From owner-freebsd-current Sun Nov 24 15:00:25 1996 Return-Path: owner-current Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id PAA17312 for current-outgoing; Sun, 24 Nov 1996 15:00:25 -0800 (PST) Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19]) by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id PAA17247 for ; Sun, 24 Nov 1996 15:00:09 -0800 (PST) Received: (from bde@localhost) by godzilla.zeta.org.au (8.8.3/8.6.9) id JAA06292; Mon, 25 Nov 1996 09:45:03 +1100 Date: Mon, 25 Nov 1996 09:45:03 +1100 From: Bruce Evans Message-Id: <199611242245.JAA06292@godzilla.zeta.org.au> To: jkh@time.cdrom.com, phk@critter.tfs.com Subject: Re: 2.2-ALPHA install failure Cc: freebsd-current@freebsd.org, joerg_wunsch@uriah.heep.sax.de Sender: owner-current@freebsd.org X-Loop: FreeBSD.org Precedence: bulk >>Or we could get Poul-Henning to DTRT with libdisk - that would >>be a concept, and even the correct place to handle it. :-) > >I actually looked at this, and there are some comments in order: > >1. I'm still not sure we know what TRT is. Although the MB/64hd/32sect >sounds like the best bet to me. No, it gratuitously breaks booting with kernels above 1GB. If you leave the partition table empty or put only invalid entries in it, then some BIOSes (ncr) will DTRT and invent a geometry with <= 1024 cylinders if possible. >2. Even if we did, I'd need a very dedicated tester with an NCR >controller to help me test it out on the only HW we know of that >fails. You mean with an NCR BIOSes and many lesser BIOSes. The NCR BIOS accepts anything reasonable. It's going in the other direction that is the problem - you can easily create configurations that work with NCR but aren't portable. Bruce