swap.c 13.2 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 24 25 26 27 28 29 30
 * 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>
P
Peter Zijlstra 已提交
31
#include <linux/backing-dev.h>
32
#include <linux/memcontrol.h>
L
Linus Torvalds 已提交
33 34 35 36

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

37 38 39
static DEFINE_PER_CPU(struct pagevec, lru_add_pvecs);
static DEFINE_PER_CPU(struct pagevec, lru_add_active_pvecs);
static DEFINE_PER_CPU(struct pagevec, lru_rotate_pvecs);
40

A
Adrian Bunk 已提交
41 42 43 44
/*
 * 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
45
static void __page_cache_release(struct page *page)
A
Adrian Bunk 已提交
46 47 48 49 50 51 52 53 54 55 56 57 58 59
{
	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_page(page);
}

N
Nick Piggin 已提交
60
static void put_compound_page(struct page *page)
L
Linus Torvalds 已提交
61
{
62
	page = compound_head(page);
N
Nick Piggin 已提交
63
	if (put_page_testzero(page)) {
64
		compound_page_dtor *dtor;
L
Linus Torvalds 已提交
65

66
		dtor = get_compound_page_dtor(page);
N
Nick Piggin 已提交
67
		(*dtor)(page);
L
Linus Torvalds 已提交
68
	}
N
Nick Piggin 已提交
69 70 71 72 73 74 75
}

void put_page(struct page *page)
{
	if (unlikely(PageCompound(page)))
		put_compound_page(page);
	else if (put_page_testzero(page))
L
Linus Torvalds 已提交
76 77 78 79
		__page_cache_release(page);
}
EXPORT_SYMBOL(put_page);

80
/**
81 82
 * put_pages_list() - release a list of pages
 * @pages: list of pages threaded on page->lru
83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98
 *
 * 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);

99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119
/*
 * 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)) {
120
			list_move_tail(&page->lru, &zone->lru[LRU_INACTIVE].list);
121 122 123 124 125 126 127 128 129 130
			pgmoved++;
		}
	}
	if (zone)
		spin_unlock(&zone->lru_lock);
	__count_vm_events(PGROTATED, pgmoved);
	release_pages(pvec->pages, pvec->nr, pvec->cold);
	pagevec_reinit(pvec);
}

L
Linus Torvalds 已提交
131 132 133
/*
 * 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
134
 * inactive list.
L
Linus Torvalds 已提交
135
 */
136
void  rotate_reclaimable_page(struct page *page)
L
Linus Torvalds 已提交
137
{
138 139 140 141 142 143 144 145 146 147 148 149
	if (!PageLocked(page) && !PageDirty(page) && !PageActive(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);
	}
L
Linus Torvalds 已提交
150 151 152 153 154
}

/*
 * FIXME: speed this up?
 */
H
Harvey Harrison 已提交
155
void activate_page(struct page *page)
L
Linus Torvalds 已提交
156 157 158 159 160 161 162 163
{
	struct zone *zone = page_zone(page);

	spin_lock_irq(&zone->lru_lock);
	if (PageLRU(page) && !PageActive(page)) {
		del_page_from_inactive_list(zone, page);
		SetPageActive(page);
		add_page_to_active_list(zone, page);
164
		__count_vm_event(PGACTIVATE);
165
		mem_cgroup_move_lists(page, true);
L
Linus Torvalds 已提交
166 167 168 169 170 171 172 173 174 175 176
	}
	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
 */
H
Harvey Harrison 已提交
177
void mark_page_accessed(struct page *page)
L
Linus Torvalds 已提交
178 179 180 181 182 183 184 185 186 187 188 189 190 191 192
{
	if (!PageActive(page) && PageReferenced(page) && PageLRU(page)) {
		activate_page(page);
		ClearPageReferenced(page);
	} else if (!PageReferenced(page)) {
		SetPageReferenced(page);
	}
}

EXPORT_SYMBOL(mark_page_accessed);

/**
 * lru_cache_add: add a page to the page lists
 * @page: the page to add
 */
H
Harvey Harrison 已提交
193
void lru_cache_add(struct page *page)
L
Linus Torvalds 已提交
194 195 196 197 198 199 200 201 202
{
	struct pagevec *pvec = &get_cpu_var(lru_add_pvecs);

	page_cache_get(page);
	if (!pagevec_add(pvec, page))
		__pagevec_lru_add(pvec);
	put_cpu_var(lru_add_pvecs);
}

H
Harvey Harrison 已提交
203
void lru_cache_add_active(struct page *page)
L
Linus Torvalds 已提交
204 205 206 207 208 209 210 211 212
{
	struct pagevec *pvec = &get_cpu_var(lru_add_active_pvecs);

	page_cache_get(page);
	if (!pagevec_add(pvec, page))
		__pagevec_lru_add_active(pvec);
	put_cpu_var(lru_add_active_pvecs);
}

213 214 215 216 217 218
/*
 * 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
219
{
220
	struct pagevec *pvec;
L
Linus Torvalds 已提交
221

222
	pvec = &per_cpu(lru_add_pvecs, cpu);
L
Linus Torvalds 已提交
223 224
	if (pagevec_count(pvec))
		__pagevec_lru_add(pvec);
225

226
	pvec = &per_cpu(lru_add_active_pvecs, cpu);
L
Linus Torvalds 已提交
227 228
	if (pagevec_count(pvec))
		__pagevec_lru_add_active(pvec);
229 230 231 232 233 234 235 236 237 238

	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);
	}
239 240 241 242
}

void lru_add_drain(void)
{
243
	drain_cpu_pagevecs(get_cpu());
244
	put_cpu();
L
Linus Torvalds 已提交
245 246
}

247
#ifdef CONFIG_NUMA
D
David Howells 已提交
248
static void lru_add_drain_per_cpu(struct work_struct *dummy)
249 250 251 252 253 254 255 256 257
{
	lru_add_drain();
}

/*
 * Returns 0 for success
 */
int lru_add_drain_all(void)
{
D
David Howells 已提交
258
	return schedule_on_each_cpu(lru_add_drain_per_cpu);
259 260 261 262 263 264 265 266 267 268 269 270 271 272
}

#else

/*
 * Returns 0 for success
 */
int lru_add_drain_all(void)
{
	lru_add_drain();
	return 0;
}
#endif

L
Linus Torvalds 已提交
273 274 275 276 277 278 279 280
/*
 * 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.
 *
281 282 283 284
 * 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
285 286 287 288 289 290
 */
void release_pages(struct page **pages, int nr, int cold)
{
	int i;
	struct pagevec pages_to_free;
	struct zone *zone = NULL;
291
	unsigned long uninitialized_var(flags);
L
Linus Torvalds 已提交
292 293 294 295 296

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

N
Nick Piggin 已提交
297 298
		if (unlikely(PageCompound(page))) {
			if (zone) {
299
				spin_unlock_irqrestore(&zone->lru_lock, flags);
N
Nick Piggin 已提交
300 301 302 303 304 305
				zone = NULL;
			}
			put_compound_page(page);
			continue;
		}

N
Nick Piggin 已提交
306
		if (!put_page_testzero(page))
L
Linus Torvalds 已提交
307 308
			continue;

309 310 311 312
		if (PageLRU(page)) {
			struct zone *pagezone = page_zone(page);
			if (pagezone != zone) {
				if (zone)
313 314
					spin_unlock_irqrestore(&zone->lru_lock,
									flags);
315
				zone = pagezone;
316
				spin_lock_irqsave(&zone->lru_lock, flags);
317
			}
N
Nick Piggin 已提交
318
			VM_BUG_ON(!PageLRU(page));
N
Nick Piggin 已提交
319
			__ClearPageLRU(page);
L
Linus Torvalds 已提交
320
			del_page_from_lru(zone, page);
321 322 323 324
		}

		if (!pagevec_add(&pages_to_free, page)) {
			if (zone) {
325
				spin_unlock_irqrestore(&zone->lru_lock, flags);
326
				zone = NULL;
L
Linus Torvalds 已提交
327
			}
328 329 330
			__pagevec_free(&pages_to_free);
			pagevec_reinit(&pages_to_free);
  		}
L
Linus Torvalds 已提交
331 332
	}
	if (zone)
333
		spin_unlock_irqrestore(&zone->lru_lock, flags);
L
Linus Torvalds 已提交
334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354

	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);
}

355 356
EXPORT_SYMBOL(__pagevec_release);

L
Linus Torvalds 已提交
357 358 359 360 361 362 363 364 365 366 367 368 369 370
/*
 * pagevec_release() for pages which are known to not be on the LRU
 *
 * This function reinitialises the caller's pagevec.
 */
void __pagevec_release_nonlru(struct pagevec *pvec)
{
	int i;
	struct pagevec pages_to_free;

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

N
Nick Piggin 已提交
371
		VM_BUG_ON(PageLRU(page));
L
Linus Torvalds 已提交
372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397
		if (put_page_testzero(page))
			pagevec_add(&pages_to_free, page);
	}
	pagevec_free(&pages_to_free);
	pagevec_reinit(pvec);
}

/*
 * 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)
{
	int i;
	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_irq(&zone->lru_lock);
			zone = pagezone;
			spin_lock_irq(&zone->lru_lock);
		}
N
Nick Piggin 已提交
398
		VM_BUG_ON(PageLRU(page));
N
Nick Piggin 已提交
399
		SetPageLRU(page);
L
Linus Torvalds 已提交
400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424
		add_page_to_inactive_list(zone, page);
	}
	if (zone)
		spin_unlock_irq(&zone->lru_lock);
	release_pages(pvec->pages, pvec->nr, pvec->cold);
	pagevec_reinit(pvec);
}

EXPORT_SYMBOL(__pagevec_lru_add);

void __pagevec_lru_add_active(struct pagevec *pvec)
{
	int i;
	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_irq(&zone->lru_lock);
			zone = pagezone;
			spin_lock_irq(&zone->lru_lock);
		}
N
Nick Piggin 已提交
425
		VM_BUG_ON(PageLRU(page));
N
Nick Piggin 已提交
426
		SetPageLRU(page);
N
Nick Piggin 已提交
427
		VM_BUG_ON(PageActive(page));
N
Nick Piggin 已提交
428
		SetPageActive(page);
L
Linus Torvalds 已提交
429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446
		add_page_to_active_list(zone, page);
	}
	if (zone)
		spin_unlock_irq(&zone->lru_lock);
	release_pages(pvec->pages, pvec->nr, pvec->cold);
	pagevec_reinit(pvec);
}

/*
 * 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];

N
Nick Piggin 已提交
447
		if (PagePrivate(page) && trylock_page(page)) {
448 449
			if (PagePrivate(page))
				try_to_release_page(page, 0);
L
Linus Torvalds 已提交
450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477
			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);
}

478 479
EXPORT_SYMBOL(pagevec_lookup);

L
Linus Torvalds 已提交
480 481 482 483 484 485 486 487
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);
}

488
EXPORT_SYMBOL(pagevec_lookup_tag);
L
Linus Torvalds 已提交
489 490 491 492 493 494 495 496

#ifdef CONFIG_SMP
/*
 * We tolerate a little inaccuracy to avoid ping-ponging the counter between
 * CPUs
 */
#define ACCT_THRESHOLD	max(16, NR_CPUS * 2)

497
static DEFINE_PER_CPU(long, committed_space);
L
Linus Torvalds 已提交
498 499 500 501 502 503 504 505 506

void vm_acct_memory(long pages)
{
	long *local;

	preempt_disable();
	local = &__get_cpu_var(committed_space);
	*local += pages;
	if (*local > ACCT_THRESHOLD || *local < -ACCT_THRESHOLD) {
A
Alan Cox 已提交
507
		atomic_long_add(*local, &vm_committed_space);
L
Linus Torvalds 已提交
508 509 510 511 512 513 514 515 516 517 518 519 520 521 522
		*local = 0;
	}
	preempt_enable();
}

#ifdef CONFIG_HOTPLUG_CPU

/* Drop the CPU's cached committed space back into the central pool. */
static int cpu_swap_callback(struct notifier_block *nfb,
			     unsigned long action,
			     void *hcpu)
{
	long *committed;

	committed = &per_cpu(committed_space, (long)hcpu);
523
	if (action == CPU_DEAD || action == CPU_DEAD_FROZEN) {
A
Alan Cox 已提交
524
		atomic_long_add(*committed, &vm_committed_space);
L
Linus Torvalds 已提交
525
		*committed = 0;
526
		drain_cpu_pagevecs((long)hcpu);
L
Linus Torvalds 已提交
527 528 529 530 531 532 533 534 535 536 537 538 539
	}
	return NOTIFY_OK;
}
#endif /* CONFIG_HOTPLUG_CPU */
#endif /* CONFIG_SMP */

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

P
Peter Zijlstra 已提交
540 541 542 543
#ifdef CONFIG_SWAP
	bdi_init(swapper_space.backing_dev_info);
#endif

L
Linus Torvalds 已提交
544 545 546 547 548 549 550 551 552
	/* 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
	 */
553
#ifdef CONFIG_HOTPLUG_CPU
L
Linus Torvalds 已提交
554
	hotcpu_notifier(cpu_swap_callback, 0);
555
#endif
L
Linus Torvalds 已提交
556
}