super.c 36.5 KB
Newer Older
M
Miklos Szeredi 已提交
1 2 3 4 5 6 7 8 9
/*
 *
 * 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.
 */

10
#include <uapi/linux/magic.h>
M
Miklos Szeredi 已提交
11 12 13 14 15 16
#include <linux/fs.h>
#include <linux/namei.h>
#include <linux/xattr.h>
#include <linux/mount.h>
#include <linux/parser.h>
#include <linux/module.h>
A
Andy Whitcroft 已提交
17
#include <linux/statfs.h>
E
Erez Zadok 已提交
18
#include <linux/seq_file.h>
M
Miklos Szeredi 已提交
19
#include <linux/posix_acl_xattr.h>
20
#include <linux/exportfs.h>
M
Miklos Szeredi 已提交
21 22 23 24 25 26 27 28 29
#include "overlayfs.h"

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


struct ovl_dir_cache;

30 31
#define OVL_MAX_STACK 500

32 33 34 35
static bool ovl_redirect_dir_def = IS_ENABLED(CONFIG_OVERLAY_FS_REDIRECT_DIR);
module_param_named(redirect_dir, ovl_redirect_dir_def, bool, 0644);
MODULE_PARM_DESC(ovl_redirect_dir_def,
		 "Default to on or off for the redirect_dir feature");
M
Miklos Szeredi 已提交
36

37 38 39 40 41 42 43
static bool ovl_redirect_always_follow =
	IS_ENABLED(CONFIG_OVERLAY_FS_REDIRECT_ALWAYS_FOLLOW);
module_param_named(redirect_always_follow, ovl_redirect_always_follow,
		   bool, 0644);
MODULE_PARM_DESC(ovl_redirect_always_follow,
		 "Follow redirects even if redirect_dir feature is turned off");

44 45 46 47 48
static bool ovl_index_def = IS_ENABLED(CONFIG_OVERLAY_FS_INDEX);
module_param_named(index, ovl_index_def, bool, 0644);
MODULE_PARM_DESC(ovl_index_def,
		 "Default to on or off for the inodes index feature");

49 50 51 52 53
static bool ovl_nfs_export_def = IS_ENABLED(CONFIG_OVERLAY_FS_NFS_EXPORT);
module_param_named(nfs_export, ovl_nfs_export_def, bool, 0644);
MODULE_PARM_DESC(ovl_nfs_export_def,
		 "Default to on or off for the NFS export feature");

54 55 56 57 58
static bool ovl_xino_auto_def = IS_ENABLED(CONFIG_OVERLAY_FS_XINO_AUTO);
module_param_named(xino_auto, ovl_xino_auto_def, bool, 0644);
MODULE_PARM_DESC(ovl_xino_auto_def,
		 "Auto enable xino feature");

59 60 61 62 63 64 65 66
static void ovl_entry_stack_free(struct ovl_entry *oe)
{
	unsigned int i;

	for (i = 0; i < oe->numlower; i++)
		dput(oe->lowerstack[i].dentry);
}

M
Miklos Szeredi 已提交
67 68 69 70 71
static void ovl_dentry_release(struct dentry *dentry)
{
	struct ovl_entry *oe = dentry->d_fsdata;

	if (oe) {
72
		ovl_entry_stack_free(oe);
M
Miklos Szeredi 已提交
73 74 75 76
		kfree_rcu(oe, rcu);
	}
}

77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92
static int ovl_check_append_only(struct inode *inode, int flag)
{
	/*
	 * This test was moot in vfs may_open() because overlay inode does
	 * not have the S_APPEND flag, so re-check on real upper inode
	 */
	if (IS_APPEND(inode)) {
		if  ((flag & O_ACCMODE) != O_RDONLY && !(flag & O_APPEND))
			return -EPERM;
		if (flag & O_TRUNC)
			return -EPERM;
	}

	return 0;
}

93 94
static struct dentry *ovl_d_real(struct dentry *dentry,
				 const struct inode *inode,
95
				 unsigned int open_flags)
M
Miklos Szeredi 已提交
96 97
{
	struct dentry *real;
98
	int err;
M
Miklos Szeredi 已提交
99

100 101 102 103
	/* It's an overlay file */
	if (inode && d_inode(dentry) == inode)
		return dentry;

104
	if (!d_is_reg(dentry)) {
M
Miklos Szeredi 已提交
105 106 107 108 109
		if (!inode || inode == d_inode(dentry))
			return dentry;
		goto bug;
	}

110
	if (open_flags) {
111
		err = ovl_open_maybe_copy_up(dentry, open_flags);
112 113 114 115
		if (err)
			return ERR_PTR(err);
	}

M
Miklos Szeredi 已提交
116
	real = ovl_dentry_upper(dentry);
117 118 119 120 121 122
	if (real && (!inode || inode == d_inode(real))) {
		if (!inode) {
			err = ovl_check_append_only(d_inode(real), open_flags);
			if (err)
				return ERR_PTR(err);
		}
M
Miklos Szeredi 已提交
123
		return real;
124
	}
M
Miklos Szeredi 已提交
125 126 127 128 129

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

M
Miklos Szeredi 已提交
130
	/* Handle recursion */
131
	real = d_real(real, inode, open_flags);
M
Miklos Szeredi 已提交
132

M
Miklos Szeredi 已提交
133 134 135
	if (!inode || inode == d_inode(real))
		return real;
bug:
M
Miklos Szeredi 已提交
136
	WARN(1, "ovl_d_real(%pd4, %s:%lu): real dentry not found\n", dentry,
M
Miklos Szeredi 已提交
137 138 139 140
	     inode ? inode->i_sb->s_id : "NULL", inode ? inode->i_ino : 0);
	return dentry;
}

141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181
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
182 183
static const struct dentry_operations ovl_dentry_operations = {
	.d_release = ovl_dentry_release,
M
Miklos Szeredi 已提交
184
	.d_real = ovl_d_real,
M
Miklos Szeredi 已提交
185 186
};

187 188
static const struct dentry_operations ovl_reval_dentry_operations = {
	.d_release = ovl_dentry_release,
M
Miklos Szeredi 已提交
189
	.d_real = ovl_d_real,
190 191 192 193
	.d_revalidate = ovl_dentry_revalidate,
	.d_weak_revalidate = ovl_dentry_weak_revalidate,
};

194 195 196 197 198 199
static struct kmem_cache *ovl_inode_cachep;

static struct inode *ovl_alloc_inode(struct super_block *sb)
{
	struct ovl_inode *oi = kmem_cache_alloc(ovl_inode_cachep, GFP_KERNEL);

200 201 202
	if (!oi)
		return NULL;

203
	oi->cache = NULL;
M
Miklos Szeredi 已提交
204
	oi->redirect = NULL;
205
	oi->version = 0;
M
Miklos Szeredi 已提交
206
	oi->flags = 0;
207
	oi->__upperdentry = NULL;
208
	oi->lower = NULL;
209
	mutex_init(&oi->lock);
210

211 212 213 214 215 216 217 218 219 220 221 222
	return &oi->vfs_inode;
}

static void ovl_i_callback(struct rcu_head *head)
{
	struct inode *inode = container_of(head, struct inode, i_rcu);

	kmem_cache_free(ovl_inode_cachep, OVL_I(inode));
}

static void ovl_destroy_inode(struct inode *inode)
{
223 224 225
	struct ovl_inode *oi = OVL_I(inode);

	dput(oi->__upperdentry);
226
	iput(oi->lower);
M
Miklos Szeredi 已提交
227
	kfree(oi->redirect);
228
	ovl_dir_cache_free(inode);
229
	mutex_destroy(&oi->lock);
230

231 232 233
	call_rcu(&inode->i_rcu, ovl_i_callback);
}

M
Miklos Szeredi 已提交
234
static void ovl_free_fs(struct ovl_fs *ofs)
M
Miklos Szeredi 已提交
235
{
236
	unsigned i;
M
Miklos Szeredi 已提交
237

M
Miklos Szeredi 已提交
238 239 240 241 242 243 244 245
	dput(ofs->indexdir);
	dput(ofs->workdir);
	if (ofs->workdir_locked)
		ovl_inuse_unlock(ofs->workbasedir);
	dput(ofs->workbasedir);
	if (ofs->upperdir_locked)
		ovl_inuse_unlock(ofs->upper_mnt->mnt_root);
	mntput(ofs->upper_mnt);
246
	for (i = 0; i < ofs->numlower; i++)
M
Miklos Szeredi 已提交
247
		mntput(ofs->lower_layers[i].mnt);
248 249
	for (i = 0; i < ofs->numlowerfs; i++)
		free_anon_bdev(ofs->lower_fs[i].pseudo_dev);
M
Miklos Szeredi 已提交
250
	kfree(ofs->lower_layers);
251
	kfree(ofs->lower_fs);
M
Miklos Szeredi 已提交
252 253 254 255

	kfree(ofs->config.lowerdir);
	kfree(ofs->config.upperdir);
	kfree(ofs->config.workdir);
256
	kfree(ofs->config.redirect_mode);
M
Miklos Szeredi 已提交
257 258 259
	if (ofs->creator_cred)
		put_cred(ofs->creator_cred);
	kfree(ofs);
M
Miklos Szeredi 已提交
260 261
}

262 263 264 265 266 267 268
static void ovl_put_super(struct super_block *sb)
{
	struct ovl_fs *ofs = sb->s_fs_info;

	ovl_free_fs(ofs);
}

269
/* Sync real dirty inodes in upper filesystem (if it exists) */
270 271
static int ovl_sync_fs(struct super_block *sb, int wait)
{
M
Miklos Szeredi 已提交
272
	struct ovl_fs *ofs = sb->s_fs_info;
273 274 275
	struct super_block *upper_sb;
	int ret;

M
Miklos Szeredi 已提交
276
	if (!ofs->upper_mnt)
277
		return 0;
278 279 280 281 282 283 284 285 286 287

	/*
	 * If this is a sync(2) call or an emergency sync, all the super blocks
	 * will be iterated, including upper_sb, so no need to do anything.
	 *
	 * If this is a syncfs(2) call, then we do need to call
	 * sync_filesystem() on upper_sb, but enough if we do it when being
	 * called with wait == 1.
	 */
	if (!wait)
288 289
		return 0;

290 291
	upper_sb = ofs->upper_mnt->mnt_sb;

292
	down_read(&upper_sb->s_umount);
293
	ret = sync_filesystem(upper_sb);
294
	up_read(&upper_sb->s_umount);
295

296 297 298
	return ret;
}

A
Andy Whitcroft 已提交
299 300 301 302 303 304
/**
 * 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
305
 * filesystem pass the statfs to the upper filesystem (if it exists)
A
Andy Whitcroft 已提交
306 307 308 309 310 311 312 313
 */
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;

314
	ovl_path_real(root_dentry, &path);
A
Andy Whitcroft 已提交
315 316 317

	err = vfs_statfs(&path, buf);
	if (!err) {
M
Miklos Szeredi 已提交
318
		buf->f_namelen = ofs->namelen;
A
Andy Whitcroft 已提交
319 320 321 322 323 324
		buf->f_type = OVERLAYFS_SUPER_MAGIC;
	}

	return err;
}

325
/* Will this overlay be forced to mount/remount ro? */
M
Miklos Szeredi 已提交
326
static bool ovl_force_readonly(struct ovl_fs *ofs)
327
{
M
Miklos Szeredi 已提交
328
	return (!ofs->upper_mnt || !ofs->workdir);
329 330
}

331 332 333 334 335
static const char *ovl_redirect_mode_def(void)
{
	return ovl_redirect_dir_def ? "on" : "off";
}

336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352
enum {
	OVL_XINO_OFF,
	OVL_XINO_AUTO,
	OVL_XINO_ON,
};

static const char * const ovl_xino_str[] = {
	"off",
	"auto",
	"on",
};

static inline int ovl_xino_def(void)
{
	return ovl_xino_auto_def ? OVL_XINO_AUTO : OVL_XINO_OFF;
}

E
Erez Zadok 已提交
353 354 355 356 357 358 359 360 361
/**
 * 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;
M
Miklos Szeredi 已提交
362
	struct ovl_fs *ofs = sb->s_fs_info;
E
Erez Zadok 已提交
363

M
Miklos Szeredi 已提交
364 365 366 367
	seq_show_option(m, "lowerdir", ofs->config.lowerdir);
	if (ofs->config.upperdir) {
		seq_show_option(m, "upperdir", ofs->config.upperdir);
		seq_show_option(m, "workdir", ofs->config.workdir);
M
Miklos Szeredi 已提交
368
	}
M
Miklos Szeredi 已提交
369
	if (ofs->config.default_permissions)
M
Miklos Szeredi 已提交
370
		seq_puts(m, ",default_permissions");
371 372
	if (strcmp(ofs->config.redirect_mode, ovl_redirect_mode_def()) != 0)
		seq_printf(m, ",redirect_dir=%s", ofs->config.redirect_mode);
M
Miklos Szeredi 已提交
373
	if (ofs->config.index != ovl_index_def)
374
		seq_printf(m, ",index=%s", ofs->config.index ? "on" : "off");
375 376 377
	if (ofs->config.nfs_export != ovl_nfs_export_def)
		seq_printf(m, ",nfs_export=%s", ofs->config.nfs_export ?
						"on" : "off");
378 379
	if (ofs->config.xino != ovl_xino_def())
		seq_printf(m, ",xino=%s", ovl_xino_str[ofs->config.xino]);
E
Erez Zadok 已提交
380 381 382
	return 0;
}

383 384
static int ovl_remount(struct super_block *sb, int *flags, char *data)
{
M
Miklos Szeredi 已提交
385
	struct ovl_fs *ofs = sb->s_fs_info;
386

387
	if (!(*flags & SB_RDONLY) && ovl_force_readonly(ofs))
388 389 390 391 392
		return -EROFS;

	return 0;
}

M
Miklos Szeredi 已提交
393
static const struct super_operations ovl_super_operations = {
394 395 396
	.alloc_inode	= ovl_alloc_inode,
	.destroy_inode	= ovl_destroy_inode,
	.drop_inode	= generic_delete_inode,
M
Miklos Szeredi 已提交
397
	.put_super	= ovl_put_super,
398
	.sync_fs	= ovl_sync_fs,
A
Andy Whitcroft 已提交
399
	.statfs		= ovl_statfs,
E
Erez Zadok 已提交
400
	.show_options	= ovl_show_options,
401
	.remount_fs	= ovl_remount,
M
Miklos Szeredi 已提交
402 403 404 405 406 407
};

enum {
	OPT_LOWERDIR,
	OPT_UPPERDIR,
	OPT_WORKDIR,
M
Miklos Szeredi 已提交
408
	OPT_DEFAULT_PERMISSIONS,
409
	OPT_REDIRECT_DIR,
410 411
	OPT_INDEX_ON,
	OPT_INDEX_OFF,
412 413
	OPT_NFS_EXPORT_ON,
	OPT_NFS_EXPORT_OFF,
414 415 416
	OPT_XINO_ON,
	OPT_XINO_OFF,
	OPT_XINO_AUTO,
M
Miklos Szeredi 已提交
417 418 419 420 421 422 423
	OPT_ERR,
};

static const match_table_t ovl_tokens = {
	{OPT_LOWERDIR,			"lowerdir=%s"},
	{OPT_UPPERDIR,			"upperdir=%s"},
	{OPT_WORKDIR,			"workdir=%s"},
M
Miklos Szeredi 已提交
424
	{OPT_DEFAULT_PERMISSIONS,	"default_permissions"},
425
	{OPT_REDIRECT_DIR,		"redirect_dir=%s"},
426 427
	{OPT_INDEX_ON,			"index=on"},
	{OPT_INDEX_OFF,			"index=off"},
428 429
	{OPT_NFS_EXPORT_ON,		"nfs_export=on"},
	{OPT_NFS_EXPORT_OFF,		"nfs_export=off"},
430 431 432
	{OPT_XINO_ON,			"xino=on"},
	{OPT_XINO_OFF,			"xino=off"},
	{OPT_XINO_AUTO,			"xino=auto"},
M
Miklos Szeredi 已提交
433 434 435
	{OPT_ERR,			NULL}
};

M
Miklos Szeredi 已提交
436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458
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;
}

459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481
static int ovl_parse_redirect_mode(struct ovl_config *config, const char *mode)
{
	if (strcmp(mode, "on") == 0) {
		config->redirect_dir = true;
		/*
		 * Does not make sense to have redirect creation without
		 * redirect following.
		 */
		config->redirect_follow = true;
	} else if (strcmp(mode, "follow") == 0) {
		config->redirect_follow = true;
	} else if (strcmp(mode, "off") == 0) {
		if (ovl_redirect_always_follow)
			config->redirect_follow = true;
	} else if (strcmp(mode, "nofollow") != 0) {
		pr_err("overlayfs: bad mount option \"redirect_dir=%s\"\n",
		       mode);
		return -EINVAL;
	}

	return 0;
}

M
Miklos Szeredi 已提交
482 483 484 485
static int ovl_parse_opt(char *opt, struct ovl_config *config)
{
	char *p;

486 487 488 489
	config->redirect_mode = kstrdup(ovl_redirect_mode_def(), GFP_KERNEL);
	if (!config->redirect_mode)
		return -ENOMEM;

M
Miklos Szeredi 已提交
490
	while ((p = ovl_next_opt(&opt)) != NULL) {
M
Miklos Szeredi 已提交
491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519
		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;

M
Miklos Szeredi 已提交
520 521 522 523
		case OPT_DEFAULT_PERMISSIONS:
			config->default_permissions = true;
			break;

524 525 526 527 528
		case OPT_REDIRECT_DIR:
			kfree(config->redirect_mode);
			config->redirect_mode = match_strdup(&args[0]);
			if (!config->redirect_mode)
				return -ENOMEM;
M
Miklos Szeredi 已提交
529 530
			break;

531 532 533 534 535 536 537 538
		case OPT_INDEX_ON:
			config->index = true;
			break;

		case OPT_INDEX_OFF:
			config->index = false;
			break;

539 540 541 542 543 544 545 546
		case OPT_NFS_EXPORT_ON:
			config->nfs_export = true;
			break;

		case OPT_NFS_EXPORT_OFF:
			config->nfs_export = false;
			break;

547 548 549 550 551 552 553 554 555 556 557 558
		case OPT_XINO_ON:
			config->xino = OVL_XINO_ON;
			break;

		case OPT_XINO_OFF:
			config->xino = OVL_XINO_OFF;
			break;

		case OPT_XINO_AUTO:
			config->xino = OVL_XINO_AUTO;
			break;

M
Miklos Szeredi 已提交
559
		default:
560
			pr_err("overlayfs: unrecognized mount option \"%s\" or missing value\n", p);
M
Miklos Szeredi 已提交
561 562 563
			return -EINVAL;
		}
	}
H
hujianyang 已提交
564 565 566 567 568 569 570 571 572

	/* 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;
	}

573
	return ovl_parse_redirect_mode(config, config->redirect_mode);
M
Miklos Szeredi 已提交
574 575 576
}

#define OVL_WORKDIR_NAME "work"
577
#define OVL_INDEXDIR_NAME "index"
M
Miklos Szeredi 已提交
578

M
Miklos Szeredi 已提交
579
static struct dentry *ovl_workdir_create(struct ovl_fs *ofs,
580
					 const char *name, bool persist)
M
Miklos Szeredi 已提交
581
{
M
Miklos Szeredi 已提交
582 583
	struct inode *dir =  ofs->workbasedir->d_inode;
	struct vfsmount *mnt = ofs->upper_mnt;
M
Miklos Szeredi 已提交
584 585 586
	struct dentry *work;
	int err;
	bool retried = false;
587
	bool locked = false;
M
Miklos Szeredi 已提交
588

A
Al Viro 已提交
589
	inode_lock_nested(dir, I_MUTEX_PARENT);
590 591
	locked = true;

M
Miklos Szeredi 已提交
592
retry:
M
Miklos Szeredi 已提交
593
	work = lookup_one_len(name, ofs->workbasedir, strlen(name));
M
Miklos Szeredi 已提交
594 595

	if (!IS_ERR(work)) {
596 597
		struct iattr attr = {
			.ia_valid = ATTR_MODE,
A
Al Viro 已提交
598
			.ia_mode = S_IFDIR | 0,
599
		};
M
Miklos Szeredi 已提交
600 601 602 603 604 605

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

606 607 608
			if (persist)
				goto out_unlock;

M
Miklos Szeredi 已提交
609
			retried = true;
M
Miklos Szeredi 已提交
610
			ovl_workdir_cleanup(dir, mnt, work, 0);
M
Miklos Szeredi 已提交
611 612 613 614
			dput(work);
			goto retry;
		}

615 616 617 618
		work = ovl_create_real(dir, work, OVL_CATTR(attr.ia_mode));
		err = PTR_ERR(work);
		if (IS_ERR(work))
			goto out_err;
619

620 621 622 623 624 625 626 627 628 629 630 631 632
		/*
		 * Try to remove POSIX ACL xattrs from workdir.  We are good if:
		 *
		 * a) success (there was a POSIX ACL xattr and was removed)
		 * b) -ENODATA (there was no POSIX ACL xattr)
		 * c) -EOPNOTSUPP (POSIX ACL xattrs are not supported)
		 *
		 * There are various other error values that could effectively
		 * mean that the xattr doesn't exist (e.g. -ERANGE is returned
		 * if the xattr name is too long), but the set of filesystems
		 * allowed as upper are limited to "normal" ones, where checking
		 * for the above two errors is sufficient.
		 */
633
		err = vfs_removexattr(work, XATTR_NAME_POSIX_ACL_DEFAULT);
M
Miklos Szeredi 已提交
634
		if (err && err != -ENODATA && err != -EOPNOTSUPP)
635 636 637
			goto out_dput;

		err = vfs_removexattr(work, XATTR_NAME_POSIX_ACL_ACCESS);
M
Miklos Szeredi 已提交
638
		if (err && err != -ENODATA && err != -EOPNOTSUPP)
639 640 641 642 643 644 645 646
			goto out_dput;

		/* Clear any inherited mode bits */
		inode_lock(work->d_inode);
		err = notify_change(work, &attr, NULL);
		inode_unlock(work->d_inode);
		if (err)
			goto out_dput;
647 648 649
	} else {
		err = PTR_ERR(work);
		goto out_err;
M
Miklos Szeredi 已提交
650 651
	}
out_unlock:
652 653
	if (locked)
		inode_unlock(dir);
M
Miklos Szeredi 已提交
654 655 656 657 658

	return work;

out_dput:
	dput(work);
659 660
out_err:
	pr_warn("overlayfs: failed to create directory %s/%s (errno: %i); mounting read-only\n",
M
Miklos Szeredi 已提交
661
		ofs->config.workdir, name, -err);
662
	work = NULL;
M
Miklos Szeredi 已提交
663 664 665
	goto out_unlock;
}

M
Miklos Szeredi 已提交
666 667 668 669 670 671 672 673 674 675 676 677 678
static void ovl_unescape(char *s)
{
	char *d = s;

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

M
Miklos Szeredi 已提交
679 680
static int ovl_mount_dir_noesc(const char *name, struct path *path)
{
681
	int err = -EINVAL;
M
Miklos Szeredi 已提交
682

683 684 685 686
	if (!*name) {
		pr_err("overlayfs: empty lowerdir\n");
		goto out;
	}
M
Miklos Szeredi 已提交
687 688 689 690 691 692
	err = kern_path(name, LOOKUP_FOLLOW, path);
	if (err) {
		pr_err("overlayfs: failed to resolve '%s': %i\n", name, err);
		goto out;
	}
	err = -EINVAL;
693
	if (ovl_dentry_weird(path->dentry)) {
M
Miklos Szeredi 已提交
694 695 696
		pr_err("overlayfs: filesystem on '%s' not supported\n", name);
		goto out_put;
	}
M
Miklos Szeredi 已提交
697
	if (!d_is_dir(path->dentry)) {
M
Miklos Szeredi 已提交
698 699 700 701 702 703
		pr_err("overlayfs: '%s' not a directory\n", name);
		goto out_put;
	}
	return 0;

out_put:
704
	path_put_init(path);
M
Miklos Szeredi 已提交
705 706 707 708 709 710 711 712 713 714 715 716
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);
717 718 719 720 721

		if (!err)
			if (ovl_dentry_remote(path->dentry)) {
				pr_err("overlayfs: filesystem on '%s' not supported as upperdir\n",
				       tmp);
722
				path_put_init(path);
723 724
				err = -EINVAL;
			}
M
Miklos Szeredi 已提交
725 726 727 728 729
		kfree(tmp);
	}
	return err;
}

M
Miklos Szeredi 已提交
730 731
static int ovl_check_namelen(struct path *path, struct ovl_fs *ofs,
			     const char *name)
M
Miklos Szeredi 已提交
732 733
{
	struct kstatfs statfs;
M
Miklos Szeredi 已提交
734 735 736 737 738 739 740 741 742 743 744 745 746
	int err = vfs_statfs(path, &statfs);

	if (err)
		pr_err("overlayfs: statfs failed on '%s'\n", name);
	else
		ofs->namelen = max(ofs->namelen, statfs.f_namelen);

	return err;
}

static int ovl_lower_dir(const char *name, struct path *path,
			 struct ovl_fs *ofs, int *stack_depth, bool *remote)
{
747
	int fh_type;
M
Miklos Szeredi 已提交
748
	int err;
M
Miklos Szeredi 已提交
749

750
	err = ovl_mount_dir_noesc(name, path);
M
Miklos Szeredi 已提交
751 752 753
	if (err)
		goto out;

M
Miklos Szeredi 已提交
754 755
	err = ovl_check_namelen(path, ofs, name);
	if (err)
M
Miklos Szeredi 已提交
756
		goto out_put;
M
Miklos Szeredi 已提交
757

M
Miklos Szeredi 已提交
758 759
	*stack_depth = max(*stack_depth, path->mnt->mnt_sb->s_stack_depth);

760 761 762
	if (ovl_dentry_remote(path->dentry))
		*remote = true;

763
	/*
764 765
	 * The inodes index feature and NFS export need to encode and decode
	 * file handles, so they require that all layers support them.
766
	 */
767
	fh_type = ovl_can_decode_fh(path->dentry->d_sb);
768
	if ((ofs->config.nfs_export ||
769
	     (ofs->config.index && ofs->config.upperdir)) && !fh_type) {
770
		ofs->config.index = false;
771 772 773
		ofs->config.nfs_export = false;
		pr_warn("overlayfs: fs on '%s' does not support file handles, falling back to index=off,nfs_export=off.\n",
			name);
774 775
	}

776 777 778 779
	/* Check if lower fs has 32bit inode numbers */
	if (fh_type != FILEID_INO32_GEN)
		ofs->xino_bits = 0;

M
Miklos Szeredi 已提交
780 781 782
	return 0;

out_put:
783
	path_put_init(path);
M
Miklos Szeredi 已提交
784 785 786 787
out:
	return err;
}

M
Miklos Szeredi 已提交
788 789 790 791 792 793 794 795 796 797 798 799
/* 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;
}

800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819
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;
}

820 821 822 823 824
static int __maybe_unused
ovl_posix_acl_xattr_get(const struct xattr_handler *handler,
			struct dentry *dentry, struct inode *inode,
			const char *name, void *buffer, size_t size)
{
825
	return ovl_xattr_get(dentry, inode, handler->name, buffer, size);
826 827
}

828 829 830 831 832
static int __maybe_unused
ovl_posix_acl_xattr_set(const struct xattr_handler *handler,
			struct dentry *dentry, struct inode *inode,
			const char *name, const void *value,
			size_t size, int flags)
M
Miklos Szeredi 已提交
833 834
{
	struct dentry *workdir = ovl_workdir(dentry);
835
	struct inode *realinode = ovl_inode_real(inode);
M
Miklos Szeredi 已提交
836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859
	struct posix_acl *acl = NULL;
	int err;

	/* Check that everything is OK before copy-up */
	if (value) {
		acl = posix_acl_from_xattr(&init_user_ns, value, size);
		if (IS_ERR(acl))
			return PTR_ERR(acl);
	}
	err = -EOPNOTSUPP;
	if (!IS_POSIXACL(d_inode(workdir)))
		goto out_acl_release;
	if (!realinode->i_op->set_acl)
		goto out_acl_release;
	if (handler->flags == ACL_TYPE_DEFAULT && !S_ISDIR(inode->i_mode)) {
		err = acl ? -EACCES : 0;
		goto out_acl_release;
	}
	err = -EPERM;
	if (!inode_owner_or_capable(inode))
		goto out_acl_release;

	posix_acl_release(acl);

860 861 862 863 864 865 866 867 868 869 870 871 872 873 874
	/*
	 * Check if sgid bit needs to be cleared (actual setacl operation will
	 * be done with mounter's capabilities and so that won't do it for us).
	 */
	if (unlikely(inode->i_mode & S_ISGID) &&
	    handler->flags == ACL_TYPE_ACCESS &&
	    !in_group_p(inode->i_gid) &&
	    !capable_wrt_inode_uidgid(inode, CAP_FSETID)) {
		struct iattr iattr = { .ia_valid = ATTR_KILL_SGID };

		err = ovl_setattr(dentry, &iattr);
		if (err)
			return err;
	}

875
	err = ovl_xattr_set(dentry, inode, handler->name, value, size, flags);
876
	if (!err)
877
		ovl_copyattr(ovl_inode_real(inode), inode);
878 879

	return err;
M
Miklos Szeredi 已提交
880 881 882 883 884 885

out_acl_release:
	posix_acl_release(acl);
	return err;
}

886 887 888 889
static int ovl_own_xattr_get(const struct xattr_handler *handler,
			     struct dentry *dentry, struct inode *inode,
			     const char *name, void *buffer, size_t size)
{
A
Amir Goldstein 已提交
890
	return -EOPNOTSUPP;
891 892
}

M
Miklos Szeredi 已提交
893 894 895 896 897
static int ovl_own_xattr_set(const struct xattr_handler *handler,
			     struct dentry *dentry, struct inode *inode,
			     const char *name, const void *value,
			     size_t size, int flags)
{
A
Amir Goldstein 已提交
898
	return -EOPNOTSUPP;
M
Miklos Szeredi 已提交
899 900
}

901 902 903 904
static int ovl_other_xattr_get(const struct xattr_handler *handler,
			       struct dentry *dentry, struct inode *inode,
			       const char *name, void *buffer, size_t size)
{
905
	return ovl_xattr_get(dentry, inode, name, buffer, size);
906 907
}

908 909 910 911 912
static int ovl_other_xattr_set(const struct xattr_handler *handler,
			       struct dentry *dentry, struct inode *inode,
			       const char *name, const void *value,
			       size_t size, int flags)
{
913
	return ovl_xattr_set(dentry, inode, name, value, size, flags);
914 915
}

916 917
static const struct xattr_handler __maybe_unused
ovl_posix_acl_access_xattr_handler = {
M
Miklos Szeredi 已提交
918 919
	.name = XATTR_NAME_POSIX_ACL_ACCESS,
	.flags = ACL_TYPE_ACCESS,
920
	.get = ovl_posix_acl_xattr_get,
M
Miklos Szeredi 已提交
921 922 923
	.set = ovl_posix_acl_xattr_set,
};

924 925
static const struct xattr_handler __maybe_unused
ovl_posix_acl_default_xattr_handler = {
M
Miklos Szeredi 已提交
926 927
	.name = XATTR_NAME_POSIX_ACL_DEFAULT,
	.flags = ACL_TYPE_DEFAULT,
928
	.get = ovl_posix_acl_xattr_get,
M
Miklos Szeredi 已提交
929 930 931 932 933
	.set = ovl_posix_acl_xattr_set,
};

static const struct xattr_handler ovl_own_xattr_handler = {
	.prefix	= OVL_XATTR_PREFIX,
934
	.get = ovl_own_xattr_get,
M
Miklos Szeredi 已提交
935 936 937 938 939
	.set = ovl_own_xattr_set,
};

static const struct xattr_handler ovl_other_xattr_handler = {
	.prefix	= "", /* catch all */
940
	.get = ovl_other_xattr_get,
M
Miklos Szeredi 已提交
941 942 943 944
	.set = ovl_other_xattr_set,
};

static const struct xattr_handler *ovl_xattr_handlers[] = {
945
#ifdef CONFIG_FS_POSIX_ACL
M
Miklos Szeredi 已提交
946 947
	&ovl_posix_acl_access_xattr_handler,
	&ovl_posix_acl_default_xattr_handler,
948
#endif
M
Miklos Szeredi 已提交
949 950 951 952 953
	&ovl_own_xattr_handler,
	&ovl_other_xattr_handler,
	NULL
};

M
Miklos Szeredi 已提交
954
static int ovl_get_upper(struct ovl_fs *ofs, struct path *upperpath)
955
{
M
Miklos Szeredi 已提交
956
	struct vfsmount *upper_mnt;
957 958
	int err;

M
Miklos Szeredi 已提交
959
	err = ovl_mount_dir(ofs->config.upperdir, upperpath);
960 961 962 963 964 965 966 967 968 969
	if (err)
		goto out;

	/* Upper fs should not be r/o */
	if (sb_rdonly(upperpath->mnt->mnt_sb)) {
		pr_err("overlayfs: upper fs is r/o, try multi-lower layers mount\n");
		err = -EINVAL;
		goto out;
	}

M
Miklos Szeredi 已提交
970
	err = ovl_check_namelen(upperpath, ofs, ofs->config.upperdir);
971 972 973 974 975
	if (err)
		goto out;

	err = -EBUSY;
	if (ovl_inuse_trylock(upperpath->dentry)) {
M
Miklos Szeredi 已提交
976 977
		ofs->upperdir_locked = true;
	} else if (ofs->config.index) {
978 979 980 981 982
		pr_err("overlayfs: upperdir is in-use by another mount, mount with '-o index=off' to override exclusive upperdir protection.\n");
		goto out;
	} else {
		pr_warn("overlayfs: upperdir is in-use by another mount, accessing files from both mounts will result in undefined behavior.\n");
	}
M
Miklos Szeredi 已提交
983 984 985 986 987 988 989 990 991 992

	upper_mnt = clone_private_mount(upperpath);
	err = PTR_ERR(upper_mnt);
	if (IS_ERR(upper_mnt)) {
		pr_err("overlayfs: failed to clone upperpath\n");
		goto out;
	}

	/* Don't inherit atime flags */
	upper_mnt->mnt_flags &= ~(MNT_NOATIME | MNT_NODIRATIME | MNT_RELATIME);
M
Miklos Szeredi 已提交
993
	ofs->upper_mnt = upper_mnt;
994 995 996 997 998
	err = 0;
out:
	return err;
}

M
Miklos Szeredi 已提交
999
static int ovl_make_workdir(struct ovl_fs *ofs, struct path *workpath)
1000
{
1001
	struct vfsmount *mnt = ofs->upper_mnt;
1002
	struct dentry *temp;
1003
	int fh_type;
1004 1005
	int err;

1006 1007 1008 1009
	err = mnt_want_write(mnt);
	if (err)
		return err;

M
Miklos Szeredi 已提交
1010 1011
	ofs->workdir = ovl_workdir_create(ofs, OVL_WORKDIR_NAME, false);
	if (!ofs->workdir)
1012
		goto out;
1013 1014 1015 1016 1017 1018 1019 1020 1021

	/*
	 * Upper should support d_type, else whiteouts are visible.  Given
	 * workdir and upper are on same fs, we can do iterate_dir() on
	 * workdir. This check requires successful creation of workdir in
	 * previous step.
	 */
	err = ovl_check_d_type_supported(workpath);
	if (err < 0)
1022
		goto out;
1023 1024 1025 1026 1027 1028 1029 1030 1031

	/*
	 * 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");

	/* Check if upper/work fs supports O_TMPFILE */
M
Miklos Szeredi 已提交
1032 1033 1034
	temp = ovl_do_tmpfile(ofs->workdir, S_IFREG | 0);
	ofs->tmpfile = !IS_ERR(temp);
	if (ofs->tmpfile)
1035 1036 1037 1038 1039 1040 1041
		dput(temp);
	else
		pr_warn("overlayfs: upper fs does not support tmpfile.\n");

	/*
	 * Check if upper/work fs supports trusted.overlay.* xattr
	 */
M
Miklos Szeredi 已提交
1042
	err = ovl_do_setxattr(ofs->workdir, OVL_XATTR_OPAQUE, "0", 1, 0);
1043
	if (err) {
M
Miklos Szeredi 已提交
1044
		ofs->noxattr = true;
1045 1046
		ofs->config.index = false;
		pr_warn("overlayfs: upper fs does not support xattr, falling back to index=off.\n");
1047
		err = 0;
1048
	} else {
M
Miklos Szeredi 已提交
1049
		vfs_removexattr(ofs->workdir, OVL_XATTR_OPAQUE);
1050 1051 1052
	}

	/* Check if upper/work fs supports file handles */
1053 1054
	fh_type = ovl_can_decode_fh(ofs->workdir->d_sb);
	if (ofs->config.index && !fh_type) {
M
Miklos Szeredi 已提交
1055
		ofs->config.index = false;
1056 1057 1058
		pr_warn("overlayfs: upper fs does not support file handles, falling back to index=off.\n");
	}

1059 1060 1061 1062
	/* Check if upper fs has 32bit inode numbers */
	if (fh_type != FILEID_INO32_GEN)
		ofs->xino_bits = 0;

1063 1064 1065 1066 1067 1068
	/* NFS export of r/w mount depends on index */
	if (ofs->config.nfs_export && !ofs->config.index) {
		pr_warn("overlayfs: NFS export requires \"index=on\", falling back to nfs_export=off.\n");
		ofs->config.nfs_export = false;
	}

1069 1070 1071
out:
	mnt_drop_write(mnt);
	return err;
1072 1073
}

M
Miklos Szeredi 已提交
1074
static int ovl_get_workdir(struct ovl_fs *ofs, struct path *upperpath)
1075 1076
{
	int err;
M
Miklos Szeredi 已提交
1077
	struct path workpath = { };
1078

M
Miklos Szeredi 已提交
1079
	err = ovl_mount_dir(ofs->config.workdir, &workpath);
1080 1081 1082 1083
	if (err)
		goto out;

	err = -EINVAL;
M
Miklos Szeredi 已提交
1084
	if (upperpath->mnt != workpath.mnt) {
1085 1086 1087
		pr_err("overlayfs: workdir and upperdir must reside under the same mount\n");
		goto out;
	}
M
Miklos Szeredi 已提交
1088
	if (!ovl_workdir_ok(workpath.dentry, upperpath->dentry)) {
1089 1090 1091 1092 1093
		pr_err("overlayfs: workdir and upperdir must be separate subtrees\n");
		goto out;
	}

	err = -EBUSY;
M
Miklos Szeredi 已提交
1094
	if (ovl_inuse_trylock(workpath.dentry)) {
M
Miklos Szeredi 已提交
1095 1096
		ofs->workdir_locked = true;
	} else if (ofs->config.index) {
1097 1098 1099 1100 1101 1102
		pr_err("overlayfs: workdir is in-use by another mount, mount with '-o index=off' to override exclusive workdir protection.\n");
		goto out;
	} else {
		pr_warn("overlayfs: workdir is in-use by another mount, accessing files from both mounts will result in undefined behavior.\n");
	}

M
Miklos Szeredi 已提交
1103 1104
	ofs->workbasedir = dget(workpath.dentry);
	err = ovl_make_workdir(ofs, &workpath);
M
Miklos Szeredi 已提交
1105 1106 1107
	if (err)
		goto out;

1108 1109
	err = 0;
out:
M
Miklos Szeredi 已提交
1110 1111
	path_put(&workpath);

1112 1113 1114
	return err;
}

M
Miklos Szeredi 已提交
1115
static int ovl_get_indexdir(struct ovl_fs *ofs, struct ovl_entry *oe,
1116
			    struct path *upperpath)
1117
{
1118
	struct vfsmount *mnt = ofs->upper_mnt;
1119 1120
	int err;

1121 1122 1123 1124
	err = mnt_want_write(mnt);
	if (err)
		return err;

1125
	/* Verify lower root is upper root origin */
1126
	err = ovl_verify_origin(upperpath->dentry, oe->lowerstack[0].dentry,
1127
				true);
1128 1129 1130 1131 1132
	if (err) {
		pr_err("overlayfs: failed to verify upper root origin\n");
		goto out;
	}

M
Miklos Szeredi 已提交
1133 1134
	ofs->indexdir = ovl_workdir_create(ofs, OVL_INDEXDIR_NAME, true);
	if (ofs->indexdir) {
1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149
		/*
		 * Verify upper root is exclusively associated with index dir.
		 * Older kernels stored upper fh in "trusted.overlay.origin"
		 * xattr. If that xattr exists, verify that it is a match to
		 * upper dir file handle. In any case, verify or set xattr
		 * "trusted.overlay.upper" to indicate that index may have
		 * directory entries.
		 */
		if (ovl_check_origin_xattr(ofs->indexdir)) {
			err = ovl_verify_set_fh(ofs->indexdir, OVL_XATTR_ORIGIN,
						upperpath->dentry, true, false);
			if (err)
				pr_err("overlayfs: failed to verify index dir 'origin' xattr\n");
		}
		err = ovl_verify_upper(ofs->indexdir, upperpath->dentry, true);
1150
		if (err)
1151
			pr_err("overlayfs: failed to verify index dir 'upper' xattr\n");
1152 1153 1154

		/* Cleanup bad/stale/orphan index entries */
		if (!err)
1155
			err = ovl_indexdir_cleanup(ofs);
1156
	}
M
Miklos Szeredi 已提交
1157
	if (err || !ofs->indexdir)
1158 1159 1160
		pr_warn("overlayfs: try deleting index dir or mounting with '-o index=off' to disable inodes index.\n");

out:
1161
	mnt_drop_write(mnt);
1162 1163 1164
	return err;
}

1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193
/* Get a unique fsid for the layer */
static int ovl_get_fsid(struct ovl_fs *ofs, struct super_block *sb)
{
	unsigned int i;
	dev_t dev;
	int err;

	/* fsid 0 is reserved for upper fs even with non upper overlay */
	if (ofs->upper_mnt && ofs->upper_mnt->mnt_sb == sb)
		return 0;

	for (i = 0; i < ofs->numlowerfs; i++) {
		if (ofs->lower_fs[i].sb == sb)
			return i + 1;
	}

	err = get_anon_bdev(&dev);
	if (err) {
		pr_err("overlayfs: failed to get anonymous bdev for lowerpath\n");
		return err;
	}

	ofs->lower_fs[ofs->numlowerfs].sb = sb;
	ofs->lower_fs[ofs->numlowerfs].pseudo_dev = dev;
	ofs->numlowerfs++;

	return ofs->numlowerfs;
}

M
Miklos Szeredi 已提交
1194
static int ovl_get_lower_layers(struct ovl_fs *ofs, struct path *stack,
1195 1196 1197 1198 1199 1200
				unsigned int numlower)
{
	int err;
	unsigned int i;

	err = -ENOMEM;
M
Miklos Szeredi 已提交
1201
	ofs->lower_layers = kcalloc(numlower, sizeof(struct ovl_layer),
1202
				    GFP_KERNEL);
M
Miklos Szeredi 已提交
1203
	if (ofs->lower_layers == NULL)
1204
		goto out;
1205 1206 1207 1208 1209 1210

	ofs->lower_fs = kcalloc(numlower, sizeof(struct ovl_sb),
				GFP_KERNEL);
	if (ofs->lower_fs == NULL)
		goto out;

1211 1212
	for (i = 0; i < numlower; i++) {
		struct vfsmount *mnt;
1213
		int fsid;
1214

1215 1216
		err = fsid = ovl_get_fsid(ofs, stack[i].mnt->mnt_sb);
		if (err < 0)
1217 1218 1219 1220 1221 1222 1223 1224
			goto out;

		mnt = clone_private_mount(&stack[i]);
		err = PTR_ERR(mnt);
		if (IS_ERR(mnt)) {
			pr_err("overlayfs: failed to clone lowerpath\n");
			goto out;
		}
1225

1226 1227 1228 1229 1230 1231
		/*
		 * Make lower layers R/O.  That way fchmod/fchown on lower file
		 * will fail instead of modifying lower fs.
		 */
		mnt->mnt_flags |= MNT_READONLY | MNT_NOATIME;

M
Miklos Szeredi 已提交
1232
		ofs->lower_layers[ofs->numlower].mnt = mnt;
1233
		ofs->lower_layers[ofs->numlower].idx = i + 1;
1234 1235 1236 1237 1238
		ofs->lower_layers[ofs->numlower].fsid = fsid;
		if (fsid) {
			ofs->lower_layers[ofs->numlower].fs =
				&ofs->lower_fs[fsid - 1];
		}
M
Miklos Szeredi 已提交
1239
		ofs->numlower++;
1240
	}
1241

1242 1243 1244 1245 1246 1247 1248 1249 1250
	/*
	 * When all layers on same fs, overlay can use real inode numbers.
	 * With mount option "xino=on", mounter declares that there are enough
	 * free high bits in underlying fs to hold the unique fsid.
	 * If overlayfs does encounter underlying inodes using the high xino
	 * bits reserved for fsid, it emits a warning and uses the original
	 * inode number.
	 */
	if (!ofs->numlowerfs || (ofs->numlowerfs == 1 && !ofs->upper_mnt)) {
1251
		ofs->xino_bits = 0;
1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266
		ofs->config.xino = OVL_XINO_OFF;
	} else if (ofs->config.xino == OVL_XINO_ON && !ofs->xino_bits) {
		/*
		 * This is a roundup of number of bits needed for numlowerfs+1
		 * (i.e. ilog2(numlowerfs+1 - 1) + 1). fsid 0 is reserved for
		 * upper fs even with non upper overlay.
		 */
		BUILD_BUG_ON(ilog2(OVL_MAX_STACK) > 31);
		ofs->xino_bits = ilog2(ofs->numlowerfs) + 1;
	}

	if (ofs->xino_bits) {
		pr_info("overlayfs: \"xino\" feature enabled using %d upper inode bits.\n",
			ofs->xino_bits);
	}
1267

1268 1269 1270 1271 1272
	err = 0;
out:
	return err;
}

1273
static struct ovl_entry *ovl_get_lowerstack(struct super_block *sb,
M
Miklos Szeredi 已提交
1274
					    struct ovl_fs *ofs)
1275 1276 1277
{
	int err;
	char *lowertmp, *lower;
1278 1279
	struct path *stack = NULL;
	unsigned int stacklen, numlower = 0, i;
1280
	bool remote = false;
1281
	struct ovl_entry *oe;
1282 1283

	err = -ENOMEM;
M
Miklos Szeredi 已提交
1284
	lowertmp = kstrdup(ofs->config.lowerdir, GFP_KERNEL);
1285
	if (!lowertmp)
1286
		goto out_err;
1287 1288 1289 1290 1291 1292

	err = -EINVAL;
	stacklen = ovl_split_lowerdirs(lowertmp);
	if (stacklen > OVL_MAX_STACK) {
		pr_err("overlayfs: too many lower directories, limit is %d\n",
		       OVL_MAX_STACK);
1293
		goto out_err;
M
Miklos Szeredi 已提交
1294
	} else if (!ofs->config.upperdir && stacklen == 1) {
1295
		pr_err("overlayfs: at least 2 lowerdir are needed while upperdir nonexistent\n");
1296
		goto out_err;
1297 1298 1299 1300
	} else if (!ofs->config.upperdir && ofs->config.nfs_export &&
		   ofs->config.redirect_follow) {
		pr_warn("overlayfs: NFS export requires \"redirect_dir=nofollow\" on non-upper mount, falling back to nfs_export=off.\n");
		ofs->config.nfs_export = false;
1301 1302 1303 1304 1305
	}

	err = -ENOMEM;
	stack = kcalloc(stacklen, sizeof(struct path), GFP_KERNEL);
	if (!stack)
1306
		goto out_err;
1307 1308 1309 1310

	err = -EINVAL;
	lower = lowertmp;
	for (numlower = 0; numlower < stacklen; numlower++) {
M
Miklos Szeredi 已提交
1311
		err = ovl_lower_dir(lower, &stack[numlower], ofs,
1312 1313
				    &sb->s_stack_depth, &remote);
		if (err)
1314
			goto out_err;
1315 1316 1317 1318 1319 1320 1321 1322

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

	err = -EINVAL;
	sb->s_stack_depth++;
	if (sb->s_stack_depth > FILESYSTEM_MAX_STACK_DEPTH) {
		pr_err("overlayfs: maximum fs stacking depth exceeded\n");
1323
		goto out_err;
1324 1325
	}

M
Miklos Szeredi 已提交
1326
	err = ovl_get_lower_layers(ofs, stack, numlower);
1327 1328 1329 1330 1331 1332 1333 1334 1335 1336
	if (err)
		goto out_err;

	err = -ENOMEM;
	oe = ovl_alloc_entry(numlower);
	if (!oe)
		goto out_err;

	for (i = 0; i < numlower; i++) {
		oe->lowerstack[i].dentry = dget(stack[i].dentry);
M
Miklos Szeredi 已提交
1337
		oe->lowerstack[i].layer = &ofs->lower_layers[i];
1338
	}
1339 1340 1341 1342 1343 1344 1345 1346 1347 1348

	if (remote)
		sb->s_d_op = &ovl_reval_dentry_operations;
	else
		sb->s_d_op = &ovl_dentry_operations;

out:
	for (i = 0; i < numlower; i++)
		path_put(&stack[i]);
	kfree(stack);
1349 1350 1351 1352 1353 1354
	kfree(lowertmp);

	return oe;

out_err:
	oe = ERR_PTR(err);
1355 1356 1357
	goto out;
}

M
Miklos Szeredi 已提交
1358 1359
static int ovl_fill_super(struct super_block *sb, void *data, int silent)
{
K
Kees Cook 已提交
1360
	struct path upperpath = { };
M
Miklos Szeredi 已提交
1361
	struct dentry *root_dentry;
1362
	struct ovl_entry *oe;
M
Miklos Szeredi 已提交
1363
	struct ovl_fs *ofs;
1364
	struct cred *cred;
M
Miklos Szeredi 已提交
1365 1366
	int err;

E
Erez Zadok 已提交
1367
	err = -ENOMEM;
M
Miklos Szeredi 已提交
1368 1369
	ofs = kzalloc(sizeof(struct ovl_fs), GFP_KERNEL);
	if (!ofs)
M
Miklos Szeredi 已提交
1370 1371
		goto out;

M
Miklos Szeredi 已提交
1372
	ofs->creator_cred = cred = prepare_creds();
1373 1374 1375
	if (!cred)
		goto out_err;

M
Miklos Szeredi 已提交
1376
	ofs->config.index = ovl_index_def;
1377
	ofs->config.nfs_export = ovl_nfs_export_def;
1378
	ofs->config.xino = ovl_xino_def();
M
Miklos Szeredi 已提交
1379
	err = ovl_parse_opt((char *) data, &ofs->config);
E
Erez Zadok 已提交
1380
	if (err)
1381
		goto out_err;
E
Erez Zadok 已提交
1382

M
Miklos Szeredi 已提交
1383
	err = -EINVAL;
M
Miklos Szeredi 已提交
1384
	if (!ofs->config.lowerdir) {
1385 1386
		if (!silent)
			pr_err("overlayfs: missing 'lowerdir'\n");
1387
		goto out_err;
M
Miklos Szeredi 已提交
1388 1389
	}

M
Miklos Szeredi 已提交
1390
	sb->s_stack_depth = 0;
1391
	sb->s_maxbytes = MAX_LFS_FILESIZE;
1392
	/* Assume underlaying fs uses 32bit inodes unless proven otherwise */
1393 1394 1395
	if (ofs->config.xino != OVL_XINO_OFF)
		ofs->xino_bits = BITS_PER_LONG - 32;

M
Miklos Szeredi 已提交
1396 1397
	if (ofs->config.upperdir) {
		if (!ofs->config.workdir) {
M
Miklos Szeredi 已提交
1398
			pr_err("overlayfs: missing 'workdir'\n");
1399
			goto out_err;
M
Miklos Szeredi 已提交
1400
		}
M
Miklos Szeredi 已提交
1401

M
Miklos Szeredi 已提交
1402
		err = ovl_get_upper(ofs, &upperpath);
M
Miklos Szeredi 已提交
1403
		if (err)
1404
			goto out_err;
1405

M
Miklos Szeredi 已提交
1406
		err = ovl_get_workdir(ofs, &upperpath);
1407
		if (err)
1408
			goto out_err;
1409

M
Miklos Szeredi 已提交
1410
		if (!ofs->workdir)
1411
			sb->s_flags |= SB_RDONLY;
1412

M
Miklos Szeredi 已提交
1413 1414
		sb->s_stack_depth = ofs->upper_mnt->mnt_sb->s_stack_depth;
		sb->s_time_gran = ofs->upper_mnt->mnt_sb->s_time_gran;
1415

M
Miklos Szeredi 已提交
1416
	}
M
Miklos Szeredi 已提交
1417
	oe = ovl_get_lowerstack(sb, ofs);
1418 1419
	err = PTR_ERR(oe);
	if (IS_ERR(oe))
1420
		goto out_err;
M
Miklos Szeredi 已提交
1421

H
hujianyang 已提交
1422
	/* If the upper fs is nonexistent, we mark overlayfs r/o too */
M
Miklos Szeredi 已提交
1423
	if (!ofs->upper_mnt)
1424
		sb->s_flags |= SB_RDONLY;
M
Miklos Szeredi 已提交
1425

M
Miklos Szeredi 已提交
1426 1427
	if (!(ovl_force_readonly(ofs)) && ofs->config.index) {
		err = ovl_get_indexdir(ofs, oe, &upperpath);
1428
		if (err)
1429
			goto out_free_oe;
1430

1431 1432 1433 1434
		/* Force r/o mount with no index dir */
		if (!ofs->indexdir) {
			dput(ofs->workdir);
			ofs->workdir = NULL;
1435
			sb->s_flags |= SB_RDONLY;
1436 1437
		}

1438 1439
	}

1440
	/* Show index=off in /proc/mounts for forced r/o mount */
1441
	if (!ofs->indexdir) {
M
Miklos Szeredi 已提交
1442
		ofs->config.index = false;
1443 1444 1445 1446 1447
		if (ofs->upper_mnt && ofs->config.nfs_export) {
			pr_warn("overlayfs: NFS export requires an index dir, falling back to nfs_export=off.\n");
			ofs->config.nfs_export = false;
		}
	}
1448

1449 1450 1451
	if (ofs->config.nfs_export)
		sb->s_export_op = &ovl_export_operations;

1452 1453 1454
	/* Never override disk quota limits or use reserved space */
	cap_lower(cred->cap_effective, CAP_SYS_RESOURCE);

1455 1456 1457
	sb->s_magic = OVERLAYFS_SUPER_MAGIC;
	sb->s_op = &ovl_super_operations;
	sb->s_xattr = ovl_xattr_handlers;
M
Miklos Szeredi 已提交
1458
	sb->s_fs_info = ofs;
1459
	sb->s_flags |= SB_POSIXACL | SB_NOREMOTELOCK;
1460

1461
	err = -ENOMEM;
1462
	root_dentry = d_make_root(ovl_new_inode(sb, S_IFDIR, 0));
M
Miklos Szeredi 已提交
1463
	if (!root_dentry)
1464
		goto out_free_oe;
M
Miklos Szeredi 已提交
1465

1466 1467
	root_dentry->d_fsdata = oe;

M
Miklos Szeredi 已提交
1468
	mntput(upperpath.mnt);
1469
	if (upperpath.dentry) {
1470
		ovl_dentry_set_upper_alias(root_dentry);
M
Miklos Szeredi 已提交
1471 1472
		if (ovl_is_impuredir(upperpath.dentry))
			ovl_set_flag(OVL_IMPURE, d_inode(root_dentry));
1473
	}
M
Miklos Szeredi 已提交
1474

1475 1476
	/* Root is always merge -> can have whiteouts */
	ovl_set_flag(OVL_WHITEOUTS, d_inode(root_dentry));
1477
	ovl_dentry_set_flag(OVL_E_CONNECTED, root_dentry);
1478 1479
	ovl_inode_init(d_inode(root_dentry), upperpath.dentry,
		       ovl_dentry_lower(root_dentry));
M
Miklos Szeredi 已提交
1480

M
Miklos Szeredi 已提交
1481 1482 1483 1484
	sb->s_root = root_dentry;

	return 0;

1485 1486
out_free_oe:
	ovl_entry_stack_free(oe);
1487
	kfree(oe);
1488
out_err:
M
Miklos Szeredi 已提交
1489
	path_put(&upperpath);
M
Miklos Szeredi 已提交
1490
	ovl_free_fs(ofs);
M
Miklos Szeredi 已提交
1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502
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,
1503
	.name		= "overlay",
M
Miklos Szeredi 已提交
1504 1505 1506
	.mount		= ovl_mount,
	.kill_sb	= kill_anon_super,
};
1507
MODULE_ALIAS_FS("overlay");
M
Miklos Szeredi 已提交
1508

1509 1510 1511 1512 1513 1514 1515
static void ovl_inode_init_once(void *foo)
{
	struct ovl_inode *oi = foo;

	inode_init_once(&oi->vfs_inode);
}

M
Miklos Szeredi 已提交
1516 1517
static int __init ovl_init(void)
{
1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532
	int err;

	ovl_inode_cachep = kmem_cache_create("ovl_inode",
					     sizeof(struct ovl_inode), 0,
					     (SLAB_RECLAIM_ACCOUNT|
					      SLAB_MEM_SPREAD|SLAB_ACCOUNT),
					     ovl_inode_init_once);
	if (ovl_inode_cachep == NULL)
		return -ENOMEM;

	err = register_filesystem(&ovl_fs_type);
	if (err)
		kmem_cache_destroy(ovl_inode_cachep);

	return err;
M
Miklos Szeredi 已提交
1533 1534 1535 1536 1537
}

static void __exit ovl_exit(void)
{
	unregister_filesystem(&ovl_fs_type);
1538 1539 1540 1541 1542 1543 1544 1545

	/*
	 * Make sure all delayed rcu free inodes are flushed before we
	 * destroy cache.
	 */
	rcu_barrier();
	kmem_cache_destroy(ovl_inode_cachep);

M
Miklos Szeredi 已提交
1546 1547 1548 1549
}

module_init(ovl_init);
module_exit(ovl_exit);