> On Thu, Feb 15, 2001 at 11:05:38AM -0500, NetBSD Bob wrote:
> > not to run Net-vs-Open on it.  Because of that silly disklabel
> > incompabibility with OpenBSD, that probably will say I need to go
> > with NetBSD on it, like on my VAXentoyz.
> > 
> These sorts of incompatibilies will complicate sharing disks between
> any ufs, but I don't expect openbsd/vax and netbsd/vax labels are
> completely incompatible yet. openbsd/vax is using a 16 filesystem
> label now, but it should still read the smaller form..  At least
> until you attempt to install boot code to it, at which point the
> structures will get plastered.

I think they are getting that way.  Even the VAX port seems that way,
so I think it may actually be a labelling incompatibility.  Drat!
I was hoping I could unroll OpenBSD from NetBSD and then compare them.
Well, it unrolled what I thought was correctly off netbsd (newfs/untar/
mkdev/etc), just like I unroll odd versions of NetBSD.  I tried every
way I could think of to get the OpenBSD boot blocks set on the NetBSD
system.  Nothing worked.  I even wrote the NetBSD boot and tried that.
Nothing.  Write the boot blocks from NetBSD and put the NetBSD kernel
on, and root fs, and away it goes into NetBSD.  Just enough of an
incompatiblity to not work.

> > > > Anyone know if the VAX ports in NetBSD or OpenBSD have this problem?
> > > > I tried unrolling a VAX OpenBSD port from NetBSD VAX, and it seemed
> > > > that the disklabels may have been off, because the system would just
> > > > not come up, at all, and jumped out to prom.
> > > 
> My guess is this was a boot problem rather than a label problem..
> Perhaps a netbsd boot{,.mop} didn't pass its information exactly as
> the openbsd kernel expected it to. Or it's also possible you tried
> a system we haven't been able to complete support for (yet). Our
> personal inventory of machines is not very great.

I was not netbooting, but actually hardwriting the file systems onto
a spare drive and then trying to get that to boot.  I would have
expected that there would have been some mix/match that would have
allowed that to happen, at least to some extent.  Heck, I use NetBSD
to boot Tahoe the first time on some of my Olden VAXen, just because
the labels and boots are close enough to work.  If nothing else, it
should come up with the boot message, and if better, the booting
kernel (until it might crash if drivers were off), or it might work
fine as is.  Right now, there is just enough of something off, that
it won't.

Hugh... can you provide a dd image of the first (boot+label+rootfs)
of OpenBSD VAX for an RZ25 disk (very common HD in VAXen).  That should
allow me to get enough of a system up to maybe hand wack it into place.
The machine is a MV3100/M20 (AFIK), with the usual MV3100 scsi controllers.
I would expect your OpenBSD suite to work on that, if it is not calling
any strange console bits, or doing something really strange in the boot.

Thanks

Bob