Maybe he understands that there are no risks in his use of S-cons. -- Shmuel (Seymour J.) Metz http://mason.gmu.edu/~smetz3 עַם יִשְׂרָאֵל חַי נֵ֣צַח יִשְׂרָאֵ֔ל לֹ֥א יְשַׁקֵּ֖ר
________________________________________ From: IBM Mainframe Discussion List <[email protected]> on behalf of Jon Perryman <[email protected]> Sent: Thursday, June 12, 2025 2:55 PM To: [email protected] <[email protected]> Subject: Re: Is there an equivalent of an S-con for a long displacement? External Message: Use Caution On Wed, 11 Jun 2025 10:14:39 +0300, Binyamin Dissen <[email protected]> wrote: >On Tue, 10 Jun 2025 15:27:34 -0500 Jon Perryman <[email protected]> wrote: >:>I've never seen a valid use case for S-CONS. > >A read only parameter list where the target routine uses the callers registers >to calculate the address of the parameters. I'd love to hear if Peter Relson would allow this use of S-CONS despite the risks. If I were reviewing this code, you would be rewriting it. There are much safer alternatives with far less complexity. An S-CON is validated where it's coded instead of where it is used. The greater the distance, the greater the risk. Something as simple as PUSH, DROP & POP fails to catch a coding error. >I chose to use LAY to get the BLLLHH and overlay the E300 and 71 with other >metadata (length,type,etc.) If you could somehow justify the S-CON risks, I would point out the inefficiency of your implementation using SY() which would be at least 10 instructions instead of a single instruction that utilizes hardware to process the long S-CON. E.g. if you left the LAY intact, then you could have used EX to execute the LAY. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
