Blame view
mm/mlock.c
22.6 KB
1da177e4c Linux-2.6.12-rc2 |
1 2 3 4 5 6 |
/* * linux/mm/mlock.c * * (C) Copyright 1995 Linus Torvalds * (C) Copyright 2002 Christoph Hellwig */ |
c59ede7b7 [PATCH] move capa... |
7 |
#include <linux/capability.h> |
1da177e4c Linux-2.6.12-rc2 |
8 9 |
#include <linux/mman.h> #include <linux/mm.h> |
b291f0003 mlock: mlocked pa... |
10 11 12 |
#include <linux/swap.h> #include <linux/swapops.h> #include <linux/pagemap.h> |
7225522bb mm: munlock: batc... |
13 |
#include <linux/pagevec.h> |
1da177e4c Linux-2.6.12-rc2 |
14 15 |
#include <linux/mempolicy.h> #include <linux/syscalls.h> |
e8edc6e03 Detach sched.h fr... |
16 |
#include <linux/sched.h> |
b95f1b31b mm: Map most file... |
17 |
#include <linux/export.h> |
b291f0003 mlock: mlocked pa... |
18 19 20 |
#include <linux/rmap.h> #include <linux/mmzone.h> #include <linux/hugetlb.h> |
7225522bb mm: munlock: batc... |
21 22 |
#include <linux/memcontrol.h> #include <linux/mm_inline.h> |
b291f0003 mlock: mlocked pa... |
23 24 |
#include "internal.h" |
1da177e4c Linux-2.6.12-rc2 |
25 |
|
e8edc6e03 Detach sched.h fr... |
26 27 28 29 |
int can_do_mlock(void) { if (capable(CAP_IPC_LOCK)) return 1; |
59e99e5b9 mm: use rlimit he... |
30 |
if (rlimit(RLIMIT_MEMLOCK) != 0) |
e8edc6e03 Detach sched.h fr... |
31 32 33 34 |
return 1; return 0; } EXPORT_SYMBOL(can_do_mlock); |
1da177e4c Linux-2.6.12-rc2 |
35 |
|
b291f0003 mlock: mlocked pa... |
36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 |
/* * Mlocked pages are marked with PageMlocked() flag for efficient testing * in vmscan and, possibly, the fault path; and to support semi-accurate * statistics. * * An mlocked page [PageMlocked(page)] is unevictable. As such, it will * be placed on the LRU "unevictable" list, rather than the [in]active lists. * The unevictable list is an LRU sibling list to the [in]active lists. * PageUnevictable is set to indicate the unevictable state. * * When lazy mlocking via vmscan, it is important to ensure that the * vma's VM_LOCKED status is not concurrently being modified, otherwise we * may have mlocked a page that is being munlocked. So lazy mlock must take * the mmap_sem for read, and verify that the vma really is locked * (see mm/rmap.c). */ /* * LRU accounting for clear_page_mlock() */ |
e6c509f85 mm: use clear_pag... |
56 |
void clear_page_mlock(struct page *page) |
b291f0003 mlock: mlocked pa... |
57 |
{ |
e6c509f85 mm: use clear_pag... |
58 |
if (!TestClearPageMlocked(page)) |
b291f0003 mlock: mlocked pa... |
59 |
return; |
b291f0003 mlock: mlocked pa... |
60 |
|
8449d21fb mm, thp: fix mloc... |
61 62 |
mod_zone_page_state(page_zone(page), NR_MLOCK, -hpage_nr_pages(page)); |
5344b7e64 vmstat: mlocked p... |
63 |
count_vm_event(UNEVICTABLE_PGCLEARED); |
b291f0003 mlock: mlocked pa... |
64 65 66 67 |
if (!isolate_lru_page(page)) { putback_lru_page(page); } else { /* |
8891d6da1 mm: remove lru_ad... |
68 |
* We lost the race. the page already moved to evictable list. |
b291f0003 mlock: mlocked pa... |
69 |
*/ |
8891d6da1 mm: remove lru_ad... |
70 |
if (PageUnevictable(page)) |
5344b7e64 vmstat: mlocked p... |
71 |
count_vm_event(UNEVICTABLE_PGSTRANDED); |
b291f0003 mlock: mlocked pa... |
72 73 74 75 76 77 78 79 80 |
} } /* * Mark page as mlocked if not already. * If page on LRU, isolate and putback to move to unevictable list. */ void mlock_vma_page(struct page *page) { |
57e68e9cd mm: try_to_unmap_... |
81 |
/* Serialize with page migration */ |
b291f0003 mlock: mlocked pa... |
82 |
BUG_ON(!PageLocked(page)); |
5344b7e64 vmstat: mlocked p... |
83 |
if (!TestSetPageMlocked(page)) { |
8449d21fb mm, thp: fix mloc... |
84 85 |
mod_zone_page_state(page_zone(page), NR_MLOCK, hpage_nr_pages(page)); |
5344b7e64 vmstat: mlocked p... |
86 87 88 89 |
count_vm_event(UNEVICTABLE_PGMLOCKED); if (!isolate_lru_page(page)) putback_lru_page(page); } |
b291f0003 mlock: mlocked pa... |
90 |
} |
7225522bb mm: munlock: batc... |
91 |
/* |
01cc2e586 mm: munlock: fix ... |
92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 |
* Isolate a page from LRU with optional get_page() pin. * Assumes lru_lock already held and page already pinned. */ static bool __munlock_isolate_lru_page(struct page *page, bool getpage) { if (PageLRU(page)) { struct lruvec *lruvec; lruvec = mem_cgroup_page_lruvec(page, page_zone(page)); if (getpage) get_page(page); ClearPageLRU(page); del_page_from_lru_list(page, lruvec, page_lru(page)); return true; } return false; } /* |
7225522bb mm: munlock: batc... |
112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 |
* Finish munlock after successful page isolation * * Page must be locked. This is a wrapper for try_to_munlock() * and putback_lru_page() with munlock accounting. */ static void __munlock_isolated_page(struct page *page) { int ret = SWAP_AGAIN; /* * Optimization: if the page was mapped just once, that's our mapping * and we don't need to check all the other vmas. */ if (page_mapcount(page) > 1) ret = try_to_munlock(page); /* Did try_to_unlock() succeed or punt? */ if (ret != SWAP_MLOCK) count_vm_event(UNEVICTABLE_PGMUNLOCKED); putback_lru_page(page); } /* * Accounting for page isolation fail during munlock * * Performs accounting when page isolation fails in munlock. There is nothing * else to do because it means some other task has already removed the page * from the LRU. putback_lru_page() will take care of removing the page from * the unevictable list, if necessary. vmscan [page_referenced()] will move * the page back to the unevictable list if some other vma has it mlocked. */ static void __munlock_isolation_failed(struct page *page) { if (PageUnevictable(page)) |
01cc2e586 mm: munlock: fix ... |
147 |
__count_vm_event(UNEVICTABLE_PGSTRANDED); |
7225522bb mm: munlock: batc... |
148 |
else |
01cc2e586 mm: munlock: fix ... |
149 |
__count_vm_event(UNEVICTABLE_PGMUNLOCKED); |
7225522bb mm: munlock: batc... |
150 |
} |
6927c1dd9 mlock: replace st... |
151 152 |
/** * munlock_vma_page - munlock a vma page |
c424be1cb mm: munlock: fix ... |
153 154 155 156 |
* @page - page to be unlocked, either a normal page or THP page head * * returns the size of the page as a page mask (0 for normal page, * HPAGE_PMD_NR - 1 for THP head page) |
b291f0003 mlock: mlocked pa... |
157 |
* |
6927c1dd9 mlock: replace st... |
158 159 160 161 162 163 164 165 166 167 |
* called from munlock()/munmap() path with page supposedly on the LRU. * When we munlock a page, because the vma where we found the page is being * munlock()ed or munmap()ed, we want to check whether other vmas hold the * page locked so that we can leave it on the unevictable lru list and not * bother vmscan with it. However, to walk the page's rmap list in * try_to_munlock() we must isolate the page from the LRU. If some other * task has removed the page from the LRU, we won't be able to do that. * So we clear the PageMlocked as we might not get another chance. If we * can't isolate the page, we leave it for putback_lru_page() and vmscan * [page_referenced()/try_to_unmap()] to deal with. |
b291f0003 mlock: mlocked pa... |
168 |
*/ |
ff6a6da60 mm: accelerate mu... |
169 |
unsigned int munlock_vma_page(struct page *page) |
b291f0003 mlock: mlocked pa... |
170 |
{ |
c424be1cb mm: munlock: fix ... |
171 |
unsigned int nr_pages; |
01cc2e586 mm: munlock: fix ... |
172 |
struct zone *zone = page_zone(page); |
ff6a6da60 mm: accelerate mu... |
173 |
|
57e68e9cd mm: try_to_unmap_... |
174 |
/* For try_to_munlock() and to serialize with page migration */ |
b291f0003 mlock: mlocked pa... |
175 |
BUG_ON(!PageLocked(page)); |
c424be1cb mm: munlock: fix ... |
176 |
/* |
01cc2e586 mm: munlock: fix ... |
177 178 179 |
* Serialize with any parallel __split_huge_page_refcount() which * might otherwise copy PageMlocked to part of the tail pages before * we clear it in the head page. It also stabilizes hpage_nr_pages(). |
c424be1cb mm: munlock: fix ... |
180 |
*/ |
01cc2e586 mm: munlock: fix ... |
181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 |
spin_lock_irq(&zone->lru_lock); nr_pages = hpage_nr_pages(page); if (!TestClearPageMlocked(page)) goto unlock_out; __mod_zone_page_state(zone, NR_MLOCK, -nr_pages); if (__munlock_isolate_lru_page(page, true)) { spin_unlock_irq(&zone->lru_lock); __munlock_isolated_page(page); goto out; } __munlock_isolation_failed(page); unlock_out: spin_unlock_irq(&zone->lru_lock); out: |
c424be1cb mm: munlock: fix ... |
200 |
return nr_pages - 1; |
b291f0003 mlock: mlocked pa... |
201 |
} |
ba470de43 mmap: handle mloc... |
202 |
/** |
408e82b78 mm: munlock use f... |
203 |
* __mlock_vma_pages_range() - mlock a range of pages in the vma. |
ba470de43 mmap: handle mloc... |
204 205 206 |
* @vma: target vma * @start: start address * @end: end address |
ba470de43 mmap: handle mloc... |
207 |
* |
408e82b78 mm: munlock use f... |
208 |
* This takes care of making the pages present too. |
b291f0003 mlock: mlocked pa... |
209 |
* |
ba470de43 mmap: handle mloc... |
210 |
* return 0 on success, negative error code on error. |
b291f0003 mlock: mlocked pa... |
211 |
* |
ba470de43 mmap: handle mloc... |
212 |
* vma->vm_mm->mmap_sem must be held for at least read. |
b291f0003 mlock: mlocked pa... |
213 |
*/ |
cea10a19b mm: directly use ... |
214 215 |
long __mlock_vma_pages_range(struct vm_area_struct *vma, unsigned long start, unsigned long end, int *nonblocking) |
b291f0003 mlock: mlocked pa... |
216 217 |
{ struct mm_struct *mm = vma->vm_mm; |
28a35716d mm: use long type... |
218 |
unsigned long nr_pages = (end - start) / PAGE_SIZE; |
408e82b78 mm: munlock use f... |
219 |
int gup_flags; |
ba470de43 mmap: handle mloc... |
220 221 222 223 224 |
VM_BUG_ON(start & ~PAGE_MASK); VM_BUG_ON(end & ~PAGE_MASK); VM_BUG_ON(start < vma->vm_start); VM_BUG_ON(end > vma->vm_end); |
408e82b78 mm: munlock use f... |
225 |
VM_BUG_ON(!rwsem_is_locked(&mm->mmap_sem)); |
b291f0003 mlock: mlocked pa... |
226 |
|
a1fde08c7 VM: skip the stac... |
227 |
gup_flags = FOLL_TOUCH | FOLL_MLOCK; |
5ecfda041 mlock: avoid dirt... |
228 229 230 231 232 233 |
/* * We want to touch writable mappings with a write fault in order * to break COW, except for shared mappings because these don't COW * and we would not want to dirty them for nothing. */ if ((vma->vm_flags & (VM_WRITE | VM_SHARED)) == VM_WRITE) |
58fa879e1 mm: FOLL flags fo... |
234 |
gup_flags |= FOLL_WRITE; |
b291f0003 mlock: mlocked pa... |
235 |
|
fdf4c587a mlock: operate on... |
236 237 238 239 240 241 |
/* * We want mlock to succeed for regions that have any permissions * other than PROT_NONE. */ if (vma->vm_flags & (VM_READ | VM_WRITE | VM_EXEC)) gup_flags |= FOLL_FORCE; |
4805b02e9 mm/mlock.c: docum... |
242 243 244 245 |
/* * We made sure addr is within a VMA, so the following will * not result in a stack expansion that recurses back here. */ |
ff6a6da60 mm: accelerate mu... |
246 |
return __get_user_pages(current, mm, start, nr_pages, gup_flags, |
53a7706d5 mlock: do not hol... |
247 |
NULL, NULL, nonblocking); |
9978ad583 mlock: make mlock... |
248 249 250 251 252 253 254 255 256 257 258 259 |
} /* * convert get_user_pages() return value to posix mlock() error */ static int __mlock_posix_error_return(long retval) { if (retval == -EFAULT) retval = -ENOMEM; else if (retval == -ENOMEM) retval = -EAGAIN; return retval; |
b291f0003 mlock: mlocked pa... |
260 |
} |
b291f0003 mlock: mlocked pa... |
261 |
/* |
56afe477d mm: munlock: bypa... |
262 263 264 265 266 267 268 269 270 271 272 273 274 275 |
* Prepare page for fast batched LRU putback via putback_lru_evictable_pagevec() * * The fast path is available only for evictable pages with single mapping. * Then we can bypass the per-cpu pvec and get better performance. * when mapcount > 1 we need try_to_munlock() which can fail. * when !page_evictable(), we need the full redo logic of putback_lru_page to * avoid leaving evictable page in unevictable list. * * In case of success, @page is added to @pvec and @pgrescued is incremented * in case that the page was previously unevictable. @page is also unlocked. */ static bool __putback_lru_fast_prepare(struct page *page, struct pagevec *pvec, int *pgrescued) { |
309381fea mm: dump page whe... |
276 277 |
VM_BUG_ON_PAGE(PageLRU(page), page); VM_BUG_ON_PAGE(!PageLocked(page), page); |
56afe477d mm: munlock: bypa... |
278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 |
if (page_mapcount(page) <= 1 && page_evictable(page)) { pagevec_add(pvec, page); if (TestClearPageUnevictable(page)) (*pgrescued)++; unlock_page(page); return true; } return false; } /* * Putback multiple evictable pages to the LRU * * Batched putback of evictable pages that bypasses the per-cpu pvec. Some of * the pages might have meanwhile become unevictable but that is OK. */ static void __putback_lru_fast(struct pagevec *pvec, int pgrescued) { count_vm_events(UNEVICTABLE_PGMUNLOCKED, pagevec_count(pvec)); /* *__pagevec_lru_add() calls release_pages() so we don't call * put_page() explicitly */ __pagevec_lru_add(pvec); count_vm_events(UNEVICTABLE_PGRESCUED, pgrescued); } /* |
7225522bb mm: munlock: batc... |
308 309 310 311 312 313 314 |
* Munlock a batch of pages from the same zone * * The work is split to two main phases. First phase clears the Mlocked flag * and attempts to isolate the pages, all under a single zone lru lock. * The second phase finishes the munlock only for pages where isolation * succeeded. * |
7a8010cd3 mm: munlock: manu... |
315 |
* Note that the pagevec may be modified during the process. |
7225522bb mm: munlock: batc... |
316 317 318 319 320 |
*/ static void __munlock_pagevec(struct pagevec *pvec, struct zone *zone) { int i; int nr = pagevec_count(pvec); |
3b25df93c mm: munlock: fix ... |
321 |
int delta_munlocked; |
56afe477d mm: munlock: bypa... |
322 323 |
struct pagevec pvec_putback; int pgrescued = 0; |
7225522bb mm: munlock: batc... |
324 |
|
3b25df93c mm: munlock: fix ... |
325 |
pagevec_init(&pvec_putback, 0); |
7225522bb mm: munlock: batc... |
326 327 328 329 330 331 |
/* Phase 1: page isolation */ spin_lock_irq(&zone->lru_lock); for (i = 0; i < nr; i++) { struct page *page = pvec->pages[i]; if (TestClearPageMlocked(page)) { |
7225522bb mm: munlock: batc... |
332 |
/* |
01cc2e586 mm: munlock: fix ... |
333 334 |
* We already have pin from follow_page_mask() * so we can spare the get_page() here. |
7225522bb mm: munlock: batc... |
335 |
*/ |
01cc2e586 mm: munlock: fix ... |
336 337 338 339 |
if (__munlock_isolate_lru_page(page, false)) continue; else __munlock_isolation_failed(page); |
7225522bb mm: munlock: batc... |
340 |
} |
01cc2e586 mm: munlock: fix ... |
341 342 343 344 345 346 347 348 349 |
/* * We won't be munlocking this page in the next phase * but we still need to release the follow_page_mask() * pin. We cannot do it under lru_lock however. If it's * the last pin, __page_cache_release() would deadlock. */ pagevec_add(&pvec_putback, pvec->pages[i]); pvec->pages[i] = NULL; |
7225522bb mm: munlock: batc... |
350 |
} |
3b25df93c mm: munlock: fix ... |
351 |
delta_munlocked = -nr + pagevec_count(&pvec_putback); |
1ebb7cc6a mm: munlock: batc... |
352 |
__mod_zone_page_state(zone, NR_MLOCK, delta_munlocked); |
7225522bb mm: munlock: batc... |
353 |
spin_unlock_irq(&zone->lru_lock); |
3b25df93c mm: munlock: fix ... |
354 355 |
/* Now we can release pins of pages that we are not munlocking */ pagevec_release(&pvec_putback); |
56afe477d mm: munlock: bypa... |
356 |
/* Phase 2: page munlock */ |
7225522bb mm: munlock: batc... |
357 358 359 360 361 |
for (i = 0; i < nr; i++) { struct page *page = pvec->pages[i]; if (page) { lock_page(page); |
56afe477d mm: munlock: bypa... |
362 363 |
if (!__putback_lru_fast_prepare(page, &pvec_putback, &pgrescued)) { |
5b40998ae mm: munlock: remo... |
364 365 366 367 368 |
/* * Slow path. We don't want to lose the last * pin before unlock_page() */ get_page(page); /* for putback_lru_page() */ |
56afe477d mm: munlock: bypa... |
369 370 |
__munlock_isolated_page(page); unlock_page(page); |
5b40998ae mm: munlock: remo... |
371 |
put_page(page); /* from follow_page_mask() */ |
56afe477d mm: munlock: bypa... |
372 |
} |
7225522bb mm: munlock: batc... |
373 374 |
} } |
56afe477d mm: munlock: bypa... |
375 |
|
5b40998ae mm: munlock: remo... |
376 377 378 379 |
/* * Phase 3: page putback for pages that qualified for the fast path * This will also call put_page() to return pin from follow_page_mask() */ |
56afe477d mm: munlock: bypa... |
380 381 |
if (pagevec_count(&pvec_putback)) __putback_lru_fast(&pvec_putback, pgrescued); |
7a8010cd3 mm: munlock: manu... |
382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 |
} /* * Fill up pagevec for __munlock_pagevec using pte walk * * The function expects that the struct page corresponding to @start address is * a non-TPH page already pinned and in the @pvec, and that it belongs to @zone. * * The rest of @pvec is filled by subsequent pages within the same pmd and same * zone, as long as the pte's are present and vm_normal_page() succeeds. These * pages also get pinned. * * Returns the address of the next page that should be scanned. This equals * @start + PAGE_SIZE when no page could be added by the pte walk. */ static unsigned long __munlock_pagevec_fill(struct pagevec *pvec, struct vm_area_struct *vma, int zoneid, unsigned long start, unsigned long end) { pte_t *pte; spinlock_t *ptl; /* * Initialize pte walk starting at the already pinned page where we |
eadb41ae8 mm/mlock.c: preve... |
406 407 |
* are sure that there is a pte, as it was pinned under the same * mmap_sem write op. |
7a8010cd3 mm: munlock: manu... |
408 409 |
*/ pte = get_locked_pte(vma->vm_mm, start, &ptl); |
eadb41ae8 mm/mlock.c: preve... |
410 411 412 413 |
/* Make sure we do not cross the page table boundary */ end = pgd_addr_end(start, end); end = pud_addr_end(start, end); end = pmd_addr_end(start, end); |
7a8010cd3 mm: munlock: manu... |
414 415 416 417 418 419 420 421 422 423 424 425 426 427 |
/* The page next to the pinned page is the first we will try to get */ start += PAGE_SIZE; while (start < end) { struct page *page = NULL; pte++; if (pte_present(*pte)) page = vm_normal_page(vma, start, *pte); /* * Break if page could not be obtained or the page's node+zone does not * match */ if (!page || page_zone_id(page) != zoneid) break; |
56afe477d mm: munlock: bypa... |
428 |
|
7a8010cd3 mm: munlock: manu... |
429 430 431 432 433 434 435 436 437 438 439 |
get_page(page); /* * Increase the address that will be returned *before* the * eventual break due to pvec becoming full by adding the page */ start += PAGE_SIZE; if (pagevec_add(pvec, page) == 0) break; } pte_unmap_unlock(pte, ptl); return start; |
7225522bb mm: munlock: batc... |
440 441 442 |
} /* |
ba470de43 mmap: handle mloc... |
443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 |
* munlock_vma_pages_range() - munlock all pages in the vma range.' * @vma - vma containing range to be munlock()ed. * @start - start address in @vma of the range * @end - end of range in @vma. * * For mremap(), munmap() and exit(). * * Called with @vma VM_LOCKED. * * Returns with VM_LOCKED cleared. Callers must be prepared to * deal with this. * * We don't save and restore VM_LOCKED here because pages are * still on lru. In unmap path, pages might be scanned by reclaim * and re-mlocked by try_to_{munlock|unmap} before we unmap and * free them. This will result in freeing mlocked pages. |
b291f0003 mlock: mlocked pa... |
459 |
*/ |
ba470de43 mmap: handle mloc... |
460 |
void munlock_vma_pages_range(struct vm_area_struct *vma, |
408e82b78 mm: munlock use f... |
461 |
unsigned long start, unsigned long end) |
b291f0003 mlock: mlocked pa... |
462 463 |
{ vma->vm_flags &= ~VM_LOCKED; |
408e82b78 mm: munlock use f... |
464 |
|
ff6a6da60 mm: accelerate mu... |
465 |
while (start < end) { |
7a8010cd3 mm: munlock: manu... |
466 |
struct page *page = NULL; |
c424be1cb mm: munlock: fix ... |
467 468 |
unsigned int page_mask; unsigned long page_increm; |
7a8010cd3 mm: munlock: manu... |
469 470 471 |
struct pagevec pvec; struct zone *zone; int zoneid; |
ff6a6da60 mm: accelerate mu... |
472 |
|
7a8010cd3 mm: munlock: manu... |
473 |
pagevec_init(&pvec, 0); |
6e919717c mm: m(un)lock avo... |
474 475 476 477 478 479 480 |
/* * Although FOLL_DUMP is intended for get_dump_page(), * it just so happens that its special treatment of the * ZERO_PAGE (returning an error instead of doing get_page) * suits munlock very well (and if somehow an abnormal page * has sneaked into the range, we won't oops here: great). */ |
ff6a6da60 mm: accelerate mu... |
481 |
page = follow_page_mask(vma, start, FOLL_GET | FOLL_DUMP, |
7a8010cd3 mm: munlock: manu... |
482 |
&page_mask); |
6e919717c mm: m(un)lock avo... |
483 |
if (page && !IS_ERR(page)) { |
7225522bb mm: munlock: batc... |
484 |
if (PageTransHuge(page)) { |
7225522bb mm: munlock: batc... |
485 486 487 488 489 490 491 492 493 494 495 496 |
lock_page(page); /* * Any THP page found by follow_page_mask() may * have gotten split before reaching * munlock_vma_page(), so we need to recompute * the page_mask here. */ page_mask = munlock_vma_page(page); unlock_page(page); put_page(page); /* follow_page_mask() */ } else { /* |
7a8010cd3 mm: munlock: manu... |
497 498 499 500 501 502 503 504 505 506 507 508 509 |
* Non-huge pages are handled in batches via * pagevec. The pin from follow_page_mask() * prevents them from collapsing by THP. */ pagevec_add(&pvec, page); zone = page_zone(page); zoneid = page_zone_id(page); /* * Try to fill the rest of pagevec using fast * pte walk. This will also update start to * the next page to process. Then munlock the * pagevec. |
7225522bb mm: munlock: batc... |
510 |
*/ |
7a8010cd3 mm: munlock: manu... |
511 512 513 514 |
start = __munlock_pagevec_fill(&pvec, vma, zoneid, start, end); __munlock_pagevec(&pvec, zone); goto next; |
7225522bb mm: munlock: batc... |
515 |
} |
408e82b78 mm: munlock use f... |
516 |
} |
c424be1cb mm: munlock: fix ... |
517 518 519 |
/* It's a bug to munlock in the middle of a THP page */ VM_BUG_ON((start >> PAGE_SHIFT) & page_mask); page_increm = 1 + page_mask; |
ff6a6da60 mm: accelerate mu... |
520 |
start += page_increm * PAGE_SIZE; |
7a8010cd3 mm: munlock: manu... |
521 |
next: |
408e82b78 mm: munlock use f... |
522 523 |
cond_resched(); } |
b291f0003 mlock: mlocked pa... |
524 525 526 527 528 529 530 |
} /* * mlock_fixup - handle mlock[all]/munlock[all] requests. * * Filters out "special" vmas -- VM_LOCKED never gets set for these, and * munlock is a no-op. However, for some special vmas, we go ahead and |
cea10a19b mm: directly use ... |
531 |
* populate the ptes. |
b291f0003 mlock: mlocked pa... |
532 533 534 |
* * For vmas that pass the filters, merge/split as appropriate. */ |
1da177e4c Linux-2.6.12-rc2 |
535 |
static int mlock_fixup(struct vm_area_struct *vma, struct vm_area_struct **prev, |
ca16d140a mm: don't access ... |
536 |
unsigned long start, unsigned long end, vm_flags_t newflags) |
1da177e4c Linux-2.6.12-rc2 |
537 |
{ |
b291f0003 mlock: mlocked pa... |
538 |
struct mm_struct *mm = vma->vm_mm; |
1da177e4c Linux-2.6.12-rc2 |
539 |
pgoff_t pgoff; |
b291f0003 mlock: mlocked pa... |
540 |
int nr_pages; |
1da177e4c Linux-2.6.12-rc2 |
541 |
int ret = 0; |
ca16d140a mm: don't access ... |
542 |
int lock = !!(newflags & VM_LOCKED); |
1da177e4c Linux-2.6.12-rc2 |
543 |
|
fed067da4 mlock: only hold ... |
544 |
if (newflags == vma->vm_flags || (vma->vm_flags & VM_SPECIAL) || |
31db58b3a mm: arch: make ge... |
545 |
is_vm_hugetlb_page(vma) || vma == get_gate_vma(current->mm)) |
b291f0003 mlock: mlocked pa... |
546 |
goto out; /* don't set VM_LOCKED, don't count */ |
1da177e4c Linux-2.6.12-rc2 |
547 548 549 550 551 552 553 |
pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT); *prev = vma_merge(mm, *prev, start, end, newflags, vma->anon_vma, vma->vm_file, pgoff, vma_policy(vma)); if (*prev) { vma = *prev; goto success; } |
1da177e4c Linux-2.6.12-rc2 |
554 555 556 557 558 559 560 561 562 563 564 565 566 567 |
if (start != vma->vm_start) { ret = split_vma(mm, vma, start, 1); if (ret) goto out; } if (end != vma->vm_end) { ret = split_vma(mm, vma, end, 0); if (ret) goto out; } success: /* |
b291f0003 mlock: mlocked pa... |
568 569 570 571 572 573 574 575 |
* Keep track of amount of locked VM. */ nr_pages = (end - start) >> PAGE_SHIFT; if (!lock) nr_pages = -nr_pages; mm->locked_vm += nr_pages; /* |
1da177e4c Linux-2.6.12-rc2 |
576 577 |
* vm_flags is protected by the mmap_sem held in write mode. * It's okay if try_to_unmap_one unmaps a page just after we |
b291f0003 mlock: mlocked pa... |
578 |
* set VM_LOCKED, __mlock_vma_pages_range will bring it back. |
1da177e4c Linux-2.6.12-rc2 |
579 |
*/ |
1da177e4c Linux-2.6.12-rc2 |
580 |
|
fed067da4 mlock: only hold ... |
581 |
if (lock) |
408e82b78 mm: munlock use f... |
582 |
vma->vm_flags = newflags; |
fed067da4 mlock: only hold ... |
583 |
else |
408e82b78 mm: munlock use f... |
584 |
munlock_vma_pages_range(vma, start, end); |
1da177e4c Linux-2.6.12-rc2 |
585 |
|
1da177e4c Linux-2.6.12-rc2 |
586 |
out: |
b291f0003 mlock: mlocked pa... |
587 |
*prev = vma; |
1da177e4c Linux-2.6.12-rc2 |
588 589 590 591 592 593 594 595 |
return ret; } static int do_mlock(unsigned long start, size_t len, int on) { unsigned long nstart, end, tmp; struct vm_area_struct * vma, * prev; int error; |
fed067da4 mlock: only hold ... |
596 597 |
VM_BUG_ON(start & ~PAGE_MASK); VM_BUG_ON(len != PAGE_ALIGN(len)); |
1da177e4c Linux-2.6.12-rc2 |
598 599 600 601 602 |
end = start + len; if (end < start) return -EINVAL; if (end == start) return 0; |
097d59106 vm: avoid using f... |
603 |
vma = find_vma(current->mm, start); |
1da177e4c Linux-2.6.12-rc2 |
604 605 |
if (!vma || vma->vm_start > start) return -ENOMEM; |
097d59106 vm: avoid using f... |
606 |
prev = vma->vm_prev; |
1da177e4c Linux-2.6.12-rc2 |
607 608 609 610 |
if (start > vma->vm_start) prev = vma; for (nstart = start ; ; ) { |
ca16d140a mm: don't access ... |
611 |
vm_flags_t newflags; |
1da177e4c Linux-2.6.12-rc2 |
612 613 |
/* Here we know that vma->vm_start <= nstart < vma->vm_end. */ |
186930500 mm: introduce VM_... |
614 615 |
newflags = vma->vm_flags & ~VM_LOCKED; if (on) |
09a9f1d27 Revert "mm: intro... |
616 |
newflags |= VM_LOCKED; |
1da177e4c Linux-2.6.12-rc2 |
617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 |
tmp = vma->vm_end; if (tmp > end) tmp = end; error = mlock_fixup(vma, &prev, nstart, tmp, newflags); if (error) break; nstart = tmp; if (nstart < prev->vm_end) nstart = prev->vm_end; if (nstart >= end) break; vma = prev->vm_next; if (!vma || vma->vm_start != nstart) { error = -ENOMEM; break; } } return error; } |
bebeb3d68 mm: introduce mm_... |
638 639 640 641 642 643 644 645 |
/* * __mm_populate - populate and/or mlock pages within a range of address space. * * This is used to implement mlock() and the MAP_POPULATE / MAP_LOCKED mmap * flags. VMAs must be already marked with the desired vm_flags, and * mmap_sem must not be held. */ int __mm_populate(unsigned long start, unsigned long len, int ignore_errors) |
fed067da4 mlock: only hold ... |
646 647 648 649 |
{ struct mm_struct *mm = current->mm; unsigned long end, nstart, nend; struct vm_area_struct *vma = NULL; |
53a7706d5 mlock: do not hol... |
650 |
int locked = 0; |
28a35716d mm: use long type... |
651 |
long ret = 0; |
fed067da4 mlock: only hold ... |
652 653 654 655 |
VM_BUG_ON(start & ~PAGE_MASK); VM_BUG_ON(len != PAGE_ALIGN(len)); end = start + len; |
fed067da4 mlock: only hold ... |
656 657 658 659 660 |
for (nstart = start; nstart < end; nstart = nend) { /* * We want to fault in pages for [nstart; end) address range. * Find first corresponding VMA. */ |
53a7706d5 mlock: do not hol... |
661 662 663 |
if (!locked) { locked = 1; down_read(&mm->mmap_sem); |
fed067da4 mlock: only hold ... |
664 |
vma = find_vma(mm, nstart); |
53a7706d5 mlock: do not hol... |
665 |
} else if (nstart >= vma->vm_end) |
fed067da4 mlock: only hold ... |
666 667 668 669 670 671 672 673 |
vma = vma->vm_next; if (!vma || vma->vm_start >= end) break; /* * Set [nstart; nend) to intersection of desired address * range with the first VMA. Also, skip undesirable VMA types. */ nend = min(end, vma->vm_end); |
09a9f1d27 Revert "mm: intro... |
674 |
if (vma->vm_flags & (VM_IO | VM_PFNMAP)) |
fed067da4 mlock: only hold ... |
675 676 677 678 |
continue; if (nstart < vma->vm_start) nstart = vma->vm_start; /* |
53a7706d5 mlock: do not hol... |
679 680 681 |
* Now fault in a range of pages. __mlock_vma_pages_range() * double checks the vma flags, so that it won't mlock pages * if the vma was already munlocked. |
fed067da4 mlock: only hold ... |
682 |
*/ |
53a7706d5 mlock: do not hol... |
683 684 685 686 687 688 |
ret = __mlock_vma_pages_range(vma, nstart, nend, &locked); if (ret < 0) { if (ignore_errors) { ret = 0; continue; /* continue at next VMA */ } |
5fdb20021 mm: move VM_LOCKE... |
689 690 691 |
ret = __mlock_posix_error_return(ret); break; } |
53a7706d5 mlock: do not hol... |
692 693 |
nend = nstart + ret * PAGE_SIZE; ret = 0; |
fed067da4 mlock: only hold ... |
694 |
} |
53a7706d5 mlock: do not hol... |
695 696 |
if (locked) up_read(&mm->mmap_sem); |
fed067da4 mlock: only hold ... |
697 698 |
return ret; /* 0 or negative error code */ } |
6a6160a7b [CVE-2009-0029] S... |
699 |
SYSCALL_DEFINE2(mlock, unsigned long, start, size_t, len) |
1da177e4c Linux-2.6.12-rc2 |
700 701 702 703 704 705 706 |
{ unsigned long locked; unsigned long lock_limit; int error = -ENOMEM; if (!can_do_mlock()) return -EPERM; |
8891d6da1 mm: remove lru_ad... |
707 |
lru_add_drain_all(); /* flush pagevec */ |
1da177e4c Linux-2.6.12-rc2 |
708 709 |
len = PAGE_ALIGN(len + (start & ~PAGE_MASK)); start &= PAGE_MASK; |
59e99e5b9 mm: use rlimit he... |
710 |
lock_limit = rlimit(RLIMIT_MEMLOCK); |
1da177e4c Linux-2.6.12-rc2 |
711 |
lock_limit >>= PAGE_SHIFT; |
1f1cd7054 mm/mlock: prepare... |
712 713 714 715 716 |
locked = len >> PAGE_SHIFT; down_write(¤t->mm->mmap_sem); locked += current->mm->locked_vm; |
1da177e4c Linux-2.6.12-rc2 |
717 718 719 720 |
/* check against resource limits */ if ((locked <= lock_limit) || capable(CAP_IPC_LOCK)) error = do_mlock(start, len, 1); |
1f1cd7054 mm/mlock: prepare... |
721 |
|
1da177e4c Linux-2.6.12-rc2 |
722 |
up_write(¤t->mm->mmap_sem); |
fed067da4 mlock: only hold ... |
723 |
if (!error) |
bebeb3d68 mm: introduce mm_... |
724 |
error = __mm_populate(start, len, 0); |
1da177e4c Linux-2.6.12-rc2 |
725 726 |
return error; } |
6a6160a7b [CVE-2009-0029] S... |
727 |
SYSCALL_DEFINE2(munlock, unsigned long, start, size_t, len) |
1da177e4c Linux-2.6.12-rc2 |
728 729 |
{ int ret; |
1da177e4c Linux-2.6.12-rc2 |
730 731 |
len = PAGE_ALIGN(len + (start & ~PAGE_MASK)); start &= PAGE_MASK; |
1f1cd7054 mm/mlock: prepare... |
732 733 |
down_write(¤t->mm->mmap_sem); |
1da177e4c Linux-2.6.12-rc2 |
734 735 |
ret = do_mlock(start, len, 0); up_write(¤t->mm->mmap_sem); |
1f1cd7054 mm/mlock: prepare... |
736 |
|
1da177e4c Linux-2.6.12-rc2 |
737 738 739 740 741 742 |
return ret; } static int do_mlockall(int flags) { struct vm_area_struct * vma, * prev = NULL; |
1da177e4c Linux-2.6.12-rc2 |
743 744 |
if (flags & MCL_FUTURE) |
09a9f1d27 Revert "mm: intro... |
745 |
current->mm->def_flags |= VM_LOCKED; |
9977f0f16 mm: don't overwri... |
746 |
else |
09a9f1d27 Revert "mm: intro... |
747 |
current->mm->def_flags &= ~VM_LOCKED; |
1da177e4c Linux-2.6.12-rc2 |
748 749 750 751 |
if (flags == MCL_FUTURE) goto out; for (vma = current->mm->mmap; vma ; vma = prev->vm_next) { |
ca16d140a mm: don't access ... |
752 |
vm_flags_t newflags; |
1da177e4c Linux-2.6.12-rc2 |
753 |
|
186930500 mm: introduce VM_... |
754 755 |
newflags = vma->vm_flags & ~VM_LOCKED; if (flags & MCL_CURRENT) |
09a9f1d27 Revert "mm: intro... |
756 |
newflags |= VM_LOCKED; |
1da177e4c Linux-2.6.12-rc2 |
757 758 759 |
/* Ignore errors */ mlock_fixup(vma, &prev, vma->vm_start, vma->vm_end, newflags); |
22356f447 mm: Place preempt... |
760 |
cond_resched(); |
1da177e4c Linux-2.6.12-rc2 |
761 762 763 764 |
} out: return 0; } |
3480b2574 [CVE-2009-0029] S... |
765 |
SYSCALL_DEFINE1(mlockall, int, flags) |
1da177e4c Linux-2.6.12-rc2 |
766 767 768 769 770 771 772 773 774 775 |
{ unsigned long lock_limit; int ret = -EINVAL; if (!flags || (flags & ~(MCL_CURRENT | MCL_FUTURE))) goto out; ret = -EPERM; if (!can_do_mlock()) goto out; |
df9d6985b mm: do not drain ... |
776 777 |
if (flags & MCL_CURRENT) lru_add_drain_all(); /* flush pagevec */ |
8891d6da1 mm: remove lru_ad... |
778 |
|
59e99e5b9 mm: use rlimit he... |
779 |
lock_limit = rlimit(RLIMIT_MEMLOCK); |
1da177e4c Linux-2.6.12-rc2 |
780 781 782 |
lock_limit >>= PAGE_SHIFT; ret = -ENOMEM; |
1f1cd7054 mm/mlock: prepare... |
783 |
down_write(¤t->mm->mmap_sem); |
1da177e4c Linux-2.6.12-rc2 |
784 785 786 787 |
if (!(flags & MCL_CURRENT) || (current->mm->total_vm <= lock_limit) || capable(CAP_IPC_LOCK)) ret = do_mlockall(flags); up_write(¤t->mm->mmap_sem); |
bebeb3d68 mm: introduce mm_... |
788 789 |
if (!ret && (flags & MCL_CURRENT)) mm_populate(0, TASK_SIZE); |
1da177e4c Linux-2.6.12-rc2 |
790 791 792 |
out: return ret; } |
3480b2574 [CVE-2009-0029] S... |
793 |
SYSCALL_DEFINE0(munlockall) |
1da177e4c Linux-2.6.12-rc2 |
794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 |
{ int ret; down_write(¤t->mm->mmap_sem); ret = do_mlockall(0); up_write(¤t->mm->mmap_sem); return ret; } /* * Objects with different lifetime than processes (SHM_LOCK and SHM_HUGETLB * shm segments) get accounted against the user_struct instead. */ static DEFINE_SPINLOCK(shmlock_user_lock); int user_shm_lock(size_t size, struct user_struct *user) { unsigned long lock_limit, locked; int allowed = 0; locked = (size + PAGE_SIZE - 1) >> PAGE_SHIFT; |
59e99e5b9 mm: use rlimit he... |
815 |
lock_limit = rlimit(RLIMIT_MEMLOCK); |
5ed44a401 do not limit lock... |
816 817 |
if (lock_limit == RLIM_INFINITY) allowed = 1; |
1da177e4c Linux-2.6.12-rc2 |
818 819 |
lock_limit >>= PAGE_SHIFT; spin_lock(&shmlock_user_lock); |
5ed44a401 do not limit lock... |
820 821 |
if (!allowed && locked + user->locked_shm > lock_limit && !capable(CAP_IPC_LOCK)) |
1da177e4c Linux-2.6.12-rc2 |
822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 |
goto out; get_uid(user); user->locked_shm += locked; allowed = 1; out: spin_unlock(&shmlock_user_lock); return allowed; } void user_shm_unlock(size_t size, struct user_struct *user) { spin_lock(&shmlock_user_lock); user->locked_shm -= (size + PAGE_SIZE - 1) >> PAGE_SHIFT; spin_unlock(&shmlock_user_lock); free_uid(user); } |