From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jakub Kicinski Subject: Re: [PATCH net] net: sched: cls: Fix offloading when ingress dev is vxlan Date: Wed, 6 Jun 2018 08:57:57 -0700 Message-ID: <20180606085757.4f17e03b@cakuba.netronome.com> References: <1528185843-18645-1-git-send-email-paulb@mellanox.com> <20180605115747.0e939ac4@cakuba.netronome.com> <20180605.150640.12956507851401975.davem@davemloft.net> <20180605142700.6033a7a0@cakuba.netronome.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: David Miller , jiri@mellanox.com, xiyou.wangcong@gmail.com, jhs@mojatatu.com, netdev@vger.kernel.org, kliteyn@mellanox.com, roid@mellanox.com, shahark@mellanox.com, markb@mellanox.com, ogerlitz@mellanox.com To: Paul Blakey Return-path: Received: from mx3.wp.pl ([212.77.101.9]:43181 "EHLO mx3.wp.pl" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932183AbeFFP6G (ORCPT ); Wed, 6 Jun 2018 11:58:06 -0400 In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On Wed, 6 Jun 2018 10:59:30 +0300, Paul Blakey wrote: > Maybe we can apply my patch logic of still trying the egress dev if the > block has a single device, and not shared. Is that ok with you? I don't remember that patch but sounds pretty bad. > You're patch seems good as an add on, but the egress hw device (sw1np0) > would go over the tc actions and see if it can offload such rule (output > to software lo device) and would fail anyway. Ack, I'm not trying to solve the skip_sw problem. It's a hard one.