pthread segfaults after upgrade to RELENG_7

David E. Thiel lx at FreeBSD.org
Wed Oct 17 21:26:52 PDT 2007


Hello,

On a couple boxes, after upgrading from 6-STABLE to RELENG_7, a fair
number of apps segfault on execution even with compat6x installed. Is
this related to the libthr changeover with 7? I don't see anything in a
(quick) check of list archives or in UPDATING. A backtrace shows:

Program received signal SIGSEGV, Segmentation fault.
[Switching to LWP 100140]
0x287469d9 in pthread_join () from /lib/libthr.so.3
(gdb) bt
#0  0x287469d9 in pthread_join () from /lib/libthr.so.3
#1  0x28743c94 in pthread_self () from /lib/libthr.so.3
#2  0x28743f5e in pthread_rwlock_timedwrlock () from /lib/libthr.so.3
#3  0x28746bf1 in pthread_join () from /lib/libthr.so.3
#4  0x28743c94 in pthread_self () from /lib/libthr.so.3
#5  0x28743f5e in pthread_rwlock_timedwrlock () from /lib/libthr.so.3
#6  0x28746bf1 in pthread_join () from /lib/libthr.so.3
#7  0x28743c94 in pthread_self () from /lib/libthr.so.3
#8  0x28743f5e in pthread_rwlock_timedwrlock () from /lib/libthr.so.3

And so on ad infinitum. I've recently reproduced this with w3m and svn.
Recompiling the app fixes the problem, so I don't mind much, but this
seems like it will bite people, and it would be no fun on a production
server.

Thanks,
David


More information about the freebsd-current mailing list