On Thu, Jun 29, 2017 at 12:47:58PM -0300, James Almer wrote:
> On 6/28/2017 10:21 AM, Clément Bœsch wrote:
> > We have AV_CPU_FLAG_ARMV8 == AV_CPU_FLAG_SSE3 which causes a trigger of
> > this MMX warning on AArch64.
> > ---
> > libavutil/cpu.c | 3 ++-
> > 1 file changed, 2 insertions(+), 1 deletion(-)
> >
> > diff --git a/libavutil/cpu.c b/libavutil/cpu.c
> > index 16e0c9278f..a22da0fa8c 100644
> > --- a/libavutil/cpu.c
> > +++ b/libavutil/cpu.c
> > @@ -61,7 +61,8 @@ static int get_cpu_flags(void)
> > }
> >
> > void av_force_cpu_flags(int arg){
> > - if ( (arg & ( AV_CPU_FLAG_3DNOW |
> > + if (ARCH_X86 &&
> > + (arg & ( AV_CPU_FLAG_3DNOW |
> > AV_CPU_FLAG_3DNOWEXT |
> > AV_CPU_FLAG_MMXEXT |
> > AV_CPU_FLAG_SSE |
> >
>
> LGTM of course. Goes to show how the CLI or anything where log messages
> are clearly visible are barely used on non-x86 targets.
>
> This should be backported as well. It's not simply a warning as it
> blindly sets the first bit of the cpuflags variable (MMX on x86, altivec
> on PPC, armv6 on arm, etc).applied, feel free to backport -- Clément B.
signature.asc
Description: PGP signature
_______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
