On Thu, 1 Sep 2016 10:47:57 -0700, Ed Jaffe <[email protected]> wrote:

>On 9/1/2016 10:18 AM, Tom Conley wrote:
>> On 9/1/2016 12:48 PM, Ed Jaffe wrote:
>>>
>>> "If you change an exit, except IRREVX01 and IRRVAF01, you must re-IPL
>>> MVS™ for the changes to take effect. IRREVX01 and IRRVAF01 are defined
>>> to the dynamic exits facility, and you can update them without
>>> re-IPLing."
>>>
>>
>> Baby steps.
>
>Haha! I ain't never messed with EVX01 or VAF01 ... only PWX01. Good
>thing I gotta sandbox and lotsa time... :-\
>
>Dynamic exit capability doesn't help much (or at all) if it's not
>implemented for the exits you actually need.
>
>Design Thinking might suggest a customer survey to find out which exits
>folks are _actually_ using and make those dynamic.

It is relatively simple to write a "stub" exit for _most_ of the non-dynamic 
RACF exits, and have that stub use MVS Dynamic Exit Services to invoke the real 
exit. That takes only as many IPLs as it takes for you to get the stub correct, 
and presumably those are IPLs of a test LPAR. IBM even provided a Red Book, 
possibly at the RACF 2.1 level, giving a working stub example with code.

That is not true for _all_ the exits though, at least unless some changes have 
been made to the MVS Dynamic Exit Service capabilities since the last time I 
looked. The problems I specifically found were with the RACROUTE 
REQUEST=FASTAUTH exits, I believe, and specifically the older ICHRFX01 and 02 
exits. It's possible that the newer ICHRFX03 and 04 would be OK, but of course 
the exit you need to use depends on the application that is making the call, 
and exactly what flavor of FASTAUTH the application issues.

The problem I remember was that the more basic FASTAUTH calls 
(non-cross-memory, profiles RACLISTed in local storage by the application) run 
in the state, key, addressing mode, and Task/SRB mode of the caller. And the 
exits are called directly by the FASTAUTH processing code. At the time I made 
my study of implementing dynamic exits via a stub routine, I could not find a 
way to use MVS Dynamic Exit services that would allow all the necessary 
environmental combinations.

As I say, the services may have improved since then. But also it is more likely 
now that ICHRFX03 and 04 are used rather than 01 and 02, and their environments 
are different and perhaps more suitable for dynamic exit usage :)

-- 
Walt

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

Reply via email to