[PATCH net v2] net: limit priority VLAN untag depth

[Kamil Sawicki <user148783812897@gmail.com>]

The priority VLAN receive path can repeatedly remove VID 0 tags and retry
packet processing. Unlike the read-only VLAN protocol parser, this path has
no limit on the number of nested tags it consumes.

Each skb_vlan_untag() advances skb->data and resets header offsets. The
offsets are stored in 16-bit fields. With enough nested tags, the network
or transport offset can wrap. A later protocol handler can then derive a
large pull length from the wrapped offsets and access memory outside the
skb allocation.

Apply the existing VLAN_MAX_DEPTH limit to the exceptional VID 0 retry.
Count the tag already stored in skb VLAN metadata as the first level, then
stop before untagging a ninth level. Keep the accounting inside this
uncommon branch so the regular VLAN receive path is unaffected.

A single VID 0 priority tag before a real VLAN tag is valid. No normal
deployment was found that intentionally emits multiple consecutive VID 0
tags; permitting eight levels is therefore conservative.

The issue was found while researching the Linux network receive path for
Google kvmCTF, with OpenAI Codex (gpt-5.6-sol) assisting the analysis. The
affected path, reproducer and fix were manually reviewed and tested.

Fixes: 36b2f61a42c2 ("net: handle 802.1P vlan 0 packets properly")
Cc: stable@vger.kernel.org
Reviewed-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Kamil Sawicki <user148783812897@gmail.com>
---
v2:
- rename vlan_depth to priority_tag_depth
- report SKB_DROP_REASON_RECURSION_LIMIT when the limit is reached
- address the expected real-world tag depth in the commit message

v1: https://lore.kernel.org/netdev/20261023003750.1133904-1-user148783812897@gmail.com/

Runtime-tested on net/main f4377bf85887 (7.3-rc7) with KASAN. The
patched kernel blocked the nested vhost/TAP trigger on five fresh boots;
each run recorded eight skb_vlan_untag() calls, no tcp_v4_rcv() call and
no KASAN report. The eight-tag packet was delivered, the nine-tag packet
was dropped, and kfree_skb recorded SKB_DROP_REASON_RECURSION_LIMIT for
all three nine-tag test transmissions.

The fragmented-head test stopped after eight untag calls without reaching
IPv4 or TCP. A 100,044-byte nonlinear GSO skb crossed two VID 0 tags, and
a 32 MiB BIG TCP transfer completed with matching hashes and skb lengths
up to 130,376 bytes. The 31-sample protocol-253 control recorded no matching
skb_copy_and_csum_bits() call and no shadow-probe hit.

Rebased onto current net/main 734bef548a96 without changing the patch-id.
Strict checkpatch and W=1 net/core/dev.o builds using allyesconfig and
allmodconfig completed without warnings.

 net/core/dev.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/net/core/dev.c b/net/core/dev.c
index 1a09e786911805..c9f07beee25782 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -6306,6 +6306,8 @@ static int __netif_receive_skb_core(struct sk_buff **pskb, bool pfmemalloc,
 	}
 
 	if (unlikely(skb_vlan_tag_present(skb)) && !netdev_uses_dsa(skb->dev)) {
+		unsigned int priority_tag_depth = 1;
+
 check_vlan_id:
 		if (skb_vlan_tag_get_id(skb)) {
 			/* Vlan id is non 0 and vlan_do_receive() above couldn't
@@ -6317,6 +6319,12 @@ static int __netif_receive_skb_core(struct sk_buff **pskb, bool pfmemalloc,
 			 * 802.1Q or 802.1AD and vlan_do_receive() above could
 			 * not find vlan dev for vlan id 0.
 			 */
+			if (unlikely(priority_tag_depth >= VLAN_MAX_DEPTH)) {
+				drop_reason = SKB_DROP_REASON_RECURSION_LIMIT;
+				goto drop;
+			}
+
+			priority_tag_depth++;
 			__vlan_hwaccel_clear_tag(skb);
 			skb = skb_vlan_untag(skb);
 			if (unlikely(!skb))

base-commit: 734bef548a96e8774d6eef5dfa2026c6edc4e80b
-- 
2.53.0