gh-91051: fix segfault when using all 8 type watchers - #107853
Merged
carljm merged 2 commits intoAug 11, 2023
Merged
Conversation
Copy link
Copy Markdown
Member
Sorry, something went wrong.
Copy link
Copy Markdown
Member
Author
|
@Yhg1s I think this fix should be backported to 3.12. |
Sorry, something went wrong.
carljm
force-pushed
the
typewatchersign
branch
from
7c249fa to
e149aed
Compare
August 11, 2023 00:47
itamaro
approved these changes
Aug 11, 2023
itamaro
left a comment
Copy link
Copy Markdown
Contributor
There was a problem hiding this comment.
lgtm! we're going to want to backport to 3.12!
Sorry, something went wrong.
Comment thread
Lib/test/test_capi/test_watchers.py
Copy link
Copy Markdown
Contributor
|
Closing and re-opening to retrigger CLA checks. Sorry for the noise. |
Sorry, something went wrong.
Copy link
Copy Markdown
Contributor
|
Thanks @carljm for the PR 🌮🎉.. I'm working now to backport this PR to: 3.12. |
Sorry, something went wrong.
miss-islington
pushed a commit
to miss-islington/cpython
that referenced
this pull request
Aug 11, 2023
Copy link
Copy Markdown
|
GH-107876 is a backport of this pull request to the 3.12 branch. |
Sorry, something went wrong.
Yhg1s
pushed a commit
that referenced
this pull request
Aug 16, 2023
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
Reviewers
5 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The type watcher implementation adds char tp_watched; to the PyTypeObject struct, and then in PyType_Modified it assigns int bits = type->tp_watched; and then treats bits as a bitset, expecting that only the low 8 bits should ever be set.
On platforms where char is signed (e.g. x86), if the top bit in tp_watched is set, the cast to int bits will sign-extend with 1s, and we will thus have a lot of high bits set in bits that are not supposed to ever be set.
On x86 in a debug build, without the fix in this PR, the new test hits the assertion in PyType_Modified that only bit positions < TYPE_MAX_WATCHERS should be set, and aborts.
The fix is simple: tp_watched should be defined as unsigned char so that it is zero-extended, not sign-extended.
Note that the test is carefully designed to not assume there are no type watchers already active. It doesn't attempt to register a fixed number of type watchers, instead it just registers type watchers until it reaches TYPE_MAX_WATCHERS - 1, the last available slot and the one that can trigger this bug (since it will set the highest bit in tp_watched when watching a type.)