dir.c 20.4 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
	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);
109 110
		if (likely(t == v)) {
			rwsem_acquire_read(&sd->dep_map, 0, 1, _RET_IP_);
111
			return sd;
112
		}
113 114 115 116
		if (t < 0)
			return NULL;

		cpu_relax();
117 118 119 120 121 122 123 124 125 126
	}
}

/**
 *	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.
 */
127
static void sysfs_put_active(struct sysfs_dirent *sd)
128
{
129 130 131 132 133 134
	struct completion *cmpl;
	int v;

	if (unlikely(!sd))
		return;

135
	rwsem_release(&sd->dep_map, 1, _RET_IP_);
136 137 138 139 140
	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
141
	 * sd->s_sibling.
142
	 */
143
	cmpl = (void *)sd->s_sibling;
144
	complete(cmpl);
145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189
}

/**
 *	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
 *
190
 *	Deny new active references and drain existing ones.
191
 */
192
static void sysfs_deactivate(struct sysfs_dirent *sd)
193
{
194 195
	DECLARE_COMPLETION_ONSTACK(wait);
	int v;
196

197
	BUG_ON(sd->s_sibling || !(sd->s_flags & SYSFS_FLAG_REMOVED));
198
	sd->s_sibling = (void *)&wait;
199

200
	rwsem_acquire(&sd->dep_map, 0, 0, _RET_IP_);
201
	/* atomic_add_return() is a mb(), put_active() will always see
202
	 * the updated sd->s_sibling.
203
	 */
204 205
	v = atomic_add_return(SD_DEACTIVATED_BIAS, &sd->s_active);

206 207
	if (v != SD_DEACTIVATED_BIAS) {
		lock_contended(&sd->dep_map, _RET_IP_);
208
		wait_for_completion(&wait);
209
	}
210

211
	sd->s_sibling = NULL;
212 213 214

	lock_acquired(&sd->dep_map, _RET_IP_);
	rwsem_release(&sd->dep_map, 1, _RET_IP_);
215 216
}

T
Tejun Heo 已提交
217
static int sysfs_alloc_ino(ino_t *pino)
218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242
{
	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);
}

243 244
void release_sysfs_dirent(struct sysfs_dirent * sd)
{
T
Tejun Heo 已提交
245 246 247
	struct sysfs_dirent *parent_sd;

 repeat:
248 249 250
	/* Moving/renaming is always done while holding reference.
	 * sd->s_parent won't change beneath us.
	 */
T
Tejun Heo 已提交
251 252
	parent_sd = sd->s_parent;

253
	if (sysfs_type(sd) == SYSFS_KOBJ_LINK)
T
Tejun Heo 已提交
254
		sysfs_put(sd->s_symlink.target_sd);
255
	if (sysfs_type(sd) & SYSFS_COPY_NAME)
T
Tejun Heo 已提交
256
		kfree(sd->s_name);
257 258 259
	if (sd->s_iattr && sd->s_iattr->ia_secdata)
		security_release_secctx(sd->s_iattr->ia_secdata,
					sd->s_iattr->ia_secdata_len);
260
	kfree(sd->s_iattr);
261
	sysfs_free_ino(sd->s_ino);
262
	kmem_cache_free(sysfs_dir_cachep, sd);
T
Tejun Heo 已提交
263 264 265 266

	sd = parent_sd;
	if (sd && atomic_dec_and_test(&sd->s_count))
		goto repeat;
267 268
}

269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285
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;

286 287 288 289 290 291 292 293
	/* 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;

294 295 296 297 298 299 300
	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.
301 302 303 304 305 306
	 *
	 * 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.
307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322
	 */
	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;
}

323
static void sysfs_dentry_iput(struct dentry *dentry, struct inode *inode)
L
Linus Torvalds 已提交
324 325 326
{
	struct sysfs_dirent * sd = dentry->d_fsdata;

E
Eric W. Biederman 已提交
327
	sysfs_put(sd);
L
Linus Torvalds 已提交
328 329 330
	iput(inode);
}

A
Al Viro 已提交
331
static const struct dentry_operations sysfs_dentry_ops = {
332 333
	.d_revalidate	= sysfs_dentry_revalidate,
	.d_delete	= sysfs_dentry_delete,
334
	.d_iput		= sysfs_dentry_iput,
L
Linus Torvalds 已提交
335 336
};

337
struct sysfs_dirent *sysfs_new_dirent(const char *name, umode_t mode, int type)
L
Linus Torvalds 已提交
338
{
T
Tejun Heo 已提交
339
	char *dup_name = NULL;
340
	struct sysfs_dirent *sd;
T
Tejun Heo 已提交
341 342 343 344

	if (type & SYSFS_COPY_NAME) {
		name = dup_name = kstrdup(name, GFP_KERNEL);
		if (!name)
345
			return NULL;
T
Tejun Heo 已提交
346
	}
L
Linus Torvalds 已提交
347

348
	sd = kmem_cache_zalloc(sysfs_dir_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
349
	if (!sd)
350
		goto err_out1;
L
Linus Torvalds 已提交
351

T
Tejun Heo 已提交
352
	if (sysfs_alloc_ino(&sd->s_ino))
353
		goto err_out2;
354

L
Linus Torvalds 已提交
355
	atomic_set(&sd->s_count, 1);
356
	atomic_set(&sd->s_active, 0);
357
	sysfs_dirent_init_lockdep(sd);
358

T
Tejun Heo 已提交
359
	sd->s_name = name;
360
	sd->s_mode = mode;
361
	sd->s_flags = type;
L
Linus Torvalds 已提交
362 363

	return sd;
T
Tejun Heo 已提交
364

365
 err_out2:
T
Tejun Heo 已提交
366
	kmem_cache_free(sysfs_dir_cachep, sd);
367 368
 err_out1:
	kfree(dup_name);
T
Tejun Heo 已提交
369
	return NULL;
L
Linus Torvalds 已提交
370 371
}

372
/**
373 374 375
 *	sysfs_addrm_start - prepare for sysfs_dirent add/remove
 *	@acxt: pointer to sysfs_addrm_cxt to be used
 *	@parent_sd: parent sysfs_dirent
376
 *
377 378
 *	This function is called when the caller is about to add or
 *	remove sysfs_dirent under @parent_sd.  This function acquires
379
 *	sysfs_mutex.  @acxt is used to keep and pass context to
380
 *	other addrm functions.
381 382
 *
 *	LOCKING:
383
 *	Kernel thread context (may sleep).  sysfs_mutex is locked on
384
 *	return.
385
 */
386 387
void sysfs_addrm_start(struct sysfs_addrm_cxt *acxt,
		       struct sysfs_dirent *parent_sd)
388
{
389 390 391 392 393 394 395
	memset(acxt, 0, sizeof(*acxt));
	acxt->parent_sd = parent_sd;

	mutex_lock(&sysfs_mutex);
}

/**
396
 *	__sysfs_add_one - add sysfs_dirent to parent without warning
397 398 399 400
 *	@acxt: addrm context to use
 *	@sd: sysfs_dirent to be added
 *
 *	Get @acxt->parent_sd and set sd->s_parent to it and increment
401 402
 *	nlink of parent inode if @sd is a directory and link into the
 *	children list of the parent.
403 404 405 406 407 408 409
 *
 *	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().
410 411 412 413
 *
 *	RETURNS:
 *	0 on success, -EEXIST if entry with the given name already
 *	exists.
414
 */
415
int __sysfs_add_one(struct sysfs_addrm_cxt *acxt, struct sysfs_dirent *sd)
416
{
417 418
	struct sysfs_inode_attrs *ps_iattr;

419
	if (sysfs_find_dirent(acxt->parent_sd, sd->s_name))
420 421
		return -EEXIST;

422 423
	sd->s_parent = sysfs_get(acxt->parent_sd);

424
	sysfs_link_sibling(sd);
425

426 427 428 429 430 431 432
	/* 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;
	}

433
	return 0;
434 435
}

436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455
/**
 *	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;
}

456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480
/**
 *	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);
481 482 483 484 485 486 487 488 489 490
	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);
	}

491 492 493
	return ret;
}

494 495 496
/**
 *	sysfs_remove_one - remove sysfs_dirent from parent
 *	@acxt: addrm context to use
497
 *	@sd: sysfs_dirent to be removed
498 499
 *
 *	Mark @sd removed and drop nlink of parent inode if @sd is a
500
 *	directory.  @sd is unlinked from the children list.
501 502 503 504 505 506 507 508 509 510
 *
 *	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)
{
511 512
	struct sysfs_inode_attrs *ps_iattr;

513 514 515
	BUG_ON(sd->s_flags & SYSFS_FLAG_REMOVED);

	sysfs_unlink_sibling(sd);
516

517 518 519 520 521 522 523
	/* 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;
	}

524 525 526
	sd->s_flags |= SYSFS_FLAG_REMOVED;
	sd->s_sibling = acxt->removed;
	acxt->removed = sd;
527 528
}

529 530 531 532 533 534
/**
 *	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
535
 *	cleaned up.
536 537
 *
 *	LOCKING:
538
 *	sysfs_mutex is released.
539
 */
540
void sysfs_addrm_finish(struct sysfs_addrm_cxt *acxt)
541 542 543 544 545 546 547 548 549 550 551 552
{
	/* 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);
553
		unmap_bin_file(sd);
554
		sysfs_put(sd);
T
Tejun Heo 已提交
555
	}
556 557
}

558 559 560 561 562 563
/**
 *	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.
564
 *
565
 *	LOCKING:
566
 *	mutex_lock(sysfs_mutex)
567
 *
568 569
 *	RETURNS:
 *	Pointer to sysfs_dirent if found, NULL if not.
570
 */
571 572
struct sysfs_dirent *sysfs_find_dirent(struct sysfs_dirent *parent_sd,
				       const unsigned char *name)
573
{
574 575
	struct sysfs_dirent *sd;

576
	for (sd = parent_sd->s_dir.children; sd; sd = sd->s_sibling)
E
Eric W. Biederman 已提交
577
		if (!strcmp(sd->s_name, name))
578 579 580
			return sd;
	return NULL;
}
581

582 583 584 585 586 587 588 589 590
/**
 *	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:
591
 *	Kernel thread context (may sleep).  Grabs sysfs_mutex.
592 593 594 595 596 597 598 599 600
 *
 *	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;

601
	mutex_lock(&sysfs_mutex);
602 603
	sd = sysfs_find_dirent(parent_sd, name);
	sysfs_get(sd);
604
	mutex_unlock(&sysfs_mutex);
605 606

	return sd;
607
}
608
EXPORT_SYMBOL_GPL(sysfs_get_dirent);
609

610 611
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
612 613
{
	umode_t mode = S_IFDIR| S_IRWXU | S_IRUGO | S_IXUGO;
614
	struct sysfs_addrm_cxt acxt;
615
	struct sysfs_dirent *sd;
616
	int rc;
L
Linus Torvalds 已提交
617

618
	/* allocate */
619
	sd = sysfs_new_dirent(name, mode, SYSFS_DIR);
620
	if (!sd)
621
		return -ENOMEM;
T
Tejun Heo 已提交
622
	sd->s_dir.kobj = kobj;
623

624
	/* link in */
625
	sysfs_addrm_start(&acxt, parent_sd);
626 627
	rc = sysfs_add_one(&acxt, sd);
	sysfs_addrm_finish(&acxt);
628

629 630 631
	if (rc == 0)
		*p_sd = sd;
	else
632
		sysfs_put(sd);
633

634
	return rc;
L
Linus Torvalds 已提交
635 636
}

637 638
int sysfs_create_subdir(struct kobject *kobj, const char *name,
			struct sysfs_dirent **p_sd)
L
Linus Torvalds 已提交
639
{
640
	return create_dir(kobj, kobj->sd, name, p_sd);
L
Linus Torvalds 已提交
641 642 643 644 645 646
}

/**
 *	sysfs_create_dir - create a directory for an object.
 *	@kobj:		object we're creating directory for. 
 */
647
int sysfs_create_dir(struct kobject * kobj)
L
Linus Torvalds 已提交
648
{
649
	struct sysfs_dirent *parent_sd, *sd;
L
Linus Torvalds 已提交
650 651 652 653
	int error = 0;

	BUG_ON(!kobj);

654
	if (kobj->parent)
655
		parent_sd = kobj->parent->sd;
L
Linus Torvalds 已提交
656
	else
E
Eric W. Biederman 已提交
657
		parent_sd = &sysfs_root;
L
Linus Torvalds 已提交
658

659
	error = create_dir(kobj, parent_sd, kobject_name(kobj), &sd);
L
Linus Torvalds 已提交
660
	if (!error)
661
		kobj->sd = sd;
L
Linus Torvalds 已提交
662 663 664 665 666 667
	return error;
}

static struct dentry * sysfs_lookup(struct inode *dir, struct dentry *dentry,
				struct nameidata *nd)
{
668
	struct dentry *ret = NULL;
669 670
	struct sysfs_dirent *parent_sd = dentry->d_parent->d_fsdata;
	struct sysfs_dirent *sd;
671
	struct inode *inode;
L
Linus Torvalds 已提交
672

673 674
	mutex_lock(&sysfs_mutex);

675
	sd = sysfs_find_dirent(parent_sd, dentry->d_name.name);
L
Linus Torvalds 已提交
676

677
	/* no such entry */
678 679
	if (!sd) {
		ret = ERR_PTR(-ENOENT);
680
		goto out_unlock;
681
	}
682 683

	/* attach dentry and inode */
684
	inode = sysfs_get_inode(sd);
685 686 687 688
	if (!inode) {
		ret = ERR_PTR(-ENOMEM);
		goto out_unlock;
	}
689

T
Tejun Heo 已提交
690
	/* instantiate and hash dentry */
691 692 693 694 695 696 697 698 699
	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);
	}
700

701
 out_unlock:
702
	mutex_unlock(&sysfs_mutex);
703
	return ret;
L
Linus Torvalds 已提交
704 705
}

706
const struct inode_operations sysfs_dir_inode_operations = {
L
Linus Torvalds 已提交
707
	.lookup		= sysfs_lookup,
708
	.permission	= sysfs_permission,
709
	.setattr	= sysfs_setattr,
710
	.getattr	= sysfs_getattr,
711
	.setxattr	= sysfs_setxattr,
L
Linus Torvalds 已提交
712 713
};

714
static void remove_dir(struct sysfs_dirent *sd)
L
Linus Torvalds 已提交
715
{
716
	struct sysfs_addrm_cxt acxt;
L
Linus Torvalds 已提交
717

718 719 720
	sysfs_addrm_start(&acxt, sd->s_parent);
	sysfs_remove_one(&acxt, sd);
	sysfs_addrm_finish(&acxt);
L
Linus Torvalds 已提交
721 722
}

723
void sysfs_remove_subdir(struct sysfs_dirent *sd)
L
Linus Torvalds 已提交
724
{
725
	remove_dir(sd);
L
Linus Torvalds 已提交
726 727 728
}


729
static void __sysfs_remove_dir(struct sysfs_dirent *dir_sd)
L
Linus Torvalds 已提交
730
{
731
	struct sysfs_addrm_cxt acxt;
732
	struct sysfs_dirent **pos;
L
Linus Torvalds 已提交
733

734
	if (!dir_sd)
L
Linus Torvalds 已提交
735 736
		return;

737
	pr_debug("sysfs %s: removing dir\n", dir_sd->s_name);
738
	sysfs_addrm_start(&acxt, dir_sd);
739
	pos = &dir_sd->s_dir.children;
740 741 742
	while (*pos) {
		struct sysfs_dirent *sd = *pos;

E
Eric W. Biederman 已提交
743
		if (sysfs_type(sd) != SYSFS_DIR)
744
			sysfs_remove_one(&acxt, sd);
745
		else
746
			pos = &(*pos)->s_sibling;
L
Linus Torvalds 已提交
747
	}
748
	sysfs_addrm_finish(&acxt);
749

750
	remove_dir(dir_sd);
751 752 753 754 755 756 757 758 759 760 761 762 763
}

/**
 *	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)
{
764
	struct sysfs_dirent *sd = kobj->sd;
765

T
Tejun Heo 已提交
766
	spin_lock(&sysfs_assoc_lock);
767
	kobj->sd = NULL;
T
Tejun Heo 已提交
768
	spin_unlock(&sysfs_assoc_lock);
769

770
	__sysfs_remove_dir(sd);
L
Linus Torvalds 已提交
771 772
}

773 774
int sysfs_rename(struct sysfs_dirent *sd,
	struct sysfs_dirent *new_parent_sd, const char *new_name)
L
Linus Torvalds 已提交
775
{
776
	const char *dup_name = NULL;
777
	int error;
L
Linus Torvalds 已提交
778

779
	mutex_lock(&sysfs_mutex);
780

781
	error = 0;
782 783
	if ((sd->s_parent == new_parent_sd) &&
	    (strcmp(sd->s_name, new_name) == 0))
784 785 786
		goto out;	/* nothing to rename */

	error = -EEXIST;
787
	if (sysfs_find_dirent(new_parent_sd, new_name))
788
		goto out;
789

790
	/* rename sysfs_dirent */
791 792 793 794 795 796 797 798 799
	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
800

801 802 803 804 805 806 807 808
	/* 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
809

810
	error = 0;
811
 out:
812
	mutex_unlock(&sysfs_mutex);
813
	kfree(dup_name);
L
Linus Torvalds 已提交
814 815 816
	return error;
}

817 818 819 820 821
int sysfs_rename_dir(struct kobject *kobj, const char *new_name)
{
	return sysfs_rename(kobj->sd, kobj->sd->s_parent, new_name);
}

822
int sysfs_move_dir(struct kobject *kobj, struct kobject *new_parent_kobj)
823
{
824 825
	struct sysfs_dirent *sd = kobj->sd;
	struct sysfs_dirent *new_parent_sd;
826

827
	BUG_ON(!sd->s_parent);
828
	new_parent_sd = new_parent_kobj && new_parent_kobj->sd ?
829
		new_parent_kobj->sd : &sysfs_root;
830

831
	return sysfs_rename(sd, new_parent_sd, sd->s_name);
832 833
}

L
Linus Torvalds 已提交
834 835 836 837 838 839 840 841
/* 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)
{
842
	struct dentry *dentry = filp->f_path.dentry;
L
Linus Torvalds 已提交
843
	struct sysfs_dirent * parent_sd = dentry->d_fsdata;
E
Eric W. Biederman 已提交
844
	struct sysfs_dirent *pos;
L
Linus Torvalds 已提交
845 846
	ino_t ino;

E
Eric W. Biederman 已提交
847 848 849
	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
850
			filp->f_pos++;
E
Eric W. Biederman 已提交
851 852 853 854 855 856 857
	}
	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
858
			filp->f_pos++;
E
Eric W. Biederman 已提交
859 860 861
	}
	if ((filp->f_pos > 1) && (filp->f_pos < INT_MAX)) {
		mutex_lock(&sysfs_mutex);
862

E
Eric W. Biederman 已提交
863
		/* Skip the dentries we have already reported */
864
		pos = parent_sd->s_dir.children;
E
Eric W. Biederman 已提交
865 866
		while (pos && (filp->f_pos > pos->s_ino))
			pos = pos->s_sibling;
867

E
Eric W. Biederman 已提交
868 869 870
		for ( ; pos; pos = pos->s_sibling) {
			const char * name;
			int len;
L
Linus Torvalds 已提交
871

E
Eric W. Biederman 已提交
872 873 874
			name = pos->s_name;
			len = strlen(name);
			filp->f_pos = ino = pos->s_ino;
L
Linus Torvalds 已提交
875

E
Eric W. Biederman 已提交
876 877
			if (filldir(dirent, name, len, filp->f_pos, ino,
					 dt_type(pos)) < 0)
L
Linus Torvalds 已提交
878 879
				break;
		}
E
Eric W. Biederman 已提交
880 881
		if (!pos)
			filp->f_pos = INT_MAX;
882
		mutex_unlock(&sysfs_mutex);
L
Linus Torvalds 已提交
883
	}
E
Eric W. Biederman 已提交
884
	return 0;
L
Linus Torvalds 已提交
885 886
}

E
Eric W. Biederman 已提交
887

888
const struct file_operations sysfs_dir_operations = {
L
Linus Torvalds 已提交
889 890
	.read		= generic_read_dir,
	.readdir	= sysfs_readdir,
891
	.llseek		= generic_file_llseek,
L
Linus Torvalds 已提交
892
};