[Intel-gfx] [PATCH i-g-t] tests/gem_softpin: Fix compiler warning on 32bit systems

[Dave Gordon]

On 10/12/15 16:43, Tvrtko Ursulin wrote:
>
> Hi,
>
> On 10/12/15 14:58, Mika Kuoppala wrote:
>> We get build error as we try to cast from ptr to integer
>> of different size on 32 bit platforms. Use unsigned long
>> as the cast, it will work with both 32 and 64 bit
>> systems.
>>
>> Cc: Vinay Belgaumkar <vinay.belgaumkar at intel.com>
>> Signed-off-by: Mika Kuoppala <mika.kuoppala at intel.com>
>> ---
>>   tests/gem_softpin.c | 6 +++---
>>   1 file changed, 3 insertions(+), 3 deletions(-)
>>
>> diff --git a/tests/gem_softpin.c b/tests/gem_softpin.c
>> index 7bee16b..0919716 100644
>> --- a/tests/gem_softpin.c
>> +++ b/tests/gem_softpin.c
>> @@ -155,7 +155,7 @@ static uint32_t init_userptr(int fd,
>> i915_gem_userptr *userptr, void *ptr,
>>
>>       memset((void*)userptr, 0, sizeof(i915_gem_userptr));
>>
>> -    userptr->user_ptr = (uint64_t)ptr; /* Need the cast to overcome
>> compiler warning */
>> +    userptr->user_ptr = (unsigned long)ptr; /* Need the cast to
>> overcome compiler warning */
>>       userptr->user_size = size;
>>       userptr->flags = 0; /* use synchronized operation */
>>
>> @@ -270,7 +270,7 @@ static void setup_execbuffer(struct
>> drm_i915_gem_execbuffer2 *execbuf,
>>                    struct drm_i915_gem_exec_object2 *exec_object,
>>                    int ring, int buffer_count, int batch_length)
>>   {
>> -    execbuf->buffers_ptr = (uint64_t)exec_object;
>> +    execbuf->buffers_ptr = (unsigned long)exec_object;
>>       execbuf->buffer_count = buffer_count;
>>       execbuf->batch_start_offset = 0;
>>       execbuf->batch_len = batch_length;
>> @@ -417,7 +417,7 @@ static void gem_pin_bo_test(void)
>>                  EXEC_OBJECT_PINNED, pinning_offset);
>>       setup_exec_obj(&exec_object2[1], batch_buf_handle, 0, 0);
>>       exec_object2[1].relocation_count = 1;
>> -    exec_object2[1].relocs_ptr = (uint64_t)reloc;
>> +    exec_object2[1].relocs_ptr = (unsigned long)reloc;
>>
>>       ring = I915_EXEC_RENDER;
>
> I had to try it to believe it and indeed gcc complains with "warning:
> cast from pointer to integer of different size [-Wpointer-to-int-cast]",
> which I don't really understand since cast is to a wider or equally wide
> data type.

I think the warning is really there for the pointer-to-signed-int 
conversion, in case that does an unexpected sign extension -- yes, I 
have seen machines where pointers were considered signed :( I suppose 
the result of converting a signed-pointer to unsigned-long-long might be 
"surprising" ...

Or maybe the gcc writers just thought it was more general to check for 
dest_size != source_size rather than < :-?

> Vinay had uintptr_t there originally which does not suffer from this
> warning so this one is my fault. I don't get the gcc logic but whatever.

'uintptr_t' should be the right type for the cast; and you should then 
be able assign that to any larger unsigned type without getting compiler 
warnings!

[Aside: I've also seen at one machine where there simply wasn't any type 
that could be used for uintptr_t; it had 4-byte 'long's and no 'long 
long', but pointers were 6 bytes!]

.Dave.

> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin at intel.com>
>
> Regards,
>
> Tvrtko
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/intel-gfx