From mboxrd@z Thu Jan 1 00:00:00 1970 From: Zoltan Kiss Subject: Re: [3.15-rc3] Bisected: xen-netback mangles packets between two guests on a bridge since merge of "TX grant mapping with SKBTX_DEV_ZEROCOPY instead of copy" series. Date: Wed, 30 Apr 2014 21:40:55 +0100 Message-ID: <53615FD7.8090803@citrix.com> References: <395225650.20140430124506@eikelenboom.it> <1398871446.29914.163.camel@edumazet-glaptop2.roam.corp.google.com> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8"; format=flowed Content-Transfer-Encoding: 7bit Cc: Ian Campbell , "David S. Miller" , , To: Eric Dumazet , Sander Eikelenboom Return-path: Received: from smtp02.citrix.com ([66.165.176.63]:26619 "EHLO SMTP02.CITRIX.COM" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1759308AbaD3Uk7 (ORCPT ); Wed, 30 Apr 2014 16:40:59 -0400 In-Reply-To: <1398871446.29914.163.camel@edumazet-glaptop2.roam.corp.google.com> Sender: netdev-owner@vger.kernel.org List-ID: On 30/04/14 16:24, Eric Dumazet wrote: > On Wed, 2014-04-30 at 12:45 +0200, Sander Eikelenboom wrote: > > This looks like skb->data_len is not properly set > > (Small note : > alloc_page(GFP_ATOMIC|__GFP_COLD) should be a plain > alloc_page(GFP_ATOMIC) : We prefer a page that has been used recently to > avoid cache misses if possible. > ) > > This probably confuses skb_try_coalesce() > > My guess is xenvif_get_requests() forgets to init nskb->len / > nskb->data_len for the skb that is attached to frag_list xenvif_get_requests() calls alloc_skb() which inits these values to zero > > So the "skb->data_len += nskb->len;" at line 1353 is probably a nop.