Blame view
mm/migrate.c
26.8 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 16 17 |
*/ #include <linux/migrate.h> #include <linux/module.h> #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> |
5a0e3ad6a
|
35 |
#include <linux/gfp.h> |
b20a35035
|
36 37 |
#include "internal.h" |
b20a35035
|
38 39 40 |
#define lru_to_page(_head) (list_entry((_head)->prev, struct page, lru)) /* |
742755a1d
|
41 |
* migrate_prep() needs to be called before we start compiling a list of pages |
748446bb6
|
42 43 |
* to be migrated using isolate_lru_page(). If scheduling work on other CPUs is * undesirable, use migrate_prep_local() |
b20a35035
|
44 45 46 |
*/ int migrate_prep(void) { |
b20a35035
|
47 48 49 50 51 52 53 54 55 56 |
/* * 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
|
57 58 59 60 61 62 63 |
/* 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
|
64 |
/* |
894bc3104
|
65 66 |
* Add isolated pages on the list back to the LRU under page lock * to avoid leaking evictable pages back onto unevictable list. |
b20a35035
|
67 |
*/ |
e13861d82
|
68 |
void putback_lru_pages(struct list_head *l) |
b20a35035
|
69 70 71 |
{ struct page *page; struct page *page2; |
b20a35035
|
72 73 |
list_for_each_entry_safe(page, page2, l, lru) { |
e24f0b8f7
|
74 |
list_del(&page->lru); |
a731286de
|
75 |
dec_zone_page_state(page, NR_ISOLATED_ANON + |
6c0b13519
|
76 |
page_is_file_cache(page)); |
894bc3104
|
77 |
putback_lru_page(page); |
b20a35035
|
78 |
} |
b20a35035
|
79 |
} |
0697212a4
|
80 81 82 |
/* * Restore a potential migration pte to a working pte entry */ |
e9995ef97
|
83 84 |
static int remove_migration_pte(struct page *new, struct vm_area_struct *vma, unsigned long addr, void *old) |
0697212a4
|
85 86 87 88 89 90 91 92 93 94 95 |
{ 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; pgd = pgd_offset(mm, addr); if (!pgd_present(*pgd)) |
e9995ef97
|
96 |
goto out; |
0697212a4
|
97 98 99 |
pud = pud_offset(pgd, addr); if (!pud_present(*pud)) |
e9995ef97
|
100 |
goto out; |
0697212a4
|
101 102 103 |
pmd = pmd_offset(pud, addr); if (!pmd_present(*pmd)) |
e9995ef97
|
104 |
goto out; |
0697212a4
|
105 106 107 108 109 |
ptep = pte_offset_map(pmd, addr); if (!is_swap_pte(*ptep)) { pte_unmap(ptep); |
e9995ef97
|
110 |
goto out; |
0697212a4
|
111 112 113 114 115 116 |
} ptl = pte_lockptr(mm, pmd); spin_lock(ptl); pte = *ptep; if (!is_swap_pte(pte)) |
e9995ef97
|
117 |
goto unlock; |
0697212a4
|
118 119 |
entry = pte_to_swp_entry(pte); |
e9995ef97
|
120 121 122 |
if (!is_migration_entry(entry) || migration_entry_to_page(entry) != old) goto unlock; |
0697212a4
|
123 |
|
0697212a4
|
124 125 126 127 |
get_page(new); pte = pte_mkold(mk_pte(new, vma->vm_page_prot)); if (is_write_migration_entry(entry)) pte = pte_mkwrite(pte); |
97ee05246
|
128 |
flush_cache_page(vma, addr, pte_pfn(pte)); |
0697212a4
|
129 |
set_pte_at(mm, addr, ptep, pte); |
04e62a29b
|
130 131 132 133 134 135 136 |
if (PageAnon(new)) page_add_anon_rmap(new, vma, addr); else page_add_file_rmap(new); /* No need to invalidate - it was non-present before */ |
4b3073e1c
|
137 |
update_mmu_cache(vma, addr, ptep); |
e9995ef97
|
138 |
unlock: |
0697212a4
|
139 |
pte_unmap_unlock(ptep, ptl); |
e9995ef97
|
140 141 |
out: return SWAP_AGAIN; |
0697212a4
|
142 143 144 |
} /* |
04e62a29b
|
145 146 147 148 149 |
* 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
|
150 |
rmap_walk(new, remove_migration_pte, old); |
04e62a29b
|
151 152 153 |
} /* |
0697212a4
|
154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 |
* 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. * * This function is called from do_swap_page(). */ 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
|
178 179 180 181 182 183 184 185 186 |
/* * 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
|
187 188 189 190 191 192 193 |
pte_unmap_unlock(ptep, ptl); wait_on_page_locked(page); put_page(page); return; out: pte_unmap_unlock(ptep, ptl); } |
b20a35035
|
194 |
/* |
c3fcf8a5d
|
195 |
* Replace the page in the mapping. |
5b5c7120e
|
196 197 198 199 |
* * The number of remaining references must be: * 1 for anonymous pages without a mapping * 2 for pages with a mapping |
266cf658e
|
200 |
* 3 for pages with a mapping and PagePrivate/PagePrivate2 set. |
b20a35035
|
201 |
*/ |
2d1db3b11
|
202 203 |
static int migrate_page_move_mapping(struct address_space *mapping, struct page *newpage, struct page *page) |
b20a35035
|
204 |
{ |
e286781d5
|
205 |
int expected_count; |
7cf9c2c76
|
206 |
void **pslot; |
b20a35035
|
207 |
|
6c5240ae7
|
208 |
if (!mapping) { |
0e8c7d0fd
|
209 |
/* Anonymous page without mapping */ |
6c5240ae7
|
210 211 212 213 |
if (page_count(page) != 1) return -EAGAIN; return 0; } |
19fd62312
|
214 |
spin_lock_irq(&mapping->tree_lock); |
b20a35035
|
215 |
|
7cf9c2c76
|
216 217 |
pslot = radix_tree_lookup_slot(&mapping->page_tree, page_index(page)); |
b20a35035
|
218 |
|
edcf4748c
|
219 |
expected_count = 2 + page_has_private(page); |
e286781d5
|
220 |
if (page_count(page) != expected_count || |
7cf9c2c76
|
221 |
(struct page *)radix_tree_deref_slot(pslot) != page) { |
19fd62312
|
222 |
spin_unlock_irq(&mapping->tree_lock); |
e23ca00bf
|
223 |
return -EAGAIN; |
b20a35035
|
224 |
} |
e286781d5
|
225 |
if (!page_freeze_refs(page, expected_count)) { |
19fd62312
|
226 |
spin_unlock_irq(&mapping->tree_lock); |
e286781d5
|
227 228 |
return -EAGAIN; } |
b20a35035
|
229 230 |
/* * Now we know that no one else is looking at the page. |
b20a35035
|
231 |
*/ |
7cf9c2c76
|
232 |
get_page(newpage); /* add cache reference */ |
b20a35035
|
233 234 235 236 |
if (PageSwapCache(page)) { SetPageSwapCache(newpage); set_page_private(newpage, page_private(page)); } |
7cf9c2c76
|
237 |
radix_tree_replace_slot(pslot, newpage); |
e286781d5
|
238 |
page_unfreeze_refs(page, expected_count); |
7cf9c2c76
|
239 240 241 242 |
/* * Drop cache reference from old page. * We know this isn't the last reference. */ |
b20a35035
|
243 |
__put_page(page); |
7cf9c2c76
|
244 |
|
0e8c7d0fd
|
245 246 247 248 249 250 251 252 253 254 255 256 |
/* * 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); |
4b02108ac
|
257 258 259 260 |
if (PageSwapBacked(page)) { __dec_zone_page_state(page, NR_SHMEM); __inc_zone_page_state(newpage, NR_SHMEM); } |
19fd62312
|
261 |
spin_unlock_irq(&mapping->tree_lock); |
b20a35035
|
262 263 264 |
return 0; } |
b20a35035
|
265 266 267 268 |
/* * Copy the page to its new location */ |
e7340f733
|
269 |
static void migrate_page_copy(struct page *newpage, struct page *page) |
b20a35035
|
270 271 272 273 274 275 276 277 278 |
{ copy_highpage(newpage, page); if (PageError(page)) SetPageError(newpage); if (PageReferenced(page)) SetPageReferenced(newpage); if (PageUptodate(page)) SetPageUptodate(newpage); |
894bc3104
|
279 280 |
if (TestClearPageActive(page)) { VM_BUG_ON(PageUnevictable(page)); |
b20a35035
|
281 |
SetPageActive(newpage); |
418b27ef5
|
282 283 |
} else if (TestClearPageUnevictable(page)) SetPageUnevictable(newpage); |
b20a35035
|
284 285 286 287 288 289 290 |
if (PageChecked(page)) SetPageChecked(newpage); if (PageMappedToDisk(page)) SetPageMappedToDisk(newpage); if (PageDirty(page)) { clear_page_dirty_for_io(page); |
3a902c5f6
|
291 292 293 294 295 296 297 298 |
/* * 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. * Wheras only part of our page may be dirty. */ __set_page_dirty_nobuffers(newpage); |
b20a35035
|
299 |
} |
b291f0003
|
300 |
mlock_migrate_page(newpage, page); |
e9995ef97
|
301 |
ksm_migrate_page(newpage, page); |
b291f0003
|
302 |
|
b20a35035
|
303 |
ClearPageSwapCache(page); |
b20a35035
|
304 305 306 307 308 309 310 311 312 313 314 |
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); } |
b20a35035
|
315 |
|
1d8b85ccf
|
316 317 318 319 320 |
/************************************************************ * Migration functions ***********************************************************/ /* Always fail migration. Used for mappings that are not movable */ |
2d1db3b11
|
321 322 |
int fail_migrate_page(struct address_space *mapping, struct page *newpage, struct page *page) |
1d8b85ccf
|
323 324 325 326 |
{ return -EIO; } EXPORT_SYMBOL(fail_migrate_page); |
b20a35035
|
327 328 |
/* * Common logic to directly migrate a single page suitable for |
266cf658e
|
329 |
* pages that do not use PagePrivate/PagePrivate2. |
b20a35035
|
330 331 332 |
* * Pages are locked upon entry and exit. */ |
2d1db3b11
|
333 334 |
int migrate_page(struct address_space *mapping, struct page *newpage, struct page *page) |
b20a35035
|
335 336 337 338 |
{ int rc; BUG_ON(PageWriteback(page)); /* Writeback must be complete */ |
2d1db3b11
|
339 |
rc = migrate_page_move_mapping(mapping, newpage, page); |
b20a35035
|
340 341 342 343 344 |
if (rc) return rc; migrate_page_copy(newpage, page); |
b20a35035
|
345 346 347 |
return 0; } EXPORT_SYMBOL(migrate_page); |
9361401eb
|
348 |
#ifdef CONFIG_BLOCK |
b20a35035
|
349 |
/* |
1d8b85ccf
|
350 351 352 353 |
* 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
|
354 355 |
int buffer_migrate_page(struct address_space *mapping, struct page *newpage, struct page *page) |
1d8b85ccf
|
356 |
{ |
1d8b85ccf
|
357 358 |
struct buffer_head *bh, *head; int rc; |
1d8b85ccf
|
359 |
if (!page_has_buffers(page)) |
2d1db3b11
|
360 |
return migrate_page(mapping, newpage, page); |
1d8b85ccf
|
361 362 |
head = page_buffers(page); |
2d1db3b11
|
363 |
rc = migrate_page_move_mapping(mapping, newpage, page); |
1d8b85ccf
|
364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 |
if (rc) return rc; bh = head; do { get_bh(bh); lock_buffer(bh); bh = bh->b_this_page; } while (bh != head); 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
|
404 |
#endif |
1d8b85ccf
|
405 |
|
04e62a29b
|
406 407 408 409 |
/* * Writeback a page to clean the dirty state */ static int writeout(struct address_space *mapping, struct page *page) |
8351a6e47
|
410 |
{ |
04e62a29b
|
411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 |
struct writeback_control wbc = { .sync_mode = WB_SYNC_NONE, .nr_to_write = 1, .range_start = 0, .range_end = LLONG_MAX, .nonblocking = 1, .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
|
428 |
/* |
04e62a29b
|
429 430 431 432 433 434 |
* 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
|
435 |
*/ |
04e62a29b
|
436 |
remove_migration_ptes(page, page); |
8351a6e47
|
437 |
|
04e62a29b
|
438 |
rc = mapping->a_ops->writepage(page, &wbc); |
8351a6e47
|
439 |
|
04e62a29b
|
440 441 442 |
if (rc != AOP_WRITEPAGE_ACTIVATE) /* unlocked. Relock */ lock_page(page); |
bda8550de
|
443 |
return (rc < 0) ? -EIO : -EAGAIN; |
04e62a29b
|
444 445 446 447 448 449 450 451 452 453 |
} /* * Default handling if a filesystem does not provide a migration function. */ static int fallback_migrate_page(struct address_space *mapping, struct page *newpage, struct page *page) { if (PageDirty(page)) return writeout(mapping, page); |
8351a6e47
|
454 455 456 457 458 |
/* * Buffers may be managed in a filesystem specific way. * We must have no buffers or drop them. */ |
266cf658e
|
459 |
if (page_has_private(page) && |
8351a6e47
|
460 461 462 463 464 |
!try_to_release_page(page, GFP_KERNEL)) return -EAGAIN; return migrate_page(mapping, newpage, page); } |
1d8b85ccf
|
465 |
/* |
e24f0b8f7
|
466 467 468 469 470 |
* 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
|
471 472 473 474 |
* * Return value: * < 0 - error code * == 0 - success |
e24f0b8f7
|
475 |
*/ |
3fe2011ff
|
476 477 |
static int move_to_new_page(struct page *newpage, struct page *page, int remap_swapcache) |
e24f0b8f7
|
478 479 480 481 482 483 484 485 486 |
{ 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
|
487 |
if (!trylock_page(newpage)) |
e24f0b8f7
|
488 489 490 491 492 |
BUG(); /* Prepare mapping for the new page.*/ newpage->index = page->index; newpage->mapping = page->mapping; |
b2e185384
|
493 494 |
if (PageSwapBacked(page)) SetPageSwapBacked(newpage); |
e24f0b8f7
|
495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 |
mapping = page_mapping(page); if (!mapping) rc = migrate_page(mapping, newpage, page); else if (mapping->a_ops->migratepage) /* * Most pages have a mapping and most filesystems * should provide a migration function. Anonymous * pages are part of swap space which also has its * own migration function. This is the most common * path for page migration. */ rc = mapping->a_ops->migratepage(mapping, newpage, page); else rc = fallback_migrate_page(mapping, newpage, page); |
3fe2011ff
|
511 |
if (rc) { |
e24f0b8f7
|
512 |
newpage->mapping = NULL; |
3fe2011ff
|
513 514 515 516 |
} else { if (remap_swapcache) remove_migration_ptes(page, newpage); } |
e24f0b8f7
|
517 518 519 520 521 522 523 524 525 526 |
unlock_page(newpage); return rc; } /* * Obtain the lock on page, remove all ptes and migrate the page * to the newly allocated page in newpage. */ |
95a402c38
|
527 |
static int unmap_and_move(new_page_t get_new_page, unsigned long private, |
62b61f611
|
528 |
struct page *page, int force, int offlining) |
e24f0b8f7
|
529 530 |
{ int rc = 0; |
742755a1d
|
531 532 |
int *result = NULL; struct page *newpage = get_new_page(page, private, &result); |
3fe2011ff
|
533 |
int remap_swapcache = 1; |
989f89c57
|
534 |
int rcu_locked = 0; |
ae41be374
|
535 |
int charge = 0; |
e00e43161
|
536 |
struct mem_cgroup *mem = NULL; |
3f6c82728
|
537 |
struct anon_vma *anon_vma = NULL; |
95a402c38
|
538 539 540 |
if (!newpage) return -ENOMEM; |
e24f0b8f7
|
541 |
|
894bc3104
|
542 |
if (page_count(page) == 1) { |
e24f0b8f7
|
543 |
/* page was freed from under us. So we are done. */ |
95a402c38
|
544 |
goto move_newpage; |
894bc3104
|
545 |
} |
e24f0b8f7
|
546 |
|
e8589cc18
|
547 |
/* prepare cgroup just returns 0 or -ENOMEM */ |
e24f0b8f7
|
548 |
rc = -EAGAIN; |
01b1ae63c
|
549 |
|
529ae9aaa
|
550 |
if (!trylock_page(page)) { |
e24f0b8f7
|
551 |
if (!force) |
95a402c38
|
552 |
goto move_newpage; |
e24f0b8f7
|
553 554 |
lock_page(page); } |
62b61f611
|
555 556 557 558 559 560 561 562 563 564 565 566 567 |
/* * 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
|
568 |
/* charge against new page */ |
ac39cf8cb
|
569 |
charge = mem_cgroup_prepare_migration(page, newpage, &mem); |
01b1ae63c
|
570 571 572 573 574 |
if (charge == -ENOMEM) { rc = -ENOMEM; goto unlock; } BUG_ON(charge); |
e24f0b8f7
|
575 576 |
if (PageWriteback(page)) { if (!force) |
01b1ae63c
|
577 |
goto uncharge; |
e24f0b8f7
|
578 579 |
wait_on_page_writeback(page); } |
e24f0b8f7
|
580 |
/* |
dc386d4d1
|
581 582 583 584 |
* 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. * This rcu_read_lock() delays freeing anon_vma pointer until the end * of migration. File cache pages are no problem because of page_lock() |
989f89c57
|
585 586 |
* File Caches may use write_page() or lock_page() in migration, then, * just care Anon page here. |
dc386d4d1
|
587 |
*/ |
989f89c57
|
588 589 590 |
if (PageAnon(page)) { rcu_read_lock(); rcu_locked = 1; |
67b9509b2
|
591 |
|
3fe2011ff
|
592 593 594 595 |
/* Determine how to safely use anon_vma */ if (!page_mapped(page)) { if (!PageSwapCache(page)) goto rcu_unlock; |
67b9509b2
|
596 |
|
3fe2011ff
|
597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 |
/* * 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 { /* * Take a reference count on the anon_vma if the * page is mapped so that it is guaranteed to * exist when the page is remapped later */ anon_vma = page_anon_vma(page); atomic_inc(&anon_vma->external_refcount); } |
989f89c57
|
619 |
} |
62e1c5530
|
620 |
|
dc386d4d1
|
621 |
/* |
62e1c5530
|
622 623 624 625 626 627 628 629 630 631 |
* 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
|
632 |
*/ |
62e1c5530
|
633 |
if (!page->mapping) { |
266cf658e
|
634 |
if (!PageAnon(page) && page_has_private(page)) { |
62e1c5530
|
635 636 637 638 639 640 641 642 |
/* * Go direct to try_to_free_buffers() here because * a) that's what try_to_release_page() would do anyway * b) we may be under rcu_read_lock() here, so we can't * use GFP_KERNEL which is what try_to_release_page() * needs to be effective. */ try_to_free_buffers(page); |
abfc34881
|
643 |
goto rcu_unlock; |
62e1c5530
|
644 |
} |
abfc34881
|
645 |
goto skip_unmap; |
62e1c5530
|
646 |
} |
dc386d4d1
|
647 |
/* Establish migration ptes or remove ptes */ |
14fa31b89
|
648 |
try_to_unmap(page, TTU_MIGRATION|TTU_IGNORE_MLOCK|TTU_IGNORE_ACCESS); |
dc386d4d1
|
649 |
|
abfc34881
|
650 |
skip_unmap: |
e6a1530d6
|
651 |
if (!page_mapped(page)) |
3fe2011ff
|
652 |
rc = move_to_new_page(newpage, page, remap_swapcache); |
e24f0b8f7
|
653 |
|
3fe2011ff
|
654 |
if (rc && remap_swapcache) |
e24f0b8f7
|
655 |
remove_migration_ptes(page, page); |
dc386d4d1
|
656 |
rcu_unlock: |
3f6c82728
|
657 658 |
/* Drop an anon_vma reference if we took one */ |
7f60c214f
|
659 |
if (anon_vma && atomic_dec_and_lock(&anon_vma->external_refcount, &anon_vma->lock)) { |
3f6c82728
|
660 661 662 663 664 |
int empty = list_empty(&anon_vma->head); spin_unlock(&anon_vma->lock); if (empty) anon_vma_free(anon_vma); } |
989f89c57
|
665 666 |
if (rcu_locked) rcu_read_unlock(); |
01b1ae63c
|
667 668 669 |
uncharge: if (!charge) mem_cgroup_end_migration(mem, page, newpage); |
e24f0b8f7
|
670 671 |
unlock: unlock_page(page); |
95a402c38
|
672 |
|
e24f0b8f7
|
673 |
if (rc != -EAGAIN) { |
aaa994b30
|
674 675 676 677 678 679 680 |
/* * 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
|
681 |
dec_zone_page_state(page, NR_ISOLATED_ANON + |
6c0b13519
|
682 |
page_is_file_cache(page)); |
894bc3104
|
683 |
putback_lru_page(page); |
e24f0b8f7
|
684 |
} |
95a402c38
|
685 686 |
move_newpage: |
894bc3104
|
687 |
|
95a402c38
|
688 689 690 691 |
/* * Move the new page to the LRU. If migration was not successful * then this will free the page. */ |
894bc3104
|
692 |
putback_lru_page(newpage); |
742755a1d
|
693 694 695 696 697 698 |
if (result) { if (rc) *result = rc; else *result = page_to_nid(newpage); } |
e24f0b8f7
|
699 700 701 702 |
return rc; } /* |
b20a35035
|
703 704 |
* migrate_pages * |
95a402c38
|
705 706 707 |
* 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
|
708 709 710 |
* * The function returns after 10 attempts or if no pages * are movable anymore because to has become empty |
aaa994b30
|
711 |
* or no retryable pages exist anymore. All pages will be |
e9534b3fd
|
712 |
* returned to the LRU or freed. |
b20a35035
|
713 |
* |
95a402c38
|
714 |
* Return: Number of pages not migrated or error code. |
b20a35035
|
715 |
*/ |
95a402c38
|
716 |
int migrate_pages(struct list_head *from, |
62b61f611
|
717 |
new_page_t get_new_page, unsigned long private, int offlining) |
b20a35035
|
718 |
{ |
e24f0b8f7
|
719 |
int retry = 1; |
b20a35035
|
720 721 722 723 724 725 726 727 728 |
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
|
729 730 |
for(pass = 0; pass < 10 && retry; pass++) { retry = 0; |
b20a35035
|
731 |
|
e24f0b8f7
|
732 |
list_for_each_entry_safe(page, page2, from, lru) { |
e24f0b8f7
|
733 |
cond_resched(); |
2d1db3b11
|
734 |
|
95a402c38
|
735 |
rc = unmap_and_move(get_new_page, private, |
62b61f611
|
736 |
page, pass > 2, offlining); |
2d1db3b11
|
737 |
|
e24f0b8f7
|
738 |
switch(rc) { |
95a402c38
|
739 740 |
case -ENOMEM: goto out; |
e24f0b8f7
|
741 |
case -EAGAIN: |
2d1db3b11
|
742 |
retry++; |
e24f0b8f7
|
743 744 |
break; case 0: |
e24f0b8f7
|
745 746 |
break; default: |
2d1db3b11
|
747 |
/* Permanent failure */ |
2d1db3b11
|
748 |
nr_failed++; |
e24f0b8f7
|
749 |
break; |
2d1db3b11
|
750 |
} |
b20a35035
|
751 752 |
} } |
95a402c38
|
753 754 |
rc = 0; out: |
b20a35035
|
755 756 |
if (!swapwrite) current->flags &= ~PF_SWAPWRITE; |
aaa994b30
|
757 |
putback_lru_pages(from); |
b20a35035
|
758 |
|
95a402c38
|
759 760 |
if (rc) return rc; |
b20a35035
|
761 |
|
95a402c38
|
762 |
return nr_failed + retry; |
b20a35035
|
763 |
} |
95a402c38
|
764 |
|
742755a1d
|
765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 |
#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
|
788 |
return alloc_pages_exact_node(pm->node, |
769848c03
|
789 |
GFP_HIGHUSER_MOVABLE | GFP_THISNODE, 0); |
742755a1d
|
790 791 792 793 794 795 |
} /* * 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
|
796 |
* The pm array ends with node = MAX_NUMNODES. |
742755a1d
|
797 |
*/ |
5e9a0f023
|
798 799 800 |
static int do_move_page_to_node_array(struct mm_struct *mm, struct page_to_node *pm, int migrate_all) |
742755a1d
|
801 802 803 804 805 806 807 808 809 810 |
{ int err; struct page_to_node *pp; LIST_HEAD(pagelist); down_read(&mm->mmap_sem); /* * Build a list of pages to migrate */ |
742755a1d
|
811 812 813 |
for (pp = pm; pp->node != MAX_NUMNODES; pp++) { struct vm_area_struct *vma; struct page *page; |
742755a1d
|
814 815 |
err = -EFAULT; vma = find_vma(mm, pp->addr); |
0dc952dc3
|
816 |
if (!vma || !vma_migratable(vma)) |
742755a1d
|
817 818 819 |
goto set_status; page = follow_page(vma, pp->addr, FOLL_GET); |
89f5b7da2
|
820 821 822 823 |
err = PTR_ERR(page); if (IS_ERR(page)) goto set_status; |
742755a1d
|
824 825 826 |
err = -ENOENT; if (!page) goto set_status; |
62b61f611
|
827 828 |
/* Use PageReserved to check for zero page */ if (PageReserved(page) || PageKsm(page)) |
742755a1d
|
829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 |
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
|
844 |
err = isolate_lru_page(page); |
6d9c285a6
|
845 |
if (!err) { |
62695a84e
|
846 |
list_add_tail(&page->lru, &pagelist); |
6d9c285a6
|
847 848 849 |
inc_zone_page_state(page, NR_ISOLATED_ANON + page_is_file_cache(page)); } |
742755a1d
|
850 851 852 853 854 855 856 857 858 859 |
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
|
860 |
err = 0; |
742755a1d
|
861 862 |
if (!list_empty(&pagelist)) err = migrate_pages(&pagelist, new_page_node, |
62b61f611
|
863 |
(unsigned long)pm, 0); |
742755a1d
|
864 865 866 867 868 869 |
up_read(&mm->mmap_sem); return err; } /* |
5e9a0f023
|
870 871 872 873 874 875 876 877 878 |
* 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) { |
3140a2273
|
879 |
struct page_to_node *pm; |
5e9a0f023
|
880 |
nodemask_t task_nodes; |
3140a2273
|
881 882 883 |
unsigned long chunk_nr_pages; unsigned long chunk_start; int err; |
5e9a0f023
|
884 885 |
task_nodes = cpuset_mems_allowed(task); |
3140a2273
|
886 887 888 |
err = -ENOMEM; pm = (struct page_to_node *)__get_free_page(GFP_KERNEL); if (!pm) |
5e9a0f023
|
889 |
goto out; |
35282a2de
|
890 891 |
migrate_prep(); |
5e9a0f023
|
892 |
/* |
3140a2273
|
893 894 |
* Store a chunk of page_to_node array in a page, * but keep the last one as a marker |
5e9a0f023
|
895 |
*/ |
3140a2273
|
896 |
chunk_nr_pages = (PAGE_SIZE / sizeof(struct page_to_node)) - 1; |
5e9a0f023
|
897 |
|
3140a2273
|
898 899 900 901 |
for (chunk_start = 0; chunk_start < nr_pages; chunk_start += chunk_nr_pages) { int j; |
5e9a0f023
|
902 |
|
3140a2273
|
903 904 905 906 907 908 |
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
|
909 |
int node; |
3140a2273
|
910 911 912 913 914 915 |
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
|
916 917 918 |
goto out_pm; err = -ENODEV; |
6f5a55f1a
|
919 920 |
if (node < 0 || node >= MAX_NUMNODES) goto out_pm; |
5e9a0f023
|
921 922 923 924 925 926 |
if (!node_state(node, N_HIGH_MEMORY)) goto out_pm; err = -EACCES; if (!node_isset(node, task_nodes)) goto out_pm; |
3140a2273
|
927 928 929 930 931 932 933 934 935 936 937 |
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
|
938 |
|
5e9a0f023
|
939 |
/* Return status information */ |
3140a2273
|
940 941 |
for (j = 0; j < chunk_nr_pages; j++) if (put_user(pm[j].status, status + j + chunk_start)) { |
5e9a0f023
|
942 |
err = -EFAULT; |
3140a2273
|
943 944 945 946 |
goto out_pm; } } err = 0; |
5e9a0f023
|
947 948 |
out_pm: |
3140a2273
|
949 |
free_page((unsigned long)pm); |
5e9a0f023
|
950 951 952 953 954 |
out: return err; } /* |
2f007e74b
|
955 |
* Determine the nodes of an array of pages and store it in an array of status. |
742755a1d
|
956 |
*/ |
80bba1290
|
957 958 |
static void do_pages_stat_array(struct mm_struct *mm, unsigned long nr_pages, const void __user **pages, int *status) |
742755a1d
|
959 |
{ |
2f007e74b
|
960 |
unsigned long i; |
2f007e74b
|
961 |
|
742755a1d
|
962 |
down_read(&mm->mmap_sem); |
2f007e74b
|
963 |
for (i = 0; i < nr_pages; i++) { |
80bba1290
|
964 |
unsigned long addr = (unsigned long)(*pages); |
742755a1d
|
965 966 |
struct vm_area_struct *vma; struct page *page; |
c095adbc2
|
967 |
int err = -EFAULT; |
2f007e74b
|
968 969 |
vma = find_vma(mm, addr); |
742755a1d
|
970 971 |
if (!vma) goto set_status; |
2f007e74b
|
972 |
page = follow_page(vma, addr, 0); |
89f5b7da2
|
973 974 975 976 |
err = PTR_ERR(page); if (IS_ERR(page)) goto set_status; |
742755a1d
|
977 978 |
err = -ENOENT; /* Use PageReserved to check for zero page */ |
62b61f611
|
979 |
if (!page || PageReserved(page) || PageKsm(page)) |
742755a1d
|
980 981 982 983 |
goto set_status; err = page_to_nid(page); set_status: |
80bba1290
|
984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 |
*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
|
1004 |
|
87b8d1ade
|
1005 1006 |
while (nr_pages) { unsigned long chunk_nr; |
80bba1290
|
1007 |
|
87b8d1ade
|
1008 1009 1010 1011 1012 1013 |
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
|
1014 1015 |
do_pages_stat_array(mm, chunk_nr, chunk_pages, chunk_status); |
87b8d1ade
|
1016 1017 |
if (copy_to_user(status, chunk_status, chunk_nr * sizeof(*status))) break; |
742755a1d
|
1018 |
|
87b8d1ade
|
1019 1020 1021 1022 1023 |
pages += chunk_nr; status += chunk_nr; nr_pages -= chunk_nr; } return nr_pages ? -EFAULT : 0; |
742755a1d
|
1024 1025 1026 1027 1028 1029 |
} /* * Move a list of pages in the address space of the currently executing * process. */ |
938bb9f5e
|
1030 1031 1032 1033 |
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
|
1034 |
{ |
c69e8d9c0
|
1035 |
const struct cred *cred = current_cred(), *tcred; |
742755a1d
|
1036 |
struct task_struct *task; |
742755a1d
|
1037 |
struct mm_struct *mm; |
5e9a0f023
|
1038 |
int err; |
742755a1d
|
1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 |
/* 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 */ read_lock(&tasklist_lock); |
228ebcbe6
|
1049 |
task = pid ? find_task_by_vpid(pid) : current; |
742755a1d
|
1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 |
if (!task) { read_unlock(&tasklist_lock); return -ESRCH; } mm = get_task_mm(task); read_unlock(&tasklist_lock); 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. */ |
c69e8d9c0
|
1066 1067 |
rcu_read_lock(); tcred = __task_cred(task); |
b6dff3ec5
|
1068 1069 |
if (cred->euid != tcred->suid && cred->euid != tcred->uid && cred->uid != tcred->suid && cred->uid != tcred->uid && |
742755a1d
|
1070 |
!capable(CAP_SYS_NICE)) { |
c69e8d9c0
|
1071 |
rcu_read_unlock(); |
742755a1d
|
1072 |
err = -EPERM; |
5e9a0f023
|
1073 |
goto out; |
742755a1d
|
1074 |
} |
c69e8d9c0
|
1075 |
rcu_read_unlock(); |
742755a1d
|
1076 |
|
86c3a7645
|
1077 1078 |
err = security_task_movememory(task); if (err) |
5e9a0f023
|
1079 |
goto out; |
86c3a7645
|
1080 |
|
5e9a0f023
|
1081 1082 1083 1084 |
if (nodes) { err = do_pages_move(mm, task, nr_pages, pages, nodes, status, flags); } else { |
2f007e74b
|
1085 |
err = do_pages_stat(mm, nr_pages, pages, status); |
742755a1d
|
1086 |
} |
742755a1d
|
1087 |
out: |
742755a1d
|
1088 1089 1090 |
mmput(mm); return err; } |
742755a1d
|
1091 |
|
7b2259b3e
|
1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 |
/* * 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
|
1102 |
for (vma = mm->mmap; vma && !err; vma = vma->vm_next) { |
7b2259b3e
|
1103 1104 1105 1106 1107 1108 1109 1110 |
if (vma->vm_ops && vma->vm_ops->migrate) { err = vma->vm_ops->migrate(vma, to, from, flags); if (err) break; } } return err; } |
83d1674a9
|
1111 |
#endif |