From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jamie Gloudon Subject: Re: BUG: TCPDUMP invalid cksum persists after disabling TCP cksum offload Date: Tue, 18 Sep 2012 18:20:40 -0400 Message-ID: <20120918222040.GA19201@darkstar> References: <20120918193208.GA19030@darkstar> <1347998905.2685.29.camel@bwh-desktop.uk.solarflarecom.com> <20120918211423.GA19115@darkstar> <5058ED9E.1000902@hp.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Vijay Subramanian , netdev@vger.kernel.org To: Rick Jones Return-path: Received: from mail-qa0-f46.google.com ([209.85.216.46]:41170 "EHLO mail-qa0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755351Ab2IRWUs (ORCPT ); Tue, 18 Sep 2012 18:20:48 -0400 Received: by qaas11 with SMTP id s11so2942670qaa.19 for ; Tue, 18 Sep 2012 15:20:47 -0700 (PDT) Content-Disposition: inline In-Reply-To: <5058ED9E.1000902@hp.com> Sender: netdev-owner@vger.kernel.org List-ID: Rick is correct. Disabling TX CKO also disable TSO. You can see that in my ethtool output. On Tue, Sep 18, 2012 at 02:54:38PM -0700, Rick Jones wrote: > On 09/18/2012 02:46 PM, Vijay Subramanian wrote: > >>The results after disabling tcp cksum offload feature: > >># ethtool -K eth1 tx off > > > >Instead of this, can you try > ># ethtool -K eth1 tso off > > Doesn't TSO depend on TX CKO being enabled? That is, if TX CKO were > disabled, shouldn't TSO have been implicitly disabled at the same time? > > rick jones