summaryrefslogtreecommitdiffstats
path: root/Python/instruction_sequence.c
diff options
context:
space:
mode:
authorBrett Simmers <swtaarrs@users.noreply.github.com>2024-05-23 20:59:35 (GMT)
committerGitHub <noreply@github.com>2024-05-23 20:59:35 (GMT)
commitbe1dfccdf2c5c7671b8a549e969b8cf7d60d9936 (patch)
tree920d35829b0fb8d51399ff69760998165da704cb /Python/instruction_sequence.c
parentb30d30c747df2bf9f1614df8e76db2ffdb24fcd8 (diff)
downloadcpython-be1dfccdf2c5c7671b8a549e969b8cf7d60d9936.zip
cpython-be1dfccdf2c5c7671b8a549e969b8cf7d60d9936.tar.gz
cpython-be1dfccdf2c5c7671b8a549e969b8cf7d60d9936.tar.bz2
gh-118727: Don't drop the GIL in `drop_gil()` unless the current thread holds it (#118745)
`drop_gil()` assumes that its caller is attached, which means that the current thread holds the GIL if and only if the GIL is enabled, and the enabled-state of the GIL won't change. This isn't true, though, because `detach_thread()` calls `_PyEval_ReleaseLock()` after detaching and `_PyThreadState_DeleteCurrent()` calls it after removing the current thread from consideration for stop-the-world requests (effectively detaching it). Fix this by remembering whether or not a thread acquired the GIL when it last attached, in `PyThreadState._status.holds_gil`, and check this in `drop_gil()` instead of `gil->enabled`. This fixes a crash in `test_multiprocessing_pool_circular_import()`, so I've reenabled it.
Diffstat (limited to 'Python/instruction_sequence.c')
0 files changed, 0 insertions, 0 deletions