[AMD Official Use Only - Internal Distribution Only]

That said, it would be easier for me to merge through the AMD tree since a relatively big AMD feature depends on it.  Not sure how much conflict potential there is if this goes through the AMD tree.

Alex

________________________________
From: amd-gfx <amd-gfx-bounces@lists.freedesktop.org> on behalf of Deucher, Alexander <Alexander.Deucher@amd.com>
Sent: Monday, April 26, 2021 3:24 PM
To: Christian König <ckoenig.leichtzumerken@gmail.com>
Cc: amd-gfx list <amd-gfx@lists.freedesktop.org>; Sun, Roy <Roy.Sun@amd.com>; Nieto, David M <David.Nieto@amd.com>
Subject: Re: [PATCH 1/2] drm/scheduler: Change scheduled fence track


[AMD Official Use Only - Internal Distribution Only]


[AMD Official Use Only - Internal Distribution Only]

No objections from me.

Thanks!

Alex

________________________________
From: Christian König <ckoenig.leichtzumerken@gmail.com>
Sent: Monday, April 26, 2021 2:49 AM
To: Deucher, Alexander <Alexander.Deucher@amd.com>
Cc: Nieto, David M <David.Nieto@amd.com>; Sun, Roy <Roy.Sun@amd.com>; amd-gfx list <amd-gfx@lists.freedesktop.org>
Subject: Re: [PATCH 1/2] drm/scheduler: Change scheduled fence track

Hey Alex,

any objections that we merge those two patches through drm-misc-next?

Thanks,
Christian.

Am 26.04.21 um 08:27 schrieb Roy Sun:
> Update the timestamp of scheduled fence on HW
> completion of the previous fences
>
> This allow more accurate tracking of the fence
> execution in HW
>
> Signed-off-by: David M Nieto <david.nieto@amd.com>
> Signed-off-by: Roy Sun <Roy.Sun@amd.com>
> ---
>   drivers/gpu/drm/scheduler/sched_main.c | 12 ++++++++++--
>   1 file changed, 10 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpu/drm/scheduler/sched_main.c b/drivers/gpu/drm/scheduler/sched_main.c
> index 92d8de24d0a1..f8e39ab0c41b 100644
> --- a/drivers/gpu/drm/scheduler/sched_main.c
> +++ b/drivers/gpu/drm/scheduler/sched_main.c
> @@ -515,7 +515,7 @@ void drm_sched_resubmit_jobs(struct drm_gpu_scheduler *sched)
>   EXPORT_SYMBOL(drm_sched_resubmit_jobs);
>
>   /**
> - * drm_sched_resubmit_jobs_ext - helper to relunch certain number of jobs from mirror ring list
> + * drm_sched_resubmit_jobs_ext - helper to relaunch certain number of jobs from pending list
>    *
>    * @sched: scheduler instance
>    * @max: job numbers to relaunch
> @@ -671,7 +671,7 @@ drm_sched_select_entity(struct drm_gpu_scheduler *sched)
>   static struct drm_sched_job *
>   drm_sched_get_cleanup_job(struct drm_gpu_scheduler *sched)
>   {
> -     struct drm_sched_job *job;
> +     struct drm_sched_job *job, *next;
>
>        /*
>         * Don't destroy jobs while the timeout worker is running  OR thread
> @@ -690,6 +690,14 @@ drm_sched_get_cleanup_job(struct drm_gpu_scheduler *sched)
>        if (job && dma_fence_is_signaled(&job->s_fence->finished)) {
>                /* remove job from pending_list */
>                list_del_init(&job->list);
> We just need to record the scheduled time of the next job. So we
> need not to check the rest job.
> +             /* account for the next fence in the queue */
> +             next = list_first_entry_or_null(&sched->pending_list,
> +                             struct drm_sched_job, list);
> +             if (next && test_bit(DMA_FENCE_FLAG_TIMESTAMP_BIT,
> +                     &job->s_fence->finished.flags)) {
> +                     next->s_fence->scheduled.timestamp =
> +                             job->s_fence->finished.timestamp;
> +             }
>        } else {
>                job = NULL;
>                /* queue timeout for next job */