Sean McGovern <[email protected]> writes:

> On Tuesday, December 13, 2011, Måns Rullgård <[email protected]> wrote:
>> Sean McGovern <[email protected]> writes:
>>
> [snip]
>
>> Try this patch:
>>
>> diff --git a/libavcodec/x86/mathops.h b/libavcodec/x86/mathops.h
>> index 50b0283..e1e7ccb 100644
>> --- a/libavcodec/x86/mathops.h
>> +++ b/libavcodec/x86/mathops.h
>> @@ -95,7 +95,7 @@ __asm__ volatile(\
>>     "cmovl %4, %1       \n\t"\
>>     "cmovl %5, %2       \n\t"\
>>     : "+&r" (x), "+&r" (a), "+r" (c)\
>> -    : "r" (y), "r" (b), "r" (d)\
>> +    : "r" (y), "r" ((int)b), "r" ((int)d)\
>>  );
>>  #endif
>
> This just moved the error up 2 lines.

Please elaborate.

> I'm starting to think I might just be wasting my time on fixing this,
> especially when it's not a portability issue on our part. It also looks
> like the remaining issues just to get it building are beyond my skill level.

There are only a few remaining issues.

> An offline message from Mike Kostylev suggested that this and h264dsp_mmx.c
> may be "unfixable" w.r.t. compiling with suncc  -- it uses %a's as well
> which is passed as a literal to the assembler (this would explain why cabac
> works again).

Yes, this is the reason that file is failing.  Ronald had a patch to
move this hideous mess to yasm, but it was shot down by the enemy.
Now that the enemy is defeated, we should recover and apply it.

-- 
Måns Rullgård
[email protected]
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to