namei.c 82.9 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
/*
 *  linux/fs/namei.c
 *
 *  Copyright (C) 1991, 1992  Linus Torvalds
 */

/*
 * Some corrections by tytso.
 */

/* [Feb 1997 T. Schoebel-Theuer] Complete rewrite of the pathname
 * lookup logic.
 */
/* [Feb-Apr 2000, AV] Rewrite to the new namespace architecture.
 */

#include <linux/init.h>
#include <linux/module.h>
#include <linux/slab.h>
#include <linux/fs.h>
#include <linux/namei.h>
#include <linux/pagemap.h>
R
Robert Love 已提交
23
#include <linux/fsnotify.h>
L
Linus Torvalds 已提交
24 25
#include <linux/personality.h>
#include <linux/security.h>
M
Mimi Zohar 已提交
26
#include <linux/ima.h>
L
Linus Torvalds 已提交
27 28 29
#include <linux/syscalls.h>
#include <linux/mount.h>
#include <linux/audit.h>
30
#include <linux/capability.h>
31
#include <linux/file.h>
32
#include <linux/fcntl.h>
33
#include <linux/device_cgroup.h>
34
#include <linux/fs_struct.h>
L
Linus Torvalds 已提交
35 36
#include <asm/uaccess.h>

37 38
#include "internal.h"

L
Linus Torvalds 已提交
39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72
/* [Feb-1997 T. Schoebel-Theuer]
 * Fundamental changes in the pathname lookup mechanisms (namei)
 * were necessary because of omirr.  The reason is that omirr needs
 * to know the _real_ pathname, not the user-supplied one, in case
 * of symlinks (and also when transname replacements occur).
 *
 * The new code replaces the old recursive symlink resolution with
 * an iterative one (in case of non-nested symlink chains).  It does
 * this with calls to <fs>_follow_link().
 * As a side effect, dir_namei(), _namei() and follow_link() are now 
 * replaced with a single function lookup_dentry() that can handle all 
 * the special cases of the former code.
 *
 * With the new dcache, the pathname is stored at each inode, at least as
 * long as the refcount of the inode is positive.  As a side effect, the
 * size of the dcache depends on the inode cache and thus is dynamic.
 *
 * [29-Apr-1998 C. Scott Ananian] Updated above description of symlink
 * resolution to correspond with current state of the code.
 *
 * Note that the symlink resolution is not *completely* iterative.
 * There is still a significant amount of tail- and mid- recursion in
 * the algorithm.  Also, note that <fs>_readlink() is not used in
 * lookup_dentry(): lookup_dentry() on the result of <fs>_readlink()
 * may return different results than <fs>_follow_link().  Many virtual
 * filesystems (including /proc) exhibit this behavior.
 */

/* [24-Feb-97 T. Schoebel-Theuer] Side effects caused by new implementation:
 * New symlink semantics: when open() is called with flags O_CREAT | O_EXCL
 * and the name already exists in form of a symlink, try to create the new
 * name indicated by the symlink. The old code always complained that the
 * name already exists, due to not following the symlink even if its target
 * is nonexistent.  The new semantics affects also mknod() and link() when
L
Lucas De Marchi 已提交
73
 * the name is a symlink pointing to a non-existent name.
L
Linus Torvalds 已提交
74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105
 *
 * I don't know which semantics is the right one, since I have no access
 * to standards. But I found by trial that HP-UX 9.0 has the full "new"
 * semantics implemented, while SunOS 4.1.1 and Solaris (SunOS 5.4) have the
 * "old" one. Personally, I think the new semantics is much more logical.
 * Note that "ln old new" where "new" is a symlink pointing to a non-existing
 * file does succeed in both HP-UX and SunOs, but not in Solaris
 * and in the old Linux semantics.
 */

/* [16-Dec-97 Kevin Buhr] For security reasons, we change some symlink
 * semantics.  See the comments in "open_namei" and "do_link" below.
 *
 * [10-Sep-98 Alan Modra] Another symlink change.
 */

/* [Feb-Apr 2000 AV] Complete rewrite. Rules for symlinks:
 *	inside the path - always follow.
 *	in the last component in creation/removal/renaming - never follow.
 *	if LOOKUP_FOLLOW passed - follow.
 *	if the pathname has trailing slashes - follow.
 *	otherwise - don't follow.
 * (applied in that order).
 *
 * [Jun 2000 AV] Inconsistent behaviour of open() in case if flags==O_CREAT
 * restored for 2.4. This is the last surviving part of old 4.2BSD bug.
 * During the 2.4 we need to fix the userland stuff depending on it -
 * hopefully we will be able to get rid of that wart in 2.5. So far only
 * XEmacs seems to be relying on it...
 */
/*
 * [Sep 2001 AV] Single-semaphore locking scheme (kudos to David Holland)
106
 * implemented.  Let's see if raised priority of ->s_vfs_rename_mutex gives
L
Linus Torvalds 已提交
107 108 109 110 111 112 113 114 115 116
 * any extra contention...
 */

/* In order to reduce some races, while at the same time doing additional
 * checking and hopefully speeding things up, we copy filenames to the
 * kernel data space before using them..
 *
 * POSIX.1 2.4: an empty pathname is invalid (ENOENT).
 * PATH_MAX includes the nul terminator --RR.
 */
117
static int do_getname(const char __user *filename, char *page)
L
Linus Torvalds 已提交
118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138
{
	int retval;
	unsigned long len = PATH_MAX;

	if (!segment_eq(get_fs(), KERNEL_DS)) {
		if ((unsigned long) filename >= TASK_SIZE)
			return -EFAULT;
		if (TASK_SIZE - (unsigned long) filename < PATH_MAX)
			len = TASK_SIZE - (unsigned long) filename;
	}

	retval = strncpy_from_user(page, filename, len);
	if (retval > 0) {
		if (retval < len)
			return 0;
		return -ENAMETOOLONG;
	} else if (!retval)
		retval = -ENOENT;
	return retval;
}

A
Al Viro 已提交
139
static char *getname_flags(const char __user * filename, int flags)
L
Linus Torvalds 已提交
140 141 142 143 144 145 146 147 148 149
{
	char *tmp, *result;

	result = ERR_PTR(-ENOMEM);
	tmp = __getname();
	if (tmp)  {
		int retval = do_getname(filename, tmp);

		result = tmp;
		if (retval < 0) {
A
Al Viro 已提交
150 151 152 153
			if (retval != -ENOENT || !(flags & LOOKUP_EMPTY)) {
				__putname(tmp);
				result = ERR_PTR(retval);
			}
L
Linus Torvalds 已提交
154 155 156 157 158 159
		}
	}
	audit_getname(result);
	return result;
}

A
Al Viro 已提交
160 161 162 163 164
char *getname(const char __user * filename)
{
	return getname_flags(filename, 0);
}

L
Linus Torvalds 已提交
165 166 167
#ifdef CONFIG_AUDITSYSCALL
void putname(const char *name)
{
168
	if (unlikely(!audit_dummy_context()))
L
Linus Torvalds 已提交
169 170 171 172 173 174 175
		audit_putname(name);
	else
		__putname(name);
}
EXPORT_SYMBOL(putname);
#endif

176 177
/*
 * This does basic POSIX ACL permission checking
L
Linus Torvalds 已提交
178
 */
179 180
static int acl_permission_check(struct inode *inode, int mask, unsigned int flags,
		int (*check_acl)(struct inode *inode, int mask, unsigned int flags))
L
Linus Torvalds 已提交
181
{
182
	unsigned int mode = inode->i_mode;
L
Linus Torvalds 已提交
183

184 185
	mask &= MAY_READ | MAY_WRITE | MAY_EXEC;

186 187 188
	if (current_user_ns() != inode_userns(inode))
		goto other_perms;

189
	if (current_fsuid() == inode->i_uid)
L
Linus Torvalds 已提交
190 191 192
		mode >>= 6;
	else {
		if (IS_POSIXACL(inode) && (mode & S_IRWXG) && check_acl) {
193 194 195
			int error = check_acl(inode, mask, flags);
			if (error != -EAGAIN)
				return error;
L
Linus Torvalds 已提交
196 197 198 199 200 201
		}

		if (in_group_p(inode->i_gid))
			mode >>= 3;
	}

202
other_perms:
L
Linus Torvalds 已提交
203 204 205
	/*
	 * If the DACs are ok we don't need any capability check.
	 */
206
	if ((mask & ~mode) == 0)
L
Linus Torvalds 已提交
207
		return 0;
208 209 210 211
	return -EACCES;
}

/**
212
 * generic_permission -  check for access rights on a Posix-like filesystem
213 214 215
 * @inode:	inode to check access rights for
 * @mask:	right to check for (%MAY_READ, %MAY_WRITE, %MAY_EXEC)
 * @check_acl:	optional callback to check for Posix ACLs
216
 * @flags:	IPERM_FLAG_ flags.
217 218 219 220
 *
 * Used to check for read/write/execute permissions on a file.
 * We use "fsuid" for this, letting us set arbitrary permissions
 * for filesystem access without changing the "normal" uids which
221 222 223 224 225
 * are used for other things.
 *
 * generic_permission is rcu-walk aware. It returns -ECHILD in case an rcu-walk
 * request cannot be satisfied (eg. requires blocking or too much complexity).
 * It would then be called again in ref-walk mode.
226
 */
227 228
int generic_permission(struct inode *inode, int mask, unsigned int flags,
	int (*check_acl)(struct inode *inode, int mask, unsigned int flags))
229 230 231 232 233 234
{
	int ret;

	/*
	 * Do the basic POSIX ACL permission checks.
	 */
235
	ret = acl_permission_check(inode, mask, flags, check_acl);
236 237
	if (ret != -EACCES)
		return ret;
L
Linus Torvalds 已提交
238 239 240

	/*
	 * Read/write DACs are always overridable.
A
Al Viro 已提交
241 242
	 * Executable DACs are overridable for all directories and
	 * for non-directories that have least one exec bit set.
L
Linus Torvalds 已提交
243
	 */
244
	if (!(mask & MAY_EXEC) || execute_ok(inode))
245
		if (ns_capable(inode_userns(inode), CAP_DAC_OVERRIDE))
L
Linus Torvalds 已提交
246 247 248 249 250
			return 0;

	/*
	 * Searching includes executable on directories, else just read.
	 */
251
	mask &= MAY_READ | MAY_WRITE | MAY_EXEC;
L
Linus Torvalds 已提交
252
	if (mask == MAY_READ || (S_ISDIR(inode->i_mode) && !(mask & MAY_WRITE)))
253
		if (ns_capable(inode_userns(inode), CAP_DAC_READ_SEARCH))
L
Linus Torvalds 已提交
254 255 256 257 258
			return 0;

	return -EACCES;
}

C
Christoph Hellwig 已提交
259 260 261 262 263 264 265 266 267 268
/**
 * inode_permission  -  check for access rights to a given inode
 * @inode:	inode to check permission on
 * @mask:	right to check for (%MAY_READ, %MAY_WRITE, %MAY_EXEC)
 *
 * Used to check for read/write/execute permissions on an inode.
 * We use "fsuid" for this, letting us set arbitrary permissions
 * for filesystem access without changing the "normal" uids which
 * are used for other things.
 */
269
int inode_permission(struct inode *inode, int mask)
L
Linus Torvalds 已提交
270
{
271
	int retval;
L
Linus Torvalds 已提交
272 273

	if (mask & MAY_WRITE) {
274
		umode_t mode = inode->i_mode;
L
Linus Torvalds 已提交
275 276 277 278 279 280 281 282 283 284 285 286 287 288 289

		/*
		 * Nobody gets write access to a read-only fs.
		 */
		if (IS_RDONLY(inode) &&
		    (S_ISREG(mode) || S_ISDIR(mode) || S_ISLNK(mode)))
			return -EROFS;

		/*
		 * Nobody gets write access to an immutable file.
		 */
		if (IS_IMMUTABLE(inode))
			return -EACCES;
	}

A
Al Viro 已提交
290
	if (inode->i_op->permission)
291
		retval = inode->i_op->permission(inode, mask, 0);
292
	else
293 294
		retval = generic_permission(inode, mask, 0,
				inode->i_op->check_acl);
295

L
Linus Torvalds 已提交
296 297 298
	if (retval)
		return retval;

299 300 301 302
	retval = devcgroup_inode_permission(inode, mask);
	if (retval)
		return retval;

303
	return security_inode_permission(inode, mask);
L
Linus Torvalds 已提交
304 305
}

306 307 308 309 310 311 312 313 314 315
/**
 * file_permission  -  check for additional access rights to a given file
 * @file:	file to check access rights for
 * @mask:	right to check for (%MAY_READ, %MAY_WRITE, %MAY_EXEC)
 *
 * Used to check for read/write/execute permissions on an already opened
 * file.
 *
 * Note:
 *	Do not use this function in new code.  All access checks should
C
Christoph Hellwig 已提交
316
 *	be done using inode_permission().
317 318 319
 */
int file_permission(struct file *file, int mask)
{
320
	return inode_permission(file->f_path.dentry->d_inode, mask);
321 322
}

L
Linus Torvalds 已提交
323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355
/*
 * get_write_access() gets write permission for a file.
 * put_write_access() releases this write permission.
 * This is used for regular files.
 * We cannot support write (and maybe mmap read-write shared) accesses and
 * MAP_DENYWRITE mmappings simultaneously. The i_writecount field of an inode
 * can have the following values:
 * 0: no writers, no VM_DENYWRITE mappings
 * < 0: (-i_writecount) vm_area_structs with VM_DENYWRITE set exist
 * > 0: (i_writecount) users are writing to the file.
 *
 * Normally we operate on that counter with atomic_{inc,dec} and it's safe
 * except for the cases where we don't hold i_writecount yet. Then we need to
 * use {get,deny}_write_access() - these functions check the sign and refuse
 * to do the change if sign is wrong. Exclusion between them is provided by
 * the inode->i_lock spinlock.
 */

int get_write_access(struct inode * inode)
{
	spin_lock(&inode->i_lock);
	if (atomic_read(&inode->i_writecount) < 0) {
		spin_unlock(&inode->i_lock);
		return -ETXTBSY;
	}
	atomic_inc(&inode->i_writecount);
	spin_unlock(&inode->i_lock);

	return 0;
}

int deny_write_access(struct file * file)
{
356
	struct inode *inode = file->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
357 358 359 360 361 362 363 364 365 366 367 368

	spin_lock(&inode->i_lock);
	if (atomic_read(&inode->i_writecount) > 0) {
		spin_unlock(&inode->i_lock);
		return -ETXTBSY;
	}
	atomic_dec(&inode->i_writecount);
	spin_unlock(&inode->i_lock);

	return 0;
}

J
Jan Blunck 已提交
369 370 371 372 373 374 375 376 377 378 379 380 381
/**
 * path_get - get a reference to a path
 * @path: path to get the reference to
 *
 * Given a path increment the reference count to the dentry and the vfsmount.
 */
void path_get(struct path *path)
{
	mntget(path->mnt);
	dget(path->dentry);
}
EXPORT_SYMBOL(path_get);

J
Jan Blunck 已提交
382 383 384 385 386 387 388
/**
 * path_put - put a reference to a path
 * @path: path to put the reference to
 *
 * Given a path decrement the reference count to the dentry and the vfsmount.
 */
void path_put(struct path *path)
L
Linus Torvalds 已提交
389
{
J
Jan Blunck 已提交
390 391
	dput(path->dentry);
	mntput(path->mnt);
L
Linus Torvalds 已提交
392
}
J
Jan Blunck 已提交
393
EXPORT_SYMBOL(path_put);
L
Linus Torvalds 已提交
394

A
Al Viro 已提交
395
/*
N
Nick Piggin 已提交
396
 * Path walking has 2 modes, rcu-walk and ref-walk (see
A
Al Viro 已提交
397 398 399 400 401 402 403
 * Documentation/filesystems/path-lookup.txt).  In situations when we can't
 * continue in RCU mode, we attempt to drop out of rcu-walk mode and grab
 * normal reference counts on dentries and vfsmounts to transition to rcu-walk
 * mode.  Refcounts are grabbed at the last known good point before rcu-walk
 * got stuck, so ref-walk may continue from there. If this is not successful
 * (eg. a seqcount has changed), then failure is returned and it's up to caller
 * to restart the path walk from the beginning in ref-walk mode.
N
Nick Piggin 已提交
404 405 406
 */

/**
A
Al Viro 已提交
407 408 409
 * unlazy_walk - try to switch to ref-walk mode.
 * @nd: nameidata pathwalk data
 * @dentry: child of nd->path.dentry or NULL
410
 * Returns: 0 on success, -ECHILD on failure
N
Nick Piggin 已提交
411
 *
A
Al Viro 已提交
412 413 414
 * unlazy_walk attempts to legitimize the current nd->path, nd->root and dentry
 * for ref-walk mode.  @dentry must be a path found by a do_lookup call on
 * @nd or NULL.  Must be called from rcu-walk context.
N
Nick Piggin 已提交
415
 */
A
Al Viro 已提交
416
static int unlazy_walk(struct nameidata *nd, struct dentry *dentry)
N
Nick Piggin 已提交
417 418 419
{
	struct fs_struct *fs = current->fs;
	struct dentry *parent = nd->path.dentry;
420
	int want_root = 0;
N
Nick Piggin 已提交
421 422

	BUG_ON(!(nd->flags & LOOKUP_RCU));
423 424
	if (nd->root.mnt && !(nd->flags & LOOKUP_ROOT)) {
		want_root = 1;
N
Nick Piggin 已提交
425 426 427 428 429 430
		spin_lock(&fs->lock);
		if (nd->root.mnt != fs->root.mnt ||
				nd->root.dentry != fs->root.dentry)
			goto err_root;
	}
	spin_lock(&parent->d_lock);
A
Al Viro 已提交
431 432 433 434 435
	if (!dentry) {
		if (!__d_rcu_to_refcount(parent, nd->seq))
			goto err_parent;
		BUG_ON(nd->inode != parent->d_inode);
	} else {
436 437
		if (dentry->d_parent != parent)
			goto err_parent;
A
Al Viro 已提交
438 439 440 441 442 443 444 445 446 447 448 449 450 451
		spin_lock_nested(&dentry->d_lock, DENTRY_D_LOCK_NESTED);
		if (!__d_rcu_to_refcount(dentry, nd->seq))
			goto err_child;
		/*
		 * If the sequence check on the child dentry passed, then
		 * the child has not been removed from its parent. This
		 * means the parent dentry must be valid and able to take
		 * a reference at this point.
		 */
		BUG_ON(!IS_ROOT(dentry) && dentry->d_parent != parent);
		BUG_ON(!parent->d_count);
		parent->d_count++;
		spin_unlock(&dentry->d_lock);
	}
N
Nick Piggin 已提交
452
	spin_unlock(&parent->d_lock);
453
	if (want_root) {
N
Nick Piggin 已提交
454 455 456 457 458 459 460 461 462
		path_get(&nd->root);
		spin_unlock(&fs->lock);
	}
	mntget(nd->path.mnt);

	rcu_read_unlock();
	br_read_unlock(vfsmount_lock);
	nd->flags &= ~LOOKUP_RCU;
	return 0;
A
Al Viro 已提交
463 464

err_child:
N
Nick Piggin 已提交
465
	spin_unlock(&dentry->d_lock);
A
Al Viro 已提交
466
err_parent:
N
Nick Piggin 已提交
467 468
	spin_unlock(&parent->d_lock);
err_root:
469
	if (want_root)
N
Nick Piggin 已提交
470 471 472 473
		spin_unlock(&fs->lock);
	return -ECHILD;
}

474 475 476 477 478 479
/**
 * release_open_intent - free up open intent resources
 * @nd: pointer to nameidata
 */
void release_open_intent(struct nameidata *nd)
{
480 481 482 483 484 485 486 487
	struct file *file = nd->intent.open.file;

	if (file && !IS_ERR(file)) {
		if (file->f_path.dentry == NULL)
			put_filp(file);
		else
			fput(file);
	}
488 489
}

A
Al Viro 已提交
490
static inline int d_revalidate(struct dentry *dentry, struct nameidata *nd)
491
{
A
Al Viro 已提交
492
	return dentry->d_op->d_revalidate(dentry, nd);
493 494
}

495
static struct dentry *
496 497
do_revalidate(struct dentry *dentry, struct nameidata *nd)
{
498
	int status = d_revalidate(dentry, nd);
499 500 501 502 503 504 505
	if (unlikely(status <= 0)) {
		/*
		 * The dentry failed validation.
		 * If d_revalidate returned 0 attempt to invalidate
		 * the dentry otherwise d_revalidate is asking us
		 * to return a fail status.
		 */
506
		if (status < 0) {
507
			dput(dentry);
508
			dentry = ERR_PTR(status);
509 510 511
		} else if (!d_invalidate(dentry)) {
			dput(dentry);
			dentry = NULL;
512 513 514 515 516
		}
	}
	return dentry;
}

517 518 519
/**
 * complete_walk - successful completion of path walk
 * @nd:  pointer nameidata
520
 *
521 522 523 524 525
 * If we had been in RCU mode, drop out of it and legitimize nd->path.
 * Revalidate the final result, unless we'd already done that during
 * the path walk or the filesystem doesn't ask for it.  Return 0 on
 * success, -error on failure.  In case of failure caller does not
 * need to drop nd->path.
526
 */
527
static int complete_walk(struct nameidata *nd)
528
{
A
Al Viro 已提交
529
	struct dentry *dentry = nd->path.dentry;
530 531
	int status;

532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549
	if (nd->flags & LOOKUP_RCU) {
		nd->flags &= ~LOOKUP_RCU;
		if (!(nd->flags & LOOKUP_ROOT))
			nd->root.mnt = NULL;
		spin_lock(&dentry->d_lock);
		if (unlikely(!__d_rcu_to_refcount(dentry, nd->seq))) {
			spin_unlock(&dentry->d_lock);
			rcu_read_unlock();
			br_read_unlock(vfsmount_lock);
			return -ECHILD;
		}
		BUG_ON(nd->inode != dentry->d_inode);
		spin_unlock(&dentry->d_lock);
		mntget(nd->path.mnt);
		rcu_read_unlock();
		br_read_unlock(vfsmount_lock);
	}

A
Al Viro 已提交
550 551 552 553
	if (likely(!(nd->flags & LOOKUP_JUMPED)))
		return 0;

	if (likely(!(dentry->d_flags & DCACHE_OP_REVALIDATE)))
554 555
		return 0;

A
Al Viro 已提交
556 557 558 559
	if (likely(!(dentry->d_sb->s_type->fs_flags & FS_REVAL_DOT)))
		return 0;

	/* Note: we do not d_invalidate() */
560
	status = d_revalidate(dentry, nd);
561 562 563
	if (status > 0)
		return 0;

A
Al Viro 已提交
564
	if (!status)
565
		status = -ESTALE;
A
Al Viro 已提交
566

567
	path_put(&nd->path);
568 569 570
	return status;
}

L
Linus Torvalds 已提交
571
/*
A
Al Viro 已提交
572 573 574
 * Short-cut version of permission(), for calling on directories
 * during pathname resolution.  Combines parts of permission()
 * and generic_permission(), and tests ONLY for MAY_EXEC permission.
L
Linus Torvalds 已提交
575 576
 *
 * If appropriate, check DAC only.  If not appropriate, or
A
Al Viro 已提交
577
 * short-cut DAC fails, then call ->permission() to do more
L
Linus Torvalds 已提交
578 579
 * complete permission check.
 */
580
static inline int exec_permission(struct inode *inode, unsigned int flags)
L
Linus Torvalds 已提交
581
{
582
	int ret;
583
	struct user_namespace *ns = inode_userns(inode);
L
Linus Torvalds 已提交
584

585
	if (inode->i_op->permission) {
586 587 588 589
		ret = inode->i_op->permission(inode, MAY_EXEC, flags);
	} else {
		ret = acl_permission_check(inode, MAY_EXEC, flags,
				inode->i_op->check_acl);
590
	}
591
	if (likely(!ret))
L
Linus Torvalds 已提交
592
		goto ok;
593
	if (ret == -ECHILD)
N
Nick Piggin 已提交
594
		return ret;
L
Linus Torvalds 已提交
595

596 597
	if (ns_capable(ns, CAP_DAC_OVERRIDE) ||
			ns_capable(ns, CAP_DAC_READ_SEARCH))
L
Linus Torvalds 已提交
598 599
		goto ok;

600
	return ret;
L
Linus Torvalds 已提交
601
ok:
602
	return security_inode_exec_permission(inode, flags);
L
Linus Torvalds 已提交
603 604
}

A
Al Viro 已提交
605 606
static __always_inline void set_root(struct nameidata *nd)
{
607 608
	if (!nd->root.mnt)
		get_fs_root(current->fs, &nd->root);
A
Al Viro 已提交
609 610
}

611 612
static int link_path_walk(const char *, struct nameidata *);

N
Nick Piggin 已提交
613 614 615 616
static __always_inline void set_root_rcu(struct nameidata *nd)
{
	if (!nd->root.mnt) {
		struct fs_struct *fs = current->fs;
N
Nick Piggin 已提交
617 618 619 620 621
		unsigned seq;

		do {
			seq = read_seqcount_begin(&fs->seq);
			nd->root = fs->root;
622
			nd->seq = __read_seqcount_begin(&nd->root.dentry->d_seq);
N
Nick Piggin 已提交
623
		} while (read_seqcount_retry(&fs->seq, seq));
N
Nick Piggin 已提交
624 625 626
	}
}

627
static __always_inline int __vfs_follow_link(struct nameidata *nd, const char *link)
L
Linus Torvalds 已提交
628
{
N
Nick Piggin 已提交
629 630
	int ret;

L
Linus Torvalds 已提交
631 632 633 634
	if (IS_ERR(link))
		goto fail;

	if (*link == '/') {
A
Al Viro 已提交
635
		set_root(nd);
J
Jan Blunck 已提交
636
		path_put(&nd->path);
A
Al Viro 已提交
637 638
		nd->path = nd->root;
		path_get(&nd->root);
A
Al Viro 已提交
639
		nd->flags |= LOOKUP_JUMPED;
L
Linus Torvalds 已提交
640
	}
N
Nick Piggin 已提交
641
	nd->inode = nd->path.dentry->d_inode;
C
Christoph Hellwig 已提交
642

N
Nick Piggin 已提交
643 644
	ret = link_path_walk(link, nd);
	return ret;
L
Linus Torvalds 已提交
645
fail:
J
Jan Blunck 已提交
646
	path_put(&nd->path);
L
Linus Torvalds 已提交
647 648 649
	return PTR_ERR(link);
}

J
Jan Blunck 已提交
650
static void path_put_conditional(struct path *path, struct nameidata *nd)
651 652
{
	dput(path->dentry);
653
	if (path->mnt != nd->path.mnt)
654 655 656
		mntput(path->mnt);
}

657 658
static inline void path_to_nameidata(const struct path *path,
					struct nameidata *nd)
659
{
N
Nick Piggin 已提交
660 661 662 663
	if (!(nd->flags & LOOKUP_RCU)) {
		dput(nd->path.dentry);
		if (nd->path.mnt != path->mnt)
			mntput(nd->path.mnt);
664
	}
N
Nick Piggin 已提交
665
	nd->path.mnt = path->mnt;
666
	nd->path.dentry = path->dentry;
667 668
}

669 670 671 672 673 674 675 676
static inline void put_link(struct nameidata *nd, struct path *link, void *cookie)
{
	struct inode *inode = link->dentry->d_inode;
	if (!IS_ERR(cookie) && inode->i_op->put_link)
		inode->i_op->put_link(link->dentry, nd, cookie);
	path_put(link);
}

A
Al Viro 已提交
677
static __always_inline int
678
follow_link(struct path *link, struct nameidata *nd, void **p)
L
Linus Torvalds 已提交
679 680
{
	int error;
681
	struct dentry *dentry = link->dentry;
L
Linus Torvalds 已提交
682

683 684
	BUG_ON(nd->flags & LOOKUP_RCU);

A
Al Viro 已提交
685 686 687
	if (link->mnt == nd->path.mnt)
		mntget(link->mnt);

688 689 690 691 692 693 694 695
	if (unlikely(current->total_link_count >= 40)) {
		*p = ERR_PTR(-ELOOP); /* no ->put_link(), please */
		path_put(&nd->path);
		return -ELOOP;
	}
	cond_resched();
	current->total_link_count++;

696
	touch_atime(link->mnt, dentry);
L
Linus Torvalds 已提交
697
	nd_set_link(nd, NULL);
A
Al Viro 已提交
698

699 700 701 702 703 704 705
	error = security_inode_follow_link(link->dentry, nd);
	if (error) {
		*p = ERR_PTR(error); /* no ->put_link(), please */
		path_put(&nd->path);
		return error;
	}

706
	nd->last_type = LAST_BIND;
A
Al Viro 已提交
707 708 709
	*p = dentry->d_inode->i_op->follow_link(dentry, nd);
	error = PTR_ERR(*p);
	if (!IS_ERR(*p)) {
L
Linus Torvalds 已提交
710
		char *s = nd_get_link(nd);
711
		error = 0;
L
Linus Torvalds 已提交
712 713
		if (s)
			error = __vfs_follow_link(nd, s);
A
Al Viro 已提交
714
		else if (nd->last_type == LAST_BIND) {
A
Al Viro 已提交
715
			nd->flags |= LOOKUP_JUMPED;
716 717
			nd->inode = nd->path.dentry->d_inode;
			if (nd->inode->i_op->follow_link) {
A
Al Viro 已提交
718 719 720 721 722
				/* stepped on a _really_ weird one */
				path_put(&nd->path);
				error = -ELOOP;
			}
		}
L
Linus Torvalds 已提交
723 724 725 726
	}
	return error;
}

N
Nick Piggin 已提交
727 728 729 730 731 732 733 734 735 736 737 738 739 740
static int follow_up_rcu(struct path *path)
{
	struct vfsmount *parent;
	struct dentry *mountpoint;

	parent = path->mnt->mnt_parent;
	if (parent == path->mnt)
		return 0;
	mountpoint = path->mnt->mnt_mountpoint;
	path->dentry = mountpoint;
	path->mnt = parent;
	return 1;
}

A
Al Viro 已提交
741
int follow_up(struct path *path)
L
Linus Torvalds 已提交
742 743 744
{
	struct vfsmount *parent;
	struct dentry *mountpoint;
N
Nick Piggin 已提交
745 746

	br_read_lock(vfsmount_lock);
A
Al Viro 已提交
747 748
	parent = path->mnt->mnt_parent;
	if (parent == path->mnt) {
N
Nick Piggin 已提交
749
		br_read_unlock(vfsmount_lock);
L
Linus Torvalds 已提交
750 751 752
		return 0;
	}
	mntget(parent);
A
Al Viro 已提交
753
	mountpoint = dget(path->mnt->mnt_mountpoint);
N
Nick Piggin 已提交
754
	br_read_unlock(vfsmount_lock);
A
Al Viro 已提交
755 756 757 758
	dput(path->dentry);
	path->dentry = mountpoint;
	mntput(path->mnt);
	path->mnt = parent;
L
Linus Torvalds 已提交
759 760 761
	return 1;
}

N
Nick Piggin 已提交
762
/*
763 764 765
 * Perform an automount
 * - return -EISDIR to tell follow_managed() to stop and return the path we
 *   were called with.
L
Linus Torvalds 已提交
766
 */
767 768
static int follow_automount(struct path *path, unsigned flags,
			    bool *need_mntput)
N
Nick Piggin 已提交
769
{
770
	struct vfsmount *mnt;
771
	int err;
772 773 774 775

	if (!path->dentry->d_op || !path->dentry->d_op->d_automount)
		return -EREMOTE;

776 777 778 779 780 781
	/* We don't want to mount if someone supplied AT_NO_AUTOMOUNT
	 * and this is the terminal part of the path.
	 */
	if ((flags & LOOKUP_NO_AUTOMOUNT) && !(flags & LOOKUP_CONTINUE))
		return -EISDIR; /* we actually want to stop here */

782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812
	/* We want to mount if someone is trying to open/create a file of any
	 * type under the mountpoint, wants to traverse through the mountpoint
	 * or wants to open the mounted directory.
	 *
	 * We don't want to mount if someone's just doing a stat and they've
	 * set AT_SYMLINK_NOFOLLOW - unless they're stat'ing a directory and
	 * appended a '/' to the name.
	 */
	if (!(flags & LOOKUP_FOLLOW) &&
	    !(flags & (LOOKUP_CONTINUE | LOOKUP_DIRECTORY |
		       LOOKUP_OPEN | LOOKUP_CREATE)))
		return -EISDIR;

	current->total_link_count++;
	if (current->total_link_count >= 40)
		return -ELOOP;

	mnt = path->dentry->d_op->d_automount(path);
	if (IS_ERR(mnt)) {
		/*
		 * The filesystem is allowed to return -EISDIR here to indicate
		 * it doesn't want to automount.  For instance, autofs would do
		 * this so that its userspace daemon can mount on this dentry.
		 *
		 * However, we can only permit this if it's a terminal point in
		 * the path being looked up; if it wasn't then the remainder of
		 * the path is inaccessible and we should say so.
		 */
		if (PTR_ERR(mnt) == -EISDIR && (flags & LOOKUP_CONTINUE))
			return -EREMOTE;
		return PTR_ERR(mnt);
N
Nick Piggin 已提交
813
	}
814

815 816
	if (!mnt) /* mount collision */
		return 0;
N
Nick Piggin 已提交
817

818 819 820 821 822
	if (!*need_mntput) {
		/* lock_mount() may release path->mnt on error */
		mntget(path->mnt);
		*need_mntput = true;
	}
823
	err = finish_automount(mnt, path);
824

825 826 827
	switch (err) {
	case -EBUSY:
		/* Someone else made a mount here whilst we were busy */
828
		return 0;
829
	case 0:
830
		path_put(path);
831 832 833
		path->mnt = mnt;
		path->dentry = dget(mnt->mnt_root);
		return 0;
834 835
	default:
		return err;
836
	}
837

A
Al Viro 已提交
838 839
}

840 841
/*
 * Handle a dentry that is managed in some way.
842
 * - Flagged for transit management (autofs)
843 844 845 846 847 848 849 850
 * - Flagged as mountpoint
 * - Flagged as automount point
 *
 * This may only be called in refwalk mode.
 *
 * Serialization is taken care of in namespace.c
 */
static int follow_managed(struct path *path, unsigned flags)
L
Linus Torvalds 已提交
851
{
852
	struct vfsmount *mnt = path->mnt; /* held by caller, must be left alone */
853 854
	unsigned managed;
	bool need_mntput = false;
855
	int ret = 0;
856 857 858 859 860 861 862

	/* Given that we're not holding a lock here, we retain the value in a
	 * local variable for each dentry as we look at it so that we don't see
	 * the components of that value change under us */
	while (managed = ACCESS_ONCE(path->dentry->d_flags),
	       managed &= DCACHE_MANAGED_DENTRY,
	       unlikely(managed != 0)) {
863 864 865 866 867
		/* Allow the filesystem to manage the transit without i_mutex
		 * being held. */
		if (managed & DCACHE_MANAGE_TRANSIT) {
			BUG_ON(!path->dentry->d_op);
			BUG_ON(!path->dentry->d_op->d_manage);
868
			ret = path->dentry->d_op->d_manage(path->dentry, false);
869
			if (ret < 0)
870
				break;
871 872
		}

873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895
		/* Transit to a mounted filesystem. */
		if (managed & DCACHE_MOUNTED) {
			struct vfsmount *mounted = lookup_mnt(path);
			if (mounted) {
				dput(path->dentry);
				if (need_mntput)
					mntput(path->mnt);
				path->mnt = mounted;
				path->dentry = dget(mounted->mnt_root);
				need_mntput = true;
				continue;
			}

			/* Something is mounted on this dentry in another
			 * namespace and/or whatever was mounted there in this
			 * namespace got unmounted before we managed to get the
			 * vfsmount_lock */
		}

		/* Handle an automount point */
		if (managed & DCACHE_NEED_AUTOMOUNT) {
			ret = follow_automount(path, flags, &need_mntput);
			if (ret < 0)
896
				break;
897 898 899 900 901
			continue;
		}

		/* We didn't change the current path point */
		break;
L
Linus Torvalds 已提交
902
	}
903 904 905 906 907 908

	if (need_mntput && path->mnt == mnt)
		mntput(path->mnt);
	if (ret == -EISDIR)
		ret = 0;
	return ret;
L
Linus Torvalds 已提交
909 910
}

911
int follow_down_one(struct path *path)
L
Linus Torvalds 已提交
912 913 914
{
	struct vfsmount *mounted;

A
Al Viro 已提交
915
	mounted = lookup_mnt(path);
L
Linus Torvalds 已提交
916
	if (mounted) {
A
Al Viro 已提交
917 918 919 920
		dput(path->dentry);
		mntput(path->mnt);
		path->mnt = mounted;
		path->dentry = dget(mounted->mnt_root);
L
Linus Torvalds 已提交
921 922 923 924 925
		return 1;
	}
	return 0;
}

926 927 928 929 930 931
static inline bool managed_dentry_might_block(struct dentry *dentry)
{
	return (dentry->d_flags & DCACHE_MANAGE_TRANSIT &&
		dentry->d_op->d_manage(dentry, true) < 0);
}

932
/*
933 934
 * Try to skip to top of mountpoint pile in rcuwalk mode.  Fail if
 * we meet a managed dentry that would need blocking.
935 936
 */
static bool __follow_mount_rcu(struct nameidata *nd, struct path *path,
937
			       struct inode **inode)
938
{
939
	for (;;) {
940
		struct vfsmount *mounted;
941 942 943 944
		/*
		 * Don't forget we might have a non-mountpoint managed dentry
		 * that wants to block transit.
		 */
945
		if (unlikely(managed_dentry_might_block(path->dentry)))
946
			return false;
947 948 949 950

		if (!d_mountpoint(path->dentry))
			break;

951 952 953 954 955 956
		mounted = __lookup_mnt(path->mnt, path->dentry, 1);
		if (!mounted)
			break;
		path->mnt = mounted;
		path->dentry = mounted->mnt_root;
		nd->seq = read_seqcount_begin(&path->dentry->d_seq);
957 958 959 960 961 962
		/*
		 * Update the inode too. We don't need to re-check the
		 * dentry sequence number here after this d_inode read,
		 * because a mount-point is always pinned.
		 */
		*inode = path->dentry->d_inode;
963 964 965 966
	}
	return true;
}

967
static void follow_mount_rcu(struct nameidata *nd)
968
{
969
	while (d_mountpoint(nd->path.dentry)) {
970
		struct vfsmount *mounted;
971
		mounted = __lookup_mnt(nd->path.mnt, nd->path.dentry, 1);
972 973
		if (!mounted)
			break;
974 975 976
		nd->path.mnt = mounted;
		nd->path.dentry = mounted->mnt_root;
		nd->seq = read_seqcount_begin(&nd->path.dentry->d_seq);
977 978 979
	}
}

N
Nick Piggin 已提交
980 981 982 983
static int follow_dotdot_rcu(struct nameidata *nd)
{
	set_root_rcu(nd);

984
	while (1) {
N
Nick Piggin 已提交
985 986 987 988 989 990 991 992 993 994 995
		if (nd->path.dentry == nd->root.dentry &&
		    nd->path.mnt == nd->root.mnt) {
			break;
		}
		if (nd->path.dentry != nd->path.mnt->mnt_root) {
			struct dentry *old = nd->path.dentry;
			struct dentry *parent = old->d_parent;
			unsigned seq;

			seq = read_seqcount_begin(&parent->d_seq);
			if (read_seqcount_retry(&old->d_seq, nd->seq))
996
				goto failed;
N
Nick Piggin 已提交
997 998 999 1000 1001 1002 1003 1004
			nd->path.dentry = parent;
			nd->seq = seq;
			break;
		}
		if (!follow_up_rcu(&nd->path))
			break;
		nd->seq = read_seqcount_begin(&nd->path.dentry->d_seq);
	}
1005 1006
	follow_mount_rcu(nd);
	nd->inode = nd->path.dentry->d_inode;
N
Nick Piggin 已提交
1007
	return 0;
1008 1009 1010

failed:
	nd->flags &= ~LOOKUP_RCU;
1011 1012
	if (!(nd->flags & LOOKUP_ROOT))
		nd->root.mnt = NULL;
1013 1014 1015
	rcu_read_unlock();
	br_read_unlock(vfsmount_lock);
	return -ECHILD;
N
Nick Piggin 已提交
1016 1017
}

1018 1019 1020 1021 1022
/*
 * Follow down to the covering mount currently visible to userspace.  At each
 * point, the filesystem owning that dentry may be queried as to whether the
 * caller is permitted to proceed or not.
 */
1023
int follow_down(struct path *path)
1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042
{
	unsigned managed;
	int ret;

	while (managed = ACCESS_ONCE(path->dentry->d_flags),
	       unlikely(managed & DCACHE_MANAGED_DENTRY)) {
		/* Allow the filesystem to manage the transit without i_mutex
		 * being held.
		 *
		 * We indicate to the filesystem if someone is trying to mount
		 * something here.  This gives autofs the chance to deny anyone
		 * other than its daemon the right to mount on its
		 * superstructure.
		 *
		 * The filesystem may sleep at this point.
		 */
		if (managed & DCACHE_MANAGE_TRANSIT) {
			BUG_ON(!path->dentry->d_op);
			BUG_ON(!path->dentry->d_op->d_manage);
1043
			ret = path->dentry->d_op->d_manage(
1044
				path->dentry, false);
1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066
			if (ret < 0)
				return ret == -EISDIR ? 0 : ret;
		}

		/* Transit to a mounted filesystem. */
		if (managed & DCACHE_MOUNTED) {
			struct vfsmount *mounted = lookup_mnt(path);
			if (!mounted)
				break;
			dput(path->dentry);
			mntput(path->mnt);
			path->mnt = mounted;
			path->dentry = dget(mounted->mnt_root);
			continue;
		}

		/* Don't handle automount points here */
		break;
	}
	return 0;
}

1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082
/*
 * Skip to top of mountpoint pile in refwalk mode for follow_dotdot()
 */
static void follow_mount(struct path *path)
{
	while (d_mountpoint(path->dentry)) {
		struct vfsmount *mounted = lookup_mnt(path);
		if (!mounted)
			break;
		dput(path->dentry);
		mntput(path->mnt);
		path->mnt = mounted;
		path->dentry = dget(mounted->mnt_root);
	}
}

N
Nick Piggin 已提交
1083
static void follow_dotdot(struct nameidata *nd)
L
Linus Torvalds 已提交
1084
{
A
Al Viro 已提交
1085
	set_root(nd);
1086

L
Linus Torvalds 已提交
1087
	while(1) {
1088
		struct dentry *old = nd->path.dentry;
L
Linus Torvalds 已提交
1089

A
Al Viro 已提交
1090 1091
		if (nd->path.dentry == nd->root.dentry &&
		    nd->path.mnt == nd->root.mnt) {
L
Linus Torvalds 已提交
1092 1093
			break;
		}
1094
		if (nd->path.dentry != nd->path.mnt->mnt_root) {
A
Al Viro 已提交
1095 1096
			/* rare case of legitimate dget_parent()... */
			nd->path.dentry = dget_parent(nd->path.dentry);
L
Linus Torvalds 已提交
1097 1098 1099
			dput(old);
			break;
		}
A
Al Viro 已提交
1100
		if (!follow_up(&nd->path))
L
Linus Torvalds 已提交
1101 1102
			break;
	}
A
Al Viro 已提交
1103
	follow_mount(&nd->path);
N
Nick Piggin 已提交
1104
	nd->inode = nd->path.dentry->d_inode;
L
Linus Torvalds 已提交
1105 1106
}

1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135
/*
 * Allocate a dentry with name and parent, and perform a parent
 * directory ->lookup on it. Returns the new dentry, or ERR_PTR
 * on error. parent->d_inode->i_mutex must be held. d_lookup must
 * have verified that no child exists while under i_mutex.
 */
static struct dentry *d_alloc_and_lookup(struct dentry *parent,
				struct qstr *name, struct nameidata *nd)
{
	struct inode *inode = parent->d_inode;
	struct dentry *dentry;
	struct dentry *old;

	/* Don't create child dentry for a dead directory. */
	if (unlikely(IS_DEADDIR(inode)))
		return ERR_PTR(-ENOENT);

	dentry = d_alloc(parent, name);
	if (unlikely(!dentry))
		return ERR_PTR(-ENOMEM);

	old = inode->i_op->lookup(inode, dentry, nd);
	if (unlikely(old)) {
		dput(dentry);
		dentry = old;
	}
	return dentry;
}

1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159
/*
 * We already have a dentry, but require a lookup to be performed on the parent
 * directory to fill in d_inode. Returns the new dentry, or ERR_PTR on error.
 * parent->d_inode->i_mutex must be held. d_lookup must have verified that no
 * child exists while under i_mutex.
 */
static struct dentry *d_inode_lookup(struct dentry *parent, struct dentry *dentry,
				     struct nameidata *nd)
{
	struct inode *inode = parent->d_inode;
	struct dentry *old;

	/* Don't create child dentry for a dead directory. */
	if (unlikely(IS_DEADDIR(inode)))
		return ERR_PTR(-ENOENT);

	old = inode->i_op->lookup(inode, dentry, nd);
	if (unlikely(old)) {
		dput(dentry);
		dentry = old;
	}
	return dentry;
}

L
Linus Torvalds 已提交
1160 1161 1162 1163 1164 1165
/*
 *  It's more convoluted than I'd like it to be, but... it's still fairly
 *  small and for now I'd prefer to have fast path as straight as possible.
 *  It _is_ time-critical.
 */
static int do_lookup(struct nameidata *nd, struct qstr *name,
N
Nick Piggin 已提交
1166
			struct path *path, struct inode **inode)
L
Linus Torvalds 已提交
1167
{
1168
	struct vfsmount *mnt = nd->path.mnt;
N
Nick Piggin 已提交
1169
	struct dentry *dentry, *parent = nd->path.dentry;
A
Al Viro 已提交
1170 1171
	int need_reval = 1;
	int status = 1;
1172 1173
	int err;

1174 1175 1176 1177 1178
	/*
	 * Rename seqlock is not required here because in the off chance
	 * of a false negative due to a concurrent rename, we're going to
	 * do the non-racy lookup, below.
	 */
N
Nick Piggin 已提交
1179 1180 1181 1182
	if (nd->flags & LOOKUP_RCU) {
		unsigned seq;
		*inode = nd->inode;
		dentry = __d_lookup_rcu(parent, name, &seq, inode);
A
Al Viro 已提交
1183 1184 1185
		if (!dentry)
			goto unlazy;

N
Nick Piggin 已提交
1186 1187 1188 1189
		/* Memory barrier in read_seqcount_begin of child is enough */
		if (__read_seqcount_retry(&parent->d_seq, nd->seq))
			return -ECHILD;
		nd->seq = seq;
A
Al Viro 已提交
1190

1191
		if (unlikely(dentry->d_flags & DCACHE_OP_REVALIDATE)) {
A
Al Viro 已提交
1192 1193 1194 1195 1196 1197
			status = d_revalidate(dentry, nd);
			if (unlikely(status <= 0)) {
				if (status != -ECHILD)
					need_reval = 0;
				goto unlazy;
			}
1198
		}
1199 1200
		if (unlikely(d_need_lookup(dentry)))
			goto unlazy;
N
Nick Piggin 已提交
1201 1202
		path->mnt = mnt;
		path->dentry = dentry;
1203 1204 1205 1206 1207
		if (unlikely(!__follow_mount_rcu(nd, path, inode)))
			goto unlazy;
		if (unlikely(path->dentry->d_flags & DCACHE_NEED_AUTOMOUNT))
			goto unlazy;
		return 0;
A
Al Viro 已提交
1208
unlazy:
A
Al Viro 已提交
1209 1210
		if (unlazy_walk(nd, dentry))
			return -ECHILD;
A
Al Viro 已提交
1211 1212
	} else {
		dentry = __d_lookup(parent, name);
1213
	}
A
Al Viro 已提交
1214

1215 1216 1217 1218
	if (dentry && unlikely(d_need_lookup(dentry))) {
		dput(dentry);
		dentry = NULL;
	}
A
Al Viro 已提交
1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234
retry:
	if (unlikely(!dentry)) {
		struct inode *dir = parent->d_inode;
		BUG_ON(nd->inode != dir);

		mutex_lock(&dir->i_mutex);
		dentry = d_lookup(parent, name);
		if (likely(!dentry)) {
			dentry = d_alloc_and_lookup(parent, name, nd);
			if (IS_ERR(dentry)) {
				mutex_unlock(&dir->i_mutex);
				return PTR_ERR(dentry);
			}
			/* known good */
			need_reval = 0;
			status = 1;
1235 1236 1237 1238 1239 1240 1241 1242 1243
		} else if (unlikely(d_need_lookup(dentry))) {
			dentry = d_inode_lookup(parent, dentry, nd);
			if (IS_ERR(dentry)) {
				mutex_unlock(&dir->i_mutex);
				return PTR_ERR(dentry);
			}
			/* known good */
			need_reval = 0;
			status = 1;
A
Al Viro 已提交
1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259
		}
		mutex_unlock(&dir->i_mutex);
	}
	if (unlikely(dentry->d_flags & DCACHE_OP_REVALIDATE) && need_reval)
		status = d_revalidate(dentry, nd);
	if (unlikely(status <= 0)) {
		if (status < 0) {
			dput(dentry);
			return status;
		}
		if (!d_invalidate(dentry)) {
			dput(dentry);
			dentry = NULL;
			need_reval = 1;
			goto retry;
		}
1260
	}
A
Al Viro 已提交
1261

1262 1263 1264
	path->mnt = mnt;
	path->dentry = dentry;
	err = follow_managed(path, nd->flags);
1265 1266
	if (unlikely(err < 0)) {
		path_put_conditional(path, nd);
1267
		return err;
1268
	}
1269
	*inode = path->dentry->d_inode;
L
Linus Torvalds 已提交
1270 1271 1272
	return 0;
}

1273 1274 1275 1276 1277 1278
static inline int may_lookup(struct nameidata *nd)
{
	if (nd->flags & LOOKUP_RCU) {
		int err = exec_permission(nd->inode, IPERM_FLAG_RCU);
		if (err != -ECHILD)
			return err;
A
Al Viro 已提交
1279
		if (unlazy_walk(nd, NULL))
1280 1281 1282 1283 1284
			return -ECHILD;
	}
	return exec_permission(nd->inode, 0);
}

1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296
static inline int handle_dots(struct nameidata *nd, int type)
{
	if (type == LAST_DOTDOT) {
		if (nd->flags & LOOKUP_RCU) {
			if (follow_dotdot_rcu(nd))
				return -ECHILD;
		} else
			follow_dotdot(nd);
	}
	return 0;
}

1297 1298 1299 1300 1301 1302
static void terminate_walk(struct nameidata *nd)
{
	if (!(nd->flags & LOOKUP_RCU)) {
		path_put(&nd->path);
	} else {
		nd->flags &= ~LOOKUP_RCU;
1303 1304
		if (!(nd->flags & LOOKUP_ROOT))
			nd->root.mnt = NULL;
1305 1306 1307 1308 1309
		rcu_read_unlock();
		br_read_unlock(vfsmount_lock);
	}
}

1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332
static inline int walk_component(struct nameidata *nd, struct path *path,
		struct qstr *name, int type, int follow)
{
	struct inode *inode;
	int err;
	/*
	 * "." and ".." are special - ".." especially so because it has
	 * to be able to know about the current root directory and
	 * parent relationships.
	 */
	if (unlikely(type != LAST_NORM))
		return handle_dots(nd, type);
	err = do_lookup(nd, name, path, &inode);
	if (unlikely(err)) {
		terminate_walk(nd);
		return err;
	}
	if (!inode) {
		path_to_nameidata(path, nd);
		terminate_walk(nd);
		return -ENOENT;
	}
	if (unlikely(inode->i_op->follow_link) && follow) {
A
Al Viro 已提交
1333 1334 1335 1336 1337 1338
		if (nd->flags & LOOKUP_RCU) {
			if (unlikely(unlazy_walk(nd, path->dentry))) {
				terminate_walk(nd);
				return -ECHILD;
			}
		}
1339 1340 1341 1342 1343 1344 1345 1346
		BUG_ON(inode != path->dentry->d_inode);
		return 1;
	}
	path_to_nameidata(path, nd);
	nd->inode = inode;
	return 0;
}

1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362
/*
 * This limits recursive symlink follows to 8, while
 * limiting consecutive symlinks to 40.
 *
 * Without that kind of total limit, nasty chains of consecutive
 * symlinks can cause almost arbitrarily long lookups.
 */
static inline int nested_symlink(struct path *path, struct nameidata *nd)
{
	int res;

	if (unlikely(current->link_count >= MAX_NESTED_LINKS)) {
		path_put_conditional(path, nd);
		path_put(&nd->path);
		return -ELOOP;
	}
1363
	BUG_ON(nd->depth >= MAX_NESTED_LINKS);
1364 1365 1366 1367 1368 1369 1370

	nd->depth++;
	current->link_count++;

	do {
		struct path link = *path;
		void *cookie;
1371 1372

		res = follow_link(&link, nd, &cookie);
1373 1374 1375
		if (!res)
			res = walk_component(nd, path, &nd->last,
					     nd->last_type, LOOKUP_FOLLOW);
1376
		put_link(nd, &link, cookie);
1377 1378 1379 1380 1381 1382 1383
	} while (res > 0);

	current->link_count--;
	nd->depth--;
	return res;
}

L
Linus Torvalds 已提交
1384 1385
/*
 * Name resolution.
1386 1387
 * This is the basic name resolution function, turning a pathname into
 * the final dentry. We expect 'base' to be positive and a directory.
L
Linus Torvalds 已提交
1388
 *
1389 1390
 * Returns 0 and nd will have valid dentry and mnt on success.
 * Returns error and drops reference to input namei data on failure.
L
Linus Torvalds 已提交
1391
 */
1392
static int link_path_walk(const char *name, struct nameidata *nd)
L
Linus Torvalds 已提交
1393 1394 1395 1396 1397 1398 1399 1400
{
	struct path next;
	int err;
	unsigned int lookup_flags = nd->flags;
	
	while (*name=='/')
		name++;
	if (!*name)
1401
		return 0;
L
Linus Torvalds 已提交
1402 1403 1404 1405 1406 1407

	/* At this point we know we have a real path component. */
	for(;;) {
		unsigned long hash;
		struct qstr this;
		unsigned int c;
A
Al Viro 已提交
1408
		int type;
L
Linus Torvalds 已提交
1409

1410
		nd->flags |= LOOKUP_CONTINUE;
1411 1412

		err = may_lookup(nd);
L
Linus Torvalds 已提交
1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427
 		if (err)
			break;

		this.name = name;
		c = *(const unsigned char *)name;

		hash = init_name_hash();
		do {
			name++;
			hash = partial_name_hash(c, hash);
			c = *(const unsigned char *)name;
		} while (c && (c != '/'));
		this.len = name - (const char *) this.name;
		this.hash = end_name_hash(hash);

A
Al Viro 已提交
1428 1429 1430
		type = LAST_NORM;
		if (this.name[0] == '.') switch (this.len) {
			case 2:
A
Al Viro 已提交
1431
				if (this.name[1] == '.') {
A
Al Viro 已提交
1432
					type = LAST_DOTDOT;
A
Al Viro 已提交
1433 1434
					nd->flags |= LOOKUP_JUMPED;
				}
A
Al Viro 已提交
1435 1436 1437 1438
				break;
			case 1:
				type = LAST_DOT;
		}
1439 1440
		if (likely(type == LAST_NORM)) {
			struct dentry *parent = nd->path.dentry;
A
Al Viro 已提交
1441
			nd->flags &= ~LOOKUP_JUMPED;
1442 1443 1444 1445 1446 1447 1448
			if (unlikely(parent->d_flags & DCACHE_OP_HASH)) {
				err = parent->d_op->d_hash(parent, nd->inode,
							   &this);
				if (err < 0)
					break;
			}
		}
A
Al Viro 已提交
1449

L
Linus Torvalds 已提交
1450 1451 1452 1453 1454
		/* remove trailing slashes? */
		if (!c)
			goto last_component;
		while (*++name == '/');
		if (!*name)
1455
			goto last_component;
L
Linus Torvalds 已提交
1456

1457 1458 1459
		err = walk_component(nd, &next, &this, type, LOOKUP_FOLLOW);
		if (err < 0)
			return err;
L
Linus Torvalds 已提交
1460

1461
		if (err) {
1462
			err = nested_symlink(&next, nd);
L
Linus Torvalds 已提交
1463
			if (err)
1464
				return err;
N
Nick Piggin 已提交
1465
		}
L
Linus Torvalds 已提交
1466
		err = -ENOTDIR; 
N
Nick Piggin 已提交
1467
		if (!nd->inode->i_op->lookup)
L
Linus Torvalds 已提交
1468 1469 1470 1471 1472
			break;
		continue;
		/* here ends the main loop */

last_component:
1473 1474
		/* Clear LOOKUP_CONTINUE iff it was previously unset */
		nd->flags &= lookup_flags | ~LOOKUP_CONTINUE;
1475 1476
		nd->last = this;
		nd->last_type = type;
1477
		return 0;
L
Linus Torvalds 已提交
1478
	}
1479
	terminate_walk(nd);
L
Linus Torvalds 已提交
1480 1481 1482
	return err;
}

A
Al Viro 已提交
1483 1484
static int path_init(int dfd, const char *name, unsigned int flags,
		     struct nameidata *nd, struct file **fp)
N
Nick Piggin 已提交
1485 1486 1487 1488 1489 1490
{
	int retval = 0;
	int fput_needed;
	struct file *file;

	nd->last_type = LAST_ROOT; /* if there are only slashes... */
A
Al Viro 已提交
1491
	nd->flags = flags | LOOKUP_JUMPED;
N
Nick Piggin 已提交
1492
	nd->depth = 0;
1493 1494
	if (flags & LOOKUP_ROOT) {
		struct inode *inode = nd->root.dentry->d_inode;
A
Al Viro 已提交
1495 1496 1497 1498 1499 1500 1501
		if (*name) {
			if (!inode->i_op->lookup)
				return -ENOTDIR;
			retval = inode_permission(inode, MAY_EXEC);
			if (retval)
				return retval;
		}
1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513
		nd->path = nd->root;
		nd->inode = inode;
		if (flags & LOOKUP_RCU) {
			br_read_lock(vfsmount_lock);
			rcu_read_lock();
			nd->seq = __read_seqcount_begin(&nd->path.dentry->d_seq);
		} else {
			path_get(&nd->path);
		}
		return 0;
	}

N
Nick Piggin 已提交
1514 1515 1516
	nd->root.mnt = NULL;

	if (*name=='/') {
A
Al Viro 已提交
1517 1518 1519 1520 1521 1522 1523 1524 1525
		if (flags & LOOKUP_RCU) {
			br_read_lock(vfsmount_lock);
			rcu_read_lock();
			set_root_rcu(nd);
		} else {
			set_root(nd);
			path_get(&nd->root);
		}
		nd->path = nd->root;
N
Nick Piggin 已提交
1526
	} else if (dfd == AT_FDCWD) {
A
Al Viro 已提交
1527 1528 1529
		if (flags & LOOKUP_RCU) {
			struct fs_struct *fs = current->fs;
			unsigned seq;
N
Nick Piggin 已提交
1530

A
Al Viro 已提交
1531 1532
			br_read_lock(vfsmount_lock);
			rcu_read_lock();
N
Nick Piggin 已提交
1533

A
Al Viro 已提交
1534 1535 1536 1537 1538 1539 1540 1541
			do {
				seq = read_seqcount_begin(&fs->seq);
				nd->path = fs->pwd;
				nd->seq = __read_seqcount_begin(&nd->path.dentry->d_seq);
			} while (read_seqcount_retry(&fs->seq, seq));
		} else {
			get_fs_pwd(current->fs, &nd->path);
		}
N
Nick Piggin 已提交
1542 1543 1544
	} else {
		struct dentry *dentry;

1545
		file = fget_raw_light(dfd, &fput_needed);
N
Nick Piggin 已提交
1546 1547 1548 1549 1550 1551
		retval = -EBADF;
		if (!file)
			goto out_fail;

		dentry = file->f_path.dentry;

A
Al Viro 已提交
1552 1553 1554 1555
		if (*name) {
			retval = -ENOTDIR;
			if (!S_ISDIR(dentry->d_inode->i_mode))
				goto fput_fail;
N
Nick Piggin 已提交
1556

A
Al Viro 已提交
1557 1558 1559 1560
			retval = file_permission(file, MAY_EXEC);
			if (retval)
				goto fput_fail;
		}
N
Nick Piggin 已提交
1561 1562

		nd->path = file->f_path;
A
Al Viro 已提交
1563 1564
		if (flags & LOOKUP_RCU) {
			if (fput_needed)
A
Al Viro 已提交
1565
				*fp = file;
A
Al Viro 已提交
1566 1567 1568 1569 1570 1571 1572
			nd->seq = __read_seqcount_begin(&nd->path.dentry->d_seq);
			br_read_lock(vfsmount_lock);
			rcu_read_lock();
		} else {
			path_get(&file->f_path);
			fput_light(file, fput_needed);
		}
N
Nick Piggin 已提交
1573 1574 1575
	}

	nd->inode = nd->path.dentry->d_inode;
1576
	return 0;
1577

1578 1579 1580 1581 1582 1583
fput_fail:
	fput_light(file, fput_needed);
out_fail:
	return retval;
}

1584 1585 1586 1587 1588 1589 1590 1591 1592 1593
static inline int lookup_last(struct nameidata *nd, struct path *path)
{
	if (nd->last_type == LAST_NORM && nd->last.name[nd->last.len])
		nd->flags |= LOOKUP_FOLLOW | LOOKUP_DIRECTORY;

	nd->flags &= ~LOOKUP_PARENT;
	return walk_component(nd, path, &nd->last, nd->last_type,
					nd->flags & LOOKUP_FOLLOW);
}

1594
/* Returns 0 and nd will be valid on success; Retuns error, otherwise. */
A
Al Viro 已提交
1595
static int path_lookupat(int dfd, const char *name,
1596 1597
				unsigned int flags, struct nameidata *nd)
{
A
Al Viro 已提交
1598
	struct file *base = NULL;
1599 1600
	struct path path;
	int err;
N
Nick Piggin 已提交
1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615

	/*
	 * Path walking is largely split up into 2 different synchronisation
	 * schemes, rcu-walk and ref-walk (explained in
	 * Documentation/filesystems/path-lookup.txt). These share much of the
	 * path walk code, but some things particularly setup, cleanup, and
	 * following mounts are sufficiently divergent that functions are
	 * duplicated. Typically there is a function foo(), and its RCU
	 * analogue, foo_rcu().
	 *
	 * -ECHILD is the error number of choice (just to avoid clashes) that
	 * is returned if some aspect of an rcu-walk fails. Such an error must
	 * be handled by restarting a traditional ref-walk (which will always
	 * be able to complete).
	 */
1616
	err = path_init(dfd, name, flags | LOOKUP_PARENT, nd, &base);
A
Al Viro 已提交
1617

1618 1619
	if (unlikely(err))
		return err;
A
Al Viro 已提交
1620 1621

	current->total_link_count = 0;
1622 1623 1624 1625 1626 1627 1628 1629
	err = link_path_walk(name, nd);

	if (!err && !(flags & LOOKUP_PARENT)) {
		err = lookup_last(nd, &path);
		while (err > 0) {
			void *cookie;
			struct path link = path;
			nd->flags |= LOOKUP_PARENT;
1630
			err = follow_link(&link, nd, &cookie);
1631 1632
			if (!err)
				err = lookup_last(nd, &path);
1633
			put_link(nd, &link, cookie);
1634 1635
		}
	}
A
Al Viro 已提交
1636

1637 1638
	if (!err)
		err = complete_walk(nd);
1639 1640 1641 1642

	if (!err && nd->flags & LOOKUP_DIRECTORY) {
		if (!nd->inode->i_op->lookup) {
			path_put(&nd->path);
A
Al Viro 已提交
1643
			err = -ENOTDIR;
1644 1645
		}
	}
A
Al Viro 已提交
1646

A
Al Viro 已提交
1647 1648
	if (base)
		fput(base);
A
Al Viro 已提交
1649

1650
	if (nd->root.mnt && !(nd->flags & LOOKUP_ROOT)) {
A
Al Viro 已提交
1651 1652 1653
		path_put(&nd->root);
		nd->root.mnt = NULL;
	}
1654
	return err;
A
Al Viro 已提交
1655
}
N
Nick Piggin 已提交
1656

A
Al Viro 已提交
1657 1658 1659 1660 1661 1662 1663 1664
static int do_path_lookup(int dfd, const char *name,
				unsigned int flags, struct nameidata *nd)
{
	int retval = path_lookupat(dfd, name, flags | LOOKUP_RCU, nd);
	if (unlikely(retval == -ECHILD))
		retval = path_lookupat(dfd, name, flags, nd);
	if (unlikely(retval == -ESTALE))
		retval = path_lookupat(dfd, name, flags | LOOKUP_REVAL, nd);
N
Nick Piggin 已提交
1665 1666 1667 1668 1669 1670 1671

	if (likely(!retval)) {
		if (unlikely(!audit_dummy_context())) {
			if (nd->path.dentry && nd->inode)
				audit_inode(name, nd->path.dentry);
		}
	}
1672
	return retval;
L
Linus Torvalds 已提交
1673 1674
}

A
Al Viro 已提交
1675
int kern_path_parent(const char *name, struct nameidata *nd)
1676
{
A
Al Viro 已提交
1677
	return do_path_lookup(AT_FDCWD, name, LOOKUP_PARENT, nd);
1678 1679
}

A
Al Viro 已提交
1680 1681 1682 1683 1684 1685 1686 1687 1688
int kern_path(const char *name, unsigned int flags, struct path *path)
{
	struct nameidata nd;
	int res = do_path_lookup(AT_FDCWD, name, flags, &nd);
	if (!res)
		*path = nd.path;
	return res;
}

1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700
/**
 * vfs_path_lookup - lookup a file path relative to a dentry-vfsmount pair
 * @dentry:  pointer to dentry of the base directory
 * @mnt: pointer to vfs mount of the base directory
 * @name: pointer to file name
 * @flags: lookup flags
 * @nd: pointer to nameidata
 */
int vfs_path_lookup(struct dentry *dentry, struct vfsmount *mnt,
		    const char *name, unsigned int flags,
		    struct nameidata *nd)
{
1701 1702 1703 1704
	nd->root.dentry = dentry;
	nd->root.mnt = mnt;
	/* the first argument of do_path_lookup() is ignored with LOOKUP_ROOT */
	return do_path_lookup(AT_FDCWD, name, flags | LOOKUP_ROOT, nd);
1705 1706
}

1707 1708
static struct dentry *__lookup_hash(struct qstr *name,
		struct dentry *base, struct nameidata *nd)
L
Linus Torvalds 已提交
1709
{
1710
	struct inode *inode = base->d_inode;
1711
	struct dentry *dentry;
L
Linus Torvalds 已提交
1712 1713
	int err;

1714
	err = exec_permission(inode, 0);
1715 1716
	if (err)
		return ERR_PTR(err);
L
Linus Torvalds 已提交
1717

1718 1719 1720 1721
	/*
	 * Don't bother with __d_lookup: callers are for creat as
	 * well as unlink, so a lot of the time it would cost
	 * a double lookup.
A
Al Viro 已提交
1722
	 */
1723
	dentry = d_lookup(base, name);
A
Al Viro 已提交
1724

1725 1726 1727 1728 1729 1730 1731 1732 1733 1734
	if (dentry && d_need_lookup(dentry)) {
		/*
		 * __lookup_hash is called with the parent dir's i_mutex already
		 * held, so we are good to go here.
		 */
		dentry = d_inode_lookup(base, dentry, nd);
		if (IS_ERR(dentry))
			return dentry;
	}

1735
	if (dentry && (dentry->d_flags & DCACHE_OP_REVALIDATE))
A
Al Viro 已提交
1736 1737
		dentry = do_revalidate(dentry, nd);

1738 1739
	if (!dentry)
		dentry = d_alloc_and_lookup(base, name, nd);
1740

L
Linus Torvalds 已提交
1741 1742 1743
	return dentry;
}

1744 1745 1746 1747 1748
/*
 * Restricted form of lookup. Doesn't follow links, single-component only,
 * needs parent already locked. Doesn't follow mounts.
 * SMP-safe.
 */
1749
static struct dentry *lookup_hash(struct nameidata *nd)
1750
{
1751
	return __lookup_hash(&nd->last, nd->path.dentry, nd);
L
Linus Torvalds 已提交
1752 1753
}

1754
/**
1755
 * lookup_one_len - filesystem helper to lookup single pathname component
1756 1757 1758 1759
 * @name:	pathname component to lookup
 * @base:	base directory to lookup from
 * @len:	maximum length @len should be interpreted to
 *
1760 1761
 * Note that this routine is purely a helper for filesystem usage and should
 * not be called by generic code.  Also note that by using this function the
1762 1763 1764
 * nameidata argument is passed to the filesystem methods and a filesystem
 * using this helper needs to be prepared for that.
 */
1765 1766 1767
struct dentry *lookup_one_len(const char *name, struct dentry *base, int len)
{
	struct qstr this;
A
Al Viro 已提交
1768 1769
	unsigned long hash;
	unsigned int c;
1770

1771 1772
	WARN_ON_ONCE(!mutex_is_locked(&base->d_inode->i_mutex));

A
Al Viro 已提交
1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785
	this.name = name;
	this.len = len;
	if (!len)
		return ERR_PTR(-EACCES);

	hash = init_name_hash();
	while (len--) {
		c = *(const unsigned char *)name++;
		if (c == '/' || c == '\0')
			return ERR_PTR(-EACCES);
		hash = partial_name_hash(c, hash);
	}
	this.hash = end_name_hash(hash);
1786 1787 1788 1789 1790 1791 1792 1793 1794
	/*
	 * See if the low-level filesystem might want
	 * to use its own hash..
	 */
	if (base->d_flags & DCACHE_OP_HASH) {
		int err = base->d_op->d_hash(base, base->d_inode, &this);
		if (err < 0)
			return ERR_PTR(err);
	}
1795

1796
	return __lookup_hash(&this, base, NULL);
1797 1798
}

1799 1800
int user_path_at(int dfd, const char __user *name, unsigned flags,
		 struct path *path)
L
Linus Torvalds 已提交
1801
{
1802
	struct nameidata nd;
A
Al Viro 已提交
1803
	char *tmp = getname_flags(name, flags);
L
Linus Torvalds 已提交
1804 1805
	int err = PTR_ERR(tmp);
	if (!IS_ERR(tmp)) {
1806 1807 1808 1809

		BUG_ON(flags & LOOKUP_PARENT);

		err = do_path_lookup(dfd, tmp, flags, &nd);
L
Linus Torvalds 已提交
1810
		putname(tmp);
1811 1812
		if (!err)
			*path = nd.path;
L
Linus Torvalds 已提交
1813 1814 1815 1816
	}
	return err;
}

1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834
static int user_path_parent(int dfd, const char __user *path,
			struct nameidata *nd, char **name)
{
	char *s = getname(path);
	int error;

	if (IS_ERR(s))
		return PTR_ERR(s);

	error = do_path_lookup(dfd, s, LOOKUP_PARENT, nd);
	if (error)
		putname(s);
	else
		*name = s;

	return error;
}

L
Linus Torvalds 已提交
1835 1836 1837 1838 1839 1840
/*
 * It's inline, so penalty for filesystems that don't use sticky bit is
 * minimal.
 */
static inline int check_sticky(struct inode *dir, struct inode *inode)
{
1841 1842
	uid_t fsuid = current_fsuid();

L
Linus Torvalds 已提交
1843 1844
	if (!(dir->i_mode & S_ISVTX))
		return 0;
1845 1846
	if (current_user_ns() != inode_userns(inode))
		goto other_userns;
1847
	if (inode->i_uid == fsuid)
L
Linus Torvalds 已提交
1848
		return 0;
1849
	if (dir->i_uid == fsuid)
L
Linus Torvalds 已提交
1850
		return 0;
1851 1852 1853

other_userns:
	return !ns_capable(inode_userns(inode), CAP_FOWNER);
L
Linus Torvalds 已提交
1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874
}

/*
 *	Check whether we can remove a link victim from directory dir, check
 *  whether the type of victim is right.
 *  1. We can't do it if dir is read-only (done in permission())
 *  2. We should have write and exec permissions on dir
 *  3. We can't remove anything from append-only dir
 *  4. We can't do anything with immutable dir (done in permission())
 *  5. If the sticky bit on dir is set we should either
 *	a. be owner of dir, or
 *	b. be owner of victim, or
 *	c. have CAP_FOWNER capability
 *  6. If the victim is append-only or immutable we can't do antyhing with
 *     links pointing to it.
 *  7. If we were asked to remove a directory and victim isn't one - ENOTDIR.
 *  8. If we were asked to remove a non-directory and victim isn't one - EISDIR.
 *  9. We can't remove a root or mountpoint.
 * 10. We don't allow removal of NFS sillyrenamed files; it's handled by
 *     nfs_async_unlink().
 */
1875
static int may_delete(struct inode *dir,struct dentry *victim,int isdir)
L
Linus Torvalds 已提交
1876 1877 1878 1879 1880 1881 1882
{
	int error;

	if (!victim->d_inode)
		return -ENOENT;

	BUG_ON(victim->d_parent->d_inode != dir);
1883
	audit_inode_child(victim, dir);
L
Linus Torvalds 已提交
1884

1885
	error = inode_permission(dir, MAY_WRITE | MAY_EXEC);
L
Linus Torvalds 已提交
1886 1887 1888 1889 1890
	if (error)
		return error;
	if (IS_APPEND(dir))
		return -EPERM;
	if (check_sticky(dir, victim->d_inode)||IS_APPEND(victim->d_inode)||
H
Hugh Dickins 已提交
1891
	    IS_IMMUTABLE(victim->d_inode) || IS_SWAPFILE(victim->d_inode))
L
Linus Torvalds 已提交
1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914
		return -EPERM;
	if (isdir) {
		if (!S_ISDIR(victim->d_inode->i_mode))
			return -ENOTDIR;
		if (IS_ROOT(victim))
			return -EBUSY;
	} else if (S_ISDIR(victim->d_inode->i_mode))
		return -EISDIR;
	if (IS_DEADDIR(dir))
		return -ENOENT;
	if (victim->d_flags & DCACHE_NFSFS_RENAMED)
		return -EBUSY;
	return 0;
}

/*	Check whether we can create an object with dentry child in directory
 *  dir.
 *  1. We can't do it if child already exists (open has special treatment for
 *     this case, but since we are inlined it's OK)
 *  2. We can't do it if dir is read-only (done in permission())
 *  3. We should have write and exec permissions on dir
 *  4. We can't do it if dir is immutable (done in permission())
 */
1915
static inline int may_create(struct inode *dir, struct dentry *child)
L
Linus Torvalds 已提交
1916 1917 1918 1919 1920
{
	if (child->d_inode)
		return -EEXIST;
	if (IS_DEADDIR(dir))
		return -ENOENT;
1921
	return inode_permission(dir, MAY_WRITE | MAY_EXEC);
L
Linus Torvalds 已提交
1922 1923 1924 1925 1926 1927 1928 1929 1930 1931
}

/*
 * p1 and p2 should be directories on the same fs.
 */
struct dentry *lock_rename(struct dentry *p1, struct dentry *p2)
{
	struct dentry *p;

	if (p1 == p2) {
I
Ingo Molnar 已提交
1932
		mutex_lock_nested(&p1->d_inode->i_mutex, I_MUTEX_PARENT);
L
Linus Torvalds 已提交
1933 1934 1935
		return NULL;
	}

1936
	mutex_lock(&p1->d_inode->i_sb->s_vfs_rename_mutex);
L
Linus Torvalds 已提交
1937

1938 1939 1940 1941 1942
	p = d_ancestor(p2, p1);
	if (p) {
		mutex_lock_nested(&p2->d_inode->i_mutex, I_MUTEX_PARENT);
		mutex_lock_nested(&p1->d_inode->i_mutex, I_MUTEX_CHILD);
		return p;
L
Linus Torvalds 已提交
1943 1944
	}

1945 1946 1947 1948 1949
	p = d_ancestor(p1, p2);
	if (p) {
		mutex_lock_nested(&p1->d_inode->i_mutex, I_MUTEX_PARENT);
		mutex_lock_nested(&p2->d_inode->i_mutex, I_MUTEX_CHILD);
		return p;
L
Linus Torvalds 已提交
1950 1951
	}

I
Ingo Molnar 已提交
1952 1953
	mutex_lock_nested(&p1->d_inode->i_mutex, I_MUTEX_PARENT);
	mutex_lock_nested(&p2->d_inode->i_mutex, I_MUTEX_CHILD);
L
Linus Torvalds 已提交
1954 1955 1956 1957 1958
	return NULL;
}

void unlock_rename(struct dentry *p1, struct dentry *p2)
{
1959
	mutex_unlock(&p1->d_inode->i_mutex);
L
Linus Torvalds 已提交
1960
	if (p1 != p2) {
1961
		mutex_unlock(&p2->d_inode->i_mutex);
1962
		mutex_unlock(&p1->d_inode->i_sb->s_vfs_rename_mutex);
L
Linus Torvalds 已提交
1963 1964 1965 1966 1967 1968
	}
}

int vfs_create(struct inode *dir, struct dentry *dentry, int mode,
		struct nameidata *nd)
{
1969
	int error = may_create(dir, dentry);
L
Linus Torvalds 已提交
1970 1971 1972 1973

	if (error)
		return error;

A
Al Viro 已提交
1974
	if (!dir->i_op->create)
L
Linus Torvalds 已提交
1975 1976 1977 1978 1979 1980 1981
		return -EACCES;	/* shouldn't it be ENOSYS? */
	mode &= S_IALLUGO;
	mode |= S_IFREG;
	error = security_inode_create(dir, dentry, mode);
	if (error)
		return error;
	error = dir->i_op->create(dir, dentry, mode, nd);
1982
	if (!error)
1983
		fsnotify_create(dir, dentry);
L
Linus Torvalds 已提交
1984 1985 1986
	return error;
}

A
Al Viro 已提交
1987
static int may_open(struct path *path, int acc_mode, int flag)
L
Linus Torvalds 已提交
1988
{
1989
	struct dentry *dentry = path->dentry;
L
Linus Torvalds 已提交
1990 1991 1992
	struct inode *inode = dentry->d_inode;
	int error;

A
Al Viro 已提交
1993 1994 1995 1996
	/* O_PATH? */
	if (!acc_mode)
		return 0;

L
Linus Torvalds 已提交
1997 1998 1999
	if (!inode)
		return -ENOENT;

C
Christoph Hellwig 已提交
2000 2001
	switch (inode->i_mode & S_IFMT) {
	case S_IFLNK:
L
Linus Torvalds 已提交
2002
		return -ELOOP;
C
Christoph Hellwig 已提交
2003 2004 2005 2006 2007 2008
	case S_IFDIR:
		if (acc_mode & MAY_WRITE)
			return -EISDIR;
		break;
	case S_IFBLK:
	case S_IFCHR:
2009
		if (path->mnt->mnt_flags & MNT_NODEV)
L
Linus Torvalds 已提交
2010
			return -EACCES;
C
Christoph Hellwig 已提交
2011 2012 2013
		/*FALLTHRU*/
	case S_IFIFO:
	case S_IFSOCK:
L
Linus Torvalds 已提交
2014
		flag &= ~O_TRUNC;
C
Christoph Hellwig 已提交
2015
		break;
2016
	}
2017

2018
	error = inode_permission(inode, acc_mode);
2019 2020
	if (error)
		return error;
M
Mimi Zohar 已提交
2021

L
Linus Torvalds 已提交
2022 2023 2024 2025
	/*
	 * An append-only file must be opened in append mode for writing.
	 */
	if (IS_APPEND(inode)) {
2026
		if  ((flag & O_ACCMODE) != O_RDONLY && !(flag & O_APPEND))
2027
			return -EPERM;
L
Linus Torvalds 已提交
2028
		if (flag & O_TRUNC)
2029
			return -EPERM;
L
Linus Torvalds 已提交
2030 2031 2032
	}

	/* O_NOATIME can only be set by the owner or superuser */
2033
	if (flag & O_NOATIME && !inode_owner_or_capable(inode))
2034
		return -EPERM;
L
Linus Torvalds 已提交
2035 2036 2037 2038

	/*
	 * Ensure there are no outstanding leases on the file.
	 */
2039
	return break_lease(inode, flag);
2040
}
L
Linus Torvalds 已提交
2041

2042
static int handle_truncate(struct file *filp)
2043
{
2044
	struct path *path = &filp->f_path;
2045 2046 2047 2048 2049 2050 2051 2052 2053
	struct inode *inode = path->dentry->d_inode;
	int error = get_write_access(inode);
	if (error)
		return error;
	/*
	 * Refuse to truncate files with mandatory locks held on them.
	 */
	error = locks_verify_locked(inode);
	if (!error)
2054
		error = security_path_truncate(path);
2055 2056 2057
	if (!error) {
		error = do_truncate(path->dentry, 0,
				    ATTR_MTIME|ATTR_CTIME|ATTR_OPEN,
2058
				    filp);
2059 2060
	}
	put_write_access(inode);
M
Mimi Zohar 已提交
2061
	return error;
L
Linus Torvalds 已提交
2062 2063
}

2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087
/*
 * Note that while the flag value (low two bits) for sys_open means:
 *	00 - read-only
 *	01 - write-only
 *	10 - read-write
 *	11 - special
 * it is changed into
 *	00 - no permissions needed
 *	01 - read-permission
 *	10 - write-permission
 *	11 - read-write
 * for the internal routines (ie open_namei()/follow_link() etc)
 * This is more logical, and also allows the 00 "no perm needed"
 * to be used for symlinks (where the permissions are checked
 * later).
 *
*/
static inline int open_to_namei_flags(int flag)
{
	if ((flag+1) & O_ACCMODE)
		flag++;
	return flag;
}

N
Nick Piggin 已提交
2088
/*
2089
 * Handle the last step of open()
N
Nick Piggin 已提交
2090
 */
2091
static struct file *do_last(struct nameidata *nd, struct path *path,
2092
			    const struct open_flags *op, const char *pathname)
2093
{
2094
	struct dentry *dir = nd->path.dentry;
2095
	struct dentry *dentry;
2096
	int open_flag = op->open_flag;
2097
	int will_truncate = open_flag & O_TRUNC;
2098
	int want_write = 0;
A
Al Viro 已提交
2099
	int acc_mode = op->acc_mode;
2100
	struct file *filp;
A
Al Viro 已提交
2101
	int error;
2102

2103 2104 2105
	nd->flags &= ~LOOKUP_PARENT;
	nd->flags |= op->intent;

2106 2107
	switch (nd->last_type) {
	case LAST_DOTDOT:
2108
	case LAST_DOT:
2109 2110 2111
		error = handle_dots(nd, nd->last_type);
		if (error)
			return ERR_PTR(error);
2112 2113
		/* fallthrough */
	case LAST_ROOT:
2114
		error = complete_walk(nd);
A
Al Viro 已提交
2115
		if (error)
2116
			return ERR_PTR(error);
2117
		audit_inode(pathname, nd->path.dentry);
2118
		if (open_flag & O_CREAT) {
2119 2120 2121 2122
			error = -EISDIR;
			goto exit;
		}
		goto ok;
2123
	case LAST_BIND:
2124
		error = complete_walk(nd);
A
Al Viro 已提交
2125
		if (error)
2126
			return ERR_PTR(error);
2127
		audit_inode(pathname, dir);
2128
		goto ok;
2129
	}
2130

2131
	if (!(open_flag & O_CREAT)) {
A
Al Viro 已提交
2132
		int symlink_ok = 0;
2133 2134
		if (nd->last.name[nd->last.len])
			nd->flags |= LOOKUP_FOLLOW | LOOKUP_DIRECTORY;
A
Al Viro 已提交
2135 2136
		if (open_flag & O_PATH && !(nd->flags & LOOKUP_FOLLOW))
			symlink_ok = 1;
2137
		/* we _can_ be in RCU mode here */
2138 2139 2140
		error = walk_component(nd, path, &nd->last, LAST_NORM,
					!symlink_ok);
		if (error < 0)
2141
			return ERR_PTR(error);
2142
		if (error) /* symlink */
2143 2144
			return NULL;
		/* sayonara */
2145 2146 2147
		error = complete_walk(nd);
		if (error)
			return ERR_PTR(-ECHILD);
2148 2149 2150

		error = -ENOTDIR;
		if (nd->flags & LOOKUP_DIRECTORY) {
2151
			if (!nd->inode->i_op->lookup)
2152 2153 2154 2155 2156 2157 2158
				goto exit;
		}
		audit_inode(pathname, nd->path.dentry);
		goto ok;
	}

	/* create side of things */
2159 2160 2161
	error = complete_walk(nd);
	if (error)
		return ERR_PTR(error);
2162 2163

	audit_inode(pathname, dir);
A
Al Viro 已提交
2164
	error = -EISDIR;
2165
	/* trailing slashes? */
N
Nick Piggin 已提交
2166 2167
	if (nd->last.name[nd->last.len])
		goto exit;
A
Al Viro 已提交
2168

2169 2170
	mutex_lock(&dir->d_inode->i_mutex);

2171 2172 2173
	dentry = lookup_hash(nd);
	error = PTR_ERR(dentry);
	if (IS_ERR(dentry)) {
2174 2175 2176 2177
		mutex_unlock(&dir->d_inode->i_mutex);
		goto exit;
	}

2178 2179 2180
	path->dentry = dentry;
	path->mnt = nd->path.mnt;

2181
	/* Negative dentry, just create the file */
2182 2183 2184 2185
	if (!dentry->d_inode) {
		int mode = op->mode;
		if (!IS_POSIXACL(dir->d_inode))
			mode &= ~current_umask();
2186 2187
		/*
		 * This write is needed to ensure that a
2188
		 * rw->ro transition does not occur between
2189 2190 2191 2192 2193 2194 2195
		 * the time when the file is created and when
		 * a permanent write count is taken through
		 * the 'struct file' in nameidata_to_filp().
		 */
		error = mnt_want_write(nd->path.mnt);
		if (error)
			goto exit_mutex_unlock;
2196
		want_write = 1;
2197
		/* Don't check for write permission, don't truncate */
2198
		open_flag &= ~O_TRUNC;
2199
		will_truncate = 0;
A
Al Viro 已提交
2200
		acc_mode = MAY_OPEN;
2201 2202 2203 2204 2205 2206 2207 2208 2209
		error = security_path_mknod(&nd->path, dentry, mode, 0);
		if (error)
			goto exit_mutex_unlock;
		error = vfs_create(dir->d_inode, dentry, mode, nd);
		if (error)
			goto exit_mutex_unlock;
		mutex_unlock(&dir->d_inode->i_mutex);
		dput(nd->path.dentry);
		nd->path.dentry = dentry;
2210
		goto common;
2211 2212 2213 2214 2215 2216 2217 2218 2219
	}

	/*
	 * It already exists.
	 */
	mutex_unlock(&dir->d_inode->i_mutex);
	audit_inode(pathname, path->dentry);

	error = -EEXIST;
2220
	if (open_flag & O_EXCL)
2221 2222
		goto exit_dput;

2223 2224 2225
	error = follow_managed(path, nd->flags);
	if (error < 0)
		goto exit_dput;
2226 2227 2228 2229

	error = -ENOENT;
	if (!path->dentry->d_inode)
		goto exit_dput;
A
Al Viro 已提交
2230 2231

	if (path->dentry->d_inode->i_op->follow_link)
2232 2233 2234
		return NULL;

	path_to_nameidata(path, nd);
N
Nick Piggin 已提交
2235
	nd->inode = path->dentry->d_inode;
2236
	error = -EISDIR;
N
Nick Piggin 已提交
2237
	if (S_ISDIR(nd->inode->i_mode))
2238
		goto exit;
2239
ok:
2240 2241 2242
	if (!S_ISREG(nd->inode->i_mode))
		will_truncate = 0;

2243 2244 2245 2246
	if (will_truncate) {
		error = mnt_want_write(nd->path.mnt);
		if (error)
			goto exit;
2247
		want_write = 1;
2248
	}
2249
common:
A
Al Viro 已提交
2250
	error = may_open(&nd->path, acc_mode, open_flag);
2251
	if (error)
2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269
		goto exit;
	filp = nameidata_to_filp(nd);
	if (!IS_ERR(filp)) {
		error = ima_file_check(filp, op->acc_mode);
		if (error) {
			fput(filp);
			filp = ERR_PTR(error);
		}
	}
	if (!IS_ERR(filp)) {
		if (will_truncate) {
			error = handle_truncate(filp);
			if (error) {
				fput(filp);
				filp = ERR_PTR(error);
			}
		}
	}
2270 2271
out:
	if (want_write)
2272 2273
		mnt_drop_write(nd->path.mnt);
	path_put(&nd->path);
2274 2275 2276 2277 2278 2279 2280
	return filp;

exit_mutex_unlock:
	mutex_unlock(&dir->d_inode->i_mutex);
exit_dput:
	path_put_conditional(path, nd);
exit:
2281 2282
	filp = ERR_PTR(error);
	goto out;
2283 2284
}

2285
static struct file *path_openat(int dfd, const char *pathname,
A
Al Viro 已提交
2286
		struct nameidata *nd, const struct open_flags *op, int flags)
L
Linus Torvalds 已提交
2287
{
2288
	struct file *base = NULL;
2289
	struct file *filp;
2290
	struct path path;
2291
	int error;
N
Nick Piggin 已提交
2292 2293 2294 2295 2296

	filp = get_empty_filp();
	if (!filp)
		return ERR_PTR(-ENFILE);

2297
	filp->f_flags = op->open_flag;
A
Al Viro 已提交
2298 2299 2300
	nd->intent.open.file = filp;
	nd->intent.open.flags = open_to_namei_flags(op->open_flag);
	nd->intent.open.create_mode = op->mode;
N
Nick Piggin 已提交
2301

A
Al Viro 已提交
2302
	error = path_init(dfd, pathname, flags | LOOKUP_PARENT, nd, &base);
N
Nick Piggin 已提交
2303
	if (unlikely(error))
2304
		goto out_filp;
N
Nick Piggin 已提交
2305

2306
	current->total_link_count = 0;
A
Al Viro 已提交
2307
	error = link_path_walk(pathname, nd);
N
Nick Piggin 已提交
2308 2309
	if (unlikely(error))
		goto out_filp;
L
Linus Torvalds 已提交
2310

A
Al Viro 已提交
2311
	filp = do_last(nd, &path, op, pathname);
2312
	while (unlikely(!filp)) { /* trailing symlink */
2313
		struct path link = path;
A
Al Viro 已提交
2314
		void *cookie;
2315
		if (!(nd->flags & LOOKUP_FOLLOW)) {
A
Al Viro 已提交
2316 2317
			path_put_conditional(&path, nd);
			path_put(&nd->path);
2318 2319 2320
			filp = ERR_PTR(-ELOOP);
			break;
		}
A
Al Viro 已提交
2321 2322
		nd->flags |= LOOKUP_PARENT;
		nd->flags &= ~(LOOKUP_OPEN|LOOKUP_CREATE|LOOKUP_EXCL);
2323
		error = follow_link(&link, nd, &cookie);
2324
		if (unlikely(error))
2325
			filp = ERR_PTR(error);
2326
		else
A
Al Viro 已提交
2327
			filp = do_last(nd, &path, op, pathname);
2328
		put_link(nd, &link, cookie);
2329
	}
A
Al Viro 已提交
2330
out:
A
Al Viro 已提交
2331 2332
	if (nd->root.mnt && !(nd->flags & LOOKUP_ROOT))
		path_put(&nd->root);
2333 2334
	if (base)
		fput(base);
A
Al Viro 已提交
2335
	release_open_intent(nd);
A
Al Viro 已提交
2336
	return filp;
L
Linus Torvalds 已提交
2337

N
Nick Piggin 已提交
2338
out_filp:
2339
	filp = ERR_PTR(error);
A
Al Viro 已提交
2340
	goto out;
L
Linus Torvalds 已提交
2341 2342
}

2343 2344 2345
struct file *do_filp_open(int dfd, const char *pathname,
		const struct open_flags *op, int flags)
{
A
Al Viro 已提交
2346
	struct nameidata nd;
2347 2348
	struct file *filp;

A
Al Viro 已提交
2349
	filp = path_openat(dfd, pathname, &nd, op, flags | LOOKUP_RCU);
2350
	if (unlikely(filp == ERR_PTR(-ECHILD)))
A
Al Viro 已提交
2351
		filp = path_openat(dfd, pathname, &nd, op, flags);
2352
	if (unlikely(filp == ERR_PTR(-ESTALE)))
A
Al Viro 已提交
2353
		filp = path_openat(dfd, pathname, &nd, op, flags | LOOKUP_REVAL);
2354 2355 2356
	return filp;
}

A
Al Viro 已提交
2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367
struct file *do_file_open_root(struct dentry *dentry, struct vfsmount *mnt,
		const char *name, const struct open_flags *op, int flags)
{
	struct nameidata nd;
	struct file *file;

	nd.root.mnt = mnt;
	nd.root.dentry = dentry;

	flags |= LOOKUP_ROOT;

A
Al Viro 已提交
2368
	if (dentry->d_inode->i_op->follow_link && op->intent & LOOKUP_OPEN)
A
Al Viro 已提交
2369 2370 2371 2372 2373 2374 2375 2376 2377 2378
		return ERR_PTR(-ELOOP);

	file = path_openat(-1, name, &nd, op, flags | LOOKUP_RCU);
	if (unlikely(file == ERR_PTR(-ECHILD)))
		file = path_openat(-1, name, &nd, op, flags);
	if (unlikely(file == ERR_PTR(-ESTALE)))
		file = path_openat(-1, name, &nd, op, flags | LOOKUP_REVAL);
	return file;
}

L
Linus Torvalds 已提交
2379 2380 2381 2382 2383 2384 2385
/**
 * lookup_create - lookup a dentry, creating it if it doesn't exist
 * @nd: nameidata info
 * @is_dir: directory flag
 *
 * Simple function to lookup and return a dentry and create it
 * if it doesn't exist.  Is SMP-safe.
2386
 *
2387
 * Returns with nd->path.dentry->d_inode->i_mutex locked.
L
Linus Torvalds 已提交
2388 2389 2390
 */
struct dentry *lookup_create(struct nameidata *nd, int is_dir)
{
2391
	struct dentry *dentry = ERR_PTR(-EEXIST);
L
Linus Torvalds 已提交
2392

2393
	mutex_lock_nested(&nd->path.dentry->d_inode->i_mutex, I_MUTEX_PARENT);
2394 2395 2396 2397
	/*
	 * Yucky last component or no last component at all?
	 * (foo/., foo/.., /////)
	 */
L
Linus Torvalds 已提交
2398 2399 2400
	if (nd->last_type != LAST_NORM)
		goto fail;
	nd->flags &= ~LOOKUP_PARENT;
2401
	nd->flags |= LOOKUP_CREATE | LOOKUP_EXCL;
2402
	nd->intent.open.flags = O_EXCL;
2403 2404 2405 2406

	/*
	 * Do the final lookup.
	 */
2407
	dentry = lookup_hash(nd);
L
Linus Torvalds 已提交
2408 2409
	if (IS_ERR(dentry))
		goto fail;
2410

2411 2412
	if (dentry->d_inode)
		goto eexist;
2413 2414 2415 2416 2417 2418
	/*
	 * Special case - lookup gave negative, but... we had foo/bar/
	 * From the vfs_mknod() POV we just have a negative dentry -
	 * all is fine. Let's be bastards - you had / on the end, you've
	 * been asking for (non-existent) directory. -ENOENT for you.
	 */
2419 2420 2421 2422
	if (unlikely(!is_dir && nd->last.name[nd->last.len])) {
		dput(dentry);
		dentry = ERR_PTR(-ENOENT);
	}
L
Linus Torvalds 已提交
2423
	return dentry;
2424
eexist:
L
Linus Torvalds 已提交
2425
	dput(dentry);
2426
	dentry = ERR_PTR(-EEXIST);
L
Linus Torvalds 已提交
2427 2428 2429
fail:
	return dentry;
}
2430
EXPORT_SYMBOL_GPL(lookup_create);
L
Linus Torvalds 已提交
2431 2432 2433

int vfs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t dev)
{
2434
	int error = may_create(dir, dentry);
L
Linus Torvalds 已提交
2435 2436 2437 2438

	if (error)
		return error;

2439 2440
	if ((S_ISCHR(mode) || S_ISBLK(mode)) &&
	    !ns_capable(inode_userns(dir), CAP_MKNOD))
L
Linus Torvalds 已提交
2441 2442
		return -EPERM;

A
Al Viro 已提交
2443
	if (!dir->i_op->mknod)
L
Linus Torvalds 已提交
2444 2445
		return -EPERM;

2446 2447 2448 2449
	error = devcgroup_inode_mknod(mode, dev);
	if (error)
		return error;

L
Linus Torvalds 已提交
2450 2451 2452 2453 2454
	error = security_inode_mknod(dir, dentry, mode, dev);
	if (error)
		return error;

	error = dir->i_op->mknod(dir, dentry, mode, dev);
2455
	if (!error)
2456
		fsnotify_create(dir, dentry);
L
Linus Torvalds 已提交
2457 2458 2459
	return error;
}

2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476
static int may_mknod(mode_t mode)
{
	switch (mode & S_IFMT) {
	case S_IFREG:
	case S_IFCHR:
	case S_IFBLK:
	case S_IFIFO:
	case S_IFSOCK:
	case 0: /* zero mode translates to S_IFREG */
		return 0;
	case S_IFDIR:
		return -EPERM;
	default:
		return -EINVAL;
	}
}

2477 2478
SYSCALL_DEFINE4(mknodat, int, dfd, const char __user *, filename, int, mode,
		unsigned, dev)
L
Linus Torvalds 已提交
2479
{
2480 2481 2482
	int error;
	char *tmp;
	struct dentry *dentry;
L
Linus Torvalds 已提交
2483 2484 2485 2486 2487
	struct nameidata nd;

	if (S_ISDIR(mode))
		return -EPERM;

2488
	error = user_path_parent(dfd, filename, &nd, &tmp);
L
Linus Torvalds 已提交
2489
	if (error)
2490 2491
		return error;

L
Linus Torvalds 已提交
2492
	dentry = lookup_create(&nd, 0);
2493 2494 2495 2496
	if (IS_ERR(dentry)) {
		error = PTR_ERR(dentry);
		goto out_unlock;
	}
2497
	if (!IS_POSIXACL(nd.path.dentry->d_inode))
A
Al Viro 已提交
2498
		mode &= ~current_umask();
2499 2500 2501 2502 2503 2504
	error = may_mknod(mode);
	if (error)
		goto out_dput;
	error = mnt_want_write(nd.path.mnt);
	if (error)
		goto out_dput;
2505 2506 2507
	error = security_path_mknod(&nd.path, dentry, mode, dev);
	if (error)
		goto out_drop_write;
2508
	switch (mode & S_IFMT) {
L
Linus Torvalds 已提交
2509
		case 0: case S_IFREG:
2510
			error = vfs_create(nd.path.dentry->d_inode,dentry,mode,&nd);
L
Linus Torvalds 已提交
2511 2512
			break;
		case S_IFCHR: case S_IFBLK:
2513
			error = vfs_mknod(nd.path.dentry->d_inode,dentry,mode,
L
Linus Torvalds 已提交
2514 2515 2516
					new_decode_dev(dev));
			break;
		case S_IFIFO: case S_IFSOCK:
2517
			error = vfs_mknod(nd.path.dentry->d_inode,dentry,mode,0);
L
Linus Torvalds 已提交
2518 2519
			break;
	}
2520
out_drop_write:
2521 2522 2523 2524
	mnt_drop_write(nd.path.mnt);
out_dput:
	dput(dentry);
out_unlock:
2525
	mutex_unlock(&nd.path.dentry->d_inode->i_mutex);
J
Jan Blunck 已提交
2526
	path_put(&nd.path);
L
Linus Torvalds 已提交
2527 2528 2529 2530 2531
	putname(tmp);

	return error;
}

2532
SYSCALL_DEFINE3(mknod, const char __user *, filename, int, mode, unsigned, dev)
2533 2534 2535 2536
{
	return sys_mknodat(AT_FDCWD, filename, mode, dev);
}

L
Linus Torvalds 已提交
2537 2538
int vfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
{
2539
	int error = may_create(dir, dentry);
L
Linus Torvalds 已提交
2540 2541 2542 2543

	if (error)
		return error;

A
Al Viro 已提交
2544
	if (!dir->i_op->mkdir)
L
Linus Torvalds 已提交
2545 2546 2547 2548 2549 2550 2551 2552
		return -EPERM;

	mode &= (S_IRWXUGO|S_ISVTX);
	error = security_inode_mkdir(dir, dentry, mode);
	if (error)
		return error;

	error = dir->i_op->mkdir(dir, dentry, mode);
2553
	if (!error)
2554
		fsnotify_mkdir(dir, dentry);
L
Linus Torvalds 已提交
2555 2556 2557
	return error;
}

2558
SYSCALL_DEFINE3(mkdirat, int, dfd, const char __user *, pathname, int, mode)
L
Linus Torvalds 已提交
2559 2560 2561
{
	int error = 0;
	char * tmp;
2562 2563
	struct dentry *dentry;
	struct nameidata nd;
L
Linus Torvalds 已提交
2564

2565 2566
	error = user_path_parent(dfd, pathname, &nd, &tmp);
	if (error)
2567
		goto out_err;
L
Linus Torvalds 已提交
2568

2569 2570 2571 2572
	dentry = lookup_create(&nd, 1);
	error = PTR_ERR(dentry);
	if (IS_ERR(dentry))
		goto out_unlock;
L
Linus Torvalds 已提交
2573

2574
	if (!IS_POSIXACL(nd.path.dentry->d_inode))
A
Al Viro 已提交
2575
		mode &= ~current_umask();
2576 2577 2578
	error = mnt_want_write(nd.path.mnt);
	if (error)
		goto out_dput;
2579 2580 2581
	error = security_path_mkdir(&nd.path, dentry, mode);
	if (error)
		goto out_drop_write;
2582
	error = vfs_mkdir(nd.path.dentry->d_inode, dentry, mode);
2583
out_drop_write:
2584 2585
	mnt_drop_write(nd.path.mnt);
out_dput:
2586 2587
	dput(dentry);
out_unlock:
2588
	mutex_unlock(&nd.path.dentry->d_inode->i_mutex);
J
Jan Blunck 已提交
2589
	path_put(&nd.path);
2590 2591
	putname(tmp);
out_err:
L
Linus Torvalds 已提交
2592 2593 2594
	return error;
}

2595
SYSCALL_DEFINE2(mkdir, const char __user *, pathname, int, mode)
2596 2597 2598 2599
{
	return sys_mkdirat(AT_FDCWD, pathname, mode);
}

L
Linus Torvalds 已提交
2600
/*
S
Sage Weil 已提交
2601 2602 2603 2604
 * The dentry_unhash() helper will try to drop the dentry early: we
 * should have a usage count of 2 if we're the only user of this
 * dentry, and if that is true (possibly after pruning the dcache),
 * then we drop the dentry now.
L
Linus Torvalds 已提交
2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616
 *
 * A low-level filesystem can, if it choses, legally
 * do a
 *
 *	if (!d_unhashed(dentry))
 *		return -EBUSY;
 *
 * if it cannot handle the case of removing a directory
 * that is still in use by something else..
 */
void dentry_unhash(struct dentry *dentry)
{
2617
	shrink_dcache_parent(dentry);
L
Linus Torvalds 已提交
2618
	spin_lock(&dentry->d_lock);
2619
	if (dentry->d_count == 1)
L
Linus Torvalds 已提交
2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630
		__d_drop(dentry);
	spin_unlock(&dentry->d_lock);
}

int vfs_rmdir(struct inode *dir, struct dentry *dentry)
{
	int error = may_delete(dir, dentry, 1);

	if (error)
		return error;

A
Al Viro 已提交
2631
	if (!dir->i_op->rmdir)
L
Linus Torvalds 已提交
2632 2633
		return -EPERM;

2634
	mutex_lock(&dentry->d_inode->i_mutex);
S
Sage Weil 已提交
2635 2636

	error = -EBUSY;
L
Linus Torvalds 已提交
2637
	if (d_mountpoint(dentry))
S
Sage Weil 已提交
2638 2639 2640 2641 2642 2643
		goto out;

	error = security_inode_rmdir(dir, dentry);
	if (error)
		goto out;

2644
	shrink_dcache_parent(dentry);
S
Sage Weil 已提交
2645 2646 2647 2648 2649 2650 2651 2652
	error = dir->i_op->rmdir(dir, dentry);
	if (error)
		goto out;

	dentry->d_inode->i_flags |= S_DEAD;
	dont_mount(dentry);

out:
2653
	mutex_unlock(&dentry->d_inode->i_mutex);
S
Sage Weil 已提交
2654
	if (!error)
L
Linus Torvalds 已提交
2655 2656 2657 2658
		d_delete(dentry);
	return error;
}

2659
static long do_rmdir(int dfd, const char __user *pathname)
L
Linus Torvalds 已提交
2660 2661 2662 2663 2664 2665
{
	int error = 0;
	char * name;
	struct dentry *dentry;
	struct nameidata nd;

2666
	error = user_path_parent(dfd, pathname, &nd, &name);
L
Linus Torvalds 已提交
2667
	if (error)
2668
		return error;
L
Linus Torvalds 已提交
2669 2670

	switch(nd.last_type) {
2671 2672 2673 2674 2675 2676 2677 2678 2679
	case LAST_DOTDOT:
		error = -ENOTEMPTY;
		goto exit1;
	case LAST_DOT:
		error = -EINVAL;
		goto exit1;
	case LAST_ROOT:
		error = -EBUSY;
		goto exit1;
L
Linus Torvalds 已提交
2680
	}
2681 2682 2683

	nd.flags &= ~LOOKUP_PARENT;

2684
	mutex_lock_nested(&nd.path.dentry->d_inode->i_mutex, I_MUTEX_PARENT);
2685
	dentry = lookup_hash(&nd);
L
Linus Torvalds 已提交
2686
	error = PTR_ERR(dentry);
2687 2688
	if (IS_ERR(dentry))
		goto exit2;
2689 2690 2691 2692
	if (!dentry->d_inode) {
		error = -ENOENT;
		goto exit3;
	}
2693 2694 2695
	error = mnt_want_write(nd.path.mnt);
	if (error)
		goto exit3;
2696 2697 2698
	error = security_path_rmdir(&nd.path, dentry);
	if (error)
		goto exit4;
2699
	error = vfs_rmdir(nd.path.dentry->d_inode, dentry);
2700
exit4:
2701 2702
	mnt_drop_write(nd.path.mnt);
exit3:
2703 2704
	dput(dentry);
exit2:
2705
	mutex_unlock(&nd.path.dentry->d_inode->i_mutex);
L
Linus Torvalds 已提交
2706
exit1:
J
Jan Blunck 已提交
2707
	path_put(&nd.path);
L
Linus Torvalds 已提交
2708 2709 2710 2711
	putname(name);
	return error;
}

2712
SYSCALL_DEFINE1(rmdir, const char __user *, pathname)
2713 2714 2715 2716
{
	return do_rmdir(AT_FDCWD, pathname);
}

L
Linus Torvalds 已提交
2717 2718 2719 2720 2721 2722 2723
int vfs_unlink(struct inode *dir, struct dentry *dentry)
{
	int error = may_delete(dir, dentry, 0);

	if (error)
		return error;

A
Al Viro 已提交
2724
	if (!dir->i_op->unlink)
L
Linus Torvalds 已提交
2725 2726
		return -EPERM;

2727
	mutex_lock(&dentry->d_inode->i_mutex);
L
Linus Torvalds 已提交
2728 2729 2730 2731
	if (d_mountpoint(dentry))
		error = -EBUSY;
	else {
		error = security_inode_unlink(dir, dentry);
2732
		if (!error) {
L
Linus Torvalds 已提交
2733
			error = dir->i_op->unlink(dir, dentry);
2734
			if (!error)
2735
				dont_mount(dentry);
2736
		}
L
Linus Torvalds 已提交
2737
	}
2738
	mutex_unlock(&dentry->d_inode->i_mutex);
L
Linus Torvalds 已提交
2739 2740 2741

	/* We don't d_delete() NFS sillyrenamed files--they still exist. */
	if (!error && !(dentry->d_flags & DCACHE_NFSFS_RENAMED)) {
2742
		fsnotify_link_count(dentry->d_inode);
J
John McCutchan 已提交
2743
		d_delete(dentry);
L
Linus Torvalds 已提交
2744
	}
R
Robert Love 已提交
2745

L
Linus Torvalds 已提交
2746 2747 2748 2749 2750
	return error;
}

/*
 * Make sure that the actual truncation of the file will occur outside its
2751
 * directory's i_mutex.  Truncate can take a long time if there is a lot of
L
Linus Torvalds 已提交
2752 2753 2754
 * writeout happening, and we don't want to prevent access to the directory
 * while waiting on the I/O.
 */
2755
static long do_unlinkat(int dfd, const char __user *pathname)
L
Linus Torvalds 已提交
2756
{
2757 2758
	int error;
	char *name;
L
Linus Torvalds 已提交
2759 2760 2761 2762
	struct dentry *dentry;
	struct nameidata nd;
	struct inode *inode = NULL;

2763
	error = user_path_parent(dfd, pathname, &nd, &name);
L
Linus Torvalds 已提交
2764
	if (error)
2765 2766
		return error;

L
Linus Torvalds 已提交
2767 2768 2769
	error = -EISDIR;
	if (nd.last_type != LAST_NORM)
		goto exit1;
2770 2771 2772

	nd.flags &= ~LOOKUP_PARENT;

2773
	mutex_lock_nested(&nd.path.dentry->d_inode->i_mutex, I_MUTEX_PARENT);
2774
	dentry = lookup_hash(&nd);
L
Linus Torvalds 已提交
2775 2776 2777
	error = PTR_ERR(dentry);
	if (!IS_ERR(dentry)) {
		/* Why not before? Because we want correct error value */
2778 2779
		if (nd.last.name[nd.last.len])
			goto slashes;
L
Linus Torvalds 已提交
2780
		inode = dentry->d_inode;
2781
		if (!inode)
2782 2783
			goto slashes;
		ihold(inode);
2784 2785 2786
		error = mnt_want_write(nd.path.mnt);
		if (error)
			goto exit2;
2787 2788 2789
		error = security_path_unlink(&nd.path, dentry);
		if (error)
			goto exit3;
2790
		error = vfs_unlink(nd.path.dentry->d_inode, dentry);
2791
exit3:
2792
		mnt_drop_write(nd.path.mnt);
L
Linus Torvalds 已提交
2793 2794 2795
	exit2:
		dput(dentry);
	}
2796
	mutex_unlock(&nd.path.dentry->d_inode->i_mutex);
L
Linus Torvalds 已提交
2797 2798 2799
	if (inode)
		iput(inode);	/* truncate the inode here */
exit1:
J
Jan Blunck 已提交
2800
	path_put(&nd.path);
L
Linus Torvalds 已提交
2801 2802 2803 2804 2805 2806 2807 2808 2809
	putname(name);
	return error;

slashes:
	error = !dentry->d_inode ? -ENOENT :
		S_ISDIR(dentry->d_inode->i_mode) ? -EISDIR : -ENOTDIR;
	goto exit2;
}

2810
SYSCALL_DEFINE3(unlinkat, int, dfd, const char __user *, pathname, int, flag)
2811 2812 2813 2814 2815 2816 2817 2818 2819 2820
{
	if ((flag & ~AT_REMOVEDIR) != 0)
		return -EINVAL;

	if (flag & AT_REMOVEDIR)
		return do_rmdir(dfd, pathname);

	return do_unlinkat(dfd, pathname);
}

2821
SYSCALL_DEFINE1(unlink, const char __user *, pathname)
2822 2823 2824 2825
{
	return do_unlinkat(AT_FDCWD, pathname);
}

2826
int vfs_symlink(struct inode *dir, struct dentry *dentry, const char *oldname)
L
Linus Torvalds 已提交
2827
{
2828
	int error = may_create(dir, dentry);
L
Linus Torvalds 已提交
2829 2830 2831 2832

	if (error)
		return error;

A
Al Viro 已提交
2833
	if (!dir->i_op->symlink)
L
Linus Torvalds 已提交
2834 2835 2836 2837 2838 2839 2840
		return -EPERM;

	error = security_inode_symlink(dir, dentry, oldname);
	if (error)
		return error;

	error = dir->i_op->symlink(dir, dentry, oldname);
2841
	if (!error)
2842
		fsnotify_create(dir, dentry);
L
Linus Torvalds 已提交
2843 2844 2845
	return error;
}

2846 2847
SYSCALL_DEFINE3(symlinkat, const char __user *, oldname,
		int, newdfd, const char __user *, newname)
L
Linus Torvalds 已提交
2848
{
2849 2850 2851
	int error;
	char *from;
	char *to;
2852 2853
	struct dentry *dentry;
	struct nameidata nd;
L
Linus Torvalds 已提交
2854 2855

	from = getname(oldname);
2856
	if (IS_ERR(from))
L
Linus Torvalds 已提交
2857 2858
		return PTR_ERR(from);

2859
	error = user_path_parent(newdfd, newname, &nd, &to);
2860
	if (error)
2861 2862
		goto out_putname;

2863 2864 2865 2866 2867
	dentry = lookup_create(&nd, 0);
	error = PTR_ERR(dentry);
	if (IS_ERR(dentry))
		goto out_unlock;

2868 2869 2870
	error = mnt_want_write(nd.path.mnt);
	if (error)
		goto out_dput;
2871 2872 2873
	error = security_path_symlink(&nd.path, dentry, from);
	if (error)
		goto out_drop_write;
2874
	error = vfs_symlink(nd.path.dentry->d_inode, dentry, from);
2875
out_drop_write:
2876 2877
	mnt_drop_write(nd.path.mnt);
out_dput:
2878 2879
	dput(dentry);
out_unlock:
2880
	mutex_unlock(&nd.path.dentry->d_inode->i_mutex);
J
Jan Blunck 已提交
2881
	path_put(&nd.path);
2882 2883
	putname(to);
out_putname:
L
Linus Torvalds 已提交
2884 2885 2886 2887
	putname(from);
	return error;
}

2888
SYSCALL_DEFINE2(symlink, const char __user *, oldname, const char __user *, newname)
2889 2890 2891 2892
{
	return sys_symlinkat(oldname, AT_FDCWD, newname);
}

L
Linus Torvalds 已提交
2893 2894 2895 2896 2897 2898 2899 2900
int vfs_link(struct dentry *old_dentry, struct inode *dir, struct dentry *new_dentry)
{
	struct inode *inode = old_dentry->d_inode;
	int error;

	if (!inode)
		return -ENOENT;

2901
	error = may_create(dir, new_dentry);
L
Linus Torvalds 已提交
2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912
	if (error)
		return error;

	if (dir->i_sb != inode->i_sb)
		return -EXDEV;

	/*
	 * A link to an append-only or immutable file cannot be created.
	 */
	if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
		return -EPERM;
A
Al Viro 已提交
2913
	if (!dir->i_op->link)
L
Linus Torvalds 已提交
2914
		return -EPERM;
2915
	if (S_ISDIR(inode->i_mode))
L
Linus Torvalds 已提交
2916 2917 2918 2919 2920 2921
		return -EPERM;

	error = security_inode_link(old_dentry, dir, new_dentry);
	if (error)
		return error;

2922
	mutex_lock(&inode->i_mutex);
2923 2924 2925 2926 2927
	/* Make sure we don't allow creating hardlink to an unlinked file */
	if (inode->i_nlink == 0)
		error =  -ENOENT;
	else
		error = dir->i_op->link(old_dentry, dir, new_dentry);
2928
	mutex_unlock(&inode->i_mutex);
2929
	if (!error)
2930
		fsnotify_link(dir, inode, new_dentry);
L
Linus Torvalds 已提交
2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942
	return error;
}

/*
 * Hardlinks are often used in delicate situations.  We avoid
 * security-related surprises by not following symlinks on the
 * newname.  --KAB
 *
 * We don't follow them on the oldname either to be compatible
 * with linux 2.0, and to avoid hard-linking to directories
 * and other special files.  --ADM
 */
2943 2944
SYSCALL_DEFINE5(linkat, int, olddfd, const char __user *, oldname,
		int, newdfd, const char __user *, newname, int, flags)
L
Linus Torvalds 已提交
2945 2946
{
	struct dentry *new_dentry;
2947 2948
	struct nameidata nd;
	struct path old_path;
2949
	int how = 0;
L
Linus Torvalds 已提交
2950
	int error;
2951
	char *to;
L
Linus Torvalds 已提交
2952

2953
	if ((flags & ~(AT_SYMLINK_FOLLOW | AT_EMPTY_PATH)) != 0)
2954
		return -EINVAL;
2955 2956 2957 2958 2959 2960 2961 2962 2963 2964 2965 2966 2967
	/*
	 * To use null names we require CAP_DAC_READ_SEARCH
	 * This ensures that not everyone will be able to create
	 * handlink using the passed filedescriptor.
	 */
	if (flags & AT_EMPTY_PATH) {
		if (!capable(CAP_DAC_READ_SEARCH))
			return -ENOENT;
		how = LOOKUP_EMPTY;
	}

	if (flags & AT_SYMLINK_FOLLOW)
		how |= LOOKUP_FOLLOW;
2968

2969
	error = user_path_at(olddfd, oldname, how, &old_path);
L
Linus Torvalds 已提交
2970
	if (error)
2971 2972 2973
		return error;

	error = user_path_parent(newdfd, newname, &nd, &to);
L
Linus Torvalds 已提交
2974 2975 2976
	if (error)
		goto out;
	error = -EXDEV;
2977
	if (old_path.mnt != nd.path.mnt)
L
Linus Torvalds 已提交
2978 2979 2980
		goto out_release;
	new_dentry = lookup_create(&nd, 0);
	error = PTR_ERR(new_dentry);
2981 2982
	if (IS_ERR(new_dentry))
		goto out_unlock;
2983 2984 2985
	error = mnt_want_write(nd.path.mnt);
	if (error)
		goto out_dput;
2986 2987 2988
	error = security_path_link(old_path.dentry, &nd.path, new_dentry);
	if (error)
		goto out_drop_write;
2989
	error = vfs_link(old_path.dentry, nd.path.dentry->d_inode, new_dentry);
2990
out_drop_write:
2991 2992
	mnt_drop_write(nd.path.mnt);
out_dput:
2993 2994
	dput(new_dentry);
out_unlock:
2995
	mutex_unlock(&nd.path.dentry->d_inode->i_mutex);
L
Linus Torvalds 已提交
2996
out_release:
J
Jan Blunck 已提交
2997
	path_put(&nd.path);
2998
	putname(to);
L
Linus Torvalds 已提交
2999
out:
3000
	path_put(&old_path);
L
Linus Torvalds 已提交
3001 3002 3003 3004

	return error;
}

3005
SYSCALL_DEFINE2(link, const char __user *, oldname, const char __user *, newname)
3006
{
3007
	return sys_linkat(AT_FDCWD, oldname, AT_FDCWD, newname, 0);
3008 3009
}

L
Linus Torvalds 已提交
3010 3011 3012 3013 3014 3015 3016
/*
 * The worst of all namespace operations - renaming directory. "Perverted"
 * doesn't even start to describe it. Somebody in UCB had a heck of a trip...
 * Problems:
 *	a) we can get into loop creation. Check is done in is_subdir().
 *	b) race potential - two innocent renames can create a loop together.
 *	   That's where 4.4 screws up. Current fix: serialization on
3017
 *	   sb->s_vfs_rename_mutex. We might be more accurate, but that's another
L
Linus Torvalds 已提交
3018 3019
 *	   story.
 *	c) we have to lock _three_ objects - parents and victim (if it exists).
3020
 *	   And that - after we got ->i_mutex on parents (until then we don't know
L
Linus Torvalds 已提交
3021 3022
 *	   whether the target exists).  Solution: try to be smart with locking
 *	   order for inodes.  We rely on the fact that tree topology may change
3023
 *	   only under ->s_vfs_rename_mutex _and_ that parent of the object we
L
Linus Torvalds 已提交
3024 3025 3026
 *	   move will be locked.  Thus we can rank directories by the tree
 *	   (ancestors first) and rank all non-directories after them.
 *	   That works since everybody except rename does "lock parent, lookup,
3027
 *	   lock child" and rename is under ->s_vfs_rename_mutex.
L
Linus Torvalds 已提交
3028 3029 3030
 *	   HOWEVER, it relies on the assumption that any object with ->lookup()
 *	   has no more than 1 dentry.  If "hybrid" objects will ever appear,
 *	   we'd better make sure that there's no link(2) for them.
3031
 *	d) conversion from fhandle to dentry may come in the wrong moment - when
3032
 *	   we are removing the target. Solution: we will have to grab ->i_mutex
L
Linus Torvalds 已提交
3033
 *	   in the fhandle_to_dentry code. [FIXME - current nfsfh.c relies on
3034
 *	   ->i_mutex on parents, which works but leads to some truly excessive
L
Linus Torvalds 已提交
3035 3036
 *	   locking].
 */
A
Adrian Bunk 已提交
3037 3038
static int vfs_rename_dir(struct inode *old_dir, struct dentry *old_dentry,
			  struct inode *new_dir, struct dentry *new_dentry)
L
Linus Torvalds 已提交
3039 3040
{
	int error = 0;
S
Sage Weil 已提交
3041
	struct inode *target = new_dentry->d_inode;
L
Linus Torvalds 已提交
3042 3043 3044 3045 3046 3047

	/*
	 * If we are going to change the parent - check write permissions,
	 * we'll need to flip '..'.
	 */
	if (new_dir != old_dir) {
3048
		error = inode_permission(old_dentry->d_inode, MAY_WRITE);
L
Linus Torvalds 已提交
3049 3050 3051 3052 3053 3054 3055 3056
		if (error)
			return error;
	}

	error = security_inode_rename(old_dir, old_dentry, new_dir, new_dentry);
	if (error)
		return error;

3057
	if (target)
3058
		mutex_lock(&target->i_mutex);
S
Sage Weil 已提交
3059 3060 3061 3062 3063

	error = -EBUSY;
	if (d_mountpoint(old_dentry) || d_mountpoint(new_dentry))
		goto out;

3064 3065
	if (target)
		shrink_dcache_parent(new_dentry);
S
Sage Weil 已提交
3066 3067 3068 3069
	error = old_dir->i_op->rename(old_dir, old_dentry, new_dir, new_dentry);
	if (error)
		goto out;

L
Linus Torvalds 已提交
3070
	if (target) {
S
Sage Weil 已提交
3071 3072
		target->i_flags |= S_DEAD;
		dont_mount(new_dentry);
L
Linus Torvalds 已提交
3073
	}
S
Sage Weil 已提交
3074 3075 3076
out:
	if (target)
		mutex_unlock(&target->i_mutex);
3077
	if (!error)
3078 3079
		if (!(old_dir->i_sb->s_type->fs_flags & FS_RENAME_DOES_D_MOVE))
			d_move(old_dentry,new_dentry);
L
Linus Torvalds 已提交
3080 3081 3082
	return error;
}

A
Adrian Bunk 已提交
3083 3084
static int vfs_rename_other(struct inode *old_dir, struct dentry *old_dentry,
			    struct inode *new_dir, struct dentry *new_dentry)
L
Linus Torvalds 已提交
3085
{
S
Sage Weil 已提交
3086
	struct inode *target = new_dentry->d_inode;
L
Linus Torvalds 已提交
3087 3088 3089 3090 3091 3092 3093 3094
	int error;

	error = security_inode_rename(old_dir, old_dentry, new_dir, new_dentry);
	if (error)
		return error;

	dget(new_dentry);
	if (target)
3095
		mutex_lock(&target->i_mutex);
S
Sage Weil 已提交
3096 3097

	error = -EBUSY;
L
Linus Torvalds 已提交
3098
	if (d_mountpoint(old_dentry)||d_mountpoint(new_dentry))
S
Sage Weil 已提交
3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109
		goto out;

	error = old_dir->i_op->rename(old_dir, old_dentry, new_dir, new_dentry);
	if (error)
		goto out;

	if (target)
		dont_mount(new_dentry);
	if (!(old_dir->i_sb->s_type->fs_flags & FS_RENAME_DOES_D_MOVE))
		d_move(old_dentry, new_dentry);
out:
L
Linus Torvalds 已提交
3110
	if (target)
3111
		mutex_unlock(&target->i_mutex);
L
Linus Torvalds 已提交
3112 3113 3114 3115 3116 3117 3118 3119 3120
	dput(new_dentry);
	return error;
}

int vfs_rename(struct inode *old_dir, struct dentry *old_dentry,
	       struct inode *new_dir, struct dentry *new_dentry)
{
	int error;
	int is_dir = S_ISDIR(old_dentry->d_inode->i_mode);
3121
	const unsigned char *old_name;
L
Linus Torvalds 已提交
3122 3123 3124 3125 3126 3127 3128 3129 3130

	if (old_dentry->d_inode == new_dentry->d_inode)
 		return 0;
 
	error = may_delete(old_dir, old_dentry, is_dir);
	if (error)
		return error;

	if (!new_dentry->d_inode)
3131
		error = may_create(new_dir, new_dentry);
L
Linus Torvalds 已提交
3132 3133 3134 3135 3136
	else
		error = may_delete(new_dir, new_dentry, is_dir);
	if (error)
		return error;

A
Al Viro 已提交
3137
	if (!old_dir->i_op->rename)
L
Linus Torvalds 已提交
3138 3139
		return -EPERM;

R
Robert Love 已提交
3140 3141
	old_name = fsnotify_oldname_init(old_dentry->d_name.name);

L
Linus Torvalds 已提交
3142 3143 3144 3145
	if (is_dir)
		error = vfs_rename_dir(old_dir,old_dentry,new_dir,new_dentry);
	else
		error = vfs_rename_other(old_dir,old_dentry,new_dir,new_dentry);
3146 3147
	if (!error)
		fsnotify_move(old_dir, new_dir, old_name, is_dir,
3148
			      new_dentry->d_inode, old_dentry);
R
Robert Love 已提交
3149 3150
	fsnotify_oldname_free(old_name);

L
Linus Torvalds 已提交
3151 3152 3153
	return error;
}

3154 3155
SYSCALL_DEFINE4(renameat, int, olddfd, const char __user *, oldname,
		int, newdfd, const char __user *, newname)
L
Linus Torvalds 已提交
3156
{
3157 3158 3159
	struct dentry *old_dir, *new_dir;
	struct dentry *old_dentry, *new_dentry;
	struct dentry *trap;
L
Linus Torvalds 已提交
3160
	struct nameidata oldnd, newnd;
3161 3162 3163
	char *from;
	char *to;
	int error;
L
Linus Torvalds 已提交
3164

3165
	error = user_path_parent(olddfd, oldname, &oldnd, &from);
L
Linus Torvalds 已提交
3166 3167 3168
	if (error)
		goto exit;

3169
	error = user_path_parent(newdfd, newname, &newnd, &to);
L
Linus Torvalds 已提交
3170 3171 3172 3173
	if (error)
		goto exit1;

	error = -EXDEV;
3174
	if (oldnd.path.mnt != newnd.path.mnt)
L
Linus Torvalds 已提交
3175 3176
		goto exit2;

3177
	old_dir = oldnd.path.dentry;
L
Linus Torvalds 已提交
3178 3179 3180 3181
	error = -EBUSY;
	if (oldnd.last_type != LAST_NORM)
		goto exit2;

3182
	new_dir = newnd.path.dentry;
L
Linus Torvalds 已提交
3183 3184 3185
	if (newnd.last_type != LAST_NORM)
		goto exit2;

3186 3187
	oldnd.flags &= ~LOOKUP_PARENT;
	newnd.flags &= ~LOOKUP_PARENT;
3188
	newnd.flags |= LOOKUP_RENAME_TARGET;
3189

L
Linus Torvalds 已提交
3190 3191
	trap = lock_rename(new_dir, old_dir);

3192
	old_dentry = lookup_hash(&oldnd);
L
Linus Torvalds 已提交
3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211
	error = PTR_ERR(old_dentry);
	if (IS_ERR(old_dentry))
		goto exit3;
	/* source must exist */
	error = -ENOENT;
	if (!old_dentry->d_inode)
		goto exit4;
	/* unless the source is a directory trailing slashes give -ENOTDIR */
	if (!S_ISDIR(old_dentry->d_inode->i_mode)) {
		error = -ENOTDIR;
		if (oldnd.last.name[oldnd.last.len])
			goto exit4;
		if (newnd.last.name[newnd.last.len])
			goto exit4;
	}
	/* source should not be ancestor of target */
	error = -EINVAL;
	if (old_dentry == trap)
		goto exit4;
3212
	new_dentry = lookup_hash(&newnd);
L
Linus Torvalds 已提交
3213 3214 3215 3216 3217 3218 3219 3220
	error = PTR_ERR(new_dentry);
	if (IS_ERR(new_dentry))
		goto exit4;
	/* target should not be an ancestor of source */
	error = -ENOTEMPTY;
	if (new_dentry == trap)
		goto exit5;

3221 3222 3223
	error = mnt_want_write(oldnd.path.mnt);
	if (error)
		goto exit5;
3224 3225 3226 3227
	error = security_path_rename(&oldnd.path, old_dentry,
				     &newnd.path, new_dentry);
	if (error)
		goto exit6;
L
Linus Torvalds 已提交
3228 3229
	error = vfs_rename(old_dir->d_inode, old_dentry,
				   new_dir->d_inode, new_dentry);
3230
exit6:
3231
	mnt_drop_write(oldnd.path.mnt);
L
Linus Torvalds 已提交
3232 3233 3234 3235 3236 3237 3238
exit5:
	dput(new_dentry);
exit4:
	dput(old_dentry);
exit3:
	unlock_rename(new_dir, old_dir);
exit2:
J
Jan Blunck 已提交
3239
	path_put(&newnd.path);
3240
	putname(to);
L
Linus Torvalds 已提交
3241
exit1:
J
Jan Blunck 已提交
3242
	path_put(&oldnd.path);
L
Linus Torvalds 已提交
3243
	putname(from);
3244
exit:
L
Linus Torvalds 已提交
3245 3246 3247
	return error;
}

3248
SYSCALL_DEFINE2(rename, const char __user *, oldname, const char __user *, newname)
3249 3250 3251 3252
{
	return sys_renameat(AT_FDCWD, oldname, AT_FDCWD, newname);
}

L
Linus Torvalds 已提交
3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277
int vfs_readlink(struct dentry *dentry, char __user *buffer, int buflen, const char *link)
{
	int len;

	len = PTR_ERR(link);
	if (IS_ERR(link))
		goto out;

	len = strlen(link);
	if (len > (unsigned) buflen)
		len = buflen;
	if (copy_to_user(buffer, link, len))
		len = -EFAULT;
out:
	return len;
}

/*
 * A helper for ->readlink().  This should be used *ONLY* for symlinks that
 * have ->follow_link() touching nd only in nd_set_link().  Using (or not
 * using) it for any given inode is up to filesystem.
 */
int generic_readlink(struct dentry *dentry, char __user *buffer, int buflen)
{
	struct nameidata nd;
3278
	void *cookie;
3279
	int res;
3280

L
Linus Torvalds 已提交
3281
	nd.depth = 0;
3282
	cookie = dentry->d_inode->i_op->follow_link(dentry, &nd);
3283 3284 3285 3286 3287 3288 3289
	if (IS_ERR(cookie))
		return PTR_ERR(cookie);

	res = vfs_readlink(dentry, buffer, buflen, nd_get_link(&nd));
	if (dentry->d_inode->i_op->put_link)
		dentry->d_inode->i_op->put_link(dentry, &nd, cookie);
	return res;
L
Linus Torvalds 已提交
3290 3291 3292 3293 3294 3295 3296 3297 3298 3299
}

int vfs_follow_link(struct nameidata *nd, const char *link)
{
	return __vfs_follow_link(nd, link);
}

/* get the link contents into pagecache */
static char *page_getlink(struct dentry * dentry, struct page **ppage)
{
3300 3301
	char *kaddr;
	struct page *page;
L
Linus Torvalds 已提交
3302
	struct address_space *mapping = dentry->d_inode->i_mapping;
3303
	page = read_mapping_page(mapping, 0, NULL);
L
Linus Torvalds 已提交
3304
	if (IS_ERR(page))
3305
		return (char*)page;
L
Linus Torvalds 已提交
3306
	*ppage = page;
3307 3308 3309
	kaddr = kmap(page);
	nd_terminate_link(kaddr, dentry->d_inode->i_size, PAGE_SIZE - 1);
	return kaddr;
L
Linus Torvalds 已提交
3310 3311 3312 3313 3314 3315 3316 3317 3318 3319 3320 3321 3322 3323
}

int page_readlink(struct dentry *dentry, char __user *buffer, int buflen)
{
	struct page *page = NULL;
	char *s = page_getlink(dentry, &page);
	int res = vfs_readlink(dentry,buffer,buflen,s);
	if (page) {
		kunmap(page);
		page_cache_release(page);
	}
	return res;
}

3324
void *page_follow_link_light(struct dentry *dentry, struct nameidata *nd)
L
Linus Torvalds 已提交
3325
{
3326
	struct page *page = NULL;
L
Linus Torvalds 已提交
3327
	nd_set_link(nd, page_getlink(dentry, &page));
3328
	return page;
L
Linus Torvalds 已提交
3329 3330
}

3331
void page_put_link(struct dentry *dentry, struct nameidata *nd, void *cookie)
L
Linus Torvalds 已提交
3332
{
3333 3334 3335
	struct page *page = cookie;

	if (page) {
L
Linus Torvalds 已提交
3336 3337 3338 3339 3340
		kunmap(page);
		page_cache_release(page);
	}
}

3341 3342 3343 3344
/*
 * The nofs argument instructs pagecache_write_begin to pass AOP_FLAG_NOFS
 */
int __page_symlink(struct inode *inode, const char *symname, int len, int nofs)
L
Linus Torvalds 已提交
3345 3346
{
	struct address_space *mapping = inode->i_mapping;
3347
	struct page *page;
3348
	void *fsdata;
3349
	int err;
L
Linus Torvalds 已提交
3350
	char *kaddr;
3351 3352 3353
	unsigned int flags = AOP_FLAG_UNINTERRUPTIBLE;
	if (nofs)
		flags |= AOP_FLAG_NOFS;
L
Linus Torvalds 已提交
3354

3355
retry:
3356
	err = pagecache_write_begin(NULL, mapping, 0, len-1,
3357
				flags, &page, &fsdata);
L
Linus Torvalds 已提交
3358
	if (err)
3359 3360
		goto fail;

L
Linus Torvalds 已提交
3361 3362 3363
	kaddr = kmap_atomic(page, KM_USER0);
	memcpy(kaddr, symname, len-1);
	kunmap_atomic(kaddr, KM_USER0);
3364 3365 3366

	err = pagecache_write_end(NULL, mapping, 0, len-1, len-1,
							page, fsdata);
L
Linus Torvalds 已提交
3367 3368
	if (err < 0)
		goto fail;
3369 3370 3371
	if (err < len-1)
		goto retry;

L
Linus Torvalds 已提交
3372 3373 3374 3375 3376 3377
	mark_inode_dirty(inode);
	return 0;
fail:
	return err;
}

3378 3379 3380
int page_symlink(struct inode *inode, const char *symname, int len)
{
	return __page_symlink(inode, symname, len,
3381
			!(mapping_gfp_mask(inode->i_mapping) & __GFP_FS));
3382 3383
}

3384
const struct inode_operations page_symlink_inode_operations = {
L
Linus Torvalds 已提交
3385 3386 3387 3388 3389
	.readlink	= generic_readlink,
	.follow_link	= page_follow_link_light,
	.put_link	= page_put_link,
};

3390
EXPORT_SYMBOL(user_path_at);
3391
EXPORT_SYMBOL(follow_down_one);
L
Linus Torvalds 已提交
3392 3393 3394 3395 3396 3397 3398 3399 3400
EXPORT_SYMBOL(follow_down);
EXPORT_SYMBOL(follow_up);
EXPORT_SYMBOL(get_write_access); /* binfmt_aout */
EXPORT_SYMBOL(getname);
EXPORT_SYMBOL(lock_rename);
EXPORT_SYMBOL(lookup_one_len);
EXPORT_SYMBOL(page_follow_link_light);
EXPORT_SYMBOL(page_put_link);
EXPORT_SYMBOL(page_readlink);
3401
EXPORT_SYMBOL(__page_symlink);
L
Linus Torvalds 已提交
3402 3403
EXPORT_SYMBOL(page_symlink);
EXPORT_SYMBOL(page_symlink_inode_operations);
A
Al Viro 已提交
3404
EXPORT_SYMBOL(kern_path_parent);
A
Al Viro 已提交
3405
EXPORT_SYMBOL(kern_path);
3406
EXPORT_SYMBOL(vfs_path_lookup);
3407
EXPORT_SYMBOL(inode_permission);
3408
EXPORT_SYMBOL(file_permission);
L
Linus Torvalds 已提交
3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422
EXPORT_SYMBOL(unlock_rename);
EXPORT_SYMBOL(vfs_create);
EXPORT_SYMBOL(vfs_follow_link);
EXPORT_SYMBOL(vfs_link);
EXPORT_SYMBOL(vfs_mkdir);
EXPORT_SYMBOL(vfs_mknod);
EXPORT_SYMBOL(generic_permission);
EXPORT_SYMBOL(vfs_readlink);
EXPORT_SYMBOL(vfs_rename);
EXPORT_SYMBOL(vfs_rmdir);
EXPORT_SYMBOL(vfs_symlink);
EXPORT_SYMBOL(vfs_unlink);
EXPORT_SYMBOL(dentry_unhash);
EXPORT_SYMBOL(generic_readlink);