From: Vincent Palatin <vpalatin@xxxxxxxxxxxx> On the resume path, wait_for_vblank is called 4 times and waits until its timeout. To reduce the delays on that path, let's compute the maximum vblank delay instead of always using the arbitrary 50 ms value. Signed-off-by: Vincent Palatin <vpalatin@xxxxxxxxxxxx> BUG=chrome-os-partner:13364 TEST=On Link, measure resume time and see a 130ms improvement. Change-Id: I88231de43573432b6d71255f622eb9860fe3c28d Reviewed-on: https://gerrit.chromium.org/gerrit/34624 Reviewed-by: Sameer Nanda <snanda@xxxxxxxxxxxx> Reviewed-by: Stéphane Marchesin <marcheu@xxxxxxxxxxxx> Commit-Ready: Vincent Palatin <vpalatin@xxxxxxxxxxxx> Tested-by: Vincent Palatin <vpalatin@xxxxxxxxxxxx> [marcheu: fixup for 3.8 rebase] --- drivers/gpu/drm/i915/intel_display.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index da1ad9c..0bc822e 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -967,6 +967,9 @@ void intel_wait_for_vblank(struct drm_device *dev, int pipe) { struct drm_i915_private *dev_priv = dev->dev_private; int pipestat_reg = PIPESTAT(pipe); + struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe]; + int timeout = crtc->hwmode.vrefresh ? + DIV_ROUND_UP(1000, crtc->hwmode.vrefresh) : 50; if (INTEL_INFO(dev)->gen >= 5) { ironlake_wait_for_vblank(dev, pipe); @@ -992,7 +995,7 @@ void intel_wait_for_vblank(struct drm_device *dev, int pipe) /* Wait for vblank interrupt bit to set */ if (wait_for(I915_READ(pipestat_reg) & PIPE_VBLANK_INTERRUPT_STATUS, - 50)) + timeout)) DRM_DEBUG_KMS("vblank wait timed out\n"); } -- 1.8.3.2 _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/intel-gfx
