diff options
| author | Ilpo Järvinen <ilpo.jarvinen@helsinki.fi> | 2009-02-28 04:44:31 +0000 |
|---|---|---|
| committer | David S. Miller <davem@davemloft.net> | 2009-03-02 03:00:11 -0800 |
| commit | 02276f3c962fd408fa9d441251067845f948bfcf (patch) | |
| tree | 040b778209f0694273a9d5d8337814b6a826c74e | |
| parent | d3d2ae454501a4dec360995649e1b002a2ad90c5 (diff) | |
tcp: fix corner case issue in segmentation during rexmitting
If cur_mss grew very recently so that the previously G/TSOed skb
now fits well into a single segment it would get send up in
parts unless we calculate # of segments again. This corner-case
could happen eg. after mtu probe completes or less than
previously sack blocks are required for the opposite direction.
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
| -rw-r--r-- | net/ipv4/tcp_output.c | 2 |
1 files changed, 2 insertions, 0 deletions
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 3feab4d6929..77af7faf38a 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -1921,6 +1921,8 @@ int tcp_retransmit_skb(struct sock *sk, struct sk_buff *skb) if (skb->len > cur_mss) { if (tcp_fragment(sk, skb, cur_mss, cur_mss)) return -ENOMEM; /* We'll try again later. */ + } else { + tcp_init_tso_segs(sk, skb, cur_mss); } tcp_retrans_try_collapse(sk, skb, cur_mss); |