Peter Relson    wrote

>>The sample showed use of ALESERV EXTRACTH and ALESERV ADD.

>>FYI, ALESERV EXTRACTH "works" as you want, but the far more direct mechanism 
>>of
>>PSAAOLD -> ASCBASSB -> ASSBSTKN is also fully supported for getting home's 
>>STOKEN.

Yes, we're a bit simple minded and tend to use IBM macros where they exist, 
although it's not a hard and fast rule. I'm sure one could argue endlessly 
about the pros and cons of using IBM CB fields  rather than the macros provided 
by IBM for lots of things, but at the end of the day maybe there can be too 
much concern for saving a few picoseconds where it is completely unimportant, 
at the cost perhaps of other aspects of (good) programming. This code is 
executed once when an address space starts. I doubt we'd make noticeable 
savings by directly accessing some CB field. This also avoids any need to check 
whether this or that CB field is an official interface, not to mention that it 
might help the less experienced when they can just look up the macro 
documentation to see what is going on. Still, another possibility to tell 
customers we've reduced instruction paths.

>>ALESERV ADD for an address space is allowed only for a non-swappable space.

Indeed we do a TRANSWAP in CAS when starting the DAS, then OKSWAP after the DAS 
xm-posts its completion of initialization.  After that things run quite happily 
swappable (in the case of the product this code came from anyway). But in fact 
it's the IARVSERV SHARE that otherwise blows up (with a 6C5), not ALESERV ADD. 

I do wonder if the sample was actually useful to someone.

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

Reply via email to