Who is affected
This vulnerability affects the Linux Kernel, specifically the ebtables SNAT target component. The flaw permits an out-of-bounds write when processing ARP sender hardware address rewrites, and can be triggered via a splice-imported file page backing a socket buffer. The record notes that affected products may be end-of-life or end-of-service; you should treat this as a signal to prioritise systems running unsupported kernel versions.
How to check whether this touches you
- Inventory which systems run Linux and identify their kernel versions using
uname -ror equivalent package management queries. - Establish whether ebtables is loaded or in use on these systems; check with
lsmod | grep ebtableoriptables-legacy -Lwhere bridging and ARP filtering are configured. - Confirm network reachability to systems that satisfy the above: determine whether they accept inbound packets from untrusted networks, or whether they bridge networks where an attacker could send crafted ARP frames.
- Check kernel version against the vendor's advisory to establish whether your running version is known to contain the flaw; note that some distributions backport fixes to older version strings, so absence of a patch tag does not guarantee vulnerability.
What to do
- Prioritise according to BOD 26-04: systems exposed to untrusted networks, particularly those bridging external and internal segments, take precedence.
- Obtain the vendor's patched kernel version and test it in a non-production environment first, paying attention to any changes in ebtables behaviour or dependencies.
- If immediate patching is not possible, reduce reachability by disabling ebtables rules that perform SNAT on ARP frames, or by isolating affected systems from networks where untrusted ARP traffic originates.
- Enable logging of kernel oops, page faults and memory corruption warnings; direct these to a central log store with sufficient retention to support later forensic review.
- If the system is end-of-life and no patch is available, escalate to your change control board for evaluation of discontinuation or transition to a supported kernel version.
If you find you were exposed
Exploitation of out-of-bounds writes in kernel memory typically crashes the system or corrupts kernel state; look backwards through your logs for unexpected reboots, kernel panics, or memory corruption messages correlating with inbound ARP traffic. Review packet captures or tcpdump logs for malformed or suspicious ARP frames arriving before any incidents. Retain all kernel logs and dmesg output; log retention is often the limiting factor in retrospective detection, so prioritise systems you suspect were exposed.