1 files changed, 63 insertions, 53 deletions
diff --git a/drivers/gpu/drm/v3d/v3d_sched.c b/drivers/gpu/drm/v3d/v3d_sched.c
index 99ac4995b5a1..34c42d6e12cd 100644
--- a/drivers/gpu/drm/v3d/v3d_sched.c
+++ b/drivers/gpu/drm/v3d/v3d_sched.c
@@ -5,16 +5,16 @@
  * DOC: Broadcom V3D scheduling
  *
  * The shared DRM GPU scheduler is used to coordinate submitting jobs
- * to the hardware.  Each DRM fd (roughly a client process) gets its
- * own scheduler entity, which will process jobs in order.  The GPU
- * scheduler will round-robin between clients to submit the next job.
+ * to the hardware. Each DRM fd (roughly a client process) gets its
+ * own scheduler entity, which will process jobs in order. The GPU
+ * scheduler will schedule the clients with a FIFO scheduling algorithm.
  *
  * For simplicity, and in order to keep latency low for interactive
  * jobs when bulk background jobs are queued up, we submit a new job
  * to the HW only when it has completed the last one, instead of
- * filling up the CT[01]Q FIFOs with jobs.  Similarly, we use
- * drm_sched_job_add_dependency() to manage the dependency between bin and
- * render, instead of having the clients submit jobs using the HW's
+ * filling up the CT[01]Q FIFOs with jobs. Similarly, we use
+ * `drm_sched_job_add_dependency()` to manage the dependency between bin
+ * and render, instead of having the clients submit jobs using the HW's
  * semaphores to interlock between them.
  */
 
@@ -120,11 +120,19 @@ v3d_cpu_job_free(struct drm_sched_job *sched_job)
 static void
 v3d_switch_perfmon(struct v3d_dev *v3d, struct v3d_job *job)
 {
-	if (job->perfmon != v3d->active_perfmon)
+	struct v3d_perfmon *perfmon = v3d->global_perfmon;
+
+	if (!perfmon)
+		perfmon = job->perfmon;
+
+	if (perfmon == v3d->active_perfmon)
+		return;
+
+	if (perfmon != v3d->active_perfmon)
 		v3d_perfmon_stop(v3d, v3d->active_perfmon, true);
 
-	if (job->perfmon && v3d->active_perfmon != job->perfmon)
-		v3d_perfmon_start(v3d, job->perfmon);
+	if (perfmon && v3d->active_perfmon != perfmon)
+		v3d_perfmon_start(v3d, perfmon);
 }
 
 static void
@@ -218,8 +226,12 @@ static struct dma_fence *v3d_bin_job_run(struct drm_sched_job *sched_job)
 	struct dma_fence *fence;
 	unsigned long irqflags;
 
-	if (unlikely(job->base.base.s_fence->finished.error))
+	if (unlikely(job->base.base.s_fence->finished.error)) {
+		spin_lock_irqsave(&v3d->job_lock, irqflags);
+		v3d->bin_job = NULL;
+		spin_unlock_irqrestore(&v3d->job_lock, irqflags);
 		return NULL;
+	}
 
 	/* Lock required around bin_job update vs
 	 * v3d_overflow_mem_work().
@@ -273,8 +285,10 @@ static struct dma_fence *v3d_render_job_run(struct drm_sched_job *sched_job)
 	struct drm_device *dev = &v3d->drm;
 	struct dma_fence *fence;
 
-	if (unlikely(job->base.base.s_fence->finished.error))
+	if (unlikely(job->base.base.s_fence->finished.error)) {
+		v3d->render_job = NULL;
 		return NULL;
+	}
 
 	v3d->render_job = job;
 
@@ -319,11 +333,17 @@ v3d_tfu_job_run(struct drm_sched_job *sched_job)
 	struct drm_device *dev = &v3d->drm;
 	struct dma_fence *fence;
 
+	if (unlikely(job->base.base.s_fence->finished.error)) {
+		v3d->tfu_job = NULL;
+		return NULL;
+	}
+
+	v3d->tfu_job = job;
+
 	fence = v3d_fence_create(v3d, V3D_TFU);
 	if (IS_ERR(fence))
 		return NULL;
 
-	v3d->tfu_job = job;
 	if (job->base.irq_fence)
 		dma_fence_put(job->base.irq_fence);
 	job->base.irq_fence = dma_fence_get(fence);
@@ -361,6 +381,11 @@ v3d_csd_job_run(struct drm_sched_job *sched_job)
 	struct dma_fence *fence;
 	int i, csd_cfg0_reg;
 
+	if (unlikely(job->base.base.s_fence->finished.error)) {
+		v3d->csd_job = NULL;
+		return NULL;
+	}
+
 	v3d->csd_job = job;
 
 	v3d_invalidate_caches(v3d);
@@ -648,8 +673,6 @@ v3d_cpu_job_run(struct drm_sched_job *sched_job)
 	struct v3d_cpu_job *job = to_cpu_job(sched_job);
 	struct v3d_dev *v3d = job->base.v3d;
 
-	v3d->cpu_job = job;
-
 	if (job->job_type >= ARRAY_SIZE(cpu_job_function)) {
 		DRM_DEBUG_DRIVER("Unknown CPU job: %d\n", job->job_type);
 		return NULL;
@@ -814,67 +837,54 @@ static const struct drm_sched_backend_ops v3d_cpu_sched_ops = {
 	.free_job = v3d_cpu_job_free
 };
 
+static int
+v3d_queue_sched_init(struct v3d_dev *v3d, const struct drm_sched_backend_ops *ops,
+		     enum v3d_queue queue, const char *name)
+{
+	struct drm_sched_init_args args = {
+		.num_rqs = DRM_SCHED_PRIORITY_COUNT,
+		.credit_limit = 1,
+		.timeout = msecs_to_jiffies(500),
+		.dev = v3d->drm.dev,
+	};
+
+	args.ops = ops;
+	args.name = name;
+
+	return drm_sched_init(&v3d->queue[queue].sched, &args);
+}
+
 int
 v3d_sched_init(struct v3d_dev *v3d)
 {
-	int hw_jobs_limit = 1;
-	int job_hang_limit = 0;
-	int hang_limit_ms = 500;
 	int ret;
 
-	ret = drm_sched_init(&v3d->queue[V3D_BIN].sched,
-			     &v3d_bin_sched_ops, NULL,
-			     DRM_SCHED_PRIORITY_COUNT,
-			     hw_jobs_limit, job_hang_limit,
-			     msecs_to_jiffies(hang_limit_ms), NULL,
-			     NULL, "v3d_bin", v3d->drm.dev);
+	ret = v3d_queue_sched_init(v3d, &v3d_bin_sched_ops, V3D_BIN, "v3d_bin");
 	if (ret)
 		return ret;
 
-	ret = drm_sched_init(&v3d->queue[V3D_RENDER].sched,
-			     &v3d_render_sched_ops, NULL,
-			     DRM_SCHED_PRIORITY_COUNT,
-			     hw_jobs_limit, job_hang_limit,
-			     msecs_to_jiffies(hang_limit_ms), NULL,
-			     NULL, "v3d_render", v3d->drm.dev);
+	ret = v3d_queue_sched_init(v3d, &v3d_render_sched_ops, V3D_RENDER,
+				   "v3d_render");
 	if (ret)
 		goto fail;
 
-	ret = drm_sched_init(&v3d->queue[V3D_TFU].sched,
-			     &v3d_tfu_sched_ops, NULL,
-			     DRM_SCHED_PRIORITY_COUNT,
-			     hw_jobs_limit, job_hang_limit,
-			     msecs_to_jiffies(hang_limit_ms), NULL,
-			     NULL, "v3d_tfu", v3d->drm.dev);
+	ret = v3d_queue_sched_init(v3d, &v3d_tfu_sched_ops, V3D_TFU, "v3d_tfu");
 	if (ret)
 		goto fail;
 
 	if (v3d_has_csd(v3d)) {
-		ret = drm_sched_init(&v3d->queue[V3D_CSD].sched,
-				     &v3d_csd_sched_ops, NULL,
-				     DRM_SCHED_PRIORITY_COUNT,
-				     hw_jobs_limit, job_hang_limit,
-				     msecs_to_jiffies(hang_limit_ms), NULL,
-				     NULL, "v3d_csd", v3d->drm.dev);
+		ret = v3d_queue_sched_init(v3d, &v3d_csd_sched_ops, V3D_CSD,
+					   "v3d_csd");
 		if (ret)
 			goto fail;
 
-		ret = drm_sched_init(&v3d->queue[V3D_CACHE_CLEAN].sched,
-				     &v3d_cache_clean_sched_ops, NULL,
-				     DRM_SCHED_PRIORITY_COUNT,
-				     hw_jobs_limit, job_hang_limit,
-				     msecs_to_jiffies(hang_limit_ms), NULL,
-				     NULL, "v3d_cache_clean", v3d->drm.dev);
+		ret = v3d_queue_sched_init(v3d, &v3d_cache_clean_sched_ops,
+					   V3D_CACHE_CLEAN, "v3d_cache_clean");
 		if (ret)
 			goto fail;
 	}
 
-	ret = drm_sched_init(&v3d->queue[V3D_CPU].sched,
-			     &v3d_cpu_sched_ops, NULL,
-			     DRM_SCHED_PRIORITY_COUNT,
-			     1, job_hang_limit,
-			     msecs_to_jiffies(hang_limit_ms), NULL,
-			     NULL, "v3d_cpu", v3d->drm.dev);
+	ret = v3d_queue_sched_init(v3d, &v3d_cpu_sched_ops, V3D_CPU, "v3d_cpu");
 	if (ret)
 		goto fail;