2 * Copyright (C) 2007 Oracle. All rights reserved.
4 * This program is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU General Public
6 * License v2 as published by the Free Software Foundation.
8 * This program is distributed in the hope that it will be useful,
9 * but WITHOUT ANY WARRANTY; without even the implied warranty of
10 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
11 * General Public License for more details.
13 * You should have received a copy of the GNU General Public
14 * License along with this program; if not, write to the
15 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
16 * Boston, MA 021110-1307, USA.
19 #include <linux/blkdev.h>
20 #include <linux/module.h>
21 #include <linux/buffer_head.h>
23 #include <linux/pagemap.h>
24 #include <linux/highmem.h>
25 #include <linux/time.h>
26 #include <linux/init.h>
27 #include <linux/string.h>
28 #include <linux/smp_lock.h>
29 #include <linux/backing-dev.h>
30 #include <linux/mount.h>
31 #include <linux/mpage.h>
32 #include <linux/swap.h>
33 #include <linux/writeback.h>
34 #include <linux/statfs.h>
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>
42 #include "transaction.h"
43 #include "btrfs_inode.h"
45 #include "print-tree.h"
51 #define BTRFS_SUPER_MAGIC 0x9123683E
53 static struct super_operations btrfs_super_ops
;
55 static void btrfs_put_super (struct super_block
* sb
)
57 struct btrfs_root
*root
= btrfs_sb(sb
);
58 struct btrfs_fs_info
*fs
= root
->fs_info
;
61 ret
= close_ctree(root
);
63 printk("close ctree returns %d\n", ret
);
65 btrfs_sysfs_del_super(fs
);
70 Opt_degraded
, Opt_subvol
, Opt_device
, Opt_nodatasum
, Opt_nodatacow
,
71 Opt_max_extent
, Opt_max_inline
, Opt_alloc_start
, Opt_nobarrier
,
72 Opt_ssd
, Opt_thread_pool
, Opt_noacl
, Opt_err
,
75 static match_table_t tokens
= {
76 {Opt_degraded
, "degraded"},
77 {Opt_subvol
, "subvol=%s"},
78 {Opt_device
, "device=%s"},
79 {Opt_nodatasum
, "nodatasum"},
80 {Opt_nodatacow
, "nodatacow"},
81 {Opt_nobarrier
, "nobarrier"},
82 {Opt_max_extent
, "max_extent=%s"},
83 {Opt_max_inline
, "max_inline=%s"},
84 {Opt_alloc_start
, "alloc_start=%s"},
85 {Opt_thread_pool
, "thread_pool=%d"},
91 u64
btrfs_parse_size(char *str
)
98 res
= simple_strtoul(str
, &end
, 10);
102 last
= tolower(last
);
117 * Regular mount options parser. Everything that is needed only when
118 * reading in a new superblock is parsed here.
120 int btrfs_parse_options(struct btrfs_root
*root
, char *options
)
122 struct btrfs_fs_info
*info
= root
->fs_info
;
123 substring_t args
[MAX_OPT_ARGS
];
131 * strsep changes the string, duplicate it because parse_options
134 options
= kstrdup(options
, GFP_NOFS
);
139 while ((p
= strsep(&options
, ",")) != NULL
) {
144 token
= match_token(p
, tokens
, args
);
147 printk(KERN_INFO
"btrfs: allowing degraded mounts\n");
148 btrfs_set_opt(info
->mount_opt
, DEGRADED
);
153 * These are parsed by btrfs_parse_early_options
154 * and can be happily ignored here.
158 printk(KERN_INFO
"btrfs: setting nodatacsum\n");
159 btrfs_set_opt(info
->mount_opt
, NODATASUM
);
162 printk(KERN_INFO
"btrfs: setting nodatacow\n");
163 btrfs_set_opt(info
->mount_opt
, NODATACOW
);
164 btrfs_set_opt(info
->mount_opt
, NODATASUM
);
167 printk(KERN_INFO
"btrfs: use ssd allocation scheme\n");
168 btrfs_set_opt(info
->mount_opt
, SSD
);
171 printk(KERN_INFO
"btrfs: turning off barriers\n");
172 btrfs_set_opt(info
->mount_opt
, NOBARRIER
);
174 case Opt_thread_pool
:
176 match_int(&args
[0], &intarg
);
178 info
->thread_pool_size
= intarg
;
179 printk(KERN_INFO
"btrfs: thread pool %d\n",
180 info
->thread_pool_size
);
184 num
= match_strdup(&args
[0]);
186 info
->max_extent
= btrfs_parse_size(num
);
189 info
->max_extent
= max_t(u64
,
190 info
->max_extent
, root
->sectorsize
);
191 printk(KERN_INFO
"btrfs: max_extent at %llu\n",
196 num
= match_strdup(&args
[0]);
198 info
->max_inline
= btrfs_parse_size(num
);
201 if (info
->max_inline
) {
202 info
->max_inline
= max_t(u64
,
206 printk(KERN_INFO
"btrfs: max_inline at %llu\n",
210 case Opt_alloc_start
:
211 num
= match_strdup(&args
[0]);
213 info
->alloc_start
= btrfs_parse_size(num
);
216 "btrfs: allocations start at %llu\n",
221 root
->fs_info
->sb
->s_flags
&= ~MS_POSIXACL
;
232 * Parse mount options that are required early in the mount process.
234 * All other options will be parsed on much later in the mount process and
235 * only when we need to allocate a new super block.
237 static int btrfs_parse_early_options(const char *options
, int flags
,
238 void *holder
, char **subvol_name
,
239 struct btrfs_fs_devices
**fs_devices
)
241 substring_t args
[MAX_OPT_ARGS
];
249 * strsep changes the string, duplicate it because parse_options
252 opts
= kstrdup(options
, GFP_KERNEL
);
256 while ((p
= strsep(&opts
, ",")) != NULL
) {
261 token
= match_token(p
, tokens
, args
);
264 *subvol_name
= match_strdup(&args
[0]);
267 error
= btrfs_scan_one_device(match_strdup(&args
[0]),
268 flags
, holder
, fs_devices
);
281 * If no subvolume name is specified we use the default one. Allocate
282 * a copy of the string "default" here so that code later in the
283 * mount path doesn't care if it's the default volume or another one.
286 *subvol_name
= kstrdup("default", GFP_KERNEL
);
293 static int btrfs_fill_super(struct super_block
* sb
,
294 struct btrfs_fs_devices
*fs_devices
,
295 void * data
, int silent
)
297 struct inode
* inode
;
298 struct dentry
* root_dentry
;
299 struct btrfs_super_block
*disk_super
;
300 struct btrfs_root
*tree_root
;
301 struct btrfs_inode
*bi
;
304 sb
->s_maxbytes
= MAX_LFS_FILESIZE
;
305 sb
->s_magic
= BTRFS_SUPER_MAGIC
;
306 sb
->s_op
= &btrfs_super_ops
;
307 sb
->s_export_op
= &btrfs_export_ops
;
308 sb
->s_xattr
= btrfs_xattr_handlers
;
310 sb
->s_flags
|= MS_POSIXACL
;
312 tree_root
= open_ctree(sb
, fs_devices
, (char *)data
);
314 if (IS_ERR(tree_root
)) {
315 printk("btrfs: open_ctree failed\n");
316 return PTR_ERR(tree_root
);
318 sb
->s_fs_info
= tree_root
;
319 disk_super
= &tree_root
->fs_info
->super_copy
;
320 inode
= btrfs_iget_locked(sb
, btrfs_super_root_dir(disk_super
),
323 bi
->location
.objectid
= inode
->i_ino
;
324 bi
->location
.offset
= 0;
325 bi
->root
= tree_root
;
327 btrfs_set_key_type(&bi
->location
, BTRFS_INODE_ITEM_KEY
);
333 if (inode
->i_state
& I_NEW
) {
334 btrfs_read_locked_inode(inode
);
335 unlock_new_inode(inode
);
338 root_dentry
= d_alloc_root(inode
);
345 /* this does the super kobj at the same time */
346 err
= btrfs_sysfs_add_super(tree_root
->fs_info
);
350 sb
->s_root
= root_dentry
;
352 save_mount_options(sb
, data
);
356 close_ctree(tree_root
);
360 int btrfs_sync_fs(struct super_block
*sb
, int wait
)
362 struct btrfs_trans_handle
*trans
;
363 struct btrfs_root
*root
;
369 filemap_flush(root
->fs_info
->btree_inode
->i_mapping
);
372 btrfs_clean_old_snapshots(root
);
373 trans
= btrfs_start_transaction(root
, 1);
374 ret
= btrfs_commit_transaction(trans
, root
);
379 static void btrfs_write_super(struct super_block
*sb
)
384 static int btrfs_test_super(struct super_block
*s
, void *data
)
386 struct btrfs_fs_devices
*test_fs_devices
= data
;
387 struct btrfs_root
*root
= btrfs_sb(s
);
389 return root
->fs_info
->fs_devices
== test_fs_devices
;
393 * Find a superblock for the given device / mount point.
395 * Note: This is based on get_sb_bdev from fs/super.c with a few additions
396 * for multiple device setup. Make sure to keep it in sync.
398 static int btrfs_get_sb(struct file_system_type
*fs_type
, int flags
,
399 const char *dev_name
, void *data
, struct vfsmount
*mnt
)
401 char *subvol_name
= NULL
;
402 struct block_device
*bdev
= NULL
;
403 struct super_block
*s
;
405 struct btrfs_fs_devices
*fs_devices
= NULL
;
408 error
= btrfs_parse_early_options(data
, flags
, fs_type
,
409 &subvol_name
, &fs_devices
);
413 error
= btrfs_scan_one_device(dev_name
, flags
, fs_type
, &fs_devices
);
415 goto error_free_subvol_name
;
417 error
= btrfs_open_devices(fs_devices
, flags
, fs_type
);
419 goto error_free_subvol_name
;
421 bdev
= fs_devices
->latest_bdev
;
422 s
= sget(fs_type
, btrfs_test_super
, set_anon_super
, fs_devices
);
427 if ((flags
^ s
->s_flags
) & MS_RDONLY
) {
428 up_write(&s
->s_umount
);
435 char b
[BDEVNAME_SIZE
];
438 strlcpy(s
->s_id
, bdevname(bdev
, b
), sizeof(s
->s_id
));
439 error
= btrfs_fill_super(s
, fs_devices
, data
,
440 flags
& MS_SILENT
? 1 : 0);
442 up_write(&s
->s_umount
);
447 btrfs_sb(s
)->fs_info
->bdev_holder
= fs_type
;
448 s
->s_flags
|= MS_ACTIVE
;
451 if (!strcmp(subvol_name
, "."))
452 root
= dget(s
->s_root
);
454 mutex_lock(&s
->s_root
->d_inode
->i_mutex
);
455 root
= lookup_one_len(subvol_name
, s
->s_root
, strlen(subvol_name
));
456 mutex_unlock(&s
->s_root
->d_inode
->i_mutex
);
458 up_write(&s
->s_umount
);
460 error
= PTR_ERR(root
);
463 if (!root
->d_inode
) {
465 up_write(&s
->s_umount
);
473 mnt
->mnt_root
= root
;
481 btrfs_close_devices(fs_devices
);
482 error_free_subvol_name
:
488 static int btrfs_statfs(struct dentry
*dentry
, struct kstatfs
*buf
)
490 struct btrfs_root
*root
= btrfs_sb(dentry
->d_sb
);
491 struct btrfs_super_block
*disk_super
= &root
->fs_info
->super_copy
;
492 int bits
= dentry
->d_sb
->s_blocksize_bits
;
493 __be32
*fsid
= (__be32
*)root
->fs_info
->fsid
;
495 buf
->f_namelen
= BTRFS_NAME_LEN
;
496 buf
->f_blocks
= btrfs_super_total_bytes(disk_super
) >> bits
;
497 buf
->f_bfree
= buf
->f_blocks
-
498 (btrfs_super_bytes_used(disk_super
) >> bits
);
499 buf
->f_bavail
= buf
->f_bfree
;
500 buf
->f_bsize
= dentry
->d_sb
->s_blocksize
;
501 buf
->f_type
= BTRFS_SUPER_MAGIC
;
502 /* We treat it as constant endianness (it doesn't matter _which_)
503 because we want the fsid to come out the same whether mounted
504 on a big-endian or little-endian host */
505 buf
->f_fsid
.val
[0] = be32_to_cpu(fsid
[0]) ^ be32_to_cpu(fsid
[2]);
506 buf
->f_fsid
.val
[1] = be32_to_cpu(fsid
[1]) ^ be32_to_cpu(fsid
[3]);
507 /* Mask in the root object ID too, to disambiguate subvols */
508 buf
->f_fsid
.val
[0] ^= BTRFS_I(dentry
->d_inode
)->root
->objectid
>> 32;
509 buf
->f_fsid
.val
[1] ^= BTRFS_I(dentry
->d_inode
)->root
->objectid
;
514 static struct file_system_type btrfs_fs_type
= {
515 .owner
= THIS_MODULE
,
517 .get_sb
= btrfs_get_sb
,
518 .kill_sb
= kill_anon_super
,
519 .fs_flags
= FS_REQUIRES_DEV
,
523 * used by btrfsctl to scan devices when no FS is mounted
525 static long btrfs_control_ioctl(struct file
*file
, unsigned int cmd
,
528 struct btrfs_ioctl_vol_args
*vol
;
529 struct btrfs_fs_devices
*fs_devices
;
533 vol
= kmalloc(sizeof(*vol
), GFP_KERNEL
);
534 if (copy_from_user(vol
, (void __user
*)arg
, sizeof(*vol
))) {
538 len
= strnlen(vol
->name
, BTRFS_PATH_NAME_MAX
);
540 case BTRFS_IOC_SCAN_DEV
:
541 ret
= btrfs_scan_one_device(vol
->name
, MS_RDONLY
,
542 &btrfs_fs_type
, &fs_devices
);
550 static void btrfs_write_super_lockfs(struct super_block
*sb
)
552 struct btrfs_root
*root
= btrfs_sb(sb
);
553 mutex_lock(&root
->fs_info
->transaction_kthread_mutex
);
554 mutex_lock(&root
->fs_info
->cleaner_mutex
);
557 static void btrfs_unlockfs(struct super_block
*sb
)
559 struct btrfs_root
*root
= btrfs_sb(sb
);
560 mutex_unlock(&root
->fs_info
->cleaner_mutex
);
561 mutex_unlock(&root
->fs_info
->transaction_kthread_mutex
);
564 static struct super_operations btrfs_super_ops
= {
565 .delete_inode
= btrfs_delete_inode
,
566 .put_super
= btrfs_put_super
,
567 .write_super
= btrfs_write_super
,
568 .sync_fs
= btrfs_sync_fs
,
569 .show_options
= generic_show_options
,
570 .write_inode
= btrfs_write_inode
,
571 .dirty_inode
= btrfs_dirty_inode
,
572 .alloc_inode
= btrfs_alloc_inode
,
573 .destroy_inode
= btrfs_destroy_inode
,
574 .statfs
= btrfs_statfs
,
575 .write_super_lockfs
= btrfs_write_super_lockfs
,
576 .unlockfs
= btrfs_unlockfs
,
579 static const struct file_operations btrfs_ctl_fops
= {
580 .unlocked_ioctl
= btrfs_control_ioctl
,
581 .compat_ioctl
= btrfs_control_ioctl
,
582 .owner
= THIS_MODULE
,
585 static struct miscdevice btrfs_misc
= {
586 .minor
= MISC_DYNAMIC_MINOR
,
587 .name
= "btrfs-control",
588 .fops
= &btrfs_ctl_fops
591 static int btrfs_interface_init(void)
593 return misc_register(&btrfs_misc
);
596 void btrfs_interface_exit(void)
598 if (misc_deregister(&btrfs_misc
) < 0)
599 printk("misc_deregister failed for control device");
602 static int __init
init_btrfs_fs(void)
606 err
= btrfs_init_sysfs();
610 err
= btrfs_init_cachep();
614 err
= extent_io_init();
618 err
= extent_map_init();
622 err
= btrfs_interface_init();
624 goto free_extent_map
;
625 err
= register_filesystem(&btrfs_fs_type
);
627 goto unregister_ioctl
;
629 printk(KERN_INFO
"%s loaded\n", BTRFS_BUILD_VERSION
);
633 btrfs_interface_exit();
639 btrfs_destroy_cachep();
645 static void __exit
exit_btrfs_fs(void)
647 btrfs_destroy_cachep();
650 btrfs_interface_exit();
651 unregister_filesystem(&btrfs_fs_type
);
653 btrfs_cleanup_fs_uuids();
656 module_init(init_btrfs_fs
)
657 module_exit(exit_btrfs_fs
)
659 MODULE_LICENSE("GPL");