Re: NetBSD/vax vm for github testing

[Mouse <mouse%Rodents-Montreal.ORG@localhost>]

>>> On a single-cpu machine, that is very much a non-problem.
>> I'd think a single-CPU machine would make it even more of a problem.
>> On a multiprocessor, [...]
> You misunderstood. I meant a single-CPU emulated CPU.

...oh!  Okay, yes, I see.

> I agree with your thought that a multi-core host system have a better
> chance at making it work, but my experience have been that it's not
> even close to good enough for at least MP PDP-11 systems.

I ran into something vaguely similar.  My MicroVAX-II emulator failed
to boot NetBSD.  Turned out that the NetBSD code (at least the version
I was using) depended on the MSCP interface taking longer to complete
its init than it takes the CPU to execute a few insturctions.  But my
emulated MSCP interface was, from the CPU's point of view, infinitely
fast.

Since I don't think it's any part of the MSCP interface contract that
there is a minimum time for _any_ operation, I treated this as a bug in
the NetBSD code and fixed it there.  So far I haven't run into any
other such.  But it does feel somewhat akin to what you describe.

It would be relatively easy to add auxiliary CPU support to my
MicroVAX-II emulator, but, without any code to run on it, there's
little point (and no way to test it).  Unfortunately I no longer have a
copy of the robotics software I mentioned, upthread, that I'd worked
on...if I ever did; I don't recall, now, whether I saved a copy myself.

/~\ The ASCII				  Mouse
\ / Ribbon Campaign
 X  Against HTML		mouse%rodents-montreal.org@localhost
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B