super.c 101.2 KB
Newer Older
1
/*
2
 *  linux/fs/ext4/super.c
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
 *
 * Copyright (C) 1992, 1993, 1994, 1995
 * Remy Card (card@masi.ibp.fr)
 * Laboratoire MASI - Institut Blaise Pascal
 * Universite Pierre et Marie Curie (Paris VI)
 *
 *  from
 *
 *  linux/fs/minix/inode.c
 *
 *  Copyright (C) 1991, 1992  Linus Torvalds
 *
 *  Big-endian to little-endian byte-swapping/bitmaps by
 *        David S. Miller (davem@caip.rutgers.edu), 1995
 */

#include <linux/module.h>
#include <linux/string.h>
#include <linux/fs.h>
#include <linux/time.h>
23
#include <linux/jbd2.h>
24 25 26 27 28 29
#include <linux/slab.h>
#include <linux/init.h>
#include <linux/blkdev.h>
#include <linux/parser.h>
#include <linux/smp_lock.h>
#include <linux/buffer_head.h>
30
#include <linux/exportfs.h>
31 32 33 34 35 36
#include <linux/vfs.h>
#include <linux/random.h>
#include <linux/mount.h>
#include <linux/namei.h>
#include <linux/quotaops.h>
#include <linux/seq_file.h>
37
#include <linux/proc_fs.h>
38
#include <linux/marker.h>
V
Vignesh Babu 已提交
39
#include <linux/log2.h>
A
Andreas Dilger 已提交
40
#include <linux/crc16.h>
41 42
#include <asm/uaccess.h>

43 44
#include "ext4.h"
#include "ext4_jbd2.h"
45 46 47
#include "xattr.h"
#include "acl.h"
#include "namei.h"
A
Andreas Dilger 已提交
48
#include "group.h"
49

50 51
struct proc_dir_entry *ext4_proc_root;

52
static int ext4_load_journal(struct super_block *, struct ext4_super_block *,
53
			     unsigned long journal_devnum);
54
static int ext4_create_journal(struct super_block *, struct ext4_super_block *,
55
			       unsigned int);
56 57 58 59 60 61
static void ext4_commit_super(struct super_block *sb,
			      struct ext4_super_block *es, int sync);
static void ext4_mark_recovery_complete(struct super_block *sb,
					struct ext4_super_block *es);
static void ext4_clear_journal_err(struct super_block *sb,
				   struct ext4_super_block *es);
62
static int ext4_sync_fs(struct super_block *sb, int wait);
63
static const char *ext4_decode_error(struct super_block *sb, int errno,
64
				     char nbuf[16]);
65 66
static int ext4_remount(struct super_block *sb, int *flags, char *data);
static int ext4_statfs(struct dentry *dentry, struct kstatfs *buf);
67
static void ext4_unlockfs(struct super_block *sb);
68
static void ext4_write_super(struct super_block *sb);
69
static void ext4_write_super_lockfs(struct super_block *sb);
70

L
Laurent Vivier 已提交
71

72 73
ext4_fsblk_t ext4_block_bitmap(struct super_block *sb,
			       struct ext4_group_desc *bg)
L
Laurent Vivier 已提交
74
{
75
	return le32_to_cpu(bg->bg_block_bitmap_lo) |
76
		(EXT4_DESC_SIZE(sb) >= EXT4_MIN_DESC_SIZE_64BIT ?
77
		(ext4_fsblk_t)le32_to_cpu(bg->bg_block_bitmap_hi) << 32 : 0);
L
Laurent Vivier 已提交
78 79
}

80 81
ext4_fsblk_t ext4_inode_bitmap(struct super_block *sb,
			       struct ext4_group_desc *bg)
L
Laurent Vivier 已提交
82
{
83
	return le32_to_cpu(bg->bg_inode_bitmap_lo) |
84
		(EXT4_DESC_SIZE(sb) >= EXT4_MIN_DESC_SIZE_64BIT ?
85
		(ext4_fsblk_t)le32_to_cpu(bg->bg_inode_bitmap_hi) << 32 : 0);
L
Laurent Vivier 已提交
86 87
}

88 89
ext4_fsblk_t ext4_inode_table(struct super_block *sb,
			      struct ext4_group_desc *bg)
L
Laurent Vivier 已提交
90
{
91
	return le32_to_cpu(bg->bg_inode_table_lo) |
92
		(EXT4_DESC_SIZE(sb) >= EXT4_MIN_DESC_SIZE_64BIT ?
93
		(ext4_fsblk_t)le32_to_cpu(bg->bg_inode_table_hi) << 32 : 0);
L
Laurent Vivier 已提交
94 95
}

96 97
void ext4_block_bitmap_set(struct super_block *sb,
			   struct ext4_group_desc *bg, ext4_fsblk_t blk)
L
Laurent Vivier 已提交
98
{
99
	bg->bg_block_bitmap_lo = cpu_to_le32((u32)blk);
100 101
	if (EXT4_DESC_SIZE(sb) >= EXT4_MIN_DESC_SIZE_64BIT)
		bg->bg_block_bitmap_hi = cpu_to_le32(blk >> 32);
L
Laurent Vivier 已提交
102 103
}

104 105
void ext4_inode_bitmap_set(struct super_block *sb,
			   struct ext4_group_desc *bg, ext4_fsblk_t blk)
L
Laurent Vivier 已提交
106
{
107
	bg->bg_inode_bitmap_lo  = cpu_to_le32((u32)blk);
108 109
	if (EXT4_DESC_SIZE(sb) >= EXT4_MIN_DESC_SIZE_64BIT)
		bg->bg_inode_bitmap_hi = cpu_to_le32(blk >> 32);
L
Laurent Vivier 已提交
110 111
}

112 113
void ext4_inode_table_set(struct super_block *sb,
			  struct ext4_group_desc *bg, ext4_fsblk_t blk)
L
Laurent Vivier 已提交
114
{
115
	bg->bg_inode_table_lo = cpu_to_le32((u32)blk);
116 117
	if (EXT4_DESC_SIZE(sb) >= EXT4_MIN_DESC_SIZE_64BIT)
		bg->bg_inode_table_hi = cpu_to_le32(blk >> 32);
L
Laurent Vivier 已提交
118 119
}

120
/*
121
 * Wrappers for jbd2_journal_start/end.
122 123 124 125 126 127
 *
 * The only special thing we need to do here is to make sure that all
 * journal_end calls result in the superblock being marked dirty, so
 * that sync() will call the filesystem's write_super callback if
 * appropriate.
 */
128
handle_t *ext4_journal_start_sb(struct super_block *sb, int nblocks)
129 130 131 132 133 134 135 136 137
{
	journal_t *journal;

	if (sb->s_flags & MS_RDONLY)
		return ERR_PTR(-EROFS);

	/* Special case here: if the journal has aborted behind our
	 * backs (eg. EIO in the commit thread), then we still need to
	 * take the FS itself readonly cleanly. */
138
	journal = EXT4_SB(sb)->s_journal;
139
	if (is_journal_aborted(journal)) {
140
		ext4_abort(sb, __func__,
141 142 143 144
			   "Detected aborted journal");
		return ERR_PTR(-EROFS);
	}

145
	return jbd2_journal_start(journal, nblocks);
146 147 148 149
}

/*
 * The only special thing we need to do here is to make sure that all
150
 * jbd2_journal_stop calls result in the superblock being marked dirty, so
151 152 153
 * that sync() will call the filesystem's write_super callback if
 * appropriate.
 */
154
int __ext4_journal_stop(const char *where, handle_t *handle)
155 156 157 158 159 160 161
{
	struct super_block *sb;
	int err;
	int rc;

	sb = handle->h_transaction->t_journal->j_private;
	err = handle->h_err;
162
	rc = jbd2_journal_stop(handle);
163 164 165 166

	if (!err)
		err = rc;
	if (err)
167
		__ext4_std_error(sb, where, err);
168 169 170
	return err;
}

171
void ext4_journal_abort_handle(const char *caller, const char *err_fn,
172 173 174
		struct buffer_head *bh, handle_t *handle, int err)
{
	char nbuf[16];
175
	const char *errstr = ext4_decode_error(NULL, err, nbuf);
176 177 178 179 180 181 182 183 184 185 186 187 188

	if (bh)
		BUFFER_TRACE(bh, "abort");

	if (!handle->h_err)
		handle->h_err = err;

	if (is_handle_aborted(handle))
		return;

	printk(KERN_ERR "%s: aborting transaction: %s in %s\n",
	       caller, errstr, err_fn);

189
	jbd2_journal_abort_handle(handle);
190 191 192 193 194 195
}

/* Deal with the reporting of failure conditions on a filesystem such as
 * inconsistencies detected or read IO failures.
 *
 * On ext2, we can store the error state of the filesystem in the
196
 * superblock.  That is not possible on ext4, because we may have other
197 198 199 200 201
 * write ordering constraints on the superblock which prevent us from
 * writing it out straight away; and given that the journal is about to
 * be aborted, we can't rely on the current, or future, transactions to
 * write out the superblock safely.
 *
202
 * We'll just use the jbd2_journal_abort() error code to record an error in
203 204 205 206
 * the journal instead.  On recovery, the journal will compain about
 * that error until we've noted it down and cleared it.
 */

207
static void ext4_handle_error(struct super_block *sb)
208
{
209
	struct ext4_super_block *es = EXT4_SB(sb)->s_es;
210

211 212
	EXT4_SB(sb)->s_mount_state |= EXT4_ERROR_FS;
	es->s_state |= cpu_to_le16(EXT4_ERROR_FS);
213 214 215 216

	if (sb->s_flags & MS_RDONLY)
		return;

217
	if (!test_opt(sb, ERRORS_CONT)) {
218
		journal_t *journal = EXT4_SB(sb)->s_journal;
219

220
		EXT4_SB(sb)->s_mount_opt |= EXT4_MOUNT_ABORT;
221
		if (journal)
222
			jbd2_journal_abort(journal, -EIO);
223
	}
224 225
	if (test_opt(sb, ERRORS_RO)) {
		printk(KERN_CRIT "Remounting filesystem read-only\n");
226 227
		sb->s_flags |= MS_RDONLY;
	}
228
	ext4_commit_super(sb, es, 1);
229
	if (test_opt(sb, ERRORS_PANIC))
230
		panic("EXT4-fs (device %s): panic forced after error\n",
231 232 233
			sb->s_id);
}

234 235
void ext4_error(struct super_block *sb, const char *function,
		const char *fmt, ...)
236 237 238 239
{
	va_list args;

	va_start(args, fmt);
240
	printk(KERN_CRIT "EXT4-fs error (device %s): %s: ", sb->s_id, function);
241 242 243 244
	vprintk(fmt, args);
	printk("\n");
	va_end(args);

245
	ext4_handle_error(sb);
246 247
}

248
static const char *ext4_decode_error(struct super_block *sb, int errno,
249 250 251 252 253 254 255 256 257 258 259 260
				     char nbuf[16])
{
	char *errstr = NULL;

	switch (errno) {
	case -EIO:
		errstr = "IO failure";
		break;
	case -ENOMEM:
		errstr = "Out of memory";
		break;
	case -EROFS:
261
		if (!sb || EXT4_SB(sb)->s_journal->j_flags & JBD2_ABORT)
262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280
			errstr = "Journal has aborted";
		else
			errstr = "Readonly filesystem";
		break;
	default:
		/* If the caller passed in an extra buffer for unknown
		 * errors, textualise them now.  Else we just return
		 * NULL. */
		if (nbuf) {
			/* Check for truncated error codes... */
			if (snprintf(nbuf, 16, "error %d", -errno) >= 0)
				errstr = nbuf;
		}
		break;
	}

	return errstr;
}

281
/* __ext4_std_error decodes expected errors from journaling functions
282 283
 * automatically and invokes the appropriate error response.  */

284
void __ext4_std_error(struct super_block *sb, const char *function, int errno)
285 286 287 288 289 290 291 292 293 294 295
{
	char nbuf[16];
	const char *errstr;

	/* Special case: if the error is EROFS, and we're not already
	 * inside a transaction, then there's really no point in logging
	 * an error. */
	if (errno == -EROFS && journal_current_handle() == NULL &&
	    (sb->s_flags & MS_RDONLY))
		return;

296
	errstr = ext4_decode_error(sb, errno, nbuf);
297 298
	printk(KERN_CRIT "EXT4-fs error (device %s) in %s: %s\n",
	       sb->s_id, function, errstr);
299

300
	ext4_handle_error(sb);
301 302 303
}

/*
304
 * ext4_abort is a much stronger failure handler than ext4_error.  The
305 306 307 308 309 310 311 312
 * abort function may be used to deal with unrecoverable failures such
 * as journal IO errors or ENOMEM at a critical moment in log management.
 *
 * We unconditionally force the filesystem into an ABORT|READONLY state,
 * unless the error response on the fs has been set to panic in which
 * case we take the easy way out and panic immediately.
 */

313 314
void ext4_abort(struct super_block *sb, const char *function,
		const char *fmt, ...)
315 316 317
{
	va_list args;

318
	printk(KERN_CRIT "ext4_abort called.\n");
319 320

	va_start(args, fmt);
321
	printk(KERN_CRIT "EXT4-fs error (device %s): %s: ", sb->s_id, function);
322 323 324 325 326
	vprintk(fmt, args);
	printk("\n");
	va_end(args);

	if (test_opt(sb, ERRORS_PANIC))
327
		panic("EXT4-fs panic from previous error\n");
328 329 330 331 332

	if (sb->s_flags & MS_RDONLY)
		return;

	printk(KERN_CRIT "Remounting filesystem read-only\n");
333
	EXT4_SB(sb)->s_mount_state |= EXT4_ERROR_FS;
334
	sb->s_flags |= MS_RDONLY;
335
	EXT4_SB(sb)->s_mount_opt |= EXT4_MOUNT_ABORT;
336 337
	if (EXT4_SB(sb)->s_journal)
		jbd2_journal_abort(EXT4_SB(sb)->s_journal, -EIO);
338 339
}

340 341
void ext4_warning(struct super_block *sb, const char *function,
		  const char *fmt, ...)
342 343 344 345
{
	va_list args;

	va_start(args, fmt);
346
	printk(KERN_WARNING "EXT4-fs warning (device %s): %s: ",
347 348 349 350 351 352
	       sb->s_id, function);
	vprintk(fmt, args);
	printk("\n");
	va_end(args);
}

353
void ext4_update_dynamic_rev(struct super_block *sb)
354
{
355
	struct ext4_super_block *es = EXT4_SB(sb)->s_es;
356

357
	if (le32_to_cpu(es->s_rev_level) > EXT4_GOOD_OLD_REV)
358 359
		return;

360
	ext4_warning(sb, __func__,
361 362
		     "updating to rev %d because of new feature flag, "
		     "running e2fsck is recommended",
363
		     EXT4_DYNAMIC_REV);
364

365 366 367
	es->s_first_ino = cpu_to_le32(EXT4_GOOD_OLD_FIRST_INO);
	es->s_inode_size = cpu_to_le16(EXT4_GOOD_OLD_INODE_SIZE);
	es->s_rev_level = cpu_to_le32(EXT4_DYNAMIC_REV);
368 369 370 371 372 373 374 375 376 377 378 379 380
	/* leave es->s_feature_*compat flags alone */
	/* es->s_uuid will be set by e2fsck if empty */

	/*
	 * The rest of the superblock fields should be zero, and if not it
	 * means they are likely already in use, so leave them alone.  We
	 * can leave it up to e2fsck to clean up any inconsistencies there.
	 */
}

/*
 * Open the external journal device
 */
381
static struct block_device *ext4_blkdev_get(dev_t dev)
382 383 384 385 386 387 388 389 390 391
{
	struct block_device *bdev;
	char b[BDEVNAME_SIZE];

	bdev = open_by_devnum(dev, FMODE_READ|FMODE_WRITE);
	if (IS_ERR(bdev))
		goto fail;
	return bdev;

fail:
392
	printk(KERN_ERR "EXT4: failed to open journal device %s: %ld\n",
393 394 395 396 397 398 399
			__bdevname(dev, b), PTR_ERR(bdev));
	return NULL;
}

/*
 * Release the journal device
 */
400
static int ext4_blkdev_put(struct block_device *bdev)
401 402
{
	bd_release(bdev);
A
Al Viro 已提交
403
	return blkdev_put(bdev, FMODE_READ|FMODE_WRITE);
404 405
}

406
static int ext4_blkdev_remove(struct ext4_sb_info *sbi)
407 408 409 410 411 412
{
	struct block_device *bdev;
	int ret = -ENODEV;

	bdev = sbi->journal_bdev;
	if (bdev) {
413
		ret = ext4_blkdev_put(bdev);
414 415 416 417 418 419 420
		sbi->journal_bdev = NULL;
	}
	return ret;
}

static inline struct inode *orphan_list_entry(struct list_head *l)
{
421
	return &list_entry(l, struct ext4_inode_info, i_orphan)->vfs_inode;
422 423
}

424
static void dump_orphan_list(struct super_block *sb, struct ext4_sb_info *sbi)
425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441
{
	struct list_head *l;

	printk(KERN_ERR "sb orphan head is %d\n",
	       le32_to_cpu(sbi->s_es->s_last_orphan));

	printk(KERN_ERR "sb_info orphan list:\n");
	list_for_each(l, &sbi->s_orphan) {
		struct inode *inode = orphan_list_entry(l);
		printk(KERN_ERR "  "
		       "inode %s:%lu at %p: mode %o, nlink %d, next %d\n",
		       inode->i_sb->s_id, inode->i_ino, inode,
		       inode->i_mode, inode->i_nlink,
		       NEXT_ORPHAN(inode));
	}
}

442
static void ext4_put_super(struct super_block *sb)
443
{
444 445
	struct ext4_sb_info *sbi = EXT4_SB(sb);
	struct ext4_super_block *es = sbi->s_es;
446
	int i, err;
447

448
	ext4_mb_release(sb);
A
Alex Tomas 已提交
449
	ext4_ext_release(sb);
450
	ext4_xattr_put_super(sb);
451
	err = jbd2_journal_destroy(sbi->s_journal);
452
	sbi->s_journal = NULL;
453 454 455
	if (err < 0)
		ext4_abort(sb, __func__, "Couldn't clean up the journal");

456
	if (!(sb->s_flags & MS_RDONLY)) {
457
		EXT4_CLEAR_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER);
458
		es->s_state = cpu_to_le16(sbi->s_mount_state);
459
		ext4_commit_super(sb, es, 1);
460
	}
461 462
	if (sbi->s_proc) {
		remove_proc_entry("inode_readahead_blks", sbi->s_proc);
463
		remove_proc_entry(sb->s_id, ext4_proc_root);
464
	}
465 466 467 468

	for (i = 0; i < sbi->s_gdb_count; i++)
		brelse(sbi->s_group_desc[i]);
	kfree(sbi->s_group_desc);
469
	kfree(sbi->s_flex_groups);
470 471 472
	percpu_counter_destroy(&sbi->s_freeblocks_counter);
	percpu_counter_destroy(&sbi->s_freeinodes_counter);
	percpu_counter_destroy(&sbi->s_dirs_counter);
473
	percpu_counter_destroy(&sbi->s_dirtyblocks_counter);
474 475 476 477 478 479 480 481 482 483 484 485 486 487
	brelse(sbi->s_sbh);
#ifdef CONFIG_QUOTA
	for (i = 0; i < MAXQUOTAS; i++)
		kfree(sbi->s_qf_names[i]);
#endif

	/* Debugging code just in case the in-memory inode orphan list
	 * isn't empty.  The on-disk one can be non-empty if we've
	 * detected an error and taken the fs readonly, but the
	 * in-memory list had better be clean by this point. */
	if (!list_empty(&sbi->s_orphan))
		dump_orphan_list(sb, sbi);
	J_ASSERT(list_empty(&sbi->s_orphan));

488
	invalidate_bdev(sb->s_bdev);
489 490 491 492 493 494 495
	if (sbi->journal_bdev && sbi->journal_bdev != sb->s_bdev) {
		/*
		 * Invalidate the journal device's buffers.  We don't want them
		 * floating about in memory - the physical journal device may
		 * hotswapped, and it breaks the `ro-after' testing code.
		 */
		sync_blockdev(sbi->journal_bdev);
496
		invalidate_bdev(sbi->journal_bdev);
497
		ext4_blkdev_remove(sbi);
498 499 500 501 502 503
	}
	sb->s_fs_info = NULL;
	kfree(sbi);
	return;
}

504
static struct kmem_cache *ext4_inode_cachep;
505 506 507 508

/*
 * Called inside transaction, so use GFP_NOFS
 */
509
static struct inode *ext4_alloc_inode(struct super_block *sb)
510
{
511
	struct ext4_inode_info *ei;
512

C
Christoph Lameter 已提交
513
	ei = kmem_cache_alloc(ext4_inode_cachep, GFP_NOFS);
514 515
	if (!ei)
		return NULL;
T
Theodore Ts'o 已提交
516
#ifdef CONFIG_EXT4_FS_POSIX_ACL
517 518
	ei->i_acl = EXT4_ACL_NOT_CACHED;
	ei->i_default_acl = EXT4_ACL_NOT_CACHED;
519 520
#endif
	ei->vfs_inode.i_version = 1;
521
	ei->vfs_inode.i_data.writeback_index = 0;
A
Alex Tomas 已提交
522
	memset(&ei->i_cached_extent, 0, sizeof(struct ext4_ext_cache));
523 524
	INIT_LIST_HEAD(&ei->i_prealloc_list);
	spin_lock_init(&ei->i_prealloc_lock);
525
	jbd2_journal_init_jbd_inode(&ei->jinode, &ei->vfs_inode);
526 527 528 529 530
	ei->i_reserved_data_blocks = 0;
	ei->i_reserved_meta_blocks = 0;
	ei->i_allocated_meta_blocks = 0;
	ei->i_delalloc_reserved_flag = 0;
	spin_lock_init(&(ei->i_block_reservation_lock));
531 532 533
	return &ei->vfs_inode;
}

534
static void ext4_destroy_inode(struct inode *inode)
535
{
536 537 538 539 540 541 542 543
	if (!list_empty(&(EXT4_I(inode)->i_orphan))) {
		printk("EXT4 Inode %p: orphan list check failed!\n",
			EXT4_I(inode));
		print_hex_dump(KERN_INFO, "", DUMP_PREFIX_ADDRESS, 16, 4,
				EXT4_I(inode), sizeof(struct ext4_inode_info),
				true);
		dump_stack();
	}
544
	kmem_cache_free(ext4_inode_cachep, EXT4_I(inode));
545 546
}

547
static void init_once(void *foo)
548
{
549
	struct ext4_inode_info *ei = (struct ext4_inode_info *) foo;
550

C
Christoph Lameter 已提交
551
	INIT_LIST_HEAD(&ei->i_orphan);
T
Theodore Ts'o 已提交
552
#ifdef CONFIG_EXT4_FS_XATTR
C
Christoph Lameter 已提交
553
	init_rwsem(&ei->xattr_sem);
554
#endif
555
	init_rwsem(&ei->i_data_sem);
C
Christoph Lameter 已提交
556
	inode_init_once(&ei->vfs_inode);
557 558 559 560
}

static int init_inodecache(void)
{
561 562
	ext4_inode_cachep = kmem_cache_create("ext4_inode_cache",
					     sizeof(struct ext4_inode_info),
563 564
					     0, (SLAB_RECLAIM_ACCOUNT|
						SLAB_MEM_SPREAD),
565
					     init_once);
566
	if (ext4_inode_cachep == NULL)
567 568 569 570 571 572
		return -ENOMEM;
	return 0;
}

static void destroy_inodecache(void)
{
573
	kmem_cache_destroy(ext4_inode_cachep);
574 575
}

576
static void ext4_clear_inode(struct inode *inode)
577
{
T
Theodore Ts'o 已提交
578
#ifdef CONFIG_EXT4_FS_POSIX_ACL
579 580 581 582 583 584 585 586 587
	if (EXT4_I(inode)->i_acl &&
			EXT4_I(inode)->i_acl != EXT4_ACL_NOT_CACHED) {
		posix_acl_release(EXT4_I(inode)->i_acl);
		EXT4_I(inode)->i_acl = EXT4_ACL_NOT_CACHED;
	}
	if (EXT4_I(inode)->i_default_acl &&
			EXT4_I(inode)->i_default_acl != EXT4_ACL_NOT_CACHED) {
		posix_acl_release(EXT4_I(inode)->i_default_acl);
		EXT4_I(inode)->i_default_acl = EXT4_ACL_NOT_CACHED;
588 589
	}
#endif
590
	ext4_discard_preallocations(inode);
591 592
	jbd2_journal_release_jbd_inode(EXT4_SB(inode->i_sb)->s_journal,
				       &EXT4_I(inode)->jinode);
593 594
}

595 596
static inline void ext4_show_quota_options(struct seq_file *seq,
					   struct super_block *sb)
597 598
{
#if defined(CONFIG_QUOTA)
599
	struct ext4_sb_info *sbi = EXT4_SB(sb);
600 601 602

	if (sbi->s_jquota_fmt)
		seq_printf(seq, ",jqfmt=%s",
603
		(sbi->s_jquota_fmt == QFMT_VFS_OLD) ? "vfsold" : "vfsv0");
604 605 606 607 608 609 610

	if (sbi->s_qf_names[USRQUOTA])
		seq_printf(seq, ",usrjquota=%s", sbi->s_qf_names[USRQUOTA]);

	if (sbi->s_qf_names[GRPQUOTA])
		seq_printf(seq, ",grpjquota=%s", sbi->s_qf_names[GRPQUOTA]);

611
	if (sbi->s_mount_opt & EXT4_MOUNT_USRQUOTA)
612 613
		seq_puts(seq, ",usrquota");

614
	if (sbi->s_mount_opt & EXT4_MOUNT_GRPQUOTA)
615 616 617 618
		seq_puts(seq, ",grpquota");
#endif
}

M
Miklos Szeredi 已提交
619 620 621 622 623
/*
 * Show an option if
 *  - it's set to a non-default value OR
 *  - if the per-sb default is different from the global default
 */
624
static int ext4_show_options(struct seq_file *seq, struct vfsmount *vfs)
625
{
626 627
	int def_errors;
	unsigned long def_mount_opts;
628
	struct super_block *sb = vfs->mnt_sb;
M
Miklos Szeredi 已提交
629 630 631 632
	struct ext4_sb_info *sbi = EXT4_SB(sb);
	struct ext4_super_block *es = sbi->s_es;

	def_mount_opts = le32_to_cpu(es->s_default_mount_opts);
633
	def_errors     = le16_to_cpu(es->s_errors);
M
Miklos Szeredi 已提交
634 635 636 637 638

	if (sbi->s_sb_block != 1)
		seq_printf(seq, ",sb=%llu", sbi->s_sb_block);
	if (test_opt(sb, MINIX_DF))
		seq_puts(seq, ",minixdf");
639
	if (test_opt(sb, GRPID) && !(def_mount_opts & EXT4_DEFM_BSDGROUPS))
M
Miklos Szeredi 已提交
640 641 642 643 644 645 646 647 648 649 650
		seq_puts(seq, ",grpid");
	if (!test_opt(sb, GRPID) && (def_mount_opts & EXT4_DEFM_BSDGROUPS))
		seq_puts(seq, ",nogrpid");
	if (sbi->s_resuid != EXT4_DEF_RESUID ||
	    le16_to_cpu(es->s_def_resuid) != EXT4_DEF_RESUID) {
		seq_printf(seq, ",resuid=%u", sbi->s_resuid);
	}
	if (sbi->s_resgid != EXT4_DEF_RESGID ||
	    le16_to_cpu(es->s_def_resgid) != EXT4_DEF_RESGID) {
		seq_printf(seq, ",resgid=%u", sbi->s_resgid);
	}
651
	if (test_opt(sb, ERRORS_RO)) {
M
Miklos Szeredi 已提交
652
		if (def_errors == EXT4_ERRORS_PANIC ||
653 654
		    def_errors == EXT4_ERRORS_CONTINUE) {
			seq_puts(seq, ",errors=remount-ro");
M
Miklos Szeredi 已提交
655 656
		}
	}
657
	if (test_opt(sb, ERRORS_CONT) && def_errors != EXT4_ERRORS_CONTINUE)
658
		seq_puts(seq, ",errors=continue");
659
	if (test_opt(sb, ERRORS_PANIC) && def_errors != EXT4_ERRORS_PANIC)
M
Miklos Szeredi 已提交
660
		seq_puts(seq, ",errors=panic");
661
	if (test_opt(sb, NO_UID32) && !(def_mount_opts & EXT4_DEFM_UID16))
M
Miklos Szeredi 已提交
662
		seq_puts(seq, ",nouid32");
663
	if (test_opt(sb, DEBUG) && !(def_mount_opts & EXT4_DEFM_DEBUG))
M
Miklos Szeredi 已提交
664 665 666
		seq_puts(seq, ",debug");
	if (test_opt(sb, OLDALLOC))
		seq_puts(seq, ",oldalloc");
T
Theodore Ts'o 已提交
667
#ifdef CONFIG_EXT4_FS_XATTR
668 669
	if (test_opt(sb, XATTR_USER) &&
		!(def_mount_opts & EXT4_DEFM_XATTR_USER))
M
Miklos Szeredi 已提交
670 671 672 673 674 675
		seq_puts(seq, ",user_xattr");
	if (!test_opt(sb, XATTR_USER) &&
	    (def_mount_opts & EXT4_DEFM_XATTR_USER)) {
		seq_puts(seq, ",nouser_xattr");
	}
#endif
T
Theodore Ts'o 已提交
676
#ifdef CONFIG_EXT4_FS_POSIX_ACL
677
	if (test_opt(sb, POSIX_ACL) && !(def_mount_opts & EXT4_DEFM_ACL))
M
Miklos Szeredi 已提交
678 679 680 681 682 683 684 685 686 687
		seq_puts(seq, ",acl");
	if (!test_opt(sb, POSIX_ACL) && (def_mount_opts & EXT4_DEFM_ACL))
		seq_puts(seq, ",noacl");
#endif
	if (!test_opt(sb, RESERVATION))
		seq_puts(seq, ",noreservation");
	if (sbi->s_commit_interval) {
		seq_printf(seq, ",commit=%u",
			   (unsigned) (sbi->s_commit_interval / HZ));
	}
E
Eric Sandeen 已提交
688 689 690 691 692 693 694
	/*
	 * We're changing the default of barrier mount option, so
	 * let's always display its mount state so it's clear what its
	 * status is.
	 */
	seq_puts(seq, ",barrier=");
	seq_puts(seq, test_opt(sb, BARRIER) ? "1" : "0");
695 696
	if (test_opt(sb, JOURNAL_ASYNC_COMMIT))
		seq_puts(seq, ",journal_async_commit");
M
Miklos Szeredi 已提交
697 698 699 700
	if (test_opt(sb, NOBH))
		seq_puts(seq, ",nobh");
	if (!test_opt(sb, EXTENTS))
		seq_puts(seq, ",noextents");
701 702
	if (test_opt(sb, I_VERSION))
		seq_puts(seq, ",i_version");
703 704 705
	if (!test_opt(sb, DELALLOC))
		seq_puts(seq, ",nodelalloc");

706

707 708
	if (sbi->s_stripe)
		seq_printf(seq, ",stripe=%lu", sbi->s_stripe);
709 710 711 712
	/*
	 * journal mode get enabled in different ways
	 * So just print the value even if we didn't specify it
	 */
713
	if (test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_JOURNAL_DATA)
714
		seq_puts(seq, ",data=journal");
715
	else if (test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_ORDERED_DATA)
716
		seq_puts(seq, ",data=ordered");
717
	else if (test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_WRITEBACK_DATA)
718 719
		seq_puts(seq, ",data=writeback");

720 721 722 723
	if (sbi->s_inode_readahead_blks != EXT4_DEF_INODE_READAHEAD_BLKS)
		seq_printf(seq, ",inode_readahead_blks=%u",
			   sbi->s_inode_readahead_blks);

724 725 726
	if (test_opt(sb, DATA_ERR_ABORT))
		seq_puts(seq, ",data_err=abort");

727
	ext4_show_quota_options(seq, sb);
728 729 730 731
	return 0;
}


C
Christoph Hellwig 已提交
732 733
static struct inode *ext4_nfs_get_inode(struct super_block *sb,
		u64 ino, u32 generation)
734 735 736
{
	struct inode *inode;

737
	if (ino < EXT4_FIRST_INO(sb) && ino != EXT4_ROOT_INO)
738
		return ERR_PTR(-ESTALE);
739
	if (ino > le32_to_cpu(EXT4_SB(sb)->s_es->s_inodes_count))
740 741 742 743
		return ERR_PTR(-ESTALE);

	/* iget isn't really right if the inode is currently unallocated!!
	 *
744
	 * ext4_read_inode will return a bad_inode if the inode had been
745 746 747 748 749
	 * deleted, so we should be safe.
	 *
	 * Currently we don't know the generation for parent directory, so
	 * a generation of 0 means "accept any"
	 */
750 751 752 753
	inode = ext4_iget(sb, ino);
	if (IS_ERR(inode))
		return ERR_CAST(inode);
	if (generation && inode->i_generation != generation) {
754 755 756
		iput(inode);
		return ERR_PTR(-ESTALE);
	}
C
Christoph Hellwig 已提交
757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772

	return inode;
}

static struct dentry *ext4_fh_to_dentry(struct super_block *sb, struct fid *fid,
		int fh_len, int fh_type)
{
	return generic_fh_to_dentry(sb, fid, fh_len, fh_type,
				    ext4_nfs_get_inode);
}

static struct dentry *ext4_fh_to_parent(struct super_block *sb, struct fid *fid,
		int fh_len, int fh_type)
{
	return generic_fh_to_parent(sb, fid, fh_len, fh_type,
				    ext4_nfs_get_inode);
773 774 775
}

#ifdef CONFIG_QUOTA
776
#define QTYPE2NAME(t) ((t) == USRQUOTA ? "user" : "group")
777
#define QTYPE2MOPT(on, t) ((t) == USRQUOTA?((on)##USRJQUOTA):((on)##GRPJQUOTA))
778

779 780 781 782 783 784 785
static int ext4_dquot_initialize(struct inode *inode, int type);
static int ext4_dquot_drop(struct inode *inode);
static int ext4_write_dquot(struct dquot *dquot);
static int ext4_acquire_dquot(struct dquot *dquot);
static int ext4_release_dquot(struct dquot *dquot);
static int ext4_mark_dquot_dirty(struct dquot *dquot);
static int ext4_write_info(struct super_block *sb, int type);
786 787
static int ext4_quota_on(struct super_block *sb, int type, int format_id,
				char *path, int remount);
788 789
static int ext4_quota_on_mount(struct super_block *sb, int type);
static ssize_t ext4_quota_read(struct super_block *sb, int type, char *data,
790
			       size_t len, loff_t off);
791
static ssize_t ext4_quota_write(struct super_block *sb, int type,
792 793
				const char *data, size_t len, loff_t off);

794 795 796
static struct dquot_operations ext4_quota_operations = {
	.initialize	= ext4_dquot_initialize,
	.drop		= ext4_dquot_drop,
797 798 799 800 801
	.alloc_space	= dquot_alloc_space,
	.alloc_inode	= dquot_alloc_inode,
	.free_space	= dquot_free_space,
	.free_inode	= dquot_free_inode,
	.transfer	= dquot_transfer,
802 803 804 805
	.write_dquot	= ext4_write_dquot,
	.acquire_dquot	= ext4_acquire_dquot,
	.release_dquot	= ext4_release_dquot,
	.mark_dirty	= ext4_mark_dquot_dirty,
806 807 808
	.write_info	= ext4_write_info,
	.alloc_dquot	= dquot_alloc,
	.destroy_dquot	= dquot_destroy,
809 810
};

811 812
static struct quotactl_ops ext4_qctl_operations = {
	.quota_on	= ext4_quota_on,
813 814 815 816 817 818 819 820 821
	.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
};
#endif

822
static const struct super_operations ext4_sops = {
823 824 825 826 827 828 829 830 831 832 833 834 835 836
	.alloc_inode	= ext4_alloc_inode,
	.destroy_inode	= ext4_destroy_inode,
	.write_inode	= ext4_write_inode,
	.dirty_inode	= ext4_dirty_inode,
	.delete_inode	= ext4_delete_inode,
	.put_super	= ext4_put_super,
	.write_super	= ext4_write_super,
	.sync_fs	= ext4_sync_fs,
	.write_super_lockfs = ext4_write_super_lockfs,
	.unlockfs	= ext4_unlockfs,
	.statfs		= ext4_statfs,
	.remount_fs	= ext4_remount,
	.clear_inode	= ext4_clear_inode,
	.show_options	= ext4_show_options,
837
#ifdef CONFIG_QUOTA
838 839
	.quota_read	= ext4_quota_read,
	.quota_write	= ext4_quota_write,
840 841 842
#endif
};

843
static const struct export_operations ext4_export_ops = {
C
Christoph Hellwig 已提交
844 845
	.fh_to_dentry = ext4_fh_to_dentry,
	.fh_to_parent = ext4_fh_to_parent,
846
	.get_parent = ext4_get_parent,
847 848 849 850 851
};

enum {
	Opt_bsd_df, Opt_minix_df, Opt_grpid, Opt_nogrpid,
	Opt_resgid, Opt_resuid, Opt_sb, Opt_err_cont, Opt_err_panic, Opt_err_ro,
852
	Opt_nouid32, Opt_debug, Opt_oldalloc, Opt_orlov,
853 854 855
	Opt_user_xattr, Opt_nouser_xattr, Opt_acl, Opt_noacl,
	Opt_reservation, Opt_noreservation, Opt_noload, Opt_nobh, Opt_bh,
	Opt_commit, Opt_journal_update, Opt_journal_inum, Opt_journal_dev,
856
	Opt_journal_checksum, Opt_journal_async_commit,
857
	Opt_abort, Opt_data_journal, Opt_data_ordered, Opt_data_writeback,
858
	Opt_data_err_abort, Opt_data_err_ignore,
859 860 861
	Opt_usrjquota, Opt_grpjquota, Opt_offusrjquota, Opt_offgrpjquota,
	Opt_jqfmt_vfsold, Opt_jqfmt_vfsv0, Opt_quota, Opt_noquota,
	Opt_ignore, Opt_barrier, Opt_err, Opt_resize, Opt_usrquota,
862
	Opt_grpquota, Opt_extents, Opt_noextents, Opt_i_version,
863
	Opt_stripe, Opt_delalloc, Opt_nodelalloc,
864
	Opt_inode_readahead_blks
865 866
};

867
static const match_table_t tokens = {
868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896
	{Opt_bsd_df, "bsddf"},
	{Opt_minix_df, "minixdf"},
	{Opt_grpid, "grpid"},
	{Opt_grpid, "bsdgroups"},
	{Opt_nogrpid, "nogrpid"},
	{Opt_nogrpid, "sysvgroups"},
	{Opt_resgid, "resgid=%u"},
	{Opt_resuid, "resuid=%u"},
	{Opt_sb, "sb=%u"},
	{Opt_err_cont, "errors=continue"},
	{Opt_err_panic, "errors=panic"},
	{Opt_err_ro, "errors=remount-ro"},
	{Opt_nouid32, "nouid32"},
	{Opt_debug, "debug"},
	{Opt_oldalloc, "oldalloc"},
	{Opt_orlov, "orlov"},
	{Opt_user_xattr, "user_xattr"},
	{Opt_nouser_xattr, "nouser_xattr"},
	{Opt_acl, "acl"},
	{Opt_noacl, "noacl"},
	{Opt_reservation, "reservation"},
	{Opt_noreservation, "noreservation"},
	{Opt_noload, "noload"},
	{Opt_nobh, "nobh"},
	{Opt_bh, "bh"},
	{Opt_commit, "commit=%u"},
	{Opt_journal_update, "journal=update"},
	{Opt_journal_inum, "journal=%u"},
	{Opt_journal_dev, "journal_dev=%u"},
897 898
	{Opt_journal_checksum, "journal_checksum"},
	{Opt_journal_async_commit, "journal_async_commit"},
899 900 901 902
	{Opt_abort, "abort"},
	{Opt_data_journal, "data=journal"},
	{Opt_data_ordered, "data=ordered"},
	{Opt_data_writeback, "data=writeback"},
903 904
	{Opt_data_err_abort, "data_err=abort"},
	{Opt_data_err_ignore, "data_err=ignore"},
905 906 907 908 909 910 911 912 913 914 915
	{Opt_offusrjquota, "usrjquota="},
	{Opt_usrjquota, "usrjquota=%s"},
	{Opt_offgrpjquota, "grpjquota="},
	{Opt_grpjquota, "grpjquota=%s"},
	{Opt_jqfmt_vfsold, "jqfmt=vfsold"},
	{Opt_jqfmt_vfsv0, "jqfmt=vfsv0"},
	{Opt_grpquota, "grpquota"},
	{Opt_noquota, "noquota"},
	{Opt_quota, "quota"},
	{Opt_usrquota, "usrquota"},
	{Opt_barrier, "barrier=%u"},
A
Alex Tomas 已提交
916
	{Opt_extents, "extents"},
M
Mingming Cao 已提交
917
	{Opt_noextents, "noextents"},
918
	{Opt_i_version, "i_version"},
919
	{Opt_stripe, "stripe=%u"},
920
	{Opt_resize, "resize"},
921
	{Opt_delalloc, "delalloc"},
922
	{Opt_nodelalloc, "nodelalloc"},
923
	{Opt_inode_readahead_blks, "inode_readahead_blks=%u"},
J
Josef Bacik 已提交
924
	{Opt_err, NULL},
925 926
};

927
static ext4_fsblk_t get_sb_block(void **data)
928
{
929
	ext4_fsblk_t	sb_block;
930 931 932 933 934
	char		*options = (char *) *data;

	if (!options || strncmp(options, "sb=", 3) != 0)
		return 1;	/* Default location */
	options += 3;
935
	/*todo: use simple_strtoll with >32bit ext4 */
936 937
	sb_block = simple_strtoul(options, &options, 0);
	if (*options && *options != ',') {
938
		printk(KERN_ERR "EXT4-fs: Invalid sb specification: %s\n",
939 940 941 942 943 944 945 946 947
		       (char *) *data);
		return 1;
	}
	if (*options == ',')
		options++;
	*data = (void *) options;
	return sb_block;
}

948 949 950
static int parse_options(char *options, struct super_block *sb,
			 unsigned int *inum, unsigned long *journal_devnum,
			 ext4_fsblk_t *n_blocks_count, int is_remount)
951
{
952
	struct ext4_sb_info *sbi = EXT4_SB(sb);
953
	char *p;
954 955 956 957
	substring_t args[MAX_OPT_ARGS];
	int data_opt = 0;
	int option;
#ifdef CONFIG_QUOTA
958
	int qtype, qfmt;
959 960
	char *qname;
#endif
961
	ext4_fsblk_t last_block;
962 963 964 965

	if (!options)
		return 1;

966
	while ((p = strsep(&options, ",")) != NULL) {
967 968 969 970 971 972 973
		int token;
		if (!*p)
			continue;

		token = match_token(p, tokens, args);
		switch (token) {
		case Opt_bsd_df:
974
			clear_opt(sbi->s_mount_opt, MINIX_DF);
975 976
			break;
		case Opt_minix_df:
977
			set_opt(sbi->s_mount_opt, MINIX_DF);
978 979
			break;
		case Opt_grpid:
980
			set_opt(sbi->s_mount_opt, GRPID);
981 982
			break;
		case Opt_nogrpid:
983
			clear_opt(sbi->s_mount_opt, GRPID);
984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999
			break;
		case Opt_resuid:
			if (match_int(&args[0], &option))
				return 0;
			sbi->s_resuid = option;
			break;
		case Opt_resgid:
			if (match_int(&args[0], &option))
				return 0;
			sbi->s_resgid = option;
			break;
		case Opt_sb:
			/* handled by get_sb_block() instead of here */
			/* *sb_block = match_int(&args[0]); */
			break;
		case Opt_err_panic:
1000 1001 1002
			clear_opt(sbi->s_mount_opt, ERRORS_CONT);
			clear_opt(sbi->s_mount_opt, ERRORS_RO);
			set_opt(sbi->s_mount_opt, ERRORS_PANIC);
1003 1004
			break;
		case Opt_err_ro:
1005 1006 1007
			clear_opt(sbi->s_mount_opt, ERRORS_CONT);
			clear_opt(sbi->s_mount_opt, ERRORS_PANIC);
			set_opt(sbi->s_mount_opt, ERRORS_RO);
1008 1009
			break;
		case Opt_err_cont:
1010 1011 1012
			clear_opt(sbi->s_mount_opt, ERRORS_RO);
			clear_opt(sbi->s_mount_opt, ERRORS_PANIC);
			set_opt(sbi->s_mount_opt, ERRORS_CONT);
1013 1014
			break;
		case Opt_nouid32:
1015
			set_opt(sbi->s_mount_opt, NO_UID32);
1016 1017
			break;
		case Opt_debug:
1018
			set_opt(sbi->s_mount_opt, DEBUG);
1019 1020
			break;
		case Opt_oldalloc:
1021
			set_opt(sbi->s_mount_opt, OLDALLOC);
1022 1023
			break;
		case Opt_orlov:
1024
			clear_opt(sbi->s_mount_opt, OLDALLOC);
1025
			break;
T
Theodore Ts'o 已提交
1026
#ifdef CONFIG_EXT4_FS_XATTR
1027
		case Opt_user_xattr:
1028
			set_opt(sbi->s_mount_opt, XATTR_USER);
1029 1030
			break;
		case Opt_nouser_xattr:
1031
			clear_opt(sbi->s_mount_opt, XATTR_USER);
1032 1033 1034 1035
			break;
#else
		case Opt_user_xattr:
		case Opt_nouser_xattr:
1036 1037
			printk(KERN_ERR "EXT4 (no)user_xattr options "
			       "not supported\n");
1038 1039
			break;
#endif
T
Theodore Ts'o 已提交
1040
#ifdef CONFIG_EXT4_FS_POSIX_ACL
1041 1042 1043 1044 1045 1046 1047 1048 1049
		case Opt_acl:
			set_opt(sbi->s_mount_opt, POSIX_ACL);
			break;
		case Opt_noacl:
			clear_opt(sbi->s_mount_opt, POSIX_ACL);
			break;
#else
		case Opt_acl:
		case Opt_noacl:
1050 1051
			printk(KERN_ERR "EXT4 (no)acl options "
			       "not supported\n");
1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066
			break;
#endif
		case Opt_reservation:
			set_opt(sbi->s_mount_opt, RESERVATION);
			break;
		case Opt_noreservation:
			clear_opt(sbi->s_mount_opt, RESERVATION);
			break;
		case Opt_journal_update:
			/* @@@ FIXME */
			/* Eventually we will want to be able to create
			   a journal file here.  For now, only allow the
			   user to specify an existing inode to be the
			   journal file. */
			if (is_remount) {
1067
				printk(KERN_ERR "EXT4-fs: cannot specify "
1068 1069 1070
				       "journal on remount\n");
				return 0;
			}
1071
			set_opt(sbi->s_mount_opt, UPDATE_JOURNAL);
1072 1073 1074
			break;
		case Opt_journal_inum:
			if (is_remount) {
1075
				printk(KERN_ERR "EXT4-fs: cannot specify "
1076 1077 1078 1079 1080 1081 1082 1083 1084
				       "journal on remount\n");
				return 0;
			}
			if (match_int(&args[0], &option))
				return 0;
			*inum = option;
			break;
		case Opt_journal_dev:
			if (is_remount) {
1085
				printk(KERN_ERR "EXT4-fs: cannot specify "
1086 1087 1088 1089 1090 1091 1092
				       "journal on remount\n");
				return 0;
			}
			if (match_int(&args[0], &option))
				return 0;
			*journal_devnum = option;
			break;
1093 1094 1095 1096 1097 1098 1099
		case Opt_journal_checksum:
			set_opt(sbi->s_mount_opt, JOURNAL_CHECKSUM);
			break;
		case Opt_journal_async_commit:
			set_opt(sbi->s_mount_opt, JOURNAL_ASYNC_COMMIT);
			set_opt(sbi->s_mount_opt, JOURNAL_CHECKSUM);
			break;
1100
		case Opt_noload:
1101
			set_opt(sbi->s_mount_opt, NOLOAD);
1102 1103 1104 1105 1106 1107 1108
			break;
		case Opt_commit:
			if (match_int(&args[0], &option))
				return 0;
			if (option < 0)
				return 0;
			if (option == 0)
1109
				option = JBD2_DEFAULT_MAX_COMMIT_AGE;
1110 1111 1112
			sbi->s_commit_interval = HZ * option;
			break;
		case Opt_data_journal:
1113
			data_opt = EXT4_MOUNT_JOURNAL_DATA;
1114 1115
			goto datacheck;
		case Opt_data_ordered:
1116
			data_opt = EXT4_MOUNT_ORDERED_DATA;
1117 1118
			goto datacheck;
		case Opt_data_writeback:
1119
			data_opt = EXT4_MOUNT_WRITEBACK_DATA;
1120 1121
		datacheck:
			if (is_remount) {
1122
				if ((sbi->s_mount_opt & EXT4_MOUNT_DATA_FLAGS)
1123 1124
						!= data_opt) {
					printk(KERN_ERR
1125
						"EXT4-fs: cannot change data "
1126 1127 1128 1129
						"mode on remount\n");
					return 0;
				}
			} else {
1130
				sbi->s_mount_opt &= ~EXT4_MOUNT_DATA_FLAGS;
1131 1132 1133
				sbi->s_mount_opt |= data_opt;
			}
			break;
1134 1135 1136 1137 1138 1139
		case Opt_data_err_abort:
			set_opt(sbi->s_mount_opt, DATA_ERR_ABORT);
			break;
		case Opt_data_err_ignore:
			clear_opt(sbi->s_mount_opt, DATA_ERR_ABORT);
			break;
1140 1141 1142 1143 1144 1145 1146
#ifdef CONFIG_QUOTA
		case Opt_usrjquota:
			qtype = USRQUOTA;
			goto set_qf_name;
		case Opt_grpjquota:
			qtype = GRPQUOTA;
set_qf_name:
1147
			if (sb_any_quota_loaded(sb) &&
1148
			    !sbi->s_qf_names[qtype]) {
1149
				printk(KERN_ERR
1150 1151
				       "EXT4-fs: Cannot change journaled "
				       "quota options when quota turned on.\n");
1152 1153 1154 1155 1156
				return 0;
			}
			qname = match_strdup(&args[0]);
			if (!qname) {
				printk(KERN_ERR
1157
					"EXT4-fs: not enough memory for "
1158 1159 1160 1161 1162 1163
					"storing quotafile name.\n");
				return 0;
			}
			if (sbi->s_qf_names[qtype] &&
			    strcmp(sbi->s_qf_names[qtype], qname)) {
				printk(KERN_ERR
1164
					"EXT4-fs: %s quota file already "
1165 1166 1167 1168 1169 1170 1171
					"specified.\n", QTYPE2NAME(qtype));
				kfree(qname);
				return 0;
			}
			sbi->s_qf_names[qtype] = qname;
			if (strchr(sbi->s_qf_names[qtype], '/')) {
				printk(KERN_ERR
1172
					"EXT4-fs: quotafile must be on "
1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185
					"filesystem root.\n");
				kfree(sbi->s_qf_names[qtype]);
				sbi->s_qf_names[qtype] = NULL;
				return 0;
			}
			set_opt(sbi->s_mount_opt, QUOTA);
			break;
		case Opt_offusrjquota:
			qtype = USRQUOTA;
			goto clear_qf_name;
		case Opt_offgrpjquota:
			qtype = GRPQUOTA;
clear_qf_name:
1186
			if (sb_any_quota_loaded(sb) &&
1187
			    sbi->s_qf_names[qtype]) {
1188
				printk(KERN_ERR "EXT4-fs: Cannot change "
1189
					"journaled quota options when "
1190 1191 1192 1193 1194 1195 1196 1197 1198 1199
					"quota turned on.\n");
				return 0;
			}
			/*
			 * The space will be released later when all options
			 * are confirmed to be correct
			 */
			sbi->s_qf_names[qtype] = NULL;
			break;
		case Opt_jqfmt_vfsold:
1200 1201
			qfmt = QFMT_VFS_OLD;
			goto set_qf_format;
1202
		case Opt_jqfmt_vfsv0:
1203 1204
			qfmt = QFMT_VFS_V0;
set_qf_format:
1205
			if (sb_any_quota_loaded(sb) &&
1206 1207 1208 1209 1210 1211 1212
			    sbi->s_jquota_fmt != qfmt) {
				printk(KERN_ERR "EXT4-fs: Cannot change "
					"journaled quota options when "
					"quota turned on.\n");
				return 0;
			}
			sbi->s_jquota_fmt = qfmt;
1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223
			break;
		case Opt_quota:
		case Opt_usrquota:
			set_opt(sbi->s_mount_opt, QUOTA);
			set_opt(sbi->s_mount_opt, USRQUOTA);
			break;
		case Opt_grpquota:
			set_opt(sbi->s_mount_opt, QUOTA);
			set_opt(sbi->s_mount_opt, GRPQUOTA);
			break;
		case Opt_noquota:
1224
			if (sb_any_quota_loaded(sb)) {
1225
				printk(KERN_ERR "EXT4-fs: Cannot change quota "
1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236
					"options when quota turned on.\n");
				return 0;
			}
			clear_opt(sbi->s_mount_opt, QUOTA);
			clear_opt(sbi->s_mount_opt, USRQUOTA);
			clear_opt(sbi->s_mount_opt, GRPQUOTA);
			break;
#else
		case Opt_quota:
		case Opt_usrquota:
		case Opt_grpquota:
1237 1238 1239
			printk(KERN_ERR
				"EXT4-fs: quota options not supported.\n");
			break;
1240 1241 1242 1243 1244 1245 1246
		case Opt_usrjquota:
		case Opt_grpjquota:
		case Opt_offusrjquota:
		case Opt_offgrpjquota:
		case Opt_jqfmt_vfsold:
		case Opt_jqfmt_vfsv0:
			printk(KERN_ERR
1247
				"EXT4-fs: journaled quota options not "
1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267
				"supported.\n");
			break;
		case Opt_noquota:
			break;
#endif
		case Opt_abort:
			set_opt(sbi->s_mount_opt, ABORT);
			break;
		case Opt_barrier:
			if (match_int(&args[0], &option))
				return 0;
			if (option)
				set_opt(sbi->s_mount_opt, BARRIER);
			else
				clear_opt(sbi->s_mount_opt, BARRIER);
			break;
		case Opt_ignore:
			break;
		case Opt_resize:
			if (!is_remount) {
1268
				printk("EXT4-fs: resize option only available "
1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281
					"for remount\n");
				return 0;
			}
			if (match_int(&args[0], &option) != 0)
				return 0;
			*n_blocks_count = option;
			break;
		case Opt_nobh:
			set_opt(sbi->s_mount_opt, NOBH);
			break;
		case Opt_bh:
			clear_opt(sbi->s_mount_opt, NOBH);
			break;
A
Alex Tomas 已提交
1282
		case Opt_extents:
1283 1284 1285 1286 1287 1288 1289
			if (!EXT4_HAS_INCOMPAT_FEATURE(sb,
					EXT4_FEATURE_INCOMPAT_EXTENTS)) {
				ext4_warning(sb, __func__,
					"extents feature not enabled "
					"on this filesystem, use tune2fs\n");
				return 0;
			}
1290
			set_opt(sbi->s_mount_opt, EXTENTS);
A
Alex Tomas 已提交
1291
			break;
M
Mingming Cao 已提交
1292
		case Opt_noextents:
1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306
			/*
			 * When e2fsprogs support resizing an already existing
			 * ext3 file system to greater than 2**32 we need to
			 * add support to block allocator to handle growing
			 * already existing block  mapped inode so that blocks
			 * allocated for them fall within 2**32
			 */
			last_block = ext4_blocks_count(sbi->s_es) - 1;
			if (last_block  > 0xffffffffULL) {
				printk(KERN_ERR "EXT4-fs: Filesystem too "
						"large to mount with "
						"-o noextents options\n");
				return 0;
			}
1307
			clear_opt(sbi->s_mount_opt, EXTENTS);
M
Mingming Cao 已提交
1308
			break;
1309 1310 1311 1312
		case Opt_i_version:
			set_opt(sbi->s_mount_opt, I_VERSION);
			sb->s_flags |= MS_I_VERSION;
			break;
1313 1314 1315
		case Opt_nodelalloc:
			clear_opt(sbi->s_mount_opt, DELALLOC);
			break;
1316 1317 1318 1319 1320 1321 1322
		case Opt_stripe:
			if (match_int(&args[0], &option))
				return 0;
			if (option < 0)
				return 0;
			sbi->s_stripe = option;
			break;
1323 1324 1325
		case Opt_delalloc:
			set_opt(sbi->s_mount_opt, DELALLOC);
			break;
1326 1327 1328 1329 1330 1331 1332
		case Opt_inode_readahead_blks:
			if (match_int(&args[0], &option))
				return 0;
			if (option < 0 || option > (1 << 30))
				return 0;
			sbi->s_inode_readahead_blks = option;
			break;
1333
		default:
1334 1335 1336
			printk(KERN_ERR
			       "EXT4-fs: Unrecognized mount option \"%s\" "
			       "or missing value\n", p);
1337 1338 1339 1340 1341
			return 0;
		}
	}
#ifdef CONFIG_QUOTA
	if (sbi->s_qf_names[USRQUOTA] || sbi->s_qf_names[GRPQUOTA]) {
1342
		if ((sbi->s_mount_opt & EXT4_MOUNT_USRQUOTA) &&
1343 1344 1345
		     sbi->s_qf_names[USRQUOTA])
			clear_opt(sbi->s_mount_opt, USRQUOTA);

1346
		if ((sbi->s_mount_opt & EXT4_MOUNT_GRPQUOTA) &&
1347 1348 1349 1350
		     sbi->s_qf_names[GRPQUOTA])
			clear_opt(sbi->s_mount_opt, GRPQUOTA);

		if ((sbi->s_qf_names[USRQUOTA] &&
1351
				(sbi->s_mount_opt & EXT4_MOUNT_GRPQUOTA)) ||
1352
		    (sbi->s_qf_names[GRPQUOTA] &&
1353 1354
				(sbi->s_mount_opt & EXT4_MOUNT_USRQUOTA))) {
			printk(KERN_ERR "EXT4-fs: old and new quota "
1355 1356 1357 1358 1359
					"format mixing.\n");
			return 0;
		}

		if (!sbi->s_jquota_fmt) {
1360
			printk(KERN_ERR "EXT4-fs: journaled quota format "
1361 1362 1363 1364 1365
					"not specified.\n");
			return 0;
		}
	} else {
		if (sbi->s_jquota_fmt) {
1366 1367
			printk(KERN_ERR "EXT4-fs: journaled quota format "
					"specified with no journaling "
1368 1369 1370 1371 1372 1373 1374 1375
					"enabled.\n");
			return 0;
		}
	}
#endif
	return 1;
}

1376
static int ext4_setup_super(struct super_block *sb, struct ext4_super_block *es,
1377 1378
			    int read_only)
{
1379
	struct ext4_sb_info *sbi = EXT4_SB(sb);
1380 1381
	int res = 0;

1382
	if (le32_to_cpu(es->s_rev_level) > EXT4_MAX_SUPP_REV) {
1383 1384
		printk(KERN_ERR "EXT4-fs warning: revision level too high, "
		       "forcing read-only mode\n");
1385 1386 1387 1388
		res = MS_RDONLY;
	}
	if (read_only)
		return res;
1389
	if (!(sbi->s_mount_state & EXT4_VALID_FS))
1390 1391
		printk(KERN_WARNING "EXT4-fs warning: mounting unchecked fs, "
		       "running e2fsck is recommended\n");
1392
	else if ((sbi->s_mount_state & EXT4_ERROR_FS))
1393 1394 1395
		printk(KERN_WARNING
		       "EXT4-fs warning: mounting fs with errors, "
		       "running e2fsck is recommended\n");
1396 1397 1398
	else if ((__s16) le16_to_cpu(es->s_max_mnt_count) >= 0 &&
		 le16_to_cpu(es->s_mnt_count) >=
		 (unsigned short) (__s16) le16_to_cpu(es->s_max_mnt_count))
1399 1400 1401
		printk(KERN_WARNING
		       "EXT4-fs warning: maximal mount count reached, "
		       "running e2fsck is recommended\n");
1402 1403 1404
	else if (le32_to_cpu(es->s_checkinterval) &&
		(le32_to_cpu(es->s_lastcheck) +
			le32_to_cpu(es->s_checkinterval) <= get_seconds()))
1405 1406 1407
		printk(KERN_WARNING
		       "EXT4-fs warning: checktime reached, "
		       "running e2fsck is recommended\n");
1408 1409
#if 0
		/* @@@ We _will_ want to clear the valid bit if we find
A
Andrew Morton 已提交
1410 1411 1412 1413
		 * inconsistencies, to force a fsck at reboot.  But for
		 * a plain journaled filesystem we can keep it set as
		 * valid forever! :)
		 */
1414
	es->s_state &= cpu_to_le16(~EXT4_VALID_FS);
1415 1416
#endif
	if (!(__s16) le16_to_cpu(es->s_max_mnt_count))
1417
		es->s_max_mnt_count = cpu_to_le16(EXT4_DFL_MAX_MNT_COUNT);
M
Marcin Slusarz 已提交
1418
	le16_add_cpu(&es->s_mnt_count, 1);
1419
	es->s_mtime = cpu_to_le32(get_seconds());
1420 1421
	ext4_update_dynamic_rev(sb);
	EXT4_SET_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER);
1422

1423
	ext4_commit_super(sb, es, 1);
1424
	if (test_opt(sb, DEBUG))
1425
		printk(KERN_INFO "[EXT4 FS bs=%lu, gc=%lu, "
1426 1427 1428
				"bpg=%lu, ipg=%lu, mo=%04lx]\n",
			sb->s_blocksize,
			sbi->s_groups_count,
1429 1430
			EXT4_BLOCKS_PER_GROUP(sb),
			EXT4_INODES_PER_GROUP(sb),
1431 1432
			sbi->s_mount_opt);

1433 1434 1435
	printk(KERN_INFO "EXT4 FS on %s, %s journal on %s\n",
	       sb->s_id, EXT4_SB(sb)->s_journal->j_inode ? "internal" :
	       "external", EXT4_SB(sb)->s_journal->j_devname);
1436 1437 1438
	return res;
}

1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457
static int ext4_fill_flex_info(struct super_block *sb)
{
	struct ext4_sb_info *sbi = EXT4_SB(sb);
	struct ext4_group_desc *gdp = NULL;
	struct buffer_head *bh;
	ext4_group_t flex_group_count;
	ext4_group_t flex_group;
	int groups_per_flex = 0;
	__u64 block_bitmap = 0;
	int i;

	if (!sbi->s_es->s_log_groups_per_flex) {
		sbi->s_log_groups_per_flex = 0;
		return 1;
	}

	sbi->s_log_groups_per_flex = sbi->s_es->s_log_groups_per_flex;
	groups_per_flex = 1 << sbi->s_log_groups_per_flex;

1458 1459
	/* We allocate both existing and potentially added groups */
	flex_group_count = ((sbi->s_groups_count + groups_per_flex - 1) +
1460 1461
			((le16_to_cpu(sbi->s_es->s_reserved_gdt_blocks) + 1) <<
			      EXT4_DESC_PER_BLOCK_BITS(sb))) / groups_per_flex;
1462
	sbi->s_flex_groups = kzalloc(flex_group_count *
1463 1464
				     sizeof(struct flex_groups), GFP_KERNEL);
	if (sbi->s_flex_groups == NULL) {
1465 1466
		printk(KERN_ERR "EXT4-fs: not enough memory for "
				"%lu flex groups\n", flex_group_count);
1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487
		goto failed;
	}

	gdp = ext4_get_group_desc(sb, 1, &bh);
	block_bitmap = ext4_block_bitmap(sb, gdp) - 1;

	for (i = 0; i < sbi->s_groups_count; i++) {
		gdp = ext4_get_group_desc(sb, i, &bh);

		flex_group = ext4_flex_group(sbi, i);
		sbi->s_flex_groups[flex_group].free_inodes +=
			le16_to_cpu(gdp->bg_free_inodes_count);
		sbi->s_flex_groups[flex_group].free_blocks +=
			le16_to_cpu(gdp->bg_free_blocks_count);
	}

	return 1;
failed:
	return 0;
}

A
Andreas Dilger 已提交
1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524
__le16 ext4_group_desc_csum(struct ext4_sb_info *sbi, __u32 block_group,
			    struct ext4_group_desc *gdp)
{
	__u16 crc = 0;

	if (sbi->s_es->s_feature_ro_compat &
	    cpu_to_le32(EXT4_FEATURE_RO_COMPAT_GDT_CSUM)) {
		int offset = offsetof(struct ext4_group_desc, bg_checksum);
		__le32 le_group = cpu_to_le32(block_group);

		crc = crc16(~0, sbi->s_es->s_uuid, sizeof(sbi->s_es->s_uuid));
		crc = crc16(crc, (__u8 *)&le_group, sizeof(le_group));
		crc = crc16(crc, (__u8 *)gdp, offset);
		offset += sizeof(gdp->bg_checksum); /* skip checksum */
		/* for checksum of struct ext4_group_desc do the rest...*/
		if ((sbi->s_es->s_feature_incompat &
		     cpu_to_le32(EXT4_FEATURE_INCOMPAT_64BIT)) &&
		    offset < le16_to_cpu(sbi->s_es->s_desc_size))
			crc = crc16(crc, (__u8 *)gdp + offset,
				    le16_to_cpu(sbi->s_es->s_desc_size) -
					offset);
	}

	return cpu_to_le16(crc);
}

int ext4_group_desc_csum_verify(struct ext4_sb_info *sbi, __u32 block_group,
				struct ext4_group_desc *gdp)
{
	if ((sbi->s_es->s_feature_ro_compat &
	     cpu_to_le32(EXT4_FEATURE_RO_COMPAT_GDT_CSUM)) &&
	    (gdp->bg_checksum != ext4_group_desc_csum(sbi, block_group, gdp)))
		return 0;

	return 1;
}

1525
/* Called at mount-time, super-block is locked */
1526
static int ext4_check_descriptors(struct super_block *sb)
1527
{
1528 1529 1530
	struct ext4_sb_info *sbi = EXT4_SB(sb);
	ext4_fsblk_t first_block = le32_to_cpu(sbi->s_es->s_first_data_block);
	ext4_fsblk_t last_block;
L
Laurent Vivier 已提交
1531 1532 1533
	ext4_fsblk_t block_bitmap;
	ext4_fsblk_t inode_bitmap;
	ext4_fsblk_t inode_table;
J
Jose R. Santos 已提交
1534
	int flexbg_flag = 0;
1535
	ext4_group_t i;
1536

J
Jose R. Santos 已提交
1537 1538 1539
	if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_FLEX_BG))
		flexbg_flag = 1;

1540
	ext4_debug("Checking group descriptors");
1541

1542 1543 1544
	for (i = 0; i < sbi->s_groups_count; i++) {
		struct ext4_group_desc *gdp = ext4_get_group_desc(sb, i, NULL);

J
Jose R. Santos 已提交
1545
		if (i == sbi->s_groups_count - 1 || flexbg_flag)
L
Laurent Vivier 已提交
1546
			last_block = ext4_blocks_count(sbi->s_es) - 1;
1547 1548
		else
			last_block = first_block +
1549
				(EXT4_BLOCKS_PER_GROUP(sb) - 1);
1550

1551
		block_bitmap = ext4_block_bitmap(sb, gdp);
1552
		if (block_bitmap < first_block || block_bitmap > last_block) {
1553 1554
			printk(KERN_ERR "EXT4-fs: ext4_check_descriptors: "
			       "Block bitmap for group %lu not in group "
1555
			       "(block %llu)!\n", i, block_bitmap);
1556 1557
			return 0;
		}
1558
		inode_bitmap = ext4_inode_bitmap(sb, gdp);
1559
		if (inode_bitmap < first_block || inode_bitmap > last_block) {
1560 1561
			printk(KERN_ERR "EXT4-fs: ext4_check_descriptors: "
			       "Inode bitmap for group %lu not in group "
1562
			       "(block %llu)!\n", i, inode_bitmap);
1563 1564
			return 0;
		}
1565
		inode_table = ext4_inode_table(sb, gdp);
L
Laurent Vivier 已提交
1566
		if (inode_table < first_block ||
1567
		    inode_table + sbi->s_itb_per_group - 1 > last_block) {
1568 1569
			printk(KERN_ERR "EXT4-fs: ext4_check_descriptors: "
			       "Inode table for group %lu not in group "
1570
			       "(block %llu)!\n", i, inode_table);
1571 1572
			return 0;
		}
1573
		spin_lock(sb_bgl_lock(sbi, i));
A
Andreas Dilger 已提交
1574
		if (!ext4_group_desc_csum_verify(sbi, i, gdp)) {
1575 1576 1577 1578
			printk(KERN_ERR "EXT4-fs: ext4_check_descriptors: "
			       "Checksum for group %lu failed (%u!=%u)\n",
			       i, le16_to_cpu(ext4_group_desc_csum(sbi, i,
			       gdp)), le16_to_cpu(gdp->bg_checksum));
1579 1580
			if (!(sb->s_flags & MS_RDONLY)) {
				spin_unlock(sb_bgl_lock(sbi, i));
1581
				return 0;
1582
			}
A
Andreas Dilger 已提交
1583
		}
1584
		spin_unlock(sb_bgl_lock(sbi, i));
J
Jose R. Santos 已提交
1585 1586
		if (!flexbg_flag)
			first_block += EXT4_BLOCKS_PER_GROUP(sb);
1587 1588
	}

L
Laurent Vivier 已提交
1589
	ext4_free_blocks_count_set(sbi->s_es, ext4_count_free_blocks(sb));
1590
	sbi->s_es->s_free_inodes_count = cpu_to_le32(ext4_count_free_inodes(sb));
1591 1592 1593
	return 1;
}

1594
/* ext4_orphan_cleanup() walks a singly-linked list of inodes (starting at
1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606
 * the superblock) which were deleted from all directories, but held open by
 * a process at the time of a crash.  We walk the list and try to delete these
 * inodes at recovery time (only with a read-write filesystem).
 *
 * In order to keep the orphan inode chain consistent during traversal (in
 * case of crash during recovery), we link each inode into the superblock
 * orphan list_head and handle it the same way as an inode deletion during
 * normal operation (which journals the operations for us).
 *
 * We only do an iget() and an iput() on each inode, which is very safe if we
 * accidentally point at an in-use or already deleted inode.  The worst that
 * can happen in this case is that we get a "bit already cleared" message from
1607
 * ext4_free_inode().  The only reason we would point at a wrong inode is if
1608 1609 1610
 * e2fsck was run on this filesystem, and it must have already done the orphan
 * inode cleanup for us, so we can safely abort without any further action.
 */
1611 1612
static void ext4_orphan_cleanup(struct super_block *sb,
				struct ext4_super_block *es)
1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623
{
	unsigned int s_flags = sb->s_flags;
	int nr_orphans = 0, nr_truncates = 0;
#ifdef CONFIG_QUOTA
	int i;
#endif
	if (!es->s_last_orphan) {
		jbd_debug(4, "no orphan inodes to clean up\n");
		return;
	}

1624 1625 1626 1627 1628 1629
	if (bdev_read_only(sb->s_bdev)) {
		printk(KERN_ERR "EXT4-fs: write access "
			"unavailable, skipping orphan cleanup.\n");
		return;
	}

1630
	if (EXT4_SB(sb)->s_mount_state & EXT4_ERROR_FS) {
1631 1632 1633 1634 1635 1636 1637 1638 1639
		if (es->s_last_orphan)
			jbd_debug(1, "Errors on filesystem, "
				  "clearing orphan list.\n");
		es->s_last_orphan = 0;
		jbd_debug(1, "Skipping orphan recovery on fs with errors.\n");
		return;
	}

	if (s_flags & MS_RDONLY) {
1640
		printk(KERN_INFO "EXT4-fs: %s: orphan cleanup on readonly fs\n",
1641 1642 1643 1644 1645 1646 1647 1648
		       sb->s_id);
		sb->s_flags &= ~MS_RDONLY;
	}
#ifdef CONFIG_QUOTA
	/* Needed for iput() to work correctly and not trash data */
	sb->s_flags |= MS_ACTIVE;
	/* Turn on quotas so that they are updated correctly */
	for (i = 0; i < MAXQUOTAS; i++) {
1649 1650
		if (EXT4_SB(sb)->s_qf_names[i]) {
			int ret = ext4_quota_on_mount(sb, i);
1651 1652
			if (ret < 0)
				printk(KERN_ERR
1653
					"EXT4-fs: Cannot turn on journaled "
1654 1655 1656 1657 1658 1659 1660 1661
					"quota: error %d\n", ret);
		}
	}
#endif

	while (es->s_last_orphan) {
		struct inode *inode;

1662 1663
		inode = ext4_orphan_get(sb, le32_to_cpu(es->s_last_orphan));
		if (IS_ERR(inode)) {
1664 1665 1666 1667
			es->s_last_orphan = 0;
			break;
		}

1668
		list_add(&EXT4_I(inode)->i_orphan, &EXT4_SB(sb)->s_orphan);
1669 1670 1671
		DQUOT_INIT(inode);
		if (inode->i_nlink) {
			printk(KERN_DEBUG
1672
				"%s: truncating inode %lu to %lld bytes\n",
1673
				__func__, inode->i_ino, inode->i_size);
1674
			jbd_debug(2, "truncating inode %lu to %lld bytes\n",
1675
				  inode->i_ino, inode->i_size);
1676
			ext4_truncate(inode);
1677 1678 1679 1680
			nr_truncates++;
		} else {
			printk(KERN_DEBUG
				"%s: deleting unreferenced inode %lu\n",
1681
				__func__, inode->i_ino);
1682 1683 1684 1685 1686 1687 1688
			jbd_debug(2, "deleting unreferenced inode %lu\n",
				  inode->i_ino);
			nr_orphans++;
		}
		iput(inode);  /* The delete magic happens here! */
	}

1689
#define PLURAL(x) (x), ((x) == 1) ? "" : "s"
1690 1691

	if (nr_orphans)
1692
		printk(KERN_INFO "EXT4-fs: %s: %d orphan inode%s deleted\n",
1693 1694
		       sb->s_id, PLURAL(nr_orphans));
	if (nr_truncates)
1695
		printk(KERN_INFO "EXT4-fs: %s: %d truncate%s cleaned up\n",
1696 1697 1698 1699 1700
		       sb->s_id, PLURAL(nr_truncates));
#ifdef CONFIG_QUOTA
	/* Turn quotas off */
	for (i = 0; i < MAXQUOTAS; i++) {
		if (sb_dqopt(sb)->files[i])
1701
			vfs_quota_off(sb, i, 0);
1702 1703 1704 1705
	}
#endif
	sb->s_flags = s_flags; /* Restore MS_RDONLY status */
}
1706 1707 1708 1709 1710 1711 1712 1713 1714
/*
 * Maximal extent format file size.
 * Resulting logical blkno at s_maxbytes must fit in our on-disk
 * extent format containers, within a sector_t, and within i_blocks
 * in the vfs.  ext4 inode has 48 bits of i_block in fsblock units,
 * so that won't be a limiting factor.
 *
 * Note, this does *not* consider any metadata overhead for vfs i_blocks.
 */
1715
static loff_t ext4_max_size(int blkbits, int has_huge_files)
1716 1717 1718 1719 1720
{
	loff_t res;
	loff_t upper_limit = MAX_LFS_FILESIZE;

	/* small i_blocks in vfs inode? */
1721
	if (!has_huge_files || sizeof(blkcnt_t) < sizeof(u64)) {
1722
		/*
J
Jens Axboe 已提交
1723
		 * CONFIG_LBD is not enabled implies the inode
1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744
		 * i_block represent total blocks in 512 bytes
		 * 32 == size of vfs inode i_blocks * 8
		 */
		upper_limit = (1LL << 32) - 1;

		/* total blocks in file system block size */
		upper_limit >>= (blkbits - 9);
		upper_limit <<= blkbits;
	}

	/* 32-bit extent-start container, ee_block */
	res = 1LL << 32;
	res <<= blkbits;
	res -= 1;

	/* Sanity check against vm- & vfs- imposed limits */
	if (res > upper_limit)
		res = upper_limit;

	return res;
}
1745 1746

/*
1747
 * Maximal bitmap file size.  There is a direct, and {,double-,triple-}indirect
1748 1749
 * block limit, and also a limit of (2^48 - 1) 512-byte sectors in i_blocks.
 * We need to be 1 filesystem block less than the 2^48 sector limit.
1750
 */
1751
static loff_t ext4_max_bitmap_size(int bits, int has_huge_files)
1752
{
1753
	loff_t res = EXT4_NDIR_BLOCKS;
1754 1755 1756
	int meta_blocks;
	loff_t upper_limit;
	/* This is calculated to be the largest file size for a
1757
	 * dense, bitmapped file such that the total number of
1758
	 * sectors in the file, including data and all indirect blocks,
1759 1760 1761 1762 1763
	 * does not exceed 2^48 -1
	 * __u32 i_blocks_lo and _u16 i_blocks_high representing the
	 * total number of  512 bytes blocks of the file
	 */

1764
	if (!has_huge_files || sizeof(blkcnt_t) < sizeof(u64)) {
1765
		/*
J
Jens Axboe 已提交
1766
		 * !has_huge_files or CONFIG_LBD is not enabled
1767 1768
		 * implies the inode i_block represent total blocks in
		 * 512 bytes 32 == size of vfs inode i_blocks * 8
1769 1770 1771 1772 1773 1774 1775
		 */
		upper_limit = (1LL << 32) - 1;

		/* total blocks in file system block size */
		upper_limit >>= (bits - 9);

	} else {
A
Aneesh Kumar K.V 已提交
1776 1777 1778 1779 1780 1781
		/*
		 * We use 48 bit ext4_inode i_blocks
		 * With EXT4_HUGE_FILE_FL set the i_blocks
		 * represent total number of blocks in
		 * file system block size
		 */
1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794
		upper_limit = (1LL << 48) - 1;

	}

	/* indirect blocks */
	meta_blocks = 1;
	/* double indirect blocks */
	meta_blocks += 1 + (1LL << (bits-2));
	/* tripple indirect blocks */
	meta_blocks += 1 + (1LL << (bits-2)) + (1LL << (2*(bits-2)));

	upper_limit -= meta_blocks;
	upper_limit <<= bits;
1795 1796 1797 1798 1799 1800 1801

	res += 1LL << (bits-2);
	res += 1LL << (2*(bits-2));
	res += 1LL << (3*(bits-2));
	res <<= bits;
	if (res > upper_limit)
		res = upper_limit;
1802 1803 1804 1805

	if (res > MAX_LFS_FILESIZE)
		res = MAX_LFS_FILESIZE;

1806 1807 1808
	return res;
}

1809
static ext4_fsblk_t descriptor_loc(struct super_block *sb,
1810
				ext4_fsblk_t logical_sb_block, int nr)
1811
{
1812
	struct ext4_sb_info *sbi = EXT4_SB(sb);
1813
	ext4_group_t bg, first_meta_bg;
1814 1815 1816 1817
	int has_super = 0;

	first_meta_bg = le32_to_cpu(sbi->s_es->s_first_meta_bg);

1818
	if (!EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_META_BG) ||
1819
	    nr < first_meta_bg)
1820
		return logical_sb_block + nr + 1;
1821
	bg = sbi->s_desc_per_block * nr;
1822
	if (ext4_bg_has_super(sb, bg))
1823
		has_super = 1;
1824
	return (has_super + ext4_group_first_block_no(sb, bg));
1825 1826
}

1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854
/**
 * ext4_get_stripe_size: Get the stripe size.
 * @sbi: In memory super block info
 *
 * If we have specified it via mount option, then
 * use the mount option value. If the value specified at mount time is
 * greater than the blocks per group use the super block value.
 * If the super block value is greater than blocks per group return 0.
 * Allocator needs it be less than blocks per group.
 *
 */
static unsigned long ext4_get_stripe_size(struct ext4_sb_info *sbi)
{
	unsigned long stride = le16_to_cpu(sbi->s_es->s_raid_stride);
	unsigned long stripe_width =
			le32_to_cpu(sbi->s_es->s_raid_stripe_width);

	if (sbi->s_stripe && sbi->s_stripe <= sbi->s_blocks_per_group)
		return sbi->s_stripe;

	if (stripe_width <= sbi->s_blocks_per_group)
		return stripe_width;

	if (stride <= sbi->s_blocks_per_group)
		return stride;

	return 0;
}
1855

1856
static int ext4_fill_super(struct super_block *sb, void *data, int silent)
A
Aneesh Kumar K.V 已提交
1857 1858
				__releases(kernel_lock)
				__acquires(kernel_lock)
A
Aneesh Kumar K.V 已提交
1859

1860
{
1861
	struct buffer_head *bh;
1862 1863 1864 1865
	struct ext4_super_block *es = NULL;
	struct ext4_sb_info *sbi;
	ext4_fsblk_t block;
	ext4_fsblk_t sb_block = get_sb_block(&data);
1866
	ext4_fsblk_t logical_sb_block;
1867 1868 1869 1870 1871
	unsigned long offset = 0;
	unsigned int journal_inum = 0;
	unsigned long journal_devnum = 0;
	unsigned long def_mount_opts;
	struct inode *root;
1872
	char *cp;
1873
	int ret = -EINVAL;
1874 1875 1876
	int blocksize;
	int db_count;
	int i;
1877
	int needs_recovery, has_huge_files;
1878
	__le32 features;
L
Laurent Vivier 已提交
1879
	__u64 blocks_count;
1880
	int err;
1881 1882 1883 1884 1885 1886

	sbi = kzalloc(sizeof(*sbi), GFP_KERNEL);
	if (!sbi)
		return -ENOMEM;
	sb->s_fs_info = sbi;
	sbi->s_mount_opt = 0;
1887 1888
	sbi->s_resuid = EXT4_DEF_RESUID;
	sbi->s_resgid = EXT4_DEF_RESGID;
1889
	sbi->s_inode_readahead_blks = EXT4_DEF_INODE_READAHEAD_BLKS;
M
Miklos Szeredi 已提交
1890
	sbi->s_sb_block = sb_block;
1891 1892 1893

	unlock_kernel();

1894 1895 1896 1897
	/* Cleanup superblock name */
	for (cp = sb->s_id; (cp = strchr(cp, '/'));)
		*cp = '!';

1898
	blocksize = sb_min_blocksize(sb, EXT4_MIN_BLOCK_SIZE);
1899
	if (!blocksize) {
1900
		printk(KERN_ERR "EXT4-fs: unable to set blocksize\n");
1901 1902 1903 1904
		goto out_fail;
	}

	/*
1905
	 * The ext4 superblock will not be buffer aligned for other than 1kB
1906 1907
	 * block sizes.  We need to calculate the offset from buffer start.
	 */
1908
	if (blocksize != EXT4_MIN_BLOCK_SIZE) {
1909 1910
		logical_sb_block = sb_block * EXT4_MIN_BLOCK_SIZE;
		offset = do_div(logical_sb_block, blocksize);
1911
	} else {
1912
		logical_sb_block = sb_block;
1913 1914
	}

1915
	if (!(bh = sb_bread(sb, logical_sb_block))) {
1916
		printk(KERN_ERR "EXT4-fs: unable to read superblock\n");
1917 1918 1919 1920
		goto out_fail;
	}
	/*
	 * Note: s_es must be initialized as soon as possible because
1921
	 *       some ext4 macro-instructions depend on its value
1922
	 */
1923
	es = (struct ext4_super_block *) (((char *)bh->b_data) + offset);
1924 1925
	sbi->s_es = es;
	sb->s_magic = le16_to_cpu(es->s_magic);
1926 1927
	if (sb->s_magic != EXT4_SUPER_MAGIC)
		goto cantfind_ext4;
1928 1929 1930

	/* Set defaults before we parse the mount options */
	def_mount_opts = le32_to_cpu(es->s_default_mount_opts);
1931
	if (def_mount_opts & EXT4_DEFM_DEBUG)
1932
		set_opt(sbi->s_mount_opt, DEBUG);
1933
	if (def_mount_opts & EXT4_DEFM_BSDGROUPS)
1934
		set_opt(sbi->s_mount_opt, GRPID);
1935
	if (def_mount_opts & EXT4_DEFM_UID16)
1936
		set_opt(sbi->s_mount_opt, NO_UID32);
T
Theodore Ts'o 已提交
1937
#ifdef CONFIG_EXT4_FS_XATTR
1938
	if (def_mount_opts & EXT4_DEFM_XATTR_USER)
1939
		set_opt(sbi->s_mount_opt, XATTR_USER);
1940
#endif
T
Theodore Ts'o 已提交
1941
#ifdef CONFIG_EXT4_FS_POSIX_ACL
1942
	if (def_mount_opts & EXT4_DEFM_ACL)
1943
		set_opt(sbi->s_mount_opt, POSIX_ACL);
1944
#endif
1945 1946 1947 1948 1949 1950 1951 1952
	if ((def_mount_opts & EXT4_DEFM_JMODE) == EXT4_DEFM_JMODE_DATA)
		sbi->s_mount_opt |= EXT4_MOUNT_JOURNAL_DATA;
	else if ((def_mount_opts & EXT4_DEFM_JMODE) == EXT4_DEFM_JMODE_ORDERED)
		sbi->s_mount_opt |= EXT4_MOUNT_ORDERED_DATA;
	else if ((def_mount_opts & EXT4_DEFM_JMODE) == EXT4_DEFM_JMODE_WBACK)
		sbi->s_mount_opt |= EXT4_MOUNT_WRITEBACK_DATA;

	if (le16_to_cpu(sbi->s_es->s_errors) == EXT4_ERRORS_PANIC)
1953
		set_opt(sbi->s_mount_opt, ERRORS_PANIC);
1954
	else if (le16_to_cpu(sbi->s_es->s_errors) == EXT4_ERRORS_CONTINUE)
D
Dmitry Mishin 已提交
1955
		set_opt(sbi->s_mount_opt, ERRORS_CONT);
1956 1957
	else
		set_opt(sbi->s_mount_opt, ERRORS_RO);
1958 1959 1960 1961 1962

	sbi->s_resuid = le16_to_cpu(es->s_def_resuid);
	sbi->s_resgid = le16_to_cpu(es->s_def_resgid);

	set_opt(sbi->s_mount_opt, RESERVATION);
E
Eric Sandeen 已提交
1963
	set_opt(sbi->s_mount_opt, BARRIER);
1964

M
Mingming Cao 已提交
1965 1966
	/*
	 * turn on extents feature by default in ext4 filesystem
1967 1968
	 * only if feature flag already set by mkfs or tune2fs.
	 * Use -o noextents to turn it off
M
Mingming Cao 已提交
1969
	 */
1970 1971 1972 1973 1974 1975
	if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_EXTENTS))
		set_opt(sbi->s_mount_opt, EXTENTS);
	else
		ext4_warning(sb, __func__,
			"extents feature not enabled on this filesystem, "
			"use tune2fs.\n");
M
Mingming Cao 已提交
1976

1977 1978 1979 1980 1981 1982 1983
	/*
	 * enable delayed allocation by default
	 * Use -o nodelalloc to turn it off
	 */
	set_opt(sbi->s_mount_opt, DELALLOC);


1984 1985
	if (!parse_options((char *) data, sb, &journal_inum, &journal_devnum,
			   NULL, 0))
1986 1987 1988
		goto failed_mount;

	sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
1989
		((sbi->s_mount_opt & EXT4_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0);
1990

1991 1992 1993 1994
	if (le32_to_cpu(es->s_rev_level) == EXT4_GOOD_OLD_REV &&
	    (EXT4_HAS_COMPAT_FEATURE(sb, ~0U) ||
	     EXT4_HAS_RO_COMPAT_FEATURE(sb, ~0U) ||
	     EXT4_HAS_INCOMPAT_FEATURE(sb, ~0U)))
1995
		printk(KERN_WARNING
1996
		       "EXT4-fs warning: feature flags set on rev 0 fs, "
1997
		       "running e2fsck is recommended\n");
1998

1999 2000 2001 2002 2003
	/*
	 * Check feature flags regardless of the revision level, since we
	 * previously didn't change the revision level when setting the flags,
	 * so there is a chance incompat flags are set on a rev 0 filesystem.
	 */
2004
	features = EXT4_HAS_INCOMPAT_FEATURE(sb, ~EXT4_FEATURE_INCOMPAT_SUPP);
2005
	if (features) {
2006
		printk(KERN_ERR "EXT4-fs: %s: couldn't mount because of "
2007 2008 2009 2010
		       "unsupported optional features (%x).\n",
		       sb->s_id, le32_to_cpu(features));
		goto failed_mount;
	}
2011
	features = EXT4_HAS_RO_COMPAT_FEATURE(sb, ~EXT4_FEATURE_RO_COMPAT_SUPP);
2012
	if (!(sb->s_flags & MS_RDONLY) && features) {
2013
		printk(KERN_ERR "EXT4-fs: %s: couldn't mount RDWR because of "
2014 2015 2016 2017
		       "unsupported optional features (%x).\n",
		       sb->s_id, le32_to_cpu(features));
		goto failed_mount;
	}
2018 2019 2020
	has_huge_files = EXT4_HAS_RO_COMPAT_FEATURE(sb,
				    EXT4_FEATURE_RO_COMPAT_HUGE_FILE);
	if (has_huge_files) {
2021 2022
		/*
		 * Large file size enabled file system can only be
J
Jens Axboe 已提交
2023
		 * mount if kernel is build with CONFIG_LBD
2024 2025 2026 2027 2028
		 */
		if (sizeof(root->i_blocks) < sizeof(u64) &&
				!(sb->s_flags & MS_RDONLY)) {
			printk(KERN_ERR "EXT4-fs: %s: Filesystem with huge "
					"files cannot be mounted read-write "
J
Jens Axboe 已提交
2029
					"without CONFIG_LBD.\n", sb->s_id);
2030 2031 2032
			goto failed_mount;
		}
	}
2033 2034
	blocksize = BLOCK_SIZE << le32_to_cpu(es->s_log_block_size);

2035 2036
	if (blocksize < EXT4_MIN_BLOCK_SIZE ||
	    blocksize > EXT4_MAX_BLOCK_SIZE) {
2037
		printk(KERN_ERR
2038
		       "EXT4-fs: Unsupported filesystem blocksize %d on %s.\n",
2039 2040 2041 2042 2043
		       blocksize, sb->s_id);
		goto failed_mount;
	}

	if (sb->s_blocksize != blocksize) {
2044 2045 2046 2047 2048

		/* Validate the filesystem blocksize */
		if (!sb_set_blocksize(sb, blocksize)) {
			printk(KERN_ERR "EXT4-fs: bad block size %d.\n",
					blocksize);
2049 2050 2051
			goto failed_mount;
		}

2052
		brelse(bh);
2053 2054 2055
		logical_sb_block = sb_block * EXT4_MIN_BLOCK_SIZE;
		offset = do_div(logical_sb_block, blocksize);
		bh = sb_bread(sb, logical_sb_block);
2056 2057
		if (!bh) {
			printk(KERN_ERR
2058
			       "EXT4-fs: Can't read superblock on 2nd try.\n");
2059 2060
			goto failed_mount;
		}
2061
		es = (struct ext4_super_block *)(((char *)bh->b_data) + offset);
2062
		sbi->s_es = es;
2063
		if (es->s_magic != cpu_to_le16(EXT4_SUPER_MAGIC)) {
2064 2065
			printk(KERN_ERR
			       "EXT4-fs: Magic mismatch, very weird !\n");
2066 2067 2068 2069
			goto failed_mount;
		}
	}

2070 2071 2072
	sbi->s_bitmap_maxbytes = ext4_max_bitmap_size(sb->s_blocksize_bits,
						      has_huge_files);
	sb->s_maxbytes = ext4_max_size(sb->s_blocksize_bits, has_huge_files);
2073

2074 2075 2076
	if (le32_to_cpu(es->s_rev_level) == EXT4_GOOD_OLD_REV) {
		sbi->s_inode_size = EXT4_GOOD_OLD_INODE_SIZE;
		sbi->s_first_ino = EXT4_GOOD_OLD_FIRST_INO;
2077 2078 2079
	} else {
		sbi->s_inode_size = le16_to_cpu(es->s_inode_size);
		sbi->s_first_ino = le32_to_cpu(es->s_first_ino);
2080
		if ((sbi->s_inode_size < EXT4_GOOD_OLD_INODE_SIZE) ||
V
Vignesh Babu 已提交
2081
		    (!is_power_of_2(sbi->s_inode_size)) ||
2082
		    (sbi->s_inode_size > blocksize)) {
2083 2084 2085
			printk(KERN_ERR
			       "EXT4-fs: unsupported inode size: %d\n",
			       sbi->s_inode_size);
2086 2087
			goto failed_mount;
		}
K
Kalpak Shah 已提交
2088 2089
		if (sbi->s_inode_size > EXT4_GOOD_OLD_INODE_SIZE)
			sb->s_time_gran = 1 << (EXT4_EPOCH_BITS - 2);
2090
	}
2091 2092
	sbi->s_desc_size = le16_to_cpu(es->s_desc_size);
	if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_64BIT)) {
2093
		if (sbi->s_desc_size < EXT4_MIN_DESC_SIZE_64BIT ||
2094
		    sbi->s_desc_size > EXT4_MAX_DESC_SIZE ||
V
vignesh babu 已提交
2095
		    !is_power_of_2(sbi->s_desc_size)) {
2096
			printk(KERN_ERR
2097
			       "EXT4-fs: unsupported descriptor size %lu\n",
2098 2099 2100 2101 2102
			       sbi->s_desc_size);
			goto failed_mount;
		}
	} else
		sbi->s_desc_size = EXT4_MIN_DESC_SIZE;
2103 2104
	sbi->s_blocks_per_group = le32_to_cpu(es->s_blocks_per_group);
	sbi->s_inodes_per_group = le32_to_cpu(es->s_inodes_per_group);
2105
	if (EXT4_INODE_SIZE(sb) == 0 || EXT4_INODES_PER_GROUP(sb) == 0)
2106 2107
		goto cantfind_ext4;
	sbi->s_inodes_per_block = blocksize / EXT4_INODE_SIZE(sb);
2108
	if (sbi->s_inodes_per_block == 0)
2109
		goto cantfind_ext4;
2110 2111
	sbi->s_itb_per_group = sbi->s_inodes_per_group /
					sbi->s_inodes_per_block;
2112
	sbi->s_desc_per_block = blocksize / EXT4_DESC_SIZE(sb);
2113 2114
	sbi->s_sbh = bh;
	sbi->s_mount_state = le16_to_cpu(es->s_state);
2115 2116
	sbi->s_addr_per_block_bits = ilog2(EXT4_ADDR_PER_BLOCK(sb));
	sbi->s_desc_per_block_bits = ilog2(EXT4_DESC_PER_BLOCK(sb));
2117
	for (i = 0; i < 4; i++)
2118 2119 2120 2121
		sbi->s_hash_seed[i] = le32_to_cpu(es->s_hash_seed[i]);
	sbi->s_def_hash_version = es->s_def_hash_version;

	if (sbi->s_blocks_per_group > blocksize * 8) {
2122 2123 2124
		printk(KERN_ERR
		       "EXT4-fs: #blocks per group too big: %lu\n",
		       sbi->s_blocks_per_group);
2125 2126 2127
		goto failed_mount;
	}
	if (sbi->s_inodes_per_group > blocksize * 8) {
2128 2129 2130
		printk(KERN_ERR
		       "EXT4-fs: #inodes per group too big: %lu\n",
		       sbi->s_inodes_per_group);
2131 2132 2133
		goto failed_mount;
	}

L
Laurent Vivier 已提交
2134
	if (ext4_blocks_count(es) >
2135
		    (sector_t)(~0ULL) >> (sb->s_blocksize_bits - 9)) {
2136
		printk(KERN_ERR "EXT4-fs: filesystem on %s:"
2137 2138
			" too large to mount safely\n", sb->s_id);
		if (sizeof(sector_t) < 8)
2139
			printk(KERN_WARNING "EXT4-fs: CONFIG_LBD not "
2140 2141 2142 2143
					"enabled\n");
		goto failed_mount;
	}

2144 2145
	if (EXT4_BLOCKS_PER_GROUP(sb) == 0)
		goto cantfind_ext4;
2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156

	/* ensure blocks_count calculation below doesn't sign-extend */
	if (ext4_blocks_count(es) + EXT4_BLOCKS_PER_GROUP(sb) <
	    le32_to_cpu(es->s_first_data_block) + 1) {
		printk(KERN_WARNING "EXT4-fs: bad geometry: block count %llu, "
		       "first data block %u, blocks per group %lu\n",
			ext4_blocks_count(es),
			le32_to_cpu(es->s_first_data_block),
			EXT4_BLOCKS_PER_GROUP(sb));
		goto failed_mount;
	}
L
Laurent Vivier 已提交
2157 2158 2159 2160 2161
	blocks_count = (ext4_blocks_count(es) -
			le32_to_cpu(es->s_first_data_block) +
			EXT4_BLOCKS_PER_GROUP(sb) - 1);
	do_div(blocks_count, EXT4_BLOCKS_PER_GROUP(sb));
	sbi->s_groups_count = blocks_count;
2162 2163
	db_count = (sbi->s_groups_count + EXT4_DESC_PER_BLOCK(sb) - 1) /
		   EXT4_DESC_PER_BLOCK(sb);
2164
	sbi->s_group_desc = kmalloc(db_count * sizeof(struct buffer_head *),
2165 2166
				    GFP_KERNEL);
	if (sbi->s_group_desc == NULL) {
2167
		printk(KERN_ERR "EXT4-fs: not enough memory\n");
2168 2169 2170
		goto failed_mount;
	}

2171
#ifdef CONFIG_PROC_FS
2172 2173 2174
	if (ext4_proc_root)
		sbi->s_proc = proc_mkdir(sb->s_id, ext4_proc_root);

2175 2176 2177 2178
	if (sbi->s_proc)
		proc_create_data("inode_readahead_blks", 0644, sbi->s_proc,
				 &ext4_ui_proc_fops,
				 &sbi->s_inode_readahead_blks);
2179
#endif
2180

2181 2182 2183
	bgl_lock_init(&sbi->s_blockgroup_lock);

	for (i = 0; i < db_count; i++) {
2184
		block = descriptor_loc(sb, logical_sb_block, i);
2185 2186
		sbi->s_group_desc[i] = sb_bread(sb, block);
		if (!sbi->s_group_desc[i]) {
2187 2188
			printk(KERN_ERR "EXT4-fs: "
			       "can't read group descriptor %d\n", i);
2189 2190 2191 2192
			db_count = i;
			goto failed_mount2;
		}
	}
2193
	if (!ext4_check_descriptors(sb)) {
2194
		printk(KERN_ERR "EXT4-fs: group descriptors corrupted!\n");
2195 2196
		goto failed_mount2;
	}
2197 2198 2199 2200 2201 2202 2203 2204
	if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_FLEX_BG))
		if (!ext4_fill_flex_info(sb)) {
			printk(KERN_ERR
			       "EXT4-fs: unable to initialize "
			       "flex_bg meta info!\n");
			goto failed_mount2;
		}

2205 2206 2207 2208
	sbi->s_gdb_count = db_count;
	get_random_bytes(&sbi->s_next_generation, sizeof(u32));
	spin_lock_init(&sbi->s_next_gen_lock);

2209 2210 2211 2212 2213 2214 2215 2216 2217 2218
	err = percpu_counter_init(&sbi->s_freeblocks_counter,
			ext4_count_free_blocks(sb));
	if (!err) {
		err = percpu_counter_init(&sbi->s_freeinodes_counter,
				ext4_count_free_inodes(sb));
	}
	if (!err) {
		err = percpu_counter_init(&sbi->s_dirs_counter,
				ext4_count_dirs(sb));
	}
2219 2220 2221
	if (!err) {
		err = percpu_counter_init(&sbi->s_dirtyblocks_counter, 0);
	}
2222 2223 2224 2225
	if (err) {
		printk(KERN_ERR "EXT4-fs: insufficient memory\n");
		goto failed_mount3;
	}
2226

2227 2228
	sbi->s_stripe = ext4_get_stripe_size(sbi);

2229 2230 2231
	/*
	 * set up enough so that it can read an inode
	 */
2232 2233 2234
	sb->s_op = &ext4_sops;
	sb->s_export_op = &ext4_export_ops;
	sb->s_xattr = ext4_xattr_handlers;
2235
#ifdef CONFIG_QUOTA
2236 2237
	sb->s_qcop = &ext4_qctl_operations;
	sb->dq_op = &ext4_quota_operations;
2238 2239 2240 2241 2242 2243
#endif
	INIT_LIST_HEAD(&sbi->s_orphan); /* unlinked but open files */

	sb->s_root = NULL;

	needs_recovery = (es->s_last_orphan != 0 ||
2244 2245
			  EXT4_HAS_INCOMPAT_FEATURE(sb,
				    EXT4_FEATURE_INCOMPAT_RECOVER));
2246 2247 2248 2249 2250 2251

	/*
	 * The first inode we look at is the journal inode.  Don't try
	 * root first: it may be modified in the journal!
	 */
	if (!test_opt(sb, NOLOAD) &&
2252 2253
	    EXT4_HAS_COMPAT_FEATURE(sb, EXT4_FEATURE_COMPAT_HAS_JOURNAL)) {
		if (ext4_load_journal(sb, es, journal_devnum))
2254
			goto failed_mount3;
2255 2256 2257 2258
		if (!(sb->s_flags & MS_RDONLY) &&
		    EXT4_SB(sb)->s_journal->j_failed_commit) {
			printk(KERN_CRIT "EXT4-fs error (device %s): "
			       "ext4_fill_super: Journal transaction "
2259
			       "%u is corrupt\n", sb->s_id,
2260
			       EXT4_SB(sb)->s_journal->j_failed_commit);
2261 2262 2263
			if (test_opt(sb, ERRORS_RO)) {
				printk(KERN_CRIT
				       "Mounting filesystem read-only\n");
2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277
				sb->s_flags |= MS_RDONLY;
				EXT4_SB(sb)->s_mount_state |= EXT4_ERROR_FS;
				es->s_state |= cpu_to_le16(EXT4_ERROR_FS);
			}
			if (test_opt(sb, ERRORS_PANIC)) {
				EXT4_SB(sb)->s_mount_state |= EXT4_ERROR_FS;
				es->s_state |= cpu_to_le16(EXT4_ERROR_FS);
				ext4_commit_super(sb, es, 1);
				printk(KERN_CRIT
				       "EXT4-fs (device %s): mount failed\n",
				      sb->s_id);
				goto failed_mount4;
			}
		}
2278
	} else if (journal_inum) {
2279
		if (ext4_create_journal(sb, es, journal_inum))
2280 2281 2282
			goto failed_mount3;
	} else {
		if (!silent)
2283 2284 2285
			printk(KERN_ERR
			       "ext4: No journal on filesystem on %s\n",
			       sb->s_id);
2286 2287 2288
		goto failed_mount3;
	}

2289 2290 2291 2292 2293 2294 2295
	if (ext4_blocks_count(es) > 0xffffffffULL &&
	    !jbd2_journal_set_features(EXT4_SB(sb)->s_journal, 0, 0,
				       JBD2_FEATURE_INCOMPAT_64BIT)) {
		printk(KERN_ERR "ext4: Failed to set 64-bit journal feature\n");
		goto failed_mount4;
	}

2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310
	if (test_opt(sb, JOURNAL_ASYNC_COMMIT)) {
		jbd2_journal_set_features(sbi->s_journal,
				JBD2_FEATURE_COMPAT_CHECKSUM, 0,
				JBD2_FEATURE_INCOMPAT_ASYNC_COMMIT);
	} else if (test_opt(sb, JOURNAL_CHECKSUM)) {
		jbd2_journal_set_features(sbi->s_journal,
				JBD2_FEATURE_COMPAT_CHECKSUM, 0, 0);
		jbd2_journal_clear_features(sbi->s_journal, 0, 0,
				JBD2_FEATURE_INCOMPAT_ASYNC_COMMIT);
	} else {
		jbd2_journal_clear_features(sbi->s_journal,
				JBD2_FEATURE_COMPAT_CHECKSUM, 0,
				JBD2_FEATURE_INCOMPAT_ASYNC_COMMIT);
	}

2311 2312 2313 2314 2315
	/* We have now updated the journal if required, so we can
	 * validate the data journaling mode. */
	switch (test_opt(sb, DATA_FLAGS)) {
	case 0:
		/* No mode set, assume a default based on the journal
A
Andrew Morton 已提交
2316 2317 2318
		 * capabilities: ORDERED_DATA if the journal can
		 * cope, else JOURNAL_DATA
		 */
2319 2320
		if (jbd2_journal_check_available_features
		    (sbi->s_journal, 0, 0, JBD2_FEATURE_INCOMPAT_REVOKE))
2321 2322 2323 2324 2325
			set_opt(sbi->s_mount_opt, ORDERED_DATA);
		else
			set_opt(sbi->s_mount_opt, JOURNAL_DATA);
		break;

2326 2327
	case EXT4_MOUNT_ORDERED_DATA:
	case EXT4_MOUNT_WRITEBACK_DATA:
2328 2329
		if (!jbd2_journal_check_available_features
		    (sbi->s_journal, 0, 0, JBD2_FEATURE_INCOMPAT_REVOKE)) {
2330
			printk(KERN_ERR "EXT4-fs: Journal does not support "
2331 2332 2333 2334 2335 2336 2337 2338
			       "requested data journaling mode\n");
			goto failed_mount4;
		}
	default:
		break;
	}

	if (test_opt(sb, NOBH)) {
2339 2340
		if (!(test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_WRITEBACK_DATA)) {
			printk(KERN_WARNING "EXT4-fs: Ignoring nobh option - "
2341 2342 2343 2344 2345
				"its supported only with writeback mode\n");
			clear_opt(sbi->s_mount_opt, NOBH);
		}
	}
	/*
2346
	 * The jbd2_journal_load will have done any necessary log recovery,
2347 2348 2349
	 * so we can safely mount the rest of the filesystem now.
	 */

2350 2351
	root = ext4_iget(sb, EXT4_ROOT_INO);
	if (IS_ERR(root)) {
2352
		printk(KERN_ERR "EXT4-fs: get root inode failed\n");
2353
		ret = PTR_ERR(root);
2354 2355 2356
		goto failed_mount4;
	}
	if (!S_ISDIR(root->i_mode) || !root->i_blocks || !root->i_size) {
2357
		iput(root);
2358
		printk(KERN_ERR "EXT4-fs: corrupt root inode, run e2fsck\n");
2359 2360
		goto failed_mount4;
	}
2361 2362 2363 2364 2365 2366 2367
	sb->s_root = d_alloc_root(root);
	if (!sb->s_root) {
		printk(KERN_ERR "EXT4-fs: get root dentry failed\n");
		iput(root);
		ret = -ENOMEM;
		goto failed_mount4;
	}
2368

2369
	ext4_setup_super(sb, es, sb->s_flags & MS_RDONLY);
K
Kalpak Shah 已提交
2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395

	/* determine the minimum size of new large inodes, if present */
	if (sbi->s_inode_size > EXT4_GOOD_OLD_INODE_SIZE) {
		sbi->s_want_extra_isize = sizeof(struct ext4_inode) -
						     EXT4_GOOD_OLD_INODE_SIZE;
		if (EXT4_HAS_RO_COMPAT_FEATURE(sb,
				       EXT4_FEATURE_RO_COMPAT_EXTRA_ISIZE)) {
			if (sbi->s_want_extra_isize <
			    le16_to_cpu(es->s_want_extra_isize))
				sbi->s_want_extra_isize =
					le16_to_cpu(es->s_want_extra_isize);
			if (sbi->s_want_extra_isize <
			    le16_to_cpu(es->s_min_extra_isize))
				sbi->s_want_extra_isize =
					le16_to_cpu(es->s_min_extra_isize);
		}
	}
	/* Check if enough inode space is available */
	if (EXT4_GOOD_OLD_INODE_SIZE + sbi->s_want_extra_isize >
							sbi->s_inode_size) {
		sbi->s_want_extra_isize = sizeof(struct ext4_inode) -
						       EXT4_GOOD_OLD_INODE_SIZE;
		printk(KERN_INFO "EXT4-fs: required extra inode space not"
			"available.\n");
	}

2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410
	if (test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_JOURNAL_DATA) {
		printk(KERN_WARNING "EXT4-fs: Ignoring delalloc option - "
				"requested data journaling mode\n");
		clear_opt(sbi->s_mount_opt, DELALLOC);
	} else if (test_opt(sb, DELALLOC))
		printk(KERN_INFO "EXT4-fs: delayed allocation enabled\n");

	ext4_ext_init(sb);
	err = ext4_mb_init(sb, needs_recovery);
	if (err) {
		printk(KERN_ERR "EXT4-fs: failed to initalize mballoc (%d)\n",
		       err);
		goto failed_mount4;
	}

2411 2412 2413 2414 2415 2416 2417 2418
	/*
	 * akpm: core read_super() calls in here with the superblock locked.
	 * That deadlocks, because orphan cleanup needs to lock the superblock
	 * in numerous places.  Here we just pop the lock - it's relatively
	 * harmless, because we are now ready to accept write_super() requests,
	 * and aviro says that's the only reason for hanging onto the
	 * superblock lock.
	 */
2419 2420 2421
	EXT4_SB(sb)->s_mount_state |= EXT4_ORPHAN_FS;
	ext4_orphan_cleanup(sb, es);
	EXT4_SB(sb)->s_mount_state &= ~EXT4_ORPHAN_FS;
2422
	if (needs_recovery)
2423
		printk(KERN_INFO "EXT4-fs: recovery complete.\n");
2424
	ext4_mark_recovery_complete(sb, es);
2425 2426 2427 2428
	printk(KERN_INFO "EXT4-fs: mounted filesystem with %s data mode.\n",
	       test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_JOURNAL_DATA ? "journal":
	       test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_ORDERED_DATA ? "ordered":
	       "writeback");
2429 2430 2431 2432

	lock_kernel();
	return 0;

2433
cantfind_ext4:
2434
	if (!silent)
2435
		printk(KERN_ERR "VFS: Can't find ext4 filesystem on dev %s.\n",
2436 2437 2438 2439
		       sb->s_id);
	goto failed_mount;

failed_mount4:
2440
	jbd2_journal_destroy(sbi->s_journal);
2441
	sbi->s_journal = NULL;
2442 2443 2444 2445
failed_mount3:
	percpu_counter_destroy(&sbi->s_freeblocks_counter);
	percpu_counter_destroy(&sbi->s_freeinodes_counter);
	percpu_counter_destroy(&sbi->s_dirs_counter);
2446
	percpu_counter_destroy(&sbi->s_dirtyblocks_counter);
2447 2448 2449 2450 2451
failed_mount2:
	for (i = 0; i < db_count; i++)
		brelse(sbi->s_group_desc[i]);
	kfree(sbi->s_group_desc);
failed_mount:
2452 2453
	if (sbi->s_proc) {
		remove_proc_entry("inode_readahead_blks", sbi->s_proc);
2454
		remove_proc_entry(sb->s_id, ext4_proc_root);
2455
	}
2456 2457 2458 2459
#ifdef CONFIG_QUOTA
	for (i = 0; i < MAXQUOTAS; i++)
		kfree(sbi->s_qf_names[i]);
#endif
2460
	ext4_blkdev_remove(sbi);
2461 2462 2463 2464 2465
	brelse(bh);
out_fail:
	sb->s_fs_info = NULL;
	kfree(sbi);
	lock_kernel();
2466
	return ret;
2467 2468 2469 2470 2471 2472 2473
}

/*
 * Setup any per-fs journal parameters now.  We'll do this both on
 * initial mount, once the journal has been initialised but before we've
 * done any recovery; and again on any subsequent remount.
 */
2474
static void ext4_init_journal_params(struct super_block *sb, journal_t *journal)
2475
{
2476
	struct ext4_sb_info *sbi = EXT4_SB(sb);
2477 2478 2479

	if (sbi->s_commit_interval)
		journal->j_commit_interval = sbi->s_commit_interval;
2480
	/* We could also set up an ext4-specific default for the commit
2481 2482 2483 2484 2485
	 * interval here, but for now we'll just fall back to the jbd
	 * default. */

	spin_lock(&journal->j_state_lock);
	if (test_opt(sb, BARRIER))
2486
		journal->j_flags |= JBD2_BARRIER;
2487
	else
2488
		journal->j_flags &= ~JBD2_BARRIER;
2489 2490 2491 2492
	if (test_opt(sb, DATA_ERR_ABORT))
		journal->j_flags |= JBD2_ABORT_ON_SYNCDATA_ERR;
	else
		journal->j_flags &= ~JBD2_ABORT_ON_SYNCDATA_ERR;
2493 2494 2495
	spin_unlock(&journal->j_state_lock);
}

2496
static journal_t *ext4_get_journal(struct super_block *sb,
2497 2498 2499 2500 2501 2502 2503 2504 2505
				   unsigned int journal_inum)
{
	struct inode *journal_inode;
	journal_t *journal;

	/* First, test for the existence of a valid inode on disk.  Bad
	 * things happen if we iget() an unused inode, as the subsequent
	 * iput() will try to delete it. */

2506 2507
	journal_inode = ext4_iget(sb, journal_inum);
	if (IS_ERR(journal_inode)) {
2508
		printk(KERN_ERR "EXT4-fs: no journal found.\n");
2509 2510 2511 2512 2513
		return NULL;
	}
	if (!journal_inode->i_nlink) {
		make_bad_inode(journal_inode);
		iput(journal_inode);
2514
		printk(KERN_ERR "EXT4-fs: journal inode is deleted.\n");
2515 2516 2517
		return NULL;
	}

2518
	jbd_debug(2, "Journal inode found at %p: %lld bytes\n",
2519
		  journal_inode, journal_inode->i_size);
2520
	if (!S_ISREG(journal_inode->i_mode)) {
2521
		printk(KERN_ERR "EXT4-fs: invalid journal inode.\n");
2522 2523 2524 2525
		iput(journal_inode);
		return NULL;
	}

2526
	journal = jbd2_journal_init_inode(journal_inode);
2527
	if (!journal) {
2528
		printk(KERN_ERR "EXT4-fs: Could not load journal inode\n");
2529 2530 2531 2532
		iput(journal_inode);
		return NULL;
	}
	journal->j_private = sb;
2533
	ext4_init_journal_params(sb, journal);
2534 2535 2536
	return journal;
}

2537
static journal_t *ext4_get_dev_journal(struct super_block *sb,
2538 2539
				       dev_t j_dev)
{
2540
	struct buffer_head *bh;
2541
	journal_t *journal;
2542 2543
	ext4_fsblk_t start;
	ext4_fsblk_t len;
2544
	int hblock, blocksize;
2545
	ext4_fsblk_t sb_block;
2546
	unsigned long offset;
2547
	struct ext4_super_block *es;
2548 2549
	struct block_device *bdev;

2550
	bdev = ext4_blkdev_get(j_dev);
2551 2552 2553 2554 2555
	if (bdev == NULL)
		return NULL;

	if (bd_claim(bdev, sb)) {
		printk(KERN_ERR
D
Dave Kleikamp 已提交
2556
			"EXT4: failed to claim external journal device.\n");
A
Al Viro 已提交
2557
		blkdev_put(bdev, FMODE_READ|FMODE_WRITE);
2558 2559 2560 2561 2562 2563 2564
		return NULL;
	}

	blocksize = sb->s_blocksize;
	hblock = bdev_hardsect_size(bdev);
	if (blocksize < hblock) {
		printk(KERN_ERR
2565
			"EXT4-fs: blocksize too small for journal device.\n");
2566 2567 2568
		goto out_bdev;
	}

2569 2570
	sb_block = EXT4_MIN_BLOCK_SIZE / blocksize;
	offset = EXT4_MIN_BLOCK_SIZE % blocksize;
2571 2572
	set_blocksize(bdev, blocksize);
	if (!(bh = __bread(bdev, sb_block, blocksize))) {
2573
		printk(KERN_ERR "EXT4-fs: couldn't read superblock of "
2574 2575 2576 2577
		       "external journal\n");
		goto out_bdev;
	}

2578 2579
	es = (struct ext4_super_block *) (((char *)bh->b_data) + offset);
	if ((le16_to_cpu(es->s_magic) != EXT4_SUPER_MAGIC) ||
2580
	    !(le32_to_cpu(es->s_feature_incompat) &
2581 2582
	      EXT4_FEATURE_INCOMPAT_JOURNAL_DEV)) {
		printk(KERN_ERR "EXT4-fs: external journal has "
2583 2584 2585 2586 2587
					"bad superblock\n");
		brelse(bh);
		goto out_bdev;
	}

2588 2589
	if (memcmp(EXT4_SB(sb)->s_es->s_journal_uuid, es->s_uuid, 16)) {
		printk(KERN_ERR "EXT4-fs: journal UUID does not match\n");
2590 2591 2592 2593
		brelse(bh);
		goto out_bdev;
	}

L
Laurent Vivier 已提交
2594
	len = ext4_blocks_count(es);
2595 2596 2597
	start = sb_block + 1;
	brelse(bh);	/* we're done with the superblock */

2598
	journal = jbd2_journal_init_dev(bdev, sb->s_bdev,
2599 2600
					start, len, blocksize);
	if (!journal) {
2601
		printk(KERN_ERR "EXT4-fs: failed to create device journal\n");
2602 2603 2604 2605 2606 2607
		goto out_bdev;
	}
	journal->j_private = sb;
	ll_rw_block(READ, 1, &journal->j_sb_buffer);
	wait_on_buffer(journal->j_sb_buffer);
	if (!buffer_uptodate(journal->j_sb_buffer)) {
2608
		printk(KERN_ERR "EXT4-fs: I/O error on journal device\n");
2609 2610 2611
		goto out_journal;
	}
	if (be32_to_cpu(journal->j_superblock->s_nr_users) != 1) {
2612
		printk(KERN_ERR "EXT4-fs: External journal has more than one "
2613 2614 2615 2616
					"user (unsupported) - %d\n",
			be32_to_cpu(journal->j_superblock->s_nr_users));
		goto out_journal;
	}
2617 2618
	EXT4_SB(sb)->journal_bdev = bdev;
	ext4_init_journal_params(sb, journal);
2619 2620
	return journal;
out_journal:
2621
	jbd2_journal_destroy(journal);
2622
out_bdev:
2623
	ext4_blkdev_put(bdev);
2624 2625 2626
	return NULL;
}

2627 2628
static int ext4_load_journal(struct super_block *sb,
			     struct ext4_super_block *es,
2629 2630 2631 2632 2633 2634 2635 2636 2637 2638
			     unsigned long journal_devnum)
{
	journal_t *journal;
	unsigned int journal_inum = le32_to_cpu(es->s_journal_inum);
	dev_t journal_dev;
	int err = 0;
	int really_read_only;

	if (journal_devnum &&
	    journal_devnum != le32_to_cpu(es->s_journal_dev)) {
2639
		printk(KERN_INFO "EXT4-fs: external journal device major/minor "
2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652
			"numbers have changed\n");
		journal_dev = new_decode_dev(journal_devnum);
	} else
		journal_dev = new_decode_dev(le32_to_cpu(es->s_journal_dev));

	really_read_only = bdev_read_only(sb->s_bdev);

	/*
	 * Are we loading a blank journal or performing recovery after a
	 * crash?  For recovery, we need to check in advance whether we
	 * can get read-write access to the device.
	 */

2653
	if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER)) {
2654
		if (sb->s_flags & MS_RDONLY) {
2655
			printk(KERN_INFO "EXT4-fs: INFO: recovery "
2656 2657
					"required on readonly filesystem.\n");
			if (really_read_only) {
2658
				printk(KERN_ERR "EXT4-fs: write access "
2659 2660 2661
					"unavailable, cannot proceed.\n");
				return -EROFS;
			}
2662 2663
			printk(KERN_INFO "EXT4-fs: write access will "
			       "be enabled during recovery.\n");
2664 2665 2666 2667
		}
	}

	if (journal_inum && journal_dev) {
2668
		printk(KERN_ERR "EXT4-fs: filesystem has both journal "
2669 2670 2671 2672 2673
		       "and inode journals!\n");
		return -EINVAL;
	}

	if (journal_inum) {
2674
		if (!(journal = ext4_get_journal(sb, journal_inum)))
2675 2676
			return -EINVAL;
	} else {
2677
		if (!(journal = ext4_get_dev_journal(sb, journal_dev)))
2678 2679 2680
			return -EINVAL;
	}

2681 2682 2683 2684 2685
	if (journal->j_flags & JBD2_BARRIER)
		printk(KERN_INFO "EXT4-fs: barriers enabled\n");
	else
		printk(KERN_INFO "EXT4-fs: barriers disabled\n");

2686
	if (!really_read_only && test_opt(sb, UPDATE_JOURNAL)) {
2687
		err = jbd2_journal_update_format(journal);
2688
		if (err)  {
2689
			printk(KERN_ERR "EXT4-fs: error updating journal.\n");
2690
			jbd2_journal_destroy(journal);
2691 2692 2693 2694
			return err;
		}
	}

2695
	if (!EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER))
2696
		err = jbd2_journal_wipe(journal, !really_read_only);
2697
	if (!err)
2698
		err = jbd2_journal_load(journal);
2699 2700

	if (err) {
2701
		printk(KERN_ERR "EXT4-fs: error loading journal.\n");
2702
		jbd2_journal_destroy(journal);
2703 2704 2705
		return err;
	}

2706 2707
	EXT4_SB(sb)->s_journal = journal;
	ext4_clear_journal_err(sb, es);
2708 2709 2710 2711 2712 2713 2714

	if (journal_devnum &&
	    journal_devnum != le32_to_cpu(es->s_journal_dev)) {
		es->s_journal_dev = cpu_to_le32(journal_devnum);
		sb->s_dirt = 1;

		/* Make sure we flush the recovery flag to disk. */
2715
		ext4_commit_super(sb, es, 1);
2716 2717 2718 2719 2720
	}

	return 0;
}

2721 2722
static int ext4_create_journal(struct super_block *sb,
			       struct ext4_super_block *es,
2723 2724 2725
			       unsigned int journal_inum)
{
	journal_t *journal;
2726
	int err;
2727 2728

	if (sb->s_flags & MS_RDONLY) {
2729
		printk(KERN_ERR "EXT4-fs: readonly filesystem when trying to "
2730 2731 2732 2733
				"create journal.\n");
		return -EROFS;
	}

2734 2735
	journal = ext4_get_journal(sb, journal_inum);
	if (!journal)
2736 2737
		return -EINVAL;

2738
	printk(KERN_INFO "EXT4-fs: creating new journal on inode %u\n",
2739 2740
	       journal_inum);

2741 2742
	err = jbd2_journal_create(journal);
	if (err) {
2743
		printk(KERN_ERR "EXT4-fs: error creating journal.\n");
2744
		jbd2_journal_destroy(journal);
2745 2746 2747
		return -EIO;
	}

2748
	EXT4_SB(sb)->s_journal = journal;
2749

2750 2751 2752
	ext4_update_dynamic_rev(sb);
	EXT4_SET_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER);
	EXT4_SET_COMPAT_FEATURE(sb, EXT4_FEATURE_COMPAT_HAS_JOURNAL);
2753 2754 2755 2756 2757

	es->s_journal_inum = cpu_to_le32(journal_inum);
	sb->s_dirt = 1;

	/* Make sure we flush the recovery flag to disk. */
2758
	ext4_commit_super(sb, es, 1);
2759 2760 2761 2762

	return 0;
}

2763 2764
static void ext4_commit_super(struct super_block *sb,
			      struct ext4_super_block *es, int sync)
2765
{
2766
	struct buffer_head *sbh = EXT4_SB(sb)->s_sbh;
2767 2768 2769

	if (!sbh)
		return;
2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783
	if (buffer_write_io_error(sbh)) {
		/*
		 * Oh, dear.  A previous attempt to write the
		 * superblock failed.  This could happen because the
		 * USB device was yanked out.  Or it could happen to
		 * be a transient write error and maybe the block will
		 * be remapped.  Nothing we can do but to retry the
		 * write and hope for the best.
		 */
		printk(KERN_ERR "ext4: previous I/O error to "
		       "superblock detected for %s.\n", sb->s_id);
		clear_buffer_write_io_error(sbh);
		set_buffer_uptodate(sbh);
	}
2784
	es->s_wtime = cpu_to_le32(get_seconds());
L
Laurent Vivier 已提交
2785
	ext4_free_blocks_count_set(es, ext4_count_free_blocks(sb));
2786
	es->s_free_inodes_count = cpu_to_le32(ext4_count_free_inodes(sb));
2787 2788
	BUFFER_TRACE(sbh, "marking dirty");
	mark_buffer_dirty(sbh);
2789
	if (sync) {
2790
		sync_dirty_buffer(sbh);
2791 2792 2793 2794 2795 2796 2797
		if (buffer_write_io_error(sbh)) {
			printk(KERN_ERR "ext4: I/O error while writing "
			       "superblock for %s.\n", sb->s_id);
			clear_buffer_write_io_error(sbh);
			set_buffer_uptodate(sbh);
		}
	}
2798 2799 2800 2801 2802 2803 2804 2805
}


/*
 * Have we just finished recovery?  If so, and if we are mounting (or
 * remounting) the filesystem readonly, then we will end up with a
 * consistent fs on disk.  Record that fact.
 */
2806 2807
static void ext4_mark_recovery_complete(struct super_block *sb,
					struct ext4_super_block *es)
2808
{
2809
	journal_t *journal = EXT4_SB(sb)->s_journal;
2810

2811
	jbd2_journal_lock_updates(journal);
2812 2813 2814
	if (jbd2_journal_flush(journal) < 0)
		goto out;

2815
	lock_super(sb);
2816
	if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER) &&
2817
	    sb->s_flags & MS_RDONLY) {
2818
		EXT4_CLEAR_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER);
2819
		sb->s_dirt = 0;
2820
		ext4_commit_super(sb, es, 1);
2821
	}
2822
	unlock_super(sb);
2823 2824

out:
2825
	jbd2_journal_unlock_updates(journal);
2826 2827 2828 2829 2830 2831 2832
}

/*
 * If we are mounting (or read-write remounting) a filesystem whose journal
 * has recorded an error from a previous lifetime, move that error to the
 * main filesystem now.
 */
2833 2834
static void ext4_clear_journal_err(struct super_block *sb,
				   struct ext4_super_block *es)
2835 2836 2837 2838 2839
{
	journal_t *journal;
	int j_errno;
	const char *errstr;

2840
	journal = EXT4_SB(sb)->s_journal;
2841 2842 2843

	/*
	 * Now check for any error status which may have been recorded in the
2844
	 * journal by a prior ext4_error() or ext4_abort()
2845 2846
	 */

2847
	j_errno = jbd2_journal_errno(journal);
2848 2849 2850
	if (j_errno) {
		char nbuf[16];

2851
		errstr = ext4_decode_error(sb, j_errno, nbuf);
2852
		ext4_warning(sb, __func__, "Filesystem error recorded "
2853
			     "from previous mount: %s", errstr);
2854
		ext4_warning(sb, __func__, "Marking fs in need of "
2855 2856
			     "filesystem check.");

2857 2858
		EXT4_SB(sb)->s_mount_state |= EXT4_ERROR_FS;
		es->s_state |= cpu_to_le16(EXT4_ERROR_FS);
2859
		ext4_commit_super(sb, es, 1);
2860

2861
		jbd2_journal_clear_err(journal);
2862 2863 2864 2865 2866 2867 2868
	}
}

/*
 * Force the running and committing transactions to commit,
 * and wait on the commit.
 */
2869
int ext4_force_commit(struct super_block *sb)
2870 2871 2872 2873 2874 2875 2876
{
	journal_t *journal;
	int ret;

	if (sb->s_flags & MS_RDONLY)
		return 0;

2877
	journal = EXT4_SB(sb)->s_journal;
2878
	sb->s_dirt = 0;
2879
	ret = ext4_journal_force_commit(journal);
2880 2881 2882 2883
	return ret;
}

/*
2884
 * Ext4 always journals updates to the superblock itself, so we don't
2885
 * have to propagate any other updates to the superblock on disk at this
2886 2887
 * point.  (We can probably nuke this function altogether, and remove
 * any mention to sb->s_dirt in all of fs/ext4; eventual cleanup...)
2888
 */
2889
static void ext4_write_super(struct super_block *sb)
2890 2891 2892 2893 2894 2895
{
	if (mutex_trylock(&sb->s_lock) != 0)
		BUG();
	sb->s_dirt = 0;
}

2896
static int ext4_sync_fs(struct super_block *sb, int wait)
2897
{
2898
	int ret = 0;
2899

2900
	trace_mark(ext4_sync_fs, "dev %s wait %d", sb->s_id, wait);
2901
	sb->s_dirt = 0;
2902 2903 2904 2905 2906
	if (wait)
		ret = ext4_force_commit(sb);
	else
		jbd2_journal_start_commit(EXT4_SB(sb)->s_journal, NULL);
	return ret;
2907 2908 2909 2910 2911 2912
}

/*
 * LVM calls this function before a (read-only) snapshot is created.  This
 * gives us a chance to flush the journal completely and mark the fs clean.
 */
2913
static void ext4_write_super_lockfs(struct super_block *sb)
2914 2915 2916 2917
{
	sb->s_dirt = 0;

	if (!(sb->s_flags & MS_RDONLY)) {
2918
		journal_t *journal = EXT4_SB(sb)->s_journal;
2919 2920

		/* Now we set up the journal barrier. */
2921
		jbd2_journal_lock_updates(journal);
2922 2923 2924 2925 2926 2927 2928

		/*
		 * We don't want to clear needs_recovery flag when we failed
		 * to flush the journal.
		 */
		if (jbd2_journal_flush(journal) < 0)
			return;
2929 2930

		/* Journal blocked and flushed, clear needs_recovery flag. */
2931 2932
		EXT4_CLEAR_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER);
		ext4_commit_super(sb, EXT4_SB(sb)->s_es, 1);
2933 2934 2935 2936 2937 2938 2939
	}
}

/*
 * Called by LVM after the snapshot is done.  We need to reset the RECOVER
 * flag here, even though the filesystem is not technically dirty yet.
 */
2940
static void ext4_unlockfs(struct super_block *sb)
2941 2942 2943 2944
{
	if (!(sb->s_flags & MS_RDONLY)) {
		lock_super(sb);
		/* Reser the needs_recovery flag before the fs is unlocked. */
2945 2946
		EXT4_SET_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER);
		ext4_commit_super(sb, EXT4_SB(sb)->s_es, 1);
2947
		unlock_super(sb);
2948
		jbd2_journal_unlock_updates(EXT4_SB(sb)->s_journal);
2949 2950 2951
	}
}

2952
static int ext4_remount(struct super_block *sb, int *flags, char *data)
2953
{
2954
	struct ext4_super_block *es;
2955 2956
	struct ext4_sb_info *sbi = EXT4_SB(sb);
	ext4_fsblk_t n_blocks_count = 0;
2957
	unsigned long old_sb_flags;
2958
	struct ext4_mount_options old_opts;
2959
	ext4_group_t g;
2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 2981 2982 2983 2984
	int err;
#ifdef CONFIG_QUOTA
	int i;
#endif

	/* Store the original options */
	old_sb_flags = sb->s_flags;
	old_opts.s_mount_opt = sbi->s_mount_opt;
	old_opts.s_resuid = sbi->s_resuid;
	old_opts.s_resgid = sbi->s_resgid;
	old_opts.s_commit_interval = sbi->s_commit_interval;
#ifdef CONFIG_QUOTA
	old_opts.s_jquota_fmt = sbi->s_jquota_fmt;
	for (i = 0; i < MAXQUOTAS; i++)
		old_opts.s_qf_names[i] = sbi->s_qf_names[i];
#endif

	/*
	 * Allow the "check" option to be passed as a remount option.
	 */
	if (!parse_options(data, sb, NULL, NULL, &n_blocks_count, 1)) {
		err = -EINVAL;
		goto restore_opts;
	}

2985
	if (sbi->s_mount_opt & EXT4_MOUNT_ABORT)
2986
		ext4_abort(sb, __func__, "Abort forced by user");
2987 2988

	sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
2989
		((sbi->s_mount_opt & EXT4_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0);
2990 2991 2992

	es = sbi->s_es;

2993
	ext4_init_journal_params(sb, sbi->s_journal);
2994 2995

	if ((*flags & MS_RDONLY) != (sb->s_flags & MS_RDONLY) ||
L
Laurent Vivier 已提交
2996
		n_blocks_count > ext4_blocks_count(es)) {
2997
		if (sbi->s_mount_opt & EXT4_MOUNT_ABORT) {
2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013
			err = -EROFS;
			goto restore_opts;
		}

		if (*flags & MS_RDONLY) {
			/*
			 * First of all, the unconditional stuff we have to do
			 * to disable replay of the journal when we next remount
			 */
			sb->s_flags |= MS_RDONLY;

			/*
			 * OK, test if we are remounting a valid rw partition
			 * readonly, and if so set the rdonly flag and then
			 * mark the partition as valid again.
			 */
3014 3015
			if (!(es->s_state & cpu_to_le16(EXT4_VALID_FS)) &&
			    (sbi->s_mount_state & EXT4_VALID_FS))
3016 3017
				es->s_state = cpu_to_le16(sbi->s_mount_state);

3018 3019 3020 3021 3022
			/*
			 * We have to unlock super so that we can wait for
			 * transactions.
			 */
			unlock_super(sb);
3023
			ext4_mark_recovery_complete(sb, es);
3024
			lock_super(sb);
3025 3026
		} else {
			__le32 ret;
3027 3028 3029
			if ((ret = EXT4_HAS_RO_COMPAT_FEATURE(sb,
					~EXT4_FEATURE_RO_COMPAT_SUPP))) {
				printk(KERN_WARNING "EXT4-fs: %s: couldn't "
3030 3031 3032 3033 3034 3035
				       "remount RDWR because of unsupported "
				       "optional features (%x).\n",
				       sb->s_id, le32_to_cpu(ret));
				err = -EROFS;
				goto restore_opts;
			}
3036

3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056
			/*
			 * Make sure the group descriptor checksums
			 * are sane.  If they aren't, refuse to
			 * remount r/w.
			 */
			for (g = 0; g < sbi->s_groups_count; g++) {
				struct ext4_group_desc *gdp =
					ext4_get_group_desc(sb, g, NULL);

				if (!ext4_group_desc_csum_verify(sbi, g, gdp)) {
					printk(KERN_ERR
	       "EXT4-fs: ext4_remount: "
		"Checksum for group %lu failed (%u!=%u)\n",
		g, le16_to_cpu(ext4_group_desc_csum(sbi, g, gdp)),
					       le16_to_cpu(gdp->bg_checksum));
					err = -EINVAL;
					goto restore_opts;
				}
			}

3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071
			/*
			 * If we have an unprocessed orphan list hanging
			 * around from a previously readonly bdev mount,
			 * require a full umount/remount for now.
			 */
			if (es->s_last_orphan) {
				printk(KERN_WARNING "EXT4-fs: %s: couldn't "
				       "remount RDWR because of unprocessed "
				       "orphan inode list.  Please "
				       "umount/remount instead.\n",
				       sb->s_id);
				err = -EINVAL;
				goto restore_opts;
			}

3072 3073 3074 3075 3076 3077
			/*
			 * Mounting a RDONLY partition read-write, so reread
			 * and store the current valid flag.  (It may have
			 * been changed by e2fsck since we originally mounted
			 * the partition.)
			 */
3078
			ext4_clear_journal_err(sb, es);
3079
			sbi->s_mount_state = le16_to_cpu(es->s_state);
3080
			if ((err = ext4_group_extend(sb, es, n_blocks_count)))
3081
				goto restore_opts;
3082
			if (!ext4_setup_super(sb, es, 0))
3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111
				sb->s_flags &= ~MS_RDONLY;
		}
	}
#ifdef CONFIG_QUOTA
	/* Release old quota file names */
	for (i = 0; i < MAXQUOTAS; i++)
		if (old_opts.s_qf_names[i] &&
		    old_opts.s_qf_names[i] != sbi->s_qf_names[i])
			kfree(old_opts.s_qf_names[i]);
#endif
	return 0;
restore_opts:
	sb->s_flags = old_sb_flags;
	sbi->s_mount_opt = old_opts.s_mount_opt;
	sbi->s_resuid = old_opts.s_resuid;
	sbi->s_resgid = old_opts.s_resgid;
	sbi->s_commit_interval = old_opts.s_commit_interval;
#ifdef CONFIG_QUOTA
	sbi->s_jquota_fmt = old_opts.s_jquota_fmt;
	for (i = 0; i < MAXQUOTAS; i++) {
		if (sbi->s_qf_names[i] &&
		    old_opts.s_qf_names[i] != sbi->s_qf_names[i])
			kfree(sbi->s_qf_names[i]);
		sbi->s_qf_names[i] = old_opts.s_qf_names[i];
	}
#endif
	return err;
}

3112
static int ext4_statfs(struct dentry *dentry, struct kstatfs *buf)
3113 3114
{
	struct super_block *sb = dentry->d_sb;
3115 3116
	struct ext4_sb_info *sbi = EXT4_SB(sb);
	struct ext4_super_block *es = sbi->s_es;
P
Pekka Enberg 已提交
3117
	u64 fsid;
3118

B
Badari Pulavarty 已提交
3119 3120
	if (test_opt(sb, MINIX_DF)) {
		sbi->s_overhead_last = 0;
3121
	} else if (sbi->s_blocks_last != ext4_blocks_count(es)) {
3122
		ext4_group_t ngroups = sbi->s_groups_count, i;
B
Badari Pulavarty 已提交
3123
		ext4_fsblk_t overhead = 0;
3124 3125 3126
		smp_rmb();

		/*
B
Badari Pulavarty 已提交
3127 3128 3129
		 * Compute the overhead (FS structures).  This is constant
		 * for a given filesystem unless the number of block groups
		 * changes so we cache the previous value until it does.
3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143
		 */

		/*
		 * All of the blocks before first_data_block are
		 * overhead
		 */
		overhead = le32_to_cpu(es->s_first_data_block);

		/*
		 * Add the overhead attributed to the superblock and
		 * block group descriptors.  If the sparse superblocks
		 * feature is turned on, then not all groups have this.
		 */
		for (i = 0; i < ngroups; i++) {
3144 3145
			overhead += ext4_bg_has_super(sb, i) +
				ext4_bg_num_gdb(sb, i);
3146 3147 3148 3149 3150 3151 3152
			cond_resched();
		}

		/*
		 * Every block group has an inode bitmap, a block
		 * bitmap, and an inode table.
		 */
B
Badari Pulavarty 已提交
3153 3154 3155
		overhead += ngroups * (2 + sbi->s_itb_per_group);
		sbi->s_overhead_last = overhead;
		smp_wmb();
3156
		sbi->s_blocks_last = ext4_blocks_count(es);
3157 3158
	}

3159
	buf->f_type = EXT4_SUPER_MAGIC;
3160
	buf->f_bsize = sb->s_blocksize;
B
Badari Pulavarty 已提交
3161
	buf->f_blocks = ext4_blocks_count(es) - sbi->s_overhead_last;
3162 3163
	buf->f_bfree = percpu_counter_sum_positive(&sbi->s_freeblocks_counter) -
		       percpu_counter_sum_positive(&sbi->s_dirtyblocks_counter);
3164
	ext4_free_blocks_count_set(es, buf->f_bfree);
L
Laurent Vivier 已提交
3165 3166
	buf->f_bavail = buf->f_bfree - ext4_r_blocks_count(es);
	if (buf->f_bfree < ext4_r_blocks_count(es))
3167 3168
		buf->f_bavail = 0;
	buf->f_files = le32_to_cpu(es->s_inodes_count);
P
Peter Zijlstra 已提交
3169
	buf->f_ffree = percpu_counter_sum_positive(&sbi->s_freeinodes_counter);
B
Badari Pulavarty 已提交
3170
	es->s_free_inodes_count = cpu_to_le32(buf->f_ffree);
3171
	buf->f_namelen = EXT4_NAME_LEN;
P
Pekka Enberg 已提交
3172 3173 3174 3175
	fsid = le64_to_cpup((void *)es->s_uuid) ^
	       le64_to_cpup((void *)es->s_uuid + sizeof(u64));
	buf->f_fsid.val[0] = fsid & 0xFFFFFFFFUL;
	buf->f_fsid.val[1] = (fsid >> 32) & 0xFFFFFFFFUL;
3176 3177 3178 3179 3180 3181
	return 0;
}

/* Helper function for writing quotas on sync - we need to start transaction before quota file
 * is locked for write. Otherwise the are possible deadlocks:
 * Process 1                         Process 2
3182
 * ext4_create()                     quota_sync()
3183
 *   jbd2_journal_start()                   write_dquot()
3184
 *   DQUOT_INIT()                        down(dqio_mutex)
3185
 *     down(dqio_mutex)                    jbd2_journal_start()
3186 3187 3188 3189 3190 3191 3192 3193 3194 3195
 *
 */

#ifdef CONFIG_QUOTA

static inline struct inode *dquot_to_inode(struct dquot *dquot)
{
	return sb_dqopt(dquot->dq_sb)->files[dquot->dq_type];
}

3196
static int ext4_dquot_initialize(struct inode *inode, int type)
3197 3198 3199 3200 3201
{
	handle_t *handle;
	int ret, err;

	/* We may create quota structure so we need to reserve enough blocks */
3202
	handle = ext4_journal_start(inode, 2*EXT4_QUOTA_INIT_BLOCKS(inode->i_sb));
3203 3204 3205
	if (IS_ERR(handle))
		return PTR_ERR(handle);
	ret = dquot_initialize(inode, type);
3206
	err = ext4_journal_stop(handle);
3207 3208 3209 3210 3211
	if (!ret)
		ret = err;
	return ret;
}

3212
static int ext4_dquot_drop(struct inode *inode)
3213 3214 3215 3216 3217
{
	handle_t *handle;
	int ret, err;

	/* We may delete quota structure so we need to reserve enough blocks */
3218
	handle = ext4_journal_start(inode, 2*EXT4_QUOTA_DEL_BLOCKS(inode->i_sb));
3219 3220 3221 3222 3223 3224
	if (IS_ERR(handle)) {
		/*
		 * We call dquot_drop() anyway to at least release references
		 * to quota structures so that umount does not hang.
		 */
		dquot_drop(inode);
3225
		return PTR_ERR(handle);
3226
	}
3227
	ret = dquot_drop(inode);
3228
	err = ext4_journal_stop(handle);
3229 3230 3231 3232 3233
	if (!ret)
		ret = err;
	return ret;
}

3234
static int ext4_write_dquot(struct dquot *dquot)
3235 3236 3237 3238 3239 3240
{
	int ret, err;
	handle_t *handle;
	struct inode *inode;

	inode = dquot_to_inode(dquot);
3241 3242
	handle = ext4_journal_start(inode,
					EXT4_QUOTA_TRANS_BLOCKS(dquot->dq_sb));
3243 3244 3245
	if (IS_ERR(handle))
		return PTR_ERR(handle);
	ret = dquot_commit(dquot);
3246
	err = ext4_journal_stop(handle);
3247 3248 3249 3250 3251
	if (!ret)
		ret = err;
	return ret;
}

3252
static int ext4_acquire_dquot(struct dquot *dquot)
3253 3254 3255 3256
{
	int ret, err;
	handle_t *handle;

3257 3258
	handle = ext4_journal_start(dquot_to_inode(dquot),
					EXT4_QUOTA_INIT_BLOCKS(dquot->dq_sb));
3259 3260 3261
	if (IS_ERR(handle))
		return PTR_ERR(handle);
	ret = dquot_acquire(dquot);
3262
	err = ext4_journal_stop(handle);
3263 3264 3265 3266 3267
	if (!ret)
		ret = err;
	return ret;
}

3268
static int ext4_release_dquot(struct dquot *dquot)
3269 3270 3271 3272
{
	int ret, err;
	handle_t *handle;

3273 3274
	handle = ext4_journal_start(dquot_to_inode(dquot),
					EXT4_QUOTA_DEL_BLOCKS(dquot->dq_sb));
J
Jan Kara 已提交
3275 3276 3277
	if (IS_ERR(handle)) {
		/* Release dquot anyway to avoid endless cycle in dqput() */
		dquot_release(dquot);
3278
		return PTR_ERR(handle);
J
Jan Kara 已提交
3279
	}
3280
	ret = dquot_release(dquot);
3281
	err = ext4_journal_stop(handle);
3282 3283 3284 3285 3286
	if (!ret)
		ret = err;
	return ret;
}

3287
static int ext4_mark_dquot_dirty(struct dquot *dquot)
3288
{
3289
	/* Are we journaling quotas? */
3290 3291
	if (EXT4_SB(dquot->dq_sb)->s_qf_names[USRQUOTA] ||
	    EXT4_SB(dquot->dq_sb)->s_qf_names[GRPQUOTA]) {
3292
		dquot_mark_dquot_dirty(dquot);
3293
		return ext4_write_dquot(dquot);
3294 3295 3296 3297 3298
	} else {
		return dquot_mark_dquot_dirty(dquot);
	}
}

3299
static int ext4_write_info(struct super_block *sb, int type)
3300 3301 3302 3303 3304
{
	int ret, err;
	handle_t *handle;

	/* Data block + inode block */
3305
	handle = ext4_journal_start(sb->s_root->d_inode, 2);
3306 3307 3308
	if (IS_ERR(handle))
		return PTR_ERR(handle);
	ret = dquot_commit_info(sb, type);
3309
	err = ext4_journal_stop(handle);
3310 3311 3312 3313 3314 3315 3316 3317 3318
	if (!ret)
		ret = err;
	return ret;
}

/*
 * Turn on quotas during mount time - we need to find
 * the quota file and such...
 */
3319
static int ext4_quota_on_mount(struct super_block *sb, int type)
3320
{
3321 3322
	return vfs_quota_on_mount(sb, EXT4_SB(sb)->s_qf_names[type],
			EXT4_SB(sb)->s_jquota_fmt, type);
3323 3324 3325 3326 3327
}

/*
 * Standard function to be called on quota_on
 */
3328
static int ext4_quota_on(struct super_block *sb, int type, int format_id,
3329
			 char *name, int remount)
3330 3331
{
	int err;
3332
	struct path path;
3333 3334 3335

	if (!test_opt(sb, QUOTA))
		return -EINVAL;
3336
	/* When remounting, no checks are needed and in fact, name is NULL */
3337
	if (remount)
3338
		return vfs_quota_on(sb, type, format_id, name, remount);
3339

3340
	err = kern_path(name, LOOKUP_FOLLOW, &path);
3341 3342
	if (err)
		return err;
3343

3344
	/* Quotafile not on the same filesystem? */
3345 3346
	if (path.mnt->mnt_sb != sb) {
		path_put(&path);
3347 3348
		return -EXDEV;
	}
3349 3350
	/* Journaling quota? */
	if (EXT4_SB(sb)->s_qf_names[type]) {
3351
		/* Quotafile not in fs root? */
3352
		if (path.dentry->d_parent != sb->s_root)
3353 3354 3355
			printk(KERN_WARNING
				"EXT4-fs: Quota file not on filesystem root. "
				"Journaled quota will not work.\n");
3356
	}
3357 3358 3359 3360 3361

	/*
	 * When we journal data on quota file, we have to flush journal to see
	 * all updates to the file when we bypass pagecache...
	 */
3362
	if (ext4_should_journal_data(path.dentry->d_inode)) {
3363 3364 3365 3366 3367
		/*
		 * We don't need to lock updates but journal_flush() could
		 * otherwise be livelocked...
		 */
		jbd2_journal_lock_updates(EXT4_SB(sb)->s_journal);
3368
		err = jbd2_journal_flush(EXT4_SB(sb)->s_journal);
3369
		jbd2_journal_unlock_updates(EXT4_SB(sb)->s_journal);
3370
		if (err) {
3371
			path_put(&path);
3372 3373
			return err;
		}
3374 3375
	}

3376 3377
	err = vfs_quota_on_path(sb, type, format_id, &path);
	path_put(&path);
3378
	return err;
3379 3380 3381 3382 3383 3384
}

/* Read data from quotafile - avoid pagecache and such because we cannot afford
 * acquiring the locks... As quota files are never truncated and quota code
 * itself serializes the operations (and noone else should touch the files)
 * we don't have to be afraid of races */
3385
static ssize_t ext4_quota_read(struct super_block *sb, int type, char *data,
3386 3387 3388
			       size_t len, loff_t off)
{
	struct inode *inode = sb_dqopt(sb)->files[type];
A
Aneesh Kumar K.V 已提交
3389
	ext4_lblk_t blk = off >> EXT4_BLOCK_SIZE_BITS(sb);
3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404
	int err = 0;
	int offset = off & (sb->s_blocksize - 1);
	int tocopy;
	size_t toread;
	struct buffer_head *bh;
	loff_t i_size = i_size_read(inode);

	if (off > i_size)
		return 0;
	if (off+len > i_size)
		len = i_size-off;
	toread = len;
	while (toread > 0) {
		tocopy = sb->s_blocksize - offset < toread ?
				sb->s_blocksize - offset : toread;
3405
		bh = ext4_bread(NULL, inode, blk, 0, &err);
3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422
		if (err)
			return err;
		if (!bh)	/* A hole? */
			memset(data, 0, tocopy);
		else
			memcpy(data, bh->b_data+offset, tocopy);
		brelse(bh);
		offset = 0;
		toread -= tocopy;
		data += tocopy;
		blk++;
	}
	return len;
}

/* Write to quotafile (we know the transaction is already started and has
 * enough credits) */
3423
static ssize_t ext4_quota_write(struct super_block *sb, int type,
3424 3425 3426
				const char *data, size_t len, loff_t off)
{
	struct inode *inode = sb_dqopt(sb)->files[type];
A
Aneesh Kumar K.V 已提交
3427
	ext4_lblk_t blk = off >> EXT4_BLOCK_SIZE_BITS(sb);
3428 3429 3430
	int err = 0;
	int offset = off & (sb->s_blocksize - 1);
	int tocopy;
3431
	int journal_quota = EXT4_SB(sb)->s_qf_names[type] != NULL;
3432 3433 3434 3435
	size_t towrite = len;
	struct buffer_head *bh;
	handle_t *handle = journal_current_handle();

J
Jan Kara 已提交
3436
	if (!handle) {
3437
		printk(KERN_WARNING "EXT4-fs: Quota write (off=%llu, len=%llu)"
J
Jan Kara 已提交
3438 3439 3440 3441
			" cancelled because transaction is not started.\n",
			(unsigned long long)off, (unsigned long long)len);
		return -EIO;
	}
3442 3443 3444 3445
	mutex_lock_nested(&inode->i_mutex, I_MUTEX_QUOTA);
	while (towrite > 0) {
		tocopy = sb->s_blocksize - offset < towrite ?
				sb->s_blocksize - offset : towrite;
3446
		bh = ext4_bread(handle, inode, blk, 1, &err);
3447 3448 3449
		if (!bh)
			goto out;
		if (journal_quota) {
3450
			err = ext4_journal_get_write_access(handle, bh);
3451 3452 3453 3454 3455 3456 3457 3458 3459 3460
			if (err) {
				brelse(bh);
				goto out;
			}
		}
		lock_buffer(bh);
		memcpy(bh->b_data+offset, data, tocopy);
		flush_dcache_page(bh->b_page);
		unlock_buffer(bh);
		if (journal_quota)
3461
			err = ext4_journal_dirty_metadata(handle, bh);
3462 3463
		else {
			/* Always do at least ordered writes for quotas */
3464
			err = ext4_jbd2_file_inode(handle, inode);
3465 3466 3467 3468 3469 3470 3471 3472 3473 3474 3475
			mark_buffer_dirty(bh);
		}
		brelse(bh);
		if (err)
			goto out;
		offset = 0;
		towrite -= tocopy;
		data += tocopy;
		blk++;
	}
out:
3476 3477
	if (len == towrite) {
		mutex_unlock(&inode->i_mutex);
3478
		return err;
3479
	}
3480 3481
	if (inode->i_size < off+len-towrite) {
		i_size_write(inode, off+len-towrite);
3482
		EXT4_I(inode)->i_disksize = inode->i_size;
3483 3484
	}
	inode->i_mtime = inode->i_ctime = CURRENT_TIME;
3485
	ext4_mark_inode_dirty(handle, inode);
3486 3487 3488 3489 3490 3491
	mutex_unlock(&inode->i_mutex);
	return len - towrite;
}

#endif

3492
static int ext4_get_sb(struct file_system_type *fs_type,
3493 3494
	int flags, const char *dev_name, void *data, struct vfsmount *mnt)
{
3495
	return get_sb_bdev(fs_type, flags, dev_name, data, ext4_fill_super, mnt);
3496 3497
}

3498 3499 3500 3501 3502 3503 3504 3505 3506 3507 3508 3509 3510 3511 3512 3513 3514 3515 3516 3517 3518 3519 3520 3521 3522 3523 3524 3525 3526 3527 3528 3529 3530 3531 3532 3533 3534 3535 3536 3537 3538 3539
#ifdef CONFIG_PROC_FS
static int ext4_ui_proc_show(struct seq_file *m, void *v)
{
	unsigned int *p = m->private;

	seq_printf(m, "%u\n", *p);
	return 0;
}

static int ext4_ui_proc_open(struct inode *inode, struct file *file)
{
	return single_open(file, ext4_ui_proc_show, PDE(inode)->data);
}

static ssize_t ext4_ui_proc_write(struct file *file, const char __user *buf,
			       size_t cnt, loff_t *ppos)
{
	unsigned int *p = PDE(file->f_path.dentry->d_inode)->data;
	char str[32];
	unsigned long value;

	if (cnt >= sizeof(str))
		return -EINVAL;
	if (copy_from_user(str, buf, cnt))
		return -EFAULT;
	value = simple_strtol(str, NULL, 0);
	if (value < 0)
		return -ERANGE;
	*p = value;
	return cnt;
}

const struct file_operations ext4_ui_proc_fops = {
	.owner		= THIS_MODULE,
	.open		= ext4_ui_proc_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
	.release	= single_release,
	.write		= ext4_ui_proc_write,
};
#endif

T
Theodore Ts'o 已提交
3540 3541 3542 3543 3544 3545 3546 3547 3548 3549 3550 3551 3552 3553 3554 3555 3556 3557 3558
static struct file_system_type ext4_fs_type = {
	.owner		= THIS_MODULE,
	.name		= "ext4",
	.get_sb		= ext4_get_sb,
	.kill_sb	= kill_block_super,
	.fs_flags	= FS_REQUIRES_DEV,
};

#ifdef CONFIG_EXT4DEV_COMPAT
static int ext4dev_get_sb(struct file_system_type *fs_type,
	int flags, const char *dev_name, void *data, struct vfsmount *mnt)
{
	printk(KERN_WARNING "EXT4-fs: Update your userspace programs "
	       "to mount using ext4\n");
	printk(KERN_WARNING "EXT4-fs: ext4dev backwards compatibility "
	       "will go away by 2.6.31\n");
	return get_sb_bdev(fs_type, flags, dev_name, data, ext4_fill_super, mnt);
}

3559
static struct file_system_type ext4dev_fs_type = {
3560
	.owner		= THIS_MODULE,
3561
	.name		= "ext4dev",
T
Theodore Ts'o 已提交
3562
	.get_sb		= ext4dev_get_sb,
3563 3564 3565
	.kill_sb	= kill_block_super,
	.fs_flags	= FS_REQUIRES_DEV,
};
T
Theodore Ts'o 已提交
3566 3567
MODULE_ALIAS("ext4dev");
#endif
3568

3569
static int __init init_ext4_fs(void)
3570
{
3571 3572
	int err;

3573
	ext4_proc_root = proc_mkdir("fs/ext4", NULL);
3574
	err = init_ext4_mballoc();
3575 3576
	if (err)
		return err;
3577 3578 3579 3580

	err = init_ext4_xattr();
	if (err)
		goto out2;
3581 3582 3583
	err = init_inodecache();
	if (err)
		goto out1;
T
Theodore Ts'o 已提交
3584
	err = register_filesystem(&ext4_fs_type);
3585 3586
	if (err)
		goto out;
T
Theodore Ts'o 已提交
3587 3588 3589 3590 3591 3592 3593
#ifdef CONFIG_EXT4DEV_COMPAT
	err = register_filesystem(&ext4dev_fs_type);
	if (err) {
		unregister_filesystem(&ext4_fs_type);
		goto out;
	}
#endif
3594 3595 3596 3597
	return 0;
out:
	destroy_inodecache();
out1:
3598
	exit_ext4_xattr();
3599 3600
out2:
	exit_ext4_mballoc();
3601 3602 3603
	return err;
}

3604
static void __exit exit_ext4_fs(void)
3605
{
T
Theodore Ts'o 已提交
3606 3607
	unregister_filesystem(&ext4_fs_type);
#ifdef CONFIG_EXT4DEV_COMPAT
3608
	unregister_filesystem(&ext4dev_fs_type);
T
Theodore Ts'o 已提交
3609
#endif
3610
	destroy_inodecache();
3611
	exit_ext4_xattr();
3612
	exit_ext4_mballoc();
3613
	remove_proc_entry("fs/ext4", NULL);
3614 3615 3616
}

MODULE_AUTHOR("Remy Card, Stephen Tweedie, Andrew Morton, Andreas Dilger, Theodore Ts'o and others");
3617
MODULE_DESCRIPTION("Fourth Extended Filesystem with extents");
3618
MODULE_LICENSE("GPL");
3619 3620
module_init(init_ext4_fs)
module_exit(exit_ext4_fs)