compaction.c 34 KB
Newer Older
1
2
3
4
5
6
7
8
9
10
11
12
13
14
/*
 * linux/mm/compaction.c
 *
 * Memory compaction for the reduction of external fragmentation. Note that
 * this heavily depends upon page migration to do all the real heavy
 * lifting
 *
 * Copyright IBM Corp. 2007-2010 Mel Gorman <mel@csn.ul.ie>
 */
#include <linux/swap.h>
#include <linux/migrate.h>
#include <linux/compaction.h>
#include <linux/mm_inline.h>
#include <linux/backing-dev.h>
15
#include <linux/sysctl.h>
16
#include <linux/sysfs.h>
17
#include <linux/balloon_compaction.h>
18
#include <linux/page-isolation.h>
19
20
#include "internal.h"

21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
#ifdef CONFIG_COMPACTION
static inline void count_compact_event(enum vm_event_item item)
{
	count_vm_event(item);
}

static inline void count_compact_events(enum vm_event_item item, long delta)
{
	count_vm_events(item, delta);
}
#else
#define count_compact_event(item) do { } while (0)
#define count_compact_events(item, delta) do { } while (0)
#endif

36
37
#if defined CONFIG_COMPACTION || defined CONFIG_CMA

38
39
40
#define CREATE_TRACE_POINTS
#include <trace/events/compaction.h>

41
42
43
44
45
46
47
48
49
50
51
52
53
54
static unsigned long release_freepages(struct list_head *freelist)
{
	struct page *page, *next;
	unsigned long count = 0;

	list_for_each_entry_safe(page, next, freelist, lru) {
		list_del(&page->lru);
		__free_page(page);
		count++;
	}

	return count;
}

55
56
57
58
59
60
61
62
63
64
static void map_pages(struct list_head *list)
{
	struct page *page;

	list_for_each_entry(page, list, lru) {
		arch_alloc_page(page, 0);
		kernel_map_pages(page, 1, 1);
	}
}

65
66
67
68
69
static inline bool migrate_async_suitable(int migratetype)
{
	return is_migrate_cma(migratetype) || migratetype == MIGRATE_MOVABLE;
}

70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
#ifdef CONFIG_COMPACTION
/* Returns true if the pageblock should be scanned for pages to isolate. */
static inline bool isolation_suitable(struct compact_control *cc,
					struct page *page)
{
	if (cc->ignore_skip_hint)
		return true;

	return !get_pageblock_skip(page);
}

/*
 * This function is called to clear all cached information on pageblocks that
 * should be skipped for page isolation when the migrate and free page scanner
 * meet.
 */
86
static void __reset_isolation_suitable(struct zone *zone)
87
88
{
	unsigned long start_pfn = zone->zone_start_pfn;
89
	unsigned long end_pfn = zone_end_pfn(zone);
90
91
	unsigned long pfn;

92
93
	zone->compact_cached_migrate_pfn[0] = start_pfn;
	zone->compact_cached_migrate_pfn[1] = start_pfn;
94
	zone->compact_cached_free_pfn = end_pfn;
95
	zone->compact_blockskip_flush = false;
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113

	/* Walk the zone and mark every pageblock as suitable for isolation */
	for (pfn = start_pfn; pfn < end_pfn; pfn += pageblock_nr_pages) {
		struct page *page;

		cond_resched();

		if (!pfn_valid(pfn))
			continue;

		page = pfn_to_page(pfn);
		if (zone != page_zone(page))
			continue;

		clear_pageblock_skip(page);
	}
}

114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
void reset_isolation_suitable(pg_data_t *pgdat)
{
	int zoneid;

	for (zoneid = 0; zoneid < MAX_NR_ZONES; zoneid++) {
		struct zone *zone = &pgdat->node_zones[zoneid];
		if (!populated_zone(zone))
			continue;

		/* Only flush if a full compaction finished recently */
		if (zone->compact_blockskip_flush)
			__reset_isolation_suitable(zone);
	}
}

129
130
/*
 * If no pages were isolated then mark this pageblock to be skipped in the
131
 * future. The information is later cleared by __reset_isolation_suitable().
132
 */
133
134
static void update_pageblock_skip(struct compact_control *cc,
			struct page *page, unsigned long nr_isolated,
135
			bool set_unsuitable, bool migrate_scanner)
136
{
137
	struct zone *zone = cc->zone;
138
	unsigned long pfn;
139
140
141
142

	if (cc->ignore_skip_hint)
		return;

143
144
145
	if (!page)
		return;

146
147
148
149
150
151
152
153
	if (nr_isolated)
		return;

	/*
	 * Only skip pageblocks when all forms of compaction will be known to
	 * fail in the near future.
	 */
	if (set_unsuitable)
154
		set_pageblock_skip(page);
155

156
157
158
159
160
161
162
163
	pfn = page_to_pfn(page);

	/* Update where async and sync compaction should restart */
	if (migrate_scanner) {
		if (cc->finished_update_migrate)
			return;
		if (pfn > zone->compact_cached_migrate_pfn[0])
			zone->compact_cached_migrate_pfn[0] = pfn;
164
165
		if (cc->mode != MIGRATE_ASYNC &&
		    pfn > zone->compact_cached_migrate_pfn[1])
166
167
168
169
170
171
			zone->compact_cached_migrate_pfn[1] = pfn;
	} else {
		if (cc->finished_update_free)
			return;
		if (pfn < zone->compact_cached_free_pfn)
			zone->compact_cached_free_pfn = pfn;
172
	}
173
174
175
176
177
178
179
180
}
#else
static inline bool isolation_suitable(struct compact_control *cc,
					struct page *page)
{
	return true;
}

181
182
static void update_pageblock_skip(struct compact_control *cc,
			struct page *page, unsigned long nr_isolated,
183
			bool set_unsuitable, bool migrate_scanner)
184
185
186
187
{
}
#endif /* CONFIG_COMPACTION */

188
189
190
191
192
static inline bool should_release_lock(spinlock_t *lock)
{
	return need_resched() || spin_is_contended(lock);
}

193
194
195
196
197
198
199
200
201
202
203
204
/*
 * Compaction requires the taking of some coarse locks that are potentially
 * very heavily contended. Check if the process needs to be scheduled or
 * if the lock is contended. For async compaction, back out in the event
 * if contention is severe. For sync compaction, schedule.
 *
 * Returns true if the lock is held.
 * Returns false if the lock is released and compaction should abort
 */
static bool compact_checklock_irqsave(spinlock_t *lock, unsigned long *flags,
				      bool locked, struct compact_control *cc)
{
205
	if (should_release_lock(lock)) {
206
207
208
209
210
211
		if (locked) {
			spin_unlock_irqrestore(lock, *flags);
			locked = false;
		}

		/* async aborts if taking too long or contended */
212
		if (cc->mode == MIGRATE_ASYNC) {
213
			cc->contended = true;
214
215
216
217
218
219
220
221
222
223
224
			return false;
		}

		cond_resched();
	}

	if (!locked)
		spin_lock_irqsave(lock, *flags);
	return true;
}

225
226
227
/* Returns true if the page is within a block suitable for migration to */
static bool suitable_migration_target(struct page *page)
{
228
	/* If the page is a large free page, then disallow migration */
229
	if (PageBuddy(page) && page_order(page) >= pageblock_order)
230
		return false;
231
232

	/* If the block is MIGRATE_MOVABLE or MIGRATE_CMA, allow migration */
233
	if (migrate_async_suitable(get_pageblock_migratetype(page)))
234
235
236
237
238
239
		return true;

	/* Otherwise skip the block */
	return false;
}

240
/*
241
242
243
 * Isolate free pages onto a private freelist. If @strict is true, will abort
 * returning 0 on any invalid PFNs or non-free pages inside of the pageblock
 * (even though it may still end up isolating some pages).
244
 */
245
246
static unsigned long isolate_freepages_block(struct compact_control *cc,
				unsigned long blockpfn,
247
248
249
				unsigned long end_pfn,
				struct list_head *freelist,
				bool strict)
250
{
251
	int nr_scanned = 0, total_isolated = 0;
252
	struct page *cursor, *valid_page = NULL;
253
254
	unsigned long flags;
	bool locked = false;
255
	bool checked_pageblock = false;
256
257
258

	cursor = pfn_to_page(blockpfn);

259
	/* Isolate free pages. */
260
261
262
263
	for (; blockpfn < end_pfn; blockpfn++, cursor++) {
		int isolated, i;
		struct page *page = cursor;

264
		nr_scanned++;
265
		if (!pfn_valid_within(blockpfn))
266
267
			goto isolate_fail;

268
269
		if (!valid_page)
			valid_page = page;
270
		if (!PageBuddy(page))
271
			goto isolate_fail;
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286

		/*
		 * The zone lock must be held to isolate freepages.
		 * Unfortunately this is a very coarse lock and can be
		 * heavily contended if there are parallel allocations
		 * or parallel compactions. For async compaction do not
		 * spin on the lock and we acquire the lock as late as
		 * possible.
		 */
		locked = compact_checklock_irqsave(&cc->zone->lock, &flags,
								locked, cc);
		if (!locked)
			break;

		/* Recheck this is a suitable migration target under lock */
287
288
289
290
291
292
293
294
295
296
		if (!strict && !checked_pageblock) {
			/*
			 * We need to check suitability of pageblock only once
			 * and this isolate_freepages_block() is called with
			 * pageblock range, so just check once is sufficient.
			 */
			checked_pageblock = true;
			if (!suitable_migration_target(page))
				break;
		}
297

298
299
		/* Recheck this is a buddy page under lock */
		if (!PageBuddy(page))
300
			goto isolate_fail;
301
302
303
304
305
306
307
308
309
310
311
312
313

		/* Found a free page, break it into order-0 pages */
		isolated = split_free_page(page);
		total_isolated += isolated;
		for (i = 0; i < isolated; i++) {
			list_add(&page->lru, freelist);
			page++;
		}

		/* If a page was split, advance to the end of it */
		if (isolated) {
			blockpfn += isolated - 1;
			cursor += isolated - 1;
314
			continue;
315
		}
316
317
318
319
320
321
322

isolate_fail:
		if (strict)
			break;
		else
			continue;

323
324
	}

325
	trace_mm_compaction_isolate_freepages(nr_scanned, total_isolated);
326
327
328
329
330
331

	/*
	 * If strict isolation is requested by CMA then check that all the
	 * pages requested were isolated. If there were any failures, 0 is
	 * returned and CMA will fail.
	 */
332
	if (strict && blockpfn < end_pfn)
333
334
335
336
337
		total_isolated = 0;

	if (locked)
		spin_unlock_irqrestore(&cc->zone->lock, flags);

338
339
	/* Update the pageblock-skip if the whole pageblock was scanned */
	if (blockpfn == end_pfn)
340
341
		update_pageblock_skip(cc, valid_page, total_isolated, true,
				      false);
342

343
	count_compact_events(COMPACTFREE_SCANNED, nr_scanned);
344
	if (total_isolated)
345
		count_compact_events(COMPACTISOLATED, total_isolated);
346
347
348
	return total_isolated;
}

349
350
351
352
353
354
355
356
357
358
359
360
361
/**
 * isolate_freepages_range() - isolate free pages.
 * @start_pfn: The first PFN to start isolating.
 * @end_pfn:   The one-past-last PFN.
 *
 * Non-free pages, invalid PFNs, or zone boundaries within the
 * [start_pfn, end_pfn) range are considered errors, cause function to
 * undo its actions and return zero.
 *
 * Otherwise, function returns one-past-the-last PFN of isolated page
 * (which may be greater then end_pfn if end fell in a middle of
 * a free page).
 */
362
unsigned long
363
364
isolate_freepages_range(struct compact_control *cc,
			unsigned long start_pfn, unsigned long end_pfn)
365
{
366
	unsigned long isolated, pfn, block_end_pfn;
367
368
369
	LIST_HEAD(freelist);

	for (pfn = start_pfn; pfn < end_pfn; pfn += isolated) {
370
		if (!pfn_valid(pfn) || cc->zone != page_zone(pfn_to_page(pfn)))
371
372
373
374
375
376
377
378
379
			break;

		/*
		 * On subsequent iterations ALIGN() is actually not needed,
		 * but we keep it that we not to complicate the code.
		 */
		block_end_pfn = ALIGN(pfn + 1, pageblock_nr_pages);
		block_end_pfn = min(block_end_pfn, end_pfn);

380
		isolated = isolate_freepages_block(cc, pfn, block_end_pfn,
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
						   &freelist, true);

		/*
		 * In strict mode, isolate_freepages_block() returns 0 if
		 * there are any holes in the block (ie. invalid PFNs or
		 * non-free pages).
		 */
		if (!isolated)
			break;

		/*
		 * If we managed to isolate pages, it is always (1 << n) *
		 * pageblock_nr_pages for some non-negative n.  (Max order
		 * page may span two pageblocks).
		 */
	}

	/* split_free_page does not map the pages */
	map_pages(&freelist);

	if (pfn < end_pfn) {
		/* Loop terminated early, cleanup. */
		release_freepages(&freelist);
		return 0;
	}

	/* We don't use freelists for anything. */
	return pfn;
}

411
/* Update the number of anon and file isolated pages in the zone */
412
static void acct_isolated(struct zone *zone, bool locked, struct compact_control *cc)
413
414
{
	struct page *page;
415
	unsigned int count[2] = { 0, };
416

417
418
	list_for_each_entry(page, &cc->migratepages, lru)
		count[!!page_is_file_cache(page)]++;
419

420
421
422
423
424
425
426
427
	/* If locked we can use the interrupt unsafe versions */
	if (locked) {
		__mod_zone_page_state(zone, NR_ISOLATED_ANON, count[0]);
		__mod_zone_page_state(zone, NR_ISOLATED_FILE, count[1]);
	} else {
		mod_zone_page_state(zone, NR_ISOLATED_ANON, count[0]);
		mod_zone_page_state(zone, NR_ISOLATED_FILE, count[1]);
	}
428
429
430
431
432
}

/* Similar to reclaim, but different enough that they don't share logic */
static bool too_many_isolated(struct zone *zone)
{
433
	unsigned long active, inactive, isolated;
434
435
436

	inactive = zone_page_state(zone, NR_INACTIVE_FILE) +
					zone_page_state(zone, NR_INACTIVE_ANON);
437
438
	active = zone_page_state(zone, NR_ACTIVE_FILE) +
					zone_page_state(zone, NR_ACTIVE_ANON);
439
440
441
	isolated = zone_page_state(zone, NR_ISOLATED_FILE) +
					zone_page_state(zone, NR_ISOLATED_ANON);

442
	return isolated > (inactive + active) / 2;
443
444
}

445
446
447
448
449
450
/**
 * isolate_migratepages_range() - isolate all migrate-able pages in range.
 * @zone:	Zone pages are in.
 * @cc:		Compaction control structure.
 * @low_pfn:	The first PFN of the range.
 * @end_pfn:	The one-past-the-last PFN of the range.
Minchan Kim's avatar
Minchan Kim committed
451
 * @unevictable: true if it allows to isolate unevictable pages
452
453
454
455
456
457
458
459
460
461
462
463
 *
 * Isolate all pages that can be migrated from the range specified by
 * [low_pfn, end_pfn).  Returns zero if there is a fatal signal
 * pending), otherwise PFN of the first page that was not scanned
 * (which may be both less, equal to or more then end_pfn).
 *
 * Assumes that cc->migratepages is empty and cc->nr_migratepages is
 * zero.
 *
 * Apart from cc->migratepages and cc->nr_migratetypes this function
 * does not modify any cc's fields, in particular it does not modify
 * (or read for that matter) cc->migrate_pfn.
464
 */
465
unsigned long
466
isolate_migratepages_range(struct zone *zone, struct compact_control *cc,
Minchan Kim's avatar
Minchan Kim committed
467
		unsigned long low_pfn, unsigned long end_pfn, bool unevictable)
468
{
469
	unsigned long last_pageblock_nr = 0, pageblock_nr;
470
	unsigned long nr_scanned = 0, nr_isolated = 0;
471
	struct list_head *migratelist = &cc->migratepages;
472
	struct lruvec *lruvec;
473
	unsigned long flags;
474
	bool locked = false;
475
	struct page *page = NULL, *valid_page = NULL;
476
	bool set_unsuitable = true;
477
478
	const isolate_mode_t mode = (cc->mode == MIGRATE_ASYNC ?
					ISOLATE_ASYNC_MIGRATE : 0) |
479
				    (unevictable ? ISOLATE_UNEVICTABLE : 0);
480
481
482
483
484
485
486

	/*
	 * Ensure that there are not too many pages isolated from the LRU
	 * list by either parallel reclaimers or compaction. If there are,
	 * delay for some time until fewer pages are isolated
	 */
	while (unlikely(too_many_isolated(zone))) {
487
		/* async migration should just abort */
488
		if (cc->mode == MIGRATE_ASYNC)
489
			return 0;
490

491
492
493
		congestion_wait(BLK_RW_ASYNC, HZ/10);

		if (fatal_signal_pending(current))
494
			return 0;
495
496
	}

497
498
499
500
501
502
	if (cond_resched()) {
		/* Async terminates prematurely on need_resched() */
		if (cc->mode == MIGRATE_ASYNC)
			return 0;
	}

503
504
	/* Time to isolate some pages for migration */
	for (; low_pfn < end_pfn; low_pfn++) {
505
		/* give a chance to irqs before checking need_resched() */
506
		if (locked && !(low_pfn % SWAP_CLUSTER_MAX)) {
507
508
509
510
			if (should_release_lock(&zone->lru_lock)) {
				spin_unlock_irqrestore(&zone->lru_lock, flags);
				locked = false;
			}
511
		}
512

513
514
515
516
517
518
519
520
521
522
523
524
525
		/*
		 * migrate_pfn does not necessarily start aligned to a
		 * pageblock. Ensure that pfn_valid is called when moving
		 * into a new MAX_ORDER_NR_PAGES range in case of large
		 * memory holes within the zone
		 */
		if ((low_pfn & (MAX_ORDER_NR_PAGES - 1)) == 0) {
			if (!pfn_valid(low_pfn)) {
				low_pfn += MAX_ORDER_NR_PAGES - 1;
				continue;
			}
		}

526
527
		if (!pfn_valid_within(low_pfn))
			continue;
528
		nr_scanned++;
529

530
531
532
533
534
535
		/*
		 * Get the page and ensure the page is within the same zone.
		 * See the comment in isolate_freepages about overlapping
		 * nodes. It is deliberate that the new zone lock is not taken
		 * as memory compaction should not move pages between nodes.
		 */
536
		page = pfn_to_page(low_pfn);
537
538
539
		if (page_zone(page) != zone)
			continue;

540
541
542
543
544
		if (!valid_page)
			valid_page = page;

		/* If isolation recently failed, do not retry */
		pageblock_nr = low_pfn >> pageblock_order;
545
546
547
548
549
550
551
552
553
554
555
556
557
		if (last_pageblock_nr != pageblock_nr) {
			int mt;

			last_pageblock_nr = pageblock_nr;
			if (!isolation_suitable(cc, page))
				goto next_pageblock;

			/*
			 * For async migration, also only scan in MOVABLE
			 * blocks. Async migration is optimistic to see if
			 * the minimum amount of work satisfies the allocation
			 */
			mt = get_pageblock_migratetype(page);
558
559
			if (cc->mode == MIGRATE_ASYNC &&
			    !migrate_async_suitable(mt)) {
560
				set_unsuitable = false;
561
562
563
				goto next_pageblock;
			}
		}
564

565
566
567
568
		/*
		 * Skip if free. page_order cannot be used without zone->lock
		 * as nothing prevents parallel allocations or buddy merging.
		 */
569
570
571
		if (PageBuddy(page))
			continue;

572
573
574
575
576
577
578
579
580
		/*
		 * Check may be lockless but that's ok as we recheck later.
		 * It's possible to migrate LRU pages and balloon pages
		 * Skip any other type of page
		 */
		if (!PageLRU(page)) {
			if (unlikely(balloon_page_movable(page))) {
				if (locked && balloon_page_isolate(page)) {
					/* Successfully isolated */
581
					goto isolate_success;
582
583
				}
			}
584
			continue;
585
		}
586
587

		/*
588
589
590
591
592
593
594
595
		 * PageLRU is set. lru_lock normally excludes isolation
		 * splitting and collapsing (collapsing has already happened
		 * if PageLRU is set) but the lock is not necessarily taken
		 * here and it is wasteful to take it just to check transhuge.
		 * Check TransHuge without lock and skip the whole pageblock if
		 * it's either a transhuge or hugetlbfs page, as calling
		 * compound_order() without preventing THP from splitting the
		 * page underneath us may return surprising results.
596
		 */
597
598
599
600
601
602
603
		if (PageTransHuge(page)) {
			if (!locked)
				goto next_pageblock;
			low_pfn += (1 << compound_order(page)) - 1;
			continue;
		}

604
605
606
607
608
609
610
611
612
		/*
		 * Migration will fail if an anonymous page is pinned in memory,
		 * so avoid taking lru_lock and isolating it unnecessarily in an
		 * admittedly racy check.
		 */
		if (!page_mapping(page) &&
		    page_count(page) > page_mapcount(page))
			continue;

613
614
615
616
617
618
619
620
621
		/* Check if it is ok to still hold the lock */
		locked = compact_checklock_irqsave(&zone->lru_lock, &flags,
								locked, cc);
		if (!locked || fatal_signal_pending(current))
			break;

		/* Recheck PageLRU and PageTransHuge under lock */
		if (!PageLRU(page))
			continue;
622
623
624
625
626
		if (PageTransHuge(page)) {
			low_pfn += (1 << compound_order(page)) - 1;
			continue;
		}

627
628
		lruvec = mem_cgroup_page_lruvec(page, zone);

629
		/* Try isolate the page */
630
		if (__isolate_lru_page(page, mode) != 0)
631
632
			continue;

633
		VM_BUG_ON_PAGE(PageTransCompound(page), page);
634

635
		/* Successfully isolated */
636
		del_page_from_lru_list(page, lruvec, page_lru(page));
637
638
639

isolate_success:
		cc->finished_update_migrate = true;
640
641
		list_add(&page->lru, migratelist);
		cc->nr_migratepages++;
642
		nr_isolated++;
643
644

		/* Avoid isolating too much */
645
646
		if (cc->nr_migratepages == COMPACT_CLUSTER_MAX) {
			++low_pfn;
647
			break;
648
		}
649
650
651
652

		continue;

next_pageblock:
653
		low_pfn = ALIGN(low_pfn + 1, pageblock_nr_pages) - 1;
654
655
	}

656
	acct_isolated(zone, locked, cc);
657

658
659
	if (locked)
		spin_unlock_irqrestore(&zone->lru_lock, flags);
660

661
662
663
664
	/*
	 * Update the pageblock-skip information and cached scanner pfn,
	 * if the whole pageblock was scanned without isolating any page.
	 */
665
666
667
	if (low_pfn == end_pfn)
		update_pageblock_skip(cc, valid_page, nr_isolated,
				      set_unsuitable, true);
668

669
670
	trace_mm_compaction_isolate_migratepages(nr_scanned, nr_isolated);

671
	count_compact_events(COMPACTMIGRATE_SCANNED, nr_scanned);
672
	if (nr_isolated)
673
		count_compact_events(COMPACTISOLATED, nr_isolated);
674

675
676
677
	return low_pfn;
}

678
679
#endif /* CONFIG_COMPACTION || CONFIG_CMA */
#ifdef CONFIG_COMPACTION
680
/*
681
682
 * Based on information in the current compact_control, find blocks
 * suitable for isolating free pages from and then isolate them.
683
 */
684
685
static void isolate_freepages(struct zone *zone,
				struct compact_control *cc)
686
{
687
	struct page *page;
688
689
690
691
	unsigned long block_start_pfn;	/* start of current pageblock */
	unsigned long block_end_pfn;	/* end of current pageblock */
	unsigned long low_pfn;	     /* lowest pfn scanner is able to scan */
	unsigned long next_free_pfn; /* start pfn for scaning at next round */
692
693
	int nr_freepages = cc->nr_freepages;
	struct list_head *freelist = &cc->freepages;
694

695
696
	/*
	 * Initialise the free scanner. The starting point is where we last
697
698
	 * successfully isolated from, zone-cached value, or the end of the
	 * zone when isolating for the first time. We need this aligned to
699
700
701
702
	 * the pageblock boundary, because we do
	 * block_start_pfn -= pageblock_nr_pages in the for loop.
	 * For ending point, take care when isolating in last pageblock of a
	 * a zone which ends in the middle of a pageblock.
703
704
	 * The low boundary is the end of the pageblock the migration scanner
	 * is using.
705
	 */
706
707
708
	block_start_pfn = cc->free_pfn & ~(pageblock_nr_pages-1);
	block_end_pfn = min(block_start_pfn + pageblock_nr_pages,
						zone_end_pfn(zone));
709
	low_pfn = ALIGN(cc->migrate_pfn + 1, pageblock_nr_pages);
710

711
	/*
712
713
	 * If no pages are isolated, the block_start_pfn < low_pfn check
	 * will kick in.
714
	 */
715
	next_free_pfn = 0;
716

717
718
719
720
721
	/*
	 * Isolate free pages until enough are available to migrate the
	 * pages on cc->migratepages. We stop searching if the migrate
	 * and free page scanners meet or enough free pages are isolated.
	 */
722
723
724
	for (; block_start_pfn >= low_pfn && cc->nr_migratepages > nr_freepages;
				block_end_pfn = block_start_pfn,
				block_start_pfn -= pageblock_nr_pages) {
725
		unsigned long isolated;
726

727
728
729
730
731
732
733
		/*
		 * This can iterate a massively long zone without finding any
		 * suitable migration targets, so periodically check if we need
		 * to schedule.
		 */
		cond_resched();

734
		if (!pfn_valid(block_start_pfn))
735
			continue;
736

737
738
739
740
741
742
743
		/*
		 * Check for overlapping nodes/zones. It's possible on some
		 * configurations to have a setup like
		 * node0 node1 node0
		 * i.e. it's possible that all pages within a zones range of
		 * pages do not belong to a single zone.
		 */
744
		page = pfn_to_page(block_start_pfn);
745
746
747
748
		if (page_zone(page) != zone)
			continue;

		/* Check the block is suitable for migration */
749
		if (!suitable_migration_target(page))
750
			continue;
751

752
753
754
755
		/* If isolation recently failed, do not retry */
		if (!isolation_suitable(cc, page))
			continue;

756
		/* Found a block suitable for isolating free pages from */
757
758
		isolated = isolate_freepages_block(cc, block_start_pfn,
					block_end_pfn, freelist, false);
759
		nr_freepages += isolated;
760
761
762
763
764
765

		/*
		 * Record the highest PFN we isolated pages from. When next
		 * looking for free pages, the search will restart here as
		 * page migration may have returned some pages to the allocator
		 */
766
		if (isolated && next_free_pfn == 0) {
767
			cc->finished_update_free = true;
768
			next_free_pfn = block_start_pfn;
769
		}
770
771
772
773
774
	}

	/* split_free_page does not map the pages */
	map_pages(freelist);

775
776
777
778
	/*
	 * If we crossed the migrate scanner, we want to keep it that way
	 * so that compact_finished() may detect this
	 */
779
780
781
782
	if (block_start_pfn < low_pfn)
		next_free_pfn = cc->migrate_pfn;

	cc->free_pfn = next_free_pfn;
783
	cc->nr_freepages = nr_freepages;
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
}

/*
 * This is a migrate-callback that "allocates" freepages by taking pages
 * from the isolated freelists in the block we are migrating to.
 */
static struct page *compaction_alloc(struct page *migratepage,
					unsigned long data,
					int **result)
{
	struct compact_control *cc = (struct compact_control *)data;
	struct page *freepage;

	/* Isolate free pages if necessary */
	if (list_empty(&cc->freepages)) {
		isolate_freepages(cc->zone, cc);

		if (list_empty(&cc->freepages))
			return NULL;
	}

	freepage = list_entry(cc->freepages.next, struct page, lru);
	list_del(&freepage->lru);
	cc->nr_freepages--;

	return freepage;
}

/*
813
814
815
816
817
818
819
820
821
822
823
824
 * This is a migrate-callback that "frees" freepages back to the isolated
 * freelist.  All pages on the freelist are from the same zone, so there is no
 * special handling needed for NUMA.
 */
static void compaction_free(struct page *page, unsigned long data)
{
	struct compact_control *cc = (struct compact_control *)data;

	list_add(&page->lru, &cc->freepages);
	cc->nr_freepages++;
}

825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
/* possible outcome of isolate_migratepages */
typedef enum {
	ISOLATE_ABORT,		/* Abort compaction now */
	ISOLATE_NONE,		/* No pages isolated, continue scanning */
	ISOLATE_SUCCESS,	/* Pages isolated, migrate */
} isolate_migrate_t;

/*
 * Isolate all pages that can be migrated from the block pointed to by
 * the migrate scanner within compact_control.
 */
static isolate_migrate_t isolate_migratepages(struct zone *zone,
					struct compact_control *cc)
{
	unsigned long low_pfn, end_pfn;

	/* Do not scan outside zone boundaries */
	low_pfn = max(cc->migrate_pfn, zone->zone_start_pfn);

	/* Only scan within a pageblock boundary */
845
	end_pfn = ALIGN(low_pfn + 1, pageblock_nr_pages);
846
847
848
849
850
851
852
853

	/* Do not cross the free scanner or scan within a memory hole */
	if (end_pfn > cc->free_pfn || !pfn_valid(low_pfn)) {
		cc->migrate_pfn = end_pfn;
		return ISOLATE_NONE;
	}

	/* Perform the isolation */
Minchan Kim's avatar
Minchan Kim committed
854
	low_pfn = isolate_migratepages_range(zone, cc, low_pfn, end_pfn, false);
855
	if (!low_pfn || cc->contended)
856
857
858
859
860
861
862
		return ISOLATE_ABORT;

	cc->migrate_pfn = low_pfn;

	return ISOLATE_SUCCESS;
}

863
static int compact_finished(struct zone *zone,
864
			    struct compact_control *cc)
865
{
866
	unsigned int order;
867
	unsigned long watermark;
868

869
870
871
	if (fatal_signal_pending(current))
		return COMPACT_PARTIAL;

872
	/* Compaction run completes if the migrate and free scanner meet */
873
	if (cc->free_pfn <= cc->migrate_pfn) {
874
		/* Let the next compaction start anew. */
875
876
		zone->compact_cached_migrate_pfn[0] = zone->zone_start_pfn;
		zone->compact_cached_migrate_pfn[1] = zone->zone_start_pfn;
877
878
		zone->compact_cached_free_pfn = zone_end_pfn(zone);

879
880
881
882
883
884
885
886
887
		/*
		 * Mark that the PG_migrate_skip information should be cleared
		 * by kswapd when it goes to sleep. kswapd does not set the
		 * flag itself as the decision to be clear should be directly
		 * based on an allocation request.
		 */
		if (!current_is_kswapd())
			zone->compact_blockskip_flush = true;

888
		return COMPACT_COMPLETE;
889
	}
890

891
892
893
894
	/*
	 * order == -1 is expected when compacting via
	 * /proc/sys/vm/compact_memory
	 */
895
896
897
	if (cc->order == -1)
		return COMPACT_CONTINUE;

898
899
900
901
902
903
904
	/* Compaction run is not finished if the watermark is not met */
	watermark = low_wmark_pages(zone);
	watermark += (1 << cc->order);

	if (!zone_watermark_ok(zone, cc->order, watermark, 0, 0))
		return COMPACT_CONTINUE;

905
	/* Direct compactor: Is a suitable page free? */
906
907
908
909
910
911
912
913
914
	for (order = cc->order; order < MAX_ORDER; order++) {
		struct free_area *area = &zone->free_area[order];

		/* Job done if page is free of the right migratetype */
		if (!list_empty(&area->free_list[cc->migratetype]))
			return COMPACT_PARTIAL;

		/* Job done if allocation would set block type */
		if (cc->order >= pageblock_order && area->nr_free)
915
916
917
			return COMPACT_PARTIAL;
	}

918
919
920
	return COMPACT_CONTINUE;
}

921
922
923
924
925
926
927
928
929
930
931
932
/*
 * compaction_suitable: Is this suitable to run compaction on this zone now?
 * Returns
 *   COMPACT_SKIPPED  - If there are too few free pages for compaction
 *   COMPACT_PARTIAL  - If the allocation would succeed without compaction
 *   COMPACT_CONTINUE - If compaction should run now
 */
unsigned long compaction_suitable(struct zone *zone, int order)
{
	int fragindex;
	unsigned long watermark;

933
934
935
936
937
938
939
	/*
	 * order == -1 is expected when compacting via
	 * /proc/sys/vm/compact_memory
	 */
	if (order == -1)
		return COMPACT_CONTINUE;

940
941
942
943
944
945
946
947
948
949
950
951
952
	/*
	 * Watermarks for order-0 must be met for compaction. Note the 2UL.
	 * This is because during migration, copies of pages need to be
	 * allocated and for a short time, the footprint is higher
	 */
	watermark = low_wmark_pages(zone) + (2UL << order);
	if (!zone_watermark_ok(zone, 0, watermark, 0, 0))
		return COMPACT_SKIPPED;

	/*
	 * fragmentation index determines if allocation failures are due to
	 * low memory or external fragmentation
	 *
953
954
	 * index of -1000 implies allocations might succeed depending on
	 * watermarks
955
956
957
958
959
960
961
962
963
	 * index towards 0 implies failure is due to lack of memory
	 * index towards 1000 implies failure is due to fragmentation
	 *
	 * Only compact if a failure would be due to fragmentation.
	 */
	fragindex = fragmentation_index(zone, order);
	if (fragindex >= 0 && fragindex <= sysctl_extfrag_threshold)
		return COMPACT_SKIPPED;

964
965
	if (fragindex == -1000 && zone_watermark_ok(zone, order, watermark,
	    0, 0))
966
967
968
969
970
		return COMPACT_PARTIAL;

	return COMPACT_CONTINUE;
}

971
972
973
static int compact_zone(struct zone *zone, struct compact_control *cc)
{
	int ret;
974
	unsigned long start_pfn = zone->zone_start_pfn;
975
	unsigned long end_pfn = zone_end_pfn(zone);
976
	const bool sync = cc->mode != MIGRATE_ASYNC;
977

978
979
980
981
982
983
984
985
986
987
988
	ret = compaction_suitable(zone, cc->order);
	switch (ret) {
	case COMPACT_PARTIAL:
	case COMPACT_SKIPPED:
		/* Compaction is likely to fail */
		return ret;
	case COMPACT_CONTINUE:
		/* Fall through to compaction */
		;
	}

989
990
991
992
993
994
995
996
	/*
	 * Clear pageblock skip if there were failures recently and compaction
	 * is about to be retried after being deferred. kswapd does not do
	 * this reset as it'll reset the cached information when going to sleep.
	 */
	if (compaction_restarting(zone, cc->order) && !current_is_kswapd())
		__reset_isolation_suitable(zone);

997
998
999
1000
1001
	/*
	 * Setup to move all movable pages to the end of the zone. Used cached
	 * information on where the scanners should start but check that it
	 * is initialised by ensuring the values are within zone boundaries.
	 */
1002
	cc->migrate_pfn = zone->compact_cached_migrate_pfn[sync];
1003
1004
1005
1006
1007
1008
1009
	cc->free_pfn = zone->compact_cached_free_pfn;
	if (cc->free_pfn < start_pfn || cc->free_pfn > end_pfn) {
		cc->free_pfn = end_pfn & ~(pageblock_nr_pages-1);
		zone->compact_cached_free_pfn = cc->free_pfn;
	}
	if (cc->migrate_pfn < start_pfn || cc->migrate_pfn > end_pfn) {
		cc->migrate_pfn = start_pfn;
1010
1011
		zone->compact_cached_migrate_pfn[0] = cc->migrate_pfn;
		zone->compact_cached_migrate_pfn[1] = cc->migrate_pfn;
1012
	}
1013

1014
1015
	trace_mm_compaction_begin(start_pfn, cc->migrate_pfn, cc->free_pfn, end_pfn);

1016
1017
1018
	migrate_prep_local();

	while ((ret = compact_finished(zone, cc)) == COMPACT_CONTINUE) {
1019
		int err;
1020

1021
1022
1023
		switch (isolate_migratepages(zone, cc)) {
		case ISOLATE_ABORT:
			ret = COMPACT_PARTIAL;
1024
			putback_movable_pages(&cc->migratepages);
1025
			cc->nr_migratepages = 0;
1026
1027
			goto out;
		case ISOLATE_NONE:
1028
			continue;
1029
1030
1031
		case ISOLATE_SUCCESS:
			;
		}
1032

1033
1034
1035
		if (!cc->nr_migratepages)
			continue;

1036
		err = migrate_pages(&cc->migratepages, compaction_alloc,
1037
				compaction_free, (unsigned long)cc, cc->mode,
1038
				MR_COMPACTION);
1039

1040
1041
		trace_mm_compaction_migratepages(cc->nr_migratepages, err,
							&cc->migratepages);
1042

1043
1044
		/* All pages were either migrated or will be released */
		cc->nr_migratepages = 0;
1045
		if (err) {
1046
			putback_movable_pages(&cc->migratepages);
1047
1048
1049
1050
1051
			/*
			 * migrate_pages() may return -ENOMEM when scanners meet
			 * and we want compact_finished() to detect it
			 */
			if (err == -ENOMEM && cc->free_pfn > cc->migrate_pfn) {
1052
1053
1054
				ret = COMPACT_PARTIAL;
				goto out;
			}
1055
1056
1057
		}
	}

1058
out:
1059
1060
1061
1062
	/* Release free pages and check accounting */
	cc->nr_freepages -= release_freepages(&cc->freepages);
	VM_BUG_ON(cc->nr_freepages != 0);

1063
1064
	trace_mm_compaction_end(ret);

1065
1066
	return ret;
}
1067

1068
1069
static unsigned long compact_zone_order(struct zone *zone, int order,
		gfp_t gfp_mask, enum migrate_mode mode, bool *contended)
1070
{
1071
	unsigned long ret;
1072
1073
1074
1075
1076
1077
	struct compact_control cc = {
		.nr_freepages = 0,
		.nr_migratepages = 0,
		.order = order,
		.migratetype = allocflags_to_migratetype(gfp_mask),
		.zone = zone,
1078
		.mode = mode,
1079
1080
1081
1082
	};
	INIT_LIST_HEAD(&cc.freepages);
	INIT_LIST_HEAD(&cc.migratepages);

1083
1084
1085
1086
1087
1088
1089
	ret = compact_zone(zone, &cc);

	VM_BUG_ON(!list_empty(&cc.freepages));
	VM_BUG_ON(!list_empty(&cc.migratepages));

	*contended = cc.contended;
	return ret;
1090
1091
}

1092
1093
int sysctl_extfrag_threshold = 500;

1094
1095
1096
1097
1098
1099
/**
 * try_to_compact_pages - Direct compact to satisfy a high-order allocation
 * @zonelist: The zonelist used for the current allocation
 * @order: The order of the current allocation
 * @gfp_mask: The GFP mask of the current allocation
 * @nodemask: The allowed nodes to allocate from
1100
 * @mode: The migration mode for async, sync light, or sync migration
1101
1102
 * @contended: Return value that is true if compaction was aborted due to lock contention
 * @page: Optionally capture a free page of the requested order during compaction
1103
1104
1105
1106
 *
 * This is the main entry point for direct page compaction.
 */
unsigned long try_to_compact_pages(struct zonelist *zonelist,
1107
			int order, gfp_t gfp_mask, nodemask_t *nodemask,
1108
			enum migrate_mode mode, bool *contended)
1109
1110
1111
1112
1113
1114
1115
{
	enum zone_type high_zoneidx = gfp_zone(gfp_mask);
	int may_enter_fs = gfp_mask & __GFP_FS;
	int may_perform_io = gfp_mask & __GFP_IO;
	struct zoneref *z;
	struct zone *zone;
	int rc = COMPACT_SKIPPED;
1116
	int alloc_flags = 0;
1117

1118
	/* Check if the GFP flags allow compaction */
1119
	if (!order || !may_enter_fs || !may_perform_io)
1120
1121
		return rc;

1122
	count_compact_event(COMPACTSTALL);
1123

1124
1125
1126
1127
#ifdef CONFIG_CMA
	if (allocflags_to_migratetype(gfp_mask) == MIGRATE_MOVABLE)
		alloc_flags |= ALLOC_CMA;
#endif
1128
1129
1130
1131
1132
	/* Compact each zone in the list */
	for_each_zone_zonelist_nodemask(zone, z, zonelist, high_zoneidx,
								nodemask) {
		int status;

1133
		status = compact_zone_order(zone, order, gfp_mask, mode,
1134
						contended);
1135
1136
		rc = max(status, rc);

1137
		/* If a normal allocation would succeed, stop compacting */
1138
1139
		if (zone_watermark_ok(zone, order, low_wmark_pages(zone), 0,
				      alloc_flags))
1140
1141
1142
1143
1144
1145
1146
			break;
	}

	return rc;
}


1147
/* Compact all zones within a node */
1148
static void __compact_pgdat(pg_data_t *pgdat, struct compact_control *cc)
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
{
	int zoneid;
	struct zone *zone;

	for (zoneid = 0; zoneid < MAX_NR_ZONES; zoneid++) {

		zone = &pgdat->node_zones[zoneid];
		if (!populated_zone(zone))
			continue;

1159
1160
1161
1162
1163
		cc->nr_freepages = 0;
		cc->nr_migratepages = 0;
		cc->zone = zone;
		INIT_LIST_HEAD(&cc->freepages);
		INIT_LIST_HEAD(&cc->migratepages);
<