page_alloc.c 208 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/memremap.h>
47
#include <linux/stop_machine.h>
48 49
#include <linux/sort.h>
#include <linux/pfn.h>
50
#include <linux/backing-dev.h>
51
#include <linux/fault-inject.h>
52
#include <linux/page-isolation.h>
53
#include <linux/page_ext.h>
54
#include <linux/debugobjects.h>
55
#include <linux/kmemleak.h>
56
#include <linux/compaction.h>
57
#include <trace/events/kmem.h>
58
#include <trace/events/oom.h>
59
#include <linux/prefetch.h>
60
#include <linux/mm_inline.h>
61
#include <linux/migrate.h>
62
#include <linux/hugetlb.h>
63
#include <linux/sched/rt.h>
64
#include <linux/sched/mm.h>
65
#include <linux/page_owner.h>
66
#include <linux/kthread.h>
67
#include <linux/memcontrol.h>
Linus Torvalds's avatar
Linus Torvalds committed
68

69
#include <asm/sections.h>
Linus Torvalds's avatar
Linus Torvalds committed
70
#include <asm/tlbflush.h>
71
#include <asm/div64.h>
Linus Torvalds's avatar
Linus Torvalds committed
72 73
#include "internal.h"

74 75
/* prevent >1 _updater_ of zone percpu pageset ->high and ->batch fields */
static DEFINE_MUTEX(pcp_batch_high_lock);
76
#define MIN_PERCPU_PAGELIST_FRACTION	(8)
77

78 79 80 81 82
#ifdef CONFIG_USE_PERCPU_NUMA_NODE_ID
DEFINE_PER_CPU(int, numa_node);
EXPORT_PER_CPU_SYMBOL(numa_node);
#endif

83 84 85 86 87 88 89 90 91
#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_);
92
int _node_numa_mem_[MAX_NUMNODES];
93 94
#endif

95 96 97 98
/* work_structs for global per-cpu drains */
DEFINE_MUTEX(pcpu_drain_mutex);
DEFINE_PER_CPU(struct work_struct, pcpu_drain);

99
#ifdef CONFIG_GCC_PLUGIN_LATENT_ENTROPY
100
volatile unsigned long latent_entropy __latent_entropy;
101 102 103
EXPORT_SYMBOL(latent_entropy);
#endif

Linus Torvalds's avatar
Linus Torvalds committed
104
/*
105
 * Array of node states.
Linus Torvalds's avatar
Linus Torvalds committed
106
 */
107 108 109 110 111 112 113
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 } },
114 115 116
#endif
#ifdef CONFIG_MOVABLE_NODE
	[N_MEMORY] = { { [0] = 1UL } },
117 118 119 120 121 122
#endif
	[N_CPU] = { { [0] = 1UL } },
#endif	/* NUMA */
};
EXPORT_SYMBOL(node_states);

123 124 125
/* Protect totalram_pages and zone->managed_pages */
static DEFINE_SPINLOCK(managed_page_count_lock);

126
unsigned long totalram_pages __read_mostly;
127
unsigned long totalreserve_pages __read_mostly;
128
unsigned long totalcma_pages __read_mostly;
129

130
int percpu_pagelist_fraction;
131
gfp_t gfp_allowed_mask __read_mostly = GFP_BOOT_MASK;
Linus Torvalds's avatar
Linus Torvalds committed
132

133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150
/*
 * A cached value of the page's pageblock's migratetype, used when the page is
 * put on a pcplist. Used to avoid the pageblock migratetype lookup when
 * freeing from pcplists in most cases, at the cost of possibly becoming stale.
 * Also the migratetype set in the page does not necessarily match the pcplist
 * index, e.g. page might have MIGRATE_CMA set but be on a pcplist with any
 * other index - this ensures that it will be put on the correct CMA freelist.
 */
static inline int get_pcppage_migratetype(struct page *page)
{
	return page->index;
}

static inline void set_pcppage_migratetype(struct page *page, int migratetype)
{
	page->index = migratetype;
}

151 152 153 154 155 156 157 158 159
#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).
 */
160 161 162 163

static gfp_t saved_gfp_mask;

void pm_restore_gfp_mask(void)
164 165
{
	WARN_ON(!mutex_is_locked(&pm_mutex));
166 167 168 169
	if (saved_gfp_mask) {
		gfp_allowed_mask = saved_gfp_mask;
		saved_gfp_mask = 0;
	}
170 171
}

172
void pm_restrict_gfp_mask(void)
173 174
{
	WARN_ON(!mutex_is_locked(&pm_mutex));
175 176
	WARN_ON(saved_gfp_mask);
	saved_gfp_mask = gfp_allowed_mask;
177
	gfp_allowed_mask &= ~(__GFP_IO | __GFP_FS);
178
}
179 180 181

bool pm_suspended_storage(void)
{
182
	if ((gfp_allowed_mask & (__GFP_IO | __GFP_FS)) == (__GFP_IO | __GFP_FS))
183 184 185
		return false;
	return true;
}
186 187
#endif /* CONFIG_PM_SLEEP */

188
#ifdef CONFIG_HUGETLB_PAGE_SIZE_VARIABLE
189
unsigned int pageblock_order __read_mostly;
190 191
#endif

192
static void __free_pages_ok(struct page *page, unsigned int order);
193

Linus Torvalds's avatar
Linus Torvalds committed
194 195 196 197 198 199
/*
 * 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
200
 *	HIGHMEM allocation will leave (224M+784M)/256 of ram reserved in ZONE_DMA
201 202 203
 *
 * 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
204
 */
205
int sysctl_lowmem_reserve_ratio[MAX_NR_ZONES-1] = {
206
#ifdef CONFIG_ZONE_DMA
207
	 256,
208
#endif
209
#ifdef CONFIG_ZONE_DMA32
210
	 256,
211
#endif
212
#ifdef CONFIG_HIGHMEM
Mel Gorman's avatar
Mel Gorman committed
213
	 32,
214
#endif
Mel Gorman's avatar
Mel Gorman committed
215
	 32,
216
};
Linus Torvalds's avatar
Linus Torvalds committed
217 218 219

EXPORT_SYMBOL(totalram_pages);

220
static char * const zone_names[MAX_NR_ZONES] = {
221
#ifdef CONFIG_ZONE_DMA
222
	 "DMA",
223
#endif
224
#ifdef CONFIG_ZONE_DMA32
225
	 "DMA32",
226
#endif
227
	 "Normal",
228
#ifdef CONFIG_HIGHMEM
Mel Gorman's avatar
Mel Gorman committed
229
	 "HighMem",
230
#endif
Mel Gorman's avatar
Mel Gorman committed
231
	 "Movable",
232 233 234
#ifdef CONFIG_ZONE_DEVICE
	 "Device",
#endif
235 236
};

237 238 239 240 241 242 243 244 245 246 247 248 249
char * const migratetype_names[MIGRATE_TYPES] = {
	"Unmovable",
	"Movable",
	"Reclaimable",
	"HighAtomic",
#ifdef CONFIG_CMA
	"CMA",
#endif
#ifdef CONFIG_MEMORY_ISOLATION
	"Isolate",
#endif
};

250 251 252 253 254 255
compound_page_dtor * const compound_page_dtors[] = {
	NULL,
	free_compound_page,
#ifdef CONFIG_HUGETLB_PAGE
	free_huge_page,
#endif
256 257 258
#ifdef CONFIG_TRANSPARENT_HUGEPAGE
	free_transhuge_page,
#endif
259 260
};

Linus Torvalds's avatar
Linus Torvalds committed
261
int min_free_kbytes = 1024;
262
int user_min_free_kbytes = -1;
263
int watermark_scale_factor = 10;
Linus Torvalds's avatar
Linus Torvalds committed
264

265 266
static unsigned long __meminitdata nr_kernel_pages;
static unsigned long __meminitdata nr_all_pages;
267
static unsigned long __meminitdata dma_reserve;
Linus Torvalds's avatar
Linus Torvalds committed
268

Tejun Heo's avatar
Tejun Heo committed
269 270 271 272 273 274
#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];
275
static bool mirrored_kernelcore;
Tejun Heo's avatar
Tejun Heo committed
276 277 278 279 280

/* 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 */
281

Miklos Szeredi's avatar
Miklos Szeredi committed
282 283
#if MAX_NUMNODES > 1
int nr_node_ids __read_mostly = MAX_NUMNODES;
284
int nr_online_nodes __read_mostly = 1;
Miklos Szeredi's avatar
Miklos Szeredi committed
285
EXPORT_SYMBOL(nr_node_ids);
286
EXPORT_SYMBOL(nr_online_nodes);
Miklos Szeredi's avatar
Miklos Szeredi committed
287 288
#endif

289 290
int page_group_by_mobility_disabled __read_mostly;

291 292 293 294 295 296 297
#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 */
298
static inline bool __meminit early_page_uninitialised(unsigned long pfn)
299
{
300 301 302
	int nid = early_pfn_to_nid(pfn);

	if (node_online(nid) && pfn >= NODE_DATA(nid)->first_deferred_pfn)
303 304 305 306 307 308 309 310 311 312 313 314 315
		return true;

	return false;
}

/*
 * 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)
{
316 317
	unsigned long max_initialise;

318 319 320
	/* Always populate low zones for address-contrained allocations */
	if (zone_end < pgdat_end_pfn(pgdat))
		return true;
321 322 323 324 325 326
	/*
	 * Initialise at least 2G of a node but also take into account that
	 * two large system hashes that can take up 1GB for 0.25TB/node.
	 */
	max_initialise = max(2UL << (30 - PAGE_SHIFT),
		(pgdat->node_spanned_pages >> 8));
327 328

	(*nr_initialised)++;
329
	if ((*nr_initialised > max_initialise) &&
330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354
	    (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;
}

static inline bool update_defer_init(pg_data_t *pgdat,
				unsigned long pfn, unsigned long zone_end,
				unsigned long *nr_initialised)
{
	return true;
}
#endif

355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454
/* Return a pointer to the bitmap storing bits affecting a block of pages */
static inline unsigned long *get_pageblock_bitmap(struct page *page,
							unsigned long pfn)
{
#ifdef CONFIG_SPARSEMEM
	return __pfn_to_section(pfn)->pageblock_flags;
#else
	return page_zone(page)->pageblock_flags;
#endif /* CONFIG_SPARSEMEM */
}

static inline int pfn_to_bitidx(struct page *page, unsigned long pfn)
{
#ifdef CONFIG_SPARSEMEM
	pfn &= (PAGES_PER_SECTION-1);
	return (pfn >> pageblock_order) * NR_PAGEBLOCK_BITS;
#else
	pfn = pfn - round_down(page_zone(page)->zone_start_pfn, pageblock_nr_pages);
	return (pfn >> pageblock_order) * NR_PAGEBLOCK_BITS;
#endif /* CONFIG_SPARSEMEM */
}

/**
 * get_pfnblock_flags_mask - Return the requested group of flags for the pageblock_nr_pages block of pages
 * @page: The page within the block of interest
 * @pfn: The target page frame number
 * @end_bitidx: The last bit of interest to retrieve
 * @mask: mask of bits that the caller is interested in
 *
 * Return: pageblock_bits flags
 */
static __always_inline unsigned long __get_pfnblock_flags_mask(struct page *page,
					unsigned long pfn,
					unsigned long end_bitidx,
					unsigned long mask)
{
	unsigned long *bitmap;
	unsigned long bitidx, word_bitidx;
	unsigned long word;

	bitmap = get_pageblock_bitmap(page, pfn);
	bitidx = pfn_to_bitidx(page, pfn);
	word_bitidx = bitidx / BITS_PER_LONG;
	bitidx &= (BITS_PER_LONG-1);

	word = bitmap[word_bitidx];
	bitidx += end_bitidx;
	return (word >> (BITS_PER_LONG - bitidx - 1)) & mask;
}

unsigned long get_pfnblock_flags_mask(struct page *page, unsigned long pfn,
					unsigned long end_bitidx,
					unsigned long mask)
{
	return __get_pfnblock_flags_mask(page, pfn, end_bitidx, mask);
}

static __always_inline int get_pfnblock_migratetype(struct page *page, unsigned long pfn)
{
	return __get_pfnblock_flags_mask(page, pfn, PB_migrate_end, MIGRATETYPE_MASK);
}

/**
 * set_pfnblock_flags_mask - Set the requested group of flags for a pageblock_nr_pages block of pages
 * @page: The page within the block of interest
 * @flags: The flags to set
 * @pfn: The target page frame number
 * @end_bitidx: The last bit of interest
 * @mask: mask of bits that the caller is interested in
 */
void set_pfnblock_flags_mask(struct page *page, unsigned long flags,
					unsigned long pfn,
					unsigned long end_bitidx,
					unsigned long mask)
{
	unsigned long *bitmap;
	unsigned long bitidx, word_bitidx;
	unsigned long old_word, word;

	BUILD_BUG_ON(NR_PAGEBLOCK_BITS != 4);

	bitmap = get_pageblock_bitmap(page, pfn);
	bitidx = pfn_to_bitidx(page, pfn);
	word_bitidx = bitidx / BITS_PER_LONG;
	bitidx &= (BITS_PER_LONG-1);

	VM_BUG_ON_PAGE(!zone_spans_pfn(page_zone(page), pfn), page);

	bitidx += end_bitidx;
	mask <<= (BITS_PER_LONG - bitidx - 1);
	flags <<= (BITS_PER_LONG - bitidx - 1);

	word = READ_ONCE(bitmap[word_bitidx]);
	for (;;) {
		old_word = cmpxchg(&bitmap[word_bitidx], word, (word & ~mask) | flags);
		if (word == old_word)
			break;
		word = old_word;
	}
}
455

456
void set_pageblock_migratetype(struct page *page, int migratetype)
457
{
458 459
	if (unlikely(page_group_by_mobility_disabled &&
		     migratetype < MIGRATE_PCPTYPES))
460 461
		migratetype = MIGRATE_UNMOVABLE;

462 463 464 465
	set_pageblock_flags_group(page, (unsigned long)migratetype,
					PB_migrate, PB_migrate_end);
}

Nick Piggin's avatar
Nick Piggin committed
466
#ifdef CONFIG_DEBUG_VM
467
static int page_outside_zone_boundaries(struct zone *zone, struct page *page)
Linus Torvalds's avatar
Linus Torvalds committed
468
{
469 470 471
	int ret = 0;
	unsigned seq;
	unsigned long pfn = page_to_pfn(page);
472
	unsigned long sp, start_pfn;
473

474 475
	do {
		seq = zone_span_seqbegin(zone);
476 477
		start_pfn = zone->zone_start_pfn;
		sp = zone->spanned_pages;
478
		if (!zone_spans_pfn(zone, pfn))
479 480 481
			ret = 1;
	} while (zone_span_seqretry(zone, seq));

482
	if (ret)
483 484 485
		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);
486

487
	return ret;
488 489 490 491
}

static int page_is_consistent(struct zone *zone, struct page *page)
{
492
	if (!pfn_valid_within(page_to_pfn(page)))
493
		return 0;
Linus Torvalds's avatar
Linus Torvalds committed
494
	if (zone != page_zone(page))
495 496 497 498 499 500 501 502 503 504
		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
505
		return 1;
506 507 508
	if (!page_is_consistent(zone, page))
		return 1;

Linus Torvalds's avatar
Linus Torvalds committed
509 510
	return 0;
}
Nick Piggin's avatar
Nick Piggin committed
511 512 513 514 515 516 517
#else
static inline int bad_range(struct zone *zone, struct page *page)
{
	return 0;
}
#endif

518 519
static void bad_page(struct page *page, const char *reason,
		unsigned long bad_flags)
Linus Torvalds's avatar
Linus Torvalds committed
520
{
521 522 523 524 525 526 527 528 529 530 531 532 533 534
	static unsigned long resume;
	static unsigned long nr_shown;
	static unsigned long nr_unshown;

	/*
	 * 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) {
535
			pr_alert(
536
			      "BUG: Bad page state: %lu messages suppressed\n",
537 538 539 540 541 542 543 544
				nr_unshown);
			nr_unshown = 0;
		}
		nr_shown = 0;
	}
	if (nr_shown++ == 0)
		resume = jiffies + 60 * HZ;

545
	pr_alert("BUG: Bad page state in process %s  pfn:%05lx\n",
546
		current->comm, page_to_pfn(page));
547 548 549 550 551
	__dump_page(page, reason);
	bad_flags &= page->flags;
	if (bad_flags)
		pr_alert("bad because of flags: %#lx(%pGp)\n",
						bad_flags, &bad_flags);
552
	dump_page_owner(page);
553

554
	print_modules();
Linus Torvalds's avatar
Linus Torvalds committed
555
	dump_stack();
556
out:
557
	/* Leave bad fields for debug, except PageBuddy could make trouble */
558
	page_mapcount_reset(page); /* remove PageBuddy */
559
	add_taint(TAINT_BAD_PAGE, LOCKDEP_NOW_UNRELIABLE);
Linus Torvalds's avatar
Linus Torvalds committed
560 561 562 563 564
}

/*
 * Higher-order pages are called "compound pages".  They are structured thusly:
 *
565
 * The first PAGE_SIZE page is called the "head page" and have PG_head set.
Linus Torvalds's avatar
Linus Torvalds committed
566
 *
567 568
 * The remaining PAGE_SIZE pages are called "tail pages". PageTail() is encoded
 * in bit 0 of page->compound_head. The rest of bits is pointer to head page.
Linus Torvalds's avatar
Linus Torvalds committed
569
 *
570 571
 * The first tail page's ->compound_dtor holds the offset in array of compound
 * page destructors. See compound_page_dtors.
Linus Torvalds's avatar
Linus Torvalds committed
572
 *
573
 * The first tail page's ->compound_order holds the order of allocation.
574
 * This usage means that zero-order pages may not be compound.
Linus Torvalds's avatar
Linus Torvalds committed
575
 */
576

577
void free_compound_page(struct page *page)
578
{
579
	__free_pages_ok(page, compound_order(page));
580 581
}

582
void prep_compound_page(struct page *page, unsigned int order)
583 584 585 586
{
	int i;
	int nr_pages = 1 << order;

587
	set_compound_page_dtor(page, COMPOUND_PAGE_DTOR);
588 589 590 591
	set_compound_order(page, order);
	__SetPageHead(page);
	for (i = 1; i < nr_pages; i++) {
		struct page *p = page + i;
592
		set_page_count(p, 0);
593
		p->mapping = TAIL_MAPPING;
594
		set_compound_head(p, page);
595
	}
596
	atomic_set(compound_mapcount_ptr(page), -1);
597 598
}

599 600
#ifdef CONFIG_DEBUG_PAGEALLOC
unsigned int _debug_guardpage_minorder;
601 602
bool _debug_pagealloc_enabled __read_mostly
			= IS_ENABLED(CONFIG_DEBUG_PAGEALLOC_ENABLE_DEFAULT);
603
EXPORT_SYMBOL(_debug_pagealloc_enabled);
604 605
bool _debug_guardpage_enabled __read_mostly;

606 607 608 609
static int __init early_debug_pagealloc(char *buf)
{
	if (!buf)
		return -EINVAL;
610
	return kstrtobool(buf, &_debug_pagealloc_enabled);
611 612 613
}
early_param("debug_pagealloc", early_debug_pagealloc);

614 615
static bool need_debug_guardpage(void)
{
616 617 618 619
	/* If we don't use debug_pagealloc, we don't need guard page */
	if (!debug_pagealloc_enabled())
		return false;

620 621 622
	if (!debug_guardpage_minorder())
		return false;

623 624 625 626 627
	return true;
}

static void init_debug_guardpage(void)
{
628 629 630
	if (!debug_pagealloc_enabled())
		return;

631 632 633
	if (!debug_guardpage_minorder())
		return;

634 635 636 637 638 639 640
	_debug_guardpage_enabled = true;
}

struct page_ext_operations debug_guardpage_ops = {
	.need = need_debug_guardpage,
	.init = init_debug_guardpage,
};
641 642 643 644 645 646

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

	if (kstrtoul(buf, 10, &res) < 0 ||  res > MAX_ORDER / 2) {
647
		pr_err("Bad debug_guardpage_minorder value\n");
648 649 650
		return 0;
	}
	_debug_guardpage_minorder = res;
651
	pr_info("Setting debug_guardpage_minorder to %lu\n", res);
652 653
	return 0;
}
654
early_param("debug_guardpage_minorder", debug_guardpage_minorder_setup);
655

656
static inline bool set_page_guard(struct zone *zone, struct page *page,
657
				unsigned int order, int migratetype)
658
{
659 660 661
	struct page_ext *page_ext;

	if (!debug_guardpage_enabled())
662 663 664 665
		return false;

	if (order >= debug_guardpage_minorder())
		return false;
666 667

	page_ext = lookup_page_ext(page);
668
	if (unlikely(!page_ext))
669
		return false;
670

671 672
	__set_bit(PAGE_EXT_DEBUG_GUARD, &page_ext->flags);

673 674 675 676
	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);
677 678

	return true;
679 680
}

681 682
static inline void clear_page_guard(struct zone *zone, struct page *page,
				unsigned int order, int migratetype)
683
{
684 685 686 687 688 689
	struct page_ext *page_ext;

	if (!debug_guardpage_enabled())
		return;

	page_ext = lookup_page_ext(page);
690 691 692
	if (unlikely(!page_ext))
		return;

693 694
	__clear_bit(PAGE_EXT_DEBUG_GUARD, &page_ext->flags);

695 696 697
	set_page_private(page, 0);
	if (!is_migrate_isolate(migratetype))
		__mod_zone_freepage_state(zone, (1 << order), migratetype);
698 699
}
#else
700
struct page_ext_operations debug_guardpage_ops;
701 702
static inline bool set_page_guard(struct zone *zone, struct page *page,
			unsigned int order, int migratetype) { return false; }
703 704
static inline void clear_page_guard(struct zone *zone, struct page *page,
				unsigned int order, int migratetype) {}
705 706
#endif

707
static inline void set_page_order(struct page *page, unsigned int order)
708
{
709
	set_page_private(page, order);
710
	__SetPageBuddy(page);
Linus Torvalds's avatar
Linus Torvalds committed
711 712 713 714
}

static inline void rmv_page_order(struct page *page)
{
715
	__ClearPageBuddy(page);
716
	set_page_private(page, 0);
Linus Torvalds's avatar
Linus Torvalds committed
717 718 719 720 721
}

/*
 * This function checks whether a page is free && is the buddy
 * we can do coalesce a page and its buddy if
722
 * (a) the buddy is not in a hole (check before calling!) &&
723
 * (b) the buddy is in the buddy system &&
724 725
 * (c) a page and its buddy have the same order &&
 * (d) a page and its buddy are in the same zone.
726
 *
727 728 729 730
 * 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
731
 *
732
 * For recording page's order, we use page_private(page).
Linus Torvalds's avatar
Linus Torvalds committed
733
 */
734
static inline int page_is_buddy(struct page *page, struct page *buddy,
735
							unsigned int order)
Linus Torvalds's avatar
Linus Torvalds committed
736
{
737
	if (page_is_guard(buddy) && page_order(buddy) == order) {
738 739 740
		if (page_zone_id(page) != page_zone_id(buddy))
			return 0;

741 742
		VM_BUG_ON_PAGE(page_count(buddy) != 0, buddy);

743 744 745
		return 1;
	}

746
	if (PageBuddy(buddy) && page_order(buddy) == order) {
747 748 749 750 751 752 753 754
		/*
		 * 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;

755 756
		VM_BUG_ON_PAGE(page_count(buddy) != 0, buddy);

757
		return 1;
758
	}
759
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
760 761 762 763 764 765 766 767 768 769 770 771 772 773 774
}

/*
 * 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
775 776 777
 * 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
778
 * So when we are allocating or freeing one, we can derive the state of the
779 780
 * 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
781
 * If a block is freed, and its buddy is also free, then this
782
 * triggers coalescing into a block of larger size.
Linus Torvalds's avatar
Linus Torvalds committed
783
 *
784
 * -- nyc
Linus Torvalds's avatar
Linus Torvalds committed
785 786
 */

Nick Piggin's avatar
Nick Piggin committed
787
static inline void __free_one_page(struct page *page,
788
		unsigned long pfn,
789 790
		struct zone *zone, unsigned int order,
		int migratetype)
Linus Torvalds's avatar
Linus Torvalds committed
791
{
792 793
	unsigned long combined_pfn;
	unsigned long uninitialized_var(buddy_pfn);
794
	struct page *buddy;
795 796 797
	unsigned int max_order;

	max_order = min_t(unsigned int, MAX_ORDER, pageblock_order + 1);
Linus Torvalds's avatar
Linus Torvalds committed
798

799
	VM_BUG_ON(!zone_is_initialized(zone));
800
	VM_BUG_ON_PAGE(page->flags & PAGE_FLAGS_CHECK_AT_PREP, page);
Linus Torvalds's avatar
Linus Torvalds committed
801

802
	VM_BUG_ON(migratetype == -1);
803
	if (likely(!is_migrate_isolate(migratetype)))
804
		__mod_zone_freepage_state(zone, 1 << order, migratetype);
805

806
	VM_BUG_ON_PAGE(pfn & ((1 << order) - 1), page);
807
	VM_BUG_ON_PAGE(bad_range(zone, page), page);
Linus Torvalds's avatar
Linus Torvalds committed
808

809
continue_merging:
810
	while (order < max_order - 1) {
811 812
		buddy_pfn = __find_buddy_pfn(pfn, order);
		buddy = page + (buddy_pfn - pfn);
813 814 815

		if (!pfn_valid_within(buddy_pfn))
			goto done_merging;
816
		if (!page_is_buddy(page, buddy, order))
817
			goto done_merging;
818 819 820 821 822
		/*
		 * 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)) {
823
			clear_page_guard(zone, buddy, order, migratetype);
824 825 826 827 828
		} else {
			list_del(&buddy->lru);
			zone->free_area[order].nr_free--;
			rmv_page_order(buddy);
		}
829 830 831
		combined_pfn = buddy_pfn & pfn;
		page = page + (combined_pfn - pfn);
		pfn = combined_pfn;
Linus Torvalds's avatar
Linus Torvalds committed
832 833
		order++;
	}
834 835 836 837 838 839 840 841 842 843 844 845
	if (max_order < MAX_ORDER) {
		/* If we are here, it means order is >= pageblock_order.
		 * We want to prevent merge between freepages on isolate
		 * pageblock and normal pageblock. Without this, pageblock
		 * isolation could cause incorrect freepage or CMA accounting.
		 *
		 * We don't want to hit this code for the more frequent
		 * low-order merging.
		 */
		if (unlikely(has_isolate_pageblock(zone))) {
			int buddy_mt;

846 847
			buddy_pfn = __find_buddy_pfn(pfn, order);
			buddy = page + (buddy_pfn - pfn);
848 849 850 851 852 853 854 855 856 857 858 859
			buddy_mt = get_pageblock_migratetype(buddy);

			if (migratetype != buddy_mt
					&& (is_migrate_isolate(migratetype) ||
						is_migrate_isolate(buddy_mt)))
				goto done_merging;
		}
		max_order++;
		goto continue_merging;
	}

done_merging:
Linus Torvalds's avatar
Linus Torvalds committed
860
	set_page_order(page, order);
861 862 863 864 865 866 867 868 869

	/*
	 * 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
	 */
870
	if ((order < MAX_ORDER-2) && pfn_valid_within(buddy_pfn)) {
871
		struct page *higher_page, *higher_buddy;
872 873 874 875
		combined_pfn = buddy_pfn & pfn;
		higher_page = page + (combined_pfn - pfn);
		buddy_pfn = __find_buddy_pfn(combined_pfn, order + 1);
		higher_buddy = higher_page + (buddy_pfn - combined_pfn);
876 877
		if (pfn_valid_within(buddy_pfn) &&
		    page_is_buddy(higher_page, higher_buddy, order + 1)) {
878 879 880 881 882 883 884 885
			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
886 887 888
	zone->free_area[order].nr_free++;
}

889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910
/*
 * A bad page could be due to a number of fields. Instead of multiple branches,
 * try and check multiple fields with one check. The caller must do a detailed
 * check if necessary.
 */
static inline bool page_expected_state(struct page *page,
					unsigned long check_flags)
{
	if (unlikely(atomic_read(&page->_mapcount) != -1))
		return false;

	if (unlikely((unsigned long)page->mapping |
			page_ref_count(page) |
#ifdef CONFIG_MEMCG
			(unsigned long)page->mem_cgroup |
#endif
			(page->flags & check_flags)))
		return false;

	return true;
}

911
static void free_pages_check_bad(struct page *page)
Linus Torvalds's avatar
Linus Torvalds committed
912
{
913 914 915 916 917
	const char *bad_reason;
	unsigned long bad_flags;

	bad_reason = NULL;
	bad_flags = 0;
918

919
	if (unlikely(atomic_read(&page->_mapcount) != -1))
920 921 922
		bad_reason = "nonzero mapcount";
	if (unlikely(page->mapping != NULL))
		bad_reason = "non-NULL mapping";
923
	if (unlikely(page_ref_count(page) != 0))
924
		bad_reason = "nonzero _refcount";
925 926 927 928
	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;
	}
929 930 931 932
#ifdef CONFIG_MEMCG
	if (unlikely(page->mem_cgroup))
		bad_reason = "page still charged to cgroup";
#endif
933
	bad_page(page, bad_reason, bad_flags);
934 935 936 937
}

static inline int free_pages_check(struct page *page)
{
938
	if (likely(page_expected_state(page, PAGE_FLAGS_CHECK_AT_FREE)))
939 940 941 942
		return 0;

	/* Something has gone sideways, find it */
	free_pages_check_bad(page);
943
	return 1;
Linus Torvalds's avatar
Linus Torvalds committed
944 945
}

946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995
static int free_tail_pages_check(struct page *head_page, struct page *page)
{
	int ret = 1;

	/*
	 * We rely page->lru.next never has bit 0 set, unless the page
	 * is PageTail(). Let's make sure that's true even for poisoned ->lru.
	 */
	BUILD_BUG_ON((unsigned long)LIST_POISON1 & 1);

	if (!IS_ENABLED(CONFIG_DEBUG_VM)) {
		ret = 0;
		goto out;
	}
	switch (page - head_page) {
	case 1:
		/* the first tail page: ->mapping is compound_mapcount() */
		if (unlikely(compound_mapcount(page))) {
			bad_page(page, "nonzero compound_mapcount", 0);
			goto out;
		}
		break;
	case 2:
		/*
		 * the second tail page: ->mapping is
		 * page_deferred_list().next -- ignore value.
		 */
		break;
	default:
		if (page->mapping != TAIL_MAPPING) {
			bad_page(page, "corrupted mapping in tail page", 0);
			goto out;
		}
		break;
	}
	if (unlikely(!PageTail(page))) {
		bad_page(page, "PageTail not set", 0);
		goto out;
	}
	if (unlikely(compound_head(page) != head_page)) {
		bad_page(page, "compound_head not consistent", 0);
		goto out;
	}
	ret = 0;
out:
	page->mapping = NULL;
	clear_compound_head(page);
	return ret;
}

996 997
static __always_inline bool free_pages_prepare(struct page *page,
					unsigned int order, bool check_free)
998
{
999
	int bad = 0;
1000 1001 1002

	VM_BUG_ON_PAGE(PageTail(page), page);

1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014
	trace_mm_page_free(page, order);
	kmemcheck_free_shadow(page, order);

	/*
	 * Check tail pages before head page information is cleared to
	 * avoid checking PageCompound for order-0 pages.
	 */
	if (unlikely(order)) {
		bool compound = PageCompound(page);
		int i;

		VM_BUG_ON_PAGE(compound && compound_order(page) != order, page);
1015

1016 1017
		if (compound)
			ClearPageDoubleMap(page);
1018 1019 1020 1021 1022 1023 1024 1025 1026 1027
		for (i = 1; i < (1 << order); i++) {
			if (compound)
				bad += free_tail_pages_check(page, page + i);
			if (unlikely(free_pages_check(page + i))) {
				bad++;
				continue;
			}
			(page + i)->flags &= ~PAGE_FLAGS_CHECK_AT_PREP;
		}
	}
1028
	if (PageMappingFlags(page))
1029
		page->mapping = NULL;