dir.c 20.1 KB
Newer Older
L
Linus Torvalds 已提交
1
/*
T
Tejun Heo 已提交
2 3 4 5 6 7 8 9 10
 * fs/sysfs/dir.c - sysfs core and dir operation implementation
 *
 * Copyright (c) 2001-3 Patrick Mochel
 * Copyright (c) 2007 SUSE Linux Products GmbH
 * Copyright (c) 2007 Tejun Heo <teheo@suse.de>
 *
 * This file is released under the GPLv2.
 *
 * Please see Documentation/filesystems/sysfs.txt for more information.
L
Linus Torvalds 已提交
11 12 13 14 15 16 17 18
 */

#undef DEBUG

#include <linux/fs.h>
#include <linux/mount.h>
#include <linux/module.h>
#include <linux/kobject.h>
19
#include <linux/namei.h>
20
#include <linux/idr.h>
21
#include <linux/completion.h>
D
Dave Young 已提交
22
#include <linux/mutex.h>
23
#include <linux/slab.h>
24
#include <linux/security.h>
L
Linus Torvalds 已提交
25 26
#include "sysfs.h"

27
DEFINE_MUTEX(sysfs_mutex);
R
Roel Kluin 已提交
28
DEFINE_SPINLOCK(sysfs_assoc_lock);
L
Linus Torvalds 已提交
29

R
Roel Kluin 已提交
30
static DEFINE_SPINLOCK(sysfs_ino_lock);
31 32
static DEFINE_IDA(sysfs_ino_ida);

33 34 35 36 37
/**
 *	sysfs_link_sibling - link sysfs_dirent into sibling list
 *	@sd: sysfs_dirent of interest
 *
 *	Link @sd into its sibling list which starts from
38
 *	sd->s_parent->s_dir.children.
39 40
 *
 *	Locking:
41
 *	mutex_lock(sysfs_mutex)
42
 */
43
static void sysfs_link_sibling(struct sysfs_dirent *sd)
44 45
{
	struct sysfs_dirent *parent_sd = sd->s_parent;
E
Eric W. Biederman 已提交
46
	struct sysfs_dirent **pos;
47 48

	BUG_ON(sd->s_sibling);
E
Eric W. Biederman 已提交
49 50 51

	/* Store directory entries in order by ino.  This allows
	 * readdir to properly restart without having to add a
52
	 * cursor into the s_dir.children list.
E
Eric W. Biederman 已提交
53
	 */
54
	for (pos = &parent_sd->s_dir.children; *pos; pos = &(*pos)->s_sibling) {
E
Eric W. Biederman 已提交
55 56 57 58 59
		if (sd->s_ino < (*pos)->s_ino)
			break;
	}
	sd->s_sibling = *pos;
	*pos = sd;
60 61 62 63 64 65 66
}

/**
 *	sysfs_unlink_sibling - unlink sysfs_dirent from sibling list
 *	@sd: sysfs_dirent of interest
 *
 *	Unlink @sd from its sibling list which starts from
67
 *	sd->s_parent->s_dir.children.
68 69
 *
 *	Locking:
70
 *	mutex_lock(sysfs_mutex)
71
 */
72
static void sysfs_unlink_sibling(struct sysfs_dirent *sd)
73 74 75
{
	struct sysfs_dirent **pos;

76 77
	for (pos = &sd->s_parent->s_dir.children; *pos;
	     pos = &(*pos)->s_sibling) {
78 79 80 81 82 83 84 85
		if (*pos == sd) {
			*pos = sd->s_sibling;
			sd->s_sibling = NULL;
			break;
		}
	}
}

86 87 88 89 90 91 92 93 94 95
/**
 *	sysfs_get_active - get an active reference to sysfs_dirent
 *	@sd: sysfs_dirent to get an active reference to
 *
 *	Get an active reference of @sd.  This function is noop if @sd
 *	is NULL.
 *
 *	RETURNS:
 *	Pointer to @sd on success, NULL on failure.
 */
96
static struct sysfs_dirent *sysfs_get_active(struct sysfs_dirent *sd)
97
{
98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114
	if (unlikely(!sd))
		return NULL;

	while (1) {
		int v, t;

		v = atomic_read(&sd->s_active);
		if (unlikely(v < 0))
			return NULL;

		t = atomic_cmpxchg(&sd->s_active, v, v + 1);
		if (likely(t == v))
			return sd;
		if (t < 0)
			return NULL;

		cpu_relax();
115 116 117 118 119 120 121 122 123 124
	}
}

/**
 *	sysfs_put_active - put an active reference to sysfs_dirent
 *	@sd: sysfs_dirent to put an active reference to
 *
 *	Put an active reference to @sd.  This function is noop if @sd
 *	is NULL.
 */
125
static void sysfs_put_active(struct sysfs_dirent *sd)
126
{
127 128 129 130 131 132 133 134 135 136 137
	struct completion *cmpl;
	int v;

	if (unlikely(!sd))
		return;

	v = atomic_dec_return(&sd->s_active);
	if (likely(v != SD_DEACTIVATED_BIAS))
		return;

	/* atomic_dec_return() is a mb(), we'll always see the updated
138
	 * sd->s_sibling.
139
	 */
140
	cmpl = (void *)sd->s_sibling;
141
	complete(cmpl);
142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186
}

/**
 *	sysfs_get_active_two - get active references to sysfs_dirent and parent
 *	@sd: sysfs_dirent of interest
 *
 *	Get active reference to @sd and its parent.  Parent's active
 *	reference is grabbed first.  This function is noop if @sd is
 *	NULL.
 *
 *	RETURNS:
 *	Pointer to @sd on success, NULL on failure.
 */
struct sysfs_dirent *sysfs_get_active_two(struct sysfs_dirent *sd)
{
	if (sd) {
		if (sd->s_parent && unlikely(!sysfs_get_active(sd->s_parent)))
			return NULL;
		if (unlikely(!sysfs_get_active(sd))) {
			sysfs_put_active(sd->s_parent);
			return NULL;
		}
	}
	return sd;
}

/**
 *	sysfs_put_active_two - put active references to sysfs_dirent and parent
 *	@sd: sysfs_dirent of interest
 *
 *	Put active references to @sd and its parent.  This function is
 *	noop if @sd is NULL.
 */
void sysfs_put_active_two(struct sysfs_dirent *sd)
{
	if (sd) {
		sysfs_put_active(sd);
		sysfs_put_active(sd->s_parent);
	}
}

/**
 *	sysfs_deactivate - deactivate sysfs_dirent
 *	@sd: sysfs_dirent to deactivate
 *
187
 *	Deny new active references and drain existing ones.
188
 */
189
static void sysfs_deactivate(struct sysfs_dirent *sd)
190
{
191 192
	DECLARE_COMPLETION_ONSTACK(wait);
	int v;
193

194
	BUG_ON(sd->s_sibling || !(sd->s_flags & SYSFS_FLAG_REMOVED));
195
	sd->s_sibling = (void *)&wait;
196 197

	/* atomic_add_return() is a mb(), put_active() will always see
198
	 * the updated sd->s_sibling.
199
	 */
200 201 202 203 204
	v = atomic_add_return(SD_DEACTIVATED_BIAS, &sd->s_active);

	if (v != SD_DEACTIVATED_BIAS)
		wait_for_completion(&wait);

205
	sd->s_sibling = NULL;
206 207
}

T
Tejun Heo 已提交
208
static int sysfs_alloc_ino(ino_t *pino)
209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233
{
	int ino, rc;

 retry:
	spin_lock(&sysfs_ino_lock);
	rc = ida_get_new_above(&sysfs_ino_ida, 2, &ino);
	spin_unlock(&sysfs_ino_lock);

	if (rc == -EAGAIN) {
		if (ida_pre_get(&sysfs_ino_ida, GFP_KERNEL))
			goto retry;
		rc = -ENOMEM;
	}

	*pino = ino;
	return rc;
}

static void sysfs_free_ino(ino_t ino)
{
	spin_lock(&sysfs_ino_lock);
	ida_remove(&sysfs_ino_ida, ino);
	spin_unlock(&sysfs_ino_lock);
}

234 235
void release_sysfs_dirent(struct sysfs_dirent * sd)
{
T
Tejun Heo 已提交
236 237 238
	struct sysfs_dirent *parent_sd;

 repeat:
239 240 241
	/* Moving/renaming is always done while holding reference.
	 * sd->s_parent won't change beneath us.
	 */
T
Tejun Heo 已提交
242 243
	parent_sd = sd->s_parent;

244
	if (sysfs_type(sd) == SYSFS_KOBJ_LINK)
T
Tejun Heo 已提交
245
		sysfs_put(sd->s_symlink.target_sd);
246
	if (sysfs_type(sd) & SYSFS_COPY_NAME)
T
Tejun Heo 已提交
247
		kfree(sd->s_name);
248 249 250
	if (sd->s_iattr && sd->s_iattr->ia_secdata)
		security_release_secctx(sd->s_iattr->ia_secdata,
					sd->s_iattr->ia_secdata_len);
251
	kfree(sd->s_iattr);
252
	sysfs_free_ino(sd->s_ino);
253
	kmem_cache_free(sysfs_dir_cachep, sd);
T
Tejun Heo 已提交
254 255 256 257

	sd = parent_sd;
	if (sd && atomic_dec_and_test(&sd->s_count))
		goto repeat;
258 259
}

260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276
static int sysfs_dentry_delete(struct dentry *dentry)
{
	struct sysfs_dirent *sd = dentry->d_fsdata;
	return !!(sd->s_flags & SYSFS_FLAG_REMOVED);
}

static int sysfs_dentry_revalidate(struct dentry *dentry, struct nameidata *nd)
{
	struct sysfs_dirent *sd = dentry->d_fsdata;
	int is_dir;

	mutex_lock(&sysfs_mutex);

	/* The sysfs dirent has been deleted */
	if (sd->s_flags & SYSFS_FLAG_REMOVED)
		goto out_bad;

277 278 279 280 281 282 283 284
	/* The sysfs dirent has been moved? */
	if (dentry->d_parent->d_fsdata != sd->s_parent)
		goto out_bad;

	/* The sysfs dirent has been renamed */
	if (strcmp(dentry->d_name.name, sd->s_name) != 0)
		goto out_bad;

285 286 287 288 289 290 291
	mutex_unlock(&sysfs_mutex);
out_valid:
	return 1;
out_bad:
	/* Remove the dentry from the dcache hashes.
	 * If this is a deleted dentry we use d_drop instead of d_delete
	 * so sysfs doesn't need to cope with negative dentries.
292 293 294 295 296 297
	 *
	 * If this is a dentry that has simply been renamed we
	 * use d_drop to remove it from the dcache lookup on its
	 * old parent.  If this dentry persists later when a lookup
	 * is performed at its new name the dentry will be readded
	 * to the dcache hashes.
298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313
	 */
	is_dir = (sysfs_type(sd) == SYSFS_DIR);
	mutex_unlock(&sysfs_mutex);
	if (is_dir) {
		/* If we have submounts we must allow the vfs caches
		 * to lie about the state of the filesystem to prevent
		 * leaks and other nasty things.
		 */
		if (have_submounts(dentry))
			goto out_valid;
		shrink_dcache_parent(dentry);
	}
	d_drop(dentry);
	return 0;
}

314
static void sysfs_dentry_iput(struct dentry *dentry, struct inode *inode)
L
Linus Torvalds 已提交
315 316 317
{
	struct sysfs_dirent * sd = dentry->d_fsdata;

E
Eric W. Biederman 已提交
318
	sysfs_put(sd);
L
Linus Torvalds 已提交
319 320 321
	iput(inode);
}

A
Al Viro 已提交
322
static const struct dentry_operations sysfs_dentry_ops = {
323 324
	.d_revalidate	= sysfs_dentry_revalidate,
	.d_delete	= sysfs_dentry_delete,
325
	.d_iput		= sysfs_dentry_iput,
L
Linus Torvalds 已提交
326 327
};

328
struct sysfs_dirent *sysfs_new_dirent(const char *name, umode_t mode, int type)
L
Linus Torvalds 已提交
329
{
T
Tejun Heo 已提交
330
	char *dup_name = NULL;
331
	struct sysfs_dirent *sd;
T
Tejun Heo 已提交
332 333 334 335

	if (type & SYSFS_COPY_NAME) {
		name = dup_name = kstrdup(name, GFP_KERNEL);
		if (!name)
336
			return NULL;
T
Tejun Heo 已提交
337
	}
L
Linus Torvalds 已提交
338

339
	sd = kmem_cache_zalloc(sysfs_dir_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
340
	if (!sd)
341
		goto err_out1;
L
Linus Torvalds 已提交
342

T
Tejun Heo 已提交
343
	if (sysfs_alloc_ino(&sd->s_ino))
344
		goto err_out2;
345

L
Linus Torvalds 已提交
346
	atomic_set(&sd->s_count, 1);
347
	atomic_set(&sd->s_active, 0);
348

T
Tejun Heo 已提交
349
	sd->s_name = name;
350
	sd->s_mode = mode;
351
	sd->s_flags = type;
L
Linus Torvalds 已提交
352 353

	return sd;
T
Tejun Heo 已提交
354

355
 err_out2:
T
Tejun Heo 已提交
356
	kmem_cache_free(sysfs_dir_cachep, sd);
357 358
 err_out1:
	kfree(dup_name);
T
Tejun Heo 已提交
359
	return NULL;
L
Linus Torvalds 已提交
360 361
}

362
/**
363 364 365
 *	sysfs_addrm_start - prepare for sysfs_dirent add/remove
 *	@acxt: pointer to sysfs_addrm_cxt to be used
 *	@parent_sd: parent sysfs_dirent
366
 *
367 368
 *	This function is called when the caller is about to add or
 *	remove sysfs_dirent under @parent_sd.  This function acquires
369
 *	sysfs_mutex.  @acxt is used to keep and pass context to
370
 *	other addrm functions.
371 372
 *
 *	LOCKING:
373
 *	Kernel thread context (may sleep).  sysfs_mutex is locked on
374
 *	return.
375
 */
376 377
void sysfs_addrm_start(struct sysfs_addrm_cxt *acxt,
		       struct sysfs_dirent *parent_sd)
378
{
379 380 381 382 383 384 385
	memset(acxt, 0, sizeof(*acxt));
	acxt->parent_sd = parent_sd;

	mutex_lock(&sysfs_mutex);
}

/**
386
 *	__sysfs_add_one - add sysfs_dirent to parent without warning
387 388 389 390
 *	@acxt: addrm context to use
 *	@sd: sysfs_dirent to be added
 *
 *	Get @acxt->parent_sd and set sd->s_parent to it and increment
391 392
 *	nlink of parent inode if @sd is a directory and link into the
 *	children list of the parent.
393 394 395 396 397 398 399
 *
 *	This function should be called between calls to
 *	sysfs_addrm_start() and sysfs_addrm_finish() and should be
 *	passed the same @acxt as passed to sysfs_addrm_start().
 *
 *	LOCKING:
 *	Determined by sysfs_addrm_start().
400 401 402 403
 *
 *	RETURNS:
 *	0 on success, -EEXIST if entry with the given name already
 *	exists.
404
 */
405
int __sysfs_add_one(struct sysfs_addrm_cxt *acxt, struct sysfs_dirent *sd)
406
{
407 408
	struct sysfs_inode_attrs *ps_iattr;

409
	if (sysfs_find_dirent(acxt->parent_sd, sd->s_name))
410 411
		return -EEXIST;

412 413
	sd->s_parent = sysfs_get(acxt->parent_sd);

414
	sysfs_link_sibling(sd);
415

416 417 418 419 420 421 422
	/* Update timestamps on the parent */
	ps_iattr = acxt->parent_sd->s_iattr;
	if (ps_iattr) {
		struct iattr *ps_iattrs = &ps_iattr->ia_iattr;
		ps_iattrs->ia_ctime = ps_iattrs->ia_mtime = CURRENT_TIME;
	}

423
	return 0;
424 425
}

426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445
/**
 *	sysfs_pathname - return full path to sysfs dirent
 *	@sd: sysfs_dirent whose path we want
 *	@path: caller allocated buffer
 *
 *	Gives the name "/" to the sysfs_root entry; any path returned
 *	is relative to wherever sysfs is mounted.
 *
 *	XXX: does no error checking on @path size
 */
static char *sysfs_pathname(struct sysfs_dirent *sd, char *path)
{
	if (sd->s_parent) {
		sysfs_pathname(sd->s_parent, path);
		strcat(path, "/");
	}
	strcat(path, sd->s_name);
	return path;
}

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
/**
 *	sysfs_add_one - add sysfs_dirent to parent
 *	@acxt: addrm context to use
 *	@sd: sysfs_dirent to be added
 *
 *	Get @acxt->parent_sd and set sd->s_parent to it and increment
 *	nlink of parent inode if @sd is a directory and link into the
 *	children list of the parent.
 *
 *	This function should be called between calls to
 *	sysfs_addrm_start() and sysfs_addrm_finish() and should be
 *	passed the same @acxt as passed to sysfs_addrm_start().
 *
 *	LOCKING:
 *	Determined by sysfs_addrm_start().
 *
 *	RETURNS:
 *	0 on success, -EEXIST if entry with the given name already
 *	exists.
 */
int sysfs_add_one(struct sysfs_addrm_cxt *acxt, struct sysfs_dirent *sd)
{
	int ret;

	ret = __sysfs_add_one(acxt, sd);
471 472 473 474 475 476 477 478 479 480
	if (ret == -EEXIST) {
		char *path = kzalloc(PATH_MAX, GFP_KERNEL);
		WARN(1, KERN_WARNING
		     "sysfs: cannot create duplicate filename '%s'\n",
		     (path == NULL) ? sd->s_name :
		     strcat(strcat(sysfs_pathname(acxt->parent_sd, path), "/"),
		            sd->s_name));
		kfree(path);
	}

481 482 483
	return ret;
}

484 485 486
/**
 *	sysfs_remove_one - remove sysfs_dirent from parent
 *	@acxt: addrm context to use
487
 *	@sd: sysfs_dirent to be removed
488 489
 *
 *	Mark @sd removed and drop nlink of parent inode if @sd is a
490
 *	directory.  @sd is unlinked from the children list.
491 492 493 494 495 496 497 498 499 500
 *
 *	This function should be called between calls to
 *	sysfs_addrm_start() and sysfs_addrm_finish() and should be
 *	passed the same @acxt as passed to sysfs_addrm_start().
 *
 *	LOCKING:
 *	Determined by sysfs_addrm_start().
 */
void sysfs_remove_one(struct sysfs_addrm_cxt *acxt, struct sysfs_dirent *sd)
{
501 502
	struct sysfs_inode_attrs *ps_iattr;

503 504 505
	BUG_ON(sd->s_flags & SYSFS_FLAG_REMOVED);

	sysfs_unlink_sibling(sd);
506

507 508 509 510 511 512 513
	/* Update timestamps on the parent */
	ps_iattr = acxt->parent_sd->s_iattr;
	if (ps_iattr) {
		struct iattr *ps_iattrs = &ps_iattr->ia_iattr;
		ps_iattrs->ia_ctime = ps_iattrs->ia_mtime = CURRENT_TIME;
	}

514 515 516
	sd->s_flags |= SYSFS_FLAG_REMOVED;
	sd->s_sibling = acxt->removed;
	acxt->removed = sd;
517 518
}

519 520 521 522 523 524
/**
 *	sysfs_addrm_finish - finish up sysfs_dirent add/remove
 *	@acxt: addrm context to finish up
 *
 *	Finish up sysfs_dirent add/remove.  Resources acquired by
 *	sysfs_addrm_start() are released and removed sysfs_dirents are
525
 *	cleaned up.
526 527
 *
 *	LOCKING:
528
 *	sysfs_mutex is released.
529
 */
530
void sysfs_addrm_finish(struct sysfs_addrm_cxt *acxt)
531 532 533 534 535 536 537 538 539 540 541 542
{
	/* release resources acquired by sysfs_addrm_start() */
	mutex_unlock(&sysfs_mutex);

	/* kill removed sysfs_dirents */
	while (acxt->removed) {
		struct sysfs_dirent *sd = acxt->removed;

		acxt->removed = sd->s_sibling;
		sd->s_sibling = NULL;

		sysfs_deactivate(sd);
543
		unmap_bin_file(sd);
544
		sysfs_put(sd);
T
Tejun Heo 已提交
545
	}
546 547
}

548 549 550 551 552 553
/**
 *	sysfs_find_dirent - find sysfs_dirent with the given name
 *	@parent_sd: sysfs_dirent to search under
 *	@name: name to look for
 *
 *	Look for sysfs_dirent with name @name under @parent_sd.
554
 *
555
 *	LOCKING:
556
 *	mutex_lock(sysfs_mutex)
557
 *
558 559
 *	RETURNS:
 *	Pointer to sysfs_dirent if found, NULL if not.
560
 */
561 562
struct sysfs_dirent *sysfs_find_dirent(struct sysfs_dirent *parent_sd,
				       const unsigned char *name)
563
{
564 565
	struct sysfs_dirent *sd;

566
	for (sd = parent_sd->s_dir.children; sd; sd = sd->s_sibling)
E
Eric W. Biederman 已提交
567
		if (!strcmp(sd->s_name, name))
568 569 570
			return sd;
	return NULL;
}
571

572 573 574 575 576 577 578 579 580
/**
 *	sysfs_get_dirent - find and get sysfs_dirent with the given name
 *	@parent_sd: sysfs_dirent to search under
 *	@name: name to look for
 *
 *	Look for sysfs_dirent with name @name under @parent_sd and get
 *	it if found.
 *
 *	LOCKING:
581
 *	Kernel thread context (may sleep).  Grabs sysfs_mutex.
582 583 584 585 586 587 588 589 590
 *
 *	RETURNS:
 *	Pointer to sysfs_dirent if found, NULL if not.
 */
struct sysfs_dirent *sysfs_get_dirent(struct sysfs_dirent *parent_sd,
				      const unsigned char *name)
{
	struct sysfs_dirent *sd;

591
	mutex_lock(&sysfs_mutex);
592 593
	sd = sysfs_find_dirent(parent_sd, name);
	sysfs_get(sd);
594
	mutex_unlock(&sysfs_mutex);
595 596

	return sd;
597
}
598
EXPORT_SYMBOL_GPL(sysfs_get_dirent);
599

600 601
static int create_dir(struct kobject *kobj, struct sysfs_dirent *parent_sd,
		      const char *name, struct sysfs_dirent **p_sd)
L
Linus Torvalds 已提交
602 603
{
	umode_t mode = S_IFDIR| S_IRWXU | S_IRUGO | S_IXUGO;
604
	struct sysfs_addrm_cxt acxt;
605
	struct sysfs_dirent *sd;
606
	int rc;
L
Linus Torvalds 已提交
607

608
	/* allocate */
609
	sd = sysfs_new_dirent(name, mode, SYSFS_DIR);
610
	if (!sd)
611
		return -ENOMEM;
T
Tejun Heo 已提交
612
	sd->s_dir.kobj = kobj;
613

614
	/* link in */
615
	sysfs_addrm_start(&acxt, parent_sd);
616 617
	rc = sysfs_add_one(&acxt, sd);
	sysfs_addrm_finish(&acxt);
618

619 620 621
	if (rc == 0)
		*p_sd = sd;
	else
622
		sysfs_put(sd);
623

624
	return rc;
L
Linus Torvalds 已提交
625 626
}

627 628
int sysfs_create_subdir(struct kobject *kobj, const char *name,
			struct sysfs_dirent **p_sd)
L
Linus Torvalds 已提交
629
{
630
	return create_dir(kobj, kobj->sd, name, p_sd);
L
Linus Torvalds 已提交
631 632 633 634 635 636
}

/**
 *	sysfs_create_dir - create a directory for an object.
 *	@kobj:		object we're creating directory for. 
 */
637
int sysfs_create_dir(struct kobject * kobj)
L
Linus Torvalds 已提交
638
{
639
	struct sysfs_dirent *parent_sd, *sd;
L
Linus Torvalds 已提交
640 641 642 643
	int error = 0;

	BUG_ON(!kobj);

644
	if (kobj->parent)
645
		parent_sd = kobj->parent->sd;
L
Linus Torvalds 已提交
646
	else
E
Eric W. Biederman 已提交
647
		parent_sd = &sysfs_root;
L
Linus Torvalds 已提交
648

649
	error = create_dir(kobj, parent_sd, kobject_name(kobj), &sd);
L
Linus Torvalds 已提交
650
	if (!error)
651
		kobj->sd = sd;
L
Linus Torvalds 已提交
652 653 654 655 656 657
	return error;
}

static struct dentry * sysfs_lookup(struct inode *dir, struct dentry *dentry,
				struct nameidata *nd)
{
658
	struct dentry *ret = NULL;
659 660
	struct sysfs_dirent *parent_sd = dentry->d_parent->d_fsdata;
	struct sysfs_dirent *sd;
661
	struct inode *inode;
L
Linus Torvalds 已提交
662

663 664
	mutex_lock(&sysfs_mutex);

665
	sd = sysfs_find_dirent(parent_sd, dentry->d_name.name);
L
Linus Torvalds 已提交
666

667
	/* no such entry */
668 669
	if (!sd) {
		ret = ERR_PTR(-ENOENT);
670
		goto out_unlock;
671
	}
672 673

	/* attach dentry and inode */
674
	inode = sysfs_get_inode(sd);
675 676 677 678
	if (!inode) {
		ret = ERR_PTR(-ENOMEM);
		goto out_unlock;
	}
679

T
Tejun Heo 已提交
680
	/* instantiate and hash dentry */
681 682 683 684 685 686 687 688 689
	ret = d_find_alias(inode);
	if (!ret) {
		dentry->d_op = &sysfs_dentry_ops;
		dentry->d_fsdata = sysfs_get(sd);
		d_add(dentry, inode);
	} else {
		d_move(ret, dentry);
		iput(inode);
	}
690

691
 out_unlock:
692
	mutex_unlock(&sysfs_mutex);
693
	return ret;
L
Linus Torvalds 已提交
694 695
}

696
const struct inode_operations sysfs_dir_inode_operations = {
L
Linus Torvalds 已提交
697
	.lookup		= sysfs_lookup,
698
	.permission	= sysfs_permission,
699
	.setattr	= sysfs_setattr,
700
	.getattr	= sysfs_getattr,
701
	.setxattr	= sysfs_setxattr,
L
Linus Torvalds 已提交
702 703
};

704
static void remove_dir(struct sysfs_dirent *sd)
L
Linus Torvalds 已提交
705
{
706
	struct sysfs_addrm_cxt acxt;
L
Linus Torvalds 已提交
707

708 709 710
	sysfs_addrm_start(&acxt, sd->s_parent);
	sysfs_remove_one(&acxt, sd);
	sysfs_addrm_finish(&acxt);
L
Linus Torvalds 已提交
711 712
}

713
void sysfs_remove_subdir(struct sysfs_dirent *sd)
L
Linus Torvalds 已提交
714
{
715
	remove_dir(sd);
L
Linus Torvalds 已提交
716 717 718
}


719
static void __sysfs_remove_dir(struct sysfs_dirent *dir_sd)
L
Linus Torvalds 已提交
720
{
721
	struct sysfs_addrm_cxt acxt;
722
	struct sysfs_dirent **pos;
L
Linus Torvalds 已提交
723

724
	if (!dir_sd)
L
Linus Torvalds 已提交
725 726
		return;

727
	pr_debug("sysfs %s: removing dir\n", dir_sd->s_name);
728
	sysfs_addrm_start(&acxt, dir_sd);
729
	pos = &dir_sd->s_dir.children;
730 731 732
	while (*pos) {
		struct sysfs_dirent *sd = *pos;

E
Eric W. Biederman 已提交
733
		if (sysfs_type(sd) != SYSFS_DIR)
734
			sysfs_remove_one(&acxt, sd);
735
		else
736
			pos = &(*pos)->s_sibling;
L
Linus Torvalds 已提交
737
	}
738
	sysfs_addrm_finish(&acxt);
739

740
	remove_dir(dir_sd);
741 742 743 744 745 746 747 748 749 750 751 752 753
}

/**
 *	sysfs_remove_dir - remove an object's directory.
 *	@kobj:	object.
 *
 *	The only thing special about this is that we remove any files in
 *	the directory before we remove the directory, and we've inlined
 *	what used to be sysfs_rmdir() below, instead of calling separately.
 */

void sysfs_remove_dir(struct kobject * kobj)
{
754
	struct sysfs_dirent *sd = kobj->sd;
755

T
Tejun Heo 已提交
756
	spin_lock(&sysfs_assoc_lock);
757
	kobj->sd = NULL;
T
Tejun Heo 已提交
758
	spin_unlock(&sysfs_assoc_lock);
759

760
	__sysfs_remove_dir(sd);
L
Linus Torvalds 已提交
761 762
}

763 764
int sysfs_rename(struct sysfs_dirent *sd,
	struct sysfs_dirent *new_parent_sd, const char *new_name)
L
Linus Torvalds 已提交
765
{
766
	const char *dup_name = NULL;
767
	int error;
L
Linus Torvalds 已提交
768

769
	mutex_lock(&sysfs_mutex);
770

771
	error = 0;
772 773
	if ((sd->s_parent == new_parent_sd) &&
	    (strcmp(sd->s_name, new_name) == 0))
774 775 776
		goto out;	/* nothing to rename */

	error = -EEXIST;
777
	if (sysfs_find_dirent(new_parent_sd, new_name))
778
		goto out;
779

780
	/* rename sysfs_dirent */
781 782 783 784 785 786 787 788 789
	if (strcmp(sd->s_name, new_name) != 0) {
		error = -ENOMEM;
		new_name = dup_name = kstrdup(new_name, GFP_KERNEL);
		if (!new_name)
			goto out;

		dup_name = sd->s_name;
		sd->s_name = new_name;
	}
T
Tejun Heo 已提交
790

791 792 793 794 795 796 797 798
	/* Remove from old parent's list and insert into new parent's list. */
	if (sd->s_parent != new_parent_sd) {
		sysfs_unlink_sibling(sd);
		sysfs_get(new_parent_sd);
		sysfs_put(sd->s_parent);
		sd->s_parent = new_parent_sd;
		sysfs_link_sibling(sd);
	}
T
Tejun Heo 已提交
799

800
	error = 0;
801
 out:
802
	mutex_unlock(&sysfs_mutex);
803
	kfree(dup_name);
L
Linus Torvalds 已提交
804 805 806
	return error;
}

807 808 809 810 811
int sysfs_rename_dir(struct kobject *kobj, const char *new_name)
{
	return sysfs_rename(kobj->sd, kobj->sd->s_parent, new_name);
}

812
int sysfs_move_dir(struct kobject *kobj, struct kobject *new_parent_kobj)
813
{
814 815
	struct sysfs_dirent *sd = kobj->sd;
	struct sysfs_dirent *new_parent_sd;
816

817
	BUG_ON(!sd->s_parent);
818
	new_parent_sd = new_parent_kobj && new_parent_kobj->sd ?
819
		new_parent_kobj->sd : &sysfs_root;
820

821
	return sysfs_rename(sd, new_parent_sd, sd->s_name);
822 823
}

L
Linus Torvalds 已提交
824 825 826 827 828 829 830 831
/* Relationship between s_mode and the DT_xxx types */
static inline unsigned char dt_type(struct sysfs_dirent *sd)
{
	return (sd->s_mode >> 12) & 15;
}

static int sysfs_readdir(struct file * filp, void * dirent, filldir_t filldir)
{
832
	struct dentry *dentry = filp->f_path.dentry;
L
Linus Torvalds 已提交
833
	struct sysfs_dirent * parent_sd = dentry->d_fsdata;
E
Eric W. Biederman 已提交
834
	struct sysfs_dirent *pos;
L
Linus Torvalds 已提交
835 836
	ino_t ino;

E
Eric W. Biederman 已提交
837 838 839
	if (filp->f_pos == 0) {
		ino = parent_sd->s_ino;
		if (filldir(dirent, ".", 1, filp->f_pos, ino, DT_DIR) == 0)
L
Linus Torvalds 已提交
840
			filp->f_pos++;
E
Eric W. Biederman 已提交
841 842 843 844 845 846 847
	}
	if (filp->f_pos == 1) {
		if (parent_sd->s_parent)
			ino = parent_sd->s_parent->s_ino;
		else
			ino = parent_sd->s_ino;
		if (filldir(dirent, "..", 2, filp->f_pos, ino, DT_DIR) == 0)
L
Linus Torvalds 已提交
848
			filp->f_pos++;
E
Eric W. Biederman 已提交
849 850 851
	}
	if ((filp->f_pos > 1) && (filp->f_pos < INT_MAX)) {
		mutex_lock(&sysfs_mutex);
852

E
Eric W. Biederman 已提交
853
		/* Skip the dentries we have already reported */
854
		pos = parent_sd->s_dir.children;
E
Eric W. Biederman 已提交
855 856
		while (pos && (filp->f_pos > pos->s_ino))
			pos = pos->s_sibling;
857

E
Eric W. Biederman 已提交
858 859 860
		for ( ; pos; pos = pos->s_sibling) {
			const char * name;
			int len;
L
Linus Torvalds 已提交
861

E
Eric W. Biederman 已提交
862 863 864
			name = pos->s_name;
			len = strlen(name);
			filp->f_pos = ino = pos->s_ino;
L
Linus Torvalds 已提交
865

E
Eric W. Biederman 已提交
866 867
			if (filldir(dirent, name, len, filp->f_pos, ino,
					 dt_type(pos)) < 0)
L
Linus Torvalds 已提交
868 869
				break;
		}
E
Eric W. Biederman 已提交
870 871
		if (!pos)
			filp->f_pos = INT_MAX;
872
		mutex_unlock(&sysfs_mutex);
L
Linus Torvalds 已提交
873
	}
E
Eric W. Biederman 已提交
874
	return 0;
L
Linus Torvalds 已提交
875 876
}

E
Eric W. Biederman 已提交
877

878
const struct file_operations sysfs_dir_operations = {
L
Linus Torvalds 已提交
879 880
	.read		= generic_read_dir,
	.readdir	= sysfs_readdir,
881
	.llseek		= generic_file_llseek,
L
Linus Torvalds 已提交
882
};