el.org ;
matthew.br...@intel.com
Subject: Re: [PATCH v2] drm/scheduler: Fix sched hang when killing app with
dependent jobs
On Tue, 2025-07-15 at 14:32 +0200, Christian König wrote:
> On 15.07.25 14:20, Philipp Stanner wrote:
> > On Tue, 2025-07-15 at 12:52 +0200, Christian König w
ion is realistic
> > here, though?
>
> No, absolutely not. It was just a micro optimization done long ago.
>
> On the other hand with the scheduler code base I'm not sure of
> anything any more.
"In my subsystem you have to run as fast as you can just to remain at
the sa
. It was just a micro optimization done long ago.
On the other hand with the scheduler code base I'm not sure of anything any
more.
Regards,
Christian.
>
>
> P.
>
>>
>> Christian.
>>
>>> I think that was the piece of the puzzle
>>> I was miss
formance regression is realistic
here, though?
P.
>
> Christian.
>
> > I think that was the piece of the puzzle
> > I was missing.
> >
> > I / DRM tools will also include a link to this thread, so I think
> > that
> > will then be sufficient.
> &g
the first dependency, not for all of them?
>>
>> Oh boy. That code base. We(tm) should look into pimping that up. Also
>> lacks documentation everywhere.
>>
>>
>> Anyways. If this solves a bug for you the patch looks fine (i.e., not
>> potentially harm
gt; From: Philipp Stanner
> Sent: Tuesday, July 15, 2025 17:12
> To: cao, lin ; Koenig, Christian ;
> pha...@kernel.org ; dri-devel@lists.freedesktop.org
>
> Cc: Yin, ZhenGuo (Chris) ; Deng, Emily
> ; d...@kernel.org ;
> matthew.br...@intel.com
> Subject: Re: [PATCH v2]
drm_sched_wakeup() to after drm_sched_fence_scheduled()
>
>
> Thanks,
> Lin
>
>
> From: Koenig, Christian
> Sent: Monday, July 14, 2025 21:39
> To: pha...@kernel.org ; cao, lin ;
> dri-devel@lists.freedesktop.org
> Cc: Yin, ZhenGuo (Chris) ; Deng, Emily
> ; d.
ence_scheduled()
>
>
> Thanks,
> Lin
>
>
> From: Koenig, Christian
> Sent: Monday, July 14, 2025 21:39
> To: pha...@kernel.org ; cao, lin ;
> dri-devel@lists.freedesktop.org
> Cc: Yin, ZhenGuo (Chris) ; Deng, Emily
> ; d...@kernel.org ;
> matthew.br...@in
, lin ;
dri-devel@lists.freedesktop.org
Cc: Yin, ZhenGuo (Chris) ; Deng, Emily
; d...@kernel.org ;
matthew.br...@intel.com
Subject: Re: [PATCH v2] drm/scheduler: Fix sched hang when killing app with
dependent jobs
On 14.07.25 15:27, Philipp Stanner wrote:
> On Mon, 2025-07-14 at 15:
On 14.07.25 15:27, Philipp Stanner wrote:
> On Mon, 2025-07-14 at 15:08 +0200, Christian König wrote:
>>
>>
>> On 14.07.25 14:46, Philipp Stanner wrote:
>>> regarding the patch subject: the prefix we use for the scheduler
>>> is:
>>> drm/sched:
>>>
>>>
>>> On Mon, 2025-07-14 at 14:23 +0800, Lin.Cao
On Mon, 2025-07-14 at 15:08 +0200, Christian König wrote:
>
>
> On 14.07.25 14:46, Philipp Stanner wrote:
> > regarding the patch subject: the prefix we use for the scheduler
> > is:
> > drm/sched:
> >
> >
> > On Mon, 2025-07-14 at 14:23 +0800, Lin.Cao wrote:
> >
> > > When Application A submi
On 14.07.25 14:46, Philipp Stanner wrote:
> regarding the patch subject: the prefix we use for the scheduler is:
> drm/sched:
>
>
> On Mon, 2025-07-14 at 14:23 +0800, Lin.Cao wrote:
>
>> When Application A submits jobs (a1, a2, a3) and application B submits
>
> s/Application/application
>
>
regarding the patch subject: the prefix we use for the scheduler is:
drm/sched:
On Mon, 2025-07-14 at 14:23 +0800, Lin.Cao wrote:
> When Application A submits jobs (a1, a2, a3) and application B submits
s/Application/application
> job b1 with a dependency on a2's scheduler fence, killing appli
On 14.07.25 08:23, Lin.Cao wrote:
> When Application A submits jobs (a1, a2, a3) and application B submits
> job b1 with a dependency on a2's scheduler fence, killing application A
> before run_job(a1) causes drm_sched_entity_kill_jobs_work() to force
> signal all jobs sequentially. However, due to
When Application A submits jobs (a1, a2, a3) and application B submits
job b1 with a dependency on a2's scheduler fence, killing application A
before run_job(a1) causes drm_sched_entity_kill_jobs_work() to force
signal all jobs sequentially. However, due to missing work_run_job or
work_free_job in
15 matches
Mail list logo
