page_alloc.c 134.8 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
/*
 *  linux/mm/page_alloc.c
 *
 *  Manages the free list, the system allocates free pages here.
 *  Note that kmalloc() lives in slab.c
 *
 *  Copyright (C) 1991, 1992, 1993, 1994  Linus Torvalds
 *  Swap reorganised 29.12.95, Stephen Tweedie
 *  Support of BIGMEM added by Gerhard Wichert, Siemens AG, July 1999
 *  Reshaped it to be a zoned allocator, Ingo Molnar, Red Hat, 1999
 *  Discontiguous memory support, Kanoj Sarcar, SGI, Nov 1999
 *  Zone balancing, Kanoj Sarcar, SGI, Jan 2000
 *  Per cpu hot/cold page lists, bulk allocation, Martin J. Bligh, Sept 2002
 *          (lots of bits borrowed from Ingo Molnar & Andrew Morton)
 */

#include <linux/stddef.h>
#include <linux/mm.h>
#include <linux/swap.h>
#include <linux/interrupt.h>
#include <linux/pagemap.h>
22
#include <linux/jiffies.h>
L
Linus Torvalds 已提交
23 24
#include <linux/bootmem.h>
#include <linux/compiler.h>
25
#include <linux/kernel.h>
26
#include <linux/kmemcheck.h>
L
Linus Torvalds 已提交
27 28 29 30 31
#include <linux/module.h>
#include <linux/suspend.h>
#include <linux/pagevec.h>
#include <linux/blkdev.h>
#include <linux/slab.h>
32
#include <linux/oom.h>
L
Linus Torvalds 已提交
33 34 35 36 37
#include <linux/notifier.h>
#include <linux/topology.h>
#include <linux/sysctl.h>
#include <linux/cpu.h>
#include <linux/cpuset.h>
38
#include <linux/memory_hotplug.h>
L
Linus Torvalds 已提交
39 40
#include <linux/nodemask.h>
#include <linux/vmalloc.h>
41
#include <linux/mempolicy.h>
42
#include <linux/stop_machine.h>
43 44
#include <linux/sort.h>
#include <linux/pfn.h>
45
#include <linux/backing-dev.h>
46
#include <linux/fault-inject.h>
K
KAMEZAWA Hiroyuki 已提交
47
#include <linux/page-isolation.h>
48
#include <linux/page_cgroup.h>
49
#include <linux/debugobjects.h>
50
#include <linux/kmemleak.h>
L
Linus Torvalds 已提交
51 52

#include <asm/tlbflush.h>
53
#include <asm/div64.h>
L
Linus Torvalds 已提交
54 55 56
#include "internal.h"

/*
57
 * Array of node states.
L
Linus Torvalds 已提交
58
 */
59 60 61 62 63 64 65 66 67 68 69 70 71
nodemask_t node_states[NR_NODE_STATES] __read_mostly = {
	[N_POSSIBLE] = NODE_MASK_ALL,
	[N_ONLINE] = { { [0] = 1UL } },
#ifndef CONFIG_NUMA
	[N_NORMAL_MEMORY] = { { [0] = 1UL } },
#ifdef CONFIG_HIGHMEM
	[N_HIGH_MEMORY] = { { [0] = 1UL } },
#endif
	[N_CPU] = { { [0] = 1UL } },
#endif	/* NUMA */
};
EXPORT_SYMBOL(node_states);

72
unsigned long totalram_pages __read_mostly;
73
unsigned long totalreserve_pages __read_mostly;
74
unsigned long highest_memmap_pfn __read_mostly;
75
int percpu_pagelist_fraction;
76
gfp_t gfp_allowed_mask __read_mostly = GFP_BOOT_MASK;
L
Linus Torvalds 已提交
77

78 79 80 81
#ifdef CONFIG_HUGETLB_PAGE_SIZE_VARIABLE
int pageblock_order __read_mostly;
#endif

82
static void __free_pages_ok(struct page *page, unsigned int order);
83

L
Linus Torvalds 已提交
84 85 86 87 88 89 90
/*
 * 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
A
Andi Kleen 已提交
91 92 93
 *
 * TBD: should special case ZONE_DMA32 machines here - in those we normally
 * don't need any ZONE_NORMAL reservation
L
Linus Torvalds 已提交
94
 */
95
int sysctl_lowmem_reserve_ratio[MAX_NR_ZONES-1] = {
96
#ifdef CONFIG_ZONE_DMA
97
	 256,
98
#endif
99
#ifdef CONFIG_ZONE_DMA32
100
	 256,
101
#endif
102
#ifdef CONFIG_HIGHMEM
M
Mel Gorman 已提交
103
	 32,
104
#endif
M
Mel Gorman 已提交
105
	 32,
106
};
L
Linus Torvalds 已提交
107 108 109

EXPORT_SYMBOL(totalram_pages);

110
static char * const zone_names[MAX_NR_ZONES] = {
111
#ifdef CONFIG_ZONE_DMA
112
	 "DMA",
113
#endif
114
#ifdef CONFIG_ZONE_DMA32
115
	 "DMA32",
116
#endif
117
	 "Normal",
118
#ifdef CONFIG_HIGHMEM
M
Mel Gorman 已提交
119
	 "HighMem",
120
#endif
M
Mel Gorman 已提交
121
	 "Movable",
122 123
};

L
Linus Torvalds 已提交
124 125
int min_free_kbytes = 1024;

126 127
unsigned long __meminitdata nr_kernel_pages;
unsigned long __meminitdata nr_all_pages;
128
static unsigned long __meminitdata dma_reserve;
L
Linus Torvalds 已提交
129

130 131
#ifdef CONFIG_ARCH_POPULATES_NODE_MAP
  /*
S
Simon Arlott 已提交
132
   * MAX_ACTIVE_REGIONS determines the maximum number of distinct
133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150
   * ranges of memory (RAM) that may be registered with add_active_range().
   * Ranges passed to add_active_range() will be merged if possible
   * so the number of times add_active_range() can be called is
   * related to the number of nodes and the number of holes
   */
  #ifdef CONFIG_MAX_ACTIVE_REGIONS
    /* Allow an architecture to set MAX_ACTIVE_REGIONS to save memory */
    #define MAX_ACTIVE_REGIONS CONFIG_MAX_ACTIVE_REGIONS
  #else
    #if MAX_NUMNODES >= 32
      /* If there can be many nodes, allow up to 50 holes per node */
      #define MAX_ACTIVE_REGIONS (MAX_NUMNODES*50)
    #else
      /* By default, allow up to 256 distinct regions */
      #define MAX_ACTIVE_REGIONS 256
    #endif
  #endif

151 152 153 154
  static struct node_active_region __meminitdata early_node_map[MAX_ACTIVE_REGIONS];
  static int __meminitdata nr_nodemap_entries;
  static unsigned long __meminitdata arch_zone_lowest_possible_pfn[MAX_NR_ZONES];
  static unsigned long __meminitdata arch_zone_highest_possible_pfn[MAX_NR_ZONES];
A
Adrian Bunk 已提交
155
  static unsigned long __initdata required_kernelcore;
A
Adrian Bunk 已提交
156
  static unsigned long __initdata required_movablecore;
A
Adrian Bunk 已提交
157
  static unsigned long __meminitdata zone_movable_pfn[MAX_NUMNODES];
M
Mel Gorman 已提交
158 159 160 161

  /* movable_zone is the "real" zone pages in ZONE_MOVABLE are taken from */
  int movable_zone;
  EXPORT_SYMBOL(movable_zone);
162 163
#endif /* CONFIG_ARCH_POPULATES_NODE_MAP */

M
Miklos Szeredi 已提交
164 165
#if MAX_NUMNODES > 1
int nr_node_ids __read_mostly = MAX_NUMNODES;
166
int nr_online_nodes __read_mostly = 1;
M
Miklos Szeredi 已提交
167
EXPORT_SYMBOL(nr_node_ids);
168
EXPORT_SYMBOL(nr_online_nodes);
M
Miklos Szeredi 已提交
169 170
#endif

171 172
int page_group_by_mobility_disabled __read_mostly;

173 174
static void set_pageblock_migratetype(struct page *page, int migratetype)
{
175 176 177 178

	if (unlikely(page_group_by_mobility_disabled))
		migratetype = MIGRATE_UNMOVABLE;

179 180 181 182
	set_pageblock_flags_group(page, (unsigned long)migratetype,
					PB_migrate, PB_migrate_end);
}

183 184
bool oom_killer_disabled __read_mostly;

N
Nick Piggin 已提交
185
#ifdef CONFIG_DEBUG_VM
186
static int page_outside_zone_boundaries(struct zone *zone, struct page *page)
L
Linus Torvalds 已提交
187
{
188 189 190
	int ret = 0;
	unsigned seq;
	unsigned long pfn = page_to_pfn(page);
191

192 193 194 195 196 197 198 199 200
	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;
201 202 203 204
}

static int page_is_consistent(struct zone *zone, struct page *page)
{
205
	if (!pfn_valid_within(page_to_pfn(page)))
206
		return 0;
L
Linus Torvalds 已提交
207
	if (zone != page_zone(page))
208 209 210 211 212 213 214 215 216 217
		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))
L
Linus Torvalds 已提交
218
		return 1;
219 220 221
	if (!page_is_consistent(zone, page))
		return 1;

L
Linus Torvalds 已提交
222 223
	return 0;
}
N
Nick Piggin 已提交
224 225 226 227 228 229 230
#else
static inline int bad_range(struct zone *zone, struct page *page)
{
	return 0;
}
#endif

N
Nick Piggin 已提交
231
static void bad_page(struct page *page)
L
Linus Torvalds 已提交
232
{
233 234 235 236 237 238 239 240 241 242 243 244 245 246
	static unsigned long resume;
	static unsigned long nr_shown;
	static unsigned long nr_unshown;

	/*
	 * Allow a burst of 60 reports, then keep quiet for that minute;
	 * or allow a steady drip of one report per second.
	 */
	if (nr_shown == 60) {
		if (time_before(jiffies, resume)) {
			nr_unshown++;
			goto out;
		}
		if (nr_unshown) {
247 248
			printk(KERN_ALERT
			      "BUG: Bad page state: %lu messages suppressed\n",
249 250 251 252 253 254 255 256
				nr_unshown);
			nr_unshown = 0;
		}
		nr_shown = 0;
	}
	if (nr_shown++ == 0)
		resume = jiffies + 60 * HZ;

257
	printk(KERN_ALERT "BUG: Bad page state in process %s  pfn:%05lx\n",
258
		current->comm, page_to_pfn(page));
259
	printk(KERN_ALERT
260 261 262 263
		"page:%p flags:%p count:%d mapcount:%d mapping:%p index:%lx\n",
		page, (void *)page->flags, page_count(page),
		page_mapcount(page), page->mapping, page->index);

L
Linus Torvalds 已提交
264
	dump_stack();
265
out:
266 267
	/* Leave bad fields for debug, except PageBuddy could make trouble */
	__ClearPageBuddy(page);
268
	add_taint(TAINT_BAD_PAGE);
L
Linus Torvalds 已提交
269 270 271 272 273 274 275 276 277 278 279 280
}

/*
 * 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).
 *
281 282 283
 * 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.
L
Linus Torvalds 已提交
284
 */
285 286 287

static void free_compound_page(struct page *page)
{
288
	__free_pages_ok(page, compound_order(page));
289 290
}

A
Andi Kleen 已提交
291
void prep_compound_page(struct page *page, unsigned long order)
292 293 294 295 296 297 298 299 300 301 302 303 304 305 306
{
	int i;
	int nr_pages = 1 << order;

	set_compound_page_dtor(page, free_compound_page);
	set_compound_order(page, order);
	__SetPageHead(page);
	for (i = 1; i < nr_pages; i++) {
		struct page *p = page + i;

		__SetPageTail(p);
		p->first_page = page;
	}
}

307
static int destroy_compound_page(struct page *page, unsigned long order)
L
Linus Torvalds 已提交
308 309 310
{
	int i;
	int nr_pages = 1 << order;
311
	int bad = 0;
L
Linus Torvalds 已提交
312

313 314
	if (unlikely(compound_order(page) != order) ||
	    unlikely(!PageHead(page))) {
N
Nick Piggin 已提交
315
		bad_page(page);
316 317
		bad++;
	}
L
Linus Torvalds 已提交
318

319
	__ClearPageHead(page);
320

321 322
	for (i = 1; i < nr_pages; i++) {
		struct page *p = page + i;
L
Linus Torvalds 已提交
323

324
		if (unlikely(!PageTail(p) || (p->first_page != page))) {
N
Nick Piggin 已提交
325
			bad_page(page);
326 327
			bad++;
		}
328
		__ClearPageTail(p);
L
Linus Torvalds 已提交
329
	}
330 331

	return bad;
L
Linus Torvalds 已提交
332 333
}

N
Nick Piggin 已提交
334 335 336 337
static inline void prep_zero_page(struct page *page, int order, gfp_t gfp_flags)
{
	int i;

338 339 340 341
	/*
	 * clear_highpage() will use KM_USER0, so it's a bug to use __GFP_ZERO
	 * and __GFP_HIGHMEM from hard or soft interrupt context.
	 */
N
Nick Piggin 已提交
342
	VM_BUG_ON((gfp_flags & __GFP_HIGHMEM) && in_interrupt());
N
Nick Piggin 已提交
343 344 345 346
	for (i = 0; i < (1 << order); i++)
		clear_highpage(page + i);
}

347 348
static inline void set_page_order(struct page *page, int order)
{
H
Hugh Dickins 已提交
349
	set_page_private(page, order);
350
	__SetPageBuddy(page);
L
Linus Torvalds 已提交
351 352 353 354
}

static inline void rmv_page_order(struct page *page)
{
355
	__ClearPageBuddy(page);
H
Hugh Dickins 已提交
356
	set_page_private(page, 0);
L
Linus Torvalds 已提交
357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373
}

/*
 * 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)
 *
A
Andreas Mohr 已提交
374
 * Assumption: *_mem_map is contiguous at least up to MAX_ORDER
L
Linus Torvalds 已提交
375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392
 */
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
N
Nick Piggin 已提交
393
 * (a) the buddy is not in a hole &&
394
 * (b) the buddy is in the buddy system &&
395 396
 * (c) a page and its buddy have the same order &&
 * (d) a page and its buddy are in the same zone.
397 398 399
 *
 * 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.
L
Linus Torvalds 已提交
400
 *
401
 * For recording page's order, we use page_private(page).
L
Linus Torvalds 已提交
402
 */
403 404
static inline int page_is_buddy(struct page *page, struct page *buddy,
								int order)
L
Linus Torvalds 已提交
405
{
406
	if (!pfn_valid_within(page_to_pfn(buddy)))
N
Nick Piggin 已提交
407 408
		return 0;

409 410 411 412
	if (page_zone_id(page) != page_zone_id(buddy))
		return 0;

	if (PageBuddy(buddy) && page_order(buddy) == order) {
413
		VM_BUG_ON(page_count(buddy) != 0);
414
		return 1;
415
	}
416
	return 0;
L
Linus Torvalds 已提交
417 418 419 420 421 422 423 424 425 426 427 428 429 430 431
}

/*
 * 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
432
 * free pages of length of (1 << order) and marked with PG_buddy. Page's
H
Hugh Dickins 已提交
433
 * order is recorded in page_private(page) field.
L
Linus Torvalds 已提交
434 435 436 437 438 439 440 441 442
 * 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
 */

N
Nick Piggin 已提交
443
static inline void __free_one_page(struct page *page,
444 445
		struct zone *zone, unsigned int order,
		int migratetype)
L
Linus Torvalds 已提交
446 447 448
{
	unsigned long page_idx;

N
Nick Piggin 已提交
449
	if (unlikely(PageCompound(page)))
450 451
		if (unlikely(destroy_compound_page(page, order)))
			return;
L
Linus Torvalds 已提交
452

453 454
	VM_BUG_ON(migratetype == -1);

L
Linus Torvalds 已提交
455 456
	page_idx = page_to_pfn(page) & ((1 << MAX_ORDER) - 1);

457
	VM_BUG_ON(page_idx & ((1 << order) - 1));
N
Nick Piggin 已提交
458
	VM_BUG_ON(bad_range(zone, page));
L
Linus Torvalds 已提交
459 460 461 462 463 464

	while (order < MAX_ORDER-1) {
		unsigned long combined_idx;
		struct page *buddy;

		buddy = __page_find_buddy(page, page_idx, order);
465
		if (!page_is_buddy(page, buddy, order))
466
			break;
N
Nick Piggin 已提交
467

468
		/* Our buddy is free, merge with it and move up one order. */
L
Linus Torvalds 已提交
469
		list_del(&buddy->lru);
470
		zone->free_area[order].nr_free--;
L
Linus Torvalds 已提交
471
		rmv_page_order(buddy);
N
Nick Piggin 已提交
472
		combined_idx = __find_combined_index(page_idx, order);
L
Linus Torvalds 已提交
473 474 475 476 477
		page = page + (combined_idx - page_idx);
		page_idx = combined_idx;
		order++;
	}
	set_page_order(page, order);
478 479
	list_add(&page->lru,
		&zone->free_area[order].free_list[migratetype]);
L
Linus Torvalds 已提交
480 481 482
	zone->free_area[order].nr_free++;
}

483 484 485 486 487 488 489 490 491 492 493 494 495 496 497
#ifdef CONFIG_HAVE_MLOCKED_PAGE_BIT
/*
 * free_page_mlock() -- clean up attempts to free and mlocked() page.
 * Page should not be on lru, so no need to fix that up.
 * free_pages_check() will verify...
 */
static inline void free_page_mlock(struct page *page)
{
	__dec_zone_page_state(page, NR_MLOCK);
	__count_vm_event(UNEVICTABLE_MLOCKFREED);
}
#else
static void free_page_mlock(struct page *page) { }
#endif

N
Nick Piggin 已提交
498
static inline int free_pages_check(struct page *page)
L
Linus Torvalds 已提交
499
{
N
Nick Piggin 已提交
500 501
	if (unlikely(page_mapcount(page) |
		(page->mapping != NULL)  |
502
		(atomic_read(&page->_count) != 0) |
503
		(page->flags & PAGE_FLAGS_CHECK_AT_FREE))) {
N
Nick Piggin 已提交
504
		bad_page(page);
505
		return 1;
506
	}
507 508 509
	if (page->flags & PAGE_FLAGS_CHECK_AT_PREP)
		page->flags &= ~PAGE_FLAGS_CHECK_AT_PREP;
	return 0;
L
Linus Torvalds 已提交
510 511 512 513 514
}

/*
 * Frees a list of pages. 
 * Assumes all pages on list are in same zone, and of same order.
515
 * count is the number of pages to free.
L
Linus Torvalds 已提交
516 517 518 519 520 521 522
 *
 * 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.
 */
N
Nick Piggin 已提交
523 524
static void free_pages_bulk(struct zone *zone, int count,
					struct list_head *list, int order)
L
Linus Torvalds 已提交
525
{
N
Nick Piggin 已提交
526
	spin_lock(&zone->lock);
527
	zone_clear_flag(zone, ZONE_ALL_UNRECLAIMABLE);
L
Linus Torvalds 已提交
528
	zone->pages_scanned = 0;
529 530

	__mod_zone_page_state(zone, NR_FREE_PAGES, count << order);
N
Nick Piggin 已提交
531 532 533
	while (count--) {
		struct page *page;

N
Nick Piggin 已提交
534
		VM_BUG_ON(list_empty(list));
L
Linus Torvalds 已提交
535
		page = list_entry(list->prev, struct page, lru);
N
Nick Piggin 已提交
536
		/* have to delete it as __free_one_page list manipulates */
L
Linus Torvalds 已提交
537
		list_del(&page->lru);
538
		__free_one_page(page, zone, order, page_private(page));
L
Linus Torvalds 已提交
539
	}
N
Nick Piggin 已提交
540
	spin_unlock(&zone->lock);
L
Linus Torvalds 已提交
541 542
}

543 544
static void free_one_page(struct zone *zone, struct page *page, int order,
				int migratetype)
L
Linus Torvalds 已提交
545
{
546
	spin_lock(&zone->lock);
547
	zone_clear_flag(zone, ZONE_ALL_UNRECLAIMABLE);
548
	zone->pages_scanned = 0;
549 550

	__mod_zone_page_state(zone, NR_FREE_PAGES, 1 << order);
551
	__free_one_page(page, zone, order, migratetype);
552
	spin_unlock(&zone->lock);
N
Nick Piggin 已提交
553 554 555 556 557
}

static void __free_pages_ok(struct page *page, unsigned int order)
{
	unsigned long flags;
L
Linus Torvalds 已提交
558
	int i;
559
	int bad = 0;
560
	int wasMlocked = TestClearPageMlocked(page);
L
Linus Torvalds 已提交
561

562 563
	kmemcheck_free_shadow(page, order);

L
Linus Torvalds 已提交
564
	for (i = 0 ; i < (1 << order) ; ++i)
565 566
		bad += free_pages_check(page + i);
	if (bad)
567 568
		return;

569
	if (!PageHighMem(page)) {
N
Nick Piggin 已提交
570
		debug_check_no_locks_freed(page_address(page),PAGE_SIZE<<order);
571 572 573
		debug_check_no_obj_freed(page_address(page),
					   PAGE_SIZE << order);
	}
N
Nick Piggin 已提交
574
	arch_free_page(page, order);
N
Nick Piggin 已提交
575
	kernel_map_pages(page, 1 << order, 0);
N
Nick Piggin 已提交
576

N
Nick Piggin 已提交
577
	local_irq_save(flags);
578
	if (unlikely(wasMlocked))
579
		free_page_mlock(page);
580
	__count_vm_events(PGFREE, 1 << order);
581 582
	free_one_page(page_zone(page), page, order,
					get_pageblock_migratetype(page));
N
Nick Piggin 已提交
583
	local_irq_restore(flags);
L
Linus Torvalds 已提交
584 585
}

586 587 588
/*
 * permit the bootmem allocator to evade page validation on high-order frees
 */
589
void __meminit __free_pages_bootmem(struct page *page, unsigned int order)
590 591 592 593
{
	if (order == 0) {
		__ClearPageReserved(page);
		set_page_count(page, 0);
594
		set_page_refcounted(page);
N
Nick Piggin 已提交
595
		__free_page(page);
596 597 598
	} else {
		int loop;

N
Nick Piggin 已提交
599
		prefetchw(page);
600 601 602
		for (loop = 0; loop < BITS_PER_LONG; loop++) {
			struct page *p = &page[loop];

N
Nick Piggin 已提交
603 604
			if (loop + 1 < BITS_PER_LONG)
				prefetchw(p + 1);
605 606 607 608
			__ClearPageReserved(p);
			set_page_count(p, 0);
		}

609
		set_page_refcounted(page);
N
Nick Piggin 已提交
610
		__free_pages(page, order);
611 612 613
	}
}

L
Linus Torvalds 已提交
614 615 616 617 618 619 620 621 622 623 624 625 626 627 628

/*
 * 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
 */
N
Nick Piggin 已提交
629
static inline void expand(struct zone *zone, struct page *page,
630 631
	int low, int high, struct free_area *area,
	int migratetype)
L
Linus Torvalds 已提交
632 633 634 635 636 637 638
{
	unsigned long size = 1 << high;

	while (high > low) {
		area--;
		high--;
		size >>= 1;
N
Nick Piggin 已提交
639
		VM_BUG_ON(bad_range(zone, &page[size]));
640
		list_add(&page[size].lru, &area->free_list[migratetype]);
L
Linus Torvalds 已提交
641 642 643 644 645 646 647 648
		area->nr_free++;
		set_page_order(&page[size], high);
	}
}

/*
 * This page is about to be returned from the page allocator
 */
N
Nick Piggin 已提交
649
static int prep_new_page(struct page *page, int order, gfp_t gfp_flags)
L
Linus Torvalds 已提交
650
{
N
Nick Piggin 已提交
651 652
	if (unlikely(page_mapcount(page) |
		(page->mapping != NULL)  |
653
		(atomic_read(&page->_count) != 0)  |
654
		(page->flags & PAGE_FLAGS_CHECK_AT_PREP))) {
N
Nick Piggin 已提交
655
		bad_page(page);
656
		return 1;
657
	}
658

H
Hugh Dickins 已提交
659
	set_page_private(page, 0);
660
	set_page_refcounted(page);
N
Nick Piggin 已提交
661 662

	arch_alloc_page(page, order);
L
Linus Torvalds 已提交
663
	kernel_map_pages(page, 1 << order, 1);
N
Nick Piggin 已提交
664 665 666 667 668 669 670

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

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

671
	return 0;
L
Linus Torvalds 已提交
672 673
}

674 675 676 677
/*
 * Go through the free lists for the given migratetype and remove
 * the smallest available page from the freelists
 */
678 679
static inline
struct page *__rmqueue_smallest(struct zone *zone, unsigned int order,
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
						int migratetype)
{
	unsigned int current_order;
	struct free_area * area;
	struct page *page;

	/* Find a page of the appropriate size in the preferred list */
	for (current_order = order; current_order < MAX_ORDER; ++current_order) {
		area = &(zone->free_area[current_order]);
		if (list_empty(&area->free_list[migratetype]))
			continue;

		page = list_entry(area->free_list[migratetype].next,
							struct page, lru);
		list_del(&page->lru);
		rmv_page_order(page);
		area->nr_free--;
		expand(zone, page, order, current_order, area, migratetype);
		return page;
	}

	return NULL;
}


705 706 707 708 709
/*
 * This array describes the order lists are fallen back to when
 * the free lists for the desirable migrate type are depleted
 */
static int fallbacks[MIGRATE_TYPES][MIGRATE_TYPES-1] = {
710 711 712 713
	[MIGRATE_UNMOVABLE]   = { MIGRATE_RECLAIMABLE, MIGRATE_MOVABLE,   MIGRATE_RESERVE },
	[MIGRATE_RECLAIMABLE] = { MIGRATE_UNMOVABLE,   MIGRATE_MOVABLE,   MIGRATE_RESERVE },
	[MIGRATE_MOVABLE]     = { MIGRATE_RECLAIMABLE, MIGRATE_UNMOVABLE, MIGRATE_RESERVE },
	[MIGRATE_RESERVE]     = { MIGRATE_RESERVE,     MIGRATE_RESERVE,   MIGRATE_RESERVE }, /* Never used */
714 715
};

716 717
/*
 * Move the free pages in a range to the free lists of the requested type.
718
 * Note that start_page and end_pages are not aligned on a pageblock
719 720
 * boundary. If alignment is required, use move_freepages_block()
 */
A
Adrian Bunk 已提交
721 722 723
static int move_freepages(struct zone *zone,
			  struct page *start_page, struct page *end_page,
			  int migratetype)
724 725 726
{
	struct page *page;
	unsigned long order;
727
	int pages_moved = 0;
728 729 730 731 732 733 734

#ifndef CONFIG_HOLES_IN_ZONE
	/*
	 * page_zone is not safe to call in this context when
	 * CONFIG_HOLES_IN_ZONE is set. This bug check is probably redundant
	 * anyway as we check zone boundaries in move_freepages_block().
	 * Remove at a later date when no bug reports exist related to
M
Mel Gorman 已提交
735
	 * grouping pages by mobility
736 737 738 739 740
	 */
	BUG_ON(page_zone(start_page) != page_zone(end_page));
#endif

	for (page = start_page; page <= end_page;) {
741 742 743
		/* Make sure we are not inadvertently changing nodes */
		VM_BUG_ON(page_to_nid(page) != zone_to_nid(zone));

744 745 746 747 748 749 750 751 752 753 754 755 756 757 758
		if (!pfn_valid_within(page_to_pfn(page))) {
			page++;
			continue;
		}

		if (!PageBuddy(page)) {
			page++;
			continue;
		}

		order = page_order(page);
		list_del(&page->lru);
		list_add(&page->lru,
			&zone->free_area[order].free_list[migratetype]);
		page += 1 << order;
759
		pages_moved += 1 << order;
760 761
	}

762
	return pages_moved;
763 764
}

A
Adrian Bunk 已提交
765 766
static int move_freepages_block(struct zone *zone, struct page *page,
				int migratetype)
767 768 769 770 771
{
	unsigned long start_pfn, end_pfn;
	struct page *start_page, *end_page;

	start_pfn = page_to_pfn(page);
772
	start_pfn = start_pfn & ~(pageblock_nr_pages-1);
773
	start_page = pfn_to_page(start_pfn);
774 775
	end_page = start_page + pageblock_nr_pages - 1;
	end_pfn = start_pfn + pageblock_nr_pages - 1;
776 777 778 779 780 781 782 783 784 785

	/* Do not cross zone boundaries */
	if (start_pfn < zone->zone_start_pfn)
		start_page = page;
	if (end_pfn >= zone->zone_start_pfn + zone->spanned_pages)
		return 0;

	return move_freepages(zone, start_page, end_page, migratetype);
}

786
/* Remove an element from the buddy allocator from the fallback list */
787 788
static inline struct page *
__rmqueue_fallback(struct zone *zone, int order, int start_migratetype)
789 790 791 792 793 794 795 796 797 798 799 800
{
	struct free_area * area;
	int current_order;
	struct page *page;
	int migratetype, i;

	/* Find the largest possible block of pages in the other list */
	for (current_order = MAX_ORDER-1; current_order >= order;
						--current_order) {
		for (i = 0; i < MIGRATE_TYPES - 1; i++) {
			migratetype = fallbacks[start_migratetype][i];

801 802 803
			/* MIGRATE_RESERVE handled later if necessary */
			if (migratetype == MIGRATE_RESERVE)
				continue;
M
Mel Gorman 已提交
804

805 806 807 808 809 810 811 812 813
			area = &(zone->free_area[current_order]);
			if (list_empty(&area->free_list[migratetype]))
				continue;

			page = list_entry(area->free_list[migratetype].next,
					struct page, lru);
			area->nr_free--;

			/*
814
			 * If breaking a large block of pages, move all free
815 816 817
			 * pages to the preferred allocation list. If falling
			 * back for a reclaimable kernel allocation, be more
			 * agressive about taking ownership of free pages
818
			 */
819
			if (unlikely(current_order >= (pageblock_order >> 1)) ||
820 821 822 823 824 825
					start_migratetype == MIGRATE_RECLAIMABLE) {
				unsigned long pages;
				pages = move_freepages_block(zone, page,
								start_migratetype);

				/* Claim the whole block if over half of it is free */
826
				if (pages >= (1 << (pageblock_order-1)))
827 828 829
					set_pageblock_migratetype(page,
								start_migratetype);

830
				migratetype = start_migratetype;
831
			}
832 833 834 835 836

			/* Remove the page from the freelists */
			list_del(&page->lru);
			rmv_page_order(page);

837
			if (current_order == pageblock_order)
838 839 840 841 842 843 844 845
				set_pageblock_migratetype(page,
							start_migratetype);

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

846
	return NULL;
847 848
}

849
/*
L
Linus Torvalds 已提交
850 851 852
 * Do the hard work of removing an element from the buddy allocator.
 * Call me with the zone->lock already held.
 */
853 854
static struct page *__rmqueue(struct zone *zone, unsigned int order,
						int migratetype)
L
Linus Torvalds 已提交
855 856 857
{
	struct page *page;

858
retry_reserve:
859
	page = __rmqueue_smallest(zone, order, migratetype);
860

861
	if (unlikely(!page) && migratetype != MIGRATE_RESERVE) {
862
		page = __rmqueue_fallback(zone, order, migratetype);
863

864 865 866 867 868 869 870 871 872 873 874
		/*
		 * Use MIGRATE_RESERVE rather than fail an allocation. goto
		 * is used because __rmqueue_smallest is an inline function
		 * and we want just one call site
		 */
		if (!page) {
			migratetype = MIGRATE_RESERVE;
			goto retry_reserve;
		}
	}

875
	return page;
L
Linus Torvalds 已提交
876 877 878 879 880 881 882 883
}

/* 
 * 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, 
884 885
			unsigned long count, struct list_head *list,
			int migratetype)
L
Linus Torvalds 已提交
886 887 888
{
	int i;
	
N
Nick Piggin 已提交
889
	spin_lock(&zone->lock);
L
Linus Torvalds 已提交
890
	for (i = 0; i < count; ++i) {
891
		struct page *page = __rmqueue(zone, order, migratetype);
N
Nick Piggin 已提交
892
		if (unlikely(page == NULL))
L
Linus Torvalds 已提交
893
			break;
894 895 896 897 898 899 900 901 902 903

		/*
		 * Split buddy pages returned by expand() are received here
		 * in physical page order. The page is added to the callers and
		 * list and the list head then moves forward. From the callers
		 * perspective, the linked list is ordered by page number in
		 * some conditions. This is useful for IO devices that can
		 * merge IO requests if the physical pages are ordered
		 * properly.
		 */
904 905
		list_add(&page->lru, list);
		set_page_private(page, migratetype);
906
		list = &page->lru;
L
Linus Torvalds 已提交
907
	}
908
	__mod_zone_page_state(zone, NR_FREE_PAGES, -(i << order));
N
Nick Piggin 已提交
909
	spin_unlock(&zone->lock);
N
Nick Piggin 已提交
910
	return i;
L
Linus Torvalds 已提交
911 912
}

913
#ifdef CONFIG_NUMA
914
/*
915 916 917 918
 * Called from the vmstat counter updater to drain pagesets of this
 * currently executing processor on remote nodes after they have
 * expired.
 *
919 920
 * Note that this function must be called with the thread pinned to
 * a single processor.
921
 */
922
void drain_zone_pages(struct zone *zone, struct per_cpu_pages *pcp)
923 924
{
	unsigned long flags;
925
	int to_drain;
926

927 928 929 930 931 932 933 934
	local_irq_save(flags);
	if (pcp->count >= pcp->batch)
		to_drain = pcp->batch;
	else
		to_drain = pcp->count;
	free_pages_bulk(zone, to_drain, &pcp->list, 0);
	pcp->count -= to_drain;
	local_irq_restore(flags);
935 936 937
}
#endif

938 939 940 941 942 943 944 945
/*
 * Drain pages of the indicated processor.
 *
 * The processor must either be the current processor and the
 * thread pinned to the current processor or a processor that
 * is not online.
 */
static void drain_pages(unsigned int cpu)
L
Linus Torvalds 已提交
946
{
N
Nick Piggin 已提交
947
	unsigned long flags;
L
Linus Torvalds 已提交
948 949
	struct zone *zone;

950
	for_each_populated_zone(zone) {
L
Linus Torvalds 已提交
951
		struct per_cpu_pageset *pset;
952
		struct per_cpu_pages *pcp;
L
Linus Torvalds 已提交
953

954
		pset = zone_pcp(zone, cpu);
955 956 957 958 959 960

		pcp = &pset->pcp;
		local_irq_save(flags);
		free_pages_bulk(zone, pcp->count, &pcp->list, 0);
		pcp->count = 0;
		local_irq_restore(flags);
L
Linus Torvalds 已提交
961 962 963
	}
}

964 965 966 967 968 969 970 971 972 973 974 975 976
/*
 * Spill all of this CPU's per-cpu pages back into the buddy allocator.
 */
void drain_local_pages(void *arg)
{
	drain_pages(smp_processor_id());
}

/*
 * Spill all the per-cpu pages from all CPUs back into the buddy allocator
 */
void drain_all_pages(void)
{
977
	on_each_cpu(drain_local_pages, NULL, 1);
978 979
}

980
#ifdef CONFIG_HIBERNATION
L
Linus Torvalds 已提交
981 982 983

void mark_free_pages(struct zone *zone)
{
984 985
	unsigned long pfn, max_zone_pfn;
	unsigned long flags;
986
	int order, t;
L
Linus Torvalds 已提交
987 988 989 990 991 992
	struct list_head *curr;

	if (!zone->spanned_pages)
		return;

	spin_lock_irqsave(&zone->lock, flags);
993 994 995 996 997 998

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

999 1000
			if (!swsusp_page_is_forbidden(page))
				swsusp_unset_page_free(page);
1001
		}
L
Linus Torvalds 已提交
1002

1003 1004
	for_each_migratetype_order(order, t) {
		list_for_each(curr, &zone->free_area[order].free_list[t]) {
1005
			unsigned long i;
L
Linus Torvalds 已提交
1006

1007 1008
			pfn = page_to_pfn(list_entry(curr, struct page, lru));
			for (i = 0; i < (1UL << order); i++)
1009
				swsusp_set_page_free(pfn_to_page(pfn + i));
1010
		}
1011
	}
L
Linus Torvalds 已提交
1012 1013
	spin_unlock_irqrestore(&zone->lock, flags);
}
1014
#endif /* CONFIG_PM */
L
Linus Torvalds 已提交
1015 1016 1017 1018

/*
 * Free a 0-order page
 */
H
Harvey Harrison 已提交
1019
static void free_hot_cold_page(struct page *page, int cold)
L
Linus Torvalds 已提交
1020 1021 1022 1023
{
	struct zone *zone = page_zone(page);
	struct per_cpu_pages *pcp;
	unsigned long flags;
1024
	int wasMlocked = TestClearPageMlocked(page);
L
Linus Torvalds 已提交
1025

1026 1027
	kmemcheck_free_shadow(page, 0);

L
Linus Torvalds 已提交
1028 1029
	if (PageAnon(page))
		page->mapping = NULL;
N
Nick Piggin 已提交
1030
	if (free_pages_check(page))
1031 1032
		return;

1033
	if (!PageHighMem(page)) {
N
Nick Piggin 已提交
1034
		debug_check_no_locks_freed(page_address(page), PAGE_SIZE);
1035 1036
		debug_check_no_obj_freed(page_address(page), PAGE_SIZE);
	}
N
Nick Piggin 已提交
1037
	arch_free_page(page, 0);
1038 1039
	kernel_map_pages(page, 1, 0);

1040
	pcp = &zone_pcp(zone, get_cpu())->pcp;
1041
	set_page_private(page, get_pageblock_migratetype(page));
L
Linus Torvalds 已提交
1042
	local_irq_save(flags);
1043
	if (unlikely(wasMlocked))
1044
		free_page_mlock(page);
1045
	__count_vm_event(PGFREE);
1046

1047 1048 1049 1050
	if (cold)
		list_add_tail(&page->lru, &pcp->list);
	else
		list_add(&page->lru, &pcp->list);
L
Linus Torvalds 已提交
1051
	pcp->count++;
N
Nick Piggin 已提交
1052 1053 1054 1055
	if (pcp->count >= pcp->high) {
		free_pages_bulk(zone, pcp->batch, &pcp->list, 0);
		pcp->count -= pcp->batch;
	}
L
Linus Torvalds 已提交
1056 1057 1058 1059
	local_irq_restore(flags);
	put_cpu();
}

H
Harvey Harrison 已提交
1060
void free_hot_page(struct page *page)
L
Linus Torvalds 已提交
1061 1062 1063 1064
{
	free_hot_cold_page(page, 0);
}
	
H
Harvey Harrison 已提交
1065
void free_cold_page(struct page *page)
L
Linus Torvalds 已提交
1066 1067 1068 1069
{
	free_hot_cold_page(page, 1);
}

N
Nick Piggin 已提交
1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081
/*
 * 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;

N
Nick Piggin 已提交
1082 1083
	VM_BUG_ON(PageCompound(page));
	VM_BUG_ON(!page_count(page));
1084 1085 1086 1087 1088 1089 1090 1091 1092 1093

#ifdef CONFIG_KMEMCHECK
	/*
	 * Split shadow pages too, because free(page[0]) would
	 * otherwise free the whole shadow.
	 */
	if (kmemcheck_page_is_tracked(page))
		split_page(virt_to_page(page[0].shadow), order);
#endif

1094 1095
	for (i = 1; i < (1 << order); i++)
		set_page_refcounted(page + i);
N
Nick Piggin 已提交
1096 1097
}

L
Linus Torvalds 已提交
1098 1099 1100 1101 1102
/*
 * 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.
 */
1103 1104
static inline
struct page *buffered_rmqueue(struct zone *preferred_zone,
1105 1106
			struct zone *zone, int order, gfp_t gfp_flags,
			int migratetype)
L
Linus Torvalds 已提交
1107 1108
{
	unsigned long flags;
1109
	struct page *page;
L
Linus Torvalds 已提交
1110
	int cold = !!(gfp_flags & __GFP_COLD);
N
Nick Piggin 已提交
1111
	int cpu;
L
Linus Torvalds 已提交
1112

1113
again:
N
Nick Piggin 已提交
1114
	cpu  = get_cpu();
N
Nick Piggin 已提交
1115
	if (likely(order == 0)) {
L
Linus Torvalds 已提交
1116 1117
		struct per_cpu_pages *pcp;

1118
		pcp = &zone_pcp(zone, cpu)->pcp;
L
Linus Torvalds 已提交
1119
		local_irq_save(flags);
N
Nick Piggin 已提交
1120
		if (!pcp->count) {
1121
			pcp->count = rmqueue_bulk(zone, 0,
1122
					pcp->batch, &pcp->list, migratetype);
N
Nick Piggin 已提交
1123 1124
			if (unlikely(!pcp->count))
				goto failed;
L
Linus Torvalds 已提交
1125
		}
1126

1127
		/* Find a page of the appropriate migrate type */
1128 1129 1130 1131 1132 1133 1134 1135 1136
		if (cold) {
			list_for_each_entry_reverse(page, &pcp->list, lru)
				if (page_private(page) == migratetype)
					break;
		} else {
			list_for_each_entry(page, &pcp->list, lru)
				if (page_private(page) == migratetype)
					break;
		}
1137

1138 1139
		/* Allocate more to the pcp list if necessary */
		if (unlikely(&page->lru == &pcp->list)) {
1140 1141 1142 1143
			pcp->count += rmqueue_bulk(zone, 0,
					pcp->batch, &pcp->list, migratetype);
			page = list_entry(pcp->list.next, struct page, lru);
		}
1144 1145 1146

		list_del(&page->lru);
		pcp->count--;
R
Rohit Seth 已提交
1147
	} else {
1148 1149 1150 1151 1152 1153 1154 1155
		if (unlikely(gfp_flags & __GFP_NOFAIL)) {
			/*
			 * __GFP_NOFAIL is not to be used in new code.
			 *
			 * All __GFP_NOFAIL callers should be fixed so that they
			 * properly detect and handle allocation failures.
			 *
			 * We most definitely don't want callers attempting to
1156
			 * allocate greater than order-1 page units with
1157 1158
			 * __GFP_NOFAIL.
			 */
1159
			WARN_ON_ONCE(order > 1);
1160
		}
L
Linus Torvalds 已提交
1161
		spin_lock_irqsave(&zone->lock, flags);
1162
		page = __rmqueue(zone, order, migratetype);
1163
		__mod_zone_page_state(zone, NR_FREE_PAGES, -(1 << order));
N
Nick Piggin 已提交
1164 1165 1166
		spin_unlock(&zone->lock);
		if (!page)
			goto failed;
L
Linus Torvalds 已提交
1167 1168
	}

1169
	__count_zone_vm_events(PGALLOC, zone, 1 << order);
1170
	zone_statistics(preferred_zone, zone);
N
Nick Piggin 已提交
1171 1172
	local_irq_restore(flags);
	put_cpu();
L
Linus Torvalds 已提交
1173

N
Nick Piggin 已提交
1174
	VM_BUG_ON(bad_range(zone, page));
N
Nick Piggin 已提交
1175
	if (prep_new_page(page, order, gfp_flags))
N
Nick Piggin 已提交
1176
		goto again;
L
Linus Torvalds 已提交
1177
	return page;
N
Nick Piggin 已提交
1178 1179 1180 1181 1182

failed:
	local_irq_restore(flags);
	put_cpu();
	return NULL;
L
Linus Torvalds 已提交
1183 1184
}

1185 1186 1187 1188 1189 1190 1191 1192 1193
/* The ALLOC_WMARK bits are used as an index to zone->watermark */
#define ALLOC_WMARK_MIN		WMARK_MIN
#define ALLOC_WMARK_LOW		WMARK_LOW
#define ALLOC_WMARK_HIGH	WMARK_HIGH
#define ALLOC_NO_WATERMARKS	0x04 /* don't check watermarks at all */

/* Mask to get the watermark bits */
#define ALLOC_WMARK_MASK	(ALLOC_NO_WATERMARKS-1)

1194 1195 1196
#define ALLOC_HARDER		0x10 /* try to alloc harder */
#define ALLOC_HIGH		0x20 /* __GFP_HIGH set */
#define ALLOC_CPUSET		0x40 /* check for correct cpuset */
R
Rohit Seth 已提交
1197

1198 1199 1200 1201 1202 1203 1204
#ifdef CONFIG_FAIL_PAGE_ALLOC

static struct fail_page_alloc_attr {
	struct fault_attr attr;

	u32 ignore_gfp_highmem;
	u32 ignore_gfp_wait;
1205
	u32 min_order;
1206 1207 1208 1209 1210

#ifdef CONFIG_FAULT_INJECTION_DEBUG_FS

	struct dentry *ignore_gfp_highmem_file;
	struct dentry *ignore_gfp_wait_file;
1211
	struct dentry *min_order_file;
1212 1213 1214 1215 1216

#endif /* CONFIG_FAULT_INJECTION_DEBUG_FS */

} fail_page_alloc = {
	.attr = FAULT_ATTR_INITIALIZER,
1217 1218
	.ignore_gfp_wait = 1,
	.ignore_gfp_highmem = 1,
1219
	.min_order = 1,
1220 1221 1222 1223 1224 1225 1226 1227 1228 1229
};

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

static int should_fail_alloc_page(gfp_t gfp_mask, unsigned int order)
{
1230 1231
	if (order < fail_page_alloc.min_order)
		return 0;
1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262
	if (gfp_mask & __GFP_NOFAIL)
		return 0;
	if (fail_page_alloc.ignore_gfp_highmem && (gfp_mask & __GFP_HIGHMEM))
		return 0;
	if (fail_page_alloc.ignore_gfp_wait && (gfp_mask & __GFP_WAIT))
		return 0;

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

#ifdef CONFIG_FAULT_INJECTION_DEBUG_FS

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

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

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

	fail_page_alloc.ignore_gfp_highmem_file =
		debugfs_create_bool("ignore-gfp-highmem", mode, dir,
				      &fail_page_alloc.ignore_gfp_highmem);
1263 1264 1265
	fail_page_alloc.min_order_file =
		debugfs_create_u32("min-order", mode, dir,
				   &fail_page_alloc.min_order);
1266 1267

	if (!fail_page_alloc.ignore_gfp_wait_file ||
1268 1269
            !fail_page_alloc.ignore_gfp_highmem_file ||
            !fail_page_alloc.min_order_file) {
1270 1271 1272
		err = -ENOMEM;
		debugfs_remove(fail_page_alloc.ignore_gfp_wait_file);
		debugfs_remove(fail_page_alloc.ignore_gfp_highmem_file);
1273
		debugfs_remove(fail_page_alloc.min_order_file);
1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292
		cleanup_fault_attr_dentries(&fail_page_alloc.attr);
	}

	return err;
}

late_initcall(fail_page_alloc_debugfs);

#endif /* CONFIG_FAULT_INJECTION_DEBUG_FS */

#else /* CONFIG_FAIL_PAGE_ALLOC */

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

#endif /* CONFIG_FAIL_PAGE_ALLOC */

L
Linus Torvalds 已提交
1293 1294 1295 1296 1297
/*
 * 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,
R
Rohit Seth 已提交
1298
		      int classzone_idx, int alloc_flags)
L
Linus Torvalds 已提交
1299 1300
{
	/* free_pages my go negative - that's OK */
1301 1302
	long min = mark;
	long free_pages = zone_page_state(z, NR_FREE_PAGES) - (1 << order) + 1;
L
Linus Torvalds 已提交
1303 1304
	int o;

R
Rohit Seth 已提交
1305
	if (alloc_flags & ALLOC_HIGH)
L
Linus Torvalds 已提交
1306
		min -= min / 2;
R
Rohit Seth 已提交
1307
	if (alloc_flags & ALLOC_HARDER)
L
Linus Torvalds 已提交
1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324
		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;
}

1325 1326 1327 1328 1329 1330
#ifdef CONFIG_NUMA
/*
 * zlc_setup - Setup for "zonelist cache".  Uses cached zone data to
 * skip over zones that are not allowed by the cpuset, or that have
 * been recently (in last second) found to be nearly full.  See further
 * comments in mmzone.h.  Reduces cache footprint of zonelist scans
S
Simon Arlott 已提交
1331
 * that have to skip over a lot of full or unallowed zones.
1332 1333 1334
 *
 * If the zonelist cache is present in the passed in zonelist, then
 * returns a pointer to the allowed node mask (either the current
1335
 * tasks mems_allowed, or node_states[N_HIGH_MEMORY].)
1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356
 *
 * If the zonelist cache is not available for this zonelist, does
 * nothing and returns NULL.
 *
 * If the fullzones BITMAP in the zonelist cache is stale (more than
 * a second since last zap'd) then we zap it out (clear its bits.)
 *
 * We hold off even calling zlc_setup, until after we've checked the
 * first zone in the zonelist, on the theory that most allocations will
 * be satisfied from that first zone, so best to examine that zone as
 * quickly as we can.
 */
static nodemask_t *zlc_setup(struct zonelist *zonelist, int alloc_flags)
{
	struct zonelist_cache *zlc;	/* cached zonelist speedup info */
	nodemask_t *allowednodes;	/* zonelist_cache approximation */

	zlc = zonelist->zlcache_ptr;
	if (!zlc)
		return NULL;

S
S.Caglar Onur 已提交
1357
	if (time_after(jiffies, zlc->last_full_zap + HZ)) {
1358 1359 1360 1361 1362 1363
		bitmap_zero(zlc->fullzones, MAX_ZONES_PER_ZONELIST);
		zlc->last_full_zap = jiffies;
	}

	allowednodes = !in_interrupt() && (alloc_flags & ALLOC_CPUSET) ?
					&cpuset_current_mems_allowed :
1364
					&node_states[N_HIGH_MEMORY];
1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389
	return allowednodes;
}

/*
 * Given 'z' scanning a zonelist, run a couple of quick checks to see
 * if it is worth looking at further for free memory:
 *  1) Check that the zone isn't thought to be full (doesn't have its
 *     bit set in the zonelist_cache fullzones BITMAP).
 *  2) Check that the zones node (obtained from the zonelist_cache
 *     z_to_n[] mapping) is allowed in the passed in allowednodes mask.
 * Return true (non-zero) if zone is worth looking at further, or
 * else return false (zero) if it is not.
 *
 * This check -ignores- the distinction between various watermarks,
 * such as GFP_HIGH, GFP_ATOMIC, PF_MEMALLOC, ...  If a zone is
 * found to be full for any variation of these watermarks, it will
 * be considered full for up to one second by all requests, unless
 * we are so low on memory on all allowed nodes that we are forced
 * into the second scan of the zonelist.
 *
 * In the second scan we ignore this zonelist cache and exactly
 * apply the watermarks to all zones, even it is slower to do so.
 * We are low on memory in the second scan, and should leave no stone
 * unturned looking for a free page.
 */
1390
static int zlc_zone_worth_trying(struct zonelist *zonelist, struct zoneref *z,
1391 1392 1393 1394 1395 1396 1397 1398 1399 1400
						nodemask_t *allowednodes)
{
	struct zonelist_cache *zlc;	/* cached zonelist speedup info */
	int i;				/* index of *z in zonelist zones */
	int n;				/* node that zone *z is on */

	zlc = zonelist->zlcache_ptr;
	if (!zlc)
		return 1;

1401
	i = z - zonelist->_zonerefs;
1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412
	n = zlc->z_to_n[i];

	/* This zone is worth trying if it is allowed but not full */
	return node_isset(n, *allowednodes) && !test_bit(i, zlc->fullzones);
}

/*
 * Given 'z' scanning a zonelist, set the corresponding bit in
 * zlc->fullzones, so that subsequent attempts to allocate a page
 * from that zone don't waste time re-examining it.
 */
1413
static void zlc_mark_zone_full(struct zonelist *zonelist, struct zoneref *z)
1414 1415 1416 1417 1418 1419 1420 1421
{
	struct zonelist_cache *zlc;	/* cached zonelist speedup info */
	int i;				/* index of *z in zonelist zones */

	zlc = zonelist->zlcache_ptr;
	if (!zlc)
		return;

1422
	i = z - zonelist->_zonerefs;
1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433

	set_bit(i, zlc->fullzones);
}

#else	/* CONFIG_NUMA */

static nodemask_t *zlc_setup(struct zonelist *zonelist, int alloc_flags)
{
	return NULL;
}

1434
static int zlc_zone_worth_trying(struct zonelist *zonelist, struct zoneref *z,
1435 1436 1437 1438 1439
				nodemask_t *allowednodes)
{
	return 1;
}

1440
static void zlc_mark_zone_full(struct zonelist *zonelist, struct zoneref *z)
1441 1442 1443 1444
{
}
#endif	/* CONFIG_NUMA */

R
Rohit Seth 已提交
1445
/*
1446
 * get_page_from_freelist goes through the zonelist trying to allocate
R
Rohit Seth 已提交
1447 1448 1449
 * a page.
 */
static struct page *
1450
get_page_from_freelist(gfp_t gfp_mask, nodemask_t *nodemask, unsigned int order,
1451
		struct zonelist *zonelist, int high_zoneidx, int alloc_flags,
1452
		struct zone *preferred_zone, int migratetype)
M
Martin Hicks 已提交
1453
{
1454
	struct zoneref *z;
R
Rohit Seth 已提交
1455
	struct page *page = NULL;
1456
	int classzone_idx;
1457
	struct zone *zone;
1458 1459 1460
	nodemask_t *allowednodes = NULL;/* zonelist_cache approximation */
	int zlc_active = 0;		/* set if using zonelist_cache */
	int did_zlc_setup = 0;		/* just call zlc_setup() one time */
1461

1462
	classzone_idx = zone_idx(preferred_zone);
1463
zonelist_scan:
R
Rohit Seth 已提交
1464
	/*
1465
	 * Scan zonelist, looking for a zone with enough free.
R
Rohit Seth 已提交
1466 1467
	 * See also cpuset_zone_allowed() comment in kernel/cpuset.c.
	 */
1468 1469
	for_each_zone_zonelist_nodemask(zone, z, zonelist,
						high_zoneidx, nodemask) {
1470 1471 1472
		if (NUMA_BUILD && zlc_active &&
			!zlc_zone_worth_trying(zonelist, z, allowednodes))
				continue;
R
Rohit Seth 已提交
1473
		if ((alloc_flags & ALLOC_CPUSET) &&
1474
			!cpuset_zone_allowed_softwall(zone, gfp_mask))
1475
				goto try_next_zone;
R
Rohit Seth 已提交
1476

1477
		BUILD_BUG_ON(ALLOC_NO_WATERMARKS < NR_WMARK);
R
Rohit Seth 已提交
1478
		if (!(alloc_flags & ALLOC_NO_WATERMARKS)) {
1479
			unsigned long mark;
1480 1481
			int ret;

1482
			mark = zone->watermark[alloc_flags & ALLOC_WMARK_MASK];
1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501
			if (zone_watermark_ok(zone, order, mark,
				    classzone_idx, alloc_flags))
				goto try_this_zone;

			if (zone_reclaim_mode == 0)
				goto this_zone_full;

			ret = zone_reclaim(zone, gfp_mask, order);
			switch (ret) {
			case ZONE_RECLAIM_NOSCAN:
				/* did not scan */
				goto try_next_zone;
			case ZONE_RECLAIM_FULL:
				/* scanned but unreclaimable */
				goto this_zone_full;
			default:
				/* did we reclaim enough */
				if (!zone_watermark_ok(zone, order, mark,
						classzone_idx, alloc_flags))
1502
					goto this_zone_full;
1503
			}
R
Rohit Seth 已提交
1504 1505
		}

1506
try_this_zone:
1507 1508
		page = buffered_rmqueue(preferred_zone, zone, order,
						gfp_mask, migratetype);
1509
		if (page)
R
Rohit Seth 已提交
1510
			break;
1511 1512 1513 1514
this_zone_full:
		if (NUMA_BUILD)
			zlc_mark_zone_full(zonelist, z);
try_next_zone:
1515
		if (NUMA_BUILD && !did_zlc_setup && nr_online_nodes > 1) {
1516 1517 1518 1519
			/*
			 * we do zlc_setup after the first zone is tried but only
			 * if there are multiple nodes make it worthwhile
			 */
1520 1521 1522 1523
			allowednodes = zlc_setup(zonelist, alloc_flags);
			zlc_active = 1;
			did_zlc_setup = 1;
		}
1524
	}
1525 1526 1527 1528 1529 1530

	if (unlikely(NUMA_BUILD && page == NULL && zlc_active)) {
		/* Disable zlc cache for second zonelist scan */
		zlc_active = 0;
		goto zonelist_scan;
	}
R
Rohit Seth 已提交
1531
	return page;
M
Martin Hicks 已提交
1532 1533
}

1534 1535 1536
static inline int
should_alloc_retry(gfp_t gfp_mask, unsigned int order,
				unsigned long pages_reclaimed)
L
Linus Torvalds 已提交
1537
{
1538 1539 1540
	/* Do not loop if specifically requested */
	if (gfp_mask & __GFP_NORETRY)
		return 0;
L
Linus Torvalds 已提交
1541

1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558
	/*
	 * In this implementation, order <= PAGE_ALLOC_COSTLY_ORDER
	 * means __GFP_NOFAIL, but that may not be true in other
	 * implementations.
	 */
	if (order <= PAGE_ALLOC_COSTLY_ORDER)
		return 1;

	/*
	 * For order > PAGE_ALLOC_COSTLY_ORDER, if __GFP_REPEAT is
	 * specified, then we retry until we no longer reclaim any pages
	 * (above), or we've reclaimed an order of pages at least as
	 * large as the allocation's order. In both cases, if the
	 * allocation still fails, we stop retrying.
	 */
	if (gfp_mask & __GFP_REPEAT && pages_reclaimed < (1 << order))
		return 1;
1559

1560 1561 1562 1563 1564 1565
	/*
	 * Don't let big-order allocations loop unless the caller
	 * explicitly requests that.
	 */
	if (gfp_mask & __GFP_NOFAIL)
		return 1;
L
Linus Torvalds 已提交
1566

1567 1568
	return 0;
}
1569

1570 1571 1572
static inline struct page *
__alloc_pages_may_oom(gfp_t gfp_mask, unsigned int order,
	struct zonelist *zonelist, enum zone_type high_zoneidx,
1573 1574
	nodemask_t *nodemask, struct zone *preferred_zone,
	int migratetype)
1575 1576 1577 1578 1579 1580
{
	struct page *page;

	/* Acquire the OOM killer lock for the zones in zonelist */
	if (!try_set_zone_oom(zonelist, gfp_mask)) {
		schedule_timeout_uninterruptible(1);
L
Linus Torvalds 已提交
1581 1582
		return NULL;
	}
1583

1584 1585 1586 1587 1588 1589 1590
	/*
	 * 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.
	 */
	page = get_page_from_freelist(gfp_mask|__GFP_HARDWALL, nodemask,
		order, zonelist, high_zoneidx,
1591
		ALLOC_WMARK_HIGH|ALLOC_CPUSET,
1592
		preferred_zone, migratetype);
R
Rohit Seth 已提交
1593
	if (page)
1594 1595 1596
		goto out;

	/* The OOM killer will not help higher order allocs */
1597
	if (order > PAGE_ALLOC_COSTLY_ORDER && !(gfp_mask & __GFP_NOFAIL))
1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611
		goto out;

	/* Exhausted what can be done so it's blamo time */
	out_of_memory(zonelist, gfp_mask, order);

out:
	clear_zonelist_oom(zonelist, gfp_mask);
	return page;
}

/* The really slow allocator path where we enter direct reclaim */
static inline struct page *
__alloc_pages_direct_reclaim(gfp_t gfp_mask, unsigned int order,
	struct zonelist *zonelist, enum zone_type high_zoneidx,
1612
	nodemask_t *nodemask, int alloc_flags, struct zone *preferred_zone,
1613
	int migratetype, unsigned long *did_some_progress)
1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644
{
	struct page *page = NULL;
	struct reclaim_state reclaim_state;
	struct task_struct *p = current;

	cond_resched();

	/* We now go into synchronous reclaim */
	cpuset_memory_pressure_bump();

	/*
	 * The task's cpuset might have expanded its set of allowable nodes
	 */
	p->flags |= PF_MEMALLOC;
	lockdep_set_current_reclaim_state(gfp_mask);
	reclaim_state.reclaimed_slab = 0;
	p->reclaim_state = &reclaim_state;

	*did_some_progress = try_to_free_pages(zonelist, order, gfp_mask, nodemask);

	p->reclaim_state = NULL;
	lockdep_clear_current_reclaim_state();
	p->flags &= ~PF_MEMALLOC;

	cond_resched();

	if (order != 0)
		drain_all_pages();

	if (likely(*did_some_progress))
		page = get_page_from_freelist(gfp_mask, nodemask, order,
1645
					zonelist, high_zoneidx,
1646 1647
					alloc_flags, preferred_zone,
					migratetype);
1648 1649 1650
	return page;
}

L
Linus Torvalds 已提交
1651
/*
1652 1653
 * This is called in the allocator slow-path if the allocation request is of
 * sufficient urgency to ignore watermarks and take other desperate measures
L
Linus Torvalds 已提交
1654
 */
1655 1656 1657
static inline struct page *
__alloc_pages_high_priority(gfp_t gfp_mask, unsigned int order,
	struct zonelist *zonelist, enum zone_type high_zoneidx,
1658 1659
	nodemask_t *nodemask, struct zone *preferred_zone,
	int migratetype)
1660 1661 1662 1663 1664
{
	struct page *page;

	do {
		page = get_page_from_freelist(gfp_mask, nodemask, order,
1665
			zonelist, high_zoneidx, ALLOC_NO_WATERMARKS,
1666
			preferred_zone, migratetype);
1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677

		if (!page && gfp_mask & __GFP_NOFAIL)
			congestion_wait(WRITE, HZ/50);
	} while (!page && (gfp_mask & __GFP_NOFAIL));

	return page;
}

static inline
void wake_all_kswapd(unsigned int order, struct zonelist *zonelist,
						enum zone_type high_zoneidx)
L
Linus Torvalds 已提交
1678
{
1679 1680
	struct zoneref *z;
	struct zone *zone;
L
Linus Torvalds 已提交
1681

1682 1683 1684
	for_each_zone_zonelist(zone, z, zonelist, high_zoneidx)
		wakeup_kswapd(zone, order);
}
1685

1686 1687 1688 1689 1690 1691
static inline int
gfp_to_alloc_flags(gfp_t gfp_mask)
{
	struct task_struct *p = current;
	int alloc_flags = ALLOC_WMARK_MIN | ALLOC_CPUSET;
	const gfp_t wait = gfp_mask & __GFP_WAIT;
L
Linus Torvalds 已提交
1692

1693 1694
	/* __GFP_HIGH is assumed to be the same as ALLOC_HIGH to save a branch. */
	BUILD_BUG_ON(__GFP_HIGH != ALLOC_HIGH);
1695

1696 1697 1698 1699 1700 1701
	/*
	 * The caller may dip into page reserves a bit more if the caller
	 * cannot run direct reclaim, or if the caller has realtime scheduling
	 * policy or is asking for __GFP_HIGH memory.  GFP_ATOMIC requests will
	 * set both ALLOC_HARDER (!wait) and ALLOC_HIGH (__GFP_HIGH).
	 */
1702
	alloc_flags |= (gfp_mask & __GFP_HIGH);
L
Linus Torvalds 已提交
1703

1704 1705
	if (!wait) {
		alloc_flags |= ALLOC_HARDER;
1706
		/*
1707 1708
		 * Ignore cpuset if GFP_ATOMIC (!wait) rather than fail alloc.
		 * See also cpuset_zone_allowed() comment in kernel/cpuset.c.
1709
		 */
1710 1711 1712 1713 1714 1715 1716 1717 1718
		alloc_flags &= ~ALLOC_CPUSET;
	} else if (unlikely(rt_task(p)))
		alloc_flags |= ALLOC_HARDER;

	if (likely(!(gfp_mask & __GFP_NOMEMALLOC))) {
		if (!in_interrupt() &&
		    ((p->flags & PF_MEMALLOC) ||
		     unlikely(test_thread_flag(TIF_MEMDIE))))
			alloc_flags |= ALLOC_NO_WATERMARKS;
L
Linus Torvalds 已提交
1719
	}
1720

1721 1722 1723
	return alloc_flags;
}

1724 1725 1726
static inline struct page *
__alloc_pages_slowpath(gfp_t gfp_mask, unsigned int order,
	struct zonelist *zonelist, enum zone_type high_zoneidx,
1727 1728
	nodemask_t *nodemask, struct zone *preferred_zone,
	int migratetype)
1729 1730 1731 1732 1733 1734 1735
{
	const gfp_t wait = gfp_mask & __GFP_WAIT;
	struct page *page = NULL;
	int alloc_flags;
	unsigned long pages_reclaimed = 0;
	unsigned long did_some_progress;
	struct task_struct *p = current;
L
Linus Torvalds 已提交
1736

1737 1738 1739 1740 1741 1742 1743 1744
	/*
	 * In the slowpath, we sanity check order to avoid ever trying to
	 * reclaim >= MAX_ORDER areas which will never succeed. Callers may
	 * be using allocators in order of preference for an area that is
	 * too large.
	 */
	if (WARN_ON_ONCE(order >= MAX_ORDER))
		return NULL;
L
Linus Torvalds 已提交
1745

1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756
	/*
	 * GFP_THISNODE (meaning __GFP_THISNODE, __GFP_NORETRY and
	 * __GFP_NOWARN set) should not cause reclaim since the subsystem
	 * (f.e. slab) using GFP_THISNODE may choose to trigger reclaim
	 * using a larger set of nodes after it has established that the
	 * allowed per node queues are empty and that nodes are
	 * over allocated.
	 */
	if (NUMA_BUILD && (gfp_mask & GFP_THISNODE) == GFP_THISNODE)
		goto nopage;

1757
	wake_all_kswapd(order, zonelist, high_zoneidx);
L
Linus Torvalds 已提交
1758

1759
	/*
R
Rohit Seth 已提交
1760 1761 1762
	 * 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.
1763
	 */
1764
	alloc_flags = gfp_to_alloc_flags(gfp_mask);
L
Linus Torvalds 已提交
1765

1766
restart:
1767
	/* This is the last chance, in general, before the goto nopage. */
1768
	page = get_page_from_freelist(gfp_mask, nodemask, order, zonelist,
1769 1770
			high_zoneidx, alloc_flags & ~ALLOC_NO_WATERMARKS,
			preferred_zone, migratetype);
R
Rohit Seth 已提交
1771 1772
	if (page)
		goto got_pg;
L
Linus Torvalds 已提交
1773

1774
rebalance:
1775
	/* Allocate without watermarks if the context allows */
1776 1777 1778 1779 1780 1781
	if (alloc_flags & ALLOC_NO_WATERMARKS) {
		page = __alloc_pages_high_priority(gfp_mask, order,
				zonelist, high_zoneidx, nodemask,
				preferred_zone, migratetype);
		if (page)
			goto got_pg;
L
Linus Torvalds 已提交
1782 1783 1784 1785 1786 1787
	}

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

1788 1789 1790 1791
	/* Avoid recursion of direct reclaim */
	if (p->flags & PF_MEMALLOC)
		goto nopage;

1792 1793 1794 1795
	/* Try direct reclaim and then allocating */
	page = __alloc_pages_direct_reclaim(gfp_mask, order,
					zonelist, high_zoneidx,
					nodemask,
1796
					alloc_flags, preferred_zone,
1797
					migratetype, &did_some_progress);
1798 1799
	if (page)
		goto got_pg;
L
Linus Torvalds 已提交
1800

1801
	/*
1802 1803
	 * If we failed to make any progress reclaiming, then we are
	 * running out of options and have to consider going OOM
1804
	 */
1805 1806
	if (!did_some_progress) {
		if ((gfp_mask & __GFP_FS) && !(gfp_mask & __GFP_NORETRY)) {
1807 1808
			if (oom_killer_disabled)
				goto nopage;
1809 1810
			page = __alloc_pages_may_oom(gfp_mask, order,
					zonelist, high_zoneidx,
1811 1812
					nodemask, preferred_zone,
					migratetype);
1813 1814
			if (page)
				goto got_pg;
L
Linus Torvalds 已提交
1815

1816
			/*
1817 1818 1819 1820
			 * The OOM killer does not trigger for high-order
			 * ~__GFP_NOFAIL allocations so if no progress is being
			 * made, there are no other options and retrying is
			 * unlikely to help.
1821
			 */
1822 1823
			if (order > PAGE_ALLOC_COSTLY_ORDER &&
						!(gfp_mask & __GFP_NOFAIL))
1824
				goto nopage;
1825

1826 1827
			goto restart;
		}
L
Linus Torvalds 已提交
1828 1829
	}

1830
	/* Check if we should retry the allocation */
1831
	pages_reclaimed += did_some_progress;
1832 1833
	if (should_alloc_retry(gfp_mask, order, pages_reclaimed)) {
		/* Wait for some write requests to complete then retry */
1834
		congestion_wait(WRITE, HZ/50);
L
Linus Torvalds 已提交
1835 1836 1837 1838 1839 1840 1841 1842 1843
		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();
J
Janet Morgan 已提交
1844
		show_mem();
L
Linus Torvalds 已提交
1845
	}
1846
	return page;
L
Linus Torvalds 已提交
1847
got_pg:
1848 1849
	if (kmemcheck_enabled)
		kmemcheck_pagealloc_alloc(page, order, gfp_mask);
L
Linus Torvalds 已提交
1850
	return page;
1851

L
Linus Torvalds 已提交
1852
}
1853 1854 1855 1856 1857 1858 1859 1860 1861

/*
 * This is the 'heart' of the zoned buddy allocator.
 */
struct page *
__alloc_pages_nodemask(gfp_t gfp_mask, unsigned int order,
			struct zonelist *zonelist, nodemask_t *nodemask)
{
	enum zone_type high_zoneidx = gfp_zone(gfp_mask);
1862
	struct zone *preferred_zone;
1863
	struct page *page;
1864
	int migratetype = allocflags_to_migratetype(gfp_mask);
1865

1866 1867
	gfp_mask &= gfp_allowed_mask;

1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882
	lockdep_trace_alloc(gfp_mask);

	might_sleep_if(gfp_mask & __GFP_WAIT);

	if (should_fail_alloc_page(gfp_mask, order))
		return NULL;

	/*
	 * Check the zones suitable for the gfp_mask contain at least one
	 * valid zone. It's possible to have an empty zonelist as a result
	 * of GFP_THISNODE and a memoryless node
	 */
	if (unlikely(!zonelist->_zonerefs->zone))
		return NULL;

1883 1884 1885 1886 1887 1888
	/* The preferred zone is used for statistics later */
	first_zones_zonelist(zonelist, high_zoneidx, nodemask, &preferred_zone);
	if (!preferred_zone)
		return NULL;

	/* First allocation attempt */
1889
	page = get_page_from_freelist(gfp_mask|__GFP_HARDWALL, nodemask, order,
1890
			zonelist, high_zoneidx, ALLOC_WMARK_LOW|ALLOC_CPUSET,
1891
			preferred_zone, migratetype);
1892 1893
	if (unlikely(!page))
		page = __alloc_pages_slowpath(gfp_mask, order,
1894
				zonelist, high_zoneidx, nodemask,
1895
				preferred_zone, migratetype);
1896 1897

	return page;
L
Linus Torvalds 已提交
1898
}
1899
EXPORT_SYMBOL(__alloc_pages_nodemask);
L
Linus Torvalds 已提交
1900 1901 1902 1903

/*
 * Common helper functions.
 */
H
Harvey Harrison 已提交
1904
unsigned long __get_free_pages(gfp_t gfp_mask, unsigned int order)
L
Linus Torvalds 已提交
1905 1906 1907 1908 1909 1910 1911 1912 1913 1914
{
	struct page * page;
	page = alloc_pages(gfp_mask, order);
	if (!page)
		return 0;
	return (unsigned long) page_address(page);
}

EXPORT_SYMBOL(__get_free_pages);

H
Harvey Harrison 已提交
1915
unsigned long get_zeroed_page(gfp_t gfp_mask)
L
Linus Torvalds 已提交
1916 1917 1918 1919 1920 1921 1922
{
	struct page * page;

	/*
	 * get_zeroed_page() returns a 32-bit address, which cannot represent
	 * a highmem page
	 */
N
Nick Piggin 已提交
1923
	VM_BUG_ON((gfp_mask & __GFP_HIGHMEM) != 0);
L
Linus Torvalds 已提交
1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940

	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);
}

H
Harvey Harrison 已提交
1941
void __free_pages(struct page *page, unsigned int order)
L
Linus Torvalds 已提交
1942
{
N
Nick Piggin 已提交
1943
	if (put_page_testzero(page)) {
L
Linus Torvalds 已提交
1944 1945 1946 1947 1948 1949 1950 1951 1952
		if (order == 0)
			free_hot_page(page);
		else
			__free_pages_ok(page, order);
	}
}

EXPORT_SYMBOL(__free_pages);

H
Harvey Harrison 已提交
1953
void free_pages(unsigned long addr, unsigned int order)
L
Linus Torvalds 已提交
1954 1955
{
	if (addr != 0) {
N
Nick Piggin 已提交
1956
		VM_BUG_ON(!virt_addr_valid((void *)addr));
L
Linus Torvalds 已提交
1957 1958 1959 1960 1961 1962
		__free_pages(virt_to_page((void *)addr), order);
	}
}

EXPORT_SYMBOL(free_pages);

1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
/**
 * alloc_pages_exact - allocate an exact number physically-contiguous pages.
 * @size: the number of bytes to allocate
 * @gfp_mask: GFP flags for the allocation
 *
 * This function is similar to alloc_pages(), except that it allocates the
 * minimum number of pages to satisfy the request.  alloc_pages() can only
 * allocate memory in power-of-two pages.
 *
 * This function is also limited by MAX_ORDER.
 *
 * Memory allocated by this function must be released by free_pages_exact().
 */
void *alloc_pages_exact(size_t size, gfp_t gfp_mask)
{
	unsigned int order = get_order(size);
	unsigned long addr;

	addr = __get_free_pages(gfp_mask, order);
	if (addr) {
		unsigned long alloc_end = addr + (PAGE_SIZE << order);
		unsigned long used = addr + PAGE_ALIGN(size);

		split_page(virt_to_page(addr), order);
		while (used < alloc_end) {
			free_page(used);
			used += PAGE_SIZE;
		}
	}

	return (void *)addr;
}
EXPORT_SYMBOL(alloc_pages_exact);

/**
 * free_pages_exact - release memory allocated via alloc_pages_exact()
 * @virt: the value returned by alloc_pages_exact.
 * @size: size of allocation, same value as passed to alloc_pages_exact().
 *
 * Release the memory allocated by a previous call to alloc_pages_exact.
 */
void free_pages_exact(void *virt, size_t size)
{
	unsigned long addr = (unsigned long)virt;
	unsigned long end = addr + PAGE_ALIGN(size);

	while (addr < end) {
		free_page(addr);
		addr += PAGE_SIZE;
	}
}
EXPORT_SYMBOL(free_pages_exact);

L
Linus Torvalds 已提交
2016 2017
static unsigned int nr_free_zone_pages(int offset)
{
2018
	struct zoneref *z;
2019 2020
	struct zone *zone;

2021
	/* Just pick one node, since fallback list is circular */
L
Linus Torvalds 已提交
2022 2023
	unsigned int sum = 0;

2024
	struct zonelist *zonelist = node_zonelist(numa_node_id(), GFP_KERNEL);
L
Linus Torvalds 已提交
2025

2026
	for_each_zone_zonelist(zone, z, zonelist, offset) {
2027
		unsigned long size = zone->present_pages;
2028
		unsigned long high = high_wmark_pages(zone);
2029 2030
		if (size > high)
			sum += size - high;
L
Linus Torvalds 已提交
2031 2032 2033 2034 2035 2036 2037 2038 2039 2040
	}

	return sum;
}

/*
 * Amount of free RAM allocatable within ZONE_DMA and ZONE_NORMAL
 */
unsigned int nr_free_buffer_pages(void)
{
A
Al Viro 已提交
2041
	return nr_free_zone_pages(gfp_zone(GFP_USER));
L
Linus Torvalds 已提交
2042
}
2043
EXPORT_SYMBOL_GPL(nr_free_buffer_pages);
L
Linus Torvalds 已提交
2044 2045 2046 2047 2048 2049

/*
 * Amount of free RAM allocatable within all zones
 */
unsigned int nr_free_pagecache_pages(void)
{
M
Mel Gorman 已提交
2050
	return nr_free_zone_pages(gfp_zone(GFP_HIGHUSER_MOVABLE));
L
Linus Torvalds 已提交
2051
}
2052 2053

static inline void show_node(struct zone *zone)
L
Linus Torvalds 已提交
2054
{
2055
	if (NUMA_BUILD)
2056
		printk("Node %d ", zone_to_nid(zone));
L
Linus Torvalds 已提交
2057 2058 2059 2060 2061 2062
}

void si_meminfo(struct sysinfo *val)
{
	val->totalram = totalram_pages;
	val->sharedram = 0;
2063
	val->freeram = global_page_state(NR_FREE_PAGES);
L
Linus Torvalds 已提交
2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077
	val->bufferram = nr_blockdev_pages();
	val->totalhigh = totalhigh_pages;
	val->freehigh = nr_free_highpages();
	val->mem_unit = PAGE_SIZE;
}

EXPORT_SYMBOL(si_meminfo);

#ifdef CONFIG_NUMA
void si_meminfo_node(struct sysinfo *val, int nid)
{
	pg_data_t *pgdat = NODE_DATA(nid);

	val->totalram = pgdat->node_present_pages;
2078
	val->freeram = node_page_state(nid, NR_FREE_PAGES);
2079
#ifdef CONFIG_HIGHMEM
L
Linus Torvalds 已提交
2080
	val->totalhigh = pgdat->node_zones[ZONE_HIGHMEM].present_pages;
2081 2082
	val->freehigh = zone_page_state(&pgdat->node_zones[ZONE_HIGHMEM],
			NR_FREE_PAGES);
2083 2084 2085 2086
#else
	val->totalhigh = 0;
	val->freehigh = 0;
#endif
L
Linus Torvalds 已提交
2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099
	val->mem_unit = PAGE_SIZE;
}
#endif

#define K(x) ((x) << (PAGE_SHIFT-10))

/*
 * Show free area list (used inside shift_scroll-lock stuff)
 * We also calculate the percentage fragmentation. We do this by counting the
 * memory on each free list with the exception of the first item on the list.
 */
void show_free_areas(void)
{
2100
	int cpu;
L
Linus Torvalds 已提交
2101 2102
	struct zone *zone;

2103
	for_each_populated_zone(zone) {
2104 2105
		show_node(zone);
		printk("%s per-cpu:\n", zone->name);
L
Linus Torvalds 已提交
2106

2107
		for_each_online_cpu(cpu) {
L
Linus Torvalds 已提交
2108 2109
			struct per_cpu_pageset *pageset;

2110
			pageset = zone_pcp(zone, cpu);
L
Linus Torvalds 已提交
2111

2112 2113 2114
			printk("CPU %4d: hi:%5d, btch:%4d usd:%4d\n",
			       cpu, pageset->pcp.high,
			       pageset->pcp.batch, pageset->pcp.count);
L
Linus Torvalds 已提交
2115 2116 2117
		}
	}

L
Lee Schermerhorn 已提交
2118 2119 2120 2121
	printk("Active_anon:%lu active_file:%lu inactive_anon:%lu\n"
		" inactive_file:%lu"
		" unevictable:%lu"
		" dirty:%lu writeback:%lu unstable:%lu\n"
2122
		" free:%lu slab:%lu mapped:%lu pagetables:%lu bounce:%lu\n",
2123 2124 2125 2126
		global_page_state(NR_ACTIVE_ANON),
		global_page_state(NR_ACTIVE_FILE),
		global_page_state(NR_INACTIVE_ANON),
		global_page_state(NR_INACTIVE_FILE),
L
Lee Schermerhorn 已提交
2127
		global_page_state(NR_UNEVICTABLE),
2128
		global_page_state(NR_FILE_DIRTY),
2129
		global_page_state(NR_WRITEBACK),
2130
		global_page_state(NR_UNSTABLE_NFS),
2131
		global_page_state(NR_FREE_PAGES),
2132 2133
		global_page_state(NR_SLAB_RECLAIMABLE) +
			global_page_state(NR_SLAB_UNRECLAIMABLE),
2134
		global_page_state(NR_FILE_MAPPED),
2135 2136
		global_page_state(NR_PAGETABLE),
		global_page_state(NR_BOUNCE));
L
Linus Torvalds 已提交
2137

2138
	for_each_populated_zone(zone) {
L
Linus Torvalds 已提交
2139 2140 2141 2142 2143 2144 2145 2146
		int i;

		show_node(zone);
		printk("%s"
			" free:%lukB"
			" min:%lukB"
			" low:%lukB"
			" high:%lukB"
2147 2148 2149 2150
			" active_anon:%lukB"
			" inactive_anon:%lukB"
			" active_file:%lukB"
			" inactive_file:%lukB"
L
Lee Schermerhorn 已提交
2151
			" unevictable:%lukB"
L
Linus Torvalds 已提交
2152 2153 2154 2155 2156
			" present:%lukB"
			" pages_scanned:%lu"
			" all_unreclaimable? %s"
			"\n",
			zone->name,
2157
			K(zone_page_state(zone, NR_FREE_PAGES)),
2158 2159 2160
			K(min_wmark_pages(zone)),
			K(low_wmark_pages(zone)),
			K(high_wmark_pages(zone)),
2161 2162 2163 2164
			K(zone_page_state(zone, NR_ACTIVE_ANON)),
			K(zone_page_state(zone, NR_INACTIVE_ANON)),
			K(zone_page_state(zone, NR_ACTIVE_FILE)),
			K(zone_page_state(zone, NR_INACTIVE_FILE)),
L
Lee Schermerhorn 已提交
2165
			K(zone_page_state(zone, NR_UNEVICTABLE)),
L
Linus Torvalds 已提交
2166 2167
			K(zone->present_pages),
			zone->pages_scanned,
2168
			(zone_is_all_unreclaimable(zone) ? "yes" : "no")
L
Linus Torvalds 已提交
2169 2170 2171 2172 2173 2174 2175
			);
		printk("lowmem_reserve[]:");
		for (i = 0; i < MAX_NR_ZONES; i++)
			printk(" %lu", zone->lowmem_reserve[i]);
		printk("\n");
	}

2176
	for_each_populated_zone(zone) {
2177
 		unsigned long nr[MAX_ORDER], flags, order, total = 0;
L
Linus Torvalds 已提交
2178 2179 2180 2181 2182 2183

		show_node(zone);
		printk("%s: ", zone->name);

		spin_lock_irqsave(&zone->lock, flags);
		for (order = 0; order < MAX_ORDER; order++) {
2184 2185
			nr[order] = zone->free_area[order].nr_free;
			total += nr[order] << order;
L
Linus Torvalds 已提交
2186 2187
		}
		spin_unlock_irqrestore(&zone->lock, flags);
2188 2189
		for (order = 0; order < MAX_ORDER; order++)
			printk("%lu*%lukB ", nr[order], K(1UL) << order);
L
Linus Torvalds 已提交
2190 2191 2192
		printk("= %lukB\n", K(total));
	}

2193 2194
	printk("%ld total pagecache pages\n", global_page_state(NR_FILE_PAGES));

L
Linus Torvalds 已提交
2195 2196 2197
	show_swap_cache_info();
}

2198 2199 2200 2201 2202 2203
static void zoneref_set_zone(struct zone *zone, struct zoneref *zoneref)
{
	zoneref->zone = zone;
	zoneref->zone_idx = zone_idx(zone);
}

L
Linus Torvalds 已提交
2204 2205
/*
 * Builds allocation fallback zone lists.
2206 2207
 *
 * Add all populated zones of a node to the zonelist.
L
Linus Torvalds 已提交
2208
 */
2209 2210
static int build_zonelists_node(pg_data_t *pgdat, struct zonelist *zonelist,
				int nr_zones, enum zone_type zone_type)
L
Linus Torvalds 已提交
2211
{
2212 2213
	struct zone *zone;

2214
	BUG_ON(zone_type >= MAX_NR_ZONES);
2215
	zone_type++;
2216 2217

	do {
2218
		zone_type--;
2219
		zone = pgdat->node_zones + zone_type;
2220
		if (populated_zone(zone)) {
2221 2222
			zoneref_set_zone(zone,
				&zonelist->_zonerefs[nr_zones++]);
2223
			check_highest_zone(zone_type);
L
Linus Torvalds 已提交
2224
		}
2225

2226
	} while (zone_type);
2227
	return nr_zones;
L
Linus Torvalds 已提交
2228 2229
}

2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250

/*
 *  zonelist_order:
 *  0 = automatic detection of better ordering.
 *  1 = order by ([node] distance, -zonetype)
 *  2 = order by (-zonetype, [node] distance)
 *
 *  If not NUMA, ZONELIST_ORDER_ZONE and ZONELIST_ORDER_NODE will create
 *  the same zonelist. So only NUMA can configure this param.
 */
#define ZONELIST_ORDER_DEFAULT  0
#define ZONELIST_ORDER_NODE     1
#define ZONELIST_ORDER_ZONE     2

/* zonelist order in the kernel.
 * set_zonelist_order() will set this to NODE or ZONE.
 */
static int current_zonelist_order = ZONELIST_ORDER_DEFAULT;
static char zonelist_order_name[3][8] = {"Default", "Node", "Zone"};


L
Linus Torvalds 已提交
2251
#ifdef CONFIG_NUMA
2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322
/* The value user specified ....changed by config */
static int user_zonelist_order = ZONELIST_ORDER_DEFAULT;
/* string for sysctl */
#define NUMA_ZONELIST_ORDER_LEN	16
char numa_zonelist_order[16] = "default";

/*
 * interface for configure zonelist ordering.
 * command line option "numa_zonelist_order"
 *	= "[dD]efault	- default, automatic configuration.
 *	= "[nN]ode 	- order by node locality, then by zone within node
 *	= "[zZ]one      - order by zone, then by locality within zone
 */

static int __parse_numa_zonelist_order(char *s)
{
	if (*s == 'd' || *s == 'D') {
		user_zonelist_order = ZONELIST_ORDER_DEFAULT;
	} else if (*s == 'n' || *s == 'N') {
		user_zonelist_order = ZONELIST_ORDER_NODE;
	} else if (*s == 'z' || *s == 'Z') {
		user_zonelist_order = ZONELIST_ORDER_ZONE;
	} else {
		printk(KERN_WARNING
			"Ignoring invalid numa_zonelist_order value:  "
			"%s\n", s);
		return -EINVAL;
	}
	return 0;
}

static __init int setup_numa_zonelist_order(char *s)
{
	if (s)
		return __parse_numa_zonelist_order(s);
	return 0;
}
early_param("numa_zonelist_order", setup_numa_zonelist_order);

/*
 * sysctl handler for numa_zonelist_order
 */
int numa_zonelist_order_handler(ctl_table *table, int write,
		struct file *file, void __user *buffer, size_t *length,
		loff_t *ppos)
{
	char saved_string[NUMA_ZONELIST_ORDER_LEN];
	int ret;

	if (write)
		strncpy(saved_string, (char*)table->data,
			NUMA_ZONELIST_ORDER_LEN);
	ret = proc_dostring(table, write, file, buffer, length, ppos);
	if (ret)
		return ret;
	if (write) {
		int oldval = user_zonelist_order;
		if (__parse_numa_zonelist_order((char*)table->data)) {
			/*
			 * bogus value.  restore saved string
			 */
			strncpy((char*)table->data, saved_string,
				NUMA_ZONELIST_ORDER_LEN);
			user_zonelist_order = oldval;
		} else if (oldval != user_zonelist_order)
			build_all_zonelists();
	}
	return 0;
}


2323
#define MAX_NODE_LOAD (nr_online_nodes)
2324 2325
static int node_load[MAX_NUMNODES];

L
Linus Torvalds 已提交
2326
/**
2327
 * find_next_best_node - find the next node that should appear in a given node's fallback list
L
Linus Torvalds 已提交
2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339
 * @node: node whose fallback list we're appending
 * @used_node_mask: nodemask_t of already used nodes
 *
 * We use a number of factors to determine which is the next node that should
 * appear on a given node's fallback list.  The node should not have appeared
 * already in @node's fallback list, and it should be the next closest node
 * according to the distance array (which contains arbitrary distance values
 * from each node to each node in the system), and should also prefer nodes
 * with no CPUs, since presumably they'll have very little allocation pressure
 * on them otherwise.
 * It returns -1 if no node is found.
 */
2340
static int find_next_best_node(int node, nodemask_t *used_node_mask)
L
Linus Torvalds 已提交
2341
{
2342
	int n, val;
L
Linus Torvalds 已提交
2343 2344
	int min_val = INT_MAX;
	int best_node = -1;
2345
	const struct cpumask *tmp = cpumask_of_node(0);
L
Linus Torvalds 已提交
2346

2347 2348 2349 2350 2351
	/* Use the local node if we haven't already */
	if (!node_isset(node, *used_node_mask)) {
		node_set(node, *used_node_mask);
		return node;
	}
L
Linus Torvalds 已提交
2352

2353
	for_each_node_state(n, N_HIGH_MEMORY) {
L
Linus Torvalds 已提交
2354 2355 2356 2357 2358 2359 2360 2361

		/* Don't want a node to appear more than once */
		if (node_isset(n, *used_node_mask))
			continue;

		/* Use the distance array to find the distance */
		val = node_distance(node, n);

2362 2363 2364
		/* Penalize nodes under us ("prefer the next node") */
		val += (n < node);

L
Linus Torvalds 已提交
2365
		/* Give preference to headless and unused nodes */
2366 2367
		tmp = cpumask_of_node(n);
		if (!cpumask_empty(tmp))
L
Linus Torvalds 已提交
2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385
			val += PENALTY_FOR_NODE_WITH_CPUS;

		/* Slight preference for less loaded node */
		val *= (MAX_NODE_LOAD*MAX_NUMNODES);
		val += node_load[n];

		if (val < min_val) {
			min_val = val;
			best_node = n;
		}
	}

	if (best_node >= 0)
		node_set(best_node, *used_node_mask);

	return best_node;
}

2386 2387 2388 2389 2390 2391 2392

/*
 * Build zonelists ordered by node and zones within node.
 * This results in maximum locality--normal zone overflows into local
 * DMA zone, if any--but risks exhausting DMA zone.
 */
static void build_zonelists_in_node_order(pg_data_t *pgdat, int node)
L
Linus Torvalds 已提交
2393
{
2394
	int j;
L
Linus Torvalds 已提交
2395
	struct zonelist *zonelist;
2396

2397
	zonelist = &pgdat->node_zonelists[0];
2398
	for (j = 0; zonelist->_zonerefs[j].zone != NULL; j++)
2399 2400 2401
		;
	j = build_zonelists_node(NODE_DATA(node), zonelist, j,
							MAX_NR_ZONES - 1);
2402 2403
	zonelist->_zonerefs[j].zone = NULL;
	zonelist->_zonerefs[j].zone_idx = 0;
2404 2405
}

2406 2407 2408 2409 2410 2411 2412 2413
/*
 * Build gfp_thisnode zonelists
 */
static void build_thisnode_zonelists(pg_data_t *pgdat)
{
	int j;
	struct zonelist *zonelist;

2414 2415
	zonelist = &pgdat->node_zonelists[1];
	j = build_zonelists_node(pgdat, zonelist, 0, MAX_NR_ZONES - 1);
2416 2417
	zonelist->_zonerefs[j].zone = NULL;
	zonelist->_zonerefs[j].zone_idx = 0;
2418 2419
}

2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434
/*
 * Build zonelists ordered by zone and nodes within zones.
 * This results in conserving DMA zone[s] until all Normal memory is
 * exhausted, but results in overflowing to remote node while memory
 * may still exist in local DMA zone.
 */
static int node_order[MAX_NUMNODES];

static void build_zonelists_in_zone_order(pg_data_t *pgdat, int nr_nodes)
{
	int pos, j, node;
	int zone_type;		/* needs to be signed */
	struct zone *z;
	struct zonelist *zonelist;

2435 2436 2437 2438 2439 2440 2441
	zonelist = &pgdat->node_zonelists[0];
	pos = 0;
	for (zone_type = MAX_NR_ZONES - 1; zone_type >= 0; zone_type--) {
		for (j = 0; j < nr_nodes; j++) {
			node = node_order[j];
			z = &NODE_DATA(node)->node_zones[zone_type];
			if (populated_zone(z)) {
2442 2443
				zoneref_set_zone(z,
					&zonelist->_zonerefs[pos++]);
2444
				check_highest_zone(zone_type);
2445 2446 2447
			}
		}
	}
2448 2449
	zonelist->_zonerefs[pos].zone = NULL;
	zonelist->_zonerefs[pos].zone_idx = 0;
2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484
}

static int default_zonelist_order(void)
{
	int nid, zone_type;
	unsigned long low_kmem_size,total_size;
	struct zone *z;
	int average_size;
	/*
         * ZONE_DMA and ZONE_DMA32 can be very small area in the sytem.
	 * If they are really small and used heavily, the system can fall
	 * into OOM very easily.
	 * This function detect ZONE_DMA/DMA32 size and confgigures zone order.
	 */
	/* Is there ZONE_NORMAL ? (ex. ppc has only DMA zone..) */
	low_kmem_size = 0;
	total_size = 0;
	for_each_online_node(nid) {
		for (zone_type = 0; zone_type < MAX_NR_ZONES; zone_type++) {
			z = &NODE_DATA(nid)->node_zones[zone_type];
			if (populated_zone(z)) {
				if (zone_type < ZONE_NORMAL)
					low_kmem_size += z->present_pages;
				total_size += z->present_pages;
			}
		}
	}
	if (!low_kmem_size ||  /* there are no DMA area. */
	    low_kmem_size > total_size/2) /* DMA/DMA32 is big. */
		return ZONELIST_ORDER_NODE;
	/*
	 * look into each node's config.
  	 * If there is a node whose DMA/DMA32 memory is very big area on
 	 * local memory, NODE_ORDER may be suitable.
         */
2485 2486
	average_size = total_size /
				(nodes_weight(node_states[N_HIGH_MEMORY]) + 1);
2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517
	for_each_online_node(nid) {
		low_kmem_size = 0;
		total_size = 0;
		for (zone_type = 0; zone_type < MAX_NR_ZONES; zone_type++) {
			z = &NODE_DATA(nid)->node_zones[zone_type];
			if (populated_zone(z)) {
				if (zone_type < ZONE_NORMAL)
					low_kmem_size += z->present_pages;
				total_size += z->present_pages;
			}
		}
		if (low_kmem_size &&
		    total_size > average_size && /* ignore small node */
		    low_kmem_size > total_size * 70/100)
			return ZONELIST_ORDER_NODE;
	}
	return ZONELIST_ORDER_ZONE;
}

static void set_zonelist_order(void)
{
	if (user_zonelist_order == ZONELIST_ORDER_DEFAULT)
		current_zonelist_order = default_zonelist_order();
	else
		current_zonelist_order = user_zonelist_order;
}

static void build_zonelists(pg_data_t *pgdat)
{
	int j, node, load;
	enum zone_type i;
L
Linus Torvalds 已提交
2518
	nodemask_t used_mask;
2519 2520 2521
	int local_node, prev_node;
	struct zonelist *zonelist;
	int order = current_zonelist_order;
L
Linus Torvalds 已提交
2522 2523

	/* initialize zonelists */
2524
	for (i = 0; i < MAX_ZONELISTS; i++) {
L
Linus Torvalds 已提交
2525
		zonelist = pgdat->node_zonelists + i;
2526 2527
		zonelist->_zonerefs[0].zone = NULL;
		zonelist->_zonerefs[0].zone_idx = 0;
L
Linus Torvalds 已提交
2528 2529 2530 2531
	}

	/* NUMA-aware ordering of nodes */
	local_node = pgdat->node_id;
2532
	load = nr_online_nodes;
L
Linus Torvalds 已提交
2533 2534
	prev_node = local_node;
	nodes_clear(used_mask);
2535 2536 2537 2538 2539

	memset(node_load, 0, sizeof(node_load));
	memset(node_order, 0, sizeof(node_order));
	j = 0;

L
Linus Torvalds 已提交
2540
	while ((node = find_next_best_node(local_node, &used_mask)) >= 0) {
2541 2542 2543 2544 2545 2546 2547 2548 2549
		int distance = node_distance(local_node, node);

		/*
		 * If another node is sufficiently far away then it is better
		 * to reclaim pages in a zone before going off node.
		 */
		if (distance > RECLAIM_DISTANCE)
			zone_reclaim_mode = 1;

L
Linus Torvalds 已提交
2550 2551 2552 2553 2554
		/*
		 * We don't want to pressure a particular node.
		 * So adding penalty to the first node in same
		 * distance group to make it round-robin.
		 */
2555
		if (distance != node_distance(local_node, prev_node))
2556 2557
			node_load[node] = load;

L
Linus Torvalds 已提交
2558 2559
		prev_node = node;
		load--;
2560 2561 2562 2563 2564
		if (order == ZONELIST_ORDER_NODE)
			build_zonelists_in_node_order(pgdat, node);
		else
			node_order[j++] = node;	/* remember order */
	}
L
Linus Torvalds 已提交
2565

2566 2567 2568
	if (order == ZONELIST_ORDER_ZONE) {
		/* calculate node order -- i.e., DMA last! */
		build_zonelists_in_zone_order(pgdat, j);
L
Linus Torvalds 已提交
2569
	}
2570 2571

	build_thisnode_zonelists(pgdat);
L
Linus Torvalds 已提交
2572 2573
}

2574
/* Construct the zonelist performance cache - see further mmzone.h */
2575
static void build_zonelist_cache(pg_data_t *pgdat)
2576
{
2577 2578
	struct zonelist *zonelist;
	struct zonelist_cache *zlc;
2579
	struct zoneref *z;
2580

2581 2582 2583
	zonelist = &pgdat->node_zonelists[0];
	zonelist->zlcache_ptr = zlc = &zonelist->zlcache;
	bitmap_zero(zlc->fullzones, MAX_ZONES_PER_ZONELIST);
2584 2585
	for (z = zonelist->_zonerefs; z->zone; z++)
		zlc->z_to_n[z - zonelist->_zonerefs] = zonelist_node_idx(z);
2586 2587
}

2588

L
Linus Torvalds 已提交
2589 2590
#else	/* CONFIG_NUMA */

2591 2592 2593 2594 2595 2596
static void set_zonelist_order(void)
{
	current_zonelist_order = ZONELIST_ORDER_ZONE;
}

static void build_zonelists(pg_data_t *pgdat)
L
Linus Torvalds 已提交
2597
{
2598
	int node, local_node;
2599 2600
	enum zone_type j;
	struct zonelist *zonelist;
L
Linus Torvalds 已提交
2601 2602 2603

	local_node = pgdat->node_id;

2604 2605
	zonelist = &pgdat->node_zonelists[0];
	j = build_zonelists_node(pgdat, zonelist, 0, MAX_NR_ZONES - 1);
L
Linus Torvalds 已提交
2606

2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619
	/*
	 * Now we build the zonelist so that it contains the zones
	 * of all the other nodes.
	 * We don't want to pressure a particular node, so when
	 * building the zones for node N, we make sure that the
	 * zones coming right after the local ones are those from
	 * node N+1 (modulo N)
	 */
	for (node = local_node + 1; node < MAX_NUMNODES; node++) {
		if (!node_online(node))
			continue;
		j = build_zonelists_node(NODE_DATA(node), zonelist, j,
							MAX_NR_ZONES - 1);
L
Linus Torvalds 已提交
2620
	}
2621 2622 2623 2624 2625 2626 2627
	for (node = 0; node < local_node; node++) {
		if (!node_online(node))
			continue;
		j = build_zonelists_node(NODE_DATA(node), zonelist, j,
							MAX_NR_ZONES - 1);
	}

2628 2629
	zonelist->_zonerefs[j].zone = NULL;
	zonelist->_zonerefs[j].zone_idx = 0;
L
Linus Torvalds 已提交
2630 2631
}

2632
/* non-NUMA variant of zonelist performance cache - just NULL zlcache_ptr */
2633
static void build_zonelist_cache(pg_data_t *pgdat)
2634
{
2635
	pgdat->node_zonelists[0].zlcache_ptr = NULL;
2636 2637
}

L
Linus Torvalds 已提交
2638 2639
#endif	/* CONFIG_NUMA */

2640
/* return values int ....just for stop_machine() */
2641
static int __build_all_zonelists(void *dummy)
L
Linus Torvalds 已提交
2642
{
2643
	int nid;
2644 2645

	for_each_online_node(nid) {
2646 2647 2648 2649
		pg_data_t *pgdat = NODE_DATA(nid);

		build_zonelists(pgdat);
		build_zonelist_cache(pgdat);
2650
	}
2651 2652 2653
	return 0;
}

2654
void build_all_zonelists(void)
2655
{
2656 2657
	set_zonelist_order();

2658
	if (system_state == SYSTEM_BOOTING) {
2659
		__build_all_zonelists(NULL);
2660
		mminit_verify_zonelist();
2661 2662
		cpuset_init_current_mems_allowed();
	} else {
S
Simon Arlott 已提交
2663
		/* we have to stop all cpus to guarantee there is no user
2664
		   of zonelist */
2665
		stop_machine(__build_all_zonelists, NULL, NULL);
2666 2667
		/* cpuset refresh routine should be here */
	}
2668
	vm_total_pages = nr_free_pagecache_pages();
2669 2670 2671 2672 2673 2674 2675
	/*
	 * Disable grouping by mobility if the number of pages in the
	 * system is too low to allow the mechanism to work. It would be
	 * more accurate, but expensive to check per-zone. This check is
	 * made on memory-hotadd so a system can start with mobility
	 * disabled and enable it later
	 */
2676
	if (vm_total_pages < (pageblock_nr_pages * MIGRATE_TYPES))
2677 2678 2679 2680 2681 2682
		page_group_by_mobility_disabled = 1;
	else
		page_group_by_mobility_disabled = 0;

	printk("Built %i zonelists in %s order, mobility grouping %s.  "
		"Total pages: %ld\n",
2683
			nr_online_nodes,
2684
			zonelist_order_name[current_zonelist_order],
2685
			page_group_by_mobility_disabled ? "off" : "on",
2686 2687 2688 2689
			vm_total_pages);
#ifdef CONFIG_NUMA
	printk("Policy zone: %s\n", zone_names[policy_zone]);
#endif
L
Linus Torvalds 已提交
2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704
}

/*
 * Helper functions to size the waitqueue hash table.
 * Essentially these want to choose hash table sizes sufficiently
 * large so that collisions trying to wait on pages are rare.
 * But in fact, the number of active page waitqueues on typical
 * systems is ridiculously low, less than 200. So this is even
 * conservative, even though it seems large.
 *
 * The constant PAGES_PER_WAITQUEUE specifies the ratio of pages to
 * waitqueues, i.e. the size of the waitq table given the number of pages.
 */
#define PAGES_PER_WAITQUEUE	256

2705
#ifndef CONFIG_MEMORY_HOTPLUG
2706
static inline unsigned long wait_table_hash_nr_entries(unsigned long pages)
L
Linus Torvalds 已提交
2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723
{
	unsigned long size = 1;

	pages /= PAGES_PER_WAITQUEUE;

	while (size < pages)
		size <<= 1;

	/*
	 * Once we have dozens or even hundreds of threads sleeping
	 * on IO we've got bigger problems than wait queue collision.
	 * Limit the size of the wait table to a reasonable size.
	 */
	size = min(size, 4096UL);

	return max(size, 4UL);
}
2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746
#else
/*
 * A zone's size might be changed by hot-add, so it is not possible to determine
 * a suitable size for its wait_table.  So we use the maximum size now.
 *
 * The max wait table size = 4096 x sizeof(wait_queue_head_t).   ie:
 *
 *    i386 (preemption config)    : 4096 x 16 = 64Kbyte.
 *    ia64, x86-64 (no preemption): 4096 x 20 = 80Kbyte.
 *    ia64, x86-64 (preemption)   : 4096 x 24 = 96Kbyte.
 *
 * The maximum entries are prepared when a zone's memory is (512K + 256) pages
 * or more by the traditional way. (See above).  It equals:
 *
 *    i386, x86-64, powerpc(4K page size) : =  ( 2G + 1M)byte.
 *    ia64(16K page size)                 : =  ( 8G + 4M)byte.
 *    powerpc (64K page size)             : =  (32G +16M)byte.
 */
static inline unsigned long wait_table_hash_nr_entries(unsigned long pages)
{
	return 4096UL;
}
#endif
L
Linus Torvalds 已提交
2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759

/*
 * This is an integer logarithm so that shifts can be used later
 * to extract the more random high bits from the multiplicative
 * hash function before the remainder is taken.
 */
static inline unsigned long wait_table_bits(unsigned long size)
{
	return ffz(~size);
}

#define LONG_ALIGN(x) (((x)+(sizeof(long))-1)&~((sizeof(long))-1))

2760
/*
2761
 * Mark a number of pageblocks as MIGRATE_RESERVE. The number
2762 2763
 * of blocks reserved is based on min_wmark_pages(zone). The memory within
 * the reserve will tend to store contiguous free pages. Setting min_free_kbytes
2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775
 * higher will lead to a bigger reserve which will get freed as contiguous
 * blocks as reclaim kicks in
 */
static void setup_zone_migrate_reserve(struct zone *zone)
{
	unsigned long start_pfn, pfn, end_pfn;
	struct page *page;
	unsigned long reserve, block_migratetype;

	/* Get the start pfn, end pfn and the number of blocks to reserve */
	start_pfn = zone->zone_start_pfn;
	end_pfn = start_pfn + zone->spanned_pages;
2776
	reserve = roundup(min_wmark_pages(zone), pageblock_nr_pages) >>
2777
							pageblock_order;
2778

2779
	for (pfn = start_pfn; pfn < end_pfn; pfn += pageblock_nr_pages) {
2780 2781 2782 2783
		if (!pfn_valid(pfn))
			continue;
		page = pfn_to_page(pfn);

2784 2785 2786 2787
		/* Watch out for overlapping nodes */
		if (page_to_nid(page) != zone_to_nid(zone))
			continue;

2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817
		/* Blocks with reserved pages will never free, skip them. */
		if (PageReserved(page))
			continue;

		block_migratetype = get_pageblock_migratetype(page);

		/* If this block is reserved, account for it */
		if (reserve > 0 && block_migratetype == MIGRATE_RESERVE) {
			reserve--;
			continue;
		}

		/* Suitable for reserving if this block is movable */
		if (reserve > 0 && block_migratetype == MIGRATE_MOVABLE) {
			set_pageblock_migratetype(page, MIGRATE_RESERVE);
			move_freepages_block(zone, page, MIGRATE_RESERVE);
			reserve--;
			continue;
		}

		/*
		 * If the reserve is met and this is a previous reserved block,
		 * take it back
		 */
		if (block_migratetype == MIGRATE_RESERVE) {
			set_pageblock_migratetype(page, MIGRATE_MOVABLE);
			move_freepages_block(zone, page, MIGRATE_MOVABLE);
		}
	}
}
M
Mel Gorman 已提交
2818

L
Linus Torvalds 已提交
2819 2820 2821 2822 2823
/*
 * Initially all pages are reserved - free ones are freed
 * up by free_all_bootmem() once the early boot process is
 * done. Non-atomic initialization, single-pass.
 */
2824
void __meminit memmap_init_zone(unsigned long size, int nid, unsigned long zone,
D
Dave Hansen 已提交
2825
		unsigned long start_pfn, enum memmap_context context)
L
Linus Torvalds 已提交
2826 2827
{
	struct page *page;
A
Andy Whitcroft 已提交
2828 2829
	unsigned long end_pfn = start_pfn + size;
	unsigned long pfn;
2830
	struct zone *z;
L
Linus Torvalds 已提交
2831

2832 2833 2834
	if (highest_memmap_pfn < end_pfn - 1)
		highest_memmap_pfn = end_pfn - 1;

2835
	z = &NODE_DATA(nid)->node_zones[zone];
2836
	for (pfn = start_pfn; pfn < end_pfn; pfn++) {
D
Dave Hansen 已提交
2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847
		/*
		 * There can be holes in boot-time mem_map[]s
		 * handed to this function.  They do not
		 * exist on hotplugged memory.
		 */
		if (context == MEMMAP_EARLY) {
			if (!early_pfn_valid(pfn))
				continue;
			if (!early_pfn_in_nid(pfn, nid))
				continue;
		}
A
Andy Whitcroft 已提交
2848 2849
		page = pfn_to_page(pfn);
		set_page_links(page, zone, nid, pfn);
2850
		mminit_verify_page_links(page, zone, nid, pfn);
2851
		init_page_count(page);
L
Linus Torvalds 已提交
2852 2853
		reset_page_mapcount(page);
		SetPageReserved(page);
2854 2855 2856 2857 2858
		/*
		 * Mark the block movable so that blocks are reserved for
		 * movable at startup. This will force kernel allocations
		 * to reserve their blocks rather than leaking throughout
		 * the address space during boot when many long-lived
2859 2860 2861
		 * kernel allocations are made. Later some blocks near
		 * the start are marked MIGRATE_RESERVE by
		 * setup_zone_migrate_reserve()
2862 2863 2864 2865 2866
		 *
		 * bitmap is created for zone's valid pfn range. but memmap
		 * can be created for invalid pages (for alignment)
		 * check here not to call set_pageblock_migratetype() against
		 * pfn out of zone.
2867
		 */
2868 2869 2870
		if ((z->zone_start_pfn <= pfn)
		    && (pfn < z->zone_start_pfn + z->spanned_pages)
		    && !(pfn & (pageblock_nr_pages - 1)))
2871
			set_pageblock_migratetype(page, MIGRATE_MOVABLE);
2872

L
Linus Torvalds 已提交
2873 2874 2875 2876
		INIT_LIST_HEAD(&page->lru);
#ifdef WANT_PAGE_VIRTUAL
		/* The shift won't overflow because ZONE_NORMAL is below 4G. */
		if (!is_highmem_idx(zone))
2877
			set_page_address(page, __va(pfn << PAGE_SHIFT));
L
Linus Torvalds 已提交
2878 2879 2880 2881
#endif
	}
}

2882
static void __meminit zone_init_free_lists(struct zone *zone)
L
Linus Torvalds 已提交
2883
{
2884 2885 2886
	int order, t;
	for_each_migratetype_order(order, t) {
		INIT_LIST_HEAD(&zone->free_area[order].free_list[t]);
L
Linus Torvalds 已提交
2887 2888 2889 2890 2891 2892
		zone->free_area[order].nr_free = 0;
	}
}

#ifndef __HAVE_ARCH_MEMMAP_INIT
#define memmap_init(size, nid, zone, start_pfn) \
D
Dave Hansen 已提交
2893
	memmap_init_zone((size), (nid), (zone), (start_pfn), MEMMAP_EARLY)
L
Linus Torvalds 已提交
2894 2895
#endif

2896
static int zone_batchsize(struct zone *zone)
2897
{
2898
#ifdef CONFIG_MMU
2899 2900 2901 2902
	int batch;

	/*
	 * The per-cpu-pages pools are set to around 1000th of the
2903
	 * size of the zone.  But no more than 1/2 of a meg.
2904 2905 2906 2907
	 *
	 * OK, so we don't know how big the cache is.  So guess.
	 */
	batch = zone->present_pages / 1024;
2908 2909
	if (batch * PAGE_SIZE > 512 * 1024)
		batch = (512 * 1024) / PAGE_SIZE;
2910 2911 2912 2913 2914
	batch /= 4;		/* We effectively *= 4 below */
	if (batch < 1)
		batch = 1;

	/*
2915 2916 2917
	 * Clamp the batch to a 2^n - 1 value. Having a power
	 * of 2 value was found to be more likely to have
	 * suboptimal cache aliasing properties in some cases.
2918
	 *
2919 2920 2921 2922
	 * For example if 2 tasks are alternately allocating
	 * batches of pages, one task can end up with a lot
	 * of pages of one half of the possible page colors
	 * and the other with pages of the other colors.
2923
	 */
2924
	batch = rounddown_pow_of_two(batch + batch/2) - 1;
2925

2926
	return batch;
2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943

#else
	/* The deferral and batching of frees should be suppressed under NOMMU
	 * conditions.
	 *
	 * The problem is that NOMMU needs to be able to allocate large chunks
	 * of contiguous memory as there's no hardware page translation to
	 * assemble apparent contiguous memory from discontiguous pages.
	 *
	 * Queueing large contiguous runs of pages for batching, however,
	 * causes the pages to actually be freed in smaller chunks.  As there
	 * can be a significant delay between the individual batches being
	 * recycled, this leads to the once large chunks of space being
	 * fragmented and becoming unavailable for high-order allocations.
	 */
	return 0;
#endif
2944 2945
}

A
Adrian Bunk 已提交
2946
static void setup_pageset(struct per_cpu_pageset *p, unsigned long batch)
2947 2948 2949
{
	struct per_cpu_pages *pcp;

2950 2951
	memset(p, 0, sizeof(*p));

2952
	pcp = &p->pcp;
2953 2954 2955 2956 2957 2958
	pcp->count = 0;
	pcp->high = 6 * batch;
	pcp->batch = max(1UL, 1 * batch);
	INIT_LIST_HEAD(&pcp->list);
}

2959 2960 2961 2962 2963 2964 2965 2966 2967 2968
/*
 * setup_pagelist_highmark() sets the high water mark for hot per_cpu_pagelist
 * to the value high for the pageset p.
 */

static void setup_pagelist_highmark(struct per_cpu_pageset *p,
				unsigned long high)
{
	struct per_cpu_pages *pcp;

2969
	pcp = &p->pcp;
2970 2971 2972 2973 2974 2975 2976
	pcp->high = high;
	pcp->batch = max(1UL, high/4);
	if ((high/4) > (PAGE_SHIFT * 8))
		pcp->batch = PAGE_SHIFT * 8;
}


2977 2978
#ifdef CONFIG_NUMA
/*
2979 2980 2981 2982 2983 2984 2985
 * Boot pageset table. One per cpu which is going to be used for all
 * zones and all nodes. The parameters will be set in such a way
 * that an item put on a list will immediately be handed over to
 * the buddy list. This is safe since pageset manipulation is done
 * with interrupts disabled.
 *
 * Some NUMA counter updates may also be caught by the boot pagesets.
2986 2987 2988 2989 2990 2991 2992 2993
 *
 * The boot_pagesets must be kept even after bootup is complete for
 * unused processors and/or zones. They do play a role for bootstrapping
 * hotplugged processors.
 *
 * zoneinfo_show() and maybe other functions do
 * not check if the processor is online before following the pageset pointer.
 * Other parts of the kernel may not check if the zone is available.
2994
 */
2995
static struct per_cpu_pageset boot_pageset[NR_CPUS];
2996 2997 2998

/*
 * Dynamically allocate memory for the
2999 3000
 * per cpu pageset array in struct zone.
 */
3001
static int __cpuinit process_zones(int cpu)
3002 3003
{
	struct zone *zone, *dzone;
3004 3005 3006
	int node = cpu_to_node(cpu);

	node_set_state(node, N_CPU);	/* this node has a cpu */
3007

3008
	for_each_populated_zone(zone) {
N
Nick Piggin 已提交
3009
		zone_pcp(zone, cpu) = kmalloc_node(sizeof(struct per_cpu_pageset),
3010
					 GFP_KERNEL, node);
N
Nick Piggin 已提交
3011
		if (!zone_pcp(zone, cpu))
3012 3013
			goto bad;

N
Nick Piggin 已提交
3014
		setup_pageset(zone_pcp(zone, cpu), zone_batchsize(zone));
3015 3016 3017 3018

		if (percpu_pagelist_fraction)
			setup_pagelist_highmark(zone_pcp(zone, cpu),
			 	(zone->present_pages / percpu_pagelist_fraction));
3019 3020 3021 3022 3023
	}

	return 0;
bad:
	for_each_zone(dzone) {
3024 3025
		if (!populated_zone(dzone))
			continue;
3026 3027
		if (dzone == zone)
			break;
N
Nick Piggin 已提交
3028
		kfree(zone_pcp(dzone, cpu));
3029
		zone_pcp(dzone, cpu) = &boot_pageset[cpu];
3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040
	}
	return -ENOMEM;
}

static inline void free_zone_pagesets(int cpu)
{
	struct zone *zone;

	for_each_zone(zone) {
		struct per_cpu_pageset *pset = zone_pcp(zone, cpu);

3041 3042 3043
		/* Free per_cpu_pageset if it is slab allocated */
		if (pset != &boot_pageset[cpu])
			kfree(pset);
3044
		zone_pcp(zone, cpu) = &boot_pageset[cpu];
3045 3046 3047
	}
}

3048
static int __cpuinit pageset_cpuup_callback(struct notifier_block *nfb,
3049 3050 3051 3052 3053 3054 3055
		unsigned long action,
		void *hcpu)
{
	int cpu = (long)hcpu;
	int ret = NOTIFY_OK;

	switch (action) {
3056
	case CPU_UP_PREPARE:
3057
	case CPU_UP_PREPARE_FROZEN:
3058 3059 3060 3061
		if (process_zones(cpu))
			ret = NOTIFY_BAD;
		break;
	case CPU_UP_CANCELED:
3062
	case CPU_UP_CANCELED_FROZEN:
3063
	case CPU_DEAD:
3064
	case CPU_DEAD_FROZEN:
3065 3066 3067 3068
		free_zone_pagesets(cpu);
		break;
	default:
		break;
3069 3070 3071 3072
	}
	return ret;
}

3073
static struct notifier_block __cpuinitdata pageset_notifier =
3074 3075
	{ &pageset_cpuup_callback, NULL, 0 };

3076
void __init setup_per_cpu_pageset(void)
3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090
{
	int err;

	/* Initialize per_cpu_pageset for cpu 0.
	 * A cpuup callback will do this for every cpu
	 * as it comes online
	 */
	err = process_zones(smp_processor_id());
	BUG_ON(err);
	register_cpu_notifier(&pageset_notifier);
}

#endif

S
Sam Ravnborg 已提交
3091
static noinline __init_refok
3092
int zone_wait_table_init(struct zone *zone, unsigned long zone_size_pages)
3093 3094 3095
{
	int i;
	struct pglist_data *pgdat = zone->zone_pgdat;
3096
	size_t alloc_size;
3097 3098 3099 3100 3101

	/*
	 * The per-page waitqueue mechanism uses hashed waitqueues
	 * per zone.
	 */
3102 3103 3104 3105
	zone->wait_table_hash_nr_entries =
		 wait_table_hash_nr_entries(zone_size_pages);
	zone->wait_table_bits =
		wait_table_bits(zone->wait_table_hash_nr_entries);
3106 3107 3108
	alloc_size = zone->wait_table_hash_nr_entries
					* sizeof(wait_queue_head_t);

3109
	if (!slab_is_available()) {
3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122
		zone->wait_table = (wait_queue_head_t *)
			alloc_bootmem_node(pgdat, alloc_size);
	} else {
		/*
		 * This case means that a zone whose size was 0 gets new memory
		 * via memory hot-add.
		 * But it may be the case that a new node was hot-added.  In
		 * this case vmalloc() will not be able to use this new node's
		 * memory - this wait_table must be initialized to use this new
		 * node itself as well.
		 * To use this new node's memory, further consideration will be
		 * necessary.
		 */
3123
		zone->wait_table = vmalloc(alloc_size);
3124 3125 3126
	}
	if (!zone->wait_table)
		return -ENOMEM;
3127

3128
	for(i = 0; i < zone->wait_table_hash_nr_entries; ++i)
3129
		init_waitqueue_head(zone->wait_table + i);
3130 3131

	return 0;
3132 3133
}

3134
static __meminit void zone_pcp_init(struct zone *zone)
3135 3136 3137 3138 3139 3140 3141
{
	int cpu;
	unsigned long batch = zone_batchsize(zone);

	for (cpu = 0; cpu < NR_CPUS; cpu++) {
#ifdef CONFIG_NUMA
		/* Early boot. Slab allocator not functional yet */
N
Nick Piggin 已提交
3142
		zone_pcp(zone, cpu) = &boot_pageset[cpu];
3143 3144 3145 3146 3147
		setup_pageset(&boot_pageset[cpu],0);
#else
		setup_pageset(zone_pcp(zone,cpu), batch);
#endif
	}
A
Anton Blanchard 已提交
3148 3149 3150
	if (zone->present_pages)
		printk(KERN_DEBUG "  %s zone: %lu pages, LIFO batch:%lu\n",
			zone->name, zone->present_pages, batch);
3151 3152
}

3153 3154
__meminit int init_currently_empty_zone(struct zone *zone,
					unsigned long zone_start_pfn,
D
Dave Hansen 已提交
3155 3156
					unsigned long size,
					enum memmap_context context)
3157 3158
{
	struct pglist_data *pgdat = zone->zone_pgdat;
3159 3160 3161 3162
	int ret;
	ret = zone_wait_table_init(zone, size);
	if (ret)
		return ret;
3163 3164 3165 3166
	pgdat->nr_zones = zone_idx(zone) + 1;

	zone->zone_start_pfn = zone_start_pfn;

3167 3168 3169 3170 3171 3172
	mminit_dprintk(MMINIT_TRACE, "memmap_init",
			"Initialising map node %d zone %lu pfns %lu -> %lu\n",
			pgdat->node_id,
			(unsigned long)zone_idx(zone),
			zone_start_pfn, (zone_start_pfn + size));

3173
	zone_init_free_lists(zone);
3174 3175

	return 0;
3176 3177
}

3178 3179 3180 3181 3182
#ifdef CONFIG_ARCH_POPULATES_NODE_MAP
/*
 * Basic iterator support. Return the first range of PFNs for a node
 * Note: nid == MAX_NUMNODES returns first region regardless of node
 */
3183
static int __meminit first_active_region_index_in_nid(int nid)
3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195
{
	int i;

	for (i = 0; i < nr_nodemap_entries; i++)
		if (nid == MAX_NUMNODES || early_node_map[i].nid == nid)
			return i;

	return -1;
}

/*
 * Basic iterator support. Return the next active range of PFNs for a node
S
Simon Arlott 已提交
3196
 * Note: nid == MAX_NUMNODES returns next region regardless of node
3197
 */
3198
static int __meminit next_active_region_index_in_nid(int index, int nid)
3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213
{
	for (index = index + 1; index < nr_nodemap_entries; index++)
		if (nid == MAX_NUMNODES || early_node_map[index].nid == nid)
			return index;

	return -1;
}

#ifndef CONFIG_HAVE_ARCH_EARLY_PFN_TO_NID
/*
 * Required by SPARSEMEM. Given a PFN, return what node the PFN is on.
 * Architectures may implement their own version but if add_active_range()
 * was used and there are no special requirements, this is a convenient
 * alternative
 */
3214
int __meminit __early_pfn_to_nid(unsigned long pfn)
3215 3216 3217 3218 3219 3220 3221 3222 3223 3224
{
	int i;

	for (i = 0; i < nr_nodemap_entries; i++) {
		unsigned long start_pfn = early_node_map[i].start_pfn;
		unsigned long end_pfn = early_node_map[i].end_pfn;

		if (start_pfn <= pfn && pfn < end_pfn)
			return early_node_map[i].nid;
	}
3225 3226
	/* This is a memory hole */
	return -1;
3227 3228 3229
}
#endif /* CONFIG_HAVE_ARCH_EARLY_PFN_TO_NID */

3230 3231
int __meminit early_pfn_to_nid(unsigned long pfn)
{
3232 3233 3234 3235 3236 3237 3238
	int nid;

	nid = __early_pfn_to_nid(pfn);
	if (nid >= 0)
		return nid;
	/* just returns 0 */
	return 0;
3239 3240
}

3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251
#ifdef CONFIG_NODES_SPAN_OTHER_NODES
bool __meminit early_pfn_in_nid(unsigned long pfn, int node)
{
	int nid;

	nid = __early_pfn_to_nid(pfn);
	if (nid >= 0 && nid != node)
		return false;
	return true;
}
#endif
3252

3253 3254 3255 3256 3257 3258 3259
/* Basic iterator support to walk early_node_map[] */
#define for_each_active_range_index_in_nid(i, nid) \
	for (i = first_active_region_index_in_nid(nid); i != -1; \
				i = next_active_region_index_in_nid(i, nid))

/**
 * free_bootmem_with_active_regions - Call free_bootmem_node for each active range
3260 3261
 * @nid: The node to free memory on. If MAX_NUMNODES, all nodes are freed.
 * @max_low_pfn: The highest PFN that will be passed to free_bootmem_node
3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288
 *
 * If an architecture guarantees that all ranges registered with
 * add_active_ranges() contain no holes and may be freed, this
 * this function may be used instead of calling free_bootmem() manually.
 */
void __init free_bootmem_with_active_regions(int nid,
						unsigned long max_low_pfn)
{
	int i;

	for_each_active_range_index_in_nid(i, nid) {
		unsigned long size_pages = 0;
		unsigned long end_pfn = early_node_map[i].end_pfn;

		if (early_node_map[i].start_pfn >= max_low_pfn)
			continue;

		if (end_pfn > max_low_pfn)
			end_pfn = max_low_pfn;

		size_pages = end_pfn - early_node_map[i].start_pfn;
		free_bootmem_node(NODE_DATA(early_node_map[i].nid),
				PFN_PHYS(early_node_map[i].start_pfn),
				size_pages << PAGE_SHIFT);
	}
}

3289 3290 3291
void __init work_with_active_regions(int nid, work_fn_t work_fn, void *data)
{
	int i;
3292
	int ret;
3293

3294 3295 3296 3297 3298 3299
	for_each_active_range_index_in_nid(i, nid) {
		ret = work_fn(early_node_map[i].start_pfn,
			      early_node_map[i].end_pfn, data);
		if (ret)
			break;
	}
3300
}
3301 3302
/**
 * sparse_memory_present_with_active_regions - Call memory_present for each active range
3303
 * @nid: The node to call memory_present for. If MAX_NUMNODES, all nodes will be used.
3304 3305 3306
 *
 * If an architecture guarantees that all ranges registered with
 * add_active_ranges() contain no holes and may be freed, this
3307
 * function may be used instead of calling memory_present() manually.
3308 3309 3310 3311 3312 3313 3314 3315 3316 3317 3318 3319 3320
 */
void __init sparse_memory_present_with_active_regions(int nid)
{
	int i;

	for_each_active_range_index_in_nid(i, nid)
		memory_present(early_node_map[i].nid,
				early_node_map[i].start_pfn,
				early_node_map[i].end_pfn);
}

/**
 * get_pfn_range_for_nid - Return the start and end page frames for a node
3321 3322 3323
 * @nid: The nid to return the range for. If MAX_NUMNODES, the min and max PFN are returned.
 * @start_pfn: Passed by reference. On return, it will have the node start_pfn.
 * @end_pfn: Passed by reference. On return, it will have the node end_pfn.
3324 3325 3326 3327
 *
 * It returns the start and end page frame of a node based on information
 * provided by an arch calling add_active_range(). If called for a node
 * with no available memory, a warning is printed and the start and end
3328
 * PFNs will be 0.
3329
 */
3330
void __meminit get_pfn_range_for_nid(unsigned int nid,
3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341
			unsigned long *start_pfn, unsigned long *end_pfn)
{
	int i;
	*start_pfn = -1UL;
	*end_pfn = 0;

	for_each_active_range_index_in_nid(i, nid) {
		*start_pfn = min(*start_pfn, early_node_map[i].start_pfn);
		*end_pfn = max(*end_pfn, early_node_map[i].end_pfn);
	}

3342
	if (*start_pfn == -1UL)
3343 3344 3345
		*start_pfn = 0;
}

M
Mel Gorman 已提交
3346 3347 3348 3349 3350
/*
 * This finds a zone that can be used for ZONE_MOVABLE pages. The
 * assumption is made that zones within a node are ordered in monotonic
 * increasing memory addresses so that the "highest" populated zone is used
 */
A
Adrian Bunk 已提交
3351
static void __init find_usable_zone_for_movable(void)
M
Mel Gorman 已提交
3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376
{
	int zone_index;
	for (zone_index = MAX_NR_ZONES - 1; zone_index >= 0; zone_index--) {
		if (zone_index == ZONE_MOVABLE)
			continue;

		if (arch_zone_highest_possible_pfn[zone_index] >
				arch_zone_lowest_possible_pfn[zone_index])
			break;
	}

	VM_BUG_ON(zone_index == -1);
	movable_zone = zone_index;
}

/*
 * The zone ranges provided by the architecture do not include ZONE_MOVABLE
 * because it is sized independant of architecture. Unlike the other zones,
 * the starting point for ZONE_MOVABLE is not fixed. It may be different
 * in each node depending on the size of each node and how evenly kernelcore
 * is distributed. This helper function adjusts the zone ranges
 * provided by the architecture for a given node by using the end of the
 * highest usable zone for ZONE_MOVABLE. This preserves the assumption that
 * zones within a node are in order of monotonic increases memory addresses
 */
A
Adrian Bunk 已提交
3377
static void __meminit adjust_zone_range_for_zone_movable(int nid,
M
Mel Gorman 已提交
3378 3379 3380 3381 3382 3383 3384 3385 3386 3387 3388 3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402
					unsigned long zone_type,
					unsigned long node_start_pfn,
					unsigned long node_end_pfn,
					unsigned long *zone_start_pfn,
					unsigned long *zone_end_pfn)
{
	/* Only adjust if ZONE_MOVABLE is on this node */
	if (zone_movable_pfn[nid]) {
		/* Size ZONE_MOVABLE */
		if (zone_type == ZONE_MOVABLE) {
			*zone_start_pfn = zone_movable_pfn[nid];
			*zone_end_pfn = min(node_end_pfn,
				arch_zone_highest_possible_pfn[movable_zone]);

		/* Adjust for ZONE_MOVABLE starting within this range */
		} else if (*zone_start_pfn < zone_movable_pfn[nid] &&
				*zone_end_pfn > zone_movable_pfn[nid]) {
			*zone_end_pfn = zone_movable_pfn[nid];

		/* Check if this whole range is within ZONE_MOVABLE */
		} else if (*zone_start_pfn >= zone_movable_pfn[nid])
			*zone_start_pfn = *zone_end_pfn;
	}
}

3403 3404 3405 3406
/*
 * Return the number of pages a zone spans in a node, including holes
 * present_pages = zone_spanned_pages_in_node() - zone_absent_pages_in_node()
 */
P
Paul Mundt 已提交
3407
static unsigned long __meminit zone_spanned_pages_in_node(int nid,
3408 3409 3410 3411 3412 3413 3414 3415 3416 3417
					unsigned long zone_type,
					unsigned long *ignored)
{
	unsigned long node_start_pfn, node_end_pfn;
	unsigned long zone_start_pfn, zone_end_pfn;

	/* Get the start and end of the node and zone */
	get_pfn_range_for_nid(nid, &node_start_pfn, &node_end_pfn);
	zone_start_pfn = arch_zone_lowest_possible_pfn[zone_type];
	zone_end_pfn = arch_zone_highest_possible_pfn[zone_type];
M
Mel Gorman 已提交
3418 3419 3420
	adjust_zone_range_for_zone_movable(nid, zone_type,
				node_start_pfn, node_end_pfn,
				&zone_start_pfn, &zone_end_pfn);
3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435

	/* Check that this node has pages within the zone's required range */
	if (zone_end_pfn < node_start_pfn || zone_start_pfn > node_end_pfn)
		return 0;

	/* Move the zone boundaries inside the node if necessary */
	zone_end_pfn = min(zone_end_pfn, node_end_pfn);
	zone_start_pfn = max(zone_start_pfn, node_start_pfn);

	/* Return the spanned pages */
	return zone_end_pfn - zone_start_pfn;
}

/*
 * Return the number of holes in a range on a node. If nid is MAX_NUMNODES,
3436
 * then all holes in the requested range will be accounted for.
3437
 */
A
Adrian Bunk 已提交
3438
static unsigned long __meminit __absent_pages_in_range(int nid,
3439 3440 3441 3442 3443 3444 3445 3446 3447 3448 3449 3450
				unsigned long range_start_pfn,
				unsigned long range_end_pfn)
{
	int i = 0;
	unsigned long prev_end_pfn = 0, hole_pages = 0;
	unsigned long start_pfn;

	/* Find the end_pfn of the first active range of pfns in the node */
	i = first_active_region_index_in_nid(nid);
	if (i == -1)
		return 0;

3451 3452
	prev_end_pfn = min(early_node_map[i].start_pfn, range_end_pfn);

3453 3454
	/* Account for ranges before physical memory on this node */
	if (early_node_map[i].start_pfn > range_start_pfn)
3455
		hole_pages = prev_end_pfn - range_start_pfn;
3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468 3469 3470 3471 3472 3473 3474 3475

	/* Find all holes for the zone within the node */
	for (; i != -1; i = next_active_region_index_in_nid(i, nid)) {

		/* No need to continue if prev_end_pfn is outside the zone */
		if (prev_end_pfn >= range_end_pfn)
			break;

		/* Make sure the end of the zone is not within the hole */
		start_pfn = min(early_node_map[i].start_pfn, range_end_pfn);
		prev_end_pfn = max(prev_end_pfn, range_start_pfn);

		/* Update the hole size cound and move on */
		if (start_pfn > range_start_pfn) {
			BUG_ON(prev_end_pfn > start_pfn);
			hole_pages += start_pfn - prev_end_pfn;
		}
		prev_end_pfn = early_node_map[i].end_pfn;
	}

3476 3477
	/* Account for ranges past physical memory on this node */
	if (range_end_pfn > prev_end_pfn)
3478
		hole_pages += range_end_pfn -
3479 3480
				max(range_start_pfn, prev_end_pfn);

3481 3482 3483 3484 3485 3486 3487 3488
	return hole_pages;
}

/**
 * absent_pages_in_range - Return number of page frames in holes within a range
 * @start_pfn: The start PFN to start searching for holes
 * @end_pfn: The end PFN to stop searching for holes
 *
3489
 * It returns the number of pages frames in memory holes within a range.
3490 3491 3492 3493 3494 3495 3496 3497
 */
unsigned long __init absent_pages_in_range(unsigned long start_pfn,
							unsigned long end_pfn)
{
	return __absent_pages_in_range(MAX_NUMNODES, start_pfn, end_pfn);
}

/* Return the number of page frames in holes in a zone on a node */
P
Paul Mundt 已提交
3498
static unsigned long __meminit zone_absent_pages_in_node(int nid,
3499 3500 3501
					unsigned long zone_type,
					unsigned long *ignored)
{
3502 3503 3504 3505 3506 3507 3508 3509 3510
	unsigned long node_start_pfn, node_end_pfn;
	unsigned long zone_start_pfn, zone_end_pfn;

	get_pfn_range_for_nid(nid, &node_start_pfn, &node_end_pfn);
	zone_start_pfn = max(arch_zone_lowest_possible_pfn[zone_type],
							node_start_pfn);
	zone_end_pfn = min(arch_zone_highest_possible_pfn[zone_type],
							node_end_pfn);

M
Mel Gorman 已提交
3511 3512 3513
	adjust_zone_range_for_zone_movable(nid, zone_type,
			node_start_pfn, node_end_pfn,
			&zone_start_pfn, &zone_end_pfn);
3514
	return __absent_pages_in_range(nid, zone_start_pfn, zone_end_pfn);
3515
}
3516

3517
#else
P
Paul Mundt 已提交
3518
static inline unsigned long __meminit zone_spanned_pages_in_node(int nid,
3519 3520 3521 3522 3523 3524
					unsigned long zone_type,
					unsigned long *zones_size)
{
	return zones_size[zone_type];
}

P
Paul Mundt 已提交
3525
static inline unsigned long __meminit zone_absent_pages_in_node(int nid,
3526 3527 3528 3529 3530 3531 3532 3533
						unsigned long zone_type,
						unsigned long *zholes_size)
{
	if (!zholes_size)
		return 0;

	return zholes_size[zone_type];
}
3534

3535 3536
#endif

3537
static void __meminit calculate_node_totalpages(struct pglist_data *pgdat,
3538 3539 3540 3541 3542 3543 3544 3545 3546 3547 3548 3549 3550 3551 3552 3553 3554 3555 3556 3557
		unsigned long *zones_size, unsigned long *zholes_size)
{
	unsigned long realtotalpages, totalpages = 0;
	enum zone_type i;

	for (i = 0; i < MAX_NR_ZONES; i++)
		totalpages += zone_spanned_pages_in_node(pgdat->node_id, i,
								zones_size);
	pgdat->node_spanned_pages = totalpages;

	realtotalpages = totalpages;
	for (i = 0; i < MAX_NR_ZONES; i++)
		realtotalpages -=
			zone_absent_pages_in_node(pgdat->node_id, i,
								zholes_size);
	pgdat->node_present_pages = realtotalpages;
	printk(KERN_DEBUG "On node %d totalpages: %lu\n", pgdat->node_id,
							realtotalpages);
}

3558 3559 3560
#ifndef CONFIG_SPARSEMEM
/*
 * Calculate the size of the zone->blockflags rounded to an unsigned long
3561 3562
 * Start by making sure zonesize is a multiple of pageblock_order by rounding
 * up. Then use 1 NR_PAGEBLOCK_BITS worth of bits per pageblock, finally
3563 3564 3565 3566 3567 3568 3569
 * round what is now in bits to nearest long in bits, then return it in
 * bytes.
 */
static unsigned long __init usemap_size(unsigned long zonesize)
{
	unsigned long usemapsize;

3570 3571
	usemapsize = roundup(zonesize, pageblock_nr_pages);
	usemapsize = usemapsize >> pageblock_order;
3572 3573 3574 3575 3576 3577 3578 3579 3580 3581 3582
	usemapsize *= NR_PAGEBLOCK_BITS;
	usemapsize = roundup(usemapsize, 8 * sizeof(unsigned long));

	return usemapsize / 8;
}

static void __init setup_usemap(struct pglist_data *pgdat,
				struct zone *zone, unsigned long zonesize)
{
	unsigned long usemapsize = usemap_size(zonesize);
	zone->pageblock_flags = NULL;
3583
	if (usemapsize)
3584 3585 3586 3587 3588 3589 3590
		zone->pageblock_flags = alloc_bootmem_node(pgdat, usemapsize);
}
#else
static void inline setup_usemap(struct pglist_data *pgdat,
				struct zone *zone, unsigned long zonesize) {}
#endif /* CONFIG_SPARSEMEM */

3591
#ifdef CONFIG_HUGETLB_PAGE_SIZE_VARIABLE
3592 3593 3594 3595 3596 3597 3598 3599 3600 3601

/* Return a sensible default order for the pageblock size. */
static inline int pageblock_default_order(void)
{
	if (HPAGE_SHIFT > PAGE_SHIFT)
		return HUGETLB_PAGE_ORDER;

	return MAX_ORDER-1;
}

3602 3603 3604 3605 3606 3607 3608 3609 3610 3611 3612 3613 3614 3615 3616
/* Initialise the number of pages represented by NR_PAGEBLOCK_BITS */
static inline void __init set_pageblock_order(unsigned int order)
{
	/* Check that pageblock_nr_pages has not already been setup */
	if (pageblock_order)
		return;

	/*
	 * Assume the largest contiguous order of interest is a huge page.
	 * This value may be variable depending on boot parameters on IA64
	 */
	pageblock_order = order;
}
#else /* CONFIG_HUGETLB_PAGE_SIZE_VARIABLE */

3617 3618 3619 3620 3621 3622 3623 3624 3625 3626
/*
 * When CONFIG_HUGETLB_PAGE_SIZE_VARIABLE is not set, set_pageblock_order()
 * and pageblock_default_order() are unused as pageblock_order is set
 * at compile-time. See include/linux/pageblock-flags.h for the values of
 * pageblock_order based on the kernel config
 */
static inline int pageblock_default_order(unsigned int order)
{
	return MAX_ORDER-1;
}
3627 3628 3629 3630
#define set_pageblock_order(x)	do {} while (0)

#endif /* CONFIG_HUGETLB_PAGE_SIZE_VARIABLE */

L
Linus Torvalds 已提交
3631 3632 3633 3634 3635 3636
/*
 * Set up the zone data structures:
 *   - mark all pages reserved
 *   - mark all memory queues empty
 *   - clear the memory bitmaps
 */
3637
static void __paginginit free_area_init_core(struct pglist_data *pgdat,
L
Linus Torvalds 已提交
3638 3639
		unsigned long *zones_size, unsigned long *zholes_size)
{
3640
	enum zone_type j;
3641
	int nid = pgdat->node_id;
L
Linus Torvalds 已提交
3642
	unsigned long zone_start_pfn = pgdat->node_start_pfn;
3643
	int ret;
L
Linus Torvalds 已提交
3644

3645
	pgdat_resize_init(pgdat);
L
Linus Torvalds 已提交
3646 3647 3648
	pgdat->nr_zones = 0;
	init_waitqueue_head(&pgdat->kswapd_wait);
	pgdat->kswapd_max_order = 0;
3649
	pgdat_page_cgroup_init(pgdat);
L
Linus Torvalds 已提交
3650 3651 3652
	
	for (j = 0; j < MAX_NR_ZONES; j++) {
		struct zone *zone = pgdat->node_zones + j;
3653
		unsigned long size, realsize, memmap_pages;
3654
		enum lru_list l;
L
Linus Torvalds 已提交
3655

3656 3657 3658
		size = zone_spanned_pages_in_node(nid, j, zones_size);
		realsize = size - zone_absent_pages_in_node(nid, j,
								zholes_size);
L
Linus Torvalds 已提交
3659

3660 3661 3662 3663 3664
		/*
		 * Adjust realsize so that it accounts for how much memory
		 * is used by this zone for memmap. This affects the watermark
		 * and per-cpu initialisations
		 */
3665 3666
		memmap_pages =
			PAGE_ALIGN(size * sizeof(struct page)) >> PAGE_SHIFT;
3667 3668
		if (realsize >= memmap_pages) {
			realsize -= memmap_pages;
3669 3670 3671 3672
			if (memmap_pages)
				printk(KERN_DEBUG
				       "  %s zone: %lu pages used for memmap\n",
				       zone_names[j], memmap_pages);
3673 3674 3675 3676 3677
		} else
			printk(KERN_WARNING
				"  %s zone: %lu pages exceeds realsize %lu\n",
				zone_names[j], memmap_pages, realsize);

3678 3679
		/* Account for reserved pages */
		if (j == 0 && realsize > dma_reserve) {
3680
			realsize -= dma_reserve;
Y
Yinghai Lu 已提交
3681
			printk(KERN_DEBUG "  %s zone: %lu pages reserved\n",
3682
					zone_names[0], dma_reserve);
3683 3684
		}

3685
		if (!is_highmem_idx(j))
L
Linus Torvalds 已提交
3686 3687 3688 3689 3690
			nr_kernel_pages += realsize;
		nr_all_pages += realsize;

		zone->spanned_pages = size;
		zone->present_pages = realsize;
3691
#ifdef CONFIG_NUMA
3692
		zone->node = nid;
3693
		zone->min_unmapped_pages = (realsize*sysctl_min_unmapped_ratio)
3694
						/ 100;
3695
		zone->min_slab_pages = (realsize * sysctl_min_slab_ratio) / 100;
3696
#endif
L
Linus Torvalds 已提交
3697 3698 3699
		zone->name = zone_names[j];
		spin_lock_init(&zone->lock);
		spin_lock_init(&zone->lru_lock);
3700
		zone_seqlock_init(zone);
L
Linus Torvalds 已提交
3701 3702
		zone->zone_pgdat = pgdat;

3703
		zone->prev_priority = DEF_PRIORITY;
L
Linus Torvalds 已提交
3704

3705
		zone_pcp_init(zone);
3706 3707
		for_each_lru(l) {
			INIT_LIST_HEAD(&zone->lru[l].list);
3708
			zone->lru[l].nr_saved_scan = 0;
3709
		}
3710 3711 3712 3713
		zone->reclaim_stat.recent_rotated[0] = 0;
		zone->reclaim_stat.recent_rotated[1] = 0;
		zone->reclaim_stat.recent_scanned[0] = 0;
		zone->reclaim_stat.recent_scanned[1] = 0;
3714
		zap_zone_vm_stats(zone);
3715
		zone->flags = 0;
L
Linus Torvalds 已提交
3716 3717 3718
		if (!size)
			continue;

3719
		set_pageblock_order(pageblock_default_order());
3720
		setup_usemap(pgdat, zone, size);
D
Dave Hansen 已提交
3721 3722
		ret = init_currently_empty_zone(zone, zone_start_pfn,
						size, MEMMAP_EARLY);
3723
		BUG_ON(ret);
3724
		memmap_init(size, nid, j, zone_start_pfn);
L
Linus Torvalds 已提交
3725 3726 3727 3728
		zone_start_pfn += size;
	}
}

S
Sam Ravnborg 已提交
3729
static void __init_refok alloc_node_mem_map(struct pglist_data *pgdat)
L
Linus Torvalds 已提交
3730 3731 3732 3733 3734
{
	/* Skip empty nodes */
	if (!pgdat->node_spanned_pages)
		return;

A
Andy Whitcroft 已提交
3735
#ifdef CONFIG_FLAT_NODE_MEM_MAP
L
Linus Torvalds 已提交
3736 3737
	/* ia64 gets its own node_mem_map, before this, without bootmem */
	if (!pgdat->node_mem_map) {
3738
		unsigned long size, start, end;
A
Andy Whitcroft 已提交
3739 3740
		struct page *map;

3741 3742 3743 3744 3745 3746 3747 3748 3749
		/*
		 * The zone's endpoints aren't required to be MAX_ORDER
		 * aligned but the node_mem_map endpoints must be in order
		 * for the buddy allocator to function correctly.
		 */
		start = pgdat->node_start_pfn & ~(MAX_ORDER_NR_PAGES - 1);
		end = pgdat->node_start_pfn + pgdat->node_spanned_pages;
		end = ALIGN(end, MAX_ORDER_NR_PAGES);
		size =  (end - start) * sizeof(struct page);
3750 3751 3752
		map = alloc_remap(pgdat->node_id, size);
		if (!map)
			map = alloc_bootmem_node(pgdat, size);
3753
		pgdat->node_mem_map = map + (pgdat->node_start_pfn - start);
L
Linus Torvalds 已提交
3754
	}
3755
#ifndef CONFIG_NEED_MULTIPLE_NODES
L
Linus Torvalds 已提交
3756 3757 3758
	/*
	 * With no DISCONTIG, the global mem_map is just set as node 0's
	 */
3759
	if (pgdat == NODE_DATA(0)) {
L
Linus Torvalds 已提交
3760
		mem_map = NODE_DATA(0)->node_mem_map;
3761 3762
#ifdef CONFIG_ARCH_POPULATES_NODE_MAP
		if (page_to_pfn(mem_map) != pgdat->node_start_pfn)
3763
			mem_map -= (pgdat->node_start_pfn - ARCH_PFN_OFFSET);
3764 3765
#endif /* CONFIG_ARCH_POPULATES_NODE_MAP */
	}
L
Linus Torvalds 已提交
3766
#endif
A
Andy Whitcroft 已提交
3767
#endif /* CONFIG_FLAT_NODE_MEM_MAP */
L
Linus Torvalds 已提交
3768 3769
}

3770 3771
void __paginginit free_area_init_node(int nid, unsigned long *zones_size,
		unsigned long node_start_pfn, unsigned long *zholes_size)
L
Linus Torvalds 已提交
3772
{
3773 3774
	pg_data_t *pgdat = NODE_DATA(nid);

L
Linus Torvalds 已提交
3775 3776
	pgdat->node_id = nid;
	pgdat->node_start_pfn = node_start_pfn;
3777
	calculate_node_totalpages(pgdat, zones_size, zholes_size);
L
Linus Torvalds 已提交
3778 3779

	alloc_node_mem_map(pgdat);
3780 3781 3782 3783 3784
#ifdef CONFIG_FLAT_NODE_MEM_MAP
	printk(KERN_DEBUG "free_area_init_node: node %d, pgdat %08lx, node_mem_map %08lx\n",
		nid, (unsigned long)pgdat,
		(unsigned long)pgdat->node_mem_map);
#endif
L
Linus Torvalds 已提交
3785 3786 3787 3788

	free_area_init_core(pgdat, zones_size, zholes_size);
}

3789
#ifdef CONFIG_ARCH_POPULATES_NODE_MAP
M
Miklos Szeredi 已提交
3790 3791 3792 3793 3794 3795 3796 3797 3798 3799 3800 3801 3802 3803 3804 3805 3806 3807 3808 3809

#if MAX_NUMNODES > 1
/*
 * Figure out the number of possible node ids.
 */
static void __init setup_nr_node_ids(void)
{
	unsigned int node;
	unsigned int highest = 0;

	for_each_node_mask(node, node_possible_map)
		highest = node;
	nr_node_ids = highest + 1;
}
#else
static inline void setup_nr_node_ids(void)
{
}
#endif

3810 3811 3812 3813 3814 3815 3816 3817 3818 3819 3820 3821 3822 3823 3824 3825 3826
/**
 * add_active_range - Register a range of PFNs backed by physical memory
 * @nid: The node ID the range resides on
 * @start_pfn: The start PFN of the available physical memory
 * @end_pfn: The end PFN of the available physical memory
 *
 * These ranges are stored in an early_node_map[] and later used by
 * free_area_init_nodes() to calculate zone sizes and holes. If the
 * range spans a memory hole, it is up to the architecture to ensure
 * the memory is not freed by the bootmem allocator. If possible
 * the range being registered will be merged with existing ranges.
 */
void __init add_active_range(unsigned int nid, unsigned long start_pfn,
						unsigned long end_pfn)
{
	int i;

3827 3828 3829 3830 3831
	mminit_dprintk(MMINIT_TRACE, "memory_register",
			"Entering add_active_range(%d, %#lx, %#lx) "
			"%d entries of %d used\n",
			nid, start_pfn, end_pfn,
			nr_nodemap_entries, MAX_ACTIVE_REGIONS);
3832

3833 3834
	mminit_validate_memmodel_limits(&start_pfn, &end_pfn);

3835 3836 3837 3838 3839 3840 3841 3842 3843 3844 3845 3846 3847 3848 3849 3850 3851 3852 3853 3854 3855 3856 3857 3858 3859 3860 3861 3862 3863 3864 3865 3866 3867 3868 3869 3870 3871 3872 3873
	/* Merge with existing active regions if possible */
	for (i = 0; i < nr_nodemap_entries; i++) {
		if (early_node_map[i].nid != nid)
			continue;

		/* Skip if an existing region covers this new one */
		if (start_pfn >= early_node_map[i].start_pfn &&
				end_pfn <= early_node_map[i].end_pfn)
			return;

		/* Merge forward if suitable */
		if (start_pfn <= early_node_map[i].end_pfn &&
				end_pfn > early_node_map[i].end_pfn) {
			early_node_map[i].end_pfn = end_pfn;
			return;
		}

		/* Merge backward if suitable */
		if (start_pfn < early_node_map[i].end_pfn &&
				end_pfn >= early_node_map[i].start_pfn) {
			early_node_map[i].start_pfn = start_pfn;
			return;
		}
	}

	/* Check that early_node_map is large enough */
	if (i >= MAX_ACTIVE_REGIONS) {
		printk(KERN_CRIT "More than %d memory regions, truncating\n",
							MAX_ACTIVE_REGIONS);
		return;
	}

	early_node_map[i].nid = nid;
	early_node_map[i].start_pfn = start_pfn;
	early_node_map[i].end_pfn = end_pfn;
	nr_nodemap_entries = i + 1;
}

/**
3874
 * remove_active_range - Shrink an existing registered range of PFNs
3875
 * @nid: The node id the range is on that should be shrunk
3876 3877
 * @start_pfn: The new PFN of the range
 * @end_pfn: The new PFN of the range
3878 3879
 *
 * i386 with NUMA use alloc_remap() to store a node_mem_map on a local node.
3880 3881 3882
 * The map is kept near the end physical page range that has already been
 * registered. This function allows an arch to shrink an existing registered
 * range.
3883
 */
3884 3885
void __init remove_active_range(unsigned int nid, unsigned long start_pfn,
				unsigned long end_pfn)
3886
{
3887 3888
	int i, j;
	int removed = 0;
3889

3890 3891 3892
	printk(KERN_DEBUG "remove_active_range (%d, %lu, %lu)\n",
			  nid, start_pfn, end_pfn);

3893
	/* Find the old active region end and shrink */
3894
	for_each_active_range_index_in_nid(i, nid) {
3895 3896
		if (early_node_map[i].start_pfn >= start_pfn &&
		    early_node_map[i].end_pfn <= end_pfn) {
3897
			/* clear it */
3898
			early_node_map[i].start_pfn = 0;
3899 3900 3901 3902
			early_node_map[i].end_pfn = 0;
			removed = 1;
			continue;
		}
3903 3904 3905 3906 3907 3908 3909 3910 3911 3912 3913 3914
		if (early_node_map[i].start_pfn < start_pfn &&
		    early_node_map[i].end_pfn > start_pfn) {
			unsigned long temp_end_pfn = early_node_map[i].end_pfn;
			early_node_map[i].end_pfn = start_pfn;
			if (temp_end_pfn > end_pfn)
				add_active_range(nid, end_pfn, temp_end_pfn);
			continue;
		}
		if (early_node_map[i].start_pfn >= start_pfn &&
		    early_node_map[i].end_pfn > end_pfn &&
		    early_node_map[i].start_pfn < end_pfn) {
			early_node_map[i].start_pfn = end_pfn;
3915
			continue;
3916
		}
3917 3918 3919 3920 3921 3922 3923 3924 3925 3926 3927 3928 3929 3930 3931 3932 3933 3934 3935
	}

	if (!removed)
		return;

	/* remove the blank ones */
	for (i = nr_nodemap_entries - 1; i > 0; i--) {
		if (early_node_map[i].nid != nid)
			continue;
		if (early_node_map[i].end_pfn)
			continue;
		/* we found it, get rid of it */
		for (j = i; j < nr_nodemap_entries - 1; j++)
			memcpy(&early_node_map[j], &early_node_map[j+1],
				sizeof(early_node_map[j]));
		j = nr_nodemap_entries - 1;
		memset(&early_node_map[j], 0, sizeof(early_node_map[j]));
		nr_nodemap_entries--;
	}
3936 3937 3938 3939
}

/**
 * remove_all_active_ranges - Remove all currently registered regions
3940
 *
3941 3942 3943 3944
 * During discovery, it may be found that a table like SRAT is invalid
 * and an alternative discovery method must be used. This function removes
 * all currently registered regions.
 */
3945
void __init remove_all_active_ranges(void)
3946 3947 3948 3949 3950 3951 3952 3953 3954 3955 3956 3957 3958 3959 3960 3961 3962 3963 3964 3965 3966 3967 3968 3969 3970 3971 3972 3973
{
	memset(early_node_map, 0, sizeof(early_node_map));
	nr_nodemap_entries = 0;
}

/* Compare two active node_active_regions */
static int __init cmp_node_active_region(const void *a, const void *b)
{
	struct node_active_region *arange = (struct node_active_region *)a;
	struct node_active_region *brange = (struct node_active_region *)b;

	/* Done this way to avoid overflows */
	if (arange->start_pfn > brange->start_pfn)
		return 1;
	if (arange->start_pfn < brange->start_pfn)
		return -1;

	return 0;
}

/* sort the node_map by start_pfn */
static void __init sort_node_map(void)
{
	sort(early_node_map, (size_t)nr_nodemap_entries,
			sizeof(struct node_active_region),
			cmp_node_active_region, NULL);
}

3974
/* Find the lowest pfn for a node */
A
Adrian Bunk 已提交
3975
static unsigned long __init find_min_pfn_for_node(int nid)
3976 3977
{
	int i;
3978
	unsigned long min_pfn = ULONG_MAX;
3979

3980 3981
	/* Assuming a sorted map, the first range found has the starting pfn */
	for_each_active_range_index_in_nid(i, nid)
3982
		min_pfn = min(min_pfn, early_node_map[i].start_pfn);
3983

3984 3985
	if (min_pfn == ULONG_MAX) {
		printk(KERN_WARNING
3986
			"Could not find start_pfn for node %d\n", nid);
3987 3988 3989 3990
		return 0;
	}

	return min_pfn;
3991 3992 3993 3994 3995 3996
}

/**
 * find_min_pfn_with_active_regions - Find the minimum PFN registered
 *
 * It returns the minimum PFN based on information provided via
3997
 * add_active_range().
3998 3999 4000 4001 4002 4003
 */
unsigned long __init find_min_pfn_with_active_regions(void)
{
	return find_min_pfn_for_node(MAX_NUMNODES);
}

4004 4005 4006 4007 4008
/*
 * early_calculate_totalpages()
 * Sum pages in active regions for movable zone.
 * Populate N_HIGH_MEMORY for calculating usable_nodes.
 */
A
Adrian Bunk 已提交
4009
static unsigned long __init early_calculate_totalpages(void)
4010 4011 4012 4013
{
	int i;
	unsigned long totalpages = 0;

4014 4015
	for (i = 0; i < nr_nodemap_entries; i++) {
		unsigned long pages = early_node_map[i].end_pfn -
4016
						early_node_map[i].start_pfn;
4017 4018 4019 4020 4021
		totalpages += pages;
		if (pages)
			node_set_state(early_node_map[i].nid, N_HIGH_MEMORY);
	}
  	return totalpages;
4022 4023
}

M
Mel Gorman 已提交
4024 4025 4026 4027 4028 4029
/*
 * Find the PFN the Movable zone begins in each node. Kernel memory
 * is spread evenly between nodes as long as the nodes have enough
 * memory. When they don't, some nodes will have more kernelcore than
 * others
 */
A
Adrian Bunk 已提交
4030
static void __init find_zone_movable_pfns_for_nodes(unsigned long *movable_pfn)
M
Mel Gorman 已提交
4031 4032 4033 4034
{
	int i, nid;
	unsigned long usable_startpfn;
	unsigned long kernelcore_node, kernelcore_remaining;
4035 4036
	/* save the state before borrow the nodemask */
	nodemask_t saved_node_state = node_states[N_HIGH_MEMORY];
4037 4038
	unsigned long totalpages = early_calculate_totalpages();
	int usable_nodes = nodes_weight(node_states[N_HIGH_MEMORY]);
M
Mel Gorman 已提交
4039

4040 4041 4042 4043 4044 4045 4046 4047 4048 4049 4050 4051 4052 4053 4054 4055 4056 4057 4058 4059 4060 4061
	/*
	 * If movablecore was specified, calculate what size of
	 * kernelcore that corresponds so that memory usable for
	 * any allocation type is evenly spread. If both kernelcore
	 * and movablecore are specified, then the value of kernelcore
	 * will be used for required_kernelcore if it's greater than
	 * what movablecore would have allowed.
	 */
	if (required_movablecore) {
		unsigned long corepages;

		/*
		 * Round-up so that ZONE_MOVABLE is at least as large as what
		 * was requested by the user
		 */
		required_movablecore =
			roundup(required_movablecore, MAX_ORDER_NR_PAGES);
		corepages = totalpages - required_movablecore;

		required_kernelcore = max(required_kernelcore, corepages);
	}

M
Mel Gorman 已提交
4062 4063
	/* If kernelcore was not specified, there is no ZONE_MOVABLE */
	if (!required_kernelcore)
4064
		goto out;
M
Mel Gorman 已提交
4065 4066 4067 4068 4069 4070 4071 4072

	/* usable_startpfn is the lowest possible pfn ZONE_MOVABLE can be at */
	find_usable_zone_for_movable();
	usable_startpfn = arch_zone_lowest_possible_pfn[movable_zone];

restart:
	/* Spread kernelcore memory as evenly as possible throughout nodes */
	kernelcore_node = required_kernelcore / usable_nodes;
4073
	for_each_node_state(nid, N_HIGH_MEMORY) {
M
Mel Gorman 已提交
4074 4075 4076 4077 4078 4079 4080 4081 4082 4083 4084 4085 4086 4087 4088 4089 4090 4091 4092 4093 4094 4095 4096 4097 4098 4099 4100 4101 4102 4103 4104 4105 4106 4107 4108 4109 4110 4111 4112 4113 4114 4115 4116 4117 4118 4119 4120 4121 4122 4123 4124 4125 4126 4127 4128 4129 4130 4131 4132 4133 4134 4135 4136 4137 4138 4139 4140 4141 4142 4143 4144 4145 4146 4147 4148 4149 4150 4151 4152 4153 4154 4155 4156 4157 4158 4159 4160 4161 4162
		/*
		 * Recalculate kernelcore_node if the division per node
		 * now exceeds what is necessary to satisfy the requested
		 * amount of memory for the kernel
		 */
		if (required_kernelcore < kernelcore_node)
			kernelcore_node = required_kernelcore / usable_nodes;

		/*
		 * As the map is walked, we track how much memory is usable
		 * by the kernel using kernelcore_remaining. When it is
		 * 0, the rest of the node is usable by ZONE_MOVABLE
		 */
		kernelcore_remaining = kernelcore_node;

		/* Go through each range of PFNs within this node */
		for_each_active_range_index_in_nid(i, nid) {
			unsigned long start_pfn, end_pfn;
			unsigned long size_pages;

			start_pfn = max(early_node_map[i].start_pfn,
						zone_movable_pfn[nid]);
			end_pfn = early_node_map[i].end_pfn;
			if (start_pfn >= end_pfn)
				continue;

			/* Account for what is only usable for kernelcore */
			if (start_pfn < usable_startpfn) {
				unsigned long kernel_pages;
				kernel_pages = min(end_pfn, usable_startpfn)
								- start_pfn;

				kernelcore_remaining -= min(kernel_pages,
							kernelcore_remaining);
				required_kernelcore -= min(kernel_pages,
							required_kernelcore);

				/* Continue if range is now fully accounted */
				if (end_pfn <= usable_startpfn) {

					/*
					 * Push zone_movable_pfn to the end so
					 * that if we have to rebalance
					 * kernelcore across nodes, we will
					 * not double account here
					 */
					zone_movable_pfn[nid] = end_pfn;
					continue;
				}
				start_pfn = usable_startpfn;
			}

			/*
			 * The usable PFN range for ZONE_MOVABLE is from
			 * start_pfn->end_pfn. Calculate size_pages as the
			 * number of pages used as kernelcore
			 */
			size_pages = end_pfn - start_pfn;
			if (size_pages > kernelcore_remaining)
				size_pages = kernelcore_remaining;
			zone_movable_pfn[nid] = start_pfn + size_pages;

			/*
			 * Some kernelcore has been met, update counts and
			 * break if the kernelcore for this node has been
			 * satisified
			 */
			required_kernelcore -= min(required_kernelcore,
								size_pages);
			kernelcore_remaining -= size_pages;
			if (!kernelcore_remaining)
				break;
		}
	}

	/*
	 * If there is still required_kernelcore, we do another pass with one
	 * less node in the count. This will push zone_movable_pfn[nid] further
	 * along on the nodes that still have memory until kernelcore is
	 * satisified
	 */
	usable_nodes--;
	if (usable_nodes && required_kernelcore > usable_nodes)
		goto restart;

	/* Align start of ZONE_MOVABLE on all nids to MAX_ORDER_NR_PAGES */
	for (nid = 0; nid < MAX_NUMNODES; nid++)
		zone_movable_pfn[nid] =
			roundup(zone_movable_pfn[nid], MAX_ORDER_NR_PAGES);
4163 4164 4165 4166

out:
	/* restore the node_state */
	node_states[N_HIGH_MEMORY] = saved_node_state;
M
Mel Gorman 已提交
4167 4168
}

4169 4170 4171 4172 4173 4174 4175 4176 4177 4178 4179 4180 4181 4182
/* Any regular memory on that node ? */
static void check_for_regular_memory(pg_data_t *pgdat)
{
#ifdef CONFIG_HIGHMEM
	enum zone_type zone_type;

	for (zone_type = 0; zone_type <= ZONE_NORMAL; zone_type++) {
		struct zone *zone = &pgdat->node_zones[zone_type];
		if (zone->present_pages)
			node_set_state(zone_to_nid(zone), N_NORMAL_MEMORY);
	}
#endif
}

4183 4184
/**
 * free_area_init_nodes - Initialise all pg_data_t and zone data
4185
 * @max_zone_pfn: an array of max PFNs for each zone
4186 4187 4188 4189 4190 4191 4192 4193 4194 4195 4196 4197 4198
 *
 * This will call free_area_init_node() for each active node in the system.
 * Using the page ranges provided by add_active_range(), the size of each
 * zone in each node and their holes is calculated. If the maximum PFN
 * between two adjacent zones match, it is assumed that the zone is empty.
 * For example, if arch_max_dma_pfn == arch_max_dma32_pfn, it is assumed
 * that arch_max_dma32_pfn has no pages. It is also assumed that a zone
 * starts where the previous one ended. For example, ZONE_DMA32 starts
 * at arch_max_dma_pfn.
 */
void __init free_area_init_nodes(unsigned long *max_zone_pfn)
{
	unsigned long nid;
4199
	int i;
4200

4201 4202 4203
	/* Sort early_node_map as initialisation assumes it is sorted */
	sort_node_map();

4204 4205 4206 4207 4208 4209 4210 4211
	/* Record where the zone boundaries are */
	memset(arch_zone_lowest_possible_pfn, 0,
				sizeof(arch_zone_lowest_possible_pfn));
	memset(arch_zone_highest_possible_pfn, 0,
				sizeof(arch_zone_highest_possible_pfn));
	arch_zone_lowest_possible_pfn[0] = find_min_pfn_with_active_regions();
	arch_zone_highest_possible_pfn[0] = max_zone_pfn[0];
	for (i = 1; i < MAX_NR_ZONES; i++) {
M
Mel Gorman 已提交
4212 4213
		if (i == ZONE_MOVABLE)
			continue;
4214 4215 4216 4217 4218
		arch_zone_lowest_possible_pfn[i] =
			arch_zone_highest_possible_pfn[i-1];
		arch_zone_highest_possible_pfn[i] =
			max(max_zone_pfn[i], arch_zone_lowest_possible_pfn[i]);
	}
M
Mel Gorman 已提交
4219 4220 4221 4222 4223 4224
	arch_zone_lowest_possible_pfn[ZONE_MOVABLE] = 0;
	arch_zone_highest_possible_pfn[ZONE_MOVABLE] = 0;

	/* Find the PFNs that ZONE_MOVABLE begins at in each node */
	memset(zone_movable_pfn, 0, sizeof(zone_movable_pfn));
	find_zone_movable_pfns_for_nodes(zone_movable_pfn);
4225 4226 4227

	/* Print out the zone ranges */
	printk("Zone PFN ranges:\n");
M
Mel Gorman 已提交
4228 4229 4230
	for (i = 0; i < MAX_NR_ZONES; i++) {
		if (i == ZONE_MOVABLE)
			continue;
4231
		printk("  %-8s %0#10lx -> %0#10lx\n",
4232 4233 4234
				zone_names[i],
				arch_zone_lowest_possible_pfn[i],
				arch_zone_highest_possible_pfn[i]);
M
Mel Gorman 已提交
4235 4236 4237 4238 4239 4240 4241 4242
	}

	/* Print out the PFNs ZONE_MOVABLE begins at in each node */
	printk("Movable zone start PFN for each node\n");
	for (i = 0; i < MAX_NUMNODES; i++) {
		if (zone_movable_pfn[i])
			printk("  Node %d: %lu\n", i, zone_movable_pfn[i]);
	}
4243 4244 4245 4246

	/* Print out the early_node_map[] */
	printk("early_node_map[%d] active PFN ranges\n", nr_nodemap_entries);
	for (i = 0; i < nr_nodemap_entries; i++)
4247
		printk("  %3d: %0#10lx -> %0#10lx\n", early_node_map[i].nid,
4248 4249 4250 4251
						early_node_map[i].start_pfn,
						early_node_map[i].end_pfn);

	/* Initialise every node */
4252
	mminit_verify_pageflags_layout();
4253
	setup_nr_node_ids();
4254 4255
	for_each_online_node(nid) {
		pg_data_t *pgdat = NODE_DATA(nid);
4256
		free_area_init_node(nid, NULL,
4257
				find_min_pfn_for_node(nid), NULL);
4258 4259 4260 4261 4262

		/* Any memory on that node */
		if (pgdat->node_present_pages)
			node_set_state(nid, N_HIGH_MEMORY);
		check_for_regular_memory(pgdat);
4263 4264
	}
}
M
Mel Gorman 已提交
4265

4266
static int __init cmdline_parse_core(char *p, unsigned long *core)
M
Mel Gorman 已提交
4267 4268 4269 4270 4271 4272
{
	unsigned long long coremem;
	if (!p)
		return -EINVAL;

	coremem = memparse(p, &p);
4273
	*core = coremem >> PAGE_SHIFT;
M
Mel Gorman 已提交
4274

4275
	/* Paranoid check that UL is enough for the coremem value */
M
Mel Gorman 已提交
4276 4277 4278 4279
	WARN_ON((coremem >> PAGE_SHIFT) > ULONG_MAX);

	return 0;
}
M
Mel Gorman 已提交
4280

4281 4282 4283 4284 4285 4286 4287 4288 4289 4290 4291 4292 4293 4294 4295 4296 4297 4298
/*
 * kernelcore=size sets the amount of memory for use for allocations that
 * cannot be reclaimed or migrated.
 */
static int __init cmdline_parse_kernelcore(char *p)
{
	return cmdline_parse_core(p, &required_kernelcore);
}

/*
 * movablecore=size sets the amount of memory for use for allocations that
 * can be reclaimed or migrated.
 */
static int __init cmdline_parse_movablecore(char *p)
{
	return cmdline_parse_core(p, &required_movablecore);
}

M
Mel Gorman 已提交
4299
early_param("kernelcore", cmdline_parse_kernelcore);
4300
early_param("movablecore", cmdline_parse_movablecore);
M
Mel Gorman 已提交
4301

4302 4303
#endif /* CONFIG_ARCH_POPULATES_NODE_MAP */

4304
/**
4305 4306
 * set_dma_reserve - set the specified number of pages reserved in the first zone
 * @new_dma_reserve: The number of pages to mark reserved
4307 4308 4309 4310
 *
 * The per-cpu batchsize and zone watermarks are determined by present_pages.
 * In the DMA zone, a significant percentage may be consumed by kernel image
 * and other unfreeable allocations which can skew the watermarks badly. This
4311 4312 4313
 * function may optionally be used to account for unfreeable pages in the
 * first zone (e.g., ZONE_DMA). The effect will be lower watermarks and
 * smaller per-cpu batchsize.
4314 4315 4316 4317 4318 4319
 */
void __init set_dma_reserve(unsigned long new_dma_reserve)
{
	dma_reserve = new_dma_reserve;
}

4320
#ifndef CONFIG_NEED_MULTIPLE_NODES
4321
struct pglist_data __refdata contig_page_data = { .bdata = &bootmem_node_data[0] };
L
Linus Torvalds 已提交
4322
EXPORT_SYMBOL(contig_page_data);
4323
#endif
L
Linus Torvalds 已提交
4324 4325 4326

void __init free_area_init(unsigned long *zones_size)
{
4327
	free_area_init_node(0, zones_size,
L
Linus Torvalds 已提交
4328 4329 4330 4331 4332 4333 4334 4335
			__pa(PAGE_OFFSET) >> PAGE_SHIFT, NULL);
}

static int page_alloc_cpu_notify(struct notifier_block *self,
				 unsigned long action, void *hcpu)
{
	int cpu = (unsigned long)hcpu;

4336
	if (action == CPU_DEAD || action == CPU_DEAD_FROZEN) {
4337 4338 4339 4340 4341 4342 4343 4344
		drain_pages(cpu);

		/*
		 * Spill the event counters of the dead processor
		 * into the current processors event counters.
		 * This artificially elevates the count of the current
		 * processor.
		 */
4345
		vm_events_fold_cpu(cpu);
4346 4347 4348 4349 4350 4351 4352 4353

		/*
		 * Zero the differential counters of the dead processor
		 * so that the vm statistics are consistent.
		 *
		 * This is only okay since the processor is dead and cannot
		 * race with what we are doing.
		 */
4354
		refresh_cpu_vm_stats(cpu);
L
Linus Torvalds 已提交
4355 4356 4357 4358 4359 4360 4361 4362 4363
	}
	return NOTIFY_OK;
}

void __init page_alloc_init(void)
{
	hotcpu_notifier(page_alloc_cpu_notify, 0);
}

4364 4365 4366 4367 4368 4369 4370 4371
/*
 * calculate_totalreserve_pages - called when sysctl_lower_zone_reserve_ratio
 *	or min_free_kbytes changes.
 */
static void calculate_totalreserve_pages(void)
{
	struct pglist_data *pgdat;
	unsigned long reserve_pages = 0;
4372
	enum zone_type i, j;
4373 4374 4375 4376 4377 4378 4379 4380 4381 4382 4383 4384

	for_each_online_pgdat(pgdat) {
		for (i = 0; i < MAX_NR_ZONES; i++) {
			struct zone *zone = pgdat->node_zones + i;
			unsigned long max = 0;

			/* Find valid and maximum lowmem_reserve in the zone */
			for (j = i; j < MAX_NR_ZONES; j++) {
				if (zone->lowmem_reserve[j] > max)
					max = zone->lowmem_reserve[j];
			}

4385 4386
			/* we treat the high watermark as reserved pages. */
			max += high_wmark_pages(zone);
4387 4388 4389 4390 4391 4392 4393 4394 4395

			if (max > zone->present_pages)
				max = zone->present_pages;
			reserve_pages += max;
		}
	}
	totalreserve_pages = reserve_pages;
}

L
Linus Torvalds 已提交
4396 4397 4398 4399 4400 4401 4402 4403 4404
/*
 * setup_per_zone_lowmem_reserve - called whenever
 *	sysctl_lower_zone_reserve_ratio changes.  Ensures that each zone
 *	has a correct pages reserved value, so an adequate number of
 *	pages are left in the zone after a successful __alloc_pages().
 */
static void setup_per_zone_lowmem_reserve(void)
{
	struct pglist_data *pgdat;
4405
	enum zone_type j, idx;
L
Linus Torvalds 已提交
4406

4407
	for_each_online_pgdat(pgdat) {
L
Linus Torvalds 已提交
4408 4409 4410 4411 4412 4413
		for (j = 0; j < MAX_NR_ZONES; j++) {
			struct zone *zone = pgdat->node_zones + j;
			unsigned long present_pages = zone->present_pages;

			zone->lowmem_reserve[j] = 0;

4414 4415
			idx = j;
			while (idx) {
L
Linus Torvalds 已提交
4416 4417
				struct zone *lower_zone;

4418 4419
				idx--;

L
Linus Torvalds 已提交
4420 4421 4422 4423 4424 4425 4426 4427 4428 4429
				if (sysctl_lowmem_reserve_ratio[idx] < 1)
					sysctl_lowmem_reserve_ratio[idx] = 1;

				lower_zone = pgdat->node_zones + idx;
				lower_zone->lowmem_reserve[j] = present_pages /
					sysctl_lowmem_reserve_ratio[idx];
				present_pages += lower_zone->present_pages;
			}
		}
	}
4430 4431 4432

	/* update totalreserve_pages */
	calculate_totalreserve_pages();
L
Linus Torvalds 已提交
4433 4434
}

4435
/**
4436
 * setup_per_zone_wmarks - called when min_free_kbytes changes
4437
 * or when memory is hot-{added|removed}
4438
 *
4439 4440
 * Ensures that the watermark[min,low,high] values for each zone are set
 * correctly with respect to min_free_kbytes.
L
Linus Torvalds 已提交
4441
 */
4442
void setup_per_zone_wmarks(void)
L
Linus Torvalds 已提交
4443 4444 4445 4446 4447 4448 4449 4450 4451 4452 4453 4454 4455
{
	unsigned long pages_min = min_free_kbytes >> (PAGE_SHIFT - 10);
	unsigned long lowmem_pages = 0;
	struct zone *zone;
	unsigned long flags;

	/* Calculate total number of !ZONE_HIGHMEM pages */
	for_each_zone(zone) {
		if (!is_highmem(zone))
			lowmem_pages += zone->present_pages;
	}

	for_each_zone(zone) {
4456 4457
		u64 tmp;

4458
		spin_lock_irqsave(&zone->lock, flags);
4459 4460
		tmp = (u64)pages_min * zone->present_pages;
		do_div(tmp, lowmem_pages);
L
Linus Torvalds 已提交
4461 4462
		if (is_highmem(zone)) {
			/*
N
Nick Piggin 已提交
4463 4464 4465 4466
			 * __GFP_HIGH and PF_MEMALLOC allocations usually don't
			 * need highmem pages, so cap pages_min to a small
			 * value here.
			 *
4467
			 * The WMARK_HIGH-WMARK_LOW and (WMARK_LOW-WMARK_MIN)
N
Nick Piggin 已提交
4468 4469
			 * deltas controls asynch page reclaim, and so should
			 * not be capped for highmem.
L
Linus Torvalds 已提交
4470 4471 4472 4473 4474 4475 4476 4477
			 */
			int min_pages;

			min_pages = zone->present_pages / 1024;
			if (min_pages < SWAP_CLUSTER_MAX)
				min_pages = SWAP_CLUSTER_MAX;
			if (min_pages > 128)
				min_pages = 128;
4478
			zone->watermark[WMARK_MIN] = min_pages;
L
Linus Torvalds 已提交
4479
		} else {
N
Nick Piggin 已提交
4480 4481
			/*
			 * If it's a lowmem zone, reserve a number of pages
L
Linus Torvalds 已提交
4482 4483
			 * proportionate to the zone's size.
			 */
4484
			zone->watermark[WMARK_MIN] = tmp;
L
Linus Torvalds 已提交
4485 4486
		}

4487 4488
		zone->watermark[WMARK_LOW]  = min_wmark_pages(zone) + (tmp >> 2);
		zone->watermark[WMARK_HIGH] = min_wmark_pages(zone) + (tmp >> 1);
4489
		setup_zone_migrate_reserve(zone);
4490
		spin_unlock_irqrestore(&zone->lock, flags);
L
Linus Torvalds 已提交
4491
	}
4492 4493 4494

	/* update totalreserve_pages */
	calculate_totalreserve_pages();
L
Linus Torvalds 已提交
4495 4496
}

4497 4498 4499 4500 4501 4502 4503 4504 4505 4506 4507 4508 4509 4510 4511 4512 4513 4514 4515 4516 4517
/**
 * The inactive anon list should be small enough that the VM never has to
 * do too much work, but large enough that each inactive page has a chance
 * to be referenced again before it is swapped out.
 *
 * The inactive_anon ratio is the target ratio of ACTIVE_ANON to
 * INACTIVE_ANON pages on this zone's LRU, maintained by the
 * pageout code. A zone->inactive_ratio of 3 means 3:1 or 25% of
 * the anonymous pages are kept on the inactive list.
 *
 * total     target    max
 * memory    ratio     inactive anon
 * -------------------------------------
 *   10MB       1         5MB
 *  100MB       1        50MB
 *    1GB       3       250MB
 *   10GB      10       0.9GB
 *  100GB      31         3GB
 *    1TB     101        10GB
 *   10TB     320        32GB
 */
4518
void calculate_zone_inactive_ratio(struct zone *zone)
4519
{
4520
	unsigned int gb, ratio;
4521

4522 4523 4524
	/* Zone size in gigabytes */
	gb = zone->present_pages >> (30 - PAGE_SHIFT);
	if (gb)
4525
		ratio = int_sqrt(10 * gb);
4526 4527
	else
		ratio = 1;
4528

4529 4530
	zone->inactive_ratio = ratio;
}
4531

4532 4533 4534 4535 4536 4537
static void __init setup_per_zone_inactive_ratio(void)
{
	struct zone *zone;

	for_each_zone(zone)
		calculate_zone_inactive_ratio(zone);
4538 4539
}

L
Linus Torvalds 已提交
4540 4541 4542 4543 4544 4545 4546 4547 4548 4549 4550 4551 4552 4553 4554 4555 4556 4557 4558 4559 4560 4561 4562 4563
/*
 * Initialise min_free_kbytes.
 *
 * For small machines we want it small (128k min).  For large machines
 * we want it large (64MB max).  But it is not linear, because network
 * bandwidth does not increase linearly with machine size.  We use
 *
 * 	min_free_kbytes = 4 * sqrt(lowmem_kbytes), for better accuracy:
 *	min_free_kbytes = sqrt(lowmem_kbytes * 16)
 *
 * which yields
 *
 * 16MB:	512k
 * 32MB:	724k
 * 64MB:	1024k
 * 128MB:	1448k
 * 256MB:	2048k
 * 512MB:	2896k
 * 1024MB:	4096k
 * 2048MB:	5792k
 * 4096MB:	8192k
 * 8192MB:	11584k
 * 16384MB:	16384k
 */
4564
static int __init init_per_zone_wmark_min(void)
L
Linus Torvalds 已提交
4565 4566 4567 4568 4569 4570 4571 4572 4573 4574
{
	unsigned long lowmem_kbytes;

	lowmem_kbytes = nr_free_buffer_pages() * (PAGE_SIZE >> 10);

	min_free_kbytes = int_sqrt(lowmem_kbytes * 16);
	if (min_free_kbytes < 128)
		min_free_kbytes = 128;
	if (min_free_kbytes > 65536)
		min_free_kbytes = 65536;
4575
	setup_per_zone_wmarks();
L
Linus Torvalds 已提交
4576
	setup_per_zone_lowmem_reserve();
4577
	setup_per_zone_inactive_ratio();
L
Linus Torvalds 已提交
4578 4579
	return 0;
}
4580
module_init(init_per_zone_wmark_min)
L
Linus Torvalds 已提交
4581 4582 4583 4584 4585 4586 4587 4588 4589 4590

/*
 * min_free_kbytes_sysctl_handler - just a wrapper around proc_dointvec() so 
 *	that we can call two helper functions whenever min_free_kbytes
 *	changes.
 */
int min_free_kbytes_sysctl_handler(ctl_table *table, int write, 
	struct file *file, void __user *buffer, size_t *length, loff_t *ppos)
{
	proc_dointvec(table, write, file, buffer, length, ppos);
4591
	if (write)
4592
		setup_per_zone_wmarks();
L
Linus Torvalds 已提交
4593 4594 4595
	return 0;
}

4596 4597 4598 4599 4600 4601 4602 4603 4604 4605 4606 4607
#ifdef CONFIG_NUMA
int sysctl_min_unmapped_ratio_sysctl_handler(ctl_table *table, int write,
	struct file *file, void __user *buffer, size_t *length, loff_t *ppos)
{
	struct zone *zone;
	int rc;

	rc = proc_dointvec_minmax(table, write, file, buffer, length, ppos);
	if (rc)
		return rc;

	for_each_zone(zone)
4608
		zone->min_unmapped_pages = (zone->present_pages *
4609 4610 4611
				sysctl_min_unmapped_ratio) / 100;
	return 0;
}
4612 4613 4614 4615 4616 4617 4618 4619 4620 4621 4622 4623 4624 4625 4626 4627

int sysctl_min_slab_ratio_sysctl_handler(ctl_table *table, int write,
	struct file *file, void __user *buffer, size_t *length, loff_t *ppos)
{
	struct zone *zone;
	int rc;

	rc = proc_dointvec_minmax(table, write, file, buffer, length, ppos);
	if (rc)
		return rc;

	for_each_zone(zone)
		zone->min_slab_pages = (zone->present_pages *
				sysctl_min_slab_ratio) / 100;
	return 0;
}
4628 4629
#endif

L
Linus Torvalds 已提交
4630 4631 4632 4633 4634 4635
/*
 * lowmem_reserve_ratio_sysctl_handler - just a wrapper around
 *	proc_dointvec() so that we can call setup_per_zone_lowmem_reserve()
 *	whenever sysctl_lowmem_reserve_ratio changes.
 *
 * The reserve ratio obviously has absolutely no relation with the
4636
 * minimum watermarks. The lowmem reserve ratio can only make sense
L
Linus Torvalds 已提交
4637 4638 4639 4640 4641 4642 4643 4644 4645 4646
 * if in function of the boot time zone sizes.
 */
int lowmem_reserve_ratio_sysctl_handler(ctl_table *table, int write,
	struct file *file, void __user *buffer, size_t *length, loff_t *ppos)
{
	proc_dointvec_minmax(table, write, file, buffer, length, ppos);
	setup_per_zone_lowmem_reserve();
	return 0;
}

4647 4648 4649 4650 4651 4652 4653 4654 4655 4656 4657 4658 4659 4660 4661 4662
/*
 * percpu_pagelist_fraction - changes the pcp->high for each zone on each
 * cpu.  It is the fraction of total pages in each zone that a hot per cpu pagelist
 * can have before it gets flushed back to buddy allocator.
 */

int percpu_pagelist_fraction_sysctl_handler(ctl_table *table, int write,
	struct file *file, void __user *buffer, size_t *length, loff_t *ppos)
{
	struct zone *zone;
	unsigned int cpu;
	int ret;

	ret = proc_dointvec_minmax(table, write, file, buffer, length, ppos);
	if (!write || (ret == -EINVAL))
		return ret;
4663
	for_each_populated_zone(zone) {
4664 4665 4666 4667 4668 4669 4670 4671 4672
		for_each_online_cpu(cpu) {
			unsigned long  high;
			high = zone->present_pages / percpu_pagelist_fraction;
			setup_pagelist_highmark(zone_pcp(zone, cpu), high);
		}
	}
	return 0;
}

4673
int hashdist = HASHDIST_DEFAULT;
L
Linus Torvalds 已提交
4674 4675 4676 4677 4678 4679 4680 4681 4682 4683 4684 4685 4686 4687 4688 4689 4690 4691 4692 4693 4694 4695 4696 4697 4698 4699 4700 4701 4702 4703 4704 4705 4706 4707

#ifdef CONFIG_NUMA
static int __init set_hashdist(char *str)
{
	if (!str)
		return 0;
	hashdist = simple_strtoul(str, &str, 0);
	return 1;
}
__setup("hashdist=", set_hashdist);
#endif

/*
 * allocate a large system hash table from bootmem
 * - it is assumed that the hash table must contain an exact power-of-2
 *   quantity of entries
 * - limit is the number of hash buckets, not the total allocation size
 */
void *__init alloc_large_system_hash(const char *tablename,
				     unsigned long bucketsize,
				     unsigned long numentries,
				     int scale,
				     int flags,
				     unsigned int *_hash_shift,
				     unsigned int *_hash_mask,
				     unsigned long limit)
{
	unsigned long long max = limit;
	unsigned long log2qty, size;
	void *table = NULL;

	/* allow the kernel cmdline to have a say */
	if (!numentries) {
		/* round applicable memory size up to nearest megabyte */
A
Andrew Morton 已提交
4708
		numentries = nr_kernel_pages;
L
Linus Torvalds 已提交
4709 4710 4711 4712 4713 4714 4715 4716 4717
		numentries += (1UL << (20 - PAGE_SHIFT)) - 1;
		numentries >>= 20 - PAGE_SHIFT;
		numentries <<= 20 - PAGE_SHIFT;

		/* limit to 1 bucket per 2^scale bytes of low memory */
		if (scale > PAGE_SHIFT)
			numentries >>= (scale - PAGE_SHIFT);
		else
			numentries <<= (PAGE_SHIFT - scale);
4718 4719 4720 4721

		/* Make sure we've got at least a 0-order allocation.. */
		if (unlikely((numentries * bucketsize) < PAGE_SIZE))
			numentries = PAGE_SIZE / bucketsize;
L
Linus Torvalds 已提交
4722
	}
4723
	numentries = roundup_pow_of_two(numentries);
L
Linus Torvalds 已提交
4724 4725 4726 4727 4728 4729 4730 4731 4732 4733

	/* limit allocation size to 1/16 total memory by default */
	if (max == 0) {
		max = ((unsigned long long)nr_all_pages << PAGE_SHIFT) >> 4;
		do_div(max, bucketsize);
	}

	if (numentries > max)
		numentries = max;

4734
	log2qty = ilog2(numentries);
L
Linus Torvalds 已提交
4735 4736 4737 4738

	do {
		size = bucketsize << log2qty;
		if (flags & HASH_EARLY)
4739
			table = alloc_bootmem_nopanic(size);
L
Linus Torvalds 已提交
4740 4741 4742
		else if (hashdist)
			table = __vmalloc(size, GFP_ATOMIC, PAGE_KERNEL);
		else {
4743 4744
			/*
			 * If bucketsize is not a power-of-two, we may free
4745 4746
			 * some pages at the end of hash table which
			 * alloc_pages_exact() automatically does
4747
			 */
4748 4749
			if (get_order(size) < MAX_ORDER)
				table = alloc_pages_exact(size, GFP_ATOMIC);
L
Linus Torvalds 已提交
4750 4751 4752 4753 4754 4755
		}
	} while (!table && size > PAGE_SIZE && --log2qty);

	if (!table)
		panic("Failed to allocate %s hash table\n", tablename);

4756
	printk(KERN_INFO "%s hash table entries: %d (order: %d, %lu bytes)\n",
L
Linus Torvalds 已提交
4757 4758
	       tablename,
	       (1U << log2qty),
4759
	       ilog2(size) - PAGE_SHIFT,
L
Linus Torvalds 已提交
4760 4761 4762 4763 4764 4765 4766
	       size);

	if (_hash_shift)
		*_hash_shift = log2qty;
	if (_hash_mask)
		*_hash_mask = (1 << log2qty) - 1;

4767 4768 4769 4770 4771 4772 4773 4774 4775 4776
	/*
	 * If hashdist is set, the table allocation is done with __vmalloc()
	 * which invokes the kmemleak_alloc() callback. This function may also
	 * be called before the slab and kmemleak are initialised when
	 * kmemleak simply buffers the request to be executed later
	 * (GFP_ATOMIC flag ignored in this case).
	 */
	if (!hashdist)
		kmemleak_alloc(table, size, 1, GFP_ATOMIC);

L
Linus Torvalds 已提交
4777 4778
	return table;
}
4779

4780 4781 4782 4783 4784 4785 4786 4787 4788 4789 4790 4791 4792 4793 4794
/* Return a pointer to the bitmap storing bits affecting a block of pages */
static inline unsigned long *get_pageblock_bitmap(struct zone *zone,
							unsigned long pfn)
{
#ifdef CONFIG_SPARSEMEM
	return __pfn_to_section(pfn)->pageblock_flags;
#else
	return zone->pageblock_flags;
#endif /* CONFIG_SPARSEMEM */
}

static inline int pfn_to_bitidx(struct zone *zone, unsigned long pfn)
{
#ifdef CONFIG_SPARSEMEM
	pfn &= (PAGES_PER_SECTION-1);
4795
	return (pfn >> pageblock_order) * NR_PAGEBLOCK_BITS;
4796 4797
#else
	pfn = pfn - zone->zone_start_pfn;
4798
	return (pfn >> pageblock_order) * NR_PAGEBLOCK_BITS;
4799 4800 4801 4802
#endif /* CONFIG_SPARSEMEM */
}

/**
4803
 * get_pageblock_flags_group - Return the requested group of flags for the pageblock_nr_pages block of pages
4804 4805 4806 4807 4808 4809 4810 4811 4812 4813 4814 4815 4816 4817 4818 4819 4820 4821 4822 4823 4824 4825
 * @page: The page within the block of interest
 * @start_bitidx: The first bit of interest to retrieve
 * @end_bitidx: The last bit of interest
 * returns pageblock_bits flags
 */
unsigned long get_pageblock_flags_group(struct page *page,
					int start_bitidx, int end_bitidx)
{
	struct zone *zone;
	unsigned long *bitmap;
	unsigned long pfn, bitidx;
	unsigned long flags = 0;
	unsigned long value = 1;

	zone = page_zone(page);
	pfn = page_to_pfn(page);
	bitmap = get_pageblock_bitmap(zone, pfn);
	bitidx = pfn_to_bitidx(zone, pfn);

	for (; start_bitidx <= end_bitidx; start_bitidx++, value <<= 1)
		if (test_bit(bitidx + start_bitidx, bitmap))
			flags |= value;
4826

4827 4828 4829 4830
	return flags;
}

/**
4831
 * set_pageblock_flags_group - Set the requested group of flags for a pageblock_nr_pages block of pages
4832 4833 4834 4835 4836 4837 4838 4839 4840 4841 4842 4843 4844 4845 4846 4847 4848
 * @page: The page within the block of interest
 * @start_bitidx: The first bit of interest
 * @end_bitidx: The last bit of interest
 * @flags: The flags to set
 */
void set_pageblock_flags_group(struct page *page, unsigned long flags,
					int start_bitidx, int end_bitidx)
{
	struct zone *zone;
	unsigned long *bitmap;
	unsigned long pfn, bitidx;
	unsigned long value = 1;

	zone = page_zone(page);
	pfn = page_to_pfn(page);
	bitmap = get_pageblock_bitmap(zone, pfn);
	bitidx = pfn_to_bitidx(zone, pfn);
4849 4850
	VM_BUG_ON(pfn < zone->zone_start_pfn);
	VM_BUG_ON(pfn >= zone->zone_start_pfn + zone->spanned_pages);
4851 4852 4853 4854 4855 4856 4857

	for (; start_bitidx <= end_bitidx; start_bitidx++, value <<= 1)
		if (flags & value)
			__set_bit(bitidx + start_bitidx, bitmap);
		else
			__clear_bit(bitidx + start_bitidx, bitmap);
}
K
KAMEZAWA Hiroyuki 已提交
4858 4859 4860 4861 4862 4863 4864 4865 4866 4867 4868 4869 4870 4871 4872 4873 4874 4875 4876 4877 4878 4879 4880 4881 4882 4883

/*
 * This is designed as sub function...plz see page_isolation.c also.
 * set/clear page block's type to be ISOLATE.
 * page allocater never alloc memory from ISOLATE block.
 */

int set_migratetype_isolate(struct page *page)
{
	struct zone *zone;
	unsigned long flags;
	int ret = -EBUSY;

	zone = page_zone(page);
	spin_lock_irqsave(&zone->lock, flags);
	/*
	 * In future, more migrate types will be able to be isolation target.
	 */
	if (get_pageblock_migratetype(page) != MIGRATE_MOVABLE)
		goto out;
	set_pageblock_migratetype(page, MIGRATE_ISOLATE);
	move_freepages_block(zone, page, MIGRATE_ISOLATE);
	ret = 0;
out:
	spin_unlock_irqrestore(&zone->lock, flags);
	if (!ret)
4884
		drain_all_pages();
K
KAMEZAWA Hiroyuki 已提交
4885 4886 4887 4888 4889 4890 4891 4892 4893 4894 4895 4896 4897 4898 4899 4900
	return ret;
}

void unset_migratetype_isolate(struct page *page)
{
	struct zone *zone;
	unsigned long flags;
	zone = page_zone(page);
	spin_lock_irqsave(&zone->lock, flags);
	if (get_pageblock_migratetype(page) != MIGRATE_ISOLATE)
		goto out;
	set_pageblock_migratetype(page, MIGRATE_MOVABLE);
	move_freepages_block(zone, page, MIGRATE_MOVABLE);
out:
	spin_unlock_irqrestore(&zone->lock, flags);
}
K
KAMEZAWA Hiroyuki 已提交
4901 4902 4903 4904 4905 4906 4907 4908 4909 4910 4911 4912 4913 4914 4915 4916 4917 4918 4919 4920 4921 4922 4923 4924 4925 4926 4927 4928 4929 4930 4931 4932 4933 4934 4935 4936 4937 4938 4939 4940 4941 4942 4943 4944 4945 4946 4947

#ifdef CONFIG_MEMORY_HOTREMOVE
/*
 * All pages in the range must be isolated before calling this.
 */
void
__offline_isolated_pages(unsigned long start_pfn, unsigned long end_pfn)
{
	struct page *page;
	struct zone *zone;
	int order, i;
	unsigned long pfn;
	unsigned long flags;
	/* find the first valid pfn */
	for (pfn = start_pfn; pfn < end_pfn; pfn++)
		if (pfn_valid(pfn))
			break;
	if (pfn == end_pfn)
		return;
	zone = page_zone(pfn_to_page(pfn));
	spin_lock_irqsave(&zone->lock, flags);
	pfn = start_pfn;
	while (pfn < end_pfn) {
		if (!pfn_valid(pfn)) {
			pfn++;
			continue;
		}
		page = pfn_to_page(pfn);
		BUG_ON(page_count(page));
		BUG_ON(!PageBuddy(page));
		order = page_order(page);
#ifdef CONFIG_DEBUG_VM
		printk(KERN_INFO "remove from free list %lx %d %lx\n",
		       pfn, 1 << order, end_pfn);
#endif
		list_del(&page->lru);
		rmv_page_order(page);
		zone->free_area[order].nr_free--;
		__mod_zone_page_state(zone, NR_FREE_PAGES,
				      - (1UL << order));
		for (i = 0; i < (1 << order); i++)
			SetPageReserved((page+i));
		pfn += (1 << order);
	}
	spin_unlock_irqrestore(&zone->lock, flags);
}
#endif