Re: bootable disk: how?

[der Mouse <mouse%Rodents.Montreal.QC.CA@localhost>]

>>> Possibly something is confused in the CHS/LBA mapping, using
>>> mbr_ext will force LBA disk reads.
>> Hmm.  Could this be related to the "can't determine number of heads"
>> message?
> If fdisk is reporting that, then (I think) the disk isn't being
> reported by the bios (in the list of known disks).  This can be
> confirmed using 'sysctl machdep.diskinfo' which should report the
> disks and their corresponding bios disk number (0x8x) together with
> the size and geometry reported by the bios.

On the 4.0 machine (where the disk is currently connected), I tried
that and got

machdep.diskinfo: 80:17921834(1023/255/63),2  wd0 sd0:80

So wd0 appears, but without anything that looks like a BIOS disk number
or geometry info.

> If the disk isn't reported by the bios, then fdisk has to solve the
> simultaenous equations from the mbr partition bounds to determine the
> geometry used to set the mbr - which it assumes is the correct bios
> geometry.

> It might be that the bios is actually trying to boot a different
> disk!

Well, when it fails, it's the only disk available, so it would be more
like "the BIOS is refusing to boot that disk, and has nothing else
available, so it errors".

So, the problem seems to be that the BIOS is doing something wrong.
Since it happens on two very different machines with very different
BIOSes, I am inclined to ascribe it to something wrong with the data on
the disk; is this plausible?  I'll try stomping /dev/zero over the
first few megabytes of the disk and using the 4.0 install CD to see if
I can render the disk bootable that way.  It'll mean another long data
copy operation, but it sure ain't working the way it is....

/~\ The ASCII                           der Mouse
\ / Ribbon Campaign
 X  Against HTML               mouse%rodents.montreal.qc.ca@localhost
/ \ Email!           7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B