dquot.c 52.9 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 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 79
/*
 * 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/smp_lock.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>
80
#include <linux/capability.h>
A
Adrian Bunk 已提交
81
#include <linux/quotaops.h>
82
#include <linux/writeback.h> /* for inode_lock, oddly enough.. */
L
Linus Torvalds 已提交
83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104

#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
105
 * for altering the flag i_mutex is also needed).  If operation is holding
L
Linus Torvalds 已提交
106
 * reference to dquot in other way (e.g. quotactl ops) it must be guarded by
I
Ingo Molnar 已提交
107
 * dqonoff_mutex.
L
Linus Torvalds 已提交
108 109 110 111
 * 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
112
 * Each dquot has its dq_lock mutex. Locked dquots might not be referenced
L
Linus Torvalds 已提交
113 114 115 116 117 118 119 120 121
 * 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
122 123 124
 *   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
125 126 127 128 129 130 131 132 133 134
 */

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 */
135
static struct kmem_cache *dquot_cachep;
L
Linus Torvalds 已提交
136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253

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)
{
254
	list_add_tail(&dquot->dq_free, &free_dquots);
L
Linus Torvalds 已提交
255 256 257 258 259 260 261 262 263 264 265 266 267 268 269
	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 */
270
	list_add_tail(&dquot->dq_inuse, &inuse_list);
L
Linus Torvalds 已提交
271 272 273 274 275 276 277 278 279 280 281 282 283 284
	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
285 286
	mutex_lock(&dquot->dq_lock);
	mutex_unlock(&dquot->dq_lock);
L
Linus Torvalds 已提交
287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324
}

#define mark_dquot_dirty(dquot) ((dquot)->dq_sb->dq_op->mark_dirty(dquot))

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
325 326
	mutex_lock(&dquot->dq_lock);
	mutex_lock(&dqopt->dqio_mutex);
L
Linus Torvalds 已提交
327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346
	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
347 348
	mutex_unlock(&dqopt->dqio_mutex);
	mutex_unlock(&dquot->dq_lock);
L
Linus Torvalds 已提交
349 350 351 352 353 354 355 356 357 358 359
	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
360
	mutex_lock(&dqopt->dqio_mutex);
L
Linus Torvalds 已提交
361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376
	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
377
	mutex_unlock(&dqopt->dqio_mutex);
L
Linus Torvalds 已提交
378 379 380 381 382 383 384 385 386 387 388
	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
389
	mutex_lock(&dquot->dq_lock);
L
Linus Torvalds 已提交
390 391 392
	/* 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
393
	mutex_lock(&dqopt->dqio_mutex);
L
Linus Torvalds 已提交
394 395 396 397 398 399 400 401 402
	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
403
	mutex_unlock(&dqopt->dqio_mutex);
L
Linus Torvalds 已提交
404
out_dqlock:
I
Ingo Molnar 已提交
405
	mutex_unlock(&dquot->dq_lock);
L
Linus Torvalds 已提交
406 407 408 409 410
	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
411 412 413 414
 * 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
415 416
static void invalidate_dquots(struct super_block *sb, int type)
{
417
	struct dquot *dquot, *tmp;
L
Linus Torvalds 已提交
418

419
restart:
L
Linus Torvalds 已提交
420
	spin_lock(&dq_list_lock);
421
	list_for_each_entry_safe(dquot, tmp, &inuse_list, dq_inuse) {
L
Linus Torvalds 已提交
422 423 424 425
		if (dquot->dq_sb != sb)
			continue;
		if (dquot->dq_type != type)
			continue;
426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453
		/* 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
454 455 456 457 458 459 460 461 462 463 464 465 466 467 468
		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
469
	mutex_lock(&dqopt->dqonoff_mutex);
L
Linus Torvalds 已提交
470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503
	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)) {
			dquot = list_entry(dirty->next, struct dquot, dq_dirty);
			/* 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
504
	mutex_unlock(&dqopt->dqonoff_mutex);
L
Linus Torvalds 已提交
505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531

	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
532
static int shrink_dqcache_memory(int nr, gfp_t gfp_mask)
L
Linus Torvalds 已提交
533 534 535 536 537 538 539 540 541 542 543 544
{
	if (nr) {
		spin_lock(&dq_list_lock);
		prune_dqcache(nr);
		spin_unlock(&dq_list_lock);
	}
	return (dqstats.free_dquots / 100) * sysctl_vfs_cache_pressure;
}

/*
 * Put reference to dquot
 * NOTE: If you change this function please check whether dqput_blocks() works right...
I
Ingo Molnar 已提交
545
 * MUST be called with either dqptr_sem or dqonoff_mutex held
L
Linus Torvalds 已提交
546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569
 */
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);
570 571 572 573
		/* 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
574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593
		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 */
594
	BUG_ON(!list_empty(&dquot->dq_free));
L
Linus Torvalds 已提交
595 596 597 598 599 600 601 602 603
#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;

604
	dquot = kmem_cache_zalloc(dquot_cachep, GFP_NOFS);
L
Linus Torvalds 已提交
605 606 607
	if(!dquot)
		return NODQUOT;

I
Ingo Molnar 已提交
608
	mutex_init(&dquot->dq_lock);
L
Linus Torvalds 已提交
609 610 611 612
	INIT_LIST_HEAD(&dquot->dq_free);
	INIT_LIST_HEAD(&dquot->dq_inuse);
	INIT_HLIST_NODE(&dquot->dq_hash);
	INIT_LIST_HEAD(&dquot->dq_dirty);
613
	init_waitqueue_head(&dquot->dq_wait_unused);
L
Linus Torvalds 已提交
614 615 616 617 618 619 620 621 622
	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
623
 * MUST be called with either dqptr_sem or dqonoff_mutex held
L
Linus Torvalds 已提交
624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667
 */
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
668
	BUG_ON(!dquot->dq_sb);	/* Has somebody invalidated entry under us? */
L
Linus Torvalds 已提交
669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687
#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
688
/* This routine is guarded by dqonoff_mutex mutex */
L
Linus Torvalds 已提交
689 690
static void add_dquot_ref(struct super_block *sb, int type)
{
691
	struct inode *inode;
L
Linus Torvalds 已提交
692 693

restart:
694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709
	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);

		sb->dq_op->initialize(inode, type);
		iput(inode);
		/* As we may have blocked we had better restart... */
		goto restart;
L
Linus Torvalds 已提交
710
	}
711
	spin_unlock(&inode_lock);
L
Linus Torvalds 已提交
712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762
}

/* 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... */
int remove_inode_dquot_ref(struct inode *inode, int type, struct list_head *tofree_head)
{
	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);
	}
}

763 764 765 766 767 768 769 770 771 772 773 774 775
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
776 777 778 779 780
/* Gather all references from inodes and drop them */
static void drop_dquot_ref(struct super_block *sb, int type)
{
	LIST_HEAD(tofree_head);

781 782 783 784 785 786
	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
787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849
}

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);
}

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;
}

/* Values of warnings */
#define NOWARN 0
#define IHARDWARN 1
#define ISOFTLONGWARN 2
#define ISOFTWARN 3
#define BHARDWARN 4
#define BSOFTLONGWARN 5
#define BSOFTWARN 6

/* Print warning to user which exceeded quota */
static void print_warning(struct dquot *dquot, const char warntype)
{
	char *msg = NULL;
850
	struct tty_struct *tty;
L
Linus Torvalds 已提交
851 852 853 854 855 856
	int flag = (warntype == BHARDWARN || warntype == BSOFTLONGWARN) ? DQ_BLKS_B :
	  ((warntype == IHARDWARN || warntype == ISOFTLONGWARN) ? DQ_INODES_B : 0);

	if (!need_print_warning(dquot) || (flag && test_and_set_bit(flag, &dquot->dq_flags)))
		return;

857
	mutex_lock(&tty_mutex);
858 859
	tty = get_current_tty();
	if (!tty)
860
		goto out_lock;
861
	tty_write_message(tty, dquot->dq_sb->s_id);
L
Linus Torvalds 已提交
862
	if (warntype == ISOFTWARN || warntype == BSOFTWARN)
863
		tty_write_message(tty, ": warning, ");
L
Linus Torvalds 已提交
864
	else
865 866
		tty_write_message(tty, ": write failed, ");
	tty_write_message(tty, quotatypes[dquot->dq_type]);
L
Linus Torvalds 已提交
867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886
	switch (warntype) {
		case IHARDWARN:
			msg = " file limit reached.\r\n";
			break;
		case ISOFTLONGWARN:
			msg = " file quota exceeded too long.\r\n";
			break;
		case ISOFTWARN:
			msg = " file quota exceeded.\r\n";
			break;
		case BHARDWARN:
			msg = " block limit reached.\r\n";
			break;
		case BSOFTLONGWARN:
			msg = " block quota exceeded too long.\r\n";
			break;
		case BSOFTWARN:
			msg = " block quota exceeded.\r\n";
			break;
	}
887
	tty_write_message(tty, msg);
888 889
out_lock:
	mutex_unlock(&tty_mutex);
L
Linus Torvalds 已提交
890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991
}

static inline void flush_warnings(struct dquot **dquots, char *warntype)
{
	int i;

	for (i = 0; i < MAXQUOTAS; i++)
		if (dquots[i] != NODQUOT && warntype[i] != NOWARN)
			print_warning(dquots[i], warntype[i]);
}

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)
{
	*warntype = NOWARN;
	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)) {
		*warntype = IHARDWARN;
		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)) {
		*warntype = ISOFTLONGWARN;
		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) {
		*warntype = ISOFTWARN;
		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)
{
	*warntype = 0;
	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)
			*warntype = BHARDWARN;
		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)
			*warntype = BSOFTLONGWARN;
		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) {
			*warntype = BSOFTWARN;
			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
992 993
	/* 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
994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055
	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
1056 1057
	/* 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
1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104
	if (IS_NOQUOTA(inode)) {
out_add:
		inode_add_bytes(inode, number);
		return QUOTA_OK;
	}
	for (cnt = 0; cnt < MAXQUOTAS; cnt++)
		warntype[cnt] = NOWARN;

	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
1105 1106
	/* 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
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
	if (IS_NOQUOTA(inode))
		return QUOTA_OK;
	for (cnt = 0; cnt < MAXQUOTAS; cnt++)
		warntype[cnt] = NOWARN;
	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]);
	flush_warnings((struct dquot **)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_free_space(struct inode *inode, qsize_t number)
{
	unsigned int cnt;

I
Ingo Molnar 已提交
1149 1150
	/* 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
1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184
	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
1185 1186
	/* 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
1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224
	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
1225 1226
	/* 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
1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 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 1288 1289 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
	if (IS_NOQUOTA(inode))
		return QUOTA_OK;
	/* Clear the arrays */
	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
		transfer_to[cnt] = transfer_from[cnt] = NODQUOT;
		warntype[cnt] = NOWARN;
	}
	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
1320
	mutex_lock(&dqopt->dqio_mutex);
L
Linus Torvalds 已提交
1321
	ret = dqopt->ops[type]->write_file_info(sb, type);
I
Ingo Molnar 已提交
1322
	mutex_unlock(&dqopt->dqio_mutex);
L
Linus Torvalds 已提交
1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 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
	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
1378
	mutex_lock(&dqopt->dqonoff_mutex);
L
Linus Torvalds 已提交
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
	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
1407
	mutex_unlock(&dqopt->dqonoff_mutex);
L
Linus Torvalds 已提交
1408
	/* Sync the superblock so that buffers with quota data are written to
1409
	 * disk (and so userspace sees correct data afterwards). */
L
Linus Torvalds 已提交
1410 1411 1412 1413 1414 1415 1416 1417 1418 1419
	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
1420
			mutex_lock(&dqopt->dqonoff_mutex);
L
Linus Torvalds 已提交
1421 1422 1423
			/* If quota was reenabled in the meantime, we have
			 * nothing to do */
			if (!sb_has_quota_enabled(sb, cnt)) {
1424
				mutex_lock(&toputinode[cnt]->i_mutex);
L
Linus Torvalds 已提交
1425 1426 1427
				toputinode[cnt]->i_flags &= ~(S_IMMUTABLE |
				  S_NOATIME | S_NOQUOTA);
				truncate_inode_pages(&toputinode[cnt]->i_data, 0);
1428
				mutex_unlock(&toputinode[cnt]->i_mutex);
L
Linus Torvalds 已提交
1429 1430 1431
				mark_inode_dirty(toputinode[cnt]);
				iput(toputinode[cnt]);
			}
I
Ingo Molnar 已提交
1432
			mutex_unlock(&dqopt->dqonoff_mutex);
L
Linus Torvalds 已提交
1433 1434
		}
	if (sb->s_bdev)
1435
		invalidate_bdev(sb->s_bdev);
L
Linus Torvalds 已提交
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
	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 */
1471
	invalidate_bdev(sb->s_bdev);
1472
	mutex_lock(&inode->i_mutex);
I
Ingo Molnar 已提交
1473
	mutex_lock(&dqopt->dqonoff_mutex);
L
Linus Torvalds 已提交
1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484
	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);
1485
	sb->dq_op->drop(inode);
L
Linus Torvalds 已提交
1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497

	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
1498
	mutex_lock(&dqopt->dqio_mutex);
L
Linus Torvalds 已提交
1499
	if ((error = dqopt->ops[type]->read_file_info(sb, type)) < 0) {
I
Ingo Molnar 已提交
1500
		mutex_unlock(&dqopt->dqio_mutex);
L
Linus Torvalds 已提交
1501 1502
		goto out_file_init;
	}
I
Ingo Molnar 已提交
1503
	mutex_unlock(&dqopt->dqio_mutex);
1504
	mutex_unlock(&inode->i_mutex);
L
Linus Torvalds 已提交
1505 1506 1507
	set_enable_flags(dqopt, type);

	add_dquot_ref(sb, type);
I
Ingo Molnar 已提交
1508
	mutex_unlock(&dqopt->dqonoff_mutex);
L
Linus Torvalds 已提交
1509 1510 1511 1512 1513 1514 1515

	return 0;

out_file_init:
	dqopt->files[type] = NULL;
	iput(inode);
out_lock:
I
Ingo Molnar 已提交
1516
	mutex_unlock(&dqopt->dqonoff_mutex);
L
Linus Torvalds 已提交
1517 1518 1519 1520 1521 1522 1523 1524
	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);
	}
1525
	mutex_unlock(&inode->i_mutex);
L
Linus Torvalds 已提交
1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546
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;
	error = security_quota_on(nd.dentry);
	if (error)
		goto out_path;
	/* Quota file not on the same filesystem? */
	if (nd.mnt->mnt_sb != sb)
		error = -EXDEV;
1547
	else
L
Linus Torvalds 已提交
1548 1549 1550 1551 1552 1553 1554 1555 1556 1557
		error = vfs_quota_on_inode(nd.dentry->d_inode, type, format_id);
out_path:
	path_release(&nd);
	return error;
}

/*
 * This function is used when filesystem needs to initialize quotas
 * during mount time.
 */
1558 1559
int vfs_quota_on_mount(struct super_block *sb, char *qf_name,
		int format_id, int type)
L
Linus Torvalds 已提交
1560
{
1561
	struct dentry *dentry;
L
Linus Torvalds 已提交
1562 1563
	int error;

1564
	dentry = lookup_one_len(qf_name, sb->s_root, strlen(qf_name));
1565 1566 1567
	if (IS_ERR(dentry))
		return PTR_ERR(dentry);

J
Jan Kara 已提交
1568 1569 1570 1571 1572
	if (!dentry->d_inode) {
		error = -ENOENT;
		goto out;
	}

L
Linus Torvalds 已提交
1573
	error = security_quota_on(dentry);
1574 1575 1576
	if (!error)
		error = vfs_quota_on_inode(dentry->d_inode, type, format_id);

J
Jan Kara 已提交
1577
out:
1578 1579
	dput(dentry);
	return error;
L
Linus Torvalds 已提交
1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603
}

/* 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
1604
	mutex_lock(&sb_dqopt(sb)->dqonoff_mutex);
L
Linus Torvalds 已提交
1605
	if (!(dquot = dqget(sb, id, type))) {
I
Ingo Molnar 已提交
1606
		mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
L
Linus Torvalds 已提交
1607 1608 1609 1610
		return -ESRCH;
	}
	do_get_dqblk(dquot, di);
	dqput(dquot);
I
Ingo Molnar 已提交
1611
	mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
L
Linus Torvalds 已提交
1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672
	return 0;
}

/* Generic routine for setting common part of quota structure */
static void do_set_dqblk(struct dquot *dquot, struct if_dqblk *di)
{
	struct mem_dqblk *dm = &dquot->dq_dqb;
	int check_blim = 0, check_ilim = 0;

	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... */
			dm->dqb_btime = get_seconds() + sb_dqopt(dquot->dq_sb)->info[dquot->dq_type].dqi_bgrace;
	}
	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... */
			dm->dqb_itime = get_seconds() + sb_dqopt(dquot->dq_sb)->info[dquot->dq_type].dqi_igrace;
	}
	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);
}

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

I
Ingo Molnar 已提交
1673
	mutex_lock(&sb_dqopt(sb)->dqonoff_mutex);
L
Linus Torvalds 已提交
1674
	if (!(dquot = dqget(sb, id, type))) {
I
Ingo Molnar 已提交
1675
		mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
L
Linus Torvalds 已提交
1676 1677 1678 1679
		return -ESRCH;
	}
	do_set_dqblk(dquot, di);
	dqput(dquot);
I
Ingo Molnar 已提交
1680
	mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
L
Linus Torvalds 已提交
1681 1682 1683 1684 1685 1686 1687 1688
	return 0;
}

/* 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
1689
	mutex_lock(&sb_dqopt(sb)->dqonoff_mutex);
L
Linus Torvalds 已提交
1690
	if (!sb_has_quota_enabled(sb, type)) {
I
Ingo Molnar 已提交
1691
		mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
L
Linus Torvalds 已提交
1692 1693 1694 1695 1696 1697 1698 1699 1700
		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
1701
	mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
L
Linus Torvalds 已提交
1702 1703 1704 1705 1706 1707 1708 1709
	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
1710
	mutex_lock(&sb_dqopt(sb)->dqonoff_mutex);
L
Linus Torvalds 已提交
1711
	if (!sb_has_quota_enabled(sb, type)) {
I
Ingo Molnar 已提交
1712
		mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
L
Linus Torvalds 已提交
1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726
		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
1727
	mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
L
Linus Torvalds 已提交
1728 1729 1730 1731 1732 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 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843
	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,
	},
	{
		.ctl_name	= FS_DQ_WARNINGS,
		.procname	= "warnings",
		.data		= &flag_print_warnings,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= &proc_dointvec,
	},
	{ .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__);

1844
	register_sysctl_table(sys_table);
L
Linus Torvalds 已提交
1845 1846 1847

	dquot_cachep = kmem_cache_create("dquot", 
			sizeof(struct dquot), sizeof(unsigned long) * 4,
1848 1849
			(SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|
				SLAB_MEM_SPREAD|SLAB_PANIC),
L
Linus Torvalds 已提交
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
			NULL, NULL);

	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));

	set_shrinker(DEFAULT_SEEKS, shrink_dqcache_memory);

	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);