Blame view
mm/memory.c
110 KB
1da177e4c
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 |
/* * linux/mm/memory.c * * Copyright (C) 1991, 1992, 1993, 1994 Linus Torvalds */ /* * demand-loading started 01.12.91 - seems it is high on the list of * things wanted, and it should be easy to implement. - Linus */ /* * Ok, demand-loading was easy, shared pages a little bit tricker. Shared * pages started 02.12.91, seems to work. - Linus. * * Tested sharing by executing about 30 /bin/sh: under the old kernel it * would have taken more than the 6M I have free, but it worked well as * far as I could see. * * Also corrected some "invalidate()"s - I wasn't doing enough of them. */ /* * Real VM (paging to/from disk) started 18.12.91. Much more work and * thought has to go into this. Oh, well.. * 19.12.91 - works, somewhat. Sometimes I get faults, don't know why. * Found it. Everything seems to work now. * 20.12.91 - Ok, making the swap-device changeable like the root. */ /* * 05.04.94 - Multi-page memory management added for v1.1. * Idea by Alex Bligh (alex@cconcepts.co.uk) * * 16.07.99 - Support of BIGMEM added by Gerhard Wichert, Siemens AG * (Gerhard.Wichert@pdb.siemens.de) * * Aug/Sep 2004 Changed to four level page tables (Andi Kleen) */ #include <linux/kernel_stat.h> #include <linux/mm.h> #include <linux/hugetlb.h> #include <linux/mman.h> #include <linux/swap.h> #include <linux/highmem.h> #include <linux/pagemap.h> |
9a8408951
|
48 |
#include <linux/ksm.h> |
1da177e4c
|
49 |
#include <linux/rmap.h> |
b95f1b31b
|
50 |
#include <linux/export.h> |
0ff922452
|
51 |
#include <linux/delayacct.h> |
1da177e4c
|
52 |
#include <linux/init.h> |
01c8f1c44
|
53 |
#include <linux/pfn_t.h> |
edc79b2a4
|
54 |
#include <linux/writeback.h> |
8a9f3ccd2
|
55 |
#include <linux/memcontrol.h> |
cddb8a5c1
|
56 |
#include <linux/mmu_notifier.h> |
3dc147414
|
57 58 59 |
#include <linux/kallsyms.h> #include <linux/swapops.h> #include <linux/elf.h> |
5a0e3ad6a
|
60 |
#include <linux/gfp.h> |
4daae3b4b
|
61 |
#include <linux/migrate.h> |
2fbc57c53
|
62 |
#include <linux/string.h> |
0abdd7a81
|
63 |
#include <linux/dma-debug.h> |
1592eef01
|
64 |
#include <linux/debugfs.h> |
6b251fc96
|
65 |
#include <linux/userfaultfd_k.h> |
bc2466e42
|
66 |
#include <linux/dax.h> |
1da177e4c
|
67 |
|
6952b61de
|
68 |
#include <asm/io.h> |
33a709b25
|
69 |
#include <asm/mmu_context.h> |
1da177e4c
|
70 71 72 73 74 |
#include <asm/pgalloc.h> #include <asm/uaccess.h> #include <asm/tlb.h> #include <asm/tlbflush.h> #include <asm/pgtable.h> |
42b777281
|
75 |
#include "internal.h" |
90572890d
|
76 77 |
#ifdef LAST_CPUPID_NOT_IN_PAGE_FLAGS #warning Unfortunate NUMA and NUMA Balancing config, growing page-frame for last_cpupid. |
75980e97d
|
78 |
#endif |
d41dee369
|
79 |
#ifndef CONFIG_NEED_MULTIPLE_NODES |
1da177e4c
|
80 81 82 83 84 85 86 |
/* use the per-pgdat data instead for discontigmem - mbligh */ unsigned long max_mapnr; struct page *mem_map; EXPORT_SYMBOL(max_mapnr); EXPORT_SYMBOL(mem_map); #endif |
1da177e4c
|
87 88 89 90 91 92 93 94 |
/* * A number of key systems in x86 including ioremap() rely on the assumption * that high_memory defines the upper bound on direct map memory, then end * of ZONE_NORMAL. Under CONFIG_DISCONTIG this means that max_low_pfn and * highstart_pfn must be the same; there must be no gap between ZONE_NORMAL * and ZONE_HIGHMEM. */ void * high_memory; |
1da177e4c
|
95 |
|
1da177e4c
|
96 |
EXPORT_SYMBOL(high_memory); |
1da177e4c
|
97 |
|
32a932332
|
98 99 100 101 102 103 104 105 106 107 108 109 |
/* * Randomize the address space (stacks, mmaps, brk, etc.). * * ( When CONFIG_COMPAT_BRK=y we exclude brk from randomization, * as ancient (libc5 based) binaries can segfault. ) */ int randomize_va_space __read_mostly = #ifdef CONFIG_COMPAT_BRK 1; #else 2; #endif |
a62eaf151
|
110 111 112 113 |
static int __init disable_randmaps(char *s) { randomize_va_space = 0; |
9b41046cd
|
114 |
return 1; |
a62eaf151
|
115 116 |
} __setup("norandmaps", disable_randmaps); |
62eede62d
|
117 |
unsigned long zero_pfn __read_mostly; |
03f6462a3
|
118 |
unsigned long highest_memmap_pfn __read_mostly; |
a13ea5b75
|
119 |
|
0b70068e4
|
120 |
EXPORT_SYMBOL(zero_pfn); |
a13ea5b75
|
121 122 123 124 125 126 127 128 129 |
/* * CONFIG_MMU architectures set up ZERO_PAGE in their paging_init() */ static int __init init_zero_pfn(void) { zero_pfn = page_to_pfn(ZERO_PAGE(0)); return 0; } core_initcall(init_zero_pfn); |
a62eaf151
|
130 |
|
d559db086
|
131 |
|
34e55232e
|
132 |
#if defined(SPLIT_RSS_COUNTING) |
ea48cf786
|
133 |
void sync_mm_rss(struct mm_struct *mm) |
34e55232e
|
134 135 136 137 |
{ int i; for (i = 0; i < NR_MM_COUNTERS; i++) { |
05af2e104
|
138 139 140 |
if (current->rss_stat.count[i]) { add_mm_counter(mm, i, current->rss_stat.count[i]); current->rss_stat.count[i] = 0; |
34e55232e
|
141 142 |
} } |
05af2e104
|
143 |
current->rss_stat.events = 0; |
34e55232e
|
144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 |
} static void add_mm_counter_fast(struct mm_struct *mm, int member, int val) { struct task_struct *task = current; if (likely(task->mm == mm)) task->rss_stat.count[member] += val; else add_mm_counter(mm, member, val); } #define inc_mm_counter_fast(mm, member) add_mm_counter_fast(mm, member, 1) #define dec_mm_counter_fast(mm, member) add_mm_counter_fast(mm, member, -1) /* sync counter once per 64 page faults */ #define TASK_RSS_EVENTS_THRESH (64) static void check_sync_rss_stat(struct task_struct *task) { if (unlikely(task != current)) return; if (unlikely(task->rss_stat.events++ > TASK_RSS_EVENTS_THRESH)) |
ea48cf786
|
165 |
sync_mm_rss(task->mm); |
34e55232e
|
166 |
} |
9547d01bf
|
167 |
#else /* SPLIT_RSS_COUNTING */ |
34e55232e
|
168 169 170 171 172 173 174 |
#define inc_mm_counter_fast(mm, member) inc_mm_counter(mm, member) #define dec_mm_counter_fast(mm, member) dec_mm_counter(mm, member) static void check_sync_rss_stat(struct task_struct *task) { } |
9547d01bf
|
175 176 177 |
#endif /* SPLIT_RSS_COUNTING */ #ifdef HAVE_GENERIC_MMU_GATHER |
ca1d6c7d9
|
178 |
static bool tlb_next_batch(struct mmu_gather *tlb) |
9547d01bf
|
179 180 181 182 183 184 |
{ struct mmu_gather_batch *batch; batch = tlb->active; if (batch->next) { tlb->active = batch->next; |
ca1d6c7d9
|
185 |
return true; |
9547d01bf
|
186 |
} |
53a59fc67
|
187 |
if (tlb->batch_count == MAX_GATHER_BATCH_COUNT) |
ca1d6c7d9
|
188 |
return false; |
53a59fc67
|
189 |
|
9547d01bf
|
190 191 |
batch = (void *)__get_free_pages(GFP_NOWAIT | __GFP_NOWARN, 0); if (!batch) |
ca1d6c7d9
|
192 |
return false; |
9547d01bf
|
193 |
|
53a59fc67
|
194 |
tlb->batch_count++; |
9547d01bf
|
195 196 197 198 199 200 |
batch->next = NULL; batch->nr = 0; batch->max = MAX_GATHER_BATCH; tlb->active->next = batch; tlb->active = batch; |
ca1d6c7d9
|
201 |
return true; |
9547d01bf
|
202 203 204 205 206 207 208 |
} /* tlb_gather_mmu * Called to initialize an (on-stack) mmu_gather structure for page-table * tear-down from @mm. The @fullmm argument is used when @mm is without * users and we're going to destroy the full address space (exit/execve). */ |
2b047252d
|
209 |
void tlb_gather_mmu(struct mmu_gather *tlb, struct mm_struct *mm, unsigned long start, unsigned long end) |
9547d01bf
|
210 211 |
{ tlb->mm = mm; |
2b047252d
|
212 213 |
/* Is it from 0 to ~0? */ tlb->fullmm = !(start | (end+1)); |
1de14c3c5
|
214 |
tlb->need_flush_all = 0; |
9547d01bf
|
215 216 217 218 |
tlb->local.next = NULL; tlb->local.nr = 0; tlb->local.max = ARRAY_SIZE(tlb->__pages); tlb->active = &tlb->local; |
53a59fc67
|
219 |
tlb->batch_count = 0; |
9547d01bf
|
220 221 222 223 |
#ifdef CONFIG_HAVE_RCU_TABLE_FREE tlb->batch = NULL; #endif |
e77b0852b
|
224 |
tlb->page_size = 0; |
fb7332a9f
|
225 226 |
__tlb_reset_range(tlb); |
9547d01bf
|
227 |
} |
1cf35d477
|
228 |
static void tlb_flush_mmu_tlbonly(struct mmu_gather *tlb) |
9547d01bf
|
229 |
{ |
721c21c17
|
230 231 |
if (!tlb->end) return; |
9547d01bf
|
232 |
tlb_flush(tlb); |
34ee645e8
|
233 |
mmu_notifier_invalidate_range(tlb->mm, tlb->start, tlb->end); |
9547d01bf
|
234 235 |
#ifdef CONFIG_HAVE_RCU_TABLE_FREE tlb_table_flush(tlb); |
34e55232e
|
236 |
#endif |
fb7332a9f
|
237 |
__tlb_reset_range(tlb); |
1cf35d477
|
238 239 240 241 242 |
} static void tlb_flush_mmu_free(struct mmu_gather *tlb) { struct mmu_gather_batch *batch; |
34e55232e
|
243 |
|
721c21c17
|
244 |
for (batch = &tlb->local; batch && batch->nr; batch = batch->next) { |
9547d01bf
|
245 246 247 248 249 |
free_pages_and_swap_cache(batch->pages, batch->nr); batch->nr = 0; } tlb->active = &tlb->local; } |
1cf35d477
|
250 251 |
void tlb_flush_mmu(struct mmu_gather *tlb) { |
1cf35d477
|
252 253 254 |
tlb_flush_mmu_tlbonly(tlb); tlb_flush_mmu_free(tlb); } |
9547d01bf
|
255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 |
/* tlb_finish_mmu * Called at the end of the shootdown operation to free up any resources * that were required. */ void tlb_finish_mmu(struct mmu_gather *tlb, unsigned long start, unsigned long end) { struct mmu_gather_batch *batch, *next; tlb_flush_mmu(tlb); /* keep the page table cache within bounds */ check_pgt_cache(); for (batch = tlb->local.next; batch; batch = next) { next = batch->next; free_pages((unsigned long)batch, 0); } tlb->local.next = NULL; } /* __tlb_remove_page * Must perform the equivalent to __free_pte(pte_get_and_clear(ptep)), while * handling the additional races in SMP caused by other CPUs caching valid * mappings in their TLBs. Returns the number of free page slots left. * When out of page slots we must call tlb_flush_mmu(). |
e9d55e157
|
280 |
*returns true if the caller should flush. |
9547d01bf
|
281 |
*/ |
e77b0852b
|
282 |
bool __tlb_remove_page_size(struct mmu_gather *tlb, struct page *page, int page_size) |
9547d01bf
|
283 284 |
{ struct mmu_gather_batch *batch; |
fb7332a9f
|
285 |
VM_BUG_ON(!tlb->end); |
9547d01bf
|
286 |
|
e77b0852b
|
287 288 289 290 291 292 |
if (!tlb->page_size) tlb->page_size = page_size; else { if (page_size != tlb->page_size) return true; } |
9547d01bf
|
293 |
batch = tlb->active; |
9547d01bf
|
294 295 |
if (batch->nr == batch->max) { if (!tlb_next_batch(tlb)) |
e9d55e157
|
296 |
return true; |
0b43c3aab
|
297 |
batch = tlb->active; |
9547d01bf
|
298 |
} |
309381fea
|
299 |
VM_BUG_ON_PAGE(batch->nr > batch->max, page); |
9547d01bf
|
300 |
|
e9d55e157
|
301 302 |
batch->pages[batch->nr++] = page; return false; |
9547d01bf
|
303 304 305 |
} #endif /* HAVE_GENERIC_MMU_GATHER */ |
267239116
|
306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 |
#ifdef CONFIG_HAVE_RCU_TABLE_FREE /* * See the comment near struct mmu_table_batch. */ static void tlb_remove_table_smp_sync(void *arg) { /* Simply deliver the interrupt */ } static void tlb_remove_table_one(void *table) { /* * This isn't an RCU grace period and hence the page-tables cannot be * assumed to be actually RCU-freed. * * It is however sufficient for software page-table walkers that rely on * IRQ disabling. See the comment near struct mmu_table_batch. */ smp_call_function(tlb_remove_table_smp_sync, NULL, 1); __tlb_remove_table(table); } static void tlb_remove_table_rcu(struct rcu_head *head) { struct mmu_table_batch *batch; int i; batch = container_of(head, struct mmu_table_batch, rcu); for (i = 0; i < batch->nr; i++) __tlb_remove_table(batch->tables[i]); free_page((unsigned long)batch); } void tlb_table_flush(struct mmu_gather *tlb) { struct mmu_table_batch **batch = &tlb->batch; if (*batch) { call_rcu_sched(&(*batch)->rcu, tlb_remove_table_rcu); *batch = NULL; } } void tlb_remove_table(struct mmu_gather *tlb, void *table) { struct mmu_table_batch **batch = &tlb->batch; |
267239116
|
356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 |
/* * When there's less then two users of this mm there cannot be a * concurrent page-table walk. */ if (atomic_read(&tlb->mm->mm_users) < 2) { __tlb_remove_table(table); return; } if (*batch == NULL) { *batch = (struct mmu_table_batch *)__get_free_page(GFP_NOWAIT | __GFP_NOWARN); if (*batch == NULL) { tlb_remove_table_one(table); return; } (*batch)->nr = 0; } (*batch)->tables[(*batch)->nr++] = table; if ((*batch)->nr == MAX_TABLE_BATCH) tlb_table_flush(tlb); } |
9547d01bf
|
377 |
#endif /* CONFIG_HAVE_RCU_TABLE_FREE */ |
267239116
|
378 |
|
1da177e4c
|
379 |
/* |
1da177e4c
|
380 381 382 |
* Note: this doesn't free the actual pages themselves. That * has been handled earlier when unmapping all the memory regions. */ |
9e1b32caa
|
383 384 |
static void free_pte_range(struct mmu_gather *tlb, pmd_t *pmd, unsigned long addr) |
1da177e4c
|
385 |
{ |
2f569afd9
|
386 |
pgtable_t token = pmd_pgtable(*pmd); |
e0da382c9
|
387 |
pmd_clear(pmd); |
9e1b32caa
|
388 |
pte_free_tlb(tlb, token, addr); |
e1f56c89b
|
389 |
atomic_long_dec(&tlb->mm->nr_ptes); |
1da177e4c
|
390 |
} |
e0da382c9
|
391 392 393 |
static inline void free_pmd_range(struct mmu_gather *tlb, pud_t *pud, unsigned long addr, unsigned long end, unsigned long floor, unsigned long ceiling) |
1da177e4c
|
394 395 396 |
{ pmd_t *pmd; unsigned long next; |
e0da382c9
|
397 |
unsigned long start; |
1da177e4c
|
398 |
|
e0da382c9
|
399 |
start = addr; |
1da177e4c
|
400 |
pmd = pmd_offset(pud, addr); |
1da177e4c
|
401 402 403 404 |
do { next = pmd_addr_end(addr, end); if (pmd_none_or_clear_bad(pmd)) continue; |
9e1b32caa
|
405 |
free_pte_range(tlb, pmd, addr); |
1da177e4c
|
406 |
} while (pmd++, addr = next, addr != end); |
e0da382c9
|
407 408 409 410 411 412 413 |
start &= PUD_MASK; if (start < floor) return; if (ceiling) { ceiling &= PUD_MASK; if (!ceiling) return; |
1da177e4c
|
414 |
} |
e0da382c9
|
415 416 417 418 419 |
if (end - 1 > ceiling - 1) return; pmd = pmd_offset(pud, start); pud_clear(pud); |
9e1b32caa
|
420 |
pmd_free_tlb(tlb, pmd, start); |
dc6c9a35b
|
421 |
mm_dec_nr_pmds(tlb->mm); |
1da177e4c
|
422 |
} |
e0da382c9
|
423 424 425 |
static inline void free_pud_range(struct mmu_gather *tlb, pgd_t *pgd, unsigned long addr, unsigned long end, unsigned long floor, unsigned long ceiling) |
1da177e4c
|
426 427 428 |
{ pud_t *pud; unsigned long next; |
e0da382c9
|
429 |
unsigned long start; |
1da177e4c
|
430 |
|
e0da382c9
|
431 |
start = addr; |
1da177e4c
|
432 |
pud = pud_offset(pgd, addr); |
1da177e4c
|
433 434 435 436 |
do { next = pud_addr_end(addr, end); if (pud_none_or_clear_bad(pud)) continue; |
e0da382c9
|
437 |
free_pmd_range(tlb, pud, addr, next, floor, ceiling); |
1da177e4c
|
438 |
} while (pud++, addr = next, addr != end); |
e0da382c9
|
439 440 441 442 443 444 445 |
start &= PGDIR_MASK; if (start < floor) return; if (ceiling) { ceiling &= PGDIR_MASK; if (!ceiling) return; |
1da177e4c
|
446 |
} |
e0da382c9
|
447 448 449 450 451 |
if (end - 1 > ceiling - 1) return; pud = pud_offset(pgd, start); pgd_clear(pgd); |
9e1b32caa
|
452 |
pud_free_tlb(tlb, pud, start); |
1da177e4c
|
453 454 455 |
} /* |
e0da382c9
|
456 |
* This function frees user-level page tables of a process. |
1da177e4c
|
457 |
*/ |
42b777281
|
458 |
void free_pgd_range(struct mmu_gather *tlb, |
e0da382c9
|
459 460 |
unsigned long addr, unsigned long end, unsigned long floor, unsigned long ceiling) |
1da177e4c
|
461 462 463 |
{ pgd_t *pgd; unsigned long next; |
e0da382c9
|
464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 |
/* * The next few lines have given us lots of grief... * * Why are we testing PMD* at this top level? Because often * there will be no work to do at all, and we'd prefer not to * go all the way down to the bottom just to discover that. * * Why all these "- 1"s? Because 0 represents both the bottom * of the address space and the top of it (using -1 for the * top wouldn't help much: the masks would do the wrong thing). * The rule is that addr 0 and floor 0 refer to the bottom of * the address space, but end 0 and ceiling 0 refer to the top * Comparisons need to use "end - 1" and "ceiling - 1" (though * that end 0 case should be mythical). * * Wherever addr is brought up or ceiling brought down, we must * be careful to reject "the opposite 0" before it confuses the * subsequent tests. But what about where end is brought down * by PMD_SIZE below? no, end can't go down to 0 there. * * Whereas we round start (addr) and ceiling down, by different * masks at different levels, in order to test whether a table * now has no other vmas using it, so can be freed, we don't * bother to round floor or end up - the tests don't need that. */ |
1da177e4c
|
490 |
|
e0da382c9
|
491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 |
addr &= PMD_MASK; if (addr < floor) { addr += PMD_SIZE; if (!addr) return; } if (ceiling) { ceiling &= PMD_MASK; if (!ceiling) return; } if (end - 1 > ceiling - 1) end -= PMD_SIZE; if (addr > end - 1) return; |
42b777281
|
506 |
pgd = pgd_offset(tlb->mm, addr); |
1da177e4c
|
507 508 509 510 |
do { next = pgd_addr_end(addr, end); if (pgd_none_or_clear_bad(pgd)) continue; |
42b777281
|
511 |
free_pud_range(tlb, pgd, addr, next, floor, ceiling); |
1da177e4c
|
512 |
} while (pgd++, addr = next, addr != end); |
e0da382c9
|
513 |
} |
42b777281
|
514 |
void free_pgtables(struct mmu_gather *tlb, struct vm_area_struct *vma, |
3bf5ee956
|
515 |
unsigned long floor, unsigned long ceiling) |
e0da382c9
|
516 517 518 519 |
{ while (vma) { struct vm_area_struct *next = vma->vm_next; unsigned long addr = vma->vm_start; |
8f4f8c164
|
520 |
/* |
25d9e2d15
|
521 522 |
* Hide vma from rmap and truncate_pagecache before freeing * pgtables |
8f4f8c164
|
523 |
*/ |
5beb49305
|
524 |
unlink_anon_vmas(vma); |
8f4f8c164
|
525 |
unlink_file_vma(vma); |
9da61aef0
|
526 |
if (is_vm_hugetlb_page(vma)) { |
3bf5ee956
|
527 |
hugetlb_free_pgd_range(tlb, addr, vma->vm_end, |
e0da382c9
|
528 |
floor, next? next->vm_start: ceiling); |
3bf5ee956
|
529 530 531 532 533 |
} else { /* * Optimization: gather nearby vmas into one call down */ while (next && next->vm_start <= vma->vm_end + PMD_SIZE |
4866920b9
|
534 |
&& !is_vm_hugetlb_page(next)) { |
3bf5ee956
|
535 536 |
vma = next; next = vma->vm_next; |
5beb49305
|
537 |
unlink_anon_vmas(vma); |
8f4f8c164
|
538 |
unlink_file_vma(vma); |
3bf5ee956
|
539 540 541 542 |
} free_pgd_range(tlb, addr, vma->vm_end, floor, next? next->vm_start: ceiling); } |
e0da382c9
|
543 544 |
vma = next; } |
1da177e4c
|
545 |
} |
3ed3a4f0d
|
546 |
int __pte_alloc(struct mm_struct *mm, pmd_t *pmd, unsigned long address) |
1da177e4c
|
547 |
{ |
c4088ebdc
|
548 |
spinlock_t *ptl; |
2f569afd9
|
549 |
pgtable_t new = pte_alloc_one(mm, address); |
1bb3630e8
|
550 551 |
if (!new) return -ENOMEM; |
362a61ad6
|
552 553 554 555 556 557 558 559 560 561 562 563 564 565 |
/* * Ensure all pte setup (eg. pte page lock and page clearing) are * visible before the pte is made visible to other CPUs by being * put into page tables. * * The other side of the story is the pointer chasing in the page * table walking code (when walking the page table without locking; * ie. most of the time). Fortunately, these data accesses consist * of a chain of data-dependent loads, meaning most CPUs (alpha * being the notable exception) will already guarantee loads are * seen in-order. See the alpha page table accessors for the * smp_read_barrier_depends() barriers in page table walking code. */ smp_wmb(); /* Could be smp_wmb__xxx(before|after)_spin_lock */ |
c4088ebdc
|
566 |
ptl = pmd_lock(mm, pmd); |
8ac1f8320
|
567 |
if (likely(pmd_none(*pmd))) { /* Has another populated it ? */ |
e1f56c89b
|
568 |
atomic_long_inc(&mm->nr_ptes); |
1da177e4c
|
569 |
pmd_populate(mm, pmd, new); |
2f569afd9
|
570 |
new = NULL; |
4b471e889
|
571 |
} |
c4088ebdc
|
572 |
spin_unlock(ptl); |
2f569afd9
|
573 574 |
if (new) pte_free(mm, new); |
1bb3630e8
|
575 |
return 0; |
1da177e4c
|
576 |
} |
1bb3630e8
|
577 |
int __pte_alloc_kernel(pmd_t *pmd, unsigned long address) |
1da177e4c
|
578 |
{ |
1bb3630e8
|
579 580 581 |
pte_t *new = pte_alloc_one_kernel(&init_mm, address); if (!new) return -ENOMEM; |
362a61ad6
|
582 |
smp_wmb(); /* See comment in __pte_alloc */ |
1bb3630e8
|
583 |
spin_lock(&init_mm.page_table_lock); |
8ac1f8320
|
584 |
if (likely(pmd_none(*pmd))) { /* Has another populated it ? */ |
1bb3630e8
|
585 |
pmd_populate_kernel(&init_mm, pmd, new); |
2f569afd9
|
586 |
new = NULL; |
4b471e889
|
587 |
} |
1bb3630e8
|
588 |
spin_unlock(&init_mm.page_table_lock); |
2f569afd9
|
589 590 |
if (new) pte_free_kernel(&init_mm, new); |
1bb3630e8
|
591 |
return 0; |
1da177e4c
|
592 |
} |
d559db086
|
593 594 595 596 597 598 |
static inline void init_rss_vec(int *rss) { memset(rss, 0, sizeof(int) * NR_MM_COUNTERS); } static inline void add_mm_rss_vec(struct mm_struct *mm, int *rss) |
ae8597623
|
599 |
{ |
d559db086
|
600 |
int i; |
34e55232e
|
601 |
if (current->mm == mm) |
05af2e104
|
602 |
sync_mm_rss(mm); |
d559db086
|
603 604 605 |
for (i = 0; i < NR_MM_COUNTERS; i++) if (rss[i]) add_mm_counter(mm, i, rss[i]); |
ae8597623
|
606 |
} |
1da177e4c
|
607 |
/* |
6aab341e0
|
608 609 610 |
* This function is called to print an error when a bad pte * is found. For example, we might have a PFN-mapped pte in * a region that doesn't allow it. |
b5810039a
|
611 612 613 |
* * The calling function must still handle the error. */ |
3dc147414
|
614 615 |
static void print_bad_pte(struct vm_area_struct *vma, unsigned long addr, pte_t pte, struct page *page) |
b5810039a
|
616 |
{ |
3dc147414
|
617 618 619 620 621 |
pgd_t *pgd = pgd_offset(vma->vm_mm, addr); pud_t *pud = pud_offset(pgd, addr); pmd_t *pmd = pmd_offset(pud, addr); struct address_space *mapping; pgoff_t index; |
d936cf9b3
|
622 623 624 625 626 627 628 629 630 631 632 633 634 635 |
static unsigned long resume; static unsigned long nr_shown; static unsigned long nr_unshown; /* * Allow a burst of 60 reports, then keep quiet for that minute; * or allow a steady drip of one report per second. */ if (nr_shown == 60) { if (time_before(jiffies, resume)) { nr_unshown++; return; } if (nr_unshown) { |
1170532bb
|
636 637 638 |
pr_alert("BUG: Bad page map: %lu messages suppressed ", nr_unshown); |
d936cf9b3
|
639 640 641 642 643 644 |
nr_unshown = 0; } nr_shown = 0; } if (nr_shown++ == 0) resume = jiffies + 60 * HZ; |
3dc147414
|
645 646 647 |
mapping = vma->vm_file ? vma->vm_file->f_mapping : NULL; index = linear_page_index(vma, addr); |
1170532bb
|
648 649 650 651 |
pr_alert("BUG: Bad page map in process %s pte:%08llx pmd:%08llx ", current->comm, (long long)pte_val(pte), (long long)pmd_val(*pmd)); |
718a38211
|
652 |
if (page) |
f0b791a34
|
653 |
dump_page(page, "bad pte"); |
1170532bb
|
654 655 656 |
pr_alert("addr:%p vm_flags:%08lx anon_vma:%p mapping:%p index:%lx ", (void *)addr, vma->vm_flags, vma->anon_vma, mapping, index); |
3dc147414
|
657 658 659 |
/* * Choose text because data symbols depend on CONFIG_KALLSYMS_ALL=y */ |
2682582a6
|
660 661 662 663 664 665 |
pr_alert("file:%pD fault:%pf mmap:%pf readpage:%pf ", vma->vm_file, vma->vm_ops ? vma->vm_ops->fault : NULL, vma->vm_file ? vma->vm_file->f_op->mmap : NULL, mapping ? mapping->a_ops->readpage : NULL); |
b5810039a
|
666 |
dump_stack(); |
373d4d099
|
667 |
add_taint(TAINT_BAD_PAGE, LOCKDEP_NOW_UNRELIABLE); |
b5810039a
|
668 669 670 |
} /* |
7e675137a
|
671 |
* vm_normal_page -- This function gets the "struct page" associated with a pte. |
6aab341e0
|
672 |
* |
7e675137a
|
673 674 675 |
* "Special" mappings do not wish to be associated with a "struct page" (either * it doesn't exist, or it exists but they don't want to touch it). In this * case, NULL is returned here. "Normal" mappings do have a struct page. |
b379d7901
|
676 |
* |
7e675137a
|
677 678 679 680 681 682 683 684 |
* There are 2 broad cases. Firstly, an architecture may define a pte_special() * pte bit, in which case this function is trivial. Secondly, an architecture * may not have a spare pte bit, which requires a more complicated scheme, * described below. * * A raw VM_PFNMAP mapping (ie. one that is not COWed) is always considered a * special mapping (even if there are underlying and valid "struct pages"). * COWed pages of a VM_PFNMAP are always normal. |
6aab341e0
|
685 |
* |
b379d7901
|
686 687 |
* The way we recognize COWed pages within VM_PFNMAP mappings is through the * rules set up by "remap_pfn_range()": the vma will have the VM_PFNMAP bit |
7e675137a
|
688 689 |
* set, and the vm_pgoff will point to the first PFN mapped: thus every special * mapping will always honor the rule |
6aab341e0
|
690 691 692 |
* * pfn_of_page == vma->vm_pgoff + ((addr - vma->vm_start) >> PAGE_SHIFT) * |
7e675137a
|
693 694 695 696 697 698 |
* And for normal mappings this is false. * * This restricts such mappings to be a linear translation from virtual address * to pfn. To get around this restriction, we allow arbitrary mappings so long * as the vma is not a COW mapping; in that case, we know that all ptes are * special (because none can have been COWed). |
b379d7901
|
699 |
* |
b379d7901
|
700 |
* |
7e675137a
|
701 |
* In order to support COW of arbitrary special mappings, we have VM_MIXEDMAP. |
b379d7901
|
702 703 704 705 706 707 708 709 710 |
* * VM_MIXEDMAP mappings can likewise contain memory with or without "struct * page" backing, however the difference is that _all_ pages with a struct * page (that is, those where pfn_valid is true) are refcounted and considered * normal pages by the VM. The disadvantage is that pages are refcounted * (which can be slower and simply not an option for some PFNMAP users). The * advantage is that we don't have to follow the strict linearity rule of * PFNMAP mappings in order to support COWable mappings. * |
ee498ed73
|
711 |
*/ |
7e675137a
|
712 713 714 715 716 717 718 |
#ifdef __HAVE_ARCH_PTE_SPECIAL # define HAVE_PTE_SPECIAL 1 #else # define HAVE_PTE_SPECIAL 0 #endif struct page *vm_normal_page(struct vm_area_struct *vma, unsigned long addr, pte_t pte) |
ee498ed73
|
719 |
{ |
22b31eec6
|
720 |
unsigned long pfn = pte_pfn(pte); |
7e675137a
|
721 722 |
if (HAVE_PTE_SPECIAL) { |
b38af4721
|
723 |
if (likely(!pte_special(pte))) |
22b31eec6
|
724 |
goto check_pfn; |
667a0a06c
|
725 726 |
if (vma->vm_ops && vma->vm_ops->find_special_page) return vma->vm_ops->find_special_page(vma, addr); |
a13ea5b75
|
727 728 |
if (vma->vm_flags & (VM_PFNMAP | VM_MIXEDMAP)) return NULL; |
62eede62d
|
729 |
if (!is_zero_pfn(pfn)) |
22b31eec6
|
730 |
print_bad_pte(vma, addr, pte, NULL); |
7e675137a
|
731 732 733 734 |
return NULL; } /* !HAVE_PTE_SPECIAL case follows: */ |
b379d7901
|
735 736 737 738 739 740 |
if (unlikely(vma->vm_flags & (VM_PFNMAP|VM_MIXEDMAP))) { if (vma->vm_flags & VM_MIXEDMAP) { if (!pfn_valid(pfn)) return NULL; goto out; } else { |
7e675137a
|
741 742 |
unsigned long off; off = (addr - vma->vm_start) >> PAGE_SHIFT; |
b379d7901
|
743 744 745 746 747 |
if (pfn == vma->vm_pgoff + off) return NULL; if (!is_cow_mapping(vma->vm_flags)) return NULL; } |
6aab341e0
|
748 |
} |
b38af4721
|
749 750 |
if (is_zero_pfn(pfn)) return NULL; |
22b31eec6
|
751 752 753 754 755 |
check_pfn: if (unlikely(pfn > highest_memmap_pfn)) { print_bad_pte(vma, addr, pte, NULL); return NULL; } |
6aab341e0
|
756 757 |
/* |
7e675137a
|
758 |
* NOTE! We still have PageReserved() pages in the page tables. |
7e675137a
|
759 |
* eg. VDSO mappings can cause them to exist. |
6aab341e0
|
760 |
*/ |
b379d7901
|
761 |
out: |
6aab341e0
|
762 |
return pfn_to_page(pfn); |
ee498ed73
|
763 |
} |
28093f9f3
|
764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 |
#ifdef CONFIG_TRANSPARENT_HUGEPAGE struct page *vm_normal_page_pmd(struct vm_area_struct *vma, unsigned long addr, pmd_t pmd) { unsigned long pfn = pmd_pfn(pmd); /* * There is no pmd_special() but there may be special pmds, e.g. * in a direct-access (dax) mapping, so let's just replicate the * !HAVE_PTE_SPECIAL case from vm_normal_page() here. */ if (unlikely(vma->vm_flags & (VM_PFNMAP|VM_MIXEDMAP))) { if (vma->vm_flags & VM_MIXEDMAP) { if (!pfn_valid(pfn)) return NULL; goto out; } else { unsigned long off; off = (addr - vma->vm_start) >> PAGE_SHIFT; if (pfn == vma->vm_pgoff + off) return NULL; if (!is_cow_mapping(vma->vm_flags)) return NULL; } } if (is_zero_pfn(pfn)) return NULL; if (unlikely(pfn > highest_memmap_pfn)) return NULL; /* * NOTE! We still have PageReserved() pages in the page tables. * eg. VDSO mappings can cause them to exist. */ out: return pfn_to_page(pfn); } #endif |
ee498ed73
|
803 |
/* |
1da177e4c
|
804 805 806 |
* copy one vm_area from one task to the other. Assumes the page tables * already present in the new task to be cleared in the whole range * covered by this vma. |
1da177e4c
|
807 |
*/ |
570a335b8
|
808 |
static inline unsigned long |
1da177e4c
|
809 |
copy_one_pte(struct mm_struct *dst_mm, struct mm_struct *src_mm, |
b5810039a
|
810 |
pte_t *dst_pte, pte_t *src_pte, struct vm_area_struct *vma, |
8c1037627
|
811 |
unsigned long addr, int *rss) |
1da177e4c
|
812 |
{ |
b5810039a
|
813 |
unsigned long vm_flags = vma->vm_flags; |
1da177e4c
|
814 815 |
pte_t pte = *src_pte; struct page *page; |
1da177e4c
|
816 817 818 |
/* pte contains position in swap or file, so copy. */ if (unlikely(!pte_present(pte))) { |
0661a3361
|
819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 |
swp_entry_t entry = pte_to_swp_entry(pte); if (likely(!non_swap_entry(entry))) { if (swap_duplicate(entry) < 0) return entry.val; /* make sure dst_mm is on swapoff's mmlist. */ if (unlikely(list_empty(&dst_mm->mmlist))) { spin_lock(&mmlist_lock); if (list_empty(&dst_mm->mmlist)) list_add(&dst_mm->mmlist, &src_mm->mmlist); spin_unlock(&mmlist_lock); } rss[MM_SWAPENTS]++; } else if (is_migration_entry(entry)) { page = migration_entry_to_page(entry); |
eca56ff90
|
836 |
rss[mm_counter(page)]++; |
0661a3361
|
837 838 839 840 841 842 843 844 845 846 847 848 |
if (is_write_migration_entry(entry) && is_cow_mapping(vm_flags)) { /* * COW mappings require pages in both * parent and child to be set to read. */ make_migration_entry_read(&entry); pte = swp_entry_to_pte(entry); if (pte_swp_soft_dirty(*src_pte)) pte = pte_swp_mksoft_dirty(pte); set_pte_at(src_mm, addr, src_pte, pte); |
0697212a4
|
849 |
} |
1da177e4c
|
850 |
} |
ae8597623
|
851 |
goto out_set_pte; |
1da177e4c
|
852 |
} |
1da177e4c
|
853 854 855 856 |
/* * If it's a COW mapping, write protect it both * in the parent and the child */ |
67121172f
|
857 |
if (is_cow_mapping(vm_flags)) { |
1da177e4c
|
858 |
ptep_set_wrprotect(src_mm, addr, src_pte); |
3dc907951
|
859 |
pte = pte_wrprotect(pte); |
1da177e4c
|
860 861 862 863 864 865 866 867 868 |
} /* * If it's a shared mapping, mark it clean in * the child */ if (vm_flags & VM_SHARED) pte = pte_mkclean(pte); pte = pte_mkold(pte); |
6aab341e0
|
869 870 871 872 |
page = vm_normal_page(vma, addr, pte); if (page) { get_page(page); |
53f9263ba
|
873 |
page_dup_rmap(page, false); |
eca56ff90
|
874 |
rss[mm_counter(page)]++; |
6aab341e0
|
875 |
} |
ae8597623
|
876 877 878 |
out_set_pte: set_pte_at(dst_mm, addr, dst_pte, pte); |
570a335b8
|
879 |
return 0; |
1da177e4c
|
880 |
} |
21bda264f
|
881 |
static int copy_pte_range(struct mm_struct *dst_mm, struct mm_struct *src_mm, |
71e3aac07
|
882 883 |
pmd_t *dst_pmd, pmd_t *src_pmd, struct vm_area_struct *vma, unsigned long addr, unsigned long end) |
1da177e4c
|
884 |
{ |
c36987e2e
|
885 |
pte_t *orig_src_pte, *orig_dst_pte; |
1da177e4c
|
886 |
pte_t *src_pte, *dst_pte; |
c74df32c7
|
887 |
spinlock_t *src_ptl, *dst_ptl; |
e040f218b
|
888 |
int progress = 0; |
d559db086
|
889 |
int rss[NR_MM_COUNTERS]; |
570a335b8
|
890 |
swp_entry_t entry = (swp_entry_t){0}; |
1da177e4c
|
891 892 |
again: |
d559db086
|
893 |
init_rss_vec(rss); |
c74df32c7
|
894 |
dst_pte = pte_alloc_map_lock(dst_mm, dst_pmd, addr, &dst_ptl); |
1da177e4c
|
895 896 |
if (!dst_pte) return -ENOMEM; |
ece0e2b64
|
897 |
src_pte = pte_offset_map(src_pmd, addr); |
4c21e2f24
|
898 |
src_ptl = pte_lockptr(src_mm, src_pmd); |
f20dc5f7c
|
899 |
spin_lock_nested(src_ptl, SINGLE_DEPTH_NESTING); |
c36987e2e
|
900 901 |
orig_src_pte = src_pte; orig_dst_pte = dst_pte; |
6606c3e0d
|
902 |
arch_enter_lazy_mmu_mode(); |
1da177e4c
|
903 |
|
1da177e4c
|
904 905 906 907 908 |
do { /* * We are holding two locks at this point - either of them * could generate latencies in another task on another CPU. */ |
e040f218b
|
909 910 911 |
if (progress >= 32) { progress = 0; if (need_resched() || |
95c354fe9
|
912 |
spin_needbreak(src_ptl) || spin_needbreak(dst_ptl)) |
e040f218b
|
913 914 |
break; } |
1da177e4c
|
915 916 917 918 |
if (pte_none(*src_pte)) { progress++; continue; } |
570a335b8
|
919 920 921 922 |
entry.val = copy_one_pte(dst_mm, src_mm, dst_pte, src_pte, vma, addr, rss); if (entry.val) break; |
1da177e4c
|
923 924 |
progress += 8; } while (dst_pte++, src_pte++, addr += PAGE_SIZE, addr != end); |
1da177e4c
|
925 |
|
6606c3e0d
|
926 |
arch_leave_lazy_mmu_mode(); |
c74df32c7
|
927 |
spin_unlock(src_ptl); |
ece0e2b64
|
928 |
pte_unmap(orig_src_pte); |
d559db086
|
929 |
add_mm_rss_vec(dst_mm, rss); |
c36987e2e
|
930 |
pte_unmap_unlock(orig_dst_pte, dst_ptl); |
c74df32c7
|
931 |
cond_resched(); |
570a335b8
|
932 933 934 935 936 937 |
if (entry.val) { if (add_swap_count_continuation(entry, GFP_KERNEL) < 0) return -ENOMEM; progress = 0; } |
1da177e4c
|
938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 |
if (addr != end) goto again; return 0; } static inline int copy_pmd_range(struct mm_struct *dst_mm, struct mm_struct *src_mm, pud_t *dst_pud, pud_t *src_pud, struct vm_area_struct *vma, unsigned long addr, unsigned long end) { pmd_t *src_pmd, *dst_pmd; unsigned long next; dst_pmd = pmd_alloc(dst_mm, dst_pud, addr); if (!dst_pmd) return -ENOMEM; src_pmd = pmd_offset(src_pud, addr); do { next = pmd_addr_end(addr, end); |
5c7fb56e5
|
956 |
if (pmd_trans_huge(*src_pmd) || pmd_devmap(*src_pmd)) { |
71e3aac07
|
957 |
int err; |
14d1a55cd
|
958 |
VM_BUG_ON(next-addr != HPAGE_PMD_SIZE); |
71e3aac07
|
959 960 961 962 963 964 965 966 |
err = copy_huge_pmd(dst_mm, src_mm, dst_pmd, src_pmd, addr, vma); if (err == -ENOMEM) return -ENOMEM; if (!err) continue; /* fall through */ } |
1da177e4c
|
967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 |
if (pmd_none_or_clear_bad(src_pmd)) continue; if (copy_pte_range(dst_mm, src_mm, dst_pmd, src_pmd, vma, addr, next)) return -ENOMEM; } while (dst_pmd++, src_pmd++, addr = next, addr != end); return 0; } static inline int copy_pud_range(struct mm_struct *dst_mm, struct mm_struct *src_mm, pgd_t *dst_pgd, pgd_t *src_pgd, struct vm_area_struct *vma, unsigned long addr, unsigned long end) { pud_t *src_pud, *dst_pud; unsigned long next; dst_pud = pud_alloc(dst_mm, dst_pgd, addr); if (!dst_pud) return -ENOMEM; src_pud = pud_offset(src_pgd, addr); do { next = pud_addr_end(addr, end); if (pud_none_or_clear_bad(src_pud)) continue; if (copy_pmd_range(dst_mm, src_mm, dst_pud, src_pud, vma, addr, next)) return -ENOMEM; } while (dst_pud++, src_pud++, addr = next, addr != end); return 0; } int copy_page_range(struct mm_struct *dst_mm, struct mm_struct *src_mm, struct vm_area_struct *vma) { pgd_t *src_pgd, *dst_pgd; unsigned long next; unsigned long addr = vma->vm_start; unsigned long end = vma->vm_end; |
2ec74c3ef
|
1005 1006 1007 |
unsigned long mmun_start; /* For mmu_notifiers */ unsigned long mmun_end; /* For mmu_notifiers */ bool is_cow; |
cddb8a5c1
|
1008 |
int ret; |
1da177e4c
|
1009 |
|
d992895ba
|
1010 1011 1012 1013 1014 1015 |
/* * Don't copy ptes where a page fault will fill them correctly. * Fork becomes much lighter when there are big shared or private * readonly mappings. The tradeoff is that copy_page_range is more * efficient than faulting. */ |
0661a3361
|
1016 1017 1018 |
if (!(vma->vm_flags & (VM_HUGETLB | VM_PFNMAP | VM_MIXEDMAP)) && !vma->anon_vma) return 0; |
d992895ba
|
1019 |
|
1da177e4c
|
1020 1021 |
if (is_vm_hugetlb_page(vma)) return copy_hugetlb_page_range(dst_mm, src_mm, vma); |
b3b9c2932
|
1022 |
if (unlikely(vma->vm_flags & VM_PFNMAP)) { |
2ab640379
|
1023 1024 1025 1026 |
/* * We do not free on error cases below as remove_vma * gets called on error from higher level routine */ |
5180da410
|
1027 |
ret = track_pfn_copy(vma); |
2ab640379
|
1028 1029 1030 |
if (ret) return ret; } |
cddb8a5c1
|
1031 1032 1033 1034 1035 1036 |
/* * We need to invalidate the secondary MMU mappings only when * there could be a permission downgrade on the ptes of the * parent mm. And a permission downgrade will only happen if * is_cow_mapping() returns true. */ |
2ec74c3ef
|
1037 1038 1039 1040 1041 1042 |
is_cow = is_cow_mapping(vma->vm_flags); mmun_start = addr; mmun_end = end; if (is_cow) mmu_notifier_invalidate_range_start(src_mm, mmun_start, mmun_end); |
cddb8a5c1
|
1043 1044 |
ret = 0; |
1da177e4c
|
1045 1046 1047 1048 1049 1050 |
dst_pgd = pgd_offset(dst_mm, addr); src_pgd = pgd_offset(src_mm, addr); do { next = pgd_addr_end(addr, end); if (pgd_none_or_clear_bad(src_pgd)) continue; |
cddb8a5c1
|
1051 1052 1053 1054 1055 |
if (unlikely(copy_pud_range(dst_mm, src_mm, dst_pgd, src_pgd, vma, addr, next))) { ret = -ENOMEM; break; } |
1da177e4c
|
1056 |
} while (dst_pgd++, src_pgd++, addr = next, addr != end); |
cddb8a5c1
|
1057 |
|
2ec74c3ef
|
1058 1059 |
if (is_cow) mmu_notifier_invalidate_range_end(src_mm, mmun_start, mmun_end); |
cddb8a5c1
|
1060 |
return ret; |
1da177e4c
|
1061 |
} |
51c6f666f
|
1062 |
static unsigned long zap_pte_range(struct mmu_gather *tlb, |
b5810039a
|
1063 |
struct vm_area_struct *vma, pmd_t *pmd, |
1da177e4c
|
1064 |
unsigned long addr, unsigned long end, |
97a894136
|
1065 |
struct zap_details *details) |
1da177e4c
|
1066 |
{ |
b5810039a
|
1067 |
struct mm_struct *mm = tlb->mm; |
d16dfc550
|
1068 |
int force_flush = 0; |
d559db086
|
1069 |
int rss[NR_MM_COUNTERS]; |
97a894136
|
1070 |
spinlock_t *ptl; |
5f1a19070
|
1071 |
pte_t *start_pte; |
97a894136
|
1072 |
pte_t *pte; |
8a5f14a23
|
1073 |
swp_entry_t entry; |
e9d55e157
|
1074 |
struct page *pending_page = NULL; |
d559db086
|
1075 |
|
d16dfc550
|
1076 |
again: |
e303297e6
|
1077 |
init_rss_vec(rss); |
5f1a19070
|
1078 1079 |
start_pte = pte_offset_map_lock(mm, pmd, addr, &ptl); pte = start_pte; |
6606c3e0d
|
1080 |
arch_enter_lazy_mmu_mode(); |
1da177e4c
|
1081 1082 |
do { pte_t ptent = *pte; |
51c6f666f
|
1083 |
if (pte_none(ptent)) { |
1da177e4c
|
1084 |
continue; |
51c6f666f
|
1085 |
} |
6f5e6b9e6
|
1086 |
|
1da177e4c
|
1087 |
if (pte_present(ptent)) { |
ee498ed73
|
1088 |
struct page *page; |
51c6f666f
|
1089 |
|
6aab341e0
|
1090 |
page = vm_normal_page(vma, addr, ptent); |
1da177e4c
|
1091 1092 1093 1094 1095 1096 1097 |
if (unlikely(details) && page) { /* * unmap_shared_mapping_pages() wants to * invalidate cache without truncating: * unmap shared but keep private pages. */ if (details->check_mapping && |
800d8c63b
|
1098 |
details->check_mapping != page_rmapping(page)) |
1da177e4c
|
1099 |
continue; |
1da177e4c
|
1100 |
} |
b5810039a
|
1101 |
ptent = ptep_get_and_clear_full(mm, addr, pte, |
a600388d2
|
1102 |
tlb->fullmm); |
1da177e4c
|
1103 1104 1105 |
tlb_remove_tlb_entry(tlb, pte, addr); if (unlikely(!page)) continue; |
eca56ff90
|
1106 1107 |
if (!PageAnon(page)) { |
1cf35d477
|
1108 |
if (pte_dirty(ptent)) { |
aac453635
|
1109 1110 1111 1112 1113 1114 |
/* * oom_reaper cannot tear down dirty * pages */ if (unlikely(details && details->ignore_dirty)) continue; |
1cf35d477
|
1115 |
force_flush = 1; |
6237bcd94
|
1116 |
set_page_dirty(page); |
1cf35d477
|
1117 |
} |
4917e5d04
|
1118 |
if (pte_young(ptent) && |
64363aad5
|
1119 |
likely(!(vma->vm_flags & VM_SEQ_READ))) |
bf3f3bc5e
|
1120 |
mark_page_accessed(page); |
6237bcd94
|
1121 |
} |
eca56ff90
|
1122 |
rss[mm_counter(page)]--; |
d281ee614
|
1123 |
page_remove_rmap(page, false); |
3dc147414
|
1124 1125 |
if (unlikely(page_mapcount(page) < 0)) print_bad_pte(vma, addr, ptent, page); |
e9d55e157
|
1126 |
if (unlikely(__tlb_remove_page(tlb, page))) { |
1cf35d477
|
1127 |
force_flush = 1; |
e9d55e157
|
1128 |
pending_page = page; |
ce9ec37bd
|
1129 |
addr += PAGE_SIZE; |
d16dfc550
|
1130 |
break; |
1cf35d477
|
1131 |
} |
1da177e4c
|
1132 1133 |
continue; } |
aac453635
|
1134 1135 |
/* only check swap_entries if explicitly asked for in details */ if (unlikely(details && !details->check_swap_entries)) |
1da177e4c
|
1136 |
continue; |
b084d4353
|
1137 |
|
8a5f14a23
|
1138 1139 1140 1141 1142 |
entry = pte_to_swp_entry(ptent); if (!non_swap_entry(entry)) rss[MM_SWAPENTS]--; else if (is_migration_entry(entry)) { struct page *page; |
9f9f1acd7
|
1143 |
|
8a5f14a23
|
1144 |
page = migration_entry_to_page(entry); |
eca56ff90
|
1145 |
rss[mm_counter(page)]--; |
b084d4353
|
1146 |
} |
8a5f14a23
|
1147 1148 |
if (unlikely(!free_swap_and_cache(entry))) print_bad_pte(vma, addr, ptent, NULL); |
9888a1cae
|
1149 |
pte_clear_not_present_full(mm, addr, pte, tlb->fullmm); |
97a894136
|
1150 |
} while (pte++, addr += PAGE_SIZE, addr != end); |
ae8597623
|
1151 |
|
d559db086
|
1152 |
add_mm_rss_vec(mm, rss); |
6606c3e0d
|
1153 |
arch_leave_lazy_mmu_mode(); |
51c6f666f
|
1154 |
|
1cf35d477
|
1155 |
/* Do the actual TLB flush before dropping ptl */ |
fb7332a9f
|
1156 |
if (force_flush) |
1cf35d477
|
1157 |
tlb_flush_mmu_tlbonly(tlb); |
1cf35d477
|
1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 |
pte_unmap_unlock(start_pte, ptl); /* * If we forced a TLB flush (either due to running out of * batch buffers or because we needed to flush dirty TLB * entries before releasing the ptl), free the batched * memory too. Restart if we didn't do everything. */ if (force_flush) { force_flush = 0; tlb_flush_mmu_free(tlb); |
e9d55e157
|
1169 1170 1171 1172 1173 |
if (pending_page) { /* remove the page with new size */ __tlb_remove_pte_page(tlb, pending_page); pending_page = NULL; } |
2b047252d
|
1174 |
if (addr != end) |
d16dfc550
|
1175 1176 |
goto again; } |
51c6f666f
|
1177 |
return addr; |
1da177e4c
|
1178 |
} |
51c6f666f
|
1179 |
static inline unsigned long zap_pmd_range(struct mmu_gather *tlb, |
b5810039a
|
1180 |
struct vm_area_struct *vma, pud_t *pud, |
1da177e4c
|
1181 |
unsigned long addr, unsigned long end, |
97a894136
|
1182 |
struct zap_details *details) |
1da177e4c
|
1183 1184 1185 1186 1187 1188 1189 |
{ pmd_t *pmd; unsigned long next; pmd = pmd_offset(pud, addr); do { next = pmd_addr_end(addr, end); |
5c7fb56e5
|
1190 |
if (pmd_trans_huge(*pmd) || pmd_devmap(*pmd)) { |
1a5a9906d
|
1191 |
if (next - addr != HPAGE_PMD_SIZE) { |
684283988
|
1192 1193 |
VM_BUG_ON_VMA(vma_is_anonymous(vma) && !rwsem_is_locked(&tlb->mm->mmap_sem), vma); |
78ddc5347
|
1194 |
split_huge_pmd(vma, pmd, addr); |
f21760b15
|
1195 |
} else if (zap_huge_pmd(tlb, vma, pmd, addr)) |
1a5a9906d
|
1196 |
goto next; |
71e3aac07
|
1197 1198 |
/* fall through */ } |
1a5a9906d
|
1199 1200 1201 1202 1203 1204 1205 1206 1207 |
/* * Here there can be other concurrent MADV_DONTNEED or * trans huge page faults running, and if the pmd is * none or trans huge it can change under us. This is * because MADV_DONTNEED holds the mmap_sem in read * mode. */ if (pmd_none_or_trans_huge_or_clear_bad(pmd)) goto next; |
97a894136
|
1208 |
next = zap_pte_range(tlb, vma, pmd, addr, next, details); |
1a5a9906d
|
1209 |
next: |
97a894136
|
1210 1211 |
cond_resched(); } while (pmd++, addr = next, addr != end); |
51c6f666f
|
1212 1213 |
return addr; |
1da177e4c
|
1214 |
} |
51c6f666f
|
1215 |
static inline unsigned long zap_pud_range(struct mmu_gather *tlb, |
b5810039a
|
1216 |
struct vm_area_struct *vma, pgd_t *pgd, |
1da177e4c
|
1217 |
unsigned long addr, unsigned long end, |
97a894136
|
1218 |
struct zap_details *details) |
1da177e4c
|
1219 1220 1221 1222 1223 1224 1225 |
{ pud_t *pud; unsigned long next; pud = pud_offset(pgd, addr); do { next = pud_addr_end(addr, end); |
97a894136
|
1226 |
if (pud_none_or_clear_bad(pud)) |
1da177e4c
|
1227 |
continue; |
97a894136
|
1228 1229 |
next = zap_pmd_range(tlb, vma, pud, addr, next, details); } while (pud++, addr = next, addr != end); |
51c6f666f
|
1230 1231 |
return addr; |
1da177e4c
|
1232 |
} |
aac453635
|
1233 |
void unmap_page_range(struct mmu_gather *tlb, |
038c7aa16
|
1234 1235 1236 |
struct vm_area_struct *vma, unsigned long addr, unsigned long end, struct zap_details *details) |
1da177e4c
|
1237 1238 1239 |
{ pgd_t *pgd; unsigned long next; |
1da177e4c
|
1240 1241 1242 1243 1244 |
BUG_ON(addr >= end); tlb_start_vma(tlb, vma); pgd = pgd_offset(vma->vm_mm, addr); do { next = pgd_addr_end(addr, end); |
97a894136
|
1245 |
if (pgd_none_or_clear_bad(pgd)) |
1da177e4c
|
1246 |
continue; |
97a894136
|
1247 1248 |
next = zap_pud_range(tlb, vma, pgd, addr, next, details); } while (pgd++, addr = next, addr != end); |
1da177e4c
|
1249 1250 |
tlb_end_vma(tlb, vma); } |
51c6f666f
|
1251 |
|
f5cc4eef9
|
1252 1253 1254 |
static void unmap_single_vma(struct mmu_gather *tlb, struct vm_area_struct *vma, unsigned long start_addr, |
4f74d2c8e
|
1255 |
unsigned long end_addr, |
f5cc4eef9
|
1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 |
struct zap_details *details) { unsigned long start = max(vma->vm_start, start_addr); unsigned long end; if (start >= vma->vm_end) return; end = min(vma->vm_end, end_addr); if (end <= vma->vm_start) return; |
cbc91f71b
|
1266 1267 |
if (vma->vm_file) uprobe_munmap(vma, start, end); |
b3b9c2932
|
1268 |
if (unlikely(vma->vm_flags & VM_PFNMAP)) |
5180da410
|
1269 |
untrack_pfn(vma, 0, 0); |
f5cc4eef9
|
1270 1271 1272 1273 1274 1275 1276 |
if (start != end) { if (unlikely(is_vm_hugetlb_page(vma))) { /* * It is undesirable to test vma->vm_file as it * should be non-null for valid hugetlb area. * However, vm_file will be NULL in the error |
7aa6b4ad5
|
1277 |
* cleanup path of mmap_region. When |
f5cc4eef9
|
1278 |
* hugetlbfs ->mmap method fails, |
7aa6b4ad5
|
1279 |
* mmap_region() nullifies vma->vm_file |
f5cc4eef9
|
1280 1281 1282 1283 |
* before calling this function to clean up. * Since no pte has actually been setup, it is * safe to do nothing in this case. */ |
24669e584
|
1284 |
if (vma->vm_file) { |
83cde9e8b
|
1285 |
i_mmap_lock_write(vma->vm_file->f_mapping); |
d833352a4
|
1286 |
__unmap_hugepage_range_final(tlb, vma, start, end, NULL); |
83cde9e8b
|
1287 |
i_mmap_unlock_write(vma->vm_file->f_mapping); |
24669e584
|
1288 |
} |
f5cc4eef9
|
1289 1290 1291 |
} else unmap_page_range(tlb, vma, start, end, details); } |
1da177e4c
|
1292 |
} |
1da177e4c
|
1293 1294 |
/** * unmap_vmas - unmap a range of memory covered by a list of vma's |
0164f69d0
|
1295 |
* @tlb: address of the caller's struct mmu_gather |
1da177e4c
|
1296 1297 1298 |
* @vma: the starting vma * @start_addr: virtual address at which to start unmapping * @end_addr: virtual address at which to end unmapping |
1da177e4c
|
1299 |
* |
508034a32
|
1300 |
* Unmap all pages in the vma list. |
1da177e4c
|
1301 |
* |
1da177e4c
|
1302 1303 1304 1305 1306 1307 1308 1309 1310 |
* Only addresses between `start' and `end' will be unmapped. * * The VMA list must be sorted in ascending virtual address order. * * unmap_vmas() assumes that the caller will flush the whole unmapped address * range after unmap_vmas() returns. So the only responsibility here is to * ensure that any thus-far unmapped pages are flushed before unmap_vmas() * drops the lock and schedules. */ |
6e8bb0193
|
1311 |
void unmap_vmas(struct mmu_gather *tlb, |
1da177e4c
|
1312 |
struct vm_area_struct *vma, unsigned long start_addr, |
4f74d2c8e
|
1313 |
unsigned long end_addr) |
1da177e4c
|
1314 |
{ |
cddb8a5c1
|
1315 |
struct mm_struct *mm = vma->vm_mm; |
1da177e4c
|
1316 |
|
cddb8a5c1
|
1317 |
mmu_notifier_invalidate_range_start(mm, start_addr, end_addr); |
f5cc4eef9
|
1318 |
for ( ; vma && vma->vm_start < end_addr; vma = vma->vm_next) |
4f74d2c8e
|
1319 |
unmap_single_vma(tlb, vma, start_addr, end_addr, NULL); |
cddb8a5c1
|
1320 |
mmu_notifier_invalidate_range_end(mm, start_addr, end_addr); |
1da177e4c
|
1321 1322 1323 1324 1325 |
} /** * zap_page_range - remove user pages in a given range * @vma: vm_area_struct holding the applicable pages |
eb4546bbb
|
1326 |
* @start: starting address of pages to zap |
1da177e4c
|
1327 |
* @size: number of bytes to zap |
8a5f14a23
|
1328 |
* @details: details of shared cache invalidation |
f5cc4eef9
|
1329 1330 |
* * Caller must protect the VMA list |
1da177e4c
|
1331 |
*/ |
7e027b14d
|
1332 |
void zap_page_range(struct vm_area_struct *vma, unsigned long start, |
1da177e4c
|
1333 1334 1335 |
unsigned long size, struct zap_details *details) { struct mm_struct *mm = vma->vm_mm; |
d16dfc550
|
1336 |
struct mmu_gather tlb; |
7e027b14d
|
1337 |
unsigned long end = start + size; |
1da177e4c
|
1338 |
|
1da177e4c
|
1339 |
lru_add_drain(); |
2b047252d
|
1340 |
tlb_gather_mmu(&tlb, mm, start, end); |
365e9c87a
|
1341 |
update_hiwater_rss(mm); |
7e027b14d
|
1342 1343 |
mmu_notifier_invalidate_range_start(mm, start, end); for ( ; vma && vma->vm_start < end; vma = vma->vm_next) |
4f74d2c8e
|
1344 |
unmap_single_vma(&tlb, vma, start, end, details); |
7e027b14d
|
1345 1346 |
mmu_notifier_invalidate_range_end(mm, start, end); tlb_finish_mmu(&tlb, start, end); |
1da177e4c
|
1347 |
} |
c627f9cc0
|
1348 |
/** |
f5cc4eef9
|
1349 1350 1351 1352 |
* zap_page_range_single - remove user pages in a given range * @vma: vm_area_struct holding the applicable pages * @address: starting address of pages to zap * @size: number of bytes to zap |
8a5f14a23
|
1353 |
* @details: details of shared cache invalidation |
f5cc4eef9
|
1354 1355 |
* * The range must fit into one VMA. |
1da177e4c
|
1356 |
*/ |
f5cc4eef9
|
1357 |
static void zap_page_range_single(struct vm_area_struct *vma, unsigned long address, |
1da177e4c
|
1358 1359 1360 |
unsigned long size, struct zap_details *details) { struct mm_struct *mm = vma->vm_mm; |
d16dfc550
|
1361 |
struct mmu_gather tlb; |
1da177e4c
|
1362 |
unsigned long end = address + size; |
1da177e4c
|
1363 |
|
1da177e4c
|
1364 |
lru_add_drain(); |
2b047252d
|
1365 |
tlb_gather_mmu(&tlb, mm, address, end); |
365e9c87a
|
1366 |
update_hiwater_rss(mm); |
f5cc4eef9
|
1367 |
mmu_notifier_invalidate_range_start(mm, address, end); |
4f74d2c8e
|
1368 |
unmap_single_vma(&tlb, vma, address, end, details); |
f5cc4eef9
|
1369 |
mmu_notifier_invalidate_range_end(mm, address, end); |
d16dfc550
|
1370 |
tlb_finish_mmu(&tlb, address, end); |
1da177e4c
|
1371 |
} |
c627f9cc0
|
1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 |
/** * zap_vma_ptes - remove ptes mapping the vma * @vma: vm_area_struct holding ptes to be zapped * @address: starting address of pages to zap * @size: number of bytes to zap * * This function only unmaps ptes assigned to VM_PFNMAP vmas. * * The entire address range must be fully contained within the vma. * * Returns 0 if successful. */ int zap_vma_ptes(struct vm_area_struct *vma, unsigned long address, unsigned long size) { if (address < vma->vm_start || address + size > vma->vm_end || !(vma->vm_flags & VM_PFNMAP)) return -1; |
f5cc4eef9
|
1390 |
zap_page_range_single(vma, address, size, NULL); |
c627f9cc0
|
1391 1392 1393 |
return 0; } EXPORT_SYMBOL_GPL(zap_vma_ptes); |
25ca1d6c0
|
1394 |
pte_t *__get_locked_pte(struct mm_struct *mm, unsigned long addr, |
920c7a5d0
|
1395 |
spinlock_t **ptl) |
c9cfcddfd
|
1396 1397 1398 1399 |
{ pgd_t * pgd = pgd_offset(mm, addr); pud_t * pud = pud_alloc(mm, pgd, addr); if (pud) { |
49c91fb01
|
1400 |
pmd_t * pmd = pmd_alloc(mm, pud, addr); |
f66055ab6
|
1401 1402 |
if (pmd) { VM_BUG_ON(pmd_trans_huge(*pmd)); |
c9cfcddfd
|
1403 |
return pte_alloc_map_lock(mm, pmd, addr, ptl); |
f66055ab6
|
1404 |
} |
c9cfcddfd
|
1405 1406 1407 |
} return NULL; } |
1da177e4c
|
1408 |
/* |
238f58d89
|
1409 1410 1411 1412 1413 1414 |
* This is the old fallback for page remapping. * * For historical reasons, it only allows reserved pages. Only * old drivers should use this, and they needed to mark their * pages reserved for the old functions anyway. */ |
423bad600
|
1415 1416 |
static int insert_page(struct vm_area_struct *vma, unsigned long addr, struct page *page, pgprot_t prot) |
238f58d89
|
1417 |
{ |
423bad600
|
1418 |
struct mm_struct *mm = vma->vm_mm; |
238f58d89
|
1419 |
int retval; |
c9cfcddfd
|
1420 |
pte_t *pte; |
8a9f3ccd2
|
1421 |
spinlock_t *ptl; |
238f58d89
|
1422 |
retval = -EINVAL; |
a145dd411
|
1423 |
if (PageAnon(page)) |
5b4e655e9
|
1424 |
goto out; |
238f58d89
|
1425 1426 |
retval = -ENOMEM; flush_dcache_page(page); |
c9cfcddfd
|
1427 |
pte = get_locked_pte(mm, addr, &ptl); |
238f58d89
|
1428 |
if (!pte) |
5b4e655e9
|
1429 |
goto out; |
238f58d89
|
1430 1431 1432 1433 1434 1435 |
retval = -EBUSY; if (!pte_none(*pte)) goto out_unlock; /* Ok, finally just insert the thing.. */ get_page(page); |
eca56ff90
|
1436 |
inc_mm_counter_fast(mm, mm_counter_file(page)); |
dd78fedde
|
1437 |
page_add_file_rmap(page, false); |
238f58d89
|
1438 1439 1440 |
set_pte_at(mm, addr, pte, mk_pte(page, prot)); retval = 0; |
8a9f3ccd2
|
1441 1442 |
pte_unmap_unlock(pte, ptl); return retval; |
238f58d89
|
1443 1444 1445 1446 1447 |
out_unlock: pte_unmap_unlock(pte, ptl); out: return retval; } |
bfa5bf6d6
|
1448 1449 1450 1451 1452 1453 |
/** * vm_insert_page - insert single page into user vma * @vma: user vma to map to * @addr: target user address of this page * @page: source kernel page * |
a145dd411
|
1454 1455 1456 1457 1458 1459 |
* This allows drivers to insert individual pages they've allocated * into a user vma. * * The page has to be a nice clean _individual_ kernel allocation. * If you allocate a compound page, you need to have marked it as * such (__GFP_COMP), or manually just split the page up yourself |
8dfcc9ba2
|
1460 |
* (see split_page()). |
a145dd411
|
1461 1462 1463 1464 1465 1466 1467 1468 |
* * NOTE! Traditionally this was done with "remap_pfn_range()" which * took an arbitrary page protection parameter. This doesn't allow * that. Your vma protection will have to be set up correctly, which * means that if you want a shared writable mapping, you'd better * ask for a shared writable mapping! * * The page does not need to be reserved. |
4b6e1e370
|
1469 1470 1471 1472 1473 |
* * Usually this function is called from f_op->mmap() handler * under mm->mmap_sem write-lock, so it can change vma->vm_flags. * Caller must set VM_MIXEDMAP on vma if it wants to call this * function from other places, for example from page-fault handler. |
a145dd411
|
1474 |
*/ |
423bad600
|
1475 1476 |
int vm_insert_page(struct vm_area_struct *vma, unsigned long addr, struct page *page) |
a145dd411
|
1477 1478 1479 1480 1481 |
{ if (addr < vma->vm_start || addr >= vma->vm_end) return -EFAULT; if (!page_count(page)) return -EINVAL; |
4b6e1e370
|
1482 1483 1484 1485 1486 |
if (!(vma->vm_flags & VM_MIXEDMAP)) { BUG_ON(down_read_trylock(&vma->vm_mm->mmap_sem)); BUG_ON(vma->vm_flags & VM_PFNMAP); vma->vm_flags |= VM_MIXEDMAP; } |
423bad600
|
1487 |
return insert_page(vma, addr, page, vma->vm_page_prot); |
a145dd411
|
1488 |
} |
e3c3374fb
|
1489 |
EXPORT_SYMBOL(vm_insert_page); |
a145dd411
|
1490 |
|
423bad600
|
1491 |
static int insert_pfn(struct vm_area_struct *vma, unsigned long addr, |
01c8f1c44
|
1492 |
pfn_t pfn, pgprot_t prot) |
423bad600
|
1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 |
{ struct mm_struct *mm = vma->vm_mm; int retval; pte_t *pte, entry; spinlock_t *ptl; retval = -ENOMEM; pte = get_locked_pte(mm, addr, &ptl); if (!pte) goto out; retval = -EBUSY; if (!pte_none(*pte)) goto out_unlock; /* Ok, finally just insert the thing.. */ |
01c8f1c44
|
1508 1509 1510 1511 |
if (pfn_t_devmap(pfn)) entry = pte_mkdevmap(pfn_t_pte(pfn, prot)); else entry = pte_mkspecial(pfn_t_pte(pfn, prot)); |
423bad600
|
1512 |
set_pte_at(mm, addr, pte, entry); |
4b3073e1c
|
1513 |
update_mmu_cache(vma, addr, pte); /* XXX: why not for insert_page? */ |
423bad600
|
1514 1515 1516 1517 1518 1519 1520 |
retval = 0; out_unlock: pte_unmap_unlock(pte, ptl); out: return retval; } |
e0dc0d8f4
|
1521 1522 1523 1524 1525 1526 |
/** * vm_insert_pfn - insert single pfn into user vma * @vma: user vma to map to * @addr: target user address of this page * @pfn: source kernel pfn * |
c462f179e
|
1527 |
* Similar to vm_insert_page, this allows drivers to insert individual pages |
e0dc0d8f4
|
1528 1529 1530 1531 |
* they've allocated into a user vma. Same comments apply. * * This function should only be called from a vm_ops->fault handler, and * in that case the handler should return NULL. |
0d71d10a4
|
1532 1533 1534 1535 1536 |
* * vma cannot be a COW mapping. * * As this is called only for pages that do not currently exist, we * do not need to flush old virtual caches or the TLB. |
e0dc0d8f4
|
1537 1538 |
*/ int vm_insert_pfn(struct vm_area_struct *vma, unsigned long addr, |
423bad600
|
1539 |
unsigned long pfn) |
e0dc0d8f4
|
1540 |
{ |
1745cbc5d
|
1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 |
return vm_insert_pfn_prot(vma, addr, pfn, vma->vm_page_prot); } EXPORT_SYMBOL(vm_insert_pfn); /** * vm_insert_pfn_prot - insert single pfn into user vma with specified pgprot * @vma: user vma to map to * @addr: target user address of this page * @pfn: source kernel pfn * @pgprot: pgprot flags for the inserted page * * This is exactly like vm_insert_pfn, except that it allows drivers to * to override pgprot on a per-page basis. * * This only makes sense for IO mappings, and it makes no sense for * cow mappings. In general, using multiple vmas is preferable; * vm_insert_pfn_prot should only be used if using multiple VMAs is * impractical. */ int vm_insert_pfn_prot(struct vm_area_struct *vma, unsigned long addr, unsigned long pfn, pgprot_t pgprot) { |
2ab640379
|
1563 |
int ret; |
7e675137a
|
1564 1565 1566 1567 1568 1569 |
/* * Technically, architectures with pte_special can avoid all these * restrictions (same for remap_pfn_range). However we would like * consistency in testing and feature parity among all, so we should * try to keep these invariants in place for everybody. */ |
b379d7901
|
1570 1571 1572 1573 1574 |
BUG_ON(!(vma->vm_flags & (VM_PFNMAP|VM_MIXEDMAP))); BUG_ON((vma->vm_flags & (VM_PFNMAP|VM_MIXEDMAP)) == (VM_PFNMAP|VM_MIXEDMAP)); BUG_ON((vma->vm_flags & VM_PFNMAP) && is_cow_mapping(vma->vm_flags)); BUG_ON((vma->vm_flags & VM_MIXEDMAP) && pfn_valid(pfn)); |
e0dc0d8f4
|
1575 |
|
423bad600
|
1576 1577 |
if (addr < vma->vm_start || addr >= vma->vm_end) return -EFAULT; |
f25748e3c
|
1578 |
if (track_pfn_insert(vma, &pgprot, __pfn_to_pfn_t(pfn, PFN_DEV))) |
2ab640379
|
1579 |
return -EINVAL; |
01c8f1c44
|
1580 |
ret = insert_pfn(vma, addr, __pfn_to_pfn_t(pfn, PFN_DEV), pgprot); |
2ab640379
|
1581 |
|
2ab640379
|
1582 |
return ret; |
423bad600
|
1583 |
} |
1745cbc5d
|
1584 |
EXPORT_SYMBOL(vm_insert_pfn_prot); |
e0dc0d8f4
|
1585 |
|
423bad600
|
1586 |
int vm_insert_mixed(struct vm_area_struct *vma, unsigned long addr, |
01c8f1c44
|
1587 |
pfn_t pfn) |
423bad600
|
1588 |
{ |
87744ab38
|
1589 |
pgprot_t pgprot = vma->vm_page_prot; |
423bad600
|
1590 |
BUG_ON(!(vma->vm_flags & VM_MIXEDMAP)); |
e0dc0d8f4
|
1591 |
|
423bad600
|
1592 1593 |
if (addr < vma->vm_start || addr >= vma->vm_end) return -EFAULT; |
87744ab38
|
1594 1595 |
if (track_pfn_insert(vma, &pgprot, pfn)) return -EINVAL; |
e0dc0d8f4
|
1596 |
|
423bad600
|
1597 1598 1599 1600 |
/* * If we don't have pte special, then we have to use the pfn_valid() * based VM_MIXEDMAP scheme (see vm_normal_page), and thus we *must* * refcount the page if pfn_valid is true (hence insert_page rather |
62eede62d
|
1601 1602 |
* than insert_pfn). If a zero_pfn were inserted into a VM_MIXEDMAP * without pte special, it would there be refcounted as a normal page. |
423bad600
|
1603 |
*/ |
03fc2da63
|
1604 |
if (!HAVE_PTE_SPECIAL && !pfn_t_devmap(pfn) && pfn_t_valid(pfn)) { |
423bad600
|
1605 |
struct page *page; |
03fc2da63
|
1606 1607 1608 1609 1610 1611 |
/* * At this point we are committed to insert_page() * regardless of whether the caller specified flags that * result in pfn_t_has_page() == false. */ page = pfn_to_page(pfn_t_to_pfn(pfn)); |
87744ab38
|
1612 |
return insert_page(vma, addr, page, pgprot); |
423bad600
|
1613 |
} |
87744ab38
|
1614 |
return insert_pfn(vma, addr, pfn, pgprot); |
e0dc0d8f4
|
1615 |
} |
423bad600
|
1616 |
EXPORT_SYMBOL(vm_insert_mixed); |
e0dc0d8f4
|
1617 |
|
a145dd411
|
1618 |
/* |
1da177e4c
|
1619 1620 1621 1622 1623 1624 1625 1626 1627 |
* maps a range of physical memory into the requested pages. the old * mappings are removed. any references to nonexistent pages results * in null mappings (currently treated as "copy-on-access") */ static int remap_pte_range(struct mm_struct *mm, pmd_t *pmd, unsigned long addr, unsigned long end, unsigned long pfn, pgprot_t prot) { pte_t *pte; |
c74df32c7
|
1628 |
spinlock_t *ptl; |
1da177e4c
|
1629 |
|
c74df32c7
|
1630 |
pte = pte_alloc_map_lock(mm, pmd, addr, &ptl); |
1da177e4c
|
1631 1632 |
if (!pte) return -ENOMEM; |
6606c3e0d
|
1633 |
arch_enter_lazy_mmu_mode(); |
1da177e4c
|
1634 1635 |
do { BUG_ON(!pte_none(*pte)); |
7e675137a
|
1636 |
set_pte_at(mm, addr, pte, pte_mkspecial(pfn_pte(pfn, prot))); |
1da177e4c
|
1637 1638 |
pfn++; } while (pte++, addr += PAGE_SIZE, addr != end); |
6606c3e0d
|
1639 |
arch_leave_lazy_mmu_mode(); |
c74df32c7
|
1640 |
pte_unmap_unlock(pte - 1, ptl); |
1da177e4c
|
1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 |
return 0; } static inline int remap_pmd_range(struct mm_struct *mm, pud_t *pud, unsigned long addr, unsigned long end, unsigned long pfn, pgprot_t prot) { pmd_t *pmd; unsigned long next; pfn -= addr >> PAGE_SHIFT; pmd = pmd_alloc(mm, pud, addr); if (!pmd) return -ENOMEM; |
f66055ab6
|
1655 |
VM_BUG_ON(pmd_trans_huge(*pmd)); |
1da177e4c
|
1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 |
do { next = pmd_addr_end(addr, end); if (remap_pte_range(mm, pmd, addr, next, pfn + (addr >> PAGE_SHIFT), prot)) return -ENOMEM; } while (pmd++, addr = next, addr != end); return 0; } static inline int remap_pud_range(struct mm_struct *mm, pgd_t *pgd, unsigned long addr, unsigned long end, unsigned long pfn, pgprot_t prot) { pud_t *pud; unsigned long next; pfn -= addr >> PAGE_SHIFT; pud = pud_alloc(mm, pgd, addr); if (!pud) return -ENOMEM; do { next = pud_addr_end(addr, end); if (remap_pmd_range(mm, pud, addr, next, pfn + (addr >> PAGE_SHIFT), prot)) return -ENOMEM; } while (pud++, addr = next, addr != end); return 0; } |
bfa5bf6d6
|
1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 |
/** * remap_pfn_range - remap kernel memory to userspace * @vma: user vma to map to * @addr: target user address to start at * @pfn: physical address of kernel memory * @size: size of map area * @prot: page protection flags for this mapping * * Note: this is only safe if the mm semaphore is held when called. */ |
1da177e4c
|
1694 1695 1696 1697 1698 |
int remap_pfn_range(struct vm_area_struct *vma, unsigned long addr, unsigned long pfn, unsigned long size, pgprot_t prot) { pgd_t *pgd; unsigned long next; |
2d15cab85
|
1699 |
unsigned long end = addr + PAGE_ALIGN(size); |
1da177e4c
|
1700 |
struct mm_struct *mm = vma->vm_mm; |
d5957d2fc
|
1701 |
unsigned long remap_pfn = pfn; |
1da177e4c
|
1702 1703 1704 1705 1706 1707 1708 |
int err; /* * Physically remapped pages are special. Tell the * rest of the world about it: * VM_IO tells people not to look at these pages * (accesses can have side effects). |
6aab341e0
|
1709 1710 1711 |
* VM_PFNMAP tells the core MM that the base pages are just * raw PFN mappings, and do not have a "struct page" associated * with them. |
314e51b98
|
1712 1713 1714 1715 |
* VM_DONTEXPAND * Disable vma merging and expanding with mremap(). * VM_DONTDUMP * Omit vma from core dump, even when VM_IO turned off. |
fb155c161
|
1716 1717 1718 1719 |
* * There's a horrible special case to handle copy-on-write * behaviour that some programs depend on. We mark the "original" * un-COW'ed pages by matching them up with "vma->vm_pgoff". |
b3b9c2932
|
1720 |
* See vm_normal_page() for details. |
1da177e4c
|
1721 |
*/ |
b3b9c2932
|
1722 1723 1724 |
if (is_cow_mapping(vma->vm_flags)) { if (addr != vma->vm_start || end != vma->vm_end) return -EINVAL; |
fb155c161
|
1725 |
vma->vm_pgoff = pfn; |
b3b9c2932
|
1726 |
} |
d5957d2fc
|
1727 |
err = track_pfn_remap(vma, &prot, remap_pfn, addr, PAGE_ALIGN(size)); |
b3b9c2932
|
1728 |
if (err) |
3c8bb73ac
|
1729 |
return -EINVAL; |
fb155c161
|
1730 |
|
314e51b98
|
1731 |
vma->vm_flags |= VM_IO | VM_PFNMAP | VM_DONTEXPAND | VM_DONTDUMP; |
1da177e4c
|
1732 1733 1734 1735 1736 |
BUG_ON(addr >= end); pfn -= addr >> PAGE_SHIFT; pgd = pgd_offset(mm, addr); flush_cache_range(vma, addr, end); |
1da177e4c
|
1737 1738 1739 1740 1741 1742 1743 |
do { next = pgd_addr_end(addr, end); err = remap_pud_range(mm, pgd, addr, next, pfn + (addr >> PAGE_SHIFT), prot); if (err) break; } while (pgd++, addr = next, addr != end); |
2ab640379
|
1744 1745 |
if (err) |
d5957d2fc
|
1746 |
untrack_pfn(vma, remap_pfn, PAGE_ALIGN(size)); |
2ab640379
|
1747 |
|
1da177e4c
|
1748 1749 1750 |
return err; } EXPORT_SYMBOL(remap_pfn_range); |
b4cbb197c
|
1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 |
/** * vm_iomap_memory - remap memory to userspace * @vma: user vma to map to * @start: start of area * @len: size of area * * This is a simplified io_remap_pfn_range() for common driver use. The * driver just needs to give us the physical memory range to be mapped, * we'll figure out the rest from the vma information. * * NOTE! Some drivers might want to tweak vma->vm_page_prot first to get * whatever write-combining details or similar. */ int vm_iomap_memory(struct vm_area_struct *vma, phys_addr_t start, unsigned long len) { unsigned long vm_len, pfn, pages; /* Check that the physical memory area passed in looks valid */ if (start + len < start) return -EINVAL; /* * You *really* shouldn't map things that aren't page-aligned, * but we've historically allowed it because IO memory might * just have smaller alignment. */ len += start & ~PAGE_MASK; pfn = start >> PAGE_SHIFT; pages = (len + ~PAGE_MASK) >> PAGE_SHIFT; if (pfn + pages < pfn) return -EINVAL; /* We start the mapping 'vm_pgoff' pages into the area */ if (vma->vm_pgoff > pages) return -EINVAL; pfn += vma->vm_pgoff; pages -= vma->vm_pgoff; /* Can we fit all of the mapping? */ vm_len = vma->vm_end - vma->vm_start; if (vm_len >> PAGE_SHIFT > pages) return -EINVAL; /* Ok, let it rip */ return io_remap_pfn_range(vma, vma->vm_start, pfn, vm_len, vma->vm_page_prot); } EXPORT_SYMBOL(vm_iomap_memory); |
aee16b3ce
|
1797 1798 1799 1800 1801 1802 |
static int apply_to_pte_range(struct mm_struct *mm, pmd_t *pmd, unsigned long addr, unsigned long end, pte_fn_t fn, void *data) { pte_t *pte; int err; |
2f569afd9
|
1803 |
pgtable_t token; |
949099148
|
1804 |
spinlock_t *uninitialized_var(ptl); |
aee16b3ce
|
1805 1806 1807 1808 1809 1810 1811 1812 |
pte = (mm == &init_mm) ? pte_alloc_kernel(pmd, addr) : pte_alloc_map_lock(mm, pmd, addr, &ptl); if (!pte) return -ENOMEM; BUG_ON(pmd_huge(*pmd)); |
38e0edb15
|
1813 |
arch_enter_lazy_mmu_mode(); |
2f569afd9
|
1814 |
token = pmd_pgtable(*pmd); |
aee16b3ce
|
1815 1816 |
do { |
c36987e2e
|
1817 |
err = fn(pte++, token, addr, data); |
aee16b3ce
|
1818 1819 |
if (err) break; |
c36987e2e
|
1820 |
} while (addr += PAGE_SIZE, addr != end); |
aee16b3ce
|
1821 |
|
38e0edb15
|
1822 |
arch_leave_lazy_mmu_mode(); |
aee16b3ce
|
1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 |
if (mm != &init_mm) pte_unmap_unlock(pte-1, ptl); return err; } static int apply_to_pmd_range(struct mm_struct *mm, pud_t *pud, unsigned long addr, unsigned long end, pte_fn_t fn, void *data) { pmd_t *pmd; unsigned long next; int err; |
ceb868796
|
1835 |
BUG_ON(pud_huge(*pud)); |
aee16b3ce
|
1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 |
pmd = pmd_alloc(mm, pud, addr); if (!pmd) return -ENOMEM; do { next = pmd_addr_end(addr, end); err = apply_to_pte_range(mm, pmd, addr, next, fn, data); if (err) break; } while (pmd++, addr = next, addr != end); return err; } static int apply_to_pud_range(struct mm_struct *mm, pgd_t *pgd, unsigned long addr, unsigned long end, pte_fn_t fn, void *data) { pud_t *pud; unsigned long next; int err; pud = pud_alloc(mm, pgd, addr); if (!pud) return -ENOMEM; do { next = pud_addr_end(addr, end); err = apply_to_pmd_range(mm, pud, addr, next, fn, data); if (err) break; } while (pud++, addr = next, addr != end); return err; } /* * Scan a region of virtual memory, filling in page tables as necessary * and calling a provided function on each leaf page table. */ int apply_to_page_range(struct mm_struct *mm, unsigned long addr, unsigned long size, pte_fn_t fn, void *data) { pgd_t *pgd; unsigned long next; |
57250a5bf
|
1877 |
unsigned long end = addr + size; |
aee16b3ce
|
1878 |
int err; |
9cb65bc3b
|
1879 1880 |
if (WARN_ON(addr >= end)) return -EINVAL; |
aee16b3ce
|
1881 1882 1883 1884 1885 1886 1887 |
pgd = pgd_offset(mm, addr); do { next = pgd_addr_end(addr, end); err = apply_to_pud_range(mm, pgd, addr, next, fn, data); if (err) break; } while (pgd++, addr = next, addr != end); |
57250a5bf
|
1888 |
|
aee16b3ce
|
1889 1890 1891 |
return err; } EXPORT_SYMBOL_GPL(apply_to_page_range); |
1da177e4c
|
1892 |
/* |
9b4bdd2ff
|
1893 1894 1895 1896 1897 |
* handle_pte_fault chooses page fault handler according to an entry which was * read non-atomically. Before making any commitment, on those architectures * or configurations (e.g. i386 with PAE) which might give a mix of unmatched * parts, do_swap_page must check under lock before unmapping the pte and * proceeding (but do_wp_page is only called after already making such a check; |
a335b2e17
|
1898 |
* and do_anonymous_page can safely check later on). |
8f4e2101f
|
1899 |
*/ |
4c21e2f24
|
1900 |
static inline int pte_unmap_same(struct mm_struct *mm, pmd_t *pmd, |
8f4e2101f
|
1901 1902 1903 1904 1905 |
pte_t *page_table, pte_t orig_pte) { int same = 1; #if defined(CONFIG_SMP) || defined(CONFIG_PREEMPT) if (sizeof(pte_t) > sizeof(unsigned long)) { |
4c21e2f24
|
1906 1907 |
spinlock_t *ptl = pte_lockptr(mm, pmd); spin_lock(ptl); |
8f4e2101f
|
1908 |
same = pte_same(*page_table, orig_pte); |
4c21e2f24
|
1909 |
spin_unlock(ptl); |
8f4e2101f
|
1910 1911 1912 1913 1914 |
} #endif pte_unmap(page_table); return same; } |
9de455b20
|
1915 |
static inline void cow_user_page(struct page *dst, struct page *src, unsigned long va, struct vm_area_struct *vma) |
6aab341e0
|
1916 |
{ |
0abdd7a81
|
1917 |
debug_dma_assert_idle(src); |
6aab341e0
|
1918 1919 1920 1921 1922 1923 1924 |
/* * If the source page was a PFN mapping, we don't have * a "struct page" for it. We do a best-effort copy by * just copying from the original user address. If that * fails, we just zero-fill it. Live with it. */ if (unlikely(!src)) { |
9b04c5fec
|
1925 |
void *kaddr = kmap_atomic(dst); |
5d2a2dbbc
|
1926 1927 1928 1929 1930 1931 1932 1933 1934 |
void __user *uaddr = (void __user *)(va & PAGE_MASK); /* * This really shouldn't fail, because the page is there * in the page tables. But it might just be unreadable, * in which case we just give up and fill the result with * zeroes. */ if (__copy_from_user_inatomic(kaddr, uaddr, PAGE_SIZE)) |
3ecb01df3
|
1935 |
clear_page(kaddr); |
9b04c5fec
|
1936 |
kunmap_atomic(kaddr); |
c4ec7b0de
|
1937 |
flush_dcache_page(dst); |
0ed361dec
|
1938 1939 |
} else copy_user_highpage(dst, src, va, vma); |
6aab341e0
|
1940 |
} |
c20cd45eb
|
1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 |
static gfp_t __get_fault_gfp_mask(struct vm_area_struct *vma) { struct file *vm_file = vma->vm_file; if (vm_file) return mapping_gfp_mask(vm_file->f_mapping) | __GFP_FS | __GFP_IO; /* * Special mappings (e.g. VDSO) do not have any file so fake * a default GFP_KERNEL for them. */ return GFP_KERNEL; } |
1da177e4c
|
1954 |
/* |
fb09a4642
|
1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 |
* Notify the address space that the page is about to become writable so that * it can prohibit this or wait for the page to get into an appropriate state. * * We do this without the lock held, so that it can sleep if it needs to. */ static int do_page_mkwrite(struct vm_area_struct *vma, struct page *page, unsigned long address) { struct vm_fault vmf; int ret; vmf.virtual_address = (void __user *)(address & PAGE_MASK); vmf.pgoff = page->index; vmf.flags = FAULT_FLAG_WRITE|FAULT_FLAG_MKWRITE; |
c20cd45eb
|
1969 |
vmf.gfp_mask = __get_fault_gfp_mask(vma); |
fb09a4642
|
1970 |
vmf.page = page; |
2e4cdab05
|
1971 |
vmf.cow_page = NULL; |
fb09a4642
|
1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 |
ret = vma->vm_ops->page_mkwrite(vma, &vmf); if (unlikely(ret & (VM_FAULT_ERROR | VM_FAULT_NOPAGE))) return ret; if (unlikely(!(ret & VM_FAULT_LOCKED))) { lock_page(page); if (!page->mapping) { unlock_page(page); return 0; /* retry */ } ret |= VM_FAULT_LOCKED; } else VM_BUG_ON_PAGE(!PageLocked(page), page); return ret; } /* |
4e047f897
|
1989 1990 1991 1992 1993 1994 1995 |
* Handle write page faults for pages that can be reused in the current vma * * This can happen either due to the mapping being with the VM_SHARED flag, * or due to us being the last reference standing to the page. In either * case, all we need to do here is to mark the page as writable and update * any related book-keeping. */ |
bae473a42
|
1996 1997 1998 |
static inline int wp_page_reuse(struct fault_env *fe, pte_t orig_pte, struct page *page, int page_mkwrite, int dirty_shared) __releases(fe->ptl) |
4e047f897
|
1999 |
{ |
bae473a42
|
2000 |
struct vm_area_struct *vma = fe->vma; |
4e047f897
|
2001 2002 2003 2004 2005 2006 2007 2008 |
pte_t entry; /* * Clear the pages cpupid information as the existing * information potentially belongs to a now completely * unrelated process. */ if (page) page_cpupid_xchg_last(page, (1 << LAST_CPUPID_SHIFT) - 1); |
bae473a42
|
2009 |
flush_cache_page(vma, fe->address, pte_pfn(orig_pte)); |
4e047f897
|
2010 2011 |
entry = pte_mkyoung(orig_pte); entry = maybe_mkwrite(pte_mkdirty(entry), vma); |
bae473a42
|
2012 2013 2014 |
if (ptep_set_access_flags(vma, fe->address, fe->pte, entry, 1)) update_mmu_cache(vma, fe->address, fe->pte); pte_unmap_unlock(fe->pte, fe->ptl); |
4e047f897
|
2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 |
if (dirty_shared) { struct address_space *mapping; int dirtied; if (!page_mkwrite) lock_page(page); dirtied = set_page_dirty(page); VM_BUG_ON_PAGE(PageAnon(page), page); mapping = page->mapping; unlock_page(page); |
09cbfeaf1
|
2027 |
put_page(page); |
4e047f897
|
2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 |
if ((dirtied || page_mkwrite) && mapping) { /* * Some device drivers do not set page.mapping * but still dirty their pages */ balance_dirty_pages_ratelimited(mapping); } if (!page_mkwrite) file_update_time(vma->vm_file); } return VM_FAULT_WRITE; } /* |
2f38ab2c3
|
2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 |
* Handle the case of a page which we actually need to copy to a new page. * * Called with mmap_sem locked and the old page referenced, but * without the ptl held. * * High level logic flow: * * - Allocate a page, copy the content of the old page to the new one. * - Handle book keeping and accounting - cgroups, mmu-notifiers, etc. * - Take the PTL. If the pte changed, bail out and release the allocated page * - If the pte is still the way we remember it, update the page table and all * relevant references. This includes dropping the reference the page-table * held to the old page, as well as updating the rmap. * - In any case, unlock the PTL and drop the reference we took to the old page. */ |
bae473a42
|
2060 2061 |
static int wp_page_copy(struct fault_env *fe, pte_t orig_pte, struct page *old_page) |
2f38ab2c3
|
2062 |
{ |
bae473a42
|
2063 2064 |
struct vm_area_struct *vma = fe->vma; struct mm_struct *mm = vma->vm_mm; |
2f38ab2c3
|
2065 |
struct page *new_page = NULL; |
2f38ab2c3
|
2066 2067 |
pte_t entry; int page_copied = 0; |
bae473a42
|
2068 2069 |
const unsigned long mmun_start = fe->address & PAGE_MASK; const unsigned long mmun_end = mmun_start + PAGE_SIZE; |
2f38ab2c3
|
2070 2071 2072 2073 2074 2075 |
struct mem_cgroup *memcg; if (unlikely(anon_vma_prepare(vma))) goto oom; if (is_zero_pfn(pte_pfn(orig_pte))) { |
bae473a42
|
2076 |
new_page = alloc_zeroed_user_highpage_movable(vma, fe->address); |
2f38ab2c3
|
2077 2078 2079 |
if (!new_page) goto oom; } else { |
bae473a42
|
2080 2081 |
new_page = alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma, fe->address); |
2f38ab2c3
|
2082 2083 |
if (!new_page) goto oom; |
bae473a42
|
2084 |
cow_user_page(new_page, old_page, fe->address, vma); |
2f38ab2c3
|
2085 |
} |
2f38ab2c3
|
2086 |
|
f627c2f53
|
2087 |
if (mem_cgroup_try_charge(new_page, mm, GFP_KERNEL, &memcg, false)) |
2f38ab2c3
|
2088 |
goto oom_free_new; |
eb3c24f30
|
2089 |
__SetPageUptodate(new_page); |
2f38ab2c3
|
2090 2091 2092 2093 2094 |
mmu_notifier_invalidate_range_start(mm, mmun_start, mmun_end); /* * Re-check the pte - we dropped the lock */ |
bae473a42
|
2095 2096 |
fe->pte = pte_offset_map_lock(mm, fe->pmd, fe->address, &fe->ptl); if (likely(pte_same(*fe->pte, orig_pte))) { |
2f38ab2c3
|
2097 2098 |
if (old_page) { if (!PageAnon(old_page)) { |
eca56ff90
|
2099 2100 |
dec_mm_counter_fast(mm, mm_counter_file(old_page)); |
2f38ab2c3
|
2101 2102 2103 2104 2105 |
inc_mm_counter_fast(mm, MM_ANONPAGES); } } else { inc_mm_counter_fast(mm, MM_ANONPAGES); } |
bae473a42
|
2106 |
flush_cache_page(vma, fe->address, pte_pfn(orig_pte)); |
2f38ab2c3
|
2107 2108 2109 2110 2111 2112 2113 2114 |
entry = mk_pte(new_page, vma->vm_page_prot); entry = maybe_mkwrite(pte_mkdirty(entry), vma); /* * Clear the pte entry and flush it first, before updating the * pte with the new entry. This will avoid a race condition * seen in the presence of one thread doing SMC and another * thread doing COW. */ |
bae473a42
|
2115 2116 |
ptep_clear_flush_notify(vma, fe->address, fe->pte); page_add_new_anon_rmap(new_page, vma, fe->address, false); |
f627c2f53
|
2117 |
mem_cgroup_commit_charge(new_page, memcg, false, false); |
2f38ab2c3
|
2118 2119 2120 2121 2122 2123 |
lru_cache_add_active_or_unevictable(new_page, vma); /* * We call the notify macro here because, when using secondary * mmu page tables (such as kvm shadow page tables), we want the * new page to be mapped directly into the secondary page table. */ |
bae473a42
|
2124 2125 |
set_pte_at_notify(mm, fe->address, fe->pte, entry); update_mmu_cache(vma, fe->address, fe->pte); |
2f38ab2c3
|
2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 |
if (old_page) { /* * Only after switching the pte to the new page may * we remove the mapcount here. Otherwise another * process may come and find the rmap count decremented * before the pte is switched to the new page, and * "reuse" the old page writing into it while our pte * here still points into it and can be read by other * threads. * * The critical issue is to order this * page_remove_rmap with the ptp_clear_flush above. * Those stores are ordered by (if nothing else,) * the barrier present in the atomic_add_negative * in page_remove_rmap. * * Then the TLB flush in ptep_clear_flush ensures that * no process can access the old page before the * decremented mapcount is visible. And the old page * cannot be reused until after the decremented * mapcount is visible. So transitively, TLBs to * old page will be flushed before it can be reused. */ |
d281ee614
|
2149 |
page_remove_rmap(old_page, false); |
2f38ab2c3
|
2150 2151 2152 2153 2154 2155 |
} /* Free the old page.. */ new_page = old_page; page_copied = 1; } else { |
f627c2f53
|
2156 |
mem_cgroup_cancel_charge(new_page, memcg, false); |
2f38ab2c3
|
2157 2158 2159 |
} if (new_page) |
09cbfeaf1
|
2160 |
put_page(new_page); |
2f38ab2c3
|
2161 |
|
bae473a42
|
2162 |
pte_unmap_unlock(fe->pte, fe->ptl); |
2f38ab2c3
|
2163 2164 2165 2166 2167 2168 2169 2170 |
mmu_notifier_invalidate_range_end(mm, mmun_start, mmun_end); if (old_page) { /* * Don't let another task, with possibly unlocked vma, * keep the mlocked page. */ if (page_copied && (vma->vm_flags & VM_LOCKED)) { lock_page(old_page); /* LRU manipulation */ |
e90309c9f
|
2171 2172 |
if (PageMlocked(old_page)) munlock_vma_page(old_page); |
2f38ab2c3
|
2173 2174 |
unlock_page(old_page); } |
09cbfeaf1
|
2175 |
put_page(old_page); |
2f38ab2c3
|
2176 2177 2178 |
} return page_copied ? VM_FAULT_WRITE : 0; oom_free_new: |
09cbfeaf1
|
2179 |
put_page(new_page); |
2f38ab2c3
|
2180 2181 |
oom: if (old_page) |
09cbfeaf1
|
2182 |
put_page(old_page); |
2f38ab2c3
|
2183 2184 |
return VM_FAULT_OOM; } |
dd9061846
|
2185 2186 2187 2188 |
/* * Handle write page faults for VM_MIXEDMAP or VM_PFNMAP for a VM_SHARED * mapping */ |
bae473a42
|
2189 |
static int wp_pfn_shared(struct fault_env *fe, pte_t orig_pte) |
dd9061846
|
2190 |
{ |
bae473a42
|
2191 |
struct vm_area_struct *vma = fe->vma; |
dd9061846
|
2192 2193 2194 |
if (vma->vm_ops && vma->vm_ops->pfn_mkwrite) { struct vm_fault vmf = { .page = NULL, |
bae473a42
|
2195 2196 2197 |
.pgoff = linear_page_index(vma, fe->address), .virtual_address = (void __user *)(fe->address & PAGE_MASK), |
dd9061846
|
2198 2199 2200 |
.flags = FAULT_FLAG_WRITE | FAULT_FLAG_MKWRITE, }; int ret; |
bae473a42
|
2201 |
pte_unmap_unlock(fe->pte, fe->ptl); |
dd9061846
|
2202 2203 2204 |
ret = vma->vm_ops->pfn_mkwrite(vma, &vmf); if (ret & VM_FAULT_ERROR) return ret; |
bae473a42
|
2205 2206 |
fe->pte = pte_offset_map_lock(vma->vm_mm, fe->pmd, fe->address, &fe->ptl); |
dd9061846
|
2207 2208 2209 2210 |
/* * We might have raced with another page fault while we * released the pte_offset_map_lock. */ |
bae473a42
|
2211 2212 |
if (!pte_same(*fe->pte, orig_pte)) { pte_unmap_unlock(fe->pte, fe->ptl); |
dd9061846
|
2213 2214 2215 |
return 0; } } |
bae473a42
|
2216 |
return wp_page_reuse(fe, orig_pte, NULL, 0, 0); |
dd9061846
|
2217 |
} |
bae473a42
|
2218 2219 2220 |
static int wp_page_shared(struct fault_env *fe, pte_t orig_pte, struct page *old_page) __releases(fe->ptl) |
93e478d4c
|
2221 |
{ |
bae473a42
|
2222 |
struct vm_area_struct *vma = fe->vma; |
93e478d4c
|
2223 |
int page_mkwrite = 0; |
09cbfeaf1
|
2224 |
get_page(old_page); |
93e478d4c
|
2225 |
|
93e478d4c
|
2226 2227 |
if (vma->vm_ops && vma->vm_ops->page_mkwrite) { int tmp; |
bae473a42
|
2228 2229 |
pte_unmap_unlock(fe->pte, fe->ptl); tmp = do_page_mkwrite(vma, old_page, fe->address); |
93e478d4c
|
2230 2231 |
if (unlikely(!tmp || (tmp & (VM_FAULT_ERROR | VM_FAULT_NOPAGE)))) { |
09cbfeaf1
|
2232 |
put_page(old_page); |
93e478d4c
|
2233 2234 2235 2236 2237 2238 2239 2240 |
return tmp; } /* * Since we dropped the lock we need to revalidate * the PTE as someone else may have changed it. If * they did, we just return, as we can count on the * MMU to tell us if they didn't also make it writable. */ |
bae473a42
|
2241 2242 2243 |
fe->pte = pte_offset_map_lock(vma->vm_mm, fe->pmd, fe->address, &fe->ptl); if (!pte_same(*fe->pte, orig_pte)) { |
93e478d4c
|
2244 |
unlock_page(old_page); |
bae473a42
|
2245 |
pte_unmap_unlock(fe->pte, fe->ptl); |
09cbfeaf1
|
2246 |
put_page(old_page); |
93e478d4c
|
2247 2248 2249 2250 |
return 0; } page_mkwrite = 1; } |
bae473a42
|
2251 |
return wp_page_reuse(fe, orig_pte, old_page, page_mkwrite, 1); |
93e478d4c
|
2252 |
} |
2f38ab2c3
|
2253 |
/* |
1da177e4c
|
2254 2255 2256 2257 |
* This routine handles present pages, when users try to write * to a shared page. It is done by copying the page to a new address * and decrementing the shared-page counter for the old page. * |
1da177e4c
|
2258 2259 2260 2261 2262 2263 2264 2265 2266 |
* Note that this routine assumes that the protection checks have been * done by the caller (the low-level page fault routine in most cases). * Thus we can safely just mark it writable once we've done any necessary * COW. * * We also mark the page dirty at this point even though the page will * change only once the write actually happens. This avoids a few races, * and potentially makes it more efficient. * |
8f4e2101f
|
2267 2268 2269 |
* We enter with non-exclusive mmap_sem (to exclude vma changes, * but allow concurrent faults), with pte both mapped and locked. * We return with mmap_sem still held, but pte unmapped and unlocked. |
1da177e4c
|
2270 |
*/ |
bae473a42
|
2271 2272 |
static int do_wp_page(struct fault_env *fe, pte_t orig_pte) __releases(fe->ptl) |
1da177e4c
|
2273 |
{ |
bae473a42
|
2274 |
struct vm_area_struct *vma = fe->vma; |
2f38ab2c3
|
2275 |
struct page *old_page; |
1da177e4c
|
2276 |
|
bae473a42
|
2277 |
old_page = vm_normal_page(vma, fe->address, orig_pte); |
251b97f55
|
2278 2279 |
if (!old_page) { /* |
64e455079
|
2280 2281 |
* VM_MIXEDMAP !pfn_valid() case, or VM_SOFTDIRTY clear on a * VM_PFNMAP VMA. |
251b97f55
|
2282 2283 |
* * We should not cow pages in a shared writeable mapping. |
dd9061846
|
2284 |
* Just mark the pages writable and/or call ops->pfn_mkwrite. |
251b97f55
|
2285 2286 2287 |
*/ if ((vma->vm_flags & (VM_WRITE|VM_SHARED)) == (VM_WRITE|VM_SHARED)) |
bae473a42
|
2288 |
return wp_pfn_shared(fe, orig_pte); |
2f38ab2c3
|
2289 |
|
bae473a42
|
2290 2291 |
pte_unmap_unlock(fe->pte, fe->ptl); return wp_page_copy(fe, orig_pte, old_page); |
251b97f55
|
2292 |
} |
1da177e4c
|
2293 |
|
d08b3851d
|
2294 |
/* |
ee6a64578
|
2295 2296 |
* Take out anonymous pages first, anonymous shared vmas are * not dirty accountable. |
d08b3851d
|
2297 |
*/ |
9a8408951
|
2298 |
if (PageAnon(old_page) && !PageKsm(old_page)) { |
6d0a07edd
|
2299 |
int total_mapcount; |
ab967d860
|
2300 |
if (!trylock_page(old_page)) { |
09cbfeaf1
|
2301 |
get_page(old_page); |
bae473a42
|
2302 |
pte_unmap_unlock(fe->pte, fe->ptl); |
ab967d860
|
2303 |
lock_page(old_page); |
bae473a42
|
2304 2305 2306 |
fe->pte = pte_offset_map_lock(vma->vm_mm, fe->pmd, fe->address, &fe->ptl); if (!pte_same(*fe->pte, orig_pte)) { |
ab967d860
|
2307 |
unlock_page(old_page); |
bae473a42
|
2308 |
pte_unmap_unlock(fe->pte, fe->ptl); |
09cbfeaf1
|
2309 |
put_page(old_page); |
287668052
|
2310 |
return 0; |
ab967d860
|
2311 |
} |
09cbfeaf1
|
2312 |
put_page(old_page); |
ee6a64578
|
2313 |
} |
6d0a07edd
|
2314 2315 2316 2317 2318 2319 2320 2321 2322 |
if (reuse_swap_page(old_page, &total_mapcount)) { if (total_mapcount == 1) { /* * The page is all ours. Move it to * our anon_vma so the rmap code will * not search our parent or siblings. * Protected against the rmap code by * the page lock. */ |
5a49973d7
|
2323 |
page_move_anon_rmap(old_page, vma); |
6d0a07edd
|
2324 |
} |
b009c024f
|
2325 |
unlock_page(old_page); |
bae473a42
|
2326 |
return wp_page_reuse(fe, orig_pte, old_page, 0, 0); |
b009c024f
|
2327 |
} |
ab967d860
|
2328 |
unlock_page(old_page); |
ee6a64578
|
2329 |
} else if (unlikely((vma->vm_flags & (VM_WRITE|VM_SHARED)) == |
d08b3851d
|
2330 |
(VM_WRITE|VM_SHARED))) { |
bae473a42
|
2331 |
return wp_page_shared(fe, orig_pte, old_page); |
1da177e4c
|
2332 |
} |
1da177e4c
|
2333 2334 2335 2336 |
/* * Ok, we need to copy. Oh, well.. */ |
09cbfeaf1
|
2337 |
get_page(old_page); |
287668052
|
2338 |
|
bae473a42
|
2339 2340 |
pte_unmap_unlock(fe->pte, fe->ptl); return wp_page_copy(fe, orig_pte, old_page); |
1da177e4c
|
2341 |
} |
97a894136
|
2342 |
static void unmap_mapping_range_vma(struct vm_area_struct *vma, |
1da177e4c
|
2343 2344 2345 |
unsigned long start_addr, unsigned long end_addr, struct zap_details *details) { |
f5cc4eef9
|
2346 |
zap_page_range_single(vma, start_addr, end_addr - start_addr, details); |
1da177e4c
|
2347 |
} |
6b2dbba8b
|
2348 |
static inline void unmap_mapping_range_tree(struct rb_root *root, |
1da177e4c
|
2349 2350 2351 |
struct zap_details *details) { struct vm_area_struct *vma; |
1da177e4c
|
2352 |
pgoff_t vba, vea, zba, zea; |
6b2dbba8b
|
2353 |
vma_interval_tree_foreach(vma, root, |
1da177e4c
|
2354 |
details->first_index, details->last_index) { |
1da177e4c
|
2355 2356 |
vba = vma->vm_pgoff; |
d6e932177
|
2357 |
vea = vba + vma_pages(vma) - 1; |
1da177e4c
|
2358 2359 2360 2361 2362 2363 |
zba = details->first_index; if (zba < vba) zba = vba; zea = details->last_index; if (zea > vea) zea = vea; |
97a894136
|
2364 |
unmap_mapping_range_vma(vma, |
1da177e4c
|
2365 2366 |
((zba - vba) << PAGE_SHIFT) + vma->vm_start, ((zea - vba + 1) << PAGE_SHIFT) + vma->vm_start, |
97a894136
|
2367 |
details); |
1da177e4c
|
2368 2369 |
} } |
1da177e4c
|
2370 |
/** |
8a5f14a23
|
2371 2372 2373 2374 |
* unmap_mapping_range - unmap the portion of all mmaps in the specified * address_space corresponding to the specified page range in the underlying * file. * |
3d41088fa
|
2375 |
* @mapping: the address space containing mmaps to be unmapped. |
1da177e4c
|
2376 2377 |
* @holebegin: byte in first page to unmap, relative to the start of * the underlying file. This will be rounded down to a PAGE_SIZE |
25d9e2d15
|
2378 |
* boundary. Note that this is different from truncate_pagecache(), which |
1da177e4c
|
2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 |
* must keep the partial page. In contrast, we must get rid of * partial pages. * @holelen: size of prospective hole in bytes. This will be rounded * up to a PAGE_SIZE boundary. A holelen of zero truncates to the * end of the file. * @even_cows: 1 when truncating a file, unmap even private COWed pages; * but 0 when invalidating pagecache, don't throw away private data. */ void unmap_mapping_range(struct address_space *mapping, loff_t const holebegin, loff_t const holelen, int even_cows) { |
aac453635
|
2390 |
struct zap_details details = { }; |
1da177e4c
|
2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 |
pgoff_t hba = holebegin >> PAGE_SHIFT; pgoff_t hlen = (holelen + PAGE_SIZE - 1) >> PAGE_SHIFT; /* Check for overflow. */ if (sizeof(holelen) > sizeof(hlen)) { long long holeend = (holebegin + holelen + PAGE_SIZE - 1) >> PAGE_SHIFT; if (holeend & ~(long long)ULONG_MAX) hlen = ULONG_MAX - hba + 1; } details.check_mapping = even_cows? NULL: mapping; |
1da177e4c
|
2403 2404 2405 2406 |
details.first_index = hba; details.last_index = hba + hlen - 1; if (details.last_index < details.first_index) details.last_index = ULONG_MAX; |
1da177e4c
|
2407 |
|
46c043ede
|
2408 |
i_mmap_lock_write(mapping); |
6b2dbba8b
|
2409 |
if (unlikely(!RB_EMPTY_ROOT(&mapping->i_mmap))) |
1da177e4c
|
2410 |
unmap_mapping_range_tree(&mapping->i_mmap, &details); |
46c043ede
|
2411 |
i_mmap_unlock_write(mapping); |
1da177e4c
|
2412 2413 |
} EXPORT_SYMBOL(unmap_mapping_range); |
1da177e4c
|
2414 |
/* |
8f4e2101f
|
2415 2416 |
* We enter with non-exclusive mmap_sem (to exclude vma changes, * but allow concurrent faults), and pte mapped but not yet locked. |
9a95f3cf7
|
2417 2418 2419 2420 |
* We return with pte unmapped and unlocked. * * We return with the mmap_sem locked or unlocked in the same cases * as does filemap_fault(). |
1da177e4c
|
2421 |
*/ |
bae473a42
|
2422 |
int do_swap_page(struct fault_env *fe, pte_t orig_pte) |
1da177e4c
|
2423 |
{ |
bae473a42
|
2424 |
struct vm_area_struct *vma = fe->vma; |
56f31801c
|
2425 |
struct page *page, *swapcache; |
00501b531
|
2426 |
struct mem_cgroup *memcg; |
65500d234
|
2427 |
swp_entry_t entry; |
1da177e4c
|
2428 |
pte_t pte; |
d065bd810
|
2429 |
int locked; |
ad8c2ee80
|
2430 |
int exclusive = 0; |
83c54070e
|
2431 |
int ret = 0; |
1da177e4c
|
2432 |
|
bae473a42
|
2433 |
if (!pte_unmap_same(vma->vm_mm, fe->pmd, fe->pte, orig_pte)) |
8f4e2101f
|
2434 |
goto out; |
65500d234
|
2435 2436 |
entry = pte_to_swp_entry(orig_pte); |
d1737fdbe
|
2437 2438 |
if (unlikely(non_swap_entry(entry))) { if (is_migration_entry(entry)) { |
bae473a42
|
2439 |
migration_entry_wait(vma->vm_mm, fe->pmd, fe->address); |
d1737fdbe
|
2440 2441 2442 |
} else if (is_hwpoison_entry(entry)) { ret = VM_FAULT_HWPOISON; } else { |
bae473a42
|
2443 |
print_bad_pte(vma, fe->address, orig_pte, NULL); |
d99be1a8e
|
2444 |
ret = VM_FAULT_SIGBUS; |
d1737fdbe
|
2445 |
} |
0697212a4
|
2446 2447 |
goto out; } |
0ff922452
|
2448 |
delayacct_set_flag(DELAYACCT_PF_SWAPIN); |
1da177e4c
|
2449 2450 |
page = lookup_swap_cache(entry); if (!page) { |
02098feaa
|
2451 |
page = swapin_readahead(entry, |
bae473a42
|
2452 |
GFP_HIGHUSER_MOVABLE, vma, fe->address); |
1da177e4c
|
2453 2454 |
if (!page) { /* |
8f4e2101f
|
2455 2456 |
* Back out if somebody else faulted in this pte * while we released the pte lock. |
1da177e4c
|
2457 |
*/ |
bae473a42
|
2458 2459 2460 |
fe->pte = pte_offset_map_lock(vma->vm_mm, fe->pmd, fe->address, &fe->ptl); if (likely(pte_same(*fe->pte, orig_pte))) |
1da177e4c
|
2461 |
ret = VM_FAULT_OOM; |
0ff922452
|
2462 |
delayacct_clear_flag(DELAYACCT_PF_SWAPIN); |
65500d234
|
2463 |
goto unlock; |
1da177e4c
|
2464 2465 2466 2467 |
} /* Had to read the page from swap area: Major fault */ ret = VM_FAULT_MAJOR; |
f8891e5e1
|
2468 |
count_vm_event(PGMAJFAULT); |
bae473a42
|
2469 |
mem_cgroup_count_vm_event(vma->vm_mm, PGMAJFAULT); |
d1737fdbe
|
2470 |
} else if (PageHWPoison(page)) { |
71f72525d
|
2471 2472 2473 2474 |
/* * hwpoisoned dirty swapcache pages are kept for killing * owner processes (which may be unknown at hwpoison time) */ |
d1737fdbe
|
2475 2476 |
ret = VM_FAULT_HWPOISON; delayacct_clear_flag(DELAYACCT_PF_SWAPIN); |
56f31801c
|
2477 |
swapcache = page; |
4779cb31c
|
2478 |
goto out_release; |
1da177e4c
|
2479 |
} |
56f31801c
|
2480 |
swapcache = page; |
bae473a42
|
2481 |
locked = lock_page_or_retry(page, vma->vm_mm, fe->flags); |
e709ffd61
|
2482 |
|
073e587ec
|
2483 |
delayacct_clear_flag(DELAYACCT_PF_SWAPIN); |
d065bd810
|
2484 2485 2486 2487 |
if (!locked) { ret |= VM_FAULT_RETRY; goto out_release; } |
073e587ec
|
2488 |
|
4969c1192
|
2489 |
/* |
31c4a3d3a
|
2490 2491 2492 2493 |
* Make sure try_to_free_swap or reuse_swap_page or swapoff did not * release the swapcache from under us. The page pin, and pte_same * test below, are not enough to exclude that. Even if it is still * swapcache, we need to check that the page's swap has not changed. |
4969c1192
|
2494 |
*/ |
31c4a3d3a
|
2495 |
if (unlikely(!PageSwapCache(page) || page_private(page) != entry.val)) |
4969c1192
|
2496 |
goto out_page; |
bae473a42
|
2497 |
page = ksm_might_need_to_copy(page, vma, fe->address); |
cbf86cfe0
|
2498 2499 2500 |
if (unlikely(!page)) { ret = VM_FAULT_OOM; page = swapcache; |
cbf86cfe0
|
2501 |
goto out_page; |
5ad646880
|
2502 |
} |
bae473a42
|
2503 2504 |
if (mem_cgroup_try_charge(page, vma->vm_mm, GFP_KERNEL, &memcg, false)) { |
8a9f3ccd2
|
2505 |
ret = VM_FAULT_OOM; |
bc43f75cd
|
2506 |
goto out_page; |
8a9f3ccd2
|
2507 |
} |
1da177e4c
|
2508 |
/* |
8f4e2101f
|
2509 |
* Back out if somebody else already faulted in this pte. |
1da177e4c
|
2510 |
*/ |
bae473a42
|
2511 2512 2513 |
fe->pte = pte_offset_map_lock(vma->vm_mm, fe->pmd, fe->address, &fe->ptl); if (unlikely(!pte_same(*fe->pte, orig_pte))) |
b81074800
|
2514 |
goto out_nomap; |
b81074800
|
2515 2516 2517 2518 |
if (unlikely(!PageUptodate(page))) { ret = VM_FAULT_SIGBUS; goto out_nomap; |
1da177e4c
|
2519 |
} |
8c7c6e34a
|
2520 2521 2522 2523 2524 2525 2526 2527 |
/* * The page isn't present yet, go ahead with the fault. * * Be careful about the sequence of operations here. * To get its accounting right, reuse_swap_page() must be called * while the page is counted on swap but not yet in mapcount i.e. * before page_add_anon_rmap() and swap_free(); try_to_free_swap() * must be called after the swap_free(), or it will never succeed. |
8c7c6e34a
|
2528 |
*/ |
1da177e4c
|
2529 |
|
bae473a42
|
2530 2531 |
inc_mm_counter_fast(vma->vm_mm, MM_ANONPAGES); dec_mm_counter_fast(vma->vm_mm, MM_SWAPENTS); |
1da177e4c
|
2532 |
pte = mk_pte(page, vma->vm_page_prot); |
bae473a42
|
2533 |
if ((fe->flags & FAULT_FLAG_WRITE) && reuse_swap_page(page, NULL)) { |
1da177e4c
|
2534 |
pte = maybe_mkwrite(pte_mkdirty(pte), vma); |
bae473a42
|
2535 |
fe->flags &= ~FAULT_FLAG_WRITE; |
9a5b489b8
|
2536 |
ret |= VM_FAULT_WRITE; |
d281ee614
|
2537 |
exclusive = RMAP_EXCLUSIVE; |
1da177e4c
|
2538 |
} |
1da177e4c
|
2539 |
flush_icache_page(vma, page); |
179ef71cb
|
2540 2541 |
if (pte_swp_soft_dirty(orig_pte)) pte = pte_mksoft_dirty(pte); |
bae473a42
|
2542 |
set_pte_at(vma->vm_mm, fe->address, fe->pte, pte); |
00501b531
|
2543 |
if (page == swapcache) { |
bae473a42
|
2544 |
do_page_add_anon_rmap(page, vma, fe->address, exclusive); |
f627c2f53
|
2545 |
mem_cgroup_commit_charge(page, memcg, true, false); |
1a8018fb4
|
2546 |
activate_page(page); |
00501b531
|
2547 |
} else { /* ksm created a completely new copy */ |
bae473a42
|
2548 |
page_add_new_anon_rmap(page, vma, fe->address, false); |
f627c2f53
|
2549 |
mem_cgroup_commit_charge(page, memcg, false, false); |
00501b531
|
2550 2551 |
lru_cache_add_active_or_unevictable(page, vma); } |
1da177e4c
|
2552 |
|
c475a8ab6
|
2553 |
swap_free(entry); |
5ccc5abaa
|
2554 2555 |
if (mem_cgroup_swap_full(page) || (vma->vm_flags & VM_LOCKED) || PageMlocked(page)) |
a2c43eed8
|
2556 |
try_to_free_swap(page); |
c475a8ab6
|
2557 |
unlock_page(page); |
56f31801c
|
2558 |
if (page != swapcache) { |
4969c1192
|
2559 2560 2561 2562 2563 2564 2565 2566 2567 |
/* * Hold the lock to avoid the swap entry to be reused * until we take the PT lock for the pte_same() check * (to avoid false positives from pte_same). For * further safety release the lock after the swap_free * so that the swap count won't change under a * parallel locked swapcache. */ unlock_page(swapcache); |
09cbfeaf1
|
2568 |
put_page(swapcache); |
4969c1192
|
2569 |
} |
c475a8ab6
|
2570 |
|
bae473a42
|
2571 2572 |
if (fe->flags & FAULT_FLAG_WRITE) { ret |= do_wp_page(fe, pte); |
61469f1d5
|
2573 2574 |
if (ret & VM_FAULT_ERROR) ret &= VM_FAULT_ERROR; |
1da177e4c
|
2575 2576 2577 2578 |
goto out; } /* No need to invalidate - it was non-present before */ |
bae473a42
|
2579 |
update_mmu_cache(vma, fe->address, fe->pte); |
65500d234
|
2580 |
unlock: |
bae473a42
|
2581 |
pte_unmap_unlock(fe->pte, fe->ptl); |
1da177e4c
|
2582 2583 |
out: return ret; |
b81074800
|
2584 |
out_nomap: |
f627c2f53
|
2585 |
mem_cgroup_cancel_charge(page, memcg, false); |
bae473a42
|
2586 |
pte_unmap_unlock(fe->pte, fe->ptl); |
bc43f75cd
|
2587 |
out_page: |
b81074800
|
2588 |
unlock_page(page); |
4779cb31c
|
2589 |
out_release: |
09cbfeaf1
|
2590 |
put_page(page); |
56f31801c
|
2591 |
if (page != swapcache) { |
4969c1192
|
2592 |
unlock_page(swapcache); |
09cbfeaf1
|
2593 |
put_page(swapcache); |
4969c1192
|
2594 |
} |
65500d234
|
2595 |
return ret; |
1da177e4c
|
2596 2597 2598 |
} /* |
8ca3eb080
|
2599 2600 |
* This is like a special single-page "expand_{down|up}wards()", * except we must first make sure that 'address{-|+}PAGE_SIZE' |
320b2b8de
|
2601 |
* doesn't hit another vma. |
320b2b8de
|
2602 2603 2604 2605 2606 |
*/ static inline int check_stack_guard_page(struct vm_area_struct *vma, unsigned long address) { address &= PAGE_MASK; if ((vma->vm_flags & VM_GROWSDOWN) && address == vma->vm_start) { |
0e8e50e20
|
2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 |
struct vm_area_struct *prev = vma->vm_prev; /* * Is there a mapping abutting this one below? * * That's only ok if it's the same stack mapping * that has gotten split.. */ if (prev && prev->vm_end == address) return prev->vm_flags & VM_GROWSDOWN ? 0 : -ENOMEM; |
320b2b8de
|
2617 |
|
fee7e49d4
|
2618 |
return expand_downwards(vma, address - PAGE_SIZE); |
320b2b8de
|
2619 |
} |
8ca3eb080
|
2620 2621 2622 2623 2624 2625 |
if ((vma->vm_flags & VM_GROWSUP) && address + PAGE_SIZE == vma->vm_end) { struct vm_area_struct *next = vma->vm_next; /* As VM_GROWSDOWN but s/below/above/ */ if (next && next->vm_start == address + PAGE_SIZE) return next->vm_flags & VM_GROWSUP ? 0 : -ENOMEM; |
fee7e49d4
|
2626 |
return expand_upwards(vma, address + PAGE_SIZE); |
8ca3eb080
|
2627 |
} |
320b2b8de
|
2628 2629 2630 2631 |
return 0; } /* |
8f4e2101f
|
2632 2633 2634 |
* We enter with non-exclusive mmap_sem (to exclude vma changes, * but allow concurrent faults), and pte mapped but not yet locked. * We return with mmap_sem still held, but pte unmapped and unlocked. |
1da177e4c
|
2635 |
*/ |
bae473a42
|
2636 |
static int do_anonymous_page(struct fault_env *fe) |
1da177e4c
|
2637 |
{ |
bae473a42
|
2638 |
struct vm_area_struct *vma = fe->vma; |
00501b531
|
2639 |
struct mem_cgroup *memcg; |
8f4e2101f
|
2640 |
struct page *page; |
1da177e4c
|
2641 |
pte_t entry; |
1da177e4c
|
2642 |
|
6b7339f4c
|
2643 2644 2645 |
/* File mapping without ->vm_ops ? */ if (vma->vm_flags & VM_SHARED) return VM_FAULT_SIGBUS; |
11ac55247
|
2646 |
/* Check if we need to add a guard page to the stack */ |
bae473a42
|
2647 |
if (check_stack_guard_page(vma, fe->address) < 0) |
9c145c56d
|
2648 |
return VM_FAULT_SIGSEGV; |
320b2b8de
|
2649 |
|
7267ec008
|
2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 |
/* * Use pte_alloc() instead of pte_alloc_map(). We can't run * pte_offset_map() on pmds where a huge pmd might be created * from a different thread. * * pte_alloc_map() is safe to use under down_write(mmap_sem) or when * parallel threads are excluded by other means. * * Here we only have down_read(mmap_sem). */ if (pte_alloc(vma->vm_mm, fe->pmd, fe->address)) return VM_FAULT_OOM; /* See the comment in pte_alloc_one_map() */ if (unlikely(pmd_trans_unstable(fe->pmd))) return 0; |
11ac55247
|
2666 |
/* Use the zero-page for reads */ |
bae473a42
|
2667 2668 2669 |
if (!(fe->flags & FAULT_FLAG_WRITE) && !mm_forbids_zeropage(vma->vm_mm)) { entry = pte_mkspecial(pfn_pte(my_zero_pfn(fe->address), |
62eede62d
|
2670 |
vma->vm_page_prot)); |
bae473a42
|
2671 2672 2673 |
fe->pte = pte_offset_map_lock(vma->vm_mm, fe->pmd, fe->address, &fe->ptl); if (!pte_none(*fe->pte)) |
a13ea5b75
|
2674 |
goto unlock; |
6b251fc96
|
2675 2676 |
/* Deliver the page fault to userland, check inside PT lock */ if (userfaultfd_missing(vma)) { |
bae473a42
|
2677 2678 |
pte_unmap_unlock(fe->pte, fe->ptl); return handle_userfault(fe, VM_UFFD_MISSING); |
6b251fc96
|
2679 |
} |
a13ea5b75
|
2680 2681 |
goto setpte; } |
557ed1fa2
|
2682 |
/* Allocate our own private page. */ |
557ed1fa2
|
2683 2684 |
if (unlikely(anon_vma_prepare(vma))) goto oom; |
bae473a42
|
2685 |
page = alloc_zeroed_user_highpage_movable(vma, fe->address); |
557ed1fa2
|
2686 2687 |
if (!page) goto oom; |
eb3c24f30
|
2688 |
|
bae473a42
|
2689 |
if (mem_cgroup_try_charge(page, vma->vm_mm, GFP_KERNEL, &memcg, false)) |
eb3c24f30
|
2690 |
goto oom_free_page; |
52f37629f
|
2691 2692 2693 2694 2695 |
/* * The memory barrier inside __SetPageUptodate makes sure that * preceeding stores to the page contents become visible before * the set_pte_at() write. */ |
0ed361dec
|
2696 |
__SetPageUptodate(page); |
8f4e2101f
|
2697 |
|
557ed1fa2
|
2698 |
entry = mk_pte(page, vma->vm_page_prot); |
1ac0cb5d0
|
2699 2700 |
if (vma->vm_flags & VM_WRITE) entry = pte_mkwrite(pte_mkdirty(entry)); |
1da177e4c
|
2701 |
|
bae473a42
|
2702 2703 2704 |
fe->pte = pte_offset_map_lock(vma->vm_mm, fe->pmd, fe->address, &fe->ptl); if (!pte_none(*fe->pte)) |
557ed1fa2
|
2705 |
goto release; |
9ba692948
|
2706 |
|
6b251fc96
|
2707 2708 |
/* Deliver the page fault to userland, check inside PT lock */ if (userfaultfd_missing(vma)) { |
bae473a42
|
2709 |
pte_unmap_unlock(fe->pte, fe->ptl); |
f627c2f53
|
2710 |
mem_cgroup_cancel_charge(page, memcg, false); |
09cbfeaf1
|
2711 |
put_page(page); |
bae473a42
|
2712 |
return handle_userfault(fe, VM_UFFD_MISSING); |
6b251fc96
|
2713 |
} |
bae473a42
|
2714 2715 |
inc_mm_counter_fast(vma->vm_mm, MM_ANONPAGES); page_add_new_anon_rmap(page, vma, fe->address, false); |
f627c2f53
|
2716 |
mem_cgroup_commit_charge(page, memcg, false, false); |
00501b531
|
2717 |
lru_cache_add_active_or_unevictable(page, vma); |
a13ea5b75
|
2718 |
setpte: |
bae473a42
|
2719 |
set_pte_at(vma->vm_mm, fe->address, fe->pte, entry); |
1da177e4c
|
2720 2721 |
/* No need to invalidate - it was non-present before */ |
bae473a42
|
2722 |
update_mmu_cache(vma, fe->address, fe->pte); |
65500d234
|
2723 |
unlock: |
bae473a42
|
2724 |
pte_unmap_unlock(fe->pte, fe->ptl); |
83c54070e
|
2725 |
return 0; |
8f4e2101f
|
2726 |
release: |
f627c2f53
|
2727 |
mem_cgroup_cancel_charge(page, memcg, false); |
09cbfeaf1
|
2728 |
put_page(page); |
8f4e2101f
|
2729 |
goto unlock; |
8a9f3ccd2
|
2730 |
oom_free_page: |
09cbfeaf1
|
2731 |
put_page(page); |
65500d234
|
2732 |
oom: |
1da177e4c
|
2733 2734 |
return VM_FAULT_OOM; } |
9a95f3cf7
|
2735 2736 2737 2738 2739 |
/* * The mmap_sem must have been held on entry, and may have been * released depending on flags and vma->vm_ops->fault() return value. * See filemap_fault() and __lock_page_retry(). */ |
bae473a42
|
2740 2741 |
static int __do_fault(struct fault_env *fe, pgoff_t pgoff, struct page *cow_page, struct page **page, void **entry) |
7eae74af3
|
2742 |
{ |
bae473a42
|
2743 |
struct vm_area_struct *vma = fe->vma; |
7eae74af3
|
2744 2745 |
struct vm_fault vmf; int ret; |
bae473a42
|
2746 |
vmf.virtual_address = (void __user *)(fe->address & PAGE_MASK); |
7eae74af3
|
2747 |
vmf.pgoff = pgoff; |
bae473a42
|
2748 |
vmf.flags = fe->flags; |
7eae74af3
|
2749 |
vmf.page = NULL; |
c20cd45eb
|
2750 |
vmf.gfp_mask = __get_fault_gfp_mask(vma); |
2e4cdab05
|
2751 |
vmf.cow_page = cow_page; |
7eae74af3
|
2752 2753 2754 2755 |
ret = vma->vm_ops->fault(vma, &vmf); if (unlikely(ret & (VM_FAULT_ERROR | VM_FAULT_NOPAGE | VM_FAULT_RETRY))) return ret; |
bc2466e42
|
2756 2757 2758 2759 |
if (ret & VM_FAULT_DAX_LOCKED) { *entry = vmf.entry; return ret; } |
7eae74af3
|
2760 2761 2762 2763 |
if (unlikely(PageHWPoison(vmf.page))) { if (ret & VM_FAULT_LOCKED) unlock_page(vmf.page); |
09cbfeaf1
|
2764 |
put_page(vmf.page); |
7eae74af3
|
2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 |
return VM_FAULT_HWPOISON; } if (unlikely(!(ret & VM_FAULT_LOCKED))) lock_page(vmf.page); else VM_BUG_ON_PAGE(!PageLocked(vmf.page), vmf.page); *page = vmf.page; return ret; } |
7267ec008
|
2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 |
static int pte_alloc_one_map(struct fault_env *fe) { struct vm_area_struct *vma = fe->vma; if (!pmd_none(*fe->pmd)) goto map_pte; if (fe->prealloc_pte) { fe->ptl = pmd_lock(vma->vm_mm, fe->pmd); if (unlikely(!pmd_none(*fe->pmd))) { spin_unlock(fe->ptl); goto map_pte; } atomic_long_inc(&vma->vm_mm->nr_ptes); pmd_populate(vma->vm_mm, fe->pmd, fe->prealloc_pte); spin_unlock(fe->ptl); fe->prealloc_pte = 0; } else if (unlikely(pte_alloc(vma->vm_mm, fe->pmd, fe->address))) { return VM_FAULT_OOM; } map_pte: /* * If a huge pmd materialized under us just retry later. Use * pmd_trans_unstable() instead of pmd_trans_huge() to ensure the pmd * didn't become pmd_trans_huge under us and then back to pmd_none, as * a result of MADV_DONTNEED running immediately after a huge pmd fault * in a different thread of this mm, in turn leading to a misleading * pmd_trans_huge() retval. All we have to ensure is that it is a * regular pmd that we can walk with pte_offset_map() and we can do that * through an atomic read in C, which is what pmd_trans_unstable() * provides. */ if (pmd_trans_unstable(fe->pmd) || pmd_devmap(*fe->pmd)) return VM_FAULT_NOPAGE; fe->pte = pte_offset_map_lock(vma->vm_mm, fe->pmd, fe->address, &fe->ptl); return 0; } |
e496cf3d7
|
2815 |
#ifdef CONFIG_TRANSPARENT_HUGE_PAGECACHE |
101024596
|
2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 |
#define HPAGE_CACHE_INDEX_MASK (HPAGE_PMD_NR - 1) static inline bool transhuge_vma_suitable(struct vm_area_struct *vma, unsigned long haddr) { if (((vma->vm_start >> PAGE_SHIFT) & HPAGE_CACHE_INDEX_MASK) != (vma->vm_pgoff & HPAGE_CACHE_INDEX_MASK)) return false; if (haddr < vma->vm_start || haddr + HPAGE_PMD_SIZE > vma->vm_end) return false; return true; } static int do_set_pmd(struct fault_env *fe, struct page *page) { struct vm_area_struct *vma = fe->vma; bool write = fe->flags & FAULT_FLAG_WRITE; unsigned long haddr = fe->address & HPAGE_PMD_MASK; pmd_t entry; int i, ret; if (!transhuge_vma_suitable(vma, haddr)) return VM_FAULT_FALLBACK; ret = VM_FAULT_FALLBACK; page = compound_head(page); fe->ptl = pmd_lock(vma->vm_mm, fe->pmd); if (unlikely(!pmd_none(*fe->pmd))) goto out; for (i = 0; i < HPAGE_PMD_NR; i++) flush_icache_page(vma, page + i); entry = mk_huge_pmd(page, vma->vm_page_prot); if (write) entry = maybe_pmd_mkwrite(pmd_mkdirty(entry), vma); add_mm_counter(vma->vm_mm, MM_FILEPAGES, HPAGE_PMD_NR); page_add_file_rmap(page, true); set_pmd_at(vma->vm_mm, haddr, fe->pmd, entry); update_mmu_cache_pmd(vma, haddr, fe->pmd); /* fault is handled */ ret = 0; |
95ecedcd6
|
2863 |
count_vm_event(THP_FILE_MAPPED); |
101024596
|
2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 |
out: spin_unlock(fe->ptl); return ret; } #else static int do_set_pmd(struct fault_env *fe, struct page *page) { BUILD_BUG(); return 0; } #endif |
8c6e50b02
|
2875 |
/** |
7267ec008
|
2876 2877 |
* alloc_set_pte - setup new PTE entry for given page and add reverse page * mapping. If needed, the fucntion allocates page table or use pre-allocated. |
8c6e50b02
|
2878 |
* |
bae473a42
|
2879 |
* @fe: fault environment |
7267ec008
|
2880 |
* @memcg: memcg to charge page (only for private mappings) |
8c6e50b02
|
2881 |
* @page: page to map |
8c6e50b02
|
2882 |
* |
7267ec008
|
2883 |
* Caller must take care of unlocking fe->ptl, if fe->pte is non-NULL on return. |
8c6e50b02
|
2884 2885 2886 2887 |
* * Target users are page handler itself and implementations of * vm_ops->map_pages. */ |
7267ec008
|
2888 2889 |
int alloc_set_pte(struct fault_env *fe, struct mem_cgroup *memcg, struct page *page) |
3bb977946
|
2890 |
{ |
bae473a42
|
2891 2892 |
struct vm_area_struct *vma = fe->vma; bool write = fe->flags & FAULT_FLAG_WRITE; |
3bb977946
|
2893 |
pte_t entry; |
101024596
|
2894 |
int ret; |
e496cf3d7
|
2895 2896 |
if (pmd_none(*fe->pmd) && PageTransCompound(page) && IS_ENABLED(CONFIG_TRANSPARENT_HUGE_PAGECACHE)) { |
101024596
|
2897 2898 2899 2900 2901 2902 2903 |
/* THP on COW? */ VM_BUG_ON_PAGE(memcg, page); ret = do_set_pmd(fe, page); if (ret != VM_FAULT_FALLBACK) return ret; } |
3bb977946
|
2904 |
|
7267ec008
|
2905 |
if (!fe->pte) { |
101024596
|
2906 |
ret = pte_alloc_one_map(fe); |
7267ec008
|
2907 2908 2909 2910 2911 2912 2913 |
if (ret) return ret; } /* Re-check under ptl */ if (unlikely(!pte_none(*fe->pte))) return VM_FAULT_NOPAGE; |
3bb977946
|
2914 2915 2916 2917 |
flush_icache_page(vma, page); entry = mk_pte(page, vma->vm_page_prot); if (write) entry = maybe_mkwrite(pte_mkdirty(entry), vma); |
bae473a42
|
2918 2919 |
/* copy-on-write page */ if (write && !(vma->vm_flags & VM_SHARED)) { |
3bb977946
|
2920 |
inc_mm_counter_fast(vma->vm_mm, MM_ANONPAGES); |
bae473a42
|
2921 |
page_add_new_anon_rmap(page, vma, fe->address, false); |
7267ec008
|
2922 2923 |
mem_cgroup_commit_charge(page, memcg, false, false); lru_cache_add_active_or_unevictable(page, vma); |
3bb977946
|
2924 |
} else { |
eca56ff90
|
2925 |
inc_mm_counter_fast(vma->vm_mm, mm_counter_file(page)); |
dd78fedde
|
2926 |
page_add_file_rmap(page, false); |
3bb977946
|
2927 |
} |
bae473a42
|
2928 |
set_pte_at(vma->vm_mm, fe->address, fe->pte, entry); |
3bb977946
|
2929 2930 |
/* no need to invalidate: a not-present page won't be cached */ |
bae473a42
|
2931 |
update_mmu_cache(vma, fe->address, fe->pte); |
7267ec008
|
2932 2933 |
return 0; |
3bb977946
|
2934 |
} |
3a91053ae
|
2935 2936 |
static unsigned long fault_around_bytes __read_mostly = rounddown_pow_of_two(65536); |
a9b0f8618
|
2937 |
|
a9b0f8618
|
2938 2939 |
#ifdef CONFIG_DEBUG_FS static int fault_around_bytes_get(void *data, u64 *val) |
1592eef01
|
2940 |
{ |
a9b0f8618
|
2941 |
*val = fault_around_bytes; |
1592eef01
|
2942 2943 |
return 0; } |
b4903d6e8
|
2944 2945 2946 2947 2948 |
/* * fault_around_pages() and fault_around_mask() expects fault_around_bytes * rounded down to nearest page order. It's what do_fault_around() expects to * see. */ |
a9b0f8618
|
2949 |
static int fault_around_bytes_set(void *data, u64 val) |
1592eef01
|
2950 |
{ |
a9b0f8618
|
2951 |
if (val / PAGE_SIZE > PTRS_PER_PTE) |
1592eef01
|
2952 |
return -EINVAL; |
b4903d6e8
|
2953 2954 2955 2956 |
if (val > PAGE_SIZE) fault_around_bytes = rounddown_pow_of_two(val); else fault_around_bytes = PAGE_SIZE; /* rounddown_pow_of_two(0) is undefined */ |
1592eef01
|
2957 2958 |
return 0; } |
a9b0f8618
|
2959 2960 2961 |
DEFINE_SIMPLE_ATTRIBUTE(fault_around_bytes_fops, fault_around_bytes_get, fault_around_bytes_set, "%llu "); |
1592eef01
|
2962 2963 2964 2965 |
static int __init fault_around_debugfs(void) { void *ret; |
a9b0f8618
|
2966 2967 |
ret = debugfs_create_file("fault_around_bytes", 0644, NULL, NULL, &fault_around_bytes_fops); |
1592eef01
|
2968 |
if (!ret) |
a9b0f8618
|
2969 |
pr_warn("Failed to create fault_around_bytes in debugfs"); |
1592eef01
|
2970 2971 2972 |
return 0; } late_initcall(fault_around_debugfs); |
1592eef01
|
2973 |
#endif |
8c6e50b02
|
2974 |
|
1fdb412bd
|
2975 2976 2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 |
/* * do_fault_around() tries to map few pages around the fault address. The hope * is that the pages will be needed soon and this will lower the number of * faults to handle. * * It uses vm_ops->map_pages() to map the pages, which skips the page if it's * not ready to be mapped: not up-to-date, locked, etc. * * This function is called with the page table lock taken. In the split ptlock * case the page table lock only protects only those entries which belong to * the page table corresponding to the fault address. * * This function doesn't cross the VMA boundaries, in order to call map_pages() * only once. * * fault_around_pages() defines how many pages we'll try to map. * do_fault_around() expects it to return a power of two less than or equal to * PTRS_PER_PTE. * * The virtual address of the area that we map is naturally aligned to the * fault_around_pages() value (and therefore to page order). This way it's * easier to guarantee that we don't cross page table boundaries. */ |
7267ec008
|
2998 |
static int do_fault_around(struct fault_env *fe, pgoff_t start_pgoff) |
8c6e50b02
|
2999 |
{ |
7267ec008
|
3000 |
unsigned long address = fe->address, nr_pages, mask; |
bae473a42
|
3001 |
pgoff_t end_pgoff; |
7267ec008
|
3002 |
int off, ret = 0; |
8c6e50b02
|
3003 |
|
4db0c3c29
|
3004 |
nr_pages = READ_ONCE(fault_around_bytes) >> PAGE_SHIFT; |
aecd6f442
|
3005 |
mask = ~(nr_pages * PAGE_SIZE - 1) & PAGE_MASK; |
7267ec008
|
3006 3007 |
fe->address = max(address & mask, fe->vma->vm_start); off = ((address - fe->address) >> PAGE_SHIFT) & (PTRS_PER_PTE - 1); |
bae473a42
|
3008 |
start_pgoff -= off; |
8c6e50b02
|
3009 3010 |
/* |
bae473a42
|
3011 3012 |
* end_pgoff is either end of page table or end of vma * or fault_around_pages() from start_pgoff, depending what is nearest. |
8c6e50b02
|
3013 |
*/ |
bae473a42
|
3014 |
end_pgoff = start_pgoff - |
7267ec008
|
3015 |
((fe->address >> PAGE_SHIFT) & (PTRS_PER_PTE - 1)) + |
8c6e50b02
|
3016 |
PTRS_PER_PTE - 1; |
bae473a42
|
3017 3018 |
end_pgoff = min3(end_pgoff, vma_pages(fe->vma) + fe->vma->vm_pgoff - 1, start_pgoff + nr_pages - 1); |
8c6e50b02
|
3019 |
|
7267ec008
|
3020 3021 |
if (pmd_none(*fe->pmd)) { fe->prealloc_pte = pte_alloc_one(fe->vma->vm_mm, fe->address); |
c5f88bd29
|
3022 3023 |
if (!fe->prealloc_pte) goto out; |
7267ec008
|
3024 |
smp_wmb(); /* See comment in __pte_alloc() */ |
8c6e50b02
|
3025 |
} |
bae473a42
|
3026 |
fe->vma->vm_ops->map_pages(fe, start_pgoff, end_pgoff); |
7267ec008
|
3027 3028 3029 3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 |
/* preallocated pagetable is unused: free it */ if (fe->prealloc_pte) { pte_free(fe->vma->vm_mm, fe->prealloc_pte); fe->prealloc_pte = 0; } /* Huge page is mapped? Page fault is solved */ if (pmd_trans_huge(*fe->pmd)) { ret = VM_FAULT_NOPAGE; goto out; } /* ->map_pages() haven't done anything useful. Cold page cache? */ if (!fe->pte) goto out; /* check if the page fault is solved */ fe->pte -= (fe->address >> PAGE_SHIFT) - (address >> PAGE_SHIFT); if (!pte_none(*fe->pte)) ret = VM_FAULT_NOPAGE; pte_unmap_unlock(fe->pte, fe->ptl); |
bae473a42
|
3048 |
out: |
bae473a42
|
3049 |
fe->address = address; |
7267ec008
|
3050 3051 |
fe->pte = NULL; return ret; |
8c6e50b02
|
3052 |
} |
7267ec008
|
3053 |
static int do_read_fault(struct fault_env *fe, pgoff_t pgoff) |
e655fb290
|
3054 |
{ |
bae473a42
|
3055 |
struct vm_area_struct *vma = fe->vma; |
e655fb290
|
3056 |
struct page *fault_page; |
8c6e50b02
|
3057 3058 3059 3060 3061 3062 3063 |
int ret = 0; /* * Let's call ->map_pages() first and use ->fault() as fallback * if page by the offset is not ready to be mapped (cold cache or * something). */ |
9b4bdd2ff
|
3064 |
if (vma->vm_ops->map_pages && fault_around_bytes >> PAGE_SHIFT > 1) { |
7267ec008
|
3065 3066 3067 |
ret = do_fault_around(fe, pgoff); if (ret) return ret; |
8c6e50b02
|
3068 |
} |
e655fb290
|
3069 |
|
bae473a42
|
3070 |
ret = __do_fault(fe, pgoff, NULL, &fault_page, NULL); |
e655fb290
|
3071 3072 |
if (unlikely(ret & (VM_FAULT_ERROR | VM_FAULT_NOPAGE | VM_FAULT_RETRY))) return ret; |
7267ec008
|
3073 3074 |
ret |= alloc_set_pte(fe, NULL, fault_page); if (fe->pte) |
bae473a42
|
3075 |
pte_unmap_unlock(fe->pte, fe->ptl); |
e655fb290
|
3076 |
unlock_page(fault_page); |
7267ec008
|
3077 3078 |
if (unlikely(ret & (VM_FAULT_ERROR | VM_FAULT_NOPAGE | VM_FAULT_RETRY))) put_page(fault_page); |
e655fb290
|
3079 3080 |
return ret; } |
7267ec008
|
3081 |
static int do_cow_fault(struct fault_env *fe, pgoff_t pgoff) |
ec47c3b95
|
3082 |
{ |
bae473a42
|
3083 |
struct vm_area_struct *vma = fe->vma; |
ec47c3b95
|
3084 |
struct page *fault_page, *new_page; |
bc2466e42
|
3085 |
void *fault_entry; |
00501b531
|
3086 |
struct mem_cgroup *memcg; |
ec47c3b95
|
3087 3088 3089 3090 |
int ret; if (unlikely(anon_vma_prepare(vma))) return VM_FAULT_OOM; |
bae473a42
|
3091 |
new_page = alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma, fe->address); |
ec47c3b95
|
3092 3093 |
if (!new_page) return VM_FAULT_OOM; |
bae473a42
|
3094 3095 |
if (mem_cgroup_try_charge(new_page, vma->vm_mm, GFP_KERNEL, &memcg, false)) { |
09cbfeaf1
|
3096 |
put_page(new_page); |
ec47c3b95
|
3097 3098 |
return VM_FAULT_OOM; } |
bae473a42
|
3099 |
ret = __do_fault(fe, pgoff, new_page, &fault_page, &fault_entry); |
ec47c3b95
|
3100 3101 |
if (unlikely(ret & (VM_FAULT_ERROR | VM_FAULT_NOPAGE | VM_FAULT_RETRY))) goto uncharge_out; |
bc2466e42
|
3102 |
if (!(ret & VM_FAULT_DAX_LOCKED)) |
bae473a42
|
3103 |
copy_user_highpage(new_page, fault_page, fe->address, vma); |
ec47c3b95
|
3104 |
__SetPageUptodate(new_page); |
7267ec008
|
3105 3106 |
ret |= alloc_set_pte(fe, memcg, new_page); if (fe->pte) |
bae473a42
|
3107 |
pte_unmap_unlock(fe->pte, fe->ptl); |
bc2466e42
|
3108 |
if (!(ret & VM_FAULT_DAX_LOCKED)) { |
2e4cdab05
|
3109 |
unlock_page(fault_page); |
09cbfeaf1
|
3110 |
put_page(fault_page); |
2e4cdab05
|
3111 |
} else { |
bc2466e42
|
3112 |
dax_unlock_mapping_entry(vma->vm_file->f_mapping, pgoff); |
2e4cdab05
|
3113 |
} |
7267ec008
|
3114 3115 |
if (unlikely(ret & (VM_FAULT_ERROR | VM_FAULT_NOPAGE | VM_FAULT_RETRY))) goto uncharge_out; |
ec47c3b95
|
3116 3117 |
return ret; uncharge_out: |
f627c2f53
|
3118 |
mem_cgroup_cancel_charge(new_page, memcg, false); |
09cbfeaf1
|
3119 |
put_page(new_page); |
ec47c3b95
|
3120 3121 |
return ret; } |
7267ec008
|
3122 |
static int do_shared_fault(struct fault_env *fe, pgoff_t pgoff) |
1da177e4c
|
3123 |
{ |
bae473a42
|
3124 |
struct vm_area_struct *vma = fe->vma; |
f0c6d4d29
|
3125 3126 |
struct page *fault_page; struct address_space *mapping; |
f0c6d4d29
|
3127 |
int dirtied = 0; |
f0c6d4d29
|
3128 |
int ret, tmp; |
1d65f86db
|
3129 |
|
bae473a42
|
3130 |
ret = __do_fault(fe, pgoff, NULL, &fault_page, NULL); |
7eae74af3
|
3131 |
if (unlikely(ret & (VM_FAULT_ERROR | VM_FAULT_NOPAGE | VM_FAULT_RETRY))) |
f0c6d4d29
|
3132 |
return ret; |
1da177e4c
|
3133 3134 |
/* |
f0c6d4d29
|
3135 3136 |
* Check if the backing address space wants to know that the page is * about to become writable |
1da177e4c
|
3137 |
*/ |
fb09a4642
|
3138 3139 |
if (vma->vm_ops->page_mkwrite) { unlock_page(fault_page); |
bae473a42
|
3140 |
tmp = do_page_mkwrite(vma, fault_page, fe->address); |
fb09a4642
|
3141 3142 |
if (unlikely(!tmp || (tmp & (VM_FAULT_ERROR | VM_FAULT_NOPAGE)))) { |
09cbfeaf1
|
3143 |
put_page(fault_page); |
fb09a4642
|
3144 |
return tmp; |
4294621f4
|
3145 |
} |
fb09a4642
|
3146 |
} |
7267ec008
|
3147 3148 |
ret |= alloc_set_pte(fe, NULL, fault_page); if (fe->pte) |
bae473a42
|
3149 |
pte_unmap_unlock(fe->pte, fe->ptl); |
7267ec008
|
3150 3151 |
if (unlikely(ret & (VM_FAULT_ERROR | VM_FAULT_NOPAGE | VM_FAULT_RETRY))) { |
f0c6d4d29
|
3152 |
unlock_page(fault_page); |
09cbfeaf1
|
3153 |
put_page(fault_page); |
f0c6d4d29
|
3154 |
return ret; |
1da177e4c
|
3155 |
} |
b827e496c
|
3156 |
|
f0c6d4d29
|
3157 3158 |
if (set_page_dirty(fault_page)) dirtied = 1; |
d82fa87d2
|
3159 3160 3161 3162 3163 3164 |
/* * Take a local copy of the address_space - page.mapping may be zeroed * by truncate after unlock_page(). The address_space itself remains * pinned by vma->vm_file's reference. We rely on unlock_page()'s * release semantics to prevent the compiler from undoing this copying. */ |
1c290f642
|
3165 |
mapping = page_rmapping(fault_page); |
f0c6d4d29
|
3166 3167 3168 3169 3170 3171 3172 |
unlock_page(fault_page); if ((dirtied || vma->vm_ops->page_mkwrite) && mapping) { /* * Some device drivers do not set page.mapping but still * dirty their pages */ balance_dirty_pages_ratelimited(mapping); |
d08b3851d
|
3173 |
} |
d00806b18
|
3174 |
|
74ec67511
|
3175 |
if (!vma->vm_ops->page_mkwrite) |
f0c6d4d29
|
3176 |
file_update_time(vma->vm_file); |
b827e496c
|
3177 |
|
1d65f86db
|
3178 |
return ret; |
54cb8821d
|
3179 |
} |
d00806b18
|
3180 |
|
9a95f3cf7
|
3181 3182 3183 3184 3185 3186 |
/* * We enter with non-exclusive mmap_sem (to exclude vma changes, * but allow concurrent faults). * The mmap_sem may have been released depending on flags and our * return value. See filemap_fault() and __lock_page_or_retry(). */ |
7267ec008
|
3187 |
static int do_fault(struct fault_env *fe) |
54cb8821d
|
3188 |
{ |
bae473a42
|
3189 3190 |
struct vm_area_struct *vma = fe->vma; pgoff_t pgoff = linear_page_index(vma, fe->address); |
54cb8821d
|
3191 |
|
6b7339f4c
|
3192 3193 3194 |
/* The VMA was not fully populated on mmap() or missing VM_DONTEXPAND */ if (!vma->vm_ops->fault) return VM_FAULT_SIGBUS; |
bae473a42
|
3195 |
if (!(fe->flags & FAULT_FLAG_WRITE)) |
7267ec008
|
3196 |
return do_read_fault(fe, pgoff); |
ec47c3b95
|
3197 |
if (!(vma->vm_flags & VM_SHARED)) |
7267ec008
|
3198 3199 |
return do_cow_fault(fe, pgoff); return do_shared_fault(fe, pgoff); |
54cb8821d
|
3200 |
} |
b19a99392
|
3201 |
static int numa_migrate_prep(struct page *page, struct vm_area_struct *vma, |
04bb2f947
|
3202 3203 |
unsigned long addr, int page_nid, int *flags) |
9532fec11
|
3204 3205 3206 3207 |
{ get_page(page); count_vm_numa_event(NUMA_HINT_FAULTS); |
04bb2f947
|
3208 |
if (page_nid == numa_node_id()) { |
9532fec11
|
3209 |
count_vm_numa_event(NUMA_HINT_FAULTS_LOCAL); |
04bb2f947
|
3210 3211 |
*flags |= TNF_FAULT_LOCAL; } |
9532fec11
|
3212 3213 3214 |
return mpol_misplaced(page, vma, addr); } |
bae473a42
|
3215 |
static int do_numa_page(struct fault_env *fe, pte_t pte) |
d10e63f29
|
3216 |
{ |
bae473a42
|
3217 |
struct vm_area_struct *vma = fe->vma; |
4daae3b4b
|
3218 |
struct page *page = NULL; |
8191acbd3
|
3219 |
int page_nid = -1; |
90572890d
|
3220 |
int last_cpupid; |
cbee9f88e
|
3221 |
int target_nid; |
b8593bfda
|
3222 |
bool migrated = false; |
b191f9b10
|
3223 |
bool was_writable = pte_write(pte); |
6688cc054
|
3224 |
int flags = 0; |
d10e63f29
|
3225 3226 3227 3228 3229 3230 |
/* * The "pte" at this point cannot be used safely without * validation through pte_unmap_same(). It's of NUMA type but * the pfn may be screwed if the read is non atomic. * |
4d9424669
|
3231 3232 3233 |
* We can safely just do a "set_pte_at()", because the old * page table entry is not accessible, so there would be no * concurrent hardware modifications to the PTE. |
d10e63f29
|
3234 |
*/ |
bae473a42
|
3235 3236 3237 3238 |
fe->ptl = pte_lockptr(vma->vm_mm, fe->pmd); spin_lock(fe->ptl); if (unlikely(!pte_same(*fe->pte, pte))) { pte_unmap_unlock(fe->pte, fe->ptl); |
4daae3b4b
|
3239 3240 |
goto out; } |
4d9424669
|
3241 3242 3243 |
/* Make it present again */ pte = pte_modify(pte, vma->vm_page_prot); pte = pte_mkyoung(pte); |
b191f9b10
|
3244 3245 |
if (was_writable) pte = pte_mkwrite(pte); |
bae473a42
|
3246 3247 |
set_pte_at(vma->vm_mm, fe->address, fe->pte, pte); update_mmu_cache(vma, fe->address, fe->pte); |
d10e63f29
|
3248 |
|
bae473a42
|
3249 |
page = vm_normal_page(vma, fe->address, pte); |
d10e63f29
|
3250 |
if (!page) { |
bae473a42
|
3251 |
pte_unmap_unlock(fe->pte, fe->ptl); |
d10e63f29
|
3252 3253 |
return 0; } |
e81c48024
|
3254 3255 |
/* TODO: handle PTE-mapped THP */ if (PageCompound(page)) { |
bae473a42
|
3256 |
pte_unmap_unlock(fe->pte, fe->ptl); |
e81c48024
|
3257 3258 |
return 0; } |
6688cc054
|
3259 |
/* |
bea66fbd1
|
3260 3261 3262 3263 3264 3265 |
* Avoid grouping on RO pages in general. RO pages shouldn't hurt as * much anyway since they can be in shared cache state. This misses * the case where a mapping is writable but the process never writes * to it but pte_write gets cleared during protection updates and * pte_dirty has unpredictable behaviour between PTE scan updates, * background writeback, dirty balancing and application behaviour. |
6688cc054
|
3266 |
*/ |
d59dc7bcf
|
3267 |
if (!pte_write(pte)) |
6688cc054
|
3268 |
flags |= TNF_NO_GROUP; |
dabe1d992
|
3269 3270 3271 3272 3273 3274 |
/* * Flag if the page is shared between multiple address spaces. This * is later used when determining whether to group tasks together */ if (page_mapcount(page) > 1 && (vma->vm_flags & VM_SHARED)) flags |= TNF_SHARED; |
90572890d
|
3275 |
last_cpupid = page_cpupid_last(page); |
8191acbd3
|
3276 |
page_nid = page_to_nid(page); |
bae473a42
|
3277 3278 3279 |
target_nid = numa_migrate_prep(page, vma, fe->address, page_nid, &flags); pte_unmap_unlock(fe->pte, fe->ptl); |
4daae3b4b
|
3280 |
if (target_nid == -1) { |
4daae3b4b
|
3281 3282 3283 3284 3285 |
put_page(page); goto out; } /* Migrate to the requested node */ |
1bc115d87
|
3286 |
migrated = migrate_misplaced_page(page, vma, target_nid); |
6688cc054
|
3287 |
if (migrated) { |
8191acbd3
|
3288 |
page_nid = target_nid; |
6688cc054
|
3289 |
flags |= TNF_MIGRATED; |
074c23817
|
3290 3291 |
} else flags |= TNF_MIGRATE_FAIL; |
4daae3b4b
|
3292 3293 |
out: |
8191acbd3
|
3294 |
if (page_nid != -1) |
6688cc054
|
3295 |
task_numa_fault(last_cpupid, page_nid, 1, flags); |
d10e63f29
|
3296 3297 |
return 0; } |
bae473a42
|
3298 |
static int create_huge_pmd(struct fault_env *fe) |
b96375f74
|
3299 |
{ |
bae473a42
|
3300 |
struct vm_area_struct *vma = fe->vma; |
fb6dd5fa4
|
3301 |
if (vma_is_anonymous(vma)) |
bae473a42
|
3302 |
return do_huge_pmd_anonymous_page(fe); |
b96375f74
|
3303 |
if (vma->vm_ops->pmd_fault) |
bae473a42
|
3304 3305 |
return vma->vm_ops->pmd_fault(vma, fe->address, fe->pmd, fe->flags); |
b96375f74
|
3306 3307 |
return VM_FAULT_FALLBACK; } |
bae473a42
|
3308 |
static int wp_huge_pmd(struct fault_env *fe, pmd_t orig_pmd) |
b96375f74
|
3309 |
{ |
bae473a42
|
3310 3311 3312 3313 3314 |
if (vma_is_anonymous(fe->vma)) return do_huge_pmd_wp_page(fe, orig_pmd); if (fe->vma->vm_ops->pmd_fault) return fe->vma->vm_ops->pmd_fault(fe->vma, fe->address, fe->pmd, fe->flags); |
af9e4d5f2
|
3315 3316 3317 3318 |
/* COW handled on pte level: split pmd */ VM_BUG_ON_VMA(fe->vma->vm_flags & VM_SHARED, fe->vma); split_huge_pmd(fe->vma, fe->pmd, fe->address); |
b96375f74
|
3319 3320 |
return VM_FAULT_FALLBACK; } |
38e088546
|
3321 3322 3323 3324 |
static inline bool vma_is_accessible(struct vm_area_struct *vma) { return vma->vm_flags & (VM_READ | VM_EXEC | VM_WRITE); } |
1da177e4c
|
3325 3326 3327 3328 3329 3330 3331 3332 3333 |
/* * These routines also need to handle stuff like marking pages dirty * and/or accessed for architectures that don't do it in hardware (most * RISC architectures). The early dirtying is also good on the i386. * * There is also a hook called "update_mmu_cache()" that architectures * with external mmu caches can use to update those (ie the Sparc or * PowerPC hashed page tables that act as extended TLBs). * |
7267ec008
|
3334 3335 |
* We enter with non-exclusive mmap_sem (to exclude vma changes, but allow * concurrent faults). |
9a95f3cf7
|
3336 |
* |
7267ec008
|
3337 3338 |
* The mmap_sem may have been released depending on flags and our return value. * See filemap_fault() and __lock_page_or_retry(). |
1da177e4c
|
3339 |
*/ |
bae473a42
|
3340 |
static int handle_pte_fault(struct fault_env *fe) |
1da177e4c
|
3341 3342 |
{ pte_t entry; |
7267ec008
|
3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 |
if (unlikely(pmd_none(*fe->pmd))) { /* * Leave __pte_alloc() until later: because vm_ops->fault may * want to allocate huge page, and if we expose page table * for an instant, it will be difficult to retract from * concurrent faults and from rmap lookups. */ fe->pte = NULL; } else { /* See comment in pte_alloc_one_map() */ if (pmd_trans_unstable(fe->pmd) || pmd_devmap(*fe->pmd)) return 0; /* * A regular pmd is established and it can't morph into a huge * pmd from under us anymore at this point because we hold the * mmap_sem read mode and khugepaged takes it in write mode. * So now it's safe to run pte_offset_map(). */ fe->pte = pte_offset_map(fe->pmd, fe->address); entry = *fe->pte; /* * some architectures can have larger ptes than wordsize, * e.g.ppc44x-defconfig has CONFIG_PTE_64BIT=y and * CONFIG_32BIT=y, so READ_ONCE or ACCESS_ONCE cannot guarantee * atomic accesses. The code below just needs a consistent * view for the ifs and we later double check anyway with the * ptl lock held. So here a barrier will do. */ barrier(); |
65500d234
|
3374 |
if (pte_none(entry)) { |
7267ec008
|
3375 3376 |
pte_unmap(fe->pte); fe->pte = NULL; |
65500d234
|
3377 |
} |
1da177e4c
|
3378 |
} |
7267ec008
|
3379 3380 3381 3382 3383 3384 3385 3386 3387 |
if (!fe->pte) { if (vma_is_anonymous(fe->vma)) return do_anonymous_page(fe); else return do_fault(fe); } if (!pte_present(entry)) return do_swap_page(fe, entry); |
38e088546
|
3388 |
if (pte_protnone(entry) && vma_is_accessible(fe->vma)) |
bae473a42
|
3389 |
return do_numa_page(fe, entry); |
d10e63f29
|
3390 |
|
bae473a42
|
3391 3392 3393 |
fe->ptl = pte_lockptr(fe->vma->vm_mm, fe->pmd); spin_lock(fe->ptl); if (unlikely(!pte_same(*fe->pte, entry))) |
8f4e2101f
|
3394 |
goto unlock; |
bae473a42
|
3395 |
if (fe->flags & FAULT_FLAG_WRITE) { |
1da177e4c
|
3396 |
if (!pte_write(entry)) |
bae473a42
|
3397 |
return do_wp_page(fe, entry); |
1da177e4c
|
3398 3399 3400 |
entry = pte_mkdirty(entry); } entry = pte_mkyoung(entry); |
bae473a42
|
3401 3402 3403 |
if (ptep_set_access_flags(fe->vma, fe->address, fe->pte, entry, fe->flags & FAULT_FLAG_WRITE)) { update_mmu_cache(fe->vma, fe->address, fe->pte); |
1a44e1490
|
3404 3405 3406 3407 3408 3409 3410 |
} else { /* * This is needed only for protection faults but the arch code * is not yet telling us if this is a protection fault or not. * This still avoids useless tlb flushes for .text page faults * with threads. */ |
bae473a42
|
3411 3412 |
if (fe->flags & FAULT_FLAG_WRITE) flush_tlb_fix_spurious_fault(fe->vma, fe->address); |
1a44e1490
|
3413 |
} |
8f4e2101f
|
3414 |
unlock: |
bae473a42
|
3415 |
pte_unmap_unlock(fe->pte, fe->ptl); |
83c54070e
|
3416 |
return 0; |
1da177e4c
|
3417 3418 3419 3420 |
} /* * By the time we get here, we already hold the mm semaphore |
9a95f3cf7
|
3421 3422 3423 |
* * The mmap_sem may have been released depending on flags and our * return value. See filemap_fault() and __lock_page_or_retry(). |
1da177e4c
|
3424 |
*/ |
dcddffd41
|
3425 3426 |
static int __handle_mm_fault(struct vm_area_struct *vma, unsigned long address, unsigned int flags) |
1da177e4c
|
3427 |
{ |
bae473a42
|
3428 3429 3430 3431 3432 |
struct fault_env fe = { .vma = vma, .address = address, .flags = flags, }; |
dcddffd41
|
3433 |
struct mm_struct *mm = vma->vm_mm; |
1da177e4c
|
3434 3435 |
pgd_t *pgd; pud_t *pud; |
1da177e4c
|
3436 |
|
1da177e4c
|
3437 |
pgd = pgd_offset(mm, address); |
1da177e4c
|
3438 3439 |
pud = pud_alloc(mm, pgd, address); if (!pud) |
c74df32c7
|
3440 |
return VM_FAULT_OOM; |
bae473a42
|
3441 3442 |
fe.pmd = pmd_alloc(mm, pud, address); if (!fe.pmd) |
c74df32c7
|
3443 |
return VM_FAULT_OOM; |
bae473a42
|
3444 3445 |
if (pmd_none(*fe.pmd) && transparent_hugepage_enabled(vma)) { int ret = create_huge_pmd(&fe); |
c02925540
|
3446 3447 |
if (!(ret & VM_FAULT_FALLBACK)) return ret; |
71e3aac07
|
3448 |
} else { |
bae473a42
|
3449 |
pmd_t orig_pmd = *fe.pmd; |
1f1d06c34
|
3450 |
int ret; |
71e3aac07
|
3451 |
barrier(); |
5c7fb56e5
|
3452 |
if (pmd_trans_huge(orig_pmd) || pmd_devmap(orig_pmd)) { |
38e088546
|
3453 |
if (pmd_protnone(orig_pmd) && vma_is_accessible(vma)) |
bae473a42
|
3454 |
return do_huge_pmd_numa_page(&fe, orig_pmd); |
d10e63f29
|
3455 |
|
bae473a42
|
3456 3457 3458 |
if ((fe.flags & FAULT_FLAG_WRITE) && !pmd_write(orig_pmd)) { ret = wp_huge_pmd(&fe, orig_pmd); |
9845cbbd1
|
3459 3460 |
if (!(ret & VM_FAULT_FALLBACK)) return ret; |
a1dd450bc
|
3461 |
} else { |
bae473a42
|
3462 |
huge_pmd_set_accessed(&fe, orig_pmd); |
9845cbbd1
|
3463 |
return 0; |
1f1d06c34
|
3464 |
} |
71e3aac07
|
3465 3466 |
} } |
bae473a42
|
3467 |
return handle_pte_fault(&fe); |
1da177e4c
|
3468 |
} |
9a95f3cf7
|
3469 3470 3471 3472 3473 3474 |
/* * By the time we get here, we already hold the mm semaphore * * The mmap_sem may have been released depending on flags and our * return value. See filemap_fault() and __lock_page_or_retry(). */ |
dcddffd41
|
3475 3476 |
int handle_mm_fault(struct vm_area_struct *vma, unsigned long address, unsigned int flags) |
519e52473
|
3477 3478 3479 3480 3481 3482 |
{ int ret; __set_current_state(TASK_RUNNING); count_vm_event(PGFAULT); |
dcddffd41
|
3483 |
mem_cgroup_count_vm_event(vma->vm_mm, PGFAULT); |
519e52473
|
3484 3485 3486 3487 3488 3489 3490 3491 3492 |
/* do counter updates before entering really critical section. */ check_sync_rss_stat(current); /* * Enable the memcg OOM handling for faults triggered in user * space. Kernel faults are handled more gracefully. */ if (flags & FAULT_FLAG_USER) |
494264208
|
3493 |
mem_cgroup_oom_enable(); |
519e52473
|
3494 |
|
bae473a42
|
3495 3496 3497 3498 3499 3500 3501 3502 3503 |
if (!arch_vma_access_permitted(vma, flags & FAULT_FLAG_WRITE, flags & FAULT_FLAG_INSTRUCTION, flags & FAULT_FLAG_REMOTE)) return VM_FAULT_SIGSEGV; if (unlikely(is_vm_hugetlb_page(vma))) ret = hugetlb_fault(vma->vm_mm, vma, address, flags); else ret = __handle_mm_fault(vma, address, flags); |
519e52473
|
3504 |
|
494264208
|
3505 3506 3507 3508 3509 3510 3511 3512 3513 3514 3515 |
if (flags & FAULT_FLAG_USER) { mem_cgroup_oom_disable(); /* * The task may have entered a memcg OOM situation but * if the allocation error was handled gracefully (no * VM_FAULT_OOM), there is no need to kill anything. * Just clean up the OOM state peacefully. */ if (task_in_memcg_oom(current) && !(ret & VM_FAULT_OOM)) mem_cgroup_oom_synchronize(false); } |
3812c8c8f
|
3516 |
|
3f70dc38c
|
3517 3518 3519 3520 3521 3522 3523 3524 3525 3526 3527 3528 |
/* * This mm has been already reaped by the oom reaper and so the * refault cannot be trusted in general. Anonymous refaults would * lose data and give a zero page instead e.g. This is especially * problem for use_mm() because regular tasks will just die and * the corrupted data will not be visible anywhere while kthread * will outlive the oom victim and potentially propagate the date * further. */ if (unlikely((current->flags & PF_KTHREAD) && !(ret & VM_FAULT_ERROR) && test_bit(MMF_UNSTABLE, &vma->vm_mm->flags))) ret = VM_FAULT_SIGBUS; |
519e52473
|
3529 3530 |
return ret; } |
e1d6d01ab
|
3531 |
EXPORT_SYMBOL_GPL(handle_mm_fault); |
519e52473
|
3532 |
|
1da177e4c
|
3533 3534 3535 |
#ifndef __PAGETABLE_PUD_FOLDED /* * Allocate page upper directory. |
872fec16d
|
3536 |
* We've already handled the fast-path in-line. |
1da177e4c
|
3537 |
*/ |
1bb3630e8
|
3538 |
int __pud_alloc(struct mm_struct *mm, pgd_t *pgd, unsigned long address) |
1da177e4c
|
3539 |
{ |
c74df32c7
|
3540 3541 |
pud_t *new = pud_alloc_one(mm, address); if (!new) |
1bb3630e8
|
3542 |
return -ENOMEM; |
1da177e4c
|
3543 |
|
362a61ad6
|
3544 |
smp_wmb(); /* See comment in __pte_alloc */ |
872fec16d
|
3545 |
spin_lock(&mm->page_table_lock); |
1bb3630e8
|
3546 |
if (pgd_present(*pgd)) /* Another has populated it */ |
5e5419734
|
3547 |
pud_free(mm, new); |
1bb3630e8
|
3548 3549 |
else pgd_populate(mm, pgd, new); |
c74df32c7
|
3550 |
spin_unlock(&mm->page_table_lock); |
1bb3630e8
|
3551 |
return 0; |
1da177e4c
|
3552 3553 3554 3555 3556 3557 |
} #endif /* __PAGETABLE_PUD_FOLDED */ #ifndef __PAGETABLE_PMD_FOLDED /* * Allocate page middle directory. |
872fec16d
|
3558 |
* We've already handled the fast-path in-line. |
1da177e4c
|
3559 |
*/ |
1bb3630e8
|
3560 |
int __pmd_alloc(struct mm_struct *mm, pud_t *pud, unsigned long address) |
1da177e4c
|
3561 |
{ |
c74df32c7
|
3562 3563 |
pmd_t *new = pmd_alloc_one(mm, address); if (!new) |
1bb3630e8
|
3564 |
return -ENOMEM; |
1da177e4c
|
3565 |
|
362a61ad6
|
3566 |
smp_wmb(); /* See comment in __pte_alloc */ |
872fec16d
|
3567 |
spin_lock(&mm->page_table_lock); |
1da177e4c
|
3568 |
#ifndef __ARCH_HAS_4LEVEL_HACK |
dc6c9a35b
|
3569 3570 |
if (!pud_present(*pud)) { mm_inc_nr_pmds(mm); |
1bb3630e8
|
3571 |
pud_populate(mm, pud, new); |
dc6c9a35b
|
3572 |
} else /* Another has populated it */ |
5e5419734
|
3573 |
pmd_free(mm, new); |
dc6c9a35b
|
3574 3575 3576 |
#else if (!pgd_present(*pud)) { mm_inc_nr_pmds(mm); |
1bb3630e8
|
3577 |
pgd_populate(mm, pud, new); |
dc6c9a35b
|
3578 3579 |
} else /* Another has populated it */ pmd_free(mm, new); |
1da177e4c
|
3580 |
#endif /* __ARCH_HAS_4LEVEL_HACK */ |
c74df32c7
|
3581 |
spin_unlock(&mm->page_table_lock); |
1bb3630e8
|
3582 |
return 0; |
e0f39591c
|
3583 |
} |
1da177e4c
|
3584 |
#endif /* __PAGETABLE_PMD_FOLDED */ |
1b36ba815
|
3585 |
static int __follow_pte(struct mm_struct *mm, unsigned long address, |
f8ad0f499
|
3586 3587 3588 3589 3590 3591 3592 3593 3594 3595 3596 3597 3598 3599 3600 3601 |
pte_t **ptepp, spinlock_t **ptlp) { pgd_t *pgd; pud_t *pud; pmd_t *pmd; pte_t *ptep; pgd = pgd_offset(mm, address); if (pgd_none(*pgd) || unlikely(pgd_bad(*pgd))) goto out; pud = pud_offset(pgd, address); if (pud_none(*pud) || unlikely(pud_bad(*pud))) goto out; pmd = pmd_offset(pud, address); |
f66055ab6
|
3602 |
VM_BUG_ON(pmd_trans_huge(*pmd)); |
f8ad0f499
|
3603 3604 3605 3606 3607 3608 3609 3610 3611 3612 3613 3614 3615 3616 3617 3618 3619 3620 3621 |
if (pmd_none(*pmd) || unlikely(pmd_bad(*pmd))) goto out; /* We cannot handle huge page PFN maps. Luckily they don't exist. */ if (pmd_huge(*pmd)) goto out; ptep = pte_offset_map_lock(mm, pmd, address, ptlp); if (!ptep) goto out; if (!pte_present(*ptep)) goto unlock; *ptepp = ptep; return 0; unlock: pte_unmap_unlock(ptep, *ptlp); out: return -EINVAL; } |
1b36ba815
|
3622 3623 3624 3625 3626 3627 3628 3629 3630 3631 |
static inline int follow_pte(struct mm_struct *mm, unsigned long address, pte_t **ptepp, spinlock_t **ptlp) { int res; /* (void) is needed to make gcc happy */ (void) __cond_lock(*ptlp, !(res = __follow_pte(mm, address, ptepp, ptlp))); return res; } |
3b6748e2d
|
3632 3633 3634 3635 3636 3637 3638 3639 3640 3641 3642 3643 3644 3645 3646 3647 3648 3649 3650 3651 3652 3653 3654 3655 3656 3657 3658 3659 |
/** * follow_pfn - look up PFN at a user virtual address * @vma: memory mapping * @address: user virtual address * @pfn: location to store found PFN * * Only IO mappings and raw PFN mappings are allowed. * * Returns zero and the pfn at @pfn on success, -ve otherwise. */ int follow_pfn(struct vm_area_struct *vma, unsigned long address, unsigned long *pfn) { int ret = -EINVAL; spinlock_t *ptl; pte_t *ptep; if (!(vma->vm_flags & (VM_IO | VM_PFNMAP))) return ret; ret = follow_pte(vma->vm_mm, address, &ptep, &ptl); if (ret) return ret; *pfn = pte_pfn(*ptep); pte_unmap_unlock(ptep, ptl); return 0; } EXPORT_SYMBOL(follow_pfn); |
28b2ee20c
|
3660 |
#ifdef CONFIG_HAVE_IOREMAP_PROT |
d87fe6607
|
3661 3662 3663 |
int follow_phys(struct vm_area_struct *vma, unsigned long address, unsigned int flags, unsigned long *prot, resource_size_t *phys) |
28b2ee20c
|
3664 |
{ |
03668a4de
|
3665 |
int ret = -EINVAL; |
28b2ee20c
|
3666 3667 |
pte_t *ptep, pte; spinlock_t *ptl; |
28b2ee20c
|
3668 |
|
d87fe6607
|
3669 3670 |
if (!(vma->vm_flags & (VM_IO | VM_PFNMAP))) goto out; |
28b2ee20c
|
3671 |
|
03668a4de
|
3672 |
if (follow_pte(vma->vm_mm, address, &ptep, &ptl)) |
d87fe6607
|
3673 |
goto out; |
28b2ee20c
|
3674 |
pte = *ptep; |
03668a4de
|
3675 |
|
28b2ee20c
|
3676 3677 |
if ((flags & FOLL_WRITE) && !pte_write(pte)) goto unlock; |
28b2ee20c
|
3678 3679 |
*prot = pgprot_val(pte_pgprot(pte)); |
03668a4de
|
3680 |
*phys = (resource_size_t)pte_pfn(pte) << PAGE_SHIFT; |
28b2ee20c
|
3681 |
|
03668a4de
|
3682 |
ret = 0; |
28b2ee20c
|
3683 3684 3685 |
unlock: pte_unmap_unlock(ptep, ptl); out: |
d87fe6607
|
3686 |
return ret; |
28b2ee20c
|
3687 3688 3689 3690 3691 3692 3693 |
} int generic_access_phys(struct vm_area_struct *vma, unsigned long addr, void *buf, int len, int write) { resource_size_t phys_addr; unsigned long prot = 0; |
2bc7273b0
|
3694 |
void __iomem *maddr; |
28b2ee20c
|
3695 |
int offset = addr & (PAGE_SIZE-1); |
d87fe6607
|
3696 |
if (follow_phys(vma, addr, write, &prot, &phys_addr)) |
28b2ee20c
|
3697 |
return -EINVAL; |
9cb12d7b4
|
3698 |
maddr = ioremap_prot(phys_addr, PAGE_ALIGN(len + offset), prot); |
28b2ee20c
|
3699 3700 3701 3702 3703 3704 3705 3706 |
if (write) memcpy_toio(maddr + offset, buf, len); else memcpy_fromio(buf, maddr + offset, len); iounmap(maddr); return len; } |
5a73633ef
|
3707 |
EXPORT_SYMBOL_GPL(generic_access_phys); |
28b2ee20c
|
3708 |
#endif |
0ec76a110
|
3709 |
/* |
206cb6365
|
3710 3711 |
* Access another process' address space as given in mm. If non-NULL, use the * given task for page fault accounting. |
0ec76a110
|
3712 |
*/ |
e71b4e061
|
3713 |
int __access_remote_vm(struct task_struct *tsk, struct mm_struct *mm, |
442486ec1
|
3714 |
unsigned long addr, void *buf, int len, unsigned int gup_flags) |
0ec76a110
|
3715 |
{ |
0ec76a110
|
3716 |
struct vm_area_struct *vma; |
0ec76a110
|
3717 |
void *old_buf = buf; |
442486ec1
|
3718 |
int write = gup_flags & FOLL_WRITE; |
0ec76a110
|
3719 |
|
0ec76a110
|
3720 |
down_read(&mm->mmap_sem); |
183ff22bb
|
3721 |
/* ignore errors, just check how much was successfully transferred */ |
0ec76a110
|
3722 3723 3724 |
while (len) { int bytes, ret, offset; void *maddr; |
28b2ee20c
|
3725 |
struct page *page = NULL; |
0ec76a110
|
3726 |
|
1e9877902
|
3727 |
ret = get_user_pages_remote(tsk, mm, addr, 1, |
442486ec1
|
3728 |
gup_flags, &page, &vma); |
28b2ee20c
|
3729 |
if (ret <= 0) { |
dbffcd03d
|
3730 3731 3732 |
#ifndef CONFIG_HAVE_IOREMAP_PROT break; #else |
28b2ee20c
|
3733 3734 3735 3736 |
/* * Check if this is a VM_IO | VM_PFNMAP VMA, which * we can access using slightly different code. */ |
28b2ee20c
|
3737 |
vma = find_vma(mm, addr); |
fe936dfc2
|
3738 |
if (!vma || vma->vm_start > addr) |
28b2ee20c
|
3739 3740 3741 3742 3743 |
break; if (vma->vm_ops && vma->vm_ops->access) ret = vma->vm_ops->access(vma, addr, buf, len, write); if (ret <= 0) |
28b2ee20c
|
3744 3745 |
break; bytes = ret; |
dbffcd03d
|
3746 |
#endif |
0ec76a110
|
3747 |
} else { |
28b2ee20c
|
3748 3749 3750 3751 3752 3753 3754 3755 3756 3757 3758 3759 3760 3761 3762 |
bytes = len; offset = addr & (PAGE_SIZE-1); if (bytes > PAGE_SIZE-offset) bytes = PAGE_SIZE-offset; maddr = kmap(page); if (write) { copy_to_user_page(vma, page, addr, maddr + offset, buf, bytes); set_page_dirty_lock(page); } else { copy_from_user_page(vma, page, addr, buf, maddr + offset, bytes); } kunmap(page); |
09cbfeaf1
|
3763 |
put_page(page); |
0ec76a110
|
3764 |
} |
0ec76a110
|
3765 3766 3767 3768 3769 |
len -= bytes; buf += bytes; addr += bytes; } up_read(&mm->mmap_sem); |
0ec76a110
|
3770 3771 3772 |
return buf - old_buf; } |
03252919b
|
3773 |
|
5ddd36b9c
|
3774 |
/** |
ae91dbfc9
|
3775 |
* access_remote_vm - access another process' address space |
5ddd36b9c
|
3776 3777 3778 3779 |
* @mm: the mm_struct of the target address space * @addr: start address to access * @buf: source or destination buffer * @len: number of bytes to transfer |
6347e8d5b
|
3780 |
* @gup_flags: flags modifying lookup behaviour |
5ddd36b9c
|
3781 3782 3783 3784 |
* * The caller must hold a reference on @mm. */ int access_remote_vm(struct mm_struct *mm, unsigned long addr, |
6347e8d5b
|
3785 |
void *buf, int len, unsigned int gup_flags) |
5ddd36b9c
|
3786 |
{ |
6347e8d5b
|
3787 |
return __access_remote_vm(NULL, mm, addr, buf, len, gup_flags); |
5ddd36b9c
|
3788 |
} |
03252919b
|
3789 |
/* |
206cb6365
|
3790 3791 3792 3793 3794 |
* Access another process' address space. * Source/target buffer must be kernel space, * Do not walk the page table directly, use get_user_pages */ int access_process_vm(struct task_struct *tsk, unsigned long addr, |
f307ab6dc
|
3795 |
void *buf, int len, unsigned int gup_flags) |
206cb6365
|
3796 3797 3798 3799 3800 3801 3802 |
{ struct mm_struct *mm; int ret; mm = get_task_mm(tsk); if (!mm) return 0; |
f307ab6dc
|
3803 |
ret = __access_remote_vm(tsk, mm, addr, buf, len, gup_flags); |
442486ec1
|
3804 |
|
206cb6365
|
3805 3806 3807 3808 |
mmput(mm); return ret; } |
03252919b
|
3809 3810 3811 3812 3813 3814 3815 |
/* * Print the name of a VMA. */ void print_vma_addr(char *prefix, unsigned long ip) { struct mm_struct *mm = current->mm; struct vm_area_struct *vma; |
e8bff74af
|
3816 3817 3818 3819 3820 3821 |
/* * Do not print if we are in atomic * contexts (in exception stacks, etc.): */ if (preempt_count()) return; |
03252919b
|
3822 3823 3824 3825 3826 3827 |
down_read(&mm->mmap_sem); vma = find_vma(mm, ip); if (vma && vma->vm_file) { struct file *f = vma->vm_file; char *buf = (char *)__get_free_page(GFP_KERNEL); if (buf) { |
2fbc57c53
|
3828 |
char *p; |
03252919b
|
3829 |
|
9bf39ab2a
|
3830 |
p = file_path(f, buf, PAGE_SIZE); |
03252919b
|
3831 3832 |
if (IS_ERR(p)) p = "?"; |
2fbc57c53
|
3833 |
printk("%s%s[%lx+%lx]", prefix, kbasename(p), |
03252919b
|
3834 3835 3836 3837 3838 |
vma->vm_start, vma->vm_end - vma->vm_start); free_page((unsigned long)buf); } } |
51a07e50b
|
3839 |
up_read(&mm->mmap_sem); |
03252919b
|
3840 |
} |
3ee1afa30
|
3841 |
|
662bbcb27
|
3842 |
#if defined(CONFIG_PROVE_LOCKING) || defined(CONFIG_DEBUG_ATOMIC_SLEEP) |
9ec23531f
|
3843 |
void __might_fault(const char *file, int line) |
3ee1afa30
|
3844 |
{ |
95156f005
|
3845 3846 3847 3848 3849 3850 3851 3852 |
/* * Some code (nfs/sunrpc) uses socket ops on kernel memory while * holding the mmap_sem, this is safe because kernel memory doesn't * get paged out, therefore we'll never actually fault, and the * below annotations will generate false positives. */ if (segment_eq(get_fs(), KERNEL_DS)) return; |
9ec23531f
|
3853 |
if (pagefault_disabled()) |
662bbcb27
|
3854 |
return; |
9ec23531f
|
3855 3856 |
__might_sleep(file, line, 0); #if defined(CONFIG_DEBUG_ATOMIC_SLEEP) |
662bbcb27
|
3857 |
if (current->mm) |
3ee1afa30
|
3858 |
might_lock_read(¤t->mm->mmap_sem); |
9ec23531f
|
3859 |
#endif |
3ee1afa30
|
3860 |
} |
9ec23531f
|
3861 |
EXPORT_SYMBOL(__might_fault); |
3ee1afa30
|
3862 |
#endif |
47ad8475c
|
3863 3864 3865 3866 3867 3868 3869 3870 3871 3872 3873 3874 3875 3876 3877 3878 3879 3880 3881 3882 3883 3884 3885 3886 3887 3888 3889 3890 3891 3892 3893 3894 3895 3896 3897 3898 3899 3900 3901 3902 3903 3904 3905 3906 3907 3908 3909 3910 3911 3912 3913 3914 3915 3916 3917 3918 3919 3920 3921 3922 3923 3924 3925 3926 3927 3928 3929 3930 3931 3932 3933 |
#if defined(CONFIG_TRANSPARENT_HUGEPAGE) || defined(CONFIG_HUGETLBFS) static void clear_gigantic_page(struct page *page, unsigned long addr, unsigned int pages_per_huge_page) { int i; struct page *p = page; might_sleep(); for (i = 0; i < pages_per_huge_page; i++, p = mem_map_next(p, page, i)) { cond_resched(); clear_user_highpage(p, addr + i * PAGE_SIZE); } } void clear_huge_page(struct page *page, unsigned long addr, unsigned int pages_per_huge_page) { int i; if (unlikely(pages_per_huge_page > MAX_ORDER_NR_PAGES)) { clear_gigantic_page(page, addr, pages_per_huge_page); return; } might_sleep(); for (i = 0; i < pages_per_huge_page; i++) { cond_resched(); clear_user_highpage(page + i, addr + i * PAGE_SIZE); } } static void copy_user_gigantic_page(struct page *dst, struct page *src, unsigned long addr, struct vm_area_struct *vma, unsigned int pages_per_huge_page) { int i; struct page *dst_base = dst; struct page *src_base = src; for (i = 0; i < pages_per_huge_page; ) { cond_resched(); copy_user_highpage(dst, src, addr + i*PAGE_SIZE, vma); i++; dst = mem_map_next(dst, dst_base, i); src = mem_map_next(src, src_base, i); } } void copy_user_huge_page(struct page *dst, struct page *src, unsigned long addr, struct vm_area_struct *vma, unsigned int pages_per_huge_page) { int i; if (unlikely(pages_per_huge_page > MAX_ORDER_NR_PAGES)) { copy_user_gigantic_page(dst, src, addr, vma, pages_per_huge_page); return; } might_sleep(); for (i = 0; i < pages_per_huge_page; i++) { cond_resched(); copy_user_highpage(dst + i, src + i, addr + i*PAGE_SIZE, vma); } } #endif /* CONFIG_TRANSPARENT_HUGEPAGE || CONFIG_HUGETLBFS */ |
49076ec2c
|
3934 |
|
40b64acd1
|
3935 |
#if USE_SPLIT_PTE_PTLOCKS && ALLOC_SPLIT_PTLOCKS |
b35f1819a
|
3936 3937 3938 3939 3940 3941 3942 3943 |
static struct kmem_cache *page_ptl_cachep; void __init ptlock_cache_init(void) { page_ptl_cachep = kmem_cache_create("page->ptl", sizeof(spinlock_t), 0, SLAB_PANIC, NULL); } |
539edb584
|
3944 |
bool ptlock_alloc(struct page *page) |
49076ec2c
|
3945 3946 |
{ spinlock_t *ptl; |
b35f1819a
|
3947 |
ptl = kmem_cache_alloc(page_ptl_cachep, GFP_KERNEL); |
49076ec2c
|
3948 3949 |
if (!ptl) return false; |
539edb584
|
3950 |
page->ptl = ptl; |
49076ec2c
|
3951 3952 |
return true; } |
539edb584
|
3953 |
void ptlock_free(struct page *page) |
49076ec2c
|
3954 |
{ |
b35f1819a
|
3955 |
kmem_cache_free(page_ptl_cachep, page->ptl); |
49076ec2c
|
3956 3957 |
} #endif |