Re: PATCH: Fix linkonce for ELF/PPC.

[Geoff Keating <geoffk at geoffk dot org>]

> Date: Mon, 29 Oct 2001 08:47:21 -0800
> From: "H . J . Lu" <hjl@lucon.org>
> Content-Disposition: inline
> User-Agent: Mutt/1.2.5i
> 
> On Sun, Oct 28, 2001 at 10:49:39PM -0800, H . J . Lu wrote:
> > On Mon, Oct 29, 2001 at 05:04:55PM +1030, Alan Modra wrote:
> > > On Sun, Oct 28, 2001 at 09:58:26PM -0800, H . J . Lu wrote:
> > > > 
> > > > I don't know how it happens on ELF/PPC. The problem can be in linker,
> > > > gcc and glibc. What I saw was the gcc couldn't catch the exception with
> > > > libstdc++.so built by the current ld. It just aborted. Also see
> > > > 
> > > > http://gcc.gnu.org/ml/gcc/2001-09/msg01248.html
> > > > 
> > > > It may or may not be related. Feel to find a real fix for the problem.
> > > > But we still need a workaround in the meantime.
> > > 
> > > Hmm, considering the above post and Geoff Keating's previous remarks
> > > (http://sources.redhat.com/ml/binutils/2001-06/msg00413.html) about
> > 
> > FWIW, binutils 2.11.92.0.10 seems ok with static libstdc++.a in gcc
> > 3.1.
> > 
> > > requiring zeros, I can't see how we can implement a correct solution
> > > in binutils.  It seems to me that this is a gcc bug, and we shouldn't
> > > be tweaking binutils at all.
> > 
> > It is ok with me as long as we all agree it is a gcc bug.
> > 
> > Franz, do you have any problems with the current Linux/PPC gcc and
> > binutils 2.11.92.0.10? If not, please report it as a gcc 3.x bug.
> > 
> > 
> 
> I have a question. Is it possible for ELF/PPC that for symbol reference,
> the assembler generates RELA relocations such that the static/dynamic
> linkers have to apply addend, even for undefined symbols?

Yes, it is possible.

-- 
- Geoffrey Keating <geoffk@geoffk.org>