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 |
|
7f43add45 mm/mlock.c: chang... |
26 |
bool can_do_mlock(void) |
e8edc6e03 Detach sched.h fr... |
27 |
{ |
59e99e5b9 mm: use rlimit he... |
28 |
if (rlimit(RLIMIT_MEMLOCK) != 0) |
7f43add45 mm/mlock.c: chang... |
29 |
return true; |
a5a6579db mm: reorder can_d... |
30 |
if (capable(CAP_IPC_LOCK)) |
7f43add45 mm/mlock.c: chang... |
31 32 |
return true; return false; |
e8edc6e03 Detach sched.h fr... |
33 34 |
} 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)); |
e90309c9f thp: allow mlocke... |
83 84 |
VM_BUG_ON_PAGE(PageTail(page), page); VM_BUG_ON_PAGE(PageCompound(page) && PageDoubleMap(page), page); |
5344b7e64 vmstat: mlocked p... |
85 |
if (!TestSetPageMlocked(page)) { |
8449d21fb mm, thp: fix mloc... |
86 87 |
mod_zone_page_state(page_zone(page), NR_MLOCK, hpage_nr_pages(page)); |
5344b7e64 vmstat: mlocked p... |
88 89 90 91 |
count_vm_event(UNEVICTABLE_PGMLOCKED); if (!isolate_lru_page(page)) putback_lru_page(page); } |
b291f0003 mlock: mlocked pa... |
92 |
} |
7225522bb mm: munlock: batc... |
93 |
/* |
01cc2e586 mm: munlock: fix ... |
94 95 96 97 98 99 100 |
* 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; |
599d0c954 mm, vmscan: move ... |
101 |
lruvec = mem_cgroup_page_lruvec(page, page_pgdat(page)); |
01cc2e586 mm: munlock: fix ... |
102 103 104 105 106 107 108 109 110 111 112 |
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... |
113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 |
* 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 ... |
148 |
__count_vm_event(UNEVICTABLE_PGSTRANDED); |
7225522bb mm: munlock: batc... |
149 |
else |
01cc2e586 mm: munlock: fix ... |
150 |
__count_vm_event(UNEVICTABLE_PGMUNLOCKED); |
7225522bb mm: munlock: batc... |
151 |
} |
6927c1dd9 mlock: replace st... |
152 153 |
/** * munlock_vma_page - munlock a vma page |
c424be1cb mm: munlock: fix ... |
154 155 156 157 |
* @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... |
158 |
* |
6927c1dd9 mlock: replace st... |
159 160 161 162 163 164 165 166 167 168 |
* 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... |
169 |
*/ |
ff6a6da60 mm: accelerate mu... |
170 |
unsigned int munlock_vma_page(struct page *page) |
b291f0003 mlock: mlocked pa... |
171 |
{ |
7162a1e87 mm: fix mlock acc... |
172 |
int nr_pages; |
01cc2e586 mm: munlock: fix ... |
173 |
struct zone *zone = page_zone(page); |
ff6a6da60 mm: accelerate mu... |
174 |
|
57e68e9cd mm: try_to_unmap_... |
175 |
/* For try_to_munlock() and to serialize with page migration */ |
b291f0003 mlock: mlocked pa... |
176 |
BUG_ON(!PageLocked(page)); |
e90309c9f thp: allow mlocke... |
177 |
VM_BUG_ON_PAGE(PageTail(page), page); |
c424be1cb mm: munlock: fix ... |
178 |
/* |
01cc2e586 mm: munlock: fix ... |
179 180 181 |
* 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 ... |
182 |
*/ |
a52633d8e mm, vmscan: move ... |
183 |
spin_lock_irq(zone_lru_lock(zone)); |
01cc2e586 mm: munlock: fix ... |
184 |
|
655548bf6 thp: fix corner c... |
185 186 187 |
if (!TestClearPageMlocked(page)) { /* Potentially, PTE-mapped THP: do not skip the rest PTEs */ nr_pages = 1; |
01cc2e586 mm: munlock: fix ... |
188 |
goto unlock_out; |
655548bf6 thp: fix corner c... |
189 |
} |
01cc2e586 mm: munlock: fix ... |
190 |
|
655548bf6 thp: fix corner c... |
191 |
nr_pages = hpage_nr_pages(page); |
01cc2e586 mm: munlock: fix ... |
192 193 194 |
__mod_zone_page_state(zone, NR_MLOCK, -nr_pages); if (__munlock_isolate_lru_page(page, true)) { |
a52633d8e mm, vmscan: move ... |
195 |
spin_unlock_irq(zone_lru_lock(zone)); |
01cc2e586 mm: munlock: fix ... |
196 197 198 199 200 201 |
__munlock_isolated_page(page); goto out; } __munlock_isolation_failed(page); unlock_out: |
a52633d8e mm, vmscan: move ... |
202 |
spin_unlock_irq(zone_lru_lock(zone)); |
01cc2e586 mm: munlock: fix ... |
203 204 |
out: |
c424be1cb mm: munlock: fix ... |
205 |
return nr_pages - 1; |
b291f0003 mlock: mlocked pa... |
206 |
} |
9978ad583 mlock: make mlock... |
207 208 209 210 211 212 213 214 215 216 |
/* * 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... |
217 |
} |
b291f0003 mlock: mlocked pa... |
218 |
/* |
56afe477d mm: munlock: bypa... |
219 220 221 222 223 224 225 226 227 228 229 230 231 232 |
* 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... |
233 234 |
VM_BUG_ON_PAGE(PageLRU(page), page); VM_BUG_ON_PAGE(!PageLocked(page), page); |
56afe477d mm: munlock: bypa... |
235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 |
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... |
265 266 267 268 269 270 271 |
* 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... |
272 |
* Note that the pagevec may be modified during the process. |
7225522bb mm: munlock: batc... |
273 274 275 276 277 |
*/ static void __munlock_pagevec(struct pagevec *pvec, struct zone *zone) { int i; int nr = pagevec_count(pvec); |
3b25df93c mm: munlock: fix ... |
278 |
int delta_munlocked; |
56afe477d mm: munlock: bypa... |
279 280 |
struct pagevec pvec_putback; int pgrescued = 0; |
7225522bb mm: munlock: batc... |
281 |
|
3b25df93c mm: munlock: fix ... |
282 |
pagevec_init(&pvec_putback, 0); |
7225522bb mm: munlock: batc... |
283 |
/* Phase 1: page isolation */ |
a52633d8e mm, vmscan: move ... |
284 |
spin_lock_irq(zone_lru_lock(zone)); |
7225522bb mm: munlock: batc... |
285 286 287 288 |
for (i = 0; i < nr; i++) { struct page *page = pvec->pages[i]; if (TestClearPageMlocked(page)) { |
7225522bb mm: munlock: batc... |
289 |
/* |
01cc2e586 mm: munlock: fix ... |
290 291 |
* We already have pin from follow_page_mask() * so we can spare the get_page() here. |
7225522bb mm: munlock: batc... |
292 |
*/ |
01cc2e586 mm: munlock: fix ... |
293 294 295 296 |
if (__munlock_isolate_lru_page(page, false)) continue; else __munlock_isolation_failed(page); |
7225522bb mm: munlock: batc... |
297 |
} |
01cc2e586 mm: munlock: fix ... |
298 299 300 301 302 303 304 305 306 |
/* * 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... |
307 |
} |
3b25df93c mm: munlock: fix ... |
308 |
delta_munlocked = -nr + pagevec_count(&pvec_putback); |
1ebb7cc6a mm: munlock: batc... |
309 |
__mod_zone_page_state(zone, NR_MLOCK, delta_munlocked); |
a52633d8e mm, vmscan: move ... |
310 |
spin_unlock_irq(zone_lru_lock(zone)); |
7225522bb mm: munlock: batc... |
311 |
|
3b25df93c mm: munlock: fix ... |
312 313 |
/* Now we can release pins of pages that we are not munlocking */ pagevec_release(&pvec_putback); |
56afe477d mm: munlock: bypa... |
314 |
/* Phase 2: page munlock */ |
7225522bb mm: munlock: batc... |
315 316 317 318 319 |
for (i = 0; i < nr; i++) { struct page *page = pvec->pages[i]; if (page) { lock_page(page); |
56afe477d mm: munlock: bypa... |
320 321 |
if (!__putback_lru_fast_prepare(page, &pvec_putback, &pgrescued)) { |
5b40998ae mm: munlock: remo... |
322 323 324 325 326 |
/* * 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... |
327 328 |
__munlock_isolated_page(page); unlock_page(page); |
5b40998ae mm: munlock: remo... |
329 |
put_page(page); /* from follow_page_mask() */ |
56afe477d mm: munlock: bypa... |
330 |
} |
7225522bb mm: munlock: batc... |
331 332 |
} } |
56afe477d mm: munlock: bypa... |
333 |
|
5b40998ae mm: munlock: remo... |
334 335 336 337 |
/* * 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... |
338 339 |
if (pagevec_count(&pvec_putback)) __putback_lru_fast(&pvec_putback, pgrescued); |
7a8010cd3 mm: munlock: manu... |
340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 |
} /* * 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... |
364 365 |
* are sure that there is a pte, as it was pinned under the same * mmap_sem write op. |
7a8010cd3 mm: munlock: manu... |
366 367 |
*/ pte = get_locked_pte(vma->vm_mm, start, &ptl); |
eadb41ae8 mm/mlock.c: preve... |
368 369 370 371 |
/* 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... |
372 373 374 375 376 377 378 379 380 381 382 383 384 385 |
/* 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... |
386 |
|
e90309c9f thp: allow mlocke... |
387 388 389 390 391 392 |
/* * Do not use pagevec for PTE-mapped THP, * munlock_vma_pages_range() will handle them. */ if (PageTransCompound(page)) break; |
7a8010cd3 mm: munlock: manu... |
393 394 395 396 397 398 399 400 401 402 403 |
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... |
404 405 406 |
} /* |
ba470de43 mmap: handle mloc... |
407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 |
* 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... |
423 |
*/ |
ba470de43 mmap: handle mloc... |
424 |
void munlock_vma_pages_range(struct vm_area_struct *vma, |
408e82b78 mm: munlock use f... |
425 |
unsigned long start, unsigned long end) |
b291f0003 mlock: mlocked pa... |
426 |
{ |
de60f5f10 mm: introduce VM_... |
427 |
vma->vm_flags &= VM_LOCKED_CLEAR_MASK; |
408e82b78 mm: munlock use f... |
428 |
|
ff6a6da60 mm: accelerate mu... |
429 |
while (start < end) { |
ab7a5af7f mm/mlock.c: drop ... |
430 |
struct page *page; |
c424be1cb mm: munlock: fix ... |
431 432 |
unsigned int page_mask; unsigned long page_increm; |
7a8010cd3 mm: munlock: manu... |
433 434 435 |
struct pagevec pvec; struct zone *zone; int zoneid; |
ff6a6da60 mm: accelerate mu... |
436 |
|
7a8010cd3 mm: munlock: manu... |
437 |
pagevec_init(&pvec, 0); |
6e919717c mm: m(un)lock avo... |
438 439 440 441 442 443 444 |
/* * 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... |
445 |
page = follow_page_mask(vma, start, FOLL_GET | FOLL_DUMP, |
7a8010cd3 mm: munlock: manu... |
446 |
&page_mask); |
e90309c9f thp: allow mlocke... |
447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 |
if (page && !IS_ERR(page)) { if (PageTransTail(page)) { VM_BUG_ON_PAGE(PageMlocked(page), page); put_page(page); /* follow_page_mask() */ } else if (PageTransHuge(page)) { 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 { /* * 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); |
7a8010cd3 mm: munlock: manu... |
471 |
|
e90309c9f thp: allow mlocke... |
472 473 474 475 476 477 478 479 480 481 482 |
/* * 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. */ start = __munlock_pagevec_fill(&pvec, vma, zoneid, start, end); __munlock_pagevec(&pvec, zone); goto next; } |
408e82b78 mm: munlock use f... |
483 |
} |
c424be1cb mm: munlock: fix ... |
484 |
page_increm = 1 + page_mask; |
ff6a6da60 mm: accelerate mu... |
485 |
start += page_increm * PAGE_SIZE; |
7a8010cd3 mm: munlock: manu... |
486 |
next: |
408e82b78 mm: munlock use f... |
487 488 |
cond_resched(); } |
b291f0003 mlock: mlocked pa... |
489 490 491 492 493 494 495 |
} /* * 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 ... |
496 |
* populate the ptes. |
b291f0003 mlock: mlocked pa... |
497 498 499 |
* * For vmas that pass the filters, merge/split as appropriate. */ |
1da177e4c Linux-2.6.12-rc2 |
500 |
static int mlock_fixup(struct vm_area_struct *vma, struct vm_area_struct **prev, |
ca16d140a mm: don't access ... |
501 |
unsigned long start, unsigned long end, vm_flags_t newflags) |
1da177e4c Linux-2.6.12-rc2 |
502 |
{ |
b291f0003 mlock: mlocked pa... |
503 |
struct mm_struct *mm = vma->vm_mm; |
1da177e4c Linux-2.6.12-rc2 |
504 |
pgoff_t pgoff; |
b291f0003 mlock: mlocked pa... |
505 |
int nr_pages; |
1da177e4c Linux-2.6.12-rc2 |
506 |
int ret = 0; |
ca16d140a mm: don't access ... |
507 |
int lock = !!(newflags & VM_LOCKED); |
b155b4fde mm: mlock: avoid ... |
508 |
vm_flags_t old_flags = vma->vm_flags; |
1da177e4c Linux-2.6.12-rc2 |
509 |
|
fed067da4 mlock: only hold ... |
510 |
if (newflags == vma->vm_flags || (vma->vm_flags & VM_SPECIAL) || |
31db58b3a mm: arch: make ge... |
511 |
is_vm_hugetlb_page(vma) || vma == get_gate_vma(current->mm)) |
b0f205c2a mm: mlock: add ml... |
512 513 |
/* don't set VM_LOCKED or VM_LOCKONFAULT and don't count */ goto out; |
b291f0003 mlock: mlocked pa... |
514 |
|
1da177e4c Linux-2.6.12-rc2 |
515 516 |
pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT); *prev = vma_merge(mm, *prev, start, end, newflags, vma->anon_vma, |
19a809afe userfaultfd: teac... |
517 518 |
vma->vm_file, pgoff, vma_policy(vma), vma->vm_userfaultfd_ctx); |
1da177e4c Linux-2.6.12-rc2 |
519 520 521 522 |
if (*prev) { vma = *prev; goto success; } |
1da177e4c Linux-2.6.12-rc2 |
523 524 525 526 527 528 529 530 531 532 533 534 535 536 |
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... |
537 538 539 540 541 |
* Keep track of amount of locked VM. */ nr_pages = (end - start) >> PAGE_SHIFT; if (!lock) nr_pages = -nr_pages; |
b155b4fde mm: mlock: avoid ... |
542 543 |
else if (old_flags & VM_LOCKED) nr_pages = 0; |
b291f0003 mlock: mlocked pa... |
544 545 546 |
mm->locked_vm += nr_pages; /* |
1da177e4c Linux-2.6.12-rc2 |
547 548 |
* 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 |
fc05f5662 mm: rename __mloc... |
549 |
* set VM_LOCKED, populate_vma_page_range will bring it back. |
1da177e4c Linux-2.6.12-rc2 |
550 |
*/ |
1da177e4c Linux-2.6.12-rc2 |
551 |
|
fed067da4 mlock: only hold ... |
552 |
if (lock) |
408e82b78 mm: munlock use f... |
553 |
vma->vm_flags = newflags; |
fed067da4 mlock: only hold ... |
554 |
else |
408e82b78 mm: munlock use f... |
555 |
munlock_vma_pages_range(vma, start, end); |
1da177e4c Linux-2.6.12-rc2 |
556 |
|
1da177e4c Linux-2.6.12-rc2 |
557 |
out: |
b291f0003 mlock: mlocked pa... |
558 |
*prev = vma; |
1da177e4c Linux-2.6.12-rc2 |
559 560 |
return ret; } |
1aab92ec3 mm: mlock: refact... |
561 562 |
static int apply_vma_lock_flags(unsigned long start, size_t len, vm_flags_t flags) |
1da177e4c Linux-2.6.12-rc2 |
563 564 565 566 |
{ unsigned long nstart, end, tmp; struct vm_area_struct * vma, * prev; int error; |
8fd9e4883 mm/mlock: use off... |
567 |
VM_BUG_ON(offset_in_page(start)); |
fed067da4 mlock: only hold ... |
568 |
VM_BUG_ON(len != PAGE_ALIGN(len)); |
1da177e4c Linux-2.6.12-rc2 |
569 570 571 572 573 |
end = start + len; if (end < start) return -EINVAL; if (end == start) return 0; |
097d59106 vm: avoid using f... |
574 |
vma = find_vma(current->mm, start); |
1da177e4c Linux-2.6.12-rc2 |
575 576 |
if (!vma || vma->vm_start > start) return -ENOMEM; |
097d59106 vm: avoid using f... |
577 |
prev = vma->vm_prev; |
1da177e4c Linux-2.6.12-rc2 |
578 579 580 581 |
if (start > vma->vm_start) prev = vma; for (nstart = start ; ; ) { |
b0f205c2a mm: mlock: add ml... |
582 |
vm_flags_t newflags = vma->vm_flags & VM_LOCKED_CLEAR_MASK; |
1da177e4c Linux-2.6.12-rc2 |
583 |
|
1aab92ec3 mm: mlock: refact... |
584 |
newflags |= flags; |
1da177e4c Linux-2.6.12-rc2 |
585 |
|
1aab92ec3 mm: mlock: refact... |
586 |
/* Here we know that vma->vm_start <= nstart < vma->vm_end. */ |
1da177e4c Linux-2.6.12-rc2 |
587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 |
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; } |
0cf2f6f6d mm: mlock: check ... |
607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 |
/* * Go through vma areas and sum size of mlocked * vma pages, as return value. * Note deferred memory locking case(mlock2(,,MLOCK_ONFAULT) * is also counted. * Return value: previously mlocked page counts */ static int count_mm_mlocked_page_nr(struct mm_struct *mm, unsigned long start, size_t len) { struct vm_area_struct *vma; int count = 0; if (mm == NULL) mm = current->mm; vma = find_vma(mm, start); if (vma == NULL) vma = mm->mmap; for (; vma ; vma = vma->vm_next) { if (start >= vma->vm_end) continue; if (start + len <= vma->vm_start) break; if (vma->vm_flags & VM_LOCKED) { if (start > vma->vm_start) count -= (start - vma->vm_start); if (start + len < vma->vm_end) { count += start + len - vma->vm_start; break; } count += vma->vm_end - vma->vm_start; } } return count >> PAGE_SHIFT; } |
dc0ef0df7 mm: make mmap_sem... |
645 |
static __must_check int do_mlock(unsigned long start, size_t len, vm_flags_t flags) |
1da177e4c Linux-2.6.12-rc2 |
646 647 648 649 650 651 652 |
{ unsigned long locked; unsigned long lock_limit; int error = -ENOMEM; if (!can_do_mlock()) return -EPERM; |
8891d6da1 mm: remove lru_ad... |
653 |
lru_add_drain_all(); /* flush pagevec */ |
8fd9e4883 mm/mlock: use off... |
654 |
len = PAGE_ALIGN(len + (offset_in_page(start))); |
1da177e4c Linux-2.6.12-rc2 |
655 |
start &= PAGE_MASK; |
59e99e5b9 mm: use rlimit he... |
656 |
lock_limit = rlimit(RLIMIT_MEMLOCK); |
1da177e4c Linux-2.6.12-rc2 |
657 |
lock_limit >>= PAGE_SHIFT; |
1f1cd7054 mm/mlock: prepare... |
658 |
locked = len >> PAGE_SHIFT; |
dc0ef0df7 mm: make mmap_sem... |
659 660 |
if (down_write_killable(¤t->mm->mmap_sem)) return -EINTR; |
1f1cd7054 mm/mlock: prepare... |
661 662 |
locked += current->mm->locked_vm; |
0cf2f6f6d mm: mlock: check ... |
663 664 665 666 667 668 669 670 671 672 |
if ((locked > lock_limit) && (!capable(CAP_IPC_LOCK))) { /* * It is possible that the regions requested intersect with * previously mlocked areas, that part area in "mm->locked_vm" * should not be counted to new mlock increment count. So check * and adjust locked count if necessary. */ locked -= count_mm_mlocked_page_nr(current->mm, start, len); } |
1da177e4c Linux-2.6.12-rc2 |
673 674 675 |
/* check against resource limits */ if ((locked <= lock_limit) || capable(CAP_IPC_LOCK)) |
1aab92ec3 mm: mlock: refact... |
676 |
error = apply_vma_lock_flags(start, len, flags); |
1f1cd7054 mm/mlock: prepare... |
677 |
|
1da177e4c Linux-2.6.12-rc2 |
678 |
up_write(¤t->mm->mmap_sem); |
c561259ca mm: move gup() ->... |
679 680 681 682 683 684 685 |
if (error) return error; error = __mm_populate(start, len, 0); if (error) return __mlock_posix_error_return(error); return 0; |
1da177e4c Linux-2.6.12-rc2 |
686 |
} |
1aab92ec3 mm: mlock: refact... |
687 688 689 690 |
SYSCALL_DEFINE2(mlock, unsigned long, start, size_t, len) { return do_mlock(start, len, VM_LOCKED); } |
a8ca5d0ec mm: mlock: add ne... |
691 692 |
SYSCALL_DEFINE3(mlock2, unsigned long, start, size_t, len, int, flags) { |
b0f205c2a mm: mlock: add ml... |
693 694 695 |
vm_flags_t vm_flags = VM_LOCKED; if (flags & ~MLOCK_ONFAULT) |
a8ca5d0ec mm: mlock: add ne... |
696 |
return -EINVAL; |
b0f205c2a mm: mlock: add ml... |
697 698 699 700 |
if (flags & MLOCK_ONFAULT) vm_flags |= VM_LOCKONFAULT; return do_mlock(start, len, vm_flags); |
a8ca5d0ec mm: mlock: add ne... |
701 |
} |
6a6160a7b [CVE-2009-0029] S... |
702 |
SYSCALL_DEFINE2(munlock, unsigned long, start, size_t, len) |
1da177e4c Linux-2.6.12-rc2 |
703 704 |
{ int ret; |
8fd9e4883 mm/mlock: use off... |
705 |
len = PAGE_ALIGN(len + (offset_in_page(start))); |
1da177e4c Linux-2.6.12-rc2 |
706 |
start &= PAGE_MASK; |
1f1cd7054 mm/mlock: prepare... |
707 |
|
dc0ef0df7 mm: make mmap_sem... |
708 709 |
if (down_write_killable(¤t->mm->mmap_sem)) return -EINTR; |
1aab92ec3 mm: mlock: refact... |
710 |
ret = apply_vma_lock_flags(start, len, 0); |
1da177e4c Linux-2.6.12-rc2 |
711 |
up_write(¤t->mm->mmap_sem); |
1f1cd7054 mm/mlock: prepare... |
712 |
|
1da177e4c Linux-2.6.12-rc2 |
713 714 |
return ret; } |
b0f205c2a mm: mlock: add ml... |
715 716 717 718 719 720 721 722 723 724 |
/* * Take the MCL_* flags passed into mlockall (or 0 if called from munlockall) * and translate into the appropriate modifications to mm->def_flags and/or the * flags for all current VMAs. * * There are a couple of subtleties with this. If mlockall() is called multiple * times with different flags, the values do not necessarily stack. If mlockall * is called once including the MCL_FUTURE flag and then a second time without * it, VM_LOCKED and VM_LOCKONFAULT will be cleared from mm->def_flags. */ |
1aab92ec3 mm: mlock: refact... |
725 |
static int apply_mlockall_flags(int flags) |
1da177e4c Linux-2.6.12-rc2 |
726 727 |
{ struct vm_area_struct * vma, * prev = NULL; |
b0f205c2a mm: mlock: add ml... |
728 |
vm_flags_t to_add = 0; |
1da177e4c Linux-2.6.12-rc2 |
729 |
|
b0f205c2a mm: mlock: add ml... |
730 731 |
current->mm->def_flags &= VM_LOCKED_CLEAR_MASK; if (flags & MCL_FUTURE) { |
09a9f1d27 Revert "mm: intro... |
732 |
current->mm->def_flags |= VM_LOCKED; |
1aab92ec3 mm: mlock: refact... |
733 |
|
b0f205c2a mm: mlock: add ml... |
734 735 736 737 738 739 740 741 742 743 744 745 |
if (flags & MCL_ONFAULT) current->mm->def_flags |= VM_LOCKONFAULT; if (!(flags & MCL_CURRENT)) goto out; } if (flags & MCL_CURRENT) { to_add |= VM_LOCKED; if (flags & MCL_ONFAULT) to_add |= VM_LOCKONFAULT; } |
1da177e4c Linux-2.6.12-rc2 |
746 747 |
for (vma = current->mm->mmap; vma ; vma = prev->vm_next) { |
ca16d140a mm: don't access ... |
748 |
vm_flags_t newflags; |
1da177e4c Linux-2.6.12-rc2 |
749 |
|
b0f205c2a mm: mlock: add ml... |
750 751 |
newflags = vma->vm_flags & VM_LOCKED_CLEAR_MASK; newflags |= to_add; |
1da177e4c Linux-2.6.12-rc2 |
752 753 754 |
/* Ignore errors */ mlock_fixup(vma, &prev, vma->vm_start, vma->vm_end, newflags); |
bde6c3aa9 rcu: Provide cond... |
755 |
cond_resched_rcu_qs(); |
1da177e4c Linux-2.6.12-rc2 |
756 757 758 759 |
} out: return 0; } |
3480b2574 [CVE-2009-0029] S... |
760 |
SYSCALL_DEFINE1(mlockall, int, flags) |
1da177e4c Linux-2.6.12-rc2 |
761 762 |
{ unsigned long lock_limit; |
86d2adccf mm/mlock.c: reorg... |
763 |
int ret; |
1da177e4c Linux-2.6.12-rc2 |
764 |
|
b0f205c2a mm: mlock: add ml... |
765 |
if (!flags || (flags & ~(MCL_CURRENT | MCL_FUTURE | MCL_ONFAULT))) |
86d2adccf mm/mlock.c: reorg... |
766 |
return -EINVAL; |
1da177e4c Linux-2.6.12-rc2 |
767 |
|
1da177e4c Linux-2.6.12-rc2 |
768 |
if (!can_do_mlock()) |
86d2adccf mm/mlock.c: reorg... |
769 |
return -EPERM; |
1da177e4c Linux-2.6.12-rc2 |
770 |
|
df9d6985b mm: do not drain ... |
771 772 |
if (flags & MCL_CURRENT) lru_add_drain_all(); /* flush pagevec */ |
8891d6da1 mm: remove lru_ad... |
773 |
|
59e99e5b9 mm: use rlimit he... |
774 |
lock_limit = rlimit(RLIMIT_MEMLOCK); |
1da177e4c Linux-2.6.12-rc2 |
775 |
lock_limit >>= PAGE_SHIFT; |
dc0ef0df7 mm: make mmap_sem... |
776 777 |
if (down_write_killable(¤t->mm->mmap_sem)) return -EINTR; |
1f1cd7054 mm/mlock: prepare... |
778 |
|
dc0ef0df7 mm: make mmap_sem... |
779 |
ret = -ENOMEM; |
1da177e4c Linux-2.6.12-rc2 |
780 781 |
if (!(flags & MCL_CURRENT) || (current->mm->total_vm <= lock_limit) || capable(CAP_IPC_LOCK)) |
1aab92ec3 mm: mlock: refact... |
782 |
ret = apply_mlockall_flags(flags); |
1da177e4c Linux-2.6.12-rc2 |
783 |
up_write(¤t->mm->mmap_sem); |
bebeb3d68 mm: introduce mm_... |
784 785 |
if (!ret && (flags & MCL_CURRENT)) mm_populate(0, TASK_SIZE); |
86d2adccf mm/mlock.c: reorg... |
786 |
|
1da177e4c Linux-2.6.12-rc2 |
787 788 |
return ret; } |
3480b2574 [CVE-2009-0029] S... |
789 |
SYSCALL_DEFINE0(munlockall) |
1da177e4c Linux-2.6.12-rc2 |
790 791 |
{ int ret; |
dc0ef0df7 mm: make mmap_sem... |
792 793 |
if (down_write_killable(¤t->mm->mmap_sem)) return -EINTR; |
1aab92ec3 mm: mlock: refact... |
794 |
ret = apply_mlockall_flags(0); |
1da177e4c Linux-2.6.12-rc2 |
795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 |
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... |
811 |
lock_limit = rlimit(RLIMIT_MEMLOCK); |
5ed44a401 do not limit lock... |
812 813 |
if (lock_limit == RLIM_INFINITY) allowed = 1; |
1da177e4c Linux-2.6.12-rc2 |
814 815 |
lock_limit >>= PAGE_SHIFT; spin_lock(&shmlock_user_lock); |
5ed44a401 do not limit lock... |
816 817 |
if (!allowed && locked + user->locked_shm > lock_limit && !capable(CAP_IPC_LOCK)) |
1da177e4c Linux-2.6.12-rc2 |
818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 |
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); } |