segfaults in libthr

[Andrew Berry]

Hi,

I just upgraded a server from 7.2-RELEASE to 7.2-RELEASE-p2. I'm  
having issues with any port which links to libthr. This includes apr,  
subversion, and sqlite, so it's pretty common for httpd to segfault. I  
rebuilt libthr with debugging symbols, and I get the following  
backtrace. This is the result of an 'svn ls' operation.

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x28e60700 (LWP 100176)]
0x28287e24 in mutex_lock_common (curthread=0x28e60700, m=0x297b3100,  
abstime=0x0) at /usr/src/lib/libthr/thread/thr_mutex.c:422
422			ENQUEUE_MUTEX(curthread, m);
(gdb) bt
#0  0x28287e24 in mutex_lock_common (curthread=0x28e60700,  
m=0x297b3100, abstime=0x0) at /usr/src/lib/libthr/thread/thr_mutex.c:422
#1  0x28287d2f in __pthread_mutex_lock (mutex=0x28d56c00) at /usr/src/ 
lib/libthr/thread/thr_mutex.c:451
#2  0x28d16862 in pthreadMutexEnter () from /usr/local/lib/ 
libsqlite3.so.8
#3  0x28d1674b in sqlite3MutexInit () from /usr/local/lib/ 
libsqlite3.so.8
#4  0x28d14875 in sqlite3_initialize () from /usr/local/lib/ 
libsqlite3.so.8
#5  0x28c993fc in init_sqlite () from /usr/local/lib/libsvn_subr-1.so.0
#6  0x28c7b259 in svn_atomic__init_once () from /usr/local/lib/ 
libsvn_subr-1.so.0
#7  0x00000001 in ?? ()
#8  0x00000000 in ?? ()
#9  0x28cb22f0 in __JCR_LIST__ () from /usr/local/lib/libsvn_subr-1.so.0
#10 0x297bf020 in ?? ()
#11 0x297bf06c in ?? ()
#12 0xbfbfe538 in ?? ()
#13 0x28c9a179 in svn_sqlite__open () from /usr/local/lib/ 
libsvn_subr-1.so.0
Previous frame inner to this frame (corrupt stack?)

Anyone have any suggestions? I've all ready rebuilt affected ports  
without any change. Can I possibly rebuild the -p1 or -RELEASE version  
of libthr without horribly breaking the system?

Thanks,
--Andrew