Re: [Intel-gfx] [PATCH v2 1/7] drm/i915: Flush submission tasklet after bumping priority

2018-05-08 Thread Tvrtko Ursulin
On 08/05/2018 10:45, Chris Wilson wrote: Quoting Tvrtko Ursulin (2018-05-08 10:40:43) On 07/05/2018 14:57, Chris Wilson wrote: When called from process context tasklet_schedule() defers itself to ksoftirqd. From experience this may cause unacceptable latencies of over 200ms in executing the

Re: [Intel-gfx] [PATCH v2 1/7] drm/i915: Flush submission tasklet after bumping priority

2018-05-08 Thread Chris Wilson
Quoting Tvrtko Ursulin (2018-05-08 10:40:43) > > On 07/05/2018 14:57, Chris Wilson wrote: > > When called from process context tasklet_schedule() defers itself to > > ksoftirqd. From experience this may cause unacceptable latencies of over > > 200ms in executing the submission tasklet, our goal

Re: [Intel-gfx] [PATCH v2 1/7] drm/i915: Flush submission tasklet after bumping priority

2018-05-08 Thread Tvrtko Ursulin
On 07/05/2018 14:57, Chris Wilson wrote: When called from process context tasklet_schedule() defers itself to ksoftirqd. From experience this may cause unacceptable latencies of over 200ms in executing the submission tasklet, our goal is to reprioritise the HW execution queue and trigger HW

[Intel-gfx] [PATCH v2 1/7] drm/i915: Flush submission tasklet after bumping priority

2018-05-07 Thread Chris Wilson
When called from process context tasklet_schedule() defers itself to ksoftirqd. From experience this may cause unacceptable latencies of over 200ms in executing the submission tasklet, our goal is to reprioritise the HW execution queue and trigger HW preemption immediately, so disable bh over the