On Fri, 26 Feb 2016 23:03:59 -0600, Tom Marchant wrote:

>On Fri, 26 Feb 2016 22:48:43 -0600, Mike Schwab wrote:
>
>>On Fri, Feb 26, 2016 at 10:37 PM, Tom Marchant
>><[email protected]> wrote:
>><deleted>
>>> In addition, when an instruction references a storage
>>> location whose upper bits match the prefix register, those bits are 
>>> replaced with 0.
>>> This last function of prefixing was not implemented on System/360 or the 
>>> earliest
>>> system/370 models.
>>>
>><deleted>
>>> --
>>> Tom Marchant
>>About 1985-1989, I remember our mainframe taking a lot of crashes from
>>jumps to low address space on a 2 processor pre-XA machine.  Was some
>>of this caused by this function?
>
>I doubt it. Do you have a dump?
>
>IIRC, the address in the prefix register is the address of a page in SQA, and 
>that any 
>reference to that SQA page would result in accessing absolute page zero.

BTW, without this reverse prefixing, that SQA page would contain the zero page 
for a 
processor, giving you twice the opportunity to overlay real page zero. And now 
that I 
think of it, I'm not at all sure that MVS maps any virtual address to that real 
address.

-- 
Tom Marchant

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to