_aft_pxl_2[2];
-	u8 v_sync_line_aft_pxl_1[2];
-	u8 v_blank_f2[2]; /* for 3D mode */
-	u8 v_blank_f3[2]; /* for 3D mode */
-	u8 v_blank_f4[2]; /* for 3D mode */
-	u8 v_blank_f5[2]; /* for 3D mode */
-	u8 v_sync_line_aft_3[2];
-	u8 v_sync_line_aft_4[2];
-	u8 v_sync_line_aft_5[2];
-	u8 v_sync_line_aft_6[2];
-	u8 v_sync_line_aft_pxl_3[2];
-	u8 v_sync_line_aft_pxl_4[2];
-	u8 v_sync_line_aft_pxl_5[2];
-	u8 v_sync_line_aft_pxl_6[2];
-	u8 vact_space_1[2];
-	u8 vact_space_2[2];
-	u8 vact_space_3[2];
-	u8 vact_space_4[2];
-	u8 vact_space_5[2];
-	u8 vact_space_6[2];
-};
-
-struct hdmi_preset_conf {
-	struct hdmi_core_regs core;
-	struct hdmi_tg_regs tg;
-};
-
-struct hdmi_conf {
-	int width;
-	int height;
-	int vrefresh;
-	bool interlace;
-	int cea_video_id;
-	const u8 *hdmiphy_data;
-	const struct hdmi_preset_conf *conf;
-};
-
-static const struct hdmi_preset_conf hdmi_conf_480p60 = {
-	.core = {
-		.h_blank = {0x8a, 0x00},
-		.v2_blank = {0x0d, 0x02},
-		.v1_blank = {0x2d, 0x00},
-		.v_line = {0x0d, 0x02},
-		.h_line = {0x5a, 0x03},
-		.hsync_pol = {0x01},
-		.vsync_pol = {0x01},
-		.int_pro_mode = {0x00},
-		.v_blank_f0 = {0xff, 0xff},
-		.v_blank_f1 = {0xff, 0xff},
-		.h_sync_start = {0x0e, 0x00},
-		.h_sync_end = {0x4c, 0x00},
-		.v_sync_line_bef_2 = {0x0f, 0x00},
-		.v_sync_line_bef_1 = {0x09, 0x00},
-		.v_sync_line_aft_2 = {0xff, 0xff},
-		.v_sync_line_aft_1 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_2 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_1 = {0xff, 0xff},
-		.v_blank_f2 = {0xff, 0xff},
-		.v_blank_f3 = {0xff, 0xff},
-		.v_blank_f4 = {0xff, 0xff},
-		.v_blank_f5 = {0xff, 0xff},
-		.v_sync_line_aft_3 = {0xff, 0xff},
-		.v_sync_line_aft_4 = {0xff, 0xff},
-		.v_sync_line_aft_5 = {0xff, 0xff},
-		.v_sync_line_aft_6 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_3 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_4 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_5 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_6 = {0xff, 0xff},
-		.vact_space_1 = {0xff, 0xff},
-		.vact_space_2 = {0xff, 0xff},
-		.vact_space_3 = {0xff, 0xff},
-		.vact_space_4 = {0xff, 0xff},
-		.vact_space_5 = {0xff, 0xff},
-		.vact_space_6 = {0xff, 0xff},
-		/* other don't care */
-	},
-	.tg = {
-		0x00, /* cmd */
-		0x5a, 0x03, /* h_fsz */
-		0x8a, 0x00, 0xd0, 0x02, /* hact */
-		0x0d, 0x02, /* v_fsz */
-		0x01, 0x00, 0x33, 0x02, /* vsync */
-		0x2d, 0x00, 0xe0, 0x01, /* vact */
-		0x33, 0x02, /* field_chg */
-		0x48, 0x02, /* vact_st2 */
-		0x00, 0x00, /* vact_st3 */
-		0x00, 0x00, /* vact_st4 */
-		0x01, 0x00, 0x01, 0x00, /* vsync top/bot */
-		0x01, 0x00, 0x33, 0x02, /* field top/bot */
-		0x00, /* 3d FP */
-	},
+struct hdmiphy_config {
+	int pixel_clock;
+	u8 conf[32];
 };
 
-static const struct hdmi_preset_conf hdmi_conf_720p50 = {
-	.core = {
-		.h_blank = {0xbc, 0x02},
-		.v2_blank = {0xee, 0x02},
-		.v1_blank = {0x1e, 0x00},
-		.v_line = {0xee, 0x02},
-		.h_line = {0xbc, 0x07},
-		.hsync_pol = {0x00},
-		.vsync_pol = {0x00},
-		.int_pro_mode = {0x00},
-		.v_blank_f0 = {0xff, 0xff},
-		.v_blank_f1 = {0xff, 0xff},
-		.h_sync_start = {0xb6, 0x01},
-		.h_sync_end = {0xde, 0x01},
-		.v_sync_line_bef_2 = {0x0a, 0x00},
-		.v_sync_line_bef_1 = {0x05, 0x00},
-		.v_sync_line_aft_2 = {0xff, 0xff},
-		.v_sync_line_aft_1 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_2 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_1 = {0xff, 0xff},
-		.v_blank_f2 = {0xff, 0xff},
-		.v_blank_f3 = {0xff, 0xff},
-		.v_blank_f4 = {0xff, 0xff},
-		.v_blank_f5 = {0xff, 0xff},
-		.v_sync_line_aft_3 = {0xff, 0xff},
-		.v_sync_line_aft_4 = {0xff, 0xff},
-		.v_sync_line_aft_5 = {0xff, 0xff},
-		.v_sync_line_aft_6 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_3 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_4 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_5 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_6 = {0xff, 0xff},
-		.vact_space_1 = {0xff, 0xff},
-		.vact_space_2 = {0xff, 0xff},
-		.vact_space_3 = {0xff, 0xff},
-		.vact_space_4 = {0xff, 0xff},
-		.vact_space_5 = {0xff, 0xff},
-		.vact_space_6 = {0xff, 0xff},
-		/* other don't care */
-	},
-	.tg = {
-		0x00, /* cmd */
-		0xbc, 0x07, /* h_fsz */
-		0xbc, 0x02, 0x00, 0x05, /* hact */
-		0xee, 0x02, /* v_fsz */
-		0x01, 0x00, 0x33, 0x02, /* vsync */
-		0x1e, 0x00, 0xd0, 0x02, /* vact */
-		0x33, 0x02, /* field_chg */
-		0x48, 0x02, /* vact_st2 */
-		0x00, 0x00, /* vact_st3 */
-		0x00, 0x00, /* vact_st4 */
-		0x01, 0x00, 0x01, 0x00, /* vsync top/bot */
-		0x01, 0x00, 0x33, 0x02, /* field top/bot */
-		0x00, /* 3d FP */
+/* list of all required phy config settings */
+static const struct hdmiphy_config hdmiphy_v14_configs[] = {
+	{
+		.pixel_clock = 25200000,
+		.conf = {
+			0x01, 0x51, 0x2A, 0x75, 0x40, 0x01, 0x00, 0x08,
+			0x82, 0x80, 0xfc, 0xd8, 0x45, 0xa0, 0xac, 0x80,
+			0x08, 0x80, 0x11, 0x04, 0x02, 0x22, 0x44, 0x86,
+			0x54, 0xf4, 0x24, 0x00, 0x00, 0x00, 0x01, 0x80,
+		},
 	},
-};
-
-static const struct hdmi_preset_conf hdmi_conf_720p60 = {
-	.core = {
-		.h_blank = {0x72, 0x01},
-		.v2_blank = {0xee, 0x02},
-		.v1_blank = {0x1e, 0x00},
-		.v_line = {0xee, 0x02},
-		.h_line = {0x72, 0x06},
-		.hsync_pol = {0x00},
-		.vsync_pol = {0x00},
-		.int_pro_mode = {0x00},
-		.v_blank_f0 = {0xff, 0xff},
-		.v_blank_f1 = {0xff, 0xff},
-		.h_sync_start = {0x6c, 0x00},
-		.h_sync_end = {0x94, 0x00},
-		.v_sync_line_bef_2 = {0x0a, 0x00},
-		.v_sync_line_bef_1 = {0x05, 0x00},
-		.v_sync_line_aft_2 = {0xff, 0xff},
-		.v_sync_line_aft_1 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_2 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_1 = {0xff, 0xff},
-		.v_blank_f2 = {0xff, 0xff},
-		.v_blank_f3 = {0xff, 0xff},
-		.v_blank_f4 = {0xff, 0xff},
-		.v_blank_f5 = {0xff, 0xff},
-		.v_sync_line_aft_3 = {0xff, 0xff},
-		.v_sync_line_aft_4 = {0xff, 0xff},
-		.v_sync_line_aft_5 = {0xff, 0xff},
-		.v_sync_line_aft_6 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_3 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_4 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_5 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_6 = {0xff, 0xff},
-		.vact_space_1 = {0xff, 0xff},
-		.vact_space_2 = {0xff, 0xff},
-		.vact_space_3 = {0xff, 0xff},
-		.vact_space_4 = {0xff, 0xff},
-		.vact_space_5 = {0xff, 0xff},
-		.vact_space_6 = {0xff, 0xff},
-		/* other don't care */
+	{
+		.pixel_clock = 27000000,
+		.conf = {
+			0x01, 0xd1, 0x22, 0x51, 0x40, 0x08, 0xfc, 0x20,
+			0x98, 0xa0, 0xcb, 0xd8, 0x45, 0xa0, 0xac, 0x80,
+			0x06, 0x80, 0x11, 0x04, 0x02, 0x22, 0x44, 0x86,
+			0x54, 0xe4, 0x24, 0x00, 0x00, 0x00, 0x01, 0x80,
+		},
 	},
-	.tg = {
-		0x00, /* cmd */
-		0x72, 0x06, /* h_fsz */
-		0x72, 0x01, 0x00, 0x05, /* hact */
-		0xee, 0x02, /* v_fsz */
-		0x01, 0x00, 0x33, 0x02, /* vsync */
-		0x1e, 0x00, 0xd0, 0x02, /* vact */
-		0x33, 0x02, /* field_chg */
-		0x48, 0x02, /* vact_st2 */
-		0x00, 0x00, /* vact_st3 */
-		0x00, 0x00, /* vact_st4 */
-		0x01, 0x00, 0x01, 0x00, /* vsync top/bot */
-		0x01, 0x00, 0x33, 0x02, /* field top/bot */
-		0x00, /* 3d FP */
+	{
+		.pixel_clock = 27027000,
+		.conf = {
+			0x01, 0xd1, 0x2d, 0x72, 0x40, 0x64, 0x12, 0x08,
+			0x43, 0xa0, 0x0e, 0xd9, 0x45, 0xa0, 0xac, 0x80,
+			0x08, 0x80, 0x11, 0x04, 0x02, 0x22, 0x44, 0x86,
+			0x54, 0xe3, 0x24, 0x00, 0x00, 0x00, 0x01, 0x00,
+		},
 	},
-};
-
-static const struct hdmi_preset_conf hdmi_conf_1080i50 = {
-	.core = {
-		.h_blank = {0xd0, 0x02},
-		.v2_blank = {0x32, 0x02},
-		.v1_blank = {0x16, 0x00},
-		.v_line = {0x65, 0x04},
-		.h_line = {0x50, 0x0a},
-		.hsync_pol = {0x00},
-		.vsync_pol = {0x00},
-		.int_pro_mode = {0x01},
-		.v_blank_f0 = {0x49, 0x02},
-		.v_blank_f1 = {0x65, 0x04},
-		.h_sync_start = {0x0e, 0x02},
-		.h_sync_end = {0x3a, 0x02},
-		.v_sync_line_bef_2 = {0x07, 0x00},
-		.v_sync_line_bef_1 = {0x02, 0x00},
-		.v_sync_line_aft_2 = {0x39, 0x02},
-		.v_sync_line_aft_1 = {0x34, 0x02},
-		.v_sync_line_aft_pxl_2 = {0x38, 0x07},
-		.v_sync_line_aft_pxl_1 = {0x38, 0x07},
-		.v_blank_f2 = {0xff, 0xff},
-		.v_blank_f3 = {0xff, 0xff},
-		.v_blank_f4 = {0xff, 0xff},
-		.v_blank_f5 = {0xff, 0xff},
-		.v_sync_line_aft_3 = {0xff, 0xff},
-		.v_sync_line_aft_4 = {0xff, 0xff},
-		.v_sync_line_aft_5 = {0xff, 0xff},
-		.v_sync_line_aft_6 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_3 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_4 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_5 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_6 = {0xff, 0xff},
-		.vact_space_1 = {0xff, 0xff},
-		.vact_space_2 = {0xff, 0xff},
-		.vact_space_3 = {0xff, 0xff},
-		.vact_space_4 = {0xff, 0xff},
-		.vact_space_5 = {0xff, 0xff},
-		.vact_space_6 = {0xff, 0xff},
-		/* other don't care */
+	{
+		.pixel_clock = 36000000,
+		.conf = {
+			0x01, 0x51, 0x2d, 0x55, 0x40, 0x01, 0x00, 0x08,
+			0x82, 0x80, 0x0e, 0xd9, 0x45, 0xa0, 0xac, 0x80,
+			0x08, 0x80, 0x11, 0x04, 0x02, 0x22, 0x44, 0x86,
+			0x54, 0xab, 0x24, 0x00, 0x00, 0x00, 0x01, 0x80,
+		},
 	},
-	.tg = {
-		0x00, /* cmd */
-		0x50, 0x0a, /* h_fsz */
-		0xd0, 0x02, 0x80, 0x07, /* hact */
-		0x65, 0x04, /* v_fsz */
-		0x01, 0x00, 0x33, 0x02, /* vsync */
-		0x16, 0x00, 0x1c, 0x02, /* vact */
-		0x33, 0x02, /* field_chg */
-		0x49, 0x02, /* vact_st2 */
-		0x00, 0x00, /* vact_st3 */
-		0x00, 0x00, /* vact_st4 */
-		0x01, 0x00, 0x33, 0x02, /* vsync top/bot */
-		0x01, 0x00, 0x33, 0x02, /* field top/bot */
-		0x00, /* 3d FP */
+	{
+		.pixel_clock = 40000000,
+		.conf = {
+			0x01, 0x51, 0x32, 0x55, 0x40, 0x01, 0x00, 0x08,
+			0x82, 0x80, 0x2c, 0xd9, 0x45, 0xa0, 0xac, 0x80,
+			0x08, 0x80, 0x11, 0x04, 0x02, 0x22, 0x44, 0x86,
+			0x54, 0x9a, 0x24, 0x00, 0x00, 0x00, 0x01, 0x80,
+		},
 	},
-};
-
-static const struct hdmi_preset_conf hdmi_conf_1080i60 = {
-	.core = {
-		.h_blank = {0x18, 0x01},
-		.v2_blank = {0x32, 0x02},
-		.v1_blank = {0x16, 0x00},
-		.v_line = {0x65, 0x04},
-		.h_line = {0x98, 0x08},
-		.hsync_pol = {0x00},
-		.vsync_pol = {0x00},
-		.int_pro_mode = {0x01},
-		.v_blank_f0 = {0x49, 0x02},
-		.v_blank_f1 = {0x65, 0x04},
-		.h_sync_start = {0x56, 0x00},
-		.h_sync_end = {0x82, 0x00},
-		.v_sync_line_bef_2 = {0x07, 0x00},
-		.v_sync_line_bef_1 = {0x02, 0x00},
-		.v_sync_line_aft_2 = {0x39, 0x02},
-		.v_sync_line_aft_1 = {0x34, 0x02},
-		.v_sync_line_aft_pxl_2 = {0xa4, 0x04},
-		.v_sync_line_aft_pxl_1 = {0xa4, 0x04},
-		.v_blank_f2 = {0xff, 0xff},
-		.v_blank_f3 = {0xff, 0xff},
-		.v_blank_f4 = {0xff, 0xff},
-		.v_blank_f5 = {0xff, 0xff},
-		.v_sync_line_aft_3 = {0xff, 0xff},
-		.v_sync_line_aft_4 = {0xff, 0xff},
-		.v_sync_line_aft_5 = {0xff, 0xff},
-		.v_sync_line_aft_6 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_3 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_4 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_5 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_6 = {0xff, 0xff},
-		.vact_space_1 = {0xff, 0xff},
-		.vact_space_2 = {0xff, 0xff},
-		.vact_space_3 = {0xff, 0xff},
-		.vact_space_4 = {0xff, 0xff},
-		.vact_space_5 = {0xff, 0xff},
-		.vact_space_6 = {0xff, 0xff},
-		/* other don't care */
+	{
+		.pixel_clock = 65000000,
+		.conf = {
+			0x01, 0xd1, 0x36, 0x34, 0x40, 0x1e, 0x0a, 0x08,
+			0x82, 0xa0, 0x45, 0xd9, 0x45, 0xa0, 0xac, 0x80,
+			0x08, 0x80, 0x11, 0x04, 0x02, 0x22, 0x44, 0x86,
+			0x54, 0xbd, 0x24, 0x01, 0x00, 0x00, 0x01, 0x80,
+		},
 	},
-	.tg = {
-		0x00, /* cmd */
-		0x98, 0x08, /* h_fsz */
-		0x18, 0x01, 0x80, 0x07, /* hact */
-		0x65, 0x04, /* v_fsz */
-		0x01, 0x00, 0x33, 0x02, /* vsync */
-		0x16, 0x00, 0x1c, 0x02, /* vact */
-		0x33, 0x02, /* field_chg */
-		0x49, 0x02, /* vact_st2 */
-		0x00, 0x00, /* vact_st3 */
-		0x00, 0x00, /* vact_st4 */
-		0x01, 0x00, 0x33, 0x02, /* vsync top/bot */
-		0x01, 0x00, 0x33, 0x02, /* field top/bot */
-		0x00, /* 3d FP */
+	{
+		.pixel_clock = 74176000,
+		.conf = {
+			0x01, 0xd1, 0x3e, 0x35, 0x40, 0x5b, 0xde, 0x08,
+			0x82, 0xa0, 0x73, 0xd9, 0x45, 0xa0, 0xac, 0x80,
+			0x56, 0x80, 0x11, 0x04, 0x02, 0x22, 0x44, 0x86,
+			0x54, 0xa6, 0x24, 0x01, 0x00, 0x00, 0x01, 0x80,
+		},
 	},
-};
-
-static const struct hdmi_preset_conf hdmi_conf_1080p30 = {
-	.core = {
-		.h_blank = {0x18, 0x01},
-		.v2_blank = {0x65, 0x04},
-		.v1_blank = {0x2d, 0x00},
-		.v_line = {0x65, 0x04},
-		.h_line = {0x98, 0x08},
-		.hsync_pol = {0x00},
-		.vsync_pol = {0x00},
-		.int_pro_mode = {0x00},
-		.v_blank_f0 = {0xff, 0xff},
-		.v_blank_f1 = {0xff, 0xff},
-		.h_sync_start = {0x56, 0x00},
-		.h_sync_end = {0x82, 0x00},
-		.v_sync_line_bef_2 = {0x09, 0x00},
-		.v_sync_line_bef_1 = {0x04, 0x00},
-		.v_sync_line_aft_2 = {0xff, 0xff},
-		.v_sync_line_aft_1 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_2 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_1 = {0xff, 0xff},
-		.v_blank_f2 = {0xff, 0xff},
-		.v_blank_f3 = {0xff, 0xff},
-		.v_blank_f4 = {0xff, 0xff},
-		.v_blank_f5 = {0xff, 0xff},
-		.v_sync_line_aft_3 = {0xff, 0xff},
-		.v_sync_line_aft_4 = {0xff, 0xff},
-		.v_sync_line_aft_5 = {0xff, 0xff},
-		.v_sync_line_aft_6 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_3 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_4 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_5 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_6 = {0xff, 0xff},
-		.vact_space_1 = {0xff, 0xff},
-		.vact_space_2 = {0xff, 0xff},
-		.vact_space_3 = {0xff, 0xff},
-		.vact_space_4 = {0xff, 0xff},
-		.vact_space_5 = {0xff, 0xff},
-		.vact_space_6 = {0xff, 0xff},
-		/* other don't care */
+	{
+		.pixel_clock = 74250000,
+		.conf = {
+			0x01, 0xd1, 0x1f, 0x10, 0x40, 0x40, 0xf8, 0x08,
+			0x81, 0xa0, 0xba, 0xd8, 0x45, 0xa0, 0xac, 0x80,
+			0x3c, 0x80, 0x11, 0x04, 0x02, 0x22, 0x44, 0x86,
+			0x54, 0xa5, 0x24, 0x01, 0x00, 0x00, 0x01, 0x00,
+		},
 	},
-	.tg = {
-		0x00, /* cmd */
-		0x98, 0x08, /* h_fsz */
-		0x18, 0x01, 0x80, 0x07, /* hact */
-		0x65, 0x04, /* v_fsz */
-		0x01, 0x00, 0x33, 0x02, /* vsync */
-		0x2d, 0x00, 0x38, 0x04, /* vact */
-		0x33, 0x02, /* field_chg */
-		0x48, 0x02, /* vact_st2 */
-		0x00, 0x00, /* vact_st3 */
-		0x00, 0x00, /* vact_st4 */
-		0x01, 0x00, 0x01, 0x00, /* vsync top/bot */
-		0x01, 0x00, 0x33, 0x02, /* field top/bot */
-		0x00, /* 3d FP */
+	{
+		.pixel_clock = 83500000,
+		.conf = {
+			0x01, 0xd1, 0x23, 0x11, 0x40, 0x0c, 0xfb, 0x08,
+			0x85, 0xa0, 0xd1, 0xd8, 0x45, 0xa0, 0xac, 0x80,
+			0x08, 0x80, 0x11, 0x04, 0x02, 0x22, 0x44, 0x86,
+			0x54, 0x93, 0x24, 0x01, 0x00, 0x00, 0x01, 0x80,
+		},
 	},
-};
-
-static const struct hdmi_preset_conf hdmi_conf_1080p50 = {
-	.core = {
-		.h_blank = {0xd0, 0x02},
-		.v2_blank = {0x65, 0x04},
-		.v1_blank = {0x2d, 0x00},
-		.v_line = {0x65, 0x04},
-		.h_line = {0x50, 0x0a},
-		.hsync_pol = {0x00},
-		.vsync_pol = {0x00},
-		.int_pro_mode = {0x00},
-		.v_blank_f0 = {0xff, 0xff},
-		.v_blank_f1 = {0xff, 0xff},
-		.h_sync_start = {0x0e, 0x02},
-		.h_sync_end = {0x3a, 0x02},
-		.v_sync_line_bef_2 = {0x09, 0x00},
-		.v_sync_line_bef_1 = {0x04, 0x00},
-		.v_sync_line_aft_2 = {0xff, 0xff},
-		.v_sync_line_aft_1 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_2 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_1 = {0xff, 0xff},
-		.v_blank_f2 = {0xff, 0xff},
-		.v_blank_f3 = {0xff, 0xff},
-		.v_blank_f4 = {0xff, 0xff},
-		.v_blank_f5 = {0xff, 0xff},
-		.v_sync_line_aft_3 = {0xff, 0xff},
-		.v_sync_line_aft_4 = {0xff, 0xff},
-		.v_sync_line_aft_5 = {0xff, 0xff},
-		.v_sync_line_aft_6 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_3 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_4 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_5 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_6 = {0xff, 0xff},
-		.vact_space_1 = {0xff, 0xff},
-		.vact_space_2 = {0xff, 0xff},
-		.vact_space_3 = {0xff, 0xff},
-		.vact_space_4 = {0xff, 0xff},
-		.vact_space_5 = {0xff, 0xff},
-		.vact_space_6 = {0xff, 0xff},
-		/* other don't care */
+	{
+		.pixel_clock = 106500000,
+		.conf = {
+			0x01, 0xd1, 0x2c, 0x12, 0x40, 0x0c, 0x09, 0x08,
+			0x84, 0xa0, 0x0a, 0xd9, 0x45, 0xa0, 0xac, 0x80,
+			0x08, 0x80, 0x11, 0x04, 0x02, 0x22, 0x44, 0x86,
+			0x54, 0x73, 0x24, 0x01, 0x00, 0x00, 0x01, 0x80,
+		},
 	},
-	.tg = {
-		0x00, /* cmd */
-		0x50, 0x0a, /* h_fsz */
-		0xd0, 0x02, 0x80, 0x07, /* hact */
-		0x65, 0x04, /* v_fsz */
-		0x01, 0x00, 0x33, 0x02, /* vsync */
-		0x2d, 0x00, 0x38, 0x04, /* vact */
-		0x33, 0x02, /* field_chg */
-		0x48, 0x02, /* vact_st2 */
-		0x00, 0x00, /* vact_st3 */
-		0x00, 0x00, /* vact_st4 */
-		0x01, 0x00, 0x01, 0x00, /* vsync top/bot */
-		0x01, 0x00, 0x33, 0x02, /* field top/bot */
-		0x00, /* 3d FP */
+	{
+		.pixel_clock = 108000000,
+		.conf = {
+			0x01, 0x51, 0x2d, 0x15, 0x40, 0x01, 0x00, 0x08,
+			0x82, 0x80, 0x0e, 0xd9, 0x45, 0xa0, 0xac, 0x80,
+			0x08, 0x80, 0x11, 0x04, 0x02, 0x22, 0x44, 0x86,
+			0x54, 0xc7, 0x25, 0x03, 0x00, 0x00, 0x01, 0x80,
+		},
 	},
-};
-
-static const struct hdmi_preset_conf hdmi_conf_1080p60 = {
-	.core = {
-		.h_blank = {0x18, 0x01},
-		.v2_blank = {0x65, 0x04},
-		.v1_blank = {0x2d, 0x00},
-		.v_line = {0x65, 0x04},
-		.h_line = {0x98, 0x08},
-		.hsync_pol = {0x00},
-		.vsync_pol = {0x00},
-		.int_pro_mode = {0x00},
-		.v_blank_f0 = {0xff, 0xff},
-		.v_blank_f1 = {0xff, 0xff},
-		.h_sync_start = {0x56, 0x00},
-		.h_sync_end = {0x82, 0x00},
-		.v_sync_line_bef_2 = {0x09, 0x00},
-		.v_sync_line_bef_1 = {0x04, 0x00},
-		.v_sync_line_aft_2 = {0xff, 0xff},
-		.v_sync_line_aft_1 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_2 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_1 = {0xff, 0xff},
-		.v_blank_f2 = {0xff, 0xff},
-		.v_blank_f3 = {0xff, 0xff},
-		.v_blank_f4 = {0xff, 0xff},
-		.v_blank_f5 = {0xff, 0xff},
-		.v_sync_line_aft_3 = {0xff, 0xff},
-		.v_sync_line_aft_4 = {0xff, 0xff},
-		.v_sync_line_aft_5 = {0xff, 0xff},
-		.v_sync_line_aft_6 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_3 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_4 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_5 = {0xff, 0xff},
-		.v_sync_line_aft_pxl_6 = {0xff, 0xff},
-		/* other don't care */
+	{
+		.pixel_clock = 146250000,
+		.conf = {
+			0x01, 0xd1, 0x3d, 0x15, 0x40, 0x18, 0xfd, 0x08,
+			0x83, 0xa0, 0x6e, 0xd9, 0x45, 0xa0, 0xac, 0x80,
+			0x08, 0x80, 0x11, 0x04, 0x02, 0x22, 0x44, 0x86,
+			0x54, 0x50, 0x25, 0x03, 0x00, 0x00, 0x01, 0x80,
+		},
 	},
-	.tg = {
-		0x00, /* cmd */
-		0x98, 0x08, /* h_fsz */
-		0x18, 0x01, 0x80, 0x07, /* hact */
-		0x65, 0x04, /* v_fsz */
-		0x01, 0x00, 0x33, 0x02, /* vsync */
-		0x2d, 0x00, 0x38, 0x04, /* vact */
-		0x33, 0x02, /* field_chg */
-		0x48, 0x02, /* vact_st2 */
-		0x00, 0x00, /* vact_st3 */
-		0x00, 0x00, /* vact_st4 */
-		0x01, 0x00, 0x01, 0x00, /* vsync top/bot */
-		0x01, 0x00, 0x33, 0x02, /* field top/bot */
-		0x00, /* 3d FP */
+	{
+		.pixel_clock = 148500000,
+		.conf = {
+			0x01, 0xd1, 0x1f, 0x00, 0x40, 0x40, 0xf8, 0x08,
+			0x81, 0xa0, 0xba, 0xd8, 0x45, 0xa0, 0xac, 0x80,
+			0x3c, 0x80, 0x11, 0x04, 0x02, 0x22, 0x44, 0x86,
+			0x54, 0x4b, 0x25, 0x03, 0x00, 0x00, 0x01, 0x00,
+		},
 	},
 };
 
-static const struct hdmi_conf hdmi_confs[] = {
-	{ 720, 480, 60, false, 3, hdmiphy_conf27_027, &hdmi_conf_480p60 },
-	{ 1280, 720, 50, false, 19, hdmiphy_conf74_25, &hdmi_conf_720p50 },
-	{ 1280, 720, 60, false, 4, hdmiphy_conf74_25, &hdmi_conf_720p60 },
-	{ 1920, 1080, 50, true, 20, hdmiphy_conf74_25, &hdmi_conf_1080i50 },
-	{ 1920, 1080, 60, true, 5, hdmiphy_conf74_25, &hdmi_conf_1080i60 },
-	{ 1920, 1080, 30, false, 34, hdmiphy_conf74_176, &hdmi_conf_1080p30 },
-	{ 1920, 1080, 50, false, 31, hdmiphy_conf148_5, &hdmi_conf_1080p50 },
-	{ 1920, 1080, 60, false, 16, hdmiphy_conf148_5, &hdmi_conf_1080p60 },
-};
-
 struct hdmi_infoframe {
 	enum HDMI_PACKET_TYPE type;
 	u8 ver;
@@ -1275,31 +889,6 @@ static int hdmi_v13_conf_index(struct drm_display_mode *mode)
 	return -EINVAL;
 }
 
-static int hdmi_v14_conf_index(struct drm_display_mode *mode)
-{
-	int i;
-
-	for (i = 0; i < ARRAY_SIZE(hdmi_confs); ++i)
-		if (hdmi_confs[i].width == mode->hdisplay &&
-				hdmi_confs[i].height == mode->vdisplay &&
-				hdmi_confs[i].vrefresh == mode->vrefresh &&
-				hdmi_confs[i].interlace ==
-				((mode->flags & DRM_MODE_FLAG_INTERLACE) ?
-				 true : false))
-			return i;
-
-	return -EINVAL;
-}
-
-static int hdmi_conf_index(struct hdmi_context *hdata,
-			   struct drm_display_mode *mode)
-{
-	if (hdata->type == HDMI_TYPE13)
-		return hdmi_v13_conf_index(mode);
-
-	return hdmi_v14_conf_index(mode);
-}
-
 static u8 hdmi_chksum(struct hdmi_context *hdata,
 			u32 start, u8 len, u32 hdr_sum)
 {
@@ -1357,7 +946,7 @@ static void hdmi_reg_infoframe(struct hdmi_context *hdata,
 		if (hdata->type == HDMI_TYPE13)
 			vic = hdmi_v13_confs[hdata->cur_conf].cea_video_id;
 		else
-			vic = hdmi_confs[hdata->cur_conf].cea_video_id;
+			vic = hdata->mode_conf.cea_video_id;
 
 		hdmi_reg_writeb(hdata, HDMI_AVI_BYTE(4), vic);
 
@@ -1434,25 +1023,33 @@ static int hdmi_v13_check_timing(struct fb_videomode *check_timing)
 	return -EINVAL;
 }
 
+static int hdmi_v14_find_phy_conf(int pixel_clock)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(hdmiphy_v14_configs); i++) {
+		if (hdmiphy_v14_configs[i].pixel_clock == pixel_clock)
+			return i;
+	}
+
+	DRM_DEBUG_KMS("Could not find phy config for %d\n", pixel_clock);
+	return -EINVAL;
+}
+
 static int hdmi_v14_check_timing(struct fb_videomode *check_timing)
 {
 	int i;
 
-	DRM_DEBUG_KMS("valid mode : xres=%d, yres=%d, refresh=%d, intl=%d\n",
+	DRM_DEBUG_KMS("mode: xres=%d, yres=%d, refresh=%d, clock=%d, intl=%d\n",
 			check_timing->xres, check_timing->yres,
-			check_timing->refresh, (check_timing->vmode &
-			FB_VMODE_INTERLACED) ? true : false);
+			check_timing->refresh, check_timing->pixclock,
+			(check_timing->vmode & FB_VMODE_INTERLACED) ?
+			true : false);
 
-	for (i = 0; i < ARRAY_SIZE(hdmi_confs); i++)
-		if (hdmi_confs[i].width == check_timing->xres &&
-			hdmi_confs[i].height == check_timing->yres &&
-			hdmi_confs[i].vrefresh == check_timing->refresh &&
-			hdmi_confs[i].interlace ==
-			((check_timing->vmode & FB_VMODE_INTERLACED) ?
-			 true : false))
-				return 0;
-
-	/* TODO */
+	for (i = 0; i < ARRAY_SIZE(hdmiphy_v14_configs); i++)
+		if (hdmiphy_v14_configs[i].pixel_clock ==
+			check_timing->pixclock)
+			return 0;
 
 	return -EINVAL;
 }
@@ -1794,9 +1391,8 @@ static void hdmi_v13_timing_apply(struct hdmi_context *hdata)
 
 static void hdmi_v14_timing_apply(struct hdmi_context *hdata)
 {
-	const struct hdmi_preset_conf *conf = hdmi_confs[hdata->cur_conf].conf;
-	const struct hdmi_core_regs *core = &conf->core;
-	const struct hdmi_tg_regs *tg = &conf->tg;
+	struct hdmi_core_regs *core = &hdata->mode_conf.core;
+	struct hdmi_tg_regs *tg = &hdata->mode_conf.tg;
 	int tries;
 
 	/* setting core registers */
@@ -1899,39 +1495,39 @@ static void hdmi_v14_timing_apply(struct hdmi_context *hdata)
 	hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_6_1, core->vact_space_6[1]);
 
 	/* Timing generator registers */
-	hdmi_reg_writeb(hdata, HDMI_TG_H_FSZ_L, tg->h_fsz_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_H_FSZ_H, tg->h_fsz_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_HACT_ST_L, tg->hact_st_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_HACT_ST_H, tg->hact_st_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_HACT_SZ_L, tg->hact_sz_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_HACT_SZ_H, tg->hact_sz_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_V_FSZ_L, tg->v_fsz_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_V_FSZ_H, tg->v_fsz_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_L, tg->vsync_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_H, tg->vsync_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC2_L, tg->vsync2_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC2_H, tg->vsync2_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST_L, tg->vact_st_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST_H, tg->vact_st_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_VACT_SZ_L, tg->vact_sz_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_VACT_SZ_H, tg->vact_sz_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_FIELD_CHG_L, tg->field_chg_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_FIELD_CHG_H, tg->field_chg_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST2_L, tg->vact_st2_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST2_H, tg->vact_st2_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST3_L, tg->vact_st3_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST3_H, tg->vact_st3_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST4_L, tg->vact_st4_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST4_H, tg->vact_st4_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_TOP_HDMI_L, tg->vsync_top_hdmi_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_TOP_HDMI_H, tg->vsync_top_hdmi_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_BOT_HDMI_L, tg->vsync_bot_hdmi_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_BOT_HDMI_H, tg->vsync_bot_hdmi_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_FIELD_TOP_HDMI_L, tg->field_top_hdmi_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_FIELD_TOP_HDMI_H, tg->field_top_hdmi_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_FIELD_BOT_HDMI_L, tg->field_bot_hdmi_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_FIELD_BOT_HDMI_H, tg->field_bot_hdmi_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_3D, tg->tg_3d);
+	hdmi_reg_writeb(hdata, HDMI_TG_H_FSZ_L, tg->h_fsz[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_H_FSZ_H, tg->h_fsz[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_HACT_ST_L, tg->hact_st[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_HACT_ST_H, tg->hact_st[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_HACT_SZ_L, tg->hact_sz[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_HACT_SZ_H, tg->hact_sz[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_V_FSZ_L, tg->v_fsz[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_V_FSZ_H, tg->v_fsz[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_L, tg->vsync[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_H, tg->vsync[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC2_L, tg->vsync2[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC2_H, tg->vsync2[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST_L, tg->vact_st[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST_H, tg->vact_st[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VACT_SZ_L, tg->vact_sz[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VACT_SZ_H, tg->vact_sz[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_FIELD_CHG_L, tg->field_chg[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_FIELD_CHG_H, tg->field_chg[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST2_L, tg->vact_st2[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST2_H, tg->vact_st2[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST3_L, tg->vact_st3[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST3_H, tg->vact_st3[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST4_L, tg->vact_st4[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST4_H, tg->vact_st4[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_TOP_HDMI_L, tg->vsync_top_hdmi[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_TOP_HDMI_H, tg->vsync_top_hdmi[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_BOT_HDMI_L, tg->vsync_bot_hdmi[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_BOT_HDMI_H, tg->vsync_bot_hdmi[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_FIELD_TOP_HDMI_L, tg->field_top_hdmi[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_FIELD_TOP_HDMI_H, tg->field_top_hdmi[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_FIELD_BOT_HDMI_L, tg->field_bot_hdmi[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_FIELD_BOT_HDMI_H, tg->field_bot_hdmi[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_3D, tg->tg_3d[0]);
 
 	/* waiting for HDMIPHY's PLL to get to steady state */
 	for (tries = 100; tries; --tries) {
@@ -2028,10 +1624,17 @@ static void hdmiphy_conf_apply(struct hdmi_context *hdata)
 	}
 
 	/* pixel clock */
-	if (hdata->type == HDMI_TYPE13)
+	if (hdata->type == HDMI_TYPE13) {
 		hdmiphy_data = hdmi_v13_confs[hdata->cur_conf].hdmiphy_data;
-	else
-		hdmiphy_data = hdmi_confs[hdata->cur_conf].hdmiphy_data;
+	} else {
+		i = hdmi_v14_find_phy_conf(hdata->mode_conf.pixel_clock);
+		if (i < 0) {
+			DRM_ERROR("failed to find hdmiphy conf\n");
+			return;
+		}
+
+		hdmiphy_data = hdmiphy_v14_configs[i].conf;
+	}
 
 	memcpy(buffer, hdmiphy_data, 32);
 	ret = i2c_master_send(hdata->hdmiphy_port, buffer, 32);
@@ -2099,7 +1702,7 @@ static void hdmi_mode_fixup(void *ctx, struct drm_connector *connector,
 	if (hdata->type == HDMI_TYPE13)
 		index = hdmi_v13_conf_index(adjusted_mode);
 	else
-		index = hdmi_v14_conf_index(adjusted_mode);
+		index = hdmi_v14_find_phy_conf(adjusted_mode->clock * 1000);
 
 	/* just return if user desired mode exists. */
 	if (index >= 0)
@@ -2113,7 +1716,7 @@ static void hdmi_mode_fixup(void *ctx, struct drm_connector *connector,
 		if (hdata->type == HDMI_TYPE13)
 			index = hdmi_v13_conf_index(m);
 		else
-			index = hdmi_v14_conf_index(m);
+			index = hdmi_v14_find_phy_conf(m->clock * 1000);
 
 		if (index >= 0) {
 			struct drm_mode_object base;
@@ -2122,6 +1725,9 @@ static void hdmi_mode_fixup(void *ctx, struct drm_connector *connector,
 			DRM_INFO("desired mode doesn't exist so\n");
 			DRM_INFO("use the most suitable mode among modes.\n");
 
+			DRM_DEBUG_KMS("Adjusted Mode: [%d]x[%d] [%d]Hz\n",
+				m->hdisplay, m->vdisplay, m->vrefresh);
+
 			/* preserve display mode header while copying. */
 			head = adjusted_mode->head;
 			base = adjusted_mode->base;
@@ -2133,6 +1739,122 @@ static void hdmi_mode_fixup(void *ctx, struct drm_connector *connector,
 	}
 }
 
+static void hdmi_set_reg(u8 *reg_pair, int num_bytes, u32 value)
+{
+	int i;
+	BUG_ON(num_bytes > 4);
+	for (i = 0; i < num_bytes; i++)
+		reg_pair[i] = (value >> (8 * i)) & 0xff;
+}
+
+static void hdmi_v14_mode_set(struct hdmi_context *hdata,
+			struct drm_display_mode *m)
+{
+	struct hdmi_core_regs *core = &hdata->mode_conf.core;
+	struct hdmi_tg_regs *tg = &hdata->mode_conf.tg;
+
+	hdata->mode_conf.cea_video_id = drm_match_cea_mode(m);
+
+	hdata->mode_conf.pixel_clock = m->clock * 1000;
+	hdmi_set_reg(core->h_blank, 2, m->htotal - m->hdisplay);
+	hdmi_set_reg(core->v_line, 2, m->vtotal);
+	hdmi_set_reg(core->h_line, 2, m->htotal);
+	hdmi_set_reg(core->hsync_pol, 1,
+			(m->flags & DRM_MODE_FLAG_NHSYNC)  ? 1 : 0);
+	hdmi_set_reg(core->vsync_pol, 1,
+			(m->flags & DRM_MODE_FLAG_NVSYNC) ? 1 : 0);
+	hdmi_set_reg(core->int_pro_mode, 1,
+			(m->flags & DRM_MODE_FLAG_INTERLACE) ? 1 : 0);
+
+	/*
+	 * Quirk requirement for exynos 5 HDMI IP design,
+	 * 2 pixels less than the actual calculation for hsync_start
+	 * and end.
+	 */
+
+	/* Following values & calculations differ for different type of modes */
+	if (m->flags & DRM_MODE_FLAG_INTERLACE) {
+		/* Interlaced Mode */
+		hdmi_set_reg(core->v_sync_line_bef_2, 2,
+			(m->vsync_end - m->vdisplay) / 2);
+		hdmi_set_reg(core->v_sync_line_bef_1, 2,
+			(m->vsync_start - m->vdisplay) / 2);
+		hdmi_set_reg(core->v2_blank, 2, m->vtotal / 2);
+		hdmi_set_reg(core->v1_blank, 2, (m->vtotal - m->vdisplay) / 2);
+		hdmi_set_reg(core->v_blank_f0, 2, (m->vtotal +
+			((m->vsync_end - m->vsync_start) * 4) + 5) / 2);
+		hdmi_set_reg(core->v_blank_f1, 2, m->vtotal);
+		hdmi_set_reg(core->v_sync_line_aft_2, 2, (m->vtotal / 2) + 7);
+		hdmi_set_reg(core->v_sync_line_aft_1, 2, (m->vtotal / 2) + 2);
+		hdmi_set_reg(core->v_sync_line_aft_pxl_2, 2,
+			(m->htotal / 2) + (m->hsync_start - m->hdisplay));
+		hdmi_set_reg(core->v_sync_line_aft_pxl_1, 2,
+			(m->htotal / 2) + (m->hsync_start - m->hdisplay));
+		hdmi_set_reg(tg->vact_st, 2, (m->vtotal - m->vdisplay) / 2);
+		hdmi_set_reg(tg->vact_sz, 2, m->vdisplay / 2);
+		hdmi_set_reg(tg->vact_st2, 2, 0x249);/* Reset value + 1*/
+		hdmi_set_reg(tg->vact_st3, 2, 0x0);
+		hdmi_set_reg(tg->vact_st4, 2, 0x0);
+	} else {
+		/* Progressive Mode */
+		hdmi_set_reg(core->v_sync_line_bef_2, 2,
+			m->vsync_end - m->vdisplay);
+		hdmi_set_reg(core->v_sync_line_bef_1, 2,
+			m->vsync_start - m->vdisplay);
+		hdmi_set_reg(core->v2_blank, 2, m->vtotal);
+		hdmi_set_reg(core->v1_blank, 2, m->vtotal - m->vdisplay);
+		hdmi_set_reg(core->v_blank_f0, 2, 0xffff);
+		hdmi_set_reg(core->v_blank_f1, 2, 0xffff);
+		hdmi_set_reg(core->v_sync_line_aft_2, 2, 0xffff);
+		hdmi_set_reg(core->v_sync_line_aft_1, 2, 0xffff);
+		hdmi_set_reg(core->v_sync_line_aft_pxl_2, 2, 0xffff);
+		hdmi_set_reg(core->v_sync_line_aft_pxl_1, 2, 0xffff);
+		hdmi_set_reg(tg->vact_st, 2, m->vtotal - m->vdisplay);
+		hdmi_set_reg(tg->vact_sz, 2, m->vdisplay);
+		hdmi_set_reg(tg->vact_st2, 2, 0x248); /* Reset value */
+		hdmi_set_reg(tg->vact_st3, 2, 0x47b); /* Reset value */
+		hdmi_set_reg(tg->vact_st4, 2, 0x6ae); /* Reset value */
+	}
+
+	/* Following values & calculations are same irrespective of mode type */
+	hdmi_set_reg(core->h_sync_start, 2, m->hsync_start - m->hdisplay - 2);
+	hdmi_set_reg(core->h_sync_end, 2, m->hsync_end - m->hdisplay - 2);
+	hdmi_set_reg(core->vact_space_1, 2, 0xffff);
+	hdmi_set_reg(core->vact_space_2, 2, 0xffff);
+	hdmi_set_reg(core->vact_space_3, 2, 0xffff);
+	hdmi_set_reg(core->vact_space_4, 2, 0xffff);
+	hdmi_set_reg(core->vact_space_5, 2, 0xffff);
+	hdmi_set_reg(core->vact_space_6, 2, 0xffff);
+	hdmi_set_reg(core->v_blank_f2, 2, 0xffff);
+	hdmi_set_reg(core->v_blank_f3, 2, 0xffff);
+	hdmi_set_reg(core->v_blank_f4, 2, 0xffff);
+	hdmi_set_reg(core->v_blank_f5, 2, 0xffff);
+	hdmi_set_reg(core->v_sync_line_aft_3, 2, 0xffff);
+	hdmi_set_reg(core->v_sync_line_aft_4, 2, 0xffff);
+	hdmi_set_reg(core->v_sync_line_aft_5, 2, 0xffff);
+	hdmi_set_reg(core->v_sync_line_aft_6, 2, 0xffff);
+	hdmi_set_reg(core->v_sync_line_aft_pxl_3, 2, 0xffff);
+	hdmi_set_reg(core->v_sync_line_aft_pxl_4, 2, 0xffff);
+	hdmi_set_reg(core->v_sync_line_aft_pxl_5, 2, 0xffff);
+	hdmi_set_reg(core->v_sync_line_aft_pxl_6, 2, 0xffff);
+
+	/* Timing generator registers */
+	hdmi_set_reg(tg->cmd, 1, 0x0);
+	hdmi_set_reg(tg->h_fsz, 2, m->htotal);
+	hdmi_set_reg(tg->hact_st, 2, m->htotal - m->hdisplay);
+	hdmi_set_reg(tg->hact_sz, 2, m->hdisplay);
+	hdmi_set_reg(tg->v_fsz, 2, m->vtotal);
+	hdmi_set_reg(tg->vsync, 2, 0x1);
+	hdmi_set_reg(tg->vsync2, 2, 0x233); /* Reset value */
+	hdmi_set_reg(tg->field_chg, 2, 0x233); /* Reset value */
+	hdmi_set_reg(tg->vsync_top_hdmi, 2, 0x1); /* Reset value */
+	hdmi_set_reg(tg->vsync_bot_hdmi, 2, 0x233); /* Reset value */
+	hdmi_set_reg(tg->field_top_hdmi, 2, 0x1); /* Reset value */
+	hdmi_set_reg(tg->field_bot_hdmi, 2, 0x233); /* Reset value */
+	hdmi_set_reg(tg->tg_3d, 1, 0x0);
+
+}
+
 static void hdmi_mode_set(void *ctx, void *mode)
 {
 	struct hdmi_context *hdata = ctx;
@@ -2140,11 +1862,15 @@ static void hdmi_mode_set(void *ctx, void *mode)
 
 	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
 
-	conf_idx = hdmi_conf_index(hdata, mode);
-	if (conf_idx >= 0)
-		hdata->cur_conf = conf_idx;
-	else
-		DRM_DEBUG_KMS("not supported mode\n");
+	if (hdata->type == HDMI_TYPE13) {
+		conf_idx = hdmi_v13_conf_index(mode);
+		if (conf_idx >= 0)
+			hdata->cur_conf = conf_idx;
+		else
+			DRM_DEBUG_KMS("not supported mode\n");
+	} else {
+		hdmi_v14_mode_set(hdata, mode);
+	}
 }
 
 static void hdmi_get_max_resol(void *ctx, unsigned int *width,
-- 
cgit v0.10.2


From 95fc633735331b3404d0ae65a3bb08d02640dbd0 Mon Sep 17 00:00:00 2001
From: Ajay Kumar <ajaykumar.rs@samsung.com>
Date: Wed, 6 Feb 2013 10:59:44 +0530
Subject: drm/exynos: Add device tree based discovery support for G2D

This patch adds device tree match table for Exynos G2D controller.

Signed-off-by: Ajay Kumar <ajaykumar.rs@samsung.com>
Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

diff --git a/drivers/gpu/drm/exynos/exynos_drm_g2d.c b/drivers/gpu/drm/exynos/exynos_drm_g2d.c
index ddcfb5d..0fcfbe4 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_g2d.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_g2d.c
@@ -19,6 +19,7 @@
 #include <linux/workqueue.h>
 #include <linux/dma-mapping.h>
 #include <linux/dma-attrs.h>
+#include <linux/of.h>
 
 #include <drm/drmP.h>
 #include <drm/exynos_drm.h>
@@ -1240,6 +1241,14 @@ static int g2d_resume(struct device *dev)
 
 static SIMPLE_DEV_PM_OPS(g2d_pm_ops, g2d_suspend, g2d_resume);
 
+#ifdef CONFIG_OF
+static const struct of_device_id exynos_g2d_match[] = {
+	{ .compatible = "samsung,exynos5250-g2d" },
+	{},
+};
+MODULE_DEVICE_TABLE(of, exynos_g2d_match);
+#endif
+
 struct platform_driver g2d_driver = {
 	.probe		= g2d_probe,
 	.remove		= g2d_remove,
@@ -1247,5 +1256,6 @@ struct platform_driver g2d_driver = {
 		.name	= "s5p-g2d",
 		.owner	= THIS_MODULE,
 		.pm	= &g2d_pm_ops,
+		.of_match_table = of_match_ptr(exynos_g2d_match),
 	},
 };
-- 
cgit v0.10.2


From bc752862170c135d6c09fb22d79eeb451023568e Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Thu, 21 Feb 2013 20:04:31 +0000
Subject: drm/i915: Handle untiled planes when computing their offsets
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

We trim the fb to fit the CRTC by computing the offset of that CRTC to
its nearest tile_row origin. This allows us to use framebuffers that are
larger than the CRTC limits without additional work.

However, we failed to compute the offset for a linear framebuffer
correctly as we treated its x-advance in whole tiles (instead of the
linear increment expected), leaving the CRTC misaligned with its
contents.

Fixes regression from commit c2c75131244507c93f812862fdbd4f3a37139401
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Thu Jul 5 12:17:30 2012 +0200

    drm/i915: adjust framebuffer base address on gen4+

v2: Adjust relative x-coordinate after linear alignment (vsyrjala)
v3: Repaint with pokadots (vsyrjala)

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=61152
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: stable@vger.kernel.org
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 6eb3882..0ff10b3 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -2001,18 +2001,29 @@ void intel_unpin_fb_obj(struct drm_i915_gem_object *obj)
 
 /* Computes the linear offset to the base tile and adjusts x, y. bytes per pixel
  * is assumed to be a power-of-two. */
-unsigned long intel_gen4_compute_offset_xtiled(int *x, int *y,
-					       unsigned int bpp,
-					       unsigned int pitch)
+unsigned long intel_gen4_compute_page_offset(int *x, int *y,
+					     unsigned int tiling_mode,
+					     unsigned int cpp,
+					     unsigned int pitch)
 {
-	int tile_rows, tiles;
+	if (tiling_mode != I915_TILING_NONE) {
+		unsigned int tile_rows, tiles;
 
-	tile_rows = *y / 8;
-	*y %= 8;
-	tiles = *x / (512/bpp);
-	*x %= 512/bpp;
+		tile_rows = *y / 8;
+		*y %= 8;
 
-	return tile_rows * pitch * 8 + tiles * 4096;
+		tiles = *x / (512/cpp);
+		*x %= 512/cpp;
+
+		return tile_rows * pitch * 8 + tiles * 4096;
+	} else {
+		unsigned int offset;
+
+		offset = *y * pitch + *x * cpp;
+		*y = 0;
+		*x = (offset & 4095) / cpp;
+		return offset & -4096;
+	}
 }
 
 static int i9xx_update_plane(struct drm_crtc *crtc, struct drm_framebuffer *fb,
@@ -2089,9 +2100,9 @@ static int i9xx_update_plane(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 
 	if (INTEL_INFO(dev)->gen >= 4) {
 		intel_crtc->dspaddr_offset =
-			intel_gen4_compute_offset_xtiled(&x, &y,
-							 fb->bits_per_pixel / 8,
-							 fb->pitches[0]);
+			intel_gen4_compute_page_offset(&x, &y, obj->tiling_mode,
+						       fb->bits_per_pixel / 8,
+						       fb->pitches[0]);
 		linear_offset -= intel_crtc->dspaddr_offset;
 	} else {
 		intel_crtc->dspaddr_offset = linear_offset;
@@ -2182,9 +2193,9 @@ static int ironlake_update_plane(struct drm_crtc *crtc,
 
 	linear_offset = y * fb->pitches[0] + x * (fb->bits_per_pixel / 8);
 	intel_crtc->dspaddr_offset =
-		intel_gen4_compute_offset_xtiled(&x, &y,
-						 fb->bits_per_pixel / 8,
-						 fb->pitches[0]);
+		intel_gen4_compute_page_offset(&x, &y, obj->tiling_mode,
+					       fb->bits_per_pixel / 8,
+					       fb->pitches[0]);
 	linear_offset -= intel_crtc->dspaddr_offset;
 
 	DRM_DEBUG_KMS("Writing base %08X %08lX %d %d %d\n",
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 005a91f..cba0261 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -649,9 +649,10 @@ extern void intel_update_sprite_watermarks(struct drm_device *dev, int pipe,
 extern void intel_update_linetime_watermarks(struct drm_device *dev, int pipe,
 			 struct drm_display_mode *mode);
 
-extern unsigned long intel_gen4_compute_offset_xtiled(int *x, int *y,
-						      unsigned int bpp,
-						      unsigned int pitch);
+extern unsigned long intel_gen4_compute_page_offset(int *x, int *y,
+						    unsigned int tiling_mode,
+						    unsigned int bpp,
+						    unsigned int pitch);
 
 extern int intel_sprite_set_colorkey(struct drm_device *dev, void *data,
 				     struct drm_file *file_priv);
diff --git a/drivers/gpu/drm/i915/intel_sprite.c b/drivers/gpu/drm/i915/intel_sprite.c
index d086e48..1b6eb76 100644
--- a/drivers/gpu/drm/i915/intel_sprite.c
+++ b/drivers/gpu/drm/i915/intel_sprite.c
@@ -122,8 +122,8 @@ ivb_update_plane(struct drm_plane *plane, struct drm_framebuffer *fb,
 
 	linear_offset = y * fb->pitches[0] + x * pixel_size;
 	sprsurf_offset =
-		intel_gen4_compute_offset_xtiled(&x, &y,
-						 pixel_size, fb->pitches[0]);
+		intel_gen4_compute_page_offset(&x, &y, obj->tiling_mode,
+					       pixel_size, fb->pitches[0]);
 	linear_offset -= sprsurf_offset;
 
 	/* HSW consolidates SPRTILEOFF and SPRLINOFF into a single SPROFFSET
@@ -295,8 +295,8 @@ ilk_update_plane(struct drm_plane *plane, struct drm_framebuffer *fb,
 
 	linear_offset = y * fb->pitches[0] + x * pixel_size;
 	dvssurf_offset =
-		intel_gen4_compute_offset_xtiled(&x, &y,
-						 pixel_size, fb->pitches[0]);
+		intel_gen4_compute_page_offset(&x, &y, obj->tiling_mode,
+					       pixel_size, fb->pitches[0]);
 	linear_offset -= dvssurf_offset;
 
 	if (obj->tiling_mode != I915_TILING_NONE)
-- 
cgit v0.10.2


From 907a773ba365daa9f86c6bb8cb11ddbf9b11a3d1 Mon Sep 17 00:00:00 2001
From: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Date: Sat, 16 Feb 2013 13:04:21 +0100
Subject: gma500: Fix n, m1 and m2 clock limits for sdvo and lvds

The values of n, m1 and m2 needs to be subtracted by 2 before writing them to
the FP register. The dot clock calculation already thinks of these values in
register form so we must also specify them as such.

Signed-off-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

diff --git a/drivers/gpu/drm/gma500/psb_intel_display.c b/drivers/gpu/drm/gma500/psb_intel_display.c
index 8033526..9edb190 100644
--- a/drivers/gpu/drm/gma500/psb_intel_display.c
+++ b/drivers/gpu/drm/gma500/psb_intel_display.c
@@ -85,14 +85,14 @@ struct psb_intel_limit_t {
 #define I9XX_DOT_MAX		 400000
 #define I9XX_VCO_MIN		1400000
 #define I9XX_VCO_MAX		2800000
-#define I9XX_N_MIN		      3
-#define I9XX_N_MAX		      8
+#define I9XX_N_MIN		      1
+#define I9XX_N_MAX		      6
 #define I9XX_M_MIN		     70
 #define I9XX_M_MAX		    120
-#define I9XX_M1_MIN		     10
-#define I9XX_M1_MAX		     20
-#define I9XX_M2_MIN		      5
-#define I9XX_M2_MAX		      9
+#define I9XX_M1_MIN		      8
+#define I9XX_M1_MAX		     18
+#define I9XX_M2_MIN		      3
+#define I9XX_M2_MAX		      7
 #define I9XX_P_SDVO_DAC_MIN	      5
 #define I9XX_P_SDVO_DAC_MAX	     80
 #define I9XX_P_LVDS_MIN		      7
-- 
cgit v0.10.2


From 18316c8c39a85c8b6e3db0a150b1bee5b6c4c053 Mon Sep 17 00:00:00 2001
From: Thierry Reding <thierry.reding@avionic-design.de>
Date: Thu, 20 Dec 2012 15:41:44 +0100
Subject: drm: Remove duplicate drm_mode_cea_vic()

The same function had already been merged with a different name. Remove
the duplicate one but reuse some of its kerneldoc fragments for the
existing implementation.

Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
index 67aa0dd..02da919 100644
--- a/drivers/gpu/drm/drm_edid.c
+++ b/drivers/gpu/drm/drm_edid.c
@@ -1516,11 +1516,14 @@ u8 *drm_find_cea_extension(struct edid *edid)
 }
 EXPORT_SYMBOL(drm_find_cea_extension);
 
-/*
- * Looks for a CEA mode matching given drm_display_mode.
- * Returns its CEA Video ID code, or 0 if not found.
+/**
+ * drm_match_cea_mode - look for a CEA mode matching given mode
+ * @to_match: display mode
+ *
+ * Returns the CEA Video ID (VIC) of the mode or 0 if it isn't a CEA-861
+ * mode.
  */
-u8 drm_match_cea_mode(struct drm_display_mode *to_match)
+u8 drm_match_cea_mode(const struct drm_display_mode *to_match)
 {
 	struct drm_display_mode *cea_mode;
 	u8 mode;
@@ -2114,22 +2117,3 @@ int drm_add_modes_noedid(struct drm_connector *connector,
 	return num_modes;
 }
 EXPORT_SYMBOL(drm_add_modes_noedid);
-
-/**
- * drm_mode_cea_vic - return the CEA-861 VIC of a given mode
- * @mode: mode
- *
- * RETURNS:
- * The VIC number, 0 in case it's not a CEA-861 mode.
- */
-uint8_t drm_mode_cea_vic(const struct drm_display_mode *mode)
-{
-	uint8_t i;
-
-	for (i = 0; i < drm_num_cea_modes; i++)
-		if (drm_mode_equal(mode, &edid_cea_modes[i]))
-			return i + 1;
-
-	return 0;
-}
-EXPORT_SYMBOL(drm_mode_cea_vic);
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 31c0205..f61cb79 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -770,7 +770,7 @@ intel_dp_mode_fixup(struct drm_encoder *encoder,
 		 * CEA-861-E - 5.1 Default Encoding Parameters
 		 * VESA DisplayPort Ver.1.2a - 5.1.1.1 Video Colorimetry
 		 */
-		if (bpp != 18 && drm_mode_cea_vic(adjusted_mode) > 1)
+		if (bpp != 18 && drm_match_cea_mode(adjusted_mode) > 1)
 			intel_dp->color_range = DP_COLOR_RANGE_16_235;
 		else
 			intel_dp->color_range = 0;
diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c
index 3ea0c8b..f131dfc 100644
--- a/drivers/gpu/drm/i915/intel_hdmi.c
+++ b/drivers/gpu/drm/i915/intel_hdmi.c
@@ -348,7 +348,7 @@ static void intel_hdmi_set_avi_infoframe(struct drm_encoder *encoder,
 			avi_if.body.avi.ITC_EC_Q_SC |= DIP_AVI_RGB_QUANT_RANGE_FULL;
 	}
 
-	avi_if.body.avi.VIC = drm_mode_cea_vic(adjusted_mode);
+	avi_if.body.avi.VIC = drm_match_cea_mode(adjusted_mode);
 
 	intel_set_infoframe(encoder, &avi_if);
 }
@@ -781,7 +781,7 @@ bool intel_hdmi_mode_fixup(struct drm_encoder *encoder,
 	if (intel_hdmi->color_range_auto) {
 		/* See CEA-861-E - 5.1 Default Encoding Parameters */
 		if (intel_hdmi->has_hdmi_sink &&
-		    drm_mode_cea_vic(adjusted_mode) > 1)
+		    drm_match_cea_mode(adjusted_mode) > 1)
 			intel_hdmi->color_range = SDVO_COLOR_RANGE_16_235;
 		else
 			intel_hdmi->color_range = 0;
diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c
index 33b46d9..d07a8cd 100644
--- a/drivers/gpu/drm/i915/intel_sdvo.c
+++ b/drivers/gpu/drm/i915/intel_sdvo.c
@@ -1077,7 +1077,7 @@ static bool intel_sdvo_mode_fixup(struct drm_encoder *encoder,
 	if (intel_sdvo->color_range_auto) {
 		/* See CEA-861-E - 5.1 Default Encoding Parameters */
 		if (intel_sdvo->has_hdmi_monitor &&
-		    drm_mode_cea_vic(adjusted_mode) > 1)
+		    drm_match_cea_mode(adjusted_mode) > 1)
 			intel_sdvo->color_range = SDVO_COLOR_RANGE_16_235;
 		else
 			intel_sdvo->color_range = 0;
diff --git a/include/drm/drm_crtc.h b/include/drm/drm_crtc.h
index 8b77627..7762718 100644
--- a/include/drm/drm_crtc.h
+++ b/include/drm/drm_crtc.h
@@ -1061,7 +1061,7 @@ extern int drm_mode_gamma_get_ioctl(struct drm_device *dev,
 extern int drm_mode_gamma_set_ioctl(struct drm_device *dev,
 				    void *data, struct drm_file *file_priv);
 extern u8 *drm_find_cea_extension(struct edid *edid);
-extern u8 drm_match_cea_mode(struct drm_display_mode *to_match);
+extern u8 drm_match_cea_mode(const struct drm_display_mode *to_match);
 extern bool drm_detect_hdmi_monitor(struct edid *edid);
 extern bool drm_detect_monitor_audio(struct edid *edid);
 extern bool drm_rgb_quant_range_selectable(struct edid *edid);
@@ -1079,7 +1079,6 @@ extern struct drm_display_mode *drm_gtf_mode_complex(struct drm_device *dev,
 				int GTF_2C, int GTF_K, int GTF_2J);
 extern int drm_add_modes_noedid(struct drm_connector *connector,
 				int hdisplay, int vdisplay);
-extern uint8_t drm_mode_cea_vic(const struct drm_display_mode *mode);
 
 extern int drm_edid_header_is_valid(const u8 *raw_edid);
 extern bool drm_edid_block_valid(u8 *raw_edid, int block, bool print_bad_edid);
-- 
cgit v0.10.2


From a6b21831772db4088bca770dc99697356e82a7e2 Mon Sep 17 00:00:00 2001
From: Thierry Reding <thierry.reding@avionic-design.de>
Date: Fri, 23 Nov 2012 15:01:42 +0100
Subject: drm: Move mode tables to drm_edid.c

The modes are only used from within drm_edid.c so we move them there to
avoid creating duplicates by multiple inclusion of drm_edid_modes.h. As
a side-effect we can also get rid of a few variables that keep track of
the number of entries in the tables and use ARRAY_SIZE() instead.

Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
index 02da919..6995d2f 100644
--- a/drivers/gpu/drm/drm_edid.c
+++ b/drivers/gpu/drm/drm_edid.c
@@ -33,7 +33,6 @@
 #include <linux/module.h>
 #include <drm/drmP.h>
 #include <drm/drm_edid.h>
-#include "drm_edid_modes.h"
 
 #define version_greater(edid, maj, min) \
 	(((edid)->version > (maj)) || \
@@ -127,6 +126,746 @@ static struct edid_quirk {
 	{ "VSC", 5020, EDID_QUIRK_FORCE_REDUCED_BLANKING },
 };
 
+/*
+ * Autogenerated from the DMT spec.
+ * This table is copied from xfree86/modes/xf86EdidModes.c.
+ */
+static const struct drm_display_mode drm_dmt_modes[] = {
+	/* 640x350@85Hz */
+	{ DRM_MODE("640x350", DRM_MODE_TYPE_DRIVER, 31500, 640, 672,
+		   736, 832, 0, 350, 382, 385, 445, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 640x400@85Hz */
+	{ DRM_MODE("640x400", DRM_MODE_TYPE_DRIVER, 31500, 640, 672,
+		   736, 832, 0, 400, 401, 404, 445, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 720x400@85Hz */
+	{ DRM_MODE("720x400", DRM_MODE_TYPE_DRIVER, 35500, 720, 756,
+		   828, 936, 0, 400, 401, 404, 446, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 640x480@60Hz */
+	{ DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 25175, 640, 656,
+		   752, 800, 0, 480, 489, 492, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 640x480@72Hz */
+	{ DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 31500, 640, 664,
+		   704, 832, 0, 480, 489, 492, 520, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 640x480@75Hz */
+	{ DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 31500, 640, 656,
+		   720, 840, 0, 480, 481, 484, 500, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 640x480@85Hz */
+	{ DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 36000, 640, 696,
+		   752, 832, 0, 480, 481, 484, 509, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 800x600@56Hz */
+	{ DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 36000, 800, 824,
+		   896, 1024, 0, 600, 601, 603, 625, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 800x600@60Hz */
+	{ DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 40000, 800, 840,
+		   968, 1056, 0, 600, 601, 605, 628, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 800x600@72Hz */
+	{ DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 50000, 800, 856,
+		   976, 1040, 0, 600, 637, 643, 666, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 800x600@75Hz */
+	{ DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 49500, 800, 816,
+		   896, 1056, 0, 600, 601, 604, 625, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 800x600@85Hz */
+	{ DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 56250, 800, 832,
+		   896, 1048, 0, 600, 601, 604, 631, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 800x600@120Hz RB */
+	{ DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 73250, 800, 848,
+		   880, 960, 0, 600, 603, 607, 636, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 848x480@60Hz */
+	{ DRM_MODE("848x480", DRM_MODE_TYPE_DRIVER, 33750, 848, 864,
+		   976, 1088, 0, 480, 486, 494, 517, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1024x768@43Hz, interlace */
+	{ DRM_MODE("1024x768i", DRM_MODE_TYPE_DRIVER, 44900, 1024, 1032,
+		   1208, 1264, 0, 768, 768, 772, 817, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC |
+			DRM_MODE_FLAG_INTERLACE) },
+	/* 1024x768@60Hz */
+	{ DRM_MODE("1024x768", DRM_MODE_TYPE_DRIVER, 65000, 1024, 1048,
+		   1184, 1344, 0, 768, 771, 777, 806, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1024x768@70Hz */
+	{ DRM_MODE("1024x768", DRM_MODE_TYPE_DRIVER, 75000, 1024, 1048,
+		   1184, 1328, 0, 768, 771, 777, 806, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1024x768@75Hz */
+	{ DRM_MODE("1024x768", DRM_MODE_TYPE_DRIVER, 78750, 1024, 1040,
+		   1136, 1312, 0, 768, 769, 772, 800, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1024x768@85Hz */
+	{ DRM_MODE("1024x768", DRM_MODE_TYPE_DRIVER, 94500, 1024, 1072,
+		   1168, 1376, 0, 768, 769, 772, 808, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1024x768@120Hz RB */
+	{ DRM_MODE("1024x768", DRM_MODE_TYPE_DRIVER, 115500, 1024, 1072,
+		   1104, 1184, 0, 768, 771, 775, 813, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1152x864@75Hz */
+	{ DRM_MODE("1152x864", DRM_MODE_TYPE_DRIVER, 108000, 1152, 1216,
+		   1344, 1600, 0, 864, 865, 868, 900, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1280x768@60Hz RB */
+	{ DRM_MODE("1280x768", DRM_MODE_TYPE_DRIVER, 68250, 1280, 1328,
+		   1360, 1440, 0, 768, 771, 778, 790, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1280x768@60Hz */
+	{ DRM_MODE("1280x768", DRM_MODE_TYPE_DRIVER, 79500, 1280, 1344,
+		   1472, 1664, 0, 768, 771, 778, 798, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1280x768@75Hz */
+	{ DRM_MODE("1280x768", DRM_MODE_TYPE_DRIVER, 102250, 1280, 1360,
+		   1488, 1696, 0, 768, 771, 778, 805, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1280x768@85Hz */
+	{ DRM_MODE("1280x768", DRM_MODE_TYPE_DRIVER, 117500, 1280, 1360,
+		   1496, 1712, 0, 768, 771, 778, 809, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1280x768@120Hz RB */
+	{ DRM_MODE("1280x768", DRM_MODE_TYPE_DRIVER, 140250, 1280, 1328,
+		   1360, 1440, 0, 768, 771, 778, 813, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1280x800@60Hz RB */
+	{ DRM_MODE("1280x800", DRM_MODE_TYPE_DRIVER, 71000, 1280, 1328,
+		   1360, 1440, 0, 800, 803, 809, 823, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1280x800@60Hz */
+	{ DRM_MODE("1280x800", DRM_MODE_TYPE_DRIVER, 83500, 1280, 1352,
+		   1480, 1680, 0, 800, 803, 809, 831, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1280x800@75Hz */
+	{ DRM_MODE("1280x800", DRM_MODE_TYPE_DRIVER, 106500, 1280, 1360,
+		   1488, 1696, 0, 800, 803, 809, 838, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1280x800@85Hz */
+	{ DRM_MODE("1280x800", DRM_MODE_TYPE_DRIVER, 122500, 1280, 1360,
+		   1496, 1712, 0, 800, 803, 809, 843, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1280x800@120Hz RB */
+	{ DRM_MODE("1280x800", DRM_MODE_TYPE_DRIVER, 146250, 1280, 1328,
+		   1360, 1440, 0, 800, 803, 809, 847, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1280x960@60Hz */
+	{ DRM_MODE("1280x960", DRM_MODE_TYPE_DRIVER, 108000, 1280, 1376,
+		   1488, 1800, 0, 960, 961, 964, 1000, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1280x960@85Hz */
+	{ DRM_MODE("1280x960", DRM_MODE_TYPE_DRIVER, 148500, 1280, 1344,
+		   1504, 1728, 0, 960, 961, 964, 1011, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1280x960@120Hz RB */
+	{ DRM_MODE("1280x960", DRM_MODE_TYPE_DRIVER, 175500, 1280, 1328,
+		   1360, 1440, 0, 960, 963, 967, 1017, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1280x1024@60Hz */
+	{ DRM_MODE("1280x1024", DRM_MODE_TYPE_DRIVER, 108000, 1280, 1328,
+		   1440, 1688, 0, 1024, 1025, 1028, 1066, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1280x1024@75Hz */
+	{ DRM_MODE("1280x1024", DRM_MODE_TYPE_DRIVER, 135000, 1280, 1296,
+		   1440, 1688, 0, 1024, 1025, 1028, 1066, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1280x1024@85Hz */
+	{ DRM_MODE("1280x1024", DRM_MODE_TYPE_DRIVER, 157500, 1280, 1344,
+		   1504, 1728, 0, 1024, 1025, 1028, 1072, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1280x1024@120Hz RB */
+	{ DRM_MODE("1280x1024", DRM_MODE_TYPE_DRIVER, 187250, 1280, 1328,
+		   1360, 1440, 0, 1024, 1027, 1034, 1084, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1360x768@60Hz */
+	{ DRM_MODE("1360x768", DRM_MODE_TYPE_DRIVER, 85500, 1360, 1424,
+		   1536, 1792, 0, 768, 771, 777, 795, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1360x768@120Hz RB */
+	{ DRM_MODE("1360x768", DRM_MODE_TYPE_DRIVER, 148250, 1360, 1408,
+		   1440, 1520, 0, 768, 771, 776, 813, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1400x1050@60Hz RB */
+	{ DRM_MODE("1400x1050", DRM_MODE_TYPE_DRIVER, 101000, 1400, 1448,
+		   1480, 1560, 0, 1050, 1053, 1057, 1080, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1400x1050@60Hz */
+	{ DRM_MODE("1400x1050", DRM_MODE_TYPE_DRIVER, 121750, 1400, 1488,
+		   1632, 1864, 0, 1050, 1053, 1057, 1089, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1400x1050@75Hz */
+	{ DRM_MODE("1400x1050", DRM_MODE_TYPE_DRIVER, 156000, 1400, 1504,
+		   1648, 1896, 0, 1050, 1053, 1057, 1099, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1400x1050@85Hz */
+	{ DRM_MODE("1400x1050", DRM_MODE_TYPE_DRIVER, 179500, 1400, 1504,
+		   1656, 1912, 0, 1050, 1053, 1057, 1105, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1400x1050@120Hz RB */
+	{ DRM_MODE("1400x1050", DRM_MODE_TYPE_DRIVER, 208000, 1400, 1448,
+		   1480, 1560, 0, 1050, 1053, 1057, 1112, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1440x900@60Hz RB */
+	{ DRM_MODE("1440x900", DRM_MODE_TYPE_DRIVER, 88750, 1440, 1488,
+		   1520, 1600, 0, 900, 903, 909, 926, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1440x900@60Hz */
+	{ DRM_MODE("1440x900", DRM_MODE_TYPE_DRIVER, 106500, 1440, 1520,
+		   1672, 1904, 0, 900, 903, 909, 934, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1440x900@75Hz */
+	{ DRM_MODE("1440x900", DRM_MODE_TYPE_DRIVER, 136750, 1440, 1536,
+		   1688, 1936, 0, 900, 903, 909, 942, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1440x900@85Hz */
+	{ DRM_MODE("1440x900", DRM_MODE_TYPE_DRIVER, 157000, 1440, 1544,
+		   1696, 1952, 0, 900, 903, 909, 948, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1440x900@120Hz RB */
+	{ DRM_MODE("1440x900", DRM_MODE_TYPE_DRIVER, 182750, 1440, 1488,
+		   1520, 1600, 0, 900, 903, 909, 953, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1600x1200@60Hz */
+	{ DRM_MODE("1600x1200", DRM_MODE_TYPE_DRIVER, 162000, 1600, 1664,
+		   1856, 2160, 0, 1200, 1201, 1204, 1250, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1600x1200@65Hz */
+	{ DRM_MODE("1600x1200", DRM_MODE_TYPE_DRIVER, 175500, 1600, 1664,
+		   1856, 2160, 0, 1200, 1201, 1204, 1250, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1600x1200@70Hz */
+	{ DRM_MODE("1600x1200", DRM_MODE_TYPE_DRIVER, 189000, 1600, 1664,
+		   1856, 2160, 0, 1200, 1201, 1204, 1250, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1600x1200@75Hz */
+	{ DRM_MODE("1600x1200", DRM_MODE_TYPE_DRIVER, 202500, 1600, 1664,
+		   1856, 2160, 0, 1200, 1201, 1204, 1250, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1600x1200@85Hz */
+	{ DRM_MODE("1600x1200", DRM_MODE_TYPE_DRIVER, 229500, 1600, 1664,
+		   1856, 2160, 0, 1200, 1201, 1204, 1250, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1600x1200@120Hz RB */
+	{ DRM_MODE("1600x1200", DRM_MODE_TYPE_DRIVER, 268250, 1600, 1648,
+		   1680, 1760, 0, 1200, 1203, 1207, 1271, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1680x1050@60Hz RB */
+	{ DRM_MODE("1680x1050", DRM_MODE_TYPE_DRIVER, 119000, 1680, 1728,
+		   1760, 1840, 0, 1050, 1053, 1059, 1080, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1680x1050@60Hz */
+	{ DRM_MODE("1680x1050", DRM_MODE_TYPE_DRIVER, 146250, 1680, 1784,
+		   1960, 2240, 0, 1050, 1053, 1059, 1089, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1680x1050@75Hz */
+	{ DRM_MODE("1680x1050", DRM_MODE_TYPE_DRIVER, 187000, 1680, 1800,
+		   1976, 2272, 0, 1050, 1053, 1059, 1099, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1680x1050@85Hz */
+	{ DRM_MODE("1680x1050", DRM_MODE_TYPE_DRIVER, 214750, 1680, 1808,
+		   1984, 2288, 0, 1050, 1053, 1059, 1105, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1680x1050@120Hz RB */
+	{ DRM_MODE("1680x1050", DRM_MODE_TYPE_DRIVER, 245500, 1680, 1728,
+		   1760, 1840, 0, 1050, 1053, 1059, 1112, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1792x1344@60Hz */
+	{ DRM_MODE("1792x1344", DRM_MODE_TYPE_DRIVER, 204750, 1792, 1920,
+		   2120, 2448, 0, 1344, 1345, 1348, 1394, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1792x1344@75Hz */
+	{ DRM_MODE("1792x1344", DRM_MODE_TYPE_DRIVER, 261000, 1792, 1888,
+		   2104, 2456, 0, 1344, 1345, 1348, 1417, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1792x1344@120Hz RB */
+	{ DRM_MODE("1792x1344", DRM_MODE_TYPE_DRIVER, 333250, 1792, 1840,
+		   1872, 1952, 0, 1344, 1347, 1351, 1423, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1856x1392@60Hz */
+	{ DRM_MODE("1856x1392", DRM_MODE_TYPE_DRIVER, 218250, 1856, 1952,
+		   2176, 2528, 0, 1392, 1393, 1396, 1439, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1856x1392@75Hz */
+	{ DRM_MODE("1856x1392", DRM_MODE_TYPE_DRIVER, 288000, 1856, 1984,
+		   2208, 2560, 0, 1392, 1395, 1399, 1500, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1856x1392@120Hz RB */
+	{ DRM_MODE("1856x1392", DRM_MODE_TYPE_DRIVER, 356500, 1856, 1904,
+		   1936, 2016, 0, 1392, 1395, 1399, 1474, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1920x1200@60Hz RB */
+	{ DRM_MODE("1920x1200", DRM_MODE_TYPE_DRIVER, 154000, 1920, 1968,
+		   2000, 2080, 0, 1200, 1203, 1209, 1235, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1920x1200@60Hz */
+	{ DRM_MODE("1920x1200", DRM_MODE_TYPE_DRIVER, 193250, 1920, 2056,
+		   2256, 2592, 0, 1200, 1203, 1209, 1245, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1920x1200@75Hz */
+	{ DRM_MODE("1920x1200", DRM_MODE_TYPE_DRIVER, 245250, 1920, 2056,
+		   2264, 2608, 0, 1200, 1203, 1209, 1255, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1920x1200@85Hz */
+	{ DRM_MODE("1920x1200", DRM_MODE_TYPE_DRIVER, 281250, 1920, 2064,
+		   2272, 2624, 0, 1200, 1203, 1209, 1262, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1920x1200@120Hz RB */
+	{ DRM_MODE("1920x1200", DRM_MODE_TYPE_DRIVER, 317000, 1920, 1968,
+		   2000, 2080, 0, 1200, 1203, 1209, 1271, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1920x1440@60Hz */
+	{ DRM_MODE("1920x1440", DRM_MODE_TYPE_DRIVER, 234000, 1920, 2048,
+		   2256, 2600, 0, 1440, 1441, 1444, 1500, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1920x1440@75Hz */
+	{ DRM_MODE("1920x1440", DRM_MODE_TYPE_DRIVER, 297000, 1920, 2064,
+		   2288, 2640, 0, 1440, 1441, 1444, 1500, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1920x1440@120Hz RB */
+	{ DRM_MODE("1920x1440", DRM_MODE_TYPE_DRIVER, 380500, 1920, 1968,
+		   2000, 2080, 0, 1440, 1443, 1447, 1525, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 2560x1600@60Hz RB */
+	{ DRM_MODE("2560x1600", DRM_MODE_TYPE_DRIVER, 268500, 2560, 2608,
+		   2640, 2720, 0, 1600, 1603, 1609, 1646, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 2560x1600@60Hz */
+	{ DRM_MODE("2560x1600", DRM_MODE_TYPE_DRIVER, 348500, 2560, 2752,
+		   3032, 3504, 0, 1600, 1603, 1609, 1658, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 2560x1600@75HZ */
+	{ DRM_MODE("2560x1600", DRM_MODE_TYPE_DRIVER, 443250, 2560, 2768,
+		   3048, 3536, 0, 1600, 1603, 1609, 1672, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 2560x1600@85HZ */
+	{ DRM_MODE("2560x1600", DRM_MODE_TYPE_DRIVER, 505250, 2560, 2768,
+		   3048, 3536, 0, 1600, 1603, 1609, 1682, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 2560x1600@120Hz RB */
+	{ DRM_MODE("2560x1600", DRM_MODE_TYPE_DRIVER, 552750, 2560, 2608,
+		   2640, 2720, 0, 1600, 1603, 1609, 1694, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+};
+
+static const struct drm_display_mode edid_est_modes[] = {
+	{ DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 40000, 800, 840,
+		   968, 1056, 0, 600, 601, 605, 628, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, /* 800x600@60Hz */
+	{ DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 36000, 800, 824,
+		   896, 1024, 0, 600, 601, 603,  625, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, /* 800x600@56Hz */
+	{ DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 31500, 640, 656,
+		   720, 840, 0, 480, 481, 484, 500, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, /* 640x480@75Hz */
+	{ DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 31500, 640, 664,
+		   704,  832, 0, 480, 489, 491, 520, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, /* 640x480@72Hz */
+	{ DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 30240, 640, 704,
+		   768,  864, 0, 480, 483, 486, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, /* 640x480@67Hz */
+	{ DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 25200, 640, 656,
+		   752, 800, 0, 480, 490, 492, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, /* 640x480@60Hz */
+	{ DRM_MODE("720x400", DRM_MODE_TYPE_DRIVER, 35500, 720, 738,
+		   846, 900, 0, 400, 421, 423,  449, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, /* 720x400@88Hz */
+	{ DRM_MODE("720x400", DRM_MODE_TYPE_DRIVER, 28320, 720, 738,
+		   846,  900, 0, 400, 412, 414, 449, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) }, /* 720x400@70Hz */
+	{ DRM_MODE("1280x1024", DRM_MODE_TYPE_DRIVER, 135000, 1280, 1296,
+		   1440, 1688, 0, 1024, 1025, 1028, 1066, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, /* 1280x1024@75Hz */
+	{ DRM_MODE("1024x768", DRM_MODE_TYPE_DRIVER, 78800, 1024, 1040,
+		   1136, 1312, 0,  768, 769, 772, 800, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, /* 1024x768@75Hz */
+	{ DRM_MODE("1024x768", DRM_MODE_TYPE_DRIVER, 75000, 1024, 1048,
+		   1184, 1328, 0,  768, 771, 777, 806, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, /* 1024x768@70Hz */
+	{ DRM_MODE("1024x768", DRM_MODE_TYPE_DRIVER, 65000, 1024, 1048,
+		   1184, 1344, 0,  768, 771, 777, 806, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, /* 1024x768@60Hz */
+	{ DRM_MODE("1024x768i", DRM_MODE_TYPE_DRIVER,44900, 1024, 1032,
+		   1208, 1264, 0, 768, 768, 776, 817, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC | DRM_MODE_FLAG_INTERLACE) }, /* 1024x768@43Hz */
+	{ DRM_MODE("832x624", DRM_MODE_TYPE_DRIVER, 57284, 832, 864,
+		   928, 1152, 0, 624, 625, 628, 667, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, /* 832x624@75Hz */
+	{ DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 49500, 800, 816,
+		   896, 1056, 0, 600, 601, 604,  625, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, /* 800x600@75Hz */
+	{ DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 50000, 800, 856,
+		   976, 1040, 0, 600, 637, 643, 666, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, /* 800x600@72Hz */
+	{ DRM_MODE("1152x864", DRM_MODE_TYPE_DRIVER, 108000, 1152, 1216,
+		   1344, 1600, 0,  864, 865, 868, 900, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, /* 1152x864@75Hz */
+};
+
+struct minimode {
+	short w;
+	short h;
+	short r;
+	short rb;
+};
+
+static const struct minimode est3_modes[] = {
+	/* byte 6 */
+	{ 640, 350, 85, 0 },
+	{ 640, 400, 85, 0 },
+	{ 720, 400, 85, 0 },
+	{ 640, 480, 85, 0 },
+	{ 848, 480, 60, 0 },
+	{ 800, 600, 85, 0 },
+	{ 1024, 768, 85, 0 },
+	{ 1152, 864, 75, 0 },
+	/* byte 7 */
+	{ 1280, 768, 60, 1 },
+	{ 1280, 768, 60, 0 },
+	{ 1280, 768, 75, 0 },
+	{ 1280, 768, 85, 0 },
+	{ 1280, 960, 60, 0 },
+	{ 1280, 960, 85, 0 },
+	{ 1280, 1024, 60, 0 },
+	{ 1280, 1024, 85, 0 },
+	/* byte 8 */
+	{ 1360, 768, 60, 0 },
+	{ 1440, 900, 60, 1 },
+	{ 1440, 900, 60, 0 },
+	{ 1440, 900, 75, 0 },
+	{ 1440, 900, 85, 0 },
+	{ 1400, 1050, 60, 1 },
+	{ 1400, 1050, 60, 0 },
+	{ 1400, 1050, 75, 0 },
+	/* byte 9 */
+	{ 1400, 1050, 85, 0 },
+	{ 1680, 1050, 60, 1 },
+	{ 1680, 1050, 60, 0 },
+	{ 1680, 1050, 75, 0 },
+	{ 1680, 1050, 85, 0 },
+	{ 1600, 1200, 60, 0 },
+	{ 1600, 1200, 65, 0 },
+	{ 1600, 1200, 70, 0 },
+	/* byte 10 */
+	{ 1600, 1200, 75, 0 },
+	{ 1600, 1200, 85, 0 },
+	{ 1792, 1344, 60, 0 },
+	{ 1792, 1344, 85, 0 },
+	{ 1856, 1392, 60, 0 },
+	{ 1856, 1392, 75, 0 },
+	{ 1920, 1200, 60, 1 },
+	{ 1920, 1200, 60, 0 },
+	/* byte 11 */
+	{ 1920, 1200, 75, 0 },
+	{ 1920, 1200, 85, 0 },
+	{ 1920, 1440, 60, 0 },
+	{ 1920, 1440, 75, 0 },
+};
+
+static const struct minimode extra_modes[] = {
+	{ 1024, 576,  60, 0 },
+	{ 1366, 768,  60, 0 },
+	{ 1600, 900,  60, 0 },
+	{ 1680, 945,  60, 0 },
+	{ 1920, 1080, 60, 0 },
+	{ 2048, 1152, 60, 0 },
+	{ 2048, 1536, 60, 0 },
+};
+
+/*
+ * Probably taken from CEA-861 spec.
+ * This table is converted from xorg's hw/xfree86/modes/xf86EdidModes.c.
+ */
+static const struct drm_display_mode edid_cea_modes[] = {
+	/* 1 - 640x480@60Hz */
+	{ DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 25175, 640, 656,
+		   752, 800, 0, 480, 490, 492, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 2 - 720x480@60Hz */
+	{ DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 27000, 720, 736,
+		   798, 858, 0, 480, 489, 495, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 3 - 720x480@60Hz */
+	{ DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 27000, 720, 736,
+		   798, 858, 0, 480, 489, 495, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 4 - 1280x720@60Hz */
+	{ DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 74250, 1280, 1390,
+		   1430, 1650, 0, 720, 725, 730, 750, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 5 - 1920x1080i@60Hz */
+	{ DRM_MODE("1920x1080i", DRM_MODE_TYPE_DRIVER, 74250, 1920, 2008,
+		   2052, 2200, 0, 1080, 1084, 1094, 1125, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC |
+			DRM_MODE_FLAG_INTERLACE) },
+	/* 6 - 1440x480i@60Hz */
+	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1478,
+		   1602, 1716, 0, 480, 488, 494, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
+			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
+	/* 7 - 1440x480i@60Hz */
+	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1478,
+		   1602, 1716, 0, 480, 488, 494, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
+			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
+	/* 8 - 1440x240@60Hz */
+	{ DRM_MODE("1440x240", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1478,
+		   1602, 1716, 0, 240, 244, 247, 262, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
+			DRM_MODE_FLAG_DBLCLK) },
+	/* 9 - 1440x240@60Hz */
+	{ DRM_MODE("1440x240", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1478,
+		   1602, 1716, 0, 240, 244, 247, 262, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
+			DRM_MODE_FLAG_DBLCLK) },
+	/* 10 - 2880x480i@60Hz */
+	{ DRM_MODE("2880x480i", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2956,
+		   3204, 3432, 0, 480, 488, 494, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
+			DRM_MODE_FLAG_INTERLACE) },
+	/* 11 - 2880x480i@60Hz */
+	{ DRM_MODE("2880x480i", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2956,
+		   3204, 3432, 0, 480, 488, 494, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
+			DRM_MODE_FLAG_INTERLACE) },
+	/* 12 - 2880x240@60Hz */
+	{ DRM_MODE("2880x240", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2956,
+		   3204, 3432, 0, 240, 244, 247, 262, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 13 - 2880x240@60Hz */
+	{ DRM_MODE("2880x240", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2956,
+		   3204, 3432, 0, 240, 244, 247, 262, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 14 - 1440x480@60Hz */
+	{ DRM_MODE("1440x480", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1472,
+		   1596, 1716, 0, 480, 489, 495, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 15 - 1440x480@60Hz */
+	{ DRM_MODE("1440x480", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1472,
+		   1596, 1716, 0, 480, 489, 495, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 16 - 1920x1080@60Hz */
+	{ DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 148500, 1920, 2008,
+		   2052, 2200, 0, 1080, 1084, 1089, 1125, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 17 - 720x576@50Hz */
+	{ DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 27000, 720, 732,
+		   796, 864, 0, 576, 581, 586, 625, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 18 - 720x576@50Hz */
+	{ DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 27000, 720, 732,
+		   796, 864, 0, 576, 581, 586, 625, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 19 - 1280x720@50Hz */
+	{ DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 74250, 1280, 1720,
+		   1760, 1980, 0, 720, 725, 730, 750, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 20 - 1920x1080i@50Hz */
+	{ DRM_MODE("1920x1080i", DRM_MODE_TYPE_DRIVER, 74250, 1920, 2448,
+		   2492, 2640, 0, 1080, 1084, 1094, 1125, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC |
+			DRM_MODE_FLAG_INTERLACE) },
+	/* 21 - 1440x576i@50Hz */
+	{ DRM_MODE("1440x576i", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1464,
+		   1590, 1728, 0, 576, 580, 586, 625, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
+			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
+	/* 22 - 1440x576i@50Hz */
+	{ DRM_MODE("1440x576i", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1464,
+		   1590, 1728, 0, 576, 580, 586, 625, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
+			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
+	/* 23 - 1440x288@50Hz */
+	{ DRM_MODE("1440x288", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1464,
+		   1590, 1728, 0, 288, 290, 293, 312, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
+			DRM_MODE_FLAG_DBLCLK) },
+	/* 24 - 1440x288@50Hz */
+	{ DRM_MODE("1440x288", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1464,
+		   1590, 1728, 0, 288, 290, 293, 312, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
+			DRM_MODE_FLAG_DBLCLK) },
+	/* 25 - 2880x576i@50Hz */
+	{ DRM_MODE("2880x576i", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2928,
+		   3180, 3456, 0, 576, 580, 586, 625, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
+			DRM_MODE_FLAG_INTERLACE) },
+	/* 26 - 2880x576i@50Hz */
+	{ DRM_MODE("2880x576i", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2928,
+		   3180, 3456, 0, 576, 580, 586, 625, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
+			DRM_MODE_FLAG_INTERLACE) },
+	/* 27 - 2880x288@50Hz */
+	{ DRM_MODE("2880x288", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2928,
+		   3180, 3456, 0, 288, 290, 293, 312, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 28 - 2880x288@50Hz */
+	{ DRM_MODE("2880x288", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2928,
+		   3180, 3456, 0, 288, 290, 293, 312, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 29 - 1440x576@50Hz */
+	{ DRM_MODE("1440x576", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1464,
+		   1592, 1728, 0, 576, 581, 586, 625, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 30 - 1440x576@50Hz */
+	{ DRM_MODE("1440x576", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1464,
+		   1592, 1728, 0, 576, 581, 586, 625, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 31 - 1920x1080@50Hz */
+	{ DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 148500, 1920, 2448,
+		   2492, 2640, 0, 1080, 1084, 1089, 1125, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 32 - 1920x1080@24Hz */
+	{ DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 74250, 1920, 2558,
+		   2602, 2750, 0, 1080, 1084, 1089, 1125, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 33 - 1920x1080@25Hz */
+	{ DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 74250, 1920, 2448,
+		   2492, 2640, 0, 1080, 1084, 1089, 1125, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 34 - 1920x1080@30Hz */
+	{ DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 74250, 1920, 2008,
+		   2052, 2200, 0, 1080, 1084, 1089, 1125, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 35 - 2880x480@60Hz */
+	{ DRM_MODE("2880x480", DRM_MODE_TYPE_DRIVER, 108000, 2880, 2944,
+		   3192, 3432, 0, 480, 489, 495, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 36 - 2880x480@60Hz */
+	{ DRM_MODE("2880x480", DRM_MODE_TYPE_DRIVER, 108000, 2880, 2944,
+		   3192, 3432, 0, 480, 489, 495, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 37 - 2880x576@50Hz */
+	{ DRM_MODE("2880x576", DRM_MODE_TYPE_DRIVER, 108000, 2880, 2928,
+		   3184, 3456, 0, 576, 581, 586, 625, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 38 - 2880x576@50Hz */
+	{ DRM_MODE("2880x576", DRM_MODE_TYPE_DRIVER, 108000, 2880, 2928,
+		   3184, 3456, 0, 576, 581, 586, 625, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 39 - 1920x1080i@50Hz */
+	{ DRM_MODE("1920x1080i", DRM_MODE_TYPE_DRIVER, 72000, 1920, 1952,
+		   2120, 2304, 0, 1080, 1126, 1136, 1250, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC |
+			DRM_MODE_FLAG_INTERLACE) },
+	/* 40 - 1920x1080i@100Hz */
+	{ DRM_MODE("1920x1080i", DRM_MODE_TYPE_DRIVER, 148500, 1920, 2448,
+		   2492, 2640, 0, 1080, 1084, 1094, 1125, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC |
+			DRM_MODE_FLAG_INTERLACE) },
+	/* 41 - 1280x720@100Hz */
+	{ DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 148500, 1280, 1720,
+		   1760, 1980, 0, 720, 725, 730, 750, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 42 - 720x576@100Hz */
+	{ DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 54000, 720, 732,
+		   796, 864, 0, 576, 581, 586, 625, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 43 - 720x576@100Hz */
+	{ DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 54000, 720, 732,
+		   796, 864, 0, 576, 581, 586, 625, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 44 - 1440x576i@100Hz */
+	{ DRM_MODE("1440x576", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1464,
+		   1590, 1728, 0, 576, 580, 586, 625, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
+			DRM_MODE_FLAG_DBLCLK) },
+	/* 45 - 1440x576i@100Hz */
+	{ DRM_MODE("1440x576", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1464,
+		   1590, 1728, 0, 576, 580, 586, 625, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
+			DRM_MODE_FLAG_DBLCLK) },
+	/* 46 - 1920x1080i@120Hz */
+	{ DRM_MODE("1920x1080i", DRM_MODE_TYPE_DRIVER, 148500, 1920, 2008,
+		   2052, 2200, 0, 1080, 1084, 1094, 1125, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC |
+			DRM_MODE_FLAG_INTERLACE) },
+	/* 47 - 1280x720@120Hz */
+	{ DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 148500, 1280, 1390,
+		   1430, 1650, 0, 720, 725, 730, 750, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 48 - 720x480@120Hz */
+	{ DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 54000, 720, 736,
+		   798, 858, 0, 480, 489, 495, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 49 - 720x480@120Hz */
+	{ DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 54000, 720, 736,
+		   798, 858, 0, 480, 489, 495, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 50 - 1440x480i@120Hz */
+	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1478,
+		   1602, 1716, 0, 480, 488, 494, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
+			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
+	/* 51 - 1440x480i@120Hz */
+	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1478,
+		   1602, 1716, 0, 480, 488, 494, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
+			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
+	/* 52 - 720x576@200Hz */
+	{ DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 108000, 720, 732,
+		   796, 864, 0, 576, 581, 586, 625, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 53 - 720x576@200Hz */
+	{ DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 108000, 720, 732,
+		   796, 864, 0, 576, 581, 586, 625, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 54 - 1440x576i@200Hz */
+	{ DRM_MODE("1440x576i", DRM_MODE_TYPE_DRIVER, 108000, 1440, 1464,
+		   1590, 1728, 0, 576, 580, 586, 625, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
+			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
+	/* 55 - 1440x576i@200Hz */
+	{ DRM_MODE("1440x576i", DRM_MODE_TYPE_DRIVER, 108000, 1440, 1464,
+		   1590, 1728, 0, 576, 580, 586, 625, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
+			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
+	/* 56 - 720x480@240Hz */
+	{ DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 108000, 720, 736,
+		   798, 858, 0, 480, 489, 495, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 57 - 720x480@240Hz */
+	{ DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 108000, 720, 736,
+		   798, 858, 0, 480, 489, 495, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 58 - 1440x480i@240 */
+	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 108000, 1440, 1478,
+		   1602, 1716, 0, 480, 488, 494, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
+			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
+	/* 59 - 1440x480i@240 */
+	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 108000, 1440, 1478,
+		   1602, 1716, 0, 480, 488, 494, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
+			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
+	/* 60 - 1280x720@24Hz */
+	{ DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 59400, 1280, 3040,
+		   3080, 3300, 0, 720, 725, 730, 750, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 61 - 1280x720@25Hz */
+	{ DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 74250, 1280, 3700,
+		   3740, 3960, 0, 720, 725, 730, 750, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 62 - 1280x720@30Hz */
+	{ DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 74250, 1280, 3040,
+		   3080, 3300, 0, 720, 725, 730, 750, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 63 - 1920x1080@120Hz */
+	{ DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 297000, 1920, 2008,
+		   2052, 2200, 0, 1080, 1084, 1089, 1125, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 64 - 1920x1080@100Hz */
+	{ DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 297000, 1920, 2448,
+		   2492, 2640, 0, 1080, 1084, 1094, 1125, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+};
+
 /*** DDC fetch and block validation ***/
 
 static const u8 edid_header[] = {
@@ -542,7 +1281,7 @@ struct drm_display_mode *drm_mode_find_dmt(struct drm_device *dev,
 {
 	int i;
 
-	for (i = 0; i < drm_num_dmt_modes; i++) {
+	for (i = 0; i < ARRAY_SIZE(drm_dmt_modes); i++) {
 		const struct drm_display_mode *ptr = &drm_dmt_modes[i];
 		if (hsize != ptr->hdisplay)
 			continue;
@@ -1083,7 +1822,7 @@ drm_dmt_modes_for_range(struct drm_connector *connector, struct edid *edid,
 	struct drm_display_mode *newmode;
 	struct drm_device *dev = connector->dev;
 
-	for (i = 0; i < drm_num_dmt_modes; i++) {
+	for (i = 0; i < ARRAY_SIZE(drm_dmt_modes); i++) {
 		if (mode_in_range(drm_dmt_modes + i, edid, timing) &&
 		    valid_inferred_mode(connector, drm_dmt_modes + i)) {
 			newmode = drm_mode_duplicate(dev, &drm_dmt_modes[i]);
@@ -1118,7 +1857,7 @@ drm_gtf_modes_for_range(struct drm_connector *connector, struct edid *edid,
 	struct drm_display_mode *newmode;
 	struct drm_device *dev = connector->dev;
 
-	for (i = 0; i < num_extra_modes; i++) {
+	for (i = 0; i < ARRAY_SIZE(extra_modes); i++) {
 		const struct minimode *m = &extra_modes[i];
 		newmode = drm_gtf_mode(dev, m->w, m->h, m->r, 0, 0);
 		if (!newmode)
@@ -1147,7 +1886,7 @@ drm_cvt_modes_for_range(struct drm_connector *connector, struct edid *edid,
 	struct drm_device *dev = connector->dev;
 	bool rb = drm_monitor_supports_rb(edid);
 
-	for (i = 0; i < num_extra_modes; i++) {
+	for (i = 0; i < ARRAY_SIZE(extra_modes); i++) {
 		const struct minimode *m = &extra_modes[i];
 		newmode = drm_cvt_mode(dev, m->w, m->h, m->r, rb, 0, 0);
 		if (!newmode)
@@ -1528,7 +2267,7 @@ u8 drm_match_cea_mode(const struct drm_display_mode *to_match)
 	struct drm_display_mode *cea_mode;
 	u8 mode;
 
-	for (mode = 0; mode < drm_num_cea_modes; mode++) {
+	for (mode = 0; mode < ARRAY_SIZE(edid_cea_modes); mode++) {
 		cea_mode = (struct drm_display_mode *)&edid_cea_modes[mode];
 
 		if (drm_mode_equal(to_match, cea_mode))
@@ -1548,7 +2287,7 @@ do_cea_modes (struct drm_connector *connector, u8 *db, u8 len)
 
 	for (mode = db; mode < db + len; mode++) {
 		cea_mode = (*mode & 127) - 1; /* CEA modes are numbered 1..127 */
-		if (cea_mode < drm_num_cea_modes) {
+		if (cea_mode < ARRAY_SIZE(edid_cea_modes)) {
 			struct drm_display_mode *newmode;
 			newmode = drm_mode_duplicate(dev,
 						     &edid_cea_modes[cea_mode]);
diff --git a/drivers/gpu/drm/drm_edid_modes.h b/drivers/gpu/drm/drm_edid_modes.h
deleted file mode 100644
index 5dbf7d2..0000000
--- a/drivers/gpu/drm/drm_edid_modes.h
+++ /dev/null
@@ -1,774 +0,0 @@
-/*
- * Copyright (c) 2007-2008 Intel Corporation
- *   Jesse Barnes <jesse.barnes@intel.com>
- * Copyright 2010 Red Hat, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sub license,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice (including the
- * next paragraph) shall be included in all copies or substantial portions
- * of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
- * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
- * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
- * DEALINGS IN THE SOFTWARE.
- */
-
-#include <linux/kernel.h>
-#include <drm/drmP.h>
-#include <drm/drm_edid.h>
-
-/*
- * Autogenerated from the DMT spec.
- * This table is copied from xfree86/modes/xf86EdidModes.c.
- */
-static const struct drm_display_mode drm_dmt_modes[] = {
-	/* 640x350@85Hz */
-	{ DRM_MODE("640x350", DRM_MODE_TYPE_DRIVER, 31500, 640, 672,
-		   736, 832, 0, 350, 382, 385, 445, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 640x400@85Hz */
-	{ DRM_MODE("640x400", DRM_MODE_TYPE_DRIVER, 31500, 640, 672,
-		   736, 832, 0, 400, 401, 404, 445, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 720x400@85Hz */
-	{ DRM_MODE("720x400", DRM_MODE_TYPE_DRIVER, 35500, 720, 756,
-		   828, 936, 0, 400, 401, 404, 446, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 640x480@60Hz */
-	{ DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 25175, 640, 656,
-		   752, 800, 0, 480, 489, 492, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 640x480@72Hz */
-	{ DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 31500, 640, 664,
-		   704, 832, 0, 480, 489, 492, 520, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 640x480@75Hz */
-	{ DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 31500, 640, 656,
-		   720, 840, 0, 480, 481, 484, 500, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 640x480@85Hz */
-	{ DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 36000, 640, 696,
-		   752, 832, 0, 480, 481, 484, 509, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 800x600@56Hz */
-	{ DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 36000, 800, 824,
-		   896, 1024, 0, 600, 601, 603, 625, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 800x600@60Hz */
-	{ DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 40000, 800, 840,
-		   968, 1056, 0, 600, 601, 605, 628, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 800x600@72Hz */
-	{ DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 50000, 800, 856,
-		   976, 1040, 0, 600, 637, 643, 666, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 800x600@75Hz */
-	{ DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 49500, 800, 816,
-		   896, 1056, 0, 600, 601, 604, 625, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 800x600@85Hz */
-	{ DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 56250, 800, 832,
-		   896, 1048, 0, 600, 601, 604, 631, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 800x600@120Hz RB */
-	{ DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 73250, 800, 848,
-		   880, 960, 0, 600, 603, 607, 636, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 848x480@60Hz */
-	{ DRM_MODE("848x480", DRM_MODE_TYPE_DRIVER, 33750, 848, 864,
-		   976, 1088, 0, 480, 486, 494, 517, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1024x768@43Hz, interlace */
-	{ DRM_MODE("1024x768i", DRM_MODE_TYPE_DRIVER, 44900, 1024, 1032,
-		   1208, 1264, 0, 768, 768, 772, 817, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC |
-			DRM_MODE_FLAG_INTERLACE) },
-	/* 1024x768@60Hz */
-	{ DRM_MODE("1024x768", DRM_MODE_TYPE_DRIVER, 65000, 1024, 1048,
-		   1184, 1344, 0, 768, 771, 777, 806, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1024x768@70Hz */
-	{ DRM_MODE("1024x768", DRM_MODE_TYPE_DRIVER, 75000, 1024, 1048,
-		   1184, 1328, 0, 768, 771, 777, 806, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1024x768@75Hz */
-	{ DRM_MODE("1024x768", DRM_MODE_TYPE_DRIVER, 78750, 1024, 1040,
-		   1136, 1312, 0, 768, 769, 772, 800, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1024x768@85Hz */
-	{ DRM_MODE("1024x768", DRM_MODE_TYPE_DRIVER, 94500, 1024, 1072,
-		   1168, 1376, 0, 768, 769, 772, 808, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1024x768@120Hz RB */
-	{ DRM_MODE("1024x768", DRM_MODE_TYPE_DRIVER, 115500, 1024, 1072,
-		   1104, 1184, 0, 768, 771, 775, 813, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1152x864@75Hz */
-	{ DRM_MODE("1152x864", DRM_MODE_TYPE_DRIVER, 108000, 1152, 1216,
-		   1344, 1600, 0, 864, 865, 868, 900, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1280x768@60Hz RB */
-	{ DRM_MODE("1280x768", DRM_MODE_TYPE_DRIVER, 68250, 1280, 1328,
-		   1360, 1440, 0, 768, 771, 778, 790, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1280x768@60Hz */
-	{ DRM_MODE("1280x768", DRM_MODE_TYPE_DRIVER, 79500, 1280, 1344,
-		   1472, 1664, 0, 768, 771, 778, 798, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1280x768@75Hz */
-	{ DRM_MODE("1280x768", DRM_MODE_TYPE_DRIVER, 102250, 1280, 1360,
-		   1488, 1696, 0, 768, 771, 778, 805, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1280x768@85Hz */
-	{ DRM_MODE("1280x768", DRM_MODE_TYPE_DRIVER, 117500, 1280, 1360,
-		   1496, 1712, 0, 768, 771, 778, 809, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1280x768@120Hz RB */
-	{ DRM_MODE("1280x768", DRM_MODE_TYPE_DRIVER, 140250, 1280, 1328,
-		   1360, 1440, 0, 768, 771, 778, 813, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1280x800@60Hz RB */
-	{ DRM_MODE("1280x800", DRM_MODE_TYPE_DRIVER, 71000, 1280, 1328,
-		   1360, 1440, 0, 800, 803, 809, 823, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1280x800@60Hz */
-	{ DRM_MODE("1280x800", DRM_MODE_TYPE_DRIVER, 83500, 1280, 1352,
-		   1480, 1680, 0, 800, 803, 809, 831, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1280x800@75Hz */
-	{ DRM_MODE("1280x800", DRM_MODE_TYPE_DRIVER, 106500, 1280, 1360,
-		   1488, 1696, 0, 800, 803, 809, 838, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1280x800@85Hz */
-	{ DRM_MODE("1280x800", DRM_MODE_TYPE_DRIVER, 122500, 1280, 1360,
-		   1496, 1712, 0, 800, 803, 809, 843, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1280x800@120Hz RB */
-	{ DRM_MODE("1280x800", DRM_MODE_TYPE_DRIVER, 146250, 1280, 1328,
-		   1360, 1440, 0, 800, 803, 809, 847, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1280x960@60Hz */
-	{ DRM_MODE("1280x960", DRM_MODE_TYPE_DRIVER, 108000, 1280, 1376,
-		   1488, 1800, 0, 960, 961, 964, 1000, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1280x960@85Hz */
-	{ DRM_MODE("1280x960", DRM_MODE_TYPE_DRIVER, 148500, 1280, 1344,
-		   1504, 1728, 0, 960, 961, 964, 1011, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1280x960@120Hz RB */
-	{ DRM_MODE("1280x960", DRM_MODE_TYPE_DRIVER, 175500, 1280, 1328,
-		   1360, 1440, 0, 960, 963, 967, 1017, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1280x1024@60Hz */
-	{ DRM_MODE("1280x1024", DRM_MODE_TYPE_DRIVER, 108000, 1280, 1328,
-		   1440, 1688, 0, 1024, 1025, 1028, 1066, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1280x1024@75Hz */
-	{ DRM_MODE("1280x1024", DRM_MODE_TYPE_DRIVER, 135000, 1280, 1296,
-		   1440, 1688, 0, 1024, 1025, 1028, 1066, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1280x1024@85Hz */
-	{ DRM_MODE("1280x1024", DRM_MODE_TYPE_DRIVER, 157500, 1280, 1344,
-		   1504, 1728, 0, 1024, 1025, 1028, 1072, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1280x1024@120Hz RB */
-	{ DRM_MODE("1280x1024", DRM_MODE_TYPE_DRIVER, 187250, 1280, 1328,
-		   1360, 1440, 0, 1024, 1027, 1034, 1084, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1360x768@60Hz */
-	{ DRM_MODE("1360x768", DRM_MODE_TYPE_DRIVER, 85500, 1360, 1424,
-		   1536, 1792, 0, 768, 771, 777, 795, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1360x768@120Hz RB */
-	{ DRM_MODE("1360x768", DRM_MODE_TYPE_DRIVER, 148250, 1360, 1408,
-		   1440, 1520, 0, 768, 771, 776, 813, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1400x1050@60Hz RB */
-	{ DRM_MODE("1400x1050", DRM_MODE_TYPE_DRIVER, 101000, 1400, 1448,
-		   1480, 1560, 0, 1050, 1053, 1057, 1080, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1400x1050@60Hz */
-	{ DRM_MODE("1400x1050", DRM_MODE_TYPE_DRIVER, 121750, 1400, 1488,
-		   1632, 1864, 0, 1050, 1053, 1057, 1089, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1400x1050@75Hz */
-	{ DRM_MODE("1400x1050", DRM_MODE_TYPE_DRIVER, 156000, 1400, 1504,
-		   1648, 1896, 0, 1050, 1053, 1057, 1099, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1400x1050@85Hz */
-	{ DRM_MODE("1400x1050", DRM_MODE_TYPE_DRIVER, 179500, 1400, 1504,
-		   1656, 1912, 0, 1050, 1053, 1057, 1105, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1400x1050@120Hz RB */
-	{ DRM_MODE("1400x1050", DRM_MODE_TYPE_DRIVER, 208000, 1400, 1448,
-		   1480, 1560, 0, 1050, 1053, 1057, 1112, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1440x900@60Hz RB */
-	{ DRM_MODE("1440x900", DRM_MODE_TYPE_DRIVER, 88750, 1440, 1488,
-		   1520, 1600, 0, 900, 903, 909, 926, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1440x900@60Hz */
-	{ DRM_MODE("1440x900", DRM_MODE_TYPE_DRIVER, 106500, 1440, 1520,
-		   1672, 1904, 0, 900, 903, 909, 934, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1440x900@75Hz */
-	{ DRM_MODE("1440x900", DRM_MODE_TYPE_DRIVER, 136750, 1440, 1536,
-		   1688, 1936, 0, 900, 903, 909, 942, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1440x900@85Hz */
-	{ DRM_MODE("1440x900", DRM_MODE_TYPE_DRIVER, 157000, 1440, 1544,
-		   1696, 1952, 0, 900, 903, 909, 948, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1440x900@120Hz RB */
-	{ DRM_MODE("1440x900", DRM_MODE_TYPE_DRIVER, 182750, 1440, 1488,
-		   1520, 1600, 0, 900, 903, 909, 953, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1600x1200@60Hz */
-	{ DRM_MODE("1600x1200", DRM_MODE_TYPE_DRIVER, 162000, 1600, 1664,
-		   1856, 2160, 0, 1200, 1201, 1204, 1250, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1600x1200@65Hz */
-	{ DRM_MODE("1600x1200", DRM_MODE_TYPE_DRIVER, 175500, 1600, 1664,
-		   1856, 2160, 0, 1200, 1201, 1204, 1250, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1600x1200@70Hz */
-	{ DRM_MODE("1600x1200", DRM_MODE_TYPE_DRIVER, 189000, 1600, 1664,
-		   1856, 2160, 0, 1200, 1201, 1204, 1250, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1600x1200@75Hz */
-	{ DRM_MODE("1600x1200", DRM_MODE_TYPE_DRIVER, 202500, 1600, 1664,
-		   1856, 2160, 0, 1200, 1201, 1204, 1250, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1600x1200@85Hz */
-	{ DRM_MODE("1600x1200", DRM_MODE_TYPE_DRIVER, 229500, 1600, 1664,
-		   1856, 2160, 0, 1200, 1201, 1204, 1250, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1600x1200@120Hz RB */
-	{ DRM_MODE("1600x1200", DRM_MODE_TYPE_DRIVER, 268250, 1600, 1648,
-		   1680, 1760, 0, 1200, 1203, 1207, 1271, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1680x1050@60Hz RB */
-	{ DRM_MODE("1680x1050", DRM_MODE_TYPE_DRIVER, 119000, 1680, 1728,
-		   1760, 1840, 0, 1050, 1053, 1059, 1080, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1680x1050@60Hz */
-	{ DRM_MODE("1680x1050", DRM_MODE_TYPE_DRIVER, 146250, 1680, 1784,
-		   1960, 2240, 0, 1050, 1053, 1059, 1089, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1680x1050@75Hz */
-	{ DRM_MODE("1680x1050", DRM_MODE_TYPE_DRIVER, 187000, 1680, 1800,
-		   1976, 2272, 0, 1050, 1053, 1059, 1099, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1680x1050@85Hz */
-	{ DRM_MODE("1680x1050", DRM_MODE_TYPE_DRIVER, 214750, 1680, 1808,
-		   1984, 2288, 0, 1050, 1053, 1059, 1105, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1680x1050@120Hz RB */
-	{ DRM_MODE("1680x1050", DRM_MODE_TYPE_DRIVER, 245500, 1680, 1728,
-		   1760, 1840, 0, 1050, 1053, 1059, 1112, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1792x1344@60Hz */
-	{ DRM_MODE("1792x1344", DRM_MODE_TYPE_DRIVER, 204750, 1792, 1920,
-		   2120, 2448, 0, 1344, 1345, 1348, 1394, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1792x1344@75Hz */
-	{ DRM_MODE("1792x1344", DRM_MODE_TYPE_DRIVER, 261000, 1792, 1888,
-		   2104, 2456, 0, 1344, 1345, 1348, 1417, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1792x1344@120Hz RB */
-	{ DRM_MODE("1792x1344", DRM_MODE_TYPE_DRIVER, 333250, 1792, 1840,
-		   1872, 1952, 0, 1344, 1347, 1351, 1423, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1856x1392@60Hz */
-	{ DRM_MODE("1856x1392", DRM_MODE_TYPE_DRIVER, 218250, 1856, 1952,
-		   2176, 2528, 0, 1392, 1393, 1396, 1439, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1856x1392@75Hz */
-	{ DRM_MODE("1856x1392", DRM_MODE_TYPE_DRIVER, 288000, 1856, 1984,
-		   2208, 2560, 0, 1392, 1395, 1399, 1500, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1856x1392@120Hz RB */
-	{ DRM_MODE("1856x1392", DRM_MODE_TYPE_DRIVER, 356500, 1856, 1904,
-		   1936, 2016, 0, 1392, 1395, 1399, 1474, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1920x1200@60Hz RB */
-	{ DRM_MODE("1920x1200", DRM_MODE_TYPE_DRIVER, 154000, 1920, 1968,
-		   2000, 2080, 0, 1200, 1203, 1209, 1235, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1920x1200@60Hz */
-	{ DRM_MODE("1920x1200", DRM_MODE_TYPE_DRIVER, 193250, 1920, 2056,
-		   2256, 2592, 0, 1200, 1203, 1209, 1245, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1920x1200@75Hz */
-	{ DRM_MODE("1920x1200", DRM_MODE_TYPE_DRIVER, 245250, 1920, 2056,
-		   2264, 2608, 0, 1200, 1203, 1209, 1255, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1920x1200@85Hz */
-	{ DRM_MODE("1920x1200", DRM_MODE_TYPE_DRIVER, 281250, 1920, 2064,
-		   2272, 2624, 0, 1200, 1203, 1209, 1262, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1920x1200@120Hz RB */
-	{ DRM_MODE("1920x1200", DRM_MODE_TYPE_DRIVER, 317000, 1920, 1968,
-		   2000, 2080, 0, 1200, 1203, 1209, 1271, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1920x1440@60Hz */
-	{ DRM_MODE("1920x1440", DRM_MODE_TYPE_DRIVER, 234000, 1920, 2048,
-		   2256, 2600, 0, 1440, 1441, 1444, 1500, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1920x1440@75Hz */
-	{ DRM_MODE("1920x1440", DRM_MODE_TYPE_DRIVER, 297000, 1920, 2064,
-		   2288, 2640, 0, 1440, 1441, 1444, 1500, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1920x1440@120Hz RB */
-	{ DRM_MODE("1920x1440", DRM_MODE_TYPE_DRIVER, 380500, 1920, 1968,
-		   2000, 2080, 0, 1440, 1443, 1447, 1525, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 2560x1600@60Hz RB */
-	{ DRM_MODE("2560x1600", DRM_MODE_TYPE_DRIVER, 268500, 2560, 2608,
-		   2640, 2720, 0, 1600, 1603, 1609, 1646, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 2560x1600@60Hz */
-	{ DRM_MODE("2560x1600", DRM_MODE_TYPE_DRIVER, 348500, 2560, 2752,
-		   3032, 3504, 0, 1600, 1603, 1609, 1658, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 2560x1600@75HZ */
-	{ DRM_MODE("2560x1600", DRM_MODE_TYPE_DRIVER, 443250, 2560, 2768,
-		   3048, 3536, 0, 1600, 1603, 1609, 1672, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 2560x1600@85HZ */
-	{ DRM_MODE("2560x1600", DRM_MODE_TYPE_DRIVER, 505250, 2560, 2768,
-		   3048, 3536, 0, 1600, 1603, 1609, 1682, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 2560x1600@120Hz RB */
-	{ DRM_MODE("2560x1600", DRM_MODE_TYPE_DRIVER, 552750, 2560, 2608,
-		   2640, 2720, 0, 1600, 1603, 1609, 1694, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-
-};
-static const int drm_num_dmt_modes =
-	sizeof(drm_dmt_modes) / sizeof(struct drm_display_mode);
-
-static const struct drm_display_mode edid_est_modes[] = {
-	{ DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 40000, 800, 840,
-		   968, 1056, 0, 600, 601, 605, 628, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, /* 800x600@60Hz */
-	{ DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 36000, 800, 824,
-		   896, 1024, 0, 600, 601, 603,  625, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, /* 800x600@56Hz */
-	{ DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 31500, 640, 656,
-		   720, 840, 0, 480, 481, 484, 500, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, /* 640x480@75Hz */
-	{ DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 31500, 640, 664,
-		   704,  832, 0, 480, 489, 491, 520, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, /* 640x480@72Hz */
-	{ DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 30240, 640, 704,
-		   768,  864, 0, 480, 483, 486, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, /* 640x480@67Hz */
-	{ DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 25200, 640, 656,
-		   752, 800, 0, 480, 490, 492, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, /* 640x480@60Hz */
-	{ DRM_MODE("720x400", DRM_MODE_TYPE_DRIVER, 35500, 720, 738,
-		   846, 900, 0, 400, 421, 423,  449, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, /* 720x400@88Hz */
-	{ DRM_MODE("720x400", DRM_MODE_TYPE_DRIVER, 28320, 720, 738,
-		   846,  900, 0, 400, 412, 414, 449, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) }, /* 720x400@70Hz */
-	{ DRM_MODE("1280x1024", DRM_MODE_TYPE_DRIVER, 135000, 1280, 1296,
-		   1440, 1688, 0, 1024, 1025, 1028, 1066, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, /* 1280x1024@75Hz */
-	{ DRM_MODE("1024x768", DRM_MODE_TYPE_DRIVER, 78800, 1024, 1040,
-		   1136, 1312, 0,  768, 769, 772, 800, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, /* 1024x768@75Hz */
-	{ DRM_MODE("1024x768", DRM_MODE_TYPE_DRIVER, 75000, 1024, 1048,
-		   1184, 1328, 0,  768, 771, 777, 806, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, /* 1024x768@70Hz */
-	{ DRM_MODE("1024x768", DRM_MODE_TYPE_DRIVER, 65000, 1024, 1048,
-		   1184, 1344, 0,  768, 771, 777, 806, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, /* 1024x768@60Hz */
-	{ DRM_MODE("1024x768i", DRM_MODE_TYPE_DRIVER,44900, 1024, 1032,
-		   1208, 1264, 0, 768, 768, 776, 817, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC | DRM_MODE_FLAG_INTERLACE) }, /* 1024x768@43Hz */
-	{ DRM_MODE("832x624", DRM_MODE_TYPE_DRIVER, 57284, 832, 864,
-		   928, 1152, 0, 624, 625, 628, 667, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, /* 832x624@75Hz */
-	{ DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 49500, 800, 816,
-		   896, 1056, 0, 600, 601, 604,  625, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, /* 800x600@75Hz */
-	{ DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 50000, 800, 856,
-		   976, 1040, 0, 600, 637, 643, 666, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, /* 800x600@72Hz */
-	{ DRM_MODE("1152x864", DRM_MODE_TYPE_DRIVER, 108000, 1152, 1216,
-		   1344, 1600, 0,  864, 865, 868, 900, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, /* 1152x864@75Hz */
-};
-
-struct minimode {
-	short w;
-	short h;
-	short r;
-	short rb;
-};
-
-static const struct minimode est3_modes[] = {
-	/* byte 6 */
-	{ 640, 350, 85, 0 },
-	{ 640, 400, 85, 0 },
-	{ 720, 400, 85, 0 },
-	{ 640, 480, 85, 0 },
-	{ 848, 480, 60, 0 },
-	{ 800, 600, 85, 0 },
-	{ 1024, 768, 85, 0 },
-	{ 1152, 864, 75, 0 },
-	/* byte 7 */
-	{ 1280, 768, 60, 1 },
-	{ 1280, 768, 60, 0 },
-	{ 1280, 768, 75, 0 },
-	{ 1280, 768, 85, 0 },
-	{ 1280, 960, 60, 0 },
-	{ 1280, 960, 85, 0 },
-	{ 1280, 1024, 60, 0 },
-	{ 1280, 1024, 85, 0 },
-	/* byte 8 */
-	{ 1360, 768, 60, 0 },
-	{ 1440, 900, 60, 1 },
-	{ 1440, 900, 60, 0 },
-	{ 1440, 900, 75, 0 },
-	{ 1440, 900, 85, 0 },
-	{ 1400, 1050, 60, 1 },
-	{ 1400, 1050, 60, 0 },
-	{ 1400, 1050, 75, 0 },
-	/* byte 9 */
-	{ 1400, 1050, 85, 0 },
-	{ 1680, 1050, 60, 1 },
-	{ 1680, 1050, 60, 0 },
-	{ 1680, 1050, 75, 0 },
-	{ 1680, 1050, 85, 0 },
-	{ 1600, 1200, 60, 0 },
-	{ 1600, 1200, 65, 0 },
-	{ 1600, 1200, 70, 0 },
-	/* byte 10 */
-	{ 1600, 1200, 75, 0 },
-	{ 1600, 1200, 85, 0 },
-	{ 1792, 1344, 60, 0 },
-	{ 1792, 1344, 85, 0 },
-	{ 1856, 1392, 60, 0 },
-	{ 1856, 1392, 75, 0 },
-	{ 1920, 1200, 60, 1 },
-	{ 1920, 1200, 60, 0 },
-	/* byte 11 */
-	{ 1920, 1200, 75, 0 },
-	{ 1920, 1200, 85, 0 },
-	{ 1920, 1440, 60, 0 },
-	{ 1920, 1440, 75, 0 },
-};
-static const int num_est3_modes = ARRAY_SIZE(est3_modes);
-
-static const struct minimode extra_modes[] = {
-	{ 1024, 576,  60, 0 },
-	{ 1366, 768,  60, 0 },
-	{ 1600, 900,  60, 0 },
-	{ 1680, 945,  60, 0 },
-	{ 1920, 1080, 60, 0 },
-	{ 2048, 1152, 60, 0 },
-	{ 2048, 1536, 60, 0 },
-};
-static const int num_extra_modes = ARRAY_SIZE(extra_modes);
-
-/*
- * Probably taken from CEA-861 spec.
- * This table is converted from xorg's hw/xfree86/modes/xf86EdidModes.c.
- */
-static const struct drm_display_mode edid_cea_modes[] = {
-	/* 1 - 640x480@60Hz */
-	{ DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 25175, 640, 656,
-		   752, 800, 0, 480, 490, 492, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 2 - 720x480@60Hz */
-	{ DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 27000, 720, 736,
-		   798, 858, 0, 480, 489, 495, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 3 - 720x480@60Hz */
-	{ DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 27000, 720, 736,
-		   798, 858, 0, 480, 489, 495, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 4 - 1280x720@60Hz */
-	{ DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 74250, 1280, 1390,
-		   1430, 1650, 0, 720, 725, 730, 750, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 5 - 1920x1080i@60Hz */
-	{ DRM_MODE("1920x1080i", DRM_MODE_TYPE_DRIVER, 74250, 1920, 2008,
-		   2052, 2200, 0, 1080, 1084, 1094, 1125, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC |
-			DRM_MODE_FLAG_INTERLACE) },
-	/* 6 - 1440x480i@60Hz */
-	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1478,
-		   1602, 1716, 0, 480, 488, 494, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
-	/* 7 - 1440x480i@60Hz */
-	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1478,
-		   1602, 1716, 0, 480, 488, 494, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
-	/* 8 - 1440x240@60Hz */
-	{ DRM_MODE("1440x240", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1478,
-		   1602, 1716, 0, 240, 244, 247, 262, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_DBLCLK) },
-	/* 9 - 1440x240@60Hz */
-	{ DRM_MODE("1440x240", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1478,
-		   1602, 1716, 0, 240, 244, 247, 262, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_DBLCLK) },
-	/* 10 - 2880x480i@60Hz */
-	{ DRM_MODE("2880x480i", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2956,
-		   3204, 3432, 0, 480, 488, 494, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE) },
-	/* 11 - 2880x480i@60Hz */
-	{ DRM_MODE("2880x480i", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2956,
-		   3204, 3432, 0, 480, 488, 494, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE) },
-	/* 12 - 2880x240@60Hz */
-	{ DRM_MODE("2880x240", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2956,
-		   3204, 3432, 0, 240, 244, 247, 262, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 13 - 2880x240@60Hz */
-	{ DRM_MODE("2880x240", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2956,
-		   3204, 3432, 0, 240, 244, 247, 262, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 14 - 1440x480@60Hz */
-	{ DRM_MODE("1440x480", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1472,
-		   1596, 1716, 0, 480, 489, 495, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 15 - 1440x480@60Hz */
-	{ DRM_MODE("1440x480", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1472,
-		   1596, 1716, 0, 480, 489, 495, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 16 - 1920x1080@60Hz */
-	{ DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 148500, 1920, 2008,
-		   2052, 2200, 0, 1080, 1084, 1089, 1125, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 17 - 720x576@50Hz */
-	{ DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 27000, 720, 732,
-		   796, 864, 0, 576, 581, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 18 - 720x576@50Hz */
-	{ DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 27000, 720, 732,
-		   796, 864, 0, 576, 581, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 19 - 1280x720@50Hz */
-	{ DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 74250, 1280, 1720,
-		   1760, 1980, 0, 720, 725, 730, 750, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 20 - 1920x1080i@50Hz */
-	{ DRM_MODE("1920x1080i", DRM_MODE_TYPE_DRIVER, 74250, 1920, 2448,
-		   2492, 2640, 0, 1080, 1084, 1094, 1125, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC |
-			DRM_MODE_FLAG_INTERLACE) },
-	/* 21 - 1440x576i@50Hz */
-	{ DRM_MODE("1440x576i", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1464,
-		   1590, 1728, 0, 576, 580, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
-	/* 22 - 1440x576i@50Hz */
-	{ DRM_MODE("1440x576i", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1464,
-		   1590, 1728, 0, 576, 580, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
-	/* 23 - 1440x288@50Hz */
-	{ DRM_MODE("1440x288", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1464,
-		   1590, 1728, 0, 288, 290, 293, 312, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_DBLCLK) },
-	/* 24 - 1440x288@50Hz */
-	{ DRM_MODE("1440x288", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1464,
-		   1590, 1728, 0, 288, 290, 293, 312, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_DBLCLK) },
-	/* 25 - 2880x576i@50Hz */
-	{ DRM_MODE("2880x576i", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2928,
-		   3180, 3456, 0, 576, 580, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE) },
-	/* 26 - 2880x576i@50Hz */
-	{ DRM_MODE("2880x576i", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2928,
-		   3180, 3456, 0, 576, 580, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE) },
-	/* 27 - 2880x288@50Hz */
-	{ DRM_MODE("2880x288", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2928,
-		   3180, 3456, 0, 288, 290, 293, 312, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 28 - 2880x288@50Hz */
-	{ DRM_MODE("2880x288", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2928,
-		   3180, 3456, 0, 288, 290, 293, 312, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 29 - 1440x576@50Hz */
-	{ DRM_MODE("1440x576", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1464,
-		   1592, 1728, 0, 576, 581, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 30 - 1440x576@50Hz */
-	{ DRM_MODE("1440x576", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1464,
-		   1592, 1728, 0, 576, 581, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 31 - 1920x1080@50Hz */
-	{ DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 148500, 1920, 2448,
-		   2492, 2640, 0, 1080, 1084, 1089, 1125, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 32 - 1920x1080@24Hz */
-	{ DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 74250, 1920, 2558,
-		   2602, 2750, 0, 1080, 1084, 1089, 1125, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 33 - 1920x1080@25Hz */
-	{ DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 74250, 1920, 2448,
-		   2492, 2640, 0, 1080, 1084, 1089, 1125, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 34 - 1920x1080@30Hz */
-	{ DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 74250, 1920, 2008,
-		   2052, 2200, 0, 1080, 1084, 1089, 1125, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 35 - 2880x480@60Hz */
-	{ DRM_MODE("2880x480", DRM_MODE_TYPE_DRIVER, 108000, 2880, 2944,
-		   3192, 3432, 0, 480, 489, 495, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 36 - 2880x480@60Hz */
-	{ DRM_MODE("2880x480", DRM_MODE_TYPE_DRIVER, 108000, 2880, 2944,
-		   3192, 3432, 0, 480, 489, 495, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 37 - 2880x576@50Hz */
-	{ DRM_MODE("2880x576", DRM_MODE_TYPE_DRIVER, 108000, 2880, 2928,
-		   3184, 3456, 0, 576, 581, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 38 - 2880x576@50Hz */
-	{ DRM_MODE("2880x576", DRM_MODE_TYPE_DRIVER, 108000, 2880, 2928,
-		   3184, 3456, 0, 576, 581, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 39 - 1920x1080i@50Hz */
-	{ DRM_MODE("1920x1080i", DRM_MODE_TYPE_DRIVER, 72000, 1920, 1952,
-		   2120, 2304, 0, 1080, 1126, 1136, 1250, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE) },
-	/* 40 - 1920x1080i@100Hz */
-	{ DRM_MODE("1920x1080i", DRM_MODE_TYPE_DRIVER, 148500, 1920, 2448,
-		   2492, 2640, 0, 1080, 1084, 1094, 1125, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC |
-			DRM_MODE_FLAG_INTERLACE) },
-	/* 41 - 1280x720@100Hz */
-	{ DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 148500, 1280, 1720,
-		   1760, 1980, 0, 720, 725, 730, 750, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 42 - 720x576@100Hz */
-	{ DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 54000, 720, 732,
-		   796, 864, 0, 576, 581, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 43 - 720x576@100Hz */
-	{ DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 54000, 720, 732,
-		   796, 864, 0, 576, 581, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 44 - 1440x576i@100Hz */
-	{ DRM_MODE("1440x576", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1464,
-		   1590, 1728, 0, 576, 580, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_DBLCLK) },
-	/* 45 - 1440x576i@100Hz */
-	{ DRM_MODE("1440x576", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1464,
-		   1590, 1728, 0, 576, 580, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_DBLCLK) },
-	/* 46 - 1920x1080i@120Hz */
-	{ DRM_MODE("1920x1080i", DRM_MODE_TYPE_DRIVER, 148500, 1920, 2008,
-		   2052, 2200, 0, 1080, 1084, 1094, 1125, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC |
-			DRM_MODE_FLAG_INTERLACE) },
-	/* 47 - 1280x720@120Hz */
-	{ DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 148500, 1280, 1390,
-		   1430, 1650, 0, 720, 725, 730, 750, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 48 - 720x480@120Hz */
-	{ DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 54000, 720, 736,
-		   798, 858, 0, 480, 489, 495, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 49 - 720x480@120Hz */
-	{ DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 54000, 720, 736,
-		   798, 858, 0, 480, 489, 495, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 50 - 1440x480i@120Hz */
-	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1478,
-		   1602, 1716, 0, 480, 488, 494, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
-	/* 51 - 1440x480i@120Hz */
-	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1478,
-		   1602, 1716, 0, 480, 488, 494, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
-	/* 52 - 720x576@200Hz */
-	{ DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 108000, 720, 732,
-		   796, 864, 0, 576, 581, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 53 - 720x576@200Hz */
-	{ DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 108000, 720, 732,
-		   796, 864, 0, 576, 581, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 54 - 1440x576i@200Hz */
-	{ DRM_MODE("1440x576i", DRM_MODE_TYPE_DRIVER, 108000, 1440, 1464,
-		   1590, 1728, 0, 576, 580, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
-	/* 55 - 1440x576i@200Hz */
-	{ DRM_MODE("1440x576i", DRM_MODE_TYPE_DRIVER, 108000, 1440, 1464,
-		   1590, 1728, 0, 576, 580, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
-	/* 56 - 720x480@240Hz */
-	{ DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 108000, 720, 736,
-		   798, 858, 0, 480, 489, 495, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 57 - 720x480@240Hz */
-	{ DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 108000, 720, 736,
-		   798, 858, 0, 480, 489, 495, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 58 - 1440x480i@240 */
-	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 108000, 1440, 1478,
-		   1602, 1716, 0, 480, 488, 494, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
-	/* 59 - 1440x480i@240 */
-	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 108000, 1440, 1478,
-		   1602, 1716, 0, 480, 488, 494, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
-	/* 60 - 1280x720@24Hz */
-	{ DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 59400, 1280, 3040,
-		   3080, 3300, 0, 720, 725, 730, 750, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 61 - 1280x720@25Hz */
-	{ DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 74250, 1280, 3700,
-		   3740, 3960, 0, 720, 725, 730, 750, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 62 - 1280x720@30Hz */
-	{ DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 74250, 1280, 3040,
-		   3080, 3300, 0, 720, 725, 730, 750, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 63 - 1920x1080@120Hz */
-	{ DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 297000, 1920, 2008,
-		   2052, 2200, 0, 1080, 1084, 1089, 1125, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 64 - 1920x1080@100Hz */
-	{ DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 297000, 1920, 2448,
-		   2492, 2640, 0, 1080, 1084, 1094, 1125, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-};
-static const int drm_num_cea_modes = ARRAY_SIZE(edid_cea_modes);
-- 
cgit v0.10.2


From 595887eb90b7e74a3e276b8f17d35236d85b0d37 Mon Sep 17 00:00:00 2001
From: Thierry Reding <thierry.reding@avionic-design.de>
Date: Wed, 21 Nov 2012 15:00:47 +0100
Subject: drm: Add some missing forward declarations

The drm_file and drm_clip_rect structures are used throughout the file
but they are never declared nor pulled in through an include. Add
forward declarations to make them available.

Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/include/drm/drm_crtc.h b/include/drm/drm_crtc.h
index 7762718..8839b3a 100644
--- a/include/drm/drm_crtc.h
+++ b/include/drm/drm_crtc.h
@@ -38,7 +38,8 @@ struct drm_device;
 struct drm_mode_set;
 struct drm_framebuffer;
 struct drm_object_properties;
-
+struct drm_file;
+struct drm_clip_rect;
 
 #define DRM_MODE_OBJECT_CRTC 0xcccccccc
 #define DRM_MODE_OBJECT_CONNECTOR 0xc0c0c0c0
-- 
cgit v0.10.2


From f142d3bd556c5e82e9bb3d33d07d6708702ea4ce Mon Sep 17 00:00:00 2001
From: Thierry Reding <thierry.reding@avionic-design.de>
Date: Wed, 21 Nov 2012 15:29:29 +0100
Subject: video: Add generic HDMI infoframe helpers

Add generic helpers to pack HDMI infoframes into binary buffers.

Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/video/Kconfig b/drivers/video/Kconfig
index 09f1a18..b11eeab 100644
--- a/drivers/video/Kconfig
+++ b/drivers/video/Kconfig
@@ -52,6 +52,9 @@ config OF_VIDEOMODE
 	help
 	  helper to get videomodes from the devicetree
 
+config HDMI
+	bool
+
 menuconfig FB
 	tristate "Support for frame buffer devices"
 	---help---
diff --git a/drivers/video/Makefile b/drivers/video/Makefile
index f592f3b..0b50082 100644
--- a/drivers/video/Makefile
+++ b/drivers/video/Makefile
@@ -5,6 +5,7 @@
 # Each configuration option enables a list of files.
 
 obj-$(CONFIG_VGASTATE)            += vgastate.o
+obj-$(CONFIG_HDMI)                += hdmi.o
 obj-y                             += fb_notify.o
 obj-$(CONFIG_FB)                  += fb.o
 fb-y                              := fbmem.o fbmon.o fbcmap.o fbsysfs.o \
diff --git a/drivers/video/hdmi.c b/drivers/video/hdmi.c
new file mode 100644
index 0000000..ab23c9b
--- /dev/null
+++ b/drivers/video/hdmi.c
@@ -0,0 +1,308 @@
+/*
+ * Copyright (C) 2012 Avionic Design GmbH
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/bitops.h>
+#include <linux/errno.h>
+#include <linux/export.h>
+#include <linux/hdmi.h>
+#include <linux/string.h>
+
+static void hdmi_infoframe_checksum(void *buffer, size_t size)
+{
+	u8 *ptr = buffer;
+	u8 csum = 0;
+	size_t i;
+
+	/* compute checksum */
+	for (i = 0; i < size; i++)
+		csum += ptr[i];
+
+	ptr[3] = 256 - csum;
+}
+
+/**
+ * hdmi_avi_infoframe_init() - initialize an HDMI AVI infoframe
+ * @frame: HDMI AVI infoframe
+ *
+ * Returns 0 on success or a negative error code on failure.
+ */
+int hdmi_avi_infoframe_init(struct hdmi_avi_infoframe *frame)
+{
+	memset(frame, 0, sizeof(*frame));
+
+	frame->type = HDMI_INFOFRAME_TYPE_AVI;
+	frame->version = 2;
+	frame->length = 13;
+
+	return 0;
+}
+EXPORT_SYMBOL(hdmi_avi_infoframe_init);
+
+/**
+ * hdmi_avi_infoframe_pack() - write HDMI AVI infoframe to binary buffer
+ * @frame: HDMI AVI infoframe
+ * @buffer: destination buffer
+ * @size: size of buffer
+ *
+ * Packs the information contained in the @frame structure into a binary
+ * representation that can be written into the corresponding controller
+ * registers. Also computes the checksum as required by section 5.3.5 of
+ * the HDMI 1.4 specification.
+ *
+ * Returns the number of bytes packed into the binary buffer or a negative
+ * error code on failure.
+ */
+ssize_t hdmi_avi_infoframe_pack(struct hdmi_avi_infoframe *frame, void *buffer,
+				size_t size)
+{
+	u8 *ptr = buffer;
+	size_t length;
+
+	length = HDMI_INFOFRAME_HEADER_SIZE + frame->length;
+
+	if (size < length)
+		return -ENOSPC;
+
+	memset(buffer, 0, length);
+
+	ptr[0] = frame->type;
+	ptr[1] = frame->version;
+	ptr[2] = frame->length;
+	ptr[3] = 0; /* checksum */
+
+	/* start infoframe payload */
+	ptr += HDMI_INFOFRAME_HEADER_SIZE;
+
+	ptr[0] = ((frame->colorspace & 0x3) << 5) | (frame->scan_mode & 0x3);
+
+	if (frame->active_info_valid)
+		ptr[0] |= BIT(4);
+
+	if (frame->horizontal_bar_valid)
+		ptr[0] |= BIT(3);
+
+	if (frame->vertical_bar_valid)
+		ptr[0] |= BIT(2);
+
+	ptr[1] = ((frame->colorimetry & 0x3) << 6) |
+		 ((frame->picture_aspect & 0x3) << 4) |
+		 (frame->active_aspect & 0xf);
+
+	ptr[2] = ((frame->extended_colorimetry & 0x7) << 4) |
+		 ((frame->quantization_range & 0x3) << 2) |
+		 (frame->nups & 0x3);
+
+	if (frame->itc)
+		ptr[2] |= BIT(7);
+
+	ptr[3] = frame->video_code & 0x7f;
+
+	ptr[4] = ((frame->ycc_quantization_range & 0x3) << 6) |
+		 ((frame->content_type & 0x3) << 4) |
+		 (frame->pixel_repeat & 0xf);
+
+	ptr[5] = frame->top_bar & 0xff;
+	ptr[6] = (frame->top_bar >> 8) & 0xff;
+	ptr[7] = frame->bottom_bar & 0xff;
+	ptr[8] = (frame->bottom_bar >> 8) & 0xff;
+	ptr[9] = frame->left_bar & 0xff;
+	ptr[10] = (frame->left_bar >> 8) & 0xff;
+	ptr[11] = frame->right_bar & 0xff;
+	ptr[12] = (frame->right_bar >> 8) & 0xff;
+
+	hdmi_infoframe_checksum(buffer, length);
+
+	return length;
+}
+EXPORT_SYMBOL(hdmi_avi_infoframe_pack);
+
+/**
+ * hdmi_spd_infoframe_init() - initialize an HDMI SPD infoframe
+ * @frame: HDMI SPD infoframe
+ * @vendor: vendor string
+ * @product: product string
+ *
+ * Returns 0 on success or a negative error code on failure.
+ */
+int hdmi_spd_infoframe_init(struct hdmi_spd_infoframe *frame,
+			    const char *vendor, const char *product)
+{
+	memset(frame, 0, sizeof(*frame));
+
+	frame->type = HDMI_INFOFRAME_TYPE_SPD;
+	frame->version = 1;
+	frame->length = 25;
+
+	strncpy(frame->vendor, vendor, sizeof(frame->vendor));
+	strncpy(frame->product, product, sizeof(frame->product));
+
+	return 0;
+}
+EXPORT_SYMBOL(hdmi_spd_infoframe_init);
+
+/**
+ * hdmi_spd_infoframe_pack() - write HDMI SPD infoframe to binary buffer
+ * @frame: HDMI SPD infoframe
+ * @buffer: destination buffer
+ * @size: size of buffer
+ *
+ * Packs the information contained in the @frame structure into a binary
+ * representation that can be written into the corresponding controller
+ * registers. Also computes the checksum as required by section 5.3.5 of
+ * the HDMI 1.4 specification.
+ *
+ * Returns the number of bytes packed into the binary buffer or a negative
+ * error code on failure.
+ */
+ssize_t hdmi_spd_infoframe_pack(struct hdmi_spd_infoframe *frame, void *buffer,
+				size_t size)
+{
+	u8 *ptr = buffer;
+	size_t length;
+
+	length = HDMI_INFOFRAME_HEADER_SIZE + frame->length;
+
+	if (size < length)
+		return -ENOSPC;
+
+	memset(buffer, 0, length);
+
+	ptr[0] = frame->type;
+	ptr[1] = frame->version;
+	ptr[2] = frame->length;
+	ptr[3] = 0; /* checksum */
+
+	/* start infoframe payload */
+	ptr += HDMI_INFOFRAME_HEADER_SIZE;
+
+	memcpy(ptr, frame->vendor, sizeof(frame->vendor));
+	memcpy(ptr + 8, frame->product, sizeof(frame->product));
+
+	ptr[24] = frame->sdi;
+
+	hdmi_infoframe_checksum(buffer, length);
+
+	return length;
+}
+EXPORT_SYMBOL(hdmi_spd_infoframe_pack);
+
+/**
+ * hdmi_audio_infoframe_init() - initialize an HDMI audio infoframe
+ * @frame: HDMI audio infoframe
+ *
+ * Returns 0 on success or a negative error code on failure.
+ */
+int hdmi_audio_infoframe_init(struct hdmi_audio_infoframe *frame)
+{
+	memset(frame, 0, sizeof(*frame));
+
+	frame->type = HDMI_INFOFRAME_TYPE_AUDIO;
+	frame->version = 1;
+	frame->length = 10;
+
+	return 0;
+}
+EXPORT_SYMBOL(hdmi_audio_infoframe_init);
+
+/**
+ * hdmi_audio_infoframe_pack() - write HDMI audio infoframe to binary buffer
+ * @frame: HDMI audio infoframe
+ * @buffer: destination buffer
+ * @size: size of buffer
+ *
+ * Packs the information contained in the @frame structure into a binary
+ * representation that can be written into the corresponding controller
+ * registers. Also computes the checksum as required by section 5.3.5 of
+ * the HDMI 1.4 specification.
+ *
+ * Returns the number of bytes packed into the binary buffer or a negative
+ * error code on failure.
+ */
+ssize_t hdmi_audio_infoframe_pack(struct hdmi_audio_infoframe *frame,
+				  void *buffer, size_t size)
+{
+	unsigned char channels;
+	u8 *ptr = buffer;
+	size_t length;
+
+	length = HDMI_INFOFRAME_HEADER_SIZE + frame->length;
+
+	if (size < length)
+		return -ENOSPC;
+
+	memset(buffer, 0, length);
+
+	if (frame->channels >= 2)
+		channels = frame->channels - 1;
+	else
+		channels = 0;
+
+	ptr[0] = frame->type;
+	ptr[1] = frame->version;
+	ptr[2] = frame->length;
+	ptr[3] = 0; /* checksum */
+
+	/* start infoframe payload */
+	ptr += HDMI_INFOFRAME_HEADER_SIZE;
+
+	ptr[0] = ((frame->coding_type & 0xf) << 4) | (channels & 0x7);
+	ptr[1] = ((frame->sample_frequency & 0x7) << 2) |
+		 (frame->sample_size & 0x3);
+	ptr[2] = frame->coding_type_ext & 0x1f;
+	ptr[3] = frame->channel_allocation;
+	ptr[4] = (frame->level_shift_value & 0xf) << 3;
+
+	if (frame->downmix_inhibit)
+		ptr[4] |= BIT(7);
+
+	hdmi_infoframe_checksum(buffer, length);
+
+	return length;
+}
+EXPORT_SYMBOL(hdmi_audio_infoframe_pack);
+
+/**
+ * hdmi_vendor_infoframe_pack() - write a HDMI vendor infoframe to binary
+ *                                buffer
+ * @frame: HDMI vendor infoframe
+ * @buffer: destination buffer
+ * @size: size of buffer
+ *
+ * Packs the information contained in the @frame structure into a binary
+ * representation that can be written into the corresponding controller
+ * registers. Also computes the checksum as required by section 5.3.5 of
+ * the HDMI 1.4 specification.
+ *
+ * Returns the number of bytes packed into the binary buffer or a negative
+ * error code on failure.
+ */
+ssize_t hdmi_vendor_infoframe_pack(struct hdmi_vendor_infoframe *frame,
+				   void *buffer, size_t size)
+{
+	u8 *ptr = buffer;
+	size_t length;
+
+	length = HDMI_INFOFRAME_HEADER_SIZE + frame->length;
+
+	if (size < length)
+		return -ENOSPC;
+
+	memset(buffer, 0, length);
+
+	ptr[0] = frame->type;
+	ptr[1] = frame->version;
+	ptr[2] = frame->length;
+	ptr[3] = 0; /* checksum */
+
+	memcpy(&ptr[HDMI_INFOFRAME_HEADER_SIZE], frame->data, frame->length);
+
+	hdmi_infoframe_checksum(buffer, length);
+
+	return length;
+}
+EXPORT_SYMBOL(hdmi_vendor_infoframe_pack);
diff --git a/include/linux/hdmi.h b/include/linux/hdmi.h
new file mode 100644
index 0000000..3b58944
--- /dev/null
+++ b/include/linux/hdmi.h
@@ -0,0 +1,231 @@
+/*
+ * Copyright (C) 2012 Avionic Design GmbH
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#ifndef __LINUX_HDMI_H_
+#define __LINUX_HDMI_H_
+
+#include <linux/types.h>
+
+enum hdmi_infoframe_type {
+	HDMI_INFOFRAME_TYPE_VENDOR = 0x81,
+	HDMI_INFOFRAME_TYPE_AVI = 0x82,
+	HDMI_INFOFRAME_TYPE_SPD = 0x83,
+	HDMI_INFOFRAME_TYPE_AUDIO = 0x84,
+};
+
+#define HDMI_INFOFRAME_HEADER_SIZE  4
+#define HDMI_AVI_INFOFRAME_SIZE    13
+#define HDMI_SPD_INFOFRAME_SIZE    25
+#define HDMI_AUDIO_INFOFRAME_SIZE  10
+
+enum hdmi_colorspace {
+	HDMI_COLORSPACE_RGB,
+	HDMI_COLORSPACE_YUV422,
+	HDMI_COLORSPACE_YUV444,
+};
+
+enum hdmi_scan_mode {
+	HDMI_SCAN_MODE_NONE,
+	HDMI_SCAN_MODE_OVERSCAN,
+	HDMI_SCAN_MODE_UNDERSCAN,
+};
+
+enum hdmi_colorimetry {
+	HDMI_COLORIMETRY_NONE,
+	HDMI_COLORIMETRY_ITU_601,
+	HDMI_COLORIMETRY_ITU_709,
+	HDMI_COLORIMETRY_EXTENDED,
+};
+
+enum hdmi_picture_aspect {
+	HDMI_PICTURE_ASPECT_NONE,
+	HDMI_PICTURE_ASPECT_4_3,
+	HDMI_PICTURE_ASPECT_16_9,
+};
+
+enum hdmi_active_aspect {
+	HDMI_ACTIVE_ASPECT_16_9_TOP = 2,
+	HDMI_ACTIVE_ASPECT_14_9_TOP = 3,
+	HDMI_ACTIVE_ASPECT_16_9_CENTER = 4,
+	HDMI_ACTIVE_ASPECT_PICTURE = 8,
+	HDMI_ACTIVE_ASPECT_4_3 = 9,
+	HDMI_ACTIVE_ASPECT_16_9 = 10,
+	HDMI_ACTIVE_ASPECT_14_9 = 11,
+	HDMI_ACTIVE_ASPECT_4_3_SP_14_9 = 13,
+	HDMI_ACTIVE_ASPECT_16_9_SP_14_9 = 14,
+	HDMI_ACTIVE_ASPECT_16_9_SP_4_3 = 15,
+};
+
+enum hdmi_extended_colorimetry {
+	HDMI_EXTENDED_COLORIMETRY_XV_YCC_601,
+	HDMI_EXTENDED_COLORIMETRY_XV_YCC_709,
+	HDMI_EXTENDED_COLORIMETRY_S_YCC_601,
+	HDMI_EXTENDED_COLORIMETRY_ADOBE_YCC_601,
+	HDMI_EXTENDED_COLORIMETRY_ADOBE_RGB,
+};
+
+enum hdmi_quantization_range {
+	HDMI_QUANTIZATION_RANGE_DEFAULT,
+	HDMI_QUANTIZATION_RANGE_LIMITED,
+	HDMI_QUANTIZATION_RANGE_FULL,
+};
+
+/* non-uniform picture scaling */
+enum hdmi_nups {
+	HDMI_NUPS_UNKNOWN,
+	HDMI_NUPS_HORIZONTAL,
+	HDMI_NUPS_VERTICAL,
+	HDMI_NUPS_BOTH,
+};
+
+enum hdmi_ycc_quantization_range {
+	HDMI_YCC_QUANTIZATION_RANGE_LIMITED,
+	HDMI_YCC_QUANTIZATION_RANGE_FULL,
+};
+
+enum hdmi_content_type {
+	HDMI_CONTENT_TYPE_NONE,
+	HDMI_CONTENT_TYPE_PHOTO,
+	HDMI_CONTENT_TYPE_CINEMA,
+	HDMI_CONTENT_TYPE_GAME,
+};
+
+struct hdmi_avi_infoframe {
+	enum hdmi_infoframe_type type;
+	unsigned char version;
+	unsigned char length;
+	enum hdmi_colorspace colorspace;
+	bool active_info_valid;
+	bool horizontal_bar_valid;
+	bool vertical_bar_valid;
+	enum hdmi_scan_mode scan_mode;
+	enum hdmi_colorimetry colorimetry;
+	enum hdmi_picture_aspect picture_aspect;
+	enum hdmi_active_aspect active_aspect;
+	bool itc;
+	enum hdmi_extended_colorimetry extended_colorimetry;
+	enum hdmi_quantization_range quantization_range;
+	enum hdmi_nups nups;
+	unsigned char video_code;
+	enum hdmi_ycc_quantization_range ycc_quantization_range;
+	enum hdmi_content_type content_type;
+	unsigned char pixel_repeat;
+	unsigned short top_bar;
+	unsigned short bottom_bar;
+	unsigned short left_bar;
+	unsigned short right_bar;
+};
+
+int hdmi_avi_infoframe_init(struct hdmi_avi_infoframe *frame);
+ssize_t hdmi_avi_infoframe_pack(struct hdmi_avi_infoframe *frame, void *buffer,
+				size_t size);
+
+enum hdmi_spd_sdi {
+	HDMI_SPD_SDI_UNKNOWN,
+	HDMI_SPD_SDI_DSTB,
+	HDMI_SPD_SDI_DVDP,
+	HDMI_SPD_SDI_DVHS,
+	HDMI_SPD_SDI_HDDVR,
+	HDMI_SPD_SDI_DVC,
+	HDMI_SPD_SDI_DSC,
+	HDMI_SPD_SDI_VCD,
+	HDMI_SPD_SDI_GAME,
+	HDMI_SPD_SDI_PC,
+	HDMI_SPD_SDI_BD,
+	HDMI_SPD_SDI_SACD,
+	HDMI_SPD_SDI_HDDVD,
+	HDMI_SPD_SDI_PMP,
+};
+
+struct hdmi_spd_infoframe {
+	enum hdmi_infoframe_type type;
+	unsigned char version;
+	unsigned char length;
+	char vendor[8];
+	char product[16];
+	enum hdmi_spd_sdi sdi;
+};
+
+int hdmi_spd_infoframe_init(struct hdmi_spd_infoframe *frame,
+			    const char *vendor, const char *product);
+ssize_t hdmi_spd_infoframe_pack(struct hdmi_spd_infoframe *frame, void *buffer,
+				size_t size);
+
+enum hdmi_audio_coding_type {
+	HDMI_AUDIO_CODING_TYPE_STREAM,
+	HDMI_AUDIO_CODING_TYPE_PCM,
+	HDMI_AUDIO_CODING_TYPE_AC3,
+	HDMI_AUDIO_CODING_TYPE_MPEG1,
+	HDMI_AUDIO_CODING_TYPE_MP3,
+	HDMI_AUDIO_CODING_TYPE_MPEG2,
+	HDMI_AUDIO_CODING_TYPE_AAC_LC,
+	HDMI_AUDIO_CODING_TYPE_DTS,
+	HDMI_AUDIO_CODING_TYPE_ATRAC,
+	HDMI_AUDIO_CODING_TYPE_DSD,
+	HDMI_AUDIO_CODING_TYPE_EAC3,
+	HDMI_AUDIO_CODING_TYPE_DTS_HD,
+	HDMI_AUDIO_CODING_TYPE_MLP,
+	HDMI_AUDIO_CODING_TYPE_DST,
+	HDMI_AUDIO_CODING_TYPE_WMA_PRO,
+};
+
+enum hdmi_audio_sample_size {
+	HDMI_AUDIO_SAMPLE_SIZE_STREAM,
+	HDMI_AUDIO_SAMPLE_SIZE_16,
+	HDMI_AUDIO_SAMPLE_SIZE_20,
+	HDMI_AUDIO_SAMPLE_SIZE_24,
+};
+
+enum hdmi_audio_sample_frequency {
+	HDMI_AUDIO_SAMPLE_FREQUENCY_STREAM,
+	HDMI_AUDIO_SAMPLE_FREQUENCY_32000,
+	HDMI_AUDIO_SAMPLE_FREQUENCY_44100,
+	HDMI_AUDIO_SAMPLE_FREQUENCY_48000,
+	HDMI_AUDIO_SAMPLE_FREQUENCY_88200,
+	HDMI_AUDIO_SAMPLE_FREQUENCY_96000,
+	HDMI_AUDIO_SAMPLE_FREQUENCY_176400,
+	HDMI_AUDIO_SAMPLE_FREQUENCY_192000,
+};
+
+enum hdmi_audio_coding_type_ext {
+	HDMI_AUDIO_CODING_TYPE_EXT_STREAM,
+	HDMI_AUDIO_CODING_TYPE_EXT_HE_AAC,
+	HDMI_AUDIO_CODING_TYPE_EXT_HE_AAC_V2,
+	HDMI_AUDIO_CODING_TYPE_EXT_MPEG_SURROUND,
+};
+
+struct hdmi_audio_infoframe {
+	enum hdmi_infoframe_type type;
+	unsigned char version;
+	unsigned char length;
+	unsigned char channels;
+	enum hdmi_audio_coding_type coding_type;
+	enum hdmi_audio_sample_size sample_size;
+	enum hdmi_audio_sample_frequency sample_frequency;
+	enum hdmi_audio_coding_type_ext coding_type_ext;
+	unsigned char channel_allocation;
+	unsigned char level_shift_value;
+	bool downmix_inhibit;
+
+};
+
+int hdmi_audio_infoframe_init(struct hdmi_audio_infoframe *frame);
+ssize_t hdmi_audio_infoframe_pack(struct hdmi_audio_infoframe *frame,
+				  void *buffer, size_t size);
+
+struct hdmi_vendor_infoframe {
+	enum hdmi_infoframe_type type;
+	unsigned char version;
+	unsigned char length;
+	u8 data[27];
+};
+
+ssize_t hdmi_vendor_infoframe_pack(struct hdmi_vendor_infoframe *frame,
+				   void *buffer, size_t size);
+
+#endif /* _DRM_HDMI_H */
-- 
cgit v0.10.2


From 10a8512008655d5ce62f8c56323a6b5bd221c920 Mon Sep 17 00:00:00 2001
From: Thierry Reding <thierry.reding@avionic-design.de>
Date: Wed, 21 Nov 2012 15:31:35 +0100
Subject: drm: Add HDMI infoframe helpers

Add a generic helper to fill in an HDMI AVI infoframe with data
extracted from a DRM display mode.

Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
index f8dae851..1e82882 100644
--- a/drivers/gpu/drm/Kconfig
+++ b/drivers/gpu/drm/Kconfig
@@ -7,6 +7,7 @@
 menuconfig DRM
 	tristate "Direct Rendering Manager (XFree86 4.1.0 and higher DRI support)"
 	depends on (AGP || AGP=n) && !EMULATED_CMPXCHG && MMU
+	select HDMI
 	select I2C
 	select I2C_ALGOBIT
 	select DMA_SHARED_BUFFER
diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
index 6995d2f..c194f4e 100644
--- a/drivers/gpu/drm/drm_edid.c
+++ b/drivers/gpu/drm/drm_edid.c
@@ -29,6 +29,7 @@
  */
 #include <linux/kernel.h>
 #include <linux/slab.h>
+#include <linux/hdmi.h>
 #include <linux/i2c.h>
 #include <linux/module.h>
 #include <drm/drmP.h>
@@ -2856,3 +2857,35 @@ int drm_add_modes_noedid(struct drm_connector *connector,
 	return num_modes;
 }
 EXPORT_SYMBOL(drm_add_modes_noedid);
+
+/**
+ * drm_hdmi_avi_infoframe_from_display_mode() - fill an HDMI AVI infoframe with
+ *                                              data from a DRM display mode
+ * @frame: HDMI AVI infoframe
+ * @mode: DRM display mode
+ *
+ * Returns 0 on success or a negative error code on failure.
+ */
+int
+drm_hdmi_avi_infoframe_from_display_mode(struct hdmi_avi_infoframe *frame,
+					 const struct drm_display_mode *mode)
+{
+	int err;
+
+	if (!frame || !mode)
+		return -EINVAL;
+
+	err = hdmi_avi_infoframe_init(frame);
+	if (err < 0)
+		return err;
+
+	frame->video_code = drm_match_cea_mode(mode);
+	if (!frame->video_code)
+		return 0;
+
+	frame->picture_aspect = HDMI_PICTURE_ASPECT_NONE;
+	frame->active_aspect = HDMI_ACTIVE_ASPECT_PICTURE;
+
+	return 0;
+}
+EXPORT_SYMBOL(drm_hdmi_avi_infoframe_from_display_mode);
diff --git a/include/drm/drm_edid.h b/include/drm/drm_edid.h
index 0cac551..5da1b4a 100644
--- a/include/drm/drm_edid.h
+++ b/include/drm/drm_edid.h
@@ -247,6 +247,8 @@ struct edid {
 struct drm_encoder;
 struct drm_connector;
 struct drm_display_mode;
+struct hdmi_avi_infoframe;
+
 void drm_edid_to_eld(struct drm_connector *connector, struct edid *edid);
 int drm_av_sync_delay(struct drm_connector *connector,
 		      struct drm_display_mode *mode);
@@ -254,4 +256,8 @@ struct drm_connector *drm_select_eld(struct drm_encoder *encoder,
 				     struct drm_display_mode *mode);
 int drm_load_edid_firmware(struct drm_connector *connector);
 
+int
+drm_hdmi_avi_infoframe_from_display_mode(struct hdmi_avi_infoframe *frame,
+					 const struct drm_display_mode *mode);
+
 #endif /* __DRM_EDID_H__ */
-- 
cgit v0.10.2


From 5e308591a887604ed4fca7e7fcd8fb18d8bdc459 Mon Sep 17 00:00:00 2001
From: Thierry Reding <thierry.reding@avionic-design.de>
Date: Mon, 14 Jan 2013 09:00:31 +0100
Subject: drm: Add EDID helper documentation

Add a reference section about the EDID helper functions to the DRM
documentation.

Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl
index 51e1904..003c18a 100644
--- a/Documentation/DocBook/drm.tmpl
+++ b/Documentation/DocBook/drm.tmpl
@@ -2151,6 +2151,10 @@ void intel_crt_init(struct drm_device *dev)
 !Iinclude/drm/drm_dp_helper.h
 !Edrivers/gpu/drm/drm_dp_helper.c
     </sect2>
+    <sect2>
+      <title>EDID Helper Functions Reference</title>
+!Edrivers/gpu/drm/drm_edid.c
+    </sect2>
   </sect1>
 
   <!-- Internals: vertical blanking -->
-- 
cgit v0.10.2


From ac24c2204a76e5b42aa103bf963ae0eda1b827f3 Mon Sep 17 00:00:00 2001
From: Thierry Reding <thierry.reding@avionic-design.de>
Date: Fri, 23 Nov 2012 15:14:00 +0100
Subject: drm/tegra: Use generic HDMI infoframe helpers

Use the generic HDMI infoframe helpers to get rid of the NVIDIA Tegra
reimplementation.

Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/tegra/Kconfig b/drivers/gpu/drm/tegra/Kconfig
index be1daf7..c92955d 100644
--- a/drivers/gpu/drm/tegra/Kconfig
+++ b/drivers/gpu/drm/tegra/Kconfig
@@ -4,6 +4,7 @@ config DRM_TEGRA
 	select DRM_KMS_HELPER
 	select DRM_GEM_CMA_HELPER
 	select DRM_KMS_CMA_HELPER
+	select DRM_HDMI
 	select FB_CFB_FILLRECT
 	select FB_CFB_COPYAREA
 	select FB_CFB_IMAGEBLIT
diff --git a/drivers/gpu/drm/tegra/hdmi.c b/drivers/gpu/drm/tegra/hdmi.c
index e060c7e..0daee8e 100644
--- a/drivers/gpu/drm/tegra/hdmi.c
+++ b/drivers/gpu/drm/tegra/hdmi.c
@@ -10,6 +10,7 @@
 #include <linux/clk.h>
 #include <linux/debugfs.h>
 #include <linux/gpio.h>
+#include <linux/hdmi.h>
 #include <linux/module.h>
 #include <linux/of.h>
 #include <linux/platform_device.h>
@@ -17,6 +18,8 @@
 
 #include <mach/clk.h>
 
+#include <drm/drm_edid.h>
+
 #include "hdmi.h"
 #include "drm.h"
 #include "dc.h"
@@ -401,54 +404,65 @@ static int tegra_hdmi_setup_audio(struct tegra_hdmi *hdmi, unsigned int pclk)
 	return 0;
 }
 
-static void tegra_hdmi_write_infopack(struct tegra_hdmi *hdmi,
-				      unsigned int offset, u8 type,
-				      u8 version, void *data, size_t size)
+static inline unsigned long tegra_hdmi_subpack(const u8 *ptr, size_t size)
 {
-	unsigned long value;
-	u8 *ptr = data;
-	u32 subpack[2];
+	unsigned long value = 0;
 	size_t i;
-	u8 csum;
 
-	/* first byte of data is the checksum */
-	csum = type + version + size - 1;
+	for (i = size; i > 0; i--)
+		value = (value << 8) | ptr[i - 1];
 
-	for (i = 1; i < size; i++)
-		csum += ptr[i];
+	return value;
+}
 
-	ptr[0] = 0x100 - csum;
+static void tegra_hdmi_write_infopack(struct tegra_hdmi *hdmi, const void *data,
+				      size_t size)
+{
+	const u8 *ptr = data;
+	unsigned long offset;
+	unsigned long value;
+	size_t i, j;
 
-	value = INFOFRAME_HEADER_TYPE(type) |
-		INFOFRAME_HEADER_VERSION(version) |
-		INFOFRAME_HEADER_LEN(size - 1);
-	tegra_hdmi_writel(hdmi, value, offset);
+	switch (ptr[0]) {
+	case HDMI_INFOFRAME_TYPE_AVI:
+		offset = HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_HEADER;
+		break;
 
-	/* The audio inforame only has one set of subpack registers.  The hdmi
-	 * block pads the rest of the data as per the spec so we have to fixup
-	 * the length before filling in the subpacks.
-	 */
-	if (offset == HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_HEADER)
-		size = 6;
+	case HDMI_INFOFRAME_TYPE_AUDIO:
+		offset = HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_HEADER;
+		break;
 
-	/* each subpack 7 bytes devided into:
-	 *   subpack_low - bytes 0 - 3
-	 *   subpack_high - bytes 4 - 6 (with byte 7 padded to 0x00)
-	 */
-	for (i = 0; i < size; i++) {
-		size_t index = i % 7;
+	case HDMI_INFOFRAME_TYPE_VENDOR:
+		offset = HDMI_NV_PDISP_HDMI_GENERIC_HEADER;
+		break;
+
+	default:
+		dev_err(hdmi->dev, "unsupported infoframe type: %02x\n",
+			ptr[0]);
+		return;
+	}
+
+	value = INFOFRAME_HEADER_TYPE(ptr[0]) |
+		INFOFRAME_HEADER_VERSION(ptr[1]) |
+		INFOFRAME_HEADER_LEN(ptr[2]);
+	tegra_hdmi_writel(hdmi, value, offset);
+	offset++;
 
-		if (index == 0)
-			memset(subpack, 0x0, sizeof(subpack));
+	/*
+	 * Each subpack contains 7 bytes, divided into:
+	 * - subpack_low: bytes 0 - 3
+	 * - subpack_high: bytes 4 - 6 (with byte 7 padded to 0x00)
+	 */
+	for (i = 3, j = 0; i < size; i += 7, j += 8) {
+		size_t rem = size - i, num = min_t(size_t, rem, 4);
 
-		((u8 *)subpack)[index] = ptr[i];
+		value = tegra_hdmi_subpack(&ptr[i], num);
+		tegra_hdmi_writel(hdmi, value, offset++);
 
-		if (index == 6 || (i + 1 == size)) {
-			unsigned int reg = offset + 1 + (i / 7) * 2;
+		num = min_t(size_t, rem - num, 3);
 
-			tegra_hdmi_writel(hdmi, subpack[0], reg);
-			tegra_hdmi_writel(hdmi, subpack[1], reg + 1);
-		}
+		value = tegra_hdmi_subpack(&ptr[i + 4], num);
+		tegra_hdmi_writel(hdmi, value, offset++);
 	}
 }
 
@@ -456,9 +470,8 @@ static void tegra_hdmi_setup_avi_infoframe(struct tegra_hdmi *hdmi,
 					   struct drm_display_mode *mode)
 {
 	struct hdmi_avi_infoframe frame;
-	unsigned int h_front_porch;
-	unsigned int hsize = 16;
-	unsigned int vsize = 9;
+	u8 buffer[17];
+	ssize_t err;
 
 	if (hdmi->dvi) {
 		tegra_hdmi_writel(hdmi, 0,
@@ -466,69 +479,19 @@ static void tegra_hdmi_setup_avi_infoframe(struct tegra_hdmi *hdmi,
 		return;
 	}
 
-	h_front_porch = mode->hsync_start - mode->hdisplay;
-	memset(&frame, 0, sizeof(frame));
-	frame.r = HDMI_AVI_R_SAME;
-
-	switch (mode->vdisplay) {
-	case 480:
-		if (mode->hdisplay == 640) {
-			frame.m = HDMI_AVI_M_4_3;
-			frame.vic = 1;
-		} else {
-			frame.m = HDMI_AVI_M_16_9;
-			frame.vic = 3;
-		}
-		break;
-
-	case 576:
-		if (((hsize * 10) / vsize) > 14) {
-			frame.m = HDMI_AVI_M_16_9;
-			frame.vic = 18;
-		} else {
-			frame.m = HDMI_AVI_M_4_3;
-			frame.vic = 17;
-		}
-		break;
-
-	case 720:
-	case 1470: /* stereo mode */
-		frame.m = HDMI_AVI_M_16_9;
-
-		if (h_front_porch == 110)
-			frame.vic = 4;
-		else
-			frame.vic = 19;
-		break;
-
-	case 1080:
-	case 2205: /* stereo mode */
-		frame.m = HDMI_AVI_M_16_9;
-
-		switch (h_front_porch) {
-		case 88:
-			frame.vic = 16;
-			break;
-
-		case 528:
-			frame.vic = 31;
-			break;
-
-		default:
-			frame.vic = 32;
-			break;
-		}
-		break;
+	err = drm_hdmi_avi_infoframe_from_display_mode(&frame, mode);
+	if (err < 0) {
+		dev_err(hdmi->dev, "failed to setup AVI infoframe: %zd\n", err);
+		return;
+	}
 
-	default:
-		frame.m = HDMI_AVI_M_16_9;
-		frame.vic = 0;
-		break;
+	err = hdmi_avi_infoframe_pack(&frame, buffer, sizeof(buffer));
+	if (err < 0) {
+		dev_err(hdmi->dev, "failed to pack AVI infoframe: %zd\n", err);
+		return;
 	}
 
-	tegra_hdmi_write_infopack(hdmi, HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_HEADER,
-				  HDMI_INFOFRAME_TYPE_AVI, HDMI_AVI_VERSION,
-				  &frame, sizeof(frame));
+	tegra_hdmi_write_infopack(hdmi, buffer, err);
 
 	tegra_hdmi_writel(hdmi, INFOFRAME_CTRL_ENABLE,
 			  HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_CTRL);
@@ -537,6 +500,8 @@ static void tegra_hdmi_setup_avi_infoframe(struct tegra_hdmi *hdmi,
 static void tegra_hdmi_setup_audio_infoframe(struct tegra_hdmi *hdmi)
 {
 	struct hdmi_audio_infoframe frame;
+	u8 buffer[14];
+	ssize_t err;
 
 	if (hdmi->dvi) {
 		tegra_hdmi_writel(hdmi, 0,
@@ -544,14 +509,29 @@ static void tegra_hdmi_setup_audio_infoframe(struct tegra_hdmi *hdmi)
 		return;
 	}
 
-	memset(&frame, 0, sizeof(frame));
-	frame.cc = HDMI_AUDIO_CC_2;
+	err = hdmi_audio_infoframe_init(&frame);
+	if (err < 0) {
+		dev_err(hdmi->dev, "failed to initialize audio infoframe: %d\n",
+			err);
+		return;
+	}
+
+	frame.channels = 2;
+
+	err = hdmi_audio_infoframe_pack(&frame, buffer, sizeof(buffer));
+	if (err < 0) {
+		dev_err(hdmi->dev, "failed to pack audio infoframe: %zd\n",
+			err);
+		return;
+	}
 
-	tegra_hdmi_write_infopack(hdmi,
-				  HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_HEADER,
-				  HDMI_INFOFRAME_TYPE_AUDIO,
-				  HDMI_AUDIO_VERSION,
-				  &frame, sizeof(frame));
+	/*
+	 * The audio infoframe has only one set of subpack registers, so the
+	 * infoframe needs to be truncated. One set of subpack registers can
+	 * contain 7 bytes. Including the 3 byte header only the first 10
+	 * bytes can be programmed.
+	 */
+	tegra_hdmi_write_infopack(hdmi, buffer, min(10, err));
 
 	tegra_hdmi_writel(hdmi, INFOFRAME_CTRL_ENABLE,
 			  HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_CTRL);
@@ -559,8 +539,10 @@ static void tegra_hdmi_setup_audio_infoframe(struct tegra_hdmi *hdmi)
 
 static void tegra_hdmi_setup_stereo_infoframe(struct tegra_hdmi *hdmi)
 {
-	struct hdmi_stereo_infoframe frame;
+	struct hdmi_vendor_infoframe frame;
 	unsigned long value;
+	u8 buffer[10];
+	ssize_t err;
 
 	if (!hdmi->stereo) {
 		value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_GENERIC_CTRL);
@@ -570,22 +552,32 @@ static void tegra_hdmi_setup_stereo_infoframe(struct tegra_hdmi *hdmi)
 	}
 
 	memset(&frame, 0, sizeof(frame));
-	frame.regid0 = 0x03;
-	frame.regid1 = 0x0c;
-	frame.regid2 = 0x00;
-	frame.hdmi_video_format = 2;
+
+	frame.type = HDMI_INFOFRAME_TYPE_VENDOR;
+	frame.version = 0x01;
+	frame.length = 6;
+
+	frame.data[0] = 0x03; /* regid0 */
+	frame.data[1] = 0x0c; /* regid1 */
+	frame.data[2] = 0x00; /* regid2 */
+	frame.data[3] = 0x02 << 5; /* video format */
 
 	/* TODO: 74 MHz limit? */
 	if (1) {
-		frame._3d_structure = 0;
+		frame.data[4] = 0x00 << 4; /* 3D structure */
 	} else {
-		frame._3d_structure = 8;
-		frame._3d_ext_data = 0;
+		frame.data[4] = 0x08 << 4; /* 3D structure */
+		frame.data[5] = 0x00 << 4; /* 3D ext. data */
+	}
+
+	err = hdmi_vendor_infoframe_pack(&frame, buffer, sizeof(buffer));
+	if (err < 0) {
+		dev_err(hdmi->dev, "failed to pack vendor infoframe: %zd\n",
+			err);
+		return;
 	}
 
-	tegra_hdmi_write_infopack(hdmi, HDMI_NV_PDISP_HDMI_GENERIC_HEADER,
-				  HDMI_INFOFRAME_TYPE_VENDOR,
-				  HDMI_VENDOR_VERSION, &frame, 6);
+	tegra_hdmi_write_infopack(hdmi, buffer, err);
 
 	value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_GENERIC_CTRL);
 	value |= GENERIC_CTRL_ENABLE;
diff --git a/drivers/gpu/drm/tegra/hdmi.h b/drivers/gpu/drm/tegra/hdmi.h
index 1477f36..52ac36e 100644
--- a/drivers/gpu/drm/tegra/hdmi.h
+++ b/drivers/gpu/drm/tegra/hdmi.h
@@ -10,195 +10,6 @@
 #ifndef TEGRA_HDMI_H
 #define TEGRA_HDMI_H 1
 
-#define HDMI_INFOFRAME_TYPE_VENDOR   0x81
-#define HDMI_INFOFRAME_TYPE_AVI      0x82
-#define HDMI_INFOFRAME_TYPE_SPD      0x83
-#define HDMI_INFOFRAME_TYPE_AUDIO    0x84
-#define HDMI_INFOFRAME_TYPE_MPEG_SRC 0x85
-#define HDMI_INFOFRAME_TYPE_NTSC_VBI 0x86
-
-/* all fields little endian */
-struct hdmi_avi_infoframe {
-	/* PB0 */
-	u8 csum;
-
-	/* PB1 */
-	unsigned s:2; /* scan information */
-	unsigned b:2; /* bar info data valid */
-	unsigned a:1; /* active info present */
-	unsigned y:2; /* RGB or YCbCr */
-	unsigned res1:1;
-
-	/* PB2 */
-	unsigned r:4; /* active format aspect ratio */
-	unsigned m:2; /* picture aspect ratio */
-	unsigned c:2; /* colorimetry */
-
-	/* PB3 */
-	unsigned sc:2;  /* scan information */
-	unsigned q:2;   /* quantization range */
-	unsigned ec:3;  /* extended colorimetry */
-	unsigned itc:1; /* it content */
-
-	/* PB4 */
-	unsigned vic:7; /* video format id code */
-	unsigned res4:1;
-
-	/* PB5 */
-	unsigned pr:4; /* pixel repetition factor */
-	unsigned cn:2; /* it content type*/
-	unsigned yq:2; /* ycc quantization range */
-
-	/* PB6-7 */
-	u16 top_bar_end_line;
-
-	/* PB8-9 */
-	u16 bot_bar_start_line;
-
-	/* PB10-11 */
-	u16 left_bar_end_pixel;
-
-	/* PB12-13 */
-	u16 right_bar_start_pixel;
-} __packed;
-
-#define HDMI_AVI_VERSION 0x02
-
-#define HDMI_AVI_Y_RGB       0x0
-#define HDMI_AVI_Y_YCBCR_422 0x1
-#define HDMI_AVI_Y_YCBCR_444 0x2
-
-#define HDMI_AVI_B_VERT  0x1
-#define HDMI_AVI_B_HORIZ 0x2
-
-#define HDMI_AVI_S_NONE      0x0
-#define HDMI_AVI_S_OVERSCAN  0x1
-#define HDMI_AVI_S_UNDERSCAN 0x2
-
-#define HDMI_AVI_C_NONE     0x0
-#define HDMI_AVI_C_SMPTE    0x1
-#define HDMI_AVI_C_ITU_R    0x2
-#define HDMI_AVI_C_EXTENDED 0x4
-
-#define HDMI_AVI_M_4_3  0x1
-#define HDMI_AVI_M_16_9 0x2
-
-#define HDMI_AVI_R_SAME        0x8
-#define HDMI_AVI_R_4_3_CENTER  0x9
-#define HDMI_AVI_R_16_9_CENTER 0xa
-#define HDMI_AVI_R_14_9_CENTER 0xb
-
-/* all fields little endian */
-struct hdmi_audio_infoframe {
-	/* PB0 */
-	u8 csum;
-
-	/* PB1 */
-	unsigned cc:3; /* channel count */
-	unsigned res1:1;
-	unsigned ct:4; /* coding type */
-
-	/* PB2 */
-	unsigned ss:2; /* sample size */
-	unsigned sf:3; /* sample frequency */
-	unsigned res2:3;
-
-	/* PB3 */
-	unsigned cxt:5; /* coding extention type */
-	unsigned res3:3;
-
-	/* PB4 */
-	u8 ca; /* channel/speaker allocation */
-
-	/* PB5 */
-	unsigned res5:3;
-	unsigned lsv:4; /* level shift value */
-	unsigned dm_inh:1; /* downmix inhibit */
-
-	/* PB6-10 reserved */
-	u8 res6;
-	u8 res7;
-	u8 res8;
-	u8 res9;
-	u8 res10;
-} __packed;
-
-#define HDMI_AUDIO_VERSION 0x01
-
-#define HDMI_AUDIO_CC_STREAM 0x0 /* specified by audio stream */
-#define HDMI_AUDIO_CC_2      0x1
-#define HDMI_AUDIO_CC_3      0x2
-#define HDMI_AUDIO_CC_4      0x3
-#define HDMI_AUDIO_CC_5      0x4
-#define HDMI_AUDIO_CC_6      0x5
-#define HDMI_AUDIO_CC_7      0x6
-#define HDMI_AUDIO_CC_8      0x7
-
-#define HDMI_AUDIO_CT_STREAM  0x0 /* specified by audio stream */
-#define HDMI_AUDIO_CT_PCM     0x1
-#define HDMI_AUDIO_CT_AC3     0x2
-#define HDMI_AUDIO_CT_MPEG1   0x3
-#define HDMI_AUDIO_CT_MP3     0x4
-#define HDMI_AUDIO_CT_MPEG2   0x5
-#define HDMI_AUDIO_CT_AAC_LC  0x6
-#define HDMI_AUDIO_CT_DTS     0x7
-#define HDMI_AUDIO_CT_ATRAC   0x8
-#define HDMI_AUDIO_CT_DSD     0x9
-#define HDMI_AUDIO_CT_E_AC3   0xa
-#define HDMI_AUDIO_CT_DTS_HD  0xb
-#define HDMI_AUDIO_CT_MLP     0xc
-#define HDMI_AUDIO_CT_DST     0xd
-#define HDMI_AUDIO_CT_WMA_PRO 0xe
-#define HDMI_AUDIO_CT_CXT     0xf
-
-#define HDMI_AUDIO_SF_STREAM 0x0 /* specified by audio stream */
-#define HDMI_AUIDO_SF_32K    0x1
-#define HDMI_AUDIO_SF_44_1K  0x2
-#define HDMI_AUDIO_SF_48K    0x3
-#define HDMI_AUDIO_SF_88_2K  0x4
-#define HDMI_AUDIO_SF_96K    0x5
-#define HDMI_AUDIO_SF_176_4K 0x6
-#define HDMI_AUDIO_SF_192K   0x7
-
-#define HDMI_AUDIO_SS_STREAM 0x0 /* specified by audio stream */
-#define HDMI_AUDIO_SS_16BIT  0x1
-#define HDMI_AUDIO_SS_20BIT  0x2
-#define HDMI_AUDIO_SS_24BIT  0x3
-
-#define HDMI_AUDIO_CXT_CT            0x0 /* refer to coding in CT */
-#define HDMI_AUDIO_CXT_HE_AAC        0x1
-#define HDMI_AUDIO_CXT_HE_AAC_V2     0x2
-#define HDMI_AUDIO_CXT_MPEG_SURROUND 0x3
-
-/* all fields little endian */
-struct hdmi_stereo_infoframe {
-	/* PB0 */
-	u8 csum;
-
-	/* PB1 */
-	u8 regid0;
-
-	/* PB2 */
-	u8 regid1;
-
-	/* PB3 */
-	u8 regid2;
-
-	/* PB4 */
-	unsigned res1:5;
-	unsigned hdmi_video_format:3;
-
-	/* PB5 */
-	unsigned res2:4;
-	unsigned _3d_structure:4;
-
-	/* PB6*/
-	unsigned res3:4;
-	unsigned _3d_ext_data:4;
-} __packed;
-
-#define HDMI_VENDOR_VERSION 0x01
-
 /* register definitions */
 #define HDMI_CTXSW						0x00
 
-- 
cgit v0.10.2


From e3b2e0347e3b3b81cc322b413abf98a349d275df Mon Sep 17 00:00:00 2001
From: Thierry Reding <thierry.reding@avionic-design.de>
Date: Mon, 14 Jan 2013 13:36:30 +0100
Subject: drm/radeon: Use generic HDMI infoframe helpers

Use the generic HDMI infoframe helpers to get rid of the duplicate
implementation in the radeon driver.

Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

diff --git a/drivers/gpu/drm/radeon/evergreen_hdmi.c b/drivers/gpu/drm/radeon/evergreen_hdmi.c
index 327c08b..4fdecc2 100644
--- a/drivers/gpu/drm/radeon/evergreen_hdmi.c
+++ b/drivers/gpu/drm/radeon/evergreen_hdmi.c
@@ -24,6 +24,7 @@
  * Authors: Christian König
  *          Rafał Miłecki
  */
+#include <linux/hdmi.h>
 #include <drm/drmP.h>
 #include <drm/radeon_drm.h>
 #include "radeon.h"
@@ -54,79 +55,18 @@ static void evergreen_hdmi_update_ACR(struct drm_encoder *encoder, uint32_t cloc
 }
 
 /*
- * calculate the crc for a given info frame
- */
-static void evergreen_hdmi_infoframe_checksum(uint8_t packetType,
-					 uint8_t versionNumber,
-					 uint8_t length,
-					 uint8_t *frame)
-{
-	int i;
-	frame[0] = packetType + versionNumber + length;
-	for (i = 1; i <= length; i++)
-		frame[0] += frame[i];
-	frame[0] = 0x100 - frame[0];
-}
-
-/*
  * build a HDMI Video Info Frame
  */
-static void evergreen_hdmi_videoinfoframe(
-	struct drm_encoder *encoder,
-	uint8_t color_format,
-	int active_information_present,
-	uint8_t active_format_aspect_ratio,
-	uint8_t scan_information,
-	uint8_t colorimetry,
-	uint8_t ex_colorimetry,
-	uint8_t quantization,
-	int ITC,
-	uint8_t picture_aspect_ratio,
-	uint8_t video_format_identification,
-	uint8_t pixel_repetition,
-	uint8_t non_uniform_picture_scaling,
-	uint8_t bar_info_data_valid,
-	uint16_t top_bar,
-	uint16_t bottom_bar,
-	uint16_t left_bar,
-	uint16_t right_bar
-)
+static void evergreen_hdmi_update_avi_infoframe(struct drm_encoder *encoder,
+						void *buffer, size_t size)
 {
 	struct drm_device *dev = encoder->dev;
 	struct radeon_device *rdev = dev->dev_private;
 	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
 	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
 	uint32_t offset = dig->afmt->offset;
+	uint8_t *frame = buffer + 3;
 
-	uint8_t frame[14];
-
-	frame[0x0] = 0;
-	frame[0x1] =
-		(scan_information & 0x3) |
-		((bar_info_data_valid & 0x3) << 2) |
-		((active_information_present & 0x1) << 4) |
-		((color_format & 0x3) << 5);
-	frame[0x2] =
-		(active_format_aspect_ratio & 0xF) |
-		((picture_aspect_ratio & 0x3) << 4) |
-		((colorimetry & 0x3) << 6);
-	frame[0x3] =
-		(non_uniform_picture_scaling & 0x3) |
-		((quantization & 0x3) << 2) |
-		((ex_colorimetry & 0x7) << 4) |
-		((ITC & 0x1) << 7);
-	frame[0x4] = (video_format_identification & 0x7F);
-	frame[0x5] = (pixel_repetition & 0xF);
-	frame[0x6] = (top_bar & 0xFF);
-	frame[0x7] = (top_bar >> 8);
-	frame[0x8] = (bottom_bar & 0xFF);
-	frame[0x9] = (bottom_bar >> 8);
-	frame[0xA] = (left_bar & 0xFF);
-	frame[0xB] = (left_bar >> 8);
-	frame[0xC] = (right_bar & 0xFF);
-	frame[0xD] = (right_bar >> 8);
-
-	evergreen_hdmi_infoframe_checksum(0x82, 0x02, 0x0D, frame);
 	/* Our header values (type, version, length) should be alright, Intel
 	 * is using the same. Checksum function also seems to be OK, it works
 	 * fine for audio infoframe. However calculated value is always lower
@@ -154,7 +94,10 @@ void evergreen_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode
 	struct radeon_device *rdev = dev->dev_private;
 	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
 	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
+	u8 buffer[HDMI_INFOFRAME_HEADER_SIZE + HDMI_AVI_INFOFRAME_SIZE];
+	struct hdmi_avi_infoframe frame;
 	uint32_t offset;
+	ssize_t err;
 
 	/* Silent, r600_hdmi_enable will raise WARN for us */
 	if (!dig->afmt->enabled)
@@ -200,9 +143,19 @@ void evergreen_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode
 
 	WREG32(HDMI_GC + offset, 0); /* unset HDMI_GC_AVMUTE */
 
-	evergreen_hdmi_videoinfoframe(encoder, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-				      0, 0, 0, 0, 0, 0);
+	err = drm_hdmi_avi_infoframe_from_display_mode(&frame, mode);
+	if (err < 0) {
+		DRM_ERROR("failed to setup AVI infoframe: %zd\n", err);
+		return;
+	}
+
+	err = hdmi_avi_infoframe_pack(&frame, buffer, sizeof(buffer));
+	if (err < 0) {
+		DRM_ERROR("failed to pack AVI infoframe: %zd\n", err);
+		return;
+	}
 
+	evergreen_hdmi_update_avi_infoframe(encoder, buffer, sizeof(buffer));
 	evergreen_hdmi_update_ACR(encoder, mode->clock);
 
 	/* it's unknown what these bits do excatly, but it's indeed quite useful for debugging */
diff --git a/drivers/gpu/drm/radeon/r600_hdmi.c b/drivers/gpu/drm/radeon/r600_hdmi.c
index 95970ec..21ecc0e 100644
--- a/drivers/gpu/drm/radeon/r600_hdmi.c
+++ b/drivers/gpu/drm/radeon/r600_hdmi.c
@@ -23,6 +23,7 @@
  *
  * Authors: Christian König
  */
+#include <linux/hdmi.h>
 #include <drm/drmP.h>
 #include <drm/radeon_drm.h>
 #include "radeon.h"
@@ -121,79 +122,18 @@ static void r600_hdmi_update_ACR(struct drm_encoder *encoder, uint32_t clock)
 }
 
 /*
- * calculate the crc for a given info frame
- */
-static void r600_hdmi_infoframe_checksum(uint8_t packetType,
-					 uint8_t versionNumber,
-					 uint8_t length,
-					 uint8_t *frame)
-{
-	int i;
-	frame[0] = packetType + versionNumber + length;
-	for (i = 1; i <= length; i++)
-		frame[0] += frame[i];
-	frame[0] = 0x100 - frame[0];
-}
-
-/*
  * build a HDMI Video Info Frame
  */
-static void r600_hdmi_videoinfoframe(
-	struct drm_encoder *encoder,
-	enum r600_hdmi_color_format color_format,
-	int active_information_present,
-	uint8_t active_format_aspect_ratio,
-	uint8_t scan_information,
-	uint8_t colorimetry,
-	uint8_t ex_colorimetry,
-	uint8_t quantization,
-	int ITC,
-	uint8_t picture_aspect_ratio,
-	uint8_t video_format_identification,
-	uint8_t pixel_repetition,
-	uint8_t non_uniform_picture_scaling,
-	uint8_t bar_info_data_valid,
-	uint16_t top_bar,
-	uint16_t bottom_bar,
-	uint16_t left_bar,
-	uint16_t right_bar
-)
+static void r600_hdmi_update_avi_infoframe(struct drm_encoder *encoder,
+					   void *buffer, size_t size)
 {
 	struct drm_device *dev = encoder->dev;
 	struct radeon_device *rdev = dev->dev_private;
 	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
 	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
 	uint32_t offset = dig->afmt->offset;
+	uint8_t *frame = buffer + 3;
 
-	uint8_t frame[14];
-
-	frame[0x0] = 0;
-	frame[0x1] =
-		(scan_information & 0x3) |
-		((bar_info_data_valid & 0x3) << 2) |
-		((active_information_present & 0x1) << 4) |
-		((color_format & 0x3) << 5);
-	frame[0x2] =
-		(active_format_aspect_ratio & 0xF) |
-		((picture_aspect_ratio & 0x3) << 4) |
-		((colorimetry & 0x3) << 6);
-	frame[0x3] =
-		(non_uniform_picture_scaling & 0x3) |
-		((quantization & 0x3) << 2) |
-		((ex_colorimetry & 0x7) << 4) |
-		((ITC & 0x1) << 7);
-	frame[0x4] = (video_format_identification & 0x7F);
-	frame[0x5] = (pixel_repetition & 0xF);
-	frame[0x6] = (top_bar & 0xFF);
-	frame[0x7] = (top_bar >> 8);
-	frame[0x8] = (bottom_bar & 0xFF);
-	frame[0x9] = (bottom_bar >> 8);
-	frame[0xA] = (left_bar & 0xFF);
-	frame[0xB] = (left_bar >> 8);
-	frame[0xC] = (right_bar & 0xFF);
-	frame[0xD] = (right_bar >> 8);
-
-	r600_hdmi_infoframe_checksum(0x82, 0x02, 0x0D, frame);
 	/* Our header values (type, version, length) should be alright, Intel
 	 * is using the same. Checksum function also seems to be OK, it works
 	 * fine for audio infoframe. However calculated value is always lower
@@ -215,39 +155,15 @@ static void r600_hdmi_videoinfoframe(
 /*
  * build a Audio Info Frame
  */
-static void r600_hdmi_audioinfoframe(
-	struct drm_encoder *encoder,
-	uint8_t channel_count,
-	uint8_t coding_type,
-	uint8_t sample_size,
-	uint8_t sample_frequency,
-	uint8_t format,
-	uint8_t channel_allocation,
-	uint8_t level_shift,
-	int downmix_inhibit
-)
+static void r600_hdmi_update_audio_infoframe(struct drm_encoder *encoder,
+					     const void *buffer, size_t size)
 {
 	struct drm_device *dev = encoder->dev;
 	struct radeon_device *rdev = dev->dev_private;
 	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
 	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
 	uint32_t offset = dig->afmt->offset;
-
-	uint8_t frame[11];
-
-	frame[0x0] = 0;
-	frame[0x1] = (channel_count & 0x7) | ((coding_type & 0xF) << 4);
-	frame[0x2] = (sample_size & 0x3) | ((sample_frequency & 0x7) << 2);
-	frame[0x3] = format;
-	frame[0x4] = channel_allocation;
-	frame[0x5] = ((level_shift & 0xF) << 3) | ((downmix_inhibit & 0x1) << 7);
-	frame[0x6] = 0;
-	frame[0x7] = 0;
-	frame[0x8] = 0;
-	frame[0x9] = 0;
-	frame[0xA] = 0;
-
-	r600_hdmi_infoframe_checksum(0x84, 0x01, 0x0A, frame);
+	const u8 *frame = buffer + 3;
 
 	WREG32(HDMI0_AUDIO_INFO0 + offset,
 		frame[0x0] | (frame[0x1] << 8) | (frame[0x2] << 16) | (frame[0x3] << 24));
@@ -320,7 +236,10 @@ void r600_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *mod
 	struct radeon_device *rdev = dev->dev_private;
 	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
 	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
+	u8 buffer[HDMI_INFOFRAME_HEADER_SIZE + HDMI_AVI_INFOFRAME_SIZE];
+	struct hdmi_avi_infoframe frame;
 	uint32_t offset;
+	ssize_t err;
 
 	/* Silent, r600_hdmi_enable will raise WARN for us */
 	if (!dig->afmt->enabled)
@@ -371,9 +290,19 @@ void r600_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *mod
 
 	WREG32(HDMI0_GC + offset, 0); /* unset HDMI0_GC_AVMUTE */
 
-	r600_hdmi_videoinfoframe(encoder, RGB, 0, 0, 0, 0,
-		0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0);
+	err = drm_hdmi_avi_infoframe_from_display_mode(&frame, mode);
+	if (err < 0) {
+		DRM_ERROR("failed to setup AVI infoframe: %zd\n", err);
+		return;
+	}
 
+	err = hdmi_avi_infoframe_pack(&frame, buffer, sizeof(buffer));
+	if (err < 0) {
+		DRM_ERROR("failed to pack AVI infoframe: %zd\n", err);
+		return;
+	}
+
+	r600_hdmi_update_avi_infoframe(encoder, buffer, sizeof(buffer));
 	r600_hdmi_update_ACR(encoder, mode->clock);
 
 	/* it's unknown what these bits do excatly, but it's indeed quite useful for debugging */
@@ -395,8 +324,11 @@ void r600_hdmi_update_audio_settings(struct drm_encoder *encoder)
 	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
 	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
 	struct r600_audio audio = r600_audio_status(rdev);
+	uint8_t buffer[HDMI_INFOFRAME_HEADER_SIZE + HDMI_AUDIO_INFOFRAME_SIZE];
+	struct hdmi_audio_infoframe frame;
 	uint32_t offset;
 	uint32_t iec;
+	ssize_t err;
 
 	if (!dig->afmt || !dig->afmt->enabled)
 		return;
@@ -462,9 +394,21 @@ void r600_hdmi_update_audio_settings(struct drm_encoder *encoder)
 		iec |= 0x5 << 16;
 	WREG32_P(HDMI0_60958_1 + offset, iec, ~0x5000f);
 
-	r600_hdmi_audioinfoframe(encoder, audio.channels - 1, 0, 0, 0, 0, 0, 0,
-				 0);
+	err = hdmi_audio_infoframe_init(&frame);
+	if (err < 0) {
+		DRM_ERROR("failed to setup audio infoframe\n");
+		return;
+	}
+
+	frame.channels = audio.channels;
+
+	err = hdmi_audio_infoframe_pack(&frame, buffer, sizeof(buffer));
+	if (err < 0) {
+		DRM_ERROR("failed to pack audio infoframe\n");
+		return;
+	}
 
+	r600_hdmi_update_audio_infoframe(encoder, buffer, sizeof(buffer));
 	r600_hdmi_audio_workaround(encoder);
 }
 
-- 
cgit v0.10.2


From 8cf1e9811471f2910fa38dc1b28e1789080ba961 Mon Sep 17 00:00:00 2001
From: Thierry Reding <thierry.reding@avionic-design.de>
Date: Wed, 13 Feb 2013 16:08:33 +0100
Subject: drm: Add consistency check for page-flipping

Driver implementations of the drm_crtc's .page_flip() function are
required to update the crtc->fb field on success to reflect that the new
framebuffer is now in use. This is important to keep reference counting
on the framebuffers balanced.

While at it, document this requirement to keep others from falling into
the same trap.

Suggested-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl
index 51e1904..a6428dd 100644
--- a/Documentation/DocBook/drm.tmpl
+++ b/Documentation/DocBook/drm.tmpl
@@ -1161,6 +1161,12 @@ int max_width, max_height;</synopsis>
             any new rendering to the frame buffer until the page flip completes.
           </para>
           <para>
+            If a page flip can be successfully scheduled the driver must set the
+            <code>drm_crtc-&lt;fb</code> field to the new framebuffer pointed to
+            by <code>fb</code>. This is important so that the reference counting
+            on framebuffers stays balanced.
+          </para>
+          <para>
             If a page flip is already pending, the
             <methodname>page_flip</methodname> operation must return
             -<errorname>EBUSY</errorname>.
diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index 781aef5..3bdf2a6 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -3792,6 +3792,13 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,
 		/* Keep the old fb, don't unref it. */
 		old_fb = NULL;
 	} else {
+		/*
+		 * Warn if the driver hasn't properly updated the crtc->fb
+		 * field to reflect that the new framebuffer is now used.
+		 * Failing to do so will screw with the reference counting
+		 * on framebuffers.
+		 */
+		WARN_ON(crtc->fb != fb);
 		/* Unref only the old framebuffer. */
 		fb = NULL;
 	}
-- 
cgit v0.10.2


From 894752bb576a26d3343cf8c674b73899f948da68 Mon Sep 17 00:00:00 2001
From: Thierry Reding <thierry.reding@avionic-design.de>
Date: Tue, 19 Feb 2013 16:22:51 +0100
Subject: drm/tegra: Remove bogus tegra_framebuffer structure

Tegra uses the CMA FB helpers so framebuffers passed to the driver need
to use the corresponding functions to access the underlying GEM objects.

This used to work because struct tegra_framebuffer was sufficiently
similar to struct drm_fb_cma but that isn't guaranteed to stay that way.

Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>

diff --git a/drivers/gpu/drm/tegra/dc.c b/drivers/gpu/drm/tegra/dc.c
index 656b2e3..d35ff8b 100644
--- a/drivers/gpu/drm/tegra/dc.c
+++ b/drivers/gpu/drm/tegra/dc.c
@@ -158,7 +158,7 @@ static int tegra_crtc_mode_set(struct drm_crtc *crtc,
 			       struct drm_display_mode *adjusted,
 			       int x, int y, struct drm_framebuffer *old_fb)
 {
-	struct tegra_framebuffer *fb = to_tegra_fb(crtc->fb);
+	struct drm_gem_cma_object *gem = drm_fb_cma_get_gem_obj(crtc->fb, 0);
 	struct tegra_dc *dc = to_tegra_dc(crtc);
 	unsigned int h_dda, v_dda, bpp;
 	struct tegra_dc_window win;
diff --git a/drivers/gpu/drm/tegra/drm.h b/drivers/gpu/drm/tegra/drm.h
index 741b5dc..3c61aab 100644
--- a/drivers/gpu/drm/tegra/drm.h
+++ b/drivers/gpu/drm/tegra/drm.h
@@ -18,16 +18,6 @@
 #include <drm/drm_fb_cma_helper.h>
 #include <drm/drm_fixed.h>
 
-struct tegra_framebuffer {
-	struct drm_framebuffer base;
-	struct drm_gem_cma_object *obj;
-};
-
-static inline struct tegra_framebuffer *to_tegra_fb(struct drm_framebuffer *fb)
-{
-	return container_of(fb, struct tegra_framebuffer, base);
-}
-
 struct host1x {
 	struct drm_device *drm;
 	struct device *dev;
@@ -44,7 +34,6 @@ struct host1x {
 	struct list_head clients;
 
 	struct drm_fbdev_cma *fbdev;
-	struct tegra_framebuffer fb;
 };
 
 struct host1x_client;
-- 
cgit v0.10.2


From f34bc78741815d0ad07298a42101a1ee2e2bcdd2 Mon Sep 17 00:00:00 2001
From: Thierry Reding <thierry.reding@avionic-design.de>
Date: Sun, 4 Nov 2012 21:47:13 +0100
Subject: drm/tegra: Add plane support

Add support for the B and C planes which support RGB and YUV pixel
formats and can be used as overlays or hardware cursor. Currently 32-bit
XRGB as well as UYVY, YUV420 and YUV422 pixel formats are advertised.
Other formats should be easy to add but these are the most common ones
and should cover the majority of use-cases.

Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>

diff --git a/drivers/gpu/drm/tegra/dc.c b/drivers/gpu/drm/tegra/dc.c
index d35ff8b..52bf63f 100644
--- a/drivers/gpu/drm/tegra/dc.c
+++ b/drivers/gpu/drm/tegra/dc.c
@@ -18,26 +18,143 @@
 #include "drm.h"
 #include "dc.h"
 
-struct tegra_dc_window {
-	fixed20_12 x;
-	fixed20_12 y;
-	fixed20_12 w;
-	fixed20_12 h;
-	unsigned int outx;
-	unsigned int outy;
-	unsigned int outw;
-	unsigned int outh;
-	unsigned int stride;
-	unsigned int fmt;
+struct tegra_plane {
+	struct drm_plane base;
+	unsigned int index;
 };
 
+static inline struct tegra_plane *to_tegra_plane(struct drm_plane *plane)
+{
+	return container_of(plane, struct tegra_plane, base);
+}
+
+static int tegra_plane_update(struct drm_plane *plane, struct drm_crtc *crtc,
+			      struct drm_framebuffer *fb, int crtc_x,
+			      int crtc_y, unsigned int crtc_w,
+			      unsigned int crtc_h, uint32_t src_x,
+			      uint32_t src_y, uint32_t src_w, uint32_t src_h)
+{
+	struct tegra_plane *p = to_tegra_plane(plane);
+	struct tegra_dc *dc = to_tegra_dc(crtc);
+	struct tegra_dc_window window;
+	unsigned int i;
+
+	memset(&window, 0, sizeof(window));
+	window.src.x = src_x >> 16;
+	window.src.y = src_y >> 16;
+	window.src.w = src_w >> 16;
+	window.src.h = src_h >> 16;
+	window.dst.x = crtc_x;
+	window.dst.y = crtc_y;
+	window.dst.w = crtc_w;
+	window.dst.h = crtc_h;
+	window.format = tegra_dc_format(fb->pixel_format);
+	window.bits_per_pixel = fb->bits_per_pixel;
+
+	for (i = 0; i < drm_format_num_planes(fb->pixel_format); i++) {
+		struct drm_gem_cma_object *gem = drm_fb_cma_get_gem_obj(fb, i);
+
+		window.base[i] = gem->paddr + fb->offsets[i];
+
+		/*
+		 * Tegra doesn't support different strides for U and V planes
+		 * so we display a warning if the user tries to display a
+		 * framebuffer with such a configuration.
+		 */
+		if (i >= 2) {
+			if (fb->pitches[i] != window.stride[1])
+				DRM_ERROR("unsupported UV-plane configuration\n");
+		} else {
+			window.stride[i] = fb->pitches[i];
+		}
+	}
+
+	return tegra_dc_setup_window(dc, p->index, &window);
+}
+
+static int tegra_plane_disable(struct drm_plane *plane)
+{
+	struct tegra_dc *dc = to_tegra_dc(plane->crtc);
+	struct tegra_plane *p = to_tegra_plane(plane);
+	unsigned long value;
+
+	value = WINDOW_A_SELECT << p->index;
+	tegra_dc_writel(dc, value, DC_CMD_DISPLAY_WINDOW_HEADER);
+
+	value = tegra_dc_readl(dc, DC_WIN_WIN_OPTIONS);
+	value &= ~WIN_ENABLE;
+	tegra_dc_writel(dc, value, DC_WIN_WIN_OPTIONS);
+
+	tegra_dc_writel(dc, WIN_A_UPDATE << p->index, DC_CMD_STATE_CONTROL);
+	tegra_dc_writel(dc, WIN_A_ACT_REQ << p->index, DC_CMD_STATE_CONTROL);
+
+	return 0;
+}
+
+static void tegra_plane_destroy(struct drm_plane *plane)
+{
+	tegra_plane_disable(plane);
+	drm_plane_cleanup(plane);
+}
+
+static const struct drm_plane_funcs tegra_plane_funcs = {
+	.update_plane = tegra_plane_update,
+	.disable_plane = tegra_plane_disable,
+	.destroy = tegra_plane_destroy,
+};
+
+static const uint32_t plane_formats[] = {
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_UYVY,
+	DRM_FORMAT_YUV420,
+	DRM_FORMAT_YUV422,
+};
+
+static int tegra_dc_add_planes(struct drm_device *drm, struct tegra_dc *dc)
+{
+	unsigned int i;
+	int err = 0;
+
+	for (i = 0; i < 2; i++) {
+		struct tegra_plane *plane;
+
+		plane = devm_kzalloc(drm->dev, sizeof(*plane), GFP_KERNEL);
+		if (!plane)
+			return -ENOMEM;
+
+		plane->index = 1 + i;
+
+		err = drm_plane_init(drm, &plane->base, 1 << dc->pipe,
+				     &tegra_plane_funcs, plane_formats,
+				     ARRAY_SIZE(plane_formats), false);
+		if (err < 0)
+			return err;
+	}
+
+	return 0;
+}
+
 static const struct drm_crtc_funcs tegra_crtc_funcs = {
 	.set_config = drm_crtc_helper_set_config,
 	.destroy = drm_crtc_cleanup,
 };
 
-static void tegra_crtc_dpms(struct drm_crtc *crtc, int mode)
+static void tegra_crtc_disable(struct drm_crtc *crtc)
 {
+	struct drm_device *drm = crtc->dev;
+	struct drm_plane *plane;
+
+	list_for_each_entry(plane, &drm->mode_config.plane_list, head) {
+		if (plane->crtc == crtc) {
+			tegra_plane_disable(plane);
+			plane->crtc = NULL;
+
+			if (plane->fb) {
+				drm_framebuffer_unreference(plane->fb);
+				plane->fb = NULL;
+			}
+		}
+	}
 }
 
 static bool tegra_crtc_mode_fixup(struct drm_crtc *crtc,
@@ -47,10 +164,11 @@ static bool tegra_crtc_mode_fixup(struct drm_crtc *crtc,
 	return true;
 }
 
-static inline u32 compute_dda_inc(fixed20_12 inf, unsigned int out, bool v,
+static inline u32 compute_dda_inc(unsigned int in, unsigned int out, bool v,
 				  unsigned int bpp)
 {
 	fixed20_12 outf = dfixed_init(out);
+	fixed20_12 inf = dfixed_init(in);
 	u32 dda_inc;
 	int max;
 
@@ -80,9 +198,10 @@ static inline u32 compute_dda_inc(fixed20_12 inf, unsigned int out, bool v,
 	return dda_inc;
 }
 
-static inline u32 compute_initial_dda(fixed20_12 in)
+static inline u32 compute_initial_dda(unsigned int in)
 {
-	return dfixed_frac(in);
+	fixed20_12 inf = dfixed_init(in);
+	return dfixed_frac(inf);
 }
 
 static int tegra_dc_set_timings(struct tegra_dc *dc,
@@ -153,6 +272,185 @@ static int tegra_crtc_setup_clk(struct drm_crtc *crtc,
 	return 0;
 }
 
+static bool tegra_dc_format_is_yuv(unsigned int format, bool *planar)
+{
+	switch (format) {
+	case WIN_COLOR_DEPTH_YCbCr422:
+	case WIN_COLOR_DEPTH_YUV422:
+		if (planar)
+			*planar = false;
+
+		return true;
+
+	case WIN_COLOR_DEPTH_YCbCr420P:
+	case WIN_COLOR_DEPTH_YUV420P:
+	case WIN_COLOR_DEPTH_YCbCr422P:
+	case WIN_COLOR_DEPTH_YUV422P:
+	case WIN_COLOR_DEPTH_YCbCr422R:
+	case WIN_COLOR_DEPTH_YUV422R:
+	case WIN_COLOR_DEPTH_YCbCr422RA:
+	case WIN_COLOR_DEPTH_YUV422RA:
+		if (planar)
+			*planar = true;
+
+		return true;
+	}
+
+	return false;
+}
+
+int tegra_dc_setup_window(struct tegra_dc *dc, unsigned int index,
+			  const struct tegra_dc_window *window)
+{
+	unsigned h_offset, v_offset, h_size, v_size, h_dda, v_dda, bpp;
+	unsigned long value;
+	bool yuv, planar;
+
+	/*
+	 * For YUV planar modes, the number of bytes per pixel takes into
+	 * account only the luma component and therefore is 1.
+	 */
+	yuv = tegra_dc_format_is_yuv(window->format, &planar);
+	if (!yuv)
+		bpp = window->bits_per_pixel / 8;
+	else
+		bpp = planar ? 1 : 2;
+
+	value = WINDOW_A_SELECT << index;
+	tegra_dc_writel(dc, value, DC_CMD_DISPLAY_WINDOW_HEADER);
+
+	tegra_dc_writel(dc, window->format, DC_WIN_COLOR_DEPTH);
+	tegra_dc_writel(dc, 0, DC_WIN_BYTE_SWAP);
+
+	value = V_POSITION(window->dst.y) | H_POSITION(window->dst.x);
+	tegra_dc_writel(dc, value, DC_WIN_POSITION);
+
+	value = V_SIZE(window->dst.h) | H_SIZE(window->dst.w);
+	tegra_dc_writel(dc, value, DC_WIN_SIZE);
+
+	h_offset = window->src.x * bpp;
+	v_offset = window->src.y;
+	h_size = window->src.w * bpp;
+	v_size = window->src.h;
+
+	value = V_PRESCALED_SIZE(v_size) | H_PRESCALED_SIZE(h_size);
+	tegra_dc_writel(dc, value, DC_WIN_PRESCALED_SIZE);
+
+	/*
+	 * For DDA computations the number of bytes per pixel for YUV planar
+	 * modes needs to take into account all Y, U and V components.
+	 */
+	if (yuv && planar)
+		bpp = 2;
+
+	h_dda = compute_dda_inc(window->src.w, window->dst.w, false, bpp);
+	v_dda = compute_dda_inc(window->src.h, window->dst.h, true, bpp);
+
+	value = V_DDA_INC(v_dda) | H_DDA_INC(h_dda);
+	tegra_dc_writel(dc, value, DC_WIN_DDA_INC);
+
+	h_dda = compute_initial_dda(window->src.x);
+	v_dda = compute_initial_dda(window->src.y);
+
+	tegra_dc_writel(dc, h_dda, DC_WIN_H_INITIAL_DDA);
+	tegra_dc_writel(dc, v_dda, DC_WIN_V_INITIAL_DDA);
+
+	tegra_dc_writel(dc, 0, DC_WIN_UV_BUF_STRIDE);
+	tegra_dc_writel(dc, 0, DC_WIN_BUF_STRIDE);
+
+	tegra_dc_writel(dc, window->base[0], DC_WINBUF_START_ADDR);
+
+	if (yuv && planar) {
+		tegra_dc_writel(dc, window->base[1], DC_WINBUF_START_ADDR_U);
+		tegra_dc_writel(dc, window->base[2], DC_WINBUF_START_ADDR_V);
+		value = window->stride[1] << 16 | window->stride[0];
+		tegra_dc_writel(dc, value, DC_WIN_LINE_STRIDE);
+	} else {
+		tegra_dc_writel(dc, window->stride[0], DC_WIN_LINE_STRIDE);
+	}
+
+	tegra_dc_writel(dc, h_offset, DC_WINBUF_ADDR_H_OFFSET);
+	tegra_dc_writel(dc, v_offset, DC_WINBUF_ADDR_V_OFFSET);
+
+	value = WIN_ENABLE;
+
+	if (yuv) {
+		/* setup default colorspace conversion coefficients */
+		tegra_dc_writel(dc, 0x00f0, DC_WIN_CSC_YOF);
+		tegra_dc_writel(dc, 0x012a, DC_WIN_CSC_KYRGB);
+		tegra_dc_writel(dc, 0x0000, DC_WIN_CSC_KUR);
+		tegra_dc_writel(dc, 0x0198, DC_WIN_CSC_KVR);
+		tegra_dc_writel(dc, 0x039b, DC_WIN_CSC_KUG);
+		tegra_dc_writel(dc, 0x032f, DC_WIN_CSC_KVG);
+		tegra_dc_writel(dc, 0x0204, DC_WIN_CSC_KUB);
+		tegra_dc_writel(dc, 0x0000, DC_WIN_CSC_KVB);
+
+		value |= CSC_ENABLE;
+	} else if (bpp < 24) {
+		value |= COLOR_EXPAND;
+	}
+
+	tegra_dc_writel(dc, value, DC_WIN_WIN_OPTIONS);
+
+	/*
+	 * Disable blending and assume Window A is the bottom-most window,
+	 * Window C is the top-most window and Window B is in the middle.
+	 */
+	tegra_dc_writel(dc, 0xffff00, DC_WIN_BLEND_NOKEY);
+	tegra_dc_writel(dc, 0xffff00, DC_WIN_BLEND_1WIN);
+
+	switch (index) {
+	case 0:
+		tegra_dc_writel(dc, 0x000000, DC_WIN_BLEND_2WIN_X);
+		tegra_dc_writel(dc, 0x000000, DC_WIN_BLEND_2WIN_Y);
+		tegra_dc_writel(dc, 0x000000, DC_WIN_BLEND_3WIN_XY);
+		break;
+
+	case 1:
+		tegra_dc_writel(dc, 0xffff00, DC_WIN_BLEND_2WIN_X);
+		tegra_dc_writel(dc, 0x000000, DC_WIN_BLEND_2WIN_Y);
+		tegra_dc_writel(dc, 0x000000, DC_WIN_BLEND_3WIN_XY);
+		break;
+
+	case 2:
+		tegra_dc_writel(dc, 0xffff00, DC_WIN_BLEND_2WIN_X);
+		tegra_dc_writel(dc, 0xffff00, DC_WIN_BLEND_2WIN_Y);
+		tegra_dc_writel(dc, 0xffff00, DC_WIN_BLEND_3WIN_XY);
+		break;
+	}
+
+	tegra_dc_writel(dc, WIN_A_UPDATE << index, DC_CMD_STATE_CONTROL);
+	tegra_dc_writel(dc, WIN_A_ACT_REQ << index, DC_CMD_STATE_CONTROL);
+
+	return 0;
+}
+
+unsigned int tegra_dc_format(uint32_t format)
+{
+	switch (format) {
+	case DRM_FORMAT_XRGB8888:
+		return WIN_COLOR_DEPTH_B8G8R8A8;
+
+	case DRM_FORMAT_RGB565:
+		return WIN_COLOR_DEPTH_B5G6R5;
+
+	case DRM_FORMAT_UYVY:
+		return WIN_COLOR_DEPTH_YCbCr422;
+
+	case DRM_FORMAT_YUV420:
+		return WIN_COLOR_DEPTH_YCbCr420P;
+
+	case DRM_FORMAT_YUV422:
+		return WIN_COLOR_DEPTH_YCbCr422P;
+
+	default:
+		break;
+	}
+
+	WARN(1, "unsupported pixel format %u, using default\n", format);
+	return WIN_COLOR_DEPTH_B8G8R8A8;
+}
+
 static int tegra_crtc_mode_set(struct drm_crtc *crtc,
 			       struct drm_display_mode *mode,
 			       struct drm_display_mode *adjusted,
@@ -160,8 +458,7 @@ static int tegra_crtc_mode_set(struct drm_crtc *crtc,
 {
 	struct drm_gem_cma_object *gem = drm_fb_cma_get_gem_obj(crtc->fb, 0);
 	struct tegra_dc *dc = to_tegra_dc(crtc);
-	unsigned int h_dda, v_dda, bpp;
-	struct tegra_dc_window win;
+	struct tegra_dc_window window;
 	unsigned long div, value;
 	int err;
 
@@ -192,81 +489,23 @@ static int tegra_crtc_mode_set(struct drm_crtc *crtc,
 	tegra_dc_writel(dc, value, DC_DISP_DISP_CLOCK_CONTROL);
 
 	/* setup window parameters */
-	memset(&win, 0, sizeof(win));
-	win.x.full = dfixed_const(0);
-	win.y.full = dfixed_const(0);
-	win.w.full = dfixed_const(mode->hdisplay);
-	win.h.full = dfixed_const(mode->vdisplay);
-	win.outx = 0;
-	win.outy = 0;
-	win.outw = mode->hdisplay;
-	win.outh = mode->vdisplay;
-
-	switch (crtc->fb->pixel_format) {
-	case DRM_FORMAT_XRGB8888:
-		win.fmt = WIN_COLOR_DEPTH_B8G8R8A8;
-		break;
-
-	case DRM_FORMAT_RGB565:
-		win.fmt = WIN_COLOR_DEPTH_B5G6R5;
-		break;
-
-	default:
-		win.fmt = WIN_COLOR_DEPTH_B8G8R8A8;
-		WARN_ON(1);
-		break;
-	}
-
-	bpp = crtc->fb->bits_per_pixel / 8;
-	win.stride = crtc->fb->pitches[0];
-
-	/* program window registers */
-	value = WINDOW_A_SELECT;
-	tegra_dc_writel(dc, value, DC_CMD_DISPLAY_WINDOW_HEADER);
-
-	tegra_dc_writel(dc, win.fmt, DC_WIN_COLOR_DEPTH);
-	tegra_dc_writel(dc, 0, DC_WIN_BYTE_SWAP);
-
-	value = V_POSITION(win.outy) | H_POSITION(win.outx);
-	tegra_dc_writel(dc, value, DC_WIN_POSITION);
-
-	value = V_SIZE(win.outh) | H_SIZE(win.outw);
-	tegra_dc_writel(dc, value, DC_WIN_SIZE);
-
-	value = V_PRESCALED_SIZE(dfixed_trunc(win.h)) |
-		H_PRESCALED_SIZE(dfixed_trunc(win.w) * bpp);
-	tegra_dc_writel(dc, value, DC_WIN_PRESCALED_SIZE);
-
-	h_dda = compute_dda_inc(win.w, win.outw, false, bpp);
-	v_dda = compute_dda_inc(win.h, win.outh, true, bpp);
-
-	value = V_DDA_INC(v_dda) | H_DDA_INC(h_dda);
-	tegra_dc_writel(dc, value, DC_WIN_DDA_INC);
-
-	h_dda = compute_initial_dda(win.x);
-	v_dda = compute_initial_dda(win.y);
-
-	tegra_dc_writel(dc, h_dda, DC_WIN_H_INITIAL_DDA);
-	tegra_dc_writel(dc, v_dda, DC_WIN_V_INITIAL_DDA);
-
-	tegra_dc_writel(dc, 0, DC_WIN_UV_BUF_STRIDE);
-	tegra_dc_writel(dc, 0, DC_WIN_BUF_STRIDE);
-
-	tegra_dc_writel(dc, fb->obj->paddr, DC_WINBUF_START_ADDR);
-	tegra_dc_writel(dc, win.stride, DC_WIN_LINE_STRIDE);
-	tegra_dc_writel(dc, dfixed_trunc(win.x) * bpp,
-			DC_WINBUF_ADDR_H_OFFSET);
-	tegra_dc_writel(dc, dfixed_trunc(win.y), DC_WINBUF_ADDR_V_OFFSET);
-
-	value = WIN_ENABLE;
-
-	if (bpp < 24)
-		value |= COLOR_EXPAND;
-
-	tegra_dc_writel(dc, value, DC_WIN_WIN_OPTIONS);
-
-	tegra_dc_writel(dc, 0xff00, DC_WIN_BLEND_NOKEY);
-	tegra_dc_writel(dc, 0xff00, DC_WIN_BLEND_1WIN);
+	memset(&window, 0, sizeof(window));
+	window.src.x = 0;
+	window.src.y = 0;
+	window.src.w = mode->hdisplay;
+	window.src.h = mode->vdisplay;
+	window.dst.x = 0;
+	window.dst.y = 0;
+	window.dst.w = mode->hdisplay;
+	window.dst.h = mode->vdisplay;
+	window.format = tegra_dc_format(crtc->fb->pixel_format);
+	window.bits_per_pixel = crtc->fb->bits_per_pixel;
+	window.stride[0] = crtc->fb->pitches[0];
+	window.base[0] = gem->paddr;
+
+	err = tegra_dc_setup_window(dc, 0, &window);
+	if (err < 0)
+		dev_err(dc->dev, "failed to enable root plane\n");
 
 	return 0;
 }
@@ -347,7 +586,7 @@ static void tegra_crtc_load_lut(struct drm_crtc *crtc)
 }
 
 static const struct drm_crtc_helper_funcs tegra_crtc_helper_funcs = {
-	.dpms = tegra_crtc_dpms,
+	.disable = tegra_crtc_disable,
 	.mode_fixup = tegra_crtc_mode_fixup,
 	.mode_set = tegra_crtc_mode_set,
 	.prepare = tegra_crtc_prepare,
@@ -588,7 +827,7 @@ static int tegra_dc_show_regs(struct seq_file *s, void *data)
 	DUMP_REG(DC_WIN_BLEND_1WIN);
 	DUMP_REG(DC_WIN_BLEND_2WIN_X);
 	DUMP_REG(DC_WIN_BLEND_2WIN_Y);
-	DUMP_REG(DC_WIN_BLEND32WIN_XY);
+	DUMP_REG(DC_WIN_BLEND_3WIN_XY);
 	DUMP_REG(DC_WIN_HP_FETCH_CONTROL);
 	DUMP_REG(DC_WINBUF_START_ADDR);
 	DUMP_REG(DC_WINBUF_START_ADDR_NS);
@@ -690,6 +929,10 @@ static int tegra_dc_drm_init(struct host1x_client *client,
 		return err;
 	}
 
+	err = tegra_dc_add_planes(drm, dc);
+	if (err < 0)
+		return err;
+
 	if (IS_ENABLED(CONFIG_DEBUG_FS)) {
 		err = tegra_dc_debugfs_init(dc, drm->primary);
 		if (err < 0)
diff --git a/drivers/gpu/drm/tegra/dc.h b/drivers/gpu/drm/tegra/dc.h
index 99977b5..e2fa328 100644
--- a/drivers/gpu/drm/tegra/dc.h
+++ b/drivers/gpu/drm/tegra/dc.h
@@ -290,8 +290,18 @@
 #define DC_DISP_SD_HW_K_VALUES			0x4dd
 #define DC_DISP_SD_MAN_K_VALUES			0x4de
 
+#define DC_WIN_CSC_YOF				0x611
+#define DC_WIN_CSC_KYRGB			0x612
+#define DC_WIN_CSC_KUR				0x613
+#define DC_WIN_CSC_KVR				0x614
+#define DC_WIN_CSC_KUG				0x615
+#define DC_WIN_CSC_KVG				0x616
+#define DC_WIN_CSC_KUB				0x617
+#define DC_WIN_CSC_KVB				0x618
+
 #define DC_WIN_WIN_OPTIONS			0x700
 #define COLOR_EXPAND (1 <<  6)
+#define CSC_ENABLE   (1 << 18)
 #define WIN_ENABLE   (1 << 30)
 
 #define DC_WIN_BYTE_SWAP			0x701
@@ -359,7 +369,7 @@
 #define DC_WIN_BLEND_1WIN			0x710
 #define DC_WIN_BLEND_2WIN_X			0x711
 #define DC_WIN_BLEND_2WIN_Y			0x712
-#define DC_WIN_BLEND32WIN_XY			0x713
+#define DC_WIN_BLEND_3WIN_XY			0x713
 
 #define DC_WIN_HP_FETCH_CONTROL			0x714
 
diff --git a/drivers/gpu/drm/tegra/drm.h b/drivers/gpu/drm/tegra/drm.h
index 3c61aab..896ff43 100644
--- a/drivers/gpu/drm/tegra/drm.h
+++ b/drivers/gpu/drm/tegra/drm.h
@@ -107,6 +107,30 @@ static inline unsigned long tegra_dc_readl(struct tegra_dc *dc,
 	return readl(dc->regs + (reg << 2));
 }
 
+struct tegra_dc_window {
+	struct {
+		unsigned int x;
+		unsigned int y;
+		unsigned int w;
+		unsigned int h;
+	} src;
+	struct {
+		unsigned int x;
+		unsigned int y;
+		unsigned int w;
+		unsigned int h;
+	} dst;
+	unsigned int bits_per_pixel;
+	unsigned int format;
+	unsigned int stride[2];
+	unsigned long base[3];
+};
+
+/* from dc.c */
+extern unsigned int tegra_dc_format(uint32_t format);
+extern int tegra_dc_setup_window(struct tegra_dc *dc, unsigned int index,
+				 const struct tegra_dc_window *window);
+
 struct tegra_output_ops {
 	int (*enable)(struct tegra_output *output);
 	int (*disable)(struct tegra_output *output);
-- 
cgit v0.10.2


From 23fb47404e6118eb53fb34c95ed98a3f4f822a76 Mon Sep 17 00:00:00 2001
From: Thierry Reding <thierry.reding@avionic-design.de>
Date: Wed, 28 Nov 2012 11:38:24 +0100
Subject: drm/tegra: Implement .mode_set_base()

The sequence for replacing the scanout buffer is much shorter than a
full mode change operation so implementing this callback considerably
speeds up cases where only a new framebuffer is to be scanned out.

Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>

diff --git a/drivers/gpu/drm/tegra/dc.c b/drivers/gpu/drm/tegra/dc.c
index 52bf63f..ceb2f52 100644
--- a/drivers/gpu/drm/tegra/dc.c
+++ b/drivers/gpu/drm/tegra/dc.c
@@ -134,6 +134,29 @@ static int tegra_dc_add_planes(struct drm_device *drm, struct tegra_dc *dc)
 	return 0;
 }
 
+static int tegra_dc_set_base(struct tegra_dc *dc, int x, int y,
+			     struct drm_framebuffer *fb)
+{
+	struct drm_gem_cma_object *gem = drm_fb_cma_get_gem_obj(fb, 0);
+	unsigned long value;
+
+	tegra_dc_writel(dc, WINDOW_A_SELECT, DC_CMD_DISPLAY_WINDOW_HEADER);
+
+	value = fb->offsets[0] + y * fb->pitches[0] +
+		x * fb->bits_per_pixel / 8;
+
+	tegra_dc_writel(dc, gem->paddr + value, DC_WINBUF_START_ADDR);
+	tegra_dc_writel(dc, fb->pitches[0], DC_WIN_LINE_STRIDE);
+
+	value = GENERAL_UPDATE | WIN_A_UPDATE;
+	tegra_dc_writel(dc, value, DC_CMD_STATE_CONTROL);
+
+	value = GENERAL_ACT_REQ | WIN_A_ACT_REQ;
+	tegra_dc_writel(dc, value, DC_CMD_STATE_CONTROL);
+
+	return 0;
+}
+
 static const struct drm_crtc_funcs tegra_crtc_funcs = {
 	.set_config = drm_crtc_helper_set_config,
 	.destroy = drm_crtc_cleanup,
@@ -510,6 +533,14 @@ static int tegra_crtc_mode_set(struct drm_crtc *crtc,
 	return 0;
 }
 
+static int tegra_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
+				    struct drm_framebuffer *old_fb)
+{
+	struct tegra_dc *dc = to_tegra_dc(crtc);
+
+	return tegra_dc_set_base(dc, x, y, crtc->fb);
+}
+
 static void tegra_crtc_prepare(struct drm_crtc *crtc)
 {
 	struct tegra_dc *dc = to_tegra_dc(crtc);
@@ -589,6 +620,7 @@ static const struct drm_crtc_helper_funcs tegra_crtc_helper_funcs = {
 	.disable = tegra_crtc_disable,
 	.mode_fixup = tegra_crtc_mode_fixup,
 	.mode_set = tegra_crtc_mode_set,
+	.mode_set_base = tegra_crtc_mode_set_base,
 	.prepare = tegra_crtc_prepare,
 	.commit = tegra_crtc_commit,
 	.load_lut = tegra_crtc_load_lut,
-- 
cgit v0.10.2


From 6e5ff998997ba7dc5ca10b6662e95a9d07f764c4 Mon Sep 17 00:00:00 2001
From: Thierry Reding <thierry.reding@avionic-design.de>
Date: Wed, 28 Nov 2012 11:45:47 +0100
Subject: drm/tegra: Implement VBLANK support

Implement support for the VBLANK IOCTL. Note that Tegra is somewhat
special in this case because it doesn't use the generic IRQ support
provided by the DRM core (DRIVER_HAVE_IRQ) but rather registers one
interrupt handler for each display controller.

While at it, clean up the way that interrupts are enabled to ensure
that the VBLANK interrupt only gets enabled when required.

Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>

diff --git a/drivers/gpu/drm/tegra/dc.c b/drivers/gpu/drm/tegra/dc.c
index ceb2f52..5f55a25 100644
--- a/drivers/gpu/drm/tegra/dc.c
+++ b/drivers/gpu/drm/tegra/dc.c
@@ -157,6 +157,32 @@ static int tegra_dc_set_base(struct tegra_dc *dc, int x, int y,
 	return 0;
 }
 
+void tegra_dc_enable_vblank(struct tegra_dc *dc)
+{
+	unsigned long value, flags;
+
+	spin_lock_irqsave(&dc->lock, flags);
+
+	value = tegra_dc_readl(dc, DC_CMD_INT_MASK);
+	value |= VBLANK_INT;
+	tegra_dc_writel(dc, value, DC_CMD_INT_MASK);
+
+	spin_unlock_irqrestore(&dc->lock, flags);
+}
+
+void tegra_dc_disable_vblank(struct tegra_dc *dc)
+{
+	unsigned long value, flags;
+
+	spin_lock_irqsave(&dc->lock, flags);
+
+	value = tegra_dc_readl(dc, DC_CMD_INT_MASK);
+	value &= ~VBLANK_INT;
+	tegra_dc_writel(dc, value, DC_CMD_INT_MASK);
+
+	spin_unlock_irqrestore(&dc->lock, flags);
+}
+
 static const struct drm_crtc_funcs tegra_crtc_funcs = {
 	.set_config = drm_crtc_helper_set_config,
 	.destroy = drm_crtc_cleanup,
@@ -485,6 +511,8 @@ static int tegra_crtc_mode_set(struct drm_crtc *crtc,
 	unsigned long div, value;
 	int err;
 
+	drm_vblank_pre_modeset(crtc->dev, dc->pipe);
+
 	err = tegra_crtc_setup_clk(crtc, mode, &div);
 	if (err) {
 		dev_err(dc->dev, "failed to setup clock for CRTC: %d\n", err);
@@ -585,31 +613,23 @@ static void tegra_crtc_prepare(struct drm_crtc *crtc)
 	tegra_dc_writel(dc, value, DC_DISP_DISP_MEM_HIGH_PRIORITY_TIMER);
 
 	value = VBLANK_INT | WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT;
-	tegra_dc_writel(dc, value, DC_CMD_INT_MASK);
-
-	value = VBLANK_INT | WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT;
 	tegra_dc_writel(dc, value, DC_CMD_INT_ENABLE);
+
+	value = WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT;
+	tegra_dc_writel(dc, value, DC_CMD_INT_MASK);
 }
 
 static void tegra_crtc_commit(struct drm_crtc *crtc)
 {
 	struct tegra_dc *dc = to_tegra_dc(crtc);
-	unsigned long update_mask;
 	unsigned long value;
 
-	update_mask = GENERAL_ACT_REQ | WIN_A_ACT_REQ;
-
-	tegra_dc_writel(dc, update_mask << 8, DC_CMD_STATE_CONTROL);
-
-	value = tegra_dc_readl(dc, DC_CMD_INT_ENABLE);
-	value |= FRAME_END_INT;
-	tegra_dc_writel(dc, value, DC_CMD_INT_ENABLE);
+	value = GENERAL_ACT_REQ | WIN_A_ACT_REQ |
+		GENERAL_UPDATE | WIN_A_UPDATE;
 
-	value = tegra_dc_readl(dc, DC_CMD_INT_MASK);
-	value |= FRAME_END_INT;
-	tegra_dc_writel(dc, value, DC_CMD_INT_MASK);
+	tegra_dc_writel(dc, value, DC_CMD_STATE_CONTROL);
 
-	tegra_dc_writel(dc, update_mask, DC_CMD_STATE_CONTROL);
+	drm_vblank_post_modeset(crtc->dev, dc->pipe);
 }
 
 static void tegra_crtc_load_lut(struct drm_crtc *crtc)
@@ -626,7 +646,7 @@ static const struct drm_crtc_helper_funcs tegra_crtc_helper_funcs = {
 	.load_lut = tegra_crtc_load_lut,
 };
 
-static irqreturn_t tegra_drm_irq(int irq, void *data)
+static irqreturn_t tegra_dc_irq(int irq, void *data)
 {
 	struct tegra_dc *dc = data;
 	unsigned long status;
@@ -971,7 +991,7 @@ static int tegra_dc_drm_init(struct host1x_client *client,
 			dev_err(dc->dev, "debugfs setup failed: %d\n", err);
 	}
 
-	err = devm_request_irq(dc->dev, dc->irq, tegra_drm_irq, 0,
+	err = devm_request_irq(dc->dev, dc->irq, tegra_dc_irq, 0,
 			       dev_name(dc->dev), dc);
 	if (err < 0) {
 		dev_err(dc->dev, "failed to request IRQ#%u: %d\n", dc->irq,
@@ -1020,6 +1040,7 @@ static int tegra_dc_probe(struct platform_device *pdev)
 	if (!dc)
 		return -ENOMEM;
 
+	spin_lock_init(&dc->lock);
 	INIT_LIST_HEAD(&dc->list);
 	dc->dev = &pdev->dev;
 
diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c
index 3a503c9..4e31dac 100644
--- a/drivers/gpu/drm/tegra/drm.c
+++ b/drivers/gpu/drm/tegra/drm.c
@@ -40,6 +40,10 @@ static int tegra_drm_load(struct drm_device *drm, unsigned long flags)
 	if (err < 0)
 		return err;
 
+	err = drm_vblank_init(drm, drm->mode_config.num_crtc);
+	if (err < 0)
+		return err;
+
 	err = tegra_drm_fb_init(drm);
 	if (err < 0)
 		return err;
@@ -89,6 +93,48 @@ static const struct file_operations tegra_drm_fops = {
 	.llseek = noop_llseek,
 };
 
+static struct drm_crtc *tegra_crtc_from_pipe(struct drm_device *drm, int pipe)
+{
+	struct drm_crtc *crtc;
+
+	list_for_each_entry(crtc, &drm->mode_config.crtc_list, head) {
+		struct tegra_dc *dc = to_tegra_dc(crtc);
+
+		if (dc->pipe == pipe)
+			return crtc;
+	}
+
+	return NULL;
+}
+
+static u32 tegra_drm_get_vblank_counter(struct drm_device *dev, int crtc)
+{
+	/* TODO: implement real hardware counter using syncpoints */
+	return drm_vblank_count(dev, crtc);
+}
+
+static int tegra_drm_enable_vblank(struct drm_device *drm, int pipe)
+{
+	struct drm_crtc *crtc = tegra_crtc_from_pipe(drm, pipe);
+	struct tegra_dc *dc = to_tegra_dc(crtc);
+
+	if (!crtc)
+		return -ENODEV;
+
+	tegra_dc_enable_vblank(dc);
+
+	return 0;
+}
+
+static void tegra_drm_disable_vblank(struct drm_device *drm, int pipe)
+{
+	struct drm_crtc *crtc = tegra_crtc_from_pipe(drm, pipe);
+	struct tegra_dc *dc = to_tegra_dc(crtc);
+
+	if (crtc)
+		tegra_dc_disable_vblank(dc);
+}
+
 struct drm_driver tegra_drm_driver = {
 	.driver_features = DRIVER_BUS_PLATFORM | DRIVER_MODESET | DRIVER_GEM,
 	.load = tegra_drm_load,
@@ -96,6 +142,10 @@ struct drm_driver tegra_drm_driver = {
 	.open = tegra_drm_open,
 	.lastclose = tegra_drm_lastclose,
 
+	.get_vblank_counter = tegra_drm_get_vblank_counter,
+	.enable_vblank = tegra_drm_enable_vblank,
+	.disable_vblank = tegra_drm_disable_vblank,
+
 	.gem_free_object = drm_gem_cma_free_object,
 	.gem_vm_ops = &drm_gem_cma_vm_ops,
 	.dumb_create = drm_gem_cma_dumb_create,
diff --git a/drivers/gpu/drm/tegra/drm.h b/drivers/gpu/drm/tegra/drm.h
index 896ff43..01f0aee 100644
--- a/drivers/gpu/drm/tegra/drm.h
+++ b/drivers/gpu/drm/tegra/drm.h
@@ -64,6 +64,7 @@ struct tegra_output;
 
 struct tegra_dc {
 	struct host1x_client client;
+	spinlock_t lock;
 
 	struct host1x *host1x;
 	struct device *dev;
@@ -130,6 +131,8 @@ struct tegra_dc_window {
 extern unsigned int tegra_dc_format(uint32_t format);
 extern int tegra_dc_setup_window(struct tegra_dc *dc, unsigned int index,
 				 const struct tegra_dc_window *window);
+extern void tegra_dc_enable_vblank(struct tegra_dc *dc);
+extern void tegra_dc_disable_vblank(struct tegra_dc *dc);
 
 struct tegra_output_ops {
 	int (*enable)(struct tegra_output *output);
-- 
cgit v0.10.2


From 3c03c46ac80ea7edc3b8f2ba85085de54aec15cd Mon Sep 17 00:00:00 2001
From: Thierry Reding <thierry.reding@avionic-design.de>
Date: Wed, 28 Nov 2012 12:00:18 +0100
Subject: drm/tegra: Implement page-flipping support

All the necessary support bits like .mode_set_base() and VBLANK are now
available, so page-flipping case easily be implemented on top.

Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>

diff --git a/drivers/gpu/drm/tegra/dc.c b/drivers/gpu/drm/tegra/dc.c
index 5f55a25..c5d825f 100644
--- a/drivers/gpu/drm/tegra/dc.c
+++ b/drivers/gpu/drm/tegra/dc.c
@@ -183,7 +183,72 @@ void tegra_dc_disable_vblank(struct tegra_dc *dc)
 	spin_unlock_irqrestore(&dc->lock, flags);
 }
 
+static void tegra_dc_finish_page_flip(struct tegra_dc *dc)
+{
+	struct drm_device *drm = dc->base.dev;
+	struct drm_crtc *crtc = &dc->base;
+	struct drm_gem_cma_object *gem;
+	unsigned long flags, base;
+
+	if (!dc->event)
+		return;
+
+	gem = drm_fb_cma_get_gem_obj(crtc->fb, 0);
+
+	/* check if new start address has been latched */
+	tegra_dc_writel(dc, READ_MUX, DC_CMD_STATE_ACCESS);
+	base = tegra_dc_readl(dc, DC_WINBUF_START_ADDR);
+	tegra_dc_writel(dc, 0, DC_CMD_STATE_ACCESS);
+
+	if (base == gem->paddr + crtc->fb->offsets[0]) {
+		spin_lock_irqsave(&drm->event_lock, flags);
+		drm_send_vblank_event(drm, dc->pipe, dc->event);
+		drm_vblank_put(drm, dc->pipe);
+		dc->event = NULL;
+		spin_unlock_irqrestore(&drm->event_lock, flags);
+	}
+}
+
+void tegra_dc_cancel_page_flip(struct drm_crtc *crtc, struct drm_file *file)
+{
+	struct tegra_dc *dc = to_tegra_dc(crtc);
+	struct drm_device *drm = crtc->dev;
+	unsigned long flags;
+
+	spin_lock_irqsave(&drm->event_lock, flags);
+
+	if (dc->event && dc->event->base.file_priv == file) {
+		dc->event->base.destroy(&dc->event->base);
+		drm_vblank_put(drm, dc->pipe);
+		dc->event = NULL;
+	}
+
+	spin_unlock_irqrestore(&drm->event_lock, flags);
+}
+
+static int tegra_dc_page_flip(struct drm_crtc *crtc, struct drm_framebuffer *fb,
+			      struct drm_pending_vblank_event *event)
+{
+	struct tegra_dc *dc = to_tegra_dc(crtc);
+	struct drm_device *drm = crtc->dev;
+
+	if (dc->event)
+		return -EBUSY;
+
+	if (event) {
+		event->pipe = dc->pipe;
+		dc->event = event;
+		drm_vblank_get(drm, dc->pipe);
+	}
+
+	tegra_dc_set_base(dc, 0, 0, fb);
+	crtc->fb = fb;
+
+	return 0;
+}
+
 static const struct drm_crtc_funcs tegra_crtc_funcs = {
+	.page_flip = tegra_dc_page_flip,
 	.set_config = drm_crtc_helper_set_config,
 	.destroy = drm_crtc_cleanup,
 };
@@ -665,6 +730,7 @@ static irqreturn_t tegra_dc_irq(int irq, void *data)
 		dev_dbg(dc->dev, "%s(): vertical blank\n", __func__);
 		*/
 		drm_handle_vblank(dc->base.dev, dc->pipe);
+		tegra_dc_finish_page_flip(dc);
 	}
 
 	if (status & (WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT)) {
diff --git a/drivers/gpu/drm/tegra/dc.h b/drivers/gpu/drm/tegra/dc.h
index e2fa328..79eaec9 100644
--- a/drivers/gpu/drm/tegra/dc.h
+++ b/drivers/gpu/drm/tegra/dc.h
@@ -58,6 +58,8 @@
 #define DC_CMD_SIGNAL_RAISE3			0x03e
 
 #define DC_CMD_STATE_ACCESS			0x040
+#define READ_MUX  (1 << 0)
+#define WRITE_MUX (1 << 2)
 
 #define DC_CMD_STATE_CONTROL			0x041
 #define GENERAL_ACT_REQ (1 <<  0)
diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c
index 4e31dac..97485af 100644
--- a/drivers/gpu/drm/tegra/drm.c
+++ b/drivers/gpu/drm/tegra/drm.c
@@ -135,11 +135,20 @@ static void tegra_drm_disable_vblank(struct drm_device *drm, int pipe)
 		tegra_dc_disable_vblank(dc);
 }
 
+static void tegra_drm_preclose(struct drm_device *drm, struct drm_file *file)
+{
+	struct drm_crtc *crtc;
+
+	list_for_each_entry(crtc, &drm->mode_config.crtc_list, head)
+		tegra_dc_cancel_page_flip(crtc, file);
+}
+
 struct drm_driver tegra_drm_driver = {
 	.driver_features = DRIVER_BUS_PLATFORM | DRIVER_MODESET | DRIVER_GEM,
 	.load = tegra_drm_load,
 	.unload = tegra_drm_unload,
 	.open = tegra_drm_open,
+	.preclose = tegra_drm_preclose,
 	.lastclose = tegra_drm_lastclose,
 
 	.get_vblank_counter = tegra_drm_get_vblank_counter,
diff --git a/drivers/gpu/drm/tegra/drm.h b/drivers/gpu/drm/tegra/drm.h
index 01f0aee..6dd75a2 100644
--- a/drivers/gpu/drm/tegra/drm.h
+++ b/drivers/gpu/drm/tegra/drm.h
@@ -84,6 +84,9 @@ struct tegra_dc {
 	struct drm_info_list *debugfs_files;
 	struct drm_minor *minor;
 	struct dentry *debugfs;
+
+	/* page-flip handling */
+	struct drm_pending_vblank_event *event;
 };
 
 static inline struct tegra_dc *host1x_client_to_dc(struct host1x_client *client)
@@ -133,6 +136,8 @@ extern int tegra_dc_setup_window(struct tegra_dc *dc, unsigned int index,
 				 const struct tegra_dc_window *window);
 extern void tegra_dc_enable_vblank(struct tegra_dc *dc);
 extern void tegra_dc_disable_vblank(struct tegra_dc *dc);
+extern void tegra_dc_cancel_page_flip(struct drm_crtc *crtc,
+				      struct drm_file *file);
 
 struct tegra_output_ops {
 	int (*enable)(struct tegra_output *output);
-- 
cgit v0.10.2


From 3b9e71eae5ac5ded28ad1dcded7c01ae73dd9581 Mon Sep 17 00:00:00 2001
From: Thierry Reding <thierry.reding@avionic-design.de>
Date: Tue, 15 Jan 2013 12:21:36 +0100
Subject: drm/tegra: Split DC_CMD_STATE_CONTROL register write

The Tegra TRM says that the ACT_REQ and UPDATE fields cannot be
programmed at the same time so they are updated in two consecutive
writes instead.

Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>

diff --git a/drivers/gpu/drm/tegra/dc.c b/drivers/gpu/drm/tegra/dc.c
index c5d825f..2086400 100644
--- a/drivers/gpu/drm/tegra/dc.c
+++ b/drivers/gpu/drm/tegra/dc.c
@@ -689,9 +689,10 @@ static void tegra_crtc_commit(struct drm_crtc *crtc)
 	struct tegra_dc *dc = to_tegra_dc(crtc);
 	unsigned long value;
 
-	value = GENERAL_ACT_REQ | WIN_A_ACT_REQ |
-		GENERAL_UPDATE | WIN_A_UPDATE;
+	value = GENERAL_UPDATE | WIN_A_UPDATE;
+	tegra_dc_writel(dc, value, DC_CMD_STATE_CONTROL);
 
+	value = GENERAL_ACT_REQ | WIN_A_ACT_REQ;
 	tegra_dc_writel(dc, value, DC_CMD_STATE_CONTROL);
 
 	drm_vblank_post_modeset(crtc->dev, dc->pipe);
-- 
cgit v0.10.2


From 84ff6b270893cd8b054bae371a315c905ffa3ac9 Mon Sep 17 00:00:00 2001
From: Thierry Reding <thierry.reding@avionic-design.de>
Date: Thu, 21 Feb 2013 08:11:57 +0100
Subject: drm/tegra: Fix color expansion

bpp stores the number of bytes per pixel, but color expansion needs to
be enabled for less than 24 bits per pixel.

Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>

diff --git a/drivers/gpu/drm/tegra/dc.c b/drivers/gpu/drm/tegra/dc.c
index 2086400..bf8095c 100644
--- a/drivers/gpu/drm/tegra/dc.c
+++ b/drivers/gpu/drm/tegra/dc.c
@@ -500,7 +500,7 @@ int tegra_dc_setup_window(struct tegra_dc *dc, unsigned int index,
 		tegra_dc_writel(dc, 0x0000, DC_WIN_CSC_KVB);
 
 		value |= CSC_ENABLE;
-	} else if (bpp < 24) {
+	} else if (window->bits_per_pixel < 24) {
 		value |= COLOR_EXPAND;
 	}
 
-- 
cgit v0.10.2


From e450fcc6669705ef49784080ac6dd8513df37763 Mon Sep 17 00:00:00 2001
From: Thierry Reding <thierry.reding@avionic-design.de>
Date: Wed, 13 Feb 2013 16:13:16 +0100
Subject: drm/tegra: Add list of framebuffers to debugfs

This list is most useful to inspect whether framebuffer reference
counting works as expected. The code is loosely based on the i915
implementation.

Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>

diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c
index 97485af..181a370 100644
--- a/drivers/gpu/drm/tegra/drm.c
+++ b/drivers/gpu/drm/tegra/drm.c
@@ -143,6 +143,45 @@ static void tegra_drm_preclose(struct drm_device *drm, struct drm_file *file)
 		tegra_dc_cancel_page_flip(crtc, file);
 }
 
+#ifdef CONFIG_DEBUG_FS
+static int tegra_debugfs_framebuffers(struct seq_file *s, void *data)
+{
+	struct drm_info_node *node = (struct drm_info_node *)s->private;
+	struct drm_device *drm = node->minor->dev;
+	struct drm_framebuffer *fb;
+
+	mutex_lock(&drm->mode_config.fb_lock);
+
+	list_for_each_entry(fb, &drm->mode_config.fb_list, head) {
+		seq_printf(s, "%3d: user size: %d x %d, depth %d, %d bpp, refcount %d\n",
+			   fb->base.id, fb->width, fb->height, fb->depth,
+			   fb->bits_per_pixel,
+			   atomic_read(&fb->refcount.refcount));
+	}
+
+	mutex_unlock(&drm->mode_config.fb_lock);
+
+	return 0;
+}
+
+static struct drm_info_list tegra_debugfs_list[] = {
+	{ "framebuffers", tegra_debugfs_framebuffers, 0 },
+};
+
+static int tegra_debugfs_init(struct drm_minor *minor)
+{
+	return drm_debugfs_create_files(tegra_debugfs_list,
+					ARRAY_SIZE(tegra_debugfs_list),
+					minor->debugfs_root, minor);
+}
+
+static void tegra_debugfs_cleanup(struct drm_minor *minor)
+{
+	drm_debugfs_remove_files(tegra_debugfs_list,
+				 ARRAY_SIZE(tegra_debugfs_list), minor);
+}
+#endif
+
 struct drm_driver tegra_drm_driver = {
 	.driver_features = DRIVER_BUS_PLATFORM | DRIVER_MODESET | DRIVER_GEM,
 	.load = tegra_drm_load,
@@ -155,6 +194,11 @@ struct drm_driver tegra_drm_driver = {
 	.enable_vblank = tegra_drm_enable_vblank,
 	.disable_vblank = tegra_drm_disable_vblank,
 
+#if defined(CONFIG_DEBUG_FS)
+	.debugfs_init = tegra_debugfs_init,
+	.debugfs_cleanup = tegra_debugfs_cleanup,
+#endif
+
 	.gem_free_object = drm_gem_cma_free_object,
 	.gem_vm_ops = &drm_gem_cma_vm_ops,
 	.dumb_create = drm_gem_cma_dumb_create,
-- 
cgit v0.10.2


From 202adf4b9f5957b26a1cb97267d78e0edb319c5e Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Fri, 22 Feb 2013 00:53:04 +0100
Subject: drm/i915: Revert hdmi HDP pin checks

This reverts

commit 8ec22b214d76773c9d89f4040505ce10f677ed9a
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Fri May 11 18:01:34 2012 +0100

    drm/i915/hdmi: Query the live connector status bit for G4x

and

commit b0ea7d37a8f63eeec5ae80b4a6403cfba01da02f
Author: Damien Lespiau <damien.lespiau@intel.com>
Date:   Thu Dec 13 16:09:00 2012 +0000

    drm/i915/hdmi: Read the HPD status before trying to read the EDID

They reliably cause HDMI to not be detected on some systems (like my
ivb or the bug reporters gm45). To fix up the very slow unplug issues
we might want to fire up a 2nd detect cycle a few hundred ms after
each hotplug. But for now at least make displays work again.

I somewhat suspect that this is confined to HDMI connectors, since all
the machines I have with DP+ outputs work correctly.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=52361
Cc: Damien Lespiau <damien.lespiau@intel.com>
Cc: Jani Nikula <jani.nikula@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: stable@vger.kernel.org.kernel.org # for 8ec22b21
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c
index 5a6138c..3647d03 100644
--- a/drivers/gpu/drm/i915/intel_hdmi.c
+++ b/drivers/gpu/drm/i915/intel_hdmi.c
@@ -793,28 +793,6 @@ bool intel_hdmi_mode_fixup(struct drm_encoder *encoder,
 	return true;
 }
 
-static bool g4x_hdmi_connected(struct intel_hdmi *intel_hdmi)
-{
-	struct drm_device *dev = intel_hdmi_to_dev(intel_hdmi);
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_digital_port *intel_dig_port = hdmi_to_dig_port(intel_hdmi);
-	uint32_t bit;
-
-	switch (intel_dig_port->port) {
-	case PORT_B:
-		bit = PORTB_HOTPLUG_LIVE_STATUS;
-		break;
-	case PORT_C:
-		bit = PORTC_HOTPLUG_LIVE_STATUS;
-		break;
-	default:
-		bit = 0;
-		break;
-	}
-
-	return I915_READ(PORT_HOTPLUG_STAT) & bit;
-}
-
 static enum drm_connector_status
 intel_hdmi_detect(struct drm_connector *connector, bool force)
 {
@@ -827,13 +805,6 @@ intel_hdmi_detect(struct drm_connector *connector, bool force)
 	struct edid *edid;
 	enum drm_connector_status status = connector_status_disconnected;
 
-
-	if (IS_G4X(dev) && !g4x_hdmi_connected(intel_hdmi))
-		return status;
-	else if (HAS_PCH_SPLIT(dev) &&
-		 !ibx_digital_port_connected(dev_priv, intel_dig_port))
-		 return status;
-
 	intel_hdmi->has_hdmi_sink = false;
 	intel_hdmi->has_audio = false;
 	intel_hdmi->rgb_quant_range_selectable = false;
-- 
cgit v0.10.2


From be88298b0a3f771a4802f20c5e66af74bfd1dff1 Mon Sep 17 00:00:00 2001
From: Stephen Rothwell <sfr@canb.auug.org.au>
Date: Tue, 26 Feb 2013 09:54:48 +1000
Subject: drm/tilcdc: only build on arm

[airlied: hack for now until we fix cma helpers on other OF platforms]
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Dave Airlie <airlied@linux.ie>

diff --git a/drivers/gpu/drm/tilcdc/Kconfig b/drivers/gpu/drm/tilcdc/Kconfig
index ae14fd6..d24d040 100644
--- a/drivers/gpu/drm/tilcdc/Kconfig
+++ b/drivers/gpu/drm/tilcdc/Kconfig
@@ -1,6 +1,6 @@
 config DRM_TILCDC
 	tristate "DRM Support for TI LCDC Display Controller"
-	depends on DRM && OF
+	depends on DRM && OF && ARM
 	select DRM_KMS_HELPER
 	select DRM_KMS_CMA_HELPER
 	select DRM_GEM_CMA_HELPER
-- 
cgit v0.10.2