On Mon, Dec 8, 2008 at 11:13 AM, Michael Schippling <[EMAIL PROTECTED]>wrote:

> I believe that all your assertions are correct, however it is
> possible for the "higher level sync events" to be executed
> directly from the interrupt handler and thus in interrupt context,
> rather than from a new task.


Any code executed from interrupt context needs to be labelled using async
and is the essence of what is meant by async.   So "higher level sync
events"
is confusing in this context.


> But the take-away point is that tasks should not be long-running
> because they may block events signaled from other "low level" tasks.


This is also stressed in Phil's book about programming TinyOS.  Has Allan
taked
a gander at it?

http://www.tinyos.net/tinyos-2.x/doc/pdf/tinyos-programming.pdf

eric


> MS
>
_______________________________________________
Tinyos-help mailing list
[email protected]
https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help

Reply via email to