Blame view
mm/migrate.c
46.7 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> |
8e6ac7fab hugetlb/cgroup: m... |
36 |
#include <linux/hugetlb_cgroup.h> |
5a0e3ad6a include cleanup: ... |
37 |
#include <linux/gfp.h> |
bf6bddf19 mm: introduce com... |
38 |
#include <linux/balloon_compaction.h> |
f714f4f20 mm: numa: call MM... |
39 |
#include <linux/mmu_notifier.h> |
b20a35035 [PATCH] page migr... |
40 |
|
0d1836c36 mm/migrate.c: fix... |
41 |
#include <asm/tlbflush.h> |
7b2a2d4a1 mm: migrate: Add ... |
42 43 |
#define CREATE_TRACE_POINTS #include <trace/events/migrate.h> |
b20a35035 [PATCH] page migr... |
44 |
#include "internal.h" |
b20a35035 [PATCH] page migr... |
45 |
/* |
742755a1d [PATCH] page migr... |
46 |
* migrate_prep() needs to be called before we start compiling a list of pages |
748446bb6 mm: compaction: m... |
47 48 |
* to be migrated using isolate_lru_page(). If scheduling work on other CPUs is * undesirable, use migrate_prep_local() |
b20a35035 [PATCH] page migr... |
49 50 51 |
*/ int migrate_prep(void) { |
b20a35035 [PATCH] page migr... |
52 53 54 55 56 57 58 59 60 61 |
/* * 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... |
62 63 64 65 66 67 68 |
/* 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... |
69 |
/* |
5733c7d11 mm: introduce put... |
70 71 72 |
* Put previously isolated pages back onto the appropriate lists * from where they were once taken off for compaction/migration. * |
59c82b70d mm/migrate: remov... |
73 74 75 |
* This function shall be used whenever the isolated pageset has been * built from lru, balloon, hugetlbfs page. See isolate_migratepages_range() * and isolate_huge_page(). |
5733c7d11 mm: introduce put... |
76 77 78 79 80 81 82 |
*/ void putback_movable_pages(struct list_head *l) { struct page *page; struct page *page2; list_for_each_entry_safe(page, page2, l, lru) { |
31caf665e mm: migrate: make... |
83 84 85 86 |
if (unlikely(PageHuge(page))) { putback_active_hugepage(page); continue; } |
5733c7d11 mm: introduce put... |
87 88 89 |
list_del(&page->lru); dec_zone_page_state(page, NR_ISOLATED_ANON + page_is_file_cache(page)); |
117aad1e9 mm: avoid reinser... |
90 |
if (unlikely(isolated_balloon_page(page))) |
bf6bddf19 mm: introduce com... |
91 92 93 |
balloon_page_putback(page); else putback_lru_page(page); |
b20a35035 [PATCH] page migr... |
94 |
} |
b20a35035 [PATCH] page migr... |
95 |
} |
0697212a4 [PATCH] Swapless ... |
96 97 98 |
/* * Restore a potential migration pte to a working pte entry */ |
e9995ef97 ksm: rmap_walk to... |
99 100 |
static int remove_migration_pte(struct page *new, struct vm_area_struct *vma, unsigned long addr, void *old) |
0697212a4 [PATCH] Swapless ... |
101 102 103 |
{ struct mm_struct *mm = vma->vm_mm; swp_entry_t entry; |
0697212a4 [PATCH] Swapless ... |
104 105 106 |
pmd_t *pmd; pte_t *ptep, pte; spinlock_t *ptl; |
290408d4a hugetlb: hugepage... |
107 108 109 110 |
if (unlikely(PageHuge(new))) { ptep = huge_pte_offset(mm, addr); if (!ptep) goto out; |
cb900f412 mm, hugetlb: conv... |
111 |
ptl = huge_pte_lockptr(hstate_vma(vma), mm, ptep); |
290408d4a hugetlb: hugepage... |
112 |
} else { |
6219049ae mm: introduce mm_... |
113 114 |
pmd = mm_find_pmd(mm, addr); if (!pmd) |
290408d4a hugetlb: hugepage... |
115 |
goto out; |
0697212a4 [PATCH] Swapless ... |
116 |
|
290408d4a hugetlb: hugepage... |
117 |
ptep = pte_offset_map(pmd, addr); |
0697212a4 [PATCH] Swapless ... |
118 |
|
486cf46f3 mm: fix race betw... |
119 120 121 122 |
/* * 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... |
123 124 125 |
ptl = pte_lockptr(mm, pmd); } |
0697212a4 [PATCH] Swapless ... |
126 |
|
0697212a4 [PATCH] Swapless ... |
127 128 129 |
spin_lock(ptl); pte = *ptep; if (!is_swap_pte(pte)) |
e9995ef97 ksm: rmap_walk to... |
130 |
goto unlock; |
0697212a4 [PATCH] Swapless ... |
131 132 |
entry = pte_to_swp_entry(pte); |
e9995ef97 ksm: rmap_walk to... |
133 134 135 |
if (!is_migration_entry(entry) || migration_entry_to_page(entry) != old) goto unlock; |
0697212a4 [PATCH] Swapless ... |
136 |
|
0697212a4 [PATCH] Swapless ... |
137 138 |
get_page(new); pte = pte_mkold(mk_pte(new, vma->vm_page_prot)); |
c3d16e165 mm: migration: do... |
139 140 |
if (pte_swp_soft_dirty(*ptep)) pte = pte_mksoft_dirty(pte); |
d3cb8bf60 mm: migrate: Clos... |
141 142 |
/* Recheck VMA as permissions can change since migration started */ |
0697212a4 [PATCH] Swapless ... |
143 |
if (is_write_migration_entry(entry)) |
d3cb8bf60 mm: migrate: Clos... |
144 |
pte = maybe_mkwrite(pte, vma); |
3ef8fd7f7 Fix migration.c c... |
145 |
#ifdef CONFIG_HUGETLB_PAGE |
be7517d6a mm/hugetlb: set P... |
146 |
if (PageHuge(new)) { |
290408d4a hugetlb: hugepage... |
147 |
pte = pte_mkhuge(pte); |
be7517d6a mm/hugetlb: set P... |
148 149 |
pte = arch_make_huge_pte(pte, vma, new, 0); } |
3ef8fd7f7 Fix migration.c c... |
150 |
#endif |
c2cc499c5 mm compaction: fi... |
151 |
flush_dcache_page(new); |
0697212a4 [PATCH] Swapless ... |
152 |
set_pte_at(mm, addr, ptep, pte); |
04e62a29b [PATCH] More page... |
153 |
|
290408d4a hugetlb: hugepage... |
154 155 156 157 158 159 |
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... |
160 161 162 163 164 |
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... |
165 |
update_mmu_cache(vma, addr, ptep); |
e9995ef97 ksm: rmap_walk to... |
166 |
unlock: |
0697212a4 [PATCH] Swapless ... |
167 |
pte_unmap_unlock(ptep, ptl); |
e9995ef97 ksm: rmap_walk to... |
168 169 |
out: return SWAP_AGAIN; |
0697212a4 [PATCH] Swapless ... |
170 171 172 |
} /* |
04e62a29b [PATCH] More page... |
173 174 175 176 177 |
* 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) { |
051ac83ad mm/rmap: make rma... |
178 179 180 181 182 183 |
struct rmap_walk_control rwc = { .rmap_one = remove_migration_pte, .arg = old, }; rmap_walk(new, &rwc); |
04e62a29b [PATCH] More page... |
184 185 186 |
} /* |
0697212a4 [PATCH] Swapless ... |
187 188 189 |
* 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 ... |
190 |
*/ |
e66f17ff7 mm/hugetlb: take ... |
191 |
void __migration_entry_wait(struct mm_struct *mm, pte_t *ptep, |
30dad3092 mm: migration: ad... |
192 |
spinlock_t *ptl) |
0697212a4 [PATCH] Swapless ... |
193 |
{ |
30dad3092 mm: migration: ad... |
194 |
pte_t pte; |
0697212a4 [PATCH] Swapless ... |
195 196 |
swp_entry_t entry; struct page *page; |
30dad3092 mm: migration: ad... |
197 |
spin_lock(ptl); |
0697212a4 [PATCH] Swapless ... |
198 199 200 201 202 203 204 205 206 |
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... |
207 208 209 210 211 212 213 214 215 |
/* * 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 ... |
216 217 218 219 220 221 222 |
pte_unmap_unlock(ptep, ptl); wait_on_page_locked(page); put_page(page); return; out: pte_unmap_unlock(ptep, ptl); } |
30dad3092 mm: migration: ad... |
223 224 225 226 227 228 229 |
void migration_entry_wait(struct mm_struct *mm, pmd_t *pmd, unsigned long address) { spinlock_t *ptl = pte_lockptr(mm, pmd); pte_t *ptep = pte_offset_map(pmd, address); __migration_entry_wait(mm, ptep, ptl); } |
cb900f412 mm, hugetlb: conv... |
230 231 |
void migration_entry_wait_huge(struct vm_area_struct *vma, struct mm_struct *mm, pte_t *pte) |
30dad3092 mm: migration: ad... |
232 |
{ |
cb900f412 mm, hugetlb: conv... |
233 |
spinlock_t *ptl = huge_pte_lockptr(hstate_vma(vma), mm, pte); |
30dad3092 mm: migration: ad... |
234 235 |
__migration_entry_wait(mm, pte, ptl); } |
b969c4ab9 mm: compaction: d... |
236 237 |
#ifdef CONFIG_BLOCK /* Returns true if all buffers are successfully locked */ |
a6bc32b89 mm: compaction: i... |
238 239 |
static bool buffer_migrate_lock_buffers(struct buffer_head *head, enum migrate_mode mode) |
b969c4ab9 mm: compaction: d... |
240 241 242 243 |
{ struct buffer_head *bh = head; /* Simple case, sync compaction */ |
a6bc32b89 mm: compaction: i... |
244 |
if (mode != MIGRATE_ASYNC) { |
b969c4ab9 mm: compaction: d... |
245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 |
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... |
280 |
enum migrate_mode mode) |
b969c4ab9 mm: compaction: d... |
281 282 283 284 |
{ return true; } #endif /* CONFIG_BLOCK */ |
b20a35035 [PATCH] page migr... |
285 |
/* |
c3fcf8a5d [PATCH] page migr... |
286 |
* Replace the page in the mapping. |
5b5c7120e [PATCH] page migr... |
287 288 289 290 |
* * The number of remaining references must be: * 1 for anonymous pages without a mapping * 2 for pages with a mapping |
266cf658e FS-Cache: Recruit... |
291 |
* 3 for pages with a mapping and PagePrivate/PagePrivate2 set. |
b20a35035 [PATCH] page migr... |
292 |
*/ |
36bc08cc0 fs/aio: Add suppo... |
293 |
int migrate_page_move_mapping(struct address_space *mapping, |
b969c4ab9 mm: compaction: d... |
294 |
struct page *newpage, struct page *page, |
8e321fefb aio/migratepages:... |
295 296 |
struct buffer_head *head, enum migrate_mode mode, int extra_count) |
b20a35035 [PATCH] page migr... |
297 |
{ |
8e321fefb aio/migratepages:... |
298 |
int expected_count = 1 + extra_count; |
7cf9c2c76 [PATCH] radix-tre... |
299 |
void **pslot; |
b20a35035 [PATCH] page migr... |
300 |
|
6c5240ae7 [PATCH] Swapless ... |
301 |
if (!mapping) { |
0e8c7d0fd page migration: f... |
302 |
/* Anonymous page without mapping */ |
8e321fefb aio/migratepages:... |
303 |
if (page_count(page) != expected_count) |
6c5240ae7 [PATCH] Swapless ... |
304 |
return -EAGAIN; |
78bd52097 mm: adjust addres... |
305 |
return MIGRATEPAGE_SUCCESS; |
6c5240ae7 [PATCH] Swapless ... |
306 |
} |
19fd62312 mm: spinlock tree... |
307 |
spin_lock_irq(&mapping->tree_lock); |
b20a35035 [PATCH] page migr... |
308 |
|
7cf9c2c76 [PATCH] radix-tre... |
309 310 |
pslot = radix_tree_lookup_slot(&mapping->page_tree, page_index(page)); |
b20a35035 [PATCH] page migr... |
311 |
|
8e321fefb aio/migratepages:... |
312 |
expected_count += 1 + page_has_private(page); |
e286781d5 mm: speculative p... |
313 |
if (page_count(page) != expected_count || |
29c1f677d mm: migration: us... |
314 |
radix_tree_deref_slot_protected(pslot, &mapping->tree_lock) != page) { |
19fd62312 mm: spinlock tree... |
315 |
spin_unlock_irq(&mapping->tree_lock); |
e23ca00bf [PATCH] Some page... |
316 |
return -EAGAIN; |
b20a35035 [PATCH] page migr... |
317 |
} |
e286781d5 mm: speculative p... |
318 |
if (!page_freeze_refs(page, expected_count)) { |
19fd62312 mm: spinlock tree... |
319 |
spin_unlock_irq(&mapping->tree_lock); |
e286781d5 mm: speculative p... |
320 321 |
return -EAGAIN; } |
b20a35035 [PATCH] page migr... |
322 |
/* |
b969c4ab9 mm: compaction: d... |
323 324 325 326 327 328 |
* 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... |
329 330 |
if (mode == MIGRATE_ASYNC && head && !buffer_migrate_lock_buffers(head, mode)) { |
b969c4ab9 mm: compaction: d... |
331 332 333 334 335 336 |
page_unfreeze_refs(page, expected_count); spin_unlock_irq(&mapping->tree_lock); return -EAGAIN; } /* |
b20a35035 [PATCH] page migr... |
337 |
* Now we know that no one else is looking at the page. |
b20a35035 [PATCH] page migr... |
338 |
*/ |
7cf9c2c76 [PATCH] radix-tre... |
339 |
get_page(newpage); /* add cache reference */ |
b20a35035 [PATCH] page migr... |
340 341 342 343 |
if (PageSwapCache(page)) { SetPageSwapCache(newpage); set_page_private(newpage, page_private(page)); } |
7cf9c2c76 [PATCH] radix-tre... |
344 345 346 |
radix_tree_replace_slot(pslot, newpage); /* |
937a94c9d mm: migrate: one ... |
347 348 |
* Drop cache reference from old page by unfreezing * to one less reference. |
7cf9c2c76 [PATCH] radix-tre... |
349 350 |
* We know this isn't the last reference. */ |
937a94c9d mm: migrate: one ... |
351 |
page_unfreeze_refs(page, expected_count - 1); |
7cf9c2c76 [PATCH] radix-tre... |
352 |
|
0e8c7d0fd page migration: f... |
353 354 355 356 357 358 359 360 361 362 363 364 |
/* * 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... |
365 |
if (!PageSwapCache(page) && PageSwapBacked(page)) { |
4b02108ac mm: oom analysis:... |
366 367 368 |
__dec_zone_page_state(page, NR_SHMEM); __inc_zone_page_state(newpage, NR_SHMEM); } |
19fd62312 mm: spinlock tree... |
369 |
spin_unlock_irq(&mapping->tree_lock); |
b20a35035 [PATCH] page migr... |
370 |
|
78bd52097 mm: adjust addres... |
371 |
return MIGRATEPAGE_SUCCESS; |
b20a35035 [PATCH] page migr... |
372 |
} |
b20a35035 [PATCH] page migr... |
373 374 |
/* |
290408d4a hugetlb: hugepage... |
375 376 377 378 379 380 381 382 383 384 385 386 |
* 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; |
78bd52097 mm: adjust addres... |
387 |
return MIGRATEPAGE_SUCCESS; |
290408d4a hugetlb: hugepage... |
388 389 390 391 392 393 394 395 396 |
} 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... |
397 |
radix_tree_deref_slot_protected(pslot, &mapping->tree_lock) != page) { |
290408d4a hugetlb: hugepage... |
398 399 400 401 402 403 404 405 406 407 408 409 |
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 ... |
410 |
page_unfreeze_refs(page, expected_count - 1); |
290408d4a hugetlb: hugepage... |
411 412 |
spin_unlock_irq(&mapping->tree_lock); |
78bd52097 mm: adjust addres... |
413 |
return MIGRATEPAGE_SUCCESS; |
290408d4a hugetlb: hugepage... |
414 415 416 |
} /* |
30b0a105d mm: thp: give tra... |
417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 |
* Gigantic pages are so large that we do not guarantee that page++ pointer * arithmetic will work across the entire page. We need something more * specialized. */ static void __copy_gigantic_page(struct page *dst, struct page *src, int nr_pages) { int i; struct page *dst_base = dst; struct page *src_base = src; for (i = 0; i < nr_pages; ) { cond_resched(); copy_highpage(dst, src); i++; dst = mem_map_next(dst, dst_base, i); src = mem_map_next(src, src_base, i); } } static void copy_huge_page(struct page *dst, struct page *src) { int i; int nr_pages; if (PageHuge(src)) { /* hugetlbfs page */ struct hstate *h = page_hstate(src); nr_pages = pages_per_huge_page(h); if (unlikely(nr_pages > MAX_ORDER_NR_PAGES)) { __copy_gigantic_page(dst, src, nr_pages); return; } } else { /* thp page */ BUG_ON(!PageTransHuge(src)); nr_pages = hpage_nr_pages(src); } for (i = 0; i < nr_pages; i++) { cond_resched(); copy_highpage(dst + i, src + i); } } /* |
b20a35035 [PATCH] page migr... |
465 466 |
* Copy the page to its new location */ |
290408d4a hugetlb: hugepage... |
467 |
void migrate_page_copy(struct page *newpage, struct page *page) |
b20a35035 [PATCH] page migr... |
468 |
{ |
7851a45cd mm: numa: Copy cp... |
469 |
int cpupid; |
b32967ff1 mm: numa: Add THP... |
470 |
if (PageHuge(page) || PageTransHuge(page)) |
290408d4a hugetlb: hugepage... |
471 472 473 |
copy_huge_page(newpage, page); else copy_highpage(newpage, page); |
b20a35035 [PATCH] page migr... |
474 475 476 477 478 479 480 |
if (PageError(page)) SetPageError(newpage); if (PageReferenced(page)) SetPageReferenced(newpage); if (PageUptodate(page)) SetPageUptodate(newpage); |
894bc3104 Unevictable LRU I... |
481 |
if (TestClearPageActive(page)) { |
309381fea mm: dump page whe... |
482 |
VM_BUG_ON_PAGE(PageUnevictable(page), page); |
b20a35035 [PATCH] page migr... |
483 |
SetPageActive(newpage); |
418b27ef5 mm: remove unevic... |
484 485 |
} else if (TestClearPageUnevictable(page)) SetPageUnevictable(newpage); |
b20a35035 [PATCH] page migr... |
486 487 488 489 490 491 492 |
if (PageChecked(page)) SetPageChecked(newpage); if (PageMappedToDisk(page)) SetPageMappedToDisk(newpage); if (PageDirty(page)) { clear_page_dirty_for_io(page); |
3a902c5f6 mm: fix warning o... |
493 494 495 496 497 |
/* * 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... |
498 |
* Whereas only part of our page may be dirty. |
3a902c5f6 mm: fix warning o... |
499 |
*/ |
752dc185d mm: fix warning i... |
500 501 502 503 |
if (PageSwapBacked(page)) SetPageDirty(newpage); else __set_page_dirty_nobuffers(newpage); |
b20a35035 [PATCH] page migr... |
504 |
} |
7851a45cd mm: numa: Copy cp... |
505 506 507 508 509 510 |
/* * Copy NUMA information to the new page, to prevent over-eager * future migrations of this same page. */ cpupid = page_cpupid_xchg_last(page, -1); page_cpupid_xchg_last(newpage, cpupid); |
b291f0003 mlock: mlocked pa... |
511 |
mlock_migrate_page(newpage, page); |
e9995ef97 ksm: rmap_walk to... |
512 |
ksm_migrate_page(newpage, page); |
c8d6553b9 ksm: make KSM pag... |
513 514 515 516 |
/* * Please do not reorder this without considering how mm/ksm.c's * get_ksm_page() depends upon ksm_migrate_page() and PageSwapCache(). */ |
b3b3a99c5 mm/migrate: check... |
517 518 |
if (PageSwapCache(page)) ClearPageSwapCache(page); |
b20a35035 [PATCH] page migr... |
519 520 |
ClearPagePrivate(page); set_page_private(page, 0); |
b20a35035 [PATCH] page migr... |
521 522 523 524 525 526 527 528 |
/* * If any waiters have accumulated on the new page then * wake them up. */ if (PageWriteback(newpage)) end_page_writeback(newpage); } |
b20a35035 [PATCH] page migr... |
529 |
|
1d8b85ccf [PATCH] page migr... |
530 531 532 |
/************************************************************ * Migration functions ***********************************************************/ |
b20a35035 [PATCH] page migr... |
533 534 |
/* * Common logic to directly migrate a single page suitable for |
266cf658e FS-Cache: Recruit... |
535 |
* pages that do not use PagePrivate/PagePrivate2. |
b20a35035 [PATCH] page migr... |
536 537 538 |
* * Pages are locked upon entry and exit. */ |
2d1db3b11 [PATCH] page migr... |
539 |
int migrate_page(struct address_space *mapping, |
a6bc32b89 mm: compaction: i... |
540 541 |
struct page *newpage, struct page *page, enum migrate_mode mode) |
b20a35035 [PATCH] page migr... |
542 543 544 545 |
{ int rc; BUG_ON(PageWriteback(page)); /* Writeback must be complete */ |
8e321fefb aio/migratepages:... |
546 |
rc = migrate_page_move_mapping(mapping, newpage, page, NULL, mode, 0); |
b20a35035 [PATCH] page migr... |
547 |
|
78bd52097 mm: adjust addres... |
548 |
if (rc != MIGRATEPAGE_SUCCESS) |
b20a35035 [PATCH] page migr... |
549 550 551 |
return rc; migrate_page_copy(newpage, page); |
78bd52097 mm: adjust addres... |
552 |
return MIGRATEPAGE_SUCCESS; |
b20a35035 [PATCH] page migr... |
553 554 |
} EXPORT_SYMBOL(migrate_page); |
9361401eb [PATCH] BLOCK: Ma... |
555 |
#ifdef CONFIG_BLOCK |
b20a35035 [PATCH] page migr... |
556 |
/* |
1d8b85ccf [PATCH] page migr... |
557 558 559 560 |
* 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... |
561 |
int buffer_migrate_page(struct address_space *mapping, |
a6bc32b89 mm: compaction: i... |
562 |
struct page *newpage, struct page *page, enum migrate_mode mode) |
1d8b85ccf [PATCH] page migr... |
563 |
{ |
1d8b85ccf [PATCH] page migr... |
564 565 |
struct buffer_head *bh, *head; int rc; |
1d8b85ccf [PATCH] page migr... |
566 |
if (!page_has_buffers(page)) |
a6bc32b89 mm: compaction: i... |
567 |
return migrate_page(mapping, newpage, page, mode); |
1d8b85ccf [PATCH] page migr... |
568 569 |
head = page_buffers(page); |
8e321fefb aio/migratepages:... |
570 |
rc = migrate_page_move_mapping(mapping, newpage, page, head, mode, 0); |
1d8b85ccf [PATCH] page migr... |
571 |
|
78bd52097 mm: adjust addres... |
572 |
if (rc != MIGRATEPAGE_SUCCESS) |
1d8b85ccf [PATCH] page migr... |
573 |
return rc; |
b969c4ab9 mm: compaction: d... |
574 575 576 577 578 |
/* * 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... |
579 580 |
if (mode != MIGRATE_ASYNC) BUG_ON(!buffer_migrate_lock_buffers(head, mode)); |
1d8b85ccf [PATCH] page migr... |
581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 |
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); |
78bd52097 mm: adjust addres... |
606 |
return MIGRATEPAGE_SUCCESS; |
1d8b85ccf [PATCH] page migr... |
607 608 |
} EXPORT_SYMBOL(buffer_migrate_page); |
9361401eb [PATCH] BLOCK: Ma... |
609 |
#endif |
1d8b85ccf [PATCH] page migr... |
610 |
|
04e62a29b [PATCH] More page... |
611 612 613 614 |
/* * Writeback a page to clean the dirty state */ static int writeout(struct address_space *mapping, struct page *page) |
8351a6e47 [PATCH] page migr... |
615 |
{ |
04e62a29b [PATCH] More page... |
616 617 618 619 620 |
struct writeback_control wbc = { .sync_mode = WB_SYNC_NONE, .nr_to_write = 1, .range_start = 0, .range_end = LLONG_MAX, |
04e62a29b [PATCH] More page... |
621 622 623 624 625 626 627 628 629 630 631 |
.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... |
632 |
/* |
04e62a29b [PATCH] More page... |
633 634 635 636 637 638 |
* 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... |
639 |
*/ |
04e62a29b [PATCH] More page... |
640 |
remove_migration_ptes(page, page); |
8351a6e47 [PATCH] page migr... |
641 |
|
04e62a29b [PATCH] More page... |
642 |
rc = mapping->a_ops->writepage(page, &wbc); |
8351a6e47 [PATCH] page migr... |
643 |
|
04e62a29b [PATCH] More page... |
644 645 646 |
if (rc != AOP_WRITEPAGE_ACTIVATE) /* unlocked. Relock */ lock_page(page); |
bda8550de migration: fix wr... |
647 |
return (rc < 0) ? -EIO : -EAGAIN; |
04e62a29b [PATCH] More page... |
648 649 650 651 652 653 |
} /* * Default handling if a filesystem does not provide a migration function. */ static int fallback_migrate_page(struct address_space *mapping, |
a6bc32b89 mm: compaction: i... |
654 |
struct page *newpage, struct page *page, enum migrate_mode mode) |
04e62a29b [PATCH] More page... |
655 |
{ |
b969c4ab9 mm: compaction: d... |
656 |
if (PageDirty(page)) { |
a6bc32b89 mm: compaction: i... |
657 658 |
/* Only writeback pages in full synchronous migration */ if (mode != MIGRATE_SYNC) |
b969c4ab9 mm: compaction: d... |
659 |
return -EBUSY; |
04e62a29b [PATCH] More page... |
660 |
return writeout(mapping, page); |
b969c4ab9 mm: compaction: d... |
661 |
} |
8351a6e47 [PATCH] page migr... |
662 663 664 665 666 |
/* * Buffers may be managed in a filesystem specific way. * We must have no buffers or drop them. */ |
266cf658e FS-Cache: Recruit... |
667 |
if (page_has_private(page) && |
8351a6e47 [PATCH] page migr... |
668 669 |
!try_to_release_page(page, GFP_KERNEL)) return -EAGAIN; |
a6bc32b89 mm: compaction: i... |
670 |
return migrate_page(mapping, newpage, page, mode); |
8351a6e47 [PATCH] page migr... |
671 |
} |
1d8b85ccf [PATCH] page migr... |
672 |
/* |
e24f0b8f7 [PATCH] page migr... |
673 674 675 676 677 |
* 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... |
678 679 680 |
* * Return value: * < 0 - error code |
78bd52097 mm: adjust addres... |
681 |
* MIGRATEPAGE_SUCCESS - success |
e24f0b8f7 [PATCH] page migr... |
682 |
*/ |
3fe2011ff mm: migration: al... |
683 |
static int move_to_new_page(struct page *newpage, struct page *page, |
2ebba6b7e mm: unmapped page... |
684 |
int page_was_mapped, enum migrate_mode mode) |
e24f0b8f7 [PATCH] page migr... |
685 686 687 688 689 690 691 692 693 |
{ 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... |
694 |
if (!trylock_page(newpage)) |
e24f0b8f7 [PATCH] page migr... |
695 696 697 698 699 |
BUG(); /* Prepare mapping for the new page.*/ newpage->index = page->index; newpage->mapping = page->mapping; |
b2e185384 define page_file_... |
700 701 |
if (PageSwapBacked(page)) SetPageSwapBacked(newpage); |
e24f0b8f7 [PATCH] page migr... |
702 703 704 |
mapping = page_mapping(page); if (!mapping) |
a6bc32b89 mm: compaction: i... |
705 |
rc = migrate_page(mapping, newpage, page, mode); |
b969c4ab9 mm: compaction: d... |
706 |
else if (mapping->a_ops->migratepage) |
e24f0b8f7 [PATCH] page migr... |
707 |
/* |
b969c4ab9 mm: compaction: d... |
708 709 710 711 |
* 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... |
712 |
*/ |
b969c4ab9 mm: compaction: d... |
713 |
rc = mapping->a_ops->migratepage(mapping, |
a6bc32b89 mm: compaction: i... |
714 |
newpage, page, mode); |
b969c4ab9 mm: compaction: d... |
715 |
else |
a6bc32b89 mm: compaction: i... |
716 |
rc = fallback_migrate_page(mapping, newpage, page, mode); |
e24f0b8f7 [PATCH] page migr... |
717 |
|
78bd52097 mm: adjust addres... |
718 |
if (rc != MIGRATEPAGE_SUCCESS) { |
e24f0b8f7 [PATCH] page migr... |
719 |
newpage->mapping = NULL; |
3fe2011ff mm: migration: al... |
720 |
} else { |
0a31bc97c mm: memcontrol: r... |
721 |
mem_cgroup_migrate(page, newpage, false); |
2ebba6b7e mm: unmapped page... |
722 |
if (page_was_mapped) |
3fe2011ff mm: migration: al... |
723 |
remove_migration_ptes(page, newpage); |
35512ecae mm: postpone migr... |
724 |
page->mapping = NULL; |
3fe2011ff mm: migration: al... |
725 |
} |
e24f0b8f7 [PATCH] page migr... |
726 727 728 729 730 |
unlock_page(newpage); return rc; } |
0dabec93d mm: migration: cl... |
731 |
static int __unmap_and_move(struct page *page, struct page *newpage, |
9c620e2bc mm: remove offlin... |
732 |
int force, enum migrate_mode mode) |
e24f0b8f7 [PATCH] page migr... |
733 |
{ |
0dabec93d mm: migration: cl... |
734 |
int rc = -EAGAIN; |
2ebba6b7e mm: unmapped page... |
735 |
int page_was_mapped = 0; |
3f6c82728 mm: migration: ta... |
736 |
struct anon_vma *anon_vma = NULL; |
95a402c38 [PATCH] page migr... |
737 |
|
529ae9aaa mm: rename page t... |
738 |
if (!trylock_page(page)) { |
a6bc32b89 mm: compaction: i... |
739 |
if (!force || mode == MIGRATE_ASYNC) |
0dabec93d mm: migration: cl... |
740 |
goto out; |
3e7d34497 mm: vmscan: recla... |
741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 |
/* * 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... |
756 |
goto out; |
3e7d34497 mm: vmscan: recla... |
757 |
|
e24f0b8f7 [PATCH] page migr... |
758 759 760 761 |
lock_page(page); } if (PageWriteback(page)) { |
11bc82d67 mm: compaction: U... |
762 |
/* |
fed5b64a9 mm/migrate: fix c... |
763 |
* Only in the case of a full synchronous migration is it |
a6bc32b89 mm: compaction: i... |
764 765 766 |
* 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... |
767 |
*/ |
a6bc32b89 mm: compaction: i... |
768 |
if (mode != MIGRATE_SYNC) { |
11bc82d67 mm: compaction: U... |
769 |
rc = -EBUSY; |
0a31bc97c mm: memcontrol: r... |
770 |
goto out_unlock; |
11bc82d67 mm: compaction: U... |
771 772 |
} if (!force) |
0a31bc97c mm: memcontrol: r... |
773 |
goto out_unlock; |
e24f0b8f7 [PATCH] page migr... |
774 775 |
wait_on_page_writeback(page); } |
e24f0b8f7 [PATCH] page migr... |
776 |
/* |
dc386d4d1 memory unplug: mi... |
777 778 |
* 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... |
779 |
* This get_anon_vma() delays freeing anon_vma pointer until the end |
dc386d4d1 memory unplug: mi... |
780 |
* of migration. File cache pages are no problem because of page_lock() |
989f89c57 fix rcu_read_lock... |
781 782 |
* File Caches may use write_page() or lock_page() in migration, then, * just care Anon page here. |
dc386d4d1 memory unplug: mi... |
783 |
*/ |
b79bc0a0c ksm: enable KSM p... |
784 |
if (PageAnon(page) && !PageKsm(page)) { |
1ce82b69e mm: fix migration... |
785 |
/* |
4fc3f1d66 mm/rmap, migratio... |
786 |
* Only page_lock_anon_vma_read() understands the subtleties of |
1ce82b69e mm: fix migration... |
787 788 |
* getting a hold on an anon_vma from outside one of its mms. */ |
746b18d42 mm: use refcounts... |
789 |
anon_vma = page_get_anon_vma(page); |
1ce82b69e mm: fix migration... |
790 791 |
if (anon_vma) { /* |
746b18d42 mm: use refcounts... |
792 |
* Anon page |
1ce82b69e mm: fix migration... |
793 |
*/ |
1ce82b69e mm: fix migration... |
794 |
} else if (PageSwapCache(page)) { |
3fe2011ff mm: migration: al... |
795 796 797 798 799 800 801 802 803 804 805 806 |
/* * 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 */ |
3fe2011ff mm: migration: al... |
807 |
} else { |
0a31bc97c mm: memcontrol: r... |
808 |
goto out_unlock; |
3fe2011ff mm: migration: al... |
809 |
} |
989f89c57 fix rcu_read_lock... |
810 |
} |
62e1c5530 page migraton: ha... |
811 |
|
d6d86c0a7 mm/balloon_compac... |
812 |
if (unlikely(isolated_balloon_page(page))) { |
bf6bddf19 mm: introduce com... |
813 814 815 816 817 818 819 820 |
/* * A ballooned page does not need any special attention from * physical to virtual reverse mapping procedures. * Skip any attempt to unmap PTEs or to remap swap cache, * in order to avoid burning cycles at rmap level, and perform * the page migration right away (proteced by page lock). */ rc = balloon_page_migrate(newpage, page, mode); |
0a31bc97c mm: memcontrol: r... |
821 |
goto out_unlock; |
bf6bddf19 mm: introduce com... |
822 |
} |
dc386d4d1 memory unplug: mi... |
823 |
/* |
62e1c5530 page migraton: ha... |
824 825 826 827 828 829 830 831 832 833 |
* 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... |
834 |
*/ |
62e1c5530 page migraton: ha... |
835 |
if (!page->mapping) { |
309381fea mm: dump page whe... |
836 |
VM_BUG_ON_PAGE(PageAnon(page), page); |
1ce82b69e mm: fix migration... |
837 |
if (page_has_private(page)) { |
62e1c5530 page migraton: ha... |
838 |
try_to_free_buffers(page); |
0a31bc97c mm: memcontrol: r... |
839 |
goto out_unlock; |
62e1c5530 page migraton: ha... |
840 |
} |
abfc34881 memory hotplug: m... |
841 |
goto skip_unmap; |
62e1c5530 page migraton: ha... |
842 |
} |
dc386d4d1 memory unplug: mi... |
843 |
/* Establish migration ptes or remove ptes */ |
2ebba6b7e mm: unmapped page... |
844 845 846 847 848 |
if (page_mapped(page)) { try_to_unmap(page, TTU_MIGRATION|TTU_IGNORE_MLOCK|TTU_IGNORE_ACCESS); page_was_mapped = 1; } |
dc386d4d1 memory unplug: mi... |
849 |
|
abfc34881 memory hotplug: m... |
850 |
skip_unmap: |
e6a1530d6 [PATCH] Allow mig... |
851 |
if (!page_mapped(page)) |
2ebba6b7e mm: unmapped page... |
852 |
rc = move_to_new_page(newpage, page, page_was_mapped, mode); |
e24f0b8f7 [PATCH] page migr... |
853 |
|
2ebba6b7e mm: unmapped page... |
854 |
if (rc && page_was_mapped) |
e24f0b8f7 [PATCH] page migr... |
855 |
remove_migration_ptes(page, page); |
3f6c82728 mm: migration: ta... |
856 857 |
/* Drop an anon_vma reference if we took one */ |
76545066c mm: extend KSM re... |
858 |
if (anon_vma) |
9e60109f1 mm: rename drop_a... |
859 |
put_anon_vma(anon_vma); |
3f6c82728 mm: migration: ta... |
860 |
|
0a31bc97c mm: memcontrol: r... |
861 |
out_unlock: |
e24f0b8f7 [PATCH] page migr... |
862 |
unlock_page(page); |
0dabec93d mm: migration: cl... |
863 864 865 |
out: return rc; } |
95a402c38 [PATCH] page migr... |
866 |
|
0dabec93d mm: migration: cl... |
867 |
/* |
ef2a5153b mm/migrate: mark ... |
868 869 870 871 872 873 874 875 876 877 |
* gcc 4.7 and 4.8 on arm get an ICEs when inlining unmap_and_move(). Work * around it. */ #if (GCC_VERSION >= 40700 && GCC_VERSION < 40900) && defined(CONFIG_ARM) #define ICE_noinline noinline #else #define ICE_noinline #endif /* |
0dabec93d mm: migration: cl... |
878 879 880 |
* Obtain the lock on page, remove all ptes and migrate the page * to the newly allocated page in newpage. */ |
ef2a5153b mm/migrate: mark ... |
881 882 883 884 |
static ICE_noinline int unmap_and_move(new_page_t get_new_page, free_page_t put_new_page, unsigned long private, struct page *page, int force, enum migrate_mode mode) |
0dabec93d mm: migration: cl... |
885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 |
{ int rc = 0; int *result = NULL; struct page *newpage = get_new_page(page, private, &result); if (!newpage) return -ENOMEM; if (page_count(page) == 1) { /* page was freed from under us. So we are done. */ goto out; } if (unlikely(PageTransHuge(page))) if (unlikely(split_huge_page(page))) goto out; |
9c620e2bc mm: remove offlin... |
901 |
rc = __unmap_and_move(page, newpage, force, mode); |
bf6bddf19 mm: introduce com... |
902 |
|
0dabec93d mm: migration: cl... |
903 |
out: |
e24f0b8f7 [PATCH] page migr... |
904 |
if (rc != -EAGAIN) { |
0dabec93d mm: migration: cl... |
905 906 907 908 909 910 911 |
/* * 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... |
912 |
dec_zone_page_state(page, NR_ISOLATED_ANON + |
6c0b13519 mm: return boolea... |
913 |
page_is_file_cache(page)); |
894bc3104 Unevictable LRU I... |
914 |
putback_lru_page(page); |
e24f0b8f7 [PATCH] page migr... |
915 |
} |
68711a746 mm, migration: ad... |
916 |
|
95a402c38 [PATCH] page migr... |
917 |
/* |
68711a746 mm, migration: ad... |
918 919 920 |
* If migration was not successful and there's a freeing callback, use * it. Otherwise, putback_lru_page() will drop the reference grabbed * during isolation. |
95a402c38 [PATCH] page migr... |
921 |
*/ |
8bdd63809 mm: fix direct re... |
922 923 |
if (rc != MIGRATEPAGE_SUCCESS && put_new_page) { ClearPageSwapBacked(newpage); |
68711a746 mm, migration: ad... |
924 |
put_new_page(newpage, private); |
d6d86c0a7 mm/balloon_compac... |
925 926 927 |
} else if (unlikely(__is_movable_balloon_page(newpage))) { /* drop our reference, page already in the balloon */ put_page(newpage); |
8bdd63809 mm: fix direct re... |
928 |
} else |
68711a746 mm, migration: ad... |
929 |
putback_lru_page(newpage); |
742755a1d [PATCH] page migr... |
930 931 932 933 934 935 |
if (result) { if (rc) *result = rc; else *result = page_to_nid(newpage); } |
e24f0b8f7 [PATCH] page migr... |
936 937 938 939 |
return rc; } /* |
290408d4a hugetlb: hugepage... |
940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 |
* 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, |
68711a746 mm, migration: ad... |
958 959 960 |
free_page_t put_new_page, unsigned long private, struct page *hpage, int force, enum migrate_mode mode) |
290408d4a hugetlb: hugepage... |
961 962 963 |
{ int rc = 0; int *result = NULL; |
2ebba6b7e mm: unmapped page... |
964 |
int page_was_mapped = 0; |
32665f2bb mm/migrate: corre... |
965 |
struct page *new_hpage; |
290408d4a hugetlb: hugepage... |
966 |
struct anon_vma *anon_vma = NULL; |
83467efbd mm: migrate: chec... |
967 968 969 970 971 972 973 |
/* * Movability of hugepages depends on architectures and hugepage size. * This check is necessary because some callers of hugepage migration * like soft offline and memory hotremove don't walk through page * tables or check whether the hugepage is pmd-based or not before * kicking migration. */ |
100873d7a hugetlb: rename h... |
974 |
if (!hugepage_migration_supported(page_hstate(hpage))) { |
32665f2bb mm/migrate: corre... |
975 |
putback_active_hugepage(hpage); |
83467efbd mm: migrate: chec... |
976 |
return -ENOSYS; |
32665f2bb mm/migrate: corre... |
977 |
} |
83467efbd mm: migrate: chec... |
978 |
|
32665f2bb mm/migrate: corre... |
979 |
new_hpage = get_new_page(hpage, private, &result); |
290408d4a hugetlb: hugepage... |
980 981 982 983 984 985 |
if (!new_hpage) return -ENOMEM; rc = -EAGAIN; if (!trylock_page(hpage)) { |
a6bc32b89 mm: compaction: i... |
986 |
if (!force || mode != MIGRATE_SYNC) |
290408d4a hugetlb: hugepage... |
987 988 989 |
goto out; lock_page(hpage); } |
746b18d42 mm: use refcounts... |
990 991 |
if (PageAnon(hpage)) anon_vma = page_get_anon_vma(hpage); |
290408d4a hugetlb: hugepage... |
992 |
|
2ebba6b7e mm: unmapped page... |
993 994 995 996 997 |
if (page_mapped(hpage)) { try_to_unmap(hpage, TTU_MIGRATION|TTU_IGNORE_MLOCK|TTU_IGNORE_ACCESS); page_was_mapped = 1; } |
290408d4a hugetlb: hugepage... |
998 999 |
if (!page_mapped(hpage)) |
2ebba6b7e mm: unmapped page... |
1000 |
rc = move_to_new_page(new_hpage, hpage, page_was_mapped, mode); |
290408d4a hugetlb: hugepage... |
1001 |
|
2ebba6b7e mm: unmapped page... |
1002 |
if (rc != MIGRATEPAGE_SUCCESS && page_was_mapped) |
290408d4a hugetlb: hugepage... |
1003 |
remove_migration_ptes(hpage, hpage); |
fd4a4663d mm: fix hugepage ... |
1004 |
if (anon_vma) |
9e60109f1 mm: rename drop_a... |
1005 |
put_anon_vma(anon_vma); |
8e6ac7fab hugetlb/cgroup: m... |
1006 |
|
68711a746 mm, migration: ad... |
1007 |
if (rc == MIGRATEPAGE_SUCCESS) |
8e6ac7fab hugetlb/cgroup: m... |
1008 |
hugetlb_cgroup_migrate(hpage, new_hpage); |
290408d4a hugetlb: hugepage... |
1009 |
unlock_page(hpage); |
09761333e mm/migrate.c: pai... |
1010 |
out: |
b8ec1cee5 mm: soft-offline:... |
1011 1012 |
if (rc != -EAGAIN) putback_active_hugepage(hpage); |
68711a746 mm, migration: ad... |
1013 1014 1015 1016 1017 1018 1019 1020 1021 |
/* * If migration was not successful and there's a freeing callback, use * it. Otherwise, put_page() will drop the reference grabbed during * isolation. */ if (rc != MIGRATEPAGE_SUCCESS && put_new_page) put_new_page(new_hpage, private); else |
9a36019b8 mm: migrate: huge... |
1022 |
putback_active_hugepage(new_hpage); |
68711a746 mm, migration: ad... |
1023 |
|
290408d4a hugetlb: hugepage... |
1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 |
if (result) { if (rc) *result = rc; else *result = page_to_nid(new_hpage); } return rc; } /* |
c73e5c9c5 mm: rewrite the c... |
1034 1035 |
* migrate_pages - migrate the pages specified in a list, to the free pages * supplied as the target for the page migration |
b20a35035 [PATCH] page migr... |
1036 |
* |
c73e5c9c5 mm: rewrite the c... |
1037 1038 1039 |
* @from: The list of pages to be migrated. * @get_new_page: The function used to allocate free pages to be used * as the target of the page migration. |
68711a746 mm, migration: ad... |
1040 1041 |
* @put_new_page: The function used to free target pages if migration * fails, or NULL if no special handling is necessary. |
c73e5c9c5 mm: rewrite the c... |
1042 1043 1044 1045 |
* @private: Private data to be passed on to get_new_page() * @mode: The migration mode that specifies the constraints for * page migration, if any. * @reason: The reason for page migration. |
b20a35035 [PATCH] page migr... |
1046 |
* |
c73e5c9c5 mm: rewrite the c... |
1047 1048 1049 |
* The function returns after 10 attempts or if no pages are movable any more * because the list has become empty or no retryable pages exist any more. * The caller should call putback_lru_pages() to return pages to the LRU |
28bd65781 mm: migration: cl... |
1050 |
* or free list only if ret != 0. |
b20a35035 [PATCH] page migr... |
1051 |
* |
c73e5c9c5 mm: rewrite the c... |
1052 |
* Returns the number of pages that were not migrated, or an error code. |
b20a35035 [PATCH] page migr... |
1053 |
*/ |
9c620e2bc mm: remove offlin... |
1054 |
int migrate_pages(struct list_head *from, new_page_t get_new_page, |
68711a746 mm, migration: ad... |
1055 1056 |
free_page_t put_new_page, unsigned long private, enum migrate_mode mode, int reason) |
b20a35035 [PATCH] page migr... |
1057 |
{ |
e24f0b8f7 [PATCH] page migr... |
1058 |
int retry = 1; |
b20a35035 [PATCH] page migr... |
1059 |
int nr_failed = 0; |
5647bc293 mm: compaction: M... |
1060 |
int nr_succeeded = 0; |
b20a35035 [PATCH] page migr... |
1061 1062 1063 1064 1065 1066 1067 1068 |
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... |
1069 1070 |
for(pass = 0; pass < 10 && retry; pass++) { retry = 0; |
b20a35035 [PATCH] page migr... |
1071 |
|
e24f0b8f7 [PATCH] page migr... |
1072 |
list_for_each_entry_safe(page, page2, from, lru) { |
e24f0b8f7 [PATCH] page migr... |
1073 |
cond_resched(); |
2d1db3b11 [PATCH] page migr... |
1074 |
|
31caf665e mm: migrate: make... |
1075 1076 |
if (PageHuge(page)) rc = unmap_and_move_huge_page(get_new_page, |
68711a746 mm, migration: ad... |
1077 1078 |
put_new_page, private, page, pass > 2, mode); |
31caf665e mm: migrate: make... |
1079 |
else |
68711a746 mm, migration: ad... |
1080 1081 |
rc = unmap_and_move(get_new_page, put_new_page, private, page, pass > 2, mode); |
2d1db3b11 [PATCH] page migr... |
1082 |
|
e24f0b8f7 [PATCH] page migr... |
1083 |
switch(rc) { |
95a402c38 [PATCH] page migr... |
1084 1085 |
case -ENOMEM: goto out; |
e24f0b8f7 [PATCH] page migr... |
1086 |
case -EAGAIN: |
2d1db3b11 [PATCH] page migr... |
1087 |
retry++; |
e24f0b8f7 [PATCH] page migr... |
1088 |
break; |
78bd52097 mm: adjust addres... |
1089 |
case MIGRATEPAGE_SUCCESS: |
5647bc293 mm: compaction: M... |
1090 |
nr_succeeded++; |
e24f0b8f7 [PATCH] page migr... |
1091 1092 |
break; default: |
354a33633 mm/migrate: add c... |
1093 1094 1095 1096 1097 1098 |
/* * Permanent failure (-EBUSY, -ENOSYS, etc.): * unlike -EAGAIN case, the failed page is * removed from migration page list and not * retried in the next outer loop. */ |
2d1db3b11 [PATCH] page migr... |
1099 |
nr_failed++; |
e24f0b8f7 [PATCH] page migr... |
1100 |
break; |
2d1db3b11 [PATCH] page migr... |
1101 |
} |
b20a35035 [PATCH] page migr... |
1102 1103 |
} } |
78bd52097 mm: adjust addres... |
1104 |
rc = nr_failed + retry; |
95a402c38 [PATCH] page migr... |
1105 |
out: |
5647bc293 mm: compaction: M... |
1106 1107 1108 1109 |
if (nr_succeeded) count_vm_events(PGMIGRATE_SUCCESS, nr_succeeded); if (nr_failed) count_vm_events(PGMIGRATE_FAIL, nr_failed); |
7b2a2d4a1 mm: migrate: Add ... |
1110 |
trace_mm_migrate_pages(nr_succeeded, nr_failed, mode, reason); |
b20a35035 [PATCH] page migr... |
1111 1112 |
if (!swapwrite) current->flags &= ~PF_SWAPWRITE; |
78bd52097 mm: adjust addres... |
1113 |
return rc; |
b20a35035 [PATCH] page migr... |
1114 |
} |
95a402c38 [PATCH] page migr... |
1115 |
|
742755a1d [PATCH] page migr... |
1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 |
#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; |
e632a938d mm: migrate: add ... |
1139 1140 1141 1142 1143 |
if (PageHuge(p)) return alloc_huge_page_node(page_hstate(compound_head(p)), pm->node); else return alloc_pages_exact_node(pm->node, |
e97ca8e5b mm: fix GFP_THISN... |
1144 |
GFP_HIGHUSER_MOVABLE | __GFP_THISNODE, 0); |
742755a1d [PATCH] page migr... |
1145 1146 1147 1148 1149 1150 |
} /* * 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... |
1151 |
* The pm array ends with node = MAX_NUMNODES. |
742755a1d [PATCH] page migr... |
1152 |
*/ |
5e9a0f023 mm: extract do_pa... |
1153 1154 1155 |
static int do_move_page_to_node_array(struct mm_struct *mm, struct page_to_node *pm, int migrate_all) |
742755a1d [PATCH] page migr... |
1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 |
{ 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... |
1166 1167 1168 |
for (pp = pm; pp->node != MAX_NUMNODES; pp++) { struct vm_area_struct *vma; struct page *page; |
742755a1d [PATCH] page migr... |
1169 1170 |
err = -EFAULT; vma = find_vma(mm, pp->addr); |
70384dc6d mm: fix error rep... |
1171 |
if (!vma || pp->addr < vma->vm_start || !vma_migratable(vma)) |
742755a1d [PATCH] page migr... |
1172 |
goto set_status; |
500d65d47 thp: pmd_trans_hu... |
1173 |
page = follow_page(vma, pp->addr, FOLL_GET|FOLL_SPLIT); |
89f5b7da2 Reinstate ZERO_PA... |
1174 1175 1176 1177 |
err = PTR_ERR(page); if (IS_ERR(page)) goto set_status; |
742755a1d [PATCH] page migr... |
1178 1179 1180 |
err = -ENOENT; if (!page) goto set_status; |
62b61f611 ksm: memory hotre... |
1181 |
/* Use PageReserved to check for zero page */ |
b79bc0a0c ksm: enable KSM p... |
1182 |
if (PageReserved(page)) |
742755a1d [PATCH] page migr... |
1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 |
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; |
e632a938d mm: migrate: add ... |
1198 |
if (PageHuge(page)) { |
e66f17ff7 mm/hugetlb: take ... |
1199 1200 |
if (PageHead(page)) isolate_huge_page(page, &pagelist); |
e632a938d mm: migrate: add ... |
1201 1202 |
goto put_and_set; } |
62695a84e vmscan: move isol... |
1203 |
err = isolate_lru_page(page); |
6d9c285a6 mm: move inc_zone... |
1204 |
if (!err) { |
62695a84e vmscan: move isol... |
1205 |
list_add_tail(&page->lru, &pagelist); |
6d9c285a6 mm: move inc_zone... |
1206 1207 1208 |
inc_zone_page_state(page, NR_ISOLATED_ANON + page_is_file_cache(page)); } |
742755a1d [PATCH] page migr... |
1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 |
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... |
1219 |
err = 0; |
cf608ac19 mm: compaction: f... |
1220 |
if (!list_empty(&pagelist)) { |
68711a746 mm, migration: ad... |
1221 |
err = migrate_pages(&pagelist, new_page_node, NULL, |
9c620e2bc mm: remove offlin... |
1222 |
(unsigned long)pm, MIGRATE_SYNC, MR_SYSCALL); |
cf608ac19 mm: compaction: f... |
1223 |
if (err) |
e632a938d mm: migrate: add ... |
1224 |
putback_movable_pages(&pagelist); |
cf608ac19 mm: compaction: f... |
1225 |
} |
742755a1d [PATCH] page migr... |
1226 1227 1228 1229 1230 1231 |
up_read(&mm->mmap_sem); return err; } /* |
5e9a0f023 mm: extract do_pa... |
1232 1233 1234 |
* Migrate an array of page address onto an array of nodes and fill * the corresponding array of status. */ |
3268c63ed mm: fix move/migr... |
1235 |
static int do_pages_move(struct mm_struct *mm, nodemask_t task_nodes, |
5e9a0f023 mm: extract do_pa... |
1236 1237 1238 1239 1240 |
unsigned long nr_pages, const void __user * __user *pages, const int __user *nodes, int __user *status, int flags) { |
3140a2273 mm: rework do_pag... |
1241 |
struct page_to_node *pm; |
3140a2273 mm: rework do_pag... |
1242 1243 1244 |
unsigned long chunk_nr_pages; unsigned long chunk_start; int err; |
5e9a0f023 mm: extract do_pa... |
1245 |
|
3140a2273 mm: rework do_pag... |
1246 1247 1248 |
err = -ENOMEM; pm = (struct page_to_node *)__get_free_page(GFP_KERNEL); if (!pm) |
5e9a0f023 mm: extract do_pa... |
1249 |
goto out; |
35282a2de migration: only m... |
1250 1251 |
migrate_prep(); |
5e9a0f023 mm: extract do_pa... |
1252 |
/* |
3140a2273 mm: rework do_pag... |
1253 1254 |
* Store a chunk of page_to_node array in a page, * but keep the last one as a marker |
5e9a0f023 mm: extract do_pa... |
1255 |
*/ |
3140a2273 mm: rework do_pag... |
1256 |
chunk_nr_pages = (PAGE_SIZE / sizeof(struct page_to_node)) - 1; |
5e9a0f023 mm: extract do_pa... |
1257 |
|
3140a2273 mm: rework do_pag... |
1258 1259 1260 1261 |
for (chunk_start = 0; chunk_start < nr_pages; chunk_start += chunk_nr_pages) { int j; |
5e9a0f023 mm: extract do_pa... |
1262 |
|
3140a2273 mm: rework do_pag... |
1263 1264 1265 1266 1267 1268 |
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... |
1269 |
int node; |
3140a2273 mm: rework do_pag... |
1270 1271 1272 1273 1274 1275 |
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... |
1276 1277 1278 |
goto out_pm; err = -ENODEV; |
6f5a55f1a Fix potential cra... |
1279 1280 |
if (node < 0 || node >= MAX_NUMNODES) goto out_pm; |
389162c22 mm,migrate: use N... |
1281 |
if (!node_state(node, N_MEMORY)) |
5e9a0f023 mm: extract do_pa... |
1282 1283 1284 1285 1286 |
goto out_pm; err = -EACCES; if (!node_isset(node, task_nodes)) goto out_pm; |
3140a2273 mm: rework do_pag... |
1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 |
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... |
1298 |
|
5e9a0f023 mm: extract do_pa... |
1299 |
/* Return status information */ |
3140a2273 mm: rework do_pag... |
1300 1301 |
for (j = 0; j < chunk_nr_pages; j++) if (put_user(pm[j].status, status + j + chunk_start)) { |
5e9a0f023 mm: extract do_pa... |
1302 |
err = -EFAULT; |
3140a2273 mm: rework do_pag... |
1303 1304 1305 1306 |
goto out_pm; } } err = 0; |
5e9a0f023 mm: extract do_pa... |
1307 1308 |
out_pm: |
3140a2273 mm: rework do_pag... |
1309 |
free_page((unsigned long)pm); |
5e9a0f023 mm: extract do_pa... |
1310 1311 1312 1313 1314 |
out: return err; } /* |
2f007e74b mm: don't vmalloc... |
1315 |
* Determine the nodes of an array of pages and store it in an array of status. |
742755a1d [PATCH] page migr... |
1316 |
*/ |
80bba1290 mm: no get_user/p... |
1317 1318 |
static void do_pages_stat_array(struct mm_struct *mm, unsigned long nr_pages, const void __user **pages, int *status) |
742755a1d [PATCH] page migr... |
1319 |
{ |
2f007e74b mm: don't vmalloc... |
1320 |
unsigned long i; |
2f007e74b mm: don't vmalloc... |
1321 |
|
742755a1d [PATCH] page migr... |
1322 |
down_read(&mm->mmap_sem); |
2f007e74b mm: don't vmalloc... |
1323 |
for (i = 0; i < nr_pages; i++) { |
80bba1290 mm: no get_user/p... |
1324 |
unsigned long addr = (unsigned long)(*pages); |
742755a1d [PATCH] page migr... |
1325 1326 |
struct vm_area_struct *vma; struct page *page; |
c095adbc2 mm: Don't touch u... |
1327 |
int err = -EFAULT; |
2f007e74b mm: don't vmalloc... |
1328 1329 |
vma = find_vma(mm, addr); |
70384dc6d mm: fix error rep... |
1330 |
if (!vma || addr < vma->vm_start) |
742755a1d [PATCH] page migr... |
1331 |
goto set_status; |
2f007e74b mm: don't vmalloc... |
1332 |
page = follow_page(vma, addr, 0); |
89f5b7da2 Reinstate ZERO_PA... |
1333 1334 1335 1336 |
err = PTR_ERR(page); if (IS_ERR(page)) goto set_status; |
742755a1d [PATCH] page migr... |
1337 1338 |
err = -ENOENT; /* Use PageReserved to check for zero page */ |
b79bc0a0c ksm: enable KSM p... |
1339 |
if (!page || PageReserved(page)) |
742755a1d [PATCH] page migr... |
1340 1341 1342 1343 |
goto set_status; err = page_to_nid(page); set_status: |
80bba1290 mm: no get_user/p... |
1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 |
*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... |
1364 |
|
87b8d1ade mm: Make copy_fro... |
1365 1366 |
while (nr_pages) { unsigned long chunk_nr; |
80bba1290 mm: no get_user/p... |
1367 |
|
87b8d1ade mm: Make copy_fro... |
1368 1369 1370 1371 1372 1373 |
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... |
1374 1375 |
do_pages_stat_array(mm, chunk_nr, chunk_pages, chunk_status); |
87b8d1ade mm: Make copy_fro... |
1376 1377 |
if (copy_to_user(status, chunk_status, chunk_nr * sizeof(*status))) break; |
742755a1d [PATCH] page migr... |
1378 |
|
87b8d1ade mm: Make copy_fro... |
1379 1380 1381 1382 1383 |
pages += chunk_nr; status += chunk_nr; nr_pages -= chunk_nr; } return nr_pages ? -EFAULT : 0; |
742755a1d [PATCH] page migr... |
1384 1385 1386 1387 1388 1389 |
} /* * Move a list of pages in the address space of the currently executing * process. */ |
938bb9f5e [CVE-2009-0029] S... |
1390 1391 1392 1393 |
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... |
1394 |
{ |
c69e8d9c0 CRED: Use RCU to ... |
1395 |
const struct cred *cred = current_cred(), *tcred; |
742755a1d [PATCH] page migr... |
1396 |
struct task_struct *task; |
742755a1d [PATCH] page migr... |
1397 |
struct mm_struct *mm; |
5e9a0f023 mm: extract do_pa... |
1398 |
int err; |
3268c63ed mm: fix move/migr... |
1399 |
nodemask_t task_nodes; |
742755a1d [PATCH] page migr... |
1400 1401 1402 1403 1404 1405 1406 1407 1408 |
/* 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... |
1409 |
rcu_read_lock(); |
228ebcbe6 Uninline find_tas... |
1410 |
task = pid ? find_task_by_vpid(pid) : current; |
742755a1d [PATCH] page migr... |
1411 |
if (!task) { |
a879bf582 mm: grab rcu read... |
1412 |
rcu_read_unlock(); |
742755a1d [PATCH] page migr... |
1413 1414 |
return -ESRCH; } |
3268c63ed mm: fix move/migr... |
1415 |
get_task_struct(task); |
742755a1d [PATCH] page migr... |
1416 1417 1418 1419 1420 1421 1422 |
/* * 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 ... |
1423 |
tcred = __task_cred(task); |
b38a86eb1 userns: Convert t... |
1424 1425 |
if (!uid_eq(cred->euid, tcred->suid) && !uid_eq(cred->euid, tcred->uid) && !uid_eq(cred->uid, tcred->suid) && !uid_eq(cred->uid, tcred->uid) && |
742755a1d [PATCH] page migr... |
1426 |
!capable(CAP_SYS_NICE)) { |
c69e8d9c0 CRED: Use RCU to ... |
1427 |
rcu_read_unlock(); |
742755a1d [PATCH] page migr... |
1428 |
err = -EPERM; |
5e9a0f023 mm: extract do_pa... |
1429 |
goto out; |
742755a1d [PATCH] page migr... |
1430 |
} |
c69e8d9c0 CRED: Use RCU to ... |
1431 |
rcu_read_unlock(); |
742755a1d [PATCH] page migr... |
1432 |
|
86c3a7645 [PATCH] SELinux: ... |
1433 1434 |
err = security_task_movememory(task); if (err) |
5e9a0f023 mm: extract do_pa... |
1435 |
goto out; |
86c3a7645 [PATCH] SELinux: ... |
1436 |
|
3268c63ed mm: fix move/migr... |
1437 1438 1439 |
task_nodes = cpuset_mems_allowed(task); mm = get_task_mm(task); put_task_struct(task); |
6e8b09eaf mm: fix NULL ptr ... |
1440 1441 1442 1443 1444 1445 1446 1447 |
if (!mm) return -EINVAL; if (nodes) err = do_pages_move(mm, task_nodes, nr_pages, pages, nodes, status, flags); else err = do_pages_stat(mm, nr_pages, pages, status); |
742755a1d [PATCH] page migr... |
1448 |
|
742755a1d [PATCH] page migr... |
1449 1450 |
mmput(mm); return err; |
3268c63ed mm: fix move/migr... |
1451 1452 1453 1454 |
out: put_task_struct(task); return err; |
742755a1d [PATCH] page migr... |
1455 |
} |
742755a1d [PATCH] page migr... |
1456 |
|
7039e1dbe mm: migrate: Intr... |
1457 1458 1459 1460 1461 1462 |
#ifdef CONFIG_NUMA_BALANCING /* * Returns true if this is a safe migration target node for misplaced NUMA * pages. Currently it only checks the watermarks which crude */ static bool migrate_balanced_pgdat(struct pglist_data *pgdat, |
3abef4e6c mm: numa: take TH... |
1463 |
unsigned long nr_migrate_pages) |
7039e1dbe mm: migrate: Intr... |
1464 1465 1466 1467 1468 1469 1470 |
{ int z; for (z = pgdat->nr_zones - 1; z >= 0; z--) { struct zone *zone = pgdat->node_zones + z; if (!populated_zone(zone)) continue; |
6e543d578 mm: vmscan: fix d... |
1471 |
if (!zone_reclaimable(zone)) |
7039e1dbe mm: migrate: Intr... |
1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 |
continue; /* Avoid waking kswapd by allocating pages_to_migrate pages. */ if (!zone_watermark_ok(zone, 0, high_wmark_pages(zone) + nr_migrate_pages, 0, 0)) continue; return true; } return false; } static struct page *alloc_misplaced_dst_page(struct page *page, unsigned long data, int **result) { int nid = (int) data; struct page *newpage; newpage = alloc_pages_exact_node(nid, |
e97ca8e5b mm: fix GFP_THISN... |
1493 1494 1495 |
(GFP_HIGHUSER_MOVABLE | __GFP_THISNODE | __GFP_NOMEMALLOC | __GFP_NORETRY | __GFP_NOWARN) & |
7039e1dbe mm: migrate: Intr... |
1496 |
~GFP_IOFS, 0); |
bac0382c6 mm: numa: migrate... |
1497 |
|
7039e1dbe mm: migrate: Intr... |
1498 1499 1500 1501 |
return newpage; } /* |
a8f607721 mm: numa: Rate li... |
1502 1503 1504 1505 1506 1507 |
* page migration rate limiting control. * Do not migrate more than @pages_to_migrate in a @migrate_interval_millisecs * window of time. Default here says do not migrate more than 1280M per second. */ static unsigned int migrate_interval_millisecs __read_mostly = 100; static unsigned int ratelimit_pages __read_mostly = 128 << (20 - PAGE_SHIFT); |
b32967ff1 mm: numa: Add THP... |
1508 |
/* Returns true if the node is migrate rate-limited after the update */ |
1c30e0177 mm: numa: make NU... |
1509 1510 |
static bool numamigrate_update_ratelimit(pg_data_t *pgdat, unsigned long nr_pages) |
7039e1dbe mm: migrate: Intr... |
1511 |
{ |
a8f607721 mm: numa: Rate li... |
1512 1513 1514 1515 1516 |
/* * Rate-limit the amount of data that is being migrated to a node. * Optimal placement is no good if the memory bus is saturated and * all the time is being spent migrating! */ |
a8f607721 mm: numa: Rate li... |
1517 |
if (time_after(jiffies, pgdat->numabalancing_migrate_next_window)) { |
1c5e9c27c mm: numa: limit s... |
1518 |
spin_lock(&pgdat->numabalancing_migrate_lock); |
a8f607721 mm: numa: Rate li... |
1519 1520 1521 |
pgdat->numabalancing_migrate_nr_pages = 0; pgdat->numabalancing_migrate_next_window = jiffies + msecs_to_jiffies(migrate_interval_millisecs); |
1c5e9c27c mm: numa: limit s... |
1522 |
spin_unlock(&pgdat->numabalancing_migrate_lock); |
a8f607721 mm: numa: Rate li... |
1523 |
} |
af1839d72 mm: numa: trace t... |
1524 1525 1526 |
if (pgdat->numabalancing_migrate_nr_pages > ratelimit_pages) { trace_mm_numa_migrate_ratelimit(current, pgdat->node_id, nr_pages); |
1c5e9c27c mm: numa: limit s... |
1527 |
return true; |
af1839d72 mm: numa: trace t... |
1528 |
} |
1c5e9c27c mm: numa: limit s... |
1529 1530 1531 1532 1533 1534 1535 1536 1537 |
/* * This is an unlocked non-atomic update so errors are possible. * The consequences are failing to migrate when we potentiall should * have which is not severe enough to warrant locking. If it is ever * a problem, it can be converted to a per-cpu counter. */ pgdat->numabalancing_migrate_nr_pages += nr_pages; return false; |
b32967ff1 mm: numa: Add THP... |
1538 |
} |
1c30e0177 mm: numa: make NU... |
1539 |
static int numamigrate_isolate_page(pg_data_t *pgdat, struct page *page) |
b32967ff1 mm: numa: Add THP... |
1540 |
{ |
340ef3902 mm: numa: cleanup... |
1541 |
int page_lru; |
a8f607721 mm: numa: Rate li... |
1542 |
|
309381fea mm: dump page whe... |
1543 |
VM_BUG_ON_PAGE(compound_order(page) && !PageTransHuge(page), page); |
3abef4e6c mm: numa: take TH... |
1544 |
|
7039e1dbe mm: migrate: Intr... |
1545 |
/* Avoid migrating to a node that is nearly full */ |
340ef3902 mm: numa: cleanup... |
1546 1547 |
if (!migrate_balanced_pgdat(pgdat, 1UL << compound_order(page))) return 0; |
7039e1dbe mm: migrate: Intr... |
1548 |
|
340ef3902 mm: numa: cleanup... |
1549 1550 |
if (isolate_lru_page(page)) return 0; |
7039e1dbe mm: migrate: Intr... |
1551 |
|
340ef3902 mm: numa: cleanup... |
1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 |
/* * migrate_misplaced_transhuge_page() skips page migration's usual * check on page_count(), so we must do it here, now that the page * has been isolated: a GUP pin, or any other pin, prevents migration. * The expected page count is 3: 1 for page's mapcount and 1 for the * caller's pin and 1 for the reference taken by isolate_lru_page(). */ if (PageTransHuge(page) && page_count(page) != 3) { putback_lru_page(page); return 0; |
7039e1dbe mm: migrate: Intr... |
1562 |
} |
340ef3902 mm: numa: cleanup... |
1563 1564 1565 |
page_lru = page_is_file_cache(page); mod_zone_page_state(page_zone(page), NR_ISOLATED_ANON + page_lru, hpage_nr_pages(page)); |
149c33e1c mm: migrate: Drop... |
1566 |
/* |
340ef3902 mm: numa: cleanup... |
1567 1568 1569 |
* Isolating the page has taken another reference, so the * caller's reference can be safely dropped without the page * disappearing underneath us during migration. |
149c33e1c mm: migrate: Drop... |
1570 1571 |
*/ put_page(page); |
340ef3902 mm: numa: cleanup... |
1572 |
return 1; |
b32967ff1 mm: numa: Add THP... |
1573 |
} |
de466bd62 mm: numa: avoid u... |
1574 1575 1576 1577 1578 |
bool pmd_trans_migrating(pmd_t pmd) { struct page *page = pmd_page(pmd); return PageLocked(page); } |
b32967ff1 mm: numa: Add THP... |
1579 1580 1581 1582 1583 |
/* * Attempt to migrate a misplaced page to the specified destination * node. Caller is expected to have an elevated reference count on * the page that will be dropped by this function before returning. */ |
1bc115d87 mm: numa: Scan pa... |
1584 1585 |
int migrate_misplaced_page(struct page *page, struct vm_area_struct *vma, int node) |
b32967ff1 mm: numa: Add THP... |
1586 1587 |
{ pg_data_t *pgdat = NODE_DATA(node); |
340ef3902 mm: numa: cleanup... |
1588 |
int isolated; |
b32967ff1 mm: numa: Add THP... |
1589 1590 1591 1592 |
int nr_remaining; LIST_HEAD(migratepages); /* |
1bc115d87 mm: numa: Scan pa... |
1593 1594 |
* Don't migrate file pages that are mapped in multiple processes * with execute permissions as they are probably shared libraries. |
b32967ff1 mm: numa: Add THP... |
1595 |
*/ |
1bc115d87 mm: numa: Scan pa... |
1596 1597 |
if (page_mapcount(page) != 1 && page_is_file_cache(page) && (vma->vm_flags & VM_EXEC)) |
b32967ff1 mm: numa: Add THP... |
1598 |
goto out; |
b32967ff1 mm: numa: Add THP... |
1599 1600 1601 1602 1603 1604 |
/* * Rate-limit the amount of data that is being migrated to a node. * Optimal placement is no good if the memory bus is saturated and * all the time is being spent migrating! */ |
340ef3902 mm: numa: cleanup... |
1605 |
if (numamigrate_update_ratelimit(pgdat, 1)) |
b32967ff1 mm: numa: Add THP... |
1606 |
goto out; |
b32967ff1 mm: numa: Add THP... |
1607 1608 1609 1610 1611 1612 |
isolated = numamigrate_isolate_page(pgdat, page); if (!isolated) goto out; list_add(&page->lru, &migratepages); |
9c620e2bc mm: remove offlin... |
1613 |
nr_remaining = migrate_pages(&migratepages, alloc_misplaced_dst_page, |
68711a746 mm, migration: ad... |
1614 1615 |
NULL, node, MIGRATE_ASYNC, MR_NUMA_MISPLACED); |
b32967ff1 mm: numa: Add THP... |
1616 |
if (nr_remaining) { |
59c82b70d mm/migrate: remov... |
1617 1618 1619 1620 1621 1622 |
if (!list_empty(&migratepages)) { list_del(&page->lru); dec_zone_page_state(page, NR_ISOLATED_ANON + page_is_file_cache(page)); putback_lru_page(page); } |
b32967ff1 mm: numa: Add THP... |
1623 1624 1625 |
isolated = 0; } else count_vm_numa_event(NUMA_PAGE_MIGRATE); |
7039e1dbe mm: migrate: Intr... |
1626 |
BUG_ON(!list_empty(&migratepages)); |
7039e1dbe mm: migrate: Intr... |
1627 |
return isolated; |
340ef3902 mm: numa: cleanup... |
1628 1629 1630 1631 |
out: put_page(page); return 0; |
7039e1dbe mm: migrate: Intr... |
1632 |
} |
220018d38 mm: numa: Add THP... |
1633 |
#endif /* CONFIG_NUMA_BALANCING */ |
b32967ff1 mm: numa: Add THP... |
1634 |
|
220018d38 mm: numa: Add THP... |
1635 |
#if defined(CONFIG_NUMA_BALANCING) && defined(CONFIG_TRANSPARENT_HUGEPAGE) |
340ef3902 mm: numa: cleanup... |
1636 1637 1638 1639 |
/* * Migrates a THP to a given target node. page must be locked and is unlocked * before returning. */ |
b32967ff1 mm: numa: Add THP... |
1640 1641 1642 1643 1644 1645 |
int migrate_misplaced_transhuge_page(struct mm_struct *mm, struct vm_area_struct *vma, pmd_t *pmd, pmd_t entry, unsigned long address, struct page *page, int node) { |
c4088ebdc mm: convert the r... |
1646 |
spinlock_t *ptl; |
b32967ff1 mm: numa: Add THP... |
1647 1648 1649 |
pg_data_t *pgdat = NODE_DATA(node); int isolated = 0; struct page *new_page = NULL; |
b32967ff1 mm: numa: Add THP... |
1650 |
int page_lru = page_is_file_cache(page); |
f714f4f20 mm: numa: call MM... |
1651 1652 |
unsigned long mmun_start = address & HPAGE_PMD_MASK; unsigned long mmun_end = mmun_start + HPAGE_PMD_SIZE; |
2b4847e73 mm: numa: seriali... |
1653 |
pmd_t orig_entry; |
b32967ff1 mm: numa: Add THP... |
1654 1655 |
/* |
b32967ff1 mm: numa: Add THP... |
1656 1657 1658 1659 |
* Rate-limit the amount of data that is being migrated to a node. * Optimal placement is no good if the memory bus is saturated and * all the time is being spent migrating! */ |
d28d43351 mm: migrate: Acco... |
1660 |
if (numamigrate_update_ratelimit(pgdat, HPAGE_PMD_NR)) |
b32967ff1 mm: numa: Add THP... |
1661 1662 1663 |
goto out_dropref; new_page = alloc_pages_node(node, |
e97ca8e5b mm: fix GFP_THISN... |
1664 1665 |
(GFP_TRANSHUGE | __GFP_THISNODE) & ~__GFP_WAIT, HPAGE_PMD_ORDER); |
340ef3902 mm: numa: cleanup... |
1666 1667 |
if (!new_page) goto out_fail; |
b32967ff1 mm: numa: Add THP... |
1668 |
isolated = numamigrate_isolate_page(pgdat, page); |
340ef3902 mm: numa: cleanup... |
1669 |
if (!isolated) { |
b32967ff1 mm: numa: Add THP... |
1670 |
put_page(new_page); |
340ef3902 mm: numa: cleanup... |
1671 |
goto out_fail; |
b32967ff1 mm: numa: Add THP... |
1672 |
} |
b0943d61b mm: numa: defer T... |
1673 1674 |
if (mm_tlb_flush_pending(mm)) flush_tlb_range(vma, mmun_start, mmun_end); |
b32967ff1 mm: numa: Add THP... |
1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 |
/* Prepare a page as a migration target */ __set_page_locked(new_page); SetPageSwapBacked(new_page); /* anon mapping, we can simply copy page->mapping to the new page: */ new_page->mapping = page->mapping; new_page->index = page->index; migrate_page_copy(new_page, page); WARN_ON(PageLRU(new_page)); /* Recheck the target PMD */ |
f714f4f20 mm: numa: call MM... |
1686 |
mmu_notifier_invalidate_range_start(mm, mmun_start, mmun_end); |
c4088ebdc mm: convert the r... |
1687 |
ptl = pmd_lock(mm, pmd); |
2b4847e73 mm: numa: seriali... |
1688 1689 |
if (unlikely(!pmd_same(*pmd, entry) || page_count(page) != 2)) { fail_putback: |
c4088ebdc mm: convert the r... |
1690 |
spin_unlock(ptl); |
f714f4f20 mm: numa: call MM... |
1691 |
mmu_notifier_invalidate_range_end(mm, mmun_start, mmun_end); |
b32967ff1 mm: numa: Add THP... |
1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 |
/* Reverse changes made by migrate_page_copy() */ if (TestClearPageActive(new_page)) SetPageActive(page); if (TestClearPageUnevictable(new_page)) SetPageUnevictable(page); mlock_migrate_page(page, new_page); unlock_page(new_page); put_page(new_page); /* Free it */ |
a54a407fb mm: Close races b... |
1702 1703 |
/* Retake the callers reference and putback on LRU */ get_page(page); |
b32967ff1 mm: numa: Add THP... |
1704 |
putback_lru_page(page); |
a54a407fb mm: Close races b... |
1705 1706 |
mod_zone_page_state(page_zone(page), NR_ISOLATED_ANON + page_lru, -HPAGE_PMD_NR); |
eb4489f69 mm: numa: avoid u... |
1707 1708 |
goto out_unlock; |
b32967ff1 mm: numa: Add THP... |
1709 |
} |
2b4847e73 mm: numa: seriali... |
1710 |
orig_entry = *pmd; |
b32967ff1 mm: numa: Add THP... |
1711 |
entry = mk_pmd(new_page, vma->vm_page_prot); |
b32967ff1 mm: numa: Add THP... |
1712 |
entry = pmd_mkhuge(entry); |
2b4847e73 mm: numa: seriali... |
1713 |
entry = maybe_pmd_mkwrite(pmd_mkdirty(entry), vma); |
b32967ff1 mm: numa: Add THP... |
1714 |
|
2b4847e73 mm: numa: seriali... |
1715 1716 1717 1718 1719 1720 1721 |
/* * Clear the old entry under pagetable lock and establish the new PTE. * Any parallel GUP will either observe the old page blocking on the * page lock, block on the page table lock or observe the new page. * The SetPageUptodate on the new page and page_add_new_anon_rmap * guarantee the copy is visible before the pagetable update. */ |
f714f4f20 mm: numa: call MM... |
1722 |
flush_cache_range(vma, mmun_start, mmun_end); |
11de9927f mm: numa: add mig... |
1723 |
page_add_anon_rmap(new_page, vma, mmun_start); |
34ee645e8 mmu_notifier: cal... |
1724 |
pmdp_clear_flush_notify(vma, mmun_start, pmd); |
f714f4f20 mm: numa: call MM... |
1725 1726 |
set_pmd_at(mm, mmun_start, pmd, entry); flush_tlb_range(vma, mmun_start, mmun_end); |
ce4a9cc57 mm,numa: fix upda... |
1727 |
update_mmu_cache_pmd(vma, address, &entry); |
2b4847e73 mm: numa: seriali... |
1728 1729 |
if (page_count(page) != 2) { |
f714f4f20 mm: numa: call MM... |
1730 1731 |
set_pmd_at(mm, mmun_start, pmd, orig_entry); flush_tlb_range(vma, mmun_start, mmun_end); |
34ee645e8 mmu_notifier: cal... |
1732 |
mmu_notifier_invalidate_range(mm, mmun_start, mmun_end); |
2b4847e73 mm: numa: seriali... |
1733 1734 1735 1736 |
update_mmu_cache_pmd(vma, address, &entry); page_remove_rmap(new_page); goto fail_putback; } |
0a31bc97c mm: memcontrol: r... |
1737 |
mem_cgroup_migrate(page, new_page, false); |
b32967ff1 mm: numa: Add THP... |
1738 |
page_remove_rmap(page); |
2b4847e73 mm: numa: seriali... |
1739 |
|
c4088ebdc mm: convert the r... |
1740 |
spin_unlock(ptl); |
f714f4f20 mm: numa: call MM... |
1741 |
mmu_notifier_invalidate_range_end(mm, mmun_start, mmun_end); |
b32967ff1 mm: numa: Add THP... |
1742 |
|
11de9927f mm: numa: add mig... |
1743 1744 1745 |
/* Take an "isolate" reference and put new page on the LRU. */ get_page(new_page); putback_lru_page(new_page); |
b32967ff1 mm: numa: Add THP... |
1746 1747 1748 1749 1750 1751 1752 |
unlock_page(new_page); unlock_page(page); put_page(page); /* Drop the rmap reference */ put_page(page); /* Drop the LRU isolation reference */ count_vm_events(PGMIGRATE_SUCCESS, HPAGE_PMD_NR); count_vm_numa_events(NUMA_PAGE_MIGRATE, HPAGE_PMD_NR); |
b32967ff1 mm: numa: Add THP... |
1753 1754 1755 1756 |
mod_zone_page_state(page_zone(page), NR_ISOLATED_ANON + page_lru, -HPAGE_PMD_NR); return isolated; |
340ef3902 mm: numa: cleanup... |
1757 1758 |
out_fail: count_vm_events(PGMIGRATE_FAIL, HPAGE_PMD_NR); |
b32967ff1 mm: numa: Add THP... |
1759 |
out_dropref: |
2b4847e73 mm: numa: seriali... |
1760 1761 |
ptl = pmd_lock(mm, pmd); if (pmd_same(*pmd, entry)) { |
4d9424669 mm: convert p[te|... |
1762 |
entry = pmd_modify(entry, vma->vm_page_prot); |
f714f4f20 mm: numa: call MM... |
1763 |
set_pmd_at(mm, mmun_start, pmd, entry); |
2b4847e73 mm: numa: seriali... |
1764 1765 1766 |
update_mmu_cache_pmd(vma, address, &entry); } spin_unlock(ptl); |
a54a407fb mm: Close races b... |
1767 |
|
eb4489f69 mm: numa: avoid u... |
1768 |
out_unlock: |
340ef3902 mm: numa: cleanup... |
1769 |
unlock_page(page); |
b32967ff1 mm: numa: Add THP... |
1770 |
put_page(page); |
b32967ff1 mm: numa: Add THP... |
1771 1772 |
return 0; } |
7039e1dbe mm: migrate: Intr... |
1773 1774 1775 |
#endif /* CONFIG_NUMA_BALANCING */ #endif /* CONFIG_NUMA */ |