From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: Fix FRTO+NewReno problem Date: Tue, 13 May 2008 02:49:15 -0700 (PDT) Message-ID: <20080513.024915.161125347.davem@davemloft.net> References: <20080512.154838.163999274.davem@davemloft.net> Mime-Version: 1.0 Content-Type: Text/Plain; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: damon@damtek.com, netdev@vger.kernel.org To: ilpo.jarvinen@helsinki.fi Return-path: Received: from 74-93-104-97-Washington.hfc.comcastbusiness.net ([74.93.104.97]:60713 "EHLO sunset.davemloft.net" rhost-flags-OK-FAIL-OK-OK) by vger.kernel.org with ESMTP id S1750814AbYEMJtV convert rfc822-to-8bit (ORCPT ); Tue, 13 May 2008 05:49:21 -0400 In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: =46rom: "Ilpo_J=E4rvinen" Date: Tue, 13 May 2008 12:42:03 +0300 (EEST) > On Mon, 12 May 2008, David Miller wrote: >=20 > > If I understand correctly, it is only non-SACK newreno case that ca= n > > get the unnecessary retransmissions, right? If so, it's not a big > > deal at all. >=20 > No, also SACK could get at least one quite easily if TCP has a small=20 > window, to get two of them with SACK one needs to have a stranger cas= e > unless I missed something. If I just remove that forward transmission= LOST=20 > marking avoidance unconditionally, it fixes that SACKFRTO window of=20 > failure too which was a pending thing to fix :-). Ok. > Btw, the first two fixes (workaround was not included) also fixed bug= zilla=20 > #10063, I just got a confirmation about that as well. >=20 > I'll send the two patches still pending separately in a minute. Thank you.