gh-135871: Fix needless spinning in _PyMutex_LockTimed (timeout==0) #135872
+3
−1
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
All commit authors signed the Contributor License Agreement. |
This comment was marked as resolved.
This comment was marked as resolved.
94f04f5 to
88e8a1c
Compare
This comment was marked as duplicate.
This comment was marked as duplicate.
88e8a1c to
45811fd
Compare
This comment was marked as duplicate.
This comment was marked as duplicate.
…==0) * Move the timeout == 0 guard outside the else so a non-blocking call returns immediately after a failed CAS instead of entering the spin loop. * Reload v on every spin iteration, allowing timed/blocking callers to notice an unlock promptly. No-GIL builds now honor the semantics of non-blocking attempts and avoid wasted CPU; GIL builds are unaffected (MAX_SPIN_COUNT == 0).
45811fd to
7296a7d
Compare
Misc/NEWS.d/next/Core and Builtins/2025-06-23-18-08-32.gh-issue-135871.50C528.rst
Outdated
Show resolved
Hide resolved
|
(Sorry Antoine, hit the wrong button!) |
vstinner
reviewed
Jun 24, 2025
Misc/NEWS.d/next/Core_and_Builtins/2025-06-23-18-08-32.gh-issue-135871.50C528.rst
Outdated
Show resolved
Hide resolved
colesbury
reviewed
Jun 24, 2025
Python/lock.c
Outdated
| @@ -89,6 +89,7 @@ _PyMutex_LockTimed(PyMutex *m, PyTime_t timeout, _PyLockFlags flags) | |||
| // Spin for a bit. | |||
| _Py_yield(); | |||
| spin_count++; | |||
| v = _Py_atomic_load_uint8_relaxed(&m->_bits); | |||
There was a problem hiding this comment.
This looks like it fixes a logic bug but it also hurts performance
See Tools/lockbench/lockbench.py
There was a problem hiding this comment.
thanks for the review! I'm convinced, this line reduces throughput on my machine by something like 40% for 8 threads. I think changing else if to if still makes sense though. I updated the PR.
colesbury
approved these changes
Jun 25, 2025
Misc/NEWS.d/next/Core_and_Builtins/2025-06-23-18-08-32.gh-issue-135871.50C528.rst
Outdated
Show resolved
Hide resolved
…e-135871.50C528.rst
|
Thanks @jtibbertsma for the PR, and @colesbury for merging it 🌮🎉.. I'm working now to backport this PR to: 3.13, 3.14. |
miss-islington
pushed a commit
to miss-islington/cpython
that referenced
this pull request
Jun 25, 2025
…ongh-135872) The free threading build could spin unnecessarily on `_Py_yield()` if the initial compare and swap failed. (cherry picked from commit cbfaf41) Co-authored-by: Joseph Tibbertsma <[email protected]>
|
Sorry, @jtibbertsma and @colesbury, I could not cleanly backport this to |
|
GH-135946 is a backport of this pull request to the 3.14 branch. |
|
GH-135946 is a backport of this pull request to the 3.14 branch. |
colesbury
pushed a commit
to colesbury/cpython
that referenced
this pull request
Jun 25, 2025
pythongh-135872) The free threading build could spin unnecessarily on `_Py_yield()` if the initial compare and swap failed. (cherry picked from commit cbfaf41) Co-authored-by: Joseph Tibbertsma <[email protected]>
|
GH-135947 is a backport of this pull request to the 3.13 branch. |
|
GH-135947 is a backport of this pull request to the 3.13 branch. |
colesbury
pushed a commit
that referenced
this pull request
Jun 25, 2025
…zero timeout (gh-135872) (gh-135946) The free threading build could spin unnecessarily on `_Py_yield()` if the initial compare and swap failed. (cherry picked from commit cbfaf41) Co-authored-by: Joseph Tibbertsma <[email protected]>
colesbury
added a commit
that referenced
this pull request
Jun 25, 2025
…zero timeout (gh-135872) (gh-135947) The free threading build could spin unnecessarily on `_Py_yield()` if the initial compare and swap failed. (cherry picked from commit cbfaf41) Co-authored-by: Joseph Tibbertsma <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Reload v on every spin iteration, allowing timed/blocking callers to notice an unlock promptly.No-GIL builds now honor the semantics of non-blocking attempts and avoid wasted CPU; GIL builds are unaffected (MAX_SPIN_COUNT == 0).
_PyMutex_LockTimedspins and may fail unnecessarily in no-GIL builds #135871