open.c 27 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
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>
Al Viro's avatar
Al Viro committed
10
#include <linux/fdtable.h>
Robert Love's avatar
Robert Love committed
11
#include <linux/fsnotify.h>
Linus Torvalds's avatar
Linus Torvalds committed
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>
Linus Torvalds's avatar
Linus Torvalds committed
18 19
#include <linux/security.h>
#include <linux/mount.h>
20
#include <linux/fcntl.h>
21
#include <linux/slab.h>
Linus Torvalds's avatar
Linus Torvalds committed
22 23
#include <asm/uaccess.h>
#include <linux/fs.h>
24
#include <linux/personality.h>
Linus Torvalds's avatar
Linus Torvalds committed
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>
Linus Torvalds's avatar
Linus Torvalds committed
33

34 35
#include "internal.h"

36 37
int do_truncate(struct dentry *dentry, loff_t length, unsigned int time_attrs,
	struct file *filp)
Linus Torvalds's avatar
Linus Torvalds committed
38
{
39
	int ret;
Linus Torvalds's avatar
Linus Torvalds committed
40 41 42 43 44 45 46
	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;
47
	newattrs.ia_valid = ATTR_SIZE | time_attrs;
48 49 50 51
	if (filp) {
		newattrs.ia_file = filp;
		newattrs.ia_valid |= ATTR_FILE;
	}
Linus Torvalds's avatar
Linus Torvalds committed
52

53
	/* Remove suid/sgid on truncate too */
54 55 56
	ret = should_remove_suid(dentry);
	if (ret)
		newattrs.ia_valid |= ret | ATTR_FORCE;
57

58
	mutex_lock(&dentry->d_inode->i_mutex);
59
	ret = notify_change(dentry, &newattrs);
60
	mutex_unlock(&dentry->d_inode->i_mutex);
61
	return ret;
Linus Torvalds's avatar
Linus Torvalds committed
62 63
}

64
static long do_sys_truncate(const char __user *pathname, loff_t length)
Linus Torvalds's avatar
Linus Torvalds committed
65
{
66 67
	struct path path;
	struct inode *inode;
Linus Torvalds's avatar
Linus Torvalds committed
68 69 70 71 72 73
	int error;

	error = -EINVAL;
	if (length < 0)	/* sorry, but loff_t says... */
		goto out;

74
	error = user_path(pathname, &path);
Linus Torvalds's avatar
Linus Torvalds committed
75 76
	if (error)
		goto out;
77
	inode = path.dentry->d_inode;
Linus Torvalds's avatar
Linus Torvalds committed
78 79 80 81 82 83 84 85 86 87

	/* For directories it's -EISDIR, for other non-regulars - -EINVAL */
	error = -EISDIR;
	if (S_ISDIR(inode->i_mode))
		goto dput_and_out;

	error = -EINVAL;
	if (!S_ISREG(inode->i_mode))
		goto dput_and_out;

88
	error = mnt_want_write(path.mnt);
Linus Torvalds's avatar
Linus Torvalds committed
89 90 91
	if (error)
		goto dput_and_out;

92
	error = inode_permission(inode, MAY_WRITE);
93 94
	if (error)
		goto mnt_drop_write_and_out;
Linus Torvalds's avatar
Linus Torvalds committed
95 96

	error = -EPERM;
97
	if (IS_APPEND(inode))
98
		goto mnt_drop_write_and_out;
Linus Torvalds's avatar
Linus Torvalds committed
99

100
	error = get_write_access(inode);
Linus Torvalds's avatar
Linus Torvalds committed
101
	if (error)
102
		goto mnt_drop_write_and_out;
Linus Torvalds's avatar
Linus Torvalds committed
103

104 105 106 107
	/*
	 * Make sure that there are no leases.  get_write_access() protects
	 * against the truncate racing with a lease-granting setlease().
	 */
108
	error = break_lease(inode, O_WRONLY);
Linus Torvalds's avatar
Linus Torvalds committed
109
	if (error)
110
		goto put_write_and_out;
Linus Torvalds's avatar
Linus Torvalds committed
111 112

	error = locks_verify_truncate(inode, NULL, length);
113
	if (!error)
114
		error = security_path_truncate(&path);
115
	if (!error)
116
		error = do_truncate(path.dentry, length, 0, NULL);
Linus Torvalds's avatar
Linus Torvalds committed
117

118 119
put_write_and_out:
	put_write_access(inode);
120
mnt_drop_write_and_out:
121
	mnt_drop_write(path.mnt);
Linus Torvalds's avatar
Linus Torvalds committed
122
dput_and_out:
123
	path_put(&path);
Linus Torvalds's avatar
Linus Torvalds committed
124 125 126 127
out:
	return error;
}

128
SYSCALL_DEFINE2(truncate, const char __user *, path, long, length)
Linus Torvalds's avatar
Linus Torvalds committed
129
{
130
	return do_sys_truncate(path, length);
Linus Torvalds's avatar
Linus Torvalds committed
131 132
}

133
static long do_sys_ftruncate(unsigned int fd, loff_t length, int small)
Linus Torvalds's avatar
Linus Torvalds committed
134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151
{
	struct inode * inode;
	struct dentry *dentry;
	struct file * file;
	int error;

	error = -EINVAL;
	if (length < 0)
		goto out;
	error = -EBADF;
	file = fget(fd);
	if (!file)
		goto out;

	/* explicitly opened as large or we are on 64-bit box */
	if (file->f_flags & O_LARGEFILE)
		small = 0;

152
	dentry = file->f_path.dentry;
Linus Torvalds's avatar
Linus Torvalds committed
153 154 155 156 157 158 159 160 161 162 163 164 165 166 167
	inode = dentry->d_inode;
	error = -EINVAL;
	if (!S_ISREG(inode->i_mode) || !(file->f_mode & FMODE_WRITE))
		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;
	if (IS_APPEND(inode))
		goto out_putf;

	error = locks_verify_truncate(inode, file, length);
168
	if (!error)
169
		error = security_path_truncate(&file->f_path);
Linus Torvalds's avatar
Linus Torvalds committed
170
	if (!error)
171
		error = do_truncate(dentry, length, ATTR_MTIME|ATTR_CTIME, file);
Linus Torvalds's avatar
Linus Torvalds committed
172 173 174 175 176 177
out_putf:
	fput(file);
out:
	return error;
}

178
SYSCALL_DEFINE2(ftruncate, unsigned int, fd, unsigned long, length)
Linus Torvalds's avatar
Linus Torvalds committed
179
{
180
	long ret = do_sys_ftruncate(fd, length, 1);
181
	/* avoid REGPARM breakage on x86: */
182
	asmlinkage_protect(2, ret, fd, length);
183
	return ret;
Linus Torvalds's avatar
Linus Torvalds committed
184 185 186 187
}

/* LFS versions of truncate are only needed on 32 bit machines */
#if BITS_PER_LONG == 32
188
SYSCALL_DEFINE(truncate64)(const char __user * path, loff_t length)
Linus Torvalds's avatar
Linus Torvalds committed
189 190 191
{
	return do_sys_truncate(path, length);
}
192 193 194 195 196 197 198
#ifdef CONFIG_HAVE_SYSCALL_WRAPPERS
asmlinkage long SyS_truncate64(long path, loff_t length)
{
	return SYSC_truncate64((const char __user *) path, length);
}
SYSCALL_ALIAS(sys_truncate64, SyS_truncate64);
#endif
Linus Torvalds's avatar
Linus Torvalds committed
199

200
SYSCALL_DEFINE(ftruncate64)(unsigned int fd, loff_t length)
Linus Torvalds's avatar
Linus Torvalds committed
201
{
202
	long ret = do_sys_ftruncate(fd, length, 0);
203
	/* avoid REGPARM breakage on x86: */
204
	asmlinkage_protect(2, ret, fd, length);
205
	return ret;
Linus Torvalds's avatar
Linus Torvalds committed
206
}
207 208 209 210 211 212
#ifdef CONFIG_HAVE_SYSCALL_WRAPPERS
asmlinkage long SyS_ftruncate64(long fd, loff_t length)
{
	return SYSC_ftruncate64((unsigned int) fd, length);
}
SYSCALL_ALIAS(sys_ftruncate64, SyS_ftruncate64);
Linus Torvalds's avatar
Linus Torvalds committed
213
#endif
214
#endif /* BITS_PER_LONG == 32 */
Linus Torvalds's avatar
Linus Torvalds committed
215

216 217

int do_fallocate(struct file *file, int mode, loff_t offset, loff_t len)
218
{
219 220
	struct inode *inode = file->f_path.dentry->d_inode;
	long ret;
221 222

	if (offset < 0 || len <= 0)
223
		return -EINVAL;
224 225

	/* Return error if mode is not supported */
226 227 228 229 230 231
	if (mode & ~(FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE))
		return -EOPNOTSUPP;

	/* Punch hole must have keep size set */
	if ((mode & FALLOC_FL_PUNCH_HOLE) &&
	    !(mode & FALLOC_FL_KEEP_SIZE))
232
		return -EOPNOTSUPP;
233 234

	if (!(file->f_mode & FMODE_WRITE))
235
		return -EBADF;
236 237 238 239 240 241 242 243

	/* It's not possible punch hole on append only file */
	if (mode & FALLOC_FL_PUNCH_HOLE && IS_APPEND(inode))
		return -EPERM;

	if (IS_IMMUTABLE(inode))
		return -EPERM;

244 245 246 247 248 249
	/*
	 * Revalidate the write permissions, in case security policy has
	 * changed since the files were opened.
	 */
	ret = security_file_permission(file, MAY_WRITE);
	if (ret)
250
		return ret;
251 252

	if (S_ISFIFO(inode->i_mode))
253
		return -ESPIPE;
254 255 256 257 258 259

	/*
	 * Let individual file system decide if it supports preallocation
	 * for directories or not.
	 */
	if (!S_ISREG(inode->i_mode) && !S_ISDIR(inode->i_mode))
260
		return -ENODEV;
261 262 263

	/* Check for wrap through zero too */
	if (((offset + len) > inode->i_sb->s_maxbytes) || ((offset + len) < 0))
264
		return -EFBIG;
265

266
	if (!file->f_op->fallocate)
267
		return -EOPNOTSUPP;
268

269
	return file->f_op->fallocate(file, mode, offset, len);
270 271 272 273 274 275 276 277 278 279 280 281 282 283
}

SYSCALL_DEFINE(fallocate)(int fd, int mode, loff_t offset, loff_t len)
{
	struct file *file;
	int error = -EBADF;

	file = fget(fd);
	if (file) {
		error = do_fallocate(file, mode, offset, len);
		fput(file);
	}

	return error;
284
}
285

286 287 288 289 290 291 292
#ifdef CONFIG_HAVE_SYSCALL_WRAPPERS
asmlinkage long SyS_fallocate(long fd, long mode, loff_t offset, loff_t len)
{
	return SYSC_fallocate((int)fd, (int)mode, offset, len);
}
SYSCALL_ALIAS(sys_fallocate, SyS_fallocate);
#endif
293

Linus Torvalds's avatar
Linus Torvalds committed
294 295 296 297 298
/*
 * 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.
 */
299
SYSCALL_DEFINE3(faccessat, int, dfd, const char __user *, filename, int, mode)
Linus Torvalds's avatar
Linus Torvalds committed
300
{
301 302
	const struct cred *old_cred;
	struct cred *override_cred;
303
	struct path path;
304
	struct inode *inode;
Linus Torvalds's avatar
Linus Torvalds committed
305 306 307 308 309
	int res;

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

310 311 312
	override_cred = prepare_creds();
	if (!override_cred)
		return -ENOMEM;
Linus Torvalds's avatar
Linus Torvalds committed
313

314 315
	override_cred->fsuid = override_cred->uid;
	override_cred->fsgid = override_cred->gid;
Linus Torvalds's avatar
Linus Torvalds committed
316

317
	if (!issecure(SECURE_NO_SETUID_FIXUP)) {
David Howells's avatar
David Howells committed
318
		/* Clear the capabilities if we switch to a non-root user */
319 320
		if (override_cred->uid)
			cap_clear(override_cred->cap_effective);
321
		else
322 323
			override_cred->cap_effective =
				override_cred->cap_permitted;
324
	}
Linus Torvalds's avatar
Linus Torvalds committed
325

326 327
	old_cred = override_creds(override_cred);

328
	res = user_path_at(dfd, filename, LOOKUP_FOLLOW, &path);
329 330 331
	if (res)
		goto out;

332
	inode = path.dentry->d_inode;
333 334

	if ((mode & MAY_EXEC) && S_ISREG(inode->i_mode)) {
335 336 337 338 339
		/*
		 * MAY_EXEC on regular files is denied if the fs is mounted
		 * with the "noexec" flag.
		 */
		res = -EACCES;
340
		if (path.mnt->mnt_flags & MNT_NOEXEC)
341 342 343
			goto out_path_release;
	}

344
	res = inode_permission(inode, mode | MAY_ACCESS);
345
	/* SuS v2 requires we report a read only fs too */
346
	if (res || !(mode & S_IWOTH) || special_file(inode->i_mode))
347
		goto out_path_release;
348 349 350 351 352 353 354 355 356 357
	/*
	 * 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.
	 */
358
	if (__mnt_is_readonly(path.mnt))
359
		res = -EROFS;
Linus Torvalds's avatar
Linus Torvalds committed
360

361
out_path_release:
362
	path_put(&path);
363
out:
364 365
	revert_creds(old_cred);
	put_cred(override_cred);
Linus Torvalds's avatar
Linus Torvalds committed
366 367 368
	return res;
}

369
SYSCALL_DEFINE2(access, const char __user *, filename, int, mode)
370 371 372 373
{
	return sys_faccessat(AT_FDCWD, filename, mode);
}

374
SYSCALL_DEFINE1(chdir, const char __user *, filename)
Linus Torvalds's avatar
Linus Torvalds committed
375
{
376
	struct path path;
Linus Torvalds's avatar
Linus Torvalds committed
377 378
	int error;

379
	error = user_path_dir(filename, &path);
Linus Torvalds's avatar
Linus Torvalds committed
380 381 382
	if (error)
		goto out;

Eric Paris's avatar
Eric Paris committed
383
	error = inode_permission(path.dentry->d_inode, MAY_EXEC | MAY_CHDIR);
Linus Torvalds's avatar
Linus Torvalds committed
384 385 386
	if (error)
		goto dput_and_out;

387
	set_fs_pwd(current->fs, &path);
Linus Torvalds's avatar
Linus Torvalds committed
388 389

dput_and_out:
390
	path_put(&path);
Linus Torvalds's avatar
Linus Torvalds committed
391 392 393 394
out:
	return error;
}

395
SYSCALL_DEFINE1(fchdir, unsigned int, fd)
Linus Torvalds's avatar
Linus Torvalds committed
396 397 398 399 400 401 402 403 404 405
{
	struct file *file;
	struct inode *inode;
	int error;

	error = -EBADF;
	file = fget(fd);
	if (!file)
		goto out;

406
	inode = file->f_path.dentry->d_inode;
Linus Torvalds's avatar
Linus Torvalds committed
407 408 409 410 411

	error = -ENOTDIR;
	if (!S_ISDIR(inode->i_mode))
		goto out_putf;

Eric Paris's avatar
Eric Paris committed
412
	error = inode_permission(inode, MAY_EXEC | MAY_CHDIR);
Linus Torvalds's avatar
Linus Torvalds committed
413
	if (!error)
414
		set_fs_pwd(current->fs, &file->f_path);
Linus Torvalds's avatar
Linus Torvalds committed
415 416 417 418 419 420
out_putf:
	fput(file);
out:
	return error;
}

421
SYSCALL_DEFINE1(chroot, const char __user *, filename)
Linus Torvalds's avatar
Linus Torvalds committed
422
{
423
	struct path path;
Linus Torvalds's avatar
Linus Torvalds committed
424 425
	int error;

426
	error = user_path_dir(filename, &path);
Linus Torvalds's avatar
Linus Torvalds committed
427 428 429
	if (error)
		goto out;

Eric Paris's avatar
Eric Paris committed
430
	error = inode_permission(path.dentry->d_inode, MAY_EXEC | MAY_CHDIR);
Linus Torvalds's avatar
Linus Torvalds committed
431 432 433 434 435 436
	if (error)
		goto dput_and_out;

	error = -EPERM;
	if (!capable(CAP_SYS_CHROOT))
		goto dput_and_out;
437 438 439
	error = security_path_chroot(&path);
	if (error)
		goto dput_and_out;
Linus Torvalds's avatar
Linus Torvalds committed
440

441
	set_fs_root(current->fs, &path);
Linus Torvalds's avatar
Linus Torvalds committed
442 443
	error = 0;
dput_and_out:
444
	path_put(&path);
Linus Torvalds's avatar
Linus Torvalds committed
445 446 447 448
out:
	return error;
}

449
SYSCALL_DEFINE2(fchmod, unsigned int, fd, mode_t, mode)
Linus Torvalds's avatar
Linus Torvalds committed
450 451 452 453 454 455 456 457 458 459 460
{
	struct inode * inode;
	struct dentry * dentry;
	struct file * file;
	int err = -EBADF;
	struct iattr newattrs;

	file = fget(fd);
	if (!file)
		goto out;

461
	dentry = file->f_path.dentry;
Linus Torvalds's avatar
Linus Torvalds committed
462 463
	inode = dentry->d_inode;

464
	audit_inode(NULL, dentry);
465

466
	err = mnt_want_write_file(file);
467
	if (err)
Linus Torvalds's avatar
Linus Torvalds committed
468
		goto out_putf;
469
	mutex_lock(&inode->i_mutex);
470 471
	err = security_path_chmod(dentry, file->f_vfsmnt, mode);
	if (err)
472
		goto out_unlock;
Linus Torvalds's avatar
Linus Torvalds committed
473 474 475 476 477
	if (mode == (mode_t) -1)
		mode = inode->i_mode;
	newattrs.ia_mode = (mode & S_IALLUGO) | (inode->i_mode & ~S_IALLUGO);
	newattrs.ia_valid = ATTR_MODE | ATTR_CTIME;
	err = notify_change(dentry, &newattrs);
478
out_unlock:
479
	mutex_unlock(&inode->i_mutex);
480
	mnt_drop_write(file->f_path.mnt);
Linus Torvalds's avatar
Linus Torvalds committed
481 482 483 484 485 486
out_putf:
	fput(file);
out:
	return err;
}

487
SYSCALL_DEFINE3(fchmodat, int, dfd, const char __user *, filename, mode_t, mode)
Linus Torvalds's avatar
Linus Torvalds committed
488
{
489 490
	struct path path;
	struct inode *inode;
Linus Torvalds's avatar
Linus Torvalds committed
491 492 493
	int error;
	struct iattr newattrs;

494
	error = user_path_at(dfd, filename, LOOKUP_FOLLOW, &path);
Linus Torvalds's avatar
Linus Torvalds committed
495 496
	if (error)
		goto out;
497
	inode = path.dentry->d_inode;
Linus Torvalds's avatar
Linus Torvalds committed
498

499
	error = mnt_want_write(path.mnt);
500
	if (error)
Linus Torvalds's avatar
Linus Torvalds committed
501
		goto dput_and_out;
502
	mutex_lock(&inode->i_mutex);
503 504
	error = security_path_chmod(path.dentry, path.mnt, mode);
	if (error)
505
		goto out_unlock;
Linus Torvalds's avatar
Linus Torvalds committed
506 507 508 509
	if (mode == (mode_t) -1)
		mode = inode->i_mode;
	newattrs.ia_mode = (mode & S_IALLUGO) | (inode->i_mode & ~S_IALLUGO);
	newattrs.ia_valid = ATTR_MODE | ATTR_CTIME;
510
	error = notify_change(path.dentry, &newattrs);
511
out_unlock:
512
	mutex_unlock(&inode->i_mutex);
513
	mnt_drop_write(path.mnt);
Linus Torvalds's avatar
Linus Torvalds committed
514
dput_and_out:
515
	path_put(&path);
Linus Torvalds's avatar
Linus Torvalds committed
516 517 518 519
out:
	return error;
}

520
SYSCALL_DEFINE2(chmod, const char __user *, filename, mode_t, mode)
521 522 523 524
{
	return sys_fchmodat(AT_FDCWD, filename, mode);
}

525
static int chown_common(struct path *path, uid_t user, gid_t group)
Linus Torvalds's avatar
Linus Torvalds committed
526
{
527
	struct inode *inode = path->dentry->d_inode;
Linus Torvalds's avatar
Linus Torvalds committed
528 529 530 531 532 533 534 535 536 537 538 539 540
	int error;
	struct iattr newattrs;

	newattrs.ia_valid =  ATTR_CTIME;
	if (user != (uid_t) -1) {
		newattrs.ia_valid |= ATTR_UID;
		newattrs.ia_uid = user;
	}
	if (group != (gid_t) -1) {
		newattrs.ia_valid |= ATTR_GID;
		newattrs.ia_gid = group;
	}
	if (!S_ISDIR(inode->i_mode))
541 542
		newattrs.ia_valid |=
			ATTR_KILL_SUID | ATTR_KILL_SGID | ATTR_KILL_PRIV;
543
	mutex_lock(&inode->i_mutex);
544 545 546
	error = security_path_chown(path, user, group);
	if (!error)
		error = notify_change(path->dentry, &newattrs);
547
	mutex_unlock(&inode->i_mutex);
548

Linus Torvalds's avatar
Linus Torvalds committed
549 550 551
	return error;
}

552
SYSCALL_DEFINE3(chown, const char __user *, filename, uid_t, user, gid_t, group)
Linus Torvalds's avatar
Linus Torvalds committed
553
{
554
	struct path path;
Linus Torvalds's avatar
Linus Torvalds committed
555 556
	int error;

557
	error = user_path(filename, &path);
558 559
	if (error)
		goto out;
560
	error = mnt_want_write(path.mnt);
561 562
	if (error)
		goto out_release;
563
	error = chown_common(&path, user, group);
564
	mnt_drop_write(path.mnt);
565
out_release:
566
	path_put(&path);
567
out:
Linus Torvalds's avatar
Linus Torvalds committed
568 569 570
	return error;
}

571 572
SYSCALL_DEFINE5(fchownat, int, dfd, const char __user *, filename, uid_t, user,
		gid_t, group, int, flag)
573
{
574
	struct path path;
575
	int error = -EINVAL;
576
	int lookup_flags;
577

578
	if ((flag & ~(AT_SYMLINK_NOFOLLOW | AT_EMPTY_PATH)) != 0)
579 580
		goto out;

581 582 583 584
	lookup_flags = (flag & AT_SYMLINK_NOFOLLOW) ? 0 : LOOKUP_FOLLOW;
	if (flag & AT_EMPTY_PATH)
		lookup_flags |= LOOKUP_EMPTY;
	error = user_path_at(dfd, filename, lookup_flags, &path);
585 586
	if (error)
		goto out;
587
	error = mnt_want_write(path.mnt);
588 589
	if (error)
		goto out_release;
590
	error = chown_common(&path, user, group);
591
	mnt_drop_write(path.mnt);
592
out_release:
593
	path_put(&path);
594 595 596 597
out:
	return error;
}

598
SYSCALL_DEFINE3(lchown, const char __user *, filename, uid_t, user, gid_t, group)
Linus Torvalds's avatar
Linus Torvalds committed
599
{
600
	struct path path;
Linus Torvalds's avatar
Linus Torvalds committed
601 602
	int error;

603
	error = user_lpath(filename, &path);
604 605
	if (error)
		goto out;
606
	error = mnt_want_write(path.mnt);
607 608
	if (error)
		goto out_release;
609
	error = chown_common(&path, user, group);
610
	mnt_drop_write(path.mnt);
611
out_release:
612
	path_put(&path);
613
out:
Linus Torvalds's avatar
Linus Torvalds committed
614 615 616
	return error;
}

617
SYSCALL_DEFINE3(fchown, unsigned int, fd, uid_t, user, gid_t, group)
Linus Torvalds's avatar
Linus Torvalds committed
618 619 620
{
	struct file * file;
	int error = -EBADF;
621
	struct dentry * dentry;
Linus Torvalds's avatar
Linus Torvalds committed
622 623

	file = fget(fd);
624 625 626
	if (!file)
		goto out;

627
	error = mnt_want_write_file(file);
628 629
	if (error)
		goto out_fput;
630
	dentry = file->f_path.dentry;
631
	audit_inode(NULL, dentry);
632
	error = chown_common(&file->f_path, user, group);
633 634
	mnt_drop_write(file->f_path.mnt);
out_fput:
635 636
	fput(file);
out:
Linus Torvalds's avatar
Linus Torvalds committed
637 638 639
	return error;
}

640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668
/*
 * You have to be very careful that these write
 * counts get cleaned up in error cases and
 * upon __fput().  This should probably never
 * be called outside of __dentry_open().
 */
static inline int __get_file_write_access(struct inode *inode,
					  struct vfsmount *mnt)
{
	int error;
	error = get_write_access(inode);
	if (error)
		return error;
	/*
	 * Do not take mount writer counts on
	 * special files since no writes to
	 * the mount itself will occur.
	 */
	if (!special_file(inode->i_mode)) {
		/*
		 * Balanced in __fput()
		 */
		error = mnt_want_write(mnt);
		if (error)
			put_write_access(inode);
	}
	return error;
}

669
static struct file *__dentry_open(struct dentry *dentry, struct vfsmount *mnt,
670
					struct file *f,
671 672
					int (*open)(struct inode *, struct file *),
					const struct cred *cred)
Linus Torvalds's avatar
Linus Torvalds committed
673
{
674
	static const struct file_operations empty_fops = {};
Linus Torvalds's avatar
Linus Torvalds committed
675 676 677
	struct inode *inode;
	int error;

Al Viro's avatar
Al Viro committed
678
	f->f_mode = OPEN_FMODE(f->f_flags) | FMODE_LSEEK |
679
				FMODE_PREAD | FMODE_PWRITE;
680 681 682 683

	if (unlikely(f->f_flags & O_PATH))
		f->f_mode = FMODE_PATH;

Linus Torvalds's avatar
Linus Torvalds committed
684 685
	inode = dentry->d_inode;
	if (f->f_mode & FMODE_WRITE) {
686
		error = __get_file_write_access(inode, mnt);
Linus Torvalds's avatar
Linus Torvalds committed
687 688
		if (error)
			goto cleanup_file;
689 690
		if (!special_file(inode->i_mode))
			file_take_write(f);
Linus Torvalds's avatar
Linus Torvalds committed
691 692 693
	}

	f->f_mapping = inode->i_mapping;
694 695
	f->f_path.dentry = dentry;
	f->f_path.mnt = mnt;
Linus Torvalds's avatar
Linus Torvalds committed
696
	f->f_pos = 0;
Nick Piggin's avatar
Nick Piggin committed
697
	file_sb_list_add(f, inode->i_sb);
Linus Torvalds's avatar
Linus Torvalds committed
698

699 700 701 702 703 704 705
	if (unlikely(f->f_mode & FMODE_PATH)) {
		f->f_op = &empty_fops;
		return f;
	}

	f->f_op = fops_get(inode->i_fop);

706
	error = security_dentry_open(f, cred);
707 708 709
	if (error)
		goto cleanup_all;

710 711 712 713
	if (!open && f->f_op)
		open = f->f_op->open;
	if (open) {
		error = open(inode, f);
Linus Torvalds's avatar
Linus Torvalds committed
714 715 716
		if (error)
			goto cleanup_all;
	}
717 718
	if ((f->f_mode & (FMODE_READ | FMODE_WRITE)) == FMODE_READ)
		i_readcount_inc(inode);
719

Linus Torvalds's avatar
Linus Torvalds committed
720 721 722 723 724 725
	f->f_flags &= ~(O_CREAT | O_EXCL | O_NOCTTY | O_TRUNC);

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

	/* NB: we're sure to have correct a_ops only after f_op->open */
	if (f->f_flags & O_DIRECT) {
726 727
		if (!f->f_mapping->a_ops ||
		    ((!f->f_mapping->a_ops->direct_IO) &&
728
		    (!f->f_mapping->a_ops->get_xip_mem))) {
Linus Torvalds's avatar
Linus Torvalds committed
729 730 731 732 733 734 735 736 737
			fput(f);
			f = ERR_PTR(-EINVAL);
		}
	}

	return f;

cleanup_all:
	fops_put(f->f_op);
738
	if (f->f_mode & FMODE_WRITE) {
Linus Torvalds's avatar
Linus Torvalds committed
739
		put_write_access(inode);
740 741 742 743 744 745 746 747
		if (!special_file(inode->i_mode)) {
			/*
			 * We don't consider this a real
			 * mnt_want/drop_write() pair
			 * because it all happenend right
			 * here, so just reset the state.
			 */
			file_reset_write(f);
748
			mnt_drop_write(mnt);
749
		}
750
	}
Nick Piggin's avatar
Nick Piggin committed
751
	file_sb_list_del(f);
752 753
	f->f_path.dentry = NULL;
	f->f_path.mnt = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
754 755 756 757 758 759 760
cleanup_file:
	put_filp(f);
	dput(dentry);
	mntput(mnt);
	return ERR_PTR(error);
}

761 762 763 764 765 766 767 768 769 770
/**
 * lookup_instantiate_filp - instantiates the open intent filp
 * @nd: pointer to nameidata
 * @dentry: pointer to dentry
 * @open: open callback
 *
 * Helper for filesystems that want to use lookup open intents and pass back
 * a fully instantiated struct file to the caller.
 * This function is meant to be called from within a filesystem's
 * lookup method.
771 772 773 774
 * Beware of calling it for non-regular files! Those ->open methods might block
 * (e.g. in fifo_open), leaving you with parent locked (and in case of fifo,
 * leading to a deadlock, as nobody can open that fifo anymore, because
 * another process to open fifo will block on locked parent when doing lookup).
775 776 777 778 779 780 781 782
 * Note that in case of error, nd->intent.open.file is destroyed, but the
 * path information remains valid.
 * If the open callback is set to NULL, then the standard f_op->open()
 * filesystem callback is substituted.
 */
struct file *lookup_instantiate_filp(struct nameidata *nd, struct dentry *dentry,
		int (*open)(struct inode *, struct file *))
{
783 784
	const struct cred *cred = current_cred();

785 786 787 788
	if (IS_ERR(nd->intent.open.file))
		goto out;
	if (IS_ERR(dentry))
		goto out_err;
789
	nd->intent.open.file = __dentry_open(dget(dentry), mntget(nd->path.mnt),
790
					     nd->intent.open.file,
791
					     open, cred);
792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807
out:
	return nd->intent.open.file;
out_err:
	release_open_intent(nd);
	nd->intent.open.file = (struct file *)dentry;
	goto out;
}
EXPORT_SYMBOL_GPL(lookup_instantiate_filp);

/**
 * nameidata_to_filp - convert a nameidata to an open filp.
 * @nd: pointer to nameidata
 * @flags: open flags
 *
 * Note that this function destroys the original nameidata
 */
808
struct file *nameidata_to_filp(struct nameidata *nd)
809
{
810
	const struct cred *cred = current_cred();
811 812 813 814
	struct file *filp;

	/* Pick up the filp from the open intent */
	filp = nd->intent.open.file;
815 816
	nd->intent.open.file = NULL;

817
	/* Has the filesystem initialised the file for us? */
Al Viro's avatar
Al Viro committed
818 819
	if (filp->f_path.dentry == NULL) {
		path_get(&nd->path);
820
		filp = __dentry_open(nd->path.dentry, nd->path.mnt, filp,
821
				     NULL, cred);
Al Viro's avatar
Al Viro committed
822
	}
823 824 825
	return filp;
}

826 827 828 829
/*
 * dentry_open() will have done dput(dentry) and mntput(mnt) if it returns an
 * error.
 */
830 831
struct file *dentry_open(struct dentry *dentry, struct vfsmount *mnt, int flags,
			 const struct cred *cred)
832 833 834 835
{
	int error;
	struct file *f;

836 837
	validate_creds(cred);

838 839 840 841 842 843 844 845 846 847 848 849
	/*
	 * We must always pass in a valid mount pointer.   Historically
	 * callers got away with not passing it, but we must enforce this at
	 * the earliest possible point now to avoid strange problems deep in the
	 * filesystem stack.
	 */
	if (!mnt) {
		printk(KERN_WARNING "%s called with NULL vfsmount\n", __func__);
		dump_stack();
		return ERR_PTR(-EINVAL);
	}

850 851
	error = -ENFILE;
	f = get_empty_filp();
852 853 854
	if (f == NULL) {
		dput(dentry);
		mntput(mnt);
855
		return ERR_PTR(error);
856
	}
857

858 859
	f->f_flags = flags;
	return __dentry_open(dentry, mnt, f, NULL, cred);
860
}
Linus Torvalds's avatar
Linus Torvalds committed
861 862
EXPORT_SYMBOL(dentry_open);

863
static void __put_unused_fd(struct files_struct *files, unsigned int fd)
Linus Torvalds's avatar
Linus Torvalds committed
864
{
865 866
	struct fdtable *fdt = files_fdtable(files);
	__FD_CLR(fd, fdt->open_fds);
867 868
	if (fd < files->next_fd)
		files->next_fd = fd;
Linus Torvalds's avatar
Linus Torvalds committed
869 870
}

871
void put_unused_fd(unsigned int fd)
Linus Torvalds's avatar
Linus Torvalds committed
872 873 874 875 876 877 878 879 880 881
{
	struct files_struct *files = current->files;
	spin_lock(&files->file_lock);
	__put_unused_fd(files, fd);
	spin_unlock(&files->file_lock);
}

EXPORT_SYMBOL(put_unused_fd);

/*
882
 * Install a file pointer in the fd array.
Linus Torvalds's avatar
Linus Torvalds committed
883 884 885 886 887 888 889 890 891 892 893
 *
 * The VFS is full of places where we drop the files lock between
 * setting the open_fds bitmap and installing the file in the file
 * array.  At any such point, we are vulnerable to a dup2() race
 * installing a file in the array before us.  We need to detect this and
 * fput() the struct file we are about to overwrite in this case.
 *
 * It should never happen - if we allow dup2() do it, _really_ bad things
 * will follow.
 */

894
void fd_install(unsigned int fd, struct file *file)
Linus Torvalds's avatar
Linus Torvalds committed
895 896
{
	struct files_struct *files = current->files;
897
	struct fdtable *fdt;
Linus Torvalds's avatar
Linus Torvalds committed
898
	spin_lock(&files->file_lock);
899
	fdt = files_fdtable(files);
900 901
	BUG_ON(fdt->fd[fd] != NULL);
	rcu_assign_pointer(fdt->fd[fd], file);
Linus Torvalds's avatar
Linus Torvalds committed
902 903 904 905 906
	spin_unlock(&files->file_lock);
}

EXPORT_SYMBOL(fd_install);

907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927
static inline int build_open_flags(int flags, int mode, struct open_flags *op)
{
	int lookup_flags = 0;
	int acc_mode;

	if (!(flags & O_CREAT))
		mode = 0;
	op->mode = mode;

	/* Must never be set by userspace */
	flags &= ~FMODE_NONOTIFY;

	/*
	 * 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;

928 929 930 931 932 933 934 935 936 937
	/*
	 * If we have O_PATH in the open flag. Then we
	 * cannot have anything other than the below set of flags
	 */
	if (flags & O_PATH) {
		flags &= O_DIRECTORY | O_NOFOLLOW | O_PATH;
		acc_mode = 0;
	} else {
		acc_mode = MAY_OPEN | ACC_MODE(flags);
	}
938

939
	op->open_flag = flags;
940 941 942 943 944 945 946 947 948 949 950 951

	/* 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;

952 953
	op->intent = flags & O_PATH ? 0 : LOOKUP_OPEN;

954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985
	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;
	return lookup_flags;
}

/**
 * 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..
 */
struct file *filp_open(const char *filename, int flags, int mode)
{
	struct open_flags op;
	int lookup = build_open_flags(flags, mode, &op);
	return do_filp_open(AT_FDCWD, filename, &op, lookup);
}
EXPORT_SYMBOL(filp_open);

986 987 988 989 990 991 992 993 994 995 996 997 998 999
struct file *file_open_root(struct dentry *dentry, struct vfsmount *mnt,
			    const char *filename, int flags)
{
	struct open_flags op;
	int lookup = build_open_flags(flags, 0, &op);
	if (flags & O_CREAT)
		return ERR_PTR(-EINVAL);
	if (!filename && (flags & O_DIRECTORY))
		if (!dentry->d_inode->i_op->lookup)
			return ERR_PTR(-ENOTDIR);
	return do_file_open_root(dentry, mnt, filename, &op, lookup);
}
EXPORT_SYMBOL(file_open_root);

1000
long do_sys_open(int dfd, const char __user *filename, int flags, int mode)
Linus Torvalds's avatar
Linus Torvalds committed
1001
{
1002 1003
	struct open_flags op;
	int lookup = build_open_flags(flags, mode, &op);
1004 1005
	char *tmp = getname(filename);
	int fd = PTR_ERR(tmp);
Linus Torvalds's avatar
Linus Torvalds committed
1006 1007

	if (!IS_ERR(tmp)) {
Ulrich Drepper's avatar
Ulrich Drepper committed
1008
		fd = get_unused_fd_flags(flags);
Linus Torvalds's avatar
Linus Torvalds committed
1009
		if (fd >= 0) {
1010
			struct file *f = do_filp_open(dfd, tmp, &op, lookup);
1011 1012 1013 1014
			if (IS_ERR(f)) {
				put_unused_fd(fd);
				fd = PTR_ERR(f);
			} else {
1015
				fsnotify_open(f);
1016 1017
				fd_install(fd, f);
			}
Linus Torvalds's avatar
Linus Torvalds committed
1018 1019 1020 1021 1022
		}
		putname(tmp);
	}
	return fd;
}
1023

1024
SYSCALL_DEFINE3(open, const char __user *, filename, int, flags, int, mode)
1025
{