This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: Move powerpc ports pieces to libc.
- From: Adhemerval Zanella <azanella at linux dot vnet dot ibm dot com>
- To: libc-alpha at sourceware dot org
- Date: Fri, 27 Sep 2013 10:09:07 -0300
- Subject: Re: Move powerpc ports pieces to libc.
- Authentication-results: sourceware.org; auth=none
- References: <Pine dot LNX dot 4 dot 64 dot 1309261728050 dot 13950 at digraph dot polyomino dot org dot uk>
On 26-09-2013 14:29, Joseph S. Myers wrote:
> This patch moves the powerpc-nofpu port and 4xx string functions from
> ports to libc. Together with the diff below for the README update and
> one case where file contents need merging into an existing file, the
> following git commands do the directory moves (and removal of the file
> from which contents were merged):
>
> git mv ports/sysdeps/powerpc/nofpu sysdeps/powerpc/nofpu
> git mv ports/sysdeps/powerpc/soft-fp sysdeps/powerpc/soft-fp
> git mv ports/sysdeps/powerpc/powerpc32/405 sysdeps/powerpc/powerpc32/405
> git mv ports/sysdeps/powerpc/powerpc32/440 sysdeps/powerpc/powerpc32/440
> git mv ports/sysdeps/powerpc/powerpc32/464 sysdeps/powerpc/powerpc32/464
> git mv ports/sysdeps/powerpc/powerpc32/476 sysdeps/powerpc/powerpc32/476
> git rm ports/sysdeps/powerpc/powerpc32/Makefile
> git mv ports/sysdeps/unix/sysv/linux/powerpc/powerpc32/nofpu sysdeps/unix/sysv/linux/powerpc/powerpc32/nofpu
> git mv ports/sysdeps/unix/sysv/linux/powerpc/powerpc32/405 sysdeps/unix/sysv/linux/powerpc/powerpc32/405
> git mv ports/sysdeps/unix/sysv/linux/powerpc/powerpc32/440 sysdeps/unix/sysv/linux/powerpc/powerpc32/440
> git mv ports/sysdeps/unix/sysv/linux/powerpc/powerpc32/464 sysdeps/unix/sysv/linux/powerpc/powerpc32/464
> git mv ports/sysdeps/unix/sysv/linux/powerpc/powerpc32/476 sysdeps/unix/sysv/linux/powerpc/powerpc32/476
>
> Tested that for both hard-float and soft-float powerpc32 builds, the
> disassembly of installed binaries is the same before and after the
> change (paths in debug info are of course different).
>
Thanks for the patch Joseph, I'm just thinking here the implication of this
change for multilib support. Although I didn't receive any reply for my
patches I still want to push them forward after Alan's LE patches. Currently
my set of patches onyl support the POWER variant and I don't think it
makes sense to add the memory/string embedded optimizations on the
POWER libc/libm.
So I open to suggestions, should I isolate the multilib to be POWER4+ only
(by consolidating the implementations on sysdeps/powerpc/powerpc32/power4)?
In this case the file organization would be somewhat strange, but it is
something I can live with. Or should we push the multilib as my initial
patch set and force the embedded build to build with an option like
'--disable-multiarch'?