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 "extent_map.h" 23 #include "extent_io.h" 24 #include "ordered-data.h" 25 #include "delayed-inode.h" 26 27 /* 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 39 #define BTRFS_INODE_HAS_ORPHAN_ITEM 5 40 #define BTRFS_INODE_HAS_ASYNC_EXTENT 6 41 #define BTRFS_INODE_NEEDS_FULL_SYNC 7 42 #define BTRFS_INODE_COPY_EVERYTHING 8 43 #define BTRFS_INODE_IN_DELALLOC_LIST 9 44 #define BTRFS_INODE_READDIO_NEED_LOCK 10 45 46 /* in memory btrfs inode */ 47 struct btrfs_inode { 48 /* which subvolume this inode belongs to */ 49 struct btrfs_root *root; 50 51 /* key used to find this inode on disk. This is used by the code 52 * to read in roots of subvolumes 53 */ 54 struct btrfs_key location; 55 56 /* Lock for counters */ 57 spinlock_t lock; 58 59 /* the extent_tree has caches of all the extent mappings to disk */ 60 struct extent_map_tree extent_tree; 61 62 /* the io_tree does range state (DIRTY, LOCKED etc) */ 63 struct extent_io_tree io_tree; 64 65 /* special utility tree used to record which mirrors have already been 66 * tried when checksums fail for a given block 67 */ 68 struct extent_io_tree io_failure_tree; 69 70 /* held while logging the inode in tree-log.c */ 71 struct mutex log_mutex; 72 73 /* held while doing delalloc reservations */ 74 struct mutex delalloc_mutex; 75 76 /* used to order data wrt metadata */ 77 struct btrfs_ordered_inode_tree ordered_tree; 78 79 /* 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 */ 83 struct list_head delalloc_inodes; 84 85 /* 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 91 /* node for the red-black tree that links inodes in subvolume root */ 92 struct rb_node rb_node; 93 94 unsigned long runtime_flags; 95 96 /* Keep track of who's O_SYNC/fsyncing currently */ 97 atomic_t sync_writers; 98 99 /* full 64 bit generation number, struct vfs_inode doesn't have a big 100 * enough field for this. 101 */ 102 u64 generation; 103 104 /* 105 * transid of the trans_handle that last modified this inode 106 */ 107 u64 last_trans; 108 109 /* 110 * log transid when this inode was last modified 111 */ 112 u64 last_sub_trans; 113 114 /* 115 * transid that last logged this inode 116 */ 117 u64 logged_trans; 118 119 /* total number of bytes pending delalloc, used by stat to calc the 120 * real block usage of the file 121 */ 122 u64 delalloc_bytes; 123 124 /* 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 */ 129 u64 disk_i_size; 130 131 /* 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; 136 137 /* 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; 143 144 /* 145 * Number of bytes outstanding that are going to need csums. This is 146 * used in ENOSPC accounting. 147 */ 148 u64 csum_bytes; 149 150 /* flags field from the on disk inode */ 151 u32 flags; 152 153 /* a local copy of root's last_log_commit */ 154 unsigned long last_log_commit; 155 156 /* 157 * 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. 161 */ 162 unsigned outstanding_extents; 163 unsigned reserved_extents; 164 165 /* 166 * always compress this one file 167 */ 168 unsigned force_compress; 169 170 struct btrfs_delayed_node *delayed_node; 171 172 struct inode vfs_inode; 173 }; 174 175 extern unsigned char btrfs_filetype_table[]; 176 177 static inline struct btrfs_inode *BTRFS_I(struct inode *inode) 178 { 179 return container_of(inode, struct btrfs_inode, vfs_inode); 180 } 181 182 static inline u64 btrfs_ino(struct inode *inode) 183 { 184 u64 ino = BTRFS_I(inode)->location.objectid; 185 186 /* 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) 191 ino = inode->i_ino; 192 return ino; 193 } 194 195 static inline void btrfs_i_size_write(struct inode *inode, u64 size) 196 { 197 i_size_write(inode, size); 198 BTRFS_I(inode)->disk_i_size = size; 199 } 200 201 static inline bool btrfs_is_free_space_inode(struct inode *inode) 202 { 203 struct btrfs_root *root = BTRFS_I(inode)->root; 204 205 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) 209 return true; 210 return false; 211 } 212 213 static inline int btrfs_inode_in_log(struct inode *inode, u64 generation) 214 { 215 if (BTRFS_I(inode)->logged_trans == generation && 216 BTRFS_I(inode)->last_sub_trans <= 217 BTRFS_I(inode)->last_log_commit && 218 BTRFS_I(inode)->last_sub_trans <= 219 BTRFS_I(inode)->root->last_log_commit) 220 return 1; 221 return 0; 222 } 223 224 struct btrfs_dio_private { 225 struct inode *inode; 226 u64 logical_offset; 227 u64 disk_bytenr; 228 u64 bytes; 229 void *private; 230 231 /* number of bios pending for this dio */ 232 atomic_t pending_bios; 233 234 /* IO errors */ 235 int errors; 236 237 /* orig_bio is our btrfs_io_bio */ 238 struct bio *orig_bio; 239 240 /* dio_bio came from fs/direct-io.c */ 241 struct bio *dio_bio; 242 u8 csum[0]; 243 }; 244 245 /* 246 * Disable DIO read nolock optimization, so new dio readers will be forced 247 * to grab i_mutex. It is used to avoid the endless truncate due to 248 * nonlocked dio read. 249 */ 250 static inline void btrfs_inode_block_unlocked_dio(struct inode *inode) 251 { 252 set_bit(BTRFS_INODE_READDIO_NEED_LOCK, &BTRFS_I(inode)->runtime_flags); 253 smp_mb(); 254 } 255 256 static inline void btrfs_inode_resume_unlocked_dio(struct inode *inode) 257 { 258 smp_mb__before_clear_bit(); 259 clear_bit(BTRFS_INODE_READDIO_NEED_LOCK, 260 &BTRFS_I(inode)->runtime_flags); 261 } 262 263 #endif 264