// SPDX-License-Identifier: GPL-2.0 /* * Copyright(C) 2005-2006, Thomas Gleixner <[email protected]> * Copyright(C) 2005-2007, Red Hat, Inc., Ingo Molnar * Copyright(C) 2006-2007 Timesys Corp., Thomas Gleixner * * NOHZ implementation for low and high resolution timers * * Started by: Thomas Gleixner and Ingo Molnar */ #include <linux/compiler.h> #include <linux/cpu.h> #include <linux/err.h> #include <linux/hrtimer.h> #include <linux/interrupt.h> #include <linux/kernel_stat.h> #include <linux/percpu.h> #include <linux/nmi.h> #include <linux/profile.h> #include <linux/sched/signal.h> #include <linux/sched/clock.h> #include <linux/sched/stat.h> #include <linux/sched/nohz.h> #include <linux/sched/loadavg.h> #include <linux/module.h> #include <linux/irq_work.h> #include <linux/posix-timers.h> #include <linux/context_tracking.h> #include <linux/mm.h> #include <asm/irq_regs.h> #include "tick-internal.h" #include <trace/events/timer.h> /* * Per-CPU nohz control structure */ static DEFINE_PER_CPU(struct tick_sched, tick_cpu_sched); struct tick_sched *tick_get_tick_sched(int cpu) { … } /* * The time when the last jiffy update happened. Write access must hold * jiffies_lock and jiffies_seq. tick_nohz_next_event() needs to get a * consistent view of jiffies and last_jiffies_update. */ static ktime_t last_jiffies_update; /* * Must be called with interrupts disabled ! */ static void tick_do_update_jiffies64(ktime_t now) { … } /* * Initialize and return retrieve the jiffies update. */ static ktime_t tick_init_jiffy_update(void) { … } static inline int tick_sched_flag_test(struct tick_sched *ts, unsigned long flag) { … } static inline void tick_sched_flag_set(struct tick_sched *ts, unsigned long flag) { … } static inline void tick_sched_flag_clear(struct tick_sched *ts, unsigned long flag) { … } #define MAX_STALLED_JIFFIES … static void tick_sched_do_timer(struct tick_sched *ts, ktime_t now) { … } static void tick_sched_handle(struct tick_sched *ts, struct pt_regs *regs) { … } /* * We rearm the timer until we get disabled by the idle code. * Called with interrupts disabled. */ static enum hrtimer_restart tick_nohz_handler(struct hrtimer *timer) { … } static void tick_sched_timer_cancel(struct tick_sched *ts) { … } #ifdef CONFIG_NO_HZ_FULL cpumask_var_t tick_nohz_full_mask; EXPORT_SYMBOL_GPL(tick_nohz_full_mask); bool tick_nohz_full_running; EXPORT_SYMBOL_GPL(tick_nohz_full_running); static atomic_t tick_dep_mask; static bool check_tick_dependency(atomic_t *dep) { int val = atomic_read(dep); if (val & TICK_DEP_MASK_POSIX_TIMER) { trace_tick_stop(0, TICK_DEP_MASK_POSIX_TIMER); return true; } if (val & TICK_DEP_MASK_PERF_EVENTS) { trace_tick_stop(0, TICK_DEP_MASK_PERF_EVENTS); return true; } if (val & TICK_DEP_MASK_SCHED) { trace_tick_stop(0, TICK_DEP_MASK_SCHED); return true; } if (val & TICK_DEP_MASK_CLOCK_UNSTABLE) { trace_tick_stop(0, TICK_DEP_MASK_CLOCK_UNSTABLE); return true; } if (val & TICK_DEP_MASK_RCU) { trace_tick_stop(0, TICK_DEP_MASK_RCU); return true; } if (val & TICK_DEP_MASK_RCU_EXP) { trace_tick_stop(0, TICK_DEP_MASK_RCU_EXP); return true; } return false; } static bool can_stop_full_tick(int cpu, struct tick_sched *ts) { lockdep_assert_irqs_disabled(); if (unlikely(!cpu_online(cpu))) return false; if (check_tick_dependency(&tick_dep_mask)) return false; if (check_tick_dependency(&ts->tick_dep_mask)) return false; if (check_tick_dependency(¤t->tick_dep_mask)) return false; if (check_tick_dependency(¤t->signal->tick_dep_mask)) return false; return true; } static void nohz_full_kick_func(struct irq_work *work) { /* Empty, the tick restart happens on tick_nohz_irq_exit() */ } static DEFINE_PER_CPU(struct irq_work, nohz_full_kick_work) = IRQ_WORK_INIT_HARD(nohz_full_kick_func); /* * Kick this CPU if it's full dynticks in order to force it to * re-evaluate its dependency on the tick and restart it if necessary. * This kick, unlike tick_nohz_full_kick_cpu() and tick_nohz_full_kick_all(), * is NMI safe. */ static void tick_nohz_full_kick(void) { if (!tick_nohz_full_cpu(smp_processor_id())) return; irq_work_queue(this_cpu_ptr(&nohz_full_kick_work)); } /* * Kick the CPU if it's full dynticks in order to force it to * re-evaluate its dependency on the tick and restart it if necessary. */ void tick_nohz_full_kick_cpu(int cpu) { if (!tick_nohz_full_cpu(cpu)) return; irq_work_queue_on(&per_cpu(nohz_full_kick_work, cpu), cpu); } static void tick_nohz_kick_task(struct task_struct *tsk) { int cpu; /* * If the task is not running, run_posix_cpu_timers() * has nothing to elapse, and an IPI can then be optimized out. * * activate_task() STORE p->tick_dep_mask * STORE p->on_rq * __schedule() (switch to task 'p') smp_mb() (atomic_fetch_or()) * LOCK rq->lock LOAD p->on_rq * smp_mb__after_spin_lock() * tick_nohz_task_switch() * LOAD p->tick_dep_mask */ if (!sched_task_on_rq(tsk)) return; /* * If the task concurrently migrates to another CPU, * we guarantee it sees the new tick dependency upon * schedule. * * set_task_cpu(p, cpu); * STORE p->cpu = @cpu * __schedule() (switch to task 'p') * LOCK rq->lock * smp_mb__after_spin_lock() STORE p->tick_dep_mask * tick_nohz_task_switch() smp_mb() (atomic_fetch_or()) * LOAD p->tick_dep_mask LOAD p->cpu */ cpu = task_cpu(tsk); preempt_disable(); if (cpu_online(cpu)) tick_nohz_full_kick_cpu(cpu); preempt_enable(); } /* * Kick all full dynticks CPUs in order to force these to re-evaluate * their dependency on the tick and restart it if necessary. */ static void tick_nohz_full_kick_all(void) { int cpu; if (!tick_nohz_full_running) return; preempt_disable(); for_each_cpu_and(cpu, tick_nohz_full_mask, cpu_online_mask) tick_nohz_full_kick_cpu(cpu); preempt_enable(); } static void tick_nohz_dep_set_all(atomic_t *dep, enum tick_dep_bits bit) { int prev; prev = atomic_fetch_or(BIT(bit), dep); if (!prev) tick_nohz_full_kick_all(); } /* * Set a global tick dependency. Used by perf events that rely on freq and * unstable clocks. */ void tick_nohz_dep_set(enum tick_dep_bits bit) { tick_nohz_dep_set_all(&tick_dep_mask, bit); } void tick_nohz_dep_clear(enum tick_dep_bits bit) { atomic_andnot(BIT(bit), &tick_dep_mask); } /* * Set per-CPU tick dependency. Used by scheduler and perf events in order to * manage event-throttling. */ void tick_nohz_dep_set_cpu(int cpu, enum tick_dep_bits bit) { int prev; struct tick_sched *ts; ts = per_cpu_ptr(&tick_cpu_sched, cpu); prev = atomic_fetch_or(BIT(bit), &ts->tick_dep_mask); if (!prev) { preempt_disable(); /* Perf needs local kick that is NMI safe */ if (cpu == smp_processor_id()) { tick_nohz_full_kick(); } else { /* Remote IRQ work not NMI-safe */ if (!WARN_ON_ONCE(in_nmi())) tick_nohz_full_kick_cpu(cpu); } preempt_enable(); } } EXPORT_SYMBOL_GPL(tick_nohz_dep_set_cpu); void tick_nohz_dep_clear_cpu(int cpu, enum tick_dep_bits bit) { struct tick_sched *ts = per_cpu_ptr(&tick_cpu_sched, cpu); atomic_andnot(BIT(bit), &ts->tick_dep_mask); } EXPORT_SYMBOL_GPL(tick_nohz_dep_clear_cpu); /* * Set a per-task tick dependency. RCU needs this. Also posix CPU timers * in order to elapse per task timers. */ void tick_nohz_dep_set_task(struct task_struct *tsk, enum tick_dep_bits bit) { if (!atomic_fetch_or(BIT(bit), &tsk->tick_dep_mask)) tick_nohz_kick_task(tsk); } EXPORT_SYMBOL_GPL(tick_nohz_dep_set_task); void tick_nohz_dep_clear_task(struct task_struct *tsk, enum tick_dep_bits bit) { atomic_andnot(BIT(bit), &tsk->tick_dep_mask); } EXPORT_SYMBOL_GPL(tick_nohz_dep_clear_task); /* * Set a per-taskgroup tick dependency. Posix CPU timers need this in order to elapse * per process timers. */ void tick_nohz_dep_set_signal(struct task_struct *tsk, enum tick_dep_bits bit) { int prev; struct signal_struct *sig = tsk->signal; prev = atomic_fetch_or(BIT(bit), &sig->tick_dep_mask); if (!prev) { struct task_struct *t; lockdep_assert_held(&tsk->sighand->siglock); __for_each_thread(sig, t) tick_nohz_kick_task(t); } } void tick_nohz_dep_clear_signal(struct signal_struct *sig, enum tick_dep_bits bit) { atomic_andnot(BIT(bit), &sig->tick_dep_mask); } /* * Re-evaluate the need for the tick as we switch the current task. * It might need the tick due to per task/process properties: * perf events, posix CPU timers, ... */ void __tick_nohz_task_switch(void) { struct tick_sched *ts; if (!tick_nohz_full_cpu(smp_processor_id())) return; ts = this_cpu_ptr(&tick_cpu_sched); if (tick_sched_flag_test(ts, TS_FLAG_STOPPED)) { if (atomic_read(¤t->tick_dep_mask) || atomic_read(¤t->signal->tick_dep_mask)) tick_nohz_full_kick(); } } /* Get the boot-time nohz CPU list from the kernel parameters. */ void __init tick_nohz_full_setup(cpumask_var_t cpumask) { alloc_bootmem_cpumask_var(&tick_nohz_full_mask); cpumask_copy(tick_nohz_full_mask, cpumask); tick_nohz_full_running = true; } bool tick_nohz_cpu_hotpluggable(unsigned int cpu) { /* * The 'tick_do_timer_cpu' CPU handles housekeeping duty (unbound * timers, workqueues, timekeeping, ...) on behalf of full dynticks * CPUs. It must remain online when nohz full is enabled. */ if (tick_nohz_full_running && READ_ONCE(tick_do_timer_cpu) == cpu) return false; return true; } static int tick_nohz_cpu_down(unsigned int cpu) { return tick_nohz_cpu_hotpluggable(cpu) ? 0 : -EBUSY; } void __init tick_nohz_init(void) { int cpu, ret; if (!tick_nohz_full_running) return; /* * Full dynticks uses IRQ work to drive the tick rescheduling on safe * locking contexts. But then we need IRQ work to raise its own * interrupts to avoid circular dependency on the tick. */ if (!arch_irq_work_has_interrupt()) { pr_warn("NO_HZ: Can't run full dynticks because arch doesn't support IRQ work self-IPIs\n"); cpumask_clear(tick_nohz_full_mask); tick_nohz_full_running = false; return; } if (IS_ENABLED(CONFIG_PM_SLEEP_SMP) && !IS_ENABLED(CONFIG_PM_SLEEP_SMP_NONZERO_CPU)) { cpu = smp_processor_id(); if (cpumask_test_cpu(cpu, tick_nohz_full_mask)) { pr_warn("NO_HZ: Clearing %d from nohz_full range " "for timekeeping\n", cpu); cpumask_clear_cpu(cpu, tick_nohz_full_mask); } } for_each_cpu(cpu, tick_nohz_full_mask) ct_cpu_track_user(cpu); ret = cpuhp_setup_state_nocalls(CPUHP_AP_ONLINE_DYN, "kernel/nohz:predown", NULL, tick_nohz_cpu_down); WARN_ON(ret < 0); pr_info("NO_HZ: Full dynticks CPUs: %*pbl.\n", cpumask_pr_args(tick_nohz_full_mask)); } #endif /* #ifdef CONFIG_NO_HZ_FULL */ /* * NOHZ - aka dynamic tick functionality */ #ifdef CONFIG_NO_HZ_COMMON /* * NO HZ enabled ? */ bool tick_nohz_enabled __read_mostly = …; unsigned long tick_nohz_active __read_mostly; /* * Enable / Disable tickless mode */ static int __init setup_tick_nohz(char *str) { … } __setup(…); bool tick_nohz_tick_stopped(void) { … } bool tick_nohz_tick_stopped_cpu(int cpu) { … } /** * tick_nohz_update_jiffies - update jiffies when idle was interrupted * @now: current ktime_t * * Called from interrupt entry when the CPU was idle * * In case the sched_tick was stopped on this CPU, we have to check if jiffies * must be updated. Otherwise an interrupt handler could use a stale jiffy * 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. */ static void tick_nohz_update_jiffies(ktime_t now) { … } static void tick_nohz_stop_idle(struct tick_sched *ts, ktime_t now) { … } static void tick_nohz_start_idle(struct tick_sched *ts) { … } static u64 get_cpu_sleep_time_us(struct tick_sched *ts, ktime_t *sleeptime, bool compute_delta, u64 *last_update_time) { … } /** * 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 cumulative idle time (since boot) for a given * CPU, in microseconds. Note that this is partially broken due to * the counter of iowait tasks that can be remotely updated without * any synchronization. Therefore it is possible to observe backward * values within two consecutive reads. * * This time is measured via accounting rather than sampling, * and is as accurate as ktime_get() is. * * Return: -1 if NOHZ is not enabled, else total idle time of the @cpu */ u64 get_cpu_idle_time_us(int cpu, u64 *last_update_time) { … } EXPORT_SYMBOL_GPL(…); /** * 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 cumulative iowait time (since boot) for a given * CPU, in microseconds. Note this is partially broken due to * the counter of iowait tasks that can be remotely updated without * any synchronization. Therefore it is possible to observe backward * values within two consecutive reads. * * This time is measured via accounting rather than sampling, * and is as accurate as ktime_get() is. * * Return: -1 if NOHZ is not enabled, else total iowait time of @cpu */ u64 get_cpu_iowait_time_us(int cpu, u64 *last_update_time) { … } EXPORT_SYMBOL_GPL(…); static void tick_nohz_restart(struct tick_sched *ts, ktime_t now) { … } static inline bool local_timer_softirq_pending(void) { … } /* * Read jiffies and the time when jiffies were updated last */ u64 get_jiffies_update(unsigned long *basej) { … } /** * tick_nohz_next_event() - return the clock monotonic based next event * @ts: pointer to tick_sched struct * @cpu: CPU number * * Return: * *%0 - When the next event is a maximum of TICK_NSEC in the future * and the tick is not stopped yet * *%next_event - Next event based on clock monotonic */ static ktime_t tick_nohz_next_event(struct tick_sched *ts, int cpu) { … } static void tick_nohz_stop_tick(struct tick_sched *ts, int cpu) { … } static void tick_nohz_retain_tick(struct tick_sched *ts) { … } #ifdef CONFIG_NO_HZ_FULL static void tick_nohz_full_stop_tick(struct tick_sched *ts, int cpu) { if (tick_nohz_next_event(ts, cpu)) tick_nohz_stop_tick(ts, cpu); else tick_nohz_retain_tick(ts); } #endif /* CONFIG_NO_HZ_FULL */ static void tick_nohz_restart_sched_tick(struct tick_sched *ts, ktime_t now) { … } static void __tick_nohz_full_update_tick(struct tick_sched *ts, ktime_t now) { … } static void tick_nohz_full_update_tick(struct tick_sched *ts) { … } /* * A pending softirq outside an IRQ (or softirq disabled section) context * should be waiting for ksoftirqd to handle it. Therefore we shouldn't * reach this code due to the need_resched() early check in can_stop_idle_tick(). * * However if we are between CPUHP_AP_SMPBOOT_THREADS and CPU_TEARDOWN_CPU on the * cpu_down() process, softirqs can still be raised while ksoftirqd is parked, * triggering the code below, since wakep_softirqd() is ignored. * */ static bool report_idle_softirq(void) { … } static bool can_stop_idle_tick(int cpu, struct tick_sched *ts) { … } /** * tick_nohz_idle_stop_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 */ void tick_nohz_idle_stop_tick(void) { … } void tick_nohz_idle_retain_tick(void) { … } /** * tick_nohz_idle_enter - prepare for entering idle on the current CPU * * Called when we start the idle loop. */ void tick_nohz_idle_enter(void) { … } /** * tick_nohz_irq_exit - Notify the tick about IRQ exit * * A timer may have been added/modified/deleted either by the current IRQ, * or by another place using this IRQ as a notification. This IRQ may have * also updated the RCU callback list. These events may require a * re-evaluation of the next tick. Depending on the context: * * 1) If the CPU is idle and no resched is pending, just proceed with idle * time accounting. The next tick will be re-evaluated on the next idle * loop iteration. * * 2) If the CPU is nohz_full: * * 2.1) If there is any tick dependency, restart the tick if stopped. * * 2.2) If there is no tick dependency, (re-)evaluate the next tick and * stop/update it accordingly. */ void tick_nohz_irq_exit(void) { … } /** * tick_nohz_idle_got_tick - Check whether or not the tick handler has run * * Return: %true if the tick handler has run, otherwise %false */ bool tick_nohz_idle_got_tick(void) { … } /** * tick_nohz_get_next_hrtimer - return the next expiration time for the hrtimer * or the tick, whichever expires first. Note that, if the tick has been * stopped, it returns the next hrtimer. * * Called from power state control code with interrupts disabled * * Return: the next expiration time */ ktime_t tick_nohz_get_next_hrtimer(void) { … } /** * tick_nohz_get_sleep_length - return the expected length of the current sleep * @delta_next: duration until the next event if the tick cannot be stopped * * Called from power state control code with interrupts disabled. * * The return value of this function and/or the value returned by it through the * @delta_next pointer can be negative which must be taken into account by its * callers. * * Return: the expected length of the current sleep */ ktime_t tick_nohz_get_sleep_length(ktime_t *delta_next) { … } /** * tick_nohz_get_idle_calls_cpu - return the current idle calls counter value * for a particular CPU. * @cpu: target CPU number * * Called from the schedutil frequency scaling governor in scheduler context. * * Return: the current idle calls counter value for @cpu */ unsigned long tick_nohz_get_idle_calls_cpu(int cpu) { … } static void tick_nohz_account_idle_time(struct tick_sched *ts, ktime_t now) { … } void tick_nohz_idle_restart_tick(void) { … } static void tick_nohz_idle_update_tick(struct tick_sched *ts, ktime_t now) { … } /** * tick_nohz_idle_exit - Update the tick upon idle task exit * * When the idle task exits, update the tick depending on the * following situations: * * 1) If the CPU is not in nohz_full mode (most cases), then * restart the tick. * * 2) If the CPU is in nohz_full mode (corner case): * 2.1) If the tick can be kept stopped (no tick dependencies) * then re-evaluate the next tick and try to keep it stopped * as long as possible. * 2.2) If the tick has dependencies, restart the tick. * */ void tick_nohz_idle_exit(void) { … } /* * In low-resolution mode, the tick handler must be implemented directly * at the clockevent level. hrtimer can't be used instead, because its * infrastructure actually relies on the tick itself as a backend in * low-resolution mode (see hrtimer_run_queues()). */ static void tick_nohz_lowres_handler(struct clock_event_device *dev) { … } static inline void tick_nohz_activate(struct tick_sched *ts) { … } /** * tick_nohz_switch_to_nohz - switch to NOHZ mode */ static void tick_nohz_switch_to_nohz(void) { … } static inline void tick_nohz_irq_enter(void) { … } #else static inline void tick_nohz_switch_to_nohz(void) { } static inline void tick_nohz_irq_enter(void) { } static inline void tick_nohz_activate(struct tick_sched *ts) { } #endif /* CONFIG_NO_HZ_COMMON */ /* * Called from irq_enter() to notify about the possible interruption of idle() */ void tick_irq_enter(void) { … } static int sched_skew_tick; static int __init skew_tick(char *str) { … } early_param(…); /** * tick_setup_sched_timer - setup the tick emulation timer * @hrtimer: whether to use the hrtimer or not */ void tick_setup_sched_timer(bool hrtimer) { … } /* * Shut down the tick and make sure the CPU won't try to retake the timekeeping * duty before disabling IRQs in idle for the last time. */ void tick_sched_timer_dying(int cpu) { … } /* * Async notification about clocksource changes */ void tick_clock_notify(void) { … } /* * Async notification about clock event changes */ void tick_oneshot_notify(void) { … } /* * Check if a change happened, which makes oneshot possible. * * Called cyclically from the hrtimer softirq (driven by the timer * softirq). 'allow_nohz' signals that we can switch into low-res NOHZ * mode, because high resolution timers are disabled (either compile * or runtime). Called with interrupts disabled. */ int tick_check_oneshot_change(int allow_nohz) { … }
Codebase Browser
linux