shmem.c 68.8 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
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.
Linus Torvalds's avatar
Linus Torvalds committed
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>
 *
Linus Torvalds's avatar
Linus Torvalds committed
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>
27
#include <linux/pagemap.h>
28 29 30 31
#include <linux/file.h>
#include <linux/mm.h>
#include <linux/module.h>
#include <linux/swap.h>
Mimi Zohar's avatar
Mimi Zohar committed
32
#include <linux/ima.h>
33 34 35 36

static struct vfsmount *shm_mnt;

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

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

Linus Torvalds's avatar
Linus Torvalds committed
64 65 66 67
#include <asm/uaccess.h>
#include <asm/div64.h>
#include <asm/pgtable.h>

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.
 */
Linus Torvalds's avatar
Linus Torvalds committed
80
#define ENTRIES_PER_PAGE (PAGE_CACHE_SIZE/sizeof(unsigned long))
Yuri Tikhonov's avatar
Yuri Tikhonov committed
81
#define ENTRIES_PER_PAGEPAGE ((unsigned long long)ENTRIES_PER_PAGE*ENTRIES_PER_PAGE)
Linus Torvalds's avatar
Linus Torvalds committed
82

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)
Linus Torvalds's avatar
Linus Torvalds committed
85

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)
Linus Torvalds's avatar
Linus Torvalds committed
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 */
Linus Torvalds's avatar
Linus Torvalds committed
107 108 109
	SGP_WRITE,	/* may exceed i_size, may allocate page */
};

Andrew Morton's avatar
Andrew Morton committed
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);
}
Andrew Morton's avatar
Andrew Morton committed
120
#endif
121

Linus Torvalds's avatar
Linus Torvalds committed
122 123 124
static int shmem_getpage(struct inode *inode, unsigned long idx,
			 struct page **pagep, enum sgp_type sgp, int *type);

Al Viro's avatar
Al Viro committed
125
static inline struct page *shmem_dir_alloc(gfp_t gfp_mask)
Linus Torvalds's avatar
Linus Torvalds committed
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
Linus Torvalds's avatar
Linus Torvalds committed
133
	 */
134
	return alloc_pages((gfp_mask & ~GFP_MOVABLE_MASK) | __GFP_ZERO,
135
				PAGE_CACHE_SHIFT-PAGE_SHIFT);
Linus Torvalds's avatar
Linus Torvalds committed
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));
Linus Torvalds's avatar
Linus Torvalds committed
190 191 192 193
}

static inline void shmem_unacct_size(unsigned long flags, loff_t size)
{
194
	if (!(flags & VM_NORESERVE))
Linus Torvalds's avatar
Linus Torvalds committed
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;
Linus Torvalds's avatar
Linus Torvalds committed
208 209 210 211
}

static inline void shmem_unacct_blocks(unsigned long flags, long pages)
{
212
	if (flags & VM_NORESERVE)
Linus Torvalds's avatar
Linus Torvalds committed
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;
Linus Torvalds's avatar
Linus Torvalds committed
222 223
static struct vm_operations_struct shmem_vm_ops;

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

static LIST_HEAD(shmem_swaplist);
231
static DEFINE_MUTEX(shmem_swaplist_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
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) {
Linus Torvalds's avatar
Linus Torvalds committed
237 238 239 240 241 242 243
		spin_lock(&sbinfo->stat_lock);
		sbinfo->free_blocks += pages;
		inode->i_blocks -= pages*BLOCKS_PER_PAGE;
		spin_unlock(&sbinfo->stat_lock);
	}
}

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

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

269
/**
Linus Torvalds's avatar
Linus Torvalds committed
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
/**
Linus Torvalds's avatar
Linus Torvalds committed
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;
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);
	}
Linus Torvalds's avatar
Linus Torvalds committed
395 396
}

397
/**
Linus Torvalds's avatar
Linus Torvalds committed
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.
Linus Torvalds's avatar
Linus Torvalds committed
404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423
 */
static swp_entry_t *shmem_swp_alloc(struct shmem_inode_info *info, unsigned long index, enum sgp_type sgp)
{
	struct inode *inode = &info->vfs_inode;
	struct shmem_sb_info *sbinfo = SHMEM_SB(inode->i_sb);
	struct page *page = NULL;
	swp_entry_t *entry;

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

	if (inode->i_op->truncate == shmem_truncate) {
826
		truncate_inode_pages(inode->i_mapping, 0);
Linus Torvalds's avatar
Linus Torvalds committed
827 828 829 830
		shmem_unacct_size(info->flags, inode->i_size);
		inode->i_size = 0;
		shmem_truncate(inode);
		if (!list_empty(&info->swaplist)) {
831
			mutex_lock(&shmem_swaplist_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
832
			list_del_init(&info->swaplist);
833
			mutex_unlock(&shmem_swaplist_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
834 835
		}
	}
836
	BUG_ON(inode->i_blocks);
837
	shmem_free_inode(inode->i_sb);
Linus Torvalds's avatar
Linus Torvalds committed
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;
Linus Torvalds's avatar
Linus Torvalds committed
864 865 866 867

	idx = 0;
	ptr = info->i_direct;
	spin_lock(&info->lock);
868 869 870 871
	if (!info->swapped) {
		list_del_init(&info->swaplist);
		goto lost2;
	}
Linus Torvalds's avatar
Linus Torvalds committed
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)
Linus Torvalds's avatar
Linus Torvalds committed
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;
				}
			}
Linus Torvalds's avatar
Linus Torvalds committed
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;
910
		if (subdir && page_private(subdir)) {
Linus Torvalds's avatar
Linus Torvalds committed
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);
Linus Torvalds's avatar
Linus Torvalds committed
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);

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;
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.
951
	 */
952
	error = mem_cgroup_cache_charge(page, current->mm, GFP_KERNEL);
Hugh Dickins's avatar
Hugh Dickins committed
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;
	}
Hugh Dickins's avatar
Hugh Dickins committed
960
	error = 1;
961 962 963

	spin_lock(&info->lock);
	ptr = shmem_swp_entry(info, idx, NULL);
964
	if (ptr && ptr->val == entry.val) {
965
		error = add_to_page_cache_locked(page, inode->i_mapping,
Hugh Dickins's avatar
Hugh Dickins committed
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);
Linus Torvalds's avatar
Linus Torvalds committed
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 */
Linus Torvalds's avatar
Linus Torvalds committed
991
	}
992 993
	if (ptr)
		shmem_swp_unmap(ptr);
Linus Torvalds's avatar
Linus Torvalds committed
994
	spin_unlock(&info->lock);
Hugh Dickins's avatar
Hugh Dickins committed
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;
Linus Torvalds's avatar
Linus Torvalds committed
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);
Linus Torvalds's avatar
Linus Torvalds committed
1013 1014
	list_for_each_safe(p, next, &shmem_swaplist) {
		info = list_entry(p, struct shmem_inode_info, swaplist);
1015
		found = shmem_unuse_inode(info, entry, page);
1016
		cond_resched();
1017 1018
		if (found)
			goto out;
Linus Torvalds's avatar
Linus Torvalds committed
1019
	}
1020
	mutex_unlock(&shmem_swaplist_mutex);
1021
out:	return found;	/* 0 or 1 or -ENOMEM */
Linus Torvalds's avatar
Linus Torvalds committed
1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041
}

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

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

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

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

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

	shmem_swp_unmap(entry);
unlock:
	spin_unlock(&info->lock);
1100
	swapcache_free(swap, NULL);
Linus Torvalds's avatar
Linus Torvalds committed
1101 1102
redirty:
	set_page_dirty(page);
1103 1104 1105 1106
	if (wbc->for_reclaim)
		return AOP_WRITEPAGE_ACTIVATE;	/* Return with page locked */
	unlock_page(page);
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
1107 1108 1109
}

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

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

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

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

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

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

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

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

1155 1156
static struct page *shmem_alloc_page(gfp_t gfp,
			struct shmem_inode_info *info, unsigned long idx)
Linus Torvalds's avatar
Linus Torvalds committed
1157 1158 1159
{
	struct vm_area_struct pvma;

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

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

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

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

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

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

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

	if (type)
Nick Piggin's avatar
Nick Piggin committed
1222
		*type = 0;
1223

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

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

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

		/* We have to do this with page locked to prevent races */
Nick Piggin's avatar
Nick Piggin committed
1293
		if (!trylock_page(swappage)) {
Linus Torvalds's avatar
Linus Torvalds committed
1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328
			shmem_swp_unmap(entry);
			spin_unlock(&info->lock);
			wait_on_page_locked(swappage);
			page_cache_release(swappage);
			goto repeat;
		}
		if (PageWriteback(swappage)) {
			shmem_swp_unmap(entry);
			spin_unlock(&info->lock);
			wait_on_page_writeback(swappage);
			unlock_page(swappage);
			page_cache_release(swappage);
			goto repeat;
		}
		if (!PageUptodate(swappage)) {
			shmem_swp_unmap(entry);
			spin_unlock(&info->lock);
			unlock_page(swappage);
			page_cache_release(swappage);
			error = -EIO;
			goto failed;
		}

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

		if (!filepage) {
1395 1396
			int ret;

Linus Torvalds's avatar
Linus Torvalds committed
1397
			spin_unlock(&info->lock);
1398
			filepage = shmem_alloc_page(gfp, info, idx);
Linus Torvalds's avatar
Linus Torvalds committed
1399 1400 1401 1402 1403 1404
			if (!filepage) {
				shmem_unacct_blocks(info->flags, 1);
				shmem_free_blocks(inode, 1);
				error = -ENOMEM;
				goto failed;
			}
1405
			SetPageSwapBacked(filepage);
Linus Torvalds's avatar
Linus Torvalds committed
1406

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

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

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

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

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

Nick Piggin's avatar
Nick Piggin committed
1475 1476
	if (((loff_t)vmf->pgoff << PAGE_CACHE_SHIFT) >= i_size_read(inode))
		return VM_FAULT_SIGBUS;
1477

1478
	error = shmem_getpage(inode, vmf->pgoff, &vmf->page, SGP_CACHE, &ret);
Nick Piggin's avatar
Nick Piggin committed
1479 1480
	if (error)
		return ((error == -ENOMEM) ? VM_FAULT_OOM : VM_FAULT_SIGBUS);
Linus Torvalds's avatar
Linus Torvalds committed
1481

Nick Piggin's avatar
Nick Piggin committed
1482
	return ret | VM_FAULT_LOCKED;
Linus Torvalds's avatar
Linus Torvalds committed
1483 1484 1485
}

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

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

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

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

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

Linus Torvalds's avatar
Linus Torvalds committed
1524 1525 1526 1527 1528
out_nomem:
	spin_unlock(&info->lock);
	return retval;
}

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

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

1544 1545
	if (shmem_reserve_inode(sb))
		return NULL;
Linus Torvalds's avatar
Linus Torvalds committed
1546 1547 1548 1549

	inode = new_inode(sb);
	if (inode) {
		inode->i_mode = mode;
1550 1551
		inode->i_uid = current_fsuid();
		inode->i_gid = current_fsgid();
Linus Torvalds's avatar
Linus Torvalds committed
1552 1553 1554
		inode->i_blocks = 0;
		inode->i_mapping->backing_dev_info = &shmem_backing_dev_info;
		inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
1555
		inode->i_generation = get_seconds();
Linus Torvalds's avatar
Linus Torvalds committed
1556 1557 1558
		info = SHMEM_I(inode);
		memset(info, 0, (char *)inode - (char *)info);
		spin_lock_init(&info->lock);
1559
		info->flags = flags & VM_NORESERVE;
Linus Torvalds's avatar
Linus Torvalds committed
1560
		INIT_LIST_HEAD(&info->swaplist);
1561
		cache_no_acl(inode);
Linus Torvalds's avatar
Linus Torvalds committed
1562 1563 1564

		switch (mode & S_IFMT) {
		default:
1565
			inode->i_op = &shmem_special_inode_operations;
Linus Torvalds's avatar
Linus Torvalds committed
1566 1567 1568
			init_special_inode(inode, mode, dev);
			break;
		case S_IFREG:
1569
			inode->i_mapping->a_ops = &shmem_aops;
Linus Torvalds's avatar
Linus Torvalds committed
1570 1571
			inode->i_op = &shmem_inode_operations;
			inode->i_fop = &shmem_file_operations;
1572 1573
			mpol_shared_policy_init(&info->policy,
						 shmem_get_sbmpol(sbinfo));
Linus Torvalds's avatar
Linus Torvalds committed
1574 1575
			break;
		case S_IFDIR:
1576
			inc_nlink(inode);
Linus Torvalds's avatar
Linus Torvalds committed
1577 1578 1579 1580