[Q] Strange stack dump when using jemalloc-3.4.0

[Taehwan Weon]

Hi, I am using jemalloc-3.4.0 on CentOs 5.4.
While running for a few hours, SEGV triggered with the following stack dump.
nc_dump_stack() is a signal handler.
As you see, a function in libc is the first jemalloc caller.
Does this mean some kind of memory corruption in jemalloc()?


------------------------ stack dump --------------------------------
nc_dump_stack at /root/trozan/netcache-core/trunk/netcache/cfs_apix.c:428
arena_chunk_purge at
/root/trozan/netcache-core/trunk/jemalloc-3.4.0/src/arena.c:783
?? at ??:0 (/lib64/libc.so.6) 0x30330)
malloc_mutex_unlock at
/root/trozan/netcache-core/trunk/jemalloc-3.4.0/include/jemalloc/internal/mutex.h:92
tcache_tsd_set at /root/trozan/netcache-core/trunk/jemalloc-3.4.0/include/
jemalloc/internal/tcache.h:143
?? at ??:0 (/lib64/libpthread.so.0) 0x5b19)
?? at ??:0 (/lib64/libc.so.6) 0xd503d)


Any hint would be highly appreciated!
Thanks in advance.

-- 
weon
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://jemalloc.net/mailman/jemalloc-discuss/attachments/20131103/a7810227/attachment.html>