Subject: Re: Earliest VAX port?
To: Mirian Crzig Lennox <mirian@cosmic.com>
From: Lord Isildur <mrfusion@vaxpower.org>
List: port-vax
Date: 12/27/2000 17:22:39
This is why the vsbus mechanism was introduced in netbsd. The small VAXen 
(vaxstations, etc) are all effecitvely busless, or at least, busless in 
the lcassical VAX sense. The devices are hooked directly to the CDAL bus 
but it's not in the VAX hierarchy that the older iron used (as per the 
design philosophy described in VARM and used inthe larger machines): 
Nexus interconnect, where processors, perhaps memories, and bus adapters 
reside. devices hung from bus adapters, perhaps through controllers on 
the busses. The busless vaxen have things hooked up directly tot he CDAL, 
which sorta breaks this hierarchy in exchange for fitting everythign on a 
small board and making it cheap. 4BSD was designed around the classic VAX 
model, and so it does not have any support for any method of connecting a 
device except thru a bus. Since this makes a really orthogonal and 
consistent model, and is really nice both for kernel organization and for 
development, the model was kept in NetBSD and the vsbus was brought into 
being
as an abstraction for the cdal-attached devices. To get 4.3 running on 
them, 
some similar mechanism needs to be worked in. I just got a vs4k60, so once
i return from vacation, im planning on seeing if i can shoehorn at least 
some kind of 4.3 (tahoe probably) onto it. NetBSD stuff cant be used 
without modification because the memory system in NetBSD was first based 
on 4.4 (which was based on mach, to some extent), and in the past year 
and a half heavily modified even from that. Still, a driver is a driver. 

if youve got time and hardware, you might want to start hacking at it too :)

happy hacking,
isildur


On 20 Dec 2000, Mirian Crzig Lennox wrote:

> In article <200012201541.KAA04436@weedcon1.cropsci.ncsu.edu>,
> NetBSD Bob <nbsdbob@weedcon1.cropsci.ncsu.edu> wrote:
> >Don't laugh, too hard.....(:+}}...
> >
> >For the fun of it I am trying to collect up the earliest ports of
> >BSD that will run on a MicroVAX II.  My goal is to roll up a set
> >of disks with each version back through 4.3BSD-Tahoe, for some
> >educational, experimental purposes.  Last night I got NetBSD-1.0A
> 
> Here's a subject that's very near and dear to my heart.  I would love to
> be able to run vanilla 4.3BSD on my VaxStation 3100, even if for nothing
> more than pure nostalgia and hack value.  I have my Ancient Unix Source
> License and Kirk McKusick's CD archive to start with.  But 4.3BSD only
> supports booting from UNIBUS and MASSBUS devices, which of course the
> VS3100 doesn't have.  So, I'm somewhat stuck.
> 
> For a while, I was hoping that the NetBSD/vax bootblock and bootloader
> were similar enough that I could at least get 4.3 to boot /vmunix to
> the point where it went looking for a root device.  But apparently,
> they aren't... I can get the NetBSD bootloader to load /vmunix, but it
> hangs hard immediately.  This was about a year ago that I tried this,
> with some variety of 1.4.
> 
> Of course, now that NetBSD has switched to elf binaries, the divergence
> with old BSD has been widened even more.  But it occurs to me that maybe
> someone else on this list has tried something similar, and may have some
> advice (or at least, maybe some interesting battle-scars).
> 
> cheers,
> --Mirian
>