From mboxrd@z Thu Jan  1 00:00:00 1970
From: Jesse Gross <jesse@nicira.com>
Subject: Re: TAHI CN-6-4-1 failed on Linux 2.6.32 kernel
Date: Fri, 20 Aug 2010 13:39:03 -0400
Message-ID: <AANLkTim9oDnu-eC31f1vE18Dd6JXMtTcxnd=ULDjAOh4@mail.gmail.com>
References: <AANLkTik9=TcMWzNyKWYyDu+bbGscPne6++gD42aju9s_@mail.gmail.com>
	<AANLkTi=qP7BO2=Uhx2Xb0XiLbUz0mPutbg0vz6QnMgYb@mail.gmail.com>
	<AANLkTim+1=4VCmY=fO5HTvDwKnveGEPsY1h1O64vnT6L@mail.gmail.com>
	<20100819.170649.212693227.davem@davemloft.net>
	<AANLkTimSishHmLTQ9RF8zMq_eOBQ+HiBroKoygFS27-x@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: QUOTED-PRINTABLE
Cc: David Miller <davem@davemloft.net>, brian.haley@hp.com,
	usagi-users-ctl@ml.linux-ipv6.org, netdev@vger.kernel.org
To: Steve Chen <schen@mvista.com>
Return-path: <netdev-owner@vger.kernel.org>
Received: from mail-iw0-f174.google.com ([209.85.214.174]:40272 "EHLO
	mail-iw0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1751734Ab0HTRjD convert rfc822-to-8bit (ORCPT
	<rfc822;netdev@vger.kernel.org>); Fri, 20 Aug 2010 13:39:03 -0400
Received: by iwn5 with SMTP id 5so1027358iwn.19
        for <netdev@vger.kernel.org>; Fri, 20 Aug 2010 10:39:03 -0700 (PDT)
In-Reply-To: <AANLkTimSishHmLTQ9RF8zMq_eOBQ+HiBroKoygFS27-x@mail.gmail.com>
Sender: netdev-owner@vger.kernel.org
List-ID: <netdev.vger.kernel.org>

On Fri, Aug 20, 2010 at 1:16 PM, Steve Chen <schen@mvista.com> wrote:
>
> On Thu, Aug 19, 2010 at 7:06 PM, David Miller <davem@davemloft.net> w=
rote:
> > From: Steve Chen <schen@mvista.com>
> > Date: Thu, 19 Aug 2010 13:35:14 -0500
> >
> >> I trace through the code. =A0It appears that the network driver (e=
1000e
> >> for my setup) always set ip_summed to CHECKSUM_UNNECESSARY. =A0I h=
ave
> >> been unsuccessful to get the driver to take the other branch where
> >> ip_summed is set to CHECKSUM_COMPLETE. =A0Even when I hard code
> >> ip_summed to CHECKSUM_COMPLETE, __skb_checksum_complete_head set
> >> ip_summed to CHECKSUM_UNNECESSARY after recomputing the checksum.
> >>
> >> So far the only way I'm able to get ICMP to recompute checksum is
> >> through the attached hack. =A0Even though I can get all the tests =
to
> >> pass, but it just seem wrong.
> >
> > If turning off hardware RX checksumming with ethtool has no effect,
> > and the problem is seen with multiple ethernet cards, the problem
> > is elsewhere.
> >
> > First of all, if you turn RX checksumming off, the checksum field
> > of the SKB should always be skb->ip_summed =3D 0. =A0If this is not
> > happening, find out why.
>
> Ahhh, thats my problem. =A0I incorrectly thought the ip_summed should=
 be
> 2. =A0The ip_summed is set to 1 in
> __skb_checksum_complete_head. =A0Looking at the code, shouldn't
>
> if (likely(!sum))
>
> be
>
> if (likely(sum))
>
> Since sum =3D=3D 0 would indicate an error?

sum =3D=3D 0 indicates that the checksum is correct.

If you compute the checksum of a packet containing the correct
checksum the result is 0.  It's like a slightly more complicated
varient of a parity bit.