Re: [FFmpeg-devel] [PATCH 1/2] aarch64: Simplify the linux runtime cpu detection code

[Sean McGovern <gseanmcg@gmail.com>]

On Tue, Oct 24, 2023 at 8:40 AM Sean McGovern <gseanmcg@gmail.com> wrote:
>
>
>
> On Tue, Oct 24, 2023, 08:23 Martin Storsjö <martin@martin.st> wrote:
>>
>> Skip doing the whole getauxval(AT_HWCAP) if HWCAP_CPUID isn't
>> defined.
>> ---
>>  libavutil/aarch64/cpu.c | 6 +-----
>>  1 file changed, 1 insertion(+), 5 deletions(-)
>>
>> diff --git a/libavutil/aarch64/cpu.c b/libavutil/aarch64/cpu.c
>> index bd780e8591..2b50c426bc 100644
>> --- a/libavutil/aarch64/cpu.c
>> +++ b/libavutil/aarch64/cpu.c
>> @@ -30,11 +30,9 @@
>>  static int detect_flags(void)
>>  {
>>      int flags = 0;
>> -    unsigned long hwcap;
>> -
>> -    hwcap = getauxval(AT_HWCAP);
>>
>>  #if defined(HWCAP_CPUID)
>> +    unsigned long hwcap = getauxval(AT_HWCAP);
>>      // We can check for DOTPROD and I8MM using HWCAP_ASIMDDP and
>>      // HWCAP2_I8MM too, avoiding to read the CPUID registers (which triggers
>>      // a trap, handled by the kernel). However the HWCAP_* defines for these
>> @@ -53,8 +51,6 @@ static int detect_flags(void)
>>          if (((tmp >> 52) & 0xf) == 0x1)
>>              flags |= AV_CPU_FLAG_I8MM;
>>      }
>> -#else
>> -    (void)hwcap;
>>  #endif
>>
>>      return flags;
>> --
>> 2.34.1
>>
>> _______________________________________________
>> ffmpeg-devel mailing list
>> ffmpeg-devel@ffmpeg.org
>> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>>
>> To unsubscribe, visit link above, or email
>> ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
>
>
> Good call, and LGTM.
>
> -- Sean McGovern

Ooops! I don't have c= so maybe someone who does could push this --
and the 2nd patch in this set needs a pair of eyes more familiar with
compiler bugs.

-- Sean McG
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".