dquot.c 56.4 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 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 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 73 74 75 76 77 78
/*
 * Implementation of the diskquota system for the LINUX operating system. QUOTA
 * is implemented using the BSD system call interface as the means of
 * communication with the user level. This file contains the generic routines
 * called by the different filesystems on allocation of an inode or block.
 * These routines take care of the administration needed to have a consistent
 * diskquota tracking system. The ideas of both user and group quotas are based
 * on the Melbourne quota system as used on BSD derived systems. The internal
 * implementation is based on one of the several variants of the LINUX
 * inode-subsystem with added complexity of the diskquota system.
 * 
 * Version: $Id: dquot.c,v 6.3 1996/11/17 18:35:34 mvw Exp mvw $
 * 
 * Author:	Marco van Wieringen <mvw@planets.elm.net>
 *
 * Fixes:   Dmitry Gorodchanin <pgmdsg@ibi.com>, 11 Feb 96
 *
 *		Revised list management to avoid races
 *		-- Bill Hawes, <whawes@star.net>, 9/98
 *
 *		Fixed races in dquot_transfer(), dqget() and dquot_alloc_...().
 *		As the consequence the locking was moved from dquot_decr_...(),
 *		dquot_incr_...() to calling functions.
 *		invalidate_dquots() now writes modified dquots.
 *		Serialized quota_off() and quota_on() for mount point.
 *		Fixed a few bugs in grow_dquots().
 *		Fixed deadlock in write_dquot() - we no longer account quotas on
 *		quota files
 *		remove_dquot_ref() moved to inode.c - it now traverses through inodes
 *		add_dquot_ref() restarts after blocking
 *		Added check for bogus uid and fixed check for group in quotactl.
 *		Jan Kara, <jack@suse.cz>, sponsored by SuSE CR, 10-11/99
 *
 *		Used struct list_head instead of own list struct
 *		Invalidation of referenced dquots is no longer possible
 *		Improved free_dquots list management
 *		Quota and i_blocks are now updated in one place to avoid races
 *		Warnings are now delayed so we won't block in critical section
 *		Write updated not to require dquot lock
 *		Jan Kara, <jack@suse.cz>, 9/2000
 *
 *		Added dynamic quota structure allocation
 *		Jan Kara <jack@suse.cz> 12/2000
 *
 *		Rewritten quota interface. Implemented new quota format and
 *		formats registering.
 *		Jan Kara, <jack@suse.cz>, 2001,2002
 *
 *		New SMP locking.
 *		Jan Kara, <jack@suse.cz>, 10/2002
 *
 *		Added journalled quota support, fix lock inversion problems
 *		Jan Kara, <jack@suse.cz>, 2003,2004
 *
 * (C) Copyright 1994 - 1997 Marco van Wieringen 
 */

#include <linux/errno.h>
#include <linux/kernel.h>
#include <linux/fs.h>
#include <linux/mount.h>
#include <linux/mm.h>
#include <linux/time.h>
#include <linux/types.h>
#include <linux/string.h>
#include <linux/fcntl.h>
#include <linux/stat.h>
#include <linux/tty.h>
#include <linux/file.h>
#include <linux/slab.h>
#include <linux/sysctl.h>
#include <linux/init.h>
#include <linux/module.h>
#include <linux/proc_fs.h>
#include <linux/security.h>
#include <linux/kmod.h>
#include <linux/namei.h>
#include <linux/buffer_head.h>
79
#include <linux/capability.h>
A
Adrian Bunk 已提交
80
#include <linux/quotaops.h>
81
#include <linux/writeback.h> /* for inode_lock, oddly enough.. */
J
Jan Kara 已提交
82 83 84 85
#ifdef CONFIG_QUOTA_NETLINK_INTERFACE
#include <net/netlink.h>
#include <net/genetlink.h>
#endif
L
Linus Torvalds 已提交
86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107

#include <asm/uaccess.h>

#define __DQUOT_PARANOIA

/*
 * There are two quota SMP locks. dq_list_lock protects all lists with quotas
 * and quota formats and also dqstats structure containing statistics about the
 * lists. dq_data_lock protects data from dq_dqb and also mem_dqinfo structures
 * and also guards consistency of dquot->dq_dqb with inode->i_blocks, i_bytes.
 * i_blocks and i_bytes updates itself are guarded by i_lock acquired directly
 * in inode_add_bytes() and inode_sub_bytes().
 *
 * The spinlock ordering is hence: dq_data_lock > dq_list_lock > i_lock
 *
 * Note that some things (eg. sb pointer, type, id) doesn't change during
 * the life of the dquot structure and so needn't to be protected by a lock
 *
 * Any operation working on dquots via inode pointers must hold dqptr_sem.  If
 * operation is just reading pointers from inode (or not using them at all) the
 * read lock is enough. If pointers are altered function must hold write lock
 * (these locking rules also apply for S_NOQUOTA flag in the inode - note that
108
 * for altering the flag i_mutex is also needed).  If operation is holding
L
Linus Torvalds 已提交
109
 * reference to dquot in other way (e.g. quotactl ops) it must be guarded by
I
Ingo Molnar 已提交
110
 * dqonoff_mutex.
L
Linus Torvalds 已提交
111 112 113 114
 * This locking assures that:
 *   a) update/access to dquot pointers in inode is serialized
 *   b) everyone is guarded against invalidate_dquots()
 *
I
Ingo Molnar 已提交
115
 * Each dquot has its dq_lock mutex. Locked dquots might not be referenced
L
Linus Torvalds 已提交
116 117 118 119 120 121 122 123 124
 * from inodes (dquot_alloc_space() and such don't check the dq_lock).
 * Currently dquot is locked only when it is being read to memory (or space for
 * it is being allocated) on the first dqget() and when it is being released on
 * the last dqput(). The allocation and release oparations are serialized by
 * the dq_lock and by checking the use count in dquot_release().  Write
 * operations on dquots don't hold dq_lock as they copy data under dq_data_lock
 * spinlock to internal buffers before writing.
 *
 * Lock ordering (including related VFS locks) is the following:
I
Ingo Molnar 已提交
125 126 127
 *   i_mutex > dqonoff_sem > journal_lock > dqptr_sem > dquot->dq_lock >
 *   dqio_mutex
 * i_mutex on quota files is special (it's below dqio_mutex)
L
Linus Torvalds 已提交
128 129 130 131 132 133 134 135 136 137
 */

static DEFINE_SPINLOCK(dq_list_lock);
DEFINE_SPINLOCK(dq_data_lock);

static char *quotatypes[] = INITQFNAMES;
static struct quota_format_type *quota_formats;	/* List of registered formats */
static struct quota_module_name module_names[] = INIT_QUOTA_MODULE_NAMES;

/* SLAB cache for dquot structures */
138
static struct kmem_cache *dquot_cachep;
L
Linus Torvalds 已提交
139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256

int register_quota_format(struct quota_format_type *fmt)
{
	spin_lock(&dq_list_lock);
	fmt->qf_next = quota_formats;
	quota_formats = fmt;
	spin_unlock(&dq_list_lock);
	return 0;
}

void unregister_quota_format(struct quota_format_type *fmt)
{
	struct quota_format_type **actqf;

	spin_lock(&dq_list_lock);
	for (actqf = &quota_formats; *actqf && *actqf != fmt; actqf = &(*actqf)->qf_next);
	if (*actqf)
		*actqf = (*actqf)->qf_next;
	spin_unlock(&dq_list_lock);
}

static struct quota_format_type *find_quota_format(int id)
{
	struct quota_format_type *actqf;

	spin_lock(&dq_list_lock);
	for (actqf = quota_formats; actqf && actqf->qf_fmt_id != id; actqf = actqf->qf_next);
	if (!actqf || !try_module_get(actqf->qf_owner)) {
		int qm;

		spin_unlock(&dq_list_lock);
		
		for (qm = 0; module_names[qm].qm_fmt_id && module_names[qm].qm_fmt_id != id; qm++);
		if (!module_names[qm].qm_fmt_id || request_module(module_names[qm].qm_mod_name))
			return NULL;

		spin_lock(&dq_list_lock);
		for (actqf = quota_formats; actqf && actqf->qf_fmt_id != id; actqf = actqf->qf_next);
		if (actqf && !try_module_get(actqf->qf_owner))
			actqf = NULL;
	}
	spin_unlock(&dq_list_lock);
	return actqf;
}

static void put_quota_format(struct quota_format_type *fmt)
{
	module_put(fmt->qf_owner);
}

/*
 * Dquot List Management:
 * The quota code uses three lists for dquot management: the inuse_list,
 * free_dquots, and dquot_hash[] array. A single dquot structure may be
 * on all three lists, depending on its current state.
 *
 * All dquots are placed to the end of inuse_list when first created, and this
 * list is used for invalidate operation, which must look at every dquot.
 *
 * Unused dquots (dq_count == 0) are added to the free_dquots list when freed,
 * and this list is searched whenever we need an available dquot.  Dquots are
 * removed from the list as soon as they are used again, and
 * dqstats.free_dquots gives the number of dquots on the list. When
 * dquot is invalidated it's completely released from memory.
 *
 * Dquots with a specific identity (device, type and id) are placed on
 * one of the dquot_hash[] hash chains. The provides an efficient search
 * mechanism to locate a specific dquot.
 */

static LIST_HEAD(inuse_list);
static LIST_HEAD(free_dquots);
static unsigned int dq_hash_bits, dq_hash_mask;
static struct hlist_head *dquot_hash;

struct dqstats dqstats;

static void dqput(struct dquot *dquot);

static inline unsigned int
hashfn(const struct super_block *sb, unsigned int id, int type)
{
	unsigned long tmp;

	tmp = (((unsigned long)sb>>L1_CACHE_SHIFT) ^ id) * (MAXQUOTAS - type);
	return (tmp + (tmp >> dq_hash_bits)) & dq_hash_mask;
}

/*
 * Following list functions expect dq_list_lock to be held
 */
static inline void insert_dquot_hash(struct dquot *dquot)
{
	struct hlist_head *head = dquot_hash + hashfn(dquot->dq_sb, dquot->dq_id, dquot->dq_type);
	hlist_add_head(&dquot->dq_hash, head);
}

static inline void remove_dquot_hash(struct dquot *dquot)
{
	hlist_del_init(&dquot->dq_hash);
}

static inline struct dquot *find_dquot(unsigned int hashent, struct super_block *sb, unsigned int id, int type)
{
	struct hlist_node *node;
	struct dquot *dquot;

	hlist_for_each (node, dquot_hash+hashent) {
		dquot = hlist_entry(node, struct dquot, dq_hash);
		if (dquot->dq_sb == sb && dquot->dq_id == id && dquot->dq_type == type)
			return dquot;
	}
	return NODQUOT;
}

/* Add a dquot to the tail of the free list */
static inline void put_dquot_last(struct dquot *dquot)
{
257
	list_add_tail(&dquot->dq_free, &free_dquots);
L
Linus Torvalds 已提交
258 259 260 261 262 263 264 265 266 267 268 269 270 271 272
	dqstats.free_dquots++;
}

static inline void remove_free_dquot(struct dquot *dquot)
{
	if (list_empty(&dquot->dq_free))
		return;
	list_del_init(&dquot->dq_free);
	dqstats.free_dquots--;
}

static inline void put_inuse(struct dquot *dquot)
{
	/* We add to the back of inuse list so we don't have to restart
	 * when traversing this list and we block */
273
	list_add_tail(&dquot->dq_inuse, &inuse_list);
L
Linus Torvalds 已提交
274 275 276 277 278 279 280 281 282 283 284 285 286 287
	dqstats.allocated_dquots++;
}

static inline void remove_inuse(struct dquot *dquot)
{
	dqstats.allocated_dquots--;
	list_del(&dquot->dq_inuse);
}
/*
 * End of list functions needing dq_list_lock
 */

static void wait_on_dquot(struct dquot *dquot)
{
I
Ingo Molnar 已提交
288 289
	mutex_lock(&dquot->dq_lock);
	mutex_unlock(&dquot->dq_lock);
L
Linus Torvalds 已提交
290 291
}

J
Jan Kara 已提交
292 293 294 295 296 297 298 299 300
static inline int dquot_dirty(struct dquot *dquot)
{
	return test_bit(DQ_MOD_B, &dquot->dq_flags);
}

static inline int mark_dquot_dirty(struct dquot *dquot)
{
	return dquot->dq_sb->dq_op->mark_dirty(dquot);
}
L
Linus Torvalds 已提交
301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335

int dquot_mark_dquot_dirty(struct dquot *dquot)
{
	spin_lock(&dq_list_lock);
	if (!test_and_set_bit(DQ_MOD_B, &dquot->dq_flags))
		list_add(&dquot->dq_dirty, &sb_dqopt(dquot->dq_sb)->
				info[dquot->dq_type].dqi_dirty_list);
	spin_unlock(&dq_list_lock);
	return 0;
}

/* This function needs dq_list_lock */
static inline int clear_dquot_dirty(struct dquot *dquot)
{
	if (!test_and_clear_bit(DQ_MOD_B, &dquot->dq_flags))
		return 0;
	list_del_init(&dquot->dq_dirty);
	return 1;
}

void mark_info_dirty(struct super_block *sb, int type)
{
	set_bit(DQF_INFO_DIRTY_B, &sb_dqopt(sb)->info[type].dqi_flags);
}
EXPORT_SYMBOL(mark_info_dirty);

/*
 *	Read dquot from disk and alloc space for it
 */

int dquot_acquire(struct dquot *dquot)
{
	int ret = 0, ret2 = 0;
	struct quota_info *dqopt = sb_dqopt(dquot->dq_sb);

I
Ingo Molnar 已提交
336 337
	mutex_lock(&dquot->dq_lock);
	mutex_lock(&dqopt->dqio_mutex);
L
Linus Torvalds 已提交
338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357
	if (!test_bit(DQ_READ_B, &dquot->dq_flags))
		ret = dqopt->ops[dquot->dq_type]->read_dqblk(dquot);
	if (ret < 0)
		goto out_iolock;
	set_bit(DQ_READ_B, &dquot->dq_flags);
	/* Instantiate dquot if needed */
	if (!test_bit(DQ_ACTIVE_B, &dquot->dq_flags) && !dquot->dq_off) {
		ret = dqopt->ops[dquot->dq_type]->commit_dqblk(dquot);
		/* Write the info if needed */
		if (info_dirty(&dqopt->info[dquot->dq_type]))
			ret2 = dqopt->ops[dquot->dq_type]->write_file_info(dquot->dq_sb, dquot->dq_type);
		if (ret < 0)
			goto out_iolock;
		if (ret2 < 0) {
			ret = ret2;
			goto out_iolock;
		}
	}
	set_bit(DQ_ACTIVE_B, &dquot->dq_flags);
out_iolock:
I
Ingo Molnar 已提交
358 359
	mutex_unlock(&dqopt->dqio_mutex);
	mutex_unlock(&dquot->dq_lock);
L
Linus Torvalds 已提交
360 361 362 363 364 365 366 367 368 369 370
	return ret;
}

/*
 *	Write dquot to disk
 */
int dquot_commit(struct dquot *dquot)
{
	int ret = 0, ret2 = 0;
	struct quota_info *dqopt = sb_dqopt(dquot->dq_sb);

I
Ingo Molnar 已提交
371
	mutex_lock(&dqopt->dqio_mutex);
L
Linus Torvalds 已提交
372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387
	spin_lock(&dq_list_lock);
	if (!clear_dquot_dirty(dquot)) {
		spin_unlock(&dq_list_lock);
		goto out_sem;
	}
	spin_unlock(&dq_list_lock);
	/* Inactive dquot can be only if there was error during read/init
	 * => we have better not writing it */
	if (test_bit(DQ_ACTIVE_B, &dquot->dq_flags)) {
		ret = dqopt->ops[dquot->dq_type]->commit_dqblk(dquot);
		if (info_dirty(&dqopt->info[dquot->dq_type]))
			ret2 = dqopt->ops[dquot->dq_type]->write_file_info(dquot->dq_sb, dquot->dq_type);
		if (ret >= 0)
			ret = ret2;
	}
out_sem:
I
Ingo Molnar 已提交
388
	mutex_unlock(&dqopt->dqio_mutex);
L
Linus Torvalds 已提交
389 390 391 392 393 394 395 396 397 398 399
	return ret;
}

/*
 *	Release dquot
 */
int dquot_release(struct dquot *dquot)
{
	int ret = 0, ret2 = 0;
	struct quota_info *dqopt = sb_dqopt(dquot->dq_sb);

I
Ingo Molnar 已提交
400
	mutex_lock(&dquot->dq_lock);
L
Linus Torvalds 已提交
401 402 403
	/* Check whether we are not racing with some other dqget() */
	if (atomic_read(&dquot->dq_count) > 1)
		goto out_dqlock;
I
Ingo Molnar 已提交
404
	mutex_lock(&dqopt->dqio_mutex);
L
Linus Torvalds 已提交
405 406 407 408 409 410 411 412 413
	if (dqopt->ops[dquot->dq_type]->release_dqblk) {
		ret = dqopt->ops[dquot->dq_type]->release_dqblk(dquot);
		/* Write the info */
		if (info_dirty(&dqopt->info[dquot->dq_type]))
			ret2 = dqopt->ops[dquot->dq_type]->write_file_info(dquot->dq_sb, dquot->dq_type);
		if (ret >= 0)
			ret = ret2;
	}
	clear_bit(DQ_ACTIVE_B, &dquot->dq_flags);
I
Ingo Molnar 已提交
414
	mutex_unlock(&dqopt->dqio_mutex);
L
Linus Torvalds 已提交
415
out_dqlock:
I
Ingo Molnar 已提交
416
	mutex_unlock(&dquot->dq_lock);
L
Linus Torvalds 已提交
417 418 419 420 421
	return ret;
}

/* Invalidate all dquots on the list. Note that this function is called after
 * quota is disabled and pointers from inodes removed so there cannot be new
422 423 424 425
 * quota users. There can still be some users of quotas due to inodes being
 * just deleted or pruned by prune_icache() (those are not attached to any
 * list). We have to wait for such users.
 */
L
Linus Torvalds 已提交
426 427
static void invalidate_dquots(struct super_block *sb, int type)
{
428
	struct dquot *dquot, *tmp;
L
Linus Torvalds 已提交
429

430
restart:
L
Linus Torvalds 已提交
431
	spin_lock(&dq_list_lock);
432
	list_for_each_entry_safe(dquot, tmp, &inuse_list, dq_inuse) {
L
Linus Torvalds 已提交
433 434 435 436
		if (dquot->dq_sb != sb)
			continue;
		if (dquot->dq_type != type)
			continue;
437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464
		/* Wait for dquot users */
		if (atomic_read(&dquot->dq_count)) {
			DEFINE_WAIT(wait);

			atomic_inc(&dquot->dq_count);
			prepare_to_wait(&dquot->dq_wait_unused, &wait,
					TASK_UNINTERRUPTIBLE);
			spin_unlock(&dq_list_lock);
			/* Once dqput() wakes us up, we know it's time to free
			 * the dquot.
			 * IMPORTANT: we rely on the fact that there is always
			 * at most one process waiting for dquot to free.
			 * Otherwise dq_count would be > 1 and we would never
			 * wake up.
			 */
			if (atomic_read(&dquot->dq_count) > 1)
				schedule();
			finish_wait(&dquot->dq_wait_unused, &wait);
			dqput(dquot);
			/* At this moment dquot() need not exist (it could be
			 * reclaimed by prune_dqcache(). Hence we must
			 * restart. */
			goto restart;
		}
		/*
		 * Quota now has no users and it has been written on last
		 * dqput()
		 */
L
Linus Torvalds 已提交
465 466 467 468 469 470 471 472 473 474 475 476 477 478 479
		remove_dquot_hash(dquot);
		remove_free_dquot(dquot);
		remove_inuse(dquot);
		kmem_cache_free(dquot_cachep, dquot);
	}
	spin_unlock(&dq_list_lock);
}

int vfs_quota_sync(struct super_block *sb, int type)
{
	struct list_head *dirty;
	struct dquot *dquot;
	struct quota_info *dqopt = sb_dqopt(sb);
	int cnt;

I
Ingo Molnar 已提交
480
	mutex_lock(&dqopt->dqonoff_mutex);
L
Linus Torvalds 已提交
481 482 483 484 485 486 487 488
	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
		if (type != -1 && cnt != type)
			continue;
		if (!sb_has_quota_enabled(sb, cnt))
			continue;
		spin_lock(&dq_list_lock);
		dirty = &dqopt->info[cnt].dqi_dirty_list;
		while (!list_empty(dirty)) {
489
			dquot = list_first_entry(dirty, struct dquot, dq_dirty);
L
Linus Torvalds 已提交
490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514
			/* Dirty and inactive can be only bad dquot... */
			if (!test_bit(DQ_ACTIVE_B, &dquot->dq_flags)) {
				clear_dquot_dirty(dquot);
				continue;
			}
			/* Now we have active dquot from which someone is
 			 * holding reference so we can safely just increase
			 * use count */
			atomic_inc(&dquot->dq_count);
			dqstats.lookups++;
			spin_unlock(&dq_list_lock);
			sb->dq_op->write_dquot(dquot);
			dqput(dquot);
			spin_lock(&dq_list_lock);
		}
		spin_unlock(&dq_list_lock);
	}

	for (cnt = 0; cnt < MAXQUOTAS; cnt++)
		if ((cnt == type || type == -1) && sb_has_quota_enabled(sb, cnt)
			&& info_dirty(&dqopt->info[cnt]))
			sb->dq_op->write_info(sb, cnt);
	spin_lock(&dq_list_lock);
	dqstats.syncs++;
	spin_unlock(&dq_list_lock);
I
Ingo Molnar 已提交
515
	mutex_unlock(&dqopt->dqonoff_mutex);
L
Linus Torvalds 已提交
516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542

	return 0;
}

/* Free unused dquots from cache */
static void prune_dqcache(int count)
{
	struct list_head *head;
	struct dquot *dquot;

	head = free_dquots.prev;
	while (head != &free_dquots && count) {
		dquot = list_entry(head, struct dquot, dq_free);
		remove_dquot_hash(dquot);
		remove_free_dquot(dquot);
		remove_inuse(dquot);
		kmem_cache_free(dquot_cachep, dquot);
		count--;
		head = free_dquots.prev;
	}
}

/*
 * This is called from kswapd when we think we need some
 * more memory
 */

A
Al Viro 已提交
543
static int shrink_dqcache_memory(int nr, gfp_t gfp_mask)
L
Linus Torvalds 已提交
544 545 546 547 548 549 550 551 552
{
	if (nr) {
		spin_lock(&dq_list_lock);
		prune_dqcache(nr);
		spin_unlock(&dq_list_lock);
	}
	return (dqstats.free_dquots / 100) * sysctl_vfs_cache_pressure;
}

553 554 555 556 557
static struct shrinker dqcache_shrinker = {
	.shrink = shrink_dqcache_memory,
	.seeks = DEFAULT_SEEKS,
};

L
Linus Torvalds 已提交
558 559 560
/*
 * Put reference to dquot
 * NOTE: If you change this function please check whether dqput_blocks() works right...
I
Ingo Molnar 已提交
561
 * MUST be called with either dqptr_sem or dqonoff_mutex held
L
Linus Torvalds 已提交
562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585
 */
static void dqput(struct dquot *dquot)
{
	if (!dquot)
		return;
#ifdef __DQUOT_PARANOIA
	if (!atomic_read(&dquot->dq_count)) {
		printk("VFS: dqput: trying to free free dquot\n");
		printk("VFS: device %s, dquot of %s %d\n",
			dquot->dq_sb->s_id,
			quotatypes[dquot->dq_type],
			dquot->dq_id);
		BUG();
	}
#endif
	
	spin_lock(&dq_list_lock);
	dqstats.drops++;
	spin_unlock(&dq_list_lock);
we_slept:
	spin_lock(&dq_list_lock);
	if (atomic_read(&dquot->dq_count) > 1) {
		/* We have more than one user... nothing to do */
		atomic_dec(&dquot->dq_count);
586 587 588 589
		/* Releasing dquot during quotaoff phase? */
		if (!sb_has_quota_enabled(dquot->dq_sb, dquot->dq_type) &&
		    atomic_read(&dquot->dq_count) == 1)
			wake_up(&dquot->dq_wait_unused);
L
Linus Torvalds 已提交
590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609
		spin_unlock(&dq_list_lock);
		return;
	}
	/* Need to release dquot? */
	if (test_bit(DQ_ACTIVE_B, &dquot->dq_flags) && dquot_dirty(dquot)) {
		spin_unlock(&dq_list_lock);
		/* Commit dquot before releasing */
		dquot->dq_sb->dq_op->write_dquot(dquot);
		goto we_slept;
	}
	/* Clear flag in case dquot was inactive (something bad happened) */
	clear_dquot_dirty(dquot);
	if (test_bit(DQ_ACTIVE_B, &dquot->dq_flags)) {
		spin_unlock(&dq_list_lock);
		dquot->dq_sb->dq_op->release_dquot(dquot);
		goto we_slept;
	}
	atomic_dec(&dquot->dq_count);
#ifdef __DQUOT_PARANOIA
	/* sanity check */
610
	BUG_ON(!list_empty(&dquot->dq_free));
L
Linus Torvalds 已提交
611 612 613 614 615 616 617 618 619
#endif
	put_dquot_last(dquot);
	spin_unlock(&dq_list_lock);
}

static struct dquot *get_empty_dquot(struct super_block *sb, int type)
{
	struct dquot *dquot;

620
	dquot = kmem_cache_zalloc(dquot_cachep, GFP_NOFS);
L
Linus Torvalds 已提交
621 622 623
	if(!dquot)
		return NODQUOT;

I
Ingo Molnar 已提交
624
	mutex_init(&dquot->dq_lock);
L
Linus Torvalds 已提交
625 626 627 628
	INIT_LIST_HEAD(&dquot->dq_free);
	INIT_LIST_HEAD(&dquot->dq_inuse);
	INIT_HLIST_NODE(&dquot->dq_hash);
	INIT_LIST_HEAD(&dquot->dq_dirty);
629
	init_waitqueue_head(&dquot->dq_wait_unused);
L
Linus Torvalds 已提交
630 631 632 633 634 635 636 637 638
	dquot->dq_sb = sb;
	dquot->dq_type = type;
	atomic_set(&dquot->dq_count, 1);

	return dquot;
}

/*
 * Get reference to dquot
I
Ingo Molnar 已提交
639
 * MUST be called with either dqptr_sem or dqonoff_mutex held
L
Linus Torvalds 已提交
640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683
 */
static struct dquot *dqget(struct super_block *sb, unsigned int id, int type)
{
	unsigned int hashent = hashfn(sb, id, type);
	struct dquot *dquot, *empty = NODQUOT;

        if (!sb_has_quota_enabled(sb, type))
		return NODQUOT;
we_slept:
	spin_lock(&dq_list_lock);
	if ((dquot = find_dquot(hashent, sb, id, type)) == NODQUOT) {
		if (empty == NODQUOT) {
			spin_unlock(&dq_list_lock);
			if ((empty = get_empty_dquot(sb, type)) == NODQUOT)
				schedule();	/* Try to wait for a moment... */
			goto we_slept;
		}
		dquot = empty;
		dquot->dq_id = id;
		/* all dquots go on the inuse_list */
		put_inuse(dquot);
		/* hash it first so it can be found */
		insert_dquot_hash(dquot);
		dqstats.lookups++;
		spin_unlock(&dq_list_lock);
	} else {
		if (!atomic_read(&dquot->dq_count))
			remove_free_dquot(dquot);
		atomic_inc(&dquot->dq_count);
		dqstats.cache_hits++;
		dqstats.lookups++;
		spin_unlock(&dq_list_lock);
		if (empty)
			kmem_cache_free(dquot_cachep, empty);
	}
	/* Wait for dq_lock - after this we know that either dquot_release() is already
	 * finished or it will be canceled due to dq_count > 1 test */
	wait_on_dquot(dquot);
	/* Read the dquot and instantiate it (everything done only if needed) */
	if (!test_bit(DQ_ACTIVE_B, &dquot->dq_flags) && sb->dq_op->acquire_dquot(dquot) < 0) {
		dqput(dquot);
		return NODQUOT;
	}
#ifdef __DQUOT_PARANOIA
684
	BUG_ON(!dquot->dq_sb);	/* Has somebody invalidated entry under us? */
L
Linus Torvalds 已提交
685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703
#endif

	return dquot;
}

static int dqinit_needed(struct inode *inode, int type)
{
	int cnt;

	if (IS_NOQUOTA(inode))
		return 0;
	if (type != -1)
		return inode->i_dquot[type] == NODQUOT;
	for (cnt = 0; cnt < MAXQUOTAS; cnt++)
		if (inode->i_dquot[cnt] == NODQUOT)
			return 1;
	return 0;
}

I
Ingo Molnar 已提交
704
/* This routine is guarded by dqonoff_mutex mutex */
L
Linus Torvalds 已提交
705 706
static void add_dquot_ref(struct super_block *sb, int type)
{
707
	struct inode *inode, *old_inode = NULL;
L
Linus Torvalds 已提交
708

709 710 711 712 713 714 715 716 717 718 719 720
	spin_lock(&inode_lock);
	list_for_each_entry(inode, &sb->s_inodes, i_sb_list) {
		if (!atomic_read(&inode->i_writecount))
			continue;
		if (!dqinit_needed(inode, type))
			continue;
		if (inode->i_state & (I_FREEING|I_WILL_FREE))
			continue;

		__iget(inode);
		spin_unlock(&inode_lock);

721
		iput(old_inode);
722
		sb->dq_op->initialize(inode, type);
723 724 725 726 727 728 729
		/* We hold a reference to 'inode' so it couldn't have been
		 * removed from s_inodes list while we dropped the inode_lock.
		 * We cannot iput the inode now as we can be holding the last
		 * reference and we cannot iput it under inode_lock. So we
		 * keep the reference and iput it later. */
		old_inode = inode;
		spin_lock(&inode_lock);
L
Linus Torvalds 已提交
730
	}
731
	spin_unlock(&inode_lock);
732
	iput(old_inode);
L
Linus Torvalds 已提交
733 734 735 736 737 738 739 740 741 742 743 744
}

/* Return 0 if dqput() won't block (note that 1 doesn't necessarily mean blocking) */
static inline int dqput_blocks(struct dquot *dquot)
{
	if (atomic_read(&dquot->dq_count) <= 1)
		return 1;
	return 0;
}

/* Remove references to dquots from inode - add dquot to list for freeing if needed */
/* We can't race with anybody because we hold dqptr_sem for writing... */
745 746
static int remove_inode_dquot_ref(struct inode *inode, int type,
				  struct list_head *tofree_head)
L
Linus Torvalds 已提交
747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784
{
	struct dquot *dquot = inode->i_dquot[type];

	inode->i_dquot[type] = NODQUOT;
	if (dquot != NODQUOT) {
		if (dqput_blocks(dquot)) {
#ifdef __DQUOT_PARANOIA
			if (atomic_read(&dquot->dq_count) != 1)
				printk(KERN_WARNING "VFS: Adding dquot with dq_count %d to dispose list.\n", atomic_read(&dquot->dq_count));
#endif
			spin_lock(&dq_list_lock);
			list_add(&dquot->dq_free, tofree_head);	/* As dquot must have currently users it can't be on the free list... */
			spin_unlock(&dq_list_lock);
			return 1;
		}
		else
			dqput(dquot);   /* We have guaranteed we won't block */
	}
	return 0;
}

/* Free list of dquots - called from inode.c */
/* dquots are removed from inodes, no new references can be got so we are the only ones holding reference */
static void put_dquot_list(struct list_head *tofree_head)
{
	struct list_head *act_head;
	struct dquot *dquot;

	act_head = tofree_head->next;
	/* So now we have dquots on the list... Just free them */
	while (act_head != tofree_head) {
		dquot = list_entry(act_head, struct dquot, dq_free);
		act_head = act_head->next;
		list_del_init(&dquot->dq_free);	/* Remove dquot from the list so we won't have problems... */
		dqput(dquot);
	}
}

785 786 787 788 789 790 791 792 793 794 795 796 797
static void remove_dquot_ref(struct super_block *sb, int type,
		struct list_head *tofree_head)
{
	struct inode *inode;

	spin_lock(&inode_lock);
	list_for_each_entry(inode, &sb->s_inodes, i_sb_list) {
		if (!IS_NOQUOTA(inode))
			remove_inode_dquot_ref(inode, type, tofree_head);
	}
	spin_unlock(&inode_lock);
}

L
Linus Torvalds 已提交
798 799 800 801 802
/* Gather all references from inodes and drop them */
static void drop_dquot_ref(struct super_block *sb, int type)
{
	LIST_HEAD(tofree_head);

803 804 805 806 807 808
	if (sb->dq_op) {
		down_write(&sb_dqopt(sb)->dqptr_sem);
		remove_dquot_ref(sb, type, &tofree_head);
		up_write(&sb_dqopt(sb)->dqptr_sem);
		put_dquot_list(&tofree_head);
	}
L
Linus Torvalds 已提交
809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842
}

static inline void dquot_incr_inodes(struct dquot *dquot, unsigned long number)
{
	dquot->dq_dqb.dqb_curinodes += number;
}

static inline void dquot_incr_space(struct dquot *dquot, qsize_t number)
{
	dquot->dq_dqb.dqb_curspace += number;
}

static inline void dquot_decr_inodes(struct dquot *dquot, unsigned long number)
{
	if (dquot->dq_dqb.dqb_curinodes > number)
		dquot->dq_dqb.dqb_curinodes -= number;
	else
		dquot->dq_dqb.dqb_curinodes = 0;
	if (dquot->dq_dqb.dqb_curinodes <= dquot->dq_dqb.dqb_isoftlimit)
		dquot->dq_dqb.dqb_itime = (time_t) 0;
	clear_bit(DQ_INODES_B, &dquot->dq_flags);
}

static inline void dquot_decr_space(struct dquot *dquot, qsize_t number)
{
	if (dquot->dq_dqb.dqb_curspace > number)
		dquot->dq_dqb.dqb_curspace -= number;
	else
		dquot->dq_dqb.dqb_curspace = 0;
	if (toqb(dquot->dq_dqb.dqb_curspace) <= dquot->dq_dqb.dqb_bsoftlimit)
		dquot->dq_dqb.dqb_btime = (time_t) 0;
	clear_bit(DQ_BLKS_B, &dquot->dq_flags);
}

843 844 845 846 847 848 849 850 851 852 853 854
static int warning_issued(struct dquot *dquot, const int warntype)
{
	int flag = (warntype == QUOTA_NL_BHARDWARN ||
		warntype == QUOTA_NL_BSOFTLONGWARN) ? DQ_BLKS_B :
		((warntype == QUOTA_NL_IHARDWARN ||
		warntype == QUOTA_NL_ISOFTLONGWARN) ? DQ_INODES_B : 0);

	if (!flag)
		return 0;
	return test_and_set_bit(flag, &dquot->dq_flags);
}

J
Jan Kara 已提交
855
#ifdef CONFIG_PRINT_QUOTA_WARNING
L
Linus Torvalds 已提交
856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872
static int flag_print_warnings = 1;

static inline int need_print_warning(struct dquot *dquot)
{
	if (!flag_print_warnings)
		return 0;

	switch (dquot->dq_type) {
		case USRQUOTA:
			return current->fsuid == dquot->dq_id;
		case GRPQUOTA:
			return in_group_p(dquot->dq_id);
	}
	return 0;
}

/* Print warning to user which exceeded quota */
873
static void print_warning(struct dquot *dquot, const int warntype)
L
Linus Torvalds 已提交
874 875
{
	char *msg = NULL;
876
	struct tty_struct *tty;
L
Linus Torvalds 已提交
877

878
	if (!need_print_warning(dquot))
L
Linus Torvalds 已提交
879 880
		return;

881
	mutex_lock(&tty_mutex);
882 883
	tty = get_current_tty();
	if (!tty)
884
		goto out_lock;
885
	tty_write_message(tty, dquot->dq_sb->s_id);
J
Jan Kara 已提交
886
	if (warntype == QUOTA_NL_ISOFTWARN || warntype == QUOTA_NL_BSOFTWARN)
887
		tty_write_message(tty, ": warning, ");
L
Linus Torvalds 已提交
888
	else
889 890
		tty_write_message(tty, ": write failed, ");
	tty_write_message(tty, quotatypes[dquot->dq_type]);
L
Linus Torvalds 已提交
891
	switch (warntype) {
J
Jan Kara 已提交
892
		case QUOTA_NL_IHARDWARN:
L
Linus Torvalds 已提交
893 894
			msg = " file limit reached.\r\n";
			break;
J
Jan Kara 已提交
895
		case QUOTA_NL_ISOFTLONGWARN:
L
Linus Torvalds 已提交
896 897
			msg = " file quota exceeded too long.\r\n";
			break;
J
Jan Kara 已提交
898
		case QUOTA_NL_ISOFTWARN:
L
Linus Torvalds 已提交
899 900
			msg = " file quota exceeded.\r\n";
			break;
J
Jan Kara 已提交
901
		case QUOTA_NL_BHARDWARN:
L
Linus Torvalds 已提交
902 903
			msg = " block limit reached.\r\n";
			break;
J
Jan Kara 已提交
904
		case QUOTA_NL_BSOFTLONGWARN:
L
Linus Torvalds 已提交
905 906
			msg = " block quota exceeded too long.\r\n";
			break;
J
Jan Kara 已提交
907
		case QUOTA_NL_BSOFTWARN:
L
Linus Torvalds 已提交
908 909 910
			msg = " block quota exceeded.\r\n";
			break;
	}
911
	tty_write_message(tty, msg);
912 913
out_lock:
	mutex_unlock(&tty_mutex);
L
Linus Torvalds 已提交
914
}
J
Jan Kara 已提交
915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934
#endif

#ifdef CONFIG_QUOTA_NETLINK_INTERFACE

/* Netlink family structure for quota */
static struct genl_family quota_genl_family = {
	.id = GENL_ID_GENERATE,
	.hdrsize = 0,
	.name = "VFS_DQUOT",
	.version = 1,
	.maxattr = QUOTA_NL_A_MAX,
};

/* Send warning to userspace about user which exceeded quota */
static void send_warning(const struct dquot *dquot, const char warntype)
{
	static atomic_t seq;
	struct sk_buff *skb;
	void *msg_head;
	int ret;
935 936
	int msg_size = 4 * nla_total_size(sizeof(u32)) +
		       2 * nla_total_size(sizeof(u64));
J
Jan Kara 已提交
937 938 939 940

	/* We have to allocate using GFP_NOFS as we are called from a
	 * filesystem performing write and thus further recursion into
	 * the fs to free some data could cause deadlocks. */
941
	skb = genlmsg_new(msg_size, GFP_NOFS);
J
Jan Kara 已提交
942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981
	if (!skb) {
		printk(KERN_ERR
		  "VFS: Not enough memory to send quota warning.\n");
		return;
	}
	msg_head = genlmsg_put(skb, 0, atomic_add_return(1, &seq),
			&quota_genl_family, 0, QUOTA_NL_C_WARNING);
	if (!msg_head) {
		printk(KERN_ERR
		  "VFS: Cannot store netlink header in quota warning.\n");
		goto err_out;
	}
	ret = nla_put_u32(skb, QUOTA_NL_A_QTYPE, dquot->dq_type);
	if (ret)
		goto attr_err_out;
	ret = nla_put_u64(skb, QUOTA_NL_A_EXCESS_ID, dquot->dq_id);
	if (ret)
		goto attr_err_out;
	ret = nla_put_u32(skb, QUOTA_NL_A_WARNING, warntype);
	if (ret)
		goto attr_err_out;
	ret = nla_put_u32(skb, QUOTA_NL_A_DEV_MAJOR,
		MAJOR(dquot->dq_sb->s_dev));
	if (ret)
		goto attr_err_out;
	ret = nla_put_u32(skb, QUOTA_NL_A_DEV_MINOR,
		MINOR(dquot->dq_sb->s_dev));
	if (ret)
		goto attr_err_out;
	ret = nla_put_u64(skb, QUOTA_NL_A_CAUSED_ID, current->user->uid);
	if (ret)
		goto attr_err_out;
	genlmsg_end(skb, msg_head);

	ret = genlmsg_multicast(skb, 0, quota_genl_family.id, GFP_NOFS);
	if (ret < 0 && ret != -ESRCH)
		printk(KERN_ERR
			"VFS: Failed to send notification message: %d\n", ret);
	return;
attr_err_out:
982
	printk(KERN_ERR "VFS: Not enough space to compose quota message!\n");
J
Jan Kara 已提交
983 984 985 986
err_out:
	kfree_skb(skb);
}
#endif
L
Linus Torvalds 已提交
987

J
Jan Kara 已提交
988
static inline void flush_warnings(struct dquot * const *dquots, char *warntype)
L
Linus Torvalds 已提交
989 990 991 992
{
	int i;

	for (i = 0; i < MAXQUOTAS; i++)
993 994
		if (dquots[i] != NODQUOT && warntype[i] != QUOTA_NL_NOWARN &&
		    !warning_issued(dquots[i], warntype[i])) {
J
Jan Kara 已提交
995
#ifdef CONFIG_PRINT_QUOTA_WARNING
L
Linus Torvalds 已提交
996
			print_warning(dquots[i], warntype[i]);
J
Jan Kara 已提交
997 998 999 1000 1001
#endif
#ifdef CONFIG_QUOTA_NETLINK_INTERFACE
			send_warning(dquots[i], warntype[i]);
#endif
		}
L
Linus Torvalds 已提交
1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014
}

static inline char ignore_hardlimit(struct dquot *dquot)
{
	struct mem_dqinfo *info = &sb_dqopt(dquot->dq_sb)->info[dquot->dq_type];

	return capable(CAP_SYS_RESOURCE) &&
	    (info->dqi_format->qf_fmt_id != QFMT_VFS_OLD || !(info->dqi_flags & V1_DQF_RSQUASH));
}

/* needs dq_data_lock */
static int check_idq(struct dquot *dquot, ulong inodes, char *warntype)
{
J
Jan Kara 已提交
1015
	*warntype = QUOTA_NL_NOWARN;
L
Linus Torvalds 已提交
1016 1017 1018 1019 1020 1021
	if (inodes <= 0 || test_bit(DQ_FAKE_B, &dquot->dq_flags))
		return QUOTA_OK;

	if (dquot->dq_dqb.dqb_ihardlimit &&
	   (dquot->dq_dqb.dqb_curinodes + inodes) > dquot->dq_dqb.dqb_ihardlimit &&
            !ignore_hardlimit(dquot)) {
J
Jan Kara 已提交
1022
		*warntype = QUOTA_NL_IHARDWARN;
L
Linus Torvalds 已提交
1023 1024 1025 1026 1027 1028 1029
		return NO_QUOTA;
	}

	if (dquot->dq_dqb.dqb_isoftlimit &&
	   (dquot->dq_dqb.dqb_curinodes + inodes) > dquot->dq_dqb.dqb_isoftlimit &&
	    dquot->dq_dqb.dqb_itime && get_seconds() >= dquot->dq_dqb.dqb_itime &&
            !ignore_hardlimit(dquot)) {
J
Jan Kara 已提交
1030
		*warntype = QUOTA_NL_ISOFTLONGWARN;
L
Linus Torvalds 已提交
1031 1032 1033 1034 1035 1036
		return NO_QUOTA;
	}

	if (dquot->dq_dqb.dqb_isoftlimit &&
	   (dquot->dq_dqb.dqb_curinodes + inodes) > dquot->dq_dqb.dqb_isoftlimit &&
	    dquot->dq_dqb.dqb_itime == 0) {
J
Jan Kara 已提交
1037
		*warntype = QUOTA_NL_ISOFTWARN;
L
Linus Torvalds 已提交
1038 1039 1040 1041 1042 1043 1044 1045 1046
		dquot->dq_dqb.dqb_itime = get_seconds() + sb_dqopt(dquot->dq_sb)->info[dquot->dq_type].dqi_igrace;
	}

	return QUOTA_OK;
}

/* needs dq_data_lock */
static int check_bdq(struct dquot *dquot, qsize_t space, int prealloc, char *warntype)
{
J
Jan Kara 已提交
1047
	*warntype = QUOTA_NL_NOWARN;
L
Linus Torvalds 已提交
1048 1049 1050 1051 1052 1053 1054
	if (space <= 0 || test_bit(DQ_FAKE_B, &dquot->dq_flags))
		return QUOTA_OK;

	if (dquot->dq_dqb.dqb_bhardlimit &&
	   toqb(dquot->dq_dqb.dqb_curspace + space) > dquot->dq_dqb.dqb_bhardlimit &&
            !ignore_hardlimit(dquot)) {
		if (!prealloc)
J
Jan Kara 已提交
1055
			*warntype = QUOTA_NL_BHARDWARN;
L
Linus Torvalds 已提交
1056 1057 1058 1059 1060 1061 1062 1063
		return NO_QUOTA;
	}

	if (dquot->dq_dqb.dqb_bsoftlimit &&
	   toqb(dquot->dq_dqb.dqb_curspace + space) > dquot->dq_dqb.dqb_bsoftlimit &&
	    dquot->dq_dqb.dqb_btime && get_seconds() >= dquot->dq_dqb.dqb_btime &&
            !ignore_hardlimit(dquot)) {
		if (!prealloc)
J
Jan Kara 已提交
1064
			*warntype = QUOTA_NL_BSOFTLONGWARN;
L
Linus Torvalds 已提交
1065 1066 1067 1068 1069 1070 1071
		return NO_QUOTA;
	}

	if (dquot->dq_dqb.dqb_bsoftlimit &&
	   toqb(dquot->dq_dqb.dqb_curspace + space) > dquot->dq_dqb.dqb_bsoftlimit &&
	    dquot->dq_dqb.dqb_btime == 0) {
		if (!prealloc) {
J
Jan Kara 已提交
1072
			*warntype = QUOTA_NL_BSOFTWARN;
L
Linus Torvalds 已提交
1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094
			dquot->dq_dqb.dqb_btime = get_seconds() + sb_dqopt(dquot->dq_sb)->info[dquot->dq_type].dqi_bgrace;
		}
		else
			/*
			 * We don't allow preallocation to exceed softlimit so exceeding will
			 * be always printed
			 */
			return NO_QUOTA;
	}

	return QUOTA_OK;
}

/*
 *	Initialize quota pointers in inode
 *	Transaction must be started at entry
 */
int dquot_initialize(struct inode *inode, int type)
{
	unsigned int id = 0;
	int cnt, ret = 0;

I
Ingo Molnar 已提交
1095 1096
	/* First test before acquiring mutex - solves deadlocks when we
         * re-enter the quota code and are already holding the mutex */
L
Linus Torvalds 已提交
1097 1098 1099 1100 1101 1102 1103 1104 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 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158
	if (IS_NOQUOTA(inode))
		return 0;
	down_write(&sb_dqopt(inode->i_sb)->dqptr_sem);
	/* Having dqptr_sem we know NOQUOTA flags can't be altered... */
	if (IS_NOQUOTA(inode))
		goto out_err;
	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
		if (type != -1 && cnt != type)
			continue;
		if (inode->i_dquot[cnt] == NODQUOT) {
			switch (cnt) {
				case USRQUOTA:
					id = inode->i_uid;
					break;
				case GRPQUOTA:
					id = inode->i_gid;
					break;
			}
			inode->i_dquot[cnt] = dqget(inode->i_sb, id, cnt);
		}
	}
out_err:
	up_write(&sb_dqopt(inode->i_sb)->dqptr_sem);
	return ret;
}

/*
 * 	Release all quotas referenced by inode
 *	Transaction must be started at an entry
 */
int dquot_drop(struct inode *inode)
{
	int cnt;

	down_write(&sb_dqopt(inode->i_sb)->dqptr_sem);
	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
		if (inode->i_dquot[cnt] != NODQUOT) {
			dqput(inode->i_dquot[cnt]);
			inode->i_dquot[cnt] = NODQUOT;
		}
	}
	up_write(&sb_dqopt(inode->i_sb)->dqptr_sem);
	return 0;
}

/*
 * Following four functions update i_blocks+i_bytes fields and
 * quota information (together with appropriate checks)
 * NOTE: We absolutely rely on the fact that caller dirties
 * the inode (usually macros in quotaops.h care about this) and
 * holds a handle for the current transaction so that dquot write and
 * inode write go into the same transaction.
 */

/*
 * This operation can block, but only after everything is updated
 */
int dquot_alloc_space(struct inode *inode, qsize_t number, int warn)
{
	int cnt, ret = NO_QUOTA;
	char warntype[MAXQUOTAS];

I
Ingo Molnar 已提交
1159 1160
	/* First test before acquiring mutex - solves deadlocks when we
         * re-enter the quota code and are already holding the mutex */
L
Linus Torvalds 已提交
1161 1162 1163 1164 1165 1166
	if (IS_NOQUOTA(inode)) {
out_add:
		inode_add_bytes(inode, number);
		return QUOTA_OK;
	}
	for (cnt = 0; cnt < MAXQUOTAS; cnt++)
J
Jan Kara 已提交
1167
		warntype[cnt] = QUOTA_NL_NOWARN;
L
Linus Torvalds 已提交
1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207

	down_read(&sb_dqopt(inode->i_sb)->dqptr_sem);
	if (IS_NOQUOTA(inode)) {	/* Now we can do reliable test... */
		up_read(&sb_dqopt(inode->i_sb)->dqptr_sem);
		goto out_add;
	}
	spin_lock(&dq_data_lock);
	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
		if (inode->i_dquot[cnt] == NODQUOT)
			continue;
		if (check_bdq(inode->i_dquot[cnt], number, warn, warntype+cnt) == NO_QUOTA)
			goto warn_put_all;
	}
	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
		if (inode->i_dquot[cnt] == NODQUOT)
			continue;
		dquot_incr_space(inode->i_dquot[cnt], number);
	}
	inode_add_bytes(inode, number);
	ret = QUOTA_OK;
warn_put_all:
	spin_unlock(&dq_data_lock);
	if (ret == QUOTA_OK)
		/* Dirtify all the dquots - this can block when journalling */
		for (cnt = 0; cnt < MAXQUOTAS; cnt++)
			if (inode->i_dquot[cnt])
				mark_dquot_dirty(inode->i_dquot[cnt]);
	flush_warnings(inode->i_dquot, warntype);
	up_read(&sb_dqopt(inode->i_sb)->dqptr_sem);
	return ret;
}

/*
 * This operation can block, but only after everything is updated
 */
int dquot_alloc_inode(const struct inode *inode, unsigned long number)
{
	int cnt, ret = NO_QUOTA;
	char warntype[MAXQUOTAS];

I
Ingo Molnar 已提交
1208 1209
	/* First test before acquiring mutex - solves deadlocks when we
         * re-enter the quota code and are already holding the mutex */
L
Linus Torvalds 已提交
1210 1211 1212
	if (IS_NOQUOTA(inode))
		return QUOTA_OK;
	for (cnt = 0; cnt < MAXQUOTAS; cnt++)
J
Jan Kara 已提交
1213
		warntype[cnt] = QUOTA_NL_NOWARN;
L
Linus Torvalds 已提交
1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239
	down_read(&sb_dqopt(inode->i_sb)->dqptr_sem);
	if (IS_NOQUOTA(inode)) {
		up_read(&sb_dqopt(inode->i_sb)->dqptr_sem);
		return QUOTA_OK;
	}
	spin_lock(&dq_data_lock);
	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
		if (inode->i_dquot[cnt] == NODQUOT)
			continue;
		if (check_idq(inode->i_dquot[cnt], number, warntype+cnt) == NO_QUOTA)
			goto warn_put_all;
	}

	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
		if (inode->i_dquot[cnt] == NODQUOT)
			continue;
		dquot_incr_inodes(inode->i_dquot[cnt], number);
	}
	ret = QUOTA_OK;
warn_put_all:
	spin_unlock(&dq_data_lock);
	if (ret == QUOTA_OK)
		/* Dirtify all the dquots - this can block when journalling */
		for (cnt = 0; cnt < MAXQUOTAS; cnt++)
			if (inode->i_dquot[cnt])
				mark_dquot_dirty(inode->i_dquot[cnt]);
J
Jan Kara 已提交
1240
	flush_warnings(inode->i_dquot, warntype);
L
Linus Torvalds 已提交
1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251
	up_read(&sb_dqopt(inode->i_sb)->dqptr_sem);
	return ret;
}

/*
 * This operation can block, but only after everything is updated
 */
int dquot_free_space(struct inode *inode, qsize_t number)
{
	unsigned int cnt;

I
Ingo Molnar 已提交
1252 1253
	/* First test before acquiring mutex - solves deadlocks when we
         * re-enter the quota code and are already holding the mutex */
L
Linus Torvalds 已提交
1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287
	if (IS_NOQUOTA(inode)) {
out_sub:
		inode_sub_bytes(inode, number);
		return QUOTA_OK;
	}
	down_read(&sb_dqopt(inode->i_sb)->dqptr_sem);
	/* Now recheck reliably when holding dqptr_sem */
	if (IS_NOQUOTA(inode)) {
		up_read(&sb_dqopt(inode->i_sb)->dqptr_sem);
		goto out_sub;
	}
	spin_lock(&dq_data_lock);
	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
		if (inode->i_dquot[cnt] == NODQUOT)
			continue;
		dquot_decr_space(inode->i_dquot[cnt], number);
	}
	inode_sub_bytes(inode, number);
	spin_unlock(&dq_data_lock);
	/* Dirtify all the dquots - this can block when journalling */
	for (cnt = 0; cnt < MAXQUOTAS; cnt++)
		if (inode->i_dquot[cnt])
			mark_dquot_dirty(inode->i_dquot[cnt]);
	up_read(&sb_dqopt(inode->i_sb)->dqptr_sem);
	return QUOTA_OK;
}

/*
 * This operation can block, but only after everything is updated
 */
int dquot_free_inode(const struct inode *inode, unsigned long number)
{
	unsigned int cnt;

I
Ingo Molnar 已提交
1288 1289
	/* First test before acquiring mutex - solves deadlocks when we
         * re-enter the quota code and are already holding the mutex */
L
Linus Torvalds 已提交
1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327
	if (IS_NOQUOTA(inode))
		return QUOTA_OK;
	down_read(&sb_dqopt(inode->i_sb)->dqptr_sem);
	/* Now recheck reliably when holding dqptr_sem */
	if (IS_NOQUOTA(inode)) {
		up_read(&sb_dqopt(inode->i_sb)->dqptr_sem);
		return QUOTA_OK;
	}
	spin_lock(&dq_data_lock);
	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
		if (inode->i_dquot[cnt] == NODQUOT)
			continue;
		dquot_decr_inodes(inode->i_dquot[cnt], number);
	}
	spin_unlock(&dq_data_lock);
	/* Dirtify all the dquots - this can block when journalling */
	for (cnt = 0; cnt < MAXQUOTAS; cnt++)
		if (inode->i_dquot[cnt])
			mark_dquot_dirty(inode->i_dquot[cnt]);
	up_read(&sb_dqopt(inode->i_sb)->dqptr_sem);
	return QUOTA_OK;
}

/*
 * Transfer the number of inode and blocks from one diskquota to an other.
 *
 * This operation can block, but only after everything is updated
 * A transaction must be started when entering this function.
 */
int dquot_transfer(struct inode *inode, struct iattr *iattr)
{
	qsize_t space;
	struct dquot *transfer_from[MAXQUOTAS];
	struct dquot *transfer_to[MAXQUOTAS];
	int cnt, ret = NO_QUOTA, chuid = (iattr->ia_valid & ATTR_UID) && inode->i_uid != iattr->ia_uid,
	    chgid = (iattr->ia_valid & ATTR_GID) && inode->i_gid != iattr->ia_gid;
	char warntype[MAXQUOTAS];

I
Ingo Molnar 已提交
1328 1329
	/* First test before acquiring mutex - solves deadlocks when we
         * re-enter the quota code and are already holding the mutex */
L
Linus Torvalds 已提交
1330 1331 1332 1333 1334
	if (IS_NOQUOTA(inode))
		return QUOTA_OK;
	/* Clear the arrays */
	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
		transfer_to[cnt] = transfer_from[cnt] = NODQUOT;
J
Jan Kara 已提交
1335
		warntype[cnt] = QUOTA_NL_NOWARN;
L
Linus Torvalds 已提交
1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422
	}
	down_write(&sb_dqopt(inode->i_sb)->dqptr_sem);
	/* Now recheck reliably when holding dqptr_sem */
	if (IS_NOQUOTA(inode)) {	/* File without quota accounting? */
		up_write(&sb_dqopt(inode->i_sb)->dqptr_sem);
		return QUOTA_OK;
	}
	/* First build the transfer_to list - here we can block on
	 * reading/instantiating of dquots.  We know that the transaction for
	 * us was already started so we don't violate lock ranking here */
	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
		switch (cnt) {
			case USRQUOTA:
				if (!chuid)
					continue;
				transfer_to[cnt] = dqget(inode->i_sb, iattr->ia_uid, cnt);
				break;
			case GRPQUOTA:
				if (!chgid)
					continue;
				transfer_to[cnt] = dqget(inode->i_sb, iattr->ia_gid, cnt);
				break;
		}
	}
	spin_lock(&dq_data_lock);
	space = inode_get_bytes(inode);
	/* Build the transfer_from list and check the limits */
	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
		if (transfer_to[cnt] == NODQUOT)
			continue;
		transfer_from[cnt] = inode->i_dquot[cnt];
		if (check_idq(transfer_to[cnt], 1, warntype+cnt) == NO_QUOTA ||
		    check_bdq(transfer_to[cnt], space, 0, warntype+cnt) == NO_QUOTA)
			goto warn_put_all;
	}

	/*
	 * Finally perform the needed transfer from transfer_from to transfer_to
	 */
	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
		/*
		 * Skip changes for same uid or gid or for turned off quota-type.
		 */
		if (transfer_to[cnt] == NODQUOT)
			continue;

		/* Due to IO error we might not have transfer_from[] structure */
		if (transfer_from[cnt]) {
			dquot_decr_inodes(transfer_from[cnt], 1);
			dquot_decr_space(transfer_from[cnt], space);
		}

		dquot_incr_inodes(transfer_to[cnt], 1);
		dquot_incr_space(transfer_to[cnt], space);

		inode->i_dquot[cnt] = transfer_to[cnt];
	}
	ret = QUOTA_OK;
warn_put_all:
	spin_unlock(&dq_data_lock);
	/* Dirtify all the dquots - this can block when journalling */
	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
		if (transfer_from[cnt])
			mark_dquot_dirty(transfer_from[cnt]);
		if (transfer_to[cnt])
			mark_dquot_dirty(transfer_to[cnt]);
	}
	flush_warnings(transfer_to, warntype);
	
	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
		if (ret == QUOTA_OK && transfer_from[cnt] != NODQUOT)
			dqput(transfer_from[cnt]);
		if (ret == NO_QUOTA && transfer_to[cnt] != NODQUOT)
			dqput(transfer_to[cnt]);
	}
	up_write(&sb_dqopt(inode->i_sb)->dqptr_sem);
	return ret;
}

/*
 * Write info of quota file to disk
 */
int dquot_commit_info(struct super_block *sb, int type)
{
	int ret;
	struct quota_info *dqopt = sb_dqopt(sb);

I
Ingo Molnar 已提交
1423
	mutex_lock(&dqopt->dqio_mutex);
L
Linus Torvalds 已提交
1424
	ret = dqopt->ops[type]->write_file_info(sb, type);
I
Ingo Molnar 已提交
1425
	mutex_unlock(&dqopt->dqio_mutex);
L
Linus Torvalds 已提交
1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480
	return ret;
}

/*
 * Definitions of diskquota operations.
 */
struct dquot_operations dquot_operations = {
	.initialize	= dquot_initialize,
	.drop		= dquot_drop,
	.alloc_space	= dquot_alloc_space,
	.alloc_inode	= dquot_alloc_inode,
	.free_space	= dquot_free_space,
	.free_inode	= dquot_free_inode,
	.transfer	= dquot_transfer,
	.write_dquot	= dquot_commit,
	.acquire_dquot	= dquot_acquire,
	.release_dquot	= dquot_release,
	.mark_dirty	= dquot_mark_dquot_dirty,
	.write_info	= dquot_commit_info
};

static inline void set_enable_flags(struct quota_info *dqopt, int type)
{
	switch (type) {
		case USRQUOTA:
			dqopt->flags |= DQUOT_USR_ENABLED;
			break;
		case GRPQUOTA:
			dqopt->flags |= DQUOT_GRP_ENABLED;
			break;
	}
}

static inline void reset_enable_flags(struct quota_info *dqopt, int type)
{
	switch (type) {
		case USRQUOTA:
			dqopt->flags &= ~DQUOT_USR_ENABLED;
			break;
		case GRPQUOTA:
			dqopt->flags &= ~DQUOT_GRP_ENABLED;
			break;
	}
}

/*
 * Turn quota off on a device. type == -1 ==> quotaoff for all types (umount)
 */
int vfs_quota_off(struct super_block *sb, int type)
{
	int cnt;
	struct quota_info *dqopt = sb_dqopt(sb);
	struct inode *toputinode[MAXQUOTAS];

	/* We need to serialize quota_off() for device */
I
Ingo Molnar 已提交
1481
	mutex_lock(&dqopt->dqonoff_mutex);
L
Linus Torvalds 已提交
1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509
	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
		toputinode[cnt] = NULL;
		if (type != -1 && cnt != type)
			continue;
		if (!sb_has_quota_enabled(sb, cnt))
			continue;
		reset_enable_flags(dqopt, cnt);

		/* Note: these are blocking operations */
		drop_dquot_ref(sb, cnt);
		invalidate_dquots(sb, cnt);
		/*
		 * Now all dquots should be invalidated, all writes done so we should be only
		 * users of the info. No locks needed.
		 */
		if (info_dirty(&dqopt->info[cnt]))
			sb->dq_op->write_info(sb, cnt);
		if (dqopt->ops[cnt]->free_file_info)
			dqopt->ops[cnt]->free_file_info(sb, cnt);
		put_quota_format(dqopt->info[cnt].dqi_format);

		toputinode[cnt] = dqopt->files[cnt];
		dqopt->files[cnt] = NULL;
		dqopt->info[cnt].dqi_flags = 0;
		dqopt->info[cnt].dqi_igrace = 0;
		dqopt->info[cnt].dqi_bgrace = 0;
		dqopt->ops[cnt] = NULL;
	}
I
Ingo Molnar 已提交
1510
	mutex_unlock(&dqopt->dqonoff_mutex);
L
Linus Torvalds 已提交
1511
	/* Sync the superblock so that buffers with quota data are written to
1512
	 * disk (and so userspace sees correct data afterwards). */
L
Linus Torvalds 已提交
1513 1514 1515 1516 1517 1518 1519 1520 1521 1522
	if (sb->s_op->sync_fs)
		sb->s_op->sync_fs(sb, 1);
	sync_blockdev(sb->s_bdev);
	/* Now the quota files are just ordinary files and we can set the
	 * inode flags back. Moreover we discard the pagecache so that
	 * userspace sees the writes we did bypassing the pagecache. We
	 * must also discard the blockdev buffers so that we see the
	 * changes done by userspace on the next quotaon() */
	for (cnt = 0; cnt < MAXQUOTAS; cnt++)
		if (toputinode[cnt]) {
I
Ingo Molnar 已提交
1523
			mutex_lock(&dqopt->dqonoff_mutex);
L
Linus Torvalds 已提交
1524 1525 1526
			/* If quota was reenabled in the meantime, we have
			 * nothing to do */
			if (!sb_has_quota_enabled(sb, cnt)) {
1527
				mutex_lock_nested(&toputinode[cnt]->i_mutex, I_MUTEX_QUOTA);
L
Linus Torvalds 已提交
1528 1529 1530
				toputinode[cnt]->i_flags &= ~(S_IMMUTABLE |
				  S_NOATIME | S_NOQUOTA);
				truncate_inode_pages(&toputinode[cnt]->i_data, 0);
1531
				mutex_unlock(&toputinode[cnt]->i_mutex);
L
Linus Torvalds 已提交
1532 1533
				mark_inode_dirty(toputinode[cnt]);
			}
1534
			iput(toputinode[cnt]);
I
Ingo Molnar 已提交
1535
			mutex_unlock(&dqopt->dqonoff_mutex);
L
Linus Torvalds 已提交
1536 1537
		}
	if (sb->s_bdev)
1538
		invalidate_bdev(sb->s_bdev);
L
Linus Torvalds 已提交
1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573
	return 0;
}

/*
 *	Turn quotas on on a device
 */

/* Helper function when we already have the inode */
static int vfs_quota_on_inode(struct inode *inode, int type, int format_id)
{
	struct quota_format_type *fmt = find_quota_format(format_id);
	struct super_block *sb = inode->i_sb;
	struct quota_info *dqopt = sb_dqopt(sb);
	int error;
	int oldflags = -1;

	if (!fmt)
		return -ESRCH;
	if (!S_ISREG(inode->i_mode)) {
		error = -EACCES;
		goto out_fmt;
	}
	if (IS_RDONLY(inode)) {
		error = -EROFS;
		goto out_fmt;
	}
	if (!sb->s_op->quota_write || !sb->s_op->quota_read) {
		error = -EINVAL;
		goto out_fmt;
	}

	/* As we bypass the pagecache we must now flush the inode so that
	 * we see all the changes from userspace... */
	write_inode_now(inode, 1);
	/* And now flush the block cache so that kernel sees the changes */
1574
	invalidate_bdev(sb->s_bdev);
1575
	mutex_lock(&inode->i_mutex);
I
Ingo Molnar 已提交
1576
	mutex_lock(&dqopt->dqonoff_mutex);
L
Linus Torvalds 已提交
1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587
	if (sb_has_quota_enabled(sb, type)) {
		error = -EBUSY;
		goto out_lock;
	}
	/* We don't want quota and atime on quota files (deadlocks possible)
	 * Also nobody should write to the file - we use special IO operations
	 * which ignore the immutable bit. */
	down_write(&dqopt->dqptr_sem);
	oldflags = inode->i_flags & (S_NOATIME | S_IMMUTABLE | S_NOQUOTA);
	inode->i_flags |= S_NOQUOTA | S_NOATIME | S_IMMUTABLE;
	up_write(&dqopt->dqptr_sem);
1588
	sb->dq_op->drop(inode);
L
Linus Torvalds 已提交
1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600

	error = -EIO;
	dqopt->files[type] = igrab(inode);
	if (!dqopt->files[type])
		goto out_lock;
	error = -EINVAL;
	if (!fmt->qf_ops->check_quota_file(sb, type))
		goto out_file_init;

	dqopt->ops[type] = fmt->qf_ops;
	dqopt->info[type].dqi_format = fmt;
	INIT_LIST_HEAD(&dqopt->info[type].dqi_dirty_list);
I
Ingo Molnar 已提交
1601
	mutex_lock(&dqopt->dqio_mutex);
L
Linus Torvalds 已提交
1602
	if ((error = dqopt->ops[type]->read_file_info(sb, type)) < 0) {
I
Ingo Molnar 已提交
1603
		mutex_unlock(&dqopt->dqio_mutex);
L
Linus Torvalds 已提交
1604 1605
		goto out_file_init;
	}
I
Ingo Molnar 已提交
1606
	mutex_unlock(&dqopt->dqio_mutex);
1607
	mutex_unlock(&inode->i_mutex);
L
Linus Torvalds 已提交
1608 1609 1610
	set_enable_flags(dqopt, type);

	add_dquot_ref(sb, type);
I
Ingo Molnar 已提交
1611
	mutex_unlock(&dqopt->dqonoff_mutex);
L
Linus Torvalds 已提交
1612 1613 1614 1615 1616 1617 1618

	return 0;

out_file_init:
	dqopt->files[type] = NULL;
	iput(inode);
out_lock:
I
Ingo Molnar 已提交
1619
	mutex_unlock(&dqopt->dqonoff_mutex);
L
Linus Torvalds 已提交
1620 1621 1622 1623 1624 1625 1626 1627
	if (oldflags != -1) {
		down_write(&dqopt->dqptr_sem);
		/* Set the flags back (in the case of accidental quotaon()
		 * on a wrong file we don't want to mess up the flags) */
		inode->i_flags &= ~(S_NOATIME | S_NOQUOTA | S_IMMUTABLE);
		inode->i_flags |= oldflags;
		up_write(&dqopt->dqptr_sem);
	}
1628
	mutex_unlock(&inode->i_mutex);
L
Linus Torvalds 已提交
1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643
out_fmt:
	put_quota_format(fmt);

	return error; 
}

/* Actual function called from quotactl() */
int vfs_quota_on(struct super_block *sb, int type, int format_id, char *path)
{
	struct nameidata nd;
	int error;

	error = path_lookup(path, LOOKUP_FOLLOW, &nd);
	if (error < 0)
		return error;
1644
	error = security_quota_on(nd.path.dentry);
L
Linus Torvalds 已提交
1645 1646 1647
	if (error)
		goto out_path;
	/* Quota file not on the same filesystem? */
1648
	if (nd.path.mnt->mnt_sb != sb)
L
Linus Torvalds 已提交
1649
		error = -EXDEV;
1650
	else
1651 1652
		error = vfs_quota_on_inode(nd.path.dentry->d_inode, type,
					   format_id);
L
Linus Torvalds 已提交
1653
out_path:
J
Jan Blunck 已提交
1654
	path_put(&nd.path);
L
Linus Torvalds 已提交
1655 1656 1657 1658 1659 1660 1661
	return error;
}

/*
 * This function is used when filesystem needs to initialize quotas
 * during mount time.
 */
1662 1663
int vfs_quota_on_mount(struct super_block *sb, char *qf_name,
		int format_id, int type)
L
Linus Torvalds 已提交
1664
{
1665
	struct dentry *dentry;
L
Linus Torvalds 已提交
1666 1667
	int error;

1668
	dentry = lookup_one_len(qf_name, sb->s_root, strlen(qf_name));
1669 1670 1671
	if (IS_ERR(dentry))
		return PTR_ERR(dentry);

J
Jan Kara 已提交
1672 1673 1674 1675 1676
	if (!dentry->d_inode) {
		error = -ENOENT;
		goto out;
	}

L
Linus Torvalds 已提交
1677
	error = security_quota_on(dentry);
1678 1679 1680
	if (!error)
		error = vfs_quota_on_inode(dentry->d_inode, type, format_id);

J
Jan Kara 已提交
1681
out:
1682 1683
	dput(dentry);
	return error;
L
Linus Torvalds 已提交
1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707
}

/* Generic routine for getting common part of quota structure */
static void do_get_dqblk(struct dquot *dquot, struct if_dqblk *di)
{
	struct mem_dqblk *dm = &dquot->dq_dqb;

	spin_lock(&dq_data_lock);
	di->dqb_bhardlimit = dm->dqb_bhardlimit;
	di->dqb_bsoftlimit = dm->dqb_bsoftlimit;
	di->dqb_curspace = dm->dqb_curspace;
	di->dqb_ihardlimit = dm->dqb_ihardlimit;
	di->dqb_isoftlimit = dm->dqb_isoftlimit;
	di->dqb_curinodes = dm->dqb_curinodes;
	di->dqb_btime = dm->dqb_btime;
	di->dqb_itime = dm->dqb_itime;
	di->dqb_valid = QIF_ALL;
	spin_unlock(&dq_data_lock);
}

int vfs_get_dqblk(struct super_block *sb, int type, qid_t id, struct if_dqblk *di)
{
	struct dquot *dquot;

I
Ingo Molnar 已提交
1708
	mutex_lock(&sb_dqopt(sb)->dqonoff_mutex);
L
Linus Torvalds 已提交
1709
	if (!(dquot = dqget(sb, id, type))) {
I
Ingo Molnar 已提交
1710
		mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
L
Linus Torvalds 已提交
1711 1712 1713 1714
		return -ESRCH;
	}
	do_get_dqblk(dquot, di);
	dqput(dquot);
I
Ingo Molnar 已提交
1715
	mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
L
Linus Torvalds 已提交
1716 1717 1718 1719
	return 0;
}

/* Generic routine for setting common part of quota structure */
1720
static int do_set_dqblk(struct dquot *dquot, struct if_dqblk *di)
L
Linus Torvalds 已提交
1721 1722 1723
{
	struct mem_dqblk *dm = &dquot->dq_dqb;
	int check_blim = 0, check_ilim = 0;
1724 1725 1726 1727 1728 1729 1730 1731 1732
	struct mem_dqinfo *dqi = &sb_dqopt(dquot->dq_sb)->info[dquot->dq_type];

	if ((di->dqb_valid & QIF_BLIMITS &&
	     (di->dqb_bhardlimit > dqi->dqi_maxblimit ||
	      di->dqb_bsoftlimit > dqi->dqi_maxblimit)) ||
	    (di->dqb_valid & QIF_ILIMITS &&
	     (di->dqb_ihardlimit > dqi->dqi_maxilimit ||
	      di->dqb_isoftlimit > dqi->dqi_maxilimit)))
		return -ERANGE;
L
Linus Torvalds 已提交
1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763

	spin_lock(&dq_data_lock);
	if (di->dqb_valid & QIF_SPACE) {
		dm->dqb_curspace = di->dqb_curspace;
		check_blim = 1;
	}
	if (di->dqb_valid & QIF_BLIMITS) {
		dm->dqb_bsoftlimit = di->dqb_bsoftlimit;
		dm->dqb_bhardlimit = di->dqb_bhardlimit;
		check_blim = 1;
	}
	if (di->dqb_valid & QIF_INODES) {
		dm->dqb_curinodes = di->dqb_curinodes;
		check_ilim = 1;
	}
	if (di->dqb_valid & QIF_ILIMITS) {
		dm->dqb_isoftlimit = di->dqb_isoftlimit;
		dm->dqb_ihardlimit = di->dqb_ihardlimit;
		check_ilim = 1;
	}
	if (di->dqb_valid & QIF_BTIME)
		dm->dqb_btime = di->dqb_btime;
	if (di->dqb_valid & QIF_ITIME)
		dm->dqb_itime = di->dqb_itime;

	if (check_blim) {
		if (!dm->dqb_bsoftlimit || toqb(dm->dqb_curspace) < dm->dqb_bsoftlimit) {
			dm->dqb_btime = 0;
			clear_bit(DQ_BLKS_B, &dquot->dq_flags);
		}
		else if (!(di->dqb_valid & QIF_BTIME))	/* Set grace only if user hasn't provided his own... */
1764
			dm->dqb_btime = get_seconds() + dqi->dqi_bgrace;
L
Linus Torvalds 已提交
1765 1766 1767 1768 1769 1770 1771
	}
	if (check_ilim) {
		if (!dm->dqb_isoftlimit || dm->dqb_curinodes < dm->dqb_isoftlimit) {
			dm->dqb_itime = 0;
			clear_bit(DQ_INODES_B, &dquot->dq_flags);
		}
		else if (!(di->dqb_valid & QIF_ITIME))	/* Set grace only if user hasn't provided his own... */
1772
			dm->dqb_itime = get_seconds() + dqi->dqi_igrace;
L
Linus Torvalds 已提交
1773 1774 1775 1776 1777 1778 1779
	}
	if (dm->dqb_bhardlimit || dm->dqb_bsoftlimit || dm->dqb_ihardlimit || dm->dqb_isoftlimit)
		clear_bit(DQ_FAKE_B, &dquot->dq_flags);
	else
		set_bit(DQ_FAKE_B, &dquot->dq_flags);
	spin_unlock(&dq_data_lock);
	mark_dquot_dirty(dquot);
1780 1781

	return 0;
L
Linus Torvalds 已提交
1782 1783 1784 1785 1786
}

int vfs_set_dqblk(struct super_block *sb, int type, qid_t id, struct if_dqblk *di)
{
	struct dquot *dquot;
1787
	int rc;
L
Linus Torvalds 已提交
1788

I
Ingo Molnar 已提交
1789
	mutex_lock(&sb_dqopt(sb)->dqonoff_mutex);
L
Linus Torvalds 已提交
1790
	if (!(dquot = dqget(sb, id, type))) {
I
Ingo Molnar 已提交
1791
		mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
L
Linus Torvalds 已提交
1792 1793
		return -ESRCH;
	}
1794
	rc = do_set_dqblk(dquot, di);
L
Linus Torvalds 已提交
1795
	dqput(dquot);
I
Ingo Molnar 已提交
1796
	mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
1797
	return rc;
L
Linus Torvalds 已提交
1798 1799 1800 1801 1802 1803 1804
}

/* Generic routine for getting common part of quota file information */
int vfs_get_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii)
{
	struct mem_dqinfo *mi;
  
I
Ingo Molnar 已提交
1805
	mutex_lock(&sb_dqopt(sb)->dqonoff_mutex);
L
Linus Torvalds 已提交
1806
	if (!sb_has_quota_enabled(sb, type)) {
I
Ingo Molnar 已提交
1807
		mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
L
Linus Torvalds 已提交
1808 1809 1810 1811 1812 1813 1814 1815 1816
		return -ESRCH;
	}
	mi = sb_dqopt(sb)->info + type;
	spin_lock(&dq_data_lock);
	ii->dqi_bgrace = mi->dqi_bgrace;
	ii->dqi_igrace = mi->dqi_igrace;
	ii->dqi_flags = mi->dqi_flags & DQF_MASK;
	ii->dqi_valid = IIF_ALL;
	spin_unlock(&dq_data_lock);
I
Ingo Molnar 已提交
1817
	mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
L
Linus Torvalds 已提交
1818 1819 1820 1821 1822 1823 1824 1825
	return 0;
}

/* Generic routine for setting common part of quota file information */
int vfs_set_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii)
{
	struct mem_dqinfo *mi;

I
Ingo Molnar 已提交
1826
	mutex_lock(&sb_dqopt(sb)->dqonoff_mutex);
L
Linus Torvalds 已提交
1827
	if (!sb_has_quota_enabled(sb, type)) {
I
Ingo Molnar 已提交
1828
		mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
L
Linus Torvalds 已提交
1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842
		return -ESRCH;
	}
	mi = sb_dqopt(sb)->info + type;
	spin_lock(&dq_data_lock);
	if (ii->dqi_valid & IIF_BGRACE)
		mi->dqi_bgrace = ii->dqi_bgrace;
	if (ii->dqi_valid & IIF_IGRACE)
		mi->dqi_igrace = ii->dqi_igrace;
	if (ii->dqi_valid & IIF_FLAGS)
		mi->dqi_flags = (mi->dqi_flags & ~DQF_MASK) | (ii->dqi_flags & DQF_MASK);
	spin_unlock(&dq_data_lock);
	mark_info_dirty(sb, type);
	/* Force write to disk */
	sb->dq_op->write_info(sb, type);
I
Ingo Molnar 已提交
1843
	mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
L
Linus Torvalds 已提交
1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921
	return 0;
}

struct quotactl_ops vfs_quotactl_ops = {
	.quota_on	= vfs_quota_on,
	.quota_off	= vfs_quota_off,
	.quota_sync	= vfs_quota_sync,
	.get_info	= vfs_get_dqinfo,
	.set_info	= vfs_set_dqinfo,
	.get_dqblk	= vfs_get_dqblk,
	.set_dqblk	= vfs_set_dqblk
};

static ctl_table fs_dqstats_table[] = {
	{
		.ctl_name	= FS_DQ_LOOKUPS,
		.procname	= "lookups",
		.data		= &dqstats.lookups,
		.maxlen		= sizeof(int),
		.mode		= 0444,
		.proc_handler	= &proc_dointvec,
	},
	{
		.ctl_name	= FS_DQ_DROPS,
		.procname	= "drops",
		.data		= &dqstats.drops,
		.maxlen		= sizeof(int),
		.mode		= 0444,
		.proc_handler	= &proc_dointvec,
	},
	{
		.ctl_name	= FS_DQ_READS,
		.procname	= "reads",
		.data		= &dqstats.reads,
		.maxlen		= sizeof(int),
		.mode		= 0444,
		.proc_handler	= &proc_dointvec,
	},
	{
		.ctl_name	= FS_DQ_WRITES,
		.procname	= "writes",
		.data		= &dqstats.writes,
		.maxlen		= sizeof(int),
		.mode		= 0444,
		.proc_handler	= &proc_dointvec,
	},
	{
		.ctl_name	= FS_DQ_CACHE_HITS,
		.procname	= "cache_hits",
		.data		= &dqstats.cache_hits,
		.maxlen		= sizeof(int),
		.mode		= 0444,
		.proc_handler	= &proc_dointvec,
	},
	{
		.ctl_name	= FS_DQ_ALLOCATED,
		.procname	= "allocated_dquots",
		.data		= &dqstats.allocated_dquots,
		.maxlen		= sizeof(int),
		.mode		= 0444,
		.proc_handler	= &proc_dointvec,
	},
	{
		.ctl_name	= FS_DQ_FREE,
		.procname	= "free_dquots",
		.data		= &dqstats.free_dquots,
		.maxlen		= sizeof(int),
		.mode		= 0444,
		.proc_handler	= &proc_dointvec,
	},
	{
		.ctl_name	= FS_DQ_SYNCS,
		.procname	= "syncs",
		.data		= &dqstats.syncs,
		.maxlen		= sizeof(int),
		.mode		= 0444,
		.proc_handler	= &proc_dointvec,
	},
J
Jan Kara 已提交
1922
#ifdef CONFIG_PRINT_QUOTA_WARNING
L
Linus Torvalds 已提交
1923 1924 1925 1926 1927 1928 1929 1930
	{
		.ctl_name	= FS_DQ_WARNINGS,
		.procname	= "warnings",
		.data		= &flag_print_warnings,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= &proc_dointvec,
	},
J
Jan Kara 已提交
1931
#endif
L
Linus Torvalds 已提交
1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961
	{ .ctl_name = 0 },
};

static ctl_table fs_table[] = {
	{
		.ctl_name	= FS_DQSTATS,
		.procname	= "quota",
		.mode		= 0555,
		.child		= fs_dqstats_table,
	},
	{ .ctl_name = 0 },
};

static ctl_table sys_table[] = {
	{
		.ctl_name	= CTL_FS,
		.procname	= "fs",
		.mode		= 0555,
		.child		= fs_table,
	},
	{ .ctl_name = 0 },
};

static int __init dquot_init(void)
{
	int i;
	unsigned long nr_hash, order;

	printk(KERN_NOTICE "VFS: Disk quotas %s\n", __DQUOT_VERSION__);

1962
	register_sysctl_table(sys_table);
L
Linus Torvalds 已提交
1963

1964
	dquot_cachep = kmem_cache_create("dquot",
L
Linus Torvalds 已提交
1965
			sizeof(struct dquot), sizeof(unsigned long) * 4,
1966 1967
			(SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|
				SLAB_MEM_SPREAD|SLAB_PANIC),
1968
			NULL);
L
Linus Torvalds 已提交
1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990

	order = 0;
	dquot_hash = (struct hlist_head *)__get_free_pages(GFP_ATOMIC, order);
	if (!dquot_hash)
		panic("Cannot create dquot hash table");

	/* Find power-of-two hlist_heads which can fit into allocation */
	nr_hash = (1UL << order) * PAGE_SIZE / sizeof(struct hlist_head);
	dq_hash_bits = 0;
	do {
		dq_hash_bits++;
	} while (nr_hash >> dq_hash_bits);
	dq_hash_bits--;

	nr_hash = 1UL << dq_hash_bits;
	dq_hash_mask = nr_hash - 1;
	for (i = 0; i < nr_hash; i++)
		INIT_HLIST_HEAD(dquot_hash + i);

	printk("Dquot-cache hash table entries: %ld (order %ld, %ld bytes)\n",
			nr_hash, order, (PAGE_SIZE << order));

1991
	register_shrinker(&dqcache_shrinker);
L
Linus Torvalds 已提交
1992

J
Jan Kara 已提交
1993 1994 1995 1996 1997
#ifdef CONFIG_QUOTA_NETLINK_INTERFACE
	if (genl_register_family(&quota_genl_family) != 0)
		printk(KERN_ERR "VFS: Failed to create quota netlink interface.\n");
#endif

L
Linus Torvalds 已提交
1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025
	return 0;
}
module_init(dquot_init);

EXPORT_SYMBOL(register_quota_format);
EXPORT_SYMBOL(unregister_quota_format);
EXPORT_SYMBOL(dqstats);
EXPORT_SYMBOL(dq_data_lock);
EXPORT_SYMBOL(vfs_quota_on);
EXPORT_SYMBOL(vfs_quota_on_mount);
EXPORT_SYMBOL(vfs_quota_off);
EXPORT_SYMBOL(vfs_quota_sync);
EXPORT_SYMBOL(vfs_get_dqinfo);
EXPORT_SYMBOL(vfs_set_dqinfo);
EXPORT_SYMBOL(vfs_get_dqblk);
EXPORT_SYMBOL(vfs_set_dqblk);
EXPORT_SYMBOL(dquot_commit);
EXPORT_SYMBOL(dquot_commit_info);
EXPORT_SYMBOL(dquot_acquire);
EXPORT_SYMBOL(dquot_release);
EXPORT_SYMBOL(dquot_mark_dquot_dirty);
EXPORT_SYMBOL(dquot_initialize);
EXPORT_SYMBOL(dquot_drop);
EXPORT_SYMBOL(dquot_alloc_space);
EXPORT_SYMBOL(dquot_alloc_inode);
EXPORT_SYMBOL(dquot_free_space);
EXPORT_SYMBOL(dquot_free_inode);
EXPORT_SYMBOL(dquot_transfer);