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.
22 #include <linux/hash.h>
23 #include "extent_map.h"
24 #include "extent_io.h"
25 #include "ordered-data.h"
26 #include "delayed-inode.h"
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.
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
40 #define BTRFS_INODE_HAS_ORPHAN_ITEM 5
41 #define BTRFS_INODE_HAS_ASYNC_EXTENT 6
42 #define BTRFS_INODE_NEEDS_FULL_SYNC 7
43 #define BTRFS_INODE_COPY_EVERYTHING 8
44 #define BTRFS_INODE_IN_DELALLOC_LIST 9
45 #define BTRFS_INODE_READDIO_NEED_LOCK 10
46 #define BTRFS_INODE_HAS_PROPS 11
48 * The following 3 bits are meant only for the btree inode.
49 * When any of them is set, it means an error happened while writing an
50 * extent buffer belonging to:
52 * 2) a log btree and first log sub-transaction
53 * 3) a log btree and second log sub-transaction
55 #define BTRFS_INODE_BTREE_ERR 12
56 #define BTRFS_INODE_BTREE_LOG1_ERR 13
57 #define BTRFS_INODE_BTREE_LOG2_ERR 14
59 /* in memory btrfs inode */
61 /* which subvolume this inode belongs to */
62 struct btrfs_root
*root
;
64 /* key used to find this inode on disk. This is used by the code
65 * to read in roots of subvolumes
67 struct btrfs_key location
;
69 /* Lock for counters */
72 /* the extent_tree has caches of all the extent mappings to disk */
73 struct extent_map_tree extent_tree
;
75 /* the io_tree does range state (DIRTY, LOCKED etc) */
76 struct extent_io_tree io_tree
;
78 /* special utility tree used to record which mirrors have already been
79 * tried when checksums fail for a given block
81 struct extent_io_tree io_failure_tree
;
83 /* held while logging the inode in tree-log.c */
84 struct mutex log_mutex
;
86 /* held while doing delalloc reservations */
87 struct mutex delalloc_mutex
;
89 /* used to order data wrt metadata */
90 struct btrfs_ordered_inode_tree ordered_tree
;
92 /* list of all the delalloc inodes in the FS. There are times we need
93 * to write all the delalloc pages to disk, and this list is used
96 struct list_head delalloc_inodes
;
98 /* node for the red-black tree that links inodes in subvolume root */
99 struct rb_node rb_node
;
101 unsigned long runtime_flags
;
103 /* Keep track of who's O_SYNC/fsyncing currently */
104 atomic_t sync_writers
;
106 /* full 64 bit generation number, struct vfs_inode doesn't have a big
107 * enough field for this.
112 * transid of the trans_handle that last modified this inode
117 * transid that last logged this inode
122 * log transid when this inode was last modified
126 /* a local copy of root's last_log_commit */
129 /* total number of bytes pending delalloc, used by stat to calc the
130 * real block usage of the file
135 * total number of bytes pending defrag, used by stat to check whether
141 * the size of the file stored in the metadata on disk. data=ordered
142 * means the in-memory i_size might be larger than the size on disk
143 * because not all the blocks are written yet.
148 * if this is a directory then index_cnt is the counter for the index
149 * number for new files that are created
153 /* Cache the directory index number to speed the dir/file remove */
156 /* the fsync log has some corner cases that mean we have to check
157 * directories to see if any unlinks have been done before
158 * the directory was logged. See tree-log.c for all the
161 u64 last_unlink_trans
;
164 * Number of bytes outstanding that are going to need csums. This is
165 * used in ENOSPC accounting.
169 /* flags field from the on disk inode */
173 * Counters to keep track of the number of extent item's we may use due
174 * to delalloc and such. outstanding_extents is the number of extent
175 * items we think we'll end up using, and reserved_extents is the number
176 * of extent items we've reserved metadata for.
178 unsigned outstanding_extents
;
179 unsigned reserved_extents
;
182 * always compress this one file
184 unsigned force_compress
;
186 struct btrfs_delayed_node
*delayed_node
;
188 struct inode vfs_inode
;
191 extern unsigned char btrfs_filetype_table
[];
193 static inline struct btrfs_inode
*BTRFS_I(struct inode
*inode
)
195 return container_of(inode
, struct btrfs_inode
, vfs_inode
);
198 static inline unsigned long btrfs_inode_hash(u64 objectid
,
199 const struct btrfs_root
*root
)
201 u64 h
= objectid
^ (root
->objectid
* GOLDEN_RATIO_PRIME
);
203 #if BITS_PER_LONG == 32
204 h
= (h
>> 32) ^ (h
& 0xffffffff);
207 return (unsigned long)h
;
210 static inline void btrfs_insert_inode_hash(struct inode
*inode
)
212 unsigned long h
= btrfs_inode_hash(inode
->i_ino
, BTRFS_I(inode
)->root
);
214 __insert_inode_hash(inode
, h
);
217 static inline u64
btrfs_ino(struct inode
*inode
)
219 u64 ino
= BTRFS_I(inode
)->location
.objectid
;
223 * type == BTRFS_ROOT_ITEM_KEY: subvol dir
225 if (!ino
|| BTRFS_I(inode
)->location
.type
== BTRFS_ROOT_ITEM_KEY
)
230 static inline void btrfs_i_size_write(struct inode
*inode
, u64 size
)
232 i_size_write(inode
, size
);
233 BTRFS_I(inode
)->disk_i_size
= size
;
236 static inline bool btrfs_is_free_space_inode(struct inode
*inode
)
238 struct btrfs_root
*root
= BTRFS_I(inode
)->root
;
240 if (root
== root
->fs_info
->tree_root
&&
241 btrfs_ino(inode
) != BTRFS_BTREE_INODE_OBJECTID
)
243 if (BTRFS_I(inode
)->location
.objectid
== BTRFS_FREE_INO_OBJECTID
)
248 static inline int btrfs_inode_in_log(struct inode
*inode
, u64 generation
)
250 if (BTRFS_I(inode
)->logged_trans
== generation
&&
251 BTRFS_I(inode
)->last_sub_trans
<=
252 BTRFS_I(inode
)->last_log_commit
&&
253 BTRFS_I(inode
)->last_sub_trans
<=
254 BTRFS_I(inode
)->root
->last_log_commit
) {
256 * After a ranged fsync we might have left some extent maps
257 * (that fall outside the fsync's range). So return false
258 * here if the list isn't empty, to make sure btrfs_log_inode()
259 * will be called and process those extent maps.
262 if (list_empty(&BTRFS_I(inode
)->extent_tree
.modified_extents
))
268 #define BTRFS_DIO_ORIG_BIO_SUBMITTED 0x1
270 struct btrfs_dio_private
{
278 /* number of bios pending for this dio */
279 atomic_t pending_bios
;
284 /* orig_bio is our btrfs_io_bio */
285 struct bio
*orig_bio
;
287 /* dio_bio came from fs/direct-io.c */
291 * The original bio may be splited to several sub-bios, this is
292 * done during endio of sub-bios
294 int (*subio_endio
)(struct inode
*, struct btrfs_io_bio
*, int);
298 * Disable DIO read nolock optimization, so new dio readers will be forced
299 * to grab i_mutex. It is used to avoid the endless truncate due to
300 * nonlocked dio read.
302 static inline void btrfs_inode_block_unlocked_dio(struct inode
*inode
)
304 set_bit(BTRFS_INODE_READDIO_NEED_LOCK
, &BTRFS_I(inode
)->runtime_flags
);
308 static inline void btrfs_inode_resume_unlocked_dio(struct inode
*inode
)
310 smp_mb__before_atomic();
311 clear_bit(BTRFS_INODE_READDIO_NEED_LOCK
,
312 &BTRFS_I(inode
)->runtime_flags
);
315 bool btrfs_page_exists_in_range(struct inode
*inode
, loff_t start
, loff_t end
);