Blame view
mm/madvise.c
28.8 KB
b24413180 License cleanup: ... |
1 |
// SPDX-License-Identifier: GPL-2.0 |
1da177e4c Linux-2.6.12-rc2 |
2 3 4 5 6 7 8 9 10 11 |
/* * linux/mm/madvise.c * * Copyright (C) 1999 Linus Torvalds * Copyright (C) 2002 Christoph Hellwig */ #include <linux/mman.h> #include <linux/pagemap.h> #include <linux/syscalls.h> |
05b743847 [PATCH] madvise: ... |
12 |
#include <linux/mempolicy.h> |
afcf938ee HWPOISON: Add a m... |
13 |
#include <linux/page-isolation.h> |
9c276cc65 mm: introduce MAD... |
14 |
#include <linux/page_idle.h> |
05ce77249 userfaultfd: non-... |
15 |
#include <linux/userfaultfd_k.h> |
1da177e4c Linux-2.6.12-rc2 |
16 |
#include <linux/hugetlb.h> |
3f31d0757 mm/fs: route MADV... |
17 |
#include <linux/falloc.h> |
692fe6243 mm: Handle MADV_W... |
18 |
#include <linux/fadvise.h> |
e8edc6e03 Detach sched.h fr... |
19 |
#include <linux/sched.h> |
f8af4da3b ksm: the mm inter... |
20 |
#include <linux/ksm.h> |
3f31d0757 mm/fs: route MADV... |
21 |
#include <linux/fs.h> |
9ab4233dd mm: Hold a file r... |
22 |
#include <linux/file.h> |
1998cc048 mm: make madvise(... |
23 |
#include <linux/blkdev.h> |
66114cad6 writeback: separa... |
24 |
#include <linux/backing-dev.h> |
a520110e4 mm: split out a n... |
25 |
#include <linux/pagewalk.h> |
1998cc048 mm: make madvise(... |
26 27 |
#include <linux/swap.h> #include <linux/swapops.h> |
3a4f8a0b3 mm: remove shmem_... |
28 |
#include <linux/shmem_fs.h> |
854e9ed09 mm: support madvi... |
29 30 31 |
#include <linux/mmu_notifier.h> #include <asm/tlb.h> |
1da177e4c Linux-2.6.12-rc2 |
32 |
|
235190738 oom-reaper: use m... |
33 |
#include "internal.h" |
d616d5126 mm: factor out co... |
34 35 36 37 |
struct madvise_walk_private { struct mmu_gather *tlb; bool pageout; }; |
1da177e4c Linux-2.6.12-rc2 |
38 |
/* |
0a27a14a6 mm: madvise avoid... |
39 40 41 42 43 44 45 46 47 48 |
* Any behaviour which results in changes to the vma->vm_flags needs to * take mmap_sem for writing. Others, which simply traverse vmas, need * to only take it for reading. */ static int madvise_need_mmap_write(int behavior) { switch (behavior) { case MADV_REMOVE: case MADV_WILLNEED: case MADV_DONTNEED: |
9c276cc65 mm: introduce MAD... |
49 |
case MADV_COLD: |
1a4e58cce mm: introduce MAD... |
50 |
case MADV_PAGEOUT: |
854e9ed09 mm: support madvi... |
51 |
case MADV_FREE: |
0a27a14a6 mm: madvise avoid... |
52 53 54 55 56 57 58 59 |
return 0; default: /* be safe, default to 1. list exceptions explicitly */ return 1; } } /* |
1da177e4c Linux-2.6.12-rc2 |
60 61 62 |
* We can potentially split a vm area into separate * areas, each area with its own behavior. */ |
ec9bed9d3 mm/madvise.c: fix... |
63 |
static long madvise_behavior(struct vm_area_struct *vma, |
05b743847 [PATCH] madvise: ... |
64 65 |
struct vm_area_struct **prev, unsigned long start, unsigned long end, int behavior) |
1da177e4c Linux-2.6.12-rc2 |
66 |
{ |
ec9bed9d3 mm/madvise.c: fix... |
67 |
struct mm_struct *mm = vma->vm_mm; |
1da177e4c Linux-2.6.12-rc2 |
68 |
int error = 0; |
05b743847 [PATCH] madvise: ... |
69 |
pgoff_t pgoff; |
3866ea90d ksm: first tidy u... |
70 |
unsigned long new_flags = vma->vm_flags; |
e798c6e87 [PATCH] madvise: ... |
71 72 |
switch (behavior) { |
f82256616 [PATCH] madvise M... |
73 74 75 |
case MADV_NORMAL: new_flags = new_flags & ~VM_RAND_READ & ~VM_SEQ_READ; break; |
e798c6e87 [PATCH] madvise: ... |
76 |
case MADV_SEQUENTIAL: |
f82256616 [PATCH] madvise M... |
77 |
new_flags = (new_flags & ~VM_RAND_READ) | VM_SEQ_READ; |
e798c6e87 [PATCH] madvise: ... |
78 79 |
break; case MADV_RANDOM: |
f82256616 [PATCH] madvise M... |
80 |
new_flags = (new_flags & ~VM_SEQ_READ) | VM_RAND_READ; |
e798c6e87 [PATCH] madvise: ... |
81 |
break; |
f82256616 [PATCH] madvise M... |
82 83 84 85 |
case MADV_DONTFORK: new_flags |= VM_DONTCOPY; break; case MADV_DOFORK: |
3866ea90d ksm: first tidy u... |
86 87 88 89 |
if (vma->vm_flags & VM_IO) { error = -EINVAL; goto out; } |
f82256616 [PATCH] madvise M... |
90 |
new_flags &= ~VM_DONTCOPY; |
e798c6e87 [PATCH] madvise: ... |
91 |
break; |
d2cd9ede6 mm,fork: introduc... |
92 93 94 95 96 97 98 99 100 101 102 |
case MADV_WIPEONFORK: /* MADV_WIPEONFORK is only supported on anonymous memory. */ if (vma->vm_file || vma->vm_flags & VM_SHARED) { error = -EINVAL; goto out; } new_flags |= VM_WIPEONFORK; break; case MADV_KEEPONFORK: new_flags &= ~VM_WIPEONFORK; break; |
accb61fe7 coredump: add VM_... |
103 |
case MADV_DONTDUMP: |
0103bd16f mm: prepare VM_DO... |
104 |
new_flags |= VM_DONTDUMP; |
accb61fe7 coredump: add VM_... |
105 106 |
break; case MADV_DODUMP: |
d41aa5252 mm: madvise(MADV_... |
107 |
if (!is_vm_hugetlb_page(vma) && new_flags & VM_SPECIAL) { |
0103bd16f mm: prepare VM_DO... |
108 109 110 111 |
error = -EINVAL; goto out; } new_flags &= ~VM_DONTDUMP; |
accb61fe7 coredump: add VM_... |
112 |
break; |
f8af4da3b ksm: the mm inter... |
113 114 115 |
case MADV_MERGEABLE: case MADV_UNMERGEABLE: error = ksm_madvise(vma, start, end, behavior, &new_flags); |
f3bc0dba3 mm/madvise: reduc... |
116 117 |
if (error) goto out_convert_errno; |
f8af4da3b ksm: the mm inter... |
118 |
break; |
0af4e98b6 thp: madvise(MADV... |
119 |
case MADV_HUGEPAGE: |
a664b2d85 thp: madvise(MADV... |
120 |
case MADV_NOHUGEPAGE: |
60ab3244e thp: khugepaged: ... |
121 |
error = hugepage_madvise(vma, &new_flags, behavior); |
f3bc0dba3 mm/madvise: reduc... |
122 123 |
if (error) goto out_convert_errno; |
0af4e98b6 thp: madvise(MADV... |
124 |
break; |
e798c6e87 [PATCH] madvise: ... |
125 |
} |
05b743847 [PATCH] madvise: ... |
126 127 |
if (new_flags == vma->vm_flags) { *prev = vma; |
836d5ffd3 [PATCH] mm: fix m... |
128 |
goto out; |
05b743847 [PATCH] madvise: ... |
129 130 131 132 |
} pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT); *prev = vma_merge(mm, *prev, start, end, new_flags, vma->anon_vma, |
19a809afe userfaultfd: teac... |
133 134 |
vma->vm_file, pgoff, vma_policy(vma), vma->vm_userfaultfd_ctx); |
05b743847 [PATCH] madvise: ... |
135 136 137 138 139 140 |
if (*prev) { vma = *prev; goto success; } *prev = vma; |
1da177e4c Linux-2.6.12-rc2 |
141 142 |
if (start != vma->vm_start) { |
def5efe03 mm, madvise: fail... |
143 144 |
if (unlikely(mm->map_count >= sysctl_max_map_count)) { error = -ENOMEM; |
1da177e4c Linux-2.6.12-rc2 |
145 |
goto out; |
def5efe03 mm, madvise: fail... |
146 147 |
} error = __split_vma(mm, vma, start, 1); |
f3bc0dba3 mm/madvise: reduc... |
148 149 |
if (error) goto out_convert_errno; |
1da177e4c Linux-2.6.12-rc2 |
150 151 152 |
} if (end != vma->vm_end) { |
def5efe03 mm, madvise: fail... |
153 154 |
if (unlikely(mm->map_count >= sysctl_max_map_count)) { error = -ENOMEM; |
1da177e4c Linux-2.6.12-rc2 |
155 |
goto out; |
def5efe03 mm, madvise: fail... |
156 157 |
} error = __split_vma(mm, vma, end, 0); |
f3bc0dba3 mm/madvise: reduc... |
158 159 |
if (error) goto out_convert_errno; |
1da177e4c Linux-2.6.12-rc2 |
160 |
} |
836d5ffd3 [PATCH] mm: fix m... |
161 |
success: |
1da177e4c Linux-2.6.12-rc2 |
162 163 164 |
/* * vm_flags is protected by the mmap_sem held in write mode. */ |
e798c6e87 [PATCH] madvise: ... |
165 |
vma->vm_flags = new_flags; |
f3bc0dba3 mm/madvise: reduc... |
166 167 168 169 170 171 172 173 |
out_convert_errno: /* * madvise() returns EAGAIN if kernel resources, such as * slab, are temporarily unavailable. */ if (error == -ENOMEM) error = -EAGAIN; |
1da177e4c Linux-2.6.12-rc2 |
174 |
out: |
1da177e4c Linux-2.6.12-rc2 |
175 176 |
return error; } |
1998cc048 mm: make madvise(... |
177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 |
#ifdef CONFIG_SWAP static int swapin_walk_pmd_entry(pmd_t *pmd, unsigned long start, unsigned long end, struct mm_walk *walk) { pte_t *orig_pte; struct vm_area_struct *vma = walk->private; unsigned long index; if (pmd_none_or_trans_huge_or_clear_bad(pmd)) return 0; for (index = start; index != end; index += PAGE_SIZE) { pte_t pte; swp_entry_t entry; struct page *page; spinlock_t *ptl; orig_pte = pte_offset_map_lock(vma->vm_mm, pmd, start, &ptl); pte = *(orig_pte + ((index - start) / PAGE_SIZE)); pte_unmap_unlock(orig_pte, ptl); |
0661a3361 mm: remove rest u... |
197 |
if (pte_present(pte) || pte_none(pte)) |
1998cc048 mm: make madvise(... |
198 199 200 201 202 203 |
continue; entry = pte_to_swp_entry(pte); if (unlikely(non_swap_entry(entry))) continue; page = read_swap_cache_async(entry, GFP_HIGHUSER_MOVABLE, |
23955622f swap: add block i... |
204 |
vma, index, false); |
1998cc048 mm: make madvise(... |
205 |
if (page) |
09cbfeaf1 mm, fs: get rid o... |
206 |
put_page(page); |
1998cc048 mm: make madvise(... |
207 208 209 210 |
} return 0; } |
7b86ac337 pagewalk: separat... |
211 212 213 |
static const struct mm_walk_ops swapin_walk_ops = { .pmd_entry = swapin_walk_pmd_entry, }; |
1998cc048 mm: make madvise(... |
214 215 216 217 218 219 220 221 222 223 224 |
static void force_shm_swapin_readahead(struct vm_area_struct *vma, unsigned long start, unsigned long end, struct address_space *mapping) { pgoff_t index; struct page *page; swp_entry_t swap; for (; start < end; start += PAGE_SIZE) { index = ((start - vma->vm_start) >> PAGE_SHIFT) + vma->vm_pgoff; |
55231e5c8 mm: madvise: fix ... |
225 |
page = find_get_entry(mapping, index); |
3159f943a xarray: Replace e... |
226 |
if (!xa_is_value(page)) { |
1998cc048 mm: make madvise(... |
227 |
if (page) |
09cbfeaf1 mm, fs: get rid o... |
228 |
put_page(page); |
1998cc048 mm: make madvise(... |
229 230 231 232 |
continue; } swap = radix_to_swp_entry(page); page = read_swap_cache_async(swap, GFP_HIGHUSER_MOVABLE, |
23955622f swap: add block i... |
233 |
NULL, 0, false); |
1998cc048 mm: make madvise(... |
234 |
if (page) |
09cbfeaf1 mm, fs: get rid o... |
235 |
put_page(page); |
1998cc048 mm: make madvise(... |
236 237 238 239 240 |
} lru_add_drain(); /* Push any new pages onto the LRU now */ } #endif /* CONFIG_SWAP */ |
1da177e4c Linux-2.6.12-rc2 |
241 242 243 |
/* * Schedule all required I/O operations. Do not wait for completion. */ |
ec9bed9d3 mm/madvise.c: fix... |
244 245 |
static long madvise_willneed(struct vm_area_struct *vma, struct vm_area_struct **prev, |
1da177e4c Linux-2.6.12-rc2 |
246 247 248 |
unsigned long start, unsigned long end) { struct file *file = vma->vm_file; |
692fe6243 mm: Handle MADV_W... |
249 |
loff_t offset; |
1da177e4c Linux-2.6.12-rc2 |
250 |
|
6ea8d958a mm/madvise.c: fix... |
251 |
*prev = vma; |
1998cc048 mm: make madvise(... |
252 |
#ifdef CONFIG_SWAP |
97b713ba3 fs: kill BDI_CAP_... |
253 |
if (!file) { |
7b86ac337 pagewalk: separat... |
254 255 |
walk_page_range(vma->vm_mm, start, end, &swapin_walk_ops, vma); lru_add_drain(); /* Push any new pages onto the LRU now */ |
1998cc048 mm: make madvise(... |
256 257 |
return 0; } |
1998cc048 mm: make madvise(... |
258 |
|
97b713ba3 fs: kill BDI_CAP_... |
259 |
if (shmem_mapping(file->f_mapping)) { |
97b713ba3 fs: kill BDI_CAP_... |
260 261 262 263 264 |
force_shm_swapin_readahead(vma, start, end, file->f_mapping); return 0; } #else |
1bef40032 [PATCH] madvise: ... |
265 266 |
if (!file) return -EBADF; |
97b713ba3 fs: kill BDI_CAP_... |
267 |
#endif |
1bef40032 [PATCH] madvise: ... |
268 |
|
e748dcd09 vfs: remove get_x... |
269 |
if (IS_DAX(file_inode(file))) { |
fe77ba6f4 [PATCH] xip: madv... |
270 271 272 |
/* no bad return value, but ignore advice */ return 0; } |
692fe6243 mm: Handle MADV_W... |
273 274 275 276 277 278 279 280 281 282 283 284 285 286 |
/* * Filesystem's fadvise may need to take various locks. We need to * explicitly grab a reference because the vma (and hence the * vma's reference to the file) can go away as soon as we drop * mmap_sem. */ *prev = NULL; /* tell sys_madvise we drop mmap_sem */ get_file(file); up_read(¤t->mm->mmap_sem); offset = (loff_t)(start - vma->vm_start) + ((loff_t)vma->vm_pgoff << PAGE_SHIFT); vfs_fadvise(file, offset, end - start, POSIX_FADV_WILLNEED); fput(file); down_read(¤t->mm->mmap_sem); |
1da177e4c Linux-2.6.12-rc2 |
287 288 |
return 0; } |
d616d5126 mm: factor out co... |
289 290 291 |
static int madvise_cold_or_pageout_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end, struct mm_walk *walk) |
9c276cc65 mm: introduce MAD... |
292 |
{ |
d616d5126 mm: factor out co... |
293 294 295 |
struct madvise_walk_private *private = walk->private; struct mmu_gather *tlb = private->tlb; bool pageout = private->pageout; |
9c276cc65 mm: introduce MAD... |
296 297 298 299 |
struct mm_struct *mm = tlb->mm; struct vm_area_struct *vma = walk->vma; pte_t *orig_pte, *pte, ptent; spinlock_t *ptl; |
d616d5126 mm: factor out co... |
300 301 302 303 304 |
struct page *page = NULL; LIST_HEAD(page_list); if (fatal_signal_pending(current)) return -EINTR; |
9c276cc65 mm: introduce MAD... |
305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 |
#ifdef CONFIG_TRANSPARENT_HUGEPAGE if (pmd_trans_huge(*pmd)) { pmd_t orig_pmd; unsigned long next = pmd_addr_end(addr, end); tlb_change_page_size(tlb, HPAGE_PMD_SIZE); ptl = pmd_trans_huge_lock(pmd, vma); if (!ptl) return 0; orig_pmd = *pmd; if (is_huge_zero_pmd(orig_pmd)) goto huge_unlock; if (unlikely(!pmd_present(orig_pmd))) { VM_BUG_ON(thp_migration_supported() && !is_pmd_migration_entry(orig_pmd)); goto huge_unlock; } page = pmd_page(orig_pmd); if (next - addr != HPAGE_PMD_SIZE) { int err; if (page_mapcount(page) != 1) goto huge_unlock; get_page(page); spin_unlock(ptl); lock_page(page); err = split_huge_page(page); unlock_page(page); put_page(page); if (!err) goto regular_page; return 0; } if (pmd_young(orig_pmd)) { pmdp_invalidate(vma, addr, pmd); orig_pmd = pmd_mkold(orig_pmd); set_pmd_at(mm, addr, pmd, orig_pmd); tlb_remove_pmd_tlb_entry(tlb, pmd, addr); } |
d616d5126 mm: factor out co... |
351 |
ClearPageReferenced(page); |
9c276cc65 mm: introduce MAD... |
352 |
test_and_clear_page_young(page); |
d616d5126 mm: factor out co... |
353 |
if (pageout) { |
820729629 mm: fix trying to... |
354 355 356 357 358 359 |
if (!isolate_lru_page(page)) { if (PageUnevictable(page)) putback_lru_page(page); else list_add(&page->lru, &page_list); } |
d616d5126 mm: factor out co... |
360 361 |
} else deactivate_page(page); |
9c276cc65 mm: introduce MAD... |
362 363 |
huge_unlock: spin_unlock(ptl); |
d616d5126 mm: factor out co... |
364 365 |
if (pageout) reclaim_pages(&page_list); |
9c276cc65 mm: introduce MAD... |
366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 |
return 0; } if (pmd_trans_unstable(pmd)) return 0; regular_page: #endif tlb_change_page_size(tlb, PAGE_SIZE); orig_pte = pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl); flush_tlb_batched_pending(mm); arch_enter_lazy_mmu_mode(); for (; addr < end; pte++, addr += PAGE_SIZE) { ptent = *pte; if (pte_none(ptent)) continue; if (!pte_present(ptent)) continue; page = vm_normal_page(vma, addr, ptent); if (!page) continue; /* * Creating a THP page is expensive so split it only if we * are sure it's worth. Split it if we are only owner. */ if (PageTransCompound(page)) { if (page_mapcount(page) != 1) break; get_page(page); if (!trylock_page(page)) { put_page(page); break; } pte_unmap_unlock(orig_pte, ptl); if (split_huge_page(page)) { unlock_page(page); put_page(page); pte_offset_map_lock(mm, pmd, addr, &ptl); break; } unlock_page(page); put_page(page); pte = pte_offset_map_lock(mm, pmd, addr, &ptl); pte--; addr -= PAGE_SIZE; continue; } VM_BUG_ON_PAGE(PageTransCompound(page), page); if (pte_young(ptent)) { ptent = ptep_get_and_clear_full(mm, addr, pte, tlb->fullmm); ptent = pte_mkold(ptent); set_pte_at(mm, addr, pte, ptent); tlb_remove_tlb_entry(tlb, pte, addr); } /* * We are deactivating a page for accelerating reclaiming. * VM couldn't reclaim the page unless we clear PG_young. * As a side effect, it makes confuse idle-page tracking * because they will miss recent referenced history. */ |
d616d5126 mm: factor out co... |
433 |
ClearPageReferenced(page); |
9c276cc65 mm: introduce MAD... |
434 |
test_and_clear_page_young(page); |
d616d5126 mm: factor out co... |
435 |
if (pageout) { |
820729629 mm: fix trying to... |
436 437 438 439 440 441 |
if (!isolate_lru_page(page)) { if (PageUnevictable(page)) putback_lru_page(page); else list_add(&page->lru, &page_list); } |
d616d5126 mm: factor out co... |
442 443 |
} else deactivate_page(page); |
9c276cc65 mm: introduce MAD... |
444 445 446 447 |
} arch_leave_lazy_mmu_mode(); pte_unmap_unlock(orig_pte, ptl); |
d616d5126 mm: factor out co... |
448 449 |
if (pageout) reclaim_pages(&page_list); |
9c276cc65 mm: introduce MAD... |
450 451 452 453 454 455 |
cond_resched(); return 0; } static const struct mm_walk_ops cold_walk_ops = { |
d616d5126 mm: factor out co... |
456 |
.pmd_entry = madvise_cold_or_pageout_pte_range, |
9c276cc65 mm: introduce MAD... |
457 458 459 460 461 462 |
}; static void madvise_cold_page_range(struct mmu_gather *tlb, struct vm_area_struct *vma, unsigned long addr, unsigned long end) { |
d616d5126 mm: factor out co... |
463 464 465 466 |
struct madvise_walk_private walk_private = { .pageout = false, .tlb = tlb, }; |
9c276cc65 mm: introduce MAD... |
467 |
tlb_start_vma(tlb, vma); |
d616d5126 mm: factor out co... |
468 |
walk_page_range(vma->vm_mm, addr, end, &cold_walk_ops, &walk_private); |
9c276cc65 mm: introduce MAD... |
469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 |
tlb_end_vma(tlb, vma); } static long madvise_cold(struct vm_area_struct *vma, struct vm_area_struct **prev, unsigned long start_addr, unsigned long end_addr) { struct mm_struct *mm = vma->vm_mm; struct mmu_gather tlb; *prev = vma; if (!can_madv_lru_vma(vma)) return -EINVAL; lru_add_drain(); tlb_gather_mmu(&tlb, mm, start_addr, end_addr); madvise_cold_page_range(&tlb, vma, start_addr, end_addr); tlb_finish_mmu(&tlb, start_addr, end_addr); return 0; } |
1a4e58cce mm: introduce MAD... |
490 491 492 493 |
static void madvise_pageout_page_range(struct mmu_gather *tlb, struct vm_area_struct *vma, unsigned long addr, unsigned long end) { |
d616d5126 mm: factor out co... |
494 495 496 497 |
struct madvise_walk_private walk_private = { .pageout = true, .tlb = tlb, }; |
1a4e58cce mm: introduce MAD... |
498 |
tlb_start_vma(tlb, vma); |
d616d5126 mm: factor out co... |
499 |
walk_page_range(vma->vm_mm, addr, end, &cold_walk_ops, &walk_private); |
1a4e58cce mm: introduce MAD... |
500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 |
tlb_end_vma(tlb, vma); } static inline bool can_do_pageout(struct vm_area_struct *vma) { if (vma_is_anonymous(vma)) return true; if (!vma->vm_file) return false; /* * paging out pagecache only for non-anonymous mappings that correspond * to the files the calling process could (if tried) open for writing; * otherwise we'd be including shared non-exclusive mappings, which * opens a side channel. */ return inode_owner_or_capable(file_inode(vma->vm_file)) || inode_permission(file_inode(vma->vm_file), MAY_WRITE) == 0; } static long madvise_pageout(struct vm_area_struct *vma, struct vm_area_struct **prev, unsigned long start_addr, unsigned long end_addr) { struct mm_struct *mm = vma->vm_mm; struct mmu_gather tlb; *prev = vma; if (!can_madv_lru_vma(vma)) return -EINVAL; if (!can_do_pageout(vma)) return 0; lru_add_drain(); tlb_gather_mmu(&tlb, mm, start_addr, end_addr); madvise_pageout_page_range(&tlb, vma, start_addr, end_addr); tlb_finish_mmu(&tlb, start_addr, end_addr); return 0; } |
854e9ed09 mm: support madvi... |
540 541 542 543 544 545 546 547 548 549 |
static int madvise_free_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end, struct mm_walk *walk) { struct mmu_gather *tlb = walk->private; struct mm_struct *mm = tlb->mm; struct vm_area_struct *vma = walk->vma; spinlock_t *ptl; pte_t *orig_pte, *pte, ptent; struct page *page; |
64b42bc1c mm/madvise.c: fre... |
550 |
int nr_swap = 0; |
b8d3c4c30 mm/huge_memory.c:... |
551 552 553 554 555 556 |
unsigned long next; next = pmd_addr_end(addr, end); if (pmd_trans_huge(*pmd)) if (madvise_free_huge_pmd(tlb, vma, pmd, addr, next)) goto next; |
854e9ed09 mm: support madvi... |
557 |
|
854e9ed09 mm: support madvi... |
558 559 |
if (pmd_trans_unstable(pmd)) return 0; |
ed6a79352 asm-generic/tlb, ... |
560 |
tlb_change_page_size(tlb, PAGE_SIZE); |
854e9ed09 mm: support madvi... |
561 |
orig_pte = pte = pte_offset_map_lock(mm, pmd, addr, &ptl); |
3ea277194 mm, mprotect: flu... |
562 |
flush_tlb_batched_pending(mm); |
854e9ed09 mm: support madvi... |
563 564 565 |
arch_enter_lazy_mmu_mode(); for (; addr != end; pte++, addr += PAGE_SIZE) { ptent = *pte; |
64b42bc1c mm/madvise.c: fre... |
566 |
if (pte_none(ptent)) |
854e9ed09 mm: support madvi... |
567 |
continue; |
64b42bc1c mm/madvise.c: fre... |
568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 |
/* * If the pte has swp_entry, just clear page table to * prevent swap-in which is more expensive rather than * (page allocation + zeroing). */ if (!pte_present(ptent)) { swp_entry_t entry; entry = pte_to_swp_entry(ptent); if (non_swap_entry(entry)) continue; nr_swap--; free_swap_and_cache(entry); pte_clear_not_present_full(mm, addr, pte, tlb->fullmm); continue; } |
854e9ed09 mm: support madvi... |
584 |
|
25b2995a3 mm: remove MEMORY... |
585 |
page = vm_normal_page(vma, addr, ptent); |
854e9ed09 mm: support madvi... |
586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 |
if (!page) continue; /* * If pmd isn't transhuge but the page is THP and * is owned by only this process, split it and * deactivate all pages. */ if (PageTransCompound(page)) { if (page_mapcount(page) != 1) goto out; get_page(page); if (!trylock_page(page)) { put_page(page); goto out; } pte_unmap_unlock(orig_pte, ptl); if (split_huge_page(page)) { unlock_page(page); put_page(page); pte_offset_map_lock(mm, pmd, addr, &ptl); goto out; } |
854e9ed09 mm: support madvi... |
609 |
unlock_page(page); |
263630e8d mm/madvise.c: fix... |
610 |
put_page(page); |
854e9ed09 mm: support madvi... |
611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 |
pte = pte_offset_map_lock(mm, pmd, addr, &ptl); pte--; addr -= PAGE_SIZE; continue; } VM_BUG_ON_PAGE(PageTransCompound(page), page); if (PageSwapCache(page) || PageDirty(page)) { if (!trylock_page(page)) continue; /* * If page is shared with others, we couldn't clear * PG_dirty of the page. */ if (page_mapcount(page) != 1) { unlock_page(page); continue; } if (PageSwapCache(page) && !try_to_free_swap(page)) { unlock_page(page); continue; } ClearPageDirty(page); unlock_page(page); } if (pte_young(ptent) || pte_dirty(ptent)) { /* * Some of architecture(ex, PPC) don't update TLB * with set_pte_at and tlb_remove_tlb_entry so for * the portability, remap the pte with old|clean * after pte clearing. */ ptent = ptep_get_and_clear_full(mm, addr, pte, tlb->fullmm); ptent = pte_mkold(ptent); ptent = pte_mkclean(ptent); set_pte_at(mm, addr, pte, ptent); tlb_remove_tlb_entry(tlb, pte, addr); } |
802a3a92a mm: reclaim MADV_... |
655 |
mark_page_lazyfree(page); |
854e9ed09 mm: support madvi... |
656 657 |
} out: |
64b42bc1c mm/madvise.c: fre... |
658 659 660 661 662 663 |
if (nr_swap) { if (current->mm == mm) sync_mm_rss(mm); add_mm_counter(mm, MM_SWAPENTS, nr_swap); } |
854e9ed09 mm: support madvi... |
664 665 666 |
arch_leave_lazy_mmu_mode(); pte_unmap_unlock(orig_pte, ptl); cond_resched(); |
b8d3c4c30 mm/huge_memory.c:... |
667 |
next: |
854e9ed09 mm: support madvi... |
668 669 |
return 0; } |
7b86ac337 pagewalk: separat... |
670 671 672 |
static const struct mm_walk_ops madvise_free_walk_ops = { .pmd_entry = madvise_free_pte_range, }; |
854e9ed09 mm: support madvi... |
673 674 675 676 |
static int madvise_free_single_vma(struct vm_area_struct *vma, unsigned long start_addr, unsigned long end_addr) { |
854e9ed09 mm: support madvi... |
677 |
struct mm_struct *mm = vma->vm_mm; |
ac46d4f3c mm/mmu_notifier: ... |
678 |
struct mmu_notifier_range range; |
854e9ed09 mm: support madvi... |
679 |
struct mmu_gather tlb; |
854e9ed09 mm: support madvi... |
680 681 682 |
/* MADV_FREE works for only anon vma at the moment */ if (!vma_is_anonymous(vma)) return -EINVAL; |
ac46d4f3c mm/mmu_notifier: ... |
683 684 |
range.start = max(vma->vm_start, start_addr); if (range.start >= vma->vm_end) |
854e9ed09 mm: support madvi... |
685 |
return -EINVAL; |
ac46d4f3c mm/mmu_notifier: ... |
686 687 |
range.end = min(vma->vm_end, end_addr); if (range.end <= vma->vm_start) |
854e9ed09 mm: support madvi... |
688 |
return -EINVAL; |
7269f9999 mm/mmu_notifier: ... |
689 |
mmu_notifier_range_init(&range, MMU_NOTIFY_CLEAR, 0, vma, mm, |
6f4f13e8d mm/mmu_notifier: ... |
690 |
range.start, range.end); |
854e9ed09 mm: support madvi... |
691 692 |
lru_add_drain(); |
ac46d4f3c mm/mmu_notifier: ... |
693 |
tlb_gather_mmu(&tlb, mm, range.start, range.end); |
854e9ed09 mm: support madvi... |
694 |
update_hiwater_rss(mm); |
ac46d4f3c mm/mmu_notifier: ... |
695 |
mmu_notifier_invalidate_range_start(&range); |
7b86ac337 pagewalk: separat... |
696 697 698 699 |
tlb_start_vma(&tlb, vma); walk_page_range(vma->vm_mm, range.start, range.end, &madvise_free_walk_ops, &tlb); tlb_end_vma(&tlb, vma); |
ac46d4f3c mm/mmu_notifier: ... |
700 701 |
mmu_notifier_invalidate_range_end(&range); tlb_finish_mmu(&tlb, range.start, range.end); |
854e9ed09 mm: support madvi... |
702 703 704 |
return 0; } |
1da177e4c Linux-2.6.12-rc2 |
705 706 707 708 |
/* * Application no longer needs these pages. If the pages are dirty, * it's OK to just throw them away. The app will be more careful about * data it wants to keep. Be sure to free swap resources too. The |
7e6cbea39 madvise: update f... |
709 |
* zap_page_range call sets things up for shrink_active_list to actually free |
1da177e4c Linux-2.6.12-rc2 |
710 711 |
* these pages later if no one else has touched them in the meantime, * although we could add these pages to a global reuse list for |
7e6cbea39 madvise: update f... |
712 |
* shrink_active_list to pick up before reclaiming other pages. |
1da177e4c Linux-2.6.12-rc2 |
713 714 715 716 717 718 719 720 721 722 723 |
* * NB: This interface discards data rather than pushes it out to swap, * as some implementations do. This has performance implications for * applications like large transactional databases which want to discard * pages in anonymous maps after committing to backing store the data * that was kept in them. There is no reason to write this data out to * the swap area if the application is discarding it. * * An interface that causes the system to free clean pages and flush * dirty pages is already available as msync(MS_INVALIDATE). */ |
230ca982b userfaultfd: non-... |
724 725 726 727 728 729 730 731 732 733 734 |
static long madvise_dontneed_single_vma(struct vm_area_struct *vma, unsigned long start, unsigned long end) { zap_page_range(vma, start, end - start); return 0; } static long madvise_dontneed_free(struct vm_area_struct *vma, struct vm_area_struct **prev, unsigned long start, unsigned long end, int behavior) |
1da177e4c Linux-2.6.12-rc2 |
735 |
{ |
05b743847 [PATCH] madvise: ... |
736 |
*prev = vma; |
9c276cc65 mm: introduce MAD... |
737 |
if (!can_madv_lru_vma(vma)) |
1da177e4c Linux-2.6.12-rc2 |
738 |
return -EINVAL; |
70ccb92fd userfaultfd: non-... |
739 740 741 742 743 744 745 746 747 748 749 750 751 752 |
if (!userfaultfd_remove(vma, start, end)) { *prev = NULL; /* mmap_sem has been dropped, prev is stale */ down_read(¤t->mm->mmap_sem); vma = find_vma(current->mm, start); if (!vma) return -ENOMEM; if (start < vma->vm_start) { /* * This "vma" under revalidation is the one * with the lowest vma->vm_start where start * is also < vma->vm_end. If start < * vma->vm_start it means an hole materialized * in the user address space within the |
230ca982b userfaultfd: non-... |
753 754 |
* virtual range passed to MADV_DONTNEED * or MADV_FREE. |
70ccb92fd userfaultfd: non-... |
755 756 757 |
*/ return -ENOMEM; } |
9c276cc65 mm: introduce MAD... |
758 |
if (!can_madv_lru_vma(vma)) |
70ccb92fd userfaultfd: non-... |
759 760 761 762 763 764 |
return -EINVAL; if (end > vma->vm_end) { /* * Don't fail if end > vma->vm_end. If the old * vma was splitted while the mmap_sem was * released the effect of the concurrent |
230ca982b userfaultfd: non-... |
765 |
* operation may not cause madvise() to |
70ccb92fd userfaultfd: non-... |
766 767 768 769 770 771 772 773 774 775 776 |
* have an undefined result. There may be an * adjacent next vma that we'll walk * next. userfaultfd_remove() will generate an * UFFD_EVENT_REMOVE repetition on the * end-vma->vm_end range, but the manager can * handle a repetition fine. */ end = vma->vm_end; } VM_WARN_ON(start >= end); } |
230ca982b userfaultfd: non-... |
777 778 779 780 781 782 783 |
if (behavior == MADV_DONTNEED) return madvise_dontneed_single_vma(vma, start, end); else if (behavior == MADV_FREE) return madvise_free_single_vma(vma, start, end); else return -EINVAL; |
1da177e4c Linux-2.6.12-rc2 |
784 |
} |
f6b3ec238 [PATCH] madvise(M... |
785 786 787 |
/* * Application wants to free up the pages and associated backing store. * This is effectively punching a hole into the middle of a file. |
f6b3ec238 [PATCH] madvise(M... |
788 789 |
*/ static long madvise_remove(struct vm_area_struct *vma, |
00e9fa2d6 [PATCH] mm: fix m... |
790 |
struct vm_area_struct **prev, |
f6b3ec238 [PATCH] madvise(M... |
791 792 |
unsigned long start, unsigned long end) { |
3f31d0757 mm/fs: route MADV... |
793 |
loff_t offset; |
90ed52ebe [PATCH] holepunch... |
794 |
int error; |
9ab4233dd mm: Hold a file r... |
795 |
struct file *f; |
f6b3ec238 [PATCH] madvise(M... |
796 |
|
90ed52ebe [PATCH] holepunch... |
797 |
*prev = NULL; /* tell sys_madvise we drop mmap_sem */ |
00e9fa2d6 [PATCH] mm: fix m... |
798 |
|
72079ba0d mm: madvise allow... |
799 |
if (vma->vm_flags & VM_LOCKED) |
f6b3ec238 [PATCH] madvise(M... |
800 |
return -EINVAL; |
9ab4233dd mm: Hold a file r... |
801 802 803 |
f = vma->vm_file; if (!f || !f->f_mapping || !f->f_mapping->host) { |
f6b3ec238 [PATCH] madvise(M... |
804 805 |
return -EINVAL; } |
69cf0fac6 [PATCH] Fix MADV_... |
806 807 |
if ((vma->vm_flags & (VM_SHARED|VM_WRITE)) != (VM_SHARED|VM_WRITE)) return -EACCES; |
f6b3ec238 [PATCH] madvise(M... |
808 809 |
offset = (loff_t)(start - vma->vm_start) + ((loff_t)vma->vm_pgoff << PAGE_SHIFT); |
90ed52ebe [PATCH] holepunch... |
810 |
|
9ab4233dd mm: Hold a file r... |
811 812 813 814 815 816 817 |
/* * Filesystem's fallocate may need to take i_mutex. We need to * explicitly grab a reference because the vma (and hence the * vma's reference to the file) can go away as soon as we drop * mmap_sem. */ get_file(f); |
70ccb92fd userfaultfd: non-... |
818 819 820 821 |
if (userfaultfd_remove(vma, start, end)) { /* mmap_sem was not released by userfaultfd_remove() */ up_read(¤t->mm->mmap_sem); } |
72c72bdf7 VFS: Rename do_fa... |
822 |
error = vfs_fallocate(f, |
3f31d0757 mm/fs: route MADV... |
823 824 |
FALLOC_FL_PUNCH_HOLE | FALLOC_FL_KEEP_SIZE, offset, end - start); |
9ab4233dd mm: Hold a file r... |
825 |
fput(f); |
0a27a14a6 mm: madvise avoid... |
826 |
down_read(¤t->mm->mmap_sem); |
90ed52ebe [PATCH] holepunch... |
827 |
return error; |
f6b3ec238 [PATCH] madvise(M... |
828 |
} |
9893e49d6 HWPOISON: Add mad... |
829 830 831 832 |
#ifdef CONFIG_MEMORY_FAILURE /* * Error injection support for memory error handling. */ |
97167a768 mm/madvise.c: cle... |
833 834 |
static int madvise_inject_error(int behavior, unsigned long start, unsigned long end) |
9893e49d6 HWPOISON: Add mad... |
835 |
{ |
97167a768 mm/madvise.c: cle... |
836 |
struct page *page; |
c461ad6a6 mm, madvise: ensu... |
837 |
struct zone *zone; |
19bfbe22f mm, hugetlb, soft... |
838 |
unsigned int order; |
97167a768 mm/madvise.c: cle... |
839 |
|
9893e49d6 HWPOISON: Add mad... |
840 841 |
if (!capable(CAP_SYS_ADMIN)) return -EPERM; |
97167a768 mm/madvise.c: cle... |
842 |
|
19bfbe22f mm, hugetlb, soft... |
843 844 |
for (; start < end; start += PAGE_SIZE << order) { |
23e7b5c2e mm, madvise_injec... |
845 |
unsigned long pfn; |
325c4ef5c mm/madvise.c:madv... |
846 |
int ret; |
97167a768 mm/madvise.c: cle... |
847 |
ret = get_user_pages_fast(start, 1, 0, &page); |
9893e49d6 HWPOISON: Add mad... |
848 849 |
if (ret != 1) return ret; |
23e7b5c2e mm, madvise_injec... |
850 |
pfn = page_to_pfn(page); |
325c4ef5c mm/madvise.c:madv... |
851 |
|
19bfbe22f mm, hugetlb, soft... |
852 853 854 855 856 857 |
/* * When soft offlining hugepages, after migrating the page * we dissolve it, therefore in the second loop "page" will * no longer be a compound page, and order will be 0. */ order = compound_order(compound_head(page)); |
97167a768 mm/madvise.c: cle... |
858 859 |
if (PageHWPoison(page)) { put_page(page); |
29b4eedee mm/hwpoison.c: fi... |
860 861 |
continue; } |
97167a768 mm/madvise.c: cle... |
862 863 864 865 |
if (behavior == MADV_SOFT_OFFLINE) { pr_info("Soft offlining pfn %#lx at process virtual address %#lx ", |
23e7b5c2e mm, madvise_injec... |
866 |
pfn, start); |
97167a768 mm/madvise.c: cle... |
867 868 |
ret = soft_offline_page(page, MF_COUNT_INCREASED); |
afcf938ee HWPOISON: Add a m... |
869 |
if (ret) |
8302423b8 mm/madvise.c: fix... |
870 |
return ret; |
afcf938ee HWPOISON: Add a m... |
871 872 |
continue; } |
23e7b5c2e mm, madvise_injec... |
873 |
|
97167a768 mm/madvise.c: cle... |
874 875 |
pr_info("Injecting memory failure for pfn %#lx at process virtual address %#lx ", |
23e7b5c2e mm, madvise_injec... |
876 |
pfn, start); |
97167a768 mm/madvise.c: cle... |
877 |
|
23e7b5c2e mm, madvise_injec... |
878 879 880 881 882 883 884 885 |
/* * Drop the page reference taken by get_user_pages_fast(). In * the absence of MF_COUNT_INCREASED the memory_failure() * routine is responsible for pinning the page to prevent it * from being released back to the page allocator. */ put_page(page); ret = memory_failure(pfn, 0); |
23a003bfd mm/madvise: pass ... |
886 887 |
if (ret) return ret; |
9893e49d6 HWPOISON: Add mad... |
888 |
} |
c461ad6a6 mm, madvise: ensu... |
889 890 891 892 |
/* Ensure that all poisoned pages are removed from per-cpu lists */ for_each_populated_zone(zone) drain_all_pages(zone); |
325c4ef5c mm/madvise.c:madv... |
893 |
return 0; |
9893e49d6 HWPOISON: Add mad... |
894 895 |
} #endif |
165cd4023 [PATCH] madvise()... |
896 897 898 |
static long madvise_vma(struct vm_area_struct *vma, struct vm_area_struct **prev, unsigned long start, unsigned long end, int behavior) |
1da177e4c Linux-2.6.12-rc2 |
899 |
{ |
1da177e4c Linux-2.6.12-rc2 |
900 |
switch (behavior) { |
f6b3ec238 [PATCH] madvise(M... |
901 |
case MADV_REMOVE: |
3866ea90d ksm: first tidy u... |
902 |
return madvise_remove(vma, prev, start, end); |
1da177e4c Linux-2.6.12-rc2 |
903 |
case MADV_WILLNEED: |
3866ea90d ksm: first tidy u... |
904 |
return madvise_willneed(vma, prev, start, end); |
9c276cc65 mm: introduce MAD... |
905 906 |
case MADV_COLD: return madvise_cold(vma, prev, start, end); |
1a4e58cce mm: introduce MAD... |
907 908 |
case MADV_PAGEOUT: return madvise_pageout(vma, prev, start, end); |
854e9ed09 mm: support madvi... |
909 |
case MADV_FREE: |
1da177e4c Linux-2.6.12-rc2 |
910 |
case MADV_DONTNEED: |
230ca982b userfaultfd: non-... |
911 |
return madvise_dontneed_free(vma, prev, start, end, behavior); |
1da177e4c Linux-2.6.12-rc2 |
912 |
default: |
3866ea90d ksm: first tidy u... |
913 |
return madvise_behavior(vma, prev, start, end, behavior); |
1da177e4c Linux-2.6.12-rc2 |
914 |
} |
1da177e4c Linux-2.6.12-rc2 |
915 |
} |
1ecef9ed0 mm/madvise.c: mak... |
916 |
static bool |
75927af8b mm: madvise(): co... |
917 918 919 920 921 922 923 924 925 926 927 |
madvise_behavior_valid(int behavior) { switch (behavior) { case MADV_DOFORK: case MADV_DONTFORK: case MADV_NORMAL: case MADV_SEQUENTIAL: case MADV_RANDOM: case MADV_REMOVE: case MADV_WILLNEED: case MADV_DONTNEED: |
854e9ed09 mm: support madvi... |
928 |
case MADV_FREE: |
9c276cc65 mm: introduce MAD... |
929 |
case MADV_COLD: |
1a4e58cce mm: introduce MAD... |
930 |
case MADV_PAGEOUT: |
f8af4da3b ksm: the mm inter... |
931 932 933 934 |
#ifdef CONFIG_KSM case MADV_MERGEABLE: case MADV_UNMERGEABLE: #endif |
0af4e98b6 thp: madvise(MADV... |
935 936 |
#ifdef CONFIG_TRANSPARENT_HUGEPAGE case MADV_HUGEPAGE: |
a664b2d85 thp: madvise(MADV... |
937 |
case MADV_NOHUGEPAGE: |
0af4e98b6 thp: madvise(MADV... |
938 |
#endif |
accb61fe7 coredump: add VM_... |
939 940 |
case MADV_DONTDUMP: case MADV_DODUMP: |
d2cd9ede6 mm,fork: introduc... |
941 942 |
case MADV_WIPEONFORK: case MADV_KEEPONFORK: |
5e451be75 mm/madvise: move ... |
943 944 945 946 |
#ifdef CONFIG_MEMORY_FAILURE case MADV_SOFT_OFFLINE: case MADV_HWPOISON: #endif |
1ecef9ed0 mm/madvise.c: mak... |
947 |
return true; |
75927af8b mm: madvise(): co... |
948 949 |
default: |
1ecef9ed0 mm/madvise.c: mak... |
950 |
return false; |
75927af8b mm: madvise(): co... |
951 952 |
} } |
3866ea90d ksm: first tidy u... |
953 |
|
1da177e4c Linux-2.6.12-rc2 |
954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 |
/* * The madvise(2) system call. * * Applications can use madvise() to advise the kernel how it should * handle paging I/O in this VM area. The idea is to help the kernel * use appropriate read-ahead and caching techniques. The information * provided is advisory only, and can be safely disregarded by the * kernel without affecting the correct operation of the application. * * behavior values: * MADV_NORMAL - the default behavior is to read clusters. This * results in some read-ahead and read-behind. * MADV_RANDOM - the system should read the minimum amount of data * on any access, since it is unlikely that the appli- * cation will need more than what it asks for. * MADV_SEQUENTIAL - pages in the given range will probably be accessed * once, so they can be aggressively read ahead, and * can be freed soon after they are accessed. * MADV_WILLNEED - the application is notifying the system to read * some pages ahead. * MADV_DONTNEED - the application is finished with the given range, * so the kernel can free resources associated with it. |
d7206a70a mm/madvise: updat... |
976 977 |
* MADV_FREE - the application marks pages in the given range as lazy free, * where actual purges are postponed until memory pressure happens. |
f6b3ec238 [PATCH] madvise(M... |
978 979 |
* MADV_REMOVE - the application wants to free up the given range of * pages and associated backing store. |
3866ea90d ksm: first tidy u... |
980 981 982 |
* MADV_DONTFORK - omit this area from child's address space when forking: * typically, to avoid COWing pages pinned by get_user_pages(). * MADV_DOFORK - cancel MADV_DONTFORK: no longer omit this area when forking. |
c02c30093 mm/madvise.c: add... |
983 984 985 |
* MADV_WIPEONFORK - present the child process with zero-filled memory in this * range after a fork. * MADV_KEEPONFORK - undo the effect of MADV_WIPEONFORK |
d7206a70a mm/madvise: updat... |
986 987 988 |
* MADV_HWPOISON - trigger memory error handler as if the given memory range * were corrupted by unrecoverable hardware memory failure. * MADV_SOFT_OFFLINE - try to soft-offline the given range of memory. |
f8af4da3b ksm: the mm inter... |
989 990 991 |
* MADV_MERGEABLE - the application recommends that KSM try to merge pages in * this area with pages of identical content from other such areas. * MADV_UNMERGEABLE- cancel MADV_MERGEABLE: no longer merge pages with others. |
d7206a70a mm/madvise: updat... |
992 993 994 995 996 997 998 999 1000 |
* MADV_HUGEPAGE - the application wants to back the given range by transparent * huge pages in the future. Existing pages might be coalesced and * new pages might be allocated as THP. * MADV_NOHUGEPAGE - mark the given range as not worth being backed by * transparent huge pages so the existing pages will not be * coalesced into THP and new pages will not be allocated as THP. * MADV_DONTDUMP - the application wants to prevent pages in the given range * from being included in its core dump. * MADV_DODUMP - cancel MADV_DONTDUMP: no longer exclude from core dump. |
1da177e4c Linux-2.6.12-rc2 |
1001 1002 1003 1004 1005 |
* * return values: * zero - success * -EINVAL - start + len < 0, start is not page-aligned, * "behavior" is not a valid value, or application |
c02c30093 mm/madvise.c: add... |
1006 1007 1008 |
* is attempting to release locked or shared pages, * or the specified address range includes file, Huge TLB, * MAP_SHARED or VMPFNMAP range. |
1da177e4c Linux-2.6.12-rc2 |
1009 1010 1011 1012 1013 1014 |
* -ENOMEM - addresses in the specified range are not currently * mapped, or are outside the AS of the process. * -EIO - an I/O error occurred while paging in data. * -EBADF - map exists, but area maps something that isn't a file. * -EAGAIN - a kernel resource was temporarily unavailable. */ |
3480b2574 [CVE-2009-0029] S... |
1015 |
SYSCALL_DEFINE3(madvise, unsigned long, start, size_t, len_in, int, behavior) |
1da177e4c Linux-2.6.12-rc2 |
1016 |
{ |
05b743847 [PATCH] madvise: ... |
1017 |
unsigned long end, tmp; |
ec9bed9d3 mm/madvise.c: fix... |
1018 |
struct vm_area_struct *vma, *prev; |
1da177e4c Linux-2.6.12-rc2 |
1019 1020 |
int unmapped_error = 0; int error = -EINVAL; |
f79777932 speed up madvise_... |
1021 |
int write; |
1da177e4c Linux-2.6.12-rc2 |
1022 |
size_t len; |
1998cc048 mm: make madvise(... |
1023 |
struct blk_plug plug; |
1da177e4c Linux-2.6.12-rc2 |
1024 |
|
057d33891 mm: untag user po... |
1025 |
start = untagged_addr(start); |
75927af8b mm: madvise(): co... |
1026 1027 |
if (!madvise_behavior_valid(behavior)) return error; |
1da177e4c Linux-2.6.12-rc2 |
1028 |
if (start & ~PAGE_MASK) |
84d96d897 mm: madvise: comp... |
1029 |
return error; |
1da177e4c Linux-2.6.12-rc2 |
1030 1031 1032 1033 |
len = (len_in + ~PAGE_MASK) & PAGE_MASK; /* Check to see whether len was rounded up from small -ve to zero */ if (len_in && !len) |
84d96d897 mm: madvise: comp... |
1034 |
return error; |
1da177e4c Linux-2.6.12-rc2 |
1035 1036 1037 |
end = start + len; if (end < start) |
84d96d897 mm: madvise: comp... |
1038 |
return error; |
1da177e4c Linux-2.6.12-rc2 |
1039 1040 1041 |
error = 0; if (end == start) |
84d96d897 mm: madvise: comp... |
1042 |
return error; |
5e451be75 mm/madvise: move ... |
1043 1044 1045 1046 |
#ifdef CONFIG_MEMORY_FAILURE if (behavior == MADV_HWPOISON || behavior == MADV_SOFT_OFFLINE) return madvise_inject_error(behavior, start, start + len_in); #endif |
84d96d897 mm: madvise: comp... |
1047 |
write = madvise_need_mmap_write(behavior); |
dc0ef0df7 mm: make mmap_sem... |
1048 1049 1050 1051 |
if (write) { if (down_write_killable(¤t->mm->mmap_sem)) return -EINTR; } else { |
84d96d897 mm: madvise: comp... |
1052 |
down_read(¤t->mm->mmap_sem); |
dc0ef0df7 mm: make mmap_sem... |
1053 |
} |
1da177e4c Linux-2.6.12-rc2 |
1054 1055 1056 1057 |
/* * If the interval [start,end) covers some unmapped address * ranges, just ignore them, but return -ENOMEM at the end. |
05b743847 [PATCH] madvise: ... |
1058 |
* - different from the way of handling in mlock etc. |
1da177e4c Linux-2.6.12-rc2 |
1059 |
*/ |
05b743847 [PATCH] madvise: ... |
1060 |
vma = find_vma_prev(current->mm, start, &prev); |
836d5ffd3 [PATCH] mm: fix m... |
1061 1062 |
if (vma && start > vma->vm_start) prev = vma; |
1998cc048 mm: make madvise(... |
1063 |
blk_start_plug(&plug); |
1da177e4c Linux-2.6.12-rc2 |
1064 1065 1066 1067 |
for (;;) { /* Still start < end. */ error = -ENOMEM; if (!vma) |
84d96d897 mm: madvise: comp... |
1068 |
goto out; |
1da177e4c Linux-2.6.12-rc2 |
1069 |
|
05b743847 [PATCH] madvise: ... |
1070 |
/* Here start < (end|vma->vm_end). */ |
1da177e4c Linux-2.6.12-rc2 |
1071 1072 1073 |
if (start < vma->vm_start) { unmapped_error = -ENOMEM; start = vma->vm_start; |
05b743847 [PATCH] madvise: ... |
1074 |
if (start >= end) |
84d96d897 mm: madvise: comp... |
1075 |
goto out; |
1da177e4c Linux-2.6.12-rc2 |
1076 |
} |
05b743847 [PATCH] madvise: ... |
1077 1078 1079 1080 |
/* Here vma->vm_start <= start < (end|vma->vm_end) */ tmp = vma->vm_end; if (end < tmp) tmp = end; |
1da177e4c Linux-2.6.12-rc2 |
1081 |
|
05b743847 [PATCH] madvise: ... |
1082 1083 |
/* Here vma->vm_start <= start < tmp <= (end|vma->vm_end). */ error = madvise_vma(vma, &prev, start, tmp, behavior); |
1da177e4c Linux-2.6.12-rc2 |
1084 |
if (error) |
84d96d897 mm: madvise: comp... |
1085 |
goto out; |
05b743847 [PATCH] madvise: ... |
1086 |
start = tmp; |
90ed52ebe [PATCH] holepunch... |
1087 |
if (prev && start < prev->vm_end) |
05b743847 [PATCH] madvise: ... |
1088 1089 1090 |
start = prev->vm_end; error = unmapped_error; if (start >= end) |
84d96d897 mm: madvise: comp... |
1091 |
goto out; |
90ed52ebe [PATCH] holepunch... |
1092 1093 1094 1095 |
if (prev) vma = prev->vm_next; else /* madvise_remove dropped mmap_sem */ vma = find_vma(current->mm, start); |
1da177e4c Linux-2.6.12-rc2 |
1096 |
} |
1da177e4c Linux-2.6.12-rc2 |
1097 |
out: |
84d96d897 mm: madvise: comp... |
1098 |
blk_finish_plug(&plug); |
f79777932 speed up madvise_... |
1099 |
if (write) |
0a27a14a6 mm: madvise avoid... |
1100 1101 1102 |
up_write(¤t->mm->mmap_sem); else up_read(¤t->mm->mmap_sem); |
1da177e4c Linux-2.6.12-rc2 |
1103 1104 |
return error; } |