> > 4.3 is lots lighter.  I run the Quasijarus port, which, after some bit
> > twiddling, runs and installs pretty well.  
> 
> I've heard good things about Quasijarus, and they have an interesting sense 
> of humor. 

The port is relatively vanilla, but requires some significant headscratching
to get it up the first time, and the fellow that runs the port is fairly
strong-willed about it.  That aside, I had to wind up using NetBSD-1.2
as the boot/label system, and then run the install of the Quasijarus port.
I could not get it up otherwise.  Once up, it is light and dainty.

> This brings up something that has been bugging me for a while.  What exactly 
> is in the recent kernels that VAX people are trying to keep up with?  Asked 
> another way, what would the VAX world loose between NetBSD 1.5 & BSD 4.3 
> kernels?  How much user-land would one loose access to by having a leaner 
> kernel?  Other than MI SCSI drivers, there isn't a lot of commonality with 
> other archs, is there? 
> 
> With the feature bloat in the NetBSD kernel mostly relevant to other 
> architectures, it seems to me that there may be value in having a "VAX/BSD" 
> based on something more suited to the hardware.  One could start with 
> something like Quasijarus, and work from there. 
> 
> Just a thought.  I know my 3100m40 is too slow to be much more than a 
> curiosity these days. 

Well, these are good issues, and from the purely technical merit, something
that needs to be looked at, for the older slower hardware, IMHO.  I put
up all the different NetBSD's on my MVII critter (1 drive for each build)
and saw the bloat/problems arise on the slow machine.  NetBSD-1.2 was
the lightest, 1.4.3 was the fastest, and 1.5 was a big dog.  If you graphed
the bloat it was most interesting to see the feeping creaturitis creep
in from 4.3 up through NetBSD-1.5.  This was especially true in the
stripped kernel sizes (MVII minimal config only).   That is not putting
the system down, but is just a fact that has happened.  It is not really
apparent on my alphas, mipses, sparcs, etc., but on the 1 vup and under
VAXentoyz, it is a real problem.  It is also, IMHO, a problem on 3 vup
and under toys (my 3100/M20, for example).  On a 4000 class machine or
a 3100/M76, and up, it really is no problem....yet.

As to answers, I don't really want to stirrrrr up the hornets nest, today.

But, thinking wistfully, musing sympathetically, and that sort of thing,
I do think someting needs to be tweaked, that is slow-VAX specific.
How we get there, I will leave with the VAXengurus aboard.

> > What specific 4.3 port were you using back then?  
> 
> Err...good question.  I don't know.  I think there was an awful lot of local 
> work done.  We had some hot-shot Unix kernel guys running around, mostly 
> working on the Clouds distributed operating system.  Next time I talk to 
> Henry Strickland, I'll ask him. 

Do, anything that surfaces is good for the historical preservation cause,
and I am big on that.

> > IFF any of that stuff is still extant,  pass it my way and I will 
> > bounce it along to the archives.
> 
> Unlikely, but I'll ping a few folks. 

Good.   ping....  ping....  ping....

> > As a demo of how light 4.3 is, the compiled kernel is around 150k
> > and it compiles in well under 30 minutes.... on an MVII 9 mb box,
> > with esdi drives, no less.....(:+}}...
> 
> Along the same lines, I recently picked up a Heurikon VME532 system (if 
> anyone knows anything about this beast, do let me know off list).  30MHz 
> NS32532, 4MB VRAM, 170MB ESDI disk.  It runs GNX 3.1, which was National 
> Semis Unix flavor, and the basic distribution + development system fits in 
> about 115MB.  The thing boots faster than any of my recent vintage BSD 
> machines, and is astonishingly responsive.  Remarkable. 

I picked up a couple of NCR towers one time, with hopes of BSDing them.
Alas, all they turned out good for were some drive spares.  Maybe you
will have better luck with your machine.  It sounds like you have all
the bits with it... mine were just rough boxes.

Bob