super.c 26.2 KB
Newer Older
M
Miklos Szeredi 已提交
1 2 3 4 5 6 7 8 9 10 11
/*
 *
 * Copyright (C) 2011 Novell Inc.
 *
 * This program is free software; you can redistribute it and/or modify it
 * under the terms of the GNU General Public License version 2 as published by
 * the Free Software Foundation.
 */

#include <linux/fs.h>
#include <linux/namei.h>
12
#include <linux/pagemap.h>
M
Miklos Szeredi 已提交
13 14 15 16 17 18
#include <linux/xattr.h>
#include <linux/security.h>
#include <linux/mount.h>
#include <linux/slab.h>
#include <linux/parser.h>
#include <linux/module.h>
19
#include <linux/pagemap.h>
M
Miklos Szeredi 已提交
20
#include <linux/sched.h>
21
#include <linux/statfs.h>
22
#include <linux/seq_file.h>
M
Miklos Szeredi 已提交
23 24 25 26 27 28
#include "overlayfs.h"

MODULE_AUTHOR("Miklos Szeredi <miklos@szeredi.hu>");
MODULE_DESCRIPTION("Overlay filesystem");
MODULE_LICENSE("GPL");

29 30 31 32
struct ovl_config {
	char *lowerdir;
	char *upperdir;
	char *workdir;
33
	bool default_permissions;
34 35
};

M
Miklos Szeredi 已提交
36 37 38
/* private information held for overlayfs's superblock */
struct ovl_fs {
	struct vfsmount *upper_mnt;
39 40
	unsigned numlower;
	struct vfsmount **lower_mnt;
M
Miklos Szeredi 已提交
41
	struct dentry *workdir;
42
	long lower_namelen;
43 44
	/* pathnames of lower and upper dirs, for show_options */
	struct ovl_config config;
45 46
	/* creds of process who forced instantiation of super block */
	const struct cred *creator_cred;
M
Miklos Szeredi 已提交
47 48 49 50 51 52 53 54 55 56 57 58 59 60 61
};

struct ovl_dir_cache;

/* private information held for every overlayfs dentry */
struct ovl_entry {
	struct dentry *__upperdentry;
	struct ovl_dir_cache *cache;
	union {
		struct {
			u64 version;
			bool opaque;
		};
		struct rcu_head rcu;
	};
62 63
	unsigned numlower;
	struct path lowerstack[];
M
Miklos Szeredi 已提交
64 65
};

66 67
#define OVL_MAX_STACK 500

68 69 70 71
static struct dentry *__ovl_dentry_lower(struct ovl_entry *oe)
{
	return oe->numlower ? oe->lowerstack[0].dentry : NULL;
}
M
Miklos Szeredi 已提交
72 73 74 75

enum ovl_path_type ovl_path_type(struct dentry *dentry)
{
	struct ovl_entry *oe = dentry->d_fsdata;
76
	enum ovl_path_type type = 0;
M
Miklos Szeredi 已提交
77 78

	if (oe->__upperdentry) {
79 80
		type = __OVL_PATH_UPPER;

81 82 83 84 85 86 87
		/*
		 * Non-dir dentry can hold lower dentry from previous
		 * location. Its purity depends only on opaque flag.
		 */
		if (oe->numlower && S_ISDIR(dentry->d_inode->i_mode))
			type |= __OVL_PATH_MERGE;
		else if (!oe->opaque)
88
			type |= __OVL_PATH_PURE;
M
Miklos Szeredi 已提交
89 90 91
	} else {
		if (oe->numlower > 1)
			type |= __OVL_PATH_MERGE;
M
Miklos Szeredi 已提交
92
	}
93
	return type;
M
Miklos Szeredi 已提交
94 95 96 97
}

static struct dentry *ovl_upperdentry_dereference(struct ovl_entry *oe)
{
98
	return lockless_dereference(oe->__upperdentry);
M
Miklos Szeredi 已提交
99 100 101 102 103 104 105 106 107 108 109 110 111 112 113
}

void ovl_path_upper(struct dentry *dentry, struct path *path)
{
	struct ovl_fs *ofs = dentry->d_sb->s_fs_info;
	struct ovl_entry *oe = dentry->d_fsdata;

	path->mnt = ofs->upper_mnt;
	path->dentry = ovl_upperdentry_dereference(oe);
}

enum ovl_path_type ovl_path_real(struct dentry *dentry, struct path *path)
{
	enum ovl_path_type type = ovl_path_type(dentry);

114
	if (!OVL_TYPE_UPPER(type))
M
Miklos Szeredi 已提交
115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
		ovl_path_lower(dentry, path);
	else
		ovl_path_upper(dentry, path);

	return type;
}

struct dentry *ovl_dentry_upper(struct dentry *dentry)
{
	struct ovl_entry *oe = dentry->d_fsdata;

	return ovl_upperdentry_dereference(oe);
}

struct dentry *ovl_dentry_lower(struct dentry *dentry)
{
	struct ovl_entry *oe = dentry->d_fsdata;

133
	return __ovl_dentry_lower(oe);
M
Miklos Szeredi 已提交
134 135 136 137 138 139 140 141 142
}

struct dentry *ovl_dentry_real(struct dentry *dentry)
{
	struct ovl_entry *oe = dentry->d_fsdata;
	struct dentry *realdentry;

	realdentry = ovl_upperdentry_dereference(oe);
	if (!realdentry)
143
		realdentry = __ovl_dentry_lower(oe);
M
Miklos Szeredi 已提交
144 145 146 147 148 149 150 151 152 153 154 155

	return realdentry;
}

struct dentry *ovl_entry_real(struct ovl_entry *oe, bool *is_upper)
{
	struct dentry *realdentry;

	realdentry = ovl_upperdentry_dereference(oe);
	if (realdentry) {
		*is_upper = true;
	} else {
156
		realdentry = __ovl_dentry_lower(oe);
M
Miklos Szeredi 已提交
157 158 159 160 161
		*is_upper = false;
	}
	return realdentry;
}

162 163 164 165 166 167 168 169 170 171 172 173
struct vfsmount *ovl_entry_mnt_real(struct ovl_entry *oe, struct inode *inode,
				    bool is_upper)
{
	if (is_upper) {
		struct ovl_fs *ofs = inode->i_sb->s_fs_info;

		return ofs->upper_mnt;
	} else {
		return oe->numlower ? oe->lowerstack[0].mnt : NULL;
	}
}

M
Miklos Szeredi 已提交
174 175 176 177 178 179 180
struct ovl_dir_cache *ovl_dir_cache(struct dentry *dentry)
{
	struct ovl_entry *oe = dentry->d_fsdata;

	return oe->cache;
}

181 182 183 184 185 186 187
bool ovl_is_default_permissions(struct inode *inode)
{
	struct ovl_fs *ofs = inode->i_sb->s_fs_info;

	return ofs->config.default_permissions;
}

M
Miklos Szeredi 已提交
188 189 190 191 192 193 194 195 196 197 198
void ovl_set_dir_cache(struct dentry *dentry, struct ovl_dir_cache *cache)
{
	struct ovl_entry *oe = dentry->d_fsdata;

	oe->cache = cache;
}

void ovl_path_lower(struct dentry *dentry, struct path *path)
{
	struct ovl_entry *oe = dentry->d_fsdata;

199
	*path = oe->numlower ? oe->lowerstack[0] : (struct path) { NULL, NULL };
M
Miklos Szeredi 已提交
200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235
}

int ovl_want_write(struct dentry *dentry)
{
	struct ovl_fs *ofs = dentry->d_sb->s_fs_info;
	return mnt_want_write(ofs->upper_mnt);
}

void ovl_drop_write(struct dentry *dentry)
{
	struct ovl_fs *ofs = dentry->d_sb->s_fs_info;
	mnt_drop_write(ofs->upper_mnt);
}

struct dentry *ovl_workdir(struct dentry *dentry)
{
	struct ovl_fs *ofs = dentry->d_sb->s_fs_info;
	return ofs->workdir;
}

bool ovl_dentry_is_opaque(struct dentry *dentry)
{
	struct ovl_entry *oe = dentry->d_fsdata;
	return oe->opaque;
}

void ovl_dentry_set_opaque(struct dentry *dentry, bool opaque)
{
	struct ovl_entry *oe = dentry->d_fsdata;
	oe->opaque = opaque;
}

void ovl_dentry_update(struct dentry *dentry, struct dentry *upperdentry)
{
	struct ovl_entry *oe = dentry->d_fsdata;

A
Al Viro 已提交
236
	WARN_ON(!inode_is_locked(upperdentry->d_parent->d_inode));
M
Miklos Szeredi 已提交
237 238 239 240 241 242 243 244 245 246 247 248 249 250
	WARN_ON(oe->__upperdentry);
	BUG_ON(!upperdentry->d_inode);
	/*
	 * Make sure upperdentry is consistent before making it visible to
	 * ovl_upperdentry_dereference().
	 */
	smp_wmb();
	oe->__upperdentry = upperdentry;
}

void ovl_dentry_version_inc(struct dentry *dentry)
{
	struct ovl_entry *oe = dentry->d_fsdata;

A
Al Viro 已提交
251
	WARN_ON(!inode_is_locked(dentry->d_inode));
M
Miklos Szeredi 已提交
252 253 254 255 256 257 258
	oe->version++;
}

u64 ovl_dentry_version_get(struct dentry *dentry)
{
	struct ovl_entry *oe = dentry->d_fsdata;

A
Al Viro 已提交
259
	WARN_ON(!inode_is_locked(dentry->d_inode));
M
Miklos Szeredi 已提交
260 261 262 263 264 265 266 267 268 269
	return oe->version;
}

bool ovl_is_whiteout(struct dentry *dentry)
{
	struct inode *inode = dentry->d_inode;

	return inode && IS_WHITEOUT(inode);
}

270 271 272 273 274 275 276
const struct cred *ovl_override_creds(struct super_block *sb)
{
	struct ovl_fs *ofs = sb->s_fs_info;

	return override_creds(ofs->creator_cred);
}

M
Miklos Szeredi 已提交
277 278 279 280 281 282 283 284 285
static bool ovl_is_opaquedir(struct dentry *dentry)
{
	int res;
	char val;
	struct inode *inode = dentry->d_inode;

	if (!S_ISDIR(inode->i_mode) || !inode->i_op->getxattr)
		return false;

286
	res = inode->i_op->getxattr(dentry, inode, OVL_XATTR_OPAQUE, &val, 1);
M
Miklos Szeredi 已提交
287 288 289 290 291 292 293 294 295 296 297
	if (res == 1 && val == 'y')
		return true;

	return false;
}

static void ovl_dentry_release(struct dentry *dentry)
{
	struct ovl_entry *oe = dentry->d_fsdata;

	if (oe) {
298 299
		unsigned int i;

M
Miklos Szeredi 已提交
300
		dput(oe->__upperdentry);
301 302
		for (i = 0; i < oe->numlower; i++)
			dput(oe->lowerstack[i].dentry);
M
Miklos Szeredi 已提交
303 304 305 306
		kfree_rcu(oe, rcu);
	}
}

307 308 309
static struct dentry *ovl_d_real(struct dentry *dentry,
				 const struct inode *inode,
				 unsigned int open_flags)
M
Miklos Szeredi 已提交
310 311 312 313 314 315 316 317 318
{
	struct dentry *real;

	if (d_is_dir(dentry)) {
		if (!inode || inode == d_inode(dentry))
			return dentry;
		goto bug;
	}

319 320 321 322 323 324 325 326 327 328
	if (d_is_negative(dentry))
		return dentry;

	if (open_flags) {
		int err = ovl_open_maybe_copy_up(dentry, open_flags);

		if (err)
			return ERR_PTR(err);
	}

M
Miklos Szeredi 已提交
329 330 331 332 333 334 335 336 337 338 339 340
	real = ovl_dentry_upper(dentry);
	if (real && (!inode || inode == d_inode(real)))
		return real;

	real = ovl_dentry_lower(dentry);
	if (!real)
		goto bug;

	if (!inode || inode == d_inode(real))
		return real;

	/* Handle recursion */
341
	return d_real(real, inode, open_flags);
M
Miklos Szeredi 已提交
342 343 344 345 346 347
bug:
	WARN(1, "ovl_d_real(%pd4, %s:%lu\n): real dentry not found\n", dentry,
	     inode ? inode->i_sb->s_id : "NULL", inode ? inode->i_ino : 0);
	return dentry;
}

348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388
static int ovl_dentry_revalidate(struct dentry *dentry, unsigned int flags)
{
	struct ovl_entry *oe = dentry->d_fsdata;
	unsigned int i;
	int ret = 1;

	for (i = 0; i < oe->numlower; i++) {
		struct dentry *d = oe->lowerstack[i].dentry;

		if (d->d_flags & DCACHE_OP_REVALIDATE) {
			ret = d->d_op->d_revalidate(d, flags);
			if (ret < 0)
				return ret;
			if (!ret) {
				if (!(flags & LOOKUP_RCU))
					d_invalidate(d);
				return -ESTALE;
			}
		}
	}
	return 1;
}

static int ovl_dentry_weak_revalidate(struct dentry *dentry, unsigned int flags)
{
	struct ovl_entry *oe = dentry->d_fsdata;
	unsigned int i;
	int ret = 1;

	for (i = 0; i < oe->numlower; i++) {
		struct dentry *d = oe->lowerstack[i].dentry;

		if (d->d_flags & DCACHE_OP_WEAK_REVALIDATE) {
			ret = d->d_op->d_weak_revalidate(d, flags);
			if (ret <= 0)
				break;
		}
	}
	return ret;
}

M
Miklos Szeredi 已提交
389 390
static const struct dentry_operations ovl_dentry_operations = {
	.d_release = ovl_dentry_release,
M
Miklos Szeredi 已提交
391
	.d_real = ovl_d_real,
M
Miklos Szeredi 已提交
392 393
};

394 395
static const struct dentry_operations ovl_reval_dentry_operations = {
	.d_release = ovl_dentry_release,
M
Miklos Szeredi 已提交
396
	.d_real = ovl_d_real,
397 398 399 400
	.d_revalidate = ovl_dentry_revalidate,
	.d_weak_revalidate = ovl_dentry_weak_revalidate,
};

401
static struct ovl_entry *ovl_alloc_entry(unsigned int numlower)
M
Miklos Szeredi 已提交
402
{
403 404 405 406 407 408 409
	size_t size = offsetof(struct ovl_entry, lowerstack[numlower]);
	struct ovl_entry *oe = kzalloc(size, GFP_KERNEL);

	if (oe)
		oe->numlower = numlower;

	return oe;
M
Miklos Szeredi 已提交
410 411
}

412 413 414 415 416 417 418 419 420 421 422 423 424 425
static bool ovl_dentry_remote(struct dentry *dentry)
{
	return dentry->d_flags &
		(DCACHE_OP_REVALIDATE | DCACHE_OP_WEAK_REVALIDATE);
}

static bool ovl_dentry_weird(struct dentry *dentry)
{
	return dentry->d_flags & (DCACHE_NEED_AUTOMOUNT |
				  DCACHE_MANAGE_TRANSIT |
				  DCACHE_OP_HASH |
				  DCACHE_OP_COMPARE);
}

M
Miklos Szeredi 已提交
426 427 428 429 430
static inline struct dentry *ovl_lookup_real(struct dentry *dir,
					     struct qstr *name)
{
	struct dentry *dentry;

431
	dentry = lookup_hash(name, dir);
M
Miklos Szeredi 已提交
432 433 434 435 436 437 438

	if (IS_ERR(dentry)) {
		if (PTR_ERR(dentry) == -ENOENT)
			dentry = NULL;
	} else if (!dentry->d_inode) {
		dput(dentry);
		dentry = NULL;
439
	} else if (ovl_dentry_weird(dentry)) {
440
		dput(dentry);
441
		/* Don't support traversing automounts and other weirdness */
442
		dentry = ERR_PTR(-EREMOTE);
M
Miklos Szeredi 已提交
443 444 445 446
	}
	return dentry;
}

447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467
/*
 * Returns next layer in stack starting from top.
 * Returns -1 if this is the last layer.
 */
int ovl_path_next(int idx, struct dentry *dentry, struct path *path)
{
	struct ovl_entry *oe = dentry->d_fsdata;

	BUG_ON(idx < 0);
	if (idx == 0) {
		ovl_path_upper(dentry, path);
		if (path->dentry)
			return oe->numlower ? 1 : -1;
		idx++;
	}
	BUG_ON(idx > oe->numlower);
	*path = oe->lowerstack[idx - 1];

	return (idx < oe->numlower) ? idx + 1 : -1;
}

M
Miklos Szeredi 已提交
468 469 470 471
struct dentry *ovl_lookup(struct inode *dir, struct dentry *dentry,
			  unsigned int flags)
{
	struct ovl_entry *oe;
M
Miklos Szeredi 已提交
472 473 474 475
	struct ovl_entry *poe = dentry->d_parent->d_fsdata;
	struct path *stack = NULL;
	struct dentry *upperdir, *upperdentry = NULL;
	unsigned int ctr = 0;
M
Miklos Szeredi 已提交
476
	struct inode *inode = NULL;
M
Miklos Szeredi 已提交
477 478 479
	bool upperopaque = false;
	struct dentry *this, *prev = NULL;
	unsigned int i;
M
Miklos Szeredi 已提交
480 481
	int err;

M
Miklos Szeredi 已提交
482
	upperdir = ovl_upperdentry_dereference(poe);
M
Miklos Szeredi 已提交
483
	if (upperdir) {
M
Miklos Szeredi 已提交
484 485 486 487 488
		this = ovl_lookup_real(upperdir, &dentry->d_name);
		err = PTR_ERR(this);
		if (IS_ERR(this))
			goto out;

489
		if (this) {
490 491 492 493 494
			if (unlikely(ovl_dentry_remote(this))) {
				dput(this);
				err = -EREMOTE;
				goto out;
			}
M
Miklos Szeredi 已提交
495 496 497 498
			if (ovl_is_whiteout(this)) {
				dput(this);
				this = NULL;
				upperopaque = true;
499
			} else if (poe->numlower && ovl_is_opaquedir(this)) {
M
Miklos Szeredi 已提交
500
				upperopaque = true;
M
Miklos Szeredi 已提交
501 502
			}
		}
M
Miklos Szeredi 已提交
503
		upperdentry = prev = this;
M
Miklos Szeredi 已提交
504
	}
M
Miklos Szeredi 已提交
505 506 507 508 509 510

	if (!upperopaque && poe->numlower) {
		err = -ENOMEM;
		stack = kcalloc(poe->numlower, sizeof(struct path), GFP_KERNEL);
		if (!stack)
			goto out_put_upper;
M
Miklos Szeredi 已提交
511 512
	}

M
Miklos Szeredi 已提交
513 514 515 516 517 518
	for (i = 0; !upperopaque && i < poe->numlower; i++) {
		bool opaque = false;
		struct path lowerpath = poe->lowerstack[i];

		this = ovl_lookup_real(lowerpath.dentry, &dentry->d_name);
		err = PTR_ERR(this);
519 520 521 522 523 524
		if (IS_ERR(this)) {
			/*
			 * If it's positive, then treat ENAMETOOLONG as ENOENT.
			 */
			if (err == -ENAMETOOLONG && (upperdentry || ctr))
				continue;
M
Miklos Szeredi 已提交
525
			goto out_put;
526
		}
M
Miklos Szeredi 已提交
527 528
		if (!this)
			continue;
529 530 531 532
		if (ovl_is_whiteout(this)) {
			dput(this);
			break;
		}
M
Miklos Szeredi 已提交
533
		/*
534 535
		 * Only makes sense to check opaque dir if this is not the
		 * lowermost layer.
M
Miklos Szeredi 已提交
536
		 */
537 538
		if (i < poe->numlower - 1 && ovl_is_opaquedir(this))
			opaque = true;
539 540 541 542 543 544 545

		if (prev && (!S_ISDIR(prev->d_inode->i_mode) ||
			     !S_ISDIR(this->d_inode->i_mode))) {
			/*
			 * FIXME: check for upper-opaqueness maybe better done
			 * in remove code.
			 */
M
Miklos Szeredi 已提交
546 547 548 549 550
			if (prev == upperdentry)
				upperopaque = true;
			dput(this);
			break;
		}
551 552 553 554 555 556
		/*
		 * If this is a non-directory then stop here.
		 */
		if (!S_ISDIR(this->d_inode->i_mode))
			opaque = true;

M
Miklos Szeredi 已提交
557 558 559 560 561 562
		stack[ctr].dentry = this;
		stack[ctr].mnt = lowerpath.mnt;
		ctr++;
		prev = this;
		if (opaque)
			break;
M
Miklos Szeredi 已提交
563 564
	}

M
Miklos Szeredi 已提交
565 566 567 568 569 570
	oe = ovl_alloc_entry(ctr);
	err = -ENOMEM;
	if (!oe)
		goto out_put;

	if (upperdentry || ctr) {
M
Miklos Szeredi 已提交
571 572
		struct dentry *realdentry;

M
Miklos Szeredi 已提交
573 574
		realdentry = upperdentry ? upperdentry : stack[0].dentry;

M
Miklos Szeredi 已提交
575 576 577 578
		err = -ENOMEM;
		inode = ovl_new_inode(dentry->d_sb, realdentry->d_inode->i_mode,
				      oe);
		if (!inode)
M
Miklos Szeredi 已提交
579
			goto out_free_oe;
M
Miklos Szeredi 已提交
580 581 582
		ovl_copyattr(realdentry->d_inode, inode);
	}

M
Miklos Szeredi 已提交
583
	oe->opaque = upperopaque;
M
Miklos Szeredi 已提交
584
	oe->__upperdentry = upperdentry;
M
Miklos Szeredi 已提交
585 586
	memcpy(oe->lowerstack, stack, sizeof(struct path) * ctr);
	kfree(stack);
M
Miklos Szeredi 已提交
587 588 589 590 591
	dentry->d_fsdata = oe;
	d_add(dentry, inode);

	return NULL;

M
Miklos Szeredi 已提交
592
out_free_oe:
M
Miklos Szeredi 已提交
593
	kfree(oe);
M
Miklos Szeredi 已提交
594 595 596 597 598 599
out_put:
	for (i = 0; i < ctr; i++)
		dput(stack[i].dentry);
	kfree(stack);
out_put_upper:
	dput(upperdentry);
M
Miklos Szeredi 已提交
600 601 602 603 604 605 606 607 608 609 610 611
out:
	return ERR_PTR(err);
}

struct file *ovl_path_open(struct path *path, int flags)
{
	return dentry_open(path, flags, current_cred());
}

static void ovl_put_super(struct super_block *sb)
{
	struct ovl_fs *ufs = sb->s_fs_info;
612
	unsigned i;
M
Miklos Szeredi 已提交
613 614 615

	dput(ufs->workdir);
	mntput(ufs->upper_mnt);
616 617
	for (i = 0; i < ufs->numlower; i++)
		mntput(ufs->lower_mnt[i]);
618
	kfree(ufs->lower_mnt);
M
Miklos Szeredi 已提交
619

620 621 622
	kfree(ufs->config.lowerdir);
	kfree(ufs->config.upperdir);
	kfree(ufs->config.workdir);
623
	put_cred(ufs->creator_cred);
M
Miklos Szeredi 已提交
624 625 626
	kfree(ufs);
}

627 628 629 630 631 632
/**
 * ovl_statfs
 * @sb: The overlayfs super block
 * @buf: The struct kstatfs to fill in with stats
 *
 * Get the filesystem statistics.  As writes always target the upper layer
633
 * filesystem pass the statfs to the upper filesystem (if it exists)
634 635 636 637 638 639 640 641
 */
static int ovl_statfs(struct dentry *dentry, struct kstatfs *buf)
{
	struct ovl_fs *ofs = dentry->d_sb->s_fs_info;
	struct dentry *root_dentry = dentry->d_sb->s_root;
	struct path path;
	int err;

642
	ovl_path_real(root_dentry, &path);
643 644 645 646 647 648 649 650 651 652

	err = vfs_statfs(&path, buf);
	if (!err) {
		buf->f_namelen = max(buf->f_namelen, ofs->lower_namelen);
		buf->f_type = OVERLAYFS_SUPER_MAGIC;
	}

	return err;
}

653 654 655 656 657 658 659 660 661 662 663
/**
 * ovl_show_options
 *
 * Prints the mount options for a given superblock.
 * Returns zero; does not fail.
 */
static int ovl_show_options(struct seq_file *m, struct dentry *dentry)
{
	struct super_block *sb = dentry->d_sb;
	struct ovl_fs *ufs = sb->s_fs_info;

664
	seq_show_option(m, "lowerdir", ufs->config.lowerdir);
665
	if (ufs->config.upperdir) {
666 667
		seq_show_option(m, "upperdir", ufs->config.upperdir);
		seq_show_option(m, "workdir", ufs->config.workdir);
668
	}
669 670
	if (ufs->config.default_permissions)
		seq_puts(m, ",default_permissions");
671 672 673
	return 0;
}

674 675 676 677
static int ovl_remount(struct super_block *sb, int *flags, char *data)
{
	struct ovl_fs *ufs = sb->s_fs_info;

678
	if (!(*flags & MS_RDONLY) && (!ufs->upper_mnt || !ufs->workdir))
679 680 681 682 683
		return -EROFS;

	return 0;
}

M
Miklos Szeredi 已提交
684 685
static const struct super_operations ovl_super_operations = {
	.put_super	= ovl_put_super,
686
	.statfs		= ovl_statfs,
687
	.show_options	= ovl_show_options,
688
	.remount_fs	= ovl_remount,
M
Miklos Szeredi 已提交
689 690 691 692 693 694
};

enum {
	OPT_LOWERDIR,
	OPT_UPPERDIR,
	OPT_WORKDIR,
695
	OPT_DEFAULT_PERMISSIONS,
M
Miklos Szeredi 已提交
696 697 698 699 700 701 702
	OPT_ERR,
};

static const match_table_t ovl_tokens = {
	{OPT_LOWERDIR,			"lowerdir=%s"},
	{OPT_UPPERDIR,			"upperdir=%s"},
	{OPT_WORKDIR,			"workdir=%s"},
703
	{OPT_DEFAULT_PERMISSIONS,	"default_permissions"},
M
Miklos Szeredi 已提交
704 705 706
	{OPT_ERR,			NULL}
};

707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729
static char *ovl_next_opt(char **s)
{
	char *sbegin = *s;
	char *p;

	if (sbegin == NULL)
		return NULL;

	for (p = sbegin; *p; p++) {
		if (*p == '\\') {
			p++;
			if (!*p)
				break;
		} else if (*p == ',') {
			*p = '\0';
			*s = p + 1;
			return sbegin;
		}
	}
	*s = NULL;
	return sbegin;
}

M
Miklos Szeredi 已提交
730 731 732 733
static int ovl_parse_opt(char *opt, struct ovl_config *config)
{
	char *p;

734
	while ((p = ovl_next_opt(&opt)) != NULL) {
M
Miklos Szeredi 已提交
735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763
		int token;
		substring_t args[MAX_OPT_ARGS];

		if (!*p)
			continue;

		token = match_token(p, ovl_tokens, args);
		switch (token) {
		case OPT_UPPERDIR:
			kfree(config->upperdir);
			config->upperdir = match_strdup(&args[0]);
			if (!config->upperdir)
				return -ENOMEM;
			break;

		case OPT_LOWERDIR:
			kfree(config->lowerdir);
			config->lowerdir = match_strdup(&args[0]);
			if (!config->lowerdir)
				return -ENOMEM;
			break;

		case OPT_WORKDIR:
			kfree(config->workdir);
			config->workdir = match_strdup(&args[0]);
			if (!config->workdir)
				return -ENOMEM;
			break;

764 765 766 767
		case OPT_DEFAULT_PERMISSIONS:
			config->default_permissions = true;
			break;

M
Miklos Szeredi 已提交
768
		default:
769
			pr_err("overlayfs: unrecognized mount option \"%s\" or missing value\n", p);
M
Miklos Szeredi 已提交
770 771 772
			return -EINVAL;
		}
	}
773 774 775 776 777 778 779 780 781

	/* Workdir is useless in non-upper mount */
	if (!config->upperdir && config->workdir) {
		pr_info("overlayfs: option \"workdir=%s\" is useless in a non-upper mount, ignore\n",
			config->workdir);
		kfree(config->workdir);
		config->workdir = NULL;
	}

M
Miklos Szeredi 已提交
782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798
	return 0;
}

#define OVL_WORKDIR_NAME "work"

static struct dentry *ovl_workdir_create(struct vfsmount *mnt,
					 struct dentry *dentry)
{
	struct inode *dir = dentry->d_inode;
	struct dentry *work;
	int err;
	bool retried = false;

	err = mnt_want_write(mnt);
	if (err)
		return ERR_PTR(err);

A
Al Viro 已提交
799
	inode_lock_nested(dir, I_MUTEX_PARENT);
M
Miklos Szeredi 已提交
800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824
retry:
	work = lookup_one_len(OVL_WORKDIR_NAME, dentry,
			      strlen(OVL_WORKDIR_NAME));

	if (!IS_ERR(work)) {
		struct kstat stat = {
			.mode = S_IFDIR | 0,
		};

		if (work->d_inode) {
			err = -EEXIST;
			if (retried)
				goto out_dput;

			retried = true;
			ovl_cleanup(dir, work);
			dput(work);
			goto retry;
		}

		err = ovl_create_real(dir, work, &stat, NULL, NULL, true);
		if (err)
			goto out_dput;
	}
out_unlock:
A
Al Viro 已提交
825
	inode_unlock(dir);
M
Miklos Szeredi 已提交
826 827 828 829 830 831 832 833 834 835
	mnt_drop_write(mnt);

	return work;

out_dput:
	dput(work);
	work = ERR_PTR(err);
	goto out_unlock;
}

836 837 838 839 840 841 842 843 844 845 846 847 848
static void ovl_unescape(char *s)
{
	char *d = s;

	for (;; s++, d++) {
		if (*s == '\\')
			s++;
		*d = *s;
		if (!*s)
			break;
	}
}

849 850
static int ovl_mount_dir_noesc(const char *name, struct path *path)
{
851
	int err = -EINVAL;
852

853 854 855 856
	if (!*name) {
		pr_err("overlayfs: empty lowerdir\n");
		goto out;
	}
857 858 859 860 861 862
	err = kern_path(name, LOOKUP_FOLLOW, path);
	if (err) {
		pr_err("overlayfs: failed to resolve '%s': %i\n", name, err);
		goto out;
	}
	err = -EINVAL;
863
	if (ovl_dentry_weird(path->dentry)) {
864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886
		pr_err("overlayfs: filesystem on '%s' not supported\n", name);
		goto out_put;
	}
	if (!S_ISDIR(path->dentry->d_inode->i_mode)) {
		pr_err("overlayfs: '%s' not a directory\n", name);
		goto out_put;
	}
	return 0;

out_put:
	path_put(path);
out:
	return err;
}

static int ovl_mount_dir(const char *name, struct path *path)
{
	int err = -ENOMEM;
	char *tmp = kstrdup(name, GFP_KERNEL);

	if (tmp) {
		ovl_unescape(tmp);
		err = ovl_mount_dir_noesc(tmp, path);
887 888 889 890 891 892 893 894

		if (!err)
			if (ovl_dentry_remote(path->dentry)) {
				pr_err("overlayfs: filesystem on '%s' not supported as upperdir\n",
				       tmp);
				path_put(path);
				err = -EINVAL;
			}
895 896 897 898 899 900
		kfree(tmp);
	}
	return err;
}

static int ovl_lower_dir(const char *name, struct path *path, long *namelen,
901
			 int *stack_depth, bool *remote)
902 903 904 905
{
	int err;
	struct kstatfs statfs;

906
	err = ovl_mount_dir_noesc(name, path);
907 908 909 910 911 912 913 914 915 916 917
	if (err)
		goto out;

	err = vfs_statfs(path, &statfs);
	if (err) {
		pr_err("overlayfs: statfs failed on '%s'\n", name);
		goto out_put;
	}
	*namelen = max(*namelen, statfs.f_namelen);
	*stack_depth = max(*stack_depth, path->mnt->mnt_sb->s_stack_depth);

918 919 920
	if (ovl_dentry_remote(path->dentry))
		*remote = true;

921 922 923 924 925 926 927 928
	return 0;

out_put:
	path_put(path);
out:
	return err;
}

M
Miklos Szeredi 已提交
929 930 931 932 933 934 935 936 937 938 939 940
/* Workdir should not be subdir of upperdir and vice versa */
static bool ovl_workdir_ok(struct dentry *workdir, struct dentry *upperdir)
{
	bool ok = false;

	if (workdir != upperdir) {
		ok = (lock_rename(workdir, upperdir) == NULL);
		unlock_rename(workdir, upperdir);
	}
	return ok;
}

941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960
static unsigned int ovl_split_lowerdirs(char *str)
{
	unsigned int ctr = 1;
	char *s, *d;

	for (s = d = str;; s++, d++) {
		if (*s == '\\') {
			s++;
		} else if (*s == ':') {
			*d = '\0';
			ctr++;
			continue;
		}
		*d = *s;
		if (!*s)
			break;
	}
	return ctr;
}

M
Miklos Szeredi 已提交
961 962
static int ovl_fill_super(struct super_block *sb, void *data, int silent)
{
963 964
	struct path upperpath = { NULL, NULL };
	struct path workpath = { NULL, NULL };
M
Miklos Szeredi 已提交
965 966 967
	struct dentry *root_dentry;
	struct ovl_entry *oe;
	struct ovl_fs *ufs;
968 969 970 971 972
	struct path *stack = NULL;
	char *lowertmp;
	char *lower;
	unsigned int numlower;
	unsigned int stacklen = 0;
973
	unsigned int i;
974
	bool remote = false;
M
Miklos Szeredi 已提交
975 976
	int err;

977 978 979
	err = -ENOMEM;
	ufs = kzalloc(sizeof(struct ovl_fs), GFP_KERNEL);
	if (!ufs)
M
Miklos Szeredi 已提交
980 981
		goto out;

982 983 984 985
	err = ovl_parse_opt((char *) data, &ufs->config);
	if (err)
		goto out_free_config;

M
Miklos Szeredi 已提交
986
	err = -EINVAL;
987
	if (!ufs->config.lowerdir) {
988 989
		if (!silent)
			pr_err("overlayfs: missing 'lowerdir'\n");
M
Miklos Szeredi 已提交
990 991 992
		goto out_free_config;
	}

993
	sb->s_stack_depth = 0;
994
	sb->s_maxbytes = MAX_LFS_FILESIZE;
995 996 997 998 999
	if (ufs->config.upperdir) {
		if (!ufs->config.workdir) {
			pr_err("overlayfs: missing 'workdir'\n");
			goto out_free_config;
		}
M
Miklos Szeredi 已提交
1000

1001 1002 1003
		err = ovl_mount_dir(ufs->config.upperdir, &upperpath);
		if (err)
			goto out_free_config;
M
Miklos Szeredi 已提交
1004

1005 1006 1007 1008 1009 1010 1011
		/* Upper fs should not be r/o */
		if (upperpath.mnt->mnt_sb->s_flags & MS_RDONLY) {
			pr_err("overlayfs: upper fs is r/o, try multi-lower layers mount\n");
			err = -EINVAL;
			goto out_put_upperpath;
		}

1012 1013 1014 1015
		err = ovl_mount_dir(ufs->config.workdir, &workpath);
		if (err)
			goto out_put_upperpath;

1016
		err = -EINVAL;
1017 1018 1019 1020 1021 1022 1023 1024 1025
		if (upperpath.mnt != workpath.mnt) {
			pr_err("overlayfs: workdir and upperdir must reside under the same mount\n");
			goto out_put_workpath;
		}
		if (!ovl_workdir_ok(workpath.dentry, upperpath.dentry)) {
			pr_err("overlayfs: workdir and upperdir must be separate subtrees\n");
			goto out_put_workpath;
		}
		sb->s_stack_depth = upperpath.mnt->mnt_sb->s_stack_depth;
1026
	}
1027 1028 1029
	err = -ENOMEM;
	lowertmp = kstrdup(ufs->config.lowerdir, GFP_KERNEL);
	if (!lowertmp)
1030
		goto out_put_workpath;
1031

1032 1033
	err = -EINVAL;
	stacklen = ovl_split_lowerdirs(lowertmp);
1034 1035 1036
	if (stacklen > OVL_MAX_STACK) {
		pr_err("overlayfs: too many lower directries, limit is %d\n",
		       OVL_MAX_STACK);
1037
		goto out_free_lowertmp;
1038 1039 1040 1041
	} else if (!ufs->config.upperdir && stacklen == 1) {
		pr_err("overlayfs: at least 2 lowerdir are needed while upperdir nonexistent\n");
		goto out_free_lowertmp;
	}
1042 1043 1044 1045 1046 1047 1048 1049

	stack = kcalloc(stacklen, sizeof(struct path), GFP_KERNEL);
	if (!stack)
		goto out_free_lowertmp;

	lower = lowertmp;
	for (numlower = 0; numlower < stacklen; numlower++) {
		err = ovl_lower_dir(lower, &stack[numlower],
1050 1051
				    &ufs->lower_namelen, &sb->s_stack_depth,
				    &remote);
1052 1053 1054 1055 1056 1057
		if (err)
			goto out_put_lowerpath;

		lower = strchr(lower, '\0') + 1;
	}

1058
	err = -EINVAL;
1059
	sb->s_stack_depth++;
1060 1061
	if (sb->s_stack_depth > FILESYSTEM_MAX_STACK_DEPTH) {
		pr_err("overlayfs: maximum fs stacking depth exceeded\n");
1062
		goto out_put_lowerpath;
1063 1064
	}

1065 1066 1067 1068 1069 1070 1071
	if (ufs->config.upperdir) {
		ufs->upper_mnt = clone_private_mount(&upperpath);
		err = PTR_ERR(ufs->upper_mnt);
		if (IS_ERR(ufs->upper_mnt)) {
			pr_err("overlayfs: failed to clone upperpath\n");
			goto out_put_lowerpath;
		}
1072

1073 1074 1075
		ufs->workdir = ovl_workdir_create(ufs->upper_mnt, workpath.dentry);
		err = PTR_ERR(ufs->workdir);
		if (IS_ERR(ufs->workdir)) {
1076 1077 1078 1079
			pr_warn("overlayfs: failed to create directory %s/%s (errno: %i); mounting read-only\n",
				ufs->config.workdir, OVL_WORKDIR_NAME, -err);
			sb->s_flags |= MS_RDONLY;
			ufs->workdir = NULL;
1080
		}
1081 1082 1083 1084

		/*
		 * Upper should support d_type, else whiteouts are visible.
		 * Given workdir and upper are on same fs, we can do
1085 1086
		 * iterate_dir() on workdir. This check requires successful
		 * creation of workdir in previous step.
1087
		 */
1088 1089 1090 1091
		if (ufs->workdir) {
			err = ovl_check_d_type_supported(&workpath);
			if (err < 0)
				goto out_put_workdir;
1092

1093 1094 1095 1096 1097 1098 1099
			/*
			 * We allowed this configuration and don't want to
			 * break users over kernel upgrade. So warn instead
			 * of erroring out.
			 */
			if (!err)
				pr_warn("overlayfs: upper fs needs to support d_type.\n");
1100
		}
M
Miklos Szeredi 已提交
1101 1102
	}

1103
	err = -ENOMEM;
1104
	ufs->lower_mnt = kcalloc(numlower, sizeof(struct vfsmount *), GFP_KERNEL);
1105
	if (ufs->lower_mnt == NULL)
1106
		goto out_put_workdir;
1107 1108
	for (i = 0; i < numlower; i++) {
		struct vfsmount *mnt = clone_private_mount(&stack[i]);
1109

1110
		err = PTR_ERR(mnt);
1111 1112 1113 1114 1115 1116 1117 1118 1119
		if (IS_ERR(mnt)) {
			pr_err("overlayfs: failed to clone lowerpath\n");
			goto out_put_lower_mnt;
		}
		/*
		 * Make lower_mnt R/O.  That way fchmod/fchown on lower file
		 * will fail instead of modifying lower fs.
		 */
		mnt->mnt_flags |= MNT_READONLY;
1120

1121 1122 1123
		ufs->lower_mnt[ufs->numlower] = mnt;
		ufs->numlower++;
	}
M
Miklos Szeredi 已提交
1124

1125 1126
	/* If the upper fs is nonexistent, we mark overlayfs r/o too */
	if (!ufs->upper_mnt)
M
Miklos Szeredi 已提交
1127 1128
		sb->s_flags |= MS_RDONLY;

1129 1130 1131 1132
	if (remote)
		sb->s_d_op = &ovl_reval_dentry_operations;
	else
		sb->s_d_op = &ovl_dentry_operations;
M
Miklos Szeredi 已提交
1133

1134 1135 1136 1137
	ufs->creator_cred = prepare_creds();
	if (!ufs->creator_cred)
		goto out_put_lower_mnt;

M
Miklos Szeredi 已提交
1138
	err = -ENOMEM;
1139
	oe = ovl_alloc_entry(numlower);
1140
	if (!oe)
1141
		goto out_put_cred;
M
Miklos Szeredi 已提交
1142

1143
	root_dentry = d_make_root(ovl_new_inode(sb, S_IFDIR, oe));
M
Miklos Szeredi 已提交
1144
	if (!root_dentry)
1145
		goto out_free_oe;
M
Miklos Szeredi 已提交
1146 1147

	mntput(upperpath.mnt);
1148 1149
	for (i = 0; i < numlower; i++)
		mntput(stack[i].mnt);
M
Miklos Szeredi 已提交
1150
	path_put(&workpath);
1151
	kfree(lowertmp);
M
Miklos Szeredi 已提交
1152 1153

	oe->__upperdentry = upperpath.dentry;
1154 1155 1156 1157
	for (i = 0; i < numlower; i++) {
		oe->lowerstack[i].dentry = stack[i].dentry;
		oe->lowerstack[i].mnt = ufs->lower_mnt[i];
	}
1158
	kfree(stack);
M
Miklos Szeredi 已提交
1159 1160 1161

	root_dentry->d_fsdata = oe;

M
Miklos Szeredi 已提交
1162 1163 1164
	ovl_copyattr(ovl_dentry_real(root_dentry)->d_inode,
		     root_dentry->d_inode);

1165
	sb->s_magic = OVERLAYFS_SUPER_MAGIC;
M
Miklos Szeredi 已提交
1166 1167 1168 1169 1170 1171
	sb->s_op = &ovl_super_operations;
	sb->s_root = root_dentry;
	sb->s_fs_info = ufs;

	return 0;

1172 1173
out_free_oe:
	kfree(oe);
1174 1175
out_put_cred:
	put_cred(ufs->creator_cred);
M
Miklos Szeredi 已提交
1176
out_put_lower_mnt:
1177 1178 1179
	for (i = 0; i < ufs->numlower; i++)
		mntput(ufs->lower_mnt[i]);
	kfree(ufs->lower_mnt);
1180 1181
out_put_workdir:
	dput(ufs->workdir);
M
Miklos Szeredi 已提交
1182 1183
	mntput(ufs->upper_mnt);
out_put_lowerpath:
1184 1185 1186 1187 1188
	for (i = 0; i < numlower; i++)
		path_put(&stack[i]);
	kfree(stack);
out_free_lowertmp:
	kfree(lowertmp);
1189 1190
out_put_workpath:
	path_put(&workpath);
M
Miklos Szeredi 已提交
1191 1192 1193
out_put_upperpath:
	path_put(&upperpath);
out_free_config:
1194 1195 1196 1197
	kfree(ufs->config.lowerdir);
	kfree(ufs->config.upperdir);
	kfree(ufs->config.workdir);
	kfree(ufs);
M
Miklos Szeredi 已提交
1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209
out:
	return err;
}

static struct dentry *ovl_mount(struct file_system_type *fs_type, int flags,
				const char *dev_name, void *raw_data)
{
	return mount_nodev(fs_type, flags, raw_data, ovl_fill_super);
}

static struct file_system_type ovl_fs_type = {
	.owner		= THIS_MODULE,
1210
	.name		= "overlay",
M
Miklos Szeredi 已提交
1211 1212 1213
	.mount		= ovl_mount,
	.kill_sb	= kill_anon_super,
};
1214
MODULE_ALIAS_FS("overlay");
M
Miklos Szeredi 已提交
1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227

static int __init ovl_init(void)
{
	return register_filesystem(&ovl_fs_type);
}

static void __exit ovl_exit(void)
{
	unregister_filesystem(&ovl_fs_type);
}

module_init(ovl_init);
module_exit(ovl_exit);
反馈
建议
客服 返回
顶部