Subject: Re: fdisk in 1.3I too trusting of disk?
To: David W Talmage <talmage@cmf.nrl.navy.mil>
From: Antti Kantee <pooka@iki.fi>
List: port-i386
Date: 04/05/1999 22:41:03
On Mon Apr 05 1999 at 12:41:10 -0400, David W Talmage wrote:

> While trying to use fdisk to reserve a partition for suspending memory
> to disk, I managed to give it bogus values (they seemed reasonable at
> the time!) and save them to disk.  Now fdisk complains of a floating
> exception and dumps core.  I recompiled fdisk with the debugger
> symbols and discovered that it thinks that the number of heads is zero
> in intuit_translated_geometry().
> 
> How can I undo this without hacking fdisk?  I tried the -b and -s
> flavors of fdisk but they insist on calling
> inuit_translated_geometry(), too.  It would be easy to make fdisk not
> call intuit_translated_geometry() if the -b flag is present.  Would
> that be a Bad Thing?

I once managed to get this behaviour out of fdisk with a disk I had
'fixed a bit'. My solution to this problem was giving the disk some of
/dev/zero, but obviously this solution isn't applicable to every
situation.

Hope this helps in some way..

-- 
  "Never underestimate the power of the Lite side of the \Source/"
 Antti Kantee   /       NetBSD        \  Citius, Altius,  \    / o
<pooka@iki.fi> / http://www.netbsd.org \     Levytys       \  /  |
______________/ Free Multiplatform UN*X \___________________\/__ |