[Intel-gfx] [PATCH] drm/i915/guc: Assert that we switch between known ggtt->invalidate functions

[Chris Wilson]

On Thu, Jun 01, 2017 at 01:24:55PM +0300, Joonas Lahtinen wrote:
> On to, 2017-06-01 at 12:03 +0200, Michal Wajdeczko wrote:
> > On Thu, Jun 01, 2017 at 10:04:46AM +0100, Chris Wilson wrote:
> > >  
> > >  void i915_ggtt_disable_guc(struct drm_i915_private *i915)
> > >  {
> > > -	if (i915->ggtt.invalidate == guc_ggtt_invalidate)
> > > -		i915->ggtt.invalidate = gen6_ggtt_invalidate;
> > > +	/* We should only be called after i915_ggtt_enable_guc() */
> > > +	GEM_BUG_ON(i915->ggtt.invalidate != guc_ggtt_invalidate);
> > > +
> > > +	i915->ggtt.invalidate = gen6_ggtt_invalidate;
> > >  }
> > 
> > While this looks correct today, it may not work in the future if we
> > will need somethig other than gen6_ggtt_invalidate() as base invalidate
> > function or guc_gtt_invalidate() as the one for the guc. Just a head up.
> 
> Currently the assignment is directly to gen6_ggtt_invalidate, no
> questions asked. So I don't think the assert could be much more :)
> 
> Maybe GuC code should backup the invalidate function before overriding.

Possibly, but I hope the alternatives are a little better known so that
we don't have to do save/restore of function pointers too often.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre