drm/xe: Take job list lock in xe_sched_add_pending_job

[ Upstream commit ed931fb40e353586f26c3327813d142f782f5f78 ]

A fragile micro optimization in xe_sched_add_pending_job relied on both
the GPU scheduler being stopped and fence signaling stopped to safely
add a job to the pending list without the job list lock in
xe_sched_add_pending_job. Remove this optimization and just take the job
list lock.

Fixes: 7ddb9403dd74 ("drm/xe: Sample ctx timestamp to determine if jobs have timed out")
Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20241003001657.3517883-2-matthew.brost@intel.com
(cherry picked from commit 90521df5fc43980e4575bd8c5b1cb62afe1a9f5f)
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

diff --git a/drivers/gpu/drm/xe/xe_gpu_scheduler.h b/drivers/gpu/drm/xe/xe_gpu_scheduler.h
index 6aac7fe686735aa24293dbbcfdfef5de4266147a..6bdd0a5b36122665445136b19ba0edc2869c33bb 100644 (file)
--- a/drivers/gpu/drm/xe/xe_gpu_scheduler.h
+++ b/drivers/gpu/drm/xe/xe_gpu_scheduler.h
@@ -51,7 +51,9 @@ xe_sched_invalidate_job(struct xe_sched_job *job, int threshold)
 static inline void xe_sched_add_pending_job(struct xe_gpu_scheduler *sched,
                                            struct xe_sched_job *job)
 {
+       spin_lock(&sched->base.job_list_lock);
        list_add(&job->drm.list, &sched->base.pending_list);
+       spin_unlock(&sched->base.job_list_lock);
 }
 
 static inline