diff options
| author | Eric Dumazet <edumazet@google.com> | 2012-07-23 09:48:52 +0200 |
|---|---|---|
| committer | David S. Miller <davem@davemloft.net> | 2012-07-23 00:58:46 -0700 |
| commit | 563d34d05786263893ba4a1042eb9b9374127cf5 (patch) | |
| tree | e9ce502c1f32bea966c81d5597d0a29eb4b9d244 /net/ipv6 | |
| parent | c3def943c7117d42caaed3478731ea7c3c87190e (diff) | |
| download | linux-3.10-563d34d05786263893ba4a1042eb9b9374127cf5.tar.gz linux-3.10-563d34d05786263893ba4a1042eb9b9374127cf5.tar.bz2 linux-3.10-563d34d05786263893ba4a1042eb9b9374127cf5.zip | |
tcp: dont drop MTU reduction indications
ICMP messages generated in output path if frame length is bigger than
mtu are actually lost because socket is owned by user (doing the xmit)
One example is the ipgre_tunnel_xmit() calling
icmp_send(skb, ICMP_DEST_UNREACH, ICMP_FRAG_NEEDED, htonl(mtu));
We had a similar case fixed in commit a34a101e1e6 (ipv6: disable GSO on
sockets hitting dst_allfrag).
Problem of such fix is that it relied on retransmit timers, so short tcp
sessions paid a too big latency increase price.
This patch uses the tcp_release_cb() infrastructure so that MTU
reduction messages (ICMP messages) are not lost, and no extra delay
is added in TCP transmits.
Reported-by: Maciej Żenczykowski <maze@google.com>
Diagnosed-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Nandita Dukkipati <nanditad@google.com>
Cc: Tom Herbert <therbert@google.com>
Cc: Tore Anderson <tore@fud.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'net/ipv6')
| -rw-r--r-- | net/ipv6/tcp_ipv6.c | 40 |
1 files changed, 24 insertions, 16 deletions
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index 0302ec3fecf..f49476e2d88 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -315,6 +315,23 @@ failure: return err; } +static void tcp_v6_mtu_reduced(struct sock *sk) +{ + struct dst_entry *dst; + + if ((1 << sk->sk_state) & (TCPF_LISTEN | TCPF_CLOSE)) + return; + + dst = inet6_csk_update_pmtu(sk, tcp_sk(sk)->mtu_info); + if (!dst) + return; + + if (inet_csk(sk)->icsk_pmtu_cookie > dst_mtu(dst)) { + tcp_sync_mss(sk, dst_mtu(dst)); + tcp_simple_retransmit(sk); + } +} + static void tcp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt, u8 type, u8 code, int offset, __be32 info) { @@ -342,7 +359,7 @@ static void tcp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt, } bh_lock_sock(sk); - if (sock_owned_by_user(sk)) + if (sock_owned_by_user(sk) && type != ICMPV6_PKT_TOOBIG) NET_INC_STATS_BH(net, LINUX_MIB_LOCKDROPPEDICMPS); if (sk->sk_state == TCP_CLOSE) @@ -371,21 +388,11 @@ static void tcp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt, } if (type == ICMPV6_PKT_TOOBIG) { - struct dst_entry *dst; - - if (sock_owned_by_user(sk)) - goto out; - if ((1 << sk->sk_state) & (TCPF_LISTEN | TCPF_CLOSE)) - goto out; - - dst = inet6_csk_update_pmtu(sk, ntohl(info)); - if (!dst) - goto out; - - if (inet_csk(sk)->icsk_pmtu_cookie > dst_mtu(dst)) { - tcp_sync_mss(sk, dst_mtu(dst)); - tcp_simple_retransmit(sk); - } + tp->mtu_info = ntohl(info); + if (!sock_owned_by_user(sk)) + tcp_v6_mtu_reduced(sk); + else + set_bit(TCP_MTU_REDUCED_DEFERRED, &tp->tsq_flags); goto out; } @@ -1949,6 +1956,7 @@ struct proto tcpv6_prot = { .sendpage = tcp_sendpage, .backlog_rcv = tcp_v6_do_rcv, .release_cb = tcp_release_cb, + .mtu_reduced = tcp_v6_mtu_reduced, .hash = tcp_v6_hash, .unhash = inet_unhash, .get_port = inet_csk_get_port, |