summaryrefslogtreecommitdiff
path: root/kernel/sched
diff options
context:
space:
mode:
authorFrederic Weisbecker <fweisbec@gmail.com>2014-06-04 18:20:21 +0400
committerFrederic Weisbecker <fweisbec@gmail.com>2014-06-16 18:26:55 +0400
commit53c5fa16b4c843f1df91f7498e3c7bf95e0eaefa (patch)
treea722c9c88d7c4dde507ffb89974d73e1222991e1 /kernel/sched
parent3d36aebc2e78923095575df954f3f3b430ac0a30 (diff)
downloadlinux-53c5fa16b4c843f1df91f7498e3c7bf95e0eaefa.tar.xz
nohz: Switch to nohz full remote kick on timer enqueue
When a new timer is enqueued on a full dynticks target, that CPU must re-evaluate the next tick to handle the timer correctly. This is currently performed through the scheduler IPI. Meanwhile this happens at the cost of off-topic workarounds in that fast path to make it call irq_exit(). As we plan to remove this hack off the scheduler IPI, lets use the nohz full kick instead. Pretty much any IPI fits for that job as long at it calls irq_exit(). The nohz full kick just happens to be handy and readily available here. If it happens to be too much an overkill in the future, we can still turn that timer kick into an empty IPI. Acked-by: Peter Zijlstra <peterz@infradead.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Ingo Molnar <mingo@kernel.org> Cc: Kevin Hilman <khilman@linaro.org> Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Diffstat (limited to 'kernel/sched')
-rw-r--r--kernel/sched/core.c8
1 files changed, 7 insertions, 1 deletions
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 3bdf01b494fe..feb54965e16f 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;
}