[linux.git] / mm / swap.c

/*
 *  linux/mm/swap.c
 *
 *  Copyright (C) 1991, 1992, 1993, 1994  Linus Torvalds
 */

/*
 * This file contains the default values for the operation of the
 * Linux VM subsystem. Fine-tuning documentation can be found in
 * Documentation/sysctl/vm.txt.
 * Started 18.12.91
 * Swap aging added 23.2.95, Stephen Tweedie.
 * Buffermem limits added 12.3.98, Rik van Riel.
 */

#include <linux/mm.h>
#include <linux/sched.h>
#include <linux/kernel_stat.h>
#include <linux/swap.h>
#include <linux/mman.h>
#include <linux/pagemap.h>
#include <linux/pagevec.h>
#include <linux/init.h>
#include <linux/module.h>
#include <linux/mm_inline.h>
#include <linux/buffer_head.h>	/* for try_to_release_page() */
#include <linux/percpu_counter.h>
#include <linux/percpu.h>
#include <linux/cpu.h>
#include <linux/notifier.h>
#include <linux/backing-dev.h>
#include <linux/memcontrol.h>
#include <linux/gfp.h>

#include "internal.h"

/* How many pages do we try to swap or page in/out together? */
int page_cluster;

static DEFINE_PER_CPU(struct pagevec[NR_LRU_LISTS], lru_add_pvecs);
static DEFINE_PER_CPU(struct pagevec, lru_rotate_pvecs);

/*
 * This path almost never happens for VM activity - pages are normally
 * freed via pagevecs.  But it gets used by networking.
 */
static void __page_cache_release(struct page *page)
{
	if (PageLRU(page)) {
		unsigned long flags;
		struct zone *zone = page_zone(page);

		spin_lock_irqsave(&zone->lru_lock, flags);
		VM_BUG_ON(!PageLRU(page));
		__ClearPageLRU(page);
		del_page_from_lru(zone, page);
		spin_unlock_irqrestore(&zone->lru_lock, flags);
	}
	free_hot_cold_page(page, 0);
}

static void put_compound_page(struct page *page)
{
	page = compound_head(page);
	if (put_page_testzero(page)) {
		compound_page_dtor *dtor;

		dtor = get_compound_page_dtor(page);
		(*dtor)(page);
	}
}

void put_page(struct page *page)
{
	if (unlikely(PageCompound(page)))
		put_compound_page(page);
	else if (put_page_testzero(page))
		__page_cache_release(page);
}
EXPORT_SYMBOL(put_page);

/**
 * put_pages_list() - release a list of pages
 * @pages: list of pages threaded on page->lru
 *
 * Release a list of pages which are strung together on page.lru.  Currently
 * used by read_cache_pages() and related error recovery code.
 */
void put_pages_list(struct list_head *pages)
{
	while (!list_empty(pages)) {
		struct page *victim;

		victim = list_entry(pages->prev, struct page, lru);
		list_del(&victim->lru);
		page_cache_release(victim);
	}
}
EXPORT_SYMBOL(put_pages_list);

/*
 * pagevec_move_tail() must be called with IRQ disabled.
 * Otherwise this may cause nasty races.
 */
static void pagevec_move_tail(struct pagevec *pvec)
{
	int i;
	int pgmoved = 0;
	struct zone *zone = NULL;

	for (i = 0; i < pagevec_count(pvec); i++) {
		struct page *page = pvec->pages[i];
		struct zone *pagezone = page_zone(page);

		if (pagezone != zone) {
			if (zone)
				spin_unlock(&zone->lru_lock);
			zone = pagezone;
			spin_lock(&zone->lru_lock);
		}
		if (PageLRU(page) && !PageActive(page) && !PageUnevictable(page)) {
			int lru = page_lru_base_type(page);
			list_move_tail(&page->lru, &zone->lru[lru].list);
			pgmoved++;
		}
	}
	if (zone)
		spin_unlock(&zone->lru_lock);
	__count_vm_events(PGROTATED, pgmoved);
	release_pages(pvec->pages, pvec->nr, pvec->cold);
	pagevec_reinit(pvec);
}

/*
 * Writeback is about to end against a page which has been marked for immediate
 * reclaim.  If it still appears to be reclaimable, move it to the tail of the
 * inactive list.
 */
void  rotate_reclaimable_page(struct page *page)
{
	if (!PageLocked(page) && !PageDirty(page) && !PageActive(page) &&
	    !PageUnevictable(page) && PageLRU(page)) {
		struct pagevec *pvec;
		unsigned long flags;

		page_cache_get(page);
		local_irq_save(flags);
		pvec = &__get_cpu_var(lru_rotate_pvecs);
		if (!pagevec_add(pvec, page))
			pagevec_move_tail(pvec);
		local_irq_restore(flags);
	}
}

static void update_page_reclaim_stat(struct zone *zone, struct page *page,
				     int file, int rotated)
{
	struct zone_reclaim_stat *reclaim_stat = &zone->reclaim_stat;
	struct zone_reclaim_stat *memcg_reclaim_stat;

	memcg_reclaim_stat = mem_cgroup_get_reclaim_stat_from_page(page);

	reclaim_stat->recent_scanned[file]++;
	if (rotated)
		reclaim_stat->recent_rotated[file]++;

	if (!memcg_reclaim_stat)
		return;

	memcg_reclaim_stat->recent_scanned[file]++;
	if (rotated)
		memcg_reclaim_stat->recent_rotated[file]++;
}

/*
 * FIXME: speed this up?
 */
void activate_page(struct page *page)
{
	struct zone *zone = page_zone(page);

	spin_lock_irq(&zone->lru_lock);
	if (PageLRU(page) && !PageActive(page) && !PageUnevictable(page)) {
		int file = page_is_file_cache(page);
		int lru = page_lru_base_type(page);
		del_page_from_lru_list(zone, page, lru);

		SetPageActive(page);
		lru += LRU_ACTIVE;
		add_page_to_lru_list(zone, page, lru);
		__count_vm_event(PGACTIVATE);

		update_page_reclaim_stat(zone, page, file, 1);
	}
	spin_unlock_irq(&zone->lru_lock);
}

/*
 * Mark a page as having seen activity.
 *
 * inactive,unreferenced	->	inactive,referenced
 * inactive,referenced		->	active,unreferenced
 * active,unreferenced		->	active,referenced
 */
void mark_page_accessed(struct page *page)
{
	if (!PageActive(page) && !PageUnevictable(page) &&
			PageReferenced(page) && PageLRU(page)) {
		activate_page(page);
		ClearPageReferenced(page);
	} else if (!PageReferenced(page)) {
		SetPageReferenced(page);
	}
}

EXPORT_SYMBOL(mark_page_accessed);

void __lru_cache_add(struct page *page, enum lru_list lru)
{
	struct pagevec *pvec = &get_cpu_var(lru_add_pvecs)[lru];

	page_cache_get(page);
	if (!pagevec_add(pvec, page))
		____pagevec_lru_add(pvec, lru);
	put_cpu_var(lru_add_pvecs);
}
EXPORT_SYMBOL(__lru_cache_add);

/**
 * lru_cache_add_lru - add a page to a page list
 * @page: the page to be added to the LRU.
 * @lru: the LRU list to which the page is added.
 */
void lru_cache_add_lru(struct page *page, enum lru_list lru)
{
	if (PageActive(page)) {
		VM_BUG_ON(PageUnevictable(page));
		ClearPageActive(page);
	} else if (PageUnevictable(page)) {
		VM_BUG_ON(PageActive(page));
		ClearPageUnevictable(page);
	}

	VM_BUG_ON(PageLRU(page) || PageActive(page) || PageUnevictable(page));
	__lru_cache_add(page, lru);
}

/**
 * add_page_to_unevictable_list - add a page to the unevictable list
 * @page:  the page to be added to the unevictable list
 *
 * Add page directly to its zone's unevictable list.  To avoid races with
 * tasks that might be making the page evictable, through eg. munlock,
 * munmap or exit, while it's not on the lru, we want to add the page
 * while it's locked or otherwise "invisible" to other tasks.  This is
 * difficult to do when using the pagevec cache, so bypass that.
 */
void add_page_to_unevictable_list(struct page *page)
{
	struct zone *zone = page_zone(page);

	spin_lock_irq(&zone->lru_lock);
	SetPageUnevictable(page);
	SetPageLRU(page);
	add_page_to_lru_list(zone, page, LRU_UNEVICTABLE);
	spin_unlock_irq(&zone->lru_lock);
}

/*
 * Drain pages out of the cpu's pagevecs.
 * Either "cpu" is the current CPU, and preemption has already been
 * disabled; or "cpu" is being hot-unplugged, and is already dead.
 */
static void drain_cpu_pagevecs(int cpu)
{
	struct pagevec *pvecs = per_cpu(lru_add_pvecs, cpu);
	struct pagevec *pvec;
	int lru;

	for_each_lru(lru) {
		pvec = &pvecs[lru - LRU_BASE];
		if (pagevec_count(pvec))
			____pagevec_lru_add(pvec, lru);
	}

	pvec = &per_cpu(lru_rotate_pvecs, cpu);
	if (pagevec_count(pvec)) {
		unsigned long flags;

		/* No harm done if a racing interrupt already did this */
		local_irq_save(flags);
		pagevec_move_tail(pvec);
		local_irq_restore(flags);
	}
}

void lru_add_drain(void)
{
	drain_cpu_pagevecs(get_cpu());
	put_cpu();
}

static void lru_add_drain_per_cpu(struct work_struct *dummy)
{
	lru_add_drain();
}

/*
 * Returns 0 for success
 */
int lru_add_drain_all(void)
{
	return schedule_on_each_cpu(lru_add_drain_per_cpu);
}

/*
 * Batched page_cache_release().  Decrement the reference count on all the
 * passed pages.  If it fell to zero then remove the page from the LRU and
 * free it.
 *
 * Avoid taking zone->lru_lock if possible, but if it is taken, retain it
 * for the remainder of the operation.
 *
 * The locking in this function is against shrink_inactive_list(): we recheck
 * the page count inside the lock to see whether shrink_inactive_list()
 * grabbed the page via the LRU.  If it did, give up: shrink_inactive_list()
 * will free it.
 */
void release_pages(struct page **pages, int nr, int cold)
{
	int i;
	struct pagevec pages_to_free;
	struct zone *zone = NULL;
	unsigned long uninitialized_var(flags);

	pagevec_init(&pages_to_free, cold);
	for (i = 0; i < nr; i++) {
		struct page *page = pages[i];

		if (unlikely(PageCompound(page))) {
			if (zone) {
				spin_unlock_irqrestore(&zone->lru_lock, flags);
				zone = NULL;
			}
			put_compound_page(page);
			continue;
		}

		if (!put_page_testzero(page))
			continue;

		if (PageLRU(page)) {
			struct zone *pagezone = page_zone(page);

			if (pagezone != zone) {
				if (zone)
					spin_unlock_irqrestore(&zone->lru_lock,
									flags);
				zone = pagezone;
				spin_lock_irqsave(&zone->lru_lock, flags);
			}
			VM_BUG_ON(!PageLRU(page));
			__ClearPageLRU(page);
			del_page_from_lru(zone, page);
		}

		if (!pagevec_add(&pages_to_free, page)) {
			if (zone) {
				spin_unlock_irqrestore(&zone->lru_lock, flags);
				zone = NULL;
			}
			__pagevec_free(&pages_to_free);
			pagevec_reinit(&pages_to_free);
  		}
	}
	if (zone)
		spin_unlock_irqrestore(&zone->lru_lock, flags);

	pagevec_free(&pages_to_free);
}

/*
 * The pages which we're about to release may be in the deferred lru-addition
 * queues.  That would prevent them from really being freed right now.  That's
 * OK from a correctness point of view but is inefficient - those pages may be
 * cache-warm and we want to give them back to the page allocator ASAP.
 *
 * So __pagevec_release() will drain those queues here.  __pagevec_lru_add()
 * and __pagevec_lru_add_active() call release_pages() directly to avoid
 * mutual recursion.
 */
void __pagevec_release(struct pagevec *pvec)
{
	lru_add_drain();
	release_pages(pvec->pages, pagevec_count(pvec), pvec->cold);
	pagevec_reinit(pvec);
}

EXPORT_SYMBOL(__pagevec_release);

/*
 * Add the passed pages to the LRU, then drop the caller's refcount
 * on them.  Reinitialises the caller's pagevec.
 */
void ____pagevec_lru_add(struct pagevec *pvec, enum lru_list lru)
{
	int i;
	struct zone *zone = NULL;

	VM_BUG_ON(is_unevictable_lru(lru));

	for (i = 0; i < pagevec_count(pvec); i++) {
		struct page *page = pvec->pages[i];
		struct zone *pagezone = page_zone(page);
		int file;
		int active;

		if (pagezone != zone) {
			if (zone)
				spin_unlock_irq(&zone->lru_lock);
			zone = pagezone;
			spin_lock_irq(&zone->lru_lock);
		}
		VM_BUG_ON(PageActive(page));
		VM_BUG_ON(PageUnevictable(page));
		VM_BUG_ON(PageLRU(page));
		SetPageLRU(page);
		active = is_active_lru(lru);
		file = is_file_lru(lru);
		if (active)
			SetPageActive(page);
		update_page_reclaim_stat(zone, page, file, active);
		add_page_to_lru_list(zone, page, lru);
	}
	if (zone)
		spin_unlock_irq(&zone->lru_lock);
	release_pages(pvec->pages, pvec->nr, pvec->cold);
	pagevec_reinit(pvec);
}

EXPORT_SYMBOL(____pagevec_lru_add);

/*
 * Try to drop buffers from the pages in a pagevec
 */
void pagevec_strip(struct pagevec *pvec)
{
	int i;

	for (i = 0; i < pagevec_count(pvec); i++) {
		struct page *page = pvec->pages[i];

		if (page_has_private(page) && trylock_page(page)) {
			if (page_has_private(page))
				try_to_release_page(page, 0);
			unlock_page(page);
		}
	}
}

/**
 * pagevec_lookup - gang pagecache lookup
 * @pvec:	Where the resulting pages are placed
 * @mapping:	The address_space to search
 * @start:	The starting page index
 * @nr_pages:	The maximum number of pages
 *
 * pagevec_lookup() will search for and return a group of up to @nr_pages pages
 * in the mapping.  The pages are placed in @pvec.  pagevec_lookup() takes a
 * reference against the pages in @pvec.
 *
 * The search returns a group of mapping-contiguous pages with ascending
 * indexes.  There may be holes in the indices due to not-present pages.
 *
 * pagevec_lookup() returns the number of pages which were found.
 */
unsigned pagevec_lookup(struct pagevec *pvec, struct address_space *mapping,
		pgoff_t start, unsigned nr_pages)
{
	pvec->nr = find_get_pages(mapping, start, nr_pages, pvec->pages);
	return pagevec_count(pvec);
}

EXPORT_SYMBOL(pagevec_lookup);

unsigned pagevec_lookup_tag(struct pagevec *pvec, struct address_space *mapping,
		pgoff_t *index, int tag, unsigned nr_pages)
{
	pvec->nr = find_get_pages_tag(mapping, index, tag,
					nr_pages, pvec->pages);
	return pagevec_count(pvec);
}

EXPORT_SYMBOL(pagevec_lookup_tag);

/*
 * Perform any setup for the swap system
 */
void __init swap_setup(void)
{
	unsigned long megs = totalram_pages >> (20 - PAGE_SHIFT);

#ifdef CONFIG_SWAP
	bdi_init(swapper_space.backing_dev_info);
#endif

	/* Use a smaller cluster for small-memory machines */
	if (megs < 16)
		page_cluster = 2;
	else
		page_cluster = 3;
	/*
	 * Right now other parts of the system means that we
	 * _really_ don't want to cluster much more
	 */
}
Commit	Line	Data
1da177e4 LT	1	/*
	2	* linux/mm/swap.c
	3	*
	4	* Copyright (C) 1991, 1992, 1993, 1994 Linus Torvalds
	5	*/
	6
	7	/*
183ff22b	8	* This file contains the default values for the operation of the
1da177e4 LT	9	* Linux VM subsystem. Fine-tuning documentation can be found in
	10	* Documentation/sysctl/vm.txt.
	11	* Started 18.12.91
	12	* Swap aging added 23.2.95, Stephen Tweedie.
	13	* Buffermem limits added 12.3.98, Rik van Riel.
	14	*/
	15
	16	#include <linux/mm.h>
	17	#include <linux/sched.h>
	18	#include <linux/kernel_stat.h>
	19	#include <linux/swap.h>
	20	#include <linux/mman.h>
	21	#include <linux/pagemap.h>
	22	#include <linux/pagevec.h>
	23	#include <linux/init.h>
	24	#include <linux/module.h>
	25	#include <linux/mm_inline.h>
	26	#include <linux/buffer_head.h> /* for try_to_release_page() */
1da177e4 LT	27	#include <linux/percpu_counter.h>
	28	#include <linux/percpu.h>
	29	#include <linux/cpu.h>
	30	#include <linux/notifier.h>
e0bf68dd	31	#include <linux/backing-dev.h>
66e1707b	32	#include <linux/memcontrol.h>
5a0e3ad6	33	#include <linux/gfp.h>
1da177e4	34
64d6519d LS	35	#include "internal.h"
64d6519d LS	36
1da177e4 LT	37	/* How many pages do we try to swap or page in/out together? */
	38	int page_cluster;
	39
f04e9ebb	40	static DEFINE_PER_CPU(struct pagevec[NR_LRU_LISTS], lru_add_pvecs);
f84f9504	41	static DEFINE_PER_CPU(struct pagevec, lru_rotate_pvecs);
902aaed0	42
b221385b AB	43	/*
	44	* This path almost never happens for VM activity - pages are normally
	45	* freed via pagevecs. But it gets used by networking.
	46	*/
920c7a5d	47	static void __page_cache_release(struct page *page)
b221385b AB	48	{
	49	if (PageLRU(page)) {
	50	unsigned long flags;
	51	struct zone *zone = page_zone(page);
	52
	53	spin_lock_irqsave(&zone->lru_lock, flags);
	54	VM_BUG_ON(!PageLRU(page));
	55	__ClearPageLRU(page);
	56	del_page_from_lru(zone, page);
	57	spin_unlock_irqrestore(&zone->lru_lock, flags);
	58	}
fc91668e	59	free_hot_cold_page(page, 0);
b221385b AB	60	}
b221385b AB	61
8519fb30	62	static void put_compound_page(struct page *page)
1da177e4	63	{
d85f3385	64	page = compound_head(page);
8519fb30	65	if (put_page_testzero(page)) {
33f2ef89	66	compound_page_dtor *dtor;
1da177e4	67
33f2ef89	68	dtor = get_compound_page_dtor(page);
8519fb30	69	(*dtor)(page);
1da177e4	70	}
8519fb30 NP	71	}
	72
	73	void put_page(struct page *page)
	74	{
	75	if (unlikely(PageCompound(page)))
	76	put_compound_page(page);
	77	else if (put_page_testzero(page))
1da177e4 LT	78	__page_cache_release(page);
	79	}
	80	EXPORT_SYMBOL(put_page);
1da177e4	81
1d7ea732	82	/**
7682486b RD	83	* put_pages_list() - release a list of pages
7682486b RD	84	* @pages: list of pages threaded on page->lru
1d7ea732 AZ	85	*
	86	* Release a list of pages which are strung together on page.lru. Currently
	87	* used by read_cache_pages() and related error recovery code.
1d7ea732 AZ	88	*/
	89	void put_pages_list(struct list_head *pages)
	90	{
	91	while (!list_empty(pages)) {
	92	struct page *victim;
	93
	94	victim = list_entry(pages->prev, struct page, lru);
	95	list_del(&victim->lru);
	96	page_cache_release(victim);
	97	}
	98	}
	99	EXPORT_SYMBOL(put_pages_list);
	100
902aaed0 HH	101	/*
	102	* pagevec_move_tail() must be called with IRQ disabled.
	103	* Otherwise this may cause nasty races.
	104	*/
	105	static void pagevec_move_tail(struct pagevec *pvec)
	106	{
	107	int i;
	108	int pgmoved = 0;
	109	struct zone *zone = NULL;
	110
	111	for (i = 0; i < pagevec_count(pvec); i++) {
	112	struct page *page = pvec->pages[i];
	113	struct zone *pagezone = page_zone(page);
	114
	115	if (pagezone != zone) {
	116	if (zone)
	117	spin_unlock(&zone->lru_lock);
	118	zone = pagezone;
	119	spin_lock(&zone->lru_lock);
	120	}
894bc310	121	if (PageLRU(page) && !PageActive(page) && !PageUnevictable(page)) {
401a8e1c	122	int lru = page_lru_base_type(page);
4f98a2fe	123	list_move_tail(&page->lru, &zone->lru[lru].list);
902aaed0 HH	124	pgmoved++;
	125	}
	126	}
	127	if (zone)
	128	spin_unlock(&zone->lru_lock);
	129	__count_vm_events(PGROTATED, pgmoved);
	130	release_pages(pvec->pages, pvec->nr, pvec->cold);
	131	pagevec_reinit(pvec);
	132	}
	133
1da177e4 LT	134	/*
	135	* Writeback is about to end against a page which has been marked for immediate
	136	* reclaim. If it still appears to be reclaimable, move it to the tail of the
902aaed0	137	* inactive list.
1da177e4	138	*/
ac6aadb2	139	void rotate_reclaimable_page(struct page *page)
1da177e4	140	{
ac6aadb2	141	if (!PageLocked(page) && !PageDirty(page) && !PageActive(page) &&
894bc310	142	!PageUnevictable(page) && PageLRU(page)) {
ac6aadb2 MS	143	struct pagevec *pvec;
	144	unsigned long flags;
	145
	146	page_cache_get(page);
	147	local_irq_save(flags);
	148	pvec = &__get_cpu_var(lru_rotate_pvecs);
	149	if (!pagevec_add(pvec, page))
	150	pagevec_move_tail(pvec);
	151	local_irq_restore(flags);
	152	}
1da177e4 LT	153	}
1da177e4 LT	154
3e2f41f1 KM	155	static void update_page_reclaim_stat(struct zone zone, struct page page,
	156	int file, int rotated)
	157	{
	158	struct zone_reclaim_stat *reclaim_stat = &zone->reclaim_stat;
	159	struct zone_reclaim_stat *memcg_reclaim_stat;
	160
	161	memcg_reclaim_stat = mem_cgroup_get_reclaim_stat_from_page(page);
	162
	163	reclaim_stat->recent_scanned[file]++;
	164	if (rotated)
	165	reclaim_stat->recent_rotated[file]++;
	166
	167	if (!memcg_reclaim_stat)
	168	return;
	169
	170	memcg_reclaim_stat->recent_scanned[file]++;
	171	if (rotated)
	172	memcg_reclaim_stat->recent_rotated[file]++;
	173	}
	174
1da177e4 LT	175	/*
	176	* FIXME: speed this up?
	177	*/
920c7a5d	178	void activate_page(struct page *page)
1da177e4 LT	179	{
	180	struct zone *zone = page_zone(page);
	181
	182	spin_lock_irq(&zone->lru_lock);
894bc310	183	if (PageLRU(page) && !PageActive(page) && !PageUnevictable(page)) {
4f98a2fe	184	int file = page_is_file_cache(page);
401a8e1c	185	int lru = page_lru_base_type(page);
4f98a2fe RR	186	del_page_from_lru_list(zone, page, lru);
4f98a2fe RR	187
1da177e4	188	SetPageActive(page);
4f98a2fe RR	189	lru += LRU_ACTIVE;
4f98a2fe RR	190	add_page_to_lru_list(zone, page, lru);
f8891e5e	191	__count_vm_event(PGACTIVATE);
4f98a2fe	192
6c0b1351	193	update_page_reclaim_stat(zone, page, file, 1);
1da177e4 LT	194	}
	195	spin_unlock_irq(&zone->lru_lock);
	196	}
	197
	198	/*
	199	* Mark a page as having seen activity.
	200	*
	201	* inactive,unreferenced -> inactive,referenced
	202	* inactive,referenced -> active,unreferenced
	203	* active,unreferenced -> active,referenced
	204	*/
920c7a5d	205	void mark_page_accessed(struct page *page)
1da177e4	206	{
894bc310 LS	207	if (!PageActive(page) && !PageUnevictable(page) &&
894bc310 LS	208	PageReferenced(page) && PageLRU(page)) {
1da177e4 LT	209	activate_page(page);
	210	ClearPageReferenced(page);
	211	} else if (!PageReferenced(page)) {
	212	SetPageReferenced(page);
	213	}
	214	}
	215
	216	EXPORT_SYMBOL(mark_page_accessed);
	217
f04e9ebb	218	void __lru_cache_add(struct page *page, enum lru_list lru)
1da177e4	219	{
f04e9ebb	220	struct pagevec *pvec = &get_cpu_var(lru_add_pvecs)[lru];
1da177e4 LT	221
	222	page_cache_get(page);
	223	if (!pagevec_add(pvec, page))
f04e9ebb	224	____pagevec_lru_add(pvec, lru);
1da177e4 LT	225	put_cpu_var(lru_add_pvecs);
1da177e4 LT	226	}
47846b06	227	EXPORT_SYMBOL(__lru_cache_add);
1da177e4	228
f04e9ebb KM	229	/**
	230	* lru_cache_add_lru - add a page to a page list
	231	* @page: the page to be added to the LRU.
	232	* @lru: the LRU list to which the page is added.
	233	*/
	234	void lru_cache_add_lru(struct page *page, enum lru_list lru)
1da177e4	235	{
f04e9ebb	236	if (PageActive(page)) {
894bc310	237	VM_BUG_ON(PageUnevictable(page));
f04e9ebb	238	ClearPageActive(page);
894bc310 LS	239	} else if (PageUnevictable(page)) {
	240	VM_BUG_ON(PageActive(page));
	241	ClearPageUnevictable(page);
f04e9ebb	242	}
1da177e4	243
894bc310	244	VM_BUG_ON(PageLRU(page) \|\| PageActive(page) \|\| PageUnevictable(page));
f04e9ebb	245	__lru_cache_add(page, lru);
1da177e4 LT	246	}
1da177e4 LT	247
894bc310 LS	248	/**
	249	* add_page_to_unevictable_list - add a page to the unevictable list
	250	* @page: the page to be added to the unevictable list
	251	*
	252	* Add page directly to its zone's unevictable list. To avoid races with
	253	* tasks that might be making the page evictable, through eg. munlock,
	254	* munmap or exit, while it's not on the lru, we want to add the page
	255	* while it's locked or otherwise "invisible" to other tasks. This is
	256	* difficult to do when using the pagevec cache, so bypass that.
	257	*/
	258	void add_page_to_unevictable_list(struct page *page)
	259	{
	260	struct zone *zone = page_zone(page);
	261
	262	spin_lock_irq(&zone->lru_lock);
	263	SetPageUnevictable(page);
	264	SetPageLRU(page);
	265	add_page_to_lru_list(zone, page, LRU_UNEVICTABLE);
	266	spin_unlock_irq(&zone->lru_lock);
	267	}
	268
902aaed0 HH	269	/*
	270	* Drain pages out of the cpu's pagevecs.
	271	* Either "cpu" is the current CPU, and preemption has already been
	272	* disabled; or "cpu" is being hot-unplugged, and is already dead.
	273	*/
	274	static void drain_cpu_pagevecs(int cpu)
1da177e4	275	{
f04e9ebb	276	struct pagevec *pvecs = per_cpu(lru_add_pvecs, cpu);
902aaed0	277	struct pagevec *pvec;
f04e9ebb	278	int lru;
1da177e4	279
f04e9ebb KM	280	for_each_lru(lru) {
	281	pvec = &pvecs[lru - LRU_BASE];
	282	if (pagevec_count(pvec))
	283	____pagevec_lru_add(pvec, lru);
	284	}
902aaed0 HH	285
	286	pvec = &per_cpu(lru_rotate_pvecs, cpu);
	287	if (pagevec_count(pvec)) {
	288	unsigned long flags;
	289
	290	/* No harm done if a racing interrupt already did this */
	291	local_irq_save(flags);
	292	pagevec_move_tail(pvec);
	293	local_irq_restore(flags);
	294	}
80bfed90 AM	295	}
	296
	297	void lru_add_drain(void)
	298	{
902aaed0	299	drain_cpu_pagevecs(get_cpu());
80bfed90	300	put_cpu();
1da177e4 LT	301	}
1da177e4 LT	302
c4028958	303	static void lru_add_drain_per_cpu(struct work_struct *dummy)
053837fc NP	304	{
	305	lru_add_drain();
	306	}
	307
	308	/*
	309	* Returns 0 for success
	310	*/
	311	int lru_add_drain_all(void)
	312	{
c4028958	313	return schedule_on_each_cpu(lru_add_drain_per_cpu);
053837fc NP	314	}
053837fc NP	315
1da177e4 LT	316	/*
	317	* Batched page_cache_release(). Decrement the reference count on all the
	318	* passed pages. If it fell to zero then remove the page from the LRU and
	319	* free it.
	320	*
	321	* Avoid taking zone->lru_lock if possible, but if it is taken, retain it
	322	* for the remainder of the operation.
	323	*
ab33dc09 FLVC	324	* The locking in this function is against shrink_inactive_list(): we recheck
	325	* the page count inside the lock to see whether shrink_inactive_list()
	326	* grabbed the page via the LRU. If it did, give up: shrink_inactive_list()
	327	* will free it.
1da177e4 LT	328	*/
	329	void release_pages(struct page **pages, int nr, int cold)
	330	{
	331	int i;
	332	struct pagevec pages_to_free;
	333	struct zone *zone = NULL;
902aaed0	334	unsigned long uninitialized_var(flags);
1da177e4 LT	335
	336	pagevec_init(&pages_to_free, cold);
	337	for (i = 0; i < nr; i++) {
	338	struct page *page = pages[i];
1da177e4	339
8519fb30 NP	340	if (unlikely(PageCompound(page))) {
8519fb30 NP	341	if (zone) {
902aaed0	342	spin_unlock_irqrestore(&zone->lru_lock, flags);
8519fb30 NP	343	zone = NULL;
	344	}
	345	put_compound_page(page);
	346	continue;
	347	}
	348
b5810039	349	if (!put_page_testzero(page))
1da177e4 LT	350	continue;
1da177e4 LT	351
46453a6e NP	352	if (PageLRU(page)) {
46453a6e NP	353	struct zone *pagezone = page_zone(page);
894bc310	354
46453a6e NP	355	if (pagezone != zone) {
46453a6e NP	356	if (zone)
902aaed0 HH	357	spin_unlock_irqrestore(&zone->lru_lock,
902aaed0 HH	358	flags);
46453a6e	359	zone = pagezone;
902aaed0	360	spin_lock_irqsave(&zone->lru_lock, flags);
46453a6e	361	}
725d704e	362	VM_BUG_ON(!PageLRU(page));
67453911	363	__ClearPageLRU(page);
1da177e4	364	del_page_from_lru(zone, page);
46453a6e NP	365	}
	366
	367	if (!pagevec_add(&pages_to_free, page)) {
	368	if (zone) {
902aaed0	369	spin_unlock_irqrestore(&zone->lru_lock, flags);
46453a6e	370	zone = NULL;
1da177e4	371	}
46453a6e NP	372	__pagevec_free(&pages_to_free);
	373	pagevec_reinit(&pages_to_free);
	374	}
1da177e4 LT	375	}
1da177e4 LT	376	if (zone)
902aaed0	377	spin_unlock_irqrestore(&zone->lru_lock, flags);
1da177e4 LT	378
	379	pagevec_free(&pages_to_free);
	380	}
	381
	382	/*
	383	* The pages which we're about to release may be in the deferred lru-addition
	384	* queues. That would prevent them from really being freed right now. That's
	385	* OK from a correctness point of view but is inefficient - those pages may be
	386	* cache-warm and we want to give them back to the page allocator ASAP.
	387	*
	388	* So __pagevec_release() will drain those queues here. __pagevec_lru_add()
	389	* and __pagevec_lru_add_active() call release_pages() directly to avoid
	390	* mutual recursion.
	391	*/
	392	void __pagevec_release(struct pagevec *pvec)
	393	{
	394	lru_add_drain();
	395	release_pages(pvec->pages, pagevec_count(pvec), pvec->cold);
	396	pagevec_reinit(pvec);
	397	}
	398
7f285701 SF	399	EXPORT_SYMBOL(__pagevec_release);
7f285701 SF	400
1da177e4 LT	401	/*
	402	* Add the passed pages to the LRU, then drop the caller's refcount
	403	* on them. Reinitialises the caller's pagevec.
	404	*/
f04e9ebb	405	void ____pagevec_lru_add(struct pagevec *pvec, enum lru_list lru)
1da177e4 LT	406	{
	407	int i;
	408	struct zone *zone = NULL;
6e901571	409
894bc310	410	VM_BUG_ON(is_unevictable_lru(lru));
1da177e4 LT	411
	412	for (i = 0; i < pagevec_count(pvec); i++) {
	413	struct page *page = pvec->pages[i];
	414	struct zone *pagezone = page_zone(page);
9ff473b9	415	int file;
3e2f41f1	416	int active;
1da177e4 LT	417
	418	if (pagezone != zone) {
	419	if (zone)
	420	spin_unlock_irq(&zone->lru_lock);
	421	zone = pagezone;
	422	spin_lock_irq(&zone->lru_lock);
	423	}
894bc310 LS	424	VM_BUG_ON(PageActive(page));
894bc310 LS	425	VM_BUG_ON(PageUnevictable(page));
725d704e	426	VM_BUG_ON(PageLRU(page));
8d438f96	427	SetPageLRU(page);
3e2f41f1	428	active = is_active_lru(lru);
9ff473b9	429	file = is_file_lru(lru);
3e2f41f1	430	if (active)
f04e9ebb	431	SetPageActive(page);
3e2f41f1	432	update_page_reclaim_stat(zone, page, file, active);
f04e9ebb	433	add_page_to_lru_list(zone, page, lru);
1da177e4 LT	434	}
	435	if (zone)
	436	spin_unlock_irq(&zone->lru_lock);
	437	release_pages(pvec->pages, pvec->nr, pvec->cold);
	438	pagevec_reinit(pvec);
	439	}
	440
f04e9ebb	441	EXPORT_SYMBOL(____pagevec_lru_add);
1da177e4 LT	442
	443	/*
	444	* Try to drop buffers from the pages in a pagevec
	445	*/
	446	void pagevec_strip(struct pagevec *pvec)
	447	{
	448	int i;
	449
	450	for (i = 0; i < pagevec_count(pvec); i++) {
	451	struct page *page = pvec->pages[i];
	452
266cf658 DH	453	if (page_has_private(page) && trylock_page(page)) {
266cf658 DH	454	if (page_has_private(page))
5b40dc78	455	try_to_release_page(page, 0);
1da177e4 LT	456	unlock_page(page);
	457	}
	458	}
	459	}
	460
	461	/**
	462	* pagevec_lookup - gang pagecache lookup
	463	* @pvec: Where the resulting pages are placed
	464	* @mapping: The address_space to search
	465	* @start: The starting page index
	466	* @nr_pages: The maximum number of pages
	467	*
	468	* pagevec_lookup() will search for and return a group of up to @nr_pages pages
	469	* in the mapping. The pages are placed in @pvec. pagevec_lookup() takes a
	470	* reference against the pages in @pvec.
	471	*
	472	* The search returns a group of mapping-contiguous pages with ascending
	473	* indexes. There may be holes in the indices due to not-present pages.
	474	*
	475	* pagevec_lookup() returns the number of pages which were found.
	476	*/
	477	unsigned pagevec_lookup(struct pagevec pvec, struct address_space mapping,
	478	pgoff_t start, unsigned nr_pages)
	479	{
	480	pvec->nr = find_get_pages(mapping, start, nr_pages, pvec->pages);
	481	return pagevec_count(pvec);
	482	}
	483
78539fdf CH	484	EXPORT_SYMBOL(pagevec_lookup);
78539fdf CH	485
1da177e4 LT	486	unsigned pagevec_lookup_tag(struct pagevec pvec, struct address_space mapping,
	487	pgoff_t *index, int tag, unsigned nr_pages)
	488	{
	489	pvec->nr = find_get_pages_tag(mapping, index, tag,
	490	nr_pages, pvec->pages);
	491	return pagevec_count(pvec);
	492	}
	493
7f285701	494	EXPORT_SYMBOL(pagevec_lookup_tag);
1da177e4	495
1da177e4 LT	496	/*
	497	* Perform any setup for the swap system
	498	*/
	499	void __init swap_setup(void)
	500	{
4481374c	501	unsigned long megs = totalram_pages >> (20 - PAGE_SHIFT);
1da177e4	502
e0bf68dd PZ	503	#ifdef CONFIG_SWAP
	504	bdi_init(swapper_space.backing_dev_info);
	505	#endif
	506
1da177e4 LT	507	/* Use a smaller cluster for small-memory machines */
	508	if (megs < 16)
	509	page_cluster = 2;
	510	else
	511	page_cluster = 3;
	512	/*
	513	* Right now other parts of the system means that we
	514	* _really_ don't want to cluster much more
	515	*/
1da177e4	516	}