drm/amdgpu: handle eviction fence race

The eviction process can get into a race condition between the eviction
fence suspend work (which replaces the old fence with new) and kms_close
(which destroys the fence and doesn't expect a new one).

This patch:
- adds a flag to indicate that fd is closing, so fence replacement is
  not required (evf_mgr->fd_closing)
- adds a flush_work() during the ev_fence_destroy routine

V2: Addressed review comments from Christian:
    - Do not use mutex to sync
    - Use flush_work and wait for suspend_work to be done

V3: Fixed state machine for queue->active, which adds into race between
    suspend/resume and queue ops

Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Shashank Sharma <shashank.sharma@amd.com>
Signed-off-by: Arvind Yadav <arvind.yadav@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

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

@@ -117,6 +117,10 @@ amdgpu_eviction_fence_suspend_worker(struct work_struct *work)
 	/* Signal old eviction fence */
 	amdgpu_eviction_fence_signal(evf_mgr);
 
+	/* Do not replace eviction fence is fd is getting closed */
+	if (evf_mgr->fd_closing)
+		return;
+
 	/* Prepare the objects to replace eviction fence */
 	drm_exec_init(&exec, DRM_EXEC_IGNORE_DUPLICATES, 0);
 	drm_exec_until_all_locked(&exec) {
@@ -199,6 +203,9 @@ void amdgpu_eviction_fence_destroy(struct amdgpu_eviction_fence_mgr *evf_mgr)
 {
 	struct amdgpu_eviction_fence *ev_fence;
 
+	/* Wait for any pending work to execute */
+	flush_delayed_work(&evf_mgr->suspend_work);
+
 	spin_lock(&evf_mgr->ev_fence_lock);
 	ev_fence = evf_mgr->ev_fence;
 	spin_unlock(&evf_mgr->ev_fence_lock);

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

@@ -1490,10 +1490,12 @@ void amdgpu_driver_postclose_kms(struct drm_device *dev,
 		amdgpu_bo_unreserve(pd);
 	}
 
+	fpriv->evf_mgr.fd_closing = true;
+	amdgpu_userq_mgr_fini(&fpriv->userq_mgr);
 	amdgpu_eviction_fence_destroy(&fpriv->evf_mgr);
+
 	amdgpu_ctx_mgr_fini(&fpriv->ctx_mgr);
 	amdgpu_vm_fini(adev, &fpriv->vm);
-	amdgpu_userq_mgr_fini(&fpriv->userq_mgr);
 
 	if (pasid)
 		amdgpu_pasid_free_delayed(pd->tbo.base.resv, pasid);

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

@@ -614,9 +614,10 @@ void amdgpu_userq_mgr_fini(struct amdgpu_userq_mgr *userq_mgr)
 
 	cancel_delayed_work(&userq_mgr->resume_work);
 
+	mutex_lock(&userq_mgr->userq_mutex);
 	idr_for_each_entry(&userq_mgr->userq_idr, queue, queue_id)
 		amdgpu_userqueue_cleanup(userq_mgr, queue, queue_id);
-
 	idr_destroy(&userq_mgr->userq_idr);
+	mutex_unlock(&userq_mgr->userq_mutex);
 	mutex_destroy(&userq_mgr->userq_mutex);
 }

drivers/gpu/drm/amd/amdgpu/mes_v11_0_userqueue.c

@@ -139,6 +139,7 @@ static int mes_v11_0_userq_map(struct amdgpu_userq_mgr *uq_mgr,
 		return r;
 	}
 
+	queue->queue_active = true;
 	DRM_DEBUG_DRIVER("Queue (doorbell:%d) mapped successfully\n", userq_props->doorbell_index);
 	return 0;
 }
@@ -160,6 +161,7 @@ static void mes_v11_0_userq_unmap(struct amdgpu_userq_mgr *uq_mgr,
 	amdgpu_mes_unlock(&adev->mes);
 	if (r)
 		DRM_ERROR("Failed to unmap queue in HW, err (%d)\n", r);
+	queue->queue_active = false;
 }
 
 static int mes_v11_0_userq_create_ctx_space(struct amdgpu_userq_mgr *uq_mgr,
@@ -331,7 +333,6 @@ static int mes_v11_0_userq_mqd_create(struct amdgpu_userq_mgr *uq_mgr,
 		goto free_ctx;
 	}
 
-	queue->queue_active = true;
 	return 0;
 
 free_ctx:
@@ -350,12 +351,12 @@ static void
 mes_v11_0_userq_mqd_destroy(struct amdgpu_userq_mgr *uq_mgr,
 			    struct amdgpu_usermode_queue *queue)
 {
-	mes_v11_0_userq_unmap(uq_mgr, queue);
-	amdgpu_bo_unref(&queue->wptr_obj.obj);
+	if (queue->queue_active)
+		mes_v11_0_userq_unmap(uq_mgr, queue);
+
 	amdgpu_userqueue_destroy_object(uq_mgr, &queue->fw_obj);
 	kfree(queue->userq_prop);
 	amdgpu_userqueue_destroy_object(uq_mgr, &queue->mqd);
-	queue->queue_active = false;
 }
 
 static int mes_v11_0_userq_suspend(struct amdgpu_userq_mgr *uq_mgr,