netfilter: xt_socket: add XT_SOCKET_NOWILDCARD flag

xt_socket module can be a nice replacement to conntrack module
in some cases (SYN filtering for example)

But it lacks the ability to match the 3rd packet of TCP
handshake (ACK coming from the client).

Add a XT_SOCKET_NOWILDCARD flag to disable the wildcard mechanism.

The wildcard is the legacy socket match behavior, that ignores
LISTEN sockets bound to INADDR_ANY (or ipv6 equivalent)

iptables -I INPUT -p tcp --syn -j SYN_CHAIN
iptables -I INPUT -m socket --nowildcard -j ACCEPT

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Patrick McHardy <kaber@trash.net>
Cc: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>

include/uapi/linux/netfilter/xt_socket.h

@@ -5,10 +5,17 @@
 
 enum {
 	XT_SOCKET_TRANSPARENT = 1 << 0,
+	XT_SOCKET_NOWILDCARD = 1 << 1,
 };
 
 struct xt_socket_mtinfo1 {
 	__u8 flags;
 };
+#define XT_SOCKET_FLAGS_V1 XT_SOCKET_TRANSPARENT
+
+struct xt_socket_mtinfo2 {
+	__u8 flags;
+};
+#define XT_SOCKET_FLAGS_V2 (XT_SOCKET_TRANSPARENT | XT_SOCKET_NOWILDCARD)
 
 #endif /* _XT_SOCKET_H */

net/netfilter/xt_socket.c

@@ -163,8 +163,11 @@ socket_match(const struct sk_buff *skb, struct xt_action_param *par,
 		bool wildcard;
 		bool transparent = true;
 
-		/* Ignore sockets listening on INADDR_ANY */
-		wildcard = (sk->sk_state != TCP_TIME_WAIT &&
+		/* Ignore sockets listening on INADDR_ANY,
+		 * unless XT_SOCKET_NOWILDCARD is set
+		 */
+		wildcard = (!(info->flags & XT_SOCKET_NOWILDCARD) &&
+			    sk->sk_state != TCP_TIME_WAIT &&
 			    inet_sk(sk)->inet_rcv_saddr == 0);
 
 		/* Ignore non-transparent sockets,
@@ -197,7 +200,7 @@ socket_mt4_v0(const struct sk_buff *skb, struct xt_action_param *par)
 }
 
 static bool
-socket_mt4_v1(const struct sk_buff *skb, struct xt_action_param *par)
+socket_mt4_v1_v2(const struct sk_buff *skb, struct xt_action_param *par)
 {
 	return socket_match(skb, par, par->matchinfo);
 }
@@ -259,7 +262,7 @@ extract_icmp6_fields(const struct sk_buff *skb,
 }
 
 static bool
-socket_mt6_v1(const struct sk_buff *skb, struct xt_action_param *par)
+socket_mt6_v1_v2(const struct sk_buff *skb, struct xt_action_param *par)
 {
 	struct ipv6hdr *iph = ipv6_hdr(skb);
 	struct udphdr _hdr, *hp = NULL;
@@ -302,8 +305,11 @@ socket_mt6_v1(const struct sk_buff *skb, struct xt_action_param *par)
 		bool wildcard;
 		bool transparent = true;
 
-		/* Ignore sockets listening on INADDR_ANY */
-		wildcard = (sk->sk_state != TCP_TIME_WAIT &&
+		/* Ignore sockets listening on INADDR_ANY
+		 * unless XT_SOCKET_NOWILDCARD is set
+		 */
+		wildcard = (!(info->flags & XT_SOCKET_NOWILDCARD) &&
+			    sk->sk_state != TCP_TIME_WAIT &&
 			    ipv6_addr_any(&inet6_sk(sk)->rcv_saddr));
 
 		/* Ignore non-transparent sockets,
@@ -331,6 +337,28 @@ socket_mt6_v1(const struct sk_buff *skb, struct xt_action_param *par)
 }
 #endif
 
+static int socket_mt_v1_check(const struct xt_mtchk_param *par)
+{
+	const struct xt_socket_mtinfo1 *info = (struct xt_socket_mtinfo1 *) par->matchinfo;
+
+	if (info->flags & ~XT_SOCKET_FLAGS_V1) {
+		pr_info("unknown flags 0x%x\n", info->flags & ~XT_SOCKET_FLAGS_V1);
+		return -EINVAL;
+	}
+	return 0;
+}
+
+static int socket_mt_v2_check(const struct xt_mtchk_param *par)
+{
+	const struct xt_socket_mtinfo2 *info = (struct xt_socket_mtinfo2 *) par->matchinfo;
+
+	if (info->flags & ~XT_SOCKET_FLAGS_V2) {
+		pr_info("unknown flags 0x%x\n", info->flags & ~XT_SOCKET_FLAGS_V2);
+		return -EINVAL;
+	}
+	return 0;
+}
+
 static struct xt_match socket_mt_reg[] __read_mostly = {
 	{
 		.name		= "socket",
@@ -345,7 +373,8 @@ static struct xt_match socket_mt_reg[] __read_mostly = {
 		.name		= "socket",
 		.revision	= 1,
 		.family		= NFPROTO_IPV4,
-		.match		= socket_mt4_v1,
+		.match		= socket_mt4_v1_v2,
+		.checkentry	= socket_mt_v1_check,
 		.matchsize	= sizeof(struct xt_socket_mtinfo1),
 		.hooks		= (1 << NF_INET_PRE_ROUTING) |
 				  (1 << NF_INET_LOCAL_IN),
@@ -356,7 +385,32 @@ static struct xt_match socket_mt_reg[] __read_mostly = {
 		.name		= "socket",
 		.revision	= 1,
 		.family		= NFPROTO_IPV6,
-		.match		= socket_mt6_v1,
+		.match		= socket_mt6_v1_v2,
+		.checkentry	= socket_mt_v1_check,
+		.matchsize	= sizeof(struct xt_socket_mtinfo1),
+		.hooks		= (1 << NF_INET_PRE_ROUTING) |
+				  (1 << NF_INET_LOCAL_IN),
+		.me		= THIS_MODULE,
+	},
+#endif
+	{
+		.name		= "socket",
+		.revision	= 2,
+		.family		= NFPROTO_IPV4,
+		.match		= socket_mt4_v1_v2,
+		.checkentry	= socket_mt_v2_check,
+		.matchsize	= sizeof(struct xt_socket_mtinfo1),
+		.hooks		= (1 << NF_INET_PRE_ROUTING) |
+				  (1 << NF_INET_LOCAL_IN),
+		.me		= THIS_MODULE,
+	},
+#ifdef XT_SOCKET_HAVE_IPV6
+	{
+		.name		= "socket",
+		.revision	= 2,
+		.family		= NFPROTO_IPV6,
+		.match		= socket_mt6_v1_v2,
+		.checkentry	= socket_mt_v2_check,
 		.matchsize	= sizeof(struct xt_socket_mtinfo1),
 		.hooks		= (1 << NF_INET_PRE_ROUTING) |
 				  (1 << NF_INET_LOCAL_IN),