drm/amdgpu/vcn: finish delay work before release resources

alexdeucher · alexdeucher · commit 622b2a0ab647 · 2019-11-13T15:29:42.000-05:00
flush/cancel delayed works before doing finalization
to avoid concurrently requests.

Reviewed-by: Christian König &lt;christian.koenig@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
@@ -193,6 +193,8 @@ int amdgpu_vcn_sw_fini(struct amdgpu_device *adev)
 {
 	int i, j;
 
+	cancel_delayed_work_sync(&adev->vcn.idle_work);
+
 	if (adev->vcn.indirect_sram) {
 		amdgpu_bo_free_kernel(&adev->vcn.dpg_sram_bo,
 				      &adev->vcn.dpg_sram_gpu_addr,

Original file line number	Diff line number	Diff line change
`@@ -193,6 +193,8 @@ int amdgpu_vcn_sw_fini(struct amdgpu_device *adev)`
`193`	`193`	`{`
`194`	`194`	`int i, j;`
`195`	`195`
	`196`	`+ cancel_delayed_work_sync(&adev->vcn.idle_work);`
	`197`	`+`
`196`	`198`	`if (adev->vcn.indirect_sram) {`
`197`	`199`	`amdgpu_bo_free_kernel(&adev->vcn.dpg_sram_bo,`
`198`	`200`	`&adev->vcn.dpg_sram_gpu_addr,`