Blame view
mm/ksm.c
54.4 KB
f8af4da3b
|
1 |
/* |
31dbd01f3
|
2 3 4 5 6 |
* Memory merging support. * * This code enables dynamic sharing of identical pages found in different * memory areas, even if they are not shared by fork() * |
36b2528dc
|
7 |
* Copyright (C) 2008-2009 Red Hat, Inc. |
31dbd01f3
|
8 9 10 11 |
* Authors: * Izik Eidus * Andrea Arcangeli * Chris Wright |
36b2528dc
|
12 |
* Hugh Dickins |
31dbd01f3
|
13 14 |
* * This work is licensed under the terms of the GNU GPL, version 2. |
f8af4da3b
|
15 16 17 |
*/ #include <linux/errno.h> |
31dbd01f3
|
18 19 |
#include <linux/mm.h> #include <linux/fs.h> |
f8af4da3b
|
20 |
#include <linux/mman.h> |
31dbd01f3
|
21 22 23 24 25 26 27 28 29 30 31 |
#include <linux/sched.h> #include <linux/rwsem.h> #include <linux/pagemap.h> #include <linux/rmap.h> #include <linux/spinlock.h> #include <linux/jhash.h> #include <linux/delay.h> #include <linux/kthread.h> #include <linux/wait.h> #include <linux/slab.h> #include <linux/rbtree.h> |
62b61f611
|
32 |
#include <linux/memory.h> |
31dbd01f3
|
33 |
#include <linux/mmu_notifier.h> |
2c6854fda
|
34 |
#include <linux/swap.h> |
f8af4da3b
|
35 |
#include <linux/ksm.h> |
d9f8984c2
|
36 |
#include <linux/hash.h> |
878aee7d6
|
37 |
#include <linux/freezer.h> |
f8af4da3b
|
38 |
|
31dbd01f3
|
39 |
#include <asm/tlbflush.h> |
73848b468
|
40 |
#include "internal.h" |
31dbd01f3
|
41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 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 84 85 |
/* * A few notes about the KSM scanning process, * to make it easier to understand the data structures below: * * In order to reduce excessive scanning, KSM sorts the memory pages by their * contents into a data structure that holds pointers to the pages' locations. * * Since the contents of the pages may change at any moment, KSM cannot just * insert the pages into a normal sorted tree and expect it to find anything. * Therefore KSM uses two data structures - the stable and the unstable tree. * * The stable tree holds pointers to all the merged pages (ksm pages), sorted * by their contents. Because each such page is write-protected, searching on * this tree is fully assured to be working (except when pages are unmapped), * and therefore this tree is called the stable tree. * * In addition to the stable tree, KSM uses a second data structure called the * unstable tree: this tree holds pointers to pages which have been found to * be "unchanged for a period of time". The unstable tree sorts these pages * by their contents, but since they are not write-protected, KSM cannot rely * upon the unstable tree to work correctly - the unstable tree is liable to * be corrupted as its contents are modified, and so it is called unstable. * * KSM solves this problem by several techniques: * * 1) The unstable tree is flushed every time KSM completes scanning all * memory areas, and then the tree is rebuilt again from the beginning. * 2) KSM will only insert into the unstable tree, pages whose hash value * has not changed since the previous scan of all memory areas. * 3) The unstable tree is a RedBlack Tree - so its balancing is based on the * colors of the nodes and not on their contents, assuring that even when * the tree gets "corrupted" it won't get out of balance, so scanning time * remains the same (also, searching and inserting nodes in an rbtree uses * the same algorithm, so we have no overhead when we flush and rebuild). * 4) KSM never flushes the stable tree, which means that even if it were to * take 10 attempts to find a page in the unstable tree, once it is found, * it is secured in the stable tree. (When we scan a new page, we first * compare it against the stable tree, and then against the unstable tree.) */ /** * struct mm_slot - ksm information per mm that is being scanned * @link: link to the mm_slots hash list * @mm_list: link into the mm_slots list, rooted in ksm_mm_head |
6514d511d
|
86 |
* @rmap_list: head for this mm_slot's singly-linked list of rmap_items |
31dbd01f3
|
87 88 89 90 91 |
* @mm: the mm that this information is valid for */ struct mm_slot { struct hlist_node link; struct list_head mm_list; |
6514d511d
|
92 |
struct rmap_item *rmap_list; |
31dbd01f3
|
93 94 95 96 97 98 99 |
struct mm_struct *mm; }; /** * struct ksm_scan - cursor for scanning * @mm_slot: the current mm_slot we are scanning * @address: the next address inside that to be scanned |
6514d511d
|
100 |
* @rmap_list: link to the next rmap to be scanned in the rmap_list |
31dbd01f3
|
101 102 103 104 105 106 107 |
* @seqnr: count of completed full scans (needed when removing unstable node) * * There is only the one ksm_scan instance of this cursor structure. */ struct ksm_scan { struct mm_slot *mm_slot; unsigned long address; |
6514d511d
|
108 |
struct rmap_item **rmap_list; |
31dbd01f3
|
109 110 111 112 |
unsigned long seqnr; }; /** |
7b6ba2c7d
|
113 114 115 |
* struct stable_node - node of the stable rbtree * @node: rb node of this ksm page in the stable tree * @hlist: hlist head of rmap_items using this ksm page |
62b61f611
|
116 |
* @kpfn: page frame number of this ksm page |
7b6ba2c7d
|
117 118 119 120 |
*/ struct stable_node { struct rb_node node; struct hlist_head hlist; |
62b61f611
|
121 |
unsigned long kpfn; |
7b6ba2c7d
|
122 123 124 |
}; /** |
31dbd01f3
|
125 |
* struct rmap_item - reverse mapping item for virtual addresses |
6514d511d
|
126 |
* @rmap_list: next rmap_item in mm_slot's singly-linked rmap_list |
db114b83a
|
127 |
* @anon_vma: pointer to anon_vma for this mm,address, when in stable tree |
31dbd01f3
|
128 129 130 |
* @mm: the memory structure this rmap_item is pointing into * @address: the virtual address this rmap_item tracks (+ flags in low bits) * @oldchecksum: previous checksum of the page at that virtual address |
7b6ba2c7d
|
131 132 133 |
* @node: rb node of this rmap_item in the unstable tree * @head: pointer to stable_node heading this list in the stable tree * @hlist: link into hlist of rmap_items hanging off that stable_node |
31dbd01f3
|
134 135 |
*/ struct rmap_item { |
6514d511d
|
136 |
struct rmap_item *rmap_list; |
db114b83a
|
137 |
struct anon_vma *anon_vma; /* when stable */ |
31dbd01f3
|
138 139 |
struct mm_struct *mm; unsigned long address; /* + low bits used for flags below */ |
7b6ba2c7d
|
140 |
unsigned int oldchecksum; /* when unstable */ |
31dbd01f3
|
141 |
union { |
7b6ba2c7d
|
142 143 144 145 146 |
struct rb_node node; /* when node of unstable tree */ struct { /* when listed from stable tree */ struct stable_node *head; struct hlist_node hlist; }; |
31dbd01f3
|
147 148 149 150 |
}; }; #define SEQNR_MASK 0x0ff /* low bits of unstable tree seqnr */ |
7b6ba2c7d
|
151 152 |
#define UNSTABLE_FLAG 0x100 /* is a node of the unstable tree */ #define STABLE_FLAG 0x200 /* is listed from the stable tree */ |
31dbd01f3
|
153 154 155 156 |
/* The stable and unstable tree heads */ static struct rb_root root_stable_tree = RB_ROOT; static struct rb_root root_unstable_tree = RB_ROOT; |
d9f8984c2
|
157 158 159 |
#define MM_SLOTS_HASH_SHIFT 10 #define MM_SLOTS_HASH_HEADS (1 << MM_SLOTS_HASH_SHIFT) static struct hlist_head mm_slots_hash[MM_SLOTS_HASH_HEADS]; |
31dbd01f3
|
160 161 162 163 164 165 166 167 168 |
static struct mm_slot ksm_mm_head = { .mm_list = LIST_HEAD_INIT(ksm_mm_head.mm_list), }; static struct ksm_scan ksm_scan = { .mm_slot = &ksm_mm_head, }; static struct kmem_cache *rmap_item_cache; |
7b6ba2c7d
|
169 |
static struct kmem_cache *stable_node_cache; |
31dbd01f3
|
170 171 172 |
static struct kmem_cache *mm_slot_cache; /* The number of nodes in the stable tree */ |
b40282603
|
173 |
static unsigned long ksm_pages_shared; |
31dbd01f3
|
174 |
|
e178dfde3
|
175 |
/* The number of page slots additionally sharing those nodes */ |
b40282603
|
176 |
static unsigned long ksm_pages_sharing; |
31dbd01f3
|
177 |
|
473b0ce4d
|
178 179 180 181 182 |
/* The number of nodes in the unstable tree */ static unsigned long ksm_pages_unshared; /* The number of rmap_items in use: to calculate pages_volatile */ static unsigned long ksm_rmap_items; |
31dbd01f3
|
183 |
/* Number of pages ksmd should scan in one batch */ |
2c6854fda
|
184 |
static unsigned int ksm_thread_pages_to_scan = 100; |
31dbd01f3
|
185 186 |
/* Milliseconds ksmd should sleep between batches */ |
2ffd8679c
|
187 |
static unsigned int ksm_thread_sleep_millisecs = 20; |
31dbd01f3
|
188 189 190 191 |
#define KSM_RUN_STOP 0 #define KSM_RUN_MERGE 1 #define KSM_RUN_UNMERGE 2 |
2c6854fda
|
192 |
static unsigned int ksm_run = KSM_RUN_STOP; |
31dbd01f3
|
193 194 195 196 197 198 199 200 201 202 203 204 205 206 |
static DECLARE_WAIT_QUEUE_HEAD(ksm_thread_wait); static DEFINE_MUTEX(ksm_thread_mutex); static DEFINE_SPINLOCK(ksm_mmlist_lock); #define KSM_KMEM_CACHE(__struct, __flags) kmem_cache_create("ksm_"#__struct,\ sizeof(struct __struct), __alignof__(struct __struct),\ (__flags), NULL) static int __init ksm_slab_init(void) { rmap_item_cache = KSM_KMEM_CACHE(rmap_item, 0); if (!rmap_item_cache) goto out; |
7b6ba2c7d
|
207 208 209 |
stable_node_cache = KSM_KMEM_CACHE(stable_node, 0); if (!stable_node_cache) goto out_free1; |
31dbd01f3
|
210 211 |
mm_slot_cache = KSM_KMEM_CACHE(mm_slot, 0); if (!mm_slot_cache) |
7b6ba2c7d
|
212 |
goto out_free2; |
31dbd01f3
|
213 214 |
return 0; |
7b6ba2c7d
|
215 216 217 |
out_free2: kmem_cache_destroy(stable_node_cache); out_free1: |
31dbd01f3
|
218 219 220 221 222 223 224 225 |
kmem_cache_destroy(rmap_item_cache); out: return -ENOMEM; } static void __init ksm_slab_free(void) { kmem_cache_destroy(mm_slot_cache); |
7b6ba2c7d
|
226 |
kmem_cache_destroy(stable_node_cache); |
31dbd01f3
|
227 228 229 230 231 232 |
kmem_cache_destroy(rmap_item_cache); mm_slot_cache = NULL; } static inline struct rmap_item *alloc_rmap_item(void) { |
473b0ce4d
|
233 234 235 236 237 238 |
struct rmap_item *rmap_item; rmap_item = kmem_cache_zalloc(rmap_item_cache, GFP_KERNEL); if (rmap_item) ksm_rmap_items++; return rmap_item; |
31dbd01f3
|
239 240 241 242 |
} static inline void free_rmap_item(struct rmap_item *rmap_item) { |
473b0ce4d
|
243 |
ksm_rmap_items--; |
31dbd01f3
|
244 245 246 |
rmap_item->mm = NULL; /* debug safety */ kmem_cache_free(rmap_item_cache, rmap_item); } |
7b6ba2c7d
|
247 248 249 250 251 252 253 254 255 |
static inline struct stable_node *alloc_stable_node(void) { return kmem_cache_alloc(stable_node_cache, GFP_KERNEL); } static inline void free_stable_node(struct stable_node *stable_node) { kmem_cache_free(stable_node_cache, stable_node); } |
31dbd01f3
|
256 257 258 259 260 261 262 263 264 265 266 |
static inline struct mm_slot *alloc_mm_slot(void) { if (!mm_slot_cache) /* initialization failed */ return NULL; return kmem_cache_zalloc(mm_slot_cache, GFP_KERNEL); } static inline void free_mm_slot(struct mm_slot *mm_slot) { kmem_cache_free(mm_slot_cache, mm_slot); } |
31dbd01f3
|
267 268 269 270 271 |
static struct mm_slot *get_mm_slot(struct mm_struct *mm) { struct mm_slot *mm_slot; struct hlist_head *bucket; struct hlist_node *node; |
d9f8984c2
|
272 |
bucket = &mm_slots_hash[hash_ptr(mm, MM_SLOTS_HASH_SHIFT)]; |
31dbd01f3
|
273 274 275 276 277 278 279 280 281 282 283 |
hlist_for_each_entry(mm_slot, node, bucket, link) { if (mm == mm_slot->mm) return mm_slot; } return NULL; } static void insert_to_mm_slots_hash(struct mm_struct *mm, struct mm_slot *mm_slot) { struct hlist_head *bucket; |
d9f8984c2
|
284 |
bucket = &mm_slots_hash[hash_ptr(mm, MM_SLOTS_HASH_SHIFT)]; |
31dbd01f3
|
285 |
mm_slot->mm = mm; |
31dbd01f3
|
286 287 288 289 290 291 292 293 294 |
hlist_add_head(&mm_slot->link, bucket); } static inline int in_stable_tree(struct rmap_item *rmap_item) { return rmap_item->address & STABLE_FLAG; } /* |
a913e182a
|
295 296 297 298 299 300 301 302 303 304 305 306 307 |
* ksmd, and unmerge_and_remove_all_rmap_items(), must not touch an mm's * page tables after it has passed through ksm_exit() - which, if necessary, * takes mmap_sem briefly to serialize against them. ksm_exit() does not set * a special flag: they can just back out as soon as mm_users goes to zero. * ksm_test_exit() is used throughout to make this test for exit: in some * places for correctness, in some places just to avoid unnecessary work. */ static inline bool ksm_test_exit(struct mm_struct *mm) { return atomic_read(&mm->mm_users) == 0; } /* |
31dbd01f3
|
308 309 310 311 312 313 314 315 316 317 |
* We use break_ksm to break COW on a ksm page: it's a stripped down * * if (get_user_pages(current, mm, addr, 1, 1, 1, &page, NULL) == 1) * put_page(page); * * but taking great care only to touch a ksm page, in a VM_MERGEABLE vma, * in case the application has unmapped and remapped mm,addr meanwhile. * Could a ksm page appear anywhere else? Actually yes, in a VM_PFNMAP * mmap of /dev/mem or /dev/kmem, where we would not want to touch it. */ |
d952b7913
|
318 |
static int break_ksm(struct vm_area_struct *vma, unsigned long addr) |
31dbd01f3
|
319 320 |
{ struct page *page; |
d952b7913
|
321 |
int ret = 0; |
31dbd01f3
|
322 323 324 325 |
do { cond_resched(); page = follow_page(vma, addr, FOLL_GET); |
22eccdd7d
|
326 |
if (IS_ERR_OR_NULL(page)) |
31dbd01f3
|
327 328 329 330 331 332 333 |
break; if (PageKsm(page)) ret = handle_mm_fault(vma->vm_mm, vma, addr, FAULT_FLAG_WRITE); else ret = VM_FAULT_WRITE; put_page(page); |
d952b7913
|
334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 |
} while (!(ret & (VM_FAULT_WRITE | VM_FAULT_SIGBUS | VM_FAULT_OOM))); /* * We must loop because handle_mm_fault() may back out if there's * any difficulty e.g. if pte accessed bit gets updated concurrently. * * VM_FAULT_WRITE is what we have been hoping for: it indicates that * COW has been broken, even if the vma does not permit VM_WRITE; * but note that a concurrent fault might break PageKsm for us. * * VM_FAULT_SIGBUS could occur if we race with truncation of the * backing file, which also invalidates anonymous pages: that's * okay, that truncation will have unmapped the PageKsm for us. * * VM_FAULT_OOM: at the time of writing (late July 2009), setting * aside mem_cgroup limits, VM_FAULT_OOM would only be set if the * current task has TIF_MEMDIE set, and will be OOM killed on return * to user; and ksmd, having no mm, would never be chosen for that. * * But if the mm is in a limited mem_cgroup, then the fault may fail * with VM_FAULT_OOM even if the current task is not TIF_MEMDIE; and * even ksmd can fail in this way - though it's usually breaking ksm * just to undo a merge it made a moment before, so unlikely to oom. * * That's a pity: we might therefore have more kernel pages allocated * than we're counting as nodes in the stable tree; but ksm_do_scan * will retry to break_cow on each pass, so should recover the page * in due course. The important thing is to not let VM_MERGEABLE * be cleared while any such pages might remain in the area. */ return (ret & VM_FAULT_OOM) ? -ENOMEM : 0; |
31dbd01f3
|
364 |
} |
8dd3557a5
|
365 |
static void break_cow(struct rmap_item *rmap_item) |
31dbd01f3
|
366 |
{ |
8dd3557a5
|
367 368 |
struct mm_struct *mm = rmap_item->mm; unsigned long addr = rmap_item->address; |
31dbd01f3
|
369 |
struct vm_area_struct *vma; |
4035c07a8
|
370 371 372 373 |
/* * It is not an accident that whenever we want to break COW * to undo, we also need to drop a reference to the anon_vma. */ |
9e60109f1
|
374 |
put_anon_vma(rmap_item->anon_vma); |
4035c07a8
|
375 |
|
81464e306
|
376 |
down_read(&mm->mmap_sem); |
9ba692948
|
377 378 |
if (ksm_test_exit(mm)) goto out; |
31dbd01f3
|
379 380 |
vma = find_vma(mm, addr); if (!vma || vma->vm_start > addr) |
81464e306
|
381 |
goto out; |
31dbd01f3
|
382 |
if (!(vma->vm_flags & VM_MERGEABLE) || !vma->anon_vma) |
81464e306
|
383 |
goto out; |
31dbd01f3
|
384 |
break_ksm(vma, addr); |
81464e306
|
385 |
out: |
31dbd01f3
|
386 387 |
up_read(&mm->mmap_sem); } |
29ad768cf
|
388 389 390 |
static struct page *page_trans_compound_anon(struct page *page) { if (PageTransCompound(page)) { |
22e5c47ee
|
391 |
struct page *head = compound_trans_head(page); |
29ad768cf
|
392 |
/* |
22e5c47ee
|
393 394 |
* head may actually be splitted and freed from under * us but it's ok here. |
29ad768cf
|
395 |
*/ |
29ad768cf
|
396 397 398 399 400 |
if (PageAnon(head)) return head; } return NULL; } |
31dbd01f3
|
401 402 403 404 405 406 407 408 |
static struct page *get_mergeable_page(struct rmap_item *rmap_item) { struct mm_struct *mm = rmap_item->mm; unsigned long addr = rmap_item->address; struct vm_area_struct *vma; struct page *page; down_read(&mm->mmap_sem); |
9ba692948
|
409 410 |
if (ksm_test_exit(mm)) goto out; |
31dbd01f3
|
411 412 413 414 415 416 417 |
vma = find_vma(mm, addr); if (!vma || vma->vm_start > addr) goto out; if (!(vma->vm_flags & VM_MERGEABLE) || !vma->anon_vma) goto out; page = follow_page(vma, addr, FOLL_GET); |
22eccdd7d
|
418 |
if (IS_ERR_OR_NULL(page)) |
31dbd01f3
|
419 |
goto out; |
29ad768cf
|
420 |
if (PageAnon(page) || page_trans_compound_anon(page)) { |
31dbd01f3
|
421 422 423 424 425 426 427 428 429 |
flush_anon_page(vma, page, addr); flush_dcache_page(page); } else { put_page(page); out: page = NULL; } up_read(&mm->mmap_sem); return page; } |
4035c07a8
|
430 431 432 433 434 435 436 437 438 439 |
static void remove_node_from_stable_tree(struct stable_node *stable_node) { struct rmap_item *rmap_item; struct hlist_node *hlist; hlist_for_each_entry(rmap_item, hlist, &stable_node->hlist, hlist) { if (rmap_item->hlist.next) ksm_pages_sharing--; else ksm_pages_shared--; |
9e60109f1
|
440 |
put_anon_vma(rmap_item->anon_vma); |
4035c07a8
|
441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 |
rmap_item->address &= PAGE_MASK; cond_resched(); } rb_erase(&stable_node->node, &root_stable_tree); free_stable_node(stable_node); } /* * get_ksm_page: checks if the page indicated by the stable node * is still its ksm page, despite having held no reference to it. * In which case we can trust the content of the page, and it * returns the gotten page; but if the page has now been zapped, * remove the stale node from the stable tree and return NULL. * * You would expect the stable_node to hold a reference to the ksm page. * But if it increments the page's count, swapping out has to wait for * ksmd to come around again before it can free the page, which may take * seconds or even minutes: much too unresponsive. So instead we use a * "keyhole reference": access to the ksm page from the stable node peeps * out through its keyhole to see if that page still holds the right key, * pointing back to this stable node. This relies on freeing a PageAnon * page to reset its page->mapping to NULL, and relies on no other use of * a page to put something that might look like our key in page->mapping. * * include/linux/pagemap.h page_cache_get_speculative() is a good reference, * but this is different - made simpler by ksm_thread_mutex being held, but * interesting for assuming that no other use of the struct page could ever * put our expected_mapping into page->mapping (or a field of the union which * coincides with page->mapping). The RCU calls are not for KSM at all, but * to keep the page_count protocol described with page_cache_get_speculative. * * Note: it is possible that get_ksm_page() will return NULL one moment, * then page the next, if the page is in between page_freeze_refs() and * page_unfreeze_refs(): this shouldn't be a problem anywhere, the page * is on its way to being freed; but it is an anomaly to bear in mind. */ static struct page *get_ksm_page(struct stable_node *stable_node) { struct page *page; void *expected_mapping; |
62b61f611
|
482 |
page = pfn_to_page(stable_node->kpfn); |
4035c07a8
|
483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 |
expected_mapping = (void *)stable_node + (PAGE_MAPPING_ANON | PAGE_MAPPING_KSM); rcu_read_lock(); if (page->mapping != expected_mapping) goto stale; if (!get_page_unless_zero(page)) goto stale; if (page->mapping != expected_mapping) { put_page(page); goto stale; } rcu_read_unlock(); return page; stale: rcu_read_unlock(); remove_node_from_stable_tree(stable_node); return NULL; } |
31dbd01f3
|
501 |
/* |
31dbd01f3
|
502 503 504 505 506 |
* Removing rmap_item from stable or unstable tree. * This function will clean the information from the stable/unstable tree. */ static void remove_rmap_item_from_tree(struct rmap_item *rmap_item) { |
7b6ba2c7d
|
507 508 |
if (rmap_item->address & STABLE_FLAG) { struct stable_node *stable_node; |
5ad646880
|
509 |
struct page *page; |
31dbd01f3
|
510 |
|
7b6ba2c7d
|
511 |
stable_node = rmap_item->head; |
4035c07a8
|
512 513 514 |
page = get_ksm_page(stable_node); if (!page) goto out; |
5ad646880
|
515 |
|
4035c07a8
|
516 |
lock_page(page); |
7b6ba2c7d
|
517 |
hlist_del(&rmap_item->hlist); |
4035c07a8
|
518 519 |
unlock_page(page); put_page(page); |
08beca44d
|
520 |
|
4035c07a8
|
521 522 523 |
if (stable_node->hlist.first) ksm_pages_sharing--; else |
7b6ba2c7d
|
524 |
ksm_pages_shared--; |
31dbd01f3
|
525 |
|
9e60109f1
|
526 |
put_anon_vma(rmap_item->anon_vma); |
93d17715a
|
527 |
rmap_item->address &= PAGE_MASK; |
31dbd01f3
|
528 |
|
7b6ba2c7d
|
529 |
} else if (rmap_item->address & UNSTABLE_FLAG) { |
31dbd01f3
|
530 531 |
unsigned char age; /* |
9ba692948
|
532 |
* Usually ksmd can and must skip the rb_erase, because |
31dbd01f3
|
533 |
* root_unstable_tree was already reset to RB_ROOT. |
9ba692948
|
534 535 536 |
* But be careful when an mm is exiting: do the rb_erase * if this rmap_item was inserted by this scan, rather * than left over from before. |
31dbd01f3
|
537 538 |
*/ age = (unsigned char)(ksm_scan.seqnr - rmap_item->address); |
cd551f975
|
539 |
BUG_ON(age > 1); |
31dbd01f3
|
540 541 |
if (!age) rb_erase(&rmap_item->node, &root_unstable_tree); |
93d17715a
|
542 |
|
473b0ce4d
|
543 |
ksm_pages_unshared--; |
93d17715a
|
544 |
rmap_item->address &= PAGE_MASK; |
31dbd01f3
|
545 |
} |
4035c07a8
|
546 |
out: |
31dbd01f3
|
547 548 |
cond_resched(); /* we're called from many long loops */ } |
31dbd01f3
|
549 |
static void remove_trailing_rmap_items(struct mm_slot *mm_slot, |
6514d511d
|
550 |
struct rmap_item **rmap_list) |
31dbd01f3
|
551 |
{ |
6514d511d
|
552 553 554 |
while (*rmap_list) { struct rmap_item *rmap_item = *rmap_list; *rmap_list = rmap_item->rmap_list; |
31dbd01f3
|
555 |
remove_rmap_item_from_tree(rmap_item); |
31dbd01f3
|
556 557 558 559 560 561 562 563 564 565 566 |
free_rmap_item(rmap_item); } } /* * Though it's very tempting to unmerge in_stable_tree(rmap_item)s rather * than check every pte of a given vma, the locking doesn't quite work for * that - an rmap_item is assigned to the stable tree after inserting ksm * page and upping mmap_sem. Nor does it fit with the way we skip dup'ing * rmap_items from parent to child at fork time (so as not to waste time * if exit comes before the next scan reaches it). |
81464e306
|
567 568 569 570 571 |
* * Similarly, although we'd like to remove rmap_items (so updating counts * and freeing memory) when unmerging an area, it's easier to leave that * to the next pass of ksmd - consider, for example, how ksmd might be * in cmp_and_merge_page on one of the rmap_items we would be removing. |
31dbd01f3
|
572 |
*/ |
d952b7913
|
573 574 |
static int unmerge_ksm_pages(struct vm_area_struct *vma, unsigned long start, unsigned long end) |
31dbd01f3
|
575 576 |
{ unsigned long addr; |
d952b7913
|
577 |
int err = 0; |
31dbd01f3
|
578 |
|
d952b7913
|
579 |
for (addr = start; addr < end && !err; addr += PAGE_SIZE) { |
9ba692948
|
580 581 |
if (ksm_test_exit(vma->vm_mm)) break; |
d952b7913
|
582 583 584 585 586 587 |
if (signal_pending(current)) err = -ERESTARTSYS; else err = break_ksm(vma, addr); } return err; |
31dbd01f3
|
588 |
} |
2ffd8679c
|
589 590 591 592 |
#ifdef CONFIG_SYSFS /* * Only called through the sysfs control interface: */ |
d952b7913
|
593 |
static int unmerge_and_remove_all_rmap_items(void) |
31dbd01f3
|
594 595 596 597 |
{ struct mm_slot *mm_slot; struct mm_struct *mm; struct vm_area_struct *vma; |
d952b7913
|
598 599 600 |
int err = 0; spin_lock(&ksm_mmlist_lock); |
9ba692948
|
601 |
ksm_scan.mm_slot = list_entry(ksm_mm_head.mm_list.next, |
d952b7913
|
602 603 |
struct mm_slot, mm_list); spin_unlock(&ksm_mmlist_lock); |
31dbd01f3
|
604 |
|
9ba692948
|
605 606 |
for (mm_slot = ksm_scan.mm_slot; mm_slot != &ksm_mm_head; mm_slot = ksm_scan.mm_slot) { |
31dbd01f3
|
607 608 609 |
mm = mm_slot->mm; down_read(&mm->mmap_sem); for (vma = mm->mmap; vma; vma = vma->vm_next) { |
9ba692948
|
610 611 |
if (ksm_test_exit(mm)) break; |
31dbd01f3
|
612 613 |
if (!(vma->vm_flags & VM_MERGEABLE) || !vma->anon_vma) continue; |
d952b7913
|
614 615 |
err = unmerge_ksm_pages(vma, vma->vm_start, vma->vm_end); |
9ba692948
|
616 617 |
if (err) goto error; |
31dbd01f3
|
618 |
} |
9ba692948
|
619 |
|
6514d511d
|
620 |
remove_trailing_rmap_items(mm_slot, &mm_slot->rmap_list); |
d952b7913
|
621 622 |
spin_lock(&ksm_mmlist_lock); |
9ba692948
|
623 |
ksm_scan.mm_slot = list_entry(mm_slot->mm_list.next, |
d952b7913
|
624 |
struct mm_slot, mm_list); |
9ba692948
|
625 626 627 628 629 630 631 632 633 634 635 636 637 |
if (ksm_test_exit(mm)) { hlist_del(&mm_slot->link); list_del(&mm_slot->mm_list); spin_unlock(&ksm_mmlist_lock); free_mm_slot(mm_slot); clear_bit(MMF_VM_MERGEABLE, &mm->flags); up_read(&mm->mmap_sem); mmdrop(mm); } else { spin_unlock(&ksm_mmlist_lock); up_read(&mm->mmap_sem); } |
31dbd01f3
|
638 |
} |
d952b7913
|
639 |
ksm_scan.seqnr = 0; |
9ba692948
|
640 641 642 643 |
return 0; error: up_read(&mm->mmap_sem); |
31dbd01f3
|
644 |
spin_lock(&ksm_mmlist_lock); |
d952b7913
|
645 |
ksm_scan.mm_slot = &ksm_mm_head; |
31dbd01f3
|
646 |
spin_unlock(&ksm_mmlist_lock); |
d952b7913
|
647 |
return err; |
31dbd01f3
|
648 |
} |
2ffd8679c
|
649 |
#endif /* CONFIG_SYSFS */ |
31dbd01f3
|
650 |
|
31dbd01f3
|
651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 |
static u32 calc_checksum(struct page *page) { u32 checksum; void *addr = kmap_atomic(page, KM_USER0); checksum = jhash2(addr, PAGE_SIZE / 4, 17); kunmap_atomic(addr, KM_USER0); return checksum; } static int memcmp_pages(struct page *page1, struct page *page2) { char *addr1, *addr2; int ret; addr1 = kmap_atomic(page1, KM_USER0); addr2 = kmap_atomic(page2, KM_USER1); ret = memcmp(addr1, addr2, PAGE_SIZE); kunmap_atomic(addr2, KM_USER1); kunmap_atomic(addr1, KM_USER0); return ret; } static inline int pages_identical(struct page *page1, struct page *page2) { return !memcmp_pages(page1, page2); } static int write_protect_page(struct vm_area_struct *vma, struct page *page, pte_t *orig_pte) { struct mm_struct *mm = vma->vm_mm; unsigned long addr; pte_t *ptep; spinlock_t *ptl; int swapped; int err = -EFAULT; addr = page_address_in_vma(page, vma); if (addr == -EFAULT) goto out; |
29ad768cf
|
691 |
BUG_ON(PageTransCompound(page)); |
31dbd01f3
|
692 693 694 |
ptep = page_check_address(page, mm, addr, &ptl, 0); if (!ptep) goto out; |
4e31635c3
|
695 |
if (pte_write(*ptep) || pte_dirty(*ptep)) { |
31dbd01f3
|
696 697 698 699 700 |
pte_t entry; swapped = PageSwapCache(page); flush_cache_page(vma, addr, page_to_pfn(page)); /* |
25985edce
|
701 |
* Ok this is tricky, when get_user_pages_fast() run it doesn't |
31dbd01f3
|
702 703 704 705 706 707 708 709 710 711 712 713 |
* take any lock, therefore the check that we are going to make * with the pagecount against the mapcount is racey and * O_DIRECT can happen right after the check. * So we clear the pte and flush the tlb before the check * this assure us that no O_DIRECT can happen after the check * or in the middle of the check. */ entry = ptep_clear_flush(vma, addr, ptep); /* * Check that no O_DIRECT or similar I/O is in progress on the * page */ |
31e855ea7
|
714 |
if (page_mapcount(page) + 1 + swapped != page_count(page)) { |
cb5323751
|
715 |
set_pte_at(mm, addr, ptep, entry); |
31dbd01f3
|
716 717 |
goto out_unlock; } |
4e31635c3
|
718 719 720 |
if (pte_dirty(entry)) set_page_dirty(page); entry = pte_mkclean(pte_wrprotect(entry)); |
31dbd01f3
|
721 722 723 724 725 726 727 728 729 730 731 732 733 |
set_pte_at_notify(mm, addr, ptep, entry); } *orig_pte = *ptep; err = 0; out_unlock: pte_unmap_unlock(ptep, ptl); out: return err; } /** * replace_page - replace page in vma by new ksm page |
8dd3557a5
|
734 735 736 |
* @vma: vma that holds the pte pointing to page * @page: the page we are replacing by kpage * @kpage: the ksm page we replace page by |
31dbd01f3
|
737 738 739 740 |
* @orig_pte: the original value of the pte * * Returns 0 on success, -EFAULT on failure. */ |
8dd3557a5
|
741 742 |
static int replace_page(struct vm_area_struct *vma, struct page *page, struct page *kpage, pte_t orig_pte) |
31dbd01f3
|
743 744 745 746 747 748 749 750 |
{ struct mm_struct *mm = vma->vm_mm; pgd_t *pgd; pud_t *pud; pmd_t *pmd; pte_t *ptep; spinlock_t *ptl; unsigned long addr; |
31dbd01f3
|
751 |
int err = -EFAULT; |
8dd3557a5
|
752 |
addr = page_address_in_vma(page, vma); |
31dbd01f3
|
753 754 755 756 757 758 759 760 761 762 763 764 |
if (addr == -EFAULT) goto out; pgd = pgd_offset(mm, addr); if (!pgd_present(*pgd)) goto out; pud = pud_offset(pgd, addr); if (!pud_present(*pud)) goto out; pmd = pmd_offset(pud, addr); |
29ad768cf
|
765 |
BUG_ON(pmd_trans_huge(*pmd)); |
31dbd01f3
|
766 767 768 769 770 771 772 773 |
if (!pmd_present(*pmd)) goto out; ptep = pte_offset_map_lock(mm, pmd, addr, &ptl); if (!pte_same(*ptep, orig_pte)) { pte_unmap_unlock(ptep, ptl); goto out; } |
8dd3557a5
|
774 |
get_page(kpage); |
5ad646880
|
775 |
page_add_anon_rmap(kpage, vma, addr); |
31dbd01f3
|
776 777 778 |
flush_cache_page(vma, addr, pte_pfn(*ptep)); ptep_clear_flush(vma, addr, ptep); |
8dd3557a5
|
779 |
set_pte_at_notify(mm, addr, ptep, mk_pte(kpage, vma->vm_page_prot)); |
31dbd01f3
|
780 |
|
8dd3557a5
|
781 |
page_remove_rmap(page); |
ae52a2adb
|
782 783 |
if (!page_mapped(page)) try_to_free_swap(page); |
8dd3557a5
|
784 |
put_page(page); |
31dbd01f3
|
785 786 787 788 789 790 |
pte_unmap_unlock(ptep, ptl); err = 0; out: return err; } |
29ad768cf
|
791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 |
static int page_trans_compound_anon_split(struct page *page) { int ret = 0; struct page *transhuge_head = page_trans_compound_anon(page); if (transhuge_head) { /* Get the reference on the head to split it. */ if (get_page_unless_zero(transhuge_head)) { /* * Recheck we got the reference while the head * was still anonymous. */ if (PageAnon(transhuge_head)) ret = split_huge_page(transhuge_head); else /* * Retry later if split_huge_page run * from under us. */ ret = 1; put_page(transhuge_head); } else /* Retry later if split_huge_page run from under us. */ ret = 1; } return ret; } |
31dbd01f3
|
817 818 |
/* * try_to_merge_one_page - take two pages and merge them into one |
8dd3557a5
|
819 820 |
* @vma: the vma that holds the pte pointing to page * @page: the PageAnon page that we want to replace with kpage |
80e148226
|
821 822 |
* @kpage: the PageKsm page that we want to map instead of page, * or NULL the first time when we want to use page as kpage. |
31dbd01f3
|
823 824 825 826 |
* * This function returns 0 if the pages were merged, -EFAULT otherwise. */ static int try_to_merge_one_page(struct vm_area_struct *vma, |
8dd3557a5
|
827 |
struct page *page, struct page *kpage) |
31dbd01f3
|
828 829 830 |
{ pte_t orig_pte = __pte(0); int err = -EFAULT; |
db114b83a
|
831 832 |
if (page == kpage) /* ksm page forked */ return 0; |
31dbd01f3
|
833 834 |
if (!(vma->vm_flags & VM_MERGEABLE)) goto out; |
29ad768cf
|
835 836 837 |
if (PageTransCompound(page) && page_trans_compound_anon_split(page)) goto out; BUG_ON(PageTransCompound(page)); |
8dd3557a5
|
838 |
if (!PageAnon(page)) |
31dbd01f3
|
839 |
goto out; |
31dbd01f3
|
840 841 842 843 844 845 846 |
/* * We need the page lock to read a stable PageSwapCache in * write_protect_page(). We use trylock_page() instead of * lock_page() because we don't want to wait here - we * prefer to continue scanning and merging different pages, * then come back to this page when it is unlocked. */ |
8dd3557a5
|
847 |
if (!trylock_page(page)) |
31e855ea7
|
848 |
goto out; |
31dbd01f3
|
849 850 851 852 853 854 |
/* * If this anonymous page is mapped only here, its pte may need * to be write-protected. If it's mapped elsewhere, all of its * ptes are necessarily already write-protected. But in either * case, we need to lock and check page_count is not raised. */ |
80e148226
|
855 856 857 858 859 860 861 862 863 864 865 866 867 |
if (write_protect_page(vma, page, &orig_pte) == 0) { if (!kpage) { /* * While we hold page lock, upgrade page from * PageAnon+anon_vma to PageKsm+NULL stable_node: * stable_tree_insert() will update stable_node. */ set_page_stable_node(page, NULL); mark_page_accessed(page); err = 0; } else if (pages_identical(page, kpage)) err = replace_page(vma, page, kpage, orig_pte); } |
31dbd01f3
|
868 |
|
80e148226
|
869 |
if ((vma->vm_flags & VM_LOCKED) && kpage && !err) { |
73848b468
|
870 |
munlock_vma_page(page); |
5ad646880
|
871 872 |
if (!PageMlocked(kpage)) { unlock_page(page); |
5ad646880
|
873 874 875 876 877 |
lock_page(kpage); mlock_vma_page(kpage); page = kpage; /* for final unlock */ } } |
73848b468
|
878 |
|
8dd3557a5
|
879 |
unlock_page(page); |
31dbd01f3
|
880 881 882 883 884 |
out: return err; } /* |
81464e306
|
885 886 |
* try_to_merge_with_ksm_page - like try_to_merge_two_pages, * but no new kernel page is allocated: kpage must already be a ksm page. |
8dd3557a5
|
887 888 |
* * This function returns 0 if the pages were merged, -EFAULT otherwise. |
81464e306
|
889 |
*/ |
8dd3557a5
|
890 891 |
static int try_to_merge_with_ksm_page(struct rmap_item *rmap_item, struct page *page, struct page *kpage) |
81464e306
|
892 |
{ |
8dd3557a5
|
893 |
struct mm_struct *mm = rmap_item->mm; |
81464e306
|
894 895 |
struct vm_area_struct *vma; int err = -EFAULT; |
8dd3557a5
|
896 897 |
down_read(&mm->mmap_sem); if (ksm_test_exit(mm)) |
9ba692948
|
898 |
goto out; |
8dd3557a5
|
899 900 |
vma = find_vma(mm, rmap_item->address); if (!vma || vma->vm_start > rmap_item->address) |
81464e306
|
901 |
goto out; |
8dd3557a5
|
902 |
err = try_to_merge_one_page(vma, page, kpage); |
db114b83a
|
903 904 905 906 |
if (err) goto out; /* Must get reference to anon_vma while still holding mmap_sem */ |
9e60109f1
|
907 908 |
rmap_item->anon_vma = vma->anon_vma; get_anon_vma(vma->anon_vma); |
81464e306
|
909 |
out: |
8dd3557a5
|
910 |
up_read(&mm->mmap_sem); |
81464e306
|
911 912 913 914 |
return err; } /* |
31dbd01f3
|
915 916 917 |
* try_to_merge_two_pages - take two identical pages and prepare them * to be merged into one page. * |
8dd3557a5
|
918 919 |
* This function returns the kpage if we successfully merged two identical * pages into one ksm page, NULL otherwise. |
31dbd01f3
|
920 |
* |
80e148226
|
921 |
* Note that this function upgrades page to ksm page: if one of the pages |
31dbd01f3
|
922 923 |
* is already a ksm page, try_to_merge_with_ksm_page should be used. */ |
8dd3557a5
|
924 925 926 927 |
static struct page *try_to_merge_two_pages(struct rmap_item *rmap_item, struct page *page, struct rmap_item *tree_rmap_item, struct page *tree_page) |
31dbd01f3
|
928 |
{ |
80e148226
|
929 |
int err; |
31dbd01f3
|
930 |
|
80e148226
|
931 |
err = try_to_merge_with_ksm_page(rmap_item, page, NULL); |
31dbd01f3
|
932 |
if (!err) { |
8dd3557a5
|
933 |
err = try_to_merge_with_ksm_page(tree_rmap_item, |
80e148226
|
934 |
tree_page, page); |
31dbd01f3
|
935 |
/* |
81464e306
|
936 937 |
* If that fails, we have a ksm page with only one pte * pointing to it: so break it. |
31dbd01f3
|
938 |
*/ |
4035c07a8
|
939 |
if (err) |
8dd3557a5
|
940 |
break_cow(rmap_item); |
31dbd01f3
|
941 |
} |
80e148226
|
942 |
return err ? NULL : page; |
31dbd01f3
|
943 944 945 |
} /* |
8dd3557a5
|
946 |
* stable_tree_search - search for page inside the stable tree |
31dbd01f3
|
947 948 949 950 |
* * This function checks if there is a page inside the stable tree * with identical content to the page that we are scanning right now. * |
7b6ba2c7d
|
951 |
* This function returns the stable tree node of identical content if found, |
31dbd01f3
|
952 953 |
* NULL otherwise. */ |
62b61f611
|
954 |
static struct page *stable_tree_search(struct page *page) |
31dbd01f3
|
955 956 |
{ struct rb_node *node = root_stable_tree.rb_node; |
7b6ba2c7d
|
957 |
struct stable_node *stable_node; |
31dbd01f3
|
958 |
|
08beca44d
|
959 960 961 |
stable_node = page_stable_node(page); if (stable_node) { /* ksm page forked */ get_page(page); |
62b61f611
|
962 |
return page; |
08beca44d
|
963 |
} |
31dbd01f3
|
964 |
while (node) { |
4035c07a8
|
965 |
struct page *tree_page; |
31dbd01f3
|
966 |
int ret; |
08beca44d
|
967 |
cond_resched(); |
7b6ba2c7d
|
968 |
stable_node = rb_entry(node, struct stable_node, node); |
4035c07a8
|
969 970 971 |
tree_page = get_ksm_page(stable_node); if (!tree_page) return NULL; |
31dbd01f3
|
972 |
|
4035c07a8
|
973 |
ret = memcmp_pages(page, tree_page); |
31dbd01f3
|
974 |
|
4035c07a8
|
975 976 |
if (ret < 0) { put_page(tree_page); |
31dbd01f3
|
977 |
node = node->rb_left; |
4035c07a8
|
978 979 |
} else if (ret > 0) { put_page(tree_page); |
31dbd01f3
|
980 |
node = node->rb_right; |
4035c07a8
|
981 |
} else |
62b61f611
|
982 |
return tree_page; |
31dbd01f3
|
983 984 985 986 987 988 989 990 991 |
} return NULL; } /* * stable_tree_insert - insert rmap_item pointing to new ksm page * into the stable tree. * |
7b6ba2c7d
|
992 993 |
* This function returns the stable tree node just allocated on success, * NULL otherwise. |
31dbd01f3
|
994 |
*/ |
7b6ba2c7d
|
995 |
static struct stable_node *stable_tree_insert(struct page *kpage) |
31dbd01f3
|
996 997 998 |
{ struct rb_node **new = &root_stable_tree.rb_node; struct rb_node *parent = NULL; |
7b6ba2c7d
|
999 |
struct stable_node *stable_node; |
31dbd01f3
|
1000 1001 |
while (*new) { |
4035c07a8
|
1002 |
struct page *tree_page; |
31dbd01f3
|
1003 |
int ret; |
08beca44d
|
1004 |
cond_resched(); |
7b6ba2c7d
|
1005 |
stable_node = rb_entry(*new, struct stable_node, node); |
4035c07a8
|
1006 1007 1008 |
tree_page = get_ksm_page(stable_node); if (!tree_page) return NULL; |
31dbd01f3
|
1009 |
|
4035c07a8
|
1010 1011 |
ret = memcmp_pages(kpage, tree_page); put_page(tree_page); |
31dbd01f3
|
1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 |
parent = *new; if (ret < 0) new = &parent->rb_left; else if (ret > 0) new = &parent->rb_right; else { /* * It is not a bug that stable_tree_search() didn't * find this node: because at that time our page was * not yet write-protected, so may have changed since. */ return NULL; } } |
7b6ba2c7d
|
1027 1028 1029 |
stable_node = alloc_stable_node(); if (!stable_node) return NULL; |
31dbd01f3
|
1030 |
|
7b6ba2c7d
|
1031 1032 1033 1034 |
rb_link_node(&stable_node->node, parent, new); rb_insert_color(&stable_node->node, &root_stable_tree); INIT_HLIST_HEAD(&stable_node->hlist); |
62b61f611
|
1035 |
stable_node->kpfn = page_to_pfn(kpage); |
08beca44d
|
1036 |
set_page_stable_node(kpage, stable_node); |
7b6ba2c7d
|
1037 |
return stable_node; |
31dbd01f3
|
1038 1039 1040 |
} /* |
8dd3557a5
|
1041 1042 |
* unstable_tree_search_insert - search for identical page, * else insert rmap_item into the unstable tree. |
31dbd01f3
|
1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 |
* * This function searches for a page in the unstable tree identical to the * page currently being scanned; and if no identical page is found in the * tree, we insert rmap_item as a new object into the unstable tree. * * This function returns pointer to rmap_item found to be identical * to the currently scanned page, NULL otherwise. * * This function does both searching and inserting, because they share * the same walking algorithm in an rbtree. */ |
8dd3557a5
|
1054 1055 1056 1057 |
static struct rmap_item *unstable_tree_search_insert(struct rmap_item *rmap_item, struct page *page, struct page **tree_pagep) |
31dbd01f3
|
1058 1059 1060 1061 1062 1063 |
{ struct rb_node **new = &root_unstable_tree.rb_node; struct rb_node *parent = NULL; while (*new) { struct rmap_item *tree_rmap_item; |
8dd3557a5
|
1064 |
struct page *tree_page; |
31dbd01f3
|
1065 |
int ret; |
d178f27fc
|
1066 |
cond_resched(); |
31dbd01f3
|
1067 |
tree_rmap_item = rb_entry(*new, struct rmap_item, node); |
8dd3557a5
|
1068 |
tree_page = get_mergeable_page(tree_rmap_item); |
22eccdd7d
|
1069 |
if (IS_ERR_OR_NULL(tree_page)) |
31dbd01f3
|
1070 1071 1072 |
return NULL; /* |
8dd3557a5
|
1073 |
* Don't substitute a ksm page for a forked page. |
31dbd01f3
|
1074 |
*/ |
8dd3557a5
|
1075 1076 |
if (page == tree_page) { put_page(tree_page); |
31dbd01f3
|
1077 1078 |
return NULL; } |
8dd3557a5
|
1079 |
ret = memcmp_pages(page, tree_page); |
31dbd01f3
|
1080 1081 1082 |
parent = *new; if (ret < 0) { |
8dd3557a5
|
1083 |
put_page(tree_page); |
31dbd01f3
|
1084 1085 |
new = &parent->rb_left; } else if (ret > 0) { |
8dd3557a5
|
1086 |
put_page(tree_page); |
31dbd01f3
|
1087 1088 |
new = &parent->rb_right; } else { |
8dd3557a5
|
1089 |
*tree_pagep = tree_page; |
31dbd01f3
|
1090 1091 1092 |
return tree_rmap_item; } } |
7b6ba2c7d
|
1093 |
rmap_item->address |= UNSTABLE_FLAG; |
31dbd01f3
|
1094 1095 1096 |
rmap_item->address |= (ksm_scan.seqnr & SEQNR_MASK); rb_link_node(&rmap_item->node, parent, new); rb_insert_color(&rmap_item->node, &root_unstable_tree); |
473b0ce4d
|
1097 |
ksm_pages_unshared++; |
31dbd01f3
|
1098 1099 1100 1101 1102 1103 1104 1105 1106 |
return NULL; } /* * stable_tree_append - add another rmap_item to the linked list of * rmap_items hanging off a given node of the stable tree, all sharing * the same ksm page. */ static void stable_tree_append(struct rmap_item *rmap_item, |
7b6ba2c7d
|
1107 |
struct stable_node *stable_node) |
31dbd01f3
|
1108 |
{ |
7b6ba2c7d
|
1109 |
rmap_item->head = stable_node; |
31dbd01f3
|
1110 |
rmap_item->address |= STABLE_FLAG; |
7b6ba2c7d
|
1111 |
hlist_add_head(&rmap_item->hlist, &stable_node->hlist); |
e178dfde3
|
1112 |
|
7b6ba2c7d
|
1113 1114 1115 1116 |
if (rmap_item->hlist.next) ksm_pages_sharing++; else ksm_pages_shared++; |
31dbd01f3
|
1117 1118 1119 |
} /* |
81464e306
|
1120 1121 1122 1123 |
* cmp_and_merge_page - first see if page can be merged into the stable tree; * if not, compare checksum to previous and if it's the same, see if page can * be inserted into the unstable tree, or merged with a page already there and * both transferred to the stable tree. |
31dbd01f3
|
1124 1125 1126 1127 1128 1129 |
* * @page: the page that we are searching identical page to. * @rmap_item: the reverse mapping into the virtual address of this page */ static void cmp_and_merge_page(struct page *page, struct rmap_item *rmap_item) { |
31dbd01f3
|
1130 |
struct rmap_item *tree_rmap_item; |
8dd3557a5
|
1131 |
struct page *tree_page = NULL; |
7b6ba2c7d
|
1132 |
struct stable_node *stable_node; |
8dd3557a5
|
1133 |
struct page *kpage; |
31dbd01f3
|
1134 1135 |
unsigned int checksum; int err; |
93d17715a
|
1136 |
remove_rmap_item_from_tree(rmap_item); |
31dbd01f3
|
1137 1138 |
/* We first start with searching the page inside the stable tree */ |
62b61f611
|
1139 1140 |
kpage = stable_tree_search(page); if (kpage) { |
08beca44d
|
1141 |
err = try_to_merge_with_ksm_page(rmap_item, page, kpage); |
31dbd01f3
|
1142 1143 1144 1145 1146 |
if (!err) { /* * The page was successfully merged: * add its rmap_item to the stable tree. */ |
5ad646880
|
1147 |
lock_page(kpage); |
62b61f611
|
1148 |
stable_tree_append(rmap_item, page_stable_node(kpage)); |
5ad646880
|
1149 |
unlock_page(kpage); |
31dbd01f3
|
1150 |
} |
8dd3557a5
|
1151 |
put_page(kpage); |
31dbd01f3
|
1152 1153 1154 1155 |
return; } /* |
4035c07a8
|
1156 1157 1158 1159 |
* If the hash value of the page has changed from the last time * we calculated it, this page is changing frequently: therefore we * don't want to insert it in the unstable tree, and we don't want * to waste our time searching for something identical to it there. |
31dbd01f3
|
1160 1161 1162 1163 1164 1165 |
*/ checksum = calc_checksum(page); if (rmap_item->oldchecksum != checksum) { rmap_item->oldchecksum = checksum; return; } |
8dd3557a5
|
1166 1167 |
tree_rmap_item = unstable_tree_search_insert(rmap_item, page, &tree_page); |
31dbd01f3
|
1168 |
if (tree_rmap_item) { |
8dd3557a5
|
1169 1170 1171 |
kpage = try_to_merge_two_pages(rmap_item, page, tree_rmap_item, tree_page); put_page(tree_page); |
31dbd01f3
|
1172 1173 1174 1175 1176 |
/* * As soon as we merge this page, we want to remove the * rmap_item of the page we have merged with from the unstable * tree, and insert it instead as new node in the stable tree. */ |
8dd3557a5
|
1177 |
if (kpage) { |
93d17715a
|
1178 |
remove_rmap_item_from_tree(tree_rmap_item); |
473b0ce4d
|
1179 |
|
5ad646880
|
1180 |
lock_page(kpage); |
7b6ba2c7d
|
1181 1182 1183 1184 1185 |
stable_node = stable_tree_insert(kpage); if (stable_node) { stable_tree_append(tree_rmap_item, stable_node); stable_tree_append(rmap_item, stable_node); } |
5ad646880
|
1186 |
unlock_page(kpage); |
7b6ba2c7d
|
1187 |
|
31dbd01f3
|
1188 1189 1190 1191 1192 1193 |
/* * If we fail to insert the page into the stable tree, * we will have 2 virtual addresses that are pointing * to a ksm page left outside the stable tree, * in which case we need to break_cow on both. */ |
7b6ba2c7d
|
1194 |
if (!stable_node) { |
8dd3557a5
|
1195 1196 |
break_cow(tree_rmap_item); break_cow(rmap_item); |
31dbd01f3
|
1197 1198 |
} } |
31dbd01f3
|
1199 1200 1201 1202 |
} } static struct rmap_item *get_next_rmap_item(struct mm_slot *mm_slot, |
6514d511d
|
1203 |
struct rmap_item **rmap_list, |
31dbd01f3
|
1204 1205 1206 |
unsigned long addr) { struct rmap_item *rmap_item; |
6514d511d
|
1207 1208 |
while (*rmap_list) { rmap_item = *rmap_list; |
93d17715a
|
1209 |
if ((rmap_item->address & PAGE_MASK) == addr) |
31dbd01f3
|
1210 |
return rmap_item; |
31dbd01f3
|
1211 1212 |
if (rmap_item->address > addr) break; |
6514d511d
|
1213 |
*rmap_list = rmap_item->rmap_list; |
31dbd01f3
|
1214 |
remove_rmap_item_from_tree(rmap_item); |
31dbd01f3
|
1215 1216 1217 1218 1219 1220 1221 1222 |
free_rmap_item(rmap_item); } rmap_item = alloc_rmap_item(); if (rmap_item) { /* It has already been zeroed */ rmap_item->mm = mm_slot->mm; rmap_item->address = addr; |
6514d511d
|
1223 1224 |
rmap_item->rmap_list = *rmap_list; *rmap_list = rmap_item; |
31dbd01f3
|
1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 |
} return rmap_item; } static struct rmap_item *scan_get_next_rmap_item(struct page **page) { struct mm_struct *mm; struct mm_slot *slot; struct vm_area_struct *vma; struct rmap_item *rmap_item; if (list_empty(&ksm_mm_head.mm_list)) return NULL; slot = ksm_scan.mm_slot; if (slot == &ksm_mm_head) { |
2919bfd07
|
1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 |
/* * A number of pages can hang around indefinitely on per-cpu * pagevecs, raised page count preventing write_protect_page * from merging them. Though it doesn't really matter much, * it is puzzling to see some stuck in pages_volatile until * other activity jostles them out, and they also prevented * LTP's KSM test from succeeding deterministically; so drain * them here (here rather than on entry to ksm_do_scan(), * so we don't IPI too often when pages_to_scan is set low). */ lru_add_drain_all(); |
31dbd01f3
|
1252 1253 1254 1255 1256 1257 1258 1259 |
root_unstable_tree = RB_ROOT; spin_lock(&ksm_mmlist_lock); slot = list_entry(slot->mm_list.next, struct mm_slot, mm_list); ksm_scan.mm_slot = slot; spin_unlock(&ksm_mmlist_lock); next_mm: ksm_scan.address = 0; |
6514d511d
|
1260 |
ksm_scan.rmap_list = &slot->rmap_list; |
31dbd01f3
|
1261 1262 1263 1264 |
} mm = slot->mm; down_read(&mm->mmap_sem); |
9ba692948
|
1265 1266 1267 1268 1269 1270 |
if (ksm_test_exit(mm)) vma = NULL; else vma = find_vma(mm, ksm_scan.address); for (; vma; vma = vma->vm_next) { |
31dbd01f3
|
1271 1272 1273 1274 1275 1276 1277 1278 |
if (!(vma->vm_flags & VM_MERGEABLE)) continue; if (ksm_scan.address < vma->vm_start) ksm_scan.address = vma->vm_start; if (!vma->anon_vma) ksm_scan.address = vma->vm_end; while (ksm_scan.address < vma->vm_end) { |
9ba692948
|
1279 1280 |
if (ksm_test_exit(mm)) break; |
31dbd01f3
|
1281 |
*page = follow_page(vma, ksm_scan.address, FOLL_GET); |
21ae5b017
|
1282 1283 1284 1285 1286 |
if (IS_ERR_OR_NULL(*page)) { ksm_scan.address += PAGE_SIZE; cond_resched(); continue; } |
29ad768cf
|
1287 1288 |
if (PageAnon(*page) || page_trans_compound_anon(*page)) { |
31dbd01f3
|
1289 1290 1291 |
flush_anon_page(vma, *page, ksm_scan.address); flush_dcache_page(*page); rmap_item = get_next_rmap_item(slot, |
6514d511d
|
1292 |
ksm_scan.rmap_list, ksm_scan.address); |
31dbd01f3
|
1293 |
if (rmap_item) { |
6514d511d
|
1294 1295 |
ksm_scan.rmap_list = &rmap_item->rmap_list; |
31dbd01f3
|
1296 1297 1298 1299 1300 1301 |
ksm_scan.address += PAGE_SIZE; } else put_page(*page); up_read(&mm->mmap_sem); return rmap_item; } |
21ae5b017
|
1302 |
put_page(*page); |
31dbd01f3
|
1303 1304 1305 1306 |
ksm_scan.address += PAGE_SIZE; cond_resched(); } } |
9ba692948
|
1307 1308 |
if (ksm_test_exit(mm)) { ksm_scan.address = 0; |
6514d511d
|
1309 |
ksm_scan.rmap_list = &slot->rmap_list; |
9ba692948
|
1310 |
} |
31dbd01f3
|
1311 1312 1313 1314 |
/* * Nuke all the rmap_items that are above this current rmap: * because there were no VM_MERGEABLE vmas with such addresses. */ |
6514d511d
|
1315 |
remove_trailing_rmap_items(slot, ksm_scan.rmap_list); |
31dbd01f3
|
1316 1317 |
spin_lock(&ksm_mmlist_lock); |
cd551f975
|
1318 1319 1320 1321 1322 1323 1324 |
ksm_scan.mm_slot = list_entry(slot->mm_list.next, struct mm_slot, mm_list); if (ksm_scan.address == 0) { /* * We've completed a full scan of all vmas, holding mmap_sem * throughout, and found no VM_MERGEABLE: so do the same as * __ksm_exit does to remove this mm from all our lists now. |
9ba692948
|
1325 1326 1327 1328 |
* This applies either when cleaning up after __ksm_exit * (but beware: we can reach here even before __ksm_exit), * or when all VM_MERGEABLE areas have been unmapped (and * mmap_sem then protects against race with MADV_MERGEABLE). |
cd551f975
|
1329 1330 1331 |
*/ hlist_del(&slot->link); list_del(&slot->mm_list); |
9ba692948
|
1332 |
spin_unlock(&ksm_mmlist_lock); |
cd551f975
|
1333 1334 |
free_mm_slot(slot); clear_bit(MMF_VM_MERGEABLE, &mm->flags); |
9ba692948
|
1335 1336 1337 1338 1339 |
up_read(&mm->mmap_sem); mmdrop(mm); } else { spin_unlock(&ksm_mmlist_lock); up_read(&mm->mmap_sem); |
cd551f975
|
1340 |
} |
31dbd01f3
|
1341 1342 |
/* Repeat until we've completed scanning the whole list */ |
cd551f975
|
1343 |
slot = ksm_scan.mm_slot; |
31dbd01f3
|
1344 1345 |
if (slot != &ksm_mm_head) goto next_mm; |
31dbd01f3
|
1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 |
ksm_scan.seqnr++; return NULL; } /** * ksm_do_scan - the ksm scanner main worker function. * @scan_npages - number of pages we want to scan before we return. */ static void ksm_do_scan(unsigned int scan_npages) { struct rmap_item *rmap_item; |
22eccdd7d
|
1357 |
struct page *uninitialized_var(page); |
31dbd01f3
|
1358 |
|
878aee7d6
|
1359 |
while (scan_npages-- && likely(!freezing(current))) { |
31dbd01f3
|
1360 1361 1362 1363 1364 1365 1366 1367 1368 |
cond_resched(); rmap_item = scan_get_next_rmap_item(&page); if (!rmap_item) return; if (!PageKsm(page) || !in_stable_tree(rmap_item)) cmp_and_merge_page(page, rmap_item); put_page(page); } } |
6e1583842
|
1369 1370 1371 1372 |
static int ksmd_should_run(void) { return (ksm_run & KSM_RUN_MERGE) && !list_empty(&ksm_mm_head.mm_list); } |
31dbd01f3
|
1373 1374 |
static int ksm_scan_thread(void *nothing) { |
878aee7d6
|
1375 |
set_freezable(); |
339aa6246
|
1376 |
set_user_nice(current, 5); |
31dbd01f3
|
1377 1378 |
while (!kthread_should_stop()) { |
6e1583842
|
1379 1380 |
mutex_lock(&ksm_thread_mutex); if (ksmd_should_run()) |
31dbd01f3
|
1381 |
ksm_do_scan(ksm_thread_pages_to_scan); |
6e1583842
|
1382 |
mutex_unlock(&ksm_thread_mutex); |
878aee7d6
|
1383 |
try_to_freeze(); |
6e1583842
|
1384 |
if (ksmd_should_run()) { |
31dbd01f3
|
1385 1386 1387 |
schedule_timeout_interruptible( msecs_to_jiffies(ksm_thread_sleep_millisecs)); } else { |
878aee7d6
|
1388 |
wait_event_freezable(ksm_thread_wait, |
6e1583842
|
1389 |
ksmd_should_run() || kthread_should_stop()); |
31dbd01f3
|
1390 1391 1392 1393 |
} } return 0; } |
f8af4da3b
|
1394 1395 1396 1397 |
int ksm_madvise(struct vm_area_struct *vma, unsigned long start, unsigned long end, int advice, unsigned long *vm_flags) { struct mm_struct *mm = vma->vm_mm; |
d952b7913
|
1398 |
int err; |
f8af4da3b
|
1399 1400 1401 1402 1403 1404 1405 1406 1407 |
switch (advice) { case MADV_MERGEABLE: /* * Be somewhat over-protective for now! */ if (*vm_flags & (VM_MERGEABLE | VM_SHARED | VM_MAYSHARE | VM_PFNMAP | VM_IO | VM_DONTEXPAND | VM_RESERVED | VM_HUGETLB | VM_INSERTPAGE | |
5ad646880
|
1408 |
VM_NONLINEAR | VM_MIXEDMAP | VM_SAO)) |
f8af4da3b
|
1409 |
return 0; /* just ignore the advice */ |
d952b7913
|
1410 1411 1412 1413 1414 |
if (!test_bit(MMF_VM_MERGEABLE, &mm->flags)) { err = __ksm_enter(mm); if (err) return err; } |
f8af4da3b
|
1415 1416 1417 1418 1419 1420 1421 |
*vm_flags |= VM_MERGEABLE; break; case MADV_UNMERGEABLE: if (!(*vm_flags & VM_MERGEABLE)) return 0; /* just ignore the advice */ |
d952b7913
|
1422 1423 1424 1425 1426 |
if (vma->anon_vma) { err = unmerge_ksm_pages(vma, start, end); if (err) return err; } |
f8af4da3b
|
1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 |
*vm_flags &= ~VM_MERGEABLE; break; } return 0; } int __ksm_enter(struct mm_struct *mm) { |
6e1583842
|
1437 1438 1439 1440 |
struct mm_slot *mm_slot; int needs_wakeup; mm_slot = alloc_mm_slot(); |
31dbd01f3
|
1441 1442 |
if (!mm_slot) return -ENOMEM; |
6e1583842
|
1443 1444 |
/* Check ksm_run too? Would need tighter locking */ needs_wakeup = list_empty(&ksm_mm_head.mm_list); |
31dbd01f3
|
1445 1446 1447 1448 1449 1450 1451 1452 1453 |
spin_lock(&ksm_mmlist_lock); insert_to_mm_slots_hash(mm, mm_slot); /* * Insert just behind the scanning cursor, to let the area settle * down a little; when fork is followed by immediate exec, we don't * want ksmd to waste time setting up and tearing down an rmap_list. */ list_add_tail(&mm_slot->mm_list, &ksm_scan.mm_slot->mm_list); spin_unlock(&ksm_mmlist_lock); |
f8af4da3b
|
1454 |
set_bit(MMF_VM_MERGEABLE, &mm->flags); |
9ba692948
|
1455 |
atomic_inc(&mm->mm_count); |
6e1583842
|
1456 1457 1458 |
if (needs_wakeup) wake_up_interruptible(&ksm_thread_wait); |
f8af4da3b
|
1459 1460 |
return 0; } |
1c2fb7a4c
|
1461 |
void __ksm_exit(struct mm_struct *mm) |
f8af4da3b
|
1462 |
{ |
cd551f975
|
1463 |
struct mm_slot *mm_slot; |
9ba692948
|
1464 |
int easy_to_free = 0; |
cd551f975
|
1465 |
|
31dbd01f3
|
1466 |
/* |
9ba692948
|
1467 1468 1469 1470 1471 1472 |
* This process is exiting: if it's straightforward (as is the * case when ksmd was never running), free mm_slot immediately. * But if it's at the cursor or has rmap_items linked to it, use * mmap_sem to synchronize with any break_cows before pagetables * are freed, and leave the mm_slot on the list for ksmd to free. * Beware: ksm may already have noticed it exiting and freed the slot. |
31dbd01f3
|
1473 |
*/ |
9ba692948
|
1474 |
|
cd551f975
|
1475 1476 |
spin_lock(&ksm_mmlist_lock); mm_slot = get_mm_slot(mm); |
9ba692948
|
1477 |
if (mm_slot && ksm_scan.mm_slot != mm_slot) { |
6514d511d
|
1478 |
if (!mm_slot->rmap_list) { |
9ba692948
|
1479 1480 1481 1482 1483 1484 1485 |
hlist_del(&mm_slot->link); list_del(&mm_slot->mm_list); easy_to_free = 1; } else { list_move(&mm_slot->mm_list, &ksm_scan.mm_slot->mm_list); } |
cd551f975
|
1486 |
} |
cd551f975
|
1487 |
spin_unlock(&ksm_mmlist_lock); |
9ba692948
|
1488 1489 1490 1491 1492 |
if (easy_to_free) { free_mm_slot(mm_slot); clear_bit(MMF_VM_MERGEABLE, &mm->flags); mmdrop(mm); } else if (mm_slot) { |
9ba692948
|
1493 1494 |
down_write(&mm->mmap_sem); up_write(&mm->mmap_sem); |
9ba692948
|
1495 |
} |
31dbd01f3
|
1496 |
} |
5ad646880
|
1497 1498 1499 1500 |
struct page *ksm_does_need_to_copy(struct page *page, struct vm_area_struct *vma, unsigned long address) { struct page *new_page; |
5ad646880
|
1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 |
new_page = alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma, address); if (new_page) { copy_user_highpage(new_page, page, address, vma); SetPageDirty(new_page); __SetPageUptodate(new_page); SetPageSwapBacked(new_page); __set_page_locked(new_page); if (page_evictable(new_page, vma)) lru_cache_add_lru(new_page, LRU_ACTIVE_ANON); else add_page_to_unevictable_list(new_page); } |
5ad646880
|
1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 |
return new_page; } int page_referenced_ksm(struct page *page, struct mem_cgroup *memcg, unsigned long *vm_flags) { struct stable_node *stable_node; struct rmap_item *rmap_item; struct hlist_node *hlist; unsigned int mapcount = page_mapcount(page); int referenced = 0; |
db114b83a
|
1526 |
int search_new_forks = 0; |
5ad646880
|
1527 1528 1529 1530 1531 1532 1533 |
VM_BUG_ON(!PageKsm(page)); VM_BUG_ON(!PageLocked(page)); stable_node = page_stable_node(page); if (!stable_node) return 0; |
db114b83a
|
1534 |
again: |
5ad646880
|
1535 |
hlist_for_each_entry(rmap_item, hlist, &stable_node->hlist, hlist) { |
db114b83a
|
1536 |
struct anon_vma *anon_vma = rmap_item->anon_vma; |
5beb49305
|
1537 |
struct anon_vma_chain *vmac; |
db114b83a
|
1538 |
struct vm_area_struct *vma; |
5ad646880
|
1539 |
|
cba48b98f
|
1540 |
anon_vma_lock(anon_vma); |
5beb49305
|
1541 1542 |
list_for_each_entry(vmac, &anon_vma->head, same_anon_vma) { vma = vmac->vma; |
db114b83a
|
1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 |
if (rmap_item->address < vma->vm_start || rmap_item->address >= vma->vm_end) continue; /* * Initially we examine only the vma which covers this * rmap_item; but later, if there is still work to do, * we examine covering vmas in other mms: in case they * were forked from the original since ksmd passed. */ if ((rmap_item->mm == vma->vm_mm) == search_new_forks) continue; if (memcg && !mm_match_cgroup(vma->vm_mm, memcg)) continue; |
5ad646880
|
1557 |
|
db114b83a
|
1558 |
referenced += page_referenced_one(page, vma, |
5ad646880
|
1559 |
rmap_item->address, &mapcount, vm_flags); |
db114b83a
|
1560 1561 1562 |
if (!search_new_forks || !mapcount) break; } |
cba48b98f
|
1563 |
anon_vma_unlock(anon_vma); |
5ad646880
|
1564 1565 1566 |
if (!mapcount) goto out; } |
db114b83a
|
1567 1568 |
if (!search_new_forks++) goto again; |
5ad646880
|
1569 |
out: |
5ad646880
|
1570 1571 1572 1573 1574 1575 1576 1577 1578 |
return referenced; } int try_to_unmap_ksm(struct page *page, enum ttu_flags flags) { struct stable_node *stable_node; struct hlist_node *hlist; struct rmap_item *rmap_item; int ret = SWAP_AGAIN; |
db114b83a
|
1579 |
int search_new_forks = 0; |
5ad646880
|
1580 1581 1582 1583 1584 1585 1586 |
VM_BUG_ON(!PageKsm(page)); VM_BUG_ON(!PageLocked(page)); stable_node = page_stable_node(page); if (!stable_node) return SWAP_FAIL; |
db114b83a
|
1587 |
again: |
5ad646880
|
1588 |
hlist_for_each_entry(rmap_item, hlist, &stable_node->hlist, hlist) { |
db114b83a
|
1589 |
struct anon_vma *anon_vma = rmap_item->anon_vma; |
5beb49305
|
1590 |
struct anon_vma_chain *vmac; |
db114b83a
|
1591 |
struct vm_area_struct *vma; |
5ad646880
|
1592 |
|
cba48b98f
|
1593 |
anon_vma_lock(anon_vma); |
5beb49305
|
1594 1595 |
list_for_each_entry(vmac, &anon_vma->head, same_anon_vma) { vma = vmac->vma; |
db114b83a
|
1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 |
if (rmap_item->address < vma->vm_start || rmap_item->address >= vma->vm_end) continue; /* * Initially we examine only the vma which covers this * rmap_item; but later, if there is still work to do, * we examine covering vmas in other mms: in case they * were forked from the original since ksmd passed. */ if ((rmap_item->mm == vma->vm_mm) == search_new_forks) continue; ret = try_to_unmap_one(page, vma, rmap_item->address, flags); if (ret != SWAP_AGAIN || !page_mapped(page)) { |
cba48b98f
|
1611 |
anon_vma_unlock(anon_vma); |
db114b83a
|
1612 1613 1614 |
goto out; } } |
cba48b98f
|
1615 |
anon_vma_unlock(anon_vma); |
5ad646880
|
1616 |
} |
db114b83a
|
1617 1618 |
if (!search_new_forks++) goto again; |
5ad646880
|
1619 |
out: |
5ad646880
|
1620 1621 |
return ret; } |
e9995ef97
|
1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 |
#ifdef CONFIG_MIGRATION int rmap_walk_ksm(struct page *page, int (*rmap_one)(struct page *, struct vm_area_struct *, unsigned long, void *), void *arg) { struct stable_node *stable_node; struct hlist_node *hlist; struct rmap_item *rmap_item; int ret = SWAP_AGAIN; int search_new_forks = 0; VM_BUG_ON(!PageKsm(page)); VM_BUG_ON(!PageLocked(page)); stable_node = page_stable_node(page); if (!stable_node) return ret; again: hlist_for_each_entry(rmap_item, hlist, &stable_node->hlist, hlist) { struct anon_vma *anon_vma = rmap_item->anon_vma; |
5beb49305
|
1641 |
struct anon_vma_chain *vmac; |
e9995ef97
|
1642 |
struct vm_area_struct *vma; |
cba48b98f
|
1643 |
anon_vma_lock(anon_vma); |
5beb49305
|
1644 1645 |
list_for_each_entry(vmac, &anon_vma->head, same_anon_vma) { vma = vmac->vma; |
e9995ef97
|
1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 |
if (rmap_item->address < vma->vm_start || rmap_item->address >= vma->vm_end) continue; /* * Initially we examine only the vma which covers this * rmap_item; but later, if there is still work to do, * we examine covering vmas in other mms: in case they * were forked from the original since ksmd passed. */ if ((rmap_item->mm == vma->vm_mm) == search_new_forks) continue; ret = rmap_one(page, vma, rmap_item->address, arg); if (ret != SWAP_AGAIN) { |
cba48b98f
|
1660 |
anon_vma_unlock(anon_vma); |
e9995ef97
|
1661 1662 1663 |
goto out; } } |
cba48b98f
|
1664 |
anon_vma_unlock(anon_vma); |
e9995ef97
|
1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 |
} if (!search_new_forks++) goto again; out: return ret; } void ksm_migrate_page(struct page *newpage, struct page *oldpage) { struct stable_node *stable_node; VM_BUG_ON(!PageLocked(oldpage)); VM_BUG_ON(!PageLocked(newpage)); VM_BUG_ON(newpage->mapping != oldpage->mapping); stable_node = page_stable_node(newpage); if (stable_node) { |
62b61f611
|
1682 1683 |
VM_BUG_ON(stable_node->kpfn != page_to_pfn(oldpage)); stable_node->kpfn = page_to_pfn(newpage); |
e9995ef97
|
1684 1685 1686 |
} } #endif /* CONFIG_MIGRATION */ |
62b61f611
|
1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 |
#ifdef CONFIG_MEMORY_HOTREMOVE static struct stable_node *ksm_check_stable_tree(unsigned long start_pfn, unsigned long end_pfn) { struct rb_node *node; for (node = rb_first(&root_stable_tree); node; node = rb_next(node)) { struct stable_node *stable_node; stable_node = rb_entry(node, struct stable_node, node); if (stable_node->kpfn >= start_pfn && stable_node->kpfn < end_pfn) return stable_node; } return NULL; } static int ksm_memory_callback(struct notifier_block *self, unsigned long action, void *arg) { struct memory_notify *mn = arg; struct stable_node *stable_node; switch (action) { case MEM_GOING_OFFLINE: /* * Keep it very simple for now: just lock out ksmd and * MADV_UNMERGEABLE while any memory is going offline. |
a0b0f58cd
|
1715 1716 1717 1718 1719 |
* mutex_lock_nested() is necessary because lockdep was alarmed * that here we take ksm_thread_mutex inside notifier chain * mutex, and later take notifier chain mutex inside * ksm_thread_mutex to unlock it. But that's safe because both * are inside mem_hotplug_mutex. |
62b61f611
|
1720 |
*/ |
a0b0f58cd
|
1721 |
mutex_lock_nested(&ksm_thread_mutex, SINGLE_DEPTH_NESTING); |
62b61f611
|
1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 |
break; case MEM_OFFLINE: /* * Most of the work is done by page migration; but there might * be a few stable_nodes left over, still pointing to struct * pages which have been offlined: prune those from the tree. */ while ((stable_node = ksm_check_stable_tree(mn->start_pfn, mn->start_pfn + mn->nr_pages)) != NULL) remove_node_from_stable_tree(stable_node); /* fallthrough */ case MEM_CANCEL_OFFLINE: mutex_unlock(&ksm_thread_mutex); break; } return NOTIFY_OK; } #endif /* CONFIG_MEMORY_HOTREMOVE */ |
2ffd8679c
|
1742 1743 1744 1745 |
#ifdef CONFIG_SYSFS /* * This all compiles without CONFIG_SYSFS, but is a waste of space. */ |
31dbd01f3
|
1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 |
#define KSM_ATTR_RO(_name) \ static struct kobj_attribute _name##_attr = __ATTR_RO(_name) #define KSM_ATTR(_name) \ static struct kobj_attribute _name##_attr = \ __ATTR(_name, 0644, _name##_show, _name##_store) static ssize_t sleep_millisecs_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf) { return sprintf(buf, "%u ", ksm_thread_sleep_millisecs); } static ssize_t sleep_millisecs_store(struct kobject *kobj, struct kobj_attribute *attr, const char *buf, size_t count) { unsigned long msecs; int err; err = strict_strtoul(buf, 10, &msecs); if (err || msecs > UINT_MAX) return -EINVAL; ksm_thread_sleep_millisecs = msecs; return count; } KSM_ATTR(sleep_millisecs); static ssize_t pages_to_scan_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf) { return sprintf(buf, "%u ", ksm_thread_pages_to_scan); } static ssize_t pages_to_scan_store(struct kobject *kobj, struct kobj_attribute *attr, const char *buf, size_t count) { int err; unsigned long nr_pages; err = strict_strtoul(buf, 10, &nr_pages); if (err || nr_pages > UINT_MAX) return -EINVAL; ksm_thread_pages_to_scan = nr_pages; return count; } KSM_ATTR(pages_to_scan); static ssize_t run_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf) { return sprintf(buf, "%u ", ksm_run); } static ssize_t run_store(struct kobject *kobj, struct kobj_attribute *attr, const char *buf, size_t count) { int err; unsigned long flags; err = strict_strtoul(buf, 10, &flags); if (err || flags > UINT_MAX) return -EINVAL; if (flags > KSM_RUN_UNMERGE) return -EINVAL; /* * KSM_RUN_MERGE sets ksmd running, and 0 stops it running. * KSM_RUN_UNMERGE stops it running and unmerges all rmap_items, |
d0f209f68
|
1822 1823 |
* breaking COW to free the pages_shared (but leaves mm_slots * on the list for when ksmd may be set running again). |
31dbd01f3
|
1824 1825 1826 1827 1828 |
*/ mutex_lock(&ksm_thread_mutex); if (ksm_run != flags) { ksm_run = flags; |
d952b7913
|
1829 |
if (flags & KSM_RUN_UNMERGE) { |
35451beec
|
1830 |
current->flags |= PF_OOM_ORIGIN; |
d952b7913
|
1831 |
err = unmerge_and_remove_all_rmap_items(); |
35451beec
|
1832 |
current->flags &= ~PF_OOM_ORIGIN; |
d952b7913
|
1833 1834 1835 1836 1837 |
if (err) { ksm_run = KSM_RUN_STOP; count = err; } } |
31dbd01f3
|
1838 1839 1840 1841 1842 1843 1844 1845 1846 |
} mutex_unlock(&ksm_thread_mutex); if (flags & KSM_RUN_MERGE) wake_up_interruptible(&ksm_thread_wait); return count; } KSM_ATTR(run); |
b40282603
|
1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 |
static ssize_t pages_shared_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf) { return sprintf(buf, "%lu ", ksm_pages_shared); } KSM_ATTR_RO(pages_shared); static ssize_t pages_sharing_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf) { |
e178dfde3
|
1858 1859 |
return sprintf(buf, "%lu ", ksm_pages_sharing); |
b40282603
|
1860 1861 |
} KSM_ATTR_RO(pages_sharing); |
473b0ce4d
|
1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 |
static ssize_t pages_unshared_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf) { return sprintf(buf, "%lu ", ksm_pages_unshared); } KSM_ATTR_RO(pages_unshared); static ssize_t pages_volatile_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf) { long ksm_pages_volatile; ksm_pages_volatile = ksm_rmap_items - ksm_pages_shared - ksm_pages_sharing - ksm_pages_unshared; /* * It was not worth any locking to calculate that statistic, * but it might therefore sometimes be negative: conceal that. */ if (ksm_pages_volatile < 0) ksm_pages_volatile = 0; return sprintf(buf, "%ld ", ksm_pages_volatile); } KSM_ATTR_RO(pages_volatile); static ssize_t full_scans_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf) { return sprintf(buf, "%lu ", ksm_scan.seqnr); } KSM_ATTR_RO(full_scans); |
31dbd01f3
|
1895 1896 1897 1898 |
static struct attribute *ksm_attrs[] = { &sleep_millisecs_attr.attr, &pages_to_scan_attr.attr, &run_attr.attr, |
b40282603
|
1899 1900 |
&pages_shared_attr.attr, &pages_sharing_attr.attr, |
473b0ce4d
|
1901 1902 1903 |
&pages_unshared_attr.attr, &pages_volatile_attr.attr, &full_scans_attr.attr, |
31dbd01f3
|
1904 1905 1906 1907 1908 1909 1910 |
NULL, }; static struct attribute_group ksm_attr_group = { .attrs = ksm_attrs, .name = "ksm", }; |
2ffd8679c
|
1911 |
#endif /* CONFIG_SYSFS */ |
31dbd01f3
|
1912 1913 1914 1915 1916 1917 1918 1919 1920 |
static int __init ksm_init(void) { struct task_struct *ksm_thread; int err; err = ksm_slab_init(); if (err) goto out; |
31dbd01f3
|
1921 1922 1923 1924 1925 |
ksm_thread = kthread_run(ksm_scan_thread, NULL, "ksmd"); if (IS_ERR(ksm_thread)) { printk(KERN_ERR "ksm: creating kthread failed "); err = PTR_ERR(ksm_thread); |
d9f8984c2
|
1926 |
goto out_free; |
31dbd01f3
|
1927 |
} |
2ffd8679c
|
1928 |
#ifdef CONFIG_SYSFS |
31dbd01f3
|
1929 1930 1931 1932 |
err = sysfs_create_group(mm_kobj, &ksm_attr_group); if (err) { printk(KERN_ERR "ksm: register sysfs failed "); |
2ffd8679c
|
1933 |
kthread_stop(ksm_thread); |
d9f8984c2
|
1934 |
goto out_free; |
31dbd01f3
|
1935 |
} |
c73602ad3
|
1936 1937 |
#else ksm_run = KSM_RUN_MERGE; /* no way for user to start it */ |
2ffd8679c
|
1938 |
#endif /* CONFIG_SYSFS */ |
31dbd01f3
|
1939 |
|
62b61f611
|
1940 1941 1942 1943 1944 1945 1946 |
#ifdef CONFIG_MEMORY_HOTREMOVE /* * Choose a high priority since the callback takes ksm_thread_mutex: * later callbacks could only be taking locks which nest within that. */ hotplug_memory_notifier(ksm_memory_callback, 100); #endif |
31dbd01f3
|
1947 |
return 0; |
d9f8984c2
|
1948 |
out_free: |
31dbd01f3
|
1949 1950 1951 |
ksm_slab_free(); out: return err; |
f8af4da3b
|
1952 |
} |
31dbd01f3
|
1953 |
module_init(ksm_init) |