开云体育

-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Jim
Snellen
Sent: Monday, October 24, 2022 7:01 PM
To: [email protected]
Subject: [H390-DOSVS] DOS/VS Handshaking as VM/CE Guest
Importance: High

If anyone chooses to re-assemble the DOS/VS supervisor, I will first
recommend creating a new one, defining a new name (ie. $$A$SUPx). After
doing so, the POWER statement to spool output to VM (PSTART LST,00E,A,,VM)
will not work. A PDZAP is required. Following is a dialog between me and
Professor Rene Ferland which corrects this situation.
--------------------------------------------------------------------

Greetings Professor Ferland,

I apologize for contacting you directly and, if you prefer, I'll open a
thread on /g/H390-DOSVS </g/H390-DOSVS>
.

On 8/14/21, in response to Steve Orso, you wrote:

On Fri, Aug 13, 2021 at 03:20 PM, <stephen.orso@...> wrote:


[...] and in particular what you need to do to run it under VM/370.

Have you run your DOS/360 under VM/370? It should not be very hard. But
the one problem I see is handshaking (automatic close of spool print/punch
files), probably not implemented on DOS/360. The OS has to be patched for
that. George Shedlock did it for DOS/VS.

Cheers,

Rene FERLAND, Montreal


I'm particularly interested in the bold statements in your response.

I generated a supervisor to change the storage allocations for each
partition. After re-IPLing with the new supervisor, I've lost the ability
to spool print to VM/CE using HercPRT. Is there a subsequent task to be
done?
--

Best Regards,

Jim Snellen

FIRST RESPONSE
Hello Jim,

Unfortunately, I am not sure I can help you. George Shedlock is the one
who implemented handshaking. He knew the operating system, assembler and
what to do. Me I can only guess from the jobs he provided.

If you logon to the DOSVS virtual machine, minidisk 191 contains all the
material of George Shedlock. There are two update jobs: SGSVC UPDATE A,
and SGATAB UPDATE A. From what I understand, the first is to update (or
add?) supervisor call SVC56 (presumably for handshaking). It will punch a
job in the reader queue that has then to be run to perform the update. The
same is true for SGATAB UPDATE A, which updates the tables and constants.
If you just changed the supervisor, I would assume you don't have to run
these jobs again. The updates they made are likely still there.

What you probably need to do is change/run the job $$A$SUP1 ZAPS A. This
job modifies the $$A$SUP1 phase in the core library for handshaking. It
makes a very small change to the supervisor binary at a specific address.
Presumably, when you re-assembled the supervisor, your new binary does not
have the modification and that's why handshaking does not work anymore. I
don't know if you can run George Shedlock's job as such for your new
supervisor, or if you need to change the address because of the changes
you made. One possible way to find out is to re-assemble George Shedlock
original supervisor ($$A$SUP1 DOSVS A) and examine the listing to
determine where the ZAP is done and find the equivalent in your new
supervisor.

That's the best I can do for you.

Good luck!

Rene FERLAND, Montreal

SECOND RESPONSE

Hello Jim,

I have assembled George Shedlock's original $$A$SUP1 supervisor and I can
see what the ZAP has done to it.

It (apparently) changes the last bit of SYSFLAG4 (check the PNG attached):
I think the ZAP says to check (VER) for 6001,6808 at address 738 and, if
it is there, replace it with 6001,6809 (REP).

So, check your own supervisor for SYSFLAG4, find the address and value,
and adjust the ZAP accordingly? :-)

Cheers,

Rene FERLAND, Montreal

REPLY
Hello Professor,

As always, you have come through again! My changes to the new supervisor
did not change the offset so I was able to run the PDZAP using the new
name. The VM handshake works wonderfully.

With your permission, I will post this dialog on /g/H390-
DOSVS </g/H390-DOSVS> so that others may benefit from
your wisdom.

Thank you.



Attachments area

On Oct 24, 2022, at 10:52 PM, Fish Fish <david.b.trout@...> wrote:

?Wouldn't it be better to just directly fix the supervisor generation macro instead? (to support a new e.g. VM=YES|NO option?)

Then no zap would be involved at all. You could generate a new supervisor at any time and would be guaranteed of always having VM handshaking support as long as VM=YES was specified. <shrug>

Or is doing that somehow undesirable?

<me: slightly confused>

--
"Fish" (David B. Trout)
Software Development Laboratories

mail: fish@...

-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Jim
Snellen
Sent: Monday, October 24, 2022 7:01 PM
To: [email protected]
Subject: [H390-DOSVS] DOS/VS Handshaking as VM/CE Guest
Importance: High

If anyone chooses to re-assemble the DOS/VS supervisor, I will first
recommend creating a new one, defining a new name (ie. $$A$SUPx). After
doing so, the POWER statement to spool output to VM (PSTART LST,00E,A,,VM)
will not work. A PDZAP is required. Following is a dialog between me and
Professor Rene Ferland which corrects this situation.
--------------------------------------------------------------------

Greetings Professor Ferland,

I apologize for contacting you directly and, if you prefer, I'll open a
thread on /g/H390-DOSVS </g/H390-DOSVS>
.

On 8/14/21, in response to Steve Orso, you wrote:

On Fri, Aug 13, 2021 at 03:20 PM, <stephen.orso@...> wrote:


[...] and in particular what you need to do to run it under VM/370.

Have you run your DOS/360 under VM/370? It should not be very hard. But
the one problem I see is handshaking (automatic close of spool print/punch
files), probably not implemented on DOS/360. The OS has to be patched for
that. George Shedlock did it for DOS/VS.

Cheers,

Rene FERLAND, Montreal


I'm particularly interested in the bold statements in your response.

I generated a supervisor to change the storage allocations for each
partition. After re-IPLing with the new supervisor, I've lost the ability
to spool print to VM/CE using HercPRT. Is there a subsequent task to be
done?
--

Best Regards,

Jim Snellen

FIRST RESPONSE
Hello Jim,

Unfortunately, I am not sure I can help you. George Shedlock is the one
who implemented handshaking. He knew the operating system, assembler and
what to do. Me I can only guess from the jobs he provided.

If you logon to the DOSVS virtual machine, minidisk 191 contains all the
material of George Shedlock. There are two update jobs: SGSVC UPDATE A,
and SGATAB UPDATE A. From what I understand, the first is to update (or
add?) supervisor call SVC56 (presumably for handshaking). It will punch a
job in the reader queue that has then to be run to perform the update. The
same is true for SGATAB UPDATE A, which updates the tables and constants.
If you just changed the supervisor, I would assume you don't have to run
these jobs again. The updates they made are likely still there.

What you probably need to do is change/run the job $$A$SUP1 ZAPS A. This
job modifies the $$A$SUP1 phase in the core library for handshaking. It
makes a very small change to the supervisor binary at a specific address.
Presumably, when you re-assembled the supervisor, your new binary does not
have the modification and that's why handshaking does not work anymore. I
don't know if you can run George Shedlock's job as such for your new
supervisor, or if you need to change the address because of the changes
you made. One possible way to find out is to re-assemble George Shedlock
original supervisor ($$A$SUP1 DOSVS A) and examine the listing to
determine where the ZAP is done and find the equivalent in your new
supervisor.

That's the best I can do for you.

Good luck!

Rene FERLAND, Montreal

SECOND RESPONSE

Hello Jim,

I have assembled George Shedlock's original $$A$SUP1 supervisor and I can
see what the ZAP has done to it.

It (apparently) changes the last bit of SYSFLAG4 (check the PNG attached):
I think the ZAP says to check (VER) for 6001,6808 at address 738 and, if
it is there, replace it with 6001,6809 (REP).

So, check your own supervisor for SYSFLAG4, find the address and value,
and adjust the ZAP accordingly? :-)

Cheers,

Rene FERLAND, Montreal

REPLY
Hello Professor,

As always, you have come through again! My changes to the new supervisor
did not change the offset so I was able to run the PDZAP using the new
name. The VM handshake works wonderfully.

With your permission, I will post this dialog on /g/H390-
DOSVS </g/H390-DOSVS> so that others may benefit from
your wisdom.

Thank you.



Attachments area

On Oct 25, 2022, at 11:33 AM, Fish Fish <david.b.trout@...> wrote:

?Jim Snellen wrote:

Fish wrote:

Wouldn't it be better to just directly fix the supervisor
generation macro instead? (to support a new e.g. VM=YES|NO
option?)

Then no zap would be involved at all. You could generate a
new supervisor at any time and would be guaranteed of always
having VM handshaking support as long as VM=YES was specified.
<shrug>

Or is doing that somehow undesirable?

<me: slightly confused>

Fish,
That certainly would be a great way of addressing this issue.
Given this was your idea, I’ll let you choose the supervisor
macro to modify and write the code to do it. I certainly don’t
have expertise to do it.

Best regards,
Jim

Well, I'm kind of busy with something else right now, but what the hell. It should be fairly easy.

To help speed things up, would you (or someone) please either upload, post the contents of, or send directly to me (preferred) the contents of the two mentioned files please? ("SGSVC UPDATE A", and "SGATAB UPDATE A").

Thanks.

--
"Fish" (David B. Trout)
Software Development Laboratories

mail: fish@...