The btrfs zoned filesystem suffered a deadlock between metadata writeback and transaction commit. During the writeback loop the guard lock "zoned_meta_io_lock" was held across a call that could block on a transaction commit, while the commit path also required the same lock. The resulting circular wait caused the system to hang in certain workloads, such as those exercised by fstests generic/475, effectively denying service to the affected system. This issue stems from an improper lock release path, a classic concurrency flaw categorized by CWE-667 and CWE-674.

Any Linux kernel that implements btrfs with zoned mode, from the generic kernel repository before the safeguard was merged. The specific kernel release containing the patch is not enumerated in the provided data; it is present in the recent mainline after commit 18577e77c2c8adaadf1f7c6e9bcd1c0b14e5dcdd. All distributions shipping an unpatched kernel or a kernel that still enables zoned btrfs are therefore impacted.

The vulnerability lacks a CVSS score and EPSS entry, and it is not listed in CISA KEV. The risk is therefore primarily operational: a system administrator could observe a hang or unresponsiveness whenever metadata intensive operations trigger a transaction commit on a zoned btrfs volume. The likely attack vector is internal; an adversary with the ability to create or manipulate large metadata workloads on a zoned btrfs filesystem could trigger the deadlock. Since the bug is a kernel-level resource contention issue and not a privilege escalation or information leakage flaw, exploitation would not require elevated privileges beyond those normally needed to interact with the filesystem.