This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: .eh_frame section
- To: Joern Rennecke <amylaar at cygnus dot co dot uk>
- Subject: Re: .eh_frame section
- From: Richard Henderson <rth at cygnus dot com>
- Date: Mon, 13 Oct 1997 11:15:19 -0700
- Cc: Bruno Haible <haible at ilog dot fr>, jason at cygnus dot com, ian at cygnus dot com, rth at cygnus dot com, meissner at cygnus dot com, egcs at cygnus dot com
- References: <199710102251.AAA25911@halles.ilog.fr> <199710131642.RAA05335@phal.cygnus.co.uk>
- Reply-To: Richard Henderson <rth at cygnus dot com>
On Mon, Oct 13, 1997 at 05:42:37PM +0100, Joern Rennecke wrote:
> > Fine. So we can replace absolute labels by label differences. I propose
> > the following: When the compiler output contains either a .gcc_except or
> > a .eh_table section, the beginning of the text section for this module
> > is marked as __TEXT_BEGIN__ and subtracted from all labels in the tables.
> > A new section .text_begin is created which only contains the absolute
> > __TEXT_BEGIN__.
>
> This doesn't sound like it'll work with linker relaxation, unless you
> want to add a lot of extra code to the linker to decode this information,
> translate it into proper relocations, and re-encode it after its
> processing is done.
Actually, it's not going to work at all except on Irix6 and OSF/1. No
one else has the relocations to subtract symbols at link time vs asm time.
r~