From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stephen Hemminger Subject: Re: [PATCH 2/7] net: add FCoE offload support through net_device Date: Fri, 20 Feb 2009 15:27:44 -0800 Message-ID: <20090220152744.6ce9b9de@s6510> References: <20090219194734.22270.8445.stgit@zychengdu.jf.intel.com> <20090219194949.22270.40834.stgit.yi.zou@intel.com> <20090219132420.4b0dfc09@extreme> <7C88852EF6F99F4EB538472FCFEBE2222427B83D@orsmsx509.amr.corp.intel.com> <7C88852EF6F99F4EB538472FCFEBE2222427BC8F@orsmsx509.amr.corp.intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: "linux-scsi@vger.kernel.org" , "netdev@vger.kernel.org" To: "Zou, Yi" Return-path: Received: from mail.vyatta.com ([76.74.103.46]:60864 "EHLO mail.vyatta.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753774AbZBTX2b (ORCPT ); Fri, 20 Feb 2009 18:28:31 -0500 In-Reply-To: <7C88852EF6F99F4EB538472FCFEBE2222427BC8F@orsmsx509.amr.corp.intel.com> Sender: netdev-owner@vger.kernel.org List-ID: On Fri, 20 Feb 2009 15:21:42 -0800 "Zou, Yi" wrote: > >>On Thu, 19 Feb 2009 12:49:49 -0700 > >>Yi Zou wrote: > >> > >>> This adds a "struct net_fcoe_ops *fcoe_ops" to net_device struct so > >>any > >>> network adapter driver can provide Fiber Channle over Ethernet (FCoE) > >>offload > >>> support through net_device. The fcoe_ops is only available when FCoE > >>is > >>> enabled in kernel as built-in or module driver. > >>> > >>> Signed-off-by: Yi Zou > >> > >>Rather than adding another _ops structure please add to existing > >>netdevice_ops. You don't need the flags that way. If netdevice_ops > >>has fcoe_setup, then device can do it... > >The comments in struct net_device_ops says for management hooks, if > >that's an ok place for net_fcoe_ops, then, I am all for it. > > > >Thanks. > > > >yi > > Hi, Stephen, > Regarding your comment about adding net_fcoe_ops to net_device_ops, > the net_fcoe_ops contains function pointers ad well as data members, > where net_device_ops seems to me is the placeholder only for function > pointers. So I think it is still better to still leave > the net_fcoe_ops to net_device. Let me know what you think. > > Thanks, > yi Data members go in net_device (because they are per device instance). Put fcoe stuff in net_device_ops. Ideally, it should look like TSO and GRO; with standard ethtool type config?