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 <[email protected]>
Cc: Christian König <[email protected]>
Reviewed-by: Christian König <[email protected]>
Signed-off-by: Shashank Sharma <[email protected]>
Signed-off-by: Arvind Yadav <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>

1 files changed, 3 insertions, 1 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
index 700442584f97..f8370da08038 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
+++ b/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);