Impact
The Linux kernel’s rxrpc implementation contains a logic error in rxrpc_recvmsg_oob() where an out‑of‑band message that is only peeked is still moved onto the pending_oobq queue unconditionally. This double reference to the same sk_buff results in a double unlink and double free, producing a use‑after‑free bug. The memory corruption can cause a crash or disrupt kernel operation.
Affected Systems
All Linux kernel releases that have not incorporated the patch referenced in the advisory commits (5801cff7d5d7b4e9d877dfb627b23eb63167f02c, 5f470cc883416fea6d3bce18ef96bf91dd49ffc3, or 9ada3931beb37068fcb725b34b0398457009f343) are impacted. This spans distributions that ship with earlier kernels; the bug resides in the core rxrpc networking code and is not confined to a specific vendor branching.
Risk and Exploitability
Based on the description, it is inferred that the trigger requires a socket performing MSG_PEEK on a stateful rxrpc connection, which an attacker can craft if the target accepts untrusted rxrpc traffic. The EPSS score of <1% indicates a very low probability of exploitation, although the use‑after‑free nature suggests a high‑impact event if triggered. The vulnerability is not listed in CISA KEV, indicating no widespread exploitation has been documented, yet the kernel’s ubiquity presents a significant attack surface. The CVSS score of 9.8 indicates a critical severity level.
OpenCVE Enrichment