Re: Why check legacy partition table for non-legacy HDs?

[Andrew Clausen]

On Sun, Sep 15, 2002 at 02:51:32PM -0700, H. J. Lu wrote:
> It doesn't make much senses to me. It is the same HD. 2 geometries are
> reported by 2 different BIOSes/drivers.

Wrong.  Only the one the BIOS is *currently* using is usable in
CHS mode.  (By currently, I mean where the HD is being used to
boot)

These numbers have nothing to do with hard disks, only BIOSes.

Of course, hardly anyone uses CHS nowadays, but still probably
enough to warn people when something is screwy.  In another couple
of years, I'd probably remove all this stuff.

> Both should be usable for that
> HD unless one of thme is bogus. Can you tell what is wrong with
> 9732/255/63 vs. 155114/16/63?

This is the nightmare scenario:

(1) person has MS-DOS (or Windows 95/98) installed on their machine,
and are using CHS
(2) they install Linux.  The Linux installer detects the BIOS
CHS geometry incorrectly.
(3) Parted uses the incorrect geometry to resize the MS-DOS
partition, and set the CHS incorrectly in the partition table
(4) Windows no longer boots

There are many variations on this theme.

Andrew