This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH] PPC64: First in the series of patches implementing POWER8 vector math.
- From: Joseph Myers <joseph at codesourcery dot com>
- To: GT <tnggil at protonmail dot com>
- Cc: "libc-alpha at sourceware dot org" <libc-alpha at sourceware dot org>
- Date: Fri, 15 Feb 2019 21:07:25 +0000
- Subject: Re: [PATCH] PPC64: First in the series of patches implementing POWER8 vector math.
- References: <7z03WSXwPIOYABvW1ZYzHobMOv90wBWKo-e0YptRFovFUpRnrSEGN-zv08kOe_fPJMFnHx1khbcfcfooubTStEAtMX96Q2U4cRVZOm_TZl8=@protonmail.com> <alpine.DEB.2.21.1902142104290.4022@digraph.polyomino.org.uk> <bP9cSHHU7t33kQuGpgbcCiszcjGQgzFwy6FpxoUcDV3cpfkeleUwmjx9BDvJexHALeLVfAgxLjo88pP8Sxyx7g4HyHDdrEnRO4rqKtXBU3k=@protonmail.com>
On Fri, 15 Feb 2019, GT wrote:
> > > +#if defined _ARCH_PPC64 && defined FAST_MATH
> >
> > Is _ARCH_PPC64 correct here - what's the status of support (in the GNU
> > toolchain, Linux kernel, etc.) for -mpowerpc64 with the 32-bit ABI (which
> > also defines _ARCH_PPC64)?
>
> To address the support issues raised, a solution would be to have
> 'configure' verify that the compiler generates a valid executable. Then
> _ARCH_PPC64 would be replaced here by a macro determined at
> configuration time.
Since installed headers need to work for all multilibs that might share a
compiler and a set of headers, a configure-time test isn't suitable here.
I think __powerpc64__ is the correct thing to test as an ABI conditional
(as opposed to an instruction set conditional) - it's what
sysdeps/unix/sysv/linux/powerpc/bits/wordsize.h uses.
--
Joseph S. Myers
joseph@codesourcery.com