[net-next-2.6.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>
#include <linux/init_task.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;
struct pid init_struct_pid = INIT_STRUCT_PID;

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);

/*
 * attach_pid() must be called with the tasklist_lock write-held.
 */
int fastcall attach_pid(struct task_struct *task, enum pid_type type,
		struct pid *pid)
{
	struct pid_link *link;

	link = &task->pids[type];
	link->pid = pid;
	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(unsigned long 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>
820e45db	30	#include <linux/init_task.h>
1da177e4 LT	31
1da177e4 LT	32	#define pid_hashfn(nr) hash_long((unsigned long)nr, pidhash_shift)
92476d7f	33	static struct hlist_head *pid_hash;
1da177e4	34	static int pidhash_shift;
e18b890b	35	static struct kmem_cache *pid_cachep;
820e45db	36	struct pid init_struct_pid = INIT_STRUCT_PID;
1da177e4 LT	37
1da177e4 LT	38	int pid_max = PID_MAX_DEFAULT;
1da177e4 LT	39
	40	#define RESERVED_PIDS 300
	41
	42	int pid_max_min = RESERVED_PIDS + 1;
	43	int pid_max_max = PID_MAX_LIMIT;
	44
1da177e4 LT	45	#define BITS_PER_PAGE (PAGE_SIZE*8)
1da177e4 LT	46	#define BITS_PER_PAGE_MASK (BITS_PER_PAGE-1)
3fbc9648	47
61a58c6c SB	48	static inline int mk_pid(struct pid_namespace *pid_ns,
61a58c6c SB	49	struct pidmap *map, int off)
3fbc9648	50	{
61a58c6c	51	return (map - pid_ns->pidmap)*BITS_PER_PAGE + off;
3fbc9648 SB	52	}
3fbc9648 SB	53
1da177e4 LT	54	#define find_next_offset(map, off) \
	55	find_next_zero_bit((map)->page, BITS_PER_PAGE, off)
	56
	57	/*
	58	* PID-map pages start out as NULL, they get allocated upon
	59	* first use and are never deallocated. This way a low pid_max
	60	* value does not cause lots of bitmaps to be allocated, but
	61	* the scheme scales to up to 4 million PIDs, runtime.
	62	*/
61a58c6c	63	struct pid_namespace init_pid_ns = {
9a575a92 CLG	64	.kref = {
	65	.refcount = ATOMIC_INIT(2),
	66	},
3fbc9648 SB	67	.pidmap = {
	68	[ 0 ... PIDMAP_ENTRIES-1] = { ATOMIC_INIT(BITS_PER_PAGE), NULL }
	69	},
84d73786 SB	70	.last_pid = 0,
84d73786 SB	71	.child_reaper = &init_task
3fbc9648	72	};
1da177e4	73
92476d7f EB	74	/*
	75	* Note: disable interrupts while the pidmap_lock is held as an
	76	* interrupt might come in and do read_lock(&tasklist_lock).
	77	*
	78	* If we don't disable interrupts there is a nasty deadlock between
	79	* detach_pid()->free_pid() and another cpu that does
	80	* spin_lock(&pidmap_lock) followed by an interrupt routine that does
	81	* read_lock(&tasklist_lock);
	82	*
	83	* After we clean up the tasklist_lock and know there are no
	84	* irq handlers that take it we can leave the interrupts enabled.
	85	* For now it is easier to be safe than to prove it can't happen.
	86	*/
3fbc9648	87
1da177e4 LT	88	static __cacheline_aligned_in_smp DEFINE_SPINLOCK(pidmap_lock);
1da177e4 LT	89
61a58c6c	90	static fastcall void free_pidmap(struct pid_namespace *pid_ns, int pid)
1da177e4	91	{
61a58c6c	92	struct pidmap *map = pid_ns->pidmap + pid / BITS_PER_PAGE;
1da177e4 LT	93	int offset = pid & BITS_PER_PAGE_MASK;
	94
	95	clear_bit(offset, map->page);
	96	atomic_inc(&map->nr_free);
	97	}
	98
61a58c6c	99	static int alloc_pidmap(struct pid_namespace *pid_ns)
1da177e4	100	{
61a58c6c	101	int i, offset, max_scan, pid, last = pid_ns->last_pid;
6a1f3b84	102	struct pidmap *map;
1da177e4 LT	103
	104	pid = last + 1;
	105	if (pid >= pid_max)
	106	pid = RESERVED_PIDS;
	107	offset = pid & BITS_PER_PAGE_MASK;
61a58c6c	108	map = &pid_ns->pidmap[pid/BITS_PER_PAGE];
1da177e4 LT	109	max_scan = (pid_max + BITS_PER_PAGE - 1)/BITS_PER_PAGE - !offset;
	110	for (i = 0; i <= max_scan; ++i) {
	111	if (unlikely(!map->page)) {
3fbc9648	112	void *page = kzalloc(PAGE_SIZE, GFP_KERNEL);
1da177e4 LT	113	/*
	114	* Free the page if someone raced with us
	115	* installing it:
	116	*/
92476d7f	117	spin_lock_irq(&pidmap_lock);
1da177e4	118	if (map->page)
3fbc9648	119	kfree(page);
1da177e4	120	else
3fbc9648	121	map->page = page;
92476d7f	122	spin_unlock_irq(&pidmap_lock);
1da177e4 LT	123	if (unlikely(!map->page))
	124	break;
	125	}
	126	if (likely(atomic_read(&map->nr_free))) {
	127	do {
	128	if (!test_and_set_bit(offset, map->page)) {
	129	atomic_dec(&map->nr_free);
61a58c6c	130	pid_ns->last_pid = pid;
1da177e4 LT	131	return pid;
	132	}
	133	offset = find_next_offset(map, offset);
61a58c6c	134	pid = mk_pid(pid_ns, map, offset);
1da177e4 LT	135	/*
	136	* find_next_offset() found a bit, the pid from it
	137	* is in-bounds, and if we fell back to the last
	138	* bitmap block and the final block was the same
	139	* as the starting point, pid is before last_pid.
	140	*/
	141	} while (offset < BITS_PER_PAGE && pid < pid_max &&
	142	(i != max_scan \|\| pid < last \|\|
	143	!((last+1) & BITS_PER_PAGE_MASK)));
	144	}
61a58c6c	145	if (map < &pid_ns->pidmap[(pid_max-1)/BITS_PER_PAGE]) {
1da177e4 LT	146	++map;
	147	offset = 0;
	148	} else {
61a58c6c	149	map = &pid_ns->pidmap[0];
1da177e4 LT	150	offset = RESERVED_PIDS;
	151	if (unlikely(last == offset))
	152	break;
	153	}
61a58c6c	154	pid = mk_pid(pid_ns, map, offset);
1da177e4 LT	155	}
	156	return -1;
	157	}
	158
61a58c6c	159	static int next_pidmap(struct pid_namespace *pid_ns, int last)
0804ef4b EB	160	{
0804ef4b EB	161	int offset;
f40f50d3	162	struct pidmap map, end;
0804ef4b EB	163
0804ef4b EB	164	offset = (last + 1) & BITS_PER_PAGE_MASK;
61a58c6c SB	165	map = &pid_ns->pidmap[(last + 1)/BITS_PER_PAGE];
61a58c6c SB	166	end = &pid_ns->pidmap[PIDMAP_ENTRIES];
f40f50d3	167	for (; map < end; map++, offset = 0) {
0804ef4b EB	168	if (unlikely(!map->page))
	169	continue;
	170	offset = find_next_bit((map)->page, BITS_PER_PAGE, offset);
	171	if (offset < BITS_PER_PAGE)
61a58c6c	172	return mk_pid(pid_ns, map, offset);
0804ef4b EB	173	}
	174	return -1;
	175	}
	176
92476d7f EB	177	fastcall void put_pid(struct pid *pid)
	178	{
	179	if (!pid)
	180	return;
	181	if ((atomic_read(&pid->count) == 1) \|\|
	182	atomic_dec_and_test(&pid->count))
	183	kmem_cache_free(pid_cachep, pid);
	184	}
bbf73147	185	EXPORT_SYMBOL_GPL(put_pid);
92476d7f EB	186
	187	static void delayed_put_pid(struct rcu_head *rhp)
	188	{
	189	struct pid *pid = container_of(rhp, struct pid, rcu);
	190	put_pid(pid);
	191	}
	192
	193	fastcall void free_pid(struct pid *pid)
	194	{
	195	/* We can be called with write_lock_irq(&tasklist_lock) held */
	196	unsigned long flags;
	197
	198	spin_lock_irqsave(&pidmap_lock, flags);
	199	hlist_del_rcu(&pid->pid_chain);
	200	spin_unlock_irqrestore(&pidmap_lock, flags);
	201
0f245285	202	free_pidmap(&init_pid_ns, pid->nr);
92476d7f EB	203	call_rcu(&pid->rcu, delayed_put_pid);
	204	}
	205
	206	struct pid *alloc_pid(void)
	207	{
	208	struct pid *pid;
	209	enum pid_type type;
	210	int nr = -1;
	211
	212	pid = kmem_cache_alloc(pid_cachep, GFP_KERNEL);
	213	if (!pid)
	214	goto out;
	215
6cc1b22a	216	nr = alloc_pidmap(current->nsproxy->pid_ns);
92476d7f EB	217	if (nr < 0)
	218	goto out_free;
	219
	220	atomic_set(&pid->count, 1);
	221	pid->nr = nr;
	222	for (type = 0; type < PIDTYPE_MAX; ++type)
	223	INIT_HLIST_HEAD(&pid->tasks[type]);
	224
	225	spin_lock_irq(&pidmap_lock);
	226	hlist_add_head_rcu(&pid->pid_chain, &pid_hash[pid_hashfn(pid->nr)]);
	227	spin_unlock_irq(&pidmap_lock);
	228
	229	out:
	230	return pid;
	231
	232	out_free:
	233	kmem_cache_free(pid_cachep, pid);
	234	pid = NULL;
	235	goto out;
	236	}
	237
	238	struct pid * fastcall find_pid(int nr)
1da177e4 LT	239	{
	240	struct hlist_node *elem;
	241	struct pid *pid;
	242
e56d0903	243	hlist_for_each_entry_rcu(pid, elem,
92476d7f	244	&pid_hash[pid_hashfn(nr)], pid_chain) {
1da177e4 LT	245	if (pid->nr == nr)
	246	return pid;
	247	}
	248	return NULL;
	249	}
bbf73147	250	EXPORT_SYMBOL_GPL(find_pid);
1da177e4	251
e713d0da SB	252	/*
	253	* attach_pid() must be called with the tasklist_lock write-held.
	254	*/
	255	int fastcall attach_pid(struct task_struct *task, enum pid_type type,
	256	struct pid *pid)
1da177e4	257	{
92476d7f	258	struct pid_link *link;
92476d7f	259
92476d7f	260	link = &task->pids[type];
e713d0da	261	link->pid = pid;
92476d7f	262	hlist_add_head_rcu(&link->node, &pid->tasks[type]);
1da177e4 LT	263
	264	return 0;
	265	}
	266
36c8b586	267	void fastcall detach_pid(struct task_struct *task, enum pid_type type)
1da177e4	268	{
92476d7f EB	269	struct pid_link *link;
	270	struct pid *pid;
	271	int tmp;
1da177e4	272
92476d7f EB	273	link = &task->pids[type];
92476d7f EB	274	pid = link->pid;
1da177e4	275
92476d7f EB	276	hlist_del_rcu(&link->node);
92476d7f EB	277	link->pid = NULL;
1da177e4	278
92476d7f EB	279	for (tmp = PIDTYPE_MAX; --tmp >= 0; )
	280	if (!hlist_empty(&pid->tasks[tmp]))
	281	return;
1da177e4	282
92476d7f	283	free_pid(pid);
1da177e4 LT	284	}
1da177e4 LT	285
c18258c6 EB	286	/* transfer_pid is an optimization of attach_pid(new), detach_pid(old) */
	287	void fastcall transfer_pid(struct task_struct old, struct task_struct new,
	288	enum pid_type type)
	289	{
	290	new->pids[type].pid = old->pids[type].pid;
	291	hlist_replace_rcu(&old->pids[type].node, &new->pids[type].node);
	292	old->pids[type].pid = NULL;
	293	}
	294
92476d7f	295	struct task_struct * fastcall pid_task(struct pid *pid, enum pid_type type)
1da177e4	296	{
92476d7f EB	297	struct task_struct *result = NULL;
	298	if (pid) {
	299	struct hlist_node *first;
	300	first = rcu_dereference(pid->tasks[type].first);
	301	if (first)
	302	result = hlist_entry(first, struct task_struct, pids[(type)].node);
	303	}
	304	return result;
	305	}
1da177e4	306
92476d7f EB	307	/*
	308	* Must be called under rcu_read_lock() or with tasklist_lock read-held.
	309	*/
36c8b586	310	struct task_struct *find_task_by_pid_type(int type, int nr)
92476d7f EB	311	{
	312	return pid_task(find_pid(nr), type);
	313	}
1da177e4	314
92476d7f	315	EXPORT_SYMBOL(find_task_by_pid_type);
1da177e4	316
1a657f78 ON	317	struct pid get_task_pid(struct task_struct task, enum pid_type type)
	318	{
	319	struct pid *pid;
	320	rcu_read_lock();
	321	pid = get_pid(task->pids[type].pid);
	322	rcu_read_unlock();
	323	return pid;
	324	}
	325
92476d7f EB	326	struct task_struct fastcall get_pid_task(struct pid pid, enum pid_type type)
	327	{
	328	struct task_struct *result;
	329	rcu_read_lock();
	330	result = pid_task(pid, type);
	331	if (result)
	332	get_task_struct(result);
	333	rcu_read_unlock();
	334	return result;
1da177e4 LT	335	}
1da177e4 LT	336
92476d7f	337	struct pid *find_get_pid(pid_t nr)
1da177e4 LT	338	{
	339	struct pid *pid;
	340
92476d7f EB	341	rcu_read_lock();
	342	pid = get_pid(find_pid(nr));
	343	rcu_read_unlock();
1da177e4	344
92476d7f	345	return pid;
1da177e4 LT	346	}
1da177e4 LT	347
0804ef4b EB	348	/*
	349	* Used by proc to find the first pid that is greater then or equal to nr.
	350	*
	351	* If there is a pid at nr this function is exactly the same as find_pid.
	352	*/
	353	struct pid *find_ge_pid(int nr)
	354	{
	355	struct pid *pid;
	356
	357	do {
	358	pid = find_pid(nr);
	359	if (pid)
	360	break;
6cc1b22a	361	nr = next_pidmap(current->nsproxy->pid_ns, nr);
0804ef4b EB	362	} while (nr > 0);
	363
	364	return pid;
	365	}
bbf73147	366	EXPORT_SYMBOL_GPL(find_get_pid);
0804ef4b	367
213dd266	368	struct pid_namespace copy_pid_ns(unsigned long flags, struct pid_namespace old_ns)
9a575a92	369	{
e3222c4e	370	BUG_ON(!old_ns);
9a575a92	371	get_pid_ns(old_ns);
e3222c4e	372	return old_ns;
9a575a92 CLG	373	}
	374
	375	void free_pid_ns(struct kref *kref)
	376	{
	377	struct pid_namespace *ns;
	378
	379	ns = container_of(kref, struct pid_namespace, kref);
	380	kfree(ns);
	381	}
	382
1da177e4 LT	383	/*
	384	* The pid hash table is scaled according to the amount of memory in the
	385	* machine. From a minimum of 16 slots up to 4096 slots at one gigabyte or
	386	* more.
	387	*/
	388	void __init pidhash_init(void)
	389	{
92476d7f	390	int i, pidhash_size;
1da177e4 LT	391	unsigned long megabytes = nr_kernel_pages >> (20 - PAGE_SHIFT);
	392
	393	pidhash_shift = max(4, fls(megabytes * 4));
	394	pidhash_shift = min(12, pidhash_shift);
	395	pidhash_size = 1 << pidhash_shift;
	396
	397	printk("PID hash table entries: %d (order: %d, %Zd bytes)\n",
	398	pidhash_size, pidhash_shift,
92476d7f EB	399	pidhash_size * sizeof(struct hlist_head));
	400
	401	pid_hash = alloc_bootmem(pidhash_size * sizeof(*(pid_hash)));
	402	if (!pid_hash)
	403	panic("Could not alloc pidhash!\n");
	404	for (i = 0; i < pidhash_size; i++)
	405	INIT_HLIST_HEAD(&pid_hash[i]);
1da177e4 LT	406	}
	407
	408	void __init pidmap_init(void)
	409	{
61a58c6c	410	init_pid_ns.pidmap[0].page = kzalloc(PAGE_SIZE, GFP_KERNEL);
73b9ebfe	411	/* Reserve PID 0. We never call free_pidmap(0) */
61a58c6c SB	412	set_bit(0, init_pid_ns.pidmap[0].page);
61a58c6c SB	413	atomic_dec(&init_pid_ns.pidmap[0].nr_free);
92476d7f	414
0a31bd5f	415	pid_cachep = KMEM_CACHE(pid, SLAB_PANIC);
1da177e4	416	}