On Wed, 25 Feb 1998, Phil Nelson wrote:
> >I just tried to install 1.3 on a Pentium with an ahc1 and disk with 1980c,
> >15t, 80s last night, and I had serious troubles getting geometry that
> >NetBSD and the BIOS agreed on, and that the BIOS could actually use to
> >boot NetBSD.  I'm just wondering if these problems have been addressed, or
> >if I should file a big PR about sysinst, or if I should ask for some help.
> 
> You should not get ONE geometry that is used by both the BIOS and
> NetBSD.  You need to use the BIOS geometry for the MBR (Master Boot
> Record).  sysinst tries to guess what your BIOS are using.  With big
> disks, it is almost guaranteed not to be the real geometry of the
> disk.  Unfortunately, many newer bios report only the real geometry
> and not the fake geometry used for the MBR.

Okay, so you're saying that, in order to do a fresh install of NetBSD on a
disk, I should do approximately the following:

1) Run pfdisk or fdisk or something, to establish that NetBSD gets all of
the disk except the first track of the first cylinder, and to find out
what the MBR geometry is.

2) Run sysinst, and tell it what the MBR geometry is.

If this is true, then why can't sysinst do it itself, correctly?  Sysinst
does have fdisk functionality.  Shouldn't there be a way to get sysinst to
write a new MBR that will work nicely with NetBSD?

Sorry if I'm missing something obvious here; I just haven't been able to
get NetBSD to boot off the hard disk yet, and it seems to me that sysinst
should be slightly more helpful than it is.

Chris

-------------------------------------------------------------------------------
Chris Jones                                      cjones@rupert.oscs.montana.edu
           Mad scientist in training...
"Is this going to be a stand-up programming session, sir, or another bug hunt?"