The Linux kernel btrfs subsystem contains a flaw that can cause the filesystem to be set to read‑only when the kernel attempts to write back metadata on a zoned btrfs device. During a writeback sweep, the active metadata block group may need to be swapped in order to continue. If that block group still has pending I/O, the kernel refuses to wait for completion to avoid a possible deadlock. The resulting error causes the write transaction to abort, propagating a fatal state that forces the filesystem into read‑only mode. The impact is a loss of write capability and service interruption, but there is no direct path to arbitrary code execution or data exfiltration. The weakness is consistent with a synchronization or resource‑management error.

Affected systems are Linux kernel users running a btrfs filesystem configured for the zoned storage feature. The issue applies to any kernel version that contains the unpatched btrfs code; particular commits identified in the advisory (e2de2989 and ecc05eda9) indicate a wide range of releases. No specific kernel versions are listed in the advisory, so if a system uses the zoned option, it is potentially impacted until a patch is applied.

The executive summary notes an EPSS score of less than 1%, indicating a very low likelihood of exploitation under normal conditions. The vulnerability is not listed in CISA’s KEV catalog, further suggesting limited known exploitation. Attack vectors are inferred to be local or privileged: the flaw is triggered during internal filesystem writeback, which typically requires kernel memory access and thus is unlikely to be exploitable by remote users. Nonetheless, the denial-of-service nature makes any exploitation objective valuable to an attacker with local or elevated access.