page_alloc.c 115 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
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
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)
{
	return ((gfp_flags & __GFP_MOVABLE) != 0);
}

Nick Piggin's avatar
Nick Piggin committed
177
#ifdef CONFIG_DEBUG_VM
178
static int page_outside_zone_boundaries(struct zone *zone, struct page *page)
Linus Torvalds's avatar
Linus Torvalds committed
179
{
180
181
182
	int ret = 0;
	unsigned seq;
	unsigned long pfn = page_to_pfn(page);
183

184
185
186
187
188
189
190
191
192
	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;
193
194
195
196
}

static int page_is_consistent(struct zone *zone, struct page *page)
{
197
	if (!pfn_valid_within(page_to_pfn(page)))
198
		return 0;
Linus Torvalds's avatar
Linus Torvalds committed
199
	if (zone != page_zone(page))
200
201
202
203
204
205
206
207
208
209
		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
210
		return 1;
211
212
213
	if (!page_is_consistent(zone, page))
		return 1;

Linus Torvalds's avatar
Linus Torvalds committed
214
215
	return 0;
}
Nick Piggin's avatar
Nick Piggin committed
216
217
218
219
220
221
222
#else
static inline int bad_range(struct zone *zone, struct page *page)
{
	return 0;
}
#endif

223
static void bad_page(struct page *page)
Linus Torvalds's avatar
Linus Torvalds committed
224
{
225
	printk(KERN_EMERG "Bad page state in process '%s'\n"
226
227
228
		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",
229
230
231
		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
232
	dump_stack();
233
234
	page->flags &= ~(1 << PG_lru	|
			1 << PG_private |
Linus Torvalds's avatar
Linus Torvalds committed
235
236
237
			1 << PG_locked	|
			1 << PG_active	|
			1 << PG_dirty	|
238
239
			1 << PG_reclaim |
			1 << PG_slab    |
Linus Torvalds's avatar
Linus Torvalds committed
240
			1 << PG_swapcache |
241
242
			1 << PG_writeback |
			1 << PG_buddy );
Linus Torvalds's avatar
Linus Torvalds committed
243
244
245
	set_page_count(page, 0);
	reset_page_mapcount(page);
	page->mapping = NULL;
246
	add_taint(TAINT_BAD_PAGE);
Linus Torvalds's avatar
Linus Torvalds committed
247
248
249
250
251
252
253
254
255
256
257
258
}

/*
 * 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).
 *
259
260
261
 * 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
262
 */
263
264
265

static void free_compound_page(struct page *page)
{
266
	__free_pages_ok(page, compound_order(page));
267
268
}

Linus Torvalds's avatar
Linus Torvalds committed
269
270
271
272
273
static void prep_compound_page(struct page *page, unsigned long order)
{
	int i;
	int nr_pages = 1 << order;

274
	set_compound_page_dtor(page, free_compound_page);
275
	set_compound_order(page, order);
276
	__SetPageHead(page);
277
	for (i = 1; i < nr_pages; i++) {
Linus Torvalds's avatar
Linus Torvalds committed
278
279
		struct page *p = page + i;

280
281
		__SetPageTail(p);
		p->first_page = page;
Linus Torvalds's avatar
Linus Torvalds committed
282
283
284
285
286
287
288
289
	}
}

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

290
	if (unlikely(compound_order(page) != order))
291
		bad_page(page);
Linus Torvalds's avatar
Linus Torvalds committed
292

293
	if (unlikely(!PageHead(page)))
294
			bad_page(page);
295
	__ClearPageHead(page);
296
	for (i = 1; i < nr_pages; i++) {
Linus Torvalds's avatar
Linus Torvalds committed
297
298
		struct page *p = page + i;

299
		if (unlikely(!PageTail(p) |
300
				(p->first_page != page)))
301
			bad_page(page);
302
		__ClearPageTail(p);
Linus Torvalds's avatar
Linus Torvalds committed
303
304
305
	}
}

Nick Piggin's avatar
Nick Piggin committed
306
307
308
309
static inline void prep_zero_page(struct page *page, int order, gfp_t gfp_flags)
{
	int i;

Nick Piggin's avatar
Nick Piggin committed
310
	VM_BUG_ON((gfp_flags & (__GFP_WAIT | __GFP_HIGHMEM)) == __GFP_HIGHMEM);
311
312
313
314
	/*
	 * 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
315
	VM_BUG_ON((gfp_flags & __GFP_HIGHMEM) && in_interrupt());
Nick Piggin's avatar
Nick Piggin committed
316
317
318
319
	for (i = 0; i < (1 << order); i++)
		clear_highpage(page + i);
}

Linus Torvalds's avatar
Linus Torvalds committed
320
321
322
323
324
/*
 * 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.
 */
325
326
static inline unsigned long page_order(struct page *page)
{
327
	return page_private(page);
Linus Torvalds's avatar
Linus Torvalds committed
328
329
}

330
331
static inline void set_page_order(struct page *page, int order)
{
332
	set_page_private(page, order);
333
	__SetPageBuddy(page);
Linus Torvalds's avatar
Linus Torvalds committed
334
335
336
337
}

static inline void rmv_page_order(struct page *page)
{
338
	__ClearPageBuddy(page);
339
	set_page_private(page, 0);
Linus Torvalds's avatar
Linus Torvalds committed
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
}

/*
 * 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
357
 * Assumption: *_mem_map is contiguous at least up to MAX_ORDER
Linus Torvalds's avatar
Linus Torvalds committed
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
 */
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
376
 * (a) the buddy is not in a hole &&
377
 * (b) the buddy is in the buddy system &&
378
379
 * (c) a page and its buddy have the same order &&
 * (d) a page and its buddy are in the same zone.
380
381
382
 *
 * 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
383
 *
384
 * For recording page's order, we use page_private(page).
Linus Torvalds's avatar
Linus Torvalds committed
385
 */
386
387
static inline int page_is_buddy(struct page *page, struct page *buddy,
								int order)
Linus Torvalds's avatar
Linus Torvalds committed
388
{
389
	if (!pfn_valid_within(page_to_pfn(buddy)))
Nick Piggin's avatar
Nick Piggin committed
390
391
		return 0;

392
393
394
395
396
	if (page_zone_id(page) != page_zone_id(buddy))
		return 0;

	if (PageBuddy(buddy) && page_order(buddy) == order) {
		BUG_ON(page_count(buddy) != 0);
397
		return 1;
398
	}
399
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
}

/*
 * 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
415
 * free pages of length of (1 << order) and marked with PG_buddy. Page's
416
 * order is recorded in page_private(page) field.
Linus Torvalds's avatar
Linus Torvalds committed
417
418
419
420
421
422
423
424
425
 * 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
426
static inline void __free_one_page(struct page *page,
Linus Torvalds's avatar
Linus Torvalds committed
427
428
429
430
		struct zone *zone, unsigned int order)
{
	unsigned long page_idx;
	int order_size = 1 << order;
431
	int migratetype = get_pageblock_migratetype(page);
Linus Torvalds's avatar
Linus Torvalds committed
432

433
	if (unlikely(PageCompound(page)))
Linus Torvalds's avatar
Linus Torvalds committed
434
435
436
437
		destroy_compound_page(page, order);

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

Nick Piggin's avatar
Nick Piggin committed
438
439
	VM_BUG_ON(page_idx & (order_size - 1));
	VM_BUG_ON(bad_range(zone, page));
Linus Torvalds's avatar
Linus Torvalds committed
440

441
	__mod_zone_page_state(zone, NR_FREE_PAGES, order_size);
Linus Torvalds's avatar
Linus Torvalds committed
442
443
444
445
446
	while (order < MAX_ORDER-1) {
		unsigned long combined_idx;
		struct page *buddy;

		buddy = __page_find_buddy(page, page_idx, order);
447
		if (!page_is_buddy(page, buddy, order))
Linus Torvalds's avatar
Linus Torvalds committed
448
			break;		/* Move the buddy up one level. */
Nick Piggin's avatar
Nick Piggin committed
449

Linus Torvalds's avatar
Linus Torvalds committed
450
		list_del(&buddy->lru);
451
		zone->free_area[order].nr_free--;
Linus Torvalds's avatar
Linus Torvalds committed
452
		rmv_page_order(buddy);
Nick Piggin's avatar
Nick Piggin committed
453
		combined_idx = __find_combined_index(page_idx, order);
Linus Torvalds's avatar
Linus Torvalds committed
454
455
456
457
458
		page = page + (combined_idx - page_idx);
		page_idx = combined_idx;
		order++;
	}
	set_page_order(page, order);
459
460
	list_add(&page->lru,
		&zone->free_area[order].free_list[migratetype]);
Linus Torvalds's avatar
Linus Torvalds committed
461
462
463
	zone->free_area[order].nr_free++;
}

464
static inline int free_pages_check(struct page *page)
Linus Torvalds's avatar
Linus Torvalds committed
465
{
Nick Piggin's avatar
Nick Piggin committed
466
467
468
	if (unlikely(page_mapcount(page) |
		(page->mapping != NULL)  |
		(page_count(page) != 0)  |
Linus Torvalds's avatar
Linus Torvalds committed
469
470
471
472
473
474
475
		(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
476
			1 << PG_writeback |
477
478
			1 << PG_reserved |
			1 << PG_buddy ))))
479
		bad_page(page);
Linus Torvalds's avatar
Linus Torvalds committed
480
	if (PageDirty(page))
Nick Piggin's avatar
Nick Piggin committed
481
		__ClearPageDirty(page);
482
483
484
485
486
487
	/*
	 * 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
488
489
490
491
492
}

/*
 * Frees a list of pages. 
 * Assumes all pages on list are in same zone, and of same order.
493
 * count is the number of pages to free.
Linus Torvalds's avatar
Linus Torvalds committed
494
495
496
497
498
499
500
 *
 * 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
501
502
static void free_pages_bulk(struct zone *zone, int count,
					struct list_head *list, int order)
Linus Torvalds's avatar
Linus Torvalds committed
503
{
Nick Piggin's avatar
Nick Piggin committed
504
	spin_lock(&zone->lock);
Linus Torvalds's avatar
Linus Torvalds committed
505
506
	zone->all_unreclaimable = 0;
	zone->pages_scanned = 0;
Nick Piggin's avatar
Nick Piggin committed
507
508
509
	while (count--) {
		struct page *page;

Nick Piggin's avatar
Nick Piggin committed
510
		VM_BUG_ON(list_empty(list));
Linus Torvalds's avatar
Linus Torvalds committed
511
		page = list_entry(list->prev, struct page, lru);
Nick Piggin's avatar
Nick Piggin committed
512
		/* have to delete it as __free_one_page list manipulates */
Linus Torvalds's avatar
Linus Torvalds committed
513
		list_del(&page->lru);
Nick Piggin's avatar
Nick Piggin committed
514
		__free_one_page(page, zone, order);
Linus Torvalds's avatar
Linus Torvalds committed
515
	}
Nick Piggin's avatar
Nick Piggin committed
516
	spin_unlock(&zone->lock);
Linus Torvalds's avatar
Linus Torvalds committed
517
518
}

Nick Piggin's avatar
Nick Piggin committed
519
static void free_one_page(struct zone *zone, struct page *page, int order)
Linus Torvalds's avatar
Linus Torvalds committed
520
{
521
522
523
	spin_lock(&zone->lock);
	zone->all_unreclaimable = 0;
	zone->pages_scanned = 0;
524
	__free_one_page(page, zone, order);
525
	spin_unlock(&zone->lock);
Nick Piggin's avatar
Nick Piggin committed
526
527
528
529
530
}

static void __free_pages_ok(struct page *page, unsigned int order)
{
	unsigned long flags;
Linus Torvalds's avatar
Linus Torvalds committed
531
	int i;
532
	int reserved = 0;
Linus Torvalds's avatar
Linus Torvalds committed
533
534

	for (i = 0 ; i < (1 << order) ; ++i)
535
		reserved += free_pages_check(page + i);
536
537
538
	if (reserved)
		return;

Nick Piggin's avatar
Nick Piggin committed
539
540
	if (!PageHighMem(page))
		debug_check_no_locks_freed(page_address(page),PAGE_SIZE<<order);
Nick Piggin's avatar
Nick Piggin committed
541
	arch_free_page(page, order);
Nick Piggin's avatar
Nick Piggin committed
542
	kernel_map_pages(page, 1 << order, 0);
Nick Piggin's avatar
Nick Piggin committed
543

Nick Piggin's avatar
Nick Piggin committed
544
	local_irq_save(flags);
545
	__count_vm_events(PGFREE, 1 << order);
Nick Piggin's avatar
Nick Piggin committed
546
	free_one_page(page_zone(page), page, order);
Nick Piggin's avatar
Nick Piggin committed
547
	local_irq_restore(flags);
Linus Torvalds's avatar
Linus Torvalds committed
548
549
}

550
551
552
553
554
555
556
557
/*
 * 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);
558
		set_page_refcounted(page);
Nick Piggin's avatar
Nick Piggin committed
559
		__free_page(page);
560
561
562
	} else {
		int loop;

Nick Piggin's avatar
Nick Piggin committed
563
		prefetchw(page);
564
565
566
		for (loop = 0; loop < BITS_PER_LONG; loop++) {
			struct page *p = &page[loop];

Nick Piggin's avatar
Nick Piggin committed
567
568
			if (loop + 1 < BITS_PER_LONG)
				prefetchw(p + 1);
569
570
571
572
			__ClearPageReserved(p);
			set_page_count(p, 0);
		}

573
		set_page_refcounted(page);
Nick Piggin's avatar
Nick Piggin committed
574
		__free_pages(page, order);
575
576
577
	}
}

Linus Torvalds's avatar
Linus Torvalds committed
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592

/*
 * 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
593
static inline void expand(struct zone *zone, struct page *page,
594
595
	int low, int high, struct free_area *area,
	int migratetype)
Linus Torvalds's avatar
Linus Torvalds committed
596
597
598
599
600
601
602
{
	unsigned long size = 1 << high;

	while (high > low) {
		area--;
		high--;
		size >>= 1;
Nick Piggin's avatar
Nick Piggin committed
603
		VM_BUG_ON(bad_range(zone, &page[size]));
604
		list_add(&page[size].lru, &area->free_list[migratetype]);
Linus Torvalds's avatar
Linus Torvalds committed
605
606
607
608
609
610
611
612
		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
613
static int prep_new_page(struct page *page, int order, gfp_t gfp_flags)
Linus Torvalds's avatar
Linus Torvalds committed
614
{
Nick Piggin's avatar
Nick Piggin committed
615
616
617
	if (unlikely(page_mapcount(page) |
		(page->mapping != NULL)  |
		(page_count(page) != 0)  |
618
619
		(page->flags & (
			1 << PG_lru	|
Linus Torvalds's avatar
Linus Torvalds committed
620
621
622
623
			1 << PG_private	|
			1 << PG_locked	|
			1 << PG_active	|
			1 << PG_dirty	|
624
			1 << PG_slab    |
Linus Torvalds's avatar
Linus Torvalds committed
625
			1 << PG_swapcache |
Nick Piggin's avatar
Nick Piggin committed
626
			1 << PG_writeback |
627
628
			1 << PG_reserved |
			1 << PG_buddy ))))
629
		bad_page(page);
Linus Torvalds's avatar
Linus Torvalds committed
630

631
632
633
634
635
636
637
	/*
	 * 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;

638
	page->flags &= ~(1 << PG_uptodate | 1 << PG_error | 1 << PG_readahead |
Linus Torvalds's avatar
Linus Torvalds committed
639
			1 << PG_referenced | 1 << PG_arch_1 |
640
			1 << PG_owner_priv_1 | 1 << PG_mappedtodisk);
641
	set_page_private(page, 0);
642
	set_page_refcounted(page);
Nick Piggin's avatar
Nick Piggin committed
643
644

	arch_alloc_page(page, order);
Linus Torvalds's avatar
Linus Torvalds committed
645
	kernel_map_pages(page, 1 << order, 1);
Nick Piggin's avatar
Nick Piggin committed
646
647
648
649
650
651
652

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

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

653
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
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
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
/*
 * 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] = {
	[MIGRATE_UNMOVABLE] = { MIGRATE_MOVABLE   },
	[MIGRATE_MOVABLE]   = { MIGRATE_UNMOVABLE },
};

/* 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--;

			/*
			 * If breaking a large block of pages, place the buddies
			 * on the preferred allocation list
			 */
			if (unlikely(current_order >= MAX_ORDER / 2))
				migratetype = start_migratetype;

			/* 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;
}

Linus Torvalds's avatar
Linus Torvalds committed
713
714
715
716
/* 
 * Do the hard work of removing an element from the buddy allocator.
 * Call me with the zone->lock already held.
 */
717
718
static struct page *__rmqueue(struct zone *zone, unsigned int order,
						int migratetype)
Linus Torvalds's avatar
Linus Torvalds committed
719
720
721
722
723
{
	struct free_area * area;
	unsigned int current_order;
	struct page *page;

724
	/* Find a page of the appropriate size in the preferred list */
Linus Torvalds's avatar
Linus Torvalds committed
725
	for (current_order = order; current_order < MAX_ORDER; ++current_order) {
726
727
		area = &(zone->free_area[current_order]);
		if (list_empty(&area->free_list[migratetype]))
Linus Torvalds's avatar
Linus Torvalds committed
728
729
			continue;

730
731
		page = list_entry(area->free_list[migratetype].next,
							struct page, lru);
Linus Torvalds's avatar
Linus Torvalds committed
732
733
734
		list_del(&page->lru);
		rmv_page_order(page);
		area->nr_free--;
735
		__mod_zone_page_state(zone, NR_FREE_PAGES, - (1UL << order));
736
737
		expand(zone, page, order, current_order, area, migratetype);
		goto got_page;
Linus Torvalds's avatar
Linus Torvalds committed
738
739
	}

740
741
742
743
744
	page = __rmqueue_fallback(zone, order, migratetype);

got_page:

	return page;
Linus Torvalds's avatar
Linus Torvalds committed
745
746
747
748
749
750
751
752
}

/* 
 * 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, 
753
754
			unsigned long count, struct list_head *list,
			int migratetype)
Linus Torvalds's avatar
Linus Torvalds committed
755
756
757
{
	int i;
	
Nick Piggin's avatar
Nick Piggin committed
758
	spin_lock(&zone->lock);
Linus Torvalds's avatar
Linus Torvalds committed
759
	for (i = 0; i < count; ++i) {
760
		struct page *page = __rmqueue(zone, order, migratetype);
Nick Piggin's avatar
Nick Piggin committed
761
		if (unlikely(page == NULL))
Linus Torvalds's avatar
Linus Torvalds committed
762
			break;
763
764
		list_add(&page->lru, list);
		set_page_private(page, migratetype);
Linus Torvalds's avatar
Linus Torvalds committed
765
	}
Nick Piggin's avatar
Nick Piggin committed
766
	spin_unlock(&zone->lock);
Nick Piggin's avatar
Nick Piggin committed
767
	return i;
Linus Torvalds's avatar
Linus Torvalds committed
768
769
}

770
#ifdef CONFIG_NUMA
771
/*
772
773
774
775
 * Called from the vmstat counter updater to drain pagesets of this
 * currently executing processor on remote nodes after they have
 * expired.
 *
776
777
 * Note that this function must be called with the thread pinned to
 * a single processor.
778
 */
779
void drain_zone_pages(struct zone *zone, struct per_cpu_pages *pcp)
780
781
{
	unsigned long flags;
782
	int to_drain;
783

784
785
786
787
788
789
790
791
	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);
792
793
794
}
#endif

Linus Torvalds's avatar
Linus Torvalds committed
795
796
static void __drain_pages(unsigned int cpu)
{
Nick Piggin's avatar
Nick Piggin committed
797
	unsigned long flags;
Linus Torvalds's avatar
Linus Torvalds committed
798
799
800
801
802
803
	struct zone *zone;
	int i;

	for_each_zone(zone) {
		struct per_cpu_pageset *pset;

804
805
806
		if (!populated_zone(zone))
			continue;

807
		pset = zone_pcp(zone, cpu);
Linus Torvalds's avatar
Linus Torvalds committed
808
809
810
811
		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
812
			local_irq_save(flags);
Nick Piggin's avatar
Nick Piggin committed
813
814
			free_pages_bulk(zone, pcp->count, &pcp->list, 0);
			pcp->count = 0;
Nick Piggin's avatar
Nick Piggin committed
815
			local_irq_restore(flags);
Linus Torvalds's avatar
Linus Torvalds committed
816
817
818
819
		}
	}
}

820
#ifdef CONFIG_HIBERNATION
Linus Torvalds's avatar
Linus Torvalds committed
821
822
823

void mark_free_pages(struct zone *zone)
{
824
825
	unsigned long pfn, max_zone_pfn;
	unsigned long flags;
826
	int order, t;
Linus Torvalds's avatar
Linus Torvalds committed
827
828
829
830
831
832
	struct list_head *curr;

	if (!zone->spanned_pages)
		return;

	spin_lock_irqsave(&zone->lock, flags);
833
834
835
836
837
838

	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);

839
840
			if (!swsusp_page_is_forbidden(page))
				swsusp_unset_page_free(page);
841
		}
Linus Torvalds's avatar
Linus Torvalds committed
842

843
844
	for_each_migratetype_order(order, t) {
		list_for_each(curr, &zone->free_area[order].free_list[t]) {
845
			unsigned long i;
Linus Torvalds's avatar
Linus Torvalds committed
846

847
848
			pfn = page_to_pfn(list_entry(curr, struct page, lru));
			for (i = 0; i < (1UL << order); i++)
849
				swsusp_set_page_free(pfn_to_page(pfn + i));
850
		}
851
	}
Linus Torvalds's avatar
Linus Torvalds committed
852
853
854
855
856
857
858
859
860
861
862
863
864
865
	spin_unlock_irqrestore(&zone->lock, flags);
}

/*
 * 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);	
}
866
#endif /* CONFIG_HIBERNATION */
Linus Torvalds's avatar
Linus Torvalds committed
867
868
869
870
871
872
873
874
875
876
877
878

/*
 * 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;
879
	if (free_pages_check(page))
880
881
		return;

Nick Piggin's avatar
Nick Piggin committed
882
883
	if (!PageHighMem(page))
		debug_check_no_locks_freed(page_address(page), PAGE_SIZE);
Nick Piggin's avatar
Nick Piggin committed
884
	arch_free_page(page, 0);
885
886
	kernel_map_pages(page, 1, 0);

887
	pcp = &zone_pcp(zone, get_cpu())->pcp[cold];
Linus Torvalds's avatar
Linus Torvalds committed
888
	local_irq_save(flags);
889
	__count_vm_event(PGFREE);
Linus Torvalds's avatar
Linus Torvalds committed
890
	list_add(&page->lru, &pcp->list);
891
	set_page_private(page, get_pageblock_migratetype(page));
Linus Torvalds's avatar
Linus Torvalds committed
892
	pcp->count++;
Nick Piggin's avatar
Nick Piggin committed
893
894
895
896
	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
897
898
899
900
901
902
903
904
905
906
907
908
909
910
	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);
}

911
912
913
914
915
916
917
918
919
920
921
922
/*
 * 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
923
924
	VM_BUG_ON(PageCompound(page));
	VM_BUG_ON(!page_count(page));
925
926
	for (i = 1; i < (1 << order); i++)
		set_page_refcounted(page + i);
927
928
}

Linus Torvalds's avatar
Linus Torvalds committed
929
930
931
932
933
/*
 * 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
934
935
static struct page *buffered_rmqueue(struct zonelist *zonelist,
			struct zone *zone, int order, gfp_t gfp_flags)
Linus Torvalds's avatar
Linus Torvalds committed
936
937
{
	unsigned long flags;
938
	struct page *page;
Linus Torvalds's avatar
Linus Torvalds committed
939
	int cold = !!(gfp_flags & __GFP_COLD);
Nick Piggin's avatar
Nick Piggin committed
940
	int cpu;
941
	int migratetype = gfpflags_to_migratetype(gfp_flags);
Linus Torvalds's avatar
Linus Torvalds committed
942

943
again:
Nick Piggin's avatar
Nick Piggin committed
944
	cpu  = get_cpu();
Nick Piggin's avatar
Nick Piggin committed
945
	if (likely(order == 0)) {
Linus Torvalds's avatar
Linus Torvalds committed
946
947
		struct per_cpu_pages *pcp;

Nick Piggin's avatar
Nick Piggin committed
948
		pcp = &zone_pcp(zone, cpu)->pcp[cold];
Linus Torvalds's avatar
Linus Torvalds committed
949
		local_irq_save(flags);
Nick Piggin's avatar
Nick Piggin committed
950
		if (!pcp->count) {
951
			pcp->count = rmqueue_bulk(zone, 0,
952
					pcp->batch, &pcp->list, migratetype);
Nick Piggin's avatar
Nick Piggin committed
953
954
			if (unlikely(!pcp->count))
				goto failed;
Linus Torvalds's avatar
Linus Torvalds committed
955
		}
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
		/* Find a page of the appropriate migrate type */
		list_for_each_entry(page, &pcp->list, lru) {
			if (page_private(page) == migratetype) {
				list_del(&page->lru);
				pcp->count--;
				break;
			}
		}

		/*
		 * Check if a page of the appropriate migrate type
		 * was found. If not, allocate more to the pcp list
		 */
		if (&page->lru == &pcp->list) {
			pcp->count += rmqueue_bulk(zone, 0,
					pcp->batch, &pcp->list, migratetype);
			page = list_entry(pcp->list.next, struct page, lru);
			VM_BUG_ON(page_private(page) != migratetype);
			list_del(&page->lru);
			pcp->count--;
		}
Rohit Seth's avatar
Rohit Seth committed
977
	} else {
Linus Torvalds's avatar
Linus Torvalds committed
978
		spin_lock_irqsave(&zone->lock, flags);
979
		page = __rmqueue(zone, order, migratetype);
Nick Piggin's avatar
Nick Piggin committed
980
981
982
		spin_unlock(&zone->lock);
		if (!page)
			goto failed;
Linus Torvalds's avatar
Linus Torvalds committed
983
984
	}

985
	__count_zone_vm_events(PGALLOC, zone, 1 << order);
986
	zone_statistics(zonelist, zone);
Nick Piggin's avatar
Nick Piggin committed
987
988
	local_irq_restore(flags);
	put_cpu();
Linus Torvalds's avatar
Linus Torvalds committed
989

Nick Piggin's avatar
Nick Piggin committed
990
	VM_BUG_ON(bad_range(zone, page));
Nick Piggin's avatar
Nick Piggin committed
991
	if (prep_new_page(page, order, gfp_flags))
Nick Piggin's avatar
Nick Piggin committed
992
		goto again;
Linus Torvalds's avatar
Linus Torvalds committed
993
	return page;
Nick Piggin's avatar
Nick Piggin committed
994
995
996
997
998

failed:
	local_irq_restore(flags);
	put_cpu();
	return NULL;
Linus Torvalds's avatar
Linus Torvalds committed
999
1000
}

Rohit Seth's avatar
Rohit Seth committed
1001
#define ALLOC_NO_WATERMARKS	0x01 /* don't check watermarks at all */
1002
1003
1004
1005
1006
1007
#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
1008

1009
1010
1011
1012
1013
1014
1015
#ifdef CONFIG_FAIL_PAGE_ALLOC

static struct fail_page_alloc_attr {
	struct fault_attr attr;

	u32 ignore_gfp_highmem;
	u32 ignore_gfp_wait;
1016
	u32 min_order;
1017
1018
1019
1020
1021

#ifdef CONFIG_FAULT_INJECTION_DEBUG_FS

	struct dentry *ignore_gfp_highmem_file;
	struct dentry *ignore_gfp_wait_file;
1022
	struct dentry *min_order_file;
1023
1024
1025
1026
1027

#endif /* CONFIG_FAULT_INJECTION_DEBUG_FS */

} fail_page_alloc = {
	.attr = FAULT_ATTR_INITIALIZER,
1028
1029
	.ignore_gfp_wait = 1,
	.ignore_gfp_highmem = 1,
1030
	.min_order = 1,
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
};

static int __init setup_fail_page_alloc(char *str)
{
	return setup_fault_attr(&fail_page_alloc.attr, str);
}
__setup("fail_page_alloc=", setup_fail_page_alloc);

static int should_fail_alloc_page(gfp_t gfp_mask, unsigned int order)
{
1041
1042
	if (order < fail_page_alloc.min_order)
		return 0;
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
	if (gfp_mask & __GFP_NOFAIL)
		return 0;
	if (fail_page_alloc.ignore_gfp_highmem && (gfp_mask & __GFP_HIGHMEM))
		return 0;
	if (fail_page_alloc.ignore_gfp_wait && (gfp_mask & __GFP_WAIT))
		return 0;

	return should_fail(&fail_page_alloc.attr, 1 << order);
}

#ifdef CONFIG_FAULT_INJECTION_DEBUG_FS

static int __init fail_page_alloc_debugfs(void)
{
	mode_t mode = S_IFREG | S_IRUSR | S_IWUSR;
	struct dentry *dir;
	int err;

	err = init_fault_attr_dentries(&fail_page_alloc.attr,
				       "fail_page_alloc");
	if (err)
		return err;
	dir = fail_page_alloc.attr.dentries.dir;

	fail_page_alloc.ignore_gfp_wait_file =
		debugfs_create_bool("ignore-gfp-wait", mode, dir,
				      &fail_page_alloc.ignore_gfp_wait);

	fail_page_alloc.ignore_gfp_highmem_file =
		debugfs_create_bool("ignore-gfp-highmem", mode, dir,
				      &fail_page_alloc.ignore_gfp_highmem);
1074
1075
1076
	fail_page_alloc.min_order_file =
		debugfs_create_u32("min-order", mode, dir,
				   &fail_page_alloc.min_order);
1077
1078

	if (!fail_page_alloc.ignore_gfp_wait_file ||
1079
1080
            !fail_page_alloc.ignore_gfp_highmem_file ||
            !fail_page_alloc.min_order_file) {
1081
1082
1083
		err = -ENOMEM;
		debugfs_remove(fail_page_alloc.ignore_gfp_wait_file);
		debugfs_remove(fail_page_alloc.ignore_gfp_highmem_file);
1084
		debugfs_remove(fail_page_alloc.min_order_file);
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
		cleanup_fault_attr_dentries(&fail_page_alloc.attr);
	}

	return err;
}

late_initcall(fail_page_alloc_debugfs);

#endif /* CONFIG_FAULT_INJECTION_DEBUG_FS */

#else /* CONFIG_FAIL_PAGE_ALLOC */

static inline int should_fail_alloc_page(gfp_t gfp_mask, unsigned int order)
{
	return 0;
}

#endif /* CONFIG_FAIL_PAGE_ALLOC */

Linus Torvalds's avatar
Linus Torvalds committed
1104
1105
1106
1107
1108
/*
 * Return 1 if free pages are above 'mark'. This takes into account the order
 * of the allocation.
 */
int zone_watermark_ok(struct zone *z, int order, unsigned long mark,
Rohit Seth's avatar
Rohit Seth committed
1109
		      int classzone_idx, int alloc_flags)
Linus Torvalds's avatar
Linus Torvalds committed
1110
1111
{
	/* free_pages my go negative - that's OK */
1112
1113
	long min = mark;
	long free_pages = zone_page_state(z, NR_FREE_PAGES) - (1 << order) + 1;
Linus Torvalds's avatar
Linus Torvalds committed
1114
1115
	int o;

Rohit Seth's avatar
Rohit Seth committed
1116
	if (alloc_flags & ALLOC_HIGH)
Linus Torvalds's avatar
Linus Torvalds committed
1117
		min -= min / 2;
Rohit Seth's avatar
Rohit Seth committed
1118
	if (alloc_flags & ALLOC_HARDER)
Linus Torvalds's avatar
Linus Torvalds committed
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
		min -= min / 4;

	if (free_pages <= min + z->lowmem_reserve[classzone_idx])
		return 0;
	for (o = 0; o < order; o++) {
		/* At the next order, this order's pages become unavailable */
		free_pages -= z->free_area[o].nr_free << o;

		/* Require fewer higher order pages to be free */
		min >>= 1;

		if (free_pages <= min)
			return 0;
	}
	return 1;
}

1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
#ifdef CONFIG_NUMA
/*
 * zlc_setup - Setup for "zonelist cache".  Uses cached zone data to
 * skip over zones that are not allowed by the cpuset, or that have
 * been recently (in last second) found to be nearly full.  See further
 * comments in mmzone.h.  Reduces cache footprint of zonelist scans
 * that have to skip over alot of full or unallowed zones.
 *
 * If the zonelist cache is present in the passed in zonelist, then
 * returns a pointer to the allowed node mask (either the current
1146
 * tasks mems_allowed, or node_states[N_HIGH_MEMORY].)
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165