[CRASH] Cannot access memory at address. Deadlock with b2b

[devoxy1]

We experience bug which causes opensips to be unresponsive for SIP requests. At some point opensips just stops processing any requests with timer errors.

OpenSIPS version you are running

version: opensips 3.4.9 (x86_64/linux)
flags: STATS: On, DISABLE_NAGLE, USE_MCAST, SHM_MMAP, PKG_MALLOC, Q_MALLOC, F_MALLOC, HP_MALLOC, DBG_MALLOC, FAST_LOCK-ADAPTIVE_WAIT
ADAPTIVE_WAIT_LOOPS=1024, MAX_RECV_BUFFER_SIZE 262144, MAX_LISTEN 16, MAX_URI_SIZE 1024, BUF_SIZE 65535
poll method support: poll, epoll, sigio_rt, select.
git revision: 448fdf959
main.c compiled on 12:37:50 Nov 19 2024 with gcc 11

Crash Core Dump
I'm attaching trap from opensips during deadlock, there are messages like "Cannot access memory".
gdb_opensips_20241119_142816.txt

Describe the traffic that generated the bug
At some point opensips starts to throw messages about timer task and stops responding to any SIP messages. Only restart helps to fix the issue.

To Reproduce

Relevant System Logs

 opensips[3230562]: WARNING:core:utimer_ticker: utimer task <tm-utimer> already scheduled 110 ms ago (now 129050 ms), delaying execution
 opensips[3230562]: WARNING:core:utimer_ticker: utimer task <tm-utimer> already scheduled 190 ms ago (now 129130 ms), delaying execution
 opensips[3230562]: WARNING:core:utimer_ticker: utimer task <tm-utimer> already scheduled 290 ms ago (now 129230 ms), delaying execution
 opensips[3230562]: WARNING:core:utimer_ticker: utimer task <tm-utimer> already scheduled 390 ms ago (now 129330 ms), delaying execution
 opensips[3230562]: WARNING:core:utimer_ticker: utimer task <tm-utimer> already scheduled 500 ms ago (now 129440 ms), delaying execution
 opensips[3230562]: WARNING:core:utimer_ticker: utimer task <tm-utimer> already scheduled 600 ms ago (now 129540 ms), delaying execution
 opensips[3230562]: WARNING:core:utimer_ticker: utimer task <tm-utimer> already scheduled 710 ms ago (now 129650 ms), delaying execution
 opensips[3230562]: WARNING:core:utimer_ticker: utimer task <tm-utimer> already scheduled 790 ms ago (now 129730 ms), delaying execution
 opensips[3230562]: WARNING:core:utimer_ticker: utimer task <tm-utimer> already scheduled 890 ms ago (now 129830 ms), delaying execution
 opensips[3230562]: WARNING:core:timer_ticker: timer task <tm-timer> already scheduled 990 ms ago (now 129830 ms), delaying execution
 opensips[3230562]: WARNING:core:utimer_ticker: utimer task <tm-utimer> already scheduled 1000 ms ago (now 129940 ms), delaying execution
 opensips[3230562]: WARNING:core:utimer_ticker: utimer task <tm-utimer> already scheduled 1190 ms ago (now 130130 ms), delaying execution
 opensips[3230562]: WARNING:core:utimer_ticker: utimer task <tm-utimer> already scheduled 1190 ms ago (now 130130 ms), delaying execution
 opensips[3230562]: WARNING:core:utimer_ticker: utimer task <tm-utimer> already scheduled 1290 ms ago (now 130230 ms), delaying execution

OS/environment information

• Operating System: Ubuntu 22.04.1 LTS
• OpenSIPS installation: git
• other relevant information:

Additional context
There were 2 opened report for previous versions, that hang on investigation and closed.
#3133
#3211

[dimebage]

i guess the issue here is use of dialog module specailly with call to create_dialog . the load_balancer module can also cause this as it internally create dialoge . this eventually lead to opensips segfault and we have been experiencing same issue for some time.