There was a deadlock originally seen by Memray when a daemon thread
enabled or disabled profiling while the interpreter was shutting down.
I think this could also happen with garbage collection, but I haven't
seen that in practice.
The daemon thread could be hung while trying acquire the global rwmutex
that prevents overlapping global and per-interpreter stop-the-world events.
Since it already held the main interpreter's stop-the-world lock, it
also deadlocked the main thread, which is trying to perform interpreter
finalization.
Swap the order of lock acquisition to prevent this deadlock.
Additionally, refactor `_PyParkingLot_Park` so that the global buckets
hashtable is left in a clean state if the thread is hung in
`PyEval_AcquireThread`.
(cherry picked from commit
90fe3250f82712b61630d636246c92df7c40c816)
Co-authored-by: Sam Gross <colesbury@gmail.com>
self.assertEqual(len(native_ids), 2)
self.assertNotEqual(native_ids[0], native_ids[1])
+ def test_stop_the_world_during_finalization(self):
+ # gh-137433: Test functions that trigger a stop-the-world in the free
+ # threading build concurrent with interpreter finalization.
+ script = """if True:
+ import gc
+ import sys
+ import threading
+ NUM_THREADS = 5
+ b = threading.Barrier(NUM_THREADS + 1)
+ def run_in_bg():
+ b.wait()
+ while True:
+ sys.setprofile(None)
+ gc.collect()
+
+ for _ in range(NUM_THREADS):
+ t = threading.Thread(target=run_in_bg, daemon=True)
+ t.start()
+
+ b.wait()
+ print("Exiting...")
+ """
+ rc, out, err = assert_python_ok('-c', script)
+ self.assertEqual(rc, 0)
+ self.assertEqual(err, b"")
+ self.assertEqual(out.strip(), b"Exiting...")
+
class ThreadJoinOnShutdown(BaseTestCase):
def _run_and_join(self, script):
--- /dev/null
+Fix a potential deadlock in the :term:`free threading` build when daemon
+threads enable or disable profiling or tracing while the main thread is
+shutting down the interpreter.
enqueue(bucket, addr, &wait);
_PyRawMutex_Unlock(&bucket->mutex);
- int res = _PySemaphore_Wait(&wait.sema, timeout_ns, detach);
+ PyThreadState *tstate = NULL;
+ if (detach) {
+ tstate = _PyThreadState_GET();
+ if (tstate && _PyThreadState_IsAttached(tstate)) {
+ // Only detach if we are attached
+ PyEval_ReleaseThread(tstate);
+ }
+ else {
+ tstate = NULL;
+ }
+ }
+
+ int res = _PySemaphore_Wait(&wait.sema, timeout_ns, 0);
if (res == Py_PARK_OK) {
goto done;
}
// Another thread has started to unpark us. Wait until we process the
// wakeup signal.
do {
- res = _PySemaphore_Wait(&wait.sema, -1, detach);
+ res = _PySemaphore_Wait(&wait.sema, -1, 0);
} while (res != Py_PARK_OK);
goto done;
}
done:
_PySemaphore_Destroy(&wait.sema);
+ if (tstate) {
+ PyEval_AcquireThread(tstate);
+ }
return res;
}
{
_PyRuntimeState *runtime = &_PyRuntime;
- PyMutex_Lock(&stw->mutex);
+ // gh-137433: Acquire the rwmutex first to avoid deadlocks with daemon
+ // threads that may hang when blocked on lock acquisition.
if (stw->is_global) {
_PyRWMutex_Lock(&runtime->stoptheworld_mutex);
}
else {
_PyRWMutex_RLock(&runtime->stoptheworld_mutex);
}
+ PyMutex_Lock(&stw->mutex);
HEAD_LOCK(runtime);
stw->requested = 1;
}
stw->requester = NULL;
HEAD_UNLOCK(runtime);
+ PyMutex_Unlock(&stw->mutex);
if (stw->is_global) {
_PyRWMutex_Unlock(&runtime->stoptheworld_mutex);
}
else {
_PyRWMutex_RUnlock(&runtime->stoptheworld_mutex);
}
- PyMutex_Unlock(&stw->mutex);
}
#endif // Py_GIL_DISABLED