[PATCH 2/2] drm/radeon/uvd: Fix forgotten unmap buffer objects

zhanglianjie zhanglianjie at uniontech.com
Tue Feb 8 03:36:36 UTC 2022


Hi,
Thanks for your review. I have resubmitted, see 
https://lkml.org/lkml/2022/2/7/2014

> 
> 
> Am 29.01.22 um 08:35 schrieb zhanglianjie:
>> after the buffer object is successfully mapped, call radeon_bo_kunmap 
>> before the function returns.
>>
>> Signed-off-by: zhanglianjie <zhanglianjie at uniontech.com>
>>
>> diff --git a/drivers/gpu/drm/radeon/radeon_uvd.c 
>> b/drivers/gpu/drm/radeon/radeon_uvd.c
>> index 377f9cdb5b53..c5482f7793db 100644
>> --- a/drivers/gpu/drm/radeon/radeon_uvd.c
>> +++ b/drivers/gpu/drm/radeon/radeon_uvd.c
>> @@ -497,6 +497,7 @@ static int radeon_uvd_cs_msg(struct 
>> radeon_cs_parser *p, struct radeon_bo *bo,
>>       handle = msg[2];
>>
>>       if (handle == 0) {
>> +        radeon_bo_kunmap(bo);
>>           DRM_ERROR("Invalid UVD handle!\n");
>>           return -EINVAL;
>>       }
>> @@ -559,11 +560,10 @@ static int radeon_uvd_cs_msg(struct 
>> radeon_cs_parser *p, struct radeon_bo *bo,
>>           return 0;
>>
>>       default:
>> -
>>           DRM_ERROR("Illegal UVD message type (%d)!\n", msg_type);
>> -        return -EINVAL;
>>       }
>>
>> +    radeon_bo_kunmap(bo);
>>       BUG();
> 
> That looks like it will trigger this BUG() now. Please also remove that 
> line.
> 
> Apart from that the patches look good to me.
> 
> Regards,
> Christian.
> 
>>       return -EINVAL;
>>   }
>> -- 
>> 2.20.1
>>
>>
>>
> 
> 




More information about the amd-gfx mailing list