> However, I'm not quite sure why does
>
> ++#ifdef S390_CRC32_VX
> ++ if (getauxval(AT_HWCAP) & HWCAP_S390_VX)
> ++ return s390_crc32_vx;
> ++#endif
>
> work. For me getauxval() causes a segfault when called from an ifunc,
> that's why I have to use the s390-specific ifunc's hwcap parameter.
Perhaps a bug in your compiler version?
In any case, I believe we can consider the current patch as working as
intended, despite this mystery.
> However, I'm not quite sure why does AT_HWCAP) & HWCAP_S390_VX)
>
> ++#ifdef S390_CRC32_VX
> ++ if (getauxval(
> ++ return s390_crc32_vx;
> ++#endif
>
> work. For me getauxval() causes a segfault when called from an ifunc,
> that's why I have to use the s390-specific ifunc's hwcap parameter.
Perhaps a bug in your compiler version?
In any case, I believe we can consider the current patch as working as
intended, despite this mystery.