bootmem.c 20.4 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
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)
{
147
	int aligned;
148
	struct page *page;
149
150
151
152
153
154
155
156
	unsigned long start, end, pages, count = 0;

	if (!bdata->node_bootmem_map)
		return 0;

	start = PFN_DOWN(bdata->node_boot_start);
	end = bdata->node_low_pfn;

157
	/*
158
159
	 * If the start is aligned to the machines wordsize, we might
	 * be able to free pages in bulks of that order.
160
	 */
161
	aligned = !(start & (BITS_PER_LONG - 1));
162

163
164
	bdebug("nid=%td start=%lx end=%lx aligned=%d\n",
		bdata - bootmem_node_data, start, end, aligned);
165

166
167
	while (start < end) {
		unsigned long *map, idx, vec;
168

169
170
171
172
173
174
175
176
		map = bdata->node_bootmem_map;
		idx = start - PFN_DOWN(bdata->node_boot_start);
		vec = ~map[idx / BITS_PER_LONG];

		if (aligned && vec == ~0UL && start + BITS_PER_LONG < end) {
			int order = ilog2(BITS_PER_LONG);

			__free_pages_bootmem(pfn_to_page(start), order);
177
			count += BITS_PER_LONG;
178
179
180
181
182
183
		} else {
			unsigned long off = 0;

			while (vec && off < BITS_PER_LONG) {
				if (vec & 1) {
					page = pfn_to_page(start + off);
184
					__free_pages_bootmem(page, 0);
185
					count++;
186
				}
187
188
				vec >>= 1;
				off++;
189
190
			}
		}
191
		start += BITS_PER_LONG;
192
193
194
	}

	page = virt_to_page(bdata->node_bootmem_map);
195
	pages = bdata->node_low_pfn - PFN_DOWN(bdata->node_boot_start);
196
197
198
199
	pages = bootmem_bootmap_pages(pages);
	count += pages;
	while (pages--)
		__free_pages_bootmem(page++, 0);
200

201
202
	bdebug("nid=%td released=%lx\n", bdata - bootmem_node_data, count);

203
204
205
	return count;
}

206
207
208
209
210
211
/**
 * 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.
 */
212
213
214
215
216
217
unsigned long __init free_all_bootmem_node(pg_data_t *pgdat)
{
	register_page_bootmem_info_node(pgdat);
	return free_all_bootmem_core(pgdat->bdata);
}

218
219
220
221
222
/**
 * free_all_bootmem - release free pages to the buddy allocator
 *
 * Returns the number of pages actually released.
 */
223
224
225
226
227
228
229
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
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);

260
261
262
263
	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));

264
265
266
267
268
269
	for (i = sidx; i < eidx; i++) {
		if (unlikely(!test_and_clear_bit(i, bdata->node_bootmem_map)))
			BUG();
	}
}

270
271
272
273
274
275
276
277
278
279
/**
 * 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.
 */
280
281
282
283
284
285
void __init free_bootmem_node(pg_data_t *pgdat, unsigned long physaddr,
			      unsigned long size)
{
	free_bootmem_core(pgdat->bdata, physaddr, size);
}

286
287
288
289
290
291
292
293
294
295
/**
 * 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.
 */
296
297
298
299
300
301
302
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
303
304
305
306
307
/*
 * 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.
 */
308
static int __init can_reserve_bootmem_core(bootmem_data_t *bdata,
309
			unsigned long addr, unsigned long size, int flags)
Linus Torvalds's avatar
Linus Torvalds committed
310
{
311
	unsigned long sidx, eidx;
Linus Torvalds's avatar
Linus Torvalds committed
312
	unsigned long i;
313
314
315
316
317
318
319

	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;
320

Linus Torvalds's avatar
Linus Torvalds committed
321
	/*
322
	 * Round up to index to the range.
Linus Torvalds's avatar
Linus Torvalds committed
323
	 */
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
	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
350
	BUG_ON(!size);
351

352
353
354
355
356
357
358
359
360
361
362
363
364
	/* 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;

365
	eidx = PFN_UP(addr + size - bdata->node_boot_start);
366
367
	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
368

369
370
371
372
373
374
375
376
377
378
	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
379
380
}

381
382
383
384
385
386
387
388
389
390
391
/**
 * 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.
 */
392
393
int __init reserve_bootmem_node(pg_data_t *pgdat, unsigned long physaddr,
				 unsigned long size, int flags)
Linus Torvalds's avatar
Linus Torvalds committed
394
{
395
	int ret;
Linus Torvalds's avatar
Linus Torvalds committed
396

397
398
399
400
401
402
	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;
}
403

404
#ifndef CONFIG_HAVE_ARCH_BOOTMEM_NODE
405
406
407
408
409
410
411
412
413
414
415
/**
 * 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.
 */
416
417
418
419
420
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
421

422
423
424
425
	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
426
	}
427
428
429
430
	list_for_each_entry(bdata, &bdata_list, list)
		reserve_bootmem_core(bdata, addr, size, flags);

	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
431
}
432
#endif /* !CONFIG_HAVE_ARCH_BOOTMEM_NODE */
Linus Torvalds's avatar
Linus Torvalds committed
433
434
435
436
437
438
439
440
441
442
443
444
445
446

/*
 * 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.
 */
447
448
449
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
450
{
451
	unsigned long areasize, preferred;
452
	unsigned long i, start = 0, incr, eidx, end_pfn;
Linus Torvalds's avatar
Linus Torvalds committed
453
	void *ret;
454
455
	unsigned long node_boot_start;
	void *node_bootmem_map;
Linus Torvalds's avatar
Linus Torvalds committed
456

457
	if (!size) {
458
		printk("alloc_bootmem_core(): zero-sized request\n");
Linus Torvalds's avatar
Linus Torvalds committed
459
460
461
462
		BUG();
	}
	BUG_ON(align & (align-1));

463
464
465
466
	/* on nodes without memory - bootmem_map is NULL */
	if (!bdata->node_bootmem_map)
		return NULL;

467
468
469
470
	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);

471
472
473
474
475
476
477
478
479
480
481
482
483
	/* 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;

484
485
	end_pfn = bdata->node_low_pfn;
	limit = PFN_DOWN(limit);
486
487
488
	if (limit && end_pfn > limit)
		end_pfn = limit;

489
	eidx = end_pfn - PFN_DOWN(node_boot_start);
Linus Torvalds's avatar
Linus Torvalds committed
490
491
492
493
494

	/*
	 * We try to allocate bootmem pages above 'goal'
	 * first, then we try to allocate lower pages.
	 */
495
496
	preferred = 0;
	if (goal && PFN_DOWN(goal) < end_pfn) {
497
498
		if (goal > node_boot_start)
			preferred = goal - node_boot_start;
Linus Torvalds's avatar
Linus Torvalds committed
499

500
501
		if (bdata->last_success > node_boot_start &&
			bdata->last_success - node_boot_start >= preferred)
502
			if (!limit || (limit && limit > bdata->last_success))
503
				preferred = bdata->last_success - node_boot_start;
504
	}
Linus Torvalds's avatar
Linus Torvalds committed
505

506
	preferred = PFN_DOWN(ALIGN(preferred, align));
507
	areasize = (size + PAGE_SIZE-1) / PAGE_SIZE;
Linus Torvalds's avatar
Linus Torvalds committed
508
509
510
	incr = align >> PAGE_SHIFT ? : 1;

restart_scan:
511
	for (i = preferred; i < eidx;) {
Linus Torvalds's avatar
Linus Torvalds committed
512
		unsigned long j;
513

514
		i = find_next_zero_bit(node_bootmem_map, eidx, i);
Linus Torvalds's avatar
Linus Torvalds committed
515
		i = ALIGN(i, incr);
516
517
		if (i >= eidx)
			break;
518
		if (test_bit(i, node_bootmem_map)) {
519
			i += incr;
Linus Torvalds's avatar
Linus Torvalds committed
520
			continue;
521
		}
Linus Torvalds's avatar
Linus Torvalds committed
522
523
524
		for (j = i + 1; j < i + areasize; ++j) {
			if (j >= eidx)
				goto fail_block;
525
			if (test_bit(j, node_bootmem_map))
Linus Torvalds's avatar
Linus Torvalds committed
526
527
528
529
530
531
				goto fail_block;
		}
		start = i;
		goto found;
	fail_block:
		i = ALIGN(j, incr);
532
533
		if (i == j)
			i += incr;
Linus Torvalds's avatar
Linus Torvalds committed
534
535
	}

536
537
	if (preferred > 0) {
		preferred = 0;
Linus Torvalds's avatar
Linus Torvalds committed
538
539
540
541
542
		goto restart_scan;
	}
	return NULL;

found:
543
	bdata->last_success = PFN_PHYS(start) + node_boot_start;
Linus Torvalds's avatar
Linus Torvalds committed
544
545
546
547
548
549
550
551
552
	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) {
553
		unsigned long offset, remaining_size;
554
		offset = ALIGN(bdata->last_offset, align);
Linus Torvalds's avatar
Linus Torvalds committed
555
		BUG_ON(offset > PAGE_SIZE);
556
		remaining_size = PAGE_SIZE - offset;
Linus Torvalds's avatar
Linus Torvalds committed
557
558
559
		if (size < remaining_size) {
			areasize = 0;
			/* last_pos unchanged */
560
561
			bdata->last_offset = offset + size;
			ret = phys_to_virt(bdata->last_pos * PAGE_SIZE +
562
					   offset + node_boot_start);
Linus Torvalds's avatar
Linus Torvalds committed
563
564
		} else {
			remaining_size = size - remaining_size;
565
566
			areasize = (remaining_size + PAGE_SIZE-1) / PAGE_SIZE;
			ret = phys_to_virt(bdata->last_pos * PAGE_SIZE +
567
					   offset + node_boot_start);
568
			bdata->last_pos = start + areasize - 1;
Linus Torvalds's avatar
Linus Torvalds committed
569
570
571
572
573
574
			bdata->last_offset = remaining_size;
		}
		bdata->last_offset &= ~PAGE_MASK;
	} else {
		bdata->last_pos = start + areasize - 1;
		bdata->last_offset = size & ~PAGE_MASK;
575
		ret = phys_to_virt(start * PAGE_SIZE + node_boot_start);
Linus Torvalds's avatar
Linus Torvalds committed
576
577
	}

578
579
580
581
582
	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
583
584
585
	/*
	 * Reserve the area now:
	 */
586
	for (i = start; i < start + areasize; i++)
587
		if (unlikely(test_and_set_bit(i, node_bootmem_map)))
Linus Torvalds's avatar
Linus Torvalds committed
588
589
590
591
592
			BUG();
	memset(ret, 0, size);
	return ret;
}

593
594
595
596
597
598
599
600
601
602
603
604
605
/**
 * __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.
 */
606
607
void * __init __alloc_bootmem_nopanic(unsigned long size, unsigned long align,
				      unsigned long goal)
Linus Torvalds's avatar
Linus Torvalds committed
608
{
609
	bootmem_data_t *bdata;
Linus Torvalds's avatar
Linus Torvalds committed
610
611
	void *ptr;

612
	list_for_each_entry(bdata, &bdata_list, list) {
613
		ptr = alloc_bootmem_core(bdata, size, align, goal, 0);
614
615
616
		if (ptr)
			return ptr;
	}
617
618
	return NULL;
}
Linus Torvalds's avatar
Linus Torvalds committed
619

620
621
622
623
624
625
626
627
628
629
630
631
632
/**
 * __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.
 */
633
634
void * __init __alloc_bootmem(unsigned long size, unsigned long align,
			      unsigned long goal)
635
636
{
	void *mem = __alloc_bootmem_nopanic(size,align,goal);
637

638
639
	if (mem)
		return mem;
Linus Torvalds's avatar
Linus Torvalds committed
640
641
642
643
644
645
646
647
	/*
	 * Whoops, we cannot satisfy the allocation request.
	 */
	printk(KERN_ALERT "bootmem alloc of %lu bytes failed!\n", size);
	panic("Out of memory");
	return NULL;
}

648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
/**
 * __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.
 */
663
664
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
665
666
667
{
	void *ptr;

668
	ptr = alloc_bootmem_core(pgdat->bdata, size, align, goal, 0);
Linus Torvalds's avatar
Linus Torvalds committed
669
	if (ptr)
670
		return ptr;
Linus Torvalds's avatar
Linus Torvalds committed
671

672
	return __alloc_bootmem(size, align, goal);
Linus Torvalds's avatar
Linus Torvalds committed
673
674
}

675
#ifdef CONFIG_SPARSEMEM
676
677
678
679
680
681
682
/**
 * 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.
 */
683
684
685
686
687
688
689
690
691
692
693
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));
694
695
	ptr = alloc_bootmem_core(pgdat->bdata, size, SMP_CACHE_BYTES, goal,
				limit);
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712

	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

713
714
715
716
717
718
719
720
721
722
723
724
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);
}

725
726
727
#ifndef ARCH_LOW_ADDRESS_LIMIT
#define ARCH_LOW_ADDRESS_LIMIT	0xffffffffUL
#endif
728

729
730
731
732
733
734
735
736
737
738
739
740
741
/**
 * __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.
 */
742
743
void * __init __alloc_bootmem_low(unsigned long size, unsigned long align,
				  unsigned long goal)
744
{
745
	bootmem_data_t *bdata;
746
747
	void *ptr;

748
	list_for_each_entry(bdata, &bdata_list, list) {
749
750
		ptr = alloc_bootmem_core(bdata, size, align, goal,
					ARCH_LOW_ADDRESS_LIMIT);
751
752
753
		if (ptr)
			return ptr;
	}
754
755
756
757
758
759
760
761
762

	/*
	 * 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;
}

763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
/**
 * __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.
 */
778
779
780
void * __init __alloc_bootmem_low_node(pg_data_t *pgdat, unsigned long size,
				       unsigned long align, unsigned long goal)
{
781
782
	return alloc_bootmem_core(pgdat->bdata, size, align, goal,
				ARCH_LOW_ADDRESS_LIMIT);
783
}