>Can PER trap on a mask setting? No. I suppose you could trap on (or trace) "all branches" or "all instructions" and then have the SLIP trap examine the time-of-PER-interrupt PSW to look at the program mask, but that would degrade things so much that you might never get there.
There is an assumption in the OP that the program itself "did it". Maybe it did. But maybe it did not do it directly. The expectation is that the (E)SPIE service manages the PSW program mask. But it is true that any program can change the program mask to whatever it chooses. If LE has asked for (E)SPIE management of program interrupt 10, then you'd expect the program mask bit for decimal overflow to be on. The (E)SPIE protocol is to save the returned token from "set" and use that on "reset" to reset things to the previous environment. I use the term "environment" loosely. If an application is "playing" with the program mask while (E)SPIE is managing it, all bets are off. I do not know what drives LE to specify what options when it uses (E)SPIE. FWIW, because of the program mask, almost all system code must use logical arithmetic rather than decimal arithmetic when managing a field that could "wrap" (AKA overflow), in case it happens to be running with a non-0 program mask. Peter Relson z/OS Core Technology Design ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
