[mirror_ubuntu-eoan-kernel.git] / fs / btrfs / super.c

/*
 * 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.
 */

#include <linux/blkdev.h>
#include <linux/module.h>
#include <linux/buffer_head.h>
#include <linux/fs.h>
#include <linux/pagemap.h>
#include <linux/highmem.h>
#include <linux/time.h>
#include <linux/init.h>
#include <linux/string.h>
#include <linux/smp_lock.h>
#include <linux/backing-dev.h>
#include <linux/mount.h>
#include <linux/mpage.h>
#include <linux/swap.h>
#include <linux/writeback.h>
#include <linux/statfs.h>
#include <linux/compat.h>
#include <linux/parser.h>
#include <linux/ctype.h>
#include <linux/namei.h>
#include <linux/miscdevice.h>
#include "ctree.h"
#include "disk-io.h"
#include "transaction.h"
#include "btrfs_inode.h"
#include "ioctl.h"
#include "print-tree.h"
#include "xattr.h"
#include "volumes.h"
#include "version.h"
#include "export.h"

#define BTRFS_SUPER_MAGIC 0x9123683E

static struct super_operations btrfs_super_ops;

static void btrfs_put_super (struct super_block * sb)
{
	struct btrfs_root *root = btrfs_sb(sb);
	struct btrfs_fs_info *fs = root->fs_info;
	int ret;

	ret = close_ctree(root);
	if (ret) {
		printk("close ctree returns %d\n", ret);
	}
	btrfs_sysfs_del_super(fs);
	sb->s_fs_info = NULL;
}

enum {
	Opt_degraded, Opt_subvol, Opt_device, Opt_nodatasum, Opt_nodatacow,
	Opt_max_extent, Opt_max_inline, Opt_alloc_start, Opt_nobarrier,
	Opt_ssd, Opt_thread_pool, Opt_noacl,  Opt_err,
};

static match_table_t tokens = {
	{Opt_degraded, "degraded"},
	{Opt_subvol, "subvol=%s"},
	{Opt_device, "device=%s"},
	{Opt_nodatasum, "nodatasum"},
	{Opt_nodatacow, "nodatacow"},
	{Opt_nobarrier, "nobarrier"},
	{Opt_max_extent, "max_extent=%s"},
	{Opt_max_inline, "max_inline=%s"},
	{Opt_alloc_start, "alloc_start=%s"},
	{Opt_thread_pool, "thread_pool=%d"},
	{Opt_ssd, "ssd"},
	{Opt_noacl, "noacl"},
	{Opt_err, NULL},
};

u64 btrfs_parse_size(char *str)
{
	u64 res;
	int mult = 1;
	char *end;
	char last;

	res = simple_strtoul(str, &end, 10);

	last = end[0];
	if (isalpha(last)) {
		last = tolower(last);
		switch (last) {
		case 'g':
			mult *= 1024;
		case 'm':
			mult *= 1024;
		case 'k':
			mult *= 1024;
		}
		res = res * mult;
	}
	return res;
}

/*
 * Regular mount options parser.  Everything that is needed only when
 * reading in a new superblock is parsed here.
 */
int btrfs_parse_options(struct btrfs_root *root, char *options)
{
	struct btrfs_fs_info *info = root->fs_info;
	substring_t args[MAX_OPT_ARGS];
	char *p, *num;
	int intarg;

	if (!options)
		return 0;

	/*
	 * strsep changes the string, duplicate it because parse_options
	 * gets called twice
	 */
	options = kstrdup(options, GFP_NOFS);
	if (!options)
		return -ENOMEM;


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

		token = match_token(p, tokens, args);
		switch (token) {
		case Opt_degraded:
			printk(KERN_INFO "btrfs: allowing degraded mounts\n");
			btrfs_set_opt(info->mount_opt, DEGRADED);
			break;
		case Opt_subvol:
		case Opt_device:
			/*
			 * These are parsed by btrfs_parse_early_options
			 * and can be happily ignored here.
			 */
			break;
		case Opt_nodatasum:
			printk(KERN_INFO "btrfs: setting nodatacsum\n");
			btrfs_set_opt(info->mount_opt, NODATASUM);
			break;
		case Opt_nodatacow:
			printk(KERN_INFO "btrfs: setting nodatacow\n");
			btrfs_set_opt(info->mount_opt, NODATACOW);
			btrfs_set_opt(info->mount_opt, NODATASUM);
			break;
		case Opt_ssd:
			printk(KERN_INFO "btrfs: use ssd allocation scheme\n");
			btrfs_set_opt(info->mount_opt, SSD);
			break;
		case Opt_nobarrier:
			printk(KERN_INFO "btrfs: turning off barriers\n");
			btrfs_set_opt(info->mount_opt, NOBARRIER);
			break;
		case Opt_thread_pool:
			intarg = 0;
			match_int(&args[0], &intarg);
			if (intarg) {
				info->thread_pool_size = intarg;
				printk(KERN_INFO "btrfs: thread pool %d\n",
				       info->thread_pool_size);
			}
			break;
		case Opt_max_extent:
			num = match_strdup(&args[0]);
			if (num) {
				info->max_extent = btrfs_parse_size(num);
				kfree(num);

				info->max_extent = max_t(u64,
					info->max_extent, root->sectorsize);
				printk(KERN_INFO "btrfs: max_extent at %llu\n",
				       info->max_extent);
			}
			break;
		case Opt_max_inline:
			num = match_strdup(&args[0]);
			if (num) {
				info->max_inline = btrfs_parse_size(num);
				kfree(num);

				if (info->max_inline) {
					info->max_inline = max_t(u64,
						info->max_inline,
						root->sectorsize);
				}
				printk(KERN_INFO "btrfs: max_inline at %llu\n",
					info->max_inline);
			}
			break;
		case Opt_alloc_start:
			num = match_strdup(&args[0]);
			if (num) {
				info->alloc_start = btrfs_parse_size(num);
				kfree(num);
				printk(KERN_INFO
					"btrfs: allocations start at %llu\n",
					info->alloc_start);
			}
			break;
		case Opt_noacl:
			root->fs_info->sb->s_flags &= ~MS_POSIXACL;
			break;
		default:
			break;
		}
	}
	kfree(options);
	return 0;
}

/*
 * 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.
 */
static int btrfs_parse_early_options(const char *options, int flags,
		void *holder, char **subvol_name,
		struct btrfs_fs_devices **fs_devices)
{
	substring_t args[MAX_OPT_ARGS];
	char *opts, *p;
	int error = 0;

	if (!options)
		goto out;

	/*
	 * strsep changes the string, duplicate it because parse_options
	 * gets called twice
	 */
	opts = kstrdup(options, GFP_KERNEL);
	if (!opts)
		return -ENOMEM;

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

		token = match_token(p, tokens, args);
		switch (token) {
		case Opt_subvol:
			*subvol_name = match_strdup(&args[0]);
			break;
		case Opt_device:
			error = btrfs_scan_one_device(match_strdup(&args[0]),
					flags, holder, fs_devices);
			if (error)
				goto out_free_opts;
			break;
		default:
			break;
		}
	}

 out_free_opts:
	kfree(opts);
 out:
	/*
	 * If no subvolume name is specified we use the default one.  Allocate
	 * a copy of the string "default" here so that code later in the
	 * mount path doesn't care if it's the default volume or another one.
	 */
	if (!*subvol_name) {
		*subvol_name = kstrdup("default", GFP_KERNEL);
		if (!*subvol_name)
			return -ENOMEM;
	}
	return error;
}

static int btrfs_fill_super(struct super_block * sb,
			    struct btrfs_fs_devices *fs_devices,
			    void * data, int silent)
{
	struct inode * inode;
	struct dentry * root_dentry;
	struct btrfs_super_block *disk_super;
	struct btrfs_root *tree_root;
	struct btrfs_inode *bi;
	int err;

	sb->s_maxbytes = MAX_LFS_FILESIZE;
	sb->s_magic = BTRFS_SUPER_MAGIC;
	sb->s_op = &btrfs_super_ops;
	sb->s_export_op = &btrfs_export_ops;
	sb->s_xattr = btrfs_xattr_handlers;
	sb->s_time_gran = 1;
	sb->s_flags |= MS_POSIXACL;

	tree_root = open_ctree(sb, fs_devices, (char *)data);

	if (IS_ERR(tree_root)) {
		printk("btrfs: open_ctree failed\n");
		return PTR_ERR(tree_root);
	}
	sb->s_fs_info = tree_root;
	disk_super = &tree_root->fs_info->super_copy;
	inode = btrfs_iget_locked(sb, btrfs_super_root_dir(disk_super),
				  tree_root);
	bi = BTRFS_I(inode);
	bi->location.objectid = inode->i_ino;
	bi->location.offset = 0;
	bi->root = tree_root;

	btrfs_set_key_type(&bi->location, BTRFS_INODE_ITEM_KEY);

	if (!inode) {
		err = -ENOMEM;
		goto fail_close;
	}
	if (inode->i_state & I_NEW) {
		btrfs_read_locked_inode(inode);
		unlock_new_inode(inode);
	}

	root_dentry = d_alloc_root(inode);
	if (!root_dentry) {
		iput(inode);
		err = -ENOMEM;
		goto fail_close;
	}

	/* this does the super kobj at the same time */
	err = btrfs_sysfs_add_super(tree_root->fs_info);
	if (err)
		goto fail_close;

	sb->s_root = root_dentry;

	save_mount_options(sb, data);
	return 0;

fail_close:
	close_ctree(tree_root);
	return err;
}

int btrfs_sync_fs(struct super_block *sb, int wait)
{
	struct btrfs_trans_handle *trans;
	struct btrfs_root *root;
	int ret;
	root = btrfs_sb(sb);

	sb->s_dirt = 0;
	if (!wait) {
		filemap_flush(root->fs_info->btree_inode->i_mapping);
		return 0;
	}
	btrfs_clean_old_snapshots(root);
	trans = btrfs_start_transaction(root, 1);
	ret = btrfs_commit_transaction(trans, root);
	sb->s_dirt = 0;
	return ret;
}

static void btrfs_write_super(struct super_block *sb)
{
	sb->s_dirt = 0;
}

static int btrfs_test_super(struct super_block *s, void *data)
{
	struct btrfs_fs_devices *test_fs_devices = data;
	struct btrfs_root *root = btrfs_sb(s);

	return root->fs_info->fs_devices == test_fs_devices;
}

/*
 * 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.
 */
static int btrfs_get_sb(struct file_system_type *fs_type, int flags,
		const char *dev_name, void *data, struct vfsmount *mnt)
{
	char *subvol_name = NULL;
	struct block_device *bdev = NULL;
	struct super_block *s;
	struct dentry *root;
	struct btrfs_fs_devices *fs_devices = NULL;
	int error = 0;

	error = btrfs_parse_early_options(data, flags, fs_type,
					  &subvol_name, &fs_devices);
	if (error)
		goto error;

	error = btrfs_scan_one_device(dev_name, flags, fs_type, &fs_devices);
	if (error)
		goto error_free_subvol_name;

	error = btrfs_open_devices(fs_devices, flags, fs_type);
	if (error)
		goto error_free_subvol_name;

	bdev = fs_devices->latest_bdev;
	s = sget(fs_type, btrfs_test_super, set_anon_super, fs_devices);
	if (IS_ERR(s))
		goto error_s;

	if (s->s_root) {
		if ((flags ^ s->s_flags) & MS_RDONLY) {
			up_write(&s->s_umount);
			deactivate_super(s);
			error = -EBUSY;
			goto error_bdev;
		}

	} else {
		char b[BDEVNAME_SIZE];

		s->s_flags = flags;
		strlcpy(s->s_id, bdevname(bdev, b), sizeof(s->s_id));
		error = btrfs_fill_super(s, fs_devices, data,
					 flags & MS_SILENT ? 1 : 0);
		if (error) {
			up_write(&s->s_umount);
			deactivate_super(s);
			goto error;
		}

		btrfs_sb(s)->fs_info->bdev_holder = fs_type;
		s->s_flags |= MS_ACTIVE;
	}

	if (!strcmp(subvol_name, "."))
		root = dget(s->s_root);
	else {
		mutex_lock(&s->s_root->d_inode->i_mutex);
		root = lookup_one_len(subvol_name, s->s_root, strlen(subvol_name));
		mutex_unlock(&s->s_root->d_inode->i_mutex);
		if (IS_ERR(root)) {
			up_write(&s->s_umount);
			deactivate_super(s);
			error = PTR_ERR(root);
			goto error;
		}
		if (!root->d_inode) {
			dput(root);
			up_write(&s->s_umount);
			deactivate_super(s);
			error = -ENXIO;
			goto error;
		}
	}

	mnt->mnt_sb = s;
	mnt->mnt_root = root;

	kfree(subvol_name);
	return 0;

error_s:
	error = PTR_ERR(s);
error_bdev:
	btrfs_close_devices(fs_devices);
error_free_subvol_name:
	kfree(subvol_name);
error:
	return error;
}

static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf)
{
	struct btrfs_root *root = btrfs_sb(dentry->d_sb);
	struct btrfs_super_block *disk_super = &root->fs_info->super_copy;
	int bits = dentry->d_sb->s_blocksize_bits;
	__be32 *fsid = (__be32 *)root->fs_info->fsid;

	buf->f_namelen = BTRFS_NAME_LEN;
	buf->f_blocks = btrfs_super_total_bytes(disk_super) >> bits;
	buf->f_bfree = buf->f_blocks -
		(btrfs_super_bytes_used(disk_super) >> bits);
	buf->f_bavail = buf->f_bfree;
	buf->f_bsize = dentry->d_sb->s_blocksize;
	buf->f_type = BTRFS_SUPER_MAGIC;
	/* We treat it as constant endianness (it doesn't matter _which_)
	   because we want the fsid to come out the same whether mounted 
	   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]);
	/* 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;

	return 0;
}

static struct file_system_type btrfs_fs_type = {
	.owner		= THIS_MODULE,
	.name		= "btrfs",
	.get_sb		= btrfs_get_sb,
	.kill_sb	= kill_anon_super,
	.fs_flags	= FS_REQUIRES_DEV,
};

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;
	int ret = 0;
	int len;

	vol = kmalloc(sizeof(*vol), GFP_KERNEL);
	if (copy_from_user(vol, (void __user *)arg, sizeof(*vol))) {
		ret = -EFAULT;
		goto out;
	}
	len = strnlen(vol->name, BTRFS_PATH_NAME_MAX);
	switch (cmd) {
	case BTRFS_IOC_SCAN_DEV:
		ret = btrfs_scan_one_device(vol->name, MS_RDONLY,
					    &btrfs_fs_type, &fs_devices);
		break;
	}
out:
	kfree(vol);
	return ret;
}

static void btrfs_write_super_lockfs(struct super_block *sb)
{
	struct btrfs_root *root = btrfs_sb(sb);
	mutex_lock(&root->fs_info->transaction_kthread_mutex);
	mutex_lock(&root->fs_info->cleaner_mutex);
}

static void btrfs_unlockfs(struct super_block *sb)
{
	struct btrfs_root *root = btrfs_sb(sb);
	mutex_unlock(&root->fs_info->cleaner_mutex);
	mutex_unlock(&root->fs_info->transaction_kthread_mutex);
}

static struct super_operations btrfs_super_ops = {
	.delete_inode	= btrfs_delete_inode,
	.put_super	= btrfs_put_super,
	.write_super	= btrfs_write_super,
	.sync_fs	= btrfs_sync_fs,
	.show_options	= generic_show_options,
	.write_inode	= btrfs_write_inode,
	.dirty_inode	= btrfs_dirty_inode,
	.alloc_inode	= btrfs_alloc_inode,
	.destroy_inode	= btrfs_destroy_inode,
	.statfs		= btrfs_statfs,
	.write_super_lockfs = btrfs_write_super_lockfs,
	.unlockfs	= btrfs_unlockfs,
};

static const struct file_operations btrfs_ctl_fops = {
	.unlocked_ioctl	 = btrfs_control_ioctl,
	.compat_ioctl = btrfs_control_ioctl,
	.owner	 = THIS_MODULE,
};

static struct miscdevice btrfs_misc = {
	.minor		= MISC_DYNAMIC_MINOR,
	.name		= "btrfs-control",
	.fops		= &btrfs_ctl_fops
};

static int btrfs_interface_init(void)
{
	return misc_register(&btrfs_misc);
}

void btrfs_interface_exit(void)
{
	if (misc_deregister(&btrfs_misc) < 0)
		printk("misc_deregister failed for control device");
}

static int __init init_btrfs_fs(void)
{
	int err;

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

	err = btrfs_init_cachep();
	if (err)
		goto free_sysfs;

	err = extent_io_init();
	if (err)
		goto free_cachep;

	err = extent_map_init();
	if (err)
		goto free_extent_io;

	err = btrfs_interface_init();
	if (err)
		goto free_extent_map;
	err = register_filesystem(&btrfs_fs_type);
	if (err)
		goto unregister_ioctl;

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

unregister_ioctl:
	btrfs_interface_exit();
free_extent_map:
	extent_map_exit();
free_extent_io:
	extent_io_exit();
free_cachep:
	btrfs_destroy_cachep();
free_sysfs:
	btrfs_exit_sysfs();
	return err;
}

static void __exit exit_btrfs_fs(void)
{
	btrfs_destroy_cachep();
	extent_map_exit();
	extent_io_exit();
	btrfs_interface_exit();
	unregister_filesystem(&btrfs_fs_type);
	btrfs_exit_sysfs();
	btrfs_cleanup_fs_uuids();
}

module_init(init_btrfs_fs)
module_exit(exit_btrfs_fs)

MODULE_LICENSE("GPL");
Commit	Line	Data
	1	/*
	2	* Copyright (C) 2007 Oracle. All rights reserved.
	3	*
	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.
	7	*
	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.
	12	*
	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.
	17	*/
	18
	19	#include <linux/blkdev.h>
	20	#include <linux/module.h>
	21	#include <linux/buffer_head.h>
	22	#include <linux/fs.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>
	40	#include "ctree.h"
	41	#include "disk-io.h"
	42	#include "transaction.h"
	43	#include "btrfs_inode.h"
	44	#include "ioctl.h"
	45	#include "print-tree.h"
	46	#include "xattr.h"
	47	#include "volumes.h"
	48	#include "version.h"
	49	#include "export.h"
	50
	51	#define BTRFS_SUPER_MAGIC 0x9123683E
	52
	53	static struct super_operations btrfs_super_ops;
	54
	55	static void btrfs_put_super (struct super_block * sb)
	56	{
	57	struct btrfs_root *root = btrfs_sb(sb);
	58	struct btrfs_fs_info *fs = root->fs_info;
	59	int ret;
	60
	61	ret = close_ctree(root);
	62	if (ret) {
	63	printk("close ctree returns %d\n", ret);
	64	}
	65	btrfs_sysfs_del_super(fs);
	66	sb->s_fs_info = NULL;
	67	}
	68
	69	enum {
	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,
	73	};
	74
	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"},
	86	{Opt_ssd, "ssd"},
	87	{Opt_noacl, "noacl"},
	88	{Opt_err, NULL},
	89	};
	90
	91	u64 btrfs_parse_size(char *str)
	92	{
	93	u64 res;
	94	int mult = 1;
	95	char *end;
	96	char last;
	97
	98	res = simple_strtoul(str, &end, 10);
	99
	100	last = end[0];
	101	if (isalpha(last)) {
	102	last = tolower(last);
	103	switch (last) {
	104	case 'g':
	105	mult *= 1024;
	106	case 'm':
	107	mult *= 1024;
	108	case 'k':
	109	mult *= 1024;
	110	}
	111	res = res * mult;
	112	}
	113	return res;
	114	}
	115
	116	/*
	117	* Regular mount options parser. Everything that is needed only when
	118	* reading in a new superblock is parsed here.
	119	*/
	120	int btrfs_parse_options(struct btrfs_root root, char options)
	121	{
	122	struct btrfs_fs_info *info = root->fs_info;
	123	substring_t args[MAX_OPT_ARGS];
	124	char p, num;
	125	int intarg;
	126
	127	if (!options)
	128	return 0;
	129
	130	/*
	131	* strsep changes the string, duplicate it because parse_options
	132	* gets called twice
	133	*/
	134	options = kstrdup(options, GFP_NOFS);
	135	if (!options)
	136	return -ENOMEM;
	137
	138
	139	while ((p = strsep(&options, ",")) != NULL) {
	140	int token;
	141	if (!*p)
	142	continue;
	143
	144	token = match_token(p, tokens, args);
	145	switch (token) {
	146	case Opt_degraded:
	147	printk(KERN_INFO "btrfs: allowing degraded mounts\n");
	148	btrfs_set_opt(info->mount_opt, DEGRADED);
	149	break;
	150	case Opt_subvol:
	151	case Opt_device:
	152	/*
	153	* These are parsed by btrfs_parse_early_options
	154	* and can be happily ignored here.
	155	*/
	156	break;
	157	case Opt_nodatasum:
	158	printk(KERN_INFO "btrfs: setting nodatacsum\n");
	159	btrfs_set_opt(info->mount_opt, NODATASUM);
	160	break;
	161	case Opt_nodatacow:
	162	printk(KERN_INFO "btrfs: setting nodatacow\n");
	163	btrfs_set_opt(info->mount_opt, NODATACOW);
	164	btrfs_set_opt(info->mount_opt, NODATASUM);
	165	break;
	166	case Opt_ssd:
	167	printk(KERN_INFO "btrfs: use ssd allocation scheme\n");
	168	btrfs_set_opt(info->mount_opt, SSD);
	169	break;
	170	case Opt_nobarrier:
	171	printk(KERN_INFO "btrfs: turning off barriers\n");
	172	btrfs_set_opt(info->mount_opt, NOBARRIER);
	173	break;
	174	case Opt_thread_pool:
	175	intarg = 0;
	176	match_int(&args[0], &intarg);
	177	if (intarg) {
	178	info->thread_pool_size = intarg;
	179	printk(KERN_INFO "btrfs: thread pool %d\n",
	180	info->thread_pool_size);
	181	}
	182	break;
	183	case Opt_max_extent:
	184	num = match_strdup(&args[0]);
	185	if (num) {
	186	info->max_extent = btrfs_parse_size(num);
	187	kfree(num);
	188
	189	info->max_extent = max_t(u64,
	190	info->max_extent, root->sectorsize);
	191	printk(KERN_INFO "btrfs: max_extent at %llu\n",
	192	info->max_extent);
	193	}
	194	break;
	195	case Opt_max_inline:
	196	num = match_strdup(&args[0]);
	197	if (num) {
	198	info->max_inline = btrfs_parse_size(num);
	199	kfree(num);
	200
	201	if (info->max_inline) {
	202	info->max_inline = max_t(u64,
	203	info->max_inline,
	204	root->sectorsize);
	205	}
	206	printk(KERN_INFO "btrfs: max_inline at %llu\n",
	207	info->max_inline);
	208	}
	209	break;
	210	case Opt_alloc_start:
	211	num = match_strdup(&args[0]);
	212	if (num) {
	213	info->alloc_start = btrfs_parse_size(num);
	214	kfree(num);
	215	printk(KERN_INFO
	216	"btrfs: allocations start at %llu\n",
	217	info->alloc_start);
	218	}
	219	break;
	220	case Opt_noacl:
	221	root->fs_info->sb->s_flags &= ~MS_POSIXACL;
	222	break;
	223	default:
	224	break;
	225	}
	226	}
	227	kfree(options);
	228	return 0;
	229	}
	230
	231	/*
	232	* Parse mount options that are required early in the mount process.
	233	*
	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.
	236	*/
	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)
	240	{
	241	substring_t args[MAX_OPT_ARGS];
	242	char opts, p;
	243	int error = 0;
	244
	245	if (!options)
	246	goto out;
	247
	248	/*
	249	* strsep changes the string, duplicate it because parse_options
	250	* gets called twice
	251	*/
	252	opts = kstrdup(options, GFP_KERNEL);
	253	if (!opts)
	254	return -ENOMEM;
	255
	256	while ((p = strsep(&opts, ",")) != NULL) {
	257	int token;
	258	if (!*p)
	259	continue;
	260
	261	token = match_token(p, tokens, args);
	262	switch (token) {
	263	case Opt_subvol:
	264	*subvol_name = match_strdup(&args[0]);
	265	break;
	266	case Opt_device:
	267	error = btrfs_scan_one_device(match_strdup(&args[0]),
	268	flags, holder, fs_devices);
	269	if (error)
	270	goto out_free_opts;
	271	break;
	272	default:
	273	break;
	274	}
	275	}
	276
	277	out_free_opts:
	278	kfree(opts);
	279	out:
	280	/*
	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.
	284	*/
	285	if (!*subvol_name) {
	286	*subvol_name = kstrdup("default", GFP_KERNEL);
	287	if (!*subvol_name)
	288	return -ENOMEM;
	289	}
	290	return error;
	291	}
	292
	293	static int btrfs_fill_super(struct super_block * sb,
	294	struct btrfs_fs_devices *fs_devices,
	295	void * data, int silent)
	296	{
	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;
	302	int err;
	303
	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;
	309	sb->s_time_gran = 1;
	310	sb->s_flags \|= MS_POSIXACL;
	311
	312	tree_root = open_ctree(sb, fs_devices, (char *)data);
	313
	314	if (IS_ERR(tree_root)) {
	315	printk("btrfs: open_ctree failed\n");
	316	return PTR_ERR(tree_root);
	317	}
	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),
	321	tree_root);
	322	bi = BTRFS_I(inode);
	323	bi->location.objectid = inode->i_ino;
	324	bi->location.offset = 0;
	325	bi->root = tree_root;
	326
	327	btrfs_set_key_type(&bi->location, BTRFS_INODE_ITEM_KEY);
	328
	329	if (!inode) {
	330	err = -ENOMEM;
	331	goto fail_close;
	332	}
	333	if (inode->i_state & I_NEW) {
	334	btrfs_read_locked_inode(inode);
	335	unlock_new_inode(inode);
	336	}
	337
	338	root_dentry = d_alloc_root(inode);
	339	if (!root_dentry) {
	340	iput(inode);
	341	err = -ENOMEM;
	342	goto fail_close;
	343	}
	344
	345	/* this does the super kobj at the same time */
	346	err = btrfs_sysfs_add_super(tree_root->fs_info);
	347	if (err)
	348	goto fail_close;
	349
	350	sb->s_root = root_dentry;
	351
	352	save_mount_options(sb, data);
	353	return 0;
	354
	355	fail_close:
	356	close_ctree(tree_root);
	357	return err;
	358	}
	359
	360	int btrfs_sync_fs(struct super_block *sb, int wait)
	361	{
	362	struct btrfs_trans_handle *trans;
	363	struct btrfs_root *root;
	364	int ret;
	365	root = btrfs_sb(sb);
	366
	367	sb->s_dirt = 0;
	368	if (!wait) {
	369	filemap_flush(root->fs_info->btree_inode->i_mapping);
	370	return 0;
	371	}
	372	btrfs_clean_old_snapshots(root);
	373	trans = btrfs_start_transaction(root, 1);
	374	ret = btrfs_commit_transaction(trans, root);
	375	sb->s_dirt = 0;
	376	return ret;
	377	}
	378
	379	static void btrfs_write_super(struct super_block *sb)
	380	{
	381	sb->s_dirt = 0;
	382	}
	383
	384	static int btrfs_test_super(struct super_block s, void data)
	385	{
	386	struct btrfs_fs_devices *test_fs_devices = data;
	387	struct btrfs_root *root = btrfs_sb(s);
	388
	389	return root->fs_info->fs_devices == test_fs_devices;
	390	}
	391
	392	/*
	393	* Find a superblock for the given device / mount point.
	394	*
	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.
	397	*/
	398	static int btrfs_get_sb(struct file_system_type *fs_type, int flags,
	399	const char dev_name, void data, struct vfsmount *mnt)
	400	{
	401	char *subvol_name = NULL;
	402	struct block_device *bdev = NULL;
	403	struct super_block *s;
	404	struct dentry *root;
	405	struct btrfs_fs_devices *fs_devices = NULL;
	406	int error = 0;
	407
	408	error = btrfs_parse_early_options(data, flags, fs_type,
	409	&subvol_name, &fs_devices);
	410	if (error)
	411	goto error;
	412
	413	error = btrfs_scan_one_device(dev_name, flags, fs_type, &fs_devices);
	414	if (error)
	415	goto error_free_subvol_name;
	416
	417	error = btrfs_open_devices(fs_devices, flags, fs_type);
	418	if (error)
	419	goto error_free_subvol_name;
	420
	421	bdev = fs_devices->latest_bdev;
	422	s = sget(fs_type, btrfs_test_super, set_anon_super, fs_devices);
	423	if (IS_ERR(s))
	424	goto error_s;
	425
	426	if (s->s_root) {
	427	if ((flags ^ s->s_flags) & MS_RDONLY) {
	428	up_write(&s->s_umount);
	429	deactivate_super(s);
	430	error = -EBUSY;
	431	goto error_bdev;
	432	}
	433
	434	} else {
	435	char b[BDEVNAME_SIZE];
	436
	437	s->s_flags = flags;
	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);
	441	if (error) {
	442	up_write(&s->s_umount);
	443	deactivate_super(s);
	444	goto error;
	445	}
	446
	447	btrfs_sb(s)->fs_info->bdev_holder = fs_type;
	448	s->s_flags \|= MS_ACTIVE;
	449	}
	450
	451	if (!strcmp(subvol_name, "."))
	452	root = dget(s->s_root);
	453	else {
	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);
	457	if (IS_ERR(root)) {
	458	up_write(&s->s_umount);
	459	deactivate_super(s);
	460	error = PTR_ERR(root);
	461	goto error;
	462	}
	463	if (!root->d_inode) {
	464	dput(root);
	465	up_write(&s->s_umount);
	466	deactivate_super(s);
	467	error = -ENXIO;
	468	goto error;
	469	}
	470	}
	471
	472	mnt->mnt_sb = s;
	473	mnt->mnt_root = root;
	474
	475	kfree(subvol_name);
	476	return 0;
	477
	478	error_s:
	479	error = PTR_ERR(s);
	480	error_bdev:
	481	btrfs_close_devices(fs_devices);
	482	error_free_subvol_name:
	483	kfree(subvol_name);
	484	error:
	485	return error;
	486	}
	487
	488	static int btrfs_statfs(struct dentry dentry, struct kstatfs buf)
	489	{
	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;
	494
	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;
	510
	511	return 0;
	512	}
	513
	514	static struct file_system_type btrfs_fs_type = {
	515	.owner = THIS_MODULE,
	516	.name = "btrfs",
	517	.get_sb = btrfs_get_sb,
	518	.kill_sb = kill_anon_super,
	519	.fs_flags = FS_REQUIRES_DEV,
	520	};
	521
	522	static long btrfs_control_ioctl(struct file *file, unsigned int cmd,
	523	unsigned long arg)
	524	{
	525	struct btrfs_ioctl_vol_args *vol;
	526	struct btrfs_fs_devices *fs_devices;
	527	int ret = 0;
	528	int len;
	529
	530	vol = kmalloc(sizeof(*vol), GFP_KERNEL);
	531	if (copy_from_user(vol, (void __user )arg, sizeof(vol))) {
	532	ret = -EFAULT;
	533	goto out;
	534	}
	535	len = strnlen(vol->name, BTRFS_PATH_NAME_MAX);
	536	switch (cmd) {
	537	case BTRFS_IOC_SCAN_DEV:
	538	ret = btrfs_scan_one_device(vol->name, MS_RDONLY,
	539	&btrfs_fs_type, &fs_devices);
	540	break;
	541	}
	542	out:
	543	kfree(vol);
	544	return ret;
	545	}
	546
	547	static void btrfs_write_super_lockfs(struct super_block *sb)
	548	{
	549	struct btrfs_root *root = btrfs_sb(sb);
	550	mutex_lock(&root->fs_info->transaction_kthread_mutex);
	551	mutex_lock(&root->fs_info->cleaner_mutex);
	552	}
	553
	554	static void btrfs_unlockfs(struct super_block *sb)
	555	{
	556	struct btrfs_root *root = btrfs_sb(sb);
	557	mutex_unlock(&root->fs_info->cleaner_mutex);
	558	mutex_unlock(&root->fs_info->transaction_kthread_mutex);
	559	}
	560
	561	static struct super_operations btrfs_super_ops = {
	562	.delete_inode = btrfs_delete_inode,
	563	.put_super = btrfs_put_super,
	564	.write_super = btrfs_write_super,
	565	.sync_fs = btrfs_sync_fs,
	566	.show_options = generic_show_options,
	567	.write_inode = btrfs_write_inode,
	568	.dirty_inode = btrfs_dirty_inode,
	569	.alloc_inode = btrfs_alloc_inode,
	570	.destroy_inode = btrfs_destroy_inode,
	571	.statfs = btrfs_statfs,
	572	.write_super_lockfs = btrfs_write_super_lockfs,
	573	.unlockfs = btrfs_unlockfs,
	574	};
	575
	576	static const struct file_operations btrfs_ctl_fops = {
	577	.unlocked_ioctl = btrfs_control_ioctl,
	578	.compat_ioctl = btrfs_control_ioctl,
	579	.owner = THIS_MODULE,
	580	};
	581
	582	static struct miscdevice btrfs_misc = {
	583	.minor = MISC_DYNAMIC_MINOR,
	584	.name = "btrfs-control",
	585	.fops = &btrfs_ctl_fops
	586	};
	587
	588	static int btrfs_interface_init(void)
	589	{
	590	return misc_register(&btrfs_misc);
	591	}
	592
	593	void btrfs_interface_exit(void)
	594	{
	595	if (misc_deregister(&btrfs_misc) < 0)
	596	printk("misc_deregister failed for control device");
	597	}
	598
	599	static int __init init_btrfs_fs(void)
	600	{
	601	int err;
	602
	603	err = btrfs_init_sysfs();
	604	if (err)
	605	return err;
	606
	607	err = btrfs_init_cachep();
	608	if (err)
	609	goto free_sysfs;
	610
	611	err = extent_io_init();
	612	if (err)
	613	goto free_cachep;
	614
	615	err = extent_map_init();
	616	if (err)
	617	goto free_extent_io;
	618
	619	err = btrfs_interface_init();
	620	if (err)
	621	goto free_extent_map;
	622	err = register_filesystem(&btrfs_fs_type);
	623	if (err)
	624	goto unregister_ioctl;
	625
	626	printk(KERN_INFO "%s loaded\n", BTRFS_BUILD_VERSION);
	627	return 0;
	628
	629	unregister_ioctl:
	630	btrfs_interface_exit();
	631	free_extent_map:
	632	extent_map_exit();
	633	free_extent_io:
	634	extent_io_exit();
	635	free_cachep:
	636	btrfs_destroy_cachep();
	637	free_sysfs:
	638	btrfs_exit_sysfs();
	639	return err;
	640	}
	641
	642	static void __exit exit_btrfs_fs(void)
	643	{
	644	btrfs_destroy_cachep();
	645	extent_map_exit();
	646	extent_io_exit();
	647	btrfs_interface_exit();
	648	unregister_filesystem(&btrfs_fs_type);
	649	btrfs_exit_sysfs();
	650	btrfs_cleanup_fs_uuids();
	651	}
	652
	653	module_init(init_btrfs_fs)
	654	module_exit(exit_btrfs_fs)
	655
	656	MODULE_LICENSE("GPL");