[FFmpeg-devel] [PATCH] avcodec/h264: mmxext 4:2:2 chroma deblock/loop filter

[James Darnley]

On 2016-01-15 21:55, James Almer wrote:
> On 1/15/2016 5:00 PM, James Darnley wrote:
>> On 2016-01-15 03:55, James Almer wrote:
>>> On 1/14/2016 11:05 PM, James Darnley wrote:
>>>> diff --git a/libavcodec/x86/h264_deblock.asm b/libavcodec/x86/h264_deblock.asm
>>>> index 5151f3c..20f0814 100644
>>>> --- a/libavcodec/x86/h264_deblock.asm
>>>> +++ b/libavcodec/x86/h264_deblock.asm
>>>> @@ -864,7 +864,47 @@ ff_chroma_inter_body_mmxext:
>>>>      DEBLOCK_P0_Q0
>>>>      ret
>>>>  
>>>> +cglobal h264_h_loop_filter_chroma422_8, 5, 7, 8, mmsize + ARCH_X86_64*2*mmsize
>>>
>>> This will not work with x86_32 compilers that don't have aligned stack (Like msvc)
>>> because r6 is needed to store the stack pointer.
>>
>> The other chroma deblock function I borrowed most of this from doesn't
>> appear to use any guard against that (see a few lines above where my
>> patch starts).  Neither in assembly or in the init function.
> 
> That's because the other chroma function doesn't allocate stack using x86inc magic.
> On x86_32 it uses r#m only and on x86_64 it handles the stack pointer by itself.

I don't understand how that works then.  It calls the cglobal macro.
Does that not just assume missing args are 0?  I'd better look and
follow that more closely.


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 603 bytes
Desc: OpenPGP digital signature
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20160115/5a6852b2/attachment.sig>