swap.c 20.4 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7
/*
 *  linux/mm/swap.c
 *
 *  Copyright (C) 1991, 1992, 1993, 1994  Linus Torvalds
 */

/*
S
Simon Arlott 已提交
8
 * This file contains the default values for the operation of the
L
Linus Torvalds 已提交
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
 * 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>
24
#include <linux/export.h>
L
Linus Torvalds 已提交
25 26 27 28 29
#include <linux/mm_inline.h>
#include <linux/percpu_counter.h>
#include <linux/percpu.h>
#include <linux/cpu.h>
#include <linux/notifier.h>
P
Peter Zijlstra 已提交
30
#include <linux/backing-dev.h>
31
#include <linux/memcontrol.h>
32
#include <linux/gfp.h>
L
Linus Torvalds 已提交
33

34 35
#include "internal.h"

L
Linus Torvalds 已提交
36 37 38
/* How many pages do we try to swap or page in/out together? */
int page_cluster;

39
static DEFINE_PER_CPU(struct pagevec[NR_LRU_LISTS], lru_add_pvecs);
40
static DEFINE_PER_CPU(struct pagevec, lru_rotate_pvecs);
M
Minchan Kim 已提交
41
static DEFINE_PER_CPU(struct pagevec, lru_deactivate_pvecs);
42

A
Adrian Bunk 已提交
43 44 45 46
/*
 * This path almost never happens for VM activity - pages are normally
 * freed via pagevecs.  But it gets used by networking.
 */
H
Harvey Harrison 已提交
47
static void __page_cache_release(struct page *page)
A
Adrian Bunk 已提交
48 49 50 51 52 53 54 55 56 57 58
{
	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);
	}
59 60 61 62 63
}

static void __put_single_page(struct page *page)
{
	__page_cache_release(page);
L
Li Hong 已提交
64
	free_hot_cold_page(page, 0);
A
Adrian Bunk 已提交
65 66
}

67
static void __put_compound_page(struct page *page)
L
Linus Torvalds 已提交
68
{
69
	compound_page_dtor *dtor;
L
Linus Torvalds 已提交
70

71 72 73 74 75 76 77 78 79
	__page_cache_release(page);
	dtor = get_compound_page_dtor(page);
	(*dtor)(page);
}

static void put_compound_page(struct page *page)
{
	if (unlikely(PageTail(page))) {
		/* __split_huge_page_refcount can run under us */
80 81 82 83
		struct page *page_head = compound_trans_head(page);

		if (likely(page != page_head &&
			   get_page_unless_zero(page_head))) {
84 85
			unsigned long flags;
			/*
86 87 88 89
			 * page_head wasn't a dangling pointer but it
			 * may not be a head page anymore by the time
			 * we obtain the lock. That is ok as long as it
			 * can't be freed from under us.
90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105
			 */
			flags = compound_lock_irqsave(page_head);
			if (unlikely(!PageTail(page))) {
				/* __split_huge_page_refcount run before us */
				compound_unlock_irqrestore(page_head, flags);
				VM_BUG_ON(PageHead(page_head));
				if (put_page_testzero(page_head))
					__put_single_page(page_head);
			out_put_single:
				if (put_page_testzero(page))
					__put_single_page(page);
				return;
			}
			VM_BUG_ON(page_head != page->first_page);
			/*
			 * We can release the refcount taken by
106 107 108
			 * get_page_unless_zero() now that
			 * __split_huge_page_refcount() is blocked on
			 * the compound_lock.
109 110 111 112
			 */
			if (put_page_testzero(page_head))
				VM_BUG_ON(1);
			/* __split_huge_page_refcount will wait now */
113 114
			VM_BUG_ON(page_mapcount(page) <= 0);
			atomic_dec(&page->_mapcount);
115
			VM_BUG_ON(atomic_read(&page_head->_count) <= 0);
116
			VM_BUG_ON(atomic_read(&page->_count) != 0);
117
			compound_unlock_irqrestore(page_head, flags);
118 119 120 121 122 123
			if (put_page_testzero(page_head)) {
				if (PageHead(page_head))
					__put_compound_page(page_head);
				else
					__put_single_page(page_head);
			}
124 125 126 127 128 129 130 131 132 133
		} else {
			/* page_head is a dangling pointer */
			VM_BUG_ON(PageTail(page));
			goto out_put_single;
		}
	} else if (put_page_testzero(page)) {
		if (PageHead(page))
			__put_compound_page(page);
		else
			__put_single_page(page);
L
Linus Torvalds 已提交
134
	}
N
Nick Piggin 已提交
135 136 137 138 139 140 141
}

void put_page(struct page *page)
{
	if (unlikely(PageCompound(page)))
		put_compound_page(page);
	else if (put_page_testzero(page))
142
		__put_single_page(page);
L
Linus Torvalds 已提交
143 144 145
}
EXPORT_SYMBOL(put_page);

146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184
/*
 * This function is exported but must not be called by anything other
 * than get_page(). It implements the slow path of get_page().
 */
bool __get_page_tail(struct page *page)
{
	/*
	 * This takes care of get_page() if run on a tail page
	 * returned by one of the get_user_pages/follow_page variants.
	 * get_user_pages/follow_page itself doesn't need the compound
	 * lock because it runs __get_page_tail_foll() under the
	 * proper PT lock that already serializes against
	 * split_huge_page().
	 */
	unsigned long flags;
	bool got = false;
	struct page *page_head = compound_trans_head(page);

	if (likely(page != page_head && get_page_unless_zero(page_head))) {
		/*
		 * page_head wasn't a dangling pointer but it
		 * may not be a head page anymore by the time
		 * we obtain the lock. That is ok as long as it
		 * can't be freed from under us.
		 */
		flags = compound_lock_irqsave(page_head);
		/* here __split_huge_page_refcount won't run anymore */
		if (likely(PageTail(page))) {
			__get_page_tail_foll(page, false);
			got = true;
		}
		compound_unlock_irqrestore(page_head, flags);
		if (unlikely(!got))
			put_page(page_head);
	}
	return got;
}
EXPORT_SYMBOL(__get_page_tail);

185
/**
186 187
 * put_pages_list() - release a list of pages
 * @pages: list of pages threaded on page->lru
188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203
 *
 * 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);

S
Shaohua Li 已提交
204 205 206
static void pagevec_lru_move_fn(struct pagevec *pvec,
				void (*move_fn)(struct page *page, void *arg),
				void *arg)
207 208 209
{
	int i;
	struct zone *zone = NULL;
S
Shaohua Li 已提交
210
	unsigned long flags = 0;
211 212 213 214 215 216 217

	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)
S
Shaohua Li 已提交
218
				spin_unlock_irqrestore(&zone->lru_lock, flags);
219
			zone = pagezone;
S
Shaohua Li 已提交
220
			spin_lock_irqsave(&zone->lru_lock, flags);
221
		}
S
Shaohua Li 已提交
222 223

		(*move_fn)(page, arg);
224 225
	}
	if (zone)
S
Shaohua Li 已提交
226
		spin_unlock_irqrestore(&zone->lru_lock, flags);
227 228
	release_pages(pvec->pages, pvec->nr, pvec->cold);
	pagevec_reinit(pvec);
S
Shaohua Li 已提交
229 230
}

S
Shaohua Li 已提交
231 232 233 234 235 236
static void pagevec_move_tail_fn(struct page *page, void *arg)
{
	int *pgmoved = arg;

	if (PageLRU(page) && !PageActive(page) && !PageUnevictable(page)) {
		enum lru_list lru = page_lru_base_type(page);
237 238 239 240 241
		struct lruvec *lruvec;

		lruvec = mem_cgroup_lru_move_lists(page_zone(page),
						   page, lru, lru);
		list_move_tail(&page->lru, &lruvec->lists[lru]);
S
Shaohua Li 已提交
242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257
		(*pgmoved)++;
	}
}

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

	pagevec_lru_move_fn(pvec, pagevec_move_tail_fn, &pgmoved);
	__count_vm_events(PGROTATED, pgmoved);
}

L
Linus Torvalds 已提交
258 259 260
/*
 * 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
261
 * inactive list.
L
Linus Torvalds 已提交
262
 */
S
Shaohua Li 已提交
263
void rotate_reclaimable_page(struct page *page)
L
Linus Torvalds 已提交
264
{
265
	if (!PageLocked(page) && !PageDirty(page) && !PageActive(page) &&
L
Lee Schermerhorn 已提交
266
	    !PageUnevictable(page) && PageLRU(page)) {
267 268 269 270 271 272 273 274 275 276
		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);
	}
L
Linus Torvalds 已提交
277 278
}

K
KOSAKI Motohiro 已提交
279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298
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]++;
}

299
static void __activate_page(struct page *page, void *arg)
L
Linus Torvalds 已提交
300 301
{
	struct zone *zone = page_zone(page);
302 303

	if (PageLRU(page) && !PageActive(page) && !PageUnevictable(page)) {
304 305 306
		int file = page_is_file_cache(page);
		int lru = page_lru_base_type(page);
		del_page_from_lru_list(zone, page, lru);
307

308 309 310 311
		SetPageActive(page);
		lru += LRU_ACTIVE;
		add_page_to_lru_list(zone, page, lru);
		__count_vm_event(PGACTIVATE);
312

313
		update_page_reclaim_stat(zone, page, file, 1);
L
Linus Torvalds 已提交
314
	}
315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350
}

#ifdef CONFIG_SMP
static DEFINE_PER_CPU(struct pagevec, activate_page_pvecs);

static void activate_page_drain(int cpu)
{
	struct pagevec *pvec = &per_cpu(activate_page_pvecs, cpu);

	if (pagevec_count(pvec))
		pagevec_lru_move_fn(pvec, __activate_page, NULL);
}

void activate_page(struct page *page)
{
	if (PageLRU(page) && !PageActive(page) && !PageUnevictable(page)) {
		struct pagevec *pvec = &get_cpu_var(activate_page_pvecs);

		page_cache_get(page);
		if (!pagevec_add(pvec, page))
			pagevec_lru_move_fn(pvec, __activate_page, NULL);
		put_cpu_var(activate_page_pvecs);
	}
}

#else
static inline void activate_page_drain(int cpu)
{
}

void activate_page(struct page *page)
{
	struct zone *zone = page_zone(page);

	spin_lock_irq(&zone->lru_lock);
	__activate_page(page, NULL);
L
Linus Torvalds 已提交
351 352
	spin_unlock_irq(&zone->lru_lock);
}
353
#endif
L
Linus Torvalds 已提交
354 355 356 357 358 359 360 361

/*
 * Mark a page as having seen activity.
 *
 * inactive,unreferenced	->	inactive,referenced
 * inactive,referenced		->	active,unreferenced
 * active,unreferenced		->	active,referenced
 */
H
Harvey Harrison 已提交
362
void mark_page_accessed(struct page *page)
L
Linus Torvalds 已提交
363
{
L
Lee Schermerhorn 已提交
364 365
	if (!PageActive(page) && !PageUnevictable(page) &&
			PageReferenced(page) && PageLRU(page)) {
L
Linus Torvalds 已提交
366 367 368 369 370 371 372 373 374
		activate_page(page);
		ClearPageReferenced(page);
	} else if (!PageReferenced(page)) {
		SetPageReferenced(page);
	}
}

EXPORT_SYMBOL(mark_page_accessed);

375
void __lru_cache_add(struct page *page, enum lru_list lru)
L
Linus Torvalds 已提交
376
{
377
	struct pagevec *pvec = &get_cpu_var(lru_add_pvecs)[lru];
L
Linus Torvalds 已提交
378 379 380

	page_cache_get(page);
	if (!pagevec_add(pvec, page))
381
		____pagevec_lru_add(pvec, lru);
L
Linus Torvalds 已提交
382 383
	put_cpu_var(lru_add_pvecs);
}
384
EXPORT_SYMBOL(__lru_cache_add);
L
Linus Torvalds 已提交
385

386 387 388 389 390 391
/**
 * 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)
L
Linus Torvalds 已提交
392
{
393
	if (PageActive(page)) {
L
Lee Schermerhorn 已提交
394
		VM_BUG_ON(PageUnevictable(page));
395
		ClearPageActive(page);
L
Lee Schermerhorn 已提交
396 397 398
	} else if (PageUnevictable(page)) {
		VM_BUG_ON(PageActive(page));
		ClearPageUnevictable(page);
399
	}
L
Linus Torvalds 已提交
400

L
Lee Schermerhorn 已提交
401
	VM_BUG_ON(PageLRU(page) || PageActive(page) || PageUnevictable(page));
402
	__lru_cache_add(page, lru);
L
Linus Torvalds 已提交
403 404
}

L
Lee Schermerhorn 已提交
405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425
/**
 * 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);
}

M
Minchan Kim 已提交
426 427 428 429 430 431
/*
 * If the page can not be invalidated, it is moved to the
 * inactive list to speed up its reclaim.  It is moved to the
 * head of the list, rather than the tail, to give the flusher
 * threads some time to write it out, as this is much more
 * effective than the single-page writeout from reclaim.
M
Minchan Kim 已提交
432 433 434 435 436 437 438 439 440 441 442 443 444 445
 *
 * If the page isn't page_mapped and dirty/writeback, the page
 * could reclaim asap using PG_reclaim.
 *
 * 1. active, mapped page -> none
 * 2. active, dirty/writeback page -> inactive, head, PG_reclaim
 * 3. inactive, mapped page -> none
 * 4. inactive, dirty/writeback page -> inactive, head, PG_reclaim
 * 5. inactive, clean -> inactive, tail
 * 6. Others -> none
 *
 * In 4, why it moves inactive's head, the VM expects the page would
 * be write it out by flusher threads as this is much more effective
 * than the single-page writeout from reclaim.
M
Minchan Kim 已提交
446
 */
S
Shaohua Li 已提交
447
static void lru_deactivate_fn(struct page *page, void *arg)
M
Minchan Kim 已提交
448 449
{
	int lru, file;
M
Minchan Kim 已提交
450
	bool active;
S
Shaohua Li 已提交
451
	struct zone *zone = page_zone(page);
M
Minchan Kim 已提交
452

M
Minchan Kim 已提交
453
	if (!PageLRU(page))
M
Minchan Kim 已提交
454 455
		return;

456 457 458
	if (PageUnevictable(page))
		return;

M
Minchan Kim 已提交
459 460 461 462
	/* Some processes are using the page */
	if (page_mapped(page))
		return;

M
Minchan Kim 已提交
463 464
	active = PageActive(page);

M
Minchan Kim 已提交
465 466
	file = page_is_file_cache(page);
	lru = page_lru_base_type(page);
M
Minchan Kim 已提交
467
	del_page_from_lru_list(zone, page, lru + active);
M
Minchan Kim 已提交
468 469 470 471
	ClearPageActive(page);
	ClearPageReferenced(page);
	add_page_to_lru_list(zone, page, lru);

M
Minchan Kim 已提交
472 473 474 475 476 477 478 479
	if (PageWriteback(page) || PageDirty(page)) {
		/*
		 * PG_reclaim could be raced with end_page_writeback
		 * It can make readahead confusing.  But race window
		 * is _really_ small and  it's non-critical problem.
		 */
		SetPageReclaim(page);
	} else {
480
		struct lruvec *lruvec;
M
Minchan Kim 已提交
481 482 483 484
		/*
		 * The page's writeback ends up during pagevec
		 * We moves tha page into tail of inactive.
		 */
485 486
		lruvec = mem_cgroup_lru_move_lists(zone, page, lru, lru);
		list_move_tail(&page->lru, &lruvec->lists[lru]);
M
Minchan Kim 已提交
487 488 489 490 491
		__count_vm_event(PGROTATED);
	}

	if (active)
		__count_vm_event(PGDEACTIVATE);
M
Minchan Kim 已提交
492 493 494
	update_page_reclaim_stat(zone, page, file, 0);
}

495 496 497 498 499 500
/*
 * 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)
L
Linus Torvalds 已提交
501
{
502
	struct pagevec *pvecs = per_cpu(lru_add_pvecs, cpu);
503
	struct pagevec *pvec;
504
	int lru;
L
Linus Torvalds 已提交
505

506 507 508 509 510
	for_each_lru(lru) {
		pvec = &pvecs[lru - LRU_BASE];
		if (pagevec_count(pvec))
			____pagevec_lru_add(pvec, lru);
	}
511 512 513 514 515 516 517 518 519 520

	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);
	}
M
Minchan Kim 已提交
521 522 523

	pvec = &per_cpu(lru_deactivate_pvecs, cpu);
	if (pagevec_count(pvec))
S
Shaohua Li 已提交
524
		pagevec_lru_move_fn(pvec, lru_deactivate_fn, NULL);
525 526

	activate_page_drain(cpu);
M
Minchan Kim 已提交
527 528 529 530 531 532 533 534 535 536 537 538
}

/**
 * deactivate_page - forcefully deactivate a page
 * @page: page to deactivate
 *
 * This function hints the VM that @page is a good reclaim candidate,
 * for example if its invalidation fails due to the page being dirty
 * or under writeback.
 */
void deactivate_page(struct page *page)
{
539 540 541 542 543 544 545
	/*
	 * In a workload with many unevictable page such as mprotect, unevictable
	 * page deactivation for accelerating reclaim is pointless.
	 */
	if (PageUnevictable(page))
		return;

M
Minchan Kim 已提交
546 547 548 549
	if (likely(get_page_unless_zero(page))) {
		struct pagevec *pvec = &get_cpu_var(lru_deactivate_pvecs);

		if (!pagevec_add(pvec, page))
S
Shaohua Li 已提交
550
			pagevec_lru_move_fn(pvec, lru_deactivate_fn, NULL);
M
Minchan Kim 已提交
551 552
		put_cpu_var(lru_deactivate_pvecs);
	}
553 554 555 556
}

void lru_add_drain(void)
{
557
	drain_cpu_pagevecs(get_cpu());
558
	put_cpu();
L
Linus Torvalds 已提交
559 560
}

D
David Howells 已提交
561
static void lru_add_drain_per_cpu(struct work_struct *dummy)
562 563 564 565 566 567 568 569 570
{
	lru_add_drain();
}

/*
 * Returns 0 for success
 */
int lru_add_drain_all(void)
{
D
David Howells 已提交
571
	return schedule_on_each_cpu(lru_add_drain_per_cpu);
572 573
}

L
Linus Torvalds 已提交
574 575 576 577 578 579 580 581
/*
 * 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.
 *
582 583 584 585
 * 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.
L
Linus Torvalds 已提交
586 587 588 589
 */
void release_pages(struct page **pages, int nr, int cold)
{
	int i;
590
	LIST_HEAD(pages_to_free);
L
Linus Torvalds 已提交
591
	struct zone *zone = NULL;
592
	unsigned long uninitialized_var(flags);
L
Linus Torvalds 已提交
593 594 595 596

	for (i = 0; i < nr; i++) {
		struct page *page = pages[i];

N
Nick Piggin 已提交
597 598
		if (unlikely(PageCompound(page))) {
			if (zone) {
599
				spin_unlock_irqrestore(&zone->lru_lock, flags);
N
Nick Piggin 已提交
600 601 602 603 604 605
				zone = NULL;
			}
			put_compound_page(page);
			continue;
		}

N
Nick Piggin 已提交
606
		if (!put_page_testzero(page))
L
Linus Torvalds 已提交
607 608
			continue;

609 610
		if (PageLRU(page)) {
			struct zone *pagezone = page_zone(page);
L
Lee Schermerhorn 已提交
611

612 613
			if (pagezone != zone) {
				if (zone)
614 615
					spin_unlock_irqrestore(&zone->lru_lock,
									flags);
616
				zone = pagezone;
617
				spin_lock_irqsave(&zone->lru_lock, flags);
618
			}
N
Nick Piggin 已提交
619
			VM_BUG_ON(!PageLRU(page));
N
Nick Piggin 已提交
620
			__ClearPageLRU(page);
L
Linus Torvalds 已提交
621
			del_page_from_lru(zone, page);
622 623
		}

624
		list_add(&page->lru, &pages_to_free);
L
Linus Torvalds 已提交
625 626
	}
	if (zone)
627
		spin_unlock_irqrestore(&zone->lru_lock, flags);
L
Linus Torvalds 已提交
628

629
	free_hot_cold_page_list(&pages_to_free, cold);
L
Linus Torvalds 已提交
630
}
M
Miklos Szeredi 已提交
631
EXPORT_SYMBOL(release_pages);
L
Linus Torvalds 已提交
632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649

/*
 * 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);
}

650 651
EXPORT_SYMBOL(__pagevec_release);

652
#ifdef CONFIG_TRANSPARENT_HUGEPAGE
653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679
/* used by __split_huge_page_refcount() */
void lru_add_page_tail(struct zone* zone,
		       struct page *page, struct page *page_tail)
{
	int active;
	enum lru_list lru;
	const int file = 0;

	VM_BUG_ON(!PageHead(page));
	VM_BUG_ON(PageCompound(page_tail));
	VM_BUG_ON(PageLRU(page_tail));
	VM_BUG_ON(!spin_is_locked(&zone->lru_lock));

	SetPageLRU(page_tail);

	if (page_evictable(page_tail, NULL)) {
		if (PageActive(page)) {
			SetPageActive(page_tail);
			active = 1;
			lru = LRU_ACTIVE_ANON;
		} else {
			active = 0;
			lru = LRU_INACTIVE_ANON;
		}
		update_page_reclaim_stat(zone, page_tail, file, active);
	} else {
		SetPageUnevictable(page_tail);
680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696
		lru = LRU_UNEVICTABLE;
	}

	if (likely(PageLRU(page)))
		list_add_tail(&page_tail->lru, &page->lru);
	else {
		struct list_head *list_head;
		/*
		 * Head page has not yet been counted, as an hpage,
		 * so we must account for each subpage individually.
		 *
		 * Use the standard add function to put page_tail on the list,
		 * but then correct its position so they all end up in order.
		 */
		add_page_to_lru_list(zone, page_tail, lru);
		list_head = page_tail->lru.prev;
		list_move_tail(&page_tail->lru, list_head);
697 698
	}
}
699
#endif /* CONFIG_TRANSPARENT_HUGEPAGE */
700

S
Shaohua Li 已提交
701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718
static void ____pagevec_lru_add_fn(struct page *page, void *arg)
{
	enum lru_list lru = (enum lru_list)arg;
	struct zone *zone = page_zone(page);
	int file = is_file_lru(lru);
	int active = is_active_lru(lru);

	VM_BUG_ON(PageActive(page));
	VM_BUG_ON(PageUnevictable(page));
	VM_BUG_ON(PageLRU(page));

	SetPageLRU(page);
	if (active)
		SetPageActive(page);
	update_page_reclaim_stat(zone, page, file, active);
	add_page_to_lru_list(zone, page, lru);
}

L
Linus Torvalds 已提交
719 720 721 722
/*
 * Add the passed pages to the LRU, then drop the caller's refcount
 * on them.  Reinitialises the caller's pagevec.
 */
723
void ____pagevec_lru_add(struct pagevec *pvec, enum lru_list lru)
L
Linus Torvalds 已提交
724
{
L
Lee Schermerhorn 已提交
725
	VM_BUG_ON(is_unevictable_lru(lru));
L
Linus Torvalds 已提交
726

S
Shaohua Li 已提交
727
	pagevec_lru_move_fn(pvec, ____pagevec_lru_add_fn, (void *)lru);
L
Linus Torvalds 已提交
728 729
}

730
EXPORT_SYMBOL(____pagevec_lru_add);
L
Linus Torvalds 已提交
731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754

/**
 * 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);
}

755 756
EXPORT_SYMBOL(pagevec_lookup);

L
Linus Torvalds 已提交
757 758 759 760 761 762 763 764
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);
}

765
EXPORT_SYMBOL(pagevec_lookup_tag);
L
Linus Torvalds 已提交
766 767 768 769 770 771

/*
 * Perform any setup for the swap system
 */
void __init swap_setup(void)
{
772
	unsigned long megs = totalram_pages >> (20 - PAGE_SHIFT);
L
Linus Torvalds 已提交
773

P
Peter Zijlstra 已提交
774 775 776 777
#ifdef CONFIG_SWAP
	bdi_init(swapper_space.backing_dev_info);
#endif

L
Linus Torvalds 已提交
778 779 780 781 782 783 784 785 786 787
	/* 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
	 */
}