Detangle lwkt_switch() and the idle thread and consolidate sleep variables

[Simon 'corecode' Schubert]

Joerg Sonnenberger wrote:
The intel docs also indicate that there can be a MWAIT 
functionality to work with interrupts disabled and as interrupt as a break 
condition.  I'm not exactly sure how this would work, maybe the interrupt 
would be serviced as soon as you enable interrupts again?
MWAIT itself would abort and the assumption is that you reenable
interrupts directly after that. So that would catch the interrupt.
Ah, I wasn't sure how interrupts are handled;  I thought it might be a 
edge triggered interrupt which just gets ignored then, but if interrupts 
are queued, then this is no issue.

cheers
  simon