drm/amdgpu: fix a potential GPU hang on cyan skillfish

Lang Yu · alexdeucher · commit bca52455a3c0 · 2022-02-02T18:35:00.000-05:00
We observed a GPU hang when querying GMC CG state(i.e.,
cat amdgpu_pm_info) on cyan skillfish. Acctually, cyan
skillfish doesn't support any CG features.

Just prevent it from accessing GMC CG registers.

Signed-off-by: Lang Yu &lt;Lang.Yu@amd.com&gt;
Reviewed-by: Lijo Lazar &lt;lijo.lazar@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
Cc: stable@vger.kernel.org
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
@@ -1140,6 +1140,9 @@ static void gmc_v10_0_get_clockgating_state(void *handle, u32 *flags)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
+	if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 1, 3))
+		return;
+
 	adev->mmhub.funcs->get_clockgating(adev, flags);
 
 	if (adev->ip_versions[ATHUB_HWIP][0] >= IP_VERSION(2, 1, 0))

Original file line number	Diff line number	Diff line change
`@@ -1140,6 +1140,9 @@ static void gmc_v10_0_get_clockgating_state(void handle, u32 flags)`
`1140`	`1140`	`{`
`1141`	`1141`	`struct amdgpu_device adev = (struct amdgpu_device )handle;`
`1142`	`1142`
	`1143`	`+ if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 1, 3))`
	`1144`	`+ return;`
	`1145`	`+`
`1143`	`1146`	`adev->mmhub.funcs->get_clockgating(adev, flags);`
`1144`	`1147`
`1145`	`1148`	`if (adev->ip_versions[ATHUB_HWIP][0] >= IP_VERSION(2, 1, 0))`