memory.c 19.6 KB
Newer Older
1
/*
2
 * Memory subsystem support
3 4 5 6 7 8 9 10 11 12 13 14 15
 *
 * Written by Matt Tolentino <matthew.e.tolentino@intel.com>
 *            Dave Hansen <haveblue@us.ibm.com>
 *
 * This file provides the necessary infrastructure to represent
 * a SPARSEMEM-memory-model system's physical memory in /sysfs.
 * All arch-independent code that assumes MEMORY_HOTPLUG requires
 * SPARSEMEM should be contained here, or in mm/memory_hotplug.c.
 */

#include <linux/module.h>
#include <linux/init.h>
#include <linux/topology.h>
16
#include <linux/capability.h>
17 18 19 20 21
#include <linux/device.h>
#include <linux/memory.h>
#include <linux/kobject.h>
#include <linux/memory_hotplug.h>
#include <linux/mm.h>
22
#include <linux/mutex.h>
23
#include <linux/stat.h>
24
#include <linux/slab.h>
25

Arun Sharma's avatar
Arun Sharma committed
26
#include <linux/atomic.h>
27 28
#include <asm/uaccess.h>

29 30
static DEFINE_MUTEX(mem_sysfs_mutex);

31
#define MEMORY_CLASS_NAME	"memory"
32 33 34 35 36 37 38

static int sections_per_block;

static inline int base_memory_block_id(int section_nr)
{
	return section_nr / sections_per_block;
}
39

40 41 42
static int memory_subsys_online(struct device *dev);
static int memory_subsys_offline(struct device *dev);

43
static struct bus_type memory_subsys = {
44
	.name = MEMORY_CLASS_NAME,
45
	.dev_name = MEMORY_CLASS_NAME,
46 47
	.online = memory_subsys_online,
	.offline = memory_subsys_offline,
48 49
};

50
static BLOCKING_NOTIFIER_HEAD(memory_chain);
51

52
int register_memory_notifier(struct notifier_block *nb)
53
{
54
        return blocking_notifier_chain_register(&memory_chain, nb);
55
}
56
EXPORT_SYMBOL(register_memory_notifier);
57

58
void unregister_memory_notifier(struct notifier_block *nb)
59
{
60
        blocking_notifier_chain_unregister(&memory_chain, nb);
61
}
62
EXPORT_SYMBOL(unregister_memory_notifier);
63

64 65 66 67 68 69 70 71 72 73 74 75 76 77
static ATOMIC_NOTIFIER_HEAD(memory_isolate_chain);

int register_memory_isolate_notifier(struct notifier_block *nb)
{
	return atomic_notifier_chain_register(&memory_isolate_chain, nb);
}
EXPORT_SYMBOL(register_memory_isolate_notifier);

void unregister_memory_isolate_notifier(struct notifier_block *nb)
{
	atomic_notifier_chain_unregister(&memory_isolate_chain, nb);
}
EXPORT_SYMBOL(unregister_memory_isolate_notifier);

78 79 80 81 82 83 84
static void memory_block_release(struct device *dev)
{
	struct memory_block *mem = container_of(dev, struct memory_block, dev);

	kfree(mem);
}

85 86 87
/*
 * register_memory - Setup a sysfs device for a memory block
 */
88
static
89
int register_memory(struct memory_block *memory)
90 91 92
{
	int error;

93 94
	memory->dev.bus = &memory_subsys;
	memory->dev.id = memory->start_section_nr / sections_per_block;
95
	memory->dev.release = memory_block_release;
96
	memory->dev.offline = memory->state == MEM_OFFLINE;
97

98
	error = device_register(&memory->dev);
99 100 101
	return error;
}

102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121
unsigned long __weak memory_block_size_bytes(void)
{
	return MIN_MEMORY_BLOCK_SIZE;
}

static unsigned long get_memory_block_size(void)
{
	unsigned long block_sz;

	block_sz = memory_block_size_bytes();

	/* Validate blk_sz is a power of 2 and not less than section size */
	if ((block_sz & (block_sz - 1)) || (block_sz < MIN_MEMORY_BLOCK_SIZE)) {
		WARN_ON(1);
		block_sz = MIN_MEMORY_BLOCK_SIZE;
	}

	return block_sz;
}

122 123 124 125 126
/*
 * use this as the physical section index that this memsection
 * uses.
 */

127 128
static ssize_t show_mem_start_phys_index(struct device *dev,
			struct device_attribute *attr, char *buf)
129 130
{
	struct memory_block *mem =
131
		container_of(dev, struct memory_block, dev);
132 133 134 135 136 137
	unsigned long phys_index;

	phys_index = mem->start_section_nr / sections_per_block;
	return sprintf(buf, "%08lx\n", phys_index);
}

138 139
static ssize_t show_mem_end_phys_index(struct device *dev,
			struct device_attribute *attr, char *buf)
140 141
{
	struct memory_block *mem =
142
		container_of(dev, struct memory_block, dev);
143 144 145 146
	unsigned long phys_index;

	phys_index = mem->end_section_nr / sections_per_block;
	return sprintf(buf, "%08lx\n", phys_index);
147 148
}

149 150 151
/*
 * Show whether the section of memory is likely to be hot-removable
 */
152 153
static ssize_t show_mem_removable(struct device *dev,
			struct device_attribute *attr, char *buf)
154
{
155 156
	unsigned long i, pfn;
	int ret = 1;
157
	struct memory_block *mem =
158
		container_of(dev, struct memory_block, dev);
159

160
	for (i = 0; i < sections_per_block; i++) {
161
		pfn = section_nr_to_pfn(mem->start_section_nr + i);
162 163 164
		ret &= is_mem_section_removable(pfn, PAGES_PER_SECTION);
	}

165 166 167
	return sprintf(buf, "%d\n", ret);
}

168 169 170
/*
 * online, offline, going offline, etc.
 */
171 172
static ssize_t show_mem_state(struct device *dev,
			struct device_attribute *attr, char *buf)
173 174
{
	struct memory_block *mem =
175
		container_of(dev, struct memory_block, dev);
176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201
	ssize_t len = 0;

	/*
	 * We can probably put these states in a nice little array
	 * so that they're not open-coded
	 */
	switch (mem->state) {
		case MEM_ONLINE:
			len = sprintf(buf, "online\n");
			break;
		case MEM_OFFLINE:
			len = sprintf(buf, "offline\n");
			break;
		case MEM_GOING_OFFLINE:
			len = sprintf(buf, "going-offline\n");
			break;
		default:
			len = sprintf(buf, "ERROR-UNKNOWN-%ld\n",
					mem->state);
			WARN_ON(1);
			break;
	}

	return len;
}

202
int memory_notify(unsigned long val, void *v)
203
{
204
	return blocking_notifier_call_chain(&memory_chain, val, v);
205 206
}

207 208 209 210 211
int memory_isolate_notify(unsigned long val, void *v)
{
	return atomic_notifier_call_chain(&memory_isolate_chain, val, v);
}

212 213 214 215
/*
 * The probe routines leave the pages reserved, just as the bootmem code does.
 * Make sure they're still that way.
 */
216
static bool pages_correctly_reserved(unsigned long start_pfn)
217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246
{
	int i, j;
	struct page *page;
	unsigned long pfn = start_pfn;

	/*
	 * memmap between sections is not contiguous except with
	 * SPARSEMEM_VMEMMAP. We lookup the page once per section
	 * and assume memmap is contiguous within each section
	 */
	for (i = 0; i < sections_per_block; i++, pfn += PAGES_PER_SECTION) {
		if (WARN_ON_ONCE(!pfn_valid(pfn)))
			return false;
		page = pfn_to_page(pfn);

		for (j = 0; j < PAGES_PER_SECTION; j++) {
			if (PageReserved(page + j))
				continue;

			printk(KERN_WARNING "section number %ld page number %d "
				"not reserved, was it already online?\n",
				pfn_to_section_nr(pfn), j);

			return false;
		}
	}

	return true;
}

247 248 249 250 251
/*
 * MEMORY_HOTPLUG depends on SPARSEMEM in mm/Kconfig, so it is
 * OK to have direct references to sparsemem variables in here.
 */
static int
252
memory_block_action(unsigned long phys_index, unsigned long action, int online_type)
253
{
254
	unsigned long start_pfn;
255
	unsigned long nr_pages = PAGES_PER_SECTION * sections_per_block;
256
	struct page *first_page;
257 258
	int ret;

259
	first_page = pfn_to_page(phys_index << PFN_SECTION_SHIFT);
260
	start_pfn = page_to_pfn(first_page);
261

262 263
	switch (action) {
		case MEM_ONLINE:
264
			if (!pages_correctly_reserved(start_pfn))
265 266
				return -EBUSY;

267
			ret = online_pages(start_pfn, nr_pages, online_type);
268 269
			break;
		case MEM_OFFLINE:
270
			ret = offline_pages(start_pfn, nr_pages);
271 272
			break;
		default:
273 274
			WARN(1, KERN_WARNING "%s(%ld, %ld) unknown action: "
			     "%ld\n", __func__, phys_index, action, action);
275 276 277 278 279 280
			ret = -EINVAL;
	}

	return ret;
}

281
static int __memory_block_change_state(struct memory_block *mem,
282 283
		unsigned long to_state, unsigned long from_state_req,
		int online_type)
284
{
285
	int ret = 0;
286

287 288
	if (mem->state != from_state_req)
		return -EINVAL;
289

290 291 292
	if (to_state == MEM_OFFLINE)
		mem->state = MEM_GOING_OFFLINE;

293
	ret = memory_block_action(mem->start_section_nr, to_state, online_type);
294
	if (ret) {
295
		mem->state = from_state_req;
296 297 298 299
	} else {
		mem->state = to_state;
		if (to_state == MEM_ONLINE)
			mem->last_online = online_type;
300
	}
301 302
	return ret;
}
303

304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349
static int memory_subsys_online(struct device *dev)
{
	struct memory_block *mem = container_of(dev, struct memory_block, dev);
	int ret;

	mutex_lock(&mem->state_mutex);

	ret = mem->state == MEM_ONLINE ? 0 :
		__memory_block_change_state(mem, MEM_ONLINE, MEM_OFFLINE,
					    mem->last_online);

	mutex_unlock(&mem->state_mutex);
	return ret;
}

static int memory_subsys_offline(struct device *dev)
{
	struct memory_block *mem = container_of(dev, struct memory_block, dev);
	int ret;

	mutex_lock(&mem->state_mutex);

	ret = mem->state == MEM_OFFLINE ? 0 :
		__memory_block_change_state(mem, MEM_OFFLINE, MEM_ONLINE, -1);

	mutex_unlock(&mem->state_mutex);
	return ret;
}

static int __memory_block_change_state_uevent(struct memory_block *mem,
		unsigned long to_state, unsigned long from_state_req,
		int online_type)
{
	int ret = __memory_block_change_state(mem, to_state, from_state_req,
					      online_type);
	if (!ret) {
		switch (mem->state) {
		case MEM_OFFLINE:
			kobject_uevent(&mem->dev.kobj, KOBJ_OFFLINE);
			break;
		case MEM_ONLINE:
			kobject_uevent(&mem->dev.kobj, KOBJ_ONLINE);
			break;
		default:
			break;
		}
350
	}
351 352 353
	return ret;
}

354
static int memory_block_change_state(struct memory_block *mem,
355 356
		unsigned long to_state, unsigned long from_state_req,
		int online_type)
357 358 359 360
{
	int ret;

	mutex_lock(&mem->state_mutex);
361 362
	ret = __memory_block_change_state_uevent(mem, to_state, from_state_req,
						 online_type);
363 364 365 366
	mutex_unlock(&mem->state_mutex);

	return ret;
}
367
static ssize_t
368 369
store_mem_state(struct device *dev,
		struct device_attribute *attr, const char *buf, size_t count)
370 371
{
	struct memory_block *mem;
372
	bool offline;
373 374
	int ret = -EINVAL;

375
	mem = container_of(dev, struct memory_block, dev);
376

377 378 379 380
	lock_device_hotplug();

	if (!strncmp(buf, "online_kernel", min_t(int, count, 13))) {
		offline = false;
381 382
		ret = memory_block_change_state(mem, MEM_ONLINE,
						MEM_OFFLINE, ONLINE_KERNEL);
383 384
	} else if (!strncmp(buf, "online_movable", min_t(int, count, 14))) {
		offline = false;
385 386
		ret = memory_block_change_state(mem, MEM_ONLINE,
						MEM_OFFLINE, ONLINE_MOVABLE);
387 388
	} else if (!strncmp(buf, "online", min_t(int, count, 6))) {
		offline = false;
389 390
		ret = memory_block_change_state(mem, MEM_ONLINE,
						MEM_OFFLINE, ONLINE_KEEP);
391 392
	} else if(!strncmp(buf, "offline", min_t(int, count, 7))) {
		offline = true;
393 394
		ret = memory_block_change_state(mem, MEM_OFFLINE,
						MEM_ONLINE, -1);
395 396 397 398 399
	}
	if (!ret)
		dev->offline = offline;

	unlock_device_hotplug();
400

401 402 403 404 405 406 407 408 409 410 411 412 413 414
	if (ret)
		return ret;
	return count;
}

/*
 * phys_device is a bad name for this.  What I really want
 * is a way to differentiate between memory ranges that
 * are part of physical devices that constitute
 * a complete removable unit or fru.
 * i.e. do these ranges belong to the same physical device,
 * s.t. if I offline all of these sections I can then
 * remove the physical device?
 */
415 416
static ssize_t show_phys_device(struct device *dev,
				struct device_attribute *attr, char *buf)
417 418
{
	struct memory_block *mem =
419
		container_of(dev, struct memory_block, dev);
420 421 422
	return sprintf(buf, "%d\n", mem->phys_device);
}

423 424 425 426 427
static DEVICE_ATTR(phys_index, 0444, show_mem_start_phys_index, NULL);
static DEVICE_ATTR(end_phys_index, 0444, show_mem_end_phys_index, NULL);
static DEVICE_ATTR(state, 0644, show_mem_state, store_mem_state);
static DEVICE_ATTR(phys_device, 0444, show_phys_device, NULL);
static DEVICE_ATTR(removable, 0444, show_mem_removable, NULL);
428 429

#define mem_create_simple_file(mem, attr_name)	\
430
	device_create_file(&mem->dev, &dev_attr_##attr_name)
431
#define mem_remove_simple_file(mem, attr_name)	\
432
	device_remove_file(&mem->dev, &dev_attr_##attr_name)
433 434 435 436 437

/*
 * Block size attribute stuff
 */
static ssize_t
438
print_block_size(struct device *dev, struct device_attribute *attr,
439
		 char *buf)
440
{
441
	return sprintf(buf, "%lx\n", get_memory_block_size());
442 443
}

444
static DEVICE_ATTR(block_size_bytes, 0444, print_block_size, NULL);
445 446 447

static int block_size_init(void)
{
448 449
	return device_create_file(memory_subsys.dev_root,
				  &dev_attr_block_size_bytes);
450 451 452 453 454 455 456 457 458 459
}

/*
 * Some architectures will have custom drivers to do this, and
 * will not need to do it from userspace.  The fake hot-add code
 * as well as ppc64 will do all of their discovery in userspace
 * and will require this interface.
 */
#ifdef CONFIG_ARCH_MEMORY_PROBE
static ssize_t
460
memory_probe_store(struct device *dev, struct device_attribute *attr,
461
		   const char *buf, size_t count)
462 463
{
	u64 phys_addr;
464
	int nid;
465
	int i, ret;
466
	unsigned long pages_per_block = PAGES_PER_SECTION * sections_per_block;
467 468 469

	phys_addr = simple_strtoull(buf, NULL, 0);

470 471 472
	if (phys_addr & ((pages_per_block << PAGE_SHIFT) - 1))
		return -EINVAL;

473 474 475 476 477
	for (i = 0; i < sections_per_block; i++) {
		nid = memory_add_physaddr_to_nid(phys_addr);
		ret = add_memory(nid, phys_addr,
				 PAGES_PER_SECTION << PAGE_SHIFT);
		if (ret)
478
			goto out;
479 480 481

		phys_addr += MIN_MEMORY_BLOCK_SIZE;
	}
482

483 484 485
	ret = count;
out:
	return ret;
486
}
487
static DEVICE_ATTR(probe, S_IWUSR, NULL, memory_probe_store);
488 489 490

static int memory_probe_init(void)
{
491
	return device_create_file(memory_subsys.dev_root, &dev_attr_probe);
492 493
}
#else
494 495 496 497
static inline int memory_probe_init(void)
{
	return 0;
}
498 499
#endif

500 501 502 503 504 505 506
#ifdef CONFIG_MEMORY_FAILURE
/*
 * Support for offlining pages of memory
 */

/* Soft offline a page */
static ssize_t
507 508
store_soft_offline_page(struct device *dev,
			struct device_attribute *attr,
509
			const char *buf, size_t count)
510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525
{
	int ret;
	u64 pfn;
	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;
	if (strict_strtoull(buf, 0, &pfn) < 0)
		return -EINVAL;
	pfn >>= PAGE_SHIFT;
	if (!pfn_valid(pfn))
		return -ENXIO;
	ret = soft_offline_page(pfn_to_page(pfn), 0);
	return ret == 0 ? count : ret;
}

/* Forcibly offline a page, including killing processes. */
static ssize_t
526 527
store_hard_offline_page(struct device *dev,
			struct device_attribute *attr,
528
			const char *buf, size_t count)
529 530 531 532 533 534 535 536
{
	int ret;
	u64 pfn;
	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;
	if (strict_strtoull(buf, 0, &pfn) < 0)
		return -EINVAL;
	pfn >>= PAGE_SHIFT;
537
	ret = memory_failure(pfn, 0, 0);
538 539 540
	return ret ? ret : count;
}

541 542
static DEVICE_ATTR(soft_offline_page, S_IWUSR, NULL, store_soft_offline_page);
static DEVICE_ATTR(hard_offline_page, S_IWUSR, NULL, store_hard_offline_page);
543 544 545 546 547

static __init int memory_fail_init(void)
{
	int err;

548 549
	err = device_create_file(memory_subsys.dev_root,
				&dev_attr_soft_offline_page);
550
	if (!err)
551 552
		err = device_create_file(memory_subsys.dev_root,
				&dev_attr_hard_offline_page);
553 554 555 556 557 558 559 560 561
	return err;
}
#else
static inline int memory_fail_init(void)
{
	return 0;
}
#endif

562 563 564 565 566
/*
 * Note that phys_device is optional.  It is here to allow for
 * differentiation between which *physical* devices each
 * section belongs to...
 */
567 568 569 570
int __weak arch_get_memory_phys_device(unsigned long start_pfn)
{
	return 0;
}
571

572 573 574 575
/*
 * A reference for the returned object is held and the reference for the
 * hinted object is released.
 */
576 577
struct memory_block *find_memory_block_hinted(struct mem_section *section,
					      struct memory_block *hint)
578
{
579
	int block_id = base_memory_block_id(__section_nr(section));
580 581
	struct device *hintdev = hint ? &hint->dev : NULL;
	struct device *dev;
582

583 584 585 586
	dev = subsys_find_device_by_id(&memory_subsys, block_id, hintdev);
	if (hint)
		put_device(&hint->dev);
	if (!dev)
587
		return NULL;
588
	return container_of(dev, struct memory_block, dev);
589 590
}

591 592 593 594 595 596
/*
 * For now, we have a linear search to go find the appropriate
 * memory_block corresponding to a particular phys_index. If
 * this gets to be a real problem, we can always use a radix
 * tree or something here.
 *
597
 * This could be made generic for all device subsystems.
598 599 600 601 602 603
 */
struct memory_block *find_memory_block(struct mem_section *section)
{
	return find_memory_block_hinted(section, NULL);
}

604 605
static int init_memory_block(struct memory_block **memory,
			     struct mem_section *section, unsigned long state)
606
{
607
	struct memory_block *mem;
608
	unsigned long start_pfn;
609
	int scn_nr;
610 611
	int ret = 0;

612
	mem = kzalloc(sizeof(*mem), GFP_KERNEL);
613 614 615
	if (!mem)
		return -ENOMEM;

616
	scn_nr = __section_nr(section);
617 618 619
	mem->start_section_nr =
			base_memory_block_id(scn_nr) * sections_per_block;
	mem->end_section_nr = mem->start_section_nr + sections_per_block - 1;
620
	mem->state = state;
621
	mem->last_online = ONLINE_KEEP;
622
	mem->section_count++;
623
	mutex_init(&mem->state_mutex);
624
	start_pfn = section_nr_to_pfn(mem->start_section_nr);
625 626
	mem->phys_device = arch_get_memory_phys_device(start_pfn);

627
	ret = register_memory(mem);
628 629
	if (!ret)
		ret = mem_create_simple_file(mem, phys_index);
630 631
	if (!ret)
		ret = mem_create_simple_file(mem, end_phys_index);
632 633 634 635 636 637
	if (!ret)
		ret = mem_create_simple_file(mem, state);
	if (!ret)
		ret = mem_create_simple_file(mem, phys_device);
	if (!ret)
		ret = mem_create_simple_file(mem, removable);
638 639 640 641 642 643

	*memory = mem;
	return ret;
}

static int add_memory_section(int nid, struct mem_section *section,
644
			struct memory_block **mem_p,
645 646
			unsigned long state, enum mem_add_context context)
{
647 648
	struct memory_block *mem = NULL;
	int scn_nr = __section_nr(section);
649 650 651 652
	int ret = 0;

	mutex_lock(&mem_sysfs_mutex);

653 654 655 656 657 658 659 660 661 662 663
	if (context == BOOT) {
		/* same memory block ? */
		if (mem_p && *mem_p)
			if (scn_nr >= (*mem_p)->start_section_nr &&
			    scn_nr <= (*mem_p)->end_section_nr) {
				mem = *mem_p;
				kobject_get(&mem->dev.kobj);
			}
	} else
		mem = find_memory_block(section);

664 665
	if (mem) {
		mem->section_count++;
666
		kobject_put(&mem->dev.kobj);
667
	} else {
668
		ret = init_memory_block(&mem, section, state);
669 670 671 672 673
		/* store memory_block pointer for next loop */
		if (!ret && context == BOOT)
			if (mem_p)
				*mem_p = mem;
	}
674

675
	if (!ret) {
676 677
		if (context == HOTPLUG &&
		    mem->section_count == sections_per_block)
678 679 680
			ret = register_mem_sect_under_node(mem, nid);
	}

681
	mutex_unlock(&mem_sysfs_mutex);
682 683 684
	return ret;
}

685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706
/*
 * need an interface for the VM to add new memory regions,
 * but without onlining it.
 */
int register_new_memory(int nid, struct mem_section *section)
{
	return add_memory_section(nid, section, NULL, MEM_OFFLINE, HOTPLUG);
}

#ifdef CONFIG_MEMORY_HOTREMOVE
static void
unregister_memory(struct memory_block *memory)
{
	BUG_ON(memory->dev.bus != &memory_subsys);

	/* drop the ref. we got in remove_memory_block() */
	kobject_put(&memory->dev.kobj);
	device_unregister(&memory->dev);
}

static int remove_memory_block(unsigned long node_id,
			       struct mem_section *section, int phys_device)
707 708 709
{
	struct memory_block *mem;

710
	mutex_lock(&mem_sysfs_mutex);
711
	mem = find_memory_block(section);
712
	unregister_mem_sect_under_nodes(mem, __section_nr(section));
713 714 715 716

	mem->section_count--;
	if (mem->section_count == 0) {
		mem_remove_simple_file(mem, phys_index);
717
		mem_remove_simple_file(mem, end_phys_index);
718 719 720
		mem_remove_simple_file(mem, state);
		mem_remove_simple_file(mem, phys_device);
		mem_remove_simple_file(mem, removable);
721 722
		unregister_memory(mem);
	} else
723
		kobject_put(&mem->dev.kobj);
724

725
	mutex_unlock(&mem_sysfs_mutex);
726 727 728 729 730
	return 0;
}

int unregister_memory_section(struct mem_section *section)
{
731
	if (!present_section(section))
732 733 734 735
		return -EINVAL;

	return remove_memory_block(0, section, 0);
}
736
#endif /* CONFIG_MEMORY_HOTREMOVE */
737

738 739
/*
 * offline one memory block. If the memory block has been offlined, do nothing.
740 741
 *
 * Call under device_hotplug_lock.
742 743 744 745 746 747
 */
int offline_memory_block(struct memory_block *mem)
{
	int ret = 0;

	mutex_lock(&mem->state_mutex);
748 749 750 751 752 753
	if (mem->state != MEM_OFFLINE) {
		ret = __memory_block_change_state_uevent(mem, MEM_OFFLINE,
							 MEM_ONLINE, -1);
		if (!ret)
			mem->dev.offline = true;
	}
754 755 756 757 758
	mutex_unlock(&mem->state_mutex);

	return ret;
}

759 760 761 762 763 764
/* return true if the memory block is offlined, otherwise, return false */
bool is_memblock_offlined(struct memory_block *mem)
{
	return mem->state == MEM_OFFLINE;
}

765 766 767 768 769 770 771
/*
 * Initialize the sysfs support for memory devices...
 */
int __init memory_dev_init(void)
{
	unsigned int i;
	int ret;
772
	int err;
773
	unsigned long block_sz;
774
	struct memory_block *mem = NULL;
775

776
	ret = subsys_system_register(&memory_subsys, NULL);
777 778
	if (ret)
		goto out;
779

780 781 782
	block_sz = get_memory_block_size();
	sections_per_block = block_sz / MIN_MEMORY_BLOCK_SIZE;

783 784 785 786 787
	/*
	 * Create entries for memory sections that were found
	 * during boot and have been initialized
	 */
	for (i = 0; i < NR_MEM_SECTIONS; i++) {
788
		if (!present_section_nr(i))
789
			continue;
790 791 792 793
		/* don't need to reuse memory_block if only one per block */
		err = add_memory_section(0, __nr_to_section(i),
				 (sections_per_block == 1) ? NULL : &mem,
					 MEM_ONLINE,
794
					 BOOT);
795 796
		if (!ret)
			ret = err;
797 798
	}

799
	err = memory_probe_init();
800 801 802
	if (!ret)
		ret = err;
	err = memory_fail_init();
803 804 805 806 807 808 809
	if (!ret)
		ret = err;
	err = block_size_init();
	if (!ret)
		ret = err;
out:
	if (ret)
810
		printk(KERN_ERR "%s() failed: %d\n", __func__, ret);
811 812
	return ret;
}