Re: [Intel-gfx] [PATCH v1 1/3] drm/i915/guc: Flush context destruction worker at suspend

2023-08-14 Thread Teres Alexis, Alan Previn
> > > Rodrigo: And why here and not some upper layer? like in prepare > alan: wait_for_suspend does both the checking for idle as well as the > potential > wedging if guc or hw has hung at this late state. if i call from the upper > layer before this wait_for_suspend, it might be too early

Re: [PATCH v1 1/3] drm/i915/guc: Flush context destruction worker at suspend

2023-08-09 Thread Teres Alexis, Alan Previn
Thanks Rodrigo for reviewing this. On Mon, 2023-08-07 at 13:52 -0400, Vivi, Rodrigo wrote: > On Wed, Aug 02, 2023 at 04:34:59PM -0700, Alan Previn wrote: > > Suspend is not like reset, it can unroll, so we have to properly > > flush pending context-guc-id deregistrations to complete before > > we

Re: [PATCH v1 1/3] drm/i915/guc: Flush context destruction worker at suspend

2023-08-07 Thread Rodrigo Vivi
On Wed, Aug 02, 2023 at 04:34:59PM -0700, Alan Previn wrote: > Suspend is not like reset, it can unroll, so we have to properly > flush pending context-guc-id deregistrations to complete before > we return from suspend calls. But if is 'unrolls' the execution should just continue, no?! In other

[PATCH v1 1/3] drm/i915/guc: Flush context destruction worker at suspend

2023-08-02 Thread Alan Previn
Suspend is not like reset, it can unroll, so we have to properly flush pending context-guc-id deregistrations to complete before we return from suspend calls. Signed-off-by: Alan Previn --- drivers/gpu/drm/i915/gt/intel_gt_pm.c | 6 +-