Thanks Chris. Yep, APF auth, SP255 and under rock are already familiar to me. It's the rock I'm trying to avoid.
Why do I ask? Because it's there. :-D Jerry On Wed, 23 Jan 2008 15:00:45 -0500, Craddock, Chris <[EMAIL PROTECTED]> wrote: >> Does that mean that a home level token pair is automatically deleted: >> >> - when the current job step terminates? (so I understand it to mean) > >Yep that's what it means. > >> <...> can anyone think of a reasonable method for >> finding a specific storage area within multiple steps of the same job? >By >> "reasonable >> method" I would exclude using a file to store the storage area's >address, >> as well as obtaining such storage in global common system storage. > >It -can- be done but your definition of "reasonable" may become a little >strained. The first issue is that the "storage area" you wanted to >retain addressability to would itself have to be owned by the initiator >TCB, i.e. the parent of your job step task or the address space itself >(effectively the same thing.) > >You can't allocate storage from that task or from any "life of address >space" subpool (like say 255) without at least being APF authorized. So >either your first step runs authorized, or some external server provides >a PC function (open to integrity issues) or you co-opt one of the step >initiation SMF exits (e.g IEFUSI) to do your dirty work. At least in the >latter case you would already be running under the right TCB to get what >you wanted. > >Beyond that, you would need to create a global name/token pair to refer >to the life-of-address-space storage because nothing else is going to >survive termination of the job step that created that name/token pair. >Alternatively you could play silly games with hiding the pointer under a >rock somewhere within the address space that WILL survive step >termination (details left as an exercise to the reader) > >Oh and then you have the problem of knowing when to clean up and release >that storage, presumably at end of job, but without using an IEFACTRT or >similar exit, I'm not sure how you would know the job had ended... All >fun and games huh? > >So why do you ask grasshopper? :-) > >CC ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
