drm/amdgpu: stop all rings before doing gpu recover

found recover_vram_from_shadow sometimes get executed
in paralle with SDMA scheduler, should stop all
schedulers before doing gpu reset/recover

Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Tested-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

1 files changed, 15 insertions, 25 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 41244858df64..64bd30075951 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -2648,22 +2648,23 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 
         /* block TTM */
         resched = ttm_bo_lock_delayed_workqueue(&adev->mman.bdev);
+
         /* store modesetting */
         if (amdgpu_device_has_dc_support(adev))
                 state = drm_atomic_helper_suspend(adev->ddev);
 
-        /* block scheduler */
+        /* block all schedulers and reset given job's ring */
         for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
                 struct amdgpu_ring *ring = adev->rings[i];
 
                 if (!ring || !ring->sched.thread)
                         continue;
 
-                /* only focus on the ring hit timeout if &job not NULL */
+                kthread_park(ring->sched.thread);
+
                 if (job && job->ring->idx != i)
                         continue;
 
-                kthread_park(ring->sched.thread);
                 drm_sched_hw_job_reset(&ring->sched, &job->base);
 
                 /* after all hw jobs are reset, hw fence is meaningless, so force_completion */
@@ -2706,33 +2707,22 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
                         }
                         dma_fence_put(fence);
                 }
+        }
 
-                for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
-                        struct amdgpu_ring *ring = adev->rings[i];
-
-                        if (!ring || !ring->sched.thread)
-                                continue;
+        for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
+                struct amdgpu_ring *ring = adev->rings[i];
 
-                        /* only focus on the ring hit timeout if &job not NULL */
-                        if (job && job->ring->idx != i)
-                                continue;
+                if (!ring || !ring->sched.thread)
+                        continue;
 
+                /* only need recovery sched of the given job's ring
+                 * or all rings (in the case @job is NULL)
+                 * after above amdgpu_reset accomplished
+                 */
+                if ((!job || job->ring->idx == i) && !r)
                         drm_sched_job_recovery(&ring->sched);
-                        kthread_unpark(ring->sched.thread);
-                }
-        } else {
-                for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
-                        struct amdgpu_ring *ring = adev->rings[i];
 
-                        if (!ring || !ring->sched.thread)
-                                continue;
-
-                        /* only focus on the ring hit timeout if &job not NULL */
-                        if (job && job->ring->idx != i)
-                                continue;
-
-                        kthread_unpark(adev->rings[i]->sched.thread);
-                }
+                kthread_unpark(ring->sched.thread);
         }
 
         if (amdgpu_device_has_dc_support(adev)) {