gh-115832: Fix instrumentation version mismatch during interpreter shutdown (#115856)

A previous commit introduced a bug to `interpreter_clear()`: it set
`interp->ceval.instrumentation_version` to 0, without making the corresponding
change to `tstate->eval_breaker` (which holds a thread-local copy of the
version). After this happens, Python code can still run due to object finalizers
during a GC, and the version check in bytecodes.c will see a different result
than the one in instrumentation.c causing an infinite loop.

The fix itself is straightforward: clear `tstate->eval_breaker` when clearing
`interp->ceval.instrumentation_version`.
This commit is contained in:
Brett Simmers 2024-03-04 08:29:39 -08:00 committed by GitHub
parent 15dc2979bc
commit 0adfa8482d
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
4 changed files with 53 additions and 2 deletions

View file

@ -9,7 +9,8 @@ import textwrap
import types
import unittest
import asyncio
from test.support import requires_specialization
from test import support
from test.support import requires_specialization, script_helper
PAIR = (0,1)
@ -1858,3 +1859,12 @@ class TestTier2Optimizer(CheckEvents):
sys.monitoring.register_callback(TEST_TOOL, E.LINE, None)
sys.monitoring.set_events(TEST_TOOL, 0)
self.assertGreater(len(events), 250)
class TestMonitoringAtShutdown(unittest.TestCase):
def test_monitoring_live_at_shutdown(self):
# gh-115832: An object destructor running during the final GC of
# interpreter shutdown triggered an infinite loop in the
# instrumentation code.
script = support.findfile("_test_monitoring_shutdown.py")
script_helper.run_test_script(script)