page_alloc.c 129 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>
Linus Torvalds's avatar
Linus Torvalds committed
26 27 28 29 30
#include <linux/module.h>
#include <linux/suspend.h>
#include <linux/pagevec.h>
#include <linux/blkdev.h>
#include <linux/slab.h>
31
#include <linux/oom.h>
Linus Torvalds's avatar
Linus Torvalds committed
32 33 34 35 36
#include <linux/notifier.h>
#include <linux/topology.h>
#include <linux/sysctl.h>
#include <linux/cpu.h>
#include <linux/cpuset.h>
37
#include <linux/memory_hotplug.h>
Linus Torvalds's avatar
Linus Torvalds committed
38 39
#include <linux/nodemask.h>
#include <linux/vmalloc.h>
40
#include <linux/mempolicy.h>
41
#include <linux/stop_machine.h>
42 43
#include <linux/sort.h>
#include <linux/pfn.h>
44
#include <linux/backing-dev.h>
45
#include <linux/fault-inject.h>
46
#include <linux/page-isolation.h>
47
#include <linux/page_cgroup.h>
48
#include <linux/debugobjects.h>
Linus Torvalds's avatar
Linus Torvalds committed
49 50

#include <asm/tlbflush.h>
51
#include <asm/div64.h>
Linus Torvalds's avatar
Linus Torvalds committed
52 53 54
#include "internal.h"

/*
55
 * Array of node states.
Linus Torvalds's avatar
Linus Torvalds committed
56
 */
57 58 59 60 61 62 63 64 65 66 67 68 69
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);

70
unsigned long totalram_pages __read_mostly;
71
unsigned long totalreserve_pages __read_mostly;
72
unsigned long highest_memmap_pfn __read_mostly;
73
int percpu_pagelist_fraction;
Linus Torvalds's avatar
Linus Torvalds committed
74

75 76 77 78
#ifdef CONFIG_HUGETLB_PAGE_SIZE_VARIABLE
int pageblock_order __read_mostly;
#endif

79
static void __free_pages_ok(struct page *page, unsigned int order);
80

Linus Torvalds's avatar
Linus Torvalds committed
81 82 83 84 85 86 87
/*
 * 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
88 89 90
 *
 * 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
91
 */
92
int sysctl_lowmem_reserve_ratio[MAX_NR_ZONES-1] = {
93
#ifdef CONFIG_ZONE_DMA
94
	 256,
95
#endif
96
#ifdef CONFIG_ZONE_DMA32
97
	 256,
98
#endif
99
#ifdef CONFIG_HIGHMEM
Mel Gorman's avatar
Mel Gorman committed
100
	 32,
101
#endif
Mel Gorman's avatar
Mel Gorman committed
102
	 32,
103
};
Linus Torvalds's avatar
Linus Torvalds committed
104 105 106

EXPORT_SYMBOL(totalram_pages);

107
static char * const zone_names[MAX_NR_ZONES] = {
108
#ifdef CONFIG_ZONE_DMA
109
	 "DMA",
110
#endif
111
#ifdef CONFIG_ZONE_DMA32
112
	 "DMA32",
113
#endif
114
	 "Normal",
115
#ifdef CONFIG_HIGHMEM
Mel Gorman's avatar
Mel Gorman committed
116
	 "HighMem",
117
#endif
Mel Gorman's avatar
Mel Gorman committed
118
	 "Movable",
119 120
};

Linus Torvalds's avatar
Linus Torvalds committed
121 122
int min_free_kbytes = 1024;

123 124
unsigned long __meminitdata nr_kernel_pages;
unsigned long __meminitdata nr_all_pages;
125
static unsigned long __meminitdata dma_reserve;
Linus Torvalds's avatar
Linus Torvalds committed
126

127 128
#ifdef CONFIG_ARCH_POPULATES_NODE_MAP
  /*
Simon Arlott's avatar
Simon Arlott committed
129
   * MAX_ACTIVE_REGIONS determines the maximum number of distinct
130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147
   * 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

148 149 150 151
  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
152
  static unsigned long __initdata required_kernelcore;
153
  static unsigned long __initdata required_movablecore;
Adrian Bunk's avatar
Adrian Bunk committed
154
  static unsigned long __meminitdata zone_movable_pfn[MAX_NUMNODES];
Mel Gorman's avatar
Mel Gorman committed
155 156 157 158

  /* movable_zone is the "real" zone pages in ZONE_MOVABLE are taken from */
  int movable_zone;
  EXPORT_SYMBOL(movable_zone);
159 160
#endif /* CONFIG_ARCH_POPULATES_NODE_MAP */

Miklos Szeredi's avatar
Miklos Szeredi committed
161 162 163 164 165
#if MAX_NUMNODES > 1
int nr_node_ids __read_mostly = MAX_NUMNODES;
EXPORT_SYMBOL(nr_node_ids);
#endif

166 167
int page_group_by_mobility_disabled __read_mostly;

168 169 170 171 172 173
static void set_pageblock_migratetype(struct page *page, int migratetype)
{
	set_pageblock_flags_group(page, (unsigned long)migratetype,
					PB_migrate, PB_migrate_end);
}

Nick Piggin's avatar
Nick Piggin committed
174
#ifdef CONFIG_DEBUG_VM
175
static int page_outside_zone_boundaries(struct zone *zone, struct page *page)
Linus Torvalds's avatar
Linus Torvalds committed
176
{
177 178 179
	int ret = 0;
	unsigned seq;
	unsigned long pfn = page_to_pfn(page);
180

181 182 183 184 185 186 187 188 189
	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;
190 191 192 193
}

static int page_is_consistent(struct zone *zone, struct page *page)
{
194
	if (!pfn_valid_within(page_to_pfn(page)))
195
		return 0;
Linus Torvalds's avatar
Linus Torvalds committed
196
	if (zone != page_zone(page))
197 198 199 200 201 202 203 204 205 206
		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
207
		return 1;
208 209 210
	if (!page_is_consistent(zone, page))
		return 1;

Linus Torvalds's avatar
Linus Torvalds committed
211 212
	return 0;
}
Nick Piggin's avatar
Nick Piggin committed
213 214 215 216 217 218 219
#else
static inline int bad_range(struct zone *zone, struct page *page)
{
	return 0;
}
#endif

220
static void bad_page(struct page *page)
Linus Torvalds's avatar
Linus Torvalds committed
221
{
222 223 224 225 226 227 228 229 230 231 232 233 234 235
	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) {
236 237
			printk(KERN_ALERT
			      "BUG: Bad page state: %lu messages suppressed\n",
238 239 240 241 242 243 244 245
				nr_unshown);
			nr_unshown = 0;
		}
		nr_shown = 0;
	}
	if (nr_shown++ == 0)
		resume = jiffies + 60 * HZ;

246
	printk(KERN_ALERT "BUG: Bad page state in process %s  pfn:%05lx\n",
247
		current->comm, page_to_pfn(page));
248
	printk(KERN_ALERT
249 250 251 252
		"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
253
	dump_stack();
254
out:
255 256
	/* Leave bad fields for debug, except PageBuddy could make trouble */
	__ClearPageBuddy(page);
257
	add_taint(TAINT_BAD_PAGE);
Linus Torvalds's avatar
Linus Torvalds committed
258 259 260 261 262 263 264 265 266 267 268 269
}

/*
 * 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).
 *
270 271 272
 * 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
273
 */
274 275 276

static void free_compound_page(struct page *page)
{
277
	__free_pages_ok(page, compound_order(page));
278 279
}

280
void prep_compound_page(struct page *page, unsigned long order)
281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297
{
	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;
	}
}

#ifdef CONFIG_HUGETLBFS
void prep_compound_gigantic_page(struct page *page, unsigned long order)
Linus Torvalds's avatar
Linus Torvalds committed
298 299 300
{
	int i;
	int nr_pages = 1 << order;
301
	struct page *p = page + 1;
Linus Torvalds's avatar
Linus Torvalds committed
302

303
	set_compound_page_dtor(page, free_compound_page);
304
	set_compound_order(page, order);
305
	__SetPageHead(page);
306
	for (i = 1; i < nr_pages; i++, p = mem_map_next(p, page, i)) {
307 308
		__SetPageTail(p);
		p->first_page = page;
Linus Torvalds's avatar
Linus Torvalds committed
309 310
	}
}
311
#endif
Linus Torvalds's avatar
Linus Torvalds committed
312

313
static int destroy_compound_page(struct page *page, unsigned long order)
Linus Torvalds's avatar
Linus Torvalds committed
314 315 316
{
	int i;
	int nr_pages = 1 << order;
317
	int bad = 0;
Linus Torvalds's avatar
Linus Torvalds committed
318

319 320
	if (unlikely(compound_order(page) != order) ||
	    unlikely(!PageHead(page))) {
321
		bad_page(page);
322 323
		bad++;
	}
Linus Torvalds's avatar
Linus Torvalds committed
324

325
	__ClearPageHead(page);
326

327 328
	for (i = 1; i < nr_pages; i++) {
		struct page *p = page + i;
Linus Torvalds's avatar
Linus Torvalds committed
329

330
		if (unlikely(!PageTail(p) || (p->first_page != page))) {
331
			bad_page(page);
332 333
			bad++;
		}
334
		__ClearPageTail(p);
Linus Torvalds's avatar
Linus Torvalds committed
335
	}
336 337

	return bad;
Linus Torvalds's avatar
Linus Torvalds committed
338 339
}

Nick Piggin's avatar
Nick Piggin committed
340 341 342 343
static inline void prep_zero_page(struct page *page, int order, gfp_t gfp_flags)
{
	int i;

344 345 346 347
	/*
	 * 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
348
	VM_BUG_ON((gfp_flags & __GFP_HIGHMEM) && in_interrupt());
Nick Piggin's avatar
Nick Piggin committed
349 350 351 352
	for (i = 0; i < (1 << order); i++)
		clear_highpage(page + i);
}

353 354
static inline void set_page_order(struct page *page, int order)
{
355
	set_page_private(page, order);
356
	__SetPageBuddy(page);
Linus Torvalds's avatar
Linus Torvalds committed
357 358 359 360
}

static inline void rmv_page_order(struct page *page)
{
361
	__ClearPageBuddy(page);
362
	set_page_private(page, 0);
Linus Torvalds's avatar
Linus Torvalds committed
363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379
}

/*
 * 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
380
 * Assumption: *_mem_map is contiguous at least up to MAX_ORDER
Linus Torvalds's avatar
Linus Torvalds committed
381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398
 */
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
399
 * (a) the buddy is not in a hole &&
400
 * (b) the buddy is in the buddy system &&
401 402
 * (c) a page and its buddy have the same order &&
 * (d) a page and its buddy are in the same zone.
403 404 405
 *
 * 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
406
 *
407
 * For recording page's order, we use page_private(page).
Linus Torvalds's avatar
Linus Torvalds committed
408
 */
409 410
static inline int page_is_buddy(struct page *page, struct page *buddy,
								int order)
Linus Torvalds's avatar
Linus Torvalds committed
411
{
412
	if (!pfn_valid_within(page_to_pfn(buddy)))
Nick Piggin's avatar
Nick Piggin committed
413 414
		return 0;

415 416 417 418 419
	if (page_zone_id(page) != page_zone_id(buddy))
		return 0;

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

/*
 * 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
438
 * free pages of length of (1 << order) and marked with PG_buddy. Page's
439
 * order is recorded in page_private(page) field.
Linus Torvalds's avatar
Linus Torvalds committed
440 441 442 443 444 445 446 447 448
 * 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
449
static inline void __free_one_page(struct page *page,
Linus Torvalds's avatar
Linus Torvalds committed
450 451 452 453
		struct zone *zone, unsigned int order)
{
	unsigned long page_idx;
	int order_size = 1 << order;
454
	int migratetype = get_pageblock_migratetype(page);
Linus Torvalds's avatar
Linus Torvalds committed
455

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

	page_idx = page_to_pfn(page) & ((1 << MAX_ORDER) - 1);

Nick Piggin's avatar
Nick Piggin committed
462 463
	VM_BUG_ON(page_idx & (order_size - 1));
	VM_BUG_ON(bad_range(zone, page));
Linus Torvalds's avatar
Linus Torvalds committed
464

465
	__mod_zone_page_state(zone, NR_FREE_PAGES, order_size);
Linus Torvalds's avatar
Linus Torvalds committed
466 467 468 469 470
	while (order < MAX_ORDER-1) {
		unsigned long combined_idx;
		struct page *buddy;

		buddy = __page_find_buddy(page, page_idx, order);
471
		if (!page_is_buddy(page, buddy, order))
472
			break;
Nick Piggin's avatar
Nick Piggin committed
473

474
		/* Our buddy is free, merge with it and move up one order. */
Linus Torvalds's avatar
Linus Torvalds committed
475
		list_del(&buddy->lru);
476
		zone->free_area[order].nr_free--;
Linus Torvalds's avatar
Linus Torvalds committed
477
		rmv_page_order(buddy);
Nick Piggin's avatar
Nick Piggin committed
478
		combined_idx = __find_combined_index(page_idx, order);
Linus Torvalds's avatar
Linus Torvalds committed
479 480 481 482 483
		page = page + (combined_idx - page_idx);
		page_idx = combined_idx;
		order++;
	}
	set_page_order(page, order);
484 485
	list_add(&page->lru,
		&zone->free_area[order].free_list[migratetype]);
Linus Torvalds's avatar
Linus Torvalds committed
486 487 488
	zone->free_area[order].nr_free++;
}

489
static inline int free_pages_check(struct page *page)
Linus Torvalds's avatar
Linus Torvalds committed
490
{
491
	free_page_mlock(page);
Nick Piggin's avatar
Nick Piggin committed
492 493 494
	if (unlikely(page_mapcount(page) |
		(page->mapping != NULL)  |
		(page_count(page) != 0)  |
495
		(page->flags & PAGE_FLAGS_CHECK_AT_FREE))) {
496
		bad_page(page);
497
		return 1;
498
	}
499 500 501
	if (page->flags & PAGE_FLAGS_CHECK_AT_PREP)
		page->flags &= ~PAGE_FLAGS_CHECK_AT_PREP;
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
502 503 504 505 506
}

/*
 * Frees a list of pages. 
 * Assumes all pages on list are in same zone, and of same order.
507
 * count is the number of pages to free.
Linus Torvalds's avatar
Linus Torvalds committed
508 509 510 511 512 513 514
 *
 * 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.
 */
Nick Piggin's avatar
Nick Piggin committed
515 516
static void free_pages_bulk(struct zone *zone, int count,
					struct list_head *list, int order)
Linus Torvalds's avatar
Linus Torvalds committed
517
{
Nick Piggin's avatar
Nick Piggin committed
518
	spin_lock(&zone->lock);
519
	zone_clear_flag(zone, ZONE_ALL_UNRECLAIMABLE);
Linus Torvalds's avatar
Linus Torvalds committed
520
	zone->pages_scanned = 0;
Nick Piggin's avatar
Nick Piggin committed
521 522 523
	while (count--) {
		struct page *page;

Nick Piggin's avatar
Nick Piggin committed
524
		VM_BUG_ON(list_empty(list));
Linus Torvalds's avatar
Linus Torvalds committed
525
		page = list_entry(list->prev, struct page, lru);
Nick Piggin's avatar
Nick Piggin committed
526
		/* have to delete it as __free_one_page list manipulates */
Linus Torvalds's avatar
Linus Torvalds committed
527
		list_del(&page->lru);
Nick Piggin's avatar
Nick Piggin committed
528
		__free_one_page(page, zone, order);
Linus Torvalds's avatar
Linus Torvalds committed
529
	}
Nick Piggin's avatar
Nick Piggin committed
530
	spin_unlock(&zone->lock);
Linus Torvalds's avatar
Linus Torvalds committed
531 532
}

Nick Piggin's avatar
Nick Piggin committed
533
static void free_one_page(struct zone *zone, struct page *page, int order)
Linus Torvalds's avatar
Linus Torvalds committed
534
{
535
	spin_lock(&zone->lock);
536
	zone_clear_flag(zone, ZONE_ALL_UNRECLAIMABLE);
537
	zone->pages_scanned = 0;
538
	__free_one_page(page, zone, order);
539
	spin_unlock(&zone->lock);
Nick Piggin's avatar
Nick Piggin committed
540 541 542 543 544
}

static void __free_pages_ok(struct page *page, unsigned int order)
{
	unsigned long flags;
Linus Torvalds's avatar
Linus Torvalds committed
545
	int i;
546
	int bad = 0;
Linus Torvalds's avatar
Linus Torvalds committed
547 548

	for (i = 0 ; i < (1 << order) ; ++i)
549 550
		bad += free_pages_check(page + i);
	if (bad)
551 552
		return;

553
	if (!PageHighMem(page)) {
Nick Piggin's avatar
Nick Piggin committed
554
		debug_check_no_locks_freed(page_address(page),PAGE_SIZE<<order);
555 556 557
		debug_check_no_obj_freed(page_address(page),
					   PAGE_SIZE << order);
	}
Nick Piggin's avatar
Nick Piggin committed
558
	arch_free_page(page, order);
Nick Piggin's avatar
Nick Piggin committed
559
	kernel_map_pages(page, 1 << order, 0);
Nick Piggin's avatar
Nick Piggin committed
560

Nick Piggin's avatar
Nick Piggin committed
561
	local_irq_save(flags);
562
	__count_vm_events(PGFREE, 1 << order);
Nick Piggin's avatar
Nick Piggin committed
563
	free_one_page(page_zone(page), page, order);
Nick Piggin's avatar
Nick Piggin committed
564
	local_irq_restore(flags);
Linus Torvalds's avatar
Linus Torvalds committed
565 566
}

567 568 569
/*
 * permit the bootmem allocator to evade page validation on high-order frees
 */
570
void __meminit __free_pages_bootmem(struct page *page, unsigned int order)
571 572 573 574
{
	if (order == 0) {
		__ClearPageReserved(page);
		set_page_count(page, 0);
575
		set_page_refcounted(page);
Nick Piggin's avatar
Nick Piggin committed
576
		__free_page(page);
577 578 579
	} else {
		int loop;

Nick Piggin's avatar
Nick Piggin committed
580
		prefetchw(page);
581 582 583
		for (loop = 0; loop < BITS_PER_LONG; loop++) {
			struct page *p = &page[loop];

Nick Piggin's avatar
Nick Piggin committed
584 585
			if (loop + 1 < BITS_PER_LONG)
				prefetchw(p + 1);
586 587 588 589
			__ClearPageReserved(p);
			set_page_count(p, 0);
		}

590
		set_page_refcounted(page);
Nick Piggin's avatar
Nick Piggin committed
591
		__free_pages(page, order);
592 593 594
	}
}

Linus Torvalds's avatar
Linus Torvalds committed
595 596 597 598 599 600 601 602 603 604 605 606 607 608 609

/*
 * 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
610
static inline void expand(struct zone *zone, struct page *page,
611 612
	int low, int high, struct free_area *area,
	int migratetype)
Linus Torvalds's avatar
Linus Torvalds committed
613 614 615 616 617 618 619
{
	unsigned long size = 1 << high;

	while (high > low) {
		area--;
		high--;
		size >>= 1;
Nick Piggin's avatar
Nick Piggin committed
620
		VM_BUG_ON(bad_range(zone, &page[size]));
621
		list_add(&page[size].lru, &area->free_list[migratetype]);
Linus Torvalds's avatar
Linus Torvalds committed
622 623 624 625 626 627 628 629
		area->nr_free++;
		set_page_order(&page[size], high);
	}
}

/*
 * This page is about to be returned from the page allocator
 */
Nick Piggin's avatar
Nick Piggin committed
630
static int prep_new_page(struct page *page, int order, gfp_t gfp_flags)
Linus Torvalds's avatar
Linus Torvalds committed
631
{
Nick Piggin's avatar
Nick Piggin committed
632 633 634
	if (unlikely(page_mapcount(page) |
		(page->mapping != NULL)  |
		(page_count(page) != 0)  |
635
		(page->flags & PAGE_FLAGS_CHECK_AT_PREP))) {
636
		bad_page(page);
637
		return 1;
638
	}
639

640
	set_page_private(page, 0);
641
	set_page_refcounted(page);
Nick Piggin's avatar
Nick Piggin committed
642 643

	arch_alloc_page(page, order);
Linus Torvalds's avatar
Linus Torvalds committed
644
	kernel_map_pages(page, 1 << order, 1);
Nick Piggin's avatar
Nick Piggin committed
645 646 647 648 649 650 651

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

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

652
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
653 654
}

655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685
/*
 * Go through the free lists for the given migratetype and remove
 * the smallest available page from the freelists
 */
static struct page *__rmqueue_smallest(struct zone *zone, unsigned int order,
						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--;
		__mod_zone_page_state(zone, NR_FREE_PAGES, - (1UL << order));
		expand(zone, page, order, current_order, area, migratetype);
		return page;
	}

	return NULL;
}


686 687 688 689 690
/*
 * 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] = {
691 692 693 694
	[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 */
695 696
};

697 698
/*
 * Move the free pages in a range to the free lists of the requested type.
699
 * Note that start_page and end_pages are not aligned on a pageblock
700 701
 * boundary. If alignment is required, use move_freepages_block()
 */
Adrian Bunk's avatar
Adrian Bunk committed
702 703 704
static int move_freepages(struct zone *zone,
			  struct page *start_page, struct page *end_page,
			  int migratetype)
705 706 707
{
	struct page *page;
	unsigned long order;
708
	int pages_moved = 0;
709 710 711 712 713 714 715

#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
716
	 * grouping pages by mobility
717 718 719 720 721
	 */
	BUG_ON(page_zone(start_page) != page_zone(end_page));
#endif

	for (page = start_page; page <= end_page;) {
722 723 724
		/* Make sure we are not inadvertently changing nodes */
		VM_BUG_ON(page_to_nid(page) != zone_to_nid(zone));

725 726 727 728 729 730 731 732 733 734 735 736 737 738 739
		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;
740
		pages_moved += 1 << order;
741 742
	}

743
	return pages_moved;
744 745
}

Adrian Bunk's avatar
Adrian Bunk committed
746 747
static int move_freepages_block(struct zone *zone, struct page *page,
				int migratetype)
748 749 750 751 752
{
	unsigned long start_pfn, end_pfn;
	struct page *start_page, *end_page;

	start_pfn = page_to_pfn(page);
753
	start_pfn = start_pfn & ~(pageblock_nr_pages-1);
754
	start_page = pfn_to_page(start_pfn);
755 756
	end_page = start_page + pageblock_nr_pages - 1;
	end_pfn = start_pfn + pageblock_nr_pages - 1;
757 758 759 760 761 762 763 764 765 766

	/* 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);
}

767 768 769 770 771 772 773 774 775 776 777 778 779 780 781
/* Remove an element from the buddy allocator from the fallback list */
static struct page *__rmqueue_fallback(struct zone *zone, int order,
						int start_migratetype)
{
	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];

782 783 784
			/* MIGRATE_RESERVE handled later if necessary */
			if (migratetype == MIGRATE_RESERVE)
				continue;
785

786 787 788 789 790 791 792 793 794
			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--;

			/*
795
			 * If breaking a large block of pages, move all free
796 797 798
			 * pages to the preferred allocation list. If falling
			 * back for a reclaimable kernel allocation, be more
			 * agressive about taking ownership of free pages
799
			 */
800
			if (unlikely(current_order >= (pageblock_order >> 1)) ||
801 802 803 804 805 806
					start_migratetype == MIGRATE_RECLAIMABLE) {
				unsigned long pages;
				pages = move_freepages_block(zone, page,
								start_migratetype);

				/* Claim the whole block if over half of it is free */
807
				if (pages >= (1 << (pageblock_order-1)))
808 809 810
					set_pageblock_migratetype(page,
								start_migratetype);

811
				migratetype = start_migratetype;
812
			}
813 814 815 816 817 818 819

			/* Remove the page from the freelists */
			list_del(&page->lru);
			rmv_page_order(page);
			__mod_zone_page_state(zone, NR_FREE_PAGES,
							-(1UL << order));

820
			if (current_order == pageblock_order)
821 822 823 824 825 826 827 828
				set_pageblock_migratetype(page,
							start_migratetype);

			expand(zone, page, order, current_order, area, migratetype);
			return page;
		}
	}

829 830
	/* Use MIGRATE_RESERVE rather than fail an allocation */
	return __rmqueue_smallest(zone, order, MIGRATE_RESERVE);
831 832
}

833
/*
Linus Torvalds's avatar
Linus Torvalds committed
834 835 836
 * Do the hard work of removing an element from the buddy allocator.
 * Call me with the zone->lock already held.
 */
837 838
static struct page *__rmqueue(struct zone *zone, unsigned int order,
						int migratetype)
Linus Torvalds's avatar
Linus Torvalds committed
839 840 841
{
	struct page *page;

842
	page = __rmqueue_smallest(zone, order, migratetype);
843

844 845
	if (unlikely(!page))
		page = __rmqueue_fallback(zone, order, migratetype);
846 847

	return page;
Linus Torvalds's avatar
Linus Torvalds committed
848 849 850 851 852 853 854 855
}

/* 
 * 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, 
856 857
			unsigned long count, struct list_head *list,
			int migratetype)
Linus Torvalds's avatar
Linus Torvalds committed
858 859 860
{
	int i;
	
Nick Piggin's avatar
Nick Piggin committed
861
	spin_lock(&zone->lock);
Linus Torvalds's avatar
Linus Torvalds committed
862
	for (i = 0; i < count; ++i) {
863
		struct page *page = __rmqueue(zone, order, migratetype);
Nick Piggin's avatar
Nick Piggin committed
864
		if (unlikely(page == NULL))
Linus Torvalds's avatar
Linus Torvalds committed
865
			break;
866 867 868 869 870 871 872 873 874 875

		/*
		 * 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.
		 */
876 877
		list_add(&page->lru, list);
		set_page_private(page, migratetype);
878
		list = &page->lru;
Linus Torvalds's avatar
Linus Torvalds committed
879
	}
Nick Piggin's avatar
Nick Piggin committed
880
	spin_unlock(&zone->lock);
Nick Piggin's avatar
Nick Piggin committed
881
	return i;
Linus Torvalds's avatar
Linus Torvalds committed
882 883
}

884
#ifdef CONFIG_NUMA
885
/*
886 887 888 889
 * Called from the vmstat counter updater to drain pagesets of this
 * currently executing processor on remote nodes after they have
 * expired.
 *
890 891
 * Note that this function must be called with the thread pinned to
 * a single processor.
892
 */
893
void drain_zone_pages(struct zone *zone, struct per_cpu_pages *pcp)
894 895
{
	unsigned long flags;
896
	int to_drain;
897

898 899 900 901 902 903 904 905
	local_irq_save(flags);
	if (pcp->count >= pcp->batch)
		to_drain = pcp->batch;
	else
		to_drain = pcp->count;
	free_pages_bulk(zone, to_drain, &pcp->list, 0);
	pcp->count -= to_drain;
	local_irq_restore(flags);
906 907 908
}
#endif

909 910 911 912 913 914 915 916
/*
 * Drain pages of the indicated processor.
 *
 * The processor must either be the current processor and the
 * thread pinned to the current processor or a processor that
 * is not online.
 */
static void drain_pages(unsigned int cpu)
Linus Torvalds's avatar
Linus Torvalds committed
917
{
Nick Piggin's avatar
Nick Piggin committed
918
	unsigned long flags;
Linus Torvalds's avatar
Linus Torvalds committed
919 920
	struct zone *zone;

921
	for_each_populated_zone(zone) {
Linus Torvalds's avatar
Linus Torvalds committed
922
		struct per_cpu_pageset *pset;
923
		struct per_cpu_pages *pcp;
Linus Torvalds's avatar
Linus Torvalds committed
924

925
		pset = zone_pcp(zone, cpu);
926 927 928 929 930 931

		pcp = &pset->pcp;
		local_irq_save(flags);
		free_pages_bulk(zone, pcp->count, &pcp->list, 0);
		pcp->count = 0;
		local_irq_restore(flags);
Linus Torvalds's avatar
Linus Torvalds committed
932 933 934
	}
}

935 936 937 938 939 940 941 942 943 944 945 946 947
/*
 * Spill all of this CPU's per-cpu pages back into the buddy allocator.
 */
void drain_local_pages(void *arg)
{
	drain_pages(smp_processor_id());
}

/*
 * Spill all the per-cpu pages from all CPUs back into the buddy allocator
 */
void drain_all_pages(void)
{
948
	on_each_cpu(drain_local_pages, NULL, 1);
949 950
}

951
#ifdef CONFIG_HIBERNATION
Linus Torvalds's avatar
Linus Torvalds committed
952 953 954

void mark_free_pages(struct zone *zone)
{
955 956
	unsigned long pfn, max_zone_pfn;
	unsigned long flags;
957
	int order, t;
Linus Torvalds's avatar
Linus Torvalds committed
958 959 960 961 962 963
	struct list_head *curr;

	if (!zone->spanned_pages)
		return;

	spin_lock_irqsave(&zone->lock, flags);
964 965 966 967 968 969

	max_zone_pfn = zone->zone_start_pfn + zone->spanned_pages;
	for (pfn = zone->zone_start_pfn; pfn < max_zone_pfn; pfn++)
		if (pfn_valid(pfn)) {
			struct page *page = pfn_to_page(pfn);

970 971
			if (!swsusp_page_is_forbidden(page))
				swsusp_unset_page_free(page);
972
		}
Linus Torvalds's avatar
Linus Torvalds committed
973

974 975
	for_each_migratetype_order(order, t) {
		list_for_each(curr, &zone->free_area[order].free_list[t]) {
976
			unsigned long i;
Linus Torvalds's avatar
Linus Torvalds committed
977

978 979
			pfn = page_to_pfn(list_entry(curr, struct page, lru));
			for (i = 0; i < (1UL << order); i++)
980
				swsusp_set_page_free(pfn_to_page(pfn + i));
981
		}
982
	}
Linus Torvalds's avatar
Linus Torvalds committed
983 984
	spin_unlock_irqrestore(&zone->lock, flags);
}