Blame view
mm/migrate.c
33.6 KB
b20a35035 [PATCH] page migr... |
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 Christoph has moved |
12 |
* Christoph Lameter |
b20a35035 [PATCH] page migr... |
13 14 15 |
*/ #include <linux/migrate.h> |
b95f1b31b mm: Map most file... |
16 |
#include <linux/export.h> |
b20a35035 [PATCH] page migr... |
17 |
#include <linux/swap.h> |
0697212a4 [PATCH] Swapless ... |
18 |
#include <linux/swapops.h> |
b20a35035 [PATCH] page migr... |
19 |
#include <linux/pagemap.h> |
e23ca00bf [PATCH] Some page... |
20 |
#include <linux/buffer_head.h> |
b20a35035 [PATCH] page migr... |
21 |
#include <linux/mm_inline.h> |
b488893a3 pid namespaces: c... |
22 |
#include <linux/nsproxy.h> |
b20a35035 [PATCH] page migr... |
23 |
#include <linux/pagevec.h> |
e9995ef97 ksm: rmap_walk to... |
24 |
#include <linux/ksm.h> |
b20a35035 [PATCH] page migr... |
25 26 27 28 |
#include <linux/rmap.h> #include <linux/topology.h> #include <linux/cpu.h> #include <linux/cpuset.h> |
04e62a29b [PATCH] More page... |
29 |
#include <linux/writeback.h> |
742755a1d [PATCH] page migr... |
30 31 |
#include <linux/mempolicy.h> #include <linux/vmalloc.h> |
86c3a7645 [PATCH] SELinux: ... |
32 |
#include <linux/security.h> |
8a9f3ccd2 Memory controller... |
33 |
#include <linux/memcontrol.h> |
4f5ca2657 mm/migrate.c shou... |
34 |
#include <linux/syscalls.h> |
290408d4a hugetlb: hugepage... |
35 |
#include <linux/hugetlb.h> |
5a0e3ad6a include cleanup: ... |
36 |
#include <linux/gfp.h> |
b20a35035 [PATCH] page migr... |
37 |
|
0d1836c36 mm/migrate.c: fix... |
38 |
#include <asm/tlbflush.h> |
b20a35035 [PATCH] page migr... |
39 |
#include "internal.h" |
b20a35035 [PATCH] page migr... |
40 |
/* |
742755a1d [PATCH] page migr... |
41 |
* migrate_prep() needs to be called before we start compiling a list of pages |
748446bb6 mm: compaction: m... |
42 43 |
* to be migrated using isolate_lru_page(). If scheduling work on other CPUs is * undesirable, use migrate_prep_local() |
b20a35035 [PATCH] page migr... |
44 45 46 |
*/ int migrate_prep(void) { |
b20a35035 [PATCH] page migr... |
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 mm: compaction: m... |
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 [PATCH] page migr... |
64 |
/* |
894bc3104 Unevictable LRU I... |
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 [PATCH] page migr... |
67 |
*/ |
e13861d82 mm: remove return... |
68 |
void putback_lru_pages(struct list_head *l) |
b20a35035 [PATCH] page migr... |
69 70 71 |
{ struct page *page; struct page *page2; |
b20a35035 [PATCH] page migr... |
72 73 |
list_for_each_entry_safe(page, page2, l, lru) { |
e24f0b8f7 [PATCH] page migr... |
74 |
list_del(&page->lru); |
a731286de mm: vmstat: add i... |
75 |
dec_zone_page_state(page, NR_ISOLATED_ANON + |
6c0b13519 mm: return boolea... |
76 |
page_is_file_cache(page)); |
894bc3104 Unevictable LRU I... |
77 |
putback_lru_page(page); |
b20a35035 [PATCH] page migr... |
78 |
} |
b20a35035 [PATCH] page migr... |
79 |
} |
0697212a4 [PATCH] Swapless ... |
80 81 82 |
/* * Restore a potential migration pte to a working pte entry */ |
e9995ef97 ksm: rmap_walk to... |
83 84 |
static int remove_migration_pte(struct page *new, struct vm_area_struct *vma, unsigned long addr, void *old) |
0697212a4 [PATCH] Swapless ... |
85 86 87 88 89 90 91 92 |
{ 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 hugetlb: hugepage... |
93 94 95 96 97 98 99 100 101 |
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 [PATCH] Swapless ... |
102 |
|
290408d4a hugetlb: hugepage... |
103 104 105 |
pud = pud_offset(pgd, addr); if (!pud_present(*pud)) goto out; |
0697212a4 [PATCH] Swapless ... |
106 |
|
290408d4a hugetlb: hugepage... |
107 |
pmd = pmd_offset(pud, addr); |
500d65d47 thp: pmd_trans_hu... |
108 109 |
if (pmd_trans_huge(*pmd)) goto out; |
290408d4a hugetlb: hugepage... |
110 111 |
if (!pmd_present(*pmd)) goto out; |
0697212a4 [PATCH] Swapless ... |
112 |
|
290408d4a hugetlb: hugepage... |
113 |
ptep = pte_offset_map(pmd, addr); |
0697212a4 [PATCH] Swapless ... |
114 |
|
486cf46f3 mm: fix race betw... |
115 116 117 118 |
/* * Peek to check is_swap_pte() before taking ptlock? No, we * can race mremap's move_ptes(), which skips anon_vma lock. */ |
290408d4a hugetlb: hugepage... |
119 120 121 |
ptl = pte_lockptr(mm, pmd); } |
0697212a4 [PATCH] Swapless ... |
122 |
|
0697212a4 [PATCH] Swapless ... |
123 124 125 |
spin_lock(ptl); pte = *ptep; if (!is_swap_pte(pte)) |
e9995ef97 ksm: rmap_walk to... |
126 |
goto unlock; |
0697212a4 [PATCH] Swapless ... |
127 128 |
entry = pte_to_swp_entry(pte); |
e9995ef97 ksm: rmap_walk to... |
129 130 131 |
if (!is_migration_entry(entry) || migration_entry_to_page(entry) != old) goto unlock; |
0697212a4 [PATCH] Swapless ... |
132 |
|
0697212a4 [PATCH] Swapless ... |
133 134 135 136 |
get_page(new); pte = pte_mkold(mk_pte(new, vma->vm_page_prot)); if (is_write_migration_entry(entry)) pte = pte_mkwrite(pte); |
3ef8fd7f7 Fix migration.c c... |
137 |
#ifdef CONFIG_HUGETLB_PAGE |
290408d4a hugetlb: hugepage... |
138 139 |
if (PageHuge(new)) pte = pte_mkhuge(pte); |
3ef8fd7f7 Fix migration.c c... |
140 |
#endif |
97ee05246 flush cache befor... |
141 |
flush_cache_page(vma, addr, pte_pfn(pte)); |
0697212a4 [PATCH] Swapless ... |
142 |
set_pte_at(mm, addr, ptep, pte); |
04e62a29b [PATCH] More page... |
143 |
|
290408d4a hugetlb: hugepage... |
144 145 146 147 148 149 |
if (PageHuge(new)) { if (PageAnon(new)) hugepage_add_anon_rmap(new, vma, addr); else page_dup_rmap(new); } else if (PageAnon(new)) |
04e62a29b [PATCH] More page... |
150 151 152 153 154 |
page_add_anon_rmap(new, vma, addr); else page_add_file_rmap(new); /* No need to invalidate - it was non-present before */ |
4b3073e1c MM: Pass a PTE po... |
155 |
update_mmu_cache(vma, addr, ptep); |
e9995ef97 ksm: rmap_walk to... |
156 |
unlock: |
0697212a4 [PATCH] Swapless ... |
157 |
pte_unmap_unlock(ptep, ptl); |
e9995ef97 ksm: rmap_walk to... |
158 159 |
out: return SWAP_AGAIN; |
0697212a4 [PATCH] Swapless ... |
160 161 162 |
} /* |
04e62a29b [PATCH] More page... |
163 164 165 166 167 |
* 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 ksm: rmap_walk to... |
168 |
rmap_walk(new, remove_migration_pte, old); |
04e62a29b [PATCH] More page... |
169 170 171 |
} /* |
0697212a4 [PATCH] Swapless ... |
172 173 174 |
* 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 [PATCH] Swapless ... |
175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 |
*/ 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 mm: speculative p... |
194 195 196 197 198 199 200 201 202 |
/* * 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 [PATCH] Swapless ... |
203 204 205 206 207 208 209 |
pte_unmap_unlock(ptep, ptl); wait_on_page_locked(page); put_page(page); return; out: pte_unmap_unlock(ptep, ptl); } |
b969c4ab9 mm: compaction: d... |
210 211 |
#ifdef CONFIG_BLOCK /* Returns true if all buffers are successfully locked */ |
a6bc32b89 mm: compaction: i... |
212 213 |
static bool buffer_migrate_lock_buffers(struct buffer_head *head, enum migrate_mode mode) |
b969c4ab9 mm: compaction: d... |
214 215 216 217 |
{ struct buffer_head *bh = head; /* Simple case, sync compaction */ |
a6bc32b89 mm: compaction: i... |
218 |
if (mode != MIGRATE_ASYNC) { |
b969c4ab9 mm: compaction: d... |
219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 |
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 mm: compaction: i... |
254 |
enum migrate_mode mode) |
b969c4ab9 mm: compaction: d... |
255 256 257 258 |
{ return true; } #endif /* CONFIG_BLOCK */ |
b20a35035 [PATCH] page migr... |
259 |
/* |
c3fcf8a5d [PATCH] page migr... |
260 |
* Replace the page in the mapping. |
5b5c7120e [PATCH] page migr... |
261 262 263 264 |
* * The number of remaining references must be: * 1 for anonymous pages without a mapping * 2 for pages with a mapping |
266cf658e FS-Cache: Recruit... |
265 |
* 3 for pages with a mapping and PagePrivate/PagePrivate2 set. |
b20a35035 [PATCH] page migr... |
266 |
*/ |
2d1db3b11 [PATCH] page migr... |
267 |
static int migrate_page_move_mapping(struct address_space *mapping, |
b969c4ab9 mm: compaction: d... |
268 |
struct page *newpage, struct page *page, |
a6bc32b89 mm: compaction: i... |
269 |
struct buffer_head *head, enum migrate_mode mode) |
b20a35035 [PATCH] page migr... |
270 |
{ |
e286781d5 mm: speculative p... |
271 |
int expected_count; |
7cf9c2c76 [PATCH] radix-tre... |
272 |
void **pslot; |
b20a35035 [PATCH] page migr... |
273 |
|
6c5240ae7 [PATCH] Swapless ... |
274 |
if (!mapping) { |
0e8c7d0fd page migration: f... |
275 |
/* Anonymous page without mapping */ |
6c5240ae7 [PATCH] Swapless ... |
276 277 278 279 |
if (page_count(page) != 1) return -EAGAIN; return 0; } |
19fd62312 mm: spinlock tree... |
280 |
spin_lock_irq(&mapping->tree_lock); |
b20a35035 [PATCH] page migr... |
281 |
|
7cf9c2c76 [PATCH] radix-tre... |
282 283 |
pslot = radix_tree_lookup_slot(&mapping->page_tree, page_index(page)); |
b20a35035 [PATCH] page migr... |
284 |
|
edcf4748c mm: return boolea... |
285 |
expected_count = 2 + page_has_private(page); |
e286781d5 mm: speculative p... |
286 |
if (page_count(page) != expected_count || |
29c1f677d mm: migration: us... |
287 |
radix_tree_deref_slot_protected(pslot, &mapping->tree_lock) != page) { |
19fd62312 mm: spinlock tree... |
288 |
spin_unlock_irq(&mapping->tree_lock); |
e23ca00bf [PATCH] Some page... |
289 |
return -EAGAIN; |
b20a35035 [PATCH] page migr... |
290 |
} |
e286781d5 mm: speculative p... |
291 |
if (!page_freeze_refs(page, expected_count)) { |
19fd62312 mm: spinlock tree... |
292 |
spin_unlock_irq(&mapping->tree_lock); |
e286781d5 mm: speculative p... |
293 294 |
return -EAGAIN; } |
b20a35035 [PATCH] page migr... |
295 |
/* |
b969c4ab9 mm: compaction: d... |
296 297 298 299 300 301 |
* 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 mm: compaction: i... |
302 303 |
if (mode == MIGRATE_ASYNC && head && !buffer_migrate_lock_buffers(head, mode)) { |
b969c4ab9 mm: compaction: d... |
304 305 306 307 308 309 |
page_unfreeze_refs(page, expected_count); spin_unlock_irq(&mapping->tree_lock); return -EAGAIN; } /* |
b20a35035 [PATCH] page migr... |
310 |
* Now we know that no one else is looking at the page. |
b20a35035 [PATCH] page migr... |
311 |
*/ |
7cf9c2c76 [PATCH] radix-tre... |
312 |
get_page(newpage); /* add cache reference */ |
b20a35035 [PATCH] page migr... |
313 314 315 316 |
if (PageSwapCache(page)) { SetPageSwapCache(newpage); set_page_private(newpage, page_private(page)); } |
7cf9c2c76 [PATCH] radix-tre... |
317 318 319 |
radix_tree_replace_slot(pslot, newpage); /* |
937a94c9d mm: migrate: one ... |
320 321 |
* Drop cache reference from old page by unfreezing * to one less reference. |
7cf9c2c76 [PATCH] radix-tre... |
322 323 |
* We know this isn't the last reference. */ |
937a94c9d mm: migrate: one ... |
324 |
page_unfreeze_refs(page, expected_count - 1); |
7cf9c2c76 [PATCH] radix-tre... |
325 |
|
0e8c7d0fd page migration: f... |
326 327 328 329 330 331 332 333 334 335 336 337 |
/* * 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 migrate: don't ac... |
338 |
if (!PageSwapCache(page) && PageSwapBacked(page)) { |
4b02108ac mm: oom analysis:... |
339 340 341 |
__dec_zone_page_state(page, NR_SHMEM); __inc_zone_page_state(newpage, NR_SHMEM); } |
19fd62312 mm: spinlock tree... |
342 |
spin_unlock_irq(&mapping->tree_lock); |
b20a35035 [PATCH] page migr... |
343 344 345 |
return 0; } |
b20a35035 [PATCH] page migr... |
346 347 |
/* |
290408d4a hugetlb: hugepage... |
348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 |
* 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 mm: migration: us... |
370 |
radix_tree_deref_slot_protected(pslot, &mapping->tree_lock) != page) { |
290408d4a hugetlb: hugepage... |
371 372 373 374 375 376 377 378 379 380 381 382 |
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 mm: migrate: one ... |
383 |
page_unfreeze_refs(page, expected_count - 1); |
290408d4a hugetlb: hugepage... |
384 385 386 387 388 389 |
spin_unlock_irq(&mapping->tree_lock); return 0; } /* |
b20a35035 [PATCH] page migr... |
390 391 |
* Copy the page to its new location */ |
290408d4a hugetlb: hugepage... |
392 |
void migrate_page_copy(struct page *newpage, struct page *page) |
b20a35035 [PATCH] page migr... |
393 |
{ |
290408d4a hugetlb: hugepage... |
394 395 396 397 |
if (PageHuge(page)) copy_huge_page(newpage, page); else copy_highpage(newpage, page); |
b20a35035 [PATCH] page migr... |
398 399 400 401 402 403 404 |
if (PageError(page)) SetPageError(newpage); if (PageReferenced(page)) SetPageReferenced(newpage); if (PageUptodate(page)) SetPageUptodate(newpage); |
894bc3104 Unevictable LRU I... |
405 406 |
if (TestClearPageActive(page)) { VM_BUG_ON(PageUnevictable(page)); |
b20a35035 [PATCH] page migr... |
407 |
SetPageActive(newpage); |
418b27ef5 mm: remove unevic... |
408 409 |
} else if (TestClearPageUnevictable(page)) SetPageUnevictable(newpage); |
b20a35035 [PATCH] page migr... |
410 411 412 413 414 415 416 |
if (PageChecked(page)) SetPageChecked(newpage); if (PageMappedToDisk(page)) SetPageMappedToDisk(newpage); if (PageDirty(page)) { clear_page_dirty_for_io(page); |
3a902c5f6 mm: fix warning o... |
417 418 419 420 421 |
/* * 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 Fix common misspe... |
422 |
* Whereas only part of our page may be dirty. |
3a902c5f6 mm: fix warning o... |
423 424 |
*/ __set_page_dirty_nobuffers(newpage); |
b20a35035 [PATCH] page migr... |
425 |
} |
b291f0003 mlock: mlocked pa... |
426 |
mlock_migrate_page(newpage, page); |
e9995ef97 ksm: rmap_walk to... |
427 |
ksm_migrate_page(newpage, page); |
b291f0003 mlock: mlocked pa... |
428 |
|
b20a35035 [PATCH] page migr... |
429 |
ClearPageSwapCache(page); |
b20a35035 [PATCH] page migr... |
430 431 432 433 434 435 436 437 438 439 440 |
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 [PATCH] page migr... |
441 |
|
1d8b85ccf [PATCH] page migr... |
442 443 444 445 446 |
/************************************************************ * Migration functions ***********************************************************/ /* Always fail migration. Used for mappings that are not movable */ |
2d1db3b11 [PATCH] page migr... |
447 448 |
int fail_migrate_page(struct address_space *mapping, struct page *newpage, struct page *page) |
1d8b85ccf [PATCH] page migr... |
449 450 451 452 |
{ return -EIO; } EXPORT_SYMBOL(fail_migrate_page); |
b20a35035 [PATCH] page migr... |
453 454 |
/* * Common logic to directly migrate a single page suitable for |
266cf658e FS-Cache: Recruit... |
455 |
* pages that do not use PagePrivate/PagePrivate2. |
b20a35035 [PATCH] page migr... |
456 457 458 |
* * Pages are locked upon entry and exit. */ |
2d1db3b11 [PATCH] page migr... |
459 |
int migrate_page(struct address_space *mapping, |
a6bc32b89 mm: compaction: i... |
460 461 |
struct page *newpage, struct page *page, enum migrate_mode mode) |
b20a35035 [PATCH] page migr... |
462 463 464 465 |
{ int rc; BUG_ON(PageWriteback(page)); /* Writeback must be complete */ |
a6bc32b89 mm: compaction: i... |
466 |
rc = migrate_page_move_mapping(mapping, newpage, page, NULL, mode); |
b20a35035 [PATCH] page migr... |
467 468 469 470 471 |
if (rc) return rc; migrate_page_copy(newpage, page); |
b20a35035 [PATCH] page migr... |
472 473 474 |
return 0; } EXPORT_SYMBOL(migrate_page); |
9361401eb [PATCH] BLOCK: Ma... |
475 |
#ifdef CONFIG_BLOCK |
b20a35035 [PATCH] page migr... |
476 |
/* |
1d8b85ccf [PATCH] page migr... |
477 478 479 480 |
* 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 [PATCH] page migr... |
481 |
int buffer_migrate_page(struct address_space *mapping, |
a6bc32b89 mm: compaction: i... |
482 |
struct page *newpage, struct page *page, enum migrate_mode mode) |
1d8b85ccf [PATCH] page migr... |
483 |
{ |
1d8b85ccf [PATCH] page migr... |
484 485 |
struct buffer_head *bh, *head; int rc; |
1d8b85ccf [PATCH] page migr... |
486 |
if (!page_has_buffers(page)) |
a6bc32b89 mm: compaction: i... |
487 |
return migrate_page(mapping, newpage, page, mode); |
1d8b85ccf [PATCH] page migr... |
488 489 |
head = page_buffers(page); |
a6bc32b89 mm: compaction: i... |
490 |
rc = migrate_page_move_mapping(mapping, newpage, page, head, mode); |
1d8b85ccf [PATCH] page migr... |
491 492 493 |
if (rc) return rc; |
b969c4ab9 mm: compaction: d... |
494 495 496 497 498 |
/* * 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 mm: compaction: i... |
499 500 |
if (mode != MIGRATE_ASYNC) BUG_ON(!buffer_migrate_lock_buffers(head, mode)); |
1d8b85ccf [PATCH] page migr... |
501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 |
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 [PATCH] BLOCK: Ma... |
530 |
#endif |
1d8b85ccf [PATCH] page migr... |
531 |
|
04e62a29b [PATCH] More page... |
532 533 534 535 |
/* * Writeback a page to clean the dirty state */ static int writeout(struct address_space *mapping, struct page *page) |
8351a6e47 [PATCH] page migr... |
536 |
{ |
04e62a29b [PATCH] More page... |
537 538 539 540 541 |
struct writeback_control wbc = { .sync_mode = WB_SYNC_NONE, .nr_to_write = 1, .range_start = 0, .range_end = LLONG_MAX, |
04e62a29b [PATCH] More page... |
542 543 544 545 546 547 548 549 550 551 552 |
.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 [PATCH] page migr... |
553 |
/* |
04e62a29b [PATCH] More page... |
554 555 556 557 558 559 |
* 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 [PATCH] page migr... |
560 |
*/ |
04e62a29b [PATCH] More page... |
561 |
remove_migration_ptes(page, page); |
8351a6e47 [PATCH] page migr... |
562 |
|
04e62a29b [PATCH] More page... |
563 |
rc = mapping->a_ops->writepage(page, &wbc); |
8351a6e47 [PATCH] page migr... |
564 |
|
04e62a29b [PATCH] More page... |
565 566 567 |
if (rc != AOP_WRITEPAGE_ACTIVATE) /* unlocked. Relock */ lock_page(page); |
bda8550de migration: fix wr... |
568 |
return (rc < 0) ? -EIO : -EAGAIN; |
04e62a29b [PATCH] More page... |
569 570 571 572 573 574 |
} /* * Default handling if a filesystem does not provide a migration function. */ static int fallback_migrate_page(struct address_space *mapping, |
a6bc32b89 mm: compaction: i... |
575 |
struct page *newpage, struct page *page, enum migrate_mode mode) |
04e62a29b [PATCH] More page... |
576 |
{ |
b969c4ab9 mm: compaction: d... |
577 |
if (PageDirty(page)) { |
a6bc32b89 mm: compaction: i... |
578 579 |
/* Only writeback pages in full synchronous migration */ if (mode != MIGRATE_SYNC) |
b969c4ab9 mm: compaction: d... |
580 |
return -EBUSY; |
04e62a29b [PATCH] More page... |
581 |
return writeout(mapping, page); |
b969c4ab9 mm: compaction: d... |
582 |
} |
8351a6e47 [PATCH] page migr... |
583 584 585 586 587 |
/* * Buffers may be managed in a filesystem specific way. * We must have no buffers or drop them. */ |
266cf658e FS-Cache: Recruit... |
588 |
if (page_has_private(page) && |
8351a6e47 [PATCH] page migr... |
589 590 |
!try_to_release_page(page, GFP_KERNEL)) return -EAGAIN; |
a6bc32b89 mm: compaction: i... |
591 |
return migrate_page(mapping, newpage, page, mode); |
8351a6e47 [PATCH] page migr... |
592 |
} |
1d8b85ccf [PATCH] page migr... |
593 |
/* |
e24f0b8f7 [PATCH] page migr... |
594 595 596 597 598 |
* 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 Unevictable LRU I... |
599 600 601 602 |
* * Return value: * < 0 - error code * == 0 - success |
e24f0b8f7 [PATCH] page migr... |
603 |
*/ |
3fe2011ff mm: migration: al... |
604 |
static int move_to_new_page(struct page *newpage, struct page *page, |
a6bc32b89 mm: compaction: i... |
605 |
int remap_swapcache, enum migrate_mode mode) |
e24f0b8f7 [PATCH] page migr... |
606 607 608 609 610 611 612 613 614 |
{ 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 mm: rename page t... |
615 |
if (!trylock_page(newpage)) |
e24f0b8f7 [PATCH] page migr... |
616 617 618 619 620 |
BUG(); /* Prepare mapping for the new page.*/ newpage->index = page->index; newpage->mapping = page->mapping; |
b2e185384 define page_file_... |
621 622 |
if (PageSwapBacked(page)) SetPageSwapBacked(newpage); |
e24f0b8f7 [PATCH] page migr... |
623 624 625 |
mapping = page_mapping(page); if (!mapping) |
a6bc32b89 mm: compaction: i... |
626 |
rc = migrate_page(mapping, newpage, page, mode); |
b969c4ab9 mm: compaction: d... |
627 |
else if (mapping->a_ops->migratepage) |
e24f0b8f7 [PATCH] page migr... |
628 |
/* |
b969c4ab9 mm: compaction: d... |
629 630 631 632 |
* 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 [PATCH] page migr... |
633 |
*/ |
b969c4ab9 mm: compaction: d... |
634 |
rc = mapping->a_ops->migratepage(mapping, |
a6bc32b89 mm: compaction: i... |
635 |
newpage, page, mode); |
b969c4ab9 mm: compaction: d... |
636 |
else |
a6bc32b89 mm: compaction: i... |
637 |
rc = fallback_migrate_page(mapping, newpage, page, mode); |
e24f0b8f7 [PATCH] page migr... |
638 |
|
3fe2011ff mm: migration: al... |
639 |
if (rc) { |
e24f0b8f7 [PATCH] page migr... |
640 |
newpage->mapping = NULL; |
3fe2011ff mm: migration: al... |
641 642 643 644 |
} else { if (remap_swapcache) remove_migration_ptes(page, newpage); } |
e24f0b8f7 [PATCH] page migr... |
645 646 647 648 649 |
unlock_page(newpage); return rc; } |
0dabec93d mm: migration: cl... |
650 |
static int __unmap_and_move(struct page *page, struct page *newpage, |
a6bc32b89 mm: compaction: i... |
651 |
int force, bool offlining, enum migrate_mode mode) |
e24f0b8f7 [PATCH] page migr... |
652 |
{ |
0dabec93d mm: migration: cl... |
653 |
int rc = -EAGAIN; |
3fe2011ff mm: migration: al... |
654 |
int remap_swapcache = 1; |
ae41be374 bugfix for memory... |
655 |
int charge = 0; |
56039efa1 memcg: fix ugly i... |
656 |
struct mem_cgroup *mem; |
3f6c82728 mm: migration: ta... |
657 |
struct anon_vma *anon_vma = NULL; |
95a402c38 [PATCH] page migr... |
658 |
|
529ae9aaa mm: rename page t... |
659 |
if (!trylock_page(page)) { |
a6bc32b89 mm: compaction: i... |
660 |
if (!force || mode == MIGRATE_ASYNC) |
0dabec93d mm: migration: cl... |
661 |
goto out; |
3e7d34497 mm: vmscan: recla... |
662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 |
/* * 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 mm: migration: cl... |
677 |
goto out; |
3e7d34497 mm: vmscan: recla... |
678 |
|
e24f0b8f7 [PATCH] page migr... |
679 680 |
lock_page(page); } |
62b61f611 ksm: memory hotre... |
681 682 683 684 685 686 687 688 689 690 691 692 693 |
/* * 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 memcg: simple mig... |
694 |
/* charge against new page */ |
ef6a3c631 mm: add replace_p... |
695 |
charge = mem_cgroup_prepare_migration(page, newpage, &mem, GFP_KERNEL); |
01b1ae63c memcg: simple mig... |
696 697 698 699 700 |
if (charge == -ENOMEM) { rc = -ENOMEM; goto unlock; } BUG_ON(charge); |
e24f0b8f7 [PATCH] page migr... |
701 |
if (PageWriteback(page)) { |
11bc82d67 mm: compaction: U... |
702 |
/* |
a6bc32b89 mm: compaction: i... |
703 704 705 706 |
* 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 mm: compaction: U... |
707 |
*/ |
a6bc32b89 mm: compaction: i... |
708 |
if (mode != MIGRATE_SYNC) { |
11bc82d67 mm: compaction: U... |
709 710 711 712 |
rc = -EBUSY; goto uncharge; } if (!force) |
01b1ae63c memcg: simple mig... |
713 |
goto uncharge; |
e24f0b8f7 [PATCH] page migr... |
714 715 |
wait_on_page_writeback(page); } |
e24f0b8f7 [PATCH] page migr... |
716 |
/* |
dc386d4d1 memory unplug: mi... |
717 718 |
* 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 mm: fix migration... |
719 |
* This get_anon_vma() delays freeing anon_vma pointer until the end |
dc386d4d1 memory unplug: mi... |
720 |
* of migration. File cache pages are no problem because of page_lock() |
989f89c57 fix rcu_read_lock... |
721 722 |
* File Caches may use write_page() or lock_page() in migration, then, * just care Anon page here. |
dc386d4d1 memory unplug: mi... |
723 |
*/ |
989f89c57 fix rcu_read_lock... |
724 |
if (PageAnon(page)) { |
1ce82b69e mm: fix migration... |
725 726 727 728 |
/* * Only page_lock_anon_vma() understands the subtleties of * getting a hold on an anon_vma from outside one of its mms. */ |
746b18d42 mm: use refcounts... |
729 |
anon_vma = page_get_anon_vma(page); |
1ce82b69e mm: fix migration... |
730 731 |
if (anon_vma) { /* |
746b18d42 mm: use refcounts... |
732 |
* Anon page |
1ce82b69e mm: fix migration... |
733 |
*/ |
1ce82b69e mm: fix migration... |
734 |
} else if (PageSwapCache(page)) { |
3fe2011ff mm: migration: al... |
735 736 737 738 739 740 741 742 743 744 745 746 747 748 |
/* * 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 mm: fix migration... |
749 |
goto uncharge; |
3fe2011ff mm: migration: al... |
750 |
} |
989f89c57 fix rcu_read_lock... |
751 |
} |
62e1c5530 page migraton: ha... |
752 |
|
dc386d4d1 memory unplug: mi... |
753 |
/* |
62e1c5530 page migraton: ha... |
754 755 756 757 758 759 760 761 762 763 |
* 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 [PATCH] page migr... |
764 |
*/ |
62e1c5530 page migraton: ha... |
765 |
if (!page->mapping) { |
1ce82b69e mm: fix migration... |
766 767 |
VM_BUG_ON(PageAnon(page)); if (page_has_private(page)) { |
62e1c5530 page migraton: ha... |
768 |
try_to_free_buffers(page); |
1ce82b69e mm: fix migration... |
769 |
goto uncharge; |
62e1c5530 page migraton: ha... |
770 |
} |
abfc34881 memory hotplug: m... |
771 |
goto skip_unmap; |
62e1c5530 page migraton: ha... |
772 |
} |
dc386d4d1 memory unplug: mi... |
773 |
/* Establish migration ptes or remove ptes */ |
14fa31b89 HWPOISON: Use bit... |
774 |
try_to_unmap(page, TTU_MIGRATION|TTU_IGNORE_MLOCK|TTU_IGNORE_ACCESS); |
dc386d4d1 memory unplug: mi... |
775 |
|
abfc34881 memory hotplug: m... |
776 |
skip_unmap: |
e6a1530d6 [PATCH] Allow mig... |
777 |
if (!page_mapped(page)) |
a6bc32b89 mm: compaction: i... |
778 |
rc = move_to_new_page(newpage, page, remap_swapcache, mode); |
e24f0b8f7 [PATCH] page migr... |
779 |
|
3fe2011ff mm: migration: al... |
780 |
if (rc && remap_swapcache) |
e24f0b8f7 [PATCH] page migr... |
781 |
remove_migration_ptes(page, page); |
3f6c82728 mm: migration: ta... |
782 783 |
/* Drop an anon_vma reference if we took one */ |
76545066c mm: extend KSM re... |
784 |
if (anon_vma) |
9e60109f1 mm: rename drop_a... |
785 |
put_anon_vma(anon_vma); |
3f6c82728 mm: migration: ta... |
786 |
|
01b1ae63c memcg: simple mig... |
787 788 |
uncharge: if (!charge) |
50de1dd96 memcg: fix memory... |
789 |
mem_cgroup_end_migration(mem, page, newpage, rc == 0); |
e24f0b8f7 [PATCH] page migr... |
790 791 |
unlock: unlock_page(page); |
0dabec93d mm: migration: cl... |
792 793 794 |
out: return rc; } |
95a402c38 [PATCH] page migr... |
795 |
|
0dabec93d mm: migration: cl... |
796 797 798 799 800 |
/* * 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 mm: compaction: i... |
801 802 |
struct page *page, int force, bool offlining, enum migrate_mode mode) |
0dabec93d mm: migration: cl... |
803 804 805 806 807 808 809 |
{ int rc = 0; int *result = NULL; struct page *newpage = get_new_page(page, private, &result); if (!newpage) return -ENOMEM; |
4e5f01c2b memcg: clear pc->... |
810 |
mem_cgroup_reset_owner(newpage); |
0dabec93d mm: migration: cl... |
811 812 813 814 815 816 817 818 |
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 mm: compaction: i... |
819 |
rc = __unmap_and_move(page, newpage, force, offlining, mode); |
0dabec93d mm: migration: cl... |
820 |
out: |
e24f0b8f7 [PATCH] page migr... |
821 |
if (rc != -EAGAIN) { |
0dabec93d mm: migration: cl... |
822 823 824 825 826 827 828 |
/* * 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 mm: vmstat: add i... |
829 |
dec_zone_page_state(page, NR_ISOLATED_ANON + |
6c0b13519 mm: return boolea... |
830 |
page_is_file_cache(page)); |
894bc3104 Unevictable LRU I... |
831 |
putback_lru_page(page); |
e24f0b8f7 [PATCH] page migr... |
832 |
} |
95a402c38 [PATCH] page migr... |
833 834 835 836 |
/* * Move the new page to the LRU. If migration was not successful * then this will free the page. */ |
894bc3104 Unevictable LRU I... |
837 |
putback_lru_page(newpage); |
742755a1d [PATCH] page migr... |
838 839 840 841 842 843 |
if (result) { if (rc) *result = rc; else *result = page_to_nid(newpage); } |
e24f0b8f7 [PATCH] page migr... |
844 845 846 847 |
return rc; } /* |
290408d4a hugetlb: hugepage... |
848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 |
* 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 mm: compaction: i... |
867 868 |
int force, bool offlining, enum migrate_mode mode) |
290408d4a hugetlb: hugepage... |
869 870 871 872 |
{ int rc = 0; int *result = NULL; struct page *new_hpage = get_new_page(hpage, private, &result); |
290408d4a hugetlb: hugepage... |
873 874 875 876 877 878 879 880 |
struct anon_vma *anon_vma = NULL; if (!new_hpage) return -ENOMEM; rc = -EAGAIN; if (!trylock_page(hpage)) { |
a6bc32b89 mm: compaction: i... |
881 |
if (!force || mode != MIGRATE_SYNC) |
290408d4a hugetlb: hugepage... |
882 883 884 |
goto out; lock_page(hpage); } |
746b18d42 mm: use refcounts... |
885 886 |
if (PageAnon(hpage)) anon_vma = page_get_anon_vma(hpage); |
290408d4a hugetlb: hugepage... |
887 888 889 890 |
try_to_unmap(hpage, TTU_MIGRATION|TTU_IGNORE_MLOCK|TTU_IGNORE_ACCESS); if (!page_mapped(hpage)) |
a6bc32b89 mm: compaction: i... |
891 |
rc = move_to_new_page(new_hpage, hpage, 1, mode); |
290408d4a hugetlb: hugepage... |
892 893 894 |
if (rc) remove_migration_ptes(hpage, hpage); |
fd4a4663d mm: fix hugepage ... |
895 |
if (anon_vma) |
9e60109f1 mm: rename drop_a... |
896 |
put_anon_vma(anon_vma); |
290408d4a hugetlb: hugepage... |
897 |
unlock_page(hpage); |
09761333e mm/migrate.c: pai... |
898 |
out: |
290408d4a hugetlb: hugepage... |
899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 |
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; } /* |
b20a35035 [PATCH] page migr... |
916 917 |
* migrate_pages * |
95a402c38 [PATCH] page migr... |
918 919 920 |
* 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 [PATCH] page migr... |
921 922 923 |
* * The function returns after 10 attempts or if no pages * are movable anymore because to has become empty |
cf608ac19 mm: compaction: f... |
924 925 |
* or no retryable pages exist anymore. * Caller should call putback_lru_pages to return pages to the LRU |
28bd65781 mm: migration: cl... |
926 |
* or free list only if ret != 0. |
b20a35035 [PATCH] page migr... |
927 |
* |
95a402c38 [PATCH] page migr... |
928 |
* Return: Number of pages not migrated or error code. |
b20a35035 [PATCH] page migr... |
929 |
*/ |
95a402c38 [PATCH] page migr... |
930 |
int migrate_pages(struct list_head *from, |
7f0f24967 mm: migration: cl... |
931 |
new_page_t get_new_page, unsigned long private, bool offlining, |
a6bc32b89 mm: compaction: i... |
932 |
enum migrate_mode mode) |
b20a35035 [PATCH] page migr... |
933 |
{ |
e24f0b8f7 [PATCH] page migr... |
934 |
int retry = 1; |
b20a35035 [PATCH] page migr... |
935 936 937 938 939 940 941 942 943 |
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 [PATCH] page migr... |
944 945 |
for(pass = 0; pass < 10 && retry; pass++) { retry = 0; |
b20a35035 [PATCH] page migr... |
946 |
|
e24f0b8f7 [PATCH] page migr... |
947 |
list_for_each_entry_safe(page, page2, from, lru) { |
e24f0b8f7 [PATCH] page migr... |
948 |
cond_resched(); |
2d1db3b11 [PATCH] page migr... |
949 |
|
95a402c38 [PATCH] page migr... |
950 |
rc = unmap_and_move(get_new_page, private, |
77f1fe6b0 mm: migration: al... |
951 |
page, pass > 2, offlining, |
a6bc32b89 mm: compaction: i... |
952 |
mode); |
2d1db3b11 [PATCH] page migr... |
953 |
|
e24f0b8f7 [PATCH] page migr... |
954 |
switch(rc) { |
95a402c38 [PATCH] page migr... |
955 956 |
case -ENOMEM: goto out; |
e24f0b8f7 [PATCH] page migr... |
957 |
case -EAGAIN: |
2d1db3b11 [PATCH] page migr... |
958 |
retry++; |
e24f0b8f7 [PATCH] page migr... |
959 960 |
break; case 0: |
e24f0b8f7 [PATCH] page migr... |
961 962 |
break; default: |
2d1db3b11 [PATCH] page migr... |
963 |
/* Permanent failure */ |
2d1db3b11 [PATCH] page migr... |
964 |
nr_failed++; |
e24f0b8f7 [PATCH] page migr... |
965 |
break; |
2d1db3b11 [PATCH] page migr... |
966 |
} |
b20a35035 [PATCH] page migr... |
967 968 |
} } |
95a402c38 [PATCH] page migr... |
969 970 |
rc = 0; out: |
b20a35035 [PATCH] page migr... |
971 972 |
if (!swapwrite) current->flags &= ~PF_SWAPWRITE; |
95a402c38 [PATCH] page migr... |
973 974 |
if (rc) return rc; |
b20a35035 [PATCH] page migr... |
975 |
|
95a402c38 [PATCH] page migr... |
976 |
return nr_failed + retry; |
b20a35035 [PATCH] page migr... |
977 |
} |
95a402c38 [PATCH] page migr... |
978 |
|
290408d4a hugetlb: hugepage... |
979 |
int migrate_huge_pages(struct list_head *from, |
7f0f24967 mm: migration: cl... |
980 |
new_page_t get_new_page, unsigned long private, bool offlining, |
a6bc32b89 mm: compaction: i... |
981 |
enum migrate_mode mode) |
290408d4a hugetlb: hugepage... |
982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 |
{ 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, |
77f1fe6b0 mm: migration: al... |
997 |
private, page, pass > 2, offlining, |
a6bc32b89 mm: compaction: i... |
998 |
mode); |
290408d4a hugetlb: hugepage... |
999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 |
switch(rc) { case -ENOMEM: goto out; case -EAGAIN: retry++; break; case 0: break; default: /* Permanent failure */ nr_failed++; break; } } } rc = 0; out: |
290408d4a hugetlb: hugepage... |
1017 1018 1019 1020 1021 |
if (rc) return rc; return nr_failed + retry; } |
742755a1d [PATCH] page migr... |
1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 |
#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 page allocator: d... |
1045 |
return alloc_pages_exact_node(pm->node, |
769848c03 Add __GFP_MOVABLE... |
1046 |
GFP_HIGHUSER_MOVABLE | GFP_THISNODE, 0); |
742755a1d [PATCH] page migr... |
1047 1048 1049 1050 1051 1052 |
} /* * 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 mm: extract do_pa... |
1053 |
* The pm array ends with node = MAX_NUMNODES. |
742755a1d [PATCH] page migr... |
1054 |
*/ |
5e9a0f023 mm: extract do_pa... |
1055 1056 1057 |
static int do_move_page_to_node_array(struct mm_struct *mm, struct page_to_node *pm, int migrate_all) |
742755a1d [PATCH] page migr... |
1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 |
{ int err; struct page_to_node *pp; LIST_HEAD(pagelist); down_read(&mm->mmap_sem); /* * Build a list of pages to migrate */ |
742755a1d [PATCH] page migr... |
1068 1069 1070 |
for (pp = pm; pp->node != MAX_NUMNODES; pp++) { struct vm_area_struct *vma; struct page *page; |
742755a1d [PATCH] page migr... |
1071 1072 |
err = -EFAULT; vma = find_vma(mm, pp->addr); |
70384dc6d mm: fix error rep... |
1073 |
if (!vma || pp->addr < vma->vm_start || !vma_migratable(vma)) |
742755a1d [PATCH] page migr... |
1074 |
goto set_status; |
500d65d47 thp: pmd_trans_hu... |
1075 |
page = follow_page(vma, pp->addr, FOLL_GET|FOLL_SPLIT); |
89f5b7da2 Reinstate ZERO_PA... |
1076 1077 1078 1079 |
err = PTR_ERR(page); if (IS_ERR(page)) goto set_status; |
742755a1d [PATCH] page migr... |
1080 1081 1082 |
err = -ENOENT; if (!page) goto set_status; |
62b61f611 ksm: memory hotre... |
1083 1084 |
/* Use PageReserved to check for zero page */ if (PageReserved(page) || PageKsm(page)) |
742755a1d [PATCH] page migr... |
1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 |
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 vmscan: move isol... |
1100 |
err = isolate_lru_page(page); |
6d9c285a6 mm: move inc_zone... |
1101 |
if (!err) { |
62695a84e vmscan: move isol... |
1102 |
list_add_tail(&page->lru, &pagelist); |
6d9c285a6 mm: move inc_zone... |
1103 1104 1105 |
inc_zone_page_state(page, NR_ISOLATED_ANON + page_is_file_cache(page)); } |
742755a1d [PATCH] page migr... |
1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 |
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 mm: stop returnin... |
1116 |
err = 0; |
cf608ac19 mm: compaction: f... |
1117 |
if (!list_empty(&pagelist)) { |
742755a1d [PATCH] page migr... |
1118 |
err = migrate_pages(&pagelist, new_page_node, |
a6bc32b89 mm: compaction: i... |
1119 |
(unsigned long)pm, 0, MIGRATE_SYNC); |
cf608ac19 mm: compaction: f... |
1120 1121 1122 |
if (err) putback_lru_pages(&pagelist); } |
742755a1d [PATCH] page migr... |
1123 1124 1125 1126 1127 1128 |
up_read(&mm->mmap_sem); return err; } /* |
5e9a0f023 mm: extract do_pa... |
1129 1130 1131 1132 1133 1134 1135 1136 1137 |
* 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 mm: rework do_pag... |
1138 |
struct page_to_node *pm; |
5e9a0f023 mm: extract do_pa... |
1139 |
nodemask_t task_nodes; |
3140a2273 mm: rework do_pag... |
1140 1141 1142 |
unsigned long chunk_nr_pages; unsigned long chunk_start; int err; |
5e9a0f023 mm: extract do_pa... |
1143 1144 |
task_nodes = cpuset_mems_allowed(task); |
3140a2273 mm: rework do_pag... |
1145 1146 1147 |
err = -ENOMEM; pm = (struct page_to_node *)__get_free_page(GFP_KERNEL); if (!pm) |
5e9a0f023 mm: extract do_pa... |
1148 |
goto out; |
35282a2de migration: only m... |
1149 1150 |
migrate_prep(); |
5e9a0f023 mm: extract do_pa... |
1151 |
/* |
3140a2273 mm: rework do_pag... |
1152 1153 |
* Store a chunk of page_to_node array in a page, * but keep the last one as a marker |
5e9a0f023 mm: extract do_pa... |
1154 |
*/ |
3140a2273 mm: rework do_pag... |
1155 |
chunk_nr_pages = (PAGE_SIZE / sizeof(struct page_to_node)) - 1; |
5e9a0f023 mm: extract do_pa... |
1156 |
|
3140a2273 mm: rework do_pag... |
1157 1158 1159 1160 |
for (chunk_start = 0; chunk_start < nr_pages; chunk_start += chunk_nr_pages) { int j; |
5e9a0f023 mm: extract do_pa... |
1161 |
|
3140a2273 mm: rework do_pag... |
1162 1163 1164 1165 1166 1167 |
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 mm: extract do_pa... |
1168 |
int node; |
3140a2273 mm: rework do_pag... |
1169 1170 1171 1172 1173 1174 |
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 mm: extract do_pa... |
1175 1176 1177 |
goto out_pm; err = -ENODEV; |
6f5a55f1a Fix potential cra... |
1178 1179 |
if (node < 0 || node >= MAX_NUMNODES) goto out_pm; |
5e9a0f023 mm: extract do_pa... |
1180 1181 1182 1183 1184 1185 |
if (!node_state(node, N_HIGH_MEMORY)) goto out_pm; err = -EACCES; if (!node_isset(node, task_nodes)) goto out_pm; |
3140a2273 mm: rework do_pag... |
1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 |
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 mm: extract do_pa... |
1197 |
|
5e9a0f023 mm: extract do_pa... |
1198 |
/* Return status information */ |
3140a2273 mm: rework do_pag... |
1199 1200 |
for (j = 0; j < chunk_nr_pages; j++) if (put_user(pm[j].status, status + j + chunk_start)) { |
5e9a0f023 mm: extract do_pa... |
1201 |
err = -EFAULT; |
3140a2273 mm: rework do_pag... |
1202 1203 1204 1205 |
goto out_pm; } } err = 0; |
5e9a0f023 mm: extract do_pa... |
1206 1207 |
out_pm: |
3140a2273 mm: rework do_pag... |
1208 |
free_page((unsigned long)pm); |
5e9a0f023 mm: extract do_pa... |
1209 1210 1211 1212 1213 |
out: return err; } /* |
2f007e74b mm: don't vmalloc... |
1214 |
* Determine the nodes of an array of pages and store it in an array of status. |
742755a1d [PATCH] page migr... |
1215 |
*/ |
80bba1290 mm: no get_user/p... |
1216 1217 |
static void do_pages_stat_array(struct mm_struct *mm, unsigned long nr_pages, const void __user **pages, int *status) |
742755a1d [PATCH] page migr... |
1218 |
{ |
2f007e74b mm: don't vmalloc... |
1219 |
unsigned long i; |
2f007e74b mm: don't vmalloc... |
1220 |
|
742755a1d [PATCH] page migr... |
1221 |
down_read(&mm->mmap_sem); |
2f007e74b mm: don't vmalloc... |
1222 |
for (i = 0; i < nr_pages; i++) { |
80bba1290 mm: no get_user/p... |
1223 |
unsigned long addr = (unsigned long)(*pages); |
742755a1d [PATCH] page migr... |
1224 1225 |
struct vm_area_struct *vma; struct page *page; |
c095adbc2 mm: Don't touch u... |
1226 |
int err = -EFAULT; |
2f007e74b mm: don't vmalloc... |
1227 1228 |
vma = find_vma(mm, addr); |
70384dc6d mm: fix error rep... |
1229 |
if (!vma || addr < vma->vm_start) |
742755a1d [PATCH] page migr... |
1230 |
goto set_status; |
2f007e74b mm: don't vmalloc... |
1231 |
page = follow_page(vma, addr, 0); |
89f5b7da2 Reinstate ZERO_PA... |
1232 1233 1234 1235 |
err = PTR_ERR(page); if (IS_ERR(page)) goto set_status; |
742755a1d [PATCH] page migr... |
1236 1237 |
err = -ENOENT; /* Use PageReserved to check for zero page */ |
62b61f611 ksm: memory hotre... |
1238 |
if (!page || PageReserved(page) || PageKsm(page)) |
742755a1d [PATCH] page migr... |
1239 1240 1241 1242 |
goto set_status; err = page_to_nid(page); set_status: |
80bba1290 mm: no get_user/p... |
1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 |
*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 mm: no get_user/p... |
1263 |
|
87b8d1ade mm: Make copy_fro... |
1264 1265 |
while (nr_pages) { unsigned long chunk_nr; |
80bba1290 mm: no get_user/p... |
1266 |
|
87b8d1ade mm: Make copy_fro... |
1267 1268 1269 1270 1271 1272 |
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 mm: no get_user/p... |
1273 1274 |
do_pages_stat_array(mm, chunk_nr, chunk_pages, chunk_status); |
87b8d1ade mm: Make copy_fro... |
1275 1276 |
if (copy_to_user(status, chunk_status, chunk_nr * sizeof(*status))) break; |
742755a1d [PATCH] page migr... |
1277 |
|
87b8d1ade mm: Make copy_fro... |
1278 1279 1280 1281 1282 |
pages += chunk_nr; status += chunk_nr; nr_pages -= chunk_nr; } return nr_pages ? -EFAULT : 0; |
742755a1d [PATCH] page migr... |
1283 1284 1285 1286 1287 1288 |
} /* * Move a list of pages in the address space of the currently executing * process. */ |
938bb9f5e [CVE-2009-0029] S... |
1289 1290 1291 1292 |
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 [PATCH] page migr... |
1293 |
{ |
c69e8d9c0 CRED: Use RCU to ... |
1294 |
const struct cred *cred = current_cred(), *tcred; |
742755a1d [PATCH] page migr... |
1295 |
struct task_struct *task; |
742755a1d [PATCH] page migr... |
1296 |
struct mm_struct *mm; |
5e9a0f023 mm: extract do_pa... |
1297 |
int err; |
742755a1d [PATCH] page migr... |
1298 1299 1300 1301 1302 1303 1304 1305 1306 |
/* 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 mm: grab rcu read... |
1307 |
rcu_read_lock(); |
228ebcbe6 Uninline find_tas... |
1308 |
task = pid ? find_task_by_vpid(pid) : current; |
742755a1d [PATCH] page migr... |
1309 |
if (!task) { |
a879bf582 mm: grab rcu read... |
1310 |
rcu_read_unlock(); |
742755a1d [PATCH] page migr... |
1311 1312 1313 |
return -ESRCH; } mm = get_task_mm(task); |
a879bf582 mm: grab rcu read... |
1314 |
rcu_read_unlock(); |
742755a1d [PATCH] page migr... |
1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 |
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 CRED: Use RCU to ... |
1325 1326 |
rcu_read_lock(); tcred = __task_cred(task); |
b6dff3ec5 CRED: Separate ta... |
1327 1328 |
if (cred->euid != tcred->suid && cred->euid != tcred->uid && cred->uid != tcred->suid && cred->uid != tcred->uid && |
742755a1d [PATCH] page migr... |
1329 |
!capable(CAP_SYS_NICE)) { |
c69e8d9c0 CRED: Use RCU to ... |
1330 |
rcu_read_unlock(); |
742755a1d [PATCH] page migr... |
1331 |
err = -EPERM; |
5e9a0f023 mm: extract do_pa... |
1332 |
goto out; |
742755a1d [PATCH] page migr... |
1333 |
} |
c69e8d9c0 CRED: Use RCU to ... |
1334 |
rcu_read_unlock(); |
742755a1d [PATCH] page migr... |
1335 |
|
86c3a7645 [PATCH] SELinux: ... |
1336 1337 |
err = security_task_movememory(task); if (err) |
5e9a0f023 mm: extract do_pa... |
1338 |
goto out; |
86c3a7645 [PATCH] SELinux: ... |
1339 |
|
5e9a0f023 mm: extract do_pa... |
1340 1341 1342 1343 |
if (nodes) { err = do_pages_move(mm, task, nr_pages, pages, nodes, status, flags); } else { |
2f007e74b mm: don't vmalloc... |
1344 |
err = do_pages_stat(mm, nr_pages, pages, status); |
742755a1d [PATCH] page migr... |
1345 |
} |
742755a1d [PATCH] page migr... |
1346 |
out: |
742755a1d [PATCH] page migr... |
1347 1348 1349 |
mmput(mm); return err; } |
742755a1d [PATCH] page migr... |
1350 |
|
7b2259b3e [PATCH] page migr... |
1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 |
/* * 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 migration: migrat... |
1361 |
for (vma = mm->mmap; vma && !err; vma = vma->vm_next) { |
7b2259b3e [PATCH] page migr... |
1362 1363 1364 1365 1366 1367 1368 1369 |
if (vma->vm_ops && vma->vm_ops->migrate) { err = vma->vm_ops->migrate(vma, to, from, flags); if (err) break; } } return err; } |
83d1674a9 mm: make CONFIG_M... |
1370 |
#endif |