Jim said > > If subpool 0 was not shared, then the EXEC PARM= parameter storage > would be the only thing in its 4K page (or its 2K page in the > old days). The rest of the page would be wasted. Sharing subpool > allows the jobsteb program to obtain the subpool 0 storage > remaining on this page. Perhaps this was a consideration in the > days of expensive, scarce storage. This is only a guess on my part. >
That would have been my guess too (and it's always 24-bit storage!) That's not the only example of combining stuff to save what would otherwise have been a wasted page. The linkage stack management code does (or did) put at least a couple of mini-stacks with 5 or 6 entries into a single LSQA frame to avoid wasting a whole frame on what was (at the time) a lightly used function. One of the BCP guys told me long ago (circa ESA V4) that the initial (mini) linkage stack depth was contrived to be just enough for the deepest calling sequence then known in the BCP itself. Don't know if that's true or an urban legend, but it seemed plausible at the time. 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
