From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christoph Hellwig Subject: Re: replace deprecated hosts.h file Date: Sun, 30 May 2004 09:46:45 +0100 Sender: linux-scsi-owner@vger.kernel.org Message-ID: <20040530084645.GA29411@infradead.org> References: <20040529142126.GA31365@devserv.devel.redhat.com> <40B8A7FA.8070802@pobox.com> <20040529151431.GD31365@devserv.devel.redhat.com> <40B8B383.4050602@pobox.com> <20040529164056.GH31365@devserv.devel.redhat.com> <1085880986.18642.351.camel@compaq.xsintricity.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from [213.146.154.40] ([213.146.154.40]:43207 "EHLO pentafluge.infradead.org") by vger.kernel.org with ESMTP id S261932AbUE3Iqt (ORCPT ); Sun, 30 May 2004 04:46:49 -0400 Content-Disposition: inline In-Reply-To: <1085880986.18642.351.camel@compaq.xsintricity.com> List-Id: linux-scsi@vger.kernel.org To: Doug Ledford Cc: Arjan Van de Ven , Jeff Garzik , linux-scsi mailing list , James Bottomley On Sat, May 29, 2004 at 09:36:26PM -0400, Doug Ledford wrote: > core kernel code == specifically tied to kernel major/minor > hardware driver code != tied to kernel major/minor In Linux it in fact is. Especially when the subsystem has big changes like scsi in the 2.4->2.6 transition. We need to clean out the letovers, and we want to finish that in early 2.7. Getting these changes in earlier tends to make life much easier, but I'll do another pass once 2.7 is out - if a maintainer wants to reject those changes then his driver will simply break.