memory.c 72.9 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>
53
#include <linux/memcontrol.h>
L
Linus Torvalds 已提交
54 55 56 57 58 59 60 61 62 63

#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
64
#ifndef CONFIG_NEED_MULTIPLE_NODES
L
Linus Torvalds 已提交
65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85
/* 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);

86 87 88 89 90 91 92 93 94 95 96 97
/*
 * Randomize the address space (stacks, mmaps, brk, etc.).
 *
 * ( When CONFIG_COMPAT_BRK=y we exclude brk from randomization,
 *   as ancient (libc5 based) binaries can segfault. )
 */
int randomize_va_space __read_mostly =
#ifdef CONFIG_COMPAT_BRK
					1;
#else
					2;
#endif
98 99 100 101

static int __init disable_randmaps(char *s)
{
	randomize_va_space = 0;
102
	return 1;
103 104 105 106
}
__setup("norandmaps", disable_randmaps);


L
Linus Torvalds 已提交
107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134
/*
 * 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.
 */
135
static void free_pte_range(struct mmu_gather *tlb, pmd_t *pmd)
L
Linus Torvalds 已提交
136
{
137
	pgtable_t token = pmd_pgtable(*pmd);
138
	pmd_clear(pmd);
139
	pte_free_tlb(tlb, token);
140
	tlb->mm->nr_ptes--;
L
Linus Torvalds 已提交
141 142
}

143 144 145
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
146 147 148
{
	pmd_t *pmd;
	unsigned long next;
149
	unsigned long start;
L
Linus Torvalds 已提交
150

151
	start = addr;
L
Linus Torvalds 已提交
152 153 154 155 156
	pmd = pmd_offset(pud, addr);
	do {
		next = pmd_addr_end(addr, end);
		if (pmd_none_or_clear_bad(pmd))
			continue;
157
		free_pte_range(tlb, pmd);
L
Linus Torvalds 已提交
158 159
	} while (pmd++, addr = next, addr != end);

160 161 162 163 164 165 166
	start &= PUD_MASK;
	if (start < floor)
		return;
	if (ceiling) {
		ceiling &= PUD_MASK;
		if (!ceiling)
			return;
L
Linus Torvalds 已提交
167
	}
168 169 170 171 172 173
	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
174 175
}

176 177 178
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
179 180 181
{
	pud_t *pud;
	unsigned long next;
182
	unsigned long start;
L
Linus Torvalds 已提交
183

184
	start = addr;
L
Linus Torvalds 已提交
185 186 187 188 189
	pud = pud_offset(pgd, addr);
	do {
		next = pud_addr_end(addr, end);
		if (pud_none_or_clear_bad(pud))
			continue;
190
		free_pmd_range(tlb, pud, addr, next, floor, ceiling);
L
Linus Torvalds 已提交
191 192
	} while (pud++, addr = next, addr != end);

193 194 195 196 197 198 199
	start &= PGDIR_MASK;
	if (start < floor)
		return;
	if (ceiling) {
		ceiling &= PGDIR_MASK;
		if (!ceiling)
			return;
L
Linus Torvalds 已提交
200
	}
201 202 203 204 205 206
	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
207 208 209
}

/*
210 211
 * This function frees user-level page tables of a process.
 *
L
Linus Torvalds 已提交
212 213
 * Must be called with pagetable lock held.
 */
214
void free_pgd_range(struct mmu_gather **tlb,
215 216
			unsigned long addr, unsigned long end,
			unsigned long floor, unsigned long ceiling)
L
Linus Torvalds 已提交
217 218 219
{
	pgd_t *pgd;
	unsigned long next;
220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246
	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
247

248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264
	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;
265
	pgd = pgd_offset((*tlb)->mm, addr);
L
Linus Torvalds 已提交
266 267 268 269
	do {
		next = pgd_addr_end(addr, end);
		if (pgd_none_or_clear_bad(pgd))
			continue;
270
		free_pud_range(*tlb, pgd, addr, next, floor, ceiling);
L
Linus Torvalds 已提交
271
	} while (pgd++, addr = next, addr != end);
272 273 274
}

void free_pgtables(struct mmu_gather **tlb, struct vm_area_struct *vma,
275
		unsigned long floor, unsigned long ceiling)
276 277 278 279 280
{
	while (vma) {
		struct vm_area_struct *next = vma->vm_next;
		unsigned long addr = vma->vm_start;

281 282 283 284 285 286
		/*
		 * Hide vma from rmap and vmtruncate before freeing pgtables
		 */
		anon_vma_unlink(vma);
		unlink_file_vma(vma);

287
		if (is_vm_hugetlb_page(vma)) {
288
			hugetlb_free_pgd_range(tlb, addr, vma->vm_end,
289
				floor, next? next->vm_start: ceiling);
290 291 292 293 294
		} else {
			/*
			 * Optimization: gather nearby vmas into one call down
			 */
			while (next && next->vm_start <= vma->vm_end + PMD_SIZE
295
			       && !is_vm_hugetlb_page(next)) {
296 297
				vma = next;
				next = vma->vm_next;
298 299
				anon_vma_unlink(vma);
				unlink_file_vma(vma);
300 301 302 303
			}
			free_pgd_range(tlb, addr, vma->vm_end,
				floor, next? next->vm_start: ceiling);
		}
304 305
		vma = next;
	}
L
Linus Torvalds 已提交
306 307
}

308
int __pte_alloc(struct mm_struct *mm, pmd_t *pmd, unsigned long address)
L
Linus Torvalds 已提交
309
{
310
	pgtable_t new = pte_alloc_one(mm, address);
311 312 313
	if (!new)
		return -ENOMEM;

H
Hugh Dickins 已提交
314
	spin_lock(&mm->page_table_lock);
315
	if (!pmd_present(*pmd)) {	/* Has another populated it ? */
L
Linus Torvalds 已提交
316 317
		mm->nr_ptes++;
		pmd_populate(mm, pmd, new);
318
		new = NULL;
L
Linus Torvalds 已提交
319
	}
H
Hugh Dickins 已提交
320
	spin_unlock(&mm->page_table_lock);
321 322
	if (new)
		pte_free(mm, new);
323
	return 0;
L
Linus Torvalds 已提交
324 325
}

326
int __pte_alloc_kernel(pmd_t *pmd, unsigned long address)
L
Linus Torvalds 已提交
327
{
328 329 330 331 332
	pte_t *new = pte_alloc_one_kernel(&init_mm, address);
	if (!new)
		return -ENOMEM;

	spin_lock(&init_mm.page_table_lock);
333
	if (!pmd_present(*pmd)) {	/* Has another populated it ? */
334
		pmd_populate_kernel(&init_mm, pmd, new);
335 336
		new = NULL;
	}
337
	spin_unlock(&init_mm.page_table_lock);
338 339
	if (new)
		pte_free_kernel(&init_mm, new);
340
	return 0;
L
Linus Torvalds 已提交
341 342
}

343 344 345 346 347 348 349 350
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
351
/*
352 353 354
 * 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
355 356 357 358 359 360 361 362 363 364 365 366 367
 *
 * 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();
}

368 369 370 371 372
static inline int is_cow_mapping(unsigned int flags)
{
	return (flags & (VM_SHARED | VM_MAYWRITE)) == VM_MAYWRITE;
}

H
Hugh Dickins 已提交
373
/*
374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391
 * 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
392
 */
393
struct page *vm_normal_page(struct vm_area_struct *vma, unsigned long addr, pte_t pte)
H
Hugh Dickins 已提交
394
{
395 396
	unsigned long pfn = pte_pfn(pte);

N
Nick Piggin 已提交
397
	if (unlikely(vma->vm_flags & VM_PFNMAP)) {
398 399 400
		unsigned long off = (addr - vma->vm_start) >> PAGE_SHIFT;
		if (pfn == vma->vm_pgoff + off)
			return NULL;
401
		if (!is_cow_mapping(vma->vm_flags))
L
Linus Torvalds 已提交
402
			return NULL;
403 404
	}

405
#ifdef CONFIG_DEBUG_VM
406 407 408 409 410 411
	/*
	 * 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.
	 */
412 413 414 415
	if (unlikely(!pfn_valid(pfn))) {
		print_bad_pte(vma, pte, addr);
		return NULL;
	}
416
#endif
417 418 419 420 421 422 423 424 425

	/*
	 * 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
426 427
}

L
Linus Torvalds 已提交
428 429 430 431 432 433
/*
 * 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
434
static inline void
L
Linus Torvalds 已提交
435
copy_one_pte(struct mm_struct *dst_mm, struct mm_struct *src_mm,
N
Nick Piggin 已提交
436
		pte_t *dst_pte, pte_t *src_pte, struct vm_area_struct *vma,
H
Hugh Dickins 已提交
437
		unsigned long addr, int *rss)
L
Linus Torvalds 已提交
438
{
N
Nick Piggin 已提交
439
	unsigned long vm_flags = vma->vm_flags;
L
Linus Torvalds 已提交
440 441 442 443 444 445
	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)) {
446 447 448
			swp_entry_t entry = pte_to_swp_entry(pte);

			swap_duplicate(entry);
L
Linus Torvalds 已提交
449 450 451
			/* make sure dst_mm is on swapoff's mmlist. */
			if (unlikely(list_empty(&dst_mm->mmlist))) {
				spin_lock(&mmlist_lock);
452 453 454
				if (list_empty(&dst_mm->mmlist))
					list_add(&dst_mm->mmlist,
						 &src_mm->mmlist);
L
Linus Torvalds 已提交
455 456
				spin_unlock(&mmlist_lock);
			}
457 458 459 460 461 462 463 464 465 466
			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
467
		}
468
		goto out_set_pte;
L
Linus Torvalds 已提交
469 470 471 472 473 474
	}

	/*
	 * If it's a COW mapping, write protect it both
	 * in the parent and the child
	 */
475
	if (is_cow_mapping(vm_flags)) {
L
Linus Torvalds 已提交
476
		ptep_set_wrprotect(src_mm, addr, src_pte);
477
		pte = pte_wrprotect(pte);
L
Linus Torvalds 已提交
478 479 480 481 482 483 484 485 486
	}

	/*
	 * 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);
487 488 489 490

	page = vm_normal_page(vma, addr, pte);
	if (page) {
		get_page(page);
N
Nick Piggin 已提交
491
		page_dup_rmap(page, vma, addr);
492 493
		rss[!!PageAnon(page)]++;
	}
494 495 496

out_set_pte:
	set_pte_at(dst_mm, addr, dst_pte, pte);
L
Linus Torvalds 已提交
497 498 499 500 501 502 503
}

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
504
	spinlock_t *src_ptl, *dst_ptl;
505
	int progress = 0;
H
Hugh Dickins 已提交
506
	int rss[2];
L
Linus Torvalds 已提交
507 508

again:
509
	rss[1] = rss[0] = 0;
H
Hugh Dickins 已提交
510
	dst_pte = pte_alloc_map_lock(dst_mm, dst_pmd, addr, &dst_ptl);
L
Linus Torvalds 已提交
511 512 513
	if (!dst_pte)
		return -ENOMEM;
	src_pte = pte_offset_map_nested(src_pmd, addr);
H
Hugh Dickins 已提交
514
	src_ptl = pte_lockptr(src_mm, src_pmd);
I
Ingo Molnar 已提交
515
	spin_lock_nested(src_ptl, SINGLE_DEPTH_NESTING);
516
	arch_enter_lazy_mmu_mode();
L
Linus Torvalds 已提交
517 518 519 520 521 522

	do {
		/*
		 * We are holding two locks at this point - either of them
		 * could generate latencies in another task on another CPU.
		 */
523 524 525
		if (progress >= 32) {
			progress = 0;
			if (need_resched() ||
N
Nick Piggin 已提交
526
			    spin_needbreak(src_ptl) || spin_needbreak(dst_ptl))
527 528
				break;
		}
L
Linus Torvalds 已提交
529 530 531 532
		if (pte_none(*src_pte)) {
			progress++;
			continue;
		}
H
Hugh Dickins 已提交
533
		copy_one_pte(dst_mm, src_mm, dst_pte, src_pte, vma, addr, rss);
L
Linus Torvalds 已提交
534 535 536
		progress += 8;
	} while (dst_pte++, src_pte++, addr += PAGE_SIZE, addr != end);

537
	arch_leave_lazy_mmu_mode();
H
Hugh Dickins 已提交
538
	spin_unlock(src_ptl);
L
Linus Torvalds 已提交
539
	pte_unmap_nested(src_pte - 1);
540
	add_mm_rss(dst_mm, rss[0], rss[1]);
H
Hugh Dickins 已提交
541 542
	pte_unmap_unlock(dst_pte - 1, dst_ptl);
	cond_resched();
L
Linus Torvalds 已提交
543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599
	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;

600 601 602 603 604 605
	/*
	 * 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.
	 */
606
	if (!(vma->vm_flags & (VM_HUGETLB|VM_NONLINEAR|VM_PFNMAP|VM_INSERTPAGE))) {
607 608 609 610
		if (!vma->anon_vma)
			return 0;
	}

L
Linus Torvalds 已提交
611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626
	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;
}

627
static unsigned long zap_pte_range(struct mmu_gather *tlb,
N
Nick Piggin 已提交
628
				struct vm_area_struct *vma, pmd_t *pmd,
L
Linus Torvalds 已提交
629
				unsigned long addr, unsigned long end,
630
				long *zap_work, struct zap_details *details)
L
Linus Torvalds 已提交
631
{
N
Nick Piggin 已提交
632
	struct mm_struct *mm = tlb->mm;
L
Linus Torvalds 已提交
633
	pte_t *pte;
634
	spinlock_t *ptl;
635 636
	int file_rss = 0;
	int anon_rss = 0;
L
Linus Torvalds 已提交
637

638
	pte = pte_offset_map_lock(mm, pmd, addr, &ptl);
639
	arch_enter_lazy_mmu_mode();
L
Linus Torvalds 已提交
640 641
	do {
		pte_t ptent = *pte;
642 643
		if (pte_none(ptent)) {
			(*zap_work)--;
L
Linus Torvalds 已提交
644
			continue;
645
		}
646 647 648

		(*zap_work) -= PAGE_SIZE;

L
Linus Torvalds 已提交
649
		if (pte_present(ptent)) {
H
Hugh Dickins 已提交
650
			struct page *page;
651

652
			page = vm_normal_page(vma, addr, ptent);
L
Linus Torvalds 已提交
653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670
			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
671
			ptent = ptep_get_and_clear_full(mm, addr, pte,
672
							tlb->fullmm);
L
Linus Torvalds 已提交
673 674 675 676 677 678
			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
679
				set_pte_at(mm, addr, pte,
L
Linus Torvalds 已提交
680 681
					   pgoff_to_pte(page->index));
			if (PageAnon(page))
H
Hugh Dickins 已提交
682
				anon_rss--;
683 684 685 686
			else {
				if (pte_dirty(ptent))
					set_page_dirty(page);
				if (pte_young(ptent))
687
					SetPageReferenced(page);
H
Hugh Dickins 已提交
688
				file_rss--;
689
			}
N
Nick Piggin 已提交
690
			page_remove_rmap(page, vma);
L
Linus Torvalds 已提交
691 692 693 694 695 696 697 698 699 700 701
			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));
702
		pte_clear_not_present_full(mm, addr, pte, tlb->fullmm);
703
	} while (pte++, addr += PAGE_SIZE, (addr != end && *zap_work > 0));
704

H
Hugh Dickins 已提交
705
	add_mm_rss(mm, file_rss, anon_rss);
706
	arch_leave_lazy_mmu_mode();
707
	pte_unmap_unlock(pte - 1, ptl);
708 709

	return addr;
L
Linus Torvalds 已提交
710 711
}

712
static inline unsigned long zap_pmd_range(struct mmu_gather *tlb,
N
Nick Piggin 已提交
713
				struct vm_area_struct *vma, pud_t *pud,
L
Linus Torvalds 已提交
714
				unsigned long addr, unsigned long end,
715
				long *zap_work, struct zap_details *details)
L
Linus Torvalds 已提交
716 717 718 719 720 721 722
{
	pmd_t *pmd;
	unsigned long next;

	pmd = pmd_offset(pud, addr);
	do {
		next = pmd_addr_end(addr, end);
723 724
		if (pmd_none_or_clear_bad(pmd)) {
			(*zap_work)--;
L
Linus Torvalds 已提交
725
			continue;
726 727 728 729 730 731
		}
		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
732 733
}

734
static inline unsigned long zap_pud_range(struct mmu_gather *tlb,
N
Nick Piggin 已提交
735
				struct vm_area_struct *vma, pgd_t *pgd,
L
Linus Torvalds 已提交
736
				unsigned long addr, unsigned long end,
737
				long *zap_work, struct zap_details *details)
L
Linus Torvalds 已提交
738 739 740 741 742 743 744
{
	pud_t *pud;
	unsigned long next;

	pud = pud_offset(pgd, addr);
	do {
		next = pud_addr_end(addr, end);
745 746
		if (pud_none_or_clear_bad(pud)) {
			(*zap_work)--;
L
Linus Torvalds 已提交
747
			continue;
748 749 750 751 752 753
		}
		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
754 755
}

756 757
static unsigned long unmap_page_range(struct mmu_gather *tlb,
				struct vm_area_struct *vma,
L
Linus Torvalds 已提交
758
				unsigned long addr, unsigned long end,
759
				long *zap_work, struct zap_details *details)
L
Linus Torvalds 已提交
760 761 762 763 764 765 766 767 768 769 770 771
{
	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);
772 773
		if (pgd_none_or_clear_bad(pgd)) {
			(*zap_work)--;
L
Linus Torvalds 已提交
774
			continue;
775 776 777 778
		}
		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
779
	tlb_end_vma(tlb, vma);
780 781

	return addr;
L
Linus Torvalds 已提交
782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799
}

#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
 *
800
 * Returns the end address of the unmapping (restart addr if interrupted).
L
Linus Torvalds 已提交
801
 *
802
 * Unmap all pages in the vma list.
L
Linus Torvalds 已提交
803
 *
804 805
 * 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
806 807 808 809 810 811 812 813 814 815 816
 * 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.
 */
817
unsigned long unmap_vmas(struct mmu_gather **tlbp,
L
Linus Torvalds 已提交
818 819 820 821
		struct vm_area_struct *vma, unsigned long start_addr,
		unsigned long end_addr, unsigned long *nr_accounted,
		struct zap_details *details)
{
822
	long zap_work = ZAP_BLOCK_SIZE;
L
Linus Torvalds 已提交
823 824
	unsigned long tlb_start = 0;	/* For tlb_finish_mmu */
	int tlb_start_valid = 0;
825
	unsigned long start = start_addr;
L
Linus Torvalds 已提交
826
	spinlock_t *i_mmap_lock = details? details->i_mmap_lock: NULL;
827
	int fullmm = (*tlbp)->fullmm;
L
Linus Torvalds 已提交
828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847

	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;
			}

848
			if (unlikely(is_vm_hugetlb_page(vma))) {
L
Linus Torvalds 已提交
849
				unmap_hugepage_range(vma, start, end);
850 851 852 853 854 855 856 857 858 859
				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
860 861 862 863 864
			}

			tlb_finish_mmu(*tlbp, tlb_start, start);

			if (need_resched() ||
N
Nick Piggin 已提交
865
				(i_mmap_lock && spin_needbreak(i_mmap_lock))) {
L
Linus Torvalds 已提交
866
				if (i_mmap_lock) {
867
					*tlbp = NULL;
L
Linus Torvalds 已提交
868 869 870 871 872
					goto out;
				}
				cond_resched();
			}

873
			*tlbp = tlb_gather_mmu(vma->vm_mm, fullmm);
L
Linus Torvalds 已提交
874
			tlb_start_valid = 0;
875
			zap_work = ZAP_BLOCK_SIZE;
L
Linus Torvalds 已提交
876 877 878
		}
	}
out:
879
	return start;	/* which is now the end (or restart) address */
L
Linus Torvalds 已提交
880 881 882 883 884 885 886 887 888
}

/**
 * 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
 */
889
unsigned long zap_page_range(struct vm_area_struct *vma, unsigned long address,
L
Linus Torvalds 已提交
890 891 892 893 894 895 896 897 898
		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);
899
	update_hiwater_rss(mm);
900 901 902
	end = unmap_vmas(&tlb, vma, address, end, &nr_accounted, details);
	if (tlb)
		tlb_finish_mmu(tlb, address, end);
903
	return end;
L
Linus Torvalds 已提交
904 905 906 907 908
}

/*
 * Do a quick page-table lookup for a single page.
 */
909
struct page *follow_page(struct vm_area_struct *vma, unsigned long address,
910
			unsigned int flags)
L
Linus Torvalds 已提交
911 912 913 914 915
{
	pgd_t *pgd;
	pud_t *pud;
	pmd_t *pmd;
	pte_t *ptep, pte;
916
	spinlock_t *ptl;
L
Linus Torvalds 已提交
917
	struct page *page;
918
	struct mm_struct *mm = vma->vm_mm;
L
Linus Torvalds 已提交
919

920 921 922 923 924
	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
925

926
	page = NULL;
L
Linus Torvalds 已提交
927 928
	pgd = pgd_offset(mm, address);
	if (pgd_none(*pgd) || unlikely(pgd_bad(*pgd)))
929
		goto no_page_table;
L
Linus Torvalds 已提交
930 931 932

	pud = pud_offset(pgd, address);
	if (pud_none(*pud) || unlikely(pud_bad(*pud)))
933
		goto no_page_table;
L
Linus Torvalds 已提交
934 935 936
	
	pmd = pmd_offset(pud, address);
	if (pmd_none(*pmd) || unlikely(pmd_bad(*pmd)))
937 938 939 940 941
		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
942
		goto out;
943
	}
L
Linus Torvalds 已提交
944

945
	ptep = pte_offset_map_lock(mm, pmd, address, &ptl);
L
Linus Torvalds 已提交
946 947 948 949
	if (!ptep)
		goto out;

	pte = *ptep;
950 951 952 953
	if (!pte_present(pte))
		goto unlock;
	if ((flags & FOLL_WRITE) && !pte_write(pte))
		goto unlock;
954 955
	page = vm_normal_page(vma, address, pte);
	if (unlikely(!page))
956
		goto unlock;
L
Linus Torvalds 已提交
957

958 959 960 961 962 963 964 965 966 967
	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
968
out:
969
	return page;
L
Linus Torvalds 已提交
970

971 972 973 974 975 976
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
977
		page = ZERO_PAGE(0);
978 979 980 981 982
		if (flags & FOLL_GET)
			get_page(page);
		BUG_ON(flags & FOLL_WRITE);
	}
	return page;
L
Linus Torvalds 已提交
983 984 985 986 987 988 989
}

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;
990
	unsigned int vm_flags;
L
Linus Torvalds 已提交
991

992 993
	if (len <= 0)
		return 0;
L
Linus Torvalds 已提交
994 995 996 997
	/* 
	 * Require read or write permissions.
	 * If 'force' is set, we only require the "MAY" flags.
	 */
998 999
	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
1000 1001 1002
	i = 0;

	do {
1003 1004
		struct vm_area_struct *vma;
		unsigned int foll_flags;
L
Linus Torvalds 已提交
1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023

		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);
1024 1025
			if (pmd_none(*pmd))
				return i ? : -EFAULT;
L
Linus Torvalds 已提交
1026
			pte = pte_offset_map(pmd, pg);
1027 1028 1029 1030
			if (pte_none(*pte)) {
				pte_unmap(pte);
				return i ? : -EFAULT;
			}
L
Linus Torvalds 已提交
1031
			if (pages) {
1032
				struct page *page = vm_normal_page(gate_vma, start, *pte);
1033 1034 1035
				pages[i] = page;
				if (page)
					get_page(page);
L
Linus Torvalds 已提交
1036 1037 1038 1039 1040 1041 1042 1043 1044 1045
			}
			pte_unmap(pte);
			if (vmas)
				vmas[i] = gate_vma;
			i++;
			start += PAGE_SIZE;
			len--;
			continue;
		}

1046
		if (!vma || (vma->vm_flags & (VM_IO | VM_PFNMAP))
1047
				|| !(vm_flags & vma->vm_flags))
L
Linus Torvalds 已提交
1048 1049 1050 1051
			return i ? : -EFAULT;

		if (is_vm_hugetlb_page(vma)) {
			i = follow_hugetlb_page(mm, vma, pages, vmas,
1052
						&start, &len, i, write);
L
Linus Torvalds 已提交
1053 1054
			continue;
		}
1055 1056 1057 1058 1059

		foll_flags = FOLL_TOUCH;
		if (pages)
			foll_flags |= FOLL_GET;
		if (!write && !(vma->vm_flags & VM_LOCKED) &&
1060 1061
		    (!vma->vm_ops || (!vma->vm_ops->nopage &&
					!vma->vm_ops->fault)))
1062 1063
			foll_flags |= FOLL_ANON;

L
Linus Torvalds 已提交
1064
		do {
1065
			struct page *page;
L
Linus Torvalds 已提交
1066

1067 1068 1069 1070 1071 1072 1073 1074
			/*
			 * 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;

1075 1076
			if (write)
				foll_flags |= FOLL_WRITE;
1077

1078
			cond_resched();
1079
			while (!(page = follow_page(vma, start, foll_flags))) {
1080
				int ret;
N
Nick Piggin 已提交
1081
				ret = handle_mm_fault(mm, vma, start,
1082
						foll_flags & FOLL_WRITE);
N
Nick Piggin 已提交
1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094
				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++;

1095
				/*
N
Nick Piggin 已提交
1096 1097 1098 1099 1100
				 * 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.
1101 1102
				 */
				if (ret & VM_FAULT_WRITE)
1103
					foll_flags &= ~FOLL_WRITE;
N
Nick Piggin 已提交
1104

1105
				cond_resched();
L
Linus Torvalds 已提交
1106 1107
			}
			if (pages) {
1108
				pages[i] = page;
1109

1110
				flush_anon_page(vma, page, start);
1111
				flush_dcache_page(page);
L
Linus Torvalds 已提交
1112 1113 1114 1115 1116 1117
			}
			if (vmas)
				vmas[i] = vma;
			i++;
			start += PAGE_SIZE;
			len--;
1118 1119
		} while (len && start < vma->vm_end);
	} while (len);
L
Linus Torvalds 已提交
1120 1121 1122 1123
	return i;
}
EXPORT_SYMBOL(get_user_pages);

H
Harvey Harrison 已提交
1124 1125
pte_t *get_locked_pte(struct mm_struct *mm, unsigned long addr,
			spinlock_t **ptl)
1126 1127 1128 1129
{
	pgd_t * pgd = pgd_offset(mm, addr);
	pud_t * pud = pud_alloc(mm, pgd, addr);
	if (pud) {
1130
		pmd_t * pmd = pmd_alloc(mm, pud, addr);
1131 1132 1133 1134 1135 1136
		if (pmd)
			return pte_alloc_map_lock(mm, pmd, addr, ptl);
	}
	return NULL;
}

1137 1138 1139 1140 1141 1142 1143 1144 1145 1146
/*
 * 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;
1147
	pte_t *pte;
1148 1149
	spinlock_t *ptl;

1150
	retval = mem_cgroup_charge(page, mm, GFP_KERNEL);
1151 1152
	if (retval)
		goto out;
1153 1154

	retval = -EINVAL;
1155
	if (PageAnon(page))
1156
		goto out_uncharge;
1157 1158
	retval = -ENOMEM;
	flush_dcache_page(page);
1159
	pte = get_locked_pte(mm, addr, &ptl);
1160
	if (!pte)
1161
		goto out_uncharge;
1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172
	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;
1173 1174
	pte_unmap_unlock(pte, ptl);
	return retval;
1175 1176
out_unlock:
	pte_unmap_unlock(pte, ptl);
1177 1178
out_uncharge:
	mem_cgroup_uncharge_page(page);
1179 1180 1181 1182
out:
	return retval;
}

1183 1184 1185 1186 1187 1188
/**
 * 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
 *
1189 1190 1191 1192 1193 1194
 * 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
1195
 * (see split_page()).
1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210
 *
 * 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;
1211
	vma->vm_flags |= VM_INSERTPAGE;
1212 1213
	return insert_page(vma->vm_mm, addr, page, vma->vm_page_prot);
}
1214
EXPORT_SYMBOL(vm_insert_page);
1215

N
Nick Piggin 已提交
1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260
/**
 * 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
1261 1262 1263 1264 1265 1266 1267 1268 1269 1270
/*
 * 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
1271
	spinlock_t *ptl;
L
Linus Torvalds 已提交
1272

H
Hugh Dickins 已提交
1273
	pte = pte_alloc_map_lock(mm, pmd, addr, &ptl);
L
Linus Torvalds 已提交
1274 1275
	if (!pte)
		return -ENOMEM;
1276
	arch_enter_lazy_mmu_mode();
L
Linus Torvalds 已提交
1277 1278
	do {
		BUG_ON(!pte_none(*pte));
N
Nick Piggin 已提交
1279
		set_pte_at(mm, addr, pte, pfn_pte(pfn, prot));
L
Linus Torvalds 已提交
1280 1281
		pfn++;
	} while (pte++, addr += PAGE_SIZE, addr != end);
1282
	arch_leave_lazy_mmu_mode();
H
Hugh Dickins 已提交
1283
	pte_unmap_unlock(pte - 1, ptl);
L
Linus Torvalds 已提交
1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326
	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;
}

1327 1328 1329 1330 1331 1332 1333 1334 1335 1336
/**
 * 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
1337 1338 1339 1340 1341
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;
1342
	unsigned long end = addr + PAGE_ALIGN(size);
L
Linus Torvalds 已提交
1343 1344 1345 1346 1347 1348 1349 1350
	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
1351 1352 1353 1354 1355
	 *   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.
1356 1357 1358
	 *   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
1359 1360 1361 1362
	 *
	 * 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
1363
	 */
1364
	if (is_cow_mapping(vma->vm_flags)) {
L
Linus Torvalds 已提交
1365
		if (addr != vma->vm_start || end != vma->vm_end)
1366
			return -EINVAL;
L
Linus Torvalds 已提交
1367 1368 1369
		vma->vm_pgoff = pfn;
	}

1370
	vma->vm_flags |= VM_IO | VM_RESERVED | VM_PFNMAP;
L
Linus Torvalds 已提交
1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386

	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);

1387 1388 1389 1390 1391 1392
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;
1393
	pgtable_t token;
1394
	spinlock_t *uninitialized_var(ptl);
1395 1396 1397 1398 1399 1400 1401 1402 1403

	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));

1404
	token = pmd_pgtable(*pmd);
1405 1406

	do {
1407
		err = fn(pte, token, addr, data);
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 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480
		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);

1481 1482 1483 1484 1485 1486 1487 1488 1489
/*
 * 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
1490
static inline int pte_unmap_same(struct mm_struct *mm, pmd_t *pmd,
1491 1492 1493 1494 1495
				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
1496 1497
		spinlock_t *ptl = pte_lockptr(mm, pmd);
		spin_lock(ptl);
1498
		same = pte_same(*page_table, orig_pte);
H
Hugh Dickins 已提交
1499
		spin_unlock(ptl);
1500 1501 1502 1503 1504 1505
	}
#endif
	pte_unmap(page_table);
	return same;
}

L
Linus Torvalds 已提交
1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518
/*
 * 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;
}

1519
static inline void cow_user_page(struct page *dst, struct page *src, unsigned long va, struct vm_area_struct *vma)
1520 1521 1522 1523 1524 1525 1526 1527 1528
{
	/*
	 * 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
1529 1530 1531 1532 1533 1534 1535 1536 1537
		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))
1538 1539
			memset(kaddr, 0, PAGE_SIZE);
		kunmap_atomic(kaddr, KM_USER0);
1540
		flush_dcache_page(dst);
N
Nick Piggin 已提交
1541 1542
	} else
		copy_user_highpage(dst, src, va, vma);
1543 1544
}

L
Linus Torvalds 已提交
1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558
/*
 * 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.
 *
1559 1560 1561
 * 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
1562
 */
1563 1564
static int do_wp_page(struct mm_struct *mm, struct vm_area_struct *vma,
		unsigned long address, pte_t *page_table, pmd_t *pmd,
1565
		spinlock_t *ptl, pte_t orig_pte)
L
Linus Torvalds 已提交
1566
{
1567
	struct page *old_page, *new_page;
L
Linus Torvalds 已提交
1568
	pte_t entry;
N
Nick Piggin 已提交
1569
	int reuse = 0, ret = 0;
1570
	int page_mkwrite = 0;
1571
	struct page *dirty_page = NULL;
L
Linus Torvalds 已提交
1572

1573 1574 1575
	old_page = vm_normal_page(vma, address, orig_pte);
	if (!old_page)
		goto gotten;
L
Linus Torvalds 已提交
1576

1577
	/*
P
Peter Zijlstra 已提交
1578 1579
	 * Take out anonymous pages first, anonymous shared vmas are
	 * not dirty accountable.
1580
	 */
P
Peter Zijlstra 已提交
1581 1582 1583 1584 1585 1586
	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)) ==
1587
					(VM_WRITE|VM_SHARED))) {
P
Peter Zijlstra 已提交
1588 1589 1590 1591 1592
		/*
		 * Only catch write-faults on shared writable pages,
		 * read-only shared pages can get COWed by
		 * get_user_pages(.write=1, .force=1).
		 */
1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615
		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);
1616
			page_cache_release(old_page);
1617 1618
			if (!pte_same(*page_table, orig_pte))
				goto unlock;
1619 1620

			page_mkwrite = 1;
L
Linus Torvalds 已提交
1621
		}
1622 1623
		dirty_page = old_page;
		get_page(dirty_page);
1624 1625 1626 1627 1628 1629 1630
		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);
1631
		if (ptep_set_access_flags(vma, address, page_table, entry,1))
1632
			update_mmu_cache(vma, address, entry);
1633 1634
		ret |= VM_FAULT_WRITE;
		goto unlock;
L
Linus Torvalds 已提交
1635 1636 1637 1638 1639
	}

	/*
	 * Ok, we need to copy. Oh, well..
	 */
N
Nick Piggin 已提交
1640
	page_cache_get(old_page);
H
Hugh Dickins 已提交
1641
gotten:
1642
	pte_unmap_unlock(page_table, ptl);
L
Linus Torvalds 已提交
1643 1644

	if (unlikely(anon_vma_prepare(vma)))
1645
		goto oom;
N
Nick Piggin 已提交
1646 1647 1648 1649 1650
	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);
N
Nick Piggin 已提交
1651
	__SetPageUptodate(new_page);
1652

1653
	if (mem_cgroup_charge(new_page, mm, GFP_KERNEL))
1654 1655
		goto oom_free_new;

L
Linus Torvalds 已提交
1656 1657 1658
	/*
	 * Re-check the pte - we dropped the lock
	 */
1659
	page_table = pte_offset_map_lock(mm, pmd, address, &ptl);
1660
	if (likely(pte_same(*page_table, orig_pte))) {
H
Hugh Dickins 已提交
1661
		if (old_page) {
N
Nick Piggin 已提交
1662
			page_remove_rmap(old_page, vma);
H
Hugh Dickins 已提交
1663 1664 1665 1666 1667
			if (!PageAnon(old_page)) {
				dec_mm_counter(mm, file_rss);
				inc_mm_counter(mm, anon_rss);
			}
		} else
1668
			inc_mm_counter(mm, anon_rss);
1669
		flush_cache_page(vma, address, pte_pfn(orig_pte));
1670 1671
		entry = mk_pte(new_page, vma->vm_page_prot);
		entry = maybe_mkwrite(pte_mkdirty(entry), vma);
1672 1673 1674 1675 1676 1677 1678 1679
		/*
		 * 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);
1680
		update_mmu_cache(vma, address, entry);
L
Linus Torvalds 已提交
1681
		lru_cache_add_active(new_page);
N
Nick Piggin 已提交
1682
		page_add_new_anon_rmap(new_page, vma, address);
L
Linus Torvalds 已提交
1683 1684 1685

		/* Free the old page.. */
		new_page = old_page;
N
Nick Piggin 已提交
1686
		ret |= VM_FAULT_WRITE;
1687 1688 1689
	} else
		mem_cgroup_uncharge_page(new_page);

H
Hugh Dickins 已提交
1690 1691 1692 1693
	if (new_page)
		page_cache_release(new_page);
	if (old_page)
		page_cache_release(old_page);
1694
unlock:
1695
	pte_unmap_unlock(page_table, ptl);
1696
	if (dirty_page) {
1697 1698 1699
		if (vma->vm_file)
			file_update_time(vma->vm_file);

1700 1701 1702 1703 1704 1705 1706 1707 1708
		/*
		 * 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);
1709
		set_page_dirty_balance(dirty_page, page_mkwrite);
1710 1711
		put_page(dirty_page);
	}
N
Nick Piggin 已提交
1712
	return ret;
1713
oom_free_new:
1714
	page_cache_release(new_page);
1715
oom:
H
Hugh Dickins 已提交
1716 1717
	if (old_page)
		page_cache_release(old_page);
L
Linus Torvalds 已提交
1718
	return VM_FAULT_OOM;
1719 1720 1721 1722

unwritable_page:
	page_cache_release(old_page);
	return VM_FAULT_SIGBUS;
L
Linus Torvalds 已提交
1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749
}

/*
 * 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
1750
 * large vma, note the restart_addr from unmap_vmas when it breaks out:
L
Linus Torvalds 已提交
1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776
 * 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;

1777 1778
	/*
	 * files that support invalidating or truncating portions of the
N
Nick Piggin 已提交
1779
	 * file from under mmaped areas must have their ->fault function
N
Nick Piggin 已提交
1780 1781
	 * return a locked page (and set VM_FAULT_LOCKED in the return).
	 * This provides synchronisation against concurrent unmapping here.
1782 1783
	 */

L
Linus Torvalds 已提交
1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794
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;
		}
	}

1795 1796
	restart_addr = zap_page_range(vma, start_addr,
					end_addr - start_addr, details);
N
Nick Piggin 已提交
1797
	need_break = need_resched() || spin_needbreak(details->i_mmap_lock);
L
Linus Torvalds 已提交
1798

1799
	if (restart_addr >= end_addr) {
L
Linus Torvalds 已提交
1800 1801 1802 1803 1804 1805
		/* 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 */
1806
		vma->vm_truncate_count = restart_addr;
L
Linus Torvalds 已提交
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 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872
		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;
	}
}

/**
1873
 * 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
1874
 * @mapping: the address space containing mmaps to be unmapped.
L
Linus Torvalds 已提交
1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910
 * @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);

1911
	/* Protect against endless unmapping loops */
L
Linus Torvalds 已提交
1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927
	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);

1928 1929 1930 1931
/**
 * 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
1932 1933 1934 1935 1936 1937 1938
 *
 * 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)
{
C
Christoph Hellwig 已提交
1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949
	if (inode->i_size < offset) {
		unsigned long limit;

		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);
	} else {
		struct address_space *mapping = inode->i_mapping;
L
Linus Torvalds 已提交
1950

C
Christoph Hellwig 已提交
1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972
		/*
		 * truncation of in-use swapfiles is disallowed - it would
		 * cause subsequent swapout to scribble on the now-freed
		 * blocks.
		 */
		if (IS_SWAPFILE(inode))
			return -ETXTBSY;
		i_size_write(inode, offset);

		/*
		 * 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.
		 */
		unmap_mapping_range(mapping, offset + PAGE_SIZE - 1, 0, 1);
		truncate_inode_pages(mapping, offset);
		unmap_mapping_range(mapping, offset + PAGE_SIZE - 1, 0, 1);
	}
1973

L
Linus Torvalds 已提交
1974 1975 1976
	if (inode->i_op && inode->i_op->truncate)
		inode->i_op->truncate(inode);
	return 0;
C
Christoph Hellwig 已提交
1977

L
Linus Torvalds 已提交
1978 1979 1980 1981 1982 1983 1984
out_sig:
	send_sig(SIGXFSZ, current, 0);
out_big:
	return -EFBIG;
}
EXPORT_SYMBOL(vmtruncate);

1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996
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;

1997
	mutex_lock(&inode->i_mutex);
1998 1999 2000
	down_write(&inode->i_alloc_sem);
	unmap_mapping_range(mapping, offset, (end - offset), 1);
	truncate_inode_pages_range(mapping, offset, end);
2001
	unmap_mapping_range(mapping, offset, (end - offset), 1);
2002 2003
	inode->i_op->truncate_range(inode, offset, end);
	up_write(&inode->i_alloc_sem);
2004
	mutex_unlock(&inode->i_mutex);
2005 2006 2007 2008

	return 0;
}

L
Linus Torvalds 已提交
2009
/*
2010 2011 2012
 * 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
2013
 */
2014 2015 2016
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
2017
{
2018
	spinlock_t *ptl;
L
Linus Torvalds 已提交
2019
	struct page *page;
2020
	swp_entry_t entry;
L
Linus Torvalds 已提交
2021
	pte_t pte;
N
Nick Piggin 已提交
2022
	int ret = 0;
L
Linus Torvalds 已提交
2023

H
Hugh Dickins 已提交
2024
	if (!pte_unmap_same(mm, pmd, page_table, orig_pte))
2025
		goto out;
2026 2027

	entry = pte_to_swp_entry(orig_pte);
2028 2029 2030 2031
	if (is_migration_entry(entry)) {
		migration_entry_wait(mm, pmd, address);
		goto out;
	}
2032
	delayacct_set_flag(DELAYACCT_PF_SWAPIN);
L
Linus Torvalds 已提交
2033 2034
	page = lookup_swap_cache(entry);
	if (!page) {
2035
		grab_swap_token(); /* Contend for token _before_ read-in */
2036 2037
		page = swapin_readahead(entry,
					GFP_HIGHUSER_MOVABLE, vma, address);
L
Linus Torvalds 已提交
2038 2039
		if (!page) {
			/*
2040 2041
			 * Back out if somebody else faulted in this pte
			 * while we released the pte lock.
L
Linus Torvalds 已提交
2042
			 */
2043
			page_table = pte_offset_map_lock(mm, pmd, address, &ptl);
L
Linus Torvalds 已提交
2044 2045
			if (likely(pte_same(*page_table, orig_pte)))
				ret = VM_FAULT_OOM;
2046
			delayacct_clear_flag(DELAYACCT_PF_SWAPIN);
2047
			goto unlock;
L
Linus Torvalds 已提交
2048 2049 2050 2051
		}

		/* Had to read the page from swap area: Major fault */
		ret = VM_FAULT_MAJOR;
2052
		count_vm_event(PGMAJFAULT);
L
Linus Torvalds 已提交
2053 2054
	}

2055
	if (mem_cgroup_charge(page, mm, GFP_KERNEL)) {
2056 2057 2058 2059 2060
		delayacct_clear_flag(DELAYACCT_PF_SWAPIN);
		ret = VM_FAULT_OOM;
		goto out;
	}

L
Linus Torvalds 已提交
2061 2062
	mark_page_accessed(page);
	lock_page(page);
2063
	delayacct_clear_flag(DELAYACCT_PF_SWAPIN);
L
Linus Torvalds 已提交
2064 2065

	/*
2066
	 * Back out if somebody else already faulted in this pte.
L
Linus Torvalds 已提交
2067
	 */
2068
	page_table = pte_offset_map_lock(mm, pmd, address, &ptl);
H
Hugh Dickins 已提交
2069
	if (unlikely(!pte_same(*page_table, orig_pte)))
2070 2071 2072 2073 2074
		goto out_nomap;

	if (unlikely(!PageUptodate(page))) {
		ret = VM_FAULT_SIGBUS;
		goto out_nomap;
L
Linus Torvalds 已提交
2075 2076 2077 2078
	}

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

2079
	inc_mm_counter(mm, anon_rss);
L
Linus Torvalds 已提交
2080 2081 2082 2083 2084 2085 2086 2087 2088 2089
	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);

2090 2091 2092 2093 2094
	swap_free(entry);
	if (vm_swap_full())
		remove_exclusive_swap_page(page);
	unlock_page(page);

L
Linus Torvalds 已提交
2095
	if (write_access) {
N
Nick Piggin 已提交
2096
		/* XXX: We could OR the do_wp_page code with this one? */
L
Linus Torvalds 已提交
2097
		if (do_wp_page(mm, vma, address,
2098 2099
				page_table, pmd, ptl, pte) & VM_FAULT_OOM) {
			mem_cgroup_uncharge_page(page);
L
Linus Torvalds 已提交
2100
			ret = VM_FAULT_OOM;
2101
		}
L
Linus Torvalds 已提交
2102 2103 2104 2105 2106
		goto out;
	}

	/* No need to invalidate - it was non-present before */
	update_mmu_cache(vma, address, pte);
2107
unlock:
2108
	pte_unmap_unlock(page_table, ptl);
L
Linus Torvalds 已提交
2109 2110
out:
	return ret;
2111
out_nomap:
2112
	mem_cgroup_uncharge_page(page);
2113
	pte_unmap_unlock(page_table, ptl);
2114 2115
	unlock_page(page);
	page_cache_release(page);
2116
	return ret;
L
Linus Torvalds 已提交
2117 2118 2119
}

/*
2120 2121 2122
 * 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
2123
 */
2124 2125 2126
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
2127
{
2128 2129
	struct page *page;
	spinlock_t *ptl;
L
Linus Torvalds 已提交
2130 2131
	pte_t entry;

N
Nick Piggin 已提交
2132 2133
	/* Allocate our own private page. */
	pte_unmap(page_table);
2134

N
Nick Piggin 已提交
2135 2136 2137 2138 2139
	if (unlikely(anon_vma_prepare(vma)))
		goto oom;
	page = alloc_zeroed_user_highpage_movable(vma, address);
	if (!page)
		goto oom;
N
Nick Piggin 已提交
2140
	__SetPageUptodate(page);
2141

2142
	if (mem_cgroup_charge(page, mm, GFP_KERNEL))
2143 2144
		goto oom_free_page;

N
Nick Piggin 已提交
2145 2146
	entry = mk_pte(page, vma->vm_page_prot);
	entry = maybe_mkwrite(pte_mkdirty(entry), vma);
L
Linus Torvalds 已提交
2147

N
Nick Piggin 已提交
2148 2149 2150 2151 2152 2153
	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);
2154
	set_pte_at(mm, address, page_table, entry);
L
Linus Torvalds 已提交
2155 2156

	/* No need to invalidate - it was non-present before */
2157 2158
	update_mmu_cache(vma, address, entry);
unlock:
2159
	pte_unmap_unlock(page_table, ptl);
N
Nick Piggin 已提交
2160
	return 0;
2161
release:
2162
	mem_cgroup_uncharge_page(page);
2163 2164
	page_cache_release(page);
	goto unlock;
2165
oom_free_page:
2166
	page_cache_release(page);
2167
oom:
L
Linus Torvalds 已提交
2168 2169 2170 2171
	return VM_FAULT_OOM;
}

/*
2172
 * __do_fault() tries to create a new page mapping. It aggressively
L
Linus Torvalds 已提交
2173
 * tries to share with existing pages, but makes a separate copy if
2174 2175
 * the FAULT_FLAG_WRITE is set in the flags parameter in order to avoid
 * the next page fault.
L
Linus Torvalds 已提交
2176 2177 2178 2179
 *
 * As this is called only for pages that do not currently exist, we
 * do not need to flush old virtual caches or the TLB.
 *
2180
 * We enter with non-exclusive mmap_sem (to exclude vma changes,
2181
 * but allow concurrent faults), and pte neither mapped nor locked.
2182
 * We return with mmap_sem still held, but pte unmapped and unlocked.
L
Linus Torvalds 已提交
2183
 */
2184
static int __do_fault(struct mm_struct *mm, struct vm_area_struct *vma,
2185
		unsigned long address, pmd_t *pmd,
2186
		pgoff_t pgoff, unsigned int flags, pte_t orig_pte)
L
Linus Torvalds 已提交
2187
{
2188
	pte_t *page_table;
2189
	spinlock_t *ptl;
N
Nick Piggin 已提交
2190
	struct page *page;
L
Linus Torvalds 已提交
2191 2192
	pte_t entry;
	int anon = 0;
2193
	struct page *dirty_page = NULL;
N
Nick Piggin 已提交
2194 2195
	struct vm_fault vmf;
	int ret;
2196
	int page_mkwrite = 0;
2197

N
Nick Piggin 已提交
2198 2199 2200 2201
	vmf.virtual_address = (void __user *)(address & PAGE_MASK);
	vmf.pgoff = pgoff;
	vmf.flags = flags;
	vmf.page = NULL;
L
Linus Torvalds 已提交
2202

2203 2204
	BUG_ON(vma->vm_flags & VM_PFNMAP);

2205
	if (likely(vma->vm_ops->fault)) {
N
Nick Piggin 已提交
2206
		ret = vma->vm_ops->fault(vma, &vmf);
N
Nick Piggin 已提交
2207 2208
		if (unlikely(ret & (VM_FAULT_ERROR | VM_FAULT_NOPAGE)))
			return ret;
2209 2210
	} else {
		/* Legacy ->nopage path */
N
Nick Piggin 已提交
2211
		ret = 0;
N
Nick Piggin 已提交
2212
		vmf.page = vma->vm_ops->nopage(vma, address & PAGE_MASK, &ret);
2213
		/* no page was available -- either SIGBUS or OOM */
N
Nick Piggin 已提交
2214
		if (unlikely(vmf.page == NOPAGE_SIGBUS))
2215
			return VM_FAULT_SIGBUS;
N
Nick Piggin 已提交
2216
		else if (unlikely(vmf.page == NOPAGE_OOM))
2217 2218
			return VM_FAULT_OOM;
	}
L
Linus Torvalds 已提交
2219

2220
	/*
N
Nick Piggin 已提交
2221
	 * For consistency in subsequent calls, make the faulted page always
2222 2223
	 * locked.
	 */
N
Nick Piggin 已提交
2224
	if (unlikely(!(ret & VM_FAULT_LOCKED)))
N
Nick Piggin 已提交
2225
		lock_page(vmf.page);
2226
	else
N
Nick Piggin 已提交
2227
		VM_BUG_ON(!PageLocked(vmf.page));
2228

L
Linus Torvalds 已提交
2229 2230 2231
	/*
	 * Should we do an early C-O-W break?
	 */
N
Nick Piggin 已提交
2232
	page = vmf.page;
2233
	if (flags & FAULT_FLAG_WRITE) {
2234
		if (!(vma->vm_flags & VM_SHARED)) {
2235
			anon = 1;
2236
			if (unlikely(anon_vma_prepare(vma))) {
N
Nick Piggin 已提交
2237
				ret = VM_FAULT_OOM;
2238
				goto out;
2239
			}
N
Nick Piggin 已提交
2240 2241
			page = alloc_page_vma(GFP_HIGHUSER_MOVABLE,
						vma, address);
2242
			if (!page) {
N
Nick Piggin 已提交
2243
				ret = VM_FAULT_OOM;
2244
				goto out;
2245
			}
N
Nick Piggin 已提交
2246
			copy_user_highpage(page, vmf.page, address, vma);
N
Nick Piggin 已提交
2247
			__SetPageUptodate(page);
2248
		} else {
2249 2250
			/*
			 * If the page will be shareable, see if the backing
2251
			 * address space wants to know that the page is about
2252 2253
			 * to become writable
			 */
2254 2255 2256
			if (vma->vm_ops->page_mkwrite) {
				unlock_page(page);
				if (vma->vm_ops->page_mkwrite(vma, page) < 0) {
N
Nick Piggin 已提交
2257 2258
					ret = VM_FAULT_SIGBUS;
					anon = 1; /* no anon but release vmf.page */
2259 2260 2261
					goto out_unlocked;
				}
				lock_page(page);
N
Nick Piggin 已提交
2262 2263 2264 2265 2266 2267 2268 2269
				/*
				 * 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
2270
					ret = 0;
N
Nick Piggin 已提交
2271 2272 2273
					anon = 1; /* no anon but release vmf.page */
					goto out;
				}
2274
				page_mkwrite = 1;
2275 2276
			}
		}
2277

L
Linus Torvalds 已提交
2278 2279
	}

2280
	if (mem_cgroup_charge(page, mm, GFP_KERNEL)) {
2281 2282 2283 2284
		ret = VM_FAULT_OOM;
		goto out;
	}

2285
	page_table = pte_offset_map_lock(mm, pmd, address, &ptl);
L
Linus Torvalds 已提交
2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297

	/*
	 * 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... */
2298
	if (likely(pte_same(*page_table, orig_pte))) {
2299 2300
		flush_icache_page(vma, page);
		entry = mk_pte(page, vma->vm_page_prot);
2301
		if (flags & FAULT_FLAG_WRITE)
L
Linus Torvalds 已提交
2302 2303 2304
			entry = maybe_mkwrite(pte_mkdirty(entry), vma);
		set_pte_at(mm, address, page_table, entry);
		if (anon) {
2305 2306 2307
                        inc_mm_counter(mm, anon_rss);
                        lru_cache_add_active(page);
                        page_add_new_anon_rmap(page, vma, address);
2308
		} else {
2309
			inc_mm_counter(mm, file_rss);
2310
			page_add_file_rmap(page);
2311
			if (flags & FAULT_FLAG_WRITE) {
2312
				dirty_page = page;
2313 2314
				get_page(dirty_page);
			}
2315
		}
2316 2317 2318

		/* no need to invalidate: a not-present page won't be cached */
		update_mmu_cache(vma, address, entry);
L
Linus Torvalds 已提交
2319
	} else {
2320
		mem_cgroup_uncharge_page(page);
2321 2322 2323
		if (anon)
			page_cache_release(page);
		else
2324
			anon = 1; /* no anon but release faulted_page */
L
Linus Torvalds 已提交
2325 2326
	}

2327
	pte_unmap_unlock(page_table, ptl);
2328 2329

out:
N
Nick Piggin 已提交
2330
	unlock_page(vmf.page);
2331
out_unlocked:
2332
	if (anon)
N
Nick Piggin 已提交
2333
		page_cache_release(vmf.page);
2334
	else if (dirty_page) {
2335 2336 2337
		if (vma->vm_file)
			file_update_time(vma->vm_file);

2338
		set_page_dirty_balance(dirty_page, page_mkwrite);
2339 2340
		put_page(dirty_page);
	}
2341

N
Nick Piggin 已提交
2342
	return ret;
2343
}
2344

2345 2346 2347 2348 2349
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)
2350
			- vma->vm_start) >> PAGE_SHIFT) + vma->vm_pgoff;
2351 2352
	unsigned int flags = (write_access ? FAULT_FLAG_WRITE : 0);

2353 2354
	pte_unmap(page_table);
	return __do_fault(mm, vma, address, pmd, pgoff, flags, orig_pte);
2355 2356
}

L
Linus Torvalds 已提交
2357

J
Jes Sorensen 已提交
2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386
/*
 * 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);
2387
	if (unlikely(pfn == NOPFN_OOM))
J
Jes Sorensen 已提交
2388
		return VM_FAULT_OOM;
2389
	else if (unlikely(pfn == NOPFN_SIGBUS))
J
Jes Sorensen 已提交
2390
		return VM_FAULT_SIGBUS;
2391
	else if (unlikely(pfn == NOPFN_REFAULT))
N
Nick Piggin 已提交
2392
		return 0;
J
Jes Sorensen 已提交
2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403

	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
2404
	return 0;
J
Jes Sorensen 已提交
2405 2406
}

L
Linus Torvalds 已提交
2407 2408 2409 2410
/*
 * Fault of a previously existing named mapping. Repopulate the pte
 * from the encoded file_pte if possible. This enables swappable
 * nonlinear vmas.
2411 2412 2413 2414
 *
 * 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
2415
 */
N
Nick Piggin 已提交
2416
static int do_nonlinear_fault(struct mm_struct *mm, struct vm_area_struct *vma,
2417 2418
		unsigned long address, pte_t *page_table, pmd_t *pmd,
		int write_access, pte_t orig_pte)
L
Linus Torvalds 已提交
2419
{
N
Nick Piggin 已提交
2420 2421
	unsigned int flags = FAULT_FLAG_NONLINEAR |
				(write_access ? FAULT_FLAG_WRITE : 0);
2422
	pgoff_t pgoff;
L
Linus Torvalds 已提交
2423

H
Hugh Dickins 已提交
2424
	if (!pte_unmap_same(mm, pmd, page_table, orig_pte))
N
Nick Piggin 已提交
2425
		return 0;
L
Linus Torvalds 已提交
2426

N
Nick Piggin 已提交
2427 2428
	if (unlikely(!(vma->vm_flags & VM_NONLINEAR) ||
			!(vma->vm_flags & VM_CAN_NONLINEAR))) {
2429 2430 2431
		/*
		 * Page table corrupted: show pte and kill process.
		 */
N
Nick Piggin 已提交
2432
		print_bad_pte(vma, orig_pte, address);
2433 2434 2435 2436
		return VM_FAULT_OOM;
	}

	pgoff = pte_to_pgoff(orig_pte);
2437
	return __do_fault(mm, vma, address, pmd, pgoff, flags, orig_pte);
L
Linus Torvalds 已提交
2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448
}

/*
 * 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
2449 2450 2451
 * 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
2452 2453
 */
static inline int handle_pte_fault(struct mm_struct *mm,
2454 2455
		struct vm_area_struct *vma, unsigned long address,
		pte_t *pte, pmd_t *pmd, int write_access)
L
Linus Torvalds 已提交
2456 2457
{
	pte_t entry;
2458
	spinlock_t *ptl;
L
Linus Torvalds 已提交
2459

2460
	entry = *pte;
L
Linus Torvalds 已提交
2461
	if (!pte_present(entry)) {
2462
		if (pte_none(entry)) {
J
Jes Sorensen 已提交
2463
			if (vma->vm_ops) {
2464 2465 2466
				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
2467 2468 2469 2470 2471 2472
				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);
2473
		}
L
Linus Torvalds 已提交
2474
		if (pte_file(entry))
N
Nick Piggin 已提交
2475
			return do_nonlinear_fault(mm, vma, address,
2476 2477 2478
					pte, pmd, write_access, entry);
		return do_swap_page(mm, vma, address,
					pte, pmd, write_access, entry);
L
Linus Torvalds 已提交
2479 2480
	}

H
Hugh Dickins 已提交
2481
	ptl = pte_lockptr(mm, pmd);
2482 2483 2484
	spin_lock(ptl);
	if (unlikely(!pte_same(*pte, entry)))
		goto unlock;
L
Linus Torvalds 已提交
2485 2486
	if (write_access) {
		if (!pte_write(entry))
2487 2488
			return do_wp_page(mm, vma, address,
					pte, pmd, ptl, entry);
L
Linus Torvalds 已提交
2489 2490 2491
		entry = pte_mkdirty(entry);
	}
	entry = pte_mkyoung(entry);
2492
	if (ptep_set_access_flags(vma, address, pte, entry, write_access)) {
2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503
		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);
	}
2504 2505
unlock:
	pte_unmap_unlock(pte, ptl);
N
Nick Piggin 已提交
2506
	return 0;
L
Linus Torvalds 已提交
2507 2508 2509 2510 2511
}

/*
 * By the time we get here, we already hold the mm semaphore
 */
N
Nick Piggin 已提交
2512
int handle_mm_fault(struct mm_struct *mm, struct vm_area_struct *vma,
L
Linus Torvalds 已提交
2513 2514 2515 2516 2517 2518 2519 2520 2521
		unsigned long address, int write_access)
{
	pgd_t *pgd;
	pud_t *pud;
	pmd_t *pmd;
	pte_t *pte;

	__set_current_state(TASK_RUNNING);

2522
	count_vm_event(PGFAULT);
L
Linus Torvalds 已提交
2523

2524 2525
	if (unlikely(is_vm_hugetlb_page(vma)))
		return hugetlb_fault(mm, vma, address, write_access);
L
Linus Torvalds 已提交
2526 2527 2528 2529

	pgd = pgd_offset(mm, address);
	pud = pud_alloc(mm, pgd, address);
	if (!pud)
H
Hugh Dickins 已提交
2530
		return VM_FAULT_OOM;
L
Linus Torvalds 已提交
2531 2532
	pmd = pmd_alloc(mm, pud, address);
	if (!pmd)
H
Hugh Dickins 已提交
2533
		return VM_FAULT_OOM;
L
Linus Torvalds 已提交
2534 2535
	pte = pte_alloc_map(mm, pmd, address);
	if (!pte)
H
Hugh Dickins 已提交
2536
		return VM_FAULT_OOM;
L
Linus Torvalds 已提交
2537

H
Hugh Dickins 已提交
2538
	return handle_pte_fault(mm, vma, address, pte, pmd, write_access);
L
Linus Torvalds 已提交
2539 2540 2541 2542 2543
}

#ifndef __PAGETABLE_PUD_FOLDED
/*
 * Allocate page upper directory.
H
Hugh Dickins 已提交
2544
 * We've already handled the fast-path in-line.
L
Linus Torvalds 已提交
2545
 */
2546
int __pud_alloc(struct mm_struct *mm, pgd_t *pgd, unsigned long address)
L
Linus Torvalds 已提交
2547
{
H
Hugh Dickins 已提交
2548 2549
	pud_t *new = pud_alloc_one(mm, address);
	if (!new)
2550
		return -ENOMEM;
L
Linus Torvalds 已提交
2551

H
Hugh Dickins 已提交
2552
	spin_lock(&mm->page_table_lock);
2553
	if (pgd_present(*pgd))		/* Another has populated it */
2554
		pud_free(mm, new);
2555 2556
	else
		pgd_populate(mm, pgd, new);
H
Hugh Dickins 已提交
2557
	spin_unlock(&mm->page_table_lock);
2558
	return 0;
L
Linus Torvalds 已提交
2559 2560 2561 2562 2563 2564
}
#endif /* __PAGETABLE_PUD_FOLDED */

#ifndef __PAGETABLE_PMD_FOLDED
/*
 * Allocate page middle directory.
H
Hugh Dickins 已提交
2565
 * We've already handled the fast-path in-line.
L
Linus Torvalds 已提交
2566
 */
2567
int __pmd_alloc(struct mm_struct *mm, pud_t *pud, unsigned long address)
L
Linus Torvalds 已提交
2568
{
H
Hugh Dickins 已提交
2569 2570
	pmd_t *new = pmd_alloc_one(mm, address);
	if (!new)
2571
		return -ENOMEM;
L
Linus Torvalds 已提交
2572

H
Hugh Dickins 已提交
2573
	spin_lock(&mm->page_table_lock);
L
Linus Torvalds 已提交
2574
#ifndef __ARCH_HAS_4LEVEL_HACK
2575
	if (pud_present(*pud))		/* Another has populated it */
2576
		pmd_free(mm, new);
2577 2578
	else
		pud_populate(mm, pud, new);
L
Linus Torvalds 已提交
2579
#else
2580
	if (pgd_present(*pud))		/* Another has populated it */
2581
		pmd_free(mm, new);
2582 2583
	else
		pgd_populate(mm, pud, new);
L
Linus Torvalds 已提交
2584
#endif /* __ARCH_HAS_4LEVEL_HACK */
H
Hugh Dickins 已提交
2585
	spin_unlock(&mm->page_table_lock);
2586
	return 0;
2587
}
L
Linus Torvalds 已提交
2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598
#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;
2599 2600
	BUG_ON(addr >= end);
	BUG_ON(end > vma->vm_end);
2601
	len = DIV_ROUND_UP(end, PAGE_SIZE) - addr/PAGE_SIZE;
L
Linus Torvalds 已提交
2602 2603 2604 2605 2606 2607 2608 2609 2610 2611
	ret = get_user_pages(current, current->mm, addr,
			len, write, 0, NULL, NULL);
	if (ret < 0)
		return ret;
	return ret == len ? 0 : -1;
}

#if !defined(__HAVE_ARCH_GATE_AREA)

#if defined(AT_SYSINFO_EHDR)
2612
static struct vm_area_struct gate_vma;
L
Linus Torvalds 已提交
2613 2614 2615 2616 2617 2618

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
2619 2620
	gate_vma.vm_flags = VM_READ | VM_MAYREAD | VM_EXEC | VM_MAYEXEC;
	gate_vma.vm_page_prot = __P101;
2621 2622 2623 2624 2625 2626 2627
	/*
	 * 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
2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651
	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 */
2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669

/*
 * 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
2670
	/* ignore errors, just check how much was successfully transferred */
2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704
	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;
}
2705 2706 2707 2708 2709 2710 2711 2712 2713

/*
 * Print the name of a VMA.
 */
void print_vma_addr(char *prefix, unsigned long ip)
{
	struct mm_struct *mm = current->mm;
	struct vm_area_struct *vma;

2714 2715 2716 2717 2718 2719 2720
	/*
	 * Do not print if we are in atomic
	 * contexts (in exception stacks, etc.):
	 */
	if (preempt_count())
		return;

2721 2722 2723 2724 2725 2726 2727 2728
	down_read(&mm->mmap_sem);
	vma = find_vma(mm, ip);
	if (vma && vma->vm_file) {
		struct file *f = vma->vm_file;
		char *buf = (char *)__get_free_page(GFP_KERNEL);
		if (buf) {
			char *p, *s;

2729
			p = d_path(&f->f_path, buf, PAGE_SIZE);
2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742
			if (IS_ERR(p))
				p = "?";
			s = strrchr(p, '/');
			if (s)
				p = s+1;
			printk("%s%s[%lx+%lx]", prefix, p,
					vma->vm_start,
					vma->vm_end - vma->vm_start);
			free_page((unsigned long)buf);
		}
	}
	up_read(&current->mm->mmap_sem);
}