MVS can not run on any machine that had an ASCII bit in the PSW. There was no 
statement of integrity for OS/360.

Does anybody know whether there was a statement of integrity for RSS/360?


--
Shmuel (Seymour J.) Metz
http://mason.gmu.edu/~smetz3


________________________________________
From: IBM Mainframe Assembler List <[email protected]> on behalf 
of Charles Mills <[email protected]>
Sent: Saturday, August 8, 2020 7:53 PM
To: [email protected]
Subject: Re: Case Study: IBM SYSTEM/360-370 ARCHITECTURE (1987)

I am not suggesting anything nefarious or in violation of the Statement. A 
program that is linkedited (sorry, I just can't get used to saying "bound") 
with AC=1 and stored into an APF-authorized library -- all documented and 
above-board and under the customer's control -- can issue MODESET MODE=SUP and 
get into supervisor state.

In addition, various "user" exits run in supervisor state. As do appendage 
routines IIRC. Etc.

APF-authorized products may have PC routines that run in supervisor state.

The point of my post was not "oh, z/OS is full of holes." It's not AFAIK, and I 
did not mean anything of the sort. My point was that there might have been code 
somewhere that set the ASCII bit on, even if the IBM OSes did not offer a 
"MODESET SIGN=ASCII" macro instruction. There are several ways that customer or 
vendor code may have done so in addition to "magic SVCs."

And yes, before someone else points it out, my memory of the history of the 
platform is not perfect. Exactly which of these methods were available back 
when the ASCII bit was still around is not something that is at the top of my 
stack. Certainly magic SVCs were less out of fashion back in those days.

I would have to think this through, but right off it would seem to me that a 
semi-magical SVC that did only one thing -- toggling the ASCII bit for its 
caller -- would not violate the integrity of the box as a whole. There could be 
some factor I have not thought of; this is an off-hand speculation of moot 
significance, not a statement to give to your auditor.

Charles


-----Original Message-----
From: IBM Mainframe Assembler List [mailto:[email protected]] On 
Behalf Of Paul Gilmartin
Sent: Saturday, August 8, 2020 2:53 PM
To: [email protected]
Subject: Re: Case Study: IBM SYSTEM/360-370 ARCHITECTURE (1987)

On 2020-08-08, at 15:19:05, Charles Mills wrote:
>
> Or a dozen or more other non-magic ways of getting into supervisor state.
>
How does this interact with IBM's Statement of Integrity:
    https://www.vm.ibm.com/devpages/SPERA/istatmnt.html
    ...
IBM will accept all APARs that describe exposures to the System Integrity of MVS
    ...
System Integrity is defined for MVS as the inability of any program not 
authorized by a mechanism under the customer's control to:
        • circumvent or disable store or fetch protection
        • access an OS password-protected or a RACF-protected resource (RACF is 
the Resource Access Control Facility), or
        • obtain control in an authorized state; that is, in supervisor state, 
with a protection key less than eight (8), or Authorized Program Facility (APF) 
authorized.

Of course, "under the customer's control" is a key phrase.

Reply via email to