{"id":"CVE-2026-98961","published":"2026-10-23T13:17:17.500","lastModified":"2026-10-23T13:17:17.500","description":"In the Linux kernel, the following vulnerability has been resolved:\n\nbtrfs: unlock inode and extent in caller when io_uring read extent fails\n\nbtrfs_uring_read_extent() runs only after btrfs_encoded_read() has taken\nthe inode shared lock and the extent lock.  On failure it used to unlock\nin out_fail, and a pages-array allocation failure returned -ENOMEM\nwithout unlocking at all.\n\nUnlock in the caller instead on all failure returns, matching the\ncopy_to_user() error path.  The deferred -EIOCBQUEUED path still unlocks\nin btrfs_uring_read_finished().","cvssScore":null,"cvssSeverity":null,"cvssVector":null,"cwes":[],"vendors":[],"products":[],"references":[{"url":"https://git.kernel.org/stable/c/2797eb38773a96914ccedeec71eb2144645400b2","tags":[]},{"url":"https://git.kernel.org/stable/c/ada1fc725a4db407294ae90a0d40ef1c46f457f2","tags":[]},{"url":"https://git.kernel.org/stable/c/b6e8add5d79d1f6f819a054c833bab452bd7365d","tags":[]}],"exploitRefs":[],"hasPoc":false,"ai":null}