Commit 7ebe183c authored Mar 24, 2013 by Yuchung Cheng Committed by David S. Miller Mar 24, 2013

tcp: undo spurious timeout after SACK reneging



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

parent 8fe7f99a

net/ipv4/tcp_input.c

+2 −5

Original line number	Diff line number	Diff line
		@@ -2059,11 +2059,8 @@ 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 {
		if (how) {
		tp->sacked_out = 0;
		tp->fackets_out = 0;
		}