Re: Key invocation detail: Bill Frantz (frantz@netcom.com)
Thu, 19 Dec 1996 21:22:48 -0800

At 11:10 AM -0800 12/19/96, Jonathan S. Shapiro wrote:
>This is a concept question concerning key invocation.
>
>If I 'fork' a kernel key, passing a start key to myself in the last
>slot, the kernel's return clearly won't happen (because I am in the
>wrong state). In this event the kernel's reply is dropped. The
>question is: is it dropped before or after the effect of invoking the
>original kernel key occurs?
>
>[I think the right answer should be before]

The right answer is after. The (virtual) returner does not check the key for resumedness until it is invoked at the end of the kernel service. Note the exact parallel with an extra-kernel service (implemented in a domain) which uses the returner. 

Bill Frantz       | I still read when I should | Periwinkle -- Consulting
(408)356-8506     | be doing something else.   | 16345 Englewood Ave.
frantz@netcom.com | It's a vice. - R. Heinlein | Los Gatos, CA 95032, USA