[PATCH 061/102] tcp: undo spurious timeout after SACK reneging

[Luis Henriques]

3.5.7.10 -stable review patch.  If anyone has any objections, please let me know.

------------------

From: Yuchung Cheng <ycheng at google.com>

commit 7ebe183c6d444ef5587d803b64a1f4734b18c564 upstream.

On SACK reneging the sender immediately retransmits and forces a
timeout but disables Eifel (undo). If the (buggy) receiver does not
drop any packet this can trigger a false slow-start retransmit storm
driven by the ACKs of the original packets. This can be detected with
undo and TCP timestamps.

Signed-off-by: Yuchung Cheng <ycheng at google.com>
Acked-by: Neal Cardwell <ncardwell at google.com>
Signed-off-by: David S. Miller <davem at davemloft.net>
Signed-off-by: Luis Henriques <luis.henriques at canonical.com>
---
 net/ipv4/tcp_input.c | 7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)

diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 410b7c4..e75e4fb 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -2267,11 +2267,8 @@ void tcp_enter_loss(struct sock *sk, int how)
 	if (tcp_is_reno(tp))
 		tcp_reset_reno_sack(tp);
 
-	if (!how) {
-		/* Push undo marker, if it was plain RTO and nothing
-		 * was retransmitted. */
-		tp->undo_marker = tp->snd_una;
-	} else {
+	tp->undo_marker = tp->snd_una;
+	if (how) {
 		tp->sacked_out = 0;
 		tp->fackets_out = 0;
 	}
-- 
1.8.1.2