On 28/08/13 10:25, Diego Biurrun wrote:
> On Wed, Aug 28, 2013 at 10:24:37AM +0200, Luca Barbato wrote:
>> On 28/08/13 10:00, Diego Biurrun wrote:
>>> On Wed, Aug 28, 2013 at 12:47:35AM +0200, Luca Barbato wrote:
>>>> On 28/08/13 00:12, Diego Biurrun wrote:
>>>>> The i686 feature really is a CPU feature and should be handled as such.
>>>>> The cpunop dependency on i686 should be expressed with a standard _deps
>>>>> declaration instead of a manual test.
>>>>
>>>> What i686 does exactly?
>>>
>>> It's supposed to cover the changes from the i586 to i686 x86 architecture,
>>> which covers cmov, long nops, and probably more that I forget.
>>
>> x86 != x86_64
>>
>> I wouldn't mix those.
> 
> Yes, sure, but what's your point here?  That x86_64 does not imply i686?
> I doubt there will ever be an x86_64 CPU w/o cmov and friends; it would
> not be backwards-compatible..

>From experience those assumption are quite short lived =)

Any x86 expert willing to chip in?

lu




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

Reply via email to