This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
Re: binutils is broken on ELF/mips (Re: binutils is broken on Linux/alpha)
> Date: Sat, 6 Oct 2001 11:13:34 +0930
> From: Alan Modra <amodra@bigpond.net.au>
> Cc: Geoff Keating <geoffk@redhat.com>, hjl@lucon.org,
> binutils@sourceware.cygnus.com
> Mail-Followup-To: Richard Henderson <rth@redhat.com>,
> Geoff Keating <geoffk@redhat.com>, hjl@lucon.org,
> binutils@sourceware.cygnus.com
> Content-Disposition: inline
> User-Agent: Mutt/1.3.17i
>
> On Fri, Oct 05, 2001 at 03:03:48PM -0700, Richard Henderson wrote:
> > On Fri, Oct 05, 2001 at 01:11:48PM -0700, Geoff Keating wrote:
> > > That is the fix I was referring to: arrange the gcc unwind mechanism
> > > so that link-once sections are used properly.
> >
> > Can we actually do this without SHT_GROUP? I'm not sure we can.
>
> All that's needed in this particular case is some sort of relocation
> referencing a symbol in the eh sections. R_NONE might even do the
> trick. We'd need a new assembler pseudo-op if we want to generate
> the relocs without adding rubbish in .text.
I think you're thinking of linker GC. This is just link-once
sections, so at least one copy should end up in the output anyway.
(Of course, yes, this will break linker GC at least initially, but
that may not be worse than the existing breakage.)
--
- Geoffrey Keating <geoffk@geoffk.org>