drm/amdgpu: add amdgpu_error_* debugfs file

This allows us to insert some error codes into the bottom of the pipeline
on an engine.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Luben Tuikov <luben.tuikov@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Author: ChristianKoenigAMD

drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c

@@ -691,6 +691,30 @@ void amdgpu_fence_driver_clear_job_fences(struct amdgpu_ring *ring)
 	}
 }
 
+/**
+ * amdgpu_fence_driver_set_error - set error code on fences
+ * @ring: the ring which contains the fences
+ * @error: the error code to set
+ *
+ * Set an error code to all the fences pending on the ring.
+ */
+void amdgpu_fence_driver_set_error(struct amdgpu_ring *ring, int error)
+{
+	struct amdgpu_fence_driver *drv = &ring->fence_drv;
+	unsigned long flags;
+
+	spin_lock_irqsave(&drv->lock, flags);
+	for (unsigned int i = 0; i <= drv->num_fences_mask; ++i) {
+		struct dma_fence *fence;
+
+		fence = rcu_dereference_protected(drv->fences[i],
+						  lockdep_is_held(&drv->lock));
+		if (fence && !dma_fence_is_signaled_locked(fence))
+			dma_fence_set_error(fence, error);
+	}
+	spin_unlock_irqrestore(&drv->lock, flags);
+}
+
 /**
  * amdgpu_fence_driver_force_completion - force signal latest fence of ring
  *

drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c

@@ -561,6 +561,17 @@ static const struct file_operations amdgpu_debugfs_mqd_fops = {
 	.llseek = default_llseek
 };
 
+static int amdgpu_debugfs_ring_error(void *data, u64 val)
+{
+	struct amdgpu_ring *ring = data;
+
+	amdgpu_fence_driver_set_error(ring, val);
+	return 0;
+}
+
+DEFINE_DEBUGFS_ATTRIBUTE_SIGNED(amdgpu_debugfs_error_fops, NULL,
+				amdgpu_debugfs_ring_error, "%lld\n");
+
 #endif
 
 void amdgpu_debugfs_ring_init(struct amdgpu_device *adev,
@@ -582,6 +593,11 @@ void amdgpu_debugfs_ring_init(struct amdgpu_device *adev,
 					 &amdgpu_debugfs_mqd_fops,
 					 ring->mqd_size);
 	}
+
+	sprintf(name, "amdgpu_error_%s", ring->name);
+	debugfs_create_file(name, 0200, root, ring,
+			    &amdgpu_debugfs_error_fops);
+
 #endif
 }
 

drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h

@@ -126,6 +126,7 @@ struct amdgpu_fence_driver {
 extern const struct drm_sched_backend_ops amdgpu_sched_ops;
 
 void amdgpu_fence_driver_clear_job_fences(struct amdgpu_ring *ring);
+void amdgpu_fence_driver_set_error(struct amdgpu_ring *ring, int error);
 void amdgpu_fence_driver_force_completion(struct amdgpu_ring *ring);
 
 int amdgpu_fence_driver_init_ring(struct amdgpu_ring *ring);