From owner-freebsd-commit  Wed Apr 12 12:24:26 1995
Return-Path: commit-owner
Received: (from majordom@localhost)
          by freefall.cdrom.com (8.6.10/8.6.6) id MAA21181
          for commit-outgoing; Wed, 12 Apr 1995 12:24:26 -0700
Received: (from majordom@localhost)
          by freefall.cdrom.com (8.6.10/8.6.6) id MAA21168
          for cvs-sys-outgoing; Wed, 12 Apr 1995 12:24:23 -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 MAA21161
          ; Wed, 12 Apr 1995 12:23:59 -0700
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.9/8.6.9) id FAA23975; Thu, 13 Apr 1995 05:22:48 +1000
Date: Thu, 13 Apr 1995 05:22:48 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199504121922.FAA23975@godzilla.zeta.org.au>
To: ache@freefall.cdrom.com, phk@ref.tfs.com
Subject: Re: cvs commit: src/sys/scsi scsi_base.c
Cc: CVS-commiters@freefall.cdrom.com, ache@astral.msk.su,
        cvs-sys@freefall.cdrom.com, dufault@hda.com, julian@tfs.com
Sender: commit-owner@FreeBSD.org
Precedence: bulk

>The point really is that we print two geometries the same way,
>wd0 and sd0, the one for wd0 is useful and true, the one for sd0
>is bogus and not used anywhere, and we don't tell people about
>the difference...

Actually, the one for wd0 is useful (1) and bogus (2), while the
one for sd0 is useless (3) and true (4).  Both should be printed.

(1) I oversimplified.  It isn't useful if it differs from the BIOS
geometry.  E.g., ESDI drives sometimes say that they have 2 more
cylinders than they have, and ESDI and IDE drives may be configured
with a different geometry than the one claimed by the drive if the
BIOS supports this.

(2) I oversimplified.  It isn't bogus if it is the same as the
physical geometry; there are various levels of bogusness depending
on how well translations are hidden from the software.

(3) I oversimplified.  It's useful if the BIOS geometry is unknown
or wrong.  E.g., for my 4.3G drive, the BIOS geometry may be is
C=1023/H=64/S=32, which loses over 3/4 of the drive, while the
geometry printed by the recently restored printf is
"(8410200 S), 4076 C 20 H 103 S/T".  The problem will be worse for
drives larger than 8GB.

(4) I oversimplified.  The number of sectors/track is usually an
average...

Bruce