Re: [PATCH] idle using PNI monitor/mwait (take 2)

[Jamie Lokier <jamie@shareable.org>]

Jamie Lokier wrote:
> local_irq_disable() isn't required in the monitor/mwait loop, because
> you check need_resched between the monitor and mwait.  (If Intel had
> implemented monitor+mwait as a single instruction, then you'd need it).
> 
> So you can remove it from your loop.
  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Excuse me, I hadn't looked at your code closely.  Everything I said up
to that last line is fine.  But the last line doesn't apply because
you don't have a local_irq_disable().

You can remove your local_irq_enable() instead, if you want, but for
a different reason.  That one _is_ defensive. :)

-- Jamie