[mirror_ubuntu-artful-kernel.git] / fs / aufs / inode.c

/*
 * Copyright (C) 2005-2017 Junjiro R. Okajima
 *
 * This program, aufs is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program.  If not, see <http://www.gnu.org/licenses/>.
 */

/*
 * inode functions
 */

#include "aufs.h"

struct inode *au_igrab(struct inode *inode)
{
	if (inode) {
		AuDebugOn(!atomic_read(&inode->i_count));
		ihold(inode);
	}
	return inode;
}

static void au_refresh_hinode_attr(struct inode *inode, int do_version)
{
	au_cpup_attr_all(inode, /*force*/0);
	au_update_iigen(inode, /*half*/1);
	if (do_version)
		inode->i_version++;
}

static int au_ii_refresh(struct inode *inode, int *update)
{
	int err, e, nbr;
	umode_t type;
	aufs_bindex_t bindex, new_bindex;
	struct super_block *sb;
	struct au_iinfo *iinfo;
	struct au_hinode *p, *q, tmp;

	AuDebugOn(au_is_bad_inode(inode));
	IiMustWriteLock(inode);

	*update = 0;
	sb = inode->i_sb;
	nbr = au_sbbot(sb) + 1;
	type = inode->i_mode & S_IFMT;
	iinfo = au_ii(inode);
	err = au_hinode_realloc(iinfo, nbr, /*may_shrink*/0);
	if (unlikely(err))
		goto out;

	AuDebugOn(iinfo->ii_btop < 0);
	p = au_hinode(iinfo, iinfo->ii_btop);
	for (bindex = iinfo->ii_btop; bindex <= iinfo->ii_bbot;
	     bindex++, p++) {
		if (!p->hi_inode)
			continue;

		AuDebugOn(type != (p->hi_inode->i_mode & S_IFMT));
		new_bindex = au_br_index(sb, p->hi_id);
		if (new_bindex == bindex)
			continue;

		if (new_bindex < 0) {
			*update = 1;
			au_hiput(p);
			p->hi_inode = NULL;
			continue;
		}

		if (new_bindex < iinfo->ii_btop)
			iinfo->ii_btop = new_bindex;
		if (iinfo->ii_bbot < new_bindex)
			iinfo->ii_bbot = new_bindex;
		/* swap two lower inode, and loop again */
		q = au_hinode(iinfo, new_bindex);
		tmp = *q;
		*q = *p;
		*p = tmp;
		if (tmp.hi_inode) {
			bindex--;
			p--;
		}
	}
	au_update_ibrange(inode, /*do_put_zero*/0);
	au_hinode_realloc(iinfo, nbr, /*may_shrink*/1); /* harmless if err */
	e = au_dy_irefresh(inode);
	if (unlikely(e && !err))
		err = e;

out:
	AuTraceErr(err);
	return err;
}

void au_refresh_iop(struct inode *inode, int force_getattr)
{
	int type;
	struct au_sbinfo *sbi = au_sbi(inode->i_sb);
	const struct inode_operations *iop
		= force_getattr ? aufs_iop : sbi->si_iop_array;

	if (inode->i_op == iop)
		return;

	switch (inode->i_mode & S_IFMT) {
	case S_IFDIR:
		type = AuIop_DIR;
		break;
	case S_IFLNK:
		type = AuIop_SYMLINK;
		break;
	default:
		type = AuIop_OTHER;
		break;
	}

	inode->i_op = iop + type;
	/* unnecessary smp_wmb() */
}

int au_refresh_hinode_self(struct inode *inode)
{
	int err, update;

	err = au_ii_refresh(inode, &update);
	if (!err)
		au_refresh_hinode_attr(inode, update && S_ISDIR(inode->i_mode));

	AuTraceErr(err);
	return err;
}

int au_refresh_hinode(struct inode *inode, struct dentry *dentry)
{
	int err, e, update;
	unsigned int flags;
	umode_t mode;
	aufs_bindex_t bindex, bbot;
	unsigned char isdir;
	struct au_hinode *p;
	struct au_iinfo *iinfo;

	err = au_ii_refresh(inode, &update);
	if (unlikely(err))
		goto out;

	update = 0;
	iinfo = au_ii(inode);
	p = au_hinode(iinfo, iinfo->ii_btop);
	mode = (inode->i_mode & S_IFMT);
	isdir = S_ISDIR(mode);
	flags = au_hi_flags(inode, isdir);
	bbot = au_dbbot(dentry);
	for (bindex = au_dbtop(dentry); bindex <= bbot; bindex++) {
		struct inode *h_i, *h_inode;
		struct dentry *h_d;

		h_d = au_h_dptr(dentry, bindex);
		if (!h_d || d_is_negative(h_d))
			continue;

		h_inode = d_inode(h_d);
		AuDebugOn(mode != (h_inode->i_mode & S_IFMT));
		if (iinfo->ii_btop <= bindex && bindex <= iinfo->ii_bbot) {
			h_i = au_h_iptr(inode, bindex);
			if (h_i) {
				if (h_i == h_inode)
					continue;
				err = -EIO;
				break;
			}
		}
		if (bindex < iinfo->ii_btop)
			iinfo->ii_btop = bindex;
		if (iinfo->ii_bbot < bindex)
			iinfo->ii_bbot = bindex;
		au_set_h_iptr(inode, bindex, au_igrab(h_inode), flags);
		update = 1;
	}
	au_update_ibrange(inode, /*do_put_zero*/0);
	e = au_dy_irefresh(inode);
	if (unlikely(e && !err))
		err = e;
	if (!err)
		au_refresh_hinode_attr(inode, update && isdir);

out:
	AuTraceErr(err);
	return err;
}

static int set_inode(struct inode *inode, struct dentry *dentry)
{
	int err;
	unsigned int flags;
	umode_t mode;
	aufs_bindex_t bindex, btop, btail;
	unsigned char isdir;
	struct dentry *h_dentry;
	struct inode *h_inode;
	struct au_iinfo *iinfo;
	struct inode_operations *iop;

	IiMustWriteLock(inode);

	err = 0;
	isdir = 0;
	iop = au_sbi(inode->i_sb)->si_iop_array;
	btop = au_dbtop(dentry);
	h_dentry = au_h_dptr(dentry, btop);
	h_inode = d_inode(h_dentry);
	mode = h_inode->i_mode;
	switch (mode & S_IFMT) {
	case S_IFREG:
		btail = au_dbtail(dentry);
		inode->i_op = iop + AuIop_OTHER;
		inode->i_fop = &aufs_file_fop;
		err = au_dy_iaop(inode, btop, h_inode);
		if (unlikely(err))
			goto out;
		break;
	case S_IFDIR:
		isdir = 1;
		btail = au_dbtaildir(dentry);
		inode->i_op = iop + AuIop_DIR;
		inode->i_fop = &aufs_dir_fop;
		break;
	case S_IFLNK:
		btail = au_dbtail(dentry);
		inode->i_op = iop + AuIop_SYMLINK;
		break;
	case S_IFBLK:
	case S_IFCHR:
	case S_IFIFO:
	case S_IFSOCK:
		btail = au_dbtail(dentry);
		inode->i_op = iop + AuIop_OTHER;
		init_special_inode(inode, mode, h_inode->i_rdev);
		break;
	default:
		AuIOErr("Unknown file type 0%o\n", mode);
		err = -EIO;
		goto out;
	}

	/* do not set hnotify for whiteouted dirs (SHWH mode) */
	flags = au_hi_flags(inode, isdir);
	if (au_opt_test(au_mntflags(dentry->d_sb), SHWH)
	    && au_ftest_hi(flags, HNOTIFY)
	    && dentry->d_name.len > AUFS_WH_PFX_LEN
	    && !memcmp(dentry->d_name.name, AUFS_WH_PFX, AUFS_WH_PFX_LEN))
		au_fclr_hi(flags, HNOTIFY);
	iinfo = au_ii(inode);
	iinfo->ii_btop = btop;
	iinfo->ii_bbot = btail;
	for (bindex = btop; bindex <= btail; bindex++) {
		h_dentry = au_h_dptr(dentry, bindex);
		if (h_dentry)
			au_set_h_iptr(inode, bindex,
				      au_igrab(d_inode(h_dentry)), flags);
	}
	au_cpup_attr_all(inode, /*force*/1);
	/*
	 * to force calling aufs_get_acl() every time,
	 * do not call cache_no_acl() for aufs inode.
	 */

out:
	return err;
}

/*
 * successful returns with iinfo write_locked
 * minus: errno
 * zero: success, matched
 * plus: no error, but unmatched
 */
static int reval_inode(struct inode *inode, struct dentry *dentry)
{
	int err;
	unsigned int gen, igflags;
	aufs_bindex_t bindex, bbot;
	struct inode *h_inode, *h_dinode;
	struct dentry *h_dentry;

	/*
	 * before this function, if aufs got any iinfo lock, it must be only
	 * one, the parent dir.
	 * it can happen by UDBA and the obsoleted inode number.
	 */
	err = -EIO;
	if (unlikely(inode->i_ino == parent_ino(dentry)))
		goto out;

	err = 1;
	ii_write_lock_new_child(inode);
	h_dentry = au_h_dptr(dentry, au_dbtop(dentry));
	h_dinode = d_inode(h_dentry);
	bbot = au_ibbot(inode);
	for (bindex = au_ibtop(inode); bindex <= bbot; bindex++) {
		h_inode = au_h_iptr(inode, bindex);
		if (!h_inode || h_inode != h_dinode)
			continue;

		err = 0;
		gen = au_iigen(inode, &igflags);
		if (gen == au_digen(dentry)
		    && !au_ig_ftest(igflags, HALF_REFRESHED))
			break;

		/* fully refresh inode using dentry */
		err = au_refresh_hinode(inode, dentry);
		if (!err)
			au_update_iigen(inode, /*half*/0);
		break;
	}

	if (unlikely(err))
		ii_write_unlock(inode);
out:
	return err;
}

int au_ino(struct super_block *sb, aufs_bindex_t bindex, ino_t h_ino,
	   unsigned int d_type, ino_t *ino)
{
	int err, idx;
	const int isnondir = d_type != DT_DIR;

	/* prevent hardlinked inode number from race condition */
	if (isnondir) {
		err = au_xinondir_enter(sb, bindex, h_ino, &idx);
		if (unlikely(err))
			goto out;
	}

	err = au_xino_read(sb, bindex, h_ino, ino);
	if (unlikely(err))
		goto out_xinondir;

	if (!*ino) {
		err = -EIO;
		*ino = au_xino_new_ino(sb);
		if (unlikely(!*ino))
			goto out_xinondir;
		err = au_xino_write(sb, bindex, h_ino, *ino);
		if (unlikely(err))
			goto out_xinondir;
	}

out_xinondir:
	if (isnondir && idx >= 0)
		au_xinondir_leave(sb, bindex, h_ino, idx);
out:
	return err;
}

/* successful returns with iinfo write_locked */
/* todo: return with unlocked? */
struct inode *au_new_inode(struct dentry *dentry, int must_new)
{
	struct inode *inode, *h_inode;
	struct dentry *h_dentry;
	struct super_block *sb;
	ino_t h_ino, ino;
	int err, idx, hlinked;
	aufs_bindex_t btop;

	sb = dentry->d_sb;
	btop = au_dbtop(dentry);
	h_dentry = au_h_dptr(dentry, btop);
	h_inode = d_inode(h_dentry);
	h_ino = h_inode->i_ino;
	hlinked = !d_is_dir(h_dentry) && h_inode->i_nlink > 1;

new_ino:
	/*
	 * stop 'race'-ing between hardlinks under different
	 * parents.
	 */
	if (hlinked) {
		err = au_xinondir_enter(sb, btop, h_ino, &idx);
		inode = ERR_PTR(err);
		if (unlikely(err))
			goto out;
	}

	err = au_xino_read(sb, btop, h_ino, &ino);
	inode = ERR_PTR(err);
	if (unlikely(err))
		goto out_xinondir;

	if (!ino) {
		ino = au_xino_new_ino(sb);
		if (unlikely(!ino)) {
			inode = ERR_PTR(-EIO);
			goto out_xinondir;
		}
	}

	AuDbg("i%lu\n", (unsigned long)ino);
	inode = au_iget_locked(sb, ino);
	err = PTR_ERR(inode);
	if (IS_ERR(inode))
		goto out_xinondir;

	AuDbg("%lx, new %d\n", inode->i_state, !!(inode->i_state & I_NEW));
	if (inode->i_state & I_NEW) {
		ii_write_lock_new_child(inode);
		err = set_inode(inode, dentry);
		if (!err) {
			unlock_new_inode(inode);
			goto out_xinondir; /* success */
		}

		/*
		 * iget_failed() calls iput(), but we need to call
		 * ii_write_unlock() after iget_failed(). so dirty hack for
		 * i_count.
		 */
		atomic_inc(&inode->i_count);
		iget_failed(inode);
		ii_write_unlock(inode);
		au_xino_write(sb, btop, h_ino, /*ino*/0);
		/* ignore this error */
		goto out_iput;
	} else if (!must_new && !IS_DEADDIR(inode) && inode->i_nlink) {
		/*
		 * horrible race condition between lookup, readdir and copyup
		 * (or something).
		 */
		if (hlinked && idx >= 0)
			au_xinondir_leave(sb, btop, h_ino, idx);
		err = reval_inode(inode, dentry);
		if (unlikely(err < 0)) {
			hlinked = 0;
			goto out_iput;
		}
		if (!err)
			goto out; /* success */
		else if (hlinked && idx >= 0) {
			err = au_xinondir_enter(sb, btop, h_ino, &idx);
			if (unlikely(err)) {
				iput(inode);
				inode = ERR_PTR(err);
				goto out;
			}
		}
	}

	if (unlikely(au_test_fs_unique_ino(h_inode)))
		AuWarn1("Warning: Un-notified UDBA or repeatedly renamed dir,"
			" b%d, %s, %pd, hi%lu, i%lu.\n",
			btop, au_sbtype(h_dentry->d_sb), dentry,
			(unsigned long)h_ino, (unsigned long)ino);
	ino = 0;
	err = au_xino_write(sb, btop, h_ino, /*ino*/0);
	if (!err) {
		iput(inode);
		if (hlinked && idx >= 0)
			au_xinondir_leave(sb, btop, h_ino, idx);
		goto new_ino;
	}

out_iput:
	iput(inode);
	inode = ERR_PTR(err);
out_xinondir:
	if (hlinked && idx >= 0)
		au_xinondir_leave(sb, btop, h_ino, idx);
out:
	return inode;
}

/* ---------------------------------------------------------------------- */

int au_test_ro(struct super_block *sb, aufs_bindex_t bindex,
	       struct inode *inode)
{
	int err;
	struct inode *hi;

	err = au_br_rdonly(au_sbr(sb, bindex));

	/* pseudo-link after flushed may happen out of bounds */
	if (!err
	    && inode
	    && au_ibtop(inode) <= bindex
	    && bindex <= au_ibbot(inode)) {
		/*
		 * permission check is unnecessary since vfsub routine
		 * will be called later
		 */
		hi = au_h_iptr(inode, bindex);
		if (hi)
			err = IS_IMMUTABLE(hi) ? -EROFS : 0;
	}

	return err;
}

int au_test_h_perm(struct inode *h_inode, int mask)
{
	if (uid_eq(current_fsuid(), GLOBAL_ROOT_UID))
		return 0;
	return inode_permission(h_inode, mask);
}

int au_test_h_perm_sio(struct inode *h_inode, int mask)
{
	if (au_test_nfs(h_inode->i_sb)
	    && (mask & MAY_WRITE)
	    && S_ISDIR(h_inode->i_mode))
		mask |= MAY_READ; /* force permission check */
	return au_test_h_perm(h_inode, mask);
}
Commit	Line	Data
b6450630 SF	1	/*
	2	* Copyright (C) 2005-2017 Junjiro R. Okajima
	3	*
	4	* This program, aufs is free software; you can redistribute it and/or modify
	5	* it under the terms of the GNU General Public License as published by
	6	* the Free Software Foundation; either version 2 of the License, or
	7	* (at your option) any later version.
	8	*
	9	* This program is distributed in the hope that it will be useful,
	10	* but WITHOUT ANY WARRANTY; without even the implied warranty of
	11	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
	12	* GNU General Public License for more details.
	13	*
	14	* You should have received a copy of the GNU General Public License
	15	* along with this program. If not, see <http://www.gnu.org/licenses/>.
	16	*/
	17
	18	/*
	19	* inode functions
	20	*/
	21
	22	#include "aufs.h"
	23
	24	struct inode au_igrab(struct inode inode)
	25	{
	26	if (inode) {
	27	AuDebugOn(!atomic_read(&inode->i_count));
	28	ihold(inode);
	29	}
	30	return inode;
	31	}
	32
	33	static void au_refresh_hinode_attr(struct inode *inode, int do_version)
	34	{
	35	au_cpup_attr_all(inode, /force/0);
	36	au_update_iigen(inode, /half/1);
	37	if (do_version)
	38	inode->i_version++;
	39	}
	40
	41	static int au_ii_refresh(struct inode inode, int update)
	42	{
	43	int err, e, nbr;
	44	umode_t type;
	45	aufs_bindex_t bindex, new_bindex;
	46	struct super_block *sb;
	47	struct au_iinfo *iinfo;
	48	struct au_hinode p, q, tmp;
	49
	50	AuDebugOn(au_is_bad_inode(inode));
	51	IiMustWriteLock(inode);
	52
	53	*update = 0;
	54	sb = inode->i_sb;
	55	nbr = au_sbbot(sb) + 1;
	56	type = inode->i_mode & S_IFMT;
	57	iinfo = au_ii(inode);
	58	err = au_hinode_realloc(iinfo, nbr, /may_shrink/0);
	59	if (unlikely(err))
	60	goto out;
	61
	62	AuDebugOn(iinfo->ii_btop < 0);
	63	p = au_hinode(iinfo, iinfo->ii_btop);
	64	for (bindex = iinfo->ii_btop; bindex <= iinfo->ii_bbot;
65	bindex++, p++) {
66	if (!p->hi_inode)
67	continue;
68
69	AuDebugOn(type != (p->hi_inode->i_mode & S_IFMT));
70	new_bindex = au_br_index(sb, p->hi_id);
71	if (new_bindex == bindex)
72	continue;
73
74	if (new_bindex < 0) {
75	*update = 1;
76	au_hiput(p);
77	p->hi_inode = NULL;
78	continue;
79	}
80
81	if (new_bindex < iinfo->ii_btop)
82	iinfo->ii_btop = new_bindex;
83	if (iinfo->ii_bbot < new_bindex)
84	iinfo->ii_bbot = new_bindex;
85	/* swap two lower inode, and loop again */
86	q = au_hinode(iinfo, new_bindex);
87	tmp = *q;
88	q = p;
89	*p = tmp;
90	if (tmp.hi_inode) {
91	bindex--;
92	p--;
93	}
94	}
95	au_update_ibrange(inode, /do_put_zero/0);
96	au_hinode_realloc(iinfo, nbr, /may_shrink/1); /* harmless if err */
97	e = au_dy_irefresh(inode);
98	if (unlikely(e && !err))
99	err = e;
100
101	out:
102	AuTraceErr(err);
103	return err;
104	}
105
106	void au_refresh_iop(struct inode *inode, int force_getattr)
107	{
108	int type;
109	struct au_sbinfo *sbi = au_sbi(inode->i_sb);
110	const struct inode_operations *iop
111	= force_getattr ? aufs_iop : sbi->si_iop_array;
112
113	if (inode->i_op == iop)
114	return;
115
116	switch (inode->i_mode & S_IFMT) {
117	case S_IFDIR:
118	type = AuIop_DIR;
119	break;
120	case S_IFLNK:
121	type = AuIop_SYMLINK;
122	break;
123	default:
124	type = AuIop_OTHER;
125	break;
126	}
127
128	inode->i_op = iop + type;
129	/* unnecessary smp_wmb() */
130	}
131
132	int au_refresh_hinode_self(struct inode *inode)
133	{
134	int err, update;
135
136	err = au_ii_refresh(inode, &update);
137	if (!err)
138	au_refresh_hinode_attr(inode, update && S_ISDIR(inode->i_mode));
139
140	AuTraceErr(err);
141	return err;
142	}
143
144	int au_refresh_hinode(struct inode inode, struct dentry dentry)
145	{
146	int err, e, update;
147	unsigned int flags;
148	umode_t mode;
149	aufs_bindex_t bindex, bbot;
150	unsigned char isdir;
151	struct au_hinode *p;
152	struct au_iinfo *iinfo;
153
154	err = au_ii_refresh(inode, &update);
155	if (unlikely(err))
156	goto out;
157
158	update = 0;
159	iinfo = au_ii(inode);
160	p = au_hinode(iinfo, iinfo->ii_btop);
161	mode = (inode->i_mode & S_IFMT);
162	isdir = S_ISDIR(mode);
163	flags = au_hi_flags(inode, isdir);
164	bbot = au_dbbot(dentry);
165	for (bindex = au_dbtop(dentry); bindex <= bbot; bindex++) {
166	struct inode h_i, h_inode;
167	struct dentry *h_d;
168
169	h_d = au_h_dptr(dentry, bindex);
170	if (!h_d \|\| d_is_negative(h_d))
171	continue;
172
173	h_inode = d_inode(h_d);
174	AuDebugOn(mode != (h_inode->i_mode & S_IFMT));
175	if (iinfo->ii_btop <= bindex && bindex <= iinfo->ii_bbot) {
176	h_i = au_h_iptr(inode, bindex);
177	if (h_i) {
178	if (h_i == h_inode)
179	continue;
180	err = -EIO;
181	break;
182	}
183	}
184	if (bindex < iinfo->ii_btop)
185	iinfo->ii_btop = bindex;
186	if (iinfo->ii_bbot < bindex)
187	iinfo->ii_bbot = bindex;
188	au_set_h_iptr(inode, bindex, au_igrab(h_inode), flags);
189	update = 1;
190	}
191	au_update_ibrange(inode, /do_put_zero/0);
192	e = au_dy_irefresh(inode);
193	if (unlikely(e && !err))
194	err = e;
195	if (!err)
196	au_refresh_hinode_attr(inode, update && isdir);
197
198	out:
199	AuTraceErr(err);
200	return err;
201	}
202
203	static int set_inode(struct inode inode, struct dentry dentry)
204	{
205	int err;
206	unsigned int flags;
207	umode_t mode;
208	aufs_bindex_t bindex, btop, btail;
209	unsigned char isdir;
210	struct dentry *h_dentry;
211	struct inode *h_inode;
212	struct au_iinfo *iinfo;
213	struct inode_operations *iop;
214
215	IiMustWriteLock(inode);
216
217	err = 0;
218	isdir = 0;
219	iop = au_sbi(inode->i_sb)->si_iop_array;
220	btop = au_dbtop(dentry);
221	h_dentry = au_h_dptr(dentry, btop);
222	h_inode = d_inode(h_dentry);
223	mode = h_inode->i_mode;
224	switch (mode & S_IFMT) {
225	case S_IFREG:
226	btail = au_dbtail(dentry);
227	inode->i_op = iop + AuIop_OTHER;
228	inode->i_fop = &aufs_file_fop;
229	err = au_dy_iaop(inode, btop, h_inode);
230	if (unlikely(err))
231	goto out;
232	break;
233	case S_IFDIR:
234	isdir = 1;
235	btail = au_dbtaildir(dentry);
236	inode->i_op = iop + AuIop_DIR;
237	inode->i_fop = &aufs_dir_fop;
238	break;
239	case S_IFLNK:
240	btail = au_dbtail(dentry);
241	inode->i_op = iop + AuIop_SYMLINK;
242	break;
243	case S_IFBLK:
244	case S_IFCHR:
245	case S_IFIFO:
246	case S_IFSOCK:
247	btail = au_dbtail(dentry);
248	inode->i_op = iop + AuIop_OTHER;
249	init_special_inode(inode, mode, h_inode->i_rdev);
250	break;
251	default:
252	AuIOErr("Unknown file type 0%o\n", mode);
253	err = -EIO;
254	goto out;
255	}
256
257	/* do not set hnotify for whiteouted dirs (SHWH mode) */
258	flags = au_hi_flags(inode, isdir);
259	if (au_opt_test(au_mntflags(dentry->d_sb), SHWH)
260	&& au_ftest_hi(flags, HNOTIFY)
261	&& dentry->d_name.len > AUFS_WH_PFX_LEN
262	&& !memcmp(dentry->d_name.name, AUFS_WH_PFX, AUFS_WH_PFX_LEN))
263	au_fclr_hi(flags, HNOTIFY);
264	iinfo = au_ii(inode);
265	iinfo->ii_btop = btop;
266	iinfo->ii_bbot = btail;
267	for (bindex = btop; bindex <= btail; bindex++) {
268	h_dentry = au_h_dptr(dentry, bindex);
269	if (h_dentry)
270	au_set_h_iptr(inode, bindex,
271	au_igrab(d_inode(h_dentry)), flags);
272	}
273	au_cpup_attr_all(inode, /force/1);
274	/*
275	* to force calling aufs_get_acl() every time,
276	* do not call cache_no_acl() for aufs inode.
277	*/
278
279	out:
280	return err;
281	}
282
283	/*
284	* successful returns with iinfo write_locked
285	* minus: errno
286	* zero: success, matched
287	* plus: no error, but unmatched
288	*/
289	static int reval_inode(struct inode inode, struct dentry dentry)
290	{
291	int err;
292	unsigned int gen, igflags;
293	aufs_bindex_t bindex, bbot;
294	struct inode h_inode, h_dinode;
295	struct dentry *h_dentry;
296
297	/*
298	* before this function, if aufs got any iinfo lock, it must be only
299	* one, the parent dir.
300	* it can happen by UDBA and the obsoleted inode number.
301	*/
302	err = -EIO;
303	if (unlikely(inode->i_ino == parent_ino(dentry)))
304	goto out;
305
306	err = 1;
307	ii_write_lock_new_child(inode);
308	h_dentry = au_h_dptr(dentry, au_dbtop(dentry));
309	h_dinode = d_inode(h_dentry);
310	bbot = au_ibbot(inode);
311	for (bindex = au_ibtop(inode); bindex <= bbot; bindex++) {
312	h_inode = au_h_iptr(inode, bindex);
313	if (!h_inode \|\| h_inode != h_dinode)
314	continue;
315
316	err = 0;
317	gen = au_iigen(inode, &igflags);
318	if (gen == au_digen(dentry)
319	&& !au_ig_ftest(igflags, HALF_REFRESHED))
320	break;
321
322	/* fully refresh inode using dentry */
323	err = au_refresh_hinode(inode, dentry);
324	if (!err)
325	au_update_iigen(inode, /half/0);
326	break;
327	}
328
329	if (unlikely(err))
330	ii_write_unlock(inode);
331	out:
332	return err;
333	}
334
335	int au_ino(struct super_block *sb, aufs_bindex_t bindex, ino_t h_ino,
336	unsigned int d_type, ino_t *ino)
337	{
338	int err, idx;
339	const int isnondir = d_type != DT_DIR;
340
341	/* prevent hardlinked inode number from race condition */
342	if (isnondir) {
343	err = au_xinondir_enter(sb, bindex, h_ino, &idx);
344	if (unlikely(err))
345	goto out;
346	}
347
348	err = au_xino_read(sb, bindex, h_ino, ino);
349	if (unlikely(err))
350	goto out_xinondir;
351
352	if (!*ino) {
353	err = -EIO;
354	*ino = au_xino_new_ino(sb);
355	if (unlikely(!*ino))
356	goto out_xinondir;
357	err = au_xino_write(sb, bindex, h_ino, *ino);
358	if (unlikely(err))
359	goto out_xinondir;
360	}
361
362	out_xinondir:
363	if (isnondir && idx >= 0)
364	au_xinondir_leave(sb, bindex, h_ino, idx);
365	out:
366	return err;
367	}
368
369	/* successful returns with iinfo write_locked */
370	/* todo: return with unlocked? */
371	struct inode au_new_inode(struct dentry dentry, int must_new)
372	{
373	struct inode inode, h_inode;
374	struct dentry *h_dentry;
375	struct super_block *sb;
376	ino_t h_ino, ino;
377	int err, idx, hlinked;
378	aufs_bindex_t btop;
379
380	sb = dentry->d_sb;
381	btop = au_dbtop(dentry);
382	h_dentry = au_h_dptr(dentry, btop);
383	h_inode = d_inode(h_dentry);
384	h_ino = h_inode->i_ino;
385	hlinked = !d_is_dir(h_dentry) && h_inode->i_nlink > 1;
386
387	new_ino:
388	/*
389	* stop 'race'-ing between hardlinks under different
390	* parents.
391	*/
392	if (hlinked) {
393	err = au_xinondir_enter(sb, btop, h_ino, &idx);
394	inode = ERR_PTR(err);
395	if (unlikely(err))
396	goto out;
397	}
398
399	err = au_xino_read(sb, btop, h_ino, &ino);
400	inode = ERR_PTR(err);
401	if (unlikely(err))
402	goto out_xinondir;
403
404	if (!ino) {
405	ino = au_xino_new_ino(sb);
406	if (unlikely(!ino)) {
407	inode = ERR_PTR(-EIO);
408	goto out_xinondir;
409	}
410	}
411
412	AuDbg("i%lu\n", (unsigned long)ino);
413	inode = au_iget_locked(sb, ino);
414	err = PTR_ERR(inode);
415	if (IS_ERR(inode))
416	goto out_xinondir;
417
418	AuDbg("%lx, new %d\n", inode->i_state, !!(inode->i_state & I_NEW));
419	if (inode->i_state & I_NEW) {
420	ii_write_lock_new_child(inode);
421	err = set_inode(inode, dentry);
422	if (!err) {
423	unlock_new_inode(inode);
424	goto out_xinondir; /* success */
425	}
426
427	/*
428	* iget_failed() calls iput(), but we need to call
429	* ii_write_unlock() after iget_failed(). so dirty hack for
430	* i_count.
431	*/
432	atomic_inc(&inode->i_count);
433	iget_failed(inode);
434	ii_write_unlock(inode);
435	au_xino_write(sb, btop, h_ino, /ino/0);
436	/* ignore this error */
437	goto out_iput;
438	} else if (!must_new && !IS_DEADDIR(inode) && inode->i_nlink) {
439	/*
440	* horrible race condition between lookup, readdir and copyup
441	* (or something).
442	*/
443	if (hlinked && idx >= 0)
444	au_xinondir_leave(sb, btop, h_ino, idx);
445	err = reval_inode(inode, dentry);
446	if (unlikely(err < 0)) {
447	hlinked = 0;
448	goto out_iput;
449	}
450	if (!err)
451	goto out; /* success */
452	else if (hlinked && idx >= 0) {
453	err = au_xinondir_enter(sb, btop, h_ino, &idx);
454	if (unlikely(err)) {
455	iput(inode);
456	inode = ERR_PTR(err);
457	goto out;
458	}
459	}
460	}
461
462	if (unlikely(au_test_fs_unique_ino(h_inode)))
463	AuWarn1("Warning: Un-notified UDBA or repeatedly renamed dir,"
464	" b%d, %s, %pd, hi%lu, i%lu.\n",
465	btop, au_sbtype(h_dentry->d_sb), dentry,
466	(unsigned long)h_ino, (unsigned long)ino);
467	ino = 0;
468	err = au_xino_write(sb, btop, h_ino, /ino/0);
469	if (!err) {
470	iput(inode);
471	if (hlinked && idx >= 0)
472	au_xinondir_leave(sb, btop, h_ino, idx);
473	goto new_ino;
474	}
475
476	out_iput:
477	iput(inode);
478	inode = ERR_PTR(err);
479	out_xinondir:
480	if (hlinked && idx >= 0)
481	au_xinondir_leave(sb, btop, h_ino, idx);
482	out:
483	return inode;
484	}
485
486	/* ---------------------------------------------------------------------- */
487
488	int au_test_ro(struct super_block *sb, aufs_bindex_t bindex,
489	struct inode *inode)
490	{
491	int err;
492	struct inode *hi;
493
494	err = au_br_rdonly(au_sbr(sb, bindex));
495
496	/* pseudo-link after flushed may happen out of bounds */
497	if (!err
498	&& inode
499	&& au_ibtop(inode) <= bindex
500	&& bindex <= au_ibbot(inode)) {
501	/*
502	* permission check is unnecessary since vfsub routine
503	* will be called later
504	*/
505	hi = au_h_iptr(inode, bindex);
506	if (hi)
507	err = IS_IMMUTABLE(hi) ? -EROFS : 0;
508	}
509
510	return err;
511	}
512
513	int au_test_h_perm(struct inode *h_inode, int mask)
514	{
515	if (uid_eq(current_fsuid(), GLOBAL_ROOT_UID))
516	return 0;
517	return inode_permission(h_inode, mask);
518	}
519
520	int au_test_h_perm_sio(struct inode *h_inode, int mask)
521	{
522	if (au_test_nfs(h_inode->i_sb)
523	&& (mask & MAY_WRITE)
524	&& S_ISDIR(h_inode->i_mode))
525	mask \|= MAY_READ; /* force permission check */
526	return au_test_h_perm(h_inode, mask);
527	}