[PATCH] drm/amdgpu: fix a warning in amdgpu_ras.c

Nirmoy nirmodas at amd.com
Fri Sep 25 19:01:09 UTC 2020


On 9/25/20 7:36 PM, Luben Tuikov wrote:
> On 2020-09-25 11:00, Nirmoy wrote:
>> Acked-by: Nirmoy Das <nirmoy.das at amd.com>
>>
>> On 9/25/20 4:31 PM, Alex Deucher wrote:
>>> drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c: In function ‘amdgpu_ras_fs_init’:
>>> drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c:1284:2: warning: ignoring return value of ‘sysfs_create_group’, declared with attribute warn_unused_result [-Wunused-result]
>>>    1284 |  sysfs_create_group(&adev->dev->kobj, &group);
>>>         |  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>>>
>>> Signed-off-by: Alex Deucher <alexander.deucher at amd.com>
>>> ---
>>>
>>> Do we care whether this fails or not?
>>
>> I think we should. Failure in sysfs_create_group() means memory starved
>>
>> system or we are doing something in the driver code.
>>
>> IMO in both cases, we should error out.
> I disagree. We should try to bring up a display as much as possible.
> sysfs failing shouldn't necessarily bring down the system.
> The system should be as resilient as possible and try to survive
> as much as possible.


True, there is no use of printing error if we can't see that.


Alex,


We should then at least throw a warning and return void in 
amdgpu_ras_fs_init().


Thanks,

Nirmoy



>
> If indeed we had had a memory starvation, other things would fail,
> before or after this sequence. And if those things are terminal,
> which sysfs is not, then let those other errors, such as no memory
> for BOs, bring the display bring-up down.
>
> Regards,
> Luben
>
>
>>
>> Nirmoy
>>
>>>    drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 4 +---
>>>    1 file changed, 1 insertion(+), 3 deletions(-)
>>>
>>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
>>> index 49d10330bf64..67724049a0fc 100644
>>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
>>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
>>> @@ -1281,9 +1281,7 @@ static int amdgpu_ras_fs_init(struct amdgpu_device *adev)
>>>    		sysfs_bin_attr_init(bin_attrs[0]);
>>>    	}
>>>    
>>> -	sysfs_create_group(&adev->dev->kobj, &group);
>>> -
>>> -	return 0;
>>> +	return sysfs_create_group(&adev->dev->kobj, &group);
>>>    }
>>>    
>>>    static int amdgpu_ras_fs_fini(struct amdgpu_device *adev)
>> _______________________________________________
>> amd-gfx mailing list
>> amd-gfx at lists.freedesktop.org
>> https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.freedesktop.org%2Fmailman%2Flistinfo%2Famd-gfx&data=02%7C01%7Cluben.tuikov%40amd.com%7Cd3a29eb3223b4b58ff6b08d86163c8a9%7C3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C637366428560245712&sdata=bYWjk2k8IeEolG2wcJLA7J%2Bx2d%2FOSXNIOpLObk%2F3jSQ%3D&reserved=0
>>


More information about the amd-gfx mailing list