tcp: skip cwnd moderation in TCP_CA_Open in tcp_try_to_open

The problem: Senders were overriding cwnd values picked during an undo
by calling tcp_moderate_cwnd() in tcp_try_to_open().

The fix: Don't moderate cwnd in tcp_try_to_open() if we're in
TCP_CA_Open, since doing so is generally unnecessary and specifically
would override a DSACK-based undo of a cwnd reduction made in fast
recovery.

Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index a4efdd7cf5a18f56f10a87a560ec7787aa645a46..78dd38cd5496d38f8b708996f17ce5d571599a96 100644 (file)
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -2881,7 +2881,8 @@ static void tcp_try_to_open(struct sock *sk, int flag)
 
        if (inet_csk(sk)->icsk_ca_state != TCP_CA_CWR) {
                tcp_try_keep_open(sk);
-               tcp_moderate_cwnd(tp);
+               if (inet_csk(sk)->icsk_ca_state != TCP_CA_Open)
+                       tcp_moderate_cwnd(tp);
        } else {
                tcp_cwnd_down(sk, flag);
        }