4.3BSD Installation (probably off-topic)

Michael Sokolov mxs46 at k2.scl.cwru.edu
Sun Nov 22 05:28:28 AEST 1998


   SHOPPA at trailing-edge.com writes:
> I've been looking over the recent 4.3-ish BSD distributions
> now available from the PUPS archive.  Thought I'd spin off
> a copy for booting on one of my spare uVax II's [...]
   
   The most important thing here is to choose the right version of BSD.
Plain 4.3 CANNOT boot on a MicroVAX II. Later versions, starting with
Tahoe, can. The patches provided in 4.3_on_uVax_instructions are nothing
more than pieces taken out of Tahoe. If you are going to use those, you
might want to use the whole Tahoe system just as well, it has some very
nice improvements, such as disk labels, better man mechanism, and MX record
support in sendmail.
   
   The problem is that we (PUPS/TUHS) haven't been able to find a Tahoe
tape with VAX binaries. I'm not sure if CSRG ever bothered to even make
one, although it's as simple as executing one script on a running system
(which they obviously had). Thus in order to run Tahoe, one would have to
cross-compile it first. It's a pain and takes a lot of expertise, so I
would strongly advise you to avoid effort duplication and wait until I do
it and put the product up in my home directory on minnie.cs.adfa.edu.au.
Actually since KA650 is all I have right now and Tahoe doesn't support it
(but the support code appears later in the SCCS tree), I'll go directly
from Ultrix (my cross-compilation base) to Quasijarus1, my first release,
and won't bother with Tahoe. But for all practical purposes Quasijarus1
will be Tahoe plus KA650 support, shadow passwords, and bugfixes.
   
   Hmm, maybe your have never heard of Quasijarus Project, so I'll explain
briefly what it is. I'm taking over UCB CSRG in terms of shepherding and
maintaining pure Berkeley UNIX(R). I will first re-create it by taking
their final SCCS tree and building my initial one, deciding piece by piece
what belongs to pure Berkeley UNIX(R) and should be kept, and what is POSIX
evil spirit or bloat and should be tossed. In general I draw the line right
around the Tahoe release (summer of 1988), but I'll include anything from
Reno and later code that's worth having, such as KA650 support and Reno's
DBM-based shadow password model. Basically, I want to create a system with
a classical (pre-Reno) look and feel which at the same time has all the
quality improvements and bugfixes ever made by Berkeley, even if they are
as late as 4.4BSD. The last classical release is Tahoe, so that's my base.
I will be using Tahoe to decide what should be included and what should the
look and feel be. Once I know from Tahoe that a given piece should be
included, I'll go to the SCCS file(s) for that piece and decide which post-
Tahoe deltas should be kept (because they are bugfixes or quality
improvements) and which deltas should be tossed (because they introduce the
evil spirit of POSIX or bloat).
   
   How soon will this happen? I'm all ready to go, but unfortunately
hardware problems are holding me back. I have solved the KA650 problem I
was having, but now I'm stuck because neither of the two TQK50 boards I
have works. (The drive SEEMS to work, though.) Thus the sooner I find a
working TQK50 board (or, alternatively, a working TK70/TQK70 pair), the
sooner will I make 4.3BSD-Quasijarus1.
   
> If there's a more appropriate forum for these questions, I'd
> appreciate being redirected to them!
   
   Right now there isn't, because my main VAX farm is currently off the
net. When I get it back on the net (no time estimate, at least several more
months), I'll set up a set of mailing lists for Quasijarus Project and
Berkeley VAX UNIX in general.
   
> OK, Before Step I, as doucmented in 4.3_on_uVax_instructions, is:
   
   Totally disregard these instructions, they are for plain 4.3 ONLY. If
you are using Tahoe or Quasijarus1, the distribution already supports
MicroVAXen as shipped. If you don't want to use Tahoe or Quasijarus1 and
want to use plain 4.3, you are on your own.
   
> Is this an actual limitation on the 43reno.vax distribution currently
> in the archive, or not?
   
   Reno doesn't have any limitations, it already supports KA630 and KA650,
just like Quasijarus1. I personally don't use it, though, because it is not
really True UNIX any more. With the evil spirit of POSIX and a bloat by a
factor of 2 in both binaries and sources, Reno is the beginning of the
destructive process that eventually (and necessarily) culminated with the
disbanding of CSRG.
   
> what non-
> microvax machines will the 43reno.vax distribution boot on?  11/750?
> 11/730?
   
   Of the big VAXen, plain 4.3 supports 11/780, 11/750, 11/730, and Venus
(should have been called 11/790, but was unfortunately named 8600). Tahoe
adds, and Quasijarus1 and Reno retain, the support for 8200.
   
   Sincerely,
   Michael Sokolov
   Phone: 440-449-0299 (Home) 216-217-2579 (Cellular)
   ARPA Internet SMTP mail: mxs46 at k2.scl.cwru.edu

Received: (from major at localhost)
	by minnie.cs.adfa.edu.au (8.9.1/8.9.1) id HAA12654
	for pups-liszt; Sun, 22 Nov 1998 07:42:18 +1100 (EST)
	(envelope-from owner-pups at minnie.cs.adfa.oz.au)
X-Authentication-Warning: minnie.cs.adfa.edu.au: major set sender to owner-pups at minnie.cs.adfa.oz.au using -f


More information about the TUHS mailing list