Blame view
mm/rmap.c
27.9 KB
1da177e4c
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 |
/* * mm/rmap.c - physical to virtual reverse mappings * * Copyright 2001, Rik van Riel <riel@conectiva.com.br> * Released under the General Public License (GPL). * * Simple, low overhead reverse mapping scheme. * Please try to keep this thing as modular as possible. * * Provides methods for unmapping each kind of mapped page: * the anon methods track anonymous pages, and * the file methods track pages belonging to an inode. * * Original design by Rik van Riel <riel@conectiva.com.br> 2001 * File methods by Dave McCracken <dmccr@us.ibm.com> 2003, 2004 * Anonymous methods by Andrea Arcangeli <andrea@suse.de> 2004 * Contributions by Hugh Dickins <hugh@veritas.com> 2003, 2004 */ /* * Lock ordering in mm: * |
1b1dcc1b5
|
23 |
* inode->i_mutex (while writing or truncating, not reading or faulting) |
82591e6ea
|
24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 |
* inode->i_alloc_sem (vmtruncate_range) * mm->mmap_sem * page->flags PG_locked (lock_page) * mapping->i_mmap_lock * anon_vma->lock * mm->page_table_lock or pte_lock * zone->lru_lock (in mark_page_accessed, isolate_lru_page) * swap_lock (in swap_duplicate, swap_info_get) * mmlist_lock (in mmput, drain_mmlist and others) * mapping->private_lock (in __set_page_dirty_buffers) * inode_lock (in set_page_dirty's __mark_inode_dirty) * sb_lock (within inode_lock in fs/fs-writeback.c) * mapping->tree_lock (widely used, in set_page_dirty, * in arch-dependent flush_dcache_mmap_lock, * within inode_lock in __sync_single_inode) |
1da177e4c
|
39 40 41 42 43 44 45 46 47 48 |
*/ #include <linux/mm.h> #include <linux/pagemap.h> #include <linux/swap.h> #include <linux/swapops.h> #include <linux/slab.h> #include <linux/init.h> #include <linux/rmap.h> #include <linux/rcupdate.h> |
a48d07afd
|
49 |
#include <linux/module.h> |
7de6b8057
|
50 |
#include <linux/kallsyms.h> |
8a9f3ccd2
|
51 |
#include <linux/memcontrol.h> |
cddb8a5c1
|
52 |
#include <linux/mmu_notifier.h> |
1da177e4c
|
53 54 |
#include <asm/tlbflush.h> |
fcc234f88
|
55 |
struct kmem_cache *anon_vma_cachep; |
1da177e4c
|
56 |
|
1da177e4c
|
57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 |
/* This must be called under the mmap_sem. */ int anon_vma_prepare(struct vm_area_struct *vma) { struct anon_vma *anon_vma = vma->anon_vma; might_sleep(); if (unlikely(!anon_vma)) { struct mm_struct *mm = vma->vm_mm; struct anon_vma *allocated, *locked; anon_vma = find_mergeable_anon_vma(vma); if (anon_vma) { allocated = NULL; locked = anon_vma; spin_lock(&locked->lock); } else { anon_vma = anon_vma_alloc(); if (unlikely(!anon_vma)) return -ENOMEM; allocated = anon_vma; locked = NULL; } /* page_table_lock to protect against threads */ spin_lock(&mm->page_table_lock); if (likely(!vma->anon_vma)) { vma->anon_vma = anon_vma; |
0697212a4
|
84 |
list_add_tail(&vma->anon_vma_node, &anon_vma->head); |
1da177e4c
|
85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 |
allocated = NULL; } spin_unlock(&mm->page_table_lock); if (locked) spin_unlock(&locked->lock); if (unlikely(allocated)) anon_vma_free(allocated); } return 0; } void __anon_vma_merge(struct vm_area_struct *vma, struct vm_area_struct *next) { BUG_ON(vma->anon_vma != next->anon_vma); list_del(&next->anon_vma_node); } void __anon_vma_link(struct vm_area_struct *vma) { struct anon_vma *anon_vma = vma->anon_vma; |
30acbabae
|
106 |
if (anon_vma) |
0697212a4
|
107 |
list_add_tail(&vma->anon_vma_node, &anon_vma->head); |
1da177e4c
|
108 109 110 111 112 113 114 115 |
} void anon_vma_link(struct vm_area_struct *vma) { struct anon_vma *anon_vma = vma->anon_vma; if (anon_vma) { spin_lock(&anon_vma->lock); |
0697212a4
|
116 |
list_add_tail(&vma->anon_vma_node, &anon_vma->head); |
1da177e4c
|
117 118 119 120 121 122 123 124 125 126 127 128 129 |
spin_unlock(&anon_vma->lock); } } void anon_vma_unlink(struct vm_area_struct *vma) { struct anon_vma *anon_vma = vma->anon_vma; int empty; if (!anon_vma) return; spin_lock(&anon_vma->lock); |
1da177e4c
|
130 131 132 133 134 135 136 137 138 |
list_del(&vma->anon_vma_node); /* We must garbage collect the anon_vma if it's empty */ empty = list_empty(&anon_vma->head); spin_unlock(&anon_vma->lock); if (empty) anon_vma_free(anon_vma); } |
51cc50685
|
139 |
static void anon_vma_ctor(void *data) |
1da177e4c
|
140 |
{ |
a35afb830
|
141 |
struct anon_vma *anon_vma = data; |
1da177e4c
|
142 |
|
a35afb830
|
143 144 |
spin_lock_init(&anon_vma->lock); INIT_LIST_HEAD(&anon_vma->head); |
1da177e4c
|
145 146 147 148 149 |
} void __init anon_vma_init(void) { anon_vma_cachep = kmem_cache_create("anon_vma", sizeof(struct anon_vma), |
20c2df83d
|
150 |
0, SLAB_DESTROY_BY_RCU|SLAB_PANIC, anon_vma_ctor); |
1da177e4c
|
151 152 153 154 155 156 157 158 |
} /* * Getting a lock on a stable anon_vma from a page off the LRU is * tricky: page_lock_anon_vma rely on RCU to guard against the races. */ static struct anon_vma *page_lock_anon_vma(struct page *page) { |
34bbd7040
|
159 |
struct anon_vma *anon_vma; |
1da177e4c
|
160 161 162 163 164 165 166 167 168 169 170 |
unsigned long anon_mapping; rcu_read_lock(); anon_mapping = (unsigned long) page->mapping; if (!(anon_mapping & PAGE_MAPPING_ANON)) goto out; if (!page_mapped(page)) goto out; anon_vma = (struct anon_vma *) (anon_mapping - PAGE_MAPPING_ANON); spin_lock(&anon_vma->lock); |
34bbd7040
|
171 |
return anon_vma; |
1da177e4c
|
172 173 |
out: rcu_read_unlock(); |
34bbd7040
|
174 175 176 177 178 179 180 |
return NULL; } static void page_unlock_anon_vma(struct anon_vma *anon_vma) { spin_unlock(&anon_vma->lock); rcu_read_unlock(); |
1da177e4c
|
181 182 183 |
} /* |
3ad33b243
|
184 185 186 |
* At what user virtual address is page expected in @vma? * Returns virtual address or -EFAULT if page's index/offset is not * within the range mapped the @vma. |
1da177e4c
|
187 188 189 190 191 192 193 194 195 |
*/ static inline unsigned long vma_address(struct page *page, struct vm_area_struct *vma) { pgoff_t pgoff = page->index << (PAGE_CACHE_SHIFT - PAGE_SHIFT); unsigned long address; address = vma->vm_start + ((pgoff - vma->vm_pgoff) << PAGE_SHIFT); if (unlikely(address < vma->vm_start || address >= vma->vm_end)) { |
3ad33b243
|
196 |
/* page should be within @vma mapping range */ |
1da177e4c
|
197 198 199 200 201 202 203 |
return -EFAULT; } return address; } /* * At what user virtual address is page expected in vma? checking that the |
ee498ed73
|
204 |
* page matches the vma: currently only used on anon pages, by unuse_vma; |
1da177e4c
|
205 206 207 208 209 210 211 212 |
*/ unsigned long page_address_in_vma(struct page *page, struct vm_area_struct *vma) { if (PageAnon(page)) { if ((void *)vma->anon_vma != (void *)page->mapping - PAGE_MAPPING_ANON) return -EFAULT; } else if (page->mapping && !(vma->vm_flags & VM_NONLINEAR)) { |
ee498ed73
|
213 214 |
if (!vma->vm_file || vma->vm_file->f_mapping != page->mapping) |
1da177e4c
|
215 216 217 218 219 220 221 |
return -EFAULT; } else return -EFAULT; return vma_address(page, vma); } /* |
81b4082dc
|
222 223 |
* Check that @page is mapped at @address into @mm. * |
479db0bf4
|
224 225 226 227 |
* If @sync is false, page_check_address may perform a racy check to avoid * the page table lock when the pte is not present (helpful when reclaiming * highly shared pages). * |
b8072f099
|
228 |
* On success returns with pte mapped and locked. |
81b4082dc
|
229 |
*/ |
ceffc0785
|
230 |
pte_t *page_check_address(struct page *page, struct mm_struct *mm, |
479db0bf4
|
231 |
unsigned long address, spinlock_t **ptlp, int sync) |
81b4082dc
|
232 233 234 235 236 |
{ pgd_t *pgd; pud_t *pud; pmd_t *pmd; pte_t *pte; |
c0718806c
|
237 |
spinlock_t *ptl; |
81b4082dc
|
238 |
|
81b4082dc
|
239 |
pgd = pgd_offset(mm, address); |
c0718806c
|
240 241 242 243 244 245 246 247 248 249 250 251 252 |
if (!pgd_present(*pgd)) return NULL; pud = pud_offset(pgd, address); if (!pud_present(*pud)) return NULL; pmd = pmd_offset(pud, address); if (!pmd_present(*pmd)) return NULL; pte = pte_offset_map(pmd, address); /* Make a quick check before getting the lock */ |
479db0bf4
|
253 |
if (!sync && !pte_present(*pte)) { |
c0718806c
|
254 255 256 |
pte_unmap(pte); return NULL; } |
4c21e2f24
|
257 |
ptl = pte_lockptr(mm, pmd); |
c0718806c
|
258 259 260 261 |
spin_lock(ptl); if (pte_present(*pte) && page_to_pfn(page) == pte_pfn(*pte)) { *ptlp = ptl; return pte; |
81b4082dc
|
262 |
} |
c0718806c
|
263 264 |
pte_unmap_unlock(pte, ptl); return NULL; |
81b4082dc
|
265 266 267 |
} /* |
1da177e4c
|
268 269 270 271 |
* Subfunctions of page_referenced: page_referenced_one called * repeatedly from either page_referenced_anon or page_referenced_file. */ static int page_referenced_one(struct page *page, |
f7b7fd8f3
|
272 |
struct vm_area_struct *vma, unsigned int *mapcount) |
1da177e4c
|
273 274 275 |
{ struct mm_struct *mm = vma->vm_mm; unsigned long address; |
1da177e4c
|
276 |
pte_t *pte; |
c0718806c
|
277 |
spinlock_t *ptl; |
1da177e4c
|
278 |
int referenced = 0; |
1da177e4c
|
279 280 281 |
address = vma_address(page, vma); if (address == -EFAULT) goto out; |
479db0bf4
|
282 |
pte = page_check_address(page, mm, address, &ptl, 0); |
c0718806c
|
283 284 |
if (!pte) goto out; |
1da177e4c
|
285 |
|
5a9bbdcd2
|
286 287 288 |
if (vma->vm_flags & VM_LOCKED) { referenced++; *mapcount = 1; /* break early from loop */ |
cddb8a5c1
|
289 |
} else if (ptep_clear_flush_young_notify(vma, address, pte)) |
c0718806c
|
290 |
referenced++; |
1da177e4c
|
291 |
|
c0718806c
|
292 293 |
/* Pretend the page is referenced if the task has the swap token and is in the middle of a page fault. */ |
f7b7fd8f3
|
294 |
if (mm != current->mm && has_swap_token(mm) && |
c0718806c
|
295 296 297 298 299 |
rwsem_is_locked(&mm->mmap_sem)) referenced++; (*mapcount)--; pte_unmap_unlock(pte, ptl); |
1da177e4c
|
300 301 302 |
out: return referenced; } |
bed7161a5
|
303 304 |
static int page_referenced_anon(struct page *page, struct mem_cgroup *mem_cont) |
1da177e4c
|
305 306 307 308 309 310 311 312 313 314 315 316 |
{ unsigned int mapcount; struct anon_vma *anon_vma; struct vm_area_struct *vma; int referenced = 0; anon_vma = page_lock_anon_vma(page); if (!anon_vma) return referenced; mapcount = page_mapcount(page); list_for_each_entry(vma, &anon_vma->head, anon_vma_node) { |
bed7161a5
|
317 318 319 320 321 |
/* * If we are reclaiming on behalf of a cgroup, skip * counting on behalf of references from different * cgroups */ |
bd845e38c
|
322 |
if (mem_cont && !mm_match_cgroup(vma->vm_mm, mem_cont)) |
bed7161a5
|
323 |
continue; |
f7b7fd8f3
|
324 |
referenced += page_referenced_one(page, vma, &mapcount); |
1da177e4c
|
325 326 327 |
if (!mapcount) break; } |
34bbd7040
|
328 329 |
page_unlock_anon_vma(anon_vma); |
1da177e4c
|
330 331 332 333 334 335 |
return referenced; } /** * page_referenced_file - referenced check for object-based rmap * @page: the page we're checking references on. |
43d8eac44
|
336 |
* @mem_cont: target memory controller |
1da177e4c
|
337 338 339 340 341 342 343 344 |
* * For an object-based mapped page, find all the places it is mapped and * check/clear the referenced flag. This is done by following the page->mapping * pointer, then walking the chain of vmas it holds. It returns the number * of references it found. * * This function is only called from page_referenced for object-based pages. */ |
bed7161a5
|
345 346 |
static int page_referenced_file(struct page *page, struct mem_cgroup *mem_cont) |
1da177e4c
|
347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 |
{ unsigned int mapcount; struct address_space *mapping = page->mapping; pgoff_t pgoff = page->index << (PAGE_CACHE_SHIFT - PAGE_SHIFT); struct vm_area_struct *vma; struct prio_tree_iter iter; int referenced = 0; /* * The caller's checks on page->mapping and !PageAnon have made * sure that this is a file page: the check for page->mapping * excludes the case just before it gets set on an anon page. */ BUG_ON(PageAnon(page)); /* * The page lock not only makes sure that page->mapping cannot * suddenly be NULLified by truncation, it makes sure that the * structure at mapping cannot be freed and reused yet, * so we can safely take mapping->i_mmap_lock. */ BUG_ON(!PageLocked(page)); spin_lock(&mapping->i_mmap_lock); /* * i_mmap_lock does not stabilize mapcount at all, but mapcount * is more likely to be accurate if we note it after spinning. */ mapcount = page_mapcount(page); vma_prio_tree_foreach(vma, &iter, &mapping->i_mmap, pgoff, pgoff) { |
bed7161a5
|
379 380 381 382 383 |
/* * If we are reclaiming on behalf of a cgroup, skip * counting on behalf of references from different * cgroups */ |
bd845e38c
|
384 |
if (mem_cont && !mm_match_cgroup(vma->vm_mm, mem_cont)) |
bed7161a5
|
385 |
continue; |
1da177e4c
|
386 387 388 389 390 |
if ((vma->vm_flags & (VM_LOCKED|VM_MAYSHARE)) == (VM_LOCKED|VM_MAYSHARE)) { referenced++; break; } |
f7b7fd8f3
|
391 |
referenced += page_referenced_one(page, vma, &mapcount); |
1da177e4c
|
392 393 394 395 396 397 398 399 400 401 402 403 |
if (!mapcount) break; } spin_unlock(&mapping->i_mmap_lock); return referenced; } /** * page_referenced - test if the page was referenced * @page: the page to test * @is_locked: caller holds lock on the page |
43d8eac44
|
404 |
* @mem_cont: target memory controller |
1da177e4c
|
405 406 407 408 |
* * Quick test_and_clear_referenced for all mappings to a page, * returns the number of ptes which referenced the page. */ |
bed7161a5
|
409 410 |
int page_referenced(struct page *page, int is_locked, struct mem_cgroup *mem_cont) |
1da177e4c
|
411 412 |
{ int referenced = 0; |
1da177e4c
|
413 414 415 416 417 |
if (TestClearPageReferenced(page)) referenced++; if (page_mapped(page) && page->mapping) { if (PageAnon(page)) |
bed7161a5
|
418 |
referenced += page_referenced_anon(page, mem_cont); |
1da177e4c
|
419 |
else if (is_locked) |
bed7161a5
|
420 |
referenced += page_referenced_file(page, mem_cont); |
529ae9aaa
|
421 |
else if (!trylock_page(page)) |
1da177e4c
|
422 423 424 |
referenced++; else { if (page->mapping) |
bed7161a5
|
425 426 |
referenced += page_referenced_file(page, mem_cont); |
1da177e4c
|
427 428 429 |
unlock_page(page); } } |
5b7baf057
|
430 431 432 |
if (page_test_and_clear_young(page)) referenced++; |
1da177e4c
|
433 434 |
return referenced; } |
d08b3851d
|
435 436 437 438 |
static int page_mkclean_one(struct page *page, struct vm_area_struct *vma) { struct mm_struct *mm = vma->vm_mm; unsigned long address; |
c2fda5fed
|
439 |
pte_t *pte; |
d08b3851d
|
440 441 442 443 444 445 |
spinlock_t *ptl; int ret = 0; address = vma_address(page, vma); if (address == -EFAULT) goto out; |
479db0bf4
|
446 |
pte = page_check_address(page, mm, address, &ptl, 1); |
d08b3851d
|
447 448 |
if (!pte) goto out; |
c2fda5fed
|
449 450 |
if (pte_dirty(*pte) || pte_write(*pte)) { pte_t entry; |
d08b3851d
|
451 |
|
c2fda5fed
|
452 |
flush_cache_page(vma, address, pte_pfn(*pte)); |
cddb8a5c1
|
453 |
entry = ptep_clear_flush_notify(vma, address, pte); |
c2fda5fed
|
454 455 |
entry = pte_wrprotect(entry); entry = pte_mkclean(entry); |
d6e88e671
|
456 |
set_pte_at(mm, address, pte, entry); |
c2fda5fed
|
457 458 |
ret = 1; } |
d08b3851d
|
459 |
|
d08b3851d
|
460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 |
pte_unmap_unlock(pte, ptl); out: return ret; } static int page_mkclean_file(struct address_space *mapping, struct page *page) { pgoff_t pgoff = page->index << (PAGE_CACHE_SHIFT - PAGE_SHIFT); struct vm_area_struct *vma; struct prio_tree_iter iter; int ret = 0; BUG_ON(PageAnon(page)); spin_lock(&mapping->i_mmap_lock); vma_prio_tree_foreach(vma, &iter, &mapping->i_mmap, pgoff, pgoff) { if (vma->vm_flags & VM_SHARED) ret += page_mkclean_one(page, vma); } spin_unlock(&mapping->i_mmap_lock); return ret; } int page_mkclean(struct page *page) { int ret = 0; BUG_ON(!PageLocked(page)); if (page_mapped(page)) { struct address_space *mapping = page_mapping(page); |
ce7e9fae8
|
491 |
if (mapping) { |
d08b3851d
|
492 |
ret = page_mkclean_file(mapping, page); |
ce7e9fae8
|
493 494 495 496 |
if (page_test_dirty(page)) { page_clear_dirty(page); ret = 1; } |
6c210482a
|
497 |
} |
d08b3851d
|
498 499 500 501 |
} return ret; } |
60b59beaf
|
502 |
EXPORT_SYMBOL_GPL(page_mkclean); |
d08b3851d
|
503 |
|
1da177e4c
|
504 |
/** |
43d8eac44
|
505 |
* __page_set_anon_rmap - setup new anonymous rmap |
9617d95e6
|
506 507 508 509 510 511 512 513 514 515 516 517 518 519 |
* @page: the page to add the mapping to * @vma: the vm area in which the mapping is added * @address: the user virtual address mapped */ static void __page_set_anon_rmap(struct page *page, struct vm_area_struct *vma, unsigned long address) { struct anon_vma *anon_vma = vma->anon_vma; BUG_ON(!anon_vma); anon_vma = (void *) anon_vma + PAGE_MAPPING_ANON; page->mapping = (struct address_space *) anon_vma; page->index = linear_page_index(vma, address); |
a74609faf
|
520 521 522 523 |
/* * nr_mapped state can be updated without turning off * interrupts because it is not modified via interrupt. */ |
f3dbd3446
|
524 |
__inc_zone_page_state(page, NR_ANON_PAGES); |
9617d95e6
|
525 526 527 |
} /** |
43d8eac44
|
528 |
* __page_check_anon_rmap - sanity check anonymous rmap addition |
c97a9e10e
|
529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 |
* @page: the page to add the mapping to * @vma: the vm area in which the mapping is added * @address: the user virtual address mapped */ static void __page_check_anon_rmap(struct page *page, struct vm_area_struct *vma, unsigned long address) { #ifdef CONFIG_DEBUG_VM /* * The page's anon-rmap details (mapping and index) are guaranteed to * be set up correctly at this point. * * We have exclusion against page_add_anon_rmap because the caller * always holds the page locked, except if called from page_dup_rmap, * in which case the page is already known to be setup. * * We have exclusion against page_add_new_anon_rmap because those pages * are initially only visible via the pagetables, and the pte is locked * over the call to page_add_new_anon_rmap. */ struct anon_vma *anon_vma = vma->anon_vma; anon_vma = (void *) anon_vma + PAGE_MAPPING_ANON; BUG_ON(page->mapping != (struct address_space *)anon_vma); BUG_ON(page->index != linear_page_index(vma, address)); #endif } /** |
1da177e4c
|
557 558 559 560 561 |
* page_add_anon_rmap - add pte mapping to an anonymous page * @page: the page to add the mapping to * @vma: the vm area in which the mapping is added * @address: the user virtual address mapped * |
c97a9e10e
|
562 |
* The caller needs to hold the pte lock and the page must be locked. |
1da177e4c
|
563 564 565 566 |
*/ void page_add_anon_rmap(struct page *page, struct vm_area_struct *vma, unsigned long address) { |
c97a9e10e
|
567 568 |
VM_BUG_ON(!PageLocked(page)); VM_BUG_ON(address < vma->vm_start || address >= vma->vm_end); |
9617d95e6
|
569 570 |
if (atomic_inc_and_test(&page->_mapcount)) __page_set_anon_rmap(page, vma, address); |
69029cd55
|
571 |
else |
c97a9e10e
|
572 |
__page_check_anon_rmap(page, vma, address); |
1da177e4c
|
573 |
} |
43d8eac44
|
574 |
/** |
9617d95e6
|
575 576 577 578 579 580 581 |
* page_add_new_anon_rmap - add pte mapping to a new anonymous page * @page: the page to add the mapping to * @vma: the vm area in which the mapping is added * @address: the user virtual address mapped * * Same as page_add_anon_rmap but must only be called on *new* pages. * This means the inc-and-test can be bypassed. |
c97a9e10e
|
582 |
* Page does not have to be locked. |
9617d95e6
|
583 584 585 586 |
*/ void page_add_new_anon_rmap(struct page *page, struct vm_area_struct *vma, unsigned long address) { |
c97a9e10e
|
587 |
BUG_ON(address < vma->vm_start || address >= vma->vm_end); |
9617d95e6
|
588 589 590 |
atomic_set(&page->_mapcount, 0); /* elevate count by 1 (starts at -1) */ __page_set_anon_rmap(page, vma, address); } |
1da177e4c
|
591 592 593 594 |
/** * page_add_file_rmap - add pte mapping to a file page * @page: the page to add the mapping to * |
b8072f099
|
595 |
* The caller needs to hold the pte lock. |
1da177e4c
|
596 597 598 |
*/ void page_add_file_rmap(struct page *page) { |
1da177e4c
|
599 |
if (atomic_inc_and_test(&page->_mapcount)) |
65ba55f50
|
600 |
__inc_zone_page_state(page, NR_FILE_MAPPED); |
1da177e4c
|
601 |
} |
c97a9e10e
|
602 603 604 605 |
#ifdef CONFIG_DEBUG_VM /** * page_dup_rmap - duplicate pte mapping to a page * @page: the page to add the mapping to |
43d8eac44
|
606 607 |
* @vma: the vm area being duplicated * @address: the user virtual address mapped |
c97a9e10e
|
608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 |
* * For copy_page_range only: minimal extract from page_add_file_rmap / * page_add_anon_rmap, avoiding unnecessary tests (already checked) so it's * quicker. * * The caller needs to hold the pte lock. */ void page_dup_rmap(struct page *page, struct vm_area_struct *vma, unsigned long address) { BUG_ON(page_mapcount(page) == 0); if (PageAnon(page)) __page_check_anon_rmap(page, vma, address); atomic_inc(&page->_mapcount); } #endif |
1da177e4c
|
623 624 625 |
/** * page_remove_rmap - take down pte mapping from a page * @page: page to remove mapping from |
43d8eac44
|
626 |
* @vma: the vm area in which the mapping is removed |
1da177e4c
|
627 |
* |
b8072f099
|
628 |
* The caller needs to hold the pte lock. |
1da177e4c
|
629 |
*/ |
7de6b8057
|
630 |
void page_remove_rmap(struct page *page, struct vm_area_struct *vma) |
1da177e4c
|
631 |
{ |
1da177e4c
|
632 |
if (atomic_add_negative(-1, &page->_mapcount)) { |
b7ab795b7
|
633 |
if (unlikely(page_mapcount(page) < 0)) { |
ef2bf0dc8
|
634 635 |
printk (KERN_EMERG "Eeek! page_mapcount(page) went negative! (%d) ", page_mapcount(page)); |
7de6b8057
|
636 637 |
printk (KERN_EMERG " page pfn = %lx ", page_to_pfn(page)); |
ef2bf0dc8
|
638 639 640 641 642 643 |
printk (KERN_EMERG " page->flags = %lx ", page->flags); printk (KERN_EMERG " page->count = %x ", page_count(page)); printk (KERN_EMERG " page->mapping = %p ", page->mapping); |
7de6b8057
|
644 645 |
print_symbol (KERN_EMERG " vma->vm_ops = %s ", (unsigned long)vma->vm_ops); |
54cb8821d
|
646 |
if (vma->vm_ops) { |
54cb8821d
|
647 648 649 |
print_symbol (KERN_EMERG " vma->vm_ops->fault = %s ", (unsigned long)vma->vm_ops->fault); } |
7de6b8057
|
650 651 652 |
if (vma->vm_file && vma->vm_file->f_op) print_symbol (KERN_EMERG " vma->vm_file->f_op->mmap = %s ", (unsigned long)vma->vm_file->f_op->mmap); |
b16bc64d1
|
653 |
BUG(); |
ef2bf0dc8
|
654 |
} |
b16bc64d1
|
655 |
|
1da177e4c
|
656 |
/* |
16f8c5b2e
|
657 658 659 660 661 |
* Now that the last pte has gone, s390 must transfer dirty * flag from storage key to struct page. We can usually skip * this if the page is anon, so about to be freed; but perhaps * not if it's in swapcache - there might be another pte slot * containing the swap entry, but page not yet written to swap. |
1da177e4c
|
662 |
*/ |
a4b526b3b
|
663 664 |
if ((!PageAnon(page) || PageSwapCache(page)) && page_test_dirty(page)) { |
6c210482a
|
665 |
page_clear_dirty(page); |
1da177e4c
|
666 |
set_page_dirty(page); |
6c210482a
|
667 |
} |
8a9f3ccd2
|
668 |
|
16f8c5b2e
|
669 |
mem_cgroup_uncharge_page(page); |
f3dbd3446
|
670 |
__dec_zone_page_state(page, |
16f8c5b2e
|
671 672 673 674 675 676 677 678 679 680 |
PageAnon(page) ? NR_ANON_PAGES : NR_FILE_MAPPED); /* * It would be tidy to reset the PageAnon mapping here, * but that might overwrite a racing page_add_anon_rmap * which increments mapcount after us but sets mapping * before us: so leave the reset to free_hot_cold_page, * and remember that it's only reliable while mapped. * Leaving it set also helps swapoff to reinstate ptes * faster for those pages still in swapcache. */ |
1da177e4c
|
681 682 683 684 685 686 687 |
} } /* * Subfunctions of try_to_unmap: try_to_unmap_one called * repeatedly from either try_to_unmap_anon or try_to_unmap_file. */ |
a48d07afd
|
688 |
static int try_to_unmap_one(struct page *page, struct vm_area_struct *vma, |
7352349a1
|
689 |
int migration) |
1da177e4c
|
690 691 692 |
{ struct mm_struct *mm = vma->vm_mm; unsigned long address; |
1da177e4c
|
693 694 |
pte_t *pte; pte_t pteval; |
c0718806c
|
695 |
spinlock_t *ptl; |
1da177e4c
|
696 |
int ret = SWAP_AGAIN; |
1da177e4c
|
697 698 699 |
address = vma_address(page, vma); if (address == -EFAULT) goto out; |
479db0bf4
|
700 |
pte = page_check_address(page, mm, address, &ptl, 0); |
c0718806c
|
701 |
if (!pte) |
81b4082dc
|
702 |
goto out; |
1da177e4c
|
703 704 705 706 707 708 |
/* * If the page is mlock()d, we cannot swap it out. * If it's recently referenced (perhaps page_referenced * skipped over this mm) then we should reactivate it. */ |
e6a1530d6
|
709 |
if (!migration && ((vma->vm_flags & VM_LOCKED) || |
cddb8a5c1
|
710 |
(ptep_clear_flush_young_notify(vma, address, pte)))) { |
1da177e4c
|
711 712 713 |
ret = SWAP_FAIL; goto out_unmap; } |
1da177e4c
|
714 715 |
/* Nuke the page table entry. */ flush_cache_page(vma, address, page_to_pfn(page)); |
cddb8a5c1
|
716 |
pteval = ptep_clear_flush_notify(vma, address, pte); |
1da177e4c
|
717 718 719 720 |
/* Move the dirty bit to the physical page now the pte is gone. */ if (pte_dirty(pteval)) set_page_dirty(page); |
365e9c87a
|
721 722 |
/* Update high watermark before we lower rss */ update_hiwater_rss(mm); |
1da177e4c
|
723 |
if (PageAnon(page)) { |
4c21e2f24
|
724 |
swp_entry_t entry = { .val = page_private(page) }; |
0697212a4
|
725 726 727 728 729 730 731 732 733 734 735 736 737 |
if (PageSwapCache(page)) { /* * Store the swap location in the pte. * See handle_pte_fault() ... */ swap_duplicate(entry); if (list_empty(&mm->mmlist)) { spin_lock(&mmlist_lock); if (list_empty(&mm->mmlist)) list_add(&mm->mmlist, &init_mm.mmlist); spin_unlock(&mmlist_lock); } |
442c9137d
|
738 |
dec_mm_counter(mm, anon_rss); |
04e62a29b
|
739 |
#ifdef CONFIG_MIGRATION |
0697212a4
|
740 741 742 743 744 745 746 747 |
} else { /* * Store the pfn of the page in a special migration * pte. do_swap_page() will wait until the migration * pte is removed and then restart fault handling. */ BUG_ON(!migration); entry = make_migration_entry(page, pte_write(pteval)); |
04e62a29b
|
748 |
#endif |
1da177e4c
|
749 750 751 |
} set_pte_at(mm, address, pte, swp_entry_to_pte(entry)); BUG_ON(pte_file(*pte)); |
4294621f4
|
752 |
} else |
04e62a29b
|
753 754 755 756 757 758 759 760 |
#ifdef CONFIG_MIGRATION if (migration) { /* Establish migration entry for a file page */ swp_entry_t entry; entry = make_migration_entry(page, pte_write(pteval)); set_pte_at(mm, address, pte, swp_entry_to_pte(entry)); } else #endif |
4294621f4
|
761 |
dec_mm_counter(mm, file_rss); |
1da177e4c
|
762 |
|
04e62a29b
|
763 |
|
7de6b8057
|
764 |
page_remove_rmap(page, vma); |
1da177e4c
|
765 766 767 |
page_cache_release(page); out_unmap: |
c0718806c
|
768 |
pte_unmap_unlock(pte, ptl); |
1da177e4c
|
769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 |
out: return ret; } /* * objrmap doesn't work for nonlinear VMAs because the assumption that * offset-into-file correlates with offset-into-virtual-addresses does not hold. * Consequently, given a particular page and its ->index, we cannot locate the * ptes which are mapping that page without an exhaustive linear search. * * So what this code does is a mini "virtual scan" of each nonlinear VMA which * maps the file to which the target page belongs. The ->vm_private_data field * holds the current cursor into that scan. Successive searches will circulate * around the vma's virtual address space. * * So as more replacement pressure is applied to the pages in a nonlinear VMA, * more scanning pressure is placed against them as well. Eventually pages * will become fully unmapped and are eligible for eviction. * * For very sparsely populated VMAs this is a little inefficient - chances are * there there won't be many ptes located within the scan cluster. In this case * maybe we could scan further - to the end of the pte page, perhaps. */ #define CLUSTER_SIZE min(32*PAGE_SIZE, PMD_SIZE) #define CLUSTER_MASK (~(CLUSTER_SIZE - 1)) static void try_to_unmap_cluster(unsigned long cursor, unsigned int *mapcount, struct vm_area_struct *vma) { struct mm_struct *mm = vma->vm_mm; pgd_t *pgd; pud_t *pud; pmd_t *pmd; |
c0718806c
|
802 |
pte_t *pte; |
1da177e4c
|
803 |
pte_t pteval; |
c0718806c
|
804 |
spinlock_t *ptl; |
1da177e4c
|
805 806 807 |
struct page *page; unsigned long address; unsigned long end; |
1da177e4c
|
808 |
|
1da177e4c
|
809 810 811 812 813 814 815 816 817 |
address = (vma->vm_start + cursor) & CLUSTER_MASK; end = address + CLUSTER_SIZE; if (address < vma->vm_start) address = vma->vm_start; if (end > vma->vm_end) end = vma->vm_end; pgd = pgd_offset(mm, address); if (!pgd_present(*pgd)) |
c0718806c
|
818 |
return; |
1da177e4c
|
819 820 821 |
pud = pud_offset(pgd, address); if (!pud_present(*pud)) |
c0718806c
|
822 |
return; |
1da177e4c
|
823 824 825 |
pmd = pmd_offset(pud, address); if (!pmd_present(*pmd)) |
c0718806c
|
826 827 828 |
return; pte = pte_offset_map_lock(mm, pmd, address, &ptl); |
1da177e4c
|
829 |
|
365e9c87a
|
830 831 |
/* Update high watermark before we lower rss */ update_hiwater_rss(mm); |
c0718806c
|
832 |
for (; address < end; pte++, address += PAGE_SIZE) { |
1da177e4c
|
833 834 |
if (!pte_present(*pte)) continue; |
6aab341e0
|
835 836 |
page = vm_normal_page(vma, address, *pte); BUG_ON(!page || PageAnon(page)); |
1da177e4c
|
837 |
|
cddb8a5c1
|
838 |
if (ptep_clear_flush_young_notify(vma, address, pte)) |
1da177e4c
|
839 840 841 |
continue; /* Nuke the page table entry. */ |
eca351336
|
842 |
flush_cache_page(vma, address, pte_pfn(*pte)); |
cddb8a5c1
|
843 |
pteval = ptep_clear_flush_notify(vma, address, pte); |
1da177e4c
|
844 845 846 847 848 849 850 851 |
/* If nonlinear, store the file page offset in the pte. */ if (page->index != linear_page_index(vma, address)) set_pte_at(mm, address, pte, pgoff_to_pte(page->index)); /* Move the dirty bit to the physical page now the pte is gone. */ if (pte_dirty(pteval)) set_page_dirty(page); |
7de6b8057
|
852 |
page_remove_rmap(page, vma); |
1da177e4c
|
853 |
page_cache_release(page); |
4294621f4
|
854 |
dec_mm_counter(mm, file_rss); |
1da177e4c
|
855 856 |
(*mapcount)--; } |
c0718806c
|
857 |
pte_unmap_unlock(pte - 1, ptl); |
1da177e4c
|
858 |
} |
7352349a1
|
859 |
static int try_to_unmap_anon(struct page *page, int migration) |
1da177e4c
|
860 861 862 863 864 865 866 867 868 869 |
{ struct anon_vma *anon_vma; struct vm_area_struct *vma; int ret = SWAP_AGAIN; anon_vma = page_lock_anon_vma(page); if (!anon_vma) return ret; list_for_each_entry(vma, &anon_vma->head, anon_vma_node) { |
7352349a1
|
870 |
ret = try_to_unmap_one(page, vma, migration); |
1da177e4c
|
871 872 873 |
if (ret == SWAP_FAIL || !page_mapped(page)) break; } |
34bbd7040
|
874 875 |
page_unlock_anon_vma(anon_vma); |
1da177e4c
|
876 877 878 879 880 881 |
return ret; } /** * try_to_unmap_file - unmap file page using the object-based rmap method * @page: the page to unmap |
43d8eac44
|
882 |
* @migration: migration flag |
1da177e4c
|
883 884 885 886 887 888 |
* * Find all the mappings of a page using the mapping pointer and the vma chains * contained in the address_space struct it points to. * * This function is only called from try_to_unmap for object-based pages. */ |
7352349a1
|
889 |
static int try_to_unmap_file(struct page *page, int migration) |
1da177e4c
|
890 891 892 893 894 895 896 897 898 899 900 901 902 |
{ struct address_space *mapping = page->mapping; pgoff_t pgoff = page->index << (PAGE_CACHE_SHIFT - PAGE_SHIFT); struct vm_area_struct *vma; struct prio_tree_iter iter; int ret = SWAP_AGAIN; unsigned long cursor; unsigned long max_nl_cursor = 0; unsigned long max_nl_size = 0; unsigned int mapcount; spin_lock(&mapping->i_mmap_lock); vma_prio_tree_foreach(vma, &iter, &mapping->i_mmap, pgoff, pgoff) { |
7352349a1
|
903 |
ret = try_to_unmap_one(page, vma, migration); |
1da177e4c
|
904 905 906 907 908 909 910 911 912 |
if (ret == SWAP_FAIL || !page_mapped(page)) goto out; } if (list_empty(&mapping->i_mmap_nonlinear)) goto out; list_for_each_entry(vma, &mapping->i_mmap_nonlinear, shared.vm_set.list) { |
e6a1530d6
|
913 |
if ((vma->vm_flags & VM_LOCKED) && !migration) |
1da177e4c
|
914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 |
continue; cursor = (unsigned long) vma->vm_private_data; if (cursor > max_nl_cursor) max_nl_cursor = cursor; cursor = vma->vm_end - vma->vm_start; if (cursor > max_nl_size) max_nl_size = cursor; } if (max_nl_size == 0) { /* any nonlinears locked or reserved */ ret = SWAP_FAIL; goto out; } /* * We don't try to search for this page in the nonlinear vmas, * and page_referenced wouldn't have found it anyway. Instead * just walk the nonlinear vmas trying to age and unmap some. * The mapcount of the page we came in with is irrelevant, * but even so use it as a guide to how hard we should try? */ mapcount = page_mapcount(page); if (!mapcount) goto out; cond_resched_lock(&mapping->i_mmap_lock); max_nl_size = (max_nl_size + CLUSTER_SIZE - 1) & CLUSTER_MASK; if (max_nl_cursor == 0) max_nl_cursor = CLUSTER_SIZE; do { list_for_each_entry(vma, &mapping->i_mmap_nonlinear, shared.vm_set.list) { |
e6a1530d6
|
947 |
if ((vma->vm_flags & VM_LOCKED) && !migration) |
1da177e4c
|
948 949 |
continue; cursor = (unsigned long) vma->vm_private_data; |
839b9685e
|
950 |
while ( cursor < max_nl_cursor && |
1da177e4c
|
951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 |
cursor < vma->vm_end - vma->vm_start) { try_to_unmap_cluster(cursor, &mapcount, vma); cursor += CLUSTER_SIZE; vma->vm_private_data = (void *) cursor; if ((int)mapcount <= 0) goto out; } vma->vm_private_data = (void *) max_nl_cursor; } cond_resched_lock(&mapping->i_mmap_lock); max_nl_cursor += CLUSTER_SIZE; } while (max_nl_cursor <= max_nl_size); /* * Don't loop forever (perhaps all the remaining pages are * in locked vmas). Reset cursor on all unreserved nonlinear * vmas, now forgetting on which ones it had fallen behind. */ |
101d2be76
|
969 970 |
list_for_each_entry(vma, &mapping->i_mmap_nonlinear, shared.vm_set.list) vma->vm_private_data = NULL; |
1da177e4c
|
971 972 973 974 975 976 977 978 |
out: spin_unlock(&mapping->i_mmap_lock); return ret; } /** * try_to_unmap - try to remove all page table mappings to a page * @page: the page to get unmapped |
43d8eac44
|
979 |
* @migration: migration flag |
1da177e4c
|
980 981 982 983 984 985 986 987 988 |
* * Tries to remove all the page table entries which are mapping this * page, used in the pageout path. Caller must hold the page lock. * Return values are: * * SWAP_SUCCESS - we succeeded in removing all mappings * SWAP_AGAIN - we missed a mapping, try again later * SWAP_FAIL - the page is unswappable */ |
7352349a1
|
989 |
int try_to_unmap(struct page *page, int migration) |
1da177e4c
|
990 991 |
{ int ret; |
1da177e4c
|
992 993 994 |
BUG_ON(!PageLocked(page)); if (PageAnon(page)) |
7352349a1
|
995 |
ret = try_to_unmap_anon(page, migration); |
1da177e4c
|
996 |
else |
7352349a1
|
997 |
ret = try_to_unmap_file(page, migration); |
1da177e4c
|
998 999 1000 1001 1002 |
if (!page_mapped(page)) ret = SWAP_SUCCESS; return ret; } |
81b4082dc
|
1003 |