PATCH: PR target/37394: [4.4 Regression] Segfault in ia64_variable_issue with -O -fschedule-insns2
H.J. Lu
hjl.tools@gmail.com
Thu Sep 18 14:12:00 GMT 2008
On Wed, Sep 17, 2008 at 10:09 PM, H.J. Lu <hjl.tools@gmail.com> wrote:
> We used to check and set flag_schedule_insns_after_reload and
> ia64_flag_schedule_insns2 in ia64_override_options after all command
> line options have been processed. It was moved to
> ia64_optimization_options to support #pragma GCC optimize and
> attribute((optimize(...))). But it breaks "-O -fschedule-insns2".
> Also #pragma GCC optimize and attribute((optimize(...))) don't
> work correctly:
>
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37565
>
> since OPTIMIZATION_OPTIONS is executed before -fschedule-insns2
> is processed when it is used in #pragma GCC optimize.
>
> OK for trunk if it passes on Linux/ia64?
>
> Thanks.
>
>
> H.J.
> ---
> gcc/
>
> 2008-09-17 H.J. Lu <hongjiu.lu@intel.com>
>
> PR target/37394
> * config/ia64/ia64.c (ia64_optimization_options): Move
> checking and setting flag_schedule_insns_after_reload and
> ia64_flag_schedule_insns2 back to ...
> (ia64_override_options): Here.
>
> gcc/testsuite/
>
> 2008-09-17 H.J. Lu <hongjiu.lu@intel.com>
>
> PR target/37394
> * g++.dg/other/pr37394.C: New.
>
Hi Steve,
It passed on Linux/ia64. OK for trunk?
Thanks.
--
H.J.
More information about the Gcc-patches
mailing list