migrate.c 33.6 KB
Newer Older
C
Christoph Lameter 已提交
1 2 3 4 5 6 7 8 9 10 11
/*
 * Memory Migration functionality - linux/mm/migration.c
 *
 * Copyright (C) 2006 Silicon Graphics, Inc., Christoph Lameter
 *
 * Page migration was first developed in the context of the memory hotplug
 * project. The main authors of the migration code are:
 *
 * IWAMOTO Toshihiro <iwamoto@valinux.co.jp>
 * Hirokazu Takahashi <taka@valinux.co.jp>
 * Dave Hansen <haveblue@us.ibm.com>
C
Christoph Lameter 已提交
12
 * Christoph Lameter
C
Christoph Lameter 已提交
13 14 15
 */

#include <linux/migrate.h>
16
#include <linux/export.h>
C
Christoph Lameter 已提交
17
#include <linux/swap.h>
18
#include <linux/swapops.h>
C
Christoph Lameter 已提交
19
#include <linux/pagemap.h>
20
#include <linux/buffer_head.h>
C
Christoph Lameter 已提交
21
#include <linux/mm_inline.h>
22
#include <linux/nsproxy.h>
C
Christoph Lameter 已提交
23
#include <linux/pagevec.h>
24
#include <linux/ksm.h>
C
Christoph Lameter 已提交
25 26 27 28
#include <linux/rmap.h>
#include <linux/topology.h>
#include <linux/cpu.h>
#include <linux/cpuset.h>
29
#include <linux/writeback.h>
30 31
#include <linux/mempolicy.h>
#include <linux/vmalloc.h>
32
#include <linux/security.h>
33
#include <linux/memcontrol.h>
34
#include <linux/syscalls.h>
N
Naoya Horiguchi 已提交
35
#include <linux/hugetlb.h>
36
#include <linux/gfp.h>
C
Christoph Lameter 已提交
37

38 39
#include <asm/tlbflush.h>

C
Christoph Lameter 已提交
40 41 42
#include "internal.h"

/*
43
 * migrate_prep() needs to be called before we start compiling a list of pages
44 45
 * to be migrated using isolate_lru_page(). If scheduling work on other CPUs is
 * undesirable, use migrate_prep_local()
C
Christoph Lameter 已提交
46 47 48 49 50 51 52 53 54 55 56 57 58 59
 */
int migrate_prep(void)
{
	/*
	 * Clear the LRU lists so pages can be isolated.
	 * Note that pages may be moved off the LRU after we have
	 * drained them. Those pages will fail to migrate like other
	 * pages that may be busy.
	 */
	lru_add_drain_all();

	return 0;
}

60 61 62 63 64 65 66 67
/* Do the necessary work of migrate_prep but not if it involves other CPUs */
int migrate_prep_local(void)
{
	lru_add_drain();

	return 0;
}

C
Christoph Lameter 已提交
68
/*
L
Lee Schermerhorn 已提交
69 70
 * Add isolated pages on the list back to the LRU under page lock
 * to avoid leaking evictable pages back onto unevictable list.
C
Christoph Lameter 已提交
71
 */
72
void putback_lru_pages(struct list_head *l)
C
Christoph Lameter 已提交
73 74 75 76 77
{
	struct page *page;
	struct page *page2;

	list_for_each_entry_safe(page, page2, l, lru) {
78
		list_del(&page->lru);
K
KOSAKI Motohiro 已提交
79
		dec_zone_page_state(page, NR_ISOLATED_ANON +
80
				page_is_file_cache(page));
L
Lee Schermerhorn 已提交
81
		putback_lru_page(page);
C
Christoph Lameter 已提交
82 83 84
	}
}

85 86 87
/*
 * Restore a potential migration pte to a working pte entry
 */
88 89
static int remove_migration_pte(struct page *new, struct vm_area_struct *vma,
				 unsigned long addr, void *old)
90 91 92 93 94 95 96 97 98
{
	struct mm_struct *mm = vma->vm_mm;
	swp_entry_t entry;
 	pgd_t *pgd;
 	pud_t *pud;
 	pmd_t *pmd;
	pte_t *ptep, pte;
 	spinlock_t *ptl;

N
Naoya Horiguchi 已提交
99 100 101 102 103 104 105 106 107
	if (unlikely(PageHuge(new))) {
		ptep = huge_pte_offset(mm, addr);
		if (!ptep)
			goto out;
		ptl = &mm->page_table_lock;
	} else {
		pgd = pgd_offset(mm, addr);
		if (!pgd_present(*pgd))
			goto out;
108

N
Naoya Horiguchi 已提交
109 110 111
		pud = pud_offset(pgd, addr);
		if (!pud_present(*pud))
			goto out;
112

N
Naoya Horiguchi 已提交
113
		pmd = pmd_offset(pud, addr);
114 115
		if (pmd_trans_huge(*pmd))
			goto out;
N
Naoya Horiguchi 已提交
116 117
		if (!pmd_present(*pmd))
			goto out;
118

N
Naoya Horiguchi 已提交
119
		ptep = pte_offset_map(pmd, addr);
120

121 122 123 124
		/*
		 * Peek to check is_swap_pte() before taking ptlock?  No, we
		 * can race mremap's move_ptes(), which skips anon_vma lock.
		 */
N
Naoya Horiguchi 已提交
125 126 127

		ptl = pte_lockptr(mm, pmd);
	}
128 129 130 131

 	spin_lock(ptl);
	pte = *ptep;
	if (!is_swap_pte(pte))
132
		goto unlock;
133 134 135

	entry = pte_to_swp_entry(pte);

136 137 138
	if (!is_migration_entry(entry) ||
	    migration_entry_to_page(entry) != old)
		goto unlock;
139 140 141 142 143

	get_page(new);
	pte = pte_mkold(mk_pte(new, vma->vm_page_prot));
	if (is_write_migration_entry(entry))
		pte = pte_mkwrite(pte);
A
Andi Kleen 已提交
144
#ifdef CONFIG_HUGETLB_PAGE
N
Naoya Horiguchi 已提交
145 146
	if (PageHuge(new))
		pte = pte_mkhuge(pte);
A
Andi Kleen 已提交
147
#endif
148
	flush_cache_page(vma, addr, pte_pfn(pte));
149
	set_pte_at(mm, addr, ptep, pte);
150

N
Naoya Horiguchi 已提交
151 152 153 154 155 156
	if (PageHuge(new)) {
		if (PageAnon(new))
			hugepage_add_anon_rmap(new, vma, addr);
		else
			page_dup_rmap(new);
	} else if (PageAnon(new))
157 158 159 160 161
		page_add_anon_rmap(new, vma, addr);
	else
		page_add_file_rmap(new);

	/* No need to invalidate - it was non-present before */
162
	update_mmu_cache(vma, addr, ptep);
163
unlock:
164
	pte_unmap_unlock(ptep, ptl);
165 166
out:
	return SWAP_AGAIN;
167 168
}

169 170 171 172 173 174
/*
 * Get rid of all migration entries and replace them by
 * references to the indicated page.
 */
static void remove_migration_ptes(struct page *old, struct page *new)
{
175
	rmap_walk(new, remove_migration_pte, old);
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
/*
 * Something used the pte of a page under migration. We need to
 * get to the page and wait until migration is finished.
 * When we return from this function the fault will be retried.
 */
void migration_entry_wait(struct mm_struct *mm, pmd_t *pmd,
				unsigned long address)
{
	pte_t *ptep, pte;
	spinlock_t *ptl;
	swp_entry_t entry;
	struct page *page;

	ptep = pte_offset_map_lock(mm, pmd, address, &ptl);
	pte = *ptep;
	if (!is_swap_pte(pte))
		goto out;

	entry = pte_to_swp_entry(pte);
	if (!is_migration_entry(entry))
		goto out;

	page = migration_entry_to_page(entry);

N
Nick Piggin 已提交
202 203 204 205 206 207 208 209 210
	/*
	 * Once radix-tree replacement of page migration started, page_count
	 * *must* be zero. And, we don't want to call wait_on_page_locked()
	 * against a page without get_page().
	 * So, we use get_page_unless_zero(), here. Even failed, page fault
	 * will occur again.
	 */
	if (!get_page_unless_zero(page))
		goto out;
211 212 213 214 215 216 217 218
	pte_unmap_unlock(ptep, ptl);
	wait_on_page_locked(page);
	put_page(page);
	return;
out:
	pte_unmap_unlock(ptep, ptl);
}

219 220
#ifdef CONFIG_BLOCK
/* Returns true if all buffers are successfully locked */
221 222
static bool buffer_migrate_lock_buffers(struct buffer_head *head,
							enum migrate_mode mode)
223 224 225 226
{
	struct buffer_head *bh = head;

	/* Simple case, sync compaction */
227
	if (mode != MIGRATE_ASYNC) {
228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262
		do {
			get_bh(bh);
			lock_buffer(bh);
			bh = bh->b_this_page;

		} while (bh != head);

		return true;
	}

	/* async case, we cannot block on lock_buffer so use trylock_buffer */
	do {
		get_bh(bh);
		if (!trylock_buffer(bh)) {
			/*
			 * We failed to lock the buffer and cannot stall in
			 * async migration. Release the taken locks
			 */
			struct buffer_head *failed_bh = bh;
			put_bh(failed_bh);
			bh = head;
			while (bh != failed_bh) {
				unlock_buffer(bh);
				put_bh(bh);
				bh = bh->b_this_page;
			}
			return false;
		}

		bh = bh->b_this_page;
	} while (bh != head);
	return true;
}
#else
static inline bool buffer_migrate_lock_buffers(struct buffer_head *head,
263
							enum migrate_mode mode)
264 265 266 267 268
{
	return true;
}
#endif /* CONFIG_BLOCK */

C
Christoph Lameter 已提交
269
/*
270
 * Replace the page in the mapping.
271 272 273 274
 *
 * The number of remaining references must be:
 * 1 for anonymous pages without a mapping
 * 2 for pages with a mapping
275
 * 3 for pages with a mapping and PagePrivate/PagePrivate2 set.
C
Christoph Lameter 已提交
276
 */
277
static int migrate_page_move_mapping(struct address_space *mapping,
278
		struct page *newpage, struct page *page,
279
		struct buffer_head *head, enum migrate_mode mode)
C
Christoph Lameter 已提交
280
{
N
Nick Piggin 已提交
281
	int expected_count;
282
	void **pslot;
C
Christoph Lameter 已提交
283

284
	if (!mapping) {
285
		/* Anonymous page without mapping */
286 287 288 289 290
		if (page_count(page) != 1)
			return -EAGAIN;
		return 0;
	}

N
Nick Piggin 已提交
291
	spin_lock_irq(&mapping->tree_lock);
C
Christoph Lameter 已提交
292

293 294
	pslot = radix_tree_lookup_slot(&mapping->page_tree,
 					page_index(page));
C
Christoph Lameter 已提交
295

296
	expected_count = 2 + page_has_private(page);
N
Nick Piggin 已提交
297
	if (page_count(page) != expected_count ||
298
		radix_tree_deref_slot_protected(pslot, &mapping->tree_lock) != page) {
N
Nick Piggin 已提交
299
		spin_unlock_irq(&mapping->tree_lock);
300
		return -EAGAIN;
C
Christoph Lameter 已提交
301 302
	}

N
Nick Piggin 已提交
303
	if (!page_freeze_refs(page, expected_count)) {
N
Nick Piggin 已提交
304
		spin_unlock_irq(&mapping->tree_lock);
N
Nick Piggin 已提交
305 306 307
		return -EAGAIN;
	}

308 309 310 311 312 313 314
	/*
	 * In the async migration case of moving a page with buffers, lock the
	 * buffers using trylock before the mapping is moved. If the mapping
	 * was moved, we later failed to lock the buffers and could not move
	 * the mapping back due to an elevated page count, we would have to
	 * block waiting on other references to be dropped.
	 */
315 316
	if (mode == MIGRATE_ASYNC && head &&
			!buffer_migrate_lock_buffers(head, mode)) {
317 318 319 320 321
		page_unfreeze_refs(page, expected_count);
		spin_unlock_irq(&mapping->tree_lock);
		return -EAGAIN;
	}

C
Christoph Lameter 已提交
322 323 324
	/*
	 * Now we know that no one else is looking at the page.
	 */
325
	get_page(newpage);	/* add cache reference */
C
Christoph Lameter 已提交
326 327 328 329 330
	if (PageSwapCache(page)) {
		SetPageSwapCache(newpage);
		set_page_private(newpage, page_private(page));
	}

331 332 333
	radix_tree_replace_slot(pslot, newpage);

	/*
334 335
	 * Drop cache reference from old page by unfreezing
	 * to one less reference.
336 337
	 * We know this isn't the last reference.
	 */
338
	page_unfreeze_refs(page, expected_count - 1);
339

340 341 342 343 344 345 346 347 348 349 350 351
	/*
	 * If moved to a different zone then also account
	 * the page for that zone. Other VM counters will be
	 * taken care of when we establish references to the
	 * new page and drop references to the old page.
	 *
	 * Note that anonymous pages are accounted for
	 * via NR_FILE_PAGES and NR_ANON_PAGES if they
	 * are mapped to swap space.
	 */
	__dec_zone_page_state(page, NR_FILE_PAGES);
	__inc_zone_page_state(newpage, NR_FILE_PAGES);
352
	if (!PageSwapCache(page) && PageSwapBacked(page)) {
353 354 355
		__dec_zone_page_state(page, NR_SHMEM);
		__inc_zone_page_state(newpage, NR_SHMEM);
	}
N
Nick Piggin 已提交
356
	spin_unlock_irq(&mapping->tree_lock);
C
Christoph Lameter 已提交
357 358 359 360

	return 0;
}

N
Naoya Horiguchi 已提交
361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383
/*
 * The expected number of remaining references is the same as that
 * of migrate_page_move_mapping().
 */
int migrate_huge_page_move_mapping(struct address_space *mapping,
				   struct page *newpage, struct page *page)
{
	int expected_count;
	void **pslot;

	if (!mapping) {
		if (page_count(page) != 1)
			return -EAGAIN;
		return 0;
	}

	spin_lock_irq(&mapping->tree_lock);

	pslot = radix_tree_lookup_slot(&mapping->page_tree,
					page_index(page));

	expected_count = 2 + page_has_private(page);
	if (page_count(page) != expected_count ||
384
		radix_tree_deref_slot_protected(pslot, &mapping->tree_lock) != page) {
N
Naoya Horiguchi 已提交
385 386 387 388 389 390 391 392 393 394 395 396 397
		spin_unlock_irq(&mapping->tree_lock);
		return -EAGAIN;
	}

	if (!page_freeze_refs(page, expected_count)) {
		spin_unlock_irq(&mapping->tree_lock);
		return -EAGAIN;
	}

	get_page(newpage);

	radix_tree_replace_slot(pslot, newpage);

398
	page_unfreeze_refs(page, expected_count - 1);
N
Naoya Horiguchi 已提交
399 400 401 402 403

	spin_unlock_irq(&mapping->tree_lock);
	return 0;
}

C
Christoph Lameter 已提交
404 405 406
/*
 * Copy the page to its new location
 */
N
Naoya Horiguchi 已提交
407
void migrate_page_copy(struct page *newpage, struct page *page)
C
Christoph Lameter 已提交
408
{
N
Naoya Horiguchi 已提交
409 410 411 412
	if (PageHuge(page))
		copy_huge_page(newpage, page);
	else
		copy_highpage(newpage, page);
C
Christoph Lameter 已提交
413 414 415 416 417 418 419

	if (PageError(page))
		SetPageError(newpage);
	if (PageReferenced(page))
		SetPageReferenced(newpage);
	if (PageUptodate(page))
		SetPageUptodate(newpage);
L
Lee Schermerhorn 已提交
420 421
	if (TestClearPageActive(page)) {
		VM_BUG_ON(PageUnevictable(page));
C
Christoph Lameter 已提交
422
		SetPageActive(newpage);
423 424
	} else if (TestClearPageUnevictable(page))
		SetPageUnevictable(newpage);
C
Christoph Lameter 已提交
425 426 427 428 429 430 431
	if (PageChecked(page))
		SetPageChecked(newpage);
	if (PageMappedToDisk(page))
		SetPageMappedToDisk(newpage);

	if (PageDirty(page)) {
		clear_page_dirty_for_io(page);
N
Nick Piggin 已提交
432 433 434 435 436
		/*
		 * Want to mark the page and the radix tree as dirty, and
		 * redo the accounting that clear_page_dirty_for_io undid,
		 * but we can't use set_page_dirty because that function
		 * is actually a signal that all of the page has become dirty.
L
Lucas De Marchi 已提交
437
		 * Whereas only part of our page may be dirty.
N
Nick Piggin 已提交
438 439
		 */
		__set_page_dirty_nobuffers(newpage);
C
Christoph Lameter 已提交
440 441
 	}

N
Nick Piggin 已提交
442
	mlock_migrate_page(newpage, page);
443
	ksm_migrate_page(newpage, page);
N
Nick Piggin 已提交
444

C
Christoph Lameter 已提交
445 446 447 448 449 450 451 452 453 454 455 456 457
	ClearPageSwapCache(page);
	ClearPagePrivate(page);
	set_page_private(page, 0);
	page->mapping = NULL;

	/*
	 * If any waiters have accumulated on the new page then
	 * wake them up.
	 */
	if (PageWriteback(newpage))
		end_page_writeback(newpage);
}

458 459 460 461 462
/************************************************************
 *                    Migration functions
 ***********************************************************/

/* Always fail migration. Used for mappings that are not movable */
463 464
int fail_migrate_page(struct address_space *mapping,
			struct page *newpage, struct page *page)
465 466 467 468 469
{
	return -EIO;
}
EXPORT_SYMBOL(fail_migrate_page);

C
Christoph Lameter 已提交
470 471
/*
 * Common logic to directly migrate a single page suitable for
472
 * pages that do not use PagePrivate/PagePrivate2.
C
Christoph Lameter 已提交
473 474 475
 *
 * Pages are locked upon entry and exit.
 */
476
int migrate_page(struct address_space *mapping,
477 478
		struct page *newpage, struct page *page,
		enum migrate_mode mode)
C
Christoph Lameter 已提交
479 480 481 482 483
{
	int rc;

	BUG_ON(PageWriteback(page));	/* Writeback must be complete */

484
	rc = migrate_page_move_mapping(mapping, newpage, page, NULL, mode);
C
Christoph Lameter 已提交
485 486 487 488 489 490 491 492 493

	if (rc)
		return rc;

	migrate_page_copy(newpage, page);
	return 0;
}
EXPORT_SYMBOL(migrate_page);

494
#ifdef CONFIG_BLOCK
495 496 497 498 499
/*
 * Migration function for pages with buffers. This function can only be used
 * if the underlying filesystem guarantees that no other references to "page"
 * exist.
 */
500
int buffer_migrate_page(struct address_space *mapping,
501
		struct page *newpage, struct page *page, enum migrate_mode mode)
502 503 504 505 506
{
	struct buffer_head *bh, *head;
	int rc;

	if (!page_has_buffers(page))
507
		return migrate_page(mapping, newpage, page, mode);
508 509 510

	head = page_buffers(page);

511
	rc = migrate_page_move_mapping(mapping, newpage, page, head, mode);
512 513 514 515

	if (rc)
		return rc;

516 517 518 519 520
	/*
	 * In the async case, migrate_page_move_mapping locked the buffers
	 * with an IRQ-safe spinlock held. In the sync case, the buffers
	 * need to be locked now
	 */
521 522
	if (mode != MIGRATE_ASYNC)
		BUG_ON(!buffer_migrate_lock_buffers(head, mode));
523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551

	ClearPagePrivate(page);
	set_page_private(newpage, page_private(page));
	set_page_private(page, 0);
	put_page(page);
	get_page(newpage);

	bh = head;
	do {
		set_bh_page(bh, newpage, bh_offset(bh));
		bh = bh->b_this_page;

	} while (bh != head);

	SetPagePrivate(newpage);

	migrate_page_copy(newpage, page);

	bh = head;
	do {
		unlock_buffer(bh);
 		put_bh(bh);
		bh = bh->b_this_page;

	} while (bh != head);

	return 0;
}
EXPORT_SYMBOL(buffer_migrate_page);
552
#endif
553

554 555 556 557
/*
 * Writeback a page to clean the dirty state
 */
static int writeout(struct address_space *mapping, struct page *page)
558
{
559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575
	struct writeback_control wbc = {
		.sync_mode = WB_SYNC_NONE,
		.nr_to_write = 1,
		.range_start = 0,
		.range_end = LLONG_MAX,
		.for_reclaim = 1
	};
	int rc;

	if (!mapping->a_ops->writepage)
		/* No write method for the address space */
		return -EINVAL;

	if (!clear_page_dirty_for_io(page))
		/* Someone else already triggered a write */
		return -EAGAIN;

576
	/*
577 578 579 580 581 582
	 * A dirty page may imply that the underlying filesystem has
	 * the page on some queue. So the page must be clean for
	 * migration. Writeout may mean we loose the lock and the
	 * page state is no longer what we checked for earlier.
	 * At this point we know that the migration attempt cannot
	 * be successful.
583
	 */
584
	remove_migration_ptes(page, page);
585

586
	rc = mapping->a_ops->writepage(page, &wbc);
587

588 589 590 591
	if (rc != AOP_WRITEPAGE_ACTIVATE)
		/* unlocked. Relock */
		lock_page(page);

H
Hugh Dickins 已提交
592
	return (rc < 0) ? -EIO : -EAGAIN;
593 594 595 596 597 598
}

/*
 * Default handling if a filesystem does not provide a migration function.
 */
static int fallback_migrate_page(struct address_space *mapping,
599
	struct page *newpage, struct page *page, enum migrate_mode mode)
600
{
601
	if (PageDirty(page)) {
602 603
		/* Only writeback pages in full synchronous migration */
		if (mode != MIGRATE_SYNC)
604
			return -EBUSY;
605
		return writeout(mapping, page);
606
	}
607 608 609 610 611

	/*
	 * Buffers may be managed in a filesystem specific way.
	 * We must have no buffers or drop them.
	 */
612
	if (page_has_private(page) &&
613 614 615
	    !try_to_release_page(page, GFP_KERNEL))
		return -EAGAIN;

616
	return migrate_page(mapping, newpage, page, mode);
617 618
}

619 620 621 622 623 624
/*
 * Move a page to a newly allocated page
 * The page is locked and all ptes have been successfully removed.
 *
 * The new page will have replaced the old page if this function
 * is successful.
L
Lee Schermerhorn 已提交
625 626 627 628
 *
 * Return value:
 *   < 0 - error code
 *  == 0 - success
629
 */
630
static int move_to_new_page(struct page *newpage, struct page *page,
631
				int remap_swapcache, enum migrate_mode mode)
632 633 634 635 636 637 638 639 640
{
	struct address_space *mapping;
	int rc;

	/*
	 * Block others from accessing the page when we get around to
	 * establishing additional references. We are the only one
	 * holding a reference to the new page at this point.
	 */
N
Nick Piggin 已提交
641
	if (!trylock_page(newpage))
642 643 644 645 646
		BUG();

	/* Prepare mapping for the new page.*/
	newpage->index = page->index;
	newpage->mapping = page->mapping;
R
Rik van Riel 已提交
647 648
	if (PageSwapBacked(page))
		SetPageSwapBacked(newpage);
649 650 651

	mapping = page_mapping(page);
	if (!mapping)
652
		rc = migrate_page(mapping, newpage, page, mode);
653
	else if (mapping->a_ops->migratepage)
654
		/*
655 656 657 658
		 * Most pages have a mapping and most filesystems provide a
		 * migratepage callback. Anonymous pages are part of swap
		 * space which also has its own migratepage callback. This
		 * is the most common path for page migration.
659
		 */
660
		rc = mapping->a_ops->migratepage(mapping,
661
						newpage, page, mode);
662
	else
663
		rc = fallback_migrate_page(mapping, newpage, page, mode);
664

665
	if (rc) {
666
		newpage->mapping = NULL;
667 668 669 670
	} else {
		if (remap_swapcache)
			remove_migration_ptes(page, newpage);
	}
671 672 673 674 675 676

	unlock_page(newpage);

	return rc;
}

677
static int __unmap_and_move(struct page *page, struct page *newpage,
678
			int force, bool offlining, enum migrate_mode mode)
679
{
680
	int rc = -EAGAIN;
681
	int remap_swapcache = 1;
682
	int charge = 0;
683
	struct mem_cgroup *mem;
684
	struct anon_vma *anon_vma = NULL;
685

N
Nick Piggin 已提交
686
	if (!trylock_page(page)) {
687
		if (!force || mode == MIGRATE_ASYNC)
688
			goto out;
689 690 691 692 693 694 695 696 697 698 699 700 701 702 703

		/*
		 * It's not safe for direct compaction to call lock_page.
		 * For example, during page readahead pages are added locked
		 * to the LRU. Later, when the IO completes the pages are
		 * marked uptodate and unlocked. However, the queueing
		 * could be merging multiple pages for one bio (e.g.
		 * mpage_readpages). If an allocation happens for the
		 * second or third page, the process can end up locking
		 * the same page twice and deadlocking. Rather than
		 * trying to be clever about what pages can be locked,
		 * avoid the use of lock_page for direct compaction
		 * altogether.
		 */
		if (current->flags & PF_MEMALLOC)
704
			goto out;
705

706 707 708
		lock_page(page);
	}

709 710 711 712 713 714 715 716 717 718 719 720 721 722
	/*
	 * Only memory hotplug's offline_pages() caller has locked out KSM,
	 * and can safely migrate a KSM page.  The other cases have skipped
	 * PageKsm along with PageReserved - but it is only now when we have
	 * the page lock that we can be certain it will not go KSM beneath us
	 * (KSM will not upgrade a page from PageAnon to PageKsm when it sees
	 * its pagecount raised, but only here do we take the page lock which
	 * serializes that).
	 */
	if (PageKsm(page) && !offlining) {
		rc = -EBUSY;
		goto unlock;
	}

723
	/* charge against new page */
724
	charge = mem_cgroup_prepare_migration(page, newpage, &mem, GFP_KERNEL);
725 726 727 728 729 730
	if (charge == -ENOMEM) {
		rc = -ENOMEM;
		goto unlock;
	}
	BUG_ON(charge);

731
	if (PageWriteback(page)) {
732
		/*
733 734 735 736
		 * Only in the case of a full syncronous migration is it
		 * necessary to wait for PageWriteback. In the async case,
		 * the retry loop is too short and in the sync-light case,
		 * the overhead of stalling is too much
737
		 */
738
		if (mode != MIGRATE_SYNC) {
739 740 741 742
			rc = -EBUSY;
			goto uncharge;
		}
		if (!force)
743
			goto uncharge;
744 745 746
		wait_on_page_writeback(page);
	}
	/*
747 748
	 * By try_to_unmap(), page->mapcount goes down to 0 here. In this case,
	 * we cannot notice that anon_vma is freed while we migrates a page.
749
	 * This get_anon_vma() delays freeing anon_vma pointer until the end
750
	 * of migration. File cache pages are no problem because of page_lock()
751 752
	 * File Caches may use write_page() or lock_page() in migration, then,
	 * just care Anon page here.
753
	 */
754
	if (PageAnon(page)) {
755 756 757 758
		/*
		 * Only page_lock_anon_vma() understands the subtleties of
		 * getting a hold on an anon_vma from outside one of its mms.
		 */
759
		anon_vma = page_get_anon_vma(page);
760 761
		if (anon_vma) {
			/*
762
			 * Anon page
763 764
			 */
		} else if (PageSwapCache(page)) {
765 766 767 768 769 770 771 772 773 774 775 776 777 778
			/*
			 * We cannot be sure that the anon_vma of an unmapped
			 * swapcache page is safe to use because we don't
			 * know in advance if the VMA that this page belonged
			 * to still exists. If the VMA and others sharing the
			 * data have been freed, then the anon_vma could
			 * already be invalid.
			 *
			 * To avoid this possibility, swapcache pages get
			 * migrated but are not remapped when migration
			 * completes
			 */
			remap_swapcache = 0;
		} else {
779
			goto uncharge;
780
		}
781
	}
782

783
	/*
784 785 786 787 788 789 790 791 792 793
	 * Corner case handling:
	 * 1. When a new swap-cache page is read into, it is added to the LRU
	 * and treated as swapcache but it has no rmap yet.
	 * Calling try_to_unmap() against a page->mapping==NULL page will
	 * trigger a BUG.  So handle it here.
	 * 2. An orphaned page (see truncate_complete_page) might have
	 * fs-private metadata. The page can be picked up due to memory
	 * offlining.  Everywhere else except page reclaim, the page is
	 * invisible to the vm, so the page can not be migrated.  So try to
	 * free the metadata, so the page can be freed.
794
	 */
795
	if (!page->mapping) {
796 797
		VM_BUG_ON(PageAnon(page));
		if (page_has_private(page)) {
798
			try_to_free_buffers(page);
799
			goto uncharge;
800
		}
801
		goto skip_unmap;
802 803
	}

804
	/* Establish migration ptes or remove ptes */
805
	try_to_unmap(page, TTU_MIGRATION|TTU_IGNORE_MLOCK|TTU_IGNORE_ACCESS);
806

807
skip_unmap:
808
	if (!page_mapped(page))
809
		rc = move_to_new_page(newpage, page, remap_swapcache, mode);
810

811
	if (rc && remap_swapcache)
812
		remove_migration_ptes(page, page);
813 814

	/* Drop an anon_vma reference if we took one */
815
	if (anon_vma)
816
		put_anon_vma(anon_vma);
817

818 819
uncharge:
	if (!charge)
820
		mem_cgroup_end_migration(mem, page, newpage, rc == 0);
821 822
unlock:
	unlock_page(page);
823 824 825
out:
	return rc;
}
826

827 828 829 830 831
/*
 * Obtain the lock on page, remove all ptes and migrate the page
 * to the newly allocated page in newpage.
 */
static int unmap_and_move(new_page_t get_new_page, unsigned long private,
832 833
			struct page *page, int force, bool offlining,
			enum migrate_mode mode)
834 835 836 837 838 839 840 841
{
	int rc = 0;
	int *result = NULL;
	struct page *newpage = get_new_page(page, private, &result);

	if (!newpage)
		return -ENOMEM;

842 843
	mem_cgroup_reset_owner(newpage);

844 845 846 847 848 849 850 851 852
	if (page_count(page) == 1) {
		/* page was freed from under us. So we are done. */
		goto out;
	}

	if (unlikely(PageTransHuge(page)))
		if (unlikely(split_huge_page(page)))
			goto out;

853
	rc = __unmap_and_move(page, newpage, force, offlining, mode);
854
out:
855
	if (rc != -EAGAIN) {
856 857 858 859 860 861 862
		/*
		 * A page that has been migrated has all references
		 * removed and will be freed. A page that has not been
		 * migrated will have kepts its references and be
		 * restored.
		 */
		list_del(&page->lru);
K
KOSAKI Motohiro 已提交
863
		dec_zone_page_state(page, NR_ISOLATED_ANON +
864
				page_is_file_cache(page));
L
Lee Schermerhorn 已提交
865
		putback_lru_page(page);
866
	}
867 868 869 870
	/*
	 * Move the new page to the LRU. If migration was not successful
	 * then this will free the page.
	 */
L
Lee Schermerhorn 已提交
871
	putback_lru_page(newpage);
872 873 874 875 876 877
	if (result) {
		if (rc)
			*result = rc;
		else
			*result = page_to_nid(newpage);
	}
878 879 880
	return rc;
}

N
Naoya Horiguchi 已提交
881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900
/*
 * Counterpart of unmap_and_move_page() for hugepage migration.
 *
 * This function doesn't wait the completion of hugepage I/O
 * because there is no race between I/O and migration for hugepage.
 * Note that currently hugepage I/O occurs only in direct I/O
 * where no lock is held and PG_writeback is irrelevant,
 * and writeback status of all subpages are counted in the reference
 * count of the head page (i.e. if all subpages of a 2MB hugepage are
 * under direct I/O, the reference of the head page is 512 and a bit more.)
 * This means that when we try to migrate hugepage whose subpages are
 * doing direct I/O, some references remain after try_to_unmap() and
 * hugepage migration fails without data corruption.
 *
 * There is also no race when direct I/O is issued on the page under migration,
 * because then pte is replaced with migration swap entry and direct I/O code
 * will wait in the page fault for migration to complete.
 */
static int unmap_and_move_huge_page(new_page_t get_new_page,
				unsigned long private, struct page *hpage,
901 902
				int force, bool offlining,
				enum migrate_mode mode)
N
Naoya Horiguchi 已提交
903 904 905 906 907 908 909 910 911 912 913 914
{
	int rc = 0;
	int *result = NULL;
	struct page *new_hpage = get_new_page(hpage, private, &result);
	struct anon_vma *anon_vma = NULL;

	if (!new_hpage)
		return -ENOMEM;

	rc = -EAGAIN;

	if (!trylock_page(hpage)) {
915
		if (!force || mode != MIGRATE_SYNC)
N
Naoya Horiguchi 已提交
916 917 918 919
			goto out;
		lock_page(hpage);
	}

920 921
	if (PageAnon(hpage))
		anon_vma = page_get_anon_vma(hpage);
N
Naoya Horiguchi 已提交
922 923 924 925

	try_to_unmap(hpage, TTU_MIGRATION|TTU_IGNORE_MLOCK|TTU_IGNORE_ACCESS);

	if (!page_mapped(hpage))
926
		rc = move_to_new_page(new_hpage, hpage, 1, mode);
N
Naoya Horiguchi 已提交
927 928 929 930

	if (rc)
		remove_migration_ptes(hpage, hpage);

H
Hugh Dickins 已提交
931
	if (anon_vma)
932
		put_anon_vma(anon_vma);
N
Naoya Horiguchi 已提交
933 934
	unlock_page(hpage);

935
out:
N
Naoya Horiguchi 已提交
936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951
	if (rc != -EAGAIN) {
		list_del(&hpage->lru);
		put_page(hpage);
	}

	put_page(new_hpage);

	if (result) {
		if (rc)
			*result = rc;
		else
			*result = page_to_nid(new_hpage);
	}
	return rc;
}

C
Christoph Lameter 已提交
952 953 954
/*
 * migrate_pages
 *
955 956 957
 * The function takes one list of pages to migrate and a function
 * that determines from the page to be migrated and the private data
 * the target of the move and allocates the page.
C
Christoph Lameter 已提交
958 959 960
 *
 * The function returns after 10 attempts or if no pages
 * are movable anymore because to has become empty
961 962
 * or no retryable pages exist anymore.
 * Caller should call putback_lru_pages to return pages to the LRU
963
 * or free list only if ret != 0.
C
Christoph Lameter 已提交
964
 *
965
 * Return: Number of pages not migrated or error code.
C
Christoph Lameter 已提交
966
 */
967
int migrate_pages(struct list_head *from,
968
		new_page_t get_new_page, unsigned long private, bool offlining,
969
		enum migrate_mode mode)
C
Christoph Lameter 已提交
970
{
971
	int retry = 1;
C
Christoph Lameter 已提交
972 973 974 975 976 977 978 979 980 981
	int nr_failed = 0;
	int pass = 0;
	struct page *page;
	struct page *page2;
	int swapwrite = current->flags & PF_SWAPWRITE;
	int rc;

	if (!swapwrite)
		current->flags |= PF_SWAPWRITE;

982 983
	for(pass = 0; pass < 10 && retry; pass++) {
		retry = 0;
C
Christoph Lameter 已提交
984

985 986
		list_for_each_entry_safe(page, page2, from, lru) {
			cond_resched();
987

988
			rc = unmap_and_move(get_new_page, private,
989
						page, pass > 2, offlining,
990
						mode);
991

992
			switch(rc) {
993 994
			case -ENOMEM:
				goto out;
995
			case -EAGAIN:
996
				retry++;
997 998 999 1000
				break;
			case 0:
				break;
			default:
1001 1002
				/* Permanent failure */
				nr_failed++;
1003
				break;
1004
			}
C
Christoph Lameter 已提交
1005 1006
		}
	}
1007 1008
	rc = 0;
out:
C
Christoph Lameter 已提交
1009 1010 1011
	if (!swapwrite)
		current->flags &= ~PF_SWAPWRITE;

1012 1013
	if (rc)
		return rc;
C
Christoph Lameter 已提交
1014

1015
	return nr_failed + retry;
C
Christoph Lameter 已提交
1016
}
1017

N
Naoya Horiguchi 已提交
1018
int migrate_huge_pages(struct list_head *from,
1019
		new_page_t get_new_page, unsigned long private, bool offlining,
1020
		enum migrate_mode mode)
N
Naoya Horiguchi 已提交
1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035
{
	int retry = 1;
	int nr_failed = 0;
	int pass = 0;
	struct page *page;
	struct page *page2;
	int rc;

	for (pass = 0; pass < 10 && retry; pass++) {
		retry = 0;

		list_for_each_entry_safe(page, page2, from, lru) {
			cond_resched();

			rc = unmap_and_move_huge_page(get_new_page,
1036
					private, page, pass > 2, offlining,
1037
					mode);
N
Naoya Horiguchi 已提交
1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061

			switch(rc) {
			case -ENOMEM:
				goto out;
			case -EAGAIN:
				retry++;
				break;
			case 0:
				break;
			default:
				/* Permanent failure */
				nr_failed++;
				break;
			}
		}
	}
	rc = 0;
out:
	if (rc)
		return rc;

	return nr_failed + retry;
}

1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085
#ifdef CONFIG_NUMA
/*
 * Move a list of individual pages
 */
struct page_to_node {
	unsigned long addr;
	struct page *page;
	int node;
	int status;
};

static struct page *new_page_node(struct page *p, unsigned long private,
		int **result)
{
	struct page_to_node *pm = (struct page_to_node *)private;

	while (pm->node != MAX_NUMNODES && pm->page != p)
		pm++;

	if (pm->node == MAX_NUMNODES)
		return NULL;

	*result = &pm->status;

1086
	return alloc_pages_exact_node(pm->node,
1087
				GFP_HIGHUSER_MOVABLE | GFP_THISNODE, 0);
1088 1089 1090 1091 1092 1093
}

/*
 * Move a set of pages as indicated in the pm array. The addr
 * field must be set to the virtual address of the page to be moved
 * and the node number must contain a valid target node.
1094
 * The pm array ends with node = MAX_NUMNODES.
1095
 */
1096 1097 1098
static int do_move_page_to_node_array(struct mm_struct *mm,
				      struct page_to_node *pm,
				      int migrate_all)
1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114
{
	int err;
	struct page_to_node *pp;
	LIST_HEAD(pagelist);

	down_read(&mm->mmap_sem);

	/*
	 * Build a list of pages to migrate
	 */
	for (pp = pm; pp->node != MAX_NUMNODES; pp++) {
		struct vm_area_struct *vma;
		struct page *page;

		err = -EFAULT;
		vma = find_vma(mm, pp->addr);
1115
		if (!vma || pp->addr < vma->vm_start || !vma_migratable(vma))
1116 1117
			goto set_status;

1118
		page = follow_page(vma, pp->addr, FOLL_GET|FOLL_SPLIT);
1119 1120 1121 1122 1123

		err = PTR_ERR(page);
		if (IS_ERR(page))
			goto set_status;

1124 1125 1126 1127
		err = -ENOENT;
		if (!page)
			goto set_status;

1128 1129
		/* Use PageReserved to check for zero page */
		if (PageReserved(page) || PageKsm(page))
1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145
			goto put_and_set;

		pp->page = page;
		err = page_to_nid(page);

		if (err == pp->node)
			/*
			 * Node already in the right place
			 */
			goto put_and_set;

		err = -EACCES;
		if (page_mapcount(page) > 1 &&
				!migrate_all)
			goto put_and_set;

1146
		err = isolate_lru_page(page);
1147
		if (!err) {
1148
			list_add_tail(&page->lru, &pagelist);
1149 1150 1151
			inc_zone_page_state(page, NR_ISOLATED_ANON +
					    page_is_file_cache(page));
		}
1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162
put_and_set:
		/*
		 * Either remove the duplicate refcount from
		 * isolate_lru_page() or drop the page ref if it was
		 * not isolated.
		 */
		put_page(page);
set_status:
		pp->status = err;
	}

1163
	err = 0;
1164
	if (!list_empty(&pagelist)) {
1165
		err = migrate_pages(&pagelist, new_page_node,
1166
				(unsigned long)pm, 0, MIGRATE_SYNC);
1167 1168 1169
		if (err)
			putback_lru_pages(&pagelist);
	}
1170 1171 1172 1173 1174

	up_read(&mm->mmap_sem);
	return err;
}

1175 1176 1177 1178 1179 1180 1181 1182 1183 1184
/*
 * Migrate an array of page address onto an array of nodes and fill
 * the corresponding array of status.
 */
static int do_pages_move(struct mm_struct *mm, struct task_struct *task,
			 unsigned long nr_pages,
			 const void __user * __user *pages,
			 const int __user *nodes,
			 int __user *status, int flags)
{
1185
	struct page_to_node *pm;
1186
	nodemask_t task_nodes;
1187 1188 1189
	unsigned long chunk_nr_pages;
	unsigned long chunk_start;
	int err;
1190 1191 1192

	task_nodes = cpuset_mems_allowed(task);

1193 1194 1195
	err = -ENOMEM;
	pm = (struct page_to_node *)__get_free_page(GFP_KERNEL);
	if (!pm)
1196
		goto out;
1197 1198 1199

	migrate_prep();

1200
	/*
1201 1202
	 * Store a chunk of page_to_node array in a page,
	 * but keep the last one as a marker
1203
	 */
1204
	chunk_nr_pages = (PAGE_SIZE / sizeof(struct page_to_node)) - 1;
1205

1206 1207 1208 1209
	for (chunk_start = 0;
	     chunk_start < nr_pages;
	     chunk_start += chunk_nr_pages) {
		int j;
1210

1211 1212 1213 1214 1215 1216
		if (chunk_start + chunk_nr_pages > nr_pages)
			chunk_nr_pages = nr_pages - chunk_start;

		/* fill the chunk pm with addrs and nodes from user-space */
		for (j = 0; j < chunk_nr_pages; j++) {
			const void __user *p;
1217 1218
			int node;

1219 1220 1221 1222 1223 1224
			err = -EFAULT;
			if (get_user(p, pages + j + chunk_start))
				goto out_pm;
			pm[j].addr = (unsigned long) p;

			if (get_user(node, nodes + j + chunk_start))
1225 1226 1227
				goto out_pm;

			err = -ENODEV;
1228 1229 1230
			if (node < 0 || node >= MAX_NUMNODES)
				goto out_pm;

1231 1232 1233 1234 1235 1236 1237
			if (!node_state(node, N_HIGH_MEMORY))
				goto out_pm;

			err = -EACCES;
			if (!node_isset(node, task_nodes))
				goto out_pm;

1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248
			pm[j].node = node;
		}

		/* End marker for this chunk */
		pm[chunk_nr_pages].node = MAX_NUMNODES;

		/* Migrate this chunk */
		err = do_move_page_to_node_array(mm, pm,
						 flags & MPOL_MF_MOVE_ALL);
		if (err < 0)
			goto out_pm;
1249 1250

		/* Return status information */
1251 1252
		for (j = 0; j < chunk_nr_pages; j++)
			if (put_user(pm[j].status, status + j + chunk_start)) {
1253
				err = -EFAULT;
1254 1255 1256 1257
				goto out_pm;
			}
	}
	err = 0;
1258 1259

out_pm:
1260
	free_page((unsigned long)pm);
1261 1262 1263 1264
out:
	return err;
}

1265
/*
1266
 * Determine the nodes of an array of pages and store it in an array of status.
1267
 */
1268 1269
static void do_pages_stat_array(struct mm_struct *mm, unsigned long nr_pages,
				const void __user **pages, int *status)
1270
{
1271 1272
	unsigned long i;

1273 1274
	down_read(&mm->mmap_sem);

1275
	for (i = 0; i < nr_pages; i++) {
1276
		unsigned long addr = (unsigned long)(*pages);
1277 1278
		struct vm_area_struct *vma;
		struct page *page;
1279
		int err = -EFAULT;
1280 1281

		vma = find_vma(mm, addr);
1282
		if (!vma || addr < vma->vm_start)
1283 1284
			goto set_status;

1285
		page = follow_page(vma, addr, 0);
1286 1287 1288 1289 1290

		err = PTR_ERR(page);
		if (IS_ERR(page))
			goto set_status;

1291 1292
		err = -ENOENT;
		/* Use PageReserved to check for zero page */
1293
		if (!page || PageReserved(page) || PageKsm(page))
1294 1295 1296 1297
			goto set_status;

		err = page_to_nid(page);
set_status:
1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318
		*status = err;

		pages++;
		status++;
	}

	up_read(&mm->mmap_sem);
}

/*
 * Determine the nodes of a user array of pages and store it in
 * a user array of status.
 */
static int do_pages_stat(struct mm_struct *mm, unsigned long nr_pages,
			 const void __user * __user *pages,
			 int __user *status)
{
#define DO_PAGES_STAT_CHUNK_NR 16
	const void __user *chunk_pages[DO_PAGES_STAT_CHUNK_NR];
	int chunk_status[DO_PAGES_STAT_CHUNK_NR];

1319 1320
	while (nr_pages) {
		unsigned long chunk_nr;
1321

1322 1323 1324 1325 1326 1327
		chunk_nr = nr_pages;
		if (chunk_nr > DO_PAGES_STAT_CHUNK_NR)
			chunk_nr = DO_PAGES_STAT_CHUNK_NR;

		if (copy_from_user(chunk_pages, pages, chunk_nr * sizeof(*chunk_pages)))
			break;
1328 1329 1330

		do_pages_stat_array(mm, chunk_nr, chunk_pages, chunk_status);

1331 1332
		if (copy_to_user(status, chunk_status, chunk_nr * sizeof(*status)))
			break;
1333

1334 1335 1336 1337 1338
		pages += chunk_nr;
		status += chunk_nr;
		nr_pages -= chunk_nr;
	}
	return nr_pages ? -EFAULT : 0;
1339 1340 1341 1342 1343 1344
}

/*
 * Move a list of pages in the address space of the currently executing
 * process.
 */
1345 1346 1347 1348
SYSCALL_DEFINE6(move_pages, pid_t, pid, unsigned long, nr_pages,
		const void __user * __user *, pages,
		const int __user *, nodes,
		int __user *, status, int, flags)
1349
{
1350
	const struct cred *cred = current_cred(), *tcred;
1351 1352
	struct task_struct *task;
	struct mm_struct *mm;
1353
	int err;
1354 1355 1356 1357 1358 1359 1360 1361 1362

	/* Check flags */
	if (flags & ~(MPOL_MF_MOVE|MPOL_MF_MOVE_ALL))
		return -EINVAL;

	if ((flags & MPOL_MF_MOVE_ALL) && !capable(CAP_SYS_NICE))
		return -EPERM;

	/* Find the mm_struct */
1363
	rcu_read_lock();
1364
	task = pid ? find_task_by_vpid(pid) : current;
1365
	if (!task) {
1366
		rcu_read_unlock();
1367 1368 1369
		return -ESRCH;
	}
	mm = get_task_mm(task);
1370
	rcu_read_unlock();
1371 1372 1373 1374 1375 1376 1377 1378 1379 1380

	if (!mm)
		return -EINVAL;

	/*
	 * Check if this process has the right to modify the specified
	 * process. The right exists if the process has administrative
	 * capabilities, superuser privileges or the same
	 * userid as the target process.
	 */
1381 1382
	rcu_read_lock();
	tcred = __task_cred(task);
1383 1384
	if (cred->euid != tcred->suid && cred->euid != tcred->uid &&
	    cred->uid  != tcred->suid && cred->uid  != tcred->uid &&
1385
	    !capable(CAP_SYS_NICE)) {
1386
		rcu_read_unlock();
1387
		err = -EPERM;
1388
		goto out;
1389
	}
1390
	rcu_read_unlock();
1391

1392 1393
 	err = security_task_movememory(task);
 	if (err)
1394
		goto out;
1395

1396 1397 1398 1399
	if (nodes) {
		err = do_pages_move(mm, task, nr_pages, pages, nodes, status,
				    flags);
	} else {
1400
		err = do_pages_stat(mm, nr_pages, pages, status);
1401 1402 1403 1404 1405 1406 1407
	}

out:
	mmput(mm);
	return err;
}

1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418
/*
 * Call migration functions in the vma_ops that may prepare
 * memory in a vm for migration. migration functions may perform
 * the migration for vmas that do not have an underlying page struct.
 */
int migrate_vmas(struct mm_struct *mm, const nodemask_t *to,
	const nodemask_t *from, unsigned long flags)
{
 	struct vm_area_struct *vma;
 	int err = 0;

1419
	for (vma = mm->mmap; vma && !err; vma = vma->vm_next) {
1420 1421 1422 1423 1424 1425 1426 1427
 		if (vma->vm_ops && vma->vm_ops->migrate) {
 			err = vma->vm_ops->migrate(vma, to, from, flags);
 			if (err)
 				break;
 		}
 	}
 	return err;
}
1428
#endif