1 files changed, 81 insertions, 77 deletions

diff --git a/net/netfilter/nf_conntrack_proto_icmp.c b/net/netfilter/nf_conntrack_proto_icmp.c
index de64d8a5fdfd..9becac953587 100644
--- a/net/netfilter/nf_conntrack_proto_icmp.c
+++ b/net/netfilter/nf_conntrack_proto_icmp.c
@@ -25,8 +25,8 @@
 
 static const unsigned int nf_ct_icmp_timeout = 30*HZ;
 
-static bool icmp_pkt_to_tuple(const struct sk_buff *skb, unsigned int dataoff,
-			      struct net *net, struct nf_conntrack_tuple *tuple)
+bool icmp_pkt_to_tuple(const struct sk_buff *skb, unsigned int dataoff,
+		       struct net *net, struct nf_conntrack_tuple *tuple)
 {
 	const struct icmphdr *hp;
 	struct icmphdr _hdr;
@@ -54,8 +54,8 @@ static const u_int8_t invmap[] = {
 	[ICMP_ADDRESSREPLY] = ICMP_ADDRESS + 1
 };
 
-static bool icmp_invert_tuple(struct nf_conntrack_tuple *tuple,
-			      const struct nf_conntrack_tuple *orig)
+bool nf_conntrack_invert_icmp_tuple(struct nf_conntrack_tuple *tuple,
+				    const struct nf_conntrack_tuple *orig)
 {
 	if (orig->dst.u.icmp.type >= sizeof(invmap) ||
 	    !invmap[orig->dst.u.icmp.type])
@@ -68,11 +68,10 @@ static bool icmp_invert_tuple(struct nf_conntrack_tuple *tuple,
 }
 
 /* Returns verdict for packet, or -1 for invalid. */
-static int icmp_packet(struct nf_conn *ct,
-		       struct sk_buff *skb,
-		       unsigned int dataoff,
-		       enum ip_conntrack_info ctinfo,
-		       const struct nf_hook_state *state)
+int nf_conntrack_icmp_packet(struct nf_conn *ct,
+			     struct sk_buff *skb,
+			     enum ip_conntrack_info ctinfo,
+			     const struct nf_hook_state *state)
 {
 	/* Do not immediately delete the connection after the first
 	   successful reply to avoid excessive conntrackd traffic
@@ -104,53 +103,94 @@ static int icmp_packet(struct nf_conn *ct,
 	return NF_ACCEPT;
 }
 
-/* Returns conntrack if it dealt with ICMP, and filled in skb fields */
-static int
-icmp_error_message(struct nf_conn *tmpl, struct sk_buff *skb,
-		   const struct nf_hook_state *state)
+/* Check inner header is related to any of the existing connections */
+int nf_conntrack_inet_error(struct nf_conn *tmpl, struct sk_buff *skb,
+			    unsigned int dataoff,
+			    const struct nf_hook_state *state,
+			    u8 l4proto, union nf_inet_addr *outer_daddr)
 {
 	struct nf_conntrack_tuple innertuple, origtuple;
-	const struct nf_conntrack_l4proto *innerproto;
 	const struct nf_conntrack_tuple_hash *h;
 	const struct nf_conntrack_zone *zone;
 	enum ip_conntrack_info ctinfo;
 	struct nf_conntrack_zone tmp;
+	union nf_inet_addr *ct_daddr;
+	enum ip_conntrack_dir dir;
+	struct nf_conn *ct;
 
 	WARN_ON(skb_nfct(skb));
 	zone = nf_ct_zone_tmpl(tmpl, skb, &tmp);
 
 	/* Are they talking about one of our connections? */
-	if (!nf_ct_get_tuplepr(skb,
-			       skb_network_offset(skb) + ip_hdrlen(skb)
-						       + sizeof(struct icmphdr),
-			       PF_INET, state->net, &origtuple)) {
-		pr_debug("icmp_error_message: failed to get tuple\n");
+	if (!nf_ct_get_tuplepr(skb, dataoff,
+			       state->pf, state->net, &origtuple))
 		return -NF_ACCEPT;
-	}
-
-	/* rcu_read_lock()ed by nf_hook_thresh */
-	innerproto = __nf_ct_l4proto_find(origtuple.dst.protonum);
 
 	/* Ordinarily, we'd expect the inverted tupleproto, but it's
 	   been preserved inside the ICMP. */
-	if (!nf_ct_invert_tuple(&innertuple, &origtuple, innerproto)) {
-		pr_debug("icmp_error_message: no match\n");
+	if (!nf_ct_invert_tuple(&innertuple, &origtuple))
 		return -NF_ACCEPT;
-	}
-
-	ctinfo = IP_CT_RELATED;
 
 	h = nf_conntrack_find_get(state->net, zone, &innertuple);
-	if (!h) {
-		pr_debug("icmp_error_message: no match\n");
+	if (!h)
+		return -NF_ACCEPT;
+
+	/* Consider: A -> T (=This machine) -> B
+	 *   Conntrack entry will look like this:
+	 *      Original:  A->B
+	 *      Reply:     B->T (SNAT case) OR A
+	 *
+	 * When this function runs, we got packet that looks like this:
+	 * iphdr|icmphdr|inner_iphdr|l4header (tcp, udp, ..).
+	 *
+	 * Above nf_conntrack_find_get() makes lookup based on inner_hdr,
+	 * so we should expect that destination of the found connection
+	 * matches outer header destination address.
+	 *
+	 * In above example, we can consider these two cases:
+	 *  1. Error coming in reply direction from B or M (middle box) to
+	 *     T (SNAT case) or A.
+	 *     Inner saddr will be B, dst will be T or A.
+	 *     The found conntrack will be reply tuple (B->T/A).
+	 *  2. Error coming in original direction from A or M to B.
+	 *     Inner saddr will be A, inner daddr will be B.
+	 *     The found conntrack will be original tuple (A->B).
+	 *
+	 * In both cases, conntrack[dir].dst == inner.dst.
+	 *
+	 * A bogus packet could look like this:
+	 *   Inner: B->T
+	 *   Outer: B->X (other machine reachable by T).
+	 *
+	 * In this case, lookup yields connection A->B and will
+	 * set packet from B->X as *RELATED*, even though no connection
+	 * from X was ever seen.
+	 */
+	ct = nf_ct_tuplehash_to_ctrack(h);
+	dir = NF_CT_DIRECTION(h);
+	ct_daddr = &ct->tuplehash[dir].tuple.dst.u3;
+	if (!nf_inet_addr_cmp(outer_daddr, ct_daddr)) {
+		if (state->pf == AF_INET) {
+			nf_l4proto_log_invalid(skb, state->net, state->pf,
+					       l4proto,
+					       "outer daddr %pI4 != inner %pI4",
+					       &outer_daddr->ip, &ct_daddr->ip);
+		} else if (state->pf == AF_INET6) {
+			nf_l4proto_log_invalid(skb, state->net, state->pf,
+					       l4proto,
+					       "outer daddr %pI6 != inner %pI6",
+					       &outer_daddr->ip6, &ct_daddr->ip6);
+		}
+		nf_ct_put(ct);
 		return -NF_ACCEPT;
 	}
 
-	if (NF_CT_DIRECTION(h) == IP_CT_DIR_REPLY)
+	ctinfo = IP_CT_RELATED;
+	if (dir == IP_CT_DIR_REPLY)
 		ctinfo += IP_CT_IS_REPLY;
 
 	/* Update skb to refer to this connection */
-	nf_ct_set(skb, nf_ct_tuplehash_to_ctrack(h), ctinfo);
+	nf_ct_set(skb, ct, ctinfo);
 	return NF_ACCEPT;
 }
 
@@ -167,11 +207,12 @@ int nf_conntrack_icmpv4_error(struct nf_conn *tmpl,
 			      struct sk_buff *skb, unsigned int dataoff,
 			      const struct nf_hook_state *state)
 {
+	union nf_inet_addr outer_daddr;
 	const struct icmphdr *icmph;
 	struct icmphdr _ih;
 
 	/* Not enough header? */
-	icmph = skb_header_pointer(skb, ip_hdrlen(skb), sizeof(_ih), &_ih);
+	icmph = skb_header_pointer(skb, dataoff, sizeof(_ih), &_ih);
 	if (icmph == NULL) {
 		icmp_error_log(skb, state, "short packet");
 		return -NF_ACCEPT;
@@ -204,7 +245,12 @@ int nf_conntrack_icmpv4_error(struct nf_conn *tmpl,
 	    icmph->type != ICMP_REDIRECT)
 		return NF_ACCEPT;
 
-	return icmp_error_message(tmpl, skb, state);
+	memset(&outer_daddr, 0, sizeof(outer_daddr));
+	outer_daddr.ip = ip_hdr(skb)->daddr;
+
+	dataoff += sizeof(*icmph);
+	return nf_conntrack_inet_error(tmpl, skb, dataoff, state,
+				       IPPROTO_ICMP, &outer_daddr);
 }
 
 #if IS_ENABLED(CONFIG_NF_CT_NETLINK)
@@ -303,56 +349,16 @@ icmp_timeout_nla_policy[CTA_TIMEOUT_ICMP_MAX+1] = {
 };
 #endif /* CONFIG_NF_CONNTRACK_TIMEOUT */
 
-#ifdef CONFIG_SYSCTL
-static struct ctl_table icmp_sysctl_table[] = {
-	{
-		.procname	= "nf_conntrack_icmp_timeout",
-		.maxlen		= sizeof(unsigned int),
-		.mode		= 0644,
-		.proc_handler	= proc_dointvec_jiffies,
-	},
-	{ }
-};
-#endif /* CONFIG_SYSCTL */
-
-static int icmp_kmemdup_sysctl_table(struct nf_proto_net *pn,
-				     struct nf_icmp_net *in)
-{
-#ifdef CONFIG_SYSCTL
-	pn->ctl_table = kmemdup(icmp_sysctl_table,
-				sizeof(icmp_sysctl_table),
-				GFP_KERNEL);
-	if (!pn->ctl_table)
-		return -ENOMEM;
-
-	pn->ctl_table[0].data = &in->timeout;
-#endif
-	return 0;
-}
-
-static int icmp_init_net(struct net *net)
+void nf_conntrack_icmp_init_net(struct net *net)
 {
 	struct nf_icmp_net *in = nf_icmp_pernet(net);
-	struct nf_proto_net *pn = &in->pn;
 
 	in->timeout = nf_ct_icmp_timeout;
-
-	return icmp_kmemdup_sysctl_table(pn, in);
-}
-
-static struct nf_proto_net *icmp_get_net_proto(struct net *net)
-{
-	return &net->ct.nf_ct_proto.icmp.pn;
 }
 
 const struct nf_conntrack_l4proto nf_conntrack_l4proto_icmp =
 {
 	.l4proto		= IPPROTO_ICMP,
-	.pkt_to_tuple		= icmp_pkt_to_tuple,
-	.invert_tuple		= icmp_invert_tuple,
-	.packet			= icmp_packet,
-	.destroy		= NULL,
-	.me			= NULL,
 #if IS_ENABLED(CONFIG_NF_CT_NETLINK)
 	.tuple_to_nlattr	= icmp_tuple_to_nlattr,
 	.nlattr_tuple_size	= icmp_nlattr_tuple_size,
@@ -368,6 +374,4 @@ const struct nf_conntrack_l4proto nf_conntrack_l4proto_icmp =
 		.nla_policy	= icmp_timeout_nla_policy,
 	},
 #endif /* CONFIG_NF_CONNTRACK_TIMEOUT */
-	.init_net		= icmp_init_net,
-	.get_net_proto		= icmp_get_net_proto,
 };