[OT] Old Timers Please? Vintage 1989 problem

Dagmar d'Surreal dagmar.wants at nospam.com
Sun Apr 11 16:31:54 PDT 2004

On Sun, 2004-04-11 at 13:14, Declan Moriarty wrote:
> Dagmar d'Surreal enlightened us
> > On Sat, 2004-04-10 at 21:56, Dagmar d'Surreal wrote:
> > 
> > > Oh christ this thing *is* oooooold.  What hardware is it on and what
> > > exactly is he trying to do with it?  More details might stimulate my
> > > memory.  (I was mainly "working" remotely on these back then.  Heh)
> At a guess, it's a 15 Mhz 68xxx machine.
> > 
> > Scratch that.  Unless he's trying desperately to get some data off the
> > machine, just shoot it.  My PDA has more power than that machine, and
> > probably better security.
> No. The query stands. It's a tester - it's not about power. That thing
> does some testing of hardware (Network stuff, I believe). Testgear for
> Electronic hardware is horribly expensive - the prices are aimed at
> multinationals, whereas small time one man outfits like myself need
> something like that to stay in business.
> What came before /proc? That's the question in a nutshell. 
> How will he find what io ports and interrupts his kernel is interested in?

He's not going to on that from within the OS...

Let me make something absolutely clear to you about that machine, which
may or may not have been obvious, but try not to take this personally.

That machine is _not_ running Linux.

It's running Unix, and it's running a really _old_ flavor of Unix.  If
you dig around in it's init scripts you will likely see sh coding so
primitive it will remind you of pictures of cavemen hunting deer
scratched onto cave walls.  1985 is a _lot_ longer ago in terms of
computing technology than most people realize.  In 1985 people were
still arguing over whether magic bytes and bangpath (#!) were the
correct way of making shell scripts execute.  There simply is no handy
interface to let you see system internals like if you had /proc around. 
Cross your fingers and start looking for header files relating to the
binaries you execute to use the testing hardware, and if you're really
lucky the binary might respond to -v or --verbose.  You're probably
going to have to flat out reverse-engineer it. 

I'm assuming that if you're looking for the ioports and interrupts that
there's either an 8-bit ISA card in the machine, or a rudimentary devel
environment on it.  If there's anything like a compiler on it, you
/might/ be able to bootstrap gcc into it and get strace running.  If you
can take the ISA card out and put it into a x86 box, it might actually
be faster to poke at the few interrupts and IO ports it could be using
to figure it out.  You can also try poking around in the machine's
BIOS.  Machines from that period of time tended to have some pretty raw
interfaces for calling a bootstrap from the hard disk, and you might be
able to reach some hardware diagnostics from within that.  Expect to
mash alphabet keys on each boot until something happens--or make a
thorough of use of Google and you might find some hardware manuals.  
The email address above is phony because my penis is already large enough, kthx. 
              AIM: evilDagmar  Jabber: evilDagmar at jabber.org

More information about the blfs-support mailing list