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>

1 files changed, 7 insertions, 0 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.c
index 189afb872775..c22767a75348 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.c
+++ b/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);