From 86b27d8050b6b2aec31063fa9f40b16fb347afb3 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Fri, 11 Feb 2011 20:47:45 +0000
Subject: drm/i915: Ignore a hung GPU when flushing the framebuffer prior to a
 switch

If the gpu is hung, then whatever was inside the render cache is lost
and there is little point waiting for it. Or complaining if we see an
EIO or EAGAIN instead. So, if the GPU is indeed in its death throes when
we need to rewrite the registers for a new framebuffer, just ignore the
error and proceed with the update.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
 drivers/gpu/drm/i915/intel_display.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

(limited to 'drivers/gpu')

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 3b006536b3d2..dcb821737f38 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -1630,19 +1630,19 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
 		struct drm_i915_gem_object *obj = to_intel_framebuffer(old_fb)->obj;
 
 		wait_event(dev_priv->pending_flip_queue,
+			   atomic_read(&dev_priv->mm.wedged) ||
 			   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
 		 * framebuffer.
+		 *
+		 * This should only fail upon a hung GPU, in which case we
+		 * can safely continue.
 		 */
 		ret = i915_gem_object_flush_gpu(obj, false);
-		if (ret) {
-			i915_gem_object_unpin(to_intel_framebuffer(crtc->fb)->obj);
-			mutex_unlock(&dev->struct_mutex);
-			return ret;
-		}
+		(void) ret;
 	}
 
 	ret = intel_pipe_set_base_atomic(crtc, crtc->fb, x, y,
-- 
cgit v1.2.3


From a36dbec57e9a665d69cd2e1a673153ddb2d62785 Mon Sep 17 00:00:00 2001
From: Jesse Barnes <jbarnes@virtuousgeek.org>
Date: Fri, 11 Feb 2011 14:44:51 -0800
Subject: drm/i915: don't enable FDI & transcoder interrupts after all

We can enable some safely, but FDI and transcoder interrupts can occur
and block other interrupts from being detected (like port hotplug
events).  So keep them disabled by default (they can be re-enabled for
debugging display bringup, but should generally be off).

Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
 drivers/gpu/drm/i915/i915_irq.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

(limited to 'drivers/gpu')

diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 97f946dcc1aa..8a9e08bf1cf7 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -316,6 +316,8 @@ static void i915_hotplug_work_func(struct work_struct *work)
 	struct drm_mode_config *mode_config = &dev->mode_config;
 	struct intel_encoder *encoder;
 
+	DRM_DEBUG_KMS("running encoder hotplug functions\n");
+
 	list_for_each_entry(encoder, &mode_config->encoder_list, base.head)
 		if (encoder->hot_plug)
 			encoder->hot_plug(encoder);
@@ -1649,9 +1651,7 @@ static int ironlake_irq_postinstall(struct drm_device *dev)
 	} else {
 		hotplug_mask = SDE_CRT_HOTPLUG | SDE_PORTB_HOTPLUG |
 			       SDE_PORTC_HOTPLUG | SDE_PORTD_HOTPLUG;
-		hotplug_mask |= SDE_AUX_MASK | SDE_FDI_MASK | SDE_TRANS_MASK;
-		I915_WRITE(FDI_RXA_IMR, 0);
-		I915_WRITE(FDI_RXB_IMR, 0);
+		hotplug_mask |= SDE_AUX_MASK;
 	}
 
 	dev_priv->pch_irq_mask = ~hotplug_mask;
-- 
cgit v1.2.3


From 011b9910bdaf2e52c48c012490ab444fceea1959 Mon Sep 17 00:00:00 2001
From: Jesse Barnes <jbarnes@virtuousgeek.org>
Date: Tue, 15 Feb 2011 15:08:02 -0800
Subject: drm/i915: skip FDI & PCH enabling for DP_A

eDP on the CPU doesn't need the PCH set up at all, it can in fact cause
problems.  So avoid FDI training and PCH PLL enabling in that case.

Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Tested-by: Andy Whitcroft <apw@canonical.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
 drivers/gpu/drm/i915/intel_display.c | 83 +++++++++++++++++++++++++++++++++++-
 1 file changed, 81 insertions(+), 2 deletions(-)

(limited to 'drivers/gpu')

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index dcb821737f38..9ca1bb2554fc 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -2045,6 +2045,31 @@ static void intel_crtc_wait_for_pending_flips(struct drm_crtc *crtc)
 		   atomic_read(&obj->pending_flip) == 0);
 }
 
+static bool intel_crtc_driving_pch(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	struct drm_mode_config *mode_config = &dev->mode_config;
+	struct intel_encoder *encoder;
+
+	/*
+	 * If there's a non-PCH eDP on this crtc, it must be DP_A, and that
+	 * must be driven by its own crtc; no sharing is possible.
+	 */
+	list_for_each_entry(encoder, &mode_config->encoder_list, base.head) {
+		if (encoder->base.crtc != crtc)
+			continue;
+
+		switch (encoder->type) {
+		case INTEL_OUTPUT_EDP:
+			if (!intel_encoder_is_pch_edp(&encoder->base))
+				return false;
+			continue;
+		}
+	}
+
+	return true;
+}
+
 static void ironlake_crtc_enable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
@@ -2053,6 +2078,7 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 	int pipe = intel_crtc->pipe;
 	int plane = intel_crtc->plane;
 	u32 reg, temp;
+	bool is_pch_port = false;
 
 	if (intel_crtc->active)
 		return;
@@ -2066,7 +2092,56 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 			I915_WRITE(PCH_LVDS, temp | LVDS_PORT_EN);
 	}
 
-	ironlake_fdi_enable(crtc);
+	is_pch_port = intel_crtc_driving_pch(crtc);
+
+	if (is_pch_port)
+		ironlake_fdi_enable(crtc);
+	else {
+		/* disable CPU FDI tx and PCH FDI rx */
+		reg = FDI_TX_CTL(pipe);
+		temp = I915_READ(reg);
+		I915_WRITE(reg, temp & ~FDI_TX_ENABLE);
+		POSTING_READ(reg);
+
+		reg = FDI_RX_CTL(pipe);
+		temp = I915_READ(reg);
+		temp &= ~(0x7 << 16);
+		temp |= (I915_READ(PIPECONF(pipe)) & PIPE_BPC_MASK) << 11;
+		I915_WRITE(reg, temp & ~FDI_RX_ENABLE);
+
+		POSTING_READ(reg);
+		udelay(100);
+
+		/* Ironlake workaround, disable clock pointer after downing FDI */
+		if (HAS_PCH_IBX(dev))
+			I915_WRITE(FDI_RX_CHICKEN(pipe),
+				   I915_READ(FDI_RX_CHICKEN(pipe) &
+					     ~FDI_RX_PHASE_SYNC_POINTER_ENABLE));
+
+		/* still set train pattern 1 */
+		reg = FDI_TX_CTL(pipe);
+		temp = I915_READ(reg);
+		temp &= ~FDI_LINK_TRAIN_NONE;
+		temp |= FDI_LINK_TRAIN_PATTERN_1;
+		I915_WRITE(reg, temp);
+
+		reg = FDI_RX_CTL(pipe);
+		temp = I915_READ(reg);
+		if (HAS_PCH_CPT(dev)) {
+			temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT;
+			temp |= FDI_LINK_TRAIN_PATTERN_1_CPT;
+		} else {
+			temp &= ~FDI_LINK_TRAIN_NONE;
+			temp |= FDI_LINK_TRAIN_PATTERN_1;
+		}
+		/* BPC in FDI rx is consistent with that in PIPECONF */
+		temp &= ~(0x07 << 16);
+		temp |= (I915_READ(PIPECONF(pipe)) & PIPE_BPC_MASK) << 11;
+		I915_WRITE(reg, temp);
+
+		POSTING_READ(reg);
+		udelay(100);
+	}
 
 	/* Enable panel fitting for LVDS */
 	if (dev_priv->pch_pf_size &&
@@ -2100,6 +2175,10 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 		intel_flush_display_plane(dev, plane);
 	}
 
+	/* Skip the PCH stuff if possible */
+	if (!is_pch_port)
+		goto done;
+
 	/* For PCH output, training FDI link */
 	if (IS_GEN6(dev))
 		gen6_fdi_link_train(crtc);
@@ -2184,7 +2263,7 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 	I915_WRITE(reg, temp | TRANS_ENABLE);
 	if (wait_for(I915_READ(reg) & TRANS_STATE_ENABLE, 100))
 		DRM_ERROR("failed to enable transcoder %d\n", pipe);
-
+done:
 	intel_crtc_load_lut(crtc);
 	intel_update_fbc(dev);
 	intel_crtc_update_cursor(crtc, true);
-- 
cgit v1.2.3


From bd6a60afeb4c9ada3ff27f1d13db1a2b5c11d8c0 Mon Sep 17 00:00:00 2001
From: Alex Deucher <alexdeucher@gmail.com>
Date: Mon, 21 Feb 2011 01:11:59 -0500
Subject: Revert "drm/radeon/kms: switch back to min->max pll post divider
 iteration"

This reverts commit a6f9761743bf35b052180f4a8bdae4d2cc0465f6.

Remove this commit as it is no longer necessary. The relevant bugs
were fixed properly in:
drm/radeon/kms: hopefully fix pll issues for real (v3)
5b40ddf888398ce4cccbf3b9d0a18d90149ed7ff
drm/radeon/kms: add missing frac fb div flag for dce4+
9f4283f49f0a96a64c5a45fe56f0f8c942885eef

This commit also broke certain ~5 Mhz modes on old arcade monitors,
so reverting this commit fixes:
https://bugzilla.kernel.org/show_bug.cgi?id=29502

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Cc: stable@kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>
---
 drivers/gpu/drm/radeon/radeon_display.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'drivers/gpu')

diff --git a/drivers/gpu/drm/radeon/radeon_display.c b/drivers/gpu/drm/radeon/radeon_display.c
index 0e657095de7c..3e7e7f9eb781 100644
--- a/drivers/gpu/drm/radeon/radeon_display.c
+++ b/drivers/gpu/drm/radeon/radeon_display.c
@@ -971,7 +971,7 @@ void radeon_compute_pll_legacy(struct radeon_pll *pll,
 		max_fractional_feed_div = pll->max_frac_feedback_div;
 	}
 
-	for (post_div = min_post_div; post_div <= max_post_div; ++post_div) {
+	for (post_div = max_post_div; post_div >= min_post_div; --post_div) {
 		uint32_t ref_div;
 
 		if ((pll->flags & RADEON_PLL_NO_ODD_POST_DIV) && (post_div & 1))
-- 
cgit v1.2.3


From e40b6fc8373314666e7853733dc0ca4049a68b95 Mon Sep 17 00:00:00 2001
From: Dave Airlie <airlied@redhat.com>
Date: Fri, 18 Feb 2011 15:51:57 +1000
Subject: drm/radeon/kms: align height of fb allocation.

this aligns the height of the fb allocation so it doesn't trip
over the size checks later when we use this from userspace to
copy the buffer at X start.

Signed-off-by: Dave Airlie <airlied@redhat.com>
---
 drivers/gpu/drm/radeon/radeon_fb.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

(limited to 'drivers/gpu')

diff --git a/drivers/gpu/drm/radeon/radeon_fb.c b/drivers/gpu/drm/radeon/radeon_fb.c
index 66324b5bb5ba..cc44bdfec80f 100644
--- a/drivers/gpu/drm/radeon/radeon_fb.c
+++ b/drivers/gpu/drm/radeon/radeon_fb.c
@@ -113,11 +113,14 @@ static int radeonfb_create_pinned_object(struct radeon_fbdev *rfbdev,
 	u32 tiling_flags = 0;
 	int ret;
 	int aligned_size, size;
+	int height = mode_cmd->height;
 
 	/* need to align pitch with crtc limits */
 	mode_cmd->pitch = radeon_align_pitch(rdev, mode_cmd->width, mode_cmd->bpp, fb_tiled) * ((mode_cmd->bpp + 1) / 8);
 
-	size = mode_cmd->pitch * mode_cmd->height;
+	if (rdev->family >= CHIP_R600)
+		height = ALIGN(mode_cmd->height, 8);
+	size = mode_cmd->pitch * height;
 	aligned_size = ALIGN(size, PAGE_SIZE);
 	ret = radeon_gem_object_create(rdev, aligned_size, 0,
 				       RADEON_GEM_DOMAIN_VRAM,
-- 
cgit v1.2.3


From c4cc383915549cf14f027f374904e30c13653dac Mon Sep 17 00:00:00 2001
From: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Date: Mon, 21 Feb 2011 05:42:00 +0100
Subject: drm/vblank: Use abs64(diff_ns) for s64 diff_ns instead of
 abs(diff_ns)

Use of abs() wrongly wrapped diff_ns to 32 bit, which gives a 1/4000
probability of a missed vblank increment at each vblank irq reenable
if the kms driver doesn't support high precision vblank timestamping.
Not a big deal in practice, but let's make it nice.

Signed-off-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
---
 drivers/gpu/drm/drm_irq.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

(limited to 'drivers/gpu')

diff --git a/drivers/gpu/drm/drm_irq.c b/drivers/gpu/drm/drm_irq.c
index 3dadfa2a8528..6d2d4faf8678 100644
--- a/drivers/gpu/drm/drm_irq.c
+++ b/drivers/gpu/drm/drm_irq.c
@@ -164,7 +164,7 @@ static void vblank_disable_and_save(struct drm_device *dev, int crtc)
 	 * available. In that case we can't account for this and just
 	 * hope for the best.
 	 */
-	if ((vblrc > 0) && (abs(diff_ns) > 1000000))
+	if ((vblrc > 0) && (abs64(diff_ns) > 1000000))
 		atomic_inc(&dev->_vblank_count[crtc]);
 
 	/* Invalidate all timestamps while vblank irq's are off. */
@@ -1293,7 +1293,7 @@ bool drm_handle_vblank(struct drm_device *dev, int crtc)
 	 * e.g., due to spurious vblank interrupts. We need to
 	 * ignore those for accounting.
 	 */
-	if (abs(diff_ns) > DRM_REDUNDANT_VBLIRQ_THRESH_NS) {
+	if (abs64(diff_ns) > DRM_REDUNDANT_VBLIRQ_THRESH_NS) {
 		/* Store new timestamp in ringbuffer. */
 		vblanktimestamp(dev, crtc, vblcount + 1) = tvblank;
 		smp_wmb();
-- 
cgit v1.2.3


From bc21512835a72bc1eab7abd7d8a1bff0435591d7 Mon Sep 17 00:00:00 2001
From: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Date: Mon, 21 Feb 2011 05:42:01 +0100
Subject: drm/vblank: Use memory barriers optimized for atomic_t instead of
 generics.

Documentation/atomic_ops.txt tells us that there are memory
barriers optimized for atomic_inc and other atomic_t ops.

Use these instead of smp_wmb(), and also to make the required
memory barriers around vblank counter increments more explicit.

Signed-off-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
---
 drivers/gpu/drm/drm_irq.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

(limited to 'drivers/gpu')

diff --git a/drivers/gpu/drm/drm_irq.c b/drivers/gpu/drm/drm_irq.c
index 6d2d4faf8678..22f3bf5ecbd2 100644
--- a/drivers/gpu/drm/drm_irq.c
+++ b/drivers/gpu/drm/drm_irq.c
@@ -164,8 +164,10 @@ static void vblank_disable_and_save(struct drm_device *dev, int crtc)
 	 * available. In that case we can't account for this and just
 	 * hope for the best.
 	 */
-	if ((vblrc > 0) && (abs64(diff_ns) > 1000000))
+	if ((vblrc > 0) && (abs64(diff_ns) > 1000000)) {
 		atomic_inc(&dev->_vblank_count[crtc]);
+		smp_mb__after_atomic_inc();
+	}
 
 	/* Invalidate all timestamps while vblank irq's are off. */
 	clear_vblank_timestamps(dev, crtc);
@@ -858,10 +860,11 @@ static void drm_update_vblank_count(struct drm_device *dev, int crtc)
 	if (rc) {
 		tslot = atomic_read(&dev->_vblank_count[crtc]) + diff;
 		vblanktimestamp(dev, crtc, tslot) = t_vblank;
-		smp_wmb();
 	}
 
+	smp_mb__before_atomic_inc();
 	atomic_add(diff, &dev->_vblank_count[crtc]);
+	smp_mb__after_atomic_inc();
 }
 
 /**
@@ -1296,12 +1299,13 @@ bool drm_handle_vblank(struct drm_device *dev, int crtc)
 	if (abs64(diff_ns) > DRM_REDUNDANT_VBLIRQ_THRESH_NS) {
 		/* Store new timestamp in ringbuffer. */
 		vblanktimestamp(dev, crtc, vblcount + 1) = tvblank;
-		smp_wmb();
 
 		/* Increment cooked vblank count. This also atomically commits
 		 * the timestamp computed above.
 		 */
+		smp_mb__before_atomic_inc();
 		atomic_inc(&dev->_vblank_count[crtc]);
+		smp_mb__after_atomic_inc();
 	} else {
 		DRM_DEBUG("crtc %d: Redundant vblirq ignored. diff_ns = %d\n",
 			  crtc, (int) diff_ns);
-- 
cgit v1.2.3


From 9be6f8a978bdcbab46474a125aa4212516b71fe7 Mon Sep 17 00:00:00 2001
From: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Date: Mon, 21 Feb 2011 05:42:02 +0100
Subject: drm/vblank: Enable precise vblank timestamps for interlaced and
 doublescan modes.

Testing showed the current code can already handle doublescan
video modes just fine. A trivial tweak makes it work for interlaced
scanout as well.

Tested and shown to be precise on Radeon rv530, r600 and
Intel 945-GME.

Signed-off-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
---
 drivers/gpu/drm/drm_irq.c | 14 ++++++--------
 1 file changed, 6 insertions(+), 8 deletions(-)

(limited to 'drivers/gpu')

diff --git a/drivers/gpu/drm/drm_irq.c b/drivers/gpu/drm/drm_irq.c
index 22f3bf5ecbd2..53120a72a48c 100644
--- a/drivers/gpu/drm/drm_irq.c
+++ b/drivers/gpu/drm/drm_irq.c
@@ -493,6 +493,12 @@ void drm_calc_timestamping_constants(struct drm_crtc *crtc)
 	/* Dot clock in Hz: */
 	dotclock = (u64) crtc->hwmode.clock * 1000;
 
+	/* Fields of interlaced scanout modes are only halve a frame duration.
+	 * Double the dotclock to get halve the frame-/line-/pixelduration.
+	 */
+	if (crtc->hwmode.flags & DRM_MODE_FLAG_INTERLACE)
+		dotclock *= 2;
+
 	/* Valid dotclock? */
 	if (dotclock > 0) {
 		/* Convert scanline length in pixels and video dot clock to
@@ -605,14 +611,6 @@ int drm_calc_vbltimestamp_from_scanoutpos(struct drm_device *dev, int crtc,
 		return -EAGAIN;
 	}
 
-	/* Don't know yet how to handle interlaced or
-	 * double scan modes. Just no-op for now.
-	 */
-	if (mode->flags & (DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLSCAN)) {
-		DRM_DEBUG("crtc %d: Noop due to unsupported mode.\n", crtc);
-		return -ENOTSUPP;
-	}
-
 	/* Get current scanout position with system timestamp.
 	 * Repeat query up to DRM_TIMESTAMP_MAXRETRIES times
 	 * if single query takes longer than max_error nanoseconds.
-- 
cgit v1.2.3


From 40f2a2fabbeffa4d47c3d904b8c94a0adb07acce Mon Sep 17 00:00:00 2001
From: Paul Bolle <pebolle@tiscali.nl>
Date: Sat, 19 Feb 2011 22:35:55 +0100
Subject: drm: drop commented out code and preceding comment

r100_gpu_init() was dropped in 90aca4d ("drm/radeon/kms: simplify &
improve GPU reset V2") but here it was only commented out.

Signed-off-by: Paul Bolle <pebolle@tiscali.nl>
Signed-off-by: Dave Airlie <airlied@redhat.com>
---
 drivers/gpu/drm/radeon/r100.c | 2 --
 1 file changed, 2 deletions(-)

(limited to 'drivers/gpu')

diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c
index 56deae5bf02e..be817f891ec8 100644
--- a/drivers/gpu/drm/radeon/r100.c
+++ b/drivers/gpu/drm/radeon/r100.c
@@ -3801,8 +3801,6 @@ static int r100_startup(struct radeon_device *rdev)
 	r100_mc_program(rdev);
 	/* Resume clock */
 	r100_clock_startup(rdev);
-	/* Initialize GPU configuration (# pipes, ...) */
-//	r100_gpu_init(rdev);
 	/* Initialize GART (initialize after TTM so we can allocate
 	 * memory through TTM but finalize after TTM) */
 	r100_enable_bm(rdev);
-- 
cgit v1.2.3


From 45e4039c3aea597ede44a264cea322908cdedfe9 Mon Sep 17 00:00:00 2001
From: Dave Airlie <airlied@redhat.com>
Date: Sun, 20 Feb 2011 21:57:32 +0000
Subject: drm/radeon: fix regression with AA resolve checking

Some userspaces can emit a whole packet without disabling AA resolve
by the looks of it, so we have to deal with them.

Signed-off-by: Dave Airlie <airlied@redhat.com>
Tested-by: Jorg Otte <jrg.otte@googlemail.com>
---
 drivers/gpu/drm/radeon/r100.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'drivers/gpu')

diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c
index be817f891ec8..93fa735c8c1a 100644
--- a/drivers/gpu/drm/radeon/r100.c
+++ b/drivers/gpu/drm/radeon/r100.c
@@ -3490,7 +3490,7 @@ void r100_cs_track_clear(struct radeon_device *rdev, struct r100_cs_track *track
 		track->num_texture = 16;
 		track->maxy = 4096;
 		track->separate_cube = 0;
-		track->aaresolve = true;
+		track->aaresolve = false;
 		track->aa.robj = NULL;
 	}
 
-- 
cgit v1.2.3


From c2e0eb167070a6e9dcb49c84c13c79a30d672431 Mon Sep 17 00:00:00 2001
From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Tue, 22 Feb 2011 18:25:49 +0100
Subject: drm/i915: fix corruptions on i8xx due to relaxed fencing

It looks like gen2 has a peculiar interleaved 2-row inter-tile
layout. Probably inherited from i81x which had 2kb tiles (which
naturally fit an even-number-of-tile-rows scheme to fit onto 4kb
pages). There is no other mention of this in any docs (also not
in the Intel internal documention according to Chris Wilson).

Problem manifests itself in corruptions in the second half of the
last tile row (if the bo has an odd number of tiles). Which can
only happen with relaxed tiling (introduced in a00b10c360b35d6431a9).

So reject set_tiling calls that don't satisfy this constrain to
prevent broken userspace from causing havoc. While at it, also
check the size for newer chipsets.

LKML: https://lkml.org/lkml/2011/2/19/5
Reported-by: Indan Zupancic <indan@nul.nu>
Tested-by: Indan Zupancic <indan@nul.nu>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
 drivers/gpu/drm/i915/i915_gem_tiling.c | 16 +++++++++++++++-
 1 file changed, 15 insertions(+), 1 deletion(-)

(limited to 'drivers/gpu')

diff --git a/drivers/gpu/drm/i915/i915_gem_tiling.c b/drivers/gpu/drm/i915/i915_gem_tiling.c
index 22a32b9932c5..79a04fde69b5 100644
--- a/drivers/gpu/drm/i915/i915_gem_tiling.c
+++ b/drivers/gpu/drm/i915/i915_gem_tiling.c
@@ -184,7 +184,7 @@ i915_gem_detect_bit_6_swizzle(struct drm_device *dev)
 static bool
 i915_tiling_ok(struct drm_device *dev, int stride, int size, int tiling_mode)
 {
-	int tile_width;
+	int tile_width, tile_height;
 
 	/* Linear is always fine */
 	if (tiling_mode == I915_TILING_NONE)
@@ -215,6 +215,20 @@ i915_tiling_ok(struct drm_device *dev, int stride, int size, int tiling_mode)
 		}
 	}
 
+	if (IS_GEN2(dev) ||
+	    (tiling_mode == I915_TILING_Y && HAS_128_BYTE_Y_TILING(dev)))
+		tile_height = 32;
+	else
+		tile_height = 8;
+	/* i8xx is strange: It has 2 interleaved rows of tiles, so needs an even
+	 * number of tile rows. */
+	if (IS_GEN2(dev))
+		tile_height *= 2;
+
+	/* Size needs to be aligned to a full tile row */
+	if (size & (tile_height * stride - 1))
+		return false;
+
 	/* 965+ just needs multiples of tile width */
 	if (INTEL_INFO(dev)->gen >= 4) {
 		if (stride & (tile_width - 1))
-- 
cgit v1.2.3


From 995073072c2ae72255b595b192cc63f43fd386ef Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Thu, 24 Feb 2011 09:42:52 +0000
Subject: drm/i915: Fix unintended recursion in ironlake_disable_rc6

After disabling, we're meant to teardown the bo used for the contexts,
not recurse into ourselves again and preventing module unload.

Reported-and-tested-by: Ben Widawsky <bwidawsk@gmail.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
 drivers/gpu/drm/i915/intel_display.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'drivers/gpu')

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 9ca1bb2554fc..e79b25bbee6c 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -6575,7 +6575,7 @@ static void ironlake_disable_rc6(struct drm_device *dev)
 		POSTING_READ(RSTDBYCTL);
 	}
 
-	ironlake_disable_rc6(dev);
+	ironlake_teardown_rc6(dev);
 }
 
 static int ironlake_setup_rc6(struct drm_device *dev)
-- 
cgit v1.2.3


From 7db26623257a16c901a4b77bfc5096ee05304932 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Mon, 28 Feb 2011 14:22:12 +1000
Subject: drm/nv50-nvc0: make sure vma is definitely unmapped when destroying
 bo

Somehow fixes a misrendering + hang at GDM startup on my NVA8...

My first guess would have been stale TLB entries laying around that a new
bo then accidentally inherits.  That doesn't make a great deal of sense
however, as when we mapped the pages for the new bo the TLBs would've
gotten flushed anyway.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
---
 drivers/gpu/drm/nouveau/nouveau_bo.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

(limited to 'drivers/gpu')

diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index d38a4d9f9b0b..a52184007f5f 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -49,7 +49,10 @@ nouveau_bo_del_ttm(struct ttm_buffer_object *bo)
 		DRM_ERROR("bo %p still attached to GEM object\n", bo);
 
 	nv10_mem_put_tile_region(dev, nvbo->tile, NULL);
-	nouveau_vm_put(&nvbo->vma);
+	if (nvbo->vma.node) {
+		nouveau_vm_unmap(&nvbo->vma);
+		nouveau_vm_put(&nvbo->vma);
+	}
 	kfree(nvbo);
 }
 
-- 
cgit v1.2.3


From 1922756124ddd53846877416d92ba4a802bc658f Mon Sep 17 00:00:00 2001
From: Dave Airlie <airlied@redhat.com>
Date: Thu, 24 Feb 2011 08:35:06 +1000
Subject: drm: fix unsigned vs signed comparison issue in modeset ctl ioctl.

This fixes CVE-2011-1013.

Reported-by: Matthiew Herrb (OpenBSD X.org team)
Cc: stable@kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>
---
 drivers/gpu/drm/drm_irq.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

(limited to 'drivers/gpu')

diff --git a/drivers/gpu/drm/drm_irq.c b/drivers/gpu/drm/drm_irq.c
index 53120a72a48c..28d1d3c24d65 100644
--- a/drivers/gpu/drm/drm_irq.c
+++ b/drivers/gpu/drm/drm_irq.c
@@ -1012,7 +1012,8 @@ int drm_modeset_ctl(struct drm_device *dev, void *data,
 		    struct drm_file *file_priv)
 {
 	struct drm_modeset_ctl *modeset = data;
-	int crtc, ret = 0;
+	int ret = 0;
+	unsigned int crtc;
 
 	/* If drm_vblank_init() hasn't been called yet, just no-op */
 	if (!dev->num_crtcs)
-- 
cgit v1.2.3


From 6927faf30920b8c03dfa007e732642a1f1f20089 Mon Sep 17 00:00:00 2001
From: Jan Niehusmann <jan@gondor.com>
Date: Tue, 1 Mar 2011 23:24:16 +0100
Subject: drm/i915: fix memory corruption with GM965 and >4GB RAM

On a Thinkpad x61s, I noticed some memory corruption when
plugging/unplugging the external VGA connection. The symptoms are that
4 bytes at the beginning of a page get overwritten by zeroes.
The address of the corruption varies when rebooting the machine, but
stays constant while it's running (so it's possible to repeatedly write
some data and then corrupt it again by plugging the cable).

Further investigation revealed that the corrupted address is
(dev_priv->status_page_dmah->busaddr & 0xffffffff), ie. the beginning of
the hardware status page of the i965 graphics card, cut to 32 bits.

So it seems that for some memory access, the hardware uses only 32 bit
addressing. If the hardware status page is located >4GB, this
corrupts unrelated memory.

Signed-off-by: Jan Niehusmann <jan@gondor.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: stable@kernel.org
---
 drivers/gpu/drm/i915/i915_dma.c | 11 +++++++++++
 1 file changed, 11 insertions(+)

(limited to 'drivers/gpu')

diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index 17bd766f2081..e33d9be7df3b 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -1895,6 +1895,17 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 	if (IS_GEN2(dev))
 		dma_set_coherent_mask(&dev->pdev->dev, DMA_BIT_MASK(30));
 
+	/* 965GM sometimes incorrectly writes to hardware status page (HWS)
+	 * using 32bit addressing, overwriting memory if HWS is located
+	 * above 4GB.
+	 *
+	 * The documentation also mentions an issue with undefined
+	 * behaviour if any general state is accessed within a page above 4GB,
+	 * which also needs to be handled carefully.
+	 */
+	if (IS_BROADWATER(dev) || IS_CRESTLINE(dev))
+		dma_set_coherent_mask(&dev->pdev->dev, DMA_BIT_MASK(32));
+
 	mmio_bar = IS_GEN2(dev) ? 1 : 0;
 	dev_priv->regs = pci_iomap(dev->pdev, mmio_bar, 0);
 	if (!dev_priv->regs) {
-- 
cgit v1.2.3


From 73412c3854c877e5f37ad944ee8977addde4d35a Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Fri, 4 Mar 2011 09:58:36 +1000
Subject: drm/nouveau: allocate kernel's notifier object at end of block

The nv30/nv40 3d driver is about to start using DMA_FENCE from the 3D
object which, it turns out, doesn't like its DMA object to not be
aligned to a 4KiB boundary.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
---
 drivers/gpu/drm/nouveau/nouveau_dma.c      |  3 ++-
 drivers/gpu/drm/nouveau/nouveau_drv.h      |  3 ++-
 drivers/gpu/drm/nouveau/nouveau_notifier.c | 11 +++++++----
 3 files changed, 11 insertions(+), 6 deletions(-)

(limited to 'drivers/gpu')

diff --git a/drivers/gpu/drm/nouveau/nouveau_dma.c b/drivers/gpu/drm/nouveau/nouveau_dma.c
index 65699bfaaaea..b368ed74aad7 100644
--- a/drivers/gpu/drm/nouveau/nouveau_dma.c
+++ b/drivers/gpu/drm/nouveau/nouveau_dma.c
@@ -83,7 +83,8 @@ nouveau_dma_init(struct nouveau_channel *chan)
 		return ret;
 
 	/* NV_MEMORY_TO_MEMORY_FORMAT requires a notifier object */
-	ret = nouveau_notifier_alloc(chan, NvNotify0, 32, &chan->m2mf_ntfy);
+	ret = nouveau_notifier_alloc(chan, NvNotify0, 32, 0xfd0, 0x1000,
+				     &chan->m2mf_ntfy);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_drv.h b/drivers/gpu/drm/nouveau/nouveau_drv.h
index 9821fcacc3d2..982d70b12722 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drv.h
+++ b/drivers/gpu/drm/nouveau/nouveau_drv.h
@@ -852,7 +852,8 @@ extern const struct ttm_mem_type_manager_func nouveau_vram_manager;
 extern int  nouveau_notifier_init_channel(struct nouveau_channel *);
 extern void nouveau_notifier_takedown_channel(struct nouveau_channel *);
 extern int  nouveau_notifier_alloc(struct nouveau_channel *, uint32_t handle,
-				   int cout, uint32_t *offset);
+				   int cout, uint32_t start, uint32_t end,
+				   uint32_t *offset);
 extern int  nouveau_notifier_offset(struct nouveau_gpuobj *, uint32_t *);
 extern int  nouveau_ioctl_notifier_alloc(struct drm_device *, void *data,
 					 struct drm_file *);
diff --git a/drivers/gpu/drm/nouveau/nouveau_notifier.c b/drivers/gpu/drm/nouveau/nouveau_notifier.c
index fe29d604b820..5ea167623a82 100644
--- a/drivers/gpu/drm/nouveau/nouveau_notifier.c
+++ b/drivers/gpu/drm/nouveau/nouveau_notifier.c
@@ -96,7 +96,8 @@ nouveau_notifier_gpuobj_dtor(struct drm_device *dev,
 
 int
 nouveau_notifier_alloc(struct nouveau_channel *chan, uint32_t handle,
-		       int size, uint32_t *b_offset)
+		       int size, uint32_t start, uint32_t end,
+		       uint32_t *b_offset)
 {
 	struct drm_device *dev = chan->dev;
 	struct nouveau_gpuobj *nobj = NULL;
@@ -104,9 +105,10 @@ nouveau_notifier_alloc(struct nouveau_channel *chan, uint32_t handle,
 	uint32_t offset;
 	int target, ret;
 
-	mem = drm_mm_search_free(&chan->notifier_heap, size, 0, 0);
+	mem = drm_mm_search_free_in_range(&chan->notifier_heap, size, 0,
+					  start, end, 0);
 	if (mem)
-		mem = drm_mm_get_block(mem, size, 0);
+		mem = drm_mm_get_block_range(mem, size, 0, start, end);
 	if (!mem) {
 		NV_ERROR(dev, "Channel %d notifier block full\n", chan->id);
 		return -ENOMEM;
@@ -177,7 +179,8 @@ nouveau_ioctl_notifier_alloc(struct drm_device *dev, void *data,
 	if (IS_ERR(chan))
 		return PTR_ERR(chan);
 
-	ret = nouveau_notifier_alloc(chan, na->handle, na->size, &na->offset);
+	ret = nouveau_notifier_alloc(chan, na->handle, na->size, 0, 0x1000,
+				     &na->offset);
 	nouveau_channel_put(&chan);
 	return ret;
 }
-- 
cgit v1.2.3


From 0ee537abbd10a9abf11e1c22ee32a68e8c12ed4a Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Sun, 6 Mar 2011 09:03:16 +0000
Subject: Revert "drm/i915: fix corruptions on i8xx due to relaxed fencing"

This reverts commit c2e0eb167070a6e9dcb49c84c13c79a30d672431.

As it turns out, userspace already depends upon being able to enable
tiling on existing bo which it promises to be large enough for its
purposes i.e. it will not access beyond the end of the last full-tile
row.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=35016
Reported-and-tested-by: Kamal Mostafa <kamal@canonical.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
 drivers/gpu/drm/i915/i915_gem_tiling.c | 16 +---------------
 1 file changed, 1 insertion(+), 15 deletions(-)

(limited to 'drivers/gpu')

diff --git a/drivers/gpu/drm/i915/i915_gem_tiling.c b/drivers/gpu/drm/i915/i915_gem_tiling.c
index 79a04fde69b5..22a32b9932c5 100644
--- a/drivers/gpu/drm/i915/i915_gem_tiling.c
+++ b/drivers/gpu/drm/i915/i915_gem_tiling.c
@@ -184,7 +184,7 @@ i915_gem_detect_bit_6_swizzle(struct drm_device *dev)
 static bool
 i915_tiling_ok(struct drm_device *dev, int stride, int size, int tiling_mode)
 {
-	int tile_width, tile_height;
+	int tile_width;
 
 	/* Linear is always fine */
 	if (tiling_mode == I915_TILING_NONE)
@@ -215,20 +215,6 @@ i915_tiling_ok(struct drm_device *dev, int stride, int size, int tiling_mode)
 		}
 	}
 
-	if (IS_GEN2(dev) ||
-	    (tiling_mode == I915_TILING_Y && HAS_128_BYTE_Y_TILING(dev)))
-		tile_height = 32;
-	else
-		tile_height = 8;
-	/* i8xx is strange: It has 2 interleaved rows of tiles, so needs an even
-	 * number of tile rows. */
-	if (IS_GEN2(dev))
-		tile_height *= 2;
-
-	/* Size needs to be aligned to a full tile row */
-	if (size & (tile_height * stride - 1))
-		return false;
-
 	/* 965+ just needs multiples of tile width */
 	if (INTEL_INFO(dev)->gen >= 4) {
 		if (stride & (tile_width - 1))
-- 
cgit v1.2.3


From 91355834646328e7edc6bd25176ae44bcd7386c7 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Fri, 4 Mar 2011 19:22:40 +0000
Subject: drm/i915: Do not overflow the MMADDR write FIFO

Whilst the GT is powered down (rc6), writes to MMADDR are placed in a
FIFO by the System Agent. This is a limited resource, only 64 entries, of
which 20 are reserved for Display and PCH writes, and so we must take
care not to queue up too many writes. To avoid this, there is counter
which we can poll to ensure there are sufficient free entries in the
fifo.

"Issuing a write to a full FIFO is not supported; at worst it could
result in corruption or a system hang."

Reported-and-Tested-by: Matt Turner <mattst88@gmail.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=34056
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
 drivers/gpu/drm/i915/i915_debugfs.c     |  4 ++--
 drivers/gpu/drm/i915/i915_drv.c         | 14 ++++++++++++--
 drivers/gpu/drm/i915/i915_drv.h         | 20 +++++++++++++++-----
 drivers/gpu/drm/i915/i915_reg.h         |  2 ++
 drivers/gpu/drm/i915/intel_display.c    |  8 ++++----
 drivers/gpu/drm/i915/intel_ringbuffer.h | 13 +++++++------
 6 files changed, 42 insertions(+), 19 deletions(-)

(limited to 'drivers/gpu')

diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index 3601466c5502..4ff9b6cc973f 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -865,7 +865,7 @@ static int i915_cur_delayinfo(struct seq_file *m, void *unused)
 		int max_freq;
 
 		/* RPSTAT1 is in the GT power well */
-		__gen6_force_wake_get(dev_priv);
+		__gen6_gt_force_wake_get(dev_priv);
 
 		seq_printf(m, "GT_PERF_STATUS: 0x%08x\n", gt_perf_status);
 		seq_printf(m, "RPSTAT1: 0x%08x\n", I915_READ(GEN6_RPSTAT1));
@@ -888,7 +888,7 @@ static int i915_cur_delayinfo(struct seq_file *m, void *unused)
 		seq_printf(m, "Max non-overclocked (RP0) frequency: %dMHz\n",
 			   max_freq * 100);
 
-		__gen6_force_wake_put(dev_priv);
+		__gen6_gt_force_wake_put(dev_priv);
 	} else {
 		seq_printf(m, "no P-state info available\n");
 	}
diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index 0ad533f06af9..37d672a116db 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -254,7 +254,7 @@ void intel_detect_pch (struct drm_device *dev)
 	}
 }
 
-void __gen6_force_wake_get(struct drm_i915_private *dev_priv)
+void __gen6_gt_force_wake_get(struct drm_i915_private *dev_priv)
 {
 	int count;
 
@@ -270,12 +270,22 @@ void __gen6_force_wake_get(struct drm_i915_private *dev_priv)
 		udelay(10);
 }
 
-void __gen6_force_wake_put(struct drm_i915_private *dev_priv)
+void __gen6_gt_force_wake_put(struct drm_i915_private *dev_priv)
 {
 	I915_WRITE_NOTRACE(FORCEWAKE, 0);
 	POSTING_READ(FORCEWAKE);
 }
 
+void __gen6_gt_wait_for_fifo(struct drm_i915_private *dev_priv)
+{
+	int loop = 500;
+	u32 fifo = I915_READ_NOTRACE(GT_FIFO_FREE_ENTRIES);
+	while (fifo < 20 && loop--) {
+		udelay(10);
+		fifo = I915_READ_NOTRACE(GT_FIFO_FREE_ENTRIES);
+	}
+}
+
 static int i915_drm_freeze(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 65dfe81d0035..549c046b4ecc 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -1353,22 +1353,32 @@ __i915_write(64, q)
  * must be set to prevent GT core from power down and stale values being
  * returned.
  */
-void __gen6_force_wake_get(struct drm_i915_private *dev_priv);
-void __gen6_force_wake_put (struct drm_i915_private *dev_priv);
-static inline u32 i915_safe_read(struct drm_i915_private *dev_priv, u32 reg)
+void __gen6_gt_force_wake_get(struct drm_i915_private *dev_priv);
+void __gen6_gt_force_wake_put(struct drm_i915_private *dev_priv);
+void __gen6_gt_wait_for_fifo(struct drm_i915_private *dev_priv);
+
+static inline u32 i915_gt_read(struct drm_i915_private *dev_priv, u32 reg)
 {
 	u32 val;
 
 	if (dev_priv->info->gen >= 6) {
-		__gen6_force_wake_get(dev_priv);
+		__gen6_gt_force_wake_get(dev_priv);
 		val = I915_READ(reg);
-		__gen6_force_wake_put(dev_priv);
+		__gen6_gt_force_wake_put(dev_priv);
 	} else
 		val = I915_READ(reg);
 
 	return val;
 }
 
+static inline void i915_gt_write(struct drm_i915_private *dev_priv,
+				u32 reg, u32 val)
+{
+	if (dev_priv->info->gen >= 6)
+		__gen6_gt_wait_for_fifo(dev_priv);
+	I915_WRITE(reg, val);
+}
+
 static inline void
 i915_write(struct drm_i915_private *dev_priv, u32 reg, u64 val, int len)
 {
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 729d4233b763..3e6f486f4605 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -3261,6 +3261,8 @@
 #define  FORCEWAKE				0xA18C
 #define  FORCEWAKE_ACK				0x130090
 
+#define  GT_FIFO_FREE_ENTRIES			0x120008
+
 #define GEN6_RPNSWREQ				0xA008
 #define   GEN6_TURBO_DISABLE			(1<<31)
 #define   GEN6_FREQUENCY(x)			((x)<<25)
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index e79b25bbee6c..49fb54fd9a18 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -1219,7 +1219,7 @@ static void sandybridge_blit_fbc_update(struct drm_device *dev)
 	u32 blt_ecoskpd;
 
 	/* Make sure blitter notifies FBC of writes */
-	__gen6_force_wake_get(dev_priv);
+	__gen6_gt_force_wake_get(dev_priv);
 	blt_ecoskpd = I915_READ(GEN6_BLITTER_ECOSKPD);
 	blt_ecoskpd |= GEN6_BLITTER_FBC_NOTIFY <<
 		GEN6_BLITTER_LOCK_SHIFT;
@@ -1230,7 +1230,7 @@ static void sandybridge_blit_fbc_update(struct drm_device *dev)
 			 GEN6_BLITTER_LOCK_SHIFT);
 	I915_WRITE(GEN6_BLITTER_ECOSKPD, blt_ecoskpd);
 	POSTING_READ(GEN6_BLITTER_ECOSKPD);
-	__gen6_force_wake_put(dev_priv);
+	__gen6_gt_force_wake_put(dev_priv);
 }
 
 static void ironlake_enable_fbc(struct drm_crtc *crtc, unsigned long interval)
@@ -6282,7 +6282,7 @@ void gen6_enable_rps(struct drm_i915_private *dev_priv)
 	 * userspace...
 	 */
 	I915_WRITE(GEN6_RC_STATE, 0);
-	__gen6_force_wake_get(dev_priv);
+	__gen6_gt_force_wake_get(dev_priv);
 
 	/* disable the counters and set deterministic thresholds */
 	I915_WRITE(GEN6_RC_CONTROL, 0);
@@ -6380,7 +6380,7 @@ void gen6_enable_rps(struct drm_i915_private *dev_priv)
 	/* enable all PM interrupts */
 	I915_WRITE(GEN6_PMINTRMSK, 0);
 
-	__gen6_force_wake_put(dev_priv);
+	__gen6_gt_force_wake_put(dev_priv);
 }
 
 void intel_enable_clock_gating(struct drm_device *dev)
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h
index 6d6fde85a636..34306865a5df 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.h
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.h
@@ -14,22 +14,23 @@ struct  intel_hw_status_page {
 	struct		drm_i915_gem_object *obj;
 };
 
-#define I915_RING_READ(reg) i915_safe_read(dev_priv, reg)
+#define I915_RING_READ(reg) i915_gt_read(dev_priv, reg)
+#define I915_RING_WRITE(reg, val) i915_gt_write(dev_priv, reg, val)
 
 #define I915_READ_TAIL(ring) I915_RING_READ(RING_TAIL((ring)->mmio_base))
-#define I915_WRITE_TAIL(ring, val) I915_WRITE(RING_TAIL((ring)->mmio_base), val)
+#define I915_WRITE_TAIL(ring, val) I915_RING_WRITE(RING_TAIL((ring)->mmio_base), val)
 
 #define I915_READ_START(ring) I915_RING_READ(RING_START((ring)->mmio_base))
-#define I915_WRITE_START(ring, val) I915_WRITE(RING_START((ring)->mmio_base), val)
+#define I915_WRITE_START(ring, val) I915_RING_WRITE(RING_START((ring)->mmio_base), val)
 
 #define I915_READ_HEAD(ring)  I915_RING_READ(RING_HEAD((ring)->mmio_base))
-#define I915_WRITE_HEAD(ring, val) I915_WRITE(RING_HEAD((ring)->mmio_base), val)
+#define I915_WRITE_HEAD(ring, val) I915_RING_WRITE(RING_HEAD((ring)->mmio_base), val)
 
 #define I915_READ_CTL(ring) I915_RING_READ(RING_CTL((ring)->mmio_base))
-#define I915_WRITE_CTL(ring, val) I915_WRITE(RING_CTL((ring)->mmio_base), val)
+#define I915_WRITE_CTL(ring, val) I915_RING_WRITE(RING_CTL((ring)->mmio_base), val)
 
-#define I915_WRITE_IMR(ring, val) I915_WRITE(RING_IMR((ring)->mmio_base), val)
 #define I915_READ_IMR(ring) I915_RING_READ(RING_IMR((ring)->mmio_base))
+#define I915_WRITE_IMR(ring, val) I915_RING_WRITE(RING_IMR((ring)->mmio_base), val)
 
 #define I915_READ_NOPID(ring) I915_RING_READ(RING_NOPID((ring)->mmio_base))
 #define I915_READ_SYNC_0(ring) I915_RING_READ(RING_SYNC_0((ring)->mmio_base))
-- 
cgit v1.2.3


From a1656b9090f7008d2941c314f5a64724bea2ae37 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Fri, 4 Mar 2011 18:48:03 +0000
Subject: drm/i915: Disable GPU semaphores by default

Andi Kleen narrowed his GPU hangs on his Sugar Bay (SNB desktop) rev 09
down to the use of GPU semaphores, and we already know that they appear
broken up to Huron River (mobile) rev 08. (I'm optimistic that disabling
GPU semaphores is simply hiding another bug by the latency and
side-effects of the additional device interaction it introduces...)

However, use of semaphores is a massive performance improvement... Only
as long as the system remains stable. Enable at your peril.

Reported-by: Andi Kleen <andi-fd@firstfloor.org>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=33921
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
 drivers/gpu/drm/i915/i915_drv.c            | 3 +++
 drivers/gpu/drm/i915/i915_drv.h            | 1 +
 drivers/gpu/drm/i915/i915_gem_execbuffer.c | 4 ++--
 3 files changed, 6 insertions(+), 2 deletions(-)

(limited to 'drivers/gpu')

diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index 37d672a116db..22ec066adae6 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -46,6 +46,9 @@ module_param_named(fbpercrtc, i915_fbpercrtc, int, 0400);
 unsigned int i915_powersave = 1;
 module_param_named(powersave, i915_powersave, int, 0600);
 
+unsigned int i915_semaphores = 0;
+module_param_named(semaphores, i915_semaphores, int, 0600);
+
 unsigned int i915_enable_rc6 = 0;
 module_param_named(i915_enable_rc6, i915_enable_rc6, int, 0600);
 
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 549c046b4ecc..d023b9b33d60 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -956,6 +956,7 @@ extern struct drm_ioctl_desc i915_ioctls[];
 extern int i915_max_ioctl;
 extern unsigned int i915_fbpercrtc;
 extern unsigned int i915_powersave;
+extern unsigned int i915_semaphores;
 extern unsigned int i915_lvds_downclock;
 extern unsigned int i915_panel_use_ssc;
 extern unsigned int i915_enable_rc6;
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index d2f445e825f2..50ab1614571c 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -772,8 +772,8 @@ i915_gem_execbuffer_sync_rings(struct drm_i915_gem_object *obj,
 	if (from == NULL || to == from)
 		return 0;
 
-	/* XXX gpu semaphores are currently causing hard hangs on SNB mobile */
-	if (INTEL_INFO(obj->base.dev)->gen < 6 || IS_MOBILE(obj->base.dev))
+	/* XXX gpu semaphores are implicated in various hard hangs on SNB */
+	if (INTEL_INFO(obj->base.dev)->gen < 6 || !i915_semaphores)
 		return i915_gem_object_wait_rendering(obj, true);
 
 	idx = intel_ring_sync_index(from, to);
-- 
cgit v1.2.3


From 467cffba85791cdfce38c124d75bd578f4bb8625 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Mon, 7 Mar 2011 10:42:03 +0000
Subject: drm/i915: Rebind the buffer if its alignment constraints changes with
 tiling

Early gen3 and gen2 chipset do not have the relaxed per-surface tiling
constraints of the later chipsets, so we need to check that the GTT
alignment is correct for the new tiling. If it is not, we need to
rebind.

Reported-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
 drivers/gpu/drm/i915/i915_drv.h        |  3 +++
 drivers/gpu/drm/i915/i915_gem.c        |  2 +-
 drivers/gpu/drm/i915/i915_gem_tiling.c | 21 +++++++++++++++++----
 3 files changed, 21 insertions(+), 5 deletions(-)

(limited to 'drivers/gpu')

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index d023b9b33d60..456f40484838 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -1178,6 +1178,9 @@ void i915_gem_detach_phys_object(struct drm_device *dev,
 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_i915_gem_object *obj);
+
 /* i915_gem_gtt.c */
 void i915_gem_restore_gtt_mappings(struct drm_device *dev);
 int __must_check i915_gem_gtt_bind_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 cf4f74c7c6fb..36e66cc5225e 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1398,7 +1398,7 @@ i915_gem_get_gtt_alignment(struct drm_i915_gem_object *obj)
  * Return the required GTT alignment for an object, only taking into account
  * unfenced tiled surface requirements.
  */
-static uint32_t
+uint32_t
 i915_gem_get_unfenced_gtt_alignment(struct drm_i915_gem_object *obj)
 {
 	struct drm_device *dev = obj->base.dev;
diff --git a/drivers/gpu/drm/i915/i915_gem_tiling.c b/drivers/gpu/drm/i915/i915_gem_tiling.c
index 22a32b9932c5..d64843e18df2 100644
--- a/drivers/gpu/drm/i915/i915_gem_tiling.c
+++ b/drivers/gpu/drm/i915/i915_gem_tiling.c
@@ -349,14 +349,27 @@ i915_gem_set_tiling(struct drm_device *dev, void *data,
 			(obj->gtt_offset + obj->base.size <= dev_priv->mm.gtt_mappable_end &&
 			 i915_gem_object_fence_ok(obj, args->tiling_mode));
 
-		obj->tiling_changed = true;
-		obj->tiling_mode = args->tiling_mode;
-		obj->stride = args->stride;
+		/* Rebind if we need a change of alignment */
+		if (!obj->map_and_fenceable) {
+			u32 unfenced_alignment =
+				i915_gem_get_unfenced_gtt_alignment(obj);
+			if (obj->gtt_offset & (unfenced_alignment - 1))
+				ret = i915_gem_object_unbind(obj);
+		}
+
+		if (ret == 0) {
+			obj->tiling_changed = true;
+			obj->tiling_mode = args->tiling_mode;
+			obj->stride = args->stride;
+		}
 	}
+	/* we have to maintain this existing ABI... */
+	args->stride = obj->stride;
+	args->tiling_mode = obj->tiling_mode;
 	drm_gem_object_unreference(&obj->base);
 	mutex_unlock(&dev->struct_mutex);
 
-	return 0;
+	return ret;
 }
 
 /**
-- 
cgit v1.2.3


From ef1b287169cd3d1e428c8ed8222e0bbf733d5dbb Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Mon, 7 Mar 2011 17:18:03 +1000
Subject: drm/nouveau: fix regression causing ttm to not be able to evict vram

TTM assumes an error condition from man->func->get_node() means that
something went horribly wrong, and causes it to bail.

The driver is supposed to return 0, and leave mm_node == NULL to
signal that it couldn't allocate any memory.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
---
 drivers/gpu/drm/nouveau/nouveau_mem.c | 6 ++++--
 drivers/gpu/drm/nouveau/nouveau_mm.c  | 2 +-
 2 files changed, 5 insertions(+), 3 deletions(-)

(limited to 'drivers/gpu')

diff --git a/drivers/gpu/drm/nouveau/nouveau_mem.c b/drivers/gpu/drm/nouveau/nouveau_mem.c
index 26347b7cd872..b0fb9bdcddb7 100644
--- a/drivers/gpu/drm/nouveau/nouveau_mem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_mem.c
@@ -725,8 +725,10 @@ nouveau_vram_manager_new(struct ttm_mem_type_manager *man,
 	ret = vram->get(dev, mem->num_pages << PAGE_SHIFT,
 			mem->page_alignment << PAGE_SHIFT, size_nc,
 			(nvbo->tile_flags >> 8) & 0xff, &node);
-	if (ret)
-		return ret;
+	if (ret) {
+		mem->mm_node = NULL;
+		return (ret == -ENOSPC) ? 0 : ret;
+	}
 
 	node->page_shift = 12;
 	if (nvbo->vma.node)
diff --git a/drivers/gpu/drm/nouveau/nouveau_mm.c b/drivers/gpu/drm/nouveau/nouveau_mm.c
index 8844b50c3e54..7609756b6faf 100644
--- a/drivers/gpu/drm/nouveau/nouveau_mm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_mm.c
@@ -123,7 +123,7 @@ nouveau_mm_get(struct nouveau_mm *rmm, int type, u32 size, u32 size_nc,
 		return 0;
 	}
 
-	return -ENOMEM;
+	return -ENOSPC;
 }
 
 int
-- 
cgit v1.2.3


From 6f70a4c3d19e8e8e1047a4dbf0ca910fed39f619 Mon Sep 17 00:00:00 2001
From: Ben Skeggs <bskeggs@redhat.com>
Date: Mon, 7 Mar 2011 17:18:04 +1000
Subject: drm/nv50-nvc0: prevent multiple vm/bar flushes occuring
 simultanenously

The per-vm mutex doesn't prevent this completely, a flush coming from the
BAR VM could potentially happen at the same time as one for the channel
VM.  Not to mention that if/when we get per-client/channel VM, this will
happen far more frequently.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
---
 drivers/gpu/drm/nouveau/nv50_instmem.c | 8 ++++++++
 drivers/gpu/drm/nouveau/nv50_vm.c      | 4 ++++
 2 files changed, 12 insertions(+)

(limited to 'drivers/gpu')

diff --git a/drivers/gpu/drm/nouveau/nv50_instmem.c b/drivers/gpu/drm/nouveau/nv50_instmem.c
index ea0041810ae3..e57caa2a00e3 100644
--- a/drivers/gpu/drm/nouveau/nv50_instmem.c
+++ b/drivers/gpu/drm/nouveau/nv50_instmem.c
@@ -403,16 +403,24 @@ nv50_instmem_unmap(struct nouveau_gpuobj *gpuobj)
 void
 nv50_instmem_flush(struct drm_device *dev)
 {
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+
+	spin_lock(&dev_priv->ramin_lock);
 	nv_wr32(dev, 0x00330c, 0x00000001);
 	if (!nv_wait(dev, 0x00330c, 0x00000002, 0x00000000))
 		NV_ERROR(dev, "PRAMIN flush timeout\n");
+	spin_unlock(&dev_priv->ramin_lock);
 }
 
 void
 nv84_instmem_flush(struct drm_device *dev)
 {
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+
+	spin_lock(&dev_priv->ramin_lock);
 	nv_wr32(dev, 0x070000, 0x00000001);
 	if (!nv_wait(dev, 0x070000, 0x00000002, 0x00000000))
 		NV_ERROR(dev, "PRAMIN flush timeout\n");
+	spin_unlock(&dev_priv->ramin_lock);
 }
 
diff --git a/drivers/gpu/drm/nouveau/nv50_vm.c b/drivers/gpu/drm/nouveau/nv50_vm.c
index 459ff08241e5..6144156f255a 100644
--- a/drivers/gpu/drm/nouveau/nv50_vm.c
+++ b/drivers/gpu/drm/nouveau/nv50_vm.c
@@ -169,7 +169,11 @@ nv50_vm_flush(struct nouveau_vm *vm)
 void
 nv50_vm_flush_engine(struct drm_device *dev, int engine)
 {
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+
+	spin_lock(&dev_priv->ramin_lock);
 	nv_wr32(dev, 0x100c80, (engine << 16) | 1);
 	if (!nv_wait(dev, 0x100c80, 0x00000001, 0x00000000))
 		NV_ERROR(dev, "vm flush timeout: engine %d\n", engine);
+	spin_unlock(&dev_priv->ramin_lock);
 }
-- 
cgit v1.2.3


From 062ac622e03a8be5f894555ece540d63a54ae8bd Mon Sep 17 00:00:00 2001
From: roel <roel.kluin@gmail.com>
Date: Mon, 7 Mar 2011 18:00:34 +0100
Subject: drm: index i shadowed in 2nd loop

Index i was already used in thhe first loop

Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
---
 drivers/gpu/drm/drm_fb_helper.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

(limited to 'drivers/gpu')

diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index 6977a1ce9d98..f73ef4390db6 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -672,7 +672,7 @@ int drm_fb_helper_setcmap(struct fb_cmap *cmap, struct fb_info *info)
 	struct drm_crtc_helper_funcs *crtc_funcs;
 	u16 *red, *green, *blue, *transp;
 	struct drm_crtc *crtc;
-	int i, rc = 0;
+	int i, j, rc = 0;
 	int start;
 
 	for (i = 0; i < fb_helper->crtc_count; i++) {
@@ -685,7 +685,7 @@ int drm_fb_helper_setcmap(struct fb_cmap *cmap, struct fb_info *info)
 		transp = cmap->transp;
 		start = cmap->start;
 
-		for (i = 0; i < cmap->len; i++) {
+		for (j = 0; j < cmap->len; j++) {
 			u16 hred, hgreen, hblue, htransp = 0xffff;
 
 			hred = *red++;
-- 
cgit v1.2.3