From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Tue, 15 May 2001 17:20:48 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Tue, 15 May 2001 17:20:38 -0400 Received: from neon-gw.transmeta.com ([209.10.217.66]:13066 "EHLO neon-gw.transmeta.com") by vger.kernel.org with ESMTP id ; Tue, 15 May 2001 17:20:26 -0400 Message-ID: <3B019D8A.86C18D1C@transmeta.com> Date: Tue, 15 May 2001 14:20:10 -0700 From: "H. Peter Anvin" Organization: Transmeta Corporation X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.4.5-pre1-zisofs i686) X-Accept-Language: en, sv, no, da, es, fr, ja MIME-Version: 1.0 To: Tim Fletcher CC: "H. Peter Anvin" , linux-kernel@vger.kernel.org Subject: Re: Device Numbers, LILO In-Reply-To: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Tim Fletcher wrote: > > > It's in, but for some strange reason you have to ask for it explicitly > > with the "lba32" option. > > Because the 32bit bios calls lilo uses in lba32 mode can cause problems > with broken or old bios's hence is defaults to a safe option, and if you > can't boot without it (over 1023 cylinders) then you turn it on at your > own risk. > > I know this from the experiance of breaking lilo on my workstation :) > The problem is that LILO tries one or the other, instead of dynamically using whichever one it needs. Heck, you may not be able to make that decision until boot time anyway! You can then either always try LBA calls and see if they are available, or default to CHS calls unless cylinder > 1023. LILO tries to do *way* too much at install time. The fact that you have to specify "linear" explicitly is a joke -- there is no excuse for !linear. -hpa -- at work, in private! "Unix gives you enough rope to shoot yourself in the foot." http://www.zytor.com/~hpa/puzzle.txt