memory.c 100.2 KB
Newer Older
L
Linus Torvalds 已提交
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>
H
Hugh Dickins 已提交
48
#include <linux/ksm.h>
L
Linus Torvalds 已提交
49
#include <linux/rmap.h>
50
#include <linux/export.h>
51
#include <linux/delayacct.h>
L
Linus Torvalds 已提交
52
#include <linux/init.h>
P
Peter Zijlstra 已提交
53
#include <linux/writeback.h>
54
#include <linux/memcontrol.h>
A
Andrea Arcangeli 已提交
55
#include <linux/mmu_notifier.h>
56 57 58
#include <linux/kallsyms.h>
#include <linux/swapops.h>
#include <linux/elf.h>
59
#include <linux/gfp.h>
60
#include <linux/migrate.h>
A
Andy Shevchenko 已提交
61
#include <linux/string.h>
62
#include <linux/dma-debug.h>
63
#include <linux/debugfs.h>
L
Linus Torvalds 已提交
64

A
Alexey Dobriyan 已提交
65
#include <asm/io.h>
L
Linus Torvalds 已提交
66 67 68 69 70 71
#include <asm/pgalloc.h>
#include <asm/uaccess.h>
#include <asm/tlb.h>
#include <asm/tlbflush.h>
#include <asm/pgtable.h>

72 73
#include "internal.h"

74 75
#ifdef LAST_CPUPID_NOT_IN_PAGE_FLAGS
#warning Unfortunate NUMA and NUMA Balancing config, growing page-frame for last_cpupid.
76 77
#endif

A
Andy Whitcroft 已提交
78
#ifndef CONFIG_NEED_MULTIPLE_NODES
L
Linus Torvalds 已提交
79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97
/* 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

/*
 * 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;

EXPORT_SYMBOL(high_memory);

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
110 111 112 113

static int __init disable_randmaps(char *s)
{
	randomize_va_space = 0;
114
	return 1;
115 116 117
}
__setup("norandmaps", disable_randmaps);

H
Hugh Dickins 已提交
118
unsigned long zero_pfn __read_mostly;
H
Hugh Dickins 已提交
119
unsigned long highest_memmap_pfn __read_mostly;
H
Hugh Dickins 已提交
120

121 122
EXPORT_SYMBOL(zero_pfn);

H
Hugh Dickins 已提交
123 124 125 126 127 128 129 130 131
/*
 * 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);
132

K
KAMEZAWA Hiroyuki 已提交
133

134 135
#if defined(SPLIT_RSS_COUNTING)

136
void sync_mm_rss(struct mm_struct *mm)
137 138 139 140
{
	int i;

	for (i = 0; i < NR_MM_COUNTERS; i++) {
141 142 143
		if (current->rss_stat.count[i]) {
			add_mm_counter(mm, i, current->rss_stat.count[i]);
			current->rss_stat.count[i] = 0;
144 145
		}
	}
146
	current->rss_stat.events = 0;
147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167
}

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))
168
		sync_mm_rss(task->mm);
169
}
170
#else /* SPLIT_RSS_COUNTING */
171 172 173 174 175 176 177 178

#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)
{
}

179 180 181 182 183 184 185 186 187 188 189 190 191 192
#endif /* SPLIT_RSS_COUNTING */

#ifdef HAVE_GENERIC_MMU_GATHER

static int tlb_next_batch(struct mmu_gather *tlb)
{
	struct mmu_gather_batch *batch;

	batch = tlb->active;
	if (batch->next) {
		tlb->active = batch->next;
		return 1;
	}

193 194 195
	if (tlb->batch_count == MAX_GATHER_BATCH_COUNT)
		return 0;

196 197 198 199
	batch = (void *)__get_free_pages(GFP_NOWAIT | __GFP_NOWARN, 0);
	if (!batch)
		return 0;

200
	tlb->batch_count++;
201 202 203 204 205 206 207 208 209 210 211 212 213 214 215
	batch->next = NULL;
	batch->nr   = 0;
	batch->max  = MAX_GATHER_BATCH;

	tlb->active->next = batch;
	tlb->active = batch;

	return 1;
}

/* 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).
 */
216
void tlb_gather_mmu(struct mmu_gather *tlb, struct mm_struct *mm, unsigned long start, unsigned long end)
217 218 219
{
	tlb->mm = mm;

220 221
	/* Is it from 0 to ~0? */
	tlb->fullmm     = !(start | (end+1));
222
	tlb->need_flush_all = 0;
223 224 225 226
	tlb->local.next = NULL;
	tlb->local.nr   = 0;
	tlb->local.max  = ARRAY_SIZE(tlb->__pages);
	tlb->active     = &tlb->local;
227
	tlb->batch_count = 0;
228 229 230 231

#ifdef CONFIG_HAVE_RCU_TABLE_FREE
	tlb->batch = NULL;
#endif
232 233

	__tlb_reset_range(tlb);
234 235
}

236
static void tlb_flush_mmu_tlbonly(struct mmu_gather *tlb)
237
{
238 239 240
	if (!tlb->end)
		return;

241
	tlb_flush(tlb);
242
	mmu_notifier_invalidate_range(tlb->mm, tlb->start, tlb->end);
243 244
#ifdef CONFIG_HAVE_RCU_TABLE_FREE
	tlb_table_flush(tlb);
245
#endif
246
	__tlb_reset_range(tlb);
247 248 249 250 251
}

static void tlb_flush_mmu_free(struct mmu_gather *tlb)
{
	struct mmu_gather_batch *batch;
252

253
	for (batch = &tlb->local; batch && batch->nr; batch = batch->next) {
254 255 256 257 258 259
		free_pages_and_swap_cache(batch->pages, batch->nr);
		batch->nr = 0;
	}
	tlb->active = &tlb->local;
}

260 261 262 263 264 265
void tlb_flush_mmu(struct mmu_gather *tlb)
{
	tlb_flush_mmu_tlbonly(tlb);
	tlb_flush_mmu_free(tlb);
}

266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295
/* 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().
 */
int __tlb_remove_page(struct mmu_gather *tlb, struct page *page)
{
	struct mmu_gather_batch *batch;

296
	VM_BUG_ON(!tlb->end);
297 298 299 300 301 302

	batch = tlb->active;
	batch->pages[batch->nr++] = page;
	if (batch->nr == batch->max) {
		if (!tlb_next_batch(tlb))
			return 0;
303
		batch = tlb->active;
304
	}
305
	VM_BUG_ON_PAGE(batch->nr > batch->max, page);
306 307 308 309 310 311

	return batch->max - batch->nr;
}

#endif /* HAVE_GENERIC_MMU_GATHER */

312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384
#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;

	/*
	 * 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);
}

385
#endif /* CONFIG_HAVE_RCU_TABLE_FREE */
386

L
Linus Torvalds 已提交
387 388 389 390
/*
 * Note: this doesn't free the actual pages themselves. That
 * has been handled earlier when unmapping all the memory regions.
 */
391 392
static void free_pte_range(struct mmu_gather *tlb, pmd_t *pmd,
			   unsigned long addr)
L
Linus Torvalds 已提交
393
{
394
	pgtable_t token = pmd_pgtable(*pmd);
395
	pmd_clear(pmd);
396
	pte_free_tlb(tlb, token, addr);
397
	atomic_long_dec(&tlb->mm->nr_ptes);
L
Linus Torvalds 已提交
398 399
}

400 401 402
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)
L
Linus Torvalds 已提交
403 404 405
{
	pmd_t *pmd;
	unsigned long next;
406
	unsigned long start;
L
Linus Torvalds 已提交
407

408
	start = addr;
L
Linus Torvalds 已提交
409 410 411 412 413
	pmd = pmd_offset(pud, addr);
	do {
		next = pmd_addr_end(addr, end);
		if (pmd_none_or_clear_bad(pmd))
			continue;
414
		free_pte_range(tlb, pmd, addr);
L
Linus Torvalds 已提交
415 416
	} while (pmd++, addr = next, addr != end);

417 418 419 420 421 422 423
	start &= PUD_MASK;
	if (start < floor)
		return;
	if (ceiling) {
		ceiling &= PUD_MASK;
		if (!ceiling)
			return;
L
Linus Torvalds 已提交
424
	}
425 426 427 428 429
	if (end - 1 > ceiling - 1)
		return;

	pmd = pmd_offset(pud, start);
	pud_clear(pud);
430
	pmd_free_tlb(tlb, pmd, start);
431
	mm_dec_nr_pmds(tlb->mm);
L
Linus Torvalds 已提交
432 433
}

434 435 436
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)
L
Linus Torvalds 已提交
437 438 439
{
	pud_t *pud;
	unsigned long next;
440
	unsigned long start;
L
Linus Torvalds 已提交
441

442
	start = addr;
L
Linus Torvalds 已提交
443 444 445 446 447
	pud = pud_offset(pgd, addr);
	do {
		next = pud_addr_end(addr, end);
		if (pud_none_or_clear_bad(pud))
			continue;
448
		free_pmd_range(tlb, pud, addr, next, floor, ceiling);
L
Linus Torvalds 已提交
449 450
	} while (pud++, addr = next, addr != end);

451 452 453 454 455 456 457
	start &= PGDIR_MASK;
	if (start < floor)
		return;
	if (ceiling) {
		ceiling &= PGDIR_MASK;
		if (!ceiling)
			return;
L
Linus Torvalds 已提交
458
	}
459 460 461 462 463
	if (end - 1 > ceiling - 1)
		return;

	pud = pud_offset(pgd, start);
	pgd_clear(pgd);
464
	pud_free_tlb(tlb, pud, start);
L
Linus Torvalds 已提交
465 466 467
}

/*
468
 * This function frees user-level page tables of a process.
L
Linus Torvalds 已提交
469
 */
470
void free_pgd_range(struct mmu_gather *tlb,
471 472
			unsigned long addr, unsigned long end,
			unsigned long floor, unsigned long ceiling)
L
Linus Torvalds 已提交
473 474 475
{
	pgd_t *pgd;
	unsigned long next;
476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501

	/*
	 * 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.
	 */
L
Linus Torvalds 已提交
502

503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518
	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;

519
	pgd = pgd_offset(tlb->mm, addr);
L
Linus Torvalds 已提交
520 521 522 523
	do {
		next = pgd_addr_end(addr, end);
		if (pgd_none_or_clear_bad(pgd))
			continue;
524
		free_pud_range(tlb, pgd, addr, next, floor, ceiling);
L
Linus Torvalds 已提交
525
	} while (pgd++, addr = next, addr != end);
526 527
}

528
void free_pgtables(struct mmu_gather *tlb, struct vm_area_struct *vma,
529
		unsigned long floor, unsigned long ceiling)
530 531 532 533 534
{
	while (vma) {
		struct vm_area_struct *next = vma->vm_next;
		unsigned long addr = vma->vm_start;

535
		/*
N
npiggin@suse.de 已提交
536 537
		 * Hide vma from rmap and truncate_pagecache before freeing
		 * pgtables
538
		 */
539
		unlink_anon_vmas(vma);
540 541
		unlink_file_vma(vma);

542
		if (is_vm_hugetlb_page(vma)) {
543
			hugetlb_free_pgd_range(tlb, addr, vma->vm_end,
544
				floor, next? next->vm_start: ceiling);
545 546 547 548 549
		} else {
			/*
			 * Optimization: gather nearby vmas into one call down
			 */
			while (next && next->vm_start <= vma->vm_end + PMD_SIZE
550
			       && !is_vm_hugetlb_page(next)) {
551 552
				vma = next;
				next = vma->vm_next;
553
				unlink_anon_vmas(vma);
554
				unlink_file_vma(vma);
555 556 557 558
			}
			free_pgd_range(tlb, addr, vma->vm_end,
				floor, next? next->vm_start: ceiling);
		}
559 560
		vma = next;
	}
L
Linus Torvalds 已提交
561 562
}

563 564
int __pte_alloc(struct mm_struct *mm, struct vm_area_struct *vma,
		pmd_t *pmd, unsigned long address)
L
Linus Torvalds 已提交
565
{
566
	spinlock_t *ptl;
567
	pgtable_t new = pte_alloc_one(mm, address);
568
	int wait_split_huge_page;
569 570 571
	if (!new)
		return -ENOMEM;

572 573 574 575 576 577 578 579 580 581 582 583 584 585 586
	/*
	 * 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 */

587
	ptl = pmd_lock(mm, pmd);
588 589
	wait_split_huge_page = 0;
	if (likely(pmd_none(*pmd))) {	/* Has another populated it ? */
590
		atomic_long_inc(&mm->nr_ptes);
L
Linus Torvalds 已提交
591
		pmd_populate(mm, pmd, new);
592
		new = NULL;
593 594
	} else if (unlikely(pmd_trans_splitting(*pmd)))
		wait_split_huge_page = 1;
595
	spin_unlock(ptl);
596 597
	if (new)
		pte_free(mm, new);
598 599
	if (wait_split_huge_page)
		wait_split_huge_page(vma->anon_vma, pmd);
600
	return 0;
L
Linus Torvalds 已提交
601 602
}

603
int __pte_alloc_kernel(pmd_t *pmd, unsigned long address)
L
Linus Torvalds 已提交
604
{
605 606 607 608
	pte_t *new = pte_alloc_one_kernel(&init_mm, address);
	if (!new)
		return -ENOMEM;

609 610
	smp_wmb(); /* See comment in __pte_alloc */

611
	spin_lock(&init_mm.page_table_lock);
612
	if (likely(pmd_none(*pmd))) {	/* Has another populated it ? */
613
		pmd_populate_kernel(&init_mm, pmd, new);
614
		new = NULL;
615 616
	} else
		VM_BUG_ON(pmd_trans_splitting(*pmd));
617
	spin_unlock(&init_mm.page_table_lock);
618 619
	if (new)
		pte_free_kernel(&init_mm, new);
620
	return 0;
L
Linus Torvalds 已提交
621 622
}

K
KAMEZAWA Hiroyuki 已提交
623 624 625 626 627 628
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)
629
{
K
KAMEZAWA Hiroyuki 已提交
630 631
	int i;

632
	if (current->mm == mm)
633
		sync_mm_rss(mm);
K
KAMEZAWA Hiroyuki 已提交
634 635 636
	for (i = 0; i < NR_MM_COUNTERS; i++)
		if (rss[i])
			add_mm_counter(mm, i, rss[i]);
637 638
}

N
Nick Piggin 已提交
639
/*
640 641 642
 * 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.
N
Nick Piggin 已提交
643 644 645
 *
 * The calling function must still handle the error.
 */
646 647
static void print_bad_pte(struct vm_area_struct *vma, unsigned long addr,
			  pte_t pte, struct page *page)
N
Nick Piggin 已提交
648
{
649 650 651 652 653
	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;
654 655 656 657 658 659 660 661 662 663 664 665 666 667
	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) {
668 669
			printk(KERN_ALERT
				"BUG: Bad page map: %lu messages suppressed\n",
670 671 672 673 674 675 676
				nr_unshown);
			nr_unshown = 0;
		}
		nr_shown = 0;
	}
	if (nr_shown++ == 0)
		resume = jiffies + 60 * HZ;
677 678 679 680

	mapping = vma->vm_file ? vma->vm_file->f_mapping : NULL;
	index = linear_page_index(vma, addr);

681 682
	printk(KERN_ALERT
		"BUG: Bad page map in process %s  pte:%08llx pmd:%08llx\n",
683 684
		current->comm,
		(long long)pte_val(pte), (long long)pmd_val(*pmd));
685
	if (page)
686
		dump_page(page, "bad pte");
687
	printk(KERN_ALERT
688 689 690 691 692 693
		"addr:%p vm_flags:%08lx anon_vma:%p mapping:%p index:%lx\n",
		(void *)addr, vma->vm_flags, vma->anon_vma, mapping, index);
	/*
	 * Choose text because data symbols depend on CONFIG_KALLSYMS_ALL=y
	 */
	if (vma->vm_ops)
J
Joe Perches 已提交
694 695
		printk(KERN_ALERT "vma->vm_ops->fault: %pSR\n",
		       vma->vm_ops->fault);
A
Al Viro 已提交
696
	if (vma->vm_file)
J
Joe Perches 已提交
697 698
		printk(KERN_ALERT "vma->vm_file->f_op->mmap: %pSR\n",
		       vma->vm_file->f_op->mmap);
N
Nick Piggin 已提交
699
	dump_stack();
700
	add_taint(TAINT_BAD_PAGE, LOCKDEP_NOW_UNRELIABLE);
N
Nick Piggin 已提交
701 702
}

H
Hugh Dickins 已提交
703
/*
N
Nick Piggin 已提交
704
 * vm_normal_page -- This function gets the "struct page" associated with a pte.
705
 *
N
Nick Piggin 已提交
706 707 708
 * "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.
J
Jared Hulbert 已提交
709
 *
N
Nick Piggin 已提交
710 711 712 713 714 715 716 717
 * 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.
718
 *
J
Jared Hulbert 已提交
719 720
 * 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
N
Nick Piggin 已提交
721 722
 * set, and the vm_pgoff will point to the first PFN mapped: thus every special
 * mapping will always honor the rule
723 724 725
 *
 *	pfn_of_page == vma->vm_pgoff + ((addr - vma->vm_start) >> PAGE_SHIFT)
 *
N
Nick Piggin 已提交
726 727 728 729 730 731
 * 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).
J
Jared Hulbert 已提交
732 733
 *
 *
N
Nick Piggin 已提交
734
 * In order to support COW of arbitrary special mappings, we have VM_MIXEDMAP.
J
Jared Hulbert 已提交
735 736 737 738 739 740 741 742 743
 *
 * 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.
 *
H
Hugh Dickins 已提交
744
 */
N
Nick Piggin 已提交
745 746 747 748 749 750 751
#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)
H
Hugh Dickins 已提交
752
{
753
	unsigned long pfn = pte_pfn(pte);
N
Nick Piggin 已提交
754 755

	if (HAVE_PTE_SPECIAL) {
756
		if (likely(!pte_special(pte)))
757
			goto check_pfn;
758 759
		if (vma->vm_ops && vma->vm_ops->find_special_page)
			return vma->vm_ops->find_special_page(vma, addr);
H
Hugh Dickins 已提交
760 761
		if (vma->vm_flags & (VM_PFNMAP | VM_MIXEDMAP))
			return NULL;
H
Hugh Dickins 已提交
762
		if (!is_zero_pfn(pfn))
763
			print_bad_pte(vma, addr, pte, NULL);
N
Nick Piggin 已提交
764 765 766 767 768
		return NULL;
	}

	/* !HAVE_PTE_SPECIAL case follows: */

J
Jared Hulbert 已提交
769 770 771 772 773 774
	if (unlikely(vma->vm_flags & (VM_PFNMAP|VM_MIXEDMAP))) {
		if (vma->vm_flags & VM_MIXEDMAP) {
			if (!pfn_valid(pfn))
				return NULL;
			goto out;
		} else {
N
Nick Piggin 已提交
775 776
			unsigned long off;
			off = (addr - vma->vm_start) >> PAGE_SHIFT;
J
Jared Hulbert 已提交
777 778 779 780 781
			if (pfn == vma->vm_pgoff + off)
				return NULL;
			if (!is_cow_mapping(vma->vm_flags))
				return NULL;
		}
782 783
	}

784 785
	if (is_zero_pfn(pfn))
		return NULL;
786 787 788 789 790
check_pfn:
	if (unlikely(pfn > highest_memmap_pfn)) {
		print_bad_pte(vma, addr, pte, NULL);
		return NULL;
	}
791 792

	/*
N
Nick Piggin 已提交
793 794
	 * NOTE! We still have PageReserved() pages in the page tables.
	 * eg. VDSO mappings can cause them to exist.
795
	 */
J
Jared Hulbert 已提交
796
out:
797
	return pfn_to_page(pfn);
H
Hugh Dickins 已提交
798 799
}

L
Linus Torvalds 已提交
800 801 802 803 804 805
/*
 * 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.
 */

H
Hugh Dickins 已提交
806
static inline unsigned long
L
Linus Torvalds 已提交
807
copy_one_pte(struct mm_struct *dst_mm, struct mm_struct *src_mm,
N
Nick Piggin 已提交
808
		pte_t *dst_pte, pte_t *src_pte, struct vm_area_struct *vma,
H
Hugh Dickins 已提交
809
		unsigned long addr, int *rss)
L
Linus Torvalds 已提交
810
{
N
Nick Piggin 已提交
811
	unsigned long vm_flags = vma->vm_flags;
L
Linus Torvalds 已提交
812 813 814 815 816
	pte_t pte = *src_pte;
	struct page *page;

	/* pte contains position in swap or file, so copy. */
	if (unlikely(!pte_present(pte))) {
817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850
		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);

			if (PageAnon(page))
				rss[MM_ANONPAGES]++;
			else
				rss[MM_FILEPAGES]++;

			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);
851
			}
L
Linus Torvalds 已提交
852
		}
853
		goto out_set_pte;
L
Linus Torvalds 已提交
854 855 856 857 858 859
	}

	/*
	 * If it's a COW mapping, write protect it both
	 * in the parent and the child
	 */
860
	if (is_cow_mapping(vm_flags)) {
L
Linus Torvalds 已提交
861
		ptep_set_wrprotect(src_mm, addr, src_pte);
862
		pte = pte_wrprotect(pte);
L
Linus Torvalds 已提交
863 864 865 866 867 868 869 870 871
	}

	/*
	 * 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);
872 873 874 875

	page = vm_normal_page(vma, addr, pte);
	if (page) {
		get_page(page);
H
Hugh Dickins 已提交
876
		page_dup_rmap(page);
K
KAMEZAWA Hiroyuki 已提交
877 878 879 880
		if (PageAnon(page))
			rss[MM_ANONPAGES]++;
		else
			rss[MM_FILEPAGES]++;
881
	}
882 883 884

out_set_pte:
	set_pte_at(dst_mm, addr, dst_pte, pte);
H
Hugh Dickins 已提交
885
	return 0;
L
Linus Torvalds 已提交
886 887
}

888
static int copy_pte_range(struct mm_struct *dst_mm, struct mm_struct *src_mm,
889 890
		   pmd_t *dst_pmd, pmd_t *src_pmd, struct vm_area_struct *vma,
		   unsigned long addr, unsigned long end)
L
Linus Torvalds 已提交
891
{
892
	pte_t *orig_src_pte, *orig_dst_pte;
L
Linus Torvalds 已提交
893
	pte_t *src_pte, *dst_pte;
H
Hugh Dickins 已提交
894
	spinlock_t *src_ptl, *dst_ptl;
895
	int progress = 0;
K
KAMEZAWA Hiroyuki 已提交
896
	int rss[NR_MM_COUNTERS];
H
Hugh Dickins 已提交
897
	swp_entry_t entry = (swp_entry_t){0};
L
Linus Torvalds 已提交
898 899

again:
K
KAMEZAWA Hiroyuki 已提交
900 901
	init_rss_vec(rss);

H
Hugh Dickins 已提交
902
	dst_pte = pte_alloc_map_lock(dst_mm, dst_pmd, addr, &dst_ptl);
L
Linus Torvalds 已提交
903 904
	if (!dst_pte)
		return -ENOMEM;
P
Peter Zijlstra 已提交
905
	src_pte = pte_offset_map(src_pmd, addr);
H
Hugh Dickins 已提交
906
	src_ptl = pte_lockptr(src_mm, src_pmd);
I
Ingo Molnar 已提交
907
	spin_lock_nested(src_ptl, SINGLE_DEPTH_NESTING);
908 909
	orig_src_pte = src_pte;
	orig_dst_pte = dst_pte;
910
	arch_enter_lazy_mmu_mode();
L
Linus Torvalds 已提交
911 912 913 914 915 916

	do {
		/*
		 * We are holding two locks at this point - either of them
		 * could generate latencies in another task on another CPU.
		 */
917 918 919
		if (progress >= 32) {
			progress = 0;
			if (need_resched() ||
N
Nick Piggin 已提交
920
			    spin_needbreak(src_ptl) || spin_needbreak(dst_ptl))
921 922
				break;
		}
L
Linus Torvalds 已提交
923 924 925 926
		if (pte_none(*src_pte)) {
			progress++;
			continue;
		}
H
Hugh Dickins 已提交
927 928 929 930
		entry.val = copy_one_pte(dst_mm, src_mm, dst_pte, src_pte,
							vma, addr, rss);
		if (entry.val)
			break;
L
Linus Torvalds 已提交
931 932 933
		progress += 8;
	} while (dst_pte++, src_pte++, addr += PAGE_SIZE, addr != end);

934
	arch_leave_lazy_mmu_mode();
H
Hugh Dickins 已提交
935
	spin_unlock(src_ptl);
P
Peter Zijlstra 已提交
936
	pte_unmap(orig_src_pte);
K
KAMEZAWA Hiroyuki 已提交
937
	add_mm_rss_vec(dst_mm, rss);
938
	pte_unmap_unlock(orig_dst_pte, dst_ptl);
H
Hugh Dickins 已提交
939
	cond_resched();
H
Hugh Dickins 已提交
940 941 942 943 944 945

	if (entry.val) {
		if (add_swap_count_continuation(entry, GFP_KERNEL) < 0)
			return -ENOMEM;
		progress = 0;
	}
L
Linus Torvalds 已提交
946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963
	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);
964 965
		if (pmd_trans_huge(*src_pmd)) {
			int err;
966
			VM_BUG_ON(next-addr != HPAGE_PMD_SIZE);
967 968 969 970 971 972 973 974
			err = copy_huge_pmd(dst_mm, src_mm,
					    dst_pmd, src_pmd, addr, vma);
			if (err == -ENOMEM)
				return -ENOMEM;
			if (!err)
				continue;
			/* fall through */
		}
L
Linus Torvalds 已提交
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 1005 1006 1007 1008 1009 1010 1011 1012
		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;
1013 1014 1015
	unsigned long mmun_start;	/* For mmu_notifiers */
	unsigned long mmun_end;		/* For mmu_notifiers */
	bool is_cow;
A
Andrea Arcangeli 已提交
1016
	int ret;
L
Linus Torvalds 已提交
1017

1018 1019 1020 1021 1022 1023
	/*
	 * 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.
	 */
1024 1025 1026
	if (!(vma->vm_flags & (VM_HUGETLB | VM_PFNMAP | VM_MIXEDMAP)) &&
			!vma->anon_vma)
		return 0;
1027

L
Linus Torvalds 已提交
1028 1029 1030
	if (is_vm_hugetlb_page(vma))
		return copy_hugetlb_page_range(dst_mm, src_mm, vma);

1031
	if (unlikely(vma->vm_flags & VM_PFNMAP)) {
1032 1033 1034 1035
		/*
		 * We do not free on error cases below as remove_vma
		 * gets called on error from higher level routine
		 */
1036
		ret = track_pfn_copy(vma);
1037 1038 1039 1040
		if (ret)
			return ret;
	}

A
Andrea Arcangeli 已提交
1041 1042 1043 1044 1045 1046
	/*
	 * 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.
	 */
1047 1048 1049 1050 1051 1052
	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);
A
Andrea Arcangeli 已提交
1053 1054

	ret = 0;
L
Linus Torvalds 已提交
1055 1056 1057 1058 1059 1060
	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;
A
Andrea Arcangeli 已提交
1061 1062 1063 1064 1065
		if (unlikely(copy_pud_range(dst_mm, src_mm, dst_pgd, src_pgd,
					    vma, addr, next))) {
			ret = -ENOMEM;
			break;
		}
L
Linus Torvalds 已提交
1066
	} while (dst_pgd++, src_pgd++, addr = next, addr != end);
A
Andrea Arcangeli 已提交
1067

1068 1069
	if (is_cow)
		mmu_notifier_invalidate_range_end(src_mm, mmun_start, mmun_end);
A
Andrea Arcangeli 已提交
1070
	return ret;
L
Linus Torvalds 已提交
1071 1072
}

1073
static unsigned long zap_pte_range(struct mmu_gather *tlb,
N
Nick Piggin 已提交
1074
				struct vm_area_struct *vma, pmd_t *pmd,
L
Linus Torvalds 已提交
1075
				unsigned long addr, unsigned long end,
1076
				struct zap_details *details)
L
Linus Torvalds 已提交
1077
{
N
Nick Piggin 已提交
1078
	struct mm_struct *mm = tlb->mm;
P
Peter Zijlstra 已提交
1079
	int force_flush = 0;
K
KAMEZAWA Hiroyuki 已提交
1080
	int rss[NR_MM_COUNTERS];
1081
	spinlock_t *ptl;
1082
	pte_t *start_pte;
1083
	pte_t *pte;
1084
	swp_entry_t entry;
K
KAMEZAWA Hiroyuki 已提交
1085

P
Peter Zijlstra 已提交
1086
again:
1087
	init_rss_vec(rss);
1088 1089
	start_pte = pte_offset_map_lock(mm, pmd, addr, &ptl);
	pte = start_pte;
1090
	arch_enter_lazy_mmu_mode();
L
Linus Torvalds 已提交
1091 1092
	do {
		pte_t ptent = *pte;
1093
		if (pte_none(ptent)) {
L
Linus Torvalds 已提交
1094
			continue;
1095
		}
1096

L
Linus Torvalds 已提交
1097
		if (pte_present(ptent)) {
H
Hugh Dickins 已提交
1098
			struct page *page;
1099

1100
			page = vm_normal_page(vma, addr, ptent);
L
Linus Torvalds 已提交
1101 1102 1103 1104 1105 1106 1107 1108 1109 1110
			if (unlikely(details) && page) {
				/*
				 * unmap_shared_mapping_pages() wants to
				 * invalidate cache without truncating:
				 * unmap shared but keep private pages.
				 */
				if (details->check_mapping &&
				    details->check_mapping != page->mapping)
					continue;
			}
N
Nick Piggin 已提交
1111
			ptent = ptep_get_and_clear_full(mm, addr, pte,
1112
							tlb->fullmm);
L
Linus Torvalds 已提交
1113 1114 1115 1116
			tlb_remove_tlb_entry(tlb, pte, addr);
			if (unlikely(!page))
				continue;
			if (PageAnon(page))
K
KAMEZAWA Hiroyuki 已提交
1117
				rss[MM_ANONPAGES]--;
1118
			else {
1119 1120
				if (pte_dirty(ptent)) {
					force_flush = 1;
1121
					set_page_dirty(page);
1122
				}
1123
				if (pte_young(ptent) &&
1124
				    likely(!(vma->vm_flags & VM_SEQ_READ)))
1125
					mark_page_accessed(page);
K
KAMEZAWA Hiroyuki 已提交
1126
				rss[MM_FILEPAGES]--;
1127
			}
1128
			page_remove_rmap(page);
1129 1130
			if (unlikely(page_mapcount(page) < 0))
				print_bad_pte(vma, addr, ptent, page);
1131 1132
			if (unlikely(!__tlb_remove_page(tlb, page))) {
				force_flush = 1;
1133
				addr += PAGE_SIZE;
P
Peter Zijlstra 已提交
1134
				break;
1135
			}
L
Linus Torvalds 已提交
1136 1137
			continue;
		}
1138
		/* If details->check_mapping, we leave swap entries. */
L
Linus Torvalds 已提交
1139 1140
		if (unlikely(details))
			continue;
K
KAMEZAWA Hiroyuki 已提交
1141

1142 1143 1144 1145 1146
		entry = pte_to_swp_entry(ptent);
		if (!non_swap_entry(entry))
			rss[MM_SWAPENTS]--;
		else if (is_migration_entry(entry)) {
			struct page *page;
1147

1148
			page = migration_entry_to_page(entry);
1149

1150 1151 1152 1153
			if (PageAnon(page))
				rss[MM_ANONPAGES]--;
			else
				rss[MM_FILEPAGES]--;
K
KAMEZAWA Hiroyuki 已提交
1154
		}
1155 1156
		if (unlikely(!free_swap_and_cache(entry)))
			print_bad_pte(vma, addr, ptent, NULL);
1157
		pte_clear_not_present_full(mm, addr, pte, tlb->fullmm);
1158
	} while (pte++, addr += PAGE_SIZE, addr != end);
1159

K
KAMEZAWA Hiroyuki 已提交
1160
	add_mm_rss_vec(mm, rss);
1161
	arch_leave_lazy_mmu_mode();
1162

1163
	/* Do the actual TLB flush before dropping ptl */
1164
	if (force_flush)
1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176
		tlb_flush_mmu_tlbonly(tlb);
	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);
1177 1178

		if (addr != end)
P
Peter Zijlstra 已提交
1179 1180 1181
			goto again;
	}

1182
	return addr;
L
Linus Torvalds 已提交
1183 1184
}

1185
static inline unsigned long zap_pmd_range(struct mmu_gather *tlb,
N
Nick Piggin 已提交
1186
				struct vm_area_struct *vma, pud_t *pud,
L
Linus Torvalds 已提交
1187
				unsigned long addr, unsigned long end,
1188
				struct zap_details *details)
L
Linus Torvalds 已提交
1189 1190 1191 1192 1193 1194 1195
{
	pmd_t *pmd;
	unsigned long next;

	pmd = pmd_offset(pud, addr);
	do {
		next = pmd_addr_end(addr, end);
1196
		if (pmd_trans_huge(*pmd)) {
1197
			if (next - addr != HPAGE_PMD_SIZE) {
1198 1199 1200 1201 1202 1203 1204 1205 1206
#ifdef CONFIG_DEBUG_VM
				if (!rwsem_is_locked(&tlb->mm->mmap_sem)) {
					pr_err("%s: mmap_sem is unlocked! addr=0x%lx end=0x%lx vma->vm_start=0x%lx vma->vm_end=0x%lx\n",
						__func__, addr, end,
						vma->vm_start,
						vma->vm_end);
					BUG();
				}
#endif
1207
				split_huge_page_pmd(vma, addr, pmd);
S
Shaohua Li 已提交
1208
			} else if (zap_huge_pmd(tlb, vma, pmd, addr))
1209
				goto next;
1210 1211
			/* fall through */
		}
1212 1213 1214 1215 1216 1217 1218 1219 1220
		/*
		 * 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;
1221
		next = zap_pte_range(tlb, vma, pmd, addr, next, details);
1222
next:
1223 1224
		cond_resched();
	} while (pmd++, addr = next, addr != end);
1225 1226

	return addr;
L
Linus Torvalds 已提交
1227 1228
}

1229
static inline unsigned long zap_pud_range(struct mmu_gather *tlb,
N
Nick Piggin 已提交
1230
				struct vm_area_struct *vma, pgd_t *pgd,
L
Linus Torvalds 已提交
1231
				unsigned long addr, unsigned long end,
1232
				struct zap_details *details)
L
Linus Torvalds 已提交
1233 1234 1235 1236 1237 1238 1239
{
	pud_t *pud;
	unsigned long next;

	pud = pud_offset(pgd, addr);
	do {
		next = pud_addr_end(addr, end);
1240
		if (pud_none_or_clear_bad(pud))
L
Linus Torvalds 已提交
1241
			continue;
1242 1243
		next = zap_pmd_range(tlb, vma, pud, addr, next, details);
	} while (pud++, addr = next, addr != end);
1244 1245

	return addr;
L
Linus Torvalds 已提交
1246 1247
}

A
Al Viro 已提交
1248 1249 1250 1251
static void unmap_page_range(struct mmu_gather *tlb,
			     struct vm_area_struct *vma,
			     unsigned long addr, unsigned long end,
			     struct zap_details *details)
L
Linus Torvalds 已提交
1252 1253 1254 1255
{
	pgd_t *pgd;
	unsigned long next;

1256
	if (details && !details->check_mapping)
L
Linus Torvalds 已提交
1257 1258 1259 1260 1261 1262 1263
		details = NULL;

	BUG_ON(addr >= end);
	tlb_start_vma(tlb, vma);
	pgd = pgd_offset(vma->vm_mm, addr);
	do {
		next = pgd_addr_end(addr, end);
1264
		if (pgd_none_or_clear_bad(pgd))
L
Linus Torvalds 已提交
1265
			continue;
1266 1267
		next = zap_pud_range(tlb, vma, pgd, addr, next, details);
	} while (pgd++, addr = next, addr != end);
L
Linus Torvalds 已提交
1268 1269
	tlb_end_vma(tlb, vma);
}
1270

1271 1272 1273

static void unmap_single_vma(struct mmu_gather *tlb,
		struct vm_area_struct *vma, unsigned long start_addr,
1274
		unsigned long end_addr,
1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285
		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;

1286 1287 1288
	if (vma->vm_file)
		uprobe_munmap(vma, start, end);

1289
	if (unlikely(vma->vm_flags & VM_PFNMAP))
1290
		untrack_pfn(vma, 0, 0);
1291 1292 1293 1294 1295 1296 1297

	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
1298
			 * cleanup path of mmap_region. When
1299
			 * hugetlbfs ->mmap method fails,
1300
			 * mmap_region() nullifies vma->vm_file
1301 1302 1303 1304
			 * before calling this function to clean up.
			 * Since no pte has actually been setup, it is
			 * safe to do nothing in this case.
			 */
1305
			if (vma->vm_file) {
1306
				i_mmap_lock_write(vma->vm_file->f_mapping);
1307
				__unmap_hugepage_range_final(tlb, vma, start, end, NULL);
1308
				i_mmap_unlock_write(vma->vm_file->f_mapping);
1309
			}
1310 1311 1312
		} else
			unmap_page_range(tlb, vma, start, end, details);
	}
L
Linus Torvalds 已提交
1313 1314 1315 1316
}

/**
 * unmap_vmas - unmap a range of memory covered by a list of vma's
1317
 * @tlb: address of the caller's struct mmu_gather
L
Linus Torvalds 已提交
1318 1319 1320 1321
 * @vma: the starting vma
 * @start_addr: virtual address at which to start unmapping
 * @end_addr: virtual address at which to end unmapping
 *
1322
 * Unmap all pages in the vma list.
L
Linus Torvalds 已提交
1323 1324 1325 1326 1327 1328 1329 1330 1331 1332
 *
 * 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.
 */
A
Al Viro 已提交
1333
void unmap_vmas(struct mmu_gather *tlb,
L
Linus Torvalds 已提交
1334
		struct vm_area_struct *vma, unsigned long start_addr,
1335
		unsigned long end_addr)
L
Linus Torvalds 已提交
1336
{
A
Andrea Arcangeli 已提交
1337
	struct mm_struct *mm = vma->vm_mm;
L
Linus Torvalds 已提交
1338

A
Andrea Arcangeli 已提交
1339
	mmu_notifier_invalidate_range_start(mm, start_addr, end_addr);
1340
	for ( ; vma && vma->vm_start < end_addr; vma = vma->vm_next)
1341
		unmap_single_vma(tlb, vma, start_addr, end_addr, NULL);
A
Andrea Arcangeli 已提交
1342
	mmu_notifier_invalidate_range_end(mm, start_addr, end_addr);
L
Linus Torvalds 已提交
1343 1344 1345 1346 1347
}

/**
 * zap_page_range - remove user pages in a given range
 * @vma: vm_area_struct holding the applicable pages
1348
 * @start: starting address of pages to zap
L
Linus Torvalds 已提交
1349
 * @size: number of bytes to zap
1350
 * @details: details of shared cache invalidation
1351 1352
 *
 * Caller must protect the VMA list
L
Linus Torvalds 已提交
1353
 */
1354
void zap_page_range(struct vm_area_struct *vma, unsigned long start,
L
Linus Torvalds 已提交
1355 1356 1357
		unsigned long size, struct zap_details *details)
{
	struct mm_struct *mm = vma->vm_mm;
P
Peter Zijlstra 已提交
1358
	struct mmu_gather tlb;
1359
	unsigned long end = start + size;
L
Linus Torvalds 已提交
1360 1361

	lru_add_drain();
1362
	tlb_gather_mmu(&tlb, mm, start, end);
1363
	update_hiwater_rss(mm);
1364 1365
	mmu_notifier_invalidate_range_start(mm, start, end);
	for ( ; vma && vma->vm_start < end; vma = vma->vm_next)
1366
		unmap_single_vma(&tlb, vma, start, end, details);
1367 1368
	mmu_notifier_invalidate_range_end(mm, start, end);
	tlb_finish_mmu(&tlb, start, end);
L
Linus Torvalds 已提交
1369 1370
}

1371 1372 1373 1374 1375
/**
 * 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
1376
 * @details: details of shared cache invalidation
1377 1378
 *
 * The range must fit into one VMA.
L
Linus Torvalds 已提交
1379
 */
1380
static void zap_page_range_single(struct vm_area_struct *vma, unsigned long address,
L
Linus Torvalds 已提交
1381 1382 1383
		unsigned long size, struct zap_details *details)
{
	struct mm_struct *mm = vma->vm_mm;
P
Peter Zijlstra 已提交
1384
	struct mmu_gather tlb;
L
Linus Torvalds 已提交
1385 1386 1387
	unsigned long end = address + size;

	lru_add_drain();
1388
	tlb_gather_mmu(&tlb, mm, address, end);
1389
	update_hiwater_rss(mm);
1390
	mmu_notifier_invalidate_range_start(mm, address, end);
1391
	unmap_single_vma(&tlb, vma, address, end, details);
1392
	mmu_notifier_invalidate_range_end(mm, address, end);
P
Peter Zijlstra 已提交
1393
	tlb_finish_mmu(&tlb, address, end);
L
Linus Torvalds 已提交
1394 1395
}

1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413
/**
 * 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;
1414
	zap_page_range_single(vma, address, size, NULL);
1415 1416 1417 1418
	return 0;
}
EXPORT_SYMBOL_GPL(zap_vma_ptes);

1419
pte_t *__get_locked_pte(struct mm_struct *mm, unsigned long addr,
H
Harvey Harrison 已提交
1420
			spinlock_t **ptl)
1421 1422 1423 1424
{
	pgd_t * pgd = pgd_offset(mm, addr);
	pud_t * pud = pud_alloc(mm, pgd, addr);
	if (pud) {
1425
		pmd_t * pmd = pmd_alloc(mm, pud, addr);
1426 1427
		if (pmd) {
			VM_BUG_ON(pmd_trans_huge(*pmd));
1428
			return pte_alloc_map_lock(mm, pmd, addr, ptl);
1429
		}
1430 1431 1432 1433
	}
	return NULL;
}

1434 1435 1436 1437 1438 1439 1440
/*
 * 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.
 */
N
Nick Piggin 已提交
1441 1442
static int insert_page(struct vm_area_struct *vma, unsigned long addr,
			struct page *page, pgprot_t prot)
1443
{
N
Nick Piggin 已提交
1444
	struct mm_struct *mm = vma->vm_mm;
1445
	int retval;
1446
	pte_t *pte;
1447 1448
	spinlock_t *ptl;

1449
	retval = -EINVAL;
1450
	if (PageAnon(page))
1451
		goto out;
1452 1453
	retval = -ENOMEM;
	flush_dcache_page(page);
1454
	pte = get_locked_pte(mm, addr, &ptl);
1455
	if (!pte)
1456
		goto out;
1457 1458 1459 1460 1461 1462
	retval = -EBUSY;
	if (!pte_none(*pte))
		goto out_unlock;

	/* Ok, finally just insert the thing.. */
	get_page(page);
1463
	inc_mm_counter_fast(mm, MM_FILEPAGES);
1464 1465 1466 1467
	page_add_file_rmap(page);
	set_pte_at(mm, addr, pte, mk_pte(page, prot));

	retval = 0;
1468 1469
	pte_unmap_unlock(pte, ptl);
	return retval;
1470 1471 1472 1473 1474 1475
out_unlock:
	pte_unmap_unlock(pte, ptl);
out:
	return retval;
}

1476 1477 1478 1479 1480 1481
/**
 * 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
 *
1482 1483 1484 1485 1486 1487
 * 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
N
Nick Piggin 已提交
1488
 * (see split_page()).
1489 1490 1491 1492 1493 1494 1495 1496
 *
 * 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.
1497 1498 1499 1500 1501
 *
 * 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.
1502
 */
N
Nick Piggin 已提交
1503 1504
int vm_insert_page(struct vm_area_struct *vma, unsigned long addr,
			struct page *page)
1505 1506 1507 1508 1509
{
	if (addr < vma->vm_start || addr >= vma->vm_end)
		return -EFAULT;
	if (!page_count(page))
		return -EINVAL;
1510 1511 1512 1513 1514
	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;
	}
N
Nick Piggin 已提交
1515
	return insert_page(vma, addr, page, vma->vm_page_prot);
1516
}
1517
EXPORT_SYMBOL(vm_insert_page);
1518

N
Nick Piggin 已提交
1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537
static int insert_pfn(struct vm_area_struct *vma, unsigned long addr,
			unsigned long pfn, pgprot_t prot)
{
	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.. */
	entry = pte_mkspecial(pfn_pte(pfn, prot));
	set_pte_at(mm, addr, pte, entry);
1538
	update_mmu_cache(vma, addr, pte); /* XXX: why not for insert_page? */
N
Nick Piggin 已提交
1539 1540 1541 1542 1543 1544 1545 1546

	retval = 0;
out_unlock:
	pte_unmap_unlock(pte, ptl);
out:
	return retval;
}

N
Nick Piggin 已提交
1547 1548 1549 1550 1551 1552
/**
 * 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
 *
1553
 * Similar to vm_insert_page, this allows drivers to insert individual pages
N
Nick Piggin 已提交
1554 1555 1556 1557
 * 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.
N
Nick Piggin 已提交
1558 1559 1560 1561 1562
 *
 * 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.
N
Nick Piggin 已提交
1563 1564
 */
int vm_insert_pfn(struct vm_area_struct *vma, unsigned long addr,
N
Nick Piggin 已提交
1565
			unsigned long pfn)
N
Nick Piggin 已提交
1566
{
1567
	int ret;
1568
	pgprot_t pgprot = vma->vm_page_prot;
N
Nick Piggin 已提交
1569 1570 1571 1572 1573 1574
	/*
	 * 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.
	 */
J
Jared Hulbert 已提交
1575 1576 1577 1578 1579
	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));
N
Nick Piggin 已提交
1580

N
Nick Piggin 已提交
1581 1582
	if (addr < vma->vm_start || addr >= vma->vm_end)
		return -EFAULT;
1583
	if (track_pfn_insert(vma, &pgprot, pfn))
1584 1585
		return -EINVAL;

1586
	ret = insert_pfn(vma, addr, pfn, pgprot);
1587 1588

	return ret;
N
Nick Piggin 已提交
1589 1590
}
EXPORT_SYMBOL(vm_insert_pfn);
N
Nick Piggin 已提交
1591

N
Nick Piggin 已提交
1592 1593 1594 1595
int vm_insert_mixed(struct vm_area_struct *vma, unsigned long addr,
			unsigned long pfn)
{
	BUG_ON(!(vma->vm_flags & VM_MIXEDMAP));
N
Nick Piggin 已提交
1596

N
Nick Piggin 已提交
1597 1598
	if (addr < vma->vm_start || addr >= vma->vm_end)
		return -EFAULT;
N
Nick Piggin 已提交
1599

N
Nick Piggin 已提交
1600 1601 1602 1603
	/*
	 * 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
H
Hugh Dickins 已提交
1604 1605
	 * 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.
N
Nick Piggin 已提交
1606 1607 1608 1609 1610 1611 1612 1613
	 */
	if (!HAVE_PTE_SPECIAL && pfn_valid(pfn)) {
		struct page *page;

		page = pfn_to_page(pfn);
		return insert_page(vma, addr, page, vma->vm_page_prot);
	}
	return insert_pfn(vma, addr, pfn, vma->vm_page_prot);
N
Nick Piggin 已提交
1614
}
N
Nick Piggin 已提交
1615
EXPORT_SYMBOL(vm_insert_mixed);
N
Nick Piggin 已提交
1616

L
Linus Torvalds 已提交
1617 1618 1619 1620 1621 1622 1623 1624 1625 1626
/*
 * 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;
H
Hugh Dickins 已提交
1627
	spinlock_t *ptl;
L
Linus Torvalds 已提交
1628

H
Hugh Dickins 已提交
1629
	pte = pte_alloc_map_lock(mm, pmd, addr, &ptl);
L
Linus Torvalds 已提交
1630 1631
	if (!pte)
		return -ENOMEM;
1632
	arch_enter_lazy_mmu_mode();
L
Linus Torvalds 已提交
1633 1634
	do {
		BUG_ON(!pte_none(*pte));
N
Nick Piggin 已提交
1635
		set_pte_at(mm, addr, pte, pte_mkspecial(pfn_pte(pfn, prot)));
L
Linus Torvalds 已提交
1636 1637
		pfn++;
	} while (pte++, addr += PAGE_SIZE, addr != end);
1638
	arch_leave_lazy_mmu_mode();
H
Hugh Dickins 已提交
1639
	pte_unmap_unlock(pte - 1, ptl);
L
Linus Torvalds 已提交
1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653
	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;
1654
	VM_BUG_ON(pmd_trans_huge(*pmd));
L
Linus Torvalds 已提交
1655 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;
}

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.
 */
L
Linus Torvalds 已提交
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;
1699
	unsigned long end = addr + PAGE_ALIGN(size);
L
Linus Torvalds 已提交
1700 1701 1702 1703 1704 1705 1706 1707
	struct mm_struct *mm = vma->vm_mm;
	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).
1708 1709 1710
	 *   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.
1711 1712 1713 1714
	 *   VM_DONTEXPAND
	 *      Disable vma merging and expanding with mremap().
	 *   VM_DONTDUMP
	 *      Omit vma from core dump, even when VM_IO turned off.
L
Linus Torvalds 已提交
1715 1716 1717 1718
	 *
	 * 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".
1719
	 * See vm_normal_page() for details.
L
Linus Torvalds 已提交
1720
	 */
1721 1722 1723
	if (is_cow_mapping(vma->vm_flags)) {
		if (addr != vma->vm_start || end != vma->vm_end)
			return -EINVAL;
L
Linus Torvalds 已提交
1724
		vma->vm_pgoff = pfn;
1725 1726 1727 1728
	}

	err = track_pfn_remap(vma, &prot, pfn, addr, PAGE_ALIGN(size));
	if (err)
1729
		return -EINVAL;
L
Linus Torvalds 已提交
1730

1731
	vma->vm_flags |= VM_IO | VM_PFNMAP | VM_DONTEXPAND | VM_DONTDUMP;
L
Linus Torvalds 已提交
1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743

	BUG_ON(addr >= end);
	pfn -= addr >> PAGE_SHIFT;
	pgd = pgd_offset(mm, addr);
	flush_cache_range(vma, addr, end);
	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);
1744 1745

	if (err)
1746
		untrack_pfn(vma, pfn, PAGE_ALIGN(size));
1747

L
Linus Torvalds 已提交
1748 1749 1750 1751
	return err;
}
EXPORT_SYMBOL(remap_pfn_range);

1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798
/**
 * 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);

1799 1800 1801 1802 1803 1804
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;
1805
	pgtable_t token;
1806
	spinlock_t *uninitialized_var(ptl);
1807 1808 1809 1810 1811 1812 1813 1814 1815

	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));

1816 1817
	arch_enter_lazy_mmu_mode();

1818
	token = pmd_pgtable(*pmd);
1819 1820

	do {
1821
		err = fn(pte++, token, addr, data);
1822 1823
		if (err)
			break;
1824
	} while (addr += PAGE_SIZE, addr != end);
1825

1826 1827
	arch_leave_lazy_mmu_mode();

1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840
	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;

A
Andi Kleen 已提交
1841 1842
	BUG_ON(pud_huge(*pud));

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 1877 1878 1879 1880 1881 1882 1883
	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;
1884
	unsigned long end = addr + size;
1885 1886 1887 1888 1889 1890 1891 1892 1893 1894
	int err;

	BUG_ON(addr >= end);
	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);
1895

1896 1897 1898 1899
	return err;
}
EXPORT_SYMBOL_GPL(apply_to_page_range);

1900
/*
1901 1902 1903 1904 1905
 * 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;
1906
 * and do_anonymous_page can safely check later on).
1907
 */
H
Hugh Dickins 已提交
1908
static inline int pte_unmap_same(struct mm_struct *mm, pmd_t *pmd,
1909 1910 1911 1912 1913
				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)) {
H
Hugh Dickins 已提交
1914 1915
		spinlock_t *ptl = pte_lockptr(mm, pmd);
		spin_lock(ptl);
1916
		same = pte_same(*page_table, orig_pte);
H
Hugh Dickins 已提交
1917
		spin_unlock(ptl);
1918 1919 1920 1921 1922 1923
	}
#endif
	pte_unmap(page_table);
	return same;
}

1924
static inline void cow_user_page(struct page *dst, struct page *src, unsigned long va, struct vm_area_struct *vma)
1925
{
1926 1927
	debug_dma_assert_idle(src);

1928 1929 1930 1931 1932 1933 1934
	/*
	 * 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)) {
1935
		void *kaddr = kmap_atomic(dst);
L
Linus Torvalds 已提交
1936 1937 1938 1939 1940 1941 1942 1943 1944
		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))
1945
			clear_page(kaddr);
1946
		kunmap_atomic(kaddr);
1947
		flush_dcache_page(dst);
N
Nick Piggin 已提交
1948 1949
	} else
		copy_user_highpage(dst, src, va, vma);
1950 1951
}

1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967
/*
 * 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;
	vmf.page = page;
1968
	vmf.cow_page = NULL;
1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984

	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;
}

L
Linus Torvalds 已提交
1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998
/*
 * 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.
 *
 * 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.
 *
1999 2000 2001
 * 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.
L
Linus Torvalds 已提交
2002
 */
2003 2004
static int do_wp_page(struct mm_struct *mm, struct vm_area_struct *vma,
		unsigned long address, pte_t *page_table, pmd_t *pmd,
2005
		spinlock_t *ptl, pte_t orig_pte)
2006
	__releases(ptl)
L
Linus Torvalds 已提交
2007
{
2008
	struct page *old_page, *new_page = NULL;
L
Linus Torvalds 已提交
2009
	pte_t entry;
2010
	int ret = 0;
2011
	int page_mkwrite = 0;
2012
	bool dirty_shared = false;
2013 2014
	unsigned long mmun_start = 0;	/* For mmu_notifiers */
	unsigned long mmun_end = 0;	/* For mmu_notifiers */
2015
	struct mem_cgroup *memcg;
L
Linus Torvalds 已提交
2016

2017
	old_page = vm_normal_page(vma, address, orig_pte);
2018 2019
	if (!old_page) {
		/*
2020 2021
		 * VM_MIXEDMAP !pfn_valid() case, or VM_SOFTDIRTY clear on a
		 * VM_PFNMAP VMA.
2022 2023 2024 2025 2026 2027 2028 2029
		 *
		 * We should not cow pages in a shared writeable mapping.
		 * Just mark the pages writable as we can't do any dirty
		 * accounting on raw pfn maps.
		 */
		if ((vma->vm_flags & (VM_WRITE|VM_SHARED)) ==
				     (VM_WRITE|VM_SHARED))
			goto reuse;
2030
		goto gotten;
2031
	}
L
Linus Torvalds 已提交
2032

2033
	/*
P
Peter Zijlstra 已提交
2034 2035
	 * Take out anonymous pages first, anonymous shared vmas are
	 * not dirty accountable.
2036
	 */
H
Hugh Dickins 已提交
2037
	if (PageAnon(old_page) && !PageKsm(old_page)) {
2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048
		if (!trylock_page(old_page)) {
			page_cache_get(old_page);
			pte_unmap_unlock(page_table, ptl);
			lock_page(old_page);
			page_table = pte_offset_map_lock(mm, pmd, address,
							 &ptl);
			if (!pte_same(*page_table, orig_pte)) {
				unlock_page(old_page);
				goto unlock;
			}
			page_cache_release(old_page);
P
Peter Zijlstra 已提交
2049
		}
2050
		if (reuse_swap_page(old_page)) {
2051 2052 2053 2054 2055 2056
			/*
			 * 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.
			 */
			page_move_anon_rmap(old_page, vma, address);
2057 2058 2059
			unlock_page(old_page);
			goto reuse;
		}
2060
		unlock_page(old_page);
P
Peter Zijlstra 已提交
2061
	} else if (unlikely((vma->vm_flags & (VM_WRITE|VM_SHARED)) ==
2062
					(VM_WRITE|VM_SHARED))) {
2063
		page_cache_get(old_page);
P
Peter Zijlstra 已提交
2064 2065 2066 2067 2068
		/*
		 * Only catch write-faults on shared writable pages,
		 * read-only shared pages can get COWed by
		 * get_user_pages(.write=1, .force=1).
		 */
2069
		if (vma->vm_ops && vma->vm_ops->page_mkwrite) {
2070
			int tmp;
2071

2072
			pte_unmap_unlock(page_table, ptl);
2073 2074 2075 2076 2077
			tmp = do_page_mkwrite(vma, old_page, address);
			if (unlikely(!tmp || (tmp &
					(VM_FAULT_ERROR | VM_FAULT_NOPAGE)))) {
				page_cache_release(old_page);
				return tmp;
2078
			}
2079 2080 2081 2082 2083 2084 2085 2086
			/*
			 * 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.
			 */
			page_table = pte_offset_map_lock(mm, pmd, address,
							 &ptl);
N
Nick Piggin 已提交
2087 2088
			if (!pte_same(*page_table, orig_pte)) {
				unlock_page(old_page);
2089
				goto unlock;
N
Nick Piggin 已提交
2090
			}
2091
			page_mkwrite = 1;
L
Linus Torvalds 已提交
2092
		}
2093 2094

		dirty_shared = true;
2095

2096
reuse:
2097 2098 2099 2100 2101 2102 2103 2104
		/*
		 * Clear the pages cpupid information as the existing
		 * information potentially belongs to a now completely
		 * unrelated process.
		 */
		if (old_page)
			page_cpupid_xchg_last(old_page, (1 << LAST_CPUPID_SHIFT) - 1);

2105 2106 2107
		flush_cache_page(vma, address, pte_pfn(orig_pte));
		entry = pte_mkyoung(orig_pte);
		entry = maybe_mkwrite(pte_mkdirty(entry), vma);
2108
		if (ptep_set_access_flags(vma, address, page_table, entry,1))
2109
			update_mmu_cache(vma, address, page_table);
2110
		pte_unmap_unlock(page_table, ptl);
2111
		ret |= VM_FAULT_WRITE;
2112

2113
		if (dirty_shared) {
2114 2115 2116
			struct address_space *mapping;
			int dirtied;

2117 2118
			if (!page_mkwrite)
				lock_page(old_page);
2119

2120 2121 2122 2123 2124
			dirtied = set_page_dirty(old_page);
			VM_BUG_ON_PAGE(PageAnon(old_page), old_page);
			mapping = old_page->mapping;
			unlock_page(old_page);
			page_cache_release(old_page);
2125

2126
			if ((dirtied || page_mkwrite) && mapping) {
2127 2128 2129 2130 2131 2132 2133
				/*
				 * Some device drivers do not set page.mapping
				 * but still dirty their pages
				 */
				balance_dirty_pages_ratelimited(mapping);
			}

2134
			if (!page_mkwrite)
2135
				file_update_time(vma->vm_file);
2136 2137 2138
		}

		return ret;
L
Linus Torvalds 已提交
2139 2140 2141 2142 2143
	}

	/*
	 * Ok, we need to copy. Oh, well..
	 */
N
Nick Piggin 已提交
2144
	page_cache_get(old_page);
H
Hugh Dickins 已提交
2145
gotten:
2146
	pte_unmap_unlock(page_table, ptl);
L
Linus Torvalds 已提交
2147 2148

	if (unlikely(anon_vma_prepare(vma)))
2149
		goto oom;
H
Hugh Dickins 已提交
2150

H
Hugh Dickins 已提交
2151
	if (is_zero_pfn(pte_pfn(orig_pte))) {
H
Hugh Dickins 已提交
2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162
		new_page = alloc_zeroed_user_highpage_movable(vma, address);
		if (!new_page)
			goto oom;
	} else {
		new_page = alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma, address);
		if (!new_page)
			goto oom;
		cow_user_page(new_page, old_page, address, vma);
	}
	__SetPageUptodate(new_page);

2163
	if (mem_cgroup_try_charge(new_page, mm, GFP_KERNEL, &memcg))
2164 2165
		goto oom_free_new;

2166
	mmun_start  = address & PAGE_MASK;
2167
	mmun_end    = mmun_start + PAGE_SIZE;
2168 2169
	mmu_notifier_invalidate_range_start(mm, mmun_start, mmun_end);

L
Linus Torvalds 已提交
2170 2171 2172
	/*
	 * Re-check the pte - we dropped the lock
	 */
2173
	page_table = pte_offset_map_lock(mm, pmd, address, &ptl);
2174
	if (likely(pte_same(*page_table, orig_pte))) {
H
Hugh Dickins 已提交
2175 2176
		if (old_page) {
			if (!PageAnon(old_page)) {
2177 2178
				dec_mm_counter_fast(mm, MM_FILEPAGES);
				inc_mm_counter_fast(mm, MM_ANONPAGES);
H
Hugh Dickins 已提交
2179 2180
			}
		} else
2181
			inc_mm_counter_fast(mm, MM_ANONPAGES);
2182
		flush_cache_page(vma, address, pte_pfn(orig_pte));
2183 2184
		entry = mk_pte(new_page, vma->vm_page_prot);
		entry = maybe_mkwrite(pte_mkdirty(entry), vma);
2185 2186 2187 2188 2189 2190
		/*
		 * 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.
		 */
2191
		ptep_clear_flush_notify(vma, address, page_table);
N
Nick Piggin 已提交
2192
		page_add_new_anon_rmap(new_page, vma, address);
2193 2194
		mem_cgroup_commit_charge(new_page, memcg, false);
		lru_cache_add_active_or_unevictable(new_page, vma);
2195 2196 2197 2198 2199 2200
		/*
		 * 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.
		 */
		set_pte_at_notify(mm, address, page_table, entry);
2201
		update_mmu_cache(vma, address, page_table);
N
Nick Piggin 已提交
2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224
		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.
			 */
2225
			page_remove_rmap(old_page);
N
Nick Piggin 已提交
2226 2227
		}

L
Linus Torvalds 已提交
2228 2229
		/* Free the old page.. */
		new_page = old_page;
N
Nick Piggin 已提交
2230
		ret |= VM_FAULT_WRITE;
2231
	} else
2232
		mem_cgroup_cancel_charge(new_page, memcg);
2233

2234 2235
	if (new_page)
		page_cache_release(new_page);
2236
unlock:
2237
	pte_unmap_unlock(page_table, ptl);
2238
	if (mmun_end > mmun_start)
2239
		mmu_notifier_invalidate_range_end(mm, mmun_start, mmun_end);
2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251
	if (old_page) {
		/*
		 * Don't let another task, with possibly unlocked vma,
		 * keep the mlocked page.
		 */
		if ((ret & VM_FAULT_WRITE) && (vma->vm_flags & VM_LOCKED)) {
			lock_page(old_page);	/* LRU manipulation */
			munlock_vma_page(old_page);
			unlock_page(old_page);
		}
		page_cache_release(old_page);
	}
N
Nick Piggin 已提交
2252
	return ret;
2253
oom_free_new:
2254
	page_cache_release(new_page);
2255
oom:
2256
	if (old_page)
H
Hugh Dickins 已提交
2257
		page_cache_release(old_page);
L
Linus Torvalds 已提交
2258 2259 2260
	return VM_FAULT_OOM;
}

2261
static void unmap_mapping_range_vma(struct vm_area_struct *vma,
L
Linus Torvalds 已提交
2262 2263 2264
		unsigned long start_addr, unsigned long end_addr,
		struct zap_details *details)
{
2265
	zap_page_range_single(vma, start_addr, end_addr - start_addr, details);
L
Linus Torvalds 已提交
2266 2267
}

2268
static inline void unmap_mapping_range_tree(struct rb_root *root,
L
Linus Torvalds 已提交
2269 2270 2271 2272 2273
					    struct zap_details *details)
{
	struct vm_area_struct *vma;
	pgoff_t vba, vea, zba, zea;

2274
	vma_interval_tree_foreach(vma, root,
L
Linus Torvalds 已提交
2275 2276 2277
			details->first_index, details->last_index) {

		vba = vma->vm_pgoff;
2278
		vea = vba + vma_pages(vma) - 1;
L
Linus Torvalds 已提交
2279 2280 2281 2282 2283 2284 2285 2286
		/* Assume for now that PAGE_CACHE_SHIFT == PAGE_SHIFT */
		zba = details->first_index;
		if (zba < vba)
			zba = vba;
		zea = details->last_index;
		if (zea > vea)
			zea = vea;

2287
		unmap_mapping_range_vma(vma,
L
Linus Torvalds 已提交
2288 2289
			((zba - vba) << PAGE_SHIFT) + vma->vm_start,
			((zea - vba + 1) << PAGE_SHIFT) + vma->vm_start,
2290
				details);
L
Linus Torvalds 已提交
2291 2292 2293 2294
	}
}

/**
2295 2296 2297 2298
 * unmap_mapping_range - unmap the portion of all mmaps in the specified
 * address_space corresponding to the specified page range in the underlying
 * file.
 *
M
Martin Waitz 已提交
2299
 * @mapping: the address space containing mmaps to be unmapped.
L
Linus Torvalds 已提交
2300 2301
 * @holebegin: byte in first page to unmap, relative to the start of
 * the underlying file.  This will be rounded down to a PAGE_SIZE
N
npiggin@suse.de 已提交
2302
 * boundary.  Note that this is different from truncate_pagecache(), which
L
Linus Torvalds 已提交
2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332
 * 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)
{
	struct zap_details details;
	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;
	details.first_index = hba;
	details.last_index = hba + hlen - 1;
	if (details.last_index < details.first_index)
		details.last_index = ULONG_MAX;


2333
	/* DAX uses i_mmap_lock to serialise file truncate vs page fault */
2334
	i_mmap_lock_write(mapping);
2335
	if (unlikely(!RB_EMPTY_ROOT(&mapping->i_mmap)))
L
Linus Torvalds 已提交
2336
		unmap_mapping_range_tree(&mapping->i_mmap, &details);
2337
	i_mmap_unlock_write(mapping);
L
Linus Torvalds 已提交
2338 2339 2340 2341
}
EXPORT_SYMBOL(unmap_mapping_range);

/*
2342 2343
 * We enter with non-exclusive mmap_sem (to exclude vma changes,
 * but allow concurrent faults), and pte mapped but not yet locked.
2344 2345 2346 2347
 * We return with pte unmapped and unlocked.
 *
 * We return with the mmap_sem locked or unlocked in the same cases
 * as does filemap_fault().
L
Linus Torvalds 已提交
2348
 */
2349 2350
static int do_swap_page(struct mm_struct *mm, struct vm_area_struct *vma,
		unsigned long address, pte_t *page_table, pmd_t *pmd,
2351
		unsigned int flags, pte_t orig_pte)
L
Linus Torvalds 已提交
2352
{
2353
	spinlock_t *ptl;
2354
	struct page *page, *swapcache;
2355
	struct mem_cgroup *memcg;
2356
	swp_entry_t entry;
L
Linus Torvalds 已提交
2357
	pte_t pte;
2358
	int locked;
2359
	int exclusive = 0;
N
Nick Piggin 已提交
2360
	int ret = 0;
L
Linus Torvalds 已提交
2361

H
Hugh Dickins 已提交
2362
	if (!pte_unmap_same(mm, pmd, page_table, orig_pte))
2363
		goto out;
2364 2365

	entry = pte_to_swp_entry(orig_pte);
2366 2367 2368 2369 2370 2371 2372
	if (unlikely(non_swap_entry(entry))) {
		if (is_migration_entry(entry)) {
			migration_entry_wait(mm, pmd, address);
		} else if (is_hwpoison_entry(entry)) {
			ret = VM_FAULT_HWPOISON;
		} else {
			print_bad_pte(vma, address, orig_pte, NULL);
H
Hugh Dickins 已提交
2373
			ret = VM_FAULT_SIGBUS;
2374
		}
2375 2376
		goto out;
	}
2377
	delayacct_set_flag(DELAYACCT_PF_SWAPIN);
L
Linus Torvalds 已提交
2378 2379
	page = lookup_swap_cache(entry);
	if (!page) {
2380 2381
		page = swapin_readahead(entry,
					GFP_HIGHUSER_MOVABLE, vma, address);
L
Linus Torvalds 已提交
2382 2383
		if (!page) {
			/*
2384 2385
			 * Back out if somebody else faulted in this pte
			 * while we released the pte lock.
L
Linus Torvalds 已提交
2386
			 */
2387
			page_table = pte_offset_map_lock(mm, pmd, address, &ptl);
L
Linus Torvalds 已提交
2388 2389
			if (likely(pte_same(*page_table, orig_pte)))
				ret = VM_FAULT_OOM;
2390
			delayacct_clear_flag(DELAYACCT_PF_SWAPIN);
2391
			goto unlock;
L
Linus Torvalds 已提交
2392 2393 2394 2395
		}

		/* Had to read the page from swap area: Major fault */
		ret = VM_FAULT_MAJOR;
2396
		count_vm_event(PGMAJFAULT);
2397
		mem_cgroup_count_vm_event(mm, PGMAJFAULT);
2398
	} else if (PageHWPoison(page)) {
2399 2400 2401 2402
		/*
		 * hwpoisoned dirty swapcache pages are kept for killing
		 * owner processes (which may be unknown at hwpoison time)
		 */
2403 2404
		ret = VM_FAULT_HWPOISON;
		delayacct_clear_flag(DELAYACCT_PF_SWAPIN);
2405
		swapcache = page;
2406
		goto out_release;
L
Linus Torvalds 已提交
2407 2408
	}

2409
	swapcache = page;
2410
	locked = lock_page_or_retry(page, mm, flags);
R
Rik van Riel 已提交
2411

2412
	delayacct_clear_flag(DELAYACCT_PF_SWAPIN);
2413 2414 2415 2416
	if (!locked) {
		ret |= VM_FAULT_RETRY;
		goto out_release;
	}
2417

A
Andrea Arcangeli 已提交
2418
	/*
2419 2420 2421 2422
	 * 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.
A
Andrea Arcangeli 已提交
2423
	 */
2424
	if (unlikely(!PageSwapCache(page) || page_private(page) != entry.val))
A
Andrea Arcangeli 已提交
2425 2426
		goto out_page;

2427 2428 2429 2430 2431
	page = ksm_might_need_to_copy(page, vma, address);
	if (unlikely(!page)) {
		ret = VM_FAULT_OOM;
		page = swapcache;
		goto out_page;
H
Hugh Dickins 已提交
2432 2433
	}

2434
	if (mem_cgroup_try_charge(page, mm, GFP_KERNEL, &memcg)) {
2435
		ret = VM_FAULT_OOM;
2436
		goto out_page;
2437 2438
	}

L
Linus Torvalds 已提交
2439
	/*
2440
	 * Back out if somebody else already faulted in this pte.
L
Linus Torvalds 已提交
2441
	 */
2442
	page_table = pte_offset_map_lock(mm, pmd, address, &ptl);
H
Hugh Dickins 已提交
2443
	if (unlikely(!pte_same(*page_table, orig_pte)))
2444 2445 2446 2447 2448
		goto out_nomap;

	if (unlikely(!PageUptodate(page))) {
		ret = VM_FAULT_SIGBUS;
		goto out_nomap;
L
Linus Torvalds 已提交
2449 2450
	}

2451 2452 2453 2454 2455 2456 2457 2458 2459
	/*
	 * 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.
	 */
L
Linus Torvalds 已提交
2460

2461
	inc_mm_counter_fast(mm, MM_ANONPAGES);
K
KAMEZAWA Hiroyuki 已提交
2462
	dec_mm_counter_fast(mm, MM_SWAPENTS);
L
Linus Torvalds 已提交
2463
	pte = mk_pte(page, vma->vm_page_prot);
2464
	if ((flags & FAULT_FLAG_WRITE) && reuse_swap_page(page)) {
L
Linus Torvalds 已提交
2465
		pte = maybe_mkwrite(pte_mkdirty(pte), vma);
2466
		flags &= ~FAULT_FLAG_WRITE;
2467
		ret |= VM_FAULT_WRITE;
2468
		exclusive = 1;
L
Linus Torvalds 已提交
2469 2470
	}
	flush_icache_page(vma, page);
2471 2472
	if (pte_swp_soft_dirty(orig_pte))
		pte = pte_mksoft_dirty(pte);
L
Linus Torvalds 已提交
2473
	set_pte_at(mm, address, page_table, pte);
2474
	if (page == swapcache) {
2475
		do_page_add_anon_rmap(page, vma, address, exclusive);
2476 2477
		mem_cgroup_commit_charge(page, memcg, true);
	} else { /* ksm created a completely new copy */
2478
		page_add_new_anon_rmap(page, vma, address);
2479 2480 2481
		mem_cgroup_commit_charge(page, memcg, false);
		lru_cache_add_active_or_unevictable(page, vma);
	}
L
Linus Torvalds 已提交
2482

2483
	swap_free(entry);
N
Nick Piggin 已提交
2484
	if (vm_swap_full() || (vma->vm_flags & VM_LOCKED) || PageMlocked(page))
2485
		try_to_free_swap(page);
2486
	unlock_page(page);
2487
	if (page != swapcache) {
A
Andrea Arcangeli 已提交
2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498
		/*
		 * 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);
		page_cache_release(swapcache);
	}
2499

2500
	if (flags & FAULT_FLAG_WRITE) {
2501 2502 2503
		ret |= do_wp_page(mm, vma, address, page_table, pmd, ptl, pte);
		if (ret & VM_FAULT_ERROR)
			ret &= VM_FAULT_ERROR;
L
Linus Torvalds 已提交
2504 2505 2506 2507
		goto out;
	}

	/* No need to invalidate - it was non-present before */
2508
	update_mmu_cache(vma, address, page_table);
2509
unlock:
2510
	pte_unmap_unlock(page_table, ptl);
L
Linus Torvalds 已提交
2511 2512
out:
	return ret;
2513
out_nomap:
2514
	mem_cgroup_cancel_charge(page, memcg);
2515
	pte_unmap_unlock(page_table, ptl);
2516
out_page:
2517
	unlock_page(page);
2518
out_release:
2519
	page_cache_release(page);
2520
	if (page != swapcache) {
A
Andrea Arcangeli 已提交
2521 2522 2523
		unlock_page(swapcache);
		page_cache_release(swapcache);
	}
2524
	return ret;
L
Linus Torvalds 已提交
2525 2526
}

2527
/*
2528 2529
 * This is like a special single-page "expand_{down|up}wards()",
 * except we must first make sure that 'address{-|+}PAGE_SIZE'
2530 2531 2532 2533 2534 2535
 * doesn't hit another vma.
 */
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) {
2536 2537 2538 2539 2540 2541 2542 2543 2544 2545
		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;
2546

2547
		return expand_downwards(vma, address - PAGE_SIZE);
2548
	}
2549 2550 2551 2552 2553 2554 2555
	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;

2556
		return expand_upwards(vma, address + PAGE_SIZE);
2557
	}
2558 2559 2560
	return 0;
}

L
Linus Torvalds 已提交
2561
/*
2562 2563 2564
 * 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.
L
Linus Torvalds 已提交
2565
 */
2566 2567
static int do_anonymous_page(struct mm_struct *mm, struct vm_area_struct *vma,
		unsigned long address, pte_t *page_table, pmd_t *pmd,
2568
		unsigned int flags)
L
Linus Torvalds 已提交
2569
{
2570
	struct mem_cgroup *memcg;
2571 2572
	struct page *page;
	spinlock_t *ptl;
L
Linus Torvalds 已提交
2573 2574
	pte_t entry;

2575 2576 2577 2578
	pte_unmap(page_table);

	/* Check if we need to add a guard page to the stack */
	if (check_stack_guard_page(vma, address) < 0)
2579
		return VM_FAULT_SIGSEGV;
2580

2581
	/* Use the zero-page for reads */
2582
	if (!(flags & FAULT_FLAG_WRITE) && !mm_forbids_zeropage(mm)) {
H
Hugh Dickins 已提交
2583 2584
		entry = pte_mkspecial(pfn_pte(my_zero_pfn(address),
						vma->vm_page_prot));
2585
		page_table = pte_offset_map_lock(mm, pmd, address, &ptl);
H
Hugh Dickins 已提交
2586 2587 2588 2589 2590
		if (!pte_none(*page_table))
			goto unlock;
		goto setpte;
	}

N
Nick Piggin 已提交
2591 2592 2593 2594 2595 2596
	/* Allocate our own private page. */
	if (unlikely(anon_vma_prepare(vma)))
		goto oom;
	page = alloc_zeroed_user_highpage_movable(vma, address);
	if (!page)
		goto oom;
2597 2598 2599 2600 2601
	/*
	 * The memory barrier inside __SetPageUptodate makes sure that
	 * preceeding stores to the page contents become visible before
	 * the set_pte_at() write.
	 */
N
Nick Piggin 已提交
2602
	__SetPageUptodate(page);
2603

2604
	if (mem_cgroup_try_charge(page, mm, GFP_KERNEL, &memcg))
2605 2606
		goto oom_free_page;

N
Nick Piggin 已提交
2607
	entry = mk_pte(page, vma->vm_page_prot);
H
Hugh Dickins 已提交
2608 2609
	if (vma->vm_flags & VM_WRITE)
		entry = pte_mkwrite(pte_mkdirty(entry));
L
Linus Torvalds 已提交
2610

N
Nick Piggin 已提交
2611
	page_table = pte_offset_map_lock(mm, pmd, address, &ptl);
2612
	if (!pte_none(*page_table))
N
Nick Piggin 已提交
2613
		goto release;
H
Hugh Dickins 已提交
2614

2615
	inc_mm_counter_fast(mm, MM_ANONPAGES);
N
Nick Piggin 已提交
2616
	page_add_new_anon_rmap(page, vma, address);
2617 2618
	mem_cgroup_commit_charge(page, memcg, false);
	lru_cache_add_active_or_unevictable(page, vma);
H
Hugh Dickins 已提交
2619
setpte:
2620
	set_pte_at(mm, address, page_table, entry);
L
Linus Torvalds 已提交
2621 2622

	/* No need to invalidate - it was non-present before */
2623
	update_mmu_cache(vma, address, page_table);
2624
unlock:
2625
	pte_unmap_unlock(page_table, ptl);
N
Nick Piggin 已提交
2626
	return 0;
2627
release:
2628
	mem_cgroup_cancel_charge(page, memcg);
2629 2630
	page_cache_release(page);
	goto unlock;
2631
oom_free_page:
2632
	page_cache_release(page);
2633
oom:
L
Linus Torvalds 已提交
2634 2635 2636
	return VM_FAULT_OOM;
}

2637 2638 2639 2640 2641
/*
 * 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().
 */
2642
static int __do_fault(struct vm_area_struct *vma, unsigned long address,
2643 2644
			pgoff_t pgoff, unsigned int flags,
			struct page *cow_page, struct page **page)
2645 2646 2647 2648 2649 2650 2651 2652
{
	struct vm_fault vmf;
	int ret;

	vmf.virtual_address = (void __user *)(address & PAGE_MASK);
	vmf.pgoff = pgoff;
	vmf.flags = flags;
	vmf.page = NULL;
2653
	vmf.cow_page = cow_page;
2654 2655 2656 2657

	ret = vma->vm_ops->fault(vma, &vmf);
	if (unlikely(ret & (VM_FAULT_ERROR | VM_FAULT_NOPAGE | VM_FAULT_RETRY)))
		return ret;
2658 2659
	if (!vmf.page)
		goto out;
2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672

	if (unlikely(PageHWPoison(vmf.page))) {
		if (ret & VM_FAULT_LOCKED)
			unlock_page(vmf.page);
		page_cache_release(vmf.page);
		return VM_FAULT_HWPOISON;
	}

	if (unlikely(!(ret & VM_FAULT_LOCKED)))
		lock_page(vmf.page);
	else
		VM_BUG_ON_PAGE(!PageLocked(vmf.page), vmf.page);

2673
 out:
2674 2675 2676 2677
	*page = vmf.page;
	return ret;
}

2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693
/**
 * do_set_pte - setup new PTE entry for given page and add reverse page mapping.
 *
 * @vma: virtual memory area
 * @address: user virtual address
 * @page: page to map
 * @pte: pointer to target page table entry
 * @write: true, if new entry is writable
 * @anon: true, if it's anonymous page
 *
 * Caller must hold page table lock relevant for @pte.
 *
 * Target users are page handler itself and implementations of
 * vm_ops->map_pages.
 */
void do_set_pte(struct vm_area_struct *vma, unsigned long address,
2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714
		struct page *page, pte_t *pte, bool write, bool anon)
{
	pte_t entry;

	flush_icache_page(vma, page);
	entry = mk_pte(page, vma->vm_page_prot);
	if (write)
		entry = maybe_mkwrite(pte_mkdirty(entry), vma);
	if (anon) {
		inc_mm_counter_fast(vma->vm_mm, MM_ANONPAGES);
		page_add_new_anon_rmap(page, vma, address);
	} else {
		inc_mm_counter_fast(vma->vm_mm, MM_FILEPAGES);
		page_add_file_rmap(page);
	}
	set_pte_at(vma->vm_mm, address, pte, entry);

	/* no need to invalidate: a not-present page won't be cached */
	update_mmu_cache(vma, address, pte);
}

2715 2716
static unsigned long fault_around_bytes __read_mostly =
	rounddown_pow_of_two(65536);
2717 2718 2719

#ifdef CONFIG_DEBUG_FS
static int fault_around_bytes_get(void *data, u64 *val)
2720
{
2721
	*val = fault_around_bytes;
2722 2723 2724
	return 0;
}

2725 2726 2727 2728 2729
/*
 * 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.
 */
2730
static int fault_around_bytes_set(void *data, u64 val)
2731
{
2732
	if (val / PAGE_SIZE > PTRS_PER_PTE)
2733
		return -EINVAL;
2734 2735 2736 2737
	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 */
2738 2739
	return 0;
}
2740 2741
DEFINE_SIMPLE_ATTRIBUTE(fault_around_bytes_fops,
		fault_around_bytes_get, fault_around_bytes_set, "%llu\n");
2742 2743 2744 2745 2746

static int __init fault_around_debugfs(void)
{
	void *ret;

2747 2748
	ret = debugfs_create_file("fault_around_bytes", 0644, NULL, NULL,
			&fault_around_bytes_fops);
2749
	if (!ret)
2750
		pr_warn("Failed to create fault_around_bytes in debugfs");
2751 2752 2753 2754
	return 0;
}
late_initcall(fault_around_debugfs);
#endif
2755

2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778
/*
 * 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.
 */
2779 2780 2781
static void do_fault_around(struct vm_area_struct *vma, unsigned long address,
		pte_t *pte, pgoff_t pgoff, unsigned int flags)
{
2782
	unsigned long start_addr, nr_pages, mask;
2783 2784 2785 2786
	pgoff_t max_pgoff;
	struct vm_fault vmf;
	int off;

2787 2788 2789 2790
	nr_pages = ACCESS_ONCE(fault_around_bytes) >> PAGE_SHIFT;
	mask = ~(nr_pages * PAGE_SIZE - 1) & PAGE_MASK;

	start_addr = max(address & mask, vma->vm_start);
2791 2792 2793 2794 2795 2796
	off = ((address - start_addr) >> PAGE_SHIFT) & (PTRS_PER_PTE - 1);
	pte -= off;
	pgoff -= off;

	/*
	 *  max_pgoff is either end of page table or end of vma
2797
	 *  or fault_around_pages() from pgoff, depending what is nearest.
2798 2799 2800 2801
	 */
	max_pgoff = pgoff - ((start_addr >> PAGE_SHIFT) & (PTRS_PER_PTE - 1)) +
		PTRS_PER_PTE - 1;
	max_pgoff = min3(max_pgoff, vma_pages(vma) + vma->vm_pgoff - 1,
2802
			pgoff + nr_pages - 1);
2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821

	/* Check if it makes any sense to call ->map_pages */
	while (!pte_none(*pte)) {
		if (++pgoff > max_pgoff)
			return;
		start_addr += PAGE_SIZE;
		if (start_addr >= vma->vm_end)
			return;
		pte++;
	}

	vmf.virtual_address = (void __user *) start_addr;
	vmf.pte = pte;
	vmf.pgoff = pgoff;
	vmf.max_pgoff = max_pgoff;
	vmf.flags = flags;
	vma->vm_ops->map_pages(vma, &vmf);
}

2822 2823 2824 2825 2826 2827
static int do_read_fault(struct mm_struct *mm, struct vm_area_struct *vma,
		unsigned long address, pmd_t *pmd,
		pgoff_t pgoff, unsigned int flags, pte_t orig_pte)
{
	struct page *fault_page;
	spinlock_t *ptl;
2828
	pte_t *pte;
2829 2830 2831 2832 2833 2834 2835
	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).
	 */
2836
	if (vma->vm_ops->map_pages && fault_around_bytes >> PAGE_SHIFT > 1) {
2837 2838 2839 2840 2841 2842
		pte = pte_offset_map_lock(mm, pmd, address, &ptl);
		do_fault_around(vma, address, pte, pgoff, flags);
		if (!pte_same(*pte, orig_pte))
			goto unlock_out;
		pte_unmap_unlock(pte, ptl);
	}
2843

2844
	ret = __do_fault(vma, address, pgoff, flags, NULL, &fault_page);
2845 2846 2847 2848 2849 2850 2851 2852 2853 2854
	if (unlikely(ret & (VM_FAULT_ERROR | VM_FAULT_NOPAGE | VM_FAULT_RETRY)))
		return ret;

	pte = pte_offset_map_lock(mm, pmd, address, &ptl);
	if (unlikely(!pte_same(*pte, orig_pte))) {
		pte_unmap_unlock(pte, ptl);
		unlock_page(fault_page);
		page_cache_release(fault_page);
		return ret;
	}
2855
	do_set_pte(vma, address, fault_page, pte, false, false);
2856
	unlock_page(fault_page);
2857 2858
unlock_out:
	pte_unmap_unlock(pte, ptl);
2859 2860 2861
	return ret;
}

2862 2863 2864 2865 2866
static int do_cow_fault(struct mm_struct *mm, struct vm_area_struct *vma,
		unsigned long address, pmd_t *pmd,
		pgoff_t pgoff, unsigned int flags, pte_t orig_pte)
{
	struct page *fault_page, *new_page;
2867
	struct mem_cgroup *memcg;
2868
	spinlock_t *ptl;
2869
	pte_t *pte;
2870 2871 2872 2873 2874 2875 2876 2877 2878
	int ret;

	if (unlikely(anon_vma_prepare(vma)))
		return VM_FAULT_OOM;

	new_page = alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma, address);
	if (!new_page)
		return VM_FAULT_OOM;

2879
	if (mem_cgroup_try_charge(new_page, mm, GFP_KERNEL, &memcg)) {
2880 2881 2882 2883
		page_cache_release(new_page);
		return VM_FAULT_OOM;
	}

2884
	ret = __do_fault(vma, address, pgoff, flags, new_page, &fault_page);
2885 2886 2887
	if (unlikely(ret & (VM_FAULT_ERROR | VM_FAULT_NOPAGE | VM_FAULT_RETRY)))
		goto uncharge_out;

2888 2889
	if (fault_page)
		copy_user_highpage(new_page, fault_page, address, vma);
2890 2891 2892 2893 2894
	__SetPageUptodate(new_page);

	pte = pte_offset_map_lock(mm, pmd, address, &ptl);
	if (unlikely(!pte_same(*pte, orig_pte))) {
		pte_unmap_unlock(pte, ptl);
2895 2896 2897 2898 2899 2900 2901 2902 2903 2904
		if (fault_page) {
			unlock_page(fault_page);
			page_cache_release(fault_page);
		} else {
			/*
			 * The fault handler has no page to lock, so it holds
			 * i_mmap_lock for read to protect against truncate.
			 */
			i_mmap_unlock_read(vma->vm_file->f_mapping);
		}
2905 2906
		goto uncharge_out;
	}
2907
	do_set_pte(vma, address, new_page, pte, true, true);
2908 2909
	mem_cgroup_commit_charge(new_page, memcg, false);
	lru_cache_add_active_or_unevictable(new_page, vma);
2910
	pte_unmap_unlock(pte, ptl);
2911 2912 2913 2914 2915 2916 2917 2918 2919 2920
	if (fault_page) {
		unlock_page(fault_page);
		page_cache_release(fault_page);
	} else {
		/*
		 * The fault handler has no page to lock, so it holds
		 * i_mmap_lock for read to protect against truncate.
		 */
		i_mmap_unlock_read(vma->vm_file->f_mapping);
	}
2921 2922
	return ret;
uncharge_out:
2923
	mem_cgroup_cancel_charge(new_page, memcg);
2924 2925 2926 2927
	page_cache_release(new_page);
	return ret;
}

2928
static int do_shared_fault(struct mm_struct *mm, struct vm_area_struct *vma,
2929
		unsigned long address, pmd_t *pmd,
2930
		pgoff_t pgoff, unsigned int flags, pte_t orig_pte)
L
Linus Torvalds 已提交
2931
{
2932 2933
	struct page *fault_page;
	struct address_space *mapping;
2934
	spinlock_t *ptl;
2935
	pte_t *pte;
2936 2937
	int dirtied = 0;
	int ret, tmp;
2938

2939
	ret = __do_fault(vma, address, pgoff, flags, NULL, &fault_page);
2940
	if (unlikely(ret & (VM_FAULT_ERROR | VM_FAULT_NOPAGE | VM_FAULT_RETRY)))
2941
		return ret;
L
Linus Torvalds 已提交
2942 2943

	/*
2944 2945
	 * Check if the backing address space wants to know that the page is
	 * about to become writable
L
Linus Torvalds 已提交
2946
	 */
2947 2948 2949 2950 2951
	if (vma->vm_ops->page_mkwrite) {
		unlock_page(fault_page);
		tmp = do_page_mkwrite(vma, fault_page, address);
		if (unlikely(!tmp ||
				(tmp & (VM_FAULT_ERROR | VM_FAULT_NOPAGE)))) {
2952
			page_cache_release(fault_page);
2953
			return tmp;
2954
		}
2955 2956
	}

2957 2958 2959 2960 2961 2962
	pte = pte_offset_map_lock(mm, pmd, address, &ptl);
	if (unlikely(!pte_same(*pte, orig_pte))) {
		pte_unmap_unlock(pte, ptl);
		unlock_page(fault_page);
		page_cache_release(fault_page);
		return ret;
L
Linus Torvalds 已提交
2963
	}
2964
	do_set_pte(vma, address, fault_page, pte, true, false);
2965
	pte_unmap_unlock(pte, ptl);
N
Nick Piggin 已提交
2966

2967 2968
	if (set_page_dirty(fault_page))
		dirtied = 1;
2969 2970 2971 2972 2973 2974
	/*
	 * 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.
	 */
2975 2976 2977 2978 2979 2980 2981 2982
	mapping = fault_page->mapping;
	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);
2983
	}
2984

2985
	if (!vma->vm_ops->page_mkwrite)
2986
		file_update_time(vma->vm_file);
N
Nick Piggin 已提交
2987

2988
	return ret;
2989
}
2990

2991 2992 2993 2994 2995 2996
/*
 * 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().
 */
2997
static int do_fault(struct mm_struct *mm, struct vm_area_struct *vma,
2998
		unsigned long address, pte_t *page_table, pmd_t *pmd,
2999
		unsigned int flags, pte_t orig_pte)
3000 3001
{
	pgoff_t pgoff = (((address & PAGE_MASK)
3002
			- vma->vm_start) >> PAGE_SHIFT) + vma->vm_pgoff;
3003

3004
	pte_unmap(page_table);
3005 3006 3007
	if (!(flags & FAULT_FLAG_WRITE))
		return do_read_fault(mm, vma, address, pmd, pgoff, flags,
				orig_pte);
3008 3009 3010
	if (!(vma->vm_flags & VM_SHARED))
		return do_cow_fault(mm, vma, address, pmd, pgoff, flags,
				orig_pte);
3011
	return do_shared_fault(mm, vma, address, pmd, pgoff, flags, orig_pte);
3012 3013
}

3014
static int numa_migrate_prep(struct page *page, struct vm_area_struct *vma,
3015 3016
				unsigned long addr, int page_nid,
				int *flags)
3017 3018 3019 3020
{
	get_page(page);

	count_vm_numa_event(NUMA_HINT_FAULTS);
3021
	if (page_nid == numa_node_id()) {
3022
		count_vm_numa_event(NUMA_HINT_FAULTS_LOCAL);
3023 3024
		*flags |= TNF_FAULT_LOCAL;
	}
3025 3026 3027 3028

	return mpol_misplaced(page, vma, addr);
}

3029
static int do_numa_page(struct mm_struct *mm, struct vm_area_struct *vma,
3030 3031
		   unsigned long addr, pte_t pte, pte_t *ptep, pmd_t *pmd)
{
3032
	struct page *page = NULL;
3033
	spinlock_t *ptl;
3034
	int page_nid = -1;
3035
	int last_cpupid;
3036
	int target_nid;
3037
	bool migrated = false;
3038
	int flags = 0;
3039

3040 3041 3042
	/* A PROT_NONE fault should not end up here */
	BUG_ON(!(vma->vm_flags & (VM_READ | VM_EXEC | VM_WRITE)));

3043 3044 3045 3046 3047
	/*
	* 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.
	*
3048 3049 3050
	* 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.
3051 3052 3053
	*/
	ptl = pte_lockptr(mm, pmd);
	spin_lock(ptl);
3054 3055 3056 3057 3058
	if (unlikely(!pte_same(*ptep, pte))) {
		pte_unmap_unlock(ptep, ptl);
		goto out;
	}

3059 3060 3061
	/* Make it present again */
	pte = pte_modify(pte, vma->vm_page_prot);
	pte = pte_mkyoung(pte);
3062 3063 3064 3065 3066 3067 3068 3069 3070
	set_pte_at(mm, addr, ptep, pte);
	update_mmu_cache(vma, addr, ptep);

	page = vm_normal_page(vma, addr, pte);
	if (!page) {
		pte_unmap_unlock(ptep, ptl);
		return 0;
	}

3071 3072 3073 3074 3075 3076 3077 3078
	/*
	 * Avoid grouping on DSO/COW pages in specific and RO pages
	 * in general, RO pages shouldn't hurt as much anyway since
	 * they can be in shared cache state.
	 */
	if (!pte_write(pte))
		flags |= TNF_NO_GROUP;

3079 3080 3081 3082 3083 3084 3085
	/*
	 * 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;

3086
	last_cpupid = page_cpupid_last(page);
3087
	page_nid = page_to_nid(page);
3088
	target_nid = numa_migrate_prep(page, vma, addr, page_nid, &flags);
3089
	pte_unmap_unlock(ptep, ptl);
3090 3091 3092 3093 3094 3095
	if (target_nid == -1) {
		put_page(page);
		goto out;
	}

	/* Migrate to the requested node */
3096
	migrated = migrate_misplaced_page(page, vma, target_nid);
3097
	if (migrated) {
3098
		page_nid = target_nid;
3099 3100
		flags |= TNF_MIGRATED;
	}
3101 3102

out:
3103
	if (page_nid != -1)
3104
		task_numa_fault(last_cpupid, page_nid, 1, flags);
3105 3106 3107
	return 0;
}

L
Linus Torvalds 已提交
3108 3109 3110 3111 3112 3113 3114 3115 3116
/*
 * 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).
 *
H
Hugh Dickins 已提交
3117 3118
 * We enter with non-exclusive mmap_sem (to exclude vma changes,
 * but allow concurrent faults), and pte mapped but not yet locked.
3119 3120 3121 3122
 * We return with pte unmapped and unlocked.
 *
 * The mmap_sem may have been released depending on flags and our
 * return value.  See filemap_fault() and __lock_page_or_retry().
L
Linus Torvalds 已提交
3123
 */
3124
static int handle_pte_fault(struct mm_struct *mm,
3125 3126
		     struct vm_area_struct *vma, unsigned long address,
		     pte_t *pte, pmd_t *pmd, unsigned int flags)
L
Linus Torvalds 已提交
3127 3128
{
	pte_t entry;
3129
	spinlock_t *ptl;
L
Linus Torvalds 已提交
3130

3131 3132 3133 3134 3135 3136 3137 3138 3139 3140
	/*
	 * 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.
	 */
	entry = *pte;
	barrier();
L
Linus Torvalds 已提交
3141
	if (!pte_present(entry)) {
3142
		if (pte_none(entry)) {
J
Jes Sorensen 已提交
3143
			if (vma->vm_ops) {
N
Nick Piggin 已提交
3144
				if (likely(vma->vm_ops->fault))
3145 3146
					return do_fault(mm, vma, address, pte,
							pmd, flags, entry);
J
Jes Sorensen 已提交
3147 3148
			}
			return do_anonymous_page(mm, vma, address,
3149
						 pte, pmd, flags);
3150 3151
		}
		return do_swap_page(mm, vma, address,
3152
					pte, pmd, flags, entry);
L
Linus Torvalds 已提交
3153 3154
	}

3155
	if (pte_protnone(entry))
3156 3157
		return do_numa_page(mm, vma, address, entry, pte, pmd);

H
Hugh Dickins 已提交
3158
	ptl = pte_lockptr(mm, pmd);
3159 3160 3161
	spin_lock(ptl);
	if (unlikely(!pte_same(*pte, entry)))
		goto unlock;
3162
	if (flags & FAULT_FLAG_WRITE) {
L
Linus Torvalds 已提交
3163
		if (!pte_write(entry))
3164 3165
			return do_wp_page(mm, vma, address,
					pte, pmd, ptl, entry);
L
Linus Torvalds 已提交
3166 3167 3168
		entry = pte_mkdirty(entry);
	}
	entry = pte_mkyoung(entry);
3169
	if (ptep_set_access_flags(vma, address, pte, entry, flags & FAULT_FLAG_WRITE)) {
3170
		update_mmu_cache(vma, address, pte);
3171 3172 3173 3174 3175 3176 3177
	} 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.
		 */
3178
		if (flags & FAULT_FLAG_WRITE)
3179
			flush_tlb_fix_spurious_fault(vma, address);
3180
	}
3181 3182
unlock:
	pte_unmap_unlock(pte, ptl);
N
Nick Piggin 已提交
3183
	return 0;
L
Linus Torvalds 已提交
3184 3185 3186 3187
}

/*
 * By the time we get here, we already hold the mm semaphore
3188 3189 3190
 *
 * The mmap_sem may have been released depending on flags and our
 * return value.  See filemap_fault() and __lock_page_or_retry().
L
Linus Torvalds 已提交
3191
 */
3192 3193
static int __handle_mm_fault(struct mm_struct *mm, struct vm_area_struct *vma,
			     unsigned long address, unsigned int flags)
L
Linus Torvalds 已提交
3194 3195 3196 3197 3198 3199
{
	pgd_t *pgd;
	pud_t *pud;
	pmd_t *pmd;
	pte_t *pte;

3200
	if (unlikely(is_vm_hugetlb_page(vma)))
3201
		return hugetlb_fault(mm, vma, address, flags);
L
Linus Torvalds 已提交
3202 3203 3204 3205

	pgd = pgd_offset(mm, address);
	pud = pud_alloc(mm, pgd, address);
	if (!pud)
H
Hugh Dickins 已提交
3206
		return VM_FAULT_OOM;
L
Linus Torvalds 已提交
3207 3208
	pmd = pmd_alloc(mm, pud, address);
	if (!pmd)
H
Hugh Dickins 已提交
3209
		return VM_FAULT_OOM;
3210
	if (pmd_none(*pmd) && transparent_hugepage_enabled(vma)) {
3211
		int ret = VM_FAULT_FALLBACK;
3212
		if (!vma->vm_ops)
3213 3214 3215 3216
			ret = do_huge_pmd_anonymous_page(mm, vma, address,
					pmd, flags);
		if (!(ret & VM_FAULT_FALLBACK))
			return ret;
3217 3218
	} else {
		pmd_t orig_pmd = *pmd;
3219 3220
		int ret;

3221 3222
		barrier();
		if (pmd_trans_huge(orig_pmd)) {
3223 3224
			unsigned int dirty = flags & FAULT_FLAG_WRITE;

3225 3226 3227 3228 3229 3230 3231 3232
			/*
			 * If the pmd is splitting, return and retry the
			 * the fault.  Alternative: wait until the split
			 * is done, and goto retry.
			 */
			if (pmd_trans_splitting(orig_pmd))
				return 0;

3233
			if (pmd_protnone(orig_pmd))
3234
				return do_huge_pmd_numa_page(mm, vma, address,
3235 3236
							     orig_pmd, pmd);

3237
			if (dirty && !pmd_write(orig_pmd)) {
3238 3239
				ret = do_huge_pmd_wp_page(mm, vma, address, pmd,
							  orig_pmd);
3240 3241
				if (!(ret & VM_FAULT_FALLBACK))
					return ret;
3242 3243 3244
			} else {
				huge_pmd_set_accessed(mm, vma, address, pmd,
						      orig_pmd, dirty);
3245
				return 0;
3246
			}
3247 3248 3249 3250 3251 3252 3253 3254
		}
	}

	/*
	 * Use __pte_alloc instead of pte_alloc_map, because we can't
	 * run pte_offset_map on the pmd, if an huge pmd could
	 * materialize from under us from a different thread.
	 */
3255 3256
	if (unlikely(pmd_none(*pmd)) &&
	    unlikely(__pte_alloc(mm, vma, pmd, address)))
H
Hugh Dickins 已提交
3257
		return VM_FAULT_OOM;
3258 3259 3260 3261 3262 3263 3264 3265 3266 3267
	/* if an huge pmd materialized from under us just retry later */
	if (unlikely(pmd_trans_huge(*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().
	 */
	pte = pte_offset_map(pmd, address);
L
Linus Torvalds 已提交
3268

3269
	return handle_pte_fault(mm, vma, address, pte, pmd, flags);
L
Linus Torvalds 已提交
3270 3271
}

3272 3273 3274 3275 3276 3277
/*
 * 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().
 */
3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295
int handle_mm_fault(struct mm_struct *mm, struct vm_area_struct *vma,
		    unsigned long address, unsigned int flags)
{
	int ret;

	__set_current_state(TASK_RUNNING);

	count_vm_event(PGFAULT);
	mem_cgroup_count_vm_event(mm, PGFAULT);

	/* 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)
3296
		mem_cgroup_oom_enable();
3297 3298 3299

	ret = __handle_mm_fault(mm, vma, address, flags);

3300 3301 3302 3303 3304 3305 3306 3307 3308 3309 3310
	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);
	}
3311

3312 3313
	return ret;
}
3314
EXPORT_SYMBOL_GPL(handle_mm_fault);
3315

L
Linus Torvalds 已提交
3316 3317 3318
#ifndef __PAGETABLE_PUD_FOLDED
/*
 * Allocate page upper directory.
H
Hugh Dickins 已提交
3319
 * We've already handled the fast-path in-line.
L
Linus Torvalds 已提交
3320
 */
3321
int __pud_alloc(struct mm_struct *mm, pgd_t *pgd, unsigned long address)
L
Linus Torvalds 已提交
3322
{
H
Hugh Dickins 已提交
3323 3324
	pud_t *new = pud_alloc_one(mm, address);
	if (!new)
3325
		return -ENOMEM;
L
Linus Torvalds 已提交
3326

3327 3328
	smp_wmb(); /* See comment in __pte_alloc */

H
Hugh Dickins 已提交
3329
	spin_lock(&mm->page_table_lock);
3330
	if (pgd_present(*pgd))		/* Another has populated it */
3331
		pud_free(mm, new);
3332 3333
	else
		pgd_populate(mm, pgd, new);
H
Hugh Dickins 已提交
3334
	spin_unlock(&mm->page_table_lock);
3335
	return 0;
L
Linus Torvalds 已提交
3336 3337 3338 3339 3340 3341
}
#endif /* __PAGETABLE_PUD_FOLDED */

#ifndef __PAGETABLE_PMD_FOLDED
/*
 * Allocate page middle directory.
H
Hugh Dickins 已提交
3342
 * We've already handled the fast-path in-line.
L
Linus Torvalds 已提交
3343
 */
3344
int __pmd_alloc(struct mm_struct *mm, pud_t *pud, unsigned long address)
L
Linus Torvalds 已提交
3345
{
H
Hugh Dickins 已提交
3346 3347
	pmd_t *new = pmd_alloc_one(mm, address);
	if (!new)
3348
		return -ENOMEM;
L
Linus Torvalds 已提交
3349

3350 3351
	smp_wmb(); /* See comment in __pte_alloc */

H
Hugh Dickins 已提交
3352
	spin_lock(&mm->page_table_lock);
L
Linus Torvalds 已提交
3353
#ifndef __ARCH_HAS_4LEVEL_HACK
3354 3355
	if (!pud_present(*pud)) {
		mm_inc_nr_pmds(mm);
3356
		pud_populate(mm, pud, new);
3357
	} else	/* Another has populated it */
3358
		pmd_free(mm, new);
3359 3360 3361
#else
	if (!pgd_present(*pud)) {
		mm_inc_nr_pmds(mm);
3362
		pgd_populate(mm, pud, new);
3363 3364
	} else /* Another has populated it */
		pmd_free(mm, new);
L
Linus Torvalds 已提交
3365
#endif /* __ARCH_HAS_4LEVEL_HACK */
H
Hugh Dickins 已提交
3366
	spin_unlock(&mm->page_table_lock);
3367
	return 0;
3368
}
L
Linus Torvalds 已提交
3369 3370
#endif /* __PAGETABLE_PMD_FOLDED */

3371
static int __follow_pte(struct mm_struct *mm, unsigned long address,
J
Johannes Weiner 已提交
3372 3373 3374 3375 3376 3377 3378 3379 3380 3381 3382 3383 3384 3385 3386 3387
		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);
3388
	VM_BUG_ON(pmd_trans_huge(*pmd));
J
Johannes Weiner 已提交
3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408
	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;
}

3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419
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;
}

J
Johannes Weiner 已提交
3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448
/**
 * 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);

3449
#ifdef CONFIG_HAVE_IOREMAP_PROT
3450 3451 3452
int follow_phys(struct vm_area_struct *vma,
		unsigned long address, unsigned int flags,
		unsigned long *prot, resource_size_t *phys)
3453
{
3454
	int ret = -EINVAL;
3455 3456 3457
	pte_t *ptep, pte;
	spinlock_t *ptl;

3458 3459
	if (!(vma->vm_flags & (VM_IO | VM_PFNMAP)))
		goto out;
3460

3461
	if (follow_pte(vma->vm_mm, address, &ptep, &ptl))
3462
		goto out;
3463
	pte = *ptep;
3464

3465 3466 3467 3468
	if ((flags & FOLL_WRITE) && !pte_write(pte))
		goto unlock;

	*prot = pgprot_val(pte_pgprot(pte));
3469
	*phys = (resource_size_t)pte_pfn(pte) << PAGE_SHIFT;
3470

3471
	ret = 0;
3472 3473 3474
unlock:
	pte_unmap_unlock(ptep, ptl);
out:
3475
	return ret;
3476 3477 3478 3479 3480 3481 3482
}

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;
K
KOSAKI Motohiro 已提交
3483
	void __iomem *maddr;
3484 3485
	int offset = addr & (PAGE_SIZE-1);

3486
	if (follow_phys(vma, addr, write, &prot, &phys_addr))
3487 3488
		return -EINVAL;

3489
	maddr = ioremap_prot(phys_addr, PAGE_ALIGN(len + offset), prot);
3490 3491 3492 3493 3494 3495 3496 3497
	if (write)
		memcpy_toio(maddr + offset, buf, len);
	else
		memcpy_fromio(buf, maddr + offset, len);
	iounmap(maddr);

	return len;
}
3498
EXPORT_SYMBOL_GPL(generic_access_phys);
3499 3500
#endif

3501
/*
3502 3503
 * Access another process' address space as given in mm.  If non-NULL, use the
 * given task for page fault accounting.
3504
 */
3505 3506
static int __access_remote_vm(struct task_struct *tsk, struct mm_struct *mm,
		unsigned long addr, void *buf, int len, int write)
3507 3508 3509 3510 3511
{
	struct vm_area_struct *vma;
	void *old_buf = buf;

	down_read(&mm->mmap_sem);
S
Simon Arlott 已提交
3512
	/* ignore errors, just check how much was successfully transferred */
3513 3514 3515
	while (len) {
		int bytes, ret, offset;
		void *maddr;
3516
		struct page *page = NULL;
3517 3518 3519

		ret = get_user_pages(tsk, mm, addr, 1,
				write, 1, &page, &vma);
3520
		if (ret <= 0) {
3521 3522 3523
#ifndef CONFIG_HAVE_IOREMAP_PROT
			break;
#else
3524 3525 3526 3527 3528
			/*
			 * Check if this is a VM_IO | VM_PFNMAP VMA, which
			 * we can access using slightly different code.
			 */
			vma = find_vma(mm, addr);
3529
			if (!vma || vma->vm_start > addr)
3530 3531 3532 3533 3534 3535 3536
				break;
			if (vma->vm_ops && vma->vm_ops->access)
				ret = vma->vm_ops->access(vma, addr, buf,
							  len, write);
			if (ret <= 0)
				break;
			bytes = ret;
3537
#endif
3538
		} else {
3539 3540 3541 3542 3543 3544 3545 3546 3547 3548 3549 3550 3551 3552 3553 3554
			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);
			page_cache_release(page);
3555 3556 3557 3558 3559 3560 3561 3562 3563
		}
		len -= bytes;
		buf += bytes;
		addr += bytes;
	}
	up_read(&mm->mmap_sem);

	return buf - old_buf;
}
3564

S
Stephen Wilson 已提交
3565
/**
3566
 * access_remote_vm - access another process' address space
S
Stephen Wilson 已提交
3567 3568 3569 3570 3571 3572 3573 3574 3575 3576 3577 3578 3579 3580
 * @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
 * @write:	whether the access is a write
 *
 * The caller must hold a reference on @mm.
 */
int access_remote_vm(struct mm_struct *mm, unsigned long addr,
		void *buf, int len, int write)
{
	return __access_remote_vm(NULL, mm, addr, buf, len, write);
}

3581 3582 3583 3584 3585 3586 3587 3588 3589 3590 3591 3592 3593 3594 3595 3596 3597 3598 3599 3600 3601
/*
 * 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,
		void *buf, int len, int write)
{
	struct mm_struct *mm;
	int ret;

	mm = get_task_mm(tsk);
	if (!mm)
		return 0;

	ret = __access_remote_vm(tsk, mm, addr, buf, len, write);
	mmput(mm);

	return ret;
}

3602 3603 3604 3605 3606 3607 3608 3609
/*
 * 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;

3610 3611 3612 3613 3614 3615 3616
	/*
	 * Do not print if we are in atomic
	 * contexts (in exception stacks, etc.):
	 */
	if (preempt_count())
		return;

3617 3618 3619 3620 3621 3622
	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) {
A
Andy Shevchenko 已提交
3623
			char *p;
3624

3625
			p = d_path(&f->f_path, buf, PAGE_SIZE);
3626 3627
			if (IS_ERR(p))
				p = "?";
A
Andy Shevchenko 已提交
3628
			printk("%s%s[%lx+%lx]", prefix, kbasename(p),
3629 3630 3631 3632 3633
					vma->vm_start,
					vma->vm_end - vma->vm_start);
			free_page((unsigned long)buf);
		}
	}
3634
	up_read(&mm->mmap_sem);
3635
}
3636

3637
#if defined(CONFIG_PROVE_LOCKING) || defined(CONFIG_DEBUG_ATOMIC_SLEEP)
3638 3639
void might_fault(void)
{
3640 3641 3642 3643 3644 3645 3646 3647 3648
	/*
	 * 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;

3649 3650 3651 3652 3653
	/*
	 * it would be nicer only to annotate paths which are not under
	 * pagefault_disable, however that requires a larger audit and
	 * providing helpers like get_user_atomic.
	 */
3654 3655 3656 3657 3658 3659
	if (in_atomic())
		return;

	__might_sleep(__FILE__, __LINE__, 0);

	if (current->mm)
3660 3661 3662 3663
		might_lock_read(&current->mm->mmap_sem);
}
EXPORT_SYMBOL(might_fault);
#endif
A
Andrea Arcangeli 已提交
3664 3665 3666 3667 3668 3669 3670 3671 3672 3673 3674 3675 3676 3677 3678 3679 3680 3681 3682 3683 3684 3685 3686 3687 3688 3689 3690 3691 3692 3693 3694 3695 3696 3697 3698 3699 3700 3701 3702 3703 3704 3705 3706 3707 3708 3709 3710 3711 3712 3713 3714 3715 3716 3717 3718 3719 3720 3721 3722 3723 3724 3725 3726 3727 3728 3729 3730 3731 3732 3733 3734

#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 */
3735

3736
#if USE_SPLIT_PTE_PTLOCKS && ALLOC_SPLIT_PTLOCKS
3737 3738 3739 3740 3741 3742 3743 3744 3745

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);
}

3746
bool ptlock_alloc(struct page *page)
3747 3748 3749
{
	spinlock_t *ptl;

3750
	ptl = kmem_cache_alloc(page_ptl_cachep, GFP_KERNEL);
3751 3752
	if (!ptl)
		return false;
3753
	page->ptl = ptl;
3754 3755 3756
	return true;
}

3757
void ptlock_free(struct page *page)
3758
{
3759
	kmem_cache_free(page_ptl_cachep, page->ptl);
3760 3761
}
#endif