This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Sparc ELF definitions of {INIT,FINI}_SECTION_ASM_OP
- From: "David O'Brien" <obrien at FreeBSD dot org>
- To: "David S. Miller" <davem at redhat dot com>
- Cc: jakub at redhat dot com, rth at redhat dot com, gcc at gcc dot gnu dot org
- Date: Wed, 3 Apr 2002 12:34:53 -0800
- Subject: Re: Sparc ELF definitions of {INIT,FINI}_SECTION_ASM_OP
- Organization: The NUXI BSD group
- References: <20020403121430.A12004@dragon.nuxi.com> <20020403.121537.40264863.davem@redhat.com>
- Reply-to: obrien at FreeBSD dot org
On Wed, Apr 03, 2002 at 12:15:37PM -0800, David S. Miller wrote:
> From: "David O'Brien" <obrien@FreeBSD.org>
> Date: Wed, 3 Apr 2002 12:14:30 -0800
>
> If so, will all ELF/SVR4 assemblers accept "'s around them?
>
> It is probably target specific being that it is set in
> config/${cpu}/sysv4.h and not config/svr4.h
Perhaps you did not know that config/elfos.h grew out of config/svr4.h.
config/svr4.h now only contains system implementation details such as the
the MD_EXEC_PREFIX path ("/usr/ccs/bin/"). All config/svr4.h consumers
must include config/elfos.h also.
Thus for this discussion, please pretend I had used config/svr4.h rather
than config/elfos.h in my original question.
config/sparc/sysv4.h says this:
/* Undefined some symbols which are defined in "svr4.h" but which are
appropriate only for typical svr4 systems, but not for the specific
case of svr4 running on a Sparc. */
#undef INIT_SECTION_ASM_OP
#undef FINI_SECTION_ASM_OP
#define INIT_SECTION_ASM_OP "\t.section\t\".init\""
#define FINI_SECTION_ASM_OP "\t.section\t\".fini\""
So why is ELF on Sparc so special?
--
-- David (obrien@FreeBSD.org)