netfilter: nfnetlink_queue: resolve clash for unconfirmed conntracks

In nfqueue, two consecutive skbuffs may race to create the conntrack
entry. Hence, the one that loses the race gets dropped due to clash in
the insertion into the hashes from the nf_conntrack_confirm() path.

This patch adds a new nf_conntrack_update() function which searches for
possible clashes and resolve them. NAT mangling for the packet losing
race is corrected by using the conntrack information that won race.

In order to avoid direct module dependencies with conntrack and NAT, the
nf_ct_hook and nf_nat_hook structures are used for this purpose.

Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>

1 files changed, 77 insertions, 0 deletions

diff --git a/net/netfilter/nf_conntrack_core.c b/net/netfilter/nf_conntrack_core.c
index 8d109d750073..3465da2a98bd 100644
--- a/net/netfilter/nf_conntrack_core.c
+++ b/net/netfilter/nf_conntrack_core.c
@@ -1607,6 +1607,82 @@ static void nf_conntrack_attach(struct sk_buff *nskb, const struct sk_buff *skb)
 	nf_conntrack_get(skb_nfct(nskb));
 }
 
+static int nf_conntrack_update(struct net *net, struct sk_buff *skb)
+{
+	const struct nf_conntrack_l3proto *l3proto;
+	const struct nf_conntrack_l4proto *l4proto;
+	struct nf_conntrack_tuple_hash *h;
+	struct nf_conntrack_tuple tuple;
+	enum ip_conntrack_info ctinfo;
+	struct nf_nat_hook *nat_hook;
+	unsigned int dataoff, status;
+	struct nf_conn *ct;
+	u16 l3num;
+	u8 l4num;
+
+	ct = nf_ct_get(skb, &ctinfo);
+	if (!ct || nf_ct_is_confirmed(ct))
+		return 0;
+
+	l3num = nf_ct_l3num(ct);
+	l3proto = nf_ct_l3proto_find_get(l3num);
+
+	if (l3proto->get_l4proto(skb, skb_network_offset(skb), &dataoff,
+				 &l4num) <= 0)
+		return -1;
+
+	l4proto = nf_ct_l4proto_find_get(l3num, l4num);
+
+	if (!nf_ct_get_tuple(skb, skb_network_offset(skb), dataoff, l3num,
+			     l4num, net, &tuple, l3proto, l4proto))
+		return -1;
+
+	if (ct->status & IPS_SRC_NAT) {
+		memcpy(tuple.src.u3.all,
+		       ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple.src.u3.all,
+		       sizeof(tuple.src.u3.all));
+		tuple.src.u.all =
+			ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple.src.u.all;
+	}
+
+	if (ct->status & IPS_DST_NAT) {
+		memcpy(tuple.dst.u3.all,
+		       ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple.dst.u3.all,
+		       sizeof(tuple.dst.u3.all));
+		tuple.dst.u.all =
+			ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple.dst.u.all;
+	}
+
+	h = nf_conntrack_find_get(net, nf_ct_zone(ct), &tuple);
+	if (!h)
+		return 0;
+
+	/* Store status bits of the conntrack that is clashing to re-do NAT
+	 * mangling according to what it has been done already to this packet.
+	 */
+	status = ct->status;
+
+	nf_ct_put(ct);
+	ct = nf_ct_tuplehash_to_ctrack(h);
+	nf_ct_set(skb, ct, ctinfo);
+
+	nat_hook = rcu_dereference(nf_nat_hook);
+	if (!nat_hook)
+		return 0;
+
+	if (status & IPS_SRC_NAT &&
+	    nat_hook->manip_pkt(skb, ct, NF_NAT_MANIP_SRC,
+				IP_CT_DIR_ORIGINAL) == NF_DROP)
+		return -1;
+
+	if (status & IPS_DST_NAT &&
+	    nat_hook->manip_pkt(skb, ct, NF_NAT_MANIP_DST,
+				IP_CT_DIR_ORIGINAL) == NF_DROP)
+		return -1;
+
+	return 0;
+}
+
 /* Bring out ya dead! */
 static struct nf_conn *
 get_next_corpse(int (*iter)(struct nf_conn *i, void *data),
@@ -2126,6 +2202,7 @@ err_cachep:
 }
 
 static struct nf_ct_hook nf_conntrack_hook = {
+	.update		= nf_conntrack_update,
 	.destroy	= destroy_conntrack,
 };