Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Thu, 7 Dec 2000 10:08:46 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Thu, 7 Dec 2000 10:08:36 -0500 Received: from router-100M.swansea.linux.org.uk ([194.168.151.17]:20491 "EHLO the-village.bc.nu") by vger.kernel.org with ESMTP id ; Thu, 7 Dec 2000 10:08:24 -0500 Subject: Re: [RFC-2] Configuring Synchronous Interfaces in Linux To: jgarzik@mandrakesoft.com (Jeff Garzik) Date: Thu, 7 Dec 2000 14:40:04 +0000 (GMT) Cc: alan@lxorguk.ukuu.org.uk (Alan Cox), khc@intrepid.pm.waw.pl (Krzysztof Halasa), linux-kernel@vger.kernel.org In-Reply-To: <3A2F9AD4.830153B4@mandrakesoft.com> from "Jeff Garzik" at Dec 07, 2000 09:12:36 AM X-Mailer: ELM [version 2.5 PL1] MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-Id: From: Alan Cox Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org > > struct hdlc_protocol > > struct fr_protocol > > struct eth_physical > > Not yet another one for eth... We now have ethtool for this. And a > generic netdevice::set_config wrapper can be created that simply calls > the ethtool ioctl with the proper info and locking. There I disagree. I would do it the other way up. Post 2.4 you make the ethtool ioctl simply build an eth_physical and run in that way. Right now the ethtool stuff is a bit hackish and cannot handle upcoming real world situations such as setting the cryptokey for onboard crypto on ethernet cards, or handling devices that present themselves as ethernet but are not at the physical layer being remotely honest about it. Alan - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org Please read the FAQ at http://www.tux.org/lkml/