When a CPU wakes up from idle and finds out that the timekeeper is sleeping, we need to kick it such that it switches from dynticks to periodic mode to maintain its timekeeping duty on behalf of the newly awoken CPU.
However we aren't using the right API for that. rcu_kick_nohz_cpu() is aimed at waking full dynticks CPUs and not the timekeeper. Moreover the timekeeper must perform a new dynticks cycle to check the new sysidle state and restart the tick if necessary. A simple call to irq_exit() isn't enough. wake_up_nohz_cpu() is a good fit for that job because it pulls the target out of the idle loop and restart the tick. Then if no other task waits for the CPU, it will reenter the idle loop and then the new sysidle state will be visible and well handled. Reviewed-by: Paul E. McKenney <[email protected]> Cc: Ingo Molnar <[email protected]> Cc: Nicolas Pitre <[email protected]> Cc: Paul E. McKenney <[email protected]> Cc: Peter Zijlstra <[email protected]> Cc: Steven Rostedt <[email protected]> Cc: Thomas Gleixner <[email protected]> Cc: Viresh Kumar <[email protected]> Signed-off-by: Frederic Weisbecker <[email protected]> --- kernel/rcu/tree_plugin.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/rcu/tree_plugin.h b/kernel/rcu/tree_plugin.h index 551398b..477d81f 100644 --- a/kernel/rcu/tree_plugin.h +++ b/kernel/rcu/tree_plugin.h @@ -2494,7 +2494,7 @@ void rcu_sysidle_force_exit(void) oldstate, RCU_SYSIDLE_NOT); if (oldstate == newoldstate && oldstate == RCU_SYSIDLE_FULL_NOTED) { - rcu_kick_nohz_cpu(tick_do_timer_cpu); + wake_up_nohz_cpu(tick_do_timer_cpu); return; /* We cleared it, done! */ } oldstate = newoldstate; -- 1.8.3.1 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
