Quoting Daniel Vetter (2018-07-11 09:19:58) > On Wed, Jul 11, 2018 at 08:36:03AM +0100, Chris Wilson wrote: > > @@ -5630,32 +5631,33 @@ i915_gem_cleanup_engines(struct drm_i915_private *dev_priv) > > dev_priv->gt.cleanup_engine(engine); > > } > > > > -void > > -i915_gem_load_init_fences(struct drm_i915_private *dev_priv) > > +void i915_ggtt_init_fences(struct i915_ggtt *ggtt) > > Follow-up patch to put this into i915_gem_fence_reg.c would be kinda cool > I think. The decl is already in i915_gem_fence_reg.h already anyway. It's caught between two worlds, iirc, I pulled it into i915_gem_gtt so that it's inline with ggtt_init. From my pov, all we do inside this function is setup the bookkeeping associated with tracking our usage of fences, with a callout to the i915_gem_fence_reg.c backend to sanitize the registers. At the moment, I'm still falling on the side of "this touches i915->ggtt therefore it should be alongside the rest of i915->ggtt initialisation". But grass, it always grows greener on the other side. -Chris _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
