super.c 51.8 KB
Newer Older
C
Chris Mason 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
/*
 * Copyright (C) 2007 Oracle.  All rights reserved.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public
 * License v2 as published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * General Public License for more details.
 *
 * You should have received a copy of the GNU General Public
 * License along with this program; if not, write to the
 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
 * Boston, MA 021110-1307, USA.
 */

Y
Yan 已提交
19
#include <linux/blkdev.h>
20
#include <linux/module.h>
C
Chris Mason 已提交
21
#include <linux/buffer_head.h>
22 23 24 25 26
#include <linux/fs.h>
#include <linux/pagemap.h>
#include <linux/highmem.h>
#include <linux/time.h>
#include <linux/init.h>
E
Eric Paris 已提交
27
#include <linux/seq_file.h>
28 29
#include <linux/string.h>
#include <linux/backing-dev.h>
Y
Yan 已提交
30
#include <linux/mount.h>
C
Chris Mason 已提交
31
#include <linux/mpage.h>
C
Chris Mason 已提交
32 33
#include <linux/swap.h>
#include <linux/writeback.h>
C
Chris Mason 已提交
34
#include <linux/statfs.h>
C
Chris Mason 已提交
35
#include <linux/compat.h>
36
#include <linux/parser.h>
37
#include <linux/ctype.h>
38
#include <linux/namei.h>
39
#include <linux/miscdevice.h>
40
#include <linux/magic.h>
41
#include <linux/slab.h>
D
Dan Magenheimer 已提交
42
#include <linux/cleancache.h>
43
#include <linux/ratelimit.h>
44
#include <linux/btrfs.h>
45
#include "delayed-inode.h"
46
#include "ctree.h"
C
Chris Mason 已提交
47
#include "disk-io.h"
48
#include "transaction.h"
C
Chris Mason 已提交
49
#include "btrfs_inode.h"
C
Chris Mason 已提交
50
#include "print-tree.h"
51
#include "hash.h"
52
#include "props.h"
J
Josef Bacik 已提交
53
#include "xattr.h"
54
#include "volumes.h"
B
Balaji Rao 已提交
55
#include "export.h"
C
Chris Mason 已提交
56
#include "compression.h"
J
Josef Bacik 已提交
57
#include "rcu-string.h"
58
#include "dev-replace.h"
59
#include "free-space-cache.h"
60
#include "backref.h"
61
#include "tests/btrfs-tests.h"
62

63 64 65
#define CREATE_TRACE_POINTS
#include <trace/events/btrfs.h>

66
static const struct super_operations btrfs_super_ops;
67
static struct file_system_type btrfs_fs_type;
C
Chris Mason 已提交
68

69
static const char *btrfs_decode_error(int errno)
L
liubo 已提交
70
{
71
	char *errstr = "unknown";
L
liubo 已提交
72 73 74 75 76 77 78 79 80 81 82

	switch (errno) {
	case -EIO:
		errstr = "IO failure";
		break;
	case -ENOMEM:
		errstr = "Out of memory";
		break;
	case -EROFS:
		errstr = "Readonly filesystem";
		break;
J
Jeff Mahoney 已提交
83 84 85
	case -EEXIST:
		errstr = "Object already exists";
		break;
86 87 88 89 90 91
	case -ENOSPC:
		errstr = "No space left";
		break;
	case -ENOENT:
		errstr = "No such entry";
		break;
L
liubo 已提交
92 93 94 95 96
	}

	return errstr;
}

97
static void save_error_info(struct btrfs_fs_info *fs_info)
L
liubo 已提交
98 99 100 101 102
{
	/*
	 * today we only save the error info into ram.  Long term we'll
	 * also send it down to the disk
	 */
103
	set_bit(BTRFS_FS_STATE_ERROR, &fs_info->fs_state);
L
liubo 已提交
104 105 106 107 108 109 110 111 112 113
}

/* btrfs handle error by forcing the filesystem readonly */
static void btrfs_handle_error(struct btrfs_fs_info *fs_info)
{
	struct super_block *sb = fs_info->sb;

	if (sb->s_flags & MS_RDONLY)
		return;

114
	if (test_bit(BTRFS_FS_STATE_ERROR, &fs_info->fs_state)) {
L
liubo 已提交
115
		sb->s_flags |= MS_RDONLY;
116
		btrfs_info(fs_info, "forced readonly");
117 118 119 120 121 122 123 124 125 126
		/*
		 * Note that a running device replace operation is not
		 * canceled here although there is no way to update
		 * the progress. It would add the risk of a deadlock,
		 * therefore the canceling is ommited. The only penalty
		 * is that some I/O remains active until the procedure
		 * completes. The next time when the filesystem is
		 * mounted writeable again, the device replace
		 * operation continues.
		 */
L
liubo 已提交
127 128 129
	}
}

130
#ifdef CONFIG_PRINTK
L
liubo 已提交
131 132 133 134 135
/*
 * __btrfs_std_error decodes expected errors from the caller and
 * invokes the approciate error response.
 */
void __btrfs_std_error(struct btrfs_fs_info *fs_info, const char *function,
J
Jeff Mahoney 已提交
136
		       unsigned int line, int errno, const char *fmt, ...)
L
liubo 已提交
137 138 139 140 141 142 143 144 145
{
	struct super_block *sb = fs_info->sb;
	const char *errstr;

	/*
	 * Special case: if the error is EROFS, and we're already
	 * under MS_RDONLY, then it is safe here.
	 */
	if (errno == -EROFS && (sb->s_flags & MS_RDONLY))
J
Jeff Mahoney 已提交
146 147
  		return;

148
	errstr = btrfs_decode_error(errno);
J
Jeff Mahoney 已提交
149
	if (fmt) {
150 151 152 153 154 155
		struct va_format vaf;
		va_list args;

		va_start(args, fmt);
		vaf.fmt = fmt;
		vaf.va = &args;
J
Jeff Mahoney 已提交
156

157 158
		printk(KERN_CRIT
			"BTRFS: error (device %s) in %s:%d: errno=%d %s (%pV)\n",
159
			sb->s_id, function, line, errno, errstr, &vaf);
160
		va_end(args);
J
Jeff Mahoney 已提交
161
	} else {
162
		printk(KERN_CRIT "BTRFS: error (device %s) in %s:%d: errno=%d %s\n",
163
			sb->s_id, function, line, errno, errstr);
J
Jeff Mahoney 已提交
164
	}
L
liubo 已提交
165

J
Jeff Mahoney 已提交
166
	/* Don't go through full error handling during mount */
167 168
	save_error_info(fs_info);
	if (sb->s_flags & MS_BORN)
J
Jeff Mahoney 已提交
169 170
		btrfs_handle_error(fs_info);
}
L
liubo 已提交
171

172
static const char * const logtypes[] = {
J
Jeff Mahoney 已提交
173 174 175 176 177 178 179 180 181 182
	"emergency",
	"alert",
	"critical",
	"error",
	"warning",
	"notice",
	"info",
	"debug",
};

183
void btrfs_printk(const struct btrfs_fs_info *fs_info, const char *fmt, ...)
J
Jeff Mahoney 已提交
184 185 186 187 188 189
{
	struct super_block *sb = fs_info->sb;
	char lvl[4];
	struct va_format vaf;
	va_list args;
	const char *type = logtypes[4];
190
	int kern_level;
J
Jeff Mahoney 已提交
191 192 193

	va_start(args, fmt);

194 195 196 197 198 199 200
	kern_level = printk_get_level(fmt);
	if (kern_level) {
		size_t size = printk_skip_level(fmt) - fmt;
		memcpy(lvl, fmt,  size);
		lvl[size] = '\0';
		fmt += size;
		type = logtypes[kern_level - '0'];
J
Jeff Mahoney 已提交
201 202 203 204 205
	} else
		*lvl = '\0';

	vaf.fmt = fmt;
	vaf.va = &args;
206

207
	printk("%sBTRFS %s (device %s): %pV\n", lvl, type, sb->s_id, &vaf);
208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230

	va_end(args);
}

#else

void __btrfs_std_error(struct btrfs_fs_info *fs_info, const char *function,
		       unsigned int line, int errno, const char *fmt, ...)
{
	struct super_block *sb = fs_info->sb;

	/*
	 * Special case: if the error is EROFS, and we're already
	 * under MS_RDONLY, then it is safe here.
	 */
	if (errno == -EROFS && (sb->s_flags & MS_RDONLY))
		return;

	/* Don't go through full error handling during mount */
	if (sb->s_flags & MS_BORN) {
		save_error_info(fs_info);
		btrfs_handle_error(fs_info);
	}
L
liubo 已提交
231
}
232
#endif
L
liubo 已提交
233

234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250
/*
 * We only mark the transaction aborted and then set the file system read-only.
 * This will prevent new transactions from starting or trying to join this
 * one.
 *
 * This means that error recovery at the call site is limited to freeing
 * any local memory allocations and passing the error code up without
 * further cleanup. The transaction should complete as it normally would
 * in the call path but will return -EIO.
 *
 * We'll complete the cleanup in btrfs_end_transaction and
 * btrfs_commit_transaction.
 */
void __btrfs_abort_transaction(struct btrfs_trans_handle *trans,
			       struct btrfs_root *root, const char *function,
			       unsigned int line, int errno)
{
251 252 253 254 255
	/*
	 * Report first abort since mount
	 */
	if (!test_and_set_bit(BTRFS_FS_STATE_TRANS_ABORTED,
				&root->fs_info->fs_state)) {
256
		WARN(1, KERN_DEBUG "BTRFS: Transaction aborted (error %d)\n",
257 258
				errno);
	}
259 260 261 262
	trans->aborted = errno;
	/* Nothing used. The other threads that have joined this
	 * transaction may be able to continue. */
	if (!trans->blocks_used) {
263 264
		const char *errstr;

265
		errstr = btrfs_decode_error(errno);
266 267 268
		btrfs_warn(root->fs_info,
		           "%s:%d: Aborting unused transaction(%s).",
		           function, line, errstr);
L
liubo 已提交
269
		return;
270
	}
271
	ACCESS_ONCE(trans->transaction->aborted) = errno;
272 273 274
	/* Wake up anybody who may be waiting on this transaction */
	wake_up(&root->fs_info->transaction_wait);
	wake_up(&root->fs_info->transaction_blocked_wait);
275 276
	__btrfs_std_error(root->fs_info, function, line, errno, NULL);
}
J
Jeff Mahoney 已提交
277 278 279 280 281 282 283 284 285 286 287
/*
 * __btrfs_panic decodes unexpected, fatal errors from the caller,
 * issues an alert, and either panics or BUGs, depending on mount options.
 */
void __btrfs_panic(struct btrfs_fs_info *fs_info, const char *function,
		   unsigned int line, int errno, const char *fmt, ...)
{
	char *s_id = "<unknown>";
	const char *errstr;
	struct va_format vaf = { .fmt = fmt };
	va_list args;
L
liubo 已提交
288

J
Jeff Mahoney 已提交
289 290
	if (fs_info)
		s_id = fs_info->sb->s_id;
L
liubo 已提交
291

J
Jeff Mahoney 已提交
292 293 294
	va_start(args, fmt);
	vaf.va = &args;

295
	errstr = btrfs_decode_error(errno);
296
	if (fs_info && (fs_info->mount_opt & BTRFS_MOUNT_PANIC_ON_FATAL_ERROR))
297 298
		panic(KERN_CRIT "BTRFS panic (device %s) in %s:%d: %pV (errno=%d %s)\n",
			s_id, function, line, &vaf, errno, errstr);
J
Jeff Mahoney 已提交
299

300 301
	btrfs_crit(fs_info, "panic in %s:%d: %pV (errno=%d %s)",
		   function, line, &vaf, errno, errstr);
J
Jeff Mahoney 已提交
302 303
	va_end(args);
	/* Caller calls BUG() */
L
liubo 已提交
304 305
}

C
Chris Mason 已提交
306
static void btrfs_put_super(struct super_block *sb)
C
Chris Mason 已提交
307
{
308
	(void)close_ctree(btrfs_sb(sb)->tree_root);
A
Al Viro 已提交
309 310 311 312 313 314
	/* FIXME: need to fix VFS to return error? */
	/* AV: return it _where_?  ->put_super() can be triggered by any number
	 * of async events, up to and including delivery of SIGKILL to the
	 * last process that kept it busy.  Or segfault in the aforementioned
	 * process...  Whom would you report that to?
	 */
C
Chris Mason 已提交
315 316
}

317
enum {
318
	Opt_degraded, Opt_subvol, Opt_subvolid, Opt_device, Opt_nodatasum,
319 320
	Opt_nodatacow, Opt_max_inline, Opt_alloc_start, Opt_nobarrier, Opt_ssd,
	Opt_nossd, Opt_ssd_spread, Opt_thread_pool, Opt_noacl, Opt_compress,
321 322
	Opt_compress_type, Opt_compress_force, Opt_compress_force_type,
	Opt_notreelog, Opt_ratio, Opt_flushoncommit, Opt_discard,
323
	Opt_space_cache, Opt_clear_cache, Opt_user_subvol_rm_allowed,
324 325
	Opt_enospc_debug, Opt_subvolrootid, Opt_defrag, Opt_inode_cache,
	Opt_no_space_cache, Opt_recovery, Opt_skip_balance,
326
	Opt_check_integrity, Opt_check_integrity_including_extent_data,
327
	Opt_check_integrity_print_mask, Opt_fatal_errors, Opt_rescan_uuid_tree,
Q
Qu Wenruo 已提交
328
	Opt_commit_interval, Opt_barrier, Opt_nodefrag, Opt_nodiscard,
Q
Qu Wenruo 已提交
329
	Opt_noenospc_debug, Opt_noflushoncommit, Opt_acl, Opt_datacow,
330
	Opt_datasum, Opt_treelog, Opt_noinode_cache,
331
	Opt_err,
332 333 334
};

static match_table_t tokens = {
335
	{Opt_degraded, "degraded"},
336
	{Opt_subvol, "subvol=%s"},
337
	{Opt_subvolid, "subvolid=%s"},
338
	{Opt_device, "device=%s"},
339
	{Opt_nodatasum, "nodatasum"},
Q
Qu Wenruo 已提交
340
	{Opt_datasum, "datasum"},
341
	{Opt_nodatacow, "nodatacow"},
Q
Qu Wenruo 已提交
342
	{Opt_datacow, "datacow"},
343
	{Opt_nobarrier, "nobarrier"},
344
	{Opt_barrier, "barrier"},
345
	{Opt_max_inline, "max_inline=%s"},
346
	{Opt_alloc_start, "alloc_start=%s"},
347
	{Opt_thread_pool, "thread_pool=%d"},
C
Chris Mason 已提交
348
	{Opt_compress, "compress"},
349
	{Opt_compress_type, "compress=%s"},
C
Chris Mason 已提交
350
	{Opt_compress_force, "compress-force"},
351
	{Opt_compress_force_type, "compress-force=%s"},
352
	{Opt_ssd, "ssd"},
353
	{Opt_ssd_spread, "ssd_spread"},
C
Chris Mason 已提交
354
	{Opt_nossd, "nossd"},
Q
Qu Wenruo 已提交
355
	{Opt_acl, "acl"},
J
Josef Bacik 已提交
356
	{Opt_noacl, "noacl"},
S
Sage Weil 已提交
357
	{Opt_notreelog, "notreelog"},
Q
Qu Wenruo 已提交
358
	{Opt_treelog, "treelog"},
359
	{Opt_flushoncommit, "flushoncommit"},
360
	{Opt_noflushoncommit, "noflushoncommit"},
361
	{Opt_ratio, "metadata_ratio=%d"},
C
Christoph Hellwig 已提交
362
	{Opt_discard, "discard"},
Q
Qu Wenruo 已提交
363
	{Opt_nodiscard, "nodiscard"},
364
	{Opt_space_cache, "space_cache"},
365
	{Opt_clear_cache, "clear_cache"},
366
	{Opt_user_subvol_rm_allowed, "user_subvol_rm_allowed"},
367
	{Opt_enospc_debug, "enospc_debug"},
368
	{Opt_noenospc_debug, "noenospc_debug"},
369
	{Opt_subvolrootid, "subvolrootid=%d"},
C
Chris Mason 已提交
370
	{Opt_defrag, "autodefrag"},
371
	{Opt_nodefrag, "noautodefrag"},
C
Chris Mason 已提交
372
	{Opt_inode_cache, "inode_cache"},
373
	{Opt_noinode_cache, "noinode_cache"},
374
	{Opt_no_space_cache, "nospace_cache"},
C
Chris Mason 已提交
375
	{Opt_recovery, "recovery"},
376
	{Opt_skip_balance, "skip_balance"},
377 378 379
	{Opt_check_integrity, "check_int"},
	{Opt_check_integrity_including_extent_data, "check_int_data"},
	{Opt_check_integrity_print_mask, "check_int_print_mask=%d"},
380
	{Opt_rescan_uuid_tree, "rescan_uuid_tree"},
J
Jeff Mahoney 已提交
381
	{Opt_fatal_errors, "fatal_errors=%s"},
382
	{Opt_commit_interval, "commit=%d"},
J
Josef Bacik 已提交
383
	{Opt_err, NULL},
384 385
};

386 387 388 389 390 391 392 393 394 395 396 397 398 399
#define btrfs_set_and_info(root, opt, fmt, args...)			\
{									\
	if (!btrfs_test_opt(root, opt))					\
		btrfs_info(root->fs_info, fmt, ##args);			\
	btrfs_set_opt(root->fs_info->mount_opt, opt);			\
}

#define btrfs_clear_and_info(root, opt, fmt, args...)			\
{									\
	if (btrfs_test_opt(root, opt))					\
		btrfs_info(root->fs_info, fmt, ##args);			\
	btrfs_clear_opt(root->fs_info->mount_opt, opt);			\
}

400 401 402
/*
 * Regular mount options parser.  Everything that is needed only when
 * reading in a new superblock is parsed here.
403
 * XXX JDM: This needs to be cleaned up for remount.
404 405
 */
int btrfs_parse_options(struct btrfs_root *root, char *options)
406
{
407
	struct btrfs_fs_info *info = root->fs_info;
408
	substring_t args[MAX_OPT_ARGS];
409 410
	char *p, *num, *orig = NULL;
	u64 cache_gen;
411
	int intarg;
S
Sage Weil 已提交
412
	int ret = 0;
413 414
	char *compress_type;
	bool compress_force = false;
415
	bool compress = false;
416

417
	cache_gen = btrfs_super_cache_generation(root->fs_info->super_copy);
418 419 420
	if (cache_gen)
		btrfs_set_opt(info->mount_opt, SPACE_CACHE);

421
	if (!options)
422
		goto out;
423

424 425 426 427 428 429 430 431
	/*
	 * strsep changes the string, duplicate it because parse_options
	 * gets called twice
	 */
	options = kstrdup(options, GFP_NOFS);
	if (!options)
		return -ENOMEM;

432
	orig = options;
433

434
	while ((p = strsep(&options, ",")) != NULL) {
435 436 437 438 439 440
		int token;
		if (!*p)
			continue;

		token = match_token(p, tokens, args);
		switch (token) {
441
		case Opt_degraded:
442
			btrfs_info(root->fs_info, "allowing degraded mounts");
443
			btrfs_set_opt(info->mount_opt, DEGRADED);
444
			break;
445
		case Opt_subvol:
446
		case Opt_subvolid:
447
		case Opt_subvolrootid:
448
		case Opt_device:
449
			/*
450
			 * These are parsed by btrfs_parse_early_options
451 452
			 * and can be happily ignored here.
			 */
453 454
			break;
		case Opt_nodatasum:
455 456
			btrfs_set_and_info(root, NODATASUM,
					   "setting nodatasum");
457
			break;
Q
Qu Wenruo 已提交
458
		case Opt_datasum:
459 460 461 462 463 464
			if (btrfs_test_opt(root, NODATASUM)) {
				if (btrfs_test_opt(root, NODATACOW))
					btrfs_info(root->fs_info, "setting datasum, datacow enabled");
				else
					btrfs_info(root->fs_info, "setting datasum");
			}
Q
Qu Wenruo 已提交
465 466 467
			btrfs_clear_opt(info->mount_opt, NODATACOW);
			btrfs_clear_opt(info->mount_opt, NODATASUM);
			break;
468
		case Opt_nodatacow:
469 470 471
			if (!btrfs_test_opt(root, NODATACOW)) {
				if (!btrfs_test_opt(root, COMPRESS) ||
				    !btrfs_test_opt(root, FORCE_COMPRESS)) {
472
					btrfs_info(root->fs_info,
473 474 475 476
						   "setting nodatacow, compression disabled");
				} else {
					btrfs_info(root->fs_info, "setting nodatacow");
				}
477 478 479
			}
			btrfs_clear_opt(info->mount_opt, COMPRESS);
			btrfs_clear_opt(info->mount_opt, FORCE_COMPRESS);
480 481
			btrfs_set_opt(info->mount_opt, NODATACOW);
			btrfs_set_opt(info->mount_opt, NODATASUM);
482
			break;
Q
Qu Wenruo 已提交
483
		case Opt_datacow:
484 485
			btrfs_clear_and_info(root, NODATACOW,
					     "setting datacow");
Q
Qu Wenruo 已提交
486
			break;
C
Chris Mason 已提交
487
		case Opt_compress_force:
488 489
		case Opt_compress_force_type:
			compress_force = true;
490
			/* Fallthrough */
491 492
		case Opt_compress:
		case Opt_compress_type:
493
			compress = true;
494 495 496 497 498
			if (token == Opt_compress ||
			    token == Opt_compress_force ||
			    strcmp(args[0].from, "zlib") == 0) {
				compress_type = "zlib";
				info->compress_type = BTRFS_COMPRESS_ZLIB;
499
				btrfs_set_opt(info->mount_opt, COMPRESS);
500 501
				btrfs_clear_opt(info->mount_opt, NODATACOW);
				btrfs_clear_opt(info->mount_opt, NODATASUM);
L
Li Zefan 已提交
502 503 504
			} else if (strcmp(args[0].from, "lzo") == 0) {
				compress_type = "lzo";
				info->compress_type = BTRFS_COMPRESS_LZO;
505
				btrfs_set_opt(info->mount_opt, COMPRESS);
506 507
				btrfs_clear_opt(info->mount_opt, NODATACOW);
				btrfs_clear_opt(info->mount_opt, NODATASUM);
508
				btrfs_set_fs_incompat(info, COMPRESS_LZO);
509 510 511 512 513
			} else if (strncmp(args[0].from, "no", 2) == 0) {
				compress_type = "no";
				btrfs_clear_opt(info->mount_opt, COMPRESS);
				btrfs_clear_opt(info->mount_opt, FORCE_COMPRESS);
				compress_force = false;
514 515 516 517 518 519
			} else {
				ret = -EINVAL;
				goto out;
			}

			if (compress_force) {
520 521 522 523 524 525 526 527
				btrfs_set_and_info(root, FORCE_COMPRESS,
						   "force %s compression",
						   compress_type);
			} else if (compress) {
				if (!btrfs_test_opt(root, COMPRESS))
					btrfs_info(root->fs_info,
						   "btrfs: use %s compression\n",
						   compress_type);
528
			}
C
Chris Mason 已提交
529
			break;
530
		case Opt_ssd:
531 532
			btrfs_set_and_info(root, SSD,
					   "use ssd allocation scheme");
533
			break;
534
		case Opt_ssd_spread:
535 536
			btrfs_set_and_info(root, SSD_SPREAD,
					   "use spread ssd allocation scheme");
537
			break;
C
Chris Mason 已提交
538
		case Opt_nossd:
539 540
			btrfs_clear_and_info(root, NOSSD,
					     "not using ssd allocation scheme");
C
Chris Mason 已提交
541 542
			btrfs_clear_opt(info->mount_opt, SSD);
			break;
543
		case Opt_barrier:
544 545
			btrfs_clear_and_info(root, NOBARRIER,
					     "turning on barriers");
546
			break;
547
		case Opt_nobarrier:
548 549
			btrfs_set_and_info(root, NOBARRIER,
					   "turning off barriers");
550
			break;
551
		case Opt_thread_pool:
552 553 554 555
			ret = match_int(&args[0], &intarg);
			if (ret) {
				goto out;
			} else if (intarg > 0) {
556
				info->thread_pool_size = intarg;
557 558 559 560
			} else {
				ret = -EINVAL;
				goto out;
			}
561
			break;
562
		case Opt_max_inline:
563 564
			num = match_strdup(&args[0]);
			if (num) {
A
Akinobu Mita 已提交
565
				info->max_inline = memparse(num, NULL);
566 567
				kfree(num);

C
Chris Mason 已提交
568 569 570 571 572
				if (info->max_inline) {
					info->max_inline = max_t(u64,
						info->max_inline,
						root->sectorsize);
				}
573
				btrfs_info(root->fs_info, "max_inline at %llu",
574
					info->max_inline);
575 576 577
			} else {
				ret = -ENOMEM;
				goto out;
578 579
			}
			break;
580
		case Opt_alloc_start:
581 582
			num = match_strdup(&args[0]);
			if (num) {
M
Miao Xie 已提交
583
				mutex_lock(&info->chunk_mutex);
A
Akinobu Mita 已提交
584
				info->alloc_start = memparse(num, NULL);
M
Miao Xie 已提交
585
				mutex_unlock(&info->chunk_mutex);
586
				kfree(num);
587
				btrfs_info(root->fs_info, "allocations start at %llu",
588
					info->alloc_start);
589 590 591
			} else {
				ret = -ENOMEM;
				goto out;
592 593
			}
			break;
Q
Qu Wenruo 已提交
594 595 596
		case Opt_acl:
			root->fs_info->sb->s_flags |= MS_POSIXACL;
			break;
J
Josef Bacik 已提交
597 598 599
		case Opt_noacl:
			root->fs_info->sb->s_flags &= ~MS_POSIXACL;
			break;
S
Sage Weil 已提交
600
		case Opt_notreelog:
601 602
			btrfs_set_and_info(root, NOTREELOG,
					   "disabling tree log");
Q
Qu Wenruo 已提交
603 604
			break;
		case Opt_treelog:
605 606
			btrfs_clear_and_info(root, NOTREELOG,
					     "enabling tree log");
S
Sage Weil 已提交
607
			break;
608
		case Opt_flushoncommit:
609 610
			btrfs_set_and_info(root, FLUSHONCOMMIT,
					   "turning on flush-on-commit");
611
			break;
612
		case Opt_noflushoncommit:
613 614
			btrfs_clear_and_info(root, FLUSHONCOMMIT,
					     "turning off flush-on-commit");
615
			break;
616
		case Opt_ratio:
617 618 619 620
			ret = match_int(&args[0], &intarg);
			if (ret) {
				goto out;
			} else if (intarg >= 0) {
621
				info->metadata_ratio = intarg;
622
				btrfs_info(root->fs_info, "metadata ratio %d",
623
				       info->metadata_ratio);
624 625 626
			} else {
				ret = -EINVAL;
				goto out;
627 628
			}
			break;
C
Christoph Hellwig 已提交
629
		case Opt_discard:
630 631
			btrfs_set_and_info(root, DISCARD,
					   "turning on discard");
C
Christoph Hellwig 已提交
632
			break;
Q
Qu Wenruo 已提交
633
		case Opt_nodiscard:
634 635
			btrfs_clear_and_info(root, DISCARD,
					     "turning off discard");
Q
Qu Wenruo 已提交
636
			break;
637
		case Opt_space_cache:
638 639
			btrfs_set_and_info(root, SPACE_CACHE,
					   "enabling disk space caching");
640
			break;
641 642 643
		case Opt_rescan_uuid_tree:
			btrfs_set_opt(info->mount_opt, RESCAN_UUID_TREE);
			break;
644
		case Opt_no_space_cache:
645 646
			btrfs_clear_and_info(root, SPACE_CACHE,
					     "disabling disk space caching");
647
			break;
C
Chris Mason 已提交
648
		case Opt_inode_cache:
649 650
			btrfs_set_and_info(root, CHANGE_INODE_CACHE,
					   "enabling inode map caching");
651 652
			break;
		case Opt_noinode_cache:
653 654
			btrfs_clear_and_info(root, CHANGE_INODE_CACHE,
					     "disabling inode map caching");
C
Chris Mason 已提交
655
			break;
656
		case Opt_clear_cache:
657 658
			btrfs_set_and_info(root, CLEAR_CACHE,
					   "force clearing of disk cache");
659
			break;
660 661 662
		case Opt_user_subvol_rm_allowed:
			btrfs_set_opt(info->mount_opt, USER_SUBVOL_RM_ALLOWED);
			break;
663 664 665
		case Opt_enospc_debug:
			btrfs_set_opt(info->mount_opt, ENOSPC_DEBUG);
			break;
666 667 668
		case Opt_noenospc_debug:
			btrfs_clear_opt(info->mount_opt, ENOSPC_DEBUG);
			break;
C
Chris Mason 已提交
669
		case Opt_defrag:
670 671
			btrfs_set_and_info(root, AUTO_DEFRAG,
					   "enabling auto defrag");
C
Chris Mason 已提交
672
			break;
673
		case Opt_nodefrag:
674 675
			btrfs_clear_and_info(root, AUTO_DEFRAG,
					     "disabling auto defrag");
676
			break;
C
Chris Mason 已提交
677
		case Opt_recovery:
678
			btrfs_info(root->fs_info, "enabling auto recovery");
C
Chris Mason 已提交
679 680
			btrfs_set_opt(info->mount_opt, RECOVERY);
			break;
681 682 683
		case Opt_skip_balance:
			btrfs_set_opt(info->mount_opt, SKIP_BALANCE);
			break;
684 685
#ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY
		case Opt_check_integrity_including_extent_data:
686 687
			btrfs_info(root->fs_info,
				   "enabling check integrity including extent data");
688 689 690 691 692
			btrfs_set_opt(info->mount_opt,
				      CHECK_INTEGRITY_INCLUDING_EXTENT_DATA);
			btrfs_set_opt(info->mount_opt, CHECK_INTEGRITY);
			break;
		case Opt_check_integrity:
693
			btrfs_info(root->fs_info, "enabling check integrity");
694 695 696
			btrfs_set_opt(info->mount_opt, CHECK_INTEGRITY);
			break;
		case Opt_check_integrity_print_mask:
697 698 699 700
			ret = match_int(&args[0], &intarg);
			if (ret) {
				goto out;
			} else if (intarg >= 0) {
701
				info->check_integrity_print_mask = intarg;
702
				btrfs_info(root->fs_info, "check_integrity_print_mask 0x%x",
703
				       info->check_integrity_print_mask);
704 705 706
			} else {
				ret = -EINVAL;
				goto out;
707 708 709 710 711 712
			}
			break;
#else
		case Opt_check_integrity_including_extent_data:
		case Opt_check_integrity:
		case Opt_check_integrity_print_mask:
713 714
			btrfs_err(root->fs_info,
				"support for check_integrity* not compiled in!");
715 716 717
			ret = -EINVAL;
			goto out;
#endif
J
Jeff Mahoney 已提交
718 719 720 721 722 723 724 725 726 727 728 729
		case Opt_fatal_errors:
			if (strcmp(args[0].from, "panic") == 0)
				btrfs_set_opt(info->mount_opt,
					      PANIC_ON_FATAL_ERROR);
			else if (strcmp(args[0].from, "bug") == 0)
				btrfs_clear_opt(info->mount_opt,
					      PANIC_ON_FATAL_ERROR);
			else {
				ret = -EINVAL;
				goto out;
			}
			break;
730 731 732 733
		case Opt_commit_interval:
			intarg = 0;
			ret = match_int(&args[0], &intarg);
			if (ret < 0) {
734
				btrfs_err(root->fs_info, "invalid commit interval");
735 736 737 738 739
				ret = -EINVAL;
				goto out;
			}
			if (intarg > 0) {
				if (intarg > 300) {
740
					btrfs_warn(root->fs_info, "excessive commit interval %d",
741 742 743 744
							intarg);
				}
				info->commit_interval = intarg;
			} else {
745
				btrfs_info(root->fs_info, "using default commit interval %ds",
746 747 748 749
				    BTRFS_DEFAULT_COMMIT_INTERVAL);
				info->commit_interval = BTRFS_DEFAULT_COMMIT_INTERVAL;
			}
			break;
S
Sage Weil 已提交
750
		case Opt_err:
751
			btrfs_info(root->fs_info, "unrecognized mount option '%s'", p);
S
Sage Weil 已提交
752 753
			ret = -EINVAL;
			goto out;
754
		default:
755
			break;
756 757
		}
	}
S
Sage Weil 已提交
758
out:
759
	if (!ret && btrfs_test_opt(root, SPACE_CACHE))
760
		btrfs_info(root->fs_info, "disk space caching is enabled");
761
	kfree(orig);
S
Sage Weil 已提交
762
	return ret;
763 764 765 766 767 768 769 770
}

/*
 * Parse mount options that are required early in the mount process.
 *
 * All other options will be parsed on much later in the mount process and
 * only when we need to allocate a new super block.
 */
771
static int btrfs_parse_early_options(const char *options, fmode_t flags,
772
		void *holder, char **subvol_name, u64 *subvol_objectid,
773
		struct btrfs_fs_devices **fs_devices)
774 775
{
	substring_t args[MAX_OPT_ARGS];
776
	char *device_name, *opts, *orig, *p;
777
	char *num = NULL;
778 779 780
	int error = 0;

	if (!options)
781
		return 0;
782 783 784 785 786 787 788 789

	/*
	 * strsep changes the string, duplicate it because parse_options
	 * gets called twice
	 */
	opts = kstrdup(options, GFP_KERNEL);
	if (!opts)
		return -ENOMEM;
790
	orig = opts;
791 792 793 794 795 796 797 798 799

	while ((p = strsep(&opts, ",")) != NULL) {
		int token;
		if (!*p)
			continue;

		token = match_token(p, tokens, args);
		switch (token) {
		case Opt_subvol:
800
			kfree(*subvol_name);
801
			*subvol_name = match_strdup(&args[0]);
802 803 804 805
			if (!*subvol_name) {
				error = -ENOMEM;
				goto out;
			}
806
			break;
807
		case Opt_subvolid:
808 809 810 811
			num = match_strdup(&args[0]);
			if (num) {
				*subvol_objectid = memparse(num, NULL);
				kfree(num);
812
				/* we want the original fs_tree */
813
				if (!*subvol_objectid)
814 815
					*subvol_objectid =
						BTRFS_FS_TREE_OBJECTID;
816 817 818
			} else {
				error = -EINVAL;
				goto out;
819
			}
820
			break;
821
		case Opt_subvolrootid:
822
			printk(KERN_WARNING
823 824
				"BTRFS: 'subvolrootid' mount option is deprecated and has "
				"no effect\n");
825
			break;
826
		case Opt_device:
827 828 829 830 831 832
			device_name = match_strdup(&args[0]);
			if (!device_name) {
				error = -ENOMEM;
				goto out;
			}
			error = btrfs_scan_one_device(device_name,
833
					flags, holder, fs_devices);
834
			kfree(device_name);
835
			if (error)
836
				goto out;
837
			break;
838 839 840 841 842
		default:
			break;
		}
	}

843
out:
844
	kfree(orig);
845
	return error;
846 847
}

848 849 850
static struct dentry *get_default_root(struct super_block *sb,
				       u64 subvol_objectid)
{
851 852
	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
	struct btrfs_root *root = fs_info->tree_root;
853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881
	struct btrfs_root *new_root;
	struct btrfs_dir_item *di;
	struct btrfs_path *path;
	struct btrfs_key location;
	struct inode *inode;
	u64 dir_id;
	int new = 0;

	/*
	 * We have a specific subvol we want to mount, just setup location and
	 * go look up the root.
	 */
	if (subvol_objectid) {
		location.objectid = subvol_objectid;
		location.type = BTRFS_ROOT_ITEM_KEY;
		location.offset = (u64)-1;
		goto find_root;
	}

	path = btrfs_alloc_path();
	if (!path)
		return ERR_PTR(-ENOMEM);
	path->leave_spinning = 1;

	/*
	 * Find the "default" dir item which points to the root item that we
	 * will mount by default if we haven't been given a specific subvolume
	 * to mount.
	 */
882
	dir_id = btrfs_super_root_dir(fs_info->super_copy);
883
	di = btrfs_lookup_dir_item(NULL, root, path, dir_id, "default", 7, 0);
884 885
	if (IS_ERR(di)) {
		btrfs_free_path(path);
886
		return ERR_CAST(di);
887
	}
888 889 890 891 892 893 894 895
	if (!di) {
		/*
		 * Ok the default dir item isn't there.  This is weird since
		 * it's always been there, but don't freak out, just try and
		 * mount to root most subvolume.
		 */
		btrfs_free_path(path);
		dir_id = BTRFS_FIRST_FREE_OBJECTID;
896
		new_root = fs_info->fs_root;
897 898 899 900 901 902 903
		goto setup_root;
	}

	btrfs_dir_item_key_to_cpu(path->nodes[0], di, &location);
	btrfs_free_path(path);

find_root:
904
	new_root = btrfs_read_fs_root_no_name(fs_info, &location);
905
	if (IS_ERR(new_root))
J
Julia Lawall 已提交
906
		return ERR_CAST(new_root);
907 908 909 910 911 912 913 914

	dir_id = btrfs_root_dirid(&new_root->root_item);
setup_root:
	location.objectid = dir_id;
	location.type = BTRFS_INODE_ITEM_KEY;
	location.offset = 0;

	inode = btrfs_iget(sb, &location, new_root, &new);
915 916
	if (IS_ERR(inode))
		return ERR_CAST(inode);
917 918 919 920 921 922 923 924 925 926 927

	/*
	 * If we're just mounting the root most subvol put the inode and return
	 * a reference to the dentry.  We will have already gotten a reference
	 * to the inode in btrfs_fill_super so we're good to go.
	 */
	if (!new && sb->s_root->d_inode == inode) {
		iput(inode);
		return dget(sb->s_root);
	}

928
	return d_obtain_alias(inode);
929 930
}

C
Chris Mason 已提交
931
static int btrfs_fill_super(struct super_block *sb,
932
			    struct btrfs_fs_devices *fs_devices,
C
Chris Mason 已提交
933
			    void *data, int silent)
C
Chris Mason 已提交
934
{
C
Chris Mason 已提交
935
	struct inode *inode;
936
	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
937
	struct btrfs_key key;
C
Chris Mason 已提交
938
	int err;
939

C
Chris Mason 已提交
940 941 942
	sb->s_maxbytes = MAX_LFS_FILESIZE;
	sb->s_magic = BTRFS_SUPER_MAGIC;
	sb->s_op = &btrfs_super_ops;
A
Al Viro 已提交
943
	sb->s_d_op = &btrfs_dentry_operations;
B
Balaji Rao 已提交
944
	sb->s_export_op = &btrfs_export_ops;
J
Josef Bacik 已提交
945
	sb->s_xattr = btrfs_xattr_handlers;
C
Chris Mason 已提交
946
	sb->s_time_gran = 1;
C
Chris Mason 已提交
947
#ifdef CONFIG_BTRFS_FS_POSIX_ACL
J
Josef Bacik 已提交
948
	sb->s_flags |= MS_POSIXACL;
949
#endif
950
	sb->s_flags |= MS_I_VERSION;
A
Al Viro 已提交
951 952
	err = open_ctree(sb, fs_devices, (char *)data);
	if (err) {
953
		printk(KERN_ERR "BTRFS: open_ctree failed\n");
A
Al Viro 已提交
954
		return err;
955 956
	}

957 958 959
	key.objectid = BTRFS_FIRST_FREE_OBJECTID;
	key.type = BTRFS_INODE_ITEM_KEY;
	key.offset = 0;
960
	inode = btrfs_iget(sb, &key, fs_info->fs_root, NULL);
961 962
	if (IS_ERR(inode)) {
		err = PTR_ERR(inode);
C
Chris Mason 已提交
963
		goto fail_close;
C
Chris Mason 已提交
964 965
	}

966 967
	sb->s_root = d_make_root(inode);
	if (!sb->s_root) {
C
Chris Mason 已提交
968 969
		err = -ENOMEM;
		goto fail_close;
C
Chris Mason 已提交
970
	}
971

C
Chris Mason 已提交
972
	save_mount_options(sb, data);
D
Dan Magenheimer 已提交
973
	cleancache_init_fs(sb);
974
	sb->s_flags |= MS_ACTIVE;
C
Chris Mason 已提交
975
	return 0;
C
Chris Mason 已提交
976 977

fail_close:
978
	close_ctree(fs_info->tree_root);
C
Chris Mason 已提交
979
	return err;
C
Chris Mason 已提交
980 981
}

S
Sage Weil 已提交
982
int btrfs_sync_fs(struct super_block *sb, int wait)
C
Chris Mason 已提交
983 984
{
	struct btrfs_trans_handle *trans;
985 986
	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
	struct btrfs_root *root = fs_info->tree_root;
C
Chris Mason 已提交
987

988 989
	trace_btrfs_sync_fs(wait);

C
Chris Mason 已提交
990
	if (!wait) {
991
		filemap_flush(fs_info->btree_inode->i_mapping);
C
Chris Mason 已提交
992 993
		return 0;
	}
994

995
	btrfs_wait_ordered_roots(fs_info, -1);
996

M
Miao Xie 已提交
997
	trans = btrfs_attach_transaction_barrier(root);
998
	if (IS_ERR(trans)) {
999 1000
		/* no transaction, don't bother */
		if (PTR_ERR(trans) == -ENOENT)
1001
			return 0;
1002
		return PTR_ERR(trans);
1003
	}
1004
	return btrfs_commit_transaction(trans, root);
C
Chris Mason 已提交
1005 1006
}

1007
static int btrfs_show_options(struct seq_file *seq, struct dentry *dentry)
E
Eric Paris 已提交
1008
{
1009 1010
	struct btrfs_fs_info *info = btrfs_sb(dentry->d_sb);
	struct btrfs_root *root = info->tree_root;
T
Tsutomu Itoh 已提交
1011
	char *compress_type;
E
Eric Paris 已提交
1012 1013 1014 1015 1016 1017 1018 1019 1020 1021

	if (btrfs_test_opt(root, DEGRADED))
		seq_puts(seq, ",degraded");
	if (btrfs_test_opt(root, NODATASUM))
		seq_puts(seq, ",nodatasum");
	if (btrfs_test_opt(root, NODATACOW))
		seq_puts(seq, ",nodatacow");
	if (btrfs_test_opt(root, NOBARRIER))
		seq_puts(seq, ",nobarrier");
	if (info->max_inline != 8192 * 1024)
1022
		seq_printf(seq, ",max_inline=%llu", info->max_inline);
E
Eric Paris 已提交
1023
	if (info->alloc_start != 0)
1024
		seq_printf(seq, ",alloc_start=%llu", info->alloc_start);
E
Eric Paris 已提交
1025 1026 1027
	if (info->thread_pool_size !=  min_t(unsigned long,
					     num_online_cpus() + 2, 8))
		seq_printf(seq, ",thread_pool=%d", info->thread_pool_size);
T
Tsutomu Itoh 已提交
1028 1029 1030 1031 1032 1033 1034 1035 1036 1037
	if (btrfs_test_opt(root, COMPRESS)) {
		if (info->compress_type == BTRFS_COMPRESS_ZLIB)
			compress_type = "zlib";
		else
			compress_type = "lzo";
		if (btrfs_test_opt(root, FORCE_COMPRESS))
			seq_printf(seq, ",compress-force=%s", compress_type);
		else
			seq_printf(seq, ",compress=%s", compress_type);
	}
C
Chris Mason 已提交
1038 1039
	if (btrfs_test_opt(root, NOSSD))
		seq_puts(seq, ",nossd");
1040 1041 1042
	if (btrfs_test_opt(root, SSD_SPREAD))
		seq_puts(seq, ",ssd_spread");
	else if (btrfs_test_opt(root, SSD))
E
Eric Paris 已提交
1043
		seq_puts(seq, ",ssd");
S
Sage Weil 已提交
1044
	if (btrfs_test_opt(root, NOTREELOG))
1045
		seq_puts(seq, ",notreelog");
1046
	if (btrfs_test_opt(root, FLUSHONCOMMIT))
1047
		seq_puts(seq, ",flushoncommit");
1048 1049
	if (btrfs_test_opt(root, DISCARD))
		seq_puts(seq, ",discard");
E
Eric Paris 已提交
1050 1051
	if (!(root->fs_info->sb->s_flags & MS_POSIXACL))
		seq_puts(seq, ",noacl");
T
Tsutomu Itoh 已提交
1052 1053
	if (btrfs_test_opt(root, SPACE_CACHE))
		seq_puts(seq, ",space_cache");
1054
	else
1055
		seq_puts(seq, ",nospace_cache");
1056 1057
	if (btrfs_test_opt(root, RESCAN_UUID_TREE))
		seq_puts(seq, ",rescan_uuid_tree");
T
Tsutomu Itoh 已提交
1058 1059 1060 1061
	if (btrfs_test_opt(root, CLEAR_CACHE))
		seq_puts(seq, ",clear_cache");
	if (btrfs_test_opt(root, USER_SUBVOL_RM_ALLOWED))
		seq_puts(seq, ",user_subvol_rm_allowed");
1062 1063 1064 1065 1066 1067
	if (btrfs_test_opt(root, ENOSPC_DEBUG))
		seq_puts(seq, ",enospc_debug");
	if (btrfs_test_opt(root, AUTO_DEFRAG))
		seq_puts(seq, ",autodefrag");
	if (btrfs_test_opt(root, INODE_MAP_CACHE))
		seq_puts(seq, ",inode_cache");
1068 1069
	if (btrfs_test_opt(root, SKIP_BALANCE))
		seq_puts(seq, ",skip_balance");
1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083
	if (btrfs_test_opt(root, RECOVERY))
		seq_puts(seq, ",recovery");
#ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY
	if (btrfs_test_opt(root, CHECK_INTEGRITY_INCLUDING_EXTENT_DATA))
		seq_puts(seq, ",check_int_data");
	else if (btrfs_test_opt(root, CHECK_INTEGRITY))
		seq_puts(seq, ",check_int");
	if (info->check_integrity_print_mask)
		seq_printf(seq, ",check_int_print_mask=%d",
				info->check_integrity_print_mask);
#endif
	if (info->metadata_ratio)
		seq_printf(seq, ",metadata_ratio=%d",
				info->metadata_ratio);
J
Jeff Mahoney 已提交
1084 1085
	if (btrfs_test_opt(root, PANIC_ON_FATAL_ERROR))
		seq_puts(seq, ",fatal_errors=panic");
1086 1087
	if (info->commit_interval != BTRFS_DEFAULT_COMMIT_INTERVAL)
		seq_printf(seq, ",commit=%d", info->commit_interval);
E
Eric Paris 已提交
1088 1089 1090
	return 0;
}

1091
static int btrfs_test_super(struct super_block *s, void *data)
Y
Yan 已提交
1092
{
1093 1094
	struct btrfs_fs_info *p = data;
	struct btrfs_fs_info *fs_info = btrfs_sb(s);
Y
Yan 已提交
1095

1096
	return fs_info->fs_devices == p->fs_devices;
Y
Yan 已提交
1097 1098
}

1099 1100
static int btrfs_set_super(struct super_block *s, void *data)
{
A
Al Viro 已提交
1101 1102 1103 1104
	int err = set_anon_super(s, data);
	if (!err)
		s->s_fs_info = data;
	return err;
Y
Yan 已提交
1105 1106
}

1107 1108 1109 1110 1111 1112 1113 1114 1115 1116
/*
 * subvolumes are identified by ino 256
 */
static inline int is_subvolume_inode(struct inode *inode)
{
	if (inode && inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)
		return 1;
	return 0;
}

1117 1118 1119 1120 1121 1122 1123
/*
 * This will strip out the subvol=%s argument for an argument string and add
 * subvolid=0 to make sure we get the actual tree root for path walking to the
 * subvol we want.
 */
static char *setup_root_args(char *args)
{
1124 1125
	unsigned len = strlen(args) + 2 + 1;
	char *src, *dst, *buf;
1126 1127

	/*
1128 1129
	 * We need the same args as before, but with this substitution:
	 * s!subvol=[^,]+!subvolid=0!
1130
	 *
1131 1132
	 * Since the replacement string is up to 2 bytes longer than the
	 * original, allocate strlen(args) + 2 + 1 bytes.
1133 1134
	 */

1135
	src = strstr(args, "subvol=");
1136
	/* This shouldn't happen, but just in case.. */
1137 1138 1139 1140 1141
	if (!src)
		return NULL;

	buf = dst = kmalloc(len, GFP_NOFS);
	if (!buf)
1142 1143 1144
		return NULL;

	/*
1145 1146
	 * If the subvol= arg is not at the start of the string,
	 * copy whatever precedes it into buf.
1147
	 */
1148 1149 1150 1151
	if (src != args) {
		*src++ = '\0';
		strcpy(buf, args);
		dst += strlen(args);
1152 1153
	}

1154 1155
	strcpy(dst, "subvolid=0");
	dst += strlen("subvolid=0");
1156 1157

	/*
1158 1159
	 * If there is a "," after the original subvol=... string,
	 * copy that suffix into our buffer.  Otherwise, we're done.
1160
	 */
1161 1162 1163
	src = strchr(src, ',');
	if (src)
		strcpy(dst, src);
1164

1165
	return buf;
1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183
}

static struct dentry *mount_subvol(const char *subvol_name, int flags,
				   const char *device_name, char *data)
{
	struct dentry *root;
	struct vfsmount *mnt;
	char *newargs;

	newargs = setup_root_args(data);
	if (!newargs)
		return ERR_PTR(-ENOMEM);
	mnt = vfs_kern_mount(&btrfs_fs_type, flags, device_name,
			     newargs);
	kfree(newargs);
	if (IS_ERR(mnt))
		return ERR_CAST(mnt);

A
Al Viro 已提交
1184
	root = mount_subtree(mnt, subvol_name);
1185

A
Al Viro 已提交
1186 1187 1188 1189 1190
	if (!IS_ERR(root) && !is_subvolume_inode(root->d_inode)) {
		struct super_block *s = root->d_sb;
		dput(root);
		root = ERR_PTR(-EINVAL);
		deactivate_locked_super(s);
1191
		printk(KERN_ERR "BTRFS: '%s' is not a valid subvolume\n",
1192 1193 1194
				subvol_name);
	}

1195 1196
	return root;
}
1197

1198 1199 1200 1201 1202 1203
/*
 * Find a superblock for the given device / mount point.
 *
 * Note:  This is based on get_sb_bdev from fs/super.c with a few additions
 *	  for multiple device setup.  Make sure to keep it in sync.
 */
A
Al Viro 已提交
1204
static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags,
1205
		const char *device_name, void *data)
Y
Yan 已提交
1206 1207 1208 1209
{
	struct block_device *bdev = NULL;
	struct super_block *s;
	struct dentry *root;
1210
	struct btrfs_fs_devices *fs_devices = NULL;
1211
	struct btrfs_fs_info *fs_info = NULL;
1212
	fmode_t mode = FMODE_READ;
1213 1214
	char *subvol_name = NULL;
	u64 subvol_objectid = 0;
Y
Yan 已提交
1215 1216
	int error = 0;

1217 1218 1219 1220
	if (!(flags & MS_RDONLY))
		mode |= FMODE_WRITE;

	error = btrfs_parse_early_options(data, mode, fs_type,
1221
					  &subvol_name, &subvol_objectid,
1222
					  &fs_devices);
1223 1224
	if (error) {
		kfree(subvol_name);
A
Al Viro 已提交
1225
		return ERR_PTR(error);
1226
	}
1227

1228 1229 1230 1231 1232 1233
	if (subvol_name) {
		root = mount_subvol(subvol_name, flags, device_name, data);
		kfree(subvol_name);
		return root;
	}

1234
	error = btrfs_scan_one_device(device_name, mode, fs_type, &fs_devices);
1235
	if (error)
1236
		return ERR_PTR(error);
Y
Yan 已提交
1237

1238 1239 1240 1241 1242 1243 1244
	/*
	 * Setup a dummy root and fs_info for test/set super.  This is because
	 * we don't actually fill this stuff out until open_ctree, but we need
	 * it for searching for existing supers, so this lets us do that and
	 * then open_ctree will properly initialize everything later.
	 */
	fs_info = kzalloc(sizeof(struct btrfs_fs_info), GFP_NOFS);
1245 1246 1247
	if (!fs_info)
		return ERR_PTR(-ENOMEM);

1248 1249
	fs_info->fs_devices = fs_devices;

1250 1251 1252 1253
	fs_info->super_copy = kzalloc(BTRFS_SUPER_INFO_SIZE, GFP_NOFS);
	fs_info->super_for_commit = kzalloc(BTRFS_SUPER_INFO_SIZE, GFP_NOFS);
	if (!fs_info->super_copy || !fs_info->super_for_commit) {
		error = -ENOMEM;
1254 1255 1256 1257 1258 1259 1260 1261 1262
		goto error_fs_info;
	}

	error = btrfs_open_devices(fs_devices, mode, fs_type);
	if (error)
		goto error_fs_info;

	if (!(flags & MS_RDONLY) && fs_devices->rw_devices == 0) {
		error = -EACCES;
1263 1264 1265
		goto error_close_devices;
	}

1266
	bdev = fs_devices->latest_bdev;
D
David Howells 已提交
1267 1268
	s = sget(fs_type, btrfs_test_super, btrfs_set_super, flags | MS_NOSEC,
		 fs_info);
1269 1270 1271 1272
	if (IS_ERR(s)) {
		error = PTR_ERR(s);
		goto error_close_devices;
	}
Y
Yan 已提交
1273 1274

	if (s->s_root) {
Y
Yan Zheng 已提交
1275
		btrfs_close_devices(fs_devices);
1276
		free_fs_info(fs_info);
1277 1278
		if ((flags ^ s->s_flags) & MS_RDONLY)
			error = -EBUSY;
Y
Yan 已提交
1279 1280 1281 1282
	} else {
		char b[BDEVNAME_SIZE];

		strlcpy(s->s_id, bdevname(bdev, b), sizeof(s->s_id));
1283
		btrfs_sb(s)->bdev_holder = fs_type;
1284 1285
		error = btrfs_fill_super(s, fs_devices, data,
					 flags & MS_SILENT ? 1 : 0);
Y
Yan 已提交
1286 1287
	}

1288 1289
	root = !error ? get_default_root(s, subvol_objectid) : ERR_PTR(error);
	if (IS_ERR(root))
1290
		deactivate_locked_super(s);
Y
Yan 已提交
1291

A
Al Viro 已提交
1292
	return root;
Y
Yan 已提交
1293

Y
Yan Zheng 已提交
1294
error_close_devices:
1295
	btrfs_close_devices(fs_devices);
1296
error_fs_info:
1297
	free_fs_info(fs_info);
A
Al Viro 已提交
1298
	return ERR_PTR(error);
Y
Yan 已提交
1299
}
1300

1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315
static void btrfs_set_max_workers(struct btrfs_workers *workers, int new_limit)
{
	spin_lock_irq(&workers->lock);
	workers->max_workers = new_limit;
	spin_unlock_irq(&workers->lock);
}

static void btrfs_resize_thread_pool(struct btrfs_fs_info *fs_info,
				     int new_pool_size, int old_pool_size)
{
	if (new_pool_size == old_pool_size)
		return;

	fs_info->thread_pool_size = new_pool_size;

1316
	btrfs_info(fs_info, "resize thread pool %d -> %d",
1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331
	       old_pool_size, new_pool_size);

	btrfs_set_max_workers(&fs_info->generic_worker, new_pool_size);
	btrfs_set_max_workers(&fs_info->workers, new_pool_size);
	btrfs_set_max_workers(&fs_info->delalloc_workers, new_pool_size);
	btrfs_set_max_workers(&fs_info->submit_workers, new_pool_size);
	btrfs_set_max_workers(&fs_info->caching_workers, new_pool_size);
	btrfs_set_max_workers(&fs_info->fixup_workers, new_pool_size);
	btrfs_set_max_workers(&fs_info->endio_workers, new_pool_size);
	btrfs_set_max_workers(&fs_info->endio_meta_workers, new_pool_size);
	btrfs_set_max_workers(&fs_info->endio_meta_write_workers, new_pool_size);
	btrfs_set_max_workers(&fs_info->endio_write_workers, new_pool_size);
	btrfs_set_max_workers(&fs_info->endio_freespace_worker, new_pool_size);
	btrfs_set_max_workers(&fs_info->delayed_workers, new_pool_size);
	btrfs_set_max_workers(&fs_info->readahead_workers, new_pool_size);
1332 1333
	btrfs_set_max_workers(&fs_info->scrub_wr_completion_workers,
			      new_pool_size);
1334 1335
}

1336
static inline void btrfs_remount_prepare(struct btrfs_fs_info *fs_info)
M
Miao Xie 已提交
1337 1338
{
	set_bit(BTRFS_FS_STATE_REMOUNTING, &fs_info->fs_state);
1339
}
M
Miao Xie 已提交
1340

1341 1342 1343
static inline void btrfs_remount_begin(struct btrfs_fs_info *fs_info,
				       unsigned long old_opts, int flags)
{
M
Miao Xie 已提交
1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370
	if (btrfs_raw_test_opt(old_opts, AUTO_DEFRAG) &&
	    (!btrfs_raw_test_opt(fs_info->mount_opt, AUTO_DEFRAG) ||
	     (flags & MS_RDONLY))) {
		/* wait for any defraggers to finish */
		wait_event(fs_info->transaction_wait,
			   (atomic_read(&fs_info->defrag_running) == 0));
		if (flags & MS_RDONLY)
			sync_filesystem(fs_info->sb);
	}
}

static inline void btrfs_remount_cleanup(struct btrfs_fs_info *fs_info,
					 unsigned long old_opts)
{
	/*
	 * We need cleanup all defragable inodes if the autodefragment is
	 * close or the fs is R/O.
	 */
	if (btrfs_raw_test_opt(old_opts, AUTO_DEFRAG) &&
	    (!btrfs_raw_test_opt(fs_info->mount_opt, AUTO_DEFRAG) ||
	     (fs_info->sb->s_flags & MS_RDONLY))) {
		btrfs_cleanup_defrag_inodes(fs_info);
	}

	clear_bit(BTRFS_FS_STATE_REMOUNTING, &fs_info->fs_state);
}

Y
Yan Zheng 已提交
1371 1372
static int btrfs_remount(struct super_block *sb, int *flags, char *data)
{
1373 1374
	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
	struct btrfs_root *root = fs_info->tree_root;
1375 1376 1377 1378 1379 1380 1381
	unsigned old_flags = sb->s_flags;
	unsigned long old_opts = fs_info->mount_opt;
	unsigned long old_compress_type = fs_info->compress_type;
	u64 old_max_inline = fs_info->max_inline;
	u64 old_alloc_start = fs_info->alloc_start;
	int old_thread_pool_size = fs_info->thread_pool_size;
	unsigned int old_metadata_ratio = fs_info->metadata_ratio;
Y
Yan Zheng 已提交
1382 1383
	int ret;

1384
	btrfs_remount_prepare(fs_info);
M
Miao Xie 已提交
1385

1386
	ret = btrfs_parse_options(root, data);
1387 1388 1389 1390
	if (ret) {
		ret = -EINVAL;
		goto restore;
	}
1391

1392
	btrfs_remount_begin(fs_info, old_opts, *flags);
1393 1394 1395
	btrfs_resize_thread_pool(fs_info,
		fs_info->thread_pool_size, old_thread_pool_size);

Y
Yan Zheng 已提交
1396
	if ((*flags & MS_RDONLY) == (sb->s_flags & MS_RDONLY))
M
Miao Xie 已提交
1397
		goto out;
Y
Yan Zheng 已提交
1398 1399

	if (*flags & MS_RDONLY) {
1400 1401 1402 1403
		/*
		 * this also happens on 'umount -rf' or on shutdown, when
		 * the filesystem is busy.
		 */
1404 1405 1406 1407 1408 1409

		/* wait for the uuid_scan task to finish */
		down(&fs_info->uuid_tree_rescan_sem);
		/* avoid complains from lockdep et al. */
		up(&fs_info->uuid_tree_rescan_sem);

Y
Yan Zheng 已提交
1410 1411
		sb->s_flags |= MS_RDONLY;

1412 1413
		btrfs_dev_replace_suspend_for_unmount(fs_info);
		btrfs_scrub_cancel(fs_info);
1414
		btrfs_pause_balance(fs_info);
1415

1416 1417 1418
		ret = btrfs_commit_super(root);
		if (ret)
			goto restore;
Y
Yan Zheng 已提交
1419
	} else {
1420 1421
		if (test_bit(BTRFS_FS_STATE_ERROR, &root->fs_info->fs_state)) {
			btrfs_err(fs_info,
1422
				"Remounting read-write after error is not allowed");
1423 1424 1425
			ret = -EINVAL;
			goto restore;
		}
1426
		if (fs_info->fs_devices->rw_devices == 0) {
1427 1428
			ret = -EACCES;
			goto restore;
1429
		}
Y
Yan Zheng 已提交
1430

1431 1432 1433
		if (fs_info->fs_devices->missing_devices >
		     fs_info->num_tolerated_disk_barrier_failures &&
		    !(*flags & MS_RDONLY)) {
1434 1435
			btrfs_warn(fs_info,
				"too many missing devices, writeable remount is not allowed");
1436 1437 1438 1439
			ret = -EACCES;
			goto restore;
		}

1440
		if (btrfs_super_log_root(fs_info->super_copy) != 0) {
1441 1442
			ret = -EINVAL;
			goto restore;
1443
		}
Y
Yan Zheng 已提交
1444

1445
		ret = btrfs_cleanup_fs_roots(fs_info);
1446 1447
		if (ret)
			goto restore;
Y
Yan Zheng 已提交
1448

1449 1450
		/* recover relocation */
		ret = btrfs_recover_relocation(root);
1451 1452
		if (ret)
			goto restore;
Y
Yan Zheng 已提交
1453

1454 1455 1456 1457
		ret = btrfs_resume_balance_async(fs_info);
		if (ret)
			goto restore;

1458 1459
		ret = btrfs_resume_dev_replace_async(fs_info);
		if (ret) {
1460
			btrfs_warn(fs_info, "failed to resume dev_replace");
1461 1462
			goto restore;
		}
1463 1464

		if (!fs_info->uuid_root) {
1465
			btrfs_info(fs_info, "creating UUID tree");
1466 1467
			ret = btrfs_create_uuid_tree(fs_info);
			if (ret) {
1468
				btrfs_warn(fs_info, "failed to create the UUID tree %d", ret);
1469 1470 1471
				goto restore;
			}
		}
Y
Yan Zheng 已提交
1472 1473
		sb->s_flags &= ~MS_RDONLY;
	}
M
Miao Xie 已提交
1474 1475
out:
	btrfs_remount_cleanup(fs_info, old_opts);
Y
Yan Zheng 已提交
1476
	return 0;
1477 1478 1479 1480 1481 1482 1483 1484 1485

restore:
	/* We've hit an error - don't reset MS_RDONLY */
	if (sb->s_flags & MS_RDONLY)
		old_flags |= MS_RDONLY;
	sb->s_flags = old_flags;
	fs_info->mount_opt = old_opts;
	fs_info->compress_type = old_compress_type;
	fs_info->max_inline = old_max_inline;
M
Miao Xie 已提交
1486
	mutex_lock(&fs_info->chunk_mutex);
1487
	fs_info->alloc_start = old_alloc_start;
M
Miao Xie 已提交
1488
	mutex_unlock(&fs_info->chunk_mutex);
1489 1490
	btrfs_resize_thread_pool(fs_info,
		old_thread_pool_size, fs_info->thread_pool_size);
1491
	fs_info->metadata_ratio = old_metadata_ratio;
M
Miao Xie 已提交
1492
	btrfs_remount_cleanup(fs_info, old_opts);
1493
	return ret;
Y
Yan Zheng 已提交
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
/* Used to sort the devices by max_avail(descending sort) */
static int btrfs_cmp_device_free_bytes(const void *dev_info1,
				       const void *dev_info2)
{
	if (((struct btrfs_device_info *)dev_info1)->max_avail >
	    ((struct btrfs_device_info *)dev_info2)->max_avail)
		return -1;
	else if (((struct btrfs_device_info *)dev_info1)->max_avail <
		 ((struct btrfs_device_info *)dev_info2)->max_avail)
		return 1;
	else
	return 0;
}

/*
 * sort the devices by max_avail, in which max free extent size of each device
 * is stored.(Descending Sort)
 */
static inline void btrfs_descending_sort_devices(
					struct btrfs_device_info *devices,
					size_t nr_devices)
{
	sort(devices, nr_devices, sizeof(struct btrfs_device_info),
	     btrfs_cmp_device_free_bytes, NULL);
}

1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536
/*
 * The helper to calc the free space on the devices that can be used to store
 * file data.
 */
static int btrfs_calc_avail_data_space(struct btrfs_root *root, u64 *free_bytes)
{
	struct btrfs_fs_info *fs_info = root->fs_info;
	struct btrfs_device_info *devices_info;
	struct btrfs_fs_devices *fs_devices = fs_info->fs_devices;
	struct btrfs_device *device;
	u64 skip_space;
	u64 type;
	u64 avail_space;
	u64 used_space;
	u64 min_stripe_size;
1537
	int min_stripes = 1, num_stripes = 1;
1538 1539 1540
	int i = 0, nr_devices;
	int ret;

1541
	nr_devices = fs_info->fs_devices->open_devices;
1542 1543
	BUG_ON(!nr_devices);

1544
	devices_info = kmalloc_array(nr_devices, sizeof(*devices_info),
1545 1546 1547 1548 1549 1550
			       GFP_NOFS);
	if (!devices_info)
		return -ENOMEM;

	/* calc min stripe number for data space alloction */
	type = btrfs_get_alloc_profile(root, 1);
1551
	if (type & BTRFS_BLOCK_GROUP_RAID0) {
1552
		min_stripes = 2;
1553 1554
		num_stripes = nr_devices;
	} else if (type & BTRFS_BLOCK_GROUP_RAID1) {
1555
		min_stripes = 2;
1556 1557
		num_stripes = 2;
	} else if (type & BTRFS_BLOCK_GROUP_RAID10) {
1558
		min_stripes = 4;
1559 1560
		num_stripes = 4;
	}
1561 1562 1563 1564 1565 1566

	if (type & BTRFS_BLOCK_GROUP_DUP)
		min_stripe_size = 2 * BTRFS_STRIPE_LEN;
	else
		min_stripe_size = BTRFS_STRIPE_LEN;

1567
	list_for_each_entry(device, &fs_devices->devices, dev_list) {
1568 1569
		if (!device->in_fs_metadata || !device->bdev ||
		    device->is_tgtdev_for_dev_replace)
1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629
			continue;

		avail_space = device->total_bytes - device->bytes_used;

		/* align with stripe_len */
		do_div(avail_space, BTRFS_STRIPE_LEN);
		avail_space *= BTRFS_STRIPE_LEN;

		/*
		 * In order to avoid overwritting the superblock on the drive,
		 * btrfs starts at an offset of at least 1MB when doing chunk
		 * allocation.
		 */
		skip_space = 1024 * 1024;

		/* user can set the offset in fs_info->alloc_start. */
		if (fs_info->alloc_start + BTRFS_STRIPE_LEN <=
		    device->total_bytes)
			skip_space = max(fs_info->alloc_start, skip_space);

		/*
		 * btrfs can not use the free space in [0, skip_space - 1],
		 * we must subtract it from the total. In order to implement
		 * it, we account the used space in this range first.
		 */
		ret = btrfs_account_dev_extents_size(device, 0, skip_space - 1,
						     &used_space);
		if (ret) {
			kfree(devices_info);
			return ret;
		}

		/* calc the free space in [0, skip_space - 1] */
		skip_space -= used_space;

		/*
		 * we can use the free space in [0, skip_space - 1], subtract
		 * it from the total.
		 */
		if (avail_space && avail_space >= skip_space)
			avail_space -= skip_space;
		else
			avail_space = 0;

		if (avail_space < min_stripe_size)
			continue;

		devices_info[i].dev = device;
		devices_info[i].max_avail = avail_space;

		i++;
	}

	nr_devices = i;

	btrfs_descending_sort_devices(devices_info, nr_devices);

	i = nr_devices - 1;
	avail_space = 0;
	while (nr_devices >= min_stripes) {
1630 1631 1632
		if (num_stripes > nr_devices)
			num_stripes = nr_devices;

1633 1634 1635 1636
		if (devices_info[i].max_avail >= min_stripe_size) {
			int j;
			u64 alloc_size;

1637
			avail_space += devices_info[i].max_avail * num_stripes;
1638
			alloc_size = devices_info[i].max_avail;
1639
			for (j = i + 1 - num_stripes; j <= i; j++)
1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650
				devices_info[j].max_avail -= alloc_size;
		}
		i--;
		nr_devices--;
	}

	kfree(devices_info);
	*free_bytes = avail_space;
	return 0;
}

C
Chris Mason 已提交
1651 1652
static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf)
{
1653 1654 1655
	struct btrfs_fs_info *fs_info = btrfs_sb(dentry->d_sb);
	struct btrfs_super_block *disk_super = fs_info->super_copy;
	struct list_head *head = &fs_info->space_info;
1656 1657
	struct btrfs_space_info *found;
	u64 total_used = 0;
1658
	u64 total_free_data = 0;
1659
	int bits = dentry->d_sb->s_blocksize_bits;
1660
	__be32 *fsid = (__be32 *)fs_info->fsid;
1661
	int ret;
C
Chris Mason 已提交
1662

1663
	/* holding chunk_muext to avoid allocating new chunks */
1664
	mutex_lock(&fs_info->chunk_mutex);
1665
	rcu_read_lock();
J
Josef Bacik 已提交
1666
	list_for_each_entry_rcu(found, head, list) {
1667 1668 1669 1670 1671 1672
		if (found->flags & BTRFS_BLOCK_GROUP_DATA) {
			total_free_data += found->disk_total - found->disk_used;
			total_free_data -=
				btrfs_account_ro_block_groups_free_space(found);
		}

1673
		total_used += found->disk_used;
J
Josef Bacik 已提交
1674
	}
1675 1676
	rcu_read_unlock();

C
Chris Mason 已提交
1677
	buf->f_namelen = BTRFS_NAME_LEN;
1678
	buf->f_blocks = btrfs_super_total_bytes(disk_super) >> bits;
1679
	buf->f_bfree = buf->f_blocks - (total_used >> bits);
C
Chris Mason 已提交
1680 1681
	buf->f_bsize = dentry->d_sb->s_blocksize;
	buf->f_type = BTRFS_SUPER_MAGIC;
1682
	buf->f_bavail = total_free_data;
1683
	ret = btrfs_calc_avail_data_space(fs_info->tree_root, &total_free_data);
1684
	if (ret) {
1685
		mutex_unlock(&fs_info->chunk_mutex);
1686 1687 1688 1689
		return ret;
	}
	buf->f_bavail += total_free_data;
	buf->f_bavail = buf->f_bavail >> bits;
1690
	mutex_unlock(&fs_info->chunk_mutex);
C
Chris Mason 已提交
1691

1692
	/* We treat it as constant endianness (it doesn't matter _which_)
C
Chris Mason 已提交
1693
	   because we want the fsid to come out the same whether mounted
1694 1695 1696
	   on a big-endian or little-endian host */
	buf->f_fsid.val[0] = be32_to_cpu(fsid[0]) ^ be32_to_cpu(fsid[2]);
	buf->f_fsid.val[1] = be32_to_cpu(fsid[1]) ^ be32_to_cpu(fsid[3]);
1697 1698 1699 1700
	/* Mask in the root object ID too, to disambiguate subvols */
	buf->f_fsid.val[0] ^= BTRFS_I(dentry->d_inode)->root->objectid >> 32;
	buf->f_fsid.val[1] ^= BTRFS_I(dentry->d_inode)->root->objectid;

C
Chris Mason 已提交
1701 1702
	return 0;
}
C
Chris Mason 已提交
1703

A
Al Viro 已提交
1704 1705
static void btrfs_kill_super(struct super_block *sb)
{
1706
	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
A
Al Viro 已提交
1707
	kill_anon_super(sb);
1708
	free_fs_info(fs_info);
A
Al Viro 已提交
1709 1710
}

1711 1712 1713
static struct file_system_type btrfs_fs_type = {
	.owner		= THIS_MODULE,
	.name		= "btrfs",
A
Al Viro 已提交
1714
	.mount		= btrfs_mount,
A
Al Viro 已提交
1715
	.kill_sb	= btrfs_kill_super,
1716 1717
	.fs_flags	= FS_REQUIRES_DEV,
};
1718
MODULE_ALIAS_FS("btrfs");
1719

C
Chris Mason 已提交
1720 1721 1722
/*
 * used by btrfsctl to scan devices when no FS is mounted
 */
1723 1724 1725 1726 1727
static long btrfs_control_ioctl(struct file *file, unsigned int cmd,
				unsigned long arg)
{
	struct btrfs_ioctl_vol_args *vol;
	struct btrfs_fs_devices *fs_devices;
1728
	int ret = -ENOTTY;
1729

1730 1731 1732
	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

L
Li Zefan 已提交
1733 1734 1735
	vol = memdup_user((void __user *)arg, sizeof(*vol));
	if (IS_ERR(vol))
		return PTR_ERR(vol);
1736

1737 1738
	switch (cmd) {
	case BTRFS_IOC_SCAN_DEV:
1739
		ret = btrfs_scan_one_device(vol->name, FMODE_READ,
1740 1741
					    &btrfs_fs_type, &fs_devices);
		break;
J
Josef Bacik 已提交
1742 1743 1744 1745 1746 1747 1748
	case BTRFS_IOC_DEVICES_READY:
		ret = btrfs_scan_one_device(vol->name, FMODE_READ,
					    &btrfs_fs_type, &fs_devices);
		if (ret)
			break;
		ret = !(fs_devices->num_devices == fs_devices->total_devices);
		break;
1749
	}
L
Li Zefan 已提交
1750

1751
	kfree(vol);
L
Linda Knippers 已提交
1752
	return ret;
1753 1754
}

1755
static int btrfs_freeze(struct super_block *sb)
Y
Yan 已提交
1756
{
1757 1758 1759
	struct btrfs_trans_handle *trans;
	struct btrfs_root *root = btrfs_sb(sb)->tree_root;

M
Miao Xie 已提交
1760
	trans = btrfs_attach_transaction_barrier(root);
1761 1762 1763 1764 1765 1766 1767
	if (IS_ERR(trans)) {
		/* no transaction, don't bother */
		if (PTR_ERR(trans) == -ENOENT)
			return 0;
		return PTR_ERR(trans);
	}
	return btrfs_commit_transaction(trans, root);
Y
Yan 已提交
1768 1769
}

1770
static int btrfs_unfreeze(struct super_block *sb)
Y
Yan 已提交
1771
{
1772
	return 0;
Y
Yan 已提交
1773
}
1774

J
Josef Bacik 已提交
1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787
static int btrfs_show_devname(struct seq_file *m, struct dentry *root)
{
	struct btrfs_fs_info *fs_info = btrfs_sb(root->d_sb);
	struct btrfs_fs_devices *cur_devices;
	struct btrfs_device *dev, *first_dev = NULL;
	struct list_head *head;
	struct rcu_string *name;

	mutex_lock(&fs_info->fs_devices->device_list_mutex);
	cur_devices = fs_info->fs_devices;
	while (cur_devices) {
		head = &cur_devices->devices;
		list_for_each_entry(dev, head, dev_list) {
1788 1789
			if (dev->missing)
				continue;
J
Josef Bacik 已提交
1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807
			if (!first_dev || dev->devid < first_dev->devid)
				first_dev = dev;
		}
		cur_devices = cur_devices->seed;
	}

	if (first_dev) {
		rcu_read_lock();
		name = rcu_dereference(first_dev->name);
		seq_escape(m, name->str, " \t\n\\");
		rcu_read_unlock();
	} else {
		WARN_ON(1);
	}
	mutex_unlock(&fs_info->fs_devices->device_list_mutex);
	return 0;
}

1808
static const struct super_operations btrfs_super_ops = {
1809
	.drop_inode	= btrfs_drop_inode,
A
Al Viro 已提交
1810
	.evict_inode	= btrfs_evict_inode,
C
Chris Mason 已提交
1811
	.put_super	= btrfs_put_super,
1812
	.sync_fs	= btrfs_sync_fs,
E
Eric Paris 已提交
1813
	.show_options	= btrfs_show_options,
J
Josef Bacik 已提交
1814
	.show_devname	= btrfs_show_devname,
C
Chris Mason 已提交
1815
	.write_inode	= btrfs_write_inode,
C
Chris Mason 已提交
1816 1817
	.alloc_inode	= btrfs_alloc_inode,
	.destroy_inode	= btrfs_destroy_inode,
C
Chris Mason 已提交
1818
	.statfs		= btrfs_statfs,
Y
Yan Zheng 已提交
1819
	.remount_fs	= btrfs_remount,
1820 1821
	.freeze_fs	= btrfs_freeze,
	.unfreeze_fs	= btrfs_unfreeze,
C
Chris Mason 已提交
1822
};
1823 1824 1825 1826 1827

static const struct file_operations btrfs_ctl_fops = {
	.unlocked_ioctl	 = btrfs_control_ioctl,
	.compat_ioctl = btrfs_control_ioctl,
	.owner	 = THIS_MODULE,
1828
	.llseek = noop_llseek,
1829 1830 1831
};

static struct miscdevice btrfs_misc = {
1832
	.minor		= BTRFS_MINOR,
1833 1834 1835 1836
	.name		= "btrfs-control",
	.fops		= &btrfs_ctl_fops
};

1837 1838 1839
MODULE_ALIAS_MISCDEV(BTRFS_MINOR);
MODULE_ALIAS("devname:btrfs-control");

1840 1841 1842 1843 1844
static int btrfs_interface_init(void)
{
	return misc_register(&btrfs_misc);
}

1845
static void btrfs_interface_exit(void)
1846 1847
{
	if (misc_deregister(&btrfs_misc) < 0)
1848
		printk(KERN_INFO "BTRFS: misc_deregister failed for control device\n");
1849 1850
}

1851 1852 1853 1854 1855 1856
static void btrfs_print_info(void)
{
	printk(KERN_INFO "Btrfs loaded"
#ifdef CONFIG_BTRFS_DEBUG
			", debug=on"
#endif
1857 1858 1859
#ifdef CONFIG_BTRFS_ASSERT
			", assert=on"
#endif
1860 1861 1862 1863 1864 1865
#ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY
			", integrity-checker=on"
#endif
			"\n");
}

1866 1867
static int btrfs_run_sanity_tests(void)
{
1868 1869
	int ret;

1870
	ret = btrfs_init_test_fs();
1871 1872
	if (ret)
		return ret;
1873 1874 1875 1876 1877 1878 1879 1880

	ret = btrfs_test_free_space_cache();
	if (ret)
		goto out;
	ret = btrfs_test_extent_buffer_operations();
	if (ret)
		goto out;
	ret = btrfs_test_extent_io();
1881 1882 1883
	if (ret)
		goto out;
	ret = btrfs_test_inodes();
1884 1885 1886
out:
	btrfs_destroy_test_fs();
	return ret;
1887 1888
}

1889 1890
static int __init init_btrfs_fs(void)
{
C
Chris Mason 已提交
1891
	int err;
1892

1893 1894 1895 1896
	err = btrfs_hash_init();
	if (err)
		return err;

1897 1898
	btrfs_props_init();

1899 1900
	err = btrfs_init_sysfs();
	if (err)
1901
		goto free_hash;
1902

1903
	btrfs_init_compress();
1904

1905 1906 1907 1908
	err = btrfs_init_cachep();
	if (err)
		goto free_compress;

1909
	err = extent_io_init();
1910 1911 1912
	if (err)
		goto free_cachep;

1913 1914 1915 1916
	err = extent_map_init();
	if (err)
		goto free_extent_io;

1917
	err = ordered_data_init();
1918 1919
	if (err)
		goto free_extent_map;
C
Chris Mason 已提交
1920

1921 1922 1923 1924
	err = btrfs_delayed_inode_init();
	if (err)
		goto free_ordered_data;

1925
	err = btrfs_auto_defrag_init();
1926 1927 1928
	if (err)
		goto free_delayed_inode;

1929
	err = btrfs_delayed_ref_init();
1930 1931 1932
	if (err)
		goto free_auto_defrag;

1933 1934 1935 1936
	err = btrfs_prelim_ref_init();
	if (err)
		goto free_prelim_ref;

1937 1938 1939 1940
	err = btrfs_interface_init();
	if (err)
		goto free_delayed_ref;

1941 1942
	btrfs_init_lockdep();

1943
	btrfs_print_info();
1944 1945 1946 1947 1948 1949 1950 1951

	err = btrfs_run_sanity_tests();
	if (err)
		goto unregister_ioctl;

	err = register_filesystem(&btrfs_fs_type);
	if (err)
		goto unregister_ioctl;
1952

1953 1954
	return 0;

1955 1956
unregister_ioctl:
	btrfs_interface_exit();
1957 1958
free_prelim_ref:
	btrfs_prelim_ref_exit();
1959 1960
free_delayed_ref:
	btrfs_delayed_ref_exit();
1961 1962
free_auto_defrag:
	btrfs_auto_defrag_exit();
1963 1964
free_delayed_inode:
	btrfs_delayed_inode_exit();
1965 1966
free_ordered_data:
	ordered_data_exit();
1967 1968
free_extent_map:
	extent_map_exit();
1969 1970
free_extent_io:
	extent_io_exit();
1971 1972
free_cachep:
	btrfs_destroy_cachep();
1973 1974
free_compress:
	btrfs_exit_compress();
1975
	btrfs_exit_sysfs();
1976 1977
free_hash:
	btrfs_hash_exit();
1978
	return err;
1979 1980 1981 1982
}

static void __exit exit_btrfs_fs(void)
{
C
Chris Mason 已提交
1983
	btrfs_destroy_cachep();
1984
	btrfs_delayed_ref_exit();
1985
	btrfs_auto_defrag_exit();
1986
	btrfs_delayed_inode_exit();
1987
	btrfs_prelim_ref_exit();
1988
	ordered_data_exit();
1989
	extent_map_exit();
1990
	extent_io_exit();
1991
	btrfs_interface_exit();
1992
	unregister_filesystem(&btrfs_fs_type);
1993
	btrfs_exit_sysfs();
1994
	btrfs_cleanup_fs_uuids();
1995
	btrfs_exit_compress();
1996
	btrfs_hash_exit();
1997 1998
}

1999
late_initcall(init_btrfs_fs);
2000 2001 2002
module_exit(exit_btrfs_fs)

MODULE_LICENSE("GPL");