inode.c 48.6 KB
Newer Older
L
Linus Torvalds 已提交
1 2
/*
 * (C) 1997 Linus Torvalds
3
 * (C) 1999 Andrea Arcangeli <andrea@suse.de> (dynamic inode allocation)
L
Linus Torvalds 已提交
4
 */
A
Al Viro 已提交
5
#include <linux/export.h>
L
Linus Torvalds 已提交
6 7 8 9 10 11 12 13
#include <linux/fs.h>
#include <linux/mm.h>
#include <linux/backing-dev.h>
#include <linux/hash.h>
#include <linux/swap.h>
#include <linux/security.h>
#include <linux/cdev.h>
#include <linux/bootmem.h>
14
#include <linux/fsnotify.h>
15
#include <linux/mount.h>
A
Al Viro 已提交
16
#include <linux/posix_acl.h>
17
#include <linux/prefetch.h>
18
#include <linux/buffer_head.h> /* for inode_has_buffers */
M
Miklos Szeredi 已提交
19
#include <linux/ratelimit.h>
20
#include "internal.h"
L
Linus Torvalds 已提交
21

22
/*
23
 * Inode locking rules:
24 25 26
 *
 * inode->i_lock protects:
 *   inode->i_state, inode->i_hash, __iget()
D
Dave Chinner 已提交
27
 * inode->i_sb->s_inode_lru_lock protects:
28
 *   inode->i_sb->s_inode_lru, inode->i_lru
29 30
 * inode_sb_list_lock protects:
 *   sb->s_inodes, inode->i_sb_list
31
 * bdi->wb.list_lock protects:
32
 *   bdi->wb.b_{dirty,io,more_io}, inode->i_wb_list
33 34
 * inode_hash_lock protects:
 *   inode_hashtable, inode->i_hash
35 36
 *
 * Lock ordering:
37 38 39
 *
 * inode_sb_list_lock
 *   inode->i_lock
D
Dave Chinner 已提交
40
 *     inode->i_sb->s_inode_lru_lock
41
 *
42
 * bdi->wb.list_lock
43
 *   inode->i_lock
44 45 46 47 48 49 50
 *
 * inode_hash_lock
 *   inode_sb_list_lock
 *   inode->i_lock
 *
 * iunique_lock
 *   inode_hash_lock
51 52
 */

53 54
static unsigned int i_hash_mask __read_mostly;
static unsigned int i_hash_shift __read_mostly;
55 56
static struct hlist_head *inode_hashtable __read_mostly;
static __cacheline_aligned_in_smp DEFINE_SPINLOCK(inode_hash_lock);
L
Linus Torvalds 已提交
57

58 59
__cacheline_aligned_in_smp DEFINE_SPINLOCK(inode_sb_list_lock);

J
Jens Axboe 已提交
60 61 62 63 64 65 66 67
/*
 * Empty aops. Can be used for the cases where the user does not
 * define any of the address_space operations.
 */
const struct address_space_operations empty_aops = {
};
EXPORT_SYMBOL(empty_aops);

L
Linus Torvalds 已提交
68 69 70 71 72
/*
 * Statistics gathering..
 */
struct inodes_stat_t inodes_stat;

73 74
static DEFINE_PER_CPU(unsigned long, nr_inodes);
static DEFINE_PER_CPU(unsigned long, nr_unused);
75

76
static struct kmem_cache *inode_cachep __read_mostly;
L
Linus Torvalds 已提交
77

78
static long get_nr_inodes(void)
79
{
80
	int i;
81
	long sum = 0;
82 83 84
	for_each_possible_cpu(i)
		sum += per_cpu(nr_inodes, i);
	return sum < 0 ? 0 : sum;
85 86
}

87
static inline long get_nr_inodes_unused(void)
88
{
89
	int i;
90
	long sum = 0;
91 92 93
	for_each_possible_cpu(i)
		sum += per_cpu(nr_unused, i);
	return sum < 0 ? 0 : sum;
94 95
}

96
long get_nr_dirty_inodes(void)
97
{
98
	/* not actually dirty inodes, but a wild approximation */
99
	long nr_dirty = get_nr_inodes() - get_nr_inodes_unused();
100 101 102 103 104 105 106 107 108 109 110
	return nr_dirty > 0 ? nr_dirty : 0;
}

/*
 * Handle nr_inode sysctl
 */
#ifdef CONFIG_SYSCTL
int proc_nr_inodes(ctl_table *table, int write,
		   void __user *buffer, size_t *lenp, loff_t *ppos)
{
	inodes_stat.nr_inodes = get_nr_inodes();
111
	inodes_stat.nr_unused = get_nr_inodes_unused();
112
	return proc_doulongvec_minmax(table, write, buffer, lenp, ppos);
113 114 115
}
#endif

D
David Chinner 已提交
116 117
/**
 * inode_init_always - perform inode structure intialisation
R
Randy Dunlap 已提交
118 119
 * @sb: superblock inode belongs to
 * @inode: inode to initialise
D
David Chinner 已提交
120 121 122 123
 *
 * These are initializations that need to be done on every inode
 * allocation as the fields are not initialised by slab allocation.
 */
124
int inode_init_always(struct super_block *sb, struct inode *inode)
L
Linus Torvalds 已提交
125
{
126
	static const struct inode_operations empty_iops;
127
	static const struct file_operations empty_fops;
128
	struct address_space *const mapping = &inode->i_data;
D
David Chinner 已提交
129 130 131 132 133 134 135

	inode->i_sb = sb;
	inode->i_blkbits = sb->s_blocksize_bits;
	inode->i_flags = 0;
	atomic_set(&inode->i_count, 1);
	inode->i_op = &empty_iops;
	inode->i_fop = &empty_fops;
M
Miklos Szeredi 已提交
136
	inode->__i_nlink = 1;
137
	inode->i_opflags = 0;
138 139
	i_uid_write(inode, 0);
	i_gid_write(inode, 0);
D
David Chinner 已提交
140 141 142 143 144
	atomic_set(&inode->i_writecount, 0);
	inode->i_size = 0;
	inode->i_blocks = 0;
	inode->i_bytes = 0;
	inode->i_generation = 0;
L
Linus Torvalds 已提交
145
#ifdef CONFIG_QUOTA
D
David Chinner 已提交
146
	memset(&inode->i_dquot, 0, sizeof(inode->i_dquot));
L
Linus Torvalds 已提交
147
#endif
D
David Chinner 已提交
148 149 150 151 152
	inode->i_pipe = NULL;
	inode->i_bdev = NULL;
	inode->i_cdev = NULL;
	inode->i_rdev = 0;
	inode->dirtied_when = 0;
M
Mimi Zohar 已提交
153 154

	if (security_inode_alloc(inode))
155
		goto out;
D
David Chinner 已提交
156 157 158 159 160 161
	spin_lock_init(&inode->i_lock);
	lockdep_set_class(&inode->i_lock, &sb->s_type->i_lock_key);

	mutex_init(&inode->i_mutex);
	lockdep_set_class(&inode->i_mutex, &sb->s_type->i_mutex_key);

C
Christoph Hellwig 已提交
162
	atomic_set(&inode->i_dio_count, 0);
D
David Chinner 已提交
163 164 165 166

	mapping->a_ops = &empty_aops;
	mapping->host = inode;
	mapping->flags = 0;
H
Hugh Dickins 已提交
167
	mapping_set_gfp_mask(mapping, GFP_HIGHUSER_MOVABLE);
168
	mapping->private_data = NULL;
D
David Chinner 已提交
169 170 171 172 173 174 175 176 177 178 179
	mapping->backing_dev_info = &default_backing_dev_info;
	mapping->writeback_index = 0;

	/*
	 * If the block_device provides a backing_dev_info for client
	 * inodes then use that.  Otherwise the inode share the bdev's
	 * backing_dev_info.
	 */
	if (sb->s_bdev) {
		struct backing_dev_info *bdi;

180
		bdi = sb->s_bdev->bd_inode->i_mapping->backing_dev_info;
D
David Chinner 已提交
181 182 183 184
		mapping->backing_dev_info = bdi;
	}
	inode->i_private = NULL;
	inode->i_mapping = mapping;
A
Al Viro 已提交
185
	INIT_HLIST_HEAD(&inode->i_dentry);	/* buggered by rcu freeing */
A
Al Viro 已提交
186 187 188
#ifdef CONFIG_FS_POSIX_ACL
	inode->i_acl = inode->i_default_acl = ACL_NOT_CACHED;
#endif
D
David Chinner 已提交
189

190 191 192 193
#ifdef CONFIG_FSNOTIFY
	inode->i_fsnotify_mask = 0;
#endif

194
	this_cpu_inc(nr_inodes);
195

196 197 198
	return 0;
out:
	return -ENOMEM;
L
Linus Torvalds 已提交
199
}
D
David Chinner 已提交
200 201 202 203 204 205 206 207 208 209 210
EXPORT_SYMBOL(inode_init_always);

static struct inode *alloc_inode(struct super_block *sb)
{
	struct inode *inode;

	if (sb->s_op->alloc_inode)
		inode = sb->s_op->alloc_inode(sb);
	else
		inode = kmem_cache_alloc(inode_cachep, GFP_KERNEL);

211 212 213 214 215 216 217 218 219 220 221 222
	if (!inode)
		return NULL;

	if (unlikely(inode_init_always(sb, inode))) {
		if (inode->i_sb->s_op->destroy_inode)
			inode->i_sb->s_op->destroy_inode(inode);
		else
			kmem_cache_free(inode_cachep, inode);
		return NULL;
	}

	return inode;
D
David Chinner 已提交
223
}
L
Linus Torvalds 已提交
224

225 226 227 228 229 230
void free_inode_nonrcu(struct inode *inode)
{
	kmem_cache_free(inode_cachep, inode);
}
EXPORT_SYMBOL(free_inode_nonrcu);

C
Christoph Hellwig 已提交
231
void __destroy_inode(struct inode *inode)
L
Linus Torvalds 已提交
232
{
233
	BUG_ON(inode_has_buffers(inode));
L
Linus Torvalds 已提交
234
	security_inode_free(inode);
235
	fsnotify_inode_delete(inode);
M
Miklos Szeredi 已提交
236 237 238 239 240
	if (!inode->i_nlink) {
		WARN_ON(atomic_long_read(&inode->i_sb->s_remove_count) == 0);
		atomic_long_dec(&inode->i_sb->s_remove_count);
	}

A
Al Viro 已提交
241 242 243 244 245 246
#ifdef CONFIG_FS_POSIX_ACL
	if (inode->i_acl && inode->i_acl != ACL_NOT_CACHED)
		posix_acl_release(inode->i_acl);
	if (inode->i_default_acl && inode->i_default_acl != ACL_NOT_CACHED)
		posix_acl_release(inode->i_default_acl);
#endif
247
	this_cpu_dec(nr_inodes);
C
Christoph Hellwig 已提交
248 249 250
}
EXPORT_SYMBOL(__destroy_inode);

N
Nick Piggin 已提交
251 252 253 254 255 256
static void i_callback(struct rcu_head *head)
{
	struct inode *inode = container_of(head, struct inode, i_rcu);
	kmem_cache_free(inode_cachep, inode);
}

C
Christoph Hellwig 已提交
257
static void destroy_inode(struct inode *inode)
C
Christoph Hellwig 已提交
258
{
N
Nick Piggin 已提交
259
	BUG_ON(!list_empty(&inode->i_lru));
C
Christoph Hellwig 已提交
260
	__destroy_inode(inode);
L
Linus Torvalds 已提交
261 262 263
	if (inode->i_sb->s_op->destroy_inode)
		inode->i_sb->s_op->destroy_inode(inode);
	else
N
Nick Piggin 已提交
264
		call_rcu(&inode->i_rcu, i_callback);
L
Linus Torvalds 已提交
265 266
}

M
Miklos Szeredi 已提交
267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335
/**
 * drop_nlink - directly drop an inode's link count
 * @inode: inode
 *
 * This is a low-level filesystem helper to replace any
 * direct filesystem manipulation of i_nlink.  In cases
 * where we are attempting to track writes to the
 * filesystem, a decrement to zero means an imminent
 * write when the file is truncated and actually unlinked
 * on the filesystem.
 */
void drop_nlink(struct inode *inode)
{
	WARN_ON(inode->i_nlink == 0);
	inode->__i_nlink--;
	if (!inode->i_nlink)
		atomic_long_inc(&inode->i_sb->s_remove_count);
}
EXPORT_SYMBOL(drop_nlink);

/**
 * clear_nlink - directly zero an inode's link count
 * @inode: inode
 *
 * This is a low-level filesystem helper to replace any
 * direct filesystem manipulation of i_nlink.  See
 * drop_nlink() for why we care about i_nlink hitting zero.
 */
void clear_nlink(struct inode *inode)
{
	if (inode->i_nlink) {
		inode->__i_nlink = 0;
		atomic_long_inc(&inode->i_sb->s_remove_count);
	}
}
EXPORT_SYMBOL(clear_nlink);

/**
 * set_nlink - directly set an inode's link count
 * @inode: inode
 * @nlink: new nlink (should be non-zero)
 *
 * This is a low-level filesystem helper to replace any
 * direct filesystem manipulation of i_nlink.
 */
void set_nlink(struct inode *inode, unsigned int nlink)
{
	if (!nlink) {
		clear_nlink(inode);
	} else {
		/* Yes, some filesystems do change nlink from zero to one */
		if (inode->i_nlink == 0)
			atomic_long_dec(&inode->i_sb->s_remove_count);

		inode->__i_nlink = nlink;
	}
}
EXPORT_SYMBOL(set_nlink);

/**
 * inc_nlink - directly increment an inode's link count
 * @inode: inode
 *
 * This is a low-level filesystem helper to replace any
 * direct filesystem manipulation of i_nlink.  Currently,
 * it is only here for parity with dec_nlink().
 */
void inc_nlink(struct inode *inode)
{
336 337
	if (unlikely(inode->i_nlink == 0)) {
		WARN_ON(!(inode->i_state & I_LINKABLE));
M
Miklos Szeredi 已提交
338
		atomic_long_dec(&inode->i_sb->s_remove_count);
339
	}
M
Miklos Szeredi 已提交
340 341 342 343 344

	inode->__i_nlink++;
}
EXPORT_SYMBOL(inc_nlink);

345 346 347 348 349
void address_space_init_once(struct address_space *mapping)
{
	memset(mapping, 0, sizeof(*mapping));
	INIT_RADIX_TREE(&mapping->page_tree, GFP_ATOMIC);
	spin_lock_init(&mapping->tree_lock);
350
	mutex_init(&mapping->i_mmap_mutex);
351 352
	INIT_LIST_HEAD(&mapping->private_list);
	spin_lock_init(&mapping->private_lock);
353
	mapping->i_mmap = RB_ROOT;
354 355 356 357
	INIT_LIST_HEAD(&mapping->i_mmap_nonlinear);
}
EXPORT_SYMBOL(address_space_init_once);

L
Linus Torvalds 已提交
358 359 360 361 362 363 364 365 366 367
/*
 * These are initializations that only need to be done
 * once, because the fields are idempotent across use
 * of the inode, so let the slab aware of that.
 */
void inode_init_once(struct inode *inode)
{
	memset(inode, 0, sizeof(*inode));
	INIT_HLIST_NODE(&inode->i_hash);
	INIT_LIST_HEAD(&inode->i_devices);
N
Nick Piggin 已提交
368 369
	INIT_LIST_HEAD(&inode->i_wb_list);
	INIT_LIST_HEAD(&inode->i_lru);
370
	address_space_init_once(&inode->i_data);
L
Linus Torvalds 已提交
371
	i_size_ordered_init(inode);
372
#ifdef CONFIG_FSNOTIFY
373
	INIT_HLIST_HEAD(&inode->i_fsnotify_marks);
374
#endif
L
Linus Torvalds 已提交
375 376 377
}
EXPORT_SYMBOL(inode_init_once);

378
static void init_once(void *foo)
L
Linus Torvalds 已提交
379
{
380
	struct inode *inode = (struct inode *) foo;
L
Linus Torvalds 已提交
381

C
Christoph Lameter 已提交
382
	inode_init_once(inode);
L
Linus Torvalds 已提交
383 384 385
}

/*
386
 * inode->i_lock must be held
L
Linus Torvalds 已提交
387
 */
388
void __iget(struct inode *inode)
L
Linus Torvalds 已提交
389
{
390 391
	atomic_inc(&inode->i_count);
}
392

A
Al Viro 已提交
393 394 395 396 397 398 399 400 401
/*
 * get additional reference to inode; caller must already hold one.
 */
void ihold(struct inode *inode)
{
	WARN_ON(atomic_inc_return(&inode->i_count) < 2);
}
EXPORT_SYMBOL(ihold);

402 403
static void inode_lru_list_add(struct inode *inode)
{
D
Dave Chinner 已提交
404
	spin_lock(&inode->i_sb->s_inode_lru_lock);
N
Nick Piggin 已提交
405
	if (list_empty(&inode->i_lru)) {
406 407
		list_add(&inode->i_lru, &inode->i_sb->s_inode_lru);
		inode->i_sb->s_nr_inodes_unused++;
408
		this_cpu_inc(nr_unused);
409
	}
D
Dave Chinner 已提交
410
	spin_unlock(&inode->i_sb->s_inode_lru_lock);
411
}
412

413 414 415 416 417 418 419 420 421 422 423 424 425
/*
 * Add inode to LRU if needed (inode is unused and clean).
 *
 * Needs inode->i_lock held.
 */
void inode_add_lru(struct inode *inode)
{
	if (!(inode->i_state & (I_DIRTY | I_SYNC | I_FREEING | I_WILL_FREE)) &&
	    !atomic_read(&inode->i_count) && inode->i_sb->s_flags & MS_ACTIVE)
		inode_lru_list_add(inode);
}


426 427
static void inode_lru_list_del(struct inode *inode)
{
D
Dave Chinner 已提交
428
	spin_lock(&inode->i_sb->s_inode_lru_lock);
N
Nick Piggin 已提交
429 430
	if (!list_empty(&inode->i_lru)) {
		list_del_init(&inode->i_lru);
431
		inode->i_sb->s_nr_inodes_unused--;
432
		this_cpu_dec(nr_unused);
433
	}
D
Dave Chinner 已提交
434
	spin_unlock(&inode->i_sb->s_inode_lru_lock);
L
Linus Torvalds 已提交
435 436
}

437 438 439 440 441 442
/**
 * inode_sb_list_add - add inode to the superblock list of inodes
 * @inode: inode to add
 */
void inode_sb_list_add(struct inode *inode)
{
443 444 445
	spin_lock(&inode_sb_list_lock);
	list_add(&inode->i_sb_list, &inode->i_sb->s_inodes);
	spin_unlock(&inode_sb_list_lock);
446 447 448
}
EXPORT_SYMBOL_GPL(inode_sb_list_add);

449
static inline void inode_sb_list_del(struct inode *inode)
450
{
451 452 453 454 455
	if (!list_empty(&inode->i_sb_list)) {
		spin_lock(&inode_sb_list_lock);
		list_del_init(&inode->i_sb_list);
		spin_unlock(&inode_sb_list_lock);
	}
456 457
}

458 459 460 461 462 463
static unsigned long hash(struct super_block *sb, unsigned long hashval)
{
	unsigned long tmp;

	tmp = (hashval * (unsigned long)sb) ^ (GOLDEN_RATIO_PRIME + hashval) /
			L1_CACHE_BYTES;
464 465
	tmp = tmp ^ ((tmp ^ GOLDEN_RATIO_PRIME) >> i_hash_shift);
	return tmp & i_hash_mask;
466 467 468 469 470 471 472 473 474 475 476 477
}

/**
 *	__insert_inode_hash - hash an inode
 *	@inode: unhashed inode
 *	@hashval: unsigned long value used to locate this object in the
 *		inode_hashtable.
 *
 *	Add an inode to the inode hash for this superblock.
 */
void __insert_inode_hash(struct inode *inode, unsigned long hashval)
{
478 479
	struct hlist_head *b = inode_hashtable + hash(inode->i_sb, hashval);

480
	spin_lock(&inode_hash_lock);
481
	spin_lock(&inode->i_lock);
482
	hlist_add_head(&inode->i_hash, b);
483
	spin_unlock(&inode->i_lock);
484
	spin_unlock(&inode_hash_lock);
485 486 487 488
}
EXPORT_SYMBOL(__insert_inode_hash);

/**
489
 *	__remove_inode_hash - remove an inode from the hash
490 491 492 493
 *	@inode: inode to unhash
 *
 *	Remove an inode from the superblock.
 */
494
void __remove_inode_hash(struct inode *inode)
495
{
496
	spin_lock(&inode_hash_lock);
497
	spin_lock(&inode->i_lock);
498
	hlist_del_init(&inode->i_hash);
499
	spin_unlock(&inode->i_lock);
500
	spin_unlock(&inode_hash_lock);
501
}
502
EXPORT_SYMBOL(__remove_inode_hash);
503

504
void clear_inode(struct inode *inode)
A
Al Viro 已提交
505 506
{
	might_sleep();
507 508 509 510 511 512
	/*
	 * We have to cycle tree_lock here because reclaim can be still in the
	 * process of removing the last page (in __delete_from_page_cache())
	 * and we must not free mapping under it.
	 */
	spin_lock_irq(&inode->i_data.tree_lock);
A
Al Viro 已提交
513
	BUG_ON(inode->i_data.nrpages);
514
	spin_unlock_irq(&inode->i_data.tree_lock);
A
Al Viro 已提交
515 516 517
	BUG_ON(!list_empty(&inode->i_data.private_list));
	BUG_ON(!(inode->i_state & I_FREEING));
	BUG_ON(inode->i_state & I_CLEAR);
N
Nick Piggin 已提交
518
	/* don't need i_lock here, no concurrent mods to i_state */
A
Al Viro 已提交
519 520
	inode->i_state = I_FREEING | I_CLEAR;
}
521
EXPORT_SYMBOL(clear_inode);
A
Al Viro 已提交
522

D
Dave Chinner 已提交
523 524 525 526 527 528 529 530 531 532 533 534 535
/*
 * Free the inode passed in, removing it from the lists it is still connected
 * to. We remove any pages still attached to the inode and wait for any IO that
 * is still in progress before finally destroying the inode.
 *
 * An inode must already be marked I_FREEING so that we avoid the inode being
 * moved back onto lists if we race with other code that manipulates the lists
 * (e.g. writeback_single_inode). The caller is responsible for setting this.
 *
 * An inode must already be removed from the LRU list before being evicted from
 * the cache. This should occur atomically with setting the I_FREEING state
 * flag, so no inodes here should ever be on the LRU when being evicted.
 */
536
static void evict(struct inode *inode)
537 538 539
{
	const struct super_operations *op = inode->i_sb->s_op;

D
Dave Chinner 已提交
540 541 542
	BUG_ON(!(inode->i_state & I_FREEING));
	BUG_ON(!list_empty(&inode->i_lru));

543 544 545
	if (!list_empty(&inode->i_wb_list))
		inode_wb_list_del(inode);

546 547
	inode_sb_list_del(inode);

548 549 550 551 552 553 554
	/*
	 * Wait for flusher thread to be done with the inode so that filesystem
	 * does not start destroying it while writeback is still running. Since
	 * the inode has I_FREEING set, flusher thread won't start new work on
	 * the inode.  We just have to wait for running writeback to finish.
	 */
	inode_wait_for_writeback(inode);
555

A
Al Viro 已提交
556 557
	if (op->evict_inode) {
		op->evict_inode(inode);
558 559 560
	} else {
		if (inode->i_data.nrpages)
			truncate_inode_pages(&inode->i_data, 0);
561
		clear_inode(inode);
562
	}
563 564 565 566
	if (S_ISBLK(inode->i_mode) && inode->i_bdev)
		bd_forget(inode);
	if (S_ISCHR(inode->i_mode) && inode->i_cdev)
		cd_forget(inode);
D
Dave Chinner 已提交
567 568 569 570 571 572 573 574 575

	remove_inode_hash(inode);

	spin_lock(&inode->i_lock);
	wake_up_bit(&inode->i_state, __I_NEW);
	BUG_ON(inode->i_state != (I_FREEING | I_CLEAR));
	spin_unlock(&inode->i_lock);

	destroy_inode(inode);
576 577
}

L
Linus Torvalds 已提交
578 579 580 581 582 583 584 585 586 587 588 589
/*
 * dispose_list - dispose of the contents of a local list
 * @head: the head of the list to free
 *
 * Dispose-list gets a local list with local inodes in it, so it doesn't
 * need to worry about list corruption and SMP locks.
 */
static void dispose_list(struct list_head *head)
{
	while (!list_empty(head)) {
		struct inode *inode;

N
Nick Piggin 已提交
590 591
		inode = list_first_entry(head, struct inode, i_lru);
		list_del_init(&inode->i_lru);
L
Linus Torvalds 已提交
592

593
		evict(inode);
L
Linus Torvalds 已提交
594 595 596
	}
}

A
Al Viro 已提交
597 598 599 600 601 602 603 604
/**
 * evict_inodes	- evict all evictable inodes for a superblock
 * @sb:		superblock to operate on
 *
 * Make sure that no inodes with zero refcount are retained.  This is
 * called by superblock shutdown after having MS_ACTIVE flag removed,
 * so any inode reaching zero refcount during or after that call will
 * be immediately evicted.
L
Linus Torvalds 已提交
605
 */
A
Al Viro 已提交
606
void evict_inodes(struct super_block *sb)
L
Linus Torvalds 已提交
607
{
A
Al Viro 已提交
608 609
	struct inode *inode, *next;
	LIST_HEAD(dispose);
L
Linus Torvalds 已提交
610

611
	spin_lock(&inode_sb_list_lock);
A
Al Viro 已提交
612 613
	list_for_each_entry_safe(inode, next, &sb->s_inodes, i_sb_list) {
		if (atomic_read(&inode->i_count))
N
Nick Piggin 已提交
614
			continue;
615 616 617 618

		spin_lock(&inode->i_lock);
		if (inode->i_state & (I_NEW | I_FREEING | I_WILL_FREE)) {
			spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
619
			continue;
620
		}
A
Al Viro 已提交
621 622

		inode->i_state |= I_FREEING;
623
		inode_lru_list_del(inode);
624
		spin_unlock(&inode->i_lock);
625
		list_add(&inode->i_lru, &dispose);
L
Linus Torvalds 已提交
626
	}
627
	spin_unlock(&inode_sb_list_lock);
A
Al Viro 已提交
628 629

	dispose_list(&dispose);
L
Linus Torvalds 已提交
630 631 632
}

/**
633 634
 * invalidate_inodes	- attempt to free all inodes on a superblock
 * @sb:		superblock to operate on
635
 * @kill_dirty: flag to guide handling of dirty inodes
L
Linus Torvalds 已提交
636
 *
637 638
 * Attempts to free all inodes for a given superblock.  If there were any
 * busy inodes return a non-zero value, else zero.
639 640
 * If @kill_dirty is set, discard dirty inodes too, otherwise treat
 * them as busy.
L
Linus Torvalds 已提交
641
 */
642
int invalidate_inodes(struct super_block *sb, bool kill_dirty)
L
Linus Torvalds 已提交
643
{
644
	int busy = 0;
645 646
	struct inode *inode, *next;
	LIST_HEAD(dispose);
L
Linus Torvalds 已提交
647

648
	spin_lock(&inode_sb_list_lock);
649
	list_for_each_entry_safe(inode, next, &sb->s_inodes, i_sb_list) {
650 651 652
		spin_lock(&inode->i_lock);
		if (inode->i_state & (I_NEW | I_FREEING | I_WILL_FREE)) {
			spin_unlock(&inode->i_lock);
N
Nick Piggin 已提交
653
			continue;
654
		}
655
		if (inode->i_state & I_DIRTY && !kill_dirty) {
656
			spin_unlock(&inode->i_lock);
657 658 659
			busy = 1;
			continue;
		}
660
		if (atomic_read(&inode->i_count)) {
661
			spin_unlock(&inode->i_lock);
662
			busy = 1;
L
Linus Torvalds 已提交
663 664
			continue;
		}
665 666

		inode->i_state |= I_FREEING;
667
		inode_lru_list_del(inode);
668
		spin_unlock(&inode->i_lock);
669
		list_add(&inode->i_lru, &dispose);
L
Linus Torvalds 已提交
670
	}
671
	spin_unlock(&inode_sb_list_lock);
L
Linus Torvalds 已提交
672

673
	dispose_list(&dispose);
L
Linus Torvalds 已提交
674 675 676 677 678 679

	return busy;
}

static int can_unuse(struct inode *inode)
{
680
	if (inode->i_state & ~I_REFERENCED)
L
Linus Torvalds 已提交
681 682 683 684 685 686 687 688 689 690 691
		return 0;
	if (inode_has_buffers(inode))
		return 0;
	if (atomic_read(&inode->i_count))
		return 0;
	if (inode->i_data.nrpages)
		return 0;
	return 1;
}

/*
692 693 694 695
 * Walk the superblock inode LRU for freeable inodes and attempt to free them.
 * This is called from the superblock shrinker function with a number of inodes
 * to trim from the LRU. Inodes to be freed are moved to a temporary list and
 * then are freed outside inode_lock by dispose_list().
L
Linus Torvalds 已提交
696 697
 *
 * Any inodes which are pinned purely because of attached pagecache have their
698 699
 * pagecache removed.  If the inode has metadata buffers attached to
 * mapping->private_list then try to remove them.
L
Linus Torvalds 已提交
700
 *
701 702 703 704 705 706 707
 * If the inode has the I_REFERENCED flag set, then it means that it has been
 * used recently - the flag is set in iput_final(). When we encounter such an
 * inode, clear the flag and move it to the back of the LRU so it gets another
 * pass through the LRU before it gets reclaimed. This is necessary because of
 * the fact we are doing lazy LRU updates to minimise lock contention so the
 * LRU does not have strict ordering. Hence we don't want to reclaim inodes
 * with this flag set because they are the inodes that are out of order.
L
Linus Torvalds 已提交
708
 */
709
void prune_icache_sb(struct super_block *sb, int nr_to_scan)
L
Linus Torvalds 已提交
710 711 712 713 714
{
	LIST_HEAD(freeable);
	int nr_scanned;
	unsigned long reap = 0;

D
Dave Chinner 已提交
715
	spin_lock(&sb->s_inode_lru_lock);
716
	for (nr_scanned = nr_to_scan; nr_scanned >= 0; nr_scanned--) {
L
Linus Torvalds 已提交
717 718
		struct inode *inode;

719
		if (list_empty(&sb->s_inode_lru))
L
Linus Torvalds 已提交
720 721
			break;

722
		inode = list_entry(sb->s_inode_lru.prev, struct inode, i_lru);
L
Linus Torvalds 已提交
723

724
		/*
D
Dave Chinner 已提交
725
		 * we are inverting the sb->s_inode_lru_lock/inode->i_lock here,
726 727 728 729
		 * so use a trylock. If we fail to get the lock, just move the
		 * inode to the back of the list so we don't spin on it.
		 */
		if (!spin_trylock(&inode->i_lock)) {
730
			list_move(&inode->i_lru, &sb->s_inode_lru);
731 732 733
			continue;
		}

734 735 736 737 738 739
		/*
		 * Referenced or dirty inodes are still in use. Give them
		 * another pass through the LRU as we canot reclaim them now.
		 */
		if (atomic_read(&inode->i_count) ||
		    (inode->i_state & ~I_REFERENCED)) {
N
Nick Piggin 已提交
740
			list_del_init(&inode->i_lru);
741
			spin_unlock(&inode->i_lock);
742
			sb->s_nr_inodes_unused--;
743
			this_cpu_dec(nr_unused);
744 745 746 747 748 749
			continue;
		}

		/* recently referenced inodes get one more pass */
		if (inode->i_state & I_REFERENCED) {
			inode->i_state &= ~I_REFERENCED;
750
			list_move(&inode->i_lru, &sb->s_inode_lru);
751
			spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
752 753 754 755
			continue;
		}
		if (inode_has_buffers(inode) || inode->i_data.nrpages) {
			__iget(inode);
756
			spin_unlock(&inode->i_lock);
D
Dave Chinner 已提交
757
			spin_unlock(&sb->s_inode_lru_lock);
L
Linus Torvalds 已提交
758
			if (remove_inode_buffers(inode))
759 760
				reap += invalidate_mapping_pages(&inode->i_data,
								0, -1);
L
Linus Torvalds 已提交
761
			iput(inode);
D
Dave Chinner 已提交
762
			spin_lock(&sb->s_inode_lru_lock);
L
Linus Torvalds 已提交
763

764
			if (inode != list_entry(sb->s_inode_lru.next,
N
Nick Piggin 已提交
765
						struct inode, i_lru))
L
Linus Torvalds 已提交
766
				continue;	/* wrong inode or list_empty */
767 768 769
			/* avoid lock inversions with trylock */
			if (!spin_trylock(&inode->i_lock))
				continue;
770 771
			if (!can_unuse(inode)) {
				spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
772
				continue;
773
			}
L
Linus Torvalds 已提交
774
		}
775
		WARN_ON(inode->i_state & I_NEW);
L
Linus Torvalds 已提交
776
		inode->i_state |= I_FREEING;
777
		spin_unlock(&inode->i_lock);
N
Nick Piggin 已提交
778 779

		list_move(&inode->i_lru, &freeable);
780
		sb->s_nr_inodes_unused--;
781
		this_cpu_dec(nr_unused);
L
Linus Torvalds 已提交
782
	}
783 784 785 786
	if (current_is_kswapd())
		__count_vm_events(KSWAPD_INODESTEAL, reap);
	else
		__count_vm_events(PGINODESTEAL, reap);
D
Dave Chinner 已提交
787
	spin_unlock(&sb->s_inode_lru_lock);
788 789
	if (current->reclaim_state)
		current->reclaim_state->reclaimed_slab += reap;
L
Linus Torvalds 已提交
790 791 792 793 794 795 796 797

	dispose_list(&freeable);
}

static void __wait_on_freeing_inode(struct inode *inode);
/*
 * Called with the inode lock held.
 */
798 799 800 801
static struct inode *find_inode(struct super_block *sb,
				struct hlist_head *head,
				int (*test)(struct inode *, void *),
				void *data)
L
Linus Torvalds 已提交
802
{
803
	struct inode *inode = NULL;
L
Linus Torvalds 已提交
804 805

repeat:
806
	hlist_for_each_entry(inode, head, i_hash) {
807 808 809
		spin_lock(&inode->i_lock);
		if (inode->i_sb != sb) {
			spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
810
			continue;
811 812 813
		}
		if (!test(inode, data)) {
			spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
814
			continue;
815
		}
A
Al Viro 已提交
816
		if (inode->i_state & (I_FREEING|I_WILL_FREE)) {
L
Linus Torvalds 已提交
817 818 819
			__wait_on_freeing_inode(inode);
			goto repeat;
		}
820
		__iget(inode);
821
		spin_unlock(&inode->i_lock);
822
		return inode;
L
Linus Torvalds 已提交
823
	}
824
	return NULL;
L
Linus Torvalds 已提交
825 826 827 828 829 830
}

/*
 * find_inode_fast is the fast path version of find_inode, see the comment at
 * iget_locked for details.
 */
831 832
static struct inode *find_inode_fast(struct super_block *sb,
				struct hlist_head *head, unsigned long ino)
L
Linus Torvalds 已提交
833
{
834
	struct inode *inode = NULL;
L
Linus Torvalds 已提交
835 836

repeat:
837
	hlist_for_each_entry(inode, head, i_hash) {
838 839 840
		spin_lock(&inode->i_lock);
		if (inode->i_ino != ino) {
			spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
841
			continue;
842 843 844
		}
		if (inode->i_sb != sb) {
			spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
845
			continue;
846
		}
A
Al Viro 已提交
847
		if (inode->i_state & (I_FREEING|I_WILL_FREE)) {
L
Linus Torvalds 已提交
848 849 850
			__wait_on_freeing_inode(inode);
			goto repeat;
		}
851
		__iget(inode);
852
		spin_unlock(&inode->i_lock);
853
		return inode;
L
Linus Torvalds 已提交
854
	}
855
	return NULL;
856 857
}

858 859 860 861
/*
 * Each cpu owns a range of LAST_INO_BATCH numbers.
 * 'shared_last_ino' is dirtied only once out of LAST_INO_BATCH allocations,
 * to renew the exhausted range.
862
 *
863 864 865 866 867 868 869 870 871
 * This does not significantly increase overflow rate because every CPU can
 * consume at most LAST_INO_BATCH-1 unused inode numbers. So there is
 * NR_CPUS*(LAST_INO_BATCH-1) wastage. At 4096 and 1024, this is ~0.1% of the
 * 2^32 range, and is a worst-case. Even a 50% wastage would only increase
 * overflow rate by 2x, which does not seem too significant.
 *
 * On a 32bit, non LFS stat() call, glibc will generate an EOVERFLOW
 * error if st_ino won't fit in target struct field. Use 32bit counter
 * here to attempt to avoid that.
872
 */
873 874 875
#define LAST_INO_BATCH 1024
static DEFINE_PER_CPU(unsigned int, last_ino);

876
unsigned int get_next_ino(void)
877
{
878 879
	unsigned int *p = &get_cpu_var(last_ino);
	unsigned int res = *p;
880

881 882 883 884 885 886 887 888 889 890 891 892
#ifdef CONFIG_SMP
	if (unlikely((res & (LAST_INO_BATCH-1)) == 0)) {
		static atomic_t shared_last_ino;
		int next = atomic_add_return(LAST_INO_BATCH, &shared_last_ino);

		res = next - LAST_INO_BATCH;
	}
#endif

	*p = ++res;
	put_cpu_var(last_ino);
	return res;
893
}
894
EXPORT_SYMBOL(get_next_ino);
895

896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918
/**
 *	new_inode_pseudo 	- obtain an inode
 *	@sb: superblock
 *
 *	Allocates a new inode for given superblock.
 *	Inode wont be chained in superblock s_inodes list
 *	This means :
 *	- fs can't be unmount
 *	- quotas, fsnotify, writeback can't work
 */
struct inode *new_inode_pseudo(struct super_block *sb)
{
	struct inode *inode = alloc_inode(sb);

	if (inode) {
		spin_lock(&inode->i_lock);
		inode->i_state = 0;
		spin_unlock(&inode->i_lock);
		INIT_LIST_HEAD(&inode->i_sb_list);
	}
	return inode;
}

L
Linus Torvalds 已提交
919 920 921 922
/**
 *	new_inode 	- obtain an inode
 *	@sb: superblock
 *
923
 *	Allocates a new inode for given superblock. The default gfp_mask
H
Hugh Dickins 已提交
924
 *	for allocations related to inode->i_mapping is GFP_HIGHUSER_MOVABLE.
925 926 927 928 929
 *	If HIGHMEM pages are unsuitable or it is known that pages allocated
 *	for the page cache are not reclaimable or migratable,
 *	mapping_set_gfp_mask() must be called with suitable flags on the
 *	newly created inode's mapping
 *
L
Linus Torvalds 已提交
930 931 932
 */
struct inode *new_inode(struct super_block *sb)
{
933
	struct inode *inode;
L
Linus Torvalds 已提交
934

935
	spin_lock_prefetch(&inode_sb_list_lock);
936

937 938
	inode = new_inode_pseudo(sb);
	if (inode)
939
		inode_sb_list_add(inode);
L
Linus Torvalds 已提交
940 941 942 943
	return inode;
}
EXPORT_SYMBOL(new_inode);

944
#ifdef CONFIG_DEBUG_LOCK_ALLOC
945 946
void lockdep_annotate_inode_mutex_key(struct inode *inode)
{
947
	if (S_ISDIR(inode->i_mode)) {
948 949
		struct file_system_type *type = inode->i_sb->s_type;

950
		/* Set new key only if filesystem hasn't already changed it */
951
		if (lockdep_match_class(&inode->i_mutex, &type->i_mutex_key)) {
952 953 954 955 956 957 958 959
			/*
			 * ensure nobody is actually holding i_mutex
			 */
			mutex_destroy(&inode->i_mutex);
			mutex_init(&inode->i_mutex);
			lockdep_set_class(&inode->i_mutex,
					  &type->i_mutex_dir_key);
		}
960
	}
961 962
}
EXPORT_SYMBOL(lockdep_annotate_inode_mutex_key);
963
#endif
964 965 966 967 968 969 970 971 972 973 974

/**
 * unlock_new_inode - clear the I_NEW state and wake up any waiters
 * @inode:	new inode to unlock
 *
 * Called when the inode is fully initialised to clear the new state of the
 * inode and wake up anyone waiting for the inode to finish initialisation.
 */
void unlock_new_inode(struct inode *inode)
{
	lockdep_annotate_inode_mutex_key(inode);
975
	spin_lock(&inode->i_lock);
C
Christoph Hellwig 已提交
976 977
	WARN_ON(!(inode->i_state & I_NEW));
	inode->i_state &= ~I_NEW;
A
Al Viro 已提交
978
	smp_mb();
979 980
	wake_up_bit(&inode->i_state, __I_NEW);
	spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
981 982 983
}
EXPORT_SYMBOL(unlock_new_inode);

C
Christoph Hellwig 已提交
984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999
/**
 * iget5_locked - obtain an inode from a mounted file system
 * @sb:		super block of file system
 * @hashval:	hash value (usually inode number) to get
 * @test:	callback used for comparisons between inodes
 * @set:	callback used to initialize a new struct inode
 * @data:	opaque data pointer to pass to @test and @set
 *
 * Search for the inode specified by @hashval and @data in the inode cache,
 * and if present it is return it with an increased reference count. This is
 * a generalized version of iget_locked() for file systems where the inode
 * number is not sufficient for unique identification of an inode.
 *
 * If the inode is not in cache, allocate a new inode and return it locked,
 * hashed, and with the I_NEW flag set. The file system gets to fill it in
 * before unlocking it via unlock_new_inode().
L
Linus Torvalds 已提交
1000
 *
C
Christoph Hellwig 已提交
1001 1002
 * Note both @test and @set are called with the inode_hash_lock held, so can't
 * sleep.
L
Linus Torvalds 已提交
1003
 */
C
Christoph Hellwig 已提交
1004 1005 1006
struct inode *iget5_locked(struct super_block *sb, unsigned long hashval,
		int (*test)(struct inode *, void *),
		int (*set)(struct inode *, void *), void *data)
L
Linus Torvalds 已提交
1007
{
C
Christoph Hellwig 已提交
1008
	struct hlist_head *head = inode_hashtable + hash(sb, hashval);
1009
	struct inode *inode;
L
Linus Torvalds 已提交
1010

C
Christoph Hellwig 已提交
1011 1012 1013 1014 1015 1016 1017 1018 1019
	spin_lock(&inode_hash_lock);
	inode = find_inode(sb, head, test, data);
	spin_unlock(&inode_hash_lock);

	if (inode) {
		wait_on_inode(inode);
		return inode;
	}

L
Linus Torvalds 已提交
1020 1021
	inode = alloc_inode(sb);
	if (inode) {
1022
		struct inode *old;
L
Linus Torvalds 已提交
1023

1024
		spin_lock(&inode_hash_lock);
L
Linus Torvalds 已提交
1025 1026 1027 1028 1029 1030
		/* We released the lock, so.. */
		old = find_inode(sb, head, test, data);
		if (!old) {
			if (set(inode, data))
				goto set_failed;

1031 1032
			spin_lock(&inode->i_lock);
			inode->i_state = I_NEW;
1033
			hlist_add_head(&inode->i_hash, head);
1034
			spin_unlock(&inode->i_lock);
1035
			inode_sb_list_add(inode);
1036
			spin_unlock(&inode_hash_lock);
L
Linus Torvalds 已提交
1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048

			/* Return the locked inode with I_NEW set, the
			 * caller is responsible for filling in the contents
			 */
			return inode;
		}

		/*
		 * Uhhuh, somebody else created the same inode under
		 * us. Use the old inode instead of the one we just
		 * allocated.
		 */
1049
		spin_unlock(&inode_hash_lock);
L
Linus Torvalds 已提交
1050 1051 1052 1053 1054 1055 1056
		destroy_inode(inode);
		inode = old;
		wait_on_inode(inode);
	}
	return inode;

set_failed:
1057
	spin_unlock(&inode_hash_lock);
L
Linus Torvalds 已提交
1058 1059 1060
	destroy_inode(inode);
	return NULL;
}
C
Christoph Hellwig 已提交
1061
EXPORT_SYMBOL(iget5_locked);
L
Linus Torvalds 已提交
1062

C
Christoph Hellwig 已提交
1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074
/**
 * iget_locked - obtain an inode from a mounted file system
 * @sb:		super block of file system
 * @ino:	inode number to get
 *
 * Search for the inode specified by @ino in the inode cache and if present
 * return it with an increased reference count. This is for file systems
 * where the inode number is sufficient for unique identification of an inode.
 *
 * If the inode is not in cache, allocate a new inode and return it locked,
 * hashed, and with the I_NEW flag set.  The file system gets to fill it in
 * before unlocking it via unlock_new_inode().
L
Linus Torvalds 已提交
1075
 */
C
Christoph Hellwig 已提交
1076
struct inode *iget_locked(struct super_block *sb, unsigned long ino)
L
Linus Torvalds 已提交
1077
{
C
Christoph Hellwig 已提交
1078
	struct hlist_head *head = inode_hashtable + hash(sb, ino);
1079
	struct inode *inode;
L
Linus Torvalds 已提交
1080

C
Christoph Hellwig 已提交
1081 1082 1083 1084 1085 1086 1087 1088
	spin_lock(&inode_hash_lock);
	inode = find_inode_fast(sb, head, ino);
	spin_unlock(&inode_hash_lock);
	if (inode) {
		wait_on_inode(inode);
		return inode;
	}

L
Linus Torvalds 已提交
1089 1090
	inode = alloc_inode(sb);
	if (inode) {
1091
		struct inode *old;
L
Linus Torvalds 已提交
1092

1093
		spin_lock(&inode_hash_lock);
L
Linus Torvalds 已提交
1094 1095 1096 1097
		/* We released the lock, so.. */
		old = find_inode_fast(sb, head, ino);
		if (!old) {
			inode->i_ino = ino;
1098 1099
			spin_lock(&inode->i_lock);
			inode->i_state = I_NEW;
1100
			hlist_add_head(&inode->i_hash, head);
1101
			spin_unlock(&inode->i_lock);
1102
			inode_sb_list_add(inode);
1103
			spin_unlock(&inode_hash_lock);
L
Linus Torvalds 已提交
1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115

			/* Return the locked inode with I_NEW set, the
			 * caller is responsible for filling in the contents
			 */
			return inode;
		}

		/*
		 * Uhhuh, somebody else created the same inode under
		 * us. Use the old inode instead of the one we just
		 * allocated.
		 */
1116
		spin_unlock(&inode_hash_lock);
L
Linus Torvalds 已提交
1117 1118 1119 1120 1121 1122
		destroy_inode(inode);
		inode = old;
		wait_on_inode(inode);
	}
	return inode;
}
C
Christoph Hellwig 已提交
1123
EXPORT_SYMBOL(iget_locked);
L
Linus Torvalds 已提交
1124

1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136
/*
 * search the inode cache for a matching inode number.
 * If we find one, then the inode number we are trying to
 * allocate is not unique and so we should not use it.
 *
 * Returns 1 if the inode number is unique, 0 if it is not.
 */
static int test_inode_iunique(struct super_block *sb, unsigned long ino)
{
	struct hlist_head *b = inode_hashtable + hash(sb, ino);
	struct inode *inode;

1137
	spin_lock(&inode_hash_lock);
1138
	hlist_for_each_entry(inode, b, i_hash) {
1139 1140
		if (inode->i_ino == ino && inode->i_sb == sb) {
			spin_unlock(&inode_hash_lock);
1141
			return 0;
1142
		}
1143
	}
1144
	spin_unlock(&inode_hash_lock);
1145 1146 1147 1148

	return 1;
}

L
Linus Torvalds 已提交
1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164
/**
 *	iunique - get a unique inode number
 *	@sb: superblock
 *	@max_reserved: highest reserved inode number
 *
 *	Obtain an inode number that is unique on the system for a given
 *	superblock. This is used by file systems that have no natural
 *	permanent inode numbering system. An inode number is returned that
 *	is higher than the reserved limit but unique.
 *
 *	BUGS:
 *	With a large number of inodes live on the file system this function
 *	currently becomes quite slow.
 */
ino_t iunique(struct super_block *sb, ino_t max_reserved)
{
1165 1166 1167 1168 1169
	/*
	 * On a 32bit, non LFS stat() call, glibc will generate an EOVERFLOW
	 * error if st_ino won't fit in target struct field. Use 32bit counter
	 * here to attempt to avoid that.
	 */
1170
	static DEFINE_SPINLOCK(iunique_lock);
1171
	static unsigned int counter;
L
Linus Torvalds 已提交
1172
	ino_t res;
1173

1174
	spin_lock(&iunique_lock);
1175 1176 1177
	do {
		if (counter <= max_reserved)
			counter = max_reserved + 1;
L
Linus Torvalds 已提交
1178
		res = counter++;
1179 1180
	} while (!test_inode_iunique(sb, res));
	spin_unlock(&iunique_lock);
L
Linus Torvalds 已提交
1181

1182 1183
	return res;
}
L
Linus Torvalds 已提交
1184 1185 1186 1187
EXPORT_SYMBOL(iunique);

struct inode *igrab(struct inode *inode)
{
1188 1189
	spin_lock(&inode->i_lock);
	if (!(inode->i_state & (I_FREEING|I_WILL_FREE))) {
L
Linus Torvalds 已提交
1190
		__iget(inode);
1191 1192 1193
		spin_unlock(&inode->i_lock);
	} else {
		spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
1194 1195 1196 1197 1198 1199
		/*
		 * Handle the case where s_op->clear_inode is not been
		 * called yet, and somebody is calling igrab
		 * while the inode is getting freed.
		 */
		inode = NULL;
1200
	}
L
Linus Torvalds 已提交
1201 1202 1203 1204 1205
	return inode;
}
EXPORT_SYMBOL(igrab);

/**
C
Christoph Hellwig 已提交
1206
 * ilookup5_nowait - search for an inode in the inode cache
L
Linus Torvalds 已提交
1207
 * @sb:		super block of file system to search
C
Christoph Hellwig 已提交
1208
 * @hashval:	hash value (usually inode number) to search for
L
Linus Torvalds 已提交
1209 1210 1211
 * @test:	callback used for comparisons between inodes
 * @data:	opaque data pointer to pass to @test
 *
C
Christoph Hellwig 已提交
1212
 * Search for the inode specified by @hashval and @data in the inode cache.
L
Linus Torvalds 已提交
1213 1214 1215
 * If the inode is in the cache, the inode is returned with an incremented
 * reference count.
 *
C
Christoph Hellwig 已提交
1216 1217
 * Note: I_NEW is not waited upon so you have to be very careful what you do
 * with the returned inode.  You probably should be using ilookup5() instead.
L
Linus Torvalds 已提交
1218
 *
1219
 * Note2: @test is called with the inode_hash_lock held, so can't sleep.
L
Linus Torvalds 已提交
1220
 */
C
Christoph Hellwig 已提交
1221 1222
struct inode *ilookup5_nowait(struct super_block *sb, unsigned long hashval,
		int (*test)(struct inode *, void *), void *data)
L
Linus Torvalds 已提交
1223
{
C
Christoph Hellwig 已提交
1224
	struct hlist_head *head = inode_hashtable + hash(sb, hashval);
L
Linus Torvalds 已提交
1225 1226
	struct inode *inode;

1227
	spin_lock(&inode_hash_lock);
L
Linus Torvalds 已提交
1228
	inode = find_inode(sb, head, test, data);
1229
	spin_unlock(&inode_hash_lock);
1230

C
Christoph Hellwig 已提交
1231
	return inode;
1232 1233 1234 1235 1236 1237 1238 1239 1240 1241
}
EXPORT_SYMBOL(ilookup5_nowait);

/**
 * ilookup5 - search for an inode in the inode cache
 * @sb:		super block of file system to search
 * @hashval:	hash value (usually inode number) to search for
 * @test:	callback used for comparisons between inodes
 * @data:	opaque data pointer to pass to @test
 *
C
Christoph Hellwig 已提交
1242 1243 1244
 * Search for the inode specified by @hashval and @data in the inode cache,
 * and if the inode is in the cache, return the inode with an incremented
 * reference count.  Waits on I_NEW before returning the inode.
1245
 * returned with an incremented reference count.
L
Linus Torvalds 已提交
1246
 *
C
Christoph Hellwig 已提交
1247 1248
 * This is a generalized version of ilookup() for file systems where the
 * inode number is not sufficient for unique identification of an inode.
L
Linus Torvalds 已提交
1249
 *
C
Christoph Hellwig 已提交
1250
 * Note: @test is called with the inode_hash_lock held, so can't sleep.
L
Linus Torvalds 已提交
1251 1252 1253 1254
 */
struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
		int (*test)(struct inode *, void *), void *data)
{
C
Christoph Hellwig 已提交
1255
	struct inode *inode = ilookup5_nowait(sb, hashval, test, data);
L
Linus Torvalds 已提交
1256

C
Christoph Hellwig 已提交
1257 1258 1259
	if (inode)
		wait_on_inode(inode);
	return inode;
L
Linus Torvalds 已提交
1260 1261 1262 1263 1264 1265 1266 1267
}
EXPORT_SYMBOL(ilookup5);

/**
 * ilookup - search for an inode in the inode cache
 * @sb:		super block of file system to search
 * @ino:	inode number to search for
 *
C
Christoph Hellwig 已提交
1268 1269
 * Search for the inode @ino in the inode cache, and if the inode is in the
 * cache, the inode is returned with an incremented reference count.
L
Linus Torvalds 已提交
1270 1271 1272 1273 1274 1275
 */
struct inode *ilookup(struct super_block *sb, unsigned long ino)
{
	struct hlist_head *head = inode_hashtable + hash(sb, ino);
	struct inode *inode;

C
Christoph Hellwig 已提交
1276 1277 1278
	spin_lock(&inode_hash_lock);
	inode = find_inode_fast(sb, head, ino);
	spin_unlock(&inode_hash_lock);
L
Linus Torvalds 已提交
1279 1280

	if (inode)
C
Christoph Hellwig 已提交
1281 1282
		wait_on_inode(inode);
	return inode;
L
Linus Torvalds 已提交
1283
}
C
Christoph Hellwig 已提交
1284
EXPORT_SYMBOL(ilookup);
L
Linus Torvalds 已提交
1285

A
Al Viro 已提交
1286 1287 1288 1289 1290 1291 1292
int insert_inode_locked(struct inode *inode)
{
	struct super_block *sb = inode->i_sb;
	ino_t ino = inode->i_ino;
	struct hlist_head *head = inode_hashtable + hash(sb, ino);

	while (1) {
1293
		struct inode *old = NULL;
1294
		spin_lock(&inode_hash_lock);
1295
		hlist_for_each_entry(old, head, i_hash) {
1296 1297 1298 1299
			if (old->i_ino != ino)
				continue;
			if (old->i_sb != sb)
				continue;
1300 1301 1302
			spin_lock(&old->i_lock);
			if (old->i_state & (I_FREEING|I_WILL_FREE)) {
				spin_unlock(&old->i_lock);
1303
				continue;
1304
			}
1305 1306
			break;
		}
1307
		if (likely(!old)) {
1308 1309
			spin_lock(&inode->i_lock);
			inode->i_state |= I_NEW;
A
Al Viro 已提交
1310
			hlist_add_head(&inode->i_hash, head);
1311
			spin_unlock(&inode->i_lock);
1312
			spin_unlock(&inode_hash_lock);
A
Al Viro 已提交
1313 1314 1315
			return 0;
		}
		__iget(old);
1316
		spin_unlock(&old->i_lock);
1317
		spin_unlock(&inode_hash_lock);
A
Al Viro 已提交
1318
		wait_on_inode(old);
A
Al Viro 已提交
1319
		if (unlikely(!inode_unhashed(old))) {
A
Al Viro 已提交
1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334
			iput(old);
			return -EBUSY;
		}
		iput(old);
	}
}
EXPORT_SYMBOL(insert_inode_locked);

int insert_inode_locked4(struct inode *inode, unsigned long hashval,
		int (*test)(struct inode *, void *), void *data)
{
	struct super_block *sb = inode->i_sb;
	struct hlist_head *head = inode_hashtable + hash(sb, hashval);

	while (1) {
1335 1336
		struct inode *old = NULL;

1337
		spin_lock(&inode_hash_lock);
1338
		hlist_for_each_entry(old, head, i_hash) {
1339 1340 1341 1342
			if (old->i_sb != sb)
				continue;
			if (!test(old, data))
				continue;
1343 1344 1345
			spin_lock(&old->i_lock);
			if (old->i_state & (I_FREEING|I_WILL_FREE)) {
				spin_unlock(&old->i_lock);
1346
				continue;
1347
			}
1348 1349
			break;
		}
1350
		if (likely(!old)) {
1351 1352
			spin_lock(&inode->i_lock);
			inode->i_state |= I_NEW;
A
Al Viro 已提交
1353
			hlist_add_head(&inode->i_hash, head);
1354
			spin_unlock(&inode->i_lock);
1355
			spin_unlock(&inode_hash_lock);
A
Al Viro 已提交
1356 1357 1358
			return 0;
		}
		__iget(old);
1359
		spin_unlock(&old->i_lock);
1360
		spin_unlock(&inode_hash_lock);
A
Al Viro 已提交
1361
		wait_on_inode(old);
A
Al Viro 已提交
1362
		if (unlikely(!inode_unhashed(old))) {
A
Al Viro 已提交
1363 1364 1365 1366 1367 1368 1369 1370
			iput(old);
			return -EBUSY;
		}
		iput(old);
	}
}
EXPORT_SYMBOL(insert_inode_locked4);

L
Linus Torvalds 已提交
1371

1372 1373 1374 1375 1376 1377 1378 1379 1380
int generic_delete_inode(struct inode *inode)
{
	return 1;
}
EXPORT_SYMBOL(generic_delete_inode);

/*
 * Called when we're dropping the last reference
 * to an inode.
1381
 *
1382 1383 1384 1385 1386
 * Call the FS "drop_inode()" function, defaulting to
 * the legacy UNIX filesystem behaviour.  If it tells
 * us to evict inode, do so.  Otherwise, retain inode
 * in cache if fs is alive, sync and evict if fs is
 * shutting down.
1387
 */
1388
static void iput_final(struct inode *inode)
L
Linus Torvalds 已提交
1389 1390
{
	struct super_block *sb = inode->i_sb;
1391 1392 1393
	const struct super_operations *op = inode->i_sb->s_op;
	int drop;

1394 1395
	WARN_ON(inode->i_state & I_NEW);

1396
	if (op->drop_inode)
1397 1398 1399
		drop = op->drop_inode(inode);
	else
		drop = generic_drop_inode(inode);
L
Linus Torvalds 已提交
1400

D
Dave Chinner 已提交
1401 1402
	if (!drop && (sb->s_flags & MS_ACTIVE)) {
		inode->i_state |= I_REFERENCED;
1403
		inode_add_lru(inode);
D
Dave Chinner 已提交
1404 1405 1406 1407
		spin_unlock(&inode->i_lock);
		return;
	}

1408
	if (!drop) {
1409
		inode->i_state |= I_WILL_FREE;
1410
		spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
1411
		write_inode_now(inode, 1);
1412
		spin_lock(&inode->i_lock);
1413
		WARN_ON(inode->i_state & I_NEW);
1414
		inode->i_state &= ~I_WILL_FREE;
L
Linus Torvalds 已提交
1415
	}
N
Nick Piggin 已提交
1416

1417
	inode->i_state |= I_FREEING;
1418 1419
	if (!list_empty(&inode->i_lru))
		inode_lru_list_del(inode);
D
Dave Chinner 已提交
1420 1421
	spin_unlock(&inode->i_lock);

1422
	evict(inode);
L
Linus Torvalds 已提交
1423 1424 1425
}

/**
1426
 *	iput	- put an inode
L
Linus Torvalds 已提交
1427 1428 1429 1430 1431 1432 1433 1434 1435 1436
 *	@inode: inode to put
 *
 *	Puts an inode, dropping its usage count. If the inode use count hits
 *	zero, the inode is then freed and may also be destroyed.
 *
 *	Consequently, iput() can sleep.
 */
void iput(struct inode *inode)
{
	if (inode) {
A
Al Viro 已提交
1437
		BUG_ON(inode->i_state & I_CLEAR);
L
Linus Torvalds 已提交
1438

1439
		if (atomic_dec_and_lock(&inode->i_count, &inode->i_lock))
L
Linus Torvalds 已提交
1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452
			iput_final(inode);
	}
}
EXPORT_SYMBOL(iput);

/**
 *	bmap	- find a block number in a file
 *	@inode: inode of file
 *	@block: block to find
 *
 *	Returns the block number on the device holding the inode that
 *	is the disk block number for the block of the file requested.
 *	That is, asked for block 4 of inode 1 the function will return the
1453
 *	disk block relative to the disk start that holds that block of the
L
Linus Torvalds 已提交
1454 1455
 *	file.
 */
1456
sector_t bmap(struct inode *inode, sector_t block)
L
Linus Torvalds 已提交
1457 1458 1459 1460 1461 1462 1463 1464
{
	sector_t res = 0;
	if (inode->i_mapping->a_ops->bmap)
		res = inode->i_mapping->a_ops->bmap(inode->i_mapping, block);
	return res;
}
EXPORT_SYMBOL(bmap);

1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498
/*
 * With relative atime, only update atime if the previous atime is
 * earlier than either the ctime or mtime or if at least a day has
 * passed since the last atime update.
 */
static int relatime_need_update(struct vfsmount *mnt, struct inode *inode,
			     struct timespec now)
{

	if (!(mnt->mnt_flags & MNT_RELATIME))
		return 1;
	/*
	 * Is mtime younger than atime? If yes, update atime:
	 */
	if (timespec_compare(&inode->i_mtime, &inode->i_atime) >= 0)
		return 1;
	/*
	 * Is ctime younger than atime? If yes, update atime:
	 */
	if (timespec_compare(&inode->i_ctime, &inode->i_atime) >= 0)
		return 1;

	/*
	 * Is the previous atime value older than a day? If yes,
	 * update atime:
	 */
	if ((long)(now.tv_sec - inode->i_atime.tv_sec) >= 24*60*60)
		return 1;
	/*
	 * Good, we can skip the atime update:
	 */
	return 0;
}

1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519
/*
 * This does the actual work of updating an inodes time or version.  Must have
 * had called mnt_want_write() before calling this.
 */
static int update_time(struct inode *inode, struct timespec *time, int flags)
{
	if (inode->i_op->update_time)
		return inode->i_op->update_time(inode, time, flags);

	if (flags & S_ATIME)
		inode->i_atime = *time;
	if (flags & S_VERSION)
		inode_inc_iversion(inode);
	if (flags & S_CTIME)
		inode->i_ctime = *time;
	if (flags & S_MTIME)
		inode->i_mtime = *time;
	mark_inode_dirty_sync(inode);
	return 0;
}

L
Linus Torvalds 已提交
1520
/**
C
Christoph Hellwig 已提交
1521
 *	touch_atime	-	update the access time
1522
 *	@path: the &struct path to update
L
Linus Torvalds 已提交
1523 1524 1525 1526 1527
 *
 *	Update the accessed time on an inode and mark it for writeback.
 *	This function automatically handles read only file systems and media,
 *	as well as the "noatime" flag and inode specific "noatime" markers.
 */
A
Al Viro 已提交
1528
void touch_atime(const struct path *path)
L
Linus Torvalds 已提交
1529
{
A
Al Viro 已提交
1530 1531
	struct vfsmount *mnt = path->mnt;
	struct inode *inode = path->dentry->d_inode;
L
Linus Torvalds 已提交
1532 1533
	struct timespec now;

1534
	if (inode->i_flags & S_NOATIME)
A
Andi Kleen 已提交
1535
		return;
1536
	if (IS_NOATIME(inode))
A
Andi Kleen 已提交
1537
		return;
A
Andrew Morton 已提交
1538
	if ((inode->i_sb->s_flags & MS_NODIRATIME) && S_ISDIR(inode->i_mode))
A
Andi Kleen 已提交
1539
		return;
V
Valerie Henson 已提交
1540

1541
	if (mnt->mnt_flags & MNT_NOATIME)
A
Andi Kleen 已提交
1542
		return;
1543
	if ((mnt->mnt_flags & MNT_NODIRATIME) && S_ISDIR(inode->i_mode))
A
Andi Kleen 已提交
1544
		return;
L
Linus Torvalds 已提交
1545 1546

	now = current_fs_time(inode->i_sb);
1547 1548

	if (!relatime_need_update(mnt, inode, now))
A
Andi Kleen 已提交
1549
		return;
1550

V
Valerie Henson 已提交
1551
	if (timespec_equal(&inode->i_atime, &now))
A
Andi Kleen 已提交
1552 1553
		return;

1554
	if (!sb_start_write_trylock(inode->i_sb))
A
Andi Kleen 已提交
1555
		return;
V
Valerie Henson 已提交
1556

1557 1558
	if (__mnt_want_write(mnt))
		goto skip_update;
1559 1560 1561 1562 1563 1564
	/*
	 * File systems can error out when updating inodes if they need to
	 * allocate new space to modify an inode (such is the case for
	 * Btrfs), but since we touch atime while walking down the path we
	 * really don't care if we failed to update the atime of the file,
	 * so just ignore the return value.
1565 1566
	 * We may also fail on filesystems that have the ability to make parts
	 * of the fs read only, e.g. subvolumes in Btrfs.
1567 1568
	 */
	update_time(inode, &now, S_ATIME);
1569 1570 1571
	__mnt_drop_write(mnt);
skip_update:
	sb_end_write(inode->i_sb);
L
Linus Torvalds 已提交
1572
}
C
Christoph Hellwig 已提交
1573
EXPORT_SYMBOL(touch_atime);
L
Linus Torvalds 已提交
1574

1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639
/*
 * The logic we want is
 *
 *	if suid or (sgid and xgrp)
 *		remove privs
 */
int should_remove_suid(struct dentry *dentry)
{
	umode_t mode = dentry->d_inode->i_mode;
	int kill = 0;

	/* suid always must be killed */
	if (unlikely(mode & S_ISUID))
		kill = ATTR_KILL_SUID;

	/*
	 * sgid without any exec bits is just a mandatory locking mark; leave
	 * it alone.  If some exec bits are set, it's a real sgid; kill it.
	 */
	if (unlikely((mode & S_ISGID) && (mode & S_IXGRP)))
		kill |= ATTR_KILL_SGID;

	if (unlikely(kill && !capable(CAP_FSETID) && S_ISREG(mode)))
		return kill;

	return 0;
}
EXPORT_SYMBOL(should_remove_suid);

static int __remove_suid(struct dentry *dentry, int kill)
{
	struct iattr newattrs;

	newattrs.ia_valid = ATTR_FORCE | kill;
	return notify_change(dentry, &newattrs);
}

int file_remove_suid(struct file *file)
{
	struct dentry *dentry = file->f_path.dentry;
	struct inode *inode = dentry->d_inode;
	int killsuid;
	int killpriv;
	int error = 0;

	/* Fast path for nothing security related */
	if (IS_NOSEC(inode))
		return 0;

	killsuid = should_remove_suid(dentry);
	killpriv = security_inode_need_killpriv(dentry);

	if (killpriv < 0)
		return killpriv;
	if (killpriv)
		error = security_inode_killpriv(dentry);
	if (!error && killsuid)
		error = __remove_suid(dentry, killsuid);
	if (!error && (inode->i_sb->s_flags & MS_NOSEC))
		inode->i_flags |= S_NOSEC;

	return error;
}
EXPORT_SYMBOL(file_remove_suid);

L
Linus Torvalds 已提交
1640
/**
1641 1642
 *	file_update_time	-	update mtime and ctime time
 *	@file: file accessed
L
Linus Torvalds 已提交
1643
 *
1644 1645 1646 1647
 *	Update the mtime and ctime members of an inode and mark the inode
 *	for writeback.  Note that this function is meant exclusively for
 *	usage in the file write path of filesystems, and filesystems may
 *	choose to explicitly ignore update via this function with the
1648
 *	S_NOCMTIME inode flag, e.g. for network filesystem where these
1649 1650
 *	timestamps are handled by the server.  This can return an error for
 *	file systems who need to allocate space in order to update an inode.
L
Linus Torvalds 已提交
1651 1652
 */

1653
int file_update_time(struct file *file)
L
Linus Torvalds 已提交
1654
{
A
Al Viro 已提交
1655
	struct inode *inode = file_inode(file);
L
Linus Torvalds 已提交
1656
	struct timespec now;
1657 1658
	int sync_it = 0;
	int ret;
L
Linus Torvalds 已提交
1659

A
Andi Kleen 已提交
1660
	/* First try to exhaust all avenues to not sync */
L
Linus Torvalds 已提交
1661
	if (IS_NOCMTIME(inode))
1662
		return 0;
1663

L
Linus Torvalds 已提交
1664
	now = current_fs_time(inode->i_sb);
A
Andi Kleen 已提交
1665 1666
	if (!timespec_equal(&inode->i_mtime, &now))
		sync_it = S_MTIME;
L
Linus Torvalds 已提交
1667

A
Andi Kleen 已提交
1668 1669
	if (!timespec_equal(&inode->i_ctime, &now))
		sync_it |= S_CTIME;
1670

A
Andi Kleen 已提交
1671 1672
	if (IS_I_VERSION(inode))
		sync_it |= S_VERSION;
1673

A
Andi Kleen 已提交
1674
	if (!sync_it)
1675
		return 0;
A
Andi Kleen 已提交
1676 1677

	/* Finally allowed to write? Takes lock. */
1678
	if (__mnt_want_write_file(file))
1679
		return 0;
A
Andi Kleen 已提交
1680

1681
	ret = update_time(inode, &now, sync_it);
1682
	__mnt_drop_write_file(file);
1683 1684

	return ret;
L
Linus Torvalds 已提交
1685
}
1686
EXPORT_SYMBOL(file_update_time);
L
Linus Torvalds 已提交
1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702

int inode_needs_sync(struct inode *inode)
{
	if (IS_SYNC(inode))
		return 1;
	if (S_ISDIR(inode->i_mode) && IS_DIRSYNC(inode))
		return 1;
	return 0;
}
EXPORT_SYMBOL(inode_needs_sync);

int inode_wait(void *word)
{
	schedule();
	return 0;
}
1703
EXPORT_SYMBOL(inode_wait);
L
Linus Torvalds 已提交
1704 1705

/*
1706 1707 1708 1709 1710 1711
 * If we try to find an inode in the inode hash while it is being
 * deleted, we have to wait until the filesystem completes its
 * deletion before reporting that it isn't found.  This function waits
 * until the deletion _might_ have completed.  Callers are responsible
 * to recheck inode state.
 *
C
Christoph Hellwig 已提交
1712
 * It doesn't matter if I_NEW is not set initially, a call to
1713 1714
 * wake_up_bit(&inode->i_state, __I_NEW) after removing from the hash list
 * will DTRT.
L
Linus Torvalds 已提交
1715 1716 1717 1718
 */
static void __wait_on_freeing_inode(struct inode *inode)
{
	wait_queue_head_t *wq;
C
Christoph Hellwig 已提交
1719 1720
	DEFINE_WAIT_BIT(wait, &inode->i_state, __I_NEW);
	wq = bit_waitqueue(&inode->i_state, __I_NEW);
L
Linus Torvalds 已提交
1721
	prepare_to_wait(wq, &wait.wait, TASK_UNINTERRUPTIBLE);
1722
	spin_unlock(&inode->i_lock);
1723
	spin_unlock(&inode_hash_lock);
L
Linus Torvalds 已提交
1724 1725
	schedule();
	finish_wait(wq, &wait.wait);
1726
	spin_lock(&inode_hash_lock);
L
Linus Torvalds 已提交
1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743
}

static __initdata unsigned long ihash_entries;
static int __init set_ihash_entries(char *str)
{
	if (!str)
		return 0;
	ihash_entries = simple_strtoul(str, &str, 0);
	return 1;
}
__setup("ihash_entries=", set_ihash_entries);

/*
 * Initialize the waitqueues and inode hash table.
 */
void __init inode_init_early(void)
{
1744
	unsigned int loop;
L
Linus Torvalds 已提交
1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759

	/* If hashes are distributed across NUMA nodes, defer
	 * hash allocation until vmalloc space is available.
	 */
	if (hashdist)
		return;

	inode_hashtable =
		alloc_large_system_hash("Inode-cache",
					sizeof(struct hlist_head),
					ihash_entries,
					14,
					HASH_EARLY,
					&i_hash_shift,
					&i_hash_mask,
1760
					0,
L
Linus Torvalds 已提交
1761 1762
					0);

1763
	for (loop = 0; loop < (1U << i_hash_shift); loop++)
L
Linus Torvalds 已提交
1764 1765 1766
		INIT_HLIST_HEAD(&inode_hashtable[loop]);
}

1767
void __init inode_init(void)
L
Linus Torvalds 已提交
1768
{
1769
	unsigned int loop;
L
Linus Torvalds 已提交
1770 1771

	/* inode slab cache */
1772 1773 1774 1775 1776
	inode_cachep = kmem_cache_create("inode_cache",
					 sizeof(struct inode),
					 0,
					 (SLAB_RECLAIM_ACCOUNT|SLAB_PANIC|
					 SLAB_MEM_SPREAD),
1777
					 init_once);
L
Linus Torvalds 已提交
1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790

	/* Hash may have been set up in inode_init_early */
	if (!hashdist)
		return;

	inode_hashtable =
		alloc_large_system_hash("Inode-cache",
					sizeof(struct hlist_head),
					ihash_entries,
					14,
					0,
					&i_hash_shift,
					&i_hash_mask,
1791
					0,
L
Linus Torvalds 已提交
1792 1793
					0);

1794
	for (loop = 0; loop < (1U << i_hash_shift); loop++)
L
Linus Torvalds 已提交
1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807
		INIT_HLIST_HEAD(&inode_hashtable[loop]);
}

void init_special_inode(struct inode *inode, umode_t mode, dev_t rdev)
{
	inode->i_mode = mode;
	if (S_ISCHR(mode)) {
		inode->i_fop = &def_chr_fops;
		inode->i_rdev = rdev;
	} else if (S_ISBLK(mode)) {
		inode->i_fop = &def_blk_fops;
		inode->i_rdev = rdev;
	} else if (S_ISFIFO(mode))
1808
		inode->i_fop = &pipefifo_fops;
L
Linus Torvalds 已提交
1809 1810 1811
	else if (S_ISSOCK(mode))
		inode->i_fop = &bad_sock_fops;
	else
1812 1813 1814
		printk(KERN_DEBUG "init_special_inode: bogus i_mode (%o) for"
				  " inode %s:%lu\n", mode, inode->i_sb->s_id,
				  inode->i_ino);
L
Linus Torvalds 已提交
1815 1816
}
EXPORT_SYMBOL(init_special_inode);
1817 1818

/**
1819
 * inode_init_owner - Init uid,gid,mode for new inode according to posix standards
1820 1821 1822 1823 1824
 * @inode: New inode
 * @dir: Directory inode
 * @mode: mode of the new inode
 */
void inode_init_owner(struct inode *inode, const struct inode *dir,
A
Al Viro 已提交
1825
			umode_t mode)
1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836
{
	inode->i_uid = current_fsuid();
	if (dir && dir->i_mode & S_ISGID) {
		inode->i_gid = dir->i_gid;
		if (S_ISDIR(mode))
			mode |= S_ISGID;
	} else
		inode->i_gid = current_fsgid();
	inode->i_mode = mode;
}
EXPORT_SYMBOL(inode_init_owner);
1837

1838 1839 1840 1841 1842 1843
/**
 * inode_owner_or_capable - check current task permissions to inode
 * @inode: inode being checked
 *
 * Return true if current either has CAP_FOWNER to the inode, or
 * owns the file.
1844
 */
1845
bool inode_owner_or_capable(const struct inode *inode)
1846
{
1847
	if (uid_eq(current_fsuid(), inode->i_uid))
1848
		return true;
1849
	if (inode_capable(inode, CAP_FOWNER))
1850 1851 1852
		return true;
	return false;
}
1853
EXPORT_SYMBOL(inode_owner_or_capable);
1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900

/*
 * Direct i/o helper functions
 */
static void __inode_dio_wait(struct inode *inode)
{
	wait_queue_head_t *wq = bit_waitqueue(&inode->i_state, __I_DIO_WAKEUP);
	DEFINE_WAIT_BIT(q, &inode->i_state, __I_DIO_WAKEUP);

	do {
		prepare_to_wait(wq, &q.wait, TASK_UNINTERRUPTIBLE);
		if (atomic_read(&inode->i_dio_count))
			schedule();
	} while (atomic_read(&inode->i_dio_count));
	finish_wait(wq, &q.wait);
}

/**
 * inode_dio_wait - wait for outstanding DIO requests to finish
 * @inode: inode to wait for
 *
 * Waits for all pending direct I/O requests to finish so that we can
 * proceed with a truncate or equivalent operation.
 *
 * Must be called under a lock that serializes taking new references
 * to i_dio_count, usually by inode->i_mutex.
 */
void inode_dio_wait(struct inode *inode)
{
	if (atomic_read(&inode->i_dio_count))
		__inode_dio_wait(inode);
}
EXPORT_SYMBOL(inode_dio_wait);

/*
 * inode_dio_done - signal finish of a direct I/O requests
 * @inode: inode the direct I/O happens on
 *
 * This is called once we've finished processing a direct I/O request,
 * and is used to wake up callers waiting for direct I/O to be quiesced.
 */
void inode_dio_done(struct inode *inode)
{
	if (atomic_dec_and_test(&inode->i_dio_count))
		wake_up_bit(&inode->i_state, __I_DIO_WAKEUP);
}
EXPORT_SYMBOL(inode_dio_done);