xfs_iops.c 33.0 KB
Newer Older
L
Linus Torvalds 已提交
1
/*
2 3
 * Copyright (c) 2000-2005 Silicon Graphics, Inc.
 * All Rights Reserved.
L
Linus Torvalds 已提交
4
 *
5 6
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License as
L
Linus Torvalds 已提交
7 8
 * published by the Free Software Foundation.
 *
9 10 11 12
 * This program is distributed in the hope that it would 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.
L
Linus Torvalds 已提交
13
 *
14 15 16
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write the Free Software Foundation,
 * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
L
Linus Torvalds 已提交
17 18 19
 */
#include "xfs.h"
#include "xfs_fs.h"
20
#include "xfs_shared.h"
21 22 23
#include "xfs_format.h"
#include "xfs_log_format.h"
#include "xfs_trans_resv.h"
L
Linus Torvalds 已提交
24
#include "xfs_mount.h"
25
#include "xfs_da_format.h"
L
Linus Torvalds 已提交
26 27
#include "xfs_inode.h"
#include "xfs_bmap.h"
D
Dave Chinner 已提交
28
#include "xfs_bmap_util.h"
29 30
#include "xfs_acl.h"
#include "xfs_quota.h"
L
Linus Torvalds 已提交
31 32
#include "xfs_error.h"
#include "xfs_attr.h"
33
#include "xfs_trans.h"
C
Christoph Hellwig 已提交
34
#include "xfs_trace.h"
35
#include "xfs_icache.h"
D
Dave Chinner 已提交
36
#include "xfs_symlink.h"
37
#include "xfs_da_btree.h"
38
#include "xfs_dir2.h"
Z
Zhi Yong Wu 已提交
39
#include "xfs_trans_space.h"
40
#include "xfs_pnfs.h"
L
Linus Torvalds 已提交
41

42
#include <linux/capability.h>
L
Linus Torvalds 已提交
43 44
#include <linux/xattr.h>
#include <linux/namei.h>
45
#include <linux/posix_acl.h>
46
#include <linux/security.h>
E
Eric Sandeen 已提交
47
#include <linux/fiemap.h>
48
#include <linux/slab.h>
L
Linus Torvalds 已提交
49

50 51 52 53 54 55 56 57 58 59 60 61
/*
 * Directories have different lock order w.r.t. mmap_sem compared to regular
 * files. This is due to readdir potentially triggering page faults on a user
 * buffer inside filldir(), and this happens with the ilock on the directory
 * held. For regular files, the lock order is the other way around - the
 * mmap_sem is taken during the page fault, and then we lock the ilock to do
 * block mapping. Hence we need a different class for the directory ilock so
 * that lockdep can tell them apart.
 */
static struct lock_class_key xfs_nondir_ilock_class;
static struct lock_class_key xfs_dir_ilock_class;

62 63 64 65 66
static int
xfs_initxattrs(
	struct inode		*inode,
	const struct xattr	*xattr_array,
	void			*fs_info)
67
{
68 69 70
	const struct xattr	*xattr;
	struct xfs_inode	*ip = XFS_I(inode);
	int			error = 0;
71 72

	for (xattr = xattr_array; xattr->name != NULL; xattr++) {
D
Dave Chinner 已提交
73
		error = xfs_attr_set(ip, xattr->name, xattr->value,
74
				      xattr->value_len, ATTR_SECURE);
75 76 77 78 79 80
		if (error < 0)
			break;
	}
	return error;
}

81 82 83 84 85 86
/*
 * Hook in SELinux.  This is not quite correct yet, what we really need
 * here (as we do for default ACLs) is a mechanism by which creation of
 * these attrs can be journalled at inode creation time (along with the
 * inode, of course, such that log replay can't cause these to be lost).
 */
87

88
STATIC int
89
xfs_init_security(
90
	struct inode	*inode,
91 92
	struct inode	*dir,
	const struct qstr *qstr)
93
{
D
Dave Chinner 已提交
94
	return security_inode_init_security(inode, dir, qstr,
95
					     &xfs_initxattrs, NULL);
96 97
}

98 99 100
static void
xfs_dentry_to_name(
	struct xfs_name	*namep,
101 102
	struct dentry	*dentry,
	int		mode)
103 104 105
{
	namep->name = dentry->d_name.name;
	namep->len = dentry->d_name.len;
106
	namep->type = xfs_mode_to_ftype[(mode & S_IFMT) >> S_SHIFT];
107 108
}

109
STATIC void
110
xfs_cleanup_inode(
111
	struct inode	*dir,
112
	struct inode	*inode,
113
	struct dentry	*dentry)
114
{
115
	struct xfs_name	teardown;
116 117

	/* Oh, the horror.
118
	 * If we can't add the ACL or we fail in
119
	 * xfs_init_security we must back out.
120 121
	 * ENOSPC can hit here, among other things.
	 */
122
	xfs_dentry_to_name(&teardown, dentry, 0);
123

124
	xfs_remove(XFS_I(dir), &teardown, XFS_I(inode));
125 126
}

L
Linus Torvalds 已提交
127
STATIC int
128
xfs_generic_create(
L
Linus Torvalds 已提交
129 130
	struct inode	*dir,
	struct dentry	*dentry,
A
Al Viro 已提交
131
	umode_t		mode,
132 133
	dev_t		rdev,
	bool		tmpfile)	/* unnamed file */
L
Linus Torvalds 已提交
134
{
C
Christoph Hellwig 已提交
135
	struct inode	*inode;
136
	struct xfs_inode *ip = NULL;
137
	struct posix_acl *default_acl, *acl;
138
	struct xfs_name	name;
L
Linus Torvalds 已提交
139 140 141 142 143 144
	int		error;

	/*
	 * Irix uses Missed'em'V split, but doesn't want to see
	 * the upper 5 bits of (14bit) major.
	 */
145 146 147 148 149 150 151
	if (S_ISCHR(mode) || S_ISBLK(mode)) {
		if (unlikely(!sysv_valid_dev(rdev) || MAJOR(rdev) & ~0x1ff))
			return -EINVAL;
		rdev = sysv_encode_dev(rdev);
	} else {
		rdev = 0;
	}
L
Linus Torvalds 已提交
152

153 154 155
	error = posix_acl_create(dir, &mode, &default_acl, &acl);
	if (error)
		return error;
L
Linus Torvalds 已提交
156

157 158 159 160 161 162
	if (!tmpfile) {
		xfs_dentry_to_name(&name, dentry, mode);
		error = xfs_create(XFS_I(dir), &name, mode, rdev, &ip);
	} else {
		error = xfs_create_tmpfile(XFS_I(dir), dentry, mode, &ip);
	}
C
Christoph Hellwig 已提交
163 164
	if (unlikely(error))
		goto out_free_acl;
165

166
	inode = VFS_I(ip);
167

168
	error = xfs_init_security(inode, dir, &dentry->d_name);
C
Christoph Hellwig 已提交
169 170 171
	if (unlikely(error))
		goto out_cleanup_inode;

172
#ifdef CONFIG_XFS_POSIX_ACL
C
Christoph Hellwig 已提交
173
	if (default_acl) {
D
Dave Chinner 已提交
174
		error = xfs_set_acl(inode, default_acl, ACL_TYPE_DEFAULT);
175
		if (error)
C
Christoph Hellwig 已提交
176
			goto out_cleanup_inode;
L
Linus Torvalds 已提交
177
	}
178
	if (acl) {
D
Dave Chinner 已提交
179
		error = xfs_set_acl(inode, acl, ACL_TYPE_ACCESS);
180 181 182 183
		if (error)
			goto out_cleanup_inode;
	}
#endif
L
Linus Torvalds 已提交
184

185 186 187 188 189
	if (tmpfile)
		d_tmpfile(dentry, inode);
	else
		d_instantiate(dentry, inode);

190 191
	xfs_finish_inode_setup(ip);

192 193 194 195 196
 out_free_acl:
	if (default_acl)
		posix_acl_release(default_acl);
	if (acl)
		posix_acl_release(acl);
D
Dave Chinner 已提交
197
	return error;
C
Christoph Hellwig 已提交
198 199

 out_cleanup_inode:
200
	xfs_finish_inode_setup(ip);
201 202 203
	if (!tmpfile)
		xfs_cleanup_inode(dir, inode, dentry);
	iput(inode);
204
	goto out_free_acl;
L
Linus Torvalds 已提交
205 206
}

207 208 209 210 211 212 213 214 215 216
STATIC int
xfs_vn_mknod(
	struct inode	*dir,
	struct dentry	*dentry,
	umode_t		mode,
	dev_t		rdev)
{
	return xfs_generic_create(dir, dentry, mode, rdev, false);
}

L
Linus Torvalds 已提交
217
STATIC int
218
xfs_vn_create(
L
Linus Torvalds 已提交
219 220
	struct inode	*dir,
	struct dentry	*dentry,
A
Al Viro 已提交
221
	umode_t		mode,
A
Al Viro 已提交
222
	bool		flags)
L
Linus Torvalds 已提交
223
{
224
	return xfs_vn_mknod(dir, dentry, mode, 0);
L
Linus Torvalds 已提交
225 226 227
}

STATIC int
228
xfs_vn_mkdir(
L
Linus Torvalds 已提交
229 230
	struct inode	*dir,
	struct dentry	*dentry,
231
	umode_t		mode)
L
Linus Torvalds 已提交
232
{
233
	return xfs_vn_mknod(dir, dentry, mode|S_IFDIR, 0);
L
Linus Torvalds 已提交
234 235 236
}

STATIC struct dentry *
237
xfs_vn_lookup(
L
Linus Torvalds 已提交
238 239
	struct inode	*dir,
	struct dentry	*dentry,
A
Al Viro 已提交
240
	unsigned int flags)
L
Linus Torvalds 已提交
241
{
242
	struct xfs_inode *cip;
243
	struct xfs_name	name;
L
Linus Torvalds 已提交
244 245 246 247 248
	int		error;

	if (dentry->d_name.len >= MAXNAMELEN)
		return ERR_PTR(-ENAMETOOLONG);

249
	xfs_dentry_to_name(&name, dentry, 0);
250
	error = xfs_lookup(XFS_I(dir), &name, &cip, NULL);
251
	if (unlikely(error)) {
D
Dave Chinner 已提交
252 253
		if (unlikely(error != -ENOENT))
			return ERR_PTR(error);
L
Linus Torvalds 已提交
254 255 256 257
		d_add(dentry, NULL);
		return NULL;
	}

258
	return d_splice_alias(VFS_I(cip), dentry);
L
Linus Torvalds 已提交
259 260
}

261 262 263 264
STATIC struct dentry *
xfs_vn_ci_lookup(
	struct inode	*dir,
	struct dentry	*dentry,
A
Al Viro 已提交
265
	unsigned int flags)
266 267 268 269 270 271 272 273 274 275
{
	struct xfs_inode *ip;
	struct xfs_name	xname;
	struct xfs_name ci_name;
	struct qstr	dname;
	int		error;

	if (dentry->d_name.len >= MAXNAMELEN)
		return ERR_PTR(-ENAMETOOLONG);

276
	xfs_dentry_to_name(&xname, dentry, 0);
277 278
	error = xfs_lookup(XFS_I(dir), &xname, &ip, &ci_name);
	if (unlikely(error)) {
D
Dave Chinner 已提交
279 280
		if (unlikely(error != -ENOENT))
			return ERR_PTR(error);
281 282 283 284 285
		/*
		 * call d_add(dentry, NULL) here when d_drop_negative_children
		 * is called in xfs_vn_mknod (ie. allow negative dentries
		 * with CI filesystems).
		 */
286 287 288 289 290
		return NULL;
	}

	/* if exact match, just splice and exit */
	if (!ci_name.name)
291
		return d_splice_alias(VFS_I(ip), dentry);
292 293 294 295

	/* else case-insensitive match... */
	dname.name = ci_name.name;
	dname.len = ci_name.len;
296
	dentry = d_add_ci(dentry, VFS_I(ip), &dname);
297 298 299 300
	kmem_free(ci_name.name);
	return dentry;
}

L
Linus Torvalds 已提交
301
STATIC int
302
xfs_vn_link(
L
Linus Torvalds 已提交
303 304 305 306
	struct dentry	*old_dentry,
	struct inode	*dir,
	struct dentry	*dentry)
{
307
	struct inode	*inode = d_inode(old_dentry);
308
	struct xfs_name	name;
L
Linus Torvalds 已提交
309 310
	int		error;

311
	xfs_dentry_to_name(&name, dentry, inode->i_mode);
L
Linus Torvalds 已提交
312

313
	error = xfs_link(XFS_I(dir), XFS_I(inode), &name);
314
	if (unlikely(error))
D
Dave Chinner 已提交
315
		return error;
316

A
Al Viro 已提交
317
	ihold(inode);
318 319
	d_instantiate(dentry, inode);
	return 0;
L
Linus Torvalds 已提交
320 321 322
}

STATIC int
323
xfs_vn_unlink(
L
Linus Torvalds 已提交
324 325 326
	struct inode	*dir,
	struct dentry	*dentry)
{
327
	struct xfs_name	name;
L
Linus Torvalds 已提交
328 329
	int		error;

330
	xfs_dentry_to_name(&name, dentry, 0);
L
Linus Torvalds 已提交
331

332
	error = xfs_remove(XFS_I(dir), &name, XFS_I(d_inode(dentry)));
333 334 335 336 337 338 339 340 341 342 343
	if (error)
		return error;

	/*
	 * With unlink, the VFS makes the dentry "negative": no inode,
	 * but still hashed. This is incompatible with case-insensitive
	 * mode, so invalidate (unhash) the dentry in CI-mode.
	 */
	if (xfs_sb_version_hasasciici(&XFS_M(dir->i_sb)->m_sb))
		d_invalidate(dentry);
	return 0;
L
Linus Torvalds 已提交
344 345 346
}

STATIC int
347
xfs_vn_symlink(
L
Linus Torvalds 已提交
348 349 350 351
	struct inode	*dir,
	struct dentry	*dentry,
	const char	*symname)
{
352 353
	struct inode	*inode;
	struct xfs_inode *cip = NULL;
354
	struct xfs_name	name;
L
Linus Torvalds 已提交
355
	int		error;
A
Al Viro 已提交
356
	umode_t		mode;
L
Linus Torvalds 已提交
357

358
	mode = S_IFLNK |
A
Al Viro 已提交
359
		(irix_symlink_mode ? 0777 & ~current_umask() : S_IRWXUGO);
360
	xfs_dentry_to_name(&name, dentry, mode);
L
Linus Torvalds 已提交
361

C
Christoph Hellwig 已提交
362
	error = xfs_symlink(XFS_I(dir), &name, symname, mode, &cip);
363 364 365
	if (unlikely(error))
		goto out;

366
	inode = VFS_I(cip);
367

368
	error = xfs_init_security(inode, dir, &dentry->d_name);
369 370 371 372
	if (unlikely(error))
		goto out_cleanup_inode;

	d_instantiate(dentry, inode);
373
	xfs_finish_inode_setup(cip);
374 375 376
	return 0;

 out_cleanup_inode:
377
	xfs_finish_inode_setup(cip);
378
	xfs_cleanup_inode(dir, inode, dentry);
379
	iput(inode);
380
 out:
D
Dave Chinner 已提交
381
	return error;
L
Linus Torvalds 已提交
382 383 384
}

STATIC int
385
xfs_vn_rename(
L
Linus Torvalds 已提交
386 387 388
	struct inode	*odir,
	struct dentry	*odentry,
	struct inode	*ndir,
389 390
	struct dentry	*ndentry,
	unsigned int	flags)
L
Linus Torvalds 已提交
391
{
392
	struct inode	*new_inode = d_inode(ndentry);
393
	int		omode = 0;
394 395
	struct xfs_name	oname;
	struct xfs_name	nname;
L
Linus Torvalds 已提交
396

D
Dave Chinner 已提交
397
	if (flags & ~(RENAME_NOREPLACE | RENAME_EXCHANGE | RENAME_WHITEOUT))
398 399
		return -EINVAL;

400 401
	/* if we are exchanging files, we need to set i_mode of both files */
	if (flags & RENAME_EXCHANGE)
402
		omode = d_inode(ndentry)->i_mode;
403 404

	xfs_dentry_to_name(&oname, odentry, omode);
405
	xfs_dentry_to_name(&nname, ndentry, d_inode(odentry)->i_mode);
406

407
	return xfs_rename(XFS_I(odir), &oname, XFS_I(d_inode(odentry)),
408
			  XFS_I(ndir), &nname,
409
			  new_inode ? XFS_I(new_inode) : NULL, flags);
L
Linus Torvalds 已提交
410 411 412 413 414 415 416
}

/*
 * careful here - this function can get called recursively, so
 * we need to be very careful about how much stack we use.
 * uio is kmalloced for this reason...
 */
417
STATIC void *
418
xfs_vn_follow_link(
L
Linus Torvalds 已提交
419 420 421 422
	struct dentry		*dentry,
	struct nameidata	*nd)
{
	char			*link;
423
	int			error = -ENOMEM;
L
Linus Torvalds 已提交
424

425
	link = kmalloc(MAXPATHLEN+1, GFP_KERNEL);
426 427
	if (!link)
		goto out_err;
L
Linus Torvalds 已提交
428

429
	error = xfs_readlink(XFS_I(d_inode(dentry)), link);
430 431
	if (unlikely(error))
		goto out_kfree;
L
Linus Torvalds 已提交
432 433

	nd_set_link(nd, link);
434
	return NULL;
435 436 437 438 439 440

 out_kfree:
	kfree(link);
 out_err:
	nd_set_link(nd, ERR_PTR(error));
	return NULL;
L
Linus Torvalds 已提交
441 442 443
}

STATIC int
444
xfs_vn_getattr(
445 446 447
	struct vfsmount		*mnt,
	struct dentry		*dentry,
	struct kstat		*stat)
L
Linus Torvalds 已提交
448
{
449
	struct inode		*inode = d_inode(dentry);
450 451 452
	struct xfs_inode	*ip = XFS_I(inode);
	struct xfs_mount	*mp = ip->i_mount;

C
Christoph Hellwig 已提交
453
	trace_xfs_getattr(ip);
454 455

	if (XFS_FORCED_SHUTDOWN(mp))
E
Eric Sandeen 已提交
456
		return -EIO;
457 458 459 460 461

	stat->size = XFS_ISIZE(ip);
	stat->dev = inode->i_sb->s_dev;
	stat->mode = ip->i_d.di_mode;
	stat->nlink = ip->i_d.di_nlink;
462 463
	stat->uid = inode->i_uid;
	stat->gid = inode->i_gid;
464 465
	stat->ino = ip->i_ino;
	stat->atime = inode->i_atime;
466 467
	stat->mtime = inode->i_mtime;
	stat->ctime = inode->i_ctime;
468 469 470 471 472 473 474 475 476 477 478 479
	stat->blocks =
		XFS_FSB_TO_BB(mp, ip->i_d.di_nblocks + ip->i_delayed_blks);


	switch (inode->i_mode & S_IFMT) {
	case S_IFBLK:
	case S_IFCHR:
		stat->blksize = BLKDEV_IOSIZE;
		stat->rdev = MKDEV(sysv_major(ip->i_df.if_u2.if_rdev) & 0x1ff,
				   sysv_minor(ip->i_df.if_u2.if_rdev));
		break;
	default:
480
		if (XFS_IS_REALTIME_INODE(ip)) {
481 482 483 484 485 486 487 488 489 490 491
			/*
			 * If the file blocks are being allocated from a
			 * realtime volume, then return the inode's realtime
			 * extent size or the realtime volume's extent size.
			 */
			stat->blksize =
				xfs_get_extsz_hint(ip) << mp->m_sb.sb_blocklog;
		} else
			stat->blksize = xfs_preferred_iosize(mp);
		stat->rdev = 0;
		break;
492
	}
493 494

	return 0;
L
Linus Torvalds 已提交
495 496
}

497 498 499 500 501
static void
xfs_setattr_mode(
	struct xfs_inode	*ip,
	struct iattr		*iattr)
{
502 503
	struct inode		*inode = VFS_I(ip);
	umode_t			mode = iattr->ia_mode;
504 505 506 507 508 509 510 511 512 513

	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));

	ip->i_d.di_mode &= S_IFMT;
	ip->i_d.di_mode |= mode & ~S_IFMT;

	inode->i_mode &= S_IFMT;
	inode->i_mode |= mode & ~S_IFMT;
}

514
void
C
Christoph Hellwig 已提交
515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539
xfs_setattr_time(
	struct xfs_inode	*ip,
	struct iattr		*iattr)
{
	struct inode		*inode = VFS_I(ip);

	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));

	if (iattr->ia_valid & ATTR_ATIME) {
		inode->i_atime = iattr->ia_atime;
		ip->i_d.di_atime.t_sec = iattr->ia_atime.tv_sec;
		ip->i_d.di_atime.t_nsec = iattr->ia_atime.tv_nsec;
	}
	if (iattr->ia_valid & ATTR_CTIME) {
		inode->i_ctime = iattr->ia_ctime;
		ip->i_d.di_ctime.t_sec = iattr->ia_ctime.tv_sec;
		ip->i_d.di_ctime.t_nsec = iattr->ia_ctime.tv_nsec;
	}
	if (iattr->ia_valid & ATTR_MTIME) {
		inode->i_mtime = iattr->ia_mtime;
		ip->i_d.di_mtime.t_sec = iattr->ia_mtime.tv_sec;
		ip->i_d.di_mtime.t_nsec = iattr->ia_mtime.tv_nsec;
	}
}

C
Christoph Hellwig 已提交
540 541 542 543 544 545 546 547 548 549 550
int
xfs_setattr_nonsize(
	struct xfs_inode	*ip,
	struct iattr		*iattr,
	int			flags)
{
	xfs_mount_t		*mp = ip->i_mount;
	struct inode		*inode = VFS_I(ip);
	int			mask = iattr->ia_valid;
	xfs_trans_t		*tp;
	int			error;
551 552
	kuid_t			uid = GLOBAL_ROOT_UID, iuid = GLOBAL_ROOT_UID;
	kgid_t			gid = GLOBAL_ROOT_GID, igid = GLOBAL_ROOT_GID;
C
Christoph Hellwig 已提交
553 554 555 556 557
	struct xfs_dquot	*udqp = NULL, *gdqp = NULL;
	struct xfs_dquot	*olddquot1 = NULL, *olddquot2 = NULL;

	trace_xfs_setattr(ip);

558 559 560
	/* If acls are being inherited, we already have this checked */
	if (!(flags & XFS_ATTR_NOACL)) {
		if (mp->m_flags & XFS_MOUNT_RDONLY)
D
Dave Chinner 已提交
561
			return -EROFS;
C
Christoph Hellwig 已提交
562

563
		if (XFS_FORCED_SHUTDOWN(mp))
D
Dave Chinner 已提交
564
			return -EIO;
C
Christoph Hellwig 已提交
565

D
Dave Chinner 已提交
566
		error = inode_change_ok(inode, iattr);
567
		if (error)
E
Eric Sandeen 已提交
568
			return error;
569
	}
C
Christoph Hellwig 已提交
570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587

	ASSERT((mask & ATTR_SIZE) == 0);

	/*
	 * If disk quotas is on, we make sure that the dquots do exist on disk,
	 * before we start any other transactions. Trying to do this later
	 * is messy. We don't care to take a readlock to look at the ids
	 * in inode here, because we can't hold it across the trans_reserve.
	 * If the IDs do change before we take the ilock, we're covered
	 * because the i_*dquot fields will get updated anyway.
	 */
	if (XFS_IS_QUOTA_ON(mp) && (mask & (ATTR_UID|ATTR_GID))) {
		uint	qflags = 0;

		if ((mask & ATTR_UID) && XFS_IS_UQUOTA_ON(mp)) {
			uid = iattr->ia_uid;
			qflags |= XFS_QMOPT_UQUOTA;
		} else {
588
			uid = inode->i_uid;
C
Christoph Hellwig 已提交
589 590 591 592 593
		}
		if ((mask & ATTR_GID) && XFS_IS_GQUOTA_ON(mp)) {
			gid = iattr->ia_gid;
			qflags |= XFS_QMOPT_GQUOTA;
		}  else {
594
			gid = inode->i_gid;
C
Christoph Hellwig 已提交
595 596 597 598 599 600 601 602 603
		}

		/*
		 * We take a reference when we initialize udqp and gdqp,
		 * so it is important that we never blindly double trip on
		 * the same variable. See xfs_create() for an example.
		 */
		ASSERT(udqp == NULL);
		ASSERT(gdqp == NULL);
604 605 606 607
		error = xfs_qm_vop_dqalloc(ip, xfs_kuid_to_uid(uid),
					   xfs_kgid_to_gid(gid),
					   xfs_get_projid(ip),
					   qflags, &udqp, &gdqp, NULL);
C
Christoph Hellwig 已提交
608 609 610 611 612
		if (error)
			return error;
	}

	tp = xfs_trans_alloc(mp, XFS_TRANS_SETATTR_NOT_SIZE);
613
	error = xfs_trans_reserve(tp, &M_RES(mp)->tr_ichange, 0, 0);
C
Christoph Hellwig 已提交
614 615 616 617 618 619 620 621 622 623 624 625 626 627 628
	if (error)
		goto out_dqrele;

	xfs_ilock(ip, XFS_ILOCK_EXCL);

	/*
	 * Change file ownership.  Must be the owner or privileged.
	 */
	if (mask & (ATTR_UID|ATTR_GID)) {
		/*
		 * These IDs could have changed since we last looked at them.
		 * But, we're assured that if the ownership did change
		 * while we didn't have the inode locked, inode's dquot(s)
		 * would have changed also.
		 */
629 630
		iuid = inode->i_uid;
		igid = inode->i_gid;
C
Christoph Hellwig 已提交
631 632 633 634 635 636 637 638
		gid = (mask & ATTR_GID) ? iattr->ia_gid : igid;
		uid = (mask & ATTR_UID) ? iattr->ia_uid : iuid;

		/*
		 * Do a quota reservation only if uid/gid is actually
		 * going to change.
		 */
		if (XFS_IS_QUOTA_RUNNING(mp) &&
639 640
		    ((XFS_IS_UQUOTA_ON(mp) && !uid_eq(iuid, uid)) ||
		     (XFS_IS_GQUOTA_ON(mp) && !gid_eq(igid, gid)))) {
C
Christoph Hellwig 已提交
641 642
			ASSERT(tp);
			error = xfs_qm_vop_chown_reserve(tp, ip, udqp, gdqp,
643
						NULL, capable(CAP_FOWNER) ?
C
Christoph Hellwig 已提交
644 645 646 647 648 649
						XFS_QMOPT_FORCE_RES : 0);
			if (error)	/* out of quota */
				goto out_trans_cancel;
		}
	}

650
	xfs_trans_ijoin(tp, ip, 0);
C
Christoph Hellwig 已提交
651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669

	/*
	 * Change file ownership.  Must be the owner or privileged.
	 */
	if (mask & (ATTR_UID|ATTR_GID)) {
		/*
		 * CAP_FSETID overrides the following restrictions:
		 *
		 * The set-user-ID and set-group-ID bits of a file will be
		 * cleared upon successful return from chown()
		 */
		if ((ip->i_d.di_mode & (S_ISUID|S_ISGID)) &&
		    !capable(CAP_FSETID))
			ip->i_d.di_mode &= ~(S_ISUID|S_ISGID);

		/*
		 * Change the ownerships and register quota modifications
		 * in the transaction.
		 */
670
		if (!uid_eq(iuid, uid)) {
C
Christoph Hellwig 已提交
671 672 673 674 675 676
			if (XFS_IS_QUOTA_RUNNING(mp) && XFS_IS_UQUOTA_ON(mp)) {
				ASSERT(mask & ATTR_UID);
				ASSERT(udqp);
				olddquot1 = xfs_qm_vop_chown(tp, ip,
							&ip->i_udquot, udqp);
			}
677
			ip->i_d.di_uid = xfs_kuid_to_uid(uid);
C
Christoph Hellwig 已提交
678 679
			inode->i_uid = uid;
		}
680
		if (!gid_eq(igid, gid)) {
C
Christoph Hellwig 已提交
681
			if (XFS_IS_QUOTA_RUNNING(mp) && XFS_IS_GQUOTA_ON(mp)) {
682 683
				ASSERT(xfs_sb_version_has_pquotino(&mp->m_sb) ||
				       !XFS_IS_PQUOTA_ON(mp));
C
Christoph Hellwig 已提交
684 685 686 687 688
				ASSERT(mask & ATTR_GID);
				ASSERT(gdqp);
				olddquot2 = xfs_qm_vop_chown(tp, ip,
							&ip->i_gdquot, gdqp);
			}
689
			ip->i_d.di_gid = xfs_kgid_to_gid(gid);
C
Christoph Hellwig 已提交
690 691 692 693
			inode->i_gid = gid;
		}
	}

694
	if (mask & ATTR_MODE)
695
		xfs_setattr_mode(ip, iattr);
C
Christoph Hellwig 已提交
696 697
	if (mask & (ATTR_ATIME|ATTR_CTIME|ATTR_MTIME))
		xfs_setattr_time(ip, iattr);
C
Christoph Hellwig 已提交
698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717

	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);

	XFS_STATS_INC(xs_ig_attrchg);

	if (mp->m_flags & XFS_MOUNT_WSYNC)
		xfs_trans_set_sync(tp);
	error = xfs_trans_commit(tp, 0);

	xfs_iunlock(ip, XFS_ILOCK_EXCL);

	/*
	 * Release any dquot(s) the inode had kept before chown.
	 */
	xfs_qm_dqrele(olddquot1);
	xfs_qm_dqrele(olddquot2);
	xfs_qm_dqrele(udqp);
	xfs_qm_dqrele(gdqp);

	if (error)
E
Eric Sandeen 已提交
718
		return error;
C
Christoph Hellwig 已提交
719 720 721 722 723 724 725 726 727

	/*
	 * XXX(hch): Updating the ACL entries is not atomic vs the i_mode
	 * 	     update.  We could avoid this with linked transactions
	 * 	     and passing down the transaction pointer all the way
	 *	     to attr_set.  No previous user of the generic
	 * 	     Posix ACL code seems to care about this issue either.
	 */
	if ((mask & ATTR_MODE) && !(flags & XFS_ATTR_NOACL)) {
D
Dave Chinner 已提交
728
		error = posix_acl_chmod(inode, inode->i_mode);
C
Christoph Hellwig 已提交
729
		if (error)
E
Eric Sandeen 已提交
730
			return error;
C
Christoph Hellwig 已提交
731 732 733 734 735
	}

	return 0;

out_trans_cancel:
736
	xfs_trans_cancel(tp);
C
Christoph Hellwig 已提交
737 738 739 740 741 742 743 744 745 746 747 748 749
	xfs_iunlock(ip, XFS_ILOCK_EXCL);
out_dqrele:
	xfs_qm_dqrele(udqp);
	xfs_qm_dqrele(gdqp);
	return error;
}

/*
 * Truncate file.  Must have write permission and not be a directory.
 */
int
xfs_setattr_size(
	struct xfs_inode	*ip,
750
	struct iattr		*iattr)
C
Christoph Hellwig 已提交
751 752 753
{
	struct xfs_mount	*mp = ip->i_mount;
	struct inode		*inode = VFS_I(ip);
754
	xfs_off_t		oldsize, newsize;
C
Christoph Hellwig 已提交
755 756
	struct xfs_trans	*tp;
	int			error;
757
	uint			lock_flags = 0;
758
	bool			did_zeroing = false;
C
Christoph Hellwig 已提交
759 760 761 762

	trace_xfs_setattr(ip);

	if (mp->m_flags & XFS_MOUNT_RDONLY)
D
Dave Chinner 已提交
763
		return -EROFS;
C
Christoph Hellwig 已提交
764 765

	if (XFS_FORCED_SHUTDOWN(mp))
D
Dave Chinner 已提交
766
		return -EIO;
C
Christoph Hellwig 已提交
767

D
Dave Chinner 已提交
768
	error = inode_change_ok(inode, iattr);
C
Christoph Hellwig 已提交
769
	if (error)
E
Eric Sandeen 已提交
770
		return error;
C
Christoph Hellwig 已提交
771

772
	ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL));
773
	ASSERT(xfs_isilocked(ip, XFS_MMAPLOCK_EXCL));
C
Christoph Hellwig 已提交
774
	ASSERT(S_ISREG(ip->i_d.di_mode));
775 776
	ASSERT((iattr->ia_valid & (ATTR_UID|ATTR_GID|ATTR_ATIME|ATTR_ATIME_SET|
		ATTR_MTIME_SET|ATTR_KILL_PRIV|ATTR_TIMES_SET)) == 0);
C
Christoph Hellwig 已提交
777

778
	oldsize = inode->i_size;
779 780
	newsize = iattr->ia_size;

C
Christoph Hellwig 已提交
781 782 783
	/*
	 * Short circuit the truncate case for zero length files.
	 */
784
	if (newsize == 0 && oldsize == 0 && ip->i_d.di_nextents == 0) {
785
		if (!(iattr->ia_valid & (ATTR_CTIME|ATTR_MTIME)))
786
			return 0;
787 788 789 790 791 792

		/*
		 * Use the regular setattr path to update the timestamps.
		 */
		iattr->ia_valid &= ~ATTR_SIZE;
		return xfs_setattr_nonsize(ip, iattr, 0);
C
Christoph Hellwig 已提交
793 794 795 796 797
	}

	/*
	 * Make sure that the dquots are attached to the inode.
	 */
798
	error = xfs_qm_dqattach(ip, 0);
C
Christoph Hellwig 已提交
799
	if (error)
800
		return error;
C
Christoph Hellwig 已提交
801 802

	/*
803 804 805 806 807 808 809
	 * File data changes must be complete before we start the transaction to
	 * modify the inode.  This needs to be done before joining the inode to
	 * the transaction because the inode cannot be unlocked once it is a
	 * part of the transaction.
	 *
	 * Start with zeroing any data block beyond EOF that we may expose on
	 * file extension.
C
Christoph Hellwig 已提交
810
	 */
811
	if (newsize > oldsize) {
812
		error = xfs_zero_eof(ip, newsize, oldsize, &did_zeroing);
C
Christoph Hellwig 已提交
813
		if (error)
814
			return error;
C
Christoph Hellwig 已提交
815 816 817 818 819 820 821
	}

	/*
	 * We are going to log the inode size change in this transaction so
	 * any previous writes that are beyond the on disk EOF and the new
	 * EOF that have not been written out need to be written here.  If we
	 * do not write the data out, we expose ourselves to the null files
822 823
	 * problem. Note that this includes any block zeroing we did above;
	 * otherwise those blocks may not be zeroed after a crash.
C
Christoph Hellwig 已提交
824
	 */
825 826
	if (newsize > ip->i_d.di_size &&
	    (oldsize != ip->i_d.di_size || did_zeroing)) {
D
Dave Chinner 已提交
827
		error = filemap_write_and_wait_range(VFS_I(ip)->i_mapping,
D
Dave Chinner 已提交
828
						      ip->i_d.di_size, newsize);
C
Christoph Hellwig 已提交
829
		if (error)
830
			return error;
C
Christoph Hellwig 已提交
831 832
	}

833
	/* Now wait for all direct I/O to complete. */
C
Christoph Hellwig 已提交
834
	inode_dio_wait(inode);
C
Christoph Hellwig 已提交
835

836
	/*
837 838 839 840 841 842
	 * We've already locked out new page faults, so now we can safely remove
	 * pages from the page cache knowing they won't get refaulted until we
	 * drop the XFS_MMAP_EXCL lock after the extent manipulations are
	 * complete. The truncate_setsize() call also cleans partial EOF page
	 * PTEs on extending truncates and hence ensures sub-page block size
	 * filesystems are correctly handled, too.
843
	 *
844 845 846 847 848 849 850 851
	 * We have to do all the page cache truncate work outside the
	 * transaction context as the "lock" order is page lock->log space
	 * reservation as defined by extent allocation in the writeback path.
	 * Hence a truncate can fail with ENOMEM from xfs_trans_reserve(), but
	 * having already truncated the in-memory version of the file (i.e. made
	 * user visible changes). There's not much we can do about this, except
	 * to hope that the caller sees ENOMEM and retries the truncate
	 * operation.
852
	 */
D
Dave Chinner 已提交
853
	error = block_truncate_page(inode->i_mapping, newsize, xfs_get_blocks);
C
Christoph Hellwig 已提交
854
	if (error)
855
		return error;
856
	truncate_setsize(inode, newsize);
C
Christoph Hellwig 已提交
857 858

	tp = xfs_trans_alloc(mp, XFS_TRANS_SETATTR_SIZE);
859
	error = xfs_trans_reserve(tp, &M_RES(mp)->tr_itruncate, 0, 0);
C
Christoph Hellwig 已提交
860 861 862 863 864
	if (error)
		goto out_trans_cancel;

	lock_flags |= XFS_ILOCK_EXCL;
	xfs_ilock(ip, XFS_ILOCK_EXCL);
865
	xfs_trans_ijoin(tp, ip, 0);
C
Christoph Hellwig 已提交
866 867 868 869 870 871 872 873 874 875 876

	/*
	 * Only change the c/mtime if we are changing the size or we are
	 * explicitly asked to change it.  This handles the semantic difference
	 * between truncate() and ftruncate() as implemented in the VFS.
	 *
	 * The regular truncate() case without ATTR_CTIME and ATTR_MTIME is a
	 * special case where we need to update the times despite not having
	 * these flags set.  For all other operations the VFS set these flags
	 * explicitly if it wants a timestamp update.
	 */
877 878
	if (newsize != oldsize &&
	    !(iattr->ia_valid & (ATTR_CTIME | ATTR_MTIME))) {
C
Christoph Hellwig 已提交
879 880
		iattr->ia_ctime = iattr->ia_mtime =
			current_fs_time(inode->i_sb);
881
		iattr->ia_valid |= ATTR_CTIME | ATTR_MTIME;
C
Christoph Hellwig 已提交
882 883
	}

884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900
	/*
	 * The first thing we do is set the size to new_size permanently on
	 * disk.  This way we don't have to worry about anyone ever being able
	 * to look at the data being freed even in the face of a crash.
	 * What we're getting around here is the case where we free a block, it
	 * is allocated to another file, it is written to, and then we crash.
	 * If the new data gets written to the file but the log buffers
	 * containing the free and reallocation don't, then we'd end up with
	 * garbage in the blocks being freed.  As long as we make the new size
	 * permanent before actually freeing any blocks it doesn't matter if
	 * they get written to.
	 */
	ip->i_d.di_size = newsize;
	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);

	if (newsize <= oldsize) {
		error = xfs_itruncate_extents(&tp, ip, XFS_DATA_FORK, newsize);
C
Christoph Hellwig 已提交
901
		if (error)
902
			goto out_trans_cancel;
C
Christoph Hellwig 已提交
903 904 905 906 907 908 909 910 911

		/*
		 * Truncated "down", so we're removing references to old data
		 * here - if we delay flushing for a long time, we expose
		 * ourselves unduly to the notorious NULL files problem.  So,
		 * we mark this inode and flush it when the file is closed,
		 * and do not wait the usual (long) time for writeout.
		 */
		xfs_iflags_set(ip, XFS_ITRUNCATED);
912 913 914

		/* A truncate down always removes post-EOF blocks. */
		xfs_inode_clear_eofblocks_tag(ip);
C
Christoph Hellwig 已提交
915 916
	}

917
	if (iattr->ia_valid & ATTR_MODE)
918
		xfs_setattr_mode(ip, iattr);
919
	if (iattr->ia_valid & (ATTR_ATIME|ATTR_CTIME|ATTR_MTIME))
C
Christoph Hellwig 已提交
920
		xfs_setattr_time(ip, iattr);
C
Christoph Hellwig 已提交
921 922 923 924 925 926 927 928 929 930 931 932 933 934 935

	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);

	XFS_STATS_INC(xs_ig_attrchg);

	if (mp->m_flags & XFS_MOUNT_WSYNC)
		xfs_trans_set_sync(tp);

	error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
out_unlock:
	if (lock_flags)
		xfs_iunlock(ip, lock_flags);
	return error;

out_trans_cancel:
936
	xfs_trans_cancel(tp);
C
Christoph Hellwig 已提交
937 938 939
	goto out_unlock;
}

L
Linus Torvalds 已提交
940
STATIC int
941
xfs_vn_setattr(
942 943
	struct dentry		*dentry,
	struct iattr		*iattr)
L
Linus Torvalds 已提交
944
{
945
	struct xfs_inode	*ip = XFS_I(d_inode(dentry));
946 947 948
	int			error;

	if (iattr->ia_valid & ATTR_SIZE) {
949 950 951
		uint		iolock = XFS_IOLOCK_EXCL;

		xfs_ilock(ip, iolock);
952
		error = xfs_break_layouts(d_inode(dentry), &iolock, true);
953 954 955 956
		if (!error) {
			xfs_ilock(ip, XFS_MMAPLOCK_EXCL);
			iolock |= XFS_MMAPLOCK_EXCL;

957
			error = xfs_setattr_size(ip, iattr);
958
		}
959
		xfs_iunlock(ip, iolock);
960 961 962 963
	} else {
		error = xfs_setattr_nonsize(ip, iattr, 0);
	}

D
Dave Chinner 已提交
964
	return error;
L
Linus Torvalds 已提交
965 966
}

C
Christoph Hellwig 已提交
967 968 969 970 971 972 973 974 975 976 977 978 979 980
STATIC int
xfs_vn_update_time(
	struct inode		*inode,
	struct timespec		*now,
	int			flags)
{
	struct xfs_inode	*ip = XFS_I(inode);
	struct xfs_mount	*mp = ip->i_mount;
	struct xfs_trans	*tp;
	int			error;

	trace_xfs_update_time(ip);

	tp = xfs_trans_alloc(mp, XFS_TRANS_FSYNC_TS);
981
	error = xfs_trans_reserve(tp, &M_RES(mp)->tr_fsyncts, 0, 0);
C
Christoph Hellwig 已提交
982
	if (error) {
983
		xfs_trans_cancel(tp);
D
Dave Chinner 已提交
984
		return error;
C
Christoph Hellwig 已提交
985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004
	}

	xfs_ilock(ip, XFS_ILOCK_EXCL);
	if (flags & S_CTIME) {
		inode->i_ctime = *now;
		ip->i_d.di_ctime.t_sec = (__int32_t)now->tv_sec;
		ip->i_d.di_ctime.t_nsec = (__int32_t)now->tv_nsec;
	}
	if (flags & S_MTIME) {
		inode->i_mtime = *now;
		ip->i_d.di_mtime.t_sec = (__int32_t)now->tv_sec;
		ip->i_d.di_mtime.t_nsec = (__int32_t)now->tv_nsec;
	}
	if (flags & S_ATIME) {
		inode->i_atime = *now;
		ip->i_d.di_atime.t_sec = (__int32_t)now->tv_sec;
		ip->i_d.di_atime.t_nsec = (__int32_t)now->tv_nsec;
	}
	xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
	xfs_trans_log_inode(tp, ip, XFS_ILOG_TIMESTAMP);
D
Dave Chinner 已提交
1005
	return xfs_trans_commit(tp, 0);
C
Christoph Hellwig 已提交
1006 1007
}

E
Eric Sandeen 已提交
1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035
#define XFS_FIEMAP_FLAGS	(FIEMAP_FLAG_SYNC|FIEMAP_FLAG_XATTR)

/*
 * Call fiemap helper to fill in user data.
 * Returns positive errors to xfs_getbmap.
 */
STATIC int
xfs_fiemap_format(
	void			**arg,
	struct getbmapx		*bmv,
	int			*full)
{
	int			error;
	struct fiemap_extent_info *fieinfo = *arg;
	u32			fiemap_flags = 0;
	u64			logical, physical, length;

	/* Do nothing for a hole */
	if (bmv->bmv_block == -1LL)
		return 0;

	logical = BBTOB(bmv->bmv_offset);
	physical = BBTOB(bmv->bmv_block);
	length = BBTOB(bmv->bmv_length);

	if (bmv->bmv_oflags & BMV_OF_PREALLOC)
		fiemap_flags |= FIEMAP_EXTENT_UNWRITTEN;
	else if (bmv->bmv_oflags & BMV_OF_DELALLOC) {
1036 1037
		fiemap_flags |= (FIEMAP_EXTENT_DELALLOC |
				 FIEMAP_EXTENT_UNKNOWN);
E
Eric Sandeen 已提交
1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049
		physical = 0;   /* no block yet */
	}
	if (bmv->bmv_oflags & BMV_OF_LAST)
		fiemap_flags |= FIEMAP_EXTENT_LAST;

	error = fiemap_fill_next_extent(fieinfo, logical, physical,
					length, fiemap_flags);
	if (error > 0) {
		error = 0;
		*full = 1;	/* user array now full */
	}

D
Dave Chinner 已提交
1050
	return error;
E
Eric Sandeen 已提交
1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068
}

STATIC int
xfs_vn_fiemap(
	struct inode		*inode,
	struct fiemap_extent_info *fieinfo,
	u64			start,
	u64			length)
{
	xfs_inode_t		*ip = XFS_I(inode);
	struct getbmapx		bm;
	int			error;

	error = fiemap_check_flags(fieinfo, XFS_FIEMAP_FLAGS);
	if (error)
		return error;

	/* Set up bmap header for xfs internal routine */
1069
	bm.bmv_offset = BTOBBT(start);
E
Eric Sandeen 已提交
1070 1071 1072 1073
	/* Special case for whole file */
	if (length == FIEMAP_MAX_OFFSET)
		bm.bmv_length = -1LL;
	else
1074
		bm.bmv_length = BTOBB(start + length) - bm.bmv_offset;
E
Eric Sandeen 已提交
1075

1076
	/* We add one because in getbmap world count includes the header */
T
Tao Ma 已提交
1077 1078 1079 1080
	bm.bmv_count = !fieinfo->fi_extents_max ? MAXEXTNUM :
					fieinfo->fi_extents_max + 1;
	bm.bmv_count = min_t(__s32, bm.bmv_count,
			     (PAGE_SIZE * 16 / sizeof(struct getbmapx)));
T
Tao Ma 已提交
1081
	bm.bmv_iflags = BMV_IF_PREALLOC | BMV_IF_NO_HOLES;
E
Eric Sandeen 已提交
1082 1083 1084 1085 1086 1087 1088
	if (fieinfo->fi_flags & FIEMAP_FLAG_XATTR)
		bm.bmv_iflags |= BMV_IF_ATTRFORK;
	if (!(fieinfo->fi_flags & FIEMAP_FLAG_SYNC))
		bm.bmv_iflags |= BMV_IF_DELALLOC;

	error = xfs_getbmap(ip, &bm, xfs_fiemap_format, fieinfo);
	if (error)
D
Dave Chinner 已提交
1089
		return error;
E
Eric Sandeen 已提交
1090 1091 1092 1093

	return 0;
}

Z
Zhi Yong Wu 已提交
1094 1095 1096 1097 1098 1099
STATIC int
xfs_vn_tmpfile(
	struct inode	*dir,
	struct dentry	*dentry,
	umode_t		mode)
{
1100
	return xfs_generic_create(dir, dentry, mode, 0, true);
Z
Zhi Yong Wu 已提交
1101 1102
}

1103
static const struct inode_operations xfs_inode_operations = {
1104
	.get_acl		= xfs_get_acl,
1105
	.set_acl		= xfs_set_acl,
1106 1107
	.getattr		= xfs_vn_getattr,
	.setattr		= xfs_vn_setattr,
1108 1109 1110
	.setxattr		= generic_setxattr,
	.getxattr		= generic_getxattr,
	.removexattr		= generic_removexattr,
1111
	.listxattr		= xfs_vn_listxattr,
E
Eric Sandeen 已提交
1112
	.fiemap			= xfs_vn_fiemap,
C
Christoph Hellwig 已提交
1113
	.update_time		= xfs_vn_update_time,
L
Linus Torvalds 已提交
1114 1115
};

1116
static const struct inode_operations xfs_dir_inode_operations = {
1117 1118 1119 1120 1121 1122
	.create			= xfs_vn_create,
	.lookup			= xfs_vn_lookup,
	.link			= xfs_vn_link,
	.unlink			= xfs_vn_unlink,
	.symlink		= xfs_vn_symlink,
	.mkdir			= xfs_vn_mkdir,
1123 1124 1125 1126 1127 1128 1129
	/*
	 * Yes, XFS uses the same method for rmdir and unlink.
	 *
	 * There are some subtile differences deeper in the code,
	 * but we use S_ISDIR to check for those.
	 */
	.rmdir			= xfs_vn_unlink,
1130
	.mknod			= xfs_vn_mknod,
1131
	.rename2		= xfs_vn_rename,
1132
	.get_acl		= xfs_get_acl,
1133
	.set_acl		= xfs_set_acl,
1134 1135
	.getattr		= xfs_vn_getattr,
	.setattr		= xfs_vn_setattr,
1136 1137 1138
	.setxattr		= generic_setxattr,
	.getxattr		= generic_getxattr,
	.removexattr		= generic_removexattr,
1139
	.listxattr		= xfs_vn_listxattr,
C
Christoph Hellwig 已提交
1140
	.update_time		= xfs_vn_update_time,
Z
Zhi Yong Wu 已提交
1141
	.tmpfile		= xfs_vn_tmpfile,
L
Linus Torvalds 已提交
1142 1143
};

1144
static const struct inode_operations xfs_dir_ci_inode_operations = {
1145 1146 1147 1148 1149 1150
	.create			= xfs_vn_create,
	.lookup			= xfs_vn_ci_lookup,
	.link			= xfs_vn_link,
	.unlink			= xfs_vn_unlink,
	.symlink		= xfs_vn_symlink,
	.mkdir			= xfs_vn_mkdir,
1151 1152 1153 1154 1155 1156 1157
	/*
	 * Yes, XFS uses the same method for rmdir and unlink.
	 *
	 * There are some subtile differences deeper in the code,
	 * but we use S_ISDIR to check for those.
	 */
	.rmdir			= xfs_vn_unlink,
1158
	.mknod			= xfs_vn_mknod,
1159
	.rename2		= xfs_vn_rename,
1160
	.get_acl		= xfs_get_acl,
1161
	.set_acl		= xfs_set_acl,
1162 1163
	.getattr		= xfs_vn_getattr,
	.setattr		= xfs_vn_setattr,
1164 1165 1166
	.setxattr		= generic_setxattr,
	.getxattr		= generic_getxattr,
	.removexattr		= generic_removexattr,
1167
	.listxattr		= xfs_vn_listxattr,
C
Christoph Hellwig 已提交
1168
	.update_time		= xfs_vn_update_time,
Z
Zhi Yong Wu 已提交
1169
	.tmpfile		= xfs_vn_tmpfile,
1170 1171
};

1172
static const struct inode_operations xfs_symlink_inode_operations = {
L
Linus Torvalds 已提交
1173
	.readlink		= generic_readlink,
1174
	.follow_link		= xfs_vn_follow_link,
A
Al Viro 已提交
1175
	.put_link		= kfree_put_link,
1176 1177
	.getattr		= xfs_vn_getattr,
	.setattr		= xfs_vn_setattr,
1178 1179 1180
	.setxattr		= generic_setxattr,
	.getxattr		= generic_getxattr,
	.removexattr		= generic_removexattr,
1181
	.listxattr		= xfs_vn_listxattr,
C
Christoph Hellwig 已提交
1182
	.update_time		= xfs_vn_update_time,
L
Linus Torvalds 已提交
1183
};
1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208

STATIC void
xfs_diflags_to_iflags(
	struct inode		*inode,
	struct xfs_inode	*ip)
{
	if (ip->i_d.di_flags & XFS_DIFLAG_IMMUTABLE)
		inode->i_flags |= S_IMMUTABLE;
	else
		inode->i_flags &= ~S_IMMUTABLE;
	if (ip->i_d.di_flags & XFS_DIFLAG_APPEND)
		inode->i_flags |= S_APPEND;
	else
		inode->i_flags &= ~S_APPEND;
	if (ip->i_d.di_flags & XFS_DIFLAG_SYNC)
		inode->i_flags |= S_SYNC;
	else
		inode->i_flags &= ~S_SYNC;
	if (ip->i_d.di_flags & XFS_DIFLAG_NOATIME)
		inode->i_flags |= S_NOATIME;
	else
		inode->i_flags &= ~S_NOATIME;
}

/*
1209
 * Initialize the Linux inode and set up the operation vectors.
1210
 *
1211 1212 1213 1214
 * When reading existing inodes from disk this is called directly from xfs_iget,
 * when creating a new inode it is called from xfs_ialloc after setting up the
 * inode. These callers have different criteria for clearing XFS_INEW, so leave
 * it up to the caller to deal with unlocking the inode appropriately.
1215 1216 1217 1218 1219
 */
void
xfs_setup_inode(
	struct xfs_inode	*ip)
{
1220
	struct inode		*inode = &ip->i_vnode;
1221
	gfp_t			gfp_mask;
1222 1223

	inode->i_ino = ip->i_ino;
C
Christoph Hellwig 已提交
1224
	inode->i_state = I_NEW;
1225 1226

	inode_sb_list_add(inode);
C
Christoph Hellwig 已提交
1227 1228
	/* make the inode look hashed for the writeback code */
	hlist_add_fake(&inode->i_hash);
1229 1230

	inode->i_mode	= ip->i_d.di_mode;
M
Miklos Szeredi 已提交
1231
	set_nlink(inode, ip->i_d.di_nlink);
1232 1233
	inode->i_uid    = xfs_uid_to_kuid(ip->i_d.di_uid);
	inode->i_gid    = xfs_gid_to_kgid(ip->i_d.di_gid);
1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256

	switch (inode->i_mode & S_IFMT) {
	case S_IFBLK:
	case S_IFCHR:
		inode->i_rdev =
			MKDEV(sysv_major(ip->i_df.if_u2.if_rdev) & 0x1ff,
			      sysv_minor(ip->i_df.if_u2.if_rdev));
		break;
	default:
		inode->i_rdev = 0;
		break;
	}

	inode->i_generation = ip->i_d.di_gen;
	i_size_write(inode, ip->i_d.di_size);
	inode->i_atime.tv_sec	= ip->i_d.di_atime.t_sec;
	inode->i_atime.tv_nsec	= ip->i_d.di_atime.t_nsec;
	inode->i_mtime.tv_sec	= ip->i_d.di_mtime.t_sec;
	inode->i_mtime.tv_nsec	= ip->i_d.di_mtime.t_nsec;
	inode->i_ctime.tv_sec	= ip->i_d.di_ctime.t_sec;
	inode->i_ctime.tv_nsec	= ip->i_d.di_ctime.t_nsec;
	xfs_diflags_to_iflags(inode, ip);

D
Dave Chinner 已提交
1257
	ip->d_ops = ip->i_mount->m_nondir_inode_ops;
1258
	lockdep_set_class(&ip->i_lock.mr_lock, &xfs_nondir_ilock_class);
1259 1260 1261 1262 1263 1264 1265
	switch (inode->i_mode & S_IFMT) {
	case S_IFREG:
		inode->i_op = &xfs_inode_operations;
		inode->i_fop = &xfs_file_operations;
		inode->i_mapping->a_ops = &xfs_address_space_operations;
		break;
	case S_IFDIR:
1266
		lockdep_set_class(&ip->i_lock.mr_lock, &xfs_dir_ilock_class);
1267 1268 1269 1270 1271
		if (xfs_sb_version_hasasciici(&XFS_M(inode->i_sb)->m_sb))
			inode->i_op = &xfs_dir_ci_inode_operations;
		else
			inode->i_op = &xfs_dir_inode_operations;
		inode->i_fop = &xfs_dir_file_operations;
1272
		ip->d_ops = ip->i_mount->m_dir_inode_ops;
1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284
		break;
	case S_IFLNK:
		inode->i_op = &xfs_symlink_inode_operations;
		if (!(ip->i_df.if_flags & XFS_IFINLINE))
			inode->i_mapping->a_ops = &xfs_address_space_operations;
		break;
	default:
		inode->i_op = &xfs_inode_operations;
		init_special_inode(inode, inode->i_mode, inode->i_rdev);
		break;
	}

1285 1286 1287 1288 1289 1290 1291 1292
	/*
	 * Ensure all page cache allocations are done from GFP_NOFS context to
	 * prevent direct reclaim recursion back into the filesystem and blowing
	 * stacks or deadlocking.
	 */
	gfp_mask = mapping_gfp_mask(inode->i_mapping);
	mapping_set_gfp_mask(inode->i_mapping, (gfp_mask & ~(__GFP_FS)));

1293 1294 1295 1296 1297 1298
	/*
	 * If there is no attribute fork no ACL can exist on this inode,
	 * and it can't have any file capabilities attached to it either.
	 */
	if (!XFS_IFORK_Q(ip)) {
		inode_has_no_xattr(inode);
1299
		cache_no_acl(inode);
1300
	}
1301
}