From owner-freebsd-hackers  Wed Jun  7 02:41:26 1995
Return-Path: hackers-owner
Received: (from majordom@localhost)
          by freefall.cdrom.com (8.6.10/8.6.6) id CAA11479
          for hackers-outgoing; Wed, 7 Jun 1995 02:41:26 -0700
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.34])
          by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id CAA11461
          for <hackers@freebsd.org>; Wed, 7 Jun 1995 02:40:57 -0700
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.9/8.6.9) id TAA11194; Wed, 7 Jun 1995 19:38:58 +1000
Date: Wed, 7 Jun 1995 19:38:58 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199506070938.TAA11194@godzilla.zeta.org.au>
To: hackers@freebsd.org, hoffman@Ipsilon.COM
Subject: Re: 2.0.5 scsi boot configuration whine
Sender: hackers-owner@freebsd.org
Precedence: bulk

>I have a single drive on an NCR SCSI with an unfortunate geometry:
>2234/7/68. It doesn't seem to matter if the FreeBSD starts at 0, 4
>(which is where 2.0.5 puts it), or 68 (which is where 2.0.5 will put
>it after being corrected about the geometry using "G" in the new
>fdisk). when corrected about the geometry, it will come up next time
>as 37978/7/4, the same total number of blocks, but using BIOSs 6 bit
>sector count.

The sector count gets truncated to 6 bits when it is stored in the
partition table, so it must be < 64 to work.  fdisk and sysinstall
should reject sector counts >= 64 (and head counts > 256, and cylinder
counts >= 65536; head counts of 256 requires special handling if they
work at all, cylinder counts >= 1024 require special handling but
should work).

Try telling sysinstall an equivalent geometry, e.g., 2234/28/17, or
boot with -v to see what the BIOS thinks is the geometry.

Bruce