Subject: Re: Why do I keep hearing 4.3 BSD these days? (was: Re: The unbearable
To: Gunther Schadow <gunther@aurora.regenstrief.org>
From: NetBSD Bob <nbsdbob@weedcon1.cropsci.ncsu.edu>
List: port-vax
Date: 05/28/2001 11:08:12
> Aha, I gather that SCO now owns Unix.

Actually, this week, the Caldera Linux House now owns UNIX.
My, has the world come full circle.....(:+}}...  SCO sold it
to them several months back.

> It went through ridiculously
> many hands after AT&T sold it. Even Novell owned it for a while,
> sure they had no clue what to do with it, hihi :-) So, now Unix
> is back with someone who know what it is :-) and SCO has done the
> right thing. Is this a correct summary of the history?

Well, it is more convoluted than that, in that it now belongs to
Caldera, who is a big Linux house, and who are ``unifying'' UNIX
and Linux for the biz environment.....whatever that really means.
IFF I can take a broad perspective, IMHO, that will be good for
Linux, and good for OpenSource and thus good for the BSD community,
in the long run.  How, in fact the dust will settle, was not exactly
clear after a read of Caldera's biz plan for it.

> Now the key question is: is 4.3 BSD (tahoe, as you say) complete?
> I mean, will it run on my VAX and your VAX just so, like Ultrix
> supposedly does? But you said tahoe "and sometimes reno", does
> that mean after signing the SCO license agreement you can choose
> from all kinds of releases? I would love to start off with 4.3
> BSD instead of Ultrix, yet I'm afraid they don't support the 
> latest in VAX 6000 400 stuff, or did they? XMI, BI? Probably not.

Well, the Quasijarus version is relatively complete, but it requires
some way to disklabel non-unix disks (I use NetBSD 1.2 for that).
But, it is all there.

You click on the http://www.sco.com/offers/antique.html and away you go!
 
> So, we agree that we love BSD. Small is beautiful, I agree. However, 
> why not discussing a few different viewpoints too :-) 

Well, I guess the blame for stirring up the hornets nets, rests on my
shoulders.  I purposely did that, BECAUSE, IMHO, I think NetBSD VAX
has gotten a bit hefty in the girth and slowed down on lesser VAXentoyz.
Bloat has surely set in, and, from the constructive criticism point
of view, it hurts on a slow 1-vup and under VAX.  Ragge posted the
11/750 sizes a few days ago, and I went back and checked my stripped
MVII builds and found:

Using what I call a stripped kernel config, MVII only (stripped to match
a 4.3BSD MVII config):

ls -l of netbsd == NetBSD-1.5                  570484 bytes
                   NetBSD-1.4.3                471793
                   NetBSD-1.3                  407533
                   NetBSD-1.2                  384105
                   4.3BSD-Quasijarus (Tahoe)   272384

I was thinking kernel size had gone 4x but it has only gone up 2x.

size netbsd.....

                             text       data      bss       dec
                             -------    ------    ------    ------
NetBSD-1.5                   458440     22856     124568    605864
NetBSD-1.4.3                 377292     19568     106396    503256
NetBSD-1.3                   319456     17408      40924    377788
NetBSD-1.2                   312320     18432      53268    384020
4.3BSD-Quasijarus (Tahoe)    182524     59952      54080    296556

For an identically configured kernel (e.g., to be stripped down to
as equivalent of a 4.3BSD MVII kernel), something has grown.
That is not really featuritis, but just plain bloat.

Speedwise, the 1.4.3 is the fastest on the net (although I have not
compared the current or the 1.5.1B2 build, yet), by 20% or so, on
my machines, talking to a relatively fast alpha box.  4.3 was the
slowest, but only by about 40% overall.  Kernel compiles took less
than 1 hour (like around 30 min) on the 4.3BSD box, from SOURCE.
We all know how long it takes an MVII to compile kernel (I get around
7 hours as my best time).  On an already slow VAX, this kind of thing
hurts!  Interestingly, the 4.3BSD box was on ESDI drives, and the
NetBSD MVII was on scsi.  Is there some scsi issue here that might
speed things up?

I dunno what the solution is, but, IMHO, something needs some twiddling
on slow VAXentoyz, else, NetBSD-1.4.3 or such is as good as it gets,
and 1.5 and later is going downhill.  I am not saying 4.3BSD is the
answer, but offer it as a reference point.  I can FEEL the difference
on the MVII class boxes.

Thanks, and with all due respect, but just constructively mumbling
under me breath, a bit.....(:+}}...

Bob