super.c 42.5 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;
399
				btrfs_set_opt(info->mount_opt, COMPRESS);
L
Li Zefan 已提交
400 401 402
			} else if (strcmp(args[0].from, "lzo") == 0) {
				compress_type = "lzo";
				info->compress_type = BTRFS_COMPRESS_LZO;
403
				btrfs_set_opt(info->mount_opt, COMPRESS);
404
				btrfs_set_fs_incompat(info, COMPRESS_LZO);
405 406 407 408 409 410
			} else if (strncmp(args[0].from, "no", 2) == 0) {
				compress_type = "no";
				info->compress_type = BTRFS_COMPRESS_NONE;
				btrfs_clear_opt(info->mount_opt, COMPRESS);
				btrfs_clear_opt(info->mount_opt, FORCE_COMPRESS);
				compress_force = false;
411 412 413 414 415 416 417 418 419 420 421 422
			} else {
				ret = -EINVAL;
				goto out;
			}

			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
423
			break;
424
		case Opt_ssd:
425 426
			printk(KERN_INFO "btrfs: use ssd allocation scheme\n");
			btrfs_set_opt(info->mount_opt, SSD);
427
			break;
428 429 430 431 432 433
		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
434
		case Opt_nossd:
435 436
			printk(KERN_INFO "btrfs: not using ssd allocation "
			       "scheme\n");
C
Chris Mason 已提交
437
			btrfs_set_opt(info->mount_opt, NOSSD);
C
Chris Mason 已提交
438
			btrfs_clear_opt(info->mount_opt, SSD);
439
			btrfs_clear_opt(info->mount_opt, SSD_SPREAD);
C
Chris Mason 已提交
440
			break;
441
		case Opt_nobarrier:
442 443
			printk(KERN_INFO "btrfs: turning off barriers\n");
			btrfs_set_opt(info->mount_opt, NOBARRIER);
444
			break;
445 446 447
		case Opt_thread_pool:
			intarg = 0;
			match_int(&args[0], &intarg);
448
			if (intarg)
449 450
				info->thread_pool_size = intarg;
			break;
451
		case Opt_max_inline:
452 453
			num = match_strdup(&args[0]);
			if (num) {
A
Akinobu Mita 已提交
454
				info->max_inline = memparse(num, NULL);
455 456
				kfree(num);

C
Chris Mason 已提交
457 458 459 460 461
				if (info->max_inline) {
					info->max_inline = max_t(u64,
						info->max_inline,
						root->sectorsize);
				}
462
				printk(KERN_INFO "btrfs: max_inline at %llu\n",
463
					(unsigned long long)info->max_inline);
464 465
			}
			break;
466
		case Opt_alloc_start:
467 468
			num = match_strdup(&args[0]);
			if (num) {
A
Akinobu Mita 已提交
469
				info->alloc_start = memparse(num, NULL);
470 471 472
				kfree(num);
				printk(KERN_INFO
					"btrfs: allocations start at %llu\n",
473
					(unsigned long long)info->alloc_start);
474 475
			}
			break;
J
Josef Bacik 已提交
476 477 478
		case Opt_noacl:
			root->fs_info->sb->s_flags &= ~MS_POSIXACL;
			break;
S
Sage Weil 已提交
479 480 481 482
		case Opt_notreelog:
			printk(KERN_INFO "btrfs: disabling tree log\n");
			btrfs_set_opt(info->mount_opt, NOTREELOG);
			break;
483 484 485 486
		case Opt_flushoncommit:
			printk(KERN_INFO "btrfs: turning on flush-on-commit\n");
			btrfs_set_opt(info->mount_opt, FLUSHONCOMMIT);
			break;
487 488 489 490 491 492 493 494 495
		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
496 497 498
		case Opt_discard:
			btrfs_set_opt(info->mount_opt, DISCARD);
			break;
499 500
		case Opt_space_cache:
			btrfs_set_opt(info->mount_opt, SPACE_CACHE);
501
			break;
502 503 504 505
		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
506 507 508 509
		case Opt_inode_cache:
			printk(KERN_INFO "btrfs: enabling inode map caching\n");
			btrfs_set_opt(info->mount_opt, INODE_MAP_CACHE);
			break;
510 511 512
		case Opt_clear_cache:
			printk(KERN_INFO "btrfs: force clearing of disk cache\n");
			btrfs_set_opt(info->mount_opt, CLEAR_CACHE);
513
			break;
514 515 516
		case Opt_user_subvol_rm_allowed:
			btrfs_set_opt(info->mount_opt, USER_SUBVOL_RM_ALLOWED);
			break;
517 518 519
		case Opt_enospc_debug:
			btrfs_set_opt(info->mount_opt, ENOSPC_DEBUG);
			break;
C
Chris Mason 已提交
520 521 522 523
		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
524 525 526 527
		case Opt_recovery:
			printk(KERN_INFO "btrfs: enabling auto recovery");
			btrfs_set_opt(info->mount_opt, RECOVERY);
			break;
528 529 530
		case Opt_skip_balance:
			btrfs_set_opt(info->mount_opt, SKIP_BALANCE);
			break;
531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561
#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
562 563 564 565 566 567 568 569 570 571 572 573
		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
574 575 576 577 578
		case Opt_err:
			printk(KERN_INFO "btrfs: unrecognized mount option "
			       "'%s'\n", p);
			ret = -EINVAL;
			goto out;
579
		default:
580
			break;
581 582
		}
	}
S
Sage Weil 已提交
583
out:
584 585
	if (!ret && btrfs_test_opt(root, SPACE_CACHE))
		printk(KERN_INFO "btrfs: disk space caching is enabled\n");
586
	kfree(orig);
S
Sage Weil 已提交
587
	return ret;
588 589 590 591 592 593 594 595
}

/*
 * 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.
 */
596
static int btrfs_parse_early_options(const char *options, fmode_t flags,
597
		void *holder, char **subvol_name, u64 *subvol_objectid,
598
		u64 *subvol_rootid, struct btrfs_fs_devices **fs_devices)
599 600
{
	substring_t args[MAX_OPT_ARGS];
601
	char *device_name, *opts, *orig, *p;
602
	int error = 0;
603
	int intarg;
604 605

	if (!options)
606
		return 0;
607 608 609 610 611 612 613 614

	/*
	 * strsep changes the string, duplicate it because parse_options
	 * gets called twice
	 */
	opts = kstrdup(options, GFP_KERNEL);
	if (!opts)
		return -ENOMEM;
615
	orig = opts;
616 617 618 619 620 621 622 623 624

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

		token = match_token(p, tokens, args);
		switch (token) {
		case Opt_subvol:
625
			kfree(*subvol_name);
626 627
			*subvol_name = match_strdup(&args[0]);
			break;
628 629
		case Opt_subvolid:
			intarg = 0;
630 631 632 633 634 635 636 637 638
			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;
			}
639
			break;
640 641 642 643 644 645 646 647 648 649 650 651
		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;
652
		case Opt_device:
653 654 655 656 657 658
			device_name = match_strdup(&args[0]);
			if (!device_name) {
				error = -ENOMEM;
				goto out;
			}
			error = btrfs_scan_one_device(device_name,
659
					flags, holder, fs_devices);
660
			kfree(device_name);
661
			if (error)
662
				goto out;
663
			break;
664 665 666 667 668
		default:
			break;
		}
	}

669
out:
670
	kfree(orig);
671
	return error;
672 673
}

674 675 676
static struct dentry *get_default_root(struct super_block *sb,
				       u64 subvol_objectid)
{
677 678
	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
	struct btrfs_root *root = fs_info->tree_root;
679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707
	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.
	 */
708
	dir_id = btrfs_super_root_dir(fs_info->super_copy);
709
	di = btrfs_lookup_dir_item(NULL, root, path, dir_id, "default", 7, 0);
710 711
	if (IS_ERR(di)) {
		btrfs_free_path(path);
712
		return ERR_CAST(di);
713
	}
714 715 716 717 718 719 720 721
	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;
722
		new_root = fs_info->fs_root;
723 724 725 726 727 728 729
		goto setup_root;
	}

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

find_root:
730
	new_root = btrfs_read_fs_root_no_name(fs_info, &location);
731
	if (IS_ERR(new_root))
J
Julia Lawall 已提交
732
		return ERR_CAST(new_root);
733 734 735 736 737 738 739 740 741 742 743

	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);
744 745
	if (IS_ERR(inode))
		return ERR_CAST(inode);
746 747 748 749 750 751 752 753 754 755 756

	/*
	 * 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);
	}

757
	return d_obtain_alias(inode);
758 759
}

C
Chris Mason 已提交
760
static int btrfs_fill_super(struct super_block *sb,
761
			    struct btrfs_fs_devices *fs_devices,
C
Chris Mason 已提交
762
			    void *data, int silent)
C
Chris Mason 已提交
763
{
C
Chris Mason 已提交
764
	struct inode *inode;
765
	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
766
	struct btrfs_key key;
C
Chris Mason 已提交
767
	int err;
768

C
Chris Mason 已提交
769 770 771
	sb->s_maxbytes = MAX_LFS_FILESIZE;
	sb->s_magic = BTRFS_SUPER_MAGIC;
	sb->s_op = &btrfs_super_ops;
A
Al Viro 已提交
772
	sb->s_d_op = &btrfs_dentry_operations;
B
Balaji Rao 已提交
773
	sb->s_export_op = &btrfs_export_ops;
J
Josef Bacik 已提交
774
	sb->s_xattr = btrfs_xattr_handlers;
C
Chris Mason 已提交
775
	sb->s_time_gran = 1;
C
Chris Mason 已提交
776
#ifdef CONFIG_BTRFS_FS_POSIX_ACL
J
Josef Bacik 已提交
777
	sb->s_flags |= MS_POSIXACL;
778
#endif
779
	sb->s_flags |= MS_I_VERSION;
A
Al Viro 已提交
780 781
	err = open_ctree(sb, fs_devices, (char *)data);
	if (err) {
C
Chris Mason 已提交
782
		printk("btrfs: open_ctree failed\n");
A
Al Viro 已提交
783
		return err;
784 785
	}

786 787 788
	key.objectid = BTRFS_FIRST_FREE_OBJECTID;
	key.type = BTRFS_INODE_ITEM_KEY;
	key.offset = 0;
789
	inode = btrfs_iget(sb, &key, fs_info->fs_root, NULL);
790 791
	if (IS_ERR(inode)) {
		err = PTR_ERR(inode);
C
Chris Mason 已提交
792
		goto fail_close;
C
Chris Mason 已提交
793 794
	}

795 796
	sb->s_root = d_make_root(inode);
	if (!sb->s_root) {
C
Chris Mason 已提交
797 798
		err = -ENOMEM;
		goto fail_close;
C
Chris Mason 已提交
799
	}
800

C
Chris Mason 已提交
801
	save_mount_options(sb, data);
D
Dan Magenheimer 已提交
802
	cleancache_init_fs(sb);
803
	sb->s_flags |= MS_ACTIVE;
C
Chris Mason 已提交
804
	return 0;
C
Chris Mason 已提交
805 806

fail_close:
807
	close_ctree(fs_info->tree_root);
C
Chris Mason 已提交
808
	return err;
C
Chris Mason 已提交
809 810
}

S
Sage Weil 已提交
811
int btrfs_sync_fs(struct super_block *sb, int wait)
C
Chris Mason 已提交
812 813
{
	struct btrfs_trans_handle *trans;
814 815
	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
	struct btrfs_root *root = fs_info->tree_root;
C
Chris Mason 已提交
816
	int ret;
C
Chris Mason 已提交
817

818 819
	trace_btrfs_sync_fs(wait);

C
Chris Mason 已提交
820
	if (!wait) {
821
		filemap_flush(fs_info->btree_inode->i_mapping);
C
Chris Mason 已提交
822 823
		return 0;
	}
824

Y
Yan, Zheng 已提交
825
	btrfs_wait_ordered_extents(root, 0, 0);
826

827
	trans = btrfs_start_transaction(root, 0);
828 829
	if (IS_ERR(trans))
		return PTR_ERR(trans);
C
Chris Mason 已提交
830
	ret = btrfs_commit_transaction(trans, root);
831
	return ret;
C
Chris Mason 已提交
832 833
}

834
static int btrfs_show_options(struct seq_file *seq, struct dentry *dentry)
E
Eric Paris 已提交
835
{
836 837
	struct btrfs_fs_info *info = btrfs_sb(dentry->d_sb);
	struct btrfs_root *root = info->tree_root;
T
Tsutomu Itoh 已提交
838
	char *compress_type;
E
Eric Paris 已提交
839 840 841 842 843 844 845 846 847 848

	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)
849 850
		seq_printf(seq, ",max_inline=%llu",
			   (unsigned long long)info->max_inline);
E
Eric Paris 已提交
851
	if (info->alloc_start != 0)
852 853
		seq_printf(seq, ",alloc_start=%llu",
			   (unsigned long long)info->alloc_start);
E
Eric Paris 已提交
854 855 856
	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
857 858 859 860 861 862 863 864 865 866
	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
867 868
	if (btrfs_test_opt(root, NOSSD))
		seq_puts(seq, ",nossd");
869 870 871
	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
872
		seq_puts(seq, ",ssd");
S
Sage Weil 已提交
873
	if (btrfs_test_opt(root, NOTREELOG))
874
		seq_puts(seq, ",notreelog");
875
	if (btrfs_test_opt(root, FLUSHONCOMMIT))
876
		seq_puts(seq, ",flushoncommit");
877 878
	if (btrfs_test_opt(root, DISCARD))
		seq_puts(seq, ",discard");
E
Eric Paris 已提交
879 880
	if (!(root->fs_info->sb->s_flags & MS_POSIXACL))
		seq_puts(seq, ",noacl");
T
Tsutomu Itoh 已提交
881 882
	if (btrfs_test_opt(root, SPACE_CACHE))
		seq_puts(seq, ",space_cache");
883
	else
884
		seq_puts(seq, ",nospace_cache");
T
Tsutomu Itoh 已提交
885 886 887 888
	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");
889 890 891 892 893 894
	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");
895 896
	if (btrfs_test_opt(root, SKIP_BALANCE))
		seq_puts(seq, ",skip_balance");
J
Jeff Mahoney 已提交
897 898
	if (btrfs_test_opt(root, PANIC_ON_FATAL_ERROR))
		seq_puts(seq, ",fatal_errors=panic");
E
Eric Paris 已提交
899 900 901
	return 0;
}

902
static int btrfs_test_super(struct super_block *s, void *data)
Y
Yan 已提交
903
{
904 905
	struct btrfs_fs_info *p = data;
	struct btrfs_fs_info *fs_info = btrfs_sb(s);
Y
Yan 已提交
906

907
	return fs_info->fs_devices == p->fs_devices;
Y
Yan 已提交
908 909
}

910 911
static int btrfs_set_super(struct super_block *s, void *data)
{
A
Al Viro 已提交
912 913 914 915
	int err = set_anon_super(s, data);
	if (!err)
		s->s_fs_info = data;
	return err;
Y
Yan 已提交
916 917
}

918 919 920 921 922 923 924 925 926 927
/*
 * 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;
}

928 929 930 931 932 933 934
/*
 * 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)
{
935 936
	unsigned len = strlen(args) + 2 + 1;
	char *src, *dst, *buf;
937 938

	/*
939 940
	 * We need the same args as before, but with this substitution:
	 * s!subvol=[^,]+!subvolid=0!
941
	 *
942 943
	 * Since the replacement string is up to 2 bytes longer than the
	 * original, allocate strlen(args) + 2 + 1 bytes.
944 945
	 */

946
	src = strstr(args, "subvol=");
947
	/* This shouldn't happen, but just in case.. */
948 949 950 951 952
	if (!src)
		return NULL;

	buf = dst = kmalloc(len, GFP_NOFS);
	if (!buf)
953 954 955
		return NULL;

	/*
956 957
	 * If the subvol= arg is not at the start of the string,
	 * copy whatever precedes it into buf.
958
	 */
959 960 961 962
	if (src != args) {
		*src++ = '\0';
		strcpy(buf, args);
		dst += strlen(args);
963 964
	}

965 966
	strcpy(dst, "subvolid=0");
	dst += strlen("subvolid=0");
967 968

	/*
969 970
	 * If there is a "," after the original subvol=... string,
	 * copy that suffix into our buffer.  Otherwise, we're done.
971
	 */
972 973 974
	src = strchr(src, ',');
	if (src)
		strcpy(dst, src);
975

976
	return buf;
977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994
}

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
995
	root = mount_subtree(mnt, subvol_name);
996

A
Al Viro 已提交
997 998 999 1000 1001
	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);
1002 1003 1004 1005
		printk(KERN_ERR "btrfs: '%s' is not a valid subvolume\n",
				subvol_name);
	}

1006 1007
	return root;
}
1008

1009 1010 1011 1012 1013 1014
/*
 * 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
1015
static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags,
1016
		const char *device_name, void *data)
Y
Yan 已提交
1017 1018 1019 1020
{
	struct block_device *bdev = NULL;
	struct super_block *s;
	struct dentry *root;
1021
	struct btrfs_fs_devices *fs_devices = NULL;
1022
	struct btrfs_fs_info *fs_info = NULL;
1023
	fmode_t mode = FMODE_READ;
1024 1025
	char *subvol_name = NULL;
	u64 subvol_objectid = 0;
1026
	u64 subvol_rootid = 0;
Y
Yan 已提交
1027 1028
	int error = 0;

1029 1030 1031 1032
	if (!(flags & MS_RDONLY))
		mode |= FMODE_WRITE;

	error = btrfs_parse_early_options(data, mode, fs_type,
1033
					  &subvol_name, &subvol_objectid,
1034
					  &subvol_rootid, &fs_devices);
1035 1036
	if (error) {
		kfree(subvol_name);
A
Al Viro 已提交
1037
		return ERR_PTR(error);
1038
	}
1039

1040 1041 1042 1043 1044 1045
	if (subvol_name) {
		root = mount_subvol(subvol_name, flags, device_name, data);
		kfree(subvol_name);
		return root;
	}

1046
	error = btrfs_scan_one_device(device_name, mode, fs_type, &fs_devices);
1047
	if (error)
1048
		return ERR_PTR(error);
Y
Yan 已提交
1049

1050 1051 1052 1053 1054 1055 1056
	/*
	 * 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);
1057 1058 1059
	if (!fs_info)
		return ERR_PTR(-ENOMEM);

1060 1061
	fs_info->fs_devices = fs_devices;

1062 1063 1064 1065
	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;
1066 1067 1068 1069 1070 1071 1072 1073 1074
		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;
1075 1076 1077
		goto error_close_devices;
	}

1078
	bdev = fs_devices->latest_bdev;
1079
	s = sget(fs_type, btrfs_test_super, btrfs_set_super, fs_info);
1080 1081 1082 1083
	if (IS_ERR(s)) {
		error = PTR_ERR(s);
		goto error_close_devices;
	}
Y
Yan 已提交
1084 1085

	if (s->s_root) {
Y
Yan Zheng 已提交
1086
		btrfs_close_devices(fs_devices);
1087
		free_fs_info(fs_info);
1088 1089
		if ((flags ^ s->s_flags) & MS_RDONLY)
			error = -EBUSY;
Y
Yan 已提交
1090 1091 1092
	} else {
		char b[BDEVNAME_SIZE];

A
Al Viro 已提交
1093
		s->s_flags = flags | MS_NOSEC;
Y
Yan 已提交
1094
		strlcpy(s->s_id, bdevname(bdev, b), sizeof(s->s_id));
1095
		btrfs_sb(s)->bdev_holder = fs_type;
1096 1097
		error = btrfs_fill_super(s, fs_devices, data,
					 flags & MS_SILENT ? 1 : 0);
Y
Yan 已提交
1098 1099
	}

1100 1101
	root = !error ? get_default_root(s, subvol_objectid) : ERR_PTR(error);
	if (IS_ERR(root))
1102
		deactivate_locked_super(s);
Y
Yan 已提交
1103

A
Al Viro 已提交
1104
	return root;
Y
Yan 已提交
1105

Y
Yan Zheng 已提交
1106
error_close_devices:
1107
	btrfs_close_devices(fs_devices);
1108
error_fs_info:
1109
	free_fs_info(fs_info);
A
Al Viro 已提交
1110
	return ERR_PTR(error);
Y
Yan 已提交
1111
}
1112

1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146
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
1147 1148
static int btrfs_remount(struct super_block *sb, int *flags, char *data)
{
1149 1150
	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
	struct btrfs_root *root = fs_info->tree_root;
1151 1152 1153 1154 1155 1156 1157
	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
1158 1159
	int ret;

1160
	ret = btrfs_parse_options(root, data);
1161 1162 1163 1164
	if (ret) {
		ret = -EINVAL;
		goto restore;
	}
1165

1166 1167 1168
	btrfs_resize_thread_pool(fs_info,
		fs_info->thread_pool_size, old_thread_pool_size);

Y
Yan Zheng 已提交
1169 1170 1171 1172 1173 1174
	if ((*flags & MS_RDONLY) == (sb->s_flags & MS_RDONLY))
		return 0;

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

1175 1176 1177
		ret = btrfs_commit_super(root);
		if (ret)
			goto restore;
Y
Yan Zheng 已提交
1178
	} else {
1179
		if (fs_info->fs_devices->rw_devices == 0) {
1180 1181
			ret = -EACCES;
			goto restore;
1182
		}
Y
Yan Zheng 已提交
1183

1184
		if (btrfs_super_log_root(fs_info->super_copy) != 0) {
1185 1186
			ret = -EINVAL;
			goto restore;
1187
		}
Y
Yan Zheng 已提交
1188

1189
		ret = btrfs_cleanup_fs_roots(fs_info);
1190 1191
		if (ret)
			goto restore;
Y
Yan Zheng 已提交
1192

1193 1194
		/* recover relocation */
		ret = btrfs_recover_relocation(root);
1195 1196
		if (ret)
			goto restore;
Y
Yan Zheng 已提交
1197

1198 1199 1200 1201
		ret = btrfs_resume_balance_async(fs_info);
		if (ret)
			goto restore;

Y
Yan Zheng 已提交
1202 1203 1204 1205
		sb->s_flags &= ~MS_RDONLY;
	}

	return 0;
1206 1207 1208 1209 1210 1211 1212 1213 1214 1215

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;
1216 1217
	btrfs_resize_thread_pool(fs_info,
		old_thread_pool_size, fs_info->thread_pool_size);
1218 1219
	fs_info->metadata_ratio = old_metadata_ratio;
	return ret;
Y
Yan Zheng 已提交
1220 1221
}

1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247
/* 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);
}

1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262
/*
 * 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;
1263
	int min_stripes = 1, num_stripes = 1;
1264 1265 1266
	int i = 0, nr_devices;
	int ret;

1267
	nr_devices = fs_info->fs_devices->open_devices;
1268 1269 1270 1271 1272 1273 1274 1275 1276
	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);
1277
	if (type & BTRFS_BLOCK_GROUP_RAID0) {
1278
		min_stripes = 2;
1279 1280
		num_stripes = nr_devices;
	} else if (type & BTRFS_BLOCK_GROUP_RAID1) {
1281
		min_stripes = 2;
1282 1283
		num_stripes = 2;
	} else if (type & BTRFS_BLOCK_GROUP_RAID10) {
1284
		min_stripes = 4;
1285 1286
		num_stripes = 4;
	}
1287 1288 1289 1290 1291 1292

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

1293 1294
	list_for_each_entry(device, &fs_devices->devices, dev_list) {
		if (!device->in_fs_metadata || !device->bdev)
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 1347 1348 1349 1350 1351 1352 1353 1354
			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) {
1355 1356 1357
		if (num_stripes > nr_devices)
			num_stripes = nr_devices;

1358 1359 1360 1361
		if (devices_info[i].max_avail >= min_stripe_size) {
			int j;
			u64 alloc_size;

1362
			avail_space += devices_info[i].max_avail * num_stripes;
1363
			alloc_size = devices_info[i].max_avail;
1364
			for (j = i + 1 - num_stripes; j <= i; j++)
1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375
				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
1376 1377
static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf)
{
1378 1379 1380
	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;
1381 1382
	struct btrfs_space_info *found;
	u64 total_used = 0;
1383
	u64 total_free_data = 0;
1384
	int bits = dentry->d_sb->s_blocksize_bits;
1385
	__be32 *fsid = (__be32 *)fs_info->fsid;
1386
	int ret;
C
Chris Mason 已提交
1387

1388
	/* holding chunk_muext to avoid allocating new chunks */
1389
	mutex_lock(&fs_info->chunk_mutex);
1390
	rcu_read_lock();
J
Josef Bacik 已提交
1391
	list_for_each_entry_rcu(found, head, list) {
1392 1393 1394 1395 1396 1397
		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);
		}

1398
		total_used += found->disk_used;
J
Josef Bacik 已提交
1399
	}
1400 1401
	rcu_read_unlock();

C
Chris Mason 已提交
1402
	buf->f_namelen = BTRFS_NAME_LEN;
1403
	buf->f_blocks = btrfs_super_total_bytes(disk_super) >> bits;
1404
	buf->f_bfree = buf->f_blocks - (total_used >> bits);
C
Chris Mason 已提交
1405 1406
	buf->f_bsize = dentry->d_sb->s_blocksize;
	buf->f_type = BTRFS_SUPER_MAGIC;
1407
	buf->f_bavail = total_free_data;
1408
	ret = btrfs_calc_avail_data_space(fs_info->tree_root, &total_free_data);
1409
	if (ret) {
1410
		mutex_unlock(&fs_info->chunk_mutex);
1411 1412 1413 1414
		return ret;
	}
	buf->f_bavail += total_free_data;
	buf->f_bavail = buf->f_bavail >> bits;
1415
	mutex_unlock(&fs_info->chunk_mutex);
C
Chris Mason 已提交
1416

1417
	/* We treat it as constant endianness (it doesn't matter _which_)
C
Chris Mason 已提交
1418
	   because we want the fsid to come out the same whether mounted
1419 1420 1421
	   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]);
1422 1423 1424 1425
	/* 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
1426 1427
	return 0;
}
C
Chris Mason 已提交
1428

A
Al Viro 已提交
1429 1430
static void btrfs_kill_super(struct super_block *sb)
{
1431
	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
A
Al Viro 已提交
1432
	kill_anon_super(sb);
1433
	free_fs_info(fs_info);
A
Al Viro 已提交
1434 1435
}

1436 1437 1438
static struct file_system_type btrfs_fs_type = {
	.owner		= THIS_MODULE,
	.name		= "btrfs",
A
Al Viro 已提交
1439
	.mount		= btrfs_mount,
A
Al Viro 已提交
1440
	.kill_sb	= btrfs_kill_super,
1441 1442
	.fs_flags	= FS_REQUIRES_DEV,
};
1443

C
Chris Mason 已提交
1444 1445 1446
/*
 * used by btrfsctl to scan devices when no FS is mounted
 */
1447 1448 1449 1450 1451
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;
1452
	int ret = -ENOTTY;
1453

1454 1455 1456
	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

L
Li Zefan 已提交
1457 1458 1459
	vol = memdup_user((void __user *)arg, sizeof(*vol));
	if (IS_ERR(vol))
		return PTR_ERR(vol);
1460

1461 1462
	switch (cmd) {
	case BTRFS_IOC_SCAN_DEV:
1463
		ret = btrfs_scan_one_device(vol->name, FMODE_READ,
1464 1465
					    &btrfs_fs_type, &fs_devices);
		break;
J
Josef Bacik 已提交
1466 1467 1468 1469 1470 1471 1472
	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;
1473
	}
L
Li Zefan 已提交
1474

1475
	kfree(vol);
L
Linda Knippers 已提交
1476
	return ret;
1477 1478
}

1479
static int btrfs_freeze(struct super_block *sb)
Y
Yan 已提交
1480
{
1481 1482 1483
	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
	mutex_lock(&fs_info->transaction_kthread_mutex);
	mutex_lock(&fs_info->cleaner_mutex);
1484
	return 0;
Y
Yan 已提交
1485 1486
}

1487
static int btrfs_unfreeze(struct super_block *sb)
Y
Yan 已提交
1488
{
1489 1490 1491
	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
	mutex_unlock(&fs_info->cleaner_mutex);
	mutex_unlock(&fs_info->transaction_kthread_mutex);
1492
	return 0;
Y
Yan 已提交
1493
}
1494

J
Josef Bacik 已提交
1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525
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;
}

1526
static const struct super_operations btrfs_super_ops = {
1527
	.drop_inode	= btrfs_drop_inode,
A
Al Viro 已提交
1528
	.evict_inode	= btrfs_evict_inode,
C
Chris Mason 已提交
1529
	.put_super	= btrfs_put_super,
1530
	.sync_fs	= btrfs_sync_fs,
E
Eric Paris 已提交
1531
	.show_options	= btrfs_show_options,
J
Josef Bacik 已提交
1532
	.show_devname	= btrfs_show_devname,
C
Chris Mason 已提交
1533
	.write_inode	= btrfs_write_inode,
C
Chris Mason 已提交
1534 1535
	.alloc_inode	= btrfs_alloc_inode,
	.destroy_inode	= btrfs_destroy_inode,
C
Chris Mason 已提交
1536
	.statfs		= btrfs_statfs,
Y
Yan Zheng 已提交
1537
	.remount_fs	= btrfs_remount,
1538 1539
	.freeze_fs	= btrfs_freeze,
	.unfreeze_fs	= btrfs_unfreeze,
C
Chris Mason 已提交
1540
};
1541 1542 1543 1544 1545

static const struct file_operations btrfs_ctl_fops = {
	.unlocked_ioctl	 = btrfs_control_ioctl,
	.compat_ioctl = btrfs_control_ioctl,
	.owner	 = THIS_MODULE,
1546
	.llseek = noop_llseek,
1547 1548 1549
};

static struct miscdevice btrfs_misc = {
1550
	.minor		= BTRFS_MINOR,
1551 1552 1553 1554
	.name		= "btrfs-control",
	.fops		= &btrfs_ctl_fops
};

1555 1556 1557
MODULE_ALIAS_MISCDEV(BTRFS_MINOR);
MODULE_ALIAS("devname:btrfs-control");

1558 1559 1560 1561 1562
static int btrfs_interface_init(void)
{
	return misc_register(&btrfs_misc);
}

1563
static void btrfs_interface_exit(void)
1564 1565
{
	if (misc_deregister(&btrfs_misc) < 0)
C
Chris Mason 已提交
1566
		printk(KERN_INFO "misc_deregister failed for control device");
1567 1568
}

1569 1570
static int __init init_btrfs_fs(void)
{
C
Chris Mason 已提交
1571
	int err;
1572 1573 1574 1575 1576

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

1577
	btrfs_init_compress();
1578

1579 1580 1581 1582
	err = btrfs_init_cachep();
	if (err)
		goto free_compress;

1583
	err = extent_io_init();
1584 1585 1586
	if (err)
		goto free_cachep;

1587 1588 1589 1590
	err = extent_map_init();
	if (err)
		goto free_extent_io;

1591
	err = btrfs_delayed_inode_init();
1592 1593
	if (err)
		goto free_extent_map;
C
Chris Mason 已提交
1594

1595 1596 1597 1598
	err = btrfs_interface_init();
	if (err)
		goto free_delayed_inode;

1599 1600 1601
	err = register_filesystem(&btrfs_fs_type);
	if (err)
		goto unregister_ioctl;
1602

1603 1604
	btrfs_init_lockdep();

1605
	printk(KERN_INFO "%s loaded\n", BTRFS_BUILD_VERSION);
1606 1607
	return 0;

1608 1609
unregister_ioctl:
	btrfs_interface_exit();
1610 1611
free_delayed_inode:
	btrfs_delayed_inode_exit();
1612 1613
free_extent_map:
	extent_map_exit();
1614 1615
free_extent_io:
	extent_io_exit();
1616 1617
free_cachep:
	btrfs_destroy_cachep();
1618 1619
free_compress:
	btrfs_exit_compress();
1620 1621
	btrfs_exit_sysfs();
	return err;
1622 1623 1624 1625
}

static void __exit exit_btrfs_fs(void)
{
C
Chris Mason 已提交
1626
	btrfs_destroy_cachep();
1627
	btrfs_delayed_inode_exit();
1628
	extent_map_exit();
1629
	extent_io_exit();
1630
	btrfs_interface_exit();
1631
	unregister_filesystem(&btrfs_fs_type);
1632
	btrfs_exit_sysfs();
1633
	btrfs_cleanup_fs_uuids();
1634
	btrfs_exit_compress();
1635 1636 1637 1638 1639 1640
}

module_init(init_btrfs_fs)
module_exit(exit_btrfs_fs)

MODULE_LICENSE("GPL");