From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stephen Hemminger Subject: Re: TCP: Peer messages Date: Sun, 19 Apr 2009 20:11:00 -0700 Message-ID: <20090419201100.55bfb6bb@nehalam> References: <20090418180303.18963ff0@nehalam> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: David Miller , Netdev To: "Ilpo =?UTF-8?B?SsOkcnZpbmVu?=" Return-path: Received: from mail.vyatta.com ([76.74.103.46]:46473 "EHLO mail.vyatta.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752490AbZDTDLI convert rfc822-to-8bit (ORCPT ); Sun, 19 Apr 2009 23:11:08 -0400 In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On Sun, 19 Apr 2009 14:19:36 +0300 (EEST) "Ilpo J=C3=A4rvinen" wrote: > On Sat, 18 Apr 2009, Stephen Hemminger wrote: >=20 > > Running some log-term tests of (using bittorrent of latest Ubuntu r= eleases). > > I see no problems with firewall rules (or sky2 driver), but I do se= veral=20 > > of these:=20 > >=20 > >=20 > > [28286.893945] TCP: Peer :63659/51413 unexpectedly shr= unk window 414881290:414894358 (repaired) > > [47312.851937] TCP: Peer :43793/51413 unexpectedly shr= unk window 888816585:888823785 (repaired) >=20 > This is what was earlier the Treason uncloacked message until somebod= y=20 > was scared to death because of it and we had to change the wording. >=20 > And it's what the old kernels do for us (and there are plenty of such= =20 > kernels around), and possibly broken middleboxes too. Unless you have= a=20 > tcpdump of a specific case or can clearly say that its frequency is m= uch=20 > larger than it used to be in some between kernel versions consistent = test=20 > setup, I'm not going to waste my time on debugging obsolete remote en= ds. >=20 > But, now that I think it a bit more, we could possibly but a WARN_ON=20 > somewhere if we ever sent past advertized window to make a distinctio= n=20 > between our and their bugs. >=20 12 over 3 days is not enough to worry about.