I am just saying that TSO has done a ton of work to move between states and maintain integrity. It does not seem to apply or be helpful in this discussion... since it breaks and attempts to maintain integrity at the same time. There are tons of examples of attempts to break and maintain integrity and many failures to do so properly.. I will mention Ray Overby's work in systematically attempting to break interfaces as an example of the many failures and the care that must be exercised when exposing MVS's authorized underbelly to the proverbial magic SVCs etc.
Rob Schramm On Tue, Jul 24, 2018 at 7:36 AM Peter Relson <[email protected]> wrote: > I'd say that the questions are not precisely-enough posed. > > <snip> > 1)Can a Problem Program (Key 8) attach a Subtask that is authorized ? > 2)Can a Problem Program attach a subtask (with the DCB parameter) that is > authorized ? The dcb is not in the steplib concatenation. > 3)Can a Problem Program invoke a Non Space Switching PC routine to Attach > a Subtask that is Authorized ? > </snip> > > What is "authorized" here? APF-authorized? Supervisor state? System key? > > A problem key supervisor state user could certainly attach a subtask that > is supervisor state. > > An authorized (state, key or APF) issuer of ATTACH can ATTACH a subtask > with pretty much any attribute it wants regarding state or key. Changing > the APF authorization is not part of the programming interface. > > So for #1: sure. But that probably is not what is being asked. A problem > state problem key non-APF-authorized task cannot attach a subtask that is > other than that same level of authorization. > > For #2, similarly above. "not in the steplib concatenation" is not > relevant. Steplib has no special connotation. A concatenation is a > concatenation. LNKLST has a special connotation depending on the LNKAUTH > system parameter. Nor does the APF state change regardless of whether the > module is AC=1, regardless of whether the module is located from an > APF-authorized concatenation. > > For #3, sure, if that PC routine results in the ATTACH being issued in > supervisor state or system key. > > 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 > -- Rob Schramm ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
