From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jiri Benc Subject: Re: [PATCH net-next v12 5/9] openvswitch: add processing of L3 packets Date: Wed, 19 Oct 2016 18:52:06 +0200 Message-ID: <20161019185206.7347e189@griffin> References: Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: Linux Kernel Network Developers , ovs dev , Lorand Jakab , Simon Horman To: Pravin Shelar Return-path: Received: from mx1.redhat.com ([209.132.183.28]:35082 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756690AbcJSQwL (ORCPT ); Wed, 19 Oct 2016 12:52:11 -0400 In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On Tue, 18 Oct 2016 22:13:45 -0700, Pravin Shelar wrote: > On Mon, Oct 17, 2016 at 6:02 AM, Jiri Benc wrote: > > - skb_reset_network_header(skb); > > + skb->protocol = parse_ethertype(skb); > > I am not sure about changing skb->protocol here. > By changing this skb loosing information about packet type. Therefore > if packet re-enters OVS (through different bridge), this packet would > look like L3 packet. function key_extract_mac_proto() would not see > TEB type packet. This should be okay. If the packet is sent out to an Ethernet interface (whatever interface it is), skb->protocol needs to contain the payload type. We're not interested in ETH_P_TEB. If the packet is sent out to an ARPHRD_NONE interface, ETH_P_TEB is pushed back. Basically, what we're doing here is unconditionally converting ETH_P_TEB packets *coming from ARPHRD_NONE interfaces* (this is important) into regular Ethernet packets. Which is exactly what we want. Am I missing something? Jiri