memory.c 73.5 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 48 49
/*
 *  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>
#include <linux/rmap.h>
#include <linux/module.h>
50
#include <linux/delayacct.h>
L
Linus Torvalds 已提交
51
#include <linux/init.h>
P
Peter Zijlstra 已提交
52
#include <linux/writeback.h>
L
Linus Torvalds 已提交
53 54 55 56 57 58 59 60 61 62

#include <asm/pgalloc.h>
#include <asm/uaccess.h>
#include <asm/tlb.h>
#include <asm/tlbflush.h>
#include <asm/pgtable.h>

#include <linux/swapops.h>
#include <linux/elf.h>

A
Andy Whitcroft 已提交
63
#ifndef CONFIG_NEED_MULTIPLE_NODES
L
Linus Torvalds 已提交
64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84
/* 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

unsigned long num_physpages;
/*
 * 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(num_physpages);
EXPORT_SYMBOL(high_memory);

85 86 87 88 89
int randomize_va_space __read_mostly = 1;

static int __init disable_randmaps(char *s)
{
	randomize_va_space = 0;
90
	return 1;
91 92 93 94
}
__setup("norandmaps", disable_randmaps);


L
Linus Torvalds 已提交
95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122
/*
 * If a p?d_bad entry is found while walking page tables, report
 * the error, before resetting entry to p?d_none.  Usually (but
 * very seldom) called out from the p?d_none_or_clear_bad macros.
 */

void pgd_clear_bad(pgd_t *pgd)
{
	pgd_ERROR(*pgd);
	pgd_clear(pgd);
}

void pud_clear_bad(pud_t *pud)
{
	pud_ERROR(*pud);
	pud_clear(pud);
}

void pmd_clear_bad(pmd_t *pmd)
{
	pmd_ERROR(*pmd);
	pmd_clear(pmd);
}

/*
 * Note: this doesn't free the actual pages themselves. That
 * has been handled earlier when unmapping all the memory regions.
 */
123
static void free_pte_range(struct mmu_gather *tlb, pmd_t *pmd)
L
Linus Torvalds 已提交
124
{
125 126
	struct page *page = pmd_page(*pmd);
	pmd_clear(pmd);
H
Hugh Dickins 已提交
127
	pte_lock_deinit(page);
128
	pte_free_tlb(tlb, page);
129
	dec_zone_page_state(page, NR_PAGETABLE);
130
	tlb->mm->nr_ptes--;
L
Linus Torvalds 已提交
131 132
}

133 134 135
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
136 137 138
{
	pmd_t *pmd;
	unsigned long next;
139
	unsigned long start;
L
Linus Torvalds 已提交
140

141
	start = addr;
L
Linus Torvalds 已提交
142 143 144 145 146
	pmd = pmd_offset(pud, addr);
	do {
		next = pmd_addr_end(addr, end);
		if (pmd_none_or_clear_bad(pmd))
			continue;
147
		free_pte_range(tlb, pmd);
L
Linus Torvalds 已提交
148 149
	} while (pmd++, addr = next, addr != end);

150 151 152 153 154 155 156
	start &= PUD_MASK;
	if (start < floor)
		return;
	if (ceiling) {
		ceiling &= PUD_MASK;
		if (!ceiling)
			return;
L
Linus Torvalds 已提交
157
	}
158 159 160 161 162 163
	if (end - 1 > ceiling - 1)
		return;

	pmd = pmd_offset(pud, start);
	pud_clear(pud);
	pmd_free_tlb(tlb, pmd);
L
Linus Torvalds 已提交
164 165
}

166 167 168
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
169 170 171
{
	pud_t *pud;
	unsigned long next;
172
	unsigned long start;
L
Linus Torvalds 已提交
173

174
	start = addr;
L
Linus Torvalds 已提交
175 176 177 178 179
	pud = pud_offset(pgd, addr);
	do {
		next = pud_addr_end(addr, end);
		if (pud_none_or_clear_bad(pud))
			continue;
180
		free_pmd_range(tlb, pud, addr, next, floor, ceiling);
L
Linus Torvalds 已提交
181 182
	} while (pud++, addr = next, addr != end);

183 184 185 186 187 188 189
	start &= PGDIR_MASK;
	if (start < floor)
		return;
	if (ceiling) {
		ceiling &= PGDIR_MASK;
		if (!ceiling)
			return;
L
Linus Torvalds 已提交
190
	}
191 192 193 194 195 196
	if (end - 1 > ceiling - 1)
		return;

	pud = pud_offset(pgd, start);
	pgd_clear(pgd);
	pud_free_tlb(tlb, pud);
L
Linus Torvalds 已提交
197 198 199
}

/*
200 201
 * This function frees user-level page tables of a process.
 *
L
Linus Torvalds 已提交
202 203
 * Must be called with pagetable lock held.
 */
204
void free_pgd_range(struct mmu_gather **tlb,
205 206
			unsigned long addr, unsigned long end,
			unsigned long floor, unsigned long ceiling)
L
Linus Torvalds 已提交
207 208 209
{
	pgd_t *pgd;
	unsigned long next;
210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236
	unsigned long start;

	/*
	 * 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
237

238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254
	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;

	start = addr;
255
	pgd = pgd_offset((*tlb)->mm, addr);
L
Linus Torvalds 已提交
256 257 258 259
	do {
		next = pgd_addr_end(addr, end);
		if (pgd_none_or_clear_bad(pgd))
			continue;
260
		free_pud_range(*tlb, pgd, addr, next, floor, ceiling);
L
Linus Torvalds 已提交
261
	} while (pgd++, addr = next, addr != end);
262 263 264
}

void free_pgtables(struct mmu_gather **tlb, struct vm_area_struct *vma,
265
		unsigned long floor, unsigned long ceiling)
266 267 268 269 270
{
	while (vma) {
		struct vm_area_struct *next = vma->vm_next;
		unsigned long addr = vma->vm_start;

271 272 273 274 275 276
		/*
		 * Hide vma from rmap and vmtruncate before freeing pgtables
		 */
		anon_vma_unlink(vma);
		unlink_file_vma(vma);

277
		if (is_vm_hugetlb_page(vma)) {
278
			hugetlb_free_pgd_range(tlb, addr, vma->vm_end,
279
				floor, next? next->vm_start: ceiling);
280 281 282 283 284
		} else {
			/*
			 * Optimization: gather nearby vmas into one call down
			 */
			while (next && next->vm_start <= vma->vm_end + PMD_SIZE
285
			       && !is_vm_hugetlb_page(next)) {
286 287
				vma = next;
				next = vma->vm_next;
288 289
				anon_vma_unlink(vma);
				unlink_file_vma(vma);
290 291 292 293
			}
			free_pgd_range(tlb, addr, vma->vm_end,
				floor, next? next->vm_start: ceiling);
		}
294 295
		vma = next;
	}
L
Linus Torvalds 已提交
296 297
}

298
int __pte_alloc(struct mm_struct *mm, pmd_t *pmd, unsigned long address)
L
Linus Torvalds 已提交
299
{
H
Hugh Dickins 已提交
300
	struct page *new = pte_alloc_one(mm, address);
301 302 303
	if (!new)
		return -ENOMEM;

H
Hugh Dickins 已提交
304
	pte_lock_init(new);
H
Hugh Dickins 已提交
305
	spin_lock(&mm->page_table_lock);
H
Hugh Dickins 已提交
306 307
	if (pmd_present(*pmd)) {	/* Another has populated it */
		pte_lock_deinit(new);
308
		pte_free(new);
H
Hugh Dickins 已提交
309
	} else {
L
Linus Torvalds 已提交
310
		mm->nr_ptes++;
311
		inc_zone_page_state(new, NR_PAGETABLE);
L
Linus Torvalds 已提交
312 313
		pmd_populate(mm, pmd, new);
	}
H
Hugh Dickins 已提交
314
	spin_unlock(&mm->page_table_lock);
315
	return 0;
L
Linus Torvalds 已提交
316 317
}

318
int __pte_alloc_kernel(pmd_t *pmd, unsigned long address)
L
Linus Torvalds 已提交
319
{
320 321 322 323 324 325 326 327 328 329 330
	pte_t *new = pte_alloc_one_kernel(&init_mm, address);
	if (!new)
		return -ENOMEM;

	spin_lock(&init_mm.page_table_lock);
	if (pmd_present(*pmd))		/* Another has populated it */
		pte_free_kernel(new);
	else
		pmd_populate_kernel(&init_mm, pmd, new);
	spin_unlock(&init_mm.page_table_lock);
	return 0;
L
Linus Torvalds 已提交
331 332
}

333 334 335 336 337 338 339 340
static inline void add_mm_rss(struct mm_struct *mm, int file_rss, int anon_rss)
{
	if (file_rss)
		add_mm_counter(mm, file_rss, file_rss);
	if (anon_rss)
		add_mm_counter(mm, anon_rss, anon_rss);
}

N
Nick Piggin 已提交
341
/*
342 343 344
 * 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
345 346 347 348 349 350 351 352 353 354 355 356 357
 *
 * The calling function must still handle the error.
 */
void print_bad_pte(struct vm_area_struct *vma, pte_t pte, unsigned long vaddr)
{
	printk(KERN_ERR "Bad pte = %08llx, process = %s, "
			"vm_flags = %lx, vaddr = %lx\n",
		(long long)pte_val(pte),
		(vma->vm_mm == current->mm ? current->comm : "???"),
		vma->vm_flags, vaddr);
	dump_stack();
}

358 359 360 361 362
static inline int is_cow_mapping(unsigned int flags)
{
	return (flags & (VM_SHARED | VM_MAYWRITE)) == VM_MAYWRITE;
}

H
Hugh Dickins 已提交
363
/*
364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381
 * This function gets the "struct page" associated with a pte.
 *
 * NOTE! Some mappings do not have "struct pages". A raw PFN mapping
 * will have each page table entry just pointing to a raw page frame
 * number, and as far as the VM layer is concerned, those do not have
 * pages associated with them - even if the PFN might point to memory
 * that otherwise is perfectly fine and has a "struct page".
 *
 * The way we recognize those mappings is through the rules set up
 * by "remap_pfn_range()": the vma will have the VM_PFNMAP bit set,
 * and the vm_pgoff will point to the first PFN mapped: thus every
 * page that is a raw mapping will always honor the rule
 *
 *	pfn_of_page == vma->vm_pgoff + ((addr - vma->vm_start) >> PAGE_SHIFT)
 *
 * and if that isn't true, the page has been COW'ed (in which case it
 * _does_ have a "struct page" associated with it even if it is in a
 * VM_PFNMAP range).
H
Hugh Dickins 已提交
382
 */
383
struct page *vm_normal_page(struct vm_area_struct *vma, unsigned long addr, pte_t pte)
H
Hugh Dickins 已提交
384
{
385 386
	unsigned long pfn = pte_pfn(pte);

N
Nick Piggin 已提交
387
	if (unlikely(vma->vm_flags & VM_PFNMAP)) {
388 389 390
		unsigned long off = (addr - vma->vm_start) >> PAGE_SHIFT;
		if (pfn == vma->vm_pgoff + off)
			return NULL;
391
		if (!is_cow_mapping(vma->vm_flags))
L
Linus Torvalds 已提交
392
			return NULL;
393 394
	}

395 396 397 398 399 400
	/*
	 * Add some anal sanity checks for now. Eventually,
	 * we should just do "return pfn_to_page(pfn)", but
	 * in the meantime we check that we get a valid pfn,
	 * and that the resulting page looks ok.
	 */
401 402 403 404 405 406 407 408 409 410 411 412 413
	if (unlikely(!pfn_valid(pfn))) {
		print_bad_pte(vma, pte, addr);
		return NULL;
	}

	/*
	 * NOTE! We still have PageReserved() pages in the page 
	 * tables. 
	 *
	 * The PAGE_ZERO() pages and various VDSO mappings can
	 * cause them to exist.
	 */
	return pfn_to_page(pfn);
H
Hugh Dickins 已提交
414 415
}

L
Linus Torvalds 已提交
416 417 418 419 420 421
/*
 * 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
422
static inline void
L
Linus Torvalds 已提交
423
copy_one_pte(struct mm_struct *dst_mm, struct mm_struct *src_mm,
N
Nick Piggin 已提交
424
		pte_t *dst_pte, pte_t *src_pte, struct vm_area_struct *vma,
H
Hugh Dickins 已提交
425
		unsigned long addr, int *rss)
L
Linus Torvalds 已提交
426
{
N
Nick Piggin 已提交
427
	unsigned long vm_flags = vma->vm_flags;
L
Linus Torvalds 已提交
428 429 430 431 432 433
	pte_t pte = *src_pte;
	struct page *page;

	/* pte contains position in swap or file, so copy. */
	if (unlikely(!pte_present(pte))) {
		if (!pte_file(pte)) {
434 435 436
			swp_entry_t entry = pte_to_swp_entry(pte);

			swap_duplicate(entry);
L
Linus Torvalds 已提交
437 438 439
			/* make sure dst_mm is on swapoff's mmlist. */
			if (unlikely(list_empty(&dst_mm->mmlist))) {
				spin_lock(&mmlist_lock);
440 441 442
				if (list_empty(&dst_mm->mmlist))
					list_add(&dst_mm->mmlist,
						 &src_mm->mmlist);
L
Linus Torvalds 已提交
443 444
				spin_unlock(&mmlist_lock);
			}
445 446 447 448 449 450 451 452 453 454
			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);
				set_pte_at(src_mm, addr, src_pte, pte);
			}
L
Linus Torvalds 已提交
455
		}
456
		goto out_set_pte;
L
Linus Torvalds 已提交
457 458 459 460 461 462
	}

	/*
	 * If it's a COW mapping, write protect it both
	 * in the parent and the child
	 */
463
	if (is_cow_mapping(vm_flags)) {
L
Linus Torvalds 已提交
464
		ptep_set_wrprotect(src_mm, addr, src_pte);
465
		pte = pte_wrprotect(pte);
L
Linus Torvalds 已提交
466 467 468 469 470 471 472 473 474
	}

	/*
	 * 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);
475 476 477 478

	page = vm_normal_page(vma, addr, pte);
	if (page) {
		get_page(page);
N
Nick Piggin 已提交
479
		page_dup_rmap(page, vma, addr);
480 481
		rss[!!PageAnon(page)]++;
	}
482 483 484

out_set_pte:
	set_pte_at(dst_mm, addr, dst_pte, pte);
L
Linus Torvalds 已提交
485 486 487 488 489 490 491
}

static int copy_pte_range(struct mm_struct *dst_mm, struct mm_struct *src_mm,
		pmd_t *dst_pmd, pmd_t *src_pmd, struct vm_area_struct *vma,
		unsigned long addr, unsigned long end)
{
	pte_t *src_pte, *dst_pte;
H
Hugh Dickins 已提交
492
	spinlock_t *src_ptl, *dst_ptl;
493
	int progress = 0;
H
Hugh Dickins 已提交
494
	int rss[2];
L
Linus Torvalds 已提交
495 496

again:
497
	rss[1] = rss[0] = 0;
H
Hugh Dickins 已提交
498
	dst_pte = pte_alloc_map_lock(dst_mm, dst_pmd, addr, &dst_ptl);
L
Linus Torvalds 已提交
499 500 501
	if (!dst_pte)
		return -ENOMEM;
	src_pte = pte_offset_map_nested(src_pmd, addr);
H
Hugh Dickins 已提交
502
	src_ptl = pte_lockptr(src_mm, src_pmd);
I
Ingo Molnar 已提交
503
	spin_lock_nested(src_ptl, SINGLE_DEPTH_NESTING);
504
	arch_enter_lazy_mmu_mode();
L
Linus Torvalds 已提交
505 506 507 508 509 510

	do {
		/*
		 * We are holding two locks at this point - either of them
		 * could generate latencies in another task on another CPU.
		 */
511 512 513
		if (progress >= 32) {
			progress = 0;
			if (need_resched() ||
H
Hugh Dickins 已提交
514 515
			    need_lockbreak(src_ptl) ||
			    need_lockbreak(dst_ptl))
516 517
				break;
		}
L
Linus Torvalds 已提交
518 519 520 521
		if (pte_none(*src_pte)) {
			progress++;
			continue;
		}
H
Hugh Dickins 已提交
522
		copy_one_pte(dst_mm, src_mm, dst_pte, src_pte, vma, addr, rss);
L
Linus Torvalds 已提交
523 524 525
		progress += 8;
	} while (dst_pte++, src_pte++, addr += PAGE_SIZE, addr != end);

526
	arch_leave_lazy_mmu_mode();
H
Hugh Dickins 已提交
527
	spin_unlock(src_ptl);
L
Linus Torvalds 已提交
528
	pte_unmap_nested(src_pte - 1);
529
	add_mm_rss(dst_mm, rss[0], rss[1]);
H
Hugh Dickins 已提交
530 531
	pte_unmap_unlock(dst_pte - 1, dst_ptl);
	cond_resched();
L
Linus Torvalds 已提交
532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588
	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);
		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;

589 590 591 592 593 594
	/*
	 * 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.
	 */
595
	if (!(vma->vm_flags & (VM_HUGETLB|VM_NONLINEAR|VM_PFNMAP|VM_INSERTPAGE))) {
596 597 598 599
		if (!vma->anon_vma)
			return 0;
	}

L
Linus Torvalds 已提交
600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615
	if (is_vm_hugetlb_page(vma))
		return copy_hugetlb_page_range(dst_mm, src_mm, vma);

	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;
		if (copy_pud_range(dst_mm, src_mm, dst_pgd, src_pgd,
						vma, addr, next))
			return -ENOMEM;
	} while (dst_pgd++, src_pgd++, addr = next, addr != end);
	return 0;
}

616
static unsigned long zap_pte_range(struct mmu_gather *tlb,
N
Nick Piggin 已提交
617
				struct vm_area_struct *vma, pmd_t *pmd,
L
Linus Torvalds 已提交
618
				unsigned long addr, unsigned long end,
619
				long *zap_work, struct zap_details *details)
L
Linus Torvalds 已提交
620
{
N
Nick Piggin 已提交
621
	struct mm_struct *mm = tlb->mm;
L
Linus Torvalds 已提交
622
	pte_t *pte;
623
	spinlock_t *ptl;
624 625
	int file_rss = 0;
	int anon_rss = 0;
L
Linus Torvalds 已提交
626

627
	pte = pte_offset_map_lock(mm, pmd, addr, &ptl);
628
	arch_enter_lazy_mmu_mode();
L
Linus Torvalds 已提交
629 630
	do {
		pte_t ptent = *pte;
631 632
		if (pte_none(ptent)) {
			(*zap_work)--;
L
Linus Torvalds 已提交
633
			continue;
634
		}
635 636 637

		(*zap_work) -= PAGE_SIZE;

L
Linus Torvalds 已提交
638
		if (pte_present(ptent)) {
H
Hugh Dickins 已提交
639
			struct page *page;
640

641
			page = vm_normal_page(vma, addr, ptent);
L
Linus Torvalds 已提交
642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659
			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;
				/*
				 * Each page->index must be checked when
				 * invalidating or truncating nonlinear.
				 */
				if (details->nonlinear_vma &&
				    (page->index < details->first_index ||
				     page->index > details->last_index))
					continue;
			}
N
Nick Piggin 已提交
660
			ptent = ptep_get_and_clear_full(mm, addr, pte,
661
							tlb->fullmm);
L
Linus Torvalds 已提交
662 663 664 665 666 667
			tlb_remove_tlb_entry(tlb, pte, addr);
			if (unlikely(!page))
				continue;
			if (unlikely(details) && details->nonlinear_vma
			    && linear_page_index(details->nonlinear_vma,
						addr) != page->index)
N
Nick Piggin 已提交
668
				set_pte_at(mm, addr, pte,
L
Linus Torvalds 已提交
669 670
					   pgoff_to_pte(page->index));
			if (PageAnon(page))
H
Hugh Dickins 已提交
671
				anon_rss--;
672 673 674 675
			else {
				if (pte_dirty(ptent))
					set_page_dirty(page);
				if (pte_young(ptent))
676
					SetPageReferenced(page);
H
Hugh Dickins 已提交
677
				file_rss--;
678
			}
N
Nick Piggin 已提交
679
			page_remove_rmap(page, vma);
L
Linus Torvalds 已提交
680 681 682 683 684 685 686 687 688 689 690
			tlb_remove_page(tlb, page);
			continue;
		}
		/*
		 * If details->check_mapping, we leave swap entries;
		 * if details->nonlinear_vma, we leave file entries.
		 */
		if (unlikely(details))
			continue;
		if (!pte_file(ptent))
			free_swap_and_cache(pte_to_swp_entry(ptent));
691
		pte_clear_not_present_full(mm, addr, pte, tlb->fullmm);
692
	} while (pte++, addr += PAGE_SIZE, (addr != end && *zap_work > 0));
693

H
Hugh Dickins 已提交
694
	add_mm_rss(mm, file_rss, anon_rss);
695
	arch_leave_lazy_mmu_mode();
696
	pte_unmap_unlock(pte - 1, ptl);
697 698

	return addr;
L
Linus Torvalds 已提交
699 700
}

701
static inline unsigned long zap_pmd_range(struct mmu_gather *tlb,
N
Nick Piggin 已提交
702
				struct vm_area_struct *vma, pud_t *pud,
L
Linus Torvalds 已提交
703
				unsigned long addr, unsigned long end,
704
				long *zap_work, struct zap_details *details)
L
Linus Torvalds 已提交
705 706 707 708 709 710 711
{
	pmd_t *pmd;
	unsigned long next;

	pmd = pmd_offset(pud, addr);
	do {
		next = pmd_addr_end(addr, end);
712 713
		if (pmd_none_or_clear_bad(pmd)) {
			(*zap_work)--;
L
Linus Torvalds 已提交
714
			continue;
715 716 717 718 719 720
		}
		next = zap_pte_range(tlb, vma, pmd, addr, next,
						zap_work, details);
	} while (pmd++, addr = next, (addr != end && *zap_work > 0));

	return addr;
L
Linus Torvalds 已提交
721 722
}

723
static inline unsigned long zap_pud_range(struct mmu_gather *tlb,
N
Nick Piggin 已提交
724
				struct vm_area_struct *vma, pgd_t *pgd,
L
Linus Torvalds 已提交
725
				unsigned long addr, unsigned long end,
726
				long *zap_work, struct zap_details *details)
L
Linus Torvalds 已提交
727 728 729 730 731 732 733
{
	pud_t *pud;
	unsigned long next;

	pud = pud_offset(pgd, addr);
	do {
		next = pud_addr_end(addr, end);
734 735
		if (pud_none_or_clear_bad(pud)) {
			(*zap_work)--;
L
Linus Torvalds 已提交
736
			continue;
737 738 739 740 741 742
		}
		next = zap_pmd_range(tlb, vma, pud, addr, next,
						zap_work, details);
	} while (pud++, addr = next, (addr != end && *zap_work > 0));

	return addr;
L
Linus Torvalds 已提交
743 744
}

745 746
static unsigned long unmap_page_range(struct mmu_gather *tlb,
				struct vm_area_struct *vma,
L
Linus Torvalds 已提交
747
				unsigned long addr, unsigned long end,
748
				long *zap_work, struct zap_details *details)
L
Linus Torvalds 已提交
749 750 751 752 753 754 755 756 757 758 759 760
{
	pgd_t *pgd;
	unsigned long next;

	if (details && !details->check_mapping && !details->nonlinear_vma)
		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);
761 762
		if (pgd_none_or_clear_bad(pgd)) {
			(*zap_work)--;
L
Linus Torvalds 已提交
763
			continue;
764 765 766 767
		}
		next = zap_pud_range(tlb, vma, pgd, addr, next,
						zap_work, details);
	} while (pgd++, addr = next, (addr != end && *zap_work > 0));
L
Linus Torvalds 已提交
768
	tlb_end_vma(tlb, vma);
769 770

	return addr;
L
Linus Torvalds 已提交
771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788
}

#ifdef CONFIG_PREEMPT
# define ZAP_BLOCK_SIZE	(8 * PAGE_SIZE)
#else
/* No preempt: go for improved straight-line efficiency */
# define ZAP_BLOCK_SIZE	(1024 * PAGE_SIZE)
#endif

/**
 * unmap_vmas - unmap a range of memory covered by a list of vma's
 * @tlbp: address of the caller's struct mmu_gather
 * @vma: the starting vma
 * @start_addr: virtual address at which to start unmapping
 * @end_addr: virtual address at which to end unmapping
 * @nr_accounted: Place number of unmapped pages in vm-accountable vma's here
 * @details: details of nonlinear truncation or shared cache invalidation
 *
789
 * Returns the end address of the unmapping (restart addr if interrupted).
L
Linus Torvalds 已提交
790
 *
791
 * Unmap all pages in the vma list.
L
Linus Torvalds 已提交
792
 *
793 794
 * We aim to not hold locks for too long (for scheduling latency reasons).
 * So zap pages in ZAP_BLOCK_SIZE bytecounts.  This means we need to
L
Linus Torvalds 已提交
795 796 797 798 799 800 801 802 803 804 805
 * return the ending mmu_gather to the caller.
 *
 * 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.
 */
806
unsigned long unmap_vmas(struct mmu_gather **tlbp,
L
Linus Torvalds 已提交
807 808 809 810
		struct vm_area_struct *vma, unsigned long start_addr,
		unsigned long end_addr, unsigned long *nr_accounted,
		struct zap_details *details)
{
811
	long zap_work = ZAP_BLOCK_SIZE;
L
Linus Torvalds 已提交
812 813
	unsigned long tlb_start = 0;	/* For tlb_finish_mmu */
	int tlb_start_valid = 0;
814
	unsigned long start = start_addr;
L
Linus Torvalds 已提交
815
	spinlock_t *i_mmap_lock = details? details->i_mmap_lock: NULL;
816
	int fullmm = (*tlbp)->fullmm;
L
Linus Torvalds 已提交
817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836

	for ( ; vma && vma->vm_start < end_addr; vma = vma->vm_next) {
		unsigned long end;

		start = max(vma->vm_start, start_addr);
		if (start >= vma->vm_end)
			continue;
		end = min(vma->vm_end, end_addr);
		if (end <= vma->vm_start)
			continue;

		if (vma->vm_flags & VM_ACCOUNT)
			*nr_accounted += (end - start) >> PAGE_SHIFT;

		while (start != end) {
			if (!tlb_start_valid) {
				tlb_start = start;
				tlb_start_valid = 1;
			}

837
			if (unlikely(is_vm_hugetlb_page(vma))) {
L
Linus Torvalds 已提交
838
				unmap_hugepage_range(vma, start, end);
839 840 841 842 843 844 845 846 847 848
				zap_work -= (end - start) /
						(HPAGE_SIZE / PAGE_SIZE);
				start = end;
			} else
				start = unmap_page_range(*tlbp, vma,
						start, end, &zap_work, details);

			if (zap_work > 0) {
				BUG_ON(start != end);
				break;
L
Linus Torvalds 已提交
849 850 851 852 853 854 855
			}

			tlb_finish_mmu(*tlbp, tlb_start, start);

			if (need_resched() ||
				(i_mmap_lock && need_lockbreak(i_mmap_lock))) {
				if (i_mmap_lock) {
856
					*tlbp = NULL;
L
Linus Torvalds 已提交
857 858 859 860 861
					goto out;
				}
				cond_resched();
			}

862
			*tlbp = tlb_gather_mmu(vma->vm_mm, fullmm);
L
Linus Torvalds 已提交
863
			tlb_start_valid = 0;
864
			zap_work = ZAP_BLOCK_SIZE;
L
Linus Torvalds 已提交
865 866 867
		}
	}
out:
868
	return start;	/* which is now the end (or restart) address */
L
Linus Torvalds 已提交
869 870 871 872 873 874 875 876 877
}

/**
 * zap_page_range - 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
 * @details: details of nonlinear truncation or shared cache invalidation
 */
878
unsigned long zap_page_range(struct vm_area_struct *vma, unsigned long address,
L
Linus Torvalds 已提交
879 880 881 882 883 884 885 886 887
		unsigned long size, struct zap_details *details)
{
	struct mm_struct *mm = vma->vm_mm;
	struct mmu_gather *tlb;
	unsigned long end = address + size;
	unsigned long nr_accounted = 0;

	lru_add_drain();
	tlb = tlb_gather_mmu(mm, 0);
888
	update_hiwater_rss(mm);
889 890 891
	end = unmap_vmas(&tlb, vma, address, end, &nr_accounted, details);
	if (tlb)
		tlb_finish_mmu(tlb, address, end);
892
	return end;
L
Linus Torvalds 已提交
893 894 895 896 897
}

/*
 * Do a quick page-table lookup for a single page.
 */
898
struct page *follow_page(struct vm_area_struct *vma, unsigned long address,
899
			unsigned int flags)
L
Linus Torvalds 已提交
900 901 902 903 904
{
	pgd_t *pgd;
	pud_t *pud;
	pmd_t *pmd;
	pte_t *ptep, pte;
905
	spinlock_t *ptl;
L
Linus Torvalds 已提交
906
	struct page *page;
907
	struct mm_struct *mm = vma->vm_mm;
L
Linus Torvalds 已提交
908

909 910 911 912 913
	page = follow_huge_addr(mm, address, flags & FOLL_WRITE);
	if (!IS_ERR(page)) {
		BUG_ON(flags & FOLL_GET);
		goto out;
	}
L
Linus Torvalds 已提交
914

915
	page = NULL;
L
Linus Torvalds 已提交
916 917
	pgd = pgd_offset(mm, address);
	if (pgd_none(*pgd) || unlikely(pgd_bad(*pgd)))
918
		goto no_page_table;
L
Linus Torvalds 已提交
919 920 921

	pud = pud_offset(pgd, address);
	if (pud_none(*pud) || unlikely(pud_bad(*pud)))
922
		goto no_page_table;
L
Linus Torvalds 已提交
923 924 925
	
	pmd = pmd_offset(pud, address);
	if (pmd_none(*pmd) || unlikely(pmd_bad(*pmd)))
926 927 928 929 930
		goto no_page_table;

	if (pmd_huge(*pmd)) {
		BUG_ON(flags & FOLL_GET);
		page = follow_huge_pmd(mm, address, pmd, flags & FOLL_WRITE);
L
Linus Torvalds 已提交
931
		goto out;
932
	}
L
Linus Torvalds 已提交
933

934
	ptep = pte_offset_map_lock(mm, pmd, address, &ptl);
L
Linus Torvalds 已提交
935 936 937 938
	if (!ptep)
		goto out;

	pte = *ptep;
939 940 941 942
	if (!pte_present(pte))
		goto unlock;
	if ((flags & FOLL_WRITE) && !pte_write(pte))
		goto unlock;
943 944
	page = vm_normal_page(vma, address, pte);
	if (unlikely(!page))
945
		goto unlock;
L
Linus Torvalds 已提交
946

947 948 949 950 951 952 953 954 955 956
	if (flags & FOLL_GET)
		get_page(page);
	if (flags & FOLL_TOUCH) {
		if ((flags & FOLL_WRITE) &&
		    !pte_dirty(pte) && !PageDirty(page))
			set_page_dirty(page);
		mark_page_accessed(page);
	}
unlock:
	pte_unmap_unlock(ptep, ptl);
L
Linus Torvalds 已提交
957
out:
958
	return page;
L
Linus Torvalds 已提交
959

960 961 962 963 964 965
no_page_table:
	/*
	 * When core dumping an enormous anonymous area that nobody
	 * has touched so far, we don't want to allocate page tables.
	 */
	if (flags & FOLL_ANON) {
N
Nick Piggin 已提交
966
		page = ZERO_PAGE(0);
967 968 969 970 971
		if (flags & FOLL_GET)
			get_page(page);
		BUG_ON(flags & FOLL_WRITE);
	}
	return page;
L
Linus Torvalds 已提交
972 973 974 975 976 977 978
}

int get_user_pages(struct task_struct *tsk, struct mm_struct *mm,
		unsigned long start, int len, int write, int force,
		struct page **pages, struct vm_area_struct **vmas)
{
	int i;
979
	unsigned int vm_flags;
L
Linus Torvalds 已提交
980 981 982 983 984

	/* 
	 * Require read or write permissions.
	 * If 'force' is set, we only require the "MAY" flags.
	 */
985 986
	vm_flags  = write ? (VM_WRITE | VM_MAYWRITE) : (VM_READ | VM_MAYREAD);
	vm_flags &= force ? (VM_MAYREAD | VM_MAYWRITE) : (VM_READ | VM_WRITE);
L
Linus Torvalds 已提交
987 988 989
	i = 0;

	do {
990 991
		struct vm_area_struct *vma;
		unsigned int foll_flags;
L
Linus Torvalds 已提交
992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010

		vma = find_extend_vma(mm, start);
		if (!vma && in_gate_area(tsk, start)) {
			unsigned long pg = start & PAGE_MASK;
			struct vm_area_struct *gate_vma = get_gate_vma(tsk);
			pgd_t *pgd;
			pud_t *pud;
			pmd_t *pmd;
			pte_t *pte;
			if (write) /* user gate pages are read-only */
				return i ? : -EFAULT;
			if (pg > TASK_SIZE)
				pgd = pgd_offset_k(pg);
			else
				pgd = pgd_offset_gate(mm, pg);
			BUG_ON(pgd_none(*pgd));
			pud = pud_offset(pgd, pg);
			BUG_ON(pud_none(*pud));
			pmd = pmd_offset(pud, pg);
1011 1012
			if (pmd_none(*pmd))
				return i ? : -EFAULT;
L
Linus Torvalds 已提交
1013
			pte = pte_offset_map(pmd, pg);
1014 1015 1016 1017
			if (pte_none(*pte)) {
				pte_unmap(pte);
				return i ? : -EFAULT;
			}
L
Linus Torvalds 已提交
1018
			if (pages) {
1019
				struct page *page = vm_normal_page(gate_vma, start, *pte);
1020 1021 1022
				pages[i] = page;
				if (page)
					get_page(page);
L
Linus Torvalds 已提交
1023 1024 1025 1026 1027 1028 1029 1030 1031 1032
			}
			pte_unmap(pte);
			if (vmas)
				vmas[i] = gate_vma;
			i++;
			start += PAGE_SIZE;
			len--;
			continue;
		}

1033
		if (!vma || (vma->vm_flags & (VM_IO | VM_PFNMAP))
1034
				|| !(vm_flags & vma->vm_flags))
L
Linus Torvalds 已提交
1035 1036 1037 1038
			return i ? : -EFAULT;

		if (is_vm_hugetlb_page(vma)) {
			i = follow_hugetlb_page(mm, vma, pages, vmas,
1039
						&start, &len, i, write);
L
Linus Torvalds 已提交
1040 1041
			continue;
		}
1042 1043 1044 1045 1046

		foll_flags = FOLL_TOUCH;
		if (pages)
			foll_flags |= FOLL_GET;
		if (!write && !(vma->vm_flags & VM_LOCKED) &&
1047 1048
		    (!vma->vm_ops || (!vma->vm_ops->nopage &&
					!vma->vm_ops->fault)))
1049 1050
			foll_flags |= FOLL_ANON;

L
Linus Torvalds 已提交
1051
		do {
1052
			struct page *page;
L
Linus Torvalds 已提交
1053

1054 1055 1056 1057 1058 1059 1060 1061
			/*
			 * If tsk is ooming, cut off its access to large memory
			 * allocations. It has a pending SIGKILL, but it can't
			 * be processed until returning to user space.
			 */
			if (unlikely(test_tsk_thread_flag(tsk, TIF_MEMDIE)))
				return -ENOMEM;

1062 1063
			if (write)
				foll_flags |= FOLL_WRITE;
1064

1065
			cond_resched();
1066
			while (!(page = follow_page(vma, start, foll_flags))) {
1067
				int ret;
N
Nick Piggin 已提交
1068
				ret = handle_mm_fault(mm, vma, start,
1069
						foll_flags & FOLL_WRITE);
N
Nick Piggin 已提交
1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081
				if (ret & VM_FAULT_ERROR) {
					if (ret & VM_FAULT_OOM)
						return i ? i : -ENOMEM;
					else if (ret & VM_FAULT_SIGBUS)
						return i ? i : -EFAULT;
					BUG();
				}
				if (ret & VM_FAULT_MAJOR)
					tsk->maj_flt++;
				else
					tsk->min_flt++;

1082
				/*
N
Nick Piggin 已提交
1083 1084 1085 1086 1087
				 * The VM_FAULT_WRITE bit tells us that
				 * do_wp_page has broken COW when necessary,
				 * even if maybe_mkwrite decided not to set
				 * pte_write. We can thus safely do subsequent
				 * page lookups as if they were reads.
1088 1089
				 */
				if (ret & VM_FAULT_WRITE)
1090
					foll_flags &= ~FOLL_WRITE;
N
Nick Piggin 已提交
1091

1092
				cond_resched();
L
Linus Torvalds 已提交
1093 1094
			}
			if (pages) {
1095
				pages[i] = page;
1096

1097
				flush_anon_page(vma, page, start);
1098
				flush_dcache_page(page);
L
Linus Torvalds 已提交
1099 1100 1101 1102 1103 1104
			}
			if (vmas)
				vmas[i] = vma;
			i++;
			start += PAGE_SIZE;
			len--;
1105 1106
		} while (len && start < vma->vm_end);
	} while (len);
L
Linus Torvalds 已提交
1107 1108 1109 1110
	return i;
}
EXPORT_SYMBOL(get_user_pages);

1111
pte_t * fastcall get_locked_pte(struct mm_struct *mm, unsigned long addr, spinlock_t **ptl)
1112 1113 1114 1115
{
	pgd_t * pgd = pgd_offset(mm, addr);
	pud_t * pud = pud_alloc(mm, pgd, addr);
	if (pud) {
1116
		pmd_t * pmd = pmd_alloc(mm, pud, addr);
1117 1118 1119 1120 1121 1122
		if (pmd)
			return pte_alloc_map_lock(mm, pmd, addr, ptl);
	}
	return NULL;
}

1123 1124 1125 1126 1127 1128 1129 1130 1131 1132
/*
 * 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.
 */
static int insert_page(struct mm_struct *mm, unsigned long addr, struct page *page, pgprot_t prot)
{
	int retval;
1133
	pte_t *pte;
1134 1135 1136
	spinlock_t *ptl;  

	retval = -EINVAL;
1137
	if (PageAnon(page))
1138 1139 1140
		goto out;
	retval = -ENOMEM;
	flush_dcache_page(page);
1141
	pte = get_locked_pte(mm, addr, &ptl);
1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160
	if (!pte)
		goto out;
	retval = -EBUSY;
	if (!pte_none(*pte))
		goto out_unlock;

	/* Ok, finally just insert the thing.. */
	get_page(page);
	inc_mm_counter(mm, file_rss);
	page_add_file_rmap(page);
	set_pte_at(mm, addr, pte, mk_pte(page, prot));

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

1161 1162 1163 1164 1165 1166
/**
 * 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
 *
1167 1168 1169 1170 1171 1172
 * 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
1173
 * (see split_page()).
1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188
 *
 * 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.
 */
int vm_insert_page(struct vm_area_struct *vma, unsigned long addr, struct page *page)
{
	if (addr < vma->vm_start || addr >= vma->vm_end)
		return -EFAULT;
	if (!page_count(page))
		return -EINVAL;
1189
	vma->vm_flags |= VM_INSERTPAGE;
1190 1191
	return insert_page(vma->vm_mm, addr, page, vma->vm_page_prot);
}
1192
EXPORT_SYMBOL(vm_insert_page);
1193

N
Nick Piggin 已提交
1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238
/**
 * 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
 *
 * Similar to vm_inert_page, this allows drivers to insert individual pages
 * 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.
 */
int vm_insert_pfn(struct vm_area_struct *vma, unsigned long addr,
		unsigned long pfn)
{
	struct mm_struct *mm = vma->vm_mm;
	int retval;
	pte_t *pte, entry;
	spinlock_t *ptl;

	BUG_ON(!(vma->vm_flags & VM_PFNMAP));
	BUG_ON(is_cow_mapping(vma->vm_flags));

	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 = pfn_pte(pfn, vma->vm_page_prot);
	set_pte_at(mm, addr, pte, entry);
	update_mmu_cache(vma, addr, entry);

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

out:
	return retval;
}
EXPORT_SYMBOL(vm_insert_pfn);

L
Linus Torvalds 已提交
1239 1240 1241 1242 1243 1244 1245 1246 1247 1248
/*
 * 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
1249
	spinlock_t *ptl;
L
Linus Torvalds 已提交
1250

H
Hugh Dickins 已提交
1251
	pte = pte_alloc_map_lock(mm, pmd, addr, &ptl);
L
Linus Torvalds 已提交
1252 1253
	if (!pte)
		return -ENOMEM;
1254
	arch_enter_lazy_mmu_mode();
L
Linus Torvalds 已提交
1255 1256
	do {
		BUG_ON(!pte_none(*pte));
N
Nick Piggin 已提交
1257
		set_pte_at(mm, addr, pte, pfn_pte(pfn, prot));
L
Linus Torvalds 已提交
1258 1259
		pfn++;
	} while (pte++, addr += PAGE_SIZE, addr != end);
1260
	arch_leave_lazy_mmu_mode();
H
Hugh Dickins 已提交
1261
	pte_unmap_unlock(pte - 1, ptl);
L
Linus Torvalds 已提交
1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304
	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;
	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;
}

1305 1306 1307 1308 1309 1310 1311 1312 1313 1314
/**
 * 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
1315 1316 1317 1318 1319
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;
1320
	unsigned long end = addr + PAGE_ALIGN(size);
L
Linus Torvalds 已提交
1321 1322 1323 1324 1325 1326 1327 1328
	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).
H
Hugh Dickins 已提交
1329 1330 1331 1332 1333
	 *   VM_RESERVED is specified all over the place, because
	 *	in 2.4 it kept swapout's vma scan off this vma; but
	 *	in 2.6 the LRU scan won't even find its pages, so this
	 *	flag means no more than count its pages in reserved_vm,
	 * 	and omit it from core dump, even when VM_IO turned off.
1334 1335 1336
	 *   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.
L
Linus Torvalds 已提交
1337 1338 1339 1340
	 *
	 * 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".
L
Linus Torvalds 已提交
1341
	 */
1342
	if (is_cow_mapping(vma->vm_flags)) {
L
Linus Torvalds 已提交
1343
		if (addr != vma->vm_start || end != vma->vm_end)
1344
			return -EINVAL;
L
Linus Torvalds 已提交
1345 1346 1347
		vma->vm_pgoff = pfn;
	}

1348
	vma->vm_flags |= VM_IO | VM_RESERVED | VM_PFNMAP;
L
Linus Torvalds 已提交
1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364

	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);
	return err;
}
EXPORT_SYMBOL(remap_pfn_range);

1365 1366 1367 1368 1369 1370 1371
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;
	struct page *pmd_page;
1372
	spinlock_t *uninitialized_var(ptl);
1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458

	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));

	pmd_page = pmd_page(*pmd);

	do {
		err = fn(pte, pmd_page, addr, data);
		if (err)
			break;
	} while (pte++, addr += PAGE_SIZE, addr != end);

	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;

	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;
	unsigned long end = addr + size;
	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);
	return err;
}
EXPORT_SYMBOL_GPL(apply_to_page_range);

1459 1460 1461 1462 1463 1464 1465 1466 1467
/*
 * 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 and do_file_page
 * must check under lock before unmapping the pte and proceeding
 * (but do_wp_page is only called after already making such a check;
 * and do_anonymous_page and do_no_page can safely check later on).
 */
H
Hugh Dickins 已提交
1468
static inline int pte_unmap_same(struct mm_struct *mm, pmd_t *pmd,
1469 1470 1471 1472 1473
				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
1474 1475
		spinlock_t *ptl = pte_lockptr(mm, pmd);
		spin_lock(ptl);
1476
		same = pte_same(*page_table, orig_pte);
H
Hugh Dickins 已提交
1477
		spin_unlock(ptl);
1478 1479 1480 1481 1482 1483
	}
#endif
	pte_unmap(page_table);
	return same;
}

L
Linus Torvalds 已提交
1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496
/*
 * Do pte_mkwrite, but only if the vma says VM_WRITE.  We do this when
 * servicing faults for write access.  In the normal case, do always want
 * pte_mkwrite.  But get_user_pages can cause write faults for mappings
 * that do not have writing enabled, when used by access_process_vm.
 */
static inline pte_t maybe_mkwrite(pte_t pte, struct vm_area_struct *vma)
{
	if (likely(vma->vm_flags & VM_WRITE))
		pte = pte_mkwrite(pte);
	return pte;
}

1497
static inline void cow_user_page(struct page *dst, struct page *src, unsigned long va, struct vm_area_struct *vma)
1498 1499 1500 1501 1502 1503 1504 1505 1506
{
	/*
	 * 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)) {
		void *kaddr = kmap_atomic(dst, KM_USER0);
L
Linus Torvalds 已提交
1507 1508 1509 1510 1511 1512 1513 1514 1515
		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))
1516 1517
			memset(kaddr, 0, PAGE_SIZE);
		kunmap_atomic(kaddr, KM_USER0);
1518
		flush_dcache_page(dst);
1519
		return;
1520

1521
	}
1522
	copy_user_highpage(dst, src, va, vma);
1523 1524
}

L
Linus Torvalds 已提交
1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538
/*
 * 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.
 *
1539 1540 1541
 * 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
1542
 */
1543 1544
static int do_wp_page(struct mm_struct *mm, struct vm_area_struct *vma,
		unsigned long address, pte_t *page_table, pmd_t *pmd,
1545
		spinlock_t *ptl, pte_t orig_pte)
L
Linus Torvalds 已提交
1546
{
1547
	struct page *old_page, *new_page;
L
Linus Torvalds 已提交
1548
	pte_t entry;
N
Nick Piggin 已提交
1549
	int reuse = 0, ret = 0;
1550
	int page_mkwrite = 0;
1551
	struct page *dirty_page = NULL;
L
Linus Torvalds 已提交
1552

1553 1554 1555
	old_page = vm_normal_page(vma, address, orig_pte);
	if (!old_page)
		goto gotten;
L
Linus Torvalds 已提交
1556

1557
	/*
P
Peter Zijlstra 已提交
1558 1559
	 * Take out anonymous pages first, anonymous shared vmas are
	 * not dirty accountable.
1560
	 */
P
Peter Zijlstra 已提交
1561 1562 1563 1564 1565 1566
	if (PageAnon(old_page)) {
		if (!TestSetPageLocked(old_page)) {
			reuse = can_share_swap_page(old_page);
			unlock_page(old_page);
		}
	} else if (unlikely((vma->vm_flags & (VM_WRITE|VM_SHARED)) ==
1567
					(VM_WRITE|VM_SHARED))) {
P
Peter Zijlstra 已提交
1568 1569 1570 1571 1572
		/*
		 * Only catch write-faults on shared writable pages,
		 * read-only shared pages can get COWed by
		 * get_user_pages(.write=1, .force=1).
		 */
1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595
		if (vma->vm_ops && vma->vm_ops->page_mkwrite) {
			/*
			 * 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.
			 */
			page_cache_get(old_page);
			pte_unmap_unlock(page_table, ptl);

			if (vma->vm_ops->page_mkwrite(vma, old_page) < 0)
				goto unwritable_page;

			/*
			 * 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);
1596
			page_cache_release(old_page);
1597 1598
			if (!pte_same(*page_table, orig_pte))
				goto unlock;
1599 1600

			page_mkwrite = 1;
L
Linus Torvalds 已提交
1601
		}
1602 1603
		dirty_page = old_page;
		get_page(dirty_page);
1604 1605 1606 1607 1608 1609 1610
		reuse = 1;
	}

	if (reuse) {
		flush_cache_page(vma, address, pte_pfn(orig_pte));
		entry = pte_mkyoung(orig_pte);
		entry = maybe_mkwrite(pte_mkdirty(entry), vma);
1611
		if (ptep_set_access_flags(vma, address, page_table, entry,1))
1612
			update_mmu_cache(vma, address, entry);
1613 1614
		ret |= VM_FAULT_WRITE;
		goto unlock;
L
Linus Torvalds 已提交
1615 1616 1617 1618 1619
	}

	/*
	 * Ok, we need to copy. Oh, well..
	 */
N
Nick Piggin 已提交
1620
	page_cache_get(old_page);
H
Hugh Dickins 已提交
1621
gotten:
1622
	pte_unmap_unlock(page_table, ptl);
L
Linus Torvalds 已提交
1623 1624

	if (unlikely(anon_vma_prepare(vma)))
1625
		goto oom;
N
Nick Piggin 已提交
1626 1627 1628 1629 1630
	VM_BUG_ON(old_page == ZERO_PAGE(0));
	new_page = alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma, address);
	if (!new_page)
		goto oom;
	cow_user_page(new_page, old_page, address, vma);
1631

L
Linus Torvalds 已提交
1632 1633 1634
	/*
	 * Re-check the pte - we dropped the lock
	 */
1635
	page_table = pte_offset_map_lock(mm, pmd, address, &ptl);
1636
	if (likely(pte_same(*page_table, orig_pte))) {
H
Hugh Dickins 已提交
1637
		if (old_page) {
N
Nick Piggin 已提交
1638
			page_remove_rmap(old_page, vma);
H
Hugh Dickins 已提交
1639 1640 1641 1642 1643
			if (!PageAnon(old_page)) {
				dec_mm_counter(mm, file_rss);
				inc_mm_counter(mm, anon_rss);
			}
		} else
1644
			inc_mm_counter(mm, anon_rss);
1645
		flush_cache_page(vma, address, pte_pfn(orig_pte));
1646 1647
		entry = mk_pte(new_page, vma->vm_page_prot);
		entry = maybe_mkwrite(pte_mkdirty(entry), vma);
1648 1649 1650 1651 1652 1653 1654 1655
		/*
		 * 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.
		 */
		ptep_clear_flush(vma, address, page_table);
		set_pte_at(mm, address, page_table, entry);
1656
		update_mmu_cache(vma, address, entry);
L
Linus Torvalds 已提交
1657
		lru_cache_add_active(new_page);
N
Nick Piggin 已提交
1658
		page_add_new_anon_rmap(new_page, vma, address);
L
Linus Torvalds 已提交
1659 1660 1661

		/* Free the old page.. */
		new_page = old_page;
N
Nick Piggin 已提交
1662
		ret |= VM_FAULT_WRITE;
L
Linus Torvalds 已提交
1663
	}
H
Hugh Dickins 已提交
1664 1665 1666 1667
	if (new_page)
		page_cache_release(new_page);
	if (old_page)
		page_cache_release(old_page);
1668
unlock:
1669
	pte_unmap_unlock(page_table, ptl);
1670
	if (dirty_page) {
1671 1672 1673 1674 1675 1676 1677 1678 1679
		/*
		 * Yes, Virginia, this is actually required to prevent a race
		 * with clear_page_dirty_for_io() from clearing the page dirty
		 * bit after it clear all dirty ptes, but before a racing
		 * do_wp_page installs a dirty pte.
		 *
		 * do_no_page is protected similarly.
		 */
		wait_on_page_locked(dirty_page);
1680
		set_page_dirty_balance(dirty_page, page_mkwrite);
1681 1682
		put_page(dirty_page);
	}
N
Nick Piggin 已提交
1683
	return ret;
1684
oom:
H
Hugh Dickins 已提交
1685 1686
	if (old_page)
		page_cache_release(old_page);
L
Linus Torvalds 已提交
1687
	return VM_FAULT_OOM;
1688 1689 1690 1691

unwritable_page:
	page_cache_release(old_page);
	return VM_FAULT_SIGBUS;
L
Linus Torvalds 已提交
1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718
}

/*
 * Helper functions for unmap_mapping_range().
 *
 * __ Notes on dropping i_mmap_lock to reduce latency while unmapping __
 *
 * We have to restart searching the prio_tree whenever we drop the lock,
 * since the iterator is only valid while the lock is held, and anyway
 * a later vma might be split and reinserted earlier while lock dropped.
 *
 * The list of nonlinear vmas could be handled more efficiently, using
 * a placeholder, but handle it in the same way until a need is shown.
 * It is important to search the prio_tree before nonlinear list: a vma
 * may become nonlinear and be shifted from prio_tree to nonlinear list
 * while the lock is dropped; but never shifted from list to prio_tree.
 *
 * In order to make forward progress despite restarting the search,
 * vm_truncate_count is used to mark a vma as now dealt with, so we can
 * quickly skip it next time around.  Since the prio_tree search only
 * shows us those vmas affected by unmapping the range in question, we
 * can't efficiently keep all vmas in step with mapping->truncate_count:
 * so instead reset them all whenever it wraps back to 0 (then go to 1).
 * mapping->truncate_count and vma->vm_truncate_count are protected by
 * i_mmap_lock.
 *
 * In order to make forward progress despite repeatedly restarting some
1719
 * large vma, note the restart_addr from unmap_vmas when it breaks out:
L
Linus Torvalds 已提交
1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745
 * and restart from that address when we reach that vma again.  It might
 * have been split or merged, shrunk or extended, but never shifted: so
 * restart_addr remains valid so long as it remains in the vma's range.
 * unmap_mapping_range forces truncate_count to leap over page-aligned
 * values so we can save vma's restart_addr in its truncate_count field.
 */
#define is_restart_addr(truncate_count) (!((truncate_count) & ~PAGE_MASK))

static void reset_vma_truncate_counts(struct address_space *mapping)
{
	struct vm_area_struct *vma;
	struct prio_tree_iter iter;

	vma_prio_tree_foreach(vma, &iter, &mapping->i_mmap, 0, ULONG_MAX)
		vma->vm_truncate_count = 0;
	list_for_each_entry(vma, &mapping->i_mmap_nonlinear, shared.vm_set.list)
		vma->vm_truncate_count = 0;
}

static int unmap_mapping_range_vma(struct vm_area_struct *vma,
		unsigned long start_addr, unsigned long end_addr,
		struct zap_details *details)
{
	unsigned long restart_addr;
	int need_break;

1746 1747
	/*
	 * files that support invalidating or truncating portions of the
N
Nick Piggin 已提交
1748
	 * file from under mmaped areas must have their ->fault function
N
Nick Piggin 已提交
1749 1750
	 * return a locked page (and set VM_FAULT_LOCKED in the return).
	 * This provides synchronisation against concurrent unmapping here.
1751 1752
	 */

L
Linus Torvalds 已提交
1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763
again:
	restart_addr = vma->vm_truncate_count;
	if (is_restart_addr(restart_addr) && start_addr < restart_addr) {
		start_addr = restart_addr;
		if (start_addr >= end_addr) {
			/* Top of vma has been split off since last time */
			vma->vm_truncate_count = details->truncate_count;
			return 0;
		}
	}

1764 1765
	restart_addr = zap_page_range(vma, start_addr,
					end_addr - start_addr, details);
L
Linus Torvalds 已提交
1766 1767 1768
	need_break = need_resched() ||
			need_lockbreak(details->i_mmap_lock);

1769
	if (restart_addr >= end_addr) {
L
Linus Torvalds 已提交
1770 1771 1772 1773 1774 1775
		/* We have now completed this vma: mark it so */
		vma->vm_truncate_count = details->truncate_count;
		if (!need_break)
			return 0;
	} else {
		/* Note restart_addr in vma's truncate_count field */
1776
		vma->vm_truncate_count = restart_addr;
L
Linus Torvalds 已提交
1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842
		if (!need_break)
			goto again;
	}

	spin_unlock(details->i_mmap_lock);
	cond_resched();
	spin_lock(details->i_mmap_lock);
	return -EINTR;
}

static inline void unmap_mapping_range_tree(struct prio_tree_root *root,
					    struct zap_details *details)
{
	struct vm_area_struct *vma;
	struct prio_tree_iter iter;
	pgoff_t vba, vea, zba, zea;

restart:
	vma_prio_tree_foreach(vma, &iter, root,
			details->first_index, details->last_index) {
		/* Skip quickly over those we have already dealt with */
		if (vma->vm_truncate_count == details->truncate_count)
			continue;

		vba = vma->vm_pgoff;
		vea = vba + ((vma->vm_end - vma->vm_start) >> PAGE_SHIFT) - 1;
		/* 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;

		if (unmap_mapping_range_vma(vma,
			((zba - vba) << PAGE_SHIFT) + vma->vm_start,
			((zea - vba + 1) << PAGE_SHIFT) + vma->vm_start,
				details) < 0)
			goto restart;
	}
}

static inline void unmap_mapping_range_list(struct list_head *head,
					    struct zap_details *details)
{
	struct vm_area_struct *vma;

	/*
	 * In nonlinear VMAs there is no correspondence between virtual address
	 * offset and file offset.  So we must perform an exhaustive search
	 * across *all* the pages in each nonlinear VMA, not just the pages
	 * whose virtual address lies outside the file truncation point.
	 */
restart:
	list_for_each_entry(vma, head, shared.vm_set.list) {
		/* Skip quickly over those we have already dealt with */
		if (vma->vm_truncate_count == details->truncate_count)
			continue;
		details->nonlinear_vma = vma;
		if (unmap_mapping_range_vma(vma, vma->vm_start,
					vma->vm_end, details) < 0)
			goto restart;
	}
}

/**
1843
 * 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
1844
 * @mapping: the address space containing mmaps to be unmapped.
L
Linus Torvalds 已提交
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
 * @holebegin: byte in first page to unmap, relative to the start of
 * the underlying file.  This will be rounded down to a PAGE_SIZE
 * boundary.  Note that this is different from vmtruncate(), which
 * 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.nonlinear_vma = NULL;
	details.first_index = hba;
	details.last_index = hba + hlen - 1;
	if (details.last_index < details.first_index)
		details.last_index = ULONG_MAX;
	details.i_mmap_lock = &mapping->i_mmap_lock;

	spin_lock(&mapping->i_mmap_lock);

1881
	/* Protect against endless unmapping loops */
L
Linus Torvalds 已提交
1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897
	mapping->truncate_count++;
	if (unlikely(is_restart_addr(mapping->truncate_count))) {
		if (mapping->truncate_count == 0)
			reset_vma_truncate_counts(mapping);
		mapping->truncate_count++;
	}
	details.truncate_count = mapping->truncate_count;

	if (unlikely(!prio_tree_empty(&mapping->i_mmap)))
		unmap_mapping_range_tree(&mapping->i_mmap, &details);
	if (unlikely(!list_empty(&mapping->i_mmap_nonlinear)))
		unmap_mapping_range_list(&mapping->i_mmap_nonlinear, &details);
	spin_unlock(&mapping->i_mmap_lock);
}
EXPORT_SYMBOL(unmap_mapping_range);

1898 1899 1900 1901
/**
 * vmtruncate - unmap mappings "freed" by truncate() syscall
 * @inode: inode of the file used
 * @offset: file offset to start truncating
L
Linus Torvalds 已提交
1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920
 *
 * NOTE! We have to be ready to update the memory sharing
 * between the file and the memory map for a potential last
 * incomplete page.  Ugly, but necessary.
 */
int vmtruncate(struct inode * inode, loff_t offset)
{
	struct address_space *mapping = inode->i_mapping;
	unsigned long limit;

	if (inode->i_size < offset)
		goto do_expand;
	/*
	 * truncation of in-use swapfiles is disallowed - it would cause
	 * subsequent swapout to scribble on the now-freed blocks.
	 */
	if (IS_SWAPFILE(inode))
		goto out_busy;
	i_size_write(inode, offset);
1921 1922 1923 1924 1925 1926 1927 1928 1929

	/*
	 * unmap_mapping_range is called twice, first simply for efficiency
	 * so that truncate_inode_pages does fewer single-page unmaps. However
	 * after this first call, and before truncate_inode_pages finishes,
	 * it is possible for private pages to be COWed, which remain after
	 * truncate_inode_pages finishes, hence the second unmap_mapping_range
	 * call must be made for correctness.
	 */
L
Linus Torvalds 已提交
1930 1931
	unmap_mapping_range(mapping, offset + PAGE_SIZE - 1, 0, 1);
	truncate_inode_pages(mapping, offset);
1932
	unmap_mapping_range(mapping, offset + PAGE_SIZE - 1, 0, 1);
L
Linus Torvalds 已提交
1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955
	goto out_truncate;

do_expand:
	limit = current->signal->rlim[RLIMIT_FSIZE].rlim_cur;
	if (limit != RLIM_INFINITY && offset > limit)
		goto out_sig;
	if (offset > inode->i_sb->s_maxbytes)
		goto out_big;
	i_size_write(inode, offset);

out_truncate:
	if (inode->i_op && inode->i_op->truncate)
		inode->i_op->truncate(inode);
	return 0;
out_sig:
	send_sig(SIGXFSZ, current, 0);
out_big:
	return -EFBIG;
out_busy:
	return -ETXTBSY;
}
EXPORT_SYMBOL(vmtruncate);

1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967
int vmtruncate_range(struct inode *inode, loff_t offset, loff_t end)
{
	struct address_space *mapping = inode->i_mapping;

	/*
	 * If the underlying filesystem is not going to provide
	 * a way to truncate a range of blocks (punch a hole) -
	 * we should return failure right now.
	 */
	if (!inode->i_op || !inode->i_op->truncate_range)
		return -ENOSYS;

1968
	mutex_lock(&inode->i_mutex);
1969 1970 1971
	down_write(&inode->i_alloc_sem);
	unmap_mapping_range(mapping, offset, (end - offset), 1);
	truncate_inode_pages_range(mapping, offset, end);
1972
	unmap_mapping_range(mapping, offset, (end - offset), 1);
1973 1974
	inode->i_op->truncate_range(inode, offset, end);
	up_write(&inode->i_alloc_sem);
1975
	mutex_unlock(&inode->i_mutex);
1976 1977 1978 1979

	return 0;
}

1980 1981 1982 1983 1984 1985
/**
 * swapin_readahead - swap in pages in hope we need them soon
 * @entry: swap entry of this memory
 * @addr: address to start
 * @vma: user vma this addresses belong to
 *
L
Linus Torvalds 已提交
1986 1987 1988
 * Primitive swap readahead code. We simply read an aligned block of
 * (1 << page_cluster) entries in the swap area. This method is chosen
 * because it doesn't cost us any seek time.  We also make sure to queue
1989
 * the 'original' request together with the readahead ones...
L
Linus Torvalds 已提交
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041
 *
 * This has been extended to use the NUMA policies from the mm triggering
 * the readahead.
 *
 * Caller must hold down_read on the vma->vm_mm if vma is not NULL.
 */
void swapin_readahead(swp_entry_t entry, unsigned long addr,struct vm_area_struct *vma)
{
#ifdef CONFIG_NUMA
	struct vm_area_struct *next_vma = vma ? vma->vm_next : NULL;
#endif
	int i, num;
	struct page *new_page;
	unsigned long offset;

	/*
	 * Get the number of handles we should do readahead io to.
	 */
	num = valid_swaphandles(entry, &offset);
	for (i = 0; i < num; offset++, i++) {
		/* Ok, do the async read-ahead now */
		new_page = read_swap_cache_async(swp_entry(swp_type(entry),
							   offset), vma, addr);
		if (!new_page)
			break;
		page_cache_release(new_page);
#ifdef CONFIG_NUMA
		/*
		 * Find the next applicable VMA for the NUMA policy.
		 */
		addr += PAGE_SIZE;
		if (addr == 0)
			vma = NULL;
		if (vma) {
			if (addr >= vma->vm_end) {
				vma = next_vma;
				next_vma = vma ? vma->vm_next : NULL;
			}
			if (vma && addr < vma->vm_start)
				vma = NULL;
		} else {
			if (next_vma && addr >= next_vma->vm_start) {
				vma = next_vma;
				next_vma = vma->vm_next;
			}
		}
#endif
	}
	lru_add_drain();	/* Push any new pages onto the LRU now */
}

/*
2042 2043 2044
 * 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
2045
 */
2046 2047 2048
static int do_swap_page(struct mm_struct *mm, struct vm_area_struct *vma,
		unsigned long address, pte_t *page_table, pmd_t *pmd,
		int write_access, pte_t orig_pte)
L
Linus Torvalds 已提交
2049
{
2050
	spinlock_t *ptl;
L
Linus Torvalds 已提交
2051
	struct page *page;
2052
	swp_entry_t entry;
L
Linus Torvalds 已提交
2053
	pte_t pte;
N
Nick Piggin 已提交
2054
	int ret = 0;
L
Linus Torvalds 已提交
2055

H
Hugh Dickins 已提交
2056
	if (!pte_unmap_same(mm, pmd, page_table, orig_pte))
2057
		goto out;
2058 2059

	entry = pte_to_swp_entry(orig_pte);
2060 2061 2062 2063
	if (is_migration_entry(entry)) {
		migration_entry_wait(mm, pmd, address);
		goto out;
	}
2064
	delayacct_set_flag(DELAYACCT_PF_SWAPIN);
L
Linus Torvalds 已提交
2065 2066
	page = lookup_swap_cache(entry);
	if (!page) {
2067
		grab_swap_token(); /* Contend for token _before_ read-in */
L
Linus Torvalds 已提交
2068 2069 2070 2071
 		swapin_readahead(entry, address, vma);
 		page = read_swap_cache_async(entry, vma, address);
		if (!page) {
			/*
2072 2073
			 * Back out if somebody else faulted in this pte
			 * while we released the pte lock.
L
Linus Torvalds 已提交
2074
			 */
2075
			page_table = pte_offset_map_lock(mm, pmd, address, &ptl);
L
Linus Torvalds 已提交
2076 2077
			if (likely(pte_same(*page_table, orig_pte)))
				ret = VM_FAULT_OOM;
2078
			delayacct_clear_flag(DELAYACCT_PF_SWAPIN);
2079
			goto unlock;
L
Linus Torvalds 已提交
2080 2081 2082 2083
		}

		/* Had to read the page from swap area: Major fault */
		ret = VM_FAULT_MAJOR;
2084
		count_vm_event(PGMAJFAULT);
L
Linus Torvalds 已提交
2085 2086 2087 2088
	}

	mark_page_accessed(page);
	lock_page(page);
2089
	delayacct_clear_flag(DELAYACCT_PF_SWAPIN);
L
Linus Torvalds 已提交
2090 2091

	/*
2092
	 * Back out if somebody else already faulted in this pte.
L
Linus Torvalds 已提交
2093
	 */
2094
	page_table = pte_offset_map_lock(mm, pmd, address, &ptl);
H
Hugh Dickins 已提交
2095
	if (unlikely(!pte_same(*page_table, orig_pte)))
2096 2097 2098 2099 2100
		goto out_nomap;

	if (unlikely(!PageUptodate(page))) {
		ret = VM_FAULT_SIGBUS;
		goto out_nomap;
L
Linus Torvalds 已提交
2101 2102 2103 2104
	}

	/* The page isn't present yet, go ahead with the fault. */

2105
	inc_mm_counter(mm, anon_rss);
L
Linus Torvalds 已提交
2106 2107 2108 2109 2110 2111 2112 2113 2114 2115
	pte = mk_pte(page, vma->vm_page_prot);
	if (write_access && can_share_swap_page(page)) {
		pte = maybe_mkwrite(pte_mkdirty(pte), vma);
		write_access = 0;
	}

	flush_icache_page(vma, page);
	set_pte_at(mm, address, page_table, pte);
	page_add_anon_rmap(page, vma, address);

2116 2117 2118 2119 2120
	swap_free(entry);
	if (vm_swap_full())
		remove_exclusive_swap_page(page);
	unlock_page(page);

L
Linus Torvalds 已提交
2121
	if (write_access) {
N
Nick Piggin 已提交
2122
		/* XXX: We could OR the do_wp_page code with this one? */
L
Linus Torvalds 已提交
2123
		if (do_wp_page(mm, vma, address,
N
Nick Piggin 已提交
2124
				page_table, pmd, ptl, pte) & VM_FAULT_OOM)
L
Linus Torvalds 已提交
2125 2126 2127 2128 2129 2130
			ret = VM_FAULT_OOM;
		goto out;
	}

	/* No need to invalidate - it was non-present before */
	update_mmu_cache(vma, address, pte);
2131
unlock:
2132
	pte_unmap_unlock(page_table, ptl);
L
Linus Torvalds 已提交
2133 2134
out:
	return ret;
2135
out_nomap:
2136
	pte_unmap_unlock(page_table, ptl);
2137 2138
	unlock_page(page);
	page_cache_release(page);
2139
	return ret;
L
Linus Torvalds 已提交
2140 2141 2142
}

/*
2143 2144 2145
 * 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
2146
 */
2147 2148 2149
static int do_anonymous_page(struct mm_struct *mm, struct vm_area_struct *vma,
		unsigned long address, pte_t *page_table, pmd_t *pmd,
		int write_access)
L
Linus Torvalds 已提交
2150
{
2151 2152
	struct page *page;
	spinlock_t *ptl;
L
Linus Torvalds 已提交
2153 2154
	pte_t entry;

N
Nick Piggin 已提交
2155 2156
	/* Allocate our own private page. */
	pte_unmap(page_table);
2157

N
Nick Piggin 已提交
2158 2159 2160 2161 2162
	if (unlikely(anon_vma_prepare(vma)))
		goto oom;
	page = alloc_zeroed_user_highpage_movable(vma, address);
	if (!page)
		goto oom;
2163

N
Nick Piggin 已提交
2164 2165
	entry = mk_pte(page, vma->vm_page_prot);
	entry = maybe_mkwrite(pte_mkdirty(entry), vma);
L
Linus Torvalds 已提交
2166

N
Nick Piggin 已提交
2167 2168 2169 2170 2171 2172
	page_table = pte_offset_map_lock(mm, pmd, address, &ptl);
	if (!pte_none(*page_table))
		goto release;
	inc_mm_counter(mm, anon_rss);
	lru_cache_add_active(page);
	page_add_new_anon_rmap(page, vma, address);
2173
	set_pte_at(mm, address, page_table, entry);
L
Linus Torvalds 已提交
2174 2175

	/* No need to invalidate - it was non-present before */
2176 2177
	update_mmu_cache(vma, address, entry);
unlock:
2178
	pte_unmap_unlock(page_table, ptl);
N
Nick Piggin 已提交
2179
	return 0;
2180 2181 2182
release:
	page_cache_release(page);
	goto unlock;
2183
oom:
L
Linus Torvalds 已提交
2184 2185 2186 2187
	return VM_FAULT_OOM;
}

/*
2188
 * __do_fault() tries to create a new page mapping. It aggressively
L
Linus Torvalds 已提交
2189
 * tries to share with existing pages, but makes a separate copy if
2190 2191
 * the FAULT_FLAG_WRITE is set in the flags parameter in order to avoid
 * the next page fault.
L
Linus Torvalds 已提交
2192 2193 2194 2195
 *
 * As this is called only for pages that do not currently exist, we
 * do not need to flush old virtual caches or the TLB.
 *
2196
 * We enter with non-exclusive mmap_sem (to exclude vma changes,
2197
 * but allow concurrent faults), and pte neither mapped nor locked.
2198
 * We return with mmap_sem still held, but pte unmapped and unlocked.
L
Linus Torvalds 已提交
2199
 */
2200
static int __do_fault(struct mm_struct *mm, struct vm_area_struct *vma,
2201
		unsigned long address, pmd_t *pmd,
2202
		pgoff_t pgoff, unsigned int flags, pte_t orig_pte)
L
Linus Torvalds 已提交
2203
{
2204
	pte_t *page_table;
2205
	spinlock_t *ptl;
N
Nick Piggin 已提交
2206
	struct page *page;
L
Linus Torvalds 已提交
2207 2208
	pte_t entry;
	int anon = 0;
2209
	struct page *dirty_page = NULL;
N
Nick Piggin 已提交
2210 2211
	struct vm_fault vmf;
	int ret;
2212
	int page_mkwrite = 0;
2213

N
Nick Piggin 已提交
2214 2215 2216 2217
	vmf.virtual_address = (void __user *)(address & PAGE_MASK);
	vmf.pgoff = pgoff;
	vmf.flags = flags;
	vmf.page = NULL;
L
Linus Torvalds 已提交
2218

2219 2220
	BUG_ON(vma->vm_flags & VM_PFNMAP);

2221
	if (likely(vma->vm_ops->fault)) {
N
Nick Piggin 已提交
2222
		ret = vma->vm_ops->fault(vma, &vmf);
N
Nick Piggin 已提交
2223 2224
		if (unlikely(ret & (VM_FAULT_ERROR | VM_FAULT_NOPAGE)))
			return ret;
2225 2226
	} else {
		/* Legacy ->nopage path */
N
Nick Piggin 已提交
2227
		ret = 0;
N
Nick Piggin 已提交
2228
		vmf.page = vma->vm_ops->nopage(vma, address & PAGE_MASK, &ret);
2229
		/* no page was available -- either SIGBUS or OOM */
N
Nick Piggin 已提交
2230
		if (unlikely(vmf.page == NOPAGE_SIGBUS))
2231
			return VM_FAULT_SIGBUS;
N
Nick Piggin 已提交
2232
		else if (unlikely(vmf.page == NOPAGE_OOM))
2233 2234
			return VM_FAULT_OOM;
	}
L
Linus Torvalds 已提交
2235

2236
	/*
N
Nick Piggin 已提交
2237
	 * For consistency in subsequent calls, make the faulted page always
2238 2239
	 * locked.
	 */
N
Nick Piggin 已提交
2240
	if (unlikely(!(ret & VM_FAULT_LOCKED)))
N
Nick Piggin 已提交
2241
		lock_page(vmf.page);
2242
	else
N
Nick Piggin 已提交
2243
		VM_BUG_ON(!PageLocked(vmf.page));
2244

L
Linus Torvalds 已提交
2245 2246 2247
	/*
	 * Should we do an early C-O-W break?
	 */
N
Nick Piggin 已提交
2248
	page = vmf.page;
2249
	if (flags & FAULT_FLAG_WRITE) {
2250
		if (!(vma->vm_flags & VM_SHARED)) {
2251
			anon = 1;
2252
			if (unlikely(anon_vma_prepare(vma))) {
N
Nick Piggin 已提交
2253
				ret = VM_FAULT_OOM;
2254
				goto out;
2255
			}
N
Nick Piggin 已提交
2256 2257
			page = alloc_page_vma(GFP_HIGHUSER_MOVABLE,
						vma, address);
2258
			if (!page) {
N
Nick Piggin 已提交
2259
				ret = VM_FAULT_OOM;
2260
				goto out;
2261
			}
N
Nick Piggin 已提交
2262
			copy_user_highpage(page, vmf.page, address, vma);
2263
		} else {
2264 2265
			/*
			 * If the page will be shareable, see if the backing
2266
			 * address space wants to know that the page is about
2267 2268
			 * to become writable
			 */
2269 2270 2271
			if (vma->vm_ops->page_mkwrite) {
				unlock_page(page);
				if (vma->vm_ops->page_mkwrite(vma, page) < 0) {
N
Nick Piggin 已提交
2272 2273
					ret = VM_FAULT_SIGBUS;
					anon = 1; /* no anon but release vmf.page */
2274 2275 2276
					goto out_unlocked;
				}
				lock_page(page);
N
Nick Piggin 已提交
2277 2278 2279 2280 2281 2282 2283 2284
				/*
				 * XXX: this is not quite right (racy vs
				 * invalidate) to unlock and relock the page
				 * like this, however a better fix requires
				 * reworking page_mkwrite locking API, which
				 * is better done later.
				 */
				if (!page->mapping) {
N
Nick Piggin 已提交
2285
					ret = 0;
N
Nick Piggin 已提交
2286 2287 2288
					anon = 1; /* no anon but release vmf.page */
					goto out;
				}
2289
				page_mkwrite = 1;
2290 2291
			}
		}
2292

L
Linus Torvalds 已提交
2293 2294
	}

2295
	page_table = pte_offset_map_lock(mm, pmd, address, &ptl);
L
Linus Torvalds 已提交
2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307

	/*
	 * This silly early PAGE_DIRTY setting removes a race
	 * due to the bad i386 page protection. But it's valid
	 * for other architectures too.
	 *
	 * Note that if write_access is true, we either now have
	 * an exclusive copy of the page, or this is a shared mapping,
	 * so we can make it writable and dirty to avoid having to
	 * handle that later.
	 */
	/* Only go through if we didn't race with anybody else... */
2308
	if (likely(pte_same(*page_table, orig_pte))) {
2309 2310
		flush_icache_page(vma, page);
		entry = mk_pte(page, vma->vm_page_prot);
2311
		if (flags & FAULT_FLAG_WRITE)
L
Linus Torvalds 已提交
2312 2313 2314
			entry = maybe_mkwrite(pte_mkdirty(entry), vma);
		set_pte_at(mm, address, page_table, entry);
		if (anon) {
2315 2316 2317
                        inc_mm_counter(mm, anon_rss);
                        lru_cache_add_active(page);
                        page_add_new_anon_rmap(page, vma, address);
2318
		} else {
2319
			inc_mm_counter(mm, file_rss);
2320
			page_add_file_rmap(page);
2321
			if (flags & FAULT_FLAG_WRITE) {
2322
				dirty_page = page;
2323 2324
				get_page(dirty_page);
			}
2325
		}
2326 2327 2328

		/* no need to invalidate: a not-present page won't be cached */
		update_mmu_cache(vma, address, entry);
L
Linus Torvalds 已提交
2329
	} else {
2330 2331 2332
		if (anon)
			page_cache_release(page);
		else
2333
			anon = 1; /* no anon but release faulted_page */
L
Linus Torvalds 已提交
2334 2335
	}

2336
	pte_unmap_unlock(page_table, ptl);
2337 2338

out:
N
Nick Piggin 已提交
2339
	unlock_page(vmf.page);
2340
out_unlocked:
2341
	if (anon)
N
Nick Piggin 已提交
2342
		page_cache_release(vmf.page);
2343
	else if (dirty_page) {
2344
		set_page_dirty_balance(dirty_page, page_mkwrite);
2345 2346
		put_page(dirty_page);
	}
2347

N
Nick Piggin 已提交
2348
	return ret;
2349
}
2350

2351 2352 2353 2354 2355
static int do_linear_fault(struct mm_struct *mm, struct vm_area_struct *vma,
		unsigned long address, pte_t *page_table, pmd_t *pmd,
		int write_access, pte_t orig_pte)
{
	pgoff_t pgoff = (((address & PAGE_MASK)
2356
			- vma->vm_start) >> PAGE_SHIFT) + vma->vm_pgoff;
2357 2358
	unsigned int flags = (write_access ? FAULT_FLAG_WRITE : 0);

2359 2360
	pte_unmap(page_table);
	return __do_fault(mm, vma, address, pmd, pgoff, flags, orig_pte);
2361 2362
}

L
Linus Torvalds 已提交
2363

J
Jes Sorensen 已提交
2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392
/*
 * do_no_pfn() tries to create a new page mapping for a page without
 * a struct_page backing it
 *
 * As this is called only for pages that do not currently exist, we
 * do not need to flush old virtual caches or the TLB.
 *
 * 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.
 *
 * It is expected that the ->nopfn handler always returns the same pfn
 * for a given virtual mapping.
 *
 * Mark this `noinline' to prevent it from bloating the main pagefault code.
 */
static noinline int do_no_pfn(struct mm_struct *mm, struct vm_area_struct *vma,
		     unsigned long address, pte_t *page_table, pmd_t *pmd,
		     int write_access)
{
	spinlock_t *ptl;
	pte_t entry;
	unsigned long pfn;

	pte_unmap(page_table);
	BUG_ON(!(vma->vm_flags & VM_PFNMAP));
	BUG_ON(is_cow_mapping(vma->vm_flags));

	pfn = vma->vm_ops->nopfn(vma, address & PAGE_MASK);
2393
	if (unlikely(pfn == NOPFN_OOM))
J
Jes Sorensen 已提交
2394
		return VM_FAULT_OOM;
2395
	else if (unlikely(pfn == NOPFN_SIGBUS))
J
Jes Sorensen 已提交
2396
		return VM_FAULT_SIGBUS;
2397
	else if (unlikely(pfn == NOPFN_REFAULT))
N
Nick Piggin 已提交
2398
		return 0;
J
Jes Sorensen 已提交
2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409

	page_table = pte_offset_map_lock(mm, pmd, address, &ptl);

	/* Only go through if we didn't race with anybody else... */
	if (pte_none(*page_table)) {
		entry = pfn_pte(pfn, vma->vm_page_prot);
		if (write_access)
			entry = maybe_mkwrite(pte_mkdirty(entry), vma);
		set_pte_at(mm, address, page_table, entry);
	}
	pte_unmap_unlock(page_table, ptl);
N
Nick Piggin 已提交
2410
	return 0;
J
Jes Sorensen 已提交
2411 2412
}

L
Linus Torvalds 已提交
2413 2414 2415 2416
/*
 * Fault of a previously existing named mapping. Repopulate the pte
 * from the encoded file_pte if possible. This enables swappable
 * nonlinear vmas.
2417 2418 2419 2420
 *
 * 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
2421
 */
N
Nick Piggin 已提交
2422
static int do_nonlinear_fault(struct mm_struct *mm, struct vm_area_struct *vma,
2423 2424
		unsigned long address, pte_t *page_table, pmd_t *pmd,
		int write_access, pte_t orig_pte)
L
Linus Torvalds 已提交
2425
{
N
Nick Piggin 已提交
2426 2427
	unsigned int flags = FAULT_FLAG_NONLINEAR |
				(write_access ? FAULT_FLAG_WRITE : 0);
2428
	pgoff_t pgoff;
L
Linus Torvalds 已提交
2429

H
Hugh Dickins 已提交
2430
	if (!pte_unmap_same(mm, pmd, page_table, orig_pte))
N
Nick Piggin 已提交
2431
		return 0;
L
Linus Torvalds 已提交
2432

N
Nick Piggin 已提交
2433 2434
	if (unlikely(!(vma->vm_flags & VM_NONLINEAR) ||
			!(vma->vm_flags & VM_CAN_NONLINEAR))) {
2435 2436 2437
		/*
		 * Page table corrupted: show pte and kill process.
		 */
N
Nick Piggin 已提交
2438
		print_bad_pte(vma, orig_pte, address);
2439 2440 2441 2442
		return VM_FAULT_OOM;
	}

	pgoff = pte_to_pgoff(orig_pte);
2443
	return __do_fault(mm, vma, address, pmd, pgoff, flags, orig_pte);
L
Linus Torvalds 已提交
2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454
}

/*
 * 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
2455 2456 2457
 * 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
2458 2459
 */
static inline int handle_pte_fault(struct mm_struct *mm,
2460 2461
		struct vm_area_struct *vma, unsigned long address,
		pte_t *pte, pmd_t *pmd, int write_access)
L
Linus Torvalds 已提交
2462 2463
{
	pte_t entry;
2464
	spinlock_t *ptl;
L
Linus Torvalds 已提交
2465

2466
	entry = *pte;
L
Linus Torvalds 已提交
2467
	if (!pte_present(entry)) {
2468
		if (pte_none(entry)) {
J
Jes Sorensen 已提交
2469
			if (vma->vm_ops) {
2470 2471 2472
				if (vma->vm_ops->fault || vma->vm_ops->nopage)
					return do_linear_fault(mm, vma, address,
						pte, pmd, write_access, entry);
J
Jes Sorensen 已提交
2473 2474 2475 2476 2477 2478
				if (unlikely(vma->vm_ops->nopfn))
					return do_no_pfn(mm, vma, address, pte,
							 pmd, write_access);
			}
			return do_anonymous_page(mm, vma, address,
						 pte, pmd, write_access);
2479
		}
L
Linus Torvalds 已提交
2480
		if (pte_file(entry))
N
Nick Piggin 已提交
2481
			return do_nonlinear_fault(mm, vma, address,
2482 2483 2484
					pte, pmd, write_access, entry);
		return do_swap_page(mm, vma, address,
					pte, pmd, write_access, entry);
L
Linus Torvalds 已提交
2485 2486
	}

H
Hugh Dickins 已提交
2487
	ptl = pte_lockptr(mm, pmd);
2488 2489 2490
	spin_lock(ptl);
	if (unlikely(!pte_same(*pte, entry)))
		goto unlock;
L
Linus Torvalds 已提交
2491 2492
	if (write_access) {
		if (!pte_write(entry))
2493 2494
			return do_wp_page(mm, vma, address,
					pte, pmd, ptl, entry);
L
Linus Torvalds 已提交
2495 2496 2497
		entry = pte_mkdirty(entry);
	}
	entry = pte_mkyoung(entry);
2498
	if (ptep_set_access_flags(vma, address, pte, entry, write_access)) {
2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509
		update_mmu_cache(vma, address, entry);
	} 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.
		 */
		if (write_access)
			flush_tlb_page(vma, address);
	}
2510 2511
unlock:
	pte_unmap_unlock(pte, ptl);
N
Nick Piggin 已提交
2512
	return 0;
L
Linus Torvalds 已提交
2513 2514 2515 2516 2517
}

/*
 * By the time we get here, we already hold the mm semaphore
 */
N
Nick Piggin 已提交
2518
int handle_mm_fault(struct mm_struct *mm, struct vm_area_struct *vma,
L
Linus Torvalds 已提交
2519 2520 2521 2522 2523 2524 2525 2526 2527
		unsigned long address, int write_access)
{
	pgd_t *pgd;
	pud_t *pud;
	pmd_t *pmd;
	pte_t *pte;

	__set_current_state(TASK_RUNNING);

2528
	count_vm_event(PGFAULT);
L
Linus Torvalds 已提交
2529

2530 2531
	if (unlikely(is_vm_hugetlb_page(vma)))
		return hugetlb_fault(mm, vma, address, write_access);
L
Linus Torvalds 已提交
2532 2533 2534 2535

	pgd = pgd_offset(mm, address);
	pud = pud_alloc(mm, pgd, address);
	if (!pud)
H
Hugh Dickins 已提交
2536
		return VM_FAULT_OOM;
L
Linus Torvalds 已提交
2537 2538
	pmd = pmd_alloc(mm, pud, address);
	if (!pmd)
H
Hugh Dickins 已提交
2539
		return VM_FAULT_OOM;
L
Linus Torvalds 已提交
2540 2541
	pte = pte_alloc_map(mm, pmd, address);
	if (!pte)
H
Hugh Dickins 已提交
2542
		return VM_FAULT_OOM;
L
Linus Torvalds 已提交
2543

H
Hugh Dickins 已提交
2544
	return handle_pte_fault(mm, vma, address, pte, pmd, write_access);
L
Linus Torvalds 已提交
2545 2546 2547 2548 2549
}

#ifndef __PAGETABLE_PUD_FOLDED
/*
 * Allocate page upper directory.
H
Hugh Dickins 已提交
2550
 * We've already handled the fast-path in-line.
L
Linus Torvalds 已提交
2551
 */
2552
int __pud_alloc(struct mm_struct *mm, pgd_t *pgd, unsigned long address)
L
Linus Torvalds 已提交
2553
{
H
Hugh Dickins 已提交
2554 2555
	pud_t *new = pud_alloc_one(mm, address);
	if (!new)
2556
		return -ENOMEM;
L
Linus Torvalds 已提交
2557

H
Hugh Dickins 已提交
2558
	spin_lock(&mm->page_table_lock);
2559
	if (pgd_present(*pgd))		/* Another has populated it */
L
Linus Torvalds 已提交
2560
		pud_free(new);
2561 2562
	else
		pgd_populate(mm, pgd, new);
H
Hugh Dickins 已提交
2563
	spin_unlock(&mm->page_table_lock);
2564
	return 0;
L
Linus Torvalds 已提交
2565 2566 2567 2568 2569 2570
}
#endif /* __PAGETABLE_PUD_FOLDED */

#ifndef __PAGETABLE_PMD_FOLDED
/*
 * Allocate page middle directory.
H
Hugh Dickins 已提交
2571
 * We've already handled the fast-path in-line.
L
Linus Torvalds 已提交
2572
 */
2573
int __pmd_alloc(struct mm_struct *mm, pud_t *pud, unsigned long address)
L
Linus Torvalds 已提交
2574
{
H
Hugh Dickins 已提交
2575 2576
	pmd_t *new = pmd_alloc_one(mm, address);
	if (!new)
2577
		return -ENOMEM;
L
Linus Torvalds 已提交
2578

H
Hugh Dickins 已提交
2579
	spin_lock(&mm->page_table_lock);
L
Linus Torvalds 已提交
2580
#ifndef __ARCH_HAS_4LEVEL_HACK
2581
	if (pud_present(*pud))		/* Another has populated it */
L
Linus Torvalds 已提交
2582
		pmd_free(new);
2583 2584
	else
		pud_populate(mm, pud, new);
L
Linus Torvalds 已提交
2585
#else
2586
	if (pgd_present(*pud))		/* Another has populated it */
L
Linus Torvalds 已提交
2587
		pmd_free(new);
2588 2589
	else
		pgd_populate(mm, pud, new);
L
Linus Torvalds 已提交
2590
#endif /* __ARCH_HAS_4LEVEL_HACK */
H
Hugh Dickins 已提交
2591
	spin_unlock(&mm->page_table_lock);
2592
	return 0;
2593
}
L
Linus Torvalds 已提交
2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604
#endif /* __PAGETABLE_PMD_FOLDED */

int make_pages_present(unsigned long addr, unsigned long end)
{
	int ret, len, write;
	struct vm_area_struct * vma;

	vma = find_vma(current->mm, addr);
	if (!vma)
		return -1;
	write = (vma->vm_flags & VM_WRITE) != 0;
2605 2606
	BUG_ON(addr >= end);
	BUG_ON(end > vma->vm_end);
2607
	len = DIV_ROUND_UP(end, PAGE_SIZE) - addr/PAGE_SIZE;
L
Linus Torvalds 已提交
2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657
	ret = get_user_pages(current, current->mm, addr,
			len, write, 0, NULL, NULL);
	if (ret < 0)
		return ret;
	return ret == len ? 0 : -1;
}

/* 
 * Map a vmalloc()-space virtual address to the physical page.
 */
struct page * vmalloc_to_page(void * vmalloc_addr)
{
	unsigned long addr = (unsigned long) vmalloc_addr;
	struct page *page = NULL;
	pgd_t *pgd = pgd_offset_k(addr);
	pud_t *pud;
	pmd_t *pmd;
	pte_t *ptep, pte;
  
	if (!pgd_none(*pgd)) {
		pud = pud_offset(pgd, addr);
		if (!pud_none(*pud)) {
			pmd = pmd_offset(pud, addr);
			if (!pmd_none(*pmd)) {
				ptep = pte_offset_map(pmd, addr);
				pte = *ptep;
				if (pte_present(pte))
					page = pte_page(pte);
				pte_unmap(ptep);
			}
		}
	}
	return page;
}

EXPORT_SYMBOL(vmalloc_to_page);

/*
 * Map a vmalloc()-space virtual address to the physical page frame number.
 */
unsigned long vmalloc_to_pfn(void * vmalloc_addr)
{
	return page_to_pfn(vmalloc_to_page(vmalloc_addr));
}

EXPORT_SYMBOL(vmalloc_to_pfn);

#if !defined(__HAVE_ARCH_GATE_AREA)

#if defined(AT_SYSINFO_EHDR)
2658
static struct vm_area_struct gate_vma;
L
Linus Torvalds 已提交
2659 2660 2661 2662 2663 2664

static int __init gate_vma_init(void)
{
	gate_vma.vm_mm = NULL;
	gate_vma.vm_start = FIXADDR_USER_START;
	gate_vma.vm_end = FIXADDR_USER_END;
R
Roland McGrath 已提交
2665 2666
	gate_vma.vm_flags = VM_READ | VM_MAYREAD | VM_EXEC | VM_MAYEXEC;
	gate_vma.vm_page_prot = __P101;
2667 2668 2669 2670 2671 2672 2673
	/*
	 * Make sure the vDSO gets into every core dump.
	 * Dumping its contents makes post-mortem fully interpretable later
	 * without matching up the same kernel and hardware config to see
	 * what PC values meant.
	 */
	gate_vma.vm_flags |= VM_ALWAYSDUMP;
L
Linus Torvalds 已提交
2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697
	return 0;
}
__initcall(gate_vma_init);
#endif

struct vm_area_struct *get_gate_vma(struct task_struct *tsk)
{
#ifdef AT_SYSINFO_EHDR
	return &gate_vma;
#else
	return NULL;
#endif
}

int in_gate_area_no_task(unsigned long addr)
{
#ifdef AT_SYSINFO_EHDR
	if ((addr >= FIXADDR_USER_START) && (addr < FIXADDR_USER_END))
		return 1;
#endif
	return 0;
}

#endif	/* __HAVE_ARCH_GATE_AREA */
2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715

/*
 * 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;
	struct vm_area_struct *vma;
	struct page *page;
	void *old_buf = buf;

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

	down_read(&mm->mmap_sem);
S
Simon Arlott 已提交
2716
	/* ignore errors, just check how much was successfully transferred */
2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750
	while (len) {
		int bytes, ret, offset;
		void *maddr;

		ret = get_user_pages(tsk, mm, addr, 1,
				write, 1, &page, &vma);
		if (ret <= 0)
			break;

		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);
		len -= bytes;
		buf += bytes;
		addr += bytes;
	}
	up_read(&mm->mmap_sem);
	mmput(mm);

	return buf - old_buf;
}