shmem.c 70.8 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>
Hugh Dickins's avatar
Hugh Dickins committed
27
#include <linux/pagemap.h>
28 29 30 31 32 33 34 35
#include <linux/file.h>
#include <linux/mm.h>
#include <linux/module.h>
#include <linux/swap.h>

static struct vfsmount *shm_mnt;

#ifdef CONFIG_SHMEM
Linus Torvalds's avatar
Linus Torvalds committed
36 37 38 39 40 41
/*
 * 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.
 */

42
#include <linux/xattr.h>
43
#include <linux/exportfs.h>
44
#include <linux/posix_acl.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
#include <linux/mman.h>
#include <linux/string.h>
#include <linux/slab.h>
#include <linux/backing-dev.h>
#include <linux/shmem_fs.h>
#include <linux/writeback.h>
#include <linux/blkdev.h>
#include <linux/security.h>
#include <linux/swapops.h>
#include <linux/mempolicy.h>
#include <linux/namei.h>
57
#include <linux/ctype.h>
58
#include <linux/migrate.h>
59
#include <linux/highmem.h>
60
#include <linux/seq_file.h>
Mimi Zohar's avatar
Mimi Zohar committed
61
#include <linux/magic.h>
62

Linus Torvalds's avatar
Linus Torvalds committed
63 64 65 66
#include <asm/uaccess.h>
#include <asm/div64.h>
#include <asm/pgtable.h>

Hugh Dickins's avatar
Hugh Dickins committed
67 68 69 70 71 72 73 74 75 76 77 78
/*
 * 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
79
#define ENTRIES_PER_PAGE (PAGE_CACHE_SIZE/sizeof(unsigned long))
Yuri Tikhonov's avatar
Yuri Tikhonov committed
80
#define ENTRIES_PER_PAGEPAGE ((unsigned long long)ENTRIES_PER_PAGE*ENTRIES_PER_PAGE)
Linus Torvalds's avatar
Linus Torvalds committed
81

Hugh Dickins's avatar
Hugh Dickins committed
82 83
#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
84

Hugh Dickins's avatar
Hugh Dickins committed
85 86 87 88
#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
89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104
#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 */
105
	SGP_DIRTY,	/* like SGP_CACHE, but set new page dirty */
Linus Torvalds's avatar
Linus Torvalds committed
106 107 108
	SGP_WRITE,	/* may exceed i_size, may allocate page */
};

Andrew Morton's avatar
Andrew Morton committed
109
#ifdef CONFIG_TMPFS
110 111 112 113 114 115 116 117 118
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
119
#endif
120

Linus Torvalds's avatar
Linus Torvalds committed
121 122 123
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
124
static inline struct page *shmem_dir_alloc(gfp_t gfp_mask)
Linus Torvalds's avatar
Linus Torvalds committed
125 126 127 128 129
{
	/*
	 * 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.
130
	 *
131
	 * Mobility flags are masked out as swap vectors cannot move
Linus Torvalds's avatar
Linus Torvalds committed
132
	 */
133
	return alloc_pages((gfp_mask & ~GFP_MOVABLE_MASK) | __GFP_ZERO,
134
				PAGE_CACHE_SHIFT-PAGE_SHIFT);
Linus Torvalds's avatar
Linus Torvalds committed
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
}

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)
{
187 188
	return (flags & VM_NORESERVE) ?
		0 : security_vm_enough_memory_kern(VM_ACCT(size));
Linus Torvalds's avatar
Linus Torvalds committed
189 190 191 192
}

static inline void shmem_unacct_size(unsigned long flags, loff_t size)
{
193
	if (!(flags & VM_NORESERVE))
Linus Torvalds's avatar
Linus Torvalds committed
194 195 196 197 198 199 200 201 202 203 204
		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)
{
205 206
	return (flags & VM_NORESERVE) ?
		security_vm_enough_memory_kern(VM_ACCT(PAGE_CACHE_SIZE)) : 0;
Linus Torvalds's avatar
Linus Torvalds committed
207 208 209 210
}

static inline void shmem_unacct_blocks(unsigned long flags, long pages)
{
211
	if (flags & VM_NORESERVE)
Linus Torvalds's avatar
Linus Torvalds committed
212 213 214
		vm_unacct_memory(pages * VM_ACCT(PAGE_CACHE_SIZE));
}

215
static const struct super_operations shmem_ops;
216
static const struct address_space_operations shmem_aops;
217
static const struct file_operations shmem_file_operations;
218 219 220
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;
221
static const struct vm_operations_struct shmem_vm_ops;
Linus Torvalds's avatar
Linus Torvalds committed
222

223
static struct backing_dev_info shmem_backing_dev_info  __read_mostly = {
Linus Torvalds's avatar
Linus Torvalds committed
224
	.ra_pages	= 0,	/* No readahead */
225
	.capabilities	= BDI_CAP_NO_ACCT_AND_WRITEBACK | BDI_CAP_SWAP_BACKED,
Linus Torvalds's avatar
Linus Torvalds committed
226 227 228 229
	.unplug_io_fn	= default_unplug_io_fn,
};

static LIST_HEAD(shmem_swaplist);
230
static DEFINE_MUTEX(shmem_swaplist_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
231 232 233 234

static void shmem_free_blocks(struct inode *inode, long pages)
{
	struct shmem_sb_info *sbinfo = SHMEM_SB(inode->i_sb);
235
	if (sbinfo->max_blocks) {
Linus Torvalds's avatar
Linus Torvalds committed
236 237 238 239 240 241 242
		spin_lock(&sbinfo->stat_lock);
		sbinfo->free_blocks += pages;
		inode->i_blocks -= pages*BLOCKS_PER_PAGE;
		spin_unlock(&sbinfo->stat_lock);
	}
}

243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267
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);
	}
}

268
/**
Linus Torvalds's avatar
Linus Torvalds committed
269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292
 * 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);
	}
}

293
/**
Linus Torvalds's avatar
Linus Torvalds committed
294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389
 * 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;
390 391 392 393
	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
394 395
}

396
/**
Linus Torvalds's avatar
Linus Torvalds committed
397 398 399 400
 * 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?
401 402
 *
 * If the entry does not exist, allocate it.
Linus Torvalds's avatar
Linus Torvalds committed
403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422
 */
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.
		 */
423
		if (sbinfo->max_blocks) {
Linus Torvalds's avatar
Linus Torvalds committed
424 425 426 427 428 429 430 431 432 433 434
			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);
435
		page = shmem_dir_alloc(mapping_gfp_mask(inode->i_mapping));
Linus Torvalds's avatar
Linus Torvalds committed
436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459
		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;
}

460
/**
Linus Torvalds's avatar
Linus Torvalds committed
461
 * shmem_free_swp - free some swap entries in a directory
462 463 464
 * @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
465
 */
466 467
static int shmem_free_swp(swp_entry_t *dir, swp_entry_t *edir,
						spinlock_t *punch_lock)
Linus Torvalds's avatar
Linus Torvalds committed
468
{
469
	spinlock_t *punch_unlock = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
470 471 472 473 474
	swp_entry_t *ptr;
	int freed = 0;

	for (ptr = dir; ptr < edir; ptr++) {
		if (ptr->val) {
475 476 477 478 479 480 481
			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
482 483 484 485 486
			free_swap_and_cache(*ptr);
			*ptr = (swp_entry_t){0};
			freed++;
		}
	}
487 488
	if (punch_unlock)
		spin_unlock(punch_unlock);
Linus Torvalds's avatar
Linus Torvalds committed
489 490 491
	return freed;
}

492 493
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
494 495 496 497 498 499 500 501 502
{
	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;
503 504
		freed += shmem_free_swp(ptr+offset, ptr+offset+size,
							punch_lock);
Linus Torvalds's avatar
Linus Torvalds committed
505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535
		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);
}

536
static void shmem_truncate_range(struct inode *inode, loff_t start, loff_t end)
Linus Torvalds's avatar
Linus Torvalds committed
537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553
{
	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;
554
	int punch_hole;
555 556
	spinlock_t *needs_lock;
	spinlock_t *punch_lock;
557
	unsigned long upper_limit;
Linus Torvalds's avatar
Linus Torvalds committed
558 559

	inode->i_ctime = inode->i_mtime = CURRENT_TIME;
560
	idx = (start + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
Linus Torvalds's avatar
Linus Torvalds committed
561 562 563 564 565
	if (idx >= info->next_index)
		return;

	spin_lock(&info->lock);
	info->flags |= SHMEM_TRUNCATE;
566 567
	if (likely(end == (loff_t) -1)) {
		limit = info->next_index;
568
		upper_limit = SHMEM_MAX_INDEX;
569
		info->next_index = idx;
570
		needs_lock = NULL;
571
		punch_hole = 0;
572
	} else {
573 574 575 576 577 578 579 580
		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;
		}
581
		needs_lock = &info->lock;
582 583 584
		punch_hole = 1;
	}

Linus Torvalds's avatar
Linus Torvalds committed
585
	topdir = info->i_indirect;
586
	if (topdir && idx <= SHMEM_NR_DIRECT && !punch_hole) {
Linus Torvalds's avatar
Linus Torvalds committed
587 588 589 590 591 592 593 594 595 596 597
		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;
598
		nr_swaps_freed = shmem_free_swp(ptr+idx, ptr+size, needs_lock);
Linus Torvalds's avatar
Linus Torvalds committed
599
	}
600 601 602 603 604

	/*
	 * If there are no indirect blocks or we are punching a hole
	 * below indirect blocks, nothing to be done.
	 */
605
	if (!topdir || limit <= SHMEM_NR_DIRECT)
Linus Torvalds's avatar
Linus Torvalds committed
606 607
		goto done2;

608 609 610 611 612 613 614 615 616 617 618 619 620
	/*
	 * 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.
	 */

621
	upper_limit -= SHMEM_NR_DIRECT;
Linus Torvalds's avatar
Linus Torvalds committed
622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640
	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;
641
			if (!diroff && !offset && upper_limit >= stage) {
642 643 644 645 646 647 648
				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
649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673
				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;
674 675
			if (punch_hole)
				needs_lock = &info->lock;
676
			if (upper_limit >= stage) {
677 678 679 680 681 682 683
				if (needs_lock) {
					spin_lock(needs_lock);
					*dir = NULL;
					spin_unlock(needs_lock);
					needs_lock = NULL;
				} else
					*dir = NULL;
684 685 686
				nr_pages_to_free++;
				list_add(&middir->lru, &pages_to_free);
			}
Linus Torvalds's avatar
Linus Torvalds committed
687 688 689 690 691
			shmem_dir_unmap(dir);
			cond_resched();
			dir = shmem_dir_map(middir);
			diroff = 0;
		}
692
		punch_lock = needs_lock;
Linus Torvalds's avatar
Linus Torvalds committed
693
		subdir = dir[diroff];
694 695 696 697 698 699 700 701 702 703 704 705
		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
706 707 708 709
			size = limit - idx;
			if (size > ENTRIES_PER_PAGE)
				size = ENTRIES_PER_PAGE;
			freed = shmem_map_and_free_swp(subdir,
710
					offset, size, &dir, punch_lock);
Linus Torvalds's avatar
Linus Torvalds committed
711 712 713
			if (!dir)
				dir = shmem_dir_map(middir);
			nr_swaps_freed += freed;
714
			if (offset || punch_lock) {
Linus Torvalds's avatar
Linus Torvalds committed
715
				spin_lock(&info->lock);
716 717
				set_page_private(subdir,
					page_private(subdir) - freed);
Linus Torvalds's avatar
Linus Torvalds committed
718
				spin_unlock(&info->lock);
719 720
			} else
				BUG_ON(page_private(subdir) != freed);
Linus Torvalds's avatar
Linus Torvalds committed
721
		}
722
		offset = 0;
Linus Torvalds's avatar
Linus Torvalds committed
723 724 725 726 727 728 729 730 731 732 733
	}
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.
734 735 736 737 738
		 *
		 * Recalling truncate_inode_pages_range and unmap_mapping_range
		 * every time for punch_hole (which never got a chance to clear
		 * SHMEM_PAGEIN at the start of vmtruncate_range) is expensive,
		 * yet hardly ever necessary: try to optimize them out later.
Linus Torvalds's avatar
Linus Torvalds committed
739
		 */
740
		truncate_inode_pages_range(inode->i_mapping, start, end);
741 742 743
		if (punch_hole)
			unmap_mapping_range(inode->i_mapping, start,
							end - start, 1);
Linus Torvalds's avatar
Linus Torvalds committed
744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762
	}

	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);
	}
}

763 764 765 766 767
static void shmem_truncate(struct inode *inode)
{
	shmem_truncate_range(inode, inode->i_size, (loff_t)-1);
}

Linus Torvalds's avatar
Linus Torvalds committed
768 769 770 771 772 773
static int shmem_notify_change(struct dentry *dentry, struct iattr *attr)
{
	struct inode *inode = dentry->d_inode;
	struct page *page = NULL;
	int error;

774
	if (S_ISREG(inode->i_mode) && (attr->ia_valid & ATTR_SIZE)) {
Linus Torvalds's avatar
Linus Torvalds committed
775 776 777 778 779 780 781 782 783 784 785 786
		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);
787 788
				if (page)
					unlock_page(page);
Linus Torvalds's avatar
Linus Torvalds committed
789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808
			}
			/*
			 * 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);
809 810
#ifdef CONFIG_TMPFS_POSIX_ACL
	if (!error && (attr->ia_valid & ATTR_MODE))
811
		error = generic_acl_chmod(inode);
812
#endif
Linus Torvalds's avatar
Linus Torvalds committed
813 814 815 816 817 818 819 820 821 822
	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) {
823
		truncate_inode_pages(inode->i_mapping, 0);
Linus Torvalds's avatar
Linus Torvalds committed
824 825 826 827
		shmem_unacct_size(info->flags, inode->i_size);
		inode->i_size = 0;
		shmem_truncate(inode);
		if (!list_empty(&info->swaplist)) {
828
			mutex_lock(&shmem_swaplist_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
829
			list_del_init(&info->swaplist);
830
			mutex_unlock(&shmem_swaplist_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
831 832
		}
	}
833
	BUG_ON(inode->i_blocks);
834
	shmem_free_inode(inode->i_sb);
Linus Torvalds's avatar
Linus Torvalds committed
835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859
	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;
860
	int error;
Linus Torvalds's avatar
Linus Torvalds committed
861 862 863 864

	idx = 0;
	ptr = info->i_direct;
	spin_lock(&info->lock);
Hugh Dickins's avatar
Hugh Dickins committed
865 866 867 868
	if (!info->swapped) {
		list_del_init(&info->swaplist);
		goto lost2;
	}
Linus Torvalds's avatar
Linus Torvalds committed
869 870 871 872 873
	limit = info->next_index;
	size = limit;
	if (size > SHMEM_NR_DIRECT)
		size = SHMEM_NR_DIRECT;
	offset = shmem_find_swp(entry, ptr, ptr+size);
874
	if (offset >= 0)
Linus Torvalds's avatar
Linus Torvalds committed
875 876 877 878 879 880 881 882 883 884
		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);
885 886 887 888 889 890 891 892
			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
893 894 895 896 897 898 899 900 901 902 903 904 905 906
			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;
907
		if (subdir && page_private(subdir)) {
Linus Torvalds's avatar
Linus Torvalds committed
908 909 910 911 912
			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);
913
			shmem_swp_unmap(ptr);
Linus Torvalds's avatar
Linus Torvalds committed
914 915 916 917 918 919 920 921 922 923 924 925 926
			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;
927 928 929
	inode = igrab(&info->vfs_inode);
	spin_unlock(&info->lock);

Hugh Dickins's avatar
Hugh Dickins committed
930 931 932 933 934 935 936 937 938
	/*
	 * 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);
939 940 941 942 943
	mutex_unlock(&shmem_swaplist_mutex);

	error = 1;
	if (!inode)
		goto out;
KAMEZAWA Hiroyuki's avatar
KAMEZAWA Hiroyuki committed
944
	/*
945 946 947
	 * 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.
KAMEZAWA Hiroyuki's avatar
KAMEZAWA Hiroyuki committed
948
	 */
949
	error = mem_cgroup_cache_charge(page, current->mm, GFP_KERNEL);
Hugh Dickins's avatar
Hugh Dickins committed
950 951
	if (error)
		goto out;
952
	error = radix_tree_preload(GFP_KERNEL);
953 954 955 956
	if (error) {
		mem_cgroup_uncharge_cache_page(page);
		goto out;
	}
Hugh Dickins's avatar
Hugh Dickins committed
957
	error = 1;
958 959 960

	spin_lock(&info->lock);
	ptr = shmem_swp_entry(info, idx, NULL);
961
	if (ptr && ptr->val == entry.val) {
Nick Piggin's avatar
Nick Piggin committed
962
		error = add_to_page_cache_locked(page, inode->i_mapping,
Hugh Dickins's avatar
Hugh Dickins committed
963
						idx, GFP_NOWAIT);
964 965 966 967
		/* does mem_cgroup_uncharge_cache_page on error */
	} else	/* we must compensate for our precharge above */
		mem_cgroup_uncharge_cache_page(page);

968 969
	if (error == -EEXIST) {
		struct page *filepage = find_get_page(inode->i_mapping, idx);
970
		error = 1;
971 972 973 974 975 976 977 978 979 980 981
		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) {
982 983
		delete_from_swap_cache(page);
		set_page_dirty(page);
Linus Torvalds's avatar
Linus Torvalds committed
984
		info->flags |= SHMEM_PAGEIN;
985 986 987
		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
988
	}
989 990
	if (ptr)
		shmem_swp_unmap(ptr);
Linus Torvalds's avatar
Linus Torvalds committed
991
	spin_unlock(&info->lock);
Hugh Dickins's avatar
Hugh Dickins committed
992
	radix_tree_preload_end();
993 994 995 996 997
out:
	unlock_page(page);
	page_cache_release(page);
	iput(inode);		/* allows for NULL */
	return error;
Linus Torvalds's avatar
Linus Torvalds committed
998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008
}

/*
 * 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;

1009
	mutex_lock(&shmem_swaplist_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
1010 1011
	list_for_each_safe(p, next, &shmem_swaplist) {
		info = list_entry(p, struct shmem_inode_info, swaplist);
Hugh Dickins's avatar
Hugh Dickins committed
1012
		found = shmem_unuse_inode(info, entry, page);
1013
		cond_resched();
1014 1015
		if (found)
			goto out;
Linus Torvalds's avatar
Linus Torvalds committed
1016
	}
1017
	mutex_unlock(&shmem_swaplist_mutex);
Hugh Dickins's avatar
Hugh Dickins committed
1018 1019 1020 1021 1022 1023 1024 1025
	/*
	 * Can some race bring us here?  We've been holding page lock,
	 * so I think not; but would rather try again later than BUG()
	 */
	unlock_page(page);
	page_cache_release(page);
out:
	return (found < 0) ? found : 0;
Linus Torvalds's avatar
Linus Torvalds committed
1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045
}

/*
 * 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;
1046
	if (!total_swap_pages)
Linus Torvalds's avatar
Linus Torvalds committed
1047 1048
		goto redirty;

1049 1050 1051 1052 1053
	/*
	 * 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,
1054 1055 1056
	 * and not for the writeback threads or sync.  However, in those cases,
	 * we do still want to check if there's a redundant swappage to be
	 * discarded.
1057 1058 1059 1060 1061 1062
	 */
	if (wbc->for_reclaim)
		swap = get_swap_page();
	else
		swap.val = 0;

Linus Torvalds's avatar
Linus Torvalds committed
1063 1064 1065 1066 1067 1068
	spin_lock(&info->lock);
	if (index >= info->next_index) {
		BUG_ON(!(info->flags & SHMEM_TRUNCATE));
		goto unlock;
	}
	entry = shmem_swp_entry(info, index, NULL);
1069 1070 1071 1072 1073 1074 1075 1076 1077
	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
1078

1079
	if (swap.val && add_to_swap_cache(page, swap, GFP_ATOMIC) == 0) {
1080
		remove_from_page_cache(page);
Linus Torvalds's avatar
Linus Torvalds committed
1081 1082
		shmem_swp_set(info, entry, swap.val);
		shmem_swp_unmap(entry);
Hugh Dickins's avatar
Hugh Dickins committed
1083 1084 1085 1086
		if (list_empty(&info->swaplist))
			inode = igrab(inode);
		else
			inode = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
1087
		spin_unlock(&info->lock);
Hugh Dickins's avatar
Hugh Dickins committed
1088
		swap_shmem_alloc(swap);
1089
		BUG_ON(page_mapped(page));
1090
		page_cache_release(page);	/* pagecache ref */
1091
		swap_writepage(page, wbc);
Hugh Dickins's avatar
Hugh Dickins committed
1092 1093 1094 1095 1096 1097 1098
		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
1099 1100 1101 1102 1103 1104
		return 0;
	}

	shmem_swp_unmap(entry);
unlock:
	spin_unlock(&info->lock);
1105 1106 1107 1108
	/*
	 * add_to_swap_cache() doesn't return -EEXIST, so we can safely
	 * clear SWAP_HAS_CACHE flag.
	 */
1109
	swapcache_free(swap, NULL);
Linus Torvalds's avatar
Linus Torvalds committed
1110 1111
redirty:
	set_page_dirty(page);
1112 1113 1114 1115
	if (wbc->for_reclaim)
		return AOP_WRITEPAGE_ACTIVATE;	/* Return with page locked */
	unlock_page(page);
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
1116 1117 1118
}

#ifdef CONFIG_NUMA
1119
#ifdef CONFIG_TMPFS
1120
static void shmem_show_mpol(struct seq_file *seq, struct mempolicy *mpol)
1121
{
1122
	char buffer[64];
1123

1124
	if (!mpol || mpol->mode == MPOL_DEFAULT)
1125
		return;		/* show nothing */
1126

1127
	mpol_to_str(buffer, sizeof(buffer), mpol, 1);
1128 1129

	seq_printf(seq, ",mpol=%s", buffer);
1130
}
1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142

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;
}
1143 1144
#endif /* CONFIG_TMPFS */

1145 1146
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
1147
{
1148
	struct mempolicy mpol, *spol;
Linus Torvalds's avatar
Linus Torvalds committed
1149
	struct vm_area_struct pvma;
1150
	struct page *page;
Linus Torvalds's avatar
Linus Torvalds committed
1151

1152 1153 1154
	spol = mpol_cond_copy(&mpol,
				mpol_shared_policy_lookup(&info->policy, idx));

Linus Torvalds's avatar
Linus Torvalds committed
1155
	/* Create a pseudo vma that just contains the policy */
1156
	pvma.vm_start = 0;
Linus Torvalds's avatar
Linus Torvalds committed
1157
	pvma.vm_pgoff = idx;
1158
	pvma.vm_ops = NULL;
1159
	pvma.vm_policy = spol;
1160
	page = swapin_readahead(entry, gfp, &pvma, 0);
Linus Torvalds's avatar
Linus Torvalds committed
1161 1162 1163
	return page;
}

1164 1165
static struct page *shmem_alloc_page(gfp_t gfp,
			struct shmem_inode_info *info, unsigned long idx)
Linus Torvalds's avatar
Linus Torvalds committed
1166 1167 1168
{
	struct vm_area_struct pvma;

1169 1170
	/* Create a pseudo vma that just contains the policy */
	pvma.vm_start = 0;
Linus Torvalds's avatar
Linus Torvalds committed
1171
	pvma.vm_pgoff = idx;
1172 1173
	pvma.vm_ops = NULL;
	pvma.vm_policy = mpol_shared_policy_lookup(&info->policy, idx);
1174 1175 1176 1177 1178

	/*
	 * alloc_page_vma() will drop the shared policy reference
	 */
	return alloc_page_vma(gfp, &pvma, 0);
Linus Torvalds's avatar
Linus Torvalds committed
1179
}
1180 1181
#else /* !CONFIG_NUMA */
#ifdef CONFIG_TMPFS
1182
static inline void shmem_show_mpol(struct seq_file *seq, struct mempolicy *p)
1183 1184 1185 1186
{
}
#endif /* CONFIG_TMPFS */

1187 1188
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
1189
{
1190
	return swapin_readahead(entry, gfp, NULL, 0);
Linus Torvalds's avatar
Linus Torvalds committed
1191 1192
}

1193 1194
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
1195
{
1196
	return alloc_page(gfp);
Linus Torvalds's avatar
Linus Torvalds committed
1197
}
1198
#endif /* CONFIG_NUMA */
Linus Torvalds's avatar
Linus Torvalds committed
1199

1200 1201 1202 1203 1204 1205 1206
#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
1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223
/*
 * 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;
1224
	gfp_t gfp;
Linus Torvalds's avatar
Linus Torvalds committed
1225 1226 1227 1228
	int error;

	if (idx >= SHMEM_MAX_INDEX)
		return -EFBIG;
1229 1230

	if (type)
Nick Piggin's avatar
Nick Piggin committed
1231
		*type = 0;
1232

Linus Torvalds's avatar
Linus Torvalds committed
1233 1234 1235 1236
	/*
	 * 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
1237
	 * But shmem_readpage (required for splice) passes in a locked
1238 1239
	 * 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
1240 1241 1242 1243 1244 1245 1246
	 */
repeat:
	if (!filepage)
		filepage = find_lock_page(mapping, idx);
	if (filepage && PageUptodate(filepage))
		goto done;
	error = 0;
1247
	gfp = mapping_gfp_mask(mapping);
Hugh Dickins's avatar
Hugh Dickins committed
1248 1249 1250 1251 1252 1253 1254 1255 1256 1257
	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
1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274

	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
1275
			if (type && !(*type & VM_FAULT_MAJOR)) {
1276
				__count_vm_event(PGMAJFAULT);
Nick Piggin's avatar