open.c 29.1 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9
/*
 *  linux/fs/open.c
 *
 *  Copyright (C) 1991, 1992  Linus Torvalds
 */

#include <linux/string.h>
#include <linux/mm.h>
#include <linux/file.h>
A
Al Viro 已提交
10
#include <linux/fdtable.h>
R
Robert Love 已提交
11
#include <linux/fsnotify.h>
L
Linus Torvalds 已提交
12 13 14 15
#include <linux/module.h>
#include <linux/tty.h>
#include <linux/namei.h>
#include <linux/backing-dev.h>
16
#include <linux/capability.h>
17
#include <linux/securebits.h>
L
Linus Torvalds 已提交
18 19
#include <linux/security.h>
#include <linux/mount.h>
20
#include <linux/fcntl.h>
21
#include <linux/slab.h>
22
#include <linux/uaccess.h>
L
Linus Torvalds 已提交
23
#include <linux/fs.h>
24
#include <linux/personality.h>
L
Linus Torvalds 已提交
25 26
#include <linux/pagemap.h>
#include <linux/syscalls.h>
27
#include <linux/rcupdate.h>
28
#include <linux/audit.h>
29
#include <linux/falloc.h>
30
#include <linux/fs_struct.h>
31
#include <linux/ima.h>
32
#include <linux/dnotify.h>
A
Al Viro 已提交
33
#include <linux/compat.h>
L
Linus Torvalds 已提交
34

35 36
#include "internal.h"

37 38
int do_truncate(struct dentry *dentry, loff_t length, unsigned int time_attrs,
	struct file *filp)
L
Linus Torvalds 已提交
39
{
40
	int ret;
L
Linus Torvalds 已提交
41 42 43 44 45 46 47
	struct iattr newattrs;

	/* Not pretty: "inode->i_size" shouldn't really be signed. But it is. */
	if (length < 0)
		return -EINVAL;

	newattrs.ia_size = length;
48
	newattrs.ia_valid = ATTR_SIZE | time_attrs;
49 50 51 52
	if (filp) {
		newattrs.ia_file = filp;
		newattrs.ia_valid |= ATTR_FILE;
	}
L
Linus Torvalds 已提交
53

54 55 56 57
	/* Remove suid, sgid, and file capabilities on truncate too */
	ret = dentry_needs_remove_privs(dentry);
	if (ret < 0)
		return ret;
58 59
	if (ret)
		newattrs.ia_valid |= ret | ATTR_FORCE;
60

A
Al Viro 已提交
61
	inode_lock(dentry->d_inode);
62 63
	/* Note any delegations or leases have already been broken: */
	ret = notify_change(dentry, &newattrs, NULL);
A
Al Viro 已提交
64
	inode_unlock(dentry->d_inode);
65
	return ret;
L
Linus Torvalds 已提交
66 67
}

A
Al Viro 已提交
68
long vfs_truncate(const struct path *path, loff_t length)
L
Linus Torvalds 已提交
69
{
70
	struct inode *inode;
71
	struct dentry *upperdentry;
72
	long error;
L
Linus Torvalds 已提交
73

74
	inode = path->dentry->d_inode;
L
Linus Torvalds 已提交
75 76 77

	/* For directories it's -EISDIR, for other non-regulars - -EINVAL */
	if (S_ISDIR(inode->i_mode))
78
		return -EISDIR;
L
Linus Torvalds 已提交
79
	if (!S_ISREG(inode->i_mode))
80
		return -EINVAL;
L
Linus Torvalds 已提交
81

82
	error = mnt_want_write(path->mnt);
L
Linus Torvalds 已提交
83
	if (error)
84
		goto out;
L
Linus Torvalds 已提交
85

86
	error = inode_permission(inode, MAY_WRITE);
87 88
	if (error)
		goto mnt_drop_write_and_out;
L
Linus Torvalds 已提交
89 90

	error = -EPERM;
91
	if (IS_APPEND(inode))
92
		goto mnt_drop_write_and_out;
L
Linus Torvalds 已提交
93

94 95 96 97 98
	/*
	 * If this is an overlayfs then do as if opening the file so we get
	 * write access on the upper inode, not on the overlay inode.  For
	 * non-overlay filesystems d_real() is an identity function.
	 */
M
Miklos Szeredi 已提交
99
	upperdentry = d_real(path->dentry, NULL, O_WRONLY, 0);
100 101 102 103 104
	error = PTR_ERR(upperdentry);
	if (IS_ERR(upperdentry))
		goto mnt_drop_write_and_out;

	error = get_write_access(upperdentry->d_inode);
L
Linus Torvalds 已提交
105
	if (error)
106
		goto mnt_drop_write_and_out;
L
Linus Torvalds 已提交
107

108 109 110 111
	/*
	 * Make sure that there are no leases.  get_write_access() protects
	 * against the truncate racing with a lease-granting setlease().
	 */
112
	error = break_lease(inode, O_WRONLY);
L
Linus Torvalds 已提交
113
	if (error)
114
		goto put_write_and_out;
L
Linus Torvalds 已提交
115 116

	error = locks_verify_truncate(inode, NULL, length);
117
	if (!error)
118
		error = security_path_truncate(path);
119
	if (!error)
120
		error = do_truncate(path->dentry, length, 0, NULL);
L
Linus Torvalds 已提交
121

122
put_write_and_out:
123
	put_write_access(upperdentry->d_inode);
124
mnt_drop_write_and_out:
125
	mnt_drop_write(path->mnt);
L
Linus Torvalds 已提交
126 127 128
out:
	return error;
}
129 130 131 132
EXPORT_SYMBOL_GPL(vfs_truncate);

static long do_sys_truncate(const char __user *pathname, loff_t length)
{
133
	unsigned int lookup_flags = LOOKUP_FOLLOW;
134 135 136 137 138 139
	struct path path;
	int error;

	if (length < 0)	/* sorry, but loff_t says... */
		return -EINVAL;

140 141
retry:
	error = user_path_at(AT_FDCWD, pathname, lookup_flags, &path);
142 143 144 145
	if (!error) {
		error = vfs_truncate(&path, length);
		path_put(&path);
	}
146 147 148 149
	if (retry_estale(error, lookup_flags)) {
		lookup_flags |= LOOKUP_REVAL;
		goto retry;
	}
150 151
	return error;
}
L
Linus Torvalds 已提交
152

153
SYSCALL_DEFINE2(truncate, const char __user *, path, long, length)
L
Linus Torvalds 已提交
154
{
155
	return do_sys_truncate(path, length);
L
Linus Torvalds 已提交
156 157
}

A
Al Viro 已提交
158 159 160 161 162 163 164
#ifdef CONFIG_COMPAT
COMPAT_SYSCALL_DEFINE2(truncate, const char __user *, path, compat_off_t, length)
{
	return do_sys_truncate(path, length);
}
#endif

165
static long do_sys_ftruncate(unsigned int fd, loff_t length, int small)
L
Linus Torvalds 已提交
166
{
A
Al Viro 已提交
167
	struct inode *inode;
L
Linus Torvalds 已提交
168
	struct dentry *dentry;
169
	struct fd f;
L
Linus Torvalds 已提交
170 171 172 173 174 175
	int error;

	error = -EINVAL;
	if (length < 0)
		goto out;
	error = -EBADF;
176 177
	f = fdget(fd);
	if (!f.file)
L
Linus Torvalds 已提交
178 179 180
		goto out;

	/* explicitly opened as large or we are on 64-bit box */
181
	if (f.file->f_flags & O_LARGEFILE)
L
Linus Torvalds 已提交
182 183
		small = 0;

184
	dentry = f.file->f_path.dentry;
L
Linus Torvalds 已提交
185 186
	inode = dentry->d_inode;
	error = -EINVAL;
187
	if (!S_ISREG(inode->i_mode) || !(f.file->f_mode & FMODE_WRITE))
L
Linus Torvalds 已提交
188 189 190 191 192 193 194 195
		goto out_putf;

	error = -EINVAL;
	/* Cannot ftruncate over 2^31 bytes without large file support */
	if (small && length > MAX_NON_LFS)
		goto out_putf;

	error = -EPERM;
196 197
	/* Check IS_APPEND on real upper inode */
	if (IS_APPEND(file_inode(f.file)))
L
Linus Torvalds 已提交
198 199
		goto out_putf;

200
	sb_start_write(inode->i_sb);
201
	error = locks_verify_truncate(inode, f.file, length);
202
	if (!error)
203
		error = security_path_truncate(&f.file->f_path);
L
Linus Torvalds 已提交
204
	if (!error)
205
		error = do_truncate(dentry, length, ATTR_MTIME|ATTR_CTIME, f.file);
206
	sb_end_write(inode->i_sb);
L
Linus Torvalds 已提交
207
out_putf:
208
	fdput(f);
L
Linus Torvalds 已提交
209 210 211 212
out:
	return error;
}

213
SYSCALL_DEFINE2(ftruncate, unsigned int, fd, unsigned long, length)
L
Linus Torvalds 已提交
214
{
215
	return do_sys_ftruncate(fd, length, 1);
L
Linus Torvalds 已提交
216 217
}

A
Al Viro 已提交
218 219 220 221 222 223 224
#ifdef CONFIG_COMPAT
COMPAT_SYSCALL_DEFINE2(ftruncate, unsigned int, fd, compat_ulong_t, length)
{
	return do_sys_ftruncate(fd, length, 1);
}
#endif

L
Linus Torvalds 已提交
225 226
/* LFS versions of truncate are only needed on 32 bit machines */
#if BITS_PER_LONG == 32
227
SYSCALL_DEFINE2(truncate64, const char __user *, path, loff_t, length)
L
Linus Torvalds 已提交
228 229 230 231
{
	return do_sys_truncate(path, length);
}

232
SYSCALL_DEFINE2(ftruncate64, unsigned int, fd, loff_t, length)
L
Linus Torvalds 已提交
233
{
234
	return do_sys_ftruncate(fd, length, 0);
L
Linus Torvalds 已提交
235
}
236
#endif /* BITS_PER_LONG == 32 */
L
Linus Torvalds 已提交
237

238

239
int vfs_fallocate(struct file *file, int mode, loff_t offset, loff_t len)
240
{
A
Al Viro 已提交
241
	struct inode *inode = file_inode(file);
242
	long ret;
243 244

	if (offset < 0 || len <= 0)
245
		return -EINVAL;
246 247

	/* Return error if mode is not supported */
248
	if (mode & ~FALLOC_FL_SUPPORTED_MASK)
249 250 251 252 253
		return -EOPNOTSUPP;

	/* Punch hole and zero range are mutually exclusive */
	if ((mode & (FALLOC_FL_PUNCH_HOLE | FALLOC_FL_ZERO_RANGE)) ==
	    (FALLOC_FL_PUNCH_HOLE | FALLOC_FL_ZERO_RANGE))
J
Josef Bacik 已提交
254 255 256 257 258
		return -EOPNOTSUPP;

	/* Punch hole must have keep size set */
	if ((mode & FALLOC_FL_PUNCH_HOLE) &&
	    !(mode & FALLOC_FL_KEEP_SIZE))
259
		return -EOPNOTSUPP;
260

261 262 263 264 265
	/* Collapse range should only be used exclusively. */
	if ((mode & FALLOC_FL_COLLAPSE_RANGE) &&
	    (mode & ~FALLOC_FL_COLLAPSE_RANGE))
		return -EINVAL;

266 267 268 269 270
	/* Insert range should only be used exclusively. */
	if ((mode & FALLOC_FL_INSERT_RANGE) &&
	    (mode & ~FALLOC_FL_INSERT_RANGE))
		return -EINVAL;

271 272 273 274 275
	/* Unshare range should only be used with allocate mode. */
	if ((mode & FALLOC_FL_UNSHARE_RANGE) &&
	    (mode & ~(FALLOC_FL_UNSHARE_RANGE | FALLOC_FL_KEEP_SIZE)))
		return -EINVAL;

276
	if (!(file->f_mode & FMODE_WRITE))
277
		return -EBADF;
278

279
	/*
280
	 * We can only allow pure fallocate on append only files
281
	 */
282
	if ((mode & ~FALLOC_FL_KEEP_SIZE) && IS_APPEND(inode))
283 284 285 286 287
		return -EPERM;

	if (IS_IMMUTABLE(inode))
		return -EPERM;

288
	/*
289
	 * We cannot allow any fallocate operation on an active swapfile
290 291
	 */
	if (IS_SWAPFILE(inode))
292
		return -ETXTBSY;
293

294 295 296 297 298 299
	/*
	 * Revalidate the write permissions, in case security policy has
	 * changed since the files were opened.
	 */
	ret = security_file_permission(file, MAY_WRITE);
	if (ret)
300
		return ret;
301 302

	if (S_ISFIFO(inode->i_mode))
303
		return -ESPIPE;
304

305 306 307 308
	if (S_ISDIR(inode->i_mode))
		return -EISDIR;

	if (!S_ISREG(inode->i_mode) && !S_ISBLK(inode->i_mode))
309
		return -ENODEV;
310 311 312

	/* Check for wrap through zero too */
	if (((offset + len) > inode->i_sb->s_maxbytes) || ((offset + len) < 0))
313
		return -EFBIG;
314

315
	if (!file->f_op->fallocate)
316
		return -EOPNOTSUPP;
317

318
	file_start_write(file);
319
	ret = file->f_op->fallocate(file, mode, offset, len);
320 321 322 323 324 325 326 327 328 329 330

	/*
	 * Create inotify and fanotify events.
	 *
	 * To keep the logic simple always create events if fallocate succeeds.
	 * This implies that events are even created if the file size remains
	 * unchanged, e.g. when using flag FALLOC_FL_KEEP_SIZE.
	 */
	if (ret == 0)
		fsnotify_modify(file);

331
	file_end_write(file);
332
	return ret;
333
}
334
EXPORT_SYMBOL_GPL(vfs_fallocate);
335

336
SYSCALL_DEFINE4(fallocate, int, fd, int, mode, loff_t, offset, loff_t, len)
337
{
338
	struct fd f = fdget(fd);
339 340
	int error = -EBADF;

341
	if (f.file) {
342
		error = vfs_fallocate(f.file, mode, offset, len);
343
		fdput(f);
344 345
	}
	return error;
346
}
347

L
Linus Torvalds 已提交
348 349 350 351 352
/*
 * access() needs to use the real uid/gid, not the effective uid/gid.
 * We do this by temporarily clearing all FS-related capabilities and
 * switching the fsuid/fsgid around to the real ones.
 */
353
long do_faccessat(int dfd, const char __user *filename, int mode)
L
Linus Torvalds 已提交
354
{
D
David Howells 已提交
355 356
	const struct cred *old_cred;
	struct cred *override_cred;
357
	struct path path;
358
	struct inode *inode;
L
Linus Torvalds 已提交
359
	int res;
360
	unsigned int lookup_flags = LOOKUP_FOLLOW;
L
Linus Torvalds 已提交
361 362 363 364

	if (mode & ~S_IRWXO)	/* where's F_OK, X_OK, W_OK, R_OK? */
		return -EINVAL;

D
David Howells 已提交
365 366 367
	override_cred = prepare_creds();
	if (!override_cred)
		return -ENOMEM;
L
Linus Torvalds 已提交
368

D
David Howells 已提交
369 370
	override_cred->fsuid = override_cred->uid;
	override_cred->fsgid = override_cred->gid;
L
Linus Torvalds 已提交
371

372
	if (!issecure(SECURE_NO_SETUID_FIXUP)) {
D
David Howells 已提交
373
		/* Clear the capabilities if we switch to a non-root user */
374 375
		kuid_t root_uid = make_kuid(override_cred->user_ns, 0);
		if (!uid_eq(override_cred->uid, root_uid))
D
David Howells 已提交
376
			cap_clear(override_cred->cap_effective);
377
		else
D
David Howells 已提交
378 379
			override_cred->cap_effective =
				override_cred->cap_permitted;
380
	}
L
Linus Torvalds 已提交
381

D
David Howells 已提交
382
	old_cred = override_creds(override_cred);
383 384
retry:
	res = user_path_at(dfd, filename, lookup_flags, &path);
385 386 387
	if (res)
		goto out;

388
	inode = d_backing_inode(path.dentry);
389 390

	if ((mode & MAY_EXEC) && S_ISREG(inode->i_mode)) {
391 392 393 394 395
		/*
		 * MAY_EXEC on regular files is denied if the fs is mounted
		 * with the "noexec" flag.
		 */
		res = -EACCES;
396
		if (path_noexec(&path))
397 398 399
			goto out_path_release;
	}

400
	res = inode_permission(inode, mode | MAY_ACCESS);
401
	/* SuS v2 requires we report a read only fs too */
402
	if (res || !(mode & S_IWOTH) || special_file(inode->i_mode))
403
		goto out_path_release;
404 405 406 407 408 409 410 411 412 413
	/*
	 * This is a rare case where using __mnt_is_readonly()
	 * is OK without a mnt_want/drop_write() pair.  Since
	 * no actual write to the fs is performed here, we do
	 * not need to telegraph to that to anyone.
	 *
	 * By doing this, we accept that this access is
	 * inherently racy and know that the fs may change
	 * state before we even see this result.
	 */
414
	if (__mnt_is_readonly(path.mnt))
415
		res = -EROFS;
L
Linus Torvalds 已提交
416

417
out_path_release:
418
	path_put(&path);
419 420 421 422
	if (retry_estale(res, lookup_flags)) {
		lookup_flags |= LOOKUP_REVAL;
		goto retry;
	}
423
out:
D
David Howells 已提交
424 425
	revert_creds(old_cred);
	put_cred(override_cred);
L
Linus Torvalds 已提交
426 427 428
	return res;
}

429 430 431 432 433
SYSCALL_DEFINE3(faccessat, int, dfd, const char __user *, filename, int, mode)
{
	return do_faccessat(dfd, filename, mode);
}

434
SYSCALL_DEFINE2(access, const char __user *, filename, int, mode)
435
{
436
	return do_faccessat(AT_FDCWD, filename, mode);
437 438
}

439
int ksys_chdir(const char __user *filename)
L
Linus Torvalds 已提交
440
{
441
	struct path path;
L
Linus Torvalds 已提交
442
	int error;
443 444 445
	unsigned int lookup_flags = LOOKUP_FOLLOW | LOOKUP_DIRECTORY;
retry:
	error = user_path_at(AT_FDCWD, filename, lookup_flags, &path);
L
Linus Torvalds 已提交
446 447 448
	if (error)
		goto out;

E
Eric Paris 已提交
449
	error = inode_permission(path.dentry->d_inode, MAY_EXEC | MAY_CHDIR);
L
Linus Torvalds 已提交
450 451 452
	if (error)
		goto dput_and_out;

453
	set_fs_pwd(current->fs, &path);
L
Linus Torvalds 已提交
454 455

dput_and_out:
456
	path_put(&path);
457 458 459 460
	if (retry_estale(error, lookup_flags)) {
		lookup_flags |= LOOKUP_REVAL;
		goto retry;
	}
L
Linus Torvalds 已提交
461 462 463 464
out:
	return error;
}

465 466 467 468 469
SYSCALL_DEFINE1(chdir, const char __user *, filename)
{
	return ksys_chdir(filename);
}

470
SYSCALL_DEFINE1(fchdir, unsigned int, fd)
L
Linus Torvalds 已提交
471
{
472
	struct fd f = fdget_raw(fd);
473
	int error;
L
Linus Torvalds 已提交
474 475

	error = -EBADF;
476
	if (!f.file)
L
Linus Torvalds 已提交
477 478 479
		goto out;

	error = -ENOTDIR;
480
	if (!d_can_lookup(f.file->f_path.dentry))
L
Linus Torvalds 已提交
481 482
		goto out_putf;

483
	error = inode_permission(file_inode(f.file), MAY_EXEC | MAY_CHDIR);
L
Linus Torvalds 已提交
484
	if (!error)
485
		set_fs_pwd(current->fs, &f.file->f_path);
L
Linus Torvalds 已提交
486
out_putf:
487
	fdput(f);
L
Linus Torvalds 已提交
488 489 490 491
out:
	return error;
}

492
int ksys_chroot(const char __user *filename)
L
Linus Torvalds 已提交
493
{
494
	struct path path;
L
Linus Torvalds 已提交
495
	int error;
496 497 498
	unsigned int lookup_flags = LOOKUP_FOLLOW | LOOKUP_DIRECTORY;
retry:
	error = user_path_at(AT_FDCWD, filename, lookup_flags, &path);
L
Linus Torvalds 已提交
499 500 501
	if (error)
		goto out;

E
Eric Paris 已提交
502
	error = inode_permission(path.dentry->d_inode, MAY_EXEC | MAY_CHDIR);
L
Linus Torvalds 已提交
503 504 505 506
	if (error)
		goto dput_and_out;

	error = -EPERM;
507
	if (!ns_capable(current_user_ns(), CAP_SYS_CHROOT))
L
Linus Torvalds 已提交
508
		goto dput_and_out;
T
Tetsuo Handa 已提交
509 510 511
	error = security_path_chroot(&path);
	if (error)
		goto dput_and_out;
L
Linus Torvalds 已提交
512

513
	set_fs_root(current->fs, &path);
L
Linus Torvalds 已提交
514 515
	error = 0;
dput_and_out:
516
	path_put(&path);
517 518 519 520
	if (retry_estale(error, lookup_flags)) {
		lookup_flags |= LOOKUP_REVAL;
		goto retry;
	}
L
Linus Torvalds 已提交
521 522 523 524
out:
	return error;
}

525 526 527 528 529
SYSCALL_DEFINE1(chroot, const char __user *, filename)
{
	return ksys_chroot(filename);
}

530
static int chmod_common(const struct path *path, umode_t mode)
L
Linus Torvalds 已提交
531
{
532
	struct inode *inode = path->dentry->d_inode;
533
	struct inode *delegated_inode = NULL;
L
Linus Torvalds 已提交
534
	struct iattr newattrs;
535
	int error;
L
Linus Torvalds 已提交
536

537 538 539
	error = mnt_want_write(path->mnt);
	if (error)
		return error;
540
retry_deleg:
A
Al Viro 已提交
541
	inode_lock(inode);
542
	error = security_path_chmod(path, mode);
543
	if (error)
544
		goto out_unlock;
L
Linus Torvalds 已提交
545 546
	newattrs.ia_mode = (mode & S_IALLUGO) | (inode->i_mode & ~S_IALLUGO);
	newattrs.ia_valid = ATTR_MODE | ATTR_CTIME;
547
	error = notify_change(path->dentry, &newattrs, &delegated_inode);
548
out_unlock:
A
Al Viro 已提交
549
	inode_unlock(inode);
550 551 552 553 554
	if (delegated_inode) {
		error = break_deleg_wait(&delegated_inode);
		if (!error)
			goto retry_deleg;
	}
555 556 557 558
	mnt_drop_write(path->mnt);
	return error;
}

559
int ksys_fchmod(unsigned int fd, umode_t mode)
560
{
A
Al Viro 已提交
561
	struct fd f = fdget(fd);
562 563
	int err = -EBADF;

A
Al Viro 已提交
564
	if (f.file) {
A
Al Viro 已提交
565
		audit_file(f.file);
A
Al Viro 已提交
566 567
		err = chmod_common(&f.file->f_path, mode);
		fdput(f);
568
	}
L
Linus Torvalds 已提交
569 570 571
	return err;
}

572 573 574 575 576 577
SYSCALL_DEFINE2(fchmod, unsigned int, fd, umode_t, mode)
{
	return ksys_fchmod(fd, mode);
}

int do_fchmodat(int dfd, const char __user *filename, umode_t mode)
L
Linus Torvalds 已提交
578
{
579
	struct path path;
L
Linus Torvalds 已提交
580
	int error;
581 582 583
	unsigned int lookup_flags = LOOKUP_FOLLOW;
retry:
	error = user_path_at(dfd, filename, lookup_flags, &path);
584 585 586
	if (!error) {
		error = chmod_common(&path, mode);
		path_put(&path);
587 588 589 590
		if (retry_estale(error, lookup_flags)) {
			lookup_flags |= LOOKUP_REVAL;
			goto retry;
		}
591
	}
L
Linus Torvalds 已提交
592 593 594
	return error;
}

595 596 597 598 599 600
SYSCALL_DEFINE3(fchmodat, int, dfd, const char __user *, filename,
		umode_t, mode)
{
	return do_fchmodat(dfd, filename, mode);
}

601
SYSCALL_DEFINE2(chmod, const char __user *, filename, umode_t, mode)
602
{
603
	return do_fchmodat(AT_FDCWD, filename, mode);
604 605
}

606
static int chown_common(const struct path *path, uid_t user, gid_t group)
L
Linus Torvalds 已提交
607
{
608
	struct inode *inode = path->dentry->d_inode;
609
	struct inode *delegated_inode = NULL;
L
Linus Torvalds 已提交
610 611
	int error;
	struct iattr newattrs;
612 613 614 615 616
	kuid_t uid;
	kgid_t gid;

	uid = make_kuid(current_user_ns(), user);
	gid = make_kgid(current_user_ns(), group);
L
Linus Torvalds 已提交
617

618
retry_deleg:
L
Linus Torvalds 已提交
619 620
	newattrs.ia_valid =  ATTR_CTIME;
	if (user != (uid_t) -1) {
621 622
		if (!uid_valid(uid))
			return -EINVAL;
L
Linus Torvalds 已提交
623
		newattrs.ia_valid |= ATTR_UID;
624
		newattrs.ia_uid = uid;
L
Linus Torvalds 已提交
625 626
	}
	if (group != (gid_t) -1) {
627 628
		if (!gid_valid(gid))
			return -EINVAL;
L
Linus Torvalds 已提交
629
		newattrs.ia_valid |= ATTR_GID;
630
		newattrs.ia_gid = gid;
L
Linus Torvalds 已提交
631 632
	}
	if (!S_ISDIR(inode->i_mode))
633 634
		newattrs.ia_valid |=
			ATTR_KILL_SUID | ATTR_KILL_SGID | ATTR_KILL_PRIV;
A
Al Viro 已提交
635
	inode_lock(inode);
636
	error = security_path_chown(path, uid, gid);
637
	if (!error)
638
		error = notify_change(path->dentry, &newattrs, &delegated_inode);
A
Al Viro 已提交
639
	inode_unlock(inode);
640 641 642 643 644
	if (delegated_inode) {
		error = break_deleg_wait(&delegated_inode);
		if (!error)
			goto retry_deleg;
	}
L
Linus Torvalds 已提交
645 646 647
	return error;
}

648 649
int do_fchownat(int dfd, const char __user *filename, uid_t user, gid_t group,
		int flag)
650
{
651
	struct path path;
652
	int error = -EINVAL;
653
	int lookup_flags;
654

655
	if ((flag & ~(AT_SYMLINK_NOFOLLOW | AT_EMPTY_PATH)) != 0)
656 657
		goto out;

658 659 660
	lookup_flags = (flag & AT_SYMLINK_NOFOLLOW) ? 0 : LOOKUP_FOLLOW;
	if (flag & AT_EMPTY_PATH)
		lookup_flags |= LOOKUP_EMPTY;
661
retry:
662
	error = user_path_at(dfd, filename, lookup_flags, &path);
663 664
	if (error)
		goto out;
665
	error = mnt_want_write(path.mnt);
666 667
	if (error)
		goto out_release;
668
	error = chown_common(&path, user, group);
669
	mnt_drop_write(path.mnt);
670
out_release:
671
	path_put(&path);
672 673 674 675
	if (retry_estale(error, lookup_flags)) {
		lookup_flags |= LOOKUP_REVAL;
		goto retry;
	}
676 677 678 679
out:
	return error;
}

680 681 682 683 684 685
SYSCALL_DEFINE5(fchownat, int, dfd, const char __user *, filename, uid_t, user,
		gid_t, group, int, flag)
{
	return do_fchownat(dfd, filename, user, group, flag);
}

686
SYSCALL_DEFINE3(chown, const char __user *, filename, uid_t, user, gid_t, group)
L
Linus Torvalds 已提交
687
{
688
	return do_fchownat(AT_FDCWD, filename, user, group, 0);
689
}
L
Linus Torvalds 已提交
690

691 692
SYSCALL_DEFINE3(lchown, const char __user *, filename, uid_t, user, gid_t, group)
{
693 694
	return do_fchownat(AT_FDCWD, filename, user, group,
			   AT_SYMLINK_NOFOLLOW);
L
Linus Torvalds 已提交
695 696
}

697
int ksys_fchown(unsigned int fd, uid_t user, gid_t group)
L
Linus Torvalds 已提交
698
{
699
	struct fd f = fdget(fd);
L
Linus Torvalds 已提交
700 701
	int error = -EBADF;

702
	if (!f.file)
703 704
		goto out;

705
	error = mnt_want_write_file_path(f.file);
706 707
	if (error)
		goto out_fput;
A
Al Viro 已提交
708
	audit_file(f.file);
709
	error = chown_common(&f.file->f_path, user, group);
710
	mnt_drop_write_file_path(f.file);
711
out_fput:
712
	fdput(f);
713
out:
L
Linus Torvalds 已提交
714 715 716
	return error;
}

717 718 719 720 721
SYSCALL_DEFINE3(fchown, unsigned int, fd, uid_t, user, gid_t, group)
{
	return ksys_fchown(fd, user, group);
}

M
Miklos Szeredi 已提交
722 723 724 725
int open_check_o_direct(struct file *f)
{
	/* NB: we're sure to have correct a_ops only after f_op->open */
	if (f->f_flags & O_DIRECT) {
M
Matthew Wilcox 已提交
726
		if (!f->f_mapping->a_ops || !f->f_mapping->a_ops->direct_IO)
M
Miklos Szeredi 已提交
727 728 729 730 731
			return -EINVAL;
	}
	return 0;
}

732
static int do_dentry_open(struct file *f,
733
			  struct inode *inode,
734 735
			  int (*open)(struct inode *, struct file *),
			  const struct cred *cred)
L
Linus Torvalds 已提交
736
{
737
	static const struct file_operations empty_fops = {};
L
Linus Torvalds 已提交
738 739
	int error;

A
Al Viro 已提交
740
	f->f_mode = OPEN_FMODE(f->f_flags) | FMODE_LSEEK |
741
				FMODE_PREAD | FMODE_PWRITE;
742

743
	path_get(&f->f_path);
744
	f->f_inode = inode;
L
Linus Torvalds 已提交
745 746
	f->f_mapping = inode->i_mapping;

747 748 749
	/* Ensure that we skip any errors that predate opening of the file */
	f->f_wb_err = filemap_sample_wb_err(f->f_mapping);

A
Al Viro 已提交
750 751
	if (unlikely(f->f_flags & O_PATH)) {
		f->f_mode = FMODE_PATH;
752
		f->f_op = &empty_fops;
753
		return 0;
754 755
	}

756
	if (f->f_mode & FMODE_WRITE && !special_file(inode->i_mode)) {
757
		error = get_write_access(inode);
A
Al Viro 已提交
758
		if (unlikely(error))
L
Linus Torvalds 已提交
759
			goto cleanup_file;
760
		error = __mnt_want_write(f->f_path.mnt);
A
Al Viro 已提交
761
		if (unlikely(error)) {
762 763 764
			put_write_access(inode);
			goto cleanup_file;
		}
765
		f->f_mode |= FMODE_WRITER;
L
Linus Torvalds 已提交
766 767
	}

768
	/* POSIX.1-2008/SUSv4 Section XSI 2.9.7 */
769
	if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode))
770 771
		f->f_mode |= FMODE_ATOMIC_POS;

772
	f->f_op = fops_get(inode->i_fop);
A
Al Viro 已提交
773 774 775 776
	if (unlikely(WARN_ON(!f->f_op))) {
		error = -ENODEV;
		goto cleanup_all;
	}
777

778
	error = security_file_open(f, cred);
779 780 781
	if (error)
		goto cleanup_all;

782
	error = break_lease(locks_inode(f), f->f_flags);
783 784 785
	if (error)
		goto cleanup_all;

A
Al Viro 已提交
786
	if (!open)
787 788 789
		open = f->f_op->open;
	if (open) {
		error = open(inode, f);
L
Linus Torvalds 已提交
790 791 792
		if (error)
			goto cleanup_all;
	}
793 794
	if ((f->f_mode & (FMODE_READ | FMODE_WRITE)) == FMODE_READ)
		i_readcount_inc(inode);
795
	if ((f->f_mode & FMODE_READ) &&
A
Al Viro 已提交
796
	     likely(f->f_op->read || f->f_op->read_iter))
797
		f->f_mode |= FMODE_CAN_READ;
798
	if ((f->f_mode & FMODE_WRITE) &&
A
Al Viro 已提交
799
	     likely(f->f_op->write || f->f_op->write_iter))
800
		f->f_mode |= FMODE_CAN_WRITE;
801

802
	f->f_write_hint = WRITE_LIFE_NOT_SET;
L
Linus Torvalds 已提交
803 804 805 806
	f->f_flags &= ~(O_CREAT | O_EXCL | O_NOCTTY | O_TRUNC);

	file_ra_state_init(&f->f_ra, f->f_mapping->host->i_mapping);

807
	return 0;
L
Linus Torvalds 已提交
808 809 810

cleanup_all:
	fops_put(f->f_op);
811
	if (f->f_mode & FMODE_WRITER) {
L
Linus Torvalds 已提交
812
		put_write_access(inode);
813
		__mnt_drop_write(f->f_path.mnt);
814
	}
L
Linus Torvalds 已提交
815
cleanup_file:
816 817 818
	path_put(&f->f_path);
	f->f_path.mnt = NULL;
	f->f_path.dentry = NULL;
819
	f->f_inode = NULL;
820
	return error;
L
Linus Torvalds 已提交
821 822
}

M
Miklos Szeredi 已提交
823 824
/**
 * finish_open - finish opening a file
825
 * @file: file pointer
M
Miklos Szeredi 已提交
826 827
 * @dentry: pointer to dentry
 * @open: open callback
828
 * @opened: state of open
M
Miklos Szeredi 已提交
829 830 831 832 833
 *
 * This can be used to finish opening a file passed to i_op->atomic_open().
 *
 * If the open callback is set to NULL, then the standard f_op->open()
 * filesystem callback is substituted.
834 835 836 837 838 839 840 841 842
 *
 * NB: the dentry reference is _not_ consumed.  If, for example, the dentry is
 * the return value of d_splice_alias(), then the caller needs to perform dput()
 * on it after finish_open().
 *
 * On successful return @file is a fully instantiated open file.  After this, if
 * an error occurs in ->atomic_open(), it needs to clean up with fput().
 *
 * Returns zero on success or -errno if the open failed.
M
Miklos Szeredi 已提交
843
 */
A
Al Viro 已提交
844 845 846
int finish_open(struct file *file, struct dentry *dentry,
		int (*open)(struct inode *, struct file *),
		int *opened)
M
Miklos Szeredi 已提交
847
{
848
	int error;
A
Al Viro 已提交
849
	BUG_ON(*opened & FILE_OPENED); /* once it's opened, it's opened */
M
Miklos Szeredi 已提交
850

851
	file->f_path.dentry = dentry;
852 853
	error = do_dentry_open(file, d_backing_inode(dentry), open,
			       current_cred());
854
	if (!error)
855
		*opened |= FILE_OPENED;
M
Miklos Szeredi 已提交
856

857
	return error;
M
Miklos Szeredi 已提交
858 859 860 861 862 863
}
EXPORT_SYMBOL(finish_open);

/**
 * finish_no_open - finish ->atomic_open() without opening the file
 *
864
 * @file: file pointer
M
Miklos Szeredi 已提交
865 866 867
 * @dentry: dentry or NULL (as returned from ->lookup())
 *
 * This can be used to set the result of a successful lookup in ->atomic_open().
868 869 870 871 872 873
 *
 * NB: unlike finish_open() this function does consume the dentry reference and
 * the caller need not dput() it.
 *
 * Returns "1" which must be the return value of ->atomic_open() after having
 * called this function.
M
Miklos Szeredi 已提交
874
 */
A
Al Viro 已提交
875
int finish_no_open(struct file *file, struct dentry *dentry)
M
Miklos Szeredi 已提交
876
{
A
Al Viro 已提交
877
	file->f_path.dentry = dentry;
A
Al Viro 已提交
878
	return 1;
M
Miklos Szeredi 已提交
879 880 881
}
EXPORT_SYMBOL(finish_no_open);

M
Miklos Szeredi 已提交
882 883 884 885 886 887
char *file_path(struct file *filp, char *buf, int buflen)
{
	return d_path(&filp->f_path, buf, buflen);
}
EXPORT_SYMBOL(file_path);

888 889 890 891 892 893 894 895 896
/**
 * vfs_open - open the file at the given path
 * @path: path to open
 * @file: newly allocated file with f_flag initialized
 * @cred: credentials to use
 */
int vfs_open(const struct path *path, struct file *file,
	     const struct cred *cred)
{
M
Miklos Szeredi 已提交
897
	struct dentry *dentry = d_real(path->dentry, NULL, file->f_flags, 0);
898

899 900
	if (IS_ERR(dentry))
		return PTR_ERR(dentry);
901

902
	file->f_path = *path;
903
	return do_dentry_open(file, d_backing_inode(dentry), NULL, cred);
904 905
}

906
struct file *dentry_open(const struct path *path, int flags,
907
			 const struct cred *cred)
908 909 910 911
{
	int error;
	struct file *f;

912 913
	validate_creds(cred);

914
	/* We must always pass in a valid mount pointer. */
915
	BUG_ON(!path->mnt);
916

917
	f = get_empty_filp();
918 919
	if (!IS_ERR(f)) {
		f->f_flags = flags;
M
Miklos Szeredi 已提交
920
		error = vfs_open(path, f, cred);
921 922 923 924 925 926 927 928 929
		if (!error) {
			/* from now on we need fput() to dispose of f */
			error = open_check_o_direct(f);
			if (error) {
				fput(f);
				f = ERR_PTR(error);
			}
		} else { 
			put_filp(f);
930 931 932 933
			f = ERR_PTR(error);
		}
	}
	return f;
934
}
L
Linus Torvalds 已提交
935 936
EXPORT_SYMBOL(dentry_open);

937
static inline int build_open_flags(int flags, umode_t mode, struct open_flags *op)
938 939
{
	int lookup_flags = 0;
A
Al Viro 已提交
940
	int acc_mode = ACC_MODE(flags);
941

942 943 944 945 946 947
	/*
	 * Clear out all open flags we don't know about so that we don't report
	 * them in fcntl(F_GETFD) or similar interfaces.
	 */
	flags &= VALID_OPEN_FLAGS;

948
	if (flags & (O_CREAT | __O_TMPFILE))
949 950 951
		op->mode = (mode & S_IALLUGO) | S_IFREG;
	else
		op->mode = 0;
952 953

	/* Must never be set by userspace */
A
Al Viro 已提交
954
	flags &= ~FMODE_NONOTIFY & ~O_CLOEXEC;
955 956 957 958 959 960 961 962 963 964

	/*
	 * O_SYNC is implemented as __O_SYNC|O_DSYNC.  As many places only
	 * check for O_DSYNC if the need any syncing at all we enforce it's
	 * always set instead of having to deal with possibly weird behaviour
	 * for malicious applications setting only __O_SYNC.
	 */
	if (flags & __O_SYNC)
		flags |= O_DSYNC;

A
Al Viro 已提交
965 966
	if (flags & __O_TMPFILE) {
		if ((flags & O_TMPFILE_MASK) != O_TMPFILE)
967
			return -EINVAL;
A
Al Viro 已提交
968 969
		if (!(acc_mode & MAY_WRITE))
			return -EINVAL;
970 971 972 973 974
	} else if (flags & O_PATH) {
		/*
		 * If we have O_PATH in the open flag. Then we
		 * cannot have anything other than the below set of flags
		 */
975 976 977
		flags &= O_DIRECTORY | O_NOFOLLOW | O_PATH;
		acc_mode = 0;
	}
978

979
	op->open_flag = flags;
980 981 982 983 984 985 986 987 988 989 990 991

	/* O_TRUNC implies we need access checks for write permissions */
	if (flags & O_TRUNC)
		acc_mode |= MAY_WRITE;

	/* Allow the LSM permission hook to distinguish append
	   access from general write access. */
	if (flags & O_APPEND)
		acc_mode |= MAY_APPEND;

	op->acc_mode = acc_mode;

992 993
	op->intent = flags & O_PATH ? 0 : LOOKUP_OPEN;

994 995 996 997 998 999 1000 1001 1002 1003
	if (flags & O_CREAT) {
		op->intent |= LOOKUP_CREATE;
		if (flags & O_EXCL)
			op->intent |= LOOKUP_EXCL;
	}

	if (flags & O_DIRECTORY)
		lookup_flags |= LOOKUP_DIRECTORY;
	if (!(flags & O_NOFOLLOW))
		lookup_flags |= LOOKUP_FOLLOW;
1004 1005
	op->lookup_flags = lookup_flags;
	return 0;
1006 1007
}

1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021
/**
 * file_open_name - open file and return file pointer
 *
 * @name:	struct filename containing path to open
 * @flags:	open flags as per the open(2) second argument
 * @mode:	mode for the new file if O_CREAT is set, else ignored
 *
 * This is the helper to open a file from kernelspace if you really
 * have to.  But in generally you should not do this, so please move
 * along, nothing to see here..
 */
struct file *file_open_name(struct filename *name, int flags, umode_t mode)
{
	struct open_flags op;
1022 1023
	int err = build_open_flags(flags, mode, &op);
	return err ? ERR_PTR(err) : do_filp_open(AT_FDCWD, name, &op);
1024 1025
}

1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036
/**
 * filp_open - open file and return file pointer
 *
 * @filename:	path to open
 * @flags:	open flags as per the open(2) second argument
 * @mode:	mode for the new file if O_CREAT is set, else ignored
 *
 * This is the helper to open a file from kernelspace if you really
 * have to.  But in generally you should not do this, so please move
 * along, nothing to see here..
 */
1037
struct file *filp_open(const char *filename, int flags, umode_t mode)
1038
{
1039 1040 1041 1042 1043 1044 1045 1046
	struct filename *name = getname_kernel(filename);
	struct file *file = ERR_CAST(name);
	
	if (!IS_ERR(name)) {
		file = file_open_name(name, flags, mode);
		putname(name);
	}
	return file;
1047 1048 1049
}
EXPORT_SYMBOL(filp_open);

A
Al Viro 已提交
1050
struct file *file_open_root(struct dentry *dentry, struct vfsmount *mnt,
1051
			    const char *filename, int flags, umode_t mode)
A
Al Viro 已提交
1052 1053
{
	struct open_flags op;
1054
	int err = build_open_flags(flags, mode, &op);
1055 1056 1057
	if (err)
		return ERR_PTR(err);
	return do_file_open_root(dentry, mnt, filename, &op);
A
Al Viro 已提交
1058 1059 1060
}
EXPORT_SYMBOL(file_open_root);

J
James Bottomley 已提交
1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080
struct file *filp_clone_open(struct file *oldfile)
{
	struct file *file;
	int retval;

	file = get_empty_filp();
	if (IS_ERR(file))
		return file;

	file->f_flags = oldfile->f_flags;
	retval = vfs_open(&oldfile->f_path, file, oldfile->f_cred);
	if (retval) {
		put_filp(file);
		return ERR_PTR(retval);
	}

	return file;
}
EXPORT_SYMBOL(filp_clone_open);

1081
long do_sys_open(int dfd, const char __user *filename, int flags, umode_t mode)
L
Linus Torvalds 已提交
1082
{
1083
	struct open_flags op;
1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102
	int fd = build_open_flags(flags, mode, &op);
	struct filename *tmp;

	if (fd)
		return fd;

	tmp = getname(filename);
	if (IS_ERR(tmp))
		return PTR_ERR(tmp);

	fd = get_unused_fd_flags(flags);
	if (fd >= 0) {
		struct file *f = do_filp_open(dfd, tmp, &op);
		if (IS_ERR(f)) {
			put_unused_fd(fd);
			fd = PTR_ERR(f);
		} else {
			fsnotify_open(f);
			fd_install(fd, f);
L
Linus Torvalds 已提交
1103 1104
		}
	}
1105
	putname(tmp);
L
Linus Torvalds 已提交
1106 1107
	return fd;
}
1108

1109
SYSCALL_DEFINE3(open, const char __user *, filename, int, flags, umode_t, mode)
1110 1111 1112 1113
{
	if (force_o_largefile())
		flags |= O_LARGEFILE;

1114
	return do_sys_open(AT_FDCWD, filename, flags, mode);
1115
}
L
Linus Torvalds 已提交
1116

1117
SYSCALL_DEFINE4(openat, int, dfd, const char __user *, filename, int, flags,
1118
		umode_t, mode)
1119 1120 1121 1122
{
	if (force_o_largefile())
		flags |= O_LARGEFILE;

1123
	return do_sys_open(dfd, filename, flags, mode);
1124 1125
}

1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145
#ifdef CONFIG_COMPAT
/*
 * Exactly like sys_open(), except that it doesn't set the
 * O_LARGEFILE flag.
 */
COMPAT_SYSCALL_DEFINE3(open, const char __user *, filename, int, flags, umode_t, mode)
{
	return do_sys_open(AT_FDCWD, filename, flags, mode);
}

/*
 * Exactly like sys_openat(), except that it doesn't set the
 * O_LARGEFILE flag.
 */
COMPAT_SYSCALL_DEFINE4(openat, int, dfd, const char __user *, filename, int, flags, umode_t, mode)
{
	return do_sys_open(dfd, filename, flags, mode);
}
#endif

L
Linus Torvalds 已提交
1146 1147 1148 1149 1150 1151
#ifndef __alpha__

/*
 * For backward compatibility?  Maybe this should be moved
 * into arch/i386 instead?
 */
1152
SYSCALL_DEFINE2(creat, const char __user *, pathname, umode_t, mode)
L
Linus Torvalds 已提交
1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164
{
	return sys_open(pathname, O_CREAT | O_WRONLY | O_TRUNC, mode);
}

#endif

/*
 * "id" is the POSIX thread ID. We use the
 * files pointer for this..
 */
int filp_close(struct file *filp, fl_owner_t id)
{
1165
	int retval = 0;
L
Linus Torvalds 已提交
1166 1167 1168

	if (!file_count(filp)) {
		printk(KERN_ERR "VFS: Close: file count is 0\n");
1169
		return 0;
L
Linus Torvalds 已提交
1170 1171
	}

A
Al Viro 已提交
1172
	if (filp->f_op->flush)
1173
		retval = filp->f_op->flush(filp, id);
L
Linus Torvalds 已提交
1174

1175 1176 1177 1178
	if (likely(!(filp->f_mode & FMODE_PATH))) {
		dnotify_flush(filp, id);
		locks_remove_posix(filp, id);
	}
L
Linus Torvalds 已提交
1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189
	fput(filp);
	return retval;
}

EXPORT_SYMBOL(filp_close);

/*
 * Careful here! We test whether the file pointer is NULL before
 * releasing the fd. This ensures that one clone task can't release
 * an fd while another clone is opening it.
 */
1190
SYSCALL_DEFINE1(close, unsigned int, fd)
L
Linus Torvalds 已提交
1191
{
1192
	int retval = __close_fd(current->files, fd);
1193 1194 1195 1196 1197 1198 1199 1200 1201

	/* can't restart close syscall because file table entry was cleared */
	if (unlikely(retval == -ERESTARTSYS ||
		     retval == -ERESTARTNOINTR ||
		     retval == -ERESTARTNOHAND ||
		     retval == -ERESTART_RESTARTBLOCK))
		retval = -EINTR;

	return retval;
L
Linus Torvalds 已提交
1202 1203 1204 1205 1206 1207 1208
}
EXPORT_SYMBOL(sys_close);

/*
 * This routine simulates a hangup on the tty, to arrange that users
 * are given clean terminals at login time.
 */
1209
SYSCALL_DEFINE0(vhangup)
L
Linus Torvalds 已提交
1210 1211
{
	if (capable(CAP_SYS_TTY_CONFIG)) {
A
Alan Cox 已提交
1212
		tty_vhangup_self();
L
Linus Torvalds 已提交
1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226
		return 0;
	}
	return -EPERM;
}

/*
 * Called when an inode is about to be open.
 * We use this to disallow opening large files on 32bit systems if
 * the caller didn't specify O_LARGEFILE.  On 64bit systems we force
 * on this flag in sys_open.
 */
int generic_file_open(struct inode * inode, struct file * filp)
{
	if (!(filp->f_flags & O_LARGEFILE) && i_size_read(inode) > MAX_NON_LFS)
1227
		return -EOVERFLOW;
L
Linus Torvalds 已提交
1228 1229 1230 1231 1232 1233 1234
	return 0;
}

EXPORT_SYMBOL(generic_file_open);

/*
 * This is used by subsystems that don't want seekable
1235 1236 1237
 * file descriptors. The function is not supposed to ever fail, the only
 * reason it returns an 'int' and not 'void' is so that it can be plugged
 * directly into file_operations structure.
L
Linus Torvalds 已提交
1238 1239 1240 1241 1242 1243 1244 1245
 */
int nonseekable_open(struct inode *inode, struct file *filp)
{
	filp->f_mode &= ~(FMODE_LSEEK | FMODE_PREAD | FMODE_PWRITE);
	return 0;
}

EXPORT_SYMBOL(nonseekable_open);