shmem.c 68.9 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8
/*
 * Resizable virtual memory filesystem for Linux.
 *
 * Copyright (C) 2000 Linus Torvalds.
 *		 2000 Transmeta Corp.
 *		 2000-2001 Christoph Rohland
 *		 2000-2001 SAP AG
 *		 2002 Red Hat Inc.
9 10
 * Copyright (C) 2002-2005 Hugh Dickins.
 * Copyright (C) 2002-2005 VERITAS Software Corporation.
L
Linus Torvalds 已提交
11 12 13 14 15 16
 * Copyright (C) 2004 Andi Kleen, SuSE Labs
 *
 * Extended attribute support for tmpfs:
 * Copyright (c) 2004, Luke Kenneth Casson Leighton <lkcl@lkcl.net>
 * Copyright (c) 2004 Red Hat, Inc., James Morris <jmorris@redhat.com>
 *
17 18 19
 * tiny-shmem:
 * Copyright (c) 2004, 2008 Matt Mackall <mpm@selenic.com>
 *
L
Linus Torvalds 已提交
20 21 22
 * This file is released under the GPL.
 */

23 24 25 26
#include <linux/fs.h>
#include <linux/init.h>
#include <linux/vfs.h>
#include <linux/mount.h>
H
Hugh Dickins 已提交
27
#include <linux/pagemap.h>
28 29 30 31
#include <linux/file.h>
#include <linux/mm.h>
#include <linux/module.h>
#include <linux/swap.h>
M
Mimi Zohar 已提交
32
#include <linux/ima.h>
33 34 35 36

static struct vfsmount *shm_mnt;

#ifdef CONFIG_SHMEM
L
Linus Torvalds 已提交
37 38 39 40 41 42
/*
 * This virtual memory filesystem is heavily based on the ramfs. It
 * extends ramfs by the ability to use swap and honor resource limits
 * which makes it a completely usable filesystem.
 */

43
#include <linux/xattr.h>
44
#include <linux/exportfs.h>
45
#include <linux/generic_acl.h>
L
Linus Torvalds 已提交
46 47 48 49 50 51 52 53 54 55 56 57
#include <linux/mman.h>
#include <linux/string.h>
#include <linux/slab.h>
#include <linux/backing-dev.h>
#include <linux/shmem_fs.h>
#include <linux/writeback.h>
#include <linux/vfs.h>
#include <linux/blkdev.h>
#include <linux/security.h>
#include <linux/swapops.h>
#include <linux/mempolicy.h>
#include <linux/namei.h>
58
#include <linux/ctype.h>
59
#include <linux/migrate.h>
60
#include <linux/highmem.h>
61
#include <linux/seq_file.h>
M
Mimi Zohar 已提交
62
#include <linux/magic.h>
63

L
Linus Torvalds 已提交
64 65 66 67
#include <asm/uaccess.h>
#include <asm/div64.h>
#include <asm/pgtable.h>

H
Hugh Dickins 已提交
68 69 70 71 72 73 74 75 76 77 78 79
/*
 * The maximum size of a shmem/tmpfs file is limited by the maximum size of
 * its triple-indirect swap vector - see illustration at shmem_swp_entry().
 *
 * With 4kB page size, maximum file size is just over 2TB on a 32-bit kernel,
 * but one eighth of that on a 64-bit kernel.  With 8kB page size, maximum
 * file size is just over 4TB on a 64-bit kernel, but 16TB on a 32-bit kernel,
 * MAX_LFS_FILESIZE being then more restrictive than swap vector layout.
 *
 * We use / and * instead of shifts in the definitions below, so that the swap
 * vector can be tested with small even values (e.g. 20) for ENTRIES_PER_PAGE.
 */
L
Linus Torvalds 已提交
80
#define ENTRIES_PER_PAGE (PAGE_CACHE_SIZE/sizeof(unsigned long))
Y
Yuri Tikhonov 已提交
81
#define ENTRIES_PER_PAGEPAGE ((unsigned long long)ENTRIES_PER_PAGE*ENTRIES_PER_PAGE)
L
Linus Torvalds 已提交
82

H
Hugh Dickins 已提交
83 84
#define SHMSWP_MAX_INDEX (SHMEM_NR_DIRECT + (ENTRIES_PER_PAGEPAGE/2) * (ENTRIES_PER_PAGE+1))
#define SHMSWP_MAX_BYTES (SHMSWP_MAX_INDEX << PAGE_CACHE_SHIFT)
L
Linus Torvalds 已提交
85

H
Hugh Dickins 已提交
86 87 88 89
#define SHMEM_MAX_BYTES  min_t(unsigned long long, SHMSWP_MAX_BYTES, MAX_LFS_FILESIZE)
#define SHMEM_MAX_INDEX  ((unsigned long)((SHMEM_MAX_BYTES+1) >> PAGE_CACHE_SHIFT))

#define BLOCKS_PER_PAGE  (PAGE_CACHE_SIZE/512)
L
Linus Torvalds 已提交
90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105
#define VM_ACCT(size)    (PAGE_CACHE_ALIGN(size) >> PAGE_SHIFT)

/* info->flags needs VM_flags to handle pagein/truncate races efficiently */
#define SHMEM_PAGEIN	 VM_READ
#define SHMEM_TRUNCATE	 VM_WRITE

/* Definition to limit shmem_truncate's steps between cond_rescheds */
#define LATENCY_LIMIT	 64

/* Pretend that each entry is of this size in directory's i_size */
#define BOGO_DIRENT_SIZE 20

/* Flag allocation requirements to shmem_getpage and shmem_swp_alloc */
enum sgp_type {
	SGP_READ,	/* don't exceed i_size, don't allocate page */
	SGP_CACHE,	/* don't exceed i_size, may allocate page */
106
	SGP_DIRTY,	/* like SGP_CACHE, but set new page dirty */
L
Linus Torvalds 已提交
107 108 109
	SGP_WRITE,	/* may exceed i_size, may allocate page */
};

A
Andrew Morton 已提交
110
#ifdef CONFIG_TMPFS
111 112 113 114 115 116 117 118 119
static unsigned long shmem_default_max_blocks(void)
{
	return totalram_pages / 2;
}

static unsigned long shmem_default_max_inodes(void)
{
	return min(totalram_pages - totalhigh_pages, totalram_pages / 2);
}
A
Andrew Morton 已提交
120
#endif
121

L
Linus Torvalds 已提交
122 123 124
static int shmem_getpage(struct inode *inode, unsigned long idx,
			 struct page **pagep, enum sgp_type sgp, int *type);

A
Al Viro 已提交
125
static inline struct page *shmem_dir_alloc(gfp_t gfp_mask)
L
Linus Torvalds 已提交
126 127 128 129 130
{
	/*
	 * The above definition of ENTRIES_PER_PAGE, and the use of
	 * BLOCKS_PER_PAGE on indirect pages, assume PAGE_CACHE_SIZE:
	 * might be reconsidered if it ever diverges from PAGE_SIZE.
131
	 *
132
	 * Mobility flags are masked out as swap vectors cannot move
L
Linus Torvalds 已提交
133
	 */
134
	return alloc_pages((gfp_mask & ~GFP_MOVABLE_MASK) | __GFP_ZERO,
135
				PAGE_CACHE_SHIFT-PAGE_SHIFT);
L
Linus Torvalds 已提交
136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187
}

static inline void shmem_dir_free(struct page *page)
{
	__free_pages(page, PAGE_CACHE_SHIFT-PAGE_SHIFT);
}

static struct page **shmem_dir_map(struct page *page)
{
	return (struct page **)kmap_atomic(page, KM_USER0);
}

static inline void shmem_dir_unmap(struct page **dir)
{
	kunmap_atomic(dir, KM_USER0);
}

static swp_entry_t *shmem_swp_map(struct page *page)
{
	return (swp_entry_t *)kmap_atomic(page, KM_USER1);
}

static inline void shmem_swp_balance_unmap(void)
{
	/*
	 * When passing a pointer to an i_direct entry, to code which
	 * also handles indirect entries and so will shmem_swp_unmap,
	 * we must arrange for the preempt count to remain in balance.
	 * What kmap_atomic of a lowmem page does depends on config
	 * and architecture, so pretend to kmap_atomic some lowmem page.
	 */
	(void) kmap_atomic(ZERO_PAGE(0), KM_USER1);
}

static inline void shmem_swp_unmap(swp_entry_t *entry)
{
	kunmap_atomic(entry, KM_USER1);
}

static inline struct shmem_sb_info *SHMEM_SB(struct super_block *sb)
{
	return sb->s_fs_info;
}

/*
 * shmem_file_setup pre-accounts the whole fixed size of a VM object,
 * for shared memory and for shared anonymous (/dev/zero) mappings
 * (unless MAP_NORESERVE and sysctl_overcommit_memory <= 1),
 * consistent with the pre-accounting of private mappings ...
 */
static inline int shmem_acct_size(unsigned long flags, loff_t size)
{
188 189
	return (flags & VM_NORESERVE) ?
		0 : security_vm_enough_memory_kern(VM_ACCT(size));
L
Linus Torvalds 已提交
190 191 192 193
}

static inline void shmem_unacct_size(unsigned long flags, loff_t size)
{
194
	if (!(flags & VM_NORESERVE))
L
Linus Torvalds 已提交
195 196 197 198 199 200 201 202 203 204 205
		vm_unacct_memory(VM_ACCT(size));
}

/*
 * ... whereas tmpfs objects are accounted incrementally as
 * pages are allocated, in order to allow huge sparse files.
 * shmem_getpage reports shmem_acct_block failure as -ENOSPC not -ENOMEM,
 * so that a failure on a sparse tmpfs mapping will give SIGBUS not OOM.
 */
static inline int shmem_acct_block(unsigned long flags)
{
206 207
	return (flags & VM_NORESERVE) ?
		security_vm_enough_memory_kern(VM_ACCT(PAGE_CACHE_SIZE)) : 0;
L
Linus Torvalds 已提交
208 209 210 211
}

static inline void shmem_unacct_blocks(unsigned long flags, long pages)
{
212
	if (flags & VM_NORESERVE)
L
Linus Torvalds 已提交
213 214 215
		vm_unacct_memory(pages * VM_ACCT(PAGE_CACHE_SIZE));
}

216
static const struct super_operations shmem_ops;
217
static const struct address_space_operations shmem_aops;
218
static const struct file_operations shmem_file_operations;
219 220 221
static const struct inode_operations shmem_inode_operations;
static const struct inode_operations shmem_dir_inode_operations;
static const struct inode_operations shmem_special_inode_operations;
L
Linus Torvalds 已提交
222 223
static struct vm_operations_struct shmem_vm_ops;

224
static struct backing_dev_info shmem_backing_dev_info  __read_mostly = {
L
Linus Torvalds 已提交
225
	.ra_pages	= 0,	/* No readahead */
226
	.capabilities	= BDI_CAP_NO_ACCT_AND_WRITEBACK | BDI_CAP_SWAP_BACKED,
L
Linus Torvalds 已提交
227 228 229 230
	.unplug_io_fn	= default_unplug_io_fn,
};

static LIST_HEAD(shmem_swaplist);
231
static DEFINE_MUTEX(shmem_swaplist_mutex);
L
Linus Torvalds 已提交
232 233 234 235

static void shmem_free_blocks(struct inode *inode, long pages)
{
	struct shmem_sb_info *sbinfo = SHMEM_SB(inode->i_sb);
236
	if (sbinfo->max_blocks) {
L
Linus Torvalds 已提交
237 238 239 240 241 242 243
		spin_lock(&sbinfo->stat_lock);
		sbinfo->free_blocks += pages;
		inode->i_blocks -= pages*BLOCKS_PER_PAGE;
		spin_unlock(&sbinfo->stat_lock);
	}
}

244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268
static int shmem_reserve_inode(struct super_block *sb)
{
	struct shmem_sb_info *sbinfo = SHMEM_SB(sb);
	if (sbinfo->max_inodes) {
		spin_lock(&sbinfo->stat_lock);
		if (!sbinfo->free_inodes) {
			spin_unlock(&sbinfo->stat_lock);
			return -ENOSPC;
		}
		sbinfo->free_inodes--;
		spin_unlock(&sbinfo->stat_lock);
	}
	return 0;
}

static void shmem_free_inode(struct super_block *sb)
{
	struct shmem_sb_info *sbinfo = SHMEM_SB(sb);
	if (sbinfo->max_inodes) {
		spin_lock(&sbinfo->stat_lock);
		sbinfo->free_inodes++;
		spin_unlock(&sbinfo->stat_lock);
	}
}

269
/**
L
Linus Torvalds 已提交
270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293
 * shmem_recalc_inode - recalculate the size of an inode
 * @inode: inode to recalc
 *
 * We have to calculate the free blocks since the mm can drop
 * undirtied hole pages behind our back.
 *
 * But normally   info->alloced == inode->i_mapping->nrpages + info->swapped
 * So mm freed is info->alloced - (inode->i_mapping->nrpages + info->swapped)
 *
 * It has to be called with the spinlock held.
 */
static void shmem_recalc_inode(struct inode *inode)
{
	struct shmem_inode_info *info = SHMEM_I(inode);
	long freed;

	freed = info->alloced - info->swapped - inode->i_mapping->nrpages;
	if (freed > 0) {
		info->alloced -= freed;
		shmem_unacct_blocks(info->flags, freed);
		shmem_free_blocks(inode, freed);
	}
}

294
/**
L
Linus Torvalds 已提交
295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390
 * shmem_swp_entry - find the swap vector position in the info structure
 * @info:  info structure for the inode
 * @index: index of the page to find
 * @page:  optional page to add to the structure. Has to be preset to
 *         all zeros
 *
 * If there is no space allocated yet it will return NULL when
 * page is NULL, else it will use the page for the needed block,
 * setting it to NULL on return to indicate that it has been used.
 *
 * The swap vector is organized the following way:
 *
 * There are SHMEM_NR_DIRECT entries directly stored in the
 * shmem_inode_info structure. So small files do not need an addional
 * allocation.
 *
 * For pages with index > SHMEM_NR_DIRECT there is the pointer
 * i_indirect which points to a page which holds in the first half
 * doubly indirect blocks, in the second half triple indirect blocks:
 *
 * For an artificial ENTRIES_PER_PAGE = 4 this would lead to the
 * following layout (for SHMEM_NR_DIRECT == 16):
 *
 * i_indirect -> dir --> 16-19
 * 	      |	     +-> 20-23
 * 	      |
 * 	      +-->dir2 --> 24-27
 * 	      |	       +-> 28-31
 * 	      |	       +-> 32-35
 * 	      |	       +-> 36-39
 * 	      |
 * 	      +-->dir3 --> 40-43
 * 	       	       +-> 44-47
 * 	      	       +-> 48-51
 * 	      	       +-> 52-55
 */
static swp_entry_t *shmem_swp_entry(struct shmem_inode_info *info, unsigned long index, struct page **page)
{
	unsigned long offset;
	struct page **dir;
	struct page *subdir;

	if (index < SHMEM_NR_DIRECT) {
		shmem_swp_balance_unmap();
		return info->i_direct+index;
	}
	if (!info->i_indirect) {
		if (page) {
			info->i_indirect = *page;
			*page = NULL;
		}
		return NULL;			/* need another page */
	}

	index -= SHMEM_NR_DIRECT;
	offset = index % ENTRIES_PER_PAGE;
	index /= ENTRIES_PER_PAGE;
	dir = shmem_dir_map(info->i_indirect);

	if (index >= ENTRIES_PER_PAGE/2) {
		index -= ENTRIES_PER_PAGE/2;
		dir += ENTRIES_PER_PAGE/2 + index/ENTRIES_PER_PAGE;
		index %= ENTRIES_PER_PAGE;
		subdir = *dir;
		if (!subdir) {
			if (page) {
				*dir = *page;
				*page = NULL;
			}
			shmem_dir_unmap(dir);
			return NULL;		/* need another page */
		}
		shmem_dir_unmap(dir);
		dir = shmem_dir_map(subdir);
	}

	dir += index;
	subdir = *dir;
	if (!subdir) {
		if (!page || !(subdir = *page)) {
			shmem_dir_unmap(dir);
			return NULL;		/* need a page */
		}
		*dir = subdir;
		*page = NULL;
	}
	shmem_dir_unmap(dir);
	return shmem_swp_map(subdir) + offset;
}

static void shmem_swp_set(struct shmem_inode_info *info, swp_entry_t *entry, unsigned long value)
{
	long incdec = value? 1: -1;

	entry->val = value;
	info->swapped += incdec;
H
Hugh Dickins 已提交
391 392 393 394
	if ((unsigned long)(entry - info->i_direct) >= SHMEM_NR_DIRECT) {
		struct page *page = kmap_atomic_to_page(entry);
		set_page_private(page, page_private(page) + incdec);
	}
L
Linus Torvalds 已提交
395 396
}

397
/**
L
Linus Torvalds 已提交
398 399 400 401
 * shmem_swp_alloc - get the position of the swap entry for the page.
 * @info:	info structure for the inode
 * @index:	index of the page to find
 * @sgp:	check and recheck i_size? skip allocation?
402 403
 *
 * If the entry does not exist, allocate it.
L
Linus Torvalds 已提交
404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423
 */
static swp_entry_t *shmem_swp_alloc(struct shmem_inode_info *info, unsigned long index, enum sgp_type sgp)
{
	struct inode *inode = &info->vfs_inode;
	struct shmem_sb_info *sbinfo = SHMEM_SB(inode->i_sb);
	struct page *page = NULL;
	swp_entry_t *entry;

	if (sgp != SGP_WRITE &&
	    ((loff_t) index << PAGE_CACHE_SHIFT) >= i_size_read(inode))
		return ERR_PTR(-EINVAL);

	while (!(entry = shmem_swp_entry(info, index, &page))) {
		if (sgp == SGP_READ)
			return shmem_swp_map(ZERO_PAGE(0));
		/*
		 * Test free_blocks against 1 not 0, since we have 1 data
		 * page (and perhaps indirect index pages) yet to allocate:
		 * a waste to allocate index if we cannot allocate data.
		 */
424
		if (sbinfo->max_blocks) {
L
Linus Torvalds 已提交
425 426 427 428 429 430 431 432 433 434 435
			spin_lock(&sbinfo->stat_lock);
			if (sbinfo->free_blocks <= 1) {
				spin_unlock(&sbinfo->stat_lock);
				return ERR_PTR(-ENOSPC);
			}
			sbinfo->free_blocks--;
			inode->i_blocks += BLOCKS_PER_PAGE;
			spin_unlock(&sbinfo->stat_lock);
		}

		spin_unlock(&info->lock);
436
		page = shmem_dir_alloc(mapping_gfp_mask(inode->i_mapping));
H
Hugh Dickins 已提交
437 438
		if (page)
			set_page_private(page, 0);
L
Linus Torvalds 已提交
439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462
		spin_lock(&info->lock);

		if (!page) {
			shmem_free_blocks(inode, 1);
			return ERR_PTR(-ENOMEM);
		}
		if (sgp != SGP_WRITE &&
		    ((loff_t) index << PAGE_CACHE_SHIFT) >= i_size_read(inode)) {
			entry = ERR_PTR(-EINVAL);
			break;
		}
		if (info->next_index <= index)
			info->next_index = index + 1;
	}
	if (page) {
		/* another task gave its page, or truncated the file */
		shmem_free_blocks(inode, 1);
		shmem_dir_free(page);
	}
	if (info->next_index <= index && !IS_ERR(entry))
		info->next_index = index + 1;
	return entry;
}

463
/**
L
Linus Torvalds 已提交
464
 * shmem_free_swp - free some swap entries in a directory
465 466 467
 * @dir:        pointer to the directory
 * @edir:       pointer after last entry of the directory
 * @punch_lock: pointer to spinlock when needed for the holepunch case
L
Linus Torvalds 已提交
468
 */
469 470
static int shmem_free_swp(swp_entry_t *dir, swp_entry_t *edir,
						spinlock_t *punch_lock)
L
Linus Torvalds 已提交
471
{
472
	spinlock_t *punch_unlock = NULL;
L
Linus Torvalds 已提交
473 474 475 476 477
	swp_entry_t *ptr;
	int freed = 0;

	for (ptr = dir; ptr < edir; ptr++) {
		if (ptr->val) {
478 479 480 481 482 483 484
			if (unlikely(punch_lock)) {
				punch_unlock = punch_lock;
				punch_lock = NULL;
				spin_lock(punch_unlock);
				if (!ptr->val)
					continue;
			}
L
Linus Torvalds 已提交
485 486 487 488 489
			free_swap_and_cache(*ptr);
			*ptr = (swp_entry_t){0};
			freed++;
		}
	}
490 491
	if (punch_unlock)
		spin_unlock(punch_unlock);
L
Linus Torvalds 已提交
492 493 494
	return freed;
}

495 496
static int shmem_map_and_free_swp(struct page *subdir, int offset,
		int limit, struct page ***dir, spinlock_t *punch_lock)
L
Linus Torvalds 已提交
497 498 499 500 501 502 503 504 505
{
	swp_entry_t *ptr;
	int freed = 0;

	ptr = shmem_swp_map(subdir);
	for (; offset < limit; offset += LATENCY_LIMIT) {
		int size = limit - offset;
		if (size > LATENCY_LIMIT)
			size = LATENCY_LIMIT;
506 507
		freed += shmem_free_swp(ptr+offset, ptr+offset+size,
							punch_lock);
L
Linus Torvalds 已提交
508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538
		if (need_resched()) {
			shmem_swp_unmap(ptr);
			if (*dir) {
				shmem_dir_unmap(*dir);
				*dir = NULL;
			}
			cond_resched();
			ptr = shmem_swp_map(subdir);
		}
	}
	shmem_swp_unmap(ptr);
	return freed;
}

static void shmem_free_pages(struct list_head *next)
{
	struct page *page;
	int freed = 0;

	do {
		page = container_of(next, struct page, lru);
		next = next->next;
		shmem_dir_free(page);
		freed++;
		if (freed >= LATENCY_LIMIT) {
			cond_resched();
			freed = 0;
		}
	} while (next);
}

539
static void shmem_truncate_range(struct inode *inode, loff_t start, loff_t end)
L
Linus Torvalds 已提交
540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556
{
	struct shmem_inode_info *info = SHMEM_I(inode);
	unsigned long idx;
	unsigned long size;
	unsigned long limit;
	unsigned long stage;
	unsigned long diroff;
	struct page **dir;
	struct page *topdir;
	struct page *middir;
	struct page *subdir;
	swp_entry_t *ptr;
	LIST_HEAD(pages_to_free);
	long nr_pages_to_free = 0;
	long nr_swaps_freed = 0;
	int offset;
	int freed;
557
	int punch_hole;
558 559
	spinlock_t *needs_lock;
	spinlock_t *punch_lock;
560
	unsigned long upper_limit;
L
Linus Torvalds 已提交
561 562

	inode->i_ctime = inode->i_mtime = CURRENT_TIME;
563
	idx = (start + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
L
Linus Torvalds 已提交
564 565 566 567 568
	if (idx >= info->next_index)
		return;

	spin_lock(&info->lock);
	info->flags |= SHMEM_TRUNCATE;
569 570
	if (likely(end == (loff_t) -1)) {
		limit = info->next_index;
571
		upper_limit = SHMEM_MAX_INDEX;
572
		info->next_index = idx;
573
		needs_lock = NULL;
574
		punch_hole = 0;
575
	} else {
576 577 578 579 580 581 582 583
		if (end + 1 >= inode->i_size) {	/* we may free a little more */
			limit = (inode->i_size + PAGE_CACHE_SIZE - 1) >>
							PAGE_CACHE_SHIFT;
			upper_limit = SHMEM_MAX_INDEX;
		} else {
			limit = (end + 1) >> PAGE_CACHE_SHIFT;
			upper_limit = limit;
		}
584
		needs_lock = &info->lock;
585 586 587
		punch_hole = 1;
	}

L
Linus Torvalds 已提交
588
	topdir = info->i_indirect;
589
	if (topdir && idx <= SHMEM_NR_DIRECT && !punch_hole) {
L
Linus Torvalds 已提交
590 591 592 593 594 595 596 597 598 599 600
		info->i_indirect = NULL;
		nr_pages_to_free++;
		list_add(&topdir->lru, &pages_to_free);
	}
	spin_unlock(&info->lock);

	if (info->swapped && idx < SHMEM_NR_DIRECT) {
		ptr = info->i_direct;
		size = limit;
		if (size > SHMEM_NR_DIRECT)
			size = SHMEM_NR_DIRECT;
601
		nr_swaps_freed = shmem_free_swp(ptr+idx, ptr+size, needs_lock);
L
Linus Torvalds 已提交
602
	}
603 604 605 606 607

	/*
	 * If there are no indirect blocks or we are punching a hole
	 * below indirect blocks, nothing to be done.
	 */
608
	if (!topdir || limit <= SHMEM_NR_DIRECT)
L
Linus Torvalds 已提交
609 610
		goto done2;

611 612 613 614 615 616 617 618 619 620 621 622 623
	/*
	 * The truncation case has already dropped info->lock, and we're safe
	 * because i_size and next_index have already been lowered, preventing
	 * access beyond.  But in the punch_hole case, we still need to take
	 * the lock when updating the swap directory, because there might be
	 * racing accesses by shmem_getpage(SGP_CACHE), shmem_unuse_inode or
	 * shmem_writepage.  However, whenever we find we can remove a whole
	 * directory page (not at the misaligned start or end of the range),
	 * we first NULLify its pointer in the level above, and then have no
	 * need to take the lock when updating its contents: needs_lock and
	 * punch_lock (either pointing to info->lock or NULL) manage this.
	 */

624
	upper_limit -= SHMEM_NR_DIRECT;
L
Linus Torvalds 已提交
625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643
	limit -= SHMEM_NR_DIRECT;
	idx = (idx > SHMEM_NR_DIRECT)? (idx - SHMEM_NR_DIRECT): 0;
	offset = idx % ENTRIES_PER_PAGE;
	idx -= offset;

	dir = shmem_dir_map(topdir);
	stage = ENTRIES_PER_PAGEPAGE/2;
	if (idx < ENTRIES_PER_PAGEPAGE/2) {
		middir = topdir;
		diroff = idx/ENTRIES_PER_PAGE;
	} else {
		dir += ENTRIES_PER_PAGE/2;
		dir += (idx - ENTRIES_PER_PAGEPAGE/2)/ENTRIES_PER_PAGEPAGE;
		while (stage <= idx)
			stage += ENTRIES_PER_PAGEPAGE;
		middir = *dir;
		if (*dir) {
			diroff = ((idx - ENTRIES_PER_PAGEPAGE/2) %
				ENTRIES_PER_PAGEPAGE) / ENTRIES_PER_PAGE;
644
			if (!diroff && !offset && upper_limit >= stage) {
645 646 647 648 649 650 651
				if (needs_lock) {
					spin_lock(needs_lock);
					*dir = NULL;
					spin_unlock(needs_lock);
					needs_lock = NULL;
				} else
					*dir = NULL;
L
Linus Torvalds 已提交
652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676
				nr_pages_to_free++;
				list_add(&middir->lru, &pages_to_free);
			}
			shmem_dir_unmap(dir);
			dir = shmem_dir_map(middir);
		} else {
			diroff = 0;
			offset = 0;
			idx = stage;
		}
	}

	for (; idx < limit; idx += ENTRIES_PER_PAGE, diroff++) {
		if (unlikely(idx == stage)) {
			shmem_dir_unmap(dir);
			dir = shmem_dir_map(topdir) +
			    ENTRIES_PER_PAGE/2 + idx/ENTRIES_PER_PAGEPAGE;
			while (!*dir) {
				dir++;
				idx += ENTRIES_PER_PAGEPAGE;
				if (idx >= limit)
					goto done1;
			}
			stage = idx + ENTRIES_PER_PAGEPAGE;
			middir = *dir;
677 678
			if (punch_hole)
				needs_lock = &info->lock;
679
			if (upper_limit >= stage) {
680 681 682 683 684 685 686
				if (needs_lock) {
					spin_lock(needs_lock);
					*dir = NULL;
					spin_unlock(needs_lock);
					needs_lock = NULL;
				} else
					*dir = NULL;
687 688 689
				nr_pages_to_free++;
				list_add(&middir->lru, &pages_to_free);
			}
L
Linus Torvalds 已提交
690 691 692 693 694
			shmem_dir_unmap(dir);
			cond_resched();
			dir = shmem_dir_map(middir);
			diroff = 0;
		}
695
		punch_lock = needs_lock;
L
Linus Torvalds 已提交
696
		subdir = dir[diroff];
697 698 699 700 701 702 703 704 705 706 707 708
		if (subdir && !offset && upper_limit-idx >= ENTRIES_PER_PAGE) {
			if (needs_lock) {
				spin_lock(needs_lock);
				dir[diroff] = NULL;
				spin_unlock(needs_lock);
				punch_lock = NULL;
			} else
				dir[diroff] = NULL;
			nr_pages_to_free++;
			list_add(&subdir->lru, &pages_to_free);
		}
		if (subdir && page_private(subdir) /* has swap entries */) {
L
Linus Torvalds 已提交
709 710 711 712
			size = limit - idx;
			if (size > ENTRIES_PER_PAGE)
				size = ENTRIES_PER_PAGE;
			freed = shmem_map_and_free_swp(subdir,
713
					offset, size, &dir, punch_lock);
L
Linus Torvalds 已提交
714 715 716
			if (!dir)
				dir = shmem_dir_map(middir);
			nr_swaps_freed += freed;
717
			if (offset || punch_lock) {
L
Linus Torvalds 已提交
718
				spin_lock(&info->lock);
719 720
				set_page_private(subdir,
					page_private(subdir) - freed);
L
Linus Torvalds 已提交
721
				spin_unlock(&info->lock);
722 723
			} else
				BUG_ON(page_private(subdir) != freed);
L
Linus Torvalds 已提交
724
		}
725
		offset = 0;
L
Linus Torvalds 已提交
726 727 728 729 730 731 732 733 734 735 736
	}
done1:
	shmem_dir_unmap(dir);
done2:
	if (inode->i_mapping->nrpages && (info->flags & SHMEM_PAGEIN)) {
		/*
		 * Call truncate_inode_pages again: racing shmem_unuse_inode
		 * may have swizzled a page in from swap since vmtruncate or
		 * generic_delete_inode did it, before we lowered next_index.
		 * Also, though shmem_getpage checks i_size before adding to
		 * cache, no recheck after: so fix the narrow window there too.
737 738 739 740 741
		 *
		 * Recalling truncate_inode_pages_range and unmap_mapping_range
		 * every time for punch_hole (which never got a chance to clear
		 * SHMEM_PAGEIN at the start of vmtruncate_range) is expensive,
		 * yet hardly ever necessary: try to optimize them out later.
L
Linus Torvalds 已提交
742
		 */
743
		truncate_inode_pages_range(inode->i_mapping, start, end);
744 745 746
		if (punch_hole)
			unmap_mapping_range(inode->i_mapping, start,
							end - start, 1);
L
Linus Torvalds 已提交
747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765
	}

	spin_lock(&info->lock);
	info->flags &= ~SHMEM_TRUNCATE;
	info->swapped -= nr_swaps_freed;
	if (nr_pages_to_free)
		shmem_free_blocks(inode, nr_pages_to_free);
	shmem_recalc_inode(inode);
	spin_unlock(&info->lock);

	/*
	 * Empty swap vector directory pages to be freed?
	 */
	if (!list_empty(&pages_to_free)) {
		pages_to_free.prev->next = NULL;
		shmem_free_pages(pages_to_free.next);
	}
}

766 767 768 769 770
static void shmem_truncate(struct inode *inode)
{
	shmem_truncate_range(inode, inode->i_size, (loff_t)-1);
}

L
Linus Torvalds 已提交
771 772 773 774 775 776
static int shmem_notify_change(struct dentry *dentry, struct iattr *attr)
{
	struct inode *inode = dentry->d_inode;
	struct page *page = NULL;
	int error;

777
	if (S_ISREG(inode->i_mode) && (attr->ia_valid & ATTR_SIZE)) {
L
Linus Torvalds 已提交
778 779 780 781 782 783 784 785 786 787 788 789
		if (attr->ia_size < inode->i_size) {
			/*
			 * If truncating down to a partial page, then
			 * if that page is already allocated, hold it
			 * in memory until the truncation is over, so
			 * truncate_partial_page cannnot miss it were
			 * it assigned to swap.
			 */
			if (attr->ia_size & (PAGE_CACHE_SIZE-1)) {
				(void) shmem_getpage(inode,
					attr->ia_size>>PAGE_CACHE_SHIFT,
						&page, SGP_READ, NULL);
H
Hugh Dickins 已提交
790 791
				if (page)
					unlock_page(page);
L
Linus Torvalds 已提交
792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811
			}
			/*
			 * Reset SHMEM_PAGEIN flag so that shmem_truncate can
			 * detect if any pages might have been added to cache
			 * after truncate_inode_pages.  But we needn't bother
			 * if it's being fully truncated to zero-length: the
			 * nrpages check is efficient enough in that case.
			 */
			if (attr->ia_size) {
				struct shmem_inode_info *info = SHMEM_I(inode);
				spin_lock(&info->lock);
				info->flags &= ~SHMEM_PAGEIN;
				spin_unlock(&info->lock);
			}
		}
	}

	error = inode_change_ok(inode, attr);
	if (!error)
		error = inode_setattr(inode, attr);
812 813 814 815
#ifdef CONFIG_TMPFS_POSIX_ACL
	if (!error && (attr->ia_valid & ATTR_MODE))
		error = generic_acl_chmod(inode, &shmem_acl_ops);
#endif
L
Linus Torvalds 已提交
816 817 818 819 820 821 822 823 824 825
	if (page)
		page_cache_release(page);
	return error;
}

static void shmem_delete_inode(struct inode *inode)
{
	struct shmem_inode_info *info = SHMEM_I(inode);

	if (inode->i_op->truncate == shmem_truncate) {
826
		truncate_inode_pages(inode->i_mapping, 0);
L
Linus Torvalds 已提交
827 828 829 830
		shmem_unacct_size(info->flags, inode->i_size);
		inode->i_size = 0;
		shmem_truncate(inode);
		if (!list_empty(&info->swaplist)) {
831
			mutex_lock(&shmem_swaplist_mutex);
L
Linus Torvalds 已提交
832
			list_del_init(&info->swaplist);
833
			mutex_unlock(&shmem_swaplist_mutex);
L
Linus Torvalds 已提交
834 835
		}
	}
836
	BUG_ON(inode->i_blocks);
837
	shmem_free_inode(inode->i_sb);
L
Linus Torvalds 已提交
838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862
	clear_inode(inode);
}

static inline int shmem_find_swp(swp_entry_t entry, swp_entry_t *dir, swp_entry_t *edir)
{
	swp_entry_t *ptr;

	for (ptr = dir; ptr < edir; ptr++) {
		if (ptr->val == entry.val)
			return ptr - dir;
	}
	return -1;
}

static int shmem_unuse_inode(struct shmem_inode_info *info, swp_entry_t entry, struct page *page)
{
	struct inode *inode;
	unsigned long idx;
	unsigned long size;
	unsigned long limit;
	unsigned long stage;
	struct page **dir;
	struct page *subdir;
	swp_entry_t *ptr;
	int offset;
863
	int error;
L
Linus Torvalds 已提交
864 865 866 867

	idx = 0;
	ptr = info->i_direct;
	spin_lock(&info->lock);
H
Hugh Dickins 已提交
868 869 870 871
	if (!info->swapped) {
		list_del_init(&info->swaplist);
		goto lost2;
	}
L
Linus Torvalds 已提交
872 873 874 875 876
	limit = info->next_index;
	size = limit;
	if (size > SHMEM_NR_DIRECT)
		size = SHMEM_NR_DIRECT;
	offset = shmem_find_swp(entry, ptr, ptr+size);
877
	if (offset >= 0)
L
Linus Torvalds 已提交
878 879 880 881 882 883 884 885 886 887
		goto found;
	if (!info->i_indirect)
		goto lost2;

	dir = shmem_dir_map(info->i_indirect);
	stage = SHMEM_NR_DIRECT + ENTRIES_PER_PAGEPAGE/2;

	for (idx = SHMEM_NR_DIRECT; idx < limit; idx += ENTRIES_PER_PAGE, dir++) {
		if (unlikely(idx == stage)) {
			shmem_dir_unmap(dir-1);
888 889 890 891 892 893 894 895
			if (cond_resched_lock(&info->lock)) {
				/* check it has not been truncated */
				if (limit > info->next_index) {
					limit = info->next_index;
					if (idx >= limit)
						goto lost2;
				}
			}
L
Linus Torvalds 已提交
896 897 898 899 900 901 902 903 904 905 906 907 908 909
			dir = shmem_dir_map(info->i_indirect) +
			    ENTRIES_PER_PAGE/2 + idx/ENTRIES_PER_PAGEPAGE;
			while (!*dir) {
				dir++;
				idx += ENTRIES_PER_PAGEPAGE;
				if (idx >= limit)
					goto lost1;
			}
			stage = idx + ENTRIES_PER_PAGEPAGE;
			subdir = *dir;
			shmem_dir_unmap(dir);
			dir = shmem_dir_map(subdir);
		}
		subdir = *dir;
H
Hugh Dickins 已提交
910
		if (subdir && page_private(subdir)) {
L
Linus Torvalds 已提交
911 912 913 914 915
			ptr = shmem_swp_map(subdir);
			size = limit - idx;
			if (size > ENTRIES_PER_PAGE)
				size = ENTRIES_PER_PAGE;
			offset = shmem_find_swp(entry, ptr, ptr+size);
916
			shmem_swp_unmap(ptr);
L
Linus Torvalds 已提交
917 918 919 920 921 922 923 924 925 926 927 928 929
			if (offset >= 0) {
				shmem_dir_unmap(dir);
				goto found;
			}
		}
	}
lost1:
	shmem_dir_unmap(dir-1);
lost2:
	spin_unlock(&info->lock);
	return 0;
found:
	idx += offset;
930 931 932
	inode = igrab(&info->vfs_inode);
	spin_unlock(&info->lock);

H
Hugh Dickins 已提交
933 934 935 936 937 938 939 940 941
	/*
	 * Move _head_ to start search for next from here.
	 * But be careful: shmem_delete_inode checks list_empty without taking
	 * mutex, and there's an instant in list_move_tail when info->swaplist
	 * would appear empty, if it were the only one on shmem_swaplist.  We
	 * could avoid doing it if inode NULL; or use this minor optimization.
	 */
	if (shmem_swaplist.next != &info->swaplist)
		list_move_tail(&shmem_swaplist, &info->swaplist);
942 943 944 945 946
	mutex_unlock(&shmem_swaplist_mutex);

	error = 1;
	if (!inode)
		goto out;
K
KAMEZAWA Hiroyuki 已提交
947
	/*
948 949 950
	 * Charge page using GFP_KERNEL while we can wait.
	 * Charged back to the user(not to caller) when swap account is used.
	 * add_to_page_cache() will be called with GFP_NOWAIT.
K
KAMEZAWA Hiroyuki 已提交
951
	 */
952
	error = mem_cgroup_cache_charge(page, current->mm, GFP_KERNEL);
H
Hugh Dickins 已提交
953 954
	if (error)
		goto out;
955
	error = radix_tree_preload(GFP_KERNEL);
956 957 958 959
	if (error) {
		mem_cgroup_uncharge_cache_page(page);
		goto out;
	}
H
Hugh Dickins 已提交
960
	error = 1;
961 962 963

	spin_lock(&info->lock);
	ptr = shmem_swp_entry(info, idx, NULL);
964
	if (ptr && ptr->val == entry.val) {
N
Nick Piggin 已提交
965
		error = add_to_page_cache_locked(page, inode->i_mapping,
H
Hugh Dickins 已提交
966
						idx, GFP_NOWAIT);
967 968 969 970
		/* does mem_cgroup_uncharge_cache_page on error */
	} else	/* we must compensate for our precharge above */
		mem_cgroup_uncharge_cache_page(page);

971 972
	if (error == -EEXIST) {
		struct page *filepage = find_get_page(inode->i_mapping, idx);
973
		error = 1;
974 975 976 977 978 979 980 981 982 983 984
		if (filepage) {
			/*
			 * There might be a more uptodate page coming down
			 * from a stacked writepage: forget our swappage if so.
			 */
			if (PageUptodate(filepage))
				error = 0;
			page_cache_release(filepage);
		}
	}
	if (!error) {
985 986
		delete_from_swap_cache(page);
		set_page_dirty(page);
L
Linus Torvalds 已提交
987
		info->flags |= SHMEM_PAGEIN;
988 989 990
		shmem_swp_set(info, ptr, 0);
		swap_free(entry);
		error = 1;	/* not an error, but entry was found */
L
Linus Torvalds 已提交
991
	}
992 993
	if (ptr)
		shmem_swp_unmap(ptr);
L
Linus Torvalds 已提交
994
	spin_unlock(&info->lock);
H
Hugh Dickins 已提交
995
	radix_tree_preload_end();
996 997 998 999 1000
out:
	unlock_page(page);
	page_cache_release(page);
	iput(inode);		/* allows for NULL */
	return error;
L
Linus Torvalds 已提交
1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011
}

/*
 * shmem_unuse() search for an eventually swapped out shmem page.
 */
int shmem_unuse(swp_entry_t entry, struct page *page)
{
	struct list_head *p, *next;
	struct shmem_inode_info *info;
	int found = 0;

1012
	mutex_lock(&shmem_swaplist_mutex);
L
Linus Torvalds 已提交
1013 1014
	list_for_each_safe(p, next, &shmem_swaplist) {
		info = list_entry(p, struct shmem_inode_info, swaplist);
H
Hugh Dickins 已提交
1015
		found = shmem_unuse_inode(info, entry, page);
1016
		cond_resched();
1017 1018
		if (found)
			goto out;
L
Linus Torvalds 已提交
1019
	}
1020
	mutex_unlock(&shmem_swaplist_mutex);
1021
out:	return found;	/* 0 or 1 or -ENOMEM */
L
Linus Torvalds 已提交
1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041
}

/*
 * Move the page from the page cache to the swap cache.
 */
static int shmem_writepage(struct page *page, struct writeback_control *wbc)
{
	struct shmem_inode_info *info;
	swp_entry_t *entry, swap;
	struct address_space *mapping;
	unsigned long index;
	struct inode *inode;

	BUG_ON(!PageLocked(page));
	mapping = page->mapping;
	index = page->index;
	inode = mapping->host;
	info = SHMEM_I(inode);
	if (info->flags & VM_LOCKED)
		goto redirty;
1042
	if (!total_swap_pages)
L
Linus Torvalds 已提交
1043 1044
		goto redirty;

1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057
	/*
	 * shmem_backing_dev_info's capabilities prevent regular writeback or
	 * sync from ever calling shmem_writepage; but a stacking filesystem
	 * may use the ->writepage of its underlying filesystem, in which case
	 * tmpfs should write out to swap only in response to memory pressure,
	 * and not for pdflush or sync.  However, in those cases, we do still
	 * want to check if there's a redundant swappage to be discarded.
	 */
	if (wbc->for_reclaim)
		swap = get_swap_page();
	else
		swap.val = 0;

L
Linus Torvalds 已提交
1058 1059 1060 1061 1062 1063
	spin_lock(&info->lock);
	if (index >= info->next_index) {
		BUG_ON(!(info->flags & SHMEM_TRUNCATE));
		goto unlock;
	}
	entry = shmem_swp_entry(info, index, NULL);
1064 1065 1066 1067 1068 1069 1070 1071 1072
	if (entry->val) {
		/*
		 * The more uptodate page coming down from a stacked
		 * writepage should replace our old swappage.
		 */
		free_swap_and_cache(*entry);
		shmem_swp_set(info, entry, 0);
	}
	shmem_recalc_inode(inode);
L
Linus Torvalds 已提交
1073

1074
	if (swap.val && add_to_swap_cache(page, swap, GFP_ATOMIC) == 0) {
1075
		remove_from_page_cache(page);
L
Linus Torvalds 已提交
1076 1077
		shmem_swp_set(info, entry, swap.val);
		shmem_swp_unmap(entry);
H
Hugh Dickins 已提交
1078 1079 1080 1081
		if (list_empty(&info->swaplist))
			inode = igrab(inode);
		else
			inode = NULL;
L
Linus Torvalds 已提交
1082
		spin_unlock(&info->lock);
1083
		swap_duplicate(swap);
1084
		BUG_ON(page_mapped(page));
1085
		page_cache_release(page);	/* pagecache ref */
H
Hugh Dickins 已提交
1086
		swap_writepage(page, wbc);
H
Hugh Dickins 已提交
1087 1088 1089 1090 1091 1092 1093
		if (inode) {
			mutex_lock(&shmem_swaplist_mutex);
			/* move instead of add in case we're racing */
			list_move_tail(&info->swaplist, &shmem_swaplist);
			mutex_unlock(&shmem_swaplist_mutex);
			iput(inode);
		}
L
Linus Torvalds 已提交
1094 1095 1096 1097 1098 1099 1100 1101 1102
		return 0;
	}

	shmem_swp_unmap(entry);
unlock:
	spin_unlock(&info->lock);
	swap_free(swap);
redirty:
	set_page_dirty(page);
1103 1104 1105 1106
	if (wbc->for_reclaim)
		return AOP_WRITEPAGE_ACTIVATE;	/* Return with page locked */
	unlock_page(page);
	return 0;
L
Linus Torvalds 已提交
1107 1108 1109
}

#ifdef CONFIG_NUMA
1110
#ifdef CONFIG_TMPFS
1111
static void shmem_show_mpol(struct seq_file *seq, struct mempolicy *mpol)
1112
{
1113
	char buffer[64];
1114

1115
	if (!mpol || mpol->mode == MPOL_DEFAULT)
1116
		return;		/* show nothing */
1117

1118
	mpol_to_str(buffer, sizeof(buffer), mpol, 1);
1119 1120

	seq_printf(seq, ",mpol=%s", buffer);
1121
}
1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133

static struct mempolicy *shmem_get_sbmpol(struct shmem_sb_info *sbinfo)
{
	struct mempolicy *mpol = NULL;
	if (sbinfo->mpol) {
		spin_lock(&sbinfo->stat_lock);	/* prevent replace/use races */
		mpol = sbinfo->mpol;
		mpol_get(mpol);
		spin_unlock(&sbinfo->stat_lock);
	}
	return mpol;
}
1134 1135
#endif /* CONFIG_TMPFS */

1136 1137
static struct page *shmem_swapin(swp_entry_t entry, gfp_t gfp,
			struct shmem_inode_info *info, unsigned long idx)
L
Linus Torvalds 已提交
1138
{
1139
	struct mempolicy mpol, *spol;
L
Linus Torvalds 已提交
1140
	struct vm_area_struct pvma;
1141
	struct page *page;
L
Linus Torvalds 已提交
1142

1143 1144 1145
	spol = mpol_cond_copy(&mpol,
				mpol_shared_policy_lookup(&info->policy, idx));

L
Linus Torvalds 已提交
1146
	/* Create a pseudo vma that just contains the policy */
1147
	pvma.vm_start = 0;
L
Linus Torvalds 已提交
1148
	pvma.vm_pgoff = idx;
1149
	pvma.vm_ops = NULL;
1150
	pvma.vm_policy = spol;
1151
	page = swapin_readahead(entry, gfp, &pvma, 0);
L
Linus Torvalds 已提交
1152 1153 1154
	return page;
}

1155 1156
static struct page *shmem_alloc_page(gfp_t gfp,
			struct shmem_inode_info *info, unsigned long idx)
L
Linus Torvalds 已提交
1157 1158 1159
{
	struct vm_area_struct pvma;

1160 1161
	/* Create a pseudo vma that just contains the policy */
	pvma.vm_start = 0;
L
Linus Torvalds 已提交
1162
	pvma.vm_pgoff = idx;
1163 1164
	pvma.vm_ops = NULL;
	pvma.vm_policy = mpol_shared_policy_lookup(&info->policy, idx);
1165 1166 1167 1168 1169

	/*
	 * alloc_page_vma() will drop the shared policy reference
	 */
	return alloc_page_vma(gfp, &pvma, 0);
L
Linus Torvalds 已提交
1170
}
1171 1172
#else /* !CONFIG_NUMA */
#ifdef CONFIG_TMPFS
1173
static inline void shmem_show_mpol(struct seq_file *seq, struct mempolicy *p)
1174 1175 1176 1177
{
}
#endif /* CONFIG_TMPFS */

1178 1179
static inline struct page *shmem_swapin(swp_entry_t entry, gfp_t gfp,
			struct shmem_inode_info *info, unsigned long idx)
L
Linus Torvalds 已提交
1180
{
1181
	return swapin_readahead(entry, gfp, NULL, 0);
L
Linus Torvalds 已提交
1182 1183
}

1184 1185
static inline struct page *shmem_alloc_page(gfp_t gfp,
			struct shmem_inode_info *info, unsigned long idx)
L
Linus Torvalds 已提交
1186
{
1187
	return alloc_page(gfp);
L
Linus Torvalds 已提交
1188
}
1189
#endif /* CONFIG_NUMA */
L
Linus Torvalds 已提交
1190

1191 1192 1193 1194 1195 1196 1197
#if !defined(CONFIG_NUMA) || !defined(CONFIG_TMPFS)
static inline struct mempolicy *shmem_get_sbmpol(struct shmem_sb_info *sbinfo)
{
	return NULL;
}
#endif

L
Linus Torvalds 已提交
1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214
/*
 * shmem_getpage - either get the page from swap or allocate a new one
 *
 * If we allocate a new one we do not mark it dirty. That's up to the
 * vm. If we swap it in we mark it dirty since we also free the swap
 * entry since a page cannot live in both the swap and page cache
 */
static int shmem_getpage(struct inode *inode, unsigned long idx,
			struct page **pagep, enum sgp_type sgp, int *type)
{
	struct address_space *mapping = inode->i_mapping;
	struct shmem_inode_info *info = SHMEM_I(inode);
	struct shmem_sb_info *sbinfo;
	struct page *filepage = *pagep;
	struct page *swappage;
	swp_entry_t *entry;
	swp_entry_t swap;
1215
	gfp_t gfp;
L
Linus Torvalds 已提交
1216 1217 1218 1219
	int error;

	if (idx >= SHMEM_MAX_INDEX)
		return -EFBIG;
1220 1221

	if (type)
N
Nick Piggin 已提交
1222
		*type = 0;
1223

L
Linus Torvalds 已提交
1224 1225 1226 1227
	/*
	 * Normally, filepage is NULL on entry, and either found
	 * uptodate immediately, or allocated and zeroed, or read
	 * in under swappage, which is then assigned to filepage.
H
Hugh Dickins 已提交
1228
	 * But shmem_readpage (required for splice) passes in a locked
1229 1230
	 * filepage, which may be found not uptodate by other callers
	 * too, and may need to be copied from the swappage read in.
L
Linus Torvalds 已提交
1231 1232 1233 1234 1235 1236 1237
	 */
repeat:
	if (!filepage)
		filepage = find_lock_page(mapping, idx);
	if (filepage && PageUptodate(filepage))
		goto done;
	error = 0;
1238
	gfp = mapping_gfp_mask(mapping);
H
Hugh Dickins 已提交
1239 1240 1241 1242 1243 1244 1245 1246 1247 1248
	if (!filepage) {
		/*
		 * Try to preload while we can wait, to not make a habit of
		 * draining atomic reserves; but don't latch on to this cpu.
		 */
		error = radix_tree_preload(gfp & ~__GFP_HIGHMEM);
		if (error)
			goto failed;
		radix_tree_preload_end();
	}
L
Linus Torvalds 已提交
1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265

	spin_lock(&info->lock);
	shmem_recalc_inode(inode);
	entry = shmem_swp_alloc(info, idx, sgp);
	if (IS_ERR(entry)) {
		spin_unlock(&info->lock);
		error = PTR_ERR(entry);
		goto failed;
	}
	swap = *entry;

	if (swap.val) {
		/* Look it up and read it in.. */
		swappage = lookup_swap_cache(swap);
		if (!swappage) {
			shmem_swp_unmap(entry);
			/* here we actually do the io */
N
Nick Piggin 已提交
1266
			if (type && !(*type & VM_FAULT_MAJOR)) {
1267
				__count_vm_event(PGMAJFAULT);
N
Nick Piggin 已提交
1268
				*type |= VM_FAULT_MAJOR;
L
Linus Torvalds 已提交
1269
			}
1270
			spin_unlock(&info->lock);
1271
			swappage = shmem_swapin(swap, gfp, info, idx);
L
Linus Torvalds 已提交
1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292
			if (!swappage) {
				spin_lock(&info->lock);
				entry = shmem_swp_alloc(info, idx, sgp);
				if (IS_ERR(entry))
					error = PTR_ERR(entry);
				else {
					if (entry->val == swap.val)
						error = -ENOMEM;
					shmem_swp_unmap(entry);
				}
				spin_unlock(&info->lock);
				if (error)
					goto failed;
				goto repeat;
			}
			wait_on_page_locked(swappage);
			page_cache_release(swappage);
			goto repeat;
		}

		/* We have to do this with page locked to prevent races */
N
Nick Piggin 已提交
1293
		if (!trylock_page(swappage)) {
L
Linus Torvalds 已提交
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 1327 1328
			shmem_swp_unmap(entry);
			spin_unlock(&info->lock);
			wait_on_page_locked(swappage);
			page_cache_release(swappage);
			goto repeat;
		}
		if (PageWriteback(swappage)) {
			shmem_swp_unmap(entry);
			spin_unlock(&info->lock);
			wait_on_page_writeback(swappage);
			unlock_page(swappage);
			page_cache_release(swappage);
			goto repeat;
		}
		if (!PageUptodate(swappage)) {
			shmem_swp_unmap(entry);
			spin_unlock(&info->lock);
			unlock_page(swappage);
			page_cache_release(swappage);
			error = -EIO;
			goto failed;
		}

		if (filepage) {
			shmem_swp_set(info, entry, 0);
			shmem_swp_unmap(entry);
			delete_from_swap_cache(swappage);
			spin_unlock(&info->lock);
			copy_highpage(filepage, swappage);
			unlock_page(swappage);
			page_cache_release(swappage);
			flush_dcache_page(filepage);
			SetPageUptodate(filepage);
			set_page_dirty(filepage);
			swap_free(swap);
N
Nick Piggin 已提交
1329 1330
		} else if (!(error = add_to_page_cache_locked(swappage, mapping,
					idx, GFP_NOWAIT))) {
L
Linus Torvalds 已提交
1331 1332 1333
			info->flags |= SHMEM_PAGEIN;
			shmem_swp_set(info, entry, 0);
			shmem_swp_unmap(entry);
1334
			delete_from_swap_cache(swappage);
L
Linus Torvalds 已提交
1335 1336
			spin_unlock(&info->lock);
			filepage = swappage;
1337
			set_page_dirty(filepage);
L
Linus Torvalds 已提交
1338 1339 1340 1341
			swap_free(swap);
		} else {
			shmem_swp_unmap(entry);
			spin_unlock(&info->lock);
1342
			if (error == -ENOMEM) {
1343 1344 1345 1346 1347 1348
				/*
				 * reclaim from proper memory cgroup and
				 * call memcg's OOM if needed.
				 */
				error = mem_cgroup_shmem_charge_fallback(
								swappage,
1349
								current->mm,
1350
								gfp);
1351 1352 1353
				if (error) {
					unlock_page(swappage);
					page_cache_release(swappage);
1354
					goto failed;
1355
				}
1356
			}
1357 1358
			unlock_page(swappage);
			page_cache_release(swappage);
L
Linus Torvalds 已提交
1359 1360 1361 1362 1363 1364
			goto repeat;
		}
	} else if (sgp == SGP_READ && !filepage) {
		shmem_swp_unmap(entry);
		filepage = find_get_page(mapping, idx);
		if (filepage &&
N
Nick Piggin 已提交
1365
		    (!PageUptodate(filepage) || !trylock_page(filepage))) {
L
Linus Torvalds 已提交
1366 1367 1368 1369 1370 1371 1372 1373 1374 1375
			spin_unlock(&info->lock);
			wait_on_page_locked(filepage);
			page_cache_release(filepage);
			filepage = NULL;
			goto repeat;
		}
		spin_unlock(&info->lock);
	} else {
		shmem_swp_unmap(entry);
		sbinfo = SHMEM_SB(inode->i_sb);
1376
		if (sbinfo->max_blocks) {
L
Linus Torvalds 已提交
1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394
			spin_lock(&sbinfo->stat_lock);
			if (sbinfo->free_blocks == 0 ||
			    shmem_acct_block(info->flags)) {
				spin_unlock(&sbinfo->stat_lock);
				spin_unlock(&info->lock);
				error = -ENOSPC;
				goto failed;
			}
			sbinfo->free_blocks--;
			inode->i_blocks += BLOCKS_PER_PAGE;
			spin_unlock(&sbinfo->stat_lock);
		} else if (shmem_acct_block(info->flags)) {
			spin_unlock(&info->lock);
			error = -ENOSPC;
			goto failed;
		}

		if (!filepage) {
1395 1396
			int ret;

L
Linus Torvalds 已提交
1397
			spin_unlock(&info->lock);
1398
			filepage = shmem_alloc_page(gfp, info, idx);
L
Linus Torvalds 已提交
1399 1400 1401 1402 1403 1404
			if (!filepage) {
				shmem_unacct_blocks(info->flags, 1);
				shmem_free_blocks(inode, 1);
				error = -ENOMEM;
				goto failed;
			}
R
Rik van Riel 已提交
1405
			SetPageSwapBacked(filepage);
L
Linus Torvalds 已提交
1406

1407 1408
			/* Precharge page while we can wait, compensate after */
			error = mem_cgroup_cache_charge(filepage, current->mm,
K
KAMEZAWA Hiroyuki 已提交
1409
					GFP_KERNEL);
1410 1411 1412 1413 1414 1415 1416 1417
			if (error) {
				page_cache_release(filepage);
				shmem_unacct_blocks(info->flags, 1);
				shmem_free_blocks(inode, 1);
				filepage = NULL;
				goto failed;
			}

L
Linus Torvalds 已提交
1418 1419 1420 1421 1422 1423 1424 1425
			spin_lock(&info->lock);
			entry = shmem_swp_alloc(info, idx, sgp);
			if (IS_ERR(entry))
				error = PTR_ERR(entry);
			else {
				swap = *entry;
				shmem_swp_unmap(entry);
			}
1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436
			ret = error || swap.val;
			if (ret)
				mem_cgroup_uncharge_cache_page(filepage);
			else
				ret = add_to_page_cache_lru(filepage, mapping,
						idx, GFP_NOWAIT);
			/*
			 * At add_to_page_cache_lru() failure, uncharge will
			 * be done automatically.
			 */
			if (ret) {
L
Linus Torvalds 已提交
1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450
				spin_unlock(&info->lock);
				page_cache_release(filepage);
				shmem_unacct_blocks(info->flags, 1);
				shmem_free_blocks(inode, 1);
				filepage = NULL;
				if (error)
					goto failed;
				goto repeat;
			}
			info->flags |= SHMEM_PAGEIN;
		}

		info->alloced++;
		spin_unlock(&info->lock);
1451
		clear_highpage(filepage);
L
Linus Torvalds 已提交
1452 1453
		flush_dcache_page(filepage);
		SetPageUptodate(filepage);
1454 1455
		if (sgp == SGP_DIRTY)
			set_page_dirty(filepage);
L
Linus Torvalds 已提交
1456 1457
	}
done:
H
Hugh Dickins 已提交
1458
	*pagep = filepage;
L
Linus Torvalds 已提交
1459 1460 1461 1462 1463 1464 1465 1466 1467 1468
	return 0;

failed:
	if (*pagep != filepage) {
		unlock_page(filepage);
		page_cache_release(filepage);
	}
	return error;
}

N
Nick Piggin 已提交
1469
static int shmem_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
L
Linus Torvalds 已提交
1470
{
1471
	struct inode *inode = vma->vm_file->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
1472
	int error;
N
Nick Piggin 已提交
1473
	int ret;
L
Linus Torvalds 已提交
1474

N
Nick Piggin 已提交
1475 1476
	if (((loff_t)vmf->pgoff << PAGE_CACHE_SHIFT) >= i_size_read(inode))
		return VM_FAULT_SIGBUS;
1477

1478
	error = shmem_getpage(inode, vmf->pgoff, &vmf->page, SGP_CACHE, &ret);
N
Nick Piggin 已提交
1479 1480
	if (error)
		return ((error == -ENOMEM) ? VM_FAULT_OOM : VM_FAULT_SIGBUS);
L
Linus Torvalds 已提交
1481

N
Nick Piggin 已提交
1482
	return ret | VM_FAULT_LOCKED;
L
Linus Torvalds 已提交
1483 1484 1485
}

#ifdef CONFIG_NUMA
A
Adrian Bunk 已提交
1486
static int shmem_set_policy(struct vm_area_struct *vma, struct mempolicy *new)
L
Linus Torvalds 已提交
1487
{
1488
	struct inode *i = vma->vm_file->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
1489 1490 1491
	return mpol_set_shared_policy(&SHMEM_I(i)->policy, vma, new);
}

A
Adrian Bunk 已提交
1492 1493
static struct mempolicy *shmem_get_policy(struct vm_area_struct *vma,
					  unsigned long addr)
L
Linus Torvalds 已提交
1494
{
1495
	struct inode *i = vma->vm_file->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
1496 1497 1498 1499 1500 1501 1502 1503 1504
	unsigned long idx;

	idx = ((addr - vma->vm_start) >> PAGE_SHIFT) + vma->vm_pgoff;
	return mpol_shared_policy_lookup(&SHMEM_I(i)->policy, idx);
}
#endif

int shmem_lock(struct file *file, int lock, struct user_struct *user)
{
1505
	struct inode *inode = file->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
1506 1507 1508 1509 1510 1511 1512 1513
	struct shmem_inode_info *info = SHMEM_I(inode);
	int retval = -ENOMEM;

	spin_lock(&info->lock);
	if (lock && !(info->flags & VM_LOCKED)) {
		if (!user_shm_lock(inode->i_size, user))
			goto out_nomem;
		info->flags |= VM_LOCKED;
1514
		mapping_set_unevictable(file->f_mapping);
L
Linus Torvalds 已提交
1515 1516 1517 1518
	}
	if (!lock && (info->flags & VM_LOCKED) && user) {
		user_shm_unlock(inode->i_size, user);
		info->flags &= ~VM_LOCKED;
1519 1520
		mapping_clear_unevictable(file->f_mapping);
		scan_mapping_unevictable_pages(file->f_mapping);
L
Linus Torvalds 已提交
1521 1522
	}
	retval = 0;
1523

L
Linus Torvalds 已提交
1524 1525 1526 1527 1528
out_nomem:
	spin_unlock(&info->lock);
	return retval;
}

1529
static int shmem_mmap(struct file *file, struct vm_area_struct *vma)
L
Linus Torvalds 已提交
1530 1531 1532
{
	file_accessed(file);
	vma->vm_ops = &shmem_vm_ops;
N
Nick Piggin 已提交
1533
	vma->vm_flags |= VM_CAN_NONLINEAR;
L
Linus Torvalds 已提交
1534 1535 1536
	return 0;
}

1537 1538
static struct inode *shmem_get_inode(struct super_block *sb, int mode,
					dev_t dev, unsigned long flags)
L
Linus Torvalds 已提交
1539 1540 1541 1542 1543
{
	struct inode *inode;
	struct shmem_inode_info *info;
	struct shmem_sb_info *sbinfo = SHMEM_SB(sb);

1544 1545
	if (shmem_reserve_inode(sb))
		return NULL;
L
Linus Torvalds 已提交
1546 1547 1548 1549

	inode = new_inode(sb);
	if (inode) {
		inode->i_mode = mode;
1550 1551
		inode->i_uid = current_fsuid();
		inode->i_gid = current_fsgid();
L
Linus Torvalds 已提交
1552 1553 1554
		inode->i_blocks = 0;
		inode->i_mapping->backing_dev_info = &shmem_backing_dev_info;
		inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
1555
		inode->i_generation = get_seconds();
L
Linus Torvalds 已提交
1556 1557 1558
		info = SHMEM_I(inode);
		memset(info, 0, (char *)inode - (char *)info);
		spin_lock_init(&info->lock);
1559
		info->flags = flags & VM_NORESERVE;
L
Linus Torvalds 已提交
1560 1561 1562 1563
		INIT_LIST_HEAD(&info->swaplist);

		switch (mode & S_IFMT) {
		default:
1564
			inode->i_op = &shmem_special_inode_operations;
L
Linus Torvalds 已提交
1565 1566 1567
			init_special_inode(inode, mode, dev);
			break;
		case S_IFREG:
1568
			inode->i_mapping->a_ops = &shmem_aops;
L
Linus Torvalds 已提交
1569 1570
			inode->i_op = &shmem_inode_operations;
			inode->i_fop = &shmem_file_operations;
1571 1572
			mpol_shared_policy_init(&info->policy,
						 shmem_get_sbmpol(sbinfo));
L
Linus Torvalds 已提交
1573 1574
			break;
		case S_IFDIR:
1575
			inc_nlink(inode);
L
Linus Torvalds 已提交
1576 1577 1578 1579 1580 1581 1582 1583 1584 1585
			/* Some things misbehave if size == 0 on a directory */
			inode->i_size = 2 * BOGO_DIRENT_SIZE;
			inode->i_op = &shmem_dir_inode_operations;
			inode->i_fop = &simple_dir_operations;
			break;
		case S_IFLNK:
			/*
			 * Must not load anything in the rbtree,
			 * mpol_free_shared_policy will not be called.
			 */
1586
			mpol_shared_policy_init(&info->policy, NULL);
L
Linus Torvalds 已提交
1587 1588
			break;
		}
1589 1590
	} else
		shmem_free_inode(sb);
L
Linus Torvalds 已提交
1591 1592 1593 1594
	return inode;
}

#ifdef CONFIG_TMPFS
1595 1596
static const struct inode_operations shmem_symlink_inode_operations;
static const struct inode_operations shmem_symlink_inline_operations;
L
Linus Torvalds 已提交
1597 1598

/*
N
Nick Piggin 已提交
1599
 * Normally tmpfs avoids the use of shmem_readpage and shmem_write_begin;
1600 1601
 * but providing them allows a tmpfs file to be used for splice, sendfile, and
 * below the loop driver, in the generic fashion that many filesystems support.
L
Linus Torvalds 已提交
1602
 */
1603 1604 1605 1606 1607 1608 1609 1610
static int shmem_readpage(struct file *file, struct page *page)
{
	struct inode *inode = page->mapping->host;
	int error = shmem_getpage(inode, page->index, &page, SGP_CACHE, NULL);
	unlock_page(page);
	return error;
}

L
Linus Torvalds 已提交
1611
static int
N
Nick Piggin 已提交
1612 1613 1614
shmem_write_begin(struct file *file, struct address_space *mapping,
			loff_t pos, unsigned len, unsigned flags,
			struct page **pagep, void **fsdata)
L
Linus Torvalds 已提交
1615
{
N
Nick Piggin 已提交
1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628
	struct inode *inode = mapping->host;
	pgoff_t index = pos >> PAGE_CACHE_SHIFT;
	*pagep = NULL;
	return shmem_getpage(inode, index, pagep, SGP_WRITE, NULL);
}

static int
shmem_write_end(struct file *file, struct address_space *mapping,
			loff_t pos, unsigned len, unsigned copied,
			struct page *page, void *fsdata)
{
	struct inode *inode = mapping->host;

H
Hugh Dickins 已提交
1629 1630 1631 1632
	if (pos + copied > inode->i_size)
		i_size_write(inode, pos + copied);

	unlock_page(page);
N
Nick Piggin 已提交
1633 1634 1635 1636
	set_page_dirty(page);
	page_cache_release(page);

	return copied;
L
Linus Torvalds 已提交
1637 1638 1639 1640
}

static void do_shmem_file_read(struct file *filp, loff_t *ppos, read_descriptor_t *desc, read_actor_t actor)
{
1641
	struct inode *inode = filp->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
1642 1643
	struct address_space *mapping = inode->i_mapping;
	unsigned long index, offset;
1644 1645 1646 1647 1648 1649 1650 1651 1652
	enum sgp_type sgp = SGP_READ;

	/*
	 * Might this read be for a stacking filesystem?  Then when reading
	 * holes of a sparse file, we actually need to allocate those pages,
	 * and even mark them dirty, so it cannot exceed the max_blocks limit.
	 */
	if (segment_eq(get_fs(), KERNEL_DS))
		sgp = SGP_DIRTY;
L
Linus Torvalds 已提交
1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670

	index = *ppos >> PAGE_CACHE_SHIFT;
	offset = *ppos & ~PAGE_CACHE_MASK;

	for (;;) {
		struct page *page = NULL;
		unsigned long end_index, nr, ret;
		loff_t i_size = i_size_read(inode);

		end_index = i_size >> PAGE_CACHE_SHIFT;
		if (index > end_index)
			break;
		if (index == end_index) {
			nr = i_size & ~PAGE_CACHE_MASK;
			if (nr <= offset)
				break;
		}

1671
		desc->error = shmem_getpage(inode, index, &page, sgp, NULL);
L
Linus Torvalds 已提交
1672 1673 1674 1675 1676
		if (desc->error) {
			if (desc->error == -EINVAL)
				desc->error = 0;
			break;
		}
H
Hugh Dickins 已提交
1677 1678
		if (page)
			unlock_page(page);
L
Linus Torvalds 已提交
1679 1680 1681

		/*
		 * We must evaluate after, since reads (unlike writes)
1682
		 * are called without i_mutex protection against truncate
L
Linus Torvalds 已提交
1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709
		 */
		nr = PAGE_CACHE_SIZE;
		i_size = i_size_read(inode);
		end_index = i_size >> PAGE_CACHE_SHIFT;
		if (index == end_index) {
			nr = i_size & ~PAGE_CACHE_MASK;
			if (nr <= offset) {
				if (page)
					page_cache_release(page);
				break;
			}
		}
		nr -= offset;

		if (page) {
			/*
			 * If users can be writing to this page using arbitrary
			 * virtual addresses, take care about potential aliasing
			 * before reading the page on the kernel side.
			 */
			if (mapping_writably_mapped(mapping))
				flush_dcache_page(page);
			/*
			 * Mark the page accessed if we read the beginning.
			 */
			if (!offset)
				mark_page_accessed(page);
N
Nick Piggin 已提交
1710
		} else {
L
Linus Torvalds 已提交
1711
			page = ZERO_PAGE(0);
N
Nick Piggin 已提交
1712 1713
			page_cache_get(page);
		}
L
Linus Torvalds 已提交
1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740

		/*
		 * Ok, we have the page, and it's up-to-date, so
		 * now we can copy it to user space...
		 *
		 * The actor routine returns how many bytes were actually used..
		 * NOTE! This may not be the same as how much of a user buffer
		 * we filled up (we may be padding etc), so we can only update
		 * "pos" here (the actor routine has to update the user buffer
		 * pointers and the remaining count).
		 */
		ret = actor(desc, page, offset, nr);
		offset += ret;
		index += offset >> PAGE_CACHE_SHIFT;
		offset &= ~PAGE_CACHE_MASK;

		page_cache_release(page);
		if (ret != nr || !desc->count)
			break;

		cond_resched();
	}

	*ppos = ((loff_t) index << PAGE_CACHE_SHIFT) + offset;
	file_accessed(filp);
}

H
Hugh Dickins 已提交
1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772
static ssize_t shmem_file_aio_read(struct kiocb *iocb,
		const struct iovec *iov, unsigned long nr_segs, loff_t pos)
{
	struct file *filp = iocb->ki_filp;
	ssize_t retval;
	unsigned long seg;
	size_t count;
	loff_t *ppos = &iocb->ki_pos;

	retval = generic_segment_checks(iov, &nr_segs, &count, VERIFY_WRITE);
	if (retval)
		return retval;

	for (seg = 0; seg < nr_segs; seg++) {
		read_descriptor_t desc;

		desc.written = 0;
		desc.arg.buf = iov[seg].iov_base;
		desc.count = iov[seg].iov_len;
		if (desc.count == 0)
			continue;
		desc.error = 0;
		do_shmem_file_read(filp, ppos, &desc, file_read_actor);
		retval += desc.written;
		if (desc.error) {
			retval = retval ?: desc.error;
			break;
		}
		if (desc.count > 0)
			break;
	}
	return retval;
L
Linus Torvalds 已提交
1773 1774
}

1775
static int shmem_statfs(struct dentry *dentry, struct kstatfs *buf)
L
Linus Torvalds 已提交
1776
{
1777
	struct shmem_sb_info *sbinfo = SHMEM_SB(dentry->d_sb);
L
Linus Torvalds 已提交
1778 1779 1780 1781

	buf->f_type = TMPFS_MAGIC;
	buf->f_bsize = PAGE_CACHE_SIZE;
	buf->f_namelen = NAME_MAX;
1782 1783
	spin_lock(&sbinfo->stat_lock);
	if (sbinfo->max_blocks) {
L
Linus Torvalds 已提交
1784 1785
		buf->f_blocks = sbinfo->max_blocks;
		buf->f_bavail = buf->f_bfree = sbinfo->free_blocks;
1786 1787
	}
	if (sbinfo->max_inodes) {
L
Linus Torvalds 已提交
1788 1789 1790 1791
		buf->f_files = sbinfo->max_inodes;
		buf->f_ffree = sbinfo->free_inodes;
	}
	/* else leave those fields 0 like simple_statfs */
1792
	spin_unlock(&sbinfo->stat_lock);
L
Linus Torvalds 已提交
1793 1794 1795 1796 1797 1798 1799 1800 1801
	return 0;
}

/*
 * File creation. Allocate an inode, and we're done..
 */
static int
shmem_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t dev)
{
1802
	struct inode *inode;
L
Linus Torvalds 已提交
1803 1804
	int error = -ENOSPC;

1805
	inode = shmem_get_inode(dir->i_sb, mode, dev, VM_NORESERVE);
L
Linus Torvalds 已提交
1806
	if (inode) {
1807 1808 1809 1810 1811 1812 1813
		error = security_inode_init_security(inode, dir, NULL, NULL,
						     NULL);
		if (error) {
			if (error != -EOPNOTSUPP) {
				iput(inode);
				return error;
			}
1814 1815 1816 1817 1818
		}
		error = shmem_acl_init(inode, dir);
		if (error) {
			iput(inode);
			return error;
1819
		}
L
Linus Torvalds 已提交
1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838
		if (dir->i_mode & S_ISGID) {
			inode->i_gid = dir->i_gid;
			if (S_ISDIR(mode))
				inode->i_mode |= S_ISGID;
		}
		dir->i_size += BOGO_DIRENT_SIZE;
		dir->i_ctime = dir->i_mtime = CURRENT_TIME;
		d_instantiate(dentry, inode);
		dget(dentry); /* Extra count - pin the dentry in core */
	}
	return error;
}

static int shmem_mkdir(struct inode *dir, struct dentry *dentry, int mode)
{
	int error;

	if ((error = shmem_mknod(dir, dentry, mode | S_IFDIR, 0)))
		return error;
1839
	inc_nlink(dir);
L
Linus Torvalds 已提交
1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854
	return 0;
}

static int shmem_create(struct inode *dir, struct dentry *dentry, int mode,
		struct nameidata *nd)
{
	return shmem_mknod(dir, dentry, mode | S_IFREG, 0);
}

/*
 * Link a file..
 */
static int shmem_link(struct dentry *old_dentry, struct inode *dir, struct dentry *dentry)
{
	struct inode *inode = old_dentry->d_inode;
1855
	int ret;
L
Linus Torvalds 已提交
1856 1857 1858 1859 1860 1861

	/*
	 * No ordinary (disk based) filesystem counts links as inodes;
	 * but each new link needs a new dentry, pinning lowmem, and
	 * tmpfs dentries cannot be pruned until they are unlinked.
	 */
1862 1863 1864
	ret = shmem_reserve_inode(inode->i_sb);
	if (ret)
		goto out;
L
Linus Torvalds 已提交
1865 1866 1867

	dir->i_size += BOGO_DIRENT_SIZE;
	inode->i_ctime = dir->i_ctime = dir->i_mtime = CURRENT_TIME;
1868
	inc_nlink(inode);
L
Linus Torvalds 已提交
1869 1870 1871
	atomic_inc(&inode->i_count);	/* New dentry reference */
	dget(dentry);		/* Extra pinning count for the created dentry */
	d_instantiate(dentry, inode);
1872 1873
out:
	return ret;
L
Linus Torvalds 已提交
1874 1875 1876 1877 1878 1879
}

static int shmem_unlink(struct inode *dir, struct dentry *dentry)
{
	struct inode *inode = dentry->d_inode;

1880 1881
	if (inode->i_nlink > 1 && !S_ISDIR(inode->i_mode))
		shmem_free_inode(inode->i_sb);
L
Linus Torvalds 已提交
1882 1883 1884

	dir->i_size -= BOGO_DIRENT_SIZE;
	inode->i_ctime = dir->i_ctime = dir->i_mtime = CURRENT_TIME;
1885
	drop_nlink(inode);
L
Linus Torvalds 已提交
1886 1887 1888 1889 1890 1891 1892 1893 1894
	dput(dentry);	/* Undo the count from "create" - this does all the work */
	return 0;
}

static int shmem_rmdir(struct inode *dir, struct dentry *dentry)
{
	if (!simple_empty(dentry))
		return -ENOTEMPTY;

1895 1896
	drop_nlink(dentry->d_inode);
	drop_nlink(dir);
L
Linus Torvalds 已提交
1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916
	return shmem_unlink(dir, dentry);
}

/*
 * The VFS layer already does all the dentry stuff for rename,
 * we just have to decrement the usage count for the target if
 * it exists so that the VFS layer correctly free's it when it
 * gets overwritten.
 */
static int shmem_rename(struct inode *old_dir, struct dentry *old_dentry, struct inode *new_dir, struct dentry *new_dentry)
{
	struct inode *inode = old_dentry->d_inode;
	int they_are_dirs = S_ISDIR(inode->i_mode);

	if (!simple_empty(new_dentry))
		return -ENOTEMPTY;

	if (new_dentry->d_inode) {
		(void) shmem_unlink(new_dir, new_dentry);
		if (they_are_dirs)
1917
			drop_nlink(old_dir);
L
Linus Torvalds 已提交
1918
	} else if (they_are_dirs) {
1919
		drop_nlink(old_dir);
1920
		inc_nlink(new_dir);
L
Linus Torvalds 已提交
1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943
	}

	old_dir->i_size -= BOGO_DIRENT_SIZE;
	new_dir->i_size += BOGO_DIRENT_SIZE;
	old_dir->i_ctime = old_dir->i_mtime =
	new_dir->i_ctime = new_dir->i_mtime =
	inode->i_ctime = CURRENT_TIME;
	return 0;
}

static int shmem_symlink(struct inode *dir, struct dentry *dentry, const char *symname)
{
	int error;
	int len;
	struct inode *inode;
	struct page *page = NULL;
	char *kaddr;
	struct shmem_inode_info *info;

	len = strlen(symname) + 1;
	if (len > PAGE_CACHE_SIZE)
		return -ENAMETOOLONG;

1944
	inode = shmem_get_inode(dir->i_sb, S_IFLNK|S_IRWXUGO, 0, VM_NORESERVE);
L
Linus Torvalds 已提交
1945 1946 1947
	if (!inode)
		return -ENOSPC;

1948 1949 1950 1951 1952 1953 1954 1955 1956 1957
	error = security_inode_init_security(inode, dir, NULL, NULL,
					     NULL);
	if (error) {
		if (error != -EOPNOTSUPP) {
			iput(inode);
			return error;
		}
		error = 0;
	}

L
Linus Torvalds 已提交
1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969
	info = SHMEM_I(inode);
	inode->i_size = len-1;
	if (len <= (char *)inode - (char *)info) {
		/* do it inline */
		memcpy(info, symname, len);
		inode->i_op = &shmem_symlink_inline_operations;
	} else {
		error = shmem_getpage(inode, 0, &page, SGP_WRITE, NULL);
		if (error) {
			iput(inode);
			return error;
		}
H
Hugh Dickins 已提交
1970
		unlock_page(page);
1971
		inode->i_mapping->a_ops = &shmem_aops;
L
Linus Torvalds 已提交
1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987
		inode->i_op = &shmem_symlink_inode_operations;
		kaddr = kmap_atomic(page, KM_USER0);
		memcpy(kaddr, symname, len);
		kunmap_atomic(kaddr, KM_USER0);
		set_page_dirty(page);
		page_cache_release(page);
	}
	if (dir->i_mode & S_ISGID)
		inode->i_gid = dir->i_gid;
	dir->i_size += BOGO_DIRENT_SIZE;
	dir->i_ctime = dir->i_mtime = CURRENT_TIME;
	d_instantiate(dentry, inode);
	dget(dentry);
	return 0;
}

1988
static void *shmem_follow_link_inline(struct dentry *dentry, struct nameidata *nd)
L
Linus Torvalds 已提交
1989 1990
{
	nd_set_link(nd, (char *)SHMEM_I(dentry->d_inode));
1991
	return NULL;
L
Linus Torvalds 已提交
1992 1993
}

1994
static void *shmem_follow_link(struct dentry *dentry, struct nameidata *nd)
L
Linus Torvalds 已提交
1995 1996 1997 1998
{
	struct page *page = NULL;
	int res = shmem_getpage(dentry->d_inode, 0, &page, SGP_READ, NULL);
	nd_set_link(nd, res ? ERR_PTR(res) : kmap(page));
H
Hugh Dickins 已提交
1999 2000
	if (page)
		unlock_page(page);
2001
	return page;
L
Linus Torvalds 已提交
2002 2003
}

2004
static void shmem_put_link(struct dentry *dentry, struct nameidata *nd, void *cookie)
L
Linus Torvalds 已提交
2005 2006
{
	if (!IS_ERR(nd_get_link(nd))) {
2007
		struct page *page = cookie;
L
Linus Torvalds 已提交
2008 2009 2010 2011 2012 2013
		kunmap(page);
		mark_page_accessed(page);
		page_cache_release(page);
	}
}

2014
static const struct inode_operations shmem_symlink_inline_operations = {
L
Linus Torvalds 已提交
2015 2016 2017 2018
	.readlink	= generic_readlink,
	.follow_link	= shmem_follow_link_inline,
};

2019
static const struct inode_operations shmem_symlink_inode_operations = {
L
Linus Torvalds 已提交
2020 2021 2022 2023 2024 2025
	.truncate	= shmem_truncate,
	.readlink	= generic_readlink,
	.follow_link	= shmem_follow_link,
	.put_link	= shmem_put_link,
};

2026
#ifdef CONFIG_TMPFS_POSIX_ACL
2027
/*
2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045
 * Superblocks without xattr inode operations will get security.* xattr
 * support from the VFS "for free". As soon as we have any other xattrs
 * like ACLs, we also need to implement the security.* handlers at
 * filesystem level, though.
 */

static size_t shmem_xattr_security_list(struct inode *inode, char *list,
					size_t list_len, const char *name,
					size_t name_len)
{
	return security_inode_listsecurity(inode, list, list_len);
}

static int shmem_xattr_security_get(struct inode *inode, const char *name,
				    void *buffer, size_t size)
{
	if (strcmp(name, "") == 0)
		return -EINVAL;
2046
	return xattr_getsecurity(inode, name, buffer, size);
2047 2048 2049 2050 2051 2052 2053 2054 2055 2056
}

static int shmem_xattr_security_set(struct inode *inode, const char *name,
				    const void *value, size_t size, int flags)
{
	if (strcmp(name, "") == 0)
		return -EINVAL;
	return security_inode_setsecurity(inode, name, value, size, flags);
}

2057
static struct xattr_handler shmem_xattr_security_handler = {
2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071
	.prefix = XATTR_SECURITY_PREFIX,
	.list   = shmem_xattr_security_list,
	.get    = shmem_xattr_security_get,
	.set    = shmem_xattr_security_set,
};

static struct xattr_handler *shmem_xattr_handlers[] = {
	&shmem_xattr_acl_access_handler,
	&shmem_xattr_acl_default_handler,
	&shmem_xattr_security_handler,
	NULL
};
#endif

2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084
static struct dentry *shmem_get_parent(struct dentry *child)
{
	return ERR_PTR(-ESTALE);
}

static int shmem_match(struct inode *ino, void *vfh)
{
	__u32 *fh = vfh;
	__u64 inum = fh[2];
	inum = (inum << 32) | fh[1];
	return ino->i_ino == inum && fh[0] == ino->i_generation;
}

C
Christoph Hellwig 已提交
2085 2086
static struct dentry *shmem_fh_to_dentry(struct super_block *sb,
		struct fid *fid, int fh_len, int fh_type)
2087 2088
{
	struct inode *inode;
C
Christoph Hellwig 已提交
2089 2090 2091 2092 2093 2094
	struct dentry *dentry = NULL;
	u64 inum = fid->raw[2];
	inum = (inum << 32) | fid->raw[1];

	if (fh_len < 3)
		return NULL;
2095

C
Christoph Hellwig 已提交
2096 2097
	inode = ilookup5(sb, (unsigned long)(inum + fid->raw[0]),
			shmem_match, fid->raw);
2098
	if (inode) {
C
Christoph Hellwig 已提交
2099
		dentry = d_find_alias(inode);
2100 2101 2102
		iput(inode);
	}

C
Christoph Hellwig 已提交
2103
	return dentry;
2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135
}

static int shmem_encode_fh(struct dentry *dentry, __u32 *fh, int *len,
				int connectable)
{
	struct inode *inode = dentry->d_inode;

	if (*len < 3)
		return 255;

	if (hlist_unhashed(&inode->i_hash)) {
		/* Unfortunately insert_inode_hash is not idempotent,
		 * so as we hash inodes here rather than at creation
		 * time, we need a lock to ensure we only try
		 * to do it once
		 */
		static DEFINE_SPINLOCK(lock);
		spin_lock(&lock);
		if (hlist_unhashed(&inode->i_hash))
			__insert_inode_hash(inode,
					    inode->i_ino + inode->i_generation);
		spin_unlock(&lock);
	}

	fh[0] = inode->i_generation;
	fh[1] = inode->i_ino;
	fh[2] = ((__u64)inode->i_ino) >> 32;

	*len = 3;
	return 1;
}

2136
static const struct export_operations shmem_export_ops = {
2137 2138
	.get_parent     = shmem_get_parent,
	.encode_fh      = shmem_encode_fh,
C
Christoph Hellwig 已提交
2139
	.fh_to_dentry	= shmem_fh_to_dentry,
2140 2141
};

2142 2143
static int shmem_parse_options(char *options, struct shmem_sb_info *sbinfo,
			       bool remount)
L
Linus Torvalds 已提交
2144 2145 2146
{
	char *this_char, *value, *rest;

2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163
	while (options != NULL) {
		this_char = options;
		for (;;) {
			/*
			 * NUL-terminate this option: unfortunately,
			 * mount options form a comma-separated list,
			 * but mpol's nodelist may also contain commas.
			 */
			options = strchr(options, ',');
			if (options == NULL)
				break;
			options++;
			if (!isdigit(*options)) {
				options[-1] = '\0';
				break;
			}
		}
L
Linus Torvalds 已提交
2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185
		if (!*this_char)
			continue;
		if ((value = strchr(this_char,'=')) != NULL) {
			*value++ = 0;
		} else {
			printk(KERN_ERR
			    "tmpfs: No value for mount option '%s'\n",
			    this_char);
			return 1;
		}

		if (!strcmp(this_char,"size")) {
			unsigned long long size;
			size = memparse(value,&rest);
			if (*rest == '%') {
				size <<= PAGE_SHIFT;
				size *= totalram_pages;
				do_div(size, 100);
				rest++;
			}
			if (*rest)
				goto bad_val;
2186 2187
			sbinfo->max_blocks =
				DIV_ROUND_UP(size, PAGE_CACHE_SIZE);
L
Linus Torvalds 已提交
2188
		} else if (!strcmp(this_char,"nr_blocks")) {
2189
			sbinfo->max_blocks = memparse(value, &rest);
L
Linus Torvalds 已提交
2190 2191 2192
			if (*rest)
				goto bad_val;
		} else if (!strcmp(this_char,"nr_inodes")) {
2193
			sbinfo->max_inodes = memparse(value, &rest);
L
Linus Torvalds 已提交
2194 2195 2196
			if (*rest)
				goto bad_val;
		} else if (!strcmp(this_char,"mode")) {
2197
			if (remount)
L
Linus Torvalds 已提交
2198
				continue;
2199
			sbinfo->mode = simple_strtoul(value, &rest, 8) & 07777;
L
Linus Torvalds 已提交
2200 2201 2202
			if (*rest)
				goto bad_val;
		} else if (!strcmp(this_char,"uid")) {
2203
			if (remount)
L
Linus Torvalds 已提交
2204
				continue;
2205
			sbinfo->uid = simple_strtoul(value, &rest, 0);
L
Linus Torvalds 已提交
2206 2207 2208
			if (*rest)
				goto bad_val;
		} else if (!strcmp(this_char,"gid")) {
2209
			if (remount)
L
Linus Torvalds 已提交
2210
				continue;
2211
			sbinfo->gid = simple_strtoul(value, &rest, 0);
L
Linus Torvalds 已提交
2212 2213
			if (*rest)
				goto bad_val;
2214
		} else if (!strcmp(this_char,"mpol")) {
2215
			if (mpol_parse_str(value, &sbinfo->mpol, 1))
2216
				goto bad_val;
L
Linus Torvalds 已提交
2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234
		} else {
			printk(KERN_ERR "tmpfs: Bad mount option %s\n",
			       this_char);
			return 1;
		}
	}
	return 0;

bad_val:
	printk(KERN_ERR "tmpfs: Bad value '%s' for mount option '%s'\n",
	       value, this_char);
	return 1;

}

static int shmem_remount_fs(struct super_block *sb, int *flags, char *data)
{
	struct shmem_sb_info *sbinfo = SHMEM_SB(sb);
2235
	struct shmem_sb_info config = *sbinfo;
2236 2237 2238 2239
	unsigned long blocks;
	unsigned long inodes;
	int error = -EINVAL;

2240
	if (shmem_parse_options(data, &config, true))
2241
		return error;
L
Linus Torvalds 已提交
2242

2243 2244 2245
	spin_lock(&sbinfo->stat_lock);
	blocks = sbinfo->max_blocks - sbinfo->free_blocks;
	inodes = sbinfo->max_inodes - sbinfo->free_inodes;
2246
	if (config.max_blocks < blocks)
2247
		goto out;
2248
	if (config.max_inodes < inodes)
2249 2250 2251 2252 2253 2254 2255
		goto out;
	/*
	 * Those tests also disallow limited->unlimited while any are in
	 * use, so i_blocks will always be zero when max_blocks is zero;
	 * but we must separately disallow unlimited->limited, because
	 * in that case we have no record of how much is already in use.
	 */
2256
	if (config.max_blocks && !sbinfo->max_blocks)
2257
		goto out;
2258
	if (config.max_inodes && !sbinfo->max_inodes)
2259 2260 2261
		goto out;

	error = 0;
2262 2263 2264 2265
	sbinfo->max_blocks  = config.max_blocks;
	sbinfo->free_blocks = config.max_blocks - blocks;
	sbinfo->max_inodes  = config.max_inodes;
	sbinfo->free_inodes = config.max_inodes - inodes;
2266 2267 2268

	mpol_put(sbinfo->mpol);
	sbinfo->mpol        = config.mpol;	/* transfers initial ref */
2269 2270 2271
out:
	spin_unlock(&sbinfo->stat_lock);
	return error;
L
Linus Torvalds 已提交
2272
}
2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288

static int shmem_show_options(struct seq_file *seq, struct vfsmount *vfs)
{
	struct shmem_sb_info *sbinfo = SHMEM_SB(vfs->mnt_sb);

	if (sbinfo->max_blocks != shmem_default_max_blocks())
		seq_printf(seq, ",size=%luk",
			sbinfo->max_blocks << (PAGE_CACHE_SHIFT - 10));
	if (sbinfo->max_inodes != shmem_default_max_inodes())
		seq_printf(seq, ",nr_inodes=%lu", sbinfo->max_inodes);
	if (sbinfo->mode != (S_IRWXUGO | S_ISVTX))
		seq_printf(seq, ",mode=%03o", sbinfo->mode);
	if (sbinfo->uid != 0)
		seq_printf(seq, ",uid=%u", sbinfo->uid);
	if (sbinfo->gid != 0)
		seq_printf(seq, ",gid=%u", sbinfo->gid);
2289
	shmem_show_mpol(seq, sbinfo->mpol);
2290 2291 2292
	return 0;
}
#endif /* CONFIG_TMPFS */
L
Linus Torvalds 已提交
2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304

static void shmem_put_super(struct super_block *sb)
{
	kfree(sb->s_fs_info);
	sb->s_fs_info = NULL;
}

static int shmem_fill_super(struct super_block *sb,
			    void *data, int silent)
{
	struct inode *inode;
	struct dentry *root;
2305
	struct shmem_sb_info *sbinfo;
2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316
	int err = -ENOMEM;

	/* Round up to L1_CACHE_BYTES to resist false sharing */
	sbinfo = kmalloc(max((int)sizeof(struct shmem_sb_info),
				L1_CACHE_BYTES), GFP_KERNEL);
	if (!sbinfo)
		return -ENOMEM;

	sbinfo->max_blocks = 0;
	sbinfo->max_inodes = 0;
	sbinfo->mode = S_IRWXUGO | S_ISVTX;
2317 2318
	sbinfo->uid = current_fsuid();
	sbinfo->gid = current_fsgid();
2319
	sbinfo->mpol = NULL;
2320
	sb->s_fs_info = sbinfo;
L
Linus Torvalds 已提交
2321

2322
#ifdef CONFIG_TMPFS
L
Linus Torvalds 已提交
2323 2324 2325 2326 2327 2328
	/*
	 * Per default we only allow half of the physical ram per
	 * tmpfs instance, limiting inodes to one per page of lowmem;
	 * but the internal instance is left unlimited.
	 */
	if (!(sb->s_flags & MS_NOUSER)) {
2329 2330 2331 2332 2333 2334
		sbinfo->max_blocks = shmem_default_max_blocks();
		sbinfo->max_inodes = shmem_default_max_inodes();
		if (shmem_parse_options(data, sbinfo, false)) {
			err = -EINVAL;
			goto failed;
		}
L
Linus Torvalds 已提交
2335
	}
2336
	sb->s_export_op = &shmem_export_ops;
L
Linus Torvalds 已提交
2337 2338 2339 2340
#else
	sb->s_flags |= MS_NOUSER;
#endif

2341
	spin_lock_init(&sbinfo->stat_lock);
2342 2343
	sbinfo->free_blocks = sbinfo->max_blocks;
	sbinfo->free_inodes = sbinfo->max_inodes;
2344

L
Linus Torvalds 已提交
2345 2346 2347 2348 2349
	sb->s_maxbytes = SHMEM_MAX_BYTES;
	sb->s_blocksize = PAGE_CACHE_SIZE;
	sb->s_blocksize_bits = PAGE_CACHE_SHIFT;
	sb->s_magic = TMPFS_MAGIC;
	sb->s_op = &shmem_ops;
2350
	sb->s_time_gran = 1;
2351 2352 2353 2354
#ifdef CONFIG_TMPFS_POSIX_ACL
	sb->s_xattr = shmem_xattr_handlers;
	sb->s_flags |= MS_POSIXACL;
#endif
2355

2356
	inode = shmem_get_inode(sb, S_IFDIR | sbinfo->mode, 0, VM_NORESERVE);
L
Linus Torvalds 已提交
2357 2358
	if (!inode)
		goto failed;
2359 2360
	inode->i_uid = sbinfo->uid;
	inode->i_gid = sbinfo->gid;
L
Linus Torvalds 已提交
2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373
	root = d_alloc_root(inode);
	if (!root)
		goto failed_iput;
	sb->s_root = root;
	return 0;

failed_iput:
	iput(inode);
failed:
	shmem_put_super(sb);
	return err;
}

2374
static struct kmem_cache *shmem_inode_cachep;
L
Linus Torvalds 已提交
2375 2376 2377 2378

static struct inode *shmem_alloc_inode(struct super_block *sb)
{
	struct shmem_inode_info *p;
2379
	p = (struct shmem_inode_info *)kmem_cache_alloc(shmem_inode_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390
	if (!p)
		return NULL;
	return &p->vfs_inode;
}

static void shmem_destroy_inode(struct inode *inode)
{
	if ((inode->i_mode & S_IFMT) == S_IFREG) {
		/* only struct inode is valid if it's an inline symlink */
		mpol_free_shared_policy(&SHMEM_I(inode)->policy);
	}
2391
	shmem_acl_destroy_inode(inode);
L
Linus Torvalds 已提交
2392 2393 2394
	kmem_cache_free(shmem_inode_cachep, SHMEM_I(inode));
}

2395
static void init_once(void *foo)
L
Linus Torvalds 已提交
2396 2397 2398
{
	struct shmem_inode_info *p = (struct shmem_inode_info *) foo;

C
Christoph Lameter 已提交
2399
	inode_init_once(&p->vfs_inode);
2400
#ifdef CONFIG_TMPFS_POSIX_ACL
C
Christoph Lameter 已提交
2401 2402
	p->i_acl = NULL;
	p->i_default_acl = NULL;
2403
#endif
L
Linus Torvalds 已提交
2404 2405 2406 2407 2408 2409
}

static int init_inodecache(void)
{
	shmem_inode_cachep = kmem_cache_create("shmem_inode_cache",
				sizeof(struct shmem_inode_info),
2410
				0, SLAB_PANIC, init_once);
L
Linus Torvalds 已提交
2411 2412 2413 2414 2415
	return 0;
}

static void destroy_inodecache(void)
{
2416
	kmem_cache_destroy(shmem_inode_cachep);
L
Linus Torvalds 已提交
2417 2418
}

2419
static const struct address_space_operations shmem_aops = {
L
Linus Torvalds 已提交
2420
	.writepage	= shmem_writepage,
2421
	.set_page_dirty	= __set_page_dirty_no_writeback,
L
Linus Torvalds 已提交
2422
#ifdef CONFIG_TMPFS
2423
	.readpage	= shmem_readpage,
N
Nick Piggin 已提交
2424 2425
	.write_begin	= shmem_write_begin,
	.write_end	= shmem_write_end,
L
Linus Torvalds 已提交
2426
#endif
2427
	.migratepage	= migrate_page,
L
Linus Torvalds 已提交
2428 2429
};

2430
static const struct file_operations shmem_file_operations = {
L
Linus Torvalds 已提交
2431 2432 2433
	.mmap		= shmem_mmap,
#ifdef CONFIG_TMPFS
	.llseek		= generic_file_llseek,
H
Hugh Dickins 已提交
2434
	.read		= do_sync_read,
H
Hugh Dickins 已提交
2435
	.write		= do_sync_write,
H
Hugh Dickins 已提交
2436
	.aio_read	= shmem_file_aio_read,
H
Hugh Dickins 已提交
2437
	.aio_write	= generic_file_aio_write,
L
Linus Torvalds 已提交
2438
	.fsync		= simple_sync_file,
2439 2440
	.splice_read	= generic_file_splice_read,
	.splice_write	= generic_file_splice_write,
L
Linus Torvalds 已提交
2441 2442 2443
#endif
};

2444
static const struct inode_operations shmem_inode_operations = {
L
Linus Torvalds 已提交
2445 2446
	.truncate	= shmem_truncate,
	.setattr	= shmem_notify_change,
2447
	.truncate_range	= shmem_truncate_range,
2448 2449 2450 2451 2452 2453 2454 2455
#ifdef CONFIG_TMPFS_POSIX_ACL
	.setxattr	= generic_setxattr,
	.getxattr	= generic_getxattr,
	.listxattr	= generic_listxattr,
	.removexattr	= generic_removexattr,
	.permission	= shmem_permission,
#endif

L
Linus Torvalds 已提交
2456 2457
};

2458
static const struct inode_operations shmem_dir_inode_operations = {
L
Linus Torvalds 已提交
2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469
#ifdef CONFIG_TMPFS
	.create		= shmem_create,
	.lookup		= simple_lookup,
	.link		= shmem_link,
	.unlink		= shmem_unlink,
	.symlink	= shmem_symlink,
	.mkdir		= shmem_mkdir,
	.rmdir		= shmem_rmdir,
	.mknod		= shmem_mknod,
	.rename		= shmem_rename,
#endif
2470 2471 2472 2473 2474 2475 2476 2477 2478 2479
#ifdef CONFIG_TMPFS_POSIX_ACL
	.setattr	= shmem_notify_change,
	.setxattr	= generic_setxattr,
	.getxattr	= generic_getxattr,
	.listxattr	= generic_listxattr,
	.removexattr	= generic_removexattr,
	.permission	= shmem_permission,
#endif
};

2480
static const struct inode_operations shmem_special_inode_operations = {
2481 2482 2483 2484 2485 2486 2487 2488
#ifdef CONFIG_TMPFS_POSIX_ACL
	.setattr	= shmem_notify_change,
	.setxattr	= generic_setxattr,
	.getxattr	= generic_getxattr,
	.listxattr	= generic_listxattr,
	.removexattr	= generic_removexattr,
	.permission	= shmem_permission,
#endif
L
Linus Torvalds 已提交
2489 2490
};

2491
static const struct super_operations shmem_ops = {
L
Linus Torvalds 已提交
2492 2493 2494 2495 2496
	.alloc_inode	= shmem_alloc_inode,
	.destroy_inode	= shmem_destroy_inode,
#ifdef CONFIG_TMPFS
	.statfs		= shmem_statfs,
	.remount_fs	= shmem_remount_fs,
2497
	.show_options	= shmem_show_options,
L
Linus Torvalds 已提交
2498 2499 2500 2501 2502 2503 2504
#endif
	.delete_inode	= shmem_delete_inode,
	.drop_inode	= generic_delete_inode,
	.put_super	= shmem_put_super,
};

static struct vm_operations_struct shmem_vm_ops = {
2505
	.fault		= shmem_fault,
L
Linus Torvalds 已提交
2506 2507 2508 2509 2510 2511 2512
#ifdef CONFIG_NUMA
	.set_policy     = shmem_set_policy,
	.get_policy     = shmem_get_policy,
#endif
};


2513 2514
static int shmem_get_sb(struct file_system_type *fs_type,
	int flags, const char *dev_name, void *data, struct vfsmount *mnt)
L
Linus Torvalds 已提交
2515
{
2516
	return get_sb_nodev(fs_type, flags, data, shmem_fill_super, mnt);
L
Linus Torvalds 已提交
2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529
}

static struct file_system_type tmpfs_fs_type = {
	.owner		= THIS_MODULE,
	.name		= "tmpfs",
	.get_sb		= shmem_get_sb,
	.kill_sb	= kill_litter_super,
};

static int __init init_tmpfs(void)
{
	int error;

P
Peter Zijlstra 已提交
2530 2531 2532 2533
	error = bdi_init(&shmem_backing_dev_info);
	if (error)
		goto out4;

L
Linus Torvalds 已提交
2534 2535 2536 2537 2538 2539 2540 2541 2542
	error = init_inodecache();
	if (error)
		goto out3;

	error = register_filesystem(&tmpfs_fs_type);
	if (error) {
		printk(KERN_ERR "Could not register tmpfs\n");
		goto out2;
	}
2543

2544
	shm_mnt = vfs_kern_mount(&tmpfs_fs_type, MS_NOUSER,
L
Linus Torvalds 已提交
2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557
				tmpfs_fs_type.name, NULL);
	if (IS_ERR(shm_mnt)) {
		error = PTR_ERR(shm_mnt);
		printk(KERN_ERR "Could not kern_mount tmpfs\n");
		goto out1;
	}
	return 0;

out1:
	unregister_filesystem(&tmpfs_fs_type);
out2:
	destroy_inodecache();
out3:
P
Peter Zijlstra 已提交
2558 2559
	bdi_destroy(&shmem_backing_dev_info);
out4:
L
Linus Torvalds 已提交
2560 2561 2562
	shm_mnt = ERR_PTR(error);
	return error;
}
2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596 2597

#else /* !CONFIG_SHMEM */

/*
 * tiny-shmem: simple shmemfs and tmpfs using ramfs code
 *
 * This is intended for small system where the benefits of the full
 * shmem code (swap-backed and resource-limited) are outweighed by
 * their complexity. On systems without swap this code should be
 * effectively equivalent, but much lighter weight.
 */

#include <linux/ramfs.h>

static struct file_system_type tmpfs_fs_type = {
	.name		= "tmpfs",
	.get_sb		= ramfs_get_sb,
	.kill_sb	= kill_litter_super,
};

static int __init init_tmpfs(void)
{
	BUG_ON(register_filesystem(&tmpfs_fs_type) != 0);

	shm_mnt = kern_mount(&tmpfs_fs_type);
	BUG_ON(IS_ERR(shm_mnt));

	return 0;
}

int shmem_unuse(swp_entry_t entry, struct page *page)
{
	return 0;
}

2598 2599 2600 2601 2602
#define shmem_vm_ops				generic_file_vm_ops
#define shmem_file_operations			ramfs_file_operations
#define shmem_get_inode(sb, mode, dev, flags)	ramfs_get_inode(sb, mode, dev)
#define shmem_acct_size(flags, size)		0
#define shmem_unacct_size(flags, size)		do {} while (0)
H
Hugh Dickins 已提交
2603
#define SHMEM_MAX_BYTES				MAX_LFS_FILESIZE
2604 2605 2606 2607

#endif /* CONFIG_SHMEM */

/* common code */
L
Linus Torvalds 已提交
2608

2609
/**
L
Linus Torvalds 已提交
2610 2611 2612
 * shmem_file_setup - get an unlinked file living in tmpfs
 * @name: name for dentry (to be seen in /proc/<pid>/maps
 * @size: size to be set for the file
2613
 * @flags: VM_NORESERVE suppresses pre-accounting of the entire object size
L
Linus Torvalds 已提交
2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646
 */
struct file *shmem_file_setup(char *name, loff_t size, unsigned long flags)
{
	int error;
	struct file *file;
	struct inode *inode;
	struct dentry *dentry, *root;
	struct qstr this;

	if (IS_ERR(shm_mnt))
		return (void *)shm_mnt;

	if (size < 0 || size > SHMEM_MAX_BYTES)
		return ERR_PTR(-EINVAL);

	if (shmem_acct_size(flags, size))
		return ERR_PTR(-ENOMEM);

	error = -ENOMEM;
	this.name = name;
	this.len = strlen(name);
	this.hash = 0; /* will go */
	root = shm_mnt->mnt_root;
	dentry = d_alloc(root, &this);
	if (!dentry)
		goto put_memory;

	error = -ENFILE;
	file = get_empty_filp();
	if (!file)
		goto put_dentry;

	error = -ENOSPC;
2647
	inode = shmem_get_inode(root->d_sb, S_IFREG | S_IRWXUGO, 0, flags);
L
Linus Torvalds 已提交
2648 2649 2650 2651 2652 2653
	if (!inode)
		goto close_file;

	d_instantiate(dentry, inode);
	inode->i_size = size;
	inode->i_nlink = 0;	/* It is unlinked */
2654
	init_file(file, shm_mnt, dentry, FMODE_WRITE | FMODE_READ,
2655 2656 2657 2658 2659 2660 2661
		  &shmem_file_operations);

#ifndef CONFIG_MMU
	error = ramfs_nommu_expand_for_mapping(inode, size);
	if (error)
		goto close_file;
#endif
L
Linus Torvalds 已提交
2662 2663 2664 2665 2666 2667 2668 2669 2670 2671
	return file;

close_file:
	put_filp(file);
put_dentry:
	dput(dentry);
put_memory:
	shmem_unacct_size(flags, size);
	return ERR_PTR(error);
}
2672
EXPORT_SYMBOL_GPL(shmem_file_setup);
L
Linus Torvalds 已提交
2673

2674
/**
L
Linus Torvalds 已提交
2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686
 * shmem_zero_setup - setup a shared anonymous mapping
 * @vma: the vma to be mmapped is prepared by do_mmap_pgoff
 */
int shmem_zero_setup(struct vm_area_struct *vma)
{
	struct file *file;
	loff_t size = vma->vm_end - vma->vm_start;

	file = shmem_file_setup("dev/zero", size, vma->vm_flags);
	if (IS_ERR(file))
		return PTR_ERR(file);

M
Mimi Zohar 已提交
2687
	ima_shm_check(file);
L
Linus Torvalds 已提交
2688 2689 2690 2691 2692 2693
	if (vma->vm_file)
		fput(vma->vm_file);
	vma->vm_file = file;
	vma->vm_ops = &shmem_vm_ops;
	return 0;
}
2694 2695

module_init(init_tmpfs)