This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH 00/19] PowerPC: multiarch support for PPC32
- From: Steven Munroe <munroesj at linux dot vnet dot ibm dot comcom>
- To: Adhemerval Zanella <azanella at linux dot vnet dot ibm dot com>
- Cc: "GNU C. Library" <libc-alpha at sourceware dot org>
- Date: Thu, 14 Nov 2013 21:30:02 -0600
- Subject: Re: [PATCH 00/19] PowerPC: multiarch support for PPC32
- Authentication-results: sourceware.org; auth=none
- References: <527248E7 dot 9070706 at linux dot vnet dot ibm dot com>
- Reply-to: munroe at us dot ibm dot com
On Thu, 2013-10-31 at 10:11 -0200, Adhemerval Zanella wrote:
> Hi all,
>
> This is respin of my patchset for multilib support to PowerPC. As previsouly
> discussed, I have changed the minimum support for PPC32 to be POWER4+
> (--with-cpu=power4). This is a large patchset so I divided in 4 parts:
>
> 1. String functions for PPC32
> 2. Math functions for PPC32
> 3. String functions for PPC64
> 4. Math functions for PPC64
>
> Some notes:
>
> * The complete pathset is at azanella/multilib-ppc branch.
>
> * I rewrote the patches to get rid of the location change of implementations.
> It was a bad strategy and the idea is keep build with --disable-multi-arch
> similar to current build.
>
> * The support is complete with both string/etc. and fpu specialized functions.
>
> * With the patches I pushed some common implementation to sysdeps/powerpc/powerX
> from sysdeps/powerpc/power32/.. and this required some cleanup on PowerPC64
> side.
>
> * Current multiarch is just enabled for --with-cpu-power4. The patch to enabled
> it for power5/power5+/power6/power6x/power7/power8 is at second patchset
> (math functions for ppc32)
>
> I tested with GCC 4.4.6 and 4.8.2 on PowerPC32.
>
It would be good to describe you strategy for overriding the entry/exit
macros then including the current --with-cpu implementations. Otherwise
it looks like empty implementations.