super.c 40.4 KB
Newer Older
1
// SPDX-License-Identifier: GPL-2.0
Linus Torvalds's avatar
Linus Torvalds committed
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
/*
 *  linux/fs/super.c
 *
 *  Copyright (C) 1991, 1992  Linus Torvalds
 *
 *  super.c contains code to handle: - mount structures
 *                                   - super-block tables
 *                                   - filesystem drivers list
 *                                   - mount system call
 *                                   - umount system call
 *                                   - ustat system call
 *
 * GK 2/5/95  -  Changed to support mounting the root fs via NFS
 *
 *  Added kerneld support: Jacques Gelinas and Bjorn Ekwall
 *  Added change_root: Werner Almesberger & Hans Lermen, Feb '96
 *  Added options to /proc/mounts:
19
 *    Torbjörn Lindh (torbjorn.lindh@gopta.se), April 14, 1996.
Linus Torvalds's avatar
Linus Torvalds committed
20 21 22 23
 *  Added devfs support: Richard Gooch <rgooch@atnf.csiro.au>, 13-JAN-1998
 *  Heavily rewritten for 'one fs - one tree' dcache architecture. AV, Mar 2000
 */

24
#include <linux/export.h>
Linus Torvalds's avatar
Linus Torvalds committed
25 26 27 28 29 30
#include <linux/slab.h>
#include <linux/blkdev.h>
#include <linux/mount.h>
#include <linux/security.h>
#include <linux/writeback.h>		/* for the emergency remount stuff */
#include <linux/idr.h>
Ingo Molnar's avatar
Ingo Molnar committed
31
#include <linux/mutex.h>
32
#include <linux/backing-dev.h>
33
#include <linux/rculist_bl.h>
34
#include <linux/cleancache.h>
Al Viro's avatar
Al Viro committed
35
#include <linux/fsnotify.h>
36
#include <linux/lockdep.h>
37
#include <linux/user_namespace.h>
38
#include "internal.h"
Linus Torvalds's avatar
Linus Torvalds committed
39

40
static int thaw_super_locked(struct super_block *sb);
Linus Torvalds's avatar
Linus Torvalds committed
41

42 43
static LIST_HEAD(super_blocks);
static DEFINE_SPINLOCK(sb_lock);
Linus Torvalds's avatar
Linus Torvalds committed
44

45 46 47 48 49 50
static char *sb_writers_name[SB_FREEZE_LEVELS] = {
	"sb_writers",
	"sb_pagefaults",
	"sb_internal",
};

51 52 53 54 55 56 57
/*
 * One thing we have to be careful of with a per-sb shrinker is that we don't
 * drop the last active reference to the superblock from within the shrinker.
 * If that happens we could trigger unregistering the shrinker from within the
 * shrinker path and that leads to deadlock on the shrinker_rwsem. Hence we
 * take a passive reference to the superblock to avoid this from occurring.
 */
58 59
static unsigned long super_cache_scan(struct shrinker *shrink,
				      struct shrink_control *sc)
60 61
{
	struct super_block *sb;
62 63 64 65 66
	long	fs_objects = 0;
	long	total_objects;
	long	freed = 0;
	long	dentries;
	long	inodes;
67 68 69 70 71 72 73

	sb = container_of(shrink, struct super_block, s_shrink);

	/*
	 * Deadlock avoidance.  We may hold various FS locks, and we don't want
	 * to recurse into the FS that called us in clear_inode() and friends..
	 */
74 75
	if (!(sc->gfp_mask & __GFP_FS))
		return SHRINK_STOP;
76

77
	if (!trylock_super(sb))
78
		return SHRINK_STOP;
79

80
	if (sb->s_op->nr_cached_objects)
81
		fs_objects = sb->s_op->nr_cached_objects(sb, sc);
82

83 84
	inodes = list_lru_shrink_count(&sb->s_inode_lru, sc);
	dentries = list_lru_shrink_count(&sb->s_dentry_lru, sc);
85
	total_objects = dentries + inodes + fs_objects + 1;
86 87
	if (!total_objects)
		total_objects = 1;
88

89
	/* proportion the scan between the caches */
90
	dentries = mult_frac(sc->nr_to_scan, dentries, total_objects);
91
	inodes = mult_frac(sc->nr_to_scan, inodes, total_objects);
92
	fs_objects = mult_frac(sc->nr_to_scan, fs_objects, total_objects);
93

94 95 96
	/*
	 * prune the dcache first as the icache is pinned by it, then
	 * prune the icache, followed by the filesystem specific caches
97 98 99
	 *
	 * Ensure that we always scan at least one object - memcg kmem
	 * accounting uses this to fully empty the caches.
100
	 */
101
	sc->nr_to_scan = dentries + 1;
102
	freed = prune_dcache_sb(sb, sc);
103
	sc->nr_to_scan = inodes + 1;
104
	freed += prune_icache_sb(sb, sc);
105 106

	if (fs_objects) {
107
		sc->nr_to_scan = fs_objects + 1;
108
		freed += sb->s_op->free_cached_objects(sb, sc);
109 110
	}

111
	up_read(&sb->s_umount);
112 113 114 115 116 117 118 119 120 121 122
	return freed;
}

static unsigned long super_cache_count(struct shrinker *shrink,
				       struct shrink_control *sc)
{
	struct super_block *sb;
	long	total_objects = 0;

	sb = container_of(shrink, struct super_block, s_shrink);

123
	/*
124 125 126 127 128 129 130 131 132 133 134 135
	 * We don't call trylock_super() here as it is a scalability bottleneck,
	 * so we're exposed to partial setup state. The shrinker rwsem does not
	 * protect filesystem operations backing list_lru_shrink_count() or
	 * s_op->nr_cached_objects(). Counts can change between
	 * super_cache_count and super_cache_scan, so we really don't need locks
	 * here.
	 *
	 * However, if we are currently mounting the superblock, the underlying
	 * filesystem might be in a state of partial construction and hence it
	 * is dangerous to access it.  trylock_super() uses a SB_BORN check to
	 * avoid this situation, so do the same here. The memory barrier is
	 * matched with the one in mount_fs() as we don't hold locks here.
136
	 */
137 138 139 140
	if (!(sb->s_flags & SB_BORN))
		return 0;
	smp_rmb();

141
	if (sb->s_op && sb->s_op->nr_cached_objects)
142
		total_objects = sb->s_op->nr_cached_objects(sb, sc);
143

144 145
	total_objects += list_lru_shrink_count(&sb->s_dentry_lru, sc);
	total_objects += list_lru_shrink_count(&sb->s_inode_lru, sc);
146

147 148 149
	if (!total_objects)
		return SHRINK_EMPTY;

150
	total_objects = vfs_pressure_ratio(total_objects);
151
	return total_objects;
152 153
}

154 155 156 157 158 159 160
static void destroy_super_work(struct work_struct *work)
{
	struct super_block *s = container_of(work, struct super_block,
							destroy_work);
	int i;

	for (i = 0; i < SB_FREEZE_LEVELS; i++)
161
		percpu_free_rwsem(&s->s_writers.rw_sem[i]);
162 163 164 165 166 167 168 169 170 171
	kfree(s);
}

static void destroy_super_rcu(struct rcu_head *head)
{
	struct super_block *s = container_of(head, struct super_block, rcu);
	INIT_WORK(&s->destroy_work, destroy_super_work);
	schedule_work(&s->destroy_work);
}

172 173
/* Free a superblock that has never been seen by anyone */
static void destroy_unused_super(struct super_block *s)
174
{
175 176 177
	if (!s)
		return;
	up_write(&s->s_umount);
178 179 180
	list_lru_destroy(&s->s_dentry_lru);
	list_lru_destroy(&s->s_inode_lru);
	security_sb_free(s);
181
	put_user_ns(s->s_user_ns);
182
	kfree(s->s_subtype);
183
	free_prealloced_shrinker(&s->s_shrink);
184 185
	/* no delays needed */
	destroy_super_work(&s->destroy_work);
186 187
}

Linus Torvalds's avatar
Linus Torvalds committed
188 189
/**
 *	alloc_super	-	create new superblock
190
 *	@type:	filesystem type superblock should belong to
191
 *	@flags: the mount flags
192
 *	@user_ns: User namespace for the super_block
Linus Torvalds's avatar
Linus Torvalds committed
193 194 195 196
 *
 *	Allocates and initializes a new &struct super_block.  alloc_super()
 *	returns a pointer new superblock or %NULL if allocation had failed.
 */
197 198
static struct super_block *alloc_super(struct file_system_type *type, int flags,
				       struct user_namespace *user_ns)
Linus Torvalds's avatar
Linus Torvalds committed
199
{
200
	struct super_block *s = kzalloc(sizeof(struct super_block),  GFP_USER);
201
	static const struct super_operations default_op;
202 203 204 205
	int i;

	if (!s)
		return NULL;
Linus Torvalds's avatar
Linus Torvalds committed
206

207
	INIT_LIST_HEAD(&s->s_mounts);
208
	s->s_user_ns = get_user_ns(user_ns);
209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226
	init_rwsem(&s->s_umount);
	lockdep_set_class(&s->s_umount, &type->s_umount_key);
	/*
	 * sget() can have s_umount recursion.
	 *
	 * When it cannot find a suitable sb, it allocates a new
	 * one (this one), and tries again to find a suitable old
	 * one.
	 *
	 * In case that succeeds, it will acquire the s_umount
	 * lock of the old one. Since these are clearly distrinct
	 * locks, and this object isn't exposed yet, there's no
	 * risk of deadlocks.
	 *
	 * Annotate this by putting this lock in a different
	 * subclass.
	 */
	down_write_nested(&s->s_umount, SINGLE_DEPTH_NESTING);
227

228 229
	if (security_sb_alloc(s))
		goto fail;
230

231
	for (i = 0; i < SB_FREEZE_LEVELS; i++) {
232 233 234
		if (__percpu_init_rwsem(&s->s_writers.rw_sem[i],
					sb_writers_name[i],
					&type->s_writers_key[i]))
235
			goto fail;
Linus Torvalds's avatar
Linus Torvalds committed
236
	}
237
	init_waitqueue_head(&s->s_writers.wait_unfrozen);
238
	s->s_bdi = &noop_backing_dev_info;
239
	s->s_flags = flags;
240
	if (s->s_user_ns != &init_user_ns)
241
		s->s_iflags |= SB_I_NODEV;
242
	INIT_HLIST_NODE(&s->s_instances);
243
	INIT_HLIST_BL_HEAD(&s->s_roots);
244
	mutex_init(&s->s_sync_lock);
245
	INIT_LIST_HEAD(&s->s_inodes);
246
	spin_lock_init(&s->s_inode_list_lock);
247 248
	INIT_LIST_HEAD(&s->s_inodes_wb);
	spin_lock_init(&s->s_inode_wblist_lock);
249 250 251 252 253

	s->s_count = 1;
	atomic_set(&s->s_active, 1);
	mutex_init(&s->s_vfs_rename_mutex);
	lockdep_set_class(&s->s_vfs_rename_mutex, &type->s_vfs_rename_key);
254
	init_rwsem(&s->s_dquot.dqio_sem);
255 256 257
	s->s_maxbytes = MAX_NON_LFS;
	s->s_op = &default_op;
	s->s_time_gran = 1000000000;
258
	s->cleancache_poolid = CLEANCACHE_NO_POOL;
259 260 261 262 263

	s->s_shrink.seeks = DEFAULT_SEEKS;
	s->s_shrink.scan_objects = super_cache_scan;
	s->s_shrink.count_objects = super_cache_count;
	s->s_shrink.batch = 1024;
264
	s->s_shrink.flags = SHRINKER_NUMA_AWARE | SHRINKER_MEMCG_AWARE;
265 266
	if (prealloc_shrinker(&s->s_shrink))
		goto fail;
267
	if (list_lru_init_memcg(&s->s_dentry_lru, &s->s_shrink))
268
		goto fail;
269
	if (list_lru_init_memcg(&s->s_inode_lru, &s->s_shrink))
270
		goto fail;
Linus Torvalds's avatar
Linus Torvalds committed
271
	return s;
272

273
fail:
274
	destroy_unused_super(s);
275
	return NULL;
Linus Torvalds's avatar
Linus Torvalds committed
276 277 278 279 280
}

/* Superblock refcounting  */

/*
281
 * Drop a superblock's refcount.  The caller must hold sb_lock.
Linus Torvalds's avatar
Linus Torvalds committed
282
 */
283
static void __put_super(struct super_block *s)
Linus Torvalds's avatar
Linus Torvalds committed
284
{
285 286 287 288 289 290 291 292 293
	if (!--s->s_count) {
		list_del_init(&s->s_list);
		WARN_ON(s->s_dentry_lru.node);
		WARN_ON(s->s_inode_lru.node);
		WARN_ON(!list_empty(&s->s_mounts));
		security_sb_free(s);
		put_user_ns(s->s_user_ns);
		kfree(s->s_subtype);
		call_rcu(&s->rcu, destroy_super_rcu);
Linus Torvalds's avatar
Linus Torvalds committed
294 295 296 297 298 299 300 301 302 303
	}
}

/**
 *	put_super	-	drop a temporary reference to superblock
 *	@sb: superblock in question
 *
 *	Drops a temporary reference, frees superblock if there's no
 *	references left.
 */
Al Viro's avatar
Al Viro committed
304
static void put_super(struct super_block *sb)
Linus Torvalds's avatar
Linus Torvalds committed
305 306 307 308 309 310 311 312
{
	spin_lock(&sb_lock);
	__put_super(sb);
	spin_unlock(&sb_lock);
}


/**
313
 *	deactivate_locked_super	-	drop an active reference to superblock
Linus Torvalds's avatar
Linus Torvalds committed
314 315
 *	@s: superblock to deactivate
 *
316
 *	Drops an active reference to superblock, converting it into a temporary
317
 *	one if there is no other active references left.  In that case we
Linus Torvalds's avatar
Linus Torvalds committed
318 319
 *	tell fs driver to shut it down and drop the temporary reference we
 *	had just acquired.
320 321
 *
 *	Caller holds exclusive lock on superblock; that lock is released.
Linus Torvalds's avatar
Linus Torvalds committed
322
 */
323
void deactivate_locked_super(struct super_block *s)
Linus Torvalds's avatar
Linus Torvalds committed
324 325
{
	struct file_system_type *fs = s->s_type;
Al Viro's avatar
Al Viro committed
326
	if (atomic_dec_and_test(&s->s_active)) {
327
		cleancache_invalidate_fs(s);
328
		unregister_shrinker(&s->s_shrink);
329
		fs->kill_sb(s);
330

331 332 333 334 335 336 337 338
		/*
		 * Since list_lru_destroy() may sleep, we cannot call it from
		 * put_super(), where we hold the sb_lock. Therefore we destroy
		 * the lru lists right now.
		 */
		list_lru_destroy(&s->s_dentry_lru);
		list_lru_destroy(&s->s_inode_lru);

Linus Torvalds's avatar
Linus Torvalds committed
339 340
		put_filesystem(fs);
		put_super(s);
341 342
	} else {
		up_write(&s->s_umount);
Linus Torvalds's avatar
Linus Torvalds committed
343 344 345
	}
}

346
EXPORT_SYMBOL(deactivate_locked_super);
Linus Torvalds's avatar
Linus Torvalds committed
347

348
/**
349
 *	deactivate_super	-	drop an active reference to superblock
350 351
 *	@s: superblock to deactivate
 *
352 353 354
 *	Variant of deactivate_locked_super(), except that superblock is *not*
 *	locked by caller.  If we are going to drop the final active reference,
 *	lock will be acquired prior to that.
355
 */
356
void deactivate_super(struct super_block *s)
357
{
358 359 360
        if (!atomic_add_unless(&s->s_active, -1, 1)) {
		down_write(&s->s_umount);
		deactivate_locked_super(s);
361 362 363
	}
}

364
EXPORT_SYMBOL(deactivate_super);
365

Linus Torvalds's avatar
Linus Torvalds committed
366 367 368 369 370 371 372 373 374
/**
 *	grab_super - acquire an active reference
 *	@s: reference we are trying to make active
 *
 *	Tries to acquire an active reference.  grab_super() is used when we
 * 	had just found a superblock in super_blocks or fs_type->fs_supers
 *	and want to turn it into a full-blown active reference.  grab_super()
 *	is called with sb_lock held and drops it.  Returns 1 in case of
 *	success, 0 if we had failed (superblock contents was already dead or
Al Viro's avatar
Al Viro committed
375 376 377
 *	dying when grab_super() had been called).  Note that this is only
 *	called for superblocks not in rundown mode (== ones still on ->fs_supers
 *	of their type), so increment of ->s_count is OK here.
Linus Torvalds's avatar
Linus Torvalds committed
378
 */
379
static int grab_super(struct super_block *s) __releases(sb_lock)
Linus Torvalds's avatar
Linus Torvalds committed
380 381 382 383
{
	s->s_count++;
	spin_unlock(&sb_lock);
	down_write(&s->s_umount);
384
	if ((s->s_flags & SB_BORN) && atomic_inc_not_zero(&s->s_active)) {
Al Viro's avatar
Al Viro committed
385 386 387
		put_super(s);
		return 1;
	}
Linus Torvalds's avatar
Linus Torvalds committed
388 389 390 391 392
	up_write(&s->s_umount);
	put_super(s);
	return 0;
}

393
/*
394
 *	trylock_super - try to grab ->s_umount shared
395
 *	@sb: reference we are trying to grab
396
 *
397
 *	Try to prevent fs shutdown.  This is used in places where we
398
 *	cannot take an active reference but we need to ensure that the
399 400 401 402 403 404 405 406 407 408
 *	filesystem is not shut down while we are working on it. It returns
 *	false if we cannot acquire s_umount or if we lose the race and
 *	filesystem already got into shutdown, and returns true with the s_umount
 *	lock held in read mode in case of success. On successful return,
 *	the caller must drop the s_umount lock when done.
 *
 *	Note that unlike get_super() et.al. this one does *not* bump ->s_count.
 *	The reason why it's safe is that we are OK with doing trylock instead
 *	of down_read().  There's a couple of places that are OK with that, but
 *	it's very much not a general-purpose interface.
409
 */
410
bool trylock_super(struct super_block *sb)
411 412
{
	if (down_read_trylock(&sb->s_umount)) {
413
		if (!hlist_unhashed(&sb->s_instances) &&
414
		    sb->s_root && (sb->s_flags & SB_BORN))
415 416 417 418 419 420 421
			return true;
		up_read(&sb->s_umount);
	}

	return false;
}

Linus Torvalds's avatar
Linus Torvalds committed
422 423 424 425 426 427 428 429 430
/**
 *	generic_shutdown_super	-	common helper for ->kill_sb()
 *	@sb: superblock to kill
 *
 *	generic_shutdown_super() does all fs-independent work on superblock
 *	shutdown.  Typical ->kill_sb() should pick all fs-specific objects
 *	that need destruction out of superblock, call generic_shutdown_super()
 *	and release aforementioned objects.  Note: dentries and inodes _are_
 *	taken care of and do not need specific handling.
431 432 433 434
 *
 *	Upon calling this function, the filesystem may no longer alter or
 *	rearrange the set of dentries belonging to this super_block, nor may it
 *	change the attachments of dentries to inodes.
Linus Torvalds's avatar
Linus Torvalds committed
435 436 437
 */
void generic_shutdown_super(struct super_block *sb)
{
438
	const struct super_operations *sop = sb->s_op;
Linus Torvalds's avatar
Linus Torvalds committed
439

440 441
	if (sb->s_root) {
		shrink_dcache_for_umount(sb);
442
		sync_filesystem(sb);
443
		sb->s_flags &= ~SB_ACTIVE;
444

445
		fsnotify_sb_delete(sb);
446
		cgroup_writeback_umount();
Al Viro's avatar
Al Viro committed
447 448

		evict_inodes(sb);
Linus Torvalds's avatar
Linus Torvalds committed
449

450 451 452 453 454
		if (sb->s_dio_done_wq) {
			destroy_workqueue(sb->s_dio_done_wq);
			sb->s_dio_done_wq = NULL;
		}

Linus Torvalds's avatar
Linus Torvalds committed
455 456 457
		if (sop->put_super)
			sop->put_super(sb);

Al Viro's avatar
Al Viro committed
458
		if (!list_empty(&sb->s_inodes)) {
459 460 461
			printk("VFS: Busy inodes after unmount of %s. "
			   "Self-destruct in 5 seconds.  Have a nice day...\n",
			   sb->s_id);
Linus Torvalds's avatar
Linus Torvalds committed
462 463 464 465
		}
	}
	spin_lock(&sb_lock);
	/* should be initialized for __put_super_and_need_restart() */
466
	hlist_del_init(&sb->s_instances);
Linus Torvalds's avatar
Linus Torvalds committed
467 468
	spin_unlock(&sb_lock);
	up_write(&sb->s_umount);
Jan Kara's avatar
Jan Kara committed
469
	if (sb->s_bdi != &noop_backing_dev_info) {
470 471 472
		bdi_put(sb->s_bdi);
		sb->s_bdi = &noop_backing_dev_info;
	}
Linus Torvalds's avatar
Linus Torvalds committed
473 474 475 476 477
}

EXPORT_SYMBOL(generic_shutdown_super);

/**
478
 *	sget_userns -	find or create a superblock
Linus Torvalds's avatar
Linus Torvalds committed
479 480 481
 *	@type:	filesystem type superblock should belong to
 *	@test:	comparison callback
 *	@set:	setup callback
482
 *	@flags:	mount flags
483
 *	@user_ns: User namespace for the super_block
Linus Torvalds's avatar
Linus Torvalds committed
484 485
 *	@data:	argument to each of them
 */
486
struct super_block *sget_userns(struct file_system_type *type,
Linus Torvalds's avatar
Linus Torvalds committed
487 488
			int (*test)(struct super_block *,void *),
			int (*set)(struct super_block *,void *),
489
			int flags, struct user_namespace *user_ns,
Linus Torvalds's avatar
Linus Torvalds committed
490 491 492
			void *data)
{
	struct super_block *s = NULL;
493
	struct super_block *old;
Linus Torvalds's avatar
Linus Torvalds committed
494 495
	int err;

496
	if (!(flags & (SB_KERNMOUNT|SB_SUBMOUNT)) &&
497 498 499
	    !(type->fs_flags & FS_USERNS_MOUNT) &&
	    !capable(CAP_SYS_ADMIN))
		return ERR_PTR(-EPERM);
Linus Torvalds's avatar
Linus Torvalds committed
500 501
retry:
	spin_lock(&sb_lock);
502
	if (test) {
503
		hlist_for_each_entry(old, &type->fs_supers, s_instances) {
504 505
			if (!test(old, data))
				continue;
506 507
			if (user_ns != old->s_user_ns) {
				spin_unlock(&sb_lock);
508
				destroy_unused_super(s);
509 510
				return ERR_PTR(-EBUSY);
			}
511 512
			if (!grab_super(old))
				goto retry;
513
			destroy_unused_super(s);
514 515
			return old;
		}
Linus Torvalds's avatar
Linus Torvalds committed
516 517 518
	}
	if (!s) {
		spin_unlock(&sb_lock);
519
		s = alloc_super(type, (flags & ~SB_SUBMOUNT), user_ns);
Linus Torvalds's avatar
Linus Torvalds committed
520 521 522 523
		if (!s)
			return ERR_PTR(-ENOMEM);
		goto retry;
	}
524

Linus Torvalds's avatar
Linus Torvalds committed
525 526 527
	err = set(s, data);
	if (err) {
		spin_unlock(&sb_lock);
528
		destroy_unused_super(s);
Linus Torvalds's avatar
Linus Torvalds committed
529 530 531 532 533
		return ERR_PTR(err);
	}
	s->s_type = type;
	strlcpy(s->s_id, type->name, sizeof(s->s_id));
	list_add_tail(&s->s_list, &super_blocks);
534
	hlist_add_head(&s->s_instances, &type->fs_supers);
Linus Torvalds's avatar
Linus Torvalds committed
535 536
	spin_unlock(&sb_lock);
	get_filesystem(type);
537
	register_shrinker_prepared(&s->s_shrink);
Linus Torvalds's avatar
Linus Torvalds committed
538 539 540
	return s;
}

541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558
EXPORT_SYMBOL(sget_userns);

/**
 *	sget	-	find or create a superblock
 *	@type:	  filesystem type superblock should belong to
 *	@test:	  comparison callback
 *	@set:	  setup callback
 *	@flags:	  mount flags
 *	@data:	  argument to each of them
 */
struct super_block *sget(struct file_system_type *type,
			int (*test)(struct super_block *,void *),
			int (*set)(struct super_block *,void *),
			int flags,
			void *data)
{
	struct user_namespace *user_ns = current_user_ns();

559 560 561 562
	/* We don't yet pass the user namespace of the parent
	 * mount through to here so always use &init_user_ns
	 * until that changes.
	 */
563
	if (flags & SB_SUBMOUNT)
564 565
		user_ns = &init_user_ns;

566
	/* Ensure the requestor has permissions over the target filesystem */
567
	if (!(flags & (SB_KERNMOUNT|SB_SUBMOUNT)) && !ns_capable(user_ns, CAP_SYS_ADMIN))
568 569 570 571 572
		return ERR_PTR(-EPERM);

	return sget_userns(type, test, set, flags, user_ns, data);
}

Linus Torvalds's avatar
Linus Torvalds committed
573 574 575 576 577 578 579 580 581 582
EXPORT_SYMBOL(sget);

void drop_super(struct super_block *sb)
{
	up_read(&sb->s_umount);
	put_super(sb);
}

EXPORT_SYMBOL(drop_super);

583 584 585 586 587 588 589
void drop_super_exclusive(struct super_block *sb)
{
	up_write(&sb->s_umount);
	put_super(sb);
}
EXPORT_SYMBOL(drop_super_exclusive);

590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611
static void __iterate_supers(void (*f)(struct super_block *))
{
	struct super_block *sb, *p = NULL;

	spin_lock(&sb_lock);
	list_for_each_entry(sb, &super_blocks, s_list) {
		if (hlist_unhashed(&sb->s_instances))
			continue;
		sb->s_count++;
		spin_unlock(&sb_lock);

		f(sb);

		spin_lock(&sb_lock);
		if (p)
			__put_super(p);
		p = sb;
	}
	if (p)
		__put_super(p);
	spin_unlock(&sb_lock);
}
Al Viro's avatar
Al Viro committed
612 613 614 615 616 617 618 619 620 621
/**
 *	iterate_supers - call function for all active superblocks
 *	@f: function to call
 *	@arg: argument to pass to it
 *
 *	Scans the superblock list and calls given function, passing it
 *	locked superblock and given argument.
 */
void iterate_supers(void (*f)(struct super_block *, void *), void *arg)
{
622
	struct super_block *sb, *p = NULL;
Al Viro's avatar
Al Viro committed
623 624

	spin_lock(&sb_lock);
625
	list_for_each_entry(sb, &super_blocks, s_list) {
626
		if (hlist_unhashed(&sb->s_instances))
Al Viro's avatar
Al Viro committed
627 628 629 630 631
			continue;
		sb->s_count++;
		spin_unlock(&sb_lock);

		down_read(&sb->s_umount);
632
		if (sb->s_root && (sb->s_flags & SB_BORN))
Al Viro's avatar
Al Viro committed
633 634 635 636
			f(sb, arg);
		up_read(&sb->s_umount);

		spin_lock(&sb_lock);
637 638 639
		if (p)
			__put_super(p);
		p = sb;
Al Viro's avatar
Al Viro committed
640
	}
641 642
	if (p)
		__put_super(p);
Al Viro's avatar
Al Viro committed
643 644 645
	spin_unlock(&sb_lock);
}

646 647 648 649 650 651 652 653 654 655 656 657 658 659 660
/**
 *	iterate_supers_type - call function for superblocks of given type
 *	@type: fs type
 *	@f: function to call
 *	@arg: argument to pass to it
 *
 *	Scans the superblock list and calls given function, passing it
 *	locked superblock and given argument.
 */
void iterate_supers_type(struct file_system_type *type,
	void (*f)(struct super_block *, void *), void *arg)
{
	struct super_block *sb, *p = NULL;

	spin_lock(&sb_lock);
661
	hlist_for_each_entry(sb, &type->fs_supers, s_instances) {
662 663 664 665
		sb->s_count++;
		spin_unlock(&sb_lock);

		down_read(&sb->s_umount);
666
		if (sb->s_root && (sb->s_flags & SB_BORN))
667 668 669 670 671 672 673 674 675 676 677 678 679 680 681
			f(sb, arg);
		up_read(&sb->s_umount);

		spin_lock(&sb_lock);
		if (p)
			__put_super(p);
		p = sb;
	}
	if (p)
		__put_super(p);
	spin_unlock(&sb_lock);
}

EXPORT_SYMBOL(iterate_supers_type);

682
static struct super_block *__get_super(struct block_device *bdev, bool excl)
Linus Torvalds's avatar
Linus Torvalds committed
683
{
684 685
	struct super_block *sb;

Linus Torvalds's avatar
Linus Torvalds committed
686 687
	if (!bdev)
		return NULL;
688

Linus Torvalds's avatar
Linus Torvalds committed
689
	spin_lock(&sb_lock);
690 691
rescan:
	list_for_each_entry(sb, &super_blocks, s_list) {
692
		if (hlist_unhashed(&sb->s_instances))
693
			continue;
694 695
		if (sb->s_bdev == bdev) {
			sb->s_count++;
Linus Torvalds's avatar
Linus Torvalds committed
696
			spin_unlock(&sb_lock);
697 698 699 700
			if (!excl)
				down_read(&sb->s_umount);
			else
				down_write(&sb->s_umount);
701
			/* still alive? */
702
			if (sb->s_root && (sb->s_flags & SB_BORN))
703
				return sb;
704 705 706 707
			if (!excl)
				up_read(&sb->s_umount);
			else
				up_write(&sb->s_umount);
708
			/* nope, got unmounted */
709
			spin_lock(&sb_lock);
710 711
			__put_super(sb);
			goto rescan;
Linus Torvalds's avatar
Linus Torvalds committed
712 713 714 715 716 717
		}
	}
	spin_unlock(&sb_lock);
	return NULL;
}

718
/**
719
 *	get_super - get the superblock of a device
720 721 722
 *	@bdev: device to get the superblock for
 *
 *	Scans the superblock list and finds the superblock of the file system
723
 *	mounted on the device given. %NULL is returned if no match is found.
724
 */
725 726 727 728 729 730 731 732
struct super_block *get_super(struct block_device *bdev)
{
	return __get_super(bdev, false);
}
EXPORT_SYMBOL(get_super);

static struct super_block *__get_super_thawed(struct block_device *bdev,
					      bool excl)
733 734
{
	while (1) {
735
		struct super_block *s = __get_super(bdev, excl);
736
		if (!s || s->s_writers.frozen == SB_UNFROZEN)
737
			return s;
738 739 740 741
		if (!excl)
			up_read(&s->s_umount);
		else
			up_write(&s->s_umount);
742 743
		wait_event(s->s_writers.wait_unfrozen,
			   s->s_writers.frozen == SB_UNFROZEN);
744 745 746
		put_super(s);
	}
}
747 748 749 750 751 752 753 754 755 756 757 758 759 760

/**
 *	get_super_thawed - get thawed superblock of a device
 *	@bdev: device to get the superblock for
 *
 *	Scans the superblock list and finds the superblock of the file system
 *	mounted on the device. The superblock is returned once it is thawed
 *	(or immediately if it was not frozen). %NULL is returned if no match
 *	is found.
 */
struct super_block *get_super_thawed(struct block_device *bdev)
{
	return __get_super_thawed(bdev, false);
}
761 762
EXPORT_SYMBOL(get_super_thawed);

763 764 765 766 767 768 769 770 771 772 773 774 775 776 777
/**
 *	get_super_exclusive_thawed - get thawed superblock of a device
 *	@bdev: device to get the superblock for
 *
 *	Scans the superblock list and finds the superblock of the file system
 *	mounted on the device. The superblock is returned once it is thawed
 *	(or immediately if it was not frozen) and s_umount semaphore is held
 *	in exclusive mode. %NULL is returned if no match is found.
 */
struct super_block *get_super_exclusive_thawed(struct block_device *bdev)
{
	return __get_super_thawed(bdev, true);
}
EXPORT_SYMBOL(get_super_exclusive_thawed);

778 779 780 781 782 783
/**
 * get_active_super - get an active reference to the superblock of a device
 * @bdev: device to get the superblock for
 *
 * Scans the superblock list and finds the superblock of the file system
 * mounted on the device given.  Returns the superblock with an active
784
 * reference or %NULL if none was found.
785 786 787 788 789 790 791 792
 */
struct super_block *get_active_super(struct block_device *bdev)
{
	struct super_block *sb;

	if (!bdev)
		return NULL;

793
restart:
794 795
	spin_lock(&sb_lock);
	list_for_each_entry(sb, &super_blocks, s_list) {
796
		if (hlist_unhashed(&sb->s_instances))
797
			continue;
798
		if (sb->s_bdev == bdev) {
Al Viro's avatar
Al Viro committed
799
			if (!grab_super(sb))
800
				goto restart;
Al Viro's avatar
Al Viro committed
801 802
			up_write(&sb->s_umount);
			return sb;
803
		}
804 805 806 807
	}
	spin_unlock(&sb_lock);
	return NULL;
}
808

809
struct super_block *user_get_super(dev_t dev)
Linus Torvalds's avatar
Linus Torvalds committed
810
{
811
	struct super_block *sb;
Linus Torvalds's avatar
Linus Torvalds committed
812 813

	spin_lock(&sb_lock);
814 815
rescan:
	list_for_each_entry(sb, &super_blocks, s_list) {
816
		if (hlist_unhashed(&sb->s_instances))
817
			continue;
818 819
		if (sb->s_dev ==  dev) {
			sb->s_count++;
Linus Torvalds's avatar
Linus Torvalds committed
820
			spin_unlock(&sb_lock);
821
			down_read(&sb->s_umount);
822
			/* still alive? */
823
			if (sb->s_root && (sb->s_flags & SB_BORN))
824 825
				return sb;
			up_read(&sb->s_umount);
826
			/* nope, got unmounted */
827
			spin_lock(&sb_lock);
828 829
			__put_super(sb);
			goto rescan;
Linus Torvalds's avatar
Linus Torvalds committed
830 831 832 833 834 835 836 837 838
		}
	}
	spin_unlock(&sb_lock);
	return NULL;
}

/**
 *	do_remount_sb - asks filesystem to change mount options.
 *	@sb:	superblock in question
839
 *	@sb_flags: revised superblock flags
Linus Torvalds's avatar
Linus Torvalds committed
840 841 842 843 844
 *	@data:	the rest of options
 *      @force: whether or not to force the change
 *
 *	Alters the mount options of a mounted file system.
 */
845
int do_remount_sb(struct super_block *sb, int sb_flags, void *data, int force)
Linus Torvalds's avatar
Linus Torvalds committed
846 847
{
	int retval;
848
	int remount_ro;
849

850
	if (sb->s_writers.frozen != SB_UNFROZEN)
851 852
		return -EBUSY;

853
#ifdef CONFIG_BLOCK
854
	if (!(sb_flags & SB_RDONLY) && bdev_read_only(sb->s_bdev))
Linus Torvalds's avatar
Linus Torvalds committed
855
		return -EACCES;
856
#endif
857

858
	remount_ro = (sb_flags & SB_RDONLY) && !sb_rdonly(sb);
859

860
	if (remount_ro) {
861
		if (!hlist_empty(&sb->s_pins)) {
862
			up_write(&sb->s_umount);
863
			group_pin_kill(&sb->s_pins);
864 865 866 867 868
			down_write(&sb->s_umount);
			if (!sb->s_root)
				return 0;
			if (sb->s_writers.frozen != SB_UNFROZEN)
				return -EBUSY;
869
			remount_ro = (sb_flags & SB_RDONLY) && !sb_rdonly(sb);
870 871 872 873
		}
	}
	shrink_dcache_sb(sb);

Linus Torvalds's avatar
Linus Torvalds committed
874 875
	/* If we are remounting RDONLY and current sb is read/write,
	   make sure there are no rw files opened */
876
	if (remount_ro) {
877
		if (force) {
878 879
			sb->s_readonly_remount = 1;
			smp_wmb();
880 881 882 883 884
		} else {
			retval = sb_prepare_remount_readonly(sb);
			if (retval)
				return retval;
		}
Linus Torvalds's avatar
Linus Torvalds committed
885 886 887
	}

	if (sb->s_op->remount_fs) {
888
		retval = sb->s_op->remount_fs(sb, &sb_flags, data);
889 890
		if (retval) {
			if (!force)
891
				goto cancel_readonly;
892 893 894 895
			/* If forced remount, go ahead despite any errors */
			WARN(1, "forced remount of a %s fs returned %i\n",
			     sb->s_type->name, retval);
		}
Linus Torvalds's avatar
Linus Torvalds committed
896
	}
897
	sb->s_flags = (sb->s_flags & ~MS_RMT_MASK) | (sb_flags & MS_RMT_MASK);
898 899 900
	/* Needs to be ordered wrt mnt_is_readonly() */
	smp_wmb();
	sb->s_readonly_remount = 0;
901

902 903 904 905 906 907 908 909 910 911
	/*
	 * Some filesystems modify their metadata via some other path than the
	 * bdev buffer cache (eg. use a private mapping, or directories in
	 * pagecache, etc). Also file data modifications go via their own
	 * mappings. So If we try to mount readonly then copy the filesystem
	 * from bdev, we could get stale data, so invalidate it to give a best
	 * effort at coherency.
	 */
	if (remount_ro && sb->s_bdev)
		invalidate_bdev(sb->s_bdev);
Linus Torvalds's avatar
Linus Torvalds committed
912
	return 0;
913 914 915 916

cancel_readonly:
	sb->s_readonly_remount = 0;
	return retval;
Linus Torvalds's avatar
Linus Torvalds committed
917 918
}

919
static void do_emergency_remount_callback(struct super_block *sb)
Linus Torvalds's avatar
Linus Torvalds committed
920
{
921 922 923 924 925 926 927
	down_write(&sb->s_umount);
	if (sb->s_root && sb->s_bdev && (sb->s_flags & SB_BORN) &&
	    !sb_rdonly(sb)) {
		/*
		 * What lock protects sb->s_flags??
		 */
		do_remount_sb(sb, SB_RDONLY, NULL, 1);
Linus Torvalds's avatar
Linus Torvalds committed
928
	}
929 930 931 932 933 934
	up_write(&sb->s_umount);
}

static void do_emergency_remount(struct work_struct *work)
{
	__iterate_supers(do_emergency_remount_callback);
935
	kfree(work);
Linus Torvalds's avatar
Linus Torvalds committed
936 937 938 939 940
	printk("Emergency Remount complete\n");
}

void emergency_remount(void)
{
941 942 943 944 945 946 947
	struct work_struct *work;

	work = kmalloc(sizeof(*work), GFP_ATOMIC);
	if (work) {
		INIT_WORK(work, do_emergency_remount);
		schedule_work(work);
	}
Linus Torvalds's avatar
Linus Torvalds committed
948 949
}

950 951 952
static void do_thaw_all_callback(struct super_block *sb)
{
	down_write(&sb->s_umount);
Al Viro's avatar
Al Viro committed
953
	if (sb->s_root && sb->s_flags & SB_BORN) {
954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983
		emergency_thaw_bdev(sb);
		thaw_super_locked(sb);
	} else {
		up_write(&sb->s_umount);
	}
}

static void do_thaw_all(struct work_struct *work)
{
	__iterate_supers(do_thaw_all_callback);
	kfree(work);
	printk(KERN_WARNING "Emergency Thaw complete\n");
}

/**
 * emergency_thaw_all -- forcibly thaw every frozen filesystem
 *
 * Used for emergency unfreeze of all filesystems via SysRq
 */
void emergency_thaw_all(void)
{
	struct work_struct *work;

	work = kmalloc(sizeof(*work), GFP_ATOMIC);
	if (work) {
		INIT_WORK(work, do_thaw_all);
		schedule_work(work);
	}
}

984
static DEFINE_IDA(unnamed_dev_ida);
Linus Torvalds's avatar
Linus Torvalds committed
985

986 987 988 989 990 991 992 993 994 995 996
/**
 * get_anon_bdev - Allocate a block device for filesystems which don't have one.
 * @p: Pointer to a dev_t.
 *
 * Filesystems which don't use real block devices can call this function
 * to allocate a virtual block device.
 *
 * Context: Any context.  Frequently called while holding sb_lock.
 * Return: 0 on success, -EMFILE if there are no anonymous bdevs left
 * or -ENOMEM if memory allocation failed.
 */
997
int get_anon_bdev(dev_t *p)
Linus Torvalds's avatar
Linus Torvalds committed
998 999
{
	int dev;
1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012

	/*
	 * Many userspace utilities consider an FSID of 0 invalid.
	 * Always return at least 1 from get_anon_bdev.
	 */
	dev = ida_alloc_range(&unnamed_dev_ida, 1, (1 << MINORBITS) - 1,
			GFP_ATOMIC);
	if (dev == -ENOSPC)
		dev = -EMFILE;
	if (dev < 0)
		return dev;

	*p = MKDEV(0, dev);
Linus Torvalds's avatar
Linus Torvalds committed
1013 1014
	return 0;
}
1015
EXPORT_SYMBOL(get_anon_bdev);
Linus Torvalds's avatar
Linus Torvalds committed
1016

1017
void free_anon_bdev(dev_t dev)
Linus Torvalds's avatar
Linus Torvalds committed
1018
{
1019
	ida_free(&unnamed_dev_ida, MINOR(dev));
Linus Torvalds's avatar
Linus Torvalds committed
1020
}
1021 1022 1023 1024
EXPORT_SYMBOL(free_anon_bdev);

int set_anon_super(struct super_block *s, void *data)
{
1025
	return get_anon_bdev(&s->s_dev);
1026 1027 1028 1029 1030 1031 1032 1033 1034
}
EXPORT_SYMBOL(set_anon_super);

void kill_anon_super(struct super_block *sb)
{
	dev_t dev = sb->s_dev;
	generic_shutdown_super(sb);
	free_anon_bdev(dev);
}
Linus Torvalds's avatar
Linus Torvalds committed
1035 1036 1037 1038 1039 1040 1041 1042 1043 1044
EXPORT_SYMBOL(kill_anon_super);

void kill_litter_super(struct super_block *sb)
{
	if (sb->s_root)
		d_genocide(sb->s_root);
	kill_anon_super(sb);
}
EXPORT_SYMBOL(kill_litter_super);

1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055
static int ns_test_super(struct super_block *sb, void *data)
{
	return sb->s_fs_info == data;
}

static int ns_set_super(struct super_block *sb, void *data)
{
	sb->s_fs_info = data;
	return set_anon_super(sb, NULL);
}

1056 1057 1058
struct dentry *mount_ns(struct file_system_type *fs_type,
	int flags, void *data, void *ns, struct user_namespace *user_ns,
	int (*fill_super)(struct super_block *, void *, int))
1059 1060 1061
{
	struct super_block *sb;

1062 1063 1064
	/* Don't allow mounting unless the caller has CAP_SYS_ADMIN
	 * over the namespace.
	 */
1065
	if (!(flags & SB_KERNMOUNT) && !ns_capable(user_ns, CAP_SYS_ADMIN))
1066 1067
		return ERR_PTR(-EPERM);

1068 1069
	sb = sget_userns(fs_type, ns_test_super, ns_set_super, flags,
			 user_ns, ns);
1070
	if (IS_ERR(sb))
Al Viro's avatar
Al Viro committed
1071
		return ERR_CAST(sb);
1072 1073 1074

	if (!sb->s_root) {
		int err;
1075
		err = fill_super(sb, data, flags & SB_SILENT ? 1 : 0);
1076
		if (err) {
1077
			deactivate_locked_super(sb);
Al Viro's avatar
Al Viro committed
1078
			return ERR_PTR(err);
1079 1080
		}

1081
		sb->s_flags |= SB_ACTIVE;
1082 1083
	}

Al Viro's avatar
Al Viro committed
1084
	return dget(sb->s_root);
1085 1086
}

Al Viro's avatar
Al Viro committed
1087
EXPORT_SYMBOL(mount_ns);
1088

1089
#ifdef CONFIG_BLOCK
Linus Torvalds's avatar
Linus Torvalds committed
1090 1091 1092 1093
static int set_bdev_super(struct super_block *s, void *data)
{
	s->s_bdev = data;
	s->s_dev = s->s_bdev->bd_dev;
1094
	s->s_bdi = bdi_get(s->s_bdev->bd_bdi);
1095

Linus Torvalds's avatar
Linus Torvalds committed
1096 1097 1098 1099 1100 1101 1102 1103
	return 0;
}

static int test_bdev_super(struct super_block *s, void *data)
{
	return (void *)s->s_bdev == data;
}

Al Viro's avatar
Al Viro committed
1104
struct dentry *mount_bdev(struct file_system_type *fs_type,
Linus Torvalds's avatar
Linus Torvalds committed
1105
	int flags, const char *dev_name, void *data,
Al Viro's avatar
Al Viro committed
1106
	int (*fill_super)(struct super_block *, void *, int))
Linus Torvalds's avatar
Linus Torvalds committed
1107 1108 1109
{
	struct block_device *bdev;
	struct super_block *s;
1110
	fmode_t mode = FMODE_READ | FMODE_EXCL;
Linus Torvalds's avatar
Linus Torvalds committed
1111 1112
	int error = 0;

1113
	if (!(flags & SB_RDONLY))
1114 1115
		mode |= FMODE_WRITE;

1116
	bdev = blkdev_get_by_path(dev_name, mode, fs_type);
Linus Torvalds's avatar
Linus Torvalds committed
1117
	if (IS_ERR(bdev))
Al Viro's avatar
Al Viro committed
1118
		return ERR_CAST(bdev);
Linus Torvalds's avatar
Linus Torvalds committed
1119 1120 1121 1122 1123 1124

	/*
	 * once the super is inserted into the list by sget, s_umount
	 * will protect the lockfs code from trying to start a snapshot
	 * while we are mounting
	 */
1125 1126 1127 1128 1129 1130
	mutex_lock(&bdev->bd_fsfreeze_mutex);
	if (bdev->bd_fsfreeze_count > 0) {
		mutex_unlock(&bdev->bd_fsfreeze_mutex);
		error = -EBUSY;
		goto error_bdev;
	}
1131
	s = sget(fs_type, test_bdev_super, set_bdev_super, flags | SB_NOSEC,
1132
		 bdev);
1133
	mutex_unlock(&bdev->bd_fsfreeze_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
1134
	if (IS_ERR(s))
1135
		goto error_s;
Linus Torvalds's avatar
Linus Torvalds committed
1136 1137

	if (s->s_root) {
1138
		if ((flags ^ s->s_flags) & SB_RDONLY) {
1139
			deactivate_locked_super(s);
1140 1141
			error = -EBUSY;
			goto error_bdev;
Linus Torvalds's avatar
Linus Torvalds committed
1142
		}
1143

1144 1145
		/*
		 * s_umount nests inside bd_mutex during
1146 1147 1148 1149
		 * __invalidate_device().  blkdev_put() acquires
		 * bd_mutex and can't be called under s_umount.  Drop
		 * s_umount temporarily.  This is safe as we're
		 * holding an active reference.
1150 1151
		 */
		up_write(&s->s_umount);
1152
		blkdev_put(bdev, mode);
1153
		down_write(&s->s_umount);
Linus Torvalds's avatar
Linus Torvalds committed
1154
	} else {
1155
		s->s_mode = mode;
1156
		snprintf(s->s_id, sizeof(s->s_id), "%pg", bdev);
1157
		sb_set_blocksize(s, block_size(bdev));
1158
		error = fill_super(s, data, flags & SB_SILENT ? 1 : 0);
Linus Torvalds's avatar
Linus Torvalds committed
1159
		if (error) {
1160
			deactivate_locked_super(s);
1161
			goto error;
1162
		}
1163

1164
		s->s_flags |= SB_ACTIVE;
1165
		bdev->bd_super = s;
Linus Torvalds's avatar
Linus Torvalds committed
1166 1167
	}

Al Viro's avatar
Al Viro committed
1168
	return dget(s->s_root);
Linus Torvalds's avatar
Linus Torvalds committed
1169

1170 1171 1172
error_s:
	error = PTR_ERR(s);
error_bdev:
1173
	blkdev_put(bdev, mode);
1174
error:
Al Viro's avatar
Al Viro committed
1175 1176 1177 1178
	return ERR_PTR(error);
}
EXPORT_SYMBOL(mount_bdev);

Linus Torvalds's avatar
Linus Torvalds committed
1179 1180 1181
void kill_block_super(struct super_block *sb)
{
	struct block_device *bdev = sb->s_bdev;
1182
	fmode_t mode = sb->s_mode;
Linus Torvalds's avatar
Linus Torvalds committed
1183

1184
	bdev->bd_super = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
1185 1186
	generic_shutdown_super(sb);
	sync_blockdev(bdev);
1187
	WARN_ON_ONCE(!(mode & FMODE_EXCL));
1188
	blkdev_put(bdev, mode | FMODE_EXCL);
Linus Torvalds's avatar
Linus Torvalds committed
1189 1190 1191
}

EXPORT_SYMBOL(kill_block_super);
1192
#endif
Linus Torvalds's avatar
Linus Torvalds committed
1193

Al Viro's avatar
Al Viro committed
1194
struct dentry *mount_nodev(struct file_system_type *fs_type,
Linus Torvalds's avatar
Linus Torvalds committed
1195
	int flags, void *data,
Al Viro's avatar
Al Viro committed
1196
	int (*fill_super)(struct super_block *, void *, int))
Linus Torvalds's avatar
Linus Torvalds committed
1197 1198
{
	int error;
1199
	struct super_block *s = sget(fs_type, NULL, set_anon_super, flags, NULL);
Linus Torvalds's avatar
Linus Torvalds committed
1200 1201

	if (IS_ERR(s))
Al Viro's avatar
Al Viro committed
1202
		return ERR_CAST(s);
Linus Torvalds's avatar
Linus Torvalds committed
1203

1204
	error = fill_super(s, data, flags & SB_SILENT ? 1 : 0);
Linus Torvalds's avatar
Linus Torvalds committed
1205
	if (error) {
1206
		deactivate_locked_super(s);
Al Viro's avatar
Al Viro committed
1207
		return ERR_PTR(error);
Linus Torvalds's avatar
Linus Torvalds committed
1208
	}
1209
	s->s_flags |= SB_ACTIVE;
Al Viro's avatar
Al Viro committed
1210
	return dget(s->s_root);
Linus Torvalds's avatar
Linus Torvalds committed
1211
}
Al Viro's avatar
Al Viro committed
1212 1213
EXPORT_SYMBOL(mount_nodev);

Linus Torvalds's avatar
Linus Torvalds committed
1214 1215 1216 1217 1218
static int compare_single(struct super_block *s, void *p)
{
	return 1;
}

Al Viro's avatar
Al Viro committed
1219
struct dentry *mount_single(struct file_system_type *fs_type,
Linus Torvalds's avatar
Linus Torvalds committed
1220
	int flags, void *data,
Al Viro's avatar
Al Viro committed
1221
	int (*fill_super)(struct super_block *, void *, int))
Linus Torvalds's avatar
Linus Torvalds committed
1222 1223 1224 1225
{
	struct super_block *s;
	int error;

1226
	s = sget(fs_type, compare_single, set_anon_super, flags, NULL);
Linus Torvalds's avatar
Linus Torvalds committed
1227
	if (IS_ERR(s))
Al Viro's avatar
Al Viro committed
1228
		return ERR_CAST(s);
Linus Torvalds's avatar
Linus Torvalds committed
1229
	if (!s->s_root) {
1230
		error = fill_super(s, data, flags & SB_SILENT ? 1 : 0);
Linus Torvalds's avatar
Linus Torvalds committed
1231
		if (error) {
1232
			deactivate_locked_super(s);
Al Viro's avatar
Al Viro committed
1233
			return ERR_PTR(error);
Linus Torvalds's avatar
Linus Torvalds committed
1234
		}
1235
		s->s_flags |= SB_ACTIVE;
1236 1237
	} else {
		do_remount_sb(s, flags, data, 0);
Linus Torvalds's avatar
Linus Torvalds committed
1238
	}
Al Viro's avatar
Al Viro committed
1239 1240 1241 1242
	return dget(s->s_root);
}
EXPORT_SYMBOL(mount_single);

1243 1244
struct dentry *
mount_fs(struct file_system_type *type, int flags, const char *name, void *data)
Linus Torvalds's avatar
Linus Torvalds committed
1245
{
1246
	struct dentry *root;
1247
	struct super_block *sb;
Linus Torvalds's avatar
Linus Torvalds committed
1248
	char *secdata = NULL;
1249
	int error = -ENOMEM;
1250

1251
	if (data && !(type->fs_flags & FS_BINARY_MOUNTDATA)) {
Linus Torvalds's avatar
Linus Torvalds committed
1252
		secdata = alloc_secdata();
1253
		if (!secdata)
1254
			goto out;
Linus Torvalds's avatar
Linus Torvalds committed
1255

1256