This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Fix for PR 12537
- From: Geoff Keating <geoffk at geoffk dot org>
- To: amodra at bigpond dot net dot au
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Sat, 29 Nov 2003 15:03:29 -0800
- Subject: Re: Fix for PR 12537
- References: <20031018055754.GO24720@bubble.sa.bigpond.net.au> <200310180717.h9I7HSj03473@desire.geoffk.org> <20031018081302.GQ24720@bubble.sa.bigpond.net.au> <jmllrixmr0.fsf@desire.geoffk.org> <20031129014848.GU2508@bubble.sa.bigpond.net.au>
> Date: Sat, 29 Nov 2003 12:18:48 +1030
> From: Alan Modra <amodra@bigpond.net.au>
> On Sat, Oct 18, 2003 at 02:24:35PM -0700, Geoff Keating wrote:
> > Alan Modra <amodra@bigpond.net.au> writes:
> >
> > ...
> > > Tue Aug 20 18:49:55 1996 Michael Meissner <meissner@tiktok.cygnus.com>
> > >
> > > * rs6000/sysv4.h (ASM_OUTPUT_SECTION_NAME): If -mrelocatable or
> > > -mrelocatable-lib, don't make read-only sections.
> >
> > This looks like a broken workaround for some bug in the middle-end or
> > elsewhere in this backend that probably doesn't exist any more. Let's
> > try taking it out and seeing what happens. If TARGET_RELOCATABLE,
> > then flag_pic will be set to 2 by SUBTARGET_OVERRIDE_OPTIONS in
> > sysv4.h, so any section that actually needs fixups should be made
> > writable by default_section_type_flags_1 anyway.
>
> Another post reminded me that I didn't get back to you on this. The
> following has been bootstrapped and regression tested (with multiple
> versions of gcc) on powerpc-linux and powerpc64-linux. OK to install?
> gcc-3.3 branch too?
>
> * config/rs6000/rs6000.c (rs6000_elf_section_type_flags): Don't
> set SECTION_WRITE on TARGET_RELOCATABLE.
This is OK. I would wait a week or two before putting it in GCC 3.3,
but it's up to you.
--
- Geoffrey Keating <geoffk@geoffk.org>