226763 - nameshm1 test hangs

Status: RESOLVED DUPLICATE of bug 200162

(NSPR :: NSPR, defect)

Description

[Wan-Teh Chang]

This is NSPR 4.4.1, 32-bit HP-UX 11.00 debug build.

The nameshm1 test hangs.  The thread stacks are:

(gdb) info threads
* 2 system thread 14356  0xc01f4708 in __ksleep () from /usr/lib/libc.2
  1 system thread 5368  warning: reading register r3: No data
0xc00f2628 in __ksem_wait () from /usr/lib/librt.2
warning: reading register r3: No data
(gdb) thread 1
warning: reading register r3: No data
[Switching to thread 1 (system thread 5368)]
#0  0xc00f2628 in __ksem_wait () from /usr/lib/librt.2
(gdb) where
#0  0xc00f2628 in __ksem_wait () from /usr/lib/librt.2
#1  0xc00f18bc in sem_wait () from /usr/lib/librt.2
#2  0xc0f56888 in PR_WaitSemaphore (sem=0x400071b8)
    at ../../../../mozilla/nsprpub/pr/src/pthreads/ptsynch.c:730
#3  0x4a88 in ClientServerTest ()
    at ../../../mozilla/nsprpub/pr/tests/nameshm1.c:462
#4  0x562c in main (argc=1, argv=0x7f7f02b4)
    at ../../../mozilla/nsprpub/pr/tests/nameshm1.c:589
(gdb) thread 2
warning: reading register r3: No data
[Switching to thread 2 (system thread 14356)]
#0  0xc01f4708 in __ksleep () from /usr/lib/libc.2
(gdb) where
#0  0xc01f4708 in __ksleep () from /usr/lib/libc.2
#1  0xc0318448 in _lwp_cond_timedwait () from /usr/lib/libpthread.1
#2  0xc0318184 in pthread_cond_wait () from /usr/lib/libpthread.1
#3  0xc0f54dc4 in PR_WaitCondVar (cvar=0x40007300, timeout=4294967295)
    at ../../../../mozilla/nsprpub/pr/src/pthreads/ptsynch.c:389
#4  0xc0f6d048 in WaitPidDaemonThread (unused=0x0)
    at ../../../../../mozilla/nsprpub/pr/src/md/unix/uxproces.c:678
#5  0xc0f63ecc in _pt_root (arg=0x400073b0 "")
    at ../../../../mozilla/nsprpub/pr/src/pthreads/ptthread.c:214
#6  0xc0319f60 in __pthread_create_system () from /usr/lib/libpthread.1
(gdb)

Comment 1

[Wan-Teh Chang]

More debug info:

(gdb) thread 1
warning: reading register r3: No data
[Switching to thread 1 (system thread 5368)]
#0  0xc00f2628 in __ksem_wait () from /usr/lib/librt.2
(gdb) up
#1  0xc00f18bc in sem_wait () from /usr/lib/librt.2
(gdb) up
#2  0xc0f56888 in PR_WaitSemaphore (sem=0x400071b8)
    at ../../../../mozilla/nsprpub/pr/src/pthreads/ptsynch.c:730
../../../../mozilla/nsprpub/pr/src/pthreads/ptsynch.c:730: No such file or direc
tory.
(gdb) up
#3  0x4a88 in ClientServerTest ()
    at ../../../mozilla/nsprpub/pr/tests/nameshm1.c:462
462             rc = PR_WaitSemaphore( sem1 );
(gdb) print i
$1 = 0
(gdb) print optPing
$2 = 1000
(gdb) print addr
$3 = (unsigned int *) 0xc00fb000
(gdb) print *addr
$4 = 1
(gdb) print sem1
$5 = (struct PRSem *) 0x400071b8
(gdb) print *sem1
$6 = {sem = 0x400071d0}
(gdb) down
#2  0xc0f56888 in PR_WaitSemaphore (sem=0x400071b8)
    at ../../../../mozilla/nsprpub/pr/src/pthreads/ptsynch.c:730
../../../../mozilla/nsprpub/pr/src/pthreads/ptsynch.c:730: No such file or direc
tory.
(gdb) print sem
$7 = (struct PRSem *) 0x400071b8
(gdb) print *sem
$8 = {sem = 0x400071d0}
(gdb) print sem->sem
$9 = (struct _posix_sem *) 0x400071d0
(gdb) up
#3  0x4a88 in ClientServerTest ()
    at ../../../mozilla/nsprpub/pr/tests/nameshm1.c:462
462             rc = PR_WaitSemaphore( sem1 );
(gdb) up
#4  0x562c in main (argc=1, argv=0x7f7f02b4)
    at ../../../mozilla/nsprpub/pr/tests/nameshm1.c:589
589             ClientServerTest();
(gdb) print failed_already
$10 = 0
(gdb) print optClient
$11 = 0
(gdb) thread 2
warning: reading register r3: No data
[Switching to thread 2 (system thread 14356)]
#0  0xc01f4708 in __ksleep () from /usr/lib/libc.2
(gdb) up
#1  0xc0318448 in _lwp_cond_timedwait () from /usr/lib/libpthread.1
(gdb) up
#2  0xc0318184 in pthread_cond_wait () from /usr/lib/libpthread.1
(gdb) up
#3  0xc0f54dc4 in PR_WaitCondVar (cvar=0x40007300, timeout=4294967295)
    at ../../../../mozilla/nsprpub/pr/src/pthreads/ptsynch.c:389
../../../../mozilla/nsprpub/pr/src/pthreads/ptsynch.c:389: No such file or direc
tory.
(gdb) up
#4  0xc0f6d048 in WaitPidDaemonThread (unused=0x0)
    at ../../../../../mozilla/nsprpub/pr/src/md/unix/uxproces.c:678
../../../../../mozilla/nsprpub/pr/src/md/unix/uxproces.c:678: No such file or di
rectory.
(gdb) print pr_wp
$12 = {once = {initialized = 1, inProgress = 1, status = PR_SUCCESS},
  thread = 0x400073b0, ml = 0x40007258, numProcs = 0, cv = 0x40007300,
  pidTable = 0x400075e8}
(gdb)

Comment 2

[Wan-Teh Chang]

And this is the only 'nameshm1' process running when
the test hangs.  It seems that the 'nameshm1' client
process was created successfully but is gone already:

(gdb) thread 1
warning: reading register r3: No data
[Switching to thread 1 (system thread 5368)]
#0  0xc00f2628 in __ksem_wait () from /usr/lib/librt.2
(gdb) up
#1  0xc00f18bc in sem_wait () from /usr/lib/librt.2
(gdb) up
#2  0xc0f56888 in PR_WaitSemaphore (sem=0x400071b8)
    at ../../../../mozilla/nsprpub/pr/src/pthreads/ptsynch.c:730
../../../../mozilla/nsprpub/pr/src/pthreads/ptsynch.c:730: No such file or direc
tory.
(gdb) up
#3  0x4a88 in ClientServerTest ()
    at ../../../mozilla/nsprpub/pr/tests/nameshm1.c:462
462             rc = PR_WaitSemaphore( sem1 );
(gdb) print proc
$13 = (struct PRProcess *) 0x40007350
(gdb)

Note the non-NULL 'proc' in the 'nameshm1' server process.

Comment 3

[Wan-Teh Chang]

I think maybe the HP-UX 11.00 machine I'm using (sbshp1)
is missing some kernel patches, so interprocess semaphores
don't work.

Comment 4

[Wan-Teh Chang]

HP tracked this down.  This is caused by a bug in sem_open.
It also causes the semaping and semaerr tests to fail.

*** This bug has been marked as a duplicate of 200162 ***