From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: [PATCH net-next v3] tcp: avoid reducing cwnd when ACK+DSACK is received Date: Tue, 13 Jan 2015 14:22:35 -0500 (EST) Message-ID: <20150113.142235.1647397796492702755.davem@davemloft.net> References: <1421055040-8732-1-git-send-email-sebastien.barre@uclouvain.be> Mime-Version: 1.0 Content-Type: Text/Plain; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: ncardwell@google.com, ycheng@google.com, eric.dumazet@gmail.com, netdev@vger.kernel.org, gregory.detal@uclouvain.be, nanditad@google.com To: sebastien.barre@uclouvain.be Return-path: Received: from shards.monkeyblade.net ([149.20.54.216]:35183 "EHLO shards.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751789AbbAMTWi convert rfc822-to-8bit (ORCPT ); Tue, 13 Jan 2015 14:22:38 -0500 In-Reply-To: <1421055040-8732-1-git-send-email-sebastien.barre@uclouvain.be> Sender: netdev-owner@vger.kernel.org List-ID: =46rom: S=E9bastien Barr=E9 Date: Mon, 12 Jan 2015 10:30:40 +0100 > With TLP, the peer may reply to a probe with an > ACK+D-SACK, with ack value set to tlp_high_seq. In the current code, > such ACK+DSACK will be missed and only at next, higher ack will the T= LP > episode be considered done. Since the DSACK is not present anymore, > this will cost a cwnd reduction. >=20 > This patch ensures that this scenario does not cause a cwnd reduction= , since > receiving an ACK+DSACK indicates that both the initial segment and th= e probe > have been received by the peer. >=20 > The following packetdrill test, from Neal Cardwell, validates this pa= tch: ... > Credits: > -Gregory helped in finding that tcp_process_tlp_ack was where the cwn= d > got reduced in our MPTCP tests. > -Neal wrote the packetdrill test above > -Yuchung reworked the patch to make it more readable. >=20 > Cc: Gregory Detal > Cc: Nandita Dukkipati > Tested-by: Neal Cardwell > Reviewed-by: Yuchung Cheng > Reviewed-by: Eric Dumazet > Signed-off-by: S=E9bastien Barr=E9 Applied, thanks everyone.