This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [rs6000] Fix compare debug failure on AIX
- From: Alan Modra <amodra at gmail dot com>
- To: David Edelsohn <dje dot gcc at gmail dot com>
- Cc: Eric Botcazou <ebotcazou at adacore dot com>, Andrew Pinski <pinskia at gmail dot com>, GCC Patches <gcc-patches at gcc dot gnu dot org>
- Date: Sat, 2 May 2015 13:54:17 +0930
- Subject: Re: [rs6000] Fix compare debug failure on AIX
- Authentication-results: sourceware.org; auth=none
- References: <1633828 dot L6eKc0dpVU at polaris> <EC523869-02C0-48DA-98B2-DFB5AE96D290 at gmail dot com> <1752984 dot z0ljzjyPbr at polaris> <CAGWvnynkc8eLtPndHRffZ-VfwgYyMD1m7m5Y37+Y84+KKP31rg at mail dot gmail dot com>
On Fri, May 01, 2015 at 09:48:51AM -0400, David Edelsohn wrote:
> On Thu, Apr 30, 2015 at 7:28 AM, Eric Botcazou <ebotcazou@adacore.com> wrote:
> >> We might want to check if doing -Og and not just -O0.
> >
> > You're right, thanks, amended patch attached, same ChangeLog.
>
> Why should GCC unnecessarily create stack frames to avoid
> compare-debug testcase failures?
More to the point, do you need a frame on AIX when debugging? If yes
then the correct change is to replace
else if (TARGET_XCOFF && write_symbols != NO_DEBUG)
info_ptr->push_p = 1;
with
else if (TARGET_XCOFF)
info_ptr->push_p = 1;
If no, then the correct change is to delete those lines from
rs6000_stack_info, and fix the comment.
https://gcc.gnu.org/ml/gcc-patches/2003-03/msg01693.html
--
Alan Modra
Australia Development Lab, IBM