[mirror_ubuntu-zesty-kernel.git] / fs / overlayfs / readdir.c

/*
 *
 * 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/slab.h>
#include <linux/namei.h>
#include <linux/file.h>
#include <linux/xattr.h>
#include <linux/rbtree.h>
#include <linux/security.h>
#include <linux/cred.h>
#include "overlayfs.h"

struct ovl_cache_entry {
	unsigned int len;
	unsigned int type;
	u64 ino;
	struct list_head l_node;
	struct rb_node node;
	struct ovl_cache_entry *next_maybe_whiteout;
	bool is_whiteout;
	char name[];
};

struct ovl_dir_cache {
	long refcount;
	u64 version;
	struct list_head entries;
};

struct ovl_readdir_data {
	struct dir_context ctx;
	bool is_merge;
	struct rb_root root;
	struct list_head *list;
	struct list_head middle;
	struct ovl_cache_entry *first_maybe_whiteout;
	int count;
	int err;
};

struct ovl_dir_file {
	bool is_real;
	bool is_upper;
	struct ovl_dir_cache *cache;
	struct list_head *cursor;
	struct file *realfile;
	struct file *upperfile;
};

static struct ovl_cache_entry *ovl_cache_entry_from_node(struct rb_node *n)
{
	return container_of(n, struct ovl_cache_entry, node);
}

static struct ovl_cache_entry *ovl_cache_entry_find(struct rb_root *root,
						    const char *name, int len)
{
	struct rb_node *node = root->rb_node;
	int cmp;

	while (node) {
		struct ovl_cache_entry *p = ovl_cache_entry_from_node(node);

		cmp = strncmp(name, p->name, len);
		if (cmp > 0)
			node = p->node.rb_right;
		else if (cmp < 0 || len < p->len)
			node = p->node.rb_left;
		else
			return p;
	}

	return NULL;
}

static struct ovl_cache_entry *ovl_cache_entry_new(struct ovl_readdir_data *rdd,
						   const char *name, int len,
						   u64 ino, unsigned int d_type)
{
	struct ovl_cache_entry *p;
	size_t size = offsetof(struct ovl_cache_entry, name[len + 1]);

	p = kmalloc(size, GFP_KERNEL);
	if (!p)
		return NULL;

	memcpy(p->name, name, len);
	p->name[len] = '\0';
	p->len = len;
	p->type = d_type;
	p->ino = ino;
	p->is_whiteout = false;

	if (d_type == DT_CHR) {
		p->next_maybe_whiteout = rdd->first_maybe_whiteout;
		rdd->first_maybe_whiteout = p;
	}
	return p;
}

static int ovl_cache_entry_add_rb(struct ovl_readdir_data *rdd,
				  const char *name, int len, u64 ino,
				  unsigned int d_type)
{
	struct rb_node **newp = &rdd->root.rb_node;
	struct rb_node *parent = NULL;
	struct ovl_cache_entry *p;

	while (*newp) {
		int cmp;
		struct ovl_cache_entry *tmp;

		parent = *newp;
		tmp = ovl_cache_entry_from_node(*newp);
		cmp = strncmp(name, tmp->name, len);
		if (cmp > 0)
			newp = &tmp->node.rb_right;
		else if (cmp < 0 || len < tmp->len)
			newp = &tmp->node.rb_left;
		else
			return 0;
	}

	p = ovl_cache_entry_new(rdd, name, len, ino, d_type);
	if (p == NULL)
		return -ENOMEM;

	list_add_tail(&p->l_node, rdd->list);
	rb_link_node(&p->node, parent, newp);
	rb_insert_color(&p->node, &rdd->root);

	return 0;
}

static int ovl_fill_lower(struct ovl_readdir_data *rdd,
			  const char *name, int namelen,
			  loff_t offset, u64 ino, unsigned int d_type)
{
	struct ovl_cache_entry *p;

	p = ovl_cache_entry_find(&rdd->root, name, namelen);
	if (p) {
		list_move_tail(&p->l_node, &rdd->middle);
	} else {
		p = ovl_cache_entry_new(rdd, name, namelen, ino, d_type);
		if (p == NULL)
			rdd->err = -ENOMEM;
		else
			list_add_tail(&p->l_node, &rdd->middle);
	}

	return rdd->err;
}

void ovl_cache_free(struct list_head *list)
{
	struct ovl_cache_entry *p;
	struct ovl_cache_entry *n;

	list_for_each_entry_safe(p, n, list, l_node)
		kfree(p);

	INIT_LIST_HEAD(list);
}

static void ovl_cache_put(struct ovl_dir_file *od, struct dentry *dentry)
{
	struct ovl_dir_cache *cache = od->cache;

	WARN_ON(cache->refcount <= 0);
	cache->refcount--;
	if (!cache->refcount) {
		if (ovl_dir_cache(dentry) == cache)
			ovl_set_dir_cache(dentry, NULL);

		ovl_cache_free(&cache->entries);
		kfree(cache);
	}
}

static int ovl_fill_merge(struct dir_context *ctx, const char *name,
			  int namelen, loff_t offset, u64 ino,
			  unsigned int d_type)
{
	struct ovl_readdir_data *rdd =
		container_of(ctx, struct ovl_readdir_data, ctx);

	rdd->count++;
	if (!rdd->is_merge)
		return ovl_cache_entry_add_rb(rdd, name, namelen, ino, d_type);
	else
		return ovl_fill_lower(rdd, name, namelen, offset, ino, d_type);
}

static int ovl_check_whiteouts(struct dentry *dir, struct ovl_readdir_data *rdd)
{
	int err;
	struct ovl_cache_entry *p;
	struct dentry *dentry;
	const struct cred *old_cred;
	struct cred *override_cred;

	override_cred = prepare_creds();
	if (!override_cred)
		return -ENOMEM;

	/*
	 * CAP_DAC_OVERRIDE for lookup
	 */
	cap_raise(override_cred->cap_effective, CAP_DAC_OVERRIDE);
	old_cred = override_creds(override_cred);

	err = mutex_lock_killable(&dir->d_inode->i_mutex);
	if (!err) {
		while (rdd->first_maybe_whiteout) {
			p = rdd->first_maybe_whiteout;
			rdd->first_maybe_whiteout = p->next_maybe_whiteout;
			dentry = lookup_one_len(p->name, dir, p->len);
			if (!IS_ERR(dentry)) {
				p->is_whiteout = ovl_is_whiteout(dentry);
				dput(dentry);
			}
		}
		inode_unlock(dir->d_inode);
	}
	revert_creds(old_cred);
	put_cred(override_cred);

	return err;
}

static inline int ovl_dir_read(struct path *realpath,
			       struct ovl_readdir_data *rdd)
{
	struct file *realfile;
	int err;

	realfile = ovl_path_open(realpath, O_RDONLY | O_DIRECTORY);
	if (IS_ERR(realfile))
		return PTR_ERR(realfile);

	rdd->first_maybe_whiteout = NULL;
	rdd->ctx.pos = 0;
	do {
		rdd->count = 0;
		rdd->err = 0;
		err = iterate_dir(realfile, &rdd->ctx);
		if (err >= 0)
			err = rdd->err;
	} while (!err && rdd->count);

	if (!err && rdd->first_maybe_whiteout)
		err = ovl_check_whiteouts(realpath->dentry, rdd);

	fput(realfile);

	return err;
}

static void ovl_dir_reset(struct file *file)
{
	struct ovl_dir_file *od = file->private_data;
	struct ovl_dir_cache *cache = od->cache;
	struct dentry *dentry = file->f_path.dentry;
	enum ovl_path_type type = ovl_path_type(dentry);

	if (cache && ovl_dentry_version_get(dentry) != cache->version) {
		ovl_cache_put(od, dentry);
		od->cache = NULL;
		od->cursor = NULL;
	}
	WARN_ON(!od->is_real && !OVL_TYPE_MERGE(type));
	if (od->is_real && OVL_TYPE_MERGE(type))
		od->is_real = false;
}

static int ovl_dir_read_merged(struct dentry *dentry, struct list_head *list)
{
	int err;
	struct path realpath;
	struct ovl_readdir_data rdd = {
		.ctx.actor = ovl_fill_merge,
		.list = list,
		.root = RB_ROOT,
		.is_merge = false,
	};
	int idx, next;

	for (idx = 0; idx != -1; idx = next) {
		next = ovl_path_next(idx, dentry, &realpath);

		if (next != -1) {
			err = ovl_dir_read(&realpath, &rdd);
			if (err)
				break;
		} else {
			/*
			 * Insert lowest layer entries before upper ones, this
			 * allows offsets to be reasonably constant
			 */
			list_add(&rdd.middle, rdd.list);
			rdd.is_merge = true;
			err = ovl_dir_read(&realpath, &rdd);
			list_del(&rdd.middle);
		}
	}
	return err;
}

static void ovl_seek_cursor(struct ovl_dir_file *od, loff_t pos)
{
	struct list_head *p;
	loff_t off = 0;

	list_for_each(p, &od->cache->entries) {
		if (off >= pos)
			break;
		off++;
	}
	/* Cursor is safe since the cache is stable */
	od->cursor = p;
}

static struct ovl_dir_cache *ovl_cache_get(struct dentry *dentry)
{
	int res;
	struct ovl_dir_cache *cache;

	cache = ovl_dir_cache(dentry);
	if (cache && ovl_dentry_version_get(dentry) == cache->version) {
		cache->refcount++;
		return cache;
	}
	ovl_set_dir_cache(dentry, NULL);

	cache = kzalloc(sizeof(struct ovl_dir_cache), GFP_KERNEL);
	if (!cache)
		return ERR_PTR(-ENOMEM);

	cache->refcount = 1;
	INIT_LIST_HEAD(&cache->entries);

	res = ovl_dir_read_merged(dentry, &cache->entries);
	if (res) {
		ovl_cache_free(&cache->entries);
		kfree(cache);
		return ERR_PTR(res);
	}

	cache->version = ovl_dentry_version_get(dentry);
	ovl_set_dir_cache(dentry, cache);

	return cache;
}

static int ovl_iterate(struct file *file, struct dir_context *ctx)
{
	struct ovl_dir_file *od = file->private_data;
	struct dentry *dentry = file->f_path.dentry;
	struct ovl_cache_entry *p;

	if (!ctx->pos)
		ovl_dir_reset(file);

	if (od->is_real)
		return iterate_dir(od->realfile, ctx);

	if (!od->cache) {
		struct ovl_dir_cache *cache;

		cache = ovl_cache_get(dentry);
		if (IS_ERR(cache))
			return PTR_ERR(cache);

		od->cache = cache;
		ovl_seek_cursor(od, ctx->pos);
	}

	while (od->cursor != &od->cache->entries) {
		p = list_entry(od->cursor, struct ovl_cache_entry, l_node);
		if (!p->is_whiteout)
			if (!dir_emit(ctx, p->name, p->len, p->ino, p->type))
				break;
		od->cursor = p->l_node.next;
		ctx->pos++;
	}
	return 0;
}

static loff_t ovl_dir_llseek(struct file *file, loff_t offset, int origin)
{
	loff_t res;
	struct ovl_dir_file *od = file->private_data;

	inode_lock(file_inode(file));
	if (!file->f_pos)
		ovl_dir_reset(file);

	if (od->is_real) {
		res = vfs_llseek(od->realfile, offset, origin);
		file->f_pos = od->realfile->f_pos;
	} else {
		res = -EINVAL;

		switch (origin) {
		case SEEK_CUR:
			offset += file->f_pos;
			break;
		case SEEK_SET:
			break;
		default:
			goto out_unlock;
		}
		if (offset < 0)
			goto out_unlock;

		if (offset != file->f_pos) {
			file->f_pos = offset;
			if (od->cache)
				ovl_seek_cursor(od, offset);
		}
		res = offset;
	}
out_unlock:
	inode_unlock(file_inode(file));

	return res;
}

static int ovl_dir_fsync(struct file *file, loff_t start, loff_t end,
			 int datasync)
{
	struct ovl_dir_file *od = file->private_data;
	struct dentry *dentry = file->f_path.dentry;
	struct file *realfile = od->realfile;

	/*
	 * Need to check if we started out being a lower dir, but got copied up
	 */
	if (!od->is_upper && OVL_TYPE_UPPER(ovl_path_type(dentry))) {
		struct inode *inode = file_inode(file);

		realfile = lockless_dereference(od->upperfile);
		if (!realfile) {
			struct path upperpath;

			ovl_path_upper(dentry, &upperpath);
			realfile = ovl_path_open(&upperpath, O_RDONLY);
			smp_mb__before_spinlock();
			inode_lock(inode);
			if (!od->upperfile) {
				if (IS_ERR(realfile)) {
					inode_unlock(inode);
					return PTR_ERR(realfile);
				}
				od->upperfile = realfile;
			} else {
				/* somebody has beaten us to it */
				if (!IS_ERR(realfile))
					fput(realfile);
				realfile = od->upperfile;
			}
			inode_unlock(inode);
		}
	}

	return vfs_fsync_range(realfile, start, end, datasync);
}

static int ovl_dir_release(struct inode *inode, struct file *file)
{
	struct ovl_dir_file *od = file->private_data;

	if (od->cache) {
		inode_lock(inode);
		ovl_cache_put(od, file->f_path.dentry);
		inode_unlock(inode);
	}
	fput(od->realfile);
	if (od->upperfile)
		fput(od->upperfile);
	kfree(od);

	return 0;
}

static int ovl_dir_open(struct inode *inode, struct file *file)
{
	struct path realpath;
	struct file *realfile;
	struct ovl_dir_file *od;
	enum ovl_path_type type;

	od = kzalloc(sizeof(struct ovl_dir_file), GFP_KERNEL);
	if (!od)
		return -ENOMEM;

	type = ovl_path_real(file->f_path.dentry, &realpath);
	realfile = ovl_path_open(&realpath, file->f_flags);
	if (IS_ERR(realfile)) {
		kfree(od);
		return PTR_ERR(realfile);
	}
	od->realfile = realfile;
	od->is_real = !OVL_TYPE_MERGE(type);
	od->is_upper = OVL_TYPE_UPPER(type);
	file->private_data = od;

	return 0;
}

const struct file_operations ovl_dir_operations = {
	.read		= generic_read_dir,
	.open		= ovl_dir_open,
	.iterate	= ovl_iterate,
	.llseek		= ovl_dir_llseek,
	.fsync		= ovl_dir_fsync,
	.release	= ovl_dir_release,
};

int ovl_check_empty_dir(struct dentry *dentry, struct list_head *list)
{
	int err;
	struct ovl_cache_entry *p;

	err = ovl_dir_read_merged(dentry, list);
	if (err)
		return err;

	err = 0;

	list_for_each_entry(p, list, l_node) {
		if (p->is_whiteout)
			continue;

		if (p->name[0] == '.') {
			if (p->len == 1)
				continue;
			if (p->len == 2 && p->name[1] == '.')
				continue;
		}
		err = -ENOTEMPTY;
		break;
	}

	return err;
}

void ovl_cleanup_whiteouts(struct dentry *upper, struct list_head *list)
{
	struct ovl_cache_entry *p;

	inode_lock_nested(upper->d_inode, I_MUTEX_CHILD);
	list_for_each_entry(p, list, l_node) {
		struct dentry *dentry;

		if (!p->is_whiteout)
			continue;

		dentry = lookup_one_len(p->name, upper, p->len);
		if (IS_ERR(dentry)) {
			pr_err("overlayfs: lookup '%s/%.*s' failed (%i)\n",
			       upper->d_name.name, p->len, p->name,
			       (int) PTR_ERR(dentry));
			continue;
		}
		if (dentry->d_inode)
			ovl_cleanup(upper->d_inode, dentry);
		dput(dentry);
	}
	inode_unlock(upper->d_inode);
}
Commit	Line	Data
e9be9d5e MS	1	/*
	2	*
	3	* Copyright (C) 2011 Novell Inc.
	4	*
	5	* This program is free software; you can redistribute it and/or modify it
	6	* under the terms of the GNU General Public License version 2 as published by
	7	* the Free Software Foundation.
	8	*/
	9
	10	#include <linux/fs.h>
	11	#include <linux/slab.h>
	12	#include <linux/namei.h>
	13	#include <linux/file.h>
	14	#include <linux/xattr.h>
	15	#include <linux/rbtree.h>
	16	#include <linux/security.h>
	17	#include <linux/cred.h>
	18	#include "overlayfs.h"
	19
	20	struct ovl_cache_entry {
e9be9d5e MS	21	unsigned int len;
	22	unsigned int type;
	23	u64 ino;
e9be9d5e MS	24	struct list_head l_node;
e9be9d5e MS	25	struct rb_node node;
cdb67279	26	struct ovl_cache_entry *next_maybe_whiteout;
c2096537	27	bool is_whiteout;
68bf8611	28	char name[];
e9be9d5e MS	29	};
	30
	31	struct ovl_dir_cache {
	32	long refcount;
	33	u64 version;
	34	struct list_head entries;
	35	};
	36
	37	struct ovl_readdir_data {
	38	struct dir_context ctx;
	39	bool is_merge;
49be4fb9	40	struct rb_root root;
e9be9d5e	41	struct list_head *list;
db6ec212	42	struct list_head middle;
cdb67279	43	struct ovl_cache_entry *first_maybe_whiteout;
e9be9d5e MS	44	int count;
	45	int err;
	46	};
	47
	48	struct ovl_dir_file {
	49	bool is_real;
	50	bool is_upper;
	51	struct ovl_dir_cache *cache;
4330397e	52	struct list_head *cursor;
e9be9d5e MS	53	struct file *realfile;
	54	struct file *upperfile;
	55	};
	56
	57	static struct ovl_cache_entry ovl_cache_entry_from_node(struct rb_node n)
	58	{
	59	return container_of(n, struct ovl_cache_entry, node);
	60	}
	61
	62	static struct ovl_cache_entry ovl_cache_entry_find(struct rb_root root,
	63	const char *name, int len)
	64	{
	65	struct rb_node *node = root->rb_node;
	66	int cmp;
	67
	68	while (node) {
	69	struct ovl_cache_entry *p = ovl_cache_entry_from_node(node);
	70
	71	cmp = strncmp(name, p->name, len);
	72	if (cmp > 0)
	73	node = p->node.rb_right;
	74	else if (cmp < 0 \|\| len < p->len)
	75	node = p->node.rb_left;
	76	else
	77	return p;
	78	}
	79
	80	return NULL;
	81	}
	82
cdb67279	83	static struct ovl_cache_entry ovl_cache_entry_new(struct ovl_readdir_data rdd,
3e01cee3	84	const char *name, int len,
e9be9d5e MS	85	u64 ino, unsigned int d_type)
	86	{
	87	struct ovl_cache_entry *p;
68bf8611	88	size_t size = offsetof(struct ovl_cache_entry, name[len + 1]);
e9be9d5e	89
68bf8611	90	p = kmalloc(size, GFP_KERNEL);
3e01cee3 MS	91	if (!p)
	92	return NULL;
	93
	94	memcpy(p->name, name, len);
	95	p->name[len] = '\0';
	96	p->len = len;
	97	p->type = d_type;
	98	p->ino = ino;
	99	p->is_whiteout = false;
3e01cee3 MS	100
3e01cee3 MS	101	if (d_type == DT_CHR) {
cdb67279 MS	102	p->next_maybe_whiteout = rdd->first_maybe_whiteout;
cdb67279 MS	103	rdd->first_maybe_whiteout = p;
3e01cee3	104	}
e9be9d5e MS	105	return p;
	106	}
	107
	108	static int ovl_cache_entry_add_rb(struct ovl_readdir_data *rdd,
	109	const char *name, int len, u64 ino,
	110	unsigned int d_type)
	111	{
49be4fb9	112	struct rb_node **newp = &rdd->root.rb_node;
e9be9d5e MS	113	struct rb_node *parent = NULL;
	114	struct ovl_cache_entry *p;
	115
	116	while (*newp) {
	117	int cmp;
	118	struct ovl_cache_entry *tmp;
	119
	120	parent = *newp;
	121	tmp = ovl_cache_entry_from_node(*newp);
	122	cmp = strncmp(name, tmp->name, len);
	123	if (cmp > 0)
	124	newp = &tmp->node.rb_right;
	125	else if (cmp < 0 \|\| len < tmp->len)
	126	newp = &tmp->node.rb_left;
	127	else
	128	return 0;
	129	}
	130
cdb67279	131	p = ovl_cache_entry_new(rdd, name, len, ino, d_type);
e9be9d5e MS	132	if (p == NULL)
	133	return -ENOMEM;
	134
	135	list_add_tail(&p->l_node, rdd->list);
	136	rb_link_node(&p->node, parent, newp);
49be4fb9	137	rb_insert_color(&p->node, &rdd->root);
e9be9d5e MS	138
	139	return 0;
	140	}
	141
	142	static int ovl_fill_lower(struct ovl_readdir_data *rdd,
	143	const char *name, int namelen,
	144	loff_t offset, u64 ino, unsigned int d_type)
	145	{
	146	struct ovl_cache_entry *p;
	147
49be4fb9	148	p = ovl_cache_entry_find(&rdd->root, name, namelen);
e9be9d5e	149	if (p) {
db6ec212	150	list_move_tail(&p->l_node, &rdd->middle);
e9be9d5e	151	} else {
cdb67279	152	p = ovl_cache_entry_new(rdd, name, namelen, ino, d_type);
e9be9d5e MS	153	if (p == NULL)
	154	rdd->err = -ENOMEM;
	155	else
db6ec212	156	list_add_tail(&p->l_node, &rdd->middle);
e9be9d5e MS	157	}
	158
	159	return rdd->err;
	160	}
	161
	162	void ovl_cache_free(struct list_head *list)
	163	{
	164	struct ovl_cache_entry *p;
	165	struct ovl_cache_entry *n;
	166
	167	list_for_each_entry_safe(p, n, list, l_node)
	168	kfree(p);
	169
	170	INIT_LIST_HEAD(list);
	171	}
	172
	173	static void ovl_cache_put(struct ovl_dir_file od, struct dentry dentry)
	174	{
	175	struct ovl_dir_cache *cache = od->cache;
	176
e9be9d5e MS	177	WARN_ON(cache->refcount <= 0);
	178	cache->refcount--;
	179	if (!cache->refcount) {
	180	if (ovl_dir_cache(dentry) == cache)
	181	ovl_set_dir_cache(dentry, NULL);
	182
	183	ovl_cache_free(&cache->entries);
	184	kfree(cache);
	185	}
	186	}
	187
ac7576f4 MS	188	static int ovl_fill_merge(struct dir_context ctx, const char name,
	189	int namelen, loff_t offset, u64 ino,
	190	unsigned int d_type)
e9be9d5e	191	{
ac7576f4 MS	192	struct ovl_readdir_data *rdd =
ac7576f4 MS	193	container_of(ctx, struct ovl_readdir_data, ctx);
e9be9d5e MS	194
	195	rdd->count++;
	196	if (!rdd->is_merge)
	197	return ovl_cache_entry_add_rb(rdd, name, namelen, ino, d_type);
	198	else
	199	return ovl_fill_lower(rdd, name, namelen, offset, ino, d_type);
	200	}
	201
cdb67279 MS	202	static int ovl_check_whiteouts(struct dentry dir, struct ovl_readdir_data rdd)
	203	{
	204	int err;
	205	struct ovl_cache_entry *p;
	206	struct dentry *dentry;
	207	const struct cred *old_cred;
	208	struct cred *override_cred;
	209
	210	override_cred = prepare_creds();
	211	if (!override_cred)
	212	return -ENOMEM;
	213
	214	/*
	215	* CAP_DAC_OVERRIDE for lookup
	216	*/
	217	cap_raise(override_cred->cap_effective, CAP_DAC_OVERRIDE);
	218	old_cred = override_creds(override_cred);
	219
	220	err = mutex_lock_killable(&dir->d_inode->i_mutex);
	221	if (!err) {
	222	while (rdd->first_maybe_whiteout) {
	223	p = rdd->first_maybe_whiteout;
	224	rdd->first_maybe_whiteout = p->next_maybe_whiteout;
	225	dentry = lookup_one_len(p->name, dir, p->len);
	226	if (!IS_ERR(dentry)) {
	227	p->is_whiteout = ovl_is_whiteout(dentry);
	228	dput(dentry);
	229	}
	230	}
5955102c	231	inode_unlock(dir->d_inode);
cdb67279 MS	232	}
	233	revert_creds(old_cred);
	234	put_cred(override_cred);
	235
	236	return err;
	237	}
	238
e9be9d5e MS	239	static inline int ovl_dir_read(struct path *realpath,
	240	struct ovl_readdir_data *rdd)
	241	{
	242	struct file *realfile;
	243	int err;
	244
	245	realfile = ovl_path_open(realpath, O_RDONLY \| O_DIRECTORY);
	246	if (IS_ERR(realfile))
	247	return PTR_ERR(realfile);
	248
cdb67279	249	rdd->first_maybe_whiteout = NULL;
e9be9d5e MS	250	rdd->ctx.pos = 0;
	251	do {
	252	rdd->count = 0;
	253	rdd->err = 0;
	254	err = iterate_dir(realfile, &rdd->ctx);
	255	if (err >= 0)
	256	err = rdd->err;
	257	} while (!err && rdd->count);
cdb67279 MS	258
	259	if (!err && rdd->first_maybe_whiteout)
	260	err = ovl_check_whiteouts(realpath->dentry, rdd);
	261
e9be9d5e MS	262	fput(realfile);
	263
	264	return err;
	265	}
	266
	267	static void ovl_dir_reset(struct file *file)
	268	{
	269	struct ovl_dir_file *od = file->private_data;
	270	struct ovl_dir_cache *cache = od->cache;
	271	struct dentry *dentry = file->f_path.dentry;
	272	enum ovl_path_type type = ovl_path_type(dentry);
	273
	274	if (cache && ovl_dentry_version_get(dentry) != cache->version) {
	275	ovl_cache_put(od, dentry);
	276	od->cache = NULL;
4330397e	277	od->cursor = NULL;
e9be9d5e	278	}
1afaba1e MS	279	WARN_ON(!od->is_real && !OVL_TYPE_MERGE(type));
1afaba1e MS	280	if (od->is_real && OVL_TYPE_MERGE(type))
e9be9d5e MS	281	od->is_real = false;
	282	}
	283
c9f00fdb	284	static int ovl_dir_read_merged(struct dentry dentry, struct list_head list)
e9be9d5e MS	285	{
e9be9d5e MS	286	int err;
9d7459d8	287	struct path realpath;
e9be9d5e MS	288	struct ovl_readdir_data rdd = {
	289	.ctx.actor = ovl_fill_merge,
	290	.list = list,
49be4fb9	291	.root = RB_ROOT,
e9be9d5e MS	292	.is_merge = false,
e9be9d5e MS	293	};
9d7459d8	294	int idx, next;
e9be9d5e	295
9d7459d8 MS	296	for (idx = 0; idx != -1; idx = next) {
9d7459d8 MS	297	next = ovl_path_next(idx, dentry, &realpath);
c9f00fdb	298
9d7459d8	299	if (next != -1) {
9d7459d8	300	err = ovl_dir_read(&realpath, &rdd);
e9be9d5e	301	if (err)
9d7459d8 MS	302	break;
	303	} else {
	304	/*
	305	* Insert lowest layer entries before upper ones, this
	306	* allows offsets to be reasonably constant
	307	*/
	308	list_add(&rdd.middle, rdd.list);
	309	rdd.is_merge = true;
	310	err = ovl_dir_read(&realpath, &rdd);
	311	list_del(&rdd.middle);
e9be9d5e MS	312	}
e9be9d5e MS	313	}
e9be9d5e	314	return err;
e9be9d5e MS	315	}
	316
	317	static void ovl_seek_cursor(struct ovl_dir_file *od, loff_t pos)
	318	{
4330397e	319	struct list_head *p;
e9be9d5e MS	320	loff_t off = 0;
e9be9d5e MS	321
4330397e	322	list_for_each(p, &od->cache->entries) {
e9be9d5e MS	323	if (off >= pos)
	324	break;
	325	off++;
	326	}
4330397e	327	/* Cursor is safe since the cache is stable */
4330397e	328	od->cursor = p;
e9be9d5e MS	329	}
	330
	331	static struct ovl_dir_cache ovl_cache_get(struct dentry dentry)
	332	{
	333	int res;
e9be9d5e MS	334	struct ovl_dir_cache *cache;
	335
	336	cache = ovl_dir_cache(dentry);
	337	if (cache && ovl_dentry_version_get(dentry) == cache->version) {
	338	cache->refcount++;
	339	return cache;
	340	}
	341	ovl_set_dir_cache(dentry, NULL);
	342
	343	cache = kzalloc(sizeof(struct ovl_dir_cache), GFP_KERNEL);
	344	if (!cache)
	345	return ERR_PTR(-ENOMEM);
	346
	347	cache->refcount = 1;
	348	INIT_LIST_HEAD(&cache->entries);
	349
c9f00fdb	350	res = ovl_dir_read_merged(dentry, &cache->entries);
e9be9d5e MS	351	if (res) {
	352	ovl_cache_free(&cache->entries);
	353	kfree(cache);
	354	return ERR_PTR(res);
	355	}
	356
	357	cache->version = ovl_dentry_version_get(dentry);
	358	ovl_set_dir_cache(dentry, cache);
	359
	360	return cache;
	361	}
	362
	363	static int ovl_iterate(struct file file, struct dir_context ctx)
	364	{
	365	struct ovl_dir_file *od = file->private_data;
	366	struct dentry *dentry = file->f_path.dentry;
4330397e	367	struct ovl_cache_entry *p;
e9be9d5e MS	368
	369	if (!ctx->pos)
	370	ovl_dir_reset(file);
	371
	372	if (od->is_real)
	373	return iterate_dir(od->realfile, ctx);
	374
	375	if (!od->cache) {
	376	struct ovl_dir_cache *cache;
	377
	378	cache = ovl_cache_get(dentry);
	379	if (IS_ERR(cache))
	380	return PTR_ERR(cache);
	381
	382	od->cache = cache;
	383	ovl_seek_cursor(od, ctx->pos);
	384	}
	385
4330397e	386	while (od->cursor != &od->cache->entries) {
	387	p = list_entry(od->cursor, struct ovl_cache_entry, l_node);
	388	if (!p->is_whiteout)
	389	if (!dir_emit(ctx, p->name, p->len, p->ino, p->type))
	390	break;
	391	od->cursor = p->l_node.next;
	392	ctx->pos++;
e9be9d5e MS	393	}
	394	return 0;
	395	}
	396
	397	static loff_t ovl_dir_llseek(struct file *file, loff_t offset, int origin)
	398	{
	399	loff_t res;
	400	struct ovl_dir_file *od = file->private_data;
	401
5955102c	402	inode_lock(file_inode(file));
e9be9d5e MS	403	if (!file->f_pos)
	404	ovl_dir_reset(file);
	405
	406	if (od->is_real) {
	407	res = vfs_llseek(od->realfile, offset, origin);
	408	file->f_pos = od->realfile->f_pos;
	409	} else {
	410	res = -EINVAL;
	411
	412	switch (origin) {
	413	case SEEK_CUR:
	414	offset += file->f_pos;
	415	break;
	416	case SEEK_SET:
	417	break;
	418	default:
	419	goto out_unlock;
	420	}
	421	if (offset < 0)
	422	goto out_unlock;
	423
	424	if (offset != file->f_pos) {
	425	file->f_pos = offset;
	426	if (od->cache)
	427	ovl_seek_cursor(od, offset);
	428	}
	429	res = offset;
	430	}
	431	out_unlock:
5955102c	432	inode_unlock(file_inode(file));
e9be9d5e MS	433
	434	return res;
	435	}
	436
	437	static int ovl_dir_fsync(struct file *file, loff_t start, loff_t end,
	438	int datasync)
	439	{
	440	struct ovl_dir_file *od = file->private_data;
	441	struct dentry *dentry = file->f_path.dentry;
	442	struct file *realfile = od->realfile;
	443
	444	/*
	445	* Need to check if we started out being a lower dir, but got copied up
	446	*/
1afaba1e	447	if (!od->is_upper && OVL_TYPE_UPPER(ovl_path_type(dentry))) {
e9be9d5e MS	448	struct inode *inode = file_inode(file);
e9be9d5e MS	449
7676895f	450	realfile = lockless_dereference(od->upperfile);
e9be9d5e MS	451	if (!realfile) {
	452	struct path upperpath;
	453
	454	ovl_path_upper(dentry, &upperpath);
	455	realfile = ovl_path_open(&upperpath, O_RDONLY);
d45f00ae	456	smp_mb__before_spinlock();
5955102c	457	inode_lock(inode);
3d268c9b AV	458	if (!od->upperfile) {
3d268c9b AV	459	if (IS_ERR(realfile)) {
5955102c	460	inode_unlock(inode);
3d268c9b AV	461	return PTR_ERR(realfile);
	462	}
	463	od->upperfile = realfile;
	464	} else {
	465	/* somebody has beaten us to it */
	466	if (!IS_ERR(realfile))
	467	fput(realfile);
	468	realfile = od->upperfile;
e9be9d5e	469	}
5955102c	470	inode_unlock(inode);
e9be9d5e	471	}
e9be9d5e MS	472	}
	473
	474	return vfs_fsync_range(realfile, start, end, datasync);
	475	}
	476
	477	static int ovl_dir_release(struct inode inode, struct file file)
	478	{
	479	struct ovl_dir_file *od = file->private_data;
	480
	481	if (od->cache) {
5955102c	482	inode_lock(inode);
e9be9d5e	483	ovl_cache_put(od, file->f_path.dentry);
5955102c	484	inode_unlock(inode);
e9be9d5e MS	485	}
	486	fput(od->realfile);
	487	if (od->upperfile)
	488	fput(od->upperfile);
	489	kfree(od);
	490
	491	return 0;
	492	}
	493
	494	static int ovl_dir_open(struct inode inode, struct file file)
	495	{
	496	struct path realpath;
	497	struct file *realfile;
	498	struct ovl_dir_file *od;
	499	enum ovl_path_type type;
	500
	501	od = kzalloc(sizeof(struct ovl_dir_file), GFP_KERNEL);
	502	if (!od)
	503	return -ENOMEM;
	504
	505	type = ovl_path_real(file->f_path.dentry, &realpath);
	506	realfile = ovl_path_open(&realpath, file->f_flags);
	507	if (IS_ERR(realfile)) {
	508	kfree(od);
	509	return PTR_ERR(realfile);
	510	}
e9be9d5e	511	od->realfile = realfile;
1afaba1e MS	512	od->is_real = !OVL_TYPE_MERGE(type);
1afaba1e MS	513	od->is_upper = OVL_TYPE_UPPER(type);
e9be9d5e MS	514	file->private_data = od;
	515
	516	return 0;
	517	}
	518
	519	const struct file_operations ovl_dir_operations = {
	520	.read = generic_read_dir,
	521	.open = ovl_dir_open,
	522	.iterate = ovl_iterate,
	523	.llseek = ovl_dir_llseek,
	524	.fsync = ovl_dir_fsync,
	525	.release = ovl_dir_release,
	526	};
	527
	528	int ovl_check_empty_dir(struct dentry dentry, struct list_head list)
	529	{
	530	int err;
e9be9d5e MS	531	struct ovl_cache_entry *p;
e9be9d5e MS	532
c9f00fdb	533	err = ovl_dir_read_merged(dentry, list);
e9be9d5e MS	534	if (err)
	535	return err;
	536
	537	err = 0;
	538
	539	list_for_each_entry(p, list, l_node) {
	540	if (p->is_whiteout)
	541	continue;
	542
	543	if (p->name[0] == '.') {
	544	if (p->len == 1)
	545	continue;
	546	if (p->len == 2 && p->name[1] == '.')
	547	continue;
	548	}
	549	err = -ENOTEMPTY;
	550	break;
	551	}
	552
	553	return err;
	554	}
	555
	556	void ovl_cleanup_whiteouts(struct dentry upper, struct list_head list)
	557	{
	558	struct ovl_cache_entry *p;
	559
5955102c	560	inode_lock_nested(upper->d_inode, I_MUTEX_CHILD);
e9be9d5e MS	561	list_for_each_entry(p, list, l_node) {
	562	struct dentry *dentry;
	563
	564	if (!p->is_whiteout)
	565	continue;
	566
	567	dentry = lookup_one_len(p->name, upper, p->len);
	568	if (IS_ERR(dentry)) {
	569	pr_err("overlayfs: lookup '%s/%.*s' failed (%i)\n",
	570	upper->d_name.name, p->len, p->name,
	571	(int) PTR_ERR(dentry));
	572	continue;
	573	}
84889d49 KK	574	if (dentry->d_inode)
84889d49 KK	575	ovl_cleanup(upper->d_inode, dentry);
e9be9d5e MS	576	dput(dentry);
e9be9d5e MS	577	}
5955102c	578	inode_unlock(upper->d_inode);
e9be9d5e	579	}