Commit | Line | Data |
---|---|---|
6cbd5570 CM |
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 | ||
2c90e5d6 CM |
19 | #ifndef __BTRFS_I__ |
20 | #define __BTRFS_I__ | |
21 | ||
778ba82b | 22 | #include <linux/hash.h> |
a52d9a80 | 23 | #include "extent_map.h" |
d1310b2e | 24 | #include "extent_io.h" |
e6dcd2dc | 25 | #include "ordered-data.h" |
16cdcec7 | 26 | #include "delayed-inode.h" |
a52d9a80 | 27 | |
72ac3c0d JB |
28 | /* |
29 | * ordered_data_close is set by truncate when a file that used | |
30 | * to have good data has been truncated to zero. When it is set | |
31 | * the btrfs file release call will add this inode to the | |
32 | * ordered operations list so that we make sure to flush out any | |
33 | * new data the application may have written before commit. | |
34 | */ | |
35 | #define BTRFS_INODE_ORDERED_DATA_CLOSE 0 | |
36 | #define BTRFS_INODE_ORPHAN_META_RESERVED 1 | |
37 | #define BTRFS_INODE_DUMMY 2 | |
38 | #define BTRFS_INODE_IN_DEFRAG 3 | |
39 | #define BTRFS_INODE_DELALLOC_META_RESERVED 4 | |
8a35d95f | 40 | #define BTRFS_INODE_HAS_ORPHAN_ITEM 5 |
7ddf5a42 | 41 | #define BTRFS_INODE_HAS_ASYNC_EXTENT 6 |
5dc562c5 | 42 | #define BTRFS_INODE_NEEDS_FULL_SYNC 7 |
e9976151 | 43 | #define BTRFS_INODE_COPY_EVERYTHING 8 |
df0af1a5 | 44 | #define BTRFS_INODE_IN_DELALLOC_LIST 9 |
2e60a51e | 45 | #define BTRFS_INODE_READDIO_NEED_LOCK 10 |
63541927 | 46 | #define BTRFS_INODE_HAS_PROPS 11 |
ddba1bfc LB |
47 | /* DIO is ready to submit */ |
48 | #define BTRFS_INODE_DIO_READY 12 | |
656f30db FM |
49 | /* |
50 | * The following 3 bits are meant only for the btree inode. | |
51 | * When any of them is set, it means an error happened while writing an | |
52 | * extent buffer belonging to: | |
53 | * 1) a non-log btree | |
54 | * 2) a log btree and first log sub-transaction | |
55 | * 3) a log btree and second log sub-transaction | |
56 | */ | |
57 | #define BTRFS_INODE_BTREE_ERR 12 | |
58 | #define BTRFS_INODE_BTREE_LOG1_ERR 13 | |
59 | #define BTRFS_INODE_BTREE_LOG2_ERR 14 | |
72ac3c0d | 60 | |
f1ace244 | 61 | /* in memory btrfs inode */ |
2c90e5d6 | 62 | struct btrfs_inode { |
d352ac68 | 63 | /* which subvolume this inode belongs to */ |
d6e4a428 | 64 | struct btrfs_root *root; |
d352ac68 | 65 | |
d352ac68 CM |
66 | /* key used to find this inode on disk. This is used by the code |
67 | * to read in roots of subvolumes | |
68 | */ | |
d6e4a428 | 69 | struct btrfs_key location; |
d352ac68 | 70 | |
2f2ff0ee FM |
71 | /* |
72 | * Lock for counters and all fields used to determine if the inode is in | |
73 | * the log or not (last_trans, last_sub_trans, last_log_commit, | |
74 | * logged_trans). | |
75 | */ | |
9e0baf60 JB |
76 | spinlock_t lock; |
77 | ||
d352ac68 | 78 | /* the extent_tree has caches of all the extent mappings to disk */ |
a52d9a80 | 79 | struct extent_map_tree extent_tree; |
d352ac68 CM |
80 | |
81 | /* the io_tree does range state (DIRTY, LOCKED etc) */ | |
d1310b2e | 82 | struct extent_io_tree io_tree; |
d352ac68 CM |
83 | |
84 | /* special utility tree used to record which mirrors have already been | |
85 | * tried when checksums fail for a given block | |
86 | */ | |
7e38326f | 87 | struct extent_io_tree io_failure_tree; |
d352ac68 | 88 | |
d352ac68 | 89 | /* held while logging the inode in tree-log.c */ |
e02119d5 | 90 | struct mutex log_mutex; |
d352ac68 | 91 | |
f248679e JB |
92 | /* held while doing delalloc reservations */ |
93 | struct mutex delalloc_mutex; | |
94 | ||
d352ac68 | 95 | /* used to order data wrt metadata */ |
e6dcd2dc | 96 | struct btrfs_ordered_inode_tree ordered_tree; |
15ee9bc7 | 97 | |
d352ac68 CM |
98 | /* list of all the delalloc inodes in the FS. There are times we need |
99 | * to write all the delalloc pages to disk, and this list is used | |
100 | * to walk them all. | |
101 | */ | |
ea8c2819 CM |
102 | struct list_head delalloc_inodes; |
103 | ||
5d4f98a2 YZ |
104 | /* node for the red-black tree that links inodes in subvolume root */ |
105 | struct rb_node rb_node; | |
106 | ||
72ac3c0d JB |
107 | unsigned long runtime_flags; |
108 | ||
9c931c5a | 109 | /* Keep track of who's O_SYNC/fsyncing currently */ |
b812ce28 JB |
110 | atomic_t sync_writers; |
111 | ||
d352ac68 CM |
112 | /* full 64 bit generation number, struct vfs_inode doesn't have a big |
113 | * enough field for this. | |
114 | */ | |
e02119d5 CM |
115 | u64 generation; |
116 | ||
15ee9bc7 JB |
117 | /* |
118 | * transid of the trans_handle that last modified this inode | |
119 | */ | |
120 | u64 last_trans; | |
257c62e1 CM |
121 | |
122 | /* | |
bb14a59b | 123 | * transid that last logged this inode |
257c62e1 | 124 | */ |
bb14a59b | 125 | u64 logged_trans; |
257c62e1 | 126 | |
e02119d5 | 127 | /* |
bb14a59b | 128 | * log transid when this inode was last modified |
e02119d5 | 129 | */ |
bb14a59b MX |
130 | int last_sub_trans; |
131 | ||
132 | /* a local copy of root's last_log_commit */ | |
133 | int last_log_commit; | |
49eb7e46 | 134 | |
d352ac68 CM |
135 | /* total number of bytes pending delalloc, used by stat to calc the |
136 | * real block usage of the file | |
137 | */ | |
9069218d | 138 | u64 delalloc_bytes; |
d352ac68 | 139 | |
47059d93 WS |
140 | /* |
141 | * total number of bytes pending defrag, used by stat to check whether | |
142 | * it needs COW. | |
143 | */ | |
144 | u64 defrag_bytes; | |
145 | ||
d352ac68 CM |
146 | /* |
147 | * the size of the file stored in the metadata on disk. data=ordered | |
148 | * means the in-memory i_size might be larger than the size on disk | |
149 | * because not all the blocks are written yet. | |
150 | */ | |
dbe674a9 | 151 | u64 disk_i_size; |
d352ac68 | 152 | |
aec7477b JB |
153 | /* |
154 | * if this is a directory then index_cnt is the counter for the index | |
155 | * number for new files that are created | |
156 | */ | |
157 | u64 index_cnt; | |
d352ac68 | 158 | |
67de1176 MX |
159 | /* Cache the directory index number to speed the dir/file remove */ |
160 | u64 dir_index; | |
161 | ||
12fcfd22 CM |
162 | /* the fsync log has some corner cases that mean we have to check |
163 | * directories to see if any unlinks have been done before | |
164 | * the directory was logged. See tree-log.c for all the | |
165 | * details | |
166 | */ | |
167 | u64 last_unlink_trans; | |
168 | ||
7709cde3 JB |
169 | /* |
170 | * Number of bytes outstanding that are going to need csums. This is | |
171 | * used in ENOSPC accounting. | |
172 | */ | |
173 | u64 csum_bytes; | |
174 | ||
f1bdcc0a JB |
175 | /* flags field from the on disk inode */ |
176 | u32 flags; | |
177 | ||
9ed74f2d | 178 | /* |
32c00aff JB |
179 | * Counters to keep track of the number of extent item's we may use due |
180 | * to delalloc and such. outstanding_extents is the number of extent | |
181 | * items we think we'll end up using, and reserved_extents is the number | |
182 | * of extent items we've reserved metadata for. | |
9ed74f2d | 183 | */ |
9e0baf60 JB |
184 | unsigned outstanding_extents; |
185 | unsigned reserved_extents; | |
9ed74f2d | 186 | |
1e701a32 CM |
187 | /* |
188 | * always compress this one file | |
189 | */ | |
72ac3c0d | 190 | unsigned force_compress; |
1e701a32 | 191 | |
16cdcec7 MX |
192 | struct btrfs_delayed_node *delayed_node; |
193 | ||
9cc97d64 | 194 | /* File creation time. */ |
195 | struct timespec i_otime; | |
196 | ||
d352ac68 | 197 | struct inode vfs_inode; |
2c90e5d6 | 198 | }; |
dbe674a9 | 199 | |
16cdcec7 MX |
200 | extern unsigned char btrfs_filetype_table[]; |
201 | ||
2c90e5d6 CM |
202 | static inline struct btrfs_inode *BTRFS_I(struct inode *inode) |
203 | { | |
204 | return container_of(inode, struct btrfs_inode, vfs_inode); | |
205 | } | |
206 | ||
778ba82b FDBM |
207 | static inline unsigned long btrfs_inode_hash(u64 objectid, |
208 | const struct btrfs_root *root) | |
209 | { | |
210 | u64 h = objectid ^ (root->objectid * GOLDEN_RATIO_PRIME); | |
211 | ||
212 | #if BITS_PER_LONG == 32 | |
213 | h = (h >> 32) ^ (h & 0xffffffff); | |
214 | #endif | |
215 | ||
216 | return (unsigned long)h; | |
217 | } | |
218 | ||
219 | static inline void btrfs_insert_inode_hash(struct inode *inode) | |
220 | { | |
221 | unsigned long h = btrfs_inode_hash(inode->i_ino, BTRFS_I(inode)->root); | |
222 | ||
223 | __insert_inode_hash(inode, h); | |
224 | } | |
225 | ||
33345d01 LZ |
226 | static inline u64 btrfs_ino(struct inode *inode) |
227 | { | |
228 | u64 ino = BTRFS_I(inode)->location.objectid; | |
229 | ||
14c7cca7 LB |
230 | /* |
231 | * !ino: btree_inode | |
232 | * type == BTRFS_ROOT_ITEM_KEY: subvol dir | |
233 | */ | |
234 | if (!ino || BTRFS_I(inode)->location.type == BTRFS_ROOT_ITEM_KEY) | |
33345d01 LZ |
235 | ino = inode->i_ino; |
236 | return ino; | |
237 | } | |
238 | ||
dbe674a9 CM |
239 | static inline void btrfs_i_size_write(struct inode *inode, u64 size) |
240 | { | |
c2167754 | 241 | i_size_write(inode, size); |
dbe674a9 CM |
242 | BTRFS_I(inode)->disk_i_size = size; |
243 | } | |
244 | ||
83eea1f1 | 245 | static inline bool btrfs_is_free_space_inode(struct inode *inode) |
2cf8572d | 246 | { |
83eea1f1 LB |
247 | struct btrfs_root *root = BTRFS_I(inode)->root; |
248 | ||
51a8cf9d LB |
249 | if (root == root->fs_info->tree_root && |
250 | btrfs_ino(inode) != BTRFS_BTREE_INODE_OBJECTID) | |
251 | return true; | |
252 | if (BTRFS_I(inode)->location.objectid == BTRFS_FREE_INO_OBJECTID) | |
2cf8572d CM |
253 | return true; |
254 | return false; | |
255 | } | |
256 | ||
22ee6985 JB |
257 | static inline int btrfs_inode_in_log(struct inode *inode, u64 generation) |
258 | { | |
2f2ff0ee FM |
259 | int ret = 0; |
260 | ||
261 | spin_lock(&BTRFS_I(inode)->lock); | |
22ee6985 | 262 | if (BTRFS_I(inode)->logged_trans == generation && |
a5874ce6 JB |
263 | BTRFS_I(inode)->last_sub_trans <= |
264 | BTRFS_I(inode)->last_log_commit && | |
265 | BTRFS_I(inode)->last_sub_trans <= | |
125c4cf9 FM |
266 | BTRFS_I(inode)->root->last_log_commit) { |
267 | /* | |
268 | * After a ranged fsync we might have left some extent maps | |
269 | * (that fall outside the fsync's range). So return false | |
270 | * here if the list isn't empty, to make sure btrfs_log_inode() | |
271 | * will be called and process those extent maps. | |
272 | */ | |
273 | smp_mb(); | |
274 | if (list_empty(&BTRFS_I(inode)->extent_tree.modified_extents)) | |
2f2ff0ee | 275 | ret = 1; |
125c4cf9 | 276 | } |
2f2ff0ee FM |
277 | spin_unlock(&BTRFS_I(inode)->lock); |
278 | return ret; | |
22ee6985 JB |
279 | } |
280 | ||
c1dc0896 MX |
281 | #define BTRFS_DIO_ORIG_BIO_SUBMITTED 0x1 |
282 | ||
facc8a22 MX |
283 | struct btrfs_dio_private { |
284 | struct inode *inode; | |
c1dc0896 | 285 | unsigned long flags; |
facc8a22 MX |
286 | u64 logical_offset; |
287 | u64 disk_bytenr; | |
288 | u64 bytes; | |
289 | void *private; | |
290 | ||
291 | /* number of bios pending for this dio */ | |
292 | atomic_t pending_bios; | |
293 | ||
294 | /* IO errors */ | |
295 | int errors; | |
296 | ||
297 | /* orig_bio is our btrfs_io_bio */ | |
298 | struct bio *orig_bio; | |
299 | ||
300 | /* dio_bio came from fs/direct-io.c */ | |
301 | struct bio *dio_bio; | |
c1dc0896 MX |
302 | |
303 | /* | |
304 | * The original bio may be splited to several sub-bios, this is | |
305 | * done during endio of sub-bios | |
306 | */ | |
8b110e39 | 307 | int (*subio_endio)(struct inode *, struct btrfs_io_bio *, int); |
facc8a22 MX |
308 | }; |
309 | ||
2e60a51e MX |
310 | /* |
311 | * Disable DIO read nolock optimization, so new dio readers will be forced | |
312 | * to grab i_mutex. It is used to avoid the endless truncate due to | |
313 | * nonlocked dio read. | |
314 | */ | |
315 | static inline void btrfs_inode_block_unlocked_dio(struct inode *inode) | |
316 | { | |
317 | set_bit(BTRFS_INODE_READDIO_NEED_LOCK, &BTRFS_I(inode)->runtime_flags); | |
318 | smp_mb(); | |
319 | } | |
320 | ||
321 | static inline void btrfs_inode_resume_unlocked_dio(struct inode *inode) | |
322 | { | |
4e857c58 | 323 | smp_mb__before_atomic(); |
2e60a51e MX |
324 | clear_bit(BTRFS_INODE_READDIO_NEED_LOCK, |
325 | &BTRFS_I(inode)->runtime_flags); | |
326 | } | |
327 | ||
fc4adbff AG |
328 | bool btrfs_page_exists_in_range(struct inode *inode, loff_t start, loff_t end); |
329 | ||
2c90e5d6 | 330 | #endif |