inode.c 45.1 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 5 6 7 8 9 10 11 12 13
 */
#include <linux/fs.h>
#include <linux/mm.h>
#include <linux/dcache.h>
#include <linux/init.h>
#include <linux/slab.h>
#include <linux/writeback.h>
#include <linux/module.h>
#include <linux/backing-dev.h>
#include <linux/wait.h>
N
Nick Piggin 已提交
14
#include <linux/rwsem.h>
L
Linus Torvalds 已提交
15 16 17 18 19 20
#include <linux/hash.h>
#include <linux/swap.h>
#include <linux/security.h>
#include <linux/pagemap.h>
#include <linux/cdev.h>
#include <linux/bootmem.h>
21
#include <linux/fsnotify.h>
22
#include <linux/mount.h>
23
#include <linux/async.h>
A
Al Viro 已提交
24
#include <linux/posix_acl.h>
25
#include <linux/prefetch.h>
26
#include <linux/ima.h>
27
#include <linux/cred.h>
28
#include <linux/buffer_head.h> /* for inode_has_buffers */
M
Miklos Szeredi 已提交
29
#include <linux/ratelimit.h>
30
#include "internal.h"
L
Linus Torvalds 已提交
31

32
/*
33
 * Inode locking rules:
34 35 36
 *
 * inode->i_lock protects:
 *   inode->i_state, inode->i_hash, __iget()
D
Dave Chinner 已提交
37
 * inode->i_sb->s_inode_lru_lock protects:
38
 *   inode->i_sb->s_inode_lru, inode->i_lru
39 40
 * inode_sb_list_lock protects:
 *   sb->s_inodes, inode->i_sb_list
41
 * bdi->wb.list_lock protects:
42
 *   bdi->wb.b_{dirty,io,more_io}, inode->i_wb_list
43 44
 * inode_hash_lock protects:
 *   inode_hashtable, inode->i_hash
45 46
 *
 * Lock ordering:
47 48 49
 *
 * inode_sb_list_lock
 *   inode->i_lock
D
Dave Chinner 已提交
50
 *     inode->i_sb->s_inode_lru_lock
51
 *
52
 * bdi->wb.list_lock
53
 *   inode->i_lock
54 55 56 57 58 59 60
 *
 * inode_hash_lock
 *   inode_sb_list_lock
 *   inode->i_lock
 *
 * iunique_lock
 *   inode_hash_lock
61 62
 */

63 64
static unsigned int i_hash_mask __read_mostly;
static unsigned int i_hash_shift __read_mostly;
65 66
static struct hlist_head *inode_hashtable __read_mostly;
static __cacheline_aligned_in_smp DEFINE_SPINLOCK(inode_hash_lock);
L
Linus Torvalds 已提交
67

68 69
__cacheline_aligned_in_smp DEFINE_SPINLOCK(inode_sb_list_lock);

J
Jens Axboe 已提交
70 71 72 73 74 75 76 77
/*
 * 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
78 79 80 81 82
/*
 * Statistics gathering..
 */
struct inodes_stat_t inodes_stat;

83
static DEFINE_PER_CPU(unsigned int, nr_inodes);
84
static DEFINE_PER_CPU(unsigned int, nr_unused);
85

86
static struct kmem_cache *inode_cachep __read_mostly;
L
Linus Torvalds 已提交
87

88
static int get_nr_inodes(void)
89
{
90 91 92 93 94
	int i;
	int sum = 0;
	for_each_possible_cpu(i)
		sum += per_cpu(nr_inodes, i);
	return sum < 0 ? 0 : sum;
95 96 97 98
}

static inline int get_nr_inodes_unused(void)
{
99 100 101 102 103
	int i;
	int sum = 0;
	for_each_possible_cpu(i)
		sum += per_cpu(nr_unused, i);
	return sum < 0 ? 0 : sum;
104 105 106 107
}

int get_nr_dirty_inodes(void)
{
108
	/* not actually dirty inodes, but a wild approximation */
109 110 111 112 113 114 115 116 117 118 119 120
	int nr_dirty = get_nr_inodes() - get_nr_inodes_unused();
	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();
121
	inodes_stat.nr_unused = get_nr_inodes_unused();
122 123 124 125
	return proc_dointvec(table, write, buffer, lenp, ppos);
}
#endif

D
David Chinner 已提交
126 127
/**
 * inode_init_always - perform inode structure intialisation
R
Randy Dunlap 已提交
128 129
 * @sb: superblock inode belongs to
 * @inode: inode to initialise
D
David Chinner 已提交
130 131 132 133
 *
 * These are initializations that need to be done on every inode
 * allocation as the fields are not initialised by slab allocation.
 */
134
int inode_init_always(struct super_block *sb, struct inode *inode)
L
Linus Torvalds 已提交
135
{
136
	static const struct inode_operations empty_iops;
137
	static const struct file_operations empty_fops;
138
	struct address_space *const mapping = &inode->i_data;
D
David Chinner 已提交
139 140 141 142 143 144 145

	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
146
	inode->__i_nlink = 1;
147
	inode->i_opflags = 0;
A
Al Viro 已提交
148 149
	inode->i_uid = 0;
	inode->i_gid = 0;
D
David Chinner 已提交
150 151 152 153 154
	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
155
#ifdef CONFIG_QUOTA
D
David Chinner 已提交
156
	memset(&inode->i_dquot, 0, sizeof(inode->i_dquot));
L
Linus Torvalds 已提交
157
#endif
D
David Chinner 已提交
158 159 160 161 162
	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
163 164

	if (security_inode_alloc(inode))
165
		goto out;
D
David Chinner 已提交
166 167 168 169 170 171
	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
172
	atomic_set(&inode->i_dio_count, 0);
D
David Chinner 已提交
173 174 175 176

	mapping->a_ops = &empty_aops;
	mapping->host = inode;
	mapping->flags = 0;
H
Hugh Dickins 已提交
177
	mapping_set_gfp_mask(mapping, GFP_HIGHUSER_MOVABLE);
D
David Chinner 已提交
178 179 180 181 182 183 184 185 186 187 188 189
	mapping->assoc_mapping = NULL;
	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;

190
		bdi = sb->s_bdev->bd_inode->i_mapping->backing_dev_info;
D
David Chinner 已提交
191 192 193 194
		mapping->backing_dev_info = bdi;
	}
	inode->i_private = NULL;
	inode->i_mapping = mapping;
195
	INIT_LIST_HEAD(&inode->i_dentry);	/* buggered by rcu freeing */
A
Al Viro 已提交
196 197 198
#ifdef CONFIG_FS_POSIX_ACL
	inode->i_acl = inode->i_default_acl = ACL_NOT_CACHED;
#endif
D
David Chinner 已提交
199

200 201 202 203
#ifdef CONFIG_FSNOTIFY
	inode->i_fsnotify_mask = 0;
#endif

204
	this_cpu_inc(nr_inodes);
205

206 207 208
	return 0;
out:
	return -ENOMEM;
L
Linus Torvalds 已提交
209
}
D
David Chinner 已提交
210 211 212 213 214 215 216 217 218 219 220
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);

221 222 223 224 225 226 227 228 229 230 231 232
	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
233
}
L
Linus Torvalds 已提交
234

235 236 237 238 239 240
void free_inode_nonrcu(struct inode *inode)
{
	kmem_cache_free(inode_cachep, inode);
}
EXPORT_SYMBOL(free_inode_nonrcu);

C
Christoph Hellwig 已提交
241
void __destroy_inode(struct inode *inode)
L
Linus Torvalds 已提交
242
{
243
	BUG_ON(inode_has_buffers(inode));
L
Linus Torvalds 已提交
244
	security_inode_free(inode);
245
	fsnotify_inode_delete(inode);
M
Miklos Szeredi 已提交
246 247 248 249 250
	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
251 252 253 254 255 256
#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
257
	this_cpu_dec(nr_inodes);
C
Christoph Hellwig 已提交
258 259 260
}
EXPORT_SYMBOL(__destroy_inode);

N
Nick Piggin 已提交
261 262 263 264 265 266
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
267
static void destroy_inode(struct inode *inode)
C
Christoph Hellwig 已提交
268
{
N
Nick Piggin 已提交
269
	BUG_ON(!list_empty(&inode->i_lru));
C
Christoph Hellwig 已提交
270
	__destroy_inode(inode);
L
Linus Torvalds 已提交
271 272 273
	if (inode->i_sb->s_op->destroy_inode)
		inode->i_sb->s_op->destroy_inode(inode);
	else
N
Nick Piggin 已提交
274
		call_rcu(&inode->i_rcu, i_callback);
L
Linus Torvalds 已提交
275 276
}

M
Miklos Szeredi 已提交
277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352
/**
 * 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)
{
	if (WARN_ON(inode->i_nlink == 0))
		atomic_long_dec(&inode->i_sb->s_remove_count);

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

353 354 355 356 357
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);
358
	mutex_init(&mapping->i_mmap_mutex);
359 360 361 362 363 364 365
	INIT_LIST_HEAD(&mapping->private_list);
	spin_lock_init(&mapping->private_lock);
	INIT_RAW_PRIO_TREE_ROOT(&mapping->i_mmap);
	INIT_LIST_HEAD(&mapping->i_mmap_nonlinear);
}
EXPORT_SYMBOL(address_space_init_once);

L
Linus Torvalds 已提交
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);
N
Nick Piggin 已提交
376 377
	INIT_LIST_HEAD(&inode->i_wb_list);
	INIT_LIST_HEAD(&inode->i_lru);
378
	address_space_init_once(&inode->i_data);
L
Linus Torvalds 已提交
379
	i_size_ordered_init(inode);
380
#ifdef CONFIG_FSNOTIFY
381
	INIT_HLIST_HEAD(&inode->i_fsnotify_marks);
382
#endif
L
Linus Torvalds 已提交
383 384 385
}
EXPORT_SYMBOL(inode_init_once);

386
static void init_once(void *foo)
L
Linus Torvalds 已提交
387
{
388
	struct inode *inode = (struct inode *) foo;
L
Linus Torvalds 已提交
389

C
Christoph Lameter 已提交
390
	inode_init_once(inode);
L
Linus Torvalds 已提交
391 392 393
}

/*
394
 * inode->i_lock must be held
L
Linus Torvalds 已提交
395
 */
396
void __iget(struct inode *inode)
L
Linus Torvalds 已提交
397
{
398 399
	atomic_inc(&inode->i_count);
}
400

A
Al Viro 已提交
401 402 403 404 405 406 407 408 409
/*
 * 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);

410 411
static void inode_lru_list_add(struct inode *inode)
{
D
Dave Chinner 已提交
412
	spin_lock(&inode->i_sb->s_inode_lru_lock);
N
Nick Piggin 已提交
413
	if (list_empty(&inode->i_lru)) {
414 415
		list_add(&inode->i_lru, &inode->i_sb->s_inode_lru);
		inode->i_sb->s_nr_inodes_unused++;
416
		this_cpu_inc(nr_unused);
417
	}
D
Dave Chinner 已提交
418
	spin_unlock(&inode->i_sb->s_inode_lru_lock);
419
}
420

421 422
static void inode_lru_list_del(struct inode *inode)
{
D
Dave Chinner 已提交
423
	spin_lock(&inode->i_sb->s_inode_lru_lock);
N
Nick Piggin 已提交
424 425
	if (!list_empty(&inode->i_lru)) {
		list_del_init(&inode->i_lru);
426
		inode->i_sb->s_nr_inodes_unused--;
427
		this_cpu_dec(nr_unused);
428
	}
D
Dave Chinner 已提交
429
	spin_unlock(&inode->i_sb->s_inode_lru_lock);
L
Linus Torvalds 已提交
430 431
}

432 433 434 435 436 437
/**
 * inode_sb_list_add - add inode to the superblock list of inodes
 * @inode: inode to add
 */
void inode_sb_list_add(struct inode *inode)
{
438 439 440
	spin_lock(&inode_sb_list_lock);
	list_add(&inode->i_sb_list, &inode->i_sb->s_inodes);
	spin_unlock(&inode_sb_list_lock);
441 442 443
}
EXPORT_SYMBOL_GPL(inode_sb_list_add);

444
static inline void inode_sb_list_del(struct inode *inode)
445
{
446 447 448 449 450
	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);
	}
451 452
}

453 454 455 456 457 458
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;
459 460
	tmp = tmp ^ ((tmp ^ GOLDEN_RATIO_PRIME) >> i_hash_shift);
	return tmp & i_hash_mask;
461 462 463 464 465 466 467 468 469 470 471 472
}

/**
 *	__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)
{
473 474
	struct hlist_head *b = inode_hashtable + hash(inode->i_sb, hashval);

475
	spin_lock(&inode_hash_lock);
476
	spin_lock(&inode->i_lock);
477
	hlist_add_head(&inode->i_hash, b);
478
	spin_unlock(&inode->i_lock);
479
	spin_unlock(&inode_hash_lock);
480 481 482 483
}
EXPORT_SYMBOL(__insert_inode_hash);

/**
484
 *	__remove_inode_hash - remove an inode from the hash
485 486 487 488
 *	@inode: inode to unhash
 *
 *	Remove an inode from the superblock.
 */
489
void __remove_inode_hash(struct inode *inode)
490
{
491
	spin_lock(&inode_hash_lock);
492
	spin_lock(&inode->i_lock);
493
	hlist_del_init(&inode->i_hash);
494
	spin_unlock(&inode->i_lock);
495
	spin_unlock(&inode_hash_lock);
496
}
497
EXPORT_SYMBOL(__remove_inode_hash);
498

A
Al Viro 已提交
499 500 501
void end_writeback(struct inode *inode)
{
	might_sleep();
502 503 504 505 506 507
	/*
	 * 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
508
	BUG_ON(inode->i_data.nrpages);
509
	spin_unlock_irq(&inode->i_data.tree_lock);
A
Al Viro 已提交
510 511 512 513
	BUG_ON(!list_empty(&inode->i_data.private_list));
	BUG_ON(!(inode->i_state & I_FREEING));
	BUG_ON(inode->i_state & I_CLEAR);
	inode_sync_wait(inode);
N
Nick Piggin 已提交
514
	/* don't need i_lock here, no concurrent mods to i_state */
A
Al Viro 已提交
515 516 517 518
	inode->i_state = I_FREEING | I_CLEAR;
}
EXPORT_SYMBOL(end_writeback);

D
Dave Chinner 已提交
519 520 521 522 523 524 525 526 527 528 529 530 531
/*
 * 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.
 */
532
static void evict(struct inode *inode)
533 534 535
{
	const struct super_operations *op = inode->i_sb->s_op;

D
Dave Chinner 已提交
536 537 538
	BUG_ON(!(inode->i_state & I_FREEING));
	BUG_ON(!list_empty(&inode->i_lru));

539 540 541
	if (!list_empty(&inode->i_wb_list))
		inode_wb_list_del(inode);

542 543
	inode_sb_list_del(inode);

A
Al Viro 已提交
544 545
	if (op->evict_inode) {
		op->evict_inode(inode);
546 547 548
	} else {
		if (inode->i_data.nrpages)
			truncate_inode_pages(&inode->i_data, 0);
A
Al Viro 已提交
549
		end_writeback(inode);
550
	}
551 552 553 554
	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
555 556 557 558 559 560 561 562 563

	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);
564 565
}

L
Linus Torvalds 已提交
566 567 568 569 570 571 572 573 574 575 576 577
/*
 * 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
578 579
		inode = list_first_entry(head, struct inode, i_lru);
		list_del_init(&inode->i_lru);
L
Linus Torvalds 已提交
580

581
		evict(inode);
L
Linus Torvalds 已提交
582 583 584
	}
}

A
Al Viro 已提交
585 586 587 588 589 590 591 592
/**
 * 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
593
 */
A
Al Viro 已提交
594
void evict_inodes(struct super_block *sb)
L
Linus Torvalds 已提交
595
{
A
Al Viro 已提交
596 597
	struct inode *inode, *next;
	LIST_HEAD(dispose);
L
Linus Torvalds 已提交
598

599
	spin_lock(&inode_sb_list_lock);
A
Al Viro 已提交
600 601
	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
602
			continue;
603 604 605 606

		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
607
			continue;
608
		}
A
Al Viro 已提交
609 610

		inode->i_state |= I_FREEING;
611
		inode_lru_list_del(inode);
612
		spin_unlock(&inode->i_lock);
613
		list_add(&inode->i_lru, &dispose);
L
Linus Torvalds 已提交
614
	}
615
	spin_unlock(&inode_sb_list_lock);
A
Al Viro 已提交
616 617

	dispose_list(&dispose);
L
Linus Torvalds 已提交
618 619 620
}

/**
621 622
 * invalidate_inodes	- attempt to free all inodes on a superblock
 * @sb:		superblock to operate on
623
 * @kill_dirty: flag to guide handling of dirty inodes
L
Linus Torvalds 已提交
624
 *
625 626
 * Attempts to free all inodes for a given superblock.  If there were any
 * busy inodes return a non-zero value, else zero.
627 628
 * If @kill_dirty is set, discard dirty inodes too, otherwise treat
 * them as busy.
L
Linus Torvalds 已提交
629
 */
630
int invalidate_inodes(struct super_block *sb, bool kill_dirty)
L
Linus Torvalds 已提交
631
{
632
	int busy = 0;
633 634
	struct inode *inode, *next;
	LIST_HEAD(dispose);
L
Linus Torvalds 已提交
635

636
	spin_lock(&inode_sb_list_lock);
637
	list_for_each_entry_safe(inode, next, &sb->s_inodes, i_sb_list) {
638 639 640
		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
641
			continue;
642
		}
643
		if (inode->i_state & I_DIRTY && !kill_dirty) {
644
			spin_unlock(&inode->i_lock);
645 646 647
			busy = 1;
			continue;
		}
648
		if (atomic_read(&inode->i_count)) {
649
			spin_unlock(&inode->i_lock);
650
			busy = 1;
L
Linus Torvalds 已提交
651 652
			continue;
		}
653 654

		inode->i_state |= I_FREEING;
655
		inode_lru_list_del(inode);
656
		spin_unlock(&inode->i_lock);
657
		list_add(&inode->i_lru, &dispose);
L
Linus Torvalds 已提交
658
	}
659
	spin_unlock(&inode_sb_list_lock);
L
Linus Torvalds 已提交
660

661
	dispose_list(&dispose);
L
Linus Torvalds 已提交
662 663 664 665 666 667

	return busy;
}

static int can_unuse(struct inode *inode)
{
668
	if (inode->i_state & ~I_REFERENCED)
L
Linus Torvalds 已提交
669 670 671 672 673 674 675 676 677 678 679
		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;
}

/*
680 681 682 683
 * 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
684 685
 *
 * Any inodes which are pinned purely because of attached pagecache have their
686 687
 * pagecache removed.  If the inode has metadata buffers attached to
 * mapping->private_list then try to remove them.
L
Linus Torvalds 已提交
688
 *
689 690 691 692 693 694 695
 * 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
696
 */
697
void prune_icache_sb(struct super_block *sb, int nr_to_scan)
L
Linus Torvalds 已提交
698 699 700 701 702
{
	LIST_HEAD(freeable);
	int nr_scanned;
	unsigned long reap = 0;

D
Dave Chinner 已提交
703
	spin_lock(&sb->s_inode_lru_lock);
704
	for (nr_scanned = nr_to_scan; nr_scanned >= 0; nr_scanned--) {
L
Linus Torvalds 已提交
705 706
		struct inode *inode;

707
		if (list_empty(&sb->s_inode_lru))
L
Linus Torvalds 已提交
708 709
			break;

710
		inode = list_entry(sb->s_inode_lru.prev, struct inode, i_lru);
L
Linus Torvalds 已提交
711

712
		/*
D
Dave Chinner 已提交
713
		 * we are inverting the sb->s_inode_lru_lock/inode->i_lock here,
714 715 716 717
		 * 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)) {
718
			list_move_tail(&inode->i_lru, &sb->s_inode_lru);
719 720 721
			continue;
		}

722 723 724 725 726 727
		/*
		 * 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
728
			list_del_init(&inode->i_lru);
729
			spin_unlock(&inode->i_lock);
730
			sb->s_nr_inodes_unused--;
731
			this_cpu_dec(nr_unused);
732 733 734 735 736 737
			continue;
		}

		/* recently referenced inodes get one more pass */
		if (inode->i_state & I_REFERENCED) {
			inode->i_state &= ~I_REFERENCED;
738
			list_move(&inode->i_lru, &sb->s_inode_lru);
739
			spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
740 741 742 743
			continue;
		}
		if (inode_has_buffers(inode) || inode->i_data.nrpages) {
			__iget(inode);
744
			spin_unlock(&inode->i_lock);
D
Dave Chinner 已提交
745
			spin_unlock(&sb->s_inode_lru_lock);
L
Linus Torvalds 已提交
746
			if (remove_inode_buffers(inode))
747 748
				reap += invalidate_mapping_pages(&inode->i_data,
								0, -1);
L
Linus Torvalds 已提交
749
			iput(inode);
D
Dave Chinner 已提交
750
			spin_lock(&sb->s_inode_lru_lock);
L
Linus Torvalds 已提交
751

752
			if (inode != list_entry(sb->s_inode_lru.next,
N
Nick Piggin 已提交
753
						struct inode, i_lru))
L
Linus Torvalds 已提交
754
				continue;	/* wrong inode or list_empty */
755 756 757
			/* avoid lock inversions with trylock */
			if (!spin_trylock(&inode->i_lock))
				continue;
758 759
			if (!can_unuse(inode)) {
				spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
760
				continue;
761
			}
L
Linus Torvalds 已提交
762
		}
763
		WARN_ON(inode->i_state & I_NEW);
L
Linus Torvalds 已提交
764
		inode->i_state |= I_FREEING;
765
		spin_unlock(&inode->i_lock);
N
Nick Piggin 已提交
766 767

		list_move(&inode->i_lru, &freeable);
768
		sb->s_nr_inodes_unused--;
769
		this_cpu_dec(nr_unused);
L
Linus Torvalds 已提交
770
	}
771 772 773 774
	if (current_is_kswapd())
		__count_vm_events(KSWAPD_INODESTEAL, reap);
	else
		__count_vm_events(PGINODESTEAL, reap);
D
Dave Chinner 已提交
775
	spin_unlock(&sb->s_inode_lru_lock);
776 777
	if (current->reclaim_state)
		current->reclaim_state->reclaimed_slab += reap;
L
Linus Torvalds 已提交
778 779 780 781 782 783 784 785

	dispose_list(&freeable);
}

static void __wait_on_freeing_inode(struct inode *inode);
/*
 * Called with the inode lock held.
 */
786 787 788 789
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
790 791
{
	struct hlist_node *node;
792
	struct inode *inode = NULL;
L
Linus Torvalds 已提交
793 794

repeat:
795
	hlist_for_each_entry(inode, node, head, i_hash) {
796 797 798
		spin_lock(&inode->i_lock);
		if (inode->i_sb != sb) {
			spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
799
			continue;
800 801 802
		}
		if (!test(inode, data)) {
			spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
803
			continue;
804
		}
A
Al Viro 已提交
805
		if (inode->i_state & (I_FREEING|I_WILL_FREE)) {
L
Linus Torvalds 已提交
806 807 808
			__wait_on_freeing_inode(inode);
			goto repeat;
		}
809
		__iget(inode);
810
		spin_unlock(&inode->i_lock);
811
		return inode;
L
Linus Torvalds 已提交
812
	}
813
	return NULL;
L
Linus Torvalds 已提交
814 815 816 817 818 819
}

/*
 * find_inode_fast is the fast path version of find_inode, see the comment at
 * iget_locked for details.
 */
820 821
static struct inode *find_inode_fast(struct super_block *sb,
				struct hlist_head *head, unsigned long ino)
L
Linus Torvalds 已提交
822 823
{
	struct hlist_node *node;
824
	struct inode *inode = NULL;
L
Linus Torvalds 已提交
825 826

repeat:
827
	hlist_for_each_entry(inode, node, head, i_hash) {
828 829 830
		spin_lock(&inode->i_lock);
		if (inode->i_ino != ino) {
			spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
831
			continue;
832 833 834
		}
		if (inode->i_sb != sb) {
			spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
835
			continue;
836
		}
A
Al Viro 已提交
837
		if (inode->i_state & (I_FREEING|I_WILL_FREE)) {
L
Linus Torvalds 已提交
838 839 840
			__wait_on_freeing_inode(inode);
			goto repeat;
		}
841
		__iget(inode);
842
		spin_unlock(&inode->i_lock);
843
		return inode;
L
Linus Torvalds 已提交
844
	}
845
	return NULL;
846 847
}

848 849 850 851
/*
 * 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.
852
 *
853 854 855 856 857 858 859 860 861
 * 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.
862
 */
863 864 865
#define LAST_INO_BATCH 1024
static DEFINE_PER_CPU(unsigned int, last_ino);

866
unsigned int get_next_ino(void)
867
{
868 869
	unsigned int *p = &get_cpu_var(last_ino);
	unsigned int res = *p;
870

871 872 873 874 875 876 877 878 879 880 881 882
#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;
883
}
884
EXPORT_SYMBOL(get_next_ino);
885

886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908
/**
 *	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
909 910 911 912
/**
 *	new_inode 	- obtain an inode
 *	@sb: superblock
 *
913
 *	Allocates a new inode for given superblock. The default gfp_mask
H
Hugh Dickins 已提交
914
 *	for allocations related to inode->i_mapping is GFP_HIGHUSER_MOVABLE.
915 916 917 918 919
 *	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
920 921 922
 */
struct inode *new_inode(struct super_block *sb)
{
923
	struct inode *inode;
L
Linus Torvalds 已提交
924

925
	spin_lock_prefetch(&inode_sb_list_lock);
926

927 928
	inode = new_inode_pseudo(sb);
	if (inode)
929
		inode_sb_list_add(inode);
L
Linus Torvalds 已提交
930 931 932 933
	return inode;
}
EXPORT_SYMBOL(new_inode);

934
#ifdef CONFIG_DEBUG_LOCK_ALLOC
935 936
void lockdep_annotate_inode_mutex_key(struct inode *inode)
{
937
	if (S_ISDIR(inode->i_mode)) {
938 939
		struct file_system_type *type = inode->i_sb->s_type;

940 941 942 943 944 945 946 947 948 949 950
		/* Set new key only if filesystem hasn't already changed it */
		if (!lockdep_match_class(&inode->i_mutex,
		    &type->i_mutex_key)) {
			/*
			 * 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);
		}
951
	}
952 953
}
EXPORT_SYMBOL(lockdep_annotate_inode_mutex_key);
954
#endif
955 956 957 958 959 960 961 962 963 964 965

/**
 * 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);
966
	spin_lock(&inode->i_lock);
C
Christoph Hellwig 已提交
967 968
	WARN_ON(!(inode->i_state & I_NEW));
	inode->i_state &= ~I_NEW;
969 970
	wake_up_bit(&inode->i_state, __I_NEW);
	spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
971 972 973
}
EXPORT_SYMBOL(unlock_new_inode);

C
Christoph Hellwig 已提交
974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989
/**
 * 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
990
 *
C
Christoph Hellwig 已提交
991 992
 * Note both @test and @set are called with the inode_hash_lock held, so can't
 * sleep.
L
Linus Torvalds 已提交
993
 */
C
Christoph Hellwig 已提交
994 995 996
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
997
{
C
Christoph Hellwig 已提交
998
	struct hlist_head *head = inode_hashtable + hash(sb, hashval);
999
	struct inode *inode;
L
Linus Torvalds 已提交
1000

C
Christoph Hellwig 已提交
1001 1002 1003 1004 1005 1006 1007 1008 1009
	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
1010 1011
	inode = alloc_inode(sb);
	if (inode) {
1012
		struct inode *old;
L
Linus Torvalds 已提交
1013

1014
		spin_lock(&inode_hash_lock);
L
Linus Torvalds 已提交
1015 1016 1017 1018 1019 1020
		/* We released the lock, so.. */
		old = find_inode(sb, head, test, data);
		if (!old) {
			if (set(inode, data))
				goto set_failed;

1021 1022
			spin_lock(&inode->i_lock);
			inode->i_state = I_NEW;
1023
			hlist_add_head(&inode->i_hash, head);
1024
			spin_unlock(&inode->i_lock);
1025
			inode_sb_list_add(inode);
1026
			spin_unlock(&inode_hash_lock);
L
Linus Torvalds 已提交
1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038

			/* 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.
		 */
1039
		spin_unlock(&inode_hash_lock);
L
Linus Torvalds 已提交
1040 1041 1042 1043 1044 1045 1046
		destroy_inode(inode);
		inode = old;
		wait_on_inode(inode);
	}
	return inode;

set_failed:
1047
	spin_unlock(&inode_hash_lock);
L
Linus Torvalds 已提交
1048 1049 1050
	destroy_inode(inode);
	return NULL;
}
C
Christoph Hellwig 已提交
1051
EXPORT_SYMBOL(iget5_locked);
L
Linus Torvalds 已提交
1052

C
Christoph Hellwig 已提交
1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064
/**
 * 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
1065
 */
C
Christoph Hellwig 已提交
1066
struct inode *iget_locked(struct super_block *sb, unsigned long ino)
L
Linus Torvalds 已提交
1067
{
C
Christoph Hellwig 已提交
1068
	struct hlist_head *head = inode_hashtable + hash(sb, ino);
1069
	struct inode *inode;
L
Linus Torvalds 已提交
1070

C
Christoph Hellwig 已提交
1071 1072 1073 1074 1075 1076 1077 1078
	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
1079 1080
	inode = alloc_inode(sb);
	if (inode) {
1081
		struct inode *old;
L
Linus Torvalds 已提交
1082

1083
		spin_lock(&inode_hash_lock);
L
Linus Torvalds 已提交
1084 1085 1086 1087
		/* We released the lock, so.. */
		old = find_inode_fast(sb, head, ino);
		if (!old) {
			inode->i_ino = ino;
1088 1089
			spin_lock(&inode->i_lock);
			inode->i_state = I_NEW;
1090
			hlist_add_head(&inode->i_hash, head);
1091
			spin_unlock(&inode->i_lock);
1092
			inode_sb_list_add(inode);
1093
			spin_unlock(&inode_hash_lock);
L
Linus Torvalds 已提交
1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105

			/* 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.
		 */
1106
		spin_unlock(&inode_hash_lock);
L
Linus Torvalds 已提交
1107 1108 1109 1110 1111 1112
		destroy_inode(inode);
		inode = old;
		wait_on_inode(inode);
	}
	return inode;
}
C
Christoph Hellwig 已提交
1113
EXPORT_SYMBOL(iget_locked);
L
Linus Torvalds 已提交
1114

1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127
/*
 * 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 hlist_node *node;
	struct inode *inode;

1128
	spin_lock(&inode_hash_lock);
1129
	hlist_for_each_entry(inode, node, b, i_hash) {
1130 1131
		if (inode->i_ino == ino && inode->i_sb == sb) {
			spin_unlock(&inode_hash_lock);
1132
			return 0;
1133
		}
1134
	}
1135
	spin_unlock(&inode_hash_lock);
1136 1137 1138 1139

	return 1;
}

L
Linus Torvalds 已提交
1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155
/**
 *	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)
{
1156 1157 1158 1159 1160
	/*
	 * 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.
	 */
1161
	static DEFINE_SPINLOCK(iunique_lock);
1162
	static unsigned int counter;
L
Linus Torvalds 已提交
1163
	ino_t res;
1164

1165
	spin_lock(&iunique_lock);
1166 1167 1168
	do {
		if (counter <= max_reserved)
			counter = max_reserved + 1;
L
Linus Torvalds 已提交
1169
		res = counter++;
1170 1171
	} while (!test_inode_iunique(sb, res));
	spin_unlock(&iunique_lock);
L
Linus Torvalds 已提交
1172

1173 1174
	return res;
}
L
Linus Torvalds 已提交
1175 1176 1177 1178
EXPORT_SYMBOL(iunique);

struct inode *igrab(struct inode *inode)
{
1179 1180
	spin_lock(&inode->i_lock);
	if (!(inode->i_state & (I_FREEING|I_WILL_FREE))) {
L
Linus Torvalds 已提交
1181
		__iget(inode);
1182 1183 1184
		spin_unlock(&inode->i_lock);
	} else {
		spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
1185 1186 1187 1188 1189 1190
		/*
		 * 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;
1191
	}
L
Linus Torvalds 已提交
1192 1193 1194 1195 1196
	return inode;
}
EXPORT_SYMBOL(igrab);

/**
C
Christoph Hellwig 已提交
1197
 * ilookup5_nowait - search for an inode in the inode cache
L
Linus Torvalds 已提交
1198
 * @sb:		super block of file system to search
C
Christoph Hellwig 已提交
1199
 * @hashval:	hash value (usually inode number) to search for
L
Linus Torvalds 已提交
1200 1201 1202
 * @test:	callback used for comparisons between inodes
 * @data:	opaque data pointer to pass to @test
 *
C
Christoph Hellwig 已提交
1203
 * Search for the inode specified by @hashval and @data in the inode cache.
L
Linus Torvalds 已提交
1204 1205 1206
 * If the inode is in the cache, the inode is returned with an incremented
 * reference count.
 *
C
Christoph Hellwig 已提交
1207 1208
 * 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
1209
 *
1210
 * Note2: @test is called with the inode_hash_lock held, so can't sleep.
L
Linus Torvalds 已提交
1211
 */
C
Christoph Hellwig 已提交
1212 1213
struct inode *ilookup5_nowait(struct super_block *sb, unsigned long hashval,
		int (*test)(struct inode *, void *), void *data)
L
Linus Torvalds 已提交
1214
{
C
Christoph Hellwig 已提交
1215
	struct hlist_head *head = inode_hashtable + hash(sb, hashval);
L
Linus Torvalds 已提交
1216 1217
	struct inode *inode;

1218
	spin_lock(&inode_hash_lock);
L
Linus Torvalds 已提交
1219
	inode = find_inode(sb, head, test, data);
1220
	spin_unlock(&inode_hash_lock);
1221

C
Christoph Hellwig 已提交
1222
	return inode;
1223 1224 1225 1226 1227 1228 1229 1230 1231 1232
}
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
1233 1234 1235
 * 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.
1236
 * returned with an incremented reference count.
L
Linus Torvalds 已提交
1237
 *
C
Christoph Hellwig 已提交
1238 1239
 * 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
1240
 *
C
Christoph Hellwig 已提交
1241
 * Note: @test is called with the inode_hash_lock held, so can't sleep.
L
Linus Torvalds 已提交
1242 1243 1244 1245
 */
struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
		int (*test)(struct inode *, void *), void *data)
{
C
Christoph Hellwig 已提交
1246
	struct inode *inode = ilookup5_nowait(sb, hashval, test, data);
L
Linus Torvalds 已提交
1247

C
Christoph Hellwig 已提交
1248 1249 1250
	if (inode)
		wait_on_inode(inode);
	return inode;
L
Linus Torvalds 已提交
1251 1252 1253 1254 1255 1256 1257 1258
}
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
1259 1260
 * 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
1261 1262 1263 1264 1265 1266
 */
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
1267 1268 1269
	spin_lock(&inode_hash_lock);
	inode = find_inode_fast(sb, head, ino);
	spin_unlock(&inode_hash_lock);
L
Linus Torvalds 已提交
1270 1271

	if (inode)
C
Christoph Hellwig 已提交
1272 1273
		wait_on_inode(inode);
	return inode;
L
Linus Torvalds 已提交
1274
}
C
Christoph Hellwig 已提交
1275
EXPORT_SYMBOL(ilookup);
L
Linus Torvalds 已提交
1276

A
Al Viro 已提交
1277 1278 1279 1280 1281 1282 1283
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) {
1284 1285
		struct hlist_node *node;
		struct inode *old = NULL;
1286
		spin_lock(&inode_hash_lock);
1287 1288 1289 1290 1291
		hlist_for_each_entry(old, node, head, i_hash) {
			if (old->i_ino != ino)
				continue;
			if (old->i_sb != sb)
				continue;
1292 1293 1294
			spin_lock(&old->i_lock);
			if (old->i_state & (I_FREEING|I_WILL_FREE)) {
				spin_unlock(&old->i_lock);
1295
				continue;
1296
			}
1297 1298 1299
			break;
		}
		if (likely(!node)) {
1300 1301
			spin_lock(&inode->i_lock);
			inode->i_state |= I_NEW;
A
Al Viro 已提交
1302
			hlist_add_head(&inode->i_hash, head);
1303
			spin_unlock(&inode->i_lock);
1304
			spin_unlock(&inode_hash_lock);
A
Al Viro 已提交
1305 1306 1307
			return 0;
		}
		__iget(old);
1308
		spin_unlock(&old->i_lock);
1309
		spin_unlock(&inode_hash_lock);
A
Al Viro 已提交
1310
		wait_on_inode(old);
A
Al Viro 已提交
1311
		if (unlikely(!inode_unhashed(old))) {
A
Al Viro 已提交
1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326
			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) {
1327 1328 1329
		struct hlist_node *node;
		struct inode *old = NULL;

1330
		spin_lock(&inode_hash_lock);
1331 1332 1333 1334 1335
		hlist_for_each_entry(old, node, head, i_hash) {
			if (old->i_sb != sb)
				continue;
			if (!test(old, data))
				continue;
1336 1337 1338
			spin_lock(&old->i_lock);
			if (old->i_state & (I_FREEING|I_WILL_FREE)) {
				spin_unlock(&old->i_lock);
1339
				continue;
1340
			}
1341 1342 1343
			break;
		}
		if (likely(!node)) {
1344 1345
			spin_lock(&inode->i_lock);
			inode->i_state |= I_NEW;
A
Al Viro 已提交
1346
			hlist_add_head(&inode->i_hash, head);
1347
			spin_unlock(&inode->i_lock);
1348
			spin_unlock(&inode_hash_lock);
A
Al Viro 已提交
1349 1350 1351
			return 0;
		}
		__iget(old);
1352
		spin_unlock(&old->i_lock);
1353
		spin_unlock(&inode_hash_lock);
A
Al Viro 已提交
1354
		wait_on_inode(old);
A
Al Viro 已提交
1355
		if (unlikely(!inode_unhashed(old))) {
A
Al Viro 已提交
1356 1357 1358 1359 1360 1361 1362 1363
			iput(old);
			return -EBUSY;
		}
		iput(old);
	}
}
EXPORT_SYMBOL(insert_inode_locked4);

L
Linus Torvalds 已提交
1364

1365 1366 1367 1368 1369 1370
int generic_delete_inode(struct inode *inode)
{
	return 1;
}
EXPORT_SYMBOL(generic_delete_inode);

L
Linus Torvalds 已提交
1371
/*
1372 1373 1374
 * Normal UNIX filesystem behaviour: delete the
 * inode when the usage count drops to zero, and
 * i_nlink is zero.
L
Linus Torvalds 已提交
1375
 */
1376
int generic_drop_inode(struct inode *inode)
L
Linus Torvalds 已提交
1377
{
A
Al Viro 已提交
1378
	return !inode->i_nlink || inode_unhashed(inode);
L
Linus Torvalds 已提交
1379
}
1380
EXPORT_SYMBOL_GPL(generic_drop_inode);
L
Linus Torvalds 已提交
1381

1382 1383 1384
/*
 * Called when we're dropping the last reference
 * to an inode.
1385
 *
1386 1387 1388 1389 1390
 * 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.
1391
 */
1392
static void iput_final(struct inode *inode)
L
Linus Torvalds 已提交
1393 1394
{
	struct super_block *sb = inode->i_sb;
1395 1396 1397
	const struct super_operations *op = inode->i_sb->s_op;
	int drop;

1398 1399
	WARN_ON(inode->i_state & I_NEW);

1400
	if (op->drop_inode)
1401 1402 1403
		drop = op->drop_inode(inode);
	else
		drop = generic_drop_inode(inode);
L
Linus Torvalds 已提交
1404

D
Dave Chinner 已提交
1405 1406 1407 1408 1409 1410 1411 1412
	if (!drop && (sb->s_flags & MS_ACTIVE)) {
		inode->i_state |= I_REFERENCED;
		if (!(inode->i_state & (I_DIRTY|I_SYNC)))
			inode_lru_list_add(inode);
		spin_unlock(&inode->i_lock);
		return;
	}

1413
	if (!drop) {
1414
		inode->i_state |= I_WILL_FREE;
1415
		spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
1416
		write_inode_now(inode, 1);
1417
		spin_lock(&inode->i_lock);
1418
		WARN_ON(inode->i_state & I_NEW);
1419
		inode->i_state &= ~I_WILL_FREE;
L
Linus Torvalds 已提交
1420
	}
N
Nick Piggin 已提交
1421

1422
	inode->i_state |= I_FREEING;
1423 1424
	if (!list_empty(&inode->i_lru))
		inode_lru_list_del(inode);
D
Dave Chinner 已提交
1425 1426
	spin_unlock(&inode->i_lock);

1427
	evict(inode);
L
Linus Torvalds 已提交
1428 1429 1430
}

/**
1431
 *	iput	- put an inode
L
Linus Torvalds 已提交
1432 1433 1434 1435 1436 1437 1438 1439 1440 1441
 *	@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
1442
		BUG_ON(inode->i_state & I_CLEAR);
L
Linus Torvalds 已提交
1443

1444
		if (atomic_dec_and_lock(&inode->i_count, &inode->i_lock))
L
Linus Torvalds 已提交
1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457
			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
1458
 *	disk block relative to the disk start that holds that block of the
L
Linus Torvalds 已提交
1459 1460
 *	file.
 */
1461
sector_t bmap(struct inode *inode, sector_t block)
L
Linus Torvalds 已提交
1462 1463 1464 1465 1466 1467 1468 1469
{
	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);

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 1499 1500 1501 1502 1503
/*
 * 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;
}

L
Linus Torvalds 已提交
1504
/**
C
Christoph Hellwig 已提交
1505 1506
 *	touch_atime	-	update the access time
 *	@mnt: mount the inode is accessed on
1507
 *	@dentry: dentry accessed
L
Linus Torvalds 已提交
1508 1509 1510 1511 1512
 *
 *	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.
 */
C
Christoph Hellwig 已提交
1513
void touch_atime(struct vfsmount *mnt, struct dentry *dentry)
L
Linus Torvalds 已提交
1514
{
C
Christoph Hellwig 已提交
1515
	struct inode *inode = dentry->d_inode;
L
Linus Torvalds 已提交
1516 1517
	struct timespec now;

1518
	if (inode->i_flags & S_NOATIME)
A
Andi Kleen 已提交
1519
		return;
1520
	if (IS_NOATIME(inode))
A
Andi Kleen 已提交
1521
		return;
A
Andrew Morton 已提交
1522
	if ((inode->i_sb->s_flags & MS_NODIRATIME) && S_ISDIR(inode->i_mode))
A
Andi Kleen 已提交
1523
		return;
V
Valerie Henson 已提交
1524

1525
	if (mnt->mnt_flags & MNT_NOATIME)
A
Andi Kleen 已提交
1526
		return;
1527
	if ((mnt->mnt_flags & MNT_NODIRATIME) && S_ISDIR(inode->i_mode))
A
Andi Kleen 已提交
1528
		return;
L
Linus Torvalds 已提交
1529 1530

	now = current_fs_time(inode->i_sb);
1531 1532

	if (!relatime_need_update(mnt, inode, now))
A
Andi Kleen 已提交
1533
		return;
1534

V
Valerie Henson 已提交
1535
	if (timespec_equal(&inode->i_atime, &now))
A
Andi Kleen 已提交
1536 1537 1538 1539
		return;

	if (mnt_want_write(mnt))
		return;
V
Valerie Henson 已提交
1540 1541 1542

	inode->i_atime = now;
	mark_inode_dirty_sync(inode);
1543
	mnt_drop_write(mnt);
L
Linus Torvalds 已提交
1544
}
C
Christoph Hellwig 已提交
1545
EXPORT_SYMBOL(touch_atime);
L
Linus Torvalds 已提交
1546 1547

/**
1548 1549
 *	file_update_time	-	update mtime and ctime time
 *	@file: file accessed
L
Linus Torvalds 已提交
1550
 *
1551 1552 1553 1554
 *	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
1555
 *	S_NOCMTIME inode flag, e.g. for network filesystem where these
1556
 *	timestamps are handled by the server.
L
Linus Torvalds 已提交
1557 1558
 */

1559
void file_update_time(struct file *file)
L
Linus Torvalds 已提交
1560
{
1561
	struct inode *inode = file->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
1562
	struct timespec now;
A
Andi Kleen 已提交
1563
	enum { S_MTIME = 1, S_CTIME = 2, S_VERSION = 4 } sync_it = 0;
L
Linus Torvalds 已提交
1564

A
Andi Kleen 已提交
1565
	/* First try to exhaust all avenues to not sync */
L
Linus Torvalds 已提交
1566 1567
	if (IS_NOCMTIME(inode))
		return;
1568

L
Linus Torvalds 已提交
1569
	now = current_fs_time(inode->i_sb);
A
Andi Kleen 已提交
1570 1571
	if (!timespec_equal(&inode->i_mtime, &now))
		sync_it = S_MTIME;
L
Linus Torvalds 已提交
1572

A
Andi Kleen 已提交
1573 1574
	if (!timespec_equal(&inode->i_ctime, &now))
		sync_it |= S_CTIME;
1575

A
Andi Kleen 已提交
1576 1577
	if (IS_I_VERSION(inode))
		sync_it |= S_VERSION;
1578

A
Andi Kleen 已提交
1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593
	if (!sync_it)
		return;

	/* Finally allowed to write? Takes lock. */
	if (mnt_want_write_file(file))
		return;

	/* Only change inode inside the lock region */
	if (sync_it & S_VERSION)
		inode_inc_iversion(inode);
	if (sync_it & S_CTIME)
		inode->i_ctime = now;
	if (sync_it & S_MTIME)
		inode->i_mtime = now;
	mark_inode_dirty_sync(inode);
A
Al Viro 已提交
1594
	mnt_drop_write_file(file);
L
Linus Torvalds 已提交
1595
}
1596
EXPORT_SYMBOL(file_update_time);
L
Linus Torvalds 已提交
1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612

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;
}
1613
EXPORT_SYMBOL(inode_wait);
L
Linus Torvalds 已提交
1614 1615

/*
1616 1617 1618 1619 1620 1621
 * 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
1622
 * It doesn't matter if I_NEW is not set initially, a call to
1623 1624
 * wake_up_bit(&inode->i_state, __I_NEW) after removing from the hash list
 * will DTRT.
L
Linus Torvalds 已提交
1625 1626 1627 1628
 */
static void __wait_on_freeing_inode(struct inode *inode)
{
	wait_queue_head_t *wq;
C
Christoph Hellwig 已提交
1629 1630
	DEFINE_WAIT_BIT(wait, &inode->i_state, __I_NEW);
	wq = bit_waitqueue(&inode->i_state, __I_NEW);
L
Linus Torvalds 已提交
1631
	prepare_to_wait(wq, &wait.wait, TASK_UNINTERRUPTIBLE);
1632
	spin_unlock(&inode->i_lock);
1633
	spin_unlock(&inode_hash_lock);
L
Linus Torvalds 已提交
1634 1635
	schedule();
	finish_wait(wq, &wait.wait);
1636
	spin_lock(&inode_hash_lock);
L
Linus Torvalds 已提交
1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653
}

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)
{
1654
	unsigned int loop;
L
Linus Torvalds 已提交
1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671

	/* 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,
					0);

1672
	for (loop = 0; loop < (1U << i_hash_shift); loop++)
L
Linus Torvalds 已提交
1673 1674 1675
		INIT_HLIST_HEAD(&inode_hashtable[loop]);
}

1676
void __init inode_init(void)
L
Linus Torvalds 已提交
1677
{
1678
	unsigned int loop;
L
Linus Torvalds 已提交
1679 1680

	/* inode slab cache */
1681 1682 1683 1684 1685
	inode_cachep = kmem_cache_create("inode_cache",
					 sizeof(struct inode),
					 0,
					 (SLAB_RECLAIM_ACCOUNT|SLAB_PANIC|
					 SLAB_MEM_SPREAD),
1686
					 init_once);
L
Linus Torvalds 已提交
1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701

	/* 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,
					0);

1702
	for (loop = 0; loop < (1U << i_hash_shift); loop++)
L
Linus Torvalds 已提交
1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719
		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))
		inode->i_fop = &def_fifo_fops;
	else if (S_ISSOCK(mode))
		inode->i_fop = &bad_sock_fops;
	else
1720 1721 1722
		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
1723 1724
}
EXPORT_SYMBOL(init_special_inode);
1725 1726

/**
1727
 * inode_init_owner - Init uid,gid,mode for new inode according to posix standards
1728 1729 1730 1731 1732
 * @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
1733
			umode_t mode)
1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744
{
	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);
1745

1746 1747 1748 1749 1750 1751
/**
 * 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.
1752
 */
1753
bool inode_owner_or_capable(const struct inode *inode)
1754 1755 1756 1757 1758 1759 1760 1761 1762
{
	struct user_namespace *ns = inode_userns(inode);

	if (current_user_ns() == ns && current_fsuid() == inode->i_uid)
		return true;
	if (ns_capable(ns, CAP_FOWNER))
		return true;
	return false;
}
1763
EXPORT_SYMBOL(inode_owner_or_capable);