[Intel-gfx] [PATCH 3/6] drm/i915: Only track live rings for retiring

2018-04-24 Thread Chris Wilson
We don't need to track every ring for its lifetime as they are managed by the contexts/engines. What we do want to track are the live rings so that we can sporadically clean up requests if userspace falls behind. We can simply restrict the gt->rings list to being only gt->live_rings. v2: s/live/ac

Re: [Intel-gfx] [PATCH 3/6] drm/i915: Only track live rings for retiring

2018-04-23 Thread Tvrtko Ursulin
On 23/04/2018 11:36, Chris Wilson wrote: Quoting Tvrtko Ursulin (2018-04-23 11:25:54) On 23/04/2018 11:13, Chris Wilson wrote: We don't need to track every ring for its lifetime as they are managed by the contexts/engines. What we do want to track are the live rings so that we can sporadicall

Re: [Intel-gfx] [PATCH 3/6] drm/i915: Only track live rings for retiring

2018-04-23 Thread Chris Wilson
Quoting Tvrtko Ursulin (2018-04-23 11:25:54) > > On 23/04/2018 11:13, Chris Wilson wrote: > > We don't need to track every ring for its lifetime as they are managed > > by the contexts/engines. What we do want to track are the live rings so > > that we can sporadically clean up requests if userspa

Re: [Intel-gfx] [PATCH 3/6] drm/i915: Only track live rings for retiring

2018-04-23 Thread Tvrtko Ursulin
On 23/04/2018 11:13, Chris Wilson wrote: We don't need to track every ring for its lifetime as they are managed by the contexts/engines. What we do want to track are the live rings so that we can sporadically clean up requests if userspace falls behind. We can simply restrict the gt->rings list

[Intel-gfx] [PATCH 3/6] drm/i915: Only track live rings for retiring

2018-04-23 Thread Chris Wilson
We don't need to track every ring for its lifetime as they are managed by the contexts/engines. What we do want to track are the live rings so that we can sporadically clean up requests if userspace falls behind. We can simply restrict the gt->rings list to being only gt->live_rings. Suggested-by: