RL02 version of UNIX6?

Noel Chiappa jnc at mercury.lcs.mit.edu
Thu Feb 2 22:06:30 CST 2017

    > From: Allison

    > for laughs I wandered over to:
    > http://ftp.math.utah.edu/pub/mirrors/minnie.tuhs.org/PDP-11/Boot_Images/
    > To see if the copy of V6 on RL02 is still there.... yep it is.  and it 
    > runs on a 11/23 just fine

Yes, that's another copy of the Shoppa disk.

So, I looked at that system, to see how it dealt with the clock issue on an
11/23 (in /sys/ken/main.c, if anyone else is interested). While looking, I
noticed something that made it extremely unlikely that it would boot on an
11/40. Sure enough, attempting to boot the /unix on it on a (simulated) -11/40
blows out.

There are a couple of other unix loads on that pack image (oldunix, unix.tmp,
etc), but all the ones I looked at had the same issue (only tried booting the
'unix' one, though); they're probably all for the same machine, so have the
same configuration issue.

V6 Unix was pretty persnickety about the hardware configuration it ran on;
while it was possible to create builds that would run on almost any
configuration, on 'vanilla' V6 that really only applies to the /40/45/70
era. And even then you still had to re-build the system to match your actual
hardware configuration, almost all the time.

The advent of the /23 (with no CSW, and no KW11-L/P), made things more
complicated. (The clock is pretty key - Unix needs one - several things,
e.g. parts of the teletype drivers, require real-time delays provided by the
clock. I've never tried to run Unix without a working clock, I'm not sure if
it would run without slowly grinding to a halt as stuff waited for clocks
delays that never happened.(

With a little work, a suite of 'universal boot' versions (one for each type of
disk controller - RK05/RL02/RX02 etc) could have been created that would boot
and run on any pretty much CPU/etc configuration - at least, well enough to
build one that did exactly match the hardware configuration at hand. (The one
on the Shoppa disk is close, from what I could see.)

I don't think anyone ever bothered, though (in part because it was much more
of a PITA to test them all, BITD, with only real hardware).


More information about the cctech mailing list