From: Ville Syrj?l? <ville.syrjala at linux.intel.com> i915_gem_execbuffer_wait_for_flips() only works on some old hardware, and it's not required by sane user space code. So assume that any insane user space is limited to <= gen5, and just skip i915_gem_execbuffer_wait_for_flips() for anything more recent. Also eliminate the extra pending flip wait in intel_finish_fb() which is based on the pending_flip counter. This wait doesn't actually do anything when used on the current crtc front buffer, so the whole thing is pointless. There's a new mechanism for waiting for pending flips, which actually does the right thing. Signed-off-by: Ville Syrj?l? <ville.syrjala at linux.intel.com> --- drivers/gpu/drm/i915/i915_gem_execbuffer.c | 4 ++++ drivers/gpu/drm/i915/intel_display.c | 4 ---- 2 files changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c index 91d43d5..14761ac 100644 --- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c @@ -616,6 +616,10 @@ i915_gem_execbuffer_wait_for_flips(struct intel_ring_buffer *ring, u32 flips) u32 plane, flip_mask; int ret; + /* This function exists only to please legacy userland. */ + if (INTEL_INFO(ring->dev)->gen >= 6) + return 0; + /* 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. diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 7bf4749..b1e8150 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -2182,10 +2182,6 @@ intel_finish_fb(struct drm_framebuffer *old_fb) bool was_interruptible = dev_priv->mm.interruptible; int ret; - 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 -- 1.7.8.6
