Looking for Stanford Ethernet AppleTalk Gateway (SEAGATE) source code

[Michael Engel]

Hi,

On 14 Dec 2014, at 16:03, Al Kossow <aek at bitsavers.org> wrote:

> On 12/13/14 9:48 PM, Michael Engel wrote:
>> 
>> according to
>> http://www.megalextoria.com/usenet-archive/news021f2/b37/fa/info-mac/00001680.html
>> Bill Croft at Stanford developed an Ethernet to Apple LocalTalk gateway
>> that ran on the original SUN 68000 CPU board.
>> 
>> Unfortunately, I cannot find the files mentioned in that post on any info-mac
>> mirror (or somewhere else on the net):
>> 
> 
> I've emailed you a copy, and I'll put the tarball up on bitsavers under bits/Stanford

Thanks, Al - got it, looks great! I also found two later versions of the seagate source,
modified to run on Kinetics FastPath routers (68008/68000-based single-board routers 
with a 82586 Ethernet controller and a Z8530) and renamed "KIP":
https://stuff.mit.edu/afs/athena/astaff/project/snmp/cmu/kip/
ftp://ftp.st.ryukoku.ac.jp:/pub/network/cap/AT-KIP/

> The code was built using the SUMEX C toolchain, derived from the MIT 68k port of PCC, which
> is even harder to find.

This is actually easy to find, SUMACC tools (actually a Mac cross compiler, which works
fine for creating b.out binaries) are included in 4.3BSD's "new.tar.gz" sources:
http://www.tuhs.org/Archive/4BSD/Distributions/4.3BSD/new.tar.gz

I have sumacc working on 4.3BSD/VAX in simh (my real VAXen are still in Germany and
are too recent to run 4.3BSD or even Quasijarus) anyways. NetBSD+emulation might work). 

The original SUN (and, presumably, also Codata) firmware was developed using sumacc, 
we are currently trying to get our own modifications to run using the sunmon sources from 
bitsavers. The sunmon sources compile after adding files from the PDF source listing of an 
older version (0.8) of sunmon (from 
http://bitsavers.trailing-edge.com/pdf/stanford/sun/SUN_Monitor_Listing_Aug81.pdf) 
and writing some missing header files (mostly nec7210.h), but we are still struggling with 
strange bus errors that keep creeping up when trying to run the code on the real machine.
It works fine in the (very incomplete) mess Codata emulator as well as a hacked-up version 
of a Verilog implementation of a small Codata subset + an attached VPI 68000 simulator 
(derived from Ben Herrenschmidt's minimigmac sources at 
http://git.ozlabs.org/?p=minimigmac.git).

I also got the sumacc compilers to compile on OS X, but these show bugs in the parser 
and code generator, resulting in crashes. I'll try to figure these out over the Christmas
break...

If anyone is interested in my simh VM image for VAX 11/780 with 4.3BSD, the sumacc compilers 
and the sunmon sources installed, please let me know.

-- Michael