This is the mail archive of the gcc-bugs@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Bug middle-end/63443] New: copyrename2 introducing bogus profile counts


https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63443

            Bug ID: 63443
           Summary: copyrename2 introducing bogus profile counts
           Product: gcc
           Version: 5.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: middle-end
          Assignee: unassigned at gcc dot gnu.org
          Reporter: tejohnson at google dot com

This problem showed up in https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63422,
where a new assert introduced in jump threading triggered because of insane
profile counts coming in. The LTO test case attached to that bug has profile
data, but the affected routine did not have profile counts and only frequencies
(main() from mozilla-xremote-client.ii which does not have a gcda file).
However, copyrename2 introduced some non-zero profile counts, which should not
happen.

To duplicate:
% g++ -flto -fprofile-use -fno-exceptions -std=gnu++0x -O2
mozilla-xremote-client.ii XRemoteClient.ii

I've attached the input files to this bug report as well.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]