pnode.c 15.3 KB
Newer Older
1 2 3 4 5 6 7 8
/*
 *  linux/fs/pnode.c
 *
 * (C) Copyright IBM Corporation 2005.
 *	Released under GPL v2.
 *	Author : Ram Pai (linuxram@us.ibm.com)
 *
 */
9
#include <linux/mnt_namespace.h>
10 11
#include <linux/mount.h>
#include <linux/fs.h>
12
#include <linux/nsproxy.h>
13
#include "internal.h"
14 15
#include "pnode.h"

16
/* return the next shared peer mount of @p */
17
static inline struct mount *next_peer(struct mount *p)
18
{
19
	return list_entry(p->mnt_share.next, struct mount, mnt_share);
20 21
}

22
static inline struct mount *first_slave(struct mount *p)
23
{
24
	return list_entry(p->mnt_slave_list.next, struct mount, mnt_slave);
25 26
}

27 28 29 30 31
static inline struct mount *last_slave(struct mount *p)
{
	return list_entry(p->mnt_slave_list.prev, struct mount, mnt_slave);
}

32
static inline struct mount *next_slave(struct mount *p)
33
{
34
	return list_entry(p->mnt_slave.next, struct mount, mnt_slave);
35 36
}

37 38 39
static struct mount *get_peer_under_root(struct mount *mnt,
					 struct mnt_namespace *ns,
					 const struct path *root)
40
{
41
	struct mount *m = mnt;
42 43 44

	do {
		/* Check the namespace first for optimization */
45
		if (m->mnt_ns == ns && is_path_reachable(m, m->mnt.mnt_root, root))
46
			return m;
47

48
		m = next_peer(m);
49
	} while (m != mnt);
50 51 52 53 54 55 56 57 58 59

	return NULL;
}

/*
 * Get ID of closest dominating peer group having a representative
 * under the given root.
 *
 * Caller must hold namespace_sem
 */
60
int get_dominating_id(struct mount *mnt, const struct path *root)
61
{
62
	struct mount *m;
63

64
	for (m = mnt->mnt_master; m != NULL; m = m->mnt_master) {
65
		struct mount *d = get_peer_under_root(m, mnt->mnt_ns, root);
66
		if (d)
Al Viro's avatar
Al Viro committed
67
			return d->mnt_group_id;
68 69 70 71 72
	}

	return 0;
}

73
static int do_make_slave(struct mount *mnt)
Ram Pai's avatar
Ram Pai committed
74
{
Al Viro's avatar
Al Viro committed
75
	struct mount *master, *slave_mnt;
Ram Pai's avatar
Ram Pai committed
76

Al Viro's avatar
Al Viro committed
77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92
	if (list_empty(&mnt->mnt_share)) {
		if (IS_MNT_SHARED(mnt)) {
			mnt_release_group_id(mnt);
			CLEAR_MNT_SHARED(mnt);
		}
		master = mnt->mnt_master;
		if (!master) {
			struct list_head *p = &mnt->mnt_slave_list;
			while (!list_empty(p)) {
				slave_mnt = list_first_entry(p,
						struct mount, mnt_slave);
				list_del_init(&slave_mnt->mnt_slave);
				slave_mnt->mnt_master = NULL;
			}
			return 0;
		}
Ram Pai's avatar
Ram Pai committed
93
	} else {
Al Viro's avatar
Al Viro committed
94 95 96 97 98 99 100 101 102 103 104
		struct mount *m;
		/*
		 * slave 'mnt' to a peer mount that has the
		 * same root dentry. If none is available then
		 * slave it to anything that is available.
		 */
		for (m = master = next_peer(mnt); m != mnt; m = next_peer(m)) {
			if (m->mnt.mnt_root == mnt->mnt.mnt_root) {
				master = m;
				break;
			}
Ram Pai's avatar
Ram Pai committed
105
		}
Al Viro's avatar
Al Viro committed
106 107 108
		list_del_init(&mnt->mnt_share);
		mnt->mnt_group_id = 0;
		CLEAR_MNT_SHARED(mnt);
Ram Pai's avatar
Ram Pai committed
109
	}
Al Viro's avatar
Al Viro committed
110 111 112 113 114
	list_for_each_entry(slave_mnt, &mnt->mnt_slave_list, mnt_slave)
		slave_mnt->mnt_master = master;
	list_move(&mnt->mnt_slave, &master->mnt_slave_list);
	list_splice(&mnt->mnt_slave_list, master->mnt_slave_list.prev);
	INIT_LIST_HEAD(&mnt->mnt_slave_list);
115
	mnt->mnt_master = master;
Ram Pai's avatar
Ram Pai committed
116 117 118
	return 0;
}

Nick Piggin's avatar
Nick Piggin committed
119 120 121
/*
 * vfsmount lock must be held for write
 */
122
void change_mnt_propagation(struct mount *mnt, int type)
123
{
124
	if (type == MS_SHARED) {
125
		set_mnt_shared(mnt);
Ram Pai's avatar
Ram Pai committed
126 127
		return;
	}
128
	do_make_slave(mnt);
Ram Pai's avatar
Ram Pai committed
129
	if (type != MS_SLAVE) {
130
		list_del_init(&mnt->mnt_slave);
131
		mnt->mnt_master = NULL;
Ram Pai's avatar
Ram Pai committed
132
		if (type == MS_UNBINDABLE)
133
			mnt->mnt.mnt_flags |= MNT_UNBINDABLE;
Andries E. Brouwer's avatar
Andries E. Brouwer committed
134
		else
135
			mnt->mnt.mnt_flags &= ~MNT_UNBINDABLE;
136
	}
137
}
138 139 140 141 142

/*
 * get the next mount in the propagation tree.
 * @m: the mount seen last
 * @origin: the original mount from where the tree walk initiated
143 144 145 146 147
 *
 * Note that peer groups form contiguous segments of slave lists.
 * We rely on that in get_source() to be able to find out if
 * vfsmount found while iterating with propagation_next() is
 * a peer of one we'd found earlier.
148
 */
149 150
static struct mount *propagation_next(struct mount *m,
					 struct mount *origin)
151
{
152
	/* are there any slaves of this mount? */
153
	if (!IS_MNT_NEW(m) && !list_empty(&m->mnt_slave_list))
154 155 156
		return first_slave(m);

	while (1) {
157
		struct mount *master = m->mnt_master;
158

159
		if (master == origin->mnt_master) {
160 161
			struct mount *next = next_peer(m);
			return (next == origin) ? NULL : next;
162
		} else if (m->mnt_slave.next != &master->mnt_slave_list)
163 164 165 166 167 168 169
			return next_slave(m);

		/* back at master */
		m = master;
	}
}

170 171 172 173 174 175 176 177 178 179 180 181 182
static struct mount *skip_propagation_subtree(struct mount *m,
						struct mount *origin)
{
	/*
	 * Advance m such that propagation_next will not return
	 * the slaves of m.
	 */
	if (!IS_MNT_NEW(m) && !list_empty(&m->mnt_slave_list))
		m = last_slave(m);

	return m;
}

Al Viro's avatar
Al Viro committed
183
static struct mount *next_group(struct mount *m, struct mount *origin)
184
{
Al Viro's avatar
Al Viro committed
185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211
	while (1) {
		while (1) {
			struct mount *next;
			if (!IS_MNT_NEW(m) && !list_empty(&m->mnt_slave_list))
				return first_slave(m);
			next = next_peer(m);
			if (m->mnt_group_id == origin->mnt_group_id) {
				if (next == origin)
					return NULL;
			} else if (m->mnt_slave.next != &next->mnt_slave)
				break;
			m = next;
		}
		/* m is the last peer */
		while (1) {
			struct mount *master = m->mnt_master;
			if (m->mnt_slave.next != &master->mnt_slave_list)
				return next_slave(m);
			m = next_peer(master);
			if (master->mnt_group_id == origin->mnt_group_id)
				break;
			if (master->mnt_slave.next == &m->mnt_slave)
				break;
			m = master;
		}
		if (m == origin)
			return NULL;
212
	}
Al Viro's avatar
Al Viro committed
213
}
214

Al Viro's avatar
Al Viro committed
215 216
/* all accesses are serialized by namespace_sem */
static struct user_namespace *user_ns;
217
static struct mount *last_dest, *first_source, *last_source, *dest_master;
Al Viro's avatar
Al Viro committed
218 219 220
static struct mountpoint *mp;
static struct hlist_head *list;

221 222 223 224 225
static inline bool peers(struct mount *m1, struct mount *m2)
{
	return m1->mnt_group_id == m2->mnt_group_id && m1->mnt_group_id;
}

Al Viro's avatar
Al Viro committed
226 227 228 229 230 231 232 233 234 235
static int propagate_one(struct mount *m)
{
	struct mount *child;
	int type;
	/* skip ones added by this propagate_mnt() */
	if (IS_MNT_NEW(m))
		return 0;
	/* skip if mountpoint isn't covered by it */
	if (!is_subdir(mp->m_dentry, m->mnt.mnt_root))
		return 0;
236
	if (peers(m, last_dest)) {
Al Viro's avatar
Al Viro committed
237 238 239
		type = CL_MAKE_SHARED;
	} else {
		struct mount *n, *p;
240
		bool done;
Al Viro's avatar
Al Viro committed
241 242
		for (n = m; ; n = p) {
			p = n->mnt_master;
243
			if (p == dest_master || IS_MNT_MARKED(p))
Al Viro's avatar
Al Viro committed
244
				break;
245
		}
246 247 248 249 250 251 252 253 254 255
		do {
			struct mount *parent = last_source->mnt_parent;
			if (last_source == first_source)
				break;
			done = parent->mnt_master == p;
			if (done && peers(n, parent))
				break;
			last_source = last_source->mnt_master;
		} while (!done);

Al Viro's avatar
Al Viro committed
256 257 258 259
		type = CL_SLAVE;
		/* beginning of peer group among the slaves? */
		if (IS_MNT_SHARED(m))
			type |= CL_MAKE_SHARED;
260
	}
Al Viro's avatar
Al Viro committed
261 262 263 264 265 266 267
		
	/* Notice when we are propagating across user namespaces */
	if (m->mnt_ns->user_ns != user_ns)
		type |= CL_UNPRIVILEGED;
	child = copy_tree(last_source, last_source->mnt.mnt_root, type);
	if (IS_ERR(child))
		return PTR_ERR(child);
268
	child->mnt.mnt_flags &= ~MNT_LOCKED;
Al Viro's avatar
Al Viro committed
269 270 271 272 273 274 275 276 277
	mnt_set_mountpoint(m, mp, child);
	last_dest = m;
	last_source = child;
	if (m->mnt_master != dest_master) {
		read_seqlock_excl(&mount_lock);
		SET_MNT_MARK(m->mnt_master);
		read_sequnlock_excl(&mount_lock);
	}
	hlist_add_head(&child->mnt_hash, list);
278
	return count_mounts(m->mnt_ns, child);
279 280 281 282 283 284 285 286 287 288 289 290 291 292 293
}

/*
 * mount 'source_mnt' under the destination 'dest_mnt' at
 * dentry 'dest_dentry'. And propagate that mount to
 * all the peer and slave mounts of 'dest_mnt'.
 * Link all the new mounts into a propagation tree headed at
 * source_mnt. Also link all the new mounts using ->mnt_list
 * headed at source_mnt's ->mnt_list
 *
 * @dest_mnt: destination mount.
 * @dest_dentry: destination dentry.
 * @source_mnt: source mount.
 * @tree_list : list of heads of trees to be attached.
 */
294
int propagate_mnt(struct mount *dest_mnt, struct mountpoint *dest_mp,
Al Viro's avatar
Al Viro committed
295
		    struct mount *source_mnt, struct hlist_head *tree_list)
296
{
Al Viro's avatar
Al Viro committed
297
	struct mount *m, *n;
298
	int ret = 0;
299

Al Viro's avatar
Al Viro committed
300 301 302 303 304 305 306
	/*
	 * we don't want to bother passing tons of arguments to
	 * propagate_one(); everything is serialized by namespace_sem,
	 * so globals will do just fine.
	 */
	user_ns = current->nsproxy->mnt_ns->user_ns;
	last_dest = dest_mnt;
307
	first_source = source_mnt;
Al Viro's avatar
Al Viro committed
308 309 310 311 312 313 314 315 316
	last_source = source_mnt;
	mp = dest_mp;
	list = tree_list;
	dest_master = dest_mnt->mnt_master;

	/* all peers of dest_mnt, except dest_mnt itself */
	for (n = next_peer(dest_mnt); n != dest_mnt; n = next_peer(n)) {
		ret = propagate_one(n);
		if (ret)
317
			goto out;
Al Viro's avatar
Al Viro committed
318
	}
319

Al Viro's avatar
Al Viro committed
320 321 322 323 324 325 326 327 328 329 330
	/* all slave groups */
	for (m = next_group(dest_mnt, dest_mnt); m;
			m = next_group(m, dest_mnt)) {
		/* everything in that slave group */
		n = m;
		do {
			ret = propagate_one(n);
			if (ret)
				goto out;
			n = next_peer(n);
		} while (n != m);
331 332
	}
out:
Al Viro's avatar
Al Viro committed
333 334 335 336 337
	read_seqlock_excl(&mount_lock);
	hlist_for_each_entry(n, tree_list, mnt_hash) {
		m = n->mnt_parent;
		if (m->mnt_master != dest_mnt->mnt_master)
			CLEAR_MNT_MARK(m->mnt_master);
338
	}
Al Viro's avatar
Al Viro committed
339
	read_sequnlock_excl(&mount_lock);
340 341
	return ret;
}
342

343 344 345 346 347 348 349 350 351 352 353 354 355 356 357
static struct mount *find_topper(struct mount *mnt)
{
	/* If there is exactly one mount covering mnt completely return it. */
	struct mount *child;

	if (!list_is_singular(&mnt->mnt_mounts))
		return NULL;

	child = list_first_entry(&mnt->mnt_mounts, struct mount, mnt_child);
	if (child->mnt_mountpoint != mnt->mnt.mnt_root)
		return NULL;

	return child;
}

358 359 360
/*
 * return true if the refcount is greater than count
 */
361
static inline int do_refcount_check(struct mount *mnt, int count)
362
{
363
	return mnt_get_count(mnt) > count;
364 365 366 367 368 369 370 371 372
}

/*
 * check if the mount 'mnt' can be unmounted successfully.
 * @mnt: the mount to be checked for unmount
 * NOTE: unmounting 'mnt' would naturally propagate to all
 * other mounts its parent propagates to.
 * Check if any of these mounts that **do not have submounts**
 * have more references than 'refcnt'. If so return busy.
Nick Piggin's avatar
Nick Piggin committed
373
 *
Nick Piggin's avatar
Nick Piggin committed
374
 * vfsmount lock must be held for write
375
 */
376
int propagate_mount_busy(struct mount *mnt, int refcnt)
377
{
378
	struct mount *m, *child, *topper;
379
	struct mount *parent = mnt->mnt_parent;
380

381
	if (mnt == parent)
382 383 384 385 386 387 388
		return do_refcount_check(mnt, refcnt);

	/*
	 * quickly check if the current mount can be unmounted.
	 * If not, we don't have to go checking for all other
	 * mounts
	 */
389
	if (!list_empty(&mnt->mnt_mounts) || do_refcount_check(mnt, refcnt))
390 391
		return 1;

392 393
	for (m = propagation_next(parent, parent); m;
	     		m = propagation_next(m, parent)) {
394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409
		int count = 1;
		child = __lookup_mnt(&m->mnt, mnt->mnt_mountpoint);
		if (!child)
			continue;

		/* Is there exactly one mount on the child that covers
		 * it completely whose reference should be ignored?
		 */
		topper = find_topper(child);
		if (topper)
			count += 1;
		else if (!list_empty(&child->mnt_mounts))
			continue;

		if (do_refcount_check(child, count))
			return 1;
410
	}
411
	return 0;
412 413
}

414 415 416 417 418 419 420 421 422 423 424 425 426 427
/*
 * Clear MNT_LOCKED when it can be shown to be safe.
 *
 * mount_lock lock must be held for write
 */
void propagate_mount_unlock(struct mount *mnt)
{
	struct mount *parent = mnt->mnt_parent;
	struct mount *m, *child;

	BUG_ON(parent == mnt);

	for (m = propagation_next(parent, parent); m;
			m = propagation_next(m, parent)) {
428
		child = __lookup_mnt(&m->mnt, mnt->mnt_mountpoint);
429 430 431 432 433
		if (child)
			child->mnt.mnt_flags &= ~MNT_LOCKED;
	}
}

434
static void umount_one(struct mount *mnt, struct list_head *to_umount)
435
{
436 437 438 439 440
	CLEAR_MNT_MARK(mnt);
	mnt->mnt.mnt_flags |= MNT_UMOUNT;
	list_del_init(&mnt->mnt_child);
	list_del_init(&mnt->mnt_umounting);
	list_move_tail(&mnt->mnt_list, to_umount);
441 442
}

443 444 445 446
/*
 * NOTE: unmounting 'mnt' naturally propagates to all other mounts its
 * parent propagates to.
 */
447 448 449
static bool __propagate_umount(struct mount *mnt,
			       struct list_head *to_umount,
			       struct list_head *to_restore)
450
{
451 452
	bool progress = false;
	struct mount *child;
453

454 455 456 457 458 459
	/*
	 * The state of the parent won't change if this mount is
	 * already unmounted or marked as without children.
	 */
	if (mnt->mnt.mnt_flags & (MNT_UMOUNT | MNT_MARKED))
		goto out;
460

461 462 463 464 465
	/* Verify topper is the only grandchild that has not been
	 * speculatively unmounted.
	 */
	list_for_each_entry(child, &mnt->mnt_mounts, mnt_child) {
		if (child->mnt_mountpoint == mnt->mnt.mnt_root)
466
			continue;
467 468 469 470 471
		if (!list_empty(&child->mnt_umounting) && IS_MNT_MARKED(child))
			continue;
		/* Found a mounted child */
		goto children;
	}
472

473 474 475
	/* Mark mounts that can be unmounted if not locked */
	SET_MNT_MARK(mnt);
	progress = true;
476

477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498
	/* If a mount is without children and not locked umount it. */
	if (!IS_MNT_LOCKED(mnt)) {
		umount_one(mnt, to_umount);
	} else {
children:
		list_move_tail(&mnt->mnt_umounting, to_restore);
	}
out:
	return progress;
}

static void umount_list(struct list_head *to_umount,
			struct list_head *to_restore)
{
	struct mount *mnt, *child, *tmp;
	list_for_each_entry(mnt, to_umount, mnt_list) {
		list_for_each_entry_safe(child, tmp, &mnt->mnt_mounts, mnt_child) {
			/* topper? */
			if (child->mnt_mountpoint == mnt->mnt.mnt_root)
				list_move_tail(&child->mnt_umounting, to_restore);
			else
				umount_one(child, to_umount);
Al Viro's avatar
Al Viro committed
499
		}
500 501 502
	}
}

503
static void restore_mounts(struct list_head *to_restore)
504
{
505 506
	/* Restore mounts to a clean working state */
	while (!list_empty(to_restore)) {
507 508 509
		struct mount *mnt, *parent;
		struct mountpoint *mp;

510 511 512
		mnt = list_first_entry(to_restore, struct mount, mnt_umounting);
		CLEAR_MNT_MARK(mnt);
		list_del_init(&mnt->mnt_umounting);
513

514
		/* Should this mount be reparented? */
515 516 517 518 519 520
		mp = mnt->mnt_mp;
		parent = mnt->mnt_parent;
		while (parent->mnt.mnt_flags & MNT_UMOUNT) {
			mp = parent->mnt_mp;
			parent = parent->mnt_parent;
		}
521 522
		if (parent != mnt->mnt_parent)
			mnt_change_mountpoint(parent, mp, mnt);
523 524 525
	}
}

526 527 528 529 530 531 532 533 534
static void cleanup_umount_visitations(struct list_head *visited)
{
	while (!list_empty(visited)) {
		struct mount *mnt =
			list_first_entry(visited, struct mount, mnt_umounting);
		list_del_init(&mnt->mnt_umounting);
	}
}

535 536 537 538
/*
 * collect all mounts that receive propagation from the mount in @list,
 * and return these additional mounts in the same list.
 * @list: the list of mounts to be unmounted.
Nick Piggin's avatar
Nick Piggin committed
539 540
 *
 * vfsmount lock must be held for write
541
 */
542
int propagate_umount(struct list_head *list)
543
{
544
	struct mount *mnt;
545 546
	LIST_HEAD(to_restore);
	LIST_HEAD(to_umount);
547
	LIST_HEAD(visited);
548

549 550
	/* Find candidates for unmounting */
	list_for_each_entry_reverse(mnt, list, mnt_list) {
551 552
		struct mount *parent = mnt->mnt_parent;
		struct mount *m;
553

554 555 556 557 558 559 560 561 562 563
		/*
		 * If this mount has already been visited it is known that it's
		 * entire peer group and all of their slaves in the propagation
		 * tree for the mountpoint has already been visited and there is
		 * no need to visit them again.
		 */
		if (!list_empty(&mnt->mnt_umounting))
			continue;

		list_add_tail(&mnt->mnt_umounting, &visited);
564 565 566 567 568 569 570
		for (m = propagation_next(parent, parent); m;
		     m = propagation_next(m, parent)) {
			struct mount *child = __lookup_mnt(&m->mnt,
							   mnt->mnt_mountpoint);
			if (!child)
				continue;

571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591
			if (!list_empty(&child->mnt_umounting)) {
				/*
				 * If the child has already been visited it is
				 * know that it's entire peer group and all of
				 * their slaves in the propgation tree for the
				 * mountpoint has already been visited and there
				 * is no need to visit this subtree again.
				 */
				m = skip_propagation_subtree(m, parent);
				continue;
			} else if (child->mnt.mnt_flags & MNT_UMOUNT) {
				/*
				 * We have come accross an partially unmounted
				 * mount in list that has not been visited yet.
				 * Remember it has been visited and continue
				 * about our merry way.
				 */
				list_add_tail(&child->mnt_umounting, &visited);
				continue;
			}

592 593 594 595 596 597 598 599 600 601
			/* Check the child and parents while progress is made */
			while (__propagate_umount(child,
						  &to_umount, &to_restore)) {
				/* Is the parent a umount candidate? */
				child = child->mnt_parent;
				if (list_empty(&child->mnt_umounting))
					break;
			}
		}
	}
602

603 604
	umount_list(&to_umount, &to_restore);
	restore_mounts(&to_restore);
605
	cleanup_umount_visitations(&visited);
606
	list_splice_tail(&to_umount, list);
607

608 609
	return 0;
}