If Lindy actually comes up with a working version of PCP, that would be
something I would like to see, just for old time's sake. It's been a
long time since I last saw it run.
I started IBM programming school the day after OS/360 release 1 (PCP)
went GA, and it was the first system ever programmed for.
My second project out of that programming school was an IBM internal
tool called MEMMAP (Memory map). It consisted of a modified assembler
which replaced all branch instructions with invalid op codes, and would
also generate a bit map that represented the instructions in the
modified module. It used a modified program check interrupt handler that
used the bit map to trace the code paths executed as it intercepted the
"bugged" branches and replaced the op codes with the correct op code
once both paths at each branch instruction had been executed.
The same modified assembler combined the resulting bit map generated
during test case execution with the assembly listing and showed which
instruction paths had been executed. The intention was to determine if
the test buckets had fully exercised the system code and to show if the
APARs and PTFs coincided with the code that had not been exercised by
the test buckets. So I saw OS/360 release 1 run sometime in mid 1968.
Writing the code in the modified PCFLIH that simulated the branch
instructions taught me about an unexpected behavior of the BXH and BXLE
instructions that occurs when a single odd-numbered register is used for
both operands, e.g. BXH R3,R3,LABEL. Loading the odd-register with a
bit pattern and then running code having either BXH or BXLE instructions
imbedded would pre-determine the order of execution, skipping some code
and executing some, depending on whether the sign bit becomes a 1 or a
0. The behavior was documented in a technical report called "OS/360
Coding Notes" by a brilliant IBMer named Bill Collier.
Mike Myers
Mentor Services Corporation
Ivan Warren wrote:
Lindy Mayfield wrote:
Lindy is only trying to understand the code that Fred and his team
spent a portion of a billion dollars writing. I fear that getting it
running on Hercules would be a wonderful challenge, but in the end a
letdown because once it is running I wouldn't be able to do much.
I'm happy traipsing through the code. I thought that starting with
the first OS that ran on S/360 would be a good starting place.
Wait wait wait...
Running OS/360 on herc is trivial (besides the ASCII bit)..
--Ivan
(PS : Gracious ! Would I like the list owner to allow S/MIME signed
messages through... I keep on getting bounced - and having to send the
message without digital signature.. Ah.. no biggy.. I'll manage).
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
