AI analysis
The Linux kernel's Open vSwitch module (net/openvswitch) contains a use-after-free in which a flow's mask is scheduled to be freed via RCU as soon as the flow is removed from the flow table, yet ovs_flow_cmd_fill_info() dereferences the stale flow->mask pointer afterwards — the RCU read lock is taken too late to protect it. A local attacker with the privileges needed to issue Open vSwitch netlink commands (CMD_DEL) can trigger this race, which widens whenever the kernel's memory allocation for the reply is slow; the result is a kernel slab use-after-free confirmed by KASAN that at minimum crashes the host and, per the CVSS 3.1 score of 7.8 (AV:L/AC:L/PR:L with C:H/I:H/A:H), could be developed into kernel memory corruption and full local privilege escalation. Affected systems are Linux machines running kernel code with the vulnerable openvswitch flow-deletion path — typically virtualization hosts and SDN data-plane nodes in OpenStack, OVN/Kubernetes, and telecom NFV environments — where an attacker already has local access and OVS administrative rights. The flaw was reported by Trend Micro's Zero Day Initiative (ZDI-CAN-32042), has no public proof of concept, is not in CISA's KEV catalog, and no exploitation in the wild is known. The fix reorders the code so that ovs_flow_tbl_remove() runs only after ovs_flow_cmd_fill_info() has finished using the flow.
What to do: Patch to a kernel release containing the upstream fix 'net: openvswitch: fix flow mask use-after-free on flow deletion' (CVE-2026-80994) as soon as your distribution ships it. If the openvswitch kernel module is not needed on a host, blacklist or unload it, and otherwise restrict which local users and containers can reach OVS netlink commands (CAP_NET_ADMIN) since exploitation requires local privileges. Monitor OVS data-plane nodes for kernel oopses or KASAN reports in ovs_flow_cmd_del/ovs_nla_put_key as an indicator of attempted triggering.
Affected
| Linux kernel (net/openvswitch — Open vSwitch kernel module) | Kernel versions containing the flow-mask RCU-free reordering introduced by the commit referenced in the upstream Fixes tag; the specific version range is not en |
Estimated exposure
mass≈ hundreds of thousands of Linux hosts with the Open vSwitch kernel module loaded (order-of-magnitude deployment estimate) — OVS/OVN is a common data plane for OpenStack, telecom NFV, and OVN-Kubernetes (the default networking in Red Hat OpenShift), so the kernel module plausibly runs on well over 100,000 hosts; no direct install counts exist, so this is a…
Order-of-magnitude estimate by the model from install counts, market share and public scan data it knows; verify before quoting.
Description
In the Linux kernel, the following vulnerability has been resolved: net: openvswitch: fix flow mask use-after-free on flow deletion The commit in the Fixes tag below made so flow->mask free is scheduled via RCU right after it is removed from the flow table. The pointer stays in the flow structure and it can be accessible while in the same RCU critical section. This is done to avoid requiring ovs_mutex for the ovs_flow_free(). However, while removing the flow during processing of CMD_DEL, we do not take RCU read lock before the removal, and ovs_flow_cmd_fill_info() uses the flow->mask pointer afterwards. The RCU read lock is taken, but it's already late at that point. The comment on that line acknowledges that the lock is cosmetic and doesn't serve a real purpose. This leads to use-after-free if the RCU grace period passes between removal and the filling. It is a short race window, but it is there and can lead to a real crash in case memory allocation for the info takes a bit longer: BUG: KASAN: slab-use-after-free in __ovs_nla_put_key net/openvswitch/flow_netlink.c:1996 BUG: KASAN: slab-use-after-free in ovs_nla_put_key+0x2463/0x2e30 net/openvswitch/flow_netlink.c:2250 Read of size 4 at addr ffff88801ee89970 by task ovs_flow_del_ec/9487 Call Trace: __ovs_nla_put_key net/openvswitch/flow_netlink.c:1996 ovs_nla_put_key+0x2463/0x2e30 net/openvswitch/flow_netlink.c:2250 ovs_flow_cmd_fill_info+0x420/0x9c0 net/openvswitch/datapath.c:930 ovs_flow_cmd_del+0x53a/0x970 net/openvswitch/datapath.c:1467 ... netlink_rcv_skb+0x156/0x420 net/netlink/af_netlink.c:2556 Allocated by task 9487: mask_alloc net/openvswitch/flow_table.c:967 flow_mask_insert net/openvswitch/flow_table.c:1012 ovs_flow_tbl_insert+0xea2/0x1a90 net/openvswitch/flow_table.c:1084 ovs_flow_cmd_new+0x7e3/0xd90 net/openvswitch/datapath.c:1086 ... netlink_rcv_skb+0x156/0x420 net/netlink/af_netlink.c:2556 Freed by task 9485: rcu_free_sheaf+0x1e/0x100 mm/slub.c:5978 rcu_do_batch kernel/rcu/tree.c:2645 rcu_core+0x59c/0x10c0 kernel/rcu/tree.c:2897 handle_softirqs+0x1e4/0x9a0 kernel/softirq.c:622 ... instr_sysvec_apic_timer_interrupt arch/x86/kernel/apic/apic.c:1062 ovs_flow_tbl_remove() must be called after the ovs_flow_cmd_fill_info() to avoid this race. This also helps with cleaning up the forced cast and the cosmetic RCU read lock. Before the commit in the Fixes tag the order did not matter as long as the flow object itself was not freed. A wider RCU critical section could be another option, but we have a GFP_KERNEL allocation in the way. Reported by Trend Micro's Zero Day Initiative as ZDI-CAN-32042.