]> git.karo-electronics.de Git - linux-beck.git/commitdiff
drm/amdgpu: stop disabling irqs when it isn't neccessary
authorChristian König <christian.koenig@amd.com>
Mon, 13 Jun 2016 14:12:43 +0000 (16:12 +0200)
committerAlex Deucher <alexander.deucher@amd.com>
Thu, 7 Jul 2016 19:06:04 +0000 (15:06 -0400)
A regular spin_lock/unlock should do here as well.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drivers/gpu/drm/amd/scheduler/gpu_scheduler.c

index 60f58f76a1ebed4923e7d3094070b35609144ccd..2baedc4e667d843d59d1448b50a156d90045f55a 100644 (file)
@@ -331,17 +331,16 @@ static void amd_sched_job_finish(struct work_struct *work)
        struct amd_sched_job *s_job = container_of(work, struct amd_sched_job,
                                                   finish_work);
        struct amd_gpu_scheduler *sched = s_job->sched;
-       unsigned long flags;
 
        /* remove job from ring_mirror_list */
-       spin_lock_irqsave(&sched->job_list_lock, flags);
+       spin_lock(&sched->job_list_lock);
        list_del_init(&s_job->node);
        if (sched->timeout != MAX_SCHEDULE_TIMEOUT) {
                struct amd_sched_job *next;
 
-               spin_unlock_irqrestore(&sched->job_list_lock, flags);
+               spin_unlock(&sched->job_list_lock);
                cancel_delayed_work_sync(&s_job->work_tdr);
-               spin_lock_irqsave(&sched->job_list_lock, flags);
+               spin_lock(&sched->job_list_lock);
 
                /* queue TDR for next job */
                next = list_first_entry_or_null(&sched->ring_mirror_list,
@@ -350,7 +349,7 @@ static void amd_sched_job_finish(struct work_struct *work)
                if (next)
                        schedule_delayed_work(&next->work_tdr, sched->timeout);
        }
-       spin_unlock_irqrestore(&sched->job_list_lock, flags);
+       spin_unlock(&sched->job_list_lock);
        sched->ops->free_job(s_job);
 }
 
@@ -364,15 +363,14 @@ static void amd_sched_job_finish_cb(struct fence *f, struct fence_cb *cb)
 static void amd_sched_job_begin(struct amd_sched_job *s_job)
 {
        struct amd_gpu_scheduler *sched = s_job->sched;
-       unsigned long flags;
 
-       spin_lock_irqsave(&sched->job_list_lock, flags);
+       spin_lock(&sched->job_list_lock);
        list_add_tail(&s_job->node, &sched->ring_mirror_list);
        if (sched->timeout != MAX_SCHEDULE_TIMEOUT &&
            list_first_entry_or_null(&sched->ring_mirror_list,
                                     struct amd_sched_job, node) == s_job)
                schedule_delayed_work(&s_job->work_tdr, sched->timeout);
-       spin_unlock_irqrestore(&sched->job_list_lock, flags);
+       spin_unlock(&sched->job_list_lock);
 }
 
 static void amd_sched_job_timedout(struct work_struct *work)