summaryrefslogtreecommitdiff
path: root/net
diff options
context:
space:
mode:
authorNeal Cardwell <ncardwell@google.com>2011-11-16 08:58:02 +0000
committerDavid S. Miller <davem@davemloft.net>2011-11-27 18:54:09 -0500
commit5628adf1a0ff39b9e76e1a8e1c94dadabfd46914 (patch)
tree99f817c26d6405ea23a7e10e34736e38d199b256 /net
parent7d2b55f80d62b6b4b72d9ef4318a2a49df3e2830 (diff)
tcp: use DSACKs that arrive when packets_out is 0
The bug: Senders ignored DSACKs after recovery when there were no outstanding packets (a common scenario for HTTP servers). The change: when there are no outstanding packets (the "no_queue" goto label), call tcp_fastretrans_alert() in order to use DSACKs to undo congestion window reductions. Other patches in this series will provide other changes that are necessary to fully fix this problem. Signed-off-by: Neal Cardwell <ncardwell@google.com> Acked-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi> Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'net')
-rw-r--r--net/ipv4/tcp_input.c4
1 files changed, 4 insertions, 0 deletions
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index f772aaa4319e..b49e41864037 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -3788,6 +3788,10 @@ static int tcp_ack(struct sock *sk, const struct sk_buff *skb, int flag)
return 1;
no_queue:
+ /* If data was DSACKed, see if we can undo a cwnd reduction. */
+ if (flag & FLAG_DSACKING_ACK)
+ tcp_fastretrans_alert(sk, pkts_acked, newly_acked_sacked,
+ is_dupack, flag);
/* If this ack opens up a zero window, clear backoff. It was
* being used to time the probes, and is probably far higher than
* it needs to be for normal retransmission.