From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jarek Poplawski Subject: Re: two other cases Re: [Bug 11721] after upgrade to 2.6.27 i cannot navigate Date: Tue, 21 Oct 2008 11:12:27 +0000 Message-ID: <20081021111227.GE4715@ff.dom.local> References: <20081021090002.21e64151@topolino> <20081021100923.GD4715@ff.dom.local> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Aldo Maggi , Netdev , bugme-daemon@bugzilla.kernel.org To: Ilpo =?iso-8859-1?Q?J=E4rvinen?= Return-path: Received: from ik-out-1112.google.com ([66.249.90.183]:9022 "EHLO ik-out-1112.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752723AbYJULMe (ORCPT ); Tue, 21 Oct 2008 07:12:34 -0400 Received: by ik-out-1112.google.com with SMTP id c30so1519692ika.5 for ; Tue, 21 Oct 2008 04:12:32 -0700 (PDT) Content-Disposition: inline In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On Tue, Oct 21, 2008 at 01:51:10PM +0300, Ilpo J=E4rvinen wrote: > On Tue, 21 Oct 2008, Jarek Poplawski wrote: >=20 > > On Tue, Oct 21, 2008 at 12:36:33PM +0300, Ilpo J=E4rvinen wrote: > > ... > > > Can you try this another debug patch below (on 2.6.27.2 is fine).= It moves=20 > > > the mss to the last position but should keep timestamps in place = by making=20 > > > wscale as first option. It is well possible that you won't get it= working=20 > > > at all except with all ts,sack and wscale set to 0 (the most like= ly=20 > > > result). Please try with all wscale,sack,ts combinations (no need= to=20 > > > provide dumps, just working/not working per case)... This should=20 > > > tell us for quite high certaintity what is the actual option whic= h is=20 > > > causing this (would it not be the mss-at-beginning which is the m= ost=20 > > > likely cause), [...] > >=20 > > I'd like to remind it seems to work with only sack off, so mss afte= r > > ts. If so, my suspicion would be around rfc's/options' dating? >=20 > Thanks, I missed that case... there are so many cases already... :-) > ...I'm not sure we'll find some very obvious reason for this, rfc dat= ing=20 > probably has very little to do here. >=20 > I still think that putting an alternative option in front is useful t= est=20 > (though my initial mss-at-beginning thought wasn't bullet-proof) but=20 > maybe we could skip all those alternative settings of wscale,sack,ts = and=20 > have them just all enabled. I agree with you, but when this test with all options enabled fails (?!), IMHO disabling sack only could be enough to get it working again. Jarek P.