nohz: Fix stale jiffies update in tick_nohz_restart()
[linux-flexiantxendom0.git] / kernel / time / tick-sched.c
index 95e41f7..c923640 100644 (file)
@@ -9,7 +9,7 @@
  *
  *  Started by: Thomas Gleixner and Ingo Molnar
  *
- *  For licencing details see kernel-base/COPYING
+ *  Distribute under GPLv2.
  */
 #include <linux/cpu.h>
 #include <linux/err.h>
@@ -19,7 +19,9 @@
 #include <linux/percpu.h>
 #include <linux/profile.h>
 #include <linux/sched.h>
-#include <linux/tick.h>
+#include <linux/module.h>
+
+#include <asm/irq_regs.h>
 
 #include "tick-internal.h"
 
@@ -46,6 +48,13 @@ static void tick_do_update_jiffies64(ktime_t now)
        unsigned long ticks = 0;
        ktime_t delta;
 
+       /*
+        * Do a quick check without holding xtime_lock:
+        */
+       delta = ktime_sub(now, last_jiffies_update);
+       if (delta.tv64 < tick_period.tv64)
+               return;
+
        /* Reevalute with xtime_lock held */
        write_seqlock(&xtime_lock);
 
@@ -66,6 +75,9 @@ static void tick_do_update_jiffies64(ktime_t now)
                                                           incr * ticks);
                }
                do_timer(++ticks);
+
+               /* Keep the tick_next_period variable up to date */
+               tick_next_period = ktime_add(last_jiffies_update, tick_period);
        }
        write_sequnlock(&xtime_lock);
 }
@@ -121,36 +133,162 @@ __setup("nohz=", setup_tick_nohz);
  * value. We do this unconditionally on any cpu, as we don't know whether the
  * cpu, which has the update task assigned is in a long sleep.
  */
-void tick_nohz_update_jiffies(void)
+static void tick_nohz_update_jiffies(ktime_t now)
 {
        int cpu = smp_processor_id();
        struct tick_sched *ts = &per_cpu(tick_cpu_sched, cpu);
        unsigned long flags;
-       ktime_t now;
 
-       if (!ts->tick_stopped)
-               return;
-
-       cpu_clear(cpu, nohz_cpu_mask);
-       now = ktime_get();
+       ts->idle_waketime = now;
 
        local_irq_save(flags);
        tick_do_update_jiffies64(now);
        local_irq_restore(flags);
+
+       touch_softlockup_watchdog();
+}
+
+/*
+ * Updates the per cpu time idle statistics counters
+ */
+static void
+update_ts_time_stats(int cpu, struct tick_sched *ts, ktime_t now, u64 *last_update_time)
+{
+       ktime_t delta;
+
+       if (ts->idle_active) {
+               delta = ktime_sub(now, ts->idle_entrytime);
+               if (nr_iowait_cpu(cpu) > 0)
+                       ts->iowait_sleeptime = ktime_add(ts->iowait_sleeptime, delta);
+               else
+                       ts->idle_sleeptime = ktime_add(ts->idle_sleeptime, delta);
+               ts->idle_entrytime = now;
+       }
+
+       if (last_update_time)
+               *last_update_time = ktime_to_us(now);
+
+}
+
+static void tick_nohz_stop_idle(int cpu, ktime_t now)
+{
+       struct tick_sched *ts = &per_cpu(tick_cpu_sched, cpu);
+
+       update_ts_time_stats(cpu, ts, now, NULL);
+       ts->idle_active = 0;
+
+       sched_clock_idle_wakeup_event(0);
+}
+
+static ktime_t tick_nohz_start_idle(int cpu, struct tick_sched *ts)
+{
+       ktime_t now;
+
+       now = ktime_get();
+
+       update_ts_time_stats(cpu, ts, now, NULL);
+
+       ts->idle_entrytime = now;
+       ts->idle_active = 1;
+       sched_clock_idle_sleep_event();
+       return now;
 }
 
 /**
+ * get_cpu_idle_time_us - get the total idle time of a cpu
+ * @cpu: CPU number to query
+ * @last_update_time: variable to store update time in. Do not update
+ * counters if NULL.
+ *
+ * Return the cummulative idle time (since boot) for a given
+ * CPU, in microseconds.
+ *
+ * This time is measured via accounting rather than sampling,
+ * and is as accurate as ktime_get() is.
+ *
+ * This function returns -1 if NOHZ is not enabled.
+ */
+u64 get_cpu_idle_time_us(int cpu, u64 *last_update_time)
+{
+       struct tick_sched *ts = &per_cpu(tick_cpu_sched, cpu);
+       ktime_t now, idle;
+
+       if (!tick_nohz_enabled)
+               return -1;
+
+       now = ktime_get();
+       if (last_update_time) {
+               update_ts_time_stats(cpu, ts, now, last_update_time);
+               idle = ts->idle_sleeptime;
+       } else {
+               if (ts->idle_active && !nr_iowait_cpu(cpu)) {
+                       ktime_t delta = ktime_sub(now, ts->idle_entrytime);
+
+                       idle = ktime_add(ts->idle_sleeptime, delta);
+               } else {
+                       idle = ts->idle_sleeptime;
+               }
+       }
+
+       return ktime_to_us(idle);
+
+}
+EXPORT_SYMBOL_GPL(get_cpu_idle_time_us);
+
+/**
+ * get_cpu_iowait_time_us - get the total iowait time of a cpu
+ * @cpu: CPU number to query
+ * @last_update_time: variable to store update time in. Do not update
+ * counters if NULL.
+ *
+ * Return the cummulative iowait time (since boot) for a given
+ * CPU, in microseconds.
+ *
+ * This time is measured via accounting rather than sampling,
+ * and is as accurate as ktime_get() is.
+ *
+ * This function returns -1 if NOHZ is not enabled.
+ */
+u64 get_cpu_iowait_time_us(int cpu, u64 *last_update_time)
+{
+       struct tick_sched *ts = &per_cpu(tick_cpu_sched, cpu);
+       ktime_t now, iowait;
+
+       if (!tick_nohz_enabled)
+               return -1;
+
+       now = ktime_get();
+       if (last_update_time) {
+               update_ts_time_stats(cpu, ts, now, last_update_time);
+               iowait = ts->iowait_sleeptime;
+       } else {
+               if (ts->idle_active && nr_iowait_cpu(cpu) > 0) {
+                       ktime_t delta = ktime_sub(now, ts->idle_entrytime);
+
+                       iowait = ktime_add(ts->iowait_sleeptime, delta);
+               } else {
+                       iowait = ts->iowait_sleeptime;
+               }
+       }
+
+       return ktime_to_us(iowait);
+}
+EXPORT_SYMBOL_GPL(get_cpu_iowait_time_us);
+
+/**
  * tick_nohz_stop_sched_tick - stop the idle tick from the idle task
  *
  * When the next event is more than a tick into the future, stop the idle tick
  * Called either from the idle loop or from irq_exit() when an idle period was
  * just interrupted by an interrupt which did not cause a reschedule.
  */
-void tick_nohz_stop_sched_tick(void)
+void tick_nohz_stop_sched_tick(int inidle)
 {
        unsigned long seq, last_jiffies, next_jiffies, delta_jiffies, flags;
        struct tick_sched *ts;
-       ktime_t last_update, expires, now, delta;
+       ktime_t last_update, expires, now;
+       struct clock_event_device *dev = __get_cpu_var(tick_cpu_device).evtdev;
+       u64 time_delta;
        int cpu;
 
        local_irq_save(flags);
@@ -158,53 +296,131 @@ void tick_nohz_stop_sched_tick(void)
        cpu = smp_processor_id();
        ts = &per_cpu(tick_cpu_sched, cpu);
 
+       /*
+        * Call to tick_nohz_start_idle stops the last_update_time from being
+        * updated. Thus, it must not be called in the event we are called from
+        * irq_exit() with the prior state different than idle.
+        */
+       if (!inidle && !ts->inidle)
+               goto end;
+
+       /*
+        * Set ts->inidle unconditionally. Even if the system did not
+        * switch to NOHZ mode the cpu frequency governers rely on the
+        * update of the idle time accounting in tick_nohz_start_idle().
+        */
+       ts->inidle = 1;
+
+       now = tick_nohz_start_idle(cpu, ts);
+
+       /*
+        * If this cpu is offline and it is the one which updates
+        * jiffies, then give up the assignment and let it be taken by
+        * the cpu which runs the tick timer next. If we don't drop
+        * this here the jiffies might be stale and do_timer() never
+        * invoked.
+        */
+       if (unlikely(!cpu_online(cpu))) {
+               if (cpu == tick_do_timer_cpu)
+                       tick_do_timer_cpu = TICK_DO_TIMER_NONE;
+       }
+
        if (unlikely(ts->nohz_mode == NOHZ_MODE_INACTIVE))
                goto end;
 
        if (need_resched())
                goto end;
 
-       cpu = smp_processor_id();
-       BUG_ON(local_softirq_pending());
+       if (unlikely(local_softirq_pending() && cpu_online(cpu))) {
+               static int ratelimit;
 
-       now = ktime_get();
-       /*
-        * When called from irq_exit we need to account the idle sleep time
-        * correctly.
-        */
-       if (ts->tick_stopped) {
-               delta = ktime_sub(now, ts->idle_entrytime);
-               ts->idle_sleeptime = ktime_add(ts->idle_sleeptime, delta);
+               if (ratelimit < 10) {
+                       printk(KERN_ERR "NOHZ: local_softirq_pending %02x\n",
+                              (unsigned int) local_softirq_pending());
+                       ratelimit++;
+               }
+               goto end;
        }
 
-       ts->idle_entrytime = now;
        ts->idle_calls++;
-
        /* Read jiffies and the time when jiffies were updated last */
        do {
                seq = read_seqbegin(&xtime_lock);
                last_update = last_jiffies_update;
                last_jiffies = jiffies;
+               time_delta = timekeeping_max_deferment();
        } while (read_seqretry(&xtime_lock, seq));
 
-       /* Get the next timer wheel timer */
-       next_jiffies = get_next_timer_interrupt(last_jiffies);
-       delta_jiffies = next_jiffies - last_jiffies;
-
+       if (rcu_needs_cpu(cpu) || printk_needs_cpu(cpu) ||
+           arch_needs_cpu(cpu)) {
+               next_jiffies = last_jiffies + 1;
+               delta_jiffies = 1;
+       } else {
+               /* Get the next timer wheel timer */
+               next_jiffies = get_next_timer_interrupt(last_jiffies);
+               delta_jiffies = next_jiffies - last_jiffies;
+       }
        /*
         * Do not stop the tick, if we are only one off
         * or if the cpu is required for rcu
         */
-       if (!ts->tick_stopped && (delta_jiffies == 1 || rcu_needs_cpu(cpu)))
+       if (!ts->tick_stopped && delta_jiffies == 1)
                goto out;
 
        /* Schedule the tick, if we are at least one jiffie off */
        if ((long)delta_jiffies >= 1) {
 
-               if (rcu_needs_cpu(cpu))
-                       delta_jiffies = 1;
+               /*
+                * If this cpu is the one which updates jiffies, then
+                * give up the assignment and let it be taken by the
+                * cpu which runs the tick timer next, which might be
+                * this cpu as well. If we don't drop this here the
+                * jiffies might be stale and do_timer() never
+                * invoked. Keep track of the fact that it was the one
+                * which had the do_timer() duty last. If this cpu is
+                * the one which had the do_timer() duty last, we
+                * limit the sleep time to the timekeeping
+                * max_deferement value which we retrieved
+                * above. Otherwise we can sleep as long as we want.
+                */
+               if (cpu == tick_do_timer_cpu) {
+                       tick_do_timer_cpu = TICK_DO_TIMER_NONE;
+                       ts->do_timer_last = 1;
+               } else if (tick_do_timer_cpu != TICK_DO_TIMER_NONE) {
+                       time_delta = KTIME_MAX;
+                       ts->do_timer_last = 0;
+               } else if (!ts->do_timer_last) {
+                       time_delta = KTIME_MAX;
+               }
+
+               /*
+                * calculate the expiry time for the next timer wheel
+                * timer. delta_jiffies >= NEXT_TIMER_MAX_DELTA signals
+                * that there is no timer pending or at least extremely
+                * far into the future (12 days for HZ=1000). In this
+                * case we set the expiry to the end of time.
+                */
+               if (likely(delta_jiffies < NEXT_TIMER_MAX_DELTA)) {
+                       /*
+                        * Calculate the time delta for the next timer event.
+                        * If the time delta exceeds the maximum time delta
+                        * permitted by the current clocksource then adjust
+                        * the time delta accordingly to ensure the
+                        * clocksource does not wrap.
+                        */
+                       time_delta = min_t(u64, time_delta,
+                                          tick_period.tv64 * delta_jiffies);
+               }
+
+               if (time_delta < KTIME_MAX)
+                       expires = ktime_add_ns(last_update, time_delta);
                else
-                       cpu_set(cpu, nohz_cpu_mask);
+                       expires.tv64 = KTIME_MAX;
+
+               /* Skip reprogram of event if its not changed */
+               if (ts->tick_stopped && ktime_equal(expires, dev->next_event))
+                       goto out;
+
                /*
                 * nohz_stop_sched_tick can be called several times before
                 * the nohz_restart_sched_tick is called. This happens when
@@ -213,26 +429,36 @@ void tick_nohz_stop_sched_tick(void)
                 * the scheduler tick in nohz_restart_sched_tick.
                 */
                if (!ts->tick_stopped) {
-                       ts->idle_tick = ts->sched_timer.expires;
+                       select_nohz_load_balancer(1);
+
+                       ts->idle_tick = hrtimer_get_expires(&ts->sched_timer);
                        ts->tick_stopped = 1;
                        ts->idle_jiffies = last_jiffies;
+                       rcu_enter_nohz();
                }
+
+               ts->idle_sleeps++;
+
+               /* Mark expires */
+               ts->idle_expires = expires;
+
                /*
-                * calculate the expiry time for the next timer wheel
-                * timer
+                * If the expiration time == KTIME_MAX, then
+                * in this case we simply stop the tick timer.
                 */
-               expires = ktime_add_ns(last_update, tick_period.tv64 *
-                                      delta_jiffies);
-               ts->idle_expires = expires;
-               ts->idle_sleeps++;
+                if (unlikely(expires.tv64 == KTIME_MAX)) {
+                       if (ts->nohz_mode == NOHZ_MODE_HIGHRES)
+                               hrtimer_cancel(&ts->sched_timer);
+                       goto out;
+               }
 
                if (ts->nohz_mode == NOHZ_MODE_HIGHRES) {
                        hrtimer_start(&ts->sched_timer, expires,
-                                     HRTIMER_MODE_ABS);
+                                     HRTIMER_MODE_ABS_PINNED);
                        /* Check, if the timer was already in the past */
                        if (hrtimer_active(&ts->sched_timer))
                                goto out;
-               } else if(!tick_program_event(expires, 0))
+               } else if (!tick_program_event(expires, 0))
                                goto out;
                /*
                 * We are past the event already. So we crossed a
@@ -240,18 +466,56 @@ void tick_nohz_stop_sched_tick(void)
                 * softirq.
                 */
                tick_do_update_jiffies64(ktime_get());
-               cpu_clear(cpu, nohz_cpu_mask);
        }
        raise_softirq_irqoff(TIMER_SOFTIRQ);
 out:
        ts->next_jiffies = next_jiffies;
        ts->last_jiffies = last_jiffies;
+       ts->sleep_length = ktime_sub(dev->next_event, now);
 end:
        local_irq_restore(flags);
 }
 
 /**
- * nohz_restart_sched_tick - restart the idle tick from the idle task
+ * tick_nohz_get_sleep_length - return the length of the current sleep
+ *
+ * Called from power state control code with interrupts disabled
+ */
+ktime_t tick_nohz_get_sleep_length(void)
+{
+       struct tick_sched *ts = &__get_cpu_var(tick_cpu_sched);
+
+       return ts->sleep_length;
+}
+
+static void tick_nohz_restart(struct tick_sched *ts, ktime_t now)
+{
+       hrtimer_cancel(&ts->sched_timer);
+       hrtimer_set_expires(&ts->sched_timer, ts->idle_tick);
+
+       while (1) {
+               /* Forward the time to expire in the future */
+               hrtimer_forward(&ts->sched_timer, now, tick_period);
+
+               if (ts->nohz_mode == NOHZ_MODE_HIGHRES) {
+                       hrtimer_start_expires(&ts->sched_timer,
+                                             HRTIMER_MODE_ABS_PINNED);
+                       /* Check, if the timer was already in the past */
+                       if (hrtimer_active(&ts->sched_timer))
+                               break;
+               } else {
+                       if (!tick_program_event(
+                               hrtimer_get_expires(&ts->sched_timer), 0))
+                               break;
+               }
+               /* Reread time and update jiffies */
+               now = ktime_get();
+               tick_do_update_jiffies64(now);
+       }
+}
+
+/**
+ * tick_nohz_restart_sched_tick - restart the idle tick from the idle task
  *
  * Restart the idle tick when the CPU is woken up from idle
  */
@@ -259,23 +523,33 @@ void tick_nohz_restart_sched_tick(void)
 {
        int cpu = smp_processor_id();
        struct tick_sched *ts = &per_cpu(tick_cpu_sched, cpu);
+#ifndef CONFIG_VIRT_CPU_ACCOUNTING
        unsigned long ticks;
-       ktime_t now, delta;
+#endif
+       ktime_t now;
+
+       local_irq_disable();
+       if (ts->idle_active || (ts->inidle && ts->tick_stopped))
+               now = ktime_get();
 
-       if (!ts->tick_stopped)
+       if (ts->idle_active)
+               tick_nohz_stop_idle(cpu, now);
+
+       if (!ts->inidle || !ts->tick_stopped) {
+               ts->inidle = 0;
+               local_irq_enable();
                return;
+       }
 
-       /* Update jiffies first */
-       now = ktime_get();
+       ts->inidle = 0;
 
-       local_irq_disable();
-       tick_do_update_jiffies64(now);
-       cpu_clear(cpu, nohz_cpu_mask);
+       rcu_exit_nohz();
 
-       /* Account the idle time */
-       delta = ktime_sub(now, ts->idle_entrytime);
-       ts->idle_sleeptime = ktime_add(ts->idle_sleeptime, delta);
+       /* Update jiffies first */
+       select_nohz_load_balancer(0);
+       tick_do_update_jiffies64(now);
 
+#ifndef CONFIG_VIRT_CPU_ACCOUNTING
        /*
         * We stopped the tick in idle. Update process times would miss the
         * time we slept as update_process_times does only a 1 tick
@@ -285,46 +559,26 @@ void tick_nohz_restart_sched_tick(void)
        /*
         * We might be one off. Do not randomly account a huge number of ticks!
         */
-       if (ticks && ticks < LONG_MAX) {
-               add_preempt_count(HARDIRQ_OFFSET);
-               account_system_time(current, HARDIRQ_OFFSET,
-                                   jiffies_to_cputime(ticks));
-               sub_preempt_count(HARDIRQ_OFFSET);
-       }
+       if (ticks && ticks < LONG_MAX)
+               account_idle_ticks(ticks);
+#endif
 
+       touch_softlockup_watchdog();
        /*
         * Cancel the scheduled timer and restore the tick
         */
        ts->tick_stopped  = 0;
-       hrtimer_cancel(&ts->sched_timer);
-       ts->sched_timer.expires = ts->idle_tick;
+       ts->idle_exittime = now;
 
-       while (1) {
-               /* Forward the time to expire in the future */
-               hrtimer_forward(&ts->sched_timer, now, tick_period);
+       tick_nohz_restart(ts, now);
 
-               if (ts->nohz_mode == NOHZ_MODE_HIGHRES) {
-                       hrtimer_start(&ts->sched_timer,
-                                     ts->sched_timer.expires,
-                                     HRTIMER_MODE_ABS);
-                       /* Check, if the timer was already in the past */
-                       if (hrtimer_active(&ts->sched_timer))
-                               break;
-               } else {
-                       if (!tick_program_event(ts->sched_timer.expires, 0))
-                               break;
-               }
-               /* Update jiffies and reread time */
-               tick_do_update_jiffies64(now);
-               now = ktime_get();
-       }
        local_irq_enable();
 }
 
 static int tick_nohz_reprogram(struct tick_sched *ts, ktime_t now)
 {
        hrtimer_forward(&ts->sched_timer, now, tick_period);
-       return tick_program_event(ts->sched_timer.expires, 0);
+       return tick_program_event(hrtimer_get_expires(&ts->sched_timer), 0);
 }
 
 /*
@@ -334,12 +588,24 @@ static void tick_nohz_handler(struct clock_event_device *dev)
 {
        struct tick_sched *ts = &__get_cpu_var(tick_cpu_sched);
        struct pt_regs *regs = get_irq_regs();
+       int cpu = smp_processor_id();
        ktime_t now = ktime_get();
 
        dev->next_event.tv64 = KTIME_MAX;
 
+       /*
+        * Check if the do_timer duty was dropped. We don't care about
+        * concurrency: This happens only when the cpu in charge went
+        * into a long sleep. If two cpus happen to assign themself to
+        * this duty, then the jiffies update is still serialized by
+        * xtime_lock.
+        */
+       if (unlikely(tick_do_timer_cpu == TICK_DO_TIMER_NONE))
+               tick_do_timer_cpu = cpu;
+
        /* Check, if the jiffies need an update */
-       tick_do_update_jiffies64(now);
+       if (tick_do_timer_cpu == cpu)
+               tick_do_update_jiffies64(now);
 
        /*
         * When we are idle and the tick is stopped, we have to touch
@@ -357,10 +623,6 @@ static void tick_nohz_handler(struct clock_event_device *dev)
        update_process_times(user_mode(regs));
        profile_tick(CPU_PROFILING);
 
-       /* Do not restart, when we are in the idle loop */
-       if (ts->tick_stopped)
-               return;
-
        while (tick_nohz_reprogram(ts, now)) {
                now = ktime_get();
                tick_do_update_jiffies64(now);
@@ -395,41 +657,108 @@ static void tick_nohz_switch_to_nohz(void)
        next = tick_init_jiffy_update();
 
        for (;;) {
-               ts->sched_timer.expires = next;
+               hrtimer_set_expires(&ts->sched_timer, next);
                if (!tick_program_event(next, 0))
                        break;
                next = ktime_add(next, tick_period);
        }
        local_irq_enable();
+}
 
-       printk(KERN_INFO "Switched to NOHz mode on CPU #%d\n",
-              smp_processor_id());
+/*
+ * When NOHZ is enabled and the tick is stopped, we need to kick the
+ * tick timer from irq_enter() so that the jiffies update is kept
+ * alive during long running softirqs. That's ugly as hell, but
+ * correctness is key even if we need to fix the offending softirq in
+ * the first place.
+ *
+ * Note, this is different to tick_nohz_restart. We just kick the
+ * timer and do not touch the other magic bits which need to be done
+ * when idle is left.
+ */
+static void tick_nohz_kick_tick(int cpu, ktime_t now)
+{
+#if 0
+       /* Switch back to 2.6.27 behaviour */
+
+       struct tick_sched *ts = &per_cpu(tick_cpu_sched, cpu);
+       ktime_t delta;
+
+       /*
+        * Do not touch the tick device, when the next expiry is either
+        * already reached or less/equal than the tick period.
+        */
+       delta = ktime_sub(hrtimer_get_expires(&ts->sched_timer), now);
+       if (delta.tv64 <= tick_period.tv64)
+               return;
+
+       tick_nohz_restart(ts, now);
+#endif
+}
+
+static inline void tick_check_nohz(int cpu)
+{
+       struct tick_sched *ts = &per_cpu(tick_cpu_sched, cpu);
+       ktime_t now;
+
+       if (!ts->idle_active && !ts->tick_stopped)
+               return;
+       now = ktime_get();
+       if (ts->idle_active)
+               tick_nohz_stop_idle(cpu, now);
+       if (ts->tick_stopped) {
+               tick_nohz_update_jiffies(now);
+               tick_nohz_kick_tick(cpu, now);
+       }
 }
 
 #else
 
 static inline void tick_nohz_switch_to_nohz(void) { }
+static inline void tick_check_nohz(int cpu) { }
 
 #endif /* NO_HZ */
 
 /*
+ * Called from irq_enter to notify about the possible interruption of idle()
+ */
+void tick_check_idle(int cpu)
+{
+       tick_check_oneshot_broadcast(cpu);
+       tick_check_nohz(cpu);
+}
+
+/*
  * High resolution timer specific code
  */
 #ifdef CONFIG_HIGH_RES_TIMERS
 /*
- * We rearm the timer until we get disabled by the idle code
+ * We rearm the timer until we get disabled by the idle code.
  * Called with interrupts disabled and timer->base->cpu_base->lock held.
  */
 static enum hrtimer_restart tick_sched_timer(struct hrtimer *timer)
 {
        struct tick_sched *ts =
                container_of(timer, struct tick_sched, sched_timer);
-       struct hrtimer_cpu_base *base = timer->base->cpu_base;
        struct pt_regs *regs = get_irq_regs();
        ktime_t now = ktime_get();
+       int cpu = smp_processor_id();
+
+#ifdef CONFIG_NO_HZ
+       /*
+        * Check if the do_timer duty was dropped. We don't care about
+        * concurrency: This happens only when the cpu in charge went
+        * into a long sleep. If two cpus happen to assign themself to
+        * this duty, then the jiffies update is still serialized by
+        * xtime_lock.
+        */
+       if (unlikely(tick_do_timer_cpu == TICK_DO_TIMER_NONE))
+               tick_do_timer_cpu = cpu;
+#endif
 
        /* Check, if the jiffies need an update */
-       tick_do_update_jiffies64(now);
+       if (tick_do_timer_cpu == cpu)
+               tick_do_update_jiffies64(now);
 
        /*
         * Do not call, when we are not in irq context and have
@@ -448,21 +777,10 @@ static enum hrtimer_restart tick_sched_timer(struct hrtimer *timer)
                        touch_softlockup_watchdog();
                        ts->idle_jiffies++;
                }
-               /*
-                * update_process_times() might take tasklist_lock, hence
-                * drop the base lock. sched-tick hrtimers are per-CPU and
-                * never accessible by userspace APIs, so this is safe to do.
-                */
-               spin_unlock(&base->lock);
                update_process_times(user_mode(regs));
                profile_tick(CPU_PROFILING);
-               spin_lock(&base->lock);
        }
 
-       /* Do not restart, when we are in the idle loop */
-       if (ts->tick_stopped)
-               return HRTIMER_NORESTART;
-
        hrtimer_forward(timer, now, tick_period);
 
        return HRTIMER_RESTART;
@@ -481,15 +799,14 @@ void tick_setup_sched_timer(void)
         */
        hrtimer_init(&ts->sched_timer, CLOCK_MONOTONIC, HRTIMER_MODE_ABS);
        ts->sched_timer.function = tick_sched_timer;
-       ts->sched_timer.cb_mode = HRTIMER_CB_IRQSAFE_NO_SOFTIRQ;
 
-       /* Get the next period */
-       ts->sched_timer.expires = tick_init_jiffy_update();
+       /* Get the next period (per cpu) */
+       hrtimer_set_expires(&ts->sched_timer, tick_init_jiffy_update());
 
        for (;;) {
                hrtimer_forward(&ts->sched_timer, now, tick_period);
-               hrtimer_start(&ts->sched_timer, ts->sched_timer.expires,
-                             HRTIMER_MODE_ABS);
+               hrtimer_start_expires(&ts->sched_timer,
+                                     HRTIMER_MODE_ABS_PINNED);
                /* Check, if the timer was already in the past */
                if (hrtimer_active(&ts->sched_timer))
                        break;
@@ -501,17 +818,21 @@ void tick_setup_sched_timer(void)
                ts->nohz_mode = NOHZ_MODE_HIGHRES;
 #endif
 }
+#endif /* HIGH_RES_TIMERS */
 
+#if defined CONFIG_NO_HZ || defined CONFIG_HIGH_RES_TIMERS
 void tick_cancel_sched_timer(int cpu)
 {
        struct tick_sched *ts = &per_cpu(tick_cpu_sched, cpu);
 
+# ifdef CONFIG_HIGH_RES_TIMERS
        if (ts->sched_timer.base)
                hrtimer_cancel(&ts->sched_timer);
-       ts->tick_stopped = 0;
+# endif
+
        ts->nohz_mode = NOHZ_MODE_INACTIVE;
 }
-#endif /* HIGH_RES_TIMERS */
+#endif
 
 /**
  * Async notification about clocksource changes
@@ -552,7 +873,7 @@ int tick_check_oneshot_change(int allow_nohz)
        if (ts->nohz_mode != NOHZ_MODE_INACTIVE)
                return 0;
 
-       if (!timekeeping_is_continuous() || !tick_is_oneshot_available())
+       if (!timekeeping_valid_for_hres() || !tick_is_oneshot_available())
                return 0;
 
        if (!allow_nohz)