This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug tree-optimization/81556] [7/8 Regression] Wrong code at -O2
- From: "jakub at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Wed, 26 Jul 2017 08:35:39 +0000
- Subject: [Bug tree-optimization/81556] [7/8 Regression] Wrong code at -O2
- Auto-submitted: auto-generated
- References: <bug-81556-4@http.gcc.gnu.org/bugzilla/>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81556
Jakub Jelinek <jakub at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |ASSIGNED
Assignee|unassigned at gcc dot gnu.org |jakub at gcc dot gnu.org
Target Milestone|--- |7.2
Summary|Wrong code at -O2 |[7/8 Regression] Wrong code
| |at -O2
--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
But it is indeed true that this seems to be reassoc. That pass shouldn't be
resetting the range info, instead it shouldn't reuse SSA_NAMEs unless they must
have the same value as before. It has code to avoid the reusing, but most
likely this is a corner case that isn't handled well.
I'll have a look.