My tests with GTF in 1987-1988 indicated that whichever application - GTF or 
TMON/MVS - was started first would succeed in using the MC control block 
structure and whichever application was started next would suffer a S0F6 (or 
maybe S0F8) ABEND.  I produced a work-around in about 3 weeks of intense 
effort.  At that time I had access to GTF microfiche.  I was very pleased a few 
years ago when I was able to start my own GTF address space when another one 
was already running and there was no ABEND.  I came to the same conclusion you 
did about the original intent's being to have a multiplicity of applications 
running simultatneously because of a common on one of the fields in one of the 
control blocks which said something like "address of next whatever control 
block in the queue of applications". 


Bill Fairchild 
Franklin, TN 

“Political language is designed to make lies sound truthful and murder 
acceptable, and to give the appearance of solidity to pure wind.” [George 
Orwell] 

----- Original Message -----
From: "Jim Mulder" <[email protected]> 
To: [email protected] 
Sent: Thursday, May 2, 2013 5:17:08 PM 
Subject: Re: GTF IHL007I  Terminating on Error 

> TMON/MVS (possibly under a newer name than when I developed this 
> part of it in 1987) sets itself up as an application using the 
> Monitor Call Trace Facility when and only when a short-duration I/O 
> trace is requested by a TMON/MVS user .  It removes all its Monitor 
> Call control blocks when the trace ends.  Unless that code has been 
> modified since I last worked with it 20 years ago. 

  That's good to know.  The Monitor Call interface in z/OS has 
never been externally documented (to my knowledge), and the internal 
macro for invoking it is PL/S(AS)(X) - only.  Althought it appeared 
that the ancients (before my time) who orginally design this interface 
were intending to allow multiple concurrent appplications, it looks like 
at some they gave up on getting that working, and restricted it to 
one application (GTF) at a time.  But, it was GTF initialization that 
enforced the restriction, not the Monitor Call interface code.  So I don't 
know what would happen if this TMON function was active at the same time 
as GTF. 

  In z/OS 1.6, there was a line item to remove this restriction, 
to allow multiple concurrent GTFs.  Which eventually (after a number 
of PTFs) mostly seems to work.  There were still some remaining 
issues relating to GTRACE TEST, and  GTFTRACE under IPCS, when 
multiple GTFs are active, that I am working on fixing in z/OS 2.1 
  
 Jim Mulder   z/OS System Test   IBM Corp.  Poughkeepsie,  NY 

---------------------------------------------------------------------- 
For IBM-MAIN subscribe / signoff / archive access instructions, 
send email to [email protected] with the message: INFO IBM-MAIN 

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to