page_alloc.c 153.5 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
#include <linux/bootmem.h>
24
#include <linux/memblock.h>
L
Linus Torvalds 已提交
25
#include <linux/compiler.h>
26
#include <linux/kernel.h>
27
#include <linux/kmemcheck.h>
L
Linus Torvalds 已提交
28 29 30 31 32
#include <linux/module.h>
#include <linux/suspend.h>
#include <linux/pagevec.h>
#include <linux/blkdev.h>
#include <linux/slab.h>
33
#include <linux/oom.h>
L
Linus Torvalds 已提交
34 35 36 37 38
#include <linux/notifier.h>
#include <linux/topology.h>
#include <linux/sysctl.h>
#include <linux/cpu.h>
#include <linux/cpuset.h>
39
#include <linux/memory_hotplug.h>
L
Linus Torvalds 已提交
40 41
#include <linux/nodemask.h>
#include <linux/vmalloc.h>
42
#include <linux/mempolicy.h>
43
#include <linux/stop_machine.h>
44 45
#include <linux/sort.h>
#include <linux/pfn.h>
46
#include <linux/backing-dev.h>
47
#include <linux/fault-inject.h>
K
KAMEZAWA Hiroyuki 已提交
48
#include <linux/page-isolation.h>
49
#include <linux/page_cgroup.h>
50
#include <linux/debugobjects.h>
51
#include <linux/kmemleak.h>
52
#include <linux/memory.h>
53
#include <linux/compaction.h>
54
#include <trace/events/kmem.h>
55
#include <linux/ftrace_event.h>
L
Linus Torvalds 已提交
56 57

#include <asm/tlbflush.h>
58
#include <asm/div64.h>
L
Linus Torvalds 已提交
59 60
#include "internal.h"

61 62 63 64 65
#ifdef CONFIG_USE_PERCPU_NUMA_NODE_ID
DEFINE_PER_CPU(int, numa_node);
EXPORT_PER_CPU_SYMBOL(numa_node);
#endif

66 67 68 69 70 71 72 73 74 75 76
#ifdef CONFIG_HAVE_MEMORYLESS_NODES
/*
 * N.B., Do NOT reference the '_numa_mem_' per cpu variable directly.
 * It will not be defined when CONFIG_HAVE_MEMORYLESS_NODES is not defined.
 * Use the accessor functions set_numa_mem(), numa_mem_id() and cpu_to_mem()
 * defined in <linux/topology.h>.
 */
DEFINE_PER_CPU(int, _numa_mem_);		/* Kernel "local memory" node */
EXPORT_PER_CPU_SYMBOL(_numa_mem_);
#endif

L
Linus Torvalds 已提交
77
/*
78
 * Array of node states.
L
Linus Torvalds 已提交
79
 */
80 81 82 83 84 85 86 87 88 89 90 91 92
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);

93
unsigned long totalram_pages __read_mostly;
94
unsigned long totalreserve_pages __read_mostly;
95
int percpu_pagelist_fraction;
96
gfp_t gfp_allowed_mask __read_mostly = GFP_BOOT_MASK;
L
Linus Torvalds 已提交
97

98 99 100 101 102 103 104 105 106
#ifdef CONFIG_PM_SLEEP
/*
 * The following functions are used by the suspend/hibernate code to temporarily
 * change gfp_allowed_mask in order to avoid using I/O during memory allocations
 * while devices are suspended.  To avoid races with the suspend/hibernate code,
 * they should always be called with pm_mutex held (gfp_allowed_mask also should
 * only be modified with pm_mutex held, unless the suspend/hibernate code is
 * guaranteed not to run in parallel with that modification).
 */
107 108 109 110

static gfp_t saved_gfp_mask;

void pm_restore_gfp_mask(void)
111 112
{
	WARN_ON(!mutex_is_locked(&pm_mutex));
113 114 115 116
	if (saved_gfp_mask) {
		gfp_allowed_mask = saved_gfp_mask;
		saved_gfp_mask = 0;
	}
117 118
}

119
void pm_restrict_gfp_mask(void)
120 121
{
	WARN_ON(!mutex_is_locked(&pm_mutex));
122 123 124
	WARN_ON(saved_gfp_mask);
	saved_gfp_mask = gfp_allowed_mask;
	gfp_allowed_mask &= ~GFP_IOFS;
125 126 127
}
#endif /* CONFIG_PM_SLEEP */

128 129 130 131
#ifdef CONFIG_HUGETLB_PAGE_SIZE_VARIABLE
int pageblock_order __read_mostly;
#endif

132
static void __free_pages_ok(struct page *page, unsigned int order);
133

L
Linus Torvalds 已提交
134 135 136 137 138 139 140
/*
 * 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
141 142 143
 *
 * 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
144
 */
145
int sysctl_lowmem_reserve_ratio[MAX_NR_ZONES-1] = {
146
#ifdef CONFIG_ZONE_DMA
147
	 256,
148
#endif
149
#ifdef CONFIG_ZONE_DMA32
150
	 256,
151
#endif
152
#ifdef CONFIG_HIGHMEM
M
Mel Gorman 已提交
153
	 32,
154
#endif
M
Mel Gorman 已提交
155
	 32,
156
};
L
Linus Torvalds 已提交
157 158 159

EXPORT_SYMBOL(totalram_pages);

160
static char * const zone_names[MAX_NR_ZONES] = {
161
#ifdef CONFIG_ZONE_DMA
162
	 "DMA",
163
#endif
164
#ifdef CONFIG_ZONE_DMA32
165
	 "DMA32",
166
#endif
167
	 "Normal",
168
#ifdef CONFIG_HIGHMEM
M
Mel Gorman 已提交
169
	 "HighMem",
170
#endif
M
Mel Gorman 已提交
171
	 "Movable",
172 173
};

L
Linus Torvalds 已提交
174 175
int min_free_kbytes = 1024;

176 177
static unsigned long __meminitdata nr_kernel_pages;
static unsigned long __meminitdata nr_all_pages;
178
static unsigned long __meminitdata dma_reserve;
L
Linus Torvalds 已提交
179

180 181
#ifdef CONFIG_ARCH_POPULATES_NODE_MAP
  /*
S
Simon Arlott 已提交
182
   * MAX_ACTIVE_REGIONS determines the maximum number of distinct
183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200
   * 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

201 202 203 204
  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
205
  static unsigned long __initdata required_kernelcore;
A
Adrian Bunk 已提交
206
  static unsigned long __initdata required_movablecore;
A
Adrian Bunk 已提交
207
  static unsigned long __meminitdata zone_movable_pfn[MAX_NUMNODES];
M
Mel Gorman 已提交
208 209 210 211

  /* movable_zone is the "real" zone pages in ZONE_MOVABLE are taken from */
  int movable_zone;
  EXPORT_SYMBOL(movable_zone);
212 213
#endif /* CONFIG_ARCH_POPULATES_NODE_MAP */

M
Miklos Szeredi 已提交
214 215
#if MAX_NUMNODES > 1
int nr_node_ids __read_mostly = MAX_NUMNODES;
216
int nr_online_nodes __read_mostly = 1;
M
Miklos Szeredi 已提交
217
EXPORT_SYMBOL(nr_node_ids);
218
EXPORT_SYMBOL(nr_online_nodes);
M
Miklos Szeredi 已提交
219 220
#endif

221 222
int page_group_by_mobility_disabled __read_mostly;

223 224
static void set_pageblock_migratetype(struct page *page, int migratetype)
{
225 226 227 228

	if (unlikely(page_group_by_mobility_disabled))
		migratetype = MIGRATE_UNMOVABLE;

229 230 231 232
	set_pageblock_flags_group(page, (unsigned long)migratetype,
					PB_migrate, PB_migrate_end);
}

233 234
bool oom_killer_disabled __read_mostly;

N
Nick Piggin 已提交
235
#ifdef CONFIG_DEBUG_VM
236
static int page_outside_zone_boundaries(struct zone *zone, struct page *page)
L
Linus Torvalds 已提交
237
{
238 239 240
	int ret = 0;
	unsigned seq;
	unsigned long pfn = page_to_pfn(page);
241

242 243 244 245 246 247 248 249 250
	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;
251 252 253 254
}

static int page_is_consistent(struct zone *zone, struct page *page)
{
255
	if (!pfn_valid_within(page_to_pfn(page)))
256
		return 0;
L
Linus Torvalds 已提交
257
	if (zone != page_zone(page))
258 259 260 261 262 263 264 265 266 267
		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
268
		return 1;
269 270 271
	if (!page_is_consistent(zone, page))
		return 1;

L
Linus Torvalds 已提交
272 273
	return 0;
}
N
Nick Piggin 已提交
274 275 276 277 278 279 280
#else
static inline int bad_range(struct zone *zone, struct page *page)
{
	return 0;
}
#endif

N
Nick Piggin 已提交
281
static void bad_page(struct page *page)
L
Linus Torvalds 已提交
282
{
283 284 285 286
	static unsigned long resume;
	static unsigned long nr_shown;
	static unsigned long nr_unshown;

287 288 289 290 291 292
	/* Don't complain about poisoned pages */
	if (PageHWPoison(page)) {
		__ClearPageBuddy(page);
		return;
	}

293 294 295 296 297 298 299 300 301 302
	/*
	 * 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) {
303 304
			printk(KERN_ALERT
			      "BUG: Bad page state: %lu messages suppressed\n",
305 306 307 308 309 310 311 312
				nr_unshown);
			nr_unshown = 0;
		}
		nr_shown = 0;
	}
	if (nr_shown++ == 0)
		resume = jiffies + 60 * HZ;

313
	printk(KERN_ALERT "BUG: Bad page state in process %s  pfn:%05lx\n",
314
		current->comm, page_to_pfn(page));
315
	dump_page(page);
316

L
Linus Torvalds 已提交
317
	dump_stack();
318
out:
319 320
	/* Leave bad fields for debug, except PageBuddy could make trouble */
	__ClearPageBuddy(page);
321
	add_taint(TAINT_BAD_PAGE);
L
Linus Torvalds 已提交
322 323 324 325 326 327 328 329 330 331 332 333
}

/*
 * 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).
 *
334 335 336
 * 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
337
 */
338 339 340

static void free_compound_page(struct page *page)
{
341
	__free_pages_ok(page, compound_order(page));
342 343
}

A
Andi Kleen 已提交
344
void prep_compound_page(struct page *page, unsigned long order)
345 346 347 348 349 350 351 352 353 354 355 356 357 358 359
{
	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;
	}
}

360
static int destroy_compound_page(struct page *page, unsigned long order)
L
Linus Torvalds 已提交
361 362 363
{
	int i;
	int nr_pages = 1 << order;
364
	int bad = 0;
L
Linus Torvalds 已提交
365

366 367
	if (unlikely(compound_order(page) != order) ||
	    unlikely(!PageHead(page))) {
N
Nick Piggin 已提交
368
		bad_page(page);
369 370
		bad++;
	}
L
Linus Torvalds 已提交
371

372
	__ClearPageHead(page);
373

374 375
	for (i = 1; i < nr_pages; i++) {
		struct page *p = page + i;
L
Linus Torvalds 已提交
376

377
		if (unlikely(!PageTail(p) || (p->first_page != page))) {
N
Nick Piggin 已提交
378
			bad_page(page);
379 380
			bad++;
		}
381
		__ClearPageTail(p);
L
Linus Torvalds 已提交
382
	}
383 384

	return bad;
L
Linus Torvalds 已提交
385 386
}

N
Nick Piggin 已提交
387 388 389 390
static inline void prep_zero_page(struct page *page, int order, gfp_t gfp_flags)
{
	int i;

391 392 393 394
	/*
	 * 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
395
	VM_BUG_ON((gfp_flags & __GFP_HIGHMEM) && in_interrupt());
N
Nick Piggin 已提交
396 397 398 399
	for (i = 0; i < (1 << order); i++)
		clear_highpage(page + i);
}

400 401
static inline void set_page_order(struct page *page, int order)
{
H
Hugh Dickins 已提交
402
	set_page_private(page, order);
403
	__SetPageBuddy(page);
L
Linus Torvalds 已提交
404 405 406 407
}

static inline void rmv_page_order(struct page *page)
{
408
	__ClearPageBuddy(page);
H
Hugh Dickins 已提交
409
	set_page_private(page, 0);
L
Linus Torvalds 已提交
410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426
}

/*
 * 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
427
 * Assumption: *_mem_map is contiguous at least up to MAX_ORDER
L
Linus Torvalds 已提交
428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445
 */
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
446
 * (a) the buddy is not in a hole &&
447
 * (b) the buddy is in the buddy system &&
448 449
 * (c) a page and its buddy have the same order &&
 * (d) a page and its buddy are in the same zone.
450 451 452
 *
 * 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
453
 *
454
 * For recording page's order, we use page_private(page).
L
Linus Torvalds 已提交
455
 */
456 457
static inline int page_is_buddy(struct page *page, struct page *buddy,
								int order)
L
Linus Torvalds 已提交
458
{
459
	if (!pfn_valid_within(page_to_pfn(buddy)))
N
Nick Piggin 已提交
460 461
		return 0;

462 463 464 465
	if (page_zone_id(page) != page_zone_id(buddy))
		return 0;

	if (PageBuddy(buddy) && page_order(buddy) == order) {
466
		VM_BUG_ON(page_count(buddy) != 0);
467
		return 1;
468
	}
469
	return 0;
L
Linus Torvalds 已提交
470 471 472 473 474 475 476 477 478 479 480 481 482 483 484
}

/*
 * 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
485
 * free pages of length of (1 << order) and marked with PG_buddy. Page's
H
Hugh Dickins 已提交
486
 * order is recorded in page_private(page) field.
L
Linus Torvalds 已提交
487 488 489 490 491 492 493 494 495
 * 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
496
static inline void __free_one_page(struct page *page,
497 498
		struct zone *zone, unsigned int order,
		int migratetype)
L
Linus Torvalds 已提交
499 500
{
	unsigned long page_idx;
501 502
	unsigned long combined_idx;
	struct page *buddy;
L
Linus Torvalds 已提交
503

N
Nick Piggin 已提交
504
	if (unlikely(PageCompound(page)))
505 506
		if (unlikely(destroy_compound_page(page, order)))
			return;
L
Linus Torvalds 已提交
507

508 509
	VM_BUG_ON(migratetype == -1);

L
Linus Torvalds 已提交
510 511
	page_idx = page_to_pfn(page) & ((1 << MAX_ORDER) - 1);

512
	VM_BUG_ON(page_idx & ((1 << order) - 1));
N
Nick Piggin 已提交
513
	VM_BUG_ON(bad_range(zone, page));
L
Linus Torvalds 已提交
514 515 516

	while (order < MAX_ORDER-1) {
		buddy = __page_find_buddy(page, page_idx, order);
517
		if (!page_is_buddy(page, buddy, order))
518
			break;
N
Nick Piggin 已提交
519

520
		/* Our buddy is free, merge with it and move up one order. */
L
Linus Torvalds 已提交
521
		list_del(&buddy->lru);
522
		zone->free_area[order].nr_free--;
L
Linus Torvalds 已提交
523
		rmv_page_order(buddy);
N
Nick Piggin 已提交
524
		combined_idx = __find_combined_index(page_idx, order);
L
Linus Torvalds 已提交
525 526 527 528 529
		page = page + (combined_idx - page_idx);
		page_idx = combined_idx;
		order++;
	}
	set_page_order(page, order);
530 531 532 533 534 535 536 537 538

	/*
	 * If this is not the largest possible page, check if the buddy
	 * of the next-highest order is free. If it is, it's possible
	 * that pages are being freed that will coalesce soon. In case,
	 * that is happening, add the free page to the tail of the list
	 * so it's less likely to be used soon and more likely to be merged
	 * as a higher order page
	 */
539
	if ((order < MAX_ORDER-2) && pfn_valid_within(page_to_pfn(buddy))) {
540 541 542 543 544 545 546 547 548 549 550 551 552
		struct page *higher_page, *higher_buddy;
		combined_idx = __find_combined_index(page_idx, order);
		higher_page = page + combined_idx - page_idx;
		higher_buddy = __page_find_buddy(higher_page, combined_idx, order + 1);
		if (page_is_buddy(higher_page, higher_buddy, order + 1)) {
			list_add_tail(&page->lru,
				&zone->free_area[order].free_list[migratetype]);
			goto out;
		}
	}

	list_add(&page->lru, &zone->free_area[order].free_list[migratetype]);
out:
L
Linus Torvalds 已提交
553 554 555
	zone->free_area[order].nr_free++;
}

556 557 558 559 560 561 562 563 564 565 566
/*
 * 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);
}

N
Nick Piggin 已提交
567
static inline int free_pages_check(struct page *page)
L
Linus Torvalds 已提交
568
{
N
Nick Piggin 已提交
569 570
	if (unlikely(page_mapcount(page) |
		(page->mapping != NULL)  |
571
		(atomic_read(&page->_count) != 0) |
572
		(page->flags & PAGE_FLAGS_CHECK_AT_FREE))) {
N
Nick Piggin 已提交
573
		bad_page(page);
574
		return 1;
575
	}
576 577 578
	if (page->flags & PAGE_FLAGS_CHECK_AT_PREP)
		page->flags &= ~PAGE_FLAGS_CHECK_AT_PREP;
	return 0;
L
Linus Torvalds 已提交
579 580 581
}

/*
582
 * Frees a number of pages from the PCP lists
L
Linus Torvalds 已提交
583
 * Assumes all pages on list are in same zone, and of same order.
584
 * count is the number of pages to free.
L
Linus Torvalds 已提交
585 586 587 588 589 590 591
 *
 * 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.
 */
592 593
static void free_pcppages_bulk(struct zone *zone, int count,
					struct per_cpu_pages *pcp)
L
Linus Torvalds 已提交
594
{
595
	int migratetype = 0;
596
	int batch_free = 0;
597
	int to_free = count;
598

N
Nick Piggin 已提交
599
	spin_lock(&zone->lock);
600
	zone->all_unreclaimable = 0;
L
Linus Torvalds 已提交
601
	zone->pages_scanned = 0;
602

603
	while (to_free) {
N
Nick Piggin 已提交
604
		struct page *page;
605 606 607
		struct list_head *list;

		/*
608 609 610 611 612
		 * Remove pages from lists in a round-robin fashion. A
		 * batch_free count is maintained that is incremented when an
		 * empty list is encountered.  This is so more pages are freed
		 * off fuller lists instead of spinning excessively around empty
		 * lists
613 614
		 */
		do {
615
			batch_free++;
616 617 618 619
			if (++migratetype == MIGRATE_PCPTYPES)
				migratetype = 0;
			list = &pcp->lists[migratetype];
		} while (list_empty(list));
N
Nick Piggin 已提交
620

621 622 623 624
		do {
			page = list_entry(list->prev, struct page, lru);
			/* must delete as __free_one_page list manipulates */
			list_del(&page->lru);
625 626 627
			/* MIGRATE_MOVABLE list may include MIGRATE_RESERVEs */
			__free_one_page(page, zone, 0, page_private(page));
			trace_mm_page_pcpu_drain(page, 0, page_private(page));
628
		} while (--to_free && --batch_free && !list_empty(list));
L
Linus Torvalds 已提交
629
	}
630
	__mod_zone_page_state(zone, NR_FREE_PAGES, count);
N
Nick Piggin 已提交
631
	spin_unlock(&zone->lock);
L
Linus Torvalds 已提交
632 633
}

634 635
static void free_one_page(struct zone *zone, struct page *page, int order,
				int migratetype)
L
Linus Torvalds 已提交
636
{
637
	spin_lock(&zone->lock);
638
	zone->all_unreclaimable = 0;
639
	zone->pages_scanned = 0;
640

641
	__free_one_page(page, zone, order, migratetype);
642
	__mod_zone_page_state(zone, NR_FREE_PAGES, 1 << order);
643
	spin_unlock(&zone->lock);
N
Nick Piggin 已提交
644 645
}

646
static bool free_pages_prepare(struct page *page, unsigned int order)
N
Nick Piggin 已提交
647
{
L
Linus Torvalds 已提交
648
	int i;
649
	int bad = 0;
L
Linus Torvalds 已提交
650

651
	trace_mm_page_free_direct(page, order);
652 653
	kmemcheck_free_shadow(page, order);

654 655 656 657 658 659 660
	for (i = 0; i < (1 << order); i++) {
		struct page *pg = page + i;

		if (PageAnon(pg))
			pg->mapping = NULL;
		bad += free_pages_check(pg);
	}
661
	if (bad)
662
		return false;
663

664
	if (!PageHighMem(page)) {
N
Nick Piggin 已提交
665
		debug_check_no_locks_freed(page_address(page),PAGE_SIZE<<order);
666 667 668
		debug_check_no_obj_freed(page_address(page),
					   PAGE_SIZE << order);
	}
N
Nick Piggin 已提交
669
	arch_free_page(page, order);
N
Nick Piggin 已提交
670
	kernel_map_pages(page, 1 << order, 0);
N
Nick Piggin 已提交
671

672 673 674 675 676 677 678 679 680 681 682
	return true;
}

static void __free_pages_ok(struct page *page, unsigned int order)
{
	unsigned long flags;
	int wasMlocked = __TestClearPageMlocked(page);

	if (!free_pages_prepare(page, order))
		return;

N
Nick Piggin 已提交
683
	local_irq_save(flags);
684
	if (unlikely(wasMlocked))
685
		free_page_mlock(page);
686
	__count_vm_events(PGFREE, 1 << order);
687 688
	free_one_page(page_zone(page), page, order,
					get_pageblock_migratetype(page));
N
Nick Piggin 已提交
689
	local_irq_restore(flags);
L
Linus Torvalds 已提交
690 691
}

692 693 694
/*
 * permit the bootmem allocator to evade page validation on high-order frees
 */
695
void __meminit __free_pages_bootmem(struct page *page, unsigned int order)
696 697 698 699
{
	if (order == 0) {
		__ClearPageReserved(page);
		set_page_count(page, 0);
700
		set_page_refcounted(page);
N
Nick Piggin 已提交
701
		__free_page(page);
702 703 704
	} else {
		int loop;

N
Nick Piggin 已提交
705
		prefetchw(page);
706 707 708
		for (loop = 0; loop < BITS_PER_LONG; loop++) {
			struct page *p = &page[loop];

N
Nick Piggin 已提交
709 710
			if (loop + 1 < BITS_PER_LONG)
				prefetchw(p + 1);
711 712 713 714
			__ClearPageReserved(p);
			set_page_count(p, 0);
		}

715
		set_page_refcounted(page);
N
Nick Piggin 已提交
716
		__free_pages(page, order);
717 718 719
	}
}

L
Linus Torvalds 已提交
720 721 722 723 724 725 726 727 728 729 730 731 732 733 734

/*
 * 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
735
static inline void expand(struct zone *zone, struct page *page,
736 737
	int low, int high, struct free_area *area,
	int migratetype)
L
Linus Torvalds 已提交
738 739 740 741 742 743 744
{
	unsigned long size = 1 << high;

	while (high > low) {
		area--;
		high--;
		size >>= 1;
N
Nick Piggin 已提交
745
		VM_BUG_ON(bad_range(zone, &page[size]));
746
		list_add(&page[size].lru, &area->free_list[migratetype]);
L
Linus Torvalds 已提交
747 748 749 750 751 752 753 754
		area->nr_free++;
		set_page_order(&page[size], high);
	}
}

/*
 * This page is about to be returned from the page allocator
 */
755
static inline int check_new_page(struct page *page)
L
Linus Torvalds 已提交
756
{
N
Nick Piggin 已提交
757 758
	if (unlikely(page_mapcount(page) |
		(page->mapping != NULL)  |
759
		(atomic_read(&page->_count) != 0)  |
760
		(page->flags & PAGE_FLAGS_CHECK_AT_PREP))) {
N
Nick Piggin 已提交
761
		bad_page(page);
762
		return 1;
763
	}
764 765 766 767 768 769 770 771 772 773 774 775
	return 0;
}

static int prep_new_page(struct page *page, int order, gfp_t gfp_flags)
{
	int i;

	for (i = 0; i < (1 << order); i++) {
		struct page *p = page + i;
		if (unlikely(check_new_page(p)))
			return 1;
	}
776

H
Hugh Dickins 已提交
777
	set_page_private(page, 0);
778
	set_page_refcounted(page);
N
Nick Piggin 已提交
779 780

	arch_alloc_page(page, order);
L
Linus Torvalds 已提交
781
	kernel_map_pages(page, 1 << order, 1);
N
Nick Piggin 已提交
782 783 784 785 786 787 788

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

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

789
	return 0;
L
Linus Torvalds 已提交
790 791
}

792 793 794 795
/*
 * Go through the free lists for the given migratetype and remove
 * the smallest available page from the freelists
 */
796 797
static inline
struct page *__rmqueue_smallest(struct zone *zone, unsigned int order,
798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822
						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;
}


823 824 825 826 827
/*
 * 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] = {
828 829 830 831
	[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 */
832 833
};

834 835
/*
 * Move the free pages in a range to the free lists of the requested type.
836
 * Note that start_page and end_pages are not aligned on a pageblock
837 838
 * boundary. If alignment is required, use move_freepages_block()
 */
A
Adrian Bunk 已提交
839 840 841
static int move_freepages(struct zone *zone,
			  struct page *start_page, struct page *end_page,
			  int migratetype)
842 843 844
{
	struct page *page;
	unsigned long order;
845
	int pages_moved = 0;
846 847 848 849 850 851 852

#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
853
	 * grouping pages by mobility
854 855 856 857 858
	 */
	BUG_ON(page_zone(start_page) != page_zone(end_page));
#endif

	for (page = start_page; page <= end_page;) {
859 860 861
		/* Make sure we are not inadvertently changing nodes */
		VM_BUG_ON(page_to_nid(page) != zone_to_nid(zone));

862 863 864 865 866 867 868 869 870 871 872 873 874 875 876
		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;
877
		pages_moved += 1 << order;
878 879
	}

880
	return pages_moved;
881 882
}

A
Adrian Bunk 已提交
883 884
static int move_freepages_block(struct zone *zone, struct page *page,
				int migratetype)
885 886 887 888 889
{
	unsigned long start_pfn, end_pfn;
	struct page *start_page, *end_page;

	start_pfn = page_to_pfn(page);
890
	start_pfn = start_pfn & ~(pageblock_nr_pages-1);
891
	start_page = pfn_to_page(start_pfn);
892 893
	end_page = start_page + pageblock_nr_pages - 1;
	end_pfn = start_pfn + pageblock_nr_pages - 1;
894 895 896 897 898 899 900 901 902 903

	/* 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);
}

904 905 906 907 908 909 910 911 912 913 914
static void change_pageblock_range(struct page *pageblock_page,
					int start_order, int migratetype)
{
	int nr_pageblocks = 1 << (start_order - pageblock_order);

	while (nr_pageblocks--) {
		set_pageblock_migratetype(pageblock_page, migratetype);
		pageblock_page += pageblock_nr_pages;
	}
}

915
/* Remove an element from the buddy allocator from the fallback list */
916 917
static inline struct page *
__rmqueue_fallback(struct zone *zone, int order, int start_migratetype)
918 919 920 921 922 923 924 925 926 927 928 929
{
	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];

930 931 932
			/* MIGRATE_RESERVE handled later if necessary */
			if (migratetype == MIGRATE_RESERVE)
				continue;
M
Mel Gorman 已提交
933

934 935 936 937 938 939 940 941 942
			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--;

			/*
943
			 * If breaking a large block of pages, move all free
944 945 946
			 * pages to the preferred allocation list. If falling
			 * back for a reclaimable kernel allocation, be more
			 * agressive about taking ownership of free pages
947
			 */
948
			if (unlikely(current_order >= (pageblock_order >> 1)) ||
949 950
					start_migratetype == MIGRATE_RECLAIMABLE ||
					page_group_by_mobility_disabled) {
951 952 953 954 955
				unsigned long pages;
				pages = move_freepages_block(zone, page,
								start_migratetype);

				/* Claim the whole block if over half of it is free */
956 957
				if (pages >= (1 << (pageblock_order-1)) ||
						page_group_by_mobility_disabled)
958 959 960
					set_pageblock_migratetype(page,
								start_migratetype);

961
				migratetype = start_migratetype;
962
			}
963 964 965 966 967

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

968 969 970
			/* Take ownership for orders >= pageblock_order */
			if (current_order >= pageblock_order)
				change_pageblock_range(page, current_order,
971 972 973
							start_migratetype);

			expand(zone, page, order, current_order, area, migratetype);
974 975 976 977

			trace_mm_page_alloc_extfrag(page, order, current_order,
				start_migratetype, migratetype);

978 979 980 981
			return page;
		}
	}

982
	return NULL;
983 984
}

985
/*
L
Linus Torvalds 已提交
986 987 988
 * Do the hard work of removing an element from the buddy allocator.
 * Call me with the zone->lock already held.
 */
989 990
static struct page *__rmqueue(struct zone *zone, unsigned int order,
						int migratetype)
L
Linus Torvalds 已提交
991 992 993
{
	struct page *page;

994
retry_reserve:
995
	page = __rmqueue_smallest(zone, order, migratetype);
996

997
	if (unlikely(!page) && migratetype != MIGRATE_RESERVE) {
998
		page = __rmqueue_fallback(zone, order, migratetype);
999

1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010
		/*
		 * 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;
		}
	}

1011
	trace_mm_page_alloc_zone_locked(page, order, migratetype);
1012
	return page;
L
Linus Torvalds 已提交
1013 1014 1015 1016 1017 1018 1019 1020
}

/* 
 * 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, 
1021
			unsigned long count, struct list_head *list,
1022
			int migratetype, int cold)
L
Linus Torvalds 已提交
1023 1024 1025
{
	int i;
	
N
Nick Piggin 已提交
1026
	spin_lock(&zone->lock);
L
Linus Torvalds 已提交
1027
	for (i = 0; i < count; ++i) {
1028
		struct page *page = __rmqueue(zone, order, migratetype);
N
Nick Piggin 已提交
1029
		if (unlikely(page == NULL))
L
Linus Torvalds 已提交
1030
			break;
1031 1032 1033 1034 1035 1036 1037 1038 1039 1040

		/*
		 * 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.
		 */
1041 1042 1043 1044
		if (likely(cold == 0))
			list_add(&page->lru, list);
		else
			list_add_tail(&page->lru, list);
1045
		set_page_private(page, migratetype);
1046
		list = &page->lru;
L
Linus Torvalds 已提交
1047
	}
1048
	__mod_zone_page_state(zone, NR_FREE_PAGES, -(i << order));
N
Nick Piggin 已提交
1049
	spin_unlock(&zone->lock);
N
Nick Piggin 已提交
1050
	return i;
L
Linus Torvalds 已提交
1051 1052
}

1053
#ifdef CONFIG_NUMA
1054
/*
1055 1056 1057 1058
 * Called from the vmstat counter updater to drain pagesets of this
 * currently executing processor on remote nodes after they have
 * expired.
 *
1059 1060
 * Note that this function must be called with the thread pinned to
 * a single processor.
1061
 */
1062
void drain_zone_pages(struct zone *zone, struct per_cpu_pages *pcp)
1063 1064
{
	unsigned long flags;
1065
	int to_drain;
1066

1067 1068 1069 1070 1071
	local_irq_save(flags);
	if (pcp->count >= pcp->batch)
		to_drain = pcp->batch;
	else
		to_drain = pcp->count;
1072
	free_pcppages_bulk(zone, to_drain, pcp);
1073 1074
	pcp->count -= to_drain;
	local_irq_restore(flags);
1075 1076 1077
}
#endif

1078 1079 1080 1081 1082 1083 1084 1085
/*
 * 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
1086
{
N
Nick Piggin 已提交
1087
	unsigned long flags;
L
Linus Torvalds 已提交
1088 1089
	struct zone *zone;

1090
	for_each_populated_zone(zone) {
L
Linus Torvalds 已提交
1091
		struct per_cpu_pageset *pset;
1092
		struct per_cpu_pages *pcp;
L
Linus Torvalds 已提交
1093

1094 1095
		local_irq_save(flags);
		pset = per_cpu_ptr(zone->pageset, cpu);
1096 1097

		pcp = &pset->pcp;
1098
		free_pcppages_bulk(zone, pcp->count, pcp);
1099 1100
		pcp->count = 0;
		local_irq_restore(flags);
L
Linus Torvalds 已提交
1101 1102 1103
	}
}

1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116
/*
 * 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)
{
1117
	on_each_cpu(drain_local_pages, NULL, 1);
1118 1119
}

1120
#ifdef CONFIG_HIBERNATION
L
Linus Torvalds 已提交
1121 1122 1123

void mark_free_pages(struct zone *zone)
{
1124 1125
	unsigned long pfn, max_zone_pfn;
	unsigned long flags;
1126
	int order, t;
L
Linus Torvalds 已提交
1127 1128 1129 1130 1131 1132
	struct list_head *curr;

	if (!zone->spanned_pages)
		return;

	spin_lock_irqsave(&zone->lock, flags);
1133 1134 1135 1136 1137 1138

	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);

1139 1140
			if (!swsusp_page_is_forbidden(page))
				swsusp_unset_page_free(page);
1141
		}
L
Linus Torvalds 已提交
1142

1143 1144
	for_each_migratetype_order(order, t) {
		list_for_each(curr, &zone->free_area[order].free_list[t]) {
1145
			unsigned long i;
L
Linus Torvalds 已提交
1146

1147 1148
			pfn = page_to_pfn(list_entry(curr, struct page, lru));
			for (i = 0; i < (1UL << order); i++)
1149
				swsusp_set_page_free(pfn_to_page(pfn + i));
1150
		}
1151
	}
L
Linus Torvalds 已提交
1152 1153
	spin_unlock_irqrestore(&zone->lock, flags);
}
1154
#endif /* CONFIG_PM */
L
Linus Torvalds 已提交
1155 1156 1157

/*
 * Free a 0-order page
L
Li Hong 已提交
1158
 * cold == 1 ? free a cold page : free a hot page
L
Linus Torvalds 已提交
1159
 */
L
Li Hong 已提交
1160
void free_hot_cold_page(struct page *page, int cold)
L
Linus Torvalds 已提交
1161 1162 1163 1164
{
	struct zone *zone = page_zone(page);
	struct per_cpu_pages *pcp;
	unsigned long flags;
1165
	int migratetype;
1166
	int wasMlocked = __TestClearPageMlocked(page);
L
Linus Torvalds 已提交
1167

1168
	if (!free_pages_prepare(page, 0))
1169 1170
		return;

1171 1172
	migratetype = get_pageblock_migratetype(page);
	set_page_private(page, migratetype);
L
Linus Torvalds 已提交
1173
	local_irq_save(flags);
1174
	if (unlikely(wasMlocked))
1175
		free_page_mlock(page);
1176
	__count_vm_event(PGFREE);
1177

1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192
	/*
	 * We only track unmovable, reclaimable and movable on pcp lists.
	 * Free ISOLATE pages back to the allocator because they are being
	 * offlined but treat RESERVE as movable pages so we can get those
	 * areas back if necessary. Otherwise, we may have to free
	 * excessively into the page allocator
	 */
	if (migratetype >= MIGRATE_PCPTYPES) {
		if (unlikely(migratetype == MIGRATE_ISOLATE)) {
			free_one_page(zone, page, 0, migratetype);
			goto out;
		}
		migratetype = MIGRATE_MOVABLE;
	}

1193
	pcp = &this_cpu_ptr(zone->pageset)->pcp;
1194
	if (cold)
1195
		list_add_tail(&page->lru, &pcp->lists[migratetype]);
1196
	else
1197
		list_add(&page->lru, &pcp->lists[migratetype]);
L
Linus Torvalds 已提交
1198
	pcp->count++;
N
Nick Piggin 已提交
1199
	if (pcp->count >= pcp->high) {
1200
		free_pcppages_bulk(zone, pcp->batch, pcp);
N
Nick Piggin 已提交
1201 1202
		pcp->count -= pcp->batch;
	}
1203 1204

out:
L
Linus Torvalds 已提交
1205 1206 1207
	local_irq_restore(flags);
}

N
Nick Piggin 已提交
1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219
/*
 * 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
1220 1221
	VM_BUG_ON(PageCompound(page));
	VM_BUG_ON(!page_count(page));
1222 1223 1224 1225 1226 1227 1228 1229 1230 1231

#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

1232 1233
	for (i = 1; i < (1 << order); i++)
		set_page_refcounted(page + i);
N
Nick Piggin 已提交
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 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280
/*
 * Similar to split_page except the page is already free. As this is only
 * being used for migration, the migratetype of the block also changes.
 * As this is called with interrupts disabled, the caller is responsible
 * for calling arch_alloc_page() and kernel_map_page() after interrupts
 * are enabled.
 *
 * Note: this is probably too low level an operation for use in drivers.
 * Please consult with lkml before using this in your driver.
 */
int split_free_page(struct page *page)
{
	unsigned int order;
	unsigned long watermark;
	struct zone *zone;

	BUG_ON(!PageBuddy(page));

	zone = page_zone(page);
	order = page_order(page);

	/* Obey watermarks as if the page was being allocated */
	watermark = low_wmark_pages(zone) + (1 << order);
	if (!zone_watermark_ok(zone, 0, watermark, 0, 0))
		return 0;

	/* Remove page from free list */
	list_del(&page->lru);
	zone->free_area[order].nr_free--;
	rmv_page_order(page);
	__mod_zone_page_state(zone, NR_FREE_PAGES, -(1UL << order));

	/* Split into individual pages */
	set_page_refcounted(page);
	split_page(page, order);

	if (order >= pageblock_order - 1) {
		struct page *endpage = page + (1 << order) - 1;
		for (; page < endpage; page += pageblock_nr_pages)
			set_pageblock_migratetype(page, MIGRATE_MOVABLE);
	}

	return 1 << order;
}

L
Linus Torvalds 已提交
1281 1282 1283 1284 1285
/*
 * 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.
 */
1286 1287
static inline
struct page *buffered_rmqueue(struct zone *preferred_zone,
1288 1289
			struct zone *zone, int order, gfp_t gfp_flags,
			int migratetype)
L
Linus Torvalds 已提交
1290 1291
{
	unsigned long flags;
1292
	struct page *page;
L
Linus Torvalds 已提交
1293 1294
	int cold = !!(gfp_flags & __GFP_COLD);

1295
again:
N
Nick Piggin 已提交
1296
	if (likely(order == 0)) {
L
Linus Torvalds 已提交
1297
		struct per_cpu_pages *pcp;
1298
		struct list_head *list;
L
Linus Torvalds 已提交
1299 1300

		local_irq_save(flags);
1301 1302
		pcp = &this_cpu_ptr(zone->pageset)->pcp;
		list = &pcp->lists[migratetype];
1303
		if (list_empty(list)) {
1304
			pcp->count += rmqueue_bulk(zone, 0,
1305
					pcp->batch, list,
1306
					migratetype, cold);
1307
			if (unlikely(list_empty(list)))
1308
				goto failed;
1309
		}
1310

1311 1312 1313 1314 1315
		if (cold)
			page = list_entry(list->prev, struct page, lru);
		else
			page = list_entry(list->next, struct page, lru);

1316 1317
		list_del(&page->lru);
		pcp->count--;
R
Rohit Seth 已提交
1318
	} else {
1319 1320 1321 1322 1323 1324 1325 1326
		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
1327
			 * allocate greater than order-1 page units with
1328 1329
			 * __GFP_NOFAIL.
			 */
1330
			WARN_ON_ONCE(order > 1);
1331
		}
L
Linus Torvalds 已提交
1332
		spin_lock_irqsave(&zone->lock, flags);
1333
		page = __rmqueue(zone, order, migratetype);
N
Nick Piggin 已提交
1334 1335 1336
		spin_unlock(&zone->lock);
		if (!page)
			goto failed;
1337
		__mod_zone_page_state(zone, NR_FREE_PAGES, -(1 << order));
L
Linus Torvalds 已提交
1338 1339
	}

1340
	__count_zone_vm_events(PGALLOC, zone, 1 << order);
1341
	zone_statistics(preferred_zone, zone);
N
Nick Piggin 已提交
1342
	local_irq_restore(flags);
L
Linus Torvalds 已提交
1343

N
Nick Piggin 已提交
1344
	VM_BUG_ON(bad_range(zone, page));
N
Nick Piggin 已提交
1345
	if (prep_new_page(page, order, gfp_flags))
N
Nick Piggin 已提交
1346
		goto again;
L
Linus Torvalds 已提交
1347
	return page;
N
Nick Piggin 已提交
1348 1349 1350 1351

failed:
	local_irq_restore(flags);
	return NULL;
L
Linus Torvalds 已提交
1352 1353
}

1354 1355 1356 1357 1358 1359 1360 1361 1362
/* 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)

1363 1364 1365
#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
1366

1367 1368 1369 1370 1371 1372 1373
#ifdef CONFIG_FAIL_PAGE_ALLOC

static struct fail_page_alloc_attr {
	struct fault_attr attr;

	u32 ignore_gfp_highmem;
	u32 ignore_gfp_wait;
1374
	u32 min_order;
1375 1376 1377 1378 1379

#ifdef CONFIG_FAULT_INJECTION_DEBUG_FS

	struct dentry *ignore_gfp_highmem_file;
	struct dentry *ignore_gfp_wait_file;
1380
	struct dentry *min_order_file;
1381 1382 1383 1384 1385

#endif /* CONFIG_FAULT_INJECTION_DEBUG_FS */

} fail_page_alloc = {
	.attr = FAULT_ATTR_INITIALIZER,
1386 1387
	.ignore_gfp_wait = 1,
	.ignore_gfp_highmem = 1,
1388
	.min_order = 1,
1389 1390 1391 1392 1393 1394 1395 1396 1397 1398
};

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)
{
1399 1400
	if (order < fail_page_alloc.min_order)
		return 0;
1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431
	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);
1432 1433 1434
	fail_page_alloc.min_order_file =
		debugfs_create_u32("min-order", mode, dir,
				   &fail_page_alloc.min_order);
1435 1436

	if (!fail_page_alloc.ignore_gfp_wait_file ||
1437 1438
            !fail_page_alloc.ignore_gfp_highmem_file ||
            !fail_page_alloc.min_order_file) {
1439 1440 1441
		err = -ENOMEM;
		debugfs_remove(fail_page_alloc.ignore_gfp_wait_file);
		debugfs_remove(fail_page_alloc.ignore_gfp_highmem_file);
1442
		debugfs_remove(fail_page_alloc.min_order_file);
1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461
		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
1462
/*
1463
 * Return true if free pages are above 'mark'. This takes into account the order
L
Linus Torvalds 已提交
1464 1465
 * of the allocation.
 */
1466 1467
static bool __zone_watermark_ok(struct zone *z, int order, unsigned long mark,
		      int classzone_idx, int alloc_flags, long free_pages)
L
Linus Torvalds 已提交
1468 1469
{
	/* free_pages my go negative - that's OK */
1470
	long min = mark;
L
Linus Torvalds 已提交
1471 1472
	int o;

1473
	free_pages -= (1 << order) + 1;
R
Rohit Seth 已提交
1474
	if (alloc_flags & ALLOC_HIGH)
L
Linus Torvalds 已提交
1475
		min -= min / 2;
R
Rohit Seth 已提交
1476
	if (alloc_flags & ALLOC_HARDER)
L
Linus Torvalds 已提交
1477 1478 1479
		min -= min / 4;

	if (free_pages <= min + z->lowmem_reserve[classzone_idx])
1480
		return false;
L
Linus Torvalds 已提交
1481 1482 1483 1484 1485 1486 1487 1488
	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)
1489
			return false;
L
Linus Torvalds 已提交
1490
	}
1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510
	return true;
}

bool zone_watermark_ok(struct zone *z, int order, unsigned long mark,
		      int classzone_idx, int alloc_flags)
{
	return __zone_watermark_ok(z, order, mark, classzone_idx, alloc_flags,
					zone_page_state(z, NR_FREE_PAGES));
}

bool zone_watermark_ok_safe(struct zone *z, int order, unsigned long mark,
		      int classzone_idx, int alloc_flags)
{
	long free_pages = zone_page_state(z, NR_FREE_PAGES);

	if (z->percpu_drift_mark && free_pages < z->percpu_drift_mark)
		free_pages = zone_page_state_snapshot(z, NR_FREE_PAGES);

	return __zone_watermark_ok(z, order, mark, classzone_idx, alloc_flags,
								free_pages);
L
Linus Torvalds 已提交
1511 1512
}

1513 1514 1515 1516 1517 1518
#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
1519
 * that have to skip over a lot of full or unallowed zones.
1520 1521 1522
 *
 * If the zonelist cache is present in the passed in zonelist, then
 * returns a pointer to the allowed node mask (either the current
1523
 * tasks mems_allowed, or node_states[N_HIGH_MEMORY].)
1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544
 *
 * 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
1545
	if (time_after(jiffies, zlc->last_full_zap + HZ)) {
1546 1547 1548 1549 1550 1551
		bitmap_zero(zlc->fullzones, MAX_ZONES_PER_ZONELIST);
		zlc->last_full_zap = jiffies;
	}

	allowednodes = !in_interrupt() && (alloc_flags & ALLOC_CPUSET) ?
					&cpuset_current_mems_allowed :
1552
					&node_states[N_HIGH_MEMORY];
1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577
	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.
 */
1578
static int zlc_zone_worth_trying(struct zonelist *zonelist, struct zoneref *z,
1579 1580 1581 1582 1583 1584 1585 1586 1587 1588
						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;

1589
	i = z - zonelist->_zonerefs;
1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600
	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.
 */
1601
static void zlc_mark_zone_full(struct zonelist *zonelist, struct zoneref *z)
1602 1603 1604 1605 1606 1607 1608 1609
{
	struct zonelist_cache *zlc;	/* cached zonelist speedup info */
	int i;				/* index of *z in zonelist zones */

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

1610
	i = z - zonelist->_zonerefs;
1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621

	set_bit(i, zlc->fullzones);
}

#else	/* CONFIG_NUMA */

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

1622
static int zlc_zone_worth_trying(struct zonelist *zonelist, struct zoneref *z,
1623 1624 1625 1626 1627
				nodemask_t *allowednodes)
{
	return 1;
}

1628
static void zlc_mark_zone_full(struct zonelist *zonelist, struct zoneref *z)
1629 1630 1631 1632
{
}
#endif	/* CONFIG_NUMA */

R
Rohit Seth 已提交
1633
/*
1634
 * get_page_from_freelist goes through the zonelist trying to allocate
R
Rohit Seth 已提交
1635 1636 1637
 * a page.
 */
static struct page *
1638
get_page_from_freelist(gfp_t gfp_mask, nodemask_t *nodemask, unsigned int order,
1639
		struct zonelist *zonelist, int high_zoneidx, int alloc_flags,
1640
		struct zone *preferred_zone, int migratetype)
M
Martin Hicks 已提交
1641
{
1642
	struct zoneref *z;
R
Rohit Seth 已提交
1643
	struct page *page = NULL;
1644
	int classzone_idx;
1645
	struct zone *zone;
1646 1647 1648
	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 */
1649

1650
	classzone_idx = zone_idx(preferred_zone);
1651
zonelist_scan:
R
Rohit Seth 已提交
1652
	/*
1653
	 * Scan zonelist, looking for a zone with enough free.
R
Rohit Seth 已提交
1654 1655
	 * See also cpuset_zone_allowed() comment in kernel/cpuset.c.
	 */
1656 1657
	for_each_zone_zonelist_nodemask(zone, z, zonelist,
						high_zoneidx, nodemask) {
1658 1659 1660
		if (NUMA_BUILD && zlc_active &&
			!zlc_zone_worth_trying(zonelist, z, allowednodes))
				continue;
R
Rohit Seth 已提交
1661
		if ((alloc_flags & ALLOC_CPUSET) &&
1662
			!cpuset_zone_allowed_softwall(zone, gfp_mask))
1663
				goto try_next_zone;
R
Rohit Seth 已提交
1664

1665
		BUILD_BUG_ON(ALLOC_NO_WATERMARKS < NR_WMARK);
R
Rohit Seth 已提交
1666
		if (!(alloc_flags & ALLOC_NO_WATERMARKS)) {
1667
			unsigned long mark;
1668 1669
			int ret;

1670
			mark = zone->watermark[alloc_flags & ALLOC_WMARK_MASK];
1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689
			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))
1690
					goto this_zone_full;
1691
			}
R
Rohit Seth 已提交
1692 1693
		}

1694
try_this_zone:
1695 1696
		page = buffered_rmqueue(preferred_zone, zone, order,
						gfp_mask, migratetype);
1697
		if (page)
R
Rohit Seth 已提交
1698
			break;
1699 1700 1701 1702
this_zone_full:
		if (NUMA_BUILD)
			zlc_mark_zone_full(zonelist, z);
try_next_zone:
1703
		if (NUMA_BUILD && !did_zlc_setup && nr_online_nodes > 1) {
1704 1705 1706 1707
			/*
			 * we do zlc_setup after the first zone is tried but only
			 * if there are multiple nodes make it worthwhile
			 */
1708 1709 1710 1711
			allowednodes = zlc_setup(zonelist, alloc_flags);
			zlc_active = 1;
			did_zlc_setup = 1;
		}
1712
	}
1713 1714 1715 1716 1717 1718

	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
1719
	return page;
M
Martin Hicks 已提交
1720 1721
}

1722 1723 1724
static inline int
should_alloc_retry(gfp_t gfp_mask, unsigned int order,
				unsigned long pages_reclaimed)
L
Linus Torvalds 已提交
1725
{
1726 1727 1728
	/* Do not loop if specifically requested */
	if (gfp_mask & __GFP_NORETRY)
		return 0;
L
Linus Torvalds 已提交
1729

1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746
	/*
	 * 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;
1747

1748 1749 1750 1751 1752 1753
	/*
	 * 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
1754

1755 1756
	return 0;
}
1757

1758 1759 1760
static inline struct page *
__alloc_pages_may_oom(gfp_t gfp_mask, unsigned int order,
	struct zonelist *zonelist, enum zone_type high_zoneidx,
1761 1762
	nodemask_t *nodemask, struct zone *preferred_zone,
	int migratetype)
1763 1764 1765 1766
{
	struct page *page;

	/* Acquire the OOM killer lock for the zones in zonelist */
1767
	if (!try_set_zonelist_oom(zonelist, gfp_mask)) {
1768
		schedule_timeout_uninterruptible(1);
L
Linus Torvalds 已提交
1769 1770
		return NULL;
	}
1771

1772 1773 1774 1775 1776 1777 1778
	/*
	 * 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,
1779
		ALLOC_WMARK_HIGH|ALLOC_CPUSET,
1780
		preferred_zone, migratetype);
R
Rohit Seth 已提交
1781
	if (page)
1782 1783
		goto out;

1784 1785 1786 1787
	if (!(gfp_mask & __GFP_NOFAIL)) {
		/* The OOM killer will not help higher order allocs */
		if (order > PAGE_ALLOC_COSTLY_ORDER)
			goto out;
1788 1789 1790
		/* The OOM killer does not needlessly kill tasks for lowmem */
		if (high_zoneidx < ZONE_NORMAL)
			goto out;
1791 1792 1793 1794 1795 1796 1797 1798 1799 1800
		/*
		 * GFP_THISNODE contains __GFP_NORETRY and we never hit this.
		 * Sanity check for bare calls of __GFP_THISNODE, not real OOM.
		 * The caller should handle page allocation failure by itself if
		 * it specifies __GFP_THISNODE.
		 * Note: Hugepage uses it but will hit PAGE_ALLOC_COSTLY_ORDER.
		 */
		if (gfp_mask & __GFP_THISNODE)
			goto out;
	}
1801
	/* Exhausted what can be done so it's blamo time */
1802
	out_of_memory(zonelist, gfp_mask, order, nodemask);
1803 1804 1805 1806 1807 1808

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

1809 1810 1811 1812 1813 1814 1815 1816 1817
#ifdef CONFIG_COMPACTION
/* Try memory compaction for high-order allocations before reclaim */
static struct page *
__alloc_pages_direct_compact(gfp_t gfp_mask, unsigned int order,
	struct zonelist *zonelist, enum zone_type high_zoneidx,
	nodemask_t *nodemask, int alloc_flags, struct zone *preferred_zone,
	int migratetype, unsigned long *did_some_progress)
{
	struct page *page;
1818
	struct task_struct *tsk = current;
1819

1820
	if (!order || compaction_deferred(preferred_zone))
1821 1822
		return NULL;

1823
	tsk->flags |= PF_MEMALLOC;
1824 1825
	*did_some_progress = try_to_compact_pages(zonelist, order, gfp_mask,
								nodemask);
1826
	tsk->flags &= ~PF_MEMALLOC;
1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837
	if (*did_some_progress != COMPACT_SKIPPED) {

		/* Page migration frees to the PCP lists but we want merging */
		drain_pages(get_cpu());
		put_cpu();

		page = get_page_from_freelist(gfp_mask, nodemask,
				order, zonelist, high_zoneidx,
				alloc_flags, preferred_zone,
				migratetype);
		if (page) {
1838 1839
			preferred_zone->compact_considered = 0;
			preferred_zone->compact_defer_shift = 0;
1840 1841 1842 1843 1844 1845 1846 1847 1848 1849
			count_vm_event(COMPACTSUCCESS);
			return page;
		}

		/*
		 * It's bad if compaction run occurs and fails.
		 * The most likely reason is that pages exist,
		 * but not enough to satisfy watermarks.
		 */
		count_vm_event(COMPACTFAIL);
1850
		defer_compaction(preferred_zone);
1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867

		cond_resched();
	}

	return NULL;
}
#else
static inline struct page *
__alloc_pages_direct_compact(gfp_t gfp_mask, unsigned int order,
	struct zonelist *zonelist, enum zone_type high_zoneidx,
	nodemask_t *nodemask, int alloc_flags, struct zone *preferred_zone,
	int migratetype, unsigned long *did_some_progress)
{
	return NULL;
}
#endif /* CONFIG_COMPACTION */

1868 1869 1870 1871
/* 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,
1872
	nodemask_t *nodemask, int alloc_flags, struct zone *preferred_zone,
1873
	int migratetype, unsigned long *did_some_progress)
1874 1875 1876 1877
{
	struct page *page = NULL;
	struct reclaim_state reclaim_state;
	struct task_struct *p = current;
1878
	bool drained = false;
1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896

	cond_resched();

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

1897 1898
	if (unlikely(!(*did_some_progress)))
		return NULL;
1899

1900 1901
retry:
	page = get_page_from_freelist(gfp_mask, nodemask, order,
1902
					zonelist, high_zoneidx,
1903 1904
					alloc_flags, preferred_zone,
					migratetype);
1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915

	/*
	 * If an allocation failed after direct reclaim, it could be because
	 * pages are pinned on the per-cpu lists. Drain them and try again
	 */
	if (!page && !drained) {
		drain_all_pages();
		drained = true;
		goto retry;
	}

1916 1917 1918
	return page;
}

L
Linus Torvalds 已提交
1919
/*
1920 1921
 * 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
1922
 */
1923 1924 1925
static inline struct page *
__alloc_pages_high_priority(gfp_t gfp_mask, unsigned int order,
	struct zonelist *zonelist, enum zone_type high_zoneidx,
1926 1927
	nodemask_t *nodemask, struct zone *preferred_zone,
	int migratetype)
1928 1929 1930 1931 1932
{
	struct page *page;

	do {
		page = get_page_from_freelist(gfp_mask, nodemask, order,
1933
			zonelist, high_zoneidx, ALLOC_NO_WATERMARKS,
1934
			preferred_zone, migratetype);
1935 1936

		if (!page && gfp_mask & __GFP_NOFAIL)
1937
			wait_iff_congested(preferred_zone, BLK_RW_ASYNC, HZ/50);
1938 1939 1940 1941 1942 1943 1944 1945
	} 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
1946
{
1947 1948
	struct zoneref *z;
	struct zone *zone;
L
Linus Torvalds 已提交
1949

1950 1951 1952
	for_each_zone_zonelist(zone, z, zonelist, high_zoneidx)
		wakeup_kswapd(zone, order);
}
1953

1954 1955 1956 1957 1958 1959
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
1960

1961
	/* __GFP_HIGH is assumed to be the same as ALLOC_HIGH to save a branch. */
1962
	BUILD_BUG_ON(__GFP_HIGH != (__force gfp_t) ALLOC_HIGH);
1963

1964 1965 1966 1967 1968 1969
	/*
	 * 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).
	 */
1970
	alloc_flags |= (__force int) (gfp_mask & __GFP_HIGH);
L
Linus Torvalds 已提交
1971

1972 1973
	if (!wait) {
		alloc_flags |= ALLOC_HARDER;
1974
		/*
1975 1976
		 * Ignore cpuset if GFP_ATOMIC (!wait) rather than fail alloc.
		 * See also cpuset_zone_allowed() comment in kernel/cpuset.c.
1977
		 */
1978
		alloc_flags &= ~ALLOC_CPUSET;
1979
	} else if (unlikely(rt_task(p)) && !in_interrupt())
1980 1981 1982 1983 1984 1985 1986
		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
1987
	}
1988

1989 1990 1991
	return alloc_flags;
}

1992 1993 1994
static inline struct page *
__alloc_pages_slowpath(gfp_t gfp_mask, unsigned int order,
	struct zonelist *zonelist, enum zone_type high_zoneidx,
1995 1996
	nodemask_t *nodemask, struct zone *preferred_zone,
	int migratetype)
1997 1998 1999 2000 2001 2002 2003
{
	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
2004

2005 2006 2007 2008 2009 2010
	/*
	 * 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.
	 */
2011 2012
	if (order >= MAX_ORDER) {
		WARN_ON_ONCE(!(gfp_mask & __GFP_NOWARN));
2013
		return NULL;
2014
	}
L
Linus Torvalds 已提交
2015

2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026
	/*
	 * 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;

2027
restart:
2028
	wake_all_kswapd(order, zonelist, high_zoneidx);
L
Linus Torvalds 已提交
2029

2030
	/*
R
Rohit Seth 已提交
2031 2032 2033
	 * 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.
2034
	 */
2035
	alloc_flags = gfp_to_alloc_flags(gfp_mask);
L
Linus Torvalds 已提交
2036

2037
	/* This is the last chance, in general, before the goto nopage. */
2038
	page = get_page_from_freelist(gfp_mask, nodemask, order, zonelist,
2039 2040
			high_zoneidx, alloc_flags & ~ALLOC_NO_WATERMARKS,
			preferred_zone, migratetype);
R
Rohit Seth 已提交
2041 2042
	if (page)
		goto got_pg;
L
Linus Torvalds 已提交
2043

2044
rebalance:
2045
	/* Allocate without watermarks if the context allows */
2046 2047 2048 2049 2050 2051
	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
2052 2053 2054 2055 2056 2057
	}

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

2058 2059 2060 2061
	/* Avoid recursion of direct reclaim */
	if (p->flags & PF_MEMALLOC)
		goto nopage;

2062 2063 2064 2065
	/* Avoid allocations with no watermarks from looping endlessly */
	if (test_thread_flag(TIF_MEMDIE) && !(gfp_mask & __GFP_NOFAIL))
		goto nopage;

2066 2067 2068 2069 2070 2071 2072 2073 2074
	/* Try direct compaction */
	page = __alloc_pages_direct_compact(gfp_mask, order,
					zonelist, high_zoneidx,
					nodemask,
					alloc_flags, preferred_zone,
					migratetype, &did_some_progress);
	if (page)
		goto got_pg;

2075 2076 2077 2078
	/* Try direct reclaim and then allocating */
	page = __alloc_pages_direct_reclaim(gfp_mask, order,
					zonelist, high_zoneidx,
					nodemask,
2079
					alloc_flags, preferred_zone,
2080
					migratetype, &did_some_progress);
2081 2082
	if (page)
		goto got_pg;
L
Linus Torvalds 已提交
2083

2084
	/*
2085 2086
	 * If we failed to make any progress reclaiming, then we are
	 * running out of options and have to consider going OOM
2087
	 */
2088 2089
	if (!did_some_progress) {
		if ((gfp_mask & __GFP_FS) && !(gfp_mask & __GFP_NORETRY)) {
2090 2091
			if (oom_killer_disabled)
				goto nopage;
2092 2093
			page = __alloc_pages_may_oom(gfp_mask, order,
					zonelist, high_zoneidx,
2094 2095
					nodemask, preferred_zone,
					migratetype);
2096 2097
			if (page)
				goto got_pg;
L
Linus Torvalds 已提交
2098

2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115
			if (!(gfp_mask & __GFP_NOFAIL)) {
				/*
				 * The oom killer is not called for high-order
				 * allocations that may fail, so if no progress
				 * is being made, there are no other options and
				 * retrying is unlikely to help.
				 */
				if (order > PAGE_ALLOC_COSTLY_ORDER)
					goto nopage;
				/*
				 * The oom killer is not called for lowmem
				 * allocations to prevent needlessly killing
				 * innocent tasks.
				 */
				if (high_zoneidx < ZONE_NORMAL)
					goto nopage;
			}
2116

2117 2118
			goto restart;
		}
L
Linus Torvalds 已提交
2119 2120
	}

2121
	/* Check if we should retry the allocation */
2122
	pages_reclaimed += did_some_progress;
2123 2124
	if (should_alloc_retry(gfp_mask, order, pages_reclaimed)) {
		/* Wait for some write requests to complete then retry */
2125
		wait_iff_congested(preferred_zone, BLK_RW_ASYNC, HZ/50);
L
Linus Torvalds 已提交
2126
		goto rebalance;
2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139
	} else {
		/*
		 * High-order allocations do not necessarily loop after
		 * direct reclaim and reclaim/compaction depends on compaction
		 * being called after reclaim so call directly if necessary
		 */
		page = __alloc_pages_direct_compact(gfp_mask, order,
					zonelist, high_zoneidx,
					nodemask,
					alloc_flags, preferred_zone,
					migratetype, &did_some_progress);
		if (page)
			goto got_pg;
L
Linus Torvalds 已提交
2140 2141 2142 2143 2144 2145 2146 2147
	}

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
2148
		show_mem();
L
Linus Torvalds 已提交
2149
	}
2150
	return page;
L
Linus Torvalds 已提交
2151
got_pg:
2152 2153
	if (kmemcheck_enabled)
		kmemcheck_pagealloc_alloc(page, order, gfp_mask);
L
Linus Torvalds 已提交
2154
	return page;
2155

L
Linus Torvalds 已提交
2156
}
2157 2158 2159 2160 2161 2162 2163 2164 2165

/*
 * 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);
2166
	struct zone *preferred_zone;
2167
	struct page *page;
2168
	int migratetype = allocflags_to_migratetype(gfp_mask);
2169

2170 2171
	gfp_mask &= gfp_allowed_mask;

2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186
	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;

2187
	get_mems_allowed();
2188 2189
	/* The preferred zone is used for statistics later */
	first_zones_zonelist(zonelist, high_zoneidx, nodemask, &preferred_zone);
2190 2191
	if (!preferred_zone) {
		put_mems_allowed();
2192
		return NULL;
2193
	}
2194 2195

	/* First allocation attempt */
2196
	page = get_page_from_freelist(gfp_mask|__GFP_HARDWALL, nodemask, order,
2197
			zonelist, high_zoneidx, ALLOC_WMARK_LOW|ALLOC_CPUSET,
2198
			preferred_zone, migratetype);
2199 2200
	if (unlikely(!page))
		page = __alloc_pages_slowpath(gfp_mask, order,
2201
				zonelist, high_zoneidx, nodemask,
2202
				preferred_zone, migratetype);
2203
	put_mems_allowed();
2204

2205
	trace_mm_page_alloc(page, order, gfp_mask, migratetype);
2206
	return page;
L
Linus Torvalds 已提交
2207
}
2208
EXPORT_SYMBOL(__alloc_pages_nodemask);
L
Linus Torvalds 已提交
2209 2210 2211 2212

/*
 * Common helper functions.
 */
H
Harvey Harrison 已提交
2213
unsigned long __get_free_pages(gfp_t gfp_mask, unsigned int order)
L
Linus Torvalds 已提交
2214
{
2215 2216 2217 2218 2219 2220 2221 2222
	struct page *page;

	/*
	 * __get_free_pages() returns a 32-bit address, which cannot represent
	 * a highmem page
	 */
	VM_BUG_ON((gfp_mask & __GFP_HIGHMEM) != 0);

L
Linus Torvalds 已提交
2223 2224 2225 2226 2227 2228 2229
	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
2230
unsigned long get_zeroed_page(gfp_t gfp_mask)
L
Linus Torvalds 已提交
2231
{
2232
	return __get_free_pages(gfp_mask | __GFP_ZERO, 0);
L
Linus Torvalds 已提交
2233 2234 2235 2236 2237 2238 2239
}
EXPORT_SYMBOL(get_zeroed_page);

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

2240 2241
	while (--i >= 0) {
		trace_mm_pagevec_free(pvec->pages[i], pvec->cold);
L
Linus Torvalds 已提交
2242
		free_hot_cold_page(pvec->pages[i], pvec->cold);
2243
	}
L
Linus Torvalds 已提交
2244 2245
}

H
Harvey Harrison 已提交
2246
void __free_pages(struct page *page, unsigned int order)
L
Linus Torvalds 已提交
2247
{
N
Nick Piggin 已提交
2248
	if (put_page_testzero(page)) {
L
Linus Torvalds 已提交
2249
		if (order == 0)
L
Li Hong 已提交
2250
			free_hot_cold_page(page, 0);
L
Linus Torvalds 已提交
2251 2252 2253 2254 2255 2256 2257
		else
			__free_pages_ok(page, order);
	}
}

EXPORT_SYMBOL(__free_pages);

H
Harvey Harrison 已提交
2258
void free_pages(unsigned long addr, unsigned int order)
L
Linus Torvalds 已提交
2259 2260
{
	if (addr != 0) {
N
Nick Piggin 已提交
2261
		VM_BUG_ON(!virt_addr_valid((void *)addr));
L
Linus Torvalds 已提交
2262 2263 2264 2265 2266 2267
		__free_pages(virt_to_page((void *)addr), order);
	}
}

EXPORT_SYMBOL(free_pages);

2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290
/**
 * 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);

K
Kevin Cernekee 已提交
2291
		split_page(virt_to_page((void *)addr), order);
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
		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
2321 2322
static unsigned int nr_free_zone_pages(int offset)
{
2323
	struct zoneref *z;
2324 2325
	struct zone *zone;

2326
	/* Just pick one node, since fallback list is circular */
L
Linus Torvalds 已提交
2327 2328
	unsigned int sum = 0;

2329
	struct zonelist *zonelist = node_zonelist(numa_node_id(), GFP_KERNEL);
L
Linus Torvalds 已提交
2330

2331
	for_each_zone_zonelist(zone, z, zonelist, offset) {
2332
		unsigned long size = zone->present_pages;
2333
		unsigned long high = high_wmark_pages(zone);
2334 2335
		if (size > high)
			sum += size - high;
L
Linus Torvalds 已提交
2336 2337 2338 2339 2340 2341 2342 2343 2344 2345
	}

	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
2346
	return nr_free_zone_pages(gfp_zone(GFP_USER));
L
Linus Torvalds 已提交
2347
}
2348
EXPORT_SYMBOL_GPL(nr_free_buffer_pages);
L
Linus Torvalds 已提交
2349 2350 2351 2352 2353 2354

/*
 * Amount of free RAM allocatable within all zones
 */
unsigned int nr_free_pagecache_pages(void)
{
M
Mel Gorman 已提交
2355
	return nr_free_zone_pages(gfp_zone(GFP_HIGHUSER_MOVABLE));
L
Linus Torvalds 已提交
2356
}
2357 2358

static inline void show_node(struct zone *zone)
L
Linus Torvalds 已提交
2359
{
2360
	if (NUMA_BUILD)
2361
		printk("Node %d ", zone_to_nid(zone));
L
Linus Torvalds 已提交
2362 2363 2364 2365 2366 2367
}

void si_meminfo(struct sysinfo *val)
{
	val->totalram = totalram_pages;
	val->sharedram = 0;
2368
	val->freeram = global_page_state(NR_FREE_PAGES);
L
Linus Torvalds 已提交
2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382
	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;
2383
	val->freeram = node_page_state(nid, NR_FREE_PAGES);
2384
#ifdef CONFIG_HIGHMEM
L
Linus Torvalds 已提交
2385
	val->totalhigh = pgdat->node_zones[ZONE_HIGHMEM].present_pages;
2386 2387
	val->freehigh = zone_page_state(&pgdat->node_zones[ZONE_HIGHMEM],
			NR_FREE_PAGES);
2388 2389 2390 2391
#else
	val->totalhigh = 0;
	val->freehigh = 0;
#endif
L
Linus Torvalds 已提交
2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404
	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)
{
2405
	int cpu;
L
Linus Torvalds 已提交
2406 2407
	struct zone *zone;

2408
	for_each_populated_zone(zone) {
2409 2410
		show_node(zone);
		printk("%s per-cpu:\n", zone->name);
L
Linus Torvalds 已提交
2411

2412
		for_each_online_cpu(cpu) {
L
Linus Torvalds 已提交
2413 2414
			struct per_cpu_pageset *pageset;

2415
			pageset = per_cpu_ptr(zone->pageset, cpu);
L
Linus Torvalds 已提交
2416

2417 2418 2419
			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
2420 2421 2422
		}
	}

K
KOSAKI Motohiro 已提交
2423 2424
	printk("active_anon:%lu inactive_anon:%lu isolated_anon:%lu\n"
		" active_file:%lu inactive_file:%lu isolated_file:%lu\n"
L
Lee Schermerhorn 已提交
2425
		" unevictable:%lu"
2426
		" dirty:%lu writeback:%lu unstable:%lu\n"
2427
		" free:%lu slab_reclaimable:%lu slab_unreclaimable:%lu\n"
2428
		" mapped:%lu shmem:%lu pagetables:%lu bounce:%lu\n",
2429 2430
		global_page_state(NR_ACTIVE_ANON),
		global_page_state(NR_INACTIVE_ANON),
K
KOSAKI Motohiro 已提交
2431 2432
		global_page_state(NR_ISOLATED_ANON),
		global_page_state(NR_ACTIVE_FILE),
2433
		global_page_state(NR_INACTIVE_FILE),
K
KOSAKI Motohiro 已提交
2434
		global_page_state(NR_ISOLATED_FILE),
L
Lee Schermerhorn 已提交
2435
		global_page_state(NR_UNEVICTABLE),
2436
		global_page_state(NR_FILE_DIRTY),
2437
		global_page_state(NR_WRITEBACK),
2438
		global_page_state(NR_UNSTABLE_NFS),
2439
		global_page_state(NR_FREE_PAGES),
2440 2441
		global_page_state(NR_SLAB_RECLAIMABLE),
		global_page_state(NR_SLAB_UNRECLAIMABLE),
2442
		global_page_state(NR_FILE_MAPPED),
2443
		global_page_state(NR_SHMEM),
2444 2445
		global_page_state(NR_PAGETABLE),
		global_page_state(NR_BOUNCE));
L
Linus Torvalds 已提交
2446

2447
	for_each_populated_zone(zone) {
L
Linus Torvalds 已提交
2448 2449 2450 2451 2452 2453 2454 2455
		int i;

		show_node(zone);
		printk("%s"
			" free:%lukB"
			" min:%lukB"
			" low:%lukB"
			" high:%lukB"
2456 2457 2458 2459
			" active_anon:%lukB"
			" inactive_anon:%lukB"
			" active_file:%lukB"
			" inactive_file:%lukB"
L
Lee Schermerhorn 已提交
2460
			" unevictable:%lukB"
K
KOSAKI Motohiro 已提交
2461 2462
			" isolated(anon):%lukB"
			" isolated(file):%lukB"
L
Linus Torvalds 已提交
2463
			" present:%lukB"
2464 2465 2466 2467
			" mlocked:%lukB"
			" dirty:%lukB"
			" writeback:%lukB"
			" mapped:%lukB"
2468
			" shmem:%lukB"
2469 2470
			" slab_reclaimable:%lukB"
			" slab_unreclaimable:%lukB"
2471
			" kernel_stack:%lukB"
2472 2473 2474 2475
			" pagetables:%lukB"
			" unstable:%lukB"
			" bounce:%lukB"
			" writeback_tmp:%lukB"
L
Linus Torvalds 已提交
2476 2477 2478 2479
			" pages_scanned:%lu"
			" all_unreclaimable? %s"
			"\n",
			zone->name,
2480
			K(zone_page_state(zone, NR_FREE_PAGES)),
2481 2482 2483
			K(min_wmark_pages(zone)),
			K(low_wmark_pages(zone)),
			K(high_wmark_pages(zone)),
2484 2485 2486 2487
			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
2488
			K(zone_page_state(zone, NR_UNEVICTABLE)),
K
KOSAKI Motohiro 已提交
2489 2490
			K(zone_page_state(zone, NR_ISOLATED_ANON)),
			K(zone_page_state(zone, NR_ISOLATED_FILE)),
L
Linus Torvalds 已提交
2491
			K(zone->present_pages),
2492 2493 2494 2495
			K(zone_page_state(zone, NR_MLOCK)),
			K(zone_page_state(zone, NR_FILE_DIRTY)),
			K(zone_page_state(zone, NR_WRITEBACK)),
			K(zone_page_state(zone, NR_FILE_MAPPED)),
2496
			K(zone_page_state(zone, NR_SHMEM)),
2497 2498
			K(zone_page_state(zone, NR_SLAB_RECLAIMABLE)),
			K(zone_page_state(zone, NR_SLAB_UNRECLAIMABLE)),
2499 2500
			zone_page_state(zone, NR_KERNEL_STACK) *
				THREAD_SIZE / 1024,
2501 2502 2503 2504
			K(zone_page_state(zone, NR_PAGETABLE)),
			K(zone_page_state(zone, NR_UNSTABLE_NFS)),
			K(zone_page_state(zone, NR_BOUNCE)),
			K(zone_page_state(zone, NR_WRITEBACK_TEMP)),
L
Linus Torvalds 已提交
2505
			zone->pages_scanned,
2506
			(zone->all_unreclaimable ? "yes" : "no")
L
Linus Torvalds 已提交
2507 2508 2509 2510 2511 2512 2513
			);
		printk("lowmem_reserve[]:");
		for (i = 0; i < MAX_NR_ZONES; i++)
			printk(" %lu", zone->lowmem_reserve[i]);
		printk("\n");
	}

2514
	for_each_populated_zone(zone) {
2515
 		unsigned long nr[MAX_ORDER], flags, order, total = 0;
L
Linus Torvalds 已提交
2516 2517 2518 2519 2520 2521

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

		spin_lock_irqsave(&zone->lock, flags);
		for (order = 0; order < MAX_ORDER; order++) {
2522 2523
			nr[order] = zone->free_area[order].nr_free;
			total += nr[order] << order;
L
Linus Torvalds 已提交
2524 2525
		}
		spin_unlock_irqrestore(&zone->lock, flags);
2526 2527
		for (order = 0; order < MAX_ORDER; order++)
			printk("%lu*%lukB ", nr[order], K(1UL) << order);
L
Linus Torvalds 已提交
2528 2529 2530
		printk("= %lukB\n", K(total));
	}

2531 2532
	printk("%ld total pagecache pages\n", global_page_state(NR_FILE_PAGES));

L
Linus Torvalds 已提交
2533 2534 2535
	show_swap_cache_info();
}

2536 2537 2538 2539 2540 2541
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
2542 2543
/*
 * Builds allocation fallback zone lists.
2544 2545
 *
 * Add all populated zones of a node to the zonelist.
L
Linus Torvalds 已提交
2546
 */
2547 2548
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
2549
{
2550 2551
	struct zone *zone;

2552
	BUG_ON(zone_type >= MAX_NR_ZONES);
2553
	zone_type++;
2554 2555

	do {
2556
		zone_type--;
2557
		zone = pgdat->node_zones + zone_type;
2558
		if (populated_zone(zone)) {
2559 2560
			zoneref_set_zone(zone,
				&zonelist->_zonerefs[nr_zones++]);
2561
			check_highest_zone(zone_type);
L
Linus Torvalds 已提交
2562
		}
2563

2564
	} while (zone_type);
2565
	return nr_zones;
L
Linus Torvalds 已提交
2566 2567
}

2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588

/*
 *  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
2589
#ifdef CONFIG_NUMA
2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632
/* 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,
2633
		void __user *buffer, size_t *length,
2634 2635 2636 2637
		loff_t *ppos)
{
	char saved_string[NUMA_ZONELIST_ORDER_LEN];
	int ret;
2638
	static DEFINE_MUTEX(zl_order_mutex);
2639

2640
	mutex_lock(&zl_order_mutex);
2641
	if (write)
2642
		strcpy(saved_string, (char*)table->data);
2643
	ret = proc_dostring(table, write, buffer, length, ppos);
2644
	if (ret)
2645
		goto out;
2646 2647 2648 2649 2650 2651 2652 2653 2654
	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;
2655 2656
		} else if (oldval != user_zonelist_order) {
			mutex_lock(&zonelists_mutex);
2657
			build_all_zonelists(NULL);
2658 2659
			mutex_unlock(&zonelists_mutex);
		}
2660
	}
2661 2662 2663
out:
	mutex_unlock(&zl_order_mutex);
	return ret;
2664 2665 2666
}


2667
#define MAX_NODE_LOAD (nr_online_nodes)
2668 2669
static int node_load[MAX_NUMNODES];

L
Linus Torvalds 已提交
2670
/**
2671
 * find_next_best_node - find the next node that should appear in a given node's fallback list
L
Linus Torvalds 已提交
2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683
 * @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.
 */
2684
static int find_next_best_node(int node, nodemask_t *used_node_mask)
L
Linus Torvalds 已提交
2685
{
2686
	int n, val;
L
Linus Torvalds 已提交
2687 2688
	int min_val = INT_MAX;
	int best_node = -1;
2689
	const struct cpumask *tmp = cpumask_of_node(0);
L
Linus Torvalds 已提交
2690

2691 2692 2693 2694 2695
	/* 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
2696

2697
	for_each_node_state(n, N_HIGH_MEMORY) {
L
Linus Torvalds 已提交
2698 2699 2700 2701 2702 2703 2704 2705

		/* 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);

2706 2707 2708
		/* Penalize nodes under us ("prefer the next node") */
		val += (n < node);

L
Linus Torvalds 已提交
2709
		/* Give preference to headless and unused nodes */
2710 2711
		tmp = cpumask_of_node(n);
		if (!cpumask_empty(tmp))
L
Linus Torvalds 已提交
2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729
			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;
}

2730 2731 2732 2733 2734 2735 2736

/*
 * 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
2737
{
2738
	int j;
L
Linus Torvalds 已提交
2739
	struct zonelist *zonelist;
2740

2741
	zonelist = &pgdat->node_zonelists[0];
2742
	for (j = 0; zonelist->_zonerefs[j].zone != NULL; j++)
2743 2744 2745
		;
	j = build_zonelists_node(NODE_DATA(node), zonelist, j,
							MAX_NR_ZONES - 1);
2746 2747
	zonelist->_zonerefs[j].zone = NULL;
	zonelist->_zonerefs[j].zone_idx = 0;
2748 2749
}

2750 2751 2752 2753 2754 2755 2756 2757
/*
 * Build gfp_thisnode zonelists
 */
static void build_thisnode_zonelists(pg_data_t *pgdat)
{
	int j;
	struct zonelist *zonelist;

2758 2759
	zonelist = &pgdat->node_zonelists[1];
	j = build_zonelists_node(pgdat, zonelist, 0, MAX_NR_ZONES - 1);
2760 2761
	zonelist->_zonerefs[j].zone = NULL;
	zonelist->_zonerefs[j].zone_idx = 0;
2762 2763
}

2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778
/*
 * 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;

2779 2780 2781 2782 2783 2784 2785
	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)) {
2786 2787
				zoneref_set_zone(z,
					&zonelist->_zonerefs[pos++]);
2788
				check_highest_zone(zone_type);
2789 2790 2791
			}
		}
	}
2792 2793
	zonelist->_zonerefs[pos].zone = NULL;
	zonelist->_zonerefs[pos].zone_idx = 0;
2794 2795 2796 2797 2798 2799 2800 2801 2802
}

static int default_zonelist_order(void)
{
	int nid, zone_type;
	unsigned long low_kmem_size,total_size;
	struct zone *z;
	int average_size;
	/*
T
Thomas Weber 已提交
2803
         * ZONE_DMA and ZONE_DMA32 can be very small area in the system.
2804 2805
	 * If they are really small and used heavily, the system can fall
	 * into OOM very easily.
2806
	 * This function detect ZONE_DMA/DMA32 size and configures zone order.
2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817
	 */
	/* 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;
2818 2819 2820 2821 2822 2823 2824 2825 2826
			} else if (zone_type == ZONE_NORMAL) {
				/*
				 * If any node has only lowmem, then node order
				 * is preferred to allow kernel allocations
				 * locally; otherwise, they can easily infringe
				 * on other nodes when there is an abundance of
				 * lowmem available to allocate from.
				 */
				return ZONELIST_ORDER_NODE;
2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837
			}
		}
	}
	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.
         */
2838 2839
	average_size = total_size /
				(nodes_weight(node_states[N_HIGH_MEMORY]) + 1);
2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870
	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
2871
	nodemask_t used_mask;
2872 2873 2874
	int local_node, prev_node;
	struct zonelist *zonelist;
	int order = current_zonelist_order;
L
Linus Torvalds 已提交
2875 2876

	/* initialize zonelists */
2877
	for (i = 0; i < MAX_ZONELISTS; i++) {
L
Linus Torvalds 已提交
2878
		zonelist = pgdat->node_zonelists + i;
2879 2880
		zonelist->_zonerefs[0].zone = NULL;
		zonelist->_zonerefs[0].zone_idx = 0;
L
Linus Torvalds 已提交
2881 2882 2883 2884
	}

	/* NUMA-aware ordering of nodes */
	local_node = pgdat->node_id;
2885
	load = nr_online_nodes;
L
Linus Torvalds 已提交
2886 2887
	prev_node = local_node;
	nodes_clear(used_mask);
2888 2889 2890 2891

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

L
Linus Torvalds 已提交
2892
	while ((node = find_next_best_node(local_node, &used_mask)) >= 0) {
2893 2894 2895 2896 2897 2898 2899 2900 2901
		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
2902 2903 2904 2905 2906
		/*
		 * 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.
		 */
2907
		if (distance != node_distance(local_node, prev_node))
2908 2909
			node_load[node] = load;

L
Linus Torvalds 已提交
2910 2911
		prev_node = node;
		load--;
2912 2913 2914 2915 2916
		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
2917

2918 2919 2920
	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
2921
	}
2922 2923

	build_thisnode_zonelists(pgdat);
L
Linus Torvalds 已提交
2924 2925
}

2926
/* Construct the zonelist performance cache - see further mmzone.h */
2927
static void build_zonelist_cache(pg_data_t *pgdat)
2928
{
2929 2930
	struct zonelist *zonelist;
	struct zonelist_cache *zlc;
2931
	struct zoneref *z;
2932

2933 2934 2935
	zonelist = &pgdat->node_zonelists[0];
	zonelist->zlcache_ptr = zlc = &zonelist->zlcache;
	bitmap_zero(zlc->fullzones, MAX_ZONES_PER_ZONELIST);
2936 2937
	for (z = zonelist->_zonerefs; z->zone; z++)
		zlc->z_to_n[z - zonelist->_zonerefs] = zonelist_node_idx(z);
2938 2939
}

2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957
#ifdef CONFIG_HAVE_MEMORYLESS_NODES
/*
 * Return node id of node used for "local" allocations.
 * I.e., first node id of first zone in arg node's generic zonelist.
 * Used for initializing percpu 'numa_mem', which is used primarily
 * for kernel allocations, so use GFP_KERNEL flags to locate zonelist.
 */
int local_memory_node(int node)
{
	struct zone *zone;

	(void)first_zones_zonelist(node_zonelist(node, GFP_KERNEL),
				   gfp_zone(GFP_KERNEL),
				   NULL,
				   &zone);
	return zone->node;
}
#endif
2958

L
Linus Torvalds 已提交
2959 2960
#else	/* CONFIG_NUMA */

2961 2962 2963 2964 2965 2966
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
2967
{
2968
	int node, local_node;
2969 2970
	enum zone_type j;
	struct zonelist *zonelist;
L
Linus Torvalds 已提交
2971 2972 2973

	local_node = pgdat->node_id;

2974 2975
	zonelist = &pgdat->node_zonelists[0];
	j = build_zonelists_node(pgdat, zonelist, 0, MAX_NR_ZONES - 1);
L
Linus Torvalds 已提交
2976

2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989
	/*
	 * 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
2990
	}
2991 2992 2993 2994 2995 2996 2997
	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);
	}

2998 2999
	zonelist->_zonerefs[j].zone = NULL;
	zonelist->_zonerefs[j].zone_idx = 0;
L
Linus Torvalds 已提交
3000 3001
}

3002
/* non-NUMA variant of zonelist performance cache - just NULL zlcache_ptr */
3003
static void build_zonelist_cache(pg_data_t *pgdat)
3004
{
3005
	pgdat->node_zonelists[0].zlcache_ptr = NULL;
3006 3007
}

L
Linus Torvalds 已提交
3008 3009
#endif	/* CONFIG_NUMA */

3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026
/*
 * 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.
 *
 * 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.
 */
static void setup_pageset(struct per_cpu_pageset *p, unsigned long batch);
static DEFINE_PER_CPU(struct per_cpu_pageset, boot_pageset);
3027
static void setup_zone_pageset(struct zone *zone);
3028

3029 3030 3031 3032 3033 3034
/*
 * Global mutex to protect against size modification of zonelists
 * as well as to serialize pageset setup for the new populated zone.
 */
DEFINE_MUTEX(zonelists_mutex);

3035
/* return values int ....just for stop_machine() */
3036
static __init_refok int __build_all_zonelists(void *data)
L
Linus Torvalds 已提交
3037
{
3038
	int nid;
3039
	int cpu;
3040

3041 3042 3043
#ifdef CONFIG_NUMA
	memset(node_load, 0, sizeof(node_load));
#endif
3044
	for_each_online_node(nid) {
3045 3046 3047 3048
		pg_data_t *pgdat = NODE_DATA(nid);

		build_zonelists(pgdat);
		build_zonelist_cache(pgdat);
3049
	}
3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063

	/*
	 * Initialize the boot_pagesets that are going to be used
	 * for bootstrapping processors. The real pagesets for
	 * each zone will be allocated later when the per cpu
	 * allocator is available.
	 *
	 * boot_pagesets are used also for bootstrapping offline
	 * cpus if the system is already booted because the pagesets
	 * are needed to initialize allocators on a specific cpu too.
	 * F.e. the percpu allocator needs the page allocator which
	 * needs the percpu allocator in order to allocate its pagesets
	 * (a chicken-egg dilemma).
	 */
3064
	for_each_possible_cpu(cpu) {
3065 3066
		setup_pageset(&per_cpu(boot_pageset, cpu), 0);

3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080
#ifdef CONFIG_HAVE_MEMORYLESS_NODES
		/*
		 * We now know the "local memory node" for each node--
		 * i.e., the node of the first zone in the generic zonelist.
		 * Set up numa_mem percpu variable for on-line cpus.  During
		 * boot, only the boot cpu should be on-line;  we'll init the
		 * secondary cpus' numa_mem as they come on-line.  During
		 * node/memory hotplug, we'll fixup all on-line cpus.
		 */
		if (cpu_online(cpu))
			set_cpu_numa_mem(cpu, local_memory_node(cpu_to_node(cpu)));
#endif
	}

3081 3082 3083
	return 0;
}

3084 3085 3086 3087
/*
 * Called with zonelists_mutex held always
 * unless system_state == SYSTEM_BOOTING.
 */
3088
void build_all_zonelists(void *data)
3089
{
3090 3091
	set_zonelist_order();

3092
	if (system_state == SYSTEM_BOOTING) {
3093
		__build_all_zonelists(NULL);
3094
		mminit_verify_zonelist();
3095 3096
		cpuset_init_current_mems_allowed();
	} else {
S
Simon Arlott 已提交
3097
		/* we have to stop all cpus to guarantee there is no user
3098
		   of zonelist */
3099 3100 3101 3102 3103
#ifdef CONFIG_MEMORY_HOTPLUG
		if (data)
			setup_zone_pageset((struct zone *)data);
#endif
		stop_machine(__build_all_zonelists, NULL, NULL);
3104 3105
		/* cpuset refresh routine should be here */
	}
3106
	vm_total_pages = nr_free_pagecache_pages();
3107 3108 3109 3110 3111 3112 3113
	/*
	 * 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
	 */
3114
	if (vm_total_pages < (pageblock_nr_pages * MIGRATE_TYPES))
3115 3116 3117 3118 3119 3120
		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",
3121
			nr_online_nodes,
3122
			zonelist_order_name[current_zonelist_order],
3123
			page_group_by_mobility_disabled ? "off" : "on",
3124 3125 3126 3127
			vm_total_pages);
#ifdef CONFIG_NUMA
	printk("Policy zone: %s\n", zone_names[policy_zone]);
#endif
L
Linus Torvalds 已提交
3128 3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142
}

/*
 * 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

3143
#ifndef CONFIG_MEMORY_HOTPLUG
3144
static inline unsigned long wait_table_hash_nr_entries(unsigned long pages)
L
Linus Torvalds 已提交
3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161
{
	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);
}
3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184
#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
3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197

/*
 * 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))

3198
/*
3199
 * Mark a number of pageblocks as MIGRATE_RESERVE. The number
3200 3201
 * 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
3202 3203 3204 3205 3206 3207 3208
 * 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;
3209 3210
	unsigned long block_migratetype;
	int reserve;
3211 3212 3213 3214

	/* 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;
3215
	reserve = roundup(min_wmark_pages(zone), pageblock_nr_pages) >>
3216
							pageblock_order;
3217

3218 3219 3220 3221 3222 3223 3224 3225 3226
	/*
	 * Reserve blocks are generally in place to help high-order atomic
	 * allocations that are short-lived. A min_free_kbytes value that
	 * would result in more than 2 reserve blocks for atomic allocations
	 * is assumed to be in place to help anti-fragmentation for the
	 * future allocation of hugepages at runtime.
	 */
	reserve = min(2, reserve);

3227
	for (pfn = start_pfn; pfn < end_pfn; pfn += pageblock_nr_pages) {
3228 3229 3230 3231
		if (!pfn_valid(pfn))
			continue;
		page = pfn_to_page(pfn);

3232 3233 3234 3235
		/* Watch out for overlapping nodes */
		if (page_to_nid(page) != zone_to_nid(zone))
			continue;

3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265
		/* 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
3266

L
Linus Torvalds 已提交
3267 3268 3269 3270 3271
/*
 * 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.
 */
3272
void __meminit memmap_init_zone(unsigned long size, int nid, unsigned long zone,
D
Dave Hansen 已提交
3273
		unsigned long start_pfn, enum memmap_context context)
L
Linus Torvalds 已提交
3274 3275
{
	struct page *page;
A
Andy Whitcroft 已提交
3276 3277
	unsigned long end_pfn = start_pfn + size;
	unsigned long pfn;
3278
	struct zone *z;
L
Linus Torvalds 已提交
3279

3280 3281 3282
	if (highest_memmap_pfn < end_pfn - 1)
		highest_memmap_pfn = end_pfn - 1;

3283
	z = &NODE_DATA(nid)->node_zones[zone];
3284
	for (pfn = start_pfn; pfn < end_pfn; pfn++) {
D
Dave Hansen 已提交
3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295
		/*
		 * 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
3296 3297
		page = pfn_to_page(pfn);
		set_page_links(page, zone, nid, pfn);
3298
		mminit_verify_page_links(page, zone, nid, pfn);
3299
		init_page_count(page);
L
Linus Torvalds 已提交
3300 3301
		reset_page_mapcount(page);
		SetPageReserved(page);
3302 3303 3304 3305 3306
		/*
		 * 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
3307 3308 3309
		 * kernel allocations are made. Later some blocks near
		 * the start are marked MIGRATE_RESERVE by
		 * setup_zone_migrate_reserve()
3310 3311 3312 3313 3314
		 *
		 * 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.
3315
		 */
3316 3317 3318
		if ((z->zone_start_pfn <= pfn)
		    && (pfn < z->zone_start_pfn + z->spanned_pages)
		    && !(pfn & (pageblock_nr_pages - 1)))
3319
			set_pageblock_migratetype(page, MIGRATE_MOVABLE);
3320

L
Linus Torvalds 已提交
3321 3322 3323 3324
		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))
3325
			set_page_address(page, __va(pfn << PAGE_SHIFT));
L
Linus Torvalds 已提交
3326 3327 3328 3329
#endif
	}
}

3330
static void __meminit zone_init_free_lists(struct zone *zone)
L
Linus Torvalds 已提交
3331
{
3332 3333 3334
	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
3335 3336 3337 3338 3339 3340
		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
3341
	memmap_init_zone((size), (nid), (zone), (start_pfn), MEMMAP_EARLY)
L
Linus Torvalds 已提交
3342 3343
#endif

3344
static int zone_batchsize(struct zone *zone)
3345
{
3346
#ifdef CONFIG_MMU
3347 3348 3349 3350
	int batch;

	/*
	 * The per-cpu-pages pools are set to around 1000th of the
3351
	 * size of the zone.  But no more than 1/2 of a meg.
3352 3353 3354 3355
	 *
	 * OK, so we don't know how big the cache is.  So guess.
	 */
	batch = zone->present_pages / 1024;
3356 3357
	if (batch * PAGE_SIZE > 512 * 1024)
		batch = (512 * 1024) / PAGE_SIZE;
3358 3359 3360 3361 3362
	batch /= 4;		/* We effectively *= 4 below */
	if (batch < 1)
		batch = 1;

	/*
3363 3364 3365
	 * 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.
3366
	 *
3367 3368 3369 3370
	 * 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.
3371
	 */
3372
	batch = rounddown_pow_of_two(batch + batch/2) - 1;
3373

3374
	return batch;
3375 3376 3377 3378 3379 3380 3381 3382 3383 3384 3385 3386 3387 3388 3389 3390 3391

#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
3392 3393
}

A
Adrian Bunk 已提交
3394
static void setup_pageset(struct per_cpu_pageset *p, unsigned long batch)
3395 3396
{
	struct per_cpu_pages *pcp;
3397
	int migratetype;
3398

3399 3400
	memset(p, 0, sizeof(*p));

3401
	pcp = &p->pcp;
3402 3403 3404
	pcp->count = 0;
	pcp->high = 6 * batch;
	pcp->batch = max(1UL, 1 * batch);
3405 3406
	for (migratetype = 0; migratetype < MIGRATE_PCPTYPES; migratetype++)
		INIT_LIST_HEAD(&pcp->lists[migratetype]);
3407 3408
}

3409 3410 3411 3412 3413 3414 3415 3416 3417 3418
/*
 * 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;

3419
	pcp = &p->pcp;
3420 3421 3422 3423 3424 3425
	pcp->high = high;
	pcp->batch = max(1UL, high/4);
	if ((high/4) > (PAGE_SHIFT * 8))
		pcp->batch = PAGE_SHIFT * 8;
}

3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443
static __meminit void setup_zone_pageset(struct zone *zone)
{
	int cpu;

	zone->pageset = alloc_percpu(struct per_cpu_pageset);

	for_each_possible_cpu(cpu) {
		struct per_cpu_pageset *pcp = per_cpu_ptr(zone->pageset, cpu);

		setup_pageset(pcp, zone_batchsize(zone));

		if (percpu_pagelist_fraction)
			setup_pagelist_highmark(pcp,
				(zone->present_pages /
					percpu_pagelist_fraction));
	}
}

3444
/*
3445 3446
 * Allocate per cpu pagesets and initialize them.
 * Before this call only boot pagesets were available.
3447
 */
3448
void __init setup_per_cpu_pageset(void)
3449
{
3450
	struct zone *zone;
3451

3452 3453
	for_each_populated_zone(zone)
		setup_zone_pageset(zone);
3454 3455
}

S
Sam Ravnborg 已提交
3456
static noinline __init_refok
3457
int zone_wait_table_init(struct zone *zone, unsigned long zone_size_pages)
3458 3459 3460
{
	int i;
	struct pglist_data *pgdat = zone->zone_pgdat;
3461
	size_t alloc_size;
3462 3463 3464 3465 3466

	/*
	 * The per-page waitqueue mechanism uses hashed waitqueues
	 * per zone.
	 */
3467 3468 3469 3470
	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);
3471 3472 3473
	alloc_size = zone->wait_table_hash_nr_entries
					* sizeof(wait_queue_head_t);

3474
	if (!slab_is_available()) {
3475 3476 3477 3478 3479 3480 3481 3482 3483 3484 3485 3486 3487
		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.
		 */
3488
		zone->wait_table = vmalloc(alloc_size);
3489 3490 3491
	}
	if (!zone->wait_table)
		return -ENOMEM;
3492

3493
	for(i = 0; i < zone->wait_table_hash_nr_entries; ++i)
3494
		init_waitqueue_head(zone->wait_table + i);
3495 3496

	return 0;
3497 3498
}

3499 3500 3501 3502 3503 3504
static int __zone_pcp_update(void *data)
{
	struct zone *zone = data;
	int cpu;
	unsigned long batch = zone_batchsize(zone), flags;

3505
	for_each_possible_cpu(cpu) {
3506 3507 3508
		struct per_cpu_pageset *pset;
		struct per_cpu_pages *pcp;

3509
		pset = per_cpu_ptr(zone->pageset, cpu);
3510 3511 3512
		pcp = &pset->pcp;

		local_irq_save(flags);
3513
		free_pcppages_bulk(zone, pcp->count, pcp);
3514 3515 3516 3517 3518 3519 3520 3521 3522 3523 3524
		setup_pageset(pset, batch);
		local_irq_restore(flags);
	}
	return 0;
}

void zone_pcp_update(struct zone *zone)
{
	stop_machine(__zone_pcp_update, zone, NULL);
}

3525
static __meminit void zone_pcp_init(struct zone *zone)
3526
{
3527 3528 3529 3530 3531 3532
	/*
	 * per cpu subsystem is not up at this point. The following code
	 * relies on the ability of the linker to provide the
	 * offset of a (static) per cpu variable into the per cpu area.
	 */
	zone->pageset = &boot_pageset;
3533

A
Anton Blanchard 已提交
3534
	if (zone->present_pages)
3535 3536 3537
		printk(KERN_DEBUG "  %s zone: %lu pages, LIFO batch:%u\n",
			zone->name, zone->present_pages,
					 zone_batchsize(zone));
3538 3539
}

3540 3541
__meminit int init_currently_empty_zone(struct zone *zone,
					unsigned long zone_start_pfn,
D
Dave Hansen 已提交
3542 3543
					unsigned long size,
					enum memmap_context context)
3544 3545
{
	struct pglist_data *pgdat = zone->zone_pgdat;
3546 3547 3548 3549
	int ret;
	ret = zone_wait_table_init(zone, size);
	if (ret)
		return ret;
3550 3551 3552 3553
	pgdat->nr_zones = zone_idx(zone) + 1;

	zone->zone_start_pfn = zone_start_pfn;

3554 3555 3556 3557 3558 3559
	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));

3560
	zone_init_free_lists(zone);
3561 3562

	return 0;
3563 3564
}

3565 3566 3567 3568 3569
#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
 */
3570
static int __meminit first_active_region_index_in_nid(int nid)
3571 3572 3573 3574 3575 3576 3577 3578 3579 3580 3581 3582
{
	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
3583
 * Note: nid == MAX_NUMNODES returns next region regardless of node
3584
 */
3585
static int __meminit next_active_region_index_in_nid(int index, int nid)
3586 3587 3588 3589 3590 3591 3592 3593 3594 3595 3596 3597 3598 3599 3600
{
	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
 */
3601
int __meminit __early_pfn_to_nid(unsigned long pfn)
3602 3603 3604 3605 3606 3607 3608 3609 3610 3611
{
	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;
	}
3612 3613
	/* This is a memory hole */
	return -1;
3614 3615 3616
}
#endif /* CONFIG_HAVE_ARCH_EARLY_PFN_TO_NID */

3617 3618
int __meminit early_pfn_to_nid(unsigned long pfn)
{
3619 3620 3621 3622 3623 3624 3625
	int nid;

	nid = __early_pfn_to_nid(pfn);
	if (nid >= 0)
		return nid;
	/* just returns 0 */
	return 0;
3626 3627
}

3628 3629 3630 3631 3632 3633 3634 3635 3636 3637 3638
#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
3639

3640 3641 3642 3643 3644 3645 3646
/* 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
3647 3648
 * @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
3649 3650 3651 3652 3653 3654 3655 3656 3657 3658 3659 3660 3661 3662 3663 3664 3665 3666 3667 3668 3669 3670 3671 3672 3673 3674 3675
 *
 * 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);
	}
}

3676 3677 3678 3679 3680 3681 3682 3683 3684 3685 3686 3687 3688 3689 3690 3691 3692 3693 3694 3695 3696 3697 3698 3699 3700 3701 3702 3703 3704 3705 3706 3707 3708 3709 3710
#ifdef CONFIG_HAVE_MEMBLOCK
u64 __init find_memory_core_early(int nid, u64 size, u64 align,
					u64 goal, u64 limit)
{
	int i;

	/* Need to go over early_node_map to find out good range for node */
	for_each_active_range_index_in_nid(i, nid) {
		u64 addr;
		u64 ei_start, ei_last;
		u64 final_start, final_end;

		ei_last = early_node_map[i].end_pfn;
		ei_last <<= PAGE_SHIFT;
		ei_start = early_node_map[i].start_pfn;
		ei_start <<= PAGE_SHIFT;

		final_start = max(ei_start, goal);
		final_end = min(ei_last, limit);

		if (final_start >= final_end)
			continue;

		addr = memblock_find_in_range(final_start, final_end, size, align);

		if (addr == MEMBLOCK_ERROR)
			continue;

		return addr;
	}

	return MEMBLOCK_ERROR;
}
#endif

3711 3712 3713 3714 3715 3716 3717 3718 3719 3720 3721 3722 3723 3724 3725
int __init add_from_early_node_map(struct range *range, int az,
				   int nr_range, int nid)
{
	int i;
	u64 start, end;

	/* need to go over early_node_map to find out good range for node */
	for_each_active_range_index_in_nid(i, nid) {
		start = early_node_map[i].start_pfn;
		end = early_node_map[i].end_pfn;
		nr_range = add_range(range, az, nr_range, start, end);
	}
	return nr_range;
}

3726
#ifdef CONFIG_NO_BOOTMEM
3727 3728 3729 3730
void * __init __alloc_memory_core_early(int nid, u64 size, u64 align,
					u64 goal, u64 limit)
{
	void *ptr;
3731
	u64 addr;
3732

3733 3734
	if (limit > memblock.current_limit)
		limit = memblock.current_limit;
3735

3736
	addr = find_memory_core_early(nid, size, align, goal, limit);
3737

3738 3739
	if (addr == MEMBLOCK_ERROR)
		return NULL;
3740

3741 3742 3743 3744 3745 3746 3747 3748 3749
	ptr = phys_to_virt(addr);
	memset(ptr, 0, size);
	memblock_x86_reserve_range(addr, addr + size, "BOOTMEM");
	/*
	 * The min_count is set to 0 so that bootmem allocated blocks
	 * are never reported as leaks.
	 */
	kmemleak_alloc(ptr, size, 0, 0);
	return ptr;
3750
}
3751
#endif
3752 3753


3754 3755 3756
void __init work_with_active_regions(int nid, work_fn_t work_fn, void *data)
{
	int i;
3757
	int ret;
3758

3759 3760 3761 3762 3763 3764
	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;
	}
3765
}
3766 3767
/**
 * sparse_memory_present_with_active_regions - Call memory_present for each active range
3768
 * @nid: The node to call memory_present for. If MAX_NUMNODES, all nodes will be used.
3769 3770 3771
 *
 * If an architecture guarantees that all ranges registered with
 * add_active_ranges() contain no holes and may be freed, this
3772
 * function may be used instead of calling memory_present() manually.
3773 3774 3775 3776 3777 3778 3779 3780 3781 3782 3783 3784 3785
 */
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
3786 3787 3788
 * @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.
3789 3790 3791 3792
 *
 * 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
3793
 * PFNs will be 0.
3794
 */
3795
void __meminit get_pfn_range_for_nid(unsigned int nid,
3796 3797 3798 3799 3800 3801 3802 3803 3804 3805 3806
			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);
	}

3807
	if (*start_pfn == -1UL)
3808 3809 3810
		*start_pfn = 0;
}

M
Mel Gorman 已提交
3811 3812 3813 3814 3815
/*
 * 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
3816
static void __init find_usable_zone_for_movable(void)
M
Mel Gorman 已提交
3817 3818 3819 3820 3821 3822 3823 3824 3825 3826 3827 3828 3829 3830 3831 3832 3833 3834 3835 3836 3837 3838 3839 3840 3841
{
	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
3842
static void __meminit adjust_zone_range_for_zone_movable(int nid,
M
Mel Gorman 已提交
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
					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;
	}
}

3868 3869 3870 3871
/*
 * 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
3872
static unsigned long __meminit zone_spanned_pages_in_node(int nid,
3873 3874 3875 3876 3877 3878 3879 3880 3881 3882
					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
3883 3884 3885
	adjust_zone_range_for_zone_movable(nid, zone_type,
				node_start_pfn, node_end_pfn,
				&zone_start_pfn, &zone_end_pfn);
3886 3887 3888 3889 3890 3891 3892 3893 3894 3895 3896 3897 3898 3899 3900

	/* 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,
3901
 * then all holes in the requested range will be accounted for.
3902
 */
3903
unsigned long __meminit __absent_pages_in_range(int nid,
3904 3905 3906 3907 3908 3909 3910 3911 3912 3913 3914 3915
				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;

3916 3917
	prev_end_pfn = min(early_node_map[i].start_pfn, range_end_pfn);

3918 3919
	/* Account for ranges before physical memory on this node */
	if (early_node_map[i].start_pfn > range_start_pfn)
3920
		hole_pages = prev_end_pfn - range_start_pfn;
3921 3922 3923 3924 3925 3926 3927 3928 3929 3930 3931 3932 3933 3934 3935 3936 3937 3938 3939 3940

	/* 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;
	}

3941 3942
	/* Account for ranges past physical memory on this node */
	if (range_end_pfn > prev_end_pfn)
3943
		hole_pages += range_end_pfn -
3944 3945
				max(range_start_pfn, prev_end_pfn);

3946 3947 3948 3949 3950 3951 3952 3953
	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
 *
3954
 * It returns the number of pages frames in memory holes within a range.
3955 3956 3957 3958 3959 3960 3961 3962
 */
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
3963
static unsigned long __meminit zone_absent_pages_in_node(int nid,
3964 3965 3966
					unsigned long zone_type,
					unsigned long *ignored)
{
3967 3968 3969 3970 3971 3972 3973 3974 3975
	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
3976 3977 3978
	adjust_zone_range_for_zone_movable(nid, zone_type,
			node_start_pfn, node_end_pfn,
			&zone_start_pfn, &zone_end_pfn);
3979
	return __absent_pages_in_range(nid, zone_start_pfn, zone_end_pfn);
3980
}
3981

3982
#else
P
Paul Mundt 已提交
3983
static inline unsigned long __meminit zone_spanned_pages_in_node(int nid,
3984 3985 3986 3987 3988 3989
					unsigned long zone_type,
					unsigned long *zones_size)
{
	return zones_size[zone_type];
}

P
Paul Mundt 已提交
3990
static inline unsigned long __meminit zone_absent_pages_in_node(int nid,
3991 3992 3993 3994 3995 3996 3997 3998
						unsigned long zone_type,
						unsigned long *zholes_size)
{
	if (!zholes_size)
		return 0;

	return zholes_size[zone_type];
}
3999

4000 4001
#endif

4002
static void __meminit calculate_node_totalpages(struct pglist_data *pgdat,
4003 4004 4005 4006 4007 4008 4009 4010 4011 4012 4013 4014 4015 4016 4017 4018 4019 4020 4021 4022
		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);
}

4023 4024 4025
#ifndef CONFIG_SPARSEMEM
/*
 * Calculate the size of the zone->blockflags rounded to an unsigned long
4026 4027
 * 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
4028 4029 4030 4031 4032 4033 4034
 * 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;

4035 4036
	usemapsize = roundup(zonesize, pageblock_nr_pages);
	usemapsize = usemapsize >> pageblock_order;
4037 4038 4039 4040 4041 4042 4043 4044 4045 4046 4047
	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;
4048
	if (usemapsize)
4049 4050 4051
		zone->pageblock_flags = alloc_bootmem_node(pgdat, usemapsize);
}
#else
4052
static inline void setup_usemap(struct pglist_data *pgdat,
4053 4054 4055
				struct zone *zone, unsigned long zonesize) {}
#endif /* CONFIG_SPARSEMEM */

4056
#ifdef CONFIG_HUGETLB_PAGE_SIZE_VARIABLE
4057 4058 4059 4060 4061 4062 4063 4064 4065 4066

/* 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;
}

4067 4068 4069 4070 4071 4072 4073 4074 4075 4076 4077 4078 4079 4080 4081
/* 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 */

4082 4083 4084 4085 4086 4087 4088 4089 4090 4091
/*
 * 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;
}
4092 4093 4094 4095
#define set_pageblock_order(x)	do {} while (0)

#endif /* CONFIG_HUGETLB_PAGE_SIZE_VARIABLE */

L
Linus Torvalds 已提交
4096 4097 4098 4099 4100 4101
/*
 * Set up the zone data structures:
 *   - mark all pages reserved
 *   - mark all memory queues empty
 *   - clear the memory bitmaps
 */
4102
static void __paginginit free_area_init_core(struct pglist_data *pgdat,
L
Linus Torvalds 已提交
4103 4104
		unsigned long *zones_size, unsigned long *zholes_size)
{
4105
	enum zone_type j;
4106
	int nid = pgdat->node_id;
L
Linus Torvalds 已提交
4107
	unsigned long zone_start_pfn = pgdat->node_start_pfn;
4108
	int ret;
L
Linus Torvalds 已提交
4109

4110
	pgdat_resize_init(pgdat);
L
Linus Torvalds 已提交
4111 4112 4113
	pgdat->nr_zones = 0;
	init_waitqueue_head(&pgdat->kswapd_wait);
	pgdat->kswapd_max_order = 0;
4114
	pgdat_page_cgroup_init(pgdat);
L
Linus Torvalds 已提交
4115 4116 4117
	
	for (j = 0; j < MAX_NR_ZONES; j++) {
		struct zone *zone = pgdat->node_zones + j;
4118
		unsigned long size, realsize, memmap_pages;
4119
		enum lru_list l;
L
Linus Torvalds 已提交
4120

4121 4122 4123
		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
4124

4125 4126 4127 4128 4129
		/*
		 * 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
		 */
4130 4131
		memmap_pages =
			PAGE_ALIGN(size * sizeof(struct page)) >> PAGE_SHIFT;
4132 4133
		if (realsize >= memmap_pages) {
			realsize -= memmap_pages;
4134 4135 4136 4137
			if (memmap_pages)
				printk(KERN_DEBUG
				       "  %s zone: %lu pages used for memmap\n",
				       zone_names[j], memmap_pages);
4138 4139 4140 4141 4142
		} else
			printk(KERN_WARNING
				"  %s zone: %lu pages exceeds realsize %lu\n",
				zone_names[j], memmap_pages, realsize);

4143 4144
		/* Account for reserved pages */
		if (j == 0 && realsize > dma_reserve) {
4145
			realsize -= dma_reserve;
Y
Yinghai Lu 已提交
4146
			printk(KERN_DEBUG "  %s zone: %lu pages reserved\n",
4147
					zone_names[0], dma_reserve);
4148 4149
		}

4150
		if (!is_highmem_idx(j))
L
Linus Torvalds 已提交
4151 4152 4153 4154 4155
			nr_kernel_pages += realsize;
		nr_all_pages += realsize;

		zone->spanned_pages = size;
		zone->present_pages = realsize;
4156
#ifdef CONFIG_NUMA
4157
		zone->node = nid;
4158
		zone->min_unmapped_pages = (realsize*sysctl_min_unmapped_ratio)
4159
						/ 100;
4160
		zone->min_slab_pages = (realsize * sysctl_min_slab_ratio) / 100;
4161
#endif
L
Linus Torvalds 已提交
4162 4163 4164
		zone->name = zone_names[j];
		spin_lock_init(&zone->lock);
		spin_lock_init(&zone->lru_lock);
4165
		zone_seqlock_init(zone);
L
Linus Torvalds 已提交
4166 4167
		zone->zone_pgdat = pgdat;

4168
		zone_pcp_init(zone);
4169 4170
		for_each_lru(l) {
			INIT_LIST_HEAD(&zone->lru[l].list);
4171
			zone->reclaim_stat.nr_saved_scan[l] = 0;
4172
		}
4173 4174 4175 4176
		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;
4177
		zap_zone_vm_stats(zone);
4178
		zone->flags = 0;
L
Linus Torvalds 已提交
4179 4180 4181
		if (!size)
			continue;

4182
		set_pageblock_order(pageblock_default_order());
4183
		setup_usemap(pgdat, zone, size);
D
Dave Hansen 已提交
4184 4185
		ret = init_currently_empty_zone(zone, zone_start_pfn,
						size, MEMMAP_EARLY);
4186
		BUG_ON(ret);
4187
		memmap_init(size, nid, j, zone_start_pfn);
L
Linus Torvalds 已提交
4188 4189 4190 4191
		zone_start_pfn += size;
	}
}

S
Sam Ravnborg 已提交
4192
static void __init_refok alloc_node_mem_map(struct pglist_data *pgdat)
L
Linus Torvalds 已提交
4193 4194 4195 4196 4197
{
	/* Skip empty nodes */
	if (!pgdat->node_spanned_pages)
		return;

A
Andy Whitcroft 已提交
4198
#ifdef CONFIG_FLAT_NODE_MEM_MAP
L
Linus Torvalds 已提交
4199 4200
	/* ia64 gets its own node_mem_map, before this, without bootmem */
	if (!pgdat->node_mem_map) {
4201
		unsigned long size, start, end;
A
Andy Whitcroft 已提交
4202 4203
		struct page *map;

4204 4205 4206 4207 4208 4209 4210 4211 4212
		/*
		 * 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);
4213 4214 4215
		map = alloc_remap(pgdat->node_id, size);
		if (!map)
			map = alloc_bootmem_node(pgdat, size);
4216
		pgdat->node_mem_map = map + (pgdat->node_start_pfn - start);
L
Linus Torvalds 已提交
4217
	}
4218
#ifndef CONFIG_NEED_MULTIPLE_NODES
L
Linus Torvalds 已提交
4219 4220 4221
	/*
	 * With no DISCONTIG, the global mem_map is just set as node 0's
	 */
4222
	if (pgdat == NODE_DATA(0)) {
L
Linus Torvalds 已提交
4223
		mem_map = NODE_DATA(0)->node_mem_map;
4224 4225
#ifdef CONFIG_ARCH_POPULATES_NODE_MAP
		if (page_to_pfn(mem_map) != pgdat->node_start_pfn)
4226
			mem_map -= (pgdat->node_start_pfn - ARCH_PFN_OFFSET);
4227 4228
#endif /* CONFIG_ARCH_POPULATES_NODE_MAP */
	}
L
Linus Torvalds 已提交
4229
#endif
A
Andy Whitcroft 已提交
4230
#endif /* CONFIG_FLAT_NODE_MEM_MAP */
L
Linus Torvalds 已提交
4231 4232
}

4233 4234
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
4235
{
4236 4237
	pg_data_t *pgdat = NODE_DATA(nid);

L
Linus Torvalds 已提交
4238 4239
	pgdat->node_id = nid;
	pgdat->node_start_pfn = node_start_pfn;
4240
	calculate_node_totalpages(pgdat, zones_size, zholes_size);
L
Linus Torvalds 已提交
4241 4242

	alloc_node_mem_map(pgdat);
4243 4244 4245 4246 4247
#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
4248 4249 4250 4251

	free_area_init_core(pgdat, zones_size, zholes_size);
}

4252
#ifdef CONFIG_ARCH_POPULATES_NODE_MAP
M
Miklos Szeredi 已提交
4253 4254 4255 4256 4257 4258 4259 4260 4261 4262 4263 4264 4265 4266 4267 4268 4269 4270 4271 4272

#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

4273 4274 4275 4276 4277 4278 4279 4280 4281 4282 4283 4284 4285 4286 4287 4288 4289
/**
 * 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;

4290 4291 4292 4293 4294
	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);
4295

4296 4297
	mminit_validate_memmodel_limits(&start_pfn, &end_pfn);

4298 4299 4300 4301 4302 4303 4304 4305 4306 4307 4308 4309 4310 4311 4312 4313 4314 4315
	/* 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 */
4316
		if (start_pfn < early_node_map[i].start_pfn &&
4317 4318 4319 4320 4321 4322 4323 4324 4325 4326 4327 4328 4329 4330 4331 4332 4333 4334 4335 4336
				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;
}

/**
4337
 * remove_active_range - Shrink an existing registered range of PFNs
4338
 * @nid: The node id the range is on that should be shrunk
4339 4340
 * @start_pfn: The new PFN of the range
 * @end_pfn: The new PFN of the range
4341 4342
 *
 * i386 with NUMA use alloc_remap() to store a node_mem_map on a local node.
4343 4344 4345
 * 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.
4346
 */
4347 4348
void __init remove_active_range(unsigned int nid, unsigned long start_pfn,
				unsigned long end_pfn)
4349
{
4350 4351
	int i, j;
	int removed = 0;
4352

4353 4354 4355
	printk(KERN_DEBUG "remove_active_range (%d, %lu, %lu)\n",
			  nid, start_pfn, end_pfn);

4356
	/* Find the old active region end and shrink */
4357
	for_each_active_range_index_in_nid(i, nid) {
4358 4359
		if (early_node_map[i].start_pfn >= start_pfn &&
		    early_node_map[i].end_pfn <= end_pfn) {
4360
			/* clear it */
4361
			early_node_map[i].start_pfn = 0;
4362 4363 4364 4365
			early_node_map[i].end_pfn = 0;
			removed = 1;
			continue;
		}
4366 4367 4368 4369 4370 4371 4372 4373 4374 4375 4376 4377
		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;
4378
			continue;
4379
		}
4380 4381 4382 4383 4384 4385 4386 4387 4388 4389 4390 4391 4392 4393 4394 4395 4396 4397 4398
	}

	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--;
	}
4399 4400 4401 4402
}

/**
 * remove_all_active_ranges - Remove all currently registered regions
4403
 *
4404 4405 4406 4407
 * 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.
 */
4408
void __init remove_all_active_ranges(void)
4409 4410 4411 4412 4413 4414 4415 4416 4417 4418 4419 4420 4421 4422 4423 4424 4425 4426 4427 4428 4429
{
	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 */
4430
void __init sort_node_map(void)
4431 4432 4433 4434 4435 4436
{
	sort(early_node_map, (size_t)nr_nodemap_entries,
			sizeof(struct node_active_region),
			cmp_node_active_region, NULL);
}

4437
/* Find the lowest pfn for a node */
A
Adrian Bunk 已提交
4438
static unsigned long __init find_min_pfn_for_node(int nid)
4439 4440
{
	int i;
4441
	unsigned long min_pfn = ULONG_MAX;
4442

4443 4444
	/* Assuming a sorted map, the first range found has the starting pfn */
	for_each_active_range_index_in_nid(i, nid)
4445
		min_pfn = min(min_pfn, early_node_map[i].start_pfn);
4446

4447 4448
	if (min_pfn == ULONG_MAX) {
		printk(KERN_WARNING
4449
			"Could not find start_pfn for node %d\n", nid);
4450 4451 4452 4453
		return 0;
	}

	return min_pfn;
4454 4455 4456 4457 4458 4459
}

/**
 * find_min_pfn_with_active_regions - Find the minimum PFN registered
 *
 * It returns the minimum PFN based on information provided via
4460
 * add_active_range().
4461 4462 4463 4464 4465 4466
 */
unsigned long __init find_min_pfn_with_active_regions(void)
{
	return find_min_pfn_for_node(MAX_NUMNODES);
}

4467 4468 4469 4470 4471
/*
 * 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
4472
static unsigned long __init early_calculate_totalpages(void)
4473 4474 4475 4476
{
	int i;
	unsigned long totalpages = 0;

4477 4478
	for (i = 0; i < nr_nodemap_entries; i++) {
		unsigned long pages = early_node_map[i].end_pfn -
4479
						early_node_map[i].start_pfn;
4480 4481 4482 4483 4484
		totalpages += pages;
		if (pages)
			node_set_state(early_node_map[i].nid, N_HIGH_MEMORY);
	}
  	return totalpages;
4485 4486
}

M
Mel Gorman 已提交
4487 4488 4489 4490 4491 4492
/*
 * 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
4493
static void __init find_zone_movable_pfns_for_nodes(unsigned long *movable_pfn)
M
Mel Gorman 已提交
4494 4495 4496 4497
{
	int i, nid;
	unsigned long usable_startpfn;
	unsigned long kernelcore_node, kernelcore_remaining;
4498 4499
	/* save the state before borrow the nodemask */
	nodemask_t saved_node_state = node_states[N_HIGH_MEMORY];
4500 4501
	unsigned long totalpages = early_calculate_totalpages();
	int usable_nodes = nodes_weight(node_states[N_HIGH_MEMORY]);
M
Mel Gorman 已提交
4502

4503 4504 4505 4506 4507 4508 4509 4510 4511 4512 4513 4514 4515 4516 4517 4518 4519 4520 4521 4522 4523 4524
	/*
	 * 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
4525 4526
	/* If kernelcore was not specified, there is no ZONE_MOVABLE */
	if (!required_kernelcore)
4527
		goto out;
M
Mel Gorman 已提交
4528 4529 4530 4531 4532 4533 4534 4535

	/* 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;
4536
	for_each_node_state(nid, N_HIGH_MEMORY) {
M
Mel Gorman 已提交
4537 4538 4539 4540 4541 4542 4543 4544 4545 4546 4547 4548 4549 4550 4551 4552 4553 4554 4555 4556 4557 4558 4559 4560 4561 4562 4563 4564 4565 4566 4567 4568 4569 4570 4571 4572 4573 4574 4575 4576 4577 4578 4579 4580 4581 4582 4583 4584 4585 4586 4587 4588 4589 4590 4591 4592 4593 4594 4595 4596 4597 4598 4599 4600 4601 4602 4603 4604 4605 4606 4607 4608 4609 4610 4611 4612 4613 4614 4615 4616 4617 4618 4619 4620 4621 4622 4623 4624 4625
		/*
		 * 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);
4626 4627 4628 4629

out:
	/* restore the node_state */
	node_states[N_HIGH_MEMORY] = saved_node_state;
M
Mel Gorman 已提交
4630 4631
}

4632 4633 4634 4635 4636 4637 4638 4639 4640 4641 4642 4643 4644 4645
/* 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
}

4646 4647
/**
 * free_area_init_nodes - Initialise all pg_data_t and zone data
4648
 * @max_zone_pfn: an array of max PFNs for each zone
4649 4650 4651 4652 4653 4654 4655 4656 4657 4658 4659 4660 4661
 *
 * 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;
4662
	int i;
4663

4664 4665 4666
	/* Sort early_node_map as initialisation assumes it is sorted */
	sort_node_map();

4667 4668 4669 4670 4671 4672 4673 4674
	/* 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
4675 4676
		if (i == ZONE_MOVABLE)
			continue;
4677 4678 4679 4680 4681
		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
4682 4683 4684 4685 4686 4687
	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);
4688 4689 4690

	/* Print out the zone ranges */
	printk("Zone PFN ranges:\n");
M
Mel Gorman 已提交
4691 4692 4693
	for (i = 0; i < MAX_NR_ZONES; i++) {
		if (i == ZONE_MOVABLE)
			continue;
4694 4695 4696 4697 4698 4699
		printk("  %-8s ", zone_names[i]);
		if (arch_zone_lowest_possible_pfn[i] ==
				arch_zone_highest_possible_pfn[i])
			printk("empty\n");
		else
			printk("%0#10lx -> %0#10lx\n",
4700 4701
				arch_zone_lowest_possible_pfn[i],
				arch_zone_highest_possible_pfn[i]);
M
Mel Gorman 已提交
4702 4703 4704 4705 4706 4707 4708 4709
	}

	/* 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]);
	}
4710 4711 4712 4713

	/* 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++)
4714
		printk("  %3d: %0#10lx -> %0#10lx\n", early_node_map[i].nid,
4715 4716 4717 4718
						early_node_map[i].start_pfn,
						early_node_map[i].end_pfn);

	/* Initialise every node */
4719
	mminit_verify_pageflags_layout();
4720
	setup_nr_node_ids();
4721 4722
	for_each_online_node(nid) {
		pg_data_t *pgdat = NODE_DATA(nid);
4723
		free_area_init_node(nid, NULL,
4724
				find_min_pfn_for_node(nid), NULL);
4725 4726 4727 4728 4729

		/* Any memory on that node */
		if (pgdat->node_present_pages)
			node_set_state(nid, N_HIGH_MEMORY);
		check_for_regular_memory(pgdat);
4730 4731
	}
}
M
Mel Gorman 已提交
4732

4733
static int __init cmdline_parse_core(char *p, unsigned long *core)
M
Mel Gorman 已提交
4734 4735 4736 4737 4738 4739
{
	unsigned long long coremem;
	if (!p)
		return -EINVAL;

	coremem = memparse(p, &p);
4740
	*core = coremem >> PAGE_SHIFT;
M
Mel Gorman 已提交
4741

4742
	/* Paranoid check that UL is enough for the coremem value */
M
Mel Gorman 已提交
4743 4744 4745 4746
	WARN_ON((coremem >> PAGE_SHIFT) > ULONG_MAX);

	return 0;
}
M
Mel Gorman 已提交
4747

4748 4749 4750 4751 4752 4753 4754 4755 4756 4757 4758 4759 4760 4761 4762 4763 4764 4765
/*
 * 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
4766
early_param("kernelcore", cmdline_parse_kernelcore);
4767
early_param("movablecore", cmdline_parse_movablecore);
M
Mel Gorman 已提交
4768

4769 4770
#endif /* CONFIG_ARCH_POPULATES_NODE_MAP */

4771
/**
4772 4773
 * set_dma_reserve - set the specified number of pages reserved in the first zone
 * @new_dma_reserve: The number of pages to mark reserved
4774 4775 4776 4777
 *
 * 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
4778 4779 4780
 * 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.
4781 4782 4783 4784 4785 4786
 */
void __init set_dma_reserve(unsigned long new_dma_reserve)
{
	dma_reserve = new_dma_reserve;
}

4787
#ifndef CONFIG_NEED_MULTIPLE_NODES
4788 4789 4790 4791 4792
struct pglist_data __refdata contig_page_data = {
#ifndef CONFIG_NO_BOOTMEM
 .bdata = &bootmem_node_data[0]
#endif
 };
L
Linus Torvalds 已提交
4793
EXPORT_SYMBOL(contig_page_data);
4794
#endif
L
Linus Torvalds 已提交
4795 4796 4797

void __init free_area_init(unsigned long *zones_size)
{
4798
	free_area_init_node(0, zones_size,
L
Linus Torvalds 已提交
4799 4800 4801 4802 4803 4804 4805 4806
			__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;

4807
	if (action == CPU_DEAD || action == CPU_DEAD_FROZEN) {
4808 4809 4810 4811 4812 4813 4814 4815
		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.
		 */
4816
		vm_events_fold_cpu(cpu);
4817 4818 4819 4820 4821 4822 4823 4824

		/*
		 * 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.
		 */
4825
		refresh_cpu_vm_stats(cpu);
L
Linus Torvalds 已提交
4826 4827 4828 4829 4830 4831 4832 4833 4834
	}
	return NOTIFY_OK;
}

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

4835 4836 4837 4838 4839 4840 4841 4842
/*
 * 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;
4843
	enum zone_type i, j;
4844 4845 4846 4847 4848 4849 4850 4851 4852 4853 4854 4855

	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];
			}

4856 4857
			/* we treat the high watermark as reserved pages. */
			max += high_wmark_pages(zone);
4858 4859 4860 4861 4862 4863 4864 4865 4866

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

L
Linus Torvalds 已提交
4867 4868 4869 4870 4871 4872 4873 4874 4875
/*
 * 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;
4876
	enum zone_type j, idx;
L
Linus Torvalds 已提交
4877

4878
	for_each_online_pgdat(pgdat) {
L
Linus Torvalds 已提交
4879 4880 4881 4882 4883 4884
		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;

4885 4886
			idx = j;
			while (idx) {
L
Linus Torvalds 已提交
4887 4888
				struct zone *lower_zone;

4889 4890
				idx--;

L
Linus Torvalds 已提交
4891 4892 4893 4894 4895 4896 4897 4898 4899 4900
				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;
			}
		}
	}
4901 4902 4903

	/* update totalreserve_pages */
	calculate_totalreserve_pages();
L
Linus Torvalds 已提交
4904 4905
}

4906
/**
4907
 * setup_per_zone_wmarks - called when min_free_kbytes changes
4908
 * or when memory is hot-{added|removed}
4909
 *
4910 4911
 * 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
4912
 */
4913
void setup_per_zone_wmarks(void)
L
Linus Torvalds 已提交
4914 4915 4916 4917 4918 4919 4920 4921 4922 4923 4924 4925 4926
{
	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) {
4927 4928
		u64 tmp;

4929
		spin_lock_irqsave(&zone->lock, flags);
4930 4931
		tmp = (u64)pages_min * zone->present_pages;
		do_div(tmp, lowmem_pages);
L
Linus Torvalds 已提交
4932 4933
		if (is_highmem(zone)) {
			/*
N
Nick Piggin 已提交
4934 4935 4936 4937
			 * __GFP_HIGH and PF_MEMALLOC allocations usually don't
			 * need highmem pages, so cap pages_min to a small
			 * value here.
			 *
4938
			 * The WMARK_HIGH-WMARK_LOW and (WMARK_LOW-WMARK_MIN)
N
Nick Piggin 已提交
4939 4940
			 * deltas controls asynch page reclaim, and so should
			 * not be capped for highmem.
L
Linus Torvalds 已提交
4941 4942 4943 4944 4945 4946 4947 4948
			 */
			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;
4949
			zone->watermark[WMARK_MIN] = min_pages;
L
Linus Torvalds 已提交
4950
		} else {
N
Nick Piggin 已提交
4951 4952
			/*
			 * If it's a lowmem zone, reserve a number of pages
L
Linus Torvalds 已提交
4953 4954
			 * proportionate to the zone's size.
			 */
4955
			zone->watermark[WMARK_MIN] = tmp;
L
Linus Torvalds 已提交
4956 4957
		}

4958 4959
		zone->watermark[WMARK_LOW]  = min_wmark_pages(zone) + (tmp >> 2);
		zone->watermark[WMARK_HIGH] = min_wmark_pages(zone) + (tmp >> 1);
4960
		setup_zone_migrate_reserve(zone);
4961
		spin_unlock_irqrestore(&zone->lock, flags);
L
Linus Torvalds 已提交
4962
	}
4963 4964 4965

	/* update totalreserve_pages */
	calculate_totalreserve_pages();
L
Linus Torvalds 已提交
4966 4967
}

4968
/*
4969 4970 4971 4972 4973 4974 4975 4976 4977 4978 4979 4980 4981 4982 4983 4984 4985 4986 4987 4988
 * 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
 */
4989
void calculate_zone_inactive_ratio(struct zone *zone)
4990
{
4991
	unsigned int gb, ratio;
4992

4993 4994 4995
	/* Zone size in gigabytes */
	gb = zone->present_pages >> (30 - PAGE_SHIFT);
	if (gb)
4996
		ratio = int_sqrt(10 * gb);
4997 4998
	else
		ratio = 1;
4999

5000 5001
	zone->inactive_ratio = ratio;
}
5002

5003 5004 5005 5006 5007 5008
static void __init setup_per_zone_inactive_ratio(void)
{
	struct zone *zone;

	for_each_zone(zone)
		calculate_zone_inactive_ratio(zone);
5009 5010
}

L
Linus Torvalds 已提交
5011 5012 5013 5014 5015 5016 5017 5018 5019 5020 5021 5022 5023 5024 5025 5026 5027 5028 5029 5030 5031 5032 5033 5034
/*
 * 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
 */
5035
static int __init init_per_zone_wmark_min(void)
L
Linus Torvalds 已提交
5036 5037 5038 5039 5040 5041 5042 5043 5044 5045
{
	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;
5046
	setup_per_zone_wmarks();
L
Linus Torvalds 已提交
5047
	setup_per_zone_lowmem_reserve();
5048
	setup_per_zone_inactive_ratio();
L
Linus Torvalds 已提交
5049 5050
	return 0;
}
5051
module_init(init_per_zone_wmark_min)
L
Linus Torvalds 已提交
5052 5053 5054 5055 5056 5057 5058

/*
 * 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, 
5059
	void __user *buffer, size_t *length, loff_t *ppos)
L
Linus Torvalds 已提交
5060
{
5061
	proc_dointvec(table, write, buffer, length, ppos);
5062
	if (write)
5063
		setup_per_zone_wmarks();
L
Linus Torvalds 已提交
5064 5065 5066
	return 0;
}

5067 5068
#ifdef CONFIG_NUMA
int sysctl_min_unmapped_ratio_sysctl_handler(ctl_table *table, int write,
5069
	void __user *buffer, size_t *length, loff_t *ppos)
5070 5071 5072 5073
{
	struct zone *zone;
	int rc;

5074
	rc = proc_dointvec_minmax(table, write, buffer, length, ppos);
5075 5076 5077 5078
	if (rc)
		return rc;

	for_each_zone(zone)
5079
		zone->min_unmapped_pages = (zone->present_pages *
5080 5081 5082
				sysctl_min_unmapped_ratio) / 100;
	return 0;
}
5083 5084

int sysctl_min_slab_ratio_sysctl_handler(ctl_table *table, int write,
5085
	void __user *buffer, size_t *length, loff_t *ppos)
5086 5087 5088 5089
{
	struct zone *zone;
	int rc;

5090
	rc = proc_dointvec_minmax(table, write, buffer, length, ppos);
5091 5092 5093 5094 5095 5096 5097 5098
	if (rc)
		return rc;

	for_each_zone(zone)
		zone->min_slab_pages = (zone->present_pages *
				sysctl_min_slab_ratio) / 100;
	return 0;
}
5099 5100
#endif

L
Linus Torvalds 已提交
5101 5102 5103 5104 5105 5106
/*
 * 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
5107
 * minimum watermarks. The lowmem reserve ratio can only make sense
L
Linus Torvalds 已提交
5108 5109 5110
 * if in function of the boot time zone sizes.
 */
int lowmem_reserve_ratio_sysctl_handler(ctl_table *table, int write,
5111
	void __user *buffer, size_t *length, loff_t *ppos)
L
Linus Torvalds 已提交
5112
{
5113
	proc_dointvec_minmax(table, write, buffer, length, ppos);
L
Linus Torvalds 已提交
5114 5115 5116 5117
	setup_per_zone_lowmem_reserve();
	return 0;
}

5118 5119 5120 5121 5122 5123 5124
/*
 * 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,
5125
	void __user *buffer, size_t *length, loff_t *ppos)
5126 5127 5128 5129 5130
{
	struct zone *zone;
	unsigned int cpu;
	int ret;

5131
	ret = proc_dointvec_minmax(table, write, buffer, length, ppos);
5132 5133
	if (!write || (ret == -EINVAL))
		return ret;
5134
	for_each_populated_zone(zone) {
5135
		for_each_possible_cpu(cpu) {
5136 5137
			unsigned long  high;
			high = zone->present_pages / percpu_pagelist_fraction;
5138 5139
			setup_pagelist_highmark(
				per_cpu_ptr(zone->pageset, cpu), high);
5140 5141 5142 5143 5144
		}
	}
	return 0;
}

5145
int hashdist = HASHDIST_DEFAULT;
L
Linus Torvalds 已提交
5146 5147 5148 5149 5150 5151 5152 5153 5154 5155 5156 5157 5158 5159 5160 5161 5162 5163 5164 5165 5166 5167 5168 5169 5170 5171 5172 5173 5174 5175 5176 5177 5178 5179

#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
5180
		numentries = nr_kernel_pages;
L
Linus Torvalds 已提交
5181 5182 5183 5184 5185 5186 5187 5188 5189
		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);
5190 5191

		/* Make sure we've got at least a 0-order allocation.. */
5192 5193 5194 5195 5196 5197 5198 5199
		if (unlikely(flags & HASH_SMALL)) {
			/* Makes no sense without HASH_EARLY */
			WARN_ON(!(flags & HASH_EARLY));
			if (!(numentries >> *_hash_shift)) {
				numentries = 1UL << *_hash_shift;
				BUG_ON(!numentries);
			}
		} else if (unlikely((numentries * bucketsize) < PAGE_SIZE))
5200
			numentries = PAGE_SIZE / bucketsize;
L
Linus Torvalds 已提交
5201
	}
5202
	numentries = roundup_pow_of_two(numentries);
L
Linus Torvalds 已提交
5203 5204 5205 5206 5207 5208 5209 5210 5211 5212

	/* 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;

5213
	log2qty = ilog2(numentries);
L
Linus Torvalds 已提交
5214 5215 5216 5217

	do {
		size = bucketsize << log2qty;
		if (flags & HASH_EARLY)
5218
			table = alloc_bootmem_nopanic(size);
L
Linus Torvalds 已提交
5219 5220 5221
		else if (hashdist)
			table = __vmalloc(size, GFP_ATOMIC, PAGE_KERNEL);
		else {
5222 5223
			/*
			 * If bucketsize is not a power-of-two, we may free
5224 5225
			 * some pages at the end of hash table which
			 * alloc_pages_exact() automatically does
5226
			 */
5227
			if (get_order(size) < MAX_ORDER) {
5228
				table = alloc_pages_exact(size, GFP_ATOMIC);
5229 5230
				kmemleak_alloc(table, size, 1, GFP_ATOMIC);
			}
L
Linus Torvalds 已提交
5231 5232 5233 5234 5235 5236
		}
	} while (!table && size > PAGE_SIZE && --log2qty);

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

5237
	printk(KERN_INFO "%s hash table entries: %ld (order: %d, %lu bytes)\n",
L
Linus Torvalds 已提交
5238
	       tablename,
5239
	       (1UL << log2qty),
5240
	       ilog2(size) - PAGE_SHIFT,
L
Linus Torvalds 已提交
5241 5242 5243 5244 5245 5246 5247 5248 5249
	       size);

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

	return table;
}
5250

5251 5252 5253 5254 5255 5256 5257 5258 5259 5260 5261 5262 5263 5264 5265
/* 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);
5266
	return (pfn >> pageblock_order) * NR_PAGEBLOCK_BITS;
5267 5268
#else
	pfn = pfn - zone->zone_start_pfn;
5269
	return (pfn >> pageblock_order) * NR_PAGEBLOCK_BITS;
5270 5271 5272 5273
#endif /* CONFIG_SPARSEMEM */
}

/**
5274
 * get_pageblock_flags_group - Return the requested group of flags for the pageblock_nr_pages block of pages
5275 5276 5277 5278 5279 5280 5281 5282 5283 5284 5285 5286 5287 5288 5289 5290 5291 5292 5293 5294 5295 5296
 * @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;
5297

5298 5299 5300 5301
	return flags;
}

/**
5302
 * set_pageblock_flags_group - Set the requested group of flags for a pageblock_nr_pages block of pages
5303 5304 5305 5306 5307 5308 5309 5310 5311 5312 5313 5314 5315 5316 5317 5318 5319
 * @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);
5320 5321
	VM_BUG_ON(pfn < zone->zone_start_pfn);
	VM_BUG_ON(pfn >= zone->zone_start_pfn + zone->spanned_pages);
5322 5323 5324 5325 5326 5327 5328

	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
5329 5330 5331 5332 5333 5334 5335

/*
 * 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.
 */

5336 5337 5338 5339 5340 5341 5342 5343 5344 5345 5346 5347 5348 5349 5350 5351 5352 5353 5354 5355 5356 5357 5358 5359 5360 5361 5362 5363 5364 5365 5366 5367 5368 5369 5370 5371 5372 5373 5374 5375 5376 5377 5378 5379 5380 5381 5382 5383 5384 5385 5386 5387 5388 5389 5390
static int
__count_immobile_pages(struct zone *zone, struct page *page, int count)
{
	unsigned long pfn, iter, found;
	/*
	 * For avoiding noise data, lru_add_drain_all() should be called
	 * If ZONE_MOVABLE, the zone never contains immobile pages
	 */
	if (zone_idx(zone) == ZONE_MOVABLE)
		return true;

	if (get_pageblock_migratetype(page) == MIGRATE_MOVABLE)
		return true;

	pfn = page_to_pfn(page);
	for (found = 0, iter = 0; iter < pageblock_nr_pages; iter++) {
		unsigned long check = pfn + iter;

		if (!pfn_valid_within(check)) {
			iter++;
			continue;
		}
		page = pfn_to_page(check);
		if (!page_count(page)) {
			if (PageBuddy(page))
				iter += (1 << page_order(page)) - 1;
			continue;
		}
		if (!PageLRU(page))
			found++;
		/*
		 * If there are RECLAIMABLE pages, we need to check it.
		 * But now, memory offline itself doesn't call shrink_slab()
		 * and it still to be fixed.
		 */
		/*
		 * If the page is not RAM, page_count()should be 0.
		 * we don't need more check. This is an _used_ not-movable page.
		 *
		 * The problematic thing here is PG_reserved pages. PG_reserved
		 * is set to both of a memory hole page and a _used_ kernel
		 * page at boot.
		 */
		if (found > count)
			return false;
	}
	return true;
}

bool is_pageblock_removable_nolock(struct page *page)
{
	struct zone *zone = page_zone(page);
	return __count_immobile_pages(zone, page, 0);
}

K
KAMEZAWA Hiroyuki 已提交
5391 5392 5393
int set_migratetype_isolate(struct page *page)
{
	struct zone *zone;
5394
	unsigned long flags, pfn;
5395 5396
	struct memory_isolate_notify arg;
	int notifier_ret;
K
KAMEZAWA Hiroyuki 已提交
5397
	int ret = -EBUSY;
5398
	int zone_idx;
K
KAMEZAWA Hiroyuki 已提交
5399 5400

	zone = page_zone(page);
5401
	zone_idx = zone_idx(zone);
5402

K
KAMEZAWA Hiroyuki 已提交
5403
	spin_lock_irqsave(&zone->lock, flags);
5404 5405 5406 5407 5408 5409

	pfn = page_to_pfn(page);
	arg.start_pfn = pfn;
	arg.nr_pages = pageblock_nr_pages;
	arg.pages_found = 0;

K
KAMEZAWA Hiroyuki 已提交
5410
	/*
5411 5412 5413 5414 5415 5416 5417 5418 5419
	 * It may be possible to isolate a pageblock even if the
	 * migratetype is not MIGRATE_MOVABLE. The memory isolation
	 * notifier chain is used by balloon drivers to return the
	 * number of pages in a range that are held by the balloon
	 * driver to shrink memory. If all the pages are accounted for
	 * by balloons, are free, or on the LRU, isolation can continue.
	 * Later, for example, when memory hotplug notifier runs, these
	 * pages reported as "can be isolated" should be isolated(freed)
	 * by the balloon driver through the memory notifier chain.
K
KAMEZAWA Hiroyuki 已提交
5420
	 */
5421 5422
	notifier_ret = memory_isolate_notify(MEM_ISOLATE_COUNT, &arg);
	notifier_ret = notifier_to_errno(notifier_ret);
5423
	if (notifier_ret)
K
KAMEZAWA Hiroyuki 已提交
5424
		goto out;
5425 5426 5427 5428 5429
	/*
	 * FIXME: Now, memory hotplug doesn't call shrink_slab() by itself.
	 * We just check MOVABLE pages.
	 */
	if (__count_immobile_pages(zone, page, arg.pages_found))
5430 5431
		ret = 0;

5432 5433 5434 5435 5436
	/*
	 * immobile means "not-on-lru" paes. If immobile is larger than
	 * removable-by-driver pages reported by notifier, we'll fail.
	 */

K
KAMEZAWA Hiroyuki 已提交
5437
out:
5438 5439 5440 5441 5442
	if (!ret) {
		set_pageblock_migratetype(page, MIGRATE_ISOLATE);
		move_freepages_block(zone, page, MIGRATE_ISOLATE);
	}

K
KAMEZAWA Hiroyuki 已提交
5443 5444
	spin_unlock_irqrestore(&zone->lock, flags);
	if (!ret)
5445
		drain_all_pages();
K
KAMEZAWA Hiroyuki 已提交
5446 5447 5448 5449 5450 5451 5452 5453 5454 5455 5456 5457 5458 5459 5460 5461
	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
5462 5463 5464 5465 5466 5467 5468 5469 5470 5471 5472 5473 5474 5475 5476 5477 5478 5479 5480 5481 5482 5483 5484 5485 5486 5487 5488 5489 5490 5491 5492 5493 5494 5495 5496 5497 5498 5499 5500 5501 5502 5503 5504 5505 5506 5507 5508

#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
5509 5510 5511 5512 5513 5514 5515 5516 5517 5518 5519 5520 5521 5522 5523 5524 5525 5526 5527 5528 5529

#ifdef CONFIG_MEMORY_FAILURE
bool is_free_buddy_page(struct page *page)
{
	struct zone *zone = page_zone(page);
	unsigned long pfn = page_to_pfn(page);
	unsigned long flags;
	int order;

	spin_lock_irqsave(&zone->lock, flags);
	for (order = 0; order < MAX_ORDER; order++) {
		struct page *page_head = page - (pfn & ((1 << order) - 1));

		if (PageBuddy(page_head) && page_order(page_head) >= order)
			break;
	}
	spin_unlock_irqrestore(&zone->lock, flags);

	return order < MAX_ORDER;
}
#endif
5530 5531 5532 5533 5534 5535 5536 5537 5538 5539 5540 5541 5542 5543 5544 5545 5546 5547 5548 5549 5550 5551 5552 5553 5554 5555 5556 5557 5558 5559 5560 5561 5562 5563 5564 5565 5566 5567 5568 5569 5570 5571 5572 5573 5574 5575 5576 5577 5578 5579 5580 5581 5582 5583 5584 5585 5586 5587 5588 5589 5590 5591 5592 5593 5594 5595 5596 5597 5598 5599 5600 5601 5602 5603 5604 5605 5606

static struct trace_print_flags pageflag_names[] = {
	{1UL << PG_locked,		"locked"	},
	{1UL << PG_error,		"error"		},
	{1UL << PG_referenced,		"referenced"	},
	{1UL << PG_uptodate,		"uptodate"	},
	{1UL << PG_dirty,		"dirty"		},
	{1UL << PG_lru,			"lru"		},
	{1UL << PG_active,		"active"	},
	{1UL << PG_slab,		"slab"		},
	{1UL << PG_owner_priv_1,	"owner_priv_1"	},
	{1UL << PG_arch_1,		"arch_1"	},
	{1UL << PG_reserved,		"reserved"	},
	{1UL << PG_private,		"private"	},
	{1UL << PG_private_2,		"private_2"	},
	{1UL << PG_writeback,		"writeback"	},
#ifdef CONFIG_PAGEFLAGS_EXTENDED
	{1UL << PG_head,		"head"		},
	{1UL << PG_tail,		"tail"		},
#else
	{1UL << PG_compound,		"compound"	},
#endif
	{1UL << PG_swapcache,		"swapcache"	},
	{1UL << PG_mappedtodisk,	"mappedtodisk"	},
	{1UL << PG_reclaim,		"reclaim"	},
	{1UL << PG_buddy,		"buddy"		},
	{1UL << PG_swapbacked,		"swapbacked"	},
	{1UL << PG_unevictable,		"unevictable"	},
#ifdef CONFIG_MMU
	{1UL << PG_mlocked,		"mlocked"	},
#endif
#ifdef CONFIG_ARCH_USES_PG_UNCACHED
	{1UL << PG_uncached,		"uncached"	},
#endif
#ifdef CONFIG_MEMORY_FAILURE
	{1UL << PG_hwpoison,		"hwpoison"	},
#endif
	{-1UL,				NULL		},
};

static void dump_page_flags(unsigned long flags)
{
	const char *delim = "";
	unsigned long mask;
	int i;

	printk(KERN_ALERT "page flags: %#lx(", flags);

	/* remove zone id */
	flags &= (1UL << NR_PAGEFLAGS) - 1;

	for (i = 0; pageflag_names[i].name && flags; i++) {

		mask = pageflag_names[i].mask;
		if ((flags & mask) != mask)
			continue;

		flags &= ~mask;
		printk("%s%s", delim, pageflag_names[i].name);
		delim = "|";
	}

	/* check for left over flags */
	if (flags)
		printk("%s%#lx", delim, flags);

	printk(")\n");
}

void dump_page(struct page *page)
{
	printk(KERN_ALERT
	       "page:%p count:%d mapcount:%d mapping:%p index:%#lx\n",
		page, page_count(page), page_mapcount(page),
		page->mapping, page->index);
	dump_page_flags(page->flags);
}