page_alloc.c 71.6 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
24
/*
 *  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/config.h>
#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>
25
#include <linux/kernel.h>
Linus Torvalds's avatar
Linus Torvalds committed
26
27
28
29
30
31
32
33
34
35
#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>
36
#include <linux/memory_hotplug.h>
Linus Torvalds's avatar
Linus Torvalds committed
37
38
#include <linux/nodemask.h>
#include <linux/vmalloc.h>
39
#include <linux/mempolicy.h>
40
#include <linux/stop_machine.h>
Linus Torvalds's avatar
Linus Torvalds committed
41
42

#include <asm/tlbflush.h>
43
#include <asm/div64.h>
Linus Torvalds's avatar
Linus Torvalds committed
44
45
46
47
48
49
#include "internal.h"

/*
 * MCD - HACK: Find somewhere to initialize this EARLY, or make this
 * initializer cleaner
 */
50
nodemask_t node_online_map __read_mostly = { { [0] = 1UL } };
51
EXPORT_SYMBOL(node_online_map);
52
nodemask_t node_possible_map __read_mostly = NODE_MASK_ALL;
53
EXPORT_SYMBOL(node_possible_map);
54
55
unsigned long totalram_pages __read_mostly;
unsigned long totalhigh_pages __read_mostly;
56
unsigned long totalreserve_pages __read_mostly;
Linus Torvalds's avatar
Linus Torvalds committed
57
long nr_swap_pages;
58
int percpu_pagelist_fraction;
Linus Torvalds's avatar
Linus Torvalds committed
59

60
static void __free_pages_ok(struct page *page, unsigned int order);
61

Linus Torvalds's avatar
Linus Torvalds committed
62
63
64
65
66
67
68
/*
 * 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
69
70
71
 *
 * 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
72
 */
73
int sysctl_lowmem_reserve_ratio[MAX_NR_ZONES-1] = { 256, 256, 32 };
Linus Torvalds's avatar
Linus Torvalds committed
74
75
76
77
78
79
80

EXPORT_SYMBOL(totalram_pages);

/*
 * Used by page_zone() to look up the address of the struct zone whose
 * id is encoded in the upper bits of page->flags
 */
81
struct zone *zone_table[1 << ZONETABLE_SHIFT] __read_mostly;
Linus Torvalds's avatar
Linus Torvalds committed
82
83
EXPORT_SYMBOL(zone_table);

84
static char *zone_names[MAX_NR_ZONES] = { "DMA", "DMA32", "Normal", "HighMem" };
Linus Torvalds's avatar
Linus Torvalds committed
85
86
int min_free_kbytes = 1024;

87
88
unsigned long __meminitdata nr_kernel_pages;
unsigned long __meminitdata nr_all_pages;
Linus Torvalds's avatar
Linus Torvalds committed
89

Nick Piggin's avatar
Nick Piggin committed
90
#ifdef CONFIG_DEBUG_VM
91
static int page_outside_zone_boundaries(struct zone *zone, struct page *page)
Linus Torvalds's avatar
Linus Torvalds committed
92
{
93
94
95
	int ret = 0;
	unsigned seq;
	unsigned long pfn = page_to_pfn(page);
96

97
98
99
100
101
102
103
104
105
	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;
106
107
108
109
}

static int page_is_consistent(struct zone *zone, struct page *page)
{
Linus Torvalds's avatar
Linus Torvalds committed
110
111
#ifdef CONFIG_HOLES_IN_ZONE
	if (!pfn_valid(page_to_pfn(page)))
112
		return 0;
Linus Torvalds's avatar
Linus Torvalds committed
113
114
#endif
	if (zone != page_zone(page))
115
116
117
118
119
120
121
122
123
124
		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
125
		return 1;
126
127
128
	if (!page_is_consistent(zone, page))
		return 1;

Linus Torvalds's avatar
Linus Torvalds committed
129
130
131
	return 0;
}

Nick Piggin's avatar
Nick Piggin committed
132
133
134
135
136
137
138
#else
static inline int bad_range(struct zone *zone, struct page *page)
{
	return 0;
}
#endif

139
static void bad_page(struct page *page)
Linus Torvalds's avatar
Linus Torvalds committed
140
{
141
	printk(KERN_EMERG "Bad page state in process '%s'\n"
142
143
144
		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",
145
146
147
		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
148
	dump_stack();
149
150
	page->flags &= ~(1 << PG_lru	|
			1 << PG_private |
Linus Torvalds's avatar
Linus Torvalds committed
151
152
153
			1 << PG_locked	|
			1 << PG_active	|
			1 << PG_dirty	|
154
155
			1 << PG_reclaim |
			1 << PG_slab    |
Linus Torvalds's avatar
Linus Torvalds committed
156
			1 << PG_swapcache |
157
158
			1 << PG_writeback |
			1 << PG_buddy );
Linus Torvalds's avatar
Linus Torvalds committed
159
160
161
	set_page_count(page, 0);
	reset_page_mapcount(page);
	page->mapping = NULL;
162
	add_taint(TAINT_BAD_PAGE);
Linus Torvalds's avatar
Linus Torvalds committed
163
164
165
166
167
168
169
170
171
172
173
174
}

/*
 * 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).
 *
175
176
177
 * 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
178
 */
179
180
181
182
183
184

static void free_compound_page(struct page *page)
{
	__free_pages_ok(page, (unsigned long)page[1].lru.prev);
}

Linus Torvalds's avatar
Linus Torvalds committed
185
186
187
188
189
static void prep_compound_page(struct page *page, unsigned long order)
{
	int i;
	int nr_pages = 1 << order;

190
	page[1].lru.next = (void *)free_compound_page;	/* set dtor */
191
	page[1].lru.prev = (void *)order;
Linus Torvalds's avatar
Linus Torvalds committed
192
193
194
	for (i = 0; i < nr_pages; i++) {
		struct page *p = page + i;

195
		__SetPageCompound(p);
196
		set_page_private(p, (unsigned long)page);
Linus Torvalds's avatar
Linus Torvalds committed
197
198
199
200
201
202
203
204
	}
}

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

205
	if (unlikely((unsigned long)page[1].lru.prev != order))
206
		bad_page(page);
Linus Torvalds's avatar
Linus Torvalds committed
207
208
209
210

	for (i = 0; i < nr_pages; i++) {
		struct page *p = page + i;

211
212
213
		if (unlikely(!PageCompound(p) |
				(page_private(p) != (unsigned long)page)))
			bad_page(page);
214
		__ClearPageCompound(p);
Linus Torvalds's avatar
Linus Torvalds committed
215
216
217
	}
}

Nick Piggin's avatar
Nick Piggin committed
218
219
220
221
222
static inline void prep_zero_page(struct page *page, int order, gfp_t gfp_flags)
{
	int i;

	BUG_ON((gfp_flags & (__GFP_WAIT | __GFP_HIGHMEM)) == __GFP_HIGHMEM);
223
224
225
226
227
	/*
	 * clear_highpage() will use KM_USER0, so it's a bug to use __GFP_ZERO
	 * and __GFP_HIGHMEM from hard or soft interrupt context.
	 */
	BUG_ON((gfp_flags & __GFP_HIGHMEM) && in_interrupt());
Nick Piggin's avatar
Nick Piggin committed
228
229
230
231
	for (i = 0; i < (1 << order); i++)
		clear_highpage(page + i);
}

Linus Torvalds's avatar
Linus Torvalds committed
232
233
234
235
236
/*
 * 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.
 */
237
238
static inline unsigned long page_order(struct page *page)
{
239
	return page_private(page);
Linus Torvalds's avatar
Linus Torvalds committed
240
241
}

242
243
static inline void set_page_order(struct page *page, int order)
{
244
	set_page_private(page, order);
245
	__SetPageBuddy(page);
Linus Torvalds's avatar
Linus Torvalds committed
246
247
248
249
}

static inline void rmv_page_order(struct page *page)
{
250
	__ClearPageBuddy(page);
251
	set_page_private(page, 0);
Linus Torvalds's avatar
Linus Torvalds committed
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
}

/*
 * 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
269
 * Assumption: *_mem_map is contiguous at least up to MAX_ORDER
Linus Torvalds's avatar
Linus Torvalds committed
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
 */
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
288
 * (a) the buddy is not in a hole &&
289
 * (b) the buddy is in the buddy system &&
290
291
 * (c) a page and its buddy have the same order &&
 * (d) a page and its buddy are in the same zone.
292
293
294
 *
 * 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
295
 *
296
 * For recording page's order, we use page_private(page).
Linus Torvalds's avatar
Linus Torvalds committed
297
 */
298
299
static inline int page_is_buddy(struct page *page, struct page *buddy,
								int order)
Linus Torvalds's avatar
Linus Torvalds committed
300
{
Nick Piggin's avatar
Nick Piggin committed
301
#ifdef CONFIG_HOLES_IN_ZONE
302
	if (!pfn_valid(page_to_pfn(buddy)))
Nick Piggin's avatar
Nick Piggin committed
303
304
305
		return 0;
#endif

306
307
308
309
310
	if (page_zone_id(page) != page_zone_id(buddy))
		return 0;

	if (PageBuddy(buddy) && page_order(buddy) == order) {
		BUG_ON(page_count(buddy) != 0);
311
		return 1;
312
	}
313
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
}

/*
 * 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
329
 * free pages of length of (1 << order) and marked with PG_buddy. Page's
330
 * order is recorded in page_private(page) field.
Linus Torvalds's avatar
Linus Torvalds committed
331
332
333
334
335
336
337
338
339
 * 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
340
static inline void __free_one_page(struct page *page,
Linus Torvalds's avatar
Linus Torvalds committed
341
342
343
344
345
		struct zone *zone, unsigned int order)
{
	unsigned long page_idx;
	int order_size = 1 << order;

346
	if (unlikely(PageCompound(page)))
Linus Torvalds's avatar
Linus Torvalds committed
347
348
349
350
351
352
353
354
355
356
357
358
359
360
		destroy_compound_page(page, order);

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

	BUG_ON(page_idx & (order_size - 1));
	BUG_ON(bad_range(zone, page));

	zone->free_pages += order_size;
	while (order < MAX_ORDER-1) {
		unsigned long combined_idx;
		struct free_area *area;
		struct page *buddy;

		buddy = __page_find_buddy(page, page_idx, order);
361
		if (!page_is_buddy(page, buddy, order))
Linus Torvalds's avatar
Linus Torvalds committed
362
			break;		/* Move the buddy up one level. */
Nick Piggin's avatar
Nick Piggin committed
363

Linus Torvalds's avatar
Linus Torvalds committed
364
365
366
367
		list_del(&buddy->lru);
		area = zone->free_area + order;
		area->nr_free--;
		rmv_page_order(buddy);
Nick Piggin's avatar
Nick Piggin committed
368
		combined_idx = __find_combined_index(page_idx, order);
Linus Torvalds's avatar
Linus Torvalds committed
369
370
371
372
373
374
375
376
377
		page = page + (combined_idx - page_idx);
		page_idx = combined_idx;
		order++;
	}
	set_page_order(page, order);
	list_add(&page->lru, &zone->free_area[order].free_list);
	zone->free_area[order].nr_free++;
}

378
static inline int free_pages_check(struct page *page)
Linus Torvalds's avatar
Linus Torvalds committed
379
{
Nick Piggin's avatar
Nick Piggin committed
380
381
382
	if (unlikely(page_mapcount(page) |
		(page->mapping != NULL)  |
		(page_count(page) != 0)  |
Linus Torvalds's avatar
Linus Torvalds committed
383
384
385
386
387
388
389
390
		(page->flags & (
			1 << PG_lru	|
			1 << PG_private |
			1 << PG_locked	|
			1 << PG_active	|
			1 << PG_reclaim	|
			1 << PG_slab	|
			1 << PG_swapcache |
Nick Piggin's avatar
Nick Piggin committed
391
			1 << PG_writeback |
392
393
			1 << PG_reserved |
			1 << PG_buddy ))))
394
		bad_page(page);
Linus Torvalds's avatar
Linus Torvalds committed
395
	if (PageDirty(page))
Nick Piggin's avatar
Nick Piggin committed
396
		__ClearPageDirty(page);
397
398
399
400
401
402
	/*
	 * 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
403
404
405
406
407
}

/*
 * Frees a list of pages. 
 * Assumes all pages on list are in same zone, and of same order.
408
 * count is the number of pages to free.
Linus Torvalds's avatar
Linus Torvalds committed
409
410
411
412
413
414
415
 *
 * 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
416
417
static void free_pages_bulk(struct zone *zone, int count,
					struct list_head *list, int order)
Linus Torvalds's avatar
Linus Torvalds committed
418
{
Nick Piggin's avatar
Nick Piggin committed
419
	spin_lock(&zone->lock);
Linus Torvalds's avatar
Linus Torvalds committed
420
421
	zone->all_unreclaimable = 0;
	zone->pages_scanned = 0;
Nick Piggin's avatar
Nick Piggin committed
422
423
424
425
	while (count--) {
		struct page *page;

		BUG_ON(list_empty(list));
Linus Torvalds's avatar
Linus Torvalds committed
426
		page = list_entry(list->prev, struct page, lru);
Nick Piggin's avatar
Nick Piggin committed
427
		/* have to delete it as __free_one_page list manipulates */
Linus Torvalds's avatar
Linus Torvalds committed
428
		list_del(&page->lru);
Nick Piggin's avatar
Nick Piggin committed
429
		__free_one_page(page, zone, order);
Linus Torvalds's avatar
Linus Torvalds committed
430
	}
Nick Piggin's avatar
Nick Piggin committed
431
	spin_unlock(&zone->lock);
Linus Torvalds's avatar
Linus Torvalds committed
432
433
}

Nick Piggin's avatar
Nick Piggin committed
434
static void free_one_page(struct zone *zone, struct page *page, int order)
Linus Torvalds's avatar
Linus Torvalds committed
435
436
{
	LIST_HEAD(list);
Nick Piggin's avatar
Nick Piggin committed
437
438
439
440
441
442
443
	list_add(&page->lru, &list);
	free_pages_bulk(zone, 1, &list, order);
}

static void __free_pages_ok(struct page *page, unsigned int order)
{
	unsigned long flags;
Linus Torvalds's avatar
Linus Torvalds committed
444
	int i;
445
	int reserved = 0;
Linus Torvalds's avatar
Linus Torvalds committed
446
447

	arch_free_page(page, order);
448
449
	if (!PageHighMem(page))
		mutex_debug_check_no_locks_freed(page_address(page),
450
						 PAGE_SIZE<<order);
Linus Torvalds's avatar
Linus Torvalds committed
451
452

	for (i = 0 ; i < (1 << order) ; ++i)
453
		reserved += free_pages_check(page + i);
454
455
456
	if (reserved)
		return;

Nick Piggin's avatar
Nick Piggin committed
457
	kernel_map_pages(page, 1 << order, 0);
Nick Piggin's avatar
Nick Piggin committed
458
	local_irq_save(flags);
Nick Piggin's avatar
Nick Piggin committed
459
	__mod_page_state(pgfree, 1 << order);
Nick Piggin's avatar
Nick Piggin committed
460
	free_one_page(page_zone(page), page, order);
Nick Piggin's avatar
Nick Piggin committed
461
	local_irq_restore(flags);
Linus Torvalds's avatar
Linus Torvalds committed
462
463
}

464
465
466
467
468
469
470
471
/*
 * 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);
472
		set_page_refcounted(page);
Nick Piggin's avatar
Nick Piggin committed
473
		__free_page(page);
474
475
476
	} else {
		int loop;

Nick Piggin's avatar
Nick Piggin committed
477
		prefetchw(page);
478
479
480
		for (loop = 0; loop < BITS_PER_LONG; loop++) {
			struct page *p = &page[loop];

Nick Piggin's avatar
Nick Piggin committed
481
482
			if (loop + 1 < BITS_PER_LONG)
				prefetchw(p + 1);
483
484
485
486
			__ClearPageReserved(p);
			set_page_count(p, 0);
		}

487
		set_page_refcounted(page);
Nick Piggin's avatar
Nick Piggin committed
488
		__free_pages(page, order);
489
490
491
	}
}

Linus Torvalds's avatar
Linus Torvalds committed
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506

/*
 * 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
507
static inline void expand(struct zone *zone, struct page *page,
Linus Torvalds's avatar
Linus Torvalds committed
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
 	int low, int high, struct free_area *area)
{
	unsigned long size = 1 << high;

	while (high > low) {
		area--;
		high--;
		size >>= 1;
		BUG_ON(bad_range(zone, &page[size]));
		list_add(&page[size].lru, &area->free_list);
		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
526
static int prep_new_page(struct page *page, int order, gfp_t gfp_flags)
Linus Torvalds's avatar
Linus Torvalds committed
527
{
Nick Piggin's avatar
Nick Piggin committed
528
529
530
	if (unlikely(page_mapcount(page) |
		(page->mapping != NULL)  |
		(page_count(page) != 0)  |
531
532
		(page->flags & (
			1 << PG_lru	|
Linus Torvalds's avatar
Linus Torvalds committed
533
534
535
536
537
			1 << PG_private	|
			1 << PG_locked	|
			1 << PG_active	|
			1 << PG_dirty	|
			1 << PG_reclaim	|
538
			1 << PG_slab    |
Linus Torvalds's avatar
Linus Torvalds committed
539
			1 << PG_swapcache |
Nick Piggin's avatar
Nick Piggin committed
540
			1 << PG_writeback |
541
542
			1 << PG_reserved |
			1 << PG_buddy ))))
543
		bad_page(page);
Linus Torvalds's avatar
Linus Torvalds committed
544

545
546
547
548
549
550
551
	/*
	 * 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;

Linus Torvalds's avatar
Linus Torvalds committed
552
553
554
	page->flags &= ~(1 << PG_uptodate | 1 << PG_error |
			1 << PG_referenced | 1 << PG_arch_1 |
			1 << PG_checked | 1 << PG_mappedtodisk);
555
	set_page_private(page, 0);
556
	set_page_refcounted(page);
Linus Torvalds's avatar
Linus Torvalds committed
557
	kernel_map_pages(page, 1 << order, 1);
Nick Piggin's avatar
Nick Piggin committed
558
559
560
561
562
563
564

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

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

565
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
}

/* 
 * Do the hard work of removing an element from the buddy allocator.
 * Call me with the zone->lock already held.
 */
static struct page *__rmqueue(struct zone *zone, unsigned int order)
{
	struct free_area * area;
	unsigned int current_order;
	struct page *page;

	for (current_order = order; current_order < MAX_ORDER; ++current_order) {
		area = zone->free_area + current_order;
		if (list_empty(&area->free_list))
			continue;

		page = list_entry(area->free_list.next, struct page, lru);
		list_del(&page->lru);
		rmv_page_order(page);
		area->nr_free--;
		zone->free_pages -= 1UL << order;
Nick Piggin's avatar
Nick Piggin committed
588
589
		expand(zone, page, order, current_order, area);
		return page;
Linus Torvalds's avatar
Linus Torvalds committed
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
	}

	return NULL;
}

/* 
 * 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, 
			unsigned long count, struct list_head *list)
{
	int i;
	
Nick Piggin's avatar
Nick Piggin committed
605
	spin_lock(&zone->lock);
Linus Torvalds's avatar
Linus Torvalds committed
606
	for (i = 0; i < count; ++i) {
Nick Piggin's avatar
Nick Piggin committed
607
608
		struct page *page = __rmqueue(zone, order);
		if (unlikely(page == NULL))
Linus Torvalds's avatar
Linus Torvalds committed
609
610
611
			break;
		list_add_tail(&page->lru, list);
	}
Nick Piggin's avatar
Nick Piggin committed
612
	spin_unlock(&zone->lock);
Nick Piggin's avatar
Nick Piggin committed
613
	return i;
Linus Torvalds's avatar
Linus Torvalds committed
614
615
}

616
#ifdef CONFIG_NUMA
617
618
619
/*
 * Called from the slab reaper to drain pagesets on a particular node that
 * belong to the currently executing processor.
620
621
 * Note that this function must be called with the thread pinned to
 * a single processor.
622
623
 */
void drain_node_pages(int nodeid)
624
{
625
	int i, z;
626
627
	unsigned long flags;

628
629
	for (z = 0; z < MAX_NR_ZONES; z++) {
		struct zone *zone = NODE_DATA(nodeid)->node_zones + z;
630
631
		struct per_cpu_pageset *pset;

Nick Piggin's avatar
Nick Piggin committed
632
		pset = zone_pcp(zone, smp_processor_id());
633
634
635
636
		for (i = 0; i < ARRAY_SIZE(pset->pcp); i++) {
			struct per_cpu_pages *pcp;

			pcp = &pset->pcp[i];
637
638
639
640
641
642
			if (pcp->count) {
				local_irq_save(flags);
				free_pages_bulk(zone, pcp->count, &pcp->list, 0);
				pcp->count = 0;
				local_irq_restore(flags);
			}
643
644
645
646
647
		}
	}
}
#endif

Linus Torvalds's avatar
Linus Torvalds committed
648
649
650
#if defined(CONFIG_PM) || defined(CONFIG_HOTPLUG_CPU)
static void __drain_pages(unsigned int cpu)
{
Nick Piggin's avatar
Nick Piggin committed
651
	unsigned long flags;
Linus Torvalds's avatar
Linus Torvalds committed
652
653
654
655
656
657
	struct zone *zone;
	int i;

	for_each_zone(zone) {
		struct per_cpu_pageset *pset;

658
		pset = zone_pcp(zone, cpu);
Linus Torvalds's avatar
Linus Torvalds committed
659
660
661
662
		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
663
			local_irq_save(flags);
Nick Piggin's avatar
Nick Piggin committed
664
665
			free_pages_bulk(zone, pcp->count, &pcp->list, 0);
			pcp->count = 0;
Nick Piggin's avatar
Nick Piggin committed
666
			local_irq_restore(flags);
Linus Torvalds's avatar
Linus Torvalds committed
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
		}
	}
}
#endif /* CONFIG_PM || CONFIG_HOTPLUG_CPU */

#ifdef CONFIG_PM

void mark_free_pages(struct zone *zone)
{
	unsigned long zone_pfn, flags;
	int order;
	struct list_head *curr;

	if (!zone->spanned_pages)
		return;

	spin_lock_irqsave(&zone->lock, flags);
	for (zone_pfn = 0; zone_pfn < zone->spanned_pages; ++zone_pfn)
		ClearPageNosaveFree(pfn_to_page(zone_pfn + zone->zone_start_pfn));

	for (order = MAX_ORDER - 1; order >= 0; --order)
		list_for_each(curr, &zone->free_area[order].free_list) {
			unsigned long start_pfn, i;

			start_pfn = page_to_pfn(list_entry(curr, struct page, lru));

			for (i=0; i < (1<<order); i++)
				SetPageNosaveFree(pfn_to_page(start_pfn+i));
	}
	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);	
}
#endif /* CONFIG_PM */

Nick Piggin's avatar
Nick Piggin committed
712
static void zone_statistics(struct zonelist *zonelist, struct zone *z, int cpu)
Linus Torvalds's avatar
Linus Torvalds committed
713
714
715
716
717
718
{
#ifdef CONFIG_NUMA
	pg_data_t *pg = z->zone_pgdat;
	pg_data_t *orig = zonelist->zones[0]->zone_pgdat;
	struct per_cpu_pageset *p;

Nick Piggin's avatar
Nick Piggin committed
719
	p = zone_pcp(z, cpu);
Linus Torvalds's avatar
Linus Torvalds committed
720
	if (pg == orig) {
721
		p->numa_hit++;
Linus Torvalds's avatar
Linus Torvalds committed
722
723
	} else {
		p->numa_miss++;
724
		zone_pcp(zonelist->zones[0], cpu)->numa_foreign++;
Linus Torvalds's avatar
Linus Torvalds committed
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
	}
	if (pg == NODE_DATA(numa_node_id()))
		p->local_node++;
	else
		p->other_node++;
#endif
}

/*
 * 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;

	arch_free_page(page, 0);

	if (PageAnon(page))
		page->mapping = NULL;
746
	if (free_pages_check(page))
747
748
749
750
		return;

	kernel_map_pages(page, 1, 0);

751
	pcp = &zone_pcp(zone, get_cpu())->pcp[cold];
Linus Torvalds's avatar
Linus Torvalds committed
752
	local_irq_save(flags);
Nick Piggin's avatar
Nick Piggin committed
753
	__inc_page_state(pgfree);
Linus Torvalds's avatar
Linus Torvalds committed
754
755
	list_add(&page->lru, &pcp->list);
	pcp->count++;
Nick Piggin's avatar
Nick Piggin committed
756
757
758
759
	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
760
761
762
763
764
765
766
767
768
769
770
771
772
773
	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);
}

774
775
776
777
778
779
780
781
782
783
784
785
786
787
/*
 * 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;

	BUG_ON(PageCompound(page));
	BUG_ON(!page_count(page));
788
789
	for (i = 1; i < (1 << order); i++)
		set_page_refcounted(page + i);
790
791
}

Linus Torvalds's avatar
Linus Torvalds committed
792
793
794
795
796
/*
 * 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
797
798
static struct page *buffered_rmqueue(struct zonelist *zonelist,
			struct zone *zone, int order, gfp_t gfp_flags)
Linus Torvalds's avatar
Linus Torvalds committed
799
800
{
	unsigned long flags;
801
	struct page *page;
Linus Torvalds's avatar
Linus Torvalds committed
802
	int cold = !!(gfp_flags & __GFP_COLD);
Nick Piggin's avatar
Nick Piggin committed
803
	int cpu;
Linus Torvalds's avatar
Linus Torvalds committed
804

805
again:
Nick Piggin's avatar
Nick Piggin committed
806
	cpu  = get_cpu();
Nick Piggin's avatar
Nick Piggin committed
807
	if (likely(order == 0)) {
Linus Torvalds's avatar
Linus Torvalds committed
808
809
		struct per_cpu_pages *pcp;

Nick Piggin's avatar
Nick Piggin committed
810
		pcp = &zone_pcp(zone, cpu)->pcp[cold];
Linus Torvalds's avatar
Linus Torvalds committed
811
		local_irq_save(flags);
Nick Piggin's avatar
Nick Piggin committed
812
		if (!pcp->count) {
Linus Torvalds's avatar
Linus Torvalds committed
813
814
			pcp->count += rmqueue_bulk(zone, 0,
						pcp->batch, &pcp->list);
Nick Piggin's avatar
Nick Piggin committed
815
816
			if (unlikely(!pcp->count))
				goto failed;
Linus Torvalds's avatar
Linus Torvalds committed
817
		}
Nick Piggin's avatar
Nick Piggin committed
818
819
820
		page = list_entry(pcp->list.next, struct page, lru);
		list_del(&page->lru);
		pcp->count--;
Rohit Seth's avatar
Rohit Seth committed
821
	} else {
Linus Torvalds's avatar
Linus Torvalds committed
822
823
		spin_lock_irqsave(&zone->lock, flags);
		page = __rmqueue(zone, order);
Nick Piggin's avatar
Nick Piggin committed
824
825
826
		spin_unlock(&zone->lock);
		if (!page)
			goto failed;
Linus Torvalds's avatar
Linus Torvalds committed
827
828
	}

Nick Piggin's avatar
Nick Piggin committed
829
830
831
832
	__mod_page_state_zone(zone, pgalloc, 1 << order);
	zone_statistics(zonelist, zone, cpu);
	local_irq_restore(flags);
	put_cpu();
Linus Torvalds's avatar
Linus Torvalds committed
833

Nick Piggin's avatar
Nick Piggin committed
834
	BUG_ON(bad_range(zone, page));
Nick Piggin's avatar
Nick Piggin committed
835
	if (prep_new_page(page, order, gfp_flags))
Nick Piggin's avatar
Nick Piggin committed
836
		goto again;
Linus Torvalds's avatar
Linus Torvalds committed
837
	return page;
Nick Piggin's avatar
Nick Piggin committed
838
839
840
841
842

failed:
	local_irq_restore(flags);
	put_cpu();
	return NULL;
Linus Torvalds's avatar
Linus Torvalds committed
843
844
}

Rohit Seth's avatar
Rohit Seth committed
845
#define ALLOC_NO_WATERMARKS	0x01 /* don't check watermarks at all */
846
847
848
849
850
851
#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
852

Linus Torvalds's avatar
Linus Torvalds committed
853
854
855
856
857
/*
 * 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
858
		      int classzone_idx, int alloc_flags)
Linus Torvalds's avatar
Linus Torvalds committed
859
860
861
862
863
{
	/* free_pages my go negative - that's OK */
	long min = mark, free_pages = z->free_pages - (1 << order) + 1;
	int o;

Rohit Seth's avatar
Rohit Seth committed
864
	if (alloc_flags & ALLOC_HIGH)
Linus Torvalds's avatar
Linus Torvalds committed
865
		min -= min / 2;
Rohit Seth's avatar
Rohit Seth committed
866
	if (alloc_flags & ALLOC_HARDER)
Linus Torvalds's avatar
Linus Torvalds committed
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
		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;
}

Rohit Seth's avatar
Rohit Seth committed
884
885
886
887
888
889
890
/*
 * get_page_from_freeliest goes through the zonelist trying to allocate
 * a page.
 */
static struct page *
get_page_from_freelist(gfp_t gfp_mask, unsigned int order,
		struct zonelist *zonelist, int alloc_flags)
Martin Hicks's avatar
Martin Hicks committed
891
{
Rohit Seth's avatar
Rohit Seth committed
892
893
894
895
896
897
898
899
900
901
902
903
904
905
	struct zone **z = zonelist->zones;
	struct page *page = NULL;
	int classzone_idx = zone_idx(*z);

	/*
	 * Go through the zonelist once, looking for a zone with enough free.
	 * See also cpuset_zone_allowed() comment in kernel/cpuset.c.
	 */
	do {
		if ((alloc_flags & ALLOC_CPUSET) &&
				!cpuset_zone_allowed(*z, gfp_mask))
			continue;

		if (!(alloc_flags & ALLOC_NO_WATERMARKS)) {
906
907
908
909
910
911
912
913
			unsigned long mark;
			if (alloc_flags & ALLOC_WMARK_MIN)
				mark = (*z)->pages_min;
			else if (alloc_flags & ALLOC_WMARK_LOW)
				mark = (*z)->pages_low;
			else
				mark = (*z)->pages_high;
			if (!zone_watermark_ok(*z, order, mark,
Rohit Seth's avatar
Rohit Seth committed
914
				    classzone_idx, alloc_flags))
915
916
917
				if (!zone_reclaim_mode ||
				    !zone_reclaim(*z, gfp_mask, order))
					continue;
Rohit Seth's avatar
Rohit Seth committed
918
919
		}

Nick Piggin's avatar
Nick Piggin committed
920
		page = buffered_rmqueue(zonelist, *z, order, gfp_mask);
Rohit Seth's avatar
Rohit Seth committed
921
922
923
924
925
		if (page) {
			break;
		}
	} while (*(++z) != NULL);
	return page;
Martin Hicks's avatar
Martin Hicks committed
926
927
}

Linus Torvalds's avatar
Linus Torvalds committed
928
929
930
931
/*
 * This is the 'heart' of the zoned buddy allocator.
 */
struct page * fastcall
932
__alloc_pages(gfp_t gfp_mask, unsigned int order,
Linus Torvalds's avatar
Linus Torvalds committed
933
934
		struct zonelist *zonelist)
{
Al Viro's avatar
Al Viro committed
935
	const gfp_t wait = gfp_mask & __GFP_WAIT;
Rohit Seth's avatar
Rohit Seth committed
936
	struct zone **z;
Linus Torvalds's avatar
Linus Torvalds committed
937
938
939
940
	struct page *page;
	struct reclaim_state reclaim_state;
	struct task_struct *p = current;
	int do_retry;
Rohit Seth's avatar
Rohit Seth committed
941
	int alloc_flags;
Linus Torvalds's avatar
Linus Torvalds committed
942
943
944
945
	int did_some_progress;

	might_sleep_if(wait);

946
restart:
Rohit Seth's avatar
Rohit Seth committed
947
	z = zonelist->zones;  /* the list of zones suitable for gfp_mask */
Linus Torvalds's avatar
Linus Torvalds committed
948

Rohit Seth's avatar
Rohit Seth committed
949
	if (unlikely(*z == NULL)) {
Linus Torvalds's avatar
Linus Torvalds committed
950
951
952
		/* Should this ever happen?? */
		return NULL;
	}
953

Rohit Seth's avatar
Rohit Seth committed
954
	page = get_page_from_freelist(gfp_mask|__GFP_HARDWALL, order,
955
				zonelist, ALLOC_WMARK_LOW|ALLOC_CPUSET);
Rohit Seth's avatar
Rohit Seth committed
956
957
	if (page)
		goto got_pg;
Linus Torvalds's avatar
Linus Torvalds committed
958

959
	do {
960
		wakeup_kswapd(*z, order);
961
	} while (*(++z));
Linus Torvalds's avatar
Linus Torvalds committed
962

963
	/*
Rohit Seth's avatar
Rohit Seth committed
964
965
966
967
968
969
	 * OK, we're below the kswapd watermark and have kicked background
	 * reclaim. Now things get more complex, so set up alloc_flags according
	 * to how we want to proceed.
	 *
	 * The caller may dip into page reserves a bit more if the caller
	 * cannot run direct reclaim, or if the caller has realtime scheduling
Paul Jackson's avatar
Paul Jackson committed
970
971
	 * policy or is asking for __GFP_HIGH memory.  GFP_ATOMIC requests will
	 * set both ALLOC_HARDER (!wait) and ALLOC_HIGH (__GFP_HIGH).
972
	 */
973
	alloc_flags = ALLOC_WMARK_MIN;
Rohit Seth's avatar
Rohit Seth committed
974
975
976
977
	if ((unlikely(rt_task(p)) && !in_interrupt()) || !wait)
		alloc_flags |= ALLOC_HARDER;
	if (gfp_mask & __GFP_HIGH)
		alloc_flags |= ALLOC_HIGH;
978
979
	if (wait)
		alloc_flags |= ALLOC_CPUSET;
Linus Torvalds's avatar
Linus Torvalds committed
980
981
982

	/*
	 * Go through the zonelist again. Let __GFP_HIGH and allocations
Rohit Seth's avatar
Rohit Seth committed
983
	 * coming from realtime tasks go deeper into reserves.
Linus Torvalds's avatar
Linus Torvalds committed
984
985
986
	 *
	 * This is the last chance, in general, before the goto nopage.
	 * Ignore cpuset if GFP_ATOMIC (!wait) rather than fail alloc.
987
	 * See also cpuset_zone_allowed() comment in kernel/cpuset.c.
Linus Torvalds's avatar
Linus Torvalds committed
988
	 */
Rohit Seth's avatar
Rohit Seth committed
989
990
991
	page = get_page_from_freelist(gfp_mask, order, zonelist, alloc_flags);
	if (page)
		goto got_pg;
Linus Torvalds's avatar
Linus Torvalds committed
992
993

	/* This allocation should allow future memory freeing. */
994
995
996
997

	if (((p->flags & PF_MEMALLOC) || unlikely(test_thread_flag(TIF_MEMDIE)))
			&& !in_interrupt()) {
		if (!(gfp_mask & __GFP_NOMEMALLOC)) {
Kirill Korotaev's avatar
Kirill Korotaev committed
998
nofail_alloc:
999
			/* go through the zonelist yet again, ignoring mins */
Rohit Seth's avatar
Rohit Seth committed
1000
			page = get_page_from_freelist(gfp_mask, order,
1001
				zonelist, ALLOC_NO_WATERMARKS);
Rohit Seth's avatar
Rohit Seth committed
1002
1003
			if (page)
				goto got_pg;
Kirill Korotaev's avatar
Kirill Korotaev committed
1004
1005
1006
1007
			if (gfp_mask & __GFP_NOFAIL) {
				blk_congestion_wait(WRITE, HZ/50);
				goto nofail_alloc;
			}
Linus Torvalds's avatar
Linus Torvalds committed
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
		}
		goto nopage;
	}

	/* Atomic allocations - we can't balance anything */
	if (!wait)
		goto nopage;

rebalance:
	cond_resched();

	/* We now go into synchronous reclaim */
1020
	cpuset_memory_pressure_bump();
Linus Torvalds's avatar
Linus Torvalds committed
1021
1022
1023
1024
	p->flags |= PF_MEMALLOC;
	reclaim_state.reclaimed_slab = 0;
	p->reclaim_state = &reclaim_state;

Rohit Seth's avatar
Rohit Seth committed
1025
	did_some_progress = try_to_free_pages(zonelist->zones, gfp_mask);
Linus Torvalds's avatar
Linus Torvalds committed
1026
1027
1028
1029
1030
1031
1032

	p->reclaim_state = NULL;
	p->flags &= ~PF_MEMALLOC;

	cond_resched();

	if (likely(did_some_progress)) {
Rohit Seth's avatar
Rohit Seth committed
1033
1034
1035
1036
		page = get_page_from_freelist(gfp_mask, order,
						zonelist, alloc_flags);
		if (page)
			goto got_pg;
Linus Torvalds's avatar
Linus Torvalds committed
1037
1038
1039
1040
1041
1042
1043
	} else if ((gfp_mask & __GFP_FS) && !(gfp_mask & __GFP_NORETRY)) {
		/*
		 * Go through the zonelist yet one more time, keep
		 * very high watermark here, this is only to catch
		 * a parallel oom killing, we must fail if we're still
		 * under heavy pressure.
		 */
Rohit Seth's avatar
Rohit Seth committed
1044
		page = get_page_from_freelist(gfp_mask|__GFP_HARDWALL, order,
1045
				zonelist, ALLOC_WMARK_HIGH|ALLOC_CPUSET);
Rohit Seth's avatar
Rohit Seth committed
1046
1047
		if (page)
			goto got_pg;
Linus Torvalds's avatar
Linus Torvalds committed
1048

1049
		out_of_memory(zonelist, gfp_mask, order);
Linus Torvalds's avatar
Linus Torvalds committed
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
		goto restart;
	}

	/*
	 * Don't let big-order allocations loop unless the caller explicitly
	 * requests that.  Wait for some write requests to complete then retry.
	 *
	 * In this implementation, __GFP_REPEAT means __GFP_NOFAIL for order
	 * <= 3, but that may not be true in other implementations.
	 */
	do_retry = 0;
	if (!(gfp_mask & __GFP_NORETRY)) {
		if ((order <= 3) || (gfp_mask & __GFP_REPEAT))
			do_retry = 1;
		if (gfp_mask & __GFP_NOFAIL)
			do_retry = 1;
	}
	if (do_retry) {
		blk_congestion_wait(WRITE, HZ/50);
		goto rebalance;
	}

nopage:
	if (!(gfp_mask & __GFP_NOWARN) && printk_ratelimit()) {
		printk(KERN_WARNING "%s: page allocation failure."
			" order:%d, mode:0x%x\n",
			p->comm, order, gfp_mask);
		dump_stack();
Janet Morgan's avatar
Janet Morgan committed
1078
		show_mem();
Linus Torvalds's avatar
Linus Torvalds committed
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
	}
got_pg:
	return page;
}

EXPORT_SYMBOL(__alloc_pages);

/*
 * Common helper functions.
 */
1089
fastcall unsigned long __get_free_pages(gfp_t gfp_mask, unsigned int order)
Linus Torvalds's avatar
Linus Torvalds committed
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
{
	struct page * page;
	page = alloc_pages(gfp_mask, order);
	if (!page)
		return 0;
	return (unsigned long) page_address(page);
}

EXPORT_SYMBOL(__get_free_pages);

1100
fastcall unsigned long get_zeroed_page(gfp_t gfp_mask)
Linus Torvalds's avatar
Linus Torvalds committed
1101
1102
1103
1104
1105
1106
1107
{
	struct page * page;

	/*
	 * get_zeroed_page() returns a 32-bit address, which cannot represent
	 * a highmem page
	 */
Al Viro's avatar
Al Viro committed
1108
	BUG_ON((gfp_mask & __GFP_HIGHMEM) != 0);
Linus Torvalds's avatar
Linus Torvalds committed
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127

	page = alloc_pages(gfp_mask | __GFP_ZERO, 0);
	if (page)
		return (unsigned long) page_address(page);
	return 0;
}

EXPORT_SYMBOL(get_zeroed_page);

void __pagevec_free(struct pagevec *pvec)
{
	int i = pagevec_count(pvec);

	while (--i >= 0)
		free_hot_cold_page(pvec->pages[i], pvec->cold);
}

fastcall void __free_pages(struct page *page, unsigned int order)
{
Nick Piggin's avatar
Nick Piggin committed
1128
	if (put_page_testzero(page)) {
Linus Torvalds's avatar
Linus Torvalds committed
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
		if (order == 0)
			free_hot_page(page);
		else
			__free_pages_ok(page, order);
	}
}

EXPORT_SYMBOL(__free_pages);

fastcall void free_pages(unsigned long addr, unsigned int order)
{
	if (addr != 0) {
		BUG_ON(!virt_addr_valid((void *)addr));
		__free_pages(virt_to_page((void *)addr), order);
	}
}

EXPORT_SYMBOL(free_pages);

/*
 * Total amount of free (allocatable) RAM:
 */
unsigned int nr_free_pages(void)
{
	unsigned int sum = 0;
	struct zone *zone;

	for_each_zone(zone)
		sum += zone->free_pages;

	return sum;
}

EXPORT_SYMBOL(nr_free_pages);

#ifdef CONFIG_NUMA
unsigned int nr_free_pages_pgdat(pg_data_t *pgdat)
{
	unsigned int i, sum = 0;

	for (i = 0; i < MAX_NR_ZONES; i++)
		sum += pgdat->node_zones[i].free_pages;

	return sum;
}
#endif

static unsigned int nr_free_zone_pages(int offset)
{
1178
1179
	/* Just pick one node, since fallback list is circular */
	pg_data_t *pgdat = NODE_DATA(numa_node_id());
Linus Torvalds's avatar
Linus Torvalds committed
1180
1181
	unsigned int sum = 0;

1182
1183
1184
	struct zonelist *zonelist = pgdat->node_zonelists + offset;
	struct zone **zonep = zonelist->zones;
	struct zone *zone;
Linus Torvalds's avatar
Linus Torvalds committed
1185

1186
1187
1188
1189
1190
	for (zone = *zonep++; zone; zone = *zonep++) {
		unsigned long size = zone->present_pages;
		unsigned long high = zone->pages_high;
		if (size > high)
			sum += size - high;
Linus Torvalds's avatar
Linus Torvalds committed
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
	}

	return sum;
}

/*
 * Amount of free RAM allocatable within ZONE_DMA and ZONE_NORMAL
 */
unsigned int nr_free_buffer_pages(void)
{
Al Viro's avatar
Al Viro committed
1201
	return nr_free_zone_pages(gfp_zone(GFP_USER));
Linus Torvalds's avatar
Linus Torvalds committed
1202
1203
1204
1205
1206
1207
1208
}

/*
 * Amount of free RAM allocatable within all zones
 */
unsigned int nr_free_pagecache_pages(void)
{
Al Viro's avatar
Al Viro committed
1209
	return nr_free_zone_pages(gfp_zone(GFP_HIGHUSER));
Linus Torvalds's avatar
Linus Torvalds committed
1210
1211
1212
1213
1214
1215
1216
1217
}

#ifdef CONFIG_HIGHMEM
unsigned int nr_free_highpages (void)
{
	pg_data_t *pgdat;
	unsigned int pages = 0;

1218
	for_each_online_pgdat(pgdat)
Linus Torvalds's avatar
Linus Torvalds committed
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
		pages += pgdat->node_zones[ZONE_HIGHMEM].free_pages;

	return pages;
}
#endif

#ifdef CONFIG_NUMA
static void show_node(struct zone *zone)
{
	printk("Node %d ", zone->zone_pgdat->node_id);
}
#else
#define show_node(zone)	do { } while (0)
#endif

/*
 * Accumulate the page_state information across all CPUs.
 * The result is unavoidably approximate - it can change
 * during and after execution of this function.
 */
static DEFINE_PER_CPU(struct page_state, page_states) = {0};

atomic_t nr_pagecache = ATOMIC_INIT(0);
EXPORT_SYMBOL(nr_pagecache);
#ifdef CONFIG_SMP
DEFINE_PER_CPU(long, nr_pagecache_local) = 0;
#endif

Nick Piggin's avatar
Nick Piggin committed
1247
static void __get_page_state(struct page_state *ret, int nr, cpumask_t *cpumask)
Linus Torvalds's avatar
Linus Torvalds committed
1248
{
1249
	unsigned cpu;
Linus Torvalds's avatar
Linus Torvalds committed
1250

1251
	memset(ret, 0, nr * sizeof(unsigned long));
1252
	cpus_and(*cpumask, *cpumask, cpu_online_map);
Linus Torvalds's avatar
Linus Torvalds committed
1253

1254
1255
1256
1257
1258
	for_each_cpu_mask(cpu, *cpumask) {
		unsigned long *in;
		unsigned long *out;
		unsigned off;
		unsigned next_cpu;
1259

Linus Torvalds's avatar
Linus Torvalds committed
1260
1261
		in = (unsigned long *)&per_cpu(page_states, cpu);

1262
1263
1264
		next_cpu = next_cpu(cpu, *cpumask);
		if (likely(next_cpu < NR_CPUS))
			prefetch(&per_cpu(page_states, next_cpu));
Linus Torvalds's avatar
Linus Torvalds committed
1265
1266
1267
1268
1269
1270
1271

		out = (unsigned long *)ret;
		for (off = 0; off < nr; off++)
			*out++ += *in++;
	}
}

1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
void get_page_state_node(struct page_state *ret, int node)
{
	int nr;
	cpumask_t mask = node_to_cpumask(node);

	nr = offsetof(struct page_state, GET_PAGE_STATE_LAST);
	nr /= sizeof(unsigned long);

	__get_page_state(ret, nr+1, &mask);
}

Linus Torvalds's avatar
Linus Torvalds committed
1283
1284
1285
void get_page_state(struct page_state *ret)
{
	int nr;
1286
	cpumask_t mask = CPU_MASK_ALL;
Linus Torvalds's avatar
Linus Torvalds committed
1287
1288
1289
1290

	nr = offsetof(struct page_state, GET_PAGE_STATE_LAST);
	nr /= sizeof(unsigned long);

1291
	__get_page_state(ret, nr + 1, &mask);
Linus Torvalds's avatar
Linus Torvalds committed
1292
1293
1294
1295
}

void get_full_page_state(struct page_state *ret)
{
1296
1297
1298
	cpumask_t mask = CPU_MASK_ALL;