From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jesse Gross Subject: Re: [PATCH net] gso: do GSO for local skb with size bigger than MTU Date: Thu, 4 Dec 2014 15:19:47 -0800 Message-ID: References: <20141203090339.GA9299@redhat.com> <20141203183859.GB16447@redhat.com> <20141203220244.GA8822@casper.infradead.org> <20141203230551.GC8822@casper.infradead.org> <20141204011515.GA15666@casper.infradead.org> <20141204092650.GA13660@casper.infradead.org> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Cc: "Michael S. Tsirkin" , "Du, Fan" , Jason Wang , "netdev@vger.kernel.org" , "davem@davemloft.net" , "fw@strlen.de" , "dev@openvswitch.org" , Pravin Shelar To: Thomas Graf Return-path: Received: from na6sys009bog037.obsmtp.com ([74.125.150.111]:37631 "HELO na6sys009bog037.obsmtp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S1753069AbaLDXUJ (ORCPT ); Thu, 4 Dec 2014 18:20:09 -0500 Received: by mail-qc0-f175.google.com with SMTP id b13so15062625qcw.6 for ; Thu, 04 Dec 2014 15:20:08 -0800 (PST) In-Reply-To: <20141204092650.GA13660@casper.infradead.org> Sender: netdev-owner@vger.kernel.org List-ID: On Thu, Dec 4, 2014 at 1:26 AM, Thomas Graf wrote: > On 12/03/14 at 05:51pm, Jesse Gross wrote: >> I think it depends on where you put the PMTU check. If routing is >> happening in OVS where it is decomposed in several discrete actions >> like set MAC and decrement TTL then perhaps there is another explicit >> action to check the MTU. If it is happening in the context of the IP >> stack, then ICMP generation occurs automatically and if you get that >> if you write a flow to send a packet there. In each case, it seems >> like a flow would be steering you by way of an action to do routing so >> you would have fine grained control. I don't see this as conflicting >> with L3 ACLs in an L2 context in the same way that you don't have to >> automatically decrement the TTL. > > OK, I was under the impression that you would detect L3 behaviour > desire in OVS without an explicit action. Hence the worry on wrong > classification for L3 ACLs. > > Whether we mark the packet and defer the MTU check or we check right > in the action is an implementation detail in the end. I think we > agree on concept level that we need an API to control behaviour. Yes, I agree.