[Bug nptl/1473] New: "pthread_cond_destroy()" hangs if called doubly on the same condvar on RHEL 4.0

[Bug nptl/1473] New: "pthread_cond_destroy()" hangs if called doubly on the same condvar on RHEL 4.0

[arati_ph at yahoo dot co dot in]

Hi,

I am facing a hang in  call : "pthread_cond_destroy()" , when it is called by 2
seperate processes on the same condition variable one after the other. Should it
give an error and exit ? What is the expected behaviour of such a scenario. ?
Also, this behaviour seems to occur only on RHEL 4.0 (gcc 3.4.3 )and does not
occur on RHEL 3.0 (gcc 3.2.3 ).

The test-case attached herewith passes on an RHEL 3.0 (gcc 3.2.3 ) machine but
hangs on RHEL 4.0 (gcc 3.4.3 )with the following stack :
To run the test-case, run it with a positive numerical parameter  indicating the
number of loops.
eg :      ./a.out 5

*** STACK ******
#0  0x00bc07a2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2
#1  0x0012522e in __lll_mutex_lock_wait () from /lib/tls/libpthread.so.0
#2  0x00122a8a in _L_mutex_lock_22 () from /lib/tls/libpthread.so.0
#3  0xbffb3294 in ?? ()
#4  0xbffb3220 in ?? ()
#5  0xbffb3208 in ?? ()
#6  0x00bcba80 in _dl_runtime_resolve () from /lib/ld-linux.so.2
#7  0x08049075 in main ()

*** STACK ******


Any pointers will be appreciated.

Thanks
Arati

-- 
           Summary: "pthread_cond_destroy()" hangs if called doubly on the
                    same condvar on RHEL 4.0
           Product: glibc
           Version: 2.2.5
            Status: NEW
          Severity: normal
          Priority: P2
         Component: nptl
        AssignedTo: drepper at redhat dot com
        ReportedBy: arati_ph at yahoo dot co dot in
                CC: glibc-bugs at sources dot redhat dot com
 GCC build triplet: gcc 3.4.3 on RHEL 4.0
  GCC host triplet: gcc 3.2.3 on RHEL 3.0
GCC target triplet: gcc 3.4.3 on RHEL 4.0


http://sourceware.org/bugzilla/show_bug.cgi?id=1473

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug nptl/1473] "pthread_cond_destroy()" hangs if called doubly on the same condvar on RHEL 4.0

[arati_ph at yahoo dot co dot in]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
            Version|2.2.5                       |2.3.4


http://sourceware.org/bugzilla/show_bug.cgi?id=1473

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug nptl/1473] "pthread_cond_destroy()" hangs if called doubly on the same condvar on RHEL 4.0

[arati_ph at yahoo dot co dot in]

------- Additional Comments From arati_ph at yahoo dot co dot in  2005-10-14 05:49 -------
Created an attachment (id=709)
 --> (http://sourceware.org/bugzilla/attachment.cgi?id=709&action=view)
Program to simulate the hang for "pthread_cond_destroy()" when called twice.


-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=1473

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug nptl/1473] "pthread_cond_destroy()" hangs if called doubly on the same condvar on RHEL 4.0

[drepper at redhat dot com]

------- Additional Comments From drepper at redhat dot com  2005-10-15 20:42 -------
It is not allowed to call pthread_cond_destroy twice.  After the first call the
memory is basically some uninitialized memory region.  It is a program bug.

Having said that, the current glibc version does try to help buggy programs a
bit.  You can check it out but if it does not work, don't ask for more changes.
 You have to fix your code instead.

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


http://sourceware.org/bugzilla/show_bug.cgi?id=1473

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.