On Wed, 24 Jan 2007 14:56:52 EST, IBM Mainframe Discussion List <[EMAIL PROTECTED]> wrote:
. . . > The poster >meant that the only architected way for a non-0 key user to switch states on >purpose was via SVC. John McKown mentions accidentally causing program check interrupts - they could also be caused deliberately, or as an integral part of the design. I seem to recall that the original "cross memory" instructions were simulated in that way on machines that did not have the hardware. > Of course, when the SVC routine (and other interrupt >classes as well) has finished the function invoked, the interrupted user is >resumed with the same PSW settings as when it was interrupted; e.g., storage >protect key, privileged state, system mask, etc. So you cannot get control back >in a more authorized state than you were before you did the SVC. > You have just proved that the MODESET SVC cannot possibly work. ---------------------------------------------------------------------- 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
