page_alloc.c 193 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
/*
 *  linux/mm/page_alloc.c
 *
 *  Manages the free list, the system allocates free pages here.
 *  Note that kmalloc() lives in slab.c
 *
 *  Copyright (C) 1991, 1992, 1993, 1994  Linus Torvalds
 *  Swap reorganised 29.12.95, Stephen Tweedie
 *  Support of BIGMEM added by Gerhard Wichert, Siemens AG, July 1999
 *  Reshaped it to be a zoned allocator, Ingo Molnar, Red Hat, 1999
 *  Discontiguous memory support, Kanoj Sarcar, SGI, Nov 1999
 *  Zone balancing, Kanoj Sarcar, SGI, Jan 2000
 *  Per cpu hot/cold page lists, bulk allocation, Martin J. Bligh, Sept 2002
 *          (lots of bits borrowed from Ingo Molnar & Andrew Morton)
 */

#include <linux/stddef.h>
#include <linux/mm.h>
#include <linux/swap.h>
#include <linux/interrupt.h>
#include <linux/pagemap.h>
22
#include <linux/jiffies.h>
Linus Torvalds's avatar
Linus Torvalds committed
23
#include <linux/bootmem.h>
24
#include <linux/memblock.h>
Linus Torvalds's avatar
Linus Torvalds committed
25
#include <linux/compiler.h>
26
#include <linux/kernel.h>
27
#include <linux/kmemcheck.h>
28
#include <linux/kasan.h>
Linus Torvalds's avatar
Linus Torvalds committed
29
30
31
32
33
#include <linux/module.h>
#include <linux/suspend.h>
#include <linux/pagevec.h>
#include <linux/blkdev.h>
#include <linux/slab.h>
34
#include <linux/ratelimit.h>
35
#include <linux/oom.h>
Linus Torvalds's avatar
Linus Torvalds committed
36
37
38
39
40
#include <linux/notifier.h>
#include <linux/topology.h>
#include <linux/sysctl.h>
#include <linux/cpu.h>
#include <linux/cpuset.h>
41
#include <linux/memory_hotplug.h>
Linus Torvalds's avatar
Linus Torvalds committed
42
43
#include <linux/nodemask.h>
#include <linux/vmalloc.h>
44
#include <linux/vmstat.h>
45
#include <linux/mempolicy.h>
46
#include <linux/stop_machine.h>
47
48
#include <linux/sort.h>
#include <linux/pfn.h>
49
#include <linux/backing-dev.h>
50
#include <linux/fault-inject.h>
51
#include <linux/page-isolation.h>
52
#include <linux/page_ext.h>
53
#include <linux/debugobjects.h>
54
#include <linux/kmemleak.h>
55
#include <linux/compaction.h>
56
#include <trace/events/kmem.h>
57
#include <linux/prefetch.h>
58
#include <linux/mm_inline.h>
59
#include <linux/migrate.h>
60
#include <linux/page_ext.h>
61
#include <linux/hugetlb.h>
62
#include <linux/sched/rt.h>
63
#include <linux/page_owner.h>
64
#include <linux/kthread.h>
Linus Torvalds's avatar
Linus Torvalds committed
65

66
#include <asm/sections.h>
Linus Torvalds's avatar
Linus Torvalds committed
67
#include <asm/tlbflush.h>
68
#include <asm/div64.h>
Linus Torvalds's avatar
Linus Torvalds committed
69
70
#include "internal.h"

71
72
/* prevent >1 _updater_ of zone percpu pageset ->high and ->batch fields */
static DEFINE_MUTEX(pcp_batch_high_lock);
73
#define MIN_PERCPU_PAGELIST_FRACTION	(8)
74

75
76
77
78
79
#ifdef CONFIG_USE_PERCPU_NUMA_NODE_ID
DEFINE_PER_CPU(int, numa_node);
EXPORT_PER_CPU_SYMBOL(numa_node);
#endif

80
81
82
83
84
85
86
87
88
#ifdef CONFIG_HAVE_MEMORYLESS_NODES
/*
 * N.B., Do NOT reference the '_numa_mem_' per cpu variable directly.
 * It will not be defined when CONFIG_HAVE_MEMORYLESS_NODES is not defined.
 * Use the accessor functions set_numa_mem(), numa_mem_id() and cpu_to_mem()
 * defined in <linux/topology.h>.
 */
DEFINE_PER_CPU(int, _numa_mem_);		/* Kernel "local memory" node */
EXPORT_PER_CPU_SYMBOL(_numa_mem_);
89
int _node_numa_mem_[MAX_NUMNODES];
90
91
#endif

Linus Torvalds's avatar
Linus Torvalds committed
92
/*
93
 * Array of node states.
Linus Torvalds's avatar
Linus Torvalds committed
94
 */
95
96
97
98
99
100
101
nodemask_t node_states[NR_NODE_STATES] __read_mostly = {
	[N_POSSIBLE] = NODE_MASK_ALL,
	[N_ONLINE] = { { [0] = 1UL } },
#ifndef CONFIG_NUMA
	[N_NORMAL_MEMORY] = { { [0] = 1UL } },
#ifdef CONFIG_HIGHMEM
	[N_HIGH_MEMORY] = { { [0] = 1UL } },
102
103
104
#endif
#ifdef CONFIG_MOVABLE_NODE
	[N_MEMORY] = { { [0] = 1UL } },
105
106
107
108
109
110
#endif
	[N_CPU] = { { [0] = 1UL } },
#endif	/* NUMA */
};
EXPORT_SYMBOL(node_states);

111
112
113
/* Protect totalram_pages and zone->managed_pages */
static DEFINE_SPINLOCK(managed_page_count_lock);

114
unsigned long totalram_pages __read_mostly;
115
unsigned long totalreserve_pages __read_mostly;
116
unsigned long totalcma_pages __read_mostly;
117
118
119
120
121
122
123
124
/*
 * When calculating the number of globally allowed dirty pages, there
 * is a certain number of per-zone reserves that should not be
 * considered dirtyable memory.  This is the sum of those reserves
 * over all existing zones that contribute dirtyable memory.
 */
unsigned long dirty_balance_reserve __read_mostly;

125
int percpu_pagelist_fraction;
126
gfp_t gfp_allowed_mask __read_mostly = GFP_BOOT_MASK;
Linus Torvalds's avatar
Linus Torvalds committed
127

128
129
130
131
132
133
134
135
136
#ifdef CONFIG_PM_SLEEP
/*
 * The following functions are used by the suspend/hibernate code to temporarily
 * change gfp_allowed_mask in order to avoid using I/O during memory allocations
 * while devices are suspended.  To avoid races with the suspend/hibernate code,
 * they should always be called with pm_mutex held (gfp_allowed_mask also should
 * only be modified with pm_mutex held, unless the suspend/hibernate code is
 * guaranteed not to run in parallel with that modification).
 */
137
138
139
140

static gfp_t saved_gfp_mask;

void pm_restore_gfp_mask(void)
141
142
{
	WARN_ON(!mutex_is_locked(&pm_mutex));
143
144
145
146
	if (saved_gfp_mask) {
		gfp_allowed_mask = saved_gfp_mask;
		saved_gfp_mask = 0;
	}
147
148
}

149
void pm_restrict_gfp_mask(void)
150
151
{
	WARN_ON(!mutex_is_locked(&pm_mutex));
152
153
154
	WARN_ON(saved_gfp_mask);
	saved_gfp_mask = gfp_allowed_mask;
	gfp_allowed_mask &= ~GFP_IOFS;
155
}
156
157
158
159
160
161
162

bool pm_suspended_storage(void)
{
	if ((gfp_allowed_mask & GFP_IOFS) == GFP_IOFS)
		return false;
	return true;
}
163
164
#endif /* CONFIG_PM_SLEEP */

165
166
167
168
#ifdef CONFIG_HUGETLB_PAGE_SIZE_VARIABLE
int pageblock_order __read_mostly;
#endif

169
static void __free_pages_ok(struct page *page, unsigned int order);
170

Linus Torvalds's avatar
Linus Torvalds committed
171
172
173
174
175
176
/*
 * results with 256, 32 in the lowmem_reserve sysctl:
 *	1G machine -> (16M dma, 800M-16M normal, 1G-800M high)
 *	1G machine -> (16M dma, 784M normal, 224M high)
 *	NORMAL allocation will leave 784M/256 of ram reserved in the ZONE_DMA
 *	HIGHMEM allocation will leave 224M/32 of ram reserved in ZONE_NORMAL
Yaowei Bai's avatar
Yaowei Bai committed
177
 *	HIGHMEM allocation will leave (224M+784M)/256 of ram reserved in ZONE_DMA
178
179
180
 *
 * TBD: should special case ZONE_DMA32 machines here - in those we normally
 * don't need any ZONE_NORMAL reservation
Linus Torvalds's avatar
Linus Torvalds committed
181
 */
182
int sysctl_lowmem_reserve_ratio[MAX_NR_ZONES-1] = {
183
#ifdef CONFIG_ZONE_DMA
184
	 256,
185
#endif
186
#ifdef CONFIG_ZONE_DMA32
187
	 256,
188
#endif
189
#ifdef CONFIG_HIGHMEM
Mel Gorman's avatar
Mel Gorman committed
190
	 32,
191
#endif
Mel Gorman's avatar
Mel Gorman committed
192
	 32,
193
};
Linus Torvalds's avatar
Linus Torvalds committed
194
195
196

EXPORT_SYMBOL(totalram_pages);

197
static char * const zone_names[MAX_NR_ZONES] = {
198
#ifdef CONFIG_ZONE_DMA
199
	 "DMA",
200
#endif
201
#ifdef CONFIG_ZONE_DMA32
202
	 "DMA32",
203
#endif
204
	 "Normal",
205
#ifdef CONFIG_HIGHMEM
Mel Gorman's avatar
Mel Gorman committed
206
	 "HighMem",
207
#endif
Mel Gorman's avatar
Mel Gorman committed
208
	 "Movable",
209
210
};

Linus Torvalds's avatar
Linus Torvalds committed
211
int min_free_kbytes = 1024;
212
int user_min_free_kbytes = -1;
Linus Torvalds's avatar
Linus Torvalds committed
213

214
215
static unsigned long __meminitdata nr_kernel_pages;
static unsigned long __meminitdata nr_all_pages;
216
static unsigned long __meminitdata dma_reserve;
Linus Torvalds's avatar
Linus Torvalds committed
217

Tejun Heo's avatar
Tejun Heo committed
218
219
220
221
222
223
224
225
226
227
228
#ifdef CONFIG_HAVE_MEMBLOCK_NODE_MAP
static unsigned long __meminitdata arch_zone_lowest_possible_pfn[MAX_NR_ZONES];
static unsigned long __meminitdata arch_zone_highest_possible_pfn[MAX_NR_ZONES];
static unsigned long __initdata required_kernelcore;
static unsigned long __initdata required_movablecore;
static unsigned long __meminitdata zone_movable_pfn[MAX_NUMNODES];

/* movable_zone is the "real" zone pages in ZONE_MOVABLE are taken from */
int movable_zone;
EXPORT_SYMBOL(movable_zone);
#endif /* CONFIG_HAVE_MEMBLOCK_NODE_MAP */
229

Miklos Szeredi's avatar
Miklos Szeredi committed
230
231
#if MAX_NUMNODES > 1
int nr_node_ids __read_mostly = MAX_NUMNODES;
232
int nr_online_nodes __read_mostly = 1;
Miklos Szeredi's avatar
Miklos Szeredi committed
233
EXPORT_SYMBOL(nr_node_ids);
234
EXPORT_SYMBOL(nr_online_nodes);
Miklos Szeredi's avatar
Miklos Szeredi committed
235
236
#endif

237
238
int page_group_by_mobility_disabled __read_mostly;

239
240
241
242
243
244
245
#ifdef CONFIG_DEFERRED_STRUCT_PAGE_INIT
static inline void reset_deferred_meminit(pg_data_t *pgdat)
{
	pgdat->first_deferred_pfn = ULONG_MAX;
}

/* Returns true if the struct page for the pfn is uninitialised */
246
static inline bool __meminit early_page_uninitialised(unsigned long pfn)
247
{
248
	if (pfn >= NODE_DATA(early_pfn_to_nid(pfn))->first_deferred_pfn)
249
250
251
252
253
		return true;

	return false;
}

254
255
256
257
258
259
260
261
static inline bool early_page_nid_uninitialised(unsigned long pfn, int nid)
{
	if (pfn >= NODE_DATA(nid)->first_deferred_pfn)
		return true;

	return false;
}

262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
/*
 * Returns false when the remaining initialisation should be deferred until
 * later in the boot cycle when it can be parallelised.
 */
static inline bool update_defer_init(pg_data_t *pgdat,
				unsigned long pfn, unsigned long zone_end,
				unsigned long *nr_initialised)
{
	/* Always populate low zones for address-contrained allocations */
	if (zone_end < pgdat_end_pfn(pgdat))
		return true;

	/* Initialise at least 2G of the highest zone */
	(*nr_initialised)++;
	if (*nr_initialised > (2UL << (30 - PAGE_SHIFT)) &&
	    (pfn & (PAGES_PER_SECTION - 1)) == 0) {
		pgdat->first_deferred_pfn = pfn;
		return false;
	}

	return true;
}
#else
static inline void reset_deferred_meminit(pg_data_t *pgdat)
{
}

static inline bool early_page_uninitialised(unsigned long pfn)
{
	return false;
}

294
295
296
297
298
static inline bool early_page_nid_uninitialised(unsigned long pfn, int nid)
{
	return false;
}

299
300
301
302
303
304
305
306
307
static inline bool update_defer_init(pg_data_t *pgdat,
				unsigned long pfn, unsigned long zone_end,
				unsigned long *nr_initialised)
{
	return true;
}
#endif


308
void set_pageblock_migratetype(struct page *page, int migratetype)
309
{
310
311
	if (unlikely(page_group_by_mobility_disabled &&
		     migratetype < MIGRATE_PCPTYPES))
312
313
		migratetype = MIGRATE_UNMOVABLE;

314
315
316
317
	set_pageblock_flags_group(page, (unsigned long)migratetype,
					PB_migrate, PB_migrate_end);
}

Nick Piggin's avatar
Nick Piggin committed
318
#ifdef CONFIG_DEBUG_VM
319
static int page_outside_zone_boundaries(struct zone *zone, struct page *page)
Linus Torvalds's avatar
Linus Torvalds committed
320
{
321
322
323
	int ret = 0;
	unsigned seq;
	unsigned long pfn = page_to_pfn(page);
324
	unsigned long sp, start_pfn;
325

326
327
	do {
		seq = zone_span_seqbegin(zone);
328
329
		start_pfn = zone->zone_start_pfn;
		sp = zone->spanned_pages;
330
		if (!zone_spans_pfn(zone, pfn))
331
332
333
			ret = 1;
	} while (zone_span_seqretry(zone, seq));

334
	if (ret)
335
336
337
		pr_err("page 0x%lx outside node %d zone %s [ 0x%lx - 0x%lx ]\n",
			pfn, zone_to_nid(zone), zone->name,
			start_pfn, start_pfn + sp);
338

339
	return ret;
340
341
342
343
}

static int page_is_consistent(struct zone *zone, struct page *page)
{
344
	if (!pfn_valid_within(page_to_pfn(page)))
345
		return 0;
Linus Torvalds's avatar
Linus Torvalds committed
346
	if (zone != page_zone(page))
347
348
349
350
351
352
353
354
355
356
		return 0;

	return 1;
}
/*
 * Temporary debugging check for pages not lying within a given zone.
 */
static int bad_range(struct zone *zone, struct page *page)
{
	if (page_outside_zone_boundaries(zone, page))
Linus Torvalds's avatar
Linus Torvalds committed
357
		return 1;
358
359
360
	if (!page_is_consistent(zone, page))
		return 1;

Linus Torvalds's avatar
Linus Torvalds committed
361
362
	return 0;
}
Nick Piggin's avatar
Nick Piggin committed
363
364
365
366
367
368
369
#else
static inline int bad_range(struct zone *zone, struct page *page)
{
	return 0;
}
#endif

370
371
static void bad_page(struct page *page, const char *reason,
		unsigned long bad_flags)
Linus Torvalds's avatar
Linus Torvalds committed
372
{
373
374
375
376
	static unsigned long resume;
	static unsigned long nr_shown;
	static unsigned long nr_unshown;

377
378
	/* Don't complain about poisoned pages */
	if (PageHWPoison(page)) {
379
		page_mapcount_reset(page); /* remove PageBuddy */
380
381
382
		return;
	}

383
384
385
386
387
388
389
390
391
392
	/*
	 * Allow a burst of 60 reports, then keep quiet for that minute;
	 * or allow a steady drip of one report per second.
	 */
	if (nr_shown == 60) {
		if (time_before(jiffies, resume)) {
			nr_unshown++;
			goto out;
		}
		if (nr_unshown) {
393
394
			printk(KERN_ALERT
			      "BUG: Bad page state: %lu messages suppressed\n",
395
396
397
398
399
400
401
402
				nr_unshown);
			nr_unshown = 0;
		}
		nr_shown = 0;
	}
	if (nr_shown++ == 0)
		resume = jiffies + 60 * HZ;

403
	printk(KERN_ALERT "BUG: Bad page state in process %s  pfn:%05lx\n",
404
		current->comm, page_to_pfn(page));
405
	dump_page_badflags(page, reason, bad_flags);
406

407
	print_modules();
Linus Torvalds's avatar
Linus Torvalds committed
408
	dump_stack();
409
out:
410
	/* Leave bad fields for debug, except PageBuddy could make trouble */
411
	page_mapcount_reset(page); /* remove PageBuddy */
412
	add_taint(TAINT_BAD_PAGE, LOCKDEP_NOW_UNRELIABLE);
Linus Torvalds's avatar
Linus Torvalds committed
413
414
415
416
417
418
419
420
421
}

/*
 * Higher-order pages are called "compound pages".  They are structured thusly:
 *
 * The first PAGE_SIZE page is called the "head page".
 *
 * The remaining PAGE_SIZE pages are called "tail pages".
 *
422
423
 * All pages have PG_compound set.  All tail pages have their ->first_page
 * pointing at the head page.
Linus Torvalds's avatar
Linus Torvalds committed
424
 *
425
426
427
 * The first tail page's ->lru.next holds the address of the compound page's
 * put_page() function.  Its ->lru.prev holds the order of allocation.
 * This usage means that zero-order pages may not be compound.
Linus Torvalds's avatar
Linus Torvalds committed
428
 */
429
430
431

static void free_compound_page(struct page *page)
{
432
	__free_pages_ok(page, compound_order(page));
433
434
}

435
void prep_compound_page(struct page *page, unsigned long order)
436
437
438
439
440
441
442
443
444
{
	int i;
	int nr_pages = 1 << order;

	set_compound_page_dtor(page, free_compound_page);
	set_compound_order(page, order);
	__SetPageHead(page);
	for (i = 1; i < nr_pages; i++) {
		struct page *p = page + i;
445
		set_page_count(p, 0);
446
		p->first_page = page;
David Rientjes's avatar
David Rientjes committed
447
448
449
		/* Make sure p->first_page is always valid for PageTail() */
		smp_wmb();
		__SetPageTail(p);
450
451
452
	}
}

453
454
#ifdef CONFIG_DEBUG_PAGEALLOC
unsigned int _debug_guardpage_minorder;
455
bool _debug_pagealloc_enabled __read_mostly;
456
457
bool _debug_guardpage_enabled __read_mostly;

458
459
460
461
462
463
464
465
466
467
468
469
static int __init early_debug_pagealloc(char *buf)
{
	if (!buf)
		return -EINVAL;

	if (strcmp(buf, "on") == 0)
		_debug_pagealloc_enabled = true;

	return 0;
}
early_param("debug_pagealloc", early_debug_pagealloc);

470
471
static bool need_debug_guardpage(void)
{
472
473
474
475
	/* If we don't use debug_pagealloc, we don't need guard page */
	if (!debug_pagealloc_enabled())
		return false;

476
477
478
479
480
	return true;
}

static void init_debug_guardpage(void)
{
481
482
483
	if (!debug_pagealloc_enabled())
		return;

484
485
486
487
488
489
490
	_debug_guardpage_enabled = true;
}

struct page_ext_operations debug_guardpage_ops = {
	.need = need_debug_guardpage,
	.init = init_debug_guardpage,
};
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505

static int __init debug_guardpage_minorder_setup(char *buf)
{
	unsigned long res;

	if (kstrtoul(buf, 10, &res) < 0 ||  res > MAX_ORDER / 2) {
		printk(KERN_ERR "Bad debug_guardpage_minorder value\n");
		return 0;
	}
	_debug_guardpage_minorder = res;
	printk(KERN_INFO "Setting debug_guardpage_minorder to %lu\n", res);
	return 0;
}
__setup("debug_guardpage_minorder=", debug_guardpage_minorder_setup);

506
507
static inline void set_page_guard(struct zone *zone, struct page *page,
				unsigned int order, int migratetype)
508
{
509
510
511
512
513
514
515
516
	struct page_ext *page_ext;

	if (!debug_guardpage_enabled())
		return;

	page_ext = lookup_page_ext(page);
	__set_bit(PAGE_EXT_DEBUG_GUARD, &page_ext->flags);

517
518
519
520
	INIT_LIST_HEAD(&page->lru);
	set_page_private(page, order);
	/* Guard pages are not available for any usage */
	__mod_zone_freepage_state(zone, -(1 << order), migratetype);
521
522
}

523
524
static inline void clear_page_guard(struct zone *zone, struct page *page,
				unsigned int order, int migratetype)
525
{
526
527
528
529
530
531
532
533
	struct page_ext *page_ext;

	if (!debug_guardpage_enabled())
		return;

	page_ext = lookup_page_ext(page);
	__clear_bit(PAGE_EXT_DEBUG_GUARD, &page_ext->flags);

534
535
536
	set_page_private(page, 0);
	if (!is_migrate_isolate(migratetype))
		__mod_zone_freepage_state(zone, (1 << order), migratetype);
537
538
}
#else
539
struct page_ext_operations debug_guardpage_ops = { NULL, };
540
541
542
543
static inline void set_page_guard(struct zone *zone, struct page *page,
				unsigned int order, int migratetype) {}
static inline void clear_page_guard(struct zone *zone, struct page *page,
				unsigned int order, int migratetype) {}
544
545
#endif

546
static inline void set_page_order(struct page *page, unsigned int order)
547
{
548
	set_page_private(page, order);
549
	__SetPageBuddy(page);
Linus Torvalds's avatar
Linus Torvalds committed
550
551
552
553
}

static inline void rmv_page_order(struct page *page)
{
554
	__ClearPageBuddy(page);
555
	set_page_private(page, 0);
Linus Torvalds's avatar
Linus Torvalds committed
556
557
558
559
560
}

/*
 * This function checks whether a page is free && is the buddy
 * we can do coalesce a page and its buddy if
Nick Piggin's avatar
Nick Piggin committed
561
 * (a) the buddy is not in a hole &&
562
 * (b) the buddy is in the buddy system &&
563
564
 * (c) a page and its buddy have the same order &&
 * (d) a page and its buddy are in the same zone.
565
 *
566
567
568
569
 * For recording whether a page is in the buddy system, we set ->_mapcount
 * PAGE_BUDDY_MAPCOUNT_VALUE.
 * Setting, clearing, and testing _mapcount PAGE_BUDDY_MAPCOUNT_VALUE is
 * serialized by zone->lock.
Linus Torvalds's avatar
Linus Torvalds committed
570
 *
571
 * For recording page's order, we use page_private(page).
Linus Torvalds's avatar
Linus Torvalds committed
572
 */
573
static inline int page_is_buddy(struct page *page, struct page *buddy,
574
							unsigned int order)
Linus Torvalds's avatar
Linus Torvalds committed
575
{
576
	if (!pfn_valid_within(page_to_pfn(buddy)))
Nick Piggin's avatar
Nick Piggin committed
577
578
		return 0;

579
	if (page_is_guard(buddy) && page_order(buddy) == order) {
580
581
582
		if (page_zone_id(page) != page_zone_id(buddy))
			return 0;

583
584
		VM_BUG_ON_PAGE(page_count(buddy) != 0, buddy);

585
586
587
		return 1;
	}

588
	if (PageBuddy(buddy) && page_order(buddy) == order) {
589
590
591
592
593
594
595
596
		/*
		 * zone check is done late to avoid uselessly
		 * calculating zone/node ids for pages that could
		 * never merge.
		 */
		if (page_zone_id(page) != page_zone_id(buddy))
			return 0;

597
598
		VM_BUG_ON_PAGE(page_count(buddy) != 0, buddy);

599
		return 1;
600
	}
601
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
}

/*
 * Freeing function for a buddy system allocator.
 *
 * The concept of a buddy system is to maintain direct-mapped table
 * (containing bit values) for memory blocks of various "orders".
 * The bottom level table contains the map for the smallest allocatable
 * units of memory (here, pages), and each level above it describes
 * pairs of units from the levels below, hence, "buddies".
 * At a high level, all that happens here is marking the table entry
 * at the bottom level available, and propagating the changes upward
 * as necessary, plus some accounting needed to play nicely with other
 * parts of the VM system.
 * At each level, we keep a list of pages, which are heads of continuous
617
618
619
 * free pages of length of (1 << order) and marked with _mapcount
 * PAGE_BUDDY_MAPCOUNT_VALUE. Page's order is recorded in page_private(page)
 * field.
Linus Torvalds's avatar
Linus Torvalds committed
620
 * So when we are allocating or freeing one, we can derive the state of the
621
622
 * other.  That is, if we allocate a small block, and both were
 * free, the remainder of the region must be split into blocks.
Linus Torvalds's avatar
Linus Torvalds committed
623
 * If a block is freed, and its buddy is also free, then this
624
 * triggers coalescing into a block of larger size.
Linus Torvalds's avatar
Linus Torvalds committed
625
 *
626
 * -- nyc
Linus Torvalds's avatar
Linus Torvalds committed
627
628
 */

Nick Piggin's avatar
Nick Piggin committed
629
static inline void __free_one_page(struct page *page,
630
		unsigned long pfn,
631
632
		struct zone *zone, unsigned int order,
		int migratetype)
Linus Torvalds's avatar
Linus Torvalds committed
633
634
{
	unsigned long page_idx;
635
	unsigned long combined_idx;
636
	unsigned long uninitialized_var(buddy_idx);
637
	struct page *buddy;
638
	int max_order = MAX_ORDER;
Linus Torvalds's avatar
Linus Torvalds committed
639

640
	VM_BUG_ON(!zone_is_initialized(zone));
641
	VM_BUG_ON_PAGE(page->flags & PAGE_FLAGS_CHECK_AT_PREP, page);
Linus Torvalds's avatar
Linus Torvalds committed
642

643
	VM_BUG_ON(migratetype == -1);
644
645
646
647
648
649
650
651
652
	if (is_migrate_isolate(migratetype)) {
		/*
		 * We restrict max order of merging to prevent merge
		 * between freepages on isolate pageblock and normal
		 * pageblock. Without this, pageblock isolation
		 * could cause incorrect freepage accounting.
		 */
		max_order = min(MAX_ORDER, pageblock_order + 1);
	} else {
653
		__mod_zone_freepage_state(zone, 1 << order, migratetype);
654
	}
655

656
	page_idx = pfn & ((1 << max_order) - 1);
Linus Torvalds's avatar
Linus Torvalds committed
657

658
659
	VM_BUG_ON_PAGE(page_idx & ((1 << order) - 1), page);
	VM_BUG_ON_PAGE(bad_range(zone, page), page);
Linus Torvalds's avatar
Linus Torvalds committed
660

661
	while (order < max_order - 1) {
662
663
		buddy_idx = __find_buddy_index(page_idx, order);
		buddy = page + (buddy_idx - page_idx);
664
		if (!page_is_buddy(page, buddy, order))
665
			break;
666
667
668
669
670
		/*
		 * Our buddy is free or it is CONFIG_DEBUG_PAGEALLOC guard page,
		 * merge with it and move up one order.
		 */
		if (page_is_guard(buddy)) {
671
			clear_page_guard(zone, buddy, order, migratetype);
672
673
674
675
676
		} else {
			list_del(&buddy->lru);
			zone->free_area[order].nr_free--;
			rmv_page_order(buddy);
		}
677
		combined_idx = buddy_idx & page_idx;
Linus Torvalds's avatar
Linus Torvalds committed
678
679
680
681
682
		page = page + (combined_idx - page_idx);
		page_idx = combined_idx;
		order++;
	}
	set_page_order(page, order);
683
684
685
686
687
688
689
690
691

	/*
	 * If this is not the largest possible page, check if the buddy
	 * of the next-highest order is free. If it is, it's possible
	 * that pages are being freed that will coalesce soon. In case,
	 * that is happening, add the free page to the tail of the list
	 * so it's less likely to be used soon and more likely to be merged
	 * as a higher order page
	 */
692
	if ((order < MAX_ORDER-2) && pfn_valid_within(page_to_pfn(buddy))) {
693
		struct page *higher_page, *higher_buddy;
694
695
696
		combined_idx = buddy_idx & page_idx;
		higher_page = page + (combined_idx - page_idx);
		buddy_idx = __find_buddy_index(combined_idx, order + 1);
697
		higher_buddy = higher_page + (buddy_idx - combined_idx);
698
699
700
701
702
703
704
705
706
		if (page_is_buddy(higher_page, higher_buddy, order + 1)) {
			list_add_tail(&page->lru,
				&zone->free_area[order].free_list[migratetype]);
			goto out;
		}
	}

	list_add(&page->lru, &zone->free_area[order].free_list[migratetype]);
out:
Linus Torvalds's avatar
Linus Torvalds committed
707
708
709
	zone->free_area[order].nr_free++;
}

710
static inline int free_pages_check(struct page *page)
Linus Torvalds's avatar
Linus Torvalds committed
711
{
712
	const char *bad_reason = NULL;
713
714
715
716
717
718
719
720
721
722
723
724
	unsigned long bad_flags = 0;

	if (unlikely(page_mapcount(page)))
		bad_reason = "nonzero mapcount";
	if (unlikely(page->mapping != NULL))
		bad_reason = "non-NULL mapping";
	if (unlikely(atomic_read(&page->_count) != 0))
		bad_reason = "nonzero _count";
	if (unlikely(page->flags & PAGE_FLAGS_CHECK_AT_FREE)) {
		bad_reason = "PAGE_FLAGS_CHECK_AT_FREE flag(s) set";
		bad_flags = PAGE_FLAGS_CHECK_AT_FREE;
	}
725
726
727
728
#ifdef CONFIG_MEMCG
	if (unlikely(page->mem_cgroup))
		bad_reason = "page still charged to cgroup";
#endif
729
730
	if (unlikely(bad_reason)) {
		bad_page(page, bad_reason, bad_flags);
731
		return 1;
732
	}
733
	page_cpupid_reset_last(page);
734
735
736
	if (page->flags & PAGE_FLAGS_CHECK_AT_PREP)
		page->flags &= ~PAGE_FLAGS_CHECK_AT_PREP;
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
737
738
739
}

/*
740
 * Frees a number of pages from the PCP lists
Linus Torvalds's avatar
Linus Torvalds committed
741
 * Assumes all pages on list are in same zone, and of same order.
742
 * count is the number of pages to free.
Linus Torvalds's avatar
Linus Torvalds committed
743
744
745
746
747
748
749
 *
 * If the zone was previously in an "all pages pinned" state then look to
 * see if this freeing clears that state.
 *
 * And clear the zone's pages_scanned counter, to hold off the "all pages are
 * pinned" detection logic.
 */
750
751
static void free_pcppages_bulk(struct zone *zone, int count,
					struct per_cpu_pages *pcp)
Linus Torvalds's avatar
Linus Torvalds committed
752
{
753
	int migratetype = 0;
754
	int batch_free = 0;
755
	int to_free = count;
756
	unsigned long nr_scanned;
757

Nick Piggin's avatar
Nick Piggin committed
758
	spin_lock(&zone->lock);
759
760
761
	nr_scanned = zone_page_state(zone, NR_PAGES_SCANNED);
	if (nr_scanned)
		__mod_zone_page_state(zone, NR_PAGES_SCANNED, -nr_scanned);
762

763
	while (to_free) {
Nick Piggin's avatar
Nick Piggin committed
764
		struct page *page;
765
766
767
		struct list_head *list;

		/*
768
769
770
771
772
		 * Remove pages from lists in a round-robin fashion. A
		 * batch_free count is maintained that is incremented when an
		 * empty list is encountered.  This is so more pages are freed
		 * off fuller lists instead of spinning excessively around empty
		 * lists
773
774
		 */
		do {
775
			batch_free++;
776
777
778
779
			if (++migratetype == MIGRATE_PCPTYPES)
				migratetype = 0;
			list = &pcp->lists[migratetype];
		} while (list_empty(list));
Nick Piggin's avatar
Nick Piggin committed
780

781
782
783
784
		/* This is the only non-empty list. Free them all. */
		if (batch_free == MIGRATE_PCPTYPES)
			batch_free = to_free;

785
		do {
786
787
			int mt;	/* migratetype of the to-be-freed page */

788
789
790
			page = list_entry(list->prev, struct page, lru);
			/* must delete as __free_one_page list manipulates */
			list_del(&page->lru);
791
			mt = get_freepage_migratetype(page);
792
			if (unlikely(has_isolate_pageblock(zone)))
793
794
				mt = get_pageblock_migratetype(page);

795
			/* MIGRATE_MOVABLE list may include MIGRATE_RESERVEs */
796
			__free_one_page(page, page_to_pfn(page), zone, 0, mt);
797
			trace_mm_page_pcpu_drain(page, 0, mt);
798
		} while (--to_free && --batch_free && !list_empty(list));
Linus Torvalds's avatar
Linus Torvalds committed
799
	}
Nick Piggin's avatar
Nick Piggin committed
800
	spin_unlock(&zone->lock);
Linus Torvalds's avatar
Linus Torvalds committed
801
802
}

803
804
static void free_one_page(struct zone *zone,
				struct page *page, unsigned long pfn,
805
				unsigned int order,
806
				int migratetype)
Linus Torvalds's avatar
Linus Torvalds committed
807
{
808
	unsigned long nr_scanned;
809
	spin_lock(&zone->lock);
810
811
812
	nr_scanned = zone_page_state(zone, NR_PAGES_SCANNED);
	if (nr_scanned)
		__mod_zone_page_state(zone, NR_PAGES_SCANNED, -nr_scanned);
813

814
815
816
817
	if (unlikely(has_isolate_pageblock(zone) ||
		is_migrate_isolate(migratetype))) {
		migratetype = get_pfnblock_migratetype(page, pfn);
	}
818
	__free_one_page(page, pfn, zone, order, migratetype);
819
	spin_unlock(&zone->lock);
Nick Piggin's avatar
Nick Piggin committed
820
821
}

822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
static int free_tail_pages_check(struct page *head_page, struct page *page)
{
	if (!IS_ENABLED(CONFIG_DEBUG_VM))
		return 0;
	if (unlikely(!PageTail(page))) {
		bad_page(page, "PageTail not set", 0);
		return 1;
	}
	if (unlikely(page->first_page != head_page)) {
		bad_page(page, "first_page not consistent", 0);
		return 1;
	}
	return 0;
}

837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
static void __meminit __init_single_page(struct page *page, unsigned long pfn,
				unsigned long zone, int nid)
{
	set_page_links(page, zone, nid, pfn);
	init_page_count(page);
	page_mapcount_reset(page);
	page_cpupid_reset_last(page);

	INIT_LIST_HEAD(&page->lru);
#ifdef WANT_PAGE_VIRTUAL
	/* The shift won't overflow because ZONE_NORMAL is below 4G. */
	if (!is_highmem_idx(zone))
		set_page_address(page, __va(pfn << PAGE_SHIFT));
#endif
}

static void __meminit __init_single_pfn(unsigned long pfn, unsigned long zone,
					int nid)
{
	return __init_single_page(pfn_to_page(pfn), pfn, zone, nid);
}

859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
#ifdef CONFIG_DEFERRED_STRUCT_PAGE_INIT
static void init_reserved_page(unsigned long pfn)
{
	pg_data_t *pgdat;
	int nid, zid;

	if (!early_page_uninitialised(pfn))
		return;

	nid = early_pfn_to_nid(pfn);
	pgdat = NODE_DATA(nid);

	for (zid = 0; zid < MAX_NR_ZONES; zid++) {
		struct zone *zone = &pgdat->node_zones[zid];

		if (pfn >= zone->zone_start_pfn && pfn < zone_end_pfn(zone))
			break;
	}
	__init_single_pfn(pfn, zid, nid);
}
#else
static inline void init_reserved_page(unsigned long pfn)
{
}
#endif /* CONFIG_DEFERRED_STRUCT_PAGE_INIT */

885
886
887
888
889
890
/*
 * Initialised pages do not have PageReserved set. This function is
 * called for each range allocated by the bootmem allocator and
 * marks the pages PageReserved. The remaining valid pages are later
 * sent to the buddy page allocator.
 */
891
void __meminit reserve_bootmem_region(unsigned long start, unsigned long end)
892
893
894
895
{
	unsigned long start_pfn = PFN_DOWN(start);
	unsigned long end_pfn = PFN_UP(end);

896
897
898
899
900
901
902
903
	for (; start_pfn < end_pfn; start_pfn++) {
		if (pfn_valid(start_pfn)) {
			struct page *page = pfn_to_page(start_pfn);

			init_reserved_page(start_pfn);
			SetPageReserved(page);
		}
	}
904
905
}

906
static bool free_pages_prepare(struct page *page, unsigned int order)
Nick Piggin's avatar
Nick Piggin committed
907
{
908
909
	bool compound = PageCompound(page);
	int i, bad = 0;
Linus Torvalds's avatar
Linus Torvalds committed
910

911
	VM_BUG_ON_PAGE(PageTail(page), page);
912
	VM_BUG_ON_PAGE(compound && compound_order(page) != order, page);
913

914
	trace_mm_page_free(page, order);
915
	kmemcheck_free_shadow(page, order);
916
	kasan_free_pages(page, order);
917

Andrea Arcangeli's avatar
Andrea Arcangeli committed
918
919
	if (PageAnon(page))
		page->mapping = NULL;
920
921
922
923
	bad += free_pages_check(page);
	for (i = 1; i < (1 << order); i++) {
		if (compound)
			bad += free_tail_pages_check(page, page + i);
Andrea Arcangeli's avatar
Andrea Arcangeli committed
924
		bad += free_pages_check(page + i);
925
	}
926
	if (bad)
927
		return false;
928

929
930
	reset_page_owner(page, order);

931
	if (!PageHighMem(page)) {
932
933
		debug_check_no_locks_freed(page_address(page),
					   PAGE_SIZE << order);
934
935
936
		debug_check_no_obj_freed(page_address(page),
					   PAGE_SIZE << order);
	}
Nick Piggin's avatar
Nick Piggin committed
937
	arch_free_page(page, order);
Nick Piggin's avatar
Nick Piggin committed
938
	kernel_map_pages(page, 1 << order, 0);
Nick Piggin's avatar
Nick Piggin committed
939

940
941
942
943
944
945
	return true;
}

static void __free_pages_ok(struct page *page, unsigned int order)
{
	unsigned long flags;
946
	int migratetype;
947
	unsigned long pfn = page_to_pfn(page);
948
949
950
951

	if (!free_pages_prepare(page, order))
		return;

952
	migratetype = get_pfnblock_migratetype(page, pfn);
Nick Piggin's avatar
Nick Piggin committed
953
	local_irq_save(flags);
954
	__count_vm_events(PGFREE, 1 << order);
955
	set_freepage_migratetype(page, migratetype);
956
	free_one_page(page_zone(page), page, pfn, order, migratetype);
Nick Piggin's avatar
Nick Piggin committed
957
	local_irq_restore(flags);
Linus Torvalds's avatar
Linus Torvalds committed
958
959
}

960
static void __init __free_pages_boot_core(struct page *page,
961
					unsigned long pfn, unsigned int order)
962
{
963
	unsigned int nr_pages = 1 << order;
964
	struct page *p = page;
965
	unsigned int loop;
966

967
968
969
	prefetchw(p);
	for (loop = 0; loop < (nr_pages - 1); loop++, p++) {
		prefetchw(p + 1);
970
971
		__ClearPageReserved(p);
		set_page_count(p, 0);
972
	}
973
974
	__ClearPageReserved(p);
	set_page_count(p, 0);
975

976
	page_zone(page)->managed_pages += nr_pages;
977
978
	set_page_refcounted(page);
	__free_pages(page, order);
979
980
}

981
982
#if defined(CONFIG_HAVE_ARCH_EARLY_PFN_TO_NID) || \
	defined(CONFIG_HAVE_MEMBLOCK_NODE_MAP)
983

984
985
986
987
static struct mminit_pfnnid_cache early_pfnnid_cache __meminitdata;

int __meminit early_pfn_to_nid(unsigned long pfn)
{
988
	static DEFINE_SPINLOCK(early_pfn_lock);
989
990
	int nid;

991
	spin_lock(&early_pfn_lock);
992
	nid = __early_pfn_to_nid(pfn, &early_pfnnid_cache);
993
994
995
996
997
	if (nid < 0)
		nid = 0;
	spin_unlock(&early_pfn_lock);

	return nid;
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
}
#endif

#ifdef CONFIG_NODES_SPAN_OTHER_NODES
static inline bool __meminit meminit_pfn_in_nid(unsigned long pfn, int node,
					struct mminit_pfnnid_cache *state)
{
	int nid;

	nid = __early_pfn_to_nid(pfn, state);
	if (nid >= 0 && nid != node)
		return false;
	return true;
}

/* Only safe to use early in boot when initialisation is single-threaded */
static inline bool __meminit early_pfn_in_nid(unsigned long pfn, int node)
{
	return meminit_pfn_in_nid(pfn, node, &early_pfnnid_cache);
}

#else

static inline bool __meminit early_pfn_in_nid(unsigned long pfn, int node)
{
	return true;
}
static inline bool __meminit meminit_pfn_in_nid(unsigned long pfn, int node,
					struct mminit_pfnnid_cache *state)
{
	return true;
}
#endif


1033
void __init __free_pages_bootmem(struct page *page, unsigned long pfn,
1034
1035
1036
1037
1038
1039
1040
							unsigned int order)
{
	if (early_page_uninitialised(pfn))
		return;
	return __free_pages_boot_core(page, pfn, order);
}

1041
#ifdef CONFIG_DEFERRED_STRUCT_PAGE_INIT
1042
static void __init deferred_free_range(struct page *page,
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
					unsigned long pfn, int nr_pages)
{
	int i;

	if (!page)
		return;

	/* Free a large naturally-aligned chunk if possible */
	if (nr_pages == MAX_ORDER_NR_PAGES &&
	    (pfn & (MAX_ORDER_NR_PAGES-1)) == 0) {
1053
		set_pageblock_migratetype(page, MIGRATE_MOVABLE);
1054
1055
1056
1057
1058
1059
1060
1061
		__free_pages_boot_core(page, pfn, MAX_ORDER-1);
		return;
	}

	for (i = 0; i < nr_pages; i++, page++, pfn++)
		__free_pages_boot_core(page, pfn, 0);
}

1062
1063
1064
1065
1066
1067
1068
1069
1070
/* Completion tracking for deferred_init_memmap() threads */
static atomic_t pgdat_init_n_undone __initdata;
static __initdata DECLARE_COMPLETION(pgdat_init_all_done_comp);

static inline void __init pgdat_init_report_one_done(void)
{
	if (atomic_dec_and_test(&pgdat_init_n_undone))
		complete(&pgdat_init_all_done_comp);
}
1071

1072
/* Initialise remaining memory on a node */
1073
static int __init deferred_init_memmap(void *data)
1074
{
1075
1076
	pg_data_t *pgdat = data;
	int nid = pgdat->node_id;
1077
1078
1079
1080
1081
1082
1083
	struct mminit_pfnnid_cache nid_init_state = { };
	unsigned long start = jiffies;
	unsigned long nr_pages = 0;
	unsigned long walk_start, walk_end;
	int i, zid;
	struct zone *zone;
	unsigned long first_init_pfn = pgdat->first_deferred_pfn;
1084
	const struct cpumask *cpumask = cpumask_of_node(pgdat->node_id);
1085

1086
	if (first_init_pfn == ULONG_MAX) {
1087
		pgdat_init_report_one_done();
1088
1089
1090
1091
1092
1093
		return 0;
	}

	/* Bind memory initialisation thread to a local node if possible */
	if (!cpumask_empty(cpumask))
		set_cpus_allowed_ptr(current, cpumask);
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108

	/* Sanity check boundaries */
	BUG_ON(pgdat->first_deferred_pfn < pgdat->node_start_pfn);
	BUG_ON(pgdat->first_deferred_pfn > pgdat_end_pfn(pgdat));
	pgdat->first_deferred_pfn = ULONG_MAX;

	/* Only the highest zone is deferred so find it */
	for (zid = 0; zid < MAX_NR_ZONES; zid++) {
		zone = pgdat->node_zones + zid;
		if (first_init_pfn < zone_end_pfn(zone))
			break;
	}

	for_each_mem_pfn_range(i, nid, &walk_start, &walk_end, NULL) {
		unsigned long pfn, end_pfn;
1109
		struct page *page = NULL;
1110
1111
1112
		struct page *free_base_page = NULL;
		unsigned long free_base_pfn = 0;
		int nr_to_free = 0;
1113
1114
1115
1116
1117
1118
1119
1120
1121

		end_pfn = min(walk_end, zone_end_pfn(zone));
		pfn = first_init_pfn;
		if (pfn < walk_start)
			pfn = walk_start;
		if (pfn < zone->zone_start_pfn)
			pfn = zone->zone_start_pfn;

		for (; pfn < end_pfn; pfn++) {
1122
			if (!pfn_valid_within(pfn))
1123
				goto free_range;
1124

1125
1126
1127
1128
1129
1130
1131
			/*
			 * Ensure pfn_valid is checked every
			 * MAX_ORDER_NR_PAGES for memory holes
			 */
			if ((pfn & (MAX_ORDER_NR_PAGES - 1)) == 0) {
				if (!pfn_valid(pfn)) {
					page = NULL;
1132
					goto free_range;
1133
1134
1135
1136
1137
				}
			}

			if (!meminit_pfn_in_nid(pfn, nid, &nid_init_state)) {
				page = NULL;