This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug rtl-optimization/68381] [6 Regression] wrong code and quality regression with __builtin_mul_overflow() @ aarch64
- From: "segher at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Tue, 24 Nov 2015 06:43:52 +0000
- Subject: [Bug rtl-optimization/68381] [6 Regression] wrong code and quality regression with __builtin_mul_overflow() @ aarch64
- Auto-submitted: auto-generated
- References: <bug-68381-4 at http dot gcc dot gnu dot org/bugzilla/>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68381
--- Comment #4 from Segher Boessenkool <segher at gcc dot gnu.org> ---
Author: segher
Date: Tue Nov 24 06:43:20 2015
New Revision: 230786
URL: https://gcc.gnu.org/viewcvs?rev=230786&root=gcc&view=rev
Log:
combine: Handle aborts in is_parallel_of_n_reg_sets (PR68381)
Some users of is_parallel_of_n_reg_sets disregard the clobbers in a
parallel after it has returned "yes, this is a parallel of N sets and
maybe some clobbers". But combine uses a clobber of const0_rtx to
indicate substitution failure, so this leads to disaster.
Fix this by checking for such special clobbers in is_parallel_of_n_reg_sets.
PR rtl-optimization/68381
* combine.c (is_parallel_of_n_reg_sets): Return false if the pattern
is poisoned.
Modified:
trunk/gcc/ChangeLog
trunk/gcc/combine.c