Merge branch 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip (87a54cae) · Commits · 戴 / test

Pull timer fixes from Thomas Gleixner: "Viresh unearthed the following three hickups in the timer/timekeeping code: - Negated check for the result of a clock event selection - A missing early exit in the jiffies update path which causes update_wall_time to be called for nothing causing lock contention and wasted cycles in the timer interrupt - Checking a variable in the NOHZ code enable code for true which can only be set by that very code after the check succeeds. That results in a rock solid runtime disablement of that feature" * 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: tick-sched: Check tick_nohz_enabled in tick_nohz_switch_to_nohz() tick-sched: Don't call update_wall_time() when delta is lesser than tick_period tick-common: Fix wrong check in tick_check_replacement()

kernel/time/tick-common.c

+1 −1

Original line number	Diff line number	Diff line
		@@ -276,7 +276,7 @@ static bool tick_check_preferred(struct clock_event_device *curdev,
		bool tick_check_replacement(struct clock_event_device *curdev,
		struct clock_event_device *newdev)
		{
		if (tick_check_percpu(curdev, newdev, smp_processor_id()))
		if (!tick_check_percpu(curdev, newdev, smp_processor_id()))
		return false;

		return tick_check_preferred(curdev, newdev);

kernel/time/tick-sched.c

+4 −1

Original line number	Diff line number	Diff line
		@@ -84,6 +84,9 @@ static void tick_do_update_jiffies64(ktime_t now)

		/* Keep the tick_next_period variable up to date */
		tick_next_period = ktime_add(last_jiffies_update, tick_period);
		} else {
		write_sequnlock(&jiffies_lock);
		return;
		}
		write_sequnlock(&jiffies_lock);
		update_wall_time();
		@@ -967,7 +970,7 @@ static void tick_nohz_switch_to_nohz(void)
		struct tick_sched *ts = &__get_cpu_var(tick_cpu_sched);
		ktime_t next;

		if (!tick_nohz_active)
		if (!tick_nohz_enabled)
		return;

		local_irq_disable();

Admin message