Audit all built-in modules for thread safety

[swtaarrs]

Feature or enhancement

Proposal:

We should audit every built-in module for thread safety and make any necessary fixes. This can be done separately from tagging the modules as safe using Py_mod_gil; any data races in built-in modules are considered bugs in the free-threaded build and not incompatibilities.

Below is a list of all source files that contain at least one module definition, as defined by files that contain either a call to PyModule_Create() or an array of PyModuleDef_Slots. If you'd like to help out with this task, take a look at one of the incomplete files. If the conversion is trivial, check it off and attach the PR to this issue. Otherwise, convert the line into an issue and assign it to yourself.

Every module is different, but here are a few high-level points to guide the work:

1. Use PyMutex as the basic unit of locking.
2. If you need to hold a lock for a longer period of time, especially across calls that may reenter Python code or acquire other locks, use critical sections.
3. If the module you're looking at is a thin wrapper around related code elsewhere in CPython (e.g., Modules/_codecsmodule.c and Python/codecs.c), you can also audit/convert the related non-module code. Otherwise, try to contain your work to just the module code, and create separate issues for any dependencies that aren't thread-safe.
4. Remember that C API functions are generally expected to handle thread-safety internally. C code that operates on Python objects using only C API calls is usually thread-safe by default (but look out for code that reads or modifies PyObject*s in C structs, since that needs synchronization).
5. Watch out for functions/macros that return borrowed references, like PyList_GetItem() or PyDict_GetItem(). If other threads could have references to the object, prefer functions like PyList_GetItemRef() or PyDict_GetItemRef() that return owned references (and will safely raise an error if the index/key doesn't exist).

Files to audit

Beta Give feedback

1. Modules/_abc.c
2. Modules/arraymodule.c
3. Modules/_asynciomodule.c
4. Modules/atexitmodule.c
5. Modules/binascii.c
6. Modules/_bisectmodule.c
7. Modules/_blake2/blake2module.c
8. Modules/_bz2module.c
9. Modules/cjkcodecs/cjkcodecs.h
10. Modules/cjkcodecs/multibytecodec.c
11. Modules/cmathmodule.c
12. Modules/_codecsmodule.c
13. Modules/_collectionsmodule.c
14. Modules/_contextvarsmodule.c
15. Modules/_csv.c
16. Modules/_ctypes/_ctypes.c
17. Modules/_ctypes/_ctypes_test.c
18. Modules/_cursesmodule.c
19. Modules/_curses_panel.c
20. Modules/_datetimemodule.c
21. Modules/_dbmmodule.c
22. Modules/_decimal/_decimal.c
23. Modules/_elementtree.c
24. Modules/errnomodule.c
25. Modules/faulthandler.c
26. Modules/fcntlmodule.c
27. Modules/_functoolsmodule.c
28. Modules/gcmodule.c
29. Modules/_gdbmmodule.c
30. Modules/grpmodule.c
31. Modules/_hashopenssl.c
32. Modules/_heapqmodule.c
33. Modules/_io/_iomodule.c
34. Modules/itertoolsmodule.c
35. Modules/_json.c
36. Modules/_localemodule.c
37. Modules/_lsprof.c
38. Modules/_lzmamodule.c
39. Modules/mathmodule.c
40. Modules/md5module.c
41. Modules/mmapmodule.c
42. Modules/_multiprocessing/multiprocessing.c
43. Modules/_multiprocessing/posixshmem.c
44. Modules/_opcode.c
45. Modules/_operator.c
46. Modules/overlapped.c
47. Modules/_pickle.c
48. Modules/posixmodule.c
49. Modules/_posixsubprocess.c
50. Modules/pwdmodule.c
51. Modules/pyexpat.c
52. Modules/_queuemodule.c
53. Modules/_randommodule.c
54. Modules/readline.c
55. Modules/resource.c
56. Modules/_scproxy.c
57. Modules/selectmodule.c
58. Modules/sha1module.c
59. Modules/sha2module.c
60. Modules/sha3module.c
61. Modules/signalmodule.c
62. Modules/socketmodule.c
63. Modules/_sqlite/module.c
64. Modules/_sre/sre.c
65. Modules/_ssl.c
66. Modules/_stat.c
67. Modules/_statisticsmodule.c
68. Modules/_struct.c
69. Modules/_suggestions.c
70. Modules/symtablemodule.c
71. Modules/_sysconfig.c
72. Modules/syslogmodule.c
73. Modules/termios.c
74. Modules/_testbuffer.c
75. Modules/_testcapimodule.c
76. Modules/_testclinic.c
77. Modules/_testclinic_limited.c
78. Modules/_testexternalinspection.c
79. Modules/_testimportmultiple.c
80. Modules/_testinternalcapi.c
81. Modules/_testlimitedcapi.c
82. Modules/_testmultiphase.c
83. Modules/_testsinglephase.c
84. Modules/_threadmodule.c
85. Modules/timemodule.c
86. Modules/_tkinter.c
87. Modules/_tracemalloc.c
88. Modules/_typingmodule.c
89. Modules/unicodedata.c
90. Modules/_uuidmodule.c
91. Modules/_weakref.c
92. Modules/_winapi.c
93. Modules/_xxinterpchannelsmodule.c
94. Modules/_xxinterpqueuesmodule.c
95. Modules/xxlimited_35.c
96. Modules/xxlimited.c
97. Modules/xxmodule.c
98. Modules/_xxsubinterpretersmodule.c
99. Modules/xxsubtype.c
100. Modules/_xxtestfuzz/_xxtestfuzz.c
101. Modules/zlibmodule.c
102. Modules/_zoneinfo.c
103. Objects/genobject.c (including ag_running_async)
104. PC/msvcrtmodule.c
105. PC/python3dll.c
106. PC/winreg.c
107. PC/winsound.c
108. Python/bltinmodule.c
109. Python/import.c
110. Python/instrumentation.c
111. Python/marshal.c
112. Python/Python-ast.c
113. Python/Python-tokenize.c
114. Python/sysmodule.c
115. Python/_warnings.c

Completed Issues

Completed Issues

Beta Give feedback

1. Make _warnings.c thread-safe in free-threaded build #116664
   topic-free-threading type-feature +2
   
2. Make _threadmodule.c thread-safe in --disable-gil builds #114271
   topic-free-threading type-feature +2
   
3. Make _struct module thread-safe in --disable-gil builds #112062
   3.13 extension-modules topic-free-threading type-feature +4
   
4. Make hashlib related modules thread-safe without the GIL #111916
   3.13 extension-modules topic-free-threading type-feature +4
   
5. socketmodule.c: use atomics to access defaulttimeout in free-threaded build #116616
   extension-modules topic-free-threading type-feature +3

Has this already been discussed elsewhere?

I have already discussed this feature proposal on Discourse

Links to previous discussion of this feature:

https://peps.python.org/pep-0703/, especially this section

Linked PRs

• gh-116738: Make _abc module thread-safe #117488
• gh-116738: Make _codecs module thread-safe #117530
• gh-116738: Make _csv module thread-safe #118344
• gh-116738: Make _json module safe in the free-threading build #119438
• gh-116738: Make the array built in module safe for free threaded python #120103

[mpage]

@swtaarrs - Is it OK if I remove you as the owner?

[swtaarrs]

Yep, I won't be doing any more work on this 😞

[eendebakpt]

I audited Modules/_statisticsmodule.c and it is already thread safe afaics (it only contains a single method with C input, conversion is handled by clinic generated code). Do we need to make a PR/issue for this module or can we check it off the list (once someone did a second look)?