page_alloc.c 68.8 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>
Linus Torvalds's avatar
Linus Torvalds committed
40 41 42 43 44 45 46 47

#include <asm/tlbflush.h>
#include "internal.h"

/*
 * MCD - HACK: Find somewhere to initialize this EARLY, or make this
 * initializer cleaner
 */
48
nodemask_t node_online_map __read_mostly = { { [0] = 1UL } };
49
EXPORT_SYMBOL(node_online_map);
50
nodemask_t node_possible_map __read_mostly = NODE_MASK_ALL;
51
EXPORT_SYMBOL(node_possible_map);
52 53
unsigned long totalram_pages __read_mostly;
unsigned long totalhigh_pages __read_mostly;
Linus Torvalds's avatar
Linus Torvalds committed
54
long nr_swap_pages;
55
int percpu_pagelist_fraction;
Linus Torvalds's avatar
Linus Torvalds committed
56

57
static void __free_pages_ok(struct page *page, unsigned int order);
58

Linus Torvalds's avatar
Linus Torvalds committed
59 60 61 62 63 64 65
/*
 * 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
66 67 68
 *
 * 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
69
 */
70
int sysctl_lowmem_reserve_ratio[MAX_NR_ZONES-1] = { 256, 256, 32 };
Linus Torvalds's avatar
Linus Torvalds committed
71 72 73 74 75 76 77

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
 */
78
struct zone *zone_table[1 << ZONETABLE_SHIFT] __read_mostly;
Linus Torvalds's avatar
Linus Torvalds committed
79 80
EXPORT_SYMBOL(zone_table);

81
static char *zone_names[MAX_NR_ZONES] = { "DMA", "DMA32", "Normal", "HighMem" };
Linus Torvalds's avatar
Linus Torvalds committed
82 83 84 85 86
int min_free_kbytes = 1024;

unsigned long __initdata nr_kernel_pages;
unsigned long __initdata nr_all_pages;

Nick Piggin's avatar
Nick Piggin committed
87
#ifdef CONFIG_DEBUG_VM
88
static int page_outside_zone_boundaries(struct zone *zone, struct page *page)
Linus Torvalds's avatar
Linus Torvalds committed
89
{
90 91 92
	int ret = 0;
	unsigned seq;
	unsigned long pfn = page_to_pfn(page);
93

94 95 96 97 98 99 100 101 102
	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;
103 104 105 106
}

static int page_is_consistent(struct zone *zone, struct page *page)
{
Linus Torvalds's avatar
Linus Torvalds committed
107 108
#ifdef CONFIG_HOLES_IN_ZONE
	if (!pfn_valid(page_to_pfn(page)))
109
		return 0;
Linus Torvalds's avatar
Linus Torvalds committed
110 111
#endif
	if (zone != page_zone(page))
112 113 114 115 116 117 118 119 120 121
		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
122
		return 1;
123 124 125
	if (!page_is_consistent(zone, page))
		return 1;

Linus Torvalds's avatar
Linus Torvalds committed
126 127 128
	return 0;
}

Nick Piggin's avatar
Nick Piggin committed
129 130 131 132 133 134 135
#else
static inline int bad_range(struct zone *zone, struct page *page)
{
	return 0;
}
#endif

136
static void bad_page(struct page *page)
Linus Torvalds's avatar
Linus Torvalds committed
137
{
138
	printk(KERN_EMERG "Bad page state in process '%s'\n"
139 140 141
		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",
142 143 144
		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
145
	dump_stack();
146 147
	page->flags &= ~(1 << PG_lru	|
			1 << PG_private |
Linus Torvalds's avatar
Linus Torvalds committed
148 149 150
			1 << PG_locked	|
			1 << PG_active	|
			1 << PG_dirty	|
151 152
			1 << PG_reclaim |
			1 << PG_slab    |
Linus Torvalds's avatar
Linus Torvalds committed
153
			1 << PG_swapcache |
154 155
			1 << PG_writeback |
			1 << PG_buddy );
Linus Torvalds's avatar
Linus Torvalds committed
156 157 158
	set_page_count(page, 0);
	reset_page_mapcount(page);
	page->mapping = NULL;
159
	add_taint(TAINT_BAD_PAGE);
Linus Torvalds's avatar
Linus Torvalds committed
160 161 162 163 164 165 166 167 168 169 170 171
}

/*
 * 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).
 *
172 173 174
 * 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
175
 */
176 177 178 179 180 181

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
182 183 184 185 186
static void prep_compound_page(struct page *page, unsigned long order)
{
	int i;
	int nr_pages = 1 << order;

187
	page[1].lru.next = (void *)free_compound_page;	/* set dtor */
188
	page[1].lru.prev = (void *)order;
Linus Torvalds's avatar
Linus Torvalds committed
189 190 191
	for (i = 0; i < nr_pages; i++) {
		struct page *p = page + i;

192
		__SetPageCompound(p);
193
		set_page_private(p, (unsigned long)page);
Linus Torvalds's avatar
Linus Torvalds committed
194 195 196 197 198 199 200 201
	}
}

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

202
	if (unlikely((unsigned long)page[1].lru.prev != order))
203
		bad_page(page);
Linus Torvalds's avatar
Linus Torvalds committed
204 205 206 207

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

208 209 210
		if (unlikely(!PageCompound(p) |
				(page_private(p) != (unsigned long)page)))
			bad_page(page);
211
		__ClearPageCompound(p);
Linus Torvalds's avatar
Linus Torvalds committed
212 213 214
	}
}

Nick Piggin's avatar
Nick Piggin committed
215 216 217 218 219
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);
220 221 222 223 224
	/*
	 * 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
225 226 227 228
	for (i = 0; i < (1 << order); i++)
		clear_highpage(page + i);
}

Linus Torvalds's avatar
Linus Torvalds committed
229 230 231 232 233 234
/*
 * 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.
 */
static inline unsigned long page_order(struct page *page) {
235
	return page_private(page);
Linus Torvalds's avatar
Linus Torvalds committed
236 237 238
}

static inline void set_page_order(struct page *page, int order) {
239
	set_page_private(page, order);
240
	__SetPageBuddy(page);
Linus Torvalds's avatar
Linus Torvalds committed
241 242 243 244
}

static inline void rmv_page_order(struct page *page)
{
245
	__ClearPageBuddy(page);
246
	set_page_private(page, 0);
Linus Torvalds's avatar
Linus Torvalds committed
247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282
}

/*
 * 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)
 *
 * Assumption: *_mem_map is contigious at least up to MAX_ORDER
 */
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
283
 * (a) the buddy is not in a hole &&
284 285 286 287 288
 * (b) the buddy is in the buddy system &&
 * (c) a page and its buddy have the same order.
 *
 * 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
289
 *
290
 * For recording page's order, we use page_private(page).
Linus Torvalds's avatar
Linus Torvalds committed
291 292 293
 */
static inline int page_is_buddy(struct page *page, int order)
{
Nick Piggin's avatar
Nick Piggin committed
294 295 296 297 298
#ifdef CONFIG_HOLES_IN_ZONE
	if (!pfn_valid(page_to_pfn(page)))
		return 0;
#endif

299 300
	if (PageBuddy(page) && page_order(page) == order) {
		BUG_ON(page_count(page) != 0);
Linus Torvalds's avatar
Linus Torvalds committed
301
               return 1;
302
	}
Linus Torvalds's avatar
Linus Torvalds committed
303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318
       return 0;
}

/*
 * 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
319
 * free pages of length of (1 << order) and marked with PG_buddy. Page's
320
 * order is recorded in page_private(page) field.
Linus Torvalds's avatar
Linus Torvalds committed
321 322 323 324 325 326 327 328 329
 * 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
330
static inline void __free_one_page(struct page *page,
Linus Torvalds's avatar
Linus Torvalds committed
331 332 333 334 335
		struct zone *zone, unsigned int order)
{
	unsigned long page_idx;
	int order_size = 1 << order;

336
	if (unlikely(PageCompound(page)))
Linus Torvalds's avatar
Linus Torvalds committed
337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352
		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);
		if (!page_is_buddy(buddy, order))
			break;		/* Move the buddy up one level. */
Nick Piggin's avatar
Nick Piggin committed
353

Linus Torvalds's avatar
Linus Torvalds committed
354 355 356 357
		list_del(&buddy->lru);
		area = zone->free_area + order;
		area->nr_free--;
		rmv_page_order(buddy);
Nick Piggin's avatar
Nick Piggin committed
358
		combined_idx = __find_combined_index(page_idx, order);
Linus Torvalds's avatar
Linus Torvalds committed
359 360 361 362 363 364 365 366 367
		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++;
}

368
static inline int free_pages_check(struct page *page)
Linus Torvalds's avatar
Linus Torvalds committed
369
{
Nick Piggin's avatar
Nick Piggin committed
370 371 372
	if (unlikely(page_mapcount(page) |
		(page->mapping != NULL)  |
		(page_count(page) != 0)  |
Linus Torvalds's avatar
Linus Torvalds committed
373 374 375 376 377 378 379 380
		(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
381
			1 << PG_writeback |
382 383
			1 << PG_reserved |
			1 << PG_buddy ))))
384
		bad_page(page);
Linus Torvalds's avatar
Linus Torvalds committed
385
	if (PageDirty(page))
Nick Piggin's avatar
Nick Piggin committed
386
		__ClearPageDirty(page);
387 388 389 390 391 392
	/*
	 * 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
393 394 395 396 397
}

/*
 * Frees a list of pages. 
 * Assumes all pages on list are in same zone, and of same order.
398
 * count is the number of pages to free.
Linus Torvalds's avatar
Linus Torvalds committed
399 400 401 402 403 404 405
 *
 * 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
406 407
static void free_pages_bulk(struct zone *zone, int count,
					struct list_head *list, int order)
Linus Torvalds's avatar
Linus Torvalds committed
408
{
Nick Piggin's avatar
Nick Piggin committed
409
	spin_lock(&zone->lock);
Linus Torvalds's avatar
Linus Torvalds committed
410 411
	zone->all_unreclaimable = 0;
	zone->pages_scanned = 0;
Nick Piggin's avatar
Nick Piggin committed
412 413 414 415
	while (count--) {
		struct page *page;

		BUG_ON(list_empty(list));
Linus Torvalds's avatar
Linus Torvalds committed
416
		page = list_entry(list->prev, struct page, lru);
Nick Piggin's avatar
Nick Piggin committed
417
		/* have to delete it as __free_one_page list manipulates */
Linus Torvalds's avatar
Linus Torvalds committed
418
		list_del(&page->lru);
Nick Piggin's avatar
Nick Piggin committed
419
		__free_one_page(page, zone, order);
Linus Torvalds's avatar
Linus Torvalds committed
420
	}
Nick Piggin's avatar
Nick Piggin committed
421
	spin_unlock(&zone->lock);
Linus Torvalds's avatar
Linus Torvalds committed
422 423
}

Nick Piggin's avatar
Nick Piggin committed
424
static void free_one_page(struct zone *zone, struct page *page, int order)
Linus Torvalds's avatar
Linus Torvalds committed
425 426
{
	LIST_HEAD(list);
Nick Piggin's avatar
Nick Piggin committed
427 428 429 430 431 432 433
	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
434
	int i;
435
	int reserved = 0;
Linus Torvalds's avatar
Linus Torvalds committed
436 437

	arch_free_page(page, order);
438 439
	if (!PageHighMem(page))
		mutex_debug_check_no_locks_freed(page_address(page),
440
						 PAGE_SIZE<<order);
Linus Torvalds's avatar
Linus Torvalds committed
441 442

	for (i = 0 ; i < (1 << order) ; ++i)
443
		reserved += free_pages_check(page + i);
444 445 446
	if (reserved)
		return;

Nick Piggin's avatar
Nick Piggin committed
447
	kernel_map_pages(page, 1 << order, 0);
Nick Piggin's avatar
Nick Piggin committed
448
	local_irq_save(flags);
Nick Piggin's avatar
Nick Piggin committed
449
	__mod_page_state(pgfree, 1 << order);
Nick Piggin's avatar
Nick Piggin committed
450
	free_one_page(page_zone(page), page, order);
Nick Piggin's avatar
Nick Piggin committed
451
	local_irq_restore(flags);
Linus Torvalds's avatar
Linus Torvalds committed
452 453
}

454 455 456 457 458 459 460 461
/*
 * 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);
462
		set_page_refcounted(page);
Nick Piggin's avatar
Nick Piggin committed
463
		__free_page(page);
464 465 466
	} else {
		int loop;

Nick Piggin's avatar
Nick Piggin committed
467
		prefetchw(page);
468 469 470
		for (loop = 0; loop < BITS_PER_LONG; loop++) {
			struct page *p = &page[loop];

Nick Piggin's avatar
Nick Piggin committed
471 472
			if (loop + 1 < BITS_PER_LONG)
				prefetchw(p + 1);
473 474 475 476
			__ClearPageReserved(p);
			set_page_count(p, 0);
		}

477
		set_page_refcounted(page);
Nick Piggin's avatar
Nick Piggin committed
478
		__free_pages(page, order);
479 480 481
	}
}

Linus Torvalds's avatar
Linus Torvalds committed
482 483 484 485 486 487 488 489 490 491 492 493 494 495 496

/*
 * 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
497
static inline void expand(struct zone *zone, struct page *page,
Linus Torvalds's avatar
Linus Torvalds committed
498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515
 	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
516
static int prep_new_page(struct page *page, int order, gfp_t gfp_flags)
Linus Torvalds's avatar
Linus Torvalds committed
517
{
Nick Piggin's avatar
Nick Piggin committed
518 519 520
	if (unlikely(page_mapcount(page) |
		(page->mapping != NULL)  |
		(page_count(page) != 0)  |
521 522
		(page->flags & (
			1 << PG_lru	|
Linus Torvalds's avatar
Linus Torvalds committed
523 524 525 526 527
			1 << PG_private	|
			1 << PG_locked	|
			1 << PG_active	|
			1 << PG_dirty	|
			1 << PG_reclaim	|
528
			1 << PG_slab    |
Linus Torvalds's avatar
Linus Torvalds committed
529
			1 << PG_swapcache |
Nick Piggin's avatar
Nick Piggin committed
530
			1 << PG_writeback |
531 532
			1 << PG_reserved |
			1 << PG_buddy ))))
533
		bad_page(page);
Linus Torvalds's avatar
Linus Torvalds committed
534

535 536 537 538 539 540 541
	/*
	 * 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
542 543 544
	page->flags &= ~(1 << PG_uptodate | 1 << PG_error |
			1 << PG_referenced | 1 << PG_arch_1 |
			1 << PG_checked | 1 << PG_mappedtodisk);
545
	set_page_private(page, 0);
546
	set_page_refcounted(page);
Linus Torvalds's avatar
Linus Torvalds committed
547
	kernel_map_pages(page, 1 << order, 1);
Nick Piggin's avatar
Nick Piggin committed
548 549 550 551 552 553 554

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

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

555
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577
}

/* 
 * 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
578 579
		expand(zone, page, order, current_order, area);
		return page;
Linus Torvalds's avatar
Linus Torvalds committed
580 581 582 583 584 585 586 587 588 589 590 591 592 593 594
	}

	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
595
	spin_lock(&zone->lock);
Linus Torvalds's avatar
Linus Torvalds committed
596
	for (i = 0; i < count; ++i) {
Nick Piggin's avatar
Nick Piggin committed
597 598
		struct page *page = __rmqueue(zone, order);
		if (unlikely(page == NULL))
Linus Torvalds's avatar
Linus Torvalds committed
599 600 601
			break;
		list_add_tail(&page->lru, list);
	}
Nick Piggin's avatar
Nick Piggin committed
602
	spin_unlock(&zone->lock);
Nick Piggin's avatar
Nick Piggin committed
603
	return i;
Linus Torvalds's avatar
Linus Torvalds committed
604 605
}

606
#ifdef CONFIG_NUMA
607 608 609
/*
 * Called from the slab reaper to drain pagesets on a particular node that
 * belong to the currently executing processor.
610 611
 * Note that this function must be called with the thread pinned to
 * a single processor.
612 613
 */
void drain_node_pages(int nodeid)
614
{
615
	int i, z;
616 617
	unsigned long flags;

618 619
	for (z = 0; z < MAX_NR_ZONES; z++) {
		struct zone *zone = NODE_DATA(nodeid)->node_zones + z;
620 621
		struct per_cpu_pageset *pset;

Nick Piggin's avatar
Nick Piggin committed
622
		pset = zone_pcp(zone, smp_processor_id());
623 624 625 626
		for (i = 0; i < ARRAY_SIZE(pset->pcp); i++) {
			struct per_cpu_pages *pcp;

			pcp = &pset->pcp[i];
627 628 629 630 631 632
			if (pcp->count) {
				local_irq_save(flags);
				free_pages_bulk(zone, pcp->count, &pcp->list, 0);
				pcp->count = 0;
				local_irq_restore(flags);
			}
633 634 635 636 637
		}
	}
}
#endif

Linus Torvalds's avatar
Linus Torvalds committed
638 639 640
#if defined(CONFIG_PM) || defined(CONFIG_HOTPLUG_CPU)
static void __drain_pages(unsigned int cpu)
{
Nick Piggin's avatar
Nick Piggin committed
641
	unsigned long flags;
Linus Torvalds's avatar
Linus Torvalds committed
642 643 644 645 646 647
	struct zone *zone;
	int i;

	for_each_zone(zone) {
		struct per_cpu_pageset *pset;

648
		pset = zone_pcp(zone, cpu);
Linus Torvalds's avatar
Linus Torvalds committed
649 650 651 652
		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
653
			local_irq_save(flags);
Nick Piggin's avatar
Nick Piggin committed
654 655
			free_pages_bulk(zone, pcp->count, &pcp->list, 0);
			pcp->count = 0;
Nick Piggin's avatar
Nick Piggin committed
656
			local_irq_restore(flags);
Linus Torvalds's avatar
Linus Torvalds committed
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
		}
	}
}
#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
702
static void zone_statistics(struct zonelist *zonelist, struct zone *z, int cpu)
Linus Torvalds's avatar
Linus Torvalds committed
703 704 705 706 707 708
{
#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
709
	p = zone_pcp(z, cpu);
Linus Torvalds's avatar
Linus Torvalds committed
710
	if (pg == orig) {
711
		p->numa_hit++;
Linus Torvalds's avatar
Linus Torvalds committed
712 713
	} else {
		p->numa_miss++;
714
		zone_pcp(zonelist->zones[0], cpu)->numa_foreign++;
Linus Torvalds's avatar
Linus Torvalds committed
715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735
	}
	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;
736
	if (free_pages_check(page))
737 738 739 740
		return;

	kernel_map_pages(page, 1, 0);

741
	pcp = &zone_pcp(zone, get_cpu())->pcp[cold];
Linus Torvalds's avatar
Linus Torvalds committed
742
	local_irq_save(flags);
Nick Piggin's avatar
Nick Piggin committed
743
	__inc_page_state(pgfree);
Linus Torvalds's avatar
Linus Torvalds committed
744 745
	list_add(&page->lru, &pcp->list);
	pcp->count++;
Nick Piggin's avatar
Nick Piggin committed
746 747 748 749
	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
750 751 752 753 754 755 756 757 758 759 760 761 762 763
	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);
}

764 765 766 767 768 769 770 771 772 773 774 775 776 777
/*
 * 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));
778 779
	for (i = 1; i < (1 << order); i++)
		set_page_refcounted(page + i);
780 781
}

Linus Torvalds's avatar
Linus Torvalds committed
782 783 784 785 786
/*
 * 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
787 788
static struct page *buffered_rmqueue(struct zonelist *zonelist,
			struct zone *zone, int order, gfp_t gfp_flags)
Linus Torvalds's avatar
Linus Torvalds committed
789 790
{
	unsigned long flags;
791
	struct page *page;
Linus Torvalds's avatar
Linus Torvalds committed
792
	int cold = !!(gfp_flags & __GFP_COLD);
Nick Piggin's avatar
Nick Piggin committed
793
	int cpu;
Linus Torvalds's avatar
Linus Torvalds committed
794

795
again:
Nick Piggin's avatar
Nick Piggin committed
796
	cpu  = get_cpu();
Nick Piggin's avatar
Nick Piggin committed
797
	if (likely(order == 0)) {
Linus Torvalds's avatar
Linus Torvalds committed
798 799
		struct per_cpu_pages *pcp;

Nick Piggin's avatar
Nick Piggin committed
800
		pcp = &zone_pcp(zone, cpu)->pcp[cold];
Linus Torvalds's avatar
Linus Torvalds committed
801
		local_irq_save(flags);
Nick Piggin's avatar
Nick Piggin committed
802
		if (!pcp->count) {
Linus Torvalds's avatar
Linus Torvalds committed
803 804
			pcp->count += rmqueue_bulk(zone, 0,
						pcp->batch, &pcp->list);
Nick Piggin's avatar
Nick Piggin committed
805 806
			if (unlikely(!pcp->count))
				goto failed;
Linus Torvalds's avatar
Linus Torvalds committed
807
		}
Nick Piggin's avatar
Nick Piggin committed
808 809 810
		page = list_entry(pcp->list.next, struct page, lru);
		list_del(&page->lru);
		pcp->count--;
Rohit Seth's avatar
Rohit Seth committed
811
	} else {
Linus Torvalds's avatar
Linus Torvalds committed
812 813
		spin_lock_irqsave(&zone->lock, flags);
		page = __rmqueue(zone, order);
Nick Piggin's avatar
Nick Piggin committed
814 815 816
		spin_unlock(&zone->lock);
		if (!page)
			goto failed;
Linus Torvalds's avatar
Linus Torvalds committed
817 818
	}

Nick Piggin's avatar
Nick Piggin committed
819 820 821 822
	__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
823

Nick Piggin's avatar
Nick Piggin committed
824
	BUG_ON(bad_range(zone, page));
Nick Piggin's avatar
Nick Piggin committed
825
	if (prep_new_page(page, order, gfp_flags))
Nick Piggin's avatar
Nick Piggin committed
826
		goto again;
Linus Torvalds's avatar
Linus Torvalds committed
827
	return page;
Nick Piggin's avatar
Nick Piggin committed
828 829 830 831 832

failed:
	local_irq_restore(flags);
	put_cpu();
	return NULL;
Linus Torvalds's avatar
Linus Torvalds committed
833 834
}

Rohit Seth's avatar
Rohit Seth committed
835
#define ALLOC_NO_WATERMARKS	0x01 /* don't check watermarks at all */
836 837 838 839 840 841
#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
842

Linus Torvalds's avatar
Linus Torvalds committed
843 844 845 846 847
/*
 * 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
848
		      int classzone_idx, int alloc_flags)
Linus Torvalds's avatar
Linus Torvalds committed
849 850 851 852 853
{
	/* 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
854
	if (alloc_flags & ALLOC_HIGH)
Linus Torvalds's avatar
Linus Torvalds committed
855
		min -= min / 2;
Rohit Seth's avatar
Rohit Seth committed
856
	if (alloc_flags & ALLOC_HARDER)
Linus Torvalds's avatar
Linus Torvalds committed
857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873
		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
874 875 876 877 878 879 880
/*
 * 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
881
{
Rohit Seth's avatar
Rohit Seth committed
882 883 884 885 886 887 888 889 890 891 892 893 894 895
	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)) {
896 897 898 899 900 901 902 903
			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
904
				    classzone_idx, alloc_flags))
905 906 907
				if (!zone_reclaim_mode ||
				    !zone_reclaim(*z, gfp_mask, order))
					continue;
Rohit Seth's avatar
Rohit Seth committed
908 909
		}

Nick Piggin's avatar
Nick Piggin committed
910
		page = buffered_rmqueue(zonelist, *z, order, gfp_mask);
Rohit Seth's avatar
Rohit Seth committed
911 912 913 914 915
		if (page) {
			break;
		}
	} while (*(++z) != NULL);
	return page;
Martin Hicks's avatar
Martin Hicks committed
916 917
}

Linus Torvalds's avatar
Linus Torvalds committed
918 919 920 921
/*
 * This is the 'heart' of the zoned buddy allocator.
 */
struct page * fastcall
922
__alloc_pages(gfp_t gfp_mask, unsigned int order,
Linus Torvalds's avatar
Linus Torvalds committed
923 924
		struct zonelist *zonelist)
{
Al Viro's avatar
Al Viro committed
925
	const gfp_t wait = gfp_mask & __GFP_WAIT;
Rohit Seth's avatar
Rohit Seth committed
926
	struct zone **z;
Linus Torvalds's avatar
Linus Torvalds committed
927 928 929 930
	struct page *page;
	struct reclaim_state reclaim_state;
	struct task_struct *p = current;
	int do_retry;
Rohit Seth's avatar
Rohit Seth committed
931
	int alloc_flags;
Linus Torvalds's avatar
Linus Torvalds committed
932 933 934 935
	int did_some_progress;

	might_sleep_if(wait);

936
restart:
Rohit Seth's avatar
Rohit Seth committed
937
	z = zonelist->zones;  /* the list of zones suitable for gfp_mask */
Linus Torvalds's avatar
Linus Torvalds committed
938

Rohit Seth's avatar
Rohit Seth committed
939
	if (unlikely(*z == NULL)) {
Linus Torvalds's avatar
Linus Torvalds committed
940 941 942
		/* Should this ever happen?? */
		return NULL;
	}
943

Rohit Seth's avatar
Rohit Seth committed
944
	page = get_page_from_freelist(gfp_mask|__GFP_HARDWALL, order,
945
				zonelist, ALLOC_WMARK_LOW|ALLOC_CPUSET);
Rohit Seth's avatar
Rohit Seth committed
946 947
	if (page)
		goto got_pg;
Linus Torvalds's avatar
Linus Torvalds committed
948

949
	do {
950 951
		if (cpuset_zone_allowed(*z, gfp_mask))
			wakeup_kswapd(*z, order);
952
	} while (*(++z));
Linus Torvalds's avatar
Linus Torvalds committed
953

954
	/*
Rohit Seth's avatar
Rohit Seth committed
955 956 957 958 959 960
	 * 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
961 962
	 * policy or is asking for __GFP_HIGH memory.  GFP_ATOMIC requests will
	 * set both ALLOC_HARDER (!wait) and ALLOC_HIGH (__GFP_HIGH).
963
	 */
964
	alloc_flags = ALLOC_WMARK_MIN;
Rohit Seth's avatar
Rohit Seth committed
965 966 967 968
	if ((unlikely(rt_task(p)) && !in_interrupt()) || !wait)
		alloc_flags |= ALLOC_HARDER;
	if (gfp_mask & __GFP_HIGH)
		alloc_flags |= ALLOC_HIGH;
969
	alloc_flags |= ALLOC_CPUSET;
Linus Torvalds's avatar
Linus Torvalds committed
970 971 972

	/*
	 * Go through the zonelist again. Let __GFP_HIGH and allocations
Rohit Seth's avatar
Rohit Seth committed
973
	 * coming from realtime tasks go deeper into reserves.
Linus Torvalds's avatar
Linus Torvalds committed
974 975 976
	 *
	 * This is the last chance, in general, before the goto nopage.
	 * Ignore cpuset if GFP_ATOMIC (!wait) rather than fail alloc.
977
	 * See also cpuset_zone_allowed() comment in kernel/cpuset.c.
Linus Torvalds's avatar
Linus Torvalds committed
978
	 */
Rohit Seth's avatar
Rohit Seth committed
979 980 981
	page = get_page_from_freelist(gfp_mask, order, zonelist, alloc_flags);
	if (page)
		goto got_pg;
Linus Torvalds's avatar
Linus Torvalds committed
982 983

	/* This allocation should allow future memory freeing. */
984 985 986 987

	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
988
nofail_alloc:
989
			/* go through the zonelist yet again, ignoring mins */
Rohit Seth's avatar
Rohit Seth committed
990
			page = get_page_from_freelist(gfp_mask, order,
991
				zonelist, ALLOC_NO_WATERMARKS);
Rohit Seth's avatar
Rohit Seth committed
992 993
			if (page)
				goto got_pg;
Kirill Korotaev's avatar
Kirill Korotaev committed
994 995 996 997
			if (gfp_mask & __GFP_NOFAIL) {
				blk_congestion_wait(WRITE, HZ/50);
				goto nofail_alloc;
			}
Linus Torvalds's avatar
Linus Torvalds committed
998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009
		}
		goto nopage;
	}

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

rebalance:
	cond_resched();

	/* We now go into synchronous reclaim */
1010
	cpuset_memory_pressure_bump();
Linus Torvalds's avatar
Linus Torvalds committed
1011 1012 1013 1014
	p->flags |= PF_MEMALLOC;
	reclaim_state.reclaimed_slab = 0;
	p->reclaim_state = &reclaim_state;

Rohit Seth's avatar
Rohit Seth committed
1015
	did_some_progress = try_to_free_pages(zonelist->zones, gfp_mask);
Linus Torvalds's avatar
Linus Torvalds committed
1016 1017 1018 1019 1020 1021 1022

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

	cond_resched();

	if (likely(did_some_progress)) {
Rohit Seth's avatar
Rohit Seth committed
1023 1024 1025 1026
		page = get_page_from_freelist(gfp_mask, order,
						zonelist, alloc_flags);
		if (page)
			goto got_pg;
Linus Torvalds's avatar
Linus Torvalds committed
1027 1028 1029 1030 1031 1032 1033
	} 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
1034
		page = get_page_from_freelist(gfp_mask|__GFP_HARDWALL, order,
1035
				zonelist, ALLOC_WMARK_HIGH|ALLOC_CPUSET);
Rohit Seth's avatar
Rohit Seth committed
1036 1037
		if (page)
			goto got_pg;
Linus Torvalds's avatar
Linus Torvalds committed
1038

1039
		out_of_memory(zonelist, gfp_mask, order);
Linus Torvalds's avatar
Linus Torvalds committed
1040 1041 1042 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
		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
1068
		show_mem();
Linus Torvalds's avatar
Linus Torvalds committed
1069 1070 1071 1072 1073 1074 1075 1076 1077 1078
	}
got_pg:
	return page;
}

EXPORT_SYMBOL(__alloc_pages);

/*
 * Common helper functions.
 */
1079
fastcall unsigned long __get_free_pages(gfp_t gfp_mask, unsigned int order)
Linus Torvalds's avatar
Linus Torvalds committed
1080 1081 1082 1083 1084 1085 1086 1087 1088 1089
{
	struct page * page;
	page = alloc_pages(gfp_mask, order);
	if (!page)
		return 0;
	return (unsigned long) page_address(page);
}

EXPORT_SYMBOL(__get_free_pages);

1090
fastcall unsigned long get_zeroed_page(gfp_t gfp_mask)
Linus Torvalds's avatar
Linus Torvalds committed
1091 1092 1093 1094 1095 1096 1097
{
	struct page * page;

	/*
	 * get_zeroed_page() returns a 32-bit address, which cannot represent
	 * a highmem page
	 */
Al Viro's avatar
Al Viro committed
1098
	BUG_ON((gfp_mask & __GFP_HIGHMEM) != 0);
Linus Torvalds's avatar
Linus Torvalds committed
1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117

	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
1118
	if (put_page_testzero(page)) {
Linus Torvalds's avatar
Linus Torvalds committed
1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 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
		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)
{