From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: [regression] 2.6.37+ commit 0363466866d9.... breaks tcp ipv6 Date: Fri, 21 Jan 2011 20:55:45 +0100 Message-ID: <1295639745.2609.29.camel@edumazet-laptop> References: <4D335417.80704@xmsnet.nl> <4D35EF64.1040906@xmsnet.nl> <4D36092B.1090406@xmsnet.nl> <1295388238.8449.10.camel@edumazet-laptop> <4D39E2EC.9020906@xmsnet.nl> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Jesse Gross , netdev To: Hans de Bruin Return-path: Received: from mail-ww0-f44.google.com ([74.125.82.44]:34207 "EHLO mail-ww0-f44.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752903Ab1AUTzu (ORCPT ); Fri, 21 Jan 2011 14:55:50 -0500 Received: by wwa36 with SMTP id 36so2261779wwa.1 for ; Fri, 21 Jan 2011 11:55:49 -0800 (PST) In-Reply-To: <4D39E2EC.9020906@xmsnet.nl> Sender: netdev-owner@vger.kernel.org List-ID: Le vendredi 21 janvier 2011 =C3=A0 20:47 +0100, Hans de Bruin a =C3=A9c= rit : > On 01/18/2011 11:03 PM, Eric Dumazet wrote: > > Le mardi 18 janvier 2011 =C3=A0 22:42 +0100, Hans de Bruin a =C3=A9= crit : > >> On 01/18/2011 09:06 PM, Jesse Gross wrote: >=20 > ... >=20 > > You could try "tcpdump -i eth0 ip6 -v" > > > > I guess you receive frames with bad checksums >=20 > While you where staring at the code, I was fooling around with tcpdum= p.=20 > And while the problem is fixed, I still have some questions: >=20 > Is there tool which shows whether a nic supports ipv6 checksum offloa= d=20 > or not? >=20 > I have captured http traffic (wget http://bootes/) between psion (my = git=20 > tree following laptop) and bootes (something running 2.6.33.7). > Attached is a capture with psion running 2.6.37 and one with this=20 > morning's git tree. Wat's with the 'chsum ... ( incorrect -> ' lines = ? > ifconfig does not show errors on either of the machines. >=20 tcpdump gets a copy of outgoing frames before NIC performs tx checksum (if tx checksum handled by NIC), so it's normal to have "bad checksums" on TX, unless you disable tx offloading (ethtool -K eth0 tx off) I was referring to check with tcpdump incoming frames, because invalid checksums in RX is sign that other peer sent wrong checksums