linux/Documentation/filesystems/erofs.rst
<<
>>
Prefs
   1.. SPDX-License-Identifier: GPL-2.0
   2
   3======================================
   4Enhanced Read-Only File System - EROFS
   5======================================
   6
   7Overview
   8========
   9
  10EROFS file-system stands for Enhanced Read-Only File System. Different
  11from other read-only file systems, it aims to be designed for flexibility,
  12scalability, but be kept simple and high performance.
  13
  14It is designed as a better filesystem solution for the following scenarios:
  15
  16 - read-only storage media or
  17
  18 - part of a fully trusted read-only solution, which means it needs to be
  19   immutable and bit-for-bit identical to the official golden image for
  20   their releases due to security and other considerations and
  21
  22 - hope to minimize extra storage space with guaranteed end-to-end performance
  23   by using compact layout, transparent file compression and direct access,
  24   especially for those embedded devices with limited memory and high-density
  25   hosts with numerous containers;
  26
  27Here is the main features of EROFS:
  28
  29 - Little endian on-disk design;
  30
  31 - Currently 4KB block size (nobh) and therefore maximum 16TB address space;
  32
  33 - Metadata & data could be mixed by design;
  34
  35 - 2 inode versions for different requirements:
  36
  37   =====================  ============  =====================================
  38                          compact (v1)  extended (v2)
  39   =====================  ============  =====================================
  40   Inode metadata size    32 bytes      64 bytes
  41   Max file size          4 GB          16 EB (also limited by max. vol size)
  42   Max uids/gids          65536         4294967296
  43   File change time       no            yes (64 + 32-bit timestamp)
  44   Max hardlinks          65536         4294967296
  45   Metadata reserved      4 bytes       14 bytes
  46   =====================  ============  =====================================
  47
  48 - Support extended attributes (xattrs) as an option;
  49
  50 - Support xattr inline and tail-end data inline for all files;
  51
  52 - Support POSIX.1e ACLs by using xattrs;
  53
  54 - Support transparent data compression as an option:
  55   LZ4 algorithm with the fixed-sized output compression for high performance;
  56
  57 - Multiple device support for multi-layer container images.
  58
  59The following git tree provides the file system user-space tools under
  60development (ex, formatting tool mkfs.erofs):
  61
  62- git://git.kernel.org/pub/scm/linux/kernel/git/xiang/erofs-utils.git
  63
  64Bugs and patches are welcome, please kindly help us and send to the following
  65linux-erofs mailing list:
  66
  67- linux-erofs mailing list   <linux-erofs@lists.ozlabs.org>
  68
  69Mount options
  70=============
  71
  72===================    =========================================================
  73(no)user_xattr         Setup Extended User Attributes. Note: xattr is enabled
  74                       by default if CONFIG_EROFS_FS_XATTR is selected.
  75(no)acl                Setup POSIX Access Control List. Note: acl is enabled
  76                       by default if CONFIG_EROFS_FS_POSIX_ACL is selected.
  77cache_strategy=%s      Select a strategy for cached decompression from now on:
  78
  79                       ==========  =============================================
  80                         disabled  In-place I/O decompression only;
  81                        readahead  Cache the last incomplete compressed physical
  82                                   cluster for further reading. It still does
  83                                   in-place I/O decompression for the rest
  84                                   compressed physical clusters;
  85                       readaround  Cache the both ends of incomplete compressed
  86                                   physical clusters for further reading.
  87                                   It still does in-place I/O decompression
  88                                   for the rest compressed physical clusters.
  89                       ==========  =============================================
  90dax={always,never}     Use direct access (no page cache).  See
  91                       Documentation/filesystems/dax.rst.
  92dax                    A legacy option which is an alias for ``dax=always``.
  93device=%s              Specify a path to an extra device to be used together.
  94===================    =========================================================
  95
  96Sysfs Entries
  97=============
  98
  99Information about mounted erofs file systems can be found in /sys/fs/erofs.
 100Each mounted filesystem will have a directory in /sys/fs/erofs based on its
 101device name (i.e., /sys/fs/erofs/sda).
 102(see also Documentation/ABI/testing/sysfs-fs-erofs)
 103
 104On-disk details
 105===============
 106
 107Summary
 108-------
 109Different from other read-only file systems, an EROFS volume is designed
 110to be as simple as possible::
 111
 112                                |-> aligned with the block size
 113   ____________________________________________________________
 114  | |SB| | ... | Metadata | ... | Data | Metadata | ... | Data |
 115  |_|__|_|_____|__________|_____|______|__________|_____|______|
 116  0 +1K
 117
 118All data areas should be aligned with the block size, but metadata areas
 119may not. All metadatas can be now observed in two different spaces (views):
 120
 121 1. Inode metadata space
 122
 123    Each valid inode should be aligned with an inode slot, which is a fixed
 124    value (32 bytes) and designed to be kept in line with compact inode size.
 125
 126    Each inode can be directly found with the following formula:
 127         inode offset = meta_blkaddr * block_size + 32 * nid
 128
 129    ::
 130
 131                                 |-> aligned with 8B
 132                                            |-> followed closely
 133     + meta_blkaddr blocks                                      |-> another slot
 134       _____________________________________________________________________
 135     |  ...   | inode |  xattrs  | extents  | data inline | ... | inode ...
 136     |________|_______|(optional)|(optional)|__(optional)_|_____|__________
 137              |-> aligned with the inode slot size
 138                   .                   .
 139                 .                         .
 140               .                              .
 141             .                                    .
 142           .                                         .
 143         .                                              .
 144       .____________________________________________________|-> aligned with 4B
 145       | xattr_ibody_header | shared xattrs | inline xattrs |
 146       |____________________|_______________|_______________|
 147       |->    12 bytes    <-|->x * 4 bytes<-|               .
 148                           .                .                 .
 149                     .                      .                   .
 150                .                           .                     .
 151            ._______________________________.______________________.
 152            | id | id | id | id |  ... | id | ent | ... | ent| ... |
 153            |____|____|____|____|______|____|_____|_____|____|_____|
 154                                            |-> aligned with 4B
 155                                                        |-> aligned with 4B
 156
 157    Inode could be 32 or 64 bytes, which can be distinguished from a common
 158    field which all inode versions have -- i_format::
 159
 160        __________________               __________________
 161       |     i_format     |             |     i_format     |
 162       |__________________|             |__________________|
 163       |        ...       |             |        ...       |
 164       |                  |             |                  |
 165       |__________________| 32 bytes    |                  |
 166                                        |                  |
 167                                        |__________________| 64 bytes
 168
 169    Xattrs, extents, data inline are followed by the corresponding inode with
 170    proper alignment, and they could be optional for different data mappings.
 171    _currently_ total 5 data layouts are supported:
 172
 173    ==  ====================================================================
 174     0  flat file data without data inline (no extent);
 175     1  fixed-sized output data compression (with non-compacted indexes);
 176     2  flat file data with tail packing data inline (no extent);
 177     3  fixed-sized output data compression (with compacted indexes, v5.3+);
 178     4  chunk-based file (v5.15+).
 179    ==  ====================================================================
 180
 181    The size of the optional xattrs is indicated by i_xattr_count in inode
 182    header. Large xattrs or xattrs shared by many different files can be
 183    stored in shared xattrs metadata rather than inlined right after inode.
 184
 185 2. Shared xattrs metadata space
 186
 187    Shared xattrs space is similar to the above inode space, started with
 188    a specific block indicated by xattr_blkaddr, organized one by one with
 189    proper align.
 190
 191    Each share xattr can also be directly found by the following formula:
 192         xattr offset = xattr_blkaddr * block_size + 4 * xattr_id
 193
 194::
 195
 196                           |-> aligned by  4 bytes
 197    + xattr_blkaddr blocks                     |-> aligned with 4 bytes
 198     _________________________________________________________________________
 199    |  ...   | xattr_entry |  xattr data | ... |  xattr_entry | xattr data  ...
 200    |________|_____________|_____________|_____|______________|_______________
 201
 202Directories
 203-----------
 204All directories are now organized in a compact on-disk format. Note that
 205each directory block is divided into index and name areas in order to support
 206random file lookup, and all directory entries are _strictly_ recorded in
 207alphabetical order in order to support improved prefix binary search
 208algorithm (could refer to the related source code).
 209
 210::
 211
 212                  ___________________________
 213                 /                           |
 214                /              ______________|________________
 215               /              /              | nameoff1       | nameoffN-1
 216  ____________.______________._______________v________________v__________
 217 | dirent | dirent | ... | dirent | filename | filename | ... | filename |
 218 |___.0___|____1___|_____|___N-1__|____0_____|____1_____|_____|___N-1____|
 219      \                           ^
 220       \                          |                           * could have
 221        \                         |                             trailing '\0'
 222         \________________________| nameoff0
 223                             Directory block
 224
 225Note that apart from the offset of the first filename, nameoff0 also indicates
 226the total number of directory entries in this block since it is no need to
 227introduce another on-disk field at all.
 228
 229Chunk-based file
 230----------------
 231In order to support chunk-based data deduplication, a new inode data layout has
 232been supported since Linux v5.15: Files are split in equal-sized data chunks
 233with ``extents`` area of the inode metadata indicating how to get the chunk
 234data: these can be simply as a 4-byte block address array or in the 8-byte
 235chunk index form (see struct erofs_inode_chunk_index in erofs_fs.h for more
 236details.)
 237
 238By the way, chunk-based files are all uncompressed for now.
 239
 240Data compression
 241----------------
 242EROFS implements LZ4 fixed-sized output compression which generates fixed-sized
 243compressed data blocks from variable-sized input in contrast to other existing
 244fixed-sized input solutions. Relatively higher compression ratios can be gotten
 245by using fixed-sized output compression since nowadays popular data compression
 246algorithms are mostly LZ77-based and such fixed-sized output approach can be
 247benefited from the historical dictionary (aka. sliding window).
 248
 249In details, original (uncompressed) data is turned into several variable-sized
 250extents and in the meanwhile, compressed into physical clusters (pclusters).
 251In order to record each variable-sized extent, logical clusters (lclusters) are
 252introduced as the basic unit of compress indexes to indicate whether a new
 253extent is generated within the range (HEAD) or not (NONHEAD). Lclusters are now
 254fixed in block size, as illustrated below::
 255
 256          |<-    variable-sized extent    ->|<-       VLE         ->|
 257        clusterofs                        clusterofs              clusterofs
 258          |                                 |                       |
 259 _________v_________________________________v_______________________v________
 260 ... |    .         |              |        .     |              |  .   ...
 261 ____|____._________|______________|________.___ _|______________|__.________
 262     |-> lcluster <-|-> lcluster <-|-> lcluster <-|-> lcluster <-|
 263          (HEAD)        (NONHEAD)       (HEAD)        (NONHEAD)    .
 264           .             CBLKCNT            .                    .
 265            .                               .                  .
 266             .                              .                .
 267       _______._____________________________.______________._________________
 268          ... |              |              |              | ...
 269       _______|______________|______________|______________|_________________
 270              |->      big pcluster       <-|-> pcluster <-|
 271
 272A physical cluster can be seen as a container of physical compressed blocks
 273which contains compressed data. Previously, only lcluster-sized (4KB) pclusters
 274were supported. After big pcluster feature is introduced (available since
 275Linux v5.13), pcluster can be a multiple of lcluster size.
 276
 277For each HEAD lcluster, clusterofs is recorded to indicate where a new extent
 278starts and blkaddr is used to seek the compressed data. For each NONHEAD
 279lcluster, delta0 and delta1 are available instead of blkaddr to indicate the
 280distance to its HEAD lcluster and the next HEAD lcluster. A PLAIN lcluster is
 281also a HEAD lcluster except that its data is uncompressed. See the comments
 282around "struct z_erofs_vle_decompressed_index" in erofs_fs.h for more details.
 283
 284If big pcluster is enabled, pcluster size in lclusters needs to be recorded as
 285well. Let the delta0 of the first NONHEAD lcluster store the compressed block
 286count with a special flag as a new called CBLKCNT NONHEAD lcluster. It's easy
 287to understand its delta0 is constantly 1, as illustrated below::
 288
 289   __________________________________________________________
 290  | HEAD |  NONHEAD  | NONHEAD | ... | NONHEAD | HEAD | HEAD |
 291  |__:___|_(CBLKCNT)_|_________|_____|_________|__:___|____:_|
 292     |<----- a big pcluster (with CBLKCNT) ------>|<--  -->|
 293           a lcluster-sized pcluster (without CBLKCNT) ^
 294
 295If another HEAD follows a HEAD lcluster, there is no room to record CBLKCNT,
 296but it's easy to know the size of such pcluster is 1 lcluster as well.
 297