> Since I had a weekend off here in Trondheim, I suddenly decided to try
> how far I could get the NetBSD/pmax port going on a DECstation-3100 we
> have here (normally running DEC/OSF 1.0). I also got some help from
> Tor Egge, so I felt we should try.
> 
> So I followed Adam Glass's instructions, and lo and behold, the # prompt!
> But, as he also said, most things didn't work. By inserting lots of
> debugging printf's in addition to the syscall tracing, we found at least
> two major problems: The obreak() syscall didn't seem to work, and also
> programs wanted to call ultrix syscall #256 (getsysinfo) to start. The
> first problem stems from the exec_ecoff module, where the original data
> segment size (or rather, original break value) didn't take into account
> the bss segment. The second we fixed by implementing a stub getsysinfo/
> setsysinfo so programs think they work (actually the getsysinfo
> implementation here could be argued to be strictly in conformance with
> documentation, implementing just one op and returning unavailable on
> all others).
> 
> I'm including the first (important) diff here, another in my next
> message.
> 
>    - gratefully yours,
>      Arne H. Juul
> 

Cool!!! Someone sending fixes instead of complaints.  I'll look at
this stuff immediately.

later,
Adam Glass

------------------------------------------------------------------------------