[PATCH 2/3] drm/sched: Keep s_fence->parent pointer

[Christian König]

Am 08.04.19 um 18:08 schrieb Andrey Grodzovsky:
> For later driver's reference to see if the fence is signaled.
>
> Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky at amd.com>
> ---
>   drivers/gpu/drm/scheduler/sched_main.c | 11 ++++++++---
>   1 file changed, 8 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/gpu/drm/scheduler/sched_main.c b/drivers/gpu/drm/scheduler/sched_main.c
> index c215cde..5bb4368 100644
> --- a/drivers/gpu/drm/scheduler/sched_main.c
> +++ b/drivers/gpu/drm/scheduler/sched_main.c
> @@ -191,8 +191,6 @@ EXPORT_SYMBOL(drm_sched_dependency_optimized);
>    */
>   static void drm_sched_start_timeout(struct drm_gpu_scheduler *sched)
>   {
> -	unsigned long flags;
> -

I think this actually belongs into patch #1.

>   	if (sched->timeout != MAX_SCHEDULE_TIMEOUT &&
>   	    !list_empty(&sched->ring_mirror_list))
>   		schedule_delayed_work(&sched->work_tdr, sched->timeout);
> @@ -371,7 +369,6 @@ void drm_sched_stop(struct drm_gpu_scheduler *sched, struct drm_sched_job *bad)
>   		    dma_fence_remove_callback(s_job->s_fence->parent,
>   					      &s_job->cb)) {
>   			dma_fence_put(s_job->s_fence->parent);
> -			s_job->s_fence->parent = NULL;

How about also moving the dma_fence_put() into 
drm_sched_resubmit_jobs(), right before we re-assign s_job->s_fence->parent?

I think that would be cleaner, but not sure if that wouldn't have any 
ugly side effects.

Christian.

>   			atomic_dec(&sched->hw_rq_count);
>   		} else {
>   			/*
> @@ -398,6 +395,14 @@ void drm_sched_stop(struct drm_gpu_scheduler *sched, struct drm_sched_job *bad)
>   				sched->ops->free_job(s_job);
>   		}
>   	}
> +
> +	/*
> +	 * Stop pending timer in flight as we rearm it in  drm_sched_start. This
> +	 * avoids the pending timeout work in progress to fire right away after
> +	 * this TDR finished and before the newly restarted jobs had a
> +	 * chance to complete.
> +	 */
> +	cancel_delayed_work(&sched->work_tdr);
>   }
>   
>   EXPORT_SYMBOL(drm_sched_stop);