Bug 4861 - uninitialized data in the profiling code
Summary: uninitialized data in the profiling code
Status: RESOLVED WORKSFORME
Alias: None
Product: glibc
Classification: Unclassified
Component: libc (show other bugs)
Version: unspecified
: P2 normal
Target Milestone: ---
Assignee: Ulrich Drepper
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2007-07-28 10:50 UTC by Thomas Richter
Modified: 2014-07-04 16:08 UTC (History)
1 user (show)

See Also:
Host: x86_64-unknown-linux-gnu
Target: x86_64-unknown-linux-gnu
Build: x86_64-unknown-linux-gnu
Last reconfirmed:
fweimer: security-

Attachments
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Thomas Richter 2007-07-28 10:50:47 UTC 
When running multithreaded code on the above architecture using valgrind, it
detects at times the following problem:

==30381== Use of uninitialised value of size 8
==30381==    at 0x564C310: profil_counter (in /lib64/libc-2.5.so)
==30381==    by 0x55B95AF: (within /lib64/libc-2.5.so)

Apparently, something's wrong here. Unfortunately, this problem is not easy to
reproduce.

A potentially related problem seems to be that semaphores and profiling do not
seem to go well together in multithreaded code. With profiling enabled, some
semaphores seem to become spuriously available?

glibc release seems to be 2.5 on a Suse 10.2 (unfortunately with a custom kernel
patched by HP), running

gcc (GCC) 4.1.2 20061115 (prerelease) (SUSE Linux)
Comment 1 Ulrich Drepper 2007-07-29 22:48:03 UTC 
Either report this to your vendor who might be able to do something with the
address or provide exact information about the line in the source file.  If you
don't provide this information I'll close the bug shortly.
Comment 2 Ulrich Drepper 2007-10-02 04:27:52 UTC 
Not enough information.