page_alloc.c 141 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 24
#include <linux/bootmem.h>
#include <linux/compiler.h>
25
#include <linux/kernel.h>
26
#include <linux/kmemcheck.h>
Linus Torvalds's avatar
Linus Torvalds committed
27 28 29 30 31
#include <linux/module.h>
#include <linux/suspend.h>
#include <linux/pagevec.h>
#include <linux/blkdev.h>
#include <linux/slab.h>
32
#include <linux/oom.h>
Linus Torvalds's avatar
Linus Torvalds committed
33 34 35 36 37
#include <linux/notifier.h>
#include <linux/topology.h>
#include <linux/sysctl.h>
#include <linux/cpu.h>
#include <linux/cpuset.h>
38
#include <linux/memory_hotplug.h>
Linus Torvalds's avatar
Linus Torvalds committed
39 40
#include <linux/nodemask.h>
#include <linux/vmalloc.h>
41
#include <linux/mempolicy.h>
42
#include <linux/stop_machine.h>
43 44
#include <linux/sort.h>
#include <linux/pfn.h>
45
#include <linux/backing-dev.h>
46
#include <linux/fault-inject.h>
47
#include <linux/page-isolation.h>
48
#include <linux/page_cgroup.h>
49
#include <linux/debugobjects.h>
50
#include <linux/kmemleak.h>
51
#include <linux/memory.h>
52
#include <trace/events/kmem.h>
Linus Torvalds's avatar
Linus Torvalds committed
53 54

#include <asm/tlbflush.h>
55
#include <asm/div64.h>
Linus Torvalds's avatar
Linus Torvalds committed
56 57 58
#include "internal.h"

/*
59
 * Array of node states.
Linus Torvalds's avatar
Linus Torvalds committed
60
 */
61 62 63 64 65 66 67 68 69 70 71 72 73
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 } },
#endif
	[N_CPU] = { { [0] = 1UL } },
#endif	/* NUMA */
};
EXPORT_SYMBOL(node_states);

74
unsigned long totalram_pages __read_mostly;
75
unsigned long totalreserve_pages __read_mostly;
76
int percpu_pagelist_fraction;
77
gfp_t gfp_allowed_mask __read_mostly = GFP_BOOT_MASK;
Linus Torvalds's avatar
Linus Torvalds committed
78

79 80 81 82
#ifdef CONFIG_HUGETLB_PAGE_SIZE_VARIABLE
int pageblock_order __read_mostly;
#endif

83
static void __free_pages_ok(struct page *page, unsigned int order);
84

Linus Torvalds's avatar
Linus Torvalds committed
85 86 87 88 89 90 91
/*
 * 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
 *	HIGHMEM allocation will (224M+784M)/256 of ram reserved in ZONE_DMA
92 93 94
 *
 * 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
95
 */
96
int sysctl_lowmem_reserve_ratio[MAX_NR_ZONES-1] = {
97
#ifdef CONFIG_ZONE_DMA
98
	 256,
99
#endif
100
#ifdef CONFIG_ZONE_DMA32
101
	 256,
102
#endif
103
#ifdef CONFIG_HIGHMEM
Mel Gorman's avatar
Mel Gorman committed
104
	 32,
105
#endif
Mel Gorman's avatar
Mel Gorman committed
106
	 32,
107
};
Linus Torvalds's avatar
Linus Torvalds committed
108 109 110

EXPORT_SYMBOL(totalram_pages);

111
static char * const zone_names[MAX_NR_ZONES] = {
112
#ifdef CONFIG_ZONE_DMA
113
	 "DMA",
114
#endif
115
#ifdef CONFIG_ZONE_DMA32
116
	 "DMA32",
117
#endif
118
	 "Normal",
119
#ifdef CONFIG_HIGHMEM
Mel Gorman's avatar
Mel Gorman committed
120
	 "HighMem",
121
#endif
Mel Gorman's avatar
Mel Gorman committed
122
	 "Movable",
123 124
};

Linus Torvalds's avatar
Linus Torvalds committed
125 126
int min_free_kbytes = 1024;

127 128
static unsigned long __meminitdata nr_kernel_pages;
static unsigned long __meminitdata nr_all_pages;
129
static unsigned long __meminitdata dma_reserve;
Linus Torvalds's avatar
Linus Torvalds committed
130

131 132
#ifdef CONFIG_ARCH_POPULATES_NODE_MAP
  /*
Simon Arlott's avatar
Simon Arlott committed
133
   * MAX_ACTIVE_REGIONS determines the maximum number of distinct
134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151
   * ranges of memory (RAM) that may be registered with add_active_range().
   * Ranges passed to add_active_range() will be merged if possible
   * so the number of times add_active_range() can be called is
   * related to the number of nodes and the number of holes
   */
  #ifdef CONFIG_MAX_ACTIVE_REGIONS
    /* Allow an architecture to set MAX_ACTIVE_REGIONS to save memory */
    #define MAX_ACTIVE_REGIONS CONFIG_MAX_ACTIVE_REGIONS
  #else
    #if MAX_NUMNODES >= 32
      /* If there can be many nodes, allow up to 50 holes per node */
      #define MAX_ACTIVE_REGIONS (MAX_NUMNODES*50)
    #else
      /* By default, allow up to 256 distinct regions */
      #define MAX_ACTIVE_REGIONS 256
    #endif
  #endif

152 153 154 155
  static struct node_active_region __meminitdata early_node_map[MAX_ACTIVE_REGIONS];
  static int __meminitdata nr_nodemap_entries;
  static unsigned long __meminitdata arch_zone_lowest_possible_pfn[MAX_NR_ZONES];
  static unsigned long __meminitdata arch_zone_highest_possible_pfn[MAX_NR_ZONES];
Adrian Bunk's avatar
Adrian Bunk committed
156
  static unsigned long __initdata required_kernelcore;
157
  static unsigned long __initdata required_movablecore;
Adrian Bunk's avatar
Adrian Bunk committed
158
  static unsigned long __meminitdata zone_movable_pfn[MAX_NUMNODES];
Mel Gorman's avatar
Mel Gorman committed
159 160 161 162

  /* movable_zone is the "real" zone pages in ZONE_MOVABLE are taken from */
  int movable_zone;
  EXPORT_SYMBOL(movable_zone);
163 164
#endif /* CONFIG_ARCH_POPULATES_NODE_MAP */

Miklos Szeredi's avatar
Miklos Szeredi committed
165 166
#if MAX_NUMNODES > 1
int nr_node_ids __read_mostly = MAX_NUMNODES;
167
int nr_online_nodes __read_mostly = 1;
Miklos Szeredi's avatar
Miklos Szeredi committed
168
EXPORT_SYMBOL(nr_node_ids);
169
EXPORT_SYMBOL(nr_online_nodes);
Miklos Szeredi's avatar
Miklos Szeredi committed
170 171
#endif

172 173
int page_group_by_mobility_disabled __read_mostly;

174 175
static void set_pageblock_migratetype(struct page *page, int migratetype)
{
176 177 178 179

	if (unlikely(page_group_by_mobility_disabled))
		migratetype = MIGRATE_UNMOVABLE;

180 181 182 183
	set_pageblock_flags_group(page, (unsigned long)migratetype,
					PB_migrate, PB_migrate_end);
}

184 185
bool oom_killer_disabled __read_mostly;

Nick Piggin's avatar
Nick Piggin committed
186
#ifdef CONFIG_DEBUG_VM
187
static int page_outside_zone_boundaries(struct zone *zone, struct page *page)
Linus Torvalds's avatar
Linus Torvalds committed
188
{
189 190 191
	int ret = 0;
	unsigned seq;
	unsigned long pfn = page_to_pfn(page);
192

193 194 195 196 197 198 199 200 201
	do {
		seq = zone_span_seqbegin(zone);
		if (pfn >= zone->zone_start_pfn + zone->spanned_pages)
			ret = 1;
		else if (pfn < zone->zone_start_pfn)
			ret = 1;
	} while (zone_span_seqretry(zone, seq));

	return ret;
202 203 204 205
}

static int page_is_consistent(struct zone *zone, struct page *page)
{
206
	if (!pfn_valid_within(page_to_pfn(page)))
207
		return 0;
Linus Torvalds's avatar
Linus Torvalds committed
208
	if (zone != page_zone(page))
209 210 211 212 213 214 215 216 217 218
		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
219
		return 1;
220 221 222
	if (!page_is_consistent(zone, page))
		return 1;

Linus Torvalds's avatar
Linus Torvalds committed
223 224
	return 0;
}
Nick Piggin's avatar
Nick Piggin committed
225 226 227 228 229 230 231
#else
static inline int bad_range(struct zone *zone, struct page *page)
{
	return 0;
}
#endif

232
static void bad_page(struct page *page)
Linus Torvalds's avatar
Linus Torvalds committed
233
{
234 235 236 237
	static unsigned long resume;
	static unsigned long nr_shown;
	static unsigned long nr_unshown;

238 239 240 241 242 243
	/* Don't complain about poisoned pages */
	if (PageHWPoison(page)) {
		__ClearPageBuddy(page);
		return;
	}

244 245 246 247 248 249 250 251 252 253
	/*
	 * 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) {
254 255
			printk(KERN_ALERT
			      "BUG: Bad page state: %lu messages suppressed\n",
256 257 258 259 260 261 262 263
				nr_unshown);
			nr_unshown = 0;
		}
		nr_shown = 0;
	}
	if (nr_shown++ == 0)
		resume = jiffies + 60 * HZ;

264
	printk(KERN_ALERT "BUG: Bad page state in process %s  pfn:%05lx\n",
265
		current->comm, page_to_pfn(page));
266
	printk(KERN_ALERT
267 268 269 270
		"page:%p flags:%p count:%d mapcount:%d mapping:%p index:%lx\n",
		page, (void *)page->flags, page_count(page),
		page_mapcount(page), page->mapping, page->index);

Linus Torvalds's avatar
Linus Torvalds committed
271
	dump_stack();
272
out:
273 274
	/* Leave bad fields for debug, except PageBuddy could make trouble */
	__ClearPageBuddy(page);
275
	add_taint(TAINT_BAD_PAGE);
Linus Torvalds's avatar
Linus Torvalds committed
276 277 278 279 280 281 282 283 284 285 286 287
}

/*
 * 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".
 *
 * All pages have PG_compound set.  All pages have their ->private pointing at
 * the head page (even the head page has this).
 *
288 289 290
 * 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
291
 */
292 293 294

static void free_compound_page(struct page *page)
{
295
	__free_pages_ok(page, compound_order(page));
296 297
}

298
void prep_compound_page(struct page *page, unsigned long order)
299 300 301 302 303 304 305 306 307 308 309 310 311 312 313
{
	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;

		__SetPageTail(p);
		p->first_page = page;
	}
}

314
static int destroy_compound_page(struct page *page, unsigned long order)
Linus Torvalds's avatar
Linus Torvalds committed
315 316 317
{
	int i;
	int nr_pages = 1 << order;
318
	int bad = 0;
Linus Torvalds's avatar
Linus Torvalds committed
319

320 321
	if (unlikely(compound_order(page) != order) ||
	    unlikely(!PageHead(page))) {
322
		bad_page(page);
323 324
		bad++;
	}
Linus Torvalds's avatar
Linus Torvalds committed
325

326
	__ClearPageHead(page);
327

328 329
	for (i = 1; i < nr_pages; i++) {
		struct page *p = page + i;
Linus Torvalds's avatar
Linus Torvalds committed
330

331
		if (unlikely(!PageTail(p) || (p->first_page != page))) {
332
			bad_page(page);
333 334
			bad++;
		}
335
		__ClearPageTail(p);
Linus Torvalds's avatar
Linus Torvalds committed
336
	}
337 338

	return bad;
Linus Torvalds's avatar
Linus Torvalds committed
339 340
}

Nick Piggin's avatar
Nick Piggin committed
341 342 343 344
static inline void prep_zero_page(struct page *page, int order, gfp_t gfp_flags)
{
	int i;

345 346 347 348
	/*
	 * clear_highpage() will use KM_USER0, so it's a bug to use __GFP_ZERO
	 * and __GFP_HIGHMEM from hard or soft interrupt context.
	 */
Nick Piggin's avatar
Nick Piggin committed
349
	VM_BUG_ON((gfp_flags & __GFP_HIGHMEM) && in_interrupt());
Nick Piggin's avatar
Nick Piggin committed
350 351 352 353
	for (i = 0; i < (1 << order); i++)
		clear_highpage(page + i);
}

354 355
static inline void set_page_order(struct page *page, int order)
{
356
	set_page_private(page, order);
357
	__SetPageBuddy(page);
Linus Torvalds's avatar
Linus Torvalds committed
358 359 360 361
}

static inline void rmv_page_order(struct page *page)
{
362
	__ClearPageBuddy(page);
363
	set_page_private(page, 0);
Linus Torvalds's avatar
Linus Torvalds committed
364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380
}

/*
 * Locate the struct page for both the matching buddy in our
 * pair (buddy1) and the combined O(n+1) page they form (page).
 *
 * 1) Any buddy B1 will have an order O twin B2 which satisfies
 * the following equation:
 *     B2 = B1 ^ (1 << O)
 * For example, if the starting buddy (buddy2) is #8 its order
 * 1 buddy is #10:
 *     B2 = 8 ^ (1 << 1) = 8 ^ 2 = 10
 *
 * 2) Any buddy B will have an order O+1 parent P which
 * satisfies the following equation:
 *     P = B & ~(1 << O)
 *
Andreas Mohr's avatar
Andreas Mohr committed
381
 * Assumption: *_mem_map is contiguous at least up to MAX_ORDER
Linus Torvalds's avatar
Linus Torvalds committed
382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399
 */
static inline struct page *
__page_find_buddy(struct page *page, unsigned long page_idx, unsigned int order)
{
	unsigned long buddy_idx = page_idx ^ (1 << order);

	return page + (buddy_idx - page_idx);
}

static inline unsigned long
__find_combined_index(unsigned long page_idx, unsigned int order)
{
	return (page_idx & ~(1 << order));
}

/*
 * 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
400
 * (a) the buddy is not in a hole &&
401
 * (b) the buddy is in the buddy system &&
402 403
 * (c) a page and its buddy have the same order &&
 * (d) a page and its buddy are in the same zone.
404 405 406
 *
 * For recording whether a page is in the buddy system, we use PG_buddy.
 * Setting, clearing, and testing PG_buddy is serialized by zone->lock.
Linus Torvalds's avatar
Linus Torvalds committed
407
 *
408
 * For recording page's order, we use page_private(page).
Linus Torvalds's avatar
Linus Torvalds committed
409
 */
410 411
static inline int page_is_buddy(struct page *page, struct page *buddy,
								int order)
Linus Torvalds's avatar
Linus Torvalds committed
412
{
413
	if (!pfn_valid_within(page_to_pfn(buddy)))
Nick Piggin's avatar
Nick Piggin committed
414 415
		return 0;

416 417 418 419
	if (page_zone_id(page) != page_zone_id(buddy))
		return 0;

	if (PageBuddy(buddy) && page_order(buddy) == order) {
420
		VM_BUG_ON(page_count(buddy) != 0);
421
		return 1;
422
	}
423
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
424 425 426 427 428 429 430 431 432 433 434 435 436 437 438
}

/*
 * 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
439
 * free pages of length of (1 << order) and marked with PG_buddy. Page's
440
 * order is recorded in page_private(page) field.
Linus Torvalds's avatar
Linus Torvalds committed
441 442 443 444 445 446 447 448 449
 * So when we are allocating or freeing one, we can derive the state of the
 * other.  That is, if we allocate a small block, and both were   
 * free, the remainder of the region must be split into blocks.   
 * If a block is freed, and its buddy is also free, then this
 * triggers coalescing into a block of larger size.            
 *
 * -- wli
 */

Nick Piggin's avatar
Nick Piggin committed
450
static inline void __free_one_page(struct page *page,
451 452
		struct zone *zone, unsigned int order,
		int migratetype)
Linus Torvalds's avatar
Linus Torvalds committed
453 454 455
{
	unsigned long page_idx;

456
	if (unlikely(PageCompound(page)))
457 458
		if (unlikely(destroy_compound_page(page, order)))
			return;
Linus Torvalds's avatar
Linus Torvalds committed
459

460 461
	VM_BUG_ON(migratetype == -1);

Linus Torvalds's avatar
Linus Torvalds committed
462 463
	page_idx = page_to_pfn(page) & ((1 << MAX_ORDER) - 1);

464
	VM_BUG_ON(page_idx & ((1 << order) - 1));
Nick Piggin's avatar
Nick Piggin committed
465
	VM_BUG_ON(bad_range(zone, page));
Linus Torvalds's avatar
Linus Torvalds committed
466 467 468 469 470 471

	while (order < MAX_ORDER-1) {
		unsigned long combined_idx;
		struct page *buddy;

		buddy = __page_find_buddy(page, page_idx, order);
472
		if (!page_is_buddy(page, buddy, order))
473
			break;
Nick Piggin's avatar
Nick Piggin committed
474

475
		/* Our buddy is free, merge with it and move up one order. */
Linus Torvalds's avatar
Linus Torvalds committed
476
		list_del(&buddy->lru);
477
		zone->free_area[order].nr_free--;
Linus Torvalds's avatar
Linus Torvalds committed
478
		rmv_page_order(buddy);
Nick Piggin's avatar
Nick Piggin committed
479
		combined_idx = __find_combined_index(page_idx, order);
Linus Torvalds's avatar
Linus Torvalds committed
480 481 482 483 484
		page = page + (combined_idx - page_idx);
		page_idx = combined_idx;
		order++;
	}
	set_page_order(page, order);
485 486
	list_add(&page->lru,
		&zone->free_area[order].free_list[migratetype]);
Linus Torvalds's avatar
Linus Torvalds committed
487 488 489
	zone->free_area[order].nr_free++;
}

490 491 492 493 494 495 496 497 498 499 500
/*
 * free_page_mlock() -- clean up attempts to free and mlocked() page.
 * Page should not be on lru, so no need to fix that up.
 * free_pages_check() will verify...
 */
static inline void free_page_mlock(struct page *page)
{
	__dec_zone_page_state(page, NR_MLOCK);
	__count_vm_event(UNEVICTABLE_MLOCKFREED);
}

501
static inline int free_pages_check(struct page *page)
Linus Torvalds's avatar
Linus Torvalds committed
502
{
Nick Piggin's avatar
Nick Piggin committed
503 504
	if (unlikely(page_mapcount(page) |
		(page->mapping != NULL)  |
505
		(atomic_read(&page->_count) != 0) |
506
		(page->flags & PAGE_FLAGS_CHECK_AT_FREE))) {
507
		bad_page(page);
508
		return 1;
509
	}
510 511 512
	if (page->flags & PAGE_FLAGS_CHECK_AT_PREP)
		page->flags &= ~PAGE_FLAGS_CHECK_AT_PREP;
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
513 514 515
}

/*
516
 * Frees a number of pages from the PCP lists
Linus Torvalds's avatar
Linus Torvalds committed
517
 * Assumes all pages on list are in same zone, and of same order.
518
 * count is the number of pages to free.
Linus Torvalds's avatar
Linus Torvalds committed
519 520 521 522 523 524 525
 *
 * 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.
 */
526 527
static void free_pcppages_bulk(struct zone *zone, int count,
					struct per_cpu_pages *pcp)
Linus Torvalds's avatar
Linus Torvalds committed
528
{
529
	int migratetype = 0;
530
	int batch_free = 0;
531

Nick Piggin's avatar
Nick Piggin committed
532
	spin_lock(&zone->lock);
533
	zone->all_unreclaimable = 0;
Linus Torvalds's avatar
Linus Torvalds committed
534
	zone->pages_scanned = 0;
535

536
	__mod_zone_page_state(zone, NR_FREE_PAGES, count);
537
	while (count) {
Nick Piggin's avatar
Nick Piggin committed
538
		struct page *page;
539 540 541
		struct list_head *list;

		/*
542 543 544 545 546
		 * 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
547 548
		 */
		do {
549
			batch_free++;
550 551 552 553
			if (++migratetype == MIGRATE_PCPTYPES)
				migratetype = 0;
			list = &pcp->lists[migratetype];
		} while (list_empty(list));
Nick Piggin's avatar
Nick Piggin committed
554

555 556 557 558
		do {
			page = list_entry(list->prev, struct page, lru);
			/* must delete as __free_one_page list manipulates */
			list_del(&page->lru);
559 560 561
			/* MIGRATE_MOVABLE list may include MIGRATE_RESERVEs */
			__free_one_page(page, zone, 0, page_private(page));
			trace_mm_page_pcpu_drain(page, 0, page_private(page));
562
		} while (--count && --batch_free && !list_empty(list));
Linus Torvalds's avatar
Linus Torvalds committed
563
	}
Nick Piggin's avatar
Nick Piggin committed
564
	spin_unlock(&zone->lock);
Linus Torvalds's avatar
Linus Torvalds committed
565 566
}

567 568
static void free_one_page(struct zone *zone, struct page *page, int order,
				int migratetype)
Linus Torvalds's avatar
Linus Torvalds committed
569
{
570
	spin_lock(&zone->lock);
571
	zone->all_unreclaimable = 0;
572
	zone->pages_scanned = 0;
573 574

	__mod_zone_page_state(zone, NR_FREE_PAGES, 1 << order);
575
	__free_one_page(page, zone, order, migratetype);
576
	spin_unlock(&zone->lock);
Nick Piggin's avatar
Nick Piggin committed
577 578 579 580 581
}

static void __free_pages_ok(struct page *page, unsigned int order)
{
	unsigned long flags;
Linus Torvalds's avatar
Linus Torvalds committed
582
	int i;
583
	int bad = 0;
584
	int wasMlocked = __TestClearPageMlocked(page);
Linus Torvalds's avatar
Linus Torvalds committed
585

586
	trace_mm_page_free_direct(page, order);
587 588
	kmemcheck_free_shadow(page, order);

Linus Torvalds's avatar
Linus Torvalds committed
589
	for (i = 0 ; i < (1 << order) ; ++i)
590 591
		bad += free_pages_check(page + i);
	if (bad)
592 593
		return;

594
	if (!PageHighMem(page)) {
Nick Piggin's avatar
Nick Piggin committed
595
		debug_check_no_locks_freed(page_address(page),PAGE_SIZE<<order);
596 597 598
		debug_check_no_obj_freed(page_address(page),
					   PAGE_SIZE << order);
	}
Nick Piggin's avatar
Nick Piggin committed
599
	arch_free_page(page, order);
Nick Piggin's avatar
Nick Piggin committed
600
	kernel_map_pages(page, 1 << order, 0);
Nick Piggin's avatar
Nick Piggin committed
601

Nick Piggin's avatar
Nick Piggin committed
602
	local_irq_save(flags);
603
	if (unlikely(wasMlocked))
604
		free_page_mlock(page);
605
	__count_vm_events(PGFREE, 1 << order);
606 607
	free_one_page(page_zone(page), page, order,
					get_pageblock_migratetype(page));
Nick Piggin's avatar
Nick Piggin committed
608
	local_irq_restore(flags);
Linus Torvalds's avatar
Linus Torvalds committed
609 610
}

611 612 613
/*
 * permit the bootmem allocator to evade page validation on high-order frees
 */
614
void __meminit __free_pages_bootmem(struct page *page, unsigned int order)
615 616 617 618
{
	if (order == 0) {
		__ClearPageReserved(page);
		set_page_count(page, 0);
619
		set_page_refcounted(page);
Nick Piggin's avatar
Nick Piggin committed
620
		__free_page(page);
621 622 623
	} else {
		int loop;

Nick Piggin's avatar
Nick Piggin committed
624
		prefetchw(page);
625 626 627
		for (loop = 0; loop < BITS_PER_LONG; loop++) {
			struct page *p = &page[loop];

Nick Piggin's avatar
Nick Piggin committed
628 629
			if (loop + 1 < BITS_PER_LONG)
				prefetchw(p + 1);
630 631 632 633
			__ClearPageReserved(p);
			set_page_count(p, 0);
		}

634
		set_page_refcounted(page);
Nick Piggin's avatar
Nick Piggin committed
635
		__free_pages(page, order);
636 637 638
	}
}

Linus Torvalds's avatar
Linus Torvalds committed
639 640 641 642 643 644 645 646 647 648 649 650 651 652 653

/*
 * The order of subdivision here is critical for the IO subsystem.
 * Please do not alter this order without good reasons and regression
 * testing. Specifically, as large blocks of memory are subdivided,
 * the order in which smaller blocks are delivered depends on the order
 * they're subdivided in this function. This is the primary factor
 * influencing the order in which pages are delivered to the IO
 * subsystem according to empirical testing, and this is also justified
 * by considering the behavior of a buddy system containing a single
 * large block of memory acted on by a series of small allocations.
 * This behavior is a critical factor in sglist merging's success.
 *
 * -- wli
 */
Nick Piggin's avatar
Nick Piggin committed
654
static inline void expand(struct zone *zone, struct page *page,
655 656
	int low, int high, struct free_area *area,
	int migratetype)
Linus Torvalds's avatar
Linus Torvalds committed
657 658 659 660 661 662 663
{
	unsigned long size = 1 << high;

	while (high > low) {
		area--;
		high--;
		size >>= 1;
Nick Piggin's avatar
Nick Piggin committed
664
		VM_BUG_ON(bad_range(zone, &page[size]));
665
		list_add(&page[size].lru, &area->free_list[migratetype]);
Linus Torvalds's avatar
Linus Torvalds committed
666 667 668 669 670 671 672 673
		area->nr_free++;
		set_page_order(&page[size], high);
	}
}

/*
 * This page is about to be returned from the page allocator
 */
674
static inline int check_new_page(struct page *page)
Linus Torvalds's avatar
Linus Torvalds committed
675
{
Nick Piggin's avatar
Nick Piggin committed
676 677
	if (unlikely(page_mapcount(page) |
		(page->mapping != NULL)  |
678
		(atomic_read(&page->_count) != 0)  |
679
		(page->flags & PAGE_FLAGS_CHECK_AT_PREP))) {
680
		bad_page(page);
681
		return 1;
682
	}
683 684 685 686 687 688 689 690 691 692 693 694
	return 0;
}

static int prep_new_page(struct page *page, int order, gfp_t gfp_flags)
{
	int i;

	for (i = 0; i < (1 << order); i++) {
		struct page *p = page + i;
		if (unlikely(check_new_page(p)))
			return 1;
	}
695

696
	set_page_private(page, 0);
697
	set_page_refcounted(page);
Nick Piggin's avatar
Nick Piggin committed
698 699

	arch_alloc_page(page, order);
Linus Torvalds's avatar
Linus Torvalds committed
700
	kernel_map_pages(page, 1 << order, 1);
Nick Piggin's avatar
Nick Piggin committed
701 702 703 704 705 706 707

	if (gfp_flags & __GFP_ZERO)
		prep_zero_page(page, order, gfp_flags);

	if (order && (gfp_flags & __GFP_COMP))
		prep_compound_page(page, order);

708
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
709 710
}

711 712 713 714
/*
 * Go through the free lists for the given migratetype and remove
 * the smallest available page from the freelists
 */
715 716
static inline
struct page *__rmqueue_smallest(struct zone *zone, unsigned int order,
717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741
						int migratetype)
{
	unsigned int current_order;
	struct free_area * area;
	struct page *page;

	/* Find a page of the appropriate size in the preferred list */
	for (current_order = order; current_order < MAX_ORDER; ++current_order) {
		area = &(zone->free_area[current_order]);
		if (list_empty(&area->free_list[migratetype]))
			continue;

		page = list_entry(area->free_list[migratetype].next,
							struct page, lru);
		list_del(&page->lru);
		rmv_page_order(page);
		area->nr_free--;
		expand(zone, page, order, current_order, area, migratetype);
		return page;
	}

	return NULL;
}


742 743 744 745 746
/*
 * This array describes the order lists are fallen back to when
 * the free lists for the desirable migrate type are depleted
 */
static int fallbacks[MIGRATE_TYPES][MIGRATE_TYPES-1] = {
747 748 749 750
	[MIGRATE_UNMOVABLE]   = { MIGRATE_RECLAIMABLE, MIGRATE_MOVABLE,   MIGRATE_RESERVE },
	[MIGRATE_RECLAIMABLE] = { MIGRATE_UNMOVABLE,   MIGRATE_MOVABLE,   MIGRATE_RESERVE },
	[MIGRATE_MOVABLE]     = { MIGRATE_RECLAIMABLE, MIGRATE_UNMOVABLE, MIGRATE_RESERVE },
	[MIGRATE_RESERVE]     = { MIGRATE_RESERVE,     MIGRATE_RESERVE,   MIGRATE_RESERVE }, /* Never used */
751 752
};

753 754
/*
 * Move the free pages in a range to the free lists of the requested type.
755
 * Note that start_page and end_pages are not aligned on a pageblock
756 757
 * boundary. If alignment is required, use move_freepages_block()
 */
Adrian Bunk's avatar
Adrian Bunk committed
758 759 760
static int move_freepages(struct zone *zone,
			  struct page *start_page, struct page *end_page,
			  int migratetype)
761 762 763
{
	struct page *page;
	unsigned long order;
764
	int pages_moved = 0;
765 766 767 768 769 770 771

#ifndef CONFIG_HOLES_IN_ZONE
	/*
	 * page_zone is not safe to call in this context when
	 * CONFIG_HOLES_IN_ZONE is set. This bug check is probably redundant
	 * anyway as we check zone boundaries in move_freepages_block().
	 * Remove at a later date when no bug reports exist related to
Mel Gorman's avatar
Mel Gorman committed
772
	 * grouping pages by mobility
773 774 775 776 777
	 */
	BUG_ON(page_zone(start_page) != page_zone(end_page));
#endif

	for (page = start_page; page <= end_page;) {
778 779 780
		/* Make sure we are not inadvertently changing nodes */
		VM_BUG_ON(page_to_nid(page) != zone_to_nid(zone));

781 782 783 784 785 786 787 788 789 790 791 792 793 794 795
		if (!pfn_valid_within(page_to_pfn(page))) {
			page++;
			continue;
		}

		if (!PageBuddy(page)) {
			page++;
			continue;
		}

		order = page_order(page);
		list_del(&page->lru);
		list_add(&page->lru,
			&zone->free_area[order].free_list[migratetype]);
		page += 1 << order;
796
		pages_moved += 1 << order;
797 798
	}

799
	return pages_moved;
800 801
}

Adrian Bunk's avatar
Adrian Bunk committed
802 803
static int move_freepages_block(struct zone *zone, struct page *page,
				int migratetype)
804 805 806 807 808
{
	unsigned long start_pfn, end_pfn;
	struct page *start_page, *end_page;

	start_pfn = page_to_pfn(page);
809
	start_pfn = start_pfn & ~(pageblock_nr_pages-1);
810
	start_page = pfn_to_page(start_pfn);
811 812
	end_page = start_page + pageblock_nr_pages - 1;
	end_pfn = start_pfn + pageblock_nr_pages - 1;
813 814 815 816 817 818 819 820 821 822

	/* Do not cross zone boundaries */
	if (start_pfn < zone->zone_start_pfn)
		start_page = page;
	if (end_pfn >= zone->zone_start_pfn + zone->spanned_pages)
		return 0;

	return move_freepages(zone, start_page, end_page, migratetype);
}

823 824 825 826 827 828 829 830 831 832 833
static void change_pageblock_range(struct page *pageblock_page,
					int start_order, int migratetype)
{
	int nr_pageblocks = 1 << (start_order - pageblock_order);

	while (nr_pageblocks--) {
		set_pageblock_migratetype(pageblock_page, migratetype);
		pageblock_page += pageblock_nr_pages;
	}
}

834
/* Remove an element from the buddy allocator from the fallback list */
835 836
static inline struct page *
__rmqueue_fallback(struct zone *zone, int order, int start_migratetype)
837 838 839 840 841 842 843 844 845 846 847 848
{
	struct free_area * area;
	int current_order;
	struct page *page;
	int migratetype, i;

	/* Find the largest possible block of pages in the other list */
	for (current_order = MAX_ORDER-1; current_order >= order;
						--current_order) {
		for (i = 0; i < MIGRATE_TYPES - 1; i++) {
			migratetype = fallbacks[start_migratetype][i];

849 850 851
			/* MIGRATE_RESERVE handled later if necessary */
			if (migratetype == MIGRATE_RESERVE)
				continue;
852

853 854 855 856 857 858 859 860 861
			area = &(zone->free_area[current_order]);
			if (list_empty(&area->free_list[migratetype]))
				continue;

			page = list_entry(area->free_list[migratetype].next,
					struct page, lru);
			area->nr_free--;

			/*
862
			 * If breaking a large block of pages, move all free
863 864 865
			 * pages to the preferred allocation list. If falling
			 * back for a reclaimable kernel allocation, be more
			 * agressive about taking ownership of free pages
866
			 */
867
			if (unlikely(current_order >= (pageblock_order >> 1)) ||
868 869
					start_migratetype == MIGRATE_RECLAIMABLE ||
					page_group_by_mobility_disabled) {
870 871 872 873 874
				unsigned long pages;
				pages = move_freepages_block(zone, page,
								start_migratetype);

				/* Claim the whole block if over half of it is free */
875 876
				if (pages >= (1 << (pageblock_order-1)) ||
						page_group_by_mobility_disabled)
877 878 879
					set_pageblock_migratetype(page,
								start_migratetype);

880
				migratetype = start_migratetype;
881
			}
882 883 884 885 886

			/* Remove the page from the freelists */
			list_del(&page->lru);
			rmv_page_order(page);

887 888 889
			/* Take ownership for orders >= pageblock_order */
			if (current_order >= pageblock_order)
				change_pageblock_range(page, current_order,
890 891 892
							start_migratetype);

			expand(zone, page, order, current_order, area, migratetype);
893 894 895 896

			trace_mm_page_alloc_extfrag(page, order, current_order,
				start_migratetype, migratetype);

897 898 899 900
			return page;
		}
	}

901
	return NULL;
902 903
}

904
/*
Linus Torvalds's avatar
Linus Torvalds committed
905 906 907
 * Do the hard work of removing an element from the buddy allocator.
 * Call me with the zone->lock already held.
 */
908 909
static struct page *__rmqueue(struct zone *zone, unsigned int order,
						int migratetype)
Linus Torvalds's avatar
Linus Torvalds committed
910 911 912
{
	struct page *page;

913
retry_reserve:
914
	page = __rmqueue_smallest(zone, order, migratetype);
915

916
	if (unlikely(!page) && migratetype != MIGRATE_RESERVE) {
917
		page = __rmqueue_fallback(zone, order, migratetype);
918

919 920 921 922 923 924 925 926 927 928 929
		/*
		 * Use MIGRATE_RESERVE rather than fail an allocation. goto
		 * is used because __rmqueue_smallest is an inline function
		 * and we want just one call site
		 */
		if (!page) {
			migratetype = MIGRATE_RESERVE;
			goto retry_reserve;
		}
	}

930
	trace_mm_page_alloc_zone_locked(page, order, migratetype);
931
	return page;
Linus Torvalds's avatar
Linus Torvalds committed
932 933 934 935 936 937 938 939
}

/* 
 * Obtain a specified number of elements from the buddy allocator, all under
 * a single hold of the lock, for efficiency.  Add them to the supplied list.
 * Returns the number of new pages which were placed at *list.
 */
static int rmqueue_bulk(struct zone *zone, unsigned int order, 
940
			unsigned long count, struct list_head *list,
941
			int migratetype, int cold)
Linus Torvalds's avatar
Linus Torvalds committed
942 943 944
{
	int i;
	
Nick Piggin's avatar
Nick Piggin committed
945
	spin_lock(&zone->lock);
Linus Torvalds's avatar
Linus Torvalds committed
946
	for (i = 0; i < count; ++i) {
947
		struct page *page = __rmqueue(zone, order, migratetype);
Nick Piggin's avatar
Nick Piggin committed
948
		if (unlikely(page == NULL))
Linus Torvalds's avatar
Linus Torvalds committed
949
			break;
950 951 952 953 954 955 956 957 958 959

		/*
		 * Split buddy pages returned by expand() are received here
		 * in physical page order. The page is added to the callers and
		 * list and the list head then moves forward. From the callers
		 * perspective, the linked list is ordered by page number in
		 * some conditions. This is useful for IO devices that can
		 * merge IO requests if the physical pages are ordered
		 * properly.
		 */
960 961 962 963
		if (likely(cold == 0))
			list_add(&page->lru, list);
		else
			list_add_tail(&page->lru, list);
964
		set_page_private(page, migratetype);
965
		list = &page->lru;
Linus Torvalds's avatar
Linus Torvalds committed
966
	}
967
	__mod_zone_page_state(zone, NR_FREE_PAGES, -(i << order));
Nick Piggin's avatar
Nick Piggin committed
968
	spin_unlock(&zone->lock);
Nick Piggin's avatar
Nick Piggin committed
969
	return i;
Linus Torvalds's avatar
Linus Torvalds committed
970 971
}