On Wed, 31 May 2006 13:20:03 -0400, J R wrote: >Judicious use of S99TIONQ? > >Maybe use S99CNENQ first?
Maybe. Or maybe it was designed with the realization that you don't actually _have_ to complete the OPEN during SSI OPEN processing, so maybe it schedules an IRB to perform the DYNALLOC with appropriate synchronization with the first GET. Or maybe the DYNALLOC doesn't happen in the user address space at all -- maybe a separate address space (STC?) does the DYNALLOC and the SSI GETs are performed cross-memory? Those are 2 ways I can think of, off hand. But Ed's probably got something else at PSI. -- Tom Schmidt Madison, WI >>From: Victor Gil >>Subject: Re: SSI experience >>Date: Wed, 31 May 2006 11:09:46 -0500 >> >>Ed, >> >>My tests indicate that DYNALLOC is in fact colliding on SYSZTIOT with the >>SSI Open/Close logic. Could you elaborate a bit on how exactly your BPAM >>subsystem avoids the collison? >> >>Thanks! >>-Victor- >> >>On Wed, 3 May 2006 14:59:35 -0700, Edward Jaffe wrote: >> >> >Tom Schmidt wrote: >> >> You aren't planning on (or needing to be) doing dynamic allocation >> >> within the SSI Open/Close routines, are you? >> >> >> > >> >We have a subsystem that emulates BPAM for a proprietary library system >> >that does just that! >> > >> >-- >> >Edward E Jaffe ---------------------------------------------------------------------- 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
