blob: d9b97d4960e654754c3e976c101618d8fc5668b0 [file] [log] [blame]
Chris Mason6cbd5572007-06-12 09:07:21 -04001/*
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
Chris Mason2c90e5d2007-04-02 10:50:19 -040019#ifndef __BTRFS_I__
20#define __BTRFS_I__
21
Chris Masona52d9a82007-08-27 16:49:44 -040022#include "extent_map.h"
Chris Masond1310b22008-01-24 16:13:08 -050023#include "extent_io.h"
Chris Masone6dcd2d2008-07-17 12:53:50 -040024#include "ordered-data.h"
Miao Xie16cdcec2011-04-22 18:12:22 +080025#include "delayed-inode.h"
Chris Masona52d9a82007-08-27 16:49:44 -040026
Josef Bacik72ac3c02012-05-23 14:13:11 -040027/*
28 * ordered_data_close is set by truncate when a file that used
29 * to have good data has been truncated to zero. When it is set
30 * the btrfs file release call will add this inode to the
31 * ordered operations list so that we make sure to flush out any
32 * new data the application may have written before commit.
33 */
34#define BTRFS_INODE_ORDERED_DATA_CLOSE 0
35#define BTRFS_INODE_ORPHAN_META_RESERVED 1
36#define BTRFS_INODE_DUMMY 2
37#define BTRFS_INODE_IN_DEFRAG 3
38#define BTRFS_INODE_DELALLOC_META_RESERVED 4
Josef Bacik8a35d952012-05-23 14:26:42 -040039#define BTRFS_INODE_HAS_ORPHAN_ITEM 5
Josef Bacik7ddf5a42012-06-08 15:26:47 -040040#define BTRFS_INODE_HAS_ASYNC_EXTENT 6
Josef Bacik5dc562c2012-08-17 13:14:17 -040041#define BTRFS_INODE_NEEDS_FULL_SYNC 7
Josef Bacike9976152012-10-11 15:53:56 -040042#define BTRFS_INODE_COPY_EVERYTHING 8
Miao Xiedf0af1a2013-01-29 10:11:59 +000043#define BTRFS_INODE_IN_DELALLOC_LIST 9
Miao Xie2e60a512013-02-08 07:01:08 +000044#define BTRFS_INODE_READDIO_NEED_LOCK 10
Josef Bacik72ac3c02012-05-23 14:13:11 -040045
Aneeshf1ace242007-06-13 16:18:26 -040046/* in memory btrfs inode */
Chris Mason2c90e5d2007-04-02 10:50:19 -040047struct btrfs_inode {
Chris Masond352ac62008-09-29 15:18:18 -040048 /* which subvolume this inode belongs to */
Chris Masond6e4a422007-04-06 15:37:36 -040049 struct btrfs_root *root;
Chris Masond352ac62008-09-29 15:18:18 -040050
Chris Masond352ac62008-09-29 15:18:18 -040051 /* key used to find this inode on disk. This is used by the code
52 * to read in roots of subvolumes
53 */
Chris Masond6e4a422007-04-06 15:37:36 -040054 struct btrfs_key location;
Chris Masond352ac62008-09-29 15:18:18 -040055
Josef Bacik9e0baf62011-07-15 15:16:44 +000056 /* Lock for counters */
57 spinlock_t lock;
58
Chris Masond352ac62008-09-29 15:18:18 -040059 /* the extent_tree has caches of all the extent mappings to disk */
Chris Masona52d9a82007-08-27 16:49:44 -040060 struct extent_map_tree extent_tree;
Chris Masond352ac62008-09-29 15:18:18 -040061
62 /* the io_tree does range state (DIRTY, LOCKED etc) */
Chris Masond1310b22008-01-24 16:13:08 -050063 struct extent_io_tree io_tree;
Chris Masond352ac62008-09-29 15:18:18 -040064
65 /* special utility tree used to record which mirrors have already been
66 * tried when checksums fail for a given block
67 */
Chris Mason7e383262008-04-09 16:28:12 -040068 struct extent_io_tree io_failure_tree;
Chris Masond352ac62008-09-29 15:18:18 -040069
Chris Masond352ac62008-09-29 15:18:18 -040070 /* held while logging the inode in tree-log.c */
Chris Masone02119d2008-09-05 16:13:11 -040071 struct mutex log_mutex;
Chris Masond352ac62008-09-29 15:18:18 -040072
Josef Bacikf2486792012-01-13 12:09:22 -050073 /* held while doing delalloc reservations */
74 struct mutex delalloc_mutex;
75
Chris Masond352ac62008-09-29 15:18:18 -040076 /* used to order data wrt metadata */
Chris Masone6dcd2d2008-07-17 12:53:50 -040077 struct btrfs_ordered_inode_tree ordered_tree;
Josef Bacik15ee9bc2007-08-10 16:22:09 -040078
Chris Masond352ac62008-09-29 15:18:18 -040079 /* list of all the delalloc inodes in the FS. There are times we need
80 * to write all the delalloc pages to disk, and this list is used
81 * to walk them all.
82 */
Chris Masonea8c2812008-08-04 23:17:27 -040083 struct list_head delalloc_inodes;
84
Chris Mason5a3f23d2009-03-31 13:27:11 -040085 /*
86 * list for tracking inodes that must be sent to disk before a
87 * rename or truncate commit
88 */
89 struct list_head ordered_operations;
90
Yan Zheng5d4f98a2009-06-10 10:45:14 -040091 /* node for the red-black tree that links inodes in subvolume root */
92 struct rb_node rb_node;
93
Josef Bacik72ac3c02012-05-23 14:13:11 -040094 unsigned long runtime_flags;
95
Josef Bacikb812ce22012-11-16 13:56:32 -050096 /* Keep track of who's O_SYNC/fsycing currently */
97 atomic_t sync_writers;
98
Chris Masond352ac62008-09-29 15:18:18 -040099 /* full 64 bit generation number, struct vfs_inode doesn't have a big
100 * enough field for this.
101 */
Chris Masone02119d2008-09-05 16:13:11 -0400102 u64 generation;
103
Josef Bacik15ee9bc2007-08-10 16:22:09 -0400104 /*
105 * transid of the trans_handle that last modified this inode
106 */
107 u64 last_trans;
Chris Mason257c62e2009-10-13 13:21:08 -0400108
109 /*
110 * log transid when this inode was last modified
111 */
112 u64 last_sub_trans;
113
Chris Masone02119d2008-09-05 16:13:11 -0400114 /*
115 * transid that last logged this inode
116 */
117 u64 logged_trans;
Chris Mason49eb7e42008-09-11 15:53:12 -0400118
Chris Masond352ac62008-09-29 15:18:18 -0400119 /* total number of bytes pending delalloc, used by stat to calc the
120 * real block usage of the file
121 */
Chris Mason90692182008-02-08 13:49:28 -0500122 u64 delalloc_bytes;
Chris Masond352ac62008-09-29 15:18:18 -0400123
124 /*
125 * the size of the file stored in the metadata on disk. data=ordered
126 * means the in-memory i_size might be larger than the size on disk
127 * because not all the blocks are written yet.
128 */
Chris Masondbe674a2008-07-17 12:54:05 -0400129 u64 disk_i_size;
Chris Masond352ac62008-09-29 15:18:18 -0400130
Josef Bacikaec74772008-07-24 12:12:38 -0400131 /*
132 * if this is a directory then index_cnt is the counter for the index
133 * number for new files that are created
134 */
135 u64 index_cnt;
Chris Masond352ac62008-09-29 15:18:18 -0400136
Chris Mason12fcfd22009-03-24 10:24:20 -0400137 /* the fsync log has some corner cases that mean we have to check
138 * directories to see if any unlinks have been done before
139 * the directory was logged. See tree-log.c for all the
140 * details
141 */
142 u64 last_unlink_trans;
143
Josef Bacik7709cde2011-08-04 10:25:02 -0400144 /*
145 * Number of bytes outstanding that are going to need csums. This is
146 * used in ENOSPC accounting.
147 */
148 u64 csum_bytes;
149
Josef Bacikf1bdcc02011-07-14 14:28:08 -0400150 /* flags field from the on disk inode */
151 u32 flags;
152
Liu Bo46d8bc32012-08-29 01:07:55 -0600153 /* a local copy of root's last_log_commit */
154 unsigned long last_log_commit;
155
Chris Mason5a3f23d2009-03-31 13:27:11 -0400156 /*
Josef Bacik32c00af2009-10-08 13:34:05 -0400157 * Counters to keep track of the number of extent item's we may use due
158 * to delalloc and such. outstanding_extents is the number of extent
159 * items we think we'll end up using, and reserved_extents is the number
160 * of extent items we've reserved metadata for.
Josef Bacik9ed74f22009-09-11 16:12:44 -0400161 */
Josef Bacik9e0baf62011-07-15 15:16:44 +0000162 unsigned outstanding_extents;
163 unsigned reserved_extents;
Josef Bacik9ed74f22009-09-11 16:12:44 -0400164
165 /*
Chris Mason1e701a32010-03-11 09:42:04 -0500166 * always compress this one file
167 */
Josef Bacik72ac3c02012-05-23 14:13:11 -0400168 unsigned force_compress;
Chris Mason1e701a32010-03-11 09:42:04 -0500169
Miao Xie16cdcec2011-04-22 18:12:22 +0800170 struct btrfs_delayed_node *delayed_node;
171
Chris Masond352ac62008-09-29 15:18:18 -0400172 struct inode vfs_inode;
Chris Mason2c90e5d2007-04-02 10:50:19 -0400173};
Chris Masondbe674a2008-07-17 12:54:05 -0400174
Miao Xie16cdcec2011-04-22 18:12:22 +0800175extern unsigned char btrfs_filetype_table[];
176
Chris Mason2c90e5d2007-04-02 10:50:19 -0400177static inline struct btrfs_inode *BTRFS_I(struct inode *inode)
178{
179 return container_of(inode, struct btrfs_inode, vfs_inode);
180}
181
Li Zefan33345d012011-04-20 10:31:50 +0800182static inline u64 btrfs_ino(struct inode *inode)
183{
184 u64 ino = BTRFS_I(inode)->location.objectid;
185
Liu Bo14c7cca2011-09-11 10:52:24 -0400186 /*
187 * !ino: btree_inode
188 * type == BTRFS_ROOT_ITEM_KEY: subvol dir
189 */
190 if (!ino || BTRFS_I(inode)->location.type == BTRFS_ROOT_ITEM_KEY)
Li Zefan33345d012011-04-20 10:31:50 +0800191 ino = inode->i_ino;
192 return ino;
193}
194
Chris Masondbe674a2008-07-17 12:54:05 -0400195static inline void btrfs_i_size_write(struct inode *inode, u64 size)
196{
Yan, Zhengc2167752009-11-12 09:34:21 +0000197 i_size_write(inode, size);
Chris Masondbe674a2008-07-17 12:54:05 -0400198 BTRFS_I(inode)->disk_i_size = size;
199}
200
Liu Bo83eea1f2012-07-10 05:28:39 -0600201static inline bool btrfs_is_free_space_inode(struct inode *inode)
Chris Mason2cf85722011-07-26 15:35:09 -0400202{
Liu Bo83eea1f2012-07-10 05:28:39 -0600203 struct btrfs_root *root = BTRFS_I(inode)->root;
204
Liu Bo51a8cf92012-07-10 05:28:38 -0600205 if (root == root->fs_info->tree_root &&
206 btrfs_ino(inode) != BTRFS_BTREE_INODE_OBJECTID)
207 return true;
208 if (BTRFS_I(inode)->location.objectid == BTRFS_FREE_INO_OBJECTID)
Chris Mason2cf85722011-07-26 15:35:09 -0400209 return true;
210 return false;
211}
212
Josef Bacik22ee6982012-05-29 16:57:49 -0400213static inline int btrfs_inode_in_log(struct inode *inode, u64 generation)
214{
Josef Bacik22ee6982012-05-29 16:57:49 -0400215 if (BTRFS_I(inode)->logged_trans == generation &&
Liu Bo46d8bc32012-08-29 01:07:55 -0600216 BTRFS_I(inode)->last_sub_trans <= BTRFS_I(inode)->last_log_commit)
217 return 1;
218 return 0;
Josef Bacik22ee6982012-05-29 16:57:49 -0400219}
220
Miao Xie2e60a512013-02-08 07:01:08 +0000221/*
222 * Disable DIO read nolock optimization, so new dio readers will be forced
223 * to grab i_mutex. It is used to avoid the endless truncate due to
224 * nonlocked dio read.
225 */
226static inline void btrfs_inode_block_unlocked_dio(struct inode *inode)
227{
228 set_bit(BTRFS_INODE_READDIO_NEED_LOCK, &BTRFS_I(inode)->runtime_flags);
229 smp_mb();
230}
231
232static inline void btrfs_inode_resume_unlocked_dio(struct inode *inode)
233{
234 smp_mb__before_clear_bit();
235 clear_bit(BTRFS_INODE_READDIO_NEED_LOCK,
236 &BTRFS_I(inode)->runtime_flags);
237}
238
Chris Mason2c90e5d2007-04-02 10:50:19 -0400239#endif