[PATCH] drm/i915: Don't clobber crtc->fb when queue_flip fails

* [PATCH] drm/i915: Don't clobber crtc->fb when queue_flip fails
@ 2013-02-22 12:17 ville.syrjala
  2013-02-22 13:31 ` [Intel-gfx] " Chris Wilson
  2013-02-22 13:46 ` Mika Kuoppala
  0 siblings, 2 replies; 6+ messages in thread
From: ville.syrjala @ 2013-02-22 12:17 UTC (permalink / raw)
  To: intel-gfx; +Cc: Ville Syrjälä, stable

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Point crtc->fb the the new framebuffer only after we know that the flip
was succesfully queued.

While at it, move the intel_fb and obj assignments a bit close to where
they're used.

Cc: stable@vger.kernel.org
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
It looks like Mika hit this rather easily in his ARB_robustness work.
Waiting for him to confirm whether this really fixes the pin_count
underflow bug he's seeing.

 drivers/gpu/drm/i915/intel_display.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 0ff10b3..e7684f1 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -7311,9 +7311,6 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc,
 	intel_crtc->unpin_work = work;
 	spin_unlock_irqrestore(&dev->event_lock, flags);
 
-	intel_fb = to_intel_framebuffer(fb);
-	obj = intel_fb->obj;
-
 	if (atomic_read(&intel_crtc->unpin_work_count) >= 2)
 		flush_workqueue(dev_priv->wq);
 
@@ -7321,12 +7318,13 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc,
 	if (ret)
 		goto cleanup;
 
+	intel_fb = to_intel_framebuffer(fb);
+	obj = intel_fb->obj;
+
 	/* Reference the objects for the scheduled work. */
 	drm_gem_object_reference(&work->old_fb_obj->base);
 	drm_gem_object_reference(&obj->base);
 
-	crtc->fb = fb;
-
 	work->pending_flip_obj = obj;
 
 	work->enable_stall_check = true;
@@ -7338,6 +7336,8 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc,
 	if (ret)
 		goto cleanup_pending;
 
+	crtc->fb = fb;
+
 	intel_disable_fbc(dev);
 	intel_mark_fb_busy(obj);
 	mutex_unlock(&dev->struct_mutex);
-- 
1.7.12.4

^ permalink raw reply related	[flat|nested] 6+ messages in thread