Subject: Re: BIOS / pfdisk / sysinst glitches ?
To: Danny Thomas <D.Thomas@vthrc.uq.edu.au>
From: None <rvb@sicily.odyssey.cs.cmu.edu>
List: port-i386
Date: 07/12/1998 16:16:55
Thank you, for the log.

D.Thomas@vthrc.uq.edu.au (Danny Thomas) writes:

> 1) Geometry translation in the BIOS is from 6256*16*63 to 782*128*63
This means you are playing LBA games

> 2) I used pfdisk to setup 4 partitions, so
pfdisk is smart, it will skip the first head if you start at cylinder 0.

>    1    7    1  176  <= NB I was leaving first cylinder free

> Now the curious thing is the installer reports geometry as
>   real geom 6256 cyl, 128 heads, 63 sec
>   BIOS geom 782 cyl, 128 heads, 63 cyl
Yes, I have seen this and the related problem you mention below.  I
will look into it.  (It has recently been explained what the idea was
behind this code.) I
can't remember, if the way to fix this was to reset the real geometry.
(or use 1.2 boot floppies to set the disklabel).

> when I think the real heads should be 16
> 
> If I change the type of partition for 4) to be not 0xA5, the real geometry
> does report 16 heads! Is this a problem with sysinst or pfdisk?
> 
> If you continue with sysinst still thinking the disk has 128 heads, doing
> the setup into the fourth partition on the disk, here's what you see in the
> end, NB partition (d), cheapest way I saw 18G appear on a disk:
>   a  153    2082  ... /
>   b  263    2236
>   c  992    2082
>   d  24,633    0
>   e  574    2500  ... /usr
I never dared let it proceed to completion here, so I never saw the
problem you indicate below.

> 
> what's more, the first partition in the MBR seems to get trashed to type 0,
> start 0, size 0.
> 
> 
> 3) when sysinst is listing the partitions, I saw:
>    0     8  1408  1416
>    1  1416  1408  2824
>    2  2824  1408  4232
>    3  4232  2016  6248
> I guess it depends on the definition of 'end', but I'd say partition 0
> spanned sectors 8 thru 1415 and so on. When I first saw this I was
> concerned that maybe partitions were overlapping.
> 
> 
> 4) I doubt it's related, but it's frustrating when new BIOS versions get
> too clever. In the case of the Award BIOS for the FIC PA-2007 motherboard,
> the later versions 109CD12 & 113CD13 truncate the disk parameters to the
> 528M 'limit'. What's bizarre is you ask the BIOS to go thru the
In my experience, if you do not set LBA, this will happen.  If you do
then it does as you indicate below.
> auto-disk-config routine and select the preferred option (782*128*63) it
> gives you*, yet that gets changed to 511*32*63 just before the BIOS hands
> off to the boot routine. This doesn't happen with the BIOS version 618J90W.
> But that version does not support 64M SDRAMs, etc.
> 
> * it provides three choices, one each for LBA, Normal & Large.
USE LBA
> 
> 
> anyway, after getting the partitions OK and NetBSD installed first, I've
> added NT and FreeBSD with no apparent problems.
Beware NT,  I've seen it do odd things to the MBR and friend.
Especially, if you go into the disk administrator.  The obvious
safeguard is to safe a copy of the MBR (fdisk) and disklabel