Subject: Re: New port
To: NetBSD Ports Discussion List <netbsd-ports@NetBSD.ORG>
From: Greg A. Woods <email@example.com>
Date: 09/15/2000 14:46:15
[ On Wednesday, September 13, 2000 at 08:46:33 (-0400), Jason R Fink wrote: ]
> Subject: Re: New port
> [ On Wednesday, September 13, 2000 at 05:37:04 (-0700), Jason R Thorpe wrote: ]
> > On Wed, Sep 13, 2000 at 11:19:59AM +0200, Frank van der Linden wrote:
> > >
> > > On Wed, Sep 13, 2000 at 10:33:39AM +0200, KukatMichael wrote:
> > > > communications controller) with SCSI, 2 serial ports, and LANCE
> > > > ethernet, this controller has its own 68000 and needs an own kernel for
> > > > it. (This could be stolen from the original MUNIX distribution first
> > > > :-).
> > >
> > > A Lance onto which you need to download 68k code to make it go?
> > > That sounds weird..
> > Not too strange. Integrated Solutions, Inc. had Ethernet cards
> > for their Qbus and VME systems that had LANCE-behind-68000, too.
> also i have heard (or read - i dunno that was a while ago) the name
> LANCE in reference to VME chassis SBC's when I did military work for the
> US. I think they were only using VxWorks _on_those_cards_ (they used
> hpux for the SBCs that users interacted with). I will check with some
> contacts i have and see if they are still using them.
Every controller on the AT&T 3B2 series machines has an onboard 80186 or
better that does the real work of talking to the hardware..... On the
serial cards (especially EPORTS) that's how they could guarantee full
speed on any of at least 'N' serial ports and very high aggregate
throughputs for the entire board, all while doing full termio processing
and without bogging down a rather slow CPU (fast at the time!). You
could easily get 16 users running WordPerfect/Unix on a 3B2/400 with
only 4MB of RAM and they'd get decent (at the time) response time from a
such a full-screen editor.
Unfortunately publicly available documentation for these boards, and
indeed even documentation about how to talk to these onboard controllers
from the kernel, is nearly non-existant so far as I can tell. Even
source licensees didn't get enough source to rebuild any of the firmware
that was sometimes downloaded to these boards at boot time....
Greg A. Woods
+1 416 218-0098 VE3TCP <firstname.lastname@example.org> <robohack!woods>
Planix, Inc. <email@example.com>; Secrets of the Weird <firstname.lastname@example.org>