vmscan.c 102.1 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
#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/backing-dev.h>
#include <linux/rmap.h>
#include <linux/topology.h>
#include <linux/cpu.h>
#include <linux/cpuset.h>
34
#include <linux/compaction.h>
L
Linus Torvalds 已提交
35 36
#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>
43
#include <linux/oom.h>
44
#include <linux/prefetch.h>
L
Linus Torvalds 已提交
45 46 47 48 49 50

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

#include <linux/swapops.h>

51 52
#include "internal.h"

53 54 55
#define CREATE_TRACE_POINTS
#include <trace/events/vmscan.h>

56
/*
57 58 59 60 61
 * reclaim_mode determines how the inactive list is shrunk
 * RECLAIM_MODE_SINGLE: Reclaim only order-0 pages
 * RECLAIM_MODE_ASYNC:  Do not block
 * RECLAIM_MODE_SYNC:   Allow blocking e.g. call wait_on_page_writeback
 * RECLAIM_MODE_LUMPYRECLAIM: For high-order allocations, take a reference
62 63
 *			page from the LRU and reclaim all pages within a
 *			naturally aligned range
64
 * RECLAIM_MODE_COMPACTION: For high-order allocations, reclaim a number of
65
 *			order-0 pages and then compact the zone
66
 */
67 68 69 70 71 72
typedef unsigned __bitwise__ reclaim_mode_t;
#define RECLAIM_MODE_SINGLE		((__force reclaim_mode_t)0x01u)
#define RECLAIM_MODE_ASYNC		((__force reclaim_mode_t)0x02u)
#define RECLAIM_MODE_SYNC		((__force reclaim_mode_t)0x04u)
#define RECLAIM_MODE_LUMPYRECLAIM	((__force reclaim_mode_t)0x08u)
#define RECLAIM_MODE_COMPACTION		((__force reclaim_mode_t)0x10u)
73

L
Linus Torvalds 已提交
74 75 76 77
struct scan_control {
	/* Incremented by the number of inactive pages that were scanned */
	unsigned long nr_scanned;

78 79 80
	/* Number of pages freed so far during a call to shrink_zones() */
	unsigned long nr_reclaimed;

81 82 83
	/* How many pages shrink_list() should reclaim */
	unsigned long nr_to_reclaim;

84 85
	unsigned long hibernation_mode;

L
Linus Torvalds 已提交
86
	/* This context's GFP mask */
A
Al Viro 已提交
87
	gfp_t gfp_mask;
L
Linus Torvalds 已提交
88 89 90

	int may_writepage;

91 92
	/* Can mapped pages be reclaimed? */
	int may_unmap;
93

94 95 96
	/* Can pages be swapped as part of reclaim? */
	int may_swap;

A
Andy Whitcroft 已提交
97
	int order;
98

99
	/*
100 101
	 * Intend to reclaim enough continuous memory rather than reclaim
	 * enough amount of memory. i.e, mode for high order allocation.
102
	 */
103
	reclaim_mode_t reclaim_mode;
104

105 106 107 108 109
	/*
	 * The memory cgroup that hit its limit and as a result is the
	 * primary target of this reclaim invocation.
	 */
	struct mem_cgroup *target_mem_cgroup;
110

111 112 113 114 115
	/*
	 * Nodemask of nodes allowed by the caller. If NULL, all nodes
	 * are scanned.
	 */
	nodemask_t	*nodemask;
L
Linus Torvalds 已提交
116 117
};

118 119 120 121 122
struct mem_cgroup_zone {
	struct mem_cgroup *mem_cgroup;
	struct zone *zone;
};

L
Linus Torvalds 已提交
123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156
#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;
157
long vm_total_pages;	/* The total number of pages which the VM controls */
L
Linus Torvalds 已提交
158 159 160 161

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

162
#ifdef CONFIG_CGROUP_MEM_RES_CTLR
163 164
static bool global_reclaim(struct scan_control *sc)
{
165
	return !sc->target_mem_cgroup;
166 167
}

168
static bool scanning_global_lru(struct mem_cgroup_zone *mz)
169
{
170
	return !mz->mem_cgroup;
171
}
172
#else
173 174 175 176 177
static bool global_reclaim(struct scan_control *sc)
{
	return true;
}

178
static bool scanning_global_lru(struct mem_cgroup_zone *mz)
179 180 181
{
	return true;
}
182 183
#endif

184
static struct zone_reclaim_stat *get_reclaim_stat(struct mem_cgroup_zone *mz)
185
{
186 187
	if (!scanning_global_lru(mz))
		return mem_cgroup_get_reclaim_stat(mz->mem_cgroup, mz->zone);
K
KOSAKI Motohiro 已提交
188

189
	return &mz->zone->reclaim_stat;
190 191
}

192 193
static unsigned long zone_nr_lru_pages(struct mem_cgroup_zone *mz,
				       enum lru_list lru)
194
{
195 196 197 198 199
	if (!scanning_global_lru(mz))
		return mem_cgroup_zone_nr_lru_pages(mz->mem_cgroup,
						    zone_to_nid(mz->zone),
						    zone_idx(mz->zone),
						    BIT(lru));
200

201
	return zone_page_state(mz->zone, NR_LRU_BASE + lru);
202 203 204
}


L
Linus Torvalds 已提交
205 206 207
/*
 * Add a shrinker callback to be called from the vm
 */
208
void register_shrinker(struct shrinker *shrinker)
L
Linus Torvalds 已提交
209
{
210
	atomic_long_set(&shrinker->nr_in_batch, 0);
211 212 213
	down_write(&shrinker_rwsem);
	list_add_tail(&shrinker->list, &shrinker_list);
	up_write(&shrinker_rwsem);
L
Linus Torvalds 已提交
214
}
215
EXPORT_SYMBOL(register_shrinker);
L
Linus Torvalds 已提交
216 217 218 219

/*
 * Remove one
 */
220
void unregister_shrinker(struct shrinker *shrinker)
L
Linus Torvalds 已提交
221 222 223 224 225
{
	down_write(&shrinker_rwsem);
	list_del(&shrinker->list);
	up_write(&shrinker_rwsem);
}
226
EXPORT_SYMBOL(unregister_shrinker);
L
Linus Torvalds 已提交
227

228 229 230 231 232 233 234 235
static inline int do_shrinker_shrink(struct shrinker *shrinker,
				     struct shrink_control *sc,
				     unsigned long nr_to_scan)
{
	sc->nr_to_scan = nr_to_scan;
	return (*shrinker->shrink)(shrinker, sc);
}

L
Linus Torvalds 已提交
236 237 238 239 240 241 242 243 244
#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
245
 * If the vm encountered mapped pages on the LRU it increase the pressure on
L
Linus Torvalds 已提交
246 247 248 249 250 251 252
 * 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.
253 254
 *
 * Returns the number of slab objects which we shrunk.
L
Linus Torvalds 已提交
255
 */
256
unsigned long shrink_slab(struct shrink_control *shrink,
257
			  unsigned long nr_pages_scanned,
258
			  unsigned long lru_pages)
L
Linus Torvalds 已提交
259 260
{
	struct shrinker *shrinker;
261
	unsigned long ret = 0;
L
Linus Torvalds 已提交
262

263 264
	if (nr_pages_scanned == 0)
		nr_pages_scanned = SWAP_CLUSTER_MAX;
L
Linus Torvalds 已提交
265

266 267 268 269 270
	if (!down_read_trylock(&shrinker_rwsem)) {
		/* Assume we'll be able to shrink next time */
		ret = 1;
		goto out;
	}
L
Linus Torvalds 已提交
271 272 273

	list_for_each_entry(shrinker, &shrinker_list, list) {
		unsigned long long delta;
274 275
		long total_scan;
		long max_pass;
276
		int shrink_ret = 0;
277 278
		long nr;
		long new_nr;
279 280
		long batch_size = shrinker->batch ? shrinker->batch
						  : SHRINK_BATCH;
L
Linus Torvalds 已提交
281

282 283 284 285
		max_pass = do_shrinker_shrink(shrinker, shrink, 0);
		if (max_pass <= 0)
			continue;

286 287 288 289 290
		/*
		 * copy the current shrinker scan count into a local variable
		 * and zero it so that other concurrent shrinker invocations
		 * don't also do this scanning work.
		 */
291
		nr = atomic_long_xchg(&shrinker->nr_in_batch, 0);
292 293

		total_scan = nr;
294
		delta = (4 * nr_pages_scanned) / shrinker->seeks;
295
		delta *= max_pass;
L
Linus Torvalds 已提交
296
		do_div(delta, lru_pages + 1);
297 298
		total_scan += delta;
		if (total_scan < 0) {
299 300
			printk(KERN_ERR "shrink_slab: %pF negative objects to "
			       "delete nr=%ld\n",
301 302
			       shrinker->shrink, total_scan);
			total_scan = max_pass;
303 304
		}

305 306 307 308 309 310 311 312 313 314 315 316 317 318 319
		/*
		 * We need to avoid excessive windup on filesystem shrinkers
		 * due to large numbers of GFP_NOFS allocations causing the
		 * shrinkers to return -1 all the time. This results in a large
		 * nr being built up so when a shrink that can do some work
		 * comes along it empties the entire cache due to nr >>>
		 * max_pass.  This is bad for sustaining a working set in
		 * memory.
		 *
		 * Hence only allow the shrinker to scan the entire cache when
		 * a large delta change is calculated directly.
		 */
		if (delta < max_pass / 4)
			total_scan = min(total_scan, max_pass / 2);

320 321 322 323 324
		/*
		 * Avoid risking looping forever due to too large nr value:
		 * never try to free more than twice the estimate number of
		 * freeable entries.
		 */
325 326
		if (total_scan > max_pass * 2)
			total_scan = max_pass * 2;
L
Linus Torvalds 已提交
327

328
		trace_mm_shrink_slab_start(shrinker, shrink, nr,
329 330 331
					nr_pages_scanned, lru_pages,
					max_pass, delta, total_scan);

332
		while (total_scan >= batch_size) {
333
			int nr_before;
L
Linus Torvalds 已提交
334

335 336
			nr_before = do_shrinker_shrink(shrinker, shrink, 0);
			shrink_ret = do_shrinker_shrink(shrinker, shrink,
337
							batch_size);
L
Linus Torvalds 已提交
338 339
			if (shrink_ret == -1)
				break;
340 341
			if (shrink_ret < nr_before)
				ret += nr_before - shrink_ret;
342 343
			count_vm_events(SLABS_SCANNED, batch_size);
			total_scan -= batch_size;
L
Linus Torvalds 已提交
344 345 346 347

			cond_resched();
		}

348 349 350 351 352
		/*
		 * move the unused scan count back into the shrinker in a
		 * manner that handles concurrent updates. If we exhausted the
		 * scan, there is no need to do an update.
		 */
353 354 355 356 357
		if (total_scan > 0)
			new_nr = atomic_long_add_return(total_scan,
					&shrinker->nr_in_batch);
		else
			new_nr = atomic_long_read(&shrinker->nr_in_batch);
358 359

		trace_mm_shrink_slab_end(shrinker, shrink_ret, nr, new_nr);
L
Linus Torvalds 已提交
360 361
	}
	up_read(&shrinker_rwsem);
362 363
out:
	cond_resched();
364
	return ret;
L
Linus Torvalds 已提交
365 366
}

367
static void set_reclaim_mode(int priority, struct scan_control *sc,
368 369
				   bool sync)
{
370
	reclaim_mode_t syncmode = sync ? RECLAIM_MODE_SYNC : RECLAIM_MODE_ASYNC;
371 372

	/*
373 374 375
	 * Initially assume we are entering either lumpy reclaim or
	 * reclaim/compaction.Depending on the order, we will either set the
	 * sync mode or just reclaim order-0 pages later.
376
	 */
377
	if (COMPACTION_BUILD)
378
		sc->reclaim_mode = RECLAIM_MODE_COMPACTION;
379
	else
380
		sc->reclaim_mode = RECLAIM_MODE_LUMPYRECLAIM;
381 382

	/*
383 384 385
	 * Avoid using lumpy reclaim or reclaim/compaction if possible by
	 * restricting when its set to either costly allocations or when
	 * under memory pressure
386 387
	 */
	if (sc->order > PAGE_ALLOC_COSTLY_ORDER)
388
		sc->reclaim_mode |= syncmode;
389
	else if (sc->order && priority < DEF_PRIORITY - 2)
390
		sc->reclaim_mode |= syncmode;
391
	else
392
		sc->reclaim_mode = RECLAIM_MODE_SINGLE | RECLAIM_MODE_ASYNC;
393 394
}

395
static void reset_reclaim_mode(struct scan_control *sc)
396
{
397
	sc->reclaim_mode = RECLAIM_MODE_SINGLE | RECLAIM_MODE_ASYNC;
398 399
}

L
Linus Torvalds 已提交
400 401
static inline int is_page_cache_freeable(struct page *page)
{
402 403 404 405 406
	/*
	 * 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.
	 */
407
	return page_count(page) - page_has_private(page) == 2;
L
Linus Torvalds 已提交
408 409
}

410 411
static int may_write_to_queue(struct backing_dev_info *bdi,
			      struct scan_control *sc)
L
Linus Torvalds 已提交
412
{
413
	if (current->flags & PF_SWAPWRITE)
L
Linus Torvalds 已提交
414 415 416 417 418
		return 1;
	if (!bdi_write_congested(bdi))
		return 1;
	if (bdi == current->backing_dev_info)
		return 1;
419 420 421 422

	/* lumpy reclaim for hugepage often need a lot of write */
	if (sc->order > PAGE_ALLOC_COSTLY_ORDER)
		return 1;
L
Linus Torvalds 已提交
423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440
	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)
{
J
Jens Axboe 已提交
441
	lock_page(page);
442 443
	if (page_mapping(page) == mapping)
		mapping_set_error(mapping, error);
L
Linus Torvalds 已提交
444 445 446
	unlock_page(page);
}

447 448 449 450 451 452 453 454 455 456 457 458
/* 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
459
/*
A
Andrew Morton 已提交
460 461
 * pageout is called by shrink_page_list() for each dirty page.
 * Calls ->writepage().
L
Linus Torvalds 已提交
462
 */
463
static pageout_t pageout(struct page *page, struct address_space *mapping,
464
			 struct scan_control *sc)
L
Linus Torvalds 已提交
465 466 467 468 469 470 471 472
{
	/*
	 * 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.
	 *
473
	 * If this process is currently in __generic_file_aio_write() against
L
Linus Torvalds 已提交
474 475 476 477 478 479 480 481 482 483 484 485 486 487 488
	 * 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.
		 */
489
		if (page_has_private(page)) {
L
Linus Torvalds 已提交
490 491
			if (try_to_free_buffers(page)) {
				ClearPageDirty(page);
492
				printk("%s: orphaned page\n", __func__);
L
Linus Torvalds 已提交
493 494 495 496 497 498 499
				return PAGE_CLEAN;
			}
		}
		return PAGE_KEEP;
	}
	if (mapping->a_ops->writepage == NULL)
		return PAGE_ACTIVATE;
500
	if (!may_write_to_queue(mapping->backing_dev_info, sc))
L
Linus Torvalds 已提交
501 502 503 504 505 506 507
		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,
508 509
			.range_start = 0,
			.range_end = LLONG_MAX,
L
Linus Torvalds 已提交
510 511 512 513 514 515 516
			.for_reclaim = 1,
		};

		SetPageReclaim(page);
		res = mapping->a_ops->writepage(page, &wbc);
		if (res < 0)
			handle_write_error(mapping, page, res);
517
		if (res == AOP_WRITEPAGE_ACTIVATE) {
L
Linus Torvalds 已提交
518 519 520
			ClearPageReclaim(page);
			return PAGE_ACTIVATE;
		}
521

L
Linus Torvalds 已提交
522 523 524 525
		if (!PageWriteback(page)) {
			/* synchronous write or broken a_ops? */
			ClearPageReclaim(page);
		}
526
		trace_mm_vmscan_writepage(page,
527
			trace_reclaim_flags(page, sc->reclaim_mode));
528
		inc_zone_page_state(page, NR_VMSCAN_WRITE);
L
Linus Torvalds 已提交
529 530 531 532 533 534
		return PAGE_SUCCESS;
	}

	return PAGE_CLEAN;
}

535
/*
N
Nick Piggin 已提交
536 537
 * Same as remove_mapping, but if the page is removed from the mapping, it
 * gets returned with a refcount of 0.
538
 */
N
Nick Piggin 已提交
539
static int __remove_mapping(struct address_space *mapping, struct page *page)
540
{
541 542
	BUG_ON(!PageLocked(page));
	BUG_ON(mapping != page_mapping(page));
543

N
Nick Piggin 已提交
544
	spin_lock_irq(&mapping->tree_lock);
545
	/*
N
Nick Piggin 已提交
546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568
	 * 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.
569
	 */
N
Nick Piggin 已提交
570
	if (!page_freeze_refs(page, 2))
571
		goto cannot_free;
N
Nick Piggin 已提交
572 573 574
	/* note: atomic_cmpxchg in page_freeze_refs provides the smp_rmb */
	if (unlikely(PageDirty(page))) {
		page_unfreeze_refs(page, 2);
575
		goto cannot_free;
N
Nick Piggin 已提交
576
	}
577 578 579 580

	if (PageSwapCache(page)) {
		swp_entry_t swap = { .val = page_private(page) };
		__delete_from_swap_cache(page);
N
Nick Piggin 已提交
581
		spin_unlock_irq(&mapping->tree_lock);
582
		swapcache_free(swap, page);
N
Nick Piggin 已提交
583
	} else {
584 585 586 587
		void (*freepage)(struct page *);

		freepage = mapping->a_ops->freepage;

588
		__delete_from_page_cache(page);
N
Nick Piggin 已提交
589
		spin_unlock_irq(&mapping->tree_lock);
590
		mem_cgroup_uncharge_cache_page(page);
591 592 593

		if (freepage != NULL)
			freepage(page);
594 595 596 597 598
	}

	return 1;

cannot_free:
N
Nick Piggin 已提交
599
	spin_unlock_irq(&mapping->tree_lock);
600 601 602
	return 0;
}

N
Nick Piggin 已提交
603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622
/*
 * 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
623 624 625 626 627 628 629 630 631 632 633 634 635
/**
 * 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);
636
	int was_unevictable = PageUnevictable(page);
L
Lee Schermerhorn 已提交
637 638 639 640 641 642 643 644 645 646 647 648 649

	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.
		 */
650
		lru = active + page_lru_base_type(page);
L
Lee Schermerhorn 已提交
651 652 653 654 655 656 657 658
		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);
659
		/*
660 661 662
		 * When racing with an mlock or AS_UNEVICTABLE clearing
		 * (page is unlocked) make sure that if the other thread
		 * does not observe our setting of PG_lru and fails
663
		 * isolation/check_move_unevictable_pages,
664
		 * we see PG_mlocked/AS_UNEVICTABLE cleared below and move
665 666
		 * the page back to the evictable list.
		 *
667
		 * The other side is TestClearPageMlocked() or shmem_lock().
668 669
		 */
		smp_mb();
L
Lee Schermerhorn 已提交
670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687
	}

	/*
	 * 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.
		 */
	}

688 689 690 691 692
	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
693 694 695
	put_page(page);		/* drop ref from isolate */
}

696 697 698
enum page_references {
	PAGEREF_RECLAIM,
	PAGEREF_RECLAIM_CLEAN,
699
	PAGEREF_KEEP,
700 701 702 703
	PAGEREF_ACTIVATE,
};

static enum page_references page_check_references(struct page *page,
704
						  struct mem_cgroup_zone *mz,
705 706
						  struct scan_control *sc)
{
707
	int referenced_ptes, referenced_page;
708 709
	unsigned long vm_flags;

710
	referenced_ptes = page_referenced(page, 1, mz->mem_cgroup, &vm_flags);
711
	referenced_page = TestClearPageReferenced(page);
712 713

	/* Lumpy reclaim - ignore references */
714
	if (sc->reclaim_mode & RECLAIM_MODE_LUMPYRECLAIM)
715 716 717 718 719 720 721 722 723
		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;

724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742
	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);

743
		if (referenced_page || referenced_ptes > 1)
744 745
			return PAGEREF_ACTIVATE;

746 747 748 749 750 751
		/*
		 * Activate file-backed executable pages after first usage.
		 */
		if (vm_flags & VM_EXEC)
			return PAGEREF_ACTIVATE;

752 753
		return PAGEREF_KEEP;
	}
754 755

	/* Reclaim if clean, defer dirty pages to writeback */
756
	if (referenced_page && !PageSwapBacked(page))
757 758 759
		return PAGEREF_RECLAIM_CLEAN;

	return PAGEREF_RECLAIM;
760 761
}

L
Linus Torvalds 已提交
762
/*
A
Andrew Morton 已提交
763
 * shrink_page_list() returns the number of reclaimed pages
L
Linus Torvalds 已提交
764
 */
A
Andrew Morton 已提交
765
static unsigned long shrink_page_list(struct list_head *page_list,
766
				      struct mem_cgroup_zone *mz,
767
				      struct scan_control *sc,
768 769 770
				      int priority,
				      unsigned long *ret_nr_dirty,
				      unsigned long *ret_nr_writeback)
L
Linus Torvalds 已提交
771 772
{
	LIST_HEAD(ret_pages);
773
	LIST_HEAD(free_pages);
L
Linus Torvalds 已提交
774
	int pgactivate = 0;
775 776
	unsigned long nr_dirty = 0;
	unsigned long nr_congested = 0;
777
	unsigned long nr_reclaimed = 0;
778
	unsigned long nr_writeback = 0;
L
Linus Torvalds 已提交
779 780 781 782

	cond_resched();

	while (!list_empty(page_list)) {
783
		enum page_references references;
L
Linus Torvalds 已提交
784 785 786 787 788 789 790 791 792
		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
793
		if (!trylock_page(page))
L
Linus Torvalds 已提交
794 795
			goto keep;

N
Nick Piggin 已提交
796
		VM_BUG_ON(PageActive(page));
797
		VM_BUG_ON(page_zone(page) != mz->zone);
L
Linus Torvalds 已提交
798 799

		sc->nr_scanned++;
800

N
Nick Piggin 已提交
801 802
		if (unlikely(!page_evictable(page, NULL)))
			goto cull_mlocked;
L
Lee Schermerhorn 已提交
803

804
		if (!sc->may_unmap && page_mapped(page))
805 806
			goto keep_locked;

L
Linus Torvalds 已提交
807 808 809 810
		/* Double the slab pressure for mapped and swapcache pages */
		if (page_mapped(page) || PageSwapCache(page))
			sc->nr_scanned++;

811 812 813 814
		may_enter_fs = (sc->gfp_mask & __GFP_FS) ||
			(PageSwapCache(page) && (sc->gfp_mask & __GFP_IO));

		if (PageWriteback(page)) {
815
			nr_writeback++;
816
			/*
817 818 819 820
			 * Synchronous reclaim cannot queue pages for
			 * writeback due to the possibility of stack overflow
			 * but if it encounters a page under writeback, wait
			 * for the IO to complete.
821
			 */
822
			if ((sc->reclaim_mode & RECLAIM_MODE_SYNC) &&
823
			    may_enter_fs)
824
				wait_on_page_writeback(page);
825 826 827 828
			else {
				unlock_page(page);
				goto keep_lumpy;
			}
829
		}
L
Linus Torvalds 已提交
830

831
		references = page_check_references(page, mz, sc);
832 833
		switch (references) {
		case PAGEREF_ACTIVATE:
L
Linus Torvalds 已提交
834
			goto activate_locked;
835 836
		case PAGEREF_KEEP:
			goto keep_locked;
837 838 839 840
		case PAGEREF_RECLAIM:
		case PAGEREF_RECLAIM_CLEAN:
			; /* try to reclaim the page below */
		}
L
Linus Torvalds 已提交
841 842 843 844 845

		/*
		 * Anonymous process memory has backing store?
		 * Try to allocate it some swap space here.
		 */
N
Nick Piggin 已提交
846
		if (PageAnon(page) && !PageSwapCache(page)) {
847 848
			if (!(sc->gfp_mask & __GFP_IO))
				goto keep_locked;
849
			if (!add_to_swap(page))
L
Linus Torvalds 已提交
850
				goto activate_locked;
851
			may_enter_fs = 1;
N
Nick Piggin 已提交
852
		}
L
Linus Torvalds 已提交
853 854 855 856 857 858 859 860

		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) {
861
			switch (try_to_unmap(page, TTU_UNMAP)) {
L
Linus Torvalds 已提交
862 863 864 865
			case SWAP_FAIL:
				goto activate_locked;
			case SWAP_AGAIN:
				goto keep_locked;
N
Nick Piggin 已提交
866 867
			case SWAP_MLOCK:
				goto cull_mlocked;
L
Linus Torvalds 已提交
868 869 870 871 872 873
			case SWAP_SUCCESS:
				; /* try to free the page below */
			}
		}

		if (PageDirty(page)) {
874 875
			nr_dirty++;

876 877
			/*
			 * Only kswapd can writeback filesystem pages to
878 879
			 * avoid risk of stack overflow but do not writeback
			 * unless under significant pressure.
880
			 */
881 882
			if (page_is_file_cache(page) &&
					(!current_is_kswapd() || priority >= DEF_PRIORITY - 2)) {
883 884 885 886 887 888 889 890 891
				/*
				 * Immediately reclaim when written back.
				 * Similar in principal to deactivate_page()
				 * except we already have the page isolated
				 * and know it's dirty
				 */
				inc_zone_page_state(page, NR_VMSCAN_IMMEDIATE);
				SetPageReclaim(page);

892 893 894
				goto keep_locked;
			}

895
			if (references == PAGEREF_RECLAIM_CLEAN)
L
Linus Torvalds 已提交
896
				goto keep_locked;
897
			if (!may_enter_fs)
L
Linus Torvalds 已提交
898
				goto keep_locked;
899
			if (!sc->may_writepage)
L
Linus Torvalds 已提交
900 901 902
				goto keep_locked;

			/* Page is dirty, try to write it out here */
903
			switch (pageout(page, mapping, sc)) {
L
Linus Torvalds 已提交
904
			case PAGE_KEEP:
905
				nr_congested++;
L
Linus Torvalds 已提交
906 907 908 909
				goto keep_locked;
			case PAGE_ACTIVATE:
				goto activate_locked;
			case PAGE_SUCCESS:
910 911 912
				if (PageWriteback(page))
					goto keep_lumpy;
				if (PageDirty(page))
L
Linus Torvalds 已提交
913
					goto keep;
914

L
Linus Torvalds 已提交
915 916 917 918
				/*
				 * A synchronous write - probably a ramdisk.  Go
				 * ahead and try to reclaim the page.
				 */
N
Nick Piggin 已提交
919
				if (!trylock_page(page))
L
Linus Torvalds 已提交
920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938
					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
939
		 * will do this, as well as the blockdev mapping.
L
Linus Torvalds 已提交
940 941 942 943 944 945 946 947 948 949
		 * 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.
		 */
950
		if (page_has_private(page)) {
L
Linus Torvalds 已提交
951 952
			if (!try_to_release_page(page, sc->gfp_mask))
				goto activate_locked;
N
Nick Piggin 已提交
953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968
			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
969 970
		}

N
Nick Piggin 已提交
971
		if (!mapping || !__remove_mapping(mapping, page))
972
			goto keep_locked;
L
Linus Torvalds 已提交
973

N
Nick Piggin 已提交
974 975 976 977 978 979 980 981
		/*
		 * 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
982
free_it:
983
		nr_reclaimed++;
984 985 986 987 988 989

		/*
		 * Is there need to periodically free_page_list? It would
		 * appear not as the counts should be low
		 */
		list_add(&page->lru, &free_pages);
L
Linus Torvalds 已提交
990 991
		continue;

N
Nick Piggin 已提交
992
cull_mlocked:
993 994
		if (PageSwapCache(page))
			try_to_free_swap(page);
N
Nick Piggin 已提交
995 996
		unlock_page(page);
		putback_lru_page(page);
997
		reset_reclaim_mode(sc);
N
Nick Piggin 已提交
998 999
		continue;

L
Linus Torvalds 已提交
1000
activate_locked:
1001 1002
		/* Not a candidate for swapping, so reclaim swap space. */
		if (PageSwapCache(page) && vm_swap_full())
1003
			try_to_free_swap(page);
L
Lee Schermerhorn 已提交
1004
		VM_BUG_ON(PageActive(page));
L
Linus Torvalds 已提交
1005 1006 1007 1008 1009
		SetPageActive(page);
		pgactivate++;
keep_locked:
		unlock_page(page);
keep:
1010
		reset_reclaim_mode(sc);
1011
keep_lumpy:
L
Linus Torvalds 已提交
1012
		list_add(&page->lru, &ret_pages);
N
Nick Piggin 已提交
1013
		VM_BUG_ON(PageLRU(page) || PageUnevictable(page));
L
Linus Torvalds 已提交
1014
	}
1015

1016 1017 1018 1019 1020 1021
	/*
	 * Tag a zone as congested if all the dirty pages encountered were
	 * backed by a congested BDI. In this case, reclaimers should just
	 * back off and wait for congestion to clear because further reclaim
	 * will encounter the same problem
	 */
1022
	if (nr_dirty && nr_dirty == nr_congested && global_reclaim(sc))
1023
		zone_set_flag(mz->zone, ZONE_CONGESTED);
1024

1025
	free_hot_cold_page_list(&free_pages, 1);
1026

L
Linus Torvalds 已提交
1027
	list_splice(&ret_pages, page_list);
1028
	count_vm_events(PGACTIVATE, pgactivate);
1029 1030
	*ret_nr_dirty += nr_dirty;
	*ret_nr_writeback += nr_writeback;
1031
	return nr_reclaimed;
L
Linus Torvalds 已提交
1032 1033
}

A
Andy Whitcroft 已提交
1034 1035 1036 1037 1038 1039 1040 1041 1042 1043
/*
 * 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.
 */
1044
int __isolate_lru_page(struct page *page, isolate_mode_t mode, int file)
A
Andy Whitcroft 已提交
1045
{
1046
	bool all_lru_mode;
A
Andy Whitcroft 已提交
1047 1048 1049 1050 1051 1052
	int ret = -EINVAL;

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

1053 1054 1055
	all_lru_mode = (mode & (ISOLATE_ACTIVE|ISOLATE_INACTIVE)) ==
		(ISOLATE_ACTIVE|ISOLATE_INACTIVE);

A
Andy Whitcroft 已提交
1056 1057 1058 1059 1060
	/*
	 * When checking the active state, we need to be sure we are
	 * dealing with comparible boolean values.  Take the logical not
	 * of each.
	 */
1061
	if (!all_lru_mode && !PageActive(page) != !(mode & ISOLATE_ACTIVE))
A
Andy Whitcroft 已提交
1062 1063
		return ret;

1064
	if (!all_lru_mode && !!page_is_file_cache(page) != file)
1065 1066
		return ret;

L
Lee Schermerhorn 已提交
1067 1068 1069 1070 1071 1072 1073 1074
	/*
	 * 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
1075
	ret = -EBUSY;
K
KAMEZAWA Hiroyuki 已提交
1076

1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109
	/*
	 * To minimise LRU disruption, the caller can indicate that it only
	 * wants to isolate pages it will be able to operate on without
	 * blocking - clean pages for the most part.
	 *
	 * ISOLATE_CLEAN means that only clean pages should be isolated. This
	 * is used by reclaim when it is cannot write to backing storage
	 *
	 * ISOLATE_ASYNC_MIGRATE is used to indicate that it only wants to pages
	 * that it is possible to migrate without blocking
	 */
	if (mode & (ISOLATE_CLEAN|ISOLATE_ASYNC_MIGRATE)) {
		/* All the caller can do on PageWriteback is block */
		if (PageWriteback(page))
			return ret;

		if (PageDirty(page)) {
			struct address_space *mapping;

			/* ISOLATE_CLEAN means only clean pages */
			if (mode & ISOLATE_CLEAN)
				return ret;

			/*
			 * Only pages without mappings or that have a
			 * ->migratepage callback are possible to migrate
			 * without blocking
			 */
			mapping = page_mapping(page);
			if (mapping && !mapping->a_ops->migratepage)
				return ret;
		}
	}
1110

1111 1112 1113
	if ((mode & ISOLATE_UNMAPPED) && page_mapped(page))
		return ret;

A
Andy Whitcroft 已提交
1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126
	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
1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137
/*
 * 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.
H
Hugh Dickins 已提交
1138
 * @mz:		The mem_cgroup_zone to pull pages from.
L
Linus Torvalds 已提交
1139
 * @dst:	The temp list to put pages on to.
H
Hugh Dickins 已提交
1140
 * @nr_scanned:	The number of pages that were scanned.
1141
 * @sc:		The scan_control struct for this reclaim session
A
Andy Whitcroft 已提交
1142
 * @mode:	One of the LRU isolation modes
H
Hugh Dickins 已提交
1143
 * @active:	True [1] if isolating active pages
1144
 * @file:	True [1] if isolating file [!anon] pages
L
Linus Torvalds 已提交
1145 1146 1147
 *
 * returns how many pages were moved onto *@dst.
 */
1148
static unsigned long isolate_lru_pages(unsigned long nr_to_scan,
H
Hugh Dickins 已提交
1149
		struct mem_cgroup_zone *mz, struct list_head *dst,
1150 1151
		unsigned long *nr_scanned, struct scan_control *sc,
		isolate_mode_t mode, int active, int file)
L
Linus Torvalds 已提交
1152
{
H
Hugh Dickins 已提交
1153 1154
	struct lruvec *lruvec;
	struct list_head *src;
1155
	unsigned long nr_taken = 0;
1156 1157 1158
	unsigned long nr_lumpy_taken = 0;
	unsigned long nr_lumpy_dirty = 0;
	unsigned long nr_lumpy_failed = 0;
1159
	unsigned long scan;
H
Hugh Dickins 已提交
1160 1161 1162 1163 1164 1165 1166 1167
	int lru = LRU_BASE;

	lruvec = mem_cgroup_zone_lruvec(mz->zone, mz->mem_cgroup);
	if (active)
		lru += LRU_ACTIVE;
	if (file)
		lru += LRU_FILE;
	src = &lruvec->lists[lru];
L
Linus Torvalds 已提交
1168

1169
	for (scan = 0; scan < nr_to_scan && !list_empty(src); scan++) {
A
Andy Whitcroft 已提交
1170 1171 1172 1173 1174 1175
		struct page *page;
		unsigned long pfn;
		unsigned long end_pfn;
		unsigned long page_pfn;
		int zone_id;

L
Linus Torvalds 已提交
1176 1177 1178
		page = lru_to_page(src);
		prefetchw_prev_lru_page(page, src, flags);

N
Nick Piggin 已提交
1179
		VM_BUG_ON(!PageLRU(page));
N
Nick Piggin 已提交
1180

1181
		switch (__isolate_lru_page(page, mode, file)) {
A
Andy Whitcroft 已提交
1182
		case 0:
1183
			mem_cgroup_lru_del(page);
A
Andy Whitcroft 已提交
1184
			list_move(&page->lru, dst);
1185
			nr_taken += hpage_nr_pages(page);
A
Andy Whitcroft 已提交
1186 1187 1188 1189 1190 1191
			break;

		case -EBUSY:
			/* else it is being freed elsewhere */
			list_move(&page->lru, src);
			continue;
1192

A
Andy Whitcroft 已提交
1193 1194 1195 1196
		default:
			BUG();
		}

1197
		if (!sc->order || !(sc->reclaim_mode & RECLAIM_MODE_LUMPYRECLAIM))
A
Andy Whitcroft 已提交
1198 1199 1200 1201 1202 1203 1204
			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
L
Lucas De Marchi 已提交
1205
		 * as the mem_map is guaranteed valid out to MAX_ORDER,
A
Andy Whitcroft 已提交
1206 1207 1208 1209 1210
		 * 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);
1211 1212
		pfn = page_pfn & ~((1 << sc->order) - 1);
		end_pfn = pfn + (1 << sc->order);
A
Andy Whitcroft 已提交
1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224
		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);
1225

A
Andy Whitcroft 已提交
1226 1227
			/* Check that we have not crossed a zone boundary. */
			if (unlikely(page_zone_id(cursor_page) != zone_id))
1228
				break;
1229 1230 1231 1232 1233 1234

			/*
			 * If we don't have enough swap space, reclaiming of
			 * anon page which don't already have a swap slot is
			 * pointless.
			 */
1235
			if (nr_swap_pages <= 0 && PageSwapBacked(cursor_page) &&
1236 1237
			    !PageSwapCache(cursor_page))
				break;
1238

1239
			if (__isolate_lru_page(cursor_page, mode, file) == 0) {
1240 1241
				unsigned int isolated_pages;

1242
				mem_cgroup_lru_del(cursor_page);
A
Andy Whitcroft 已提交
1243
				list_move(&cursor_page->lru, dst);
1244 1245 1246
				isolated_pages = hpage_nr_pages(cursor_page);
				nr_taken += isolated_pages;
				nr_lumpy_taken += isolated_pages;
1247
				if (PageDirty(cursor_page))
1248
					nr_lumpy_dirty += isolated_pages;
A
Andy Whitcroft 已提交
1249
				scan++;
1250
				pfn += isolated_pages - 1;
1251
			} else {
1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265
				/*
				 * Check if the page is freed already.
				 *
				 * We can't use page_count() as that
				 * requires compound_head and we don't
				 * have a pin on the page here. If a
				 * page is tail, we may or may not
				 * have isolated the head, so assume
				 * it's not free, it'd be tricky to
				 * track the head status without a
				 * page pin.
				 */
				if (!PageTail(cursor_page) &&
				    !atomic_read(&cursor_page->_count))
1266 1267
					continue;
				break;
A
Andy Whitcroft 已提交
1268 1269
			}
		}
1270 1271 1272 1273

		/* If we break out of the loop above, lumpy reclaim failed */
		if (pfn < end_pfn)
			nr_lumpy_failed++;
L
Linus Torvalds 已提交
1274 1275
	}

H
Hugh Dickins 已提交
1276
	*nr_scanned = scan;
1277

1278
	trace_mm_vmscan_lru_isolate(sc->order,
1279 1280 1281
			nr_to_scan, scan,
			nr_taken,
			nr_lumpy_taken, nr_lumpy_dirty, nr_lumpy_failed,
1282
			mode, file);
L
Linus Torvalds 已提交
1283 1284 1285
	return nr_taken;
}

1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296
/**
 * 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
1297 1298 1299
 * 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.
1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314
 *
 * 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;

1315 1316
	VM_BUG_ON(!page_count(page));

1317 1318 1319 1320
	if (PageLRU(page)) {
		struct zone *zone = page_zone(page);

		spin_lock_irq(&zone->lru_lock);
1321
		if (PageLRU(page)) {
L
Lee Schermerhorn 已提交
1322
			int lru = page_lru(page);
1323
			ret = 0;
1324
			get_page(page);
1325
			ClearPageLRU(page);
1326 1327

			del_page_from_lru_list(zone, page, lru);
1328 1329 1330 1331 1332 1333
		}
		spin_unlock_irq(&zone->lru_lock);
	}
	return ret;
}

1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344
/*
 * 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;

1345
	if (!global_reclaim(sc))
1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358
		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;
}

1359
static noinline_for_stack void
1360 1361
putback_inactive_pages(struct mem_cgroup_zone *mz,
		       struct list_head *page_list)
1362
{
1363
	struct zone_reclaim_stat *reclaim_stat = get_reclaim_stat(mz);
1364 1365
	struct zone *zone = mz->zone;
	LIST_HEAD(pages_to_free);
1366 1367 1368 1369 1370

	/*
	 * Put back any unfreeable pages.
	 */
	while (!list_empty(page_list)) {
1371
		struct page *page = lru_to_page(page_list);
1372
		int lru;
1373

1374 1375 1376 1377 1378 1379 1380 1381
		VM_BUG_ON(PageLRU(page));
		list_del(&page->lru);
		if (unlikely(!page_evictable(page, NULL))) {
			spin_unlock_irq(&zone->lru_lock);
			putback_lru_page(page);
			spin_lock_irq(&zone->lru_lock);
			continue;
		}
1382
		SetPageLRU(page);
1383
		lru = page_lru(page);
1384
		add_page_to_lru_list(zone, page, lru);
1385 1386
		if (is_active_lru(lru)) {
			int file = is_file_lru(lru);
1387 1388
			int numpages = hpage_nr_pages(page);
			reclaim_stat->recent_rotated[file] += numpages;
1389
		}
1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400
		if (put_page_testzero(page)) {
			__ClearPageLRU(page);
			__ClearPageActive(page);
			del_page_from_lru_list(zone, page, lru);

			if (unlikely(PageCompound(page))) {
				spin_unlock_irq(&zone->lru_lock);
				(*get_compound_page_dtor(page))(page);
				spin_lock_irq(&zone->lru_lock);
			} else
				list_add(&page->lru, &pages_to_free);
1401 1402 1403
		}
	}

1404 1405 1406 1407
	/*
	 * To save our caller's stack, now use input list for pages to free.
	 */
	list_splice(&pages_to_free, page_list);
1408 1409
}

1410 1411
static noinline_for_stack void
update_isolated_counts(struct mem_cgroup_zone *mz,
1412
		       struct list_head *page_list,
1413
		       unsigned long *nr_anon,
1414
		       unsigned long *nr_file)
1415
{
1416
	struct zone_reclaim_stat *reclaim_stat = get_reclaim_stat(mz);
1417
	struct zone *zone = mz->zone;
1418
	unsigned int count[NR_LRU_LISTS] = { 0, };
1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435
	unsigned long nr_active = 0;
	struct page *page;
	int lru;

	/*
	 * Count pages and clear active flags
	 */
	list_for_each_entry(page, page_list, lru) {
		int numpages = hpage_nr_pages(page);
		lru = page_lru_base_type(page);
		if (PageActive(page)) {
			lru += LRU_ACTIVE;
			ClearPageActive(page);
			nr_active += numpages;
		}
		count[lru] += numpages;
	}
1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454

	__count_vm_events(PGDEACTIVATE, nr_active);

	__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]);

	*nr_anon = count[LRU_ACTIVE_ANON] + count[LRU_INACTIVE_ANON];
	*nr_file = count[LRU_ACTIVE_FILE] + count[LRU_INACTIVE_FILE];

	reclaim_stat->recent_scanned[0] += *nr_anon;
	reclaim_stat->recent_scanned[1] += *nr_file;
}

1455
/*
1456
 * Returns true if a direct reclaim should wait on pages under writeback.
1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474
 *
 * If we are direct reclaiming for contiguous pages and we do not reclaim
 * everything in the list, try again and wait for writeback IO to complete.
 * This will stall high-order allocations noticeably. Only do that when really
 * need to free the pages under high memory pressure.
 */
static inline bool should_reclaim_stall(unsigned long nr_taken,
					unsigned long nr_freed,
					int priority,
					struct scan_control *sc)
{
	int lumpy_stall_priority;

	/* kswapd should not stall on sync IO */
	if (current_is_kswapd())
		return false;

	/* Only stall on lumpy reclaim */
1475
	if (sc->reclaim_mode & RECLAIM_MODE_SINGLE)
1476 1477
		return false;

1478
	/* If we have reclaimed everything on the isolated list, no stall */
1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495
	if (nr_freed == nr_taken)
		return false;

	/*
	 * For high-order allocations, there are two stall thresholds.
	 * High-cost allocations stall immediately where as lower
	 * order allocations such as stacks require the scanning
	 * priority to be much higher before stalling.
	 */
	if (sc->order > PAGE_ALLOC_COSTLY_ORDER)
		lumpy_stall_priority = DEF_PRIORITY;
	else
		lumpy_stall_priority = DEF_PRIORITY / 3;

	return priority <= lumpy_stall_priority;
}

L
Linus Torvalds 已提交
1496
/*
A
Andrew Morton 已提交
1497 1498
 * shrink_inactive_list() is a helper for shrink_zone().  It returns the number
 * of reclaimed pages
L
Linus Torvalds 已提交
1499
 */
1500
static noinline_for_stack unsigned long
1501 1502
shrink_inactive_list(unsigned long nr_to_scan, struct mem_cgroup_zone *mz,
		     struct scan_control *sc, int priority, int file)
L
Linus Torvalds 已提交
1503 1504
{
	LIST_HEAD(page_list);
1505
	unsigned long nr_scanned;
1506
	unsigned long nr_reclaimed = 0;
1507 1508 1509
	unsigned long nr_taken;
	unsigned long nr_anon;
	unsigned long nr_file;
1510 1511
	unsigned long nr_dirty = 0;
	unsigned long nr_writeback = 0;
1512
	isolate_mode_t isolate_mode = ISOLATE_INACTIVE;
1513
	struct zone *zone = mz->zone;
1514

1515
	while (unlikely(too_many_isolated(zone, file, sc))) {
1516
		congestion_wait(BLK_RW_ASYNC, HZ/10);
1517 1518 1519 1520 1521 1522

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

1523
	set_reclaim_mode(priority, sc, false);
1524
	if (sc->reclaim_mode & RECLAIM_MODE_LUMPYRECLAIM)
1525
		isolate_mode |= ISOLATE_ACTIVE;
1526

L
Linus Torvalds 已提交
1527
	lru_add_drain();
1528 1529

	if (!sc->may_unmap)
1530
		isolate_mode |= ISOLATE_UNMAPPED;
1531
	if (!sc->may_writepage)
1532
		isolate_mode |= ISOLATE_CLEAN;
1533

L
Linus Torvalds 已提交
1534
	spin_lock_irq(&zone->lru_lock);
1535

1536 1537
	nr_taken = isolate_lru_pages(nr_to_scan, mz, &page_list, &nr_scanned,
				     sc, isolate_mode, 0, file);
1538
	if (global_reclaim(sc)) {
1539 1540 1541 1542 1543 1544 1545 1546
		zone->pages_scanned += nr_scanned;
		if (current_is_kswapd())
			__count_zone_vm_events(PGSCAN_KSWAPD, zone,
					       nr_scanned);
		else
			__count_zone_vm_events(PGSCAN_DIRECT, zone,
					       nr_scanned);
	}
1547

1548 1549 1550 1551
	if (nr_taken == 0) {
		spin_unlock_irq(&zone->lru_lock);
		return 0;
	}
A
Andy Whitcroft 已提交
1552

1553 1554 1555 1556
	update_isolated_counts(mz, &page_list, &nr_anon, &nr_file);

	__mod_zone_page_state(zone, NR_ISOLATED_ANON, nr_anon);
	__mod_zone_page_state(zone, NR_ISOLATED_FILE, nr_file);
L
Linus Torvalds 已提交
1557

1558
	spin_unlock_irq(&zone->lru_lock);
1559

1560
	nr_reclaimed = shrink_page_list(&page_list, mz, sc, priority,
1561
						&nr_dirty, &nr_writeback);
1562

1563 1564
	/* Check if we should syncronously wait for writeback */
	if (should_reclaim_stall(nr_taken, nr_reclaimed, priority, sc)) {
1565
		set_reclaim_mode(priority, sc, true);
1566
		nr_reclaimed += shrink_page_list(&page_list, mz, sc,
1567
					priority, &nr_dirty, &nr_writeback);
1568
	}
1569

1570 1571
	spin_lock_irq(&zone->lru_lock);

1572 1573 1574
	if (current_is_kswapd())
		__count_vm_events(KSWAPD_STEAL, nr_reclaimed);
	__count_zone_vm_events(PGSTEAL, zone, nr_reclaimed);
N
Nick Piggin 已提交
1575

1576 1577 1578 1579 1580 1581 1582 1583
	putback_inactive_pages(mz, &page_list);

	__mod_zone_page_state(zone, NR_ISOLATED_ANON, -nr_anon);
	__mod_zone_page_state(zone, NR_ISOLATED_FILE, -nr_file);

	spin_unlock_irq(&zone->lru_lock);

	free_hot_cold_page_list(&page_list, 1);
1584

1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610
	/*
	 * If reclaim is isolating dirty pages under writeback, it implies
	 * that the long-lived page allocation rate is exceeding the page
	 * laundering rate. Either the global limits are not being effective
	 * at throttling processes due to the page distribution throughout
	 * zones or there is heavy usage of a slow backing device. The
	 * only option is to throttle from reclaim context which is not ideal
	 * as there is no guarantee the dirtying process is throttled in the
	 * same way balance_dirty_pages() manages.
	 *
	 * This scales the number of dirty pages that must be under writeback
	 * before throttling depending on priority. It is a simple backoff
	 * function that has the most effect in the range DEF_PRIORITY to
	 * DEF_PRIORITY-2 which is the priority reclaim is considered to be
	 * in trouble and reclaim is considered to be in trouble.
	 *
	 * DEF_PRIORITY   100% isolated pages must be PageWriteback to throttle
	 * DEF_PRIORITY-1  50% must be PageWriteback
	 * DEF_PRIORITY-2  25% must be PageWriteback, kswapd in trouble
	 * ...
	 * DEF_PRIORITY-6 For SWAP_CLUSTER_MAX isolated pages, throttle if any
	 *                     isolated page is PageWriteback
	 */
	if (nr_writeback && nr_writeback >= (nr_taken >> (DEF_PRIORITY-priority)))
		wait_iff_congested(zone, BLK_RW_ASYNC, HZ/10);

1611 1612 1613 1614
	trace_mm_vmscan_lru_shrink_inactive(zone->zone_pgdat->node_id,
		zone_idx(zone),
		nr_scanned, nr_reclaimed,
		priority,
1615
		trace_shrink_flags(file, sc->reclaim_mode));
1616
	return nr_reclaimed;
L
Linus Torvalds 已提交
1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635
}

/*
 * 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.
 */
1636

1637 1638
static void move_active_pages_to_lru(struct zone *zone,
				     struct list_head *list,
1639
				     struct list_head *pages_to_free,
1640 1641 1642 1643 1644
				     enum lru_list lru)
{
	unsigned long pgmoved = 0;
	struct page *page;

1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655
	if (buffer_heads_over_limit) {
		spin_unlock_irq(&zone->lru_lock);
		list_for_each_entry(page, list, lru) {
			if (page_has_private(page) && trylock_page(page)) {
				if (page_has_private(page))
					try_to_release_page(page, 0);
				unlock_page(page);
			}
		}
		spin_lock_irq(&zone->lru_lock);
	}
1656 1657

	while (!list_empty(list)) {
1658 1659
		struct lruvec *lruvec;

1660 1661 1662 1663 1664
		page = lru_to_page(list);

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

1665 1666
		lruvec = mem_cgroup_lru_add_list(zone, page, lru);
		list_move(&page->lru, &lruvec->lists[lru]);
1667
		pgmoved += hpage_nr_pages(page);
1668

1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679
		if (put_page_testzero(page)) {
			__ClearPageLRU(page);
			__ClearPageActive(page);
			del_page_from_lru_list(zone, page, lru);

			if (unlikely(PageCompound(page))) {
				spin_unlock_irq(&zone->lru_lock);
				(*get_compound_page_dtor(page))(page);
				spin_lock_irq(&zone->lru_lock);
			} else
				list_add(&page->lru, pages_to_free);
1680 1681 1682 1683 1684 1685
		}
	}
	__mod_zone_page_state(zone, NR_LRU_BASE + lru, pgmoved);
	if (!is_active_lru(lru))
		__count_vm_events(PGDEACTIVATE, pgmoved);
}
1686

H
Hugh Dickins 已提交
1687
static void shrink_active_list(unsigned long nr_to_scan,
1688 1689 1690
			       struct mem_cgroup_zone *mz,
			       struct scan_control *sc,
			       int priority, int file)
L
Linus Torvalds 已提交
1691
{
1692
	unsigned long nr_taken;
H
Hugh Dickins 已提交
1693
	unsigned long nr_scanned;
1694
	unsigned long vm_flags;
L
Linus Torvalds 已提交
1695
	LIST_HEAD(l_hold);	/* The pages which were snipped off */
1696
	LIST_HEAD(l_active);
1697
	LIST_HEAD(l_inactive);
L
Linus Torvalds 已提交
1698
	struct page *page;
1699
	struct zone_reclaim_stat *reclaim_stat = get_reclaim_stat(mz);
1700
	unsigned long nr_rotated = 0;
1701
	isolate_mode_t isolate_mode = ISOLATE_ACTIVE;
1702
	struct zone *zone = mz->zone;
L
Linus Torvalds 已提交
1703 1704

	lru_add_drain();
1705 1706

	if (!sc->may_unmap)
1707
		isolate_mode |= ISOLATE_UNMAPPED;
1708
	if (!sc->may_writepage)
1709
		isolate_mode |= ISOLATE_CLEAN;
1710

L
Linus Torvalds 已提交
1711
	spin_lock_irq(&zone->lru_lock);
1712

1713
	nr_taken = isolate_lru_pages(nr_to_scan, mz, &l_hold, &nr_scanned, sc,
1714
				     isolate_mode, 1, file);
1715
	if (global_reclaim(sc))
H
Hugh Dickins 已提交
1716
		zone->pages_scanned += nr_scanned;
1717

1718
	reclaim_stat->recent_scanned[file] += nr_taken;
1719

H
Hugh Dickins 已提交
1720
	__count_zone_vm_events(PGREFILL, zone, nr_scanned);
1721
	if (file)
1722
		__mod_zone_page_state(zone, NR_ACTIVE_FILE, -nr_taken);
1723
	else
1724
		__mod_zone_page_state(zone, NR_ACTIVE_ANON, -nr_taken);
K
KOSAKI Motohiro 已提交
1725
	__mod_zone_page_state(zone, NR_ISOLATED_ANON + file, nr_taken);
L
Linus Torvalds 已提交
1726 1727 1728 1729 1730 1731
	spin_unlock_irq(&zone->lru_lock);

	while (!list_empty(&l_hold)) {
		cond_resched();
		page = lru_to_page(&l_hold);
		list_del(&page->lru);
1732

L
Lee Schermerhorn 已提交
1733 1734 1735 1736 1737
		if (unlikely(!page_evictable(page, NULL))) {
			putback_lru_page(page);
			continue;
		}

1738
		if (page_referenced(page, 0, mz->mem_cgroup, &vm_flags)) {
1739
			nr_rotated += hpage_nr_pages(page);
1740 1741 1742 1743 1744 1745 1746 1747 1748
			/*
			 * 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.
			 */
1749
			if ((vm_flags & VM_EXEC) && page_is_file_cache(page)) {
1750 1751 1752 1753
				list_add(&page->lru, &l_active);
				continue;
			}
		}
1754

1755
		ClearPageActive(page);	/* we are de-activating */
L
Linus Torvalds 已提交
1756 1757 1758
		list_add(&page->lru, &l_inactive);
	}

1759
	/*
1760
	 * Move pages back to the lru list.
1761
	 */
1762
	spin_lock_irq(&zone->lru_lock);
1763
	/*
1764 1765 1766 1767
	 * 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.
1768
	 */
1769
	reclaim_stat->recent_rotated[file] += nr_rotated;
1770

1771
	move_active_pages_to_lru(zone, &l_active, &l_hold,
1772
						LRU_ACTIVE + file * LRU_FILE);
1773
	move_active_pages_to_lru(zone, &l_inactive, &l_hold,
1774
						LRU_BASE   + file * LRU_FILE);
K
KOSAKI Motohiro 已提交
1775
	__mod_zone_page_state(zone, NR_ISOLATED_ANON + file, -nr_taken);
1776
	spin_unlock_irq(&zone->lru_lock);
1777 1778

	free_hot_cold_page_list(&l_hold, 1);
L
Linus Torvalds 已提交
1779 1780
}

1781
#ifdef CONFIG_SWAP
1782
static int inactive_anon_is_low_global(struct zone *zone)
1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794
{
	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;
}

1795 1796 1797 1798 1799 1800 1801 1802
/**
 * 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.
 */
1803
static int inactive_anon_is_low(struct mem_cgroup_zone *mz)
1804
{
1805 1806 1807 1808 1809 1810 1811
	/*
	 * If we don't have swap space, anonymous page deactivation
	 * is pointless.
	 */
	if (!total_swap_pages)
		return 0;

1812 1813 1814 1815 1816
	if (!scanning_global_lru(mz))
		return mem_cgroup_inactive_anon_is_low(mz->mem_cgroup,
						       mz->zone);

	return inactive_anon_is_low_global(mz->zone);
1817
}
1818
#else
1819
static inline int inactive_anon_is_low(struct mem_cgroup_zone *mz)
1820 1821 1822 1823
{
	return 0;
}
#endif
1824

1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836
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
1837
 * @mz: memory cgroup and zone to check
1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848
 *
 * 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.
 */
1849
static int inactive_file_is_low(struct mem_cgroup_zone *mz)
1850
{
1851 1852 1853
	if (!scanning_global_lru(mz))
		return mem_cgroup_inactive_file_is_low(mz->mem_cgroup,
						       mz->zone);
1854

1855
	return inactive_file_is_low_global(mz->zone);
1856 1857
}

1858
static int inactive_list_is_low(struct mem_cgroup_zone *mz, int file)
1859 1860
{
	if (file)
1861
		return inactive_file_is_low(mz);
1862
	else
1863
		return inactive_anon_is_low(mz);
1864 1865
}

1866
static unsigned long shrink_list(enum lru_list lru, unsigned long nr_to_scan,
1867 1868
				 struct mem_cgroup_zone *mz,
				 struct scan_control *sc, int priority)
1869
{
1870 1871
	int file = is_file_lru(lru);

1872
	if (is_active_lru(lru)) {
1873 1874
		if (inactive_list_is_low(mz, file))
			shrink_active_list(nr_to_scan, mz, sc, priority, file);
1875 1876 1877
		return 0;
	}

1878
	return shrink_inactive_list(nr_to_scan, mz, sc, priority, file);
1879 1880
}

1881 1882
static int vmscan_swappiness(struct mem_cgroup_zone *mz,
			     struct scan_control *sc)
1883
{
1884
	if (global_reclaim(sc))
1885
		return vm_swappiness;
1886
	return mem_cgroup_swappiness(mz->mem_cgroup);
1887 1888
}

1889 1890 1891 1892 1893 1894
/*
 * 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.
 *
1895
 * nr[0] = anon pages to scan; nr[1] = file pages to scan
1896
 */
1897 1898
static void get_scan_count(struct mem_cgroup_zone *mz, struct scan_control *sc,
			   unsigned long *nr, int priority)
1899 1900 1901 1902
{
	unsigned long anon, file, free;
	unsigned long anon_prio, file_prio;
	unsigned long ap, fp;
1903
	struct zone_reclaim_stat *reclaim_stat = get_reclaim_stat(mz);
1904
	u64 fraction[2], denominator;
H
Hugh Dickins 已提交
1905
	enum lru_list lru;
1906
	int noswap = 0;
1907
	bool force_scan = false;
1908

1909 1910 1911 1912 1913 1914 1915 1916 1917 1918
	/*
	 * If the zone or memcg is small, nr[l] can be 0.  This
	 * results in no scanning on this priority and a potential
	 * priority drop.  Global direct reclaim can go to the next
	 * zone and tends to have no problems. Global kswapd is for
	 * zone balancing and it needs to scan a minimum amount. When
	 * reclaiming for a memcg, a priority drop can cause high
	 * latencies, so it's better to scan a minimum amount there as
	 * well.
	 */
1919
	if (current_is_kswapd() && mz->zone->all_unreclaimable)
1920
		force_scan = true;
1921
	if (!global_reclaim(sc))
1922
		force_scan = true;
1923 1924 1925 1926 1927 1928 1929 1930 1931

	/* 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;
	}
1932

1933 1934 1935 1936
	anon  = zone_nr_lru_pages(mz, LRU_ACTIVE_ANON) +
		zone_nr_lru_pages(mz, LRU_INACTIVE_ANON);
	file  = zone_nr_lru_pages(mz, LRU_ACTIVE_FILE) +
		zone_nr_lru_pages(mz, LRU_INACTIVE_FILE);
1937

1938
	if (global_reclaim(sc)) {
1939
		free  = zone_page_state(mz->zone, NR_FREE_PAGES);
1940 1941
		/* If we have very few page cache pages,
		   force-scan anon pages. */
1942
		if (unlikely(file + free <= high_wmark_pages(mz->zone))) {
1943 1944 1945 1946
			fraction[0] = 1;
			fraction[1] = 0;
			denominator = 1;
			goto out;
1947
		}
1948 1949
	}

1950 1951 1952 1953
	/*
	 * With swappiness at 100, anonymous and file have the same priority.
	 * This scanning priority is essentially the inverse of IO cost.
	 */
1954 1955
	anon_prio = vmscan_swappiness(mz, sc);
	file_prio = 200 - vmscan_swappiness(mz, sc);
1956

1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967
	/*
	 * 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]
	 */
1968
	spin_lock_irq(&mz->zone->lru_lock);
1969 1970 1971
	if (unlikely(reclaim_stat->recent_scanned[0] > anon / 4)) {
		reclaim_stat->recent_scanned[0] /= 2;
		reclaim_stat->recent_rotated[0] /= 2;
1972 1973
	}

1974 1975 1976
	if (unlikely(reclaim_stat->recent_scanned[1] > file / 4)) {
		reclaim_stat->recent_scanned[1] /= 2;
		reclaim_stat->recent_rotated[1] /= 2;
1977 1978 1979
	}

	/*
1980 1981 1982
	 * 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.
1983
	 */
1984 1985
	ap = (anon_prio + 1) * (reclaim_stat->recent_scanned[0] + 1);
	ap /= reclaim_stat->recent_rotated[0] + 1;
1986

1987 1988
	fp = (file_prio + 1) * (reclaim_stat->recent_scanned[1] + 1);
	fp /= reclaim_stat->recent_rotated[1] + 1;
1989
	spin_unlock_irq(&mz->zone->lru_lock);
1990

1991 1992 1993 1994
	fraction[0] = ap;
	fraction[1] = fp;
	denominator = ap + fp + 1;
out:
H
Hugh Dickins 已提交
1995 1996
	for_each_evictable_lru(lru) {
		int file = is_file_lru(lru);
1997
		unsigned long scan;
1998

H
Hugh Dickins 已提交
1999
		scan = zone_nr_lru_pages(mz, lru);
2000 2001
		if (priority || noswap) {
			scan >>= priority;
2002 2003
			if (!scan && force_scan)
				scan = SWAP_CLUSTER_MAX;
2004 2005
			scan = div64_u64(scan * fraction[file], denominator);
		}
H
Hugh Dickins 已提交
2006
		nr[lru] = scan;
2007
	}
2008
}
2009

2010 2011 2012 2013 2014 2015 2016
/*
 * Reclaim/compaction depends on a number of pages being freed. To avoid
 * disruption to the system, a small number of order-0 pages continue to be
 * rotated and reclaimed in the normal fashion. However, by the time we get
 * back to the allocator and call try_to_compact_zone(), we ensure that
 * there are enough free pages for it to be likely successful
 */
2017
static inline bool should_continue_reclaim(struct mem_cgroup_zone *mz,
2018 2019 2020 2021 2022 2023 2024 2025
					unsigned long nr_reclaimed,
					unsigned long nr_scanned,
					struct scan_control *sc)
{
	unsigned long pages_for_compaction;
	unsigned long inactive_lru_pages;

	/* If not in reclaim/compaction mode, stop */
2026
	if (!(sc->reclaim_mode & RECLAIM_MODE_COMPACTION))
2027 2028
		return false;

2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050
	/* Consider stopping depending on scan and reclaim activity */
	if (sc->gfp_mask & __GFP_REPEAT) {
		/*
		 * For __GFP_REPEAT allocations, stop reclaiming if the
		 * full LRU list has been scanned and we are still failing
		 * to reclaim pages. This full LRU scan is potentially
		 * expensive but a __GFP_REPEAT caller really wants to succeed
		 */
		if (!nr_reclaimed && !nr_scanned)
			return false;
	} else {
		/*
		 * For non-__GFP_REPEAT allocations which can presumably
		 * fail without consequence, stop if we failed to reclaim
		 * any pages from the last SWAP_CLUSTER_MAX number of
		 * pages that were scanned. This will return to the
		 * caller faster at the risk reclaim/compaction and
		 * the resulting allocation attempt fails
		 */
		if (!nr_reclaimed)
			return false;
	}
2051 2052 2053 2054 2055 2056

	/*
	 * If we have not reclaimed enough pages for compaction and the
	 * inactive lists are large enough, continue reclaiming
	 */
	pages_for_compaction = (2UL << sc->order);
2057
	inactive_lru_pages = zone_nr_lru_pages(mz, LRU_INACTIVE_FILE);
2058
	if (nr_swap_pages > 0)
2059
		inactive_lru_pages += zone_nr_lru_pages(mz, LRU_INACTIVE_ANON);
2060 2061 2062 2063 2064
	if (sc->nr_reclaimed < pages_for_compaction &&
			inactive_lru_pages > pages_for_compaction)
		return true;

	/* If compaction would go ahead or the allocation would succeed, stop */
2065
	switch (compaction_suitable(mz->zone, sc->order)) {
2066 2067 2068 2069 2070 2071 2072 2073
	case COMPACT_PARTIAL:
	case COMPACT_CONTINUE:
		return false;
	default:
		return true;
	}
}

L
Linus Torvalds 已提交
2074 2075 2076
/*
 * This is a basic per-zone page freer.  Used by both kswapd and direct reclaim.
 */
2077 2078
static void shrink_mem_cgroup_zone(int priority, struct mem_cgroup_zone *mz,
				   struct scan_control *sc)
L
Linus Torvalds 已提交
2079
{
2080
	unsigned long nr[NR_LRU_LISTS];
2081
	unsigned long nr_to_scan;
H
Hugh Dickins 已提交
2082
	enum lru_list lru;
2083
	unsigned long nr_reclaimed, nr_scanned;
2084
	unsigned long nr_to_reclaim = sc->nr_to_reclaim;
2085
	struct blk_plug plug;
2086

2087 2088
restart:
	nr_reclaimed = 0;
2089
	nr_scanned = sc->nr_scanned;
2090
	get_scan_count(mz, sc, nr, priority);
L
Linus Torvalds 已提交
2091

2092
	blk_start_plug(&plug);
2093 2094
	while (nr[LRU_INACTIVE_ANON] || nr[LRU_ACTIVE_FILE] ||
					nr[LRU_INACTIVE_FILE]) {
H
Hugh Dickins 已提交
2095 2096
		for_each_evictable_lru(lru) {
			if (nr[lru]) {
K
KOSAKI Motohiro 已提交
2097
				nr_to_scan = min_t(unsigned long,
H
Hugh Dickins 已提交
2098 2099
						   nr[lru], SWAP_CLUSTER_MAX);
				nr[lru] -= nr_to_scan;
L
Linus Torvalds 已提交
2100

H
Hugh Dickins 已提交
2101
				nr_reclaimed += shrink_list(lru, nr_to_scan,
2102
							    mz, sc, priority);
2103
			}
L
Linus Torvalds 已提交
2104
		}
2105 2106 2107 2108 2109 2110 2111 2112
		/*
		 * 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.
		 */
2113 2114 2115 2116 2117 2118
		if (nr_reclaimed >= nr_to_reclaim)
			nr_to_reclaim = 0;
		else
			nr_to_reclaim -= nr_reclaimed;

		if (!nr_to_reclaim && priority < DEF_PRIORITY)
2119
			break;
L
Linus Torvalds 已提交
2120
	}
2121
	blk_finish_plug(&plug);
2122
	sc->nr_reclaimed += nr_reclaimed;
2123

2124 2125 2126 2127
	/*
	 * Even if we did not try to evict anon pages at all, we want to
	 * rebalance the anon lru active/inactive ratio.
	 */
2128 2129
	if (inactive_anon_is_low(mz))
		shrink_active_list(SWAP_CLUSTER_MAX, mz, sc, priority, 0);
2130

2131
	/* reclaim/compaction might need reclaim to continue */
2132
	if (should_continue_reclaim(mz, nr_reclaimed,
2133 2134 2135
					sc->nr_scanned - nr_scanned, sc))
		goto restart;

2136
	throttle_vm_writeout(sc->gfp_mask);
L
Linus Torvalds 已提交
2137 2138
}

2139 2140 2141
static void shrink_zone(int priority, struct zone *zone,
			struct scan_control *sc)
{
2142 2143
	struct mem_cgroup *root = sc->target_mem_cgroup;
	struct mem_cgroup_reclaim_cookie reclaim = {
2144
		.zone = zone,
2145
		.priority = priority,
2146
	};
2147 2148 2149 2150 2151 2152 2153 2154
	struct mem_cgroup *memcg;

	memcg = mem_cgroup_iter(root, NULL, &reclaim);
	do {
		struct mem_cgroup_zone mz = {
			.mem_cgroup = memcg,
			.zone = zone,
		};
2155

2156 2157 2158 2159 2160 2161
		shrink_mem_cgroup_zone(priority, &mz, sc);
		/*
		 * Limit reclaim has historically picked one memcg and
		 * scanned it with decreasing priority levels until
		 * nr_to_reclaim had been reclaimed.  This priority
		 * cycle is thus over after a single memcg.
2162 2163 2164 2165
		 *
		 * Direct reclaim and kswapd, on the other hand, have
		 * to scan all memory cgroups to fulfill the overall
		 * scan target for the zone.
2166 2167 2168 2169 2170 2171 2172
		 */
		if (!global_reclaim(sc)) {
			mem_cgroup_iter_break(root, memcg);
			break;
		}
		memcg = mem_cgroup_iter(root, memcg, &reclaim);
	} while (memcg);
2173 2174
}

2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200
/* Returns true if compaction should go ahead for a high-order request */
static inline bool compaction_ready(struct zone *zone, struct scan_control *sc)
{
	unsigned long balance_gap, watermark;
	bool watermark_ok;

	/* Do not consider compaction for orders reclaim is meant to satisfy */
	if (sc->order <= PAGE_ALLOC_COSTLY_ORDER)
		return false;

	/*
	 * Compaction takes time to run and there are potentially other
	 * callers using the pages just freed. Continue reclaiming until
	 * there is a buffer of free pages available to give compaction
	 * a reasonable chance of completing and allocating the page
	 */
	balance_gap = min(low_wmark_pages(zone),
		(zone->present_pages + KSWAPD_ZONE_BALANCE_GAP_RATIO-1) /
			KSWAPD_ZONE_BALANCE_GAP_RATIO);
	watermark = high_wmark_pages(zone) + balance_gap + (2UL << sc->order);
	watermark_ok = zone_watermark_ok_safe(zone, 0, watermark, 0, 0);

	/*
	 * If compaction is deferred, reclaim up to a point where
	 * compaction will have a chance of success when re-enabled
	 */
2201
	if (compaction_deferred(zone, sc->order))
2202 2203 2204 2205 2206 2207 2208 2209 2210
		return watermark_ok;

	/* If compaction is not ready to start, keep reclaiming */
	if (!compaction_suitable(zone, sc->order))
		return false;

	return watermark_ok;
}

L
Linus Torvalds 已提交
2211 2212 2213 2214 2215
/*
 * 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.
 *
2216 2217
 * We reclaim from a zone even if that zone is over high_wmark_pages(zone).
 * Because:
L
Linus Torvalds 已提交
2218 2219
 * a) The caller may be trying to free *extra* pages to satisfy a higher-order
 *    allocation or
2220 2221 2222
 * 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
2223 2224 2225
 *
 * If a zone is deemed to be full of pinned pages then just give it a light
 * scan then give up on it.
2226 2227
 *
 * This function returns true if a zone is being reclaimed for a costly
2228
 * high-order allocation and compaction is ready to begin. This indicates to
2229 2230
 * the caller that it should consider retrying the allocation instead of
 * further reclaim.
L
Linus Torvalds 已提交
2231
 */
2232
static bool shrink_zones(int priority, struct zonelist *zonelist,
2233
					struct scan_control *sc)
L
Linus Torvalds 已提交
2234
{
2235
	struct zoneref *z;
2236
	struct zone *zone;
2237 2238
	unsigned long nr_soft_reclaimed;
	unsigned long nr_soft_scanned;
2239
	bool aborted_reclaim = false;
2240

2241 2242
	for_each_zone_zonelist_nodemask(zone, z, zonelist,
					gfp_zone(sc->gfp_mask), sc->nodemask) {
2243
		if (!populated_zone(zone))
L
Linus Torvalds 已提交
2244
			continue;
2245 2246 2247 2248
		/*
		 * Take care memory controller reclaiming has small influence
		 * to global LRU.
		 */
2249
		if (global_reclaim(sc)) {
2250 2251
			if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL))
				continue;
2252
			if (zone->all_unreclaimable && priority != DEF_PRIORITY)
2253
				continue;	/* Let kswapd poll it */
2254 2255
			if (COMPACTION_BUILD) {
				/*
2256 2257 2258 2259 2260 2261 2262
				 * If we already have plenty of memory free for
				 * compaction in this zone, don't free any more.
				 * Even though compaction is invoked for any
				 * non-zero order, only frequent costly order
				 * reclamation is disruptive enough to become a
				 * noticable problem, like transparent huge page
				 * allocations.
2263
				 */
2264
				if (compaction_ready(zone, sc)) {
2265
					aborted_reclaim = true;
2266
					continue;
2267
				}
2268
			}
2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281
			/*
			 * This steals pages from memory cgroups over softlimit
			 * and returns the number of reclaimed pages and
			 * scanned pages. This works for global memory pressure
			 * and balancing, not for a memcg's limit.
			 */
			nr_soft_scanned = 0;
			nr_soft_reclaimed = mem_cgroup_soft_limit_reclaim(zone,
						sc->order, sc->gfp_mask,
						&nr_soft_scanned);
			sc->nr_reclaimed += nr_soft_reclaimed;
			sc->nr_scanned += nr_soft_scanned;
			/* need some check for avoid more shrink_zone() */
2282
		}
2283

2284
		shrink_zone(priority, zone, sc);
L
Linus Torvalds 已提交
2285
	}
2286

2287
	return aborted_reclaim;
2288 2289 2290 2291 2292 2293 2294
}

static bool zone_reclaimable(struct zone *zone)
{
	return zone->pages_scanned < zone_reclaimable_pages(zone) * 6;
}

2295
/* All zones in zonelist are unreclaimable? */
2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307
static bool all_unreclaimable(struct zonelist *zonelist,
		struct scan_control *sc)
{
	struct zoneref *z;
	struct zone *zone;

	for_each_zone_zonelist_nodemask(zone, z, zonelist,
			gfp_zone(sc->gfp_mask), sc->nodemask) {
		if (!populated_zone(zone))
			continue;
		if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL))
			continue;
2308 2309
		if (!zone->all_unreclaimable)
			return false;
2310 2311
	}

2312
	return true;
L
Linus Torvalds 已提交
2313
}
2314

L
Linus Torvalds 已提交
2315 2316 2317 2318 2319 2320 2321 2322
/*
 * 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
2323 2324 2325 2326
 * 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.
2327 2328 2329
 *
 * returns:	0, if no pages reclaimed
 * 		else, the number of pages reclaimed
L
Linus Torvalds 已提交
2330
 */
2331
static unsigned long do_try_to_free_pages(struct zonelist *zonelist,
2332 2333
					struct scan_control *sc,
					struct shrink_control *shrink)
L
Linus Torvalds 已提交
2334 2335
{
	int priority;
2336
	unsigned long total_scanned = 0;
L
Linus Torvalds 已提交
2337
	struct reclaim_state *reclaim_state = current->reclaim_state;
2338
	struct zoneref *z;
2339
	struct zone *zone;
2340
	unsigned long writeback_threshold;
2341
	bool aborted_reclaim;
L
Linus Torvalds 已提交
2342

2343
	get_mems_allowed();
2344 2345
	delayacct_freepages_start();

2346
	if (global_reclaim(sc))
2347
		count_vm_event(ALLOCSTALL);
L
Linus Torvalds 已提交
2348 2349

	for (priority = DEF_PRIORITY; priority >= 0; priority--) {
2350
		sc->nr_scanned = 0;
2351
		if (!priority)
2352
			disable_swap_token(sc->target_mem_cgroup);
2353
		aborted_reclaim = shrink_zones(priority, zonelist, sc);
2354

2355 2356 2357 2358
		/*
		 * Don't shrink slabs when reclaiming memory from
		 * over limit cgroups
		 */
2359
		if (global_reclaim(sc)) {
2360
			unsigned long lru_pages = 0;
2361 2362
			for_each_zone_zonelist(zone, z, zonelist,
					gfp_zone(sc->gfp_mask)) {
2363 2364 2365 2366 2367 2368
				if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL))
					continue;

				lru_pages += zone_reclaimable_pages(zone);
			}

2369
			shrink_slab(shrink, sc->nr_scanned, lru_pages);
2370
			if (reclaim_state) {
2371
				sc->nr_reclaimed += reclaim_state->reclaimed_slab;
2372 2373
				reclaim_state->reclaimed_slab = 0;
			}
L
Linus Torvalds 已提交
2374
		}
2375
		total_scanned += sc->nr_scanned;
2376
		if (sc->nr_reclaimed >= sc->nr_to_reclaim)
L
Linus Torvalds 已提交
2377 2378 2379 2380 2381 2382 2383 2384 2385
			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.
		 */
2386 2387
		writeback_threshold = sc->nr_to_reclaim + sc->nr_to_reclaim / 2;
		if (total_scanned > writeback_threshold) {
2388 2389
			wakeup_flusher_threads(laptop_mode ? 0 : total_scanned,
						WB_REASON_TRY_TO_FREE_PAGES);
2390
			sc->may_writepage = 1;
L
Linus Torvalds 已提交
2391 2392 2393
		}

		/* Take a nap, wait for some writeback to complete */
2394
		if (!sc->hibernation_mode && sc->nr_scanned &&
2395 2396 2397 2398
		    priority < DEF_PRIORITY - 2) {
			struct zone *preferred_zone;

			first_zones_zonelist(zonelist, gfp_zone(sc->gfp_mask),
2399 2400
						&cpuset_current_mems_allowed,
						&preferred_zone);
2401 2402
			wait_iff_congested(preferred_zone, BLK_RW_ASYNC, HZ/10);
		}
L
Linus Torvalds 已提交
2403
	}
2404

L
Linus Torvalds 已提交
2405
out:
2406
	delayacct_freepages_end();
2407
	put_mems_allowed();
2408

2409 2410 2411
	if (sc->nr_reclaimed)
		return sc->nr_reclaimed;

2412 2413 2414 2415 2416 2417 2418 2419
	/*
	 * As hibernation is going on, kswapd is freezed so that it can't mark
	 * the zone into all_unreclaimable. Thus bypassing all_unreclaimable
	 * check.
	 */
	if (oom_killer_disabled)
		return 0;

2420 2421
	/* Aborted reclaim to try compaction? don't OOM, then */
	if (aborted_reclaim)
2422 2423
		return 1;

2424
	/* top priority shrink_zones still had more to do? don't OOM, then */
2425
	if (global_reclaim(sc) && !all_unreclaimable(zonelist, sc))
2426 2427 2428
		return 1;

	return 0;
L
Linus Torvalds 已提交
2429 2430
}

2431
unsigned long try_to_free_pages(struct zonelist *zonelist, int order,
2432
				gfp_t gfp_mask, nodemask_t *nodemask)
2433
{
2434
	unsigned long nr_reclaimed;
2435 2436 2437
	struct scan_control sc = {
		.gfp_mask = gfp_mask,
		.may_writepage = !laptop_mode,
2438
		.nr_to_reclaim = SWAP_CLUSTER_MAX,
2439
		.may_unmap = 1,
2440
		.may_swap = 1,
2441
		.order = order,
2442
		.target_mem_cgroup = NULL,
2443
		.nodemask = nodemask,
2444
	};
2445 2446 2447
	struct shrink_control shrink = {
		.gfp_mask = sc.gfp_mask,
	};
2448

2449 2450 2451 2452
	trace_mm_vmscan_direct_reclaim_begin(order,
				sc.may_writepage,
				gfp_mask);

2453
	nr_reclaimed = do_try_to_free_pages(zonelist, &sc, &shrink);
2454 2455 2456 2457

	trace_mm_vmscan_direct_reclaim_end(nr_reclaimed);

	return nr_reclaimed;
2458 2459
}

2460
#ifdef CONFIG_CGROUP_MEM_RES_CTLR
2461

2462
unsigned long mem_cgroup_shrink_node_zone(struct mem_cgroup *memcg,
2463
						gfp_t gfp_mask, bool noswap,
2464 2465
						struct zone *zone,
						unsigned long *nr_scanned)
2466 2467
{
	struct scan_control sc = {
2468
		.nr_scanned = 0,
2469
		.nr_to_reclaim = SWAP_CLUSTER_MAX,
2470 2471 2472 2473
		.may_writepage = !laptop_mode,
		.may_unmap = 1,
		.may_swap = !noswap,
		.order = 0,
2474
		.target_mem_cgroup = memcg,
2475
	};
2476
	struct mem_cgroup_zone mz = {
2477
		.mem_cgroup = memcg,
2478 2479
		.zone = zone,
	};
2480

2481 2482
	sc.gfp_mask = (gfp_mask & GFP_RECLAIM_MASK) |
			(GFP_HIGHUSER_MOVABLE & ~GFP_RECLAIM_MASK);
2483 2484 2485 2486 2487

	trace_mm_vmscan_memcg_softlimit_reclaim_begin(0,
						      sc.may_writepage,
						      sc.gfp_mask);

2488 2489 2490 2491 2492 2493 2494
	/*
	 * 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.
	 */
2495
	shrink_mem_cgroup_zone(0, &mz, &sc);
2496 2497 2498

	trace_mm_vmscan_memcg_softlimit_reclaim_end(sc.nr_reclaimed);

2499
	*nr_scanned = sc.nr_scanned;
2500 2501 2502
	return sc.nr_reclaimed;
}

2503
unsigned long try_to_free_mem_cgroup_pages(struct mem_cgroup *memcg,
K
KOSAKI Motohiro 已提交
2504
					   gfp_t gfp_mask,
2505
					   bool noswap)
2506
{
2507
	struct zonelist *zonelist;
2508
	unsigned long nr_reclaimed;
2509
	int nid;
2510 2511
	struct scan_control sc = {
		.may_writepage = !laptop_mode,
2512
		.may_unmap = 1,
2513
		.may_swap = !noswap,
2514
		.nr_to_reclaim = SWAP_CLUSTER_MAX,
2515
		.order = 0,
2516
		.target_mem_cgroup = memcg,
2517
		.nodemask = NULL, /* we don't care the placement */
2518 2519 2520 2521 2522
		.gfp_mask = (gfp_mask & GFP_RECLAIM_MASK) |
				(GFP_HIGHUSER_MOVABLE & ~GFP_RECLAIM_MASK),
	};
	struct shrink_control shrink = {
		.gfp_mask = sc.gfp_mask,
2523 2524
	};

2525 2526 2527 2528 2529
	/*
	 * Unlike direct reclaim via alloc_pages(), memcg's reclaim doesn't
	 * take care of from where we get pages. So the node where we start the
	 * scan does not need to be the current node.
	 */
2530
	nid = mem_cgroup_select_victim_node(memcg);
2531 2532

	zonelist = NODE_DATA(nid)->node_zonelists;
2533 2534 2535 2536 2537

	trace_mm_vmscan_memcg_reclaim_begin(0,
					    sc.may_writepage,
					    sc.gfp_mask);

2538
	nr_reclaimed = do_try_to_free_pages(zonelist, &sc, &shrink);
2539 2540 2541 2542

	trace_mm_vmscan_memcg_reclaim_end(nr_reclaimed);

	return nr_reclaimed;
2543 2544 2545
}
#endif

2546 2547 2548
static void age_active_anon(struct zone *zone, struct scan_control *sc,
			    int priority)
{
2549
	struct mem_cgroup *memcg;
2550

2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566
	if (!total_swap_pages)
		return;

	memcg = mem_cgroup_iter(NULL, NULL, NULL);
	do {
		struct mem_cgroup_zone mz = {
			.mem_cgroup = memcg,
			.zone = zone,
		};

		if (inactive_anon_is_low(&mz))
			shrink_active_list(SWAP_CLUSTER_MAX, &mz,
					   sc, priority, 0);

		memcg = mem_cgroup_iter(NULL, memcg, NULL);
	} while (memcg);
2567 2568
}

2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579
/*
 * pgdat_balanced is used when checking if a node is balanced for high-order
 * allocations. Only zones that meet watermarks and are in a zone allowed
 * by the callers classzone_idx are added to balanced_pages. The total of
 * balanced pages must be at least 25% of the zones allowed by classzone_idx
 * for the node to be considered balanced. Forcing all zones to be balanced
 * for high orders can cause excessive reclaim when there are imbalanced zones.
 * The choice of 25% is due to
 *   o a 16M DMA zone that is balanced will not balance a zone on any
 *     reasonable sized machine
 *   o On all other machines, the top zone must be at least a reasonable
L
Lucas De Marchi 已提交
2580
 *     percentage of the middle zones. For example, on 32-bit x86, highmem
2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593
 *     would need to be at least 256M for it to be balance a whole node.
 *     Similarly, on x86-64 the Normal zone would need to be at least 1G
 *     to balance a node on its own. These seemed like reasonable ratios.
 */
static bool pgdat_balanced(pg_data_t *pgdat, unsigned long balanced_pages,
						int classzone_idx)
{
	unsigned long present_pages = 0;
	int i;

	for (i = 0; i <= classzone_idx; i++)
		present_pages += pgdat->node_zones[i].present_pages;

S
Shaohua Li 已提交
2594 2595
	/* A special case here: if zone has no page, we think it's balanced */
	return balanced_pages >= (present_pages >> 2);
2596 2597
}

2598
/* is kswapd sleeping prematurely? */
2599 2600
static bool sleeping_prematurely(pg_data_t *pgdat, int order, long remaining,
					int classzone_idx)
2601
{
2602
	int i;
2603 2604
	unsigned long balanced = 0;
	bool all_zones_ok = true;
2605 2606 2607

	/* If a direct reclaimer woke kswapd within HZ/10, it's premature */
	if (remaining)
2608
		return true;
2609

2610
	/* Check the watermark levels */
2611
	for (i = 0; i <= classzone_idx; i++) {
2612 2613 2614 2615 2616
		struct zone *zone = pgdat->node_zones + i;

		if (!populated_zone(zone))
			continue;

2617 2618 2619 2620 2621 2622 2623 2624
		/*
		 * balance_pgdat() skips over all_unreclaimable after
		 * DEF_PRIORITY. Effectively, it considers them balanced so
		 * they must be considered balanced here as well if kswapd
		 * is to sleep
		 */
		if (zone->all_unreclaimable) {
			balanced += zone->present_pages;
2625
			continue;
2626
		}
2627

2628
		if (!zone_watermark_ok_safe(zone, order, high_wmark_pages(zone),
2629
							i, 0))
2630 2631 2632
			all_zones_ok = false;
		else
			balanced += zone->present_pages;
2633
	}
2634

2635 2636 2637 2638 2639 2640
	/*
	 * For high-order requests, the balanced zones must contain at least
	 * 25% of the nodes pages for kswapd to sleep. For order-0, all zones
	 * must be balanced
	 */
	if (order)
2641
		return !pgdat_balanced(pgdat, balanced, classzone_idx);
2642 2643
	else
		return !all_zones_ok;
2644 2645
}

L
Linus Torvalds 已提交
2646 2647
/*
 * For kswapd, balance_pgdat() will work across all this node's zones until
2648
 * they are all at high_wmark_pages(zone).
L
Linus Torvalds 已提交
2649
 *
2650
 * Returns the final order kswapd was reclaiming at
L
Linus Torvalds 已提交
2651 2652 2653 2654 2655 2656 2657 2658 2659 2660
 *
 * 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
2661 2662 2663 2664 2665
 * 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
2666
 */
2667
static unsigned long balance_pgdat(pg_data_t *pgdat, int order,
2668
							int *classzone_idx)
L
Linus Torvalds 已提交
2669 2670
{
	int all_zones_ok;
2671
	unsigned long balanced;
L
Linus Torvalds 已提交
2672 2673
	int priority;
	int i;
2674
	int end_zone = 0;	/* Inclusive.  0 = ZONE_DMA */
2675
	unsigned long total_scanned;
L
Linus Torvalds 已提交
2676
	struct reclaim_state *reclaim_state = current->reclaim_state;
2677 2678
	unsigned long nr_soft_reclaimed;
	unsigned long nr_soft_scanned;
2679 2680
	struct scan_control sc = {
		.gfp_mask = GFP_KERNEL,
2681
		.may_unmap = 1,
2682
		.may_swap = 1,
2683 2684 2685 2686 2687
		/*
		 * 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,
A
Andy Whitcroft 已提交
2688
		.order = order,
2689
		.target_mem_cgroup = NULL,
2690
	};
2691 2692 2693
	struct shrink_control shrink = {
		.gfp_mask = sc.gfp_mask,
	};
L
Linus Torvalds 已提交
2694 2695
loop_again:
	total_scanned = 0;
2696
	sc.nr_reclaimed = 0;
C
Christoph Lameter 已提交
2697
	sc.may_writepage = !laptop_mode;
2698
	count_vm_event(PAGEOUTRUN);
L
Linus Torvalds 已提交
2699 2700 2701

	for (priority = DEF_PRIORITY; priority >= 0; priority--) {
		unsigned long lru_pages = 0;
2702
		int has_under_min_watermark_zone = 0;
L
Linus Torvalds 已提交
2703

2704 2705
		/* The swap token gets in the way of swapout... */
		if (!priority)
2706
			disable_swap_token(NULL);
2707

L
Linus Torvalds 已提交
2708
		all_zones_ok = 1;
2709
		balanced = 0;
L
Linus Torvalds 已提交
2710

2711 2712 2713 2714 2715 2716
		/*
		 * 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
2717

2718 2719
			if (!populated_zone(zone))
				continue;
L
Linus Torvalds 已提交
2720

2721
			if (zone->all_unreclaimable && priority != DEF_PRIORITY)
2722
				continue;
L
Linus Torvalds 已提交
2723

2724 2725 2726 2727
			/*
			 * Do some background aging of the anon list, to give
			 * pages a chance to be referenced before reclaiming.
			 */
2728
			age_active_anon(zone, &sc, priority);
2729

2730
			if (!zone_watermark_ok_safe(zone, order,
2731
					high_wmark_pages(zone), 0, 0)) {
2732
				end_zone = i;
A
Andrew Morton 已提交
2733
				break;
2734 2735 2736
			} else {
				/* If balanced, clear the congested flag */
				zone_clear_flag(zone, ZONE_CONGESTED);
L
Linus Torvalds 已提交
2737 2738
			}
		}
A
Andrew Morton 已提交
2739 2740 2741
		if (i < 0)
			goto out;

L
Linus Torvalds 已提交
2742 2743 2744
		for (i = 0; i <= end_zone; i++) {
			struct zone *zone = pgdat->node_zones + i;

2745
			lru_pages += zone_reclaimable_pages(zone);
L
Linus Torvalds 已提交
2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758
		}

		/*
		 * 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;
2759
			int nr_slab, testorder;
2760
			unsigned long balance_gap;
L
Linus Torvalds 已提交
2761

2762
			if (!populated_zone(zone))
L
Linus Torvalds 已提交
2763 2764
				continue;

2765
			if (zone->all_unreclaimable && priority != DEF_PRIORITY)
L
Linus Torvalds 已提交
2766 2767 2768
				continue;

			sc.nr_scanned = 0;
2769

2770
			nr_soft_scanned = 0;
2771 2772 2773
			/*
			 * Call soft limit reclaim before calling shrink_zone.
			 */
2774 2775 2776 2777 2778
			nr_soft_reclaimed = mem_cgroup_soft_limit_reclaim(zone,
							order, sc.gfp_mask,
							&nr_soft_scanned);
			sc.nr_reclaimed += nr_soft_reclaimed;
			total_scanned += nr_soft_scanned;
2779

2780
			/*
2781 2782 2783 2784 2785 2786
			 * We put equal pressure on every zone, unless
			 * one zone has way too many pages free
			 * already. The "too many pages" is defined
			 * as the high wmark plus a "gap" where the
			 * gap is either the low watermark or 1%
			 * of the zone, whichever is smaller.
2787
			 */
2788 2789 2790 2791
			balance_gap = min(low_wmark_pages(zone),
				(zone->present_pages +
					KSWAPD_ZONE_BALANCE_GAP_RATIO-1) /
				KSWAPD_ZONE_BALANCE_GAP_RATIO);
2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805
			/*
			 * Kswapd reclaims only single pages with compaction
			 * enabled. Trying too hard to reclaim until contiguous
			 * free pages have become available can hurt performance
			 * by evicting too much useful data from memory.
			 * Do not reclaim more than needed for compaction.
			 */
			testorder = order;
			if (COMPACTION_BUILD && order &&
					compaction_suitable(zone, order) !=
						COMPACT_SKIPPED)
				testorder = 0;

			if (!zone_watermark_ok_safe(zone, testorder,
2806
					high_wmark_pages(zone) + balance_gap,
2807
					end_zone, 0)) {
2808
				shrink_zone(priority, zone, &sc);
2809

2810 2811 2812 2813 2814 2815 2816 2817 2818
				reclaim_state->reclaimed_slab = 0;
				nr_slab = shrink_slab(&shrink, sc.nr_scanned, lru_pages);
				sc.nr_reclaimed += reclaim_state->reclaimed_slab;
				total_scanned += sc.nr_scanned;

				if (nr_slab == 0 && !zone_reclaimable(zone))
					zone->all_unreclaimable = 1;
			}

L
Linus Torvalds 已提交
2819 2820 2821 2822 2823 2824
			/*
			 * 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 &&
2825
			    total_scanned > sc.nr_reclaimed + sc.nr_reclaimed / 2)
L
Linus Torvalds 已提交
2826
				sc.may_writepage = 1;
2827

2828 2829 2830
			if (zone->all_unreclaimable) {
				if (end_zone && end_zone == i)
					end_zone--;
2831
				continue;
2832
			}
2833

2834
			if (!zone_watermark_ok_safe(zone, testorder,
2835 2836 2837 2838 2839 2840 2841
					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!
				 */
2842
				if (!zone_watermark_ok_safe(zone, order,
2843 2844
					    min_wmark_pages(zone), end_zone, 0))
					has_under_min_watermark_zone = 1;
2845 2846 2847 2848 2849 2850 2851 2852 2853
			} else {
				/*
				 * If a zone reaches its high watermark,
				 * consider it to be no longer congested. It's
				 * possible there are dirty pages backed by
				 * congested BDIs but as pressure is relieved,
				 * spectulatively avoid congestion waits
				 */
				zone_clear_flag(zone, ZONE_CONGESTED);
2854
				if (i <= *classzone_idx)
2855
					balanced += zone->present_pages;
2856
			}
2857

L
Linus Torvalds 已提交
2858
		}
2859
		if (all_zones_ok || (order && pgdat_balanced(pgdat, balanced, *classzone_idx)))
L
Linus Torvalds 已提交
2860 2861 2862 2863 2864
			break;		/* kswapd: all done */
		/*
		 * OK, kswapd is getting into trouble.  Take a nap, then take
		 * another pass across the zones.
		 */
2865 2866 2867 2868 2869 2870
		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
2871 2872 2873 2874 2875 2876 2877

		/*
		 * 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.
		 */
2878
		if (sc.nr_reclaimed >= SWAP_CLUSTER_MAX)
L
Linus Torvalds 已提交
2879 2880 2881
			break;
	}
out:
2882 2883 2884

	/*
	 * order-0: All zones must meet high watermark for a balanced node
2885 2886
	 * high-order: Balanced zones must make up at least 25% of the node
	 *             for the node to be balanced
2887
	 */
2888
	if (!(all_zones_ok || (order && pgdat_balanced(pgdat, balanced, *classzone_idx)))) {
L
Linus Torvalds 已提交
2889
		cond_resched();
2890 2891 2892

		try_to_freeze();

2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909
		/*
		 * 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
2910 2911 2912
		goto loop_again;
	}

2913 2914 2915 2916 2917 2918 2919 2920 2921
	/*
	 * If kswapd was reclaiming at a higher order, it has the option of
	 * sleeping without all zones being balanced. Before it does, it must
	 * ensure that the watermarks for order-0 on *all* zones are met and
	 * that the congestion flags are cleared. The congestion flag must
	 * be cleared as kswapd is the only mechanism that clears the flag
	 * and it is potentially going to sleep here.
	 */
	if (order) {
2922 2923
		int zones_need_compaction = 1;

2924 2925 2926 2927 2928 2929 2930 2931 2932
		for (i = 0; i <= end_zone; i++) {
			struct zone *zone = pgdat->node_zones + i;

			if (!populated_zone(zone))
				continue;

			if (zone->all_unreclaimable && priority != DEF_PRIORITY)
				continue;

2933 2934 2935 2936
			/* Would compaction fail due to lack of free memory? */
			if (compaction_suitable(zone, order) == COMPACT_SKIPPED)
				goto loop_again;

2937 2938 2939 2940 2941 2942 2943
			/* Confirm the zone is balanced for order-0 */
			if (!zone_watermark_ok(zone, 0,
					high_wmark_pages(zone), 0, 0)) {
				order = sc.order = 0;
				goto loop_again;
			}

2944 2945 2946 2947 2948
			/* Check if the memory needs to be defragmented. */
			if (zone_watermark_ok(zone, order,
				    low_wmark_pages(zone), *classzone_idx, 0))
				zones_need_compaction = 0;

2949 2950 2951
			/* If balanced, clear the congested flag */
			zone_clear_flag(zone, ZONE_CONGESTED);
		}
2952 2953 2954

		if (zones_need_compaction)
			compact_pgdat(pgdat, order);
2955 2956
	}

2957 2958 2959 2960 2961 2962
	/*
	 * Return the order we were reclaiming at so sleeping_prematurely()
	 * makes a decision on the order we were last reclaiming at. However,
	 * if another caller entered the allocator slow path while kswapd
	 * was awake, order will remain at the higher level
	 */
2963
	*classzone_idx = end_zone;
2964
	return order;
L
Linus Torvalds 已提交
2965 2966
}

2967
static void kswapd_try_to_sleep(pg_data_t *pgdat, int order, int classzone_idx)
2968 2969 2970 2971 2972 2973 2974 2975 2976 2977
{
	long remaining = 0;
	DEFINE_WAIT(wait);

	if (freezing(current) || kthread_should_stop())
		return;

	prepare_to_wait(&pgdat->kswapd_wait, &wait, TASK_INTERRUPTIBLE);

	/* Try to sleep for a short interval */
2978
	if (!sleeping_prematurely(pgdat, order, remaining, classzone_idx)) {
2979 2980 2981 2982 2983 2984 2985 2986 2987
		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.
	 */
2988
	if (!sleeping_prematurely(pgdat, order, remaining, classzone_idx)) {
2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010
		trace_mm_vmscan_kswapd_sleep(pgdat->node_id);

		/*
		 * vmstat counters are not perfectly accurate and the estimated
		 * value for counters such as NR_FREE_PAGES can deviate from the
		 * true value by nr_online_cpus * threshold. To avoid the zone
		 * watermarks being breached while under pressure, we reduce the
		 * per-cpu vmstat threshold while kswapd is awake and restore
		 * them before going back to sleep.
		 */
		set_pgdat_percpu_threshold(pgdat, calculate_normal_threshold);
		schedule();
		set_pgdat_percpu_threshold(pgdat, calculate_pressure_threshold);
	} else {
		if (remaining)
			count_vm_event(KSWAPD_LOW_WMARK_HIT_QUICKLY);
		else
			count_vm_event(KSWAPD_HIGH_WMARK_HIT_QUICKLY);
	}
	finish_wait(&pgdat->kswapd_wait, &wait);
}

L
Linus Torvalds 已提交
3011 3012
/*
 * The background pageout daemon, started as a kernel thread
3013
 * from the init process.
L
Linus Torvalds 已提交
3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025
 *
 * 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)
{
3026
	unsigned long order, new_order;
3027
	unsigned balanced_order;
3028
	int classzone_idx, new_classzone_idx;
3029
	int balanced_classzone_idx;
L
Linus Torvalds 已提交
3030 3031
	pg_data_t *pgdat = (pg_data_t*)p;
	struct task_struct *tsk = current;
3032

L
Linus Torvalds 已提交
3033 3034 3035
	struct reclaim_state reclaim_state = {
		.reclaimed_slab = 0,
	};
3036
	const struct cpumask *cpumask = cpumask_of_node(pgdat->node_id);
L
Linus Torvalds 已提交
3037

3038 3039
	lockdep_set_current_reclaim_state(GFP_KERNEL);

R
Rusty Russell 已提交
3040
	if (!cpumask_empty(cpumask))
3041
		set_cpus_allowed_ptr(tsk, cpumask);
L
Linus Torvalds 已提交
3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055
	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).
	 */
3056
	tsk->flags |= PF_MEMALLOC | PF_SWAPWRITE | PF_KSWAPD;
3057
	set_freezable();
L
Linus Torvalds 已提交
3058

3059
	order = new_order = 0;
3060
	balanced_order = 0;
3061
	classzone_idx = new_classzone_idx = pgdat->nr_zones - 1;
3062
	balanced_classzone_idx = classzone_idx;
L
Linus Torvalds 已提交
3063
	for ( ; ; ) {
3064
		int ret;
3065

3066 3067 3068 3069 3070
		/*
		 * If the last balance_pgdat was unsuccessful it's unlikely a
		 * new request of a similar or harder type will succeed soon
		 * so consider going to sleep on the basis we reclaimed at
		 */
3071 3072
		if (balanced_classzone_idx >= new_classzone_idx &&
					balanced_order == new_order) {
3073 3074 3075 3076 3077 3078
			new_order = pgdat->kswapd_max_order;
			new_classzone_idx = pgdat->classzone_idx;
			pgdat->kswapd_max_order =  0;
			pgdat->classzone_idx = pgdat->nr_zones - 1;
		}

3079
		if (order < new_order || classzone_idx > new_classzone_idx) {
L
Linus Torvalds 已提交
3080 3081
			/*
			 * Don't sleep if someone wants a larger 'order'
3082
			 * allocation or has tigher zone constraints
L
Linus Torvalds 已提交
3083 3084
			 */
			order = new_order;
3085
			classzone_idx = new_classzone_idx;
L
Linus Torvalds 已提交
3086
		} else {
3087 3088
			kswapd_try_to_sleep(pgdat, balanced_order,
						balanced_classzone_idx);
L
Linus Torvalds 已提交
3089
			order = pgdat->kswapd_max_order;
3090
			classzone_idx = pgdat->classzone_idx;
3091 3092
			new_order = order;
			new_classzone_idx = classzone_idx;
3093
			pgdat->kswapd_max_order = 0;
3094
			pgdat->classzone_idx = pgdat->nr_zones - 1;
L
Linus Torvalds 已提交
3095 3096
		}

3097 3098 3099 3100 3101 3102 3103 3104
		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
		 */
3105 3106
		if (!ret) {
			trace_mm_vmscan_kswapd_wake(pgdat->node_id, order);
3107 3108 3109
			balanced_classzone_idx = classzone_idx;
			balanced_order = balance_pgdat(pgdat, order,
						&balanced_classzone_idx);
3110
		}
L
Linus Torvalds 已提交
3111 3112 3113 3114 3115 3116 3117
	}
	return 0;
}

/*
 * A zone is low on free memory, so wake its kswapd task to service it.
 */
3118
void wakeup_kswapd(struct zone *zone, int order, enum zone_type classzone_idx)
L
Linus Torvalds 已提交
3119 3120 3121
{
	pg_data_t *pgdat;

3122
	if (!populated_zone(zone))
L
Linus Torvalds 已提交
3123 3124
		return;

3125
	if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL))
L
Linus Torvalds 已提交
3126
		return;
3127
	pgdat = zone->zone_pgdat;
3128
	if (pgdat->kswapd_max_order < order) {
L
Linus Torvalds 已提交
3129
		pgdat->kswapd_max_order = order;
3130 3131
		pgdat->classzone_idx = min(pgdat->classzone_idx, classzone_idx);
	}
3132
	if (!waitqueue_active(&pgdat->kswapd_wait))
L
Linus Torvalds 已提交
3133
		return;
3134 3135 3136 3137
	if (zone_watermark_ok_safe(zone, order, low_wmark_pages(zone), 0, 0))
		return;

	trace_mm_vmscan_wakeup_kswapd(pgdat->node_id, zone_idx(zone), order);
3138
	wake_up_interruptible(&pgdat->kswapd_wait);
L
Linus Torvalds 已提交
3139 3140
}

3141 3142 3143 3144 3145 3146 3147 3148
/*
 * 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)
3149
{
3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173
	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;
3174 3175
}

3176
#ifdef CONFIG_HIBERNATION
L
Linus Torvalds 已提交
3177
/*
3178
 * Try to free `nr_to_reclaim' of memory, system-wide, and return the number of
3179 3180 3181 3182 3183
 * 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
3184
 */
3185
unsigned long shrink_all_memory(unsigned long nr_to_reclaim)
L
Linus Torvalds 已提交
3186
{
3187 3188
	struct reclaim_state reclaim_state;
	struct scan_control sc = {
3189 3190 3191
		.gfp_mask = GFP_HIGHUSER_MOVABLE,
		.may_swap = 1,
		.may_unmap = 1,
3192
		.may_writepage = 1,
3193 3194 3195
		.nr_to_reclaim = nr_to_reclaim,
		.hibernation_mode = 1,
		.order = 0,
L
Linus Torvalds 已提交
3196
	};
3197 3198 3199 3200
	struct shrink_control shrink = {
		.gfp_mask = sc.gfp_mask,
	};
	struct zonelist *zonelist = node_zonelist(numa_node_id(), sc.gfp_mask);
3201 3202
	struct task_struct *p = current;
	unsigned long nr_reclaimed;
L
Linus Torvalds 已提交
3203

3204 3205 3206 3207
	p->flags |= PF_MEMALLOC;
	lockdep_set_current_reclaim_state(sc.gfp_mask);
	reclaim_state.reclaimed_slab = 0;
	p->reclaim_state = &reclaim_state;
3208

3209
	nr_reclaimed = do_try_to_free_pages(zonelist, &sc, &shrink);
3210

3211 3212 3213
	p->reclaim_state = NULL;
	lockdep_clear_current_reclaim_state();
	p->flags &= ~PF_MEMALLOC;
3214

3215
	return nr_reclaimed;
L
Linus Torvalds 已提交
3216
}
3217
#endif /* CONFIG_HIBERNATION */
L
Linus Torvalds 已提交
3218 3219 3220 3221 3222

/* 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. */
3223
static int __devinit cpu_callback(struct notifier_block *nfb,
3224
				  unsigned long action, void *hcpu)
L
Linus Torvalds 已提交
3225
{
3226
	int nid;
L
Linus Torvalds 已提交
3227

3228
	if (action == CPU_ONLINE || action == CPU_ONLINE_FROZEN) {
3229
		for_each_node_state(nid, N_HIGH_MEMORY) {
3230
			pg_data_t *pgdat = NODE_DATA(nid);
3231 3232 3233
			const struct cpumask *mask;

			mask = cpumask_of_node(pgdat->node_id);
3234

3235
			if (cpumask_any_and(cpu_online_mask, mask) < nr_cpu_ids)
L
Linus Torvalds 已提交
3236
				/* One of our CPUs online: restore mask */
3237
				set_cpus_allowed_ptr(pgdat->kswapd, mask);
L
Linus Torvalds 已提交
3238 3239 3240 3241 3242
		}
	}
	return NOTIFY_OK;
}

3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264
/*
 * 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;
}

3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275
/*
 * 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
3276 3277
static int __init kswapd_init(void)
{
3278
	int nid;
3279

L
Linus Torvalds 已提交
3280
	swap_setup();
3281
	for_each_node_state(nid, N_HIGH_MEMORY)
3282
 		kswapd_run(nid);
L
Linus Torvalds 已提交
3283 3284 3285 3286 3287
	hotcpu_notifier(cpu_callback, 0);
	return 0;
}

module_init(kswapd_init)
3288 3289 3290 3291 3292 3293 3294 3295 3296 3297

#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;

3298
#define RECLAIM_OFF 0
3299
#define RECLAIM_ZONE (1<<0)	/* Run shrink_inactive_list on the zone */
3300 3301 3302
#define RECLAIM_WRITE (1<<1)	/* Writeout pages during reclaim */
#define RECLAIM_SWAP (1<<2)	/* Swap pages out during reclaim */

3303 3304 3305 3306 3307 3308 3309
/*
 * 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

3310 3311 3312 3313 3314 3315
/*
 * Percentage of pages in a zone that must be unmapped for zone_reclaim to
 * occur.
 */
int sysctl_min_unmapped_ratio = 1;

3316 3317 3318 3319 3320 3321
/*
 * 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;

3322 3323 3324 3325 3326 3327 3328 3329 3330 3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363
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;
}

3364 3365 3366
/*
 * Try to free up some pages from this zone through reclaim.
 */
3367
static int __zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order)
3368
{
3369
	/* Minimum pages needed in order to stay on node */
3370
	const unsigned long nr_pages = 1 << order;
3371 3372
	struct task_struct *p = current;
	struct reclaim_state reclaim_state;
3373
	int priority;
3374 3375
	struct scan_control sc = {
		.may_writepage = !!(zone_reclaim_mode & RECLAIM_WRITE),
3376
		.may_unmap = !!(zone_reclaim_mode & RECLAIM_SWAP),
3377
		.may_swap = 1,
3378 3379
		.nr_to_reclaim = max_t(unsigned long, nr_pages,
				       SWAP_CLUSTER_MAX),
3380
		.gfp_mask = gfp_mask,
3381
		.order = order,
3382
	};
3383 3384 3385
	struct shrink_control shrink = {
		.gfp_mask = sc.gfp_mask,
	};
3386
	unsigned long nr_slab_pages0, nr_slab_pages1;
3387 3388

	cond_resched();
3389 3390 3391 3392 3393 3394
	/*
	 * 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;
3395
	lockdep_set_current_reclaim_state(gfp_mask);
3396 3397
	reclaim_state.reclaimed_slab = 0;
	p->reclaim_state = &reclaim_state;
3398

3399
	if (zone_pagecache_reclaimable(zone) > zone->min_unmapped_pages) {
3400 3401 3402 3403 3404 3405
		/*
		 * Free memory by calling shrink zone with increasing
		 * priorities until we have enough memory freed.
		 */
		priority = ZONE_RECLAIM_PRIORITY;
		do {
3406
			shrink_zone(priority, zone, &sc);
3407
			priority--;
3408
		} while (priority >= 0 && sc.nr_reclaimed < nr_pages);
3409
	}
3410

3411 3412
	nr_slab_pages0 = zone_page_state(zone, NR_SLAB_RECLAIMABLE);
	if (nr_slab_pages0 > zone->min_slab_pages) {
3413
		/*
3414
		 * shrink_slab() does not currently allow us to determine how
3415 3416 3417 3418
		 * 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.
3419
		 *
3420 3421
		 * Note that shrink_slab will free memory on all zones and may
		 * take a long time.
3422
		 */
3423 3424 3425 3426
		for (;;) {
			unsigned long lru_pages = zone_reclaimable_pages(zone);

			/* No reclaimable slab or very low memory pressure */
3427
			if (!shrink_slab(&shrink, sc.nr_scanned, lru_pages))
3428 3429 3430 3431 3432 3433 3434 3435
				break;

			/* Freed enough memory */
			nr_slab_pages1 = zone_page_state(zone,
							NR_SLAB_RECLAIMABLE);
			if (nr_slab_pages1 + nr_pages <= nr_slab_pages0)
				break;
		}
3436 3437 3438 3439 3440

		/*
		 * Update nr_reclaimed by the number of slab pages we
		 * reclaimed from this zone.
		 */
3441 3442 3443
		nr_slab_pages1 = zone_page_state(zone, NR_SLAB_RECLAIMABLE);
		if (nr_slab_pages1 < nr_slab_pages0)
			sc.nr_reclaimed += nr_slab_pages0 - nr_slab_pages1;
3444 3445
	}

3446
	p->reclaim_state = NULL;
3447
	current->flags &= ~(PF_MEMALLOC | PF_SWAPWRITE);
3448
	lockdep_clear_current_reclaim_state();
3449
	return sc.nr_reclaimed >= nr_pages;
3450
}
3451 3452 3453 3454

int zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order)
{
	int node_id;
3455
	int ret;
3456 3457

	/*
3458 3459
	 * Zone reclaim reclaims unmapped file backed pages and
	 * slab pages if we are over the defined limits.
3460
	 *
3461 3462 3463 3464 3465
	 * 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.
3466
	 */
3467 3468
	if (zone_pagecache_reclaimable(zone) <= zone->min_unmapped_pages &&
	    zone_page_state(zone, NR_SLAB_RECLAIMABLE) <= zone->min_slab_pages)
3469
		return ZONE_RECLAIM_FULL;
3470

3471
	if (zone->all_unreclaimable)
3472
		return ZONE_RECLAIM_FULL;
3473

3474
	/*
3475
	 * Do not scan if the allocation should not be delayed.
3476
	 */
3477
	if (!(gfp_mask & __GFP_WAIT) || (current->flags & PF_MEMALLOC))
3478
		return ZONE_RECLAIM_NOSCAN;
3479 3480 3481 3482 3483 3484 3485

	/*
	 * 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.
	 */
3486
	node_id = zone_to_nid(zone);
3487
	if (node_state(node_id, N_CPU) && node_id != numa_node_id())
3488
		return ZONE_RECLAIM_NOSCAN;
3489 3490

	if (zone_test_and_set_flag(zone, ZONE_RECLAIM_LOCKED))
3491 3492
		return ZONE_RECLAIM_NOSCAN;

3493 3494 3495
	ret = __zone_reclaim(zone, gfp_mask, order);
	zone_clear_flag(zone, ZONE_RECLAIM_LOCKED);

3496 3497 3498
	if (!ret)
		count_vm_event(PGSCAN_ZONE_RECLAIM_FAILED);

3499
	return ret;
3500
}
3501
#endif
L
Lee Schermerhorn 已提交
3502 3503 3504 3505 3506 3507 3508

/*
 * 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
3509 3510
 * 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
3511 3512
 *
 * Reasons page might not be evictable:
3513
 * (1) page's mapping marked unevictable
N
Nick Piggin 已提交
3514
 * (2) page is part of an mlocked VMA
3515
 *
L
Lee Schermerhorn 已提交
3516 3517 3518 3519
 */
int page_evictable(struct page *page, struct vm_area_struct *vma)
{

3520 3521 3522
	if (mapping_unevictable(page_mapping(page)))
		return 0;

N
Nick Piggin 已提交
3523 3524
	if (PageMlocked(page) || (vma && is_mlocked_vma(vma, page)))
		return 0;
L
Lee Schermerhorn 已提交
3525 3526 3527

	return 1;
}
3528

3529
#ifdef CONFIG_SHMEM
3530
/**
3531 3532 3533
 * check_move_unevictable_pages - check pages for evictability and move to appropriate zone lru list
 * @pages:	array of pages to check
 * @nr_pages:	number of pages to check
3534
 *
3535
 * Checks pages for evictability and moves them to the appropriate lru list.
3536 3537
 *
 * This function is only used for SysV IPC SHM_UNLOCK.
3538
 */
3539
void check_move_unevictable_pages(struct page **pages, int nr_pages)
3540
{
3541
	struct lruvec *lruvec;
3542 3543 3544 3545
	struct zone *zone = NULL;
	int pgscanned = 0;
	int pgrescued = 0;
	int i;
3546

3547 3548 3549
	for (i = 0; i < nr_pages; i++) {
		struct page *page = pages[i];
		struct zone *pagezone;
3550

3551 3552 3553 3554 3555 3556 3557 3558
		pgscanned++;
		pagezone = page_zone(page);
		if (pagezone != zone) {
			if (zone)
				spin_unlock_irq(&zone->lru_lock);
			zone = pagezone;
			spin_lock_irq(&zone->lru_lock);
		}
3559

3560 3561
		if (!PageLRU(page) || !PageUnevictable(page))
			continue;
3562

3563 3564 3565 3566 3567 3568 3569 3570 3571 3572 3573
		if (page_evictable(page, NULL)) {
			enum lru_list lru = page_lru_base_type(page);

			VM_BUG_ON(PageActive(page));
			ClearPageUnevictable(page);
			__dec_zone_state(zone, NR_UNEVICTABLE);
			lruvec = mem_cgroup_lru_move_lists(zone, page,
						LRU_UNEVICTABLE, lru);
			list_move(&page->lru, &lruvec->lists[lru]);
			__inc_zone_state(zone, NR_INACTIVE_ANON + lru);
			pgrescued++;
3574
		}
3575
	}
3576

3577 3578 3579 3580
	if (zone) {
		__count_vm_events(UNEVICTABLE_PGRESCUED, pgrescued);
		__count_vm_events(UNEVICTABLE_PGSCANNED, pgscanned);
		spin_unlock_irq(&zone->lru_lock);
3581 3582
	}
}
3583
#endif /* CONFIG_SHMEM */
3584

3585
static void warn_scan_unevictable_pages(void)
3586
{
3587
	printk_once(KERN_WARNING
3588
		    "%s: The scan_unevictable_pages sysctl/node-interface has been "
3589
		    "disabled for lack of a legitimate use case.  If you have "
3590 3591
		    "one, please send an email to linux-mm@kvack.org.\n",
		    current->comm);
3592 3593 3594 3595 3596 3597 3598 3599 3600
}

/*
 * 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,
3601
			   void __user *buffer,
3602 3603
			   size_t *length, loff_t *ppos)
{
3604
	warn_scan_unevictable_pages();
3605
	proc_doulongvec_minmax(table, write, buffer, length, ppos);
3606 3607 3608 3609
	scan_unevictable_pages = 0;
	return 0;
}

3610
#ifdef CONFIG_NUMA
3611 3612 3613 3614 3615
/*
 * per node 'scan_unevictable_pages' attribute.  On demand re-scan of
 * a specified node's per zone unevictable lists for evictable pages.
 */

3616 3617
static ssize_t read_scan_unevictable_node(struct device *dev,
					  struct device_attribute *attr,
3618 3619
					  char *buf)
{
3620
	warn_scan_unevictable_pages();
3621 3622 3623
	return sprintf(buf, "0\n");	/* always zero; should fit... */
}

3624 3625
static ssize_t write_scan_unevictable_node(struct device *dev,
					   struct device_attribute *attr,
3626 3627
					const char *buf, size_t count)
{
3628
	warn_scan_unevictable_pages();
3629 3630 3631 3632
	return 1;
}


3633
static DEVICE_ATTR(scan_unevictable_pages, S_IRUGO | S_IWUSR,
3634 3635 3636 3637 3638
			read_scan_unevictable_node,
			write_scan_unevictable_node);

int scan_unevictable_register_node(struct node *node)
{
3639
	return device_create_file(&node->dev, &dev_attr_scan_unevictable_pages);
3640 3641 3642 3643
}

void scan_unevictable_unregister_node(struct node *node)
{
3644
	device_remove_file(&node->dev, &dev_attr_scan_unevictable_pages);
3645
}
3646
#endif