Re: [Intel-gfx] [PATCH] drm/i915: Performed deferred clflush inside set-cache-level

2015-01-21 Thread Daniel Vetter
On Thu, Jan 15, 2015 at 08:54:54AM +, Chris Wilson wrote: > On Wed, Jan 14, 2015 at 08:46:09PM +0100, Daniel Vetter wrote: > > > + if (obj->cache_dirty && > > > +obj->base.write_domain != I915_GEM_DOMAIN_CPU && > > > +cpu_write_needs_clflush(obj) && > > > +i915_gem_clflush_object(ob

Re: [Intel-gfx] [PATCH] drm/i915: Performed deferred clflush inside set-cache-level

2015-01-21 Thread Chris Wilson
On Thu, Jan 15, 2015 at 08:54:54AM +, Chris Wilson wrote: > On Wed, Jan 14, 2015 at 08:46:09PM +0100, Daniel Vetter wrote: > > Also, can we please have a testcase to at lest exercise the > > codepath? It sounds like a real functional tests using crc is a bit more > > work, but just poking at th

Re: [Intel-gfx] [PATCH] drm/i915: Performed deferred clflush inside set-cache-level

2015-01-15 Thread Chris Wilson
On Wed, Jan 14, 2015 at 08:46:09PM +0100, Daniel Vetter wrote: > > + if (obj->cache_dirty && > > +obj->base.write_domain != I915_GEM_DOMAIN_CPU && > > +cpu_write_needs_clflush(obj) && > > +i915_gem_clflush_object(obj, true)) > > Imo hiding the actual action in the if condition like thi

Re: [Intel-gfx] [PATCH] drm/i915: Performed deferred clflush inside set-cache-level

2015-01-14 Thread Daniel Vetter
On Tue, Jan 13, 2015 at 01:32:52PM +, Chris Wilson wrote: > Currently we are hitting the WARN inside > i915_gem_object_set_cache_level() as we can now have an unbound object > in the GTT write domain (due to 43566dedde54f9 "drm/i915: Broaden > application of set-domain(GTT)"). To avoid the warn

Re: [Intel-gfx] [PATCH] drm/i915: Performed deferred clflush inside set-cache-level

2015-01-13 Thread shuang . he
Tested-By: PRC QA PRTS (Patch Regression Test System Contact: shuang...@intel.com) -Summary- Platform Delta drm-intel-nightly Series Applied PNV 354/354

Re: [Intel-gfx] [PATCH] drm/i915: Performed deferred clflush inside set-cache-level

2015-01-13 Thread Chris Wilson
On Tue, Jan 13, 2015 at 10:23:55PM +0200, Ville Syrjälä wrote: > On Tue, Jan 13, 2015 at 01:32:52PM +, Chris Wilson wrote: > > Currently we are hitting the WARN inside > > i915_gem_object_set_cache_level() as we can now have an unbound object > > in the GTT write domain (due to 43566dedde54f9 "

Re: [Intel-gfx] [PATCH] drm/i915: Performed deferred clflush inside set-cache-level

2015-01-13 Thread Ville Syrjälä
On Tue, Jan 13, 2015 at 01:32:52PM +, Chris Wilson wrote: > Currently we are hitting the WARN inside > i915_gem_object_set_cache_level() as we can now have an unbound object > in the GTT write domain (due to 43566dedde54f9 "drm/i915: Broaden > application of set-domain(GTT)"). To avoid the warn

Re: [Intel-gfx] [PATCH] drm/i915: Performed deferred clflush inside set-cache-level

2015-01-13 Thread Jani Nikula
On Tue, 13 Jan 2015, Chris Wilson wrote: > Currently we are hitting the WARN inside > i915_gem_object_set_cache_level() as we can now have an unbound object > in the GTT write domain (due to 43566dedde54f9 "drm/i915: Broaden > application of set-domain(GTT)"). To avoid the warning, we need to trac

[Intel-gfx] [PATCH] drm/i915: Performed deferred clflush inside set-cache-level

2015-01-13 Thread Chris Wilson
Currently we are hitting the WARN inside i915_gem_object_set_cache_level() as we can now have an unbound object in the GTT write domain (due to 43566dedde54f9 "drm/i915: Broaden application of set-domain(GTT)"). To avoid the warning, we need to track when we elided the clflush on a cacheable object