[PATCH libdrm 2/2] amdgpu: Use char* pointers in amdgpu_find_bo_by_cpu_mapping

Michel Dänzer michel at daenzer.net
Thu Aug 16 13:55:12 UTC 2018


On 2018-08-15 03:07 AM, Zhang, Jerry (Junwei) wrote:
> On 08/14/2018 05:58 PM, Michel Dänzer wrote:
>> From: Michel Dänzer <michel.daenzer at amd.com>
>>
>> Arithmetic using void* pointers isn't defined by the C standard, only as
>> a GCC extension. Avoids compiler warnings:
>>
>> ../../amdgpu/amdgpu_bo.c: In function ‘amdgpu_find_bo_by_cpu_mapping’:
>> ../../amdgpu/amdgpu_bo.c:554:48: warning: pointer of type ‘void *’
>> used in arithmetic [-Wpointer-arith]
>>     if (cpu >= bo->cpu_ptr && cpu < (bo->cpu_ptr + bo->alloc_size))
>>                                                  ^
>> ../../amdgpu/amdgpu_bo.c:561:23: warning: pointer of type ‘void *’
>> used in subtraction [-Wpointer-arith]
>>     *offset_in_bo = cpu - bo->cpu_ptr;
>>                         ^
>>
>> Fixes: 4d454424e1f2 ("amdgpu: add a function to find bo by cpu mapping
>>                       (v2)")
>> Signed-off-by: Michel Dänzer <michel.daenzer at amd.com>
>> ---
>>   amdgpu/amdgpu.h    | 2 +-
>>   amdgpu/amdgpu_bo.c | 7 ++++---
>>   2 files changed, 5 insertions(+), 4 deletions(-)
>>
>> diff --git a/amdgpu/amdgpu.h b/amdgpu/amdgpu.h
>> index a8c353c6..f2bdeb95 100644
>> --- a/amdgpu/amdgpu.h
>> +++ b/amdgpu/amdgpu.h
>> @@ -695,7 +695,7 @@ int
>> amdgpu_create_bo_from_user_mem(amdgpu_device_handle dev,
>>    *
>>   */
>>   int amdgpu_find_bo_by_cpu_mapping(amdgpu_device_handle dev,
>> -                  void *cpu,
>> +                  char *cpu,
> 
> Shall we cast the cpu pointer when do arithmetic and keep the arguments
> as other cpu ponter?
> 
> e.g.
> @@ -565,14 +565,15 @@ int
> amdgpu_find_bo_by_cpu_mapping(amdgpu_device_handle dev,
>                 bo = handle_table_lookup(&dev->bo_handles, i);
>                 if (!bo || !bo->cpu_ptr || size > bo->alloc_size)
>                         continue;
> -               if (cpu >= bo->cpu_ptr && cpu < (bo->cpu_ptr +
> bo->alloc_size))
> +               if (cpu >= bo->cpu_ptr &&
> +                  (uint64_t)cpu < ((uint64_t)bo->cpu_ptr +
> bo->alloc_size))
>                         break;
>         }
> 
>         if (i < dev->bo_handles.max_key) {
>                 atomic_inc(&bo->refcount);
>                 *buf_handle = bo;
> -               *offset_in_bo = cpu - bo->cpu_ptr;
> +               *offset_in_bo = (uint64_t)cpu - (uint64_t)bo->cpu_ptr;
>         } else {
>                 *buf_handle = NULL;
>                 *offset_in_bo = 0;

Thanks for the suggestion, I sent out a v2 patch along those lines
(though slightly different).


-- 
Earthling Michel Dänzer               |               http://www.amd.com
Libre software enthusiast             |             Mesa and X developer


More information about the amd-gfx mailing list