Alto Ether

Juergen Buchmueller pullmoll at t-online.de
Sun Jul 15 14:20:09 PDT 2007


On Sun, 15 Jul 2007 11:33:23 -0700
Al Kossow <aek at spies.com> wrote:

> All sorts of interesting things happen once the net gets running.

How are the chances to get the Ethernet card PROM a49 dumped? It seems my
guessed PROM just isn't good enough. For example reading of incoming packets
is oftentimes terminated before the packet actually ends.

A first try to do the whole Ethernet timing based on the (invented) PROM was
even worse than what I have now. You get used to how the designers did
their job, but without the real data you're ultimately lost.

Another question I have is if you perhaps know where to find a description
of the F9401 CRC chip? It doesn't seem to be fully compatible with the
74401, just similiar. That's not too important until I have to create and
parse packets including the exact CRC polynomial.

> Something you will also want to simulate is the Trident disk interface.

I saw triex.run on diag.dsk.Z, but I haven't seen the Trident hardware
mentioned, except as a range of memory addresses in some "regusage" file.

> Once you have that working, you can try bringing up the IFS (interim file server)
> which lets you do things like network booting. We would REALLY like to have a file
> server of some form on the 3mb net at the Museum we're going to put together for
> the Alto restoration.

L'll try to do a IPv4 packet translation back and forth inside Salto, and
if you can then route packets between your LAN and the 3MB net at the
museum, that should really be all you need. SDL is fairly portable. Salto
compiles and runs on CYGWIN with just one change: some missing error number
in errno.h. An IFS server could probably even run without any display at
all.

Right now IFS.RUN does nothing but show a hourglass cursor.

Over the next couple of weeks I'll have devote some time to a paid job. I
am, however, optimistic to be able to get the Ethernet simulation working
well enough. It was so far easy, compared to the obstacles in the DIABLO
drive emulation.

> There are some unix-side services written at Stanford for net booting. That's where
> the boot images came from. That stuff ran off of Nowiki's packet filter code.

It may help to be able to peek at this. It'll be the first time I do this
either way, using existing packet filter code, or creating my own raw IP
packets. I haven't written any packet filter accessing code yet, while I
don't fear it. It's just that it will take some time.

BTW: Sorry Eric for abusing the list. I am, of course, willing to help, if
you want to use any of the info and findings for Altogether.


More information about the Altogether-devel mailing list