• gevent version:
  21.12.0, wheels from pypi
• Python version: python 3.10.2
• Operating System: docker image 3.10.2-slim-bullseye

Description:

Hi, the problem appears suddenly today. Build stage has stall on manage.py compilemessages. strace shows only a infinite futex waiting.
Usually it's always worked ok. I don't know why exactly a behaviour is changed. But I found why it is stall.

Simple reproducible example ((not) worked only under docker)

from gevent import monkey
import gevent

if __name__ == '__main__':
    monkey.patch_thread();
    monkey.patch_subprocess();

    import subprocess
    from concurrent.futures import ThreadPoolExecutor

    def func():
        # gevent.sleep(0) # uncomment to fail
        return subprocess.run(["ls", "-l", "/dev/null"], capture_output=True)

    with ThreadPoolExecutor(max_workers=5) as executor:
        fut = executor.submit(func)
        print(fut.result())

This code will stall, but ctr+c give traceback:

^CException ignored in: <built-in method acquire of _thread.lock object at 0x7f1af2ab8800>
Traceback (most recent call last):
  File "/usr/local/lib/python3.10/site-packages/gevent/os.py", line 431, in fork_and_watch
    pid = fork()
KeyboardInterrupt:
Traceback (most recent call last):
  File "/var/www/cabinet/test.py", line 16, in <module>
    fut = executor.submit(func)
  File "/usr/local/lib/python3.10/concurrent/futures/thread.py", line 162, in submit
    with self._shutdown_lock, _global_shutdown_lock:
RuntimeError: release unlocked lock

If uncomment sleep in func all works fine. So problem occurs due to the code is not concurrent really.
ThreadPoolExecutor acquires a several locks in submit. CPython calls a some handlers before fork here. Particularly concurrent.futures.thread sets one. As the code are not runned concurrently and _global_shutdown_lock is the simple Lock, not RLock, so it will be deadlocked.