Capability Based Addressing

[John Carr <74711.77@xxxxxxxxxxxxxx>]

RE: Intel delays Merced

Thought you's guy's would like to hear what Dr. Frank said. (See below) ....

>>Hello Neil,
>>Not easily.  An AS/400 pointer is really a capability and part of the 128 
>>bits in the pointer is used to hold authority to objects.  Extending the 
>>address space to 96 bits can be done with minimal pain (as in about the 
>>same pain as CISC to RISC); anything beyond that will be just as hard for
>>the AS/400 as it is for lesser systems.
>>Simon Coulter.


>What do you think Frank?
>John Carr

John,
The answer is not totally correct. The S/38 was a capability-based machine
with the capabilities in the high-order bits of the pointer as described. 
In order to get C2 certification, we had to take the capabilities out of 
the pointers for all user code. Thus, the AS/400 is NOT a capability-based 
machine.

The operating system code still uses the capabilities in the pointers, so 
if we moved to 128-bit addresses, we would have make some operating system
changes similar to what we did when we removed the capabilities from the 
user code. 

The point is: no user code changes as we increase the address size. 
This is what technology independence is all about.

Frank 

(AKA Dr. Frank S.) - JC

+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---