From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: From: Greg Kroah-Hartman To: linux-kernel@vger.kernel.org Cc: Greg Kroah-Hartman , stable@vger.kernel.org, Yuchung Cheng , Neal Cardwell , "David S. Miller" Subject: [ 44/56] tcp: undo spurious timeout after SACK reneging Date: Tue, 2 Apr 2013 15:50:11 -0700 Message-Id: <20130402224716.968437540@linuxfoundation.org> In-Reply-To: <20130402224711.840825715@linuxfoundation.org> References: <20130402224711.840825715@linuxfoundation.org> Sender: linux-kernel-owner@vger.kernel.org List-ID: 3.0-stable review patch. If anyone has any objections, please let me know. ------------------ From: Yuchung Cheng [ Upstream commit 7ebe183c6d444ef5587d803b64a1f4734b18c564 ] 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 Acked-by: Neal Cardwell Signed-off-by: David S. Miller Signed-off-by: Greg Kroah-Hartman --- net/ipv4/tcp_input.c | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -2244,11 +2244,8 @@ void tcp_enter_loss(struct sock *sk, int 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; }