From owner-freebsd-bugs  Wed Oct 16 02:40:06 1996
Return-Path: owner-bugs
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id CAA00856
          for bugs-outgoing; Wed, 16 Oct 1996 02:40:06 -0700 (PDT)
Received: (from gnats@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id CAA00818;
          Wed, 16 Oct 1996 02:40:03 -0700 (PDT)
Resent-Date: Wed, 16 Oct 1996 02:40:03 -0700 (PDT)
Resent-Message-Id: <199610160940.CAA00818@freefall.freebsd.org>
Resent-From: gnats (GNATS Management)
Resent-To: freebsd-bugs
Resent-Reply-To: FreeBSD-gnats@freefall.FreeBSD.org,
        Received:  (from nobody@localhost)by.freefall.freebsd.org.id.CAA00287;Wed; (8.7.5/8.7.3);,
        16 Oct 1996 02:36:05.-0700 (PDT)
Message-Id: <199610160936.CAA00287@freefall.freebsd.org>
Date: Wed, 16 Oct 1996 02:36:05 -0700 (PDT)
From: chad@dcfinc.com
To: freebsd-gnats-submit@freebsd.org
X-Send-Pr-Version: www-1.0
Subject: i386/1821: boot fails if bad144 selected on large partition
Sender: owner-bugs@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk


>Number:         1821
>Category:       i386
>Synopsis:       boot fails if bad144 selected on large partition
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    freebsd-bugs
>State:          open
>Class:          doc-bug
>Submitter-Id:   current-users
>Arrival-Date:   Wed Oct 16 02:40:01 PDT 1996
>Last-Modified:
>Originator:     Chad R. Larson
>Organization:
DCF, Inc.
>Release:        2.1.0
>Environment:
>Description:
During the "Novice" install, when the "fdisk" utility was run, I 
selected the "bad block check" option.  This caused the BIOS booter
to fail with a "Error C:3197 > 1023 (BIOS limit)" message.

This turns out to be caused by the BIOS booter attempting to read the
bad block table, which is at the end of the "c" slice of the partition.

I fixed it by using the fixdisk floppy's fdisk to write a new partition
entry that does not specify bad block processing.

This is perhaps a documentation problem.  The bad track processing
option is incompatable with a FreeBSD partition that extends beyond
cylinder 1023, regardless of where the root slice resides within that
partition.  I don't find that limitation documented anywhere.  I found
it by reading the source code to the BIOS booter.

Better might be a second-level booter that doesn't use the BIOS to do
the disk reads (Interactive UNIX does this), so that only the 
second-level booter itself need be below the 1023 cylinder fence.
>How-To-Repeat:
Select the "bad block mapping" option from the fdisk menu on a
partition that extends beyond the 1023 cylinder boundry, regardless
of the size of the root slice within that partition.
>Fix:
Don't use bad144 mapping.  Tell your friends...
>Audit-Trail:
>Unformatted: