From 3810631332465d967ba5e27ea2c7dff2c9afac6c Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Thu, 12 Feb 2015 23:33:15 +0100 Subject: PM / sleep: Re-implement suspend-to-idle handling In preparation for adding support for quiescing timers in the final stage of suspend-to-idle transitions, rework the freeze_enter() function making the system wait on a wakeup event, the freeze_wake() function terminating the suspend-to-idle loop and the mechanism by which deep idle states are entered during suspend-to-idle. First of all, introduce a simple state machine for suspend-to-idle and make the code in question use it. Second, prevent freeze_enter() from losing wakeup events due to race conditions and ensure that the number of online CPUs won't change while it is being executed. In addition to that, make it force all of the CPUs re-enter the idle loop in case they are in idle states already (so they can enter deeper idle states if possible). Next, drop cpuidle_use_deepest_state() and replace use_deepest_state checks in cpuidle_select() and cpuidle_reflect() with a single suspend-to-idle state check in cpuidle_idle_call(). Finally, introduce cpuidle_enter_freeze() that will simply find the deepest idle state available to the given CPU and enter it using cpuidle_enter(). Signed-off-by: Rafael J. Wysocki Acked-by: Peter Zijlstra (Intel) --- kernel/power/suspend.c | 43 ++++++++++++++++++++++++++++++++++++------- kernel/sched/idle.c | 16 ++++++++++++++++ 2 files changed, 52 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/power/suspend.c b/kernel/power/suspend.c index c347e3ce3a55..b7d6b3a721b1 100644 --- a/kernel/power/suspend.c +++ b/kernel/power/suspend.c @@ -37,7 +37,9 @@ const char *pm_states[PM_SUSPEND_MAX]; static const struct platform_suspend_ops *suspend_ops; static const struct platform_freeze_ops *freeze_ops; static DECLARE_WAIT_QUEUE_HEAD(suspend_freeze_wait_head); -static bool suspend_freeze_wake; + +enum freeze_state __read_mostly suspend_freeze_state; +static DEFINE_SPINLOCK(suspend_freeze_lock); void freeze_set_ops(const struct platform_freeze_ops *ops) { @@ -48,22 +50,49 @@ void freeze_set_ops(const struct platform_freeze_ops *ops) static void freeze_begin(void) { - suspend_freeze_wake = false; + suspend_freeze_state = FREEZE_STATE_NONE; } static void freeze_enter(void) { - cpuidle_use_deepest_state(true); + spin_lock_irq(&suspend_freeze_lock); + if (pm_wakeup_pending()) + goto out; + + suspend_freeze_state = FREEZE_STATE_ENTER; + spin_unlock_irq(&suspend_freeze_lock); + + get_online_cpus(); cpuidle_resume(); - wait_event(suspend_freeze_wait_head, suspend_freeze_wake); + + /* Push all the CPUs into the idle loop. */ + wake_up_all_idle_cpus(); + pr_debug("PM: suspend-to-idle\n"); + /* Make the current CPU wait so it can enter the idle loop too. */ + wait_event(suspend_freeze_wait_head, + suspend_freeze_state == FREEZE_STATE_WAKE); + pr_debug("PM: resume from suspend-to-idle\n"); + cpuidle_pause(); - cpuidle_use_deepest_state(false); + put_online_cpus(); + + spin_lock_irq(&suspend_freeze_lock); + + out: + suspend_freeze_state = FREEZE_STATE_NONE; + spin_unlock_irq(&suspend_freeze_lock); } void freeze_wake(void) { - suspend_freeze_wake = true; - wake_up(&suspend_freeze_wait_head); + unsigned long flags; + + spin_lock_irqsave(&suspend_freeze_lock, flags); + if (suspend_freeze_state > FREEZE_STATE_NONE) { + suspend_freeze_state = FREEZE_STATE_WAKE; + wake_up(&suspend_freeze_wait_head); + } + spin_unlock_irqrestore(&suspend_freeze_lock, flags); } EXPORT_SYMBOL_GPL(freeze_wake); diff --git a/kernel/sched/idle.c b/kernel/sched/idle.c index aaf1c1d5cf5d..94b2d7b88a27 100644 --- a/kernel/sched/idle.c +++ b/kernel/sched/idle.c @@ -7,6 +7,7 @@ #include #include #include +#include #include @@ -104,6 +105,21 @@ static void cpuidle_idle_call(void) */ rcu_idle_enter(); + /* + * Suspend-to-idle ("freeze") is a system state in which all user space + * has been frozen, all I/O devices have been suspended and the only + * activity happens here and in iterrupts (if any). In that case bypass + * the cpuidle governor and go stratight for the deepest idle state + * available. Possibly also suspend the local tick and the entire + * timekeeping to prevent timer interrupts from kicking us out of idle + * until a proper wakeup interrupt happens. + */ + if (idle_should_freeze()) { + cpuidle_enter_freeze(); + local_irq_enable(); + goto exit_idle; + } + /* * Ask the cpuidle framework to choose a convenient idle state. * Fall back to the default arch idle method on errors. -- cgit v1.2.3 From affe3e85ae78507cc953f3f700e0644e50844cff Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Wed, 11 Feb 2015 05:01:52 +0100 Subject: timekeeping: Pass readout base to update_fast_timekeeper() Modify update_fast_timekeeper() to take a struct tk_read_base pointer as its argument (instead of a struct timekeeper pointer) and update its kerneldoc comment to reflect that. That will allow a struct tk_read_base that is not part of a struct timekeeper to be passed to it in the next patch. Signed-off-by: Rafael J. Wysocki Acked-by: Peter Zijlstra (Intel) Acked-by: John Stultz --- kernel/time/timekeeping.c | 14 ++++++-------- 1 file changed, 6 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c index b124af259800..abf08f4366c1 100644 --- a/kernel/time/timekeeping.c +++ b/kernel/time/timekeeping.c @@ -230,9 +230,7 @@ static inline s64 timekeeping_get_ns_raw(struct timekeeper *tk) /** * update_fast_timekeeper - Update the fast and NMI safe monotonic timekeeper. - * @tk: The timekeeper from which we take the update - * @tkf: The fast timekeeper to update - * @tbase: The time base for the fast timekeeper (mono/raw) + * @tkr: Timekeeping readout base from which we take the update * * We want to use this from any context including NMI and tracing / * instrumenting the timekeeping code itself. @@ -244,11 +242,11 @@ static inline s64 timekeeping_get_ns_raw(struct timekeeper *tk) * smp_wmb(); <- Ensure that the last base[1] update is visible * tkf->seq++; * smp_wmb(); <- Ensure that the seqcount update is visible - * update(tkf->base[0], tk); + * update(tkf->base[0], tkr); * smp_wmb(); <- Ensure that the base[0] update is visible * tkf->seq++; * smp_wmb(); <- Ensure that the seqcount update is visible - * update(tkf->base[1], tk); + * update(tkf->base[1], tkr); * * The reader side does: * @@ -269,7 +267,7 @@ static inline s64 timekeeping_get_ns_raw(struct timekeeper *tk) * slightly wrong timestamp (a few nanoseconds). See * @ktime_get_mono_fast_ns. */ -static void update_fast_timekeeper(struct timekeeper *tk) +static void update_fast_timekeeper(struct tk_read_base *tkr) { struct tk_read_base *base = tk_fast_mono.base; @@ -277,7 +275,7 @@ static void update_fast_timekeeper(struct timekeeper *tk) raw_write_seqcount_latch(&tk_fast_mono.seq); /* Update base[0] */ - memcpy(base, &tk->tkr, sizeof(*base)); + memcpy(base, tkr, sizeof(*base)); /* Force readers back to base[0] */ raw_write_seqcount_latch(&tk_fast_mono.seq); @@ -462,7 +460,7 @@ static void timekeeping_update(struct timekeeper *tk, unsigned int action) memcpy(&shadow_timekeeper, &tk_core.timekeeper, sizeof(tk_core.timekeeper)); - update_fast_timekeeper(tk); + update_fast_timekeeper(&tk->tkr); } /** -- cgit v1.2.3 From 060407aed56c00960c9b5f70f5d19b2823adffd7 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Fri, 13 Feb 2015 14:49:02 +0100 Subject: timekeeping: Make it safe to use the fast timekeeper while suspended Theoretically, ktime_get_mono_fast_ns() may be executed after timekeeping has been suspended (or before it is resumed) which in turn may lead to undefined behavior, for example, when the clocksource read from timekeeping_get_ns() called by it is not accessible at that time. Prevent that from happening by setting up a dummy readout base for the fast timekeeper during timekeeping_suspend() such that it will always return the same number of cycles. After the last timekeeping_update() in timekeeping_suspend() the clocksource is read and the result is stored as cycles_at_suspend. The readout base from the current timekeeper is copied onto the dummy and the ->read pointer of the dummy is set to a routine unconditionally returning cycles_at_suspend. Next, the dummy is passed to update_fast_timekeeper(). Then, ktime_get_mono_fast_ns() will work until the subsequent timekeeping_resume() and the proper readout base for the fast timekeeper will be restored by the timekeeping_update() called right after clearing timekeeping_suspended. Signed-off-by: Rafael J. Wysocki Acked-by: John Stultz Acked-by: Peter Zijlstra (Intel) --- kernel/time/timekeeping.c | 30 ++++++++++++++++++++++++++++++ 1 file changed, 30 insertions(+) (limited to 'kernel') diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c index abf08f4366c1..aef5dc722abf 100644 --- a/kernel/time/timekeeping.c +++ b/kernel/time/timekeeping.c @@ -332,6 +332,35 @@ u64 notrace ktime_get_mono_fast_ns(void) } EXPORT_SYMBOL_GPL(ktime_get_mono_fast_ns); +/* Suspend-time cycles value for halted fast timekeeper. */ +static cycle_t cycles_at_suspend; + +static cycle_t dummy_clock_read(struct clocksource *cs) +{ + return cycles_at_suspend; +} + +/** + * halt_fast_timekeeper - Prevent fast timekeeper from accessing clocksource. + * @tk: Timekeeper to snapshot. + * + * It generally is unsafe to access the clocksource after timekeeping has been + * suspended, so take a snapshot of the readout base of @tk and use it as the + * fast timekeeper's readout base while suspended. It will return the same + * number of cycles every time until timekeeping is resumed at which time the + * proper readout base for the fast timekeeper will be restored automatically. + */ +static void halt_fast_timekeeper(struct timekeeper *tk) +{ + static struct tk_read_base tkr_dummy; + struct tk_read_base *tkr = &tk->tkr; + + memcpy(&tkr_dummy, tkr, sizeof(tkr_dummy)); + cycles_at_suspend = tkr->read(tkr->clock); + tkr_dummy.read = dummy_clock_read; + update_fast_timekeeper(&tkr_dummy); +} + #ifdef CONFIG_GENERIC_TIME_VSYSCALL_OLD static inline void update_vsyscall(struct timekeeper *tk) @@ -1294,6 +1323,7 @@ static int timekeeping_suspend(void) } timekeeping_update(tk, TK_MIRROR); + halt_fast_timekeeper(tk); write_seqcount_end(&tk_core.seq); raw_spin_unlock_irqrestore(&timekeeper_lock, flags); -- cgit v1.2.3 From 124cf9117c5f93cc5b324530b7e105b09c729d5d Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Fri, 13 Feb 2015 23:50:43 +0100 Subject: PM / sleep: Make it possible to quiesce timers during suspend-to-idle The efficiency of suspend-to-idle depends on being able to keep CPUs in the deepest available idle states for as much time as possible. Ideally, they should only be brought out of idle by system wakeup interrupts. However, timer interrupts occurring periodically prevent that from happening and it is not practical to chase all of the "misbehaving" timers in a whack-a-mole fashion. A much more effective approach is to suspend the local ticks for all CPUs and the entire timekeeping along the lines of what is done during full suspend, which also helps to keep suspend-to-idle and full suspend reasonably similar. The idea is to suspend the local tick on each CPU executing cpuidle_enter_freeze() and to make the last of them suspend the entire timekeeping. That should prevent timer interrupts from triggering until an IO interrupt wakes up one of the CPUs. It needs to be done with interrupts disabled on all of the CPUs, though, because otherwise the suspended clocksource might be accessed by an interrupt handler which might lead to fatal consequences. Unfortunately, the existing ->enter callbacks provided by cpuidle drivers generally cannot be used for implementing that, because some of them re-enable interrupts temporarily and some idle entry methods cause interrupts to be re-enabled automatically on exit. Also some of these callbacks manipulate local clock event devices of the CPUs which really shouldn't be done after suspending their ticks. To overcome that difficulty, introduce a new cpuidle state callback, ->enter_freeze, that will be guaranteed (1) to keep interrupts disabled all the time (and return with interrupts disabled) and (2) not to touch the CPU timer devices. Modify cpuidle_enter_freeze() to look for the deepest available idle state with ->enter_freeze present and to make the CPU execute that callback with suspended tick (and the last of the online CPUs to execute it with suspended timekeeping). Suggested-by: Thomas Gleixner Signed-off-by: Rafael J. Wysocki Acked-by: Peter Zijlstra (Intel) --- drivers/cpuidle/cpuidle.c | 49 +++++++++++++++++++++++++++++++++++++++++----- include/linux/cpuidle.h | 9 +++++++++ include/linux/tick.h | 6 +++++- kernel/time/tick-common.c | 50 +++++++++++++++++++++++++++++++++++++++++++++++ kernel/time/timekeeping.c | 4 ++-- kernel/time/timekeeping.h | 2 ++ 6 files changed, 112 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/drivers/cpuidle/cpuidle.c b/drivers/cpuidle/cpuidle.c index 23a8d6cc8d30..4d534582514e 100644 --- a/drivers/cpuidle/cpuidle.c +++ b/drivers/cpuidle/cpuidle.c @@ -20,6 +20,7 @@ #include #include #include +#include #include #include "cpuidle.h" @@ -69,18 +70,20 @@ int cpuidle_play_dead(void) * cpuidle_find_deepest_state - Find deepest state meeting specific conditions. * @drv: cpuidle driver for the given CPU. * @dev: cpuidle device for the given CPU. + * @freeze: Whether or not the state should be suitable for suspend-to-idle. */ static int cpuidle_find_deepest_state(struct cpuidle_driver *drv, - struct cpuidle_device *dev) + struct cpuidle_device *dev, bool freeze) { unsigned int latency_req = 0; - int i, ret = CPUIDLE_DRIVER_STATE_START - 1; + int i, ret = freeze ? -1 : CPUIDLE_DRIVER_STATE_START - 1; for (i = CPUIDLE_DRIVER_STATE_START; i < drv->state_count; i++) { struct cpuidle_state *s = &drv->states[i]; struct cpuidle_state_usage *su = &dev->states_usage[i]; - if (s->disabled || su->disable || s->exit_latency <= latency_req) + if (s->disabled || su->disable || s->exit_latency <= latency_req + || (freeze && !s->enter_freeze)) continue; latency_req = s->exit_latency; @@ -89,10 +92,31 @@ static int cpuidle_find_deepest_state(struct cpuidle_driver *drv, return ret; } +static void enter_freeze_proper(struct cpuidle_driver *drv, + struct cpuidle_device *dev, int index) +{ + tick_freeze(); + /* + * The state used here cannot be a "coupled" one, because the "coupled" + * cpuidle mechanism enables interrupts and doing that with timekeeping + * suspended is generally unsafe. + */ + drv->states[index].enter_freeze(dev, drv, index); + WARN_ON(!irqs_disabled()); + /* + * timekeeping_resume() that will be called by tick_unfreeze() for the + * last CPU executing it calls functions containing RCU read-side + * critical sections, so tell RCU about that. + */ + RCU_NONIDLE(tick_unfreeze()); +} + /** * cpuidle_enter_freeze - Enter an idle state suitable for suspend-to-idle. * - * Find the deepest state available and enter it. + * If there are states with the ->enter_freeze callback, find the deepest of + * them and enter it with frozen tick. Otherwise, find the deepest state + * available and enter it normally. */ void cpuidle_enter_freeze(void) { @@ -100,7 +124,22 @@ void cpuidle_enter_freeze(void) struct cpuidle_driver *drv = cpuidle_get_cpu_driver(dev); int index; - index = cpuidle_find_deepest_state(drv, dev); + /* + * Find the deepest state with ->enter_freeze present, which guarantees + * that interrupts won't be enabled when it exits and allows the tick to + * be frozen safely. + */ + index = cpuidle_find_deepest_state(drv, dev, true); + if (index >= 0) { + enter_freeze_proper(drv, dev, index); + return; + } + + /* + * It is not safe to freeze the tick, find the deepest state available + * at all and try to enter it normally. + */ + index = cpuidle_find_deepest_state(drv, dev, false); if (index >= 0) cpuidle_enter(drv, dev, index); else diff --git a/include/linux/cpuidle.h b/include/linux/cpuidle.h index f63aabf4ee90..f551a9299ac9 100644 --- a/include/linux/cpuidle.h +++ b/include/linux/cpuidle.h @@ -50,6 +50,15 @@ struct cpuidle_state { int index); int (*enter_dead) (struct cpuidle_device *dev, int index); + + /* + * CPUs execute ->enter_freeze with the local tick or entire timekeeping + * suspended, so it must not re-enable interrupts at any point (even + * temporarily) or attempt to change states of clock event devices. + */ + void (*enter_freeze) (struct cpuidle_device *dev, + struct cpuidle_driver *drv, + int index); }; /* Idle State Flags */ diff --git a/include/linux/tick.h b/include/linux/tick.h index eda850ca757a..9c085dc12ae9 100644 --- a/include/linux/tick.h +++ b/include/linux/tick.h @@ -79,6 +79,9 @@ extern void __init tick_init(void); extern int tick_is_oneshot_available(void); extern struct tick_device *tick_get_device(int cpu); +extern void tick_freeze(void); +extern void tick_unfreeze(void); + # ifdef CONFIG_HIGH_RES_TIMERS extern int tick_init_highres(void); extern int tick_program_event(ktime_t expires, int force); @@ -119,6 +122,8 @@ static inline int tick_oneshot_mode_active(void) { return 0; } #else /* CONFIG_GENERIC_CLOCKEVENTS */ static inline void tick_init(void) { } +static inline void tick_freeze(void) { } +static inline void tick_unfreeze(void) { } static inline void tick_cancel_sched_timer(int cpu) { } static inline void tick_clock_notify(void) { } static inline int tick_check_oneshot_change(int allow_nohz) { return 0; } @@ -226,5 +231,4 @@ static inline void tick_nohz_task_switch(struct task_struct *tsk) __tick_nohz_task_switch(tsk); } - #endif diff --git a/kernel/time/tick-common.c b/kernel/time/tick-common.c index 7efeedf53ebd..f7c515595b42 100644 --- a/kernel/time/tick-common.c +++ b/kernel/time/tick-common.c @@ -394,6 +394,56 @@ void tick_resume(void) } } +static DEFINE_RAW_SPINLOCK(tick_freeze_lock); +static unsigned int tick_freeze_depth; + +/** + * tick_freeze - Suspend the local tick and (possibly) timekeeping. + * + * Check if this is the last online CPU executing the function and if so, + * suspend timekeeping. Otherwise suspend the local tick. + * + * Call with interrupts disabled. Must be balanced with %tick_unfreeze(). + * Interrupts must not be enabled before the subsequent %tick_unfreeze(). + */ +void tick_freeze(void) +{ + raw_spin_lock(&tick_freeze_lock); + + tick_freeze_depth++; + if (tick_freeze_depth == num_online_cpus()) { + timekeeping_suspend(); + } else { + tick_suspend(); + tick_suspend_broadcast(); + } + + raw_spin_unlock(&tick_freeze_lock); +} + +/** + * tick_unfreeze - Resume the local tick and (possibly) timekeeping. + * + * Check if this is the first CPU executing the function and if so, resume + * timekeeping. Otherwise resume the local tick. + * + * Call with interrupts disabled. Must be balanced with %tick_freeze(). + * Interrupts must not be enabled after the preceding %tick_freeze(). + */ +void tick_unfreeze(void) +{ + raw_spin_lock(&tick_freeze_lock); + + if (tick_freeze_depth == num_online_cpus()) + timekeeping_resume(); + else + tick_resume(); + + tick_freeze_depth--; + + raw_spin_unlock(&tick_freeze_lock); +} + /** * tick_init - initialize the tick control */ diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c index aef5dc722abf..91db94136c10 100644 --- a/kernel/time/timekeeping.c +++ b/kernel/time/timekeeping.c @@ -1197,7 +1197,7 @@ void timekeeping_inject_sleeptime64(struct timespec64 *delta) * xtime/wall_to_monotonic/jiffies/etc are * still managed by arch specific suspend/resume code. */ -static void timekeeping_resume(void) +void timekeeping_resume(void) { struct timekeeper *tk = &tk_core.timekeeper; struct clocksource *clock = tk->tkr.clock; @@ -1278,7 +1278,7 @@ static void timekeeping_resume(void) hrtimers_resume(); } -static int timekeeping_suspend(void) +int timekeeping_suspend(void) { struct timekeeper *tk = &tk_core.timekeeper; unsigned long flags; diff --git a/kernel/time/timekeeping.h b/kernel/time/timekeeping.h index adc1fc98bde3..1d91416055d5 100644 --- a/kernel/time/timekeeping.h +++ b/kernel/time/timekeeping.h @@ -16,5 +16,7 @@ extern int timekeeping_inject_offset(struct timespec *ts); extern s32 timekeeping_get_tai_offset(void); extern void timekeeping_set_tai_offset(s32 tai_offset); extern void timekeeping_clocktai(struct timespec *ts); +extern int timekeeping_suspend(void); +extern void timekeeping_resume(void); #endif -- cgit v1.2.3