On 2017/7/12 0:09, Frederic Weisbecker wrote: > On Tue, Jul 11, 2017 at 11:41:57AM +0200, Peter Zijlstra wrote: >> >>> - totally from arch_cpu_idle_enter entry to arch_cpu_idle_exit return costs >>> 9122ns - 15318ns. >>> --In this period, rcu_idle_enter costs 1985ns - 2262ns, rcu_idle_exit >>> costs >>> 1813ns - 3507ns >> >> Is that the POPF being painful? or something else? > > Probably that and the atomic_add_return(). > We thought RCU idle cost is high, but it seems not. But it still has few > microseconds can be saved if we can remove them from fast idle path.
Thanks, -Aubrey
