shmem.c 70.8 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
#include <linux/file.h>
#include <linux/mm.h>
#include <linux/module.h>
31
#include <linux/percpu_counter.h>
32 33 34 35 36
#include <linux/swap.h>

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/posix_acl.h>
46
#include <linux/generic_acl.h>
L
Linus Torvalds 已提交
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/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;
222
static const struct vm_operations_struct shmem_vm_ops;
L
Linus Torvalds 已提交
223

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) {
237 238
		percpu_counter_add(&sbinfo->used_blocks, -pages);
		spin_lock(&inode->i_lock);
L
Linus Torvalds 已提交
239
		inode->i_blocks -= pages*BLOCKS_PER_PAGE;
240
		spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
241 242 243
	}
}

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
 */
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));
		/*
420
		 * Test used_blocks against 1 less max_blocks, since we have 1 data
L
Linus Torvalds 已提交
421 422 423
		 * page (and perhaps indirect index pages) yet to allocate:
		 * a waste to allocate index if we cannot allocate data.
		 */
424
		if (sbinfo->max_blocks) {
425
			if (percpu_counter_compare(&sbinfo->used_blocks, (sbinfo->max_blocks - 1)) > 0)
L
Linus Torvalds 已提交
426
				return ERR_PTR(-ENOSPC);
427 428
			percpu_counter_inc(&sbinfo->used_blocks);
			spin_lock(&inode->i_lock);
L
Linus Torvalds 已提交
429
			inode->i_blocks += BLOCKS_PER_PAGE;
430
			spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
431 432 433
		}

		spin_unlock(&info->lock);
434
		page = shmem_dir_alloc(mapping_gfp_mask(inode->i_mapping));
L
Linus Torvalds 已提交
435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458
		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;
}

459
/**
L
Linus Torvalds 已提交
460
 * shmem_free_swp - free some swap entries in a directory
461 462 463
 * @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
464
 */
465 466
static int shmem_free_swp(swp_entry_t *dir, swp_entry_t *edir,
						spinlock_t *punch_lock)
L
Linus Torvalds 已提交
467
{
468
	spinlock_t *punch_unlock = NULL;
L
Linus Torvalds 已提交
469 470 471 472 473
	swp_entry_t *ptr;
	int freed = 0;

	for (ptr = dir; ptr < edir; ptr++) {
		if (ptr->val) {
474 475 476 477 478 479 480
			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
481 482 483 484 485
			free_swap_and_cache(*ptr);
			*ptr = (swp_entry_t){0};
			freed++;
		}
	}
486 487
	if (punch_unlock)
		spin_unlock(punch_unlock);
L
Linus Torvalds 已提交
488 489 490
	return freed;
}

491 492
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
493 494 495 496 497 498 499 500 501
{
	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;
502 503
		freed += shmem_free_swp(ptr+offset, ptr+offset+size,
							punch_lock);
L
Linus Torvalds 已提交
504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534
		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);
}

535
static void shmem_truncate_range(struct inode *inode, loff_t start, loff_t end)
L
Linus Torvalds 已提交
536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552
{
	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;
553
	int punch_hole;
554 555
	spinlock_t *needs_lock;
	spinlock_t *punch_lock;
556
	unsigned long upper_limit;
L
Linus Torvalds 已提交
557 558

	inode->i_ctime = inode->i_mtime = CURRENT_TIME;
559
	idx = (start + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
L
Linus Torvalds 已提交
560 561 562 563 564
	if (idx >= info->next_index)
		return;

	spin_lock(&info->lock);
	info->flags |= SHMEM_TRUNCATE;
565 566
	if (likely(end == (loff_t) -1)) {
		limit = info->next_index;
567
		upper_limit = SHMEM_MAX_INDEX;
568
		info->next_index = idx;
569
		needs_lock = NULL;
570
		punch_hole = 0;
571
	} else {
572 573 574 575 576 577 578 579
		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;
		}
580
		needs_lock = &info->lock;
581 582 583
		punch_hole = 1;
	}

L
Linus Torvalds 已提交
584
	topdir = info->i_indirect;
585
	if (topdir && idx <= SHMEM_NR_DIRECT && !punch_hole) {
L
Linus Torvalds 已提交
586 587 588 589 590 591 592 593 594 595 596
		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;
597
		nr_swaps_freed = shmem_free_swp(ptr+idx, ptr+size, needs_lock);
L
Linus Torvalds 已提交
598
	}
599 600 601 602 603

	/*
	 * If there are no indirect blocks or we are punching a hole
	 * below indirect blocks, nothing to be done.
	 */
604
	if (!topdir || limit <= SHMEM_NR_DIRECT)
L
Linus Torvalds 已提交
605 606
		goto done2;

607 608 609 610 611 612 613 614 615 616 617 618 619
	/*
	 * 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.
	 */

620
	upper_limit -= SHMEM_NR_DIRECT;
L
Linus Torvalds 已提交
621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639
	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;
640
			if (!diroff && !offset && upper_limit >= stage) {
641 642 643 644 645 646 647
				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
648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672
				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;
673 674
			if (punch_hole)
				needs_lock = &info->lock;
675
			if (upper_limit >= stage) {
676 677 678 679 680 681 682
				if (needs_lock) {
					spin_lock(needs_lock);
					*dir = NULL;
					spin_unlock(needs_lock);
					needs_lock = NULL;
				} else
					*dir = NULL;
683 684 685
				nr_pages_to_free++;
				list_add(&middir->lru, &pages_to_free);
			}
L
Linus Torvalds 已提交
686 687 688 689 690
			shmem_dir_unmap(dir);
			cond_resched();
			dir = shmem_dir_map(middir);
			diroff = 0;
		}
691
		punch_lock = needs_lock;
L
Linus Torvalds 已提交
692
		subdir = dir[diroff];
693 694 695 696 697 698 699 700 701 702 703 704
		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
705 706 707 708
			size = limit - idx;
			if (size > ENTRIES_PER_PAGE)
				size = ENTRIES_PER_PAGE;
			freed = shmem_map_and_free_swp(subdir,
709
					offset, size, &dir, punch_lock);
L
Linus Torvalds 已提交
710 711 712
			if (!dir)
				dir = shmem_dir_map(middir);
			nr_swaps_freed += freed;
713
			if (offset || punch_lock) {
L
Linus Torvalds 已提交
714
				spin_lock(&info->lock);
715 716
				set_page_private(subdir,
					page_private(subdir) - freed);
L
Linus Torvalds 已提交
717
				spin_unlock(&info->lock);
718 719
			} else
				BUG_ON(page_private(subdir) != freed);
L
Linus Torvalds 已提交
720
		}
721
		offset = 0;
L
Linus Torvalds 已提交
722 723 724 725 726 727 728
	}
done1:
	shmem_dir_unmap(dir);
done2:
	if (inode->i_mapping->nrpages && (info->flags & SHMEM_PAGEIN)) {
		/*
		 * Call truncate_inode_pages again: racing shmem_unuse_inode
729 730 731 732 733
		 * may have swizzled a page in from swap since
		 * truncate_pagecache 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.
734 735 736 737 738
		 *
		 * 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
739
		 */
740
		truncate_inode_pages_range(inode->i_mapping, start, end);
741 742 743
		if (punch_hole)
			unmap_mapping_range(inode->i_mapping, start,
							end - start, 1);
L
Linus Torvalds 已提交
744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765
	}

	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);
	}
}

static int shmem_notify_change(struct dentry *dentry, struct iattr *attr)
{
	struct inode *inode = dentry->d_inode;
766
	loff_t newsize = attr->ia_size;
L
Linus Torvalds 已提交
767 768
	int error;

769 770
	if (S_ISREG(inode->i_mode) && (attr->ia_valid & ATTR_SIZE)
					&& newsize != inode->i_size) {
771 772 773
		struct page *page = NULL;

		if (newsize < inode->i_size) {
L
Linus Torvalds 已提交
774 775 776 777 778 779 780
			/*
			 * 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.
			 */
781
			if (newsize & (PAGE_CACHE_SIZE-1)) {
L
Linus Torvalds 已提交
782
				(void) shmem_getpage(inode,
783
					newsize >> PAGE_CACHE_SHIFT,
L
Linus Torvalds 已提交
784
						&page, SGP_READ, NULL);
H
Hugh Dickins 已提交
785 786
				if (page)
					unlock_page(page);
L
Linus Torvalds 已提交
787 788 789 790 791 792 793 794
			}
			/*
			 * 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.
			 */
795
			if (newsize) {
L
Linus Torvalds 已提交
796 797 798 799 800 801
				struct shmem_inode_info *info = SHMEM_I(inode);
				spin_lock(&info->lock);
				info->flags &= ~SHMEM_PAGEIN;
				spin_unlock(&info->lock);
			}
		}
802 803 804 805 806 807 808

		error = simple_setsize(inode, newsize);
		if (page)
			page_cache_release(page);
		if (error)
			return error;
		shmem_truncate_range(inode, newsize, (loff_t)-1);
L
Linus Torvalds 已提交
809 810 811 812
	}

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

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

825
	if (inode->i_mapping->a_ops == &shmem_aops) {
826
		truncate_inode_pages(inode->i_mapping, 0);
L
Linus Torvalds 已提交
827 828
		shmem_unacct_size(info->flags, inode->i_size);
		inode->i_size = 0;
829
		shmem_truncate_range(inode, 0, (loff_t)-1);
L
Linus Torvalds 已提交
830
		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);
H
Hugh Dickins 已提交
1021 1022 1023 1024 1025 1026 1027 1028
	/*
	 * Can some race bring us here?  We've been holding page lock,
	 * so I think not; but would rather try again later than BUG()
	 */
	unlock_page(page);
	page_cache_release(page);
out:
	return (found < 0) ? found : 0;
L
Linus Torvalds 已提交
1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048
}

/*
 * 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;
1049
	if (!total_swap_pages)
L
Linus Torvalds 已提交
1050 1051
		goto redirty;

1052 1053 1054 1055 1056
	/*
	 * 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,
1057 1058 1059
	 * and not for the writeback threads or sync.  However, in those cases,
	 * we do still want to check if there's a redundant swappage to be
	 * discarded.
1060 1061 1062 1063 1064 1065
	 */
	if (wbc->for_reclaim)
		swap = get_swap_page();
	else
		swap.val = 0;

L
Linus Torvalds 已提交
1066 1067 1068 1069 1070 1071
	spin_lock(&info->lock);
	if (index >= info->next_index) {
		BUG_ON(!(info->flags & SHMEM_TRUNCATE));
		goto unlock;
	}
	entry = shmem_swp_entry(info, index, NULL);
1072 1073 1074 1075 1076 1077 1078 1079 1080
	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
1081

1082
	if (swap.val && add_to_swap_cache(page, swap, GFP_ATOMIC) == 0) {
1083
		remove_from_page_cache(page);
L
Linus Torvalds 已提交
1084 1085
		shmem_swp_set(info, entry, swap.val);
		shmem_swp_unmap(entry);
H
Hugh Dickins 已提交
1086 1087 1088 1089
		if (list_empty(&info->swaplist))
			inode = igrab(inode);
		else
			inode = NULL;
L
Linus Torvalds 已提交
1090
		spin_unlock(&info->lock);
H
Hugh Dickins 已提交
1091
		swap_shmem_alloc(swap);
1092
		BUG_ON(page_mapped(page));
1093
		page_cache_release(page);	/* pagecache ref */
H
Hugh Dickins 已提交
1094
		swap_writepage(page, wbc);
H
Hugh Dickins 已提交
1095 1096 1097 1098 1099 1100 1101
		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
1102 1103 1104 1105 1106 1107
		return 0;
	}

	shmem_swp_unmap(entry);
unlock:
	spin_unlock(&info->lock);
1108 1109 1110 1111
	/*
	 * add_to_swap_cache() doesn't return -EEXIST, so we can safely
	 * clear SWAP_HAS_CACHE flag.
	 */
1112
	swapcache_free(swap, NULL);
L
Linus Torvalds 已提交
1113 1114
redirty:
	set_page_dirty(page);
1115 1116 1117 1118
	if (wbc->for_reclaim)
		return AOP_WRITEPAGE_ACTIVATE;	/* Return with page locked */
	unlock_page(page);
	return 0;
L
Linus Torvalds 已提交
1119 1120 1121
}

#ifdef CONFIG_NUMA
1122
#ifdef CONFIG_TMPFS
1123
static void shmem_show_mpol(struct seq_file *seq, struct mempolicy *mpol)
1124
{
1125
	char buffer[64];
1126

1127
	if (!mpol || mpol->mode == MPOL_DEFAULT)
1128
		return;		/* show nothing */
1129

1130
	mpol_to_str(buffer, sizeof(buffer), mpol, 1);
1131 1132

	seq_printf(seq, ",mpol=%s", buffer);
1133
}
1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145

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;
}
1146 1147
#endif /* CONFIG_TMPFS */

1148 1149
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
1150
{
1151
	struct mempolicy mpol, *spol;
L
Linus Torvalds 已提交
1152
	struct vm_area_struct pvma;
1153
	struct page *page;
L
Linus Torvalds 已提交
1154

1155 1156 1157
	spol = mpol_cond_copy(&mpol,
				mpol_shared_policy_lookup(&info->policy, idx));

L
Linus Torvalds 已提交
1158
	/* Create a pseudo vma that just contains the policy */
1159
	pvma.vm_start = 0;
L
Linus Torvalds 已提交
1160
	pvma.vm_pgoff = idx;
1161
	pvma.vm_ops = NULL;
1162
	pvma.vm_policy = spol;
1163
	page = swapin_readahead(entry, gfp, &pvma, 0);
L
Linus Torvalds 已提交
1164 1165 1166
	return page;
}

1167 1168
static struct page *shmem_alloc_page(gfp_t gfp,
			struct shmem_inode_info *info, unsigned long idx)
L
Linus Torvalds 已提交
1169 1170 1171
{
	struct vm_area_struct pvma;

1172 1173
	/* Create a pseudo vma that just contains the policy */
	pvma.vm_start = 0;
L
Linus Torvalds 已提交
1174
	pvma.vm_pgoff = idx;
1175 1176
	pvma.vm_ops = NULL;
	pvma.vm_policy = mpol_shared_policy_lookup(&info->policy, idx);
1177 1178 1179 1180 1181

	/*
	 * alloc_page_vma() will drop the shared policy reference
	 */
	return alloc_page_vma(gfp, &pvma, 0);
L
Linus Torvalds 已提交
1182
}
1183 1184
#else /* !CONFIG_NUMA */
#ifdef CONFIG_TMPFS
1185
static inline void shmem_show_mpol(struct seq_file *seq, struct mempolicy *p)
1186 1187 1188 1189
{
}
#endif /* CONFIG_TMPFS */

1190 1191
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
1192
{
1193
	return swapin_readahead(entry, gfp, NULL, 0);
L
Linus Torvalds 已提交
1194 1195
}

1196 1197
static inline struct page *shmem_alloc_page(gfp_t gfp,
			struct shmem_inode_info *info, unsigned long idx)
L
Linus Torvalds 已提交
1198
{
1199
	return alloc_page(gfp);
L
Linus Torvalds 已提交
1200
}
1201
#endif /* CONFIG_NUMA */
L
Linus Torvalds 已提交
1202

1203 1204 1205 1206 1207 1208 1209
#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
1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226
/*
 * 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;
1227
	gfp_t gfp;
L
Linus Torvalds 已提交
1228 1229 1230 1231
	int error;

	if (idx >= SHMEM_MAX_INDEX)
		return -EFBIG;
1232 1233

	if (type)
N
Nick Piggin 已提交
1234
		*type = 0;
1235

L
Linus Torvalds 已提交
1236 1237 1238 1239
	/*
	 * 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
1240
	 * But shmem_readpage (required for splice) passes in a locked
1241 1242
	 * 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
1243 1244 1245 1246 1247 1248 1249
	 */
repeat:
	if (!filepage)
		filepage = find_lock_page(mapping, idx);
	if (filepage && PageUptodate(filepage))
		goto done;
	error = 0;
1250
	gfp = mapping_gfp_mask(mapping);
H
Hugh Dickins 已提交
1251 1252 1253 1254 1255 1256 1257 1258 1259 1260
	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
1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277

	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
1278
			if (type && !(*type & VM_FAULT_MAJOR)) {
1279
				__count_vm_event(PGMAJFAULT);
N
Nick Piggin 已提交
1280
				*type |= VM_FAULT_MAJOR;
L
Linus Torvalds 已提交
1281
			}
1282
			spin_unlock(&info->lock);
1283
			swappage = shmem_swapin(swap, gfp, info, idx);
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
			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
1305
		if (!trylock_page(swappage)) {
L
Linus Torvalds 已提交
1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340
			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
1341 1342
		} else if (!(error = add_to_page_cache_locked(swappage, mapping,
					idx, GFP_NOWAIT))) {
L
Linus Torvalds 已提交
1343 1344 1345
			info->flags |= SHMEM_PAGEIN;
			shmem_swp_set(info, entry, 0);
			shmem_swp_unmap(entry);
1346
			delete_from_swap_cache(swappage);
L
Linus Torvalds 已提交
1347 1348
			spin_unlock(&info->lock);
			filepage = swappage;
1349
			set_page_dirty(filepage);
L
Linus Torvalds 已提交
1350 1351 1352 1353
			swap_free(swap);
		} else {
			shmem_swp_unmap(entry);
			spin_unlock(&info->lock);
1354
			if (error == -ENOMEM) {
1355 1356 1357 1358 1359 1360
				/*
				 * reclaim from proper memory cgroup and
				 * call memcg's OOM if needed.
				 */
				error = mem_cgroup_shmem_charge_fallback(
								swappage,
1361
								current->mm,
1362
								gfp);
1363 1364 1365
				if (error) {
					unlock_page(swappage);
					page_cache_release(swappage);
1366
					goto failed;
1367
				}
1368
			}
1369 1370
			unlock_page(swappage);
			page_cache_release(swappage);
L
Linus Torvalds 已提交
1371 1372 1373 1374 1375 1376
			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
1377
		    (!PageUptodate(filepage) || !trylock_page(filepage))) {
L
Linus Torvalds 已提交
1378 1379 1380 1381 1382 1383 1384 1385 1386 1387
			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);
1388
		if (sbinfo->max_blocks) {
1389
			if ((percpu_counter_compare(&sbinfo->used_blocks, sbinfo->max_blocks) > 0) ||
L
Linus Torvalds 已提交
1390 1391 1392 1393 1394
			    shmem_acct_block(info->flags)) {
				spin_unlock(&info->lock);
				error = -ENOSPC;
				goto failed;
			}
1395 1396
			percpu_counter_inc(&sbinfo->used_blocks);
			spin_lock(&inode->i_lock);
L
Linus Torvalds 已提交
1397
			inode->i_blocks += BLOCKS_PER_PAGE;
1398
			spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
1399 1400 1401 1402 1403 1404 1405
		} else if (shmem_acct_block(info->flags)) {
			spin_unlock(&info->lock);
			error = -ENOSPC;
			goto failed;
		}

		if (!filepage) {
1406 1407
			int ret;

L
Linus Torvalds 已提交
1408
			spin_unlock(&info->lock);
1409
			filepage = shmem_alloc_page(gfp, info, idx);
L
Linus Torvalds 已提交
1410 1411 1412 1413 1414 1415
			if (!filepage) {
				shmem_unacct_blocks(info->flags, 1);
				shmem_free_blocks(inode, 1);
				error = -ENOMEM;
				goto failed;
			}
R
Rik van Riel 已提交
1416
			SetPageSwapBacked(filepage);
L
Linus Torvalds 已提交
1417

1418 1419
			/* Precharge page while we can wait, compensate after */
			error = mem_cgroup_cache_charge(filepage, current->mm,
K
KAMEZAWA Hiroyuki 已提交
1420
					GFP_KERNEL);
1421 1422 1423 1424 1425 1426 1427 1428
			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
1429 1430 1431 1432 1433 1434 1435 1436
			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);
			}
1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447
			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
1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461
				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);
1462
		clear_highpage(filepage);
L
Linus Torvalds 已提交
1463 1464
		flush_dcache_page(filepage);
		SetPageUptodate(filepage);
1465 1466
		if (sgp == SGP_DIRTY)
			set_page_dirty(filepage);
L
Linus Torvalds 已提交
1467 1468
	}
done:
H
Hugh Dickins 已提交
1469
	*pagep = filepage;
L
Linus Torvalds 已提交
1470 1471 1472 1473 1474 1475 1476 1477 1478 1479
	return 0;

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

N
Nick Piggin 已提交
1480
static int shmem_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
L
Linus Torvalds 已提交
1481
{
1482
	struct inode *inode = vma->vm_file->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
1483
	int error;
N
Nick Piggin 已提交
1484
	int ret;
L
Linus Torvalds 已提交
1485

N
Nick Piggin 已提交
1486 1487
	if (((loff_t)vmf->pgoff << PAGE_CACHE_SHIFT) >= i_size_read(inode))
		return VM_FAULT_SIGBUS;
1488

1489
	error = shmem_getpage(inode, vmf->pgoff, &vmf->page, SGP_CACHE, &ret);
N
Nick Piggin 已提交
1490 1491
	if (error)
		return ((error == -ENOMEM) ? VM_FAULT_OOM : VM_FAULT_SIGBUS);
L
Linus Torvalds 已提交
1492

N
Nick Piggin 已提交
1493
	return ret | VM_FAULT_LOCKED;
L
Linus Torvalds 已提交
1494 1495 1496
}

#ifdef CONFIG_NUMA
A
Adrian Bunk 已提交
1497
static int shmem_set_policy(struct vm_area_struct *vma, struct mempolicy *new)
L
Linus Torvalds 已提交
1498
{
1499
	struct inode *i = vma->vm_file->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
1500 1501 1502
	return mpol_set_shared_policy(&SHMEM_I(i)->policy, vma, new);
}

A
Adrian Bunk 已提交
1503 1504
static struct mempolicy *shmem_get_policy(struct vm_area_struct *vma,
					  unsigned long addr)
L
Linus Torvalds 已提交
1505
{
1506
	struct inode *i = vma->vm_file->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
1507 1508 1509 1510 1511 1512 1513 1514 1515
	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)
{
1516
	struct inode *inode = file->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
1517 1518 1519 1520 1521 1522 1523 1524
	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;
1525
		mapping_set_unevictable(file->f_mapping);
L
Linus Torvalds 已提交
1526 1527 1528 1529
	}
	if (!lock && (info->flags & VM_LOCKED) && user) {
		user_shm_unlock(inode->i_size, user);
		info->flags &= ~VM_LOCKED;
1530 1531
		mapping_clear_unevictable(file->f_mapping);
		scan_mapping_unevictable_pages(file->f_mapping);
L
Linus Torvalds 已提交
1532 1533
	}
	retval = 0;
1534

L
Linus Torvalds 已提交
1535 1536 1537 1538 1539
out_nomem:
	spin_unlock(&info->lock);
	return retval;
}

1540
static int shmem_mmap(struct file *file, struct vm_area_struct *vma)
L
Linus Torvalds 已提交
1541 1542 1543
{
	file_accessed(file);
	vma->vm_ops = &shmem_vm_ops;
N
Nick Piggin 已提交
1544
	vma->vm_flags |= VM_CAN_NONLINEAR;
L
Linus Torvalds 已提交
1545 1546 1547
	return 0;
}

1548 1549
static struct inode *shmem_get_inode(struct super_block *sb, const struct inode *dir,
				     int mode, dev_t dev, unsigned long flags)
L
Linus Torvalds 已提交
1550 1551 1552 1553 1554
{
	struct inode *inode;
	struct shmem_inode_info *info;
	struct shmem_sb_info *sbinfo = SHMEM_SB(sb);

1555 1556
	if (shmem_reserve_inode(sb))
		return NULL;
L
Linus Torvalds 已提交
1557 1558 1559

	inode = new_inode(sb);
	if (inode) {
1560
		inode_init_owner(inode, dir, mode);
L
Linus Torvalds 已提交
1561 1562 1563
		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;
1564
		inode->i_generation = get_seconds();
L
Linus Torvalds 已提交
1565 1566 1567
		info = SHMEM_I(inode);
		memset(info, 0, (char *)inode - (char *)info);
		spin_lock_init(&info->lock);
1568
		info->flags = flags & VM_NORESERVE;
L
Linus Torvalds 已提交
1569
		INIT_LIST_HEAD(&info->swaplist);
1570
		cache_no_acl(inode);
L
Linus Torvalds 已提交
1571 1572 1573

		switch (mode & S_IFMT) {
		default:
1574
			inode->i_op = &shmem_special_inode_operations;
L
Linus Torvalds 已提交
1575 1576 1577
			init_special_inode(inode, mode, dev);
			break;
		case S_IFREG:
1578
			inode->i_mapping->a_ops = &shmem_aops;
L
Linus Torvalds 已提交
1579 1580
			inode->i_op = &shmem_inode_operations;
			inode->i_fop = &shmem_file_operations;
1581 1582
			mpol_shared_policy_init(&info->policy,
						 shmem_get_sbmpol(sbinfo));
L
Linus Torvalds 已提交
1583 1584
			break;
		case S_IFDIR:
1585
			inc_nlink(inode);
L
Linus Torvalds 已提交
1586 1587 1588 1589 1590 1591 1592 1593 1594 1595
			/* 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.
			 */
1596
			mpol_shared_policy_init(&info->policy, NULL);
L
Linus Torvalds 已提交
1597 1598
			break;
		}
1599 1600
	} else
		shmem_free_inode(sb);
L
Linus Torvalds 已提交
1601 1602 1603 1604
	return inode;
}

#ifdef CONFIG_TMPFS
1605 1606
static const struct inode_operations shmem_symlink_inode_operations;
static const struct inode_operations shmem_symlink_inline_operations;
L
Linus Torvalds 已提交
1607 1608

/*
N
Nick Piggin 已提交
1609
 * Normally tmpfs avoids the use of shmem_readpage and shmem_write_begin;
1610 1611
 * 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
1612
 */
1613 1614 1615 1616 1617 1618 1619 1620
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
1621
static int
N
Nick Piggin 已提交
1622 1623 1624
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
1625
{
N
Nick Piggin 已提交
1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638
	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
1639 1640 1641
	if (pos + copied > inode->i_size)
		i_size_write(inode, pos + copied);

N
Nick Piggin 已提交
1642
	set_page_dirty(page);
1643
	unlock_page(page);
N
Nick Piggin 已提交
1644 1645 1646
	page_cache_release(page);

	return copied;
L
Linus Torvalds 已提交
1647 1648 1649 1650
}

static void do_shmem_file_read(struct file *filp, loff_t *ppos, read_descriptor_t *desc, read_actor_t actor)
{
1651
	struct inode *inode = filp->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
1652 1653
	struct address_space *mapping = inode->i_mapping;
	unsigned long index, offset;
1654 1655 1656 1657 1658 1659 1660 1661 1662
	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
1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680

	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;
		}

1681
		desc->error = shmem_getpage(inode, index, &page, sgp, NULL);
L
Linus Torvalds 已提交
1682 1683 1684 1685 1686
		if (desc->error) {
			if (desc->error == -EINVAL)
				desc->error = 0;
			break;
		}
H
Hugh Dickins 已提交
1687 1688
		if (page)
			unlock_page(page);
L
Linus Torvalds 已提交
1689 1690 1691

		/*
		 * We must evaluate after, since reads (unlike writes)
1692
		 * are called without i_mutex protection against truncate
L
Linus Torvalds 已提交
1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719
		 */
		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
1720
		} else {
L
Linus Torvalds 已提交
1721
			page = ZERO_PAGE(0);
N
Nick Piggin 已提交
1722 1723
			page_cache_get(page);
		}
L
Linus Torvalds 已提交
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 1750

		/*
		 * 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
1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782
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
1783 1784
}

1785
static int shmem_statfs(struct dentry *dentry, struct kstatfs *buf)
L
Linus Torvalds 已提交
1786
{
1787
	struct shmem_sb_info *sbinfo = SHMEM_SB(dentry->d_sb);
L
Linus Torvalds 已提交
1788 1789 1790 1791

	buf->f_type = TMPFS_MAGIC;
	buf->f_bsize = PAGE_CACHE_SIZE;
	buf->f_namelen = NAME_MAX;
1792
	if (sbinfo->max_blocks) {
L
Linus Torvalds 已提交
1793
		buf->f_blocks = sbinfo->max_blocks;
1794 1795
		buf->f_bavail = buf->f_bfree =
				sbinfo->max_blocks - percpu_counter_sum(&sbinfo->used_blocks);
1796 1797
	}
	if (sbinfo->max_inodes) {
L
Linus Torvalds 已提交
1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810
		buf->f_files = sbinfo->max_inodes;
		buf->f_ffree = sbinfo->free_inodes;
	}
	/* else leave those fields 0 like simple_statfs */
	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)
{
1811
	struct inode *inode;
L
Linus Torvalds 已提交
1812 1813
	int error = -ENOSPC;

1814
	inode = shmem_get_inode(dir->i_sb, dir, mode, dev, VM_NORESERVE);
L
Linus Torvalds 已提交
1815
	if (inode) {
1816 1817 1818 1819 1820 1821 1822
		error = security_inode_init_security(inode, dir, NULL, NULL,
						     NULL);
		if (error) {
			if (error != -EOPNOTSUPP) {
				iput(inode);
				return error;
			}
1823
		}
1824 1825
#ifdef CONFIG_TMPFS_POSIX_ACL
		error = generic_acl_init(inode, dir);
1826 1827 1828
		if (error) {
			iput(inode);
			return error;
1829
		}
A
Al Viro 已提交
1830 1831
#else
		error = 0;
1832
#endif
L
Linus Torvalds 已提交
1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846
		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;
1847
	inc_nlink(dir);
L
Linus Torvalds 已提交
1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862
	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;
1863
	int ret;
L
Linus Torvalds 已提交
1864 1865 1866 1867 1868 1869

	/*
	 * 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.
	 */
1870 1871 1872
	ret = shmem_reserve_inode(inode->i_sb);
	if (ret)
		goto out;
L
Linus Torvalds 已提交
1873 1874 1875

	dir->i_size += BOGO_DIRENT_SIZE;
	inode->i_ctime = dir->i_ctime = dir->i_mtime = CURRENT_TIME;
1876
	inc_nlink(inode);
L
Linus Torvalds 已提交
1877 1878 1879
	atomic_inc(&inode->i_count);	/* New dentry reference */
	dget(dentry);		/* Extra pinning count for the created dentry */
	d_instantiate(dentry, inode);
1880 1881
out:
	return ret;
L
Linus Torvalds 已提交
1882 1883 1884 1885 1886 1887
}

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

1888 1889
	if (inode->i_nlink > 1 && !S_ISDIR(inode->i_mode))
		shmem_free_inode(inode->i_sb);
L
Linus Torvalds 已提交
1890 1891 1892

	dir->i_size -= BOGO_DIRENT_SIZE;
	inode->i_ctime = dir->i_ctime = dir->i_mtime = CURRENT_TIME;
1893
	drop_nlink(inode);
L
Linus Torvalds 已提交
1894 1895 1896 1897 1898 1899 1900 1901 1902
	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;

1903 1904
	drop_nlink(dentry->d_inode);
	drop_nlink(dir);
L
Linus Torvalds 已提交
1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924
	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)
1925
			drop_nlink(old_dir);
L
Linus Torvalds 已提交
1926
	} else if (they_are_dirs) {
1927
		drop_nlink(old_dir);
1928
		inc_nlink(new_dir);
L
Linus Torvalds 已提交
1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951
	}

	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;

1952
	inode = shmem_get_inode(dir->i_sb, dir, S_IFLNK|S_IRWXUGO, 0, VM_NORESERVE);
L
Linus Torvalds 已提交
1953 1954 1955
	if (!inode)
		return -ENOSPC;

1956 1957 1958 1959 1960 1961 1962 1963 1964 1965
	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
1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977
	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;
		}
1978
		inode->i_mapping->a_ops = &shmem_aops;
L
Linus Torvalds 已提交
1979 1980 1981 1982 1983
		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);
1984
		unlock_page(page);
L
Linus Torvalds 已提交
1985 1986 1987 1988 1989 1990 1991 1992 1993
		page_cache_release(page);
	}
	dir->i_size += BOGO_DIRENT_SIZE;
	dir->i_ctime = dir->i_mtime = CURRENT_TIME;
	d_instantiate(dentry, inode);
	dget(dentry);
	return 0;
}

1994
static void *shmem_follow_link_inline(struct dentry *dentry, struct nameidata *nd)
L
Linus Torvalds 已提交
1995 1996
{
	nd_set_link(nd, (char *)SHMEM_I(dentry->d_inode));
1997
	return NULL;
L
Linus Torvalds 已提交
1998 1999
}

2000
static void *shmem_follow_link(struct dentry *dentry, struct nameidata *nd)
L
Linus Torvalds 已提交
2001 2002 2003 2004
{
	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
2005 2006
	if (page)
		unlock_page(page);
2007
	return page;
L
Linus Torvalds 已提交
2008 2009
}

2010
static void shmem_put_link(struct dentry *dentry, struct nameidata *nd, void *cookie)
L
Linus Torvalds 已提交
2011 2012
{
	if (!IS_ERR(nd_get_link(nd))) {
2013
		struct page *page = cookie;
L
Linus Torvalds 已提交
2014 2015 2016 2017 2018 2019
		kunmap(page);
		mark_page_accessed(page);
		page_cache_release(page);
	}
}

2020
static const struct inode_operations shmem_symlink_inline_operations = {
L
Linus Torvalds 已提交
2021 2022 2023 2024
	.readlink	= generic_readlink,
	.follow_link	= shmem_follow_link_inline,
};

2025
static const struct inode_operations shmem_symlink_inode_operations = {
L
Linus Torvalds 已提交
2026 2027 2028 2029 2030
	.readlink	= generic_readlink,
	.follow_link	= shmem_follow_link,
	.put_link	= shmem_put_link,
};

2031
#ifdef CONFIG_TMPFS_POSIX_ACL
2032
/*
2033 2034 2035 2036 2037 2038
 * 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.
 */

2039
static size_t shmem_xattr_security_list(struct dentry *dentry, char *list,
2040
					size_t list_len, const char *name,
2041
					size_t name_len, int handler_flags)
2042
{
2043
	return security_inode_listsecurity(dentry->d_inode, list, list_len);
2044 2045
}

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

2054 2055
static int shmem_xattr_security_set(struct dentry *dentry, const char *name,
		const void *value, size_t size, int flags, int handler_flags)
2056 2057 2058
{
	if (strcmp(name, "") == 0)
		return -EINVAL;
2059 2060
	return security_inode_setsecurity(dentry->d_inode, name, value,
					  size, flags);
2061 2062
}

2063
static const struct xattr_handler shmem_xattr_security_handler = {
2064 2065 2066 2067 2068 2069
	.prefix = XATTR_SECURITY_PREFIX,
	.list   = shmem_xattr_security_list,
	.get    = shmem_xattr_security_get,
	.set    = shmem_xattr_security_set,
};

2070
static const struct xattr_handler *shmem_xattr_handlers[] = {
2071 2072
	&generic_acl_access_handler,
	&generic_acl_default_handler,
2073 2074 2075 2076 2077
	&shmem_xattr_security_handler,
	NULL
};
#endif

2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090
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
2091 2092
static struct dentry *shmem_fh_to_dentry(struct super_block *sb,
		struct fid *fid, int fh_len, int fh_type)
2093 2094
{
	struct inode *inode;
C
Christoph Hellwig 已提交
2095 2096 2097 2098 2099 2100
	struct dentry *dentry = NULL;
	u64 inum = fid->raw[2];
	inum = (inum << 32) | fid->raw[1];

	if (fh_len < 3)
		return NULL;
2101

C
Christoph Hellwig 已提交
2102 2103
	inode = ilookup5(sb, (unsigned long)(inum + fid->raw[0]),
			shmem_match, fid->raw);
2104
	if (inode) {
C
Christoph Hellwig 已提交
2105
		dentry = d_find_alias(inode);
2106 2107 2108
		iput(inode);
	}

C
Christoph Hellwig 已提交
2109
	return dentry;
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 2136 2137 2138 2139 2140 2141
}

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;
}

2142
static const struct export_operations shmem_export_ops = {
2143 2144
	.get_parent     = shmem_get_parent,
	.encode_fh      = shmem_encode_fh,
C
Christoph Hellwig 已提交
2145
	.fh_to_dentry	= shmem_fh_to_dentry,
2146 2147
};

2148 2149
static int shmem_parse_options(char *options, struct shmem_sb_info *sbinfo,
			       bool remount)
L
Linus Torvalds 已提交
2150 2151 2152
{
	char *this_char, *value, *rest;

2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169
	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
2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191
		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;
2192 2193
			sbinfo->max_blocks =
				DIV_ROUND_UP(size, PAGE_CACHE_SIZE);
L
Linus Torvalds 已提交
2194
		} else if (!strcmp(this_char,"nr_blocks")) {
2195
			sbinfo->max_blocks = memparse(value, &rest);
L
Linus Torvalds 已提交
2196 2197 2198
			if (*rest)
				goto bad_val;
		} else if (!strcmp(this_char,"nr_inodes")) {
2199
			sbinfo->max_inodes = memparse(value, &rest);
L
Linus Torvalds 已提交
2200 2201 2202
			if (*rest)
				goto bad_val;
		} else if (!strcmp(this_char,"mode")) {
2203
			if (remount)
L
Linus Torvalds 已提交
2204
				continue;
2205
			sbinfo->mode = simple_strtoul(value, &rest, 8) & 07777;
L
Linus Torvalds 已提交
2206 2207 2208
			if (*rest)
				goto bad_val;
		} else if (!strcmp(this_char,"uid")) {
2209
			if (remount)
L
Linus Torvalds 已提交
2210
				continue;
2211
			sbinfo->uid = simple_strtoul(value, &rest, 0);
L
Linus Torvalds 已提交
2212 2213 2214
			if (*rest)
				goto bad_val;
		} else if (!strcmp(this_char,"gid")) {
2215
			if (remount)
L
Linus Torvalds 已提交
2216
				continue;
2217
			sbinfo->gid = simple_strtoul(value, &rest, 0);
L
Linus Torvalds 已提交
2218 2219
			if (*rest)
				goto bad_val;
2220
		} else if (!strcmp(this_char,"mpol")) {
2221
			if (mpol_parse_str(value, &sbinfo->mpol, 1))
2222
				goto bad_val;
L
Linus Torvalds 已提交
2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240
		} 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);
2241
	struct shmem_sb_info config = *sbinfo;
2242 2243 2244
	unsigned long inodes;
	int error = -EINVAL;

2245
	if (shmem_parse_options(data, &config, true))
2246
		return error;
L
Linus Torvalds 已提交
2247

2248 2249
	spin_lock(&sbinfo->stat_lock);
	inodes = sbinfo->max_inodes - sbinfo->free_inodes;
2250
	if (percpu_counter_compare(&sbinfo->used_blocks, config.max_blocks) > 0)
2251
		goto out;
2252
	if (config.max_inodes < inodes)
2253 2254 2255 2256 2257 2258 2259
		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.
	 */
2260
	if (config.max_blocks && !sbinfo->max_blocks)
2261
		goto out;
2262
	if (config.max_inodes && !sbinfo->max_inodes)
2263 2264 2265
		goto out;

	error = 0;
2266 2267 2268
	sbinfo->max_blocks  = config.max_blocks;
	sbinfo->max_inodes  = config.max_inodes;
	sbinfo->free_inodes = config.max_inodes - inodes;
2269 2270 2271

	mpol_put(sbinfo->mpol);
	sbinfo->mpol        = config.mpol;	/* transfers initial ref */
2272 2273 2274
out:
	spin_unlock(&sbinfo->stat_lock);
	return error;
L
Linus Torvalds 已提交
2275
}
2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291

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);
2292
	shmem_show_mpol(seq, sbinfo->mpol);
2293 2294 2295
	return 0;
}
#endif /* CONFIG_TMPFS */
L
Linus Torvalds 已提交
2296 2297 2298 2299 2300 2301 2302

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

2303
int shmem_fill_super(struct super_block *sb, void *data, int silent)
L
Linus Torvalds 已提交
2304 2305 2306
{
	struct inode *inode;
	struct dentry *root;
2307
	struct shmem_sb_info *sbinfo;
2308 2309 2310
	int err = -ENOMEM;

	/* Round up to L1_CACHE_BYTES to resist false sharing */
2311
	sbinfo = kzalloc(max((int)sizeof(struct shmem_sb_info),
2312 2313 2314 2315 2316
				L1_CACHE_BYTES), GFP_KERNEL);
	if (!sbinfo)
		return -ENOMEM;

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

2321
#ifdef CONFIG_TMPFS
L
Linus Torvalds 已提交
2322 2323 2324 2325 2326 2327
	/*
	 * 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)) {
2328 2329 2330 2331 2332 2333
		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
2334
	}
2335
	sb->s_export_op = &shmem_export_ops;
L
Linus Torvalds 已提交
2336 2337 2338 2339
#else
	sb->s_flags |= MS_NOUSER;
#endif

2340
	spin_lock_init(&sbinfo->stat_lock);
2341
	percpu_counter_init(&sbinfo->used_blocks, 0);
2342
	sbinfo->free_inodes = sbinfo->max_inodes;
2343

L
Linus Torvalds 已提交
2344 2345 2346 2347 2348
	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;
2349
	sb->s_time_gran = 1;
2350 2351 2352 2353
#ifdef CONFIG_TMPFS_POSIX_ACL
	sb->s_xattr = shmem_xattr_handlers;
	sb->s_flags |= MS_POSIXACL;
#endif
2354

2355
	inode = shmem_get_inode(sb, NULL, S_IFDIR | sbinfo->mode, 0, VM_NORESERVE);
L
Linus Torvalds 已提交
2356 2357
	if (!inode)
		goto failed;
2358 2359
	inode->i_uid = sbinfo->uid;
	inode->i_gid = sbinfo->gid;
L
Linus Torvalds 已提交
2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372
	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;
}

2373
static struct kmem_cache *shmem_inode_cachep;
L
Linus Torvalds 已提交
2374 2375 2376 2377

static struct inode *shmem_alloc_inode(struct super_block *sb)
{
	struct shmem_inode_info *p;
2378
	p = (struct shmem_inode_info *)kmem_cache_alloc(shmem_inode_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392
	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);
	}
	kmem_cache_free(shmem_inode_cachep, SHMEM_I(inode));
}

2393
static void init_once(void *foo)
L
Linus Torvalds 已提交
2394 2395 2396
{
	struct shmem_inode_info *p = (struct shmem_inode_info *) foo;

C
Christoph Lameter 已提交
2397
	inode_init_once(&p->vfs_inode);
L
Linus Torvalds 已提交
2398 2399 2400 2401 2402 2403
}

static int init_inodecache(void)
{
	shmem_inode_cachep = kmem_cache_create("shmem_inode_cache",
				sizeof(struct shmem_inode_info),
2404
				0, SLAB_PANIC, init_once);
L
Linus Torvalds 已提交
2405 2406 2407 2408 2409
	return 0;
}

static void destroy_inodecache(void)
{
2410
	kmem_cache_destroy(shmem_inode_cachep);
L
Linus Torvalds 已提交
2411 2412
}

2413
static const struct address_space_operations shmem_aops = {
L
Linus Torvalds 已提交
2414
	.writepage	= shmem_writepage,
2415
	.set_page_dirty	= __set_page_dirty_no_writeback,
L
Linus Torvalds 已提交
2416
#ifdef CONFIG_TMPFS
2417
	.readpage	= shmem_readpage,
N
Nick Piggin 已提交
2418 2419
	.write_begin	= shmem_write_begin,
	.write_end	= shmem_write_end,
L
Linus Torvalds 已提交
2420
#endif
2421
	.migratepage	= migrate_page,
2422
	.error_remove_page = generic_error_remove_page,
L
Linus Torvalds 已提交
2423 2424
};

2425
static const struct file_operations shmem_file_operations = {
L
Linus Torvalds 已提交
2426 2427 2428
	.mmap		= shmem_mmap,
#ifdef CONFIG_TMPFS
	.llseek		= generic_file_llseek,
H
Hugh Dickins 已提交
2429
	.read		= do_sync_read,
H
Hugh Dickins 已提交
2430
	.write		= do_sync_write,
H
Hugh Dickins 已提交
2431
	.aio_read	= shmem_file_aio_read,
H
Hugh Dickins 已提交
2432
	.aio_write	= generic_file_aio_write,
2433
	.fsync		= noop_fsync,
2434 2435
	.splice_read	= generic_file_splice_read,
	.splice_write	= generic_file_splice_write,
L
Linus Torvalds 已提交
2436 2437 2438
#endif
};

2439
static const struct inode_operations shmem_inode_operations = {
L
Linus Torvalds 已提交
2440
	.setattr	= shmem_notify_change,
2441
	.truncate_range	= shmem_truncate_range,
2442 2443 2444 2445 2446
#ifdef CONFIG_TMPFS_POSIX_ACL
	.setxattr	= generic_setxattr,
	.getxattr	= generic_getxattr,
	.listxattr	= generic_listxattr,
	.removexattr	= generic_removexattr,
2447
	.check_acl	= generic_check_acl,
2448 2449
#endif

L
Linus Torvalds 已提交
2450 2451
};

2452
static const struct inode_operations shmem_dir_inode_operations = {
L
Linus Torvalds 已提交
2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463
#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
2464 2465 2466 2467 2468 2469
#ifdef CONFIG_TMPFS_POSIX_ACL
	.setattr	= shmem_notify_change,
	.setxattr	= generic_setxattr,
	.getxattr	= generic_getxattr,
	.listxattr	= generic_listxattr,
	.removexattr	= generic_removexattr,
2470
	.check_acl	= generic_check_acl,
2471 2472 2473
#endif
};

2474
static const struct inode_operations shmem_special_inode_operations = {
2475 2476 2477 2478 2479 2480
#ifdef CONFIG_TMPFS_POSIX_ACL
	.setattr	= shmem_notify_change,
	.setxattr	= generic_setxattr,
	.getxattr	= generic_getxattr,
	.listxattr	= generic_listxattr,
	.removexattr	= generic_removexattr,
2481
	.check_acl	= generic_check_acl,
2482
#endif
L
Linus Torvalds 已提交
2483 2484
};

2485
static const struct super_operations shmem_ops = {
L
Linus Torvalds 已提交
2486 2487 2488 2489 2490
	.alloc_inode	= shmem_alloc_inode,
	.destroy_inode	= shmem_destroy_inode,
#ifdef CONFIG_TMPFS
	.statfs		= shmem_statfs,
	.remount_fs	= shmem_remount_fs,
2491
	.show_options	= shmem_show_options,
L
Linus Torvalds 已提交
2492 2493 2494 2495 2496 2497
#endif
	.delete_inode	= shmem_delete_inode,
	.drop_inode	= generic_delete_inode,
	.put_super	= shmem_put_super,
};

2498
static const struct vm_operations_struct shmem_vm_ops = {
2499
	.fault		= shmem_fault,
L
Linus Torvalds 已提交
2500 2501 2502 2503 2504 2505 2506
#ifdef CONFIG_NUMA
	.set_policy     = shmem_set_policy,
	.get_policy     = shmem_get_policy,
#endif
};


2507 2508
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
2509
{
2510
	return get_sb_nodev(fs_type, flags, data, shmem_fill_super, mnt);
L
Linus Torvalds 已提交
2511 2512 2513 2514 2515 2516 2517 2518 2519
}

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

2520
int __init init_tmpfs(void)
L
Linus Torvalds 已提交
2521 2522 2523
{
	int error;

P
Peter Zijlstra 已提交
2524 2525 2526 2527
	error = bdi_init(&shmem_backing_dev_info);
	if (error)
		goto out4;

L
Linus Torvalds 已提交
2528 2529 2530 2531 2532 2533 2534 2535 2536
	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;
	}
2537

2538
	shm_mnt = vfs_kern_mount(&tmpfs_fs_type, MS_NOUSER,
L
Linus Torvalds 已提交
2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551
				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
2552 2553
	bdi_destroy(&shmem_backing_dev_info);
out4:
L
Linus Torvalds 已提交
2554 2555 2556
	shm_mnt = ERR_PTR(error);
	return error;
}
2557

2558 2559 2560 2561 2562 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
#ifdef CONFIG_CGROUP_MEM_RES_CTLR
/**
 * mem_cgroup_get_shmem_target - find a page or entry assigned to the shmem file
 * @inode: the inode to be searched
 * @pgoff: the offset to be searched
 * @pagep: the pointer for the found page to be stored
 * @ent: the pointer for the found swap entry to be stored
 *
 * If a page is found, refcount of it is incremented. Callers should handle
 * these refcount.
 */
void mem_cgroup_get_shmem_target(struct inode *inode, pgoff_t pgoff,
					struct page **pagep, swp_entry_t *ent)
{
	swp_entry_t entry = { .val = 0 }, *ptr;
	struct page *page = NULL;
	struct shmem_inode_info *info = SHMEM_I(inode);

	if ((pgoff << PAGE_CACHE_SHIFT) >= i_size_read(inode))
		goto out;

	spin_lock(&info->lock);
	ptr = shmem_swp_entry(info, pgoff, NULL);
#ifdef CONFIG_SWAP
	if (ptr && ptr->val) {
		entry.val = ptr->val;
		page = find_get_page(&swapper_space, entry.val);
	} else
#endif
		page = find_get_page(inode->i_mapping, pgoff);
	if (ptr)
		shmem_swp_unmap(ptr);
	spin_unlock(&info->lock);
out:
	*pagep = page;
	*ent = entry;
}
#endif

2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615
#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,
};

2616
int __init init_tmpfs(void)
2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630
{
	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;
}

H
Hugh Dickins 已提交
2631 2632 2633 2634 2635
int shmem_lock(struct file *file, int lock, struct user_struct *user)
{
	return 0;
}

2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660
#ifdef CONFIG_CGROUP_MEM_RES_CTLR
/**
 * mem_cgroup_get_shmem_target - find a page or entry assigned to the shmem file
 * @inode: the inode to be searched
 * @pgoff: the offset to be searched
 * @pagep: the pointer for the found page to be stored
 * @ent: the pointer for the found swap entry to be stored
 *
 * If a page is found, refcount of it is incremented. Callers should handle
 * these refcount.
 */
void mem_cgroup_get_shmem_target(struct inode *inode, pgoff_t pgoff,
					struct page **pagep, swp_entry_t *ent)
{
	struct page *page = NULL;

	if ((pgoff << PAGE_CACHE_SHIFT) >= i_size_read(inode))
		goto out;
	page = find_get_page(inode->i_mapping, pgoff);
out:
	*pagep = page;
	*ent = (swp_entry_t){ .val = 0 };
}
#endif

2661 2662
#define shmem_vm_ops				generic_file_vm_ops
#define shmem_file_operations			ramfs_file_operations
2663
#define shmem_get_inode(sb, dir, mode, dev, flags)	ramfs_get_inode(sb, dir, mode, dev)
2664 2665
#define shmem_acct_size(flags, size)		0
#define shmem_unacct_size(flags, size)		do {} while (0)
H
Hugh Dickins 已提交
2666
#define SHMEM_MAX_BYTES				MAX_LFS_FILESIZE
2667 2668 2669 2670

#endif /* CONFIG_SHMEM */

/* common code */
L
Linus Torvalds 已提交
2671

2672
/**
L
Linus Torvalds 已提交
2673 2674 2675
 * 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
2676
 * @flags: VM_NORESERVE suppresses pre-accounting of the entire object size
L
Linus Torvalds 已提交
2677
 */
2678
struct file *shmem_file_setup(const char *name, loff_t size, unsigned long flags)
L
Linus Torvalds 已提交
2679 2680 2681 2682
{
	int error;
	struct file *file;
	struct inode *inode;
2683 2684
	struct path path;
	struct dentry *root;
L
Linus Torvalds 已提交
2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700
	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;
2701 2702
	path.dentry = d_alloc(root, &this);
	if (!path.dentry)
L
Linus Torvalds 已提交
2703
		goto put_memory;
2704
	path.mnt = mntget(shm_mnt);
L
Linus Torvalds 已提交
2705 2706

	error = -ENOSPC;
2707
	inode = shmem_get_inode(root->d_sb, NULL, S_IFREG | S_IRWXUGO, 0, flags);
L
Linus Torvalds 已提交
2708
	if (!inode)
2709
		goto put_dentry;
L
Linus Torvalds 已提交
2710

2711
	d_instantiate(path.dentry, inode);
L
Linus Torvalds 已提交
2712 2713
	inode->i_size = size;
	inode->i_nlink = 0;	/* It is unlinked */
2714 2715 2716
#ifndef CONFIG_MMU
	error = ramfs_nommu_expand_for_mapping(inode, size);
	if (error)
2717
		goto put_dentry;
2718
#endif
2719 2720

	error = -ENFILE;
2721
	file = alloc_file(&path, FMODE_WRITE | FMODE_READ,
2722 2723 2724 2725
		  &shmem_file_operations);
	if (!file)
		goto put_dentry;

L
Linus Torvalds 已提交
2726 2727 2728
	return file;

put_dentry:
2729
	path_put(&path);
L
Linus Torvalds 已提交
2730 2731 2732 2733
put_memory:
	shmem_unacct_size(flags, size);
	return ERR_PTR(error);
}
2734
EXPORT_SYMBOL_GPL(shmem_file_setup);
L
Linus Torvalds 已提交
2735

2736
/**
L
Linus Torvalds 已提交
2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754
 * 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);

	if (vma->vm_file)
		fput(vma->vm_file);
	vma->vm_file = file;
	vma->vm_ops = &shmem_vm_ops;
	return 0;
}