Merge branch 'timers/nohz-irq-work-v7' of git://git.kernel.org/pub/scm/linux/kernel/git/frederic/linux-dynticks into timers/nohz

Pull nohz updates from Frederic Weisbecker:

 " This set moves the nohz kick, used to notify a full dynticks CPU when
   events require tick rescheduling, out of the scheduler tick to a
   dedicated IPI.

   This debloats a bit the scheduler IPI from off-topic work that was
   abusing that scheduler fast path for its convenient asynchronous
   properties. Now the nohz kick uses irq-work for its own needs.

   Of course this implied quite some background infrastructure rework,
   including:

     * Clean up some irq-work internals
     * Implement remote irq-work
     * Implement nohz kick on top of remote irq-work
     * Move full dynticks timer enqueue notification to new kick
     * Move multi-task notification to new kick
     * Remove unecessary barriers on multi-task notification
 "

Signed-off-by: Ingo Molnar <mingo@kernel.org>

1 files changed, 13 insertions, 9 deletions

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 3bdf01b494fe..7f3063c153d8 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -684,10 +684,16 @@ static void wake_up_idle_cpu(int cpu)
 
 static bool wake_up_full_nohz_cpu(int cpu)
 {
+        /*
+         * We just need the target to call irq_exit() and re-evaluate
+         * the next tick. The nohz full kick at least implies that.
+         * If needed we can still optimize that later with an
+         * empty IRQ.
+         */
         if (tick_nohz_full_cpu(cpu)) {
                 if (cpu != smp_processor_id() ||
                     tick_nohz_tick_stopped())
-                        smp_send_reschedule(cpu);
+                        tick_nohz_full_kick_cpu(cpu);
                 return true;
         }
 
@@ -734,10 +740,11 @@ bool sched_can_stop_tick(void)
 
        rq = this_rq();
 
-       /* Make sure rq->nr_running update is visible after the IPI */
-       smp_rmb();
-
-       /* More than one running task need preemption */
+        /*
+         * More than one running task need preemption.
+         * nr_running update is assumed to be visible
+         * after IPI is sent from wakers.
+         */
        if (rq->nr_running > 1)
                return false;
 
@@ -1568,9 +1575,7 @@ void scheduler_ipi(void)
          */
         preempt_fold_need_resched();
 
-        if (llist_empty(&this_rq()->wake_list)
-                        && !tick_nohz_full_cpu(smp_processor_id())
-                        && !got_nohz_idle_kick())
+        if (llist_empty(&this_rq()->wake_list) && !got_nohz_idle_kick())
                 return;
 
         /*
@@ -1587,7 +1592,6 @@ void scheduler_ipi(void)
          * somewhat pessimize the simple resched case.
          */
         irq_enter();
-        tick_nohz_full_check();
         sched_ttwu_pending();
 
         /*