From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757489AbXD0WUs (ORCPT ); Fri, 27 Apr 2007 18:20:48 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1757497AbXD0WUs (ORCPT ); Fri, 27 Apr 2007 18:20:48 -0400 Received: from gw.goop.org ([64.81.55.164]:54707 "EHLO mail.goop.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757489AbXD0WTw (ORCPT ); Fri, 27 Apr 2007 18:19:52 -0400 Message-ID: <46327706.80508@goop.org> Date: Fri, 27 Apr 2007 15:19:50 -0700 From: Jeremy Fitzhardinge User-Agent: Thunderbird 1.5.0.10 (X11/20070302) MIME-Version: 1.0 To: Herbert Xu CC: Andi Kleen , Andrew Morton , virtualization@lists.osdl.org, lkml , Chris Wright , Christian Limpach Subject: Re: [PATCH 22/25] xen: xen-netfront: use skb.cb for storing private data References: <20070423215638.563901986@goop.org> <20070423215713.550883195@goop.org> <20070424014528.GA27789@gondor.apana.org.au> <462D88EF.9030007@goop.org> <20070424055731.GA30018@gondor.apana.org.au> In-Reply-To: <20070424055731.GA30018@gondor.apana.org.au> Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Herbert Xu wrote: > Sure thing. I'll look over it soon. > > Actually there is one thing I'd like to see changed first up: I noticed > that you've stripped out the checksum hack which is in the main Xen tree. > We actually have the code in net-2.6.22 (which is also in mm) that lets > you use CHECKSUM_PARTIAL on received packets without having to do that > hack. > > Here's the patch that I've been testing so far. It's against the Xen > source, but should be easy to adapt to your version as well. > > I just thought about this again, and in fact we need this change for > correctness as well as performance. Because not setting ip_summed > to CHECKSUM_PARTIAL in netfront is not going to stop netback from > sending CHECKSUM_PARTIAL packets to us. If these packets are then > routed/bridged back to netback, they'll have the wrong checksum. > OK, I've been sitting on this in the hope that I'll suddenly see the light and work out what you're talking about - but apparently that's not going to happen. So, some questions: 1. Does this patch change the dom0 <-> domU interface? or does it fix something that's currently broken? 2. Can just apply the netfront part to the pv_ops kernel, or does it require the corresponding dom0 patch to be applied as well? Thanks, J