Nuclear meltdown? 
Author Message
 Nuclear meltdown?

Crossposted to comp.lang.asm370 and comp.lang.cobol, come to c.s.y2k for
the signal, stay for the noise.

On Fri, 29 Jan 1999 12:13:52, The COBOL programmer known as Goobers

Quote:


> > S0C7 mean anything to you?

> More than to most folks... time to recompile NOOBJ, NOLOAD and a PMAP
> (or LIST).

> (at times like this it would have had much better comic effect were I to
> have allowed myself the obfuscational banality of answering a question
> with a question, viz.,

> 'S0C7 mean anything to you?'

> 'You mean it doesn't to someone?')

> DD

But the problem won't be in one of those  magical 14 RR instructions.
You know those RR instructions, from the days of yore when programmers
wrote entire reservations systems in a solution space of 16 32 bit
registers using at most 12 RR instructions; oh, go ahead use 14.  We
can't make bricks without straw, it's Moshe's fault.

Aye, now those were programmers, hairy, smelly yes, but capable of feats
of geekdom such as we can't imagine.  And the geek-grrrls, sultry, raven
haired beauties, dancing the dance of welcome in their father's tent.

It'll be in one of the bloated SS instructions, maybe FC or FA.

Dump Reading 101 - You'll need to know this in 336 Days.

Find the PSW, where does it point?

Find the current registers, oh my, there are saveareas all over the
place.  Well, those look good.

Thought of the day to C mechanics, S/390 machine language is a high
level language, higher than C, and more precise, and expressive.

Take the instruction before the address, S/390 updates the location
address counter before executing the current instruction.  There was
probably a reason for this at one time.

Disassemble the instruction, use the current registers to find the date,

base+offset  S
base+index+offset  X

Use the 4 bit L to find the end of the data.  There it'll be.

MP d1(l1,b1),d1(l2,b2)

where d1 = displacement (or offset) 1
       l1  = length of data 1
       b1 = register

FA32 5000 6001   <---<<< previous (failing) instruction

FA  32    5 000   6 001    <---<<< broken out
MP  l1 l2 b1 d1   b2 d2     <---<<< operands

of the 32, 3 is l1, 2 is l2.

if register 5 contains the value 00000400
   register 6 contains the value 00000310

Look at 310+1 for a length of 3.  Verify that it contains a packed
decimal number.

Y2K denialists, this is what Moshe's S/390 beeper mechanics deal with on
a regular basis.  And from all indications do it for less money than I
happily pay my automobile mechanic, $75/flat hour.

I'll write about rates and fair compensation for geek boyz and grrrls in
a future WRP.

cory hamasaki 336 Days, 8,079 Hours

WRPs are at http://www.*-*-*.com/



Tue, 17 Jul 2001 03:00:00 GMT  
 Nuclear meltdown?

Quote:

>Crossposted to comp.lang.asm370 and comp.lang.cobol, come to c.s.y2k for
>the signal, stay for the noise.

>On Fri, 29 Jan 1999 12:13:52, The COBOL programmer known as Goobers


>> > S0C7 mean anything to you?

>> More than to most folks... time to recompile NOOBJ, NOLOAD and a PMAP
>> (or LIST).

>> (at times like this it would have had much better comic effect were I to
>> have allowed myself the obfuscational banality of answering a question
>> with a question, viz.,

>> 'S0C7 mean anything to you?'

>> 'You mean it doesn't to someone?')

>> DD

>But the problem won't be in one of those  magical 14 RR instructions.
>You know those RR instructions, from the days of yore when programmers
>wrote entire reservations systems in a solution space of 16 32 bit
>registers using at most 12 RR instructions; oh, go ahead use 14.  We
>can't make bricks without straw, it's Moshe's fault.

Now, now, Mr Hamasaki, you pointed out the imprecision of my language and
I accepted the correction.

Quote:

>Aye, now those were programmers, hairy, smelly yes, but capable of feats
>of geekdom such as we can't imagine.  And the geek-grrrls, sultry, raven
>haired beauties, dancing the dance of welcome in their father's tent.

>It'll be in one of the bloated SS instructions, maybe FC or FA.

Really?  I've usually found that a S0C7 occurs in a ZAP (opcode F8)... but
remember, I'm a COBOL jockey; data-errors are often the result of poor
initialisation.

Quote:

>Dump Reading 101 - You'll need to know this in 336 Days.

>Find the PSW, where does it point?

Where it *always* points... to the next instruction to be executed.

Quote:

>Find the current registers, oh my, there are saveareas all over the
>place.  Well, those look good.

>Thought of the day to C mechanics, S/390 machine language is a high
>level language, higher than C, and more precise, and expressive.

>Take the instruction before the address, S/390 updates the location
>address counter before executing the current instruction.  There was
>probably a reason for this at one time.

Same reason as there is for it nowadays... because the S360 did it.

Quote:

>Disassemble the instruction, use the current registers to find the date,

>base+offset  S
>base+index+offset  X

>Use the 4 bit L to find the end of the data.  There it'll be.

>MP d1(l1,b1),d1(l2,b2)

So, not a ZAP.. still another packed-data-addressing instruction; not bad
gfor a COBOL-jockey.

- Show quoted text -

Quote:

>where d1 = displacement (or offset) 1
>       l1  = length of data 1
>       b1 = register

>FA32 5000 6001   <---<<< previous (failing) instruction

>FA  32    5 000   6 001    <---<<< broken out
>MP  l1 l2 b1 d1   b2 d2     <---<<< operands

>of the 32, 3 is l1, 2 is l2.

>if register 5 contains the value 00000400
>   register 6 contains the value 00000310

>Look at 310+1 for a length of 3.  Verify that it contains a packed
>decimal number.

Dollars to Navy beans says it contains LOW-VALUES... oopsie, binary zeroes
for you 2GL types.

DD



Tue, 17 Jul 2001 03:00:00 GMT  
 Nuclear meltdown?

Quote:

> Dollars to Navy beans says it contains LOW-VALUES... oopsie, binary zeroes
> for you 2GL types.

> DD

Speaking as one of those who's STILL writing massive reservation systems w/16
registers. (But we're only allowed to use 11 if them unless you REALLY like to
see the guys in coverage go nuts...)  We proud and wonderful 2GL types actually
DO know low-values.  We even know HIGH values!!!  (-;

I don't think we could get over 4000 messages a second w/1.5 second average
response time and over 99% availability 24/7.  Let Gates and his bloatware try
to approach THOSE numbers!!!

--
Chuck Rush
The Sabre Group
Senior Programmer
Sabre Cars Development
(817) 963-3613

Et casu Latine loqui{*filter*}Societatis Jesu ne umquam conaris...
And may you never accidentally try out your Latin on a Jesuit.



Tue, 17 Jul 2001 03:00:00 GMT  
 Nuclear meltdown?

Quote:

>if register 5 contains the value 00000400
>   register 6 contains the value 00000310

>Look at 310+1 for a length of 3.  Verify that it contains a packed
>decimal number.

Here's a thing you reminded me of.  For anyone testing programs on a
standalone home P390, make sure you test them with high addressing.  I
fell foul of that one when a program loaded with a base address of say
00008000 and length 3000 worked perfectly.  But when someone else was
using the machine and it loaded at 00012000 (or whatever) it failed.
Using add halfword to index address areas isn't always a good idea ...

:Dave  [I didn't explain that very well but then it is Friday after all]
--
Dave Eastabrook. 25+ years IBM Mainframes, and general Y2k Consultant.
IBM Assembler from highly experienced specialists, telecommuting only.
Elmbronze Ltd. 1-111 fully equipped telework associates available NOW.
http://www.elmbronze.demon.co.uk/telework/    also /IBM/ or /year2000/



Tue, 17 Jul 2001 03:00:00 GMT  
 Nuclear meltdown?


Quote:

>> Dollars to Navy beans says it contains LOW-VALUES... oopsie, binary zeroes
>> for you 2GL types.

>> DD

>Speaking as one of those who's STILL writing massive reservation systems w/16
>registers. (But we're only allowed to use 11 if them unless you REALLY like to
>see the guys in coverage go nuts...)  We proud and wonderful 2GL types actually
>DO know low-values.  We even know HIGH values!!!  (-;

Mr Rsuh, it is a pleasure to be made aware of folks out there who posess
broader learning, aye; my experience, though, has shown me that it is
better to explain something to a point of mind-numbing tedium than to run
the rsik of finding out that someone was smiling and nodding,
uncomprehendingly.

DD



Tue, 17 Jul 2001 03:00:00 GMT  
 Nuclear meltdown?
Or your English with DD :=]

Grrreat quote Chuck!

JB



Quote:

snip>
>Et casu Latine loqui{*filter*}Societatis Jesu ne umquam conaris...
>And may you never accidentally try out your Latin on a Jesuit.



Tue, 17 Jul 2001 03:00:00 GMT  
 Nuclear meltdown?

. . .

Quote:

>Really?  I've usually found that a S0C7 occurs in a ZAP (opcode F8)... but
>remember, I'm a COBOL jockey; data-errors are often the result of poor
>initialisation.

Hmm... A S0C7 due to a ZAP being missing? Or is it that common to
initialize fields using variable data of untrustworthy origin?

Quote:
>>Find the PSW, where does it point?

>Where it *always* points... to the next instruction to be executed.

For S0C7s perhaps, but this is not "always" the case for S0C4s.
Andy Wood



Tue, 17 Jul 2001 03:00:00 GMT  
 Nuclear meltdown?

Quote:


>. . .

>>Really?  I've usually found that a S0C7 occurs in a ZAP (opcode F8)... but
>>remember, I'm a COBOL jockey; data-errors are often the result of poor
>>initialisation.

>Hmm... A S0C7 due to a ZAP being missing? Or is it that common to
>initialize fields using variable data of untrustworthy origin?

Frequently enough... data-initialisation is often neglected (in COBOL, the
VALUE clause in WORKING-STORAGE).

As a feller once told me 'Of *course* you blew up on a S0C7 in a
PERFORM... VARYING, your subscript was fulla air!' = numeric field was not
appropriately initialised.

Quote:

>>>Find the PSW, where does it point?

>>Where it *always* points... to the next instruction to be executed.

>For S0C7s perhaps, but this is not "always" the case for S0C4s.

True enough, and I should have been more specific in addressing the matter
at hand... S0C4s can fill just about *any* location with 'air'.

DD



Tue, 17 Jul 2001 03:00:00 GMT  
 Nuclear meltdown?

Quote:


>>if register 5 contains the value 00000400
>>   register 6 contains the value 00000310

>>Look at 310+1 for a length of 3.  Verify that it contains a packed
>>decimal number.

>Here's a thing you reminded me of.  For anyone testing programs on a
>standalone home P390, make sure you test them with high addressing.  I

How do you do that, and why is it something unique to P390s?

Quote:
>fell foul of that one when a program loaded with a base address of say
>00008000 and length 3000 worked perfectly.  But when someone else was
>using the machine and it loaded at 00012000 (or whatever) it failed.
>Using add halfword to index address areas isn't always a good idea ...

Not sure I understand the issue there, but then there are so many
boudaries to test and so little time. Some years ago we made a change
to a system which caused the address of the bottom of the extended
private area to increase. A program which had been in use for many
years started to abend. It turned out that if a particular control
block got allocated at an address greater than X'05F5E0FF' then the
program would fail. Just the sort of thing you might expect. That is
if you knew that the program took the address, converted it to a
decimal character string in an 8 byte field, and then converted it
back into an address.

Andy Wood



Tue, 17 Jul 2001 03:00:00 GMT  
 Nuclear meltdown?


Hi, Cory -- Hi, DD...

Don't see your postings that often anymore.  I got a life and
dumped c.s.y-2.  :-)

Anyway....
:>
:>Find the PSW, where does it point?

: Where it *always* points... to the next instruction to be executed.

Cory, you were the 'most wrong' here.  DD, you were imprecise.
It points to the instruction that has not yet been executed.
As well, --- it MAY point to the bad instruction.
Check your ILC -- ALWAYS check your ILC.
If it's 0 (zero) then the Pgm Chk interrupt PSW points to
the 'bad' instruction.

Jonesy
MainFrame (and dump readin') since 1966
--
--
Marvin L. Jones  jonz<AT>rmi.net | jonz<AT>ieee.org
Gunnison, Colorado
336 days to go until the Year 2000
701 days to go until the 3rd Millennium of the C.E.



Tue, 17 Jul 2001 03:00:00 GMT  
 Nuclear meltdown?

Quote:
> Speaking as one of those who's STILL writing massive reservation systems w/16
> The Sabre Group
> Sabre Cars Development

Could you suggest the titles of any books or articles that describe
the early (1950s/1960s) Sabre development?  (in terms for a lay person.)


Wed, 18 Jul 2001 03:00:00 GMT  
 
 [ 23 post ]  Go to page: [1] [2]

 Relevant Pages 

1. Nuclear meltdown?

2. Nuclear meltdown?

3. Nuclear meltdown?

4. Smalltalk MT goes to Microsoft Meltdown 2001

5. Form meltdown

6. Printing Speed Meltdown

7. -Conventional Memory Meltdown

8. Automated Nuclear Support

9. Meltdown Australia (crosspost from c.s.y2k)

10. Off topic: October 23, 1997 Nuclear Report

11. Image Display program for nuclear medicine images.

12. OT: ANWR, Energy, Nuclear fuel, Greenies, tree huggers, NRA, etc......

 

 
Powered by phpBB® Forum Software