Chris,
Just to make sure, you want to just remove write_domain check from
if statement before clflush in execbuffer_move_to_gpu. Am I right?
I will try both (cache_dirty only vs cache_dirty & !cache_coherent)
and get back to you shortly.
On Wed, Apr 19, 2017 at 07:08:33PM +0100, Chris Wilson wrote:
> On Wed, Apr 19, 2017 at 09:52:28AM -0700, Dongwon Kim wrote:
> > I tried your patch but it didn't fix the original
> > problem. I think it is somehow related to the flushing condition
> > here:
> >
> > @@ -1129,10 +1129,8 @@ i915_gem_execbuffer_move_to_gpu(struct drm_i915_gem_request *req,
> > if (vma->exec_entry->flags & EXEC_OBJECT_ASYNC)
> > continue;
> >
> > if (obj->base.write_domain & I915_GEM_DOMAIN_CPU) {
> > + if (obj->base.write_domain & obj->cache_dirty)
> > i915_gem_clflush_object(obj, 0);
> > - obj->base.write_domain = 0;
> > - }
> >
> > here, we do clflush only if write_domain is not 0 even if cache_dirty
> > flag is set after your patch is applied.
>
> This can be just reduced to if (obj->cache_dirty) clflush().
>
> We're slightly better in that we don't set obj->cache_dirty so often for
> normal gpu rendering (just on transitions away from the gpu now), but it
> still means we will be redundantly checking for clflushes prior to
> rendering.
>
> Can you double check that this patch + if (obj->cache_dirty) works for you?
>
> What I guess I really want here is
> if (obj->cache_dirty & !obj->cache_coherent)
> essentially inlining the check from clflush.
> -Chris
>
> --
> Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
