shmem.c 67.7 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 17 18 19 20 21 22 23 24 25 26 27 28
 * 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>
 *
 * This file is released under the GPL.
 */

/*
 * 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.
 */

#include <linux/module.h>
#include <linux/init.h>
#include <linux/fs.h>
29
#include <linux/xattr.h>
30
#include <linux/exportfs.h>
31
#include <linux/generic_acl.h>
L
Linus Torvalds 已提交
32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48
#include <linux/mm.h>
#include <linux/mman.h>
#include <linux/file.h>
#include <linux/swap.h>
#include <linux/pagemap.h>
#include <linux/string.h>
#include <linux/slab.h>
#include <linux/backing-dev.h>
#include <linux/shmem_fs.h>
#include <linux/mount.h>
#include <linux/writeback.h>
#include <linux/vfs.h>
#include <linux/blkdev.h>
#include <linux/security.h>
#include <linux/swapops.h>
#include <linux/mempolicy.h>
#include <linux/namei.h>
49
#include <linux/ctype.h>
50
#include <linux/migrate.h>
51
#include <linux/highmem.h>
52
#include <linux/seq_file.h>
53

L
Linus Torvalds 已提交
54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83
#include <asm/uaccess.h>
#include <asm/div64.h>
#include <asm/pgtable.h>

/* This magic number is used in glibc for posix shared memory */
#define TMPFS_MAGIC	0x01021994

#define ENTRIES_PER_PAGE (PAGE_CACHE_SIZE/sizeof(unsigned long))
#define ENTRIES_PER_PAGEPAGE (ENTRIES_PER_PAGE*ENTRIES_PER_PAGE)
#define BLOCKS_PER_PAGE  (PAGE_CACHE_SIZE/512)

#define SHMEM_MAX_INDEX  (SHMEM_NR_DIRECT + (ENTRIES_PER_PAGEPAGE/2) * (ENTRIES_PER_PAGE+1))
#define SHMEM_MAX_BYTES  ((unsigned long long)SHMEM_MAX_INDEX << PAGE_CACHE_SHIFT)

#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 */
84
	SGP_DIRTY,	/* like SGP_CACHE, but set new page dirty */
L
Linus Torvalds 已提交
85 86 87
	SGP_WRITE,	/* may exceed i_size, may allocate page */
};

A
Andrew Morton 已提交
88
#ifdef CONFIG_TMPFS
89 90 91 92 93 94 95 96 97
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
98
#endif
99

L
Linus Torvalds 已提交
100 101 102
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
103
static inline struct page *shmem_dir_alloc(gfp_t gfp_mask)
L
Linus Torvalds 已提交
104 105 106 107 108
{
	/*
	 * 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.
109
	 *
110
	 * Mobility flags are masked out as swap vectors cannot move
L
Linus Torvalds 已提交
111
	 */
112
	return alloc_pages((gfp_mask & ~GFP_MOVABLE_MASK) | __GFP_ZERO,
113
				PAGE_CACHE_SHIFT-PAGE_SHIFT);
L
Linus Torvalds 已提交
114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193
}

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)
{
	return (flags & VM_ACCOUNT)?
		security_vm_enough_memory(VM_ACCT(size)): 0;
}

static inline void shmem_unacct_size(unsigned long flags, loff_t size)
{
	if (flags & VM_ACCOUNT)
		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)
{
	return (flags & VM_ACCOUNT)?
		0: security_vm_enough_memory(VM_ACCT(PAGE_CACHE_SIZE));
}

static inline void shmem_unacct_blocks(unsigned long flags, long pages)
{
	if (!(flags & VM_ACCOUNT))
		vm_unacct_memory(pages * VM_ACCT(PAGE_CACHE_SIZE));
}

194
static const struct super_operations shmem_ops;
195
static const struct address_space_operations shmem_aops;
196
static const struct file_operations shmem_file_operations;
197 198 199
static const struct inode_operations shmem_inode_operations;
static const struct inode_operations shmem_dir_inode_operations;
static const struct inode_operations shmem_special_inode_operations;
L
Linus Torvalds 已提交
200 201
static struct vm_operations_struct shmem_vm_ops;

202
static struct backing_dev_info shmem_backing_dev_info  __read_mostly = {
L
Linus Torvalds 已提交
203 204 205 206 207 208
	.ra_pages	= 0,	/* No readahead */
	.capabilities	= BDI_CAP_NO_ACCT_DIRTY | BDI_CAP_NO_WRITEBACK,
	.unplug_io_fn	= default_unplug_io_fn,
};

static LIST_HEAD(shmem_swaplist);
209
static DEFINE_MUTEX(shmem_swaplist_mutex);
L
Linus Torvalds 已提交
210 211 212 213

static void shmem_free_blocks(struct inode *inode, long pages)
{
	struct shmem_sb_info *sbinfo = SHMEM_SB(inode->i_sb);
214
	if (sbinfo->max_blocks) {
L
Linus Torvalds 已提交
215 216 217 218 219 220 221
		spin_lock(&sbinfo->stat_lock);
		sbinfo->free_blocks += pages;
		inode->i_blocks -= pages*BLOCKS_PER_PAGE;
		spin_unlock(&sbinfo->stat_lock);
	}
}

222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246
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);
	}
}

247
/**
L
Linus Torvalds 已提交
248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271
 * 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);
	}
}

272
/**
L
Linus Torvalds 已提交
273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368
 * 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
369 370 371 372
	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
373 374
}

375
/**
L
Linus Torvalds 已提交
376 377 378 379
 * 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?
380 381
 *
 * If the entry does not exist, allocate it.
L
Linus Torvalds 已提交
382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401
 */
static swp_entry_t *shmem_swp_alloc(struct shmem_inode_info *info, unsigned long index, enum sgp_type sgp)
{
	struct inode *inode = &info->vfs_inode;
	struct shmem_sb_info *sbinfo = SHMEM_SB(inode->i_sb);
	struct page *page = NULL;
	swp_entry_t *entry;

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

	while (!(entry = shmem_swp_entry(info, index, &page))) {
		if (sgp == SGP_READ)
			return shmem_swp_map(ZERO_PAGE(0));
		/*
		 * Test free_blocks against 1 not 0, since we have 1 data
		 * page (and perhaps indirect index pages) yet to allocate:
		 * a waste to allocate index if we cannot allocate data.
		 */
402
		if (sbinfo->max_blocks) {
L
Linus Torvalds 已提交
403 404 405 406 407 408 409 410 411 412 413
			spin_lock(&sbinfo->stat_lock);
			if (sbinfo->free_blocks <= 1) {
				spin_unlock(&sbinfo->stat_lock);
				return ERR_PTR(-ENOSPC);
			}
			sbinfo->free_blocks--;
			inode->i_blocks += BLOCKS_PER_PAGE;
			spin_unlock(&sbinfo->stat_lock);
		}

		spin_unlock(&info->lock);
414
		page = shmem_dir_alloc(mapping_gfp_mask(inode->i_mapping));
H
Hugh Dickins 已提交
415 416
		if (page)
			set_page_private(page, 0);
L
Linus Torvalds 已提交
417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440
		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;
}

441
/**
L
Linus Torvalds 已提交
442
 * shmem_free_swp - free some swap entries in a directory
443 444 445
 * @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
446
 */
447 448
static int shmem_free_swp(swp_entry_t *dir, swp_entry_t *edir,
						spinlock_t *punch_lock)
L
Linus Torvalds 已提交
449
{
450
	spinlock_t *punch_unlock = NULL;
L
Linus Torvalds 已提交
451 452 453 454 455
	swp_entry_t *ptr;
	int freed = 0;

	for (ptr = dir; ptr < edir; ptr++) {
		if (ptr->val) {
456 457 458 459 460 461 462
			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
463 464 465 466 467
			free_swap_and_cache(*ptr);
			*ptr = (swp_entry_t){0};
			freed++;
		}
	}
468 469
	if (punch_unlock)
		spin_unlock(punch_unlock);
L
Linus Torvalds 已提交
470 471 472
	return freed;
}

473 474
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
475 476 477 478 479 480 481 482 483
{
	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;
484 485
		freed += shmem_free_swp(ptr+offset, ptr+offset+size,
							punch_lock);
L
Linus Torvalds 已提交
486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516
		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);
}

517
static void shmem_truncate_range(struct inode *inode, loff_t start, loff_t end)
L
Linus Torvalds 已提交
518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534
{
	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;
535
	int punch_hole;
536 537
	spinlock_t *needs_lock;
	spinlock_t *punch_lock;
538
	unsigned long upper_limit;
L
Linus Torvalds 已提交
539 540

	inode->i_ctime = inode->i_mtime = CURRENT_TIME;
541
	idx = (start + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
L
Linus Torvalds 已提交
542 543 544 545 546
	if (idx >= info->next_index)
		return;

	spin_lock(&info->lock);
	info->flags |= SHMEM_TRUNCATE;
547 548
	if (likely(end == (loff_t) -1)) {
		limit = info->next_index;
549
		upper_limit = SHMEM_MAX_INDEX;
550
		info->next_index = idx;
551
		needs_lock = NULL;
552
		punch_hole = 0;
553
	} else {
554 555 556 557 558 559 560 561
		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;
		}
562
		needs_lock = &info->lock;
563 564 565
		punch_hole = 1;
	}

L
Linus Torvalds 已提交
566
	topdir = info->i_indirect;
567
	if (topdir && idx <= SHMEM_NR_DIRECT && !punch_hole) {
L
Linus Torvalds 已提交
568 569 570 571 572 573 574 575 576 577 578
		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;
579
		nr_swaps_freed = shmem_free_swp(ptr+idx, ptr+size, needs_lock);
L
Linus Torvalds 已提交
580
	}
581 582 583 584 585

	/*
	 * If there are no indirect blocks or we are punching a hole
	 * below indirect blocks, nothing to be done.
	 */
586
	if (!topdir || limit <= SHMEM_NR_DIRECT)
L
Linus Torvalds 已提交
587 588
		goto done2;

589 590 591 592 593 594 595 596 597 598 599 600 601
	/*
	 * 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.
	 */

602
	upper_limit -= SHMEM_NR_DIRECT;
L
Linus Torvalds 已提交
603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621
	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;
622
			if (!diroff && !offset && upper_limit >= stage) {
623 624 625 626 627 628 629
				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
630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654
				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;
655 656
			if (punch_hole)
				needs_lock = &info->lock;
657
			if (upper_limit >= stage) {
658 659 660 661 662 663 664
				if (needs_lock) {
					spin_lock(needs_lock);
					*dir = NULL;
					spin_unlock(needs_lock);
					needs_lock = NULL;
				} else
					*dir = NULL;
665 666 667
				nr_pages_to_free++;
				list_add(&middir->lru, &pages_to_free);
			}
L
Linus Torvalds 已提交
668 669 670 671 672
			shmem_dir_unmap(dir);
			cond_resched();
			dir = shmem_dir_map(middir);
			diroff = 0;
		}
673
		punch_lock = needs_lock;
L
Linus Torvalds 已提交
674
		subdir = dir[diroff];
675 676 677 678 679 680 681 682 683 684 685 686
		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
687 688 689 690
			size = limit - idx;
			if (size > ENTRIES_PER_PAGE)
				size = ENTRIES_PER_PAGE;
			freed = shmem_map_and_free_swp(subdir,
691
					offset, size, &dir, punch_lock);
L
Linus Torvalds 已提交
692 693 694
			if (!dir)
				dir = shmem_dir_map(middir);
			nr_swaps_freed += freed;
695
			if (offset || punch_lock) {
L
Linus Torvalds 已提交
696
				spin_lock(&info->lock);
697 698
				set_page_private(subdir,
					page_private(subdir) - freed);
L
Linus Torvalds 已提交
699
				spin_unlock(&info->lock);
700 701
			} else
				BUG_ON(page_private(subdir) != freed);
L
Linus Torvalds 已提交
702
		}
703
		offset = 0;
L
Linus Torvalds 已提交
704 705 706 707 708 709 710 711 712 713 714
	}
done1:
	shmem_dir_unmap(dir);
done2:
	if (inode->i_mapping->nrpages && (info->flags & SHMEM_PAGEIN)) {
		/*
		 * Call truncate_inode_pages again: racing shmem_unuse_inode
		 * may have swizzled a page in from swap since vmtruncate or
		 * generic_delete_inode did it, before we lowered next_index.
		 * Also, though shmem_getpage checks i_size before adding to
		 * cache, no recheck after: so fix the narrow window there too.
715 716 717 718 719
		 *
		 * 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
720
		 */
721
		truncate_inode_pages_range(inode->i_mapping, start, end);
722 723 724
		if (punch_hole)
			unmap_mapping_range(inode->i_mapping, start,
							end - start, 1);
L
Linus Torvalds 已提交
725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743
	}

	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);
	}
}

744 745 746 747 748
static void shmem_truncate(struct inode *inode)
{
	shmem_truncate_range(inode, inode->i_size, (loff_t)-1);
}

L
Linus Torvalds 已提交
749 750 751 752 753 754
static int shmem_notify_change(struct dentry *dentry, struct iattr *attr)
{
	struct inode *inode = dentry->d_inode;
	struct page *page = NULL;
	int error;

755
	if (S_ISREG(inode->i_mode) && (attr->ia_valid & ATTR_SIZE)) {
L
Linus Torvalds 已提交
756 757 758 759 760 761 762 763 764 765 766 767
		if (attr->ia_size < inode->i_size) {
			/*
			 * If truncating down to a partial page, then
			 * if that page is already allocated, hold it
			 * in memory until the truncation is over, so
			 * truncate_partial_page cannnot miss it were
			 * it assigned to swap.
			 */
			if (attr->ia_size & (PAGE_CACHE_SIZE-1)) {
				(void) shmem_getpage(inode,
					attr->ia_size>>PAGE_CACHE_SHIFT,
						&page, SGP_READ, NULL);
H
Hugh Dickins 已提交
768 769
				if (page)
					unlock_page(page);
L
Linus Torvalds 已提交
770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789
			}
			/*
			 * Reset SHMEM_PAGEIN flag so that shmem_truncate can
			 * detect if any pages might have been added to cache
			 * after truncate_inode_pages.  But we needn't bother
			 * if it's being fully truncated to zero-length: the
			 * nrpages check is efficient enough in that case.
			 */
			if (attr->ia_size) {
				struct shmem_inode_info *info = SHMEM_I(inode);
				spin_lock(&info->lock);
				info->flags &= ~SHMEM_PAGEIN;
				spin_unlock(&info->lock);
			}
		}
	}

	error = inode_change_ok(inode, attr);
	if (!error)
		error = inode_setattr(inode, attr);
790 791 792 793
#ifdef CONFIG_TMPFS_POSIX_ACL
	if (!error && (attr->ia_valid & ATTR_MODE))
		error = generic_acl_chmod(inode, &shmem_acl_ops);
#endif
L
Linus Torvalds 已提交
794 795 796 797 798 799 800 801 802 803
	if (page)
		page_cache_release(page);
	return error;
}

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

	if (inode->i_op->truncate == shmem_truncate) {
804
		truncate_inode_pages(inode->i_mapping, 0);
L
Linus Torvalds 已提交
805 806 807 808
		shmem_unacct_size(info->flags, inode->i_size);
		inode->i_size = 0;
		shmem_truncate(inode);
		if (!list_empty(&info->swaplist)) {
809
			mutex_lock(&shmem_swaplist_mutex);
L
Linus Torvalds 已提交
810
			list_del_init(&info->swaplist);
811
			mutex_unlock(&shmem_swaplist_mutex);
L
Linus Torvalds 已提交
812 813
		}
	}
814
	BUG_ON(inode->i_blocks);
815
	shmem_free_inode(inode->i_sb);
L
Linus Torvalds 已提交
816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840
	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;
841
	int error;
L
Linus Torvalds 已提交
842 843 844 845

	idx = 0;
	ptr = info->i_direct;
	spin_lock(&info->lock);
H
Hugh Dickins 已提交
846 847 848 849
	if (!info->swapped) {
		list_del_init(&info->swaplist);
		goto lost2;
	}
L
Linus Torvalds 已提交
850 851 852 853 854
	limit = info->next_index;
	size = limit;
	if (size > SHMEM_NR_DIRECT)
		size = SHMEM_NR_DIRECT;
	offset = shmem_find_swp(entry, ptr, ptr+size);
855
	if (offset >= 0)
L
Linus Torvalds 已提交
856 857 858 859 860 861 862 863 864 865
		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);
866 867 868 869 870 871 872 873
			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
874 875 876 877 878 879 880 881 882 883 884 885 886 887
			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
888
		if (subdir && page_private(subdir)) {
L
Linus Torvalds 已提交
889 890 891 892 893
			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);
894
			shmem_swp_unmap(ptr);
L
Linus Torvalds 已提交
895 896 897 898 899 900 901 902 903 904 905 906 907
			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;
908 909 910
	inode = igrab(&info->vfs_inode);
	spin_unlock(&info->lock);

H
Hugh Dickins 已提交
911 912 913 914 915 916 917 918 919
	/*
	 * 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);
920 921 922 923 924
	mutex_unlock(&shmem_swaplist_mutex);

	error = 1;
	if (!inode)
		goto out;
925 926
	/* Precharge page while we can wait, compensate afterwards */
	error = mem_cgroup_cache_charge(page, current->mm, GFP_KERNEL);
H
Hugh Dickins 已提交
927 928
	if (error)
		goto out;
929 930 931
	error = radix_tree_preload(GFP_KERNEL);
	if (error)
		goto uncharge;
H
Hugh Dickins 已提交
932
	error = 1;
933 934 935 936 937

	spin_lock(&info->lock);
	ptr = shmem_swp_entry(info, idx, NULL);
	if (ptr && ptr->val == entry.val)
		error = add_to_page_cache(page, inode->i_mapping,
H
Hugh Dickins 已提交
938
						idx, GFP_NOWAIT);
939 940
	if (error == -EEXIST) {
		struct page *filepage = find_get_page(inode->i_mapping, idx);
941
		error = 1;
942 943 944 945 946 947 948 949 950 951 952
		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) {
953 954
		delete_from_swap_cache(page);
		set_page_dirty(page);
L
Linus Torvalds 已提交
955
		info->flags |= SHMEM_PAGEIN;
956 957 958
		shmem_swp_set(info, ptr, 0);
		swap_free(entry);
		error = 1;	/* not an error, but entry was found */
L
Linus Torvalds 已提交
959
	}
960 961
	if (ptr)
		shmem_swp_unmap(ptr);
L
Linus Torvalds 已提交
962
	spin_unlock(&info->lock);
H
Hugh Dickins 已提交
963
	radix_tree_preload_end();
964 965
uncharge:
	mem_cgroup_uncharge_page(page);
966 967 968 969 970
out:
	unlock_page(page);
	page_cache_release(page);
	iput(inode);		/* allows for NULL */
	return error;
L
Linus Torvalds 已提交
971 972 973 974 975 976 977 978 979 980 981
}

/*
 * 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;

982
	mutex_lock(&shmem_swaplist_mutex);
L
Linus Torvalds 已提交
983 984
	list_for_each_safe(p, next, &shmem_swaplist) {
		info = list_entry(p, struct shmem_inode_info, swaplist);
H
Hugh Dickins 已提交
985
		found = shmem_unuse_inode(info, entry, page);
986
		cond_resched();
987 988
		if (found)
			goto out;
L
Linus Torvalds 已提交
989
	}
990
	mutex_unlock(&shmem_swaplist_mutex);
991
out:	return found;	/* 0 or 1 or -ENOMEM */
L
Linus Torvalds 已提交
992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011
}

/*
 * 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;
1012
	if (!total_swap_pages)
L
Linus Torvalds 已提交
1013 1014
		goto redirty;

1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027
	/*
	 * shmem_backing_dev_info's capabilities prevent regular writeback or
	 * sync from ever calling shmem_writepage; but a stacking filesystem
	 * may use the ->writepage of its underlying filesystem, in which case
	 * tmpfs should write out to swap only in response to memory pressure,
	 * and not for pdflush or sync.  However, in those cases, we do still
	 * want to check if there's a redundant swappage to be discarded.
	 */
	if (wbc->for_reclaim)
		swap = get_swap_page();
	else
		swap.val = 0;

L
Linus Torvalds 已提交
1028 1029 1030 1031 1032 1033
	spin_lock(&info->lock);
	if (index >= info->next_index) {
		BUG_ON(!(info->flags & SHMEM_TRUNCATE));
		goto unlock;
	}
	entry = shmem_swp_entry(info, index, NULL);
1034 1035 1036 1037 1038 1039 1040 1041 1042
	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
1043

1044
	if (swap.val && add_to_swap_cache(page, swap, GFP_ATOMIC) == 0) {
1045
		remove_from_page_cache(page);
L
Linus Torvalds 已提交
1046 1047
		shmem_swp_set(info, entry, swap.val);
		shmem_swp_unmap(entry);
H
Hugh Dickins 已提交
1048 1049 1050 1051
		if (list_empty(&info->swaplist))
			inode = igrab(inode);
		else
			inode = NULL;
L
Linus Torvalds 已提交
1052
		spin_unlock(&info->lock);
1053
		swap_duplicate(swap);
1054
		BUG_ON(page_mapped(page));
1055 1056
		page_cache_release(page);	/* pagecache ref */
		set_page_dirty(page);
L
Linus Torvalds 已提交
1057
		unlock_page(page);
H
Hugh Dickins 已提交
1058 1059 1060 1061 1062 1063 1064
		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
1065 1066 1067 1068 1069 1070 1071 1072 1073
		return 0;
	}

	shmem_swp_unmap(entry);
unlock:
	spin_unlock(&info->lock);
	swap_free(swap);
redirty:
	set_page_dirty(page);
1074 1075 1076 1077
	if (wbc->for_reclaim)
		return AOP_WRITEPAGE_ACTIVATE;	/* Return with page locked */
	unlock_page(page);
	return 0;
L
Linus Torvalds 已提交
1078 1079 1080
}

#ifdef CONFIG_NUMA
1081
#ifdef CONFIG_TMPFS
1082
static int shmem_parse_mpol(char *value, unsigned short *policy,
1083
			unsigned short *mode_flags, nodemask_t *policy_nodes)
1084 1085
{
	char *nodelist = strchr(value, ':');
1086
	char *flags = strchr(value, '=');
1087 1088 1089 1090 1091 1092 1093
	int err = 1;

	if (nodelist) {
		/* NUL-terminate policy string */
		*nodelist++ = '\0';
		if (nodelist_parse(nodelist, *policy_nodes))
			goto out;
1094
		if (!nodes_subset(*policy_nodes, node_states[N_HIGH_MEMORY]))
1095
			goto out;
1096
	}
1097 1098
	if (flags)
		*flags++ = '\0';
1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120
	if (!strcmp(value, "default")) {
		*policy = MPOL_DEFAULT;
		/* Don't allow a nodelist */
		if (!nodelist)
			err = 0;
	} else if (!strcmp(value, "prefer")) {
		*policy = MPOL_PREFERRED;
		/* Insist on a nodelist of one node only */
		if (nodelist) {
			char *rest = nodelist;
			while (isdigit(*rest))
				rest++;
			if (!*rest)
				err = 0;
		}
	} else if (!strcmp(value, "bind")) {
		*policy = MPOL_BIND;
		/* Insist on a nodelist */
		if (nodelist)
			err = 0;
	} else if (!strcmp(value, "interleave")) {
		*policy = MPOL_INTERLEAVE;
1121 1122 1123
		/*
		 * Default to online nodes with memory if no nodelist
		 */
1124
		if (!nodelist)
1125
			*policy_nodes = node_states[N_HIGH_MEMORY];
1126 1127
		err = 0;
	}
1128 1129

	*mode_flags = 0;
1130
	if (flags) {
1131 1132 1133 1134
		/*
		 * Currently, we only support two mutually exclusive
		 * mode flags.
		 */
1135 1136
		if (!strcmp(flags, "static"))
			*mode_flags |= MPOL_F_STATIC_NODES;
1137
		else if (!strcmp(flags, "relative"))
1138
			*mode_flags |= MPOL_F_RELATIVE_NODES;
1139 1140
		else
			err = 1;	/* unrecognized flag */
1141
	}
1142 1143 1144 1145
out:
	/* Restore string for error message */
	if (nodelist)
		*--nodelist = ':';
1146 1147
	if (flags)
		*--flags = '=';
1148 1149 1150
	return err;
}

1151
static void shmem_show_mpol(struct seq_file *seq, unsigned short policy,
1152
			unsigned short flags, const nodemask_t policy_nodes)
1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186
{
	char *policy_string;

	switch (policy) {
	case MPOL_PREFERRED:
		policy_string = "prefer";
		break;
	case MPOL_BIND:
		policy_string = "bind";
		break;
	case MPOL_INTERLEAVE:
		policy_string = "interleave";
		break;
	default:
		/* MPOL_DEFAULT */
		return;
	}

	seq_printf(seq, ",mpol=%s", policy_string);

	if (policy != MPOL_INTERLEAVE ||
	    !nodes_equal(policy_nodes, node_states[N_HIGH_MEMORY])) {
		char buffer[64];
		int len;

		len = nodelist_scnprintf(buffer, sizeof(buffer), policy_nodes);
		if (len < sizeof(buffer))
			seq_printf(seq, ":%s", buffer);
		else
			seq_printf(seq, ":?");
	}
}
#endif /* CONFIG_TMPFS */

1187 1188
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
1189 1190
{
	struct vm_area_struct pvma;
1191
	struct page *page;
L
Linus Torvalds 已提交
1192 1193

	/* Create a pseudo vma that just contains the policy */
1194
	pvma.vm_start = 0;
L
Linus Torvalds 已提交
1195
	pvma.vm_pgoff = idx;
1196 1197
	pvma.vm_ops = NULL;
	pvma.vm_policy = mpol_shared_policy_lookup(&info->policy, idx);
1198
	page = swapin_readahead(entry, gfp, &pvma, 0);
L
Linus Torvalds 已提交
1199 1200 1201 1202
	mpol_free(pvma.vm_policy);
	return page;
}

1203 1204
static struct page *shmem_alloc_page(gfp_t gfp,
			struct shmem_inode_info *info, unsigned long idx)
L
Linus Torvalds 已提交
1205 1206 1207 1208
{
	struct vm_area_struct pvma;
	struct page *page;

1209 1210
	/* Create a pseudo vma that just contains the policy */
	pvma.vm_start = 0;
L
Linus Torvalds 已提交
1211
	pvma.vm_pgoff = idx;
1212 1213
	pvma.vm_ops = NULL;
	pvma.vm_policy = mpol_shared_policy_lookup(&info->policy, idx);
1214
	page = alloc_page_vma(gfp, &pvma, 0);
L
Linus Torvalds 已提交
1215 1216 1217
	mpol_free(pvma.vm_policy);
	return page;
}
1218 1219
#else /* !CONFIG_NUMA */
#ifdef CONFIG_TMPFS
1220
static inline int shmem_parse_mpol(char *value, unsigned short *policy,
1221
			unsigned short *mode_flags, nodemask_t *policy_nodes)
1222 1223 1224 1225
{
	return 1;
}

1226
static inline void shmem_show_mpol(struct seq_file *seq, unsigned short policy,
1227
			unsigned short flags, const nodemask_t policy_nodes)
1228 1229 1230 1231
{
}
#endif /* CONFIG_TMPFS */

1232 1233
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
1234
{
1235
	return swapin_readahead(entry, gfp, NULL, 0);
L
Linus Torvalds 已提交
1236 1237
}

1238 1239
static inline struct page *shmem_alloc_page(gfp_t gfp,
			struct shmem_inode_info *info, unsigned long idx)
L
Linus Torvalds 已提交
1240
{
1241
	return alloc_page(gfp);
L
Linus Torvalds 已提交
1242
}
1243
#endif /* CONFIG_NUMA */
L
Linus Torvalds 已提交
1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261

/*
 * 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;
1262
	gfp_t gfp;
L
Linus Torvalds 已提交
1263 1264 1265 1266
	int error;

	if (idx >= SHMEM_MAX_INDEX)
		return -EFBIG;
1267 1268

	if (type)
N
Nick Piggin 已提交
1269
		*type = 0;
1270

L
Linus Torvalds 已提交
1271 1272 1273 1274
	/*
	 * 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
1275
	 * But shmem_readpage (required for splice) passes in a locked
1276 1277
	 * 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
1278 1279 1280 1281 1282 1283 1284
	 */
repeat:
	if (!filepage)
		filepage = find_lock_page(mapping, idx);
	if (filepage && PageUptodate(filepage))
		goto done;
	error = 0;
1285
	gfp = mapping_gfp_mask(mapping);
H
Hugh Dickins 已提交
1286 1287 1288 1289 1290 1291 1292 1293 1294 1295
	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
1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312

	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
1313
			if (type && !(*type & VM_FAULT_MAJOR)) {
1314
				__count_vm_event(PGMAJFAULT);
N
Nick Piggin 已提交
1315
				*type |= VM_FAULT_MAJOR;
L
Linus Torvalds 已提交
1316
			}
1317
			spin_unlock(&info->lock);
1318
			swappage = shmem_swapin(swap, gfp, info, idx);
L
Linus Torvalds 已提交
1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375
			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 */
		if (TestSetPageLocked(swappage)) {
			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);
1376
		} else if (!(error = add_to_page_cache(
H
Hugh Dickins 已提交
1377
				swappage, mapping, idx, GFP_NOWAIT))) {
L
Linus Torvalds 已提交
1378 1379 1380
			info->flags |= SHMEM_PAGEIN;
			shmem_swp_set(info, entry, 0);
			shmem_swp_unmap(entry);
1381
			delete_from_swap_cache(swappage);
L
Linus Torvalds 已提交
1382 1383
			spin_unlock(&info->lock);
			filepage = swappage;
1384
			set_page_dirty(filepage);
L
Linus Torvalds 已提交
1385 1386 1387 1388 1389
			swap_free(swap);
		} else {
			shmem_swp_unmap(entry);
			spin_unlock(&info->lock);
			unlock_page(swappage);
1390 1391
			if (error == -ENOMEM) {
				/* allow reclaim from this memory cgroup */
1392
				error = mem_cgroup_cache_charge(swappage,
1393
					current->mm, gfp & ~__GFP_HIGHMEM);
1394 1395
				if (error) {
					page_cache_release(swappage);
1396
					goto failed;
1397 1398
				}
				mem_cgroup_uncharge_page(swappage);
1399
			}
1400
			page_cache_release(swappage);
L
Linus Torvalds 已提交
1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417
			goto repeat;
		}
	} else if (sgp == SGP_READ && !filepage) {
		shmem_swp_unmap(entry);
		filepage = find_get_page(mapping, idx);
		if (filepage &&
		    (!PageUptodate(filepage) || TestSetPageLocked(filepage))) {
			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);
1418
		if (sbinfo->max_blocks) {
L
Linus Torvalds 已提交
1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437
			spin_lock(&sbinfo->stat_lock);
			if (sbinfo->free_blocks == 0 ||
			    shmem_acct_block(info->flags)) {
				spin_unlock(&sbinfo->stat_lock);
				spin_unlock(&info->lock);
				error = -ENOSPC;
				goto failed;
			}
			sbinfo->free_blocks--;
			inode->i_blocks += BLOCKS_PER_PAGE;
			spin_unlock(&sbinfo->stat_lock);
		} else if (shmem_acct_block(info->flags)) {
			spin_unlock(&info->lock);
			error = -ENOSPC;
			goto failed;
		}

		if (!filepage) {
			spin_unlock(&info->lock);
1438
			filepage = shmem_alloc_page(gfp, info, idx);
L
Linus Torvalds 已提交
1439 1440 1441 1442 1443 1444 1445
			if (!filepage) {
				shmem_unacct_blocks(info->flags, 1);
				shmem_free_blocks(inode, 1);
				error = -ENOMEM;
				goto failed;
			}

1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456
			/* Precharge page while we can wait, compensate after */
			error = mem_cgroup_cache_charge(filepage, current->mm,
							gfp & ~__GFP_HIGHMEM);
			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
1457 1458 1459 1460 1461 1462 1463 1464 1465
			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);
			}
			if (error || swap.val || 0 != add_to_page_cache_lru(
H
Hugh Dickins 已提交
1466
					filepage, mapping, idx, GFP_NOWAIT)) {
L
Linus Torvalds 已提交
1467
				spin_unlock(&info->lock);
1468
				mem_cgroup_uncharge_page(filepage);
L
Linus Torvalds 已提交
1469 1470 1471 1472 1473 1474 1475 1476
				page_cache_release(filepage);
				shmem_unacct_blocks(info->flags, 1);
				shmem_free_blocks(inode, 1);
				filepage = NULL;
				if (error)
					goto failed;
				goto repeat;
			}
1477
			mem_cgroup_uncharge_page(filepage);
L
Linus Torvalds 已提交
1478 1479 1480 1481 1482
			info->flags |= SHMEM_PAGEIN;
		}

		info->alloced++;
		spin_unlock(&info->lock);
1483
		clear_highpage(filepage);
L
Linus Torvalds 已提交
1484 1485
		flush_dcache_page(filepage);
		SetPageUptodate(filepage);
1486 1487
		if (sgp == SGP_DIRTY)
			set_page_dirty(filepage);
L
Linus Torvalds 已提交
1488 1489
	}
done:
H
Hugh Dickins 已提交
1490
	*pagep = filepage;
L
Linus Torvalds 已提交
1491 1492 1493 1494 1495 1496 1497 1498 1499 1500
	return 0;

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

N
Nick Piggin 已提交
1501
static int shmem_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
L
Linus Torvalds 已提交
1502
{
1503
	struct inode *inode = vma->vm_file->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
1504
	int error;
N
Nick Piggin 已提交
1505
	int ret;
L
Linus Torvalds 已提交
1506

N
Nick Piggin 已提交
1507 1508
	if (((loff_t)vmf->pgoff << PAGE_CACHE_SHIFT) >= i_size_read(inode))
		return VM_FAULT_SIGBUS;
1509

1510
	error = shmem_getpage(inode, vmf->pgoff, &vmf->page, SGP_CACHE, &ret);
N
Nick Piggin 已提交
1511 1512
	if (error)
		return ((error == -ENOMEM) ? VM_FAULT_OOM : VM_FAULT_SIGBUS);
L
Linus Torvalds 已提交
1513

N
Nick Piggin 已提交
1514
	mark_page_accessed(vmf->page);
N
Nick Piggin 已提交
1515
	return ret | VM_FAULT_LOCKED;
L
Linus Torvalds 已提交
1516 1517 1518
}

#ifdef CONFIG_NUMA
A
Adrian Bunk 已提交
1519
static int shmem_set_policy(struct vm_area_struct *vma, struct mempolicy *new)
L
Linus Torvalds 已提交
1520
{
1521
	struct inode *i = vma->vm_file->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
1522 1523 1524
	return mpol_set_shared_policy(&SHMEM_I(i)->policy, vma, new);
}

A
Adrian Bunk 已提交
1525 1526
static struct mempolicy *shmem_get_policy(struct vm_area_struct *vma,
					  unsigned long addr)
L
Linus Torvalds 已提交
1527
{
1528
	struct inode *i = vma->vm_file->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
1529 1530 1531 1532 1533 1534 1535 1536 1537
	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)
{
1538
	struct inode *inode = file->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557
	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;
	}
	if (!lock && (info->flags & VM_LOCKED) && user) {
		user_shm_unlock(inode->i_size, user);
		info->flags &= ~VM_LOCKED;
	}
	retval = 0;
out_nomem:
	spin_unlock(&info->lock);
	return retval;
}

1558
static int shmem_mmap(struct file *file, struct vm_area_struct *vma)
L
Linus Torvalds 已提交
1559 1560 1561
{
	file_accessed(file);
	vma->vm_ops = &shmem_vm_ops;
N
Nick Piggin 已提交
1562
	vma->vm_flags |= VM_CAN_NONLINEAR;
L
Linus Torvalds 已提交
1563 1564 1565 1566 1567 1568 1569 1570 1571 1572
	return 0;
}

static struct inode *
shmem_get_inode(struct super_block *sb, int mode, dev_t dev)
{
	struct inode *inode;
	struct shmem_inode_info *info;
	struct shmem_sb_info *sbinfo = SHMEM_SB(sb);

1573 1574
	if (shmem_reserve_inode(sb))
		return NULL;
L
Linus Torvalds 已提交
1575 1576 1577 1578 1579 1580 1581 1582 1583 1584

	inode = new_inode(sb);
	if (inode) {
		inode->i_mode = mode;
		inode->i_uid = current->fsuid;
		inode->i_gid = current->fsgid;
		inode->i_blocks = 0;
		inode->i_mapping->a_ops = &shmem_aops;
		inode->i_mapping->backing_dev_info = &shmem_backing_dev_info;
		inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
1585
		inode->i_generation = get_seconds();
L
Linus Torvalds 已提交
1586 1587 1588 1589 1590 1591 1592
		info = SHMEM_I(inode);
		memset(info, 0, (char *)inode - (char *)info);
		spin_lock_init(&info->lock);
		INIT_LIST_HEAD(&info->swaplist);

		switch (mode & S_IFMT) {
		default:
1593
			inode->i_op = &shmem_special_inode_operations;
L
Linus Torvalds 已提交
1594 1595 1596 1597 1598
			init_special_inode(inode, mode, dev);
			break;
		case S_IFREG:
			inode->i_op = &shmem_inode_operations;
			inode->i_fop = &shmem_file_operations;
1599
			mpol_shared_policy_init(&info->policy, sbinfo->policy,
1600
					sbinfo->flags, &sbinfo->policy_nodes);
L
Linus Torvalds 已提交
1601 1602
			break;
		case S_IFDIR:
1603
			inc_nlink(inode);
L
Linus Torvalds 已提交
1604 1605 1606 1607 1608 1609 1610 1611 1612 1613
			/* 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.
			 */
1614
			mpol_shared_policy_init(&info->policy, MPOL_DEFAULT, 0,
1615
						NULL);
L
Linus Torvalds 已提交
1616 1617
			break;
		}
1618 1619
	} else
		shmem_free_inode(sb);
L
Linus Torvalds 已提交
1620 1621 1622 1623
	return inode;
}

#ifdef CONFIG_TMPFS
1624 1625
static const struct inode_operations shmem_symlink_inode_operations;
static const struct inode_operations shmem_symlink_inline_operations;
L
Linus Torvalds 已提交
1626 1627

/*
N
Nick Piggin 已提交
1628
 * Normally tmpfs avoids the use of shmem_readpage and shmem_write_begin;
1629 1630
 * 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
1631
 */
1632 1633 1634 1635 1636 1637 1638 1639
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
1640
static int
N
Nick Piggin 已提交
1641 1642 1643
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
1644
{
N
Nick Piggin 已提交
1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657
	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
1658 1659 1660 1661
	if (pos + copied > inode->i_size)
		i_size_write(inode, pos + copied);

	unlock_page(page);
N
Nick Piggin 已提交
1662 1663 1664 1665
	set_page_dirty(page);
	page_cache_release(page);

	return copied;
L
Linus Torvalds 已提交
1666 1667 1668 1669
}

static void do_shmem_file_read(struct file *filp, loff_t *ppos, read_descriptor_t *desc, read_actor_t actor)
{
1670
	struct inode *inode = filp->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
1671 1672
	struct address_space *mapping = inode->i_mapping;
	unsigned long index, offset;
1673 1674 1675 1676 1677 1678 1679 1680 1681
	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
1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699

	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;
		}

1700
		desc->error = shmem_getpage(inode, index, &page, sgp, NULL);
L
Linus Torvalds 已提交
1701 1702 1703 1704 1705
		if (desc->error) {
			if (desc->error == -EINVAL)
				desc->error = 0;
			break;
		}
H
Hugh Dickins 已提交
1706 1707
		if (page)
			unlock_page(page);
L
Linus Torvalds 已提交
1708 1709 1710

		/*
		 * We must evaluate after, since reads (unlike writes)
1711
		 * are called without i_mutex protection against truncate
L
Linus Torvalds 已提交
1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738
		 */
		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
1739
		} else {
L
Linus Torvalds 已提交
1740
			page = ZERO_PAGE(0);
N
Nick Piggin 已提交
1741 1742
			page_cache_get(page);
		}
L
Linus Torvalds 已提交
1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791

		/*
		 * 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);
}

static ssize_t shmem_file_read(struct file *filp, char __user *buf, size_t count, loff_t *ppos)
{
	read_descriptor_t desc;

	if ((ssize_t) count < 0)
		return -EINVAL;
	if (!access_ok(VERIFY_WRITE, buf, count))
		return -EFAULT;
	if (!count)
		return 0;

	desc.written = 0;
	desc.count = count;
	desc.arg.buf = buf;
	desc.error = 0;

	do_shmem_file_read(filp, ppos, &desc, file_read_actor);
	if (desc.written)
		return desc.written;
	return desc.error;
}

1792
static int shmem_statfs(struct dentry *dentry, struct kstatfs *buf)
L
Linus Torvalds 已提交
1793
{
1794
	struct shmem_sb_info *sbinfo = SHMEM_SB(dentry->d_sb);
L
Linus Torvalds 已提交
1795 1796 1797 1798

	buf->f_type = TMPFS_MAGIC;
	buf->f_bsize = PAGE_CACHE_SIZE;
	buf->f_namelen = NAME_MAX;
1799 1800
	spin_lock(&sbinfo->stat_lock);
	if (sbinfo->max_blocks) {
L
Linus Torvalds 已提交
1801 1802
		buf->f_blocks = sbinfo->max_blocks;
		buf->f_bavail = buf->f_bfree = sbinfo->free_blocks;
1803 1804
	}
	if (sbinfo->max_inodes) {
L
Linus Torvalds 已提交
1805 1806 1807 1808
		buf->f_files = sbinfo->max_inodes;
		buf->f_ffree = sbinfo->free_inodes;
	}
	/* else leave those fields 0 like simple_statfs */
1809
	spin_unlock(&sbinfo->stat_lock);
L
Linus Torvalds 已提交
1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822
	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)
{
	struct inode *inode = shmem_get_inode(dir->i_sb, mode, dev);
	int error = -ENOSPC;

	if (inode) {
1823 1824 1825 1826 1827 1828 1829
		error = security_inode_init_security(inode, dir, NULL, NULL,
						     NULL);
		if (error) {
			if (error != -EOPNOTSUPP) {
				iput(inode);
				return error;
			}
1830 1831 1832 1833 1834
		}
		error = shmem_acl_init(inode, dir);
		if (error) {
			iput(inode);
			return error;
1835
		}
L
Linus Torvalds 已提交
1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854
		if (dir->i_mode & S_ISGID) {
			inode->i_gid = dir->i_gid;
			if (S_ISDIR(mode))
				inode->i_mode |= S_ISGID;
		}
		dir->i_size += BOGO_DIRENT_SIZE;
		dir->i_ctime = dir->i_mtime = CURRENT_TIME;
		d_instantiate(dentry, inode);
		dget(dentry); /* Extra count - pin the dentry in core */
	}
	return error;
}

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

	if ((error = shmem_mknod(dir, dentry, mode | S_IFDIR, 0)))
		return error;
1855
	inc_nlink(dir);
L
Linus Torvalds 已提交
1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870
	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;
1871
	int ret;
L
Linus Torvalds 已提交
1872 1873 1874 1875 1876 1877

	/*
	 * 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.
	 */
1878 1879 1880
	ret = shmem_reserve_inode(inode->i_sb);
	if (ret)
		goto out;
L
Linus Torvalds 已提交
1881 1882 1883

	dir->i_size += BOGO_DIRENT_SIZE;
	inode->i_ctime = dir->i_ctime = dir->i_mtime = CURRENT_TIME;
1884
	inc_nlink(inode);
L
Linus Torvalds 已提交
1885 1886 1887
	atomic_inc(&inode->i_count);	/* New dentry reference */
	dget(dentry);		/* Extra pinning count for the created dentry */
	d_instantiate(dentry, inode);
1888 1889
out:
	return ret;
L
Linus Torvalds 已提交
1890 1891 1892 1893 1894 1895
}

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

1896 1897
	if (inode->i_nlink > 1 && !S_ISDIR(inode->i_mode))
		shmem_free_inode(inode->i_sb);
L
Linus Torvalds 已提交
1898 1899 1900

	dir->i_size -= BOGO_DIRENT_SIZE;
	inode->i_ctime = dir->i_ctime = dir->i_mtime = CURRENT_TIME;
1901
	drop_nlink(inode);
L
Linus Torvalds 已提交
1902 1903 1904 1905 1906 1907 1908 1909 1910
	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;

1911 1912
	drop_nlink(dentry->d_inode);
	drop_nlink(dir);
L
Linus Torvalds 已提交
1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932
	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)
1933
			drop_nlink(old_dir);
L
Linus Torvalds 已提交
1934
	} else if (they_are_dirs) {
1935
		drop_nlink(old_dir);
1936
		inc_nlink(new_dir);
L
Linus Torvalds 已提交
1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963
	}

	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;

	inode = shmem_get_inode(dir->i_sb, S_IFLNK|S_IRWXUGO, 0);
	if (!inode)
		return -ENOSPC;

1964 1965 1966 1967 1968 1969 1970 1971 1972 1973
	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
1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985
	info = SHMEM_I(inode);
	inode->i_size = len-1;
	if (len <= (char *)inode - (char *)info) {
		/* do it inline */
		memcpy(info, symname, len);
		inode->i_op = &shmem_symlink_inline_operations;
	} else {
		error = shmem_getpage(inode, 0, &page, SGP_WRITE, NULL);
		if (error) {
			iput(inode);
			return error;
		}
H
Hugh Dickins 已提交
1986
		unlock_page(page);
L
Linus Torvalds 已提交
1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
		inode->i_op = &shmem_symlink_inode_operations;
		kaddr = kmap_atomic(page, KM_USER0);
		memcpy(kaddr, symname, len);
		kunmap_atomic(kaddr, KM_USER0);
		set_page_dirty(page);
		page_cache_release(page);
	}
	if (dir->i_mode & S_ISGID)
		inode->i_gid = dir->i_gid;
	dir->i_size += BOGO_DIRENT_SIZE;
	dir->i_ctime = dir->i_mtime = CURRENT_TIME;
	d_instantiate(dentry, inode);
	dget(dentry);
	return 0;
}

2003
static void *shmem_follow_link_inline(struct dentry *dentry, struct nameidata *nd)
L
Linus Torvalds 已提交
2004 2005
{
	nd_set_link(nd, (char *)SHMEM_I(dentry->d_inode));
2006
	return NULL;
L
Linus Torvalds 已提交
2007 2008
}

2009
static void *shmem_follow_link(struct dentry *dentry, struct nameidata *nd)
L
Linus Torvalds 已提交
2010 2011 2012 2013
{
	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
2014 2015
	if (page)
		unlock_page(page);
2016
	return page;
L
Linus Torvalds 已提交
2017 2018
}

2019
static void shmem_put_link(struct dentry *dentry, struct nameidata *nd, void *cookie)
L
Linus Torvalds 已提交
2020 2021
{
	if (!IS_ERR(nd_get_link(nd))) {
2022
		struct page *page = cookie;
L
Linus Torvalds 已提交
2023 2024 2025 2026 2027 2028
		kunmap(page);
		mark_page_accessed(page);
		page_cache_release(page);
	}
}

2029
static const struct inode_operations shmem_symlink_inline_operations = {
L
Linus Torvalds 已提交
2030 2031 2032 2033
	.readlink	= generic_readlink,
	.follow_link	= shmem_follow_link_inline,
};

2034
static const struct inode_operations shmem_symlink_inode_operations = {
L
Linus Torvalds 已提交
2035 2036 2037 2038 2039 2040
	.truncate	= shmem_truncate,
	.readlink	= generic_readlink,
	.follow_link	= shmem_follow_link,
	.put_link	= shmem_put_link,
};

2041
#ifdef CONFIG_TMPFS_POSIX_ACL
2042
/*
2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060
 * Superblocks without xattr inode operations will get security.* xattr
 * support from the VFS "for free". As soon as we have any other xattrs
 * like ACLs, we also need to implement the security.* handlers at
 * filesystem level, though.
 */

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

static int shmem_xattr_security_get(struct inode *inode, const char *name,
				    void *buffer, size_t size)
{
	if (strcmp(name, "") == 0)
		return -EINVAL;
2061
	return xattr_getsecurity(inode, name, buffer, size);
2062 2063 2064 2065 2066 2067 2068 2069 2070 2071
}

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

2072
static struct xattr_handler shmem_xattr_security_handler = {
2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086
	.prefix = XATTR_SECURITY_PREFIX,
	.list   = shmem_xattr_security_list,
	.get    = shmem_xattr_security_get,
	.set    = shmem_xattr_security_set,
};

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

2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099
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
2100 2101
static struct dentry *shmem_fh_to_dentry(struct super_block *sb,
		struct fid *fid, int fh_len, int fh_type)
2102 2103
{
	struct inode *inode;
C
Christoph Hellwig 已提交
2104 2105 2106 2107 2108 2109
	struct dentry *dentry = NULL;
	u64 inum = fid->raw[2];
	inum = (inum << 32) | fid->raw[1];

	if (fh_len < 3)
		return NULL;
2110

C
Christoph Hellwig 已提交
2111 2112
	inode = ilookup5(sb, (unsigned long)(inum + fid->raw[0]),
			shmem_match, fid->raw);
2113
	if (inode) {
C
Christoph Hellwig 已提交
2114
		dentry = d_find_alias(inode);
2115 2116 2117
		iput(inode);
	}

C
Christoph Hellwig 已提交
2118
	return dentry;
2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150
}

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;
}

2151
static const struct export_operations shmem_export_ops = {
2152 2153
	.get_parent     = shmem_get_parent,
	.encode_fh      = shmem_encode_fh,
C
Christoph Hellwig 已提交
2154
	.fh_to_dentry	= shmem_fh_to_dentry,
2155 2156
};

2157 2158
static int shmem_parse_options(char *options, struct shmem_sb_info *sbinfo,
			       bool remount)
L
Linus Torvalds 已提交
2159 2160 2161
{
	char *this_char, *value, *rest;

2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178
	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
2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200
		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;
2201 2202
			sbinfo->max_blocks =
				DIV_ROUND_UP(size, PAGE_CACHE_SIZE);
L
Linus Torvalds 已提交
2203
		} else if (!strcmp(this_char,"nr_blocks")) {
2204
			sbinfo->max_blocks = memparse(value, &rest);
L
Linus Torvalds 已提交
2205 2206 2207
			if (*rest)
				goto bad_val;
		} else if (!strcmp(this_char,"nr_inodes")) {
2208
			sbinfo->max_inodes = memparse(value, &rest);
L
Linus Torvalds 已提交
2209 2210 2211
			if (*rest)
				goto bad_val;
		} else if (!strcmp(this_char,"mode")) {
2212
			if (remount)
L
Linus Torvalds 已提交
2213
				continue;
2214
			sbinfo->mode = simple_strtoul(value, &rest, 8) & 07777;
L
Linus Torvalds 已提交
2215 2216 2217
			if (*rest)
				goto bad_val;
		} else if (!strcmp(this_char,"uid")) {
2218
			if (remount)
L
Linus Torvalds 已提交
2219
				continue;
2220
			sbinfo->uid = simple_strtoul(value, &rest, 0);
L
Linus Torvalds 已提交
2221 2222 2223
			if (*rest)
				goto bad_val;
		} else if (!strcmp(this_char,"gid")) {
2224
			if (remount)
L
Linus Torvalds 已提交
2225
				continue;
2226
			sbinfo->gid = simple_strtoul(value, &rest, 0);
L
Linus Torvalds 已提交
2227 2228
			if (*rest)
				goto bad_val;
2229
		} else if (!strcmp(this_char,"mpol")) {
2230
			if (shmem_parse_mpol(value, &sbinfo->policy,
2231
				&sbinfo->flags, &sbinfo->policy_nodes))
2232
				goto bad_val;
L
Linus Torvalds 已提交
2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250
		} 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);
2251
	struct shmem_sb_info config = *sbinfo;
2252 2253 2254 2255
	unsigned long blocks;
	unsigned long inodes;
	int error = -EINVAL;

2256
	if (shmem_parse_options(data, &config, true))
2257
		return error;
L
Linus Torvalds 已提交
2258

2259 2260 2261
	spin_lock(&sbinfo->stat_lock);
	blocks = sbinfo->max_blocks - sbinfo->free_blocks;
	inodes = sbinfo->max_inodes - sbinfo->free_inodes;
2262
	if (config.max_blocks < blocks)
2263
		goto out;
2264
	if (config.max_inodes < inodes)
2265 2266 2267 2268 2269 2270 2271
		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.
	 */
2272
	if (config.max_blocks && !sbinfo->max_blocks)
2273
		goto out;
2274
	if (config.max_inodes && !sbinfo->max_inodes)
2275 2276 2277
		goto out;

	error = 0;
2278 2279 2280 2281 2282
	sbinfo->max_blocks  = config.max_blocks;
	sbinfo->free_blocks = config.max_blocks - blocks;
	sbinfo->max_inodes  = config.max_inodes;
	sbinfo->free_inodes = config.max_inodes - inodes;
	sbinfo->policy      = config.policy;
2283
	sbinfo->flags	    = config.flags;
2284
	sbinfo->policy_nodes = config.policy_nodes;
2285 2286 2287
out:
	spin_unlock(&sbinfo->stat_lock);
	return error;
L
Linus Torvalds 已提交
2288
}
2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304

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);
2305 2306
	shmem_show_mpol(seq, sbinfo->policy, sbinfo->flags,
			sbinfo->policy_nodes);
2307 2308 2309
	return 0;
}
#endif /* CONFIG_TMPFS */
L
Linus Torvalds 已提交
2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321

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

static int shmem_fill_super(struct super_block *sb,
			    void *data, int silent)
{
	struct inode *inode;
	struct dentry *root;
2322
	struct shmem_sb_info *sbinfo;
2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336
	int err = -ENOMEM;

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

	sbinfo->max_blocks = 0;
	sbinfo->max_inodes = 0;
	sbinfo->mode = S_IRWXUGO | S_ISVTX;
	sbinfo->uid = current->fsuid;
	sbinfo->gid = current->fsgid;
	sbinfo->policy = MPOL_DEFAULT;
2337
	sbinfo->flags = 0;
2338 2339
	sbinfo->policy_nodes = node_states[N_HIGH_MEMORY];
	sb->s_fs_info = sbinfo;
L
Linus Torvalds 已提交
2340

2341
#ifdef CONFIG_TMPFS
L
Linus Torvalds 已提交
2342 2343 2344 2345 2346 2347
	/*
	 * 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)) {
2348 2349 2350 2351 2352 2353
		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
2354
	}
2355
	sb->s_export_op = &shmem_export_ops;
L
Linus Torvalds 已提交
2356 2357 2358 2359
#else
	sb->s_flags |= MS_NOUSER;
#endif

2360
	spin_lock_init(&sbinfo->stat_lock);
2361 2362
	sbinfo->free_blocks = sbinfo->max_blocks;
	sbinfo->free_inodes = sbinfo->max_inodes;
2363

L
Linus Torvalds 已提交
2364 2365 2366 2367 2368
	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;
2369
	sb->s_time_gran = 1;
2370 2371 2372 2373
#ifdef CONFIG_TMPFS_POSIX_ACL
	sb->s_xattr = shmem_xattr_handlers;
	sb->s_flags |= MS_POSIXACL;
#endif
2374

2375
	inode = shmem_get_inode(sb, S_IFDIR | sbinfo->mode, 0);
L
Linus Torvalds 已提交
2376 2377
	if (!inode)
		goto failed;
2378 2379
	inode->i_uid = sbinfo->uid;
	inode->i_gid = sbinfo->gid;
L
Linus Torvalds 已提交
2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392
	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;
}

2393
static struct kmem_cache *shmem_inode_cachep;
L
Linus Torvalds 已提交
2394 2395 2396 2397

static struct inode *shmem_alloc_inode(struct super_block *sb)
{
	struct shmem_inode_info *p;
2398
	p = (struct shmem_inode_info *)kmem_cache_alloc(shmem_inode_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409
	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);
	}
2410
	shmem_acl_destroy_inode(inode);
L
Linus Torvalds 已提交
2411 2412 2413
	kmem_cache_free(shmem_inode_cachep, SHMEM_I(inode));
}

2414
static void init_once(struct kmem_cache *cachep, void *foo)
L
Linus Torvalds 已提交
2415 2416 2417
{
	struct shmem_inode_info *p = (struct shmem_inode_info *) foo;

C
Christoph Lameter 已提交
2418
	inode_init_once(&p->vfs_inode);
2419
#ifdef CONFIG_TMPFS_POSIX_ACL
C
Christoph Lameter 已提交
2420 2421
	p->i_acl = NULL;
	p->i_default_acl = NULL;
2422
#endif
L
Linus Torvalds 已提交
2423 2424 2425 2426 2427 2428
}

static int init_inodecache(void)
{
	shmem_inode_cachep = kmem_cache_create("shmem_inode_cache",
				sizeof(struct shmem_inode_info),
2429
				0, SLAB_PANIC, init_once);
L
Linus Torvalds 已提交
2430 2431 2432 2433 2434
	return 0;
}

static void destroy_inodecache(void)
{
2435
	kmem_cache_destroy(shmem_inode_cachep);
L
Linus Torvalds 已提交
2436 2437
}

2438
static const struct address_space_operations shmem_aops = {
L
Linus Torvalds 已提交
2439
	.writepage	= shmem_writepage,
2440
	.set_page_dirty	= __set_page_dirty_no_writeback,
L
Linus Torvalds 已提交
2441
#ifdef CONFIG_TMPFS
2442
	.readpage	= shmem_readpage,
N
Nick Piggin 已提交
2443 2444
	.write_begin	= shmem_write_begin,
	.write_end	= shmem_write_end,
L
Linus Torvalds 已提交
2445
#endif
2446
	.migratepage	= migrate_page,
L
Linus Torvalds 已提交
2447 2448
};

2449
static const struct file_operations shmem_file_operations = {
L
Linus Torvalds 已提交
2450 2451 2452 2453
	.mmap		= shmem_mmap,
#ifdef CONFIG_TMPFS
	.llseek		= generic_file_llseek,
	.read		= shmem_file_read,
H
Hugh Dickins 已提交
2454 2455
	.write		= do_sync_write,
	.aio_write	= generic_file_aio_write,
L
Linus Torvalds 已提交
2456
	.fsync		= simple_sync_file,
2457 2458
	.splice_read	= generic_file_splice_read,
	.splice_write	= generic_file_splice_write,
L
Linus Torvalds 已提交
2459 2460 2461
#endif
};

2462
static const struct inode_operations shmem_inode_operations = {
L
Linus Torvalds 已提交
2463 2464
	.truncate	= shmem_truncate,
	.setattr	= shmem_notify_change,
2465
	.truncate_range	= shmem_truncate_range,
2466 2467 2468 2469 2470 2471 2472 2473
#ifdef CONFIG_TMPFS_POSIX_ACL
	.setxattr	= generic_setxattr,
	.getxattr	= generic_getxattr,
	.listxattr	= generic_listxattr,
	.removexattr	= generic_removexattr,
	.permission	= shmem_permission,
#endif

L
Linus Torvalds 已提交
2474 2475
};

2476
static const struct inode_operations shmem_dir_inode_operations = {
L
Linus Torvalds 已提交
2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487
#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
2488 2489 2490 2491 2492 2493 2494 2495 2496 2497
#ifdef CONFIG_TMPFS_POSIX_ACL
	.setattr	= shmem_notify_change,
	.setxattr	= generic_setxattr,
	.getxattr	= generic_getxattr,
	.listxattr	= generic_listxattr,
	.removexattr	= generic_removexattr,
	.permission	= shmem_permission,
#endif
};

2498
static const struct inode_operations shmem_special_inode_operations = {
2499 2500 2501 2502 2503 2504 2505 2506
#ifdef CONFIG_TMPFS_POSIX_ACL
	.setattr	= shmem_notify_change,
	.setxattr	= generic_setxattr,
	.getxattr	= generic_getxattr,
	.listxattr	= generic_listxattr,
	.removexattr	= generic_removexattr,
	.permission	= shmem_permission,
#endif
L
Linus Torvalds 已提交
2507 2508
};

2509
static const struct super_operations shmem_ops = {
L
Linus Torvalds 已提交
2510 2511 2512 2513 2514
	.alloc_inode	= shmem_alloc_inode,
	.destroy_inode	= shmem_destroy_inode,
#ifdef CONFIG_TMPFS
	.statfs		= shmem_statfs,
	.remount_fs	= shmem_remount_fs,
2515
	.show_options	= shmem_show_options,
L
Linus Torvalds 已提交
2516 2517 2518 2519 2520 2521 2522
#endif
	.delete_inode	= shmem_delete_inode,
	.drop_inode	= generic_delete_inode,
	.put_super	= shmem_put_super,
};

static struct vm_operations_struct shmem_vm_ops = {
2523
	.fault		= shmem_fault,
L
Linus Torvalds 已提交
2524 2525 2526 2527 2528 2529 2530
#ifdef CONFIG_NUMA
	.set_policy     = shmem_set_policy,
	.get_policy     = shmem_get_policy,
#endif
};


2531 2532
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
2533
{
2534
	return get_sb_nodev(fs_type, flags, data, shmem_fill_super, mnt);
L
Linus Torvalds 已提交
2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548
}

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

static int __init init_tmpfs(void)
{
	int error;

P
Peter Zijlstra 已提交
2549 2550 2551 2552
	error = bdi_init(&shmem_backing_dev_info);
	if (error)
		goto out4;

L
Linus Torvalds 已提交
2553 2554 2555 2556 2557 2558 2559 2560 2561
	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;
	}
2562

2563
	shm_mnt = vfs_kern_mount(&tmpfs_fs_type, MS_NOUSER,
L
Linus Torvalds 已提交
2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576
				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
2577 2578
	bdi_destroy(&shmem_backing_dev_info);
out4:
L
Linus Torvalds 已提交
2579 2580 2581 2582 2583
	shm_mnt = ERR_PTR(error);
	return error;
}
module_init(init_tmpfs)

2584
/**
L
Linus Torvalds 已提交
2585 2586 2587
 * 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
2588
 * @flags: vm_flags
L
Linus Torvalds 已提交
2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629
 */
struct file *shmem_file_setup(char *name, loff_t size, unsigned long flags)
{
	int error;
	struct file *file;
	struct inode *inode;
	struct dentry *dentry, *root;
	struct qstr this;

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

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

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

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

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

	error = -ENOSPC;
	inode = shmem_get_inode(root->d_sb, S_IFREG | S_IRWXUGO, 0);
	if (!inode)
		goto close_file;

	SHMEM_I(inode)->flags = flags & VM_ACCOUNT;
	d_instantiate(dentry, inode);
	inode->i_size = size;
	inode->i_nlink = 0;	/* It is unlinked */
2630 2631
	init_file(file, shm_mnt, dentry, FMODE_WRITE | FMODE_READ,
			&shmem_file_operations);
L
Linus Torvalds 已提交
2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642
	return file;

close_file:
	put_filp(file);
put_dentry:
	dput(dentry);
put_memory:
	shmem_unacct_size(flags, size);
	return ERR_PTR(error);
}

2643
/**
L
Linus Torvalds 已提交
2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661
 * 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;
}