vmscan.c 80.3 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
/*
 *  linux/mm/vmscan.c
 *
 *  Copyright (C) 1991, 1992, 1993, 1994  Linus Torvalds
 *
 *  Swap reorganised 29.12.95, Stephen Tweedie.
 *  kswapd added: 7.1.96  sct
 *  Removed kswapd_ctl limits, and swap out as many pages as needed
 *  to bring the system back to freepages.high: 2.4.97, Rik van Riel.
 *  Zone aware kswapd started 02/00, Kanoj Sarcar (kanoj@sgi.com).
 *  Multiqueue VM started 5.8.00, Rik van Riel.
 */

#include <linux/mm.h>
#include <linux/module.h>
16
#include <linux/gfp.h>
L
Linus Torvalds 已提交
17 18 19 20 21
#include <linux/kernel_stat.h>
#include <linux/swap.h>
#include <linux/pagemap.h>
#include <linux/init.h>
#include <linux/highmem.h>
22
#include <linux/vmstat.h>
L
Linus Torvalds 已提交
23 24 25 26 27 28 29 30 31 32 33 34 35 36
#include <linux/file.h>
#include <linux/writeback.h>
#include <linux/blkdev.h>
#include <linux/buffer_head.h>	/* for try_to_release_page(),
					buffer_heads_over_limit */
#include <linux/mm_inline.h>
#include <linux/pagevec.h>
#include <linux/backing-dev.h>
#include <linux/rmap.h>
#include <linux/topology.h>
#include <linux/cpu.h>
#include <linux/cpuset.h>
#include <linux/notifier.h>
#include <linux/rwsem.h>
37
#include <linux/delay.h>
38
#include <linux/kthread.h>
39
#include <linux/freezer.h>
40
#include <linux/memcontrol.h>
41
#include <linux/delayacct.h>
42
#include <linux/sysctl.h>
L
Linus Torvalds 已提交
43 44 45 46 47 48

#include <asm/tlbflush.h>
#include <asm/div64.h>

#include <linux/swapops.h>

49 50
#include "internal.h"

L
Linus Torvalds 已提交
51 52 53 54
struct scan_control {
	/* Incremented by the number of inactive pages that were scanned */
	unsigned long nr_scanned;

55 56 57
	/* Number of pages freed so far during a call to shrink_zones() */
	unsigned long nr_reclaimed;

58 59 60
	/* How many pages shrink_list() should reclaim */
	unsigned long nr_to_reclaim;

61 62
	unsigned long hibernation_mode;

L
Linus Torvalds 已提交
63
	/* This context's GFP mask */
A
Al Viro 已提交
64
	gfp_t gfp_mask;
L
Linus Torvalds 已提交
65 66 67

	int may_writepage;

68 69
	/* Can mapped pages be reclaimed? */
	int may_unmap;
70

71 72 73
	/* Can pages be swapped as part of reclaim? */
	int may_swap;

74
	int swappiness;
75 76

	int all_unreclaimable;
A
Andy Whitcroft 已提交
77 78

	int order;
79 80 81 82

	/* Which cgroup do we reclaim from */
	struct mem_cgroup *mem_cgroup;

83 84 85 86 87 88
	/*
	 * Nodemask of nodes allowed by the caller. If NULL, all nodes
	 * are scanned.
	 */
	nodemask_t	*nodemask;

89 90 91 92
	/* Pluggable isolate pages callback */
	unsigned long (*isolate_pages)(unsigned long nr, struct list_head *dst,
			unsigned long *scanned, int order, int mode,
			struct zone *z, struct mem_cgroup *mem_cont,
93
			int active, int file);
L
Linus Torvalds 已提交
94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129
};

#define lru_to_page(_head) (list_entry((_head)->prev, struct page, lru))

#ifdef ARCH_HAS_PREFETCH
#define prefetch_prev_lru_page(_page, _base, _field)			\
	do {								\
		if ((_page)->lru.prev != _base) {			\
			struct page *prev;				\
									\
			prev = lru_to_page(&(_page->lru));		\
			prefetch(&prev->_field);			\
		}							\
	} while (0)
#else
#define prefetch_prev_lru_page(_page, _base, _field) do { } while (0)
#endif

#ifdef ARCH_HAS_PREFETCHW
#define prefetchw_prev_lru_page(_page, _base, _field)			\
	do {								\
		if ((_page)->lru.prev != _base) {			\
			struct page *prev;				\
									\
			prev = lru_to_page(&(_page->lru));		\
			prefetchw(&prev->_field);			\
		}							\
	} while (0)
#else
#define prefetchw_prev_lru_page(_page, _base, _field) do { } while (0)
#endif

/*
 * From 0 .. 100.  Higher means more swappy.
 */
int vm_swappiness = 60;
130
long vm_total_pages;	/* The total number of pages which the VM controls */
L
Linus Torvalds 已提交
131 132 133 134

static LIST_HEAD(shrinker_list);
static DECLARE_RWSEM(shrinker_rwsem);

135
#ifdef CONFIG_CGROUP_MEM_RES_CTLR
136
#define scanning_global_lru(sc)	(!(sc)->mem_cgroup)
137
#else
138
#define scanning_global_lru(sc)	(1)
139 140
#endif

141 142 143
static struct zone_reclaim_stat *get_reclaim_stat(struct zone *zone,
						  struct scan_control *sc)
{
144
	if (!scanning_global_lru(sc))
K
KOSAKI Motohiro 已提交
145 146
		return mem_cgroup_get_reclaim_stat(sc->mem_cgroup, zone);

147 148 149
	return &zone->reclaim_stat;
}

150 151
static unsigned long zone_nr_lru_pages(struct zone *zone,
				struct scan_control *sc, enum lru_list lru)
152
{
153
	if (!scanning_global_lru(sc))
154 155
		return mem_cgroup_zone_nr_pages(sc->mem_cgroup, zone, lru);

156 157 158 159
	return zone_page_state(zone, NR_LRU_BASE + lru);
}


L
Linus Torvalds 已提交
160 161 162
/*
 * Add a shrinker callback to be called from the vm
 */
163
void register_shrinker(struct shrinker *shrinker)
L
Linus Torvalds 已提交
164
{
165 166 167 168
	shrinker->nr = 0;
	down_write(&shrinker_rwsem);
	list_add_tail(&shrinker->list, &shrinker_list);
	up_write(&shrinker_rwsem);
L
Linus Torvalds 已提交
169
}
170
EXPORT_SYMBOL(register_shrinker);
L
Linus Torvalds 已提交
171 172 173 174

/*
 * Remove one
 */
175
void unregister_shrinker(struct shrinker *shrinker)
L
Linus Torvalds 已提交
176 177 178 179 180
{
	down_write(&shrinker_rwsem);
	list_del(&shrinker->list);
	up_write(&shrinker_rwsem);
}
181
EXPORT_SYMBOL(unregister_shrinker);
L
Linus Torvalds 已提交
182 183 184 185 186 187 188 189 190 191

#define SHRINK_BATCH 128
/*
 * Call the shrink functions to age shrinkable caches
 *
 * Here we assume it costs one seek to replace a lru page and that it also
 * takes a seek to recreate a cache object.  With this in mind we age equal
 * percentages of the lru and ageable caches.  This should balance the seeks
 * generated by these structures.
 *
S
Simon Arlott 已提交
192
 * If the vm encountered mapped pages on the LRU it increase the pressure on
L
Linus Torvalds 已提交
193 194 195 196 197 198 199
 * slab to avoid swapping.
 *
 * We do weird things to avoid (scanned*seeks*entries) overflowing 32 bits.
 *
 * `lru_pages' represents the number of on-LRU pages in all the zones which
 * are eligible for the caller's allocation attempt.  It is used for balancing
 * slab reclaim versus page reclaim.
200 201
 *
 * Returns the number of slab objects which we shrunk.
L
Linus Torvalds 已提交
202
 */
203 204
unsigned long shrink_slab(unsigned long scanned, gfp_t gfp_mask,
			unsigned long lru_pages)
L
Linus Torvalds 已提交
205 206
{
	struct shrinker *shrinker;
207
	unsigned long ret = 0;
L
Linus Torvalds 已提交
208 209 210 211 212

	if (scanned == 0)
		scanned = SWAP_CLUSTER_MAX;

	if (!down_read_trylock(&shrinker_rwsem))
213
		return 1;	/* Assume we'll be able to shrink next time */
L
Linus Torvalds 已提交
214 215 216 217

	list_for_each_entry(shrinker, &shrinker_list, list) {
		unsigned long long delta;
		unsigned long total_scan;
218
		unsigned long max_pass = (*shrinker->shrink)(0, gfp_mask);
L
Linus Torvalds 已提交
219 220

		delta = (4 * scanned) / shrinker->seeks;
221
		delta *= max_pass;
L
Linus Torvalds 已提交
222 223
		do_div(delta, lru_pages + 1);
		shrinker->nr += delta;
224
		if (shrinker->nr < 0) {
225 226 227
			printk(KERN_ERR "shrink_slab: %pF negative objects to "
			       "delete nr=%ld\n",
			       shrinker->shrink, shrinker->nr);
228 229 230 231 232 233 234 235 236 237
			shrinker->nr = max_pass;
		}

		/*
		 * Avoid risking looping forever due to too large nr value:
		 * never try to free more than twice the estimate number of
		 * freeable entries.
		 */
		if (shrinker->nr > max_pass * 2)
			shrinker->nr = max_pass * 2;
L
Linus Torvalds 已提交
238 239 240 241 242 243 244

		total_scan = shrinker->nr;
		shrinker->nr = 0;

		while (total_scan >= SHRINK_BATCH) {
			long this_scan = SHRINK_BATCH;
			int shrink_ret;
245
			int nr_before;
L
Linus Torvalds 已提交
246

247 248
			nr_before = (*shrinker->shrink)(0, gfp_mask);
			shrink_ret = (*shrinker->shrink)(this_scan, gfp_mask);
L
Linus Torvalds 已提交
249 250
			if (shrink_ret == -1)
				break;
251 252
			if (shrink_ret < nr_before)
				ret += nr_before - shrink_ret;
253
			count_vm_events(SLABS_SCANNED, this_scan);
L
Linus Torvalds 已提交
254 255 256 257 258 259 260 261
			total_scan -= this_scan;

			cond_resched();
		}

		shrinker->nr += total_scan;
	}
	up_read(&shrinker_rwsem);
262
	return ret;
L
Linus Torvalds 已提交
263 264 265 266
}

static inline int is_page_cache_freeable(struct page *page)
{
267 268 269 270 271
	/*
	 * A freeable page cache page is referenced only by the caller
	 * that isolated the page, the page cache radix tree and
	 * optional buffer heads at page->private.
	 */
272
	return page_count(page) - page_has_private(page) == 2;
L
Linus Torvalds 已提交
273 274 275 276
}

static int may_write_to_queue(struct backing_dev_info *bdi)
{
277
	if (current->flags & PF_SWAPWRITE)
L
Linus Torvalds 已提交
278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301
		return 1;
	if (!bdi_write_congested(bdi))
		return 1;
	if (bdi == current->backing_dev_info)
		return 1;
	return 0;
}

/*
 * We detected a synchronous write error writing a page out.  Probably
 * -ENOSPC.  We need to propagate that into the address_space for a subsequent
 * fsync(), msync() or close().
 *
 * The tricky part is that after writepage we cannot touch the mapping: nothing
 * prevents it from being freed up.  But we have a ref on the page and once
 * that page is locked, the mapping is pinned.
 *
 * We're allowed to run sleeping lock_page() here because we know the caller has
 * __GFP_FS.
 */
static void handle_write_error(struct address_space *mapping,
				struct page *page, int error)
{
	lock_page(page);
302 303
	if (page_mapping(page) == mapping)
		mapping_set_error(mapping, error);
L
Linus Torvalds 已提交
304 305 306
	unlock_page(page);
}

307 308 309 310 311 312
/* Request for sync pageout. */
enum pageout_io {
	PAGEOUT_IO_ASYNC,
	PAGEOUT_IO_SYNC,
};

313 314 315 316 317 318 319 320 321 322 323 324
/* possible outcome of pageout() */
typedef enum {
	/* failed to write page out, page is locked */
	PAGE_KEEP,
	/* move page to the active list, page is locked */
	PAGE_ACTIVATE,
	/* page has been sent to the disk successfully, page is unlocked */
	PAGE_SUCCESS,
	/* page is clean and locked */
	PAGE_CLEAN,
} pageout_t;

L
Linus Torvalds 已提交
325
/*
A
Andrew Morton 已提交
326 327
 * pageout is called by shrink_page_list() for each dirty page.
 * Calls ->writepage().
L
Linus Torvalds 已提交
328
 */
329 330
static pageout_t pageout(struct page *page, struct address_space *mapping,
						enum pageout_io sync_writeback)
L
Linus Torvalds 已提交
331 332 333 334 335 336 337 338
{
	/*
	 * If the page is dirty, only perform writeback if that write
	 * will be non-blocking.  To prevent this allocation from being
	 * stalled by pagecache activity.  But note that there may be
	 * stalls if we need to run get_block().  We could test
	 * PagePrivate for that.
	 *
339
	 * If this process is currently in __generic_file_aio_write() against
L
Linus Torvalds 已提交
340 341 342 343 344 345 346 347 348 349 350 351 352 353 354
	 * this page's queue, we can perform writeback even if that
	 * will block.
	 *
	 * If the page is swapcache, write it back even if that would
	 * block, for some throttling. This happens by accident, because
	 * swap_backing_dev_info is bust: it doesn't reflect the
	 * congestion state of the swapdevs.  Easy to fix, if needed.
	 */
	if (!is_page_cache_freeable(page))
		return PAGE_KEEP;
	if (!mapping) {
		/*
		 * Some data journaling orphaned pages can have
		 * page->mapping == NULL while being dirty with clean buffers.
		 */
355
		if (page_has_private(page)) {
L
Linus Torvalds 已提交
356 357
			if (try_to_free_buffers(page)) {
				ClearPageDirty(page);
358
				printk("%s: orphaned page\n", __func__);
L
Linus Torvalds 已提交
359 360 361 362 363 364 365 366 367 368 369 370 371 372 373
				return PAGE_CLEAN;
			}
		}
		return PAGE_KEEP;
	}
	if (mapping->a_ops->writepage == NULL)
		return PAGE_ACTIVATE;
	if (!may_write_to_queue(mapping->backing_dev_info))
		return PAGE_KEEP;

	if (clear_page_dirty_for_io(page)) {
		int res;
		struct writeback_control wbc = {
			.sync_mode = WB_SYNC_NONE,
			.nr_to_write = SWAP_CLUSTER_MAX,
374 375
			.range_start = 0,
			.range_end = LLONG_MAX,
L
Linus Torvalds 已提交
376 377 378 379 380 381 382 383
			.nonblocking = 1,
			.for_reclaim = 1,
		};

		SetPageReclaim(page);
		res = mapping->a_ops->writepage(page, &wbc);
		if (res < 0)
			handle_write_error(mapping, page, res);
384
		if (res == AOP_WRITEPAGE_ACTIVATE) {
L
Linus Torvalds 已提交
385 386 387
			ClearPageReclaim(page);
			return PAGE_ACTIVATE;
		}
388 389 390 391 392 393 394 395 396

		/*
		 * Wait on writeback if requested to. This happens when
		 * direct reclaiming a large contiguous area and the
		 * first attempt to free a range of pages fails.
		 */
		if (PageWriteback(page) && sync_writeback == PAGEOUT_IO_SYNC)
			wait_on_page_writeback(page);

L
Linus Torvalds 已提交
397 398 399 400
		if (!PageWriteback(page)) {
			/* synchronous write or broken a_ops? */
			ClearPageReclaim(page);
		}
401
		inc_zone_page_state(page, NR_VMSCAN_WRITE);
L
Linus Torvalds 已提交
402 403 404 405 406 407
		return PAGE_SUCCESS;
	}

	return PAGE_CLEAN;
}

408
/*
N
Nick Piggin 已提交
409 410
 * Same as remove_mapping, but if the page is removed from the mapping, it
 * gets returned with a refcount of 0.
411
 */
N
Nick Piggin 已提交
412
static int __remove_mapping(struct address_space *mapping, struct page *page)
413
{
414 415
	BUG_ON(!PageLocked(page));
	BUG_ON(mapping != page_mapping(page));
416

N
Nick Piggin 已提交
417
	spin_lock_irq(&mapping->tree_lock);
418
	/*
N
Nick Piggin 已提交
419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441
	 * The non racy check for a busy page.
	 *
	 * Must be careful with the order of the tests. When someone has
	 * a ref to the page, it may be possible that they dirty it then
	 * drop the reference. So if PageDirty is tested before page_count
	 * here, then the following race may occur:
	 *
	 * get_user_pages(&page);
	 * [user mapping goes away]
	 * write_to(page);
	 *				!PageDirty(page)    [good]
	 * SetPageDirty(page);
	 * put_page(page);
	 *				!page_count(page)   [good, discard it]
	 *
	 * [oops, our write_to data is lost]
	 *
	 * Reversing the order of the tests ensures such a situation cannot
	 * escape unnoticed. The smp_rmb is needed to ensure the page->flags
	 * load is not satisfied before that of page->_count.
	 *
	 * Note that if SetPageDirty is always performed via set_page_dirty,
	 * and thus under tree_lock, then this ordering is not required.
442
	 */
N
Nick Piggin 已提交
443
	if (!page_freeze_refs(page, 2))
444
		goto cannot_free;
N
Nick Piggin 已提交
445 446 447
	/* note: atomic_cmpxchg in page_freeze_refs provides the smp_rmb */
	if (unlikely(PageDirty(page))) {
		page_unfreeze_refs(page, 2);
448
		goto cannot_free;
N
Nick Piggin 已提交
449
	}
450 451 452 453

	if (PageSwapCache(page)) {
		swp_entry_t swap = { .val = page_private(page) };
		__delete_from_swap_cache(page);
N
Nick Piggin 已提交
454
		spin_unlock_irq(&mapping->tree_lock);
455
		swapcache_free(swap, page);
N
Nick Piggin 已提交
456 457
	} else {
		__remove_from_page_cache(page);
N
Nick Piggin 已提交
458
		spin_unlock_irq(&mapping->tree_lock);
459
		mem_cgroup_uncharge_cache_page(page);
460 461 462 463 464
	}

	return 1;

cannot_free:
N
Nick Piggin 已提交
465
	spin_unlock_irq(&mapping->tree_lock);
466 467 468
	return 0;
}

N
Nick Piggin 已提交
469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488
/*
 * Attempt to detach a locked page from its ->mapping.  If it is dirty or if
 * someone else has a ref on the page, abort and return 0.  If it was
 * successfully detached, return 1.  Assumes the caller has a single ref on
 * this page.
 */
int remove_mapping(struct address_space *mapping, struct page *page)
{
	if (__remove_mapping(mapping, page)) {
		/*
		 * Unfreezing the refcount with 1 rather than 2 effectively
		 * drops the pagecache ref for us without requiring another
		 * atomic operation.
		 */
		page_unfreeze_refs(page, 1);
		return 1;
	}
	return 0;
}

L
Lee Schermerhorn 已提交
489 490 491 492 493 494 495 496 497 498 499 500 501
/**
 * putback_lru_page - put previously isolated page onto appropriate LRU list
 * @page: page to be put back to appropriate lru list
 *
 * Add previously isolated @page to appropriate LRU list.
 * Page may still be unevictable for other reasons.
 *
 * lru_lock must not be held, interrupts must be enabled.
 */
void putback_lru_page(struct page *page)
{
	int lru;
	int active = !!TestClearPageActive(page);
502
	int was_unevictable = PageUnevictable(page);
L
Lee Schermerhorn 已提交
503 504 505 506 507 508 509 510 511 512 513 514 515

	VM_BUG_ON(PageLRU(page));

redo:
	ClearPageUnevictable(page);

	if (page_evictable(page, NULL)) {
		/*
		 * For evictable pages, we can use the cache.
		 * In event of a race, worst case is we end up with an
		 * unevictable page on [in]active list.
		 * We know how to handle that.
		 */
516
		lru = active + page_lru_base_type(page);
L
Lee Schermerhorn 已提交
517 518 519 520 521 522 523 524
		lru_cache_add_lru(page, lru);
	} else {
		/*
		 * Put unevictable pages directly on zone's unevictable
		 * list.
		 */
		lru = LRU_UNEVICTABLE;
		add_page_to_unevictable_list(page);
525 526 527 528 529 530 531 532 533 534
		/*
		 * When racing with an mlock clearing (page is
		 * unlocked), make sure that if the other thread does
		 * not observe our setting of PG_lru and fails
		 * isolation, we see PG_mlocked cleared below and move
		 * the page back to the evictable list.
		 *
		 * The other side is TestClearPageMlocked().
		 */
		smp_mb();
L
Lee Schermerhorn 已提交
535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552
	}

	/*
	 * page's status can change while we move it among lru. If an evictable
	 * page is on unevictable list, it never be freed. To avoid that,
	 * check after we added it to the list, again.
	 */
	if (lru == LRU_UNEVICTABLE && page_evictable(page, NULL)) {
		if (!isolate_lru_page(page)) {
			put_page(page);
			goto redo;
		}
		/* This means someone else dropped this page from LRU
		 * So, it will be freed or putback to LRU again. There is
		 * nothing to do here.
		 */
	}

553 554 555 556 557
	if (was_unevictable && lru != LRU_UNEVICTABLE)
		count_vm_event(UNEVICTABLE_PGRESCUED);
	else if (!was_unevictable && lru == LRU_UNEVICTABLE)
		count_vm_event(UNEVICTABLE_PGCULLED);

L
Lee Schermerhorn 已提交
558 559 560
	put_page(page);		/* drop ref from isolate */
}

561 562 563
enum page_references {
	PAGEREF_RECLAIM,
	PAGEREF_RECLAIM_CLEAN,
564
	PAGEREF_KEEP,
565 566 567 568 569 570
	PAGEREF_ACTIVATE,
};

static enum page_references page_check_references(struct page *page,
						  struct scan_control *sc)
{
571
	int referenced_ptes, referenced_page;
572 573
	unsigned long vm_flags;

574 575
	referenced_ptes = page_referenced(page, 1, sc->mem_cgroup, &vm_flags);
	referenced_page = TestClearPageReferenced(page);
576 577 578 579 580 581 582 583 584 585 586 587

	/* Lumpy reclaim - ignore references */
	if (sc->order > PAGE_ALLOC_COSTLY_ORDER)
		return PAGEREF_RECLAIM;

	/*
	 * Mlock lost the isolation race with us.  Let try_to_unmap()
	 * move the page to the unevictable list.
	 */
	if (vm_flags & VM_LOCKED)
		return PAGEREF_RECLAIM;

588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611
	if (referenced_ptes) {
		if (PageAnon(page))
			return PAGEREF_ACTIVATE;
		/*
		 * All mapped pages start out with page table
		 * references from the instantiating fault, so we need
		 * to look twice if a mapped file page is used more
		 * than once.
		 *
		 * Mark it and spare it for another trip around the
		 * inactive list.  Another page table reference will
		 * lead to its activation.
		 *
		 * Note: the mark is set for activated pages as well
		 * so that recently deactivated but used pages are
		 * quickly recovered.
		 */
		SetPageReferenced(page);

		if (referenced_page)
			return PAGEREF_ACTIVATE;

		return PAGEREF_KEEP;
	}
612 613

	/* Reclaim if clean, defer dirty pages to writeback */
614 615 616 617
	if (referenced_page)
		return PAGEREF_RECLAIM_CLEAN;

	return PAGEREF_RECLAIM;
618 619
}

L
Linus Torvalds 已提交
620
/*
A
Andrew Morton 已提交
621
 * shrink_page_list() returns the number of reclaimed pages
L
Linus Torvalds 已提交
622
 */
A
Andrew Morton 已提交
623
static unsigned long shrink_page_list(struct list_head *page_list,
624 625
					struct scan_control *sc,
					enum pageout_io sync_writeback)
L
Linus Torvalds 已提交
626 627 628 629
{
	LIST_HEAD(ret_pages);
	struct pagevec freed_pvec;
	int pgactivate = 0;
630
	unsigned long nr_reclaimed = 0;
L
Linus Torvalds 已提交
631 632 633 634 635

	cond_resched();

	pagevec_init(&freed_pvec, 1);
	while (!list_empty(page_list)) {
636
		enum page_references references;
L
Linus Torvalds 已提交
637 638 639 640 641 642 643 644 645
		struct address_space *mapping;
		struct page *page;
		int may_enter_fs;

		cond_resched();

		page = lru_to_page(page_list);
		list_del(&page->lru);

N
Nick Piggin 已提交
646
		if (!trylock_page(page))
L
Linus Torvalds 已提交
647 648
			goto keep;

N
Nick Piggin 已提交
649
		VM_BUG_ON(PageActive(page));
L
Linus Torvalds 已提交
650 651

		sc->nr_scanned++;
652

N
Nick Piggin 已提交
653 654
		if (unlikely(!page_evictable(page, NULL)))
			goto cull_mlocked;
L
Lee Schermerhorn 已提交
655

656
		if (!sc->may_unmap && page_mapped(page))
657 658
			goto keep_locked;

L
Linus Torvalds 已提交
659 660 661 662
		/* Double the slab pressure for mapped and swapcache pages */
		if (page_mapped(page) || PageSwapCache(page))
			sc->nr_scanned++;

663 664 665 666 667 668 669 670 671 672 673 674 675 676
		may_enter_fs = (sc->gfp_mask & __GFP_FS) ||
			(PageSwapCache(page) && (sc->gfp_mask & __GFP_IO));

		if (PageWriteback(page)) {
			/*
			 * Synchronous reclaim is performed in two passes,
			 * first an asynchronous pass over the list to
			 * start parallel writeback, and a second synchronous
			 * pass to wait for the IO to complete.  Wait here
			 * for any page for which writeback has already
			 * started.
			 */
			if (sync_writeback == PAGEOUT_IO_SYNC && may_enter_fs)
				wait_on_page_writeback(page);
677
			else
678 679
				goto keep_locked;
		}
L
Linus Torvalds 已提交
680

681 682 683
		references = page_check_references(page, sc);
		switch (references) {
		case PAGEREF_ACTIVATE:
L
Linus Torvalds 已提交
684
			goto activate_locked;
685 686
		case PAGEREF_KEEP:
			goto keep_locked;
687 688 689 690
		case PAGEREF_RECLAIM:
		case PAGEREF_RECLAIM_CLEAN:
			; /* try to reclaim the page below */
		}
L
Linus Torvalds 已提交
691 692 693 694 695

		/*
		 * Anonymous process memory has backing store?
		 * Try to allocate it some swap space here.
		 */
N
Nick Piggin 已提交
696
		if (PageAnon(page) && !PageSwapCache(page)) {
697 698
			if (!(sc->gfp_mask & __GFP_IO))
				goto keep_locked;
699
			if (!add_to_swap(page))
L
Linus Torvalds 已提交
700
				goto activate_locked;
701
			may_enter_fs = 1;
N
Nick Piggin 已提交
702
		}
L
Linus Torvalds 已提交
703 704 705 706 707 708 709 710

		mapping = page_mapping(page);

		/*
		 * The page is mapped into the page tables of one or more
		 * processes. Try to unmap it here.
		 */
		if (page_mapped(page) && mapping) {
711
			switch (try_to_unmap(page, TTU_UNMAP)) {
L
Linus Torvalds 已提交
712 713 714 715
			case SWAP_FAIL:
				goto activate_locked;
			case SWAP_AGAIN:
				goto keep_locked;
N
Nick Piggin 已提交
716 717
			case SWAP_MLOCK:
				goto cull_mlocked;
L
Linus Torvalds 已提交
718 719 720 721 722 723
			case SWAP_SUCCESS:
				; /* try to free the page below */
			}
		}

		if (PageDirty(page)) {
724
			if (references == PAGEREF_RECLAIM_CLEAN)
L
Linus Torvalds 已提交
725
				goto keep_locked;
726
			if (!may_enter_fs)
L
Linus Torvalds 已提交
727
				goto keep_locked;
728
			if (!sc->may_writepage)
L
Linus Torvalds 已提交
729 730 731
				goto keep_locked;

			/* Page is dirty, try to write it out here */
732
			switch (pageout(page, mapping, sync_writeback)) {
L
Linus Torvalds 已提交
733 734 735 736 737
			case PAGE_KEEP:
				goto keep_locked;
			case PAGE_ACTIVATE:
				goto activate_locked;
			case PAGE_SUCCESS:
738
				if (PageWriteback(page) || PageDirty(page))
L
Linus Torvalds 已提交
739 740 741 742 743
					goto keep;
				/*
				 * A synchronous write - probably a ramdisk.  Go
				 * ahead and try to reclaim the page.
				 */
N
Nick Piggin 已提交
744
				if (!trylock_page(page))
L
Linus Torvalds 已提交
745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763
					goto keep;
				if (PageDirty(page) || PageWriteback(page))
					goto keep_locked;
				mapping = page_mapping(page);
			case PAGE_CLEAN:
				; /* try to free the page below */
			}
		}

		/*
		 * If the page has buffers, try to free the buffer mappings
		 * associated with this page. If we succeed we try to free
		 * the page as well.
		 *
		 * We do this even if the page is PageDirty().
		 * try_to_release_page() does not perform I/O, but it is
		 * possible for a page to have PageDirty set, but it is actually
		 * clean (all its buffers are clean).  This happens if the
		 * buffers were written out directly, with submit_bh(). ext3
L
Lee Schermerhorn 已提交
764
		 * will do this, as well as the blockdev mapping.
L
Linus Torvalds 已提交
765 766 767 768 769 770 771 772 773 774
		 * try_to_release_page() will discover that cleanness and will
		 * drop the buffers and mark the page clean - it can be freed.
		 *
		 * Rarely, pages can have buffers and no ->mapping.  These are
		 * the pages which were not successfully invalidated in
		 * truncate_complete_page().  We try to drop those buffers here
		 * and if that worked, and the page is no longer mapped into
		 * process address space (page_count == 1) it can be freed.
		 * Otherwise, leave the page on the LRU so it is swappable.
		 */
775
		if (page_has_private(page)) {
L
Linus Torvalds 已提交
776 777
			if (!try_to_release_page(page, sc->gfp_mask))
				goto activate_locked;
N
Nick Piggin 已提交
778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793
			if (!mapping && page_count(page) == 1) {
				unlock_page(page);
				if (put_page_testzero(page))
					goto free_it;
				else {
					/*
					 * rare race with speculative reference.
					 * the speculative reference will free
					 * this page shortly, so we may
					 * increment nr_reclaimed here (and
					 * leave it off the LRU).
					 */
					nr_reclaimed++;
					continue;
				}
			}
L
Linus Torvalds 已提交
794 795
		}

N
Nick Piggin 已提交
796
		if (!mapping || !__remove_mapping(mapping, page))
797
			goto keep_locked;
L
Linus Torvalds 已提交
798

N
Nick Piggin 已提交
799 800 801 802 803 804 805 806
		/*
		 * At this point, we have no other references and there is
		 * no way to pick any more up (removed from LRU, removed
		 * from pagecache). Can use non-atomic bitops now (and
		 * we obviously don't have to worry about waking up a process
		 * waiting on the page lock, because there are no references.
		 */
		__clear_page_locked(page);
N
Nick Piggin 已提交
807
free_it:
808
		nr_reclaimed++;
N
Nick Piggin 已提交
809 810 811 812
		if (!pagevec_add(&freed_pvec, page)) {
			__pagevec_free(&freed_pvec);
			pagevec_reinit(&freed_pvec);
		}
L
Linus Torvalds 已提交
813 814
		continue;

N
Nick Piggin 已提交
815
cull_mlocked:
816 817
		if (PageSwapCache(page))
			try_to_free_swap(page);
N
Nick Piggin 已提交
818 819 820 821
		unlock_page(page);
		putback_lru_page(page);
		continue;

L
Linus Torvalds 已提交
822
activate_locked:
823 824
		/* Not a candidate for swapping, so reclaim swap space. */
		if (PageSwapCache(page) && vm_swap_full())
825
			try_to_free_swap(page);
L
Lee Schermerhorn 已提交
826
		VM_BUG_ON(PageActive(page));
L
Linus Torvalds 已提交
827 828 829 830 831 832
		SetPageActive(page);
		pgactivate++;
keep_locked:
		unlock_page(page);
keep:
		list_add(&page->lru, &ret_pages);
N
Nick Piggin 已提交
833
		VM_BUG_ON(PageLRU(page) || PageUnevictable(page));
L
Linus Torvalds 已提交
834 835 836
	}
	list_splice(&ret_pages, page_list);
	if (pagevec_count(&freed_pvec))
N
Nick Piggin 已提交
837
		__pagevec_free(&freed_pvec);
838
	count_vm_events(PGACTIVATE, pgactivate);
839
	return nr_reclaimed;
L
Linus Torvalds 已提交
840 841
}

A
Andy Whitcroft 已提交
842 843 844 845 846 847 848 849 850 851
/*
 * Attempt to remove the specified page from its LRU.  Only take this page
 * if it is of the appropriate PageActive status.  Pages which are being
 * freed elsewhere are also ignored.
 *
 * page:	page to consider
 * mode:	one of the LRU isolation modes defined above
 *
 * returns 0 on success, -ve errno on failure.
 */
852
int __isolate_lru_page(struct page *page, int mode, int file)
A
Andy Whitcroft 已提交
853 854 855 856 857 858 859 860 861 862 863 864 865 866 867
{
	int ret = -EINVAL;

	/* Only take pages on the LRU. */
	if (!PageLRU(page))
		return ret;

	/*
	 * When checking the active state, we need to be sure we are
	 * dealing with comparible boolean values.  Take the logical not
	 * of each.
	 */
	if (mode != ISOLATE_BOTH && (!PageActive(page) != !mode))
		return ret;

868
	if (mode != ISOLATE_BOTH && page_is_file_cache(page) != file)
869 870
		return ret;

L
Lee Schermerhorn 已提交
871 872 873 874 875 876 877 878
	/*
	 * When this function is being called for lumpy reclaim, we
	 * initially look into all LRU pages, active, inactive and
	 * unevictable; only give shrink_page_list evictable pages.
	 */
	if (PageUnevictable(page))
		return ret;

A
Andy Whitcroft 已提交
879
	ret = -EBUSY;
K
KAMEZAWA Hiroyuki 已提交
880

A
Andy Whitcroft 已提交
881 882 883 884 885 886 887 888 889 890 891 892 893
	if (likely(get_page_unless_zero(page))) {
		/*
		 * Be careful not to clear PageLRU until after we're
		 * sure the page is not being freed elsewhere -- the
		 * page release code relies on it.
		 */
		ClearPageLRU(page);
		ret = 0;
	}

	return ret;
}

L
Linus Torvalds 已提交
894 895 896 897 898 899 900 901 902 903 904 905 906 907
/*
 * zone->lru_lock is heavily contended.  Some of the functions that
 * shrink the lists perform better by taking out a batch of pages
 * and working on them outside the LRU lock.
 *
 * For pagecache intensive workloads, this function is the hottest
 * spot in the kernel (apart from copy_*_user functions).
 *
 * Appropriate locks must be held before calling this function.
 *
 * @nr_to_scan:	The number of pages to look through on the list.
 * @src:	The LRU list to pull pages off.
 * @dst:	The temp list to put pages on to.
 * @scanned:	The number of pages that were scanned.
A
Andy Whitcroft 已提交
908 909
 * @order:	The caller's attempted allocation order
 * @mode:	One of the LRU isolation modes
910
 * @file:	True [1] if isolating file [!anon] pages
L
Linus Torvalds 已提交
911 912 913
 *
 * returns how many pages were moved onto *@dst.
 */
914 915
static unsigned long isolate_lru_pages(unsigned long nr_to_scan,
		struct list_head *src, struct list_head *dst,
916
		unsigned long *scanned, int order, int mode, int file)
L
Linus Torvalds 已提交
917
{
918
	unsigned long nr_taken = 0;
919
	unsigned long scan;
L
Linus Torvalds 已提交
920

921
	for (scan = 0; scan < nr_to_scan && !list_empty(src); scan++) {
A
Andy Whitcroft 已提交
922 923 924 925 926 927
		struct page *page;
		unsigned long pfn;
		unsigned long end_pfn;
		unsigned long page_pfn;
		int zone_id;

L
Linus Torvalds 已提交
928 929 930
		page = lru_to_page(src);
		prefetchw_prev_lru_page(page, src, flags);

N
Nick Piggin 已提交
931
		VM_BUG_ON(!PageLRU(page));
N
Nick Piggin 已提交
932

933
		switch (__isolate_lru_page(page, mode, file)) {
A
Andy Whitcroft 已提交
934 935
		case 0:
			list_move(&page->lru, dst);
936
			mem_cgroup_del_lru(page);
937
			nr_taken++;
A
Andy Whitcroft 已提交
938 939 940 941 942
			break;

		case -EBUSY:
			/* else it is being freed elsewhere */
			list_move(&page->lru, src);
943
			mem_cgroup_rotate_lru_list(page, page_lru(page));
A
Andy Whitcroft 已提交
944
			continue;
945

A
Andy Whitcroft 已提交
946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977
		default:
			BUG();
		}

		if (!order)
			continue;

		/*
		 * Attempt to take all pages in the order aligned region
		 * surrounding the tag page.  Only take those pages of
		 * the same active state as that tag page.  We may safely
		 * round the target page pfn down to the requested order
		 * as the mem_map is guarenteed valid out to MAX_ORDER,
		 * where that page is in a different zone we will detect
		 * it from its zone id and abort this block scan.
		 */
		zone_id = page_zone_id(page);
		page_pfn = page_to_pfn(page);
		pfn = page_pfn & ~((1 << order) - 1);
		end_pfn = pfn + (1 << order);
		for (; pfn < end_pfn; pfn++) {
			struct page *cursor_page;

			/* The target page is in the block, ignore it. */
			if (unlikely(pfn == page_pfn))
				continue;

			/* Avoid holes within the zone. */
			if (unlikely(!pfn_valid_within(pfn)))
				break;

			cursor_page = pfn_to_page(pfn);
978

A
Andy Whitcroft 已提交
979 980 981
			/* Check that we have not crossed a zone boundary. */
			if (unlikely(page_zone_id(cursor_page) != zone_id))
				continue;
982 983 984 985 986 987 988 989 990 991

			/*
			 * If we don't have enough swap space, reclaiming of
			 * anon page which don't already have a swap slot is
			 * pointless.
			 */
			if (nr_swap_pages <= 0 && PageAnon(cursor_page) &&
					!PageSwapCache(cursor_page))
				continue;

992
			if (__isolate_lru_page(cursor_page, mode, file) == 0) {
A
Andy Whitcroft 已提交
993
				list_move(&cursor_page->lru, dst);
994
				mem_cgroup_del_lru(cursor_page);
A
Andy Whitcroft 已提交
995 996 997 998
				nr_taken++;
				scan++;
			}
		}
L
Linus Torvalds 已提交
999 1000 1001 1002 1003 1004
	}

	*scanned = scan;
	return nr_taken;
}

1005 1006 1007 1008 1009
static unsigned long isolate_pages_global(unsigned long nr,
					struct list_head *dst,
					unsigned long *scanned, int order,
					int mode, struct zone *z,
					struct mem_cgroup *mem_cont,
1010
					int active, int file)
1011
{
1012
	int lru = LRU_BASE;
1013
	if (active)
1014 1015 1016 1017
		lru += LRU_ACTIVE;
	if (file)
		lru += LRU_FILE;
	return isolate_lru_pages(nr, &z->lru[lru].list, dst, scanned, order,
1018
								mode, file);
1019 1020
}

A
Andy Whitcroft 已提交
1021 1022 1023 1024
/*
 * clear_active_flags() is a helper for shrink_active_list(), clearing
 * any active bits from the pages in the list.
 */
1025 1026
static unsigned long clear_active_flags(struct list_head *page_list,
					unsigned int *count)
A
Andy Whitcroft 已提交
1027 1028
{
	int nr_active = 0;
1029
	int lru;
A
Andy Whitcroft 已提交
1030 1031
	struct page *page;

1032
	list_for_each_entry(page, page_list, lru) {
1033
		lru = page_lru_base_type(page);
A
Andy Whitcroft 已提交
1034
		if (PageActive(page)) {
1035
			lru += LRU_ACTIVE;
A
Andy Whitcroft 已提交
1036 1037 1038
			ClearPageActive(page);
			nr_active++;
		}
1039 1040
		count[lru]++;
	}
A
Andy Whitcroft 已提交
1041 1042 1043 1044

	return nr_active;
}

1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055
/**
 * isolate_lru_page - tries to isolate a page from its LRU list
 * @page: page to isolate from its LRU list
 *
 * Isolates a @page from an LRU list, clears PageLRU and adjusts the
 * vmstat statistic corresponding to whatever LRU list the page was on.
 *
 * Returns 0 if the page was removed from an LRU list.
 * Returns -EBUSY if the page was not on an LRU list.
 *
 * The returned page will have PageLRU() cleared.  If it was found on
L
Lee Schermerhorn 已提交
1056 1057 1058
 * the active list, it will have PageActive set.  If it was found on
 * the unevictable list, it will have the PageUnevictable bit set. That flag
 * may need to be cleared by the caller before letting the page go.
1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078
 *
 * The vmstat statistic corresponding to the list on which the page was
 * found will be decremented.
 *
 * Restrictions:
 * (1) Must be called with an elevated refcount on the page. This is a
 *     fundamentnal difference from isolate_lru_pages (which is called
 *     without a stable reference).
 * (2) the lru_lock must not be held.
 * (3) interrupts must be enabled.
 */
int isolate_lru_page(struct page *page)
{
	int ret = -EBUSY;

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

		spin_lock_irq(&zone->lru_lock);
		if (PageLRU(page) && get_page_unless_zero(page)) {
L
Lee Schermerhorn 已提交
1079
			int lru = page_lru(page);
1080 1081
			ret = 0;
			ClearPageLRU(page);
1082 1083

			del_page_from_lru_list(zone, page, lru);
1084 1085 1086 1087 1088 1089
		}
		spin_unlock_irq(&zone->lru_lock);
	}
	return ret;
}

1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114
/*
 * Are there way too many processes in the direct reclaim path already?
 */
static int too_many_isolated(struct zone *zone, int file,
		struct scan_control *sc)
{
	unsigned long inactive, isolated;

	if (current_is_kswapd())
		return 0;

	if (!scanning_global_lru(sc))
		return 0;

	if (file) {
		inactive = zone_page_state(zone, NR_INACTIVE_FILE);
		isolated = zone_page_state(zone, NR_ISOLATED_FILE);
	} else {
		inactive = zone_page_state(zone, NR_INACTIVE_ANON);
		isolated = zone_page_state(zone, NR_ISOLATED_ANON);
	}

	return isolated > inactive;
}

L
Linus Torvalds 已提交
1115
/*
A
Andrew Morton 已提交
1116 1117
 * shrink_inactive_list() is a helper for shrink_zone().  It returns the number
 * of reclaimed pages
L
Linus Torvalds 已提交
1118
 */
A
Andrew Morton 已提交
1119
static unsigned long shrink_inactive_list(unsigned long max_scan,
R
Rik van Riel 已提交
1120 1121
			struct zone *zone, struct scan_control *sc,
			int priority, int file)
L
Linus Torvalds 已提交
1122 1123 1124
{
	LIST_HEAD(page_list);
	struct pagevec pvec;
1125
	unsigned long nr_scanned = 0;
1126
	unsigned long nr_reclaimed = 0;
1127
	struct zone_reclaim_stat *reclaim_stat = get_reclaim_stat(zone, sc);
1128 1129
	int lumpy_reclaim = 0;

1130
	while (unlikely(too_many_isolated(zone, file, sc))) {
1131
		congestion_wait(BLK_RW_ASYNC, HZ/10);
1132 1133 1134 1135 1136 1137

		/* We are about to die and free our memory. Return now. */
		if (fatal_signal_pending(current))
			return SWAP_CLUSTER_MAX;
	}

1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148
	/*
	 * If we need a large contiguous chunk of memory, or have
	 * trouble getting a small set of contiguous pages, we
	 * will reclaim both active and inactive pages.
	 *
	 * We use the same threshold as pageout congestion_wait below.
	 */
	if (sc->order > PAGE_ALLOC_COSTLY_ORDER)
		lumpy_reclaim = 1;
	else if (sc->order && priority < DEF_PRIORITY - 2)
		lumpy_reclaim = 1;
L
Linus Torvalds 已提交
1149 1150 1151 1152 1153

	pagevec_init(&pvec, 1);

	lru_add_drain();
	spin_lock_irq(&zone->lru_lock);
1154
	do {
L
Linus Torvalds 已提交
1155
		struct page *page;
1156 1157 1158
		unsigned long nr_taken;
		unsigned long nr_scan;
		unsigned long nr_freed;
A
Andy Whitcroft 已提交
1159
		unsigned long nr_active;
1160
		unsigned int count[NR_LRU_LISTS] = { 0, };
1161
		int mode = lumpy_reclaim ? ISOLATE_BOTH : ISOLATE_INACTIVE;
K
KOSAKI Motohiro 已提交
1162 1163
		unsigned long nr_anon;
		unsigned long nr_file;
L
Linus Torvalds 已提交
1164

K
KOSAKI Motohiro 已提交
1165
		nr_taken = sc->isolate_pages(SWAP_CLUSTER_MAX,
1166 1167
			     &page_list, &nr_scan, sc->order, mode,
				zone, sc->mem_cgroup, 0, file);
1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181

		if (scanning_global_lru(sc)) {
			zone->pages_scanned += nr_scan;
			if (current_is_kswapd())
				__count_zone_vm_events(PGSCAN_KSWAPD, zone,
						       nr_scan);
			else
				__count_zone_vm_events(PGSCAN_DIRECT, zone,
						       nr_scan);
		}

		if (nr_taken == 0)
			goto done;

1182
		nr_active = clear_active_flags(&page_list, count);
1183
		__count_vm_events(PGDEACTIVATE, nr_active);
A
Andy Whitcroft 已提交
1184

1185 1186 1187 1188 1189 1190 1191 1192 1193
		__mod_zone_page_state(zone, NR_ACTIVE_FILE,
						-count[LRU_ACTIVE_FILE]);
		__mod_zone_page_state(zone, NR_INACTIVE_FILE,
						-count[LRU_INACTIVE_FILE]);
		__mod_zone_page_state(zone, NR_ACTIVE_ANON,
						-count[LRU_ACTIVE_ANON]);
		__mod_zone_page_state(zone, NR_INACTIVE_ANON,
						-count[LRU_INACTIVE_ANON]);

K
KOSAKI Motohiro 已提交
1194 1195 1196 1197
		nr_anon = count[LRU_ACTIVE_ANON] + count[LRU_INACTIVE_ANON];
		nr_file = count[LRU_ACTIVE_FILE] + count[LRU_INACTIVE_FILE];
		__mod_zone_page_state(zone, NR_ISOLATED_ANON, nr_anon);
		__mod_zone_page_state(zone, NR_ISOLATED_FILE, nr_file);
K
KOSAKI Motohiro 已提交
1198

H
Huang Shijie 已提交
1199 1200
		reclaim_stat->recent_scanned[0] += nr_anon;
		reclaim_stat->recent_scanned[1] += nr_file;
K
KOSAKI Motohiro 已提交
1201

L
Linus Torvalds 已提交
1202 1203
		spin_unlock_irq(&zone->lru_lock);

1204
		nr_scanned += nr_scan;
1205 1206 1207 1208 1209 1210 1211 1212 1213
		nr_freed = shrink_page_list(&page_list, sc, PAGEOUT_IO_ASYNC);

		/*
		 * If we are direct reclaiming for contiguous pages and we do
		 * not reclaim everything in the list, try again and wait
		 * for IO to complete. This will stall high-order allocations
		 * but that should be acceptable to the caller
		 */
		if (nr_freed < nr_taken && !current_is_kswapd() &&
1214
		    lumpy_reclaim) {
1215
			congestion_wait(BLK_RW_ASYNC, HZ/10);
1216 1217 1218 1219 1220

			/*
			 * The attempt at page out may have made some
			 * of the pages active, mark them inactive again.
			 */
1221
			nr_active = clear_active_flags(&page_list, count);
1222 1223 1224 1225 1226 1227
			count_vm_events(PGDEACTIVATE, nr_active);

			nr_freed += shrink_page_list(&page_list, sc,
							PAGEOUT_IO_SYNC);
		}

1228
		nr_reclaimed += nr_freed;
1229

N
Nick Piggin 已提交
1230
		local_irq_disable();
1231
		if (current_is_kswapd())
1232
			__count_vm_events(KSWAPD_STEAL, nr_freed);
S
Shantanu Goel 已提交
1233
		__count_zone_vm_events(PGSTEAL, zone, nr_freed);
N
Nick Piggin 已提交
1234 1235

		spin_lock(&zone->lru_lock);
L
Linus Torvalds 已提交
1236 1237 1238 1239
		/*
		 * Put back any unfreeable pages.
		 */
		while (!list_empty(&page_list)) {
L
Lee Schermerhorn 已提交
1240
			int lru;
L
Linus Torvalds 已提交
1241
			page = lru_to_page(&page_list);
N
Nick Piggin 已提交
1242
			VM_BUG_ON(PageLRU(page));
L
Linus Torvalds 已提交
1243
			list_del(&page->lru);
L
Lee Schermerhorn 已提交
1244 1245 1246 1247 1248 1249 1250 1251 1252
			if (unlikely(!page_evictable(page, NULL))) {
				spin_unlock_irq(&zone->lru_lock);
				putback_lru_page(page);
				spin_lock_irq(&zone->lru_lock);
				continue;
			}
			SetPageLRU(page);
			lru = page_lru(page);
			add_page_to_lru_list(zone, page, lru);
1253
			if (is_active_lru(lru)) {
1254
				int file = is_file_lru(lru);
1255
				reclaim_stat->recent_rotated[file]++;
1256
			}
L
Linus Torvalds 已提交
1257 1258 1259 1260 1261 1262
			if (!pagevec_add(&pvec, page)) {
				spin_unlock_irq(&zone->lru_lock);
				__pagevec_release(&pvec);
				spin_lock_irq(&zone->lru_lock);
			}
		}
K
KOSAKI Motohiro 已提交
1263 1264 1265
		__mod_zone_page_state(zone, NR_ISOLATED_ANON, -nr_anon);
		__mod_zone_page_state(zone, NR_ISOLATED_FILE, -nr_file);

1266
  	} while (nr_scanned < max_scan);
1267

L
Linus Torvalds 已提交
1268
done:
1269
	spin_unlock_irq(&zone->lru_lock);
L
Linus Torvalds 已提交
1270
	pagevec_release(&pvec);
1271
	return nr_reclaimed;
L
Linus Torvalds 已提交
1272 1273
}

1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287
/*
 * We are about to scan this zone at a certain priority level.  If that priority
 * level is smaller (ie: more urgent) than the previous priority, then note
 * that priority level within the zone.  This is done so that when the next
 * process comes in to scan this zone, it will immediately start out at this
 * priority level rather than having to build up its own scanning priority.
 * Here, this priority affects only the reclaim-mapped threshold.
 */
static inline void note_zone_scanning_priority(struct zone *zone, int priority)
{
	if (priority < zone->prev_priority)
		zone->prev_priority = priority;
}

L
Linus Torvalds 已提交
1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304
/*
 * This moves pages from the active list to the inactive list.
 *
 * We move them the other way if the page is referenced by one or more
 * processes, from rmap.
 *
 * If the pages are mostly unmapped, the processing is fast and it is
 * appropriate to hold zone->lru_lock across the whole operation.  But if
 * the pages are mapped, the processing is slow (page_referenced()) so we
 * should drop zone->lru_lock around each page.  It's impossible to balance
 * this, so instead we remove the pages from the LRU while processing them.
 * It is safe to rely on PG_active against the non-LRU pages in here because
 * nobody will play with that bit on a non-LRU page.
 *
 * The downside is that we have to touch page->_count against each page.
 * But we had to alter page->flags anyway.
 */
1305

1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337
static void move_active_pages_to_lru(struct zone *zone,
				     struct list_head *list,
				     enum lru_list lru)
{
	unsigned long pgmoved = 0;
	struct pagevec pvec;
	struct page *page;

	pagevec_init(&pvec, 1);

	while (!list_empty(list)) {
		page = lru_to_page(list);

		VM_BUG_ON(PageLRU(page));
		SetPageLRU(page);

		list_move(&page->lru, &zone->lru[lru].list);
		mem_cgroup_add_lru_list(page, lru);
		pgmoved++;

		if (!pagevec_add(&pvec, page) || list_empty(list)) {
			spin_unlock_irq(&zone->lru_lock);
			if (buffer_heads_over_limit)
				pagevec_strip(&pvec);
			__pagevec_release(&pvec);
			spin_lock_irq(&zone->lru_lock);
		}
	}
	__mod_zone_page_state(zone, NR_LRU_BASE + lru, pgmoved);
	if (!is_active_lru(lru))
		__count_vm_events(PGDEACTIVATE, pgmoved);
}
1338

A
Andrew Morton 已提交
1339
static void shrink_active_list(unsigned long nr_pages, struct zone *zone,
1340
			struct scan_control *sc, int priority, int file)
L
Linus Torvalds 已提交
1341
{
1342
	unsigned long nr_taken;
1343
	unsigned long pgscanned;
1344
	unsigned long vm_flags;
L
Linus Torvalds 已提交
1345
	LIST_HEAD(l_hold);	/* The pages which were snipped off */
1346
	LIST_HEAD(l_active);
1347
	LIST_HEAD(l_inactive);
L
Linus Torvalds 已提交
1348
	struct page *page;
1349
	struct zone_reclaim_stat *reclaim_stat = get_reclaim_stat(zone, sc);
1350
	unsigned long nr_rotated = 0;
L
Linus Torvalds 已提交
1351 1352 1353

	lru_add_drain();
	spin_lock_irq(&zone->lru_lock);
1354
	nr_taken = sc->isolate_pages(nr_pages, &l_hold, &pgscanned, sc->order,
1355
					ISOLATE_ACTIVE, zone,
1356
					sc->mem_cgroup, 1, file);
1357 1358 1359 1360
	/*
	 * zone->pages_scanned is used for detect zone's oom
	 * mem_cgroup remembers nr_scan by itself.
	 */
1361
	if (scanning_global_lru(sc)) {
1362
		zone->pages_scanned += pgscanned;
1363
	}
1364
	reclaim_stat->recent_scanned[file] += nr_taken;
1365

1366
	__count_zone_vm_events(PGREFILL, zone, pgscanned);
1367
	if (file)
1368
		__mod_zone_page_state(zone, NR_ACTIVE_FILE, -nr_taken);
1369
	else
1370
		__mod_zone_page_state(zone, NR_ACTIVE_ANON, -nr_taken);
K
KOSAKI Motohiro 已提交
1371
	__mod_zone_page_state(zone, NR_ISOLATED_ANON + file, nr_taken);
L
Linus Torvalds 已提交
1372 1373 1374 1375 1376 1377
	spin_unlock_irq(&zone->lru_lock);

	while (!list_empty(&l_hold)) {
		cond_resched();
		page = lru_to_page(&l_hold);
		list_del(&page->lru);
1378

L
Lee Schermerhorn 已提交
1379 1380 1381 1382 1383
		if (unlikely(!page_evictable(page, NULL))) {
			putback_lru_page(page);
			continue;
		}

1384
		if (page_referenced(page, 0, sc->mem_cgroup, &vm_flags)) {
1385
			nr_rotated++;
1386 1387 1388 1389 1390 1391 1392 1393 1394
			/*
			 * Identify referenced, file-backed active pages and
			 * give them one more trip around the active list. So
			 * that executable code get better chances to stay in
			 * memory under moderate memory pressure.  Anon pages
			 * are not likely to be evicted by use-once streaming
			 * IO, plus JVM can create lots of anon VM_EXEC pages,
			 * so we ignore them here.
			 */
1395
			if ((vm_flags & VM_EXEC) && page_is_file_cache(page)) {
1396 1397 1398 1399
				list_add(&page->lru, &l_active);
				continue;
			}
		}
1400

1401
		ClearPageActive(page);	/* we are de-activating */
L
Linus Torvalds 已提交
1402 1403 1404
		list_add(&page->lru, &l_inactive);
	}

1405
	/*
1406
	 * Move pages back to the lru list.
1407
	 */
1408
	spin_lock_irq(&zone->lru_lock);
1409
	/*
1410 1411 1412 1413
	 * Count referenced pages from currently used mappings as rotated,
	 * even though only some of them are actually re-activated.  This
	 * helps balance scan pressure between file and anonymous pages in
	 * get_scan_ratio.
1414
	 */
1415
	reclaim_stat->recent_rotated[file] += nr_rotated;
1416

1417 1418 1419 1420
	move_active_pages_to_lru(zone, &l_active,
						LRU_ACTIVE + file * LRU_FILE);
	move_active_pages_to_lru(zone, &l_inactive,
						LRU_BASE   + file * LRU_FILE);
K
KOSAKI Motohiro 已提交
1421
	__mod_zone_page_state(zone, NR_ISOLATED_ANON + file, -nr_taken);
1422
	spin_unlock_irq(&zone->lru_lock);
L
Linus Torvalds 已提交
1423 1424
}

1425
static int inactive_anon_is_low_global(struct zone *zone)
1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437
{
	unsigned long active, inactive;

	active = zone_page_state(zone, NR_ACTIVE_ANON);
	inactive = zone_page_state(zone, NR_INACTIVE_ANON);

	if (inactive * zone->inactive_ratio < active)
		return 1;

	return 0;
}

1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449
/**
 * inactive_anon_is_low - check if anonymous pages need to be deactivated
 * @zone: zone to check
 * @sc:   scan control of this context
 *
 * Returns true if the zone does not have enough inactive anon pages,
 * meaning some active anon pages need to be deactivated.
 */
static int inactive_anon_is_low(struct zone *zone, struct scan_control *sc)
{
	int low;

1450
	if (scanning_global_lru(sc))
1451 1452
		low = inactive_anon_is_low_global(zone);
	else
1453
		low = mem_cgroup_inactive_anon_is_low(sc->mem_cgroup);
1454 1455 1456
	return low;
}

1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492
static int inactive_file_is_low_global(struct zone *zone)
{
	unsigned long active, inactive;

	active = zone_page_state(zone, NR_ACTIVE_FILE);
	inactive = zone_page_state(zone, NR_INACTIVE_FILE);

	return (active > inactive);
}

/**
 * inactive_file_is_low - check if file pages need to be deactivated
 * @zone: zone to check
 * @sc:   scan control of this context
 *
 * When the system is doing streaming IO, memory pressure here
 * ensures that active file pages get deactivated, until more
 * than half of the file pages are on the inactive list.
 *
 * Once we get to that situation, protect the system's working
 * set from being evicted by disabling active file page aging.
 *
 * This uses a different ratio than the anonymous pages, because
 * the page cache uses a use-once replacement algorithm.
 */
static int inactive_file_is_low(struct zone *zone, struct scan_control *sc)
{
	int low;

	if (scanning_global_lru(sc))
		low = inactive_file_is_low_global(zone);
	else
		low = mem_cgroup_inactive_file_is_low(sc->mem_cgroup);
	return low;
}

1493 1494 1495 1496 1497 1498 1499 1500 1501
static int inactive_list_is_low(struct zone *zone, struct scan_control *sc,
				int file)
{
	if (file)
		return inactive_file_is_low(zone, sc);
	else
		return inactive_anon_is_low(zone, sc);
}

1502
static unsigned long shrink_list(enum lru_list lru, unsigned long nr_to_scan,
1503 1504
	struct zone *zone, struct scan_control *sc, int priority)
{
1505 1506
	int file = is_file_lru(lru);

1507 1508 1509
	if (is_active_lru(lru)) {
		if (inactive_list_is_low(zone, sc, file))
		    shrink_active_list(nr_to_scan, zone, sc, priority, file);
1510 1511 1512
		return 0;
	}

R
Rik van Riel 已提交
1513
	return shrink_inactive_list(nr_to_scan, zone, sc, priority, file);
1514 1515
}

1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535
/*
 * Smallish @nr_to_scan's are deposited in @nr_saved_scan,
 * until we collected @swap_cluster_max pages to scan.
 */
static unsigned long nr_scan_try_batch(unsigned long nr_to_scan,
				       unsigned long *nr_saved_scan)
{
	unsigned long nr;

	*nr_saved_scan += nr_to_scan;
	nr = *nr_saved_scan;

	if (nr >= SWAP_CLUSTER_MAX)
		*nr_saved_scan = 0;
	else
		nr = 0;

	return nr;
}

1536 1537 1538 1539 1540 1541
/*
 * Determine how aggressively the anon and file LRU lists should be
 * scanned.  The relative value of each set of LRU lists is determined
 * by looking at the fraction of the pages scanned we did rotate back
 * onto the active list instead of evict.
 *
1542
 * nr[0] = anon pages to scan; nr[1] = file pages to scan
1543
 */
1544 1545
static void get_scan_count(struct zone *zone, struct scan_control *sc,
					unsigned long *nr, int priority)
1546 1547 1548 1549
{
	unsigned long anon, file, free;
	unsigned long anon_prio, file_prio;
	unsigned long ap, fp;
1550
	struct zone_reclaim_stat *reclaim_stat = get_reclaim_stat(zone, sc);
1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562
	u64 fraction[2], denominator;
	enum lru_list l;
	int noswap = 0;

	/* If we have no swap space, do not bother scanning anon pages. */
	if (!sc->may_swap || (nr_swap_pages <= 0)) {
		noswap = 1;
		fraction[0] = 0;
		fraction[1] = 1;
		denominator = 1;
		goto out;
	}
1563

1564 1565 1566 1567
	anon  = zone_nr_lru_pages(zone, sc, LRU_ACTIVE_ANON) +
		zone_nr_lru_pages(zone, sc, LRU_INACTIVE_ANON);
	file  = zone_nr_lru_pages(zone, sc, LRU_ACTIVE_FILE) +
		zone_nr_lru_pages(zone, sc, LRU_INACTIVE_FILE);
1568

1569
	if (scanning_global_lru(sc)) {
1570 1571 1572
		free  = zone_page_state(zone, NR_FREE_PAGES);
		/* If we have very few page cache pages,
		   force-scan anon pages. */
1573
		if (unlikely(file + free <= high_wmark_pages(zone))) {
1574 1575 1576 1577
			fraction[0] = 1;
			fraction[1] = 0;
			denominator = 1;
			goto out;
1578
		}
1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591
	}

	/*
	 * OK, so we have swap space and a fair amount of page cache
	 * pages.  We use the recently rotated / recently scanned
	 * ratios to determine how valuable each cache is.
	 *
	 * Because workloads change over time (and to avoid overflow)
	 * we keep these statistics as a floating average, which ends
	 * up weighing recent references more than old ones.
	 *
	 * anon in [0], file in [1]
	 */
1592
	if (unlikely(reclaim_stat->recent_scanned[0] > anon / 4)) {
1593
		spin_lock_irq(&zone->lru_lock);
1594 1595
		reclaim_stat->recent_scanned[0] /= 2;
		reclaim_stat->recent_rotated[0] /= 2;
1596 1597 1598
		spin_unlock_irq(&zone->lru_lock);
	}

1599
	if (unlikely(reclaim_stat->recent_scanned[1] > file / 4)) {
1600
		spin_lock_irq(&zone->lru_lock);
1601 1602
		reclaim_stat->recent_scanned[1] /= 2;
		reclaim_stat->recent_rotated[1] /= 2;
1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613
		spin_unlock_irq(&zone->lru_lock);
	}

	/*
	 * With swappiness at 100, anonymous and file have the same priority.
	 * This scanning priority is essentially the inverse of IO cost.
	 */
	anon_prio = sc->swappiness;
	file_prio = 200 - sc->swappiness;

	/*
1614 1615 1616
	 * The amount of pressure on anon vs file pages is inversely
	 * proportional to the fraction of recently scanned pages on
	 * each list that were recently referenced and in active use.
1617
	 */
1618 1619
	ap = (anon_prio + 1) * (reclaim_stat->recent_scanned[0] + 1);
	ap /= reclaim_stat->recent_rotated[0] + 1;
1620

1621 1622
	fp = (file_prio + 1) * (reclaim_stat->recent_scanned[1] + 1);
	fp /= reclaim_stat->recent_rotated[1] + 1;
1623

1624 1625 1626 1627 1628 1629 1630
	fraction[0] = ap;
	fraction[1] = fp;
	denominator = ap + fp + 1;
out:
	for_each_evictable_lru(l) {
		int file = is_file_lru(l);
		unsigned long scan;
1631

1632 1633 1634 1635 1636 1637 1638 1639
		scan = zone_nr_lru_pages(zone, sc, l);
		if (priority || noswap) {
			scan >>= priority;
			scan = div64_u64(scan * fraction[file], denominator);
		}
		nr[l] = nr_scan_try_batch(scan,
					  &reclaim_stat->nr_saved_scan[l]);
	}
1640
}
1641

L
Linus Torvalds 已提交
1642 1643 1644
/*
 * This is a basic per-zone page freer.  Used by both kswapd and direct reclaim.
 */
1645
static void shrink_zone(int priority, struct zone *zone,
1646
				struct scan_control *sc)
L
Linus Torvalds 已提交
1647
{
1648
	unsigned long nr[NR_LRU_LISTS];
1649
	unsigned long nr_to_scan;
1650
	enum lru_list l;
1651
	unsigned long nr_reclaimed = sc->nr_reclaimed;
1652
	unsigned long nr_to_reclaim = sc->nr_to_reclaim;
1653

1654
	get_scan_count(zone, sc, nr, priority);
L
Linus Torvalds 已提交
1655

1656 1657
	while (nr[LRU_INACTIVE_ANON] || nr[LRU_ACTIVE_FILE] ||
					nr[LRU_INACTIVE_FILE]) {
L
Lee Schermerhorn 已提交
1658
		for_each_evictable_lru(l) {
1659
			if (nr[l]) {
K
KOSAKI Motohiro 已提交
1660 1661
				nr_to_scan = min_t(unsigned long,
						   nr[l], SWAP_CLUSTER_MAX);
1662
				nr[l] -= nr_to_scan;
L
Linus Torvalds 已提交
1663

1664 1665
				nr_reclaimed += shrink_list(l, nr_to_scan,
							    zone, sc, priority);
1666
			}
L
Linus Torvalds 已提交
1667
		}
1668 1669 1670 1671 1672 1673 1674 1675
		/*
		 * On large memory systems, scan >> priority can become
		 * really large. This is fine for the starting priority;
		 * we want to put equal scanning pressure on each zone.
		 * However, if the VM has a harder time of freeing pages,
		 * with multiple processes reclaiming pages, the total
		 * freeing target can get unreasonably large.
		 */
1676
		if (nr_reclaimed >= nr_to_reclaim && priority < DEF_PRIORITY)
1677
			break;
L
Linus Torvalds 已提交
1678 1679
	}

1680 1681
	sc->nr_reclaimed = nr_reclaimed;

1682 1683 1684 1685
	/*
	 * Even if we did not try to evict anon pages at all, we want to
	 * rebalance the anon lru active/inactive ratio.
	 */
1686
	if (inactive_anon_is_low(zone, sc) && nr_swap_pages > 0)
1687 1688
		shrink_active_list(SWAP_CLUSTER_MAX, zone, sc, priority, 0);

1689
	throttle_vm_writeout(sc->gfp_mask);
L
Linus Torvalds 已提交
1690 1691 1692 1693 1694 1695 1696
}

/*
 * This is the direct reclaim path, for page-allocating processes.  We only
 * try to reclaim pages from zones which will satisfy the caller's allocation
 * request.
 *
1697 1698
 * We reclaim from a zone even if that zone is over high_wmark_pages(zone).
 * Because:
L
Linus Torvalds 已提交
1699 1700
 * a) The caller may be trying to free *extra* pages to satisfy a higher-order
 *    allocation or
1701 1702 1703
 * b) The target zone may be at high_wmark_pages(zone) but the lower zones
 *    must go *over* high_wmark_pages(zone) to satisfy the `incremental min'
 *    zone defense algorithm.
L
Linus Torvalds 已提交
1704 1705 1706 1707
 *
 * If a zone is deemed to be full of pinned pages then just give it a light
 * scan then give up on it.
 */
1708
static void shrink_zones(int priority, struct zonelist *zonelist,
1709
					struct scan_control *sc)
L
Linus Torvalds 已提交
1710
{
1711
	enum zone_type high_zoneidx = gfp_zone(sc->gfp_mask);
1712
	struct zoneref *z;
1713
	struct zone *zone;
1714

1715
	sc->all_unreclaimable = 1;
1716 1717
	for_each_zone_zonelist_nodemask(zone, z, zonelist, high_zoneidx,
					sc->nodemask) {
1718
		if (!populated_zone(zone))
L
Linus Torvalds 已提交
1719
			continue;
1720 1721 1722 1723
		/*
		 * Take care memory controller reclaiming has small influence
		 * to global LRU.
		 */
1724
		if (scanning_global_lru(sc)) {
1725 1726 1727
			if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL))
				continue;
			note_zone_scanning_priority(zone, priority);
L
Linus Torvalds 已提交
1728

1729
			if (zone->all_unreclaimable && priority != DEF_PRIORITY)
1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740
				continue;	/* Let kswapd poll it */
			sc->all_unreclaimable = 0;
		} else {
			/*
			 * Ignore cpuset limitation here. We just want to reduce
			 * # of used pages by us regardless of memory shortage.
			 */
			sc->all_unreclaimable = 0;
			mem_cgroup_note_reclaim_priority(sc->mem_cgroup,
							priority);
		}
1741

1742
		shrink_zone(priority, zone, sc);
L
Linus Torvalds 已提交
1743 1744
	}
}
1745

L
Linus Torvalds 已提交
1746 1747 1748 1749 1750 1751 1752 1753
/*
 * This is the main entry point to direct page reclaim.
 *
 * If a full scan of the inactive list fails to free enough memory then we
 * are "out of memory" and something needs to be killed.
 *
 * If the caller is !__GFP_FS then the probability of a failure is reasonably
 * high - the zone may be full of dirty or under-writeback pages, which this
1754 1755 1756 1757
 * caller can't do much about.  We kick the writeback threads and take explicit
 * naps in the hope that some of these pages can be written.  But if the
 * allocating task holds filesystem locks which prevent writeout this might not
 * work, and the allocation attempt will fail.
1758 1759 1760
 *
 * returns:	0, if no pages reclaimed
 * 		else, the number of pages reclaimed
L
Linus Torvalds 已提交
1761
 */
1762
static unsigned long do_try_to_free_pages(struct zonelist *zonelist,
1763
					struct scan_control *sc)
L
Linus Torvalds 已提交
1764 1765
{
	int priority;
1766
	unsigned long ret = 0;
1767
	unsigned long total_scanned = 0;
L
Linus Torvalds 已提交
1768 1769
	struct reclaim_state *reclaim_state = current->reclaim_state;
	unsigned long lru_pages = 0;
1770
	struct zoneref *z;
1771
	struct zone *zone;
1772
	enum zone_type high_zoneidx = gfp_zone(sc->gfp_mask);
1773
	unsigned long writeback_threshold;
L
Linus Torvalds 已提交
1774

1775
	get_mems_allowed();
1776 1777
	delayacct_freepages_start();

1778
	if (scanning_global_lru(sc))
1779 1780 1781 1782
		count_vm_event(ALLOCSTALL);
	/*
	 * mem_cgroup will not do shrink_slab.
	 */
1783
	if (scanning_global_lru(sc)) {
1784
		for_each_zone_zonelist(zone, z, zonelist, high_zoneidx) {
L
Linus Torvalds 已提交
1785

1786 1787
			if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL))
				continue;
L
Linus Torvalds 已提交
1788

1789
			lru_pages += zone_reclaimable_pages(zone);
1790
		}
L
Linus Torvalds 已提交
1791 1792 1793
	}

	for (priority = DEF_PRIORITY; priority >= 0; priority--) {
1794
		sc->nr_scanned = 0;
1795 1796
		if (!priority)
			disable_swap_token();
1797
		shrink_zones(priority, zonelist, sc);
1798 1799 1800 1801
		/*
		 * Don't shrink slabs when reclaiming memory from
		 * over limit cgroups
		 */
1802
		if (scanning_global_lru(sc)) {
1803
			shrink_slab(sc->nr_scanned, sc->gfp_mask, lru_pages);
1804
			if (reclaim_state) {
1805
				sc->nr_reclaimed += reclaim_state->reclaimed_slab;
1806 1807
				reclaim_state->reclaimed_slab = 0;
			}
L
Linus Torvalds 已提交
1808
		}
1809
		total_scanned += sc->nr_scanned;
1810
		if (sc->nr_reclaimed >= sc->nr_to_reclaim) {
1811
			ret = sc->nr_reclaimed;
L
Linus Torvalds 已提交
1812 1813 1814 1815 1816 1817 1818 1819 1820 1821
			goto out;
		}

		/*
		 * Try to write back as many pages as we just scanned.  This
		 * tends to cause slow streaming writers to write data to the
		 * disk smoothly, at the dirtying rate, which is nice.   But
		 * that's undesirable in laptop mode, where we *want* lumpy
		 * writeout.  So in laptop mode, write out the whole world.
		 */
1822 1823
		writeback_threshold = sc->nr_to_reclaim + sc->nr_to_reclaim / 2;
		if (total_scanned > writeback_threshold) {
1824
			wakeup_flusher_threads(laptop_mode ? 0 : total_scanned);
1825
			sc->may_writepage = 1;
L
Linus Torvalds 已提交
1826 1827 1828
		}

		/* Take a nap, wait for some writeback to complete */
1829 1830
		if (!sc->hibernation_mode && sc->nr_scanned &&
		    priority < DEF_PRIORITY - 2)
1831
			congestion_wait(BLK_RW_ASYNC, HZ/10);
L
Linus Torvalds 已提交
1832
	}
1833
	/* top priority shrink_zones still had more to do? don't OOM, then */
1834
	if (!sc->all_unreclaimable && scanning_global_lru(sc))
1835
		ret = sc->nr_reclaimed;
L
Linus Torvalds 已提交
1836
out:
1837 1838 1839 1840 1841 1842 1843 1844 1845
	/*
	 * Now that we've scanned all the zones at this priority level, note
	 * that level within the zone so that the next thread which performs
	 * scanning of this zone will immediately start out at this priority
	 * level.  This affects only the decision whether or not to bring
	 * mapped pages onto the inactive list.
	 */
	if (priority < 0)
		priority = 0;
L
Linus Torvalds 已提交
1846

1847
	if (scanning_global_lru(sc)) {
1848
		for_each_zone_zonelist(zone, z, zonelist, high_zoneidx) {
1849 1850 1851 1852 1853 1854 1855 1856

			if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL))
				continue;

			zone->prev_priority = priority;
		}
	} else
		mem_cgroup_record_reclaim_priority(sc->mem_cgroup, priority);
L
Linus Torvalds 已提交
1857

1858
	delayacct_freepages_end();
1859
	put_mems_allowed();
1860

L
Linus Torvalds 已提交
1861 1862 1863
	return ret;
}

1864
unsigned long try_to_free_pages(struct zonelist *zonelist, int order,
1865
				gfp_t gfp_mask, nodemask_t *nodemask)
1866 1867 1868 1869
{
	struct scan_control sc = {
		.gfp_mask = gfp_mask,
		.may_writepage = !laptop_mode,
1870
		.nr_to_reclaim = SWAP_CLUSTER_MAX,
1871
		.may_unmap = 1,
1872
		.may_swap = 1,
1873 1874 1875 1876
		.swappiness = vm_swappiness,
		.order = order,
		.mem_cgroup = NULL,
		.isolate_pages = isolate_pages_global,
1877
		.nodemask = nodemask,
1878 1879
	};

1880
	return do_try_to_free_pages(zonelist, &sc);
1881 1882
}

1883
#ifdef CONFIG_CGROUP_MEM_RES_CTLR
1884

1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916
unsigned long mem_cgroup_shrink_node_zone(struct mem_cgroup *mem,
						gfp_t gfp_mask, bool noswap,
						unsigned int swappiness,
						struct zone *zone, int nid)
{
	struct scan_control sc = {
		.may_writepage = !laptop_mode,
		.may_unmap = 1,
		.may_swap = !noswap,
		.swappiness = swappiness,
		.order = 0,
		.mem_cgroup = mem,
		.isolate_pages = mem_cgroup_isolate_pages,
	};
	nodemask_t nm  = nodemask_of_node(nid);

	sc.gfp_mask = (gfp_mask & GFP_RECLAIM_MASK) |
			(GFP_HIGHUSER_MOVABLE & ~GFP_RECLAIM_MASK);
	sc.nodemask = &nm;
	sc.nr_reclaimed = 0;
	sc.nr_scanned = 0;
	/*
	 * NOTE: Although we can get the priority field, using it
	 * here is not a good idea, since it limits the pages we can scan.
	 * if we don't reclaim here, the shrink_zone from balance_pgdat
	 * will pick up pages from other mem cgroup's as well. We hack
	 * the priority and make it zero.
	 */
	shrink_zone(0, zone, &sc);
	return sc.nr_reclaimed;
}

1917
unsigned long try_to_free_mem_cgroup_pages(struct mem_cgroup *mem_cont,
K
KOSAKI Motohiro 已提交
1918 1919 1920
					   gfp_t gfp_mask,
					   bool noswap,
					   unsigned int swappiness)
1921
{
1922
	struct zonelist *zonelist;
1923 1924
	struct scan_control sc = {
		.may_writepage = !laptop_mode,
1925
		.may_unmap = 1,
1926
		.may_swap = !noswap,
1927
		.nr_to_reclaim = SWAP_CLUSTER_MAX,
K
KOSAKI Motohiro 已提交
1928
		.swappiness = swappiness,
1929 1930 1931
		.order = 0,
		.mem_cgroup = mem_cont,
		.isolate_pages = mem_cgroup_isolate_pages,
1932
		.nodemask = NULL, /* we don't care the placement */
1933 1934
	};

1935 1936 1937 1938
	sc.gfp_mask = (gfp_mask & GFP_RECLAIM_MASK) |
			(GFP_HIGHUSER_MOVABLE & ~GFP_RECLAIM_MASK);
	zonelist = NODE_DATA(numa_node_id())->node_zonelists;
	return do_try_to_free_pages(zonelist, &sc);
1939 1940 1941
}
#endif

1942
/* is kswapd sleeping prematurely? */
1943
static int sleeping_prematurely(pg_data_t *pgdat, int order, long remaining)
1944
{
1945
	int i;
1946 1947 1948 1949 1950 1951

	/* If a direct reclaimer woke kswapd within HZ/10, it's premature */
	if (remaining)
		return 1;

	/* If after HZ/10, a zone is below the high mark, it's premature */
1952 1953 1954 1955 1956 1957
	for (i = 0; i < pgdat->nr_zones; i++) {
		struct zone *zone = pgdat->node_zones + i;

		if (!populated_zone(zone))
			continue;

1958
		if (zone->all_unreclaimable)
1959 1960
			continue;

1961 1962 1963
		if (!zone_watermark_ok(zone, order, high_wmark_pages(zone),
								0, 0))
			return 1;
1964
	}
1965 1966 1967 1968

	return 0;
}

L
Linus Torvalds 已提交
1969 1970
/*
 * For kswapd, balance_pgdat() will work across all this node's zones until
1971
 * they are all at high_wmark_pages(zone).
L
Linus Torvalds 已提交
1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983
 *
 * Returns the number of pages which were actually freed.
 *
 * There is special handling here for zones which are full of pinned pages.
 * This can happen if the pages are all mlocked, or if they are all used by
 * device drivers (say, ZONE_DMA).  Or if they are all in use by hugetlb.
 * What we do is to detect the case where all pages in the zone have been
 * scanned twice and there has been zero successful reclaim.  Mark the zone as
 * dead and from now on, only perform a short scan.  Basically we're polling
 * the zone for when the problem goes away.
 *
 * kswapd scans the zones in the highmem->normal->dma direction.  It skips
1984 1985 1986 1987 1988
 * zones which have free_pages > high_wmark_pages(zone), but once a zone is
 * found to have free_pages <= high_wmark_pages(zone), we scan that zone and the
 * lower zones regardless of the number of free pages in the lower zones. This
 * interoperates with the page allocator fallback scheme to ensure that aging
 * of pages is balanced across the zones.
L
Linus Torvalds 已提交
1989
 */
1990
static unsigned long balance_pgdat(pg_data_t *pgdat, int order)
L
Linus Torvalds 已提交
1991 1992 1993 1994
{
	int all_zones_ok;
	int priority;
	int i;
1995
	unsigned long total_scanned;
L
Linus Torvalds 已提交
1996
	struct reclaim_state *reclaim_state = current->reclaim_state;
1997 1998
	struct scan_control sc = {
		.gfp_mask = GFP_KERNEL,
1999
		.may_unmap = 1,
2000
		.may_swap = 1,
2001 2002 2003 2004 2005
		/*
		 * kswapd doesn't want to be bailed out while reclaim. because
		 * we want to put equal scanning pressure on each zone.
		 */
		.nr_to_reclaim = ULONG_MAX,
2006
		.swappiness = vm_swappiness,
A
Andy Whitcroft 已提交
2007
		.order = order,
2008 2009
		.mem_cgroup = NULL,
		.isolate_pages = isolate_pages_global,
2010
	};
2011 2012
	/*
	 * temp_priority is used to remember the scanning priority at which
2013 2014
	 * this zone was successfully refilled to
	 * free_pages == high_wmark_pages(zone).
2015 2016
	 */
	int temp_priority[MAX_NR_ZONES];
L
Linus Torvalds 已提交
2017 2018 2019

loop_again:
	total_scanned = 0;
2020
	sc.nr_reclaimed = 0;
C
Christoph Lameter 已提交
2021
	sc.may_writepage = !laptop_mode;
2022
	count_vm_event(PAGEOUTRUN);
L
Linus Torvalds 已提交
2023

2024 2025
	for (i = 0; i < pgdat->nr_zones; i++)
		temp_priority[i] = DEF_PRIORITY;
L
Linus Torvalds 已提交
2026 2027 2028 2029

	for (priority = DEF_PRIORITY; priority >= 0; priority--) {
		int end_zone = 0;	/* Inclusive.  0 = ZONE_DMA */
		unsigned long lru_pages = 0;
2030
		int has_under_min_watermark_zone = 0;
L
Linus Torvalds 已提交
2031

2032 2033 2034 2035
		/* The swap token gets in the way of swapout... */
		if (!priority)
			disable_swap_token();

L
Linus Torvalds 已提交
2036 2037
		all_zones_ok = 1;

2038 2039 2040 2041 2042 2043
		/*
		 * Scan in the highmem->dma direction for the highest
		 * zone which needs scanning
		 */
		for (i = pgdat->nr_zones - 1; i >= 0; i--) {
			struct zone *zone = pgdat->node_zones + i;
L
Linus Torvalds 已提交
2044

2045 2046
			if (!populated_zone(zone))
				continue;
L
Linus Torvalds 已提交
2047

2048
			if (zone->all_unreclaimable && priority != DEF_PRIORITY)
2049
				continue;
L
Linus Torvalds 已提交
2050

2051 2052 2053 2054
			/*
			 * Do some background aging of the anon list, to give
			 * pages a chance to be referenced before reclaiming.
			 */
2055
			if (inactive_anon_is_low(zone, &sc))
2056 2057 2058
				shrink_active_list(SWAP_CLUSTER_MAX, zone,
							&sc, priority, 0);

2059 2060
			if (!zone_watermark_ok(zone, order,
					high_wmark_pages(zone), 0, 0)) {
2061
				end_zone = i;
A
Andrew Morton 已提交
2062
				break;
L
Linus Torvalds 已提交
2063 2064
			}
		}
A
Andrew Morton 已提交
2065 2066 2067
		if (i < 0)
			goto out;

L
Linus Torvalds 已提交
2068 2069 2070
		for (i = 0; i <= end_zone; i++) {
			struct zone *zone = pgdat->node_zones + i;

2071
			lru_pages += zone_reclaimable_pages(zone);
L
Linus Torvalds 已提交
2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084
		}

		/*
		 * Now scan the zone in the dma->highmem direction, stopping
		 * at the last zone which needs scanning.
		 *
		 * We do this because the page allocator works in the opposite
		 * direction.  This prevents the page allocator from allocating
		 * pages behind kswapd's direction of progress, which would
		 * cause too much scanning of the lower zones.
		 */
		for (i = 0; i <= end_zone; i++) {
			struct zone *zone = pgdat->node_zones + i;
2085
			int nr_slab;
2086
			int nid, zid;
L
Linus Torvalds 已提交
2087

2088
			if (!populated_zone(zone))
L
Linus Torvalds 已提交
2089 2090
				continue;

2091
			if (zone->all_unreclaimable && priority != DEF_PRIORITY)
L
Linus Torvalds 已提交
2092 2093
				continue;

2094
			temp_priority[i] = priority;
L
Linus Torvalds 已提交
2095
			sc.nr_scanned = 0;
2096
			note_zone_scanning_priority(zone, priority);
2097 2098 2099 2100 2101 2102 2103 2104 2105

			nid = pgdat->node_id;
			zid = zone_idx(zone);
			/*
			 * Call soft limit reclaim before calling shrink_zone.
			 * For now we ignore the return value
			 */
			mem_cgroup_soft_limit_reclaim(zone, order, sc.gfp_mask,
							nid, zid);
2106 2107 2108 2109
			/*
			 * We put equal pressure on every zone, unless one
			 * zone has way too many pages free already.
			 */
2110 2111
			if (!zone_watermark_ok(zone, order,
					8*high_wmark_pages(zone), end_zone, 0))
2112
				shrink_zone(priority, zone, &sc);
L
Linus Torvalds 已提交
2113
			reclaim_state->reclaimed_slab = 0;
2114 2115
			nr_slab = shrink_slab(sc.nr_scanned, GFP_KERNEL,
						lru_pages);
2116
			sc.nr_reclaimed += reclaim_state->reclaimed_slab;
L
Linus Torvalds 已提交
2117
			total_scanned += sc.nr_scanned;
2118
			if (zone->all_unreclaimable)
L
Linus Torvalds 已提交
2119
				continue;
2120 2121 2122
			if (nr_slab == 0 &&
			    zone->pages_scanned >= (zone_reclaimable_pages(zone) * 6))
				zone->all_unreclaimable = 1;
L
Linus Torvalds 已提交
2123 2124 2125 2126 2127 2128
			/*
			 * If we've done a decent amount of scanning and
			 * the reclaim ratio is low, start doing writepage
			 * even in laptop mode
			 */
			if (total_scanned > SWAP_CLUSTER_MAX * 2 &&
2129
			    total_scanned > sc.nr_reclaimed + sc.nr_reclaimed / 2)
L
Linus Torvalds 已提交
2130
				sc.may_writepage = 1;
2131

2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143
			if (!zone_watermark_ok(zone, order,
					high_wmark_pages(zone), end_zone, 0)) {
				all_zones_ok = 0;
				/*
				 * We are still under min water mark.  This
				 * means that we have a GFP_ATOMIC allocation
				 * failure risk. Hurry up!
				 */
				if (!zone_watermark_ok(zone, order,
					    min_wmark_pages(zone), end_zone, 0))
					has_under_min_watermark_zone = 1;
			}
2144

L
Linus Torvalds 已提交
2145 2146 2147 2148 2149 2150 2151
		}
		if (all_zones_ok)
			break;		/* kswapd: all done */
		/*
		 * OK, kswapd is getting into trouble.  Take a nap, then take
		 * another pass across the zones.
		 */
2152 2153 2154 2155 2156 2157
		if (total_scanned && (priority < DEF_PRIORITY - 2)) {
			if (has_under_min_watermark_zone)
				count_vm_event(KSWAPD_SKIP_CONGESTION_WAIT);
			else
				congestion_wait(BLK_RW_ASYNC, HZ/10);
		}
L
Linus Torvalds 已提交
2158 2159 2160 2161 2162 2163 2164

		/*
		 * We do this so kswapd doesn't build up large priorities for
		 * example when it is freeing in parallel with allocators. It
		 * matches the direct reclaim path behaviour in terms of impact
		 * on zone->*_priority.
		 */
2165
		if (sc.nr_reclaimed >= SWAP_CLUSTER_MAX)
L
Linus Torvalds 已提交
2166 2167 2168
			break;
	}
out:
2169 2170 2171 2172 2173
	/*
	 * Note within each zone the priority level at which this zone was
	 * brought into a happy state.  So that the next thread which scans this
	 * zone will start out at that priority level.
	 */
L
Linus Torvalds 已提交
2174 2175 2176
	for (i = 0; i < pgdat->nr_zones; i++) {
		struct zone *zone = pgdat->node_zones + i;

2177
		zone->prev_priority = temp_priority[i];
L
Linus Torvalds 已提交
2178 2179 2180
	}
	if (!all_zones_ok) {
		cond_resched();
2181 2182 2183

		try_to_freeze();

2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200
		/*
		 * Fragmentation may mean that the system cannot be
		 * rebalanced for high-order allocations in all zones.
		 * At this point, if nr_reclaimed < SWAP_CLUSTER_MAX,
		 * it means the zones have been fully scanned and are still
		 * not balanced. For high-order allocations, there is
		 * little point trying all over again as kswapd may
		 * infinite loop.
		 *
		 * Instead, recheck all watermarks at order-0 as they
		 * are the most important. If watermarks are ok, kswapd will go
		 * back to sleep. High-order users can still perform direct
		 * reclaim if they wish.
		 */
		if (sc.nr_reclaimed < SWAP_CLUSTER_MAX)
			order = sc.order = 0;

L
Linus Torvalds 已提交
2201 2202 2203
		goto loop_again;
	}

2204
	return sc.nr_reclaimed;
L
Linus Torvalds 已提交
2205 2206 2207 2208
}

/*
 * The background pageout daemon, started as a kernel thread
2209
 * from the init process.
L
Linus Torvalds 已提交
2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228
 *
 * This basically trickles out pages so that we have _some_
 * free memory available even if there is no other activity
 * that frees anything up. This is needed for things like routing
 * etc, where we otherwise might have all activity going on in
 * asynchronous contexts that cannot page things out.
 *
 * If there are applications that are active memory-allocators
 * (most normal use), this basically shouldn't matter.
 */
static int kswapd(void *p)
{
	unsigned long order;
	pg_data_t *pgdat = (pg_data_t*)p;
	struct task_struct *tsk = current;
	DEFINE_WAIT(wait);
	struct reclaim_state reclaim_state = {
		.reclaimed_slab = 0,
	};
2229
	const struct cpumask *cpumask = cpumask_of_node(pgdat->node_id);
L
Linus Torvalds 已提交
2230

2231 2232
	lockdep_set_current_reclaim_state(GFP_KERNEL);

R
Rusty Russell 已提交
2233
	if (!cpumask_empty(cpumask))
2234
		set_cpus_allowed_ptr(tsk, cpumask);
L
Linus Torvalds 已提交
2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248
	current->reclaim_state = &reclaim_state;

	/*
	 * Tell the memory management that we're a "memory allocator",
	 * and that if we need more memory we should get access to it
	 * regardless (see "__alloc_pages()"). "kswapd" should
	 * never get caught in the normal page freeing logic.
	 *
	 * (Kswapd normally doesn't need memory anyway, but sometimes
	 * you need a small amount of memory in order to be able to
	 * page out something else, and this flag essentially protects
	 * us from recursively trying to free more memory as we're
	 * trying to free the first piece of memory in the first place).
	 */
2249
	tsk->flags |= PF_MEMALLOC | PF_SWAPWRITE | PF_KSWAPD;
2250
	set_freezable();
L
Linus Torvalds 已提交
2251 2252 2253 2254

	order = 0;
	for ( ; ; ) {
		unsigned long new_order;
2255
		int ret;
2256

L
Linus Torvalds 已提交
2257 2258 2259 2260 2261 2262 2263 2264 2265 2266
		prepare_to_wait(&pgdat->kswapd_wait, &wait, TASK_INTERRUPTIBLE);
		new_order = pgdat->kswapd_max_order;
		pgdat->kswapd_max_order = 0;
		if (order < new_order) {
			/*
			 * Don't sleep if someone wants a larger 'order'
			 * allocation
			 */
			order = new_order;
		} else {
2267 2268 2269 2270
			if (!freezing(current) && !kthread_should_stop()) {
				long remaining = 0;

				/* Try to sleep for a short interval */
2271
				if (!sleeping_prematurely(pgdat, order, remaining)) {
2272 2273 2274 2275 2276 2277 2278 2279 2280 2281
					remaining = schedule_timeout(HZ/10);
					finish_wait(&pgdat->kswapd_wait, &wait);
					prepare_to_wait(&pgdat->kswapd_wait, &wait, TASK_INTERRUPTIBLE);
				}

				/*
				 * After a short sleep, check if it was a
				 * premature sleep. If not, then go fully
				 * to sleep until explicitly woken up
				 */
2282
				if (!sleeping_prematurely(pgdat, order, remaining))
2283 2284 2285
					schedule();
				else {
					if (remaining)
2286
						count_vm_event(KSWAPD_LOW_WMARK_HIT_QUICKLY);
2287
					else
2288
						count_vm_event(KSWAPD_HIGH_WMARK_HIT_QUICKLY);
2289 2290
				}
			}
2291

L
Linus Torvalds 已提交
2292 2293 2294 2295
			order = pgdat->kswapd_max_order;
		}
		finish_wait(&pgdat->kswapd_wait, &wait);

2296 2297 2298 2299 2300 2301 2302 2303 2304
		ret = try_to_freeze();
		if (kthread_should_stop())
			break;

		/*
		 * We can speed up thawing tasks if we don't call balance_pgdat
		 * after returning from the refrigerator
		 */
		if (!ret)
2305
			balance_pgdat(pgdat, order);
L
Linus Torvalds 已提交
2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316
	}
	return 0;
}

/*
 * A zone is low on free memory, so wake its kswapd task to service it.
 */
void wakeup_kswapd(struct zone *zone, int order)
{
	pg_data_t *pgdat;

2317
	if (!populated_zone(zone))
L
Linus Torvalds 已提交
2318 2319 2320
		return;

	pgdat = zone->zone_pgdat;
2321
	if (zone_watermark_ok(zone, order, low_wmark_pages(zone), 0, 0))
L
Linus Torvalds 已提交
2322 2323 2324
		return;
	if (pgdat->kswapd_max_order < order)
		pgdat->kswapd_max_order = order;
2325
	if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL))
L
Linus Torvalds 已提交
2326
		return;
2327
	if (!waitqueue_active(&pgdat->kswapd_wait))
L
Linus Torvalds 已提交
2328
		return;
2329
	wake_up_interruptible(&pgdat->kswapd_wait);
L
Linus Torvalds 已提交
2330 2331
}

2332 2333 2334 2335 2336 2337 2338 2339
/*
 * The reclaimable count would be mostly accurate.
 * The less reclaimable pages may be
 * - mlocked pages, which will be moved to unevictable list when encountered
 * - mapped pages, which may require several travels to be reclaimed
 * - dirty pages, which is not "instantly" reclaimable
 */
unsigned long global_reclaimable_pages(void)
2340
{
2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364
	int nr;

	nr = global_page_state(NR_ACTIVE_FILE) +
	     global_page_state(NR_INACTIVE_FILE);

	if (nr_swap_pages > 0)
		nr += global_page_state(NR_ACTIVE_ANON) +
		      global_page_state(NR_INACTIVE_ANON);

	return nr;
}

unsigned long zone_reclaimable_pages(struct zone *zone)
{
	int nr;

	nr = zone_page_state(zone, NR_ACTIVE_FILE) +
	     zone_page_state(zone, NR_INACTIVE_FILE);

	if (nr_swap_pages > 0)
		nr += zone_page_state(zone, NR_ACTIVE_ANON) +
		      zone_page_state(zone, NR_INACTIVE_ANON);

	return nr;
2365 2366
}

2367
#ifdef CONFIG_HIBERNATION
L
Linus Torvalds 已提交
2368
/*
2369
 * Try to free `nr_to_reclaim' of memory, system-wide, and return the number of
2370 2371 2372 2373 2374
 * freed pages.
 *
 * Rather than trying to age LRUs the aim is to preserve the overall
 * LRU order by reclaiming preferentially
 * inactive > active > active referenced > active mapped
L
Linus Torvalds 已提交
2375
 */
2376
unsigned long shrink_all_memory(unsigned long nr_to_reclaim)
L
Linus Torvalds 已提交
2377
{
2378 2379
	struct reclaim_state reclaim_state;
	struct scan_control sc = {
2380 2381 2382
		.gfp_mask = GFP_HIGHUSER_MOVABLE,
		.may_swap = 1,
		.may_unmap = 1,
2383
		.may_writepage = 1,
2384 2385 2386 2387
		.nr_to_reclaim = nr_to_reclaim,
		.hibernation_mode = 1,
		.swappiness = vm_swappiness,
		.order = 0,
2388
		.isolate_pages = isolate_pages_global,
L
Linus Torvalds 已提交
2389
	};
2390 2391 2392
	struct zonelist * zonelist = node_zonelist(numa_node_id(), sc.gfp_mask);
	struct task_struct *p = current;
	unsigned long nr_reclaimed;
L
Linus Torvalds 已提交
2393

2394 2395 2396 2397
	p->flags |= PF_MEMALLOC;
	lockdep_set_current_reclaim_state(sc.gfp_mask);
	reclaim_state.reclaimed_slab = 0;
	p->reclaim_state = &reclaim_state;
2398

2399
	nr_reclaimed = do_try_to_free_pages(zonelist, &sc);
2400

2401 2402 2403
	p->reclaim_state = NULL;
	lockdep_clear_current_reclaim_state();
	p->flags &= ~PF_MEMALLOC;
2404

2405
	return nr_reclaimed;
L
Linus Torvalds 已提交
2406
}
2407
#endif /* CONFIG_HIBERNATION */
L
Linus Torvalds 已提交
2408 2409 2410 2411 2412

/* It's optimal to keep kswapds on the same CPUs as their memory, but
   not required for correctness.  So if the last cpu in a node goes
   away, we get changed to run anywhere: as the first one comes back,
   restore their cpu bindings. */
2413
static int __devinit cpu_callback(struct notifier_block *nfb,
2414
				  unsigned long action, void *hcpu)
L
Linus Torvalds 已提交
2415
{
2416
	int nid;
L
Linus Torvalds 已提交
2417

2418
	if (action == CPU_ONLINE || action == CPU_ONLINE_FROZEN) {
2419
		for_each_node_state(nid, N_HIGH_MEMORY) {
2420
			pg_data_t *pgdat = NODE_DATA(nid);
2421 2422 2423
			const struct cpumask *mask;

			mask = cpumask_of_node(pgdat->node_id);
2424

2425
			if (cpumask_any_and(cpu_online_mask, mask) < nr_cpu_ids)
L
Linus Torvalds 已提交
2426
				/* One of our CPUs online: restore mask */
2427
				set_cpus_allowed_ptr(pgdat->kswapd, mask);
L
Linus Torvalds 已提交
2428 2429 2430 2431 2432
		}
	}
	return NOTIFY_OK;
}

2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454
/*
 * This kswapd start function will be called by init and node-hot-add.
 * On node-hot-add, kswapd will moved to proper cpus if cpus are hot-added.
 */
int kswapd_run(int nid)
{
	pg_data_t *pgdat = NODE_DATA(nid);
	int ret = 0;

	if (pgdat->kswapd)
		return 0;

	pgdat->kswapd = kthread_run(kswapd, pgdat, "kswapd%d", nid);
	if (IS_ERR(pgdat->kswapd)) {
		/* failure at boot is fatal */
		BUG_ON(system_state == SYSTEM_BOOTING);
		printk("Failed to start kswapd on node %d\n",nid);
		ret = -1;
	}
	return ret;
}

2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465
/*
 * Called by memory hotplug when all memory in a node is offlined.
 */
void kswapd_stop(int nid)
{
	struct task_struct *kswapd = NODE_DATA(nid)->kswapd;

	if (kswapd)
		kthread_stop(kswapd);
}

L
Linus Torvalds 已提交
2466 2467
static int __init kswapd_init(void)
{
2468
	int nid;
2469

L
Linus Torvalds 已提交
2470
	swap_setup();
2471
	for_each_node_state(nid, N_HIGH_MEMORY)
2472
 		kswapd_run(nid);
L
Linus Torvalds 已提交
2473 2474 2475 2476 2477
	hotcpu_notifier(cpu_callback, 0);
	return 0;
}

module_init(kswapd_init)
2478 2479 2480 2481 2482 2483 2484 2485 2486 2487

#ifdef CONFIG_NUMA
/*
 * Zone reclaim mode
 *
 * If non-zero call zone_reclaim when the number of free pages falls below
 * the watermarks.
 */
int zone_reclaim_mode __read_mostly;

2488
#define RECLAIM_OFF 0
2489
#define RECLAIM_ZONE (1<<0)	/* Run shrink_inactive_list on the zone */
2490 2491 2492
#define RECLAIM_WRITE (1<<1)	/* Writeout pages during reclaim */
#define RECLAIM_SWAP (1<<2)	/* Swap pages out during reclaim */

2493 2494 2495 2496 2497 2498 2499
/*
 * Priority for ZONE_RECLAIM. This determines the fraction of pages
 * of a node considered for each zone_reclaim. 4 scans 1/16th of
 * a zone.
 */
#define ZONE_RECLAIM_PRIORITY 4

2500 2501 2502 2503 2504 2505
/*
 * Percentage of pages in a zone that must be unmapped for zone_reclaim to
 * occur.
 */
int sysctl_min_unmapped_ratio = 1;

2506 2507 2508 2509 2510 2511
/*
 * If the number of slab pages in a zone grows beyond this percentage then
 * slab reclaim needs to occur.
 */
int sysctl_min_slab_ratio = 5;

2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553
static inline unsigned long zone_unmapped_file_pages(struct zone *zone)
{
	unsigned long file_mapped = zone_page_state(zone, NR_FILE_MAPPED);
	unsigned long file_lru = zone_page_state(zone, NR_INACTIVE_FILE) +
		zone_page_state(zone, NR_ACTIVE_FILE);

	/*
	 * It's possible for there to be more file mapped pages than
	 * accounted for by the pages on the file LRU lists because
	 * tmpfs pages accounted for as ANON can also be FILE_MAPPED
	 */
	return (file_lru > file_mapped) ? (file_lru - file_mapped) : 0;
}

/* Work out how many page cache pages we can reclaim in this reclaim_mode */
static long zone_pagecache_reclaimable(struct zone *zone)
{
	long nr_pagecache_reclaimable;
	long delta = 0;

	/*
	 * If RECLAIM_SWAP is set, then all file pages are considered
	 * potentially reclaimable. Otherwise, we have to worry about
	 * pages like swapcache and zone_unmapped_file_pages() provides
	 * a better estimate
	 */
	if (zone_reclaim_mode & RECLAIM_SWAP)
		nr_pagecache_reclaimable = zone_page_state(zone, NR_FILE_PAGES);
	else
		nr_pagecache_reclaimable = zone_unmapped_file_pages(zone);

	/* If we can't clean pages, remove dirty pages from consideration */
	if (!(zone_reclaim_mode & RECLAIM_WRITE))
		delta += zone_page_state(zone, NR_FILE_DIRTY);

	/* Watch for any possible underflows due to delta */
	if (unlikely(delta > nr_pagecache_reclaimable))
		delta = nr_pagecache_reclaimable;

	return nr_pagecache_reclaimable - delta;
}

2554 2555 2556
/*
 * Try to free up some pages from this zone through reclaim.
 */
2557
static int __zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order)
2558
{
2559
	/* Minimum pages needed in order to stay on node */
2560
	const unsigned long nr_pages = 1 << order;
2561 2562
	struct task_struct *p = current;
	struct reclaim_state reclaim_state;
2563
	int priority;
2564 2565
	struct scan_control sc = {
		.may_writepage = !!(zone_reclaim_mode & RECLAIM_WRITE),
2566
		.may_unmap = !!(zone_reclaim_mode & RECLAIM_SWAP),
2567
		.may_swap = 1,
2568 2569
		.nr_to_reclaim = max_t(unsigned long, nr_pages,
				       SWAP_CLUSTER_MAX),
2570
		.gfp_mask = gfp_mask,
2571
		.swappiness = vm_swappiness,
2572
		.order = order,
2573
		.isolate_pages = isolate_pages_global,
2574
	};
2575
	unsigned long slab_reclaimable;
2576 2577 2578

	disable_swap_token();
	cond_resched();
2579 2580 2581 2582 2583 2584
	/*
	 * We need to be able to allocate from the reserves for RECLAIM_SWAP
	 * and we also need to be able to write out pages for RECLAIM_WRITE
	 * and RECLAIM_SWAP.
	 */
	p->flags |= PF_MEMALLOC | PF_SWAPWRITE;
2585
	lockdep_set_current_reclaim_state(gfp_mask);
2586 2587
	reclaim_state.reclaimed_slab = 0;
	p->reclaim_state = &reclaim_state;
2588

2589
	if (zone_pagecache_reclaimable(zone) > zone->min_unmapped_pages) {
2590 2591 2592 2593 2594 2595
		/*
		 * Free memory by calling shrink zone with increasing
		 * priorities until we have enough memory freed.
		 */
		priority = ZONE_RECLAIM_PRIORITY;
		do {
2596
			note_zone_scanning_priority(zone, priority);
2597
			shrink_zone(priority, zone, &sc);
2598
			priority--;
2599
		} while (priority >= 0 && sc.nr_reclaimed < nr_pages);
2600
	}
2601

2602 2603
	slab_reclaimable = zone_page_state(zone, NR_SLAB_RECLAIMABLE);
	if (slab_reclaimable > zone->min_slab_pages) {
2604
		/*
2605
		 * shrink_slab() does not currently allow us to determine how
2606 2607 2608 2609
		 * many pages were freed in this zone. So we take the current
		 * number of slab pages and shake the slab until it is reduced
		 * by the same nr_pages that we used for reclaiming unmapped
		 * pages.
2610
		 *
2611 2612
		 * Note that shrink_slab will free memory on all zones and may
		 * take a long time.
2613
		 */
2614
		while (shrink_slab(sc.nr_scanned, gfp_mask, order) &&
2615 2616
			zone_page_state(zone, NR_SLAB_RECLAIMABLE) >
				slab_reclaimable - nr_pages)
2617
			;
2618 2619 2620 2621 2622

		/*
		 * Update nr_reclaimed by the number of slab pages we
		 * reclaimed from this zone.
		 */
2623
		sc.nr_reclaimed += slab_reclaimable -
2624
			zone_page_state(zone, NR_SLAB_RECLAIMABLE);
2625 2626
	}

2627
	p->reclaim_state = NULL;
2628
	current->flags &= ~(PF_MEMALLOC | PF_SWAPWRITE);
2629
	lockdep_clear_current_reclaim_state();
2630
	return sc.nr_reclaimed >= nr_pages;
2631
}
2632 2633 2634 2635

int zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order)
{
	int node_id;
2636
	int ret;
2637 2638

	/*
2639 2640
	 * Zone reclaim reclaims unmapped file backed pages and
	 * slab pages if we are over the defined limits.
2641
	 *
2642 2643 2644 2645 2646
	 * A small portion of unmapped file backed pages is needed for
	 * file I/O otherwise pages read by file I/O will be immediately
	 * thrown out if the zone is overallocated. So we do not reclaim
	 * if less than a specified percentage of the zone is used by
	 * unmapped file backed pages.
2647
	 */
2648 2649
	if (zone_pagecache_reclaimable(zone) <= zone->min_unmapped_pages &&
	    zone_page_state(zone, NR_SLAB_RECLAIMABLE) <= zone->min_slab_pages)
2650
		return ZONE_RECLAIM_FULL;
2651

2652
	if (zone->all_unreclaimable)
2653
		return ZONE_RECLAIM_FULL;
2654

2655
	/*
2656
	 * Do not scan if the allocation should not be delayed.
2657
	 */
2658
	if (!(gfp_mask & __GFP_WAIT) || (current->flags & PF_MEMALLOC))
2659
		return ZONE_RECLAIM_NOSCAN;
2660 2661 2662 2663 2664 2665 2666

	/*
	 * Only run zone reclaim on the local zone or on zones that do not
	 * have associated processors. This will favor the local processor
	 * over remote processors and spread off node memory allocations
	 * as wide as possible.
	 */
2667
	node_id = zone_to_nid(zone);
2668
	if (node_state(node_id, N_CPU) && node_id != numa_node_id())
2669
		return ZONE_RECLAIM_NOSCAN;
2670 2671

	if (zone_test_and_set_flag(zone, ZONE_RECLAIM_LOCKED))
2672 2673
		return ZONE_RECLAIM_NOSCAN;

2674 2675 2676
	ret = __zone_reclaim(zone, gfp_mask, order);
	zone_clear_flag(zone, ZONE_RECLAIM_LOCKED);

2677 2678 2679
	if (!ret)
		count_vm_event(PGSCAN_ZONE_RECLAIM_FAILED);

2680
	return ret;
2681
}
2682
#endif
L
Lee Schermerhorn 已提交
2683 2684 2685 2686 2687 2688 2689

/*
 * page_evictable - test whether a page is evictable
 * @page: the page to test
 * @vma: the VMA in which the page is or will be mapped, may be NULL
 *
 * Test whether page is evictable--i.e., should be placed on active/inactive
N
Nick Piggin 已提交
2690 2691
 * lists vs unevictable list.  The vma argument is !NULL when called from the
 * fault path to determine how to instantate a new page.
L
Lee Schermerhorn 已提交
2692 2693
 *
 * Reasons page might not be evictable:
2694
 * (1) page's mapping marked unevictable
N
Nick Piggin 已提交
2695
 * (2) page is part of an mlocked VMA
2696
 *
L
Lee Schermerhorn 已提交
2697 2698 2699 2700
 */
int page_evictable(struct page *page, struct vm_area_struct *vma)
{

2701 2702 2703
	if (mapping_unevictable(page_mapping(page)))
		return 0;

N
Nick Piggin 已提交
2704 2705
	if (PageMlocked(page) || (vma && is_mlocked_vma(vma, page)))
		return 0;
L
Lee Schermerhorn 已提交
2706 2707 2708

	return 1;
}
2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727

/**
 * check_move_unevictable_page - check page for evictability and move to appropriate zone lru list
 * @page: page to check evictability and move to appropriate lru list
 * @zone: zone page is in
 *
 * Checks a page for evictability and moves the page to the appropriate
 * zone lru list.
 *
 * Restrictions: zone->lru_lock must be held, page must be on LRU and must
 * have PageUnevictable set.
 */
static void check_move_unevictable_page(struct page *page, struct zone *zone)
{
	VM_BUG_ON(PageActive(page));

retry:
	ClearPageUnevictable(page);
	if (page_evictable(page, NULL)) {
2728
		enum lru_list l = page_lru_base_type(page);
2729

2730 2731
		__dec_zone_state(zone, NR_UNEVICTABLE);
		list_move(&page->lru, &zone->lru[l].list);
K
KAMEZAWA Hiroyuki 已提交
2732
		mem_cgroup_move_lists(page, LRU_UNEVICTABLE, l);
2733 2734 2735 2736 2737 2738 2739 2740
		__inc_zone_state(zone, NR_INACTIVE_ANON + l);
		__count_vm_event(UNEVICTABLE_PGRESCUED);
	} else {
		/*
		 * rotate unevictable list
		 */
		SetPageUnevictable(page);
		list_move(&page->lru, &zone->lru[LRU_UNEVICTABLE].list);
K
KAMEZAWA Hiroyuki 已提交
2741
		mem_cgroup_rotate_lru_list(page, LRU_UNEVICTABLE);
2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800
		if (page_evictable(page, NULL))
			goto retry;
	}
}

/**
 * scan_mapping_unevictable_pages - scan an address space for evictable pages
 * @mapping: struct address_space to scan for evictable pages
 *
 * Scan all pages in mapping.  Check unevictable pages for
 * evictability and move them to the appropriate zone lru list.
 */
void scan_mapping_unevictable_pages(struct address_space *mapping)
{
	pgoff_t next = 0;
	pgoff_t end   = (i_size_read(mapping->host) + PAGE_CACHE_SIZE - 1) >>
			 PAGE_CACHE_SHIFT;
	struct zone *zone;
	struct pagevec pvec;

	if (mapping->nrpages == 0)
		return;

	pagevec_init(&pvec, 0);
	while (next < end &&
		pagevec_lookup(&pvec, mapping, next, PAGEVEC_SIZE)) {
		int i;
		int pg_scanned = 0;

		zone = NULL;

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

			pg_scanned++;
			if (page_index > next)
				next = page_index;
			next++;

			if (pagezone != zone) {
				if (zone)
					spin_unlock_irq(&zone->lru_lock);
				zone = pagezone;
				spin_lock_irq(&zone->lru_lock);
			}

			if (PageLRU(page) && PageUnevictable(page))
				check_move_unevictable_page(page, zone);
		}
		if (zone)
			spin_unlock_irq(&zone->lru_lock);
		pagevec_release(&pvec);

		count_vm_events(UNEVICTABLE_PGSCANNED, pg_scanned);
	}

}
2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812

/**
 * scan_zone_unevictable_pages - check unevictable list for evictable pages
 * @zone - zone of which to scan the unevictable list
 *
 * Scan @zone's unevictable LRU lists to check for pages that have become
 * evictable.  Move those that have to @zone's inactive list where they
 * become candidates for reclaim, unless shrink_inactive_zone() decides
 * to reactivate them.  Pages that are still unevictable are rotated
 * back onto @zone's unevictable list.
 */
#define SCAN_UNEVICTABLE_BATCH_SIZE 16UL /* arbitrary lock hold batch size */
2813
static void scan_zone_unevictable_pages(struct zone *zone)
2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854
{
	struct list_head *l_unevictable = &zone->lru[LRU_UNEVICTABLE].list;
	unsigned long scan;
	unsigned long nr_to_scan = zone_page_state(zone, NR_UNEVICTABLE);

	while (nr_to_scan > 0) {
		unsigned long batch_size = min(nr_to_scan,
						SCAN_UNEVICTABLE_BATCH_SIZE);

		spin_lock_irq(&zone->lru_lock);
		for (scan = 0;  scan < batch_size; scan++) {
			struct page *page = lru_to_page(l_unevictable);

			if (!trylock_page(page))
				continue;

			prefetchw_prev_lru_page(page, l_unevictable, flags);

			if (likely(PageLRU(page) && PageUnevictable(page)))
				check_move_unevictable_page(page, zone);

			unlock_page(page);
		}
		spin_unlock_irq(&zone->lru_lock);

		nr_to_scan -= batch_size;
	}
}


/**
 * scan_all_zones_unevictable_pages - scan all unevictable lists for evictable pages
 *
 * A really big hammer:  scan all zones' unevictable LRU lists to check for
 * pages that have become evictable.  Move those back to the zones'
 * inactive list where they become candidates for reclaim.
 * This occurs when, e.g., we have unswappable pages on the unevictable lists,
 * and we add swap to the system.  As such, it runs in the context of a task
 * that has possibly/probably made some previously unevictable pages
 * evictable.
 */
2855
static void scan_all_zones_unevictable_pages(void)
2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870
{
	struct zone *zone;

	for_each_zone(zone) {
		scan_zone_unevictable_pages(zone);
	}
}

/*
 * scan_unevictable_pages [vm] sysctl handler.  On demand re-scan of
 * all nodes' unevictable lists for evictable pages
 */
unsigned long scan_unevictable_pages;

int scan_unevictable_handler(struct ctl_table *table, int write,
2871
			   void __user *buffer,
2872 2873
			   size_t *length, loff_t *ppos)
{
2874
	proc_doulongvec_minmax(table, write, buffer, length, ppos);
2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929

	if (write && *(unsigned long *)table->data)
		scan_all_zones_unevictable_pages();

	scan_unevictable_pages = 0;
	return 0;
}

/*
 * per node 'scan_unevictable_pages' attribute.  On demand re-scan of
 * a specified node's per zone unevictable lists for evictable pages.
 */

static ssize_t read_scan_unevictable_node(struct sys_device *dev,
					  struct sysdev_attribute *attr,
					  char *buf)
{
	return sprintf(buf, "0\n");	/* always zero; should fit... */
}

static ssize_t write_scan_unevictable_node(struct sys_device *dev,
					   struct sysdev_attribute *attr,
					const char *buf, size_t count)
{
	struct zone *node_zones = NODE_DATA(dev->id)->node_zones;
	struct zone *zone;
	unsigned long res;
	unsigned long req = strict_strtoul(buf, 10, &res);

	if (!req)
		return 1;	/* zero is no-op */

	for (zone = node_zones; zone - node_zones < MAX_NR_ZONES; ++zone) {
		if (!populated_zone(zone))
			continue;
		scan_zone_unevictable_pages(zone);
	}
	return 1;
}


static SYSDEV_ATTR(scan_unevictable_pages, S_IRUGO | S_IWUSR,
			read_scan_unevictable_node,
			write_scan_unevictable_node);

int scan_unevictable_register_node(struct node *node)
{
	return sysdev_create_file(&node->sysdev, &attr_scan_unevictable_pages);
}

void scan_unevictable_unregister_node(struct node *node)
{
	sysdev_remove_file(&node->sysdev, &attr_scan_unevictable_pages);
}