bootmem.c 20.5 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
1
/*
2
 *  bootmem - A boot-time physical memory allocator and configurator
Linus Torvalds's avatar
Linus Torvalds committed
3
4
 *
 *  Copyright (C) 1999 Ingo Molnar
5
6
 *                1999 Kanoj Sarcar, SGI
 *                2008 Johannes Weiner
Linus Torvalds's avatar
Linus Torvalds committed
7
 *
8
9
 * Access to this subsystem has to be serialized externally (which is true
 * for the boot process anyway).
Linus Torvalds's avatar
Linus Torvalds committed
10
11
 */
#include <linux/init.h>
12
#include <linux/pfn.h>
Linus Torvalds's avatar
Linus Torvalds committed
13
14
#include <linux/bootmem.h>
#include <linux/module.h>
15
16

#include <asm/bug.h>
Linus Torvalds's avatar
Linus Torvalds committed
17
#include <asm/io.h>
18
#include <asm/processor.h>
19

Linus Torvalds's avatar
Linus Torvalds committed
20
21
22
23
24
25
#include "internal.h"

unsigned long max_low_pfn;
unsigned long min_low_pfn;
unsigned long max_pfn;

26
27
28
29
30
31
32
33
#ifdef CONFIG_CRASH_DUMP
/*
 * If we have booted due to a crash, max_pfn will be a very low value. We need
 * to know the amount of memory that the previous kernel used.
 */
unsigned long saved_max_pfn;
#endif

34
35
bootmem_data_t bootmem_node_data[MAX_NUMNODES] __initdata;

36
37
static struct list_head bdata_list __initdata = LIST_HEAD_INIT(bdata_list);

38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
static int bootmem_debug;

static int __init bootmem_debug_setup(char *buf)
{
	bootmem_debug = 1;
	return 0;
}
early_param("bootmem_debug", bootmem_debug_setup);

#define bdebug(fmt, args...) ({				\
	if (unlikely(bootmem_debug))			\
		printk(KERN_INFO			\
			"bootmem::%s " fmt,		\
			__FUNCTION__, ## args);		\
})

54
static unsigned long __init bootmap_bytes(unsigned long pages)
55
{
56
	unsigned long bytes = (pages + 7) / 8;
57

58
	return ALIGN(bytes, sizeof(long));
59
60
}

61
62
63
64
/**
 * bootmem_bootmap_pages - calculate bitmap size in pages
 * @pages: number of pages the bitmap has to represent
 */
65
unsigned long __init bootmem_bootmap_pages(unsigned long pages)
Linus Torvalds's avatar
Linus Torvalds committed
66
{
67
	unsigned long bytes = bootmap_bytes(pages);
Linus Torvalds's avatar
Linus Torvalds committed
68

69
	return PAGE_ALIGN(bytes) >> PAGE_SHIFT;
Linus Torvalds's avatar
Linus Torvalds committed
70
}
71

72
73
74
/*
 * link bdata in order
 */
75
static void __init link_bootmem(bootmem_data_t *bdata)
76
{
77
	struct list_head *iter;
78

79
80
81
82
83
84
	list_for_each(iter, &bdata_list) {
		bootmem_data_t *ent;

		ent = list_entry(iter, bootmem_data_t, list);
		if (bdata->node_boot_start < ent->node_boot_start)
			break;
85
	}
86
	list_add_tail(&bdata->list, iter);
87
88
}

Linus Torvalds's avatar
Linus Torvalds committed
89
90
91
/*
 * Called once to set up the allocator itself.
 */
92
static unsigned long __init init_bootmem_core(bootmem_data_t *bdata,
Linus Torvalds's avatar
Linus Torvalds committed
93
94
	unsigned long mapstart, unsigned long start, unsigned long end)
{
95
	unsigned long mapsize;
Linus Torvalds's avatar
Linus Torvalds committed
96

97
	mminit_validate_memmodel_limits(&start, &end);
98
99
	bdata->node_bootmem_map = phys_to_virt(PFN_PHYS(mapstart));
	bdata->node_boot_start = PFN_PHYS(start);
Linus Torvalds's avatar
Linus Torvalds committed
100
	bdata->node_low_pfn = end;
101
	link_bootmem(bdata);
Linus Torvalds's avatar
Linus Torvalds committed
102
103
104
105
106

	/*
	 * Initially all pages are reserved - setup_arch() has to
	 * register free RAM areas explicitly.
	 */
107
	mapsize = bootmap_bytes(end - start);
Linus Torvalds's avatar
Linus Torvalds committed
108
109
	memset(bdata->node_bootmem_map, 0xff, mapsize);

110
111
112
	bdebug("nid=%td start=%lx map=%lx end=%lx mapsize=%lx\n",
		bdata - bootmem_node_data, start, mapstart, end, mapsize);

Linus Torvalds's avatar
Linus Torvalds committed
113
114
115
	return mapsize;
}

116
117
118
119
120
121
122
123
124
/**
 * init_bootmem_node - register a node as boot memory
 * @pgdat: node to register
 * @freepfn: pfn where the bitmap for this node is to be placed
 * @startpfn: first pfn on the node
 * @endpfn: first pfn after the node
 *
 * Returns the number of bytes needed to hold the bitmap for this node.
 */
125
126
127
128
129
130
unsigned long __init init_bootmem_node(pg_data_t *pgdat, unsigned long freepfn,
				unsigned long startpfn, unsigned long endpfn)
{
	return init_bootmem_core(pgdat->bdata, freepfn, startpfn, endpfn);
}

131
132
133
134
135
136
137
/**
 * init_bootmem - register boot memory
 * @start: pfn where the bitmap is to be placed
 * @pages: number of available physical pages
 *
 * Returns the number of bytes needed to hold the bitmap.
 */
138
139
140
141
142
143
144
145
146
147
148
149
unsigned long __init init_bootmem(unsigned long start, unsigned long pages)
{
	max_low_pfn = pages;
	min_low_pfn = start;
	return init_bootmem_core(NODE_DATA(0)->bdata, start, 0, pages);
}

static unsigned long __init free_all_bootmem_core(bootmem_data_t *bdata)
{
	struct page *page;
	unsigned long pfn;
	unsigned long i, count;
150
	unsigned long idx, pages;
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
	unsigned long *map;
	int gofast = 0;

	BUG_ON(!bdata->node_bootmem_map);

	count = 0;
	/* first extant page of the node */
	pfn = PFN_DOWN(bdata->node_boot_start);
	idx = bdata->node_low_pfn - pfn;
	map = bdata->node_bootmem_map;
	/*
	 * Check if we are aligned to BITS_PER_LONG pages.  If so, we might
	 * be able to free page orders of that size at once.
	 */
	if (!(pfn & (BITS_PER_LONG-1)))
		gofast = 1;

	for (i = 0; i < idx; ) {
		unsigned long v = ~map[i / BITS_PER_LONG];

		if (gofast && v == ~0UL) {
			int order;

			page = pfn_to_page(pfn);
			count += BITS_PER_LONG;
			order = ffs(BITS_PER_LONG) - 1;
			__free_pages_bootmem(page, order);
			i += BITS_PER_LONG;
			page += BITS_PER_LONG;
		} else if (v) {
			unsigned long m;

			page = pfn_to_page(pfn);
			for (m = 1; m && i < idx; m<<=1, page++, i++) {
				if (v & m) {
					count++;
					__free_pages_bootmem(page, 0);
				}
			}
		} else {
			i += BITS_PER_LONG;
		}
		pfn += BITS_PER_LONG;
	}

	/*
	 * Now free the allocator bitmap itself, it's not
	 * needed anymore:
	 */
	page = virt_to_page(bdata->node_bootmem_map);
201
202
	pages = bdata->node_low_pfn - PFN_DOWN(bdata->node_boot_start);
	idx = bootmem_bootmap_pages(pages);
203
204
205
206
207
	for (i = 0; i < idx; i++, page++)
		__free_pages_bootmem(page, 0);
	count += i;
	bdata->node_bootmem_map = NULL;

208
209
	bdebug("nid=%td released=%lx\n", bdata - bootmem_node_data, count);

210
211
212
	return count;
}

213
214
215
216
217
218
/**
 * free_all_bootmem_node - release a node's free pages to the buddy allocator
 * @pgdat: node to be released
 *
 * Returns the number of pages actually released.
 */
219
220
221
222
223
224
unsigned long __init free_all_bootmem_node(pg_data_t *pgdat)
{
	register_page_bootmem_info_node(pgdat);
	return free_all_bootmem_core(pgdat->bdata);
}

225
226
227
228
229
/**
 * free_all_bootmem - release free pages to the buddy allocator
 *
 * Returns the number of pages actually released.
 */
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
unsigned long __init free_all_bootmem(void)
{
	return free_all_bootmem_core(NODE_DATA(0)->bdata);
}

static void __init free_bootmem_core(bootmem_data_t *bdata, unsigned long addr,
				     unsigned long size)
{
	unsigned long sidx, eidx;
	unsigned long i;

	BUG_ON(!size);

	/* out range */
	if (addr + size < bdata->node_boot_start ||
		PFN_DOWN(addr) > bdata->node_low_pfn)
		return;
	/*
	 * round down end of usable mem, partially free pages are
	 * considered reserved.
	 */

	if (addr >= bdata->node_boot_start && addr < bdata->last_success)
		bdata->last_success = addr;

	/*
	 * Round up to index to the range.
	 */
	if (PFN_UP(addr) > PFN_DOWN(bdata->node_boot_start))
		sidx = PFN_UP(addr) - PFN_DOWN(bdata->node_boot_start);
	else
		sidx = 0;

	eidx = PFN_DOWN(addr + size - bdata->node_boot_start);
	if (eidx > bdata->node_low_pfn - PFN_DOWN(bdata->node_boot_start))
		eidx = bdata->node_low_pfn - PFN_DOWN(bdata->node_boot_start);

267
268
269
270
	bdebug("nid=%td start=%lx end=%lx\n", bdata - bootmem_node_data,
		sidx + PFN_DOWN(bdata->node_boot_start),
		eidx + PFN_DOWN(bdata->node_boot_start));

271
272
273
274
275
276
	for (i = sidx; i < eidx; i++) {
		if (unlikely(!test_and_clear_bit(i, bdata->node_bootmem_map)))
			BUG();
	}
}

277
278
279
280
281
282
283
284
285
286
/**
 * free_bootmem_node - mark a page range as usable
 * @pgdat: node the range resides on
 * @physaddr: starting address of the range
 * @size: size of the range in bytes
 *
 * Partial pages will be considered reserved and left as they are.
 *
 * Only physical pages that actually reside on @pgdat are marked.
 */
287
288
289
290
291
292
void __init free_bootmem_node(pg_data_t *pgdat, unsigned long physaddr,
			      unsigned long size)
{
	free_bootmem_core(pgdat->bdata, physaddr, size);
}

293
294
295
296
297
298
299
300
301
302
/**
 * free_bootmem - mark a page range as usable
 * @addr: starting address of the range
 * @size: size of the range in bytes
 *
 * Partial pages will be considered reserved and left as they are.
 *
 * All physical pages within the range are marked, no matter what
 * node they reside on.
 */
303
304
305
306
307
308
309
void __init free_bootmem(unsigned long addr, unsigned long size)
{
	bootmem_data_t *bdata;
	list_for_each_entry(bdata, &bdata_list, list)
		free_bootmem_core(bdata, addr, size);
}

Linus Torvalds's avatar
Linus Torvalds committed
310
311
312
313
314
/*
 * Marks a particular physical memory range as unallocatable. Usable RAM
 * might be used for boot-time allocations - or it might get added
 * to the free page pool later on.
 */
315
static int __init can_reserve_bootmem_core(bootmem_data_t *bdata,
316
			unsigned long addr, unsigned long size, int flags)
Linus Torvalds's avatar
Linus Torvalds committed
317
{
318
	unsigned long sidx, eidx;
Linus Torvalds's avatar
Linus Torvalds committed
319
	unsigned long i;
320
321
322
323
324
325
326

	BUG_ON(!size);

	/* out of range, don't hold other */
	if (addr + size < bdata->node_boot_start ||
		PFN_DOWN(addr) > bdata->node_low_pfn)
		return 0;
327

Linus Torvalds's avatar
Linus Torvalds committed
328
	/*
329
	 * Round up to index to the range.
Linus Torvalds's avatar
Linus Torvalds committed
330
	 */
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
	if (addr > bdata->node_boot_start)
		sidx= PFN_DOWN(addr - bdata->node_boot_start);
	else
		sidx = 0;

	eidx = PFN_UP(addr + size - bdata->node_boot_start);
	if (eidx > bdata->node_low_pfn - PFN_DOWN(bdata->node_boot_start))
		eidx = bdata->node_low_pfn - PFN_DOWN(bdata->node_boot_start);

	for (i = sidx; i < eidx; i++) {
		if (test_bit(i, bdata->node_bootmem_map)) {
			if (flags & BOOTMEM_EXCLUSIVE)
				return -EBUSY;
		}
	}

	return 0;

}

static void __init reserve_bootmem_core(bootmem_data_t *bdata,
			unsigned long addr, unsigned long size, int flags)
{
	unsigned long sidx, eidx;
	unsigned long i;

Linus Torvalds's avatar
Linus Torvalds committed
357
	BUG_ON(!size);
358

359
360
361
362
363
364
365
366
367
368
369
370
371
	/* out of range */
	if (addr + size < bdata->node_boot_start ||
		PFN_DOWN(addr) > bdata->node_low_pfn)
		return;

	/*
	 * Round up to index to the range.
	 */
	if (addr > bdata->node_boot_start)
		sidx= PFN_DOWN(addr - bdata->node_boot_start);
	else
		sidx = 0;

372
	eidx = PFN_UP(addr + size - bdata->node_boot_start);
373
374
	if (eidx > bdata->node_low_pfn - PFN_DOWN(bdata->node_boot_start))
		eidx = bdata->node_low_pfn - PFN_DOWN(bdata->node_boot_start);
Linus Torvalds's avatar
Linus Torvalds committed
375

376
377
378
379
380
381
382
383
384
385
	bdebug("nid=%td start=%lx end=%lx flags=%x\n",
		bdata - bootmem_node_data,
		sidx + PFN_DOWN(bdata->node_boot_start),
		eidx + PFN_DOWN(bdata->node_boot_start),
		flags);

	for (i = sidx; i < eidx; i++)
		if (test_and_set_bit(i, bdata->node_bootmem_map))
			bdebug("hm, page %lx reserved twice.\n",
				PFN_DOWN(bdata->node_boot_start) + i);
Linus Torvalds's avatar
Linus Torvalds committed
386
387
}

388
389
390
391
392
393
394
395
396
397
398
/**
 * reserve_bootmem_node - mark a page range as reserved
 * @pgdat: node the range resides on
 * @physaddr: starting address of the range
 * @size: size of the range in bytes
 * @flags: reservation flags (see linux/bootmem.h)
 *
 * Partial pages will be reserved.
 *
 * Only physical pages that actually reside on @pgdat are marked.
 */
399
400
int __init reserve_bootmem_node(pg_data_t *pgdat, unsigned long physaddr,
				 unsigned long size, int flags)
Linus Torvalds's avatar
Linus Torvalds committed
401
{
402
	int ret;
Linus Torvalds's avatar
Linus Torvalds committed
403

404
405
406
407
408
409
	ret = can_reserve_bootmem_core(pgdat->bdata, physaddr, size, flags);
	if (ret < 0)
		return -ENOMEM;
	reserve_bootmem_core(pgdat->bdata, physaddr, size, flags);
	return 0;
}
410

411
#ifndef CONFIG_HAVE_ARCH_BOOTMEM_NODE
412
413
414
415
416
417
418
419
420
421
422
/**
 * reserve_bootmem - mark a page range as usable
 * @addr: starting address of the range
 * @size: size of the range in bytes
 * @flags: reservation flags (see linux/bootmem.h)
 *
 * Partial pages will be reserved.
 *
 * All physical pages within the range are marked, no matter what
 * node they reside on.
 */
423
424
425
426
427
int __init reserve_bootmem(unsigned long addr, unsigned long size,
			    int flags)
{
	bootmem_data_t *bdata;
	int ret;
Linus Torvalds's avatar
Linus Torvalds committed
428

429
430
431
432
	list_for_each_entry(bdata, &bdata_list, list) {
		ret = can_reserve_bootmem_core(bdata, addr, size, flags);
		if (ret < 0)
			return ret;
Linus Torvalds's avatar
Linus Torvalds committed
433
	}
434
435
436
437
	list_for_each_entry(bdata, &bdata_list, list)
		reserve_bootmem_core(bdata, addr, size, flags);

	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
438
}
439
#endif /* !CONFIG_HAVE_ARCH_BOOTMEM_NODE */
Linus Torvalds's avatar
Linus Torvalds committed
440
441
442
443
444
445
446
447
448
449
450
451
452
453

/*
 * We 'merge' subsequent allocations to save space. We might 'lose'
 * some fraction of a page if allocations cannot be satisfied due to
 * size constraints on boxes where there is physical RAM space
 * fragmentation - in these cases (mostly large memory boxes) this
 * is not a problem.
 *
 * On low memory boxes we get it right in 100% of the cases.
 *
 * alignment has to be a power of 2 value.
 *
 * NOTE:  This function is _not_ reentrant.
 */
454
455
456
static void * __init
alloc_bootmem_core(struct bootmem_data *bdata, unsigned long size,
		unsigned long align, unsigned long goal, unsigned long limit)
Linus Torvalds's avatar
Linus Torvalds committed
457
{
458
	unsigned long areasize, preferred;
459
	unsigned long i, start = 0, incr, eidx, end_pfn;
Linus Torvalds's avatar
Linus Torvalds committed
460
	void *ret;
461
462
	unsigned long node_boot_start;
	void *node_bootmem_map;
Linus Torvalds's avatar
Linus Torvalds committed
463

464
	if (!size) {
465
		printk("alloc_bootmem_core(): zero-sized request\n");
Linus Torvalds's avatar
Linus Torvalds committed
466
467
468
469
		BUG();
	}
	BUG_ON(align & (align-1));

470
471
472
473
	/* on nodes without memory - bootmem_map is NULL */
	if (!bdata->node_bootmem_map)
		return NULL;

474
475
476
477
	bdebug("nid=%td size=%lx [%lu pages] align=%lx goal=%lx limit=%lx\n",
		bdata - bootmem_node_data, size, PAGE_ALIGN(size) >> PAGE_SHIFT,
		align, goal, limit);

478
479
480
481
482
483
484
485
486
487
488
489
490
	/* bdata->node_boot_start is supposed to be (12+6)bits alignment on x86_64 ? */
	node_boot_start = bdata->node_boot_start;
	node_bootmem_map = bdata->node_bootmem_map;
	if (align) {
		node_boot_start = ALIGN(bdata->node_boot_start, align);
		if (node_boot_start > bdata->node_boot_start)
			node_bootmem_map = (unsigned long *)bdata->node_bootmem_map +
			    PFN_DOWN(node_boot_start - bdata->node_boot_start)/BITS_PER_LONG;
	}

	if (limit && node_boot_start >= limit)
		return NULL;

491
492
	end_pfn = bdata->node_low_pfn;
	limit = PFN_DOWN(limit);
493
494
495
	if (limit && end_pfn > limit)
		end_pfn = limit;

496
	eidx = end_pfn - PFN_DOWN(node_boot_start);
Linus Torvalds's avatar
Linus Torvalds committed
497
498
499
500
501

	/*
	 * We try to allocate bootmem pages above 'goal'
	 * first, then we try to allocate lower pages.
	 */
502
503
	preferred = 0;
	if (goal && PFN_DOWN(goal) < end_pfn) {
504
505
		if (goal > node_boot_start)
			preferred = goal - node_boot_start;
Linus Torvalds's avatar
Linus Torvalds committed
506

507
508
		if (bdata->last_success > node_boot_start &&
			bdata->last_success - node_boot_start >= preferred)
509
			if (!limit || (limit && limit > bdata->last_success))
510
				preferred = bdata->last_success - node_boot_start;
511
	}
Linus Torvalds's avatar
Linus Torvalds committed
512

513
	preferred = PFN_DOWN(ALIGN(preferred, align));
514
	areasize = (size + PAGE_SIZE-1) / PAGE_SIZE;
Linus Torvalds's avatar
Linus Torvalds committed
515
516
517
	incr = align >> PAGE_SHIFT ? : 1;

restart_scan:
518
	for (i = preferred; i < eidx;) {
Linus Torvalds's avatar
Linus Torvalds committed
519
		unsigned long j;
520

521
		i = find_next_zero_bit(node_bootmem_map, eidx, i);
Linus Torvalds's avatar
Linus Torvalds committed
522
		i = ALIGN(i, incr);
523
524
		if (i >= eidx)
			break;
525
		if (test_bit(i, node_bootmem_map)) {
526
			i += incr;
Linus Torvalds's avatar
Linus Torvalds committed
527
			continue;
528
		}
Linus Torvalds's avatar
Linus Torvalds committed
529
530
531
		for (j = i + 1; j < i + areasize; ++j) {
			if (j >= eidx)
				goto fail_block;
532
			if (test_bit(j, node_bootmem_map))
Linus Torvalds's avatar
Linus Torvalds committed
533
534
535
536
537
538
				goto fail_block;
		}
		start = i;
		goto found;
	fail_block:
		i = ALIGN(j, incr);
539
540
		if (i == j)
			i += incr;
Linus Torvalds's avatar
Linus Torvalds committed
541
542
	}

543
544
	if (preferred > 0) {
		preferred = 0;
Linus Torvalds's avatar
Linus Torvalds committed
545
546
547
548
549
		goto restart_scan;
	}
	return NULL;

found:
550
	bdata->last_success = PFN_PHYS(start) + node_boot_start;
Linus Torvalds's avatar
Linus Torvalds committed
551
552
553
554
555
556
557
558
559
	BUG_ON(start >= eidx);

	/*
	 * Is the next page of the previous allocation-end the start
	 * of this allocation's buffer? If yes then we can 'merge'
	 * the previous partial page with this allocation.
	 */
	if (align < PAGE_SIZE &&
	    bdata->last_offset && bdata->last_pos+1 == start) {
560
		unsigned long offset, remaining_size;
561
		offset = ALIGN(bdata->last_offset, align);
Linus Torvalds's avatar
Linus Torvalds committed
562
		BUG_ON(offset > PAGE_SIZE);
563
		remaining_size = PAGE_SIZE - offset;
Linus Torvalds's avatar
Linus Torvalds committed
564
565
566
		if (size < remaining_size) {
			areasize = 0;
			/* last_pos unchanged */
567
568
			bdata->last_offset = offset + size;
			ret = phys_to_virt(bdata->last_pos * PAGE_SIZE +
569
					   offset + node_boot_start);
Linus Torvalds's avatar
Linus Torvalds committed
570
571
		} else {
			remaining_size = size - remaining_size;
572
573
			areasize = (remaining_size + PAGE_SIZE-1) / PAGE_SIZE;
			ret = phys_to_virt(bdata->last_pos * PAGE_SIZE +
574
					   offset + node_boot_start);
575
			bdata->last_pos = start + areasize - 1;
Linus Torvalds's avatar
Linus Torvalds committed
576
577
578
579
580
581
			bdata->last_offset = remaining_size;
		}
		bdata->last_offset &= ~PAGE_MASK;
	} else {
		bdata->last_pos = start + areasize - 1;
		bdata->last_offset = size & ~PAGE_MASK;
582
		ret = phys_to_virt(start * PAGE_SIZE + node_boot_start);
Linus Torvalds's avatar
Linus Torvalds committed
583
584
	}

585
586
587
588
589
	bdebug("nid=%td start=%lx end=%lx\n",
		bdata - bootmem_node_data,
		start + PFN_DOWN(bdata->node_boot_start),
		start + areasize + PFN_DOWN(bdata->node_boot_start));

Linus Torvalds's avatar
Linus Torvalds committed
590
591
592
	/*
	 * Reserve the area now:
	 */
593
	for (i = start; i < start + areasize; i++)
594
		if (unlikely(test_and_set_bit(i, node_bootmem_map)))
Linus Torvalds's avatar
Linus Torvalds committed
595
596
597
598
599
			BUG();
	memset(ret, 0, size);
	return ret;
}

600
601
602
603
604
605
606
607
608
609
610
611
612
/**
 * __alloc_bootmem_nopanic - allocate boot memory without panicking
 * @size: size of the request in bytes
 * @align: alignment of the region
 * @goal: preferred starting address of the region
 *
 * The goal is dropped if it can not be satisfied and the allocation will
 * fall back to memory below @goal.
 *
 * Allocation may happen on any node in the system.
 *
 * Returns NULL on failure.
 */
613
614
void * __init __alloc_bootmem_nopanic(unsigned long size, unsigned long align,
				      unsigned long goal)
Linus Torvalds's avatar
Linus Torvalds committed
615
{
616
	bootmem_data_t *bdata;
Linus Torvalds's avatar
Linus Torvalds committed
617
618
	void *ptr;

619
	list_for_each_entry(bdata, &bdata_list, list) {
620
		ptr = alloc_bootmem_core(bdata, size, align, goal, 0);
621
622
623
		if (ptr)
			return ptr;
	}
624
625
	return NULL;
}
Linus Torvalds's avatar
Linus Torvalds committed
626

627
628
629
630
631
632
633
634
635
636
637
638
639
/**
 * __alloc_bootmem - allocate boot memory
 * @size: size of the request in bytes
 * @align: alignment of the region
 * @goal: preferred starting address of the region
 *
 * The goal is dropped if it can not be satisfied and the allocation will
 * fall back to memory below @goal.
 *
 * Allocation may happen on any node in the system.
 *
 * The function panics if the request can not be satisfied.
 */
640
641
void * __init __alloc_bootmem(unsigned long size, unsigned long align,
			      unsigned long goal)
642
643
{
	void *mem = __alloc_bootmem_nopanic(size,align,goal);
644

645
646
	if (mem)
		return mem;
Linus Torvalds's avatar
Linus Torvalds committed
647
648
649
650
651
652
653
654
	/*
	 * Whoops, we cannot satisfy the allocation request.
	 */
	printk(KERN_ALERT "bootmem alloc of %lu bytes failed!\n", size);
	panic("Out of memory");
	return NULL;
}

655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
/**
 * __alloc_bootmem_node - allocate boot memory from a specific node
 * @pgdat: node to allocate from
 * @size: size of the request in bytes
 * @align: alignment of the region
 * @goal: preferred starting address of the region
 *
 * The goal is dropped if it can not be satisfied and the allocation will
 * fall back to memory below @goal.
 *
 * Allocation may fall back to any node in the system if the specified node
 * can not hold the requested memory.
 *
 * The function panics if the request can not be satisfied.
 */
670
671
void * __init __alloc_bootmem_node(pg_data_t *pgdat, unsigned long size,
				   unsigned long align, unsigned long goal)
Linus Torvalds's avatar
Linus Torvalds committed
672
673
674
{
	void *ptr;

675
	ptr = alloc_bootmem_core(pgdat->bdata, size, align, goal, 0);
Linus Torvalds's avatar
Linus Torvalds committed
676
	if (ptr)
677
		return ptr;
Linus Torvalds's avatar
Linus Torvalds committed
678

679
	return __alloc_bootmem(size, align, goal);
Linus Torvalds's avatar
Linus Torvalds committed
680
681
}

682
#ifdef CONFIG_SPARSEMEM
683
684
685
686
687
688
689
/**
 * alloc_bootmem_section - allocate boot memory from a specific section
 * @size: size of the request in bytes
 * @section_nr: sparse map section to allocate from
 *
 * Return NULL on failure.
 */
690
691
692
693
694
695
696
697
698
699
700
void * __init alloc_bootmem_section(unsigned long size,
				    unsigned long section_nr)
{
	void *ptr;
	unsigned long limit, goal, start_nr, end_nr, pfn;
	struct pglist_data *pgdat;

	pfn = section_nr_to_pfn(section_nr);
	goal = PFN_PHYS(pfn);
	limit = PFN_PHYS(section_nr_to_pfn(section_nr + 1)) - 1;
	pgdat = NODE_DATA(early_pfn_to_nid(pfn));
701
702
	ptr = alloc_bootmem_core(pgdat->bdata, size, SMP_CACHE_BYTES, goal,
				limit);
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719

	if (!ptr)
		return NULL;

	start_nr = pfn_to_section_nr(PFN_DOWN(__pa(ptr)));
	end_nr = pfn_to_section_nr(PFN_DOWN(__pa(ptr) + size));
	if (start_nr != section_nr || end_nr != section_nr) {
		printk(KERN_WARNING "alloc_bootmem failed on section %ld.\n",
		       section_nr);
		free_bootmem_core(pgdat->bdata, __pa(ptr), size);
		ptr = NULL;
	}

	return ptr;
}
#endif

720
721
722
723
724
725
726
727
728
729
730
731
void * __init __alloc_bootmem_node_nopanic(pg_data_t *pgdat, unsigned long size,
				   unsigned long align, unsigned long goal)
{
	void *ptr;

	ptr = alloc_bootmem_core(pgdat->bdata, size, align, goal, 0);
	if (ptr)
		return ptr;

	return __alloc_bootmem_nopanic(size, align, goal);
}

732
733
734
#ifndef ARCH_LOW_ADDRESS_LIMIT
#define ARCH_LOW_ADDRESS_LIMIT	0xffffffffUL
#endif
735

736
737
738
739
740
741
742
743
744
745
746
747
748
/**
 * __alloc_bootmem_low - allocate low boot memory
 * @size: size of the request in bytes
 * @align: alignment of the region
 * @goal: preferred starting address of the region
 *
 * The goal is dropped if it can not be satisfied and the allocation will
 * fall back to memory below @goal.
 *
 * Allocation may happen on any node in the system.
 *
 * The function panics if the request can not be satisfied.
 */
749
750
void * __init __alloc_bootmem_low(unsigned long size, unsigned long align,
				  unsigned long goal)
751
{
752
	bootmem_data_t *bdata;
753
754
	void *ptr;

755
	list_for_each_entry(bdata, &bdata_list, list) {
756
757
		ptr = alloc_bootmem_core(bdata, size, align, goal,
					ARCH_LOW_ADDRESS_LIMIT);
758
759
760
		if (ptr)
			return ptr;
	}
761
762
763
764
765
766
767
768
769

	/*
	 * Whoops, we cannot satisfy the allocation request.
	 */
	printk(KERN_ALERT "low bootmem alloc of %lu bytes failed!\n", size);
	panic("Out of low memory");
	return NULL;
}

770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
/**
 * __alloc_bootmem_low_node - allocate low boot memory from a specific node
 * @pgdat: node to allocate from
 * @size: size of the request in bytes
 * @align: alignment of the region
 * @goal: preferred starting address of the region
 *
 * The goal is dropped if it can not be satisfied and the allocation will
 * fall back to memory below @goal.
 *
 * Allocation may fall back to any node in the system if the specified node
 * can not hold the requested memory.
 *
 * The function panics if the request can not be satisfied.
 */
785
786
787
void * __init __alloc_bootmem_low_node(pg_data_t *pgdat, unsigned long size,
				       unsigned long align, unsigned long goal)
{
788
789
	return alloc_bootmem_core(pgdat->bdata, size, align, goal,
				ARCH_LOW_ADDRESS_LIMIT);
790
}