IIUC, tick_nohz_full_kick_cpu() is intended to wakeup idle CPUs that will not be poked by scheduler because they are actually nohz_full.
But in fact this function kicks all CPUs listed in tick_nohz_full_mask, namely: - idle CPUs; - CPUs runnung normal tasks; - CPUs running isolated tasks [1]; For normal tasks it introduces unneeded latency, and for isolated tasks it's fatal because isolation gets broken and task receives SIGKILL. The patch below makes tick_nohz_full_kick_cpu() kicking only idle CPUs. Non-idle nohz_full CPUs will observe changed system settings just like non-idle normal (i.e. not nohz_full) CPUs, at next reschedule. [1] https://lkml.org/lkml/2017/11/3/589 Signed-off-by: Yury Norov <yno...@caviumnetworks.com> --- kernel/time/tick-sched.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index c026145eba2f..1c24c700e75a 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -247,7 +247,7 @@ static void tick_nohz_full_kick(void) */ void tick_nohz_full_kick_cpu(int cpu) { - if (!tick_nohz_full_cpu(cpu)) + if (!(tick_nohz_full_cpu(cpu) && idle_cpu(cpu))) return; irq_work_queue_on(&per_cpu(nohz_full_kick_work, cpu), cpu); -- 2.17.1
