drm/i915: don't call intel_fbc_update() at intel_unpin_work_fn()
authorPaulo Zanoni <paulo.r.zanoni@intel.com>
Tue, 14 Jul 2015 19:29:11 +0000 (16:29 -0300)
committerDaniel Vetter <daniel.vetter@ffwll.ch>
Wed, 5 Aug 2015 08:00:16 +0000 (10:00 +0200)
Because intel_unpin_work_fn() already calls
intel_frontbuffer_flip_complete() which will call intel_fbc_flush()
which will call intel_fbc_update() when needed.

We couldn't fix this previously due to the fact that FBC was not
properly behaving as intended on frontbuffer flushes, but now that
this is fixed, we can remove the additional call.

Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
drivers/gpu/drm/i915/intel_display.c

index af0bcfee4771fd1af6ed0d2f06a5a9d4cf500f65..3dbe3162c848c2ffe8400d2c489fe12ed45f6e15 100644 (file)
@@ -10762,15 +10762,12 @@ static void intel_unpin_work_fn(struct work_struct *__work)
                container_of(__work, struct intel_unpin_work, work);
        struct intel_crtc *crtc = to_intel_crtc(work->crtc);
        struct drm_device *dev = crtc->base.dev;
-       struct drm_i915_private *dev_priv = dev->dev_private;
        struct drm_plane *primary = crtc->base.primary;
 
        mutex_lock(&dev->struct_mutex);
        intel_unpin_fb_obj(work->old_fb, primary->state);
        drm_gem_object_unreference(&work->pending_flip_obj->base);
 
-       intel_fbc_update(dev_priv);
-
        if (work->flip_queued_req)
                i915_gem_request_assign(&work->flip_queued_req, NULL);
        mutex_unlock(&dev->struct_mutex);