From mboxrd@z Thu Jan 1 00:00:00 1970 From: Du Fan Subject: [Discussion] About over-MTU-sized skb in virtualized env Date: Wed, 03 Dec 2014 14:39:37 +0800 Message-ID: <547EB029.5010102@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Thomas Graf , "Michael S. Tsirkin" , Jesse Gross , Flavio Leitner , "davem@davemloft.net" , pshelar@nicira.com, netdev , "dev@openvswitch.org" , "Du, Fan" To: Florian Westphal Return-path: Received: from mail-pa0-f50.google.com ([209.85.220.50]:58232 "EHLO mail-pa0-f50.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750707AbaLCGmU (ORCPT ); Wed, 3 Dec 2014 01:42:20 -0500 Received: by mail-pa0-f50.google.com with SMTP id bj1so15149866pad.37 for ; Tue, 02 Dec 2014 22:42:20 -0800 (PST) Sender: netdev-owner@vger.kernel.org List-ID: Sorry for resend this mail, because my company email is rejected by net= dev. Hi Florian 214 static int ip_finish_output_gso(struct sk_buff *skb) 215 { 216 netdev_features_t features; 217 struct sk_buff *segs; 218 int ret =3D 0; 219 220 /* common case: locally created skb or seglen is <=3D mtu */ 221 if (((IPCB(skb)->flags & IPSKB_FORWARDED) =3D=3D 0) || 222 skb_gso_network_seglen(skb) <=3D ip_skb_dst_mtu(skb)) 223 return ip_finish_output2(skb); Could you please state _concrete_ reason why locally created skb length is =EF=BC=BFalways=EF=BC=BF fitting into MTU size? or why we nee= ds this checking. Especially in below scenario: http://docs.openstack.org/admin-guide-cloud/content/figures/14/a/a/comm= on/figures/under-the-hood-scenario-1-ovs-compute.png As many people have discuss it in this thread: http://marc.info/?t=3D141715669700004&r=3D1&w=3D2