This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
Re: [RFA:] test-case for relocs for linkonce-excluded sections don't become zero, causing g++ EH failure
- To: hans-peter dot nilsson at axis dot com
- Subject: Re: [RFA:] test-case for relocs for linkonce-excluded sections don't become zero, causing g++ EH failure
- From: Geoff Keating <geoffk at geoffk dot org>
- Date: Fri, 28 Sep 2001 16:04:56 -0700
- CC: binutils at sources dot redhat dot com
- References: <200109282213.AAA30150@ignucius.axis.se>
- Reply-to: Geoff Keating <geoffk at redhat dot com>
> Date: Sat, 29 Sep 2001 00:13:01 +0200
> From: Hans-Peter Nilsson <hans-peter.nilsson@axis.com>
>
> I started this writeup as a bug report for gcc, but it turns out
> to be a bug in binutils. Relocs for linkonce-excluded sections
> don't become zero, as gcc expects, and as happens in binutils
> 2.9.5/Debian 2.2, 2.9.5.0.22/RH 6.2 but not in "current" CVS for
> ix86 -- and worse, for CRIS too. ;-) This bug should be fixed
> before 2.12 in all ports serving a gcc port using DWARF 2 EH, or
> people will not be able to use C++ and exceptions, or will be
> forced to configure gcc with --enable-sjlj-exceptions.
This was looked at earlier, and the conclusion that was generally
agreed was that ld's behaviour is correct, and that GCC should instead,
for functions in linkonce sections, also put their EH data in linkonce
sections.
See <http://sources.redhat.com/ml/binutils/2001-06/msg00413.html>
and the following thread.
--
- Geoffrey Keating <geoffk@geoffk.org>