CVE-2026-63921: ip6: vti: Use ip6_tnl.net in vti6_siocdevprivate().
In the Linux kernel, the following vulnerability has been resolved:
ip6: vti: Use ip6tnl.net in vti6siocdevprivate().
After patch 1/2 in this series, vti6update() unlinks and relinks the tunnel through t->net. vti6siocdevprivate() still uses devnet(dev) for the collision lookup. For a tunnel moved through IFLANETNSFD, devnet(dev) is the new netns, not t->net.
SIOCCHGTUNNEL on a migrated tunnel then runs:
net = devnet(dev) / migrated netns / t = vti6locate(net, &p1, false) / misses target in t->net / ... t = netdevpriv(dev) vti6update(t, &p1, false) / mutates t->net's hash /
A caller in the migrated netns picks params that match a tunnel in the creation netns. The lookup in devnet(dev) finds nothing. vti6update() prepends the migrated tunnel at the head of the creation netns hash bucket for those params. Later lookups in the creation netns resolve to the migrated device. xfrm receive delivers the matched packets through a device the caller controls.
Reachable from an unprivileged user namespace (unshare --user --map-root-user --net). Cross tenant scope on container hosts.
Switch the SIOCCHGTUNNEL path on a non fallback device to use t->net for the lookup. The lookup now matches the netns vti6update() operates on.
Also add nscapable(self->net->userns, CAPNETADMIN) before the lookup. The check at the top of the case is against devnet(dev)->userns, which after migration is the attacker's netns. A caller there can pick params absent from self->net, the lookup returns NULL, t becomes self, and vti6update() inserts the device into the creation netns hash. The new check requires CAPNETADMIN in the creation netns userns too.
SIOCADDTUNNEL and SIOCCHGTUNNEL on the fallback device keep devnet(dev), which equals initnet there.
Other sources
In the Linux kernel, the following vulnerability has been resolved:
ip6: vti: Use ip6tnl.net in vti6siocdevprivate().
After patch 1/2 in this series, vti6update() unlinks and relinks the tunnel through t-net. vti6siocdevprivate() still uses devnet(dev) for the collision lookup. For a tunnel moved through IFLANETNSFD, devnet(dev) is the new netns, not t-net.
SIOCCHGTUNNEL on a migrated tunnel then runs:
net = devnet(dev) / migrated netns / t = vti6locate(net, &p1, false) / misses target in t-net / ... t = netdevpriv(dev) vti6update(t, &p1, false) / mutates t-net's hash /
A caller in the migrated netns picks params that match a tunnel in the creation netns. The lookup in devnet(dev) finds nothing. vti6update() prepends the migrated tunnel at the head of the creation netns hash bucket for those params. Later lookups in the creation netns resolve to the migrated device. xfrm receive delivers the matched packets through a device the caller controls.
Reachable from an unprivileged user namespace (unshare --user --map-root-user --net). Cross tenant scope on container hosts.
Switch the SIOCCHGTUNNEL path on a non fallback device to use t-net for the lookup. The lookup now matches the netns vti6update() operates on.
Also add nscapable(self-net-userns, CAPNETADMIN) before the lookup. The check at the top of the case is against devnet(dev)-userns, which after migration is the attacker's netns. A caller there can pick params absent from self-net, the lookup returns NULL, t becomes self, and vti6update() inserts the device into the creation netns hash. The new check requires CAPNETADMIN in the creation netns userns too.
SIOCADDTUNNEL and SIOCCHGTUNNEL on the fallback device keep devnet(dev), which equals initnet there.
— IBM
Affected Software
Remediation
Recommended actions to resolve this vulnerability, in priority order.
- Compensating control
Require CAP_NET_ADMIN in the tunnel's creation network namespace user namespace before allowing SIOCCHGTUNNEL or related tunnel updates.