This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: patch to improve i386 epilogue (version 4)
- To: Ulrich Drepper <drepper at cygnus dot com>, egcs-patches at cygnus dot com
- Subject: Re: patch to improve i386 epilogue (version 4)
- From: Richard Henderson <rth at cygnus dot com>
- Date: Fri, 9 Oct 1998 11:43:36 -0700
- Cc: mark at markmitchell dot com
- References: <199810081643.MAA13935@jwlab.FEITH.COM>, <199810081857.LAA07960.cygnus.egcs.patches@smtp.earthlink.net> <r2u31dolhd.fsf@happy.cygnus.com>
- Reply-To: Richard Henderson <rth at cygnus dot com>
On Fri, Oct 09, 1998 at 07:31:26AM -0700, Ulrich Drepper wrote:
> But this cannot happen until we can switch over to DWARF for
> all Linux (which are ready for wider use). This currently is not
> possible. And once it is possible it requires careful handling. All
> libraries at this point will have stabs information.
You know, gdb can and does properly handle the situation where
there are multiple debug formats in the same program. It slurps
all of the info it can find in any format into its own internal
data structures.
That is not an issue for blocking a move to dwarf2.
What does need to happen though, is for someone to teach gdb
how to use the .debug_frame info to unwind the stack. At the
moment it ignores that info and just uses its traditional code
walk and frame pointer.
r~