[net-next-2.6.git] / kernel / cpu.c

/* CPU control.
 * (C) 2001, 2002, 2003, 2004 Rusty Russell
 *
 * This code is licenced under the GPL.
 */
#include <linux/proc_fs.h>
#include <linux/smp.h>
#include <linux/init.h>
#include <linux/notifier.h>
#include <linux/sched.h>
#include <linux/unistd.h>
#include <linux/cpu.h>
#include <linux/module.h>
#include <linux/kthread.h>
#include <linux/stop_machine.h>
#include <linux/mutex.h>
#include <linux/gfp.h>

#ifdef CONFIG_SMP
/* Serializes the updates to cpu_online_mask, cpu_present_mask */
static DEFINE_MUTEX(cpu_add_remove_lock);

/*
 * The following two API's must be used when attempting
 * to serialize the updates to cpu_online_mask, cpu_present_mask.
 */
void cpu_maps_update_begin(void)
{
	mutex_lock(&cpu_add_remove_lock);
}

void cpu_maps_update_done(void)
{
	mutex_unlock(&cpu_add_remove_lock);
}

static RAW_NOTIFIER_HEAD(cpu_chain);

/* If set, cpu_up and cpu_down will return -EBUSY and do nothing.
 * Should always be manipulated under cpu_add_remove_lock
 */
static int cpu_hotplug_disabled;

#ifdef CONFIG_HOTPLUG_CPU

static struct {
	struct task_struct *active_writer;
	struct mutex lock; /* Synchronizes accesses to refcount, */
	/*
	 * Also blocks the new readers during
	 * an ongoing cpu hotplug operation.
	 */
	int refcount;
} cpu_hotplug = {
	.active_writer = NULL,
	.lock = __MUTEX_INITIALIZER(cpu_hotplug.lock),
	.refcount = 0,
};

void get_online_cpus(void)
{
	might_sleep();
	if (cpu_hotplug.active_writer == current)
		return;
	mutex_lock(&cpu_hotplug.lock);
	cpu_hotplug.refcount++;
	mutex_unlock(&cpu_hotplug.lock);

}
EXPORT_SYMBOL_GPL(get_online_cpus);

void put_online_cpus(void)
{
	if (cpu_hotplug.active_writer == current)
		return;
	mutex_lock(&cpu_hotplug.lock);
	if (!--cpu_hotplug.refcount && unlikely(cpu_hotplug.active_writer))
		wake_up_process(cpu_hotplug.active_writer);
	mutex_unlock(&cpu_hotplug.lock);

}
EXPORT_SYMBOL_GPL(put_online_cpus);

/*
 * This ensures that the hotplug operation can begin only when the
 * refcount goes to zero.
 *
 * Note that during a cpu-hotplug operation, the new readers, if any,
 * will be blocked by the cpu_hotplug.lock
 *
 * Since cpu_hotplug_begin() is always called after invoking
 * cpu_maps_update_begin(), we can be sure that only one writer is active.
 *
 * Note that theoretically, there is a possibility of a livelock:
 * - Refcount goes to zero, last reader wakes up the sleeping
 *   writer.
 * - Last reader unlocks the cpu_hotplug.lock.
 * - A new reader arrives at this moment, bumps up the refcount.
 * - The writer acquires the cpu_hotplug.lock finds the refcount
 *   non zero and goes to sleep again.
 *
 * However, this is very difficult to achieve in practice since
 * get_online_cpus() not an api which is called all that often.
 *
 */
static void cpu_hotplug_begin(void)
{
	cpu_hotplug.active_writer = current;

	for (;;) {
		mutex_lock(&cpu_hotplug.lock);
		if (likely(!cpu_hotplug.refcount))
			break;
		__set_current_state(TASK_UNINTERRUPTIBLE);
		mutex_unlock(&cpu_hotplug.lock);
		schedule();
	}
}

static void cpu_hotplug_done(void)
{
	cpu_hotplug.active_writer = NULL;
	mutex_unlock(&cpu_hotplug.lock);
}

#else /* #if CONFIG_HOTPLUG_CPU */
static void cpu_hotplug_begin(void) {}
static void cpu_hotplug_done(void) {}
#endif	/* #esle #if CONFIG_HOTPLUG_CPU */

/* Need to know about CPUs going up/down? */
int __ref register_cpu_notifier(struct notifier_block *nb)
{
	int ret;
	cpu_maps_update_begin();
	ret = raw_notifier_chain_register(&cpu_chain, nb);
	cpu_maps_update_done();
	return ret;
}

static int __cpu_notify(unsigned long val, void *v, int nr_to_call,
			int *nr_calls)
{
	int ret;

	ret = __raw_notifier_call_chain(&cpu_chain, val, v, nr_to_call,
					nr_calls);

	return notifier_to_errno(ret);
}

static int cpu_notify(unsigned long val, void *v)
{
	return __cpu_notify(val, v, -1, NULL);
}

#ifdef CONFIG_HOTPLUG_CPU

static void cpu_notify_nofail(unsigned long val, void *v)
{
	BUG_ON(cpu_notify(val, v));
}

EXPORT_SYMBOL(register_cpu_notifier);

void __ref unregister_cpu_notifier(struct notifier_block *nb)
{
	cpu_maps_update_begin();
	raw_notifier_chain_unregister(&cpu_chain, nb);
	cpu_maps_update_done();
}
EXPORT_SYMBOL(unregister_cpu_notifier);

static inline void check_for_tasks(int cpu)
{
	struct task_struct *p;

	write_lock_irq(&tasklist_lock);
	for_each_process(p) {
		if (task_cpu(p) == cpu && p->state == TASK_RUNNING &&
		    (!cputime_eq(p->utime, cputime_zero) ||
		     !cputime_eq(p->stime, cputime_zero)))
			printk(KERN_WARNING "Task %s (pid = %d) is on cpu %d "
				"(state = %ld, flags = %x)\n",
				p->comm, task_pid_nr(p), cpu,
				p->state, p->flags);
	}
	write_unlock_irq(&tasklist_lock);
}

struct take_cpu_down_param {
	struct task_struct *caller;
	unsigned long mod;
	void *hcpu;
};

/* Take this CPU down. */
static int __ref take_cpu_down(void *_param)
{
	struct take_cpu_down_param *param = _param;
	unsigned int cpu = (unsigned long)param->hcpu;
	int err;

	/* Ensure this CPU doesn't handle any more interrupts. */
	err = __cpu_disable();
	if (err < 0)
		return err;

	cpu_notify(CPU_DYING | param->mod, param->hcpu);

	if (task_cpu(param->caller) == cpu)
		move_task_off_dead_cpu(cpu, param->caller);
	/* Force idle task to run as soon as we yield: it should
	   immediately notice cpu is offline and die quickly. */
	sched_idle_next();
	return 0;
}

/* Requires cpu_add_remove_lock to be held */
static int __ref _cpu_down(unsigned int cpu, int tasks_frozen)
{
	int err, nr_calls = 0;
	void *hcpu = (void *)(long)cpu;
	unsigned long mod = tasks_frozen ? CPU_TASKS_FROZEN : 0;
	struct take_cpu_down_param tcd_param = {
		.caller = current,
		.mod = mod,
		.hcpu = hcpu,
	};

	if (num_online_cpus() == 1)
		return -EBUSY;

	if (!cpu_online(cpu))
		return -EINVAL;

	cpu_hotplug_begin();
	err = __cpu_notify(CPU_DOWN_PREPARE | mod, hcpu, -1, &nr_calls);
	if (err) {
		nr_calls--;
		__cpu_notify(CPU_DOWN_FAILED | mod, hcpu, nr_calls, NULL);
		printk("%s: attempt to take down CPU %u failed\n",
				__func__, cpu);
		goto out_release;
	}

	err = __stop_machine(take_cpu_down, &tcd_param, cpumask_of(cpu));
	if (err) {
		/* CPU didn't die: tell everyone.  Can't complain. */
		cpu_notify_nofail(CPU_DOWN_FAILED | mod, hcpu);

		goto out_release;
	}
	BUG_ON(cpu_online(cpu));

	/* Wait for it to sleep (leaving idle task). */
	while (!idle_cpu(cpu))
		yield();

	/* This actually kills the CPU. */
	__cpu_die(cpu);

	/* CPU is completely dead: tell everyone.  Too late to complain. */
	cpu_notify_nofail(CPU_DEAD | mod, hcpu);

	check_for_tasks(cpu);

out_release:
	cpu_hotplug_done();
	if (!err)
		cpu_notify_nofail(CPU_POST_DEAD | mod, hcpu);
	return err;
}

int __ref cpu_down(unsigned int cpu)
{
	int err;

	cpu_maps_update_begin();

	if (cpu_hotplug_disabled) {
		err = -EBUSY;
		goto out;
	}

	err = _cpu_down(cpu, 0);

out:
	cpu_maps_update_done();
	return err;
}
EXPORT_SYMBOL(cpu_down);
#endif /*CONFIG_HOTPLUG_CPU*/

/* Requires cpu_add_remove_lock to be held */
static int __cpuinit _cpu_up(unsigned int cpu, int tasks_frozen)
{
	int ret, nr_calls = 0;
	void *hcpu = (void *)(long)cpu;
	unsigned long mod = tasks_frozen ? CPU_TASKS_FROZEN : 0;

	if (cpu_online(cpu) || !cpu_present(cpu))
		return -EINVAL;

	cpu_hotplug_begin();
	ret = __cpu_notify(CPU_UP_PREPARE | mod, hcpu, -1, &nr_calls);
	if (ret) {
		nr_calls--;
		printk("%s: attempt to bring up CPU %u failed\n",
				__func__, cpu);
		goto out_notify;
	}

	/* Arch-specific enabling code. */
	ret = __cpu_up(cpu);
	if (ret != 0)
		goto out_notify;
	BUG_ON(!cpu_online(cpu));

	/* Now call notifier in preparation. */
	cpu_notify(CPU_ONLINE | mod, hcpu);

out_notify:
	if (ret != 0)
		__cpu_notify(CPU_UP_CANCELED | mod, hcpu, nr_calls, NULL);
	cpu_hotplug_done();

	return ret;
}

int __cpuinit cpu_up(unsigned int cpu)
{
	int err = 0;

#ifdef	CONFIG_MEMORY_HOTPLUG
	int nid;
	pg_data_t	*pgdat;
#endif

	if (!cpu_possible(cpu)) {
		printk(KERN_ERR "can't online cpu %d because it is not "
			"configured as may-hotadd at boot time\n", cpu);
#if defined(CONFIG_IA64)
		printk(KERN_ERR "please check additional_cpus= boot "
				"parameter\n");
#endif
		return -EINVAL;
	}

#ifdef	CONFIG_MEMORY_HOTPLUG
	nid = cpu_to_node(cpu);
	if (!node_online(nid)) {
		err = mem_online_node(nid);
		if (err)
			return err;
	}

	pgdat = NODE_DATA(nid);
	if (!pgdat) {
		printk(KERN_ERR
			"Can't online cpu %d due to NULL pgdat\n", cpu);
		return -ENOMEM;
	}

	if (pgdat->node_zonelists->_zonerefs->zone == NULL) {
		mutex_lock(&zonelists_mutex);
		build_all_zonelists(NULL);
		mutex_unlock(&zonelists_mutex);
	}
#endif

	cpu_maps_update_begin();

	if (cpu_hotplug_disabled) {
		err = -EBUSY;
		goto out;
	}

	err = _cpu_up(cpu, 0);

out:
	cpu_maps_update_done();
	return err;
}

#ifdef CONFIG_PM_SLEEP_SMP
static cpumask_var_t frozen_cpus;

int disable_nonboot_cpus(void)
{
	int cpu, first_cpu, error = 0;

	cpu_maps_update_begin();
	first_cpu = cpumask_first(cpu_online_mask);
	/*
	 * We take down all of the non-boot CPUs in one shot to avoid races
	 * with the userspace trying to use the CPU hotplug at the same time
	 */
	cpumask_clear(frozen_cpus);

	printk("Disabling non-boot CPUs ...\n");
	for_each_online_cpu(cpu) {
		if (cpu == first_cpu)
			continue;
		error = _cpu_down(cpu, 1);
		if (!error)
			cpumask_set_cpu(cpu, frozen_cpus);
		else {
			printk(KERN_ERR "Error taking CPU%d down: %d\n",
				cpu, error);
			break;
		}
	}

	if (!error) {
		BUG_ON(num_online_cpus() > 1);
		/* Make sure the CPUs won't be enabled by someone else */
		cpu_hotplug_disabled = 1;
	} else {
		printk(KERN_ERR "Non-boot CPUs are not disabled\n");
	}
	cpu_maps_update_done();
	return error;
}

void __weak arch_enable_nonboot_cpus_begin(void)
{
}

void __weak arch_enable_nonboot_cpus_end(void)
{
}

void __ref enable_nonboot_cpus(void)
{
	int cpu, error;

	/* Allow everyone to use the CPU hotplug again */
	cpu_maps_update_begin();
	cpu_hotplug_disabled = 0;
	if (cpumask_empty(frozen_cpus))
		goto out;

	printk("Enabling non-boot CPUs ...\n");

	arch_enable_nonboot_cpus_begin();

	for_each_cpu(cpu, frozen_cpus) {
		error = _cpu_up(cpu, 1);
		if (!error) {
			printk("CPU%d is up\n", cpu);
			continue;
		}
		printk(KERN_WARNING "Error taking CPU%d up: %d\n", cpu, error);
	}

	arch_enable_nonboot_cpus_end();

	cpumask_clear(frozen_cpus);
out:
	cpu_maps_update_done();
}

static int alloc_frozen_cpus(void)
{
	if (!alloc_cpumask_var(&frozen_cpus, GFP_KERNEL|__GFP_ZERO))
		return -ENOMEM;
	return 0;
}
core_initcall(alloc_frozen_cpus);
#endif /* CONFIG_PM_SLEEP_SMP */

/**
 * notify_cpu_starting(cpu) - call the CPU_STARTING notifiers
 * @cpu: cpu that just started
 *
 * This function calls the cpu_chain notifiers with CPU_STARTING.
 * It must be called by the arch code on the new cpu, before the new cpu
 * enables interrupts and before the "boot" cpu returns from __cpu_up().
 */
void __cpuinit notify_cpu_starting(unsigned int cpu)
{
	unsigned long val = CPU_STARTING;

#ifdef CONFIG_PM_SLEEP_SMP
	if (frozen_cpus != NULL && cpumask_test_cpu(cpu, frozen_cpus))
		val = CPU_STARTING_FROZEN;
#endif /* CONFIG_PM_SLEEP_SMP */
	cpu_notify(val, (void *)(long)cpu);
}

#endif /* CONFIG_SMP */

/*
 * cpu_bit_bitmap[] is a special, "compressed" data structure that
 * represents all NR_CPUS bits binary values of 1<<nr.
 *
 * It is used by cpumask_of() to get a constant address to a CPU
 * mask value that has a single bit set only.
 */

/* cpu_bit_bitmap[0] is empty - so we can back into it */
#define MASK_DECLARE_1(x)	[x+1][0] = 1UL << (x)
#define MASK_DECLARE_2(x)	MASK_DECLARE_1(x), MASK_DECLARE_1(x+1)
#define MASK_DECLARE_4(x)	MASK_DECLARE_2(x), MASK_DECLARE_2(x+2)
#define MASK_DECLARE_8(x)	MASK_DECLARE_4(x), MASK_DECLARE_4(x+4)

const unsigned long cpu_bit_bitmap[BITS_PER_LONG+1][BITS_TO_LONGS(NR_CPUS)] = {

	MASK_DECLARE_8(0),	MASK_DECLARE_8(8),
	MASK_DECLARE_8(16),	MASK_DECLARE_8(24),
#if BITS_PER_LONG > 32
	MASK_DECLARE_8(32),	MASK_DECLARE_8(40),
	MASK_DECLARE_8(48),	MASK_DECLARE_8(56),
#endif
};
EXPORT_SYMBOL_GPL(cpu_bit_bitmap);

const DECLARE_BITMAP(cpu_all_bits, NR_CPUS) = CPU_BITS_ALL;
EXPORT_SYMBOL(cpu_all_bits);

#ifdef CONFIG_INIT_ALL_POSSIBLE
static DECLARE_BITMAP(cpu_possible_bits, CONFIG_NR_CPUS) __read_mostly
	= CPU_BITS_ALL;
#else
static DECLARE_BITMAP(cpu_possible_bits, CONFIG_NR_CPUS) __read_mostly;
#endif
const struct cpumask *const cpu_possible_mask = to_cpumask(cpu_possible_bits);
EXPORT_SYMBOL(cpu_possible_mask);

static DECLARE_BITMAP(cpu_online_bits, CONFIG_NR_CPUS) __read_mostly;
const struct cpumask *const cpu_online_mask = to_cpumask(cpu_online_bits);
EXPORT_SYMBOL(cpu_online_mask);

static DECLARE_BITMAP(cpu_present_bits, CONFIG_NR_CPUS) __read_mostly;
const struct cpumask *const cpu_present_mask = to_cpumask(cpu_present_bits);
EXPORT_SYMBOL(cpu_present_mask);

static DECLARE_BITMAP(cpu_active_bits, CONFIG_NR_CPUS) __read_mostly;
const struct cpumask *const cpu_active_mask = to_cpumask(cpu_active_bits);
EXPORT_SYMBOL(cpu_active_mask);

void set_cpu_possible(unsigned int cpu, bool possible)
{
	if (possible)
		cpumask_set_cpu(cpu, to_cpumask(cpu_possible_bits));
	else
		cpumask_clear_cpu(cpu, to_cpumask(cpu_possible_bits));
}

void set_cpu_present(unsigned int cpu, bool present)
{
	if (present)
		cpumask_set_cpu(cpu, to_cpumask(cpu_present_bits));
	else
		cpumask_clear_cpu(cpu, to_cpumask(cpu_present_bits));
}

void set_cpu_online(unsigned int cpu, bool online)
{
	if (online)
		cpumask_set_cpu(cpu, to_cpumask(cpu_online_bits));
	else
		cpumask_clear_cpu(cpu, to_cpumask(cpu_online_bits));
}

void set_cpu_active(unsigned int cpu, bool active)
{
	if (active)
		cpumask_set_cpu(cpu, to_cpumask(cpu_active_bits));
	else
		cpumask_clear_cpu(cpu, to_cpumask(cpu_active_bits));
}

void init_cpu_present(const struct cpumask *src)
{
	cpumask_copy(to_cpumask(cpu_present_bits), src);
}

void init_cpu_possible(const struct cpumask *src)
{
	cpumask_copy(to_cpumask(cpu_possible_bits), src);
}

void init_cpu_online(const struct cpumask *src)
{
	cpumask_copy(to_cpumask(cpu_online_bits), src);
}
Commit	Line	Data
1da177e4 LT	1	/* CPU control.
	2	* (C) 2001, 2002, 2003, 2004 Rusty Russell
	3	*
	4	* This code is licenced under the GPL.
	5	*/
	6	#include <linux/proc_fs.h>
	7	#include <linux/smp.h>
	8	#include <linux/init.h>
	9	#include <linux/notifier.h>
	10	#include <linux/sched.h>
	11	#include <linux/unistd.h>
	12	#include <linux/cpu.h>
	13	#include <linux/module.h>
	14	#include <linux/kthread.h>
	15	#include <linux/stop_machine.h>
81615b62	16	#include <linux/mutex.h>
5a0e3ad6	17	#include <linux/gfp.h>
1da177e4	18
98a79d6a	19	#ifdef CONFIG_SMP
b3199c02	20	/* Serializes the updates to cpu_online_mask, cpu_present_mask */
aa953877	21	static DEFINE_MUTEX(cpu_add_remove_lock);
1da177e4	22
79a6cdeb LJ	23	/*
	24	* The following two API's must be used when attempting
	25	* to serialize the updates to cpu_online_mask, cpu_present_mask.
	26	*/
	27	void cpu_maps_update_begin(void)
	28	{
	29	mutex_lock(&cpu_add_remove_lock);
	30	}
	31
	32	void cpu_maps_update_done(void)
	33	{
	34	mutex_unlock(&cpu_add_remove_lock);
	35	}
	36
5c113fbe	37	static RAW_NOTIFIER_HEAD(cpu_chain);
1da177e4	38
e3920fb4 RW	39	/* If set, cpu_up and cpu_down will return -EBUSY and do nothing.
	40	* Should always be manipulated under cpu_add_remove_lock
	41	*/
	42	static int cpu_hotplug_disabled;
	43
79a6cdeb LJ	44	#ifdef CONFIG_HOTPLUG_CPU
79a6cdeb LJ	45
d221938c GS	46	static struct {
	47	struct task_struct *active_writer;
	48	struct mutex lock; /* Synchronizes accesses to refcount, */
	49	/*
	50	* Also blocks the new readers during
	51	* an ongoing cpu hotplug operation.
	52	*/
	53	int refcount;
31950eb6 LT	54	} cpu_hotplug = {
	55	.active_writer = NULL,
	56	.lock = __MUTEX_INITIALIZER(cpu_hotplug.lock),
	57	.refcount = 0,
	58	};
d221938c	59
86ef5c9a	60	void get_online_cpus(void)
a9d9baa1	61	{
d221938c GS	62	might_sleep();
d221938c GS	63	if (cpu_hotplug.active_writer == current)
aa953877	64	return;
d221938c GS	65	mutex_lock(&cpu_hotplug.lock);
	66	cpu_hotplug.refcount++;
	67	mutex_unlock(&cpu_hotplug.lock);
	68
a9d9baa1	69	}
86ef5c9a	70	EXPORT_SYMBOL_GPL(get_online_cpus);
90d45d17	71
86ef5c9a	72	void put_online_cpus(void)
a9d9baa1	73	{
d221938c	74	if (cpu_hotplug.active_writer == current)
aa953877	75	return;
d221938c	76	mutex_lock(&cpu_hotplug.lock);
d2ba7e2a ON	77	if (!--cpu_hotplug.refcount && unlikely(cpu_hotplug.active_writer))
d2ba7e2a ON	78	wake_up_process(cpu_hotplug.active_writer);
d221938c GS	79	mutex_unlock(&cpu_hotplug.lock);
d221938c GS	80
a9d9baa1	81	}
86ef5c9a	82	EXPORT_SYMBOL_GPL(put_online_cpus);
a9d9baa1	83
d221938c GS	84	/*
	85	* This ensures that the hotplug operation can begin only when the
	86	* refcount goes to zero.
	87	*
	88	* Note that during a cpu-hotplug operation, the new readers, if any,
	89	* will be blocked by the cpu_hotplug.lock
	90	*
d2ba7e2a ON	91	* Since cpu_hotplug_begin() is always called after invoking
d2ba7e2a ON	92	* cpu_maps_update_begin(), we can be sure that only one writer is active.
d221938c GS	93	*
	94	* Note that theoretically, there is a possibility of a livelock:
	95	* - Refcount goes to zero, last reader wakes up the sleeping
	96	* writer.
	97	* - Last reader unlocks the cpu_hotplug.lock.
	98	* - A new reader arrives at this moment, bumps up the refcount.
	99	* - The writer acquires the cpu_hotplug.lock finds the refcount
	100	* non zero and goes to sleep again.
	101	*
	102	* However, this is very difficult to achieve in practice since
86ef5c9a	103	* get_online_cpus() not an api which is called all that often.
d221938c GS	104	*
	105	*/
	106	static void cpu_hotplug_begin(void)
	107	{
d221938c	108	cpu_hotplug.active_writer = current;
d2ba7e2a ON	109
	110	for (;;) {
	111	mutex_lock(&cpu_hotplug.lock);
	112	if (likely(!cpu_hotplug.refcount))
	113	break;
	114	__set_current_state(TASK_UNINTERRUPTIBLE);
d221938c GS	115	mutex_unlock(&cpu_hotplug.lock);
d221938c GS	116	schedule();
d221938c	117	}
d221938c GS	118	}
	119
	120	static void cpu_hotplug_done(void)
	121	{
	122	cpu_hotplug.active_writer = NULL;
	123	mutex_unlock(&cpu_hotplug.lock);
	124	}
79a6cdeb LJ	125
	126	#else /* #if CONFIG_HOTPLUG_CPU */
	127	static void cpu_hotplug_begin(void) {}
	128	static void cpu_hotplug_done(void) {}
	129	#endif /* #esle #if CONFIG_HOTPLUG_CPU */
	130
1da177e4	131	/* Need to know about CPUs going up/down? */
f7b16c10	132	int __ref register_cpu_notifier(struct notifier_block *nb)
1da177e4	133	{
bd5349cf	134	int ret;
d221938c	135	cpu_maps_update_begin();
bd5349cf	136	ret = raw_notifier_chain_register(&cpu_chain, nb);
d221938c	137	cpu_maps_update_done();
bd5349cf	138	return ret;
1da177e4	139	}
65edc68c	140
e9fb7631 AM	141	static int __cpu_notify(unsigned long val, void *v, int nr_to_call,
	142	int *nr_calls)
	143	{
e6bde73b AM	144	int ret;
	145
	146	ret = __raw_notifier_call_chain(&cpu_chain, val, v, nr_to_call,
e9fb7631	147	nr_calls);
e6bde73b AM	148
e6bde73b AM	149	return notifier_to_errno(ret);
e9fb7631 AM	150	}
	151
	152	static int cpu_notify(unsigned long val, void *v)
	153	{
	154	return __cpu_notify(val, v, -1, NULL);
	155	}
	156
00b9b0af LT	157	#ifdef CONFIG_HOTPLUG_CPU
00b9b0af LT	158
e9fb7631 AM	159	static void cpu_notify_nofail(unsigned long val, void *v)
e9fb7631 AM	160	{
00b9b0af	161	BUG_ON(cpu_notify(val, v));
e9fb7631 AM	162	}
e9fb7631 AM	163
1da177e4 LT	164	EXPORT_SYMBOL(register_cpu_notifier);
1da177e4 LT	165
9647155f	166	void __ref unregister_cpu_notifier(struct notifier_block *nb)
1da177e4	167	{
d221938c	168	cpu_maps_update_begin();
bd5349cf	169	raw_notifier_chain_unregister(&cpu_chain, nb);
d221938c	170	cpu_maps_update_done();
1da177e4 LT	171	}
	172	EXPORT_SYMBOL(unregister_cpu_notifier);
	173
1da177e4 LT	174	static inline void check_for_tasks(int cpu)
	175	{
	176	struct task_struct *p;
	177
	178	write_lock_irq(&tasklist_lock);
	179	for_each_process(p) {
11854247	180	if (task_cpu(p) == cpu && p->state == TASK_RUNNING &&
1da177e4 LT	181	(!cputime_eq(p->utime, cputime_zero) \|\|
1da177e4 LT	182	!cputime_eq(p->stime, cputime_zero)))
9d3cfc4c FP	183	printk(KERN_WARNING "Task %s (pid = %d) is on cpu %d "
	184	"(state = %ld, flags = %x)\n",
	185	p->comm, task_pid_nr(p), cpu,
	186	p->state, p->flags);
1da177e4 LT	187	}
	188	write_unlock_irq(&tasklist_lock);
	189	}
	190
db912f96	191	struct take_cpu_down_param {
6a1bdc1b	192	struct task_struct *caller;
db912f96 AK	193	unsigned long mod;
	194	void *hcpu;
	195	};
	196
1da177e4	197	/* Take this CPU down. */
514a20a5	198	static int __ref take_cpu_down(void *_param)
1da177e4	199	{
db912f96	200	struct take_cpu_down_param *param = _param;
6a1bdc1b	201	unsigned int cpu = (unsigned long)param->hcpu;
1da177e4 LT	202	int err;
1da177e4 LT	203
1da177e4 LT	204	/* Ensure this CPU doesn't handle any more interrupts. */
	205	err = __cpu_disable();
	206	if (err < 0)
f3705136	207	return err;
1da177e4	208
e9fb7631	209	cpu_notify(CPU_DYING \| param->mod, param->hcpu);
3ba35573	210
6a1bdc1b ON	211	if (task_cpu(param->caller) == cpu)
6a1bdc1b ON	212	move_task_off_dead_cpu(cpu, param->caller);
f3705136 ZM	213	/* Force idle task to run as soon as we yield: it should
	214	immediately notice cpu is offline and die quickly. */
	215	sched_idle_next();
	216	return 0;
1da177e4 LT	217	}
1da177e4 LT	218
e3920fb4	219	/* Requires cpu_add_remove_lock to be held */
514a20a5	220	static int __ref _cpu_down(unsigned int cpu, int tasks_frozen)
1da177e4	221	{
e7407dcc	222	int err, nr_calls = 0;
e7407dcc	223	void hcpu = (void )(long)cpu;
8bb78442	224	unsigned long mod = tasks_frozen ? CPU_TASKS_FROZEN : 0;
db912f96	225	struct take_cpu_down_param tcd_param = {
6a1bdc1b	226	.caller = current,
db912f96 AK	227	.mod = mod,
	228	.hcpu = hcpu,
	229	};
1da177e4	230
e3920fb4 RW	231	if (num_online_cpus() == 1)
e3920fb4 RW	232	return -EBUSY;
1da177e4	233
e3920fb4 RW	234	if (!cpu_online(cpu))
e3920fb4 RW	235	return -EINVAL;
1da177e4	236
d221938c	237	cpu_hotplug_begin();
e9fb7631	238	err = __cpu_notify(CPU_DOWN_PREPARE \| mod, hcpu, -1, &nr_calls);
e6bde73b	239	if (err) {
a0d8cdb6	240	nr_calls--;
e9fb7631	241	__cpu_notify(CPU_DOWN_FAILED \| mod, hcpu, nr_calls, NULL);
1da177e4	242	printk("%s: attempt to take down CPU %u failed\n",
af1f16d0	243	__func__, cpu);
baaca49f	244	goto out_release;
1da177e4 LT	245	}
1da177e4 LT	246
e0b582ec	247	err = __stop_machine(take_cpu_down, &tcd_param, cpumask_of(cpu));
04321587	248	if (err) {
1da177e4	249	/* CPU didn't die: tell everyone. Can't complain. */
e9fb7631	250	cpu_notify_nofail(CPU_DOWN_FAILED \| mod, hcpu);
1da177e4	251
6a1bdc1b	252	goto out_release;
8fa1d7d3	253	}
04321587	254	BUG_ON(cpu_online(cpu));
1da177e4 LT	255
	256	/* Wait for it to sleep (leaving idle task). */
	257	while (!idle_cpu(cpu))
	258	yield();
	259
	260	/* This actually kills the CPU. */
	261	__cpu_die(cpu);
	262
1da177e4	263	/* CPU is completely dead: tell everyone. Too late to complain. */
e9fb7631	264	cpu_notify_nofail(CPU_DEAD \| mod, hcpu);
1da177e4 LT	265
	266	check_for_tasks(cpu);
	267
baaca49f	268	out_release:
d221938c	269	cpu_hotplug_done();
e9fb7631 AM	270	if (!err)
e9fb7631 AM	271	cpu_notify_nofail(CPU_POST_DEAD \| mod, hcpu);
e3920fb4 RW	272	return err;
	273	}
	274
514a20a5	275	int __ref cpu_down(unsigned int cpu)
e3920fb4	276	{
9ea09af3	277	int err;
e3920fb4	278
d221938c	279	cpu_maps_update_begin();
e761b772 MK	280
e761b772 MK	281	if (cpu_hotplug_disabled) {
e3920fb4	282	err = -EBUSY;
e761b772 MK	283	goto out;
	284	}
	285
e761b772	286	err = _cpu_down(cpu, 0);
e3920fb4	287
e761b772	288	out:
d221938c	289	cpu_maps_update_done();
1da177e4 LT	290	return err;
1da177e4 LT	291	}
b62b8ef9	292	EXPORT_SYMBOL(cpu_down);
1da177e4 LT	293	#endif /CONFIG_HOTPLUG_CPU/
1da177e4 LT	294
e3920fb4	295	/* Requires cpu_add_remove_lock to be held */
8bb78442	296	static int __cpuinit _cpu_up(unsigned int cpu, int tasks_frozen)
1da177e4	297	{
baaca49f	298	int ret, nr_calls = 0;
1da177e4	299	void hcpu = (void )(long)cpu;
8bb78442	300	unsigned long mod = tasks_frozen ? CPU_TASKS_FROZEN : 0;
1da177e4	301
e3920fb4 RW	302	if (cpu_online(cpu) \|\| !cpu_present(cpu))
e3920fb4 RW	303	return -EINVAL;
90d45d17	304
d221938c	305	cpu_hotplug_begin();
e9fb7631	306	ret = __cpu_notify(CPU_UP_PREPARE \| mod, hcpu, -1, &nr_calls);
e6bde73b	307	if (ret) {
a0d8cdb6	308	nr_calls--;
1da177e4	309	printk("%s: attempt to bring up CPU %u failed\n",
af1f16d0	310	__func__, cpu);
1da177e4 LT	311	goto out_notify;
	312	}
	313
	314	/* Arch-specific enabling code. */
	315	ret = __cpu_up(cpu);
	316	if (ret != 0)
	317	goto out_notify;
6978c705	318	BUG_ON(!cpu_online(cpu));
1da177e4 LT	319
1da177e4 LT	320	/* Now call notifier in preparation. */
e9fb7631	321	cpu_notify(CPU_ONLINE \| mod, hcpu);
1da177e4 LT	322
	323	out_notify:
	324	if (ret != 0)
e9fb7631	325	__cpu_notify(CPU_UP_CANCELED \| mod, hcpu, nr_calls, NULL);
d221938c	326	cpu_hotplug_done();
e3920fb4 RW	327
	328	return ret;
	329	}
	330
b282b6f8	331	int __cpuinit cpu_up(unsigned int cpu)
e3920fb4 RW	332	{
e3920fb4 RW	333	int err = 0;
cf23422b	334
	335	#ifdef CONFIG_MEMORY_HOTPLUG
	336	int nid;
	337	pg_data_t *pgdat;
	338	#endif
	339
e0b582ec	340	if (!cpu_possible(cpu)) {
73e753a5 KH	341	printk(KERN_ERR "can't online cpu %d because it is not "
73e753a5 KH	342	"configured as may-hotadd at boot time\n", cpu);
87d5e023	343	#if defined(CONFIG_IA64)
73e753a5 KH	344	printk(KERN_ERR "please check additional_cpus= boot "
	345	"parameter\n");
	346	#endif
	347	return -EINVAL;
	348	}
e3920fb4	349
cf23422b	350	#ifdef CONFIG_MEMORY_HOTPLUG
	351	nid = cpu_to_node(cpu);
	352	if (!node_online(nid)) {
	353	err = mem_online_node(nid);
	354	if (err)
	355	return err;
	356	}
	357
	358	pgdat = NODE_DATA(nid);
	359	if (!pgdat) {
	360	printk(KERN_ERR
	361	"Can't online cpu %d due to NULL pgdat\n", cpu);
	362	return -ENOMEM;
	363	}
	364
4eaf3f64 HL	365	if (pgdat->node_zonelists->_zonerefs->zone == NULL) {
4eaf3f64 HL	366	mutex_lock(&zonelists_mutex);
1f522509	367	build_all_zonelists(NULL);
4eaf3f64 HL	368	mutex_unlock(&zonelists_mutex);
4eaf3f64 HL	369	}
cf23422b	370	#endif
cf23422b	371
d221938c	372	cpu_maps_update_begin();
e761b772 MK	373
e761b772 MK	374	if (cpu_hotplug_disabled) {
e3920fb4	375	err = -EBUSY;
e761b772 MK	376	goto out;
	377	}
	378
	379	err = _cpu_up(cpu, 0);
	380
e761b772	381	out:
d221938c	382	cpu_maps_update_done();
e3920fb4 RW	383	return err;
	384	}
	385
f3de4be9	386	#ifdef CONFIG_PM_SLEEP_SMP
e0b582ec	387	static cpumask_var_t frozen_cpus;
e3920fb4 RW	388
	389	int disable_nonboot_cpus(void)
	390	{
e9a5f426	391	int cpu, first_cpu, error = 0;
e3920fb4	392
d221938c	393	cpu_maps_update_begin();
e0b582ec	394	first_cpu = cpumask_first(cpu_online_mask);
9ee349ad XF	395	/*
9ee349ad XF	396	* We take down all of the non-boot CPUs in one shot to avoid races
e3920fb4 RW	397	* with the userspace trying to use the CPU hotplug at the same time
e3920fb4 RW	398	*/
e0b582ec	399	cpumask_clear(frozen_cpus);
6ad4c188	400
e3920fb4 RW	401	printk("Disabling non-boot CPUs ...\n");
	402	for_each_online_cpu(cpu) {
	403	if (cpu == first_cpu)
	404	continue;
8bb78442	405	error = _cpu_down(cpu, 1);
feae3203	406	if (!error)
e0b582ec	407	cpumask_set_cpu(cpu, frozen_cpus);
feae3203	408	else {
e3920fb4 RW	409	printk(KERN_ERR "Error taking CPU%d down: %d\n",
	410	cpu, error);
	411	break;
	412	}
	413	}
86886e55	414
e3920fb4 RW	415	if (!error) {
	416	BUG_ON(num_online_cpus() > 1);
	417	/* Make sure the CPUs won't be enabled by someone else */
	418	cpu_hotplug_disabled = 1;
	419	} else {
e1d9fd2e	420	printk(KERN_ERR "Non-boot CPUs are not disabled\n");
e3920fb4	421	}
d221938c	422	cpu_maps_update_done();
e3920fb4 RW	423	return error;
	424	}
	425
d0af9eed SS	426	void __weak arch_enable_nonboot_cpus_begin(void)
	427	{
	428	}
	429
	430	void __weak arch_enable_nonboot_cpus_end(void)
	431	{
	432	}
	433
fa7303e2	434	void __ref enable_nonboot_cpus(void)
e3920fb4 RW	435	{
	436	int cpu, error;
	437
	438	/* Allow everyone to use the CPU hotplug again */
d221938c	439	cpu_maps_update_begin();
e3920fb4	440	cpu_hotplug_disabled = 0;
e0b582ec	441	if (cpumask_empty(frozen_cpus))
1d64b9cb	442	goto out;
e3920fb4 RW	443
e3920fb4 RW	444	printk("Enabling non-boot CPUs ...\n");
d0af9eed SS	445
	446	arch_enable_nonboot_cpus_begin();
	447
e0b582ec	448	for_each_cpu(cpu, frozen_cpus) {
8bb78442	449	error = _cpu_up(cpu, 1);
e3920fb4 RW	450	if (!error) {
	451	printk("CPU%d is up\n", cpu);
	452	continue;
	453	}
1d64b9cb	454	printk(KERN_WARNING "Error taking CPU%d up: %d\n", cpu, error);
e3920fb4	455	}
d0af9eed SS	456
	457	arch_enable_nonboot_cpus_end();
	458
e0b582ec	459	cpumask_clear(frozen_cpus);
1d64b9cb	460	out:
d221938c	461	cpu_maps_update_done();
1da177e4	462	}
e0b582ec RR	463
	464	static int alloc_frozen_cpus(void)
	465	{
	466	if (!alloc_cpumask_var(&frozen_cpus, GFP_KERNEL\|__GFP_ZERO))
	467	return -ENOMEM;
	468	return 0;
	469	}
	470	core_initcall(alloc_frozen_cpus);
f3de4be9	471	#endif /* CONFIG_PM_SLEEP_SMP */
68f4f1ec	472
e545a614 MS	473	/**
	474	* notify_cpu_starting(cpu) - call the CPU_STARTING notifiers
	475	* @cpu: cpu that just started
	476	*
	477	* This function calls the cpu_chain notifiers with CPU_STARTING.
	478	* It must be called by the arch code on the new cpu, before the new cpu
	479	* enables interrupts and before the "boot" cpu returns from __cpu_up().
	480	*/
84196414	481	void __cpuinit notify_cpu_starting(unsigned int cpu)
e545a614 MS	482	{
	483	unsigned long val = CPU_STARTING;
	484
	485	#ifdef CONFIG_PM_SLEEP_SMP
e0b582ec	486	if (frozen_cpus != NULL && cpumask_test_cpu(cpu, frozen_cpus))
e545a614 MS	487	val = CPU_STARTING_FROZEN;
e545a614 MS	488	#endif /* CONFIG_PM_SLEEP_SMP */
e9fb7631	489	cpu_notify(val, (void *)(long)cpu);
e545a614 MS	490	}
e545a614 MS	491
68f4f1ec	492	#endif /* CONFIG_SMP */
b8d317d1	493
e56b3bc7 LT	494	/*
	495	* cpu_bit_bitmap[] is a special, "compressed" data structure that
	496	* represents all NR_CPUS bits binary values of 1<<nr.
	497	*
e0b582ec	498	* It is used by cpumask_of() to get a constant address to a CPU
e56b3bc7 LT	499	* mask value that has a single bit set only.
e56b3bc7 LT	500	*/
b8d317d1	501
e56b3bc7 LT	502	/* cpu_bit_bitmap[0] is empty - so we can back into it */
	503	#define MASK_DECLARE_1(x) [x+1][0] = 1UL << (x)
	504	#define MASK_DECLARE_2(x) MASK_DECLARE_1(x), MASK_DECLARE_1(x+1)
	505	#define MASK_DECLARE_4(x) MASK_DECLARE_2(x), MASK_DECLARE_2(x+2)
	506	#define MASK_DECLARE_8(x) MASK_DECLARE_4(x), MASK_DECLARE_4(x+4)
b8d317d1	507
e56b3bc7 LT	508	const unsigned long cpu_bit_bitmap[BITS_PER_LONG+1][BITS_TO_LONGS(NR_CPUS)] = {
	509
	510	MASK_DECLARE_8(0), MASK_DECLARE_8(8),
	511	MASK_DECLARE_8(16), MASK_DECLARE_8(24),
	512	#if BITS_PER_LONG > 32
	513	MASK_DECLARE_8(32), MASK_DECLARE_8(40),
	514	MASK_DECLARE_8(48), MASK_DECLARE_8(56),
b8d317d1 MT	515	#endif
b8d317d1 MT	516	};
e56b3bc7	517	EXPORT_SYMBOL_GPL(cpu_bit_bitmap);
2d3854a3 RR	518
	519	const DECLARE_BITMAP(cpu_all_bits, NR_CPUS) = CPU_BITS_ALL;
	520	EXPORT_SYMBOL(cpu_all_bits);
b3199c02 RR	521
	522	#ifdef CONFIG_INIT_ALL_POSSIBLE
	523	static DECLARE_BITMAP(cpu_possible_bits, CONFIG_NR_CPUS) __read_mostly
	524	= CPU_BITS_ALL;
	525	#else
	526	static DECLARE_BITMAP(cpu_possible_bits, CONFIG_NR_CPUS) __read_mostly;
	527	#endif
	528	const struct cpumask *const cpu_possible_mask = to_cpumask(cpu_possible_bits);
	529	EXPORT_SYMBOL(cpu_possible_mask);
	530
	531	static DECLARE_BITMAP(cpu_online_bits, CONFIG_NR_CPUS) __read_mostly;
	532	const struct cpumask *const cpu_online_mask = to_cpumask(cpu_online_bits);
	533	EXPORT_SYMBOL(cpu_online_mask);
	534
	535	static DECLARE_BITMAP(cpu_present_bits, CONFIG_NR_CPUS) __read_mostly;
	536	const struct cpumask *const cpu_present_mask = to_cpumask(cpu_present_bits);
	537	EXPORT_SYMBOL(cpu_present_mask);
	538
	539	static DECLARE_BITMAP(cpu_active_bits, CONFIG_NR_CPUS) __read_mostly;
	540	const struct cpumask *const cpu_active_mask = to_cpumask(cpu_active_bits);
	541	EXPORT_SYMBOL(cpu_active_mask);
3fa41520 RR	542
	543	void set_cpu_possible(unsigned int cpu, bool possible)
	544	{
	545	if (possible)
	546	cpumask_set_cpu(cpu, to_cpumask(cpu_possible_bits));
	547	else
	548	cpumask_clear_cpu(cpu, to_cpumask(cpu_possible_bits));
	549	}
	550
	551	void set_cpu_present(unsigned int cpu, bool present)
	552	{
	553	if (present)
	554	cpumask_set_cpu(cpu, to_cpumask(cpu_present_bits));
	555	else
	556	cpumask_clear_cpu(cpu, to_cpumask(cpu_present_bits));
	557	}
	558
	559	void set_cpu_online(unsigned int cpu, bool online)
	560	{
	561	if (online)
	562	cpumask_set_cpu(cpu, to_cpumask(cpu_online_bits));
	563	else
	564	cpumask_clear_cpu(cpu, to_cpumask(cpu_online_bits));
	565	}
	566
	567	void set_cpu_active(unsigned int cpu, bool active)
	568	{
	569	if (active)
	570	cpumask_set_cpu(cpu, to_cpumask(cpu_active_bits));
	571	else
	572	cpumask_clear_cpu(cpu, to_cpumask(cpu_active_bits));
	573	}
	574
	575	void init_cpu_present(const struct cpumask *src)
	576	{
	577	cpumask_copy(to_cpumask(cpu_present_bits), src);
	578	}
	579
	580	void init_cpu_possible(const struct cpumask *src)
	581	{
	582	cpumask_copy(to_cpumask(cpu_possible_bits), src);
	583	}
	584
	585	void init_cpu_online(const struct cpumask *src)
	586	{
	587	cpumask_copy(to_cpumask(cpu_online_bits), src);
	588	}