i read russ' sleep history of the discussion of
sleep/wakeup on 9fans and didn't see any
references to this.

i've been wondering why wakeup requires that
some process be in state Wakeme and that
it be waiting for the particular Rendez passed
to wakeup.

i don't think this is required for correctness 
(the model works with this commented out)
and only disallows waking up a process that's not
sleeping or sleeping on another Rendez.

it would seem to me useful to be able to not
make fancy provisions for avoiding double
wakeups.  is there any reason to avoid this
that i do not see?

- erik

Reply via email to