Re: [RFT][PATCH v4 1/7] time: tick-sched: Reorganize idle tick management code

[Frederic Weisbecker]

On Wed, Mar 14, 2018 at 06:20:16PM +0100, Peter Zijlstra wrote:
> On Wed, Mar 14, 2018 at 04:49:39PM +0100, Frederic Weisbecker wrote:
> > On Mon, Mar 12, 2018 at 10:47:41AM +0100, Rafael J. Wysocki wrote:
> 
> > > @@ -136,6 +138,13 @@ static inline u64 get_cpu_idle_time_us(i
> > >  static inline u64 get_cpu_iowait_time_us(int cpu, u64 *unused) { return -1; }
> > >  #endif /* !CONFIG_NO_HZ_COMMON */
> > >  
> > > +static inline void tick_nohz_idle_stop_tick_protected(void)
> > > +{
> > > +	local_irq_disable();
> > > +	tick_nohz_idle_stop_tick();
> > > +	local_irq_enable();
> > > +}
> > 
> > It seems that even if we have CONFIG_NO_HZ_COMMON=n,
> > tick_nohz_idle_stop_tick_protected() will have overhead, right?
> 
> IIRC the only callsite of the _protected thing that remains at the end
> is in the hotplug path. So who cares ;-)

Hey, I see you trying to sneak in overhead out of hatred against the hotplug
code ;-)