From mboxrd@z Thu Jan 1 00:00:00 1970 From: Vincent JARDIN Subject: Re: [PATCH v5 00/29] Support VFD and DPDK PF + kernel VF on i40e Date: Mon, 16 Jan 2017 00:32:21 +0100 Message-ID: <9cc4dbde-bc9c-7557-bcff-c88a8d04c709@6wind.com> References: <4341B239C0EFF9468EE453F9E9F4604D3C5CA3DA@shsmsx102.ccr.corp.intel.com> <586D647A.5040607@research.att.com> <4341B239C0EFF9468EE453F9E9F4604D3C5D9598@shsmsx102.ccr.corp.intel.com> <58795CE3.4080403@gmail.com> <6FE891CC-F927-4719-B7F8-BCF8EB4DEF9F@research.att.com> Mime-Version: 1.0 Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 8bit Cc: "Chen, Jing D" , "DANIELS, EDWARD S (EDWARD)" , "dev@dpdk.org" , "ZELEZNIAK, ALEX" , "Fastabend, John R" , Ferruh Yigit To: "JOSHI, KAUSTUBH (KAUSTUBH)" , John Fastabend Return-path: Received: from mail-wm0-f45.google.com (mail-wm0-f45.google.com [74.125.82.45]) by dpdk.org (Postfix) with ESMTP id 462FBF967 for ; Mon, 16 Jan 2017 00:32:24 +0100 (CET) Received: by mail-wm0-f45.google.com with SMTP id r126so137991789wmr.0 for ; Sun, 15 Jan 2017 15:32:24 -0800 (PST) In-Reply-To: <6FE891CC-F927-4719-B7F8-BCF8EB4DEF9F@research.att.com> List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" Thanks John for the kernel support. I'd like to add one thought on this need from Joshi: > Also, the kernel drivers have no concept of passing VF messages to upstream "decision making” (or policy enforcement) software like VFd. It shall be possible to define some Netlink or policy management messages over the PF, so a userland process can police and provide the "decision making" in a standard way that would be generic. Joshi, John is confirming that the kernel path is possible. Please, if you see other requirements, keep listing them. Le 14/01/2017 à 00:49, JOSHI, KAUSTUBH (KAUSTUBH) a écrit : > However, in theory, this *can* be done through an extra VF as well, so I didn't bring up that particular point as being a deciding factor in this debate. It remains an important item: it shows that thru the kernel, all technical requirements are manageable, including the steering of traffic. I'll continue the comments on the patch on the v9's thread. http://dpdk.org/ml/archives/dev/2017-January/054953.html Best regards, Vincent