[Intel-gfx] [PATCH 14/15] drm/i915: Track frontbuffer invalidation/flushing

[Daniel Vetter]

On Tue, Jun 17, 2014 at 07:52:25AM +0100, Chris Wilson wrote:
> On Mon, Jun 16, 2014 at 07:51:34PM +0200, Daniel Vetter wrote:
> > +static void intel_mark_fb_busy(struct drm_device *dev,
> > +			       unsigned frontbuffer_bits,
> > +			       struct intel_engine_cs *ring)
> >  {
> > -	struct drm_device *dev = obj->base.dev;
> >  	enum pipe pipe;
> >  
> > -	intel_edp_psr_exit(dev);
> > -
> >  	if (!i915.powersave)
> >  		return;
> 
> I think this wants to be moved from here down to
> intel_increase_pllclock. It's a little dangerous here as we use it
> inconsistently, and we certainly want to keep tracking fb bits for other
> reasons.

powersave was originally the overall blocker for all things power saving
related. We already have separate knobs for all real features, so I guess
we could just rip this out - it doesn't actually guard much of anything,
is inconsistent and always enabled anyway.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch