Re: [PATCH RFC] futex: Fix might_sleep() warning in futex_pivot_pending()
Yao Kai <[email protected]>
| Newsgroups | dev.linux.lists.syzbot |
|---|---|
| Message-ID | <[email protected]> |
The MP ordering itself looks sound. However, hash_new is now intentionally accessed locklessly. Please use WRITE_ONCE() for all post-publication stores to hash_new so that the access annotations are consistent and the code remains clean under strict KCSAN. Thanks, Yao Kai