From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: [PATCH] tcp: frto should not set snd_cwnd to 0 Date: Wed, 06 Feb 2013 16:19:19 -0500 (EST) Message-ID: <20130206.161919.1732889798702305226.davem@davemloft.net> References: <20130206.155504.1553906005129825120.davem@davemloft.net> <20130206211343.GW8912@reaktio.net> Mime-Version: 1.0 Content-Type: Text/Plain; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: ncardwell@google.com, ilpo.jarvinen@helsinki.fi, eric.dumazet@gmail.com, hannes@stressinduktion.org, stephen@networkplumber.org, netdev@vger.kernel.org, ycheng@google.com To: pasik@iki.fi Return-path: Received: from shards.monkeyblade.net ([149.20.54.216]:43909 "EHLO shards.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1758026Ab3BFVTW convert rfc822-to-8bit (ORCPT ); Wed, 6 Feb 2013 16:19:22 -0500 In-Reply-To: <20130206211343.GW8912@reaktio.net> Sender: netdev-owner@vger.kernel.org List-ID: =46rom: Pasi K=E4rkk=E4inen Date: Wed, 6 Feb 2013 23:13:43 +0200 > On Wed, Feb 06, 2013 at 03:55:04PM -0500, David Miller wrote: >> From: Neal Cardwell >> Date: Tue, 5 Feb 2013 14:49:04 -0500 >>=20 >> > On Mon, Feb 4, 2013 at 7:14 AM, Ilpo J=E4rvinen wrote: >> >> [PATCH 1/1] tcp: fix for zero packets_in_flight was too broad >> >> >> >> There are transients during normal FRTO procedure during which >> >> the packets_in_flight can go to zero between write_queue state >> >> updates and firing the resulting segments out. As FRTO processing >> >> occurs during that window the check must be more precise to >> >> not match "spuriously" :-). More specificly, e.g., when >> >> packets_in_flight is zero but FLAG_DATA_ACKED is true the problem= atic >> >> branch that set cwnd into zero would not be taken and new segment= s >> >> might be sent out later. >> >> >> >> Only compile tested. >> >> >> >> Signed-off-by: Ilpo J=E4rvinen >> ... >> > Acked-by: Neal Cardwell >>=20 >> Applied, thanks everyone. >> >=20 > Hmm.. are we missing CC stable@kernel.org in these patches?=20 > I guess 3.6.x is already EOL, but it'd be nice to get this bug fixed = also in 3.7.x .. We never CC: stable on networking patches, I queue them up manually and submit them at a time of my own choosing.