Repository navigation
PyEval_AcquireLock() and PyEval_AcquireThread() do not handle runtime finalization properly. #80656
Description
Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.
Show more details
GitHub fields:
assignee = None
closed_at = <Date 2019-04-29.09:23:39.689>
created_at = <Date 2019-03-29.19:26:20.924>
labels = ['interpreter-core', 'type-bug', '3.8']
title = 'PyEval_AcquireLock() and PyEval_AcquireThread() do not handle runtime finalization properly.'
updated_at = <Date 2020-03-09.10:53:27.812>
user = 'https://github.com/ericsnowcurrently'bugs.python.org fields:
activity = <Date 2020-03-09.10:53:27.812>
actor = 'vstinner'
assignee = 'none'
closed = True
closed_date = <Date 2019-04-29.09:23:39.689>
closer = 'vstinner'
components = ['Interpreter Core']
creation = <Date 2019-03-29.19:26:20.924>
creator = 'eric.snow'
dependencies = []
files = []
hgrepos = []
issue_num = 36475
keywords = ['patch']
message_count = 7.0
messages = ['339138', '339365', '339876', '341054', '341055', '341389', '363714']
nosy_count = 3.0
nosy_names = ['vstinner', 'eric.snow', 'nanjekyejoannah']
pr_nums = ['12667', '13068']
priority = 'normal'
resolution = 'fixed'
stage = 'resolved'
status = 'closed'
superseder = None
type = 'behavior'
url = 'https://bugs.python.org/issue36475'
versions = ['Python 3.8']
Activity
ericsnowcurrently commentedon Mar 29, 2019
Daemon threads keep running until they finish or until finalization starts. For the latter, there is a check right after the thread acquires the GIL which causes the thread to exit if runtime finalization has started. [1] However, there are functions in the C-API that facilitate acquiring the GIL, but do not cause the thread to exit during finalization:
Daemon threads that acquire the GIL through these can cause a deadlock during finalization. (See issue bpo-36469.) They should probably be updated to match what PyEval_RestoreThread() does.
[1] see PyEval_RestoreThread() and the eval loop, in PyEval_EvalFrameEx()
nanjekyejoannah commentedon Apr 2, 2019
@eric.snow , you can review the PR I submitted for this.
nanjekyejoannah commentedon Apr 10, 2019
@eric do we need any tests for this?
vstinner commentedon Apr 29, 2019
New changeset f781d20 by Victor Stinner (Joannah Nanjekye) in branch 'master':
bpo-36475: Finalize PyEval_AcquireLock() and PyEval_AcquireThread() properly (GH-12667)
f781d20
vstinner commentedon Apr 29, 2019
I am not comfortable to backport this change to Python 3.7. It's too early to know how it will impact applications and how many complains we will get :-) If someone really wants to backport this scary change to 3.7, I would suggest to wait for 1 month after Python 3.8.0 final release.
I close the issue.
See bpo-36479 for the follow-up.
vstinner commentedon May 4, 2019
New changeset c664b34 by Victor Stinner in branch 'master':
bpo-36475: Make PyThread_exit_thread with _Py_NO_RETURN (GH-13068)
c664b34
vstinner commentedon Mar 9, 2020
I marked bpo-23592 as duplicate of this issue.