[linux-2.6-block.git] / kernel / pid.c

/*
 * Generic pidhash and scalable, time-bounded PID allocator
 *
 * (C) 2002-2003 William Irwin, IBM
 * (C) 2004 William Irwin, Oracle
 * (C) 2002-2004 Ingo Molnar, Red Hat
 *
 * pid-structures are backing objects for tasks sharing a given ID to chain
 * against. There is very little to them aside from hashing them and
 * parking tasks using given ID's on a list.
 *
 * The hash is always changed with the tasklist_lock write-acquired,
 * and the hash is only accessed with the tasklist_lock at least
 * read-acquired, so there's no additional SMP locking needed here.
 *
 * We have a list of bitmap pages, which bitmaps represent the PID space.
 * Allocating and freeing PIDs is completely lockless. The worst-case
 * allocation scenario when all but one out of 1 million PIDs possible are
 * allocated already: the scanning of 32 list entries and at most PAGE_SIZE
 * bytes. The typical fastpath is a single successful setbit. Freeing is O(1).
 */

#include <linux/mm.h>
#include <linux/module.h>
#include <linux/slab.h>
#include <linux/init.h>
#include <linux/bootmem.h>
#include <linux/hash.h>
#include <linux/pid_namespace.h>

#define pid_hashfn(nr) hash_long((unsigned long)nr, pidhash_shift)
static struct hlist_head *pid_hash;
static int pidhash_shift;
static struct kmem_cache *pid_cachep;

int pid_max = PID_MAX_DEFAULT;

#define RESERVED_PIDS		300

int pid_max_min = RESERVED_PIDS + 1;
int pid_max_max = PID_MAX_LIMIT;

#define BITS_PER_PAGE		(PAGE_SIZE*8)
#define BITS_PER_PAGE_MASK	(BITS_PER_PAGE-1)

static inline int mk_pid(struct pid_namespace *pid_ns,
		struct pidmap *map, int off)
{
	return (map - pid_ns->pidmap)*BITS_PER_PAGE + off;
}

#define find_next_offset(map, off)					\
		find_next_zero_bit((map)->page, BITS_PER_PAGE, off)

/*
 * PID-map pages start out as NULL, they get allocated upon
 * first use and are never deallocated. This way a low pid_max
 * value does not cause lots of bitmaps to be allocated, but
 * the scheme scales to up to 4 million PIDs, runtime.
 */
struct pid_namespace init_pid_ns = {
	.kref = {
		.refcount       = ATOMIC_INIT(2),
	},
	.pidmap = {
		[ 0 ... PIDMAP_ENTRIES-1] = { ATOMIC_INIT(BITS_PER_PAGE), NULL }
	},
	.last_pid = 0
};

/*
 * Note: disable interrupts while the pidmap_lock is held as an
 * interrupt might come in and do read_lock(&tasklist_lock).
 *
 * If we don't disable interrupts there is a nasty deadlock between
 * detach_pid()->free_pid() and another cpu that does
 * spin_lock(&pidmap_lock) followed by an interrupt routine that does
 * read_lock(&tasklist_lock);
 *
 * After we clean up the tasklist_lock and know there are no
 * irq handlers that take it we can leave the interrupts enabled.
 * For now it is easier to be safe than to prove it can't happen.
 */

static  __cacheline_aligned_in_smp DEFINE_SPINLOCK(pidmap_lock);

static fastcall void free_pidmap(struct pid_namespace *pid_ns, int pid)
{
	struct pidmap *map = pid_ns->pidmap + pid / BITS_PER_PAGE;
	int offset = pid & BITS_PER_PAGE_MASK;

	clear_bit(offset, map->page);
	atomic_inc(&map->nr_free);
}

static int alloc_pidmap(struct pid_namespace *pid_ns)
{
	int i, offset, max_scan, pid, last = pid_ns->last_pid;
	struct pidmap *map;

	pid = last + 1;
	if (pid >= pid_max)
		pid = RESERVED_PIDS;
	offset = pid & BITS_PER_PAGE_MASK;
	map = &pid_ns->pidmap[pid/BITS_PER_PAGE];
	max_scan = (pid_max + BITS_PER_PAGE - 1)/BITS_PER_PAGE - !offset;
	for (i = 0; i <= max_scan; ++i) {
		if (unlikely(!map->page)) {
			void *page = kzalloc(PAGE_SIZE, GFP_KERNEL);
			/*
			 * Free the page if someone raced with us
			 * installing it:
			 */
			spin_lock_irq(&pidmap_lock);
			if (map->page)
				kfree(page);
			else
				map->page = page;
			spin_unlock_irq(&pidmap_lock);
			if (unlikely(!map->page))
				break;
		}
		if (likely(atomic_read(&map->nr_free))) {
			do {
				if (!test_and_set_bit(offset, map->page)) {
					atomic_dec(&map->nr_free);
					pid_ns->last_pid = pid;
					return pid;
				}
				offset = find_next_offset(map, offset);
				pid = mk_pid(pid_ns, map, offset);
			/*
			 * find_next_offset() found a bit, the pid from it
			 * is in-bounds, and if we fell back to the last
			 * bitmap block and the final block was the same
			 * as the starting point, pid is before last_pid.
			 */
			} while (offset < BITS_PER_PAGE && pid < pid_max &&
					(i != max_scan || pid < last ||
					    !((last+1) & BITS_PER_PAGE_MASK)));
		}
		if (map < &pid_ns->pidmap[(pid_max-1)/BITS_PER_PAGE]) {
			++map;
			offset = 0;
		} else {
			map = &pid_ns->pidmap[0];
			offset = RESERVED_PIDS;
			if (unlikely(last == offset))
				break;
		}
		pid = mk_pid(pid_ns, map, offset);
	}
	return -1;
}

static int next_pidmap(struct pid_namespace *pid_ns, int last)
{
	int offset;
	struct pidmap *map, *end;

	offset = (last + 1) & BITS_PER_PAGE_MASK;
	map = &pid_ns->pidmap[(last + 1)/BITS_PER_PAGE];
	end = &pid_ns->pidmap[PIDMAP_ENTRIES];
	for (; map < end; map++, offset = 0) {
		if (unlikely(!map->page))
			continue;
		offset = find_next_bit((map)->page, BITS_PER_PAGE, offset);
		if (offset < BITS_PER_PAGE)
			return mk_pid(pid_ns, map, offset);
	}
	return -1;
}

fastcall void put_pid(struct pid *pid)
{
	if (!pid)
		return;
	if ((atomic_read(&pid->count) == 1) ||
	     atomic_dec_and_test(&pid->count))
		kmem_cache_free(pid_cachep, pid);
}
EXPORT_SYMBOL_GPL(put_pid);

static void delayed_put_pid(struct rcu_head *rhp)
{
	struct pid *pid = container_of(rhp, struct pid, rcu);
	put_pid(pid);
}

fastcall void free_pid(struct pid *pid)
{
	/* We can be called with write_lock_irq(&tasklist_lock) held */
	unsigned long flags;

	spin_lock_irqsave(&pidmap_lock, flags);
	hlist_del_rcu(&pid->pid_chain);
	spin_unlock_irqrestore(&pidmap_lock, flags);

	free_pidmap(current->nsproxy->pid_ns, pid->nr);
	call_rcu(&pid->rcu, delayed_put_pid);
}

struct pid *alloc_pid(void)
{
	struct pid *pid;
	enum pid_type type;
	int nr = -1;

	pid = kmem_cache_alloc(pid_cachep, GFP_KERNEL);
	if (!pid)
		goto out;

	nr = alloc_pidmap(current->nsproxy->pid_ns);
	if (nr < 0)
		goto out_free;

	atomic_set(&pid->count, 1);
	pid->nr = nr;
	for (type = 0; type < PIDTYPE_MAX; ++type)
		INIT_HLIST_HEAD(&pid->tasks[type]);

	spin_lock_irq(&pidmap_lock);
	hlist_add_head_rcu(&pid->pid_chain, &pid_hash[pid_hashfn(pid->nr)]);
	spin_unlock_irq(&pidmap_lock);

out:
	return pid;

out_free:
	kmem_cache_free(pid_cachep, pid);
	pid = NULL;
	goto out;
}

struct pid * fastcall find_pid(int nr)
{
	struct hlist_node *elem;
	struct pid *pid;

	hlist_for_each_entry_rcu(pid, elem,
			&pid_hash[pid_hashfn(nr)], pid_chain) {
		if (pid->nr == nr)
			return pid;
	}
	return NULL;
}
EXPORT_SYMBOL_GPL(find_pid);

int fastcall attach_pid(struct task_struct *task, enum pid_type type, int nr)
{
	struct pid_link *link;
	struct pid *pid;

	link = &task->pids[type];
	link->pid = pid = find_pid(nr);
	hlist_add_head_rcu(&link->node, &pid->tasks[type]);

	return 0;
}

void fastcall detach_pid(struct task_struct *task, enum pid_type type)
{
	struct pid_link *link;
	struct pid *pid;
	int tmp;

	link = &task->pids[type];
	pid = link->pid;

	hlist_del_rcu(&link->node);
	link->pid = NULL;

	for (tmp = PIDTYPE_MAX; --tmp >= 0; )
		if (!hlist_empty(&pid->tasks[tmp]))
			return;

	free_pid(pid);
}

/* transfer_pid is an optimization of attach_pid(new), detach_pid(old) */
void fastcall transfer_pid(struct task_struct *old, struct task_struct *new,
			   enum pid_type type)
{
	new->pids[type].pid = old->pids[type].pid;
	hlist_replace_rcu(&old->pids[type].node, &new->pids[type].node);
	old->pids[type].pid = NULL;
}

struct task_struct * fastcall pid_task(struct pid *pid, enum pid_type type)
{
	struct task_struct *result = NULL;
	if (pid) {
		struct hlist_node *first;
		first = rcu_dereference(pid->tasks[type].first);
		if (first)
			result = hlist_entry(first, struct task_struct, pids[(type)].node);
	}
	return result;
}

/*
 * Must be called under rcu_read_lock() or with tasklist_lock read-held.
 */
struct task_struct *find_task_by_pid_type(int type, int nr)
{
	return pid_task(find_pid(nr), type);
}

EXPORT_SYMBOL(find_task_by_pid_type);

struct pid *get_task_pid(struct task_struct *task, enum pid_type type)
{
	struct pid *pid;
	rcu_read_lock();
	pid = get_pid(task->pids[type].pid);
	rcu_read_unlock();
	return pid;
}

struct task_struct *fastcall get_pid_task(struct pid *pid, enum pid_type type)
{
	struct task_struct *result;
	rcu_read_lock();
	result = pid_task(pid, type);
	if (result)
		get_task_struct(result);
	rcu_read_unlock();
	return result;
}

struct pid *find_get_pid(pid_t nr)
{
	struct pid *pid;

	rcu_read_lock();
	pid = get_pid(find_pid(nr));
	rcu_read_unlock();

	return pid;
}

/*
 * Used by proc to find the first pid that is greater then or equal to nr.
 *
 * If there is a pid at nr this function is exactly the same as find_pid.
 */
struct pid *find_ge_pid(int nr)
{
	struct pid *pid;

	do {
		pid = find_pid(nr);
		if (pid)
			break;
		nr = next_pidmap(current->nsproxy->pid_ns, nr);
	} while (nr > 0);

	return pid;
}
EXPORT_SYMBOL_GPL(find_get_pid);

int copy_pid_ns(int flags, struct task_struct *tsk)
{
	struct pid_namespace *old_ns = tsk->nsproxy->pid_ns;
	int err = 0;

	if (!old_ns)
		return 0;

	get_pid_ns(old_ns);
	return err;
}

void free_pid_ns(struct kref *kref)
{
	struct pid_namespace *ns;

	ns = container_of(kref, struct pid_namespace, kref);
	kfree(ns);
}

/*
 * The pid hash table is scaled according to the amount of memory in the
 * machine.  From a minimum of 16 slots up to 4096 slots at one gigabyte or
 * more.
 */
void __init pidhash_init(void)
{
	int i, pidhash_size;
	unsigned long megabytes = nr_kernel_pages >> (20 - PAGE_SHIFT);

	pidhash_shift = max(4, fls(megabytes * 4));
	pidhash_shift = min(12, pidhash_shift);
	pidhash_size = 1 << pidhash_shift;

	printk("PID hash table entries: %d (order: %d, %Zd bytes)\n",
		pidhash_size, pidhash_shift,
		pidhash_size * sizeof(struct hlist_head));

	pid_hash = alloc_bootmem(pidhash_size *	sizeof(*(pid_hash)));
	if (!pid_hash)
		panic("Could not alloc pidhash!\n");
	for (i = 0; i < pidhash_size; i++)
		INIT_HLIST_HEAD(&pid_hash[i]);
}

void __init pidmap_init(void)
{
	init_pid_ns.pidmap[0].page = kzalloc(PAGE_SIZE, GFP_KERNEL);
	/* Reserve PID 0. We never call free_pidmap(0) */
	set_bit(0, init_pid_ns.pidmap[0].page);
	atomic_dec(&init_pid_ns.pidmap[0].nr_free);

	pid_cachep = kmem_cache_create("pid", sizeof(struct pid),
					__alignof__(struct pid),
					SLAB_PANIC, NULL, NULL);
}
Commit	Line	Data
1da177e4 LT	1	/*
	2	* Generic pidhash and scalable, time-bounded PID allocator
	3	*
	4	* (C) 2002-2003 William Irwin, IBM
	5	* (C) 2004 William Irwin, Oracle
	6	* (C) 2002-2004 Ingo Molnar, Red Hat
	7	*
	8	* pid-structures are backing objects for tasks sharing a given ID to chain
	9	* against. There is very little to them aside from hashing them and
	10	* parking tasks using given ID's on a list.
	11	*
	12	* The hash is always changed with the tasklist_lock write-acquired,
	13	* and the hash is only accessed with the tasklist_lock at least
	14	* read-acquired, so there's no additional SMP locking needed here.
	15	*
	16	* We have a list of bitmap pages, which bitmaps represent the PID space.
	17	* Allocating and freeing PIDs is completely lockless. The worst-case
	18	* allocation scenario when all but one out of 1 million PIDs possible are
	19	* allocated already: the scanning of 32 list entries and at most PAGE_SIZE
	20	* bytes. The typical fastpath is a single successful setbit. Freeing is O(1).
	21	*/
	22
	23	#include <linux/mm.h>
	24	#include <linux/module.h>
	25	#include <linux/slab.h>
	26	#include <linux/init.h>
	27	#include <linux/bootmem.h>
	28	#include <linux/hash.h>
61a58c6c	29	#include <linux/pid_namespace.h>
1da177e4 LT	30
1da177e4 LT	31	#define pid_hashfn(nr) hash_long((unsigned long)nr, pidhash_shift)
92476d7f	32	static struct hlist_head *pid_hash;
1da177e4	33	static int pidhash_shift;
e18b890b	34	static struct kmem_cache *pid_cachep;
1da177e4 LT	35
1da177e4 LT	36	int pid_max = PID_MAX_DEFAULT;
1da177e4 LT	37
	38	#define RESERVED_PIDS 300
	39
	40	int pid_max_min = RESERVED_PIDS + 1;
	41	int pid_max_max = PID_MAX_LIMIT;
	42
1da177e4 LT	43	#define BITS_PER_PAGE (PAGE_SIZE*8)
1da177e4 LT	44	#define BITS_PER_PAGE_MASK (BITS_PER_PAGE-1)
3fbc9648	45
61a58c6c SB	46	static inline int mk_pid(struct pid_namespace *pid_ns,
61a58c6c SB	47	struct pidmap *map, int off)
3fbc9648	48	{
61a58c6c	49	return (map - pid_ns->pidmap)*BITS_PER_PAGE + off;
3fbc9648 SB	50	}
3fbc9648 SB	51
1da177e4 LT	52	#define find_next_offset(map, off) \
	53	find_next_zero_bit((map)->page, BITS_PER_PAGE, off)
	54
	55	/*
	56	* PID-map pages start out as NULL, they get allocated upon
	57	* first use and are never deallocated. This way a low pid_max
	58	* value does not cause lots of bitmaps to be allocated, but
	59	* the scheme scales to up to 4 million PIDs, runtime.
	60	*/
61a58c6c	61	struct pid_namespace init_pid_ns = {
9a575a92 CLG	62	.kref = {
	63	.refcount = ATOMIC_INIT(2),
	64	},
3fbc9648 SB	65	.pidmap = {
	66	[ 0 ... PIDMAP_ENTRIES-1] = { ATOMIC_INIT(BITS_PER_PAGE), NULL }
	67	},
	68	.last_pid = 0
	69	};
1da177e4	70
92476d7f EB	71	/*
	72	* Note: disable interrupts while the pidmap_lock is held as an
	73	* interrupt might come in and do read_lock(&tasklist_lock).
	74	*
	75	* If we don't disable interrupts there is a nasty deadlock between
	76	* detach_pid()->free_pid() and another cpu that does
	77	* spin_lock(&pidmap_lock) followed by an interrupt routine that does
	78	* read_lock(&tasklist_lock);
	79	*
	80	* After we clean up the tasklist_lock and know there are no
	81	* irq handlers that take it we can leave the interrupts enabled.
	82	* For now it is easier to be safe than to prove it can't happen.
	83	*/
3fbc9648	84
1da177e4 LT	85	static __cacheline_aligned_in_smp DEFINE_SPINLOCK(pidmap_lock);
1da177e4 LT	86
61a58c6c	87	static fastcall void free_pidmap(struct pid_namespace *pid_ns, int pid)
1da177e4	88	{
61a58c6c	89	struct pidmap *map = pid_ns->pidmap + pid / BITS_PER_PAGE;
1da177e4 LT	90	int offset = pid & BITS_PER_PAGE_MASK;
	91
	92	clear_bit(offset, map->page);
	93	atomic_inc(&map->nr_free);
	94	}
	95
61a58c6c	96	static int alloc_pidmap(struct pid_namespace *pid_ns)
1da177e4	97	{
61a58c6c	98	int i, offset, max_scan, pid, last = pid_ns->last_pid;
6a1f3b84	99	struct pidmap *map;
1da177e4 LT	100
	101	pid = last + 1;
	102	if (pid >= pid_max)
	103	pid = RESERVED_PIDS;
	104	offset = pid & BITS_PER_PAGE_MASK;
61a58c6c	105	map = &pid_ns->pidmap[pid/BITS_PER_PAGE];
1da177e4 LT	106	max_scan = (pid_max + BITS_PER_PAGE - 1)/BITS_PER_PAGE - !offset;
	107	for (i = 0; i <= max_scan; ++i) {
	108	if (unlikely(!map->page)) {
3fbc9648	109	void *page = kzalloc(PAGE_SIZE, GFP_KERNEL);
1da177e4 LT	110	/*
	111	* Free the page if someone raced with us
	112	* installing it:
	113	*/
92476d7f	114	spin_lock_irq(&pidmap_lock);
1da177e4	115	if (map->page)
3fbc9648	116	kfree(page);
1da177e4	117	else
3fbc9648	118	map->page = page;
92476d7f	119	spin_unlock_irq(&pidmap_lock);
1da177e4 LT	120	if (unlikely(!map->page))
	121	break;
	122	}
	123	if (likely(atomic_read(&map->nr_free))) {
	124	do {
	125	if (!test_and_set_bit(offset, map->page)) {
	126	atomic_dec(&map->nr_free);
61a58c6c	127	pid_ns->last_pid = pid;
1da177e4 LT	128	return pid;
	129	}
	130	offset = find_next_offset(map, offset);
61a58c6c	131	pid = mk_pid(pid_ns, map, offset);
1da177e4 LT	132	/*
	133	* find_next_offset() found a bit, the pid from it
	134	* is in-bounds, and if we fell back to the last
	135	* bitmap block and the final block was the same
	136	* as the starting point, pid is before last_pid.
	137	*/
	138	} while (offset < BITS_PER_PAGE && pid < pid_max &&
	139	(i != max_scan \|\| pid < last \|\|
	140	!((last+1) & BITS_PER_PAGE_MASK)));
	141	}
61a58c6c	142	if (map < &pid_ns->pidmap[(pid_max-1)/BITS_PER_PAGE]) {
1da177e4 LT	143	++map;
	144	offset = 0;
	145	} else {
61a58c6c	146	map = &pid_ns->pidmap[0];
1da177e4 LT	147	offset = RESERVED_PIDS;
	148	if (unlikely(last == offset))
	149	break;
	150	}
61a58c6c	151	pid = mk_pid(pid_ns, map, offset);
1da177e4 LT	152	}
	153	return -1;
	154	}
	155
61a58c6c	156	static int next_pidmap(struct pid_namespace *pid_ns, int last)
0804ef4b EB	157	{
0804ef4b EB	158	int offset;
f40f50d3	159	struct pidmap map, end;
0804ef4b EB	160
0804ef4b EB	161	offset = (last + 1) & BITS_PER_PAGE_MASK;
61a58c6c SB	162	map = &pid_ns->pidmap[(last + 1)/BITS_PER_PAGE];
61a58c6c SB	163	end = &pid_ns->pidmap[PIDMAP_ENTRIES];
f40f50d3	164	for (; map < end; map++, offset = 0) {
0804ef4b EB	165	if (unlikely(!map->page))
	166	continue;
	167	offset = find_next_bit((map)->page, BITS_PER_PAGE, offset);
	168	if (offset < BITS_PER_PAGE)
61a58c6c	169	return mk_pid(pid_ns, map, offset);
0804ef4b EB	170	}
	171	return -1;
	172	}
	173
92476d7f EB	174	fastcall void put_pid(struct pid *pid)
	175	{
	176	if (!pid)
	177	return;
	178	if ((atomic_read(&pid->count) == 1) \|\|
	179	atomic_dec_and_test(&pid->count))
	180	kmem_cache_free(pid_cachep, pid);
	181	}
bbf73147	182	EXPORT_SYMBOL_GPL(put_pid);
92476d7f EB	183
	184	static void delayed_put_pid(struct rcu_head *rhp)
	185	{
	186	struct pid *pid = container_of(rhp, struct pid, rcu);
	187	put_pid(pid);
	188	}
	189
	190	fastcall void free_pid(struct pid *pid)
	191	{
	192	/* We can be called with write_lock_irq(&tasklist_lock) held */
	193	unsigned long flags;
	194
	195	spin_lock_irqsave(&pidmap_lock, flags);
	196	hlist_del_rcu(&pid->pid_chain);
	197	spin_unlock_irqrestore(&pidmap_lock, flags);
	198
6cc1b22a	199	free_pidmap(current->nsproxy->pid_ns, pid->nr);
92476d7f EB	200	call_rcu(&pid->rcu, delayed_put_pid);
	201	}
	202
	203	struct pid *alloc_pid(void)
	204	{
	205	struct pid *pid;
	206	enum pid_type type;
	207	int nr = -1;
	208
	209	pid = kmem_cache_alloc(pid_cachep, GFP_KERNEL);
	210	if (!pid)
	211	goto out;
	212
6cc1b22a	213	nr = alloc_pidmap(current->nsproxy->pid_ns);
92476d7f EB	214	if (nr < 0)
	215	goto out_free;
	216
	217	atomic_set(&pid->count, 1);
	218	pid->nr = nr;
	219	for (type = 0; type < PIDTYPE_MAX; ++type)
	220	INIT_HLIST_HEAD(&pid->tasks[type]);
	221
	222	spin_lock_irq(&pidmap_lock);
	223	hlist_add_head_rcu(&pid->pid_chain, &pid_hash[pid_hashfn(pid->nr)]);
	224	spin_unlock_irq(&pidmap_lock);
	225
	226	out:
	227	return pid;
	228
	229	out_free:
	230	kmem_cache_free(pid_cachep, pid);
	231	pid = NULL;
	232	goto out;
	233	}
	234
	235	struct pid * fastcall find_pid(int nr)
1da177e4 LT	236	{
	237	struct hlist_node *elem;
	238	struct pid *pid;
	239
e56d0903	240	hlist_for_each_entry_rcu(pid, elem,
92476d7f	241	&pid_hash[pid_hashfn(nr)], pid_chain) {
1da177e4 LT	242	if (pid->nr == nr)
	243	return pid;
	244	}
	245	return NULL;
	246	}
bbf73147	247	EXPORT_SYMBOL_GPL(find_pid);
1da177e4	248
36c8b586	249	int fastcall attach_pid(struct task_struct *task, enum pid_type type, int nr)
1da177e4	250	{
92476d7f EB	251	struct pid_link *link;
	252	struct pid *pid;
	253
92476d7f EB	254	link = &task->pids[type];
	255	link->pid = pid = find_pid(nr);
	256	hlist_add_head_rcu(&link->node, &pid->tasks[type]);
1da177e4 LT	257
	258	return 0;
	259	}
	260
36c8b586	261	void fastcall detach_pid(struct task_struct *task, enum pid_type type)
1da177e4	262	{
92476d7f EB	263	struct pid_link *link;
	264	struct pid *pid;
	265	int tmp;
1da177e4	266
92476d7f EB	267	link = &task->pids[type];
92476d7f EB	268	pid = link->pid;
1da177e4	269
92476d7f EB	270	hlist_del_rcu(&link->node);
92476d7f EB	271	link->pid = NULL;
1da177e4	272
92476d7f EB	273	for (tmp = PIDTYPE_MAX; --tmp >= 0; )
	274	if (!hlist_empty(&pid->tasks[tmp]))
	275	return;
1da177e4	276
92476d7f	277	free_pid(pid);
1da177e4 LT	278	}
1da177e4 LT	279
c18258c6 EB	280	/* transfer_pid is an optimization of attach_pid(new), detach_pid(old) */
	281	void fastcall transfer_pid(struct task_struct old, struct task_struct new,
	282	enum pid_type type)
	283	{
	284	new->pids[type].pid = old->pids[type].pid;
	285	hlist_replace_rcu(&old->pids[type].node, &new->pids[type].node);
	286	old->pids[type].pid = NULL;
	287	}
	288
92476d7f	289	struct task_struct * fastcall pid_task(struct pid *pid, enum pid_type type)
1da177e4	290	{
92476d7f EB	291	struct task_struct *result = NULL;
	292	if (pid) {
	293	struct hlist_node *first;
	294	first = rcu_dereference(pid->tasks[type].first);
	295	if (first)
	296	result = hlist_entry(first, struct task_struct, pids[(type)].node);
	297	}
	298	return result;
	299	}
1da177e4	300
92476d7f EB	301	/*
	302	* Must be called under rcu_read_lock() or with tasklist_lock read-held.
	303	*/
36c8b586	304	struct task_struct *find_task_by_pid_type(int type, int nr)
92476d7f EB	305	{
	306	return pid_task(find_pid(nr), type);
	307	}
1da177e4	308
92476d7f	309	EXPORT_SYMBOL(find_task_by_pid_type);
1da177e4	310
1a657f78 ON	311	struct pid get_task_pid(struct task_struct task, enum pid_type type)
	312	{
	313	struct pid *pid;
	314	rcu_read_lock();
	315	pid = get_pid(task->pids[type].pid);
	316	rcu_read_unlock();
	317	return pid;
	318	}
	319
92476d7f EB	320	struct task_struct fastcall get_pid_task(struct pid pid, enum pid_type type)
	321	{
	322	struct task_struct *result;
	323	rcu_read_lock();
	324	result = pid_task(pid, type);
	325	if (result)
	326	get_task_struct(result);
	327	rcu_read_unlock();
	328	return result;
1da177e4 LT	329	}
1da177e4 LT	330
92476d7f	331	struct pid *find_get_pid(pid_t nr)
1da177e4 LT	332	{
	333	struct pid *pid;
	334
92476d7f EB	335	rcu_read_lock();
	336	pid = get_pid(find_pid(nr));
	337	rcu_read_unlock();
1da177e4	338
92476d7f	339	return pid;
1da177e4 LT	340	}
1da177e4 LT	341
0804ef4b EB	342	/*
	343	* Used by proc to find the first pid that is greater then or equal to nr.
	344	*
	345	* If there is a pid at nr this function is exactly the same as find_pid.
	346	*/
	347	struct pid *find_ge_pid(int nr)
	348	{
	349	struct pid *pid;
	350
	351	do {
	352	pid = find_pid(nr);
	353	if (pid)
	354	break;
6cc1b22a	355	nr = next_pidmap(current->nsproxy->pid_ns, nr);
0804ef4b EB	356	} while (nr > 0);
	357
	358	return pid;
	359	}
bbf73147	360	EXPORT_SYMBOL_GPL(find_get_pid);
0804ef4b	361
9a575a92 CLG	362	int copy_pid_ns(int flags, struct task_struct *tsk)
	363	{
	364	struct pid_namespace *old_ns = tsk->nsproxy->pid_ns;
	365	int err = 0;
	366
	367	if (!old_ns)
	368	return 0;
	369
	370	get_pid_ns(old_ns);
	371	return err;
	372	}
	373
	374	void free_pid_ns(struct kref *kref)
	375	{
	376	struct pid_namespace *ns;
	377
	378	ns = container_of(kref, struct pid_namespace, kref);
	379	kfree(ns);
	380	}
	381
1da177e4 LT	382	/*
	383	* The pid hash table is scaled according to the amount of memory in the
	384	* machine. From a minimum of 16 slots up to 4096 slots at one gigabyte or
	385	* more.
	386	*/
	387	void __init pidhash_init(void)
	388	{
92476d7f	389	int i, pidhash_size;
1da177e4 LT	390	unsigned long megabytes = nr_kernel_pages >> (20 - PAGE_SHIFT);
	391
	392	pidhash_shift = max(4, fls(megabytes * 4));
	393	pidhash_shift = min(12, pidhash_shift);
	394	pidhash_size = 1 << pidhash_shift;
	395
	396	printk("PID hash table entries: %d (order: %d, %Zd bytes)\n",
	397	pidhash_size, pidhash_shift,
92476d7f EB	398	pidhash_size * sizeof(struct hlist_head));
	399
	400	pid_hash = alloc_bootmem(pidhash_size * sizeof(*(pid_hash)));
	401	if (!pid_hash)
	402	panic("Could not alloc pidhash!\n");
	403	for (i = 0; i < pidhash_size; i++)
	404	INIT_HLIST_HEAD(&pid_hash[i]);
1da177e4 LT	405	}
	406
	407	void __init pidmap_init(void)
	408	{
61a58c6c	409	init_pid_ns.pidmap[0].page = kzalloc(PAGE_SIZE, GFP_KERNEL);
73b9ebfe	410	/* Reserve PID 0. We never call free_pidmap(0) */
61a58c6c SB	411	set_bit(0, init_pid_ns.pidmap[0].page);
61a58c6c SB	412	atomic_dec(&init_pid_ns.pidmap[0].nr_free);
92476d7f EB	413
	414	pid_cachep = kmem_cache_create("pid", sizeof(struct pid),
	415	__alignof__(struct pid),
	416	SLAB_PANIC, NULL, NULL);
1da177e4	417	}