I had the sane problem with my CATS install of 1.5 from the distribution
CDROM.
I had to boot my CATS system from a homemade cdrom and then manually gunzip
the .tgz files to get all of them to install.

----- Original Message -----
From: "Chris Gilbert" <chris@paradox.demon.co.uk>
To: "Jeremy Cooper" <jeremy@baymoo.org>; <port-arm32@netbsd.org>
Cc: "Mark Whittington" <markc@liquidev.com>
Sent: Sunday, January 28, 2001 5:10 AM
Subject: Re: DNARD, NetBSD 1.5, IDE disk: Data Abort when booting


> On Sunday 28 January 2001  7:55 am, Jeremy Cooper wrote:
> > On Sat, 27 Jan 2001, Mark Whittington wrote:
> > >   Boot device: /isa/ide/disk@0:\netbsd  Arguments:
> > >   Data Abort
> > >   ok
> >
> > I talked with Mark and we found the source of the problem: the
> > SHARKINST kernel for 1.5 mistakenly installs the GENERIC arm32
> > kernel, whereas it should install the SHARK arm32 kernel.  I've added an
> > entry to the arm32 FAQ about this.  We should find out how to fix it in
> > the next release.
>
> There's been some discussion that we'll be splitting out the arm32 port
into
> a generic bit, and then the platform parts so that we can have GENERIC
> kernels for each bit of hardware (currently that's about 5 different
ones..)
> Which, although it doesn't help now, should in the long term make this
less
> of an issue :)
>
> Chris