From owner-freebsd-net Thu Aug 8 17:40:44 2002 Delivered-To: freebsd-net@freebsd.org Received: from mx1.FreeBSD.org (mx1.FreeBSD.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id E82D837B400 for ; Thu, 8 Aug 2002 17:40:40 -0700 (PDT) Received: from blv-smtpout-01.boeing.com (blv-smtpout-01.boeing.com [192.161.36.5]) by mx1.FreeBSD.org (Postfix) with ESMTP id 2E40F43E65 for ; Thu, 8 Aug 2002 17:40:40 -0700 (PDT) (envelope-from thomas.r.henderson@boeing.com) Received: from slb-av-02.boeing.com ([129.172.13.7]) by blv-smtpout-01.boeing.com (8.9.2/8.8.5-M2) with ESMTP id RAA26919; Thu, 8 Aug 2002 17:40:38 -0700 (PDT) Received: from blv-hub-01.boeing.com (localhost [127.0.0.1]) by slb-av-02.boeing.com (8.9.3/8.9.2/MBS-AV-02) with ESMTP id RAA22086; Thu, 8 Aug 2002 17:40:37 -0700 (PDT) Received: from xch-nwbh-02.nw.nos.boeing.com (xch-nwbh-02.nw.nos.boeing.com [192.54.12.28]) by blv-hub-01.boeing.com (8.11.3/8.11.3/MBS-LDAP-01) with ESMTP id g790eam18833; Thu, 8 Aug 2002 17:40:36 -0700 (PDT) Received: by xch-nwbh-02.nw.nos.boeing.com with Internet Mail Service (5.5.2650.21) id ; Thu, 8 Aug 2002 17:40:38 -0700 Message-ID: <00EBC850E752CC46B8509DAB4D0D2CB910692D@xch-nw-29.nw.nos.boeing.com> From: "Henderson, Thomas R" To: "'Mike Silbersack'" Cc: "'freebsd-net@freebsd.org'" Subject: RE: [patch] Possible newreno fix, please test Date: Thu, 8 Aug 2002 17:40:35 -0700 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Type: text/plain; charset="iso-8859-1" Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org > Personally, I don't believe that the code in question was > correct in the > first place. Isn't the congestion window supposed to be increased for > every dupack received, not just every third? > I agree-- but it seems to me that the problem line in there is t_dupacks = 0; which causes two possible problems i) cwnd incremented for every third dupack ii) (more serious) you never do the window deflation in certain loss scenarios, because your dupack counter doesn't indicate that you were previously in fast retransmission. I wonder if the patch should be instead: +++ tcp_input.c Wed Aug 7 23:03:38 2002 @@ -1662,9 +1662,7 @@ /* False retransmit, should not * cut window */ tp->snd_cwnd += tp->t_maxseg; - tp->t_dupacks = 0; (void) tcp_output(tp); goto drop; } if (win < 2) To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message