inode.c 56.6 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
1 2
/*
 * (C) 1997 Linus Torvalds
3
 * (C) 1999 Andrea Arcangeli <andrea@suse.de> (dynamic inode allocation)
Linus Torvalds's avatar
Linus Torvalds committed
4
 */
Al Viro's avatar
Al Viro committed
5
#include <linux/export.h>
Linus Torvalds's avatar
Linus Torvalds committed
6 7 8 9 10 11 12
#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>
13
#include <linux/memblock.h>
14
#include <linux/fsnotify.h>
15
#include <linux/mount.h>
16
#include <linux/posix_acl.h>
17
#include <linux/prefetch.h>
18
#include <linux/buffer_head.h> /* for inode_has_buffers */
19
#include <linux/ratelimit.h>
20
#include <linux/list_lru.h>
21
#include <linux/iversion.h>
22
#include <trace/events/writeback.h>
23
#include "internal.h"
Linus Torvalds's avatar
Linus Torvalds committed
24

25
/*
26
 * Inode locking rules:
27 28 29
 *
 * inode->i_lock protects:
 *   inode->i_state, inode->i_hash, __iget()
30
 * Inode LRU list locks protect:
31
 *   inode->i_sb->s_inode_lru, inode->i_lru
32 33
 * inode->i_sb->s_inode_list_lock protects:
 *   inode->i_sb->s_inodes, inode->i_sb_list
34
 * bdi->wb.list_lock protects:
35
 *   bdi->wb.b_{dirty,io,more_io,dirty_time}, inode->i_io_list
36 37
 * inode_hash_lock protects:
 *   inode_hashtable, inode->i_hash
38 39
 *
 * Lock ordering:
40
 *
41
 * inode->i_sb->s_inode_list_lock
42
 *   inode->i_lock
43
 *     Inode LRU list locks
44
 *
45
 * bdi->wb.list_lock
46
 *   inode->i_lock
47 48
 *
 * inode_hash_lock
49
 *   inode->i_sb->s_inode_list_lock
50 51 52 53
 *   inode->i_lock
 *
 * iunique_lock
 *   inode_hash_lock
54 55
 */

56 57
static unsigned int i_hash_mask __read_mostly;
static unsigned int i_hash_shift __read_mostly;
58 59
static struct hlist_head *inode_hashtable __read_mostly;
static __cacheline_aligned_in_smp DEFINE_SPINLOCK(inode_hash_lock);
Linus Torvalds's avatar
Linus Torvalds committed
60

Jens Axboe's avatar
Jens Axboe committed
61 62 63 64 65 66 67 68
/*
 * 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);

Linus Torvalds's avatar
Linus Torvalds committed
69 70 71 72 73
/*
 * Statistics gathering..
 */
struct inodes_stat_t inodes_stat;

74 75
static DEFINE_PER_CPU(unsigned long, nr_inodes);
static DEFINE_PER_CPU(unsigned long, nr_unused);
76

77
static struct kmem_cache *inode_cachep __read_mostly;
Linus Torvalds's avatar
Linus Torvalds committed
78

79
static long get_nr_inodes(void)
80
{
81
	int i;
82
	long sum = 0;
83 84 85
	for_each_possible_cpu(i)
		sum += per_cpu(nr_inodes, i);
	return sum < 0 ? 0 : sum;
86 87
}

88
static inline long get_nr_inodes_unused(void)
89
{
90
	int i;
91
	long sum = 0;
92 93 94
	for_each_possible_cpu(i)
		sum += per_cpu(nr_unused, i);
	return sum < 0 ? 0 : sum;
95 96
}

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

/*
 * Handle nr_inode sysctl
 */
#ifdef CONFIG_SYSCTL
108
int proc_nr_inodes(struct ctl_table *table, int write,
109 110 111
		   void __user *buffer, size_t *lenp, loff_t *ppos)
{
	inodes_stat.nr_inodes = get_nr_inodes();
112
	inodes_stat.nr_unused = get_nr_inodes_unused();
113
	return proc_doulongvec_minmax(table, write, buffer, lenp, ppos);
114 115 116
}
#endif

117 118 119 120 121
static int no_open(struct inode *inode, struct file *file)
{
	return -ENXIO;
}

122
/**
123
 * inode_init_always - perform inode structure initialisation
124 125
 * @sb: superblock inode belongs to
 * @inode: inode to initialise
126 127 128 129
 *
 * These are initializations that need to be done on every inode
 * allocation as the fields are not initialised by slab allocation.
 */
130
int inode_init_always(struct super_block *sb, struct inode *inode)
Linus Torvalds's avatar
Linus Torvalds committed
131
{
132
	static const struct inode_operations empty_iops;
133
	static const struct file_operations no_open_fops = {.open = no_open};
134
	struct address_space *const mapping = &inode->i_data;
135 136 137 138 139 140

	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;
141
	inode->i_fop = &no_open_fops;
Miklos Szeredi's avatar
Miklos Szeredi committed
142
	inode->__i_nlink = 1;
143
	inode->i_opflags = 0;
144 145
	if (sb->s_xattr)
		inode->i_opflags |= IOP_XATTR;
146 147
	i_uid_write(inode, 0);
	i_gid_write(inode, 0);
148 149
	atomic_set(&inode->i_writecount, 0);
	inode->i_size = 0;
150
	inode->i_write_hint = WRITE_LIFE_NOT_SET;
151 152 153 154 155 156
	inode->i_blocks = 0;
	inode->i_bytes = 0;
	inode->i_generation = 0;
	inode->i_pipe = NULL;
	inode->i_bdev = NULL;
	inode->i_cdev = NULL;
Al Viro's avatar
Al Viro committed
157
	inode->i_link = NULL;
158
	inode->i_dir_seq = 0;
159 160
	inode->i_rdev = 0;
	inode->dirtied_when = 0;
Mimi Zohar's avatar
Mimi Zohar committed
161

162 163 164 165 166 167
#ifdef CONFIG_CGROUP_WRITEBACK
	inode->i_wb_frn_winner = 0;
	inode->i_wb_frn_avg_time = 0;
	inode->i_wb_frn_history = 0;
#endif

Mimi Zohar's avatar
Mimi Zohar committed
168
	if (security_inode_alloc(inode))
169
		goto out;
170 171 172
	spin_lock_init(&inode->i_lock);
	lockdep_set_class(&inode->i_lock, &sb->s_type->i_lock_key);

173 174
	init_rwsem(&inode->i_rwsem);
	lockdep_set_class(&inode->i_rwsem, &sb->s_type->i_mutex_key);
175

Christoph Hellwig's avatar
Christoph Hellwig committed
176
	atomic_set(&inode->i_dio_count, 0);
177 178 179 180

	mapping->a_ops = &empty_aops;
	mapping->host = inode;
	mapping->flags = 0;
181
	mapping->wb_err = 0;
182
	atomic_set(&mapping->i_mmap_writable, 0);
183
	mapping_set_gfp_mask(mapping, GFP_HIGHUSER_MOVABLE);
184
	mapping->private_data = NULL;
185 186 187
	mapping->writeback_index = 0;
	inode->i_private = NULL;
	inode->i_mapping = mapping;
188
	INIT_HLIST_HEAD(&inode->i_dentry);	/* buggered by rcu freeing */
189 190 191
#ifdef CONFIG_FS_POSIX_ACL
	inode->i_acl = inode->i_default_acl = ACL_NOT_CACHED;
#endif
192

193 194 195
#ifdef CONFIG_FSNOTIFY
	inode->i_fsnotify_mask = 0;
#endif
196
	inode->i_flctx = NULL;
197
	this_cpu_inc(nr_inodes);
198

199 200 201
	return 0;
out:
	return -ENOMEM;
Linus Torvalds's avatar
Linus Torvalds committed
202
}
203 204 205 206 207 208 209 210 211 212 213
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);

214 215 216 217 218 219 220 221 222 223 224 225
	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;
226
}
Linus Torvalds's avatar
Linus Torvalds committed
227

228 229 230 231 232 233
void free_inode_nonrcu(struct inode *inode)
{
	kmem_cache_free(inode_cachep, inode);
}
EXPORT_SYMBOL(free_inode_nonrcu);

234
void __destroy_inode(struct inode *inode)
Linus Torvalds's avatar
Linus Torvalds committed
235
{
236
	BUG_ON(inode_has_buffers(inode));
237
	inode_detach_wb(inode);
Linus Torvalds's avatar
Linus Torvalds committed
238
	security_inode_free(inode);
239
	fsnotify_inode_delete(inode);
240
	locks_free_lock_context(inode);
241 242 243 244 245
	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);
	}

246
#ifdef CONFIG_FS_POSIX_ACL
247
	if (inode->i_acl && !is_uncached_acl(inode->i_acl))
248
		posix_acl_release(inode->i_acl);
249
	if (inode->i_default_acl && !is_uncached_acl(inode->i_default_acl))
250 251
		posix_acl_release(inode->i_default_acl);
#endif
252
	this_cpu_dec(nr_inodes);
253 254 255
}
EXPORT_SYMBOL(__destroy_inode);

Nick Piggin's avatar
Nick Piggin committed
256 257 258 259 260 261
static void i_callback(struct rcu_head *head)
{
	struct inode *inode = container_of(head, struct inode, i_rcu);
	kmem_cache_free(inode_cachep, inode);
}

262
static void destroy_inode(struct inode *inode)
263
{
264
	BUG_ON(!list_empty(&inode->i_lru));
265
	__destroy_inode(inode);
Linus Torvalds's avatar
Linus Torvalds committed
266 267 268
	if (inode->i_sb->s_op->destroy_inode)
		inode->i_sb->s_op->destroy_inode(inode);
	else
Nick Piggin's avatar
Nick Piggin committed
269
		call_rcu(&inode->i_rcu, i_callback);
Linus Torvalds's avatar
Linus Torvalds committed
270 271
}

272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340
/**
 * 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)
{
341 342
	if (unlikely(inode->i_nlink == 0)) {
		WARN_ON(!(inode->i_state & I_LINKABLE));
343
		atomic_long_dec(&inode->i_sb->s_remove_count);
344
	}
345 346 347 348 349

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

350
static void __address_space_init_once(struct address_space *mapping)
351
{
352
	xa_init_flags(&mapping->i_pages, XA_FLAGS_LOCK_IRQ);
353
	init_rwsem(&mapping->i_mmap_rwsem);
354 355
	INIT_LIST_HEAD(&mapping->private_list);
	spin_lock_init(&mapping->private_lock);
356
	mapping->i_mmap = RB_ROOT_CACHED;
357
}
358 359 360 361 362 363

void address_space_init_once(struct address_space *mapping)
{
	memset(mapping, 0, sizeof(*mapping));
	__address_space_init_once(mapping);
}
364 365
EXPORT_SYMBOL(address_space_init_once);

Linus Torvalds's avatar
Linus Torvalds committed
366 367 368 369 370 371 372 373 374 375
/*
 * 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);
376
	INIT_LIST_HEAD(&inode->i_io_list);
377
	INIT_LIST_HEAD(&inode->i_wb_list);
378
	INIT_LIST_HEAD(&inode->i_lru);
379
	__address_space_init_once(&inode->i_data);
Linus Torvalds's avatar
Linus Torvalds committed
380 381 382 383
	i_size_ordered_init(inode);
}
EXPORT_SYMBOL(inode_init_once);

384
static void init_once(void *foo)
Linus Torvalds's avatar
Linus Torvalds committed
385
{
386
	struct inode *inode = (struct inode *) foo;
Linus Torvalds's avatar
Linus Torvalds committed
387

388
	inode_init_once(inode);
Linus Torvalds's avatar
Linus Torvalds committed
389 390 391
}

/*
392
 * inode->i_lock must be held
Linus Torvalds's avatar
Linus Torvalds committed
393
 */
394
void __iget(struct inode *inode)
Linus Torvalds's avatar
Linus Torvalds committed
395
{
396 397
	atomic_inc(&inode->i_count);
}
398

Al Viro's avatar
Al Viro committed
399 400 401 402 403 404 405 406 407
/*
 * 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);

408 409
static void inode_lru_list_add(struct inode *inode)
{
410
	if (list_lru_add(&inode->i_sb->s_inode_lru, &inode->i_lru))
411
		this_cpu_inc(nr_unused);
412 413
	else
		inode->i_state |= I_REFERENCED;
414
}
415

416 417 418 419 420 421 422
/*
 * Add inode to LRU if needed (inode is unused and clean).
 *
 * Needs inode->i_lock held.
 */
void inode_add_lru(struct inode *inode)
{
423 424
	if (!(inode->i_state & (I_DIRTY_ALL | I_SYNC |
				I_FREEING | I_WILL_FREE)) &&
425
	    !atomic_read(&inode->i_count) && inode->i_sb->s_flags & SB_ACTIVE)
426 427 428 429
		inode_lru_list_add(inode);
}


430 431
static void inode_lru_list_del(struct inode *inode)
{
432 433

	if (list_lru_del(&inode->i_sb->s_inode_lru, &inode->i_lru))
434
		this_cpu_dec(nr_unused);
Linus Torvalds's avatar
Linus Torvalds committed
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
	spin_lock(&inode->i_sb->s_inode_list_lock);
444
	list_add(&inode->i_sb_list, &inode->i_sb->s_inodes);
445
	spin_unlock(&inode->i_sb->s_inode_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
	if (!list_empty(&inode->i_sb_list)) {
452
		spin_lock(&inode->i_sb->s_inode_list_lock);
453
		list_del_init(&inode->i_sb_list);
454
		spin_unlock(&inode->i_sb->s_inode_list_lock);
455
	}
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)
Al Viro's avatar
Al Viro committed
505
{
506
	/*
Matthew Wilcox's avatar
Matthew Wilcox committed
507
	 * We have to cycle the i_pages lock here because reclaim can be in the
508
	 * process of removing the last page (in __delete_from_page_cache())
Matthew Wilcox's avatar
Matthew Wilcox committed
509
	 * and we must not free the mapping under it.
510
	 */
Matthew Wilcox's avatar
Matthew Wilcox committed
511
	xa_lock_irq(&inode->i_data.i_pages);
Al Viro's avatar
Al Viro committed
512
	BUG_ON(inode->i_data.nrpages);
513
	BUG_ON(inode->i_data.nrexceptional);
Matthew Wilcox's avatar
Matthew Wilcox committed
514
	xa_unlock_irq(&inode->i_data.i_pages);
Al Viro's avatar
Al Viro committed
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);
518
	BUG_ON(!list_empty(&inode->i_wb_list));
Nick Piggin's avatar
Nick Piggin committed
519
	/* don't need i_lock here, no concurrent mods to i_state */
Al Viro's avatar
Al Viro committed
520 521
	inode->i_state = I_FREEING | I_CLEAR;
}
522
EXPORT_SYMBOL(clear_inode);
Al Viro's avatar
Al Viro committed
523

Dave Chinner's avatar
Dave Chinner committed
524 525 526 527 528 529 530 531 532 533 534 535 536
/*
 * 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.
 */
537
static void evict(struct inode *inode)
538 539 540
{
	const struct super_operations *op = inode->i_sb->s_op;

Dave Chinner's avatar
Dave Chinner committed
541 542 543
	BUG_ON(!(inode->i_state & I_FREEING));
	BUG_ON(!list_empty(&inode->i_lru));

544 545
	if (!list_empty(&inode->i_io_list))
		inode_io_list_del(inode);
546

547 548
	inode_sb_list_del(inode);

549 550 551 552 553 554 555
	/*
	 * 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);
556

Al Viro's avatar
Al Viro committed
557 558
	if (op->evict_inode) {
		op->evict_inode(inode);
559
	} else {
560
		truncate_inode_pages_final(&inode->i_data);
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);
Dave Chinner's avatar
Dave Chinner committed
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
}

Linus Torvalds's avatar
Linus Torvalds committed
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;

590 591
		inode = list_first_entry(head, struct inode, i_lru);
		list_del_init(&inode->i_lru);
Linus Torvalds's avatar
Linus Torvalds committed
592

593
		evict(inode);
594
		cond_resched();
Linus Torvalds's avatar
Linus Torvalds committed
595 596 597
	}
}

Al Viro's avatar
Al Viro committed
598 599 600 601 602
/**
 * 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
603
 * called by superblock shutdown after having SB_ACTIVE flag removed,
Al Viro's avatar
Al Viro committed
604 605
 * so any inode reaching zero refcount during or after that call will
 * be immediately evicted.
Linus Torvalds's avatar
Linus Torvalds committed
606
 */
Al Viro's avatar
Al Viro committed
607
void evict_inodes(struct super_block *sb)
Linus Torvalds's avatar
Linus Torvalds committed
608
{
Al Viro's avatar
Al Viro committed
609 610
	struct inode *inode, *next;
	LIST_HEAD(dispose);
Linus Torvalds's avatar
Linus Torvalds committed
611

612
again:
613
	spin_lock(&sb->s_inode_list_lock);
Al Viro's avatar
Al Viro committed
614 615
	list_for_each_entry_safe(inode, next, &sb->s_inodes, i_sb_list) {
		if (atomic_read(&inode->i_count))
Nick Piggin's avatar
Nick Piggin committed
616
			continue;
617 618 619 620

		spin_lock(&inode->i_lock);
		if (inode->i_state & (I_NEW | I_FREEING | I_WILL_FREE)) {
			spin_unlock(&inode->i_lock);
Linus Torvalds's avatar
Linus Torvalds committed
621
			continue;
622
		}
Al Viro's avatar
Al Viro committed
623 624

		inode->i_state |= I_FREEING;
625
		inode_lru_list_del(inode);
626
		spin_unlock(&inode->i_lock);
627
		list_add(&inode->i_lru, &dispose);
628 629 630 631 632 633 634 635 636 637 638 639

		/*
		 * We can have a ton of inodes to evict at unmount time given
		 * enough memory, check to see if we need to go to sleep for a
		 * bit so we don't livelock.
		 */
		if (need_resched()) {
			spin_unlock(&sb->s_inode_list_lock);
			cond_resched();
			dispose_list(&dispose);
			goto again;
		}
Linus Torvalds's avatar
Linus Torvalds committed
640
	}
641
	spin_unlock(&sb->s_inode_list_lock);
Al Viro's avatar
Al Viro committed
642 643

	dispose_list(&dispose);
Linus Torvalds's avatar
Linus Torvalds committed
644
}
645
EXPORT_SYMBOL_GPL(evict_inodes);
Linus Torvalds's avatar
Linus Torvalds committed
646 647

/**
648 649
 * invalidate_inodes	- attempt to free all inodes on a superblock
 * @sb:		superblock to operate on
650
 * @kill_dirty: flag to guide handling of dirty inodes
Linus Torvalds's avatar
Linus Torvalds committed
651
 *
652 653
 * Attempts to free all inodes for a given superblock.  If there were any
 * busy inodes return a non-zero value, else zero.
654 655
 * If @kill_dirty is set, discard dirty inodes too, otherwise treat
 * them as busy.
Linus Torvalds's avatar
Linus Torvalds committed
656
 */
657
int invalidate_inodes(struct super_block *sb, bool kill_dirty)
Linus Torvalds's avatar
Linus Torvalds committed
658
{
659
	int busy = 0;
660 661
	struct inode *inode, *next;
	LIST_HEAD(dispose);
Linus Torvalds's avatar
Linus Torvalds committed
662

663
	spin_lock(&sb->s_inode_list_lock);
664
	list_for_each_entry_safe(inode, next, &sb->s_inodes, i_sb_list) {
665 666 667
		spin_lock(&inode->i_lock);
		if (inode->i_state & (I_NEW | I_FREEING | I_WILL_FREE)) {
			spin_unlock(&inode->i_lock);
Nick Piggin's avatar
Nick Piggin committed
668
			continue;
669
		}
670
		if (inode->i_state & I_DIRTY_ALL && !kill_dirty) {
671
			spin_unlock(&inode->i_lock);
672 673 674
			busy = 1;
			continue;
		}
675
		if (atomic_read(&inode->i_count)) {
676
			spin_unlock(&inode->i_lock);
677
			busy = 1;
Linus Torvalds's avatar
Linus Torvalds committed
678 679
			continue;
		}
680 681

		inode->i_state |= I_FREEING;
682
		inode_lru_list_del(inode);
683
		spin_unlock(&inode->i_lock);
684
		list_add(&inode->i_lru, &dispose);
Linus Torvalds's avatar
Linus Torvalds committed
685
	}
686
	spin_unlock(&sb->s_inode_list_lock);
Linus Torvalds's avatar
Linus Torvalds committed
687

688
	dispose_list(&dispose);
Linus Torvalds's avatar
Linus Torvalds committed
689 690 691 692 693

	return busy;
}

/*
694
 * Isolate the inode from the LRU in preparation for freeing it.
Linus Torvalds's avatar
Linus Torvalds committed
695 696
 *
 * Any inodes which are pinned purely because of attached pagecache have their
697 698
 * pagecache removed.  If the inode has metadata buffers attached to
 * mapping->private_list then try to remove them.
Linus Torvalds's avatar
Linus Torvalds committed
699
 *
700 701 702 703 704 705 706
 * 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.
Linus Torvalds's avatar
Linus Torvalds committed
707
 */
708 709
static enum lru_status inode_lru_isolate(struct list_head *item,
		struct list_lru_one *lru, spinlock_t *lru_lock, void *arg)
Linus Torvalds's avatar
Linus Torvalds committed
710
{
711 712
	struct list_head *freeable = arg;
	struct inode	*inode = container_of(item, struct inode, i_lru);
Linus Torvalds's avatar
Linus Torvalds committed
713

714 715 716 717 718 719
	/*
	 * we are inverting the lru lock/inode->i_lock here, so use a trylock.
	 * If we fail to get the lock, just skip it.
	 */
	if (!spin_trylock(&inode->i_lock))
		return LRU_SKIP;
Linus Torvalds's avatar
Linus Torvalds committed
720

721 722 723 724 725 726
	/*
	 * 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)) {
727
		list_lru_isolate(lru, &inode->i_lru);
728 729 730 731
		spin_unlock(&inode->i_lock);
		this_cpu_dec(nr_unused);
		return LRU_REMOVED;
	}
Linus Torvalds's avatar
Linus Torvalds committed
732

733 734 735 736 737
	/*
	 * Recently referenced inodes and inodes with many attached pages
	 * get one more pass.
	 */
	if (inode->i_state & I_REFERENCED || inode->i_data.nrpages > 1) {
738 739 740 741
		inode->i_state &= ~I_REFERENCED;
		spin_unlock(&inode->i_lock);
		return LRU_ROTATE;
	}
Linus Torvalds's avatar
Linus Torvalds committed
742

743 744 745 746 747 748 749 750 751 752 753 754 755
	if (inode_has_buffers(inode) || inode->i_data.nrpages) {
		__iget(inode);
		spin_unlock(&inode->i_lock);
		spin_unlock(lru_lock);
		if (remove_inode_buffers(inode)) {
			unsigned long reap;
			reap = invalidate_mapping_pages(&inode->i_data, 0, -1);
			if (current_is_kswapd())
				__count_vm_events(KSWAPD_INODESTEAL, reap);
			else
				__count_vm_events(PGINODESTEAL, reap);
			if (current->reclaim_state)
				current->reclaim_state->reclaimed_slab += reap;
756
		}
757 758 759 760
		iput(inode);
		spin_lock(lru_lock);
		return LRU_RETRY;
	}
761

762 763
	WARN_ON(inode->i_state & I_NEW);
	inode->i_state |= I_FREEING;
764
	list_lru_isolate_move(lru, &inode->i_lru, freeable);
765
	spin_unlock(&inode->i_lock);
766

767 768 769
	this_cpu_dec(nr_unused);
	return LRU_REMOVED;
}
770

771 772 773 774 775 776
/*
 * 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().
 */
777
long prune_icache_sb(struct super_block *sb, struct shrink_control *sc)
778 779 780
{
	LIST_HEAD(freeable);
	long freed;
Linus Torvalds's avatar
Linus Torvalds committed
781

782 783
	freed = list_lru_shrink_walk(&sb->s_inode_lru, sc,
				     inode_lru_isolate, &freeable);
Linus Torvalds's avatar
Linus Torvalds committed
784
	dispose_list(&freeable);
785
	return freed;
Linus Torvalds's avatar
Linus Torvalds committed
786 787 788 789 790 791
}

static void __wait_on_freeing_inode(struct inode *inode);
/*
 * Called with the inode lock held.
 */
792 793 794 795
static struct inode *find_inode(struct super_block *sb,
				struct hlist_head *head,
				int (*test)(struct inode *, void *),
				void *data)
Linus Torvalds's avatar
Linus Torvalds committed
796
{
797
	struct inode *inode = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
798 799

repeat:
800
	hlist_for_each_entry(inode, head, i_hash) {
801
		if (inode->i_sb != sb)
Linus Torvalds's avatar
Linus Torvalds committed
802
			continue;
803
		if (!test(inode, data))
Linus Torvalds's avatar
Linus Torvalds committed
804
			continue;
805
		spin_lock(&inode->i_lock);
806
		if (inode->i_state & (I_FREEING|I_WILL_FREE)) {
Linus Torvalds's avatar
Linus Torvalds committed
807 808 809
			__wait_on_freeing_inode(inode);
			goto repeat;
		}
810 811 812 813
		if (unlikely(inode->i_state & I_CREATING)) {
			spin_unlock(&inode->i_lock);
			return ERR_PTR(-ESTALE);
		}
814
		__iget(inode);
815
		spin_unlock(&inode->i_lock);
816
		return inode;
Linus Torvalds's avatar
Linus Torvalds committed
817
	}
818
	return NULL;
Linus Torvalds's avatar
Linus Torvalds committed
819 820 821 822 823 824
}

/*
 * find_inode_fast is the fast path version of find_inode, see the comment at
 * iget_locked for details.
 */
825 826
static struct inode *find_inode_fast(struct super_block *sb,
				struct hlist_head *head, unsigned long ino)
Linus Torvalds's avatar
Linus Torvalds committed
827
{
828
	struct inode *inode = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
829 830

repeat:
831
	hlist_for_each_entry(inode, head, i_hash) {
832
		if (inode->i_ino != ino)
Linus Torvalds's avatar
Linus Torvalds committed
833
			continue;
834
		if (inode->i_sb != sb)
Linus Torvalds's avatar
Linus Torvalds committed
835
			continue;
836
		spin_lock(&inode->i_lock);
837
		if (inode->i_state & (I_FREEING|I_WILL_FREE)) {
Linus Torvalds's avatar
Linus Torvalds committed
838 839 840
			__wait_on_freeing_inode(inode);
			goto repeat;
		}
841 842 843 844
		if (unlikely(inode->i_state & I_CREATING)) {
			spin_unlock(&inode->i_lock);
			return ERR_PTR(-ESTALE);
		}
845
		__iget(inode);
846
		spin_unlock(&inode->i_lock);
847
		return inode;
Linus Torvalds's avatar
Linus Torvalds committed
848
	}
849
	return NULL;
850 851
}

852 853 854 855
/*
 * 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.
856
 *
857 858 859 860 861 862 863 864 865
 * 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.
866
 */
867 868 869
#define LAST_INO_BATCH 1024
static DEFINE_PER_CPU(unsigned int, last_ino);

870
unsigned int get_next_ino(void)
871
{
872 873
	unsigned int *p = &get_cpu_var(last_ino);
	unsigned int res = *p;
874

875 876 877 878 879 880 881 882 883
#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

884 885 886 887 888
	res++;
	/* get_next_ino should not provide a 0 inode number */
	if (unlikely(!res))
		res++;
	*p = res;
889 890
	put_cpu_var(last_ino);
	return res;
891
}
892
EXPORT_SYMBOL(get_next_ino);
893

894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916
/**
 *	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;
}

Linus Torvalds's avatar
Linus Torvalds committed
917 918 919 920
/**
 *	new_inode 	- obtain an inode
 *	@sb: superblock
 *
921
 *	Allocates a new inode for given superblock. The default gfp_mask
922
 *	for allocations related to inode->i_mapping is GFP_HIGHUSER_MOVABLE.
923 924 925 926 927
 *	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
 *
Linus Torvalds's avatar
Linus Torvalds committed
928 929 930
 */
struct inode *new_inode(struct super_block *sb)
{
931
	struct inode *inode;
Linus Torvalds's avatar
Linus Torvalds committed
932

933
	spin_lock_prefetch(&sb->s_inode_list_lock);
934

935 936
	inode = new_inode_pseudo(sb);
	if (inode)
937
		inode_sb_list_add(inode);
Linus Torvalds's avatar
Linus Torvalds committed
938 939 940 941
	return inode;
}
EXPORT_SYMBOL(new_inode);

942
#ifdef CONFIG_DEBUG_LOCK_ALLOC
943 944
void lockdep_annotate_inode_mutex_key(struct inode *inode)
{
945
	if (S_ISDIR(inode->i_mode)) {
946 947
		struct file_system_type *type = inode->i_sb->s_type;

948
		/* Set new key only if filesystem hasn't already changed it */
949
		if (lockdep_match_class(&inode->i_rwsem, &type->i_mutex_key)) {
950 951 952
			/*
			 * ensure nobody is actually holding i_mutex
			 */
953 954 955
			// mutex_destroy(&inode->i_mutex);
			init_rwsem(&inode->i_rwsem);
			lockdep_set_class(&inode->i_rwsem,
956 957
					  &type->i_mutex_dir_key);
		}
958
	}
959 960
}
EXPORT_SYMBOL(lockdep_annotate_inode_mutex_key);
961
#endif
962 963 964 965 966 967 968 969 970 971 972

/**
 * 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);
973
	spin_lock(&inode->i_lock);
Christoph Hellwig's avatar
Christoph Hellwig committed
974
	WARN_ON(!(inode->i_state & I_NEW));
975
	inode->i_state &= ~I_NEW & ~I_CREATING;
976
	smp_mb();
977 978
	wake_up_bit(&inode->i_state, __I_NEW);
	spin_unlock(&inode->i_lock);
Linus Torvalds's avatar
Linus Torvalds committed
979 980 981
}
EXPORT_SYMBOL(unlock_new_inode);

982 983 984 985 986 987 988 989 990 991 992 993 994
void discard_new_inode(struct inode *inode)
{
	lockdep_annotate_inode_mutex_key(inode);
	spin_lock(&inode->i_lock);
	WARN_ON(!(inode->i_state & I_NEW));
	inode->i_state &= ~I_NEW;
	smp_mb();
	wake_up_bit(&inode->i_state, __I_NEW);
	spin_unlock(&inode->i_lock);
	iput(inode);
}
EXPORT_SYMBOL(discard_new_inode);

995 996
/**
 * lock_two_nondirectories - take two i_mutexes on non-directory objects
997 998 999 1000
 *
 * Lock any non-NULL argument that is not a directory.
 * Zero, one or two objects may be locked by this function.
 *
1001 1002 1003 1004 1005
 * @inode1: first inode to lock
 * @inode2: second inode to lock
 */
void lock_two_nondirectories(struct inode *inode1, struct inode *inode2)
{
1006 1007 1008 1009
	if (inode1 > inode2)
		swap(inode1, inode2);

	if (inode1 && !S_ISDIR(inode1->i_mode))
Al Viro's avatar
Al Viro committed
1010
		inode_lock(inode1);
1011
	if (inode2 && !S_ISDIR(inode2->i_mode) && inode2 != inode1)
Al Viro's avatar
Al Viro committed
1012
		inode_lock_nested(inode2, I_MUTEX_NONDIR2);
1013 1014 1015 1016 1017 1018 1019 1020 1021 1022
}
EXPORT_SYMBOL(lock_two_nondirectories);

/**
 * unlock_two_nondirectories - release locks from lock_two_nondirectories()
 * @inode1: first inode to unlock
 * @inode2: second inode to unlock
 */
void unlock_two_nondirectories(struct inode *inode1, struct inode *inode2)
{
1023
	if (inode1 && !S_ISDIR(inode1->i_mode))
Al Viro's avatar
Al Viro committed
1024
		inode_unlock(inode1);
1025
	if (inode2 && !S_ISDIR(inode2->i_mode) && inode2 != inode1)
Al Viro's avatar
Al Viro committed
1026
		inode_unlock(inode2);
1027 1028 1029
}
EXPORT_SYMBOL(unlock_two_nondirectories);

1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055
/**
 * inode_insert5 - obtain an inode from a mounted file system
 * @inode:	pre-allocated inode to use for insert to cache
 * @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 variant of iget5_locked() for callers that don't want to fail on memory
 * allocation of inode.
 *
 * If the inode is not in cache, insert the pre-allocated inode to cache 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().
 *
 * Note both @test and @set are called with the inode_hash_lock held, so can't
 * sleep.
 */
struct inode *inode_insert5(struct inode *inode, unsigned long hashval,
			    int (*test)(struct inode *, void *),
			    int (*set)(struct inode *, void *), void *data)
{
	struct hlist_head *head = inode_hashtable + hash(inode->i_sb, hashval);
	struct inode *old;
1056
	bool creating = inode->i_state & I_CREATING;
1057 1058 1059 1060 1061 1062 1063 1064 1065 1066

again:
	spin_lock(&inode_hash_lock);
	old = find_inode(inode->i_sb, head, test, data);
	if (unlikely(old)) {
		/*
		 * Uhhuh, somebody else created the same inode under us.
		 * Use the old inode instead of the preallocated one.
		 */
		spin_unlock(&inode_hash_lock);
1067 1068
		if (IS_ERR(old))
			return NULL;
1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089
		wait_on_inode(old);
		if (unlikely(inode_unhashed(old))) {
			iput(old);
			goto again;
		}
		return old;
	}

	if (set && unlikely(set(inode, data))) {
		inode = NULL;
		goto unlock;
	}

	/*
	 * Return the locked inode with I_NEW set, the
	 * caller is responsible for filling in the contents
	 */
	spin_lock(&inode->i_lock);
	inode->i_state |= I_NEW;
	hlist_add_head(&inode->i_hash, head);
	spin_unlock(&inode->i_lock);
1090 1091
	if (!creating)
		inode_sb_list_add(inode);
1092 1093 1094 1095 1096 1097 1098
unlock:
	spin_unlock(&inode_hash_lock);

	return inode;
}
EXPORT_SYMBOL(inode_insert5);

1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114
/**
 * 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().
Linus Torvalds's avatar
Linus Torvalds committed
1115
 *
1116 1117
 * Note both @test and @set are called with the inode_hash_lock held, so can't
 * sleep.
Linus Torvalds's avatar
Linus Torvalds committed
1118
 */
1119 1120 1121
struct inode *iget5_locked(struct super_block *sb, unsigned long hashval,
		int (*test)(struct inode *, void *),
		int (*set)(struct inode *, void *), void *data)
Linus Torvalds's avatar
Linus Torvalds committed
1122
{
1123
	struct inode *inode = ilookup5(sb, hashval, test, data);
1124

1125
	if (!inode) {
1126
		struct inode *new = alloc_inode(sb);
1127

1128
		if (new) {
1129
			new->i_state = 0;
1130 1131
			inode = inode_insert5(new, hashval, test, set, data);
			if (unlikely(inode != new))
1132
				destroy_inode(new);
1133
		}
Linus Torvalds's avatar
Linus Torvalds committed
1134 1135 1136
	}
	return inode;
}
1137
EXPORT_SYMBOL(iget5_locked);
Linus Torvalds's avatar
Linus Torvalds committed
1138

1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150
/**
 * 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().
Linus Torvalds's avatar
Linus Torvalds committed
1151
 */
1152
struct inode *iget_locked(struct super_block *sb, unsigned long ino)
Linus Torvalds's avatar
Linus Torvalds committed
1153
{
1154
	struct hlist_head *head = inode_hashtable + hash(sb, ino);
1155
	struct inode *inode;
1156
again:
1157 1158 1159 1160
	spin_lock(&inode_hash_lock);
	inode = find_inode_fast(sb, head, ino);
	spin_unlock(&inode_hash_lock);
	if (inode) {
1161 1162
		if (IS_ERR(inode))
			return NULL;
1163
		wait_on_inode(inode);
1164 1165 1166 1167
		if (unlikely(inode_unhashed(inode))) {
			iput(inode);
			goto again;
		}
1168 1169 1170
		return inode;
	}

Linus Torvalds's avatar
Linus Torvalds committed
1171 1172
	inode = alloc_inode(sb);
	if (inode) {
1173
		struct inode *old;
Linus Torvalds's avatar
Linus Torvalds committed
1174

1175
		spin_lock(&inode_hash_lock);
Linus Torvalds's avatar
Linus Torvalds committed
1176 1177 1178 1179
		/* We released the lock, so.. */
		old = find_inode_fast(sb, head, ino);
		if (!old) {
			inode->i_ino = ino;
1180 1181
			spin_lock(&inode->i_lock);
			inode->i_state = I_NEW;
1182
			hlist_add_head(&inode->i_hash, head);
1183
			spin_unlock(&inode->i_lock);
1184
			inode_sb_list_add(inode);
1185
			spin_unlock(&inode_hash_lock);
Linus Torvalds's avatar
Linus Torvalds committed
1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197

			/* 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.
		 */
1198
		spin_unlock(&inode_hash_lock);
Linus Torvalds's avatar
Linus Torvalds committed
1199
		destroy_inode(inode);
1200 1201
		if (IS_ERR(old))
			return NULL;
Linus Torvalds's avatar
Linus Torvalds committed
1202 1203
		inode = old;
		wait_on_inode(inode);
1204 1205 1206 1207
		if (unlikely(inode_unhashed(inode))) {
			iput(inode);
			goto again;
		}
Linus Torvalds's avatar
Linus Torvalds committed
1208 1209 1210
	}
	return inode;
}
1211
EXPORT_SYMBOL(iget_locked);
Linus Torvalds's avatar
Linus Torvalds committed
1212

1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224
/*
 * 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;

1225
	spin_lock(&inode_hash_lock);
1226
	hlist_for_each_entry(inode, b, i_hash) {
1227 1228
		if (inode->i_ino == ino && inode->i_sb == sb) {
			spin_unlock(&inode_hash_lock);
1229
			return 0;
1230
		}
1231
	}
1232
	spin_unlock(&inode_hash_lock);
1233 1234 1235 1236

	return 1;
}

Linus Torvalds's avatar
Linus Torvalds committed
1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252
/**
 *	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)
{
1253 1254 1255 1256 1257
	/*
	 * 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.
	 */
1258
	static DEFINE_SPINLOCK(iunique_lock);
1259
	static unsigned int counter;
Linus Torvalds's avatar
Linus Torvalds committed
1260
	ino_t res;
1261

1262
	spin_lock(&iunique_lock);
1263 1264 1265
	do {
		if (counter <= max_reserved)
			counter = max_reserved + 1;
Linus Torvalds's avatar
Linus Torvalds committed
1266
		res = counter++;
1267 1268
	} while (!test_inode_iunique(sb, res));
	spin_unlock(&iunique_lock);
Linus Torvalds's avatar
Linus Torvalds committed
1269

1270 1271
	return res;
}
Linus Torvalds's avatar
Linus Torvalds committed
1272 1273 1274 1275
EXPORT_SYMBOL(iunique);

struct inode *igrab(struct inode *inode)
{
1276 1277
	spin_lock(&inode->i_lock);
	if (!(inode->i_state & (I_FREEING|I_WILL_FREE))) {
Linus Torvalds's avatar
Linus Torvalds committed
1278
		__iget(inode);
1279 1280 1281
		spin_unlock(&inode->i_lock);
	} else {
		spin_unlock(&inode->i_lock);
Linus Torvalds's avatar
Linus Torvalds committed
1282 1283 1284 1285 1286 1287
		/*
		 * 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;
1288
	}
Linus Torvalds's avatar
Linus Torvalds committed
1289