From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga09.intel.com (mga09.intel.com [134.134.136.24]) by dpdk.org (Postfix) with ESMTP id 81D9E9ACD for ; Tue, 1 Mar 2016 23:40:31 +0100 (CET) Received: from orsmga001.jf.intel.com ([10.7.209.18]) by orsmga102.jf.intel.com with ESMTP; 01 Mar 2016 14:40:31 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.22,524,1449561600"; d="scan'208";a="898862001" Received: from bricha3-mobl3.ger.corp.intel.com ([10.122.33.133]) by orsmga001.jf.intel.com with SMTP; 01 Mar 2016 14:40:25 -0800 Received: by (sSMTP sendmail emulation); Tue, 01 Mar 2016 22:40:24 +0025 Date: Tue, 1 Mar 2016 22:40:24 +0000 From: Bruce Richardson To: Jay Rolette Message-ID: <20160301224023.GA30516@bricha3-MOBL3> References: <1453911849-16562-1-git-send-email-ferruh.yigit@intel.com> <56D420E5.9010802@intel.com> <56D42462.3020905@scylladb.com> <1945473.Tiatd2m80T@xps13> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: Organization: Intel Shannon Ltd. User-Agent: Mutt/1.5.23 (2014-03-12) Cc: DPDK , Avi Kivity Subject: Re: [dpdk-dev] [PATCH 1/3] kcp: add kernel control path kernel module X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: patches and discussions about DPDK List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 01 Mar 2016 22:40:31 -0000 On Mon, Feb 29, 2016 at 08:33:25AM -0600, Jay Rolette wrote: > On Mon, Feb 29, 2016 at 5:06 AM, Thomas Monjalon > wrote: > > > Hi, > > I totally agree with Avi's comments. > > This topic is really important for the future of DPDK. > > So I think we must give some time to continue the discussion > > and have netdev involved in the choices done. > > As a consequence, these series should not be merged in the release 16.04. > > Thanks for continuing the work. > > > > I know you guys are very interested in getting rid of the out-of-tree > drivers, but please do not block incremental improvements to DPDK in the > meantime. Ferruh's patch improves the usability of KNI. Don't throw out > good and useful enhancements just because it isn't where you want to be in > the end. > > I'd like to see these be merged. > > Jay +1 to this. While this may not eliminate out of tree kernel modules, and solve all problems, I think taking in kcp and kdp and removing KNI will leave DPDK in a better state than it was. Also, with regards to having the kernel data path, and the port control part inside the same module, our experience with KNI has led us to explicitly separating them out. The path from user to kernel space should be completely separated from the netdevs which back dpdk ports. Consider stats reporting alone: netdevs backed by dpdk ports can report out packet rx/tx counts for the hw ports while the user-kernel path can report out packet rx/tx from kernel. Regards, /Bruce