3. killed this process, will call to drm_sched_entity_cleanup, will
register drm_sched_entity_kill_jobs_cb to jobB->finished, but jobB isn't
submitted at all.
I think that point is your misunderstanding. Killing process won’t stop jobB to
be submitted.
—
Sincerely Yours,
Pixel
O
That change is indeed fixing a problem. When drm_sched_job_recovery() is
called s_job->entity should already be NULL.
And even when the process is killed we should still either re-submit the
jobs or set the error.
BTW: There is another bug in that function: guilty_context needs to be
kept al
NAK on it.
First of all, without this patch, Does it cause any issue?
second,
entity->last_scheduled present the last submiting job.
Your this change will break this meaning and don't work, e.g.
1. mirror list has jobA and jobB, assuming they are belonged to same
entity, then the entity->las
Reviewed-by : Monk Liu
-Original Message-
From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf Of Pixel
Ding
Sent: 2018年4月25日 16:40
To: amd-gfx@lists.freedesktop.org
Cc: Ding, Pixel
Subject: [PATCH] drm/scheduler: don't update last scheduled fence in TDR
The current s
Hi Monk,
Please review it. Thanks.
—
Sincerely Yours,
Pixel
On 2018/4/25, 4:39 PM, "Pixel Ding" wrote:
The current sequence in scheduler thread is:
1. update last sched fence
2. job begin (adding to mirror list)
3. job finish (remove from mirror list)
4. back to 1