Re: Overlaid VCON in load module

2015-01-23 Thread Peter Relson
Two directed loads into the same area would clearly be an error (unless the user of the first load was fully done with it). There is no reason to fully serialize two directed loads any more than there is reason to fully serialize two loads of non-reentrant modules. Each will get their own

Re: Overlaid VCON in load module

2015-01-22 Thread Peter Relson
I don't see anything in the manuals to indicate that we have to prevent LOAD from running at the same time. You do not need to prevent that. Is there an undocumented requirement to single thread multiple LOADs for the same module within a region? There is no such requirement. Only the first

Re: Overlaid VCON in load module

2015-01-22 Thread Chase, John
-Original Message- From: IBM Mainframe Discussion List On Behalf Of Paul Gilmartin On 2015-01-21 15:12, Greg Dyck wrote: Normal (non-directed) LOAD processing serializes fetch processing for a module and only performs relocation of ACONs/VCONs once when the module is fetched.

Re: Overlaid VCON in load module

2015-01-22 Thread Gary Weinhold
Thanks for all responses. Sam Siegal asked: 1) Does this only happen on second and subsequent load? Since it loads correctly all the time except for rare situations in a multitasking environment, we are assuming it is related to a second load running concurrently with the first load. 2) What

Re: Overlaid VCON in load module

2015-01-21 Thread Abe Kornelis
Gary, In one of my first programs, I inadvertently defined a Vcon double. It gave me exactly the result you describe. It led me to conclude that the loader simply adds the offset to the base value. If you define (or load) twice, the addition is performed twice as well... Looks like you will

Re: Overlaid VCON in load module

2015-01-21 Thread Sam Siegel
Gary - I have an idea of what is occurring, but some additional info is needed. 1) Does this only happen on second and subsequent load? 2) What attributes (rent, reus, etc.) are present on the load module? 3) is this under CICS? Or a proprietary multitasking region? 4) Is code all ASM? Or is

Re: Overlaid VCON in load module

2015-01-21 Thread Greg Dyck
From what we can see the address that ends up in the VCON is the displacement from the beginning of the module (the value presumably placed their by the binder) plus twice the load point. Has anyone seen this happen before? Is there an undocumented requirement to single thread multiple LOADs

Re: Overlaid VCON in load module

2015-01-21 Thread Paul Gilmartin
On 2015-01-21 15:12, Greg Dyck wrote: Normal (non-directed) LOAD processing serializes fetch processing for a module and only performs relocation of ACONs/VCONs once when the module is fetched. The only way I could see what you describe happen is if there were two RLD entries for the

Re: Overlaid VCON in load module

2015-01-21 Thread John McKown
No. CICS has another TCB which uses z/OS directed LOAD to fetch load modules program objects. On Jan 21, 2015 5:00 PM, Paul Gilmartin 000433f07816-dmarc-requ...@listserv.ua.edu wrote: On 2015-01-21 15:12, Greg Dyck wrote: Normal (non-directed) LOAD processing serializes fetch

Overlaid VCON in load module

2015-01-21 Thread Gary Weinhold
We have a situation where it appears that a LOAD of a module may not be resolving a VCON correctly. The first requirement for this module in a region will cause us to issue a LOAD for the module. In a multitasking environment we know that our code could be running on two TCBs simultaneously,