Re: Documentation for reading JES2

On 3/12/2010 3:32 PM, scott wrote:
> I have looked at the Q source code. Was wondering if there was some
> document that talks about it to give some insight as to why and when
> somethings were done.

I haven't looked at Q, or JES2/HASP internals for a few decades, 
but the basic spool organization  is based on disk block 
addressing by MTTR. M ranges from 0 to (max spool packs + 
checkpoint number), TT is the absolute track number on the 
volume (not as usual the track number in data set extents), and 
R the conventional record number. The checkpoint data set 
contain an image of the job queue (chain of $JQE elements, one 
per job, arranged by JES2 queue type), and the $JOT with $JOEs 
(job output elements, one per combination of print/punch 
characteristics [e.g., STD. form, PN, 6 LPI, Class). An 
individual job has a $JCT (or two during execution - one input, 
one output), then $IOTs containing $PDDBs (print/punch/input 
data sets; one or more per JES controlled DD). I/O is done with 
EXCP (Q may use BSAM, I don't remember). There are a few other 
blocks, but those are the basic ones. Additional ones reside in 
user address spaces and are not accessible to you (e.g., $SDB). 
In MVT, the fundamental control block is the $HCT, readable by 
all running tasks; it has the pointers to the various queues, 
counters, and other goodies. In MVS, the $HCT is in the JES2 
address space, and not readable unless you schedule an SRB to 
extract information; links to the $HCT, etc. are in a new $SVT 
block.

As far as I know, there is no available documentation on JES2 
internals; I'm not even aware of anything for HASP, but I never 
looked for any. Your best bet is to assemble the modules in 
SYS1.HASPSRC (doubles as a macro library), start with HASPINIT, 
then HASPRDR, then HASPPRPU. Those should show you the basics of 
how job information is created and used.

Gerhard Postpischil
Bradford, VT