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 #ifndef __BTRFS_I__ 20 #define __BTRFS_I__ 21 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" 27 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 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 47 /* in memory btrfs inode */ 48 struct btrfs_inode { 49 /* which subvolume this inode belongs to */ 50 struct btrfs_root *root; 51 52 /* key used to find this inode on disk. This is used by the code 53 * to read in roots of subvolumes 54 */ 55 struct btrfs_key location; 56 57 /* Lock for counters */ 58 spinlock_t lock; 59 60 /* the extent_tree has caches of all the extent mappings to disk */ 61 struct extent_map_tree extent_tree; 62 63 /* the io_tree does range state (DIRTY, LOCKED etc) */ 64 struct extent_io_tree io_tree; 65 66 /* special utility tree used to record which mirrors have already been 67 * tried when checksums fail for a given block 68 */ 69 struct extent_io_tree io_failure_tree; 70 71 /* held while logging the inode in tree-log.c */ 72 struct mutex log_mutex; 73 74 /* held while doing delalloc reservations */ 75 struct mutex delalloc_mutex; 76 77 /* used to order data wrt metadata */ 78 struct btrfs_ordered_inode_tree ordered_tree; 79 80 /* list of all the delalloc inodes in the FS. There are times we need 81 * to write all the delalloc pages to disk, and this list is used 82 * to walk them all. 83 */ 84 struct list_head delalloc_inodes; 85 86 /* 87 * list for tracking inodes that must be sent to disk before a 88 * rename or truncate commit 89 */ 90 struct list_head ordered_operations; 91 92 /* node for the red-black tree that links inodes in subvolume root */ 93 struct rb_node rb_node; 94 95 unsigned long runtime_flags; 96 97 /* Keep track of who's O_SYNC/fsyncing currently */ 98 atomic_t sync_writers; 99 100 /* full 64 bit generation number, struct vfs_inode doesn't have a big 101 * enough field for this. 102 */ 103 u64 generation; 104 105 /* 106 * transid of the trans_handle that last modified this inode 107 */ 108 u64 last_trans; 109 110 /* 111 * log transid when this inode was last modified 112 */ 113 u64 last_sub_trans; 114 115 /* 116 * transid that last logged this inode 117 */ 118 u64 logged_trans; 119 120 /* total number of bytes pending delalloc, used by stat to calc the 121 * real block usage of the file 122 */ 123 u64 delalloc_bytes; 124 125 /* 126 * the size of the file stored in the metadata on disk. data=ordered 127 * means the in-memory i_size might be larger than the size on disk 128 * because not all the blocks are written yet. 129 */ 130 u64 disk_i_size; 131 132 /* 133 * if this is a directory then index_cnt is the counter for the index 134 * number for new files that are created 135 */ 136 u64 index_cnt; 137 138 /* the fsync log has some corner cases that mean we have to check 139 * directories to see if any unlinks have been done before 140 * the directory was logged. See tree-log.c for all the 141 * details 142 */ 143 u64 last_unlink_trans; 144 145 /* 146 * Number of bytes outstanding that are going to need csums. This is 147 * used in ENOSPC accounting. 148 */ 149 u64 csum_bytes; 150 151 /* flags field from the on disk inode */ 152 u32 flags; 153 154 /* a local copy of root's last_log_commit */ 155 unsigned long last_log_commit; 156 157 /* 158 * Counters to keep track of the number of extent item's we may use due 159 * to delalloc and such. outstanding_extents is the number of extent 160 * items we think we'll end up using, and reserved_extents is the number 161 * of extent items we've reserved metadata for. 162 */ 163 unsigned outstanding_extents; 164 unsigned reserved_extents; 165 166 /* 167 * always compress this one file 168 */ 169 unsigned force_compress; 170 171 struct btrfs_delayed_node *delayed_node; 172 173 struct inode vfs_inode; 174 }; 175 176 extern unsigned char btrfs_filetype_table[]; 177 178 static inline struct btrfs_inode *BTRFS_I(struct inode *inode) 179 { 180 return container_of(inode, struct btrfs_inode, vfs_inode); 181 } 182 183 static inline unsigned long btrfs_inode_hash(u64 objectid, 184 const struct btrfs_root *root) 185 { 186 u64 h = objectid ^ (root->objectid * GOLDEN_RATIO_PRIME); 187 188 #if BITS_PER_LONG == 32 189 h = (h >> 32) ^ (h & 0xffffffff); 190 #endif 191 192 return (unsigned long)h; 193 } 194 195 static inline void btrfs_insert_inode_hash(struct inode *inode) 196 { 197 unsigned long h = btrfs_inode_hash(inode->i_ino, BTRFS_I(inode)->root); 198 199 __insert_inode_hash(inode, h); 200 } 201 202 static inline u64 btrfs_ino(struct inode *inode) 203 { 204 u64 ino = BTRFS_I(inode)->location.objectid; 205 206 /* 207 * !ino: btree_inode 208 * type == BTRFS_ROOT_ITEM_KEY: subvol dir 209 */ 210 if (!ino || BTRFS_I(inode)->location.type == BTRFS_ROOT_ITEM_KEY) 211 ino = inode->i_ino; 212 return ino; 213 } 214 215 static inline void btrfs_i_size_write(struct inode *inode, u64 size) 216 { 217 i_size_write(inode, size); 218 BTRFS_I(inode)->disk_i_size = size; 219 } 220 221 static inline bool btrfs_is_free_space_inode(struct inode *inode) 222 { 223 struct btrfs_root *root = BTRFS_I(inode)->root; 224 225 if (root == root->fs_info->tree_root && 226 btrfs_ino(inode) != BTRFS_BTREE_INODE_OBJECTID) 227 return true; 228 if (BTRFS_I(inode)->location.objectid == BTRFS_FREE_INO_OBJECTID) 229 return true; 230 return false; 231 } 232 233 static inline int btrfs_inode_in_log(struct inode *inode, u64 generation) 234 { 235 if (BTRFS_I(inode)->logged_trans == generation && 236 BTRFS_I(inode)->last_sub_trans <= 237 BTRFS_I(inode)->last_log_commit && 238 BTRFS_I(inode)->last_sub_trans <= 239 BTRFS_I(inode)->root->last_log_commit) 240 return 1; 241 return 0; 242 } 243 244 struct btrfs_dio_private { 245 struct inode *inode; 246 u64 logical_offset; 247 u64 disk_bytenr; 248 u64 bytes; 249 void *private; 250 251 /* number of bios pending for this dio */ 252 atomic_t pending_bios; 253 254 /* IO errors */ 255 int errors; 256 257 /* orig_bio is our btrfs_io_bio */ 258 struct bio *orig_bio; 259 260 /* dio_bio came from fs/direct-io.c */ 261 struct bio *dio_bio; 262 u8 csum[0]; 263 }; 264 265 /* 266 * Disable DIO read nolock optimization, so new dio readers will be forced 267 * to grab i_mutex. It is used to avoid the endless truncate due to 268 * nonlocked dio read. 269 */ 270 static inline void btrfs_inode_block_unlocked_dio(struct inode *inode) 271 { 272 set_bit(BTRFS_INODE_READDIO_NEED_LOCK, &BTRFS_I(inode)->runtime_flags); 273 smp_mb(); 274 } 275 276 static inline void btrfs_inode_resume_unlocked_dio(struct inode *inode) 277 { 278 smp_mb__before_clear_bit(); 279 clear_bit(BTRFS_INODE_READDIO_NEED_LOCK, 280 &BTRFS_I(inode)->runtime_flags); 281 } 282 283 #endif 284