Hello from Gregg C Levine
I am, (Would all of you believe?), running the currently released version of
NetBSD-VAX under the Simh simulator. And this was version of the Simh
simulator was built on a Windows box, the "cpu" in this case is indeed a
KA655, at least that's the name of the binary file that gets loaded before
anything else happens. Now the question, does anyone have a copy of the
KA655 tech manual, as a PDF file? Or know where I can obtain one? Or the
actual manual for purchase. I would like to better understand the processor
behind the simulator's origins. The reason, by the way, that I joined this
group, is indeed what I am doing on that simulator. Incidentally folks, I
applaud the successes that have been reported so far. It means that we, this
community, is doing good.
Gregg C Levine drwho8@worldnet.att.net
"How many floors does this TARDIS of yours have, anyway?"

----- Original Message -----
From: "Antonio Carlini" <Antonio.Carlini@riverstonenet.com>
To: "Michael L. Hitch" <mhitch@lightning.msu.montana.edu>; "Hugh Graham"
<hugh@openbsd.org>
Cc: <port-vax@netbsd.org>
Sent: Saturday, July 20, 2002 3:42 PM
Subject: RE: VAXstation 4000/90 Success!


>
>  I have asked our DEC/Compaq/HP service person if he could find out
> anything through his channels.

All the VS4000 systems (including VLC IIRC) and the NVAX
based MicroVAX 3100 systems have firmware in flash. Updates
were available in various forms: at least a MOP bootable and an
OpenVMS executable were available - there may have been other
forms. The only VAX system for which I recall firmware upgrades
being available publically were (some of) the VAX 7000 systems.
In fact, when last I looked, this upgrade (and those for pretty much
all the Alphas) could be found with a search on the OpenVMS web
page.

I did ask (some time ago) whether the firmware could be released
as part of the Freeware CDs, but it needed approval. It may happen,
or it may not.

>  Unfortunately, the AMD chip doesn't work that way.  You have to erase
> the chip to all ones, and then clear the bits to program it.  Once a bit
> is cleared, it can only be set by erasing the entire memory.

That's certainly true for all the AMD chips for which I have
had to write flash upgrade code (never on a VAX though).

>  I've tried a simple 32 bit sum, and didn't get any useful result.
> Somewhere in all the code in those images should be the test it uses to
> verify the PROM.  I've disassembled it with objdump and scanned it very
> briefly, but didn't spot anything.  The objdump output is 145,000 lines,
> which is a lot to grovel through, and objdump probably messes up some of
> the disassembly.

If you look at the KA655 tech manual, on p4-81 it shows that the
EPROM checksum does not live at the end of the EPROM. This
is admittedly a much older system (EPROM, no flash) but the
same could well be true of the VS4000-90.


Antonio