This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug tree-optimization/69376] [6 Regression] wrong code at -Os and above on x86_64-linux-gnu
- From: "bernds at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Thu, 21 Jan 2016 19:02:40 +0000
- Subject: [Bug tree-optimization/69376] [6 Regression] wrong code at -Os and above on x86_64-linux-gnu
- Auto-submitted: auto-generated
- References: <bug-69376-4 at http dot gcc dot gnu dot org/bugzilla/>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69376
Bernd Schmidt <bernds at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |bernds at gcc dot gnu.org,
| |rguenther at suse dot de
--- Comment #2 from Bernd Schmidt <bernds at gcc dot gnu.org> ---
Either one of the PR69378 or PR69345 fixes makes this disappear, but I'm not
sure it's fixed. The key difference for this testcase here seems to be in the
vrp2 dumps:
Intersecting
[0, 127]
and
- [0, 127]
+ ~[0, 127]
to
- [0, 127]
+ UNDEFINED
I noticed that the 69117 patch adds copies of SSA_NAME_RANGE_INFO, but does not
copy SSA_NAME_RANGE_TYPE. Could that be related to the problem?
I'm somewhat confused by the lack of a range type in struct range_info_def.