timekeeping: Distangle resume and clock-was-set events

Resuming timekeeping is a clock-was-set event and uses the clock-was-set
notification mechanism. This is in the way of making the clock-was-set
update for hrtimers selective so unnecessary IPIs are avoided when a CPU
base does not have timers queued which are affected by the clock setting.

Distangle it by invoking hrtimer_resume() on each unfreezing CPU and invoke
the new timerfd_resume() function from timekeeping_resume() which is the
only place where this is needed.

Rename hrtimer_resume() to hrtimer_resume_local() to reflect the change.

With this the clock_was_set*() functions are not longer required to IPI all
CPUs unconditionally and can get some smarts to avoid them.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20210713135158.488853478@linutronix.de

5 files changed, 18 insertions, 11 deletions

diff --git a/include/linux/hrtimer.h b/include/linux/hrtimer.h
index 253c6e25f331..0ee140176f10 100644
--- a/include/linux/hrtimer.h
+++ b/include/linux/hrtimer.h
@@ -354,7 +354,6 @@ extern void timerfd_resume(void);
 static inline void timerfd_clock_was_set(void) { }
 static inline void timerfd_resume(void) { }
 #endif
-extern void hrtimers_resume(void);
 
 DECLARE_PER_CPU(struct tick_device, tick_cpu_device);
 
diff --git a/kernel/time/hrtimer.c b/kernel/time/hrtimer.c
index 214fd65a9597..68e56f0ecb09 100644
--- a/kernel/time/hrtimer.c
+++ b/kernel/time/hrtimer.c
@@ -900,8 +900,8 @@ static void clock_was_set_work(struct work_struct *work)
 static DECLARE_WORK(hrtimer_work, clock_was_set_work);
 
 /*
- * Called from timekeeping and resume code to reprogram the hrtimer
- * interrupt device on all cpus and to notify timerfd.
+ * Called from timekeeping code to reprogram the hrtimer interrupt device
+ * on all cpus and to notify timerfd.
  */
 void clock_was_set_delayed(void)
 {
@@ -909,18 +909,15 @@ void clock_was_set_delayed(void)
 }
 
 /*
- * During resume we might have to reprogram the high resolution timer
- * interrupt on all online CPUs.  However, all other CPUs will be
- * stopped with IRQs interrupts disabled so the clock_was_set() call
- * must be deferred.
+ * Called during resume either directly from via timekeeping_resume()
+ * or in the case of s2idle from tick_unfreeze() to ensure that the
+ * hrtimers are up to date.
  */
-void hrtimers_resume(void)
+void hrtimers_resume_local(void)
 {
 	lockdep_assert_irqs_disabled();
 	/* Retrigger on the local CPU */
 	retrigger_next_event(NULL);
-	/* And schedule a retrigger for all others */
-	clock_was_set_delayed();
 }
 
 /*
diff --git a/kernel/time/tick-common.c b/kernel/time/tick-common.c
index d663249652ef..46789356f856 100644
--- a/kernel/time/tick-common.c
+++ b/kernel/time/tick-common.c
@@ -470,6 +470,13 @@ void tick_resume_local(void)
 		else
 			tick_resume_oneshot();
 	}
+
+	/*
+	 * Ensure that hrtimers are up to date and the clockevents device
+	 * is reprogrammed correctly when high resolution timers are
+	 * enabled.
+	 */
+	hrtimers_resume_local();
 }
 
 /**
diff --git a/kernel/time/tick-internal.h b/kernel/time/tick-internal.h
index cd610faa2523..22de98cc6dd8 100644
--- a/kernel/time/tick-internal.h
+++ b/kernel/time/tick-internal.h
@@ -168,3 +168,5 @@ void timer_clear_idle(void);
 
 void clock_was_set(void);
 void clock_was_set_delayed(void);
+
+void hrtimers_resume_local(void);
diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c
index 8a364aa9881a..c8a9b9e54c9d 100644
--- a/kernel/time/timekeeping.c
+++ b/kernel/time/timekeeping.c
@@ -1810,8 +1810,10 @@ void timekeeping_resume(void)
 
 	touch_softlockup_watchdog();
 
+	/* Resume the clockevent device(s) and hrtimers */
 	tick_resume();
-	hrtimers_resume();
+	/* Notify timerfd as resume is equivalent to clock_was_set() */
+	timerfd_resume();
 }
 
 int timekeeping_suspend(void)