Skip to contentSkip to conversation

net: ipv6: Validate GSO SKB before finish IPv6 processing

PATCH Superseded Aya Levin 2 participants 2 messages Jan 5, 2021 – Jan 6, 2021

Subsystems (3)

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore

Aya Levin

author
#0 lore raw

PATCH v2 1/1 net: ipv6: Validate GSO SKB before finish IPv6 processing

There are cases where GSO segment's length exceeds the egress MTU. If so:
- Consume the SKB and its segments.
- Issue an ICMP packet with 'Packet Too Big' message containing the MTU, allowing the source host to reduce its Path MTU appropriately.

Note: These cases are handled in the same manner in IPv4 output finish. This patch aligns the behavior of IPv6 and the one of IPv4.

Fixes 9e50849054a4 ("netfilter: ipv6: move POSTROUTING invocation before fragmentation")
---
 net/ipv6/ip6_output.c | 41 ++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 40 insertions(+), 1 deletion(-)

Hi,

Please queue to -stable >= v2.6.35.

Thanks,
Aya

v2:
Addressed error: uninitialized symbol 'ret', reported by
kernel test robot <[email protected]> / Dan Carpenter <[email protected]>.

net/ipv6/ip6_output.c +40 -1
index 749ad72386b2..077d43af8226 100644
@@ -125,8 +125,43 @@ static int ip6_finish_output2(struct net *net, struct sock *sk, struct sk_buff *
 	return -EINVAL;
 }
 
+static int
+ip6_finish_output_gso_slowpath_drop(struct net *net, struct sock *sk,
+				    struct sk_buff *skb, unsigned int mtu)
+{
+	struct sk_buff *segs, *nskb;
+	netdev_features_t features;
+	int ret = 0;
+
+	/* Please see corresponding comment in ip_finish_output_gso
+	 * describing the cases where GSO segment length exceeds the
+	 * egress MTU.
+	 */
+	features = netif_skb_features(skb);
+	segs = skb_gso_segment(skb, features & ~NETIF_F_GSO_MASK);
+	if (IS_ERR_OR_NULL(segs)) {
+		kfree_skb(skb);
+		return -ENOMEM;
+	}
+
+	consume_skb(skb);
+
+	skb_list_walk_safe(segs, segs, nskb) {
+		int err;
+
+		skb_mark_not_on_list(segs);
+		err = ip6_fragment(net, sk, segs, ip6_finish_output2);
+		if (err && ret == 0)
+			ret = err;
+	}
+
+	return ret;
+}
+
 static int __ip6_finish_output(struct net *net, struct sock *sk, struct sk_buff *skb)
 {
+	unsigned int mtu;
+
 #if defined(CONFIG_NETFILTER) && defined(CONFIG_XFRM)
 	/* Policy lookup after SNAT yielded a new policy */
 	if (skb_dst(skb)->xfrm) {
@@ -135,7 +170,11 @@ static int __ip6_finish_output(struct net *net, struct sock *sk, struct sk_buff
 	}
 #endif
 
-	if ((skb->len > ip6_skb_dst_mtu(skb) && !skb_is_gso(skb)) ||
+	mtu = ip6_skb_dst_mtu(skb);
+	if (skb_is_gso(skb) && !skb_gso_validate_network_len(skb, mtu))
+		return ip6_finish_output_gso_slowpath_drop(net, sk, skb, mtu);
+
+	if ((skb->len > mtu && !skb_is_gso(skb)) ||
 	    dst_allfrag(skb_dst(skb)) ||
 	    (IP6CB(skb)->frag_max_size && skb->len > IP6CB(skb)->frag_max_size))
 		return ip6_fragment(net, sk, skb, ip6_finish_output2);
Signature · 1 line
2.14.1
maintainer
#1 lore raw
Quoted 1 line · On Tue, 5 Jan 2021 15:43:21 +0200 Aya Levin wrote:

There are cases where GSO segment's length exceeds the egress MTU.

Please name same for posterity.

Please widen the CC list.

Quoted 9 lines

If so:
- Consume the SKB and its segments.
- Issue an ICMP packet with 'Packet Too Big' message containing the MTU, allowing the source host to reduce its Path MTU appropriately.

Note: These cases are handled in the same manner in IPv4 output finish. This patch aligns the behavior of IPv6 and the one of IPv4.

Fixes 9e50849054a4 ("netfilter: ipv6: move POSTROUTING invocation before fragmentation")

Not sure if this commit actually adds the problem or just moves the problematic snippet around. But probably doesn't matter that much for 10+ years old code.

Quoted 2 lines

The code itself looks reasonable AFAICT.