expfs.c 13.5 KB
Newer Older
1 2 3 4 5 6 7 8
/*
 * Copyright (C) Neil Brown 2002
 * Copyright (C) Christoph Hellwig 2007
 *
 * This file contains the code mapping from inodes to NFS file handles,
 * and for mapping back from file handles to dentries.
 *
 * For details on why we do all the strange and hairy things in here
9
 * take a look at Documentation/filesystems/nfs/Exporting.
10
 */
11
#include <linux/exportfs.h>
Linus Torvalds's avatar
Linus Torvalds committed
12 13 14
#include <linux/fs.h>
#include <linux/file.h>
#include <linux/module.h>
15
#include <linux/mount.h>
Linus Torvalds's avatar
Linus Torvalds committed
16
#include <linux/namei.h>
17
#include <linux/sched.h>
Linus Torvalds's avatar
Linus Torvalds committed
18

19
#define dprintk(fmt, args...) do{}while(0)
Linus Torvalds's avatar
Linus Torvalds committed
20 21


22
static int get_name(const struct path *path, char *name, struct dentry *child);
23 24


25 26
static int exportfs_get_name(struct vfsmount *mnt, struct dentry *dir,
		char *name, struct dentry *child)
27
{
28
	const struct export_operations *nop = dir->d_sb->s_export_op;
29
	struct path path = {.mnt = mnt, .dentry = dir};
30 31 32 33

	if (nop->get_name)
		return nop->get_name(dir, name, child);
	else
34
		return get_name(&path, name, child);
35
}
Linus Torvalds's avatar
Linus Torvalds committed
36

37 38 39
/*
 * Check if the dentry or any of it's aliases is acceptable.
 */
40 41 42 43 44 45
static struct dentry *
find_acceptable_alias(struct dentry *result,
		int (*acceptable)(void *context, struct dentry *dentry),
		void *context)
{
	struct dentry *dentry, *toput = NULL;
46
	struct inode *inode;
47

48 49 50
	if (acceptable(context, result))
		return result;

51 52
	inode = result->d_inode;
	spin_lock(&inode->i_lock);
53
	hlist_for_each_entry(dentry, &inode->i_dentry, d_alias) {
54
		dget(dentry);
55
		spin_unlock(&inode->i_lock);
56 57 58 59 60 61
		if (toput)
			dput(toput);
		if (dentry != result && acceptable(context, dentry)) {
			dput(result);
			return dentry;
		}
62
		spin_lock(&inode->i_lock);
63 64
		toput = dentry;
	}
65
	spin_unlock(&inode->i_lock);
66 67 68 69 70 71

	if (toput)
		dput(toput);
	return NULL;
}

72 73 74 75 76 77 78
/*
 * Find root of a disconnected subtree and return a reference to it.
 */
static struct dentry *
find_disconnected_root(struct dentry *dentry)
{
	dget(dentry);
79 80 81 82 83 84 85 86
	while (!IS_ROOT(dentry)) {
		struct dentry *parent = dget_parent(dentry);

		if (!(parent->d_flags & DCACHE_DISCONNECTED)) {
			dput(parent);
			break;
		}

87 88 89 90 91 92
		dput(dentry);
		dentry = parent;
	}
	return dentry;
}

93 94
/*
 * Make sure target_dir is fully connected to the dentry tree.
Linus Torvalds's avatar
Linus Torvalds committed
95
 *
96 97 98 99 100 101 102 103 104 105 106 107 108
 * On successful return, DCACHE_DISCONNECTED will be cleared on
 * target_dir, and target_dir->d_parent->...->d_parent will reach the
 * root of the filesystem.
 *
 * Whenever DCACHE_DISCONNECTED is unset, target_dir is fully connected.
 * But the converse is not true: target_dir may have DCACHE_DISCONNECTED
 * set but already be connected.  In that case we'll verify the
 * connection to root and then clear the flag.
 *
 * Note that target_dir could be removed by a concurrent operation.  In
 * that case reconnect_path may still succeed with target_dir fully
 * connected, but further operations using the filehandle will fail when
 * necessary (due to S_DEAD being set on the directory).
Linus Torvalds's avatar
Linus Torvalds committed
109
 */
110
static int
111
reconnect_path(struct vfsmount *mnt, struct dentry *target_dir, char *nbuf)
Linus Torvalds's avatar
Linus Torvalds committed
112
{
113 114
	int noprogress = 0;
	int err = -ESTALE;
Linus Torvalds's avatar
Linus Torvalds committed
115 116

	/*
117
	 * It is possible that a confused file system might not let us complete
Linus Torvalds's avatar
Linus Torvalds committed
118 119 120 121 122 123 124
	 * the path to the root.  For example, if get_parent returns a directory
	 * in which we cannot find a name for the child.  While this implies a
	 * very sick filesystem we don't want it to cause knfsd to spin.  Hence
	 * the noprogress counter.  If we go through the loop 10 times (2 is
	 * probably enough) without getting anywhere, we just give up
	 */
	while (target_dir->d_flags & DCACHE_DISCONNECTED && noprogress++ < 10) {
125
		struct dentry *pd = find_disconnected_root(target_dir);
Linus Torvalds's avatar
Linus Torvalds committed
126

127 128
		BUG_ON(pd == mnt->mnt_sb->s_root);

Linus Torvalds's avatar
Linus Torvalds committed
129 130 131 132 133 134 135
		if (!IS_ROOT(pd)) {
			/* must have found a connected parent - great */
			spin_lock(&pd->d_lock);
			pd->d_flags &= ~DCACHE_DISCONNECTED;
			spin_unlock(&pd->d_lock);
			noprogress = 0;
		} else {
136 137 138 139 140 141 142 143 144 145 146 147 148 149
			/*
			 * We have hit the top of a disconnected path, try to
			 * find parent and connect.
			 *
			 * Racing with some other process renaming a directory
			 * isn't much of a problem here.  If someone renames
			 * the directory, it will end up properly connected,
			 * which is what we want
			 *
			 * Getting the parent can't be supported generically,
			 * the locking is too icky.
			 *
			 * Instead we just return EACCES.  If server reboots
			 * or inodes get flushed, you lose
Linus Torvalds's avatar
Linus Torvalds committed
150
			 */
151
			struct dentry *ppd = ERR_PTR(-EACCES);
Linus Torvalds's avatar
Linus Torvalds committed
152 153
			struct dentry *npd;

154
			mutex_lock(&pd->d_inode->i_mutex);
155 156
			if (mnt->mnt_sb->s_export_op->get_parent)
				ppd = mnt->mnt_sb->s_export_op->get_parent(pd);
157
			mutex_unlock(&pd->d_inode->i_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
158 159 160

			if (IS_ERR(ppd)) {
				err = PTR_ERR(ppd);
161
				dprintk("%s: get_parent of %ld failed, err %d\n",
162
					__func__, pd->d_inode->i_ino, err);
Linus Torvalds's avatar
Linus Torvalds committed
163 164 165
				dput(pd);
				break;
			}
166

167
			dprintk("%s: find name of %lu in %lu\n", __func__,
168
				pd->d_inode->i_ino, ppd->d_inode->i_ino);
169
			err = exportfs_get_name(mnt, ppd, nbuf, pd);
Linus Torvalds's avatar
Linus Torvalds committed
170 171 172 173 174 175 176 177 178 179
			if (err) {
				dput(ppd);
				dput(pd);
				if (err == -ENOENT)
					/* some race between get_parent and
					 * get_name?  just try again
					 */
					continue;
				break;
			}
180
			dprintk("%s: found name: %s\n", __func__, nbuf);
181
			mutex_lock(&ppd->d_inode->i_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
182
			npd = lookup_one_len(nbuf, ppd, strlen(nbuf));
183
			mutex_unlock(&ppd->d_inode->i_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
184 185
			if (IS_ERR(npd)) {
				err = PTR_ERR(npd);
186
				dprintk("%s: lookup failed: %d\n",
187
					__func__, err);
Linus Torvalds's avatar
Linus Torvalds committed
188 189 190 191 192 193 194 195 196 197 198 199
				dput(ppd);
				dput(pd);
				break;
			}
			/* we didn't really want npd, we really wanted
			 * a side-effect of the lookup.
			 * hopefully, npd == pd, though it isn't really
			 * a problem if it isn't
			 */
			if (npd == pd)
				noprogress = 0;
			else
200
				printk("%s: npd != pd\n", __func__);
Linus Torvalds's avatar
Linus Torvalds committed
201 202 203 204 205 206 207 208 209 210 211 212 213 214 215
			dput(npd);
			dput(ppd);
			if (IS_ROOT(pd)) {
				/* something went wrong, we have to give up */
				dput(pd);
				break;
			}
		}
		dput(pd);
	}

	if (target_dir->d_flags & DCACHE_DISCONNECTED) {
		/* something went wrong - oh-well */
		if (!err)
			err = -ESTALE;
216
		return err;
Linus Torvalds's avatar
Linus Torvalds committed
217
	}
218 219 220 221

	return 0;
}

Linus Torvalds's avatar
Linus Torvalds committed
222
struct getdents_callback {
223
	struct dir_context ctx;
Linus Torvalds's avatar
Linus Torvalds committed
224 225
	char *name;		/* name that was found. It already points to a
				   buffer NAME_MAX+1 is size */
226
	u64 ino;		/* the inum we are looking for */
Linus Torvalds's avatar
Linus Torvalds committed
227 228 229 230 231 232 233 234 235
	int found;		/* inode matched? */
	int sequence;		/* sequence counter */
};

/*
 * A rather strange filldir function to capture
 * the name matching the specified inode number.
 */
static int filldir_one(void * __buf, const char * name, int len,
236
			loff_t pos, u64 ino, unsigned int d_type)
Linus Torvalds's avatar
Linus Torvalds committed
237 238 239 240 241
{
	struct getdents_callback *buf = __buf;
	int result = 0;

	buf->sequence++;
242
	if (buf->ino == ino && len <= NAME_MAX) {
Linus Torvalds's avatar
Linus Torvalds committed
243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259
		memcpy(buf->name, name, len);
		buf->name[len] = '\0';
		buf->found = 1;
		result = -1;
	}
	return result;
}

/**
 * get_name - default export_operations->get_name function
 * @dentry: the directory in which to find a name
 * @name:   a pointer to a %NAME_MAX+1 char buffer to store the name
 * @child:  the dentry for the child directory.
 *
 * calls readdir on the parent until it finds an entry with
 * the same inode number as the child, and returns that.
 */
260
static int get_name(const struct path *path, char *name, struct dentry *child)
Linus Torvalds's avatar
Linus Torvalds committed
261
{
262
	const struct cred *cred = current_cred();
263
	struct inode *dir = path->dentry->d_inode;
Linus Torvalds's avatar
Linus Torvalds committed
264 265
	int error;
	struct file *file;
266 267 268 269 270
	struct kstat stat;
	struct path child_path = {
		.mnt = path->mnt,
		.dentry = child,
	};
Al Viro's avatar
Al Viro committed
271 272 273 274
	struct getdents_callback buffer = {
		.ctx.actor = filldir_one,
		.name = name,
	};
Linus Torvalds's avatar
Linus Torvalds committed
275 276 277 278 279 280 281

	error = -ENOTDIR;
	if (!dir || !S_ISDIR(dir->i_mode))
		goto out;
	error = -EINVAL;
	if (!dir->i_fop)
		goto out;
282 283 284 285 286 287 288 289 290 291
	/*
	 * inode->i_ino is unsigned long, kstat->ino is u64, so the
	 * former would be insufficient on 32-bit hosts when the
	 * filesystem supports 64-bit inode numbers.  So we need to
	 * actually call ->getattr, not just read i_ino:
	 */
	error = vfs_getattr_nosec(&child_path, &stat);
	if (error)
		return error;
	buffer.ino = stat.ino;
Linus Torvalds's avatar
Linus Torvalds committed
292 293 294
	/*
	 * Open the directory ...
	 */
295
	file = dentry_open(path, O_RDONLY, cred);
Linus Torvalds's avatar
Linus Torvalds committed
296 297 298 299 300
	error = PTR_ERR(file);
	if (IS_ERR(file))
		goto out;

	error = -EINVAL;
Al Viro's avatar
Al Viro committed
301
	if (!file->f_op->iterate)
Linus Torvalds's avatar
Linus Torvalds committed
302 303 304 305 306 307
		goto out_close;

	buffer.sequence = 0;
	while (1) {
		int old_seq = buffer.sequence;

308
		error = iterate_dir(file, &buffer.ctx);
309 310 311 312
		if (buffer.found) {
			error = 0;
			break;
		}
Linus Torvalds's avatar
Linus Torvalds committed
313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329

		if (error < 0)
			break;

		error = -ENOENT;
		if (old_seq == buffer.sequence)
			break;
	}

out_close:
	fput(file);
out:
	return error;
}

/**
 * export_encode_fh - default export_operations->encode_fh function
Al Viro's avatar
Al Viro committed
330
 * @inode:   the object to encode
Linus Torvalds's avatar
Linus Torvalds committed
331 332
 * @fh:      where to store the file handle fragment
 * @max_len: maximum length to store there
Al Viro's avatar
Al Viro committed
333
 * @parent:  parent directory inode, if wanted
Linus Torvalds's avatar
Linus Torvalds committed
334 335 336 337 338 339
 *
 * This default encode_fh function assumes that the 32 inode number
 * is suitable for locating an inode, and that the generation number
 * can be used to check that it is still valid.  It places them in the
 * filehandle fragment where export_decode_fh expects to find them.
 */
Al Viro's avatar
Al Viro committed
340 341
static int export_encode_fh(struct inode *inode, struct fid *fid,
		int *max_len, struct inode *parent)
Linus Torvalds's avatar
Linus Torvalds committed
342 343
{
	int len = *max_len;
Christoph Hellwig's avatar
Christoph Hellwig committed
344
	int type = FILEID_INO32_GEN;
345

Al Viro's avatar
Al Viro committed
346
	if (parent && (len < 4)) {
347
		*max_len = 4;
348
		return FILEID_INVALID;
349 350
	} else if (len < 2) {
		*max_len = 2;
351
		return FILEID_INVALID;
352
	}
Linus Torvalds's avatar
Linus Torvalds committed
353 354

	len = 2;
Christoph Hellwig's avatar
Christoph Hellwig committed
355 356
	fid->i32.ino = inode->i_ino;
	fid->i32.gen = inode->i_generation;
Al Viro's avatar
Al Viro committed
357
	if (parent) {
Christoph Hellwig's avatar
Christoph Hellwig committed
358 359
		fid->i32.parent_ino = parent->i_ino;
		fid->i32.parent_gen = parent->i_generation;
Linus Torvalds's avatar
Linus Torvalds committed
360
		len = 4;
Christoph Hellwig's avatar
Christoph Hellwig committed
361
		type = FILEID_INO32_GEN_PARENT;
Linus Torvalds's avatar
Linus Torvalds committed
362 363 364 365 366
	}
	*max_len = len;
	return type;
}

367 368 369 370 371 372 373 374 375 376 377 378
int exportfs_encode_inode_fh(struct inode *inode, struct fid *fid,
			     int *max_len, struct inode *parent)
{
	const struct export_operations *nop = inode->i_sb->s_export_op;

	if (nop && nop->encode_fh)
		return nop->encode_fh(inode, fid->raw, max_len, parent);

	return export_encode_fh(inode, fid, max_len, parent);
}
EXPORT_SYMBOL_GPL(exportfs_encode_inode_fh);

Christoph Hellwig's avatar
Christoph Hellwig committed
379
int exportfs_encode_fh(struct dentry *dentry, struct fid *fid, int *max_len,
380 381
		int connectable)
{
382
	int error;
Al Viro's avatar
Al Viro committed
383 384
	struct dentry *p = NULL;
	struct inode *inode = dentry->d_inode, *parent = NULL;
385

Al Viro's avatar
Al Viro committed
386 387 388 389 390 391 392 393
	if (connectable && !S_ISDIR(inode->i_mode)) {
		p = dget_parent(dentry);
		/*
		 * note that while p might've ceased to be our parent already,
		 * it's still pinned by and still positive.
		 */
		parent = p->d_inode;
	}
394 395

	error = exportfs_encode_inode_fh(inode, fid, max_len, parent);
Al Viro's avatar
Al Viro committed
396
	dput(p);
397 398

	return error;
399 400 401
}
EXPORT_SYMBOL_GPL(exportfs_encode_fh);

Christoph Hellwig's avatar
Christoph Hellwig committed
402 403 404
struct dentry *exportfs_decode_fh(struct vfsmount *mnt, struct fid *fid,
		int fh_len, int fileid_type,
		int (*acceptable)(void *, struct dentry *), void *context)
405
{
406
	const struct export_operations *nop = mnt->mnt_sb->s_export_op;
407
	struct dentry *result, *alias;
408
	char nbuf[NAME_MAX+1];
409
	int err;
410

411 412 413
	/*
	 * Try to get any dentry for the given file handle from the filesystem.
	 */
414 415
	if (!nop || !nop->fh_to_dentry)
		return ERR_PTR(-ESTALE);
416
	result = nop->fh_to_dentry(mnt->mnt_sb, fid, fh_len, fileid_type);
417 418
	if (!result)
		result = ERR_PTR(-ESTALE);
419 420 421 422 423 424 425 426 427 428 429 430 431
	if (IS_ERR(result))
		return result;

	if (S_ISDIR(result->d_inode->i_mode)) {
		/*
		 * This request is for a directory.
		 *
		 * On the positive side there is only one dentry for each
		 * directory inode.  On the negative side this implies that we
		 * to ensure our dentry is connected all the way up to the
		 * filesystem root.
		 */
		if (result->d_flags & DCACHE_DISCONNECTED) {
432
			err = reconnect_path(mnt, result, nbuf);
433 434 435 436 437 438 439 440 441 442
			if (err)
				goto err_result;
		}

		if (!acceptable(context, result)) {
			err = -EACCES;
			goto err_result;
		}

		return result;
443
	} else {
444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471
		/*
		 * It's not a directory.  Life is a little more complicated.
		 */
		struct dentry *target_dir, *nresult;

		/*
		 * See if either the dentry we just got from the filesystem
		 * or any alias for it is acceptable.  This is always true
		 * if this filesystem is exported without the subtreecheck
		 * option.  If the filesystem is exported with the subtree
		 * check option there's a fair chance we need to look at
		 * the parent directory in the file handle and make sure
		 * it's connected to the filesystem root.
		 */
		alias = find_acceptable_alias(result, acceptable, context);
		if (alias)
			return alias;

		/*
		 * Try to extract a dentry for the parent directory from the
		 * file handle.  If this fails we'll have to give up.
		 */
		err = -ESTALE;
		if (!nop->fh_to_parent)
			goto err_result;

		target_dir = nop->fh_to_parent(mnt->mnt_sb, fid,
				fh_len, fileid_type);
472 473
		if (!target_dir)
			goto err_result;
474 475 476 477 478 479 480 481 482
		err = PTR_ERR(target_dir);
		if (IS_ERR(target_dir))
			goto err_result;

		/*
		 * And as usual we need to make sure the parent directory is
		 * connected to the filesystem root.  The VFS really doesn't
		 * like disconnected directories..
		 */
483
		err = reconnect_path(mnt, target_dir, nbuf);
484 485 486 487 488 489 490 491 492 493
		if (err) {
			dput(target_dir);
			goto err_result;
		}

		/*
		 * Now that we've got both a well-connected parent and a
		 * dentry for the inode we're after, make sure that our
		 * inode is actually connected to the parent.
		 */
494
		err = exportfs_get_name(mnt, target_dir, nbuf, result);
495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525
		if (!err) {
			mutex_lock(&target_dir->d_inode->i_mutex);
			nresult = lookup_one_len(nbuf, target_dir,
						 strlen(nbuf));
			mutex_unlock(&target_dir->d_inode->i_mutex);
			if (!IS_ERR(nresult)) {
				if (nresult->d_inode) {
					dput(result);
					result = nresult;
				} else
					dput(nresult);
			}
		}

		/*
		 * At this point we are done with the parent, but it's pinned
		 * by the child dentry anyway.
		 */
		dput(target_dir);

		/*
		 * And finally make sure the dentry is actually acceptable
		 * to NFSD.
		 */
		alias = find_acceptable_alias(result, acceptable, context);
		if (!alias) {
			err = -EACCES;
			goto err_result;
		}

		return alias;
526 527
	}

528 529 530
 err_result:
	dput(result);
	return ERR_PTR(err);
531 532 533
}
EXPORT_SYMBOL_GPL(exportfs_decode_fh);

Linus Torvalds's avatar
Linus Torvalds committed
534
MODULE_LICENSE("GPL");