[PATCH RFC 3/3] net: tcp_ipv4: re-run netfilter chains for marked skbs

[Daniel Mack <daniel@zonque.org>]

When an skb has been marked for later re-iteration through netfilter,
do that after __inet_lookup_skb() has been called. This allows packets
sent to unconnected sockets to be filtered reliably.

Note that this will never happen for subsequent packets in the same
stream, as skb->sk will be set due to early demux, and hence
skb->nf_postponed will remain 0.

Signed-off-by: Daniel Mack <daniel@zonque.org>
---
 net/ipv4/tcp_ipv4.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 93898e0..61e0cb4 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -78,6 +78,7 @@
 
 #include <linux/inet.h>
 #include <linux/ipv6.h>
+#include <linux/netfilter.h>
 #include <linux/stddef.h>
 #include <linux/proc_fs.h>
 #include <linux/seq_file.h>
@@ -1594,6 +1595,15 @@ int tcp_v4_rcv(struct sk_buff *skb)
 	if (!sk)
 		goto no_tcp_socket;
 
+	if (unlikely(skb->nf_postponed)) {
+		ret = nf_hook(NFPROTO_IPV4, NF_INET_LOCAL_IN, sk,
+			      skb, skb->dev, NULL, NULL);
+		if (ret != 1) {
+			sock_put(sk);
+			return 0;
+		}
+	}
+
 process:
 	if (sk->sk_state == TCP_TIME_WAIT)
 		goto do_time_wait;
-- 
2.5.0