page_alloc.c 118 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
22
23
/*
 *  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>
#include <linux/bootmem.h>
#include <linux/compiler.h>
24
#include <linux/kernel.h>
Linus Torvalds's avatar
Linus Torvalds committed
25
26
27
28
29
30
31
32
33
34
#include <linux/module.h>
#include <linux/suspend.h>
#include <linux/pagevec.h>
#include <linux/blkdev.h>
#include <linux/slab.h>
#include <linux/notifier.h>
#include <linux/topology.h>
#include <linux/sysctl.h>
#include <linux/cpu.h>
#include <linux/cpuset.h>
35
#include <linux/memory_hotplug.h>
Linus Torvalds's avatar
Linus Torvalds committed
36
37
#include <linux/nodemask.h>
#include <linux/vmalloc.h>
38
#include <linux/mempolicy.h>
39
#include <linux/stop_machine.h>
40
41
#include <linux/sort.h>
#include <linux/pfn.h>
42
#include <linux/backing-dev.h>
43
#include <linux/fault-inject.h>
Linus Torvalds's avatar
Linus Torvalds committed
44
45

#include <asm/tlbflush.h>
46
#include <asm/div64.h>
Linus Torvalds's avatar
Linus Torvalds committed
47
48
49
#include "internal.h"

/*
50
 * Array of node states.
Linus Torvalds's avatar
Linus Torvalds committed
51
 */
52
53
54
55
56
57
58
59
60
61
62
63
64
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);

65
unsigned long totalram_pages __read_mostly;
66
unsigned long totalreserve_pages __read_mostly;
Linus Torvalds's avatar
Linus Torvalds committed
67
long nr_swap_pages;
68
int percpu_pagelist_fraction;
Linus Torvalds's avatar
Linus Torvalds committed
69

70
static void __free_pages_ok(struct page *page, unsigned int order);
71

Linus Torvalds's avatar
Linus Torvalds committed
72
73
74
75
76
77
78
/*
 * 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
79
80
81
 *
 * 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
82
 */
83
int sysctl_lowmem_reserve_ratio[MAX_NR_ZONES-1] = {
84
#ifdef CONFIG_ZONE_DMA
85
	 256,
86
#endif
87
#ifdef CONFIG_ZONE_DMA32
88
	 256,
89
#endif
90
#ifdef CONFIG_HIGHMEM
Mel Gorman's avatar
Mel Gorman committed
91
	 32,
92
#endif
Mel Gorman's avatar
Mel Gorman committed
93
	 32,
94
};
Linus Torvalds's avatar
Linus Torvalds committed
95
96
97

EXPORT_SYMBOL(totalram_pages);

98
static char * const zone_names[MAX_NR_ZONES] = {
99
#ifdef CONFIG_ZONE_DMA
100
	 "DMA",
101
#endif
102
#ifdef CONFIG_ZONE_DMA32
103
	 "DMA32",
104
#endif
105
	 "Normal",
106
#ifdef CONFIG_HIGHMEM
Mel Gorman's avatar
Mel Gorman committed
107
	 "HighMem",
108
#endif
Mel Gorman's avatar
Mel Gorman committed
109
	 "Movable",
110
111
};

Linus Torvalds's avatar
Linus Torvalds committed
112
113
int min_free_kbytes = 1024;

114
115
unsigned long __meminitdata nr_kernel_pages;
unsigned long __meminitdata nr_all_pages;
116
static unsigned long __meminitdata dma_reserve;
Linus Torvalds's avatar
Linus Torvalds committed
117

118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
#ifdef CONFIG_ARCH_POPULATES_NODE_MAP
  /*
   * MAX_ACTIVE_REGIONS determines the maxmimum number of distinct
   * 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

139
140
141
142
  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];
143
#ifdef CONFIG_MEMORY_HOTPLUG_RESERVE
144
145
  static unsigned long __meminitdata node_boundary_start_pfn[MAX_NUMNODES];
  static unsigned long __meminitdata node_boundary_end_pfn[MAX_NUMNODES];
146
#endif /* CONFIG_MEMORY_HOTPLUG_RESERVE */
Mel Gorman's avatar
Mel Gorman committed
147
  unsigned long __initdata required_kernelcore;
148
  unsigned long __initdata required_movablecore;
149
  unsigned long __meminitdata zone_movable_pfn[MAX_NUMNODES];
Mel Gorman's avatar
Mel Gorman committed
150
151
152
153

  /* movable_zone is the "real" zone pages in ZONE_MOVABLE are taken from */
  int movable_zone;
  EXPORT_SYMBOL(movable_zone);
154
155
#endif /* CONFIG_ARCH_POPULATES_NODE_MAP */

Miklos Szeredi's avatar
Miklos Szeredi committed
156
157
158
159
160
#if MAX_NUMNODES > 1
int nr_node_ids __read_mostly = MAX_NUMNODES;
EXPORT_SYMBOL(nr_node_ids);
#endif

161
#ifdef CONFIG_PAGE_GROUP_BY_MOBILITY
162
163
164
165
166
167
168
169
170
171
172
173
174
static inline int get_pageblock_migratetype(struct page *page)
{
	return get_pageblock_flags_group(page, PB_migrate, PB_migrate_end);
}

static void set_pageblock_migratetype(struct page *page, int migratetype)
{
	set_pageblock_flags_group(page, (unsigned long)migratetype,
					PB_migrate, PB_migrate_end);
}

static inline int gfpflags_to_migratetype(gfp_t gfp_flags)
{
175
176
177
178
	WARN_ON((gfp_flags & GFP_MOVABLE_MASK) == GFP_MOVABLE_MASK);

	return (((gfp_flags & __GFP_MOVABLE) != 0) << 1) |
		((gfp_flags & __GFP_RECLAIMABLE) != 0);
179
180
}

181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
#else
static inline int get_pageblock_migratetype(struct page *page)
{
	return MIGRATE_UNMOVABLE;
}

static void set_pageblock_migratetype(struct page *page, int migratetype)
{
}

static inline int gfpflags_to_migratetype(gfp_t gfp_flags)
{
	return MIGRATE_UNMOVABLE;
}
#endif /* CONFIG_PAGE_GROUP_BY_MOBILITY */

Nick Piggin's avatar
Nick Piggin committed
197
#ifdef CONFIG_DEBUG_VM
198
static int page_outside_zone_boundaries(struct zone *zone, struct page *page)
Linus Torvalds's avatar
Linus Torvalds committed
199
{
200
201
202
	int ret = 0;
	unsigned seq;
	unsigned long pfn = page_to_pfn(page);
203

204
205
206
207
208
209
210
211
212
	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;
213
214
215
216
}

static int page_is_consistent(struct zone *zone, struct page *page)
{
217
	if (!pfn_valid_within(page_to_pfn(page)))
218
		return 0;
Linus Torvalds's avatar
Linus Torvalds committed
219
	if (zone != page_zone(page))
220
221
222
223
224
225
226
227
228
229
		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
230
		return 1;
231
232
233
	if (!page_is_consistent(zone, page))
		return 1;

Linus Torvalds's avatar
Linus Torvalds committed
234
235
	return 0;
}
Nick Piggin's avatar
Nick Piggin committed
236
237
238
239
240
241
242
#else
static inline int bad_range(struct zone *zone, struct page *page)
{
	return 0;
}
#endif

243
static void bad_page(struct page *page)
Linus Torvalds's avatar
Linus Torvalds committed
244
{
245
	printk(KERN_EMERG "Bad page state in process '%s'\n"
246
247
248
		KERN_EMERG "page:%p flags:0x%0*lx mapping:%p mapcount:%d count:%d\n"
		KERN_EMERG "Trying to fix it up, but a reboot is needed\n"
		KERN_EMERG "Backtrace:\n",
249
250
251
		current->comm, page, (int)(2*sizeof(unsigned long)),
		(unsigned long)page->flags, page->mapping,
		page_mapcount(page), page_count(page));
Linus Torvalds's avatar
Linus Torvalds committed
252
	dump_stack();
253
254
	page->flags &= ~(1 << PG_lru	|
			1 << PG_private |
Linus Torvalds's avatar
Linus Torvalds committed
255
256
257
			1 << PG_locked	|
			1 << PG_active	|
			1 << PG_dirty	|
258
259
			1 << PG_reclaim |
			1 << PG_slab    |
Linus Torvalds's avatar
Linus Torvalds committed
260
			1 << PG_swapcache |
261
262
			1 << PG_writeback |
			1 << PG_buddy );
Linus Torvalds's avatar
Linus Torvalds committed
263
264
265
	set_page_count(page, 0);
	reset_page_mapcount(page);
	page->mapping = NULL;
266
	add_taint(TAINT_BAD_PAGE);
Linus Torvalds's avatar
Linus Torvalds committed
267
268
269
270
271
272
273
274
275
276
277
278
}

/*
 * 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).
 *
279
280
281
 * 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
282
 */
283
284
285

static void free_compound_page(struct page *page)
{
286
	__free_pages_ok(page, compound_order(page));
287
288
}

Linus Torvalds's avatar
Linus Torvalds committed
289
290
291
292
293
static void prep_compound_page(struct page *page, unsigned long order)
{
	int i;
	int nr_pages = 1 << order;

294
	set_compound_page_dtor(page, free_compound_page);
295
	set_compound_order(page, order);
296
	__SetPageHead(page);
297
	for (i = 1; i < nr_pages; i++) {
Linus Torvalds's avatar
Linus Torvalds committed
298
299
		struct page *p = page + i;

300
301
		__SetPageTail(p);
		p->first_page = page;
Linus Torvalds's avatar
Linus Torvalds committed
302
303
304
305
306
307
308
309
	}
}

static void destroy_compound_page(struct page *page, unsigned long order)
{
	int i;
	int nr_pages = 1 << order;

310
	if (unlikely(compound_order(page) != order))
311
		bad_page(page);
Linus Torvalds's avatar
Linus Torvalds committed
312

313
	if (unlikely(!PageHead(page)))
314
			bad_page(page);
315
	__ClearPageHead(page);
316
	for (i = 1; i < nr_pages; i++) {
Linus Torvalds's avatar
Linus Torvalds committed
317
318
		struct page *p = page + i;

319
		if (unlikely(!PageTail(p) |
320
				(p->first_page != page)))
321
			bad_page(page);
322
		__ClearPageTail(p);
Linus Torvalds's avatar
Linus Torvalds committed
323
324
325
	}
}

Nick Piggin's avatar
Nick Piggin committed
326
327
328
329
static inline void prep_zero_page(struct page *page, int order, gfp_t gfp_flags)
{
	int i;

Nick Piggin's avatar
Nick Piggin committed
330
	VM_BUG_ON((gfp_flags & (__GFP_WAIT | __GFP_HIGHMEM)) == __GFP_HIGHMEM);
331
332
333
334
	/*
	 * 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
335
	VM_BUG_ON((gfp_flags & __GFP_HIGHMEM) && in_interrupt());
Nick Piggin's avatar
Nick Piggin committed
336
337
338
339
	for (i = 0; i < (1 << order); i++)
		clear_highpage(page + i);
}

Linus Torvalds's avatar
Linus Torvalds committed
340
341
342
343
344
/*
 * function for dealing with page's order in buddy system.
 * zone->lock is already acquired when we use these.
 * So, we don't need atomic page->flags operations here.
 */
345
346
static inline unsigned long page_order(struct page *page)
{
347
	return page_private(page);
Linus Torvalds's avatar
Linus Torvalds committed
348
349
}

350
351
static inline void set_page_order(struct page *page, int order)
{
352
	set_page_private(page, order);
353
	__SetPageBuddy(page);
Linus Torvalds's avatar
Linus Torvalds committed
354
355
356
357
}

static inline void rmv_page_order(struct page *page)
{
358
	__ClearPageBuddy(page);
359
	set_page_private(page, 0);
Linus Torvalds's avatar
Linus Torvalds committed
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
}

/*
 * 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
377
 * Assumption: *_mem_map is contiguous at least up to MAX_ORDER
Linus Torvalds's avatar
Linus Torvalds committed
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
 */
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
396
 * (a) the buddy is not in a hole &&
397
 * (b) the buddy is in the buddy system &&
398
399
 * (c) a page and its buddy have the same order &&
 * (d) a page and its buddy are in the same zone.
400
401
402
 *
 * 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
403
 *
404
 * For recording page's order, we use page_private(page).
Linus Torvalds's avatar
Linus Torvalds committed
405
 */
406
407
static inline int page_is_buddy(struct page *page, struct page *buddy,
								int order)
Linus Torvalds's avatar
Linus Torvalds committed
408
{
409
	if (!pfn_valid_within(page_to_pfn(buddy)))
Nick Piggin's avatar
Nick Piggin committed
410
411
		return 0;

412
413
414
415
416
	if (page_zone_id(page) != page_zone_id(buddy))
		return 0;

	if (PageBuddy(buddy) && page_order(buddy) == order) {
		BUG_ON(page_count(buddy) != 0);
417
		return 1;
418
	}
419
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
}

/*
 * 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
435
 * free pages of length of (1 << order) and marked with PG_buddy. Page's
436
 * order is recorded in page_private(page) field.
Linus Torvalds's avatar
Linus Torvalds committed
437
438
439
440
441
442
443
444
445
 * 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
446
static inline void __free_one_page(struct page *page,
Linus Torvalds's avatar
Linus Torvalds committed
447
448
449
450
		struct zone *zone, unsigned int order)
{
	unsigned long page_idx;
	int order_size = 1 << order;
451
	int migratetype = get_pageblock_migratetype(page);
Linus Torvalds's avatar
Linus Torvalds committed
452

453
	if (unlikely(PageCompound(page)))
Linus Torvalds's avatar
Linus Torvalds committed
454
455
456
457
		destroy_compound_page(page, order);

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

Nick Piggin's avatar
Nick Piggin committed
458
459
	VM_BUG_ON(page_idx & (order_size - 1));
	VM_BUG_ON(bad_range(zone, page));
Linus Torvalds's avatar
Linus Torvalds committed
460

461
	__mod_zone_page_state(zone, NR_FREE_PAGES, order_size);
Linus Torvalds's avatar
Linus Torvalds committed
462
463
464
465
466
	while (order < MAX_ORDER-1) {
		unsigned long combined_idx;
		struct page *buddy;

		buddy = __page_find_buddy(page, page_idx, order);
467
		if (!page_is_buddy(page, buddy, order))
Linus Torvalds's avatar
Linus Torvalds committed
468
			break;		/* Move the buddy up one level. */
Nick Piggin's avatar
Nick Piggin committed
469

Linus Torvalds's avatar
Linus Torvalds committed
470
		list_del(&buddy->lru);
471
		zone->free_area[order].nr_free--;
Linus Torvalds's avatar
Linus Torvalds committed
472
		rmv_page_order(buddy);
Nick Piggin's avatar
Nick Piggin committed
473
		combined_idx = __find_combined_index(page_idx, order);
Linus Torvalds's avatar
Linus Torvalds committed
474
475
476
477
478
		page = page + (combined_idx - page_idx);
		page_idx = combined_idx;
		order++;
	}
	set_page_order(page, order);
479
480
	list_add(&page->lru,
		&zone->free_area[order].free_list[migratetype]);
Linus Torvalds's avatar
Linus Torvalds committed
481
482
483
	zone->free_area[order].nr_free++;
}

484
static inline int free_pages_check(struct page *page)
Linus Torvalds's avatar
Linus Torvalds committed
485
{
Nick Piggin's avatar
Nick Piggin committed
486
487
488
	if (unlikely(page_mapcount(page) |
		(page->mapping != NULL)  |
		(page_count(page) != 0)  |
Linus Torvalds's avatar
Linus Torvalds committed
489
490
491
492
493
494
495
		(page->flags & (
			1 << PG_lru	|
			1 << PG_private |
			1 << PG_locked	|
			1 << PG_active	|
			1 << PG_slab	|
			1 << PG_swapcache |
Nick Piggin's avatar
Nick Piggin committed
496
			1 << PG_writeback |
497
498
			1 << PG_reserved |
			1 << PG_buddy ))))
499
		bad_page(page);
Linus Torvalds's avatar
Linus Torvalds committed
500
	if (PageDirty(page))
Nick Piggin's avatar
Nick Piggin committed
501
		__ClearPageDirty(page);
502
503
504
505
506
507
	/*
	 * For now, we report if PG_reserved was found set, but do not
	 * clear it, and do not free the page.  But we shall soon need
	 * to do more, for when the ZERO_PAGE count wraps negative.
	 */
	return PageReserved(page);
Linus Torvalds's avatar
Linus Torvalds committed
508
509
510
511
512
}

/*
 * Frees a list of pages. 
 * Assumes all pages on list are in same zone, and of same order.
513
 * count is the number of pages to free.
Linus Torvalds's avatar
Linus Torvalds committed
514
515
516
517
518
519
520
 *
 * 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
521
522
static void free_pages_bulk(struct zone *zone, int count,
					struct list_head *list, int order)
Linus Torvalds's avatar
Linus Torvalds committed
523
{
Nick Piggin's avatar
Nick Piggin committed
524
	spin_lock(&zone->lock);
Linus Torvalds's avatar
Linus Torvalds committed
525
526
	zone->all_unreclaimable = 0;
	zone->pages_scanned = 0;
Nick Piggin's avatar
Nick Piggin committed
527
528
529
	while (count--) {
		struct page *page;

Nick Piggin's avatar
Nick Piggin committed
530
		VM_BUG_ON(list_empty(list));
Linus Torvalds's avatar
Linus Torvalds committed
531
		page = list_entry(list->prev, struct page, lru);
Nick Piggin's avatar
Nick Piggin committed
532
		/* have to delete it as __free_one_page list manipulates */
Linus Torvalds's avatar
Linus Torvalds committed
533
		list_del(&page->lru);
Nick Piggin's avatar
Nick Piggin committed
534
		__free_one_page(page, zone, order);
Linus Torvalds's avatar
Linus Torvalds committed
535
	}
Nick Piggin's avatar
Nick Piggin committed
536
	spin_unlock(&zone->lock);
Linus Torvalds's avatar
Linus Torvalds committed
537
538
}

Nick Piggin's avatar
Nick Piggin committed
539
static void free_one_page(struct zone *zone, struct page *page, int order)
Linus Torvalds's avatar
Linus Torvalds committed
540
{
541
542
543
	spin_lock(&zone->lock);
	zone->all_unreclaimable = 0;
	zone->pages_scanned = 0;
544
	__free_one_page(page, zone, order);
545
	spin_unlock(&zone->lock);
Nick Piggin's avatar
Nick Piggin committed
546
547
548
549
550
}

static void __free_pages_ok(struct page *page, unsigned int order)
{
	unsigned long flags;
Linus Torvalds's avatar
Linus Torvalds committed
551
	int i;
552
	int reserved = 0;
Linus Torvalds's avatar
Linus Torvalds committed
553
554

	for (i = 0 ; i < (1 << order) ; ++i)
555
		reserved += free_pages_check(page + i);
556
557
558
	if (reserved)
		return;

Nick Piggin's avatar
Nick Piggin committed
559
560
	if (!PageHighMem(page))
		debug_check_no_locks_freed(page_address(page),PAGE_SIZE<<order);
Nick Piggin's avatar
Nick Piggin committed
561
	arch_free_page(page, order);
Nick Piggin's avatar
Nick Piggin committed
562
	kernel_map_pages(page, 1 << order, 0);
Nick Piggin's avatar
Nick Piggin committed
563

Nick Piggin's avatar
Nick Piggin committed
564
	local_irq_save(flags);
565
	__count_vm_events(PGFREE, 1 << order);
Nick Piggin's avatar
Nick Piggin committed
566
	free_one_page(page_zone(page), page, order);
Nick Piggin's avatar
Nick Piggin committed
567
	local_irq_restore(flags);
Linus Torvalds's avatar
Linus Torvalds committed
568
569
}

570
571
572
573
574
575
576
577
/*
 * permit the bootmem allocator to evade page validation on high-order frees
 */
void fastcall __init __free_pages_bootmem(struct page *page, unsigned int order)
{
	if (order == 0) {
		__ClearPageReserved(page);
		set_page_count(page, 0);
578
		set_page_refcounted(page);
Nick Piggin's avatar
Nick Piggin committed
579
		__free_page(page);
580
581
582
	} else {
		int loop;

Nick Piggin's avatar
Nick Piggin committed
583
		prefetchw(page);
584
585
586
		for (loop = 0; loop < BITS_PER_LONG; loop++) {
			struct page *p = &page[loop];

Nick Piggin's avatar
Nick Piggin committed
587
588
			if (loop + 1 < BITS_PER_LONG)
				prefetchw(p + 1);
589
590
591
592
			__ClearPageReserved(p);
			set_page_count(p, 0);
		}

593
		set_page_refcounted(page);
Nick Piggin's avatar
Nick Piggin committed
594
		__free_pages(page, order);
595
596
597
	}
}

Linus Torvalds's avatar
Linus Torvalds committed
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612

/*
 * 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
613
static inline void expand(struct zone *zone, struct page *page,
614
615
	int low, int high, struct free_area *area,
	int migratetype)
Linus Torvalds's avatar
Linus Torvalds committed
616
617
618
619
620
621
622
{
	unsigned long size = 1 << high;

	while (high > low) {
		area--;
		high--;
		size >>= 1;
Nick Piggin's avatar
Nick Piggin committed
623
		VM_BUG_ON(bad_range(zone, &page[size]));
624
		list_add(&page[size].lru, &area->free_list[migratetype]);
Linus Torvalds's avatar
Linus Torvalds committed
625
626
627
628
629
630
631
632
		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
633
static int prep_new_page(struct page *page, int order, gfp_t gfp_flags)
Linus Torvalds's avatar
Linus Torvalds committed
634
{
Nick Piggin's avatar
Nick Piggin committed
635
636
637
	if (unlikely(page_mapcount(page) |
		(page->mapping != NULL)  |
		(page_count(page) != 0)  |
638
639
		(page->flags & (
			1 << PG_lru	|
Linus Torvalds's avatar
Linus Torvalds committed
640
641
642
643
			1 << PG_private	|
			1 << PG_locked	|
			1 << PG_active	|
			1 << PG_dirty	|
644
			1 << PG_slab    |
Linus Torvalds's avatar
Linus Torvalds committed
645
			1 << PG_swapcache |
Nick Piggin's avatar
Nick Piggin committed
646
			1 << PG_writeback |
647
648
			1 << PG_reserved |
			1 << PG_buddy ))))
649
		bad_page(page);
Linus Torvalds's avatar
Linus Torvalds committed
650

651
652
653
654
655
656
657
	/*
	 * For now, we report if PG_reserved was found set, but do not
	 * clear it, and do not allocate the page: as a safety net.
	 */
	if (PageReserved(page))
		return 1;

658
	page->flags &= ~(1 << PG_uptodate | 1 << PG_error | 1 << PG_readahead |
Linus Torvalds's avatar
Linus Torvalds committed
659
			1 << PG_referenced | 1 << PG_arch_1 |
660
			1 << PG_owner_priv_1 | 1 << PG_mappedtodisk);
661
	set_page_private(page, 0);
662
	set_page_refcounted(page);
Nick Piggin's avatar
Nick Piggin committed
663
664

	arch_alloc_page(page, order);
Linus Torvalds's avatar
Linus Torvalds committed
665
	kernel_map_pages(page, 1 << order, 1);
Nick Piggin's avatar
Nick Piggin committed
666
667
668
669
670
671
672

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

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

673
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
674
675
}

676
#ifdef CONFIG_PAGE_GROUP_BY_MOBILITY
677
678
679
680
681
/*
 * 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] = {
682
683
684
	[MIGRATE_UNMOVABLE]   = { MIGRATE_RECLAIMABLE, MIGRATE_MOVABLE   },
	[MIGRATE_RECLAIMABLE] = { MIGRATE_UNMOVABLE,   MIGRATE_MOVABLE   },
	[MIGRATE_MOVABLE]     = { MIGRATE_RECLAIMABLE, MIGRATE_UNMOVABLE },
685
686
};

687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
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
742
743
744
745
746
747
748
749
750
751
752
/*
 * Move the free pages in a range to the free lists of the requested type.
 * Note that start_page and end_pages are not aligned in a MAX_ORDER_NR_PAGES
 * boundary. If alignment is required, use move_freepages_block()
 */
int move_freepages(struct zone *zone,
			struct page *start_page, struct page *end_page,
			int migratetype)
{
	struct page *page;
	unsigned long order;
	int blocks_moved = 0;

#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
	 * CONFIG_PAGE_GROUP_BY_MOBILITY
	 */
	BUG_ON(page_zone(start_page) != page_zone(end_page));
#endif

	for (page = start_page; page <= end_page;) {
		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;
		blocks_moved++;
	}

	return blocks_moved;
}

int move_freepages_block(struct zone *zone, struct page *page, int migratetype)
{
	unsigned long start_pfn, end_pfn;
	struct page *start_page, *end_page;

	start_pfn = page_to_pfn(page);
	start_pfn = start_pfn & ~(MAX_ORDER_NR_PAGES-1);
	start_page = pfn_to_page(start_pfn);
	end_page = start_page + MAX_ORDER_NR_PAGES - 1;
	end_pfn = start_pfn + MAX_ORDER_NR_PAGES - 1;

	/* 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);
}

753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
/* 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];

			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--;

			/*
777
778
			 * If breaking a large block of pages, move all free
			 * pages to the preferred allocation list
779
			 */
780
			if (unlikely(current_order >= MAX_ORDER / 2)) {
781
				migratetype = start_migratetype;
782
783
				move_freepages_block(zone, page, migratetype);
			}
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801

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

			if (current_order == MAX_ORDER - 1)
				set_pageblock_migratetype(page,
							start_migratetype);

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

	return NULL;
}
802
803
804
805
806
807
808
#else
static struct page *__rmqueue_fallback(struct zone *zone, int order,
						int start_migratetype)
{
	return NULL;
}
#endif /* CONFIG_PAGE_GROUP_BY_MOBILITY */
809

Linus Torvalds's avatar
Linus Torvalds committed
810
811
812
813
/* 
 * Do the hard work of removing an element from the buddy allocator.
 * Call me with the zone->lock already held.
 */
814
815
static struct page *__rmqueue(struct zone *zone, unsigned int order,
						int migratetype)
Linus Torvalds's avatar
Linus Torvalds committed
816
817
818
819
820
{
	struct free_area * area;
	unsigned int current_order;
	struct page *page;

821
	/* Find a page of the appropriate size in the preferred list */
Linus Torvalds's avatar
Linus Torvalds committed
822
	for (current_order = order; current_order < MAX_ORDER; ++current_order) {
823
824
		area = &(zone->free_area[current_order]);
		if (list_empty(&area->free_list[migratetype]))
Linus Torvalds's avatar
Linus Torvalds committed
825
826
			continue;

827
828
		page = list_entry(area->free_list[migratetype].next,
							struct page, lru);
Linus Torvalds's avatar
Linus Torvalds committed
829
830
831
		list_del(&page->lru);
		rmv_page_order(page);
		area->nr_free--;
832
		__mod_zone_page_state(zone, NR_FREE_PAGES, - (1UL << order));
833
834
		expand(zone, page, order, current_order, area, migratetype);
		goto got_page;
Linus Torvalds's avatar
Linus Torvalds committed
835
836
	}

837
838
839
840
841
	page = __rmqueue_fallback(zone, order, migratetype);

got_page:

	return page;
Linus Torvalds's avatar
Linus Torvalds committed
842
843
844
845
846
847
848
849
}

/* 
 * 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, 
850
851
			unsigned long count, struct list_head *list,
			int migratetype)
Linus Torvalds's avatar
Linus Torvalds committed
852
853
854
{
	int i;
	
Nick Piggin's avatar
Nick Piggin committed
855
	spin_lock(&zone->lock);
Linus Torvalds's avatar
Linus Torvalds committed
856
	for (i = 0; i < count; ++i) {
857
		struct page *page = __rmqueue(zone, order, migratetype);
Nick Piggin's avatar
Nick Piggin committed
858
		if (unlikely(page == NULL))
Linus Torvalds's avatar
Linus Torvalds committed
859
			break;
860
861
		list_add(&page->lru, list);
		set_page_private(page, migratetype);
Linus Torvalds's avatar
Linus Torvalds committed
862
	}
Nick Piggin's avatar
Nick Piggin committed
863
	spin_unlock(&zone->lock);
Nick Piggin's avatar
Nick Piggin committed
864
	return i;
Linus Torvalds's avatar
Linus Torvalds committed
865
866
}

867
#ifdef CONFIG_NUMA
868
/*
869
870
871
872
 * Called from the vmstat counter updater to drain pagesets of this
 * currently executing processor on remote nodes after they have
 * expired.
 *
873
874
 * Note that this function must be called with the thread pinned to
 * a single processor.
875
 */
876
void drain_zone_pages(struct zone *zone, struct per_cpu_pages *pcp)
877
878
{
	unsigned long flags;
879
	int to_drain;
880

881
882
883
884
885
886
887
888
	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);
889
890
891
}
#endif

Linus Torvalds's avatar
Linus Torvalds committed
892
893
static void __drain_pages(unsigned int cpu)
{
Nick Piggin's avatar
Nick Piggin committed
894
	unsigned long flags;
Linus Torvalds's avatar
Linus Torvalds committed
895
896
897
898
899
900
	struct zone *zone;
	int i;

	for_each_zone(zone) {
		struct per_cpu_pageset *pset;

901
902
903
		if (!populated_zone(zone))
			continue;

904
		pset = zone_pcp(zone, cpu);
Linus Torvalds's avatar
Linus Torvalds committed
905
906
907
908
		for (i = 0; i < ARRAY_SIZE(pset->pcp); i++) {
			struct per_cpu_pages *pcp;

			pcp = &pset->pcp[i];
Nick Piggin's avatar
Nick Piggin committed
909
			local_irq_save(flags);
Nick Piggin's avatar
Nick Piggin committed
910
911
			free_pages_bulk(zone, pcp->count, &pcp->list, 0);
			pcp->count = 0;
Nick Piggin's avatar
Nick Piggin committed
912
			local_irq_restore(flags);
Linus Torvalds's avatar
Linus Torvalds committed
913
914
915
916
		}
	}
}

917
#ifdef CONFIG_HIBERNATION
Linus Torvalds's avatar
Linus Torvalds committed
918
919
920

void mark_free_pages(struct zone *zone)
{
921
922
	unsigned long pfn, max_zone_pfn;
	unsigned long flags;
923
	int order, t;
Linus Torvalds's avatar
Linus Torvalds committed
924
925
926
927
928
929
	struct list_head *curr;

	if (!zone->spanned_pages)
		return;

	spin_lock_irqsave(&zone->lock, flags);
930
931
932
933
934
935

	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);

936
937
			if (!swsusp_page_is_forbidden(page))
				swsusp_unset_page_free(page);
938
		}
Linus Torvalds's avatar
Linus Torvalds committed
939

940
941
	for_each_migratetype_order(order, t) {
		list_for_each(curr, &zone->free_area[order].free_list[t]) {
942
			unsigned long i;
Linus Torvalds's avatar
Linus Torvalds committed
943

944
945
			pfn = page_to_pfn(list_entry(curr, struct page, lru));
			for (i = 0; i < (1UL << order); i++)
946
				swsusp_set_page_free(pfn_to_page(pfn + i));
947
		}
948
	}
Linus Torvalds's avatar
Linus Torvalds committed
949
950
	spin_unlock_irqrestore(&zone->lock, flags);
}
951
#endif /* CONFIG_PM */
Linus Torvalds's avatar
Linus Torvalds committed
952

953
#if defined(CONFIG_HIBERNATION) || defined(CONFIG_PAGE_GROUP_BY_MOBILITY)
Linus Torvalds's avatar
Linus Torvalds committed
954
955
956
957
958
959
960
961
962
963
964
/*
 * Spill all of this CPU's per-cpu pages back into the buddy allocator.
 */
void drain_local_pages(void)
{
	unsigned long flags;

	local_irq_save(flags);	
	__drain_pages(smp_processor_id());
	local_irq_restore(flags);	
}
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986

void smp_drain_local_pages(void *arg)
{
	drain_local_pages();
}

/*
 * Spill all the per-cpu pages from all CPUs back into the buddy allocator
 */
void drain_all_local_pages(void)
{
	unsigned long flags;

	local_irq_save(flags);
	__drain_pages(smp_processor_id());
	local_irq_restore(flags);

	smp_call_function(smp_drain_local_pages, NULL, 0, 1);
}
#else
void drain_all_local_pages(void) {}
#endif /* CONFIG_HIBERNATION || CONFIG_PAGE_GROUP_BY_MOBILITY */
Linus Torvalds's avatar
Linus Torvalds committed
987
988
989
990
991
992
993
994
995
996
997
998

/*
 * Free a 0-order page
 */
static void fastcall free_hot_cold_page(struct page *page, int cold)
{
	struct zone *zone = page_zone(page);
	struct per_cpu_pages *pcp;
	unsigned long flags;

	if (PageAnon(page))
		page->mapping = NULL;
999
	if (free_pages_check(page))
1000
1001
		return;

Nick Piggin's avatar
Nick Piggin committed
1002
1003
	if (!PageHighMem(page))
		debug_check_no_locks_freed(page_address(page), PAGE_SIZE);
Nick Piggin's avatar
Nick Piggin committed
1004
	arch_free_page(page, 0);
1005
1006
	kernel_map_pages(page, 1, 0);

1007
	pcp = &zone_pcp(zone, get_cpu())->pcp[cold];
Linus Torvalds's avatar
Linus Torvalds committed
1008
	local_irq_save(flags);
1009
	__count_vm_event(PGFREE);
Linus Torvalds's avatar
Linus Torvalds committed
1010
	list_add(&page->lru, &pcp->list);
1011
	set_page_private(page, get_pageblock_migratetype(page));
Linus Torvalds's avatar
Linus Torvalds committed
1012
	pcp->count++;
Nick Piggin's avatar
Nick Piggin committed
1013
1014
1015
1016
	if (pcp->count >= pcp->high) {
		free_pages_bulk(zone, pcp->batch, &pcp->list, 0);
		pcp->count -= pcp->batch;
	}
Linus Torvalds's avatar
Linus Torvalds committed
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
	local_irq_restore(flags);
	put_cpu();
}

void fastcall free_hot_page(struct page *page)
{
	free_hot_cold_page(page, 0);
}
	
void fastcall free_cold_page(struct page *page)
{
	free_hot_cold_page(page, 1);
}

1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
/*
 * split_page takes a non-compound higher-order page, and splits it into
 * n (1<<order) sub-pages: page[0..n]
 * Each sub-page must be freed individually.
 *
 * Note: this is probably too low level an operation for use in drivers.
 * Please consult with lkml before using this in your driver.
 */
void split_page(struct page *page, unsigned int order)
{
	int i;

Nick Piggin's avatar
Nick Piggin committed
1043
1044
	VM_BUG_ON(PageCompound(page));
	VM_BUG_ON(!page_count(page));
1045
1046
	for (i = 1; i < (1 << order); i++)
		set_page_refcounted(page + i);
1047
1048
}

Linus Torvalds's avatar
Linus Torvalds committed
1049
1050
1051
1052
1053
/*
 * Really, prep_compound_page() should be called from __rmqueue_bulk().  But
 * we cheat by calling it from here, in the order > 0 path.  Saves a branch
 * or two.
 */
Nick Piggin's avatar
Nick Piggin committed
1054
1055
static struct page *buffered_rmqueue(struct zonelist *zonelist,
			struct zone *zone, int order, gfp_t gfp_flags)
Linus Torvalds's avatar
Linus Torvalds committed
1056
1057
{
	unsigned long flags;
1058
	struct page *page;
Linus Torvalds's avatar
Linus Torvalds committed
1059
	int cold = !!(gfp_flags & __GFP_COLD);
Nick Piggin's avatar
Nick Piggin committed
1060
	int cpu;
1061
	int migratetype = gfpflags_to_migratetype(gfp_flags);
Linus Torvalds's avatar
Linus Torvalds committed
1062

1063
again:
Nick Piggin's avatar
Nick Piggin committed
1064
	cpu  = get_cpu();
Nick Piggin's avatar
Nick Piggin committed
1065
	if (likely(order == 0)) {
Linus Torvalds's avatar
Linus Torvalds committed
1066
1067
		struct per_cpu_pages *pcp;

Nick Piggin's avatar
Nick Piggin committed
1068
		pcp = &zone_pcp(zone, cpu)->pcp[cold];
Linus Torvalds's avatar
Linus Torvalds committed
1069
		local_irq_save(flags);
Nick Piggin's avatar
Nick Piggin committed
1070
		if (!pcp->count) {
1071
			pcp->count = rmqueue_bulk(zone, 0,
1072
					pcp->batch, &pcp->list, migratetype);
Nick Piggin's avatar
Nick Piggin committed
1073
1074
			if (unlikely(!pcp->count))
				goto failed;
Linus Torvalds's avatar
Linus Torvalds committed
1075
		}
1076
1077

#ifdef CONFIG_PAGE_GROUP_BY_MOBILITY
1078
		/* Find a page of the appropriate migrate type */
1079
1080
		list_for_each_entry(page, &pcp->list, lru)
			if (page_private(page) == migratetype)
1081
1082
				break;

1083
1084
		/* Allocate more to the pcp list if necessary */
		if (unlikely(&page->lru == &pcp->list)) {
1085
1086
1087
1088
			pcp->count += rmqueue_bulk(zone, 0,
					pcp->batch, &pcp->list, migratetype);
			page = list_entry(pcp->list.next, struct page, lru);
		}
1089
1090
1091
1092
1093
1094
#else
		page = list_entry(pcp->list.next, struct page, lru);
#endif /* CONFIG_PAGE_GROUP_BY_MOBILITY */

		list_del(&page->lru);
		pcp->count--;
Rohit Seth's avatar
Rohit Seth committed
1095
	} else {
Linus Torvalds's avatar
Linus Torvalds committed
1096
		spin_lock_irqsave(&zone->lock, flags);
1097
		page = __rmqueue(zone, order, migratetype);
Nick Piggin's avatar
Nick Piggin committed
1098
1099
1100
		spin_unlock(&zone->lock);
		if (!page)
			goto failed;
Linus Torvalds's avatar
Linus Torvalds committed
1101
1102
	}

1103
	__count_zone_vm_events(PGALLOC, zone, 1 << order);
1104
	zone_statistics(zonelist, zone);
Nick Piggin's avatar
Nick Piggin committed
1105
1106
	local_irq_restore(flags);
	put_cpu();
Linus Torvalds's avatar
Linus Torvalds committed
1107

Nick Piggin's avatar
Nick Piggin committed
1108
	VM_BUG_ON(bad_range(zone, page));
Nick Piggin's avatar
Nick Piggin committed
1109
	if (prep_new_page(page, order, gfp_flags))
Nick Piggin's avatar
Nick Piggin committed
1110
		goto again;
Linus Torvalds's avatar
Linus Torvalds committed
1111
	return page;
Nick Piggin's avatar
Nick Piggin committed
1112
1113
1114
1115
1116

failed:
	local_irq_restore(flags);
	put_cpu();
	return NULL;
Linus Torvalds's avatar
Linus Torvalds committed
1117
1118
}

Rohit Seth's avatar
Rohit Seth committed
1119
#define ALLOC_NO_WATERMARKS	0x01 /* don't check watermarks at all */
1120
1121
1122
1123
1124
1125
#define ALLOC_WMARK_MIN		0x02 /* use pages_min watermark */
#define ALLOC_WMARK_LOW		0x04 /* use pages_low watermark */
#define ALLOC_WMARK_HIGH	0x08 /* use pages_high watermark */
#define ALLOC_HARDER		0x10 /* try to alloc harder */
#define ALLOC_HIGH		0x20 /* __GFP_HIGH set */
#define ALLOC_CPUSET		0x40 /* check for correct cpuset */
Rohit Seth's avatar
Rohit Seth committed
1126

1127
1128
1129
1130
1131
1132
1133
#ifdef CONFIG_FAIL_PAGE_ALLOC

static struct fail_page_alloc_attr {
	struct fault_attr attr;

	u32 ignore_gfp_highmem;
	u32 ignore_gfp_wait;
1134
	u32 min_order;
1135
1136
1137
1138
1139

#ifdef CONFIG_FAULT_INJECTION_DEBUG_FS

	struct dentry *ignore_gfp_highmem_file;
	struct dentry *ignore_gfp_wait_file;