[FFmpeg-devel] [PATCH 1/3] diracdec: add 10-bit Haar SIMD functions

[James Darnley]

On 2018-07-27 15:05, Henrik Gramner wrote:
> On Fri, Jul 27, 2018 at 1:47 PM, James Darnley <jdarnley at obe.tv> wrote:
>> On 2018-07-26 17:29, Rostislav Pehlivanov wrote:
>>>> +cglobal horizontal_compose_haar_10bit, 3, 6+ARCH_X86_64, 4, b, temp_, w,
>>>> x, b2
>>>> +    DECLARE_REG_TMP 2,5
>>>> +    %if ARCH_X86_64
>>>> +        %define tail r6d
>>>> +    %else
>>>> +        %define tail dword wm
>>>> +    %endif
>>>> +
>>>>
>>>
>>> You can remove this whole bit, the init function only gets called if
>>> ARCH_X86_64 is true.
>>
>> Where did you get that from?  I don't require 64-bit for this.
> 
> Can't you just use 7 GPR:s on x86-32 as well?

I'm sure I've done that in the past and at least 1 platform has always
complained due to PIE or stack alignment or whatever, I think.  I went
looking for an old email but couldn't find it.

If you want me to try it I can.