[cap-talk] kernel object knowledge

[Bill Frantz]

alan.karp at hp.com (Karp, Alan H) on Tuesday, May 29, 2007 wrote:

>Note that Norm was not talking
>about an object capability system, in which there is only the object
>reference, but one in which there are extra control bits in the
>capability.

Note that object capability systems can have a very similar problem.  If
we want to define a general capability storage object, and have a facet
on it which only returns "read only" authority to the capabilities
stored, then if the system is like E, and uses part of the message as a
method selector, the storage object needs to:

  (1) Understand the protocol for parsing the method selector, and

  (2) Understand enough about the method semantics for each object to
      know which methods are indeed "read only", so it can build a facet
      front-end object for the underlying object.

Cheers - Bill

-----------------------------------------------------------------------
Bill Frantz        | gets() remains as a monument | Periwinkle
(408)356-8506      | to C's continuing support of | 16345 Englewood Ave
www.pwpconsult.com | buffer overruns.             | Los Gatos, CA 95032