Re: [Intel-gfx] [PATCH] drm/i915: fix SFC reset flow

2019-09-18 Thread Tvrtko Ursulin
On 17/09/2019 19:29, Daniele Ceraolo Spurio wrote: On 9/17/2019 3:22 AM, Tvrtko Ursulin wrote: On 16/09/2019 22:41, Daniele Ceraolo Spurio wrote: Our assumption that the we can ask the HW to lock the SFC even if not currently in use does not match the HW commitment. The expectation from the

Re: [Intel-gfx] [PATCH] drm/i915: fix SFC reset flow

2019-09-17 Thread Daniele Ceraolo Spurio
On 9/17/2019 12:49 PM, Chris Wilson wrote: Quoting Daniele Ceraolo Spurio (2019-09-17 20:45:02) On 9/17/2019 11:57 AM, Chris Wilson wrote: Quoting Daniele Ceraolo Spurio (2019-09-17 19:36:35) On 9/17/2019 12:57 AM, Chris Wilson wrote: Quoting Daniele Ceraolo Spurio (2019-09-16 22:41:04) O

Re: [Intel-gfx] [PATCH] drm/i915: fix SFC reset flow

2019-09-17 Thread Chris Wilson
Quoting Daniele Ceraolo Spurio (2019-09-17 20:45:02) > > > On 9/17/2019 11:57 AM, Chris Wilson wrote: > > Quoting Daniele Ceraolo Spurio (2019-09-17 19:36:35) > >> > >> On 9/17/2019 12:57 AM, Chris Wilson wrote: > >>> Quoting Daniele Ceraolo Spurio (2019-09-16 22:41:04) > Our assumption that

Re: [Intel-gfx] [PATCH] drm/i915: fix SFC reset flow

2019-09-17 Thread Daniele Ceraolo Spurio
On 9/17/2019 11:57 AM, Chris Wilson wrote: Quoting Daniele Ceraolo Spurio (2019-09-17 19:36:35) On 9/17/2019 12:57 AM, Chris Wilson wrote: Quoting Daniele Ceraolo Spurio (2019-09-16 22:41:04) Our assumption that the we can ask the HW to lock the SFC even if not currently in use does not mat

Re: [Intel-gfx] [PATCH] drm/i915: fix SFC reset flow

2019-09-17 Thread Chris Wilson
Quoting Daniele Ceraolo Spurio (2019-09-17 19:36:35) > > > On 9/17/2019 12:57 AM, Chris Wilson wrote: > > Quoting Daniele Ceraolo Spurio (2019-09-16 22:41:04) > >> Our assumption that the we can ask the HW to lock the SFC even if not > >> currently in use does not match the HW commitment. The exp

Re: [Intel-gfx] [PATCH] drm/i915: fix SFC reset flow

2019-09-17 Thread Daniele Ceraolo Spurio
On 9/17/2019 12:57 AM, Chris Wilson wrote: Quoting Daniele Ceraolo Spurio (2019-09-16 22:41:04) Our assumption that the we can ask the HW to lock the SFC even if not currently in use does not match the HW commitment. The expectation from the HW is that SW will not try to lock the SFC if the en

Re: [Intel-gfx] [PATCH] drm/i915: fix SFC reset flow

2019-09-17 Thread Daniele Ceraolo Spurio
On 9/17/2019 3:22 AM, Tvrtko Ursulin wrote: On 16/09/2019 22:41, Daniele Ceraolo Spurio wrote: Our assumption that the we can ask the HW to lock the SFC even if not currently in use does not match the HW commitment. The expectation from the HW is that SW will not try to lock the SFC if the en

Re: [Intel-gfx] [PATCH] drm/i915: fix SFC reset flow

2019-09-17 Thread Chris Wilson
Quoting Daniele Ceraolo Spurio (2019-09-16 22:41:04) > @@ -401,7 +407,10 @@ static void gen11_unlock_sfc(struct intel_engine_cs > *engine) > return; > } > > - rmw_clear_fw(uncore, sfc_forced_lock, sfc_forced_lock_bit); > + lock = intel_uncore_read_fw(uncore, s

Re: [Intel-gfx] [PATCH] drm/i915: fix SFC reset flow

2019-09-17 Thread Tvrtko Ursulin
On 16/09/2019 22:41, Daniele Ceraolo Spurio wrote: Our assumption that the we can ask the HW to lock the SFC even if not currently in use does not match the HW commitment. The expectation from the HW is that SW will not try to lock the SFC if the engine is not using it and if we do that the beha

Re: [Intel-gfx] [PATCH] drm/i915: fix SFC reset flow

2019-09-17 Thread Chris Wilson
Quoting Daniele Ceraolo Spurio (2019-09-16 22:41:04) > Our assumption that the we can ask the HW to lock the SFC even if not > currently in use does not match the HW commitment. The expectation from > the HW is that SW will not try to lock the SFC if the engine is not > using it and if we do that t