[PATCH] hrtimers: cleanup locking

Improve kernel/hrtimers.c locking: use a per-CPU base with a lock to control
locking of all clocks belonging to a CPU.  This simplifies code that needs to
lock all clocks at once.  This makes life easier for high-res timers and
dyntick.

No functional changes.

[ optimization change from Andrew Morton <akpm@osdl.org> ]

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: john stultz <johnstul@us.ibm.com>
Cc: Roman Zippel <zippel@linux-m68k.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This commit is contained in:
Thomas Gleixner 2007-02-16 01:27:50 -08:00 committed by Linus Torvalds
parent c9cb2e3d7c
commit 3c8aa39d7c
2 changed files with 126 additions and 101 deletions

View file

@ -21,6 +21,9 @@
#include <linux/list.h> #include <linux/list.h>
#include <linux/wait.h> #include <linux/wait.h>
struct hrtimer_clock_base;
struct hrtimer_cpu_base;
/* /*
* Mode arguments of xxx_hrtimer functions: * Mode arguments of xxx_hrtimer functions:
*/ */
@ -37,8 +40,6 @@ enum hrtimer_restart {
HRTIMER_RESTART, /* Timer must be restarted */ HRTIMER_RESTART, /* Timer must be restarted */
}; };
struct hrtimer_base;
/** /**
* struct hrtimer - the basic hrtimer structure * struct hrtimer - the basic hrtimer structure
* @node: red black tree node for time ordered insertion * @node: red black tree node for time ordered insertion
@ -51,10 +52,10 @@ struct hrtimer_base;
* The hrtimer structure must be initialized by init_hrtimer_#CLOCKTYPE() * The hrtimer structure must be initialized by init_hrtimer_#CLOCKTYPE()
*/ */
struct hrtimer { struct hrtimer {
struct rb_node node; struct rb_node node;
ktime_t expires; ktime_t expires;
enum hrtimer_restart (*function)(struct hrtimer *); enum hrtimer_restart (*function)(struct hrtimer *);
struct hrtimer_base *base; struct hrtimer_clock_base *base;
}; };
/** /**
@ -71,29 +72,41 @@ struct hrtimer_sleeper {
/** /**
* struct hrtimer_base - the timer base for a specific clock * struct hrtimer_base - the timer base for a specific clock
* @index: clock type index for per_cpu support when moving a timer * @index: clock type index for per_cpu support when moving a
* to a base on another cpu. * timer to a base on another cpu.
* @lock: lock protecting the base and associated timers
* @active: red black tree root node for the active timers * @active: red black tree root node for the active timers
* @first: pointer to the timer node which expires first * @first: pointer to the timer node which expires first
* @resolution: the resolution of the clock, in nanoseconds * @resolution: the resolution of the clock, in nanoseconds
* @get_time: function to retrieve the current time of the clock * @get_time: function to retrieve the current time of the clock
* @get_softirq_time: function to retrieve the current time from the softirq * @get_softirq_time: function to retrieve the current time from the softirq
* @curr_timer: the timer which is executing a callback right now
* @softirq_time: the time when running the hrtimer queue in the softirq * @softirq_time: the time when running the hrtimer queue in the softirq
* @lock_key: the lock_class_key for use with lockdep
*/ */
struct hrtimer_base { struct hrtimer_clock_base {
struct hrtimer_cpu_base *cpu_base;
clockid_t index; clockid_t index;
spinlock_t lock;
struct rb_root active; struct rb_root active;
struct rb_node *first; struct rb_node *first;
ktime_t resolution; ktime_t resolution;
ktime_t (*get_time)(void); ktime_t (*get_time)(void);
ktime_t (*get_softirq_time)(void); ktime_t (*get_softirq_time)(void);
struct hrtimer *curr_timer;
ktime_t softirq_time; ktime_t softirq_time;
struct lock_class_key lock_key; };
#define HRTIMER_MAX_CLOCK_BASES 2
/*
* struct hrtimer_cpu_base - the per cpu clock bases
* @lock: lock protecting the base and associated clock bases
* and timers
* @lock_key: the lock_class_key for use with lockdep
* @clock_base: array of clock bases for this cpu
* @curr_timer: the timer which is executing a callback right now
*/
struct hrtimer_cpu_base {
spinlock_t lock;
struct lock_class_key lock_key;
struct hrtimer_clock_base clock_base[HRTIMER_MAX_CLOCK_BASES];
struct hrtimer *curr_timer;
}; };
/* /*

View file

@ -1,8 +1,9 @@
/* /*
* linux/kernel/hrtimer.c * linux/kernel/hrtimer.c
* *
* Copyright(C) 2005, Thomas Gleixner <tglx@linutronix.de> * Copyright(C) 2005-2006, Thomas Gleixner <tglx@linutronix.de>
* Copyright(C) 2005, Red Hat, Inc., Ingo Molnar * Copyright(C) 2005-2006, Red Hat, Inc., Ingo Molnar
* Copyright(C) 2006 Timesys Corp., Thomas Gleixner <tglx@timesys.com>
* *
* High-resolution kernel timers * High-resolution kernel timers
* *
@ -79,21 +80,22 @@ EXPORT_SYMBOL_GPL(ktime_get_real);
* This ensures that we capture erroneous accesses to these clock ids * This ensures that we capture erroneous accesses to these clock ids
* rather than moving them into the range of valid clock id's. * rather than moving them into the range of valid clock id's.
*/ */
static DEFINE_PER_CPU(struct hrtimer_cpu_base, hrtimer_bases) =
#define MAX_HRTIMER_BASES 2
static DEFINE_PER_CPU(struct hrtimer_base, hrtimer_bases[MAX_HRTIMER_BASES]) =
{ {
.clock_base =
{ {
.index = CLOCK_REALTIME, {
.get_time = &ktime_get_real, .index = CLOCK_REALTIME,
.resolution = KTIME_REALTIME_RES, .get_time = &ktime_get_real,
}, .resolution = KTIME_REALTIME_RES,
{ },
.index = CLOCK_MONOTONIC, {
.get_time = &ktime_get, .index = CLOCK_MONOTONIC,
.resolution = KTIME_MONOTONIC_RES, .get_time = &ktime_get,
}, .resolution = KTIME_MONOTONIC_RES,
},
}
}; };
/** /**
@ -125,7 +127,7 @@ EXPORT_SYMBOL_GPL(ktime_get_ts);
* Get the coarse grained time at the softirq based on xtime and * Get the coarse grained time at the softirq based on xtime and
* wall_to_monotonic. * wall_to_monotonic.
*/ */
static void hrtimer_get_softirq_time(struct hrtimer_base *base) static void hrtimer_get_softirq_time(struct hrtimer_cpu_base *base)
{ {
ktime_t xtim, tomono; ktime_t xtim, tomono;
struct timespec xts; struct timespec xts;
@ -142,8 +144,9 @@ static void hrtimer_get_softirq_time(struct hrtimer_base *base)
xtim = timespec_to_ktime(xts); xtim = timespec_to_ktime(xts);
tomono = timespec_to_ktime(wall_to_monotonic); tomono = timespec_to_ktime(wall_to_monotonic);
base[CLOCK_REALTIME].softirq_time = xtim; base->clock_base[CLOCK_REALTIME].softirq_time = xtim;
base[CLOCK_MONOTONIC].softirq_time = ktime_add(xtim, tomono); base->clock_base[CLOCK_MONOTONIC].softirq_time =
ktime_add(xtim, tomono);
} }
/* /*
@ -166,19 +169,20 @@ static void hrtimer_get_softirq_time(struct hrtimer_base *base)
* possible to set timer->base = NULL and drop the lock: the timer remains * possible to set timer->base = NULL and drop the lock: the timer remains
* locked. * locked.
*/ */
static struct hrtimer_base *lock_hrtimer_base(const struct hrtimer *timer, static
unsigned long *flags) struct hrtimer_clock_base *lock_hrtimer_base(const struct hrtimer *timer,
unsigned long *flags)
{ {
struct hrtimer_base *base; struct hrtimer_clock_base *base;
for (;;) { for (;;) {
base = timer->base; base = timer->base;
if (likely(base != NULL)) { if (likely(base != NULL)) {
spin_lock_irqsave(&base->lock, *flags); spin_lock_irqsave(&base->cpu_base->lock, *flags);
if (likely(base == timer->base)) if (likely(base == timer->base))
return base; return base;
/* The timer has migrated to another CPU: */ /* The timer has migrated to another CPU: */
spin_unlock_irqrestore(&base->lock, *flags); spin_unlock_irqrestore(&base->cpu_base->lock, *flags);
} }
cpu_relax(); cpu_relax();
} }
@ -187,12 +191,14 @@ static struct hrtimer_base *lock_hrtimer_base(const struct hrtimer *timer,
/* /*
* Switch the timer base to the current CPU when possible. * Switch the timer base to the current CPU when possible.
*/ */
static inline struct hrtimer_base * static inline struct hrtimer_clock_base *
switch_hrtimer_base(struct hrtimer *timer, struct hrtimer_base *base) switch_hrtimer_base(struct hrtimer *timer, struct hrtimer_clock_base *base)
{ {
struct hrtimer_base *new_base; struct hrtimer_clock_base *new_base;
struct hrtimer_cpu_base *new_cpu_base;
new_base = &__get_cpu_var(hrtimer_bases)[base->index]; new_cpu_base = &__get_cpu_var(hrtimer_bases);
new_base = &new_cpu_base->clock_base[base->index];
if (base != new_base) { if (base != new_base) {
/* /*
@ -204,13 +210,13 @@ switch_hrtimer_base(struct hrtimer *timer, struct hrtimer_base *base)
* completed. There is no conflict as we hold the lock until * completed. There is no conflict as we hold the lock until
* the timer is enqueued. * the timer is enqueued.
*/ */
if (unlikely(base->curr_timer == timer)) if (unlikely(base->cpu_base->curr_timer == timer))
return base; return base;
/* See the comment in lock_timer_base() */ /* See the comment in lock_timer_base() */
timer->base = NULL; timer->base = NULL;
spin_unlock(&base->lock); spin_unlock(&base->cpu_base->lock);
spin_lock(&new_base->lock); spin_lock(&new_base->cpu_base->lock);
timer->base = new_base; timer->base = new_base;
} }
return new_base; return new_base;
@ -220,12 +226,12 @@ switch_hrtimer_base(struct hrtimer *timer, struct hrtimer_base *base)
#define set_curr_timer(b, t) do { } while (0) #define set_curr_timer(b, t) do { } while (0)
static inline struct hrtimer_base * static inline struct hrtimer_clock_base *
lock_hrtimer_base(const struct hrtimer *timer, unsigned long *flags) lock_hrtimer_base(const struct hrtimer *timer, unsigned long *flags)
{ {
struct hrtimer_base *base = timer->base; struct hrtimer_clock_base *base = timer->base;
spin_lock_irqsave(&base->lock, *flags); spin_lock_irqsave(&base->cpu_base->lock, *flags);
return base; return base;
} }
@ -305,7 +311,7 @@ void hrtimer_notify_resume(void)
static inline static inline
void unlock_hrtimer_base(const struct hrtimer *timer, unsigned long *flags) void unlock_hrtimer_base(const struct hrtimer *timer, unsigned long *flags)
{ {
spin_unlock_irqrestore(&timer->base->lock, *flags); spin_unlock_irqrestore(&timer->base->cpu_base->lock, *flags);
} }
/** /**
@ -355,7 +361,8 @@ hrtimer_forward(struct hrtimer *timer, ktime_t now, ktime_t interval)
* The timer is inserted in expiry order. Insertion into the * The timer is inserted in expiry order. Insertion into the
* red black tree is O(log(n)). Must hold the base lock. * red black tree is O(log(n)). Must hold the base lock.
*/ */
static void enqueue_hrtimer(struct hrtimer *timer, struct hrtimer_base *base) static void enqueue_hrtimer(struct hrtimer *timer,
struct hrtimer_clock_base *base)
{ {
struct rb_node **link = &base->active.rb_node; struct rb_node **link = &base->active.rb_node;
struct rb_node *parent = NULL; struct rb_node *parent = NULL;
@ -394,7 +401,8 @@ static void enqueue_hrtimer(struct hrtimer *timer, struct hrtimer_base *base)
* *
* Caller must hold the base lock. * Caller must hold the base lock.
*/ */
static void __remove_hrtimer(struct hrtimer *timer, struct hrtimer_base *base) static void __remove_hrtimer(struct hrtimer *timer,
struct hrtimer_clock_base *base)
{ {
/* /*
* Remove the timer from the rbtree and replace the * Remove the timer from the rbtree and replace the
@ -410,7 +418,7 @@ static void __remove_hrtimer(struct hrtimer *timer, struct hrtimer_base *base)
* remove hrtimer, called with base lock held * remove hrtimer, called with base lock held
*/ */
static inline int static inline int
remove_hrtimer(struct hrtimer *timer, struct hrtimer_base *base) remove_hrtimer(struct hrtimer *timer, struct hrtimer_clock_base *base)
{ {
if (hrtimer_active(timer)) { if (hrtimer_active(timer)) {
__remove_hrtimer(timer, base); __remove_hrtimer(timer, base);
@ -432,7 +440,7 @@ remove_hrtimer(struct hrtimer *timer, struct hrtimer_base *base)
int int
hrtimer_start(struct hrtimer *timer, ktime_t tim, const enum hrtimer_mode mode) hrtimer_start(struct hrtimer *timer, ktime_t tim, const enum hrtimer_mode mode)
{ {
struct hrtimer_base *base, *new_base; struct hrtimer_clock_base *base, *new_base;
unsigned long flags; unsigned long flags;
int ret; int ret;
@ -479,13 +487,13 @@ EXPORT_SYMBOL_GPL(hrtimer_start);
*/ */
int hrtimer_try_to_cancel(struct hrtimer *timer) int hrtimer_try_to_cancel(struct hrtimer *timer)
{ {
struct hrtimer_base *base; struct hrtimer_clock_base *base;
unsigned long flags; unsigned long flags;
int ret = -1; int ret = -1;
base = lock_hrtimer_base(timer, &flags); base = lock_hrtimer_base(timer, &flags);
if (base->curr_timer != timer) if (base->cpu_base->curr_timer != timer)
ret = remove_hrtimer(timer, base); ret = remove_hrtimer(timer, base);
unlock_hrtimer_base(timer, &flags); unlock_hrtimer_base(timer, &flags);
@ -521,12 +529,12 @@ EXPORT_SYMBOL_GPL(hrtimer_cancel);
*/ */
ktime_t hrtimer_get_remaining(const struct hrtimer *timer) ktime_t hrtimer_get_remaining(const struct hrtimer *timer)
{ {
struct hrtimer_base *base; struct hrtimer_clock_base *base;
unsigned long flags; unsigned long flags;
ktime_t rem; ktime_t rem;
base = lock_hrtimer_base(timer, &flags); base = lock_hrtimer_base(timer, &flags);
rem = ktime_sub(timer->expires, timer->base->get_time()); rem = ktime_sub(timer->expires, base->get_time());
unlock_hrtimer_base(timer, &flags); unlock_hrtimer_base(timer, &flags);
return rem; return rem;
@ -542,26 +550,29 @@ EXPORT_SYMBOL_GPL(hrtimer_get_remaining);
*/ */
ktime_t hrtimer_get_next_event(void) ktime_t hrtimer_get_next_event(void)
{ {
struct hrtimer_base *base = __get_cpu_var(hrtimer_bases); struct hrtimer_cpu_base *cpu_base = &__get_cpu_var(hrtimer_bases);
struct hrtimer_clock_base *base = cpu_base->clock_base;
ktime_t delta, mindelta = { .tv64 = KTIME_MAX }; ktime_t delta, mindelta = { .tv64 = KTIME_MAX };
unsigned long flags; unsigned long flags;
int i; int i;
for (i = 0; i < MAX_HRTIMER_BASES; i++, base++) { spin_lock_irqsave(&cpu_base->lock, flags);
for (i = 0; i < HRTIMER_MAX_CLOCK_BASES; i++, base++) {
struct hrtimer *timer; struct hrtimer *timer;
spin_lock_irqsave(&base->lock, flags); if (!base->first)
if (!base->first) {
spin_unlock_irqrestore(&base->lock, flags);
continue; continue;
}
timer = rb_entry(base->first, struct hrtimer, node); timer = rb_entry(base->first, struct hrtimer, node);
delta.tv64 = timer->expires.tv64; delta.tv64 = timer->expires.tv64;
spin_unlock_irqrestore(&base->lock, flags);
delta = ktime_sub(delta, base->get_time()); delta = ktime_sub(delta, base->get_time());
if (delta.tv64 < mindelta.tv64) if (delta.tv64 < mindelta.tv64)
mindelta.tv64 = delta.tv64; mindelta.tv64 = delta.tv64;
} }
spin_unlock_irqrestore(&cpu_base->lock, flags);
if (mindelta.tv64 < 0) if (mindelta.tv64 < 0)
mindelta.tv64 = 0; mindelta.tv64 = 0;
return mindelta; return mindelta;
@ -577,16 +588,16 @@ ktime_t hrtimer_get_next_event(void)
void hrtimer_init(struct hrtimer *timer, clockid_t clock_id, void hrtimer_init(struct hrtimer *timer, clockid_t clock_id,
enum hrtimer_mode mode) enum hrtimer_mode mode)
{ {
struct hrtimer_base *bases; struct hrtimer_cpu_base *cpu_base;
memset(timer, 0, sizeof(struct hrtimer)); memset(timer, 0, sizeof(struct hrtimer));
bases = __raw_get_cpu_var(hrtimer_bases); cpu_base = &__raw_get_cpu_var(hrtimer_bases);
if (clock_id == CLOCK_REALTIME && mode != HRTIMER_MODE_ABS) if (clock_id == CLOCK_REALTIME && mode != HRTIMER_MODE_ABS)
clock_id = CLOCK_MONOTONIC; clock_id = CLOCK_MONOTONIC;
timer->base = &bases[clock_id]; timer->base = &cpu_base->clock_base[clock_id];
rb_set_parent(&timer->node, &timer->node); rb_set_parent(&timer->node, &timer->node);
} }
EXPORT_SYMBOL_GPL(hrtimer_init); EXPORT_SYMBOL_GPL(hrtimer_init);
@ -601,10 +612,10 @@ EXPORT_SYMBOL_GPL(hrtimer_init);
*/ */
int hrtimer_get_res(const clockid_t which_clock, struct timespec *tp) int hrtimer_get_res(const clockid_t which_clock, struct timespec *tp)
{ {
struct hrtimer_base *bases; struct hrtimer_cpu_base *cpu_base;
bases = __raw_get_cpu_var(hrtimer_bases); cpu_base = &__raw_get_cpu_var(hrtimer_bases);
*tp = ktime_to_timespec(bases[which_clock].resolution); *tp = ktime_to_timespec(cpu_base->clock_base[which_clock].resolution);
return 0; return 0;
} }
@ -613,9 +624,11 @@ EXPORT_SYMBOL_GPL(hrtimer_get_res);
/* /*
* Expire the per base hrtimer-queue: * Expire the per base hrtimer-queue:
*/ */
static inline void run_hrtimer_queue(struct hrtimer_base *base) static inline void run_hrtimer_queue(struct hrtimer_cpu_base *cpu_base,
int index)
{ {
struct rb_node *node; struct rb_node *node;
struct hrtimer_clock_base *base = &cpu_base->clock_base[index];
if (!base->first) if (!base->first)
return; return;
@ -623,7 +636,7 @@ static inline void run_hrtimer_queue(struct hrtimer_base *base)
if (base->get_softirq_time) if (base->get_softirq_time)
base->softirq_time = base->get_softirq_time(); base->softirq_time = base->get_softirq_time();
spin_lock_irq(&base->lock); spin_lock_irq(&cpu_base->lock);
while ((node = base->first)) { while ((node = base->first)) {
struct hrtimer *timer; struct hrtimer *timer;
@ -635,21 +648,21 @@ static inline void run_hrtimer_queue(struct hrtimer_base *base)
break; break;
fn = timer->function; fn = timer->function;
set_curr_timer(base, timer); set_curr_timer(cpu_base, timer);
__remove_hrtimer(timer, base); __remove_hrtimer(timer, base);
spin_unlock_irq(&base->lock); spin_unlock_irq(&cpu_base->lock);
restart = fn(timer); restart = fn(timer);
spin_lock_irq(&base->lock); spin_lock_irq(&cpu_base->lock);
if (restart != HRTIMER_NORESTART) { if (restart != HRTIMER_NORESTART) {
BUG_ON(hrtimer_active(timer)); BUG_ON(hrtimer_active(timer));
enqueue_hrtimer(timer, base); enqueue_hrtimer(timer, base);
} }
} }
set_curr_timer(base, NULL); set_curr_timer(cpu_base, NULL);
spin_unlock_irq(&base->lock); spin_unlock_irq(&cpu_base->lock);
} }
/* /*
@ -657,13 +670,13 @@ static inline void run_hrtimer_queue(struct hrtimer_base *base)
*/ */
void hrtimer_run_queues(void) void hrtimer_run_queues(void)
{ {
struct hrtimer_base *base = __get_cpu_var(hrtimer_bases); struct hrtimer_cpu_base *cpu_base = &__get_cpu_var(hrtimer_bases);
int i; int i;
hrtimer_get_softirq_time(base); hrtimer_get_softirq_time(cpu_base);
for (i = 0; i < MAX_HRTIMER_BASES; i++) for (i = 0; i < HRTIMER_MAX_CLOCK_BASES; i++)
run_hrtimer_queue(&base[i]); run_hrtimer_queue(cpu_base, i);
} }
/* /*
@ -792,19 +805,21 @@ sys_nanosleep(struct timespec __user *rqtp, struct timespec __user *rmtp)
*/ */
static void __devinit init_hrtimers_cpu(int cpu) static void __devinit init_hrtimers_cpu(int cpu)
{ {
struct hrtimer_base *base = per_cpu(hrtimer_bases, cpu); struct hrtimer_cpu_base *cpu_base = &per_cpu(hrtimer_bases, cpu);
int i; int i;
for (i = 0; i < MAX_HRTIMER_BASES; i++, base++) { spin_lock_init(&cpu_base->lock);
spin_lock_init(&base->lock); lockdep_set_class(&cpu_base->lock, &cpu_base->lock_key);
lockdep_set_class(&base->lock, &base->lock_key);
} for (i = 0; i < HRTIMER_MAX_CLOCK_BASES; i++)
cpu_base->clock_base[i].cpu_base = cpu_base;
} }
#ifdef CONFIG_HOTPLUG_CPU #ifdef CONFIG_HOTPLUG_CPU
static void migrate_hrtimer_list(struct hrtimer_base *old_base, static void migrate_hrtimer_list(struct hrtimer_clock_base *old_base,
struct hrtimer_base *new_base) struct hrtimer_clock_base *new_base)
{ {
struct hrtimer *timer; struct hrtimer *timer;
struct rb_node *node; struct rb_node *node;
@ -819,29 +834,26 @@ static void migrate_hrtimer_list(struct hrtimer_base *old_base,
static void migrate_hrtimers(int cpu) static void migrate_hrtimers(int cpu)
{ {
struct hrtimer_base *old_base, *new_base; struct hrtimer_cpu_base *old_base, *new_base;
int i; int i;
BUG_ON(cpu_online(cpu)); BUG_ON(cpu_online(cpu));
old_base = per_cpu(hrtimer_bases, cpu); old_base = &per_cpu(hrtimer_bases, cpu);
new_base = get_cpu_var(hrtimer_bases); new_base = &get_cpu_var(hrtimer_bases);
local_irq_disable(); local_irq_disable();
for (i = 0; i < MAX_HRTIMER_BASES; i++) { spin_lock(&new_base->lock);
spin_lock(&old_base->lock);
spin_lock(&new_base->lock);
spin_lock(&old_base->lock);
for (i = 0; i < HRTIMER_MAX_CLOCK_BASES; i++) {
BUG_ON(old_base->curr_timer); BUG_ON(old_base->curr_timer);
migrate_hrtimer_list(old_base, new_base); migrate_hrtimer_list(&old_base->clock_base[i],
&new_base->clock_base[i]);
spin_unlock(&old_base->lock);
spin_unlock(&new_base->lock);
old_base++;
new_base++;
} }
spin_unlock(&old_base->lock);
spin_unlock(&new_base->lock);
local_irq_enable(); local_irq_enable();
put_cpu_var(hrtimer_bases); put_cpu_var(hrtimer_bases);