Based on kernel version 4.0. Page generated on 2015-04-14 21:24 EST.
1 2 The Second Extended Filesystem 3 ============================== 4 5 ext2 was originally released in January 1993. Written by R\'emy Card, 6 Theodore Ts'o and Stephen Tweedie, it was a major rewrite of the 7 Extended Filesystem. It is currently still (April 2001) the predominant 8 filesystem in use by Linux. There are also implementations available 9 for NetBSD, FreeBSD, the GNU HURD, Windows 95/98/NT, OS/2 and RISC OS. 10 11 Options 12 ======= 13 14 Most defaults are determined by the filesystem superblock, and can be 15 set using tune2fs(8). Kernel-determined defaults are indicated by (*). 16 17 bsddf (*) Makes `df' act like BSD. 18 minixdf Makes `df' act like Minix. 19 20 check=none, nocheck (*) Don't do extra checking of bitmaps on mount 21 (check=normal and check=strict options removed) 22 23 dax Use direct access (no page cache). See 24 Documentation/filesystems/dax.txt. 25 26 debug Extra debugging information is sent to the 27 kernel syslog. Useful for developers. 28 29 errors=continue Keep going on a filesystem error. 30 errors=remount-ro Remount the filesystem read-only on an error. 31 errors=panic Panic and halt the machine if an error occurs. 32 33 grpid, bsdgroups Give objects the same group ID as their parent. 34 nogrpid, sysvgroups New objects have the group ID of their creator. 35 36 nouid32 Use 16-bit UIDs and GIDs. 37 38 oldalloc Enable the old block allocator. Orlov should 39 have better performance, we'd like to get some 40 feedback if it's the contrary for you. 41 orlov (*) Use the Orlov block allocator. 42 (See http://lwn.net/Articles/14633/ and 43 http://lwn.net/Articles/14446/.) 44 45 resuid=n The user ID which may use the reserved blocks. 46 resgid=n The group ID which may use the reserved blocks. 47 48 sb=n Use alternate superblock at this location. 49 50 user_xattr Enable "user." POSIX Extended Attributes 51 (requires CONFIG_EXT2_FS_XATTR). 52 See also http://acl.bestbits.at 53 nouser_xattr Don't support "user." extended attributes. 54 55 acl Enable POSIX Access Control Lists support 56 (requires CONFIG_EXT2_FS_POSIX_ACL). 57 See also http://acl.bestbits.at 58 noacl Don't support POSIX ACLs. 59 60 nobh Do not attach buffer_heads to file pagecache. 61 62 grpquota,noquota,quota,usrquota Quota options are silently ignored by ext2. 63 64 65 Specification 66 ============= 67 68 ext2 shares many properties with traditional Unix filesystems. It has 69 the concepts of blocks, inodes and directories. It has space in the 70 specification for Access Control Lists (ACLs), fragments, undeletion and 71 compression though these are not yet implemented (some are available as 72 separate patches). There is also a versioning mechanism to allow new 73 features (such as journalling) to be added in a maximally compatible 74 manner. 75 76 Blocks 77 ------ 78 79 The space in the device or file is split up into blocks. These are 80 a fixed size, of 1024, 2048 or 4096 bytes (8192 bytes on Alpha systems), 81 which is decided when the filesystem is created. Smaller blocks mean 82 less wasted space per file, but require slightly more accounting overhead, 83 and also impose other limits on the size of files and the filesystem. 84 85 Block Groups 86 ------------ 87 88 Blocks are clustered into block groups in order to reduce fragmentation 89 and minimise the amount of head seeking when reading a large amount 90 of consecutive data. Information about each block group is kept in a 91 descriptor table stored in the block(s) immediately after the superblock. 92 Two blocks near the start of each group are reserved for the block usage 93 bitmap and the inode usage bitmap which show which blocks and inodes 94 are in use. Since each bitmap is limited to a single block, this means 95 that the maximum size of a block group is 8 times the size of a block. 96 97 The block(s) following the bitmaps in each block group are designated 98 as the inode table for that block group and the remainder are the data 99 blocks. The block allocation algorithm attempts to allocate data blocks 100 in the same block group as the inode which contains them. 101 102 The Superblock 103 -------------- 104 105 The superblock contains all the information about the configuration of 106 the filing system. The primary copy of the superblock is stored at an 107 offset of 1024 bytes from the start of the device, and it is essential 108 to mounting the filesystem. Since it is so important, backup copies of 109 the superblock are stored in block groups throughout the filesystem. 110 The first version of ext2 (revision 0) stores a copy at the start of 111 every block group, along with backups of the group descriptor block(s). 112 Because this can consume a considerable amount of space for large 113 filesystems, later revisions can optionally reduce the number of backup 114 copies by only putting backups in specific groups (this is the sparse 115 superblock feature). The groups chosen are 0, 1 and powers of 3, 5 and 7. 116 117 The information in the superblock contains fields such as the total 118 number of inodes and blocks in the filesystem and how many are free, 119 how many inodes and blocks are in each block group, when the filesystem 120 was mounted (and if it was cleanly unmounted), when it was modified, 121 what version of the filesystem it is (see the Revisions section below) 122 and which OS created it. 123 124 If the filesystem is revision 1 or higher, then there are extra fields, 125 such as a volume name, a unique identification number, the inode size, 126 and space for optional filesystem features to store configuration info. 127 128 All fields in the superblock (as in all other ext2 structures) are stored 129 on the disc in little endian format, so a filesystem is portable between 130 machines without having to know what machine it was created on. 131 132 Inodes 133 ------ 134 135 The inode (index node) is a fundamental concept in the ext2 filesystem. 136 Each object in the filesystem is represented by an inode. The inode 137 structure contains pointers to the filesystem blocks which contain the 138 data held in the object and all of the metadata about an object except 139 its name. The metadata about an object includes the permissions, owner, 140 group, flags, size, number of blocks used, access time, change time, 141 modification time, deletion time, number of links, fragments, version 142 (for NFS) and extended attributes (EAs) and/or Access Control Lists (ACLs). 143 144 There are some reserved fields which are currently unused in the inode 145 structure and several which are overloaded. One field is reserved for the 146 directory ACL if the inode is a directory and alternately for the top 32 147 bits of the file size if the inode is a regular file (allowing file sizes 148 larger than 2GB). The translator field is unused under Linux, but is used 149 by the HURD to reference the inode of a program which will be used to 150 interpret this object. Most of the remaining reserved fields have been 151 used up for both Linux and the HURD for larger owner and group fields, 152 The HURD also has a larger mode field so it uses another of the remaining 153 fields to store the extra more bits. 154 155 There are pointers to the first 12 blocks which contain the file's data 156 in the inode. There is a pointer to an indirect block (which contains 157 pointers to the next set of blocks), a pointer to a doubly-indirect 158 block (which contains pointers to indirect blocks) and a pointer to a 159 trebly-indirect block (which contains pointers to doubly-indirect blocks). 160 161 The flags field contains some ext2-specific flags which aren't catered 162 for by the standard chmod flags. These flags can be listed with lsattr 163 and changed with the chattr command, and allow specific filesystem 164 behaviour on a per-file basis. There are flags for secure deletion, 165 undeletable, compression, synchronous updates, immutability, append-only, 166 dumpable, no-atime, indexed directories, and data-journaling. Not all 167 of these are supported yet. 168 169 Directories 170 ----------- 171 172 A directory is a filesystem object and has an inode just like a file. 173 It is a specially formatted file containing records which associate 174 each name with an inode number. Later revisions of the filesystem also 175 encode the type of the object (file, directory, symlink, device, fifo, 176 socket) to avoid the need to check the inode itself for this information 177 (support for taking advantage of this feature does not yet exist in 178 Glibc 2.2). 179 180 The inode allocation code tries to assign inodes which are in the same 181 block group as the directory in which they are first created. 182 183 The current implementation of ext2 uses a singly-linked list to store 184 the filenames in the directory; a pending enhancement uses hashing of the 185 filenames to allow lookup without the need to scan the entire directory. 186 187 The current implementation never removes empty directory blocks once they 188 have been allocated to hold more files. 189 190 Special files 191 ------------- 192 193 Symbolic links are also filesystem objects with inodes. They deserve 194 special mention because the data for them is stored within the inode 195 itself if the symlink is less than 60 bytes long. It uses the fields 196 which would normally be used to store the pointers to data blocks. 197 This is a worthwhile optimisation as it we avoid allocating a full 198 block for the symlink, and most symlinks are less than 60 characters long. 199 200 Character and block special devices never have data blocks assigned to 201 them. Instead, their device number is stored in the inode, again reusing 202 the fields which would be used to point to the data blocks. 203 204 Reserved Space 205 -------------- 206 207 In ext2, there is a mechanism for reserving a certain number of blocks 208 for a particular user (normally the super-user). This is intended to 209 allow for the system to continue functioning even if non-privileged users 210 fill up all the space available to them (this is independent of filesystem 211 quotas). It also keeps the filesystem from filling up entirely which 212 helps combat fragmentation. 213 214 Filesystem check 215 ---------------- 216 217 At boot time, most systems run a consistency check (e2fsck) on their 218 filesystems. The superblock of the ext2 filesystem contains several 219 fields which indicate whether fsck should actually run (since checking 220 the filesystem at boot can take a long time if it is large). fsck will 221 run if the filesystem was not cleanly unmounted, if the maximum mount 222 count has been exceeded or if the maximum time between checks has been 223 exceeded. 224 225 Feature Compatibility 226 --------------------- 227 228 The compatibility feature mechanism used in ext2 is sophisticated. 229 It safely allows features to be added to the filesystem, without 230 unnecessarily sacrificing compatibility with older versions of the 231 filesystem code. The feature compatibility mechanism is not supported by 232 the original revision 0 (EXT2_GOOD_OLD_REV) of ext2, but was introduced in 233 revision 1. There are three 32-bit fields, one for compatible features 234 (COMPAT), one for read-only compatible (RO_COMPAT) features and one for 235 incompatible (INCOMPAT) features. 236 237 These feature flags have specific meanings for the kernel as follows: 238 239 A COMPAT flag indicates that a feature is present in the filesystem, 240 but the on-disk format is 100% compatible with older on-disk formats, so 241 a kernel which didn't know anything about this feature could read/write 242 the filesystem without any chance of corrupting the filesystem (or even 243 making it inconsistent). This is essentially just a flag which says 244 "this filesystem has a (hidden) feature" that the kernel or e2fsck may 245 want to be aware of (more on e2fsck and feature flags later). The ext3 246 HAS_JOURNAL feature is a COMPAT flag because the ext3 journal is simply 247 a regular file with data blocks in it so the kernel does not need to 248 take any special notice of it if it doesn't understand ext3 journaling. 249 250 An RO_COMPAT flag indicates that the on-disk format is 100% compatible 251 with older on-disk formats for reading (i.e. the feature does not change 252 the visible on-disk format). However, an old kernel writing to such a 253 filesystem would/could corrupt the filesystem, so this is prevented. The 254 most common such feature, SPARSE_SUPER, is an RO_COMPAT feature because 255 sparse groups allow file data blocks where superblock/group descriptor 256 backups used to live, and ext2_free_blocks() refuses to free these blocks, 257 which would leading to inconsistent bitmaps. An old kernel would also 258 get an error if it tried to free a series of blocks which crossed a group 259 boundary, but this is a legitimate layout in a SPARSE_SUPER filesystem. 260 261 An INCOMPAT flag indicates the on-disk format has changed in some 262 way that makes it unreadable by older kernels, or would otherwise 263 cause a problem if an old kernel tried to mount it. FILETYPE is an 264 INCOMPAT flag because older kernels would think a filename was longer 265 than 256 characters, which would lead to corrupt directory listings. 266 The COMPRESSION flag is an obvious INCOMPAT flag - if the kernel 267 doesn't understand compression, you would just get garbage back from 268 read() instead of it automatically decompressing your data. The ext3 269 RECOVER flag is needed to prevent a kernel which does not understand the 270 ext3 journal from mounting the filesystem without replaying the journal. 271 272 For e2fsck, it needs to be more strict with the handling of these 273 flags than the kernel. If it doesn't understand ANY of the COMPAT, 274 RO_COMPAT, or INCOMPAT flags it will refuse to check the filesystem, 275 because it has no way of verifying whether a given feature is valid 276 or not. Allowing e2fsck to succeed on a filesystem with an unknown 277 feature is a false sense of security for the user. Refusing to check 278 a filesystem with unknown features is a good incentive for the user to 279 update to the latest e2fsck. This also means that anyone adding feature 280 flags to ext2 also needs to update e2fsck to verify these features. 281 282 Metadata 283 -------- 284 285 It is frequently claimed that the ext2 implementation of writing 286 asynchronous metadata is faster than the ffs synchronous metadata 287 scheme but less reliable. Both methods are equally resolvable by their 288 respective fsck programs. 289 290 If you're exceptionally paranoid, there are 3 ways of making metadata 291 writes synchronous on ext2: 292 293 per-file if you have the program source: use the O_SYNC flag to open() 294 per-file if you don't have the source: use "chattr +S" on the file 295 per-filesystem: add the "sync" option to mount (or in /etc/fstab) 296 297 the first and last are not ext2 specific but do force the metadata to 298 be written synchronously. See also Journaling below. 299 300 Limitations 301 ----------- 302 303 There are various limits imposed by the on-disk layout of ext2. Other 304 limits are imposed by the current implementation of the kernel code. 305 Many of the limits are determined at the time the filesystem is first 306 created, and depend upon the block size chosen. The ratio of inodes to 307 data blocks is fixed at filesystem creation time, so the only way to 308 increase the number of inodes is to increase the size of the filesystem. 309 No tools currently exist which can change the ratio of inodes to blocks. 310 311 Most of these limits could be overcome with slight changes in the on-disk 312 format and using a compatibility flag to signal the format change (at 313 the expense of some compatibility). 314 315 Filesystem block size: 1kB 2kB 4kB 8kB 316 317 File size limit: 16GB 256GB 2048GB 2048GB 318 Filesystem size limit: 2047GB 8192GB 16384GB 32768GB 319 320 There is a 2.4 kernel limit of 2048GB for a single block device, so no 321 filesystem larger than that can be created at this time. There is also 322 an upper limit on the block size imposed by the page size of the kernel, 323 so 8kB blocks are only allowed on Alpha systems (and other architectures 324 which support larger pages). 325 326 There is an upper limit of 32000 subdirectories in a single directory. 327 328 There is a "soft" upper limit of about 10-15k files in a single directory 329 with the current linear linked-list directory implementation. This limit 330 stems from performance problems when creating and deleting (and also 331 finding) files in such large directories. Using a hashed directory index 332 (under development) allows 100k-1M+ files in a single directory without 333 performance problems (although RAM size becomes an issue at this point). 334 335 The (meaningless) absolute upper limit of files in a single directory 336 (imposed by the file size, the realistic limit is obviously much less) 337 is over 130 trillion files. It would be higher except there are not 338 enough 4-character names to make up unique directory entries, so they 339 have to be 8 character filenames, even then we are fairly close to 340 running out of unique filenames. 341 342 Journaling 343 ---------- 344 345 A journaling extension to the ext2 code has been developed by Stephen 346 Tweedie. It avoids the risks of metadata corruption and the need to 347 wait for e2fsck to complete after a crash, without requiring a change 348 to the on-disk ext2 layout. In a nutshell, the journal is a regular 349 file which stores whole metadata (and optionally data) blocks that have 350 been modified, prior to writing them into the filesystem. This means 351 it is possible to add a journal to an existing ext2 filesystem without 352 the need for data conversion. 353 354 When changes to the filesystem (e.g. a file is renamed) they are stored in 355 a transaction in the journal and can either be complete or incomplete at 356 the time of a crash. If a transaction is complete at the time of a crash 357 (or in the normal case where the system does not crash), then any blocks 358 in that transaction are guaranteed to represent a valid filesystem state, 359 and are copied into the filesystem. If a transaction is incomplete at 360 the time of the crash, then there is no guarantee of consistency for 361 the blocks in that transaction so they are discarded (which means any 362 filesystem changes they represent are also lost). 363 Check Documentation/filesystems/ext3.txt if you want to read more about 364 ext3 and journaling. 365 366 References 367 ========== 368 369 The kernel source file:/usr/src/linux/fs/ext2/ 370 e2fsprogs (e2fsck) http://e2fsprogs.sourceforge.net/ 371 Design & Implementation http://e2fsprogs.sourceforge.net/ext2intro.html 372 Journaling (ext3) ftp://ftp.uk.linux.org/pub/linux/sct/fs/jfs/ 373 Filesystem Resizing http://ext2resize.sourceforge.net/ 374 Compression (*) http://e2compr.sourceforge.net/ 375 376 Implementations for: 377 Windows 95/98/NT/2000 http://www.chrysocome.net/explore2fs 378 Windows 95 (*) http://www.yipton.net/content.html#FSDEXT2 379 DOS client (*) ftp://metalab.unc.edu/pub/Linux/system/filesystems/ext2/ 380 OS/2 (+) ftp://metalab.unc.edu/pub/Linux/system/filesystems/ext2/ 381 RISC OS client http://www.esw-heim.tu-clausthal.de/~marco/smorbrod/IscaFS/ 382 383 (*) no longer actively developed/supported (as of Apr 2001) 384 (+) no longer actively developed/supported (as of Mar 2009)