ipv6: tcp: handle ICMP messages on TCP_NEW_SYN_RECV request sockets
authorEric Dumazet <edumazet@google.com>
Sun, 22 Mar 2015 17:22:23 +0000 (10:22 -0700)
committerDavid S. Miller <davem@davemloft.net>
Mon, 23 Mar 2015 20:52:26 +0000 (16:52 -0400)
tcp_v6_err() can restrict lookups to ehash table, and not to listeners.

Note this patch creates the infrastructure, but this means that ICMP
messages for request sockets are ignored until complete conversion.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
net/ipv6/tcp_ipv6.c

index 6e3f90db038cb001dad5c4dddef88d93ecdbf5f3..4a4e6d30c448b1ee8ec4948e025ed0a19d505553 100644 (file)
@@ -324,18 +324,20 @@ static void tcp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
 {
        const struct ipv6hdr *hdr = (const struct ipv6hdr *)skb->data;
        const struct tcphdr *th = (struct tcphdr *)(skb->data+offset);
+       struct net *net = dev_net(skb->dev);
+       struct request_sock *fastopen;
        struct ipv6_pinfo *np;
-       struct sock *sk;
-       int err;
        struct tcp_sock *tp;
-       struct request_sock *fastopen;
        __u32 seq, snd_una;
-       struct net *net = dev_net(skb->dev);
+       struct sock *sk;
+       int err;
 
-       sk = inet6_lookup(net, &tcp_hashinfo, &hdr->daddr,
-                       th->dest, &hdr->saddr, th->source, skb->dev->ifindex);
+       sk = __inet6_lookup_established(net, &tcp_hashinfo,
+                                       &hdr->daddr, th->dest,
+                                       &hdr->saddr, ntohs(th->source),
+                                       skb->dev->ifindex);
 
-       if (sk == NULL) {
+       if (!sk) {
                ICMP6_INC_STATS_BH(net, __in6_dev_get(skb->dev),
                                   ICMP6_MIB_INERRORS);
                return;
@@ -345,6 +347,9 @@ static void tcp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
                inet_twsk_put(inet_twsk(sk));
                return;
        }
+       seq = ntohl(th->seq);
+       if (sk->sk_state == TCP_NEW_SYN_RECV)
+               return tcp_req_err(sk, seq);
 
        bh_lock_sock(sk);
        if (sock_owned_by_user(sk) && type != ICMPV6_PKT_TOOBIG)
@@ -359,7 +364,6 @@ static void tcp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
        }
 
        tp = tcp_sk(sk);
-       seq = ntohl(th->seq);
        /* XXX (TFO) - tp->snd_una should be ISN (tcp_create_openreq_child() */
        fastopen = tp->fastopen_rsk;
        snd_una = fastopen ? tcp_rsk(fastopen)->snt_isn : tp->snd_una;
@@ -403,33 +407,6 @@ static void tcp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
 
        /* Might be for an request_sock */
        switch (sk->sk_state) {
-               struct request_sock *req;
-       case TCP_LISTEN:
-               if (sock_owned_by_user(sk))
-                       goto out;
-
-               /* Note : We use inet6_iif() here, not tcp_v6_iif() */
-               req = inet6_csk_search_req(sk, th->dest, &hdr->daddr,
-                                          &hdr->saddr, inet6_iif(skb));
-               if (!req)
-                       goto out;
-
-               /* ICMPs are not backlogged, hence we cannot get
-                * an established socket here.
-                */
-               WARN_ON(req->sk != NULL);
-
-               if (seq != tcp_rsk(req)->snt_isn) {
-                       NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS);
-                       reqsk_put(req);
-                       goto out;
-               }
-
-               inet_csk_reqsk_queue_drop(sk, req);
-               NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENDROPS);
-               reqsk_put(req);
-               goto out;
-
        case TCP_SYN_SENT:
        case TCP_SYN_RECV:
                /* Only in fast or simultaneous open. If a fast open socket is