tcp: fix TCP_MAXSEG for established IPv6 passive sockets
authorNeal Cardwell <ncardwell@google.com>
Sun, 22 Apr 2012 09:45:47 +0000 (09:45 +0000)
committerDavid S. Miller <davem@davemloft.net>
Sun, 22 Apr 2012 21:09:35 +0000 (17:09 -0400)
Commit f5fff5d forgot to fix TCP_MAXSEG behavior IPv6 sockets, so IPv6
TCP server sockets that used TCP_MAXSEG would find that the advmss of
child sockets would be incorrect. This commit mirrors the advmss logic
from tcp_v4_syn_recv_sock in tcp_v6_syn_recv_sock. Eventually this
logic should probably be shared between IPv4 and IPv6, but this at
least fixes this issue.

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

index 86cfe6005f40a76c716676161efa521cbd808171..98256cf72f9dbca22dc80b3405ac80ec4fdb158f 100644 (file)
@@ -1383,6 +1383,10 @@ static struct sock * tcp_v6_syn_recv_sock(struct sock *sk, struct sk_buff *skb,
        tcp_mtup_init(newsk);
        tcp_sync_mss(newsk, dst_mtu(dst));
        newtp->advmss = dst_metric_advmss(dst);
+       if (tcp_sk(sk)->rx_opt.user_mss &&
+           tcp_sk(sk)->rx_opt.user_mss < newtp->advmss)
+               newtp->advmss = tcp_sk(sk)->rx_opt.user_mss;
+
        tcp_initialize_rcv_mss(newsk);
        if (tcp_rsk(req)->snt_synack)
                tcp_valid_rtt_meas(newsk,