[Python-checkins] bpo-45274: Fix Thread._wait_for_tstate_lock() race condition (GH-28532) (GH-28580)

[vstinner]

https://github.com/python/cpython/commit/fae2694bea5e9e5a114af8cb40b60e7131a6340c
commit: fae2694bea5e9e5a114af8cb40b60e7131a6340c
branch: 3.10
author: Miss Islington (bot) <31488909+miss-islington at users.noreply.github.com>
committer: vstinner <vstinner at python.org>
date: 2021-09-27T14:53:31+02:00
summary:

bpo-45274: Fix Thread._wait_for_tstate_lock() race condition (GH-28532) (GH-28580)

Fix a race condition in the Thread.join() method of the threading
module. If the function is interrupted by a signal and the signal
handler raises an exception, make sure that the thread remains in a
consistent state to prevent a deadlock.
(cherry picked from commit a22be4943c119fecf5433d999227ff78fc2e5741)

Co-authored-by: Victor Stinner <vstinner at python.org>

files:
A Misc/NEWS.d/next/Library/2021-09-23-22-17-26.bpo-45274.gPpa4E.rst
M Lib/threading.py

diff --git a/Lib/threading.py b/Lib/threading.py
index 766011fa0312b..f39da145cd978 100644
--- a/Lib/threading.py
+++ b/Lib/threading.py
@@ -1100,11 +1100,24 @@ def _wait_for_tstate_lock(self, block=True, timeout=-1):
         # If the lock is acquired, the C code is done, and self._stop() is
         # called.  That sets ._is_stopped to True, and ._tstate_lock to None.
         lock = self._tstate_lock
-        if lock is None:  # already determined that the C code is done
+        if lock is None:
+            # already determined that the C code is done
             assert self._is_stopped
-        elif lock.acquire(block, timeout):
-            lock.release()
-            self._stop()
+            return
+
+        try:
+            if lock.acquire(block, timeout):
+                lock.release()
+                self._stop()
+        except:
+            if lock.locked():
+                # bpo-45274: lock.acquire() acquired the lock, but the function
+                # was interrupted with an exception before reaching the
+                # lock.release(). It can happen if a signal handler raises an
+                # exception, like CTRL+C which raises KeyboardInterrupt.
+                lock.release()
+                self._stop()
+            raise
 
     @property
     def name(self):
diff --git a/Misc/NEWS.d/next/Library/2021-09-23-22-17-26.bpo-45274.gPpa4E.rst b/Misc/NEWS.d/next/Library/2021-09-23-22-17-26.bpo-45274.gPpa4E.rst
new file mode 100644
index 0000000000000..94d06cef89b7b
--- /dev/null
+++ b/Misc/NEWS.d/next/Library/2021-09-23-22-17-26.bpo-45274.gPpa4E.rst
@@ -0,0 +1,5 @@
+Fix a race condition in the :meth:`Thread.join() <threading.Thread.join>`
+method of the :mod:`threading` module. If the function is interrupted by a
+signal and the signal handler raises an exception, make sure that the thread
+remains in a consistent state to prevent a deadlock. Patch by Victor
+Stinner.