[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>

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

int pid_max = PID_MAX_DEFAULT;
int last_pid;

#define RESERVED_PIDS		300

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

#define PIDMAP_ENTRIES		((PID_MAX_LIMIT + 8*PAGE_SIZE - 1)/PAGE_SIZE/8)
#define BITS_PER_PAGE		(PAGE_SIZE*8)
#define BITS_PER_PAGE_MASK	(BITS_PER_PAGE-1)
#define mk_pid(map, off)	(((map) - pidmap_array)*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.
 */
typedef struct pidmap {
	atomic_t nr_free;
	void *page;
} pidmap_t;

static pidmap_t pidmap_array[PIDMAP_ENTRIES] =
	 { [ 0 ... PIDMAP_ENTRIES-1 ] = { ATOMIC_INIT(BITS_PER_PAGE), NULL } };

/*
 * 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(int pid)
{
	pidmap_t *map = pidmap_array + 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(void)
{
	int i, offset, max_scan, pid, last = last_pid;
	pidmap_t *map;

	pid = last + 1;
	if (pid >= pid_max)
		pid = RESERVED_PIDS;
	offset = pid & BITS_PER_PAGE_MASK;
	map = &pidmap_array[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)) {
			unsigned long page = get_zeroed_page(GFP_KERNEL);
			/*
			 * Free the page if someone raced with us
			 * installing it:
			 */
			spin_lock_irq(&pidmap_lock);
			if (map->page)
				free_page(page);
			else
				map->page = (void *)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);
					last_pid = pid;
					return pid;
				}
				offset = find_next_offset(map, offset);
				pid = mk_pid(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 < &pidmap_array[(pid_max-1)/BITS_PER_PAGE]) {
			++map;
			offset = 0;
		} else {
			map = &pidmap_array[0];
			offset = RESERVED_PIDS;
			if (unlikely(last == offset))
				break;
		}
		pid = mk_pid(map, offset);
	}
	return -1;
}

static int next_pidmap(int last)
{
	int offset;
	pidmap_t *map;

	offset = (last + 1) & BITS_PER_PAGE_MASK;
	map = &pidmap_array[(last + 1)/BITS_PER_PAGE];
	for (; map < &pidmap_array[PIDMAP_ENTRIES]; 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(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);
}

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(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();
	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;
}

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 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(nr);
	} while (nr > 0);

	return pid;
}

/*
 * 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)
{
	pidmap_array->page = (void *)get_zeroed_page(GFP_KERNEL);
	/* Reserve PID 0. We never call free_pidmap(0) */
	set_bit(0, pidmap_array->page);
	atomic_dec(&pidmap_array->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>
	29
	30	#define pid_hashfn(nr) hash_long((unsigned long)nr, pidhash_shift)
92476d7f	31	static struct hlist_head *pid_hash;
1da177e4	32	static int pidhash_shift;
92476d7f	33	static kmem_cache_t *pid_cachep;
1da177e4 LT	34
	35	int pid_max = PID_MAX_DEFAULT;
	36	int last_pid;
	37
	38	#define RESERVED_PIDS 300
	39
	40	int pid_max_min = RESERVED_PIDS + 1;
	41	int pid_max_max = PID_MAX_LIMIT;
	42
	43	#define PIDMAP_ENTRIES ((PID_MAX_LIMIT + 8*PAGE_SIZE - 1)/PAGE_SIZE/8)
	44	#define BITS_PER_PAGE (PAGE_SIZE*8)
	45	#define BITS_PER_PAGE_MASK (BITS_PER_PAGE-1)
	46	#define mk_pid(map, off) (((map) - pidmap_array)*BITS_PER_PAGE + (off))
	47	#define find_next_offset(map, off) \
	48	find_next_zero_bit((map)->page, BITS_PER_PAGE, off)
	49
	50	/*
	51	* PID-map pages start out as NULL, they get allocated upon
	52	* first use and are never deallocated. This way a low pid_max
	53	* value does not cause lots of bitmaps to be allocated, but
	54	* the scheme scales to up to 4 million PIDs, runtime.
	55	*/
	56	typedef struct pidmap {
	57	atomic_t nr_free;
	58	void *page;
	59	} pidmap_t;
	60
	61	static pidmap_t pidmap_array[PIDMAP_ENTRIES] =
	62	{ [ 0 ... PIDMAP_ENTRIES-1 ] = { ATOMIC_INIT(BITS_PER_PAGE), NULL } };
	63
92476d7f EB	64	/*
	65	* Note: disable interrupts while the pidmap_lock is held as an
	66	* interrupt might come in and do read_lock(&tasklist_lock).
	67	*
	68	* If we don't disable interrupts there is a nasty deadlock between
	69	* detach_pid()->free_pid() and another cpu that does
	70	* spin_lock(&pidmap_lock) followed by an interrupt routine that does
	71	* read_lock(&tasklist_lock);
	72	*
	73	* After we clean up the tasklist_lock and know there are no
	74	* irq handlers that take it we can leave the interrupts enabled.
	75	* For now it is easier to be safe than to prove it can't happen.
	76	*/
1da177e4 LT	77	static __cacheline_aligned_in_smp DEFINE_SPINLOCK(pidmap_lock);
1da177e4 LT	78
92476d7f	79	static fastcall void free_pidmap(int pid)
1da177e4 LT	80	{
	81	pidmap_t *map = pidmap_array + pid / BITS_PER_PAGE;
	82	int offset = pid & BITS_PER_PAGE_MASK;
	83
	84	clear_bit(offset, map->page);
	85	atomic_inc(&map->nr_free);
	86	}
	87
92476d7f	88	static int alloc_pidmap(void)
1da177e4 LT	89	{
	90	int i, offset, max_scan, pid, last = last_pid;
	91	pidmap_t *map;
	92
	93	pid = last + 1;
	94	if (pid >= pid_max)
	95	pid = RESERVED_PIDS;
	96	offset = pid & BITS_PER_PAGE_MASK;
	97	map = &pidmap_array[pid/BITS_PER_PAGE];
	98	max_scan = (pid_max + BITS_PER_PAGE - 1)/BITS_PER_PAGE - !offset;
	99	for (i = 0; i <= max_scan; ++i) {
	100	if (unlikely(!map->page)) {
	101	unsigned long page = get_zeroed_page(GFP_KERNEL);
	102	/*
	103	* Free the page if someone raced with us
	104	* installing it:
	105	*/
92476d7f	106	spin_lock_irq(&pidmap_lock);
1da177e4 LT	107	if (map->page)
	108	free_page(page);
	109	else
	110	map->page = (void *)page;
92476d7f	111	spin_unlock_irq(&pidmap_lock);
1da177e4 LT	112	if (unlikely(!map->page))
	113	break;
	114	}
	115	if (likely(atomic_read(&map->nr_free))) {
	116	do {
	117	if (!test_and_set_bit(offset, map->page)) {
	118	atomic_dec(&map->nr_free);
	119	last_pid = pid;
	120	return pid;
	121	}
	122	offset = find_next_offset(map, offset);
	123	pid = mk_pid(map, offset);
	124	/*
	125	* find_next_offset() found a bit, the pid from it
	126	* is in-bounds, and if we fell back to the last
	127	* bitmap block and the final block was the same
	128	* as the starting point, pid is before last_pid.
	129	*/
	130	} while (offset < BITS_PER_PAGE && pid < pid_max &&
	131	(i != max_scan \|\| pid < last \|\|
	132	!((last+1) & BITS_PER_PAGE_MASK)));
	133	}
	134	if (map < &pidmap_array[(pid_max-1)/BITS_PER_PAGE]) {
	135	++map;
	136	offset = 0;
	137	} else {
	138	map = &pidmap_array[0];
	139	offset = RESERVED_PIDS;
	140	if (unlikely(last == offset))
	141	break;
	142	}
	143	pid = mk_pid(map, offset);
	144	}
	145	return -1;
	146	}
	147
0804ef4b EB	148	static int next_pidmap(int last)
	149	{
	150	int offset;
	151	pidmap_t *map;
	152
	153	offset = (last + 1) & BITS_PER_PAGE_MASK;
	154	map = &pidmap_array[(last + 1)/BITS_PER_PAGE];
	155	for (; map < &pidmap_array[PIDMAP_ENTRIES]; map++, offset = 0) {
	156	if (unlikely(!map->page))
	157	continue;
	158	offset = find_next_bit((map)->page, BITS_PER_PAGE, offset);
	159	if (offset < BITS_PER_PAGE)
	160	return mk_pid(map, offset);
	161	}
	162	return -1;
	163	}
	164
92476d7f EB	165	fastcall void put_pid(struct pid *pid)
	166	{
	167	if (!pid)
	168	return;
	169	if ((atomic_read(&pid->count) == 1) \|\|
	170	atomic_dec_and_test(&pid->count))
	171	kmem_cache_free(pid_cachep, pid);
	172	}
	173
	174	static void delayed_put_pid(struct rcu_head *rhp)
	175	{
	176	struct pid *pid = container_of(rhp, struct pid, rcu);
	177	put_pid(pid);
	178	}
	179
	180	fastcall void free_pid(struct pid *pid)
	181	{
	182	/* We can be called with write_lock_irq(&tasklist_lock) held */
	183	unsigned long flags;
	184
	185	spin_lock_irqsave(&pidmap_lock, flags);
	186	hlist_del_rcu(&pid->pid_chain);
	187	spin_unlock_irqrestore(&pidmap_lock, flags);
	188
	189	free_pidmap(pid->nr);
	190	call_rcu(&pid->rcu, delayed_put_pid);
	191	}
	192
	193	struct pid *alloc_pid(void)
	194	{
	195	struct pid *pid;
	196	enum pid_type type;
	197	int nr = -1;
	198
	199	pid = kmem_cache_alloc(pid_cachep, GFP_KERNEL);
	200	if (!pid)
	201	goto out;
	202
	203	nr = alloc_pidmap();
	204	if (nr < 0)
	205	goto out_free;
	206
	207	atomic_set(&pid->count, 1);
	208	pid->nr = nr;
	209	for (type = 0; type < PIDTYPE_MAX; ++type)
	210	INIT_HLIST_HEAD(&pid->tasks[type]);
	211
	212	spin_lock_irq(&pidmap_lock);
	213	hlist_add_head_rcu(&pid->pid_chain, &pid_hash[pid_hashfn(pid->nr)]);
	214	spin_unlock_irq(&pidmap_lock);
	215
	216	out:
	217	return pid;
	218
	219	out_free:
	220	kmem_cache_free(pid_cachep, pid);
	221	pid = NULL;
	222	goto out;
	223	}
	224
	225	struct pid * fastcall find_pid(int nr)
1da177e4 LT	226	{
	227	struct hlist_node *elem;
	228	struct pid *pid;
	229
e56d0903	230	hlist_for_each_entry_rcu(pid, elem,
92476d7f	231	&pid_hash[pid_hashfn(nr)], pid_chain) {
1da177e4 LT	232	if (pid->nr == nr)
	233	return pid;
	234	}
	235	return NULL;
	236	}
	237
36c8b586	238	int fastcall attach_pid(struct task_struct *task, enum pid_type type, int nr)
1da177e4	239	{
92476d7f EB	240	struct pid_link *link;
	241	struct pid *pid;
	242
92476d7f EB	243	link = &task->pids[type];
	244	link->pid = pid = find_pid(nr);
	245	hlist_add_head_rcu(&link->node, &pid->tasks[type]);
1da177e4 LT	246
	247	return 0;
	248	}
	249
36c8b586	250	void fastcall detach_pid(struct task_struct *task, enum pid_type type)
1da177e4	251	{
92476d7f EB	252	struct pid_link *link;
	253	struct pid *pid;
	254	int tmp;
1da177e4	255
92476d7f EB	256	link = &task->pids[type];
92476d7f EB	257	pid = link->pid;
1da177e4	258
92476d7f EB	259	hlist_del_rcu(&link->node);
92476d7f EB	260	link->pid = NULL;
1da177e4	261
92476d7f EB	262	for (tmp = PIDTYPE_MAX; --tmp >= 0; )
	263	if (!hlist_empty(&pid->tasks[tmp]))
	264	return;
1da177e4	265
92476d7f	266	free_pid(pid);
1da177e4 LT	267	}
1da177e4 LT	268
c18258c6 EB	269	/* transfer_pid is an optimization of attach_pid(new), detach_pid(old) */
	270	void fastcall transfer_pid(struct task_struct old, struct task_struct new,
	271	enum pid_type type)
	272	{
	273	new->pids[type].pid = old->pids[type].pid;
	274	hlist_replace_rcu(&old->pids[type].node, &new->pids[type].node);
	275	old->pids[type].pid = NULL;
	276	}
	277
92476d7f	278	struct task_struct * fastcall pid_task(struct pid *pid, enum pid_type type)
1da177e4	279	{
92476d7f EB	280	struct task_struct *result = NULL;
	281	if (pid) {
	282	struct hlist_node *first;
	283	first = rcu_dereference(pid->tasks[type].first);
	284	if (first)
	285	result = hlist_entry(first, struct task_struct, pids[(type)].node);
	286	}
	287	return result;
	288	}
1da177e4	289
92476d7f EB	290	/*
	291	* Must be called under rcu_read_lock() or with tasklist_lock read-held.
	292	*/
36c8b586	293	struct task_struct *find_task_by_pid_type(int type, int nr)
92476d7f EB	294	{
	295	return pid_task(find_pid(nr), type);
	296	}
1da177e4	297
92476d7f	298	EXPORT_SYMBOL(find_task_by_pid_type);
1da177e4	299
92476d7f EB	300	struct task_struct fastcall get_pid_task(struct pid pid, enum pid_type type)
	301	{
	302	struct task_struct *result;
	303	rcu_read_lock();
	304	result = pid_task(pid, type);
	305	if (result)
	306	get_task_struct(result);
	307	rcu_read_unlock();
	308	return result;
1da177e4 LT	309	}
1da177e4 LT	310
92476d7f	311	struct pid *find_get_pid(pid_t nr)
1da177e4 LT	312	{
	313	struct pid *pid;
	314
92476d7f EB	315	rcu_read_lock();
	316	pid = get_pid(find_pid(nr));
	317	rcu_read_unlock();
1da177e4	318
92476d7f	319	return pid;
1da177e4 LT	320	}
1da177e4 LT	321
0804ef4b EB	322	/*
	323	* Used by proc to find the first pid that is greater then or equal to nr.
	324	*
	325	* If there is a pid at nr this function is exactly the same as find_pid.
	326	*/
	327	struct pid *find_ge_pid(int nr)
	328	{
	329	struct pid *pid;
	330
	331	do {
	332	pid = find_pid(nr);
	333	if (pid)
	334	break;
	335	nr = next_pidmap(nr);
	336	} while (nr > 0);
	337
	338	return pid;
	339	}
	340
1da177e4 LT	341	/*
	342	* The pid hash table is scaled according to the amount of memory in the
	343	* machine. From a minimum of 16 slots up to 4096 slots at one gigabyte or
	344	* more.
	345	*/
	346	void __init pidhash_init(void)
	347	{
92476d7f	348	int i, pidhash_size;
1da177e4 LT	349	unsigned long megabytes = nr_kernel_pages >> (20 - PAGE_SHIFT);
	350
	351	pidhash_shift = max(4, fls(megabytes * 4));
	352	pidhash_shift = min(12, pidhash_shift);
	353	pidhash_size = 1 << pidhash_shift;
	354
	355	printk("PID hash table entries: %d (order: %d, %Zd bytes)\n",
	356	pidhash_size, pidhash_shift,
92476d7f EB	357	pidhash_size * sizeof(struct hlist_head));
	358
	359	pid_hash = alloc_bootmem(pidhash_size * sizeof(*(pid_hash)));
	360	if (!pid_hash)
	361	panic("Could not alloc pidhash!\n");
	362	for (i = 0; i < pidhash_size; i++)
	363	INIT_HLIST_HEAD(&pid_hash[i]);
1da177e4 LT	364	}
	365
	366	void __init pidmap_init(void)
	367	{
1da177e4	368	pidmap_array->page = (void *)get_zeroed_page(GFP_KERNEL);
73b9ebfe	369	/* Reserve PID 0. We never call free_pidmap(0) */
1da177e4 LT	370	set_bit(0, pidmap_array->page);
1da177e4 LT	371	atomic_dec(&pidmap_array->nr_free);
92476d7f EB	372
	373	pid_cachep = kmem_cache_create("pid", sizeof(struct pid),
	374	__alignof__(struct pid),
	375	SLAB_PANIC, NULL, NULL);
1da177e4	376	}