Segmentation fault in 3.11.5, frame and multithreading conflict

[vkhodygo]

Crash report

What happened?

The code is a stripped down version of https://github.com/mckinsey/causalnex. I tried to make it as short as possible, but that takes time I'm afraid.

test.zip

lldb output when I tun the script:

Process 94644 launched: '/home/vlad/.conda/envs/test_311_valid/bin/python' (x86_64)
Process 94644 stopped and restarted: thread 1 received signal: SIGCHLD
Process 94644 stopped and restarted: thread 1 received signal: SIGCHLD
Process 94644 stopped and restarted: thread 1 received signal: SIGCHLD
Process 94644 stopped and restarted: thread 1 received signal: SIGCHLD
Process 94644 stopped and restarted: thread 1 received signal: SIGCHLD
Process 94644 stopped and restarted: thread 1 received signal: SIGCHLD
Process 94644 stopped and restarted: thread 1 received signal: SIGCHLD
Process 94644 stopped and restarted: thread 1 received signal: SIGCHLD
Process 94644 stopped and restarted: thread 1 received signal: SIGCHLD
Process 94644 stopped and restarted: thread 1 received signal: SIGCHLD
Process 94644 stopped and restarted: thread 1 received signal: SIGCHLD
Process 94644 stopped
* thread #1, name = 'python', stop reason = signal SIGSEGV: invalid address (fault address: 0x10)
    frame #0: 0x000000000050ed50 python`_PyEval_EvalFrameDefault at ceval.c:3187:23

gdb extracts this from a coredump:

[Thread debugging using libthread_db enabled]                                                                                                                                                                        
Using host libthread_db library "/usr/lib/libthread_db.so.1".
Core was generated by `/home/vlad/.conda/envs/test_311_valid/bin/python /home/vlad/Work/IHW/_mwe/test.'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0  __pthread_kill_implementation (threadid=<optimized out>, signo=signo@entry=11, no_tid=no_tid@entry=0) at pthread_kill.c:44
44	     return INTERNAL_SYSCALL_ERROR_P (ret) ? INTERNAL_SYSCALL_ERRNO (ret) : 0;                                                                                                                              
[Current thread is 1 (Thread 0x7fd0437c2340 (LWP 92633))]

I have a feeling this might be related to #106883 and probably to #99729. The code works fine with the latest 3.10, I failed to test it with 3.12 though: there is an implicit torch dependency but no build for 3.12 yet.

CPython versions tested on:

3.10, 3.11

Operating systems tested on:

Linux

Output from running 'python -VV' on the command line:

Python 3.11.5 (main, Sep 11 2023, 13:54:46) [GCC 11.2.0]

[novialriptide]

Hello! I have narrowed it down from the python file that potential *= bbn_node.func(*arg_list) is the line of code that causes the segfault.

[vkhodygo]

Thanks, I appreciate that. However, it goes a bit deeper than that, and I'm not sure how to narrow it down better, without this much extra code.