vmscan.c 70.4 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
/*
 *  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>
#include <linux/slab.h>
#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 55
struct scan_control {
	/* Incremented by the number of inactive pages that were scanned */
	unsigned long nr_scanned;

	/* This context's GFP mask */
A
Al Viro 已提交
56
	gfp_t gfp_mask;
L
Linus Torvalds 已提交
57 58 59

	int may_writepage;

60 61 62
	/* Can pages be swapped as part of reclaim? */
	int may_swap;

L
Linus Torvalds 已提交
63 64 65 66 67
	/* This context's SWAP_CLUSTER_MAX. If freeing memory for
	 * suspend, we effectively ignore SWAP_CLUSTER_MAX.
	 * In this context, it doesn't matter that we scan the
	 * whole list at once. */
	int swap_cluster_max;
68 69

	int swappiness;
70 71

	int all_unreclaimable;
A
Andy Whitcroft 已提交
72 73

	int order;
74 75 76 77 78 79 80 81

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

	/* 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,
82
			int active, int file);
L
Linus Torvalds 已提交
83 84 85 86 87 88 89 90 91 92 93 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
};

#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;
119
long vm_total_pages;	/* The total number of pages which the VM controls */
L
Linus Torvalds 已提交
120 121 122 123

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

124
#ifdef CONFIG_CGROUP_MEM_RES_CTLR
125 126 127 128 129
#define scan_global_lru(sc)	(!(sc)->mem_cgroup)
#else
#define scan_global_lru(sc)	(1)
#endif

L
Linus Torvalds 已提交
130 131 132
/*
 * Add a shrinker callback to be called from the vm
 */
133
void register_shrinker(struct shrinker *shrinker)
L
Linus Torvalds 已提交
134
{
135 136 137 138
	shrinker->nr = 0;
	down_write(&shrinker_rwsem);
	list_add_tail(&shrinker->list, &shrinker_list);
	up_write(&shrinker_rwsem);
L
Linus Torvalds 已提交
139
}
140
EXPORT_SYMBOL(register_shrinker);
L
Linus Torvalds 已提交
141 142 143 144

/*
 * Remove one
 */
145
void unregister_shrinker(struct shrinker *shrinker)
L
Linus Torvalds 已提交
146 147 148 149 150
{
	down_write(&shrinker_rwsem);
	list_del(&shrinker->list);
	up_write(&shrinker_rwsem);
}
151
EXPORT_SYMBOL(unregister_shrinker);
L
Linus Torvalds 已提交
152 153 154 155 156 157 158 159 160 161

#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
162
 * If the vm encountered mapped pages on the LRU it increase the pressure on
L
Linus Torvalds 已提交
163 164 165 166 167 168 169
 * 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.
170 171
 *
 * Returns the number of slab objects which we shrunk.
L
Linus Torvalds 已提交
172
 */
173 174
unsigned long shrink_slab(unsigned long scanned, gfp_t gfp_mask,
			unsigned long lru_pages)
L
Linus Torvalds 已提交
175 176
{
	struct shrinker *shrinker;
177
	unsigned long ret = 0;
L
Linus Torvalds 已提交
178 179 180 181 182

	if (scanned == 0)
		scanned = SWAP_CLUSTER_MAX;

	if (!down_read_trylock(&shrinker_rwsem))
183
		return 1;	/* Assume we'll be able to shrink next time */
L
Linus Torvalds 已提交
184 185 186 187

	list_for_each_entry(shrinker, &shrinker_list, list) {
		unsigned long long delta;
		unsigned long total_scan;
188
		unsigned long max_pass = (*shrinker->shrink)(0, gfp_mask);
L
Linus Torvalds 已提交
189 190

		delta = (4 * scanned) / shrinker->seeks;
191
		delta *= max_pass;
L
Linus Torvalds 已提交
192 193
		do_div(delta, lru_pages + 1);
		shrinker->nr += delta;
194 195
		if (shrinker->nr < 0) {
			printk(KERN_ERR "%s: nr=%ld\n",
196
					__func__, shrinker->nr);
197 198 199 200 201 202 203 204 205 206
			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
207 208 209 210 211 212 213

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

		while (total_scan >= SHRINK_BATCH) {
			long this_scan = SHRINK_BATCH;
			int shrink_ret;
214
			int nr_before;
L
Linus Torvalds 已提交
215

216 217
			nr_before = (*shrinker->shrink)(0, gfp_mask);
			shrink_ret = (*shrinker->shrink)(this_scan, gfp_mask);
L
Linus Torvalds 已提交
218 219
			if (shrink_ret == -1)
				break;
220 221
			if (shrink_ret < nr_before)
				ret += nr_before - shrink_ret;
222
			count_vm_events(SLABS_SCANNED, this_scan);
L
Linus Torvalds 已提交
223 224 225 226 227 228 229 230
			total_scan -= this_scan;

			cond_resched();
		}

		shrinker->nr += total_scan;
	}
	up_read(&shrinker_rwsem);
231
	return ret;
L
Linus Torvalds 已提交
232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261
}

/* Called without lock on whether page is mapped, so answer is unstable */
static inline int page_mapping_inuse(struct page *page)
{
	struct address_space *mapping;

	/* Page is in somebody's page tables. */
	if (page_mapped(page))
		return 1;

	/* Be more reluctant to reclaim swapcache than pagecache */
	if (PageSwapCache(page))
		return 1;

	mapping = page_mapping(page);
	if (!mapping)
		return 0;

	/* File is mmap'd by somebody? */
	return mapping_mapped(mapping);
}

static inline int is_page_cache_freeable(struct page *page)
{
	return page_count(page) - !!PagePrivate(page) == 2;
}

static int may_write_to_queue(struct backing_dev_info *bdi)
{
262
	if (current->flags & PF_SWAPWRITE)
L
Linus Torvalds 已提交
263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286
		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);
287 288
	if (page_mapping(page) == mapping)
		mapping_set_error(mapping, error);
L
Linus Torvalds 已提交
289 290 291
	unlock_page(page);
}

292 293 294 295 296 297
/* Request for sync pageout. */
enum pageout_io {
	PAGEOUT_IO_ASYNC,
	PAGEOUT_IO_SYNC,
};

298 299 300 301 302 303 304 305 306 307 308 309
/* 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
310
/*
A
Andrew Morton 已提交
311 312
 * pageout is called by shrink_page_list() for each dirty page.
 * Calls ->writepage().
L
Linus Torvalds 已提交
313
 */
314 315
static pageout_t pageout(struct page *page, struct address_space *mapping,
						enum pageout_io sync_writeback)
L
Linus Torvalds 已提交
316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340
{
	/*
	 * 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.
	 *
	 * If this process is currently in generic_file_write() against
	 * 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.
	 * See swapfile.c:page_queue_congested().
	 */
	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.
		 */
341
		if (PagePrivate(page)) {
L
Linus Torvalds 已提交
342 343
			if (try_to_free_buffers(page)) {
				ClearPageDirty(page);
344
				printk("%s: orphaned page\n", __func__);
L
Linus Torvalds 已提交
345 346 347 348 349 350 351 352 353 354 355 356 357 358 359
				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,
360 361
			.range_start = 0,
			.range_end = LLONG_MAX,
L
Linus Torvalds 已提交
362 363 364 365 366 367 368 369
			.nonblocking = 1,
			.for_reclaim = 1,
		};

		SetPageReclaim(page);
		res = mapping->a_ops->writepage(page, &wbc);
		if (res < 0)
			handle_write_error(mapping, page, res);
370
		if (res == AOP_WRITEPAGE_ACTIVATE) {
L
Linus Torvalds 已提交
371 372 373
			ClearPageReclaim(page);
			return PAGE_ACTIVATE;
		}
374 375 376 377 378 379 380 381 382

		/*
		 * 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
383 384 385 386
		if (!PageWriteback(page)) {
			/* synchronous write or broken a_ops? */
			ClearPageReclaim(page);
		}
387
		inc_zone_page_state(page, NR_VMSCAN_WRITE);
L
Linus Torvalds 已提交
388 389 390 391 392 393
		return PAGE_SUCCESS;
	}

	return PAGE_CLEAN;
}

394
/*
N
Nick Piggin 已提交
395 396
 * Same as remove_mapping, but if the page is removed from the mapping, it
 * gets returned with a refcount of 0.
397
 */
N
Nick Piggin 已提交
398
static int __remove_mapping(struct address_space *mapping, struct page *page)
399
{
400 401
	BUG_ON(!PageLocked(page));
	BUG_ON(mapping != page_mapping(page));
402

N
Nick Piggin 已提交
403
	spin_lock_irq(&mapping->tree_lock);
404
	/*
N
Nick Piggin 已提交
405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427
	 * 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.
428
	 */
N
Nick Piggin 已提交
429
	if (!page_freeze_refs(page, 2))
430
		goto cannot_free;
N
Nick Piggin 已提交
431 432 433
	/* note: atomic_cmpxchg in page_freeze_refs provides the smp_rmb */
	if (unlikely(PageDirty(page))) {
		page_unfreeze_refs(page, 2);
434
		goto cannot_free;
N
Nick Piggin 已提交
435
	}
436 437 438 439

	if (PageSwapCache(page)) {
		swp_entry_t swap = { .val = page_private(page) };
		__delete_from_swap_cache(page);
N
Nick Piggin 已提交
440
		spin_unlock_irq(&mapping->tree_lock);
441
		swap_free(swap);
N
Nick Piggin 已提交
442 443
	} else {
		__remove_from_page_cache(page);
N
Nick Piggin 已提交
444
		spin_unlock_irq(&mapping->tree_lock);
445 446 447 448 449
	}

	return 1;

cannot_free:
N
Nick Piggin 已提交
450
	spin_unlock_irq(&mapping->tree_lock);
451 452 453
	return 0;
}

N
Nick Piggin 已提交
454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473
/*
 * 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
474 475 476 477 478 479 480 481 482 483 484 485 486 487
/**
 * 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.
 */
#ifdef CONFIG_UNEVICTABLE_LRU
void putback_lru_page(struct page *page)
{
	int lru;
	int active = !!TestClearPageActive(page);
488
	int was_unevictable = PageUnevictable(page);
L
Lee Schermerhorn 已提交
489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529

	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.
		 */
		lru = active + page_is_file_cache(page);
		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);
	}
	mem_cgroup_move_lists(page, lru);

	/*
	 * 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.
		 */
	}

530 531 532 533 534
	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
535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552
	put_page(page);		/* drop ref from isolate */
}

#else /* CONFIG_UNEVICTABLE_LRU */

void putback_lru_page(struct page *page)
{
	int lru;
	VM_BUG_ON(PageLRU(page));

	lru = !!TestClearPageActive(page) + page_is_file_cache(page);
	lru_cache_add_lru(page, lru);
	mem_cgroup_move_lists(page, lru);
	put_page(page);
}
#endif /* CONFIG_UNEVICTABLE_LRU */


L
Linus Torvalds 已提交
553
/*
A
Andrew Morton 已提交
554
 * shrink_page_list() returns the number of reclaimed pages
L
Linus Torvalds 已提交
555
 */
A
Andrew Morton 已提交
556
static unsigned long shrink_page_list(struct list_head *page_list,
557 558
					struct scan_control *sc,
					enum pageout_io sync_writeback)
L
Linus Torvalds 已提交
559 560 561 562
{
	LIST_HEAD(ret_pages);
	struct pagevec freed_pvec;
	int pgactivate = 0;
563
	unsigned long nr_reclaimed = 0;
L
Linus Torvalds 已提交
564 565 566 567 568 569 570 571 572 573 574 575 576 577 578

	cond_resched();

	pagevec_init(&freed_pvec, 1);
	while (!list_empty(page_list)) {
		struct address_space *mapping;
		struct page *page;
		int may_enter_fs;
		int referenced;

		cond_resched();

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

N
Nick Piggin 已提交
579
		if (!trylock_page(page))
L
Linus Torvalds 已提交
580 581
			goto keep;

N
Nick Piggin 已提交
582
		VM_BUG_ON(PageActive(page));
L
Linus Torvalds 已提交
583 584

		sc->nr_scanned++;
585

N
Nick Piggin 已提交
586 587
		if (unlikely(!page_evictable(page, NULL)))
			goto cull_mlocked;
L
Lee Schermerhorn 已提交
588

589 590 591
		if (!sc->may_swap && page_mapped(page))
			goto keep_locked;

L
Linus Torvalds 已提交
592 593 594 595
		/* Double the slab pressure for mapped and swapcache pages */
		if (page_mapped(page) || PageSwapCache(page))
			sc->nr_scanned++;

596 597 598 599 600 601 602 603 604 605 606 607 608 609
		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);
610
			else
611 612
				goto keep_locked;
		}
L
Linus Torvalds 已提交
613

614
		referenced = page_referenced(page, 1, sc->mem_cgroup);
L
Linus Torvalds 已提交
615
		/* In active use or really unfreeable?  Activate it. */
A
Andy Whitcroft 已提交
616 617
		if (sc->order <= PAGE_ALLOC_COSTLY_ORDER &&
					referenced && page_mapping_inuse(page))
L
Linus Torvalds 已提交
618 619 620 621 622 623
			goto activate_locked;

		/*
		 * Anonymous process memory has backing store?
		 * Try to allocate it some swap space here.
		 */
N
Nick Piggin 已提交
624
		if (PageAnon(page) && !PageSwapCache(page)) {
625 626
			if (!(sc->gfp_mask & __GFP_IO))
				goto keep_locked;
627
			if (!add_to_swap(page))
L
Linus Torvalds 已提交
628
				goto activate_locked;
629
			may_enter_fs = 1;
N
Nick Piggin 已提交
630
		}
L
Linus Torvalds 已提交
631 632 633 634 635 636 637 638

		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) {
639
			switch (try_to_unmap(page, 0)) {
L
Linus Torvalds 已提交
640 641 642 643
			case SWAP_FAIL:
				goto activate_locked;
			case SWAP_AGAIN:
				goto keep_locked;
N
Nick Piggin 已提交
644 645
			case SWAP_MLOCK:
				goto cull_mlocked;
L
Linus Torvalds 已提交
646 647 648 649 650 651
			case SWAP_SUCCESS:
				; /* try to free the page below */
			}
		}

		if (PageDirty(page)) {
A
Andy Whitcroft 已提交
652
			if (sc->order <= PAGE_ALLOC_COSTLY_ORDER && referenced)
L
Linus Torvalds 已提交
653
				goto keep_locked;
654
			if (!may_enter_fs)
L
Linus Torvalds 已提交
655
				goto keep_locked;
656
			if (!sc->may_writepage)
L
Linus Torvalds 已提交
657 658 659
				goto keep_locked;

			/* Page is dirty, try to write it out here */
660
			switch (pageout(page, mapping, sync_writeback)) {
L
Linus Torvalds 已提交
661 662 663 664 665
			case PAGE_KEEP:
				goto keep_locked;
			case PAGE_ACTIVATE:
				goto activate_locked;
			case PAGE_SUCCESS:
666
				if (PageWriteback(page) || PageDirty(page))
L
Linus Torvalds 已提交
667 668 669 670 671
					goto keep;
				/*
				 * A synchronous write - probably a ramdisk.  Go
				 * ahead and try to reclaim the page.
				 */
N
Nick Piggin 已提交
672
				if (!trylock_page(page))
L
Linus Torvalds 已提交
673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691
					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
692
		 * will do this, as well as the blockdev mapping.
L
Linus Torvalds 已提交
693 694 695 696 697 698 699 700 701 702 703 704 705
		 * 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.
		 */
		if (PagePrivate(page)) {
			if (!try_to_release_page(page, sc->gfp_mask))
				goto activate_locked;
N
Nick Piggin 已提交
706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721
			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
722 723
		}

N
Nick Piggin 已提交
724
		if (!mapping || !__remove_mapping(mapping, page))
725
			goto keep_locked;
L
Linus Torvalds 已提交
726

N
Nick Piggin 已提交
727 728 729 730 731 732 733 734
		/*
		 * 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
735
free_it:
736
		nr_reclaimed++;
N
Nick Piggin 已提交
737 738 739 740
		if (!pagevec_add(&freed_pvec, page)) {
			__pagevec_free(&freed_pvec);
			pagevec_reinit(&freed_pvec);
		}
L
Linus Torvalds 已提交
741 742
		continue;

N
Nick Piggin 已提交
743
cull_mlocked:
744 745
		if (PageSwapCache(page))
			try_to_free_swap(page);
N
Nick Piggin 已提交
746 747 748 749
		unlock_page(page);
		putback_lru_page(page);
		continue;

L
Linus Torvalds 已提交
750
activate_locked:
751 752
		/* Not a candidate for swapping, so reclaim swap space. */
		if (PageSwapCache(page) && vm_swap_full())
753
			try_to_free_swap(page);
L
Lee Schermerhorn 已提交
754
		VM_BUG_ON(PageActive(page));
L
Linus Torvalds 已提交
755 756 757 758 759 760
		SetPageActive(page);
		pgactivate++;
keep_locked:
		unlock_page(page);
keep:
		list_add(&page->lru, &ret_pages);
N
Nick Piggin 已提交
761
		VM_BUG_ON(PageLRU(page) || PageUnevictable(page));
L
Linus Torvalds 已提交
762 763 764
	}
	list_splice(&ret_pages, page_list);
	if (pagevec_count(&freed_pvec))
N
Nick Piggin 已提交
765
		__pagevec_free(&freed_pvec);
766
	count_vm_events(PGACTIVATE, pgactivate);
767
	return nr_reclaimed;
L
Linus Torvalds 已提交
768 769
}

A
Andy Whitcroft 已提交
770 771 772 773 774 775 776 777 778 779 780 781 782 783 784
/* LRU Isolation modes. */
#define ISOLATE_INACTIVE 0	/* Isolate inactive pages. */
#define ISOLATE_ACTIVE 1	/* Isolate active pages. */
#define ISOLATE_BOTH 2		/* Isolate both active and inactive pages. */

/*
 * 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.
 */
785
int __isolate_lru_page(struct page *page, int mode, int file)
A
Andy Whitcroft 已提交
786 787 788 789 790 791 792 793 794 795 796 797 798 799 800
{
	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;

801 802 803
	if (mode != ISOLATE_BOTH && (!page_is_file_cache(page) != !file))
		return ret;

L
Lee Schermerhorn 已提交
804 805 806 807 808 809 810 811
	/*
	 * 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
812 813 814 815 816 817 818 819 820 821 822 823 824 825
	ret = -EBUSY;
	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
826 827 828 829 830 831 832 833 834 835 836 837 838 839
/*
 * 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
840 841
 * @order:	The caller's attempted allocation order
 * @mode:	One of the LRU isolation modes
842
 * @file:	True [1] if isolating file [!anon] pages
L
Linus Torvalds 已提交
843 844 845
 *
 * returns how many pages were moved onto *@dst.
 */
846 847
static unsigned long isolate_lru_pages(unsigned long nr_to_scan,
		struct list_head *src, struct list_head *dst,
848
		unsigned long *scanned, int order, int mode, int file)
L
Linus Torvalds 已提交
849
{
850
	unsigned long nr_taken = 0;
851
	unsigned long scan;
L
Linus Torvalds 已提交
852

853
	for (scan = 0; scan < nr_to_scan && !list_empty(src); scan++) {
A
Andy Whitcroft 已提交
854 855 856 857 858 859
		struct page *page;
		unsigned long pfn;
		unsigned long end_pfn;
		unsigned long page_pfn;
		int zone_id;

L
Linus Torvalds 已提交
860 861 862
		page = lru_to_page(src);
		prefetchw_prev_lru_page(page, src, flags);

N
Nick Piggin 已提交
863
		VM_BUG_ON(!PageLRU(page));
N
Nick Piggin 已提交
864

865
		switch (__isolate_lru_page(page, mode, file)) {
A
Andy Whitcroft 已提交
866 867
		case 0:
			list_move(&page->lru, dst);
868
			nr_taken++;
A
Andy Whitcroft 已提交
869 870 871 872 873 874
			break;

		case -EBUSY:
			/* else it is being freed elsewhere */
			list_move(&page->lru, src);
			continue;
875

A
Andy Whitcroft 已提交
876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907
		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);
908

A
Andy Whitcroft 已提交
909 910 911
			/* Check that we have not crossed a zone boundary. */
			if (unlikely(page_zone_id(cursor_page) != zone_id))
				continue;
912
			switch (__isolate_lru_page(cursor_page, mode, file)) {
A
Andy Whitcroft 已提交
913 914 915 916 917 918 919 920 921 922
			case 0:
				list_move(&cursor_page->lru, dst);
				nr_taken++;
				scan++;
				break;

			case -EBUSY:
				/* else it is being freed elsewhere */
				list_move(&cursor_page->lru, src);
			default:
L
Lee Schermerhorn 已提交
923
				break;	/* ! on LRU or wrong list */
A
Andy Whitcroft 已提交
924 925
			}
		}
L
Linus Torvalds 已提交
926 927 928 929 930 931
	}

	*scanned = scan;
	return nr_taken;
}

932 933 934 935 936
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,
937
					int active, int file)
938
{
939
	int lru = LRU_BASE;
940
	if (active)
941 942 943 944 945
		lru += LRU_ACTIVE;
	if (file)
		lru += LRU_FILE;
	return isolate_lru_pages(nr, &z->lru[lru].list, dst, scanned, order,
								mode, !!file);
946 947
}

A
Andy Whitcroft 已提交
948 949 950 951
/*
 * clear_active_flags() is a helper for shrink_active_list(), clearing
 * any active bits from the pages in the list.
 */
952 953
static unsigned long clear_active_flags(struct list_head *page_list,
					unsigned int *count)
A
Andy Whitcroft 已提交
954 955
{
	int nr_active = 0;
956
	int lru;
A
Andy Whitcroft 已提交
957 958
	struct page *page;

959 960
	list_for_each_entry(page, page_list, lru) {
		lru = page_is_file_cache(page);
A
Andy Whitcroft 已提交
961
		if (PageActive(page)) {
962
			lru += LRU_ACTIVE;
A
Andy Whitcroft 已提交
963 964 965
			ClearPageActive(page);
			nr_active++;
		}
966 967
		count[lru]++;
	}
A
Andy Whitcroft 已提交
968 969 970 971

	return nr_active;
}

972 973 974 975 976 977 978 979 980 981 982
/**
 * 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
983 984 985
 * 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.
986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005
 *
 * 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
1006
			int lru = page_lru(page);
1007 1008
			ret = 0;
			ClearPageLRU(page);
1009 1010

			del_page_from_lru_list(zone, page, lru);
1011 1012 1013 1014 1015 1016
		}
		spin_unlock_irq(&zone->lru_lock);
	}
	return ret;
}

L
Linus Torvalds 已提交
1017
/*
A
Andrew Morton 已提交
1018 1019
 * shrink_inactive_list() is a helper for shrink_zone().  It returns the number
 * of reclaimed pages
L
Linus Torvalds 已提交
1020
 */
A
Andrew Morton 已提交
1021
static unsigned long shrink_inactive_list(unsigned long max_scan,
R
Rik van Riel 已提交
1022 1023
			struct zone *zone, struct scan_control *sc,
			int priority, int file)
L
Linus Torvalds 已提交
1024 1025 1026
{
	LIST_HEAD(page_list);
	struct pagevec pvec;
1027
	unsigned long nr_scanned = 0;
1028
	unsigned long nr_reclaimed = 0;
L
Linus Torvalds 已提交
1029 1030 1031 1032 1033

	pagevec_init(&pvec, 1);

	lru_add_drain();
	spin_lock_irq(&zone->lru_lock);
1034
	do {
L
Linus Torvalds 已提交
1035
		struct page *page;
1036 1037 1038
		unsigned long nr_taken;
		unsigned long nr_scan;
		unsigned long nr_freed;
A
Andy Whitcroft 已提交
1039
		unsigned long nr_active;
1040
		unsigned int count[NR_LRU_LISTS] = { 0, };
R
Rik van Riel 已提交
1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053
		int mode = ISOLATE_INACTIVE;

		/*
		 * 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)
			mode = ISOLATE_BOTH;
		else if (sc->order && priority < DEF_PRIORITY - 2)
			mode = ISOLATE_BOTH;
L
Linus Torvalds 已提交
1054

1055
		nr_taken = sc->isolate_pages(sc->swap_cluster_max,
1056 1057 1058
			     &page_list, &nr_scan, sc->order, mode,
				zone, sc->mem_cgroup, 0, file);
		nr_active = clear_active_flags(&page_list, count);
1059
		__count_vm_events(PGDEACTIVATE, nr_active);
A
Andy Whitcroft 已提交
1060

1061 1062 1063 1064 1065 1066 1067 1068 1069 1070
		__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]);

		if (scan_global_lru(sc)) {
1071
			zone->pages_scanned += nr_scan;
1072 1073 1074 1075 1076
			zone->recent_scanned[0] += count[LRU_INACTIVE_ANON];
			zone->recent_scanned[0] += count[LRU_ACTIVE_ANON];
			zone->recent_scanned[1] += count[LRU_INACTIVE_FILE];
			zone->recent_scanned[1] += count[LRU_ACTIVE_FILE];
		}
L
Linus Torvalds 已提交
1077 1078
		spin_unlock_irq(&zone->lru_lock);

1079
		nr_scanned += nr_scan;
1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095
		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() &&
					sc->order > PAGE_ALLOC_COSTLY_ORDER) {
			congestion_wait(WRITE, HZ/10);

			/*
			 * The attempt at page out may have made some
			 * of the pages active, mark them inactive again.
			 */
1096
			nr_active = clear_active_flags(&page_list, count);
1097 1098 1099 1100 1101 1102
			count_vm_events(PGDEACTIVATE, nr_active);

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

1103
		nr_reclaimed += nr_freed;
N
Nick Piggin 已提交
1104 1105
		local_irq_disable();
		if (current_is_kswapd()) {
1106 1107
			__count_zone_vm_events(PGSCAN_KSWAPD, zone, nr_scan);
			__count_vm_events(KSWAPD_STEAL, nr_freed);
1108
		} else if (scan_global_lru(sc))
1109
			__count_zone_vm_events(PGSCAN_DIRECT, zone, nr_scan);
1110

S
Shantanu Goel 已提交
1111
		__count_zone_vm_events(PGSTEAL, zone, nr_freed);
N
Nick Piggin 已提交
1112

1113 1114 1115
		if (nr_taken == 0)
			goto done;

N
Nick Piggin 已提交
1116
		spin_lock(&zone->lru_lock);
L
Linus Torvalds 已提交
1117 1118 1119 1120
		/*
		 * Put back any unfreeable pages.
		 */
		while (!list_empty(&page_list)) {
L
Lee Schermerhorn 已提交
1121
			int lru;
L
Linus Torvalds 已提交
1122
			page = lru_to_page(&page_list);
N
Nick Piggin 已提交
1123
			VM_BUG_ON(PageLRU(page));
L
Linus Torvalds 已提交
1124
			list_del(&page->lru);
L
Lee Schermerhorn 已提交
1125 1126 1127 1128 1129 1130 1131 1132 1133 1134
			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);
			mem_cgroup_move_lists(page, lru);
1135 1136 1137 1138
			if (PageActive(page) && scan_global_lru(sc)) {
				int file = !!page_is_file_cache(page);
				zone->recent_rotated[file]++;
			}
L
Linus Torvalds 已提交
1139 1140 1141 1142 1143 1144
			if (!pagevec_add(&pvec, page)) {
				spin_unlock_irq(&zone->lru_lock);
				__pagevec_release(&pvec);
				spin_lock_irq(&zone->lru_lock);
			}
		}
1145
  	} while (nr_scanned < max_scan);
1146
	spin_unlock(&zone->lru_lock);
L
Linus Torvalds 已提交
1147
done:
1148
	local_irq_enable();
L
Linus Torvalds 已提交
1149
	pagevec_release(&pvec);
1150
	return nr_reclaimed;
L
Linus Torvalds 已提交
1151 1152
}

1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166
/*
 * 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;
}

N
Nick Piggin 已提交
1167 1168
static inline int zone_is_near_oom(struct zone *zone)
{
1169
	return zone->pages_scanned >= (zone_lru_pages(zone) * 3);
1170 1171
}

L
Linus Torvalds 已提交
1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188
/*
 * 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.
 */
1189 1190


A
Andrew Morton 已提交
1191
static void shrink_active_list(unsigned long nr_pages, struct zone *zone,
1192
			struct scan_control *sc, int priority, int file)
L
Linus Torvalds 已提交
1193
{
1194
	unsigned long pgmoved;
L
Linus Torvalds 已提交
1195
	int pgdeactivate = 0;
1196
	unsigned long pgscanned;
L
Linus Torvalds 已提交
1197
	LIST_HEAD(l_hold);	/* The pages which were snipped off */
1198
	LIST_HEAD(l_inactive);
L
Linus Torvalds 已提交
1199 1200
	struct page *page;
	struct pagevec pvec;
1201
	enum lru_list lru;
L
Linus Torvalds 已提交
1202 1203 1204

	lru_add_drain();
	spin_lock_irq(&zone->lru_lock);
1205 1206
	pgmoved = sc->isolate_pages(nr_pages, &l_hold, &pgscanned, sc->order,
					ISOLATE_ACTIVE, zone,
1207
					sc->mem_cgroup, 1, file);
1208 1209 1210 1211
	/*
	 * zone->pages_scanned is used for detect zone's oom
	 * mem_cgroup remembers nr_scan by itself.
	 */
1212
	if (scan_global_lru(sc)) {
1213
		zone->pages_scanned += pgscanned;
1214 1215
		zone->recent_scanned[!!file] += pgmoved;
	}
1216

1217 1218 1219 1220
	if (file)
		__mod_zone_page_state(zone, NR_ACTIVE_FILE, -pgmoved);
	else
		__mod_zone_page_state(zone, NR_ACTIVE_ANON, -pgmoved);
L
Linus Torvalds 已提交
1221 1222
	spin_unlock_irq(&zone->lru_lock);

1223
	pgmoved = 0;
L
Linus Torvalds 已提交
1224 1225 1226 1227
	while (!list_empty(&l_hold)) {
		cond_resched();
		page = lru_to_page(&l_hold);
		list_del(&page->lru);
1228

L
Lee Schermerhorn 已提交
1229 1230 1231 1232 1233
		if (unlikely(!page_evictable(page, NULL))) {
			putback_lru_page(page);
			continue;
		}

1234 1235 1236 1237 1238
		/* page_referenced clears PageReferenced */
		if (page_mapping_inuse(page) &&
		    page_referenced(page, 0, sc->mem_cgroup))
			pgmoved++;

L
Linus Torvalds 已提交
1239 1240 1241
		list_add(&page->lru, &l_inactive);
	}

1242
	spin_lock_irq(&zone->lru_lock);
1243
	/*
1244 1245 1246 1247 1248
	 * Count referenced pages from currently used mappings as
	 * rotated, even though they are moved to the inactive list.
	 * This helps balance scan pressure between file and anonymous
	 * pages in get_scan_ratio.
	 */
1249 1250
	if (scan_global_lru(sc))
		zone->recent_rotated[!!file] += pgmoved;
1251

1252
	/*
1253
	 * Move the pages to the [file or anon] inactive list.
1254
	 */
L
Linus Torvalds 已提交
1255
	pagevec_init(&pvec, 1);
1256

L
Linus Torvalds 已提交
1257
	pgmoved = 0;
1258
	lru = LRU_BASE + file * LRU_FILE;
L
Linus Torvalds 已提交
1259 1260 1261
	while (!list_empty(&l_inactive)) {
		page = lru_to_page(&l_inactive);
		prefetchw_prev_lru_page(page, &l_inactive, flags);
N
Nick Piggin 已提交
1262
		VM_BUG_ON(PageLRU(page));
N
Nick Piggin 已提交
1263
		SetPageLRU(page);
N
Nick Piggin 已提交
1264
		VM_BUG_ON(!PageActive(page));
N
Nick Piggin 已提交
1265 1266
		ClearPageActive(page);

1267
		list_move(&page->lru, &zone->lru[lru].list);
L
Lee Schermerhorn 已提交
1268
		mem_cgroup_move_lists(page, lru);
L
Linus Torvalds 已提交
1269 1270
		pgmoved++;
		if (!pagevec_add(&pvec, page)) {
1271
			__mod_zone_page_state(zone, NR_LRU_BASE + lru, pgmoved);
L
Linus Torvalds 已提交
1272 1273 1274 1275 1276 1277 1278 1279 1280
			spin_unlock_irq(&zone->lru_lock);
			pgdeactivate += pgmoved;
			pgmoved = 0;
			if (buffer_heads_over_limit)
				pagevec_strip(&pvec);
			__pagevec_release(&pvec);
			spin_lock_irq(&zone->lru_lock);
		}
	}
1281
	__mod_zone_page_state(zone, NR_LRU_BASE + lru, pgmoved);
L
Linus Torvalds 已提交
1282 1283 1284 1285 1286 1287
	pgdeactivate += pgmoved;
	if (buffer_heads_over_limit) {
		spin_unlock_irq(&zone->lru_lock);
		pagevec_strip(&pvec);
		spin_lock_irq(&zone->lru_lock);
	}
1288 1289 1290
	__count_zone_vm_events(PGREFILL, zone, pgscanned);
	__count_vm_events(PGDEACTIVATE, pgdeactivate);
	spin_unlock_irq(&zone->lru_lock);
1291 1292
	if (vm_swap_full())
		pagevec_swap_free(&pvec);
L
Linus Torvalds 已提交
1293

N
Nick Piggin 已提交
1294
	pagevec_release(&pvec);
L
Linus Torvalds 已提交
1295 1296
}

1297
static unsigned long shrink_list(enum lru_list lru, unsigned long nr_to_scan,
1298 1299
	struct zone *zone, struct scan_control *sc, int priority)
{
1300 1301
	int file = is_file_lru(lru);

1302 1303 1304 1305 1306 1307 1308
	if (lru == LRU_ACTIVE_FILE) {
		shrink_active_list(nr_to_scan, zone, sc, priority, file);
		return 0;
	}

	if (lru == LRU_ACTIVE_ANON &&
	    (!scan_global_lru(sc) || inactive_anon_is_low(zone))) {
1309
		shrink_active_list(nr_to_scan, zone, sc, priority, file);
1310 1311
		return 0;
	}
R
Rik van Riel 已提交
1312
	return shrink_inactive_list(nr_to_scan, zone, sc, priority, file);
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
}

/*
 * 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.
 *
 * percent[0] specifies how much pressure to put on ram/swap backed
 * memory, while percent[1] determines pressure on the file LRUs.
 */
static void get_scan_ratio(struct zone *zone, struct scan_control *sc,
					unsigned long *percent)
{
	unsigned long anon, file, free;
	unsigned long anon_prio, file_prio;
	unsigned long ap, fp;

	/* If we have no swap space, do not bother scanning anon pages. */
	if (nr_swap_pages <= 0) {
		percent[0] = 0;
		percent[1] = 100;
		return;
	}

1338 1339 1340 1341 1342 1343
	anon  = zone_page_state(zone, NR_ACTIVE_ANON) +
		zone_page_state(zone, NR_INACTIVE_ANON);
	file  = zone_page_state(zone, NR_ACTIVE_FILE) +
		zone_page_state(zone, NR_INACTIVE_FILE);
	free  = zone_page_state(zone, NR_FREE_PAGES);

1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383
	/* If we have very few page cache pages, force-scan anon pages. */
	if (unlikely(file + free <= zone->pages_high)) {
		percent[0] = 100;
		percent[1] = 0;
		return;
	}

	/*
	 * 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]
	 */
	if (unlikely(zone->recent_scanned[0] > anon / 4)) {
		spin_lock_irq(&zone->lru_lock);
		zone->recent_scanned[0] /= 2;
		zone->recent_rotated[0] /= 2;
		spin_unlock_irq(&zone->lru_lock);
	}

	if (unlikely(zone->recent_scanned[1] > file / 4)) {
		spin_lock_irq(&zone->lru_lock);
		zone->recent_scanned[1] /= 2;
		zone->recent_rotated[1] /= 2;
		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;

	/*
1384 1385 1386
	 * 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.
1387 1388 1389 1390 1391 1392 1393 1394 1395 1396
	 */
	ap = (anon_prio + 1) * (zone->recent_scanned[0] + 1);
	ap /= zone->recent_rotated[0] + 1;

	fp = (file_prio + 1) * (zone->recent_scanned[1] + 1);
	fp /= zone->recent_rotated[1] + 1;

	/* Normalize to percentages */
	percent[0] = 100 * ap / (ap + fp + 1);
	percent[1] = 100 - percent[0];
1397 1398
}

1399

L
Linus Torvalds 已提交
1400 1401 1402
/*
 * This is a basic per-zone page freer.  Used by both kswapd and direct reclaim.
 */
1403 1404
static unsigned long shrink_zone(int priority, struct zone *zone,
				struct scan_control *sc)
L
Linus Torvalds 已提交
1405
{
1406
	unsigned long nr[NR_LRU_LISTS];
1407
	unsigned long nr_to_scan;
1408
	unsigned long nr_reclaimed = 0;
1409
	unsigned long percent[2];	/* anon @ 0; file @ 1 */
1410
	enum lru_list l;
L
Linus Torvalds 已提交
1411

1412 1413
	get_scan_ratio(zone, sc, percent);

L
Lee Schermerhorn 已提交
1414
	for_each_evictable_lru(l) {
1415 1416 1417
		if (scan_global_lru(sc)) {
			int file = is_file_lru(l);
			int scan;
1418

1419 1420 1421 1422 1423
			scan = zone_page_state(zone, NR_LRU_BASE + l);
			if (priority) {
				scan >>= priority;
				scan = (scan * percent[file]) / 100;
			}
1424
			zone->lru[l].nr_scan += scan;
1425 1426 1427 1428 1429
			nr[l] = zone->lru[l].nr_scan;
			if (nr[l] >= sc->swap_cluster_max)
				zone->lru[l].nr_scan = 0;
			else
				nr[l] = 0;
1430 1431 1432 1433 1434 1435 1436 1437
		} else {
			/*
			 * This reclaim occurs not because zone memory shortage
			 * but because memory controller hits its limit.
			 * Don't modify zone reclaim related data.
			 */
			nr[l] = mem_cgroup_calc_reclaim(sc->mem_cgroup, zone,
								priority, l);
1438
		}
1439
	}
L
Linus Torvalds 已提交
1440

1441 1442
	while (nr[LRU_INACTIVE_ANON] || nr[LRU_ACTIVE_FILE] ||
					nr[LRU_INACTIVE_FILE]) {
L
Lee Schermerhorn 已提交
1443
		for_each_evictable_lru(l) {
1444 1445
			if (nr[l]) {
				nr_to_scan = min(nr[l],
L
Linus Torvalds 已提交
1446
					(unsigned long)sc->swap_cluster_max);
1447
				nr[l] -= nr_to_scan;
L
Linus Torvalds 已提交
1448

1449 1450 1451
				nr_reclaimed += shrink_list(l, nr_to_scan,
							zone, sc, priority);
			}
L
Linus Torvalds 已提交
1452 1453 1454
		}
	}

1455 1456 1457 1458 1459 1460 1461 1462 1463
	/*
	 * Even if we did not try to evict anon pages at all, we want to
	 * rebalance the anon lru active/inactive ratio.
	 */
	if (!scan_global_lru(sc) || inactive_anon_is_low(zone))
		shrink_active_list(SWAP_CLUSTER_MAX, zone, sc, priority, 0);
	else if (!scan_global_lru(sc))
		shrink_active_list(SWAP_CLUSTER_MAX, zone, sc, priority, 0);

1464
	throttle_vm_writeout(sc->gfp_mask);
1465
	return nr_reclaimed;
L
Linus Torvalds 已提交
1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483
}

/*
 * 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.
 *
 * We reclaim from a zone even if that zone is over pages_high.  Because:
 * a) The caller may be trying to free *extra* pages to satisfy a higher-order
 *    allocation or
 * b) The zones may be over pages_high but they must go *over* pages_high to
 *    satisfy the `incremental min' zone defense algorithm.
 *
 * Returns the number of reclaimed pages.
 *
 * If a zone is deemed to be full of pinned pages then just give it a light
 * scan then give up on it.
 */
1484
static unsigned long shrink_zones(int priority, struct zonelist *zonelist,
1485
					struct scan_control *sc)
L
Linus Torvalds 已提交
1486
{
1487
	enum zone_type high_zoneidx = gfp_zone(sc->gfp_mask);
1488
	unsigned long nr_reclaimed = 0;
1489
	struct zoneref *z;
1490
	struct zone *zone;
1491

1492
	sc->all_unreclaimable = 1;
1493
	for_each_zone_zonelist(zone, z, zonelist, high_zoneidx) {
1494
		if (!populated_zone(zone))
L
Linus Torvalds 已提交
1495
			continue;
1496 1497 1498 1499 1500 1501 1502 1503
		/*
		 * Take care memory controller reclaiming has small influence
		 * to global LRU.
		 */
		if (scan_global_lru(sc)) {
			if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL))
				continue;
			note_zone_scanning_priority(zone, priority);
L
Linus Torvalds 已提交
1504

1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517
			if (zone_is_all_unreclaimable(zone) &&
						priority != DEF_PRIORITY)
				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);
		}
1518

1519
		nr_reclaimed += shrink_zone(priority, zone, sc);
L
Linus Torvalds 已提交
1520
	}
1521

1522
	return nr_reclaimed;
L
Linus Torvalds 已提交
1523
}
1524

L
Linus Torvalds 已提交
1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536
/*
 * 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
 * caller can't do much about.  We kick pdflush 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.
1537 1538 1539
 *
 * returns:	0, if no pages reclaimed
 * 		else, the number of pages reclaimed
L
Linus Torvalds 已提交
1540
 */
1541
static unsigned long do_try_to_free_pages(struct zonelist *zonelist,
1542
					struct scan_control *sc)
L
Linus Torvalds 已提交
1543 1544
{
	int priority;
1545
	unsigned long ret = 0;
1546
	unsigned long total_scanned = 0;
1547
	unsigned long nr_reclaimed = 0;
L
Linus Torvalds 已提交
1548 1549
	struct reclaim_state *reclaim_state = current->reclaim_state;
	unsigned long lru_pages = 0;
1550
	struct zoneref *z;
1551
	struct zone *zone;
1552
	enum zone_type high_zoneidx = gfp_zone(sc->gfp_mask);
L
Linus Torvalds 已提交
1553

1554 1555
	delayacct_freepages_start();

1556 1557 1558 1559 1560 1561
	if (scan_global_lru(sc))
		count_vm_event(ALLOCSTALL);
	/*
	 * mem_cgroup will not do shrink_slab.
	 */
	if (scan_global_lru(sc)) {
1562
		for_each_zone_zonelist(zone, z, zonelist, high_zoneidx) {
L
Linus Torvalds 已提交
1563

1564 1565
			if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL))
				continue;
L
Linus Torvalds 已提交
1566

1567
			lru_pages += zone_lru_pages(zone);
1568
		}
L
Linus Torvalds 已提交
1569 1570 1571
	}

	for (priority = DEF_PRIORITY; priority >= 0; priority--) {
1572
		sc->nr_scanned = 0;
1573 1574
		if (!priority)
			disable_swap_token();
1575
		nr_reclaimed += shrink_zones(priority, zonelist, sc);
1576 1577 1578 1579
		/*
		 * Don't shrink slabs when reclaiming memory from
		 * over limit cgroups
		 */
1580
		if (scan_global_lru(sc)) {
1581
			shrink_slab(sc->nr_scanned, sc->gfp_mask, lru_pages);
1582 1583 1584 1585
			if (reclaim_state) {
				nr_reclaimed += reclaim_state->reclaimed_slab;
				reclaim_state->reclaimed_slab = 0;
			}
L
Linus Torvalds 已提交
1586
		}
1587 1588
		total_scanned += sc->nr_scanned;
		if (nr_reclaimed >= sc->swap_cluster_max) {
1589
			ret = nr_reclaimed;
L
Linus Torvalds 已提交
1590 1591 1592 1593 1594 1595 1596 1597 1598 1599
			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.
		 */
1600 1601
		if (total_scanned > sc->swap_cluster_max +
					sc->swap_cluster_max / 2) {
1602
			wakeup_pdflush(laptop_mode ? 0 : total_scanned);
1603
			sc->may_writepage = 1;
L
Linus Torvalds 已提交
1604 1605 1606
		}

		/* Take a nap, wait for some writeback to complete */
1607
		if (sc->nr_scanned && priority < DEF_PRIORITY - 2)
1608
			congestion_wait(WRITE, HZ/10);
L
Linus Torvalds 已提交
1609
	}
1610
	/* top priority shrink_zones still had more to do? don't OOM, then */
1611
	if (!sc->all_unreclaimable && scan_global_lru(sc))
1612
		ret = nr_reclaimed;
L
Linus Torvalds 已提交
1613
out:
1614 1615 1616 1617 1618 1619 1620 1621 1622
	/*
	 * 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
1623

1624
	if (scan_global_lru(sc)) {
1625
		for_each_zone_zonelist(zone, z, zonelist, high_zoneidx) {
1626 1627 1628 1629 1630 1631 1632 1633

			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
1634

1635 1636
	delayacct_freepages_end();

L
Linus Torvalds 已提交
1637 1638 1639
	return ret;
}

1640 1641
unsigned long try_to_free_pages(struct zonelist *zonelist, int order,
								gfp_t gfp_mask)
1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653
{
	struct scan_control sc = {
		.gfp_mask = gfp_mask,
		.may_writepage = !laptop_mode,
		.swap_cluster_max = SWAP_CLUSTER_MAX,
		.may_swap = 1,
		.swappiness = vm_swappiness,
		.order = order,
		.mem_cgroup = NULL,
		.isolate_pages = isolate_pages_global,
	};

1654
	return do_try_to_free_pages(zonelist, &sc);
1655 1656
}

1657
#ifdef CONFIG_CGROUP_MEM_RES_CTLR
1658

1659 1660
unsigned long try_to_free_mem_cgroup_pages(struct mem_cgroup *mem_cont,
						gfp_t gfp_mask)
1661 1662 1663 1664 1665 1666 1667 1668 1669 1670
{
	struct scan_control sc = {
		.may_writepage = !laptop_mode,
		.may_swap = 1,
		.swap_cluster_max = SWAP_CLUSTER_MAX,
		.swappiness = vm_swappiness,
		.order = 0,
		.mem_cgroup = mem_cont,
		.isolate_pages = mem_cgroup_isolate_pages,
	};
1671
	struct zonelist *zonelist;
1672

1673 1674 1675 1676
	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);
1677 1678 1679
}
#endif

L
Linus Torvalds 已提交
1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700
/*
 * For kswapd, balance_pgdat() will work across all this node's zones until
 * they are all at pages_high.
 *
 * 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
 * zones which have free_pages > pages_high, but once a zone is found to have
 * free_pages <= pages_high, 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.
 */
1701
static unsigned long balance_pgdat(pg_data_t *pgdat, int order)
L
Linus Torvalds 已提交
1702 1703 1704 1705
{
	int all_zones_ok;
	int priority;
	int i;
1706
	unsigned long total_scanned;
1707
	unsigned long nr_reclaimed;
L
Linus Torvalds 已提交
1708
	struct reclaim_state *reclaim_state = current->reclaim_state;
1709 1710 1711
	struct scan_control sc = {
		.gfp_mask = GFP_KERNEL,
		.may_swap = 1,
1712 1713
		.swap_cluster_max = SWAP_CLUSTER_MAX,
		.swappiness = vm_swappiness,
A
Andy Whitcroft 已提交
1714
		.order = order,
1715 1716
		.mem_cgroup = NULL,
		.isolate_pages = isolate_pages_global,
1717
	};
1718 1719 1720 1721 1722
	/*
	 * temp_priority is used to remember the scanning priority at which
	 * this zone was successfully refilled to free_pages == pages_high.
	 */
	int temp_priority[MAX_NR_ZONES];
L
Linus Torvalds 已提交
1723 1724 1725

loop_again:
	total_scanned = 0;
1726
	nr_reclaimed = 0;
C
Christoph Lameter 已提交
1727
	sc.may_writepage = !laptop_mode;
1728
	count_vm_event(PAGEOUTRUN);
L
Linus Torvalds 已提交
1729

1730 1731
	for (i = 0; i < pgdat->nr_zones; i++)
		temp_priority[i] = DEF_PRIORITY;
L
Linus Torvalds 已提交
1732 1733 1734 1735 1736

	for (priority = DEF_PRIORITY; priority >= 0; priority--) {
		int end_zone = 0;	/* Inclusive.  0 = ZONE_DMA */
		unsigned long lru_pages = 0;

1737 1738 1739 1740
		/* The swap token gets in the way of swapout... */
		if (!priority)
			disable_swap_token();

L
Linus Torvalds 已提交
1741 1742
		all_zones_ok = 1;

1743 1744 1745 1746 1747 1748
		/*
		 * 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
1749

1750 1751
			if (!populated_zone(zone))
				continue;
L
Linus Torvalds 已提交
1752

1753 1754
			if (zone_is_all_unreclaimable(zone) &&
			    priority != DEF_PRIORITY)
1755
				continue;
L
Linus Torvalds 已提交
1756

1757 1758 1759 1760 1761 1762 1763 1764
			/*
			 * Do some background aging of the anon list, to give
			 * pages a chance to be referenced before reclaiming.
			 */
			if (inactive_anon_is_low(zone))
				shrink_active_list(SWAP_CLUSTER_MAX, zone,
							&sc, priority, 0);

1765 1766 1767
			if (!zone_watermark_ok(zone, order, zone->pages_high,
					       0, 0)) {
				end_zone = i;
A
Andrew Morton 已提交
1768
				break;
L
Linus Torvalds 已提交
1769 1770
			}
		}
A
Andrew Morton 已提交
1771 1772 1773
		if (i < 0)
			goto out;

L
Linus Torvalds 已提交
1774 1775 1776
		for (i = 0; i <= end_zone; i++) {
			struct zone *zone = pgdat->node_zones + i;

1777
			lru_pages += zone_lru_pages(zone);
L
Linus Torvalds 已提交
1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790
		}

		/*
		 * 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;
1791
			int nr_slab;
L
Linus Torvalds 已提交
1792

1793
			if (!populated_zone(zone))
L
Linus Torvalds 已提交
1794 1795
				continue;

1796 1797
			if (zone_is_all_unreclaimable(zone) &&
					priority != DEF_PRIORITY)
L
Linus Torvalds 已提交
1798 1799
				continue;

1800 1801 1802
			if (!zone_watermark_ok(zone, order, zone->pages_high,
					       end_zone, 0))
				all_zones_ok = 0;
1803
			temp_priority[i] = priority;
L
Linus Torvalds 已提交
1804
			sc.nr_scanned = 0;
1805
			note_zone_scanning_priority(zone, priority);
1806 1807 1808 1809 1810 1811 1812
			/*
			 * We put equal pressure on every zone, unless one
			 * zone has way too many pages free already.
			 */
			if (!zone_watermark_ok(zone, order, 8*zone->pages_high,
						end_zone, 0))
				nr_reclaimed += shrink_zone(priority, zone, &sc);
L
Linus Torvalds 已提交
1813
			reclaim_state->reclaimed_slab = 0;
1814 1815
			nr_slab = shrink_slab(sc.nr_scanned, GFP_KERNEL,
						lru_pages);
1816
			nr_reclaimed += reclaim_state->reclaimed_slab;
L
Linus Torvalds 已提交
1817
			total_scanned += sc.nr_scanned;
1818
			if (zone_is_all_unreclaimable(zone))
L
Linus Torvalds 已提交
1819
				continue;
1820
			if (nr_slab == 0 && zone->pages_scanned >=
1821
						(zone_lru_pages(zone) * 6))
1822 1823
					zone_set_flag(zone,
						      ZONE_ALL_UNRECLAIMABLE);
L
Linus Torvalds 已提交
1824 1825 1826 1827 1828 1829
			/*
			 * 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 &&
1830
			    total_scanned > nr_reclaimed + nr_reclaimed / 2)
L
Linus Torvalds 已提交
1831 1832 1833 1834 1835 1836 1837 1838
				sc.may_writepage = 1;
		}
		if (all_zones_ok)
			break;		/* kswapd: all done */
		/*
		 * OK, kswapd is getting into trouble.  Take a nap, then take
		 * another pass across the zones.
		 */
1839
		if (total_scanned && priority < DEF_PRIORITY - 2)
1840
			congestion_wait(WRITE, HZ/10);
L
Linus Torvalds 已提交
1841 1842 1843 1844 1845 1846 1847

		/*
		 * 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.
		 */
1848
		if (nr_reclaimed >= SWAP_CLUSTER_MAX)
L
Linus Torvalds 已提交
1849 1850 1851
			break;
	}
out:
1852 1853 1854 1855 1856
	/*
	 * 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
1857 1858 1859
	for (i = 0; i < pgdat->nr_zones; i++) {
		struct zone *zone = pgdat->node_zones + i;

1860
		zone->prev_priority = temp_priority[i];
L
Linus Torvalds 已提交
1861 1862 1863
	}
	if (!all_zones_ok) {
		cond_resched();
1864 1865 1866

		try_to_freeze();

L
Linus Torvalds 已提交
1867 1868 1869
		goto loop_again;
	}

1870
	return nr_reclaimed;
L
Linus Torvalds 已提交
1871 1872 1873 1874
}

/*
 * The background pageout daemon, started as a kernel thread
1875
 * from the init process.
L
Linus Torvalds 已提交
1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894
 *
 * 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,
	};
1895
	node_to_cpumask_ptr(cpumask, pgdat->node_id);
L
Linus Torvalds 已提交
1896

R
Rusty Russell 已提交
1897
	if (!cpumask_empty(cpumask))
1898
		set_cpus_allowed_ptr(tsk, cpumask);
L
Linus Torvalds 已提交
1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912
	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).
	 */
1913
	tsk->flags |= PF_MEMALLOC | PF_SWAPWRITE | PF_KSWAPD;
1914
	set_freezable();
L
Linus Torvalds 已提交
1915 1916 1917 1918

	order = 0;
	for ( ; ; ) {
		unsigned long new_order;
1919

L
Linus Torvalds 已提交
1920 1921 1922 1923 1924 1925 1926 1927 1928 1929
		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 {
1930 1931 1932
			if (!freezing(current))
				schedule();

L
Linus Torvalds 已提交
1933 1934 1935 1936
			order = pgdat->kswapd_max_order;
		}
		finish_wait(&pgdat->kswapd_wait, &wait);

1937 1938 1939 1940 1941 1942
		if (!try_to_freeze()) {
			/* We can speed up thawing tasks if we don't call
			 * balance_pgdat after returning from the refrigerator
			 */
			balance_pgdat(pgdat, order);
		}
L
Linus Torvalds 已提交
1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953
	}
	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;

1954
	if (!populated_zone(zone))
L
Linus Torvalds 已提交
1955 1956 1957
		return;

	pgdat = zone->zone_pgdat;
R
Rohit Seth 已提交
1958
	if (zone_watermark_ok(zone, order, zone->pages_low, 0, 0))
L
Linus Torvalds 已提交
1959 1960 1961
		return;
	if (pgdat->kswapd_max_order < order)
		pgdat->kswapd_max_order = order;
1962
	if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL))
L
Linus Torvalds 已提交
1963
		return;
1964
	if (!waitqueue_active(&pgdat->kswapd_wait))
L
Linus Torvalds 已提交
1965
		return;
1966
	wake_up_interruptible(&pgdat->kswapd_wait);
L
Linus Torvalds 已提交
1967 1968
}

1969 1970 1971 1972 1973 1974 1975 1976
unsigned long global_lru_pages(void)
{
	return global_page_state(NR_ACTIVE_ANON)
		+ global_page_state(NR_ACTIVE_FILE)
		+ global_page_state(NR_INACTIVE_ANON)
		+ global_page_state(NR_INACTIVE_FILE);
}

L
Linus Torvalds 已提交
1977 1978
#ifdef CONFIG_PM
/*
1979 1980 1981 1982 1983 1984
 * Helper function for shrink_all_memory().  Tries to reclaim 'nr_pages' pages
 * from LRU lists system-wide, for given pass and priority, and returns the
 * number of reclaimed pages
 *
 * For pass > 3 we also try to shrink the LRU lists that contain a few pages
 */
1985 1986
static unsigned long shrink_all_zones(unsigned long nr_pages, int prio,
				      int pass, struct scan_control *sc)
1987 1988 1989
{
	struct zone *zone;
	unsigned long nr_to_scan, ret = 0;
1990
	enum lru_list l;
1991 1992 1993 1994 1995 1996

	for_each_zone(zone) {

		if (!populated_zone(zone))
			continue;

1997
		if (zone_is_all_unreclaimable(zone) && prio != DEF_PRIORITY)
1998 1999
			continue;

L
Lee Schermerhorn 已提交
2000 2001
		for_each_evictable_lru(l) {
			/* For pass = 0, we don't shrink the active list */
2002 2003
			if (pass == 0 &&
				(l == LRU_ACTIVE || l == LRU_ACTIVE_FILE))
2004 2005 2006 2007 2008 2009 2010
				continue;

			zone->lru[l].nr_scan +=
				(zone_page_state(zone, NR_LRU_BASE + l)
								>> prio) + 1;
			if (zone->lru[l].nr_scan >= nr_pages || pass > 3) {
				zone->lru[l].nr_scan = 0;
2011
				nr_to_scan = min(nr_pages,
2012 2013 2014 2015 2016 2017
					zone_page_state(zone,
							NR_LRU_BASE + l));
				ret += shrink_list(l, nr_to_scan, zone,
								sc, prio);
				if (ret >= nr_pages)
					return ret;
2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031
			}
		}
	}

	return ret;
}

/*
 * Try to free `nr_pages' of memory, system-wide, and return the number of
 * 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
2032
 */
2033
unsigned long shrink_all_memory(unsigned long nr_pages)
L
Linus Torvalds 已提交
2034
{
2035
	unsigned long lru_pages, nr_slab;
2036
	unsigned long ret = 0;
2037 2038 2039 2040 2041 2042 2043 2044
	int pass;
	struct reclaim_state reclaim_state;
	struct scan_control sc = {
		.gfp_mask = GFP_KERNEL,
		.may_swap = 0,
		.swap_cluster_max = nr_pages,
		.may_writepage = 1,
		.swappiness = vm_swappiness,
2045
		.isolate_pages = isolate_pages_global,
L
Linus Torvalds 已提交
2046 2047 2048
	};

	current->reclaim_state = &reclaim_state;
2049

2050
	lru_pages = global_lru_pages();
2051
	nr_slab = global_page_state(NR_SLAB_RECLAIMABLE);
2052 2053 2054 2055 2056
	/* If slab caches are huge, it's better to hit them first */
	while (nr_slab >= lru_pages) {
		reclaim_state.reclaimed_slab = 0;
		shrink_slab(nr_pages, sc.gfp_mask, lru_pages);
		if (!reclaim_state.reclaimed_slab)
L
Linus Torvalds 已提交
2057
			break;
2058 2059 2060 2061 2062 2063

		ret += reclaim_state.reclaimed_slab;
		if (ret >= nr_pages)
			goto out;

		nr_slab -= reclaim_state.reclaimed_slab;
L
Linus Torvalds 已提交
2064
	}
2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091

	/*
	 * We try to shrink LRUs in 5 passes:
	 * 0 = Reclaim from inactive_list only
	 * 1 = Reclaim from active list but don't reclaim mapped
	 * 2 = 2nd pass of type 1
	 * 3 = Reclaim mapped (normal reclaim)
	 * 4 = 2nd pass of type 3
	 */
	for (pass = 0; pass < 5; pass++) {
		int prio;

		/* Force reclaiming mapped pages in the passes #3 and #4 */
		if (pass > 2) {
			sc.may_swap = 1;
			sc.swappiness = 100;
		}

		for (prio = DEF_PRIORITY; prio >= 0; prio--) {
			unsigned long nr_to_scan = nr_pages - ret;

			sc.nr_scanned = 0;
			ret += shrink_all_zones(nr_to_scan, prio, pass, &sc);
			if (ret >= nr_pages)
				goto out;

			reclaim_state.reclaimed_slab = 0;
2092
			shrink_slab(sc.nr_scanned, sc.gfp_mask,
2093
					global_lru_pages());
2094 2095 2096 2097 2098
			ret += reclaim_state.reclaimed_slab;
			if (ret >= nr_pages)
				goto out;

			if (sc.nr_scanned && prio < DEF_PRIORITY - 2)
2099
				congestion_wait(WRITE, HZ / 10);
2100
		}
2101
	}
2102 2103 2104 2105 2106

	/*
	 * If ret = 0, we could not shrink LRUs, but there may be something
	 * in slab caches
	 */
2107
	if (!ret) {
2108 2109
		do {
			reclaim_state.reclaimed_slab = 0;
2110
			shrink_slab(nr_pages, sc.gfp_mask, global_lru_pages());
2111 2112
			ret += reclaim_state.reclaimed_slab;
		} while (ret < nr_pages && reclaim_state.reclaimed_slab > 0);
2113
	}
2114 2115

out:
L
Linus Torvalds 已提交
2116
	current->reclaim_state = NULL;
2117

L
Linus Torvalds 已提交
2118 2119 2120 2121 2122 2123 2124 2125
	return ret;
}
#endif

/* 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. */
2126
static int __devinit cpu_callback(struct notifier_block *nfb,
2127
				  unsigned long action, void *hcpu)
L
Linus Torvalds 已提交
2128
{
2129
	int nid;
L
Linus Torvalds 已提交
2130

2131
	if (action == CPU_ONLINE || action == CPU_ONLINE_FROZEN) {
2132
		for_each_node_state(nid, N_HIGH_MEMORY) {
2133 2134 2135
			pg_data_t *pgdat = NODE_DATA(nid);
			node_to_cpumask_ptr(mask, pgdat->node_id);

2136
			if (cpumask_any_and(cpu_online_mask, mask) < nr_cpu_ids)
L
Linus Torvalds 已提交
2137
				/* One of our CPUs online: restore mask */
2138
				set_cpus_allowed_ptr(pgdat->kswapd, mask);
L
Linus Torvalds 已提交
2139 2140 2141 2142 2143
		}
	}
	return NOTIFY_OK;
}

2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165
/*
 * 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;
}

L
Linus Torvalds 已提交
2166 2167
static int __init kswapd_init(void)
{
2168
	int nid;
2169

L
Linus Torvalds 已提交
2170
	swap_setup();
2171
	for_each_node_state(nid, N_HIGH_MEMORY)
2172
 		kswapd_run(nid);
L
Linus Torvalds 已提交
2173 2174 2175 2176 2177
	hotcpu_notifier(cpu_callback, 0);
	return 0;
}

module_init(kswapd_init)
2178 2179 2180 2181 2182 2183 2184 2185 2186 2187

#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;

2188
#define RECLAIM_OFF 0
2189
#define RECLAIM_ZONE (1<<0)	/* Run shrink_inactive_list on the zone */
2190 2191 2192
#define RECLAIM_WRITE (1<<1)	/* Writeout pages during reclaim */
#define RECLAIM_SWAP (1<<2)	/* Swap pages out during reclaim */

2193 2194 2195 2196 2197 2198 2199
/*
 * 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

2200 2201 2202 2203 2204 2205
/*
 * Percentage of pages in a zone that must be unmapped for zone_reclaim to
 * occur.
 */
int sysctl_min_unmapped_ratio = 1;

2206 2207 2208 2209 2210 2211
/*
 * 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;

2212 2213 2214
/*
 * Try to free up some pages from this zone through reclaim.
 */
2215
static int __zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order)
2216
{
2217
	/* Minimum pages needed in order to stay on node */
2218
	const unsigned long nr_pages = 1 << order;
2219 2220
	struct task_struct *p = current;
	struct reclaim_state reclaim_state;
2221
	int priority;
2222
	unsigned long nr_reclaimed = 0;
2223 2224 2225
	struct scan_control sc = {
		.may_writepage = !!(zone_reclaim_mode & RECLAIM_WRITE),
		.may_swap = !!(zone_reclaim_mode & RECLAIM_SWAP),
2226 2227
		.swap_cluster_max = max_t(unsigned long, nr_pages,
					SWAP_CLUSTER_MAX),
2228
		.gfp_mask = gfp_mask,
2229
		.swappiness = vm_swappiness,
2230
		.isolate_pages = isolate_pages_global,
2231
	};
2232
	unsigned long slab_reclaimable;
2233 2234 2235

	disable_swap_token();
	cond_resched();
2236 2237 2238 2239 2240 2241
	/*
	 * 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;
2242 2243
	reclaim_state.reclaimed_slab = 0;
	p->reclaim_state = &reclaim_state;
2244

2245 2246 2247 2248 2249 2250 2251 2252 2253
	if (zone_page_state(zone, NR_FILE_PAGES) -
		zone_page_state(zone, NR_FILE_MAPPED) >
		zone->min_unmapped_pages) {
		/*
		 * Free memory by calling shrink zone with increasing
		 * priorities until we have enough memory freed.
		 */
		priority = ZONE_RECLAIM_PRIORITY;
		do {
2254
			note_zone_scanning_priority(zone, priority);
2255 2256 2257 2258
			nr_reclaimed += shrink_zone(priority, zone, &sc);
			priority--;
		} while (priority >= 0 && nr_reclaimed < nr_pages);
	}
2259

2260 2261
	slab_reclaimable = zone_page_state(zone, NR_SLAB_RECLAIMABLE);
	if (slab_reclaimable > zone->min_slab_pages) {
2262
		/*
2263
		 * shrink_slab() does not currently allow us to determine how
2264 2265 2266 2267
		 * 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.
2268
		 *
2269 2270
		 * Note that shrink_slab will free memory on all zones and may
		 * take a long time.
2271
		 */
2272
		while (shrink_slab(sc.nr_scanned, gfp_mask, order) &&
2273 2274
			zone_page_state(zone, NR_SLAB_RECLAIMABLE) >
				slab_reclaimable - nr_pages)
2275
			;
2276 2277 2278 2279 2280 2281 2282

		/*
		 * Update nr_reclaimed by the number of slab pages we
		 * reclaimed from this zone.
		 */
		nr_reclaimed += slab_reclaimable -
			zone_page_state(zone, NR_SLAB_RECLAIMABLE);
2283 2284
	}

2285
	p->reclaim_state = NULL;
2286
	current->flags &= ~(PF_MEMALLOC | PF_SWAPWRITE);
2287
	return nr_reclaimed >= nr_pages;
2288
}
2289 2290 2291 2292

int zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order)
{
	int node_id;
2293
	int ret;
2294 2295

	/*
2296 2297
	 * Zone reclaim reclaims unmapped file backed pages and
	 * slab pages if we are over the defined limits.
2298
	 *
2299 2300 2301 2302 2303
	 * 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.
2304
	 */
2305
	if (zone_page_state(zone, NR_FILE_PAGES) -
2306 2307 2308
	    zone_page_state(zone, NR_FILE_MAPPED) <= zone->min_unmapped_pages
	    && zone_page_state(zone, NR_SLAB_RECLAIMABLE)
			<= zone->min_slab_pages)
2309
		return 0;
2310

2311 2312 2313
	if (zone_is_all_unreclaimable(zone))
		return 0;

2314
	/*
2315
	 * Do not scan if the allocation should not be delayed.
2316
	 */
2317
	if (!(gfp_mask & __GFP_WAIT) || (current->flags & PF_MEMALLOC))
2318 2319 2320 2321 2322 2323 2324 2325
			return 0;

	/*
	 * 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.
	 */
2326
	node_id = zone_to_nid(zone);
2327
	if (node_state(node_id, N_CPU) && node_id != numa_node_id())
2328
		return 0;
2329 2330 2331 2332 2333 2334 2335

	if (zone_test_and_set_flag(zone, ZONE_RECLAIM_LOCKED))
		return 0;
	ret = __zone_reclaim(zone, gfp_mask, order);
	zone_clear_flag(zone, ZONE_RECLAIM_LOCKED);

	return ret;
2336
}
2337
#endif
L
Lee Schermerhorn 已提交
2338 2339 2340 2341 2342 2343 2344 2345

#ifdef CONFIG_UNEVICTABLE_LRU
/*
 * 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
2346 2347
 * 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
2348 2349
 *
 * Reasons page might not be evictable:
2350
 * (1) page's mapping marked unevictable
N
Nick Piggin 已提交
2351
 * (2) page is part of an mlocked VMA
2352
 *
L
Lee Schermerhorn 已提交
2353 2354 2355 2356
 */
int page_evictable(struct page *page, struct vm_area_struct *vma)
{

2357 2358 2359
	if (mapping_unevictable(page_mapping(page)))
		return 0;

N
Nick Piggin 已提交
2360 2361
	if (PageMlocked(page) || (vma && is_mlocked_vma(vma, page)))
		return 0;
L
Lee Schermerhorn 已提交
2362 2363 2364

	return 1;
}
2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384

/**
 * 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)) {
		enum lru_list l = LRU_INACTIVE_ANON + page_is_file_cache(page);
2385

2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454
		__dec_zone_state(zone, NR_UNEVICTABLE);
		list_move(&page->lru, &zone->lru[l].list);
		__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);
		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);
	}

}
2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508

/**
 * 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 */
void scan_zone_unevictable_pages(struct zone *zone)
{
	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.
 */
2509
static void scan_all_zones_unevictable_pages(void)
2510 2511 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 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583
{
	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,
			   struct file *file, void __user *buffer,
			   size_t *length, loff_t *ppos)
{
	proc_doulongvec_minmax(table, write, file, buffer, length, ppos);

	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);
}

L
Lee Schermerhorn 已提交
2584
#endif