SIGSEGV when access a fork Event in a spawn Process

[kemingy]

BPO	45298
Nosy	@iritkatriel, @kemingy

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = None
created_at = <Date 2021-09-27.01:33:14.936>
labels = ['type-crash', 'docs']
title = 'SIGSEGV when access a fork Event in a spawn Process'
updated_at = <Date 2022-01-23.19:45:31.931>
user = 'https://github.com/kemingy'

bugs.python.org fields:

activity = <Date 2022-01-23.19:45:31.931>
actor = 'iritkatriel'
assignee = 'docs@python'
closed = False
closed_date = None
closer = None
components = ['Documentation']
creation = <Date 2021-09-27.01:33:14.936>
creator = 'kemingy'
dependencies = []
files = []
hgrepos = []
issue_num = 45298
keywords = []
message_count = 2.0
messages = ['402687', '411412']
nosy_count = 3.0
nosy_names = ['docs@python', 'iritkatriel', 'kemingy']
pr_nums = []
priority = 'normal'
resolution = None
stage = None
status = 'open'
superseder = None
type = 'crash'
url = 'https://bugs.python.org/issue45298'
versions = []

[kemingy]

Code to trigger this problem:

import multiprocessing as mp
from time import sleep


def wait_for_event(event):
    while not event.is_set():
        sleep(0.1)


def trigger_segment_fault():
    event = mp.get_context("fork").Event()
    p = mp.get_context("spawn").Process(target=wait_for_event, args=(event,))
    p.start()  # this will show the exitcode=-SIGSEGV
    sleep(1)
    print(p)
    event.set()
    p.terminate()


if __name__ == "__main__":
    trigger_segment_fault()

Accessing this forked event in a spawned process will result in a segment fault.

I have found a related report: https://bugs.python.org/issue43832. But I think it's not well documented in the Python 3 multiprocessing doc.

Will it be better to explicit indicate that the event is related to the start method context in the documentation?

[iritkatriel]

Which python version and system are you seeing this on?

On 3.11 on a Mac I don't get a segfault but rather

Process SpawnProcess-1:
Traceback (most recent call last):
  File "/Users/iritkatriel/src/cpython/Lib/multiprocessing/process.py", line 315, in _bootstrap
    self.run()
    ^^^^^^^^^^
  File "/Users/iritkatriel/src/cpython/Lib/multiprocessing/process.py", line 108, in run
    self._target(*self._args, **self._kwargs)
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/iritkatriel/src/cpython/tt.py", line 6, in wait_for_event
    while not event.is_set():
              ^^^^^^^^^^^^^^
  File "/Users/iritkatriel/src/cpython/Lib/multiprocessing/synchronize.py", line 328, in is_set
    with self._cond:
    ^^^^^^^^^^^^^^^^
  File "/Users/iritkatriel/src/cpython/Lib/multiprocessing/synchronize.py", line 230, in __enter__
    return self._lock.__enter__()
           ^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/iritkatriel/src/cpython/Lib/multiprocessing/synchronize.py", line 95, in __enter__
    return self._semlock.__enter__()
           ^^^^^^^^^^^^^^^^^^^^^^^^^
OSError: [Errno 9] Bad file descriptor
<SpawnProcess name='SpawnProcess-1' pid=29612 parent=29610 stopped exitcode=1>

[duaneg]

On current HEAD this prints a useful error, describing exactly what went wrong:

RuntimeError: A SemLock created in a fork context is being shared with a process in a spawn context. This is not supported. Please use the same context to create multiprocessing objects and Process.

I think this issue can be closed.