On Sun, Feb 9, 2014 at 8:57 PM, Kyle Owen <kylevowen at gmail.com> wrote:
Greetings,
I spent much of the weekend improving my disk server and OS/8 handlers.
Very cool.
https://github.com/drovak/os8diskserver
Here's some more notes:
- The server has been compiled successfully under Cygwin, Debian on
Raspberry Pi, and OS X Snow Leopard.
Awesome.
- The OS/8 system handlers use some PDP-8/E and later
instructions. If
someone wants to help go back through and track these down, we can see if a
more compliant version will still fit in the allotted space.
The one I know about is BSW. It doesn't swap the upper and lower 6
bits in the AC on pre-OMNIBUS machines. Some of the more
"interesting" single-instruction-AC-load microprogrammed IOTs depend
on it.
I was looking around for a list of which instructions are specific to
which models of CPU since part of that is how the CPU detection in
FOCAL works. I know it's documented somewhere, but a search of a few
minutes did not turn up specific information, just general discussion
that there are differences.
Since OS/8 requires 8K minimum, it's probably reasonable to consider
supporting the PDP-8/L and PDP-8/i but not to get too worked up about
supporting older models. If it can be done, sure, there's no reason
not to, but if there just aren't enough locations to bum for all
models, but there is for the PDP-8/i and newer, I'd say that's
probably an adequate target. There just aren't that many systems from
the mid-1960s that have enough RAM to run OS/8. Personally, one of my
long-term goals is to either upgrade my 4K PDP-8/i to 8K (or more!) or
to find a mass storage device for my 12K PDP-8/L so I can run OS/8.
It's not impossible to add an extra serial port to the -8/L, so the
os8diskserver might be an easy way to do it. I have a several DF32s,
but they are all Negibus (not DF32D) and I don't have a
Negibus-Posibus converter (DW08?) Some day, I *will* have OS/8
running on a 1968-era machine!
Also, if an OS/8 guru can help me out, I'd
appreciate it. Will a read/write
from SYS: on an RK05 past block 3248 (decimal) succeed? That is, can a read
or write carry over from one side of the platter to the other? After
looking at the RK8E system handler, I would say this is a yes.
I _think_ the RK05 does a seamless seek across sides. I'm far more
experienced with the RL8A and RL01, but I think I remember a note
saying the RL8A was *not* like the RK8E/RK05 because it _did_ require
an explicit seek across the boundary.
-ethan