This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug c++/82338] New: valgrind error in inherit_in_ebb
- From: "dcb314 at hotmail dot com" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Wed, 27 Sep 2017 07:49:23 +0000
- Subject: [Bug c++/82338] New: valgrind error in inherit_in_ebb
- Auto-submitted: auto-generated
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82338
Bug ID: 82338
Summary: valgrind error in inherit_in_ebb
Product: gcc
Version: 8.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: c++
Assignee: unassigned at gcc dot gnu.org
Reporter: dcb314 at hotmail dot com
Target Milestone: ---
Created attachment 42242
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42242&action=edit
C++ source code
I built a recent trunk version of gcc with valgrind.
I compiled the attached C++ code with flags -O3 -fPIC -fno-strict-aliasing -c
and got the following:
$ ~/gcc/results.253187.valgrind/bin/gcc -O3 -fPIC -fno-strict-aliasing -c
bug386.cc
==17152== Conditional jump or move depends on uninitialised value(s)
==17152== at 0x9DCE01: inherit_in_ebb (lra-constraints.c:6224)
==17152== by 0x9DCE01: lra_inheritance() (lra-constraints.c:6474)
==17152== by 0x9C8518: lra(_IO_FILE*) (lra.c:2430)
==17152== by 0x986161: do_reload (ira.c:5440)
The bug seems to have been introduced sometime before revision 249539.
svn blame claims that lra-constraints.c:6224 is as follows:
192719 vmakarov && usage_insns[regno].calls_num == calls_num - 1
I'll have a go at reducing the code.