From owner-freebsd-current Mon Dec 14 17:37:04 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id RAA16241 for freebsd-current-outgoing; Mon, 14 Dec 1998 17:37:04 -0800 (PST) (envelope-from owner-freebsd-current@FreeBSD.ORG) Received: from dingo.cdrom.com (goldfish.pht.co.jp [210.171.55.12]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id RAA16033 for ; Mon, 14 Dec 1998 17:36:54 -0800 (PST) (envelope-from mike@dingo.cdrom.com) Received: from dingo.cdrom.com (localhost [127.0.0.1]) by dingo.cdrom.com (8.9.1/8.8.8) with ESMTP id DAA00486; Mon, 14 Dec 1998 03:18:44 -0800 (PST) (envelope-from mike@dingo.cdrom.com) Message-Id: <199812141118.DAA00486@dingo.cdrom.com> X-Mailer: exmh version 2.0.2 2/24/98 To: "David E. Cross" cc: "Jordan K. Hubbard" , Mikael Karpberg , Mike Smith , obrien@NUXI.com, current@FreeBSD.ORG Subject: Re: New drivers and install floppy space In-reply-to: Your message of "Tue, 08 Dec 1998 10:01:30 EST." Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Mon, 14 Dec 1998 03:18:44 -0800 From: Mike Smith Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG > On Tue, 8 Dec 1998, Jordan K. Hubbard wrote: > > > > > If we're going to go off the beaten track, we might as well do it > > > > properly. A 10M harddisk image is going to be a much better idea in > > > > that case. > > > > > > Probably... If all BIOSes that handle CDROM booting can handle that... > > > > Which is as good a time as any to make a *practical* suggestion and > > get us moving back in the direction of progress, I think. :-) > > > > Let's postpone this discussion for another 15 days or so, until Mike > > can return (to the US) from the Australian location he's currently > > occupying. He and I will look over his spec and see if we can't hack > > mkisofs into using a 10MB image rather than the 1.44MB image it uses > > My understanding of this has been that the BIOS treats the CDROM as a > virtual floppy drive for the purposes of booting off of it (which matches > my experience of specifying a floppy disk image as the boot file when > using mkisofs, and having that work). It treats it as a virtual BIOS disk of some sort, using a linear region on the disk to provide the raw sectors for the disk. > I think a perhaps better solution, > and more usefull to a wider variety of people would be to have a bootdisk > that would then mount its root filesystem off of the CDROM in a > conventional manor, and make that the 'bootfile' on the CDROM. So a > person with a BIOS that cannot boot off of CDROM (they still exist in > great numbers), would place the CD in the drive, have the CDboot floppy, > and still be in buisness. (This of course does not address those w/o a > CDROM). There are some gymnastic reasons why this isn't as easy as it might be, but theoretically it would be possible to suck a root filesystem image in from a temporarily-mounted, supported CDROM and then mount that image as the MFS root. Your code diffs to implement this, cleanly, would be very welcome. You might want to consider doing it in a fashion consistent with the way that we would be doing this for the bootable CDROM, ie. read the raw CDROM to find the boot image file, 'mount' the boot image file as a UFS filesystem, locate the 'mfsroot' file and read that; this would make things somewhat easier I expect. You would *definitely* want to write this code as a KLD module using a SYSINIT() that caused it to run between the device probes and the attemt to mount the root filesystem. -- \\ Sometimes you're ahead, \\ Mike Smith \\ sometimes you're behind. \\ mike@smith.net.au \\ The race is long, and in the \\ msmith@freebsd.org \\ end it's only with yourself. \\ msmith@cdrom.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message