[3.13] gh-126914: Store the Preallocated Thread State's Pointer in a PyInterpreterState Field - #127114
Merged
ericsnowcurrently merged 2 commits intoDec 2, 2024
Merged
Conversation
Copy link
Copy Markdown
Member
Sorry, something went wrong.
…yInterpreterState Field (pythongh-126989) This approach eliminates the originally reported race. It also gets rid of the deadlock reported in pythongh-96071, so we can remove the workaround added then.
Copy link
Copy Markdown
Member
Author
|
@Yhg1s, this changes only internal ABI, so we should be fine updating the ABI data file, right? |
Sorry, something went wrong.
Copy link
Copy Markdown
Member
|
Yes, the struct is always allocated by the runtime, and the new field is used only by the runtime, never extension modules, so updating the ABI definition is fine. |
Sorry, something went wrong.
Copy link
Copy Markdown
Member
|
Do you want to get this into 3.13.1? (It's scheduled for tomorrow :) |
Sorry, something went wrong.
ericsnowcurrently
enabled auto-merge (squash)
December 2, 2024 18:39
colesbury
added a commit
to colesbury/cpython
that referenced
this pull request
Apr 1, 2025
The 3.13 free threaded build immortalizes certain objects to avoid reference count contention. In pythongh-127114 the condition was unintentionally changed to happen when the first thread was created instead of the first non-main thread. The `interp->gc.immortalize` field is then cleared again during `_PyGC_Init()`. Change the condition so that we check if we should immortalize objects using deferred reference counting whenever a non-main thread is created.
Yhg1s
pushed a commit
that referenced
this pull request
Apr 7, 2025
gh-131988: Fix a multithreaded scaling regression The 3.13 free threaded build immortalizes certain objects to avoid reference count contention. In gh-127114 the condition was unintentionally changed to happen when the first thread was created instead of the first non-main thread. The `interp->gc.immortalize` field is then cleared again during `_PyGC_Init()`. Change the condition so that we check if we should immortalize objects using deferred reference counting whenever a non-main thread is created.
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
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.
This approach eliminates the originally reported race. It also gets rid of the deadlock reported in gh-96071, so we can remove the workaround added then.
This is mostly a cherry-pick of 1c0a104 (AKA gh-126989). The difference is we add PyInterpreterState.threads_preallocated at the end of PyInterpreterState, instead of adding PyInterpreterState.threads.preallocated. That avoids ABI disruption.