In the Linux kernel, the following vulnerability has been resolved:
gtp: add synchronize_net() in gtp_newlink() error path to prevent use-after-free
gtp_newlink()'s error path frees tid_hash and addr_hash without
waiting for an RCU grace period after clearing sk_user_data. A
concurrent gtp_encap_recv() in softirq may still hold the gtp_dev
pointer obtained via rcu_dereference_sk_user_data() and access the
freed memory.
BUG: KASAN: slab-use-after-free in gtp0_pdp_find+0x1f6/0x200 (gtp.c:152)
Call Trace:
<IRQ>
gtp0_pdp_find+0x1f6/0x200
gtp_encap_recv+0x527/0x24b0
udp_queue_rcv_one_skb+0x75f/0xc10
Add synchronize_net() before the kfree calls in out_hashtable, which
covers all error paths from both gtp_encap_enable() and
gtp_create_sockets().
CVSS Vector: CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H
CVSS Score: 7.8
AV:L - The UAF is triggered from gtp_newlink()'s RTM_NEWLINK error path (ip link add type gtp), a local rtnetlink configuration interface. Concurrent gtp_encap_recv() in softirq only uses already-freed tid_hash/addr_hash; a remote GTP-U peer cannot cause the free.
AC:L - The attacker controls both sides of the RCU race: one thread issues RTM_NEWLINK that fails after setup_udp_tunnel_sock() (invalid second FD, IPv4/IPv6 family mismatch, or register_netdevice failure) while another floods UDP datagrams at the encap socket so gtp_encap_recv() still holds the gtp_dev pointer across the kfree.
PR:L - Creating a GTP device requires CAP_NET_ADMIN in the target netns via rtnl_link_get_net_capable()/netlink_ns_capable(). Unprivileged users obtain this through user+network namespaces (unshare -Urn); gtp_link_ops has no extra init-namespace check and MODULE_ALIAS_RTNL_LINK("gtp") permits module autoload.
UI:N - The attacker opens its own UDP sockets, sends packets to them, and issues the failing RTM_NEWLINK from its own process. No separate victim action such as mounting a filesystem or opening a file is required.
S:U - Impact is a host-kernel slab use-after-free in drivers/net/gtp.c, i.e. standard local privilege escalation within the same kernel security authority, not a KVM/Xen guest-to-host escape or IOMMU/DMA boundary bypass.
C:H - gtp0_pdp_find()/gtp1_pdp_find() walk the attacker-sized freed tid_hash hlist under RCU. Heap spraying that slab yields a fake pdp_ctx and an arbitrary kernel read primitive, so Confidentiality is High per UAF guidance.
I:H - A sprayed fake pdp_ctx is passed to gtp_rx(), which writes device stats, assigns skb->dev, and calls __netif_rx(); combined with UAF heap reuse this is an arbitrary write and control-flow hijack primitive, so Integrity is High.
A:H - KASAN reports slab-use-after-free in gtp0_pdp_find from gtp_encap_recv in softirq (udp_queue_rcv_one_skb). The UAF causes a kernel oops or panic even when not fully exploited, so Availability is High.
| Attack Vector |
Local |
Scope |
Unchanged |
| Attack Complexity |
Low |
Confidentiality Impact |
High |
| Privileges Required |
Low |
Integrity Impact |
High |
| User Interaction |
None |
Availability Impact |
High |
AV:L - The UAF is triggered from gtp_newlink()'s RTM_NEWLINK error path (ip link add type gtp), a local rtnetlink configuration interface. Concurrent gtp_encap_recv() in softirq only uses already-freed tid_hash/addr_hash; a remote GTP-U peer cannot cause the free.
AC:L - The attacker controls both sides of the RCU race: one thread issues RTM_NEWLINK that fails after setup_udp_tunnel_sock() (invalid second FD, IPv4/IPv6 family mismatch, or register_netdevice failure) while another floods UDP datagrams at the encap socket so gtp_encap_recv() still holds the gtp_dev pointer across the kfree.
PR:L - Creating a GTP device requires CAP_NET_ADMIN in the target netns via rtnl_link_get_net_capable()/netlink_ns_capable(). Unprivileged users obtain this through user+network namespaces (unshare -Urn); gtp_link_ops has no extra init-namespace check and MODULE_ALIAS_RTNL_LINK("gtp") permits module autoload.
UI:N - The attacker opens its own UDP sockets, sends packets to them, and issues the failing RTM_NEWLINK from its own process. No separate victim action such as mounting a filesystem or opening a file is required.
S:U - Impact is a host-kernel slab use-after-free in drivers/net/gtp.c, i.e. standard local privilege escalation within the same kernel security authority, not a KVM/Xen guest-to-host escape or IOMMU/DMA boundary bypass.
C:H - gtp0_pdp_find()/gtp1_pdp_find() walk the attacker-sized freed tid_hash hlist under RCU. Heap spraying that slab yields a fake pdp_ctx and an arbitrary kernel read primitive, so Confidentiality is High per UAF guidance.
I:H - A sprayed fake pdp_ctx is passed to gtp_rx(), which writes device stats, assigns skb->dev, and calls __netif_rx(); combined with UAF heap reuse this is an arbitrary write and control-flow hijack primitive, so Integrity is High.
A:H - KASAN reports slab-use-after-free in gtp0_pdp_find from gtp_encap_recv in softirq (udp_queue_rcv_one_skb). The UAF causes a kernel oops or panic even when not fully exploited, so Availability is High.
CVSS 3.1