Em Qua, 2016-03-23 às 14:33 +0100, Maarten Lankhorst escreveu: > intel_post_plane_update did an extra vblank wait that's no longer > needed when enabling ips. > > Changes since v1: > - Add comment explaining why vblank wait is performed. (Paulo) > > Signed-off-by: Maarten Lankhorst <[email protected]> > --- > Noticed this one still in my branch, does this look ok? > > ~Maarten > > diff --git a/drivers/gpu/drm/i915/intel_display.c > b/drivers/gpu/drm/i915/intel_display.c > index dc920cfc3fd7..6dcc159ce6ac 100644 > --- a/drivers/gpu/drm/i915/intel_display.c > +++ b/drivers/gpu/drm/i915/intel_display.c > @@ -4454,8 +4454,11 @@ void hsw_enable_ips(struct intel_crtc *crtc) > if (!crtc->config->ips_enabled) > return; > > - /* We can only enable IPS after we enable a plane and wait > for a vblank */ > - intel_wait_for_vblank(dev, crtc->pipe); > + /* > + * We can only enable IPS after we enable a plane and wait > for a vblank > + * This function is called from post_plane_update, which is > run after > + * a vblank wait. > + */
There are actually other callers, such as intel_dp.c, but I suppose they are correct since they can only be called when the planes are already enabled, so: Reviewed-by: Paulo Zanoni <[email protected]> > > assert_plane_enabled(dev_priv, crtc->plane); > if (IS_BROADWELL(dev)) { > _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
