[coyotos-dev] Status and roadmap
Sam Mason
sam at samason.me.uk
Mon Jan 22 10:50:10 CST 2007
On Mon, Jan 22, 2007 at 10:20:06AM -0500, Jonathan S. Shapiro wrote:
> On Mon, 2007-01-22 at 00:12 +0000, Sam Mason wrote:
> > My memory wasn't quite what it should be! I was remembering section
> > 3.3.1 "Event Arrival and Run-In". It only explicitly specifies what
> > happens when the activated field is non-zero. I thought this was
> > missing but re-reading more slowly I now realise that this hasn't been
> > missed and is most definitly the intended behaviour!
> >
> > It may help people, especially on a first reading, to put something in
> > here to say that doing nothing in this case is correct.
>
> Sam: there was an update between your two readings. Given the new text,
> do you still think that a note is appropriate, and if so, where?
You currently have the following in the section titled "Event Arrival
and Run-In".
1. The Activated field is tested.
o If the Activated field is zero, the process is currently in the
normal state. An arbitrary enqueued event is chosen for delivery
(as described below), and its payload is transferred to the
dispatch registers.
The Activated field is set to 1.
2. If further events are queued for delivery, the PendingEvents
field is set to 1. Otherwise, it is set to zero.
3. The process is dispatched to a CPU and execution begins.
I would be tempted to put something after "If the Activated field is
zero" paragraphs to the effect of "If the Activated field is non-zero
the dispatch registers remain untouched". I'm not sure if I'm being
overly lazy in thinking this is needed, but it seems like the sort of
place you'd have a comment in the code to say nothing needs to be done
in this case.
This is a bit of a break from the distinct normative and descriptive
sections you have at the moment, so I'd understand if it stays the way
it is at the moment.
Sam
More information about the coyotos-dev
mailing list