Blame view
mm/migrate.c
33.4 KB
b20a35035
|
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> |
cde535359
|
12 |
* Christoph Lameter |
b20a35035
|
13 14 15 |
*/ #include <linux/migrate.h> |
b95f1b31b
|
16 |
#include <linux/export.h> |
b20a35035
|
17 |
#include <linux/swap.h> |
0697212a4
|
18 |
#include <linux/swapops.h> |
b20a35035
|
19 |
#include <linux/pagemap.h> |
e23ca00bf
|
20 |
#include <linux/buffer_head.h> |
b20a35035
|
21 |
#include <linux/mm_inline.h> |
b488893a3
|
22 |
#include <linux/nsproxy.h> |
b20a35035
|
23 |
#include <linux/pagevec.h> |
e9995ef97
|
24 |
#include <linux/ksm.h> |
b20a35035
|
25 26 27 28 |
#include <linux/rmap.h> #include <linux/topology.h> #include <linux/cpu.h> #include <linux/cpuset.h> |
04e62a29b
|
29 |
#include <linux/writeback.h> |
742755a1d
|
30 31 |
#include <linux/mempolicy.h> #include <linux/vmalloc.h> |
86c3a7645
|
32 |
#include <linux/security.h> |
8a9f3ccd2
|
33 |
#include <linux/memcontrol.h> |
4f5ca2657
|
34 |
#include <linux/syscalls.h> |
290408d4a
|
35 |
#include <linux/hugetlb.h> |
8e6ac7fab
|
36 |
#include <linux/hugetlb_cgroup.h> |
5a0e3ad6a
|
37 |
#include <linux/gfp.h> |
b20a35035
|
38 |
|
0d1836c36
|
39 |
#include <asm/tlbflush.h> |
b20a35035
|
40 |
#include "internal.h" |
b20a35035
|
41 |
/* |
742755a1d
|
42 |
* migrate_prep() needs to be called before we start compiling a list of pages |
748446bb6
|
43 44 |
* to be migrated using isolate_lru_page(). If scheduling work on other CPUs is * undesirable, use migrate_prep_local() |
b20a35035
|
45 46 47 |
*/ int migrate_prep(void) { |
b20a35035
|
48 49 50 51 52 53 54 55 56 57 |
/* * 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; } |
748446bb6
|
58 59 60 61 62 63 64 |
/* Do the necessary work of migrate_prep but not if it involves other CPUs */ int migrate_prep_local(void) { lru_add_drain(); return 0; } |
b20a35035
|
65 |
/* |
894bc3104
|
66 67 |
* Add isolated pages on the list back to the LRU under page lock * to avoid leaking evictable pages back onto unevictable list. |
b20a35035
|
68 |
*/ |
e13861d82
|
69 |
void putback_lru_pages(struct list_head *l) |
b20a35035
|
70 71 72 |
{ struct page *page; struct page *page2; |
b20a35035
|
73 74 |
list_for_each_entry_safe(page, page2, l, lru) { |
e24f0b8f7
|
75 |
list_del(&page->lru); |
a731286de
|
76 |
dec_zone_page_state(page, NR_ISOLATED_ANON + |
6c0b13519
|
77 |
page_is_file_cache(page)); |
894bc3104
|
78 |
putback_lru_page(page); |
b20a35035
|
79 |
} |
b20a35035
|
80 |
} |
0697212a4
|
81 82 83 |
/* * Restore a potential migration pte to a working pte entry */ |
e9995ef97
|
84 85 |
static int remove_migration_pte(struct page *new, struct vm_area_struct *vma, unsigned long addr, void *old) |
0697212a4
|
86 87 88 89 90 91 92 93 |
{ 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; |
290408d4a
|
94 95 96 97 98 99 100 101 102 |
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; |
0697212a4
|
103 |
|
290408d4a
|
104 105 106 |
pud = pud_offset(pgd, addr); if (!pud_present(*pud)) goto out; |
0697212a4
|
107 |
|
290408d4a
|
108 |
pmd = pmd_offset(pud, addr); |
500d65d47
|
109 110 |
if (pmd_trans_huge(*pmd)) goto out; |
290408d4a
|
111 112 |
if (!pmd_present(*pmd)) goto out; |
0697212a4
|
113 |
|
290408d4a
|
114 |
ptep = pte_offset_map(pmd, addr); |
0697212a4
|
115 |
|
486cf46f3
|
116 117 118 119 |
/* * Peek to check is_swap_pte() before taking ptlock? No, we * can race mremap's move_ptes(), which skips anon_vma lock. */ |
290408d4a
|
120 121 122 |
ptl = pte_lockptr(mm, pmd); } |
0697212a4
|
123 |
|
0697212a4
|
124 125 126 |
spin_lock(ptl); pte = *ptep; if (!is_swap_pte(pte)) |
e9995ef97
|
127 |
goto unlock; |
0697212a4
|
128 129 |
entry = pte_to_swp_entry(pte); |
e9995ef97
|
130 131 132 |
if (!is_migration_entry(entry) || migration_entry_to_page(entry) != old) goto unlock; |
0697212a4
|
133 |
|
0697212a4
|
134 135 136 137 |
get_page(new); pte = pte_mkold(mk_pte(new, vma->vm_page_prot)); if (is_write_migration_entry(entry)) pte = pte_mkwrite(pte); |
3ef8fd7f7
|
138 |
#ifdef CONFIG_HUGETLB_PAGE |
290408d4a
|
139 140 |
if (PageHuge(new)) pte = pte_mkhuge(pte); |
3ef8fd7f7
|
141 |
#endif |
97ee05246
|
142 |
flush_cache_page(vma, addr, pte_pfn(pte)); |
0697212a4
|
143 |
set_pte_at(mm, addr, ptep, pte); |
04e62a29b
|
144 |
|
290408d4a
|
145 146 147 148 149 150 |
if (PageHuge(new)) { if (PageAnon(new)) hugepage_add_anon_rmap(new, vma, addr); else page_dup_rmap(new); } else if (PageAnon(new)) |
04e62a29b
|
151 152 153 154 155 |
page_add_anon_rmap(new, vma, addr); else page_add_file_rmap(new); /* No need to invalidate - it was non-present before */ |
4b3073e1c
|
156 |
update_mmu_cache(vma, addr, ptep); |
e9995ef97
|
157 |
unlock: |
0697212a4
|
158 |
pte_unmap_unlock(ptep, ptl); |
e9995ef97
|
159 160 |
out: return SWAP_AGAIN; |
0697212a4
|
161 162 163 |
} /* |
04e62a29b
|
164 165 166 167 168 |
* 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) { |
e9995ef97
|
169 |
rmap_walk(new, remove_migration_pte, old); |
04e62a29b
|
170 171 172 |
} /* |
0697212a4
|
173 174 175 |
* 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. |
0697212a4
|
176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 |
*/ 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); |
e286781d5
|
195 196 197 198 199 200 201 202 203 |
/* * 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; |
0697212a4
|
204 205 206 207 208 209 210 |
pte_unmap_unlock(ptep, ptl); wait_on_page_locked(page); put_page(page); return; out: pte_unmap_unlock(ptep, ptl); } |
b969c4ab9
|
211 212 |
#ifdef CONFIG_BLOCK /* Returns true if all buffers are successfully locked */ |
a6bc32b89
|
213 214 |
static bool buffer_migrate_lock_buffers(struct buffer_head *head, enum migrate_mode mode) |
b969c4ab9
|
215 216 217 218 |
{ struct buffer_head *bh = head; /* Simple case, sync compaction */ |
a6bc32b89
|
219 |
if (mode != MIGRATE_ASYNC) { |
b969c4ab9
|
220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 |
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, |
a6bc32b89
|
255 |
enum migrate_mode mode) |
b969c4ab9
|
256 257 258 259 |
{ return true; } #endif /* CONFIG_BLOCK */ |
b20a35035
|
260 |
/* |
c3fcf8a5d
|
261 |
* Replace the page in the mapping. |
5b5c7120e
|
262 263 264 265 |
* * The number of remaining references must be: * 1 for anonymous pages without a mapping * 2 for pages with a mapping |
266cf658e
|
266 |
* 3 for pages with a mapping and PagePrivate/PagePrivate2 set. |
b20a35035
|
267 |
*/ |
2d1db3b11
|
268 |
static int migrate_page_move_mapping(struct address_space *mapping, |
b969c4ab9
|
269 |
struct page *newpage, struct page *page, |
a6bc32b89
|
270 |
struct buffer_head *head, enum migrate_mode mode) |
b20a35035
|
271 |
{ |
e286781d5
|
272 |
int expected_count; |
7cf9c2c76
|
273 |
void **pslot; |
b20a35035
|
274 |
|
6c5240ae7
|
275 |
if (!mapping) { |
0e8c7d0fd
|
276 |
/* Anonymous page without mapping */ |
6c5240ae7
|
277 278 279 280 |
if (page_count(page) != 1) return -EAGAIN; return 0; } |
19fd62312
|
281 |
spin_lock_irq(&mapping->tree_lock); |
b20a35035
|
282 |
|
7cf9c2c76
|
283 284 |
pslot = radix_tree_lookup_slot(&mapping->page_tree, page_index(page)); |
b20a35035
|
285 |
|
edcf4748c
|
286 |
expected_count = 2 + page_has_private(page); |
e286781d5
|
287 |
if (page_count(page) != expected_count || |
29c1f677d
|
288 |
radix_tree_deref_slot_protected(pslot, &mapping->tree_lock) != page) { |
19fd62312
|
289 |
spin_unlock_irq(&mapping->tree_lock); |
e23ca00bf
|
290 |
return -EAGAIN; |
b20a35035
|
291 |
} |
e286781d5
|
292 |
if (!page_freeze_refs(page, expected_count)) { |
19fd62312
|
293 |
spin_unlock_irq(&mapping->tree_lock); |
e286781d5
|
294 295 |
return -EAGAIN; } |
b20a35035
|
296 |
/* |
b969c4ab9
|
297 298 299 300 301 302 |
* 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. */ |
a6bc32b89
|
303 304 |
if (mode == MIGRATE_ASYNC && head && !buffer_migrate_lock_buffers(head, mode)) { |
b969c4ab9
|
305 306 307 308 309 310 |
page_unfreeze_refs(page, expected_count); spin_unlock_irq(&mapping->tree_lock); return -EAGAIN; } /* |
b20a35035
|
311 |
* Now we know that no one else is looking at the page. |
b20a35035
|
312 |
*/ |
7cf9c2c76
|
313 |
get_page(newpage); /* add cache reference */ |
b20a35035
|
314 315 316 317 |
if (PageSwapCache(page)) { SetPageSwapCache(newpage); set_page_private(newpage, page_private(page)); } |
7cf9c2c76
|
318 319 320 |
radix_tree_replace_slot(pslot, newpage); /* |
937a94c9d
|
321 322 |
* Drop cache reference from old page by unfreezing * to one less reference. |
7cf9c2c76
|
323 324 |
* We know this isn't the last reference. */ |
937a94c9d
|
325 |
page_unfreeze_refs(page, expected_count - 1); |
7cf9c2c76
|
326 |
|
0e8c7d0fd
|
327 328 329 330 331 332 333 334 335 336 337 338 |
/* * 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); |
99a15e21d
|
339 |
if (!PageSwapCache(page) && PageSwapBacked(page)) { |
4b02108ac
|
340 341 342 |
__dec_zone_page_state(page, NR_SHMEM); __inc_zone_page_state(newpage, NR_SHMEM); } |
19fd62312
|
343 |
spin_unlock_irq(&mapping->tree_lock); |
b20a35035
|
344 345 346 |
return 0; } |
b20a35035
|
347 348 |
/* |
290408d4a
|
349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 |
* 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 || |
29c1f677d
|
371 |
radix_tree_deref_slot_protected(pslot, &mapping->tree_lock) != page) { |
290408d4a
|
372 373 374 375 376 377 378 379 380 381 382 383 |
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); |
937a94c9d
|
384 |
page_unfreeze_refs(page, expected_count - 1); |
290408d4a
|
385 386 387 388 389 390 |
spin_unlock_irq(&mapping->tree_lock); return 0; } /* |
b20a35035
|
391 392 |
* Copy the page to its new location */ |
290408d4a
|
393 |
void migrate_page_copy(struct page *newpage, struct page *page) |
b20a35035
|
394 |
{ |
290408d4a
|
395 396 397 398 |
if (PageHuge(page)) copy_huge_page(newpage, page); else copy_highpage(newpage, page); |
b20a35035
|
399 400 401 402 403 404 405 |
if (PageError(page)) SetPageError(newpage); if (PageReferenced(page)) SetPageReferenced(newpage); if (PageUptodate(page)) SetPageUptodate(newpage); |
894bc3104
|
406 407 |
if (TestClearPageActive(page)) { VM_BUG_ON(PageUnevictable(page)); |
b20a35035
|
408 |
SetPageActive(newpage); |
418b27ef5
|
409 410 |
} else if (TestClearPageUnevictable(page)) SetPageUnevictable(newpage); |
b20a35035
|
411 412 413 414 415 416 417 |
if (PageChecked(page)) SetPageChecked(newpage); if (PageMappedToDisk(page)) SetPageMappedToDisk(newpage); if (PageDirty(page)) { clear_page_dirty_for_io(page); |
3a902c5f6
|
418 419 420 421 422 |
/* * 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. |
25985edce
|
423 |
* Whereas only part of our page may be dirty. |
3a902c5f6
|
424 |
*/ |
752dc185d
|
425 426 427 428 |
if (PageSwapBacked(page)) SetPageDirty(newpage); else __set_page_dirty_nobuffers(newpage); |
b20a35035
|
429 |
} |
b291f0003
|
430 |
mlock_migrate_page(newpage, page); |
e9995ef97
|
431 |
ksm_migrate_page(newpage, page); |
b291f0003
|
432 |
|
b20a35035
|
433 |
ClearPageSwapCache(page); |
b20a35035
|
434 435 |
ClearPagePrivate(page); set_page_private(page, 0); |
b20a35035
|
436 437 438 439 440 441 442 443 |
/* * If any waiters have accumulated on the new page then * wake them up. */ if (PageWriteback(newpage)) end_page_writeback(newpage); } |
b20a35035
|
444 |
|
1d8b85ccf
|
445 446 447 448 449 |
/************************************************************ * Migration functions ***********************************************************/ /* Always fail migration. Used for mappings that are not movable */ |
2d1db3b11
|
450 451 |
int fail_migrate_page(struct address_space *mapping, struct page *newpage, struct page *page) |
1d8b85ccf
|
452 453 454 455 |
{ return -EIO; } EXPORT_SYMBOL(fail_migrate_page); |
b20a35035
|
456 457 |
/* * Common logic to directly migrate a single page suitable for |
266cf658e
|
458 |
* pages that do not use PagePrivate/PagePrivate2. |
b20a35035
|
459 460 461 |
* * Pages are locked upon entry and exit. */ |
2d1db3b11
|
462 |
int migrate_page(struct address_space *mapping, |
a6bc32b89
|
463 464 |
struct page *newpage, struct page *page, enum migrate_mode mode) |
b20a35035
|
465 466 467 468 |
{ int rc; BUG_ON(PageWriteback(page)); /* Writeback must be complete */ |
a6bc32b89
|
469 |
rc = migrate_page_move_mapping(mapping, newpage, page, NULL, mode); |
b20a35035
|
470 471 472 473 474 |
if (rc) return rc; migrate_page_copy(newpage, page); |
b20a35035
|
475 476 477 |
return 0; } EXPORT_SYMBOL(migrate_page); |
9361401eb
|
478 |
#ifdef CONFIG_BLOCK |
b20a35035
|
479 |
/* |
1d8b85ccf
|
480 481 482 483 |
* Migration function for pages with buffers. This function can only be used * if the underlying filesystem guarantees that no other references to "page" * exist. */ |
2d1db3b11
|
484 |
int buffer_migrate_page(struct address_space *mapping, |
a6bc32b89
|
485 |
struct page *newpage, struct page *page, enum migrate_mode mode) |
1d8b85ccf
|
486 |
{ |
1d8b85ccf
|
487 488 |
struct buffer_head *bh, *head; int rc; |
1d8b85ccf
|
489 |
if (!page_has_buffers(page)) |
a6bc32b89
|
490 |
return migrate_page(mapping, newpage, page, mode); |
1d8b85ccf
|
491 492 |
head = page_buffers(page); |
a6bc32b89
|
493 |
rc = migrate_page_move_mapping(mapping, newpage, page, head, mode); |
1d8b85ccf
|
494 495 496 |
if (rc) return rc; |
b969c4ab9
|
497 498 499 500 501 |
/* * 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 */ |
a6bc32b89
|
502 503 |
if (mode != MIGRATE_ASYNC) BUG_ON(!buffer_migrate_lock_buffers(head, mode)); |
1d8b85ccf
|
504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 |
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); |
9361401eb
|
533 |
#endif |
1d8b85ccf
|
534 |
|
04e62a29b
|
535 536 537 538 |
/* * Writeback a page to clean the dirty state */ static int writeout(struct address_space *mapping, struct page *page) |
8351a6e47
|
539 |
{ |
04e62a29b
|
540 541 542 543 544 |
struct writeback_control wbc = { .sync_mode = WB_SYNC_NONE, .nr_to_write = 1, .range_start = 0, .range_end = LLONG_MAX, |
04e62a29b
|
545 546 547 548 549 550 551 552 553 554 555 |
.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; |
8351a6e47
|
556 |
/* |
04e62a29b
|
557 558 559 560 561 562 |
* 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. |
8351a6e47
|
563 |
*/ |
04e62a29b
|
564 |
remove_migration_ptes(page, page); |
8351a6e47
|
565 |
|
04e62a29b
|
566 |
rc = mapping->a_ops->writepage(page, &wbc); |
8351a6e47
|
567 |
|
04e62a29b
|
568 569 570 |
if (rc != AOP_WRITEPAGE_ACTIVATE) /* unlocked. Relock */ lock_page(page); |
bda8550de
|
571 |
return (rc < 0) ? -EIO : -EAGAIN; |
04e62a29b
|
572 573 574 575 576 577 |
} /* * Default handling if a filesystem does not provide a migration function. */ static int fallback_migrate_page(struct address_space *mapping, |
a6bc32b89
|
578 |
struct page *newpage, struct page *page, enum migrate_mode mode) |
04e62a29b
|
579 |
{ |
b969c4ab9
|
580 |
if (PageDirty(page)) { |
a6bc32b89
|
581 582 |
/* Only writeback pages in full synchronous migration */ if (mode != MIGRATE_SYNC) |
b969c4ab9
|
583 |
return -EBUSY; |
04e62a29b
|
584 |
return writeout(mapping, page); |
b969c4ab9
|
585 |
} |
8351a6e47
|
586 587 588 589 590 |
/* * Buffers may be managed in a filesystem specific way. * We must have no buffers or drop them. */ |
266cf658e
|
591 |
if (page_has_private(page) && |
8351a6e47
|
592 593 |
!try_to_release_page(page, GFP_KERNEL)) return -EAGAIN; |
a6bc32b89
|
594 |
return migrate_page(mapping, newpage, page, mode); |
8351a6e47
|
595 |
} |
1d8b85ccf
|
596 |
/* |
e24f0b8f7
|
597 598 599 600 601 |
* 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. |
894bc3104
|
602 603 604 605 |
* * Return value: * < 0 - error code * == 0 - success |
e24f0b8f7
|
606 |
*/ |
3fe2011ff
|
607 |
static int move_to_new_page(struct page *newpage, struct page *page, |
a6bc32b89
|
608 |
int remap_swapcache, enum migrate_mode mode) |
e24f0b8f7
|
609 610 611 612 613 614 615 616 617 |
{ 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. */ |
529ae9aaa
|
618 |
if (!trylock_page(newpage)) |
e24f0b8f7
|
619 620 621 622 623 |
BUG(); /* Prepare mapping for the new page.*/ newpage->index = page->index; newpage->mapping = page->mapping; |
b2e185384
|
624 625 |
if (PageSwapBacked(page)) SetPageSwapBacked(newpage); |
e24f0b8f7
|
626 627 628 |
mapping = page_mapping(page); if (!mapping) |
a6bc32b89
|
629 |
rc = migrate_page(mapping, newpage, page, mode); |
b969c4ab9
|
630 |
else if (mapping->a_ops->migratepage) |
e24f0b8f7
|
631 |
/* |
b969c4ab9
|
632 633 634 635 |
* 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. |
e24f0b8f7
|
636 |
*/ |
b969c4ab9
|
637 |
rc = mapping->a_ops->migratepage(mapping, |
a6bc32b89
|
638 |
newpage, page, mode); |
b969c4ab9
|
639 |
else |
a6bc32b89
|
640 |
rc = fallback_migrate_page(mapping, newpage, page, mode); |
e24f0b8f7
|
641 |
|
3fe2011ff
|
642 |
if (rc) { |
e24f0b8f7
|
643 |
newpage->mapping = NULL; |
3fe2011ff
|
644 645 646 |
} else { if (remap_swapcache) remove_migration_ptes(page, newpage); |
35512ecae
|
647 |
page->mapping = NULL; |
3fe2011ff
|
648 |
} |
e24f0b8f7
|
649 650 651 652 653 |
unlock_page(newpage); return rc; } |
0dabec93d
|
654 |
static int __unmap_and_move(struct page *page, struct page *newpage, |
a6bc32b89
|
655 |
int force, bool offlining, enum migrate_mode mode) |
e24f0b8f7
|
656 |
{ |
0dabec93d
|
657 |
int rc = -EAGAIN; |
3fe2011ff
|
658 |
int remap_swapcache = 1; |
56039efa1
|
659 |
struct mem_cgroup *mem; |
3f6c82728
|
660 |
struct anon_vma *anon_vma = NULL; |
95a402c38
|
661 |
|
529ae9aaa
|
662 |
if (!trylock_page(page)) { |
a6bc32b89
|
663 |
if (!force || mode == MIGRATE_ASYNC) |
0dabec93d
|
664 |
goto out; |
3e7d34497
|
665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 |
/* * 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) |
0dabec93d
|
680 |
goto out; |
3e7d34497
|
681 |
|
e24f0b8f7
|
682 683 |
lock_page(page); } |
62b61f611
|
684 685 686 687 688 689 690 691 692 693 694 695 696 |
/* * 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; } |
01b1ae63c
|
697 |
/* charge against new page */ |
0030f535a
|
698 |
mem_cgroup_prepare_migration(page, newpage, &mem); |
01b1ae63c
|
699 |
|
e24f0b8f7
|
700 |
if (PageWriteback(page)) { |
11bc82d67
|
701 |
/* |
a6bc32b89
|
702 703 704 705 |
* 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 |
11bc82d67
|
706 |
*/ |
a6bc32b89
|
707 |
if (mode != MIGRATE_SYNC) { |
11bc82d67
|
708 709 710 711 |
rc = -EBUSY; goto uncharge; } if (!force) |
01b1ae63c
|
712 |
goto uncharge; |
e24f0b8f7
|
713 714 |
wait_on_page_writeback(page); } |
e24f0b8f7
|
715 |
/* |
dc386d4d1
|
716 717 |
* 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. |
1ce82b69e
|
718 |
* This get_anon_vma() delays freeing anon_vma pointer until the end |
dc386d4d1
|
719 |
* of migration. File cache pages are no problem because of page_lock() |
989f89c57
|
720 721 |
* File Caches may use write_page() or lock_page() in migration, then, * just care Anon page here. |
dc386d4d1
|
722 |
*/ |
989f89c57
|
723 |
if (PageAnon(page)) { |
1ce82b69e
|
724 725 726 727 |
/* * Only page_lock_anon_vma() understands the subtleties of * getting a hold on an anon_vma from outside one of its mms. */ |
746b18d42
|
728 |
anon_vma = page_get_anon_vma(page); |
1ce82b69e
|
729 730 |
if (anon_vma) { /* |
746b18d42
|
731 |
* Anon page |
1ce82b69e
|
732 |
*/ |
1ce82b69e
|
733 |
} else if (PageSwapCache(page)) { |
3fe2011ff
|
734 735 736 737 738 739 740 741 742 743 744 745 746 747 |
/* * 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 { |
1ce82b69e
|
748 |
goto uncharge; |
3fe2011ff
|
749 |
} |
989f89c57
|
750 |
} |
62e1c5530
|
751 |
|
dc386d4d1
|
752 |
/* |
62e1c5530
|
753 754 755 756 757 758 759 760 761 762 |
* 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. |
e24f0b8f7
|
763 |
*/ |
62e1c5530
|
764 |
if (!page->mapping) { |
1ce82b69e
|
765 766 |
VM_BUG_ON(PageAnon(page)); if (page_has_private(page)) { |
62e1c5530
|
767 |
try_to_free_buffers(page); |
1ce82b69e
|
768 |
goto uncharge; |
62e1c5530
|
769 |
} |
abfc34881
|
770 |
goto skip_unmap; |
62e1c5530
|
771 |
} |
dc386d4d1
|
772 |
/* Establish migration ptes or remove ptes */ |
14fa31b89
|
773 |
try_to_unmap(page, TTU_MIGRATION|TTU_IGNORE_MLOCK|TTU_IGNORE_ACCESS); |
dc386d4d1
|
774 |
|
abfc34881
|
775 |
skip_unmap: |
e6a1530d6
|
776 |
if (!page_mapped(page)) |
a6bc32b89
|
777 |
rc = move_to_new_page(newpage, page, remap_swapcache, mode); |
e24f0b8f7
|
778 |
|
3fe2011ff
|
779 |
if (rc && remap_swapcache) |
e24f0b8f7
|
780 |
remove_migration_ptes(page, page); |
3f6c82728
|
781 782 |
/* Drop an anon_vma reference if we took one */ |
76545066c
|
783 |
if (anon_vma) |
9e60109f1
|
784 |
put_anon_vma(anon_vma); |
3f6c82728
|
785 |
|
01b1ae63c
|
786 |
uncharge: |
0030f535a
|
787 |
mem_cgroup_end_migration(mem, page, newpage, rc == 0); |
e24f0b8f7
|
788 789 |
unlock: unlock_page(page); |
0dabec93d
|
790 791 792 |
out: return rc; } |
95a402c38
|
793 |
|
0dabec93d
|
794 795 796 797 798 |
/* * 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, |
a6bc32b89
|
799 800 |
struct page *page, int force, bool offlining, enum migrate_mode mode) |
0dabec93d
|
801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 |
{ int rc = 0; int *result = NULL; struct page *newpage = get_new_page(page, private, &result); if (!newpage) return -ENOMEM; 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; |
a6bc32b89
|
817 |
rc = __unmap_and_move(page, newpage, force, offlining, mode); |
0dabec93d
|
818 |
out: |
e24f0b8f7
|
819 |
if (rc != -EAGAIN) { |
0dabec93d
|
820 821 822 823 824 825 826 |
/* * 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); |
a731286de
|
827 |
dec_zone_page_state(page, NR_ISOLATED_ANON + |
6c0b13519
|
828 |
page_is_file_cache(page)); |
894bc3104
|
829 |
putback_lru_page(page); |
e24f0b8f7
|
830 |
} |
95a402c38
|
831 832 833 834 |
/* * Move the new page to the LRU. If migration was not successful * then this will free the page. */ |
894bc3104
|
835 |
putback_lru_page(newpage); |
742755a1d
|
836 837 838 839 840 841 |
if (result) { if (rc) *result = rc; else *result = page_to_nid(newpage); } |
e24f0b8f7
|
842 843 844 845 |
return rc; } /* |
290408d4a
|
846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 |
* 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, |
a6bc32b89
|
865 866 |
int force, bool offlining, enum migrate_mode mode) |
290408d4a
|
867 868 869 870 |
{ int rc = 0; int *result = NULL; struct page *new_hpage = get_new_page(hpage, private, &result); |
290408d4a
|
871 872 873 874 875 876 877 878 |
struct anon_vma *anon_vma = NULL; if (!new_hpage) return -ENOMEM; rc = -EAGAIN; if (!trylock_page(hpage)) { |
a6bc32b89
|
879 |
if (!force || mode != MIGRATE_SYNC) |
290408d4a
|
880 881 882 |
goto out; lock_page(hpage); } |
746b18d42
|
883 884 |
if (PageAnon(hpage)) anon_vma = page_get_anon_vma(hpage); |
290408d4a
|
885 886 887 888 |
try_to_unmap(hpage, TTU_MIGRATION|TTU_IGNORE_MLOCK|TTU_IGNORE_ACCESS); if (!page_mapped(hpage)) |
a6bc32b89
|
889 |
rc = move_to_new_page(new_hpage, hpage, 1, mode); |
290408d4a
|
890 891 892 |
if (rc) remove_migration_ptes(hpage, hpage); |
fd4a4663d
|
893 |
if (anon_vma) |
9e60109f1
|
894 |
put_anon_vma(anon_vma); |
8e6ac7fab
|
895 896 897 |
if (!rc) hugetlb_cgroup_migrate(hpage, new_hpage); |
290408d4a
|
898 |
unlock_page(hpage); |
09761333e
|
899 |
out: |
290408d4a
|
900 |
put_page(new_hpage); |
290408d4a
|
901 902 903 904 905 906 907 908 909 910 |
if (result) { if (rc) *result = rc; else *result = page_to_nid(new_hpage); } return rc; } /* |
b20a35035
|
911 912 |
* migrate_pages * |
95a402c38
|
913 914 915 |
* 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. |
b20a35035
|
916 917 918 |
* * The function returns after 10 attempts or if no pages * are movable anymore because to has become empty |
cf608ac19
|
919 920 |
* or no retryable pages exist anymore. * Caller should call putback_lru_pages to return pages to the LRU |
28bd65781
|
921 |
* or free list only if ret != 0. |
b20a35035
|
922 |
* |
95a402c38
|
923 |
* Return: Number of pages not migrated or error code. |
b20a35035
|
924 |
*/ |
95a402c38
|
925 |
int migrate_pages(struct list_head *from, |
7f0f24967
|
926 |
new_page_t get_new_page, unsigned long private, bool offlining, |
a6bc32b89
|
927 |
enum migrate_mode mode) |
b20a35035
|
928 |
{ |
e24f0b8f7
|
929 |
int retry = 1; |
b20a35035
|
930 931 932 933 934 935 936 937 938 |
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; |
e24f0b8f7
|
939 940 |
for(pass = 0; pass < 10 && retry; pass++) { retry = 0; |
b20a35035
|
941 |
|
e24f0b8f7
|
942 |
list_for_each_entry_safe(page, page2, from, lru) { |
e24f0b8f7
|
943 |
cond_resched(); |
2d1db3b11
|
944 |
|
95a402c38
|
945 |
rc = unmap_and_move(get_new_page, private, |
77f1fe6b0
|
946 |
page, pass > 2, offlining, |
a6bc32b89
|
947 |
mode); |
2d1db3b11
|
948 |
|
e24f0b8f7
|
949 |
switch(rc) { |
95a402c38
|
950 951 |
case -ENOMEM: goto out; |
e24f0b8f7
|
952 |
case -EAGAIN: |
2d1db3b11
|
953 |
retry++; |
e24f0b8f7
|
954 955 |
break; case 0: |
e24f0b8f7
|
956 957 |
break; default: |
2d1db3b11
|
958 |
/* Permanent failure */ |
2d1db3b11
|
959 |
nr_failed++; |
e24f0b8f7
|
960 |
break; |
2d1db3b11
|
961 |
} |
b20a35035
|
962 963 |
} } |
95a402c38
|
964 965 |
rc = 0; out: |
b20a35035
|
966 967 |
if (!swapwrite) current->flags &= ~PF_SWAPWRITE; |
95a402c38
|
968 969 |
if (rc) return rc; |
b20a35035
|
970 |
|
95a402c38
|
971 |
return nr_failed + retry; |
b20a35035
|
972 |
} |
95a402c38
|
973 |
|
189ebff28
|
974 975 976 |
int migrate_huge_page(struct page *hpage, new_page_t get_new_page, unsigned long private, bool offlining, enum migrate_mode mode) |
290408d4a
|
977 |
{ |
189ebff28
|
978 979 980 981 982 983 984 985 986 987 988 |
int pass, rc; for (pass = 0; pass < 10; pass++) { rc = unmap_and_move_huge_page(get_new_page, private, hpage, pass > 2, offlining, mode); switch (rc) { case -ENOMEM: goto out; case -EAGAIN: /* try again */ |
290408d4a
|
989 |
cond_resched(); |
189ebff28
|
990 991 992 993 994 995 |
break; case 0: goto out; default: rc = -EIO; goto out; |
290408d4a
|
996 997 |
} } |
290408d4a
|
998 |
out: |
189ebff28
|
999 |
return rc; |
290408d4a
|
1000 |
} |
742755a1d
|
1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 |
#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; |
6484eb3e2
|
1024 |
return alloc_pages_exact_node(pm->node, |
769848c03
|
1025 |
GFP_HIGHUSER_MOVABLE | GFP_THISNODE, 0); |
742755a1d
|
1026 1027 1028 1029 1030 1031 |
} /* * 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. |
5e9a0f023
|
1032 |
* The pm array ends with node = MAX_NUMNODES. |
742755a1d
|
1033 |
*/ |
5e9a0f023
|
1034 1035 1036 |
static int do_move_page_to_node_array(struct mm_struct *mm, struct page_to_node *pm, int migrate_all) |
742755a1d
|
1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 |
{ int err; struct page_to_node *pp; LIST_HEAD(pagelist); down_read(&mm->mmap_sem); /* * Build a list of pages to migrate */ |
742755a1d
|
1047 1048 1049 |
for (pp = pm; pp->node != MAX_NUMNODES; pp++) { struct vm_area_struct *vma; struct page *page; |
742755a1d
|
1050 1051 |
err = -EFAULT; vma = find_vma(mm, pp->addr); |
70384dc6d
|
1052 |
if (!vma || pp->addr < vma->vm_start || !vma_migratable(vma)) |
742755a1d
|
1053 |
goto set_status; |
500d65d47
|
1054 |
page = follow_page(vma, pp->addr, FOLL_GET|FOLL_SPLIT); |
89f5b7da2
|
1055 1056 1057 1058 |
err = PTR_ERR(page); if (IS_ERR(page)) goto set_status; |
742755a1d
|
1059 1060 1061 |
err = -ENOENT; if (!page) goto set_status; |
62b61f611
|
1062 1063 |
/* Use PageReserved to check for zero page */ if (PageReserved(page) || PageKsm(page)) |
742755a1d
|
1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 |
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; |
62695a84e
|
1079 |
err = isolate_lru_page(page); |
6d9c285a6
|
1080 |
if (!err) { |
62695a84e
|
1081 |
list_add_tail(&page->lru, &pagelist); |
6d9c285a6
|
1082 1083 1084 |
inc_zone_page_state(page, NR_ISOLATED_ANON + page_is_file_cache(page)); } |
742755a1d
|
1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 |
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; } |
e78bbfa82
|
1095 |
err = 0; |
cf608ac19
|
1096 |
if (!list_empty(&pagelist)) { |
742755a1d
|
1097 |
err = migrate_pages(&pagelist, new_page_node, |
a6bc32b89
|
1098 |
(unsigned long)pm, 0, MIGRATE_SYNC); |
cf608ac19
|
1099 1100 1101 |
if (err) putback_lru_pages(&pagelist); } |
742755a1d
|
1102 1103 1104 1105 1106 1107 |
up_read(&mm->mmap_sem); return err; } /* |
5e9a0f023
|
1108 1109 1110 |
* Migrate an array of page address onto an array of nodes and fill * the corresponding array of status. */ |
3268c63ed
|
1111 |
static int do_pages_move(struct mm_struct *mm, nodemask_t task_nodes, |
5e9a0f023
|
1112 1113 1114 1115 1116 |
unsigned long nr_pages, const void __user * __user *pages, const int __user *nodes, int __user *status, int flags) { |
3140a2273
|
1117 |
struct page_to_node *pm; |
3140a2273
|
1118 1119 1120 |
unsigned long chunk_nr_pages; unsigned long chunk_start; int err; |
5e9a0f023
|
1121 |
|
3140a2273
|
1122 1123 1124 |
err = -ENOMEM; pm = (struct page_to_node *)__get_free_page(GFP_KERNEL); if (!pm) |
5e9a0f023
|
1125 |
goto out; |
35282a2de
|
1126 1127 |
migrate_prep(); |
5e9a0f023
|
1128 |
/* |
3140a2273
|
1129 1130 |
* Store a chunk of page_to_node array in a page, * but keep the last one as a marker |
5e9a0f023
|
1131 |
*/ |
3140a2273
|
1132 |
chunk_nr_pages = (PAGE_SIZE / sizeof(struct page_to_node)) - 1; |
5e9a0f023
|
1133 |
|
3140a2273
|
1134 1135 1136 1137 |
for (chunk_start = 0; chunk_start < nr_pages; chunk_start += chunk_nr_pages) { int j; |
5e9a0f023
|
1138 |
|
3140a2273
|
1139 1140 1141 1142 1143 1144 |
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; |
5e9a0f023
|
1145 |
int node; |
3140a2273
|
1146 1147 1148 1149 1150 1151 |
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)) |
5e9a0f023
|
1152 1153 1154 |
goto out_pm; err = -ENODEV; |
6f5a55f1a
|
1155 1156 |
if (node < 0 || node >= MAX_NUMNODES) goto out_pm; |
5e9a0f023
|
1157 1158 1159 1160 1161 1162 |
if (!node_state(node, N_HIGH_MEMORY)) goto out_pm; err = -EACCES; if (!node_isset(node, task_nodes)) goto out_pm; |
3140a2273
|
1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 |
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; |
5e9a0f023
|
1174 |
|
5e9a0f023
|
1175 |
/* Return status information */ |
3140a2273
|
1176 1177 |
for (j = 0; j < chunk_nr_pages; j++) if (put_user(pm[j].status, status + j + chunk_start)) { |
5e9a0f023
|
1178 |
err = -EFAULT; |
3140a2273
|
1179 1180 1181 1182 |
goto out_pm; } } err = 0; |
5e9a0f023
|
1183 1184 |
out_pm: |
3140a2273
|
1185 |
free_page((unsigned long)pm); |
5e9a0f023
|
1186 1187 1188 1189 1190 |
out: return err; } /* |
2f007e74b
|
1191 |
* Determine the nodes of an array of pages and store it in an array of status. |
742755a1d
|
1192 |
*/ |
80bba1290
|
1193 1194 |
static void do_pages_stat_array(struct mm_struct *mm, unsigned long nr_pages, const void __user **pages, int *status) |
742755a1d
|
1195 |
{ |
2f007e74b
|
1196 |
unsigned long i; |
2f007e74b
|
1197 |
|
742755a1d
|
1198 |
down_read(&mm->mmap_sem); |
2f007e74b
|
1199 |
for (i = 0; i < nr_pages; i++) { |
80bba1290
|
1200 |
unsigned long addr = (unsigned long)(*pages); |
742755a1d
|
1201 1202 |
struct vm_area_struct *vma; struct page *page; |
c095adbc2
|
1203 |
int err = -EFAULT; |
2f007e74b
|
1204 1205 |
vma = find_vma(mm, addr); |
70384dc6d
|
1206 |
if (!vma || addr < vma->vm_start) |
742755a1d
|
1207 |
goto set_status; |
2f007e74b
|
1208 |
page = follow_page(vma, addr, 0); |
89f5b7da2
|
1209 1210 1211 1212 |
err = PTR_ERR(page); if (IS_ERR(page)) goto set_status; |
742755a1d
|
1213 1214 |
err = -ENOENT; /* Use PageReserved to check for zero page */ |
62b61f611
|
1215 |
if (!page || PageReserved(page) || PageKsm(page)) |
742755a1d
|
1216 1217 1218 1219 |
goto set_status; err = page_to_nid(page); set_status: |
80bba1290
|
1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 |
*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]; |
80bba1290
|
1240 |
|
87b8d1ade
|
1241 1242 |
while (nr_pages) { unsigned long chunk_nr; |
80bba1290
|
1243 |
|
87b8d1ade
|
1244 1245 1246 1247 1248 1249 |
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; |
80bba1290
|
1250 1251 |
do_pages_stat_array(mm, chunk_nr, chunk_pages, chunk_status); |
87b8d1ade
|
1252 1253 |
if (copy_to_user(status, chunk_status, chunk_nr * sizeof(*status))) break; |
742755a1d
|
1254 |
|
87b8d1ade
|
1255 1256 1257 1258 1259 |
pages += chunk_nr; status += chunk_nr; nr_pages -= chunk_nr; } return nr_pages ? -EFAULT : 0; |
742755a1d
|
1260 1261 1262 1263 1264 1265 |
} /* * Move a list of pages in the address space of the currently executing * process. */ |
938bb9f5e
|
1266 1267 1268 1269 |
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) |
742755a1d
|
1270 |
{ |
c69e8d9c0
|
1271 |
const struct cred *cred = current_cred(), *tcred; |
742755a1d
|
1272 |
struct task_struct *task; |
742755a1d
|
1273 |
struct mm_struct *mm; |
5e9a0f023
|
1274 |
int err; |
3268c63ed
|
1275 |
nodemask_t task_nodes; |
742755a1d
|
1276 1277 1278 1279 1280 1281 1282 1283 1284 |
/* 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 */ |
a879bf582
|
1285 |
rcu_read_lock(); |
228ebcbe6
|
1286 |
task = pid ? find_task_by_vpid(pid) : current; |
742755a1d
|
1287 |
if (!task) { |
a879bf582
|
1288 |
rcu_read_unlock(); |
742755a1d
|
1289 1290 |
return -ESRCH; } |
3268c63ed
|
1291 |
get_task_struct(task); |
742755a1d
|
1292 1293 1294 1295 1296 1297 1298 |
/* * 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. */ |
c69e8d9c0
|
1299 |
tcred = __task_cred(task); |
b38a86eb1
|
1300 1301 |
if (!uid_eq(cred->euid, tcred->suid) && !uid_eq(cred->euid, tcred->uid) && !uid_eq(cred->uid, tcred->suid) && !uid_eq(cred->uid, tcred->uid) && |
742755a1d
|
1302 |
!capable(CAP_SYS_NICE)) { |
c69e8d9c0
|
1303 |
rcu_read_unlock(); |
742755a1d
|
1304 |
err = -EPERM; |
5e9a0f023
|
1305 |
goto out; |
742755a1d
|
1306 |
} |
c69e8d9c0
|
1307 |
rcu_read_unlock(); |
742755a1d
|
1308 |
|
86c3a7645
|
1309 1310 |
err = security_task_movememory(task); if (err) |
5e9a0f023
|
1311 |
goto out; |
86c3a7645
|
1312 |
|
3268c63ed
|
1313 1314 1315 |
task_nodes = cpuset_mems_allowed(task); mm = get_task_mm(task); put_task_struct(task); |
6e8b09eaf
|
1316 1317 1318 1319 1320 1321 1322 1323 |
if (!mm) return -EINVAL; if (nodes) err = do_pages_move(mm, task_nodes, nr_pages, pages, nodes, status, flags); else err = do_pages_stat(mm, nr_pages, pages, status); |
742755a1d
|
1324 |
|
742755a1d
|
1325 1326 |
mmput(mm); return err; |
3268c63ed
|
1327 1328 1329 1330 |
out: put_task_struct(task); return err; |
742755a1d
|
1331 |
} |
742755a1d
|
1332 |
|
7b2259b3e
|
1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 |
/* * 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; |
1001c9fb8
|
1343 |
for (vma = mm->mmap; vma && !err; vma = vma->vm_next) { |
7b2259b3e
|
1344 1345 1346 1347 1348 1349 1350 1351 |
if (vma->vm_ops && vma->vm_ops->migrate) { err = vma->vm_ops->migrate(vma, to, from, flags); if (err) break; } } return err; } |
83d1674a9
|
1352 |
#endif |