sched: idle: Do not stop the tick upfront in the idle loop

Push the decision whether or not to stop the tick somewhat deeper
into the idle loop.

Stopping the tick upfront leads to unpleasant outcomes in case the
idle governor doesn't agree with the nohz code on the duration of the
upcoming idle period.  Specifically, if the tick has been stopped and
the idle governor predicts short idle, the situation is bad regardless
of whether or not the prediction is accurate.  If it is accurate, the
tick has been stopped unnecessarily which means excessive overhead.
If it is not accurate, the CPU is likely to spend too much time in
the (shallow, because short idle has been predicted) idle state
selected by the governor [1].

As the first step towards addressing this problem, change the code
to make the tick stopping decision inside of the loop in do_idle().
In particular, do not stop the tick in the cpu_idle_poll() code path.
Also don't do that in tick_nohz_irq_exit() which doesn't really have
enough information on whether or not to stop the tick.

Link: https://marc.info/?l=linux-pm&m=150116085925208&w=2 # [1]
Link: https://tu-dresden.de/zih/forschung/ressourcen/dateien/projekte/haec/powernightmares.pdf
Suggested-by: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>

3 files changed, 26 insertions, 11 deletions

diff --git a/include/linux/tick.h b/include/linux/tick.h
index 1d253df9ea3c..fccebfba167e 100644
--- a/include/linux/tick.h
+++ b/include/linux/tick.h
@@ -116,6 +116,7 @@ extern bool tick_nohz_enabled;
 extern bool tick_nohz_tick_stopped(void);
 extern bool tick_nohz_tick_stopped_cpu(int cpu);
 extern void tick_nohz_idle_stop_tick(void);
+extern void tick_nohz_idle_restart_tick(void);
 extern void tick_nohz_idle_enter(void);
 extern void tick_nohz_idle_exit(void);
 extern void tick_nohz_irq_exit(void);
@@ -137,6 +138,7 @@ static inline void tick_nohz_idle_stop_tick_protected(void)
 static inline int tick_nohz_tick_stopped(void) { return 0; }
 static inline int tick_nohz_tick_stopped_cpu(int cpu) { return 0; }
 static inline void tick_nohz_idle_stop_tick(void) { }
+static inline void tick_nohz_idle_restart_tick(void) { }
 static inline void tick_nohz_idle_enter(void) { }
 static inline void tick_nohz_idle_exit(void) { }
 
diff --git a/kernel/sched/idle.c b/kernel/sched/idle.c
index c0bc111878e6..3777e83c0b5a 100644
--- a/kernel/sched/idle.c
+++ b/kernel/sched/idle.c
@@ -216,13 +216,13 @@ static void do_idle(void)
 
         __current_set_polling();
         tick_nohz_idle_enter();
-        tick_nohz_idle_stop_tick_protected();
 
         while (!need_resched()) {
                 check_pgt_cache();
                 rmb();
 
                 if (cpu_is_offline(cpu)) {
+                        tick_nohz_idle_stop_tick_protected();
                         cpuhp_report_idle_dead();
                         arch_cpu_idle_dead();
                 }
@@ -236,10 +236,13 @@ static void do_idle(void)
                  * broadcast device expired for us, we don't want to go deep
                  * idle as we know that the IPI is going to arrive right away.
                  */
-                if (cpu_idle_force_poll || tick_check_broadcast_expired())
+                if (cpu_idle_force_poll || tick_check_broadcast_expired()) {
+                        tick_nohz_idle_restart_tick();
                         cpu_idle_poll();
-                else
+                } else {
+                        tick_nohz_idle_stop_tick();
                         cpuidle_idle_call();
+                }
                 arch_cpu_idle_exit();
         }
 
diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c
index 678349aec483..f5d37788ea85 100644
--- a/kernel/time/tick-sched.c
+++ b/kernel/time/tick-sched.c
@@ -960,12 +960,10 @@ void tick_nohz_irq_exit(void)
 {
         struct tick_sched *ts = this_cpu_ptr(&tick_cpu_sched);
 
-        if (ts->inidle) {
+        if (ts->inidle)
                 tick_nohz_start_idle(ts);
-                __tick_nohz_idle_stop_tick(ts);
-        } else {
+        else
                 tick_nohz_full_update_tick(ts);
-        }
 }
 
 /**
@@ -1026,6 +1024,20 @@ static void tick_nohz_account_idle_ticks(struct tick_sched *ts)
 #endif
 }
 
+static void __tick_nohz_idle_restart_tick(struct tick_sched *ts, ktime_t now)
+{
+        tick_nohz_restart_sched_tick(ts, now);
+        tick_nohz_account_idle_ticks(ts);
+}
+
+void tick_nohz_idle_restart_tick(void)
+{
+        struct tick_sched *ts = this_cpu_ptr(&tick_cpu_sched);
+
+        if (ts->tick_stopped)
+                __tick_nohz_idle_restart_tick(ts, ktime_get());
+}
+
 /**
  * tick_nohz_idle_exit - restart the idle tick from the idle task
  *
@@ -1050,10 +1062,8 @@ void tick_nohz_idle_exit(void)
         if (ts->idle_active)
                 tick_nohz_stop_idle(ts, now);
 
-        if (ts->tick_stopped) {
-                tick_nohz_restart_sched_tick(ts, now);
-                tick_nohz_account_idle_ticks(ts);
-        }
+        if (ts->tick_stopped)
+                __tick_nohz_idle_restart_tick(ts, now);
 
         local_irq_enable();
 }