[PATCH] hrtimer: hrtimer core code
authorThomas Gleixner <tglx@linutronix.de>
Tue, 10 Jan 2006 04:52:32 +0000 (20:52 -0800)
committerLinus Torvalds <torvalds@g5.osdl.org>
Tue, 10 Jan 2006 16:01:37 +0000 (08:01 -0800)
hrtimer subsystem core.  It is initialized at bootup and expired by the timer
interrupt, but is otherwise not utilized by any other subsystem yet.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
include/linux/hrtimer.h [new file with mode: 0644]
include/linux/ktime.h
init/main.c
kernel/Makefile
kernel/hrtimer.c [new file with mode: 0644]
kernel/timer.c

diff --git a/include/linux/hrtimer.h b/include/linux/hrtimer.h
new file mode 100644 (file)
index 0000000..64f8d55
--- /dev/null
@@ -0,0 +1,130 @@
+/*
+ *  include/linux/hrtimer.h
+ *
+ *  hrtimers - High-resolution kernel timers
+ *
+ *   Copyright(C) 2005, Thomas Gleixner <tglx@linutronix.de>
+ *   Copyright(C) 2005, Red Hat, Inc., Ingo Molnar
+ *
+ *  data type definitions, declarations, prototypes
+ *
+ *  Started by: Thomas Gleixner and Ingo Molnar
+ *
+ *  For licencing details see kernel-base/COPYING
+ */
+#ifndef _LINUX_HRTIMER_H
+#define _LINUX_HRTIMER_H
+
+#include <linux/rbtree.h>
+#include <linux/ktime.h>
+#include <linux/init.h>
+#include <linux/list.h>
+#include <linux/wait.h>
+
+/*
+ * Mode arguments of xxx_hrtimer functions:
+ */
+enum hrtimer_mode {
+       HRTIMER_ABS,    /* Time value is absolute */
+       HRTIMER_REL,    /* Time value is relative to now */
+};
+
+enum hrtimer_restart {
+       HRTIMER_NORESTART,
+       HRTIMER_RESTART,
+};
+
+/*
+ * Timer states:
+ */
+enum hrtimer_state {
+       HRTIMER_INACTIVE,       /* Timer is inactive */
+       HRTIMER_EXPIRED,                /* Timer is expired */
+       HRTIMER_PENDING,                /* Timer is pending */
+};
+
+struct hrtimer_base;
+
+/**
+ * struct hrtimer - the basic hrtimer structure
+ *
+ * @node:      red black tree node for time ordered insertion
+ * @list:      list head for easier access to the time ordered list,
+ *             without walking the red black tree.
+ * @expires:   the absolute expiry time in the hrtimers internal
+ *             representation. The time is related to the clock on
+ *             which the timer is based.
+ * @state:     state of the timer
+ * @function:  timer expiry callback function
+ * @data:      argument for the callback function
+ * @base:      pointer to the timer base (per cpu and per clock)
+ *
+ * The hrtimer structure must be initialized by init_hrtimer_#CLOCKTYPE()
+ */
+struct hrtimer {
+       struct rb_node          node;
+       struct list_head        list;
+       ktime_t                 expires;
+       enum hrtimer_state      state;
+       int                     (*function)(void *);
+       void                    *data;
+       struct hrtimer_base     *base;
+};
+
+/**
+ * struct hrtimer_base - the timer base for a specific clock
+ *
+ * @index:     clock type index for per_cpu support when moving a 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
+ * @pending:   list of pending timers for simple time ordered access
+ * @resolution:        the resolution of the clock, in nanoseconds
+ * @get_time:  function to retrieve the current time of the clock
+ * @curr_timer:        the timer which is executing a callback right now
+ */
+struct hrtimer_base {
+       clockid_t               index;
+       spinlock_t              lock;
+       struct rb_root          active;
+       struct list_head        pending;
+       unsigned long           resolution;
+       ktime_t                 (*get_time)(void);
+       struct hrtimer          *curr_timer;
+};
+
+/* Exported timer functions: */
+
+/* Initialize timers: */
+extern void hrtimer_init(struct hrtimer *timer, const clockid_t which_clock);
+extern void hrtimer_rebase(struct hrtimer *timer, const clockid_t which_clock);
+
+
+/* Basic timer operations: */
+extern int hrtimer_start(struct hrtimer *timer, ktime_t tim,
+                        const enum hrtimer_mode mode);
+extern int hrtimer_cancel(struct hrtimer *timer);
+extern int hrtimer_try_to_cancel(struct hrtimer *timer);
+
+#define hrtimer_restart(timer) hrtimer_start((timer), (timer)->expires, HRTIMER_ABS)
+
+/* Query timers: */
+extern ktime_t hrtimer_get_remaining(const struct hrtimer *timer);
+extern int hrtimer_get_res(const clockid_t which_clock, struct timespec *tp);
+
+static inline int hrtimer_active(const struct hrtimer *timer)
+{
+       return timer->state == HRTIMER_PENDING;
+}
+
+/* Forward a hrtimer so it expires after now: */
+extern unsigned long hrtimer_forward(struct hrtimer *timer,
+                                    const ktime_t interval);
+
+/* Soft interrupt function to run the hrtimer queues: */
+extern void hrtimer_run_queues(void);
+
+/* Bootup initialization: */
+extern void __init hrtimers_init(void);
+
+#endif
index 5b9a9eb..222a047 100644 (file)
@@ -266,4 +266,19 @@ static inline u64 ktime_to_ns(const ktime_t kt)
 
 #endif
 
+/*
+ * The resolution of the clocks. The resolution value is returned in
+ * the clock_getres() system call to give application programmers an
+ * idea of the (in)accuracy of timers. Timer values are rounded up to
+ * this resolution values.
+ */
+#define KTIME_REALTIME_RES     (NSEC_PER_SEC/HZ)
+#define KTIME_MONOTONIC_RES    (NSEC_PER_SEC/HZ)
+
+/* Get the monotonic time in timespec format: */
+extern void ktime_get_ts(struct timespec *ts);
+
+/* Get the real (wall-) time in timespec format: */
+#define ktime_get_real_ts(ts)  getnstimeofday(ts)
+
 #endif
index 8342c28..e092b19 100644 (file)
@@ -485,6 +485,7 @@ asmlinkage void __init start_kernel(void)
        init_IRQ();
        pidhash_init();
        init_timers();
+       hrtimers_init();
        softirq_init();
        time_init();
 
index 1e03970..3551266 100644 (file)
@@ -7,7 +7,8 @@ obj-y     = sched.o fork.o exec_domain.o panic.o printk.o profile.o \
            sysctl.o capability.o ptrace.o timer.o user.o \
            signal.o sys.o kmod.o workqueue.o pid.o \
            rcupdate.o intermodule.o extable.o params.o posix-timers.o \
-           kthread.o wait.o kfifo.o sys_ni.o posix-cpu-timers.o mutex.o
+           kthread.o wait.o kfifo.o sys_ni.o posix-cpu-timers.o mutex.o \
+           hrtimer.o
 
 obj-$(CONFIG_DEBUG_MUTEXES) += mutex-debug.o
 obj-$(CONFIG_FUTEX) += futex.o
diff --git a/kernel/hrtimer.c b/kernel/hrtimer.c
new file mode 100644 (file)
index 0000000..690efd9
--- /dev/null
@@ -0,0 +1,679 @@
+/*
+ *  linux/kernel/hrtimer.c
+ *
+ *  Copyright(C) 2005, Thomas Gleixner <tglx@linutronix.de>
+ *  Copyright(C) 2005, Red Hat, Inc., Ingo Molnar
+ *
+ *  High-resolution kernel timers
+ *
+ *  In contrast to the low-resolution timeout API implemented in
+ *  kernel/timer.c, hrtimers provide finer resolution and accuracy
+ *  depending on system configuration and capabilities.
+ *
+ *  These timers are currently used for:
+ *   - itimers
+ *   - POSIX timers
+ *   - nanosleep
+ *   - precise in-kernel timing
+ *
+ *  Started by: Thomas Gleixner and Ingo Molnar
+ *
+ *  Credits:
+ *     based on kernel/timer.c
+ *
+ *  For licencing details see kernel-base/COPYING
+ */
+
+#include <linux/cpu.h>
+#include <linux/module.h>
+#include <linux/percpu.h>
+#include <linux/hrtimer.h>
+#include <linux/notifier.h>
+#include <linux/syscalls.h>
+#include <linux/interrupt.h>
+
+#include <asm/uaccess.h>
+
+/**
+ * ktime_get - get the monotonic time in ktime_t format
+ *
+ * returns the time in ktime_t format
+ */
+static ktime_t ktime_get(void)
+{
+       struct timespec now;
+
+       ktime_get_ts(&now);
+
+       return timespec_to_ktime(now);
+}
+
+/**
+ * ktime_get_real - get the real (wall-) time in ktime_t format
+ *
+ * returns the time in ktime_t format
+ */
+static ktime_t ktime_get_real(void)
+{
+       struct timespec now;
+
+       getnstimeofday(&now);
+
+       return timespec_to_ktime(now);
+}
+
+EXPORT_SYMBOL_GPL(ktime_get_real);
+
+/*
+ * The timer bases:
+ */
+
+#define MAX_HRTIMER_BASES 2
+
+static DEFINE_PER_CPU(struct hrtimer_base, hrtimer_bases[MAX_HRTIMER_BASES]) =
+{
+       {
+               .index = CLOCK_REALTIME,
+               .get_time = &ktime_get_real,
+               .resolution = KTIME_REALTIME_RES,
+       },
+       {
+               .index = CLOCK_MONOTONIC,
+               .get_time = &ktime_get,
+               .resolution = KTIME_MONOTONIC_RES,
+       },
+};
+
+/**
+ * ktime_get_ts - get the monotonic clock in timespec format
+ *
+ * @ts:                pointer to timespec variable
+ *
+ * The function calculates the monotonic clock from the realtime
+ * clock and the wall_to_monotonic offset and stores the result
+ * in normalized timespec format in the variable pointed to by ts.
+ */
+void ktime_get_ts(struct timespec *ts)
+{
+       struct timespec tomono;
+       unsigned long seq;
+
+       do {
+               seq = read_seqbegin(&xtime_lock);
+               getnstimeofday(ts);
+               tomono = wall_to_monotonic;
+
+       } while (read_seqretry(&xtime_lock, seq));
+
+       set_normalized_timespec(ts, ts->tv_sec + tomono.tv_sec,
+                               ts->tv_nsec + tomono.tv_nsec);
+}
+
+/*
+ * Functions and macros which are different for UP/SMP systems are kept in a
+ * single place
+ */
+#ifdef CONFIG_SMP
+
+#define set_curr_timer(b, t)           do { (b)->curr_timer = (t); } while (0)
+
+/*
+ * We are using hashed locking: holding per_cpu(hrtimer_bases)[n].lock
+ * means that all timers which are tied to this base via timer->base are
+ * locked, and the base itself is locked too.
+ *
+ * So __run_timers/migrate_timers can safely modify all timers which could
+ * be found on the lists/queues.
+ *
+ * When the timer's base is locked, and the timer removed from list, it is
+ * possible to set timer->base = NULL and drop the lock: the timer remains
+ * locked.
+ */
+static struct hrtimer_base *lock_hrtimer_base(const struct hrtimer *timer,
+                                             unsigned long *flags)
+{
+       struct hrtimer_base *base;
+
+       for (;;) {
+               base = timer->base;
+               if (likely(base != NULL)) {
+                       spin_lock_irqsave(&base->lock, *flags);
+                       if (likely(base == timer->base))
+                               return base;
+                       /* The timer has migrated to another CPU: */
+                       spin_unlock_irqrestore(&base->lock, *flags);
+               }
+               cpu_relax();
+       }
+}
+
+/*
+ * Switch the timer base to the current CPU when possible.
+ */
+static inline struct hrtimer_base *
+switch_hrtimer_base(struct hrtimer *timer, struct hrtimer_base *base)
+{
+       struct hrtimer_base *new_base;
+
+       new_base = &__get_cpu_var(hrtimer_bases[base->index]);
+
+       if (base != new_base) {
+               /*
+                * We are trying to schedule the timer on the local CPU.
+                * However we can't change timer's base while it is running,
+                * so we keep it on the same CPU. No hassle vs. reprogramming
+                * the event source in the high resolution case. The softirq
+                * code will take care of this when the timer function has
+                * completed. There is no conflict as we hold the lock until
+                * the timer is enqueued.
+                */
+               if (unlikely(base->curr_timer == timer))
+                       return base;
+
+               /* See the comment in lock_timer_base() */
+               timer->base = NULL;
+               spin_unlock(&base->lock);
+               spin_lock(&new_base->lock);
+               timer->base = new_base;
+       }
+       return new_base;
+}
+
+#else /* CONFIG_SMP */
+
+#define set_curr_timer(b, t)           do { } while (0)
+
+static inline struct hrtimer_base *
+lock_hrtimer_base(const struct hrtimer *timer, unsigned long *flags)
+{
+       struct hrtimer_base *base = timer->base;
+
+       spin_lock_irqsave(&base->lock, *flags);
+
+       return base;
+}
+
+#define switch_hrtimer_base(t, b)      (b)
+
+#endif /* !CONFIG_SMP */
+
+/*
+ * Functions for the union type storage format of ktime_t which are
+ * too large for inlining:
+ */
+#if BITS_PER_LONG < 64
+# ifndef CONFIG_KTIME_SCALAR
+/**
+ * ktime_add_ns - Add a scalar nanoseconds value to a ktime_t variable
+ *
+ * @kt:                addend
+ * @nsec:      the scalar nsec value to add
+ *
+ * Returns the sum of kt and nsec in ktime_t format
+ */
+ktime_t ktime_add_ns(const ktime_t kt, u64 nsec)
+{
+       ktime_t tmp;
+
+       if (likely(nsec < NSEC_PER_SEC)) {
+               tmp.tv64 = nsec;
+       } else {
+               unsigned long rem = do_div(nsec, NSEC_PER_SEC);
+
+               tmp = ktime_set((long)nsec, rem);
+       }
+
+       return ktime_add(kt, tmp);
+}
+
+#else /* CONFIG_KTIME_SCALAR */
+
+# endif /* !CONFIG_KTIME_SCALAR */
+
+/*
+ * Divide a ktime value by a nanosecond value
+ */
+static unsigned long ktime_divns(const ktime_t kt, nsec_t div)
+{
+       u64 dclc, inc, dns;
+       int sft = 0;
+
+       dclc = dns = ktime_to_ns(kt);
+       inc = div;
+       /* Make sure the divisor is less than 2^32: */
+       while (div >> 32) {
+               sft++;
+               div >>= 1;
+       }
+       dclc >>= sft;
+       do_div(dclc, (unsigned long) div);
+
+       return (unsigned long) dclc;
+}
+
+#else /* BITS_PER_LONG < 64 */
+# define ktime_divns(kt, div)          (unsigned long)((kt).tv64 / (div))
+#endif /* BITS_PER_LONG >= 64 */
+
+/*
+ * Counterpart to lock_timer_base above:
+ */
+static inline
+void unlock_hrtimer_base(const struct hrtimer *timer, unsigned long *flags)
+{
+       spin_unlock_irqrestore(&timer->base->lock, *flags);
+}
+
+/**
+ * hrtimer_forward - forward the timer expiry
+ *
+ * @timer:     hrtimer to forward
+ * @interval:  the interval to forward
+ *
+ * Forward the timer expiry so it will expire in the future.
+ * The number of overruns is added to the overrun field.
+ */
+unsigned long
+hrtimer_forward(struct hrtimer *timer, const ktime_t interval)
+{
+       unsigned long orun = 1;
+       ktime_t delta, now;
+
+       now = timer->base->get_time();
+
+       delta = ktime_sub(now, timer->expires);
+
+       if (delta.tv64 < 0)
+               return 0;
+
+       if (unlikely(delta.tv64 >= interval.tv64)) {
+               nsec_t incr = ktime_to_ns(interval);
+
+               orun = ktime_divns(delta, incr);
+               timer->expires = ktime_add_ns(timer->expires, incr * orun);
+               if (timer->expires.tv64 > now.tv64)
+                       return orun;
+               /*
+                * This (and the ktime_add() below) is the
+                * correction for exact:
+                */
+               orun++;
+       }
+       timer->expires = ktime_add(timer->expires, interval);
+
+       return orun;
+}
+
+/*
+ * enqueue_hrtimer - internal function to (re)start a timer
+ *
+ * The timer is inserted in expiry order. Insertion into the
+ * red black tree is O(log(n)). Must hold the base lock.
+ */
+static void enqueue_hrtimer(struct hrtimer *timer, struct hrtimer_base *base)
+{
+       struct rb_node **link = &base->active.rb_node;
+       struct list_head *prev = &base->pending;
+       struct rb_node *parent = NULL;
+       struct hrtimer *entry;
+
+       /*
+        * Find the right place in the rbtree:
+        */
+       while (*link) {
+               parent = *link;
+               entry = rb_entry(parent, struct hrtimer, node);
+               /*
+                * We dont care about collisions. Nodes with
+                * the same expiry time stay together.
+                */
+               if (timer->expires.tv64 < entry->expires.tv64)
+                       link = &(*link)->rb_left;
+               else {
+                       link = &(*link)->rb_right;
+                       prev = &entry->list;
+               }
+       }
+
+       /*
+        * Insert the timer to the rbtree and to the sorted list:
+        */
+       rb_link_node(&timer->node, parent, link);
+       rb_insert_color(&timer->node, &base->active);
+       list_add(&timer->list, prev);
+
+       timer->state = HRTIMER_PENDING;
+}
+
+
+/*
+ * __remove_hrtimer - internal function to remove a timer
+ *
+ * Caller must hold the base lock.
+ */
+static void __remove_hrtimer(struct hrtimer *timer, struct hrtimer_base *base)
+{
+       /*
+        * Remove the timer from the sorted list and from the rbtree:
+        */
+       list_del(&timer->list);
+       rb_erase(&timer->node, &base->active);
+}
+
+/*
+ * remove hrtimer, called with base lock held
+ */
+static inline int
+remove_hrtimer(struct hrtimer *timer, struct hrtimer_base *base)
+{
+       if (hrtimer_active(timer)) {
+               __remove_hrtimer(timer, base);
+               timer->state = HRTIMER_INACTIVE;
+               return 1;
+       }
+       return 0;
+}
+
+/**
+ * hrtimer_start - (re)start an relative timer on the current CPU
+ *
+ * @timer:     the timer to be added
+ * @tim:       expiry time
+ * @mode:      expiry mode: absolute (HRTIMER_ABS) or relative (HRTIMER_REL)
+ *
+ * Returns:
+ *  0 on success
+ *  1 when the timer was active
+ */
+int
+hrtimer_start(struct hrtimer *timer, ktime_t tim, const enum hrtimer_mode mode)
+{
+       struct hrtimer_base *base, *new_base;
+       unsigned long flags;
+       int ret;
+
+       base = lock_hrtimer_base(timer, &flags);
+
+       /* Remove an active timer from the queue: */
+       ret = remove_hrtimer(timer, base);
+
+       /* Switch the timer base, if necessary: */
+       new_base = switch_hrtimer_base(timer, base);
+
+       if (mode == HRTIMER_REL)
+               tim = ktime_add(tim, new_base->get_time());
+       timer->expires = tim;
+
+       enqueue_hrtimer(timer, new_base);
+
+       unlock_hrtimer_base(timer, &flags);
+
+       return ret;
+}
+
+/**
+ * hrtimer_try_to_cancel - try to deactivate a timer
+ *
+ * @timer:     hrtimer to stop
+ *
+ * Returns:
+ *  0 when the timer was not active
+ *  1 when the timer was active
+ * -1 when the timer is currently excuting the callback function and
+ *    can not be stopped
+ */
+int hrtimer_try_to_cancel(struct hrtimer *timer)
+{
+       struct hrtimer_base *base;
+       unsigned long flags;
+       int ret = -1;
+
+       base = lock_hrtimer_base(timer, &flags);
+
+       if (base->curr_timer != timer)
+               ret = remove_hrtimer(timer, base);
+
+       unlock_hrtimer_base(timer, &flags);
+
+       return ret;
+
+}
+
+/**
+ * hrtimer_cancel - cancel a timer and wait for the handler to finish.
+ *
+ * @timer:     the timer to be cancelled
+ *
+ * Returns:
+ *  0 when the timer was not active
+ *  1 when the timer was active
+ */
+int hrtimer_cancel(struct hrtimer *timer)
+{
+       for (;;) {
+               int ret = hrtimer_try_to_cancel(timer);
+
+               if (ret >= 0)
+                       return ret;
+       }
+}
+
+/**
+ * hrtimer_get_remaining - get remaining time for the timer
+ *
+ * @timer:     the timer to read
+ */
+ktime_t hrtimer_get_remaining(const struct hrtimer *timer)
+{
+       struct hrtimer_base *base;
+       unsigned long flags;
+       ktime_t rem;
+
+       base = lock_hrtimer_base(timer, &flags);
+       rem = ktime_sub(timer->expires, timer->base->get_time());
+       unlock_hrtimer_base(timer, &flags);
+
+       return rem;
+}
+
+/**
+ * hrtimer_rebase - rebase an initialized hrtimer to a different base
+ *
+ * @timer:     the timer to be rebased
+ * @clock_id:  the clock to be used
+ */
+void hrtimer_rebase(struct hrtimer *timer, const clockid_t clock_id)
+{
+       struct hrtimer_base *bases;
+
+       bases = per_cpu(hrtimer_bases, raw_smp_processor_id());
+       timer->base = &bases[clock_id];
+}
+
+/**
+ * hrtimer_init - initialize a timer to the given clock
+ *
+ * @timer:     the timer to be initialized
+ * @clock_id:  the clock to be used
+ */
+void hrtimer_init(struct hrtimer *timer, const clockid_t clock_id)
+{
+       memset(timer, 0, sizeof(struct hrtimer));
+       hrtimer_rebase(timer, clock_id);
+}
+
+/**
+ * hrtimer_get_res - get the timer resolution for a clock
+ *
+ * @which_clock: which clock to query
+ * @tp:                 pointer to timespec variable to store the resolution
+ *
+ * Store the resolution of the clock selected by which_clock in the
+ * variable pointed to by tp.
+ */
+int hrtimer_get_res(const clockid_t which_clock, struct timespec *tp)
+{
+       struct hrtimer_base *bases;
+
+       tp->tv_sec = 0;
+       bases = per_cpu(hrtimer_bases, raw_smp_processor_id());
+       tp->tv_nsec = bases[which_clock].resolution;
+
+       return 0;
+}
+
+/*
+ * Expire the per base hrtimer-queue:
+ */
+static inline void run_hrtimer_queue(struct hrtimer_base *base)
+{
+       ktime_t now = base->get_time();
+
+       spin_lock_irq(&base->lock);
+
+       while (!list_empty(&base->pending)) {
+               struct hrtimer *timer;
+               int (*fn)(void *);
+               int restart;
+               void *data;
+
+               timer = list_entry(base->pending.next, struct hrtimer, list);
+               if (now.tv64 <= timer->expires.tv64)
+                       break;
+
+               fn = timer->function;
+               data = timer->data;
+               set_curr_timer(base, timer);
+               __remove_hrtimer(timer, base);
+               spin_unlock_irq(&base->lock);
+
+               /*
+                * fn == NULL is special case for the simplest timer
+                * variant - wake up process and do not restart:
+                */
+               if (!fn) {
+                       wake_up_process(data);
+                       restart = HRTIMER_NORESTART;
+               } else
+                       restart = fn(data);
+
+               spin_lock_irq(&base->lock);
+
+               if (restart == HRTIMER_RESTART)
+                       enqueue_hrtimer(timer, base);
+               else
+                       timer->state = HRTIMER_EXPIRED;
+       }
+       set_curr_timer(base, NULL);
+       spin_unlock_irq(&base->lock);
+}
+
+/*
+ * Called from timer softirq every jiffy, expire hrtimers:
+ */
+void hrtimer_run_queues(void)
+{
+       struct hrtimer_base *base = __get_cpu_var(hrtimer_bases);
+       int i;
+
+       for (i = 0; i < MAX_HRTIMER_BASES; i++)
+               run_hrtimer_queue(&base[i]);
+}
+
+/*
+ * Functions related to boot-time initialization:
+ */
+static void __devinit init_hrtimers_cpu(int cpu)
+{
+       struct hrtimer_base *base = per_cpu(hrtimer_bases, cpu);
+       int i;
+
+       for (i = 0; i < MAX_HRTIMER_BASES; i++) {
+               spin_lock_init(&base->lock);
+               INIT_LIST_HEAD(&base->pending);
+               base++;
+       }
+}
+
+#ifdef CONFIG_HOTPLUG_CPU
+
+static void migrate_hrtimer_list(struct hrtimer_base *old_base,
+                               struct hrtimer_base *new_base)
+{
+       struct hrtimer *timer;
+       struct rb_node *node;
+
+       while ((node = rb_first(&old_base->active))) {
+               timer = rb_entry(node, struct hrtimer, node);
+               __remove_hrtimer(timer, old_base);
+               timer->base = new_base;
+               enqueue_hrtimer(timer, new_base);
+       }
+}
+
+static void migrate_hrtimers(int cpu)
+{
+       struct hrtimer_base *old_base, *new_base;
+       int i;
+
+       BUG_ON(cpu_online(cpu));
+       old_base = per_cpu(hrtimer_bases, cpu);
+       new_base = get_cpu_var(hrtimer_bases);
+
+       local_irq_disable();
+
+       for (i = 0; i < MAX_HRTIMER_BASES; i++) {
+
+               spin_lock(&new_base->lock);
+               spin_lock(&old_base->lock);
+
+               BUG_ON(old_base->curr_timer);
+
+               migrate_hrtimer_list(old_base, new_base);
+
+               spin_unlock(&old_base->lock);
+               spin_unlock(&new_base->lock);
+               old_base++;
+               new_base++;
+       }
+
+       local_irq_enable();
+       put_cpu_var(hrtimer_bases);
+}
+#endif /* CONFIG_HOTPLUG_CPU */
+
+static int __devinit hrtimer_cpu_notify(struct notifier_block *self,
+                                       unsigned long action, void *hcpu)
+{
+       long cpu = (long)hcpu;
+
+       switch (action) {
+
+       case CPU_UP_PREPARE:
+               init_hrtimers_cpu(cpu);
+               break;
+
+#ifdef CONFIG_HOTPLUG_CPU
+       case CPU_DEAD:
+               migrate_hrtimers(cpu);
+               break;
+#endif
+
+       default:
+               break;
+       }
+
+       return NOTIFY_OK;
+}
+
+static struct notifier_block __devinitdata hrtimers_nb = {
+       .notifier_call = hrtimer_cpu_notify,
+};
+
+void __init hrtimers_init(void)
+{
+       hrtimer_cpu_notify(&hrtimers_nb, (unsigned long)CPU_UP_PREPARE,
+                         (void *)(long)smp_processor_id());
+       register_cpu_notifier(&hrtimers_nb);
+}
+
index 074b4bd..80bf2ac 100644 (file)
@@ -858,6 +858,7 @@ static void run_timer_softirq(struct softirq_action *h)
 {
        tvec_base_t *base = &__get_cpu_var(tvec_bases);
 
+       hrtimer_run_queues();
        if (time_after_eq(jiffies, base->timer_jiffies))
                __run_timers(base);
 }