7 Mar
2022
7 Mar
'22
11:06 p.m.
Since we're moving on the front panel operation from console terminal...
On the PDP-11, on a running system, the front panel is very limited. You
can see some state and (if you're really fast) can see data that is
passing over the bus. You can also set the switch register. But both
EXAMINE and DEPOSIT is non-functional. The only thing you can really do
that could be considered harmful would be pulling HALT. Once the machine
is halted, you can then fool around in memory. But on a running system,
you can not. Also, if the key is in the lock position, then HALT does
not work.
Johnny
On 2022-03-07 23:03, Thomas DeBellis wrote:
I think these observations are basically correct; you
can only go so far
with software based security until you really have to start thinking
about hardening the physical environment.
Columbia was perhaps kind of an odd case in that a rather large number
of people could come waltzing through the machine room and some didn't
really understand, "Don't Touch". We systems programmers REALLY didn't
feel comfortable with this.
But the security question really was, once you are in normal operation,
just how 'normal' should the console be? Tops-10's console was pretty
normal except for the INITIA automatic nice things. Tops-20's is closer
as you normally can't get to the auto-magic. My recollection is that on
at least at least some IBM mainframes, you couldn't use the console for
a foreground session. I think this was the case for OS/360 (which we
ran on our 91) and maybe for MVS. I can't remember about VM.
What part of the KA10 console are you talking about, actually? I
thought a few switches did work even with run on. My 1971 version of
SYSREF (from the Orange Book, pages 138 to 143) states that some
switches, such as READ IN, START, EXAMINE NEXT and DEPOSIT NEXT don't
work if run is on. However, EXAMINE THIS and DEPOSIT do work with run
on, which is what I remember.
I can't remember where I put the later version of SYSREF that has KI
operations. What do you remember?
The switch register (obviously) works and WPI used it for a fun
application. We have less than 37 operations personnel assigned to our
KA, and each operator had their own bit. When they came in to work,
they pressed the switch for their 'bit'. You could run a program which
would read the switch register and tell you how many operators were on
duty and who they were.
--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: bqt(a)softjar.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol
------------------------------------------------------------------------
On 3/7/22 4:26 PM, Peter Lothberg wrote:
I think we are "over board", if you are standing next to the machine all bets
are off.
Om the KI10, the front panel works while the CPU is running, not like KA10 where you have
to halt the CPU.
On the KI you can examine/deposit anywhere, use physical addresses or go through the
"MMU". You can put an instruction on the switches and press XCT and the CPU will
execute that as next instruction in current context.
Yes, there is a "front panel lock" toggle switch, but no key-lock.....
(Or walk to the back of the CPU and turn power off...)
(or shut down the air conditioner)
Even Inter CPU's have a "front end" that can be accessed if you have access
to the motherboard....
-P
> ------------------------------------------------------------------------
> On 3/72022 3:56 PM, Thomas DeBellis wrote
> Under Tops-10, the situation might be considered to be worse (but I believe that term
to be unfair). That is, the operating system has no control over what might be examined
or deposited from the front panel.
> There, the CTY is even more special: it doesn't need a password for/any/ login.
You can do an _I [1,2]_ and you are now an operator (I.E.,
> root). However, for many years, I thought this was just neat...
>
> What about other DEC multi-user operating systems? Do you have 'magic'
consoles?
_______________________________________________
HECnet mailing list -- hecnet(a)lists.dfupdate.se
To unsubscribe send an email to hecnet-leave(a)lists.dfupdate.se