Messages in this thread | | | Date | Thu, 12 Feb 2015 14:14:20 +0100 | From | Peter Zijlstra <> | Subject | Re: [PATCH 1/6] PM / sleep: Re-implement suspend-to-idle handling |
| |
On Wed, Feb 11, 2015 at 05:01:09AM +0100, Rafael J. Wysocki wrote: > +/* Suspend-to-idle state machnine. */ > +enum freeze_state { > + FREEZE_STATE_NONE, /* Not suspended/suspending. */ > + FREEZE_STATE_ENTER, /* Enter suspend-to-idle. */ > + FREEZE_STATE_WAKE, /* Wake up from suspend-to-idle. */ > +}; > + > +static enum freeze_state __read_mostly suspend_freeze_state; > +static DEFINE_SPINLOCK(suspend_freeze_lock); > + > +bool idle_should_freeze(void) > +{ > + return unlikely(suspend_freeze_state == FREEZE_STATE_ENTER); > +}
I don't see how a compiler can propagate the unlikely through an actual function call. AFAICT that needs to be an inline function for that to work.
It would mean exposing suspend_freeze_state and the enum; is there a reason not to want to do that?
| |