Blame view
mm/rmap.c
52.2 KB
1da177e4c
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 |
/* * 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 |
98f32602d
|
17 |
* Contributions by Hugh Dickins 2003, 2004 |
1da177e4c
|
18 19 20 21 22 |
*/ /* * Lock ordering in mm: * |
1b1dcc1b5
|
23 |
* inode->i_mutex (while writing or truncating, not reading or faulting) |
82591e6ea
|
24 25 |
* mm->mmap_sem * page->flags PG_locked (lock_page) |
3d48ae45e
|
26 |
* mapping->i_mmap_mutex |
2b575eb64
|
27 |
* anon_vma->mutex |
82591e6ea
|
28 29 30 31 32 |
* 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) |
250df6ed2
|
33 |
* inode->i_lock (in set_page_dirty's __mark_inode_dirty) |
f758eeabe
|
34 |
* bdi.wb->list_lock (in set_page_dirty's __mark_inode_dirty) |
82591e6ea
|
35 36 37 |
* 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, |
f758eeabe
|
38 |
* within bdi.wb->list_lock in __sync_single_inode) |
6a46079cf
|
39 |
* |
9b679320a
|
40 41 |
* anon_vma->mutex,mapping->i_mutex (memory_failure, collect_procs_anon) * ->tasklist_lock |
6a46079cf
|
42 |
* pte map lock |
1da177e4c
|
43 44 45 46 47 48 49 50 |
*/ #include <linux/mm.h> #include <linux/pagemap.h> #include <linux/swap.h> #include <linux/swapops.h> #include <linux/slab.h> #include <linux/init.h> |
5ad646880
|
51 |
#include <linux/ksm.h> |
1da177e4c
|
52 53 |
#include <linux/rmap.h> #include <linux/rcupdate.h> |
b95f1b31b
|
54 |
#include <linux/export.h> |
8a9f3ccd2
|
55 |
#include <linux/memcontrol.h> |
cddb8a5c1
|
56 |
#include <linux/mmu_notifier.h> |
64cdd548f
|
57 |
#include <linux/migrate.h> |
0fe6e20b9
|
58 |
#include <linux/hugetlb.h> |
1da177e4c
|
59 60 |
#include <asm/tlbflush.h> |
b291f0003
|
61 |
#include "internal.h" |
fdd2e5f88
|
62 |
static struct kmem_cache *anon_vma_cachep; |
5beb49305
|
63 |
static struct kmem_cache *anon_vma_chain_cachep; |
fdd2e5f88
|
64 65 66 |
static inline struct anon_vma *anon_vma_alloc(void) { |
01d8b20de
|
67 68 69 70 71 72 73 74 75 76 77 78 79 |
struct anon_vma *anon_vma; anon_vma = kmem_cache_alloc(anon_vma_cachep, GFP_KERNEL); if (anon_vma) { atomic_set(&anon_vma->refcount, 1); /* * Initialise the anon_vma root to point to itself. If called * from fork, the root will be reset to the parents anon_vma. */ anon_vma->root = anon_vma; } return anon_vma; |
fdd2e5f88
|
80 |
} |
01d8b20de
|
81 |
static inline void anon_vma_free(struct anon_vma *anon_vma) |
fdd2e5f88
|
82 |
{ |
01d8b20de
|
83 |
VM_BUG_ON(atomic_read(&anon_vma->refcount)); |
88c22088b
|
84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 |
/* * Synchronize against page_lock_anon_vma() such that * we can safely hold the lock without the anon_vma getting * freed. * * Relies on the full mb implied by the atomic_dec_and_test() from * put_anon_vma() against the acquire barrier implied by * mutex_trylock() from page_lock_anon_vma(). This orders: * * page_lock_anon_vma() VS put_anon_vma() * mutex_trylock() atomic_dec_and_test() * LOCK MB * atomic_read() mutex_is_locked() * * LOCK should suffice since the actual taking of the lock must * happen _before_ what follows. */ if (mutex_is_locked(&anon_vma->root->mutex)) { anon_vma_lock(anon_vma); anon_vma_unlock(anon_vma); } |
fdd2e5f88
|
106 107 |
kmem_cache_free(anon_vma_cachep, anon_vma); } |
1da177e4c
|
108 |
|
dd34739c0
|
109 |
static inline struct anon_vma_chain *anon_vma_chain_alloc(gfp_t gfp) |
5beb49305
|
110 |
{ |
dd34739c0
|
111 |
return kmem_cache_alloc(anon_vma_chain_cachep, gfp); |
5beb49305
|
112 |
} |
e574b5fd2
|
113 |
static void anon_vma_chain_free(struct anon_vma_chain *anon_vma_chain) |
5beb49305
|
114 115 116 |
{ kmem_cache_free(anon_vma_chain_cachep, anon_vma_chain); } |
d9d332e08
|
117 118 119 120 121 122 123 124 125 |
/** * anon_vma_prepare - attach an anon_vma to a memory region * @vma: the memory region in question * * This makes sure the memory mapping described by 'vma' has * an 'anon_vma' attached to it, so that we can associate the * anonymous pages mapped into it with that anon_vma. * * The common case will be that we already have one, but if |
23a0790af
|
126 |
* not we either need to find an adjacent mapping that we |
d9d332e08
|
127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 |
* can re-use the anon_vma from (very common when the only * reason for splitting a vma has been mprotect()), or we * allocate a new one. * * Anon-vma allocations are very subtle, because we may have * optimistically looked up an anon_vma in page_lock_anon_vma() * and that may actually touch the spinlock even in the newly * allocated vma (it depends on RCU to make sure that the * anon_vma isn't actually destroyed). * * As a result, we need to do proper anon_vma locking even * for the new allocation. At the same time, we do not want * to do any locking for the common case of already having * an anon_vma. * * This must be called with the mmap_sem held for reading. */ |
1da177e4c
|
144 145 146 |
int anon_vma_prepare(struct vm_area_struct *vma) { struct anon_vma *anon_vma = vma->anon_vma; |
5beb49305
|
147 |
struct anon_vma_chain *avc; |
1da177e4c
|
148 149 150 151 |
might_sleep(); if (unlikely(!anon_vma)) { struct mm_struct *mm = vma->vm_mm; |
d9d332e08
|
152 |
struct anon_vma *allocated; |
1da177e4c
|
153 |
|
dd34739c0
|
154 |
avc = anon_vma_chain_alloc(GFP_KERNEL); |
5beb49305
|
155 156 |
if (!avc) goto out_enomem; |
1da177e4c
|
157 |
anon_vma = find_mergeable_anon_vma(vma); |
d9d332e08
|
158 159 |
allocated = NULL; if (!anon_vma) { |
1da177e4c
|
160 161 |
anon_vma = anon_vma_alloc(); if (unlikely(!anon_vma)) |
5beb49305
|
162 |
goto out_enomem_free_avc; |
1da177e4c
|
163 |
allocated = anon_vma; |
1da177e4c
|
164 |
} |
cba48b98f
|
165 |
anon_vma_lock(anon_vma); |
1da177e4c
|
166 167 168 169 |
/* page_table_lock to protect against threads */ spin_lock(&mm->page_table_lock); if (likely(!vma->anon_vma)) { vma->anon_vma = anon_vma; |
5beb49305
|
170 171 172 |
avc->anon_vma = anon_vma; avc->vma = vma; list_add(&avc->same_vma, &vma->anon_vma_chain); |
26ba0cb63
|
173 |
list_add_tail(&avc->same_anon_vma, &anon_vma->head); |
1da177e4c
|
174 |
allocated = NULL; |
31f2b0ebc
|
175 |
avc = NULL; |
1da177e4c
|
176 177 |
} spin_unlock(&mm->page_table_lock); |
cba48b98f
|
178 |
anon_vma_unlock(anon_vma); |
31f2b0ebc
|
179 180 |
if (unlikely(allocated)) |
01d8b20de
|
181 |
put_anon_vma(allocated); |
31f2b0ebc
|
182 |
if (unlikely(avc)) |
5beb49305
|
183 |
anon_vma_chain_free(avc); |
1da177e4c
|
184 185 |
} return 0; |
5beb49305
|
186 187 188 189 190 |
out_enomem_free_avc: anon_vma_chain_free(avc); out_enomem: return -ENOMEM; |
1da177e4c
|
191 |
} |
bb4aa3967
|
192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 |
/* * This is a useful helper function for locking the anon_vma root as * we traverse the vma->anon_vma_chain, looping over anon_vma's that * have the same vma. * * Such anon_vma's should have the same root, so you'd expect to see * just a single mutex_lock for the whole traversal. */ static inline struct anon_vma *lock_anon_vma_root(struct anon_vma *root, struct anon_vma *anon_vma) { struct anon_vma *new_root = anon_vma->root; if (new_root != root) { if (WARN_ON_ONCE(root)) mutex_unlock(&root->mutex); root = new_root; mutex_lock(&root->mutex); } return root; } static inline void unlock_anon_vma_root(struct anon_vma *root) { if (root) mutex_unlock(&root->mutex); } |
5beb49305
|
217 218 219 |
static void anon_vma_chain_link(struct vm_area_struct *vma, struct anon_vma_chain *avc, struct anon_vma *anon_vma) |
1da177e4c
|
220 |
{ |
5beb49305
|
221 222 223 |
avc->vma = vma; avc->anon_vma = anon_vma; list_add(&avc->same_vma, &vma->anon_vma_chain); |
05759d380
|
224 225 226 227 |
/* * It's critical to add new vmas to the tail of the anon_vma, * see comment in huge_memory.c:__split_huge_page(). */ |
5beb49305
|
228 |
list_add_tail(&avc->same_anon_vma, &anon_vma->head); |
1da177e4c
|
229 |
} |
5beb49305
|
230 231 232 233 234 |
/* * Attach the anon_vmas from src to dst. * Returns 0 on success, -ENOMEM on failure. */ int anon_vma_clone(struct vm_area_struct *dst, struct vm_area_struct *src) |
1da177e4c
|
235 |
{ |
5beb49305
|
236 |
struct anon_vma_chain *avc, *pavc; |
bb4aa3967
|
237 |
struct anon_vma *root = NULL; |
5beb49305
|
238 |
|
646d87b48
|
239 |
list_for_each_entry_reverse(pavc, &src->anon_vma_chain, same_vma) { |
bb4aa3967
|
240 |
struct anon_vma *anon_vma; |
dd34739c0
|
241 242 243 244 245 246 247 248 |
avc = anon_vma_chain_alloc(GFP_NOWAIT | __GFP_NOWARN); if (unlikely(!avc)) { unlock_anon_vma_root(root); root = NULL; avc = anon_vma_chain_alloc(GFP_KERNEL); if (!avc) goto enomem_failure; } |
bb4aa3967
|
249 250 251 |
anon_vma = pavc->anon_vma; root = lock_anon_vma_root(root, anon_vma); anon_vma_chain_link(dst, avc, anon_vma); |
5beb49305
|
252 |
} |
bb4aa3967
|
253 |
unlock_anon_vma_root(root); |
5beb49305
|
254 |
return 0; |
1da177e4c
|
255 |
|
5beb49305
|
256 257 258 |
enomem_failure: unlink_anon_vmas(dst); return -ENOMEM; |
1da177e4c
|
259 |
} |
5beb49305
|
260 |
/* |
948f017b0
|
261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 |
* Some rmap walk that needs to find all ptes/hugepmds without false * negatives (like migrate and split_huge_page) running concurrent * with operations that copy or move pagetables (like mremap() and * fork()) to be safe. They depend on the anon_vma "same_anon_vma" * list to be in a certain order: the dst_vma must be placed after the * src_vma in the list. This is always guaranteed by fork() but * mremap() needs to call this function to enforce it in case the * dst_vma isn't newly allocated and chained with the anon_vma_clone() * function but just an extension of a pre-existing vma through * vma_merge. * * NOTE: the same_anon_vma list can still be changed by other * processes while mremap runs because mremap doesn't hold the * anon_vma mutex to prevent modifications to the list while it * runs. All we need to enforce is that the relative order of this * process vmas isn't changing (we don't care about other vmas * order). Each vma corresponds to an anon_vma_chain structure so * there's no risk that other processes calling anon_vma_moveto_tail() * and changing the same_anon_vma list under mremap() will screw with * the relative order of this process vmas in the list, because we * they can't alter the order of any vma that belongs to this * process. And there can't be another anon_vma_moveto_tail() running * concurrently with mremap() coming from this process because we hold * the mmap_sem for the whole mremap(). fork() ordering dependency * also shouldn't be affected because fork() only cares that the * parent vmas are placed in the list before the child vmas and * anon_vma_moveto_tail() won't reorder vmas from either the fork() * parent or child. */ void anon_vma_moveto_tail(struct vm_area_struct *dst) { struct anon_vma_chain *pavc; struct anon_vma *root = NULL; list_for_each_entry_reverse(pavc, &dst->anon_vma_chain, same_vma) { struct anon_vma *anon_vma = pavc->anon_vma; VM_BUG_ON(pavc->vma != dst); root = lock_anon_vma_root(root, anon_vma); list_del(&pavc->same_anon_vma); list_add_tail(&pavc->same_anon_vma, &anon_vma->head); } unlock_anon_vma_root(root); } /* |
5beb49305
|
306 307 308 309 310 |
* Attach vma to its own anon_vma, as well as to the anon_vmas that * the corresponding VMA in the parent process is attached to. * Returns 0 on success, non-zero on failure. */ int anon_vma_fork(struct vm_area_struct *vma, struct vm_area_struct *pvma) |
1da177e4c
|
311 |
{ |
5beb49305
|
312 313 |
struct anon_vma_chain *avc; struct anon_vma *anon_vma; |
1da177e4c
|
314 |
|
5beb49305
|
315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 |
/* Don't bother if the parent process has no anon_vma here. */ if (!pvma->anon_vma) return 0; /* * First, attach the new VMA to the parent VMA's anon_vmas, * so rmap can find non-COWed pages in child processes. */ if (anon_vma_clone(vma, pvma)) return -ENOMEM; /* Then add our own anon_vma. */ anon_vma = anon_vma_alloc(); if (!anon_vma) goto out_error; |
dd34739c0
|
330 |
avc = anon_vma_chain_alloc(GFP_KERNEL); |
5beb49305
|
331 332 |
if (!avc) goto out_error_free_anon_vma; |
5c341ee1d
|
333 334 335 336 337 338 |
/* * The root anon_vma's spinlock is the lock actually used when we * lock any of the anon_vmas in this anon_vma tree. */ anon_vma->root = pvma->anon_vma->root; |
76545066c
|
339 |
/* |
01d8b20de
|
340 341 342 |
* With refcounts, an anon_vma can stay around longer than the * process it belongs to. The root anon_vma needs to be pinned until * this anon_vma is freed, because the lock lives in the root. |
76545066c
|
343 344 |
*/ get_anon_vma(anon_vma->root); |
5beb49305
|
345 346 |
/* Mark this anon_vma as the one where our new (COWed) pages go. */ vma->anon_vma = anon_vma; |
bb4aa3967
|
347 |
anon_vma_lock(anon_vma); |
5c341ee1d
|
348 |
anon_vma_chain_link(vma, avc, anon_vma); |
bb4aa3967
|
349 |
anon_vma_unlock(anon_vma); |
5beb49305
|
350 351 352 353 |
return 0; out_error_free_anon_vma: |
01d8b20de
|
354 |
put_anon_vma(anon_vma); |
5beb49305
|
355 |
out_error: |
4946d54cb
|
356 |
unlink_anon_vmas(vma); |
5beb49305
|
357 |
return -ENOMEM; |
1da177e4c
|
358 |
} |
5beb49305
|
359 360 361 |
void unlink_anon_vmas(struct vm_area_struct *vma) { struct anon_vma_chain *avc, *next; |
eee2acbae
|
362 |
struct anon_vma *root = NULL; |
5beb49305
|
363 |
|
5c341ee1d
|
364 365 366 367 |
/* * Unlink each anon_vma chained to the VMA. This list is ordered * from newest to oldest, ensuring the root anon_vma gets freed last. */ |
5beb49305
|
368 |
list_for_each_entry_safe(avc, next, &vma->anon_vma_chain, same_vma) { |
eee2acbae
|
369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 |
struct anon_vma *anon_vma = avc->anon_vma; root = lock_anon_vma_root(root, anon_vma); list_del(&avc->same_anon_vma); /* * Leave empty anon_vmas on the list - we'll need * to free them outside the lock. */ if (list_empty(&anon_vma->head)) continue; list_del(&avc->same_vma); anon_vma_chain_free(avc); } unlock_anon_vma_root(root); /* * Iterate the list once more, it now only contains empty and unlinked * anon_vmas, destroy them. Could not do before due to __put_anon_vma() * needing to acquire the anon_vma->root->mutex. */ list_for_each_entry_safe(avc, next, &vma->anon_vma_chain, same_vma) { struct anon_vma *anon_vma = avc->anon_vma; put_anon_vma(anon_vma); |
5beb49305
|
395 396 397 398 |
list_del(&avc->same_vma); anon_vma_chain_free(avc); } } |
51cc50685
|
399 |
static void anon_vma_ctor(void *data) |
1da177e4c
|
400 |
{ |
a35afb830
|
401 |
struct anon_vma *anon_vma = data; |
1da177e4c
|
402 |
|
2b575eb64
|
403 |
mutex_init(&anon_vma->mutex); |
83813267c
|
404 |
atomic_set(&anon_vma->refcount, 0); |
a35afb830
|
405 |
INIT_LIST_HEAD(&anon_vma->head); |
1da177e4c
|
406 407 408 409 410 |
} void __init anon_vma_init(void) { anon_vma_cachep = kmem_cache_create("anon_vma", sizeof(struct anon_vma), |
20c2df83d
|
411 |
0, SLAB_DESTROY_BY_RCU|SLAB_PANIC, anon_vma_ctor); |
5beb49305
|
412 |
anon_vma_chain_cachep = KMEM_CACHE(anon_vma_chain, SLAB_PANIC); |
1da177e4c
|
413 414 415 |
} /* |
6111e4ca6
|
416 417 418 419 420 421 422 423 424 |
* Getting a lock on a stable anon_vma from a page off the LRU is tricky! * * Since there is no serialization what so ever against page_remove_rmap() * the best this function can do is return a locked anon_vma that might * have been relevant to this page. * * The page might have been remapped to a different anon_vma or the anon_vma * returned may already be freed (and even reused). * |
bc658c960
|
425 426 427 428 429 |
* In case it was remapped to a different anon_vma, the new anon_vma will be a * child of the old anon_vma, and the anon_vma lifetime rules will therefore * ensure that any anon_vma obtained from the page will still be valid for as * long as we observe page_mapped() [ hence all those page_mapped() tests ]. * |
6111e4ca6
|
430 431 432 433 434 435 436 |
* All users of this function must be very careful when walking the anon_vma * chain and verify that the page in question is indeed mapped in it * [ something equivalent to page_mapped_in_vma() ]. * * Since anon_vma's slab is DESTROY_BY_RCU and we know from page_remove_rmap() * that the anon_vma pointer from page->mapping is valid if there is a * mapcount, we can dereference the anon_vma after observing those. |
1da177e4c
|
437 |
*/ |
746b18d42
|
438 |
struct anon_vma *page_get_anon_vma(struct page *page) |
1da177e4c
|
439 |
{ |
746b18d42
|
440 |
struct anon_vma *anon_vma = NULL; |
1da177e4c
|
441 442 443 |
unsigned long anon_mapping; rcu_read_lock(); |
80e148226
|
444 |
anon_mapping = (unsigned long) ACCESS_ONCE(page->mapping); |
3ca7b3c5b
|
445 |
if ((anon_mapping & PAGE_MAPPING_FLAGS) != PAGE_MAPPING_ANON) |
1da177e4c
|
446 447 448 449 450 |
goto out; if (!page_mapped(page)) goto out; anon_vma = (struct anon_vma *) (anon_mapping - PAGE_MAPPING_ANON); |
746b18d42
|
451 452 453 454 |
if (!atomic_inc_not_zero(&anon_vma->refcount)) { anon_vma = NULL; goto out; } |
f18194275
|
455 456 457 |
/* * If this page is still mapped, then its anon_vma cannot have been |
746b18d42
|
458 459 460 461 |
* freed. But if it has been unmapped, we have no security against the * anon_vma structure being freed and reused (for another anon_vma: * SLAB_DESTROY_BY_RCU guarantees that - so the atomic_inc_not_zero() * above cannot corrupt). |
f18194275
|
462 |
*/ |
746b18d42
|
463 464 465 466 |
if (!page_mapped(page)) { put_anon_vma(anon_vma); anon_vma = NULL; } |
1da177e4c
|
467 468 |
out: rcu_read_unlock(); |
746b18d42
|
469 470 471 |
return anon_vma; } |
88c22088b
|
472 473 474 475 476 477 478 |
/* * Similar to page_get_anon_vma() except it locks the anon_vma. * * Its a little more complex as it tries to keep the fast path to a single * atomic op -- the trylock. If we fail the trylock, we fall back to getting a * reference like with page_get_anon_vma() and then block on the mutex. */ |
746b18d42
|
479 480 |
struct anon_vma *page_lock_anon_vma(struct page *page) { |
88c22088b
|
481 |
struct anon_vma *anon_vma = NULL; |
eee0f252c
|
482 |
struct anon_vma *root_anon_vma; |
88c22088b
|
483 |
unsigned long anon_mapping; |
746b18d42
|
484 |
|
88c22088b
|
485 486 487 488 489 490 491 492 |
rcu_read_lock(); anon_mapping = (unsigned long) ACCESS_ONCE(page->mapping); if ((anon_mapping & PAGE_MAPPING_FLAGS) != PAGE_MAPPING_ANON) goto out; if (!page_mapped(page)) goto out; anon_vma = (struct anon_vma *) (anon_mapping - PAGE_MAPPING_ANON); |
eee0f252c
|
493 494 |
root_anon_vma = ACCESS_ONCE(anon_vma->root); if (mutex_trylock(&root_anon_vma->mutex)) { |
88c22088b
|
495 |
/* |
eee0f252c
|
496 497 |
* If the page is still mapped, then this anon_vma is still * its anon_vma, and holding the mutex ensures that it will |
bc658c960
|
498 |
* not go away, see anon_vma_free(). |
88c22088b
|
499 |
*/ |
eee0f252c
|
500 501 |
if (!page_mapped(page)) { mutex_unlock(&root_anon_vma->mutex); |
88c22088b
|
502 503 504 505 |
anon_vma = NULL; } goto out; } |
746b18d42
|
506 |
|
88c22088b
|
507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 |
/* trylock failed, we got to sleep */ if (!atomic_inc_not_zero(&anon_vma->refcount)) { anon_vma = NULL; goto out; } if (!page_mapped(page)) { put_anon_vma(anon_vma); anon_vma = NULL; goto out; } /* we pinned the anon_vma, its safe to sleep */ rcu_read_unlock(); anon_vma_lock(anon_vma); if (atomic_dec_and_test(&anon_vma->refcount)) { /* * Oops, we held the last refcount, release the lock * and bail -- can't simply use put_anon_vma() because * we'll deadlock on the anon_vma_lock() recursion. */ anon_vma_unlock(anon_vma); __put_anon_vma(anon_vma); anon_vma = NULL; } return anon_vma; out: rcu_read_unlock(); |
746b18d42
|
538 |
return anon_vma; |
34bbd7040
|
539 |
} |
10be22dfe
|
540 |
void page_unlock_anon_vma(struct anon_vma *anon_vma) |
34bbd7040
|
541 |
{ |
cba48b98f
|
542 |
anon_vma_unlock(anon_vma); |
1da177e4c
|
543 544 545 |
} /* |
3ad33b243
|
546 547 548 |
* 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
|
549 |
*/ |
71e3aac07
|
550 |
inline unsigned long |
1da177e4c
|
551 552 553 554 |
vma_address(struct page *page, struct vm_area_struct *vma) { pgoff_t pgoff = page->index << (PAGE_CACHE_SHIFT - PAGE_SHIFT); unsigned long address; |
0fe6e20b9
|
555 556 |
if (unlikely(is_vm_hugetlb_page(vma))) pgoff = page->index << huge_page_order(page_hstate(page)); |
1da177e4c
|
557 558 |
address = vma->vm_start + ((pgoff - vma->vm_pgoff) << PAGE_SHIFT); if (unlikely(address < vma->vm_start || address >= vma->vm_end)) { |
3ad33b243
|
559 |
/* page should be within @vma mapping range */ |
1da177e4c
|
560 561 562 563 564 565 |
return -EFAULT; } return address; } /* |
bf89c8c86
|
566 |
* At what user virtual address is page expected in vma? |
ab941e0ff
|
567 |
* Caller should check the page is actually part of the vma. |
1da177e4c
|
568 569 570 |
*/ unsigned long page_address_in_vma(struct page *page, struct vm_area_struct *vma) { |
21d0d443c
|
571 |
if (PageAnon(page)) { |
4829b906c
|
572 573 574 575 576 577 578 |
struct anon_vma *page__anon_vma = page_anon_vma(page); /* * Note: swapoff's unuse_vma() is more efficient with this * check, and needs it to match anon_vma when KSM is active. */ if (!vma->anon_vma || !page__anon_vma || vma->anon_vma->root != page__anon_vma->root) |
21d0d443c
|
579 580 |
return -EFAULT; } else if (page->mapping && !(vma->vm_flags & VM_NONLINEAR)) { |
ee498ed73
|
581 582 |
if (!vma->vm_file || vma->vm_file->f_mapping != page->mapping) |
1da177e4c
|
583 584 585 586 587 588 589 |
return -EFAULT; } else return -EFAULT; return vma_address(page, vma); } /* |
81b4082dc
|
590 591 |
* Check that @page is mapped at @address into @mm. * |
479db0bf4
|
592 593 594 595 |
* 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
|
596 |
* On success returns with pte mapped and locked. |
81b4082dc
|
597 |
*/ |
e9a81a821
|
598 |
pte_t *__page_check_address(struct page *page, struct mm_struct *mm, |
479db0bf4
|
599 |
unsigned long address, spinlock_t **ptlp, int sync) |
81b4082dc
|
600 601 602 603 604 |
{ pgd_t *pgd; pud_t *pud; pmd_t *pmd; pte_t *pte; |
c0718806c
|
605 |
spinlock_t *ptl; |
81b4082dc
|
606 |
|
0fe6e20b9
|
607 608 609 610 611 |
if (unlikely(PageHuge(page))) { pte = huge_pte_offset(mm, address); ptl = &mm->page_table_lock; goto check; } |
81b4082dc
|
612 |
pgd = pgd_offset(mm, address); |
c0718806c
|
613 614 615 616 617 618 619 620 621 622 |
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; |
71e3aac07
|
623 624 |
if (pmd_trans_huge(*pmd)) return NULL; |
c0718806c
|
625 626 627 |
pte = pte_offset_map(pmd, address); /* Make a quick check before getting the lock */ |
479db0bf4
|
628 |
if (!sync && !pte_present(*pte)) { |
c0718806c
|
629 630 631 |
pte_unmap(pte); return NULL; } |
4c21e2f24
|
632 |
ptl = pte_lockptr(mm, pmd); |
0fe6e20b9
|
633 |
check: |
c0718806c
|
634 635 636 637 |
spin_lock(ptl); if (pte_present(*pte) && page_to_pfn(page) == pte_pfn(*pte)) { *ptlp = ptl; return pte; |
81b4082dc
|
638 |
} |
c0718806c
|
639 640 |
pte_unmap_unlock(pte, ptl); return NULL; |
81b4082dc
|
641 |
} |
b291f0003
|
642 643 644 645 646 647 648 649 650 |
/** * page_mapped_in_vma - check whether a page is really mapped in a VMA * @page: the page to test * @vma: the VMA to test * * Returns 1 if the page is mapped into the page tables of the VMA, 0 * if the page is not mapped into the page tables of this VMA. Only * valid for normal file or anonymous VMAs. */ |
6a46079cf
|
651 |
int page_mapped_in_vma(struct page *page, struct vm_area_struct *vma) |
b291f0003
|
652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 |
{ unsigned long address; pte_t *pte; spinlock_t *ptl; address = vma_address(page, vma); if (address == -EFAULT) /* out of vma range */ return 0; pte = page_check_address(page, vma->vm_mm, address, &ptl, 1); if (!pte) /* the page is not in this mm */ return 0; pte_unmap_unlock(pte, ptl); return 1; } |
81b4082dc
|
667 |
/* |
1da177e4c
|
668 669 670 |
* Subfunctions of page_referenced: page_referenced_one called * repeatedly from either page_referenced_anon or page_referenced_file. */ |
5ad646880
|
671 672 673 |
int page_referenced_one(struct page *page, struct vm_area_struct *vma, unsigned long address, unsigned int *mapcount, unsigned long *vm_flags) |
1da177e4c
|
674 675 |
{ struct mm_struct *mm = vma->vm_mm; |
1da177e4c
|
676 |
int referenced = 0; |
71e3aac07
|
677 678 679 680 |
if (unlikely(PageTransHuge(page))) { pmd_t *pmd; spin_lock(&mm->page_table_lock); |
2da28bfd9
|
681 682 683 684 |
/* * rmap might return false positives; we must filter * these out using page_check_address_pmd(). */ |
71e3aac07
|
685 686 |
pmd = page_check_address_pmd(page, mm, address, PAGE_CHECK_ADDRESS_PMD_FLAG); |
2da28bfd9
|
687 688 689 690 691 692 693 694 695 696 697 698 699 700 |
if (!pmd) { spin_unlock(&mm->page_table_lock); goto out; } if (vma->vm_flags & VM_LOCKED) { spin_unlock(&mm->page_table_lock); *mapcount = 0; /* break early from loop */ *vm_flags |= VM_LOCKED; goto out; } /* go ahead even if the pmd is pmd_trans_splitting() */ if (pmdp_clear_flush_young_notify(vma, address, pmd)) |
71e3aac07
|
701 702 703 704 705 |
referenced++; spin_unlock(&mm->page_table_lock); } else { pte_t *pte; spinlock_t *ptl; |
2da28bfd9
|
706 707 708 709 |
/* * rmap might return false positives; we must filter * these out using page_check_address(). */ |
71e3aac07
|
710 711 712 |
pte = page_check_address(page, mm, address, &ptl, 0); if (!pte) goto out; |
2da28bfd9
|
713 714 715 716 717 718 |
if (vma->vm_flags & VM_LOCKED) { pte_unmap_unlock(pte, ptl); *mapcount = 0; /* break early from loop */ *vm_flags |= VM_LOCKED; goto out; } |
71e3aac07
|
719 720 721 722 723 724 725 726 727 728 729 730 731 |
if (ptep_clear_flush_young_notify(vma, address, pte)) { /* * Don't treat a reference through a sequentially read * mapping as such. If the page has been used in * another mapping, we will catch it; if this other * mapping is already gone, the unmap path will have * set PG_referenced or activated the page. */ if (likely(!VM_SequentialReadHint(vma))) referenced++; } pte_unmap_unlock(pte, ptl); } |
2da28bfd9
|
732 733 734 735 736 |
/* Pretend the page is referenced if the task has the swap token and is in the middle of a page fault. */ if (mm != current->mm && has_swap_token(mm) && rwsem_is_locked(&mm->mmap_sem)) referenced++; |
c0718806c
|
737 |
(*mapcount)--; |
273f047e3
|
738 |
|
6fe6b7e35
|
739 740 |
if (referenced) *vm_flags |= vma->vm_flags; |
273f047e3
|
741 |
out: |
1da177e4c
|
742 743 |
return referenced; } |
bed7161a5
|
744 |
static int page_referenced_anon(struct page *page, |
72835c86c
|
745 |
struct mem_cgroup *memcg, |
6fe6b7e35
|
746 |
unsigned long *vm_flags) |
1da177e4c
|
747 748 749 |
{ unsigned int mapcount; struct anon_vma *anon_vma; |
5beb49305
|
750 |
struct anon_vma_chain *avc; |
1da177e4c
|
751 752 753 754 755 756 757 |
int referenced = 0; anon_vma = page_lock_anon_vma(page); if (!anon_vma) return referenced; mapcount = page_mapcount(page); |
5beb49305
|
758 759 |
list_for_each_entry(avc, &anon_vma->head, same_anon_vma) { struct vm_area_struct *vma = avc->vma; |
1cb1729b1
|
760 761 762 |
unsigned long address = vma_address(page, vma); if (address == -EFAULT) continue; |
bed7161a5
|
763 764 765 766 767 |
/* * If we are reclaiming on behalf of a cgroup, skip * counting on behalf of references from different * cgroups */ |
72835c86c
|
768 |
if (memcg && !mm_match_cgroup(vma->vm_mm, memcg)) |
bed7161a5
|
769 |
continue; |
1cb1729b1
|
770 |
referenced += page_referenced_one(page, vma, address, |
6fe6b7e35
|
771 |
&mapcount, vm_flags); |
1da177e4c
|
772 773 774 |
if (!mapcount) break; } |
34bbd7040
|
775 776 |
page_unlock_anon_vma(anon_vma); |
1da177e4c
|
777 778 779 780 781 782 |
return referenced; } /** * page_referenced_file - referenced check for object-based rmap * @page: the page we're checking references on. |
72835c86c
|
783 |
* @memcg: target memory control group |
6fe6b7e35
|
784 |
* @vm_flags: collect encountered vma->vm_flags who actually referenced the page |
1da177e4c
|
785 786 787 788 789 790 791 792 |
* * 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
|
793 |
static int page_referenced_file(struct page *page, |
72835c86c
|
794 |
struct mem_cgroup *memcg, |
6fe6b7e35
|
795 |
unsigned long *vm_flags) |
1da177e4c
|
796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 |
{ 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, |
3d48ae45e
|
815 |
* so we can safely take mapping->i_mmap_mutex. |
1da177e4c
|
816 817 |
*/ BUG_ON(!PageLocked(page)); |
3d48ae45e
|
818 |
mutex_lock(&mapping->i_mmap_mutex); |
1da177e4c
|
819 820 |
/* |
3d48ae45e
|
821 |
* i_mmap_mutex does not stabilize mapcount at all, but mapcount |
1da177e4c
|
822 823 824 825 826 |
* 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) { |
1cb1729b1
|
827 828 829 |
unsigned long address = vma_address(page, vma); if (address == -EFAULT) continue; |
bed7161a5
|
830 831 832 833 834 |
/* * If we are reclaiming on behalf of a cgroup, skip * counting on behalf of references from different * cgroups */ |
72835c86c
|
835 |
if (memcg && !mm_match_cgroup(vma->vm_mm, memcg)) |
bed7161a5
|
836 |
continue; |
1cb1729b1
|
837 |
referenced += page_referenced_one(page, vma, address, |
6fe6b7e35
|
838 |
&mapcount, vm_flags); |
1da177e4c
|
839 840 841 |
if (!mapcount) break; } |
3d48ae45e
|
842 |
mutex_unlock(&mapping->i_mmap_mutex); |
1da177e4c
|
843 844 845 846 847 848 849 |
return referenced; } /** * page_referenced - test if the page was referenced * @page: the page to test * @is_locked: caller holds lock on the page |
72835c86c
|
850 |
* @memcg: target memory cgroup |
6fe6b7e35
|
851 |
* @vm_flags: collect encountered vma->vm_flags who actually referenced the page |
1da177e4c
|
852 853 854 855 |
* * Quick test_and_clear_referenced for all mappings to a page, * returns the number of ptes which referenced the page. */ |
6fe6b7e35
|
856 857 |
int page_referenced(struct page *page, int is_locked, |
72835c86c
|
858 |
struct mem_cgroup *memcg, |
6fe6b7e35
|
859 |
unsigned long *vm_flags) |
1da177e4c
|
860 861 |
{ int referenced = 0; |
5ad646880
|
862 |
int we_locked = 0; |
1da177e4c
|
863 |
|
6fe6b7e35
|
864 |
*vm_flags = 0; |
3ca7b3c5b
|
865 |
if (page_mapped(page) && page_rmapping(page)) { |
5ad646880
|
866 867 868 869 870 871 872 873 |
if (!is_locked && (!PageAnon(page) || PageKsm(page))) { we_locked = trylock_page(page); if (!we_locked) { referenced++; goto out; } } if (unlikely(PageKsm(page))) |
72835c86c
|
874 |
referenced += page_referenced_ksm(page, memcg, |
5ad646880
|
875 876 |
vm_flags); else if (PageAnon(page)) |
72835c86c
|
877 |
referenced += page_referenced_anon(page, memcg, |
6fe6b7e35
|
878 |
vm_flags); |
5ad646880
|
879 |
else if (page->mapping) |
72835c86c
|
880 |
referenced += page_referenced_file(page, memcg, |
6fe6b7e35
|
881 |
vm_flags); |
5ad646880
|
882 |
if (we_locked) |
1da177e4c
|
883 |
unlock_page(page); |
50a15981a
|
884 885 886 |
if (page_test_and_clear_young(page_to_pfn(page))) referenced++; |
1da177e4c
|
887 |
} |
5ad646880
|
888 |
out: |
1da177e4c
|
889 890 |
return referenced; } |
1cb1729b1
|
891 892 |
static int page_mkclean_one(struct page *page, struct vm_area_struct *vma, unsigned long address) |
d08b3851d
|
893 894 |
{ struct mm_struct *mm = vma->vm_mm; |
c2fda5fed
|
895 |
pte_t *pte; |
d08b3851d
|
896 897 |
spinlock_t *ptl; int ret = 0; |
479db0bf4
|
898 |
pte = page_check_address(page, mm, address, &ptl, 1); |
d08b3851d
|
899 900 |
if (!pte) goto out; |
c2fda5fed
|
901 902 |
if (pte_dirty(*pte) || pte_write(*pte)) { pte_t entry; |
d08b3851d
|
903 |
|
c2fda5fed
|
904 |
flush_cache_page(vma, address, pte_pfn(*pte)); |
cddb8a5c1
|
905 |
entry = ptep_clear_flush_notify(vma, address, pte); |
c2fda5fed
|
906 907 |
entry = pte_wrprotect(entry); entry = pte_mkclean(entry); |
d6e88e671
|
908 |
set_pte_at(mm, address, pte, entry); |
c2fda5fed
|
909 910 |
ret = 1; } |
d08b3851d
|
911 |
|
d08b3851d
|
912 913 914 915 916 917 918 919 920 921 922 923 924 |
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)); |
3d48ae45e
|
925 |
mutex_lock(&mapping->i_mmap_mutex); |
d08b3851d
|
926 |
vma_prio_tree_foreach(vma, &iter, &mapping->i_mmap, pgoff, pgoff) { |
1cb1729b1
|
927 928 929 930 931 932 |
if (vma->vm_flags & VM_SHARED) { unsigned long address = vma_address(page, vma); if (address == -EFAULT) continue; ret += page_mkclean_one(page, vma, address); } |
d08b3851d
|
933 |
} |
3d48ae45e
|
934 |
mutex_unlock(&mapping->i_mmap_mutex); |
d08b3851d
|
935 936 937 938 939 940 941 942 943 944 945 |
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
|
946 |
if (mapping) { |
d08b3851d
|
947 |
ret = page_mkclean_file(mapping, page); |
2d42552d1
|
948 |
if (page_test_and_clear_dirty(page_to_pfn(page), 1)) |
ce7e9fae8
|
949 |
ret = 1; |
6c210482a
|
950 |
} |
d08b3851d
|
951 952 953 954 |
} return ret; } |
60b59beaf
|
955 |
EXPORT_SYMBOL_GPL(page_mkclean); |
d08b3851d
|
956 |
|
1da177e4c
|
957 |
/** |
c44b67432
|
958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 |
* page_move_anon_rmap - move a page to our anon_vma * @page: the page to move to our anon_vma * @vma: the vma the page belongs to * @address: the user virtual address mapped * * When a page belongs exclusively to one process after a COW event, * that page can be moved into the anon_vma that belongs to just that * process, so the rmap code will not search the parent or sibling * processes. */ void page_move_anon_rmap(struct page *page, struct vm_area_struct *vma, unsigned long address) { struct anon_vma *anon_vma = vma->anon_vma; VM_BUG_ON(!PageLocked(page)); VM_BUG_ON(!anon_vma); VM_BUG_ON(page->index != linear_page_index(vma, address)); anon_vma = (void *) anon_vma + PAGE_MAPPING_ANON; page->mapping = (struct address_space *) anon_vma; } /** |
4e1c19750
|
982 983 984 985 |
* __page_set_anon_rmap - set up new anonymous rmap * @page: Page to add to rmap * @vma: VM area to add page to. * @address: User virtual address of the mapping |
e8a03feb5
|
986 |
* @exclusive: the page is exclusively owned by the current process |
9617d95e6
|
987 988 |
*/ static void __page_set_anon_rmap(struct page *page, |
e8a03feb5
|
989 |
struct vm_area_struct *vma, unsigned long address, int exclusive) |
9617d95e6
|
990 |
{ |
e8a03feb5
|
991 |
struct anon_vma *anon_vma = vma->anon_vma; |
ea90002b0
|
992 |
|
e8a03feb5
|
993 |
BUG_ON(!anon_vma); |
ea90002b0
|
994 |
|
4e1c19750
|
995 996 |
if (PageAnon(page)) return; |
ea90002b0
|
997 |
/* |
e8a03feb5
|
998 999 1000 |
* If the page isn't exclusively mapped into this vma, * we must use the _oldest_ possible anon_vma for the * page mapping! |
ea90002b0
|
1001 |
*/ |
4e1c19750
|
1002 |
if (!exclusive) |
288468c33
|
1003 |
anon_vma = anon_vma->root; |
9617d95e6
|
1004 |
|
9617d95e6
|
1005 1006 |
anon_vma = (void *) anon_vma + PAGE_MAPPING_ANON; page->mapping = (struct address_space *) anon_vma; |
9617d95e6
|
1007 |
page->index = linear_page_index(vma, address); |
9617d95e6
|
1008 1009 1010 |
} /** |
43d8eac44
|
1011 |
* __page_check_anon_rmap - sanity check anonymous rmap addition |
c97a9e10e
|
1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 |
* @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. */ |
44ab57a06
|
1032 |
BUG_ON(page_anon_vma(page)->root != vma->anon_vma->root); |
c97a9e10e
|
1033 1034 1035 1036 1037 |
BUG_ON(page->index != linear_page_index(vma, address)); #endif } /** |
1da177e4c
|
1038 1039 1040 1041 1042 |
* 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 * |
5ad646880
|
1043 |
* The caller needs to hold the pte lock, and the page must be locked in |
80e148226
|
1044 1045 1046 |
* the anon_vma case: to serialize mapping,index checking after setting, * and to ensure that PageAnon is not being upgraded racily to PageKsm * (but PageKsm is never downgraded to PageAnon). |
1da177e4c
|
1047 1048 1049 1050 |
*/ void page_add_anon_rmap(struct page *page, struct vm_area_struct *vma, unsigned long address) { |
ad8c2ee80
|
1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 |
do_page_add_anon_rmap(page, vma, address, 0); } /* * Special version of the above for do_swap_page, which often runs * into pages that are exclusively owned by the current process. * Everybody else should continue to use page_add_anon_rmap above. */ void do_page_add_anon_rmap(struct page *page, struct vm_area_struct *vma, unsigned long address, int exclusive) { |
5ad646880
|
1062 |
int first = atomic_inc_and_test(&page->_mapcount); |
79134171d
|
1063 1064 1065 1066 1067 1068 1069 |
if (first) { if (!PageTransHuge(page)) __inc_zone_page_state(page, NR_ANON_PAGES); else __inc_zone_page_state(page, NR_ANON_TRANSPARENT_HUGEPAGES); } |
5ad646880
|
1070 1071 |
if (unlikely(PageKsm(page))) return; |
c97a9e10e
|
1072 |
VM_BUG_ON(!PageLocked(page)); |
5dbe0af47
|
1073 |
/* address might be in next vma when migration races vma_adjust */ |
5ad646880
|
1074 |
if (first) |
ad8c2ee80
|
1075 |
__page_set_anon_rmap(page, vma, address, exclusive); |
69029cd55
|
1076 |
else |
c97a9e10e
|
1077 |
__page_check_anon_rmap(page, vma, address); |
1da177e4c
|
1078 |
} |
43d8eac44
|
1079 |
/** |
9617d95e6
|
1080 1081 1082 1083 1084 1085 1086 |
* 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
|
1087 |
* Page does not have to be locked. |
9617d95e6
|
1088 1089 1090 1091 |
*/ void page_add_new_anon_rmap(struct page *page, struct vm_area_struct *vma, unsigned long address) { |
b5934c531
|
1092 |
VM_BUG_ON(address < vma->vm_start || address >= vma->vm_end); |
cbf84b7ad
|
1093 1094 |
SetPageSwapBacked(page); atomic_set(&page->_mapcount, 0); /* increment count (starts at -1) */ |
79134171d
|
1095 1096 1097 1098 |
if (!PageTransHuge(page)) __inc_zone_page_state(page, NR_ANON_PAGES); else __inc_zone_page_state(page, NR_ANON_TRANSPARENT_HUGEPAGES); |
e8a03feb5
|
1099 |
__page_set_anon_rmap(page, vma, address, 1); |
b5934c531
|
1100 |
if (page_evictable(page, vma)) |
cbf84b7ad
|
1101 |
lru_cache_add_lru(page, LRU_ACTIVE_ANON); |
b5934c531
|
1102 1103 |
else add_page_to_unevictable_list(page); |
9617d95e6
|
1104 |
} |
1da177e4c
|
1105 1106 1107 1108 |
/** * page_add_file_rmap - add pte mapping to a file page * @page: the page to add the mapping to * |
b8072f099
|
1109 |
* The caller needs to hold the pte lock. |
1da177e4c
|
1110 1111 1112 |
*/ void page_add_file_rmap(struct page *page) { |
d69b042f3
|
1113 |
if (atomic_inc_and_test(&page->_mapcount)) { |
65ba55f50
|
1114 |
__inc_zone_page_state(page, NR_FILE_MAPPED); |
2a7106f2c
|
1115 |
mem_cgroup_inc_page_stat(page, MEMCG_NR_FILE_MAPPED); |
d69b042f3
|
1116 |
} |
1da177e4c
|
1117 1118 1119 1120 1121 1122 |
} /** * page_remove_rmap - take down pte mapping from a page * @page: page to remove mapping from * |
b8072f099
|
1123 |
* The caller needs to hold the pte lock. |
1da177e4c
|
1124 |
*/ |
edc315fd2
|
1125 |
void page_remove_rmap(struct page *page) |
1da177e4c
|
1126 |
{ |
b904dcfed
|
1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 |
/* page still mapped by someone else? */ if (!atomic_add_negative(-1, &page->_mapcount)) return; /* * 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. */ |
2d42552d1
|
1138 1139 |
if ((!PageAnon(page) || PageSwapCache(page)) && page_test_and_clear_dirty(page_to_pfn(page), 1)) |
b904dcfed
|
1140 |
set_page_dirty(page); |
0fe6e20b9
|
1141 1142 1143 1144 1145 1146 |
/* * Hugepages are not counted in NR_ANON_PAGES nor NR_FILE_MAPPED * and not charged by memcg for now. */ if (unlikely(PageHuge(page))) return; |
b904dcfed
|
1147 1148 |
if (PageAnon(page)) { mem_cgroup_uncharge_page(page); |
79134171d
|
1149 1150 1151 1152 1153 |
if (!PageTransHuge(page)) __dec_zone_page_state(page, NR_ANON_PAGES); else __dec_zone_page_state(page, NR_ANON_TRANSPARENT_HUGEPAGES); |
b904dcfed
|
1154 1155 |
} else { __dec_zone_page_state(page, NR_FILE_MAPPED); |
2a7106f2c
|
1156 |
mem_cgroup_dec_page_stat(page, MEMCG_NR_FILE_MAPPED); |
b904dcfed
|
1157 |
} |
b904dcfed
|
1158 1159 1160 1161 1162 1163 1164 1165 1166 |
/* * 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
|
1167 1168 1169 1170 |
} /* * Subfunctions of try_to_unmap: try_to_unmap_one called |
99ef0315f
|
1171 |
* repeatedly from try_to_unmap_ksm, try_to_unmap_anon or try_to_unmap_file. |
1da177e4c
|
1172 |
*/ |
5ad646880
|
1173 1174 |
int try_to_unmap_one(struct page *page, struct vm_area_struct *vma, unsigned long address, enum ttu_flags flags) |
1da177e4c
|
1175 1176 |
{ struct mm_struct *mm = vma->vm_mm; |
1da177e4c
|
1177 1178 |
pte_t *pte; pte_t pteval; |
c0718806c
|
1179 |
spinlock_t *ptl; |
1da177e4c
|
1180 |
int ret = SWAP_AGAIN; |
479db0bf4
|
1181 |
pte = page_check_address(page, mm, address, &ptl, 0); |
c0718806c
|
1182 |
if (!pte) |
81b4082dc
|
1183 |
goto out; |
1da177e4c
|
1184 1185 1186 1187 1188 1189 |
/* * 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. */ |
14fa31b89
|
1190 |
if (!(flags & TTU_IGNORE_MLOCK)) { |
caed0f486
|
1191 1192 |
if (vma->vm_flags & VM_LOCKED) goto out_mlock; |
af8e3354b
|
1193 |
if (TTU_ACTION(flags) == TTU_MUNLOCK) |
53f79acb6
|
1194 |
goto out_unmap; |
14fa31b89
|
1195 1196 |
} if (!(flags & TTU_IGNORE_ACCESS)) { |
b291f0003
|
1197 1198 1199 1200 1201 |
if (ptep_clear_flush_young_notify(vma, address, pte)) { ret = SWAP_FAIL; goto out_unmap; } } |
1da177e4c
|
1202 |
|
1da177e4c
|
1203 1204 |
/* Nuke the page table entry. */ flush_cache_page(vma, address, page_to_pfn(page)); |
cddb8a5c1
|
1205 |
pteval = ptep_clear_flush_notify(vma, address, pte); |
1da177e4c
|
1206 1207 1208 1209 |
/* Move the dirty bit to the physical page now the pte is gone. */ if (pte_dirty(pteval)) set_page_dirty(page); |
365e9c87a
|
1210 1211 |
/* Update high watermark before we lower rss */ update_hiwater_rss(mm); |
888b9f7c5
|
1212 1213 |
if (PageHWPoison(page) && !(flags & TTU_IGNORE_HWPOISON)) { if (PageAnon(page)) |
d559db086
|
1214 |
dec_mm_counter(mm, MM_ANONPAGES); |
888b9f7c5
|
1215 |
else |
d559db086
|
1216 |
dec_mm_counter(mm, MM_FILEPAGES); |
888b9f7c5
|
1217 1218 1219 |
set_pte_at(mm, address, pte, swp_entry_to_pte(make_hwpoison_entry(page))); } else if (PageAnon(page)) { |
4c21e2f24
|
1220 |
swp_entry_t entry = { .val = page_private(page) }; |
0697212a4
|
1221 1222 1223 1224 1225 1226 |
if (PageSwapCache(page)) { /* * Store the swap location in the pte. * See handle_pte_fault() ... */ |
570a335b8
|
1227 1228 1229 1230 1231 |
if (swap_duplicate(entry) < 0) { set_pte_at(mm, address, pte, pteval); ret = SWAP_FAIL; goto out_unmap; } |
0697212a4
|
1232 1233 1234 1235 1236 1237 |
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); } |
d559db086
|
1238 |
dec_mm_counter(mm, MM_ANONPAGES); |
b084d4353
|
1239 |
inc_mm_counter(mm, MM_SWAPENTS); |
64cdd548f
|
1240 |
} else if (PAGE_MIGRATION) { |
0697212a4
|
1241 1242 1243 1244 1245 |
/* * 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. */ |
14fa31b89
|
1246 |
BUG_ON(TTU_ACTION(flags) != TTU_MIGRATION); |
0697212a4
|
1247 |
entry = make_migration_entry(page, pte_write(pteval)); |
1da177e4c
|
1248 1249 1250 |
} set_pte_at(mm, address, pte, swp_entry_to_pte(entry)); BUG_ON(pte_file(*pte)); |
14fa31b89
|
1251 |
} else if (PAGE_MIGRATION && (TTU_ACTION(flags) == TTU_MIGRATION)) { |
04e62a29b
|
1252 1253 1254 1255 1256 |
/* 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 |
d559db086
|
1257 |
dec_mm_counter(mm, MM_FILEPAGES); |
1da177e4c
|
1258 |
|
edc315fd2
|
1259 |
page_remove_rmap(page); |
1da177e4c
|
1260 1261 1262 |
page_cache_release(page); out_unmap: |
c0718806c
|
1263 |
pte_unmap_unlock(pte, ptl); |
caed0f486
|
1264 1265 |
out: return ret; |
53f79acb6
|
1266 |
|
caed0f486
|
1267 1268 1269 1270 1271 1272 1273 |
out_mlock: pte_unmap_unlock(pte, ptl); /* * We need mmap_sem locking, Otherwise VM_LOCKED check makes * unstable result and race. Plus, We can't wait here because |
2b575eb64
|
1274 |
* we now hold anon_vma->mutex or mapping->i_mmap_mutex. |
caed0f486
|
1275 1276 1277 1278 1279 1280 1281 1282 |
* if trylock failed, the page remain in evictable lru and later * vmscan could retry to move the page to unevictable lru if the * page is actually mlocked. */ if (down_read_trylock(&vma->vm_mm->mmap_sem)) { if (vma->vm_flags & VM_LOCKED) { mlock_vma_page(page); ret = SWAP_MLOCK; |
53f79acb6
|
1283 |
} |
caed0f486
|
1284 |
up_read(&vma->vm_mm->mmap_sem); |
53f79acb6
|
1285 |
} |
1da177e4c
|
1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 |
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. |
b291f0003
|
1307 1308 1309 1310 1311 |
* * Mlocked pages: check VM_LOCKED under mmap_sem held for read, if we can * acquire it without blocking. If vma locked, mlock the pages in the cluster, * rather than unmapping them. If we encounter the "check_page" that vmscan is * trying to unmap, return SWAP_MLOCK, else default SWAP_AGAIN. |
1da177e4c
|
1312 1313 1314 |
*/ #define CLUSTER_SIZE min(32*PAGE_SIZE, PMD_SIZE) #define CLUSTER_MASK (~(CLUSTER_SIZE - 1)) |
b291f0003
|
1315 1316 |
static int try_to_unmap_cluster(unsigned long cursor, unsigned int *mapcount, struct vm_area_struct *vma, struct page *check_page) |
1da177e4c
|
1317 1318 1319 1320 1321 |
{ struct mm_struct *mm = vma->vm_mm; pgd_t *pgd; pud_t *pud; pmd_t *pmd; |
c0718806c
|
1322 |
pte_t *pte; |
1da177e4c
|
1323 |
pte_t pteval; |
c0718806c
|
1324 |
spinlock_t *ptl; |
1da177e4c
|
1325 1326 1327 |
struct page *page; unsigned long address; unsigned long end; |
b291f0003
|
1328 1329 |
int ret = SWAP_AGAIN; int locked_vma = 0; |
1da177e4c
|
1330 |
|
1da177e4c
|
1331 1332 1333 1334 1335 1336 1337 1338 1339 |
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)) |
b291f0003
|
1340 |
return ret; |
1da177e4c
|
1341 1342 1343 |
pud = pud_offset(pgd, address); if (!pud_present(*pud)) |
b291f0003
|
1344 |
return ret; |
1da177e4c
|
1345 1346 1347 |
pmd = pmd_offset(pud, address); if (!pmd_present(*pmd)) |
b291f0003
|
1348 1349 1350 |
return ret; /* |
af8e3354b
|
1351 |
* If we can acquire the mmap_sem for read, and vma is VM_LOCKED, |
b291f0003
|
1352 1353 |
* keep the sem while scanning the cluster for mlocking pages. */ |
af8e3354b
|
1354 |
if (down_read_trylock(&vma->vm_mm->mmap_sem)) { |
b291f0003
|
1355 1356 1357 1358 |
locked_vma = (vma->vm_flags & VM_LOCKED); if (!locked_vma) up_read(&vma->vm_mm->mmap_sem); /* don't need it */ } |
c0718806c
|
1359 1360 |
pte = pte_offset_map_lock(mm, pmd, address, &ptl); |
1da177e4c
|
1361 |
|
365e9c87a
|
1362 1363 |
/* Update high watermark before we lower rss */ update_hiwater_rss(mm); |
c0718806c
|
1364 |
for (; address < end; pte++, address += PAGE_SIZE) { |
1da177e4c
|
1365 1366 |
if (!pte_present(*pte)) continue; |
6aab341e0
|
1367 1368 |
page = vm_normal_page(vma, address, *pte); BUG_ON(!page || PageAnon(page)); |
1da177e4c
|
1369 |
|
b291f0003
|
1370 1371 1372 1373 1374 1375 |
if (locked_vma) { mlock_vma_page(page); /* no-op if already mlocked */ if (page == check_page) ret = SWAP_MLOCK; continue; /* don't unmap */ } |
cddb8a5c1
|
1376 |
if (ptep_clear_flush_young_notify(vma, address, pte)) |
1da177e4c
|
1377 1378 1379 |
continue; /* Nuke the page table entry. */ |
eca351336
|
1380 |
flush_cache_page(vma, address, pte_pfn(*pte)); |
cddb8a5c1
|
1381 |
pteval = ptep_clear_flush_notify(vma, address, pte); |
1da177e4c
|
1382 1383 1384 1385 1386 1387 1388 1389 |
/* 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); |
edc315fd2
|
1390 |
page_remove_rmap(page); |
1da177e4c
|
1391 |
page_cache_release(page); |
d559db086
|
1392 |
dec_mm_counter(mm, MM_FILEPAGES); |
1da177e4c
|
1393 1394 |
(*mapcount)--; } |
c0718806c
|
1395 |
pte_unmap_unlock(pte - 1, ptl); |
b291f0003
|
1396 1397 1398 |
if (locked_vma) up_read(&vma->vm_mm->mmap_sem); return ret; |
1da177e4c
|
1399 |
} |
71e3aac07
|
1400 |
bool is_vma_temporary_stack(struct vm_area_struct *vma) |
a8bef8ff6
|
1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 |
{ int maybe_stack = vma->vm_flags & (VM_GROWSDOWN | VM_GROWSUP); if (!maybe_stack) return false; if ((vma->vm_flags & VM_STACK_INCOMPLETE_SETUP) == VM_STACK_INCOMPLETE_SETUP) return true; return false; } |
b291f0003
|
1413 1414 1415 1416 |
/** * try_to_unmap_anon - unmap or unlock anonymous page using the object-based * rmap method * @page: the page to unmap/unlock |
8051be5e6
|
1417 |
* @flags: action and flags |
b291f0003
|
1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 |
* * Find all the mappings of a page using the mapping pointer and the vma chains * contained in the anon_vma struct it points to. * * This function is only called from try_to_unmap/try_to_munlock for * anonymous pages. * When called from try_to_munlock(), the mmap_sem of the mm containing the vma * where the page was found will be held for write. So, we won't recheck * vm_flags for that VMA. That should be OK, because that vma shouldn't be * 'LOCKED. */ |
14fa31b89
|
1429 |
static int try_to_unmap_anon(struct page *page, enum ttu_flags flags) |
1da177e4c
|
1430 1431 |
{ struct anon_vma *anon_vma; |
5beb49305
|
1432 |
struct anon_vma_chain *avc; |
1da177e4c
|
1433 |
int ret = SWAP_AGAIN; |
b291f0003
|
1434 |
|
1da177e4c
|
1435 1436 1437 |
anon_vma = page_lock_anon_vma(page); if (!anon_vma) return ret; |
5beb49305
|
1438 1439 |
list_for_each_entry(avc, &anon_vma->head, same_anon_vma) { struct vm_area_struct *vma = avc->vma; |
a8bef8ff6
|
1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 |
unsigned long address; /* * During exec, a temporary VMA is setup and later moved. * The VMA is moved under the anon_vma lock but not the * page tables leading to a race where migration cannot * find the migration ptes. Rather than increasing the * locking requirements of exec(), migration skips * temporary VMAs until after exec() completes. */ if (PAGE_MIGRATION && (flags & TTU_MIGRATION) && is_vma_temporary_stack(vma)) continue; address = vma_address(page, vma); |
1cb1729b1
|
1455 1456 1457 |
if (address == -EFAULT) continue; ret = try_to_unmap_one(page, vma, address, flags); |
53f79acb6
|
1458 1459 |
if (ret != SWAP_AGAIN || !page_mapped(page)) break; |
1da177e4c
|
1460 |
} |
34bbd7040
|
1461 1462 |
page_unlock_anon_vma(anon_vma); |
1da177e4c
|
1463 1464 1465 1466 |
return ret; } /** |
b291f0003
|
1467 1468 |
* try_to_unmap_file - unmap/unlock file page using the object-based rmap method * @page: the page to unmap/unlock |
14fa31b89
|
1469 |
* @flags: action and flags |
1da177e4c
|
1470 1471 1472 1473 |
* * Find all the mappings of a page using the mapping pointer and the vma chains * contained in the address_space struct it points to. * |
b291f0003
|
1474 1475 1476 1477 1478 1479 |
* This function is only called from try_to_unmap/try_to_munlock for * object-based pages. * When called from try_to_munlock(), the mmap_sem of the mm containing the vma * where the page was found will be held for write. So, we won't recheck * vm_flags for that VMA. That should be OK, because that vma shouldn't be * 'LOCKED. |
1da177e4c
|
1480 |
*/ |
14fa31b89
|
1481 |
static int try_to_unmap_file(struct page *page, enum ttu_flags flags) |
1da177e4c
|
1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 |
{ 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; |
3d48ae45e
|
1492 |
mutex_lock(&mapping->i_mmap_mutex); |
1da177e4c
|
1493 |
vma_prio_tree_foreach(vma, &iter, &mapping->i_mmap, pgoff, pgoff) { |
1cb1729b1
|
1494 1495 1496 1497 |
unsigned long address = vma_address(page, vma); if (address == -EFAULT) continue; ret = try_to_unmap_one(page, vma, address, flags); |
53f79acb6
|
1498 1499 |
if (ret != SWAP_AGAIN || !page_mapped(page)) goto out; |
1da177e4c
|
1500 1501 1502 1503 |
} if (list_empty(&mapping->i_mmap_nonlinear)) goto out; |
53f79acb6
|
1504 1505 1506 1507 1508 1509 1510 |
/* * We don't bother to try to find the munlocked page in nonlinears. * It's costly. Instead, later, page reclaim logic may call * try_to_unmap(TTU_MUNLOCK) and recover PG_mlocked lazily. */ if (TTU_ACTION(flags) == TTU_MUNLOCK) goto out; |
1da177e4c
|
1511 1512 |
list_for_each_entry(vma, &mapping->i_mmap_nonlinear, shared.vm_set.list) { |
1da177e4c
|
1513 1514 1515 1516 1517 1518 1519 |
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; } |
b291f0003
|
1520 |
if (max_nl_size == 0) { /* all nonlinears locked or reserved ? */ |
1da177e4c
|
1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 |
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; |
3d48ae45e
|
1535 |
cond_resched(); |
1da177e4c
|
1536 1537 1538 1539 1540 1541 1542 1543 |
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) { |
1da177e4c
|
1544 |
cursor = (unsigned long) vma->vm_private_data; |
839b9685e
|
1545 |
while ( cursor < max_nl_cursor && |
1da177e4c
|
1546 |
cursor < vma->vm_end - vma->vm_start) { |
53f79acb6
|
1547 1548 1549 |
if (try_to_unmap_cluster(cursor, &mapcount, vma, page) == SWAP_MLOCK) ret = SWAP_MLOCK; |
1da177e4c
|
1550 1551 1552 1553 1554 1555 1556 |
cursor += CLUSTER_SIZE; vma->vm_private_data = (void *) cursor; if ((int)mapcount <= 0) goto out; } vma->vm_private_data = (void *) max_nl_cursor; } |
3d48ae45e
|
1557 |
cond_resched(); |
1da177e4c
|
1558 1559 1560 1561 1562 1563 1564 1565 |
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
|
1566 1567 |
list_for_each_entry(vma, &mapping->i_mmap_nonlinear, shared.vm_set.list) vma->vm_private_data = NULL; |
1da177e4c
|
1568 |
out: |
3d48ae45e
|
1569 |
mutex_unlock(&mapping->i_mmap_mutex); |
1da177e4c
|
1570 1571 1572 1573 1574 1575 |
return ret; } /** * try_to_unmap - try to remove all page table mappings to a page * @page: the page to get unmapped |
14fa31b89
|
1576 |
* @flags: action and flags |
1da177e4c
|
1577 1578 1579 1580 1581 1582 1583 1584 |
* * 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 |
b291f0003
|
1585 |
* SWAP_MLOCK - page is mlocked. |
1da177e4c
|
1586 |
*/ |
14fa31b89
|
1587 |
int try_to_unmap(struct page *page, enum ttu_flags flags) |
1da177e4c
|
1588 1589 |
{ int ret; |
1da177e4c
|
1590 |
BUG_ON(!PageLocked(page)); |
91600e9e5
|
1591 |
VM_BUG_ON(!PageHuge(page) && PageTransHuge(page)); |
1da177e4c
|
1592 |
|
5ad646880
|
1593 1594 1595 |
if (unlikely(PageKsm(page))) ret = try_to_unmap_ksm(page, flags); else if (PageAnon(page)) |
14fa31b89
|
1596 |
ret = try_to_unmap_anon(page, flags); |
1da177e4c
|
1597 |
else |
14fa31b89
|
1598 |
ret = try_to_unmap_file(page, flags); |
b291f0003
|
1599 |
if (ret != SWAP_MLOCK && !page_mapped(page)) |
1da177e4c
|
1600 1601 1602 |
ret = SWAP_SUCCESS; return ret; } |
81b4082dc
|
1603 |
|
b291f0003
|
1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 |
/** * try_to_munlock - try to munlock a page * @page: the page to be munlocked * * Called from munlock code. Checks all of the VMAs mapping the page * to make sure nobody else has this page mlocked. The page will be * returned with PG_mlocked cleared if no other vmas have it mlocked. * * Return values are: * |
53f79acb6
|
1614 |
* SWAP_AGAIN - no vma is holding page mlocked, or, |
b291f0003
|
1615 |
* SWAP_AGAIN - page mapped in mlocked vma -- couldn't acquire mmap sem |
5ad646880
|
1616 |
* SWAP_FAIL - page cannot be located at present |
b291f0003
|
1617 1618 1619 1620 1621 |
* SWAP_MLOCK - page is now mlocked. */ int try_to_munlock(struct page *page) { VM_BUG_ON(!PageLocked(page) || PageLRU(page)); |
5ad646880
|
1622 1623 1624 |
if (unlikely(PageKsm(page))) return try_to_unmap_ksm(page, TTU_MUNLOCK); else if (PageAnon(page)) |
14fa31b89
|
1625 |
return try_to_unmap_anon(page, TTU_MUNLOCK); |
b291f0003
|
1626 |
else |
14fa31b89
|
1627 |
return try_to_unmap_file(page, TTU_MUNLOCK); |
b291f0003
|
1628 |
} |
e9995ef97
|
1629 |
|
01d8b20de
|
1630 |
void __put_anon_vma(struct anon_vma *anon_vma) |
76545066c
|
1631 |
{ |
01d8b20de
|
1632 |
struct anon_vma *root = anon_vma->root; |
76545066c
|
1633 |
|
01d8b20de
|
1634 1635 |
if (root != anon_vma && atomic_dec_and_test(&root->refcount)) anon_vma_free(root); |
76545066c
|
1636 |
|
01d8b20de
|
1637 |
anon_vma_free(anon_vma); |
76545066c
|
1638 |
} |
76545066c
|
1639 |
|
e9995ef97
|
1640 1641 1642 1643 1644 1645 1646 1647 1648 |
#ifdef CONFIG_MIGRATION /* * rmap_walk() and its helpers rmap_walk_anon() and rmap_walk_file(): * Called by migrate.c to remove migration ptes, but might be used more later. */ static int rmap_walk_anon(struct page *page, int (*rmap_one)(struct page *, struct vm_area_struct *, unsigned long, void *), void *arg) { struct anon_vma *anon_vma; |
5beb49305
|
1649 |
struct anon_vma_chain *avc; |
e9995ef97
|
1650 1651 1652 1653 1654 |
int ret = SWAP_AGAIN; /* * Note: remove_migration_ptes() cannot use page_lock_anon_vma() * because that depends on page_mapped(); but not all its usages |
3f6c82728
|
1655 1656 |
* are holding mmap_sem. Users without mmap_sem are required to * take a reference count to prevent the anon_vma disappearing |
e9995ef97
|
1657 1658 1659 1660 |
*/ anon_vma = page_anon_vma(page); if (!anon_vma) return ret; |
cba48b98f
|
1661 |
anon_vma_lock(anon_vma); |
5beb49305
|
1662 1663 |
list_for_each_entry(avc, &anon_vma->head, same_anon_vma) { struct vm_area_struct *vma = avc->vma; |
e9995ef97
|
1664 1665 1666 1667 1668 1669 1670 |
unsigned long address = vma_address(page, vma); if (address == -EFAULT) continue; ret = rmap_one(page, vma, address, arg); if (ret != SWAP_AGAIN) break; } |
cba48b98f
|
1671 |
anon_vma_unlock(anon_vma); |
e9995ef97
|
1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 |
return ret; } static int rmap_walk_file(struct page *page, int (*rmap_one)(struct page *, struct vm_area_struct *, unsigned long, void *), void *arg) { 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; if (!mapping) return ret; |
3d48ae45e
|
1686 |
mutex_lock(&mapping->i_mmap_mutex); |
e9995ef97
|
1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 |
vma_prio_tree_foreach(vma, &iter, &mapping->i_mmap, pgoff, pgoff) { unsigned long address = vma_address(page, vma); if (address == -EFAULT) continue; ret = rmap_one(page, vma, address, arg); if (ret != SWAP_AGAIN) break; } /* * No nonlinear handling: being always shared, nonlinear vmas * never contain migration ptes. Decide what to do about this * limitation to linear when we need rmap_walk() on nonlinear. */ |
3d48ae45e
|
1700 |
mutex_unlock(&mapping->i_mmap_mutex); |
e9995ef97
|
1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 |
return ret; } int rmap_walk(struct page *page, int (*rmap_one)(struct page *, struct vm_area_struct *, unsigned long, void *), void *arg) { VM_BUG_ON(!PageLocked(page)); if (unlikely(PageKsm(page))) return rmap_walk_ksm(page, rmap_one, arg); else if (PageAnon(page)) return rmap_walk_anon(page, rmap_one, arg); else return rmap_walk_file(page, rmap_one, arg); } #endif /* CONFIG_MIGRATION */ |
0fe6e20b9
|
1717 |
|
e3390f67a
|
1718 |
#ifdef CONFIG_HUGETLB_PAGE |
0fe6e20b9
|
1719 1720 1721 1722 1723 1724 1725 1726 1727 |
/* * The following three functions are for anonymous (private mapped) hugepages. * Unlike common anonymous pages, anonymous hugepages have no accounting code * and no lru code, because we handle hugepages differently from common pages. */ static void __hugepage_set_anon_rmap(struct page *page, struct vm_area_struct *vma, unsigned long address, int exclusive) { struct anon_vma *anon_vma = vma->anon_vma; |
433abed6c
|
1728 |
|
0fe6e20b9
|
1729 |
BUG_ON(!anon_vma); |
433abed6c
|
1730 1731 1732 1733 1734 |
if (PageAnon(page)) return; if (!exclusive) anon_vma = anon_vma->root; |
0fe6e20b9
|
1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 |
anon_vma = (void *) anon_vma + PAGE_MAPPING_ANON; page->mapping = (struct address_space *) anon_vma; page->index = linear_page_index(vma, address); } void hugepage_add_anon_rmap(struct page *page, struct vm_area_struct *vma, unsigned long address) { struct anon_vma *anon_vma = vma->anon_vma; int first; |
a850ea303
|
1745 1746 |
BUG_ON(!PageLocked(page)); |
0fe6e20b9
|
1747 |
BUG_ON(!anon_vma); |
5dbe0af47
|
1748 |
/* address might be in next vma when migration races vma_adjust */ |
0fe6e20b9
|
1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 |
first = atomic_inc_and_test(&page->_mapcount); if (first) __hugepage_set_anon_rmap(page, vma, address, 0); } void hugepage_add_new_anon_rmap(struct page *page, struct vm_area_struct *vma, unsigned long address) { BUG_ON(address < vma->vm_start || address >= vma->vm_end); atomic_set(&page->_mapcount, 0); __hugepage_set_anon_rmap(page, vma, address, 1); } |
e3390f67a
|
1761 |
#endif /* CONFIG_HUGETLB_PAGE */ |