[Bug libc/16897] New: pthread_cond_timedwait wakes up early

[Bug libc/16897] New: pthread_cond_timedwait wakes up early

[te.anton at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=16897

            Bug ID: 16897
           Summary: pthread_cond_timedwait wakes up early
           Product: glibc
           Version: 2.18
            Status: NEW
          Severity: normal
          Priority: P2
         Component: libc
          Assignee: unassigned at sourceware dot org
          Reporter: te.anton at gmail dot com
                CC: drepper.fsp at gmail dot com

Created attachment 7583
  --> https://sourceware.org/bugzilla/attachment.cgi?id=7583&action=edit
test application in C to reproduce the bug

pthread_cond_timedwait wakes up early (1 second early it seems).

The bug is reproducible on:

* Fedora 20

$ cat /etc/issue
Fedora release 20 (Heisenbug)
Kernel \r on an \m (\l)

$ ldd --version
ldd (GNU libc) 2.18
Copyright (C) 2013 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
Written by Roland McGrath and Ulrich Drepper.

* Ubuntu 12.04.4 LTS

$ cat /etc/issue
Ubuntu 12.04.4 LTS \n \l

$ ldd --version
ldd (Ubuntu EGLIBC 2.15-0ubuntu10.5) 2.15
Copyright (C) 2012 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
Written by Roland McGrath and Ulrich Drepper.

* Ubuntu 13.04

$ cat /etc/issue
Ubuntu 13.04 \n \l

$ ldd --version
ldd (Ubuntu EGLIBC 2.17-0ubuntu5) 2.17
Copyright (C) 2012 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
Written by Roland McGrath and Ulrich Drepper.

Looks like libc/pthread 2.13 (fedora 14) and 2.10 does not have this bug. Also
the issued is not happening on my Mac Book Air.

Steps to reproduce:

* compile the test application
$ gcc -Wall -pipe -g -O0 bug_pthread_cond_timedwait.c -lpthread -o
bug_pthread_cond_timedwait
* run the application
./bug_pthread_cond_timedwait
* Result:
wait:1399055635 wakeup:1399055634
bug: main.c:54: main: Assertion `!"early wakeup"' failed.
Aborted (core dumped)
* Expected result:
OK
OK
switch, sleep 89383 ms
OK
OK
OK
switch, sleep 30886 ms
OK
...

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug libc/16897] pthread_cond_timedwait wakes up early

[sourceware at linglink dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=16897

Maurice Ling <sourceware at linglink dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sourceware at linglink dot com

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug libc/16897] pthread_cond_timedwait wakes up early

[sourceware at linglink dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=16897

--- Comment #1 from Maurice Ling <sourceware at linglink dot com> ---
Cross compiled for ARM does not have the timed wait problem:

$ ldd --version
ldd (crosstool-NG 1.18.0) 2.17
Copyright (C) 2012 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
Written by Roland McGrath and Ulrich Drepper.
$ uname -a
Linux nc1 2.6.37+ #2 Thu May 1 12:59:11 PDT 2014 armv7l GNU/Linux

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/16897] pthread_cond_timedwait wakes up early

[schwab@linux-m68k.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=16897

Andreas Schwab <schwab@linux-m68k.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|libc                        |nptl

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/16897] pthread_cond_timedwait wakes up early

[schwab@linux-m68k.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=16897

Andreas Schwab <schwab@linux-m68k.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|---                         |INVALID

--- Comment #2 from Andreas Schwab <schwab@linux-m68k.org> ---
The time function returns a value that is not precise enough to measure the
exact wakup time.  Use gettimeofday instead.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/16897] pthread_cond_timedwait wakes up early

[te.anton at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=16897

--- Comment #3 from Anton Te <te.anton at gmail dot com> ---
The issue is reproducible with gettimeofday. See the source code in the
attachment. There is the function mytime, you can change the implementation
from time() to gettimeofday().

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/16897] pthread_cond_timedwait wakes up early

[schwab@linux-m68k.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=16897

--- Comment #4 from Andreas Schwab <schwab@linux-m68k.org> ---
Must be a bug in your kernel then.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/16897] pthread_cond_timedwait wakes up early

[te.anton at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=16897

--- Comment #5 from Anton Te <te.anton at gmail dot com> ---
As you can see I tested it with different versions: Fedora 20, Ubuntu 12.04,
Ubuntu 13.04.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/16897] pthread_cond_timedwait wakes up early

[fweimer at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=16897

Florian Weimer <fweimer at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
              Flags|                            |security-

-- 
You are receiving this mail because:
You are on the CC list for the bug.