drm/amdgpu: Fix vce work queue was not cancelled when suspend
authorRex Zhu <Rex.Zhu@amd.com>
Thu, 27 Sep 2018 12:48:39 +0000 (20:48 +0800)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Sat, 13 Oct 2018 07:27:23 +0000 (09:27 +0200)
commit 61ea6f5831974ebd1a57baffd7cc30600a2e26fc upstream.

The vce cancel_delayed_work_sync never be called.
driver call the function in error path.

This caused the A+A suspend hang when runtime pm enebled.
As we will visit the smu in the idle queue. this will cause
smu hang because the dgpu has been suspend, and the dgpu also
will be waked up. As the smu has been hang, so the dgpu resume
will failed.

Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Feifei Xu <Feifei.Xu@amd.com>
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c
drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c

index 9fc3d38..fb36425 100644 (file)
@@ -231,6 +231,8 @@ int amdgpu_vce_suspend(struct amdgpu_device *adev)
 {
        int i;
 
+       cancel_delayed_work_sync(&adev->vce.idle_work);
+
        if (adev->vce.vcpu_bo == NULL)
                return 0;
 
@@ -241,7 +243,6 @@ int amdgpu_vce_suspend(struct amdgpu_device *adev)
        if (i == AMDGPU_MAX_VCE_HANDLES)
                return 0;
 
-       cancel_delayed_work_sync(&adev->vce.idle_work);
        /* TODO: suspending running encoding sessions isn't supported */
        return -EINVAL;
 }
index 1612d8a..fca1b10 100644 (file)
@@ -155,11 +155,11 @@ int amdgpu_vcn_suspend(struct amdgpu_device *adev)
        unsigned size;
        void *ptr;
 
+       cancel_delayed_work_sync(&adev->vcn.idle_work);
+
        if (adev->vcn.vcpu_bo == NULL)
                return 0;
 
-       cancel_delayed_work_sync(&adev->vcn.idle_work);
-
        size = amdgpu_bo_size(adev->vcn.vcpu_bo);
        ptr = adev->vcn.cpu_addr;