This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] powerpc64-linux bi-arch support
- From: Jakub Jelinek <jakub at redhat dot com>
- To: David Edelsohn <dje at watson dot ibm dot com>
- Cc: Alan Modra <amodra at bigpond dot net dot au>, Janis Johnson <janis187 at us dot ibm dot com>, gcc-patches at gcc dot gnu dot org
- Date: Tue, 27 May 2003 11:44:23 -0400
- Subject: Re: [PATCH] powerpc64-linux bi-arch support
- References: <jakub@redhat.com> <200305271506.LAA26090@makai.watson.ibm.com>
- Reply-to: Jakub Jelinek <jakub at redhat dot com>
On Tue, May 27, 2003 at 11:06:00AM -0400, David Edelsohn wrote:
> >>>>> Jakub Jelinek writes:
>
> >> Why are you changing RS6000_CALL_GLUE based on TARGET_64BIT? It is a new
> >> mnemonic versus old mnemonic issue. PPC32 Linux does not use it and PPC64
> >> only uses "nop".
>
> Jakub> I think -m32 -mcall-aixdesc still uses it.
> Jakub> If it is ok for -m32 -mcall-aixdesc to emit nop instead of cror 31,31,31,
> Jakub> then this hunk can certainly go away.
>
> If that's the reason, then it certainly is not reflected in any
> comment and the test should be based on aix-calldesc, not TARGET_64BIT.
A comment can be certainly added (most of such linux64.h changes were simple
if sysv4.h/linux.h version differs from linux64.h, use
previous linux64.h definition for TARGET_64BIT and sysv4.h/linux.h
definition for !TARGET_64BIT).
But I don't see how the test can be based on aixdesc (ie. DEFAULT_ABI == ABI_AIX)
because ABI_AIX is set for both -m64 (which is -m64 -mcall-aixdesc with no
other -mcall-* variants allowed) and -m32 -mcall-aixdesc.
ppc64-linux -m64 only compiler hardcodes DEFAULT_ABI to ABI_AIX while with
the bi-arch patch it is variable initialized to ABI_AIX with an error if
cmdline switches override it in -m64 mode.
Jakub