From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: [PATCH 06/30] solos: Handle attribute show/store in kernel more sanely Date: Sat, 21 Mar 2009 13:22:39 -0700 (PDT) Message-ID: <20090321.132239.74116822.davem@davemloft.net> References: <1237325355.27681.335.camel@macbook.infradead.org> <20090317154411.598b9f00@nehalam> <1237330172.27829.14.camel@macbook.infradead.org> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: shemminger@vyatta.com, netdev@vger.kernel.org To: dwmw2@infradead.org Return-path: Received: from 74-93-104-97-Washington.hfc.comcastbusiness.net ([74.93.104.97]:51504 "EHLO sunset.davemloft.net" rhost-flags-OK-FAIL-OK-OK) by vger.kernel.org with ESMTP id S1752605AbZCUUWw (ORCPT ); Sat, 21 Mar 2009 16:22:52 -0400 In-Reply-To: <1237330172.27829.14.camel@macbook.infradead.org> Sender: netdev-owner@vger.kernel.org List-ID: From: David Woodhouse Date: Tue, 17 Mar 2009 22:49:32 +0000 > On Tue, 2009-03-17 at 15:44 -0700, Stephen Hemminger wrote: > > Since these are generic, please use some form of netlink for this > > instead of sysfs. > > > > Bonding was the first mistake (trying to use sysfs for a transactional > > API). > > Pls don't repeat the mistake. > > It's not a transactional API. > > The sysfs model actually fits the communication with the hardware almost > perfectly -- there are a bunch of named parameters, all of which can be > read and a few of which can be written to. A sysfs read or write > translates directly into a single exchange with the device's firmware. Fair enough, I'll pull this into net-next-2.6