linux-mips-fnet
[Top] [All Lists]

Re: interrupt code

To: linux-mips
Subject: Re: interrupt code
From: Michael Engel <engel@unix-ag.uni-siegen.de>
Date: Tue, 25 Feb 1997 08:52:15 +0100 (MET)
In-reply-to: <199702250723.AA05442@internet-mail2.ford.com> from "hkoerfg1@ford.com" at Feb 25, 97 08:23:31 am
Hi,

> >   We would like to know the current status of the IRQ code for the
> >DEC 5000/125. How close is it to being completed? 

I have a timer interrupt working on my DS5000/200 and will (hopefully) soon
have it running on the 2100/3100. IIRC, the dispatching of the several
interrupt sources (they all jump to the same address) still makes some 
problems.

> >Also, if anyone out
> >there has worked on the console code for the said machine, I'd much
> >appreciate a response as that's what we'd like to work on. 

Basically, the REX routines should work for you - at least they work for 
my DS5000/120 here. This may, however, not work related to PROM revision 
level differences.

I have a console driver that works on the 2100/3100 and _should_ work on 
a standard CX graphics board (haven't tried yet). For the DS5000/1xx, one
should only have to adapt the serial driver code for the Z8530 that the 
5000/1xx uses (instead of the DZ11 chip in the other DECstations). I have 
the Zilog data book for the Z8530 here, but haven't yet searched for the
I/O addresses of the 8530 [and I'm afraid the keyboard port on my 5000/120
is not working :-( - well, I got it for free :-)]

> >So the
> >sharing of knowledge can be a good thing. Also, has anyone looked
> >around in the NetBSD code for information about the hardware, I've
> >looked around in there. You'd be surprised about the knowledge in
> >there. 

Yes ... and also by the chaos and badly documented code in NetBSD :-( ...
Three other sources of information that may be relevant are
- Sprite by John Ousterhout (Berkeley) from ftp.cs.berkeley.edu
- Mach3
- Xinu/DECstation (runs only on 3100 (maybe 2100, too) with serial console)

> >Also, there are references to DEC manuals that were used to
> >create the code. 

Yeah ... John "Maddog" Hall is still working on getting the documentation -
any news on that, Paul (I'm afraid not :-() ? DEC sometimes seems to have
lots of trouble finding their own documentation.

> That is exactly the point, where i'd like to start. As far as i can see 
> the NetBSD Sources are the only source of information available yet, 
> at least for kmins and maxines. I know this can be more difficult 
> than starting from scratch, but why don't we use it, is there 
> something i am missing? 

Nope, try it. We may not reuse code from NetBSD because of some silly 
copyright restrictions (but we could always use code from OpenBSD or 
4.4BSD Lite, I think). If you find out information on the 5000/1xx and Maxine 
(a list of I/O adresses etc.), that'd be great !

regards,
        Michael Engel   (engel@unix-ag.uni-siegen.de)

<Prev in Thread] Current Thread [Next in Thread>