// SPDX-License-Identifier: GPL-2.0/* * Copyright(C) 2005-2006, Linutronix GmbH, Thomas Gleixner <tglx@kernel.org> * 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){ return &per_cpu(tick_cpu_sched, 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){ unsigned long ticks = 1; ktime_t delta, nextp; /* * 64-bit can do a quick check without holding the jiffies lock and * without looking at the sequence count. The smp_load_acquire() * pairs with the update done later in this function. * * 32-bit cannot do that because the store of 'tick_next_period' * consists of two 32-bit stores, and the first store could be * moved by the CPU to a random point in the future. */ if (IS_ENABLED(CONFIG_64BIT)) { if (ktime_before(now, smp_load_acquire(&tick_next_period))) return; } else { unsigned int seq; /* * Avoid contention on 'jiffies_lock' and protect the quick * check with the sequence count. */ do { seq = read_seqcount_begin(&jiffies_seq); nextp = tick_next_period; } while (read_seqcount_retry(&jiffies_seq, seq)); if (ktime_before(now, nextp)) return; } /* Quick check failed, i.e. update is required. */ raw_spin_lock(&jiffies_lock); /* * Re-evaluate with the lock held. Another CPU might have done the * update already. */ if (ktime_before(now, tick_next_period)) { raw_spin_unlock(&jiffies_lock); return; } write_seqcount_begin(&jiffies_seq); delta = ktime_sub(now, tick_next_period); if (unlikely(delta >= TICK_NSEC)) { /* Slow path for long idle sleep times */ s64 incr = TICK_NSEC; ticks += ktime_divns(delta, incr); last_jiffies_update = ktime_add_ns(last_jiffies_update, incr * ticks); } else { last_jiffies_update = ktime_add_ns(last_jiffies_update, TICK_NSEC); } /* Advance jiffies to complete the 'jiffies_seq' protected job */ jiffies_64 += ticks; /* Keep the tick_next_period variable up to date */ nextp = ktime_add_ns(last_jiffies_update, TICK_NSEC); if (IS_ENABLED(CONFIG_64BIT)) { /* * Pairs with smp_load_acquire() in the lockless quick * check above, and ensures that the update to 'jiffies_64' is * not reordered vs. the store to 'tick_next_period', neither * by the compiler nor by the CPU. */ smp_store_release(&tick_next_period, nextp); } else { /* * A plain store is good enough on 32-bit, as the quick check * above is protected by the sequence count. */ tick_next_period = nextp; } /* * Release the sequence count. calc_global_load() below is not * protected by it, but 'jiffies_lock' needs to be held to prevent * concurrent invocations. */ write_seqcount_end(&jiffies_seq); calc_global_load(); raw_spin_unlock(&jiffies_lock); update_wall_time();} /* * Initialize and return retrieve the jiffies update. */static ktime_t tick_init_jiffy_update(void){ ktime_t period; raw_spin_lock(&jiffies_lock); write_seqcount_begin(&jiffies_seq); /* Have we started the jiffies update yet ? */ if (last_jiffies_update == 0) { u32 rem; /* * Ensure that the tick is aligned to a multiple of * TICK_NSEC. */ div_u64_rem(tick_next_period, TICK_NSEC, &rem); if (rem) tick_next_period += TICK_NSEC - rem; last_jiffies_update = tick_next_period; } period = last_jiffies_update; write_seqcount_end(&jiffies_seq); raw_spin_unlock(&jiffies_lock); return period;} static inline int tick_sched_flag_test(struct tick_sched *ts, unsigned long flag){ return !!(ts->flags & flag);} static inline void tick_sched_flag_set(struct tick_sched *ts, unsigned long flag){ lockdep_assert_irqs_disabled(); ts->flags |= flag;} static inline void tick_sched_flag_clear(struct tick_sched *ts, unsigned long flag){ lockdep_assert_irqs_disabled(); ts->flags &= ~flag;} /* * Allow only one non-timekeeper CPU at a time update jiffies from * the timer tick. * * Returns true if update was run. */static bool tick_limited_update_jiffies64(struct tick_sched *ts, ktime_t now){ static atomic_t in_progress; int inp; inp = atomic_read(&in_progress); if (inp || !atomic_try_cmpxchg(&in_progress, &inp, 1)) return false; if (ts->last_tick_jiffies == jiffies) tick_do_update_jiffies64(now); atomic_set(&in_progress, 0); return true;} #define MAX_STALLED_JIFFIES 5 static void tick_sched_do_timer(struct tick_sched *ts, ktime_t now){ int tick_cpu, cpu = smp_processor_id(); /* * 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 themselves to * this duty, then the jiffies update is still serialized by * 'jiffies_lock'. * * If nohz_full is enabled, this should not happen because the * 'tick_do_timer_cpu' CPU never relinquishes. */ tick_cpu = READ_ONCE(tick_do_timer_cpu); if (IS_ENABLED(CONFIG_NO_HZ_COMMON) && unlikely(tick_cpu == TICK_DO_TIMER_NONE)) {#ifdef CONFIG_NO_HZ_FULL WARN_ON_ONCE(tick_nohz_full_running);#endif WRITE_ONCE(tick_do_timer_cpu, cpu); tick_cpu = cpu; } /* Check if jiffies need an update */ if (tick_cpu == cpu) tick_do_update_jiffies64(now); /* * If the jiffies update stalled for too long (timekeeper in stop_machine() * or VMEXIT'ed for several msecs), force an update. */ if (ts->last_tick_jiffies != jiffies) { ts->stalled_jiffies = 0; ts->last_tick_jiffies = READ_ONCE(jiffies); } else { if (++ts->stalled_jiffies >= MAX_STALLED_JIFFIES) { if (tick_limited_update_jiffies64(ts, now)) { ts->stalled_jiffies = 0; ts->last_tick_jiffies = READ_ONCE(jiffies); } } } if (tick_sched_flag_test(ts, TS_FLAG_INIDLE)) ts->got_idle_tick = 1;} static void tick_sched_handle(struct tick_sched *ts, struct pt_regs *regs){ /* * When we are idle and the tick is stopped, we have to touch * the watchdog as we might not schedule for a really long * time. This happens on completely idle SMP systems while * waiting on the login prompt. We also increment the "start of * idle" jiffy stamp so the idle accounting adjustment we do * when we go busy again does not account too many ticks. */ if (IS_ENABLED(CONFIG_NO_HZ_COMMON) && tick_sched_flag_test(ts, TS_FLAG_STOPPED)) { touch_softlockup_watchdog_sched(); /* * In case the current tick fired too early past its expected * expiration, make sure we don't bypass the next clock reprogramming * to the same deadline. */ ts->next_tick = 0; } update_process_times(user_mode(regs)); profile_tick(CPU_PROFILING);} /* * 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){ struct tick_sched *ts = container_of(timer, struct tick_sched, sched_timer); struct pt_regs *regs = get_irq_regs(); ktime_t now = ktime_get(); tick_sched_do_timer(ts, now); /* * Do not call when we are not in IRQ context and have * no valid 'regs' pointer */ if (regs) tick_sched_handle(ts, regs); else ts->next_tick = 0; /* * In dynticks mode, tick reprogram is deferred: * - to the idle task if in dynticks-idle * - to IRQ exit if in full-dynticks. */ if (unlikely(tick_sched_flag_test(ts, TS_FLAG_STOPPED))) return HRTIMER_NORESTART; hrtimer_forward(timer, now, TICK_NSEC); return HRTIMER_RESTART;} #ifdef CONFIG_NO_HZ_FULLcpumask_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 (likely(!tracepoint_enabled(tick_stop))) return !!val; 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 * * XXX given a task picks up the dependency on schedule(), should we * only care about tasks that are currently on the CPU instead of all * that are on the runqueue? * * That is, does this want to be: task_on_cpu() / task_curr()? */ 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,Lines 1–500 of 1590. Load more
// SPDX-License-Identifier: GPL-2.0 /* * Copyright(C) 2005-2006, Linutronix GmbH, Thomas Gleixner <tglx@kernel.org> * 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) { return &per_cpu(tick_cpu_sched, 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) { unsigned long ticks = 1; ktime_t delta, nextp; /* * 64-bit can do a quick check without holding the jiffies lock and * without looking at the sequence count. The smp_load_acquire() * pairs with the update done later in this function. * * 32-bit cannot do that because the store of 'tick_next_period' * consists of two 32-bit stores, and the first store could be * moved by the CPU to a random point in the future. */ if (IS_ENABLED(CONFIG_64BIT)) { if (ktime_before(now, smp_load_acquire(&tick_next_period))) return; } else { unsigned int seq; /* * Avoid contention on 'jiffies_lock' and protect the quick * check with the sequence count. */ do { seq = read_seqcount_begin(&jiffies_seq); nextp = tick_next_period; } while (read_seqcount_retry(&jiffies_seq, seq)); if (ktime_before(now, nextp)) return; } /* Quick check failed, i.e. update is required. */ raw_spin_lock(&jiffies_lock); /* * Re-evaluate with the lock held. Another CPU might have done the * update already. */ if (ktime_before(now, tick_next_period)) { raw_spin_unlock(&jiffies_lock); return; } write_seqcount_begin(&jiffies_seq); delta = ktime_sub(now, tick_next_period); if (unlikely(delta >= TICK_NSEC)) { /* Slow path for long idle sleep times */ s64 incr = TICK_NSEC; ticks += ktime_divns(delta, incr); last_jiffies_update = ktime_add_ns(last_jiffies_update, incr * ticks); } else { last_jiffies_update = ktime_add_ns(last_jiffies_update, TICK_NSEC); } /* Advance jiffies to complete the 'jiffies_seq' protected job */ jiffies_64 += ticks; /* Keep the tick_next_period variable up to date */ nextp = ktime_add_ns(last_jiffies_update, TICK_NSEC); if (IS_ENABLED(CONFIG_64BIT)) { /* * Pairs with smp_load_acquire() in the lockless quick * check above, and ensures that the update to 'jiffies_64' is * not reordered vs. the store to 'tick_next_period', neither * by the compiler nor by the CPU. */ smp_store_release(&tick_next_period, nextp); } else { /* * A plain store is good enough on 32-bit, as the quick check * above is protected by the sequence count. */ tick_next_period = nextp; } /* * Release the sequence count. calc_global_load() below is not * protected by it, but 'jiffies_lock' needs to be held to prevent * concurrent invocations. */ write_seqcount_end(&jiffies_seq); calc_global_load(); raw_spin_unlock(&jiffies_lock); update_wall_time(); } /* * Initialize and return retrieve the jiffies update. */ static ktime_t tick_init_jiffy_update(void) { ktime_t period; raw_spin_lock(&jiffies_lock); write_seqcount_begin(&jiffies_seq); /* Have we started the jiffies update yet ? */ if (last_jiffies_update == 0) { u32 rem; /* * Ensure that the tick is aligned to a multiple of * TICK_NSEC. */ div_u64_rem(tick_next_period, TICK_NSEC, &rem); if (rem) tick_next_period += TICK_NSEC - rem; last_jiffies_update = tick_next_period; } period = last_jiffies_update; write_seqcount_end(&jiffies_seq); raw_spin_unlock(&jiffies_lock); return period; } static inline int tick_sched_flag_test(struct tick_sched *ts, unsigned long flag) { return !!(ts->flags & flag); } static inline void tick_sched_flag_set(struct tick_sched *ts, unsigned long flag) { lockdep_assert_irqs_disabled(); ts->flags |= flag; } static inline void tick_sched_flag_clear(struct tick_sched *ts, unsigned long flag) { lockdep_assert_irqs_disabled(); ts->flags &= ~flag; } /* * Allow only one non-timekeeper CPU at a time update jiffies from * the timer tick. * * Returns true if update was run. */ static bool tick_limited_update_jiffies64(struct tick_sched *ts, ktime_t now) { static atomic_t in_progress; int inp; inp = atomic_read(&in_progress); if (inp || !atomic_try_cmpxchg(&in_progress, &inp, 1)) return false; if (ts->last_tick_jiffies == jiffies) tick_do_update_jiffies64(now); atomic_set(&in_progress, 0); return true; } #define MAX_STALLED_JIFFIES 5 static void tick_sched_do_timer(struct tick_sched *ts, ktime_t now) { int tick_cpu, cpu = smp_processor_id(); /* * 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 themselves to * this duty, then the jiffies update is still serialized by * 'jiffies_lock'. * * If nohz_full is enabled, this should not happen because the * 'tick_do_timer_cpu' CPU never relinquishes. */ tick_cpu = READ_ONCE(tick_do_timer_cpu); if (IS_ENABLED(CONFIG_NO_HZ_COMMON) && unlikely(tick_cpu == TICK_DO_TIMER_NONE)) { #ifdef CONFIG_NO_HZ_FULL WARN_ON_ONCE(tick_nohz_full_running); #endif WRITE_ONCE(tick_do_timer_cpu, cpu); tick_cpu = cpu; } /* Check if jiffies need an update */ if (tick_cpu == cpu) tick_do_update_jiffies64(now); /* * If the jiffies update stalled for too long (timekeeper in stop_machine() * or VMEXIT'ed for several msecs), force an update. */ if (ts->last_tick_jiffies != jiffies) { ts->stalled_jiffies = 0; ts->last_tick_jiffies = READ_ONCE(jiffies); } else { if (++ts->stalled_jiffies >= MAX_STALLED_JIFFIES) { if (tick_limited_update_jiffies64(ts, now)) { ts->stalled_jiffies = 0; ts->last_tick_jiffies = READ_ONCE(jiffies); } } } if (tick_sched_flag_test(ts, TS_FLAG_INIDLE)) ts->got_idle_tick = 1; } static void tick_sched_handle(struct tick_sched *ts, struct pt_regs *regs) { /* * When we are idle and the tick is stopped, we have to touch * the watchdog as we might not schedule for a really long * time. This happens on completely idle SMP systems while * waiting on the login prompt. We also increment the "start of * idle" jiffy stamp so the idle accounting adjustment we do * when we go busy again does not account too many ticks. */ if (IS_ENABLED(CONFIG_NO_HZ_COMMON) && tick_sched_flag_test(ts, TS_FLAG_STOPPED)) { touch_softlockup_watchdog_sched(); /* * In case the current tick fired too early past its expected * expiration, make sure we don't bypass the next clock reprogramming * to the same deadline. */ ts->next_tick = 0; } update_process_times(user_mode(regs)); profile_tick(CPU_PROFILING); } /* * 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) { struct tick_sched *ts = container_of(timer, struct tick_sched, sched_timer); struct pt_regs *regs = get_irq_regs(); ktime_t now = ktime_get(); tick_sched_do_timer(ts, now); /* * Do not call when we are not in IRQ context and have * no valid 'regs' pointer */ if (regs) tick_sched_handle(ts, regs); else ts->next_tick = 0; /* * In dynticks mode, tick reprogram is deferred: * - to the idle task if in dynticks-idle * - to IRQ exit if in full-dynticks. */ if (unlikely(tick_sched_flag_test(ts, TS_FLAG_STOPPED))) return HRTIMER_NORESTART; hrtimer_forward(timer, now, TICK_NSEC); return HRTIMER_RESTART; } #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 (likely(!tracepoint_enabled(tick_stop))) return !!val; 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 * * XXX given a task picks up the dependency on schedule(), should we * only care about tasks that are currently on the CPU instead of all * that are on the runqueue? * * That is, does this want to be: task_on_cpu() / task_curr()? */ 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,