Segfault with jemalloc 3.4

Ofer Samocha ofers at iMesh.com
Mon Oct 21 01:32:11 PDT 2013


Hi Jason,

Thanks for your prompt reply.

Here is some additional info on this matter:


*         We get crashes with different stack traces in jemalloc. Here is another, different example then the original one we sent you. Here it looks like we start with area_dalloc and not area_ralloc.


#0  0x0000000000aa4abe in arena_chunk_purge (arena=0x7f6bb14000c0, all=<value optimized out>) at src/arena.c:783
#1  arena_purge (arena=0x7f6bb14000c0, all=<value optimized out>) at src/arena.c:952
#2  0x0000000000aa56e7 in arena_dalloc_large_locked (arena=0x7f6bb14000c0, chunk=<value optimized out>, ptr=0x7f6b52248000) at src/arena.c:1770
#3  arena_dalloc_large (arena=0x7f6bb14000c0, chunk=<value optimized out>, ptr=0x7f6b52248000) at src/arena.c:1778
#4  0x00000000008ee9e8 in checked_delete<IMsgDecoder> (this=<value optimized out>, __in_chrg=<value optimized out>) at /usr/local/include/boost/checked_delete.hpp:34
#5  CMsgSockHandlerImpl::~CMsgSockHandlerImpl (this=<value optimized out>, __in_chrg=<value optimized out>) at /home/imesh/SFIM2Rel/src/comm/MsgSockHandler.cpp:252
#6  0x00000000008f2942 in checked_delete<CMsgSockHandlerImpl> (this=<value optimized out>) at /usr/local/include/boost/checked_delete.hpp:34
#7  operator() (this=<value optimized out>) at /home/imesh/SFIM2Rel/src/../ViberCore/include/sti/SmartPointers.h:21
#8  boost::detail::sp_counted_impl_pd<CMsgSockHandlerImpl*, sti::sti_deleter<CMsgSockHandlerImpl> >::dispose (this=<value optimized out>) at /usr/local/include/boost/smart_ptr/detail/sp_counted_impl.hpp:148
#9  0x00000000008e543c in release (this=0x7f6b8bf52780, __in_chrg=<value optimized out>) at /usr/local/include/boost/smart_ptr/detail/sp_counted_base_gcc_x86.hpp:145
#10 ~shared_count (this=0x7f6b8bf52780, __in_chrg=<value optimized out>) at /usr/local/include/boost/smart_ptr/detail/shared_count.hpp:305
#11 ~shared_ptr (this=0x7f6b8bf52780, __in_chrg=<value optimized out>) at /usr/local/include/boost/smart_ptr/shared_ptr.hpp:165
#12 reset (this=0x7f6b8bf52780, __in_chrg=<value optimized out>) at /usr/local/include/boost/smart_ptr/shared_ptr.hpp:382
#13 reset (this=0x7f6b8bf52780, __in_chrg=<value optimized out>) at /home/imesh/SFIM2Rel/src/../ViberCore/include/sti/SmartPointers.h:101
#14 CMsgSockHandler::~CMsgSockHandler (this=0x7f6b8bf52780, __in_chrg=<value optimized out>) at /home/imesh/SFIM2Rel/src/comm/MsgSockHandler.cpp:40
#15 0x00000000009cb591 in release (this=0x7f6bb102b300, sock=...) at /usr/local/include/boost/smart_ptr/detail/sp_counted_base_gcc_x86.hpp:145
#16 ~shared_count (this=0x7f6bb102b300, sock=...) at /usr/local/include/boost/smart_ptr/detail/shared_count.hpp:305
#17 ~shared_ptr (this=0x7f6bb102b300, sock=...) at /usr/local/include/boost/smart_ptr/shared_ptr.hpp:165
#18 ~shared_ptr (this=0x7f6bb102b300, sock=...) at /home/imesh/SFIM2Rel/src/../ViberCore/include/sti/SmartPointers.h:57
#19 sti::CDispatcher::doRemoveSocket (this=0x7f6bb102b300, sock=...) at /home/imesh/SFIM2Rel/src/infra/Dispatcher.cpp:863
#20 0x00000000009cb839 in sti::CDispatcher::RemoveDetachedSockets (this=0x7f6bb102b300) at /home/imesh/SFIM2Rel/src/infra/Dispatcher.cpp:900
#21 0x00000000009d7b09 in sti::CDispatcher::HandleEvents (this=0x7f6bb102b300, timeout=6) at /home/imesh/SFIM2Rel/src/infra/Dispatcher.cpp:590
#22 0x000000000080db3b in CServerHelper::HandleEvents (this=0x7f6bb1021020) at /home/imesh/SFIM2Rel/src/servercomm/ServerHelper.cpp:397
#23 0x00000000005c8e6e in main (argc=<value optimized out>, argv=<value optimized out>) at /home/imesh/SFIM2Rel/src/SN/SN.cpp:34



*         Our application in single threaded. To be more accurate, there are two threads - one the main thread and the other the log4cplus thread that samples the logger file every minute.



*         The compile flags we use are:


enable_autogen='0'
enable_debug='0'
enable_dss='0'
enable_experimental='0'
enable_fill='0'
enable_lazy_lock='0'
enable_mremap='0'
enable_munmap='0'
enable_prof='0'
enable_stats='0'
enable_tcache='1'
enable_tls='1'
enable_utrace='0'
enable_valgrind='0'
enable_xmalloc='0'
enable_zone_allocator=''



Thanks,
Ofer


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://jemalloc.net/mailman/jemalloc-discuss/attachments/20131021/bd5a739b/attachment.html>


More information about the jemalloc-discuss mailing list