Monday, May 9, 2016

SimH IRIS "v2" with fewer crashes

This page is a part of the "Understanding IRIS" collection.  Many thanks to David Takle, for figuring this out, and sharing this with us:

NOTE:  See also this page, which contains the foundational resources and our "first-attempt" parameters and sources that we used before making the below discovery.  

------------------------------------------------------------------

Found the crash problem !! 

The crashes I have been experiencing, so far do NOT seem to be related to Pico-N issues.
What happened is that I inadvertently set the wrong status for the CPU in the startup fie.

What I had in there was:
SET CPU 64K

You would think that might work, and there is no real clue in the docs that it won't, but apparently only in some limited fashion. Words above 077777 octal were not addressable, and simply dropped the top bit in addressing.  So while IRIS thought my partition was at 160000, when it tried to copy my Swap-File header into the partition, it landed at 60000, which wiped out a big chunk of the $DEC driver.
Which meant that any application requiring $DEC, such as LIBR or QUERY to display decimal numbers would crash when it got to that part of the driver.

The correct setup is

SET CPU EXT64KW

It's a lot more fun now!

Here's the updated nova-iris.simh (renamed nova-iris.v2.simh)  You may need to right-click and save as.

------------------------------------------------------------------

So, I now get the following TRAP #122 notification when running the nova-iris.v2.simh file.  Is this expected?

NOVA simulator V3.9-0
sim> do nova-iris.v2.simh

PLEASE WAIT . . .

TRAP #122   STATUS = 177777 37777 600 20234 0

PLEASE WAIT . . .

CONFIGURATION PROBLEM, ONLY MASTER PORT ACTIVE !

ENTER YEAR,MONTH,DAY,HOUR,MINUTE   !
%\

THIS IS "IRIS" R7.5 !

COPYRIGHT 1974,1981 POINT 4 DATA CORPORATION


ACCOUNT ID ?   PORT #0      GROUP 0 USER 1     AUG 19, 1986  11:58:33

TIME LEFT  --  CPU: NO LIMIT       CONNECT: NO LIMIT

1168 BLOCKS IN USE, 3500 AVAILABLE ON UNIT #0

#

-------------------------------------------------------------

OK, it looks like we need both commands in the command file.

I have done a lot of minor cleanup on my environment here, so I am re-sending it. See attached.
This will fix the trap. I also disabled the $MMUX driver, since we do not have a multiplexor.
Fixed a program on LU 0. If you type "IPL" at the '#' prompt, it will reboot IRIS. (the old version of IPL used a bad boot sector).

I also changed the names of our files, so we have something of a naming convention.
The drive is now iris0c.d31. The first part, 'iris0' identifies the content on the drive, the letter 'c' identifies a matched collection of LU's. Thus, iris1c.d31 would be a companion LU1. The part after the dot identifies it as a diablo 31 drive of type DKP. That way if we know how to define it to simH. If we later port this particular LU to another type of drive, the last few letters will change.

Correspondingly, the simh command file is now iris-c-d31.sim. That tells us the file will attach Diablo drives containing the 'c' collection of LU's.

~David

[Please click on the hyperlinks in the filenames above, to download those files and test for yourself!]

------------------------------------------------------------------

Cprossu:

stumbled on another command

#install
? LOGICAL UNIT TABLE IS FULL !
SEE "MILU" IN SECTION 2.4 OF MANAGER MANUAL.
INSTALLATION TERMINATED.
#



This page is a part of the "Understanding IRIS" collection.  

No comments:

Post a Comment