bootmem.c 20.9 KB
Newer Older
L
Linus Torvalds 已提交
1
/*
2
 *  bootmem - A boot-time physical memory allocator and configurator
L
Linus Torvalds 已提交
3 4
 *
 *  Copyright (C) 1999 Ingo Molnar
5 6
 *                1999 Kanoj Sarcar, SGI
 *                2008 Johannes Weiner
L
Linus Torvalds 已提交
7
 *
8 9
 * Access to this subsystem has to be serialized externally (which is true
 * for the boot process anyway).
L
Linus Torvalds 已提交
10 11
 */
#include <linux/init.h>
12
#include <linux/pfn.h>
13
#include <linux/slab.h>
L
Linus Torvalds 已提交
14 15
#include <linux/bootmem.h>
#include <linux/module.h>
16
#include <linux/kmemleak.h>
17
#include <linux/range.h>
18
#include <linux/memblock.h>
19 20

#include <asm/bug.h>
L
Linus Torvalds 已提交
21
#include <asm/io.h>
22
#include <asm/processor.h>
23

L
Linus Torvalds 已提交
24 25 26 27 28 29
#include "internal.h"

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

30 31 32 33 34 35 36 37
#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

38 39
bootmem_data_t bootmem_node_data[MAX_NUMNODES] __initdata;

40 41
static struct list_head bdata_list __initdata = LIST_HEAD_INIT(bdata_list);

42 43 44 45 46 47 48 49 50 51 52 53 54
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,		\
55
			__func__, ## args);		\
56 57
})

58
static unsigned long __init bootmap_bytes(unsigned long pages)
59
{
60
	unsigned long bytes = (pages + 7) / 8;
61

62
	return ALIGN(bytes, sizeof(long));
63 64
}

65 66 67 68
/**
 * bootmem_bootmap_pages - calculate bitmap size in pages
 * @pages: number of pages the bitmap has to represent
 */
69
unsigned long __init bootmem_bootmap_pages(unsigned long pages)
L
Linus Torvalds 已提交
70
{
71
	unsigned long bytes = bootmap_bytes(pages);
L
Linus Torvalds 已提交
72

73
	return PAGE_ALIGN(bytes) >> PAGE_SHIFT;
L
Linus Torvalds 已提交
74
}
75

76 77 78
/*
 * link bdata in order
 */
79
static void __init link_bootmem(bootmem_data_t *bdata)
80
{
81
	struct list_head *iter;
82

83 84 85 86
	list_for_each(iter, &bdata_list) {
		bootmem_data_t *ent;

		ent = list_entry(iter, bootmem_data_t, list);
87
		if (bdata->node_min_pfn < ent->node_min_pfn)
88
			break;
89
	}
90
	list_add_tail(&bdata->list, iter);
91 92
}

L
Linus Torvalds 已提交
93 94 95
/*
 * Called once to set up the allocator itself.
 */
96
static unsigned long __init init_bootmem_core(bootmem_data_t *bdata,
L
Linus Torvalds 已提交
97 98
	unsigned long mapstart, unsigned long start, unsigned long end)
{
99
	unsigned long mapsize;
L
Linus Torvalds 已提交
100

101
	mminit_validate_memmodel_limits(&start, &end);
102
	bdata->node_bootmem_map = phys_to_virt(PFN_PHYS(mapstart));
103
	bdata->node_min_pfn = start;
L
Linus Torvalds 已提交
104
	bdata->node_low_pfn = end;
105
	link_bootmem(bdata);
L
Linus Torvalds 已提交
106 107 108 109 110

	/*
	 * Initially all pages are reserved - setup_arch() has to
	 * register free RAM areas explicitly.
	 */
111
	mapsize = bootmap_bytes(end - start);
L
Linus Torvalds 已提交
112 113
	memset(bdata->node_bootmem_map, 0xff, mapsize);

114 115 116
	bdebug("nid=%td start=%lx map=%lx end=%lx mapsize=%lx\n",
		bdata - bootmem_node_data, start, mapstart, end, mapsize);

L
Linus Torvalds 已提交
117 118 119
	return mapsize;
}

120 121 122 123 124 125 126 127 128
/**
 * 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.
 */
129 130 131 132 133 134
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);
}

135 136 137 138 139 140 141
/**
 * 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.
 */
142 143 144 145 146 147
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);
}
148

F
FUJITA Tomonori 已提交
149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172
/*
 * free_bootmem_late - free bootmem pages directly to page allocator
 * @addr: starting address of the range
 * @size: size of the range in bytes
 *
 * This is only useful when the bootmem allocator has already been torn
 * down, but we are still initializing the system.  Pages are given directly
 * to the page allocator, no bootmem metadata is updated because it is gone.
 */
void __init free_bootmem_late(unsigned long addr, unsigned long size)
{
	unsigned long cursor, end;

	kmemleak_free_part(__va(addr), size);

	cursor = PFN_UP(addr);
	end = PFN_DOWN(addr + size);

	for (; cursor < end; cursor++) {
		__free_pages_bootmem(pfn_to_page(cursor), 0);
		totalram_pages++;
	}
}

173 174
static unsigned long __init free_all_bootmem_core(bootmem_data_t *bdata)
{
175
	int aligned;
176
	struct page *page;
177 178 179 180 181
	unsigned long start, end, pages, count = 0;

	if (!bdata->node_bootmem_map)
		return 0;

182
	start = bdata->node_min_pfn;
183 184
	end = bdata->node_low_pfn;

185
	/*
186 187
	 * If the start is aligned to the machines wordsize, we might
	 * be able to free pages in bulks of that order.
188
	 */
189
	aligned = !(start & (BITS_PER_LONG - 1));
190

191 192
	bdebug("nid=%td start=%lx end=%lx aligned=%d\n",
		bdata - bootmem_node_data, start, end, aligned);
193

194 195
	while (start < end) {
		unsigned long *map, idx, vec;
196

197
		map = bdata->node_bootmem_map;
198
		idx = start - bdata->node_min_pfn;
199 200 201 202 203 204
		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);
205
			count += BITS_PER_LONG;
206 207 208 209 210 211
		} else {
			unsigned long off = 0;

			while (vec && off < BITS_PER_LONG) {
				if (vec & 1) {
					page = pfn_to_page(start + off);
212
					__free_pages_bootmem(page, 0);
213
					count++;
214
				}
215 216
				vec >>= 1;
				off++;
217 218
			}
		}
219
		start += BITS_PER_LONG;
220 221 222
	}

	page = virt_to_page(bdata->node_bootmem_map);
223
	pages = bdata->node_low_pfn - bdata->node_min_pfn;
224 225 226 227
	pages = bootmem_bootmap_pages(pages);
	count += pages;
	while (pages--)
		__free_pages_bootmem(page++, 0);
228

229 230
	bdebug("nid=%td released=%lx\n", bdata - bootmem_node_data, count);

231 232 233
	return count;
}

234 235 236 237 238 239
/**
 * 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.
 */
240 241 242 243 244 245
unsigned long __init free_all_bootmem_node(pg_data_t *pgdat)
{
	register_page_bootmem_info_node(pgdat);
	return free_all_bootmem_core(pgdat->bdata);
}

246 247 248 249 250
/**
 * free_all_bootmem - release free pages to the buddy allocator
 *
 * Returns the number of pages actually released.
 */
251 252
unsigned long __init free_all_bootmem(void)
{
253 254 255 256 257 258 259
	unsigned long total_pages = 0;
	bootmem_data_t *bdata;

	list_for_each_entry(bdata, &bdata_list, list)
		total_pages += free_all_bootmem_core(bdata);

	return total_pages;
260 261
}

J
Johannes Weiner 已提交
262 263 264 265 266 267
static void __init __free(bootmem_data_t *bdata,
			unsigned long sidx, unsigned long eidx)
{
	unsigned long idx;

	bdebug("nid=%td start=%lx end=%lx\n", bdata - bootmem_node_data,
268 269
		sidx + bdata->node_min_pfn,
		eidx + bdata->node_min_pfn);
J
Johannes Weiner 已提交
270

271 272 273
	if (bdata->hint_idx > sidx)
		bdata->hint_idx = sidx;

J
Johannes Weiner 已提交
274 275 276 277 278 279 280 281 282 283 284 285 286
	for (idx = sidx; idx < eidx; idx++)
		if (!test_and_clear_bit(idx, bdata->node_bootmem_map))
			BUG();
}

static int __init __reserve(bootmem_data_t *bdata, unsigned long sidx,
			unsigned long eidx, int flags)
{
	unsigned long idx;
	int exclusive = flags & BOOTMEM_EXCLUSIVE;

	bdebug("nid=%td start=%lx end=%lx flags=%x\n",
		bdata - bootmem_node_data,
287 288
		sidx + bdata->node_min_pfn,
		eidx + bdata->node_min_pfn,
J
Johannes Weiner 已提交
289 290 291 292 293 294 295 296 297
		flags);

	for (idx = sidx; idx < eidx; idx++)
		if (test_and_set_bit(idx, bdata->node_bootmem_map)) {
			if (exclusive) {
				__free(bdata, sidx, idx);
				return -EBUSY;
			}
			bdebug("silent double reserve of PFN %lx\n",
298
				idx + bdata->node_min_pfn);
J
Johannes Weiner 已提交
299 300 301 302
		}
	return 0;
}

303 304 305
static int __init mark_bootmem_node(bootmem_data_t *bdata,
				unsigned long start, unsigned long end,
				int reserve, int flags)
306 307 308
{
	unsigned long sidx, eidx;

309 310
	bdebug("nid=%td start=%lx end=%lx reserve=%d flags=%x\n",
		bdata - bootmem_node_data, start, end, reserve, flags);
311

312
	BUG_ON(start < bdata->node_min_pfn);
313
	BUG_ON(end > bdata->node_low_pfn);
314

315 316
	sidx = start - bdata->node_min_pfn;
	eidx = end - bdata->node_min_pfn;
317

318 319
	if (reserve)
		return __reserve(bdata, sidx, eidx, flags);
320
	else
321 322 323 324 325 326 327 328 329 330 331 332 333 334 335
		__free(bdata, sidx, eidx);
	return 0;
}

static int __init mark_bootmem(unsigned long start, unsigned long end,
				int reserve, int flags)
{
	unsigned long pos;
	bootmem_data_t *bdata;

	pos = start;
	list_for_each_entry(bdata, &bdata_list, list) {
		int err;
		unsigned long max;

336 337
		if (pos < bdata->node_min_pfn ||
		    pos >= bdata->node_low_pfn) {
338 339 340 341 342
			BUG_ON(pos != start);
			continue;
		}

		max = min(bdata->node_low_pfn, end);
343

344 345 346 347 348
		err = mark_bootmem_node(bdata, pos, max, reserve, flags);
		if (reserve && err) {
			mark_bootmem(start, pos, 0, 0);
			return err;
		}
349

350 351 352 353 354
		if (max == end)
			return 0;
		pos = bdata->node_low_pfn;
	}
	BUG();
355 356
}

357 358 359 360 361 362 363 364
/**
 * 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.
 *
365
 * The range must reside completely on the specified node.
366
 */
367 368 369
void __init free_bootmem_node(pg_data_t *pgdat, unsigned long physaddr,
			      unsigned long size)
{
370 371
	unsigned long start, end;

372 373
	kmemleak_free_part(__va(physaddr), size);

374 375 376 377
	start = PFN_UP(physaddr);
	end = PFN_DOWN(physaddr + size);

	mark_bootmem_node(pgdat->bdata, start, end, 0, 0);
378 379
}

380 381 382 383 384 385 386
/**
 * 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.
 *
387
 * The range must be contiguous but may span node boundaries.
388
 */
389 390
void __init free_bootmem(unsigned long addr, unsigned long size)
{
391
	unsigned long start, end;
392

393 394
	kmemleak_free_part(__va(addr), size);

395 396
	start = PFN_UP(addr);
	end = PFN_DOWN(addr + size);
L
Linus Torvalds 已提交
397

398
	mark_bootmem(start, end, 0, 0);
L
Linus Torvalds 已提交
399 400
}

401 402 403 404 405 406 407 408 409
/**
 * 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.
 *
410
 * The range must reside completely on the specified node.
411
 */
412 413
int __init reserve_bootmem_node(pg_data_t *pgdat, unsigned long physaddr,
				 unsigned long size, int flags)
L
Linus Torvalds 已提交
414
{
415
	unsigned long start, end;
L
Linus Torvalds 已提交
416

417 418 419 420
	start = PFN_DOWN(physaddr);
	end = PFN_UP(physaddr + size);

	return mark_bootmem_node(pgdat->bdata, start, end, 1, flags);
421
}
422

423 424 425 426 427 428 429 430
/**
 * 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.
 *
431
 * The range must be contiguous but may span node boundaries.
432
 */
433 434 435
int __init reserve_bootmem(unsigned long addr, unsigned long size,
			    int flags)
{
436
	unsigned long start, end;
L
Linus Torvalds 已提交
437

438 439
	start = PFN_DOWN(addr);
	end = PFN_UP(addr + size);
440

441
	return mark_bootmem(start, end, 1, flags);
L
Linus Torvalds 已提交
442 443
}

444 445 446 447 448 449
int __weak __init reserve_bootmem_generic(unsigned long phys, unsigned long len,
				   int flags)
{
	return reserve_bootmem(phys, len, flags);
}

450 451
static unsigned long __init align_idx(struct bootmem_data *bdata,
				      unsigned long idx, unsigned long step)
452 453 454 455 456 457 458 459 460 461 462
{
	unsigned long base = bdata->node_min_pfn;

	/*
	 * Align the index with respect to the node start so that the
	 * combination of both satisfies the requested alignment.
	 */

	return ALIGN(base + idx, step) - base;
}

463 464
static unsigned long __init align_off(struct bootmem_data *bdata,
				      unsigned long off, unsigned long align)
465 466 467 468 469 470 471 472
{
	unsigned long base = PFN_PHYS(bdata->node_min_pfn);

	/* Same as align_idx for byte offsets */

	return ALIGN(base + off, align) - base;
}

473 474 475
static void * __init alloc_bootmem_core(struct bootmem_data *bdata,
					unsigned long size, unsigned long align,
					unsigned long goal, unsigned long limit)
L
Linus Torvalds 已提交
476
{
477
	unsigned long fallback = 0;
478 479
	unsigned long min, max, start, sidx, midx, step;

480 481 482 483
	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);

484 485 486
	BUG_ON(!size);
	BUG_ON(align & (align - 1));
	BUG_ON(limit && goal + size > limit);
L
Linus Torvalds 已提交
487

488 489 490
	if (!bdata->node_bootmem_map)
		return NULL;

491
	min = bdata->node_min_pfn;
492
	max = bdata->node_low_pfn;
Y
Yinghai Lu 已提交
493

494 495 496 497 498 499
	goal >>= PAGE_SHIFT;
	limit >>= PAGE_SHIFT;

	if (limit && max > limit)
		max = limit;
	if (max <= min)
Y
Yinghai Lu 已提交
500 501
		return NULL;

502
	step = max(align >> PAGE_SHIFT, 1UL);
503

504 505 506 507
	if (goal && min < goal && goal < max)
		start = ALIGN(goal, step);
	else
		start = ALIGN(min, step);
L
Linus Torvalds 已提交
508

509
	sidx = start - bdata->node_min_pfn;
510
	midx = max - bdata->node_min_pfn;
L
Linus Torvalds 已提交
511

512
	if (bdata->hint_idx > sidx) {
513 514 515 516 517
		/*
		 * Handle the valid case of sidx being zero and still
		 * catch the fallback below.
		 */
		fallback = sidx + 1;
518
		sidx = align_idx(bdata, bdata->hint_idx, step);
519
	}
L
Linus Torvalds 已提交
520

521 522 523 524 525 526
	while (1) {
		int merge;
		void *region;
		unsigned long eidx, i, start_off, end_off;
find_block:
		sidx = find_next_zero_bit(bdata->node_bootmem_map, midx, sidx);
527
		sidx = align_idx(bdata, sidx, step);
528
		eidx = sidx + PFN_UP(size);
529

530
		if (sidx >= midx || eidx > midx)
531
			break;
L
Linus Torvalds 已提交
532

533 534
		for (i = sidx; i < eidx; i++)
			if (test_bit(i, bdata->node_bootmem_map)) {
535
				sidx = align_idx(bdata, i, step);
536 537 538 539
				if (sidx == i)
					sidx += step;
				goto find_block;
			}
L
Linus Torvalds 已提交
540

541
		if (bdata->last_end_off & (PAGE_SIZE - 1) &&
542
				PFN_DOWN(bdata->last_end_off) + 1 == sidx)
543
			start_off = align_off(bdata, bdata->last_end_off, align);
544 545 546 547 548 549 550 551 552 553 554 555
		else
			start_off = PFN_PHYS(sidx);

		merge = PFN_DOWN(start_off) < sidx;
		end_off = start_off + size;

		bdata->last_end_off = end_off;
		bdata->hint_idx = PFN_UP(end_off);

		/*
		 * Reserve the area now:
		 */
J
Johannes Weiner 已提交
556 557 558
		if (__reserve(bdata, PFN_DOWN(start_off) + merge,
				PFN_UP(end_off), BOOTMEM_EXCLUSIVE))
			BUG();
559

560 561
		region = phys_to_virt(PFN_PHYS(bdata->node_min_pfn) +
				start_off);
562
		memset(region, 0, size);
563 564 565 566 567
		/*
		 * The min_count is set to 0 so that bootmem allocated blocks
		 * are never reported as leaks.
		 */
		kmemleak_alloc(region, size, 0, 0);
568
		return region;
L
Linus Torvalds 已提交
569 570
	}

571
	if (fallback) {
572
		sidx = align_idx(bdata, fallback - 1, step);
573 574 575 576 577 578 579
		fallback = 0;
		goto find_block;
	}

	return NULL;
}

580 581 582 583
static void * __init alloc_arch_preferred_bootmem(bootmem_data_t *bdata,
					unsigned long size, unsigned long align,
					unsigned long goal, unsigned long limit)
{
584 585 586
	if (WARN_ON_ONCE(slab_is_available()))
		return kzalloc(size, GFP_NOWAIT);

587
#ifdef CONFIG_HAVE_ARCH_BOOTMEM
588 589 590 591 592 593 594 595 596
	{
		bootmem_data_t *p_bdata;

		p_bdata = bootmem_arch_preferred_node(bdata, size, align,
							goal, limit);
		if (p_bdata)
			return alloc_bootmem_core(p_bdata, size, align,
							goal, limit);
	}
597 598 599 600
#endif
	return NULL;
}

601 602 603 604 605 606
static void * __init ___alloc_bootmem_nopanic(unsigned long size,
					unsigned long align,
					unsigned long goal,
					unsigned long limit)
{
	bootmem_data_t *bdata;
607
	void *region;
608 609

restart:
610 611 612
	region = alloc_arch_preferred_bootmem(NULL, size, align, goal, limit);
	if (region)
		return region;
613

614
	list_for_each_entry(bdata, &bdata_list, list) {
615 616
		if (goal && bdata->node_low_pfn <= PFN_DOWN(goal))
			continue;
617
		if (limit && bdata->node_min_pfn >= PFN_DOWN(limit))
618 619 620 621 622 623 624
			break;

		region = alloc_bootmem_core(bdata, size, align, goal, limit);
		if (region)
			return region;
	}

625 626
	if (goal) {
		goal = 0;
627
		goto restart;
628
	}
629

630
	return NULL;
L
Linus Torvalds 已提交
631 632
}

633 634 635 636 637 638 639 640 641 642 643 644 645
/**
 * __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.
 */
646
void * __init __alloc_bootmem_nopanic(unsigned long size, unsigned long align,
647
					unsigned long goal)
L
Linus Torvalds 已提交
648
{
649 650 651
	unsigned long limit = 0;

	return ___alloc_bootmem_nopanic(size, align, goal, limit);
652
}
L
Linus Torvalds 已提交
653

654 655 656 657 658 659 660 661 662 663 664 665
static void * __init ___alloc_bootmem(unsigned long size, unsigned long align,
					unsigned long goal, unsigned long limit)
{
	void *mem = ___alloc_bootmem_nopanic(size, align, goal, limit);

	if (mem)
		return mem;
	/*
	 * Whoops, we cannot satisfy the allocation request.
	 */
	printk(KERN_ALERT "bootmem alloc of %lu bytes failed!\n", size);
	panic("Out of memory");
666 667
	return NULL;
}
L
Linus Torvalds 已提交
668

669 670 671 672 673 674 675 676 677 678 679 680 681
/**
 * __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.
 */
682 683
void * __init __alloc_bootmem(unsigned long size, unsigned long align,
			      unsigned long goal)
684
{
685 686 687
	unsigned long limit = 0;

	return ___alloc_bootmem(size, align, goal, limit);
L
Linus Torvalds 已提交
688 689
}

690 691 692 693 694 695
static void * __init ___alloc_bootmem_node(bootmem_data_t *bdata,
				unsigned long size, unsigned long align,
				unsigned long goal, unsigned long limit)
{
	void *ptr;

696 697 698 699
	ptr = alloc_arch_preferred_bootmem(bdata, size, align, goal, limit);
	if (ptr)
		return ptr;

700 701 702 703 704 705 706
	ptr = alloc_bootmem_core(bdata, size, align, goal, limit);
	if (ptr)
		return ptr;

	return ___alloc_bootmem(size, align, goal, limit);
}

707 708 709 710 711 712 713 714 715 716 717 718 719 720 721
/**
 * __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.
 */
722 723
void * __init __alloc_bootmem_node(pg_data_t *pgdat, unsigned long size,
				   unsigned long align, unsigned long goal)
L
Linus Torvalds 已提交
724
{
725 726 727
	if (WARN_ON_ONCE(slab_is_available()))
		return kzalloc_node(size, GFP_NOWAIT, pgdat->node_id);

728
	return  ___alloc_bootmem_node(pgdat->bdata, size, align, goal, 0);
729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757
}

void * __init __alloc_bootmem_node_high(pg_data_t *pgdat, unsigned long size,
				   unsigned long align, unsigned long goal)
{
#ifdef MAX_DMA32_PFN
	unsigned long end_pfn;

	if (WARN_ON_ONCE(slab_is_available()))
		return kzalloc_node(size, GFP_NOWAIT, pgdat->node_id);

	/* update goal according ...MAX_DMA32_PFN */
	end_pfn = pgdat->node_start_pfn + pgdat->node_spanned_pages;

	if (end_pfn > MAX_DMA32_PFN + (128 >> (20 - PAGE_SHIFT)) &&
	    (goal >> PAGE_SHIFT) < MAX_DMA32_PFN) {
		void *ptr;
		unsigned long new_goal;

		new_goal = MAX_DMA32_PFN << PAGE_SHIFT;
		ptr = alloc_bootmem_core(pgdat->bdata, size, align,
						 new_goal, 0);
		if (ptr)
			return ptr;
	}
#endif

	return __alloc_bootmem_node(pgdat, size, align, goal);

L
Linus Torvalds 已提交
758 759
}

760
#ifdef CONFIG_SPARSEMEM
761 762 763 764 765 766 767
/**
 * 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.
 */
768 769 770
void * __init alloc_bootmem_section(unsigned long size,
				    unsigned long section_nr)
{
771 772
	bootmem_data_t *bdata;
	unsigned long pfn, goal, limit;
773 774

	pfn = section_nr_to_pfn(section_nr);
775 776 777
	goal = pfn << PAGE_SHIFT;
	limit = section_nr_to_pfn(section_nr + 1) << PAGE_SHIFT;
	bdata = &bootmem_node_data[early_pfn_to_nid(pfn)];
778

779
	return alloc_bootmem_core(bdata, size, SMP_CACHE_BYTES, goal, limit);
780 781 782
}
#endif

783 784 785 786 787
void * __init __alloc_bootmem_node_nopanic(pg_data_t *pgdat, unsigned long size,
				   unsigned long align, unsigned long goal)
{
	void *ptr;

788 789 790
	if (WARN_ON_ONCE(slab_is_available()))
		return kzalloc_node(size, GFP_NOWAIT, pgdat->node_id);

791 792 793 794
	ptr = alloc_arch_preferred_bootmem(pgdat->bdata, size, align, goal, 0);
	if (ptr)
		return ptr;

795 796 797 798 799 800 801
	ptr = alloc_bootmem_core(pgdat->bdata, size, align, goal, 0);
	if (ptr)
		return ptr;

	return __alloc_bootmem_nopanic(size, align, goal);
}

802 803 804
#ifndef ARCH_LOW_ADDRESS_LIMIT
#define ARCH_LOW_ADDRESS_LIMIT	0xffffffffUL
#endif
805

806 807 808 809 810 811 812 813 814 815 816 817 818
/**
 * __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.
 */
819 820
void * __init __alloc_bootmem_low(unsigned long size, unsigned long align,
				  unsigned long goal)
821
{
822
	return ___alloc_bootmem(size, align, goal, ARCH_LOW_ADDRESS_LIMIT);
823 824
}

825 826 827 828 829 830 831 832 833 834 835 836 837 838 839
/**
 * __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.
 */
840 841 842
void * __init __alloc_bootmem_low_node(pg_data_t *pgdat, unsigned long size,
				       unsigned long align, unsigned long goal)
{
843 844 845
	if (WARN_ON_ONCE(slab_is_available()))
		return kzalloc_node(size, GFP_NOWAIT, pgdat->node_id);

846
	return ___alloc_bootmem_node(pgdat->bdata, size, align,
847
				goal, ARCH_LOW_ADDRESS_LIMIT);
848
}