[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,
	.child_reaper = &init_task
};

/*
 * 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(&init_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);

struct pid_namespace *copy_pid_ns(int flags, struct pid_namespace *old_ns)
{
	BUG_ON(!old_ns);
	get_pid_ns(old_ns);
	return old_ns;
}

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(pid, SLAB_PANIC);
}
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	},
84d73786 SB	68	.last_pid = 0,
84d73786 SB	69	.child_reaper = &init_task
3fbc9648	70	};
1da177e4	71
92476d7f EB	72	/*
	73	* Note: disable interrupts while the pidmap_lock is held as an
	74	* interrupt might come in and do read_lock(&tasklist_lock).
	75	*
	76	* If we don't disable interrupts there is a nasty deadlock between
	77	* detach_pid()->free_pid() and another cpu that does
	78	* spin_lock(&pidmap_lock) followed by an interrupt routine that does
	79	* read_lock(&tasklist_lock);
	80	*
	81	* After we clean up the tasklist_lock and know there are no
	82	* irq handlers that take it we can leave the interrupts enabled.
	83	* For now it is easier to be safe than to prove it can't happen.
	84	*/
3fbc9648	85
1da177e4 LT	86	static __cacheline_aligned_in_smp DEFINE_SPINLOCK(pidmap_lock);
1da177e4 LT	87
61a58c6c	88	static fastcall void free_pidmap(struct pid_namespace *pid_ns, int pid)
1da177e4	89	{
61a58c6c	90	struct pidmap *map = pid_ns->pidmap + pid / BITS_PER_PAGE;
1da177e4 LT	91	int offset = pid & BITS_PER_PAGE_MASK;
	92
	93	clear_bit(offset, map->page);
	94	atomic_inc(&map->nr_free);
	95	}
	96
61a58c6c	97	static int alloc_pidmap(struct pid_namespace *pid_ns)
1da177e4	98	{
61a58c6c	99	int i, offset, max_scan, pid, last = pid_ns->last_pid;
6a1f3b84	100	struct pidmap *map;
1da177e4 LT	101
	102	pid = last + 1;
	103	if (pid >= pid_max)
	104	pid = RESERVED_PIDS;
	105	offset = pid & BITS_PER_PAGE_MASK;
61a58c6c	106	map = &pid_ns->pidmap[pid/BITS_PER_PAGE];
1da177e4 LT	107	max_scan = (pid_max + BITS_PER_PAGE - 1)/BITS_PER_PAGE - !offset;
	108	for (i = 0; i <= max_scan; ++i) {
	109	if (unlikely(!map->page)) {
3fbc9648	110	void *page = kzalloc(PAGE_SIZE, GFP_KERNEL);
1da177e4 LT	111	/*
	112	* Free the page if someone raced with us
	113	* installing it:
	114	*/
92476d7f	115	spin_lock_irq(&pidmap_lock);
1da177e4	116	if (map->page)
3fbc9648	117	kfree(page);
1da177e4	118	else
3fbc9648	119	map->page = page;
92476d7f	120	spin_unlock_irq(&pidmap_lock);
1da177e4 LT	121	if (unlikely(!map->page))
	122	break;
	123	}
	124	if (likely(atomic_read(&map->nr_free))) {
	125	do {
	126	if (!test_and_set_bit(offset, map->page)) {
	127	atomic_dec(&map->nr_free);
61a58c6c	128	pid_ns->last_pid = pid;
1da177e4 LT	129	return pid;
	130	}
	131	offset = find_next_offset(map, offset);
61a58c6c	132	pid = mk_pid(pid_ns, map, offset);
1da177e4 LT	133	/*
	134	* find_next_offset() found a bit, the pid from it
	135	* is in-bounds, and if we fell back to the last
	136	* bitmap block and the final block was the same
	137	* as the starting point, pid is before last_pid.
	138	*/
	139	} while (offset < BITS_PER_PAGE && pid < pid_max &&
	140	(i != max_scan \|\| pid < last \|\|
	141	!((last+1) & BITS_PER_PAGE_MASK)));
	142	}
61a58c6c	143	if (map < &pid_ns->pidmap[(pid_max-1)/BITS_PER_PAGE]) {
1da177e4 LT	144	++map;
	145	offset = 0;
	146	} else {
61a58c6c	147	map = &pid_ns->pidmap[0];
1da177e4 LT	148	offset = RESERVED_PIDS;
	149	if (unlikely(last == offset))
	150	break;
	151	}
61a58c6c	152	pid = mk_pid(pid_ns, map, offset);
1da177e4 LT	153	}
	154	return -1;
	155	}
	156
61a58c6c	157	static int next_pidmap(struct pid_namespace *pid_ns, int last)
0804ef4b EB	158	{
0804ef4b EB	159	int offset;
f40f50d3	160	struct pidmap map, end;
0804ef4b EB	161
0804ef4b EB	162	offset = (last + 1) & BITS_PER_PAGE_MASK;
61a58c6c SB	163	map = &pid_ns->pidmap[(last + 1)/BITS_PER_PAGE];
61a58c6c SB	164	end = &pid_ns->pidmap[PIDMAP_ENTRIES];
f40f50d3	165	for (; map < end; map++, offset = 0) {
0804ef4b EB	166	if (unlikely(!map->page))
	167	continue;
	168	offset = find_next_bit((map)->page, BITS_PER_PAGE, offset);
	169	if (offset < BITS_PER_PAGE)
61a58c6c	170	return mk_pid(pid_ns, map, offset);
0804ef4b EB	171	}
	172	return -1;
	173	}
	174
92476d7f EB	175	fastcall void put_pid(struct pid *pid)
	176	{
	177	if (!pid)
	178	return;
	179	if ((atomic_read(&pid->count) == 1) \|\|
	180	atomic_dec_and_test(&pid->count))
	181	kmem_cache_free(pid_cachep, pid);
	182	}
bbf73147	183	EXPORT_SYMBOL_GPL(put_pid);
92476d7f EB	184
	185	static void delayed_put_pid(struct rcu_head *rhp)
	186	{
	187	struct pid *pid = container_of(rhp, struct pid, rcu);
	188	put_pid(pid);
	189	}
	190
	191	fastcall void free_pid(struct pid *pid)
	192	{
	193	/* We can be called with write_lock_irq(&tasklist_lock) held */
	194	unsigned long flags;
	195
	196	spin_lock_irqsave(&pidmap_lock, flags);
	197	hlist_del_rcu(&pid->pid_chain);
	198	spin_unlock_irqrestore(&pidmap_lock, flags);
	199
0f245285	200	free_pidmap(&init_pid_ns, pid->nr);
92476d7f EB	201	call_rcu(&pid->rcu, delayed_put_pid);
	202	}
	203
	204	struct pid *alloc_pid(void)
	205	{
	206	struct pid *pid;
	207	enum pid_type type;
	208	int nr = -1;
	209
	210	pid = kmem_cache_alloc(pid_cachep, GFP_KERNEL);
	211	if (!pid)
	212	goto out;
	213
6cc1b22a	214	nr = alloc_pidmap(current->nsproxy->pid_ns);
92476d7f EB	215	if (nr < 0)
	216	goto out_free;
	217
	218	atomic_set(&pid->count, 1);
	219	pid->nr = nr;
	220	for (type = 0; type < PIDTYPE_MAX; ++type)
	221	INIT_HLIST_HEAD(&pid->tasks[type]);
	222
	223	spin_lock_irq(&pidmap_lock);
	224	hlist_add_head_rcu(&pid->pid_chain, &pid_hash[pid_hashfn(pid->nr)]);
	225	spin_unlock_irq(&pidmap_lock);
	226
	227	out:
	228	return pid;
	229
	230	out_free:
	231	kmem_cache_free(pid_cachep, pid);
	232	pid = NULL;
	233	goto out;
	234	}
	235
	236	struct pid * fastcall find_pid(int nr)
1da177e4 LT	237	{
	238	struct hlist_node *elem;
	239	struct pid *pid;
	240
e56d0903	241	hlist_for_each_entry_rcu(pid, elem,
92476d7f	242	&pid_hash[pid_hashfn(nr)], pid_chain) {
1da177e4 LT	243	if (pid->nr == nr)
	244	return pid;
	245	}
	246	return NULL;
	247	}
bbf73147	248	EXPORT_SYMBOL_GPL(find_pid);
1da177e4	249
36c8b586	250	int fastcall attach_pid(struct task_struct *task, enum pid_type type, int nr)
1da177e4	251	{
92476d7f EB	252	struct pid_link *link;
	253	struct pid *pid;
	254
92476d7f EB	255	link = &task->pids[type];
	256	link->pid = pid = find_pid(nr);
	257	hlist_add_head_rcu(&link->node, &pid->tasks[type]);
1da177e4 LT	258
	259	return 0;
	260	}
	261
36c8b586	262	void fastcall detach_pid(struct task_struct *task, enum pid_type type)
1da177e4	263	{
92476d7f EB	264	struct pid_link *link;
	265	struct pid *pid;
	266	int tmp;
1da177e4	267
92476d7f EB	268	link = &task->pids[type];
92476d7f EB	269	pid = link->pid;
1da177e4	270
92476d7f EB	271	hlist_del_rcu(&link->node);
92476d7f EB	272	link->pid = NULL;
1da177e4	273
92476d7f EB	274	for (tmp = PIDTYPE_MAX; --tmp >= 0; )
	275	if (!hlist_empty(&pid->tasks[tmp]))
	276	return;
1da177e4	277
92476d7f	278	free_pid(pid);
1da177e4 LT	279	}
1da177e4 LT	280
c18258c6 EB	281	/* transfer_pid is an optimization of attach_pid(new), detach_pid(old) */
	282	void fastcall transfer_pid(struct task_struct old, struct task_struct new,
	283	enum pid_type type)
	284	{
	285	new->pids[type].pid = old->pids[type].pid;
	286	hlist_replace_rcu(&old->pids[type].node, &new->pids[type].node);
	287	old->pids[type].pid = NULL;
	288	}
	289
92476d7f	290	struct task_struct * fastcall pid_task(struct pid *pid, enum pid_type type)
1da177e4	291	{
92476d7f EB	292	struct task_struct *result = NULL;
	293	if (pid) {
	294	struct hlist_node *first;
	295	first = rcu_dereference(pid->tasks[type].first);
	296	if (first)
	297	result = hlist_entry(first, struct task_struct, pids[(type)].node);
	298	}
	299	return result;
	300	}
1da177e4	301
92476d7f EB	302	/*
	303	* Must be called under rcu_read_lock() or with tasklist_lock read-held.
	304	*/
36c8b586	305	struct task_struct *find_task_by_pid_type(int type, int nr)
92476d7f EB	306	{
	307	return pid_task(find_pid(nr), type);
	308	}
1da177e4	309
92476d7f	310	EXPORT_SYMBOL(find_task_by_pid_type);
1da177e4	311
1a657f78 ON	312	struct pid get_task_pid(struct task_struct task, enum pid_type type)
	313	{
	314	struct pid *pid;
	315	rcu_read_lock();
	316	pid = get_pid(task->pids[type].pid);
	317	rcu_read_unlock();
	318	return pid;
	319	}
	320
92476d7f EB	321	struct task_struct fastcall get_pid_task(struct pid pid, enum pid_type type)
	322	{
	323	struct task_struct *result;
	324	rcu_read_lock();
	325	result = pid_task(pid, type);
	326	if (result)
	327	get_task_struct(result);
	328	rcu_read_unlock();
	329	return result;
1da177e4 LT	330	}
1da177e4 LT	331
92476d7f	332	struct pid *find_get_pid(pid_t nr)
1da177e4 LT	333	{
	334	struct pid *pid;
	335
92476d7f EB	336	rcu_read_lock();
	337	pid = get_pid(find_pid(nr));
	338	rcu_read_unlock();
1da177e4	339
92476d7f	340	return pid;
1da177e4 LT	341	}
1da177e4 LT	342
0804ef4b EB	343	/*
	344	* Used by proc to find the first pid that is greater then or equal to nr.
	345	*
	346	* If there is a pid at nr this function is exactly the same as find_pid.
	347	*/
	348	struct pid *find_ge_pid(int nr)
	349	{
	350	struct pid *pid;
	351
	352	do {
	353	pid = find_pid(nr);
	354	if (pid)
	355	break;
6cc1b22a	356	nr = next_pidmap(current->nsproxy->pid_ns, nr);
0804ef4b EB	357	} while (nr > 0);
	358
	359	return pid;
	360	}
bbf73147	361	EXPORT_SYMBOL_GPL(find_get_pid);
0804ef4b	362
e3222c4e	363	struct pid_namespace copy_pid_ns(int flags, struct pid_namespace old_ns)
9a575a92	364	{
e3222c4e	365	BUG_ON(!old_ns);
9a575a92	366	get_pid_ns(old_ns);
e3222c4e	367	return old_ns;
9a575a92 CLG	368	}
	369
	370	void free_pid_ns(struct kref *kref)
	371	{
	372	struct pid_namespace *ns;
	373
	374	ns = container_of(kref, struct pid_namespace, kref);
	375	kfree(ns);
	376	}
	377
1da177e4 LT	378	/*
	379	* The pid hash table is scaled according to the amount of memory in the
	380	* machine. From a minimum of 16 slots up to 4096 slots at one gigabyte or
	381	* more.
	382	*/
	383	void __init pidhash_init(void)
	384	{
92476d7f	385	int i, pidhash_size;
1da177e4 LT	386	unsigned long megabytes = nr_kernel_pages >> (20 - PAGE_SHIFT);
	387
	388	pidhash_shift = max(4, fls(megabytes * 4));
	389	pidhash_shift = min(12, pidhash_shift);
	390	pidhash_size = 1 << pidhash_shift;
	391
	392	printk("PID hash table entries: %d (order: %d, %Zd bytes)\n",
	393	pidhash_size, pidhash_shift,
92476d7f EB	394	pidhash_size * sizeof(struct hlist_head));
	395
	396	pid_hash = alloc_bootmem(pidhash_size * sizeof(*(pid_hash)));
	397	if (!pid_hash)
	398	panic("Could not alloc pidhash!\n");
	399	for (i = 0; i < pidhash_size; i++)
	400	INIT_HLIST_HEAD(&pid_hash[i]);
1da177e4 LT	401	}
	402
	403	void __init pidmap_init(void)
	404	{
61a58c6c	405	init_pid_ns.pidmap[0].page = kzalloc(PAGE_SIZE, GFP_KERNEL);
73b9ebfe	406	/* Reserve PID 0. We never call free_pidmap(0) */
61a58c6c SB	407	set_bit(0, init_pid_ns.pidmap[0].page);
61a58c6c SB	408	atomic_dec(&init_pid_ns.pidmap[0].nr_free);
92476d7f	409
0a31bd5f	410	pid_cachep = KMEM_CACHE(pid, SLAB_PANIC);
1da177e4	411	}