OSDN Git Service

drm/amdgpu: Correct the amdgpu_ucode_fini_bo place for Tonga
authorEmily Deng <Emily.Deng@amd.com>
Thu, 8 Mar 2018 02:49:09 +0000 (10:49 +0800)
committerAlex Deucher <alexander.deucher@amd.com>
Wed, 14 Mar 2018 19:38:25 +0000 (14:38 -0500)
The amdgpu_ucode_fini_bo should be called after gfx_v8_0_hw_fini,
or it will have KCQ disable failed issue.

For Tonga, as it firstly finishes SMC block, and the SMC hw fini
will call amdgpu_ucode_fini, which will lead the amdgpu_ucode_fini_bo
called before gfx_v8_0_hw_fini, this is incorrect.

Signed-off-by: Emily Deng <Emily.Deng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
drivers/gpu/drm/amd/amdgpu/amdgpu_powerplay.c

index 9a2c2c3..cc582e2 100644 (file)
@@ -1457,6 +1457,9 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev)
        }
 
        for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
+               if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_SMC &&
+                       adev->firmware.load_type == AMDGPU_FW_LOAD_SMU)
+                       amdgpu_ucode_fini_bo(adev);
                if (!adev->ip_blocks[i].status.hw)
                        continue;
 
index 5c2e2d5..825c9b9 100644 (file)
@@ -162,9 +162,6 @@ static int amdgpu_pp_hw_fini(void *handle)
                ret = adev->powerplay.ip_funcs->hw_fini(
                                        adev->powerplay.pp_handle);
 
-       if (adev->firmware.load_type == AMDGPU_FW_LOAD_SMU)
-               amdgpu_ucode_fini_bo(adev);
-
        return ret;
 }