super.c 42.0 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>
C
Chris Mason 已提交
44
#include "compat.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 "ioctl.h"
C
Chris Mason 已提交
51
#include "print-tree.h"
J
Josef Bacik 已提交
52
#include "xattr.h"
53
#include "volumes.h"
54
#include "version.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

59 60 61
#define CREATE_TRACE_POINTS
#include <trace/events/btrfs.h>

62
static const struct super_operations btrfs_super_ops;
63
static struct file_system_type btrfs_fs_type;
C
Chris Mason 已提交
64

L
liubo 已提交
65 66 67 68 69 70 71 72 73 74 75 76 77 78 79
static const char *btrfs_decode_error(struct btrfs_fs_info *fs_info, int errno,
				      char nbuf[16])
{
	char *errstr = NULL;

	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
80 81 82
	case -EEXIST:
		errstr = "Object already exists";
		break;
L
liubo 已提交
83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122
	default:
		if (nbuf) {
			if (snprintf(nbuf, 16, "error %d", -errno) >= 0)
				errstr = nbuf;
		}
		break;
	}

	return errstr;
}

static void __save_error_info(struct btrfs_fs_info *fs_info)
{
	/*
	 * today we only save the error info into ram.  Long term we'll
	 * also send it down to the disk
	 */
	fs_info->fs_state = BTRFS_SUPER_FLAG_ERROR;
}

/* NOTE:
 *	We move write_super stuff at umount in order to avoid deadlock
 *	for umount hold all lock.
 */
static void save_error_info(struct btrfs_fs_info *fs_info)
{
	__save_error_info(fs_info);
}

/* 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;

	if (fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR) {
		sb->s_flags |= MS_RDONLY;
		printk(KERN_INFO "btrfs is forced readonly\n");
123 124
		__btrfs_scrub_cancel(fs_info);
//		WARN_ON(1);
L
liubo 已提交
125 126 127 128 129 130 131 132
	}
}

/*
 * __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
133
		       unsigned int line, int errno, const char *fmt, ...)
L
liubo 已提交
134 135 136 137
{
	struct super_block *sb = fs_info->sb;
	char nbuf[16];
	const char *errstr;
J
Jeff Mahoney 已提交
138 139
	va_list args;
	va_start(args, fmt);
L
liubo 已提交
140 141 142 143 144 145

	/*
	 * 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
146 147 148 149 150 151 152 153 154 155 156 157 158 159 160
  		return;

  	errstr = btrfs_decode_error(fs_info, errno, nbuf);
	if (fmt) {
		struct va_format vaf = {
			.fmt = fmt,
			.va = &args,
		};

		printk(KERN_CRIT "BTRFS error (device %s) in %s:%d: %s (%pV)\n",
			sb->s_id, function, line, errstr, &vaf);
	} else {
		printk(KERN_CRIT "BTRFS error (device %s) in %s:%d: %s\n",
			sb->s_id, function, line, errstr);
	}
L
liubo 已提交
161

J
Jeff Mahoney 已提交
162 163 164 165 166 167 168
	/* 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);
	}
	va_end(args);
}
L
liubo 已提交
169

J
Jeff Mahoney 已提交
170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191
const char *logtypes[] = {
	"emergency",
	"alert",
	"critical",
	"error",
	"warning",
	"notice",
	"info",
	"debug",
};

void btrfs_printk(struct btrfs_fs_info *fs_info, const char *fmt, ...)
{
	struct super_block *sb = fs_info->sb;
	char lvl[4];
	struct va_format vaf;
	va_list args;
	const char *type = logtypes[4];

	va_start(args, fmt);

	if (fmt[0] == '<' && isdigit(fmt[1]) && fmt[2] == '>') {
192 193
		memcpy(lvl, fmt, 3);
		lvl[3] = '\0';
J
Jeff Mahoney 已提交
194 195 196 197 198 199 200 201
		fmt += 3;
		type = logtypes[fmt[1] - '0'];
	} else
		*lvl = '\0';

	vaf.fmt = fmt;
	vaf.va = &args;
	printk("%sBTRFS %s (device %s): %pV", lvl, type, sb->s_id, &vaf);
L
liubo 已提交
202 203
}

204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220
/*
 * 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)
{
221
	WARN_ONCE(1, KERN_DEBUG "btrfs: Transaction aborted");
222 223 224 225 226
	trans->aborted = errno;
	/* Nothing used. The other threads that have joined this
	 * transaction may be able to continue. */
	if (!trans->blocks_used) {
		btrfs_printk(root->fs_info, "Aborting unused transaction.\n");
L
liubo 已提交
227
		return;
228 229 230 231
	}
	trans->transaction->aborted = errno;
	__btrfs_std_error(root->fs_info, function, line, errno, NULL);
}
J
Jeff Mahoney 已提交
232 233 234 235 236 237 238 239 240 241 242 243
/*
 * __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 nbuf[16];
	char *s_id = "<unknown>";
	const char *errstr;
	struct va_format vaf = { .fmt = fmt };
	va_list args;
L
liubo 已提交
244

J
Jeff Mahoney 已提交
245 246
	if (fs_info)
		s_id = fs_info->sb->s_id;
L
liubo 已提交
247

J
Jeff Mahoney 已提交
248 249 250 251 252 253 254 255 256 257 258 259
	va_start(args, fmt);
	vaf.va = &args;

	errstr = btrfs_decode_error(fs_info, errno, nbuf);
	if (fs_info->mount_opt & BTRFS_MOUNT_PANIC_ON_FATAL_ERROR)
		panic(KERN_CRIT "BTRFS panic (device %s) in %s:%d: %pV (%s)\n",
			s_id, function, line, &vaf, errstr);

	printk(KERN_CRIT "BTRFS panic (device %s) in %s:%d: %pV (%s)\n",
	       s_id, function, line, &vaf, errstr);
	va_end(args);
	/* Caller calls BUG() */
L
liubo 已提交
260 261
}

C
Chris Mason 已提交
262
static void btrfs_put_super(struct super_block *sb)
C
Chris Mason 已提交
263
{
264
	(void)close_ctree(btrfs_sb(sb)->tree_root);
A
Al Viro 已提交
265 266 267 268 269 270
	/* 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
271 272
}

273
enum {
274
	Opt_degraded, Opt_subvol, Opt_subvolid, Opt_device, Opt_nodatasum,
275 276
	Opt_nodatacow, Opt_max_inline, Opt_alloc_start, Opt_nobarrier, Opt_ssd,
	Opt_nossd, Opt_ssd_spread, Opt_thread_pool, Opt_noacl, Opt_compress,
277 278
	Opt_compress_type, Opt_compress_force, Opt_compress_force_type,
	Opt_notreelog, Opt_ratio, Opt_flushoncommit, Opt_discard,
279
	Opt_space_cache, Opt_clear_cache, Opt_user_subvol_rm_allowed,
280 281
	Opt_enospc_debug, Opt_subvolrootid, Opt_defrag, Opt_inode_cache,
	Opt_no_space_cache, Opt_recovery, Opt_skip_balance,
282
	Opt_check_integrity, Opt_check_integrity_including_extent_data,
J
Jeff Mahoney 已提交
283
	Opt_check_integrity_print_mask, Opt_fatal_errors,
284
	Opt_err,
285 286 287
};

static match_table_t tokens = {
288
	{Opt_degraded, "degraded"},
289
	{Opt_subvol, "subvol=%s"},
290
	{Opt_subvolid, "subvolid=%d"},
291
	{Opt_device, "device=%s"},
292
	{Opt_nodatasum, "nodatasum"},
293
	{Opt_nodatacow, "nodatacow"},
294
	{Opt_nobarrier, "nobarrier"},
295
	{Opt_max_inline, "max_inline=%s"},
296
	{Opt_alloc_start, "alloc_start=%s"},
297
	{Opt_thread_pool, "thread_pool=%d"},
C
Chris Mason 已提交
298
	{Opt_compress, "compress"},
299
	{Opt_compress_type, "compress=%s"},
C
Chris Mason 已提交
300
	{Opt_compress_force, "compress-force"},
301
	{Opt_compress_force_type, "compress-force=%s"},
302
	{Opt_ssd, "ssd"},
303
	{Opt_ssd_spread, "ssd_spread"},
C
Chris Mason 已提交
304
	{Opt_nossd, "nossd"},
J
Josef Bacik 已提交
305
	{Opt_noacl, "noacl"},
S
Sage Weil 已提交
306
	{Opt_notreelog, "notreelog"},
307
	{Opt_flushoncommit, "flushoncommit"},
308
	{Opt_ratio, "metadata_ratio=%d"},
C
Christoph Hellwig 已提交
309
	{Opt_discard, "discard"},
310
	{Opt_space_cache, "space_cache"},
311
	{Opt_clear_cache, "clear_cache"},
312
	{Opt_user_subvol_rm_allowed, "user_subvol_rm_allowed"},
313
	{Opt_enospc_debug, "enospc_debug"},
314
	{Opt_subvolrootid, "subvolrootid=%d"},
C
Chris Mason 已提交
315
	{Opt_defrag, "autodefrag"},
C
Chris Mason 已提交
316
	{Opt_inode_cache, "inode_cache"},
317
	{Opt_no_space_cache, "nospace_cache"},
C
Chris Mason 已提交
318
	{Opt_recovery, "recovery"},
319
	{Opt_skip_balance, "skip_balance"},
320 321 322
	{Opt_check_integrity, "check_int"},
	{Opt_check_integrity_including_extent_data, "check_int_data"},
	{Opt_check_integrity_print_mask, "check_int_print_mask=%d"},
J
Jeff Mahoney 已提交
323
	{Opt_fatal_errors, "fatal_errors=%s"},
J
Josef Bacik 已提交
324
	{Opt_err, NULL},
325 326
};

327 328 329
/*
 * Regular mount options parser.  Everything that is needed only when
 * reading in a new superblock is parsed here.
330
 * XXX JDM: This needs to be cleaned up for remount.
331 332
 */
int btrfs_parse_options(struct btrfs_root *root, char *options)
333
{
334
	struct btrfs_fs_info *info = root->fs_info;
335
	substring_t args[MAX_OPT_ARGS];
336 337
	char *p, *num, *orig = NULL;
	u64 cache_gen;
338
	int intarg;
S
Sage Weil 已提交
339
	int ret = 0;
340 341
	char *compress_type;
	bool compress_force = false;
342

343
	cache_gen = btrfs_super_cache_generation(root->fs_info->super_copy);
344 345 346
	if (cache_gen)
		btrfs_set_opt(info->mount_opt, SPACE_CACHE);

347
	if (!options)
348
		goto out;
349

350 351 352 353 354 355 356 357
	/*
	 * strsep changes the string, duplicate it because parse_options
	 * gets called twice
	 */
	options = kstrdup(options, GFP_NOFS);
	if (!options)
		return -ENOMEM;

358
	orig = options;
359

360
	while ((p = strsep(&options, ",")) != NULL) {
361 362 363 364 365 366
		int token;
		if (!*p)
			continue;

		token = match_token(p, tokens, args);
		switch (token) {
367
		case Opt_degraded:
368 369
			printk(KERN_INFO "btrfs: allowing degraded mounts\n");
			btrfs_set_opt(info->mount_opt, DEGRADED);
370
			break;
371
		case Opt_subvol:
372
		case Opt_subvolid:
373
		case Opt_subvolrootid:
374
		case Opt_device:
375
			/*
376
			 * These are parsed by btrfs_parse_early_options
377 378
			 * and can be happily ignored here.
			 */
379 380
			break;
		case Opt_nodatasum:
381
			printk(KERN_INFO "btrfs: setting nodatasum\n");
382
			btrfs_set_opt(info->mount_opt, NODATASUM);
383 384
			break;
		case Opt_nodatacow:
385 386 387
			printk(KERN_INFO "btrfs: setting nodatacow\n");
			btrfs_set_opt(info->mount_opt, NODATACOW);
			btrfs_set_opt(info->mount_opt, NODATASUM);
388
			break;
C
Chris Mason 已提交
389
		case Opt_compress_force:
390 391 392 393 394 395 396 397 398
		case Opt_compress_force_type:
			compress_force = true;
		case Opt_compress:
		case Opt_compress_type:
			if (token == Opt_compress ||
			    token == Opt_compress_force ||
			    strcmp(args[0].from, "zlib") == 0) {
				compress_type = "zlib";
				info->compress_type = BTRFS_COMPRESS_ZLIB;
L
Li Zefan 已提交
399 400 401
			} else if (strcmp(args[0].from, "lzo") == 0) {
				compress_type = "lzo";
				info->compress_type = BTRFS_COMPRESS_LZO;
402 403 404 405 406
			} else {
				ret = -EINVAL;
				goto out;
			}

C
Chris Mason 已提交
407
			btrfs_set_opt(info->mount_opt, COMPRESS);
408 409 410 411 412 413 414
			if (compress_force) {
				btrfs_set_opt(info->mount_opt, FORCE_COMPRESS);
				pr_info("btrfs: force %s compression\n",
					compress_type);
			} else
				pr_info("btrfs: use %s compression\n",
					compress_type);
C
Chris Mason 已提交
415
			break;
416
		case Opt_ssd:
417 418
			printk(KERN_INFO "btrfs: use ssd allocation scheme\n");
			btrfs_set_opt(info->mount_opt, SSD);
419
			break;
420 421 422 423 424 425
		case Opt_ssd_spread:
			printk(KERN_INFO "btrfs: use spread ssd "
			       "allocation scheme\n");
			btrfs_set_opt(info->mount_opt, SSD);
			btrfs_set_opt(info->mount_opt, SSD_SPREAD);
			break;
C
Chris Mason 已提交
426
		case Opt_nossd:
427 428
			printk(KERN_INFO "btrfs: not using ssd allocation "
			       "scheme\n");
C
Chris Mason 已提交
429
			btrfs_set_opt(info->mount_opt, NOSSD);
C
Chris Mason 已提交
430
			btrfs_clear_opt(info->mount_opt, SSD);
431
			btrfs_clear_opt(info->mount_opt, SSD_SPREAD);
C
Chris Mason 已提交
432
			break;
433
		case Opt_nobarrier:
434 435
			printk(KERN_INFO "btrfs: turning off barriers\n");
			btrfs_set_opt(info->mount_opt, NOBARRIER);
436
			break;
437 438 439
		case Opt_thread_pool:
			intarg = 0;
			match_int(&args[0], &intarg);
440
			if (intarg)
441 442
				info->thread_pool_size = intarg;
			break;
443
		case Opt_max_inline:
444 445
			num = match_strdup(&args[0]);
			if (num) {
A
Akinobu Mita 已提交
446
				info->max_inline = memparse(num, NULL);
447 448
				kfree(num);

C
Chris Mason 已提交
449 450 451 452 453
				if (info->max_inline) {
					info->max_inline = max_t(u64,
						info->max_inline,
						root->sectorsize);
				}
454
				printk(KERN_INFO "btrfs: max_inline at %llu\n",
455
					(unsigned long long)info->max_inline);
456 457
			}
			break;
458
		case Opt_alloc_start:
459 460
			num = match_strdup(&args[0]);
			if (num) {
A
Akinobu Mita 已提交
461
				info->alloc_start = memparse(num, NULL);
462 463 464
				kfree(num);
				printk(KERN_INFO
					"btrfs: allocations start at %llu\n",
465
					(unsigned long long)info->alloc_start);
466 467
			}
			break;
J
Josef Bacik 已提交
468 469 470
		case Opt_noacl:
			root->fs_info->sb->s_flags &= ~MS_POSIXACL;
			break;
S
Sage Weil 已提交
471 472 473 474
		case Opt_notreelog:
			printk(KERN_INFO "btrfs: disabling tree log\n");
			btrfs_set_opt(info->mount_opt, NOTREELOG);
			break;
475 476 477 478
		case Opt_flushoncommit:
			printk(KERN_INFO "btrfs: turning on flush-on-commit\n");
			btrfs_set_opt(info->mount_opt, FLUSHONCOMMIT);
			break;
479 480 481 482 483 484 485 486 487
		case Opt_ratio:
			intarg = 0;
			match_int(&args[0], &intarg);
			if (intarg) {
				info->metadata_ratio = intarg;
				printk(KERN_INFO "btrfs: metadata ratio %d\n",
				       info->metadata_ratio);
			}
			break;
C
Christoph Hellwig 已提交
488 489 490
		case Opt_discard:
			btrfs_set_opt(info->mount_opt, DISCARD);
			break;
491 492
		case Opt_space_cache:
			btrfs_set_opt(info->mount_opt, SPACE_CACHE);
493
			break;
494 495 496 497
		case Opt_no_space_cache:
			printk(KERN_INFO "btrfs: disabling disk space caching\n");
			btrfs_clear_opt(info->mount_opt, SPACE_CACHE);
			break;
C
Chris Mason 已提交
498 499 500 501
		case Opt_inode_cache:
			printk(KERN_INFO "btrfs: enabling inode map caching\n");
			btrfs_set_opt(info->mount_opt, INODE_MAP_CACHE);
			break;
502 503 504
		case Opt_clear_cache:
			printk(KERN_INFO "btrfs: force clearing of disk cache\n");
			btrfs_set_opt(info->mount_opt, CLEAR_CACHE);
505
			break;
506 507 508
		case Opt_user_subvol_rm_allowed:
			btrfs_set_opt(info->mount_opt, USER_SUBVOL_RM_ALLOWED);
			break;
509 510 511
		case Opt_enospc_debug:
			btrfs_set_opt(info->mount_opt, ENOSPC_DEBUG);
			break;
C
Chris Mason 已提交
512 513 514 515
		case Opt_defrag:
			printk(KERN_INFO "btrfs: enabling auto defrag");
			btrfs_set_opt(info->mount_opt, AUTO_DEFRAG);
			break;
C
Chris Mason 已提交
516 517 518 519
		case Opt_recovery:
			printk(KERN_INFO "btrfs: enabling auto recovery");
			btrfs_set_opt(info->mount_opt, RECOVERY);
			break;
520 521 522
		case Opt_skip_balance:
			btrfs_set_opt(info->mount_opt, SKIP_BALANCE);
			break;
523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553
#ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY
		case Opt_check_integrity_including_extent_data:
			printk(KERN_INFO "btrfs: enabling check integrity"
			       " including extent data\n");
			btrfs_set_opt(info->mount_opt,
				      CHECK_INTEGRITY_INCLUDING_EXTENT_DATA);
			btrfs_set_opt(info->mount_opt, CHECK_INTEGRITY);
			break;
		case Opt_check_integrity:
			printk(KERN_INFO "btrfs: enabling check integrity\n");
			btrfs_set_opt(info->mount_opt, CHECK_INTEGRITY);
			break;
		case Opt_check_integrity_print_mask:
			intarg = 0;
			match_int(&args[0], &intarg);
			if (intarg) {
				info->check_integrity_print_mask = intarg;
				printk(KERN_INFO "btrfs:"
				       " check_integrity_print_mask 0x%x\n",
				       info->check_integrity_print_mask);
			}
			break;
#else
		case Opt_check_integrity_including_extent_data:
		case Opt_check_integrity:
		case Opt_check_integrity_print_mask:
			printk(KERN_ERR "btrfs: support for check_integrity*"
			       " not compiled in!\n");
			ret = -EINVAL;
			goto out;
#endif
J
Jeff Mahoney 已提交
554 555 556 557 558 559 560 561 562 563 564 565
		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;
S
Sage Weil 已提交
566 567 568 569 570
		case Opt_err:
			printk(KERN_INFO "btrfs: unrecognized mount option "
			       "'%s'\n", p);
			ret = -EINVAL;
			goto out;
571
		default:
572
			break;
573 574
		}
	}
S
Sage Weil 已提交
575
out:
576 577
	if (!ret && btrfs_test_opt(root, SPACE_CACHE))
		printk(KERN_INFO "btrfs: disk space caching is enabled\n");
578
	kfree(orig);
S
Sage Weil 已提交
579
	return ret;
580 581 582 583 584 585 586 587
}

/*
 * 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.
 */
588
static int btrfs_parse_early_options(const char *options, fmode_t flags,
589
		void *holder, char **subvol_name, u64 *subvol_objectid,
590
		u64 *subvol_rootid, struct btrfs_fs_devices **fs_devices)
591 592
{
	substring_t args[MAX_OPT_ARGS];
593
	char *device_name, *opts, *orig, *p;
594
	int error = 0;
595
	int intarg;
596 597

	if (!options)
598
		return 0;
599 600 601 602 603 604 605 606

	/*
	 * strsep changes the string, duplicate it because parse_options
	 * gets called twice
	 */
	opts = kstrdup(options, GFP_KERNEL);
	if (!opts)
		return -ENOMEM;
607
	orig = opts;
608 609 610 611 612 613 614 615 616

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

		token = match_token(p, tokens, args);
		switch (token) {
		case Opt_subvol:
617
			kfree(*subvol_name);
618 619
			*subvol_name = match_strdup(&args[0]);
			break;
620 621
		case Opt_subvolid:
			intarg = 0;
622 623 624 625 626 627 628 629 630
			error = match_int(&args[0], &intarg);
			if (!error) {
				/* we want the original fs_tree */
				if (!intarg)
					*subvol_objectid =
						BTRFS_FS_TREE_OBJECTID;
				else
					*subvol_objectid = intarg;
			}
631
			break;
632 633 634 635 636 637 638 639 640 641 642 643
		case Opt_subvolrootid:
			intarg = 0;
			error = match_int(&args[0], &intarg);
			if (!error) {
				/* we want the original fs_tree */
				if (!intarg)
					*subvol_rootid =
						BTRFS_FS_TREE_OBJECTID;
				else
					*subvol_rootid = intarg;
			}
			break;
644
		case Opt_device:
645 646 647 648 649 650
			device_name = match_strdup(&args[0]);
			if (!device_name) {
				error = -ENOMEM;
				goto out;
			}
			error = btrfs_scan_one_device(device_name,
651
					flags, holder, fs_devices);
652
			kfree(device_name);
653
			if (error)
654
				goto out;
655
			break;
656 657 658 659 660
		default:
			break;
		}
	}

661
out:
662
	kfree(orig);
663
	return error;
664 665
}

666 667 668
static struct dentry *get_default_root(struct super_block *sb,
				       u64 subvol_objectid)
{
669 670
	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
	struct btrfs_root *root = fs_info->tree_root;
671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699
	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.
	 */
700
	dir_id = btrfs_super_root_dir(fs_info->super_copy);
701
	di = btrfs_lookup_dir_item(NULL, root, path, dir_id, "default", 7, 0);
702 703
	if (IS_ERR(di)) {
		btrfs_free_path(path);
704
		return ERR_CAST(di);
705
	}
706 707 708 709 710 711 712 713
	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;
714
		new_root = fs_info->fs_root;
715 716 717 718 719 720 721
		goto setup_root;
	}

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

find_root:
722
	new_root = btrfs_read_fs_root_no_name(fs_info, &location);
723
	if (IS_ERR(new_root))
J
Julia Lawall 已提交
724
		return ERR_CAST(new_root);
725 726 727 728 729 730 731 732 733 734 735

	if (btrfs_root_refs(&new_root->root_item) == 0)
		return ERR_PTR(-ENOENT);

	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);
736 737
	if (IS_ERR(inode))
		return ERR_CAST(inode);
738 739 740 741 742 743 744 745 746 747 748

	/*
	 * 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);
	}

749
	return d_obtain_alias(inode);
750 751
}

C
Chris Mason 已提交
752
static int btrfs_fill_super(struct super_block *sb,
753
			    struct btrfs_fs_devices *fs_devices,
C
Chris Mason 已提交
754
			    void *data, int silent)
C
Chris Mason 已提交
755
{
C
Chris Mason 已提交
756
	struct inode *inode;
757
	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
758
	struct btrfs_key key;
C
Chris Mason 已提交
759
	int err;
760

C
Chris Mason 已提交
761 762 763
	sb->s_maxbytes = MAX_LFS_FILESIZE;
	sb->s_magic = BTRFS_SUPER_MAGIC;
	sb->s_op = &btrfs_super_ops;
A
Al Viro 已提交
764
	sb->s_d_op = &btrfs_dentry_operations;
B
Balaji Rao 已提交
765
	sb->s_export_op = &btrfs_export_ops;
J
Josef Bacik 已提交
766
	sb->s_xattr = btrfs_xattr_handlers;
C
Chris Mason 已提交
767
	sb->s_time_gran = 1;
C
Chris Mason 已提交
768
#ifdef CONFIG_BTRFS_FS_POSIX_ACL
J
Josef Bacik 已提交
769
	sb->s_flags |= MS_POSIXACL;
770
#endif
771
	sb->s_flags |= MS_I_VERSION;
A
Al Viro 已提交
772 773
	err = open_ctree(sb, fs_devices, (char *)data);
	if (err) {
C
Chris Mason 已提交
774
		printk("btrfs: open_ctree failed\n");
A
Al Viro 已提交
775
		return err;
776 777
	}

778 779 780
	key.objectid = BTRFS_FIRST_FREE_OBJECTID;
	key.type = BTRFS_INODE_ITEM_KEY;
	key.offset = 0;
781
	inode = btrfs_iget(sb, &key, fs_info->fs_root, NULL);
782 783
	if (IS_ERR(inode)) {
		err = PTR_ERR(inode);
C
Chris Mason 已提交
784
		goto fail_close;
C
Chris Mason 已提交
785 786
	}

787 788
	sb->s_root = d_make_root(inode);
	if (!sb->s_root) {
C
Chris Mason 已提交
789 790
		err = -ENOMEM;
		goto fail_close;
C
Chris Mason 已提交
791
	}
792

C
Chris Mason 已提交
793
	save_mount_options(sb, data);
D
Dan Magenheimer 已提交
794
	cleancache_init_fs(sb);
795
	sb->s_flags |= MS_ACTIVE;
C
Chris Mason 已提交
796
	return 0;
C
Chris Mason 已提交
797 798

fail_close:
799
	close_ctree(fs_info->tree_root);
C
Chris Mason 已提交
800
	return err;
C
Chris Mason 已提交
801 802
}

S
Sage Weil 已提交
803
int btrfs_sync_fs(struct super_block *sb, int wait)
C
Chris Mason 已提交
804 805
{
	struct btrfs_trans_handle *trans;
806 807
	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
	struct btrfs_root *root = fs_info->tree_root;
C
Chris Mason 已提交
808
	int ret;
C
Chris Mason 已提交
809

810 811
	trace_btrfs_sync_fs(wait);

C
Chris Mason 已提交
812
	if (!wait) {
813
		filemap_flush(fs_info->btree_inode->i_mapping);
C
Chris Mason 已提交
814 815
		return 0;
	}
816

Y
Yan, Zheng 已提交
817
	btrfs_wait_ordered_extents(root, 0, 0);
818

819
	trans = btrfs_start_transaction(root, 0);
820 821
	if (IS_ERR(trans))
		return PTR_ERR(trans);
C
Chris Mason 已提交
822
	ret = btrfs_commit_transaction(trans, root);
823
	return ret;
C
Chris Mason 已提交
824 825
}

826
static int btrfs_show_options(struct seq_file *seq, struct dentry *dentry)
E
Eric Paris 已提交
827
{
828 829
	struct btrfs_fs_info *info = btrfs_sb(dentry->d_sb);
	struct btrfs_root *root = info->tree_root;
T
Tsutomu Itoh 已提交
830
	char *compress_type;
E
Eric Paris 已提交
831 832 833 834 835 836 837 838 839 840

	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)
841 842
		seq_printf(seq, ",max_inline=%llu",
			   (unsigned long long)info->max_inline);
E
Eric Paris 已提交
843
	if (info->alloc_start != 0)
844 845
		seq_printf(seq, ",alloc_start=%llu",
			   (unsigned long long)info->alloc_start);
E
Eric Paris 已提交
846 847 848
	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
849 850 851 852 853 854 855 856 857 858
	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
859 860
	if (btrfs_test_opt(root, NOSSD))
		seq_puts(seq, ",nossd");
861 862 863
	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
864
		seq_puts(seq, ",ssd");
S
Sage Weil 已提交
865
	if (btrfs_test_opt(root, NOTREELOG))
866
		seq_puts(seq, ",notreelog");
867
	if (btrfs_test_opt(root, FLUSHONCOMMIT))
868
		seq_puts(seq, ",flushoncommit");
869 870
	if (btrfs_test_opt(root, DISCARD))
		seq_puts(seq, ",discard");
E
Eric Paris 已提交
871 872
	if (!(root->fs_info->sb->s_flags & MS_POSIXACL))
		seq_puts(seq, ",noacl");
T
Tsutomu Itoh 已提交
873 874
	if (btrfs_test_opt(root, SPACE_CACHE))
		seq_puts(seq, ",space_cache");
875
	else
876
		seq_puts(seq, ",nospace_cache");
T
Tsutomu Itoh 已提交
877 878 879 880
	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");
881 882 883 884 885 886
	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");
887 888
	if (btrfs_test_opt(root, SKIP_BALANCE))
		seq_puts(seq, ",skip_balance");
J
Jeff Mahoney 已提交
889 890
	if (btrfs_test_opt(root, PANIC_ON_FATAL_ERROR))
		seq_puts(seq, ",fatal_errors=panic");
E
Eric Paris 已提交
891 892 893
	return 0;
}

894
static int btrfs_test_super(struct super_block *s, void *data)
Y
Yan 已提交
895
{
896 897
	struct btrfs_fs_info *p = data;
	struct btrfs_fs_info *fs_info = btrfs_sb(s);
Y
Yan 已提交
898

899
	return fs_info->fs_devices == p->fs_devices;
Y
Yan 已提交
900 901
}

902 903
static int btrfs_set_super(struct super_block *s, void *data)
{
A
Al Viro 已提交
904 905 906 907
	int err = set_anon_super(s, data);
	if (!err)
		s->s_fs_info = data;
	return err;
Y
Yan 已提交
908 909
}

910 911 912 913 914 915 916 917 918 919
/*
 * 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;
}

920 921 922 923 924 925 926
/*
 * 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)
{
927 928
	unsigned len = strlen(args) + 2 + 1;
	char *src, *dst, *buf;
929 930

	/*
931 932
	 * We need the same args as before, but with this substitution:
	 * s!subvol=[^,]+!subvolid=0!
933
	 *
934 935
	 * Since the replacement string is up to 2 bytes longer than the
	 * original, allocate strlen(args) + 2 + 1 bytes.
936 937
	 */

938
	src = strstr(args, "subvol=");
939
	/* This shouldn't happen, but just in case.. */
940 941 942 943 944
	if (!src)
		return NULL;

	buf = dst = kmalloc(len, GFP_NOFS);
	if (!buf)
945 946 947
		return NULL;

	/*
948 949
	 * If the subvol= arg is not at the start of the string,
	 * copy whatever precedes it into buf.
950
	 */
951 952 953 954
	if (src != args) {
		*src++ = '\0';
		strcpy(buf, args);
		dst += strlen(args);
955 956
	}

957 958
	strcpy(dst, "subvolid=0");
	dst += strlen("subvolid=0");
959 960

	/*
961 962
	 * If there is a "," after the original subvol=... string,
	 * copy that suffix into our buffer.  Otherwise, we're done.
963
	 */
964 965 966
	src = strchr(src, ',');
	if (src)
		strcpy(dst, src);
967

968
	return buf;
969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986
}

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
987
	root = mount_subtree(mnt, subvol_name);
988

A
Al Viro 已提交
989 990 991 992 993
	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);
994 995 996 997
		printk(KERN_ERR "btrfs: '%s' is not a valid subvolume\n",
				subvol_name);
	}

998 999
	return root;
}
1000

1001 1002 1003 1004 1005 1006
/*
 * 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
1007
static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags,
1008
		const char *device_name, void *data)
Y
Yan 已提交
1009 1010 1011 1012
{
	struct block_device *bdev = NULL;
	struct super_block *s;
	struct dentry *root;
1013
	struct btrfs_fs_devices *fs_devices = NULL;
1014
	struct btrfs_fs_info *fs_info = NULL;
1015
	fmode_t mode = FMODE_READ;
1016 1017
	char *subvol_name = NULL;
	u64 subvol_objectid = 0;
1018
	u64 subvol_rootid = 0;
Y
Yan 已提交
1019 1020
	int error = 0;

1021 1022 1023 1024
	if (!(flags & MS_RDONLY))
		mode |= FMODE_WRITE;

	error = btrfs_parse_early_options(data, mode, fs_type,
1025
					  &subvol_name, &subvol_objectid,
1026
					  &subvol_rootid, &fs_devices);
1027 1028
	if (error) {
		kfree(subvol_name);
A
Al Viro 已提交
1029
		return ERR_PTR(error);
1030
	}
1031

1032 1033 1034 1035 1036 1037
	if (subvol_name) {
		root = mount_subvol(subvol_name, flags, device_name, data);
		kfree(subvol_name);
		return root;
	}

1038
	error = btrfs_scan_one_device(device_name, mode, fs_type, &fs_devices);
1039
	if (error)
1040
		return ERR_PTR(error);
Y
Yan 已提交
1041

1042 1043 1044 1045 1046 1047 1048
	/*
	 * 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);
1049 1050 1051
	if (!fs_info)
		return ERR_PTR(-ENOMEM);

1052 1053
	fs_info->fs_devices = fs_devices;

1054 1055 1056 1057
	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;
1058 1059 1060 1061 1062 1063 1064 1065 1066
		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;
1067 1068 1069
		goto error_close_devices;
	}

1070
	bdev = fs_devices->latest_bdev;
1071
	s = sget(fs_type, btrfs_test_super, btrfs_set_super, fs_info);
1072 1073 1074 1075
	if (IS_ERR(s)) {
		error = PTR_ERR(s);
		goto error_close_devices;
	}
Y
Yan 已提交
1076 1077

	if (s->s_root) {
Y
Yan Zheng 已提交
1078
		btrfs_close_devices(fs_devices);
1079
		free_fs_info(fs_info);
1080 1081
		if ((flags ^ s->s_flags) & MS_RDONLY)
			error = -EBUSY;
Y
Yan 已提交
1082 1083 1084
	} else {
		char b[BDEVNAME_SIZE];

A
Al Viro 已提交
1085
		s->s_flags = flags | MS_NOSEC;
Y
Yan 已提交
1086
		strlcpy(s->s_id, bdevname(bdev, b), sizeof(s->s_id));
1087
		btrfs_sb(s)->bdev_holder = fs_type;
1088 1089
		error = btrfs_fill_super(s, fs_devices, data,
					 flags & MS_SILENT ? 1 : 0);
Y
Yan 已提交
1090 1091
	}

1092 1093
	root = !error ? get_default_root(s, subvol_objectid) : ERR_PTR(error);
	if (IS_ERR(root))
1094
		deactivate_locked_super(s);
Y
Yan 已提交
1095

A
Al Viro 已提交
1096
	return root;
Y
Yan 已提交
1097

Y
Yan Zheng 已提交
1098
error_close_devices:
1099
	btrfs_close_devices(fs_devices);
1100
error_fs_info:
1101
	free_fs_info(fs_info);
A
Al Viro 已提交
1102
	return ERR_PTR(error);
Y
Yan 已提交
1103
}
1104

1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138
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;

	printk(KERN_INFO "btrfs: resize thread pool %d -> %d\n",
	       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);
	btrfs_set_max_workers(&fs_info->scrub_workers, new_pool_size);
}

Y
Yan Zheng 已提交
1139 1140
static int btrfs_remount(struct super_block *sb, int *flags, char *data)
{
1141 1142
	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
	struct btrfs_root *root = fs_info->tree_root;
1143 1144 1145 1146 1147 1148 1149
	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
1150 1151
	int ret;

1152
	ret = btrfs_parse_options(root, data);
1153 1154 1155 1156
	if (ret) {
		ret = -EINVAL;
		goto restore;
	}
1157

1158 1159 1160
	btrfs_resize_thread_pool(fs_info,
		fs_info->thread_pool_size, old_thread_pool_size);

Y
Yan Zheng 已提交
1161 1162 1163 1164 1165 1166
	if ((*flags & MS_RDONLY) == (sb->s_flags & MS_RDONLY))
		return 0;

	if (*flags & MS_RDONLY) {
		sb->s_flags |= MS_RDONLY;

1167 1168 1169
		ret = btrfs_commit_super(root);
		if (ret)
			goto restore;
Y
Yan Zheng 已提交
1170
	} else {
1171
		if (fs_info->fs_devices->rw_devices == 0) {
1172 1173
			ret = -EACCES;
			goto restore;
1174
		}
Y
Yan Zheng 已提交
1175

1176
		if (btrfs_super_log_root(fs_info->super_copy) != 0) {
1177 1178
			ret = -EINVAL;
			goto restore;
1179
		}
Y
Yan Zheng 已提交
1180

1181
		ret = btrfs_cleanup_fs_roots(fs_info);
1182 1183
		if (ret)
			goto restore;
Y
Yan Zheng 已提交
1184

1185 1186
		/* recover relocation */
		ret = btrfs_recover_relocation(root);
1187 1188
		if (ret)
			goto restore;
Y
Yan Zheng 已提交
1189

1190 1191 1192 1193
		ret = btrfs_resume_balance_async(fs_info);
		if (ret)
			goto restore;

Y
Yan Zheng 已提交
1194 1195 1196 1197
		sb->s_flags &= ~MS_RDONLY;
	}

	return 0;
1198 1199 1200 1201 1202 1203 1204 1205 1206 1207

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;
	fs_info->alloc_start = old_alloc_start;
1208 1209
	btrfs_resize_thread_pool(fs_info,
		old_thread_pool_size, fs_info->thread_pool_size);
1210 1211
	fs_info->metadata_ratio = old_metadata_ratio;
	return ret;
Y
Yan Zheng 已提交
1212 1213
}

1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239
/* 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);
}

1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254
/*
 * 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;
1255
	int min_stripes = 1, num_stripes = 1;
1256 1257 1258
	int i = 0, nr_devices;
	int ret;

1259
	nr_devices = fs_info->fs_devices->open_devices;
1260 1261 1262 1263 1264 1265 1266 1267 1268
	BUG_ON(!nr_devices);

	devices_info = kmalloc(sizeof(*devices_info) * nr_devices,
			       GFP_NOFS);
	if (!devices_info)
		return -ENOMEM;

	/* calc min stripe number for data space alloction */
	type = btrfs_get_alloc_profile(root, 1);
1269
	if (type & BTRFS_BLOCK_GROUP_RAID0) {
1270
		min_stripes = 2;
1271 1272
		num_stripes = nr_devices;
	} else if (type & BTRFS_BLOCK_GROUP_RAID1) {
1273
		min_stripes = 2;
1274 1275
		num_stripes = 2;
	} else if (type & BTRFS_BLOCK_GROUP_RAID10) {
1276
		min_stripes = 4;
1277 1278
		num_stripes = 4;
	}
1279 1280 1281 1282 1283 1284

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

1285 1286
	list_for_each_entry(device, &fs_devices->devices, dev_list) {
		if (!device->in_fs_metadata || !device->bdev)
1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346
			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) {
1347 1348 1349
		if (num_stripes > nr_devices)
			num_stripes = nr_devices;

1350 1351 1352 1353
		if (devices_info[i].max_avail >= min_stripe_size) {
			int j;
			u64 alloc_size;

1354
			avail_space += devices_info[i].max_avail * num_stripes;
1355
			alloc_size = devices_info[i].max_avail;
1356
			for (j = i + 1 - num_stripes; j <= i; j++)
1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367
				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
1368 1369
static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf)
{
1370 1371 1372
	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;
1373 1374
	struct btrfs_space_info *found;
	u64 total_used = 0;
1375
	u64 total_free_data = 0;
1376
	int bits = dentry->d_sb->s_blocksize_bits;
1377
	__be32 *fsid = (__be32 *)fs_info->fsid;
1378
	int ret;
C
Chris Mason 已提交
1379

1380
	/* holding chunk_muext to avoid allocating new chunks */
1381
	mutex_lock(&fs_info->chunk_mutex);
1382
	rcu_read_lock();
J
Josef Bacik 已提交
1383
	list_for_each_entry_rcu(found, head, list) {
1384 1385 1386 1387 1388 1389
		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);
		}

1390
		total_used += found->disk_used;
J
Josef Bacik 已提交
1391
	}
1392 1393
	rcu_read_unlock();

C
Chris Mason 已提交
1394
	buf->f_namelen = BTRFS_NAME_LEN;
1395
	buf->f_blocks = btrfs_super_total_bytes(disk_super) >> bits;
1396
	buf->f_bfree = buf->f_blocks - (total_used >> bits);
C
Chris Mason 已提交
1397 1398
	buf->f_bsize = dentry->d_sb->s_blocksize;
	buf->f_type = BTRFS_SUPER_MAGIC;
1399
	buf->f_bavail = total_free_data;
1400
	ret = btrfs_calc_avail_data_space(fs_info->tree_root, &total_free_data);
1401
	if (ret) {
1402
		mutex_unlock(&fs_info->chunk_mutex);
1403 1404 1405 1406
		return ret;
	}
	buf->f_bavail += total_free_data;
	buf->f_bavail = buf->f_bavail >> bits;
1407
	mutex_unlock(&fs_info->chunk_mutex);
C
Chris Mason 已提交
1408

1409
	/* We treat it as constant endianness (it doesn't matter _which_)
C
Chris Mason 已提交
1410
	   because we want the fsid to come out the same whether mounted
1411 1412 1413
	   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]);
1414 1415 1416 1417
	/* 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
1418 1419
	return 0;
}
C
Chris Mason 已提交
1420

A
Al Viro 已提交
1421 1422
static void btrfs_kill_super(struct super_block *sb)
{
1423
	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
A
Al Viro 已提交
1424
	kill_anon_super(sb);
1425
	free_fs_info(fs_info);
A
Al Viro 已提交
1426 1427
}

1428 1429 1430
static struct file_system_type btrfs_fs_type = {
	.owner		= THIS_MODULE,
	.name		= "btrfs",
A
Al Viro 已提交
1431
	.mount		= btrfs_mount,
A
Al Viro 已提交
1432
	.kill_sb	= btrfs_kill_super,
1433 1434
	.fs_flags	= FS_REQUIRES_DEV,
};
1435

C
Chris Mason 已提交
1436 1437 1438
/*
 * used by btrfsctl to scan devices when no FS is mounted
 */
1439 1440 1441 1442 1443
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;
1444
	int ret = -ENOTTY;
1445

1446 1447 1448
	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

L
Li Zefan 已提交
1449 1450 1451
	vol = memdup_user((void __user *)arg, sizeof(*vol));
	if (IS_ERR(vol))
		return PTR_ERR(vol);
1452

1453 1454
	switch (cmd) {
	case BTRFS_IOC_SCAN_DEV:
1455
		ret = btrfs_scan_one_device(vol->name, FMODE_READ,
1456 1457 1458
					    &btrfs_fs_type, &fs_devices);
		break;
	}
L
Li Zefan 已提交
1459

1460
	kfree(vol);
L
Linda Knippers 已提交
1461
	return ret;
1462 1463
}

1464
static int btrfs_freeze(struct super_block *sb)
Y
Yan 已提交
1465
{
1466 1467 1468
	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
	mutex_lock(&fs_info->transaction_kthread_mutex);
	mutex_lock(&fs_info->cleaner_mutex);
1469
	return 0;
Y
Yan 已提交
1470 1471
}

1472
static int btrfs_unfreeze(struct super_block *sb)
Y
Yan 已提交
1473
{
1474 1475 1476
	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
	mutex_unlock(&fs_info->cleaner_mutex);
	mutex_unlock(&fs_info->transaction_kthread_mutex);
1477
	return 0;
Y
Yan 已提交
1478
}
1479

J
Josef Bacik 已提交
1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510
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) {
			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;
}

1511
static const struct super_operations btrfs_super_ops = {
1512
	.drop_inode	= btrfs_drop_inode,
A
Al Viro 已提交
1513
	.evict_inode	= btrfs_evict_inode,
C
Chris Mason 已提交
1514
	.put_super	= btrfs_put_super,
1515
	.sync_fs	= btrfs_sync_fs,
E
Eric Paris 已提交
1516
	.show_options	= btrfs_show_options,
J
Josef Bacik 已提交
1517
	.show_devname	= btrfs_show_devname,
C
Chris Mason 已提交
1518
	.write_inode	= btrfs_write_inode,
C
Chris Mason 已提交
1519 1520
	.alloc_inode	= btrfs_alloc_inode,
	.destroy_inode	= btrfs_destroy_inode,
C
Chris Mason 已提交
1521
	.statfs		= btrfs_statfs,
Y
Yan Zheng 已提交
1522
	.remount_fs	= btrfs_remount,
1523 1524
	.freeze_fs	= btrfs_freeze,
	.unfreeze_fs	= btrfs_unfreeze,
C
Chris Mason 已提交
1525
};
1526 1527 1528 1529 1530

static const struct file_operations btrfs_ctl_fops = {
	.unlocked_ioctl	 = btrfs_control_ioctl,
	.compat_ioctl = btrfs_control_ioctl,
	.owner	 = THIS_MODULE,
1531
	.llseek = noop_llseek,
1532 1533 1534
};

static struct miscdevice btrfs_misc = {
1535
	.minor		= BTRFS_MINOR,
1536 1537 1538 1539
	.name		= "btrfs-control",
	.fops		= &btrfs_ctl_fops
};

1540 1541 1542
MODULE_ALIAS_MISCDEV(BTRFS_MINOR);
MODULE_ALIAS("devname:btrfs-control");

1543 1544 1545 1546 1547
static int btrfs_interface_init(void)
{
	return misc_register(&btrfs_misc);
}

1548
static void btrfs_interface_exit(void)
1549 1550
{
	if (misc_deregister(&btrfs_misc) < 0)
C
Chris Mason 已提交
1551
		printk(KERN_INFO "misc_deregister failed for control device");
1552 1553
}

1554 1555
static int __init init_btrfs_fs(void)
{
C
Chris Mason 已提交
1556
	int err;
1557 1558 1559 1560 1561

	err = btrfs_init_sysfs();
	if (err)
		return err;

1562
	btrfs_init_compress();
1563

1564 1565 1566 1567
	err = btrfs_init_cachep();
	if (err)
		goto free_compress;

1568
	err = extent_io_init();
1569 1570 1571
	if (err)
		goto free_cachep;

1572 1573 1574 1575
	err = extent_map_init();
	if (err)
		goto free_extent_io;

1576
	err = btrfs_delayed_inode_init();
1577 1578
	if (err)
		goto free_extent_map;
C
Chris Mason 已提交
1579

1580 1581 1582 1583
	err = btrfs_interface_init();
	if (err)
		goto free_delayed_inode;

1584 1585 1586
	err = register_filesystem(&btrfs_fs_type);
	if (err)
		goto unregister_ioctl;
1587

1588 1589
	btrfs_init_lockdep();

1590
	printk(KERN_INFO "%s loaded\n", BTRFS_BUILD_VERSION);
1591 1592
	return 0;

1593 1594
unregister_ioctl:
	btrfs_interface_exit();
1595 1596
free_delayed_inode:
	btrfs_delayed_inode_exit();
1597 1598
free_extent_map:
	extent_map_exit();
1599 1600
free_extent_io:
	extent_io_exit();
1601 1602
free_cachep:
	btrfs_destroy_cachep();
1603 1604
free_compress:
	btrfs_exit_compress();
1605 1606
	btrfs_exit_sysfs();
	return err;
1607 1608 1609 1610
}

static void __exit exit_btrfs_fs(void)
{
C
Chris Mason 已提交
1611
	btrfs_destroy_cachep();
1612
	btrfs_delayed_inode_exit();
1613
	extent_map_exit();
1614
	extent_io_exit();
1615
	btrfs_interface_exit();
1616
	unregister_filesystem(&btrfs_fs_type);
1617
	btrfs_exit_sysfs();
1618
	btrfs_cleanup_fs_uuids();
1619
	btrfs_exit_compress();
1620 1621 1622 1623 1624 1625
}

module_init(init_btrfs_fs)
module_exit(exit_btrfs_fs)

MODULE_LICENSE("GPL");