From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jiri Pirko Subject: Re: [PATCH net-next 1/8] net/flower: Introduce hardware offload support Date: Tue, 1 Mar 2016 18:01:28 +0100 Message-ID: <20160301170128.GG2098@nanopsycho.orion> References: <1456842290-7844-1-git-send-email-amir@vadai.me> <1456842290-7844-2-git-send-email-amir@vadai.me> <20160301144719.GA2098@nanopsycho.orion> <56d5c71f.a151c20a.cf0a8.ffffd4ad@mx.google.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: "David S. Miller" , netdev@vger.kernel.org, Or Gerlitz , John Fastabend , Saeed Mahameed , Hadar Har-Zion , Jiri Pirko To: Amir Vadai Return-path: Received: from mail-wm0-f68.google.com ([74.125.82.68]:36449 "EHLO mail-wm0-f68.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751359AbcCARBb (ORCPT ); Tue, 1 Mar 2016 12:01:31 -0500 Received: by mail-wm0-f68.google.com with SMTP id l68so5142580wml.3 for ; Tue, 01 Mar 2016 09:01:31 -0800 (PST) Content-Disposition: inline In-Reply-To: <56d5c71f.a151c20a.cf0a8.ffffd4ad@mx.google.com> Sender: netdev-owner@vger.kernel.org List-ID: Tue, Mar 01, 2016 at 05:49:27PM CET, amir@vadai.me wrote: >On Tue, Mar 01, 2016 at 03:47:19PM +0100, Jiri Pirko wrote: >> Tue, Mar 01, 2016 at 03:24:43PM CET, amir@vadai.me wrote: >> >This patch is based on a patch made by John Fastabend. >> >It adds support for offloading cls_flower. >> >A filter that is offloaded successfuly by hardware, will not be added to >> >the hashtable and won't be processed by software. >> >> That is wrong. User should explitly specify to not include rule into sw >> by SKIP_KERNEL flag (does not exist now, with John's recent patch we'll >> have only SKIP_HW). Please add that in this patchset. >Why? If a rule is offloaded, why would the user want to reprocess it by >software? >If the user use SKIP_HW, it will be processed by SW. Else, the user >would want it to be processed by HW or fallback to SW. I don't >understand in which case the user would like to have it done twice. For example if you turn on the offloading by unsetting NETIF_F_HW_TC. Or if someone inserts skbs into rx path directly, for example pktgen. We need SKIP_KERNEL to be set by user, not implicit.