1-minute read
What a defender needs to know
A public proof of concept for CVE-2026-64531, or OVSwrap, gives local users root access on Linux hosts with Open vSwitch and conntrack kernel support. CIQ published its guidance on July 27. Sixteen days later, pre-derived offsets for approximately 760 kernel builds keep the issue actionable. A local user can reach the flaw through private namespaces; a container with CAP_NET_ADMIN can use an attacker-controlled network namespace. DPDK-only datapaths are outside this kernel path.
The kernel stores nested Open vSwitch actions as Netlink attributes with a 16-bit length field. A crafted action stream can make one nested attribute exceed that limit, corrupt kernel memory, and let the exploit change process credentials. The demonstration also creates a passwordless-root file under /etc/sudoers.d/.
Inventory the running kernel plus Open vSwitch and conntrack support. Apply the distribution’s fixed kernel and reboot. CIQ names fixed builds for RLC Pro, CIQ SIG/Cloud Next, and CIQ Linux Kernel; Rocky community users need the matching errata. If kernel OVS is unnecessary, block the module and reboot. Restricting unprivileged network namespaces blocks only the local-user path and does not protect containers retaining CAP_NET_ADMIN.
Review /etc/sudoers.d/, lsns -t net, process telemetry, audit logs, and OVS-related kernel events. uname -r must show the applicable fixed-or-newer build. A module block should leave lsmod empty for Open vSwitch and make modprobe -n -v openvswitch print install /bin/false; the namespace control should report user.max_net_namespaces = 0. Exposure review must account for unexplained privileged state and namespaces. A patched kernel prevents the next attempt; it cannot retroactively attest to the old one.